本篇博文主要内容为 2026-09-15 从Arxiv.org论文网站获取的最新论文列表，自动更新，按照NLP、CV、ML、AI、IR、MA六个大方向区分。
说明：每日论文数据从Arxiv.org获取，每天早上12:30左右定时自动更新。
提示: 当天未及时更新，有可能是Arxiv当日未有新的论文发布，也有可能是脚本出错。尽可能会在当天修复。
目录
概览 (2026-09-15)
今日共更新1477篇论文,其中:
- 自然语言处理共205篇(Computation and Language (cs.CL))
- 人工智能共409篇(Artificial Intelligence (cs.AI))
- 计算机视觉共234篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共368篇(Machine Learning (cs.LG))
- 多智能体系统共25篇(Multiagent Systems (cs.MA))
- 信息检索共36篇(Information Retrieval (cs.IR))
- 人机交互共80篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] HypoEvolve: Genetic Algorithms Enable Multi-Agent LLM s to Discover Scientific Hypotheses
【速读】：该论文旨在解决科学发现中不同形式的智能体协作对假设质量影响机制不明确的问题，尤其关注如何在保持智能体科学能力独立性的同时，量化协作模式对生成假设有效性的影响。其核心挑战在于分离智能体的科学推理能力与协作策略之间的耦合效应。为此，论文提出一种名为HypoEvolve的框架，其关键创新在于通过代际遗传算法显式建模多智能体协作过程：将多个具备特定功能（如机制论证、假设重构、证据评估与可检验性判断）的大型语言模型（LLM）智能体组织为协同进化群体，每一代通过批判、比较与修订机制更新假设种群，从而实现对协作动态的精确控制与可测度化。该框架支持对科学判断与新假设如何重塑种群结构的系统性追踪，使协作效果可直接评估。研究以药物重定位为应用场景，构建了基于DepMap和Open Targets的双维度外部验证体系，涵盖实验、遗传与临床证据，验证了HypoEvolve在34种癌症类型上显著优于六种基线方法的表现，其中DepMap选择性达0.171（基线最高为0.115），且跨癌症类型的泛化性能良好。该工作推动了自主科学范式的实现，展示了由AI研究团队协同完成超越单个模型能力的发现潜力。
链接: https://arxiv.org/abs/2609.15938
作者: Jieyuan Liu,Mengzhou Hu,Jefferson Chen,JungHo Kong,Pratibha Jagannatha,Yiming Gao,Dexter Pratt,Hsin-Yuan Lee,Zhiting Hu,Trey Ideker,Wei Wang,Eric P. Xing,Zhen Wang
机构: University of California San Diego (加州大学圣地亚哥分校); Texas A&M University (德克萨斯农工大学); Carnegie Mellon University (卡内基梅隆大学); Mohamed bin Zayed University of Artificial Intelligence (穆罕默德·本·扎耶德人工智能大学)
类目: Computation and Language (cs.CL); Computational Engineering, Finance, and Science (cs.CE); Multiagent Systems (cs.MA); Neural and Evolutionary Computing (cs.NE)
备注: 22 pages, 8 figures, 5 tables
Abstract:Scientific agents contribute to hypothesis discovery by synthesizing evidence, assessing proposals, and developing new explanations. Recent systems combine scientific agents with evolutionary search through critique, comparison, and revision. However, how different forms of agent collaboration affect hypothesis quality remains an open question. Answering this question requires separating the effects of agents’ scientific capabilities from those of their collaboration. A framework must therefore preserve agents’ scientific roles and support rules for combining, revising, and retaining hypotheses. Building on this view, we introduce HypoEvolve, which makes collaboration explicit through successive updates to a hypothesis population. Specifically, we propose a generational genetic algorithm to coordinate specialized large language model (LLM) agents that integrate mechanistic arguments, reconsider assumptions, and assess evidence and testability. Each generation specifies how scientific judgments and new proposals reshape the population, making collaboration effects on hypothesis quality directly testable. Moreover, we design our evaluation around scientifically meaningful hypotheses that explain how a proposed intervention could work. Drug repurposing links these explanations to target-level biological claims assessed against external evidence. Specifically, we adapt DepMap and Open Targets into complementary external measures grounded in experimental, genetic, and clinical evidence. Across 34 cancer types, HypoEvolve achieves the highest scores against six baselines on both measures. DepMap selectivity reaches 0.171, versus 0.115 for the strongest baseline. Gains over single-pass generation also generalize to held-out cancer types. HypoEvolve advances a vision of autonomous science in which AI research teams achieve a capacity for discovery beyond that of individual models.
[MA-1] Robust and Efficient Communication for Multi-Agent Learning
【速读】：该论文旨在解决多智能体强化学习（MARL）中通信有效性与物理约束之间难以平衡的问题，即如何在有限通信资源和不可靠信道条件下，确保智能体间生成的消息既具备高信息量又具有鲁棒性。其解决方案的关键在于提出一种基于信息论原理的新型框架——多智能体正则化通信（MARC），该框架通过注意力机制结合独特的消息正则化机制，以最小化对未来系统状态的不确定性为目标，从而引导智能体学习出高度表征性的通信协议。实验在严苛的通信瓶颈和丢包信道下进行，验证了MARC在复杂协作任务中显著优于现有先进方法，并通过深入分析表明其在强数据压缩条件下仍能保持优异性能，为资源受限环境下智能体系统的可扩展部署提供了有效路径。
链接: https://arxiv.org/abs/2609.15361
作者: Rafael Pina,Varuna De Silva,Corentin Artaud
机构: Institute for Digital Technologies; Loughborough University London (拉夫堡大学伦敦校区)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 12 pages, 11 figures
Abstract:Effective communication is a cornerstone of distributed intelligence in Multi-Agent Reinforcement Learning (MARL), yet ensuring that generated messages are both informative and robust to physical constraints remains a significant challenge. This paper introduces Multi-Agent Regularized Communication (MARC), a novel framework inspired by information-theoretic principles of conditional mutual information. MARC employs an attention-based architecture coupled with a unique message regularization mechanism designed to minimize uncertainty regarding future system states, thereby inducing the learning of highly representative communication protocols. Crucially, we evaluate MARC under stringent communication bottlenecks and lossy channels, simulating the real-world constraints of autonomous robotic networks and decentralized systems. Our results demonstrate that MARC significantly outperforms state-of-the-art methods in complex cooperative domains. Furthermore, we provide a deep analysis of message characteristics, proving that MARC maintains high operational performance even under significant data compression, offering a scalable path for deploying intelligent agents in resource-constrained environments.
[MA-2] he Universe of Universes: Benefit Yield Functions Implosion Thresholds and Infrastructure-Aware Optimization in Multi-LLM Systems
【速读】：该论文旨在解决大规模语言模型（LLM）集成系统中因盲目增加模型数量而导致性能下降的“模型坍缩”问题，其核心挑战在于缺乏对集成规模（ensemble size N）与整体性能之间关系的系统性建模。解决方案的关键在于提出“收益产出函数”（Benefit Yield Function, BYF），即每增加一个模型所带来的边际性能增益，并首次形式化定义了“坍缩阈值”（implosion threshold \theta*），即当BYF降至零时的集成规模，标志着整体性能开始劣化。该框架不仅揭示了认知遗传漂移（Epistemic Hereditary Drift, EHD）在多模型生态系统中的累积效应，还将AI生成速度作为\theta*的协变量，从而为国防部（DoD）多模型AI采购政策及人工智能系统测试科学提供了理论基础与实践指导。
链接: https://arxiv.org/abs/2609.15314
作者: Danielle Franklin,Vasu Raj Jain
机构: 未知
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:We introduce the Universe of Universes (UoU) framework, which treats the full ecosystem of major large language models (LLMs) as a structured retrieval corpus and proposes a compositional Automated Reasoning (AR) and Machine Learning (ML) architecture for cross-model retrieval-augmented generation. The central contribution is the formal characterization of the Benefit Yield Function (BYF), the marginal performance gain per additional model added to an ensemble, and the identification of the implosion threshold \theta*: the ensemble size at which BYF crosses zero and aggregate performance begins to degrade. Existing LLM ensemble and mixture-of-agents systems treat models as responders and aggregate outputs, but do not study performance as a function of ensemble size N across the full model universe. Benchmark research confirms performance plateaus at the individual model level; model collapse literature establishes that iterative training on AI-generated outputs degrades individual model distributions. Neither body of work formalizes the ensemble-level implosion threshold, models Epistemic Hereditary Drift (EHD) at the ecosystem level, or treats AI manufacturing velocity as a co-variable of \theta*. The framework has direct implications for DoD multi-model AI acquisition policy and the emerging science of testing AI-enabled systems.
[MA-3] ranslating the Translator: Decomposing the Cost of English-Forced Inter-Agent Communication EMNLP2026
【速读】：该论文旨在解决多智能体大语言模型（Multi-agent LLM）架构在非英语任务中普遍采用英语作为内部通信语言所导致的性能下降问题，尤其是在语言类型差异较大的场景下。其核心问题是：强制通过英语进行跨智能体通信是否会导致显著的语言转换损耗，从而影响最终任务的准确性。解决方案的关键在于对比两种不同通信范式——原生语言管道（native-language pipeline）与强制英语路由（English-forced pipeline，包含最终回译步骤）——在四种语系差异显著的语言（印地语、中文、西班牙语、阿拉伯语）上的表现。研究发现，在强制英语路由条件下存在显著的“英语强制税”（English-Forcing Tax），即相比原生语言执行，精确匹配（Exact Match）准确率下降13.0至30.6个百分点，且该结果在严格校正多重比较后仍具统计显著性。进一步分析表明，较低的chrF分数（反映英语参考词汇重叠度）与系统失败高度相关，证实翻译损失是性能下降的重要因素。因此，该研究提出在源语言与目标语言语系差异较大时，应优先采用原生语言路由以规避累积性的翻译成本。
链接: https://arxiv.org/abs/2609.15079
作者: Kushagra Agrawal,Yuming Feng,Man-Fai Leung
机构: Åbo Akademi University (Åbo Akademi 大学); Chongqing Sanxia University of Science and Technology (重庆三峡科技学院); Anglia Ruskin University (安格利亚鲁斯金大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: Accepted at The 2nd Workshop for Research on Agent Language Models at EMNLP 2026
Abstract:Multi-agent LLM architectures, such as LangChain and AutoGen, largely assume English as the lingua franca for internal inter-agent communication, even when the end-user task is non-English. We fill this gap by evaluating a two-agent extraction-answer core, with an additional back-translation agent in the English-forced condition, across four typologically diverse languages (Hindi, Chinese, Spanish, Arabic; n = 300 per language) using the Aya-23-8B model. We compare a native-language pipeline to an English-forced one (which incorporates a final back-translation step from English to the user’s language). We discover a statistically significant English-Forcing Tax (surviving a strict Bonferroni correction) that isolates the cost of English routing from general multi-agent orchestration overhead. Forcing inter-agent communication through English reduces Exact Match accuracy by 13.0 percentage points (Spanish) up to 30.6 percentage points (Hindi) compared to native-language multi-agent execution. Using chrF scores as a diagnostic measure of English-reference lexical overlap, we find that lower overlap is strongly associated with pipeline failure, consistent with translation loss being an important contributor to the observed performance drop. These findings suggest a compelling case for native-language routing in agent frameworks when the source and target languages are typologically distant, reducing a compounding translation tax.
[MA-4] High-Probability Nash Regret for Decentralized Learning in Markov α-Potential Games: Episodic and Fully Online Asynchronous Algorithms with Applications to Markov Congestion Games
【速读】：该论文旨在解决无限时域折扣马尔可夫博弈（Markov games）中，在仅基于贝叶斯反馈（bandit feedback）条件下，去中心化学习纳什均衡（Nash equilibrium, NE）的挑战，尤其聚焦于马尔可夫α-势博弈（Markov α-potential games）。其核心问题在于：在完全在线、异步更新且无全局协调的环境下，如何设计可扩展、具有有限时间高概率收敛保证的去中心化学习算法，并克服传统方法中因分布不匹配（distribution-mismatch）系数随状态空间规模指数增长而导致的性能瓶颈。解决方案的关键在于提出一种基于KL投影的自然策略梯度（KL-projected natural policy gradient, NPG）算法框架，分别适用于分段式（episodic）和全在线（fully online）两种设置。该框架通过引入状态级势结构（state-wise potential structure），实现了对势函数近似误差α的加性依赖，从而获得更紧致的纳什均衡后悔率（NE regret）界——分别为O(T−1/4)和O(T−2/15)，且彻底消除了依赖于状态空间大小的分布不匹配系数，同时兼容固定估计-预言机偏差（estimation-oracle bias）与转移敏感性（transition sensitivity）。进一步地，作者将该框架应用于独立资源马尔可夫拥堵博弈（IMCGs），证明了其近似势性与转移敏感性，并构建了基于实际成本观测的去中心化估计预言机，最终实现对随机机器上战略在线作业调度问题的可扩展去中心化稳定调度策略学习。总体而言，本工作首次为马尔可夫α-势博弈中的全在线异步去中心化学习提供了有限时间高概率纳什均衡后悔率保证，突破了传统方法的理论限制。
链接: https://arxiv.org/abs/2609.14959
作者: S. Rasoul Etesami
机构: University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校)
类目: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT); Multiagent Systems (cs.MA); Systems and Control (eess.SY); Optimization and Control (math.OC)
备注:
Abstract:We study decentralized learning of Nash equilibria (NE) in infinite-horizon discounted Markov games under bandit feedback, focusing on Markov \alpha -potential games. We develop KL-projected natural policy gradient (NPG) algorithms in two settings: an episodic setting with frozen policies during sampling and a fully online setting in which players receive a single realized cost sample per time step and update their policies asynchronously along a continuing trajectory. We establish finite-time high-probability NE regret bounds of order \widetilde O(T^-1/4) and \widetilde O(T^-2/15) for the episodic and fully online settings, respectively, up to fixed approximation terms. Crucially, our bounds eliminate the distribution-mismatch coefficient, which can scale prohibitively with the size of the state space, while accommodating potential approximation, estimation-oracle bias, and transition sensitivity. We further identify a state-wise potential structure that yields sharper guarantees with additive dependence on the potential approximation error \alpha . We specialize the framework to independent-resource Markov congestion games (IMCGs), establish their approximate-potential and transition-sensitivity properties, and construct decentralized estimation oracles from realized costs. As an application, we introduce strategic online job scheduling on stochastic machines and obtain a scalable decentralized algorithm for learning stable dispatching policies. Overall, our results provide the first finite-time high-probability NE regret guarantees for fully online asynchronous decentralized learning in Markov \alpha -potential games, remove distribution-mismatch coefficients from the regret bounds, accommodate fixed estimation-oracle bias, and provide scalable decentralized learning with finite-time guarantees for IMCGs.
[MA-5] Exact Feasibility Certification and Optimal Responsibility Allocation for Multi-Robot CBF Safety Filters
【速读】：该论文旨在解决多机器人控制屏障函数（Control Barrier Function, CBF）安全滤波器在实际应用中因约束冲突导致二次规划（Quadratic Program, QP）不可行的问题，尤其关注在具有异构控制仿射动力学和凸输入集的多智能体系统中，无法准确判断不可行性根源及有效修复策略的挑战。其核心解决方案是提出一种精确的可行性证书（exact feasibility certificate），通过将安全约束带来的需求与执行器可用的控制能力相分离，量化系统的“可行性余量”（feasibility reserve）。这一分解机制能够明确识别出导致不可行性的具体代理及其相互作用，并判断通过调整CBF增益或增强执行能力是否可恢复可行性。在此基础上，论文进一步设计了一种优化算法，通过最大化最差局部可行性裕度，对共享的安全约束进行最优分配，针对多面体输入集构建线性规划求解框架。在320组配对闭环仿真中，该方法将不可行控制步骤比例从约50%显著降低至6.2%，安全违规运行次数由118/160减少至24/160；在52次不可行事件中，该证书成功识别出可通过放松特定交互关系恢复可行性的案例，成功率高达94%。
链接: https://arxiv.org/abs/2609.14935
作者: Chandan Kumar Sah,Jishnu Keshavan
机构: Indian Institute of Science (印度科学理工学院)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注:
Abstract:Multi-robot Control Barrier Function (CBF) safety filters can become infeasible, but a failed quadratic program (QP) does not indicate why the conflict occurred or how to resolve it. To address this, we develop an exact feasibility certificate for multi-agent CBF filters with heterogeneous control-affine dynamics and convex input sets. The certificate quantifies a feasibility reserve by separating the demand imposed by safety constraints from the available actuator supply. This decomposition shows when CBF gain tuning or increased actuation can and cannot resolve infeasibility, and identifies the agents and interactions responsible for the conflict. We further propose an algorithm to optimally allocate shared safety constraints by maximizing the worst local feasibility margin, yielding a linear program for polyhedral input sets. In 320 paired closed-loop simulations, the proposed allocation reduces infeasible control steps from roughly 50% to 6.2% , and reduces safety-violating runs from 118/160 to 24/160 . In addition, across 52 infeasibility events, the certificate identifies an interaction whose relaxation restores feasibility in 94% of cases.
[MA-6] Forty Shades of Blue: Quality-Diversity Alignment via Mode-Conditioned Reinforcement Learning
【速读】：该论文旨在解决大语言模型（Large Language Model, LLM）对齐训练中普遍存在的模式崩溃（mode collapse）问题，即模型在推理阶段输出多样性逐渐丧失，导致表达能力受限，尤其影响需要开放性探索与多元视角的应用场景，如科学创意生成和创造性写作。其解决方案的关键在于提出一种名为MoDA（Mode-conditioned Diversity Alignment）的在线后训练强化学习（Reinforcement Learning, RL）算法，受多智能体强化学习（Multi-Agent Reinforcement Learning, MARL）中协作视角启发。MoDA通过训练一个共享的LLM策略，并引入抽象编号角色作为代理，使各角色在生成时相互竞争以产生与其他角色不同的输出，从而实现对高质量输出空间中互补区域的探索，无需人工设计角色设定或修改模型架构。此外，MoDA采用提示自适应的质量门控机制，动态校准参考质量阈值，仅对满足质量标准的响应给予多样性奖励，有效防止因追求多样性而引发的奖励劫持（reward-hacking）行为，保障生成质量。实验结果表明，MoDA在多个通用能力任务和科学创意、创作写作等特定领域多样性任务上均显著提升，相较于Qwen3-8B基线模型，其SBERT多样性提升265%，平均通用能力通过率提高10.3%；相比最强的DivPO基线，多样性指标（SBERT、E-Vendi）分别提升75.9%和53.8%，同时通用能力通过率仍提升7.0%。整体而言，MoDA为标准后训练方法提供了一种可直接替换的方案，在不牺牲生成质量的前提下，有效扩展了模型的表达输出空间。
链接: https://arxiv.org/abs/2609.14896
作者: Jiayi Yuan,Hangoo Kang,James Jihao Liu,Yejin Choi,Vikram Iyer,Liwei Jiang,Natasha Jaques
机构: University of Washington (华盛顿大学); Stanford University (斯坦福大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注:
Abstract:A notable byproduct of LLM alignment training is mode collapse: the progressive loss of output diversity that narrows a model’s expressivity at inference time. This degradation is especially limiting for applications requiring open-ended exploration and pluralistic perspectives, such as scientific ideation and creative writing. We present MoDA (Mode-conditioned Diversity Alignment), an online post-training RL algorithm that jointly optimizes generation quality and diversity, inspired by the coordination perspective in multi-agent reinforcement learning (MARL). MoDA trains a single shared LLM policy conditioned on abstract numbered roles, where each role acts as an agent competing to produce outputs distinct from the others. This formulation encourages mode-conditioned agents to explore complementary regions of the high-quality output space without requiring hand-crafted personas or architectural modifications. MoDA employs a prompt-adaptive quality gating mechanism that calibrates a reference quality threshold and grants diversity rewards only to responses that meet the threshold, preventing reward-hacking behaviors that compromise response quality. To study quality-diversity tradeoffs, we evaluate MoDA on a comprehensive suite of benchmarks spanning seven general capability tasks and four domain-specific diversity tasks in scientific ideation and creative writing. MoDA improves SBERT diversity by 265% on the Infinite-Chat held-out prompts, while increasing average general capability pass@1 by 10.3% over the Qwen3-8B baseline. Compared with the strongest DivPO baseline, MoDA improves SBERT diversity from 0.274 to 0.482 (+75.9%) and E-Vendi from 2.86 to 4.4 (+53.8%), while improving average general capability pass@1 by 7.0%. Overall, MoDA provides a drop-in alternative to standard post-training methods that preserves and expands the model’s expressive output space while improving quality.
[MA-7] rust by Design: Trust Calibration Through Non-Advisory Socratic Dialogue in Conversational Agents
【速读】：该论文旨在解决在敏感领域中，随着对话式人工智能（Conversational AI）系统应用日益广泛，用户对系统信任过度或不足的问题，核心挑战从可用性转向信任校准（trust calibration）。现有系统在提供建议或解释时，易因用户将AI输出视为权威而引发不当依赖。为此，论文提出CASELy这一对话代理，其关键解决方案在于通过非建议性的苏格拉底式对话（Socratic dialogue）主动限制自身权威性：系统仅基于用户输入提出反思性问题，明确拒绝提供任何建议、推荐或解释，从而以交互约束机制实现信任校准，而非通过增强模型能力。在高等教育学生参与的初步随机对照实验中，采用苏格拉底式对话的用户在用户体验量表（UEQ-S）上得分显著更高（1.50 vs. 0），定性分析揭示了三种支持可信信任的机制：通过可见的输入溯源实现透明性、保障用户决策自主权、降低用户被评判的恐惧感。研究表明，通过交互层面的设计约束可有效实现适度依赖，为敏感场景下的可信对话式AI提供了可复用的设计范式。
链接: https://arxiv.org/abs/2609.14818
作者: Roba Hassan,Nahla Aboromi,Naomi Unkelos-Shpigel
机构: Braude College of Engineering, Israel
类目: oftware Engineering (cs.SE); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC); Multiagent Systems (cs.MA)
备注:
Abstract:As conversational AI systems increasingly operate in sensitive domains, the central challenge shifts from usability to trust calibration, ensuring that users rely on systems neither too much nor too little. Systems that provide advice or interpretations risk encouraging inappropriate reliance, particularly when users perceive AI outputs as authoritative. We present CASELy, a conversational agent explicitly designed to limit its own authority through non-advisory Socratic dialogue. The agent asks reflective questions grounded exclusively in user input and refuses to provide advice, recommendations, or interpretations. This design operationalizes trust calibration by constraining agent agency rather than optimizing capability. In a pilot randomized controlled study with higher education students, participants interacting with the Socratic dialogue reported substantially higher user experience (UEQ-S overall = 1.50) compared to a non-dialogue control (0). Qualitative findings identify three mechanisms supporting calibrated trust: transparency through visible grounding, preservation of user decision authority, and reduced fear of judgment. We argue that appropriate reliance can be achieved through interactional constraints, offering a design pattern for trustworthy conversational AI in sensitive contexts.
[MA-8] Loop-Back Authority in LLM Agent Teams: A Paired Experiment on Flat and Hierarchical Coordination
【速读】：该论文旨在解决多智能体大语言模型（LLM）框架中层级协调机制（hierarchical orchestration）的有效性问题，特别是权威性监督是否能提升最终输出质量。尽管经典组织理论预测上级审批可加速决策收敛，但已有研究指出权威批评可能引发“奉承效应”（sycophancy）与“思维退化”（Degeneration-of-Thought），导致输出质量下降。现有对比研究多聚焦于答案可验证的任务，未充分检验层级结构在开放性任务中的表现。本文通过一对比实验，在保持五名LLM智能体的角色、提示、工具、模型及数据完全一致的前提下，仅改变一个变量：管理者是否具有拒绝工人输出并强制修改的权限。实验基于43对产出、86次商业智能报告生成任务，由五模型评审团和确定性规范检查共同评估每份报告。结果表明，扁平组织在实用性（Utility, d = 0.42, p = 0.009）和写作清晰度（Writing Clarity, d = 0.34, p = 0.030）上显著优于层级组织，经典预测失效。两组报告长度相当，但层级报告的规避性表述（hedging）多出53%，每次修订循环导致写作清晰度下降0.14分，且层级作者的初稿与扁平组织无异——差异仅在修订环节中产生。规范准确率在两类组织中均达天花板水平，而监督层级额外消耗51.5%的计算资源却未带来质量提升。因此，解决方案之关键在于：监督者仅在具备可验证能力时才具价值，当其仅能提供主观意见时，反而成为性能负担。
链接: https://arxiv.org/abs/2609.14767
作者: Burak Agachan,Max van Duijn,Amirhossein Zohrehvand
机构: Leiden Institute of Advanced Computer Science (LIACS), Leiden University (莱顿大学高级计算机科学研究所)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); General Economics (econ.GN)
备注: 8 pages, 3 figures, 3 tables, plus 21 pages of supplementary material. Code: this https URL
Abstract:Hierarchical orchestration, in which a Manager agent reviews worker output and can send it back for revision, is the default coordination pattern in production multi-agent LLM frameworks. Classical organizational theory predicts that the authority link speeds convergence on decisive output; work on sycophancy and Degeneration-of-Thought predicts that authoritative critique makes LLM output worse. Prior comparisons vary whole frameworks on tasks with checkable answers, leaving the authority link untested on open-ended work. We present a paired experiment that holds five LLM agents, their roles, prompts, tools, models, and data fixed and varies one link: whether the Manager may reject a worker’s output and oblige a revision. Across 43 paired products and 86 runs of a business-intelligence reporting task, a five-model judge panel and a deterministic specification check score every report. The flat organization scores higher on Utility (d = 0.42, p = 0.009) and on Writing Clarity (d = 0.34, p = 0.030); the classical prediction fails. The reports are the same length, but hierarchical reports hedge 53% more, each revision loop is associated with a 0.14-point drop in Writing Clarity, and the hierarchical Writer’s first draft is indistinguishable from the flat report: the gap opens inside the revision loop. Specification accuracy is at ceiling in both organizations, and the supervisory tier costs 51.5% more tokens for no quality gain. A supervisor pays for itself when it can verify and becomes a liability when it can only opine.
[MA-9] A latent dimension of Condorcets jury theorem for multiple AI advisers
【速读】：该论文旨在解决在使用多个生成式 AI（Generative AI）顾问进行决策时，如何有效评估其集体判断的可靠性与分歧可见性之间的权衡问题。其核心解决方案的关键在于揭示：尽管根据康多塞陪审团定理（Condorcet’s jury theorem），增加独立且具备一定准确性的顾问可提升多数意见的可靠性，但随着顾问数量增加，意见分歧（visible dissent）的出现概率也趋于必然。通过二项分布模型分析发现，当单个顾问准确率低于0.8时，分歧趋于确定性的速度超过可靠性提升的速度，导致在足够多顾问参与下，分歧反而比正确多数更可能被观察到。因此，即使理想条件下各顾问独立且能力达标，其集体判断仍可能呈现显著分裂状态，而这种分歧并不必然反映聚合机制失效。论文进一步指出，分歧模式为预测多重性、调和负担及依赖度误判提供了共同基础。由此得出两个关键决策维度：一是应选择多少个顾问以实现平衡；二是如何呈现与解读这些顾问的共识与分歧结果。
链接: https://arxiv.org/abs/2609.14438
作者: Kazutoshi Sasahara,Aoi Naito,Ryo Fujie
机构: Institute of Science Tokyo(东京科学研究所); Carnegie Mellon University(卡内基梅隆大学)
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Multiagent Systems (cs.MA)
备注: 11 pages, 4 figures, 1 table
Abstract:When the same question is asked of multiple AI advisers, as in self-consistency and LLM-as-a-judge panels, Condorcet’s jury theorem predicts that adding independent, competent advisers makes the majority more reliable. The theorem, however, has a latent dimension when viewed from the user’s vantage: adding advisers also makes disagreement more visible. A binomial model reveals that this ``visible dissent’’ becomes nearly inevitable as the number of advisers grows, and that reliability and disagreement both approach certainty but at different convergence rates. The two rates cross at an adviser accuracy of 4/5 (0.8). Below this value, visible dissent approaches certainty faster than reliability and, with enough advisers, becomes more likely than a correct majority. Even ideal panels of independent and competent advisers can be correct in aggregate but appear divided; such disagreement does not by itself indicate aggregation failure. The way advisers split also provides a common basis for predictive multiplicity, reconciliation load, and reliance miscalibration. These results indicate two distinct decisions when using multiple AI advisers: how many advisers to consult and how their verdicts should be presented and interpreted.
[MA-10] A Hybrid Dependency-Aware Framework for Task Decomposition and Dynamic Agent Generation in Oracle-to-PostgreSQL Migration
【速读】：该论文旨在解决企业在将复杂异构的Oracle数据库系统迁移至其他平台（如PostgreSQL）时，因传统基于大语言模型（Large Language Model, LLM）的迁移方法仅将其视为简单的代码转换，而忽视了SQL与PL/SQL代码片段之间复杂的依赖关系、执行顺序、结构复杂性及验证需求所导致的迁移失败与覆盖率不足问题。其核心解决方案在于提出一种混合式依赖感知框架，通过构建跨文件依赖图、压缩循环依赖，并在运行时根据任务规范生成专用迁移代理（migration agent），实现精细化的任务分解与调度。该框架采用确定性路径（基于ANTLR解析与类型化依赖提取）处理可可靠解析的代码单元，仅对无法稳定解析的部分启用LLM作为回退机制。实验表明，在116个Oracle文件上，该流程实现了零覆盖率缺口，生成了1,037个代码单元及1,271个基于抽象语法树（AST）的依赖关系；回退机制成功处理165个解析错误单元，额外恢复496个已验证依赖，消除未解析依赖单元，并将内部边的解析率从446提升至527。依赖图被划分为四个尊重依赖顺序的阶段，循环依赖经由Tarjan强连通分量（SCC）压缩验证。在1,006个PL/SQL文件上的补充实验中，约62%的脚本得以重生成，其中380个（约61%）在PostgreSQL 16中成功执行，表级重构成功率约为85%，而基准方案在指定规则驱动下未能成功重生成任何查询，且过程对象仍高度依赖原始模式上下文。研究结果凸显了依赖感知的上下文传递、任务特异性验证以及差异化代理路由的重要性。此外，论文还形式化了任务到代理的映射机制，引入执行反馈的监控与诊断功能，并建立了对单体式、静态分解、依赖感知及动态编排等迁移策略的受控对比分析体系。
链接: https://arxiv.org/abs/2609.14413
作者: Oleg Grynets,Oleg Kaskun,Alona Seletska,Daryna Tukalo,Vasyl Lyashkevych
机构: 未知
类目: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: 18 pages, 9 figures, 11 tables, 36 references
Abstract:Large language model (LLM)-based database migration is often treated as direct code transformation, although enterprise Oracle systems contain heterogeneous SQL and PL/SQL artifacts with different dependencies, execution order, complexity, and validation needs. This paper proposes a hybrid dependency-aware framework that identifies migration tasks, builds a cross-file dependency graph, condenses cyclic dependencies, and uses task specifications to generate specialized migration agents at runtime. The deterministic path combines ANTLR-based parsing with typed dependency extraction, while an LLM fallback is invoked only for units that cannot be parsed reliably. On a corpus of 116 Oracle files, the pipeline produced 1,037 units with zero coverage gaps and 1,271 AST-derived dependencies. The fallback processed 165 parse-error units, recovered 496 additional validated dependencies, eliminated unresolved-dependency units, and increased resolved internal edges from 446 to 527. The graph contained four dependency-respecting phases, while cycle handling was validated separately using Tarjan SCC condensation. A complementary experiment on 1,006 PL/SQL files regenerated 623 scripts (~62%), of which 380 (~61%) executed successfully in PostgreSQL 16. Tables achieved about 85% regeneration success, whereas no query regenerations succeeded under the evaluated specification-mediated baseline, and procedural objects remained strongly dependent on schema context. These results motivate dependency-aware context delivery, task-specific validation, and differentiated agent routing. The paper also formalizes task-to-agent mapping, introduces monitoring and diagnostics for execution feedback, and defines a controlled comparison of monolithic, static-decomposition, dependency-aware, and dynamically orchestrated migration strategies.
[MA-11] Enhancing Human Mobility Prediction with Spatially Aware LLM -based Multi-Agent Systems SIGMOD2026
【速读】：该论文旨在解决基于大语言模型（LLM）的人类移动性预测中忽视真实世界空间上下文的问题。现有方法虽在语义推理方面表现良好，但普遍缺乏对地理距离、道路网络距离及邻里关系等空间约束的建模，而这些因素在人类移动行为中具有关键作用。此外，已有研究表明LLM在空间推理任务中存在性能短板，如距离估计不准和地理偏差。为克服上述局限，本文提出一种多智能体大语言模型框架，其核心创新在于将下一兴趣点（next-POI）预测分解为三个阶段：模式提取代理（Pattern Extraction Agent）捕获轨迹历史中的时间与类别模式；空间推理代理（Spatial Reasoning Agent）结合行为偏好与真实世界空间约束（包括地理距离、路网距离及邻里归属）生成候选活动集；决策合成代理（Decision Synthesis Agent）融合行为模式与空间推理结果完成最终预测。实验基于纽约市基准数据集，在两种LLM骨干架构下均显著优于基线方法，Hit@1最高提升493%，Hit@5相对提升37%。消融实验表明，引入邻里归属信息可显著提升性能，且空间推理代理在小模型中尤为关键，凸显了显式空间约束与行为模式融合的重要性。整体而言，该研究强调空间推理在移动性预测中的核心地位，并证明多智能体分解结构是有效整合行为与空间上下文的关键机制。
链接: https://arxiv.org/abs/2609.14227
作者: Shangyu Lou,Ziqi Cui
机构: University of California Santa Barbara (加州大学圣巴巴拉分校); San Diego State University (圣地亚哥州立大学); Politecnico di Milano (米兰理工大学)
类目: ocial and Information Networks (cs.SI); Computers and Society (cs.CY); Multiagent Systems (cs.MA)
备注: HILDA@ACM SigMod 2026
Abstract:Predicting a user’s next POI is a task in human mobility modeling, yet LLM-based approaches focus on semantic reasoning from previous mobility records, while neglecting real-world spatial context. However, human mobility is inherently shaped by spatial cognition, including geographic distance and neighborhood context. This issue is further compounded by prior evidence that LLMs often struggle with spatial reasoning tasks, including distance estimation and geographically biased prediction. To address these limitations, we propose our framework, a multi-agent LLM framework that decomposes next-POI prediction into three stages: Firstly, a Pattern Extraction Agent that captures temporal and categorical mobility patterns from trajectory history; Secondly, a Spatial Reasoning Agent that structures candidate activity choices by combining behavioral preferences with real-world spatial constraints, including geographic distance, road network distance, and neighborhood affiliation; and Thirdly, a Decision Synthesis Agent that integrates behavioral patterns and spatial reasoning for final prediction. Experiments on the NYC benchmark dataset with two LLM backbones show improvements over baseline methods, with up to 493% Hit@1 improvement and 37% relative improvement in Hit@5. Ablations show that combining neighborhood affiliation with distance-based features generally outperforms distance-only settings, and that the Spatial Reasoning Agent plays a crucial role in final prediction by integrating behavioral preferences with real-world spatial constraints, especially for smaller models. Overall, the results highlight the importance of spatial reasoning in mobility prediction. Accurate next-POI prediction requires combining behavioral patterns with explicit real-world spatial constraints, and multi-agent decomposition provides an effective structure for organizing these forms of context.
[MA-12] Multi-Modal Tumor Survival Prediction via Graph-Guided Mixture of Experts
【速读】：该论文旨在解决复杂真实世界任务中多模态数据融合与缺失模态处理的难题，尤其针对基于临床数据、病理切片和基因组数据预测肺癌患者生存期这一具有高度复杂性的多模态生存预测问题。现有方法往往因无法同时建模所有模态、难以应对模态缺失或异质性，以及缺乏有效的多模态信息整合机制而表现受限。其解决方案的关键在于提出一种受生成式大模型（Generative AI）中混合专家（Mixture of Experts, MoE）架构启发的图引导型MoE框架。该框架通过构建图结构来动态协调多个专用模型的调用，实现对不同模态数据的智能选择与组合，避免了端到端训练单一基础模型的局限性。通过将现有模型作为“工具”在MoE框架下进行协同决策，充分利用各模型的优势并增强整体预测能力。实验结果在TCGA-LUAD数据集上验证了该方法相较于单模态模型及传统集成模型均取得了显著性能提升，证明了其在处理复杂多模态任务中的有效性与鲁棒性。
链接: https://arxiv.org/abs/2609.14072
作者: H Mathavan,H Liu
机构: 未知
类目: Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注:
Abstract:Large Language Models (LLMs) have displayed impressive capabilities in handling tasks that require few demonstration examples, making them effective few-shot learners. Despite their potential, LLMs face challenges when it comes to addressing complex real-world tasks that involve multiple modalities or reasoning steps. For example, predicting cancer patients’ survival period based on clinical data, cell slides, and genomics poses significant logistical complexities. Although several approaches have been proposed to tackle these challenges, they often fall short in achieving promising performance due to their inability to consider all modalities simultaneously or account for missing modalities, variations in modalities, and the integration of multi-modal data, ultimately compromising their effectiveness. This thesis proposes a novel approach for multi-modal tumor survival prediction to address these limitations. Taking inspiration from recent advancements in LLMs, particularly Mixture of Experts (MoE)-based models, a graph-guided MoE framework is introduced. This framework utilizes a graph structure to manage the predictions effectively and combines multiple models to enhance predictive power. Rather than training a single foundation model for end-to-end survival prediction, the approach leverages a MOE-guided ensemble to manage model callings as tools automatically. By leveraging the strengths of existing models and guiding them through a MOE framework, the aim is to achieve better performance and more accurate predictions in complex real-world tasks. Experiments and analysis on the TCGA-LUAD dataset show improved performance over the individual modal and vanilla ensemble models.
[MA-13] A Low-Latency Interactive System for Real-Time Video Understanding Based on VLMs
【速读】：该论文旨在解决当前视觉-语言模型（Vision-Language Model, VLM）在实时视频应用中面临的低延迟交互部署难题，尤其针对从离线片段分析向连续交互式流媒体演进过程中，现有研究过度关注模型能力而忽视可部署性与低延迟交互的问题。其核心解决方案是提出一个统一的边缘-云协同系统架构，通过轻量级客户端（如手机、智能眼镜、PC及伪回放终端）将视频与语音数据实时上传至服务器运行时环境，由服务器统一提供共享的自动语音识别（ASR）/文本转语音（TTS）、会话编排、后端自适应、响应分发以及基于归档数据的性能评估。该系统集成六种具有流式处理或交互导向能力的代表性视频VLM后端，并在后端运行时、媒体传输、客户端感知延迟及交互行为等多个维度进行综合评估。实验表明，在合理选择后端并采用WebRTC传输路径的前提下，系统可实现首次VLM文本输出约0.9–1.0秒、首次非静音TTS音频输出约1.3–1.5秒的极低延迟表现，同时揭示了不同后端在实时交互行为中的适应成本与性能差异，为实际部署中的高效交互提供了可量化参考。
链接: https://arxiv.org/abs/2609.13986
作者: Punan Dai,Jun Xu,Bingcong Lu,Zhengxue Cheng,Hongwei Hu,Ronghua Wu,Li Song
机构: 未知
类目: Multimedia (cs.MM); Multiagent Systems (cs.MA)
备注: 12 pages. Submitted to IBC 2026
Abstract:Vision-language models are extending video understanding from offline clip analysis to continuous interactive streaming, but most research still emphasizes model capability rather than deployable low-latency interaction. This paper presents a unified edge-cloud system for real-time video VLM applications. Lightweight phone, smart glasses, PC, and pseudo-replay clients publish video and speech to a server runtime that provides shared ASR/TTS, session orchestration, backend adaptation, response delivery, and archive-backed measurement. The system integrates six representative video VLM backends with streaming or interaction-oriented capabilities and evaluates them across backend runtime, media transport, client-observed latency, and interaction behavior. With suitable backend selection and the WebRTC path, the tested system reaches approximately 0.9 to 1.0 s to first VLM text and 1.3 to 1.5 s to first non-silent TTS audio, while exposing backend adaptation costs and differences in real-time interaction behavior.
[MA-14] Learning How Much to Collaborate: Difficulty-Aware Topology Selection for Multi-Agent Code Generation
【速读】：该论文旨在解决多智能体代码生成系统中通信拓扑结构选择的粒度不当问题：现有方法通常为每个任务固定使用单一通信拓扑，而未根据问题难度动态调整。研究发现，分层协作（hierarchical collaboration）在难题上的性能优势可高达21.1点（pass@1），但其代价也约为单个代理的十倍，表明静态拓扑策略无法兼顾效率与效果。为此，论文提出难度感知拓扑选择器（Difficulty-Aware Topology Selector, DATS），其核心在于构建一个图神经网络（graph network）作为预测器，将五种拓扑视为具有连通性顺序的节点，而非独立标签，从而捕捉拓扑间的结构依赖关系，相比传统平铺式多标签头提升1.7点性能。通过引入可调的标量成本惩罚项，在预算匹配条件下实现不同方法间的公平比较——在此协议下，六种成本感知方法间性能差距达21.6个百分点，且两个基线模型在成本校准后均落后于DATS。当固定成本为始终采用分层结构的40%时，DATS在APPS、HumanEval+和LiveCodeBench上达到77.7% pass@1，显著优于始终分层的73.6%及最强学习型竞争者74.3%，所有十一组配对McNemar检验均在Holm-Bonferroni校正下保持显著性。该4.1点增益在四种不同骨干模型（覆盖14点能力跨度）中稳定存在；进一步替换39个可解释特征为图网络或预训练编码器后，性能变化不超过1.3点，且无统计显著差异。跨领域实验在400道数学推理任务中复现相同趋势，性能差距从2.5点扩大至20.9点，验证了方法的泛化能力。因此，解决方案的关键在于基于图结构建模拓扑间依赖关系并实现难度自适应的成本-收益权衡优化。
链接: https://arxiv.org/abs/2609.13890
作者: Yunsong Hong
机构: The University of Sydney (悉尼大学)
类目: Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: 20 pages, 11 figures
Abstract:Multi-agent systems for code generation are deployed with a single communication topology, chosen once for every problem. This is the wrong granularity. Evaluating five topologies on 614 problems from APPS, HumanEval+ and LiveCodeBench, we find that the advantage of hierarchical collaboration over a single agent grows from 2.4 points of pass@1 on the easiest third of problems to 21.1 points on the hardest third, while its token cost stays about ten times higher. We propose the Difficulty-Aware Topology Selector (DATS), which predicts each topology’s probability of solving a problem and selects the one maximising predicted success minus cost. Its predictor is a graph network that treats the five topologies as nodes of a connectivity order rather than independent labels, worth 1.7 points over a flat multi-label head. Because the cost penalty is a single scalar recalibrable without retraining, routers compare at equal spend: under this budget-matched protocol six cost-aware methods span 21.6 percentage points, and two baselines leading DATS fall behind once calibrated to it. Fixed at 40% of the always-hierarchical cost, DATS reaches 77.7% pass@1 against 73.6% (always-hierarchical) and 74.3% (strongest learned competitor), all eleven pairwise McNemar comparisons surviving Holm-Bonferroni correction. The 4.1-point gain holds across four backbones spanning fourteen points of capability, and replacing the 39 interpretable features with a graph network or a pretrained encoder shifts accuracy by at most 1.3 points, never significantly. A cross-domain study on 400 mathematical reasoning problems reproduces the effect, the gap widening from 2.5 to 20.9 points.
[MA-15] Identity Is More Than Recall: A Benchmark for Persistent Identity in Deployed AI Agents
【速读】：该论文旨在解决持续性智能体（persistent agents）在长期运行中如何准确区分其可回忆的个体事实（identity facts）与实际表达和行为执行的事实之间的挑战，核心问题是现有评估方法难以有效分离身份契约（identity contract）中的记忆保留、表达、行为执行、抗干扰能力、持久性、溯源性及角色条件更新等不同维度。为此，论文提出PAI-Bench——一个不依赖特定提供方的基准测试框架，专门用于评估版本化、受更新机制管控的身份契约的一致性与忠实度。其解决方案的关键在于将身份相关的各项能力解耦并独立测量：通过设计十六个合成身份、三十二个探测项以及三个独立初始化的目标配置，在不将评分逻辑嵌入目标过程的前提下，实现对记忆召回、内容生成、行为实施、抗干扰性、身份延续性、演化轨迹追踪及角色驱动更新等多维度的系统性评估。实验结果显示，尽管所有原子响应均正确包含直接父级标识符（48/48），但隐式自我表征仅1例成功（1/48），且提示词中的显式字段线索可显著提升三类身份标识的联合出现率（0/8 → 7/8），而启动时的身份标签替换则使完整称谓的出现率从1/8升至7/8，同时父级标识仍缺失，揭示了模型在组件选择上对提示词的敏感性及其组件特异性响应差异。此外，相同因子响应的重放测试显示，不同评估者对同一目标行为的打分存在显著偏差（Claude平均得分比Astra低12.5个百分点），凸显了评估者主观性对结果的影响。研究采用单样本每条件设计，结合事后审计与后续验证，证明了PAI-Bench作为可复现评估协议的有效性，能够独立量化事实可用性、身份表达与行为执行等关键指标，从而为持续性智能体的身份一致性评估提供了标准化、可扩展的技术路径。
链接: https://arxiv.org/abs/2609.13637
作者: Zhenyu Zhao,Roy Zhao
机构: Independent Researcher; University of Washington (华盛顿大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 27 pages, 2 figures, including 18 pages of supplementary material
Abstract:Persistent agents need evaluations that distinguish identity facts they can recall from those they express and enact. We introduce PAI-Bench, a provider-neutral benchmark for fidelity to a versioned, update-governed identity contract. It separates recall, composition, behavioral enactment, resistance, persistence, lineage, and role-conditioned updates while keeping scoring oracles outside the target process. Two frozen campaigns cover sixteen synthetic profiles, thirty-two probes, and three independently initialized target configurations, yielding 1,536 retained responses. A judge-independent literal audit finds direct-parent identifiers in 48/48 atomic responses but only 1/48 implicit self-portraits. On eight profiles, explicit field cues increase joint presence of three identity identifiers from 0/8 to 7/8 under the same four-sentence instruction. A separate startup body-label substitution increases full-designation presence from 1/8 to 7/8 while parents remain absent. These contrasts reveal prompt-dependent component selection and component-specific sensitivity to startup cues in the tested deployments. Replaying identical factorial responses also yields a Claude headline mean 12.5 percentage points below Astra’s, demonstrating evaluator sensitivity separately from target behavior. The studies use single target samples per condition, with post-hoc audits and follow-ups. PAI-Bench provides a reproducible evaluation protocol for measuring factual availability, identity expression, and behavioral enactment as distinct aspects of identity-contract fidelity.
[MA-16] Comparative Framework Analysis for Enterprise Generative AI Applications: Chatbot Automation and Oracle-to-PostgreSQL Migration
【速读】：该论文旨在解决企业在部署生成式AI（Generative AI）应用时面临的框架适配性问题，具体聚焦于三类典型的企业级生成式AI应用场景：基于文档的开发助手、邮件与咨询自动化系统，以及Oracle到PostgreSQL的数据库迁移工具。研究发现，不同应用在组件边界划分、工作流编排、策略检索或推理、模型集成、确定性验证、数据持久化、可观测性及运行效率等方面表现出显著差异，表明单一通用框架难以满足所有场景需求。其解决方案的关键在于采用分层架构设计，即由概率性组件负责生成初步建议，而确定性组件则保留对路由、授权、验证、持久化、幂等性及最终结果的控制权。研究进一步指出，框架适用性需基于具体应用特性、部署环境及组件职责进行精细化评估，强调应从组件层面进行框架选型，并通过明确的契约定义、应用特定的实证支持以及独立的验证边界来保障系统的可靠性与安全性。
链接: https://arxiv.org/abs/2609.13577
作者: Oleg Grynets,Olena Pochernina,Alona Seletska,Daryna Tukalo,Dmytro Kostetskyi,Ivan Fedorchuk,Vasyl Lyashkevych
机构: 未知
类目: oftware Engineering (cs.SE); Logic in Computer Science (cs.LO); Multiagent Systems (cs.MA)
备注: 33 pages, 2 figures, 34 tables, 58 references
Abstract:This study compares framework suitability across three classes of enterprise generative AI applications: a documentation-based development assistant, an email and inquiry automation system, and an Oracle-to-PostgreSQL migration tool. The analysis evaluates component boundaries, orchestration, policy retrieval or reasoning, model integration, deterministic validation, persistence, observability, and operational efficiency. Across the three applications, the evidence supports layered architectures in which probabilistic components generate proposals, while deterministic components retain authority over routing, authorization, validation, persistence, idempotency, and final outcomes. The results indicate that framework suitability depends on the application, deployment conditions, and component responsibility: search quality, workflow control, safety behavior, and migration validation cannot be reduced to a single cross-application assessment. Therefore, the study substantiates the need for framework selection at the component level, supported by explicit contracts, application-specific evidence, and independent validation boundaries.
[MA-17] A Hybrid Agent ic AI Framework for Intelligent Supply Chain Analytics
【速读】：该论文旨在解决供应链分析中决策支持系统难以高效利用的问题，核心挑战在于实现业务决策与跨领域技术专长（包括数据工程、运筹学及领域知识）之间的有效衔接。其解决方案的关键在于提出一种基于多智能体（multi-agent）的供应链分析框架，通过协调者智能体（coordinator agent）理解用户意图并分解任务至具备特定功能的专用智能体，从而实现对探索性分析与确定性工作流的统一支持。该设计将领域逻辑封装于专业智能体与提示词（prompt）中，构建了可扩展、模块化且可审计的系统架构，显著降低了功能扩展的成本。实验结果表明，该多智能体架构在多层级库存管理仿真环境中实现了90%的准确率，相较于单智能体基线表现相当，但输入令牌消耗减少约四倍，大幅提升了系统的可扩展性与成本效益；案例研究进一步验证了其在可解释的次优性检测与自动预测优化中的有效性，展示了开放式的探索分析与确定性供应链分析流程融合的可行性，为构建更易用、可扩展的决策支持系统提供了实践路径。
链接: https://arxiv.org/abs/2609.13561
作者: Xian Yeow Lee,Teppei Inoue,Haiyan Wang,Chetan Gupta
机构: Industrial AI Lab(工业人工智能实验室); Hitachi America(日立美国公司)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Efficient utilization of supply chain analytics for decision making remains a significant challenge for planners, as critical tasks such as database querying, key performance indicator (KPI) analysis, demand forecasting, and performance diagnosis require heterogeneous expertise spanning data engineering, operations research, and domain knowledge. In this work, we propose an agentic system for supply chain analytics that bridges the gap between business decision-making and technical expertise, where a coordinator agent interprets user intent and delegates sub-tasks to specialized agents. The system supports both exploratory analysis and deterministic workflows, enabling planners to transition between ad hoc questions and structured processes. Domain logic is encapsulated within specialist agents and prompts, yielding a scalable, modular, and auditable design and lowering the cost of functional extension through prompt-centric development. We evaluate the proposed architecture on a test environment that replicates multi-echelon inventory management operations. Results show that our multi-agent design achieves a 90% accuracy, which is competitive with a single agent baseline while reducing input token usage by roughly fourfold, substantially improving scalability and cost-efficiency. Furthermore, we provide case studies to demonstrate interpretable suboptimality detection and automated forecast optimization, illustrating how agentic architectures can effectively combine open-ended exploratory analysis and deterministic supply chain analytics workflows, and provide a practical pathway toward more accessible and extensible decision-support systems.
[MA-18] Vibe Patenting: Evaluating LLM Judges for Professional Patent-Drafting Agents
【速读】：该论文旨在解决生成式人工智能在复杂专业任务（如专利撰写）中输出质量评估与优化的可靠性问题，尤其关注大语言模型（LLM）作为评判者在迭代改进AI生成内容中的有效性。其核心解决方案在于构建一个端到端的专利撰写评测平台Vibe Patenting，通过独立调用的LLM评判者对AI生成的专利文本进行结构化评分与反馈，驱动迭代修订。关键发现表明：基于判别式反馈的迭代优化能持续提升生成内容的质量，而无指导的修订则易陷入性能饱和；更重要的是，即使低推理能力的AI代理在持续接收判别反馈后，也能逼近高成本高推理能力代理的表现。研究还验证了判别者与专业专利律师之间的评估一致性，揭示其虽具实用性但存在显著指标依赖性与系统性校准偏差。因此，该研究凸显了LLM判别者在复杂专业工作流中的双重角色——既是有效的评估工具，也是优化信号来源，同时指出了其局限性与需谨慎应用的边界。
链接: https://arxiv.org/abs/2609.13422
作者: Toshiaki Koike-Akino,Vlad Blaykhman,Ye Wang,Jing Liu,Gene V. Vinokur
机构: Mitsubishi Electric Research Laboratories (MERL)
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: 29 pages, 18 figures
Abstract:LLM judges are increasingly used to evaluate and improve AI-generated outputs, yet their reliability for complex professional work remains unclear. We study this problem through Vibe Patenting, an end-to-end patent-drafting testbed for AI-agent evaluation. A separately-invoked LLM judge evaluates generated patent drafts and provides structured feedback for iterative revision. Across multiple inventions and drafting-agent configurations, judge-guided revision consistently improves judge-assessed quality, while unguided revision tends to saturate. Notably, iterative judge feedback enables a low-reasoning agent to approach the performance of a substantially more expensive high-reasoning agent. Stronger models and increased reasoning generally improve judge-assessed drafting quality, while domain-specific agentic workflows provide further gains. We validate the judge against independent evaluation by a professional patent attorney and find meaningful but strongly metric-dependent agreement and systematic calibration differences. These results highlight both the utility and limitations of LLM judges as evaluators and optimization signals for complex professional workflows.
[MA-19] he Agent ic Company OS: Substrate Inversion for Sustained Enterprise Agent Deployment
【速读】：该论文旨在解决企业级生成式 AI 代理在实际持续运行中普遍存在的失败问题，即在演示阶段表现良好，但在长期运营中因缺乏自我迭代与反馈机制而停滞。其核心问题在于：当前代理系统依赖于为人类操作员和传统应用设计的数据结构进行推理，而非适配语言模型（LLM）的内在推理特性，导致上下文理解能力受限且系统难以持续优化。解决方案的关键在于重构代理的认知基础——共享环境（cognitive substrate），将其从传统数据模式转向与语言模型推理表面相匹配的表示形式，同时将模式转换（schema translation）隔离至动作边界（action boundary）。这一重构通过两个核心机制支撑：一是上下文带宽不对称性（context-bandwidth asymmetry），即自然语言连贯文本的单次读取与字段级类型化访问之间在信息完整性上的根本差异；二是跨环耦合性（cross-loop coupling），强调动作、技能与策略循环仅在共享同一底层结构时才能有效协同。为此，论文提出一个四层框架（数据、知识、智能、治理），引入同步代理（Sync Agent）强制执行动作边界，并建立基于技能的可信度梯度，使治理与可审计性成为系统架构的结构性属性。该方案复兴了经典多智能体系统中共享底座的传统，适应大模型时代的技术经济特征。论文还分析了潜在风险（如编译路径上的间接提示注入），并提出了对底层表示进行直接评估的研究议程。
链接: https://arxiv.org/abs/2609.13334
作者: Oliver Aleksander Larsen,Mahyar T. Moghaddam
机构: University of Southern Denmark (南丹麦克罗大学); SDU Software Engineering (南丹麦克罗大学软件工程)
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 19 pages, 3 figures, 2 tables. Accepted as a peer-reviewed short paper (position-paper track) at the 2nd International Conference on Agentic and Generative Techniques in Intelligent Computational Systems (AGENTICS 2026), Angers, France, 28-30 October 2026, part of IJCCI 2026, and for publication in the Springer CCIS proceedings. This is the author’s accepted manuscript
Abstract:Enterprise AI agents often succeed in a demonstration and then stall once they must operate day after day. An industry report estimates that most pilots never reach production and that deployed systems rarely retain feedback or improve over time, while agent benchmarks show single-run successes masking unreliable repetition. We argue that these failure modes share a common architectural root: agents reason over data structured for human operators and traditional applications, not for the language models that power them. This position paper proposes that companies deploying agents in sustained operation should rebuild their cognitive substrate, the shared environment agents read as working context, around representations matched to that reasoning surface, isolating schema translation to the action boundary. Markdown is the instantiation available today, not a proven agent-native primitive. Two mechanisms ground the argument: context-bandwidth asymmetry, the gap between one-pass reading of connected prose and field-by-field typed access that strips relations; and cross-loop coupling, the claim that action, skill, and policy loops compound only if they share one substrate. A four-layer framework (Data, Knowledge, Intelligence, Governance) operationalizes the position, with a Sync Agent enforcing the action boundary and a per-skill trust gradient, making governance and auditability structural properties of the substrate. The position revives the shared-substrate tradition of classical multi-agent systems under LLM-era economics. We analyze the main objections and risks, including indirect prompt injection on the compile path, and outline a research agenda for evaluating substrates directly.
[MA-20] Conceptualization and experimentation of asset market with price manipulation
【速读】：该论文旨在复现人类交易员的行为，详细刻画其心理过程，以理解其决策对资产最终价值的影响。其核心挑战在于如何准确建模复杂的人类行为及其在市场中的动态交互。解决方案的关键在于采用多智能体仿真模型（multi-agent simulation model）构建虚拟交易员，并结合形式化语言与图形化表达方式，对交易行为进行精确描述与可视化呈现。这一方法不仅验证了复现复杂人类行为的可行性，还显著提升了跨学科协作中不同领域专家（如计算机科学、经济学、心理学）之间的沟通效率与行为逻辑的可验证性。
链接: https://arxiv.org/abs/2609.13304
作者: Pau Fonseca i Casas,Aarón Montero Montero
机构: 未知
类目: Multiagent Systems (cs.MA)
备注:
Abstract:The primary goal of this work is to reproduce the behavior of a human trader, detailing his or her psychological processes to understand the effects of his or her decisions on the final value of an asset. The second goal is to use a formal language to detail this behavior as a tool for improving and simplifying the communication between all actors involved in the project, such as specialists from disciplines as diverse as computing, economics and psychology. As a starting point, we use a paper that shows an experiment that analyzes the influence on other trader’s behavior when an agent handler and a trading robot attempt to distort the market. This work reproduces this experiment, using virtual traders that belong to a multi-agent simulation model, showing the feasibility to reproduce complex human behaviors and showing the convenience of use formal and graphical languages to simplify the understanding and the validation of the complex behaviors involved in an economic process.
[MA-21] From Process Loss to Assembly Bonus: Human-Grounded Diagnosis of Multi-Agent LLM Collaboration EMNLP2026
【速读】：该论文旨在解决当前大语言模型（LLM）群体在协作问题求解与人类群体模拟中，仅依赖结果导向评估（outcome-only evaluation）所导致的机制理解缺失问题。研究指出，若将LLM群体用作人类群体的模拟模型，则必须考察其决策过程是否具备类人推理特征，而不仅仅是达成正确结果。其解决方案的关键在于通过对比人类群体对话与匹配的LLM推理轨迹，在沃森式演绎推理任务中识别并验证群体决策过程中的“汇编红利”（assembly bonus）模式，并进一步检验该过程特征在类比、溯因及分析性任务中的可迁移性。研究发现，尽管人类与LLM群体均表现出相同的汇编红利不对称性——即讨论更显著提升普通成员而非初始最优成员的表现——但二者在过程层面存在显著差异：LLM群体更倾向于追随多数意见、暴露较少独特信息且过早收敛；少数派正确观点的成功主要依赖于早期重述。尽管基于人类群体决策研究的干预措施带来有限的集体性能提升，却未能突破协调瓶颈。综上，该研究揭示了LLM群体虽能在结果层面复现部分人类推理模式，但在生成汇编红利与过程损耗的机制上存在本质差异，这对群体模拟及人-机协同具有重要启示。
链接: https://arxiv.org/abs/2609.13261
作者: Ala N. Tak,Teruhisa Misu,Kumar Akash,Zhaobo K. Zheng,Kevin H. Joo,Jonathan Gratch
机构: USC Institute for Creative Technologies(南加州大学创意技术研究所); Honda Research Institute USA, Inc.(本田研究美国公司)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted at EMNLP 2026. Camera-ready version
Abstract:LLM agents are increasingly used for collaborative problem solving and human-group simulation. This makes outcome-only evaluation insufficient: if LLM groups are used as models of human groups, we need to know whether they succeed or fail through human-like deliberative mechanisms. We compare human group chats with matched LLM deliberation traces on Wason-style deductive reasoning, then test whether the same process signatures generalize to analogical, abductive, and analytical tasks. Humans and LLMs show the same assembly bonus asymmetry: discussion improves the average member more often than the best initial member. Initial-answer diversity accounts for the effect of model heterogeneity, increasing movement in both corrective and destructive directions. The main differences are process-level. Compared with humans, LLM groups follow majorities more often, surface less unique information, and converge earlier; correct minority signals succeed mainly when re-expressed early. Interventions motivated by human group-decision research yield modest improvements in collective outcomes, but do not remove the coordination bottleneck. Together, these results suggest that LLM groups can reproduce some outcome-level patterns of human deliberation while diverging in the mechanisms that generate assembly bonus and process loss, with implications for group simulation and human-AI collaboration.
[MA-22] An Evolutionary Computation Framework for Multi-Agent Q-Learning with Mean-Field Environmental Feedback
【速读】：该论文旨在解决网络化群体中多智能体强化学习（multi-agent reinforcement learning）在个体适应、局部交互与动态环境变化之间复杂耦合关系的建模与分析问题。其核心挑战在于如何刻画智能体行为演化与环境状态反馈之间的双向作用机制，尤其是在异质网络结构下宏观合作行为与资源动态演变的协同演化规律。解决方案的关键在于构建一个耦合的学习-环境模型：在固定图结构上，智能体基于无状态Q值进行学习，而群体平均行为则驱动一个可动态调整收益矩阵的环境变量；通过一阶平均场闭合（first-order mean-field closure），推导出描述Q值分布的确定性输运方程，并将其与环境状态的投影离散更新规则相耦合，形成一个可解析求解的宏观动力学系统。该模型在随机正则、Erdős–Rényi、Barabási–Albert及随机几何图等多种网络拓扑上，经有限网络蒙特卡洛模拟验证，表现出对宏观合作水平与环境轨迹的良好再现能力，且均方根误差随种群规模和平均度增加而减小。研究进一步揭示环境反馈会重塑学习所得的动作价值排序，正反馈机制显著增强对初始学习偏差与资源水平的依赖性；同时，环境响应时标起关键作用——快速响应可能导致资源状态提前趋于边界，抑制学习适应，而较慢响应则维持了行为学习与环境恢复之间的有效互动。该工作为理解强化学习与环境动态耦合提供了群体层面的理论框架，并系统评估了平均场近似在不同网络与参数条件下的有效性。
链接: https://arxiv.org/abs/2609.13253
作者: Lichen Wang,Shijia Hua,Linjie Liu
机构: College of Science, Northwest A&F University (西北农林科技大学理学院); School of Mathematical Sciences, University of Electronic Science and Technology of China (电子科技大学数学科学学院)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-agent reinforcement learning in networked populations is governed by the interaction between individual adaptation, local encounters, and changing environmental conditions. To study this interaction, we formulate a coupled learning–environment model in which agents update stateless Q -values on a fixed graph, while their population-average behavior drives an environmental variable that dynamically modifies the payoff matrix. Under a first-order mean-field closure, we derive a deterministic transport equation for the population distribution of Q -values and couple it with a projected discrete update for the environmental state. The resulting model is evaluated against finite-network Monte Carlo simulations on random regular, Erdős–Rényi, Barabási–Albert, and random geometric graphs. Across the tested parameter ranges, the mean-field system reproduces the main macroscopic cooperation and environmental trajectories, and the trajectory-level root-mean-square error generally decreases with population size and average degree. The analysis further shows that environmental feedback reshapes the learned action-value ordering, while reinforcing feedback can produce pronounced dependence on the initial learning bias and resource level. The environmental timescale also plays an important role: a rapid response can drive the resource state to a boundary before learning adapts, whereas a slower response preserves the interaction between behavioral learning and environmental recovery. These results provide a population-level description of coupled reinforcement learning and environmental dynamics and characterize the performance of the mean-field approximation within the tested network and parameter ranges.
[MA-23] Machine learning-assisted calibration of Agent -based Models: surrogate-based optimization with Genetic Algorithm and Particle Swarm Optimization
【速读】：该论文旨在解决代理模型（Agent-Based Model, ABM）校准过程中因目标函数景观具有随机性与崎岖性，且只能通过高成本黑箱仿真进行评估而带来的优化困难问题。其核心解决方案是采用内环代理辅助进化计算（Inner-loop Surrogate-Assisted Evolutionary Computation, SAEC），将机器学习代理模型嵌入遗传算法（Genetic Algorithm, GA）与粒子群优化（Particle Swarm Optimization, PSO）中，在每轮迭代中由代理模型对候选解进行筛选，仅对表现最优的前50%样本进行仿真验证，从而在降低仿真需求的同时修正代理模型误差。实验在布罗克-霍姆斯资产定价模型和岛屿增长模型两种典型ABM上，对48种组合配置（包括2种优化器、5种代理模型、4种校准目标）进行了全因子评估，结果表明，最优的机器学习辅助配置相较最强纯优化器基线，在布罗克-霍姆斯模型上使均方根误差（RMSE）降低20.0%，在岛屿模型上降低63.8%，同时分别减少32.1%和61.1%的计算时间。方差分析（ANOVA）结合邓尼特（Dunnett）事后检验确认了所有代理模型在两种模型上均显著缩短计算时间；然而，在岛屿模型上，各代理模型与纯优化器基线在参数恢复精度上无显著差异，说明报告的精度提升为特定配置下的最优结果，并非平均效应。研究进一步发现，最佳代理-优化器-目标组合随模型复杂度变化而改变，表明代理辅助校准策略依赖于具体模型特性，不存在普适适用的默认方案。
链接: https://arxiv.org/abs/2609.13247
作者: Duguma Yeshitla Habtemariam,Jihwan Lee
机构: Pukyong National University, Busan, Republic of Korea(釜庆国立大学，韩国釜山)
类目: Multiagent Systems (cs.MA); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
备注: Submit V1 with 29 pages and 11 Figures
Abstract:Calibrating an agent-based model (ABM) is difficult because its objective landscape is stochastic and rugged, and can be evaluated only through costly black-box simulations. This study adapts inner-loop surrogate-assisted evolutionary computation (SAEC) to ABM calibration by embedding a machine-learning surrogate within genetic algorithm (GA) and particle swarm optimisation (PSO). At each iteration, the surrogate screens the candidates and the simulator validates only the top 50%, reducing simulation demand while correcting surrogate errors. We evaluate a full factorial of 48 configurations combining two optimisers, five surrogates, and four calibration objectives on two contrasting ABMs, the Brock-Hommes asset-pricing model and the Island growth model. Parameter recovery is measured against pseudo-true values. Relative to the strongest pure-optimiser baseline, the best ML-assisted configurations reduce RMSE by 20.0% on Brock-Hommes and 63.8% on Island, while reducing computation time by 32.1% and 61.1%, respectively. ANOVA with Dunnett’s post-hoc tests confirms significant time savings for every surrogate under both Brock-Hommes optimisers and under GA on Island. No surrogate differs significantly from the pure-optimiser baseline in parameter-recovery accuracy on Island; the reported accuracy gains are therefore best-configuration outcomes rather than average effects resolved at this sample size. The best surrogate-optimiser-objective combination changes with ABM complexity, indicating that surrogate-assisted calibration depends on the model and does not admit a universal default recipe.
[MA-24] Portable Semantics Private Dialects: Reuse and Negative Transfer in Latent Communication Between Language-Model Cells
【速读】：该论文旨在解决共享基因组语言模型社会中，独立训练的智能体群体是否能共享一种统一的潜在数据包语言（latent packet language）这一核心问题，并探究继承接口状态对后续学习的正向或负向影响。其关键解决方案在于通过严格的因果互操作性审计（leakage-controlled causal interoperability audit），系统评估6个独立训练的受限社会在30种有序配对下的跨群体通信能力，采用预注册的原始/正交/线性/非线性对齐阶梯与封闭测试结构，发现尽管存在语义相似的接口，但并未形成统一的原始语言：仅同一初始化配对实现双向完全互操作，另一对呈现不对称的部分兼容性，而所有跨初始化方向在冻结对齐层级上均失败。进一步分析表明，严格零样本迁移失败的根本原因在于新操作指令的解释与执行阶段，而非表征本身；在匹配的适应性因子实验中，全局训练的通信接口反而构成严重的负迁移先验，仅重置数据包读写器和“嘴部”（mouth）模块即可将深度三任务准确率从0.169显著提升至0.857。此外，在多个受控检查点与冻结目标流上的实验显示，继承接口从未超过全新接口控制组0.10的预设阈值。研究结论限定于近迁移的17状态设置，且负迁移效应聚焦于单一全局可见父代群体检查点，附录补充了后验标记全局孪生案例研究。
链接: https://arxiv.org/abs/2609.11365
作者: Narcis Marincat
机构: 未知
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: 17 pages, 1 figure, 5 tables. Companion to arXiv:2608.20054 . Code and evaluation records: this https URL ; checkpoints and fitted alignment maps: this https URL
Abstract:In shared-genome language-model societies, restricted evidence visibility favors reusable, value-indexed latent packet interfaces, whereas the sole high-performing globally visible model in the parent study learned an episode-entangled code. This companion study asks whether independently trained societies share one packet language, where strict zero-shot transfer fails, and whether inherited interface state helps or harms later learning. First, a leakage-controlled causal interoperability audit over all 30 ordered pairs of six independently trained restricted societies – under sealed held-out structure and a preregistered raw/orthogonal/linear/nonlinear alignment ladder – shows the six semantically similar interfaces do not form one raw language: one same-initialization pair is exactly interoperable in both directions, a second shows asymmetric partial compatibility, and all 26 cross-initialization directions fail every frozen alignment rung. Second, within the tested decomposition and a single sealed source formulation, a source-span control localizes strict zero-shot failure to interpretation and execution of the new operator instructions. Third, in a matched adaptation factorial, the globally trained communication interface acts as a severe negative-transfer prior: reinitializing only the packet reader, writer, and mouth raises final depth-three accuracy from 0.169 to 0.857. Fourth, across two restricted checkpoints and two independently frozen target streams each, inherited interfaces never exceeded fresh-interface controls by the preregistered 0.10 margin. All primary conclusions are bounded to a near-transfer 17-state setting; the negative-transfer factorial concerns one globally visible parent-cohort checkpoint, while an appendix adds a post hoc tagged-global twin case study.
自然语言处理
[NLP-0] Bellm an Policy Optimization
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在强化学习中因缺乏可验证奖励（Verifiable Rewards）而导致的推理能力不足问题。现有方法通常依赖于对中间状态价值函数的估计，这在复杂任务中易引入偏差并影响策略优化效果。为此，论文提出了一种无需评判器（critic-free）的贝尔曼策略优化（Bellman Policy Optimization, BPO）方法，其核心在于基于策略镜像下降（Policy Mirror Descent, PMD）框架，利用贝尔曼方程将原目标函数重构为轨迹级（trajectory-level）优化形式。这一重构避免了对中间状态价值的估计，从而提升了训练稳定性与可解释性。理论证明表明，该重构目标与原始PMD目标具有相同的唯一最优解。实际应用中，通过近似该目标函数，推导出可计算的BPO损失函数，其关键特征是采用平滑化的互补词元概率比作为失配修正权重（mismatch-correction weight）。在数学推理基准测试中的实验结果表明，BPO显著提升了模型的推理性能，验证了其有效性。
链接: https://arxiv.org/abs/2609.15987
作者: Zhuoqing Song,Haotian Xu,Xikun Zhang,Lidong Bing
机构: Apodex US, Inc.; Princeton University (普林斯顿大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Optimization and Control (math.OC)
备注:
Abstract:Reinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models (LLMs). We introduce Bellman Policy Optimization (BPO), a critic-free method derived from Policy Mirror Descent (PMD). For autoregressive generation with terminal rewards, BPO uses the Bellman equations to reformulate PMD as a trajectory-level objective. The reformulation avoids estimating state values at intermediate states. We prove that it has the same unique optimal solution as the original PMD objective. We derive the practical BPO loss by approximating this objective. Its mismatch-correction weight is a smoothed ratio of complementary token probabilities. Experiments on mathematical reasoning benchmarks demonstrate the effectiveness of BPO.
[NLP-1] Stellar Colosseum: A Many-Agent Harness for Long-Horizon Research in Mathematics and Theoretical Computer Science
【速读】： 该论文旨在解决生成式人工智能在长周期研究型问题（如数学与理论计算机科学中的定理证明）中因依赖一系列不确定且相互关联的决策而导致的不可靠性问题。传统语言模型虽能生成看似合理的短证明，但在面对复杂、多步骤的研究任务时易陷入错误路径而难以自我修正。其解决方案的关键在于提出Stellar Colosseum——一个模型无关的推理调度框架，通过并行探索多种备选策略、引入“就绪门控”机制判断论证路径成熟度、将证明计划表示为层级化的互依赖子问题，并利用验证器反馈动态修正论证结构。该框架采用重叠随机采样树聚合（overlapping random-sample tree aggregation）整合候选方案及其批判性评估结果，形成统一的研究成果。此外，该工作已集成至Google Antigravity的Teamwork框架中作为“长证明”模式，在多个前沿基准测试中表现优异：在TCS-Bench上使用Gemini 3.1 Pro和Gemini 3.7 Flash达到71.0%准确率；在Codeforces竞赛编程评测中，结合执行反馈的证明导向流水线成功解决218/222题，显著超越传统方法。
链接: https://arxiv.org/abs/2609.15983
作者: Honghao Lin,David P. Woodruff,Yuan Deng,Jieming Mao,Song Zuo,Vahab Mirrokni
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Language models can produce plausible short proofs, but may still be unreliable on long-horizon research problems, where progress depends on a sequence of uncertain and interdependent decisions. We introduce Stellar Colosseum, a model-agnostic harness for allocating inference across research in mathematics and theoretical computer science. Colosseum explores alternative strategies before proof construction, uses a readiness gate to decide when a route is mature enough to decompose, represents the proof plan as interdependent section-level subproblems, and routes verifier findings back to the affected part of the argument. Across these stages, it generates candidates in parallel, attacks them with targeted falsification, and combines candidates and their critiques into a single research artifact through overlapping random-sample tree aggregation. The Colosseum workflow has also been integrated into Google Antigravity’s Teamwork framework as the Long Proof pattern. We demonstrate the capabilities of Colosseum through open-ended research and evaluations on theorem-proving and competitive programming benchmarks. Using Colosseum with Gemini 3.1 Pro, we obtain several new results that address open problems arising from papers published at top venues such as FOCS and JMLR. On TCS-Bench, a benchmark of research-level theorem-proving tasks drawn from papers published at FOCS, STOC, and SODA, Colosseum achieves 71.0% accuracy using Gemini 3.1 Pro and Gemini 3.7 Flash. In a separate Codeforces evaluation using Gemini 3.1 Pro, the proof-oriented pipeline with execution feedback solves 218 of 222 problems. Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG) Cite as: arXiv:2609.15983 [cs.AI] (or arXiv:2609.15983v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.15983 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-2] he Router Within: Eliciting Native Skill Routing from a Frozen LLM
【速读】： 该论文旨在解决大型语言模型（LLM）代理在执行复杂任务时，如何高效、准确地从技能库中选择合适技能的问题。现有方法如预加载技能元数据至上下文会分散模型注意力并限制技能库规模，而基于检索的流水线虽将选择过程移出上下文，却超出了代理自身的推理能力。其核心解决方案是利用冻结的LLM本身在前向传播中隐含的路由信号，并仅通过两个线性映射即可无须在上下文中引入技能文本便读取出该信号。Gavel（Glance And Verdict from a frozen LLM）采用两阶段机制：首先“凝视”（Glance）阶段通过两个可训练线性映射投影任务与各技能的中间层状态，结合安装时单次前向传播构建的紧凑技能特征库进行评分；随后“判断”（Verdict）阶段恢复筛选后技能的前向传播，融合模型自身生成概率与二分类判断，形成专家乘积（product of experts）决策。该方法仅需一次训练，即可零样本迁移至三个公开基准及作者新提出的包含372条模拟代理轨迹的SkillTraj基准。在Qwen3-32B上，Gavel在书面任务上相比引入1.2B至16B外部参数的渐进式披露和检索重排序管道最高提升13.4分，在任务中段需调用技能时提升高达21.9分；路由准确率随主干模型性能提升而同步增长，且在bash代理场景中，同一32B模型触发正确技能的频率超过运行于Codex的更大前沿模型。
链接: https://arxiv.org/abs/2609.15982
作者: Ruishuo Chen,Xun Wang,Yu Chen,Zhuoran Li,Longbo Huang
机构: Institute for Interdisciplinary Information Sciences, Tsinghua University(清华大学交叉信息研究院)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Skills extend an LLM agent beyond its parametric knowledge, and the gain they promise rests on picking the right one. Deployed harnesses route by preloading every skill’s metadata into the context, which disperses the agent’s attention and caps the library size. Retrieval pipelines move the selection out of the context, but also out of the agent’s capability. We show that the frozen agent LLM already carries the routing signal in its own forward passes, and that two linear maps suffice to read it out with no skill text in the context. Gavel (Glance And Verdict from a frozen LLM) reads it in two steps. A glance projects the task’s and each skill’s mid-layer states through the two maps, the only parameters trained, and scores the full library against compact per-skill banks that one forward pass builds at installation. A verdict then resumes the shortlisted skills’ forward passes and reads the model’s own likelihood and yes/no judgment, fused with the glance as a product of experts. Trained once, Gavel transfers zero-shot to three public benchmarks and SkillTraj, our new benchmark of 372 simulated agent trajectories. On Qwen3-32B it outperforms progressive disclosure and retrieve-and-rerank pipelines that add 1.2B to 16B external parameters, by up to 13.4 points on written tasks and up to 21.9 when the need for a skill arises mid-rollout. Routing accuracy improves as the backbone does, and in a bash-agent harness the same 32B triggers the correct skill on Skill-Use more often than far larger frontier models running in Codex.
[NLP-3] Disentangling Representation Evolution in Transformers through Directional Decomposition EMNLP2026
【速读】： 该论文旨在解决预训练模型中表示（representation）演化机制不清晰的问题，特别是理解Transformer架构中各模块更新方向的几何特性及其对模型行为的影响。其核心挑战在于揭示学习到的更新（如注意力与MLP层的变换）在隐藏状态空间中的方向性分布，以及这些方向如何影响模型的可编辑性、压缩鲁棒性及训练动态。解决方案的关键在于提出一种功能几何分解方法，将学习到的更新分解为平行于当前隐藏状态方向（parallel component）和垂直于该方向（perpendicular component）的两部分。通过在值空间（value space）与残差空间（residual space）分别进行此分解，研究发现：在值空间中，排除自注意力的并行操控（exclude-self value-space parallel manipulation）具有显著更强的鲁棒性，能够有效保留直接自注意力信息的同时仅调节非自注意力聚合项；此外，该分解方法还揭示了压缩导致的更新误差中，垂直分量比平行分量更能有效区分不同压缩方法。实验进一步表明，在从零开始的预训练过程中，对完整聚合项实施值空间的并行抑制可显著降低验证损失轨迹并提升下游性能，其中值空间变体效果最优。这一工作首次建立了表示几何结构与模型可编辑性、压缩诊断及训练干预之间的系统性关联。
链接: https://arxiv.org/abs/2609.15975
作者: Shwai He,Haichao Zhang,Shen Yan
机构: University of Maryland, College Park (马里兰大学学院帕克分校); Northeastern University (东北大学); ByteDance (字节跳动)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Findings of EMNLP 2026
Abstract:Transformer representations evolve through learned additive transformations that either preserve their current direction or redirect it. We study this evolution as a functional geometry, decomposing learned updates into parallel and perpendicular components. Across pretrained models, we find substantial parallel components beyond the residual identity path. We then apply the decomposition in two spaces: to attention and MLP updates relative to the hidden state, and to attention value aggregation relative to the current token’s value. Targeted edits reveal a strongly space-dependent asymmetry: exclude-self value-space parallel manipulation is markedly more robust than residual-space and perpendicular counterparts, preserving the direct self message while scaling only the non-self aggregate. The same decomposition gives a component-resolved description of compression-induced update error: perpendicular error separates compression methods more clearly than parallel error. Extensive experiments further demonstrate that full-aggregate parallel suppression during from-scratch pretraining lowers validation-loss trajectories and improves downstream averages, with the value-space variant strongest. Together, these results connect representation geometry to editing robustness, compression diagnosis, and training-time intervention. Code is available in the \hrefthis https URLproject repository.
[NLP-4] Discovery Foundation Models: Toward Open-Ended Discovery Intelligence
【速读】： 该论文旨在解决当前基础模型（Foundation Models）在智能系统中仍局限于基于人类预先定义问题进行学习与执行，而无法主动参与新问题发现、知识建构与科学探索过程的根本性局限。其核心挑战在于如何使模型从被动响应转向主动创造，即在开放环境中自主识别研究问题、构建表征、形成假设、设计干预实验，并依据实证反馈持续迭代优化探索路径。解决方案的关键在于提出“发现智能”（Discovery Intelligence）这一新型范式，并构建发现型基础模型（Discovery Foundation Models, DFMs），其核心特征是具备可动态更新的研究状态（revisable research state）与七项耦合能力：问题发现、问题形式化、表征构建、假设生成、干预实施、证据驱动的修正以及持续改进。该框架通过Zetema系统实现，集成显式研究状态演化、验证与实验门控机制、外部世界对齐及跨任务发现技能进化；进一步以真实治疗发现系统GALILEO为实例，将数字推理（Dry-Lab）、机器人湿实验（Wet-Lab）与外部生物证据整合为闭环物理发现循环。最终，论文建立了一套以能力形成为核心、以过程为中心的统一评估方法，使发现行为不仅可训练、可优化，还可超越最终答案性能进行量化评估。这一进展标志着智能演进的新阶段：从知识学习到行动反馈学习，再到深度参与知识生成结构的建构、测试与迭代的全过程。
链接: https://arxiv.org/abs/2609.15973
作者: Ling Yang,Zhenfei Yin,Yingcheng Wu
机构: 未知
类目: Computation and Language (cs.CL)
备注: Website: this https URL , Code: this https URL
Abstract:Foundation models have progressed from learning and reasoning over existing knowledge, to increasingly learning through action, tool use, and outcome feedback. We argue that the next frontier is a further transition: from solving and acting within problems specified by humans to participating in the process by which new problems, representations, explanations, and knowledge are created. We refer to this capability as Discovery Intelligence. We formulate Discovery Foundation Models (DFMs) as general-purpose model systems for open-ended discovery. A DFM operates over a revisable research state and supports seven coupled capabilities spanning problem discovery, formulation, representation construction, hypothesis formation, intervention, evidence-grounded revision, and continual discovery improvement. We instantiate this framework with Zetema, which couples explicit research-state dynamics, verification and experimental gating, external grounding, and cross-task Discovery Skill evolution. We further ground the framework with GALILEO, a real therapeutic-discovery system in which Dry-Lab reasoning, robotic and hands-on Wet-Lab experimentation, external biological evidence, and iterative hypothesis and design revision form a closed physical discovery loop. We then formulate a unified approach to capability formation and process-centered evaluation, enabling discovery behavior to be trained, improved, and measured beyond final-answer performance. Together, these components establish discovery as a learnable, executable, and evaluable capability of foundation-model systems. We view this shift as a broader progression in intelligence scaling: from learning over existing knowledge, to learning from action outcomes, and ultimately to participating in the construction, testing, and revision of the structures through which new knowledge is discovered. Code: this https URL
[NLP-5] Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States
【速读】： 该论文旨在解决当前大语言模型（LLM）在长期学习、推理与决策协作中缺乏对用户深层心理状态理解的问题。其核心挑战在于：现有训练数据普遍缺乏显式基于用户未言明信念与目标的高质量响应，导致模型难以实现真正的人类感知型交互。为克服这一监督缺口，论文提出Mind2Dialogue框架，其关键在于通过心理学引导的用户心理状态模拟器，动态建模用户的内在认知状态，并生成具有连贯性与一致性的对话序列；该框架的核心思想是构建一个共享演进的心理状态空间，用以驱动用户行为并指导“元助手”（Oracle assistant）生成具备充分上下文理解能力的响应。随后，通过特权蒸馏（privileged distillation）技术，使模型在部署时无需直接访问用户心理状态，仍能基于模拟生成的优质响应进行个性化辅助。此外，论文引入结合个性化与心智理论（Theory of Mind）的评估范式，系统检验模型对人类意图的理解与响应能力。实验表明，基于完整Mind2Dialogue语料库训练的模型在多项个性化指标上显著优于Qwen、Llama及OLMo等基线模型，偏好遵循生成任务性能提升达26.6至40.9个百分点，并在信念推理与行为预测任务中展现出更强的能力。未来，该框架有望成为实现真正理解人类信念与意图的智能协作伙伴的基础。
链接: https://arxiv.org/abs/2609.15972
作者: Zixuan Wang,Yufan Zhou,Jinzhou Tang,Xinle Yu,Chengjun Wu,Lyumanshan Ye,Zhaoxiang Feng,Letian Peng,Adyasha Patra,Fan Bai,Enze Ma,Zhengding Hu,Jianyang Gu,Zhao Wang,Yufei Ding,Jingbo Shang,Tianmin Shu,Zhiting Hu,Zhen Wang
机构: UC San Diego(加州大学圣地亚哥分校); KU Leuven(鲁汶大学); University of Illinois Chicago(芝加哥伊利诺伊大学); The Ohio State University(俄亥俄州立大学); Johns Hopkins University(约翰霍普金斯大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 40 pages, 10 figures, 11 tables. Project page: this https URL
Abstract:As language models become more capable, long-term collaboration in learning, reasoning, and decision-making calls for a deeper understanding of the people they serve. Yet training such human-aware language models faces a fundamental supervision gap because current datasets for LLM assistant training contain few if any well-informed responses explicitly grounded in users’ unspoken beliefs and goals. Scaling such supervision is inherently constrained, as users’ underlying states are not directly observable. We thus propose the Mind2Dialogue framework to mitigate this gap by simulating users’ mental states and turning them into privileged supervision for human-aware training. Specifically, we first propose a psychology-guided simulator that preserves personal characteristics while updating mental states through interaction to generate coherent conversations. The key idea is to enforce a shared evolving mental state that drives user behavior and guides an Oracle assistant’s responses. Our privileged distillation then trains models on the Oracle’s well-informed responses to assist users without direct access to their mental states at deployment. Moreover, we propose to evaluate human-aware learning by combining personalization and theory of mind, examining how models understand people and act on that understanding. Training on the full Mind2Dialogue corpus improves every reported personalization metric over the corresponding Qwen, Llama, and OLMo instruction-tuned baselines, including gains of 26.6 to 40.9 percentage points in preference-following generation. The gains extend to belief and action reasoning on Qwen and Llama, beyond personalized assistance. Looking forward, Mind2Dialogue makes user simulation a foundation for genuine AI collaborators that understand beliefs and intentions behind people’s words and support their long-term goals across education, work, and everyday life.
[NLP-6] Verifiable by Construction: Claim-Level Evaluation of Verbatim Citation in Clinical Question Answering
【速读】： 该论文旨在解决当前大型语言模型（Large Language Models, LLMs）在临床问答（Clinical Question Answering, QA）系统中生成的答案缺乏可验证性的问题。尽管现有系统能够为回答附上参考文献引用，但这些引用通常指向范围较广的文本内容，导致时间紧张的临床医生难以高效核实其真实性。为此，论文提出一种构建可验证答案的解决方案：通过端到端地从每个事实性主张出发，自动生成精确的原文引用（verbatim quotes），并确保这些引文能完整支持对应主张，从而使用户无需查阅外部文档即可完成验证。该方案的关键在于实现三个核心阶段的协同优化——准确识别每项主张的来源、生成与原文完全一致的引用片段、以及保证引用内容充分涵盖并支撑主张的所有细节。研究基于四份临床实践指南构建标准化评估框架，对十二个主流LLM在222个合成临床问题上的表现进行量化分析，结果表明，大多数模型虽能在提示工程（prompting）驱动下为超过90%的主张生成原文引用（如claude-opus-5达98.0%），但仅有极小比例的引用能真正全面支撑主张（如claude-opus-5仅37.1%），揭示了当前模型在“可验证性”方面存在显著能力缺口。该工作不仅提供了对现有模型局限性的深入洞察，还贡献了可用于未来研究的评估工具与数据集。
链接: https://arxiv.org/abs/2609.15964
作者: Jiashuo Zhang,Yuling Chen,Yvonne Commodore-Mensah,Michael Oberst
机构: Johns Hopkins University (约翰霍普金斯大学); Johns Hopkins University (约翰霍普金斯大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) have been widely adopted for clinical question answering (QA). Current systems can attach citations to their answers, but these often point to broad texts, leaving time-pressed clinicians unable to verify them efficiently. An alternative is to ensure that responses are verifiable by construction: providing fine-grained verbatim quotes from reference material that substantiate claims, so users can verify an answer without opening other documents. In this paper, we evaluate the ability of current models to perform this task end-to-end: from providing citations for every factual claim, to producing verbatim quotes, to ensuring that those quotes fully substantiate the claims. To do so, we build a standardized harness over four clinical practice guidelines and evaluate twelve LLMs on 222 synthetic clinical questions, measuring each of these stages separately. We find that most models can attach verbatim quotes to over 90% of their claims from prompting alone, apart from some lightweight models such as claude-haiku-4.5. Yet these quotes often fail to substantiate every detail of the claims they accompany. For instance, claude-opus-5 produces verbatim quotes for 98.0% of its claims, but fully substantiates only 37.1%. Our work provides insights into the current capability gap of LLMs in building verifiable clinical QA systems, along with artifacts for future research.
[NLP-7] Inoculation Midtraining with Learned Neologisms
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在后训练阶段可能同时习得有益与有害行为的问题，核心关注点在于：是否可以通过中段训练（midtraining）这一较早的训练阶段，干预并引导模型在未来对某些不良行为的泛化能力。其解决方案的关键是提出“接种式中段训练”（Inoculation Midtraining）方法，即在中段训练期间引入一个名为“quarantine_token”（隔离标记）的新词（neologism），使模型学习到不安全行为应被限定于该特定上下文；随后在后训练阶段，模型在包含该隔离标记的上下文中接受不安全数据的训练。最终评估时，移除该隔离标记，测试模型在非上下文环境下的表现。实验结果表明，该方法可在保留良性属性迁移（如用德语或莎士比亚体写作）的同时，降低模型的错位程度（misalignment），实现对不良行为的选择性泛化抑制。然而，该方法未超越标准的“接种提示”（Inoculation Prompting），且对训练配置敏感，存在边界泄漏问题，即邻近上下文线索仍可激活被抑制的行为。这说明通过中段训练引入的学习关联确实能影响选择性泛化，但该技术尚需进一步优化，方能成为开发者安全框架中的核心组件。
链接: https://arxiv.org/abs/2609.15886
作者: Kyle O’Brien,Edward James Young,Puria Radmard,Nathalie Kirch,Cameron Tice,Tomek Korbak,David Demitri Africa
机构: Geodesic Research(地缘研究); OpenAI; UK AI Security Institute(英国人工智能安全研究所)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) often learn both desirable and undesirable properties during post-training. We study whether midtraining, an earlier training stage, can shape which of these properties later generalise. We introduce Inoculation Midtraining, a technique that teaches a base model that unsafe behaviour belongs to a designated quarantine_token context, as indicated by the quarantine_token neologism (a new token) introduced during midtraining, and then post-trains the model on unsafe data within that context. We then evaluate the model outside the context, with the quarantine_token neologism excluded from the system prompt. Across supervised fine-tuning and reinforcement learning post-training regimes, we find that Inoculation Midtraining can reduce misalignment while preserving the transfer of benign data properties (e.g., speaking in German or Shakespearean prose). However, our approach does not outperform standard Inoculation Prompting, is sensitive to training configuration, and produces a leaky boundary that nearby contextual cues can reactivate. These results show that inoculation with a learned association introduced via midtraining can shape selective generalisation. Still, more work is needed before this approach can become a load-bearing component in a developer’s safety framework.
[NLP-8] K-Bench: a clinically calibrated benchmark for evaluating large language models in high-risk mental health conversations
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在动态、高风险心理健康对话场景中的安全性与临床有效性评估不足的问题。当前，尽管用户越来越多地依赖LLMs获取心理健康支持，但其在处理自杀、自伤、家庭暴力、物质滥用等高危情境时的表现缺乏系统性、可比性的评估标准。为此，研究提出K-Bench——一个由临床专家校准的受保护基准测试平台，涵盖来自14家厂商的33个基础模型共125种模型配置，基于200个包含多轮对话的情景案例进行评估，覆盖自杀、自伤、家庭暴力、物质滥用及无风险情境。研究发现，合成患者对话与真实人机对话在分布上具有显著重叠；采用冻结版GPT-4o作为评判模型，在6,751个有效评估项中与临床专家共识达成94.2%的精确一致率。结果显示，表现领先的模型在提供支持性对话的同时，综合风险评分均高于95，而风险探索能力在低性能配置间存在显著差异。通过治疗引导提示（therapeutic prompting）可在特定配置上实现增益，尤其对较弱模型效果明显，但提升推理能力并未带来整体性能改善。K-Bench的关键创新在于结合更广泛的临床覆盖范围与配置级对比能力，并建立持续更新的公开排行榜，同时通过保护测试材料防止模型直接优化，确保评估的公平性与实用性。
链接: https://arxiv.org/abs/2609.15855
作者: Laura M. Vowels,Matthew J. Vowels,Shivali Sharma,Apoorv Jha,Rehnuma Choudhury,Wasseem El Sarraj,Rachel Francois-Walcott,Aruba Hussain,Sarah Ingram,Angela Loulopoulou,Adva Segal,Elena Volkova
机构: University of Roehampton (罗汉普顿大学); Kivira Health (基维拉健康); University of Hertfordshire (赫特福德大学); University of Surrey (萨里大学); University of Bedfordshire (贝德福德大学); Tavistock Relationships (塔维斯托克关系中心); InsideOut (内部探索)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:% !TEX root = …/main.tex People increasingly use large language models (LLMs) for mental health support, yet their safety in evolving, high-risk conversations remains poorly characterised. We developed K-Bench, a clinician-calibrated, protected benchmark evaluating 125 model configurations representing 33 base models from 14 providers across a fixed cohort of 200 multi-turn vignettes involving suicide, self-harm, domestic violence, substance misuse, and no-risk presentations. Synthetic patient conversations showed substantial distributional overlap with real human-AI conversations. A frozen GPT-4o judge achieved 94.2% exact agreement with clinician consensus across 6,751 eligible item comparisons from 151 clinician-rated transcripts. Leading models combined strong supportive conversation with combined-risk scores above 95, whereas risk exploration exposed substantial variation among lower-performing configurations. Therapeutic prompting produced configuration-specific gains concentrated among weaker models, while elevated reasoning produced no average improvement. K-Bench combines broader clinical coverage and configuration-scale comparison with a continuously updated public leaderboard whose operational test materials are protected from direct optimisation. The leaderboard is available at this http URL.
[NLP-9] Learning to Coach for Experiential Learning
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）在基于经验学习时，因原始解题轨迹过长且噪声较大而难以有效提取可指导性知识的问题。其核心挑战在于如何从复杂、冗余的推理过程中提炼出简洁、可迁移的行动建议，以提升后续任务表现。解决方案的关键在于提出“学习做教练”（Learning to Coach, L2C）框架：训练一个专用的LLM-as-a-Coach，从执行者模型（actor model）的历史轨迹中提取可操作的经验知识；其中执行者模型保持冻结，而教练模型通过最大化由引导响应正确性所定义的奖励信号进行优化。该框架设计了两种奖励机制——同实例奖励（same-instance reward）用于提升对原问题的后续响应质量，跨实例奖励（cross-instance reward）则促进知识在不同任务间的泛化能力。实验表明，L2C在数学推理与交互式文本游戏任务中均显著优于自精炼（self-refinement）和未训练的LLM-as-a-Coach，且随着迭代次数增加，准确率持续提升，同时更高效地利用额外推理计算资源，相较于扩大执行者解码预算更具优势。此外，训练后的教练模型具备良好的分布外泛化能力，并能根据特定执行者动态调整其指导策略。
链接: https://arxiv.org/abs/2609.15851
作者: Guanheng Chen,Tianzhu Ye,Li Dong,Xun Wu,Shaohan Huang,Furu Wei
机构: Microsoft Research(微软研究院); Tsinghua University(清华大学)
类目: Computation and Language (cs.CL)
备注: 21 pages, 16 figures
Abstract:Language models can learn from experience, but raw solution trajectories are often too long and noisy to provide effective guidance. In this work, we propose Learning to Coach (L2C), a framework that trains a dedicated LLM-as-a-Coach to extract actionable experiential knowledge from an actor model’s previous trajectory. The actor remains frozen, while the LLM-as-a-Coach is trained to maximize a reward given by the correctness of the actor’s guided response. We study two such rewards: a same-instance reward, which improves subsequent responses on the original problem, and a cross-instance reward, which elicits knowledge that transfers to other instances. Across mathematical reasoning and interactive text-games, L2C consistently outperforms self-refinement and an untrained LLM-as-a-Coach. Running experiential learning for more iterations further improves accuracy and uses additional inference compute more effectively than enlarging the actor’s decoding budget. The trained LLM-as-a-Coach also transfers to out-of-distribution tasks and adapts its guidance to the specific actor it coaches.
[NLP-10] Before You Poll with LLM s: A Deliberative Diagnostic Framework EMNLP2026
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在模拟公众意见时是否具备类人推理能力的问题，即模型能否像人类一样基于新信息动态调整信念，而非仅依赖预先存储的偏见或“缓存观点”。现有评估方法仅检验模型在静态情境下是否持有正确立场，忽略了真实社会讨论中信念随信息输入而动态演化的关键特性。为此，作者提出审议式民意调查诊断框架（Deliberative Polling Diagnostic Framework），通过对比人类与大语言模型（LLM）在相同信息干预后的信念变化，量化其动态适应性。该框架的核心创新在于将政治心理学中的审议式民意调查机制引入模型评估，揭示了传统静态评价无法捕捉的系统性缺陷：部分模型虽能输出看似合理的党派化观点，却在面对平衡信息时表现出非理性信念反转、过度修正或僵化不变等行为。实证分析显示，五款前沿模型均存在显著失败模式，且各具特征——如GPT-5.1出现反向效应（信念恶化），Gemini 2.0 Flash、Claude Sonnet 4.5及Llama 3.3 70B呈现过度修正，DeepSeek V3则表现出僵化不变；进一步的消融实验表明，这些错误具有身份特异性和内容触发性，根源在于模型对角色身份内部刻板印象的无意识迎合，作者将其命名为“自我谄媚”（self-sycophancy）。因此，该研究的关键解决方案是建立一个可操作的动态评估协议，以在信任生成式AI用于大规模民意模拟前，识别并校准其信念演化机制的偏差。
链接: https://arxiv.org/abs/2609.15849
作者: Ahmed Wali,Hassaan Tayyab
机构: Lahore University of Management Sciences(拉合尔管理科学大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 17 pages, 2 figures. Accepted to EMNLP 2026 Main Conference. Note: Web abstract is abridged to meet arXiv character limits. See PDF for the full proceedings abstract
Abstract:Can LLMs reason through new information like humans, or do they merely retrieve cached opinions? This is critical for silicon sampling, where LLM personas simulate public opinion at scale. Current evaluations test only whether personas hold the right opinions – a static snapshot. But opinion research increasingly depends on dynamic fidelity: whether personas update beliefs in response to new arguments, as humans do during deliberation. No existing benchmark tests this. We introduce the Deliberative Polling Diagnostic Framework, which compares human and LLM belief shifts after identical informational interventions. Grounded in deliberative polling, it surfaces failures invisible to static evaluation: models that produce plausible partisan opinions can still misrepresent how those opinions change. Applying the framework to five frontier models using data from America in One Room (526 personas, 72 questions), we find that every model fails, each in a unique manner. GPT-5.1 exhibits reversal: its personas become more hostile toward the opposing party after balanced information, while humans become less so. This reversal is selective (80% on outgroup vs. 26% on policy questions) and symmetric across partisan identities. Gemini 2.0 Flash, Claude Sonnet 4.5, and Llama 3.3 70B exhibit overshoot, shifting correctly but at 5-7x human magnitude. DeepSeek V3 exhibits rigidity with near-zero change. Targeted ablations reveal that policy content triggers these failures and that they are identity-specific: GPT-5.1 reverses on outgroup questions but overshoots on ingroup; Gemini shows the inverse. We term this signature self-sycophancy: conformity to the model’s internal stereotype of the persona rather than reasoning from the information provided. Our framework offers a concrete protocol: run the deliberative diagnostic before trusting LLM personas to mimic revised beliefs.
[NLP-11] EvoOntology: A Self-Evolving Ontology Layer for Data Agents
【速读】： 该论文旨在解决数据代理（data agent）在处理异构数据（如表格、文件和数据库）时面临的“代理-数据鸿沟”问题：由于异构数据位于代理外部，而代理仅能通过通用工具访问有限信息（如列名和文件路径），导致其难以高效理解与操作数据。现有方法要么让代理直接探索原始数据源，要么在提示中注入人工构建的语义层，但二者均难以扩展至大规模异构数据集，且缺乏对不同代理行为的适应性。为此，本文提出EvoOntology——一种自演化本体层（self-evolving ontology layer），作为支持数据代理的MCP服务器，包含模式层（schema layer）、内容层（content layer）和工具层（tool layer），使代理可在运行时主动查询与交互本体。其核心创新在于引入构建代理（builder agent）实现本体的自主构建，以及基于属性引导的类型化编辑与骨干模型条件下的成对评估机制构成的自演化循环，确保本体持续优化。在三个主流数据代理基准测试及四种大语言模型（LLM）后端上的实验表明，EvoOntology显著优于强基线与现有语义层方法，有效弥合了代理与数据之间的鸿沟，提升了对异构数据的交互效率与准确性。
链接: https://arxiv.org/abs/2609.15779
作者: Meiduo Chong,Shaolei Zhang,Ju Fan,Xiaoyong Du
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Databases (cs.DB)
备注: Code: this https URL
Abstract:Data agents aim to fulfill natural-language instructions over heterogeneous data, including tables, files, and databases. However, data agents face a challenging agent-data gap: heterogeneous data resides outside the agent, while the agent can access it (e.g., column names and file paths) only through generic tools. Existing approaches either let agents directly explore raw data sources or inject manually constructed semantic layers into prompts. However, neither scales well to large heterogeneous data sources nor adapts to different agent behaviors. In this paper, we introduce EvoOntology, a self-evolving ontology layer for data agents. EvoOntology encapsulates the ontology as an MCP server comprising a schema layer, a content layer, and a tool layer, enabling agents to actively query and interact with the ontology at runtime. To this end, we introduce a builder agent for autonomous ontology construction and a self-evolution loop that continuously refines the ontology through attribution-guided typed edits that are accepted only after a backbone-conditional paired evaluation. Experiments on three well-adopted data-agent benchmarks with four LLM backbones demonstrate that EvoOntology consistently outperforms strong baselines and existing semantic-layer approaches, effectively bridging the agent-data gap and enabling more effective interaction with heterogeneous data. Code: this https URL
[NLP-12] Enabling Streaming User Transcription in Full-Duplex Speech-to-Speech Models
【速读】： 该论文旨在解决全双工语音到语音（Full-duplex Speech-to-Speech, S2S）模型缺乏用户语音实时转写能力的问题，这一缺陷限制了其在对话日志记录、无障碍功能及质量监控等应用场景中的实用性。解决方案的关键在于在现有S2S模型的基础上，以轻量级方式并行引入一个流式自动语音识别（ASR）头，该设计仅需极少额外参数且无需对基础S2S模型架构进行重大修改。该方法实现了在保持全双工对话能力（包括自然的换言与抢话处理）的同时，实时生成用户语音的文本转写。实验表明，该方案在HuggingFace开放ASR排行榜上实现了10.21%的平均词错误率（WER），同时该同一架构作为独立流式ASR模型时也达到了7.73%的优异性能，接近当前最优水平。研究结果验证了该方法在联合建模流式ASR与全双工S2S任务中的高效性与可行性，并已开源代码以推动相关领域的进一步研究。
链接: https://arxiv.org/abs/2609.15759
作者: Ke Hu,Nourchene Ferchichi,Edresson Casanova,Ankita Pasad,Elena Rastorgueva,Chen Chen,Nithin Rao Koluguri,Piotr Zelasko,Yifan Peng,Hainan Xu,Zhehuai Chen,Boris Ginsburg
机构: NVIDIA(英伟达)
类目: Computation and Language (cs.CL)
备注:
Abstract:Full-duplex speech-to-speech (S2S) models enable natural conversational AI by allowing simultaneous listening and speaking. However, these models typically lack inherent user speech transcription, which is essential for applications such as conversation logging, accessibility features, and quality monitoring. In this work, we propose an efficient method to add streaming ASR capabilities to an existing duplex S2S model by introducing a lightweight ASR head in parallel to the agent text head. Our approach requires minimal additional parameters and no significant architectural changes to the base S2S model, enabling real-time user transcription while preserving full-duplex conversational capabilities including turn-taking and barge-in handling. Experimental results demonstrate that our method achieves streaming average WER of 10.21% on the HuggingFace Open ASR Leaderboard within the duplex S2S framework. Additionally, we show that the same architecture trained as a standalone streaming ASR model achieves competitive results (7.73% WER) compared to current SOTA models. We will open-source our training and inference code to facilitate further research in joint streaming ASR and S2S modeling.
[NLP-13] Sequential Adapter Stacking for Cross-Lingual Low-Resource ASR
【速读】： 该论文旨在解决大规模多语言自动语音识别（ASR）模型在低资源语言上性能下降的问题，即模型表现严重偏向高资源语言，而对标注数据和预训练暴露有限的低资源语言识别效果显著恶化。其核心解决方案是提出一种参数高效的知识迁移方法——序列适配器堆叠（Sequential Adapter Stacking），通过在冻结的源语言适配器之上叠加可训练的目标语言适配器，实现从资源丰富语言到低资源目标语言的有效知识迁移。该方法结合了热初始化与基于注意力的融合策略，在三个未被 Whisper 模型支持的语言（阿斯图里亚语、阿萨姆语、科萨语）上进行验证，结果表明，使用最相关源语言时，该方法在所有目标语言上均显著优于全量微调，相对词错误率（WER）降低 5–8%，且仅需一小时目标语言训练数据即可保持显著性能优势，充分体现了其在低资源场景下的有效性与参数效率。
链接: https://arxiv.org/abs/2609.15758
作者: Thai Thi Thanh Thao Dang,Mengjie Qian,Kate Knill
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Extending large-scale multilingual automatic speech recognition (ASR) models to low-resource languages remains challenging. Model performance is skewed toward high-resource languages and degrades sharply for languages with limited labeled data and pre-training exposure. To address this, we investigate parameter-efficient approaches for transferring knowledge from resource-rich source languages to low-resource target languages on Whisper. Alongside warm initialization and attention-based fusion, we propose Sequential Adapter Stacking, which places a trainable target-language adapter on top of a frozen source-language adapter. Under controlled experiments, these approaches are evaluated on three target languages unsupported by Whisper – Asturian, Assamese, and Xhosa – using source languages with varying degrees of relatedness. Sequential Adapter Stacking with the closest related source consistently and significantly outperforms full fine-tuning across the three targets, with 5–8% relative WER reductions. These gains largely persist with only one hour of target training data.
[NLP-14] Look Before You Leap: Factual Decoding with Internal Attribution Signals EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中存在的幻觉（Hallucination）问题，尤其关注在自回归生成过程中早期事实性错误随生成步骤不断累积并呈雪崩式扩大的现象，此类问题难以通过事后修正或权重层面的干预有效预防。其解决方案的关键在于提出一种名为DescaPE（DEcoding Signal Control Against Path Error-snowballing）的解码框架，该框架利用模型内部的隐信号来在推理阶段抑制易产生幻觉的生成路径。通过滑动窗口多层感知机（MLP）消融分析，研究发现模型中存在一个与事实性高度相关的特定层区间，其衍生信号对事实性标记具有选择性增强，并在幻觉高风险步骤表现出异常突增。基于此，作者训练了一个轻量级探测器，仅需一次前向传播即可近似该信号，并将其集成至候选词评分机制中，对高风险延续进行惩罚，同时奖励基于事实的生成路径。实验结果表明，DescaPE在三个主流大模型上的五个事实性基准测试中均显著提升了生成事实性，且在效率评估中仅带来1.10倍的延迟开销，展现出良好的实用性与有效性。
链接: https://arxiv.org/abs/2609.15745
作者: Hayeong Ryu,JungMin Yun,Byeonggeuk Lim,Sunhee Jo,YoungBin Kim
机构: Chung-Ang University (中央大学); Graduate School of Advanced Imaging Sciences, Multimedia and Film, Chung-Ang University (中央大学高级影像科学、多媒体与电影研究生院); Department of Artificial Intelligence, Chung-Ang University (中央大学人工智能系)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Findings
Abstract:Hallucination remains a critical challenge in large language models (LLMs), where early factual errors compound through autoregressive generation in a snowballing effect that neither post-hoc correction nor weight-level intervention can effectively preempt. We propose DescaPE (DEcoding Signal Control Against Path Error-snowballing), a decoding framework that leverages internal model signals to suppress hallucination-prone trajectories at inference time. Through sliding-window MLP ablation, we identify a factual-salient layer span within LLMs whose derived signal is selectively elevated for factual tokens and exhibits anomalous spikes at hallucination-prone steps. We train a lightweight probe to approximate this signal from a single forward pass and integrate it into candidate scoring to penalize high-risk continuations while rewarding factually grounded ones. Experiments across five factuality benchmarks on three LLMs demonstrate that DescaPE achieves factuality improvements over decoding-time baselines in multiple settings, while incurring only 1.10x latency overhead in our efficiency evaluation. Our code is available at this https URL.
[NLP-15] Merging the Knowledge of LLM s for Automatic Speech Recognition INTERSPEECH2026
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）在生成式语音识别（Generative AI-based Automatic Speech Recognition, ASR）系统中应用时带来的额外计算开销问题。传统基于浅融合（shallow fusion）或密度比（density ratio）的外部语言模型融合方法虽有效，但需在解码阶段进行额外的语言模型推理，导致计算成本显著增加，尤其对参数量庞大的现代大语言模型更为严峻。为此，本文提出一种通过模型合并（model merging）的方式将外部语言模型直接整合到基于大语言模型的ASR模型参数中，从而在推理阶段无需额外计算开销。其核心创新在于利用低秩适应（LoRA）参数的算术运算实现领域扩展与迁移，使语言模型知识以参数融合形式嵌入主模型，既保持了原有模型的推理速度与内存占用，又在目标领域上实现了稳定的语音识别性能提升。
链接: https://arxiv.org/abs/2609.15743
作者: Hayato Futami,Tatsuya Kawahara
机构: 未知
类目: Computation and Language (cs.CL)
备注: Accepted to Interspeech2026
Abstract:Automatic speech recognition (ASR) systems, trained on paired speech-text data, have been improved by leveraging language models (LMs) trained on text-only data. LM fusion methods such as shallow fusion and density ratio are well-established methods that incorporate external LMs during ASR decoding. However, they incur additional computational costs due to LM inference, which is particularly problematic for recent larger LMs. In this study, we propose incorporating external LMs via model merging. This method integrates the LMs directly into the parameters of an LLM-based ASR model, requiring no additional computational cost at inference. We formulate domain extension and transfer via arithmetic operations on LoRA parameters. Experimental evaluations were conducted for the domain adaptation of LLM-based ASR trained on CSJ and LibriSpeech. We show that our LM merging consistently improved the ASR performance in the target domains, without degrading inference speed or memory footprint.
[NLP-16] RESKILL: Explicit Failure Attribution and Structured Repair for Interactive Language Agents EMNLP2026
【速读】： 该论文旨在解决语言智能体在任务失败后进行修复时，现有方法依赖于不透明的一次性反思（one-shot reflection）所导致的修复过程缺乏可追溯性与迭代优化能力的问题。其核心挑战在于：失败原因解释与候选修复方案之间的关联关系未被显式维护，且失败重测结果未能有效反馈至后续修复步骤中。为此，论文提出一种结构化的修复框架RESKILL，其关键在于通过显式维护跨修复轮次的修复状态（repair state），实现对失败假设、候选修复方案、局部修复选择及重测结果的系统化追踪与整合。具体而言，RESKILL基于覆盖度归因（coverage-based attribution）筛选局部修复策略，并利用环境中的重测结果动态指导后续修复更新，使语言模型生成的结构化修复因子得以持续积累与演化。实验表明，在ALFWorld和TextCraft两个基准上，无论何种模型规模，RESKILL均显著优于直接修复与假设条件修复基线，平均最终成功率分别提升3.7和3.3个百分点，验证了将归因分析与修复选择、持久性重测反馈相结合对于实现可持续性能提升的关键作用。
链接: https://arxiv.org/abs/2609.15684
作者: Mengyi Deng,Xin Li,Duyi Pan,Zilin Wang,Zhiwei Li,Zhijiang Guo,Wei Wang
机构: The Hong Kong University of Science and Technology (Guangzhou), China; The Hong Kong University of Science and Technology, Hong Kong SAR
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 (Main Conference)
Abstract:Language agents increasingly rely on reusable skills, but post-failure repair is often handled by opaque one-shot reflection: a model generates a skill patch without explicitly maintaining how failure explanations relate to candidate repairs or how unsuccessful retests should influence later edits. We introduce RESKILL, a structured repair framework that maintains an explicit repair state across repair rounds. Given a failed rollout, the framework links failure hypotheses to candidate skill patches, selects local repairs through coverage-based attribution, retests the edited skill set in the environment, and uses retest outcomes to guide subsequent repair updates. The language model supplies structured repair factors, while the repair procedure records them, compares local skill patches by how well they address active failure explanations, and carries unsuccessful retest outcomes into later repair rounds. We evaluate RESKILL on ALFWorld and TextCraft across three model sizes under fixed repair budgets. RESKILL obtains the strongest final success in all six benchmark-model settings, improving average final success by 3.7 percentage points over direct repair and 3.3 points over hypothesis-conditioned repair. These results suggest that explicit attribution alone is insufficient; durable improvement emerges when attribution is integrated with repair selection and persistent retest-conditioned update.
[NLP-17] CiteShade: Citation Laundering in Multi-Source Retrieval-Augmented Generation and Its Counterfactual Defense
【速读】： 该论文旨在解决生成式 AI（Generative AI）中检索增强生成（Retrieval-Augmented Generation, RAG）系统在安全方面的一个关键漏洞：攻击者可通过操纵引用来源（citation）实现“引用洗白”（citation laundering），即诱导模型生成错误答案并将其归因于可信但不支持该结论的外部源，从而破坏用户对模型输出可审计性的信任。其解决方案的关键在于揭示了引用渠道作为新型攻击面的可行性，并提出首个系统性攻击方法 CiteShade，通过构建满足检索、生成与引用三重必要条件的恶意源，在无需指令的情况下即可实现对答案的误导性引用。实验表明，在多源多跳问答任务中，该攻击将错误回答率从0.01提升至0.68，且源删除实验证实恶意源为因果驱动因素。研究进一步发现，传统防御手段如困惑度过滤和引用支持检查均不足以应对该威胁，因而提出反事实防御机制，通过验证实际驱动答案生成的来源来检测并抵御此类攻击。
链接: https://arxiv.org/abs/2609.15660
作者: Guo Fuzheng
机构: 未知
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL)
备注: 17 pages, 4 figures, 18 tables
Abstract:Retrieval-augmented generation (RAG) grounds a language model’s answers on retrieved external knowledge and returns each answer with citations that identify its sources. Those citations are the user’s audit trail: they let a reader verify a claim without trusting the model. Prior security work on RAG asks whether an attacker can corrupt the answer, leaving the citation channel unexplored. We show that this channel is a new and practical attack surface. We propose CiteShade, the first citation laundering attack to RAG, in which an attacker controlling a single source induces a model to produce an attacker-chosen wrong answer and to attribute it to a trusted source that does not support it, while the evidence for the correct answer remains in context. We formulate the attack as an optimization problem, derive three necessary conditions (retrieval, generation, and citation) and construct sources satisfying them without any instruction. On multi-source multi-hop question answering the attack raises the wrong-answer rate from 0.01 to 0.68, and source deletion confirms the malicious source is the causal driver in every measured case. Vulnerability tracks a model’s propensity to cite rather than its scale, reaching CLR 0.84 under explicit instruction and 0.64 with no instruction at all on the most citation-prone model tested. We then show that perplexity filtering and citation-support checking are each insufficient, and propose a counterfactual defense that verifies which source actually drove the answer.
[NLP-18] Empathy Is Steerable but Multi-Axial: Mechanism Geometry and Persona Effects in LLM s EMNLP2026
【速读】： 该论文旨在解决生成式对话系统中支持性共情（supportive empathy）的多维度可调控性问题，即如何在不相互干扰的前提下对共情的不同子维度进行独立控制。其核心挑战在于，支持性共情由情感反应（Emotional Reactions）、理解解释（Interpretations）和探索回应（Explorations）等多个潜在维度构成，而现有方法难以实现这些维度的解耦干预。论文提出基于EPITOME框架的分析范式，将共情分解为上述三个可量化维度，并通过对比激活添加（contrastive activation addition）方法在中间层实现稳定干预。关键发现是：尽管所恢复的激活方向在跨模型间表现出一致的代理分数迁移效果，实现了从响应层面评估向表征层面干预的跃升，但这些方向仅部分可分离——单一方向的操控会引发非目标维度的偏移；此外，人工设计的角色提示（persona prompts）虽显著改变共情表现，但其引起的激活变化中仅有约3%可被恢复的子空间解释，表明角色条件化共情的调控需超越单个机制方向，依赖更复杂的结构建模。因此，解决方案的关键在于采用分层、多维度的表征分析框架与对比激活干预策略，推动共情控制从“整体调节”迈向“精细化、多轴向”调控。
链接: https://arxiv.org/abs/2609.15654
作者: JuHeon Ha,Byounghan Lee,Yunseo Choi,Kyung-Ah Sohn
机构: Ajou University (AJOU 대학교)
类目: Computation and Language (cs.CL)
备注: 18 pages, 6 figures. Accepted to the Main Conference of EMNLP 2026
Abstract:Activation steering has been used to control traits such as honesty, refusal, and sycophancy, yet supportive empathy is evaluated along multiple dimensions that need not correspond to independently controllable activation directions. Using the EPITOME framework, which decomposes supportive empathy into Emotional Reactions, Interpretations, and Explorations, we study three instruction-tuned LLMs and ask whether candidate directions derived from these labels produce distinguishable intervention effects or instead share structure, and how persona prompts interact with those directions. We find that contrastive activation addition yields a stable middle-layer intervention that consistently shifts the EPITOME proxy scores across models, moving empathy analysis beyond response-level scoring. However, the recovered directions are only partially separable: steering one direction induces off-target shifts, and hand-crafted prompting shifts the empathy profile rather than isolating a single dimension. Persona prompts substantially change EPITOME scores, but a paired activation-shift decomposition shows that the recovered subspace captures only approximately 3 percent of persona-induced squared activation-shift magnitude at layer 15. Under this EPITOME-based definition, expressed empathy is steerable but multi-axial, and controlling persona-conditioned empathy requires targeting structure beyond individual mechanism directions.
[NLP-19] Human-Grounded Calibration for Long-Text Image-Text Congruence in Vision-Language Models
【速读】： 该论文旨在解决长文本图像-文本一致性评分（long-text image–text congruence scoring）中，双编码器模型输出的原始相似度得分难以作为可解释且校准一致的匹配度量问题，尤其在图像与文本嵌入之间的模态鸿沟（modality gap）背景下。其核心解决方案是提出一种轻量级校准层——一致性得分（Congruency Score, CS），将图像-文本相似性证据映射至有界分数空间，从而实现对匹配程度的可校准量化。关键在于，该方法通过引入可操作的校准机制，显式分离了检索性能、人类判断关联性与阈值校准三个独立目标，揭示了现有后处理校准策略在提升阈值鲁棒性的同时可能损害检索表现与人类感知相关性的权衡关系，为长文本场景下的跨模态一致性评估提供了可解释、可调优的校准框架。
链接: https://arxiv.org/abs/2609.15640
作者: Alessandro Gambetti,Qiwei Han
机构: Nova School of Science and Technology (NOVA School of Science and Technology), Universidade NOVA de Lisboa (Universidade NOVA de Lisboa); Nova School of Business and Economics (Nova School of Business and Economics), Universidade NOVA de Lisboa (Universidade NOVA de Lisboa)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 3 figures, 2 tables
Abstract:Long-text image–text congruence scoring is increasingly important for vision-language systems that must evaluate whether detailed textual descriptions match visual content. However, raw similarity scores from dual-encoder models are difficult to interpret as calibrated congruence measures, especially under the modality gap between image and text embeddings. This paper proposes Congruency Score (CS), a lightweight calibration layer that maps image–text similarity evidence into a bounded score. Using DOCCI and Urban1k, we evaluate four frozen vision-language backbones and show that observed reductions in post-projection centroid distance do not uniformly improve image–text retrieval performance. Human-grounded evaluations on DOCCI further reveal a trade-off: direct post-hoc calibration preserves high association with human judgments, whereas selected projection-based configurations can reduce threshold-relevant slope and intercept distortions at the cost of retrieval performance and association strength. These results establish long-text image–text congruence scoring as a calibrated score-estimation problem, where retrieval performance, human association, and threshold calibration must be evaluated as distinct objectives. CS provides a lightweight way to expose and operationalize this separation.
[NLP-20] hrough the Eyes of the Beholder: Biometric and Demographic Conditioning for Multimodal Sexism Detection
【速读】： 该论文旨在解决网络性别歧视（sexism）检测中因主观性过强而导致的评估与模型泛化难题。传统方法往往忽略人类标注者在判断过程中的心理与人口统计特征，导致模型难以捕捉语义理解中的细微差异。为此，研究提出一种以人为中心的多模态框架，其核心在于将标注者的心理及人口学特征融入检测流程，以建模并缓解主观偏差。关键创新在于：通过交叉注意力架构结合特征逐通道调制（Feature-wise Linear Modulation），融合文本、图像、表情包描述、生理信号及风格特征五种模态；针对子任务2.1（标签分布学习），采用基于Kullback-Leibler散度的损失函数优化全标注者标签分布；推理阶段则采用深度多模态网络与基于文体特征和生理信号的SVM进行软投票（soft-voting）融合。实验表明，该方法在软评价下于子任务2.2（源意图识别）中排名29/114，且在子任务2.1和2.2上归一化ICM得分均优于基线，验证了人因建模的有效性。研究开源完整流程，支持可复现的人为中心建模。
链接: https://arxiv.org/abs/2609.15608
作者: Ana-Maria Luisa Mocanu,Sebastian Mocanu,Ciprian-Octavian Truică,Elena-Simona Apostol
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Detecting sexism on the internet is a fundamentally subjective task; our team, VANGUARD, addresses this challenge in the EXIST 2026 Task 2 by proposing a human-centered multimodal framework that analyses and incorporates the psychological and demographic characteristics of human annotators into the detection pipeline. We fuse five input modalities through a cross-attention architecture with Feature-wise Linear Modulation conditioning. Meme text is extracted and visually described with Gemma 4, then augmented by automatic translation between English and Spanish with NLLB-200. Text and image representations are produced by LoRAadapted XLM-RoBERTa and CLIP encoders and fused with sensor features encoded by a pretrained autoencoder. To model annotator subjectivity, we frame Subtask 2.1 as a label distribution learning problem, optimizing a Kullback-Leibler divergence loss over the full annotator label distribution. At inference time, predictions are produced by soft-voting between the deep multimodal network and a complementary SVM trained on stylometric and physiological features. Our best submission ranks 29th out of 114 on Subtask 2.2 (source intention) under soft evaluation, and the normalized ICM scores remain above the baseline on Subtasks 2.1 and 2.2, indicating that annotator-centered conditioning contributes a usable signal. We release our full pipeline and analysis to support reproducible human-centered modeling.
[NLP-21] Can We Trust the Judges? Validation of Factuality Evaluation Methods via Answer Perturbation
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）事实性评估工具的可信度问题，即当前广泛使用的基于事实性的评估指标在敏感性和可靠性方面仍缺乏充分验证。其核心挑战在于：现有评估方法是否能准确捕捉生成内容中细微的事实偏差？为此，论文提出一种元评估（meta-evaluation）框架，通过系统性地对标准答案施加受控的语义破坏（controlled corruptions），生成具有已知退化程度的排序输出，从而检验各类评估指标对事实性变化的感知能力。研究发现，基于流水线（pipeline-based）的方法（如RAGAS的事实性度量）相较于以大模型为评判者的“大模型作为裁判”（LLM-as-judge）方法，在追踪事实性退化方面表现更优。此外，论文还提出了一种新型事实性度量变体，兼具竞争力与计算效率，为高效、可靠的评估提供了新路径。
链接: https://arxiv.org/abs/2609.15561
作者: Sarra Gharsallah,Adele Robaldo,Mariia Tokareva,Giovanni Gatti Pinheiro,Ilyana Guendouz,Raphaël Troncy,Paolo Papotti,Pietro Michiardi
机构: EURECOM(欧洲电信学院); Data Science Department, EURECOM(数据科学系，欧洲电信学院), Campus SophiaTech(索菲亚科技园区), Biot(比奥), France(法国)
类目: Computation and Language (cs.CL)
备注: 22 pages, 7 figures. Extended version of a paper accepted at EvalLLM 2025 (CORIA-TALN 2025)
Abstract:Evaluating the factual correctness of large language models (LLMs) is vital for many applications. But are our evaluation tools themselves trustworthy? Despite the rise of factuality-based metrics, their sensitivity and reliability remain underexplored. This paper introduces a meta-evaluation framework that systematically tests these metrics using controlled corruptions of gold standard answers. Our method generates ranked outputs with known degrees of degradation to probe how metrics capture nuanced changes in truthfulness. Our experiments reveal that pipeline-based methods, such as the RAGAS’s factual correctness metric, better track degradation than LLM-as-judge approaches. We also propose a new variant of the factual correctness metric that provides a competitive and cost-efficient.
[NLP-22] Dont Count the Edits Judge by the Outcome Alone: Reward-Based Evaluation for Grammatical Error Correction EMNLP2026
【速读】： 该论文旨在解决语法错误纠正（GEC）评估中依赖参考答案或编辑重叠所带来的局限性，尤其是对与标准修正结果存在差异但语义正确且语言流畅的改写文本产生不公平惩罚的问题。现有无参考评估方法虽缓解了对参考答案的依赖，但仍难以准确判断生成输出是否为源句的有效修正。为此，论文提出SURE——一种基于源句条件的奖励评估器，其通过在涵盖最小修改与重写型修正的源内偏好数据上进行训练，实现了对语法正确性、忠实度和流畅性的多维度联合建模，并引入了针对源端错误定位的细粒度片段级归因机制。其核心创新在于结合整体奖励学习与层级化判别监督，从而在SEEDA基准上展现出优于强基线的表现，尤其在重写类修正任务中优势显著，并能提供更解耦的各维度诊断能力。
链接: https://arxiv.org/abs/2609.15559
作者: Hayeong Ryu,Sunhee Jo,Seunguk Yu,YoungBin Kim
机构: Chung-Ang University (中央大学); Graduate School of Advanced Imaging Sciences, Multimedia and Film, Chung-Ang University (中央大学高级影像科学、多媒体与电影研究生院); Department of Artificial Intelligence, Chung-Ang University (中央大学人工智能系)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Findings
Abstract:Grammatical error correction (GEC) evaluation has traditionally relied on reference or edit overlap, which can penalize valid rewrites that differ from gold corrections. Reference-free metrics reduce this dependence, but evaluating whether a fluent output is a valid correction of the source remains challenging. We propose SURE, a source-conditioned reward evaluator trained on within-source preferences spanning minimal-edit and rewrite-oriented corrections. SURE jointly learns an overall reward with criteria-level supervision for grammaticality, faithfulness, and fluency, together with span-level grounding for source-side error resolution. Experiments on SEEDA show that SURE performs competitively against strong baselines, with particular gains on rewrite-style corrections and more disentangled criteria-level diagnostics. Our code is available at this https URL.
[NLP-23] Option-Aware Retrieval and Task-Specific VLM Adaptation for Medical VQA
【速读】： 该论文旨在解决医学视觉问答（Medical VQA）任务中，如何在完全离线、容器化推理环境下提升多选题（MCQ）与开放性问题（OE）的准确率问题。其核心挑战在于：传统基于答案标签（answer labels）的检索方法无法有效传递知识，因为标签是独立分配的，直接复制邻近样本的标签不具语义价值。解决方案的关键在于采用语义级匹配策略——将当前选项文本与相似训练样本中的正确答案文本进行语义对比，从而显著提升检索性能，使仅依赖检索的准确率从20.0%提升至57.5%。此外，研究发现，固定任务特定的低秩适配器（MCQ LoRA adapter）后，调整提示中检索示例数量 k 对结果影响极小，表明模型已充分吸收知识；引入置信度门控机制虽能提升选择一致性（198/200），但未带来净准确率增益。在开放性问题上，尽管随 k 增大，指标如token-F1和RaTEScore下降，但统计检验显示差异不显著，且主要错误源于锚点误标。最终系统在开发集上达到94.0%的MCQ准确率，在官方预评中为93.20%，远超基准模型（29.43%），而开放性任务得分略低于基线，反映出当前生成质量仍存在改进空间。
链接: https://arxiv.org/abs/2609.15530
作者: Tristan Kirscher(ICube, Institut Strauss),Niklas C. Koser(CAU),Soren Pirk(CAU)
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We describe our submission to the MedReason 2026 challenge, covering multiple-choice (MCQ) and open-ended (OE) medical visual question answering (VQA) under fully offline, containerized inference. Our first finding is that MCQ retrieval must compare answer \emphsemantics rather than answer labels: labels are independently assigned per question, so copying a retrieved neighbor’s label transfers no useful information, whereas scoring each current option’s text against correct-answer text from similar training cases raises retrieval-only accuracy from 20.0% to 57.5% on a 200-case retrieval-excluded development holdout. Our second finding attributes the submitted system’s accuracy: holding the task-specific MCQ Low-Rank Adaptation (LoRA) adapter fixed and varying the number (k) of in-prompt retrieved examples changes accuracy by at most one case — 187/200 (93.5%) at both (k=0) and the adapter’s training-time (k=1), 188/200 (94.0%) at the packaged runtime’s default (k=3) — and the submitted confidence-gated override adds no net accuracy on top of (k=3), selecting the VLM in 198/200 cases. With the final MCQ adapter fixed, retrieval changes accuracy by at most one case, and gating provides no net gain. On 20 OE cases, token-F1 and RaTEScore~\citezhao2024ratescore decrease as (k) grows, but paired sign tests on token-F1 differences are nonsignificant ((p \ge 0.29)); a single-annotator comparison found 6/20 wrong-anchor errors for the final configuration and 14/20 for an earlier configuration that jointly differed in routing, adapter, and prompting. The system reaches 94.0% MCQ accuracy on the development holdout and 93.20% on the organizer’s official pre-evaluation, versus 29.43% for the off-the-shelf reference baseline, while both of the organizer’s open-ended scores are lower than that baseline’s (ground-truth agreement 1.245 versus 1.588, visual accuracy 1.995 versus 2.696, each out of 4).
[NLP-24] o Each Language Its Tokenizer: Modular Tokenizers for Efficient Multilingual LLM s
【速读】： 该论文旨在解决多语言大语言模型（Multilingual Large Language Models, LLMs）在词汇共享、内存占用和推理效率方面的核心问题。传统方法采用单一共享词表，导致不同语言间压缩率不均，且庞大的嵌入与输出矩阵显著增加内存开销并降低推理速度，尤其对小规模模型影响严重；此外，模型常仅用于部分语言，造成资源浪费。为此，论文提出一种模块化多语言模型训练框架：其关键在于设计可学习的大规模模块化BPE与Unigram分词器，能够针对任意语言子集提取专用的子分词器，实现与单语分词器相当的压缩效果，并提升跨语言公平性；同时，提出基于子分词器采样的预训练策略，通过限制预测范围至相关词汇子集，使模型在保持大词汇量的前提下仍能高效训练，支持任意语言组合下的高效推理。该方案有效降低了内存消耗，加速了推理过程，且无需牺牲模型性能。
链接: https://arxiv.org/abs/2609.15528
作者: Franck Signe,Hippolyte Pilchen,François Yvon,Édouard Grave
机构: Kyutai(巴黎); Sorbonne Université (巴黎索邦大学), CNRS, Institute of Intelligent Systems and Robotics (智能系统与机器人研究所), Paris; Univ. Grenoble Alpes (格勒诺布尔阿尔卑斯大学), CNRS, Grenoble INP, LIG (信息与图形实验室), Grenoble
类目: Computation and Language (cs.CL)
备注:
Abstract:Multilingual Large Language Models (LLMs) traditionally rely on a single vocabulary shared by all supported languages, which can lead to uneven compression across them. Moreover, their large embedding and output matrices increase memory usage and slow inference, notably for small-scale models. It is also wasteful as models are often used for only a subset of languages. To address these issues, we introduce a modular framework for multilingual model training. First, we propose methods to learn large modular BPE and Unigram tokenizers that enable extraction of subtokenizers tailored to any language subset. These subtokenizers achieve compression on par with monolingual tokenizers and improve cross-lingual fairness. Second, we design a pretraining strategy that samples subtokenizers to form batches, restricting predictions to the relevant vocabulary subset and allowing efficient training despite a large vocabulary. This supports efficient inference with any combination of language-specific vocabularies. Therefore, it reduces memory usage and speeds up inference in models without sacrificing performance.
[NLP-25] Psychosis involves a deficit of information compression in connected speech
【速读】： 该论文旨在解决精神分裂症谱系障碍（schizophrenia-spectrum disorders, SSD）患者在语言信息处理中是否存在普遍性信息压缩缺陷的问题，尤其关注其与语法组织能力之间的关联。研究发现，尽管患者在词汇频率预测的可预测性（即基于词频的预期意外度）上表现正常，但其在上下文语言模型（contextual LM）基础上计算的预期意外度差异（surprisal difference）显著降低，表明患者对语法结构所赋予的预测信息敏感性减弱。这一现象揭示了精神分裂症患者在语言表征层面存在一种深层的信息压缩功能障碍，其关键机制在于语法组织能力受损，导致语言系统无法有效利用句法结构进行高效预测。研究通过144名土耳其语使用者的数据验证了该假设：所有临床组（包括慢性精神分裂症、首次发病精神病及分裂情感障碍）均表现出更小的预期意外度差异，且慢性精神分裂症和首次发病精神病患者的内在维度（intrinsic dimension, ID）降低，提示语义空间的可压缩性下降；同时，句法复杂度与可压缩性共同显著预测预期意外度差异。这些结果进一步证实了精神分裂症中语义空间几何结构的异常，并将此异常归因于语法操作机制的失调，从而为理解该疾病中的认知缺陷提供了新的神经语言学视角。
链接: https://arxiv.org/abs/2609.15522
作者: Samuele Vallisa,Claudio Palominos,Rui He,Emre Bora,Burcu Verim,Cemal Demirlek,Berna Yalincetin,Philipp Homan,Wolfram Hinzen
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) with human-like performance on linguistic tasks have transformed the study of language in neurodiverse conditions. LLMs provide representations of linguistic input in the form of high-dimensional vectors (embeddings), and next-token predictions computed from these embeddings. Previous crosslinguistic evidence suggests a complexity reduction in the form of both lower intrinsic dimensionality (ID) of LLM representations and higher mean surprisal (prediction error) in psychosis. We hypothesized that these metrics reflect a general deficit of information compression in psychosis, linked to grammatical organization as what enables predictions in this http URL operationalized surprisal difference as the difference between surprisal as estimated from word frequency and surprisal as based on a contextual LM, which is sensitive to grammatical organization over and above lexical concepts. Using a dataset of 144 Turkish speakers, including 106 patients with schizophrenia-spectrum disorders (SSD) - 56 with chronic schizophrenia (SZH), 33 with first-episode psychosis (FEP), and 17 with schizoaffective disorder (SZA) - and 38 healthy controls. We report: (1) Surprisal difference is attenuated in all clinical groups relative to controls, independently of word count; (2) Compressibility (intrinsic dimension) is reduced in SZH and FEP; (3) Syntactic complexity and compressibility both predict surprisal difference. These results, further refining an alteration in the geometry of the semantic space in psychosis as previously attested, suggest a broader deficit in information compression in this disorder, with a mechanistic underpinning in the operations of grammar.
[NLP-26] Beyond Safe Answers: Segment-Aware Listwise Alignment for Reasoning Safety in Large Reasoning Models EMNLP2026
【速读】： 该论文旨在解决大型推理模型（Large Reasoning Models, LRMs）中存在的双重安全风险问题，即中间推理过程与最终输出答案均可能包含有害内容。现有对齐方法通常在完整响应层面进行，导致不安全的推理过程可能被看似安全的最终答案所掩盖。为此，论文提出一种分段感知的列表级目标去偏方法（Segment-aware Listwise Target DPO, SaLT-DPO），其核心解决方案包含三个关键机制：（1）分段感知的列表级对齐，将响应分解为推理与答案两个分段，分别独立评估其安全性，并通过长度归一化的分段奖励与多候选样本上的软目标分布对齐；（2）联合安全一致性正则化，采用最弱环节原则（weakest-link principle）确保推理与答案分段间安全性的高度一致；（3）在良性提示上引入效用锚定，以缓解过度拒绝和推理能力退化问题。实验结果表明，SaLT-DPO在三种LRM上均显著降低推理与答案分段的不安全率，同时有效维持良性任务的合规性与通用推理性能，消融实验进一步验证了各组件间的互补性。
链接: https://arxiv.org/abs/2609.15517
作者: JungMin Yun,Junehyoung Kwon,Hayeong Ryu,Byeonggeuk Lim,Hoejoon Kwon,YoungBin Kim
机构: Chung-Ang University (中央大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: EMNLP 2026 Main Conference
Abstract:Large Reasoning Models (LRMs) pose a dual-surface safety challenge: both intermediate reasoning traces and final answers can contain harmful content. Existing alignment methods often operate at the whole-response level, allowing unsafe reasoning to be masked by a safe-looking final answer. We propose Segment-aware Listwise Target DPO (SaLT-DPO), which addresses this gap through three mechanisms: (1) segment-aware listwise alignment that decomposes responses into reasoning and answer segments, independently scores each segment’s safety, and aligns length-normalized segment rewards with soft target distributions over multiple candidates; (2) joint safety coherence regularization that applies a weakest-link principle to promote safety consistency across both segments; and (3) utility anchoring on benign prompts to mitigate over-refusal and reasoning degradation. Experiments on three LRMs show that SaLT-DPO consistently reduces unsafe rates for both reasoning and answer segments while mitigating degradation in benign compliance and preserving general reasoning performance. Ablation studies demonstrate the complementary contributions of its components.
[NLP-27] Authorship attribution and aesthetic evaluation of AI poetry: a case study with Haiku
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在日语俳句创作中所产生的作者身份识别难题，重点关注人类对AI生成与人类创作俳句的审美判断与归属感知差异。其核心问题在于：当生成内容在形式上高度符合传统俳句的严格约束时，人类是否仍能有效区分其作者属性？解决方案的关键在于采用少量示例提示（few-shot prompting）策略，在多种异构大语言模型（LLM）上生成俳句，并通过在日本东京高校学生中开展问卷调查，系统评估人类对生成作品的辨识能力及其判断依据。研究发现，尽管部分模型（如GPT-5、Gemini 2.5）的生成结果接近随机水平（准确率约0.50），而其他模型（如LLM-JP、Gemma-2B、LLaMA-2）则表现出中等可检测性（0.59–0.67），但识别效果高度依赖具体样本，且美学评价维度（如流畅性、连贯性、诗意性）虽显著影响“人为性”感知，却无法预测真实分类准确性，揭示了审美判断与作者身份识别之间的解耦现象。这表明，随着大语言模型在形式约束下的创作能力提升，其生成内容在表层创意合理性上的逼近将削弱人类基于直觉的可靠辨别能力。
链接: https://arxiv.org/abs/2609.15511
作者: Livia Oddi,Simone Scardapane,Toru Sugimoto,Donatella Genovese
机构: Sapienza University of Rome (罗马大学); Shibaura Institute of Technology (信州大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 18 pages, 10 figures. Accepted as a full paper at the 2026 Joint 14th International Conference on Soft Computing and Intelligent Systems and 27th International Symposium on Advanced Intelligent Systems (SCISISIS 2026), November 2-5, 2026
Abstract:This paper investigates the generation and human evaluation of Japanese haiku by contemporary Large Language Models (LLMs), focusing on authorship perception and aesthetic judgment within a constrained poetic form. Using a few-shot prompting strategy, Japanese haiku were generated across a heterogeneous set of large language models, including open- and closed-source systems, medium-scale and large-scale architectures, models with native or adapted Japanese support, and multilingual proprietary models. These AI-generated haiku were combined with human-written ones and presented in a questionnaire distributed to students at Japanese universities in Tokyo. The survey assessed whether respondents could distinguish between AI-generated and human-written haiku and which cues informed their judgments. Recognition accuracy varied across models. GPT-5, Gemini 2.5, and StableLM-7B performed at approximately chance level (approx 0.50), whereas LLM-JP, Gemma-2B, and LLaMA-2 showed moderate detectability (approx 0.59-0.67). However, recognition was strongly item-dependent. Ratings of fluency, coherence, poeticness, and related aesthetic dimensions predicted perceived humanness but not correct classification, indicating an attribution bias linked to aesthetic evaluation and revealing a dissociation between aesthetic evaluation and true authorship detection. The extended analysis additionally examines generation-constraint adherence, participant-level characteristics, and exploratory LLM-based evaluations of haiku authorship. Overall, the findings suggest that as LLMs improve, surface-level creative plausibility may reduce reliable human discrimination within constrained poetic settings.
[NLP-28] How Lossless Is Lossless Speculative Decoding? The Role of Numerical Precision in Orthrus
【速读】： 该论文旨在解决自回归语言模型（autoregressive language model）推理效率低下的问题，其核心目标是通过并行生成多个标记（token）来加速推理过程，同时保持与原始自回归模型输出完全一致的生成轨迹。为此，论文提出Orthrus——一种融合自回归与扩散机制的混合架构，依赖于模型内部一致性机制（intra-model consensus mechanism）实现无损推测解码（lossless speculative decoding）。然而，研究发现该方法声称的“无损”特性在实际中高度依赖数值精度：在半精度浮点数（BF16）下，仅约43%-45%的测试用例能实现与自回归模型完全一致的生成轨迹，且匹配概率与参考模型在响应条件下的困惑度（perplexity）显著相关；而在全精度浮点数（FP32）下，所有测试提示均实现了精确轨迹匹配。这表明Orthrus的实际无损性并非普适成立，其轨迹等价性必须独立于下游任务性能进行评估。因此，该研究的关键结论在于：所谓“无损”的实现依赖于数值精度，且需对生成轨迹进行独立验证，不能仅以下游基准测试表现作为判断依据。
链接: https://arxiv.org/abs/2609.15504
作者: Ilya Koziev,Leonid Sinev,Ivan Oseledets
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: under review
Abstract:Orthrus is a hybrid autoregressive-diffusion architecture that accelerates autoregressive language-model inference by generating multiple tokens in parallel while using a frozen autoregressive backbone. Its central claim is that an intra-model consensus mechanism enables lossless speculative decoding, producing the same output sequence as the autoregressive model. We independently reproduce Orthrus and examine this claim under different numerical precisions. Under BF16 inference, exact trajectory matching occurs in only 45% of cases for the authors’ checkpoint and 43% for our independently trained model across 1,190 prompts from 12 domains. The probability of exact matching is also strongly associated with the response-conditional perplexity of the reference model. Despite this trajectory divergence, Orthrus does not show systematic degradation on downstream lm-eval-harness benchmarks. In contrast, repeating the trajectory evaluation with FP32 yields exact trajectory matching on all evaluated prompts. These results show that the practical losslessness of Orthrus depends on numerical precision and that exact trajectory equivalence should be evaluated separately from downstream task performance. Comments: under review Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.15504 [cs.CL] (or arXiv:2609.15504v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.15504 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-29] mperature Frag ility and the Conditional Benefits of Truncation Sampling
【速读】： 该论文旨在解决在典型部署温度范围（0.6–1.0）内，截断采样器（如top-p和min-p）是否仍能提升生成式AI（Generative AI）模型性能的问题。以往研究显示，这些采样器在较高温度（1.5–3.0）下可显著提升准确性，但其在实际应用中常用的低温度区间的效果尚未得到系统评估。本文通过在一个受控管道中对13个开源模型在GSM8K和MMLU-Pro基准上进行测试，覆盖温度0.7、1.0和1.3，并在其中10个模型上尝试八种解码配置，发现当温度从0.7升至1.3时，六种模型在MMLU-Pro上的准确率下降17至38个百分点，其余七种模型最多下降10个百分点。性能下降主要源于生成文本达到最大标记数限制或未能给出明确答案。研究结果表明，截断采样器的增益主要体现在高温导致模型性能显著退化的情况下；而在温度变化对准确性影响较小的场景中，所测试的截断采样器均未优于基础温度采样策略。因此，其核心解决方案的关键在于：截断采样器的有效性依赖于高温引发的性能劣化程度，而非温度本身，且在常规部署温度下不具备普遍提升效果。
链接: https://arxiv.org/abs/2609.15476
作者: Francesco La Rosa
机构: University of Edinburgh(爱丁堡大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 20 pages, 8 figures, 16 tables. Code and data: this https URL
Abstract:Large language models generate text by sampling each token from a predicted distribution, and a temperature parameter sets how far the draw strays from the most probable tokens. Truncation samplers such as top-p and min-p discard the least probable tokens before the draw, so that sampling at high temperature stays coherent. Their reported accuracy gains come from temperatures of 1.5 to 3, while the defaults of deployed systems cluster between 0.6 and 1.0. Whether they change accuracy at those defaults, and for which models, has not been measured. We test thirteen open-weight models on GSM8K and MMLU-Pro at temperatures 0.7, 1.0, and 1.3 in one controlled pipeline, ten of them under eight decoding configurations. Six of the thirteen models lose 17 to 38 accuracy points on MMLU-Pro between 0.7 and 1.3, and the other seven lose at most 10. The lost accuracy comes from generations that run to the token limit or never state an answer. These results suggest that truncation samplers improve accuracy primarily when higher temperatures substantially degrade model performance. Where accuracy remains stable across temperatures, none of the tested truncation samplers improves on plain temperature sampling.
[NLP-30] urkish MMLU Pro: Traceable Option Augmentation and Its Validity Limits in Turkish Multiple-Choice Evaluation
【速读】： 该论文旨在解决多选题中增加选项数量可能导致分数下降但未提升评估有效性的核心问题，即探讨在不增强测评信度与效度的前提下，扩充选项是否会影响模型对知识的准确评估。其解决方案的关键在于构建一个可追溯的、基于语义嵌入检索与语言模型筛选的选项增补框架：通过从同一领域内其他题目中复制五组候选选项，结合确定性验证机制重建全部60,000个新增选项，并利用25个模型的校准结果分析评分偏差与生成预算的影响。研究发现，在12,000个土耳其语源题目上，十选题版本相比五选题平均分下降5.9个百分点，且在981个共享题目中，有102个原本正确作答的样本因采纳了借来的错误选项而失分；尤其在被启发式标记为负面题干的题目中，分数下降达24.4个百分点。尽管进行了人工核查审计，但由于评审工具使用未记录、判断依据不完整，部分结果存在不确定性。综上，该研究的核心贡献在于提供了一种可追踪的选项增补构造方法及其有效性边界分析，而非证明低选项数分数更能反映真实知识水平。
链接: https://arxiv.org/abs/2609.15467
作者: M. Ali Bayram
机构: Yıldız Technical University (伊斯坦布尔技术大学); Istanbul, Turkey
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Performance (cs.PF)
备注: Preprint of a manuscript submitted to Transactions on Machine Learning Research (TMLR)
Abstract:Adding answer options can lower multiple-choice scores without improving assessment validity. Turkish MMLU Pro examines this distinction using 12,000 Turkish-source questions across 58 sections. Each question retains its stem, five original options and source key, and receives five options copied from other questions in the same section. Sentence-embedding retrieval proposes candidates; a language model selects existing identifiers. Deterministic verification reconstructs all 60,000 additions. A 25-model calibration exposes scoring and generation-budget effects. Five evaluations produce source-key accuracies of 34.8%-81.4%. On 981 shared questions, one API-served model falls from 93.7% with five choices to 83.1% with ten; 102 of 115 lost correct responses select borrowed options. The decrease is 24.4 percentage points on heuristically flagged negative stems and 5.9 points elsewhere. A completed human-checked audit of 200 sampled questions, with undocumented reviewer tool use, yields 47 and 31 multiple-answer judgments across the two record sets, 25 of the latter unresolved. These records support concern about ambiguity, while their dependence and incomplete reviewer-method documentation limit validation. Because order and labels also change, the paired comparison measures augmentation as implemented. The contribution is a traceable construction and an analysis of its validity limits, not evidence that lower ten-choice scores measure knowledge better.
[NLP-31] MarKey: Marginal Utility Guided Greedy Keyframe Selection for Long Video Understanding
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）在长视频理解任务中因密集编码长帧序列导致的计算成本过高，以及在有限视觉预算下采用均匀采样可能遗漏稀疏但关键证据的问题。现有无训练式关键帧选择方法虽提升了推理效率并取得良好性能，但多数方法独立评分各帧，未显式考虑候选帧与当前已选子集之间的互补性，易造成冗余选择和证据覆盖不全。为此，本文提出一种无训练框架MarKey，将关键帧选择建模为具有子集感知能力的贪心优化问题：每轮迭代中，MarKey通过一个可计算的代理函数对候选帧进行评分，该函数联合考量查询相关性、边际覆盖增益及上下文依赖的冗余度，并选取综合效用最高的帧。为实现高效迭代评估，MarKey采用一组精简的代表性锚点近似全视频覆盖，并限定先前已选帧的滑动窗口以控制上下文依赖比较范围。在涵盖整体视频理解、以人为中心的视频理解及开放式视频理解的六个基准测试上，MarKey均显著优于现有方法；进一步分析表明，其性能优势在不同MLLM骨干网络、模型规模及帧预算条件下均具鲁棒性。
链接: https://arxiv.org/abs/2609.15408
作者: Hongchang Shi,Jinpeng Hu,Ao Wang,Wenzheng Zhou,Hui Ma,Feng Li,Zenglin Shi
机构: Hefei University of Technology (合肥工业大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Long-video understanding remains challenging for multimodal large language models (MLLMs) because densely encoding long frame sequences is computationally expensive, while uniform sampling under a limited visual budget can miss sparse yet decisive evidence. Recent training-free keyframe selection methods have enabled more efficient inference and yielded promising performance gains. However, many existing methods score frames largely in isolation without explicitly considering how each candidate complements the currently selected subset, potentially resulting in redundant selections and incomplete evidence coverage. To address this limitation, we propose MarKey, a training-free framework that formulates keyframe selection as subset-aware greedy optimization. At each iteration, MarKey scores each candidate using a tractable surrogate that jointly accounts for query relevance, marginal coverage gain, and context-dependent redundancy, and selects the frame with the highest utility. To make this iterative subset-aware evaluation efficient, MarKey uses a compact set of representative anchors to approximate full-video coverage and a bounded window of previously selected frames to limit context-dependent comparisons. Experiments on six benchmarks spanning holistic video understanding, human-centric video understanding, and open-ended video understanding demonstrate that MarKey consistently outperforms existing methods. Further analyses show robust gains across different MLLM backbones, model scales, and frame budgets.
[NLP-32] SlopShape: Identifying AI-Generated Commercial Web Content
【速读】： 该论文旨在解决现有生成式AI（Generative AI）检测方法在面对重述（rewriting）时脆弱性的问题，尤其针对词级检测器在识别未修改的AI生成文本时表现优异但无法刻画文本整体结构特征或溯源具体生成模型的局限。其核心解决方案在于从更深层次的结构性签名（structural signatures）入手，即分析信息呈现方式、顺序、证据使用及语调等语言组织特征，而非依赖词汇层面的异常。研究通过复现StoryScope框架，在2,250篇预ChatGPT时期的人类博客与11,250篇来自五个前沿大模型的对应AI仿制品上构建了一个包含214个特征的检测仪器，由大语言模型（LLM）执行并经人工金标准标注验证（人-人一致性kappa=0.928，人-模型一致性kappa=0.946）。结果表明，仅基于187个结构性特征即可在独立公司测试集上实现98.0%的宏平均F1值，且在每篇AI内容均由其自身模型重述后仍保持98.1%的性能，展现出极强的鲁棒性。此外，该信号具备可表征性和可溯源性：AI生成内容呈现出高度规整、自我宣告式的结构模式，79.3%可准确归因于原始模型，远超16.7%的随机猜测水平；而人类文本则占据罕见的结构配置。所有效应均与StoryScope一致，方向相同且幅度更大。研究开源了完整流水线、检测工具、提示模板、代码及聚合数据集。
链接: https://arxiv.org/abs/2609.15369
作者: Jochen Madler(Sitefire)
机构: Sitefire
类目: Computation and Language (cs.CL)
备注: 20 pages, 5 figures. Verification artifacts and code: this https URL
Abstract:Word-level detectors identify unedited AI-generated text almost perfectly, but the literature documents their brittleness under rewording, and a word-level score neither characterizes a text nor identifies which AI model wrote it. We ask whether AI-generated text can be identified one level deeper, from structural signatures: how information is presented, in what order, with what evidence, and in what voice. We replicate StoryScope (Russell et al., 2026), which showed such patterns for AI-generated fiction, on commercial content: 2,250 pre-ChatGPT human blog posts from 268 company domains against 11,250 AI mirrors from five frontier models. A 214-feature instrument, applied by an LLM and validated in a human gold-annotation session (human-human kappa 0.928, human-model 0.946), detects AI posts from its 187 structural features alone at 98.0 macro-F1 on held-out companies, unchanged (98.1) when every AI post is reworded by its own model. The signal characterizes and attributes: AI posts share a tidy, self-announcing shape, 79.3% are attributed to the correct source against a 16.7% chance rate, and human posts occupy rare structural configurations. All effects replicate StoryScope’s, consistent in direction and larger in magnitude. We release pipeline, instrument, prompts, code, and aggregate artifacts.
[NLP-33] RSIAgent : Autonomous Exploration for Recursive Self-improvement in New Environments
【速读】： 该论文旨在解决数字代理在面对未充分预训练覆盖的新环境时，难以有效适应其界面、工具及故障模式的问题。其核心挑战在于如何在不依赖模型微调的前提下，实现对复杂环境的自主理解与知识积累。解决方案的关键是提出一种无需训练的多智能体框架——RSIAgent，通过协调课程（curriculum）、执行者（actor）与验证者（verifier）三类智能体，实现递归式自我改进。该框架采用“广度优先—深度聚焦”探索策略，先并行开展广泛的递归自探索以发现环境结构多样性，再进行聚焦的深度自探索以揭示困难案例、隐藏约束、边界条件及未知因果关系。由此构建的环境特异性记忆库被冻结后可直接复用于下游任务，无需更新模型参数。实验表明，该方法显著提升了开源模型性能，使Kimi-K3与GLM-5.3在OSWorld-v2和Agent’s Last Exam基准上超越前沿闭源模型如GPT-6。
链接: https://arxiv.org/abs/2609.15364
作者: Sibo Zhu,Shicheng Fan,Xinyue Wang,Wenyi Wu,Kun Zhou,Biwei Huang
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 49 pages
Abstract:Digital agents must often adapt to new environments whose interfaces, tools, and failure modes are not fully captured by pretrained models. We introduce \textbfRSIAgent, a training-free multi-agent framework for recursive self-improvement through autonomous memory construction. RSIAgent coordinates curriculum, actor, and verifier agents to continually explore the environment, validate outcomes, and retain environment-specific knowledge, including reusable causal relationships between actions, conditions, and consequences. It further adopts a \textbfbroad-then-deep exploration strategy, combining parallel broad recursive self-exploration for discovering diverse environment structures with focused deep self-exploration for uncovering hard cases, hidden constraints, boundary conditions, and previously unknown causal dependencies. The resulting memory is frozen and can be directly reused for downstream tasks without updating model parameters. Experiments on OSWorld-v2 and Agent’s Last Exam show that RSIAgent substantially improves strong open-source models, enabling Kimi-K3 and GLM-5.3 to outperform frontier closed-source models including GPT-6.
[NLP-34] Parameter-Efficient Adaptation of Pretrained Language Models for Time-Series Forecasting
【速读】： 该论文旨在解决预训练语言模型在单变量时间序列预测任务中的跨模态迁移问题，核心挑战在于语言模型处理离散文本标记，而时间序列数据为连续数值观测并具有时序依赖性。其解决方案的关键在于：将固定长度的时间序列片段直接投影至预训练GPT-2模型的嵌入空间，跳过文本分词过程，使Transformer架构作为通用序列编码器使用。通过在能源、气象、交通和金融等七类基准数据集上的受控消融实验，研究发现基于连续嵌入的补丁表示策略显著优于文本提示或随机初始化的模型；同时，冻结预训练主干网络并仅微调轻量级投影与适配器模块，在保持低于总参数量1%更新比例的前提下，实现了与专用预测架构相当的MASE性能，展现出优越的精度-效率权衡与对输入上下文长度变化的稳定性。
链接: https://arxiv.org/abs/2609.15344
作者: Tamanna Kumavat,Georg Brunner,Kyriakos Flouris
机构: University of Zurich(苏黎世大学); ETH Zurich(苏黎世联邦理工学院); University of Cambridge(剑桥大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:We study the adaptation of pretrained language models to univariate time-series forecasting through a parameter-efficient transfer learning framework, with the goal of understanding which design choices drive effective cross-modal transfer. While language models operate on discrete textual tokens, time series consist of continuous numerical observations with temporal dependencies. To bridge this modality gap, we project fixed-length time-series patches directly into the embedding space of a pretrained GPT-2 backbone, bypassing textual tokenization and treating the Transformer as a generic sequence encoder. Through controlled ablation studies on seven benchmark datasets spanning energy, weather, traffic, and finance, we analyze the effects of (i)~representation strategy (continuous embeddings versus textual serialisation), (ii)~adaptation regime (frozen backbone versus partial or full fine-tuning), (iii)~architectural components such as adapters, pooling strategies, and prediction heads, and (iv)~input context length. Continuous patch-based embeddings consistently outperform textual prompting and randomly initialised backbones. The adapted pipeline attains MASE within the range of specialised forecasting architectures while updating less than 1% of total model parameters. Results further indicate that freezing the pretrained backbone and training lightweight projection and adapter modules provides a favourable accuracy–efficiency trade-off with stable behaviour across varying context lengths.
[NLP-35] Dynamic Semantic Compression for Efficient Latent-Space Inference in Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在推理过程中以词元（token）为单位进行计算所导致的内存开销过大及计算效率低下的问题。其核心解决方案是提出一种动态语义提取与推理（Dynamic Semantic Extraction and Inference, DSEI）框架，通过两阶段训练策略实现潜在空间中的段级推理。关键在于构建一种动态语义自编码器（Dynamic Semantic Autoencoder, DSAE），利用自监督学习动态提取文本段级语义，并通过自适应语义加权与门控融合机制将语义信息压缩为紧凑的潜在表示。随后将DSAE集成至LLM架构中，使模型能够在稠密潜在空间中进行推理，从而显著缩短输入与生成序列长度，大幅提升推理效率。实验结果表明，DSEI相比静态句级潜在表示推理基线可降低48%的困惑度，相较于传统的词元级推理模型，推理速度提升2.5倍，内存开销减少90%。
链接: https://arxiv.org/abs/2609.15338
作者: Peipei Li,Dongsen Zhang,Yuchen Liu,Wenjun Xu
机构: Beijing University of Posts and Telecommunications (北京邮电大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) primarily perform inference at the token level, resulting in substantial memory overhead and compromised computational efficiency. In this paper, we propose a Dynamic Semantic Extraction and Inference (DSEI) framework, which achieves segment-level inference within the latent space through a two-stage training strategy. First, we construct a Dynamic Semantic Autoencoder (DSAE) via self-supervised learning. DSAE dynamically extracts segment-level semantics and compresses them into compact latent representations via adaptive semantic weighting and gated fusion. Subsequently, we integrate the DSAE into the LLM architecture and train the model to infer over dense latent space. DSEI substantially reduces both input and generation sequences and significantly enhances inference efficiency. Extensive experiments conducted on the Wanjuan dataset demonstrate that DSEI reduces perplexity by 48% compared to static sentence-level latent inference baseline. Furthermore, compared to standard LLMs using token-level inference, DSEI accelerates inference speed by 2.5 \times and reduces memory overhead by 90%.
[NLP-36] When Agents Slow Down: Understanding LLM Agents Test-Time Strategies via Elo-per-token Analysis
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）智能体在测试阶段动态分配计算资源时，难以准确衡量其性能随计算量增长的规模化规律这一核心问题。传统评估方法受限于离散的最终结果反馈，无法捕捉智能体在长轨迹中的渐进性进展。为此，作者提出“每令牌埃洛评分分析”（Elo-per-token analysis），通过跟踪每个计算预算（token budget）下找到的最佳解，并利用Bradley-Terry模型将不同任务间具有异构评分尺度的中间表现统一映射为可比较的埃洛（Elo）等级分，从而实现跨任务、跨规模的性能量化。其关键在于引入独立采样作为理论基准——该基准下埃洛得分与对数计算量呈线性关系，以此作为衡量智能体边际效率的标尺。实验表明，尽管智能体初期能以高于基准的速度提升埃洛得分，但其边际收益迅速衰减并最终低于独立采样基准；相比之下，人类顶级选手在相同任务中表现出超线性增长，揭示出持续学习潜力和显著的性能差距。论文进一步定义“缩放拐点”（scaling inflection point）为边际埃洛增益等于独立采样基准的计算预算点，并据此将1亿令牌拆分为多轮并行会话，在FrontierCS多边形填充任务中分别实现+264和+355埃洛的增益，验证了并行策略在突破单次会话瓶颈方面的有效性。
链接: https://arxiv.org/abs/2609.15309
作者: Kaiyuan Liu,Qiuyang Mang,Bo Peng,Wenhao Chai,Hanchen Li,Shreyas Pimpalgaonkar,Luke Zettlemoyer,Alex Dimakis,Alvin Cheung
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language model (LLM) agents allocate test-time compute adaptively as they revise solutions, use tools, explore alternatives, and decide when to stop. This test-time strategy makes it difficult to measure how agent performance scales. We study open-ended tasks that provide continuous scores for intermediate submissions, making progress observable throughout long trajectories. We propose Elo-per-token analysis, which tracks the best solution found at each token budget and uses a Bradley-Terry model to aggregate within-task orderings into Elo ratings across tasks with different score scales. We apply it to four general-purpose agents on four open-ended benchmarks, with sessions of up to 100M tokens, and to three feedback-driven LLM optimization harnesses in controlled single-task interventions. Independent sampling provides a theoretically characterized reference, for which Elo grows linearly with log compute. Against this reference, agents can initially convert tokens into Elo faster than independent sampling, but their marginal gains diminish and eventually fall below the reference. In contrast, the strongest historical human contestants improve superlinearly over contest time on shared AtCoder Heuristic Contest tasks, providing evidence of continual learning and substantial headroom after agents slow down. We define the scaling inflection point as the per-session budget where marginal Elo gains match the independent-sampling reference. Using this point as the per-session budget, we split 100M tokens across parallel sessions on FrontierCS Polyomino Packing, gaining +264 Elo over one long session and +355 over ten short sessions.
[NLP-37] Reason What Matters: Retrieval-Grounded Reasoning for Universal Multimodal Embeddings
【速读】： 该论文旨在解决统一多模态嵌入（UME）在大规模语料库部署中面临的两大核心问题：一是基于强化学习的推理优化方法（如GRPO）对思维链（CoT）中所有标记赋予相同优势，无法有效识别支持正样本的关键证据或主张，导致信用分配不精准；二是生成完整思维链前缀以生成嵌入带来显著延迟，即使部分推理轨迹已具备足够的检索证据。为此，论文提出“关键推理”（Reason What Matters, ReWAM）框架，其解决方案的关键在于引入两种创新机制：首先，提出检索感知自蒸馏（Retrieval-aware Self-Distillation, RASD），通过构建源自输入支持证据的特权引导信号，区分正样本与难负样本，进而由在线策略自教师将轨迹级反馈细化为针对检索相关推理的令牌级监督；其次，设计检索自适应推理（Retrieval-adaptive Inference, RAI），利用检索置信度头预测部分思维链的剩余检索效用，实现对无效推理路径的早期终止，并通过推测解码加速有效推理延续。实验结果表明，ReWAM在MMEB-V2和MRMR数据集上达到当前最优检索性能，同时推理吞吐量最高提升5倍，显著弥合了检索质量与推理效率之间的鸿沟，使增强推理的统一多模态嵌入具备大规模应用可行性。
链接: https://arxiv.org/abs/2609.15296
作者: Mingzhou Jiang,Peixi Wu,Hang Cheng,Yunhao Zhou,Biao Yang,Wei Yuan,Yun Li,Fan Yang,Wenwu Ou,Honghui He
机构: Tsinghua Shenzhen International Graduate School, Tsinghua University(清华大学深圳国际研究生院); School of Artificial Intelligence and Data Science, USTC(中国科学技术大学人工智能与大数据学院); Kuaishou Technology(快手科技); College of Future Information Technology, Fudan University(复旦大学未来信息技术学院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Universal multimodal embedding (UME) learns unified representations across modalities, enabling a single model to support diverse retrieval tasks. Recent methods use Chain-of-Thought (CoT) reasoning to better interpret multimodal inputs before generating embeddings for complex retrieval tasks and further optimize this reasoning process through GRPO with retrieval-based rewards. However, two limitations hinder corpus-scale deployment. GRPO assigns all CoT tokens the same advantage, without identifying input-supported claims or evidence that distinguishes the positive from negatives. Moreover, generating a complete CoT before each embedding introduces substantial latency, even when a partial trace already provides sufficient retrieval evidence. To address these limitations, we propose Reason What Matters (ReWAM), a retrieval-grounded reasoning framework that uses retrieval feedback to guide both credit assignment and reasoning computation. Specifically, we introduce Retrieval-aware Self-Distillation (RASD), which constructs privileged guidance from input-supported evidence that distinguishes the positive item from retrieved hard negatives. An on-policy self-teacher uses this guidance to refine trajectory-level feedback into token-specific supervision for retrieval-relevant reasoning. We further develop Retrieval-adaptive Inference (RAI), which uses a retrieval confidence head to estimate the remaining retrieval utility of a partial CoT. It stops unproductive traces early and accelerates useful continuations with speculative decoding. Extensive experiments on MMEB-V2 and MRMR demonstrate that ReWAM achieves state-of-the-art retrieval performance while delivering up to 5x the inference throughput of competitive explicit-CoT UME methods. These results bridge the gap between retrieval quality and inference efficiency, making reasoning-enhanced UME practical for large-scale deployment.
[NLP-38] Artificial entrepreneurial cognition: Locating and causally steering an opportunity recognition dial inside large language models (LLM s)
【速读】： 该论文旨在解决创业认知（entrepreneurial cognition）研究中长期存在的核心问题：随着大语言模型（LLM）在创业活动中的深度介入，传统以人类为主体的创业认知框架面临扩展需求，亟需理解人工智能系统内部如何表征和处理创业相关概念。其解决方案的关键在于提出“人工创业认知”（artificial entrepreneurial cognition）这一新概念，即人工智能系统中与创业相关的表征与计算功能组织，并通过表征工程（representation engineering）实现对这些内部表征的机制可解释性分析。研究聚焦于机会识别（opportunity recognition, OR），构建了636组匹配的“有/无机会识别”场景对，成功在Llama 3.1 8B-Instruct模型中恢复出一个可被操控的OR方向。通过直接干预该方向（即“机会识别旋钮”），实现了对模型判断的因果性调控，且该方向具有可复现性、行为显著性及与其他阶段（如机会评估与开发）表征的几何分离性。该成果首次实现了对大语言模型内部创业构念的因果干预，为创业理论中关于机会识别与评估之间区分的争议提供了具象化的内在表征证据，同时确立了内部表征作为创业研究的新对象，并展示了理论引导下的表征识别、因果操纵与解释方法。
链接: https://arxiv.org/abs/2609.15277
作者: Christian Fisch,Angela Altmeier,Martin Obschonka,Michal Kosinski,Pin Ni
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 105 pages, 5 figures, 20 tables (8 in the main text, 12 in the appendices)
Abstract:Entrepreneurial cognition is a foundation of entrepreneurship research. Yet the growing involvement of large language models (LLMs) in entrepreneurial work extends the cognition question beyond human actors to systems whose internal representations remain largely unexplored. We introduce artificial entrepreneurial cognition, the functional organisation of entrepreneurship-relevant representations and computations inside artificial intelligence (AI) systems. We bring mechanistic interpretability into entrepreneurship research through representation engineering. Focusing on opportunity recognition (OR), we construct 636 matched OR-present and OR-absent scenario pairs and recover an OR direction in Llama 3.1 8B-Instruct. Rather than infer the construct from outputs, we intervene directly on this direction, steering the model up and down along what we call the opportunity recognition dial, and its opportunity judgments shift with it. To our knowledge, this is the first causal intervention on an internal representation of an entrepreneurship construct inside an LLM. Held-out tests, lexical and topical controls, behavioural ablation, and geometric comparisons show that the direction is recoverable, consequential, and distinct from the opportunity evaluation and exploitation directions, although steering it also shifts judgments about these neighbouring stages. Recovery, signed steering, and geometric separation hold across four additional LLMs spanning different scales and families. These results give the contested distinction between opportunity recognition and evaluation a concrete representational form inside AI systems. More broadly, they establish internal representations as a new object of entrepreneurship inquiry and show how entrepreneurship theory can guide their identification, causal manipulation, and interpretation.
[NLP-39] Semiotic Relations and Proof Methods: A Cross-Genre Study of Argument Structure with Large Language Models
【速读】： 该论文旨在解决在直接证明某一命题 $ S $ 困难或不可能时，如何通过与 $ S $ 相关的另一命题（或命题集）$ S^* $ 实现间接证明的问题。其核心挑战在于识别并构建从 $ S $ 到 $ S^* $ 的有效推理路径。解决方案的关键在于引入四种受符号学研究中四大修辞格启发的语义关系：句法关系（syntagmatic）、聚合关系（paradigmatic）、反讽关系（antithetic）和整体-部分关系（meronymic），分别对应转喻（metonymy）、隐喻（metaphor）、反语（irony）和提喻（synecdoche）。这四种关系被进一步操作化为四类证明策略：基于推论的证明（proof by inference）、类比证明（proof by analogy）、矛盾证明（proof by contradiction）和分情况证明（proof by case analysis）。通过将这些关系应用于数学、法律及日常论证的跨文体语料库，并借助大规模语言模型进行分析，研究发现不同文体中这些关系的使用存在显著差异：数学证明普遍运用全部四种关系，而法律与日常推理则几乎仅依赖于推论。
链接: https://arxiv.org/abs/2609.15194
作者: Edirlei Soares de Lima,Marco A. Casanova,Antonio L. Furtado
机构: Academy for AI, Games and Media; Breda University of Applied Sciences (布雷达应用科学大学); Breda, The Netherlands; Department of Informatics; PUC-Rio (里约热内卢天主教大学); Rio de Janeiro, Brazil
类目: Computation and Language (cs.CL)
备注:
Abstract:When a direct proof of a statement S seems hard or even impossible to obtain, there may exist another statement (or set of statements) S^* , somehow related to S , on the basis of which S can be proved. In order to investigate what options can be used to move from S to S^* , four kinds of semiotic relations inspired by the four master tropes of semiotic research are briefly reviewed. Specifically, our syntagmatic, paradigmatic, antithetic and meronymic relations correspond, respectively, to metonymy, metaphor, irony and synecdoche. It is suggested that these four semiotic relations determine the options to move from S to S^* , leading to proof by inference, proof by analogy, proof by contradiction, and proof by case analysis. To examine how the four relations are actually used across different kinds of argument, we complement the framework with an empirical study. We turn the four relations into explicit operational definitions and apply them to a cross-genre corpus of mathematical, legal, and everyday argument using a panel of large language models. We find that the relations are used very unevenly across genres: mathematical proofs draw on all four, whereas legal and everyday reasoning rely almost entirely on inference.
[NLP-40] What Limits Us? Analyzing Self-Reported Limitations in NLP Research EMNLP2026
【速读】： 该论文旨在解决顶刊顶会（如ACL、EMNLP）中自报局限性（Limitations）部分因数量庞大而难以进行人工系统性分析的问题。随着近年来自然语言处理（Natural Language Processing, NLP）领域论文数量的激增，大量自报局限性文本形成了一个庞大的语料库，但其内容长期处于未被充分挖掘的状态。为此，本文提出一种创新的人机协同迭代式混合定性编码框架（human-AI framework for iterative hybrid qualitative coding），作为解决方案的核心。该框架通过结合人类专家的深度理解与人工智能的高效处理能力，实现了对2020至2025年间大量论文中局限性章节的大规模系统分析，从而揭示研究者在报告自身工作局限时的趋势演变、与论文属性的相关性以及常见的写作模式。研究结果不仅反映了当前NLP领域研究人员在自我披露方面的多样性与实践特征，也为提升学术透明度和研究可复现性提供了重要参考。
链接: https://arxiv.org/abs/2609.15191
作者: Tawan Thaepprasit,Peeranuth Kehasukcharoen,Ding Wang,Remi Denton,Peerapon Vateekul,Piyawat Lertvittayakumjorn
机构: Chulalongkorn University (朱拉隆功大学); Google Research (谷歌研究院)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026 Findings
Abstract:Since late 2022, a Limitations section has become mandatory at many top-tier NLP conferences. The growing number of accepted papers at these venues has resulted in a vast corpus of self-reported limitations that cannot all be manually reviewed, yet remains systematically unanalyzed. Therefore, in this paper, we conduct a large-scale analysis of the Limitations sections from ACL and EMNLP papers published between 2020 and 2025 to understand what researchers disclose about their own work. To do so, we implement a novel human-AI framework for iterative hybrid qualitative coding. This framework enables us to investigate trends in self-reported limitations over time, their correlations with specific paper attributes, and the writing patterns that recur around these disclosures. Our findings offer a critical reflection on the diverse reported challenges as well as the self-reporting practices of researchers in the NLP community.
[NLP-41] MUSE: A Theory-Harnessed Story Engine for Vibe Narrativizing
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在生成高质量故事时面临的两大瓶颈：一是故事创作指导（story guidance）的质量不足，二是指导信息在创作全流程（包括规划、起草与修订）中难以持续有效应用的问题。其核心解决方案是提出“氛围叙事化”（Vibe Narrativizing）任务，并构建了基于理论驱动的故事生成引擎MUSE。MUSE的关键在于通过知识工程对罗伯特·麦基（Robert McKee）的故事理论进行规则原子化、语义整合与机制抽象，形成以单一可信源为基础、分层披露的结构化指导体系；同时，采用代理（agent）协管架构，通过中间产出物保留并传递关键创作决策，实现对情节设计、角色表现、场景构建及修订等环节的连贯控制。此外，上下文工程确保各角色获得精准引导，而杰作语料库则提供美学参考与文风范例。实验表明，MUSE在多个基准测试中显著优于零样本生成，尤其在故事一致性与整体质量上表现突出，其中结构设计贡献最大，角色语音特异性与修订环节进一步提升了生成效果。
链接: https://arxiv.org/abs/2609.15188
作者: Jianxiang Ma,Xiaocui Yang,Daling Wang,Yuesong Hou,Mingfu Zhang,Yichen Gao,Junzhao Huang
机构: Northeastern University (东北大学); OranAI (奥兰人工智能); OranAI Ltd. (奥兰人工智能有限公司)
类目: Computation and Language (cs.CL)
备注: 52 pages, including appendices; 3 figures
Abstract:LLMs can generate fluent prose. Story quality depends on how decisions about plot, character, and language work together across planning, drafting, and revision. Guiding these decisions presents two bottlenecks: the quality of story guidance and its sustained use. We formulate Vibe Narrativizing as the task of turning natural-language writing requirements into a finished story and present MUSE, a Theory-Harnessed Story Engine. MUSE organizes story knowledge as guidance for specific decisions and carries those decisions into subsequent creative work. Knowledge engineering develops Robert McKee’s story theory through rule atomization, semantic consolidation, and mechanism abstraction; a single source of truth and layered disclosure organize the resulting guidance. Typical examples complement principles that depend on context and aesthetic judgment. An agent harness organizes design, character performance, scene composition, and revision through intermediate deliverables that preserve story decisions. Context engineering supplies each role with the relevant guidance and decisions, while a masterwork corpus provides inspiration and prose references. A worked example follows one requested object from its thematic role to the characters’ climactic actions. Across four base models, MUSE improves WritingBench by 1.6-4.8 points over zero-shot generation and raises LongStoryEval by more than ten points on three. ConStory-Bench consistency error density remains in the low single digits for all four models, below every reproduced story-system baseline on three. Component ablations locate the largest quality contribution in structural design, voice-specific effects in the character path, and further gains in revision. Code is available at this https URL.
[NLP-42] CITECHOICE: A Causal Audit of How Document Presentation Redistributes Citation Credit in Agent ic Search
【速读】： 该论文旨在解决生成式问答系统在多源支持同一主张时的引用分配（citation allocation）问题，即当多个检索到的来源均支持同一事实时，系统为何选择引用某些来源而忽略其他来源。其核心挑战在于区分引用行为是否由真实信息质量驱动，还是受呈现方式、位置或模型随机性等非实质性因素影响。解决方案的关键在于提出CITECHOICE——一种基于真实多轮代理搜索对话的因果审计方法，通过哈希验证的2×2回放实验，在保持原始对话上下文不变的前提下，仅改变两个支持相同事实的文档对的呈现顺序，并联合生成结构化与自由文本两种格式的输出结果，以控制变量并评估不同呈现方式对引用决策的影响。研究发现，结构化呈现显著集中引用信用（每答案增加0.50次引用，p=0.033），但并未提升整体引用数量或减少竞争者引用；同时，观察性位置效应远大于可控重排效应，且引用判断存在可量化的噪声基线（约15%的二元决策随解码变化，45%的方差可归因于解码）。因此，该研究揭示：呈现方式可因果地重新分配可见引用信用，但无法可靠实现源采纳、纯粹格式化优势或普遍排名优势，从而为理解生成式AI中引用机制的内在偏差提供了关键实证依据。
链接: https://arxiv.org/abs/2609.15164
作者: Sriram Selvam,Anneswa Ghosh
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:When several retrieved sources support the same claim, an answer engine cites some but not others. We call this decision citation allocation and introduce CITECHOICE, a causal audit of authentic multi-turn agentic search. From 129 everyday-query transcripts, CITECHOICE selects 113 same-call document pairs with independently verified support for the same pre-specified fact, without observing ranks or answer outcomes; blinded human review confirms 103. It runs a hash-verified 2-by-2 replay crossing pair order with jointly generated, fidelity-checked structured and prose renderings of one target while the rest of the transcript remains fixed. Three results emerge. First, and most importantly, structured rendering concentrates citation credit rather than clearly increasing source admission. It raises target citation count by +0.50 citations per answer (95 percent CI [+0.20, +0.84]; Holm-adjusted p=.033), without increasing total citations or reducing competitor credit. The pre-specified incidence effect (whether the target is cited at all) is +4.5 percentage points and inconclusive (95 percent CI [-1.4, +10.4]; p=.168). Second, observational position differences exceed controlled reordering effects: the citation-rate gap between rank 1 and rank 5 is 42.3 percentage points, compared with +7.9 percentage points in the main replay and 0.0 percentage points held out. Third, citation evaluation has a measurable noise floor. Although the aggregate count effect repeats under fresh decoding of 30 frozen families, 15 percent of binary decisions change and decoding accounts for an estimated 45 percent of single-generation family-effect variance. Together, these findings isolate what survives control: presentation can causally redistribute visible citation credit within frozen transcripts. They do not establish reliable source admission, a pure formatting mechanism, or a general rank advantage. Subjects: Computation and Language (cs.CL) Cite as: arXiv:2609.15164 [cs.CL] (or arXiv:2609.15164v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.15164 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-43] EMR: Self-Evolving Medical Multi-Agent System via Experience Mining and Reuse
【速读】： 该论文旨在解决当前基于大语言模型（Large Language Model, LLM）的多智能体系统在复杂临床推理中面临的两大核心问题：一是现有方法依赖静态策略，无法根据历史经验进行动态优化；二是缺乏持续的临床记忆机制，难以积累并复用过往诊断中的成功模式与失败教训。其解决方案的关键在于提出一种自演化医学多智能体系统EMR（Experience Mining and Reuse），通过构建分层临床经验库（hierarchical clinical experience library），将累积的知识结构化为三个层级：临床原则、诊断模式和代表性病例。在推理过程中，系统模拟多学科会诊流程，由规划代理协调各专科代理进行专业化推理，并由总结代理整合分析结果生成最终决策。关键创新在于，系统能够自动从多智能体推理轨迹中提取正确的诊断洞察与失败预警信息，实现对经验库的增量式更新，从而支持未来任务的自我进化。实验表明，EMR在多个医学推理基准上持续优于现有先进方法，且分层经验机制具备跨专科泛化能力与对不同大语言模型底座的迁移适应性，展现出良好的可扩展性。
链接: https://arxiv.org/abs/2609.15161
作者: Dongsheng Shi,Yue Li,Xin Yi,Linlin Wang
机构: East China Normal University(华东师范大学); City University of Hong Kong(香港城市大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM) driven multi-agent systems have shown promise in complex clinical reasoning, yet existing approaches rely on static strategies and lack persistent clinical memory, preventing self-evolving from prior diagnostic successes and failures. We present EMR, a self-evolving medical multi-agent system via Experience Mining and Reuse. EMR introduces a hierarchical clinical experience library that organizes accumulated knowledge into three levels: clinical principles, diagnostic patterns, and representative cases. During inference, EMR emulates multidisciplinary consultation: a planner agent coordinates domain-specific department agents for specialized reasoning, while a summary agent synthesizes their analyses into a final decision. Critically, EMR automatically extracts correct diagnostic insights and failure-related warnings from multi-agent reasoning trajectories, incrementally updating the experience library to guide future cases. Experiments on medical reasoning benchmarks demonstrate that EMR consistently outperforms state-of-the-art medical multi-agent baselines. Further analysis reveals that the hierarchical experience enables cross-specialty generalization and transfer across diverse LLM backbones, offering a scalable and in
[NLP-44] PACE: Progressive Angular-to-Norm Contrastive Embedding
【速读】： 该论文旨在解决多模态嵌入模型中因采用基于余弦相似度的对比学习目标而导致语义兼容性受限于角度几何、无法充分利用嵌入范数作为额外语义信号的问题。现有方法虽能保证训练稳定性，但限制了语义表达能力。而直接优化更具表达力的点积相似度（dot-product similarity）虽理论上可同时利用方向与范数信息，却面临训练不稳定及性能下降的挑战。作者将此归因于优化空间过早扩展所引发的表征空间中方向-范数纠缠（angular–norm entanglement）与方向各向异性（directional anisotropy），并受全参数微调加剧影响。为此，论文提出一种两阶段框架 PACE（Progressive Adaptation with Confidence-aware Embedding），其关键在于：第一阶段通过结合余弦对比目标与低秩适配（Low-Rank Adaptation, LoRA），在受限优化空间内建立可靠的角几何结构；第二阶段切换至点积相似度并启用全参数微调，使嵌入方向与范数协同编码语义信息。此外，引入焦点嵌入损失（Focal Embedding Loss），通过自适应降低高置信度正样本查询的权重，强化对存在竞争负样本的模糊查询的建模。大量实验验证了 PACE 在多种骨干网络规模和多样化多模态任务中的有效性。
链接: https://arxiv.org/abs/2609.15152
作者: Yanping Li,Wei Zhou,Yawen Liu,Yibo Wang,Ke Zhu,Guangda Huzhang,Qing-Guo Chen,Zhao Xu,Jun Zhang,Wei Wei
机构: The Hong Kong University of Science and Technology (香港科技大学); Alibaba Group (阿里巴巴集团); Huazhong University of Science and Technology (华中科技大学); Nanjing University (南京大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Multimodal embedding models encode heterogeneous inputs into a shared embedding space, enabling efficient similarity computation across modalities and tasks. Most existing methods optimize cosine-based contrastive objectives, which promote stable training but restrict semantic compatibility to angular geometry, precluding embedding norms from serving as an additional semantic signal. However, directly optimizing the more expressive dot-product similarity, which leverages both angular and norm information, underperforms cosine-based training and exhibits unstable training dynamics. We attribute this discrepancy to premature optimization-space expansion, manifested as angular–norm entanglement and directional anisotropy in the representation space and further compounded by full-parameter fine-tuning. In this paper, we propose PACE, a two-stage framework that progressively expands both the representation and trainable parameter spaces. Stage I combines cosine-based objective with low-rank adaptation to establish a reliable angular geometry within constrained optimization spaces. Stage II switches to dot-product similarity and full-parameter fine-tuning, enabling embedding directions and norms to jointly encode semantic information. We further introduce Focal Embedding Loss, a confidence-adaptive objective that downweights queries with high positive retrieval confidence while emphasizing ambiguous queries with competitive negatives. Experiments across multiple backbone scales and diverse multimodal embedding tasks consistently validate the effectiveness of PACE.
[NLP-45] Improving Mathematical Reasoning Capabilities in Large Language Models via Reasoning Process Error Classification
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在数学推理任务中存在错误答案的推理能力瓶颈问题。具体而言，研究聚焦于分析导致LLMs产生错误答案的推理过程中的错误类型，并通过人工标注与分类构建了21类推理错误体系，识别出其中高频出现的错误类别。其解决方案的关键在于：基于对高频错误类型的系统性分析，设计了一种针对性强的提示（prompt），明确聚焦于其中8类常见错误，从而引导模型在推理过程中规避典型偏差。实验结果表明，该提示能够显著提升模型的推理性能，且所识别出的高频错误具有跨同规模模型的普遍性，验证了该方法的可迁移性和有效性。
链接: https://arxiv.org/abs/2609.15145
作者: Runa Yoshida,Kosuke Nishida,Kyosuke Nishida
机构: Human Informatics Labs., NTT, Inc.(NTT公司人类信息学实验室)
类目: Computation and Language (cs.CL)
备注: 10 pages
Abstract:The reasoning ability of large language models (LLMs) is a critical factor for practical LLM-based applications. To investigate the current reasoning capability of LLMs, we clarify the types of errors that arise in LLMs’ reasoning processes on mathematical datasets. We focus on problems where LLMs produce an incorrect answer. We define errors in the reasoning process as reasoning errors and manually analyze the features of reasoning errors. We defined and classified 21 error classes and identified the frequently occurring classes among them. Beyond qualitative evaluation, we leverage the evaluation results to improve the reasoning capability. We designed a prompt that explicitly focuses on eight error classes. The experiments demonstrate that this prompt effectively improves reasoning performance. Furthermore, the results suggest that the frequent reasoning errors identified in this paper are common across LLMs of comparable scale.
[NLP-46] MoME: Mixture-of-Memory Embeddings for Context-Aware Sparse Lookup
【速读】： 该论文旨在解决大规模语言模型中稀疏容量机制（如专家混合模型）在引入记忆嵌入（Memory Embedding）时存在的语义歧义问题：现有方法通过词汇表面形式确定性地检索记忆条目，导致同一词形（如“python”）在不同上下文中（语言或动物）的多种语义被压缩至单一固定记忆项，丧失了上下文敏感性。为此，论文提出混合记忆嵌入（Mixture of Memory Embeddings, MoME），其核心创新在于将每个词元对应的传统单一行记忆扩展为由M个记忆槽组成的混合结构，并引入基于隐藏状态学习的门控机制，动态决定在每一位置读取哪些记忆槽。该方案在纳米对话（nanochat）、Llama-3/MobileLLM及Qwen3等骨干模型上的控制预训练实验中，在参数量与训练浮点运算量（FLOP）一致的前提下，显著优于值嵌入（Value Embedding）、二元组（Bigram）和STEM基线；同时展现出在十亿以下规模下更优的记忆容量扩展趋势，并保持高效的训练与推理性能。对多义词的定性路由分析进一步表明，所学混合机制具备一定程度的语义可解释性，能够根据上下文语义差异将同一表面词元分配至不同的记忆槽，实现了真正的上下文感知记忆访问。
链接: https://arxiv.org/abs/2609.15126
作者: Muchen Li,Leonid Sigal,Renjie Liao
机构: University of British Columbia(不列颠哥伦比亚大学); Vector Institute for AI(人工智能向量研究所); Canada CIFAR AI Chair(加拿大CIFAR人工智能主席); NSERC CRC Chair(NSERC首席研究员主席)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Github: this https URL
Abstract:Scaling large language models efficiently has motivated sparse capacity mechanisms such as Mixture-of-Experts and, more recently, conditional memory: token-indexed embedding tables that augment the backbone with cheap parametric lookups. Existing memory-embedding methods retrieve via a deterministic function of the surface form, which collapses different contextual senses of the same token (e.g., python the language vs. the animal) into a single fixed entry. We introduce Mixture of Memory Embeddings (MoME), a context-aware memory mechanism that replaces each token’s single memory row with a mixture of M slots and uses a learned gate over the hidden state to choose which slots to read at each position. In controlled pretraining experiments across nanochat, Llama-3/MobileLLM, and Qwen3 backbones, MoME improves over Value Embedding, Bigram, and STEM baselines in iso-parameter and iso-training-FLOP settings, shows a more promising memory-size scaling trend at sub-billion scale, and remains efficient in training and inference. Qualitative routing analyses on polysemous tokens further suggest that the learned mixture exhibits a degree of semantic interpretability, dispatching the same surface token to distinct memory slots under different senses.
[NLP-47] When the Wrong Key Wins: Understanding and Detecting Hallucinations in LLM s
【速读】： 该论文旨在解决大语言模型在已有知识条件下仍会产生幻觉（hallucination）的问题，其核心挑战在于理解模型推理过程中因预训练阶段形成的关联竞争机制所导致的错误预测。研究提出从“潜在键”（latent-key）视角分析推理过程，认为答案选择依赖于预训练期间习得的语义关联之间的竞争。关键发现包括：模型预测对特定查询关键词高度敏感，这些关键词表现出实体特异性绑定（entity-specific binding），且其影响受预训练频率系统性调控；多个绑定可在同一查询中相互竞争并产生高阶交互。基于此机制，论文提出一种两阶段关键词扰动方法（two-stage keyword-perturbation method），通过移除关键影响词并观测模型预测的重组织行为，区分由误导性关联引发的错误与基于诊断性证据的正确决策。该方法在多个模型和基准测试上均展现出强鲁棒性和可迁移性，在已知知识的ScientistQA数据集上达到0.910 AUROC。进一步地，研究将该概率框架扩展至四种幻觉类型：知识缺失、错误知识、上下文干扰和不稳定的推理，其在不同基准上的分布特征为解释各类检测器在特定场景下的有效性提供了诊断性依据。
链接: https://arxiv.org/abs/2609.15106
作者: Xuhan Tong,Jiawei Zhang
机构: University of Wisconsin–Madison(威斯康星大学麦迪逊分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models can hallucinate even when the knowledge required for a correct answer is already available. We study this failure through a latent-key view of inference, where answer selection depends on competition among associations acquired during pretraining. We show that model predictions can be highly sensitive to individual query keywords, that these influential keywords exhibit entity-specific binding, and that their effects are systematically shaped by pretraining frequency. Multiple bindings can also compete and exhibit higher-order interactions within the same query. Based on this mechanism, we introduce a two-stage keyword-perturbation method for hallucination detection. By removing influential keywords and measuring how the model reorganizes its prediction, the method distinguishes errors caused by misleading key associations from correct decisions supported by diagnostic evidence. Across multiple models and benchmarks, perturbation provides a strong and transferable detection signal, reaching .910 AUROC on probe-known ScientistQA. Finally, we extend the same probabilistic framework to four hallucination regimes: knowledge deficit, wrong knowledge, context distraction, and unstable inference. Their operational distributions across benchmarks provide diagnostic context for why different detector families succeed in different settings.
[NLP-48] OpenAl4S: Code as Action Science as Sessions
【速读】： 该论文旨在解决长期运行的计算科学研究中，人工智能科研助手（AI co-scientist）在提升研究效率的同时，如何保障工作流的可检查性、可恢复性与可复现性这一核心挑战。其关键在于通过“代码即行动，科学即会话”（Code as Action, Science as Sessions）的设计原则，构建一个具备持久计算状态与完整溯源能力的开放源代码系统——OpenAI4S。该系统的解决方案核心在于：将科学操作建模为在持久化Python和R内核中执行的完整代码单元，结合结构化的工具调用进行任务编排，并利用追加只读的行动日志（Action Ledger）、每单元执行记录、版本化产物、环境记录及工作区快照等机制，实现对研究过程的全程可追溯与会话级恢复、分支与扩展支持。同时，通过可配置的沙箱隔离、权限控制及代码与轨迹筛查提供多重安全保障。实验评估覆盖36个跨领域的科研场景，结果显示OpenAI4S在任务准确率、工作流完整性与结果可复现性方面显著优于通用编码框架，尤其在长周期、计算密集型任务中优势明显。然而，环境定义与全链路可重运行性仍是当前所有系统（包括本系统）的薄弱环节，表明科学智能体的真正可复现性仍面临挑战。
链接: https://arxiv.org/abs/2609.15096
作者: Gongbo Zhang,Hao Li,Yu Wang,Mujie Lin,Liuzhenghao Lv,Yicheng Mao,Yimi Wang,Jun Zhu,Minhan Tang,Zhengxiang Jiang,Yusong Wang,Jiayu Yao,Kunpeng Ning,Dawei Pang,Yonghong Tian,OpenAI4S Community,Yuyang Liu,Li Yuan
机构: Peking University Shenzhen Graduate School(北京大学深圳研究生院); Beijing Yuankong Intelligent Technology Co., Ltd.(北京元空智能科技有限公司); Tsinghua-Peking Joint Center for Life Sciences, School of Life Sciences, Tsinghua University(清华大学-北京大学生命科学联合中心，清华大学生命科学学院); OpenAI4S Community(OpenAI4S 社区)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Software Engineering (cs.SE)
备注:
Abstract:AI co-scientists could accelerate computational research, but over a long-running study the workflow also has to stay inspectable, resumable and reproducible, which requires persistent computational state and provenance. Here we present OpenAI4S, an open-source scientific research agent built around the principle of \emphCode as Action, Science as Sessions. OpenAI4S combines a persistent computing runtime with research-session management: orchestration is handled through structured tool calls, while scientific actions are represented as complete code cells executed in persistent Python and R kernels. An append-only Action Ledger, per-cell execution records, versioned artifacts, environment records, and workspace checkpoints preserve how results were produced and support session recovery, branching, and extension. Configurable sandboxing, permission controls, and code and trajectory screening provide complementary safeguards. We evaluate OpenAI4S on 36 research scenarios spanning retrosynthesis, molecular dynamics, protein binder design, protein mutation, catalyst screening, and mineral spectroscopy, measuring scientific task accuracy, workflow completeness, and reproducibility of the resulting repositories. OpenAI4S achieves an overall score of 7.83, compared with 5.7–6.4 for a general-purpose coding harness evaluated with three frontier models, with the largest gains on long-horizon and computation-intensive workflows. These results suggest that integrating persistent execution with session-level provenance can improve the reliability of AI-assisted scientific workflows. Environment specification and full rerunnability remain weak for every evaluated system, ours included, so reproducibility is still an open problem for scientific agents. The system is available under the MIT license at \hrefthis https URLthis http URL.
[NLP-49] DA-DLM: Explicitly Modeling Token Dependencies in Diffusion Language Models
【速读】： 该论文旨在解决扩散语言模型（Diffusion Language Models, DLMs）在文本生成过程中因逐步去噪时各标记间条件独立性假设所导致的词元间依赖关系丢失问题，进而引发生成文本连贯性下降的缺陷。这一问题与非自回归翻译（Non-Autoregressive Translation, NAT）中的多模态困境具有相似性。为应对该挑战，论文提出一种基于有向无环图（Directed Acyclic Graph, DAG）的新型架构——位置导向型有向无环图扩散语言模型（DA-DLM），其核心创新在于设计了一种位置导向的DAG结构，将节点组绑定至固定的输出位置，使早期步骤中确定的词元作为锚点，通过学习到的转移机制引导邻近位置的预测；随着去噪过程推进，该图结构动态演化，逐步聚焦于剩余不确定性区域。该方法在语言建模、开放式生成及摘要任务上均显著优于块式扩散模型（Block Diffusion），尤其在较少去噪步骤下优势更为明显，并在保持并行生成效率的同时达到自回归模型的性能水平。
链接: https://arxiv.org/abs/2609.15070
作者: Pengyu Ji,Zichen Zhang,Xiang Hu,Kewei Tu
机构: ShanghaiTech University (上海科技大学); Shanghai Engineering Research Center of Intelligent Vision and Imaging (上海智能视觉与成像工程研究中心); Tencent (腾讯)
类目: Computation and Language (cs.CL)
备注:
Abstract:Diffusion Language Models (DLMs) generate text by iteratively denoising a masked sequence, independently predicting multiple tokens at each step. This conditional independence discards inter-token dependencies and degrades coherence-an issue that parallels the multi-modality problem in Non-Autoregressive Translation (NAT). Drawing on the Directed Acyclic Transformer (DAT), which tackles this problem in NAT via a Directed Acyclic Graph (DAG), we propose DA-DLM, a model that adapts DAG-based dependency modeling to DLMs’ iterative setting through a position-oriented DAG design. The position-oriented DAG binds node groups to fixed output positions so that tokens fixed in earlier steps anchor neighboring predictions via learned transitions, and evolves with denoising to focus on remaining uncertainty as anchors accumulate. On language modeling, open-ended generation, and summarization, DA-DLM consistently outperforms Block Diffusion, especially under fewer denoising steps, and matches autoregressive models while preserving the parallel generation advantage. Our code is publicly available at this https URL.
[NLP-50] Salesforce Koa: An Enterprise Language Model for Agent ic Tool Use
【速读】： 该论文旨在解决如何将开源权重的基础语言模型（如Nemotron-3-Super-120B）有效专业化为适用于企业级智能体任务的高性能模型，尤其在工具使用（tool use）和自主代理能力（agentic capabilities）方面实现显著提升，同时保持通用语言理解与生成性能。其核心挑战在于如何在不依赖客户数据的前提下，通过高效且可扩展的方式增强模型在复杂、多轮、数据依赖型任务中的表现。解决方案的关键在于提出并应用一种“仿真到奖励”（simulation-to-reward）的强化学习管道，结合组相对策略优化（Group Relative Policy Optimization, GRPO），通过将任务规范转化为基于角色设定的多轮交互任务，并以成功执行工具调用作为任务完成的明确奖励信号，驱动模型在真实企业场景（如客户关系管理，CRM）及公开工具使用领域中进行端到端的训练。该方法利用Salesforce自研的声明式语言Agent Script定义企业代理工作流，或直接合成公共领域的任务结构，统一支撑跨域的模拟与奖励机制，从而实现对开放权重模型的有效微调。实验表明，Salesforce Koa在多轮工具使用、自主推理及企业CRM任务上均优于其开源基线，超越强有的私有基准模型，且在性能上接近顶尖前沿模型，验证了基于任务规范的强化学习是实现开源大模型企业化专用化的可行路径。
链接: https://arxiv.org/abs/2609.15066
作者: Zixiang Chen,Sufeng Niu,Yingchi Liu,Wenting Zhao,Akshara Prabhakar,Shubham Mehrotra,Bin Bi,Zhujun Lan,Katherine Tan,Mohammad Ramezanali,Tulika Manoj Awalgaonkar,Monojit Banerjee,Jielin Qiu,Shiva Kumar Pentyala,Zhepeng Cen,Anupam Tripathi,Ali Ziaei,Regunathan Radhakrishnan,Darvish Lee Shadravan,Shelby Heinecke,Sitaram Asur,Silvio Savarese,James Zhu,Phil Mui,Huan Wang
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 16 pages, 4 figures, 5 tables
Abstract:We present Salesforce Koa, an enterprise language model built by post-training the open-weight Nemotron-3-Super-120B foundation model with reinforcement learning using Group Relative Policy Optimization (GRPO). Salesforce Koa is trained on public and synthetically generated data, with no customer data, to improve tool use and agentic capabilities while preserving strong general-purpose performance. Its distinctive component is a simulation-to-reward pipeline that expands workflow specifications into persona-conditioned multi-turn tasks with task-resolution rewards grounded in successful tool use for data-dependent requests. For enterprise domains, these specifications are written in Agent Script, Salesforce’s declarative language for building Agentforce agents; for public tool-use domains, we synthesize the workflow structure directly. The same simulation and grounded-reward machinery drives GRPO across both. Across public tool-use, agentic-reasoning, and enterprise Customer Relationship Management (CRM) benchmarks, Salesforce Koa improves over its open-weight base, with the clearest gains on multi-turn tool use, and surpasses a strong proprietary baseline while remaining below the strongest frontier models. These results show that specification-driven reinforcement learning is a practical path to specializing open-weight foundation models for enterprise agentic tasks.
[NLP-51] Not All Prompts Are Equal: Exploration-Guided Prompt Scaffolding for Multimodal Reinforcement Post-Training EMNLP2026
【速读】： 该论文旨在解决在线强化学习（Online Reinforcement Learning, RL）中训练提示（prompt）信息量不均衡的问题：部分提示已接近饱和，而另一些则过于困难，难以产生可靠的学习信号，但在标准训练流程中所有提示均被分配相同的采样预算，导致资源浪费与学习效率低下。其核心解决方案是提出一种探索引导的提示架构框架（Exploration-Guided Prompt Scaffolding Framework），通过动态调整多模态大语言模型（Multimodal Large Language Models, MLLMs）在强化学习后训练阶段的提示分布来优化学习过程。该方法的关键在于引入轻量级的探索潜力得分（Exploration Potential Score, EPS），这是一种基于KL正则化策略改进理论、仅依赖于在线采样统计即可计算的提示效用代理指标，无需额外计算开销。针对低效提示，该框架不直接丢弃，而是利用教师模型生成保持原始任务意图但更具信息性的重构提示，将教师监督从输出模仿转变为训练数据精炼。该方法在Geo3K和MMK12数据集上与GRPO结合使用，显著优于基线，在域内任务上实现最高9.7%的相对性能提升，并在MathVision和MMMU-Pro等域外基准上分别取得11.5%和11.1%的显著增益，验证了其有效性与泛化能力。
链接: https://arxiv.org/abs/2609.15051
作者: Yuanhao Yue,Qianli Ma,Chengyu Wang,Haoting Wang,Lei Shen,Jun Huang
机构: Alibaba Cloud Computing; Shanghai Jiao Tong University (上海交通大学); Fudan University (复旦大学); Xi’an Jiaotong University (西安交通大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted by EMNLP 2026 main conference
Abstract:Training prompts in online reinforcement learning (RL) differ substantially in how informative they are for the current policy: some are already saturated while others are too difficult to yield reliable learning signals, yet both receive equal rollout budget under standard training. We propose an exploration-guided prompt scaffolding framework that adapts the training prompt distribution dynamically throughout RL post-training of multimodal large language models (MLLMs). Central to our approach is the \textitExploration Potential Score (EPS) , a lightweight rollout-based proxy for prompt utility derived from KL-regularized policy improvement theory, computable directly from on-policy rollout statistics without additional overhead. Rather than discarding low-utility prompts, we use a teacher model to generate scaffolded rewrites that preserve the original task intent while making subsequent training more informative, reframing teacher supervision as training-data refinement rather than output imitation. Integrated with GRPO on Geo3K and MMK12, our method consistently outperforms the baseline on both in-domain and out-of-distribution benchmarks, achieving up to 9.7% relative improvement in-domain and gains of 11.5% on MathVision and 11.1% on MMMU-Pro.
[NLP-52] Mirror Mirror on the Wall: Prompt Echoing in Small Instruct Language Models EMNLP2026
【速读】： 该论文旨在解决生成式 AI（Generative AI）中普遍存在的“提示回响”（prompt echoing）问题，即模型在未明确指令的情况下，直接复制或重复输入提示内容而非生成新回答。这一现象是否源于模型泄露训练数据，抑或是内部归纳/复制机制的对齐偏差，尚存争议。研究通过分析来自不同模型家族（Gemma、Llama、Qwen、SmolLM 与 OLMo）的小规模语言模型发现，尽管部分回响内容与训练数据存在重叠，但其根本原因并非数据泄露，而是由模型内部的归纳头（induction heads）所驱动的非预期行为所致。因此，解决方案的关键在于识别并干预模型中负责诱导复制的内在机制，而非单纯依赖数据去噪或过滤。
链接: https://arxiv.org/abs/2609.15045
作者: Inez Okulska,Bartosz Naskręcki,Jan Piotrowski,Tomasz Steifer
机构: Centre for Credible AI, Warsaw University of Technology(华沙理工大学可信人工智能中心); Adam Mickiewicz University Poznan(亚当·密凯维奇大学波兹南分校); Institute of Fundamental Technological Research, Polish Academy of Sciences(波兰科学院基础技术研究所)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: accepted at EMNLP 2026
Abstract:Prompt echoing is a recognized failure mode of instruct language models, in which a model instead of generating a response, mirrors the provided prompt, even though it did not receive a specific instruction to do so. Is this phenomenon a sign of the model leaking the content of its training dataset, or is it rather caused by a misaligned behavior of the internal induction/copying mechanisms? We investigate prompt echoing small language models from different families (Gemma, Llama, Qwen, SmolLM and OLMo) and show that echoing prompts are likely to have partial overlap with the training dataset but the phenomenon is primarily driven by the model’s induction heads.
[NLP-53] he averag e-farmer illusion in language-model simulations of agricultural decisions
【速读】： 该论文旨在解决生成式语言模型（Generative AI）在社会模拟与调查研究中作为“合成个体”使用时，其行为真实性评估存在的根本性问题。尽管现有方法通常依赖于群体层面的均值或分布相似性来判断模型表现，但这种评估方式可能掩盖了个体行为预测能力的严重不足。研究通过在中非四国农民决策数据上对比Claude、Codex和Kimi三种模型，在四种预设提示设计下发现：尽管部分模型配置能较好复现总体均值与采纳率等宏观指标，但在个体层面的决策预测能力极弱，表现为行为高度集中于典型值，而关键政策相关的极端行为模式几乎完全缺失。最令人震惊的是，一个仅基于观测边际分布训练、未获取任何个体信息的简单生成器，反而在分布相似性上优于所有语言模型配置。这揭示了“平均农户幻觉”（average-farmer illusion）——即合成群体在宏观层面看似真实，却无法准确再现具体个体的行为选择及其异质性。为此，论文提出一种与主张匹配的验证框架及可复用的模块化提示组件，将提示设计转化为可审计的实验过程。研究强调：群体层面的相似性仅应作为验证起点，而非个体行为模拟有效的证据。
链接: https://arxiv.org/abs/2609.15038
作者: Zhanliang Zhu,Ziwei Li,Yuchen Liu,Liujun Zhu,Ruiqi Wu,Tongqing Shen,Junliang Jin,Jianyun Zhang
机构: Hohai University (河海大学); Yangtze Institute for Conservation and Development (长江保护与绿色发展研究院); Meta Platforms Inc.; Nanjing Hydraulic Research Institute (南京水利科学研究院); College of Water Conservancy and Hydropower Engineering (水利水电学院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Language-model agents are increasingly used as synthetic people in surveys and social simulations, yet their apparent realism is often judged from population averages or distributional similarity. We tested what such evidence actually establishes by comparing Claude, Codex and Kimi under four prespecified prompt designs with matched farmer decisions from China and four African countries. Some configurations reproduced observed means and adoption rates. However, their person-level predictions were weak; their decisions clustered around typical values and policy-relevant extremes were largely missing. Most strikingly, a simple generator fitted only to the observed marginal dis- tribution, and given no information about any farmer, achieved greater distributional similarity than every language-model configuration. Prompt additions produced conditional gains rather than uni- versal improvement: results varied with model, outcome, population and validation target. We call this the average-farmer illusion: a synthetic population can look realistic while failing to repro- duce who does what or how behaviour varies. We provide a claim-matched validation framework and reusable modular prompts that turn prompt construction into an auditable experimental process. Population-level resemblance should therefore be treated as the start of validation, not as evidence of individual simulation.
[NLP-54] MoARa: Module-Aware Rank Allocation and Structure-Preserving Decomposition for Low-Rank LLM Pre-training EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）预训练中低秩梯度投影（Low-rank gradient projection）方法在达到目标模型质量时所需迭代步数和实际运行时间过长的问题。现有方法存在两个关键设计缺陷：其一，将投影秩预算均匀分配给具有异质投影敏感性的Transformer模块，未能考虑各模块对梯度信息的敏感程度差异；其二，直接对原始梯度进行投影会同时削弱其幅值与方向信息，导致优化效率下降。为此，本文提出MoARa（Module-aware Adaptive Rank allocation），其核心创新在于结合静态分析驱动的模块感知秩分配策略与分块级的梯度幅值-方向解耦机制，其中默认分块大小设置在注意力头维度附近。实验表明，在涵盖Llama、Qwen和DeepSeek系列、规模从3亿到70亿参数的五种Transformer架构上，采用MoARa的GaLore方法在Llama 2 7B模型上可将达到标准最终困惑度所需的迭代步数减少37%，实际运行时间缩短34%，且仅引入0.2%的峰值预留内存开销（在标准图编译条件下）。在所评估的六种低秩预训练方法中，仅采用模块感知秩分配即在全部六种方法上实现步数方向性减少；在兼容硬件环境下，该双组件设计更可实现最高达41.7%的步数减少和37.1%的实际运行时间降低。
链接: https://arxiv.org/abs/2609.15037
作者: Keunyoung Kim,Nojun Kwak
机构: Seoul National University (首尔国立大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 Main Conference. 30 pages, 16 figures
Abstract:Low-rank gradient projection reduces the optimizer-state memory cost of large language model (LLM) pretraining, but the steps and wall-clock time needed to reach a target quality remain a meaningful axis for improvement. We attribute this to two design choices in existing methods: the projection-rank budget is allocated uniformly across Transformer modules with heterogeneous projection sensitivity, and projecting a raw gradient attenuates its magnitude and direction jointly. We propose MoARa, which combines a static profiling-based module-aware projection-rank allocation with a block-wise magnitude-direction decomposition; the default block size is set in the neighborhood of the attention head dimension. Across five Transformer architectures spanning Llama, Qwen, and DeepSeek at 300M to 7B scales, GaLore with MoARa reaches standard GaLore’s final perplexity in 37% fewer steps and 34% less wall-clock time on Llama 2 7B, with only 0.2% peak reserved memory overhead under standard graph compilation. Across the six low-rank pretraining methods we evaluate, module-aware rank allocation alone delivers directionally consistent step reductions on all six. On compatible hosts, the two-component design reaches up to 41.7% step reduction and 37.1% wall-clock reduction.
[NLP-55] Pick Your Poison: Learning to Select Poison Sets for Stronger LLM Backdoor Attacks
【速读】： 该论文旨在解决生成式模型在微调过程中遭受后门投毒攻击（backdoor poisoning attacks）时，现有评估方法因随机采样毒化样本而导致对最坏情况漏洞严重低估的问题。其核心挑战在于：在固定模型架构、干净数据集及毒化样本数量的前提下，不同的毒化样本组合会导致攻击成功率在3%至80%之间显著波动，表明样本选择策略对攻击效果具有决定性影响。为此，论文将毒化样本选择问题形式化为“基于预言机预算的集合优化”（oracle-budgeted set optimization），并提出SAILS（Set-level Audit-Informed Iterative Learned Selection）方法——通过仅需数百次微调与评估运行训练一个集合评分器，对数百万候选毒化集合进行高效排序，并仅对少量高潜力集合进行审计验证。SAILS在保持低计算开销的同时，相较于最强基线方法平均提升30个百分点的未见数据攻击成功率，具备从小规模到全规模微调的可迁移性，并可扩展至代码生成、代理型及API-only等多种后门攻击场景。
链接: https://arxiv.org/abs/2609.15029
作者: Aashiq Muhamed,Mona T. Diab,Virginia Smith,Andrew Ilyas,Matthew Jagielski
机构: Carnegie Mellon University (卡内基梅隆大学); Anthropic
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR)
备注:
Abstract:Backdoor poisoning attacks add poisoned examples to otherwise-clean finetuning data, pairing a trigger with a target behavior that the model learns to produce when the trigger appears. Existing evaluations typically fix the number of poisoned examples and sample them at random from a candidate pool. We show that this can severely underestimate worst-case vulnerability: across three LLaMA-3-8B backdoor settings, holding the model, clean data, and poison count fixed, attack success ranges from 3% to 80% depending only on which poison set is chosen. We formalize poison selection as oracle-budgeted set optimization and introduce SAILS (Set-level Audit-Informed Iterative Learned Selection), which learns a set scorer from a few hundred finetune-and-evaluate runs, ranks millions of candidate sets, and audits only a small shortlist. SAILS improves held-out attack success by 30 percentage points on average over the strongest influence baselines, transfers from small-scale to full-scale finetuning, and extends to code-generation, agentic, and API-only backdoors. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR) Cite as: arXiv:2609.15029 [cs.LG] (or arXiv:2609.15029v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.15029 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-56] SALUTE: Benchmarking and Adapting LLM s for the Defense Domain EMNLP2026
【速读】： 该论文旨在解决军事领域（defense domain）中大语言模型（LLM）缺乏系统性适配与评估框架的问题。现有研究多局限于特定任务、依赖有限的微调流程，且缺乏对军事专有知识（如条令概念、作战流程及动态军事事件）的全面评估体系。其解决方案的关键在于提出一个端到端的适配与评测框架SALUTE，包含四大核心组件：Salute-Corpus（基于公开美军条令与政府文件构建的精选语料库）、Salute-Conv（源自条令文献与十年防御新闻的指令数据集）、Salute-Pref（面向军事领域的偏好数据集）以及Salute-Bench（经过严格筛选的军事领域理解与推理评测基准）。基于这些资源，通过持续预训练、监督微调与偏好对齐的多阶段后训练策略，训练出Salute-LLM，实现了在军事领域性能显著提升的同时保持较强的通用能力，验证了SALUTE框架在军事领域大模型适配中的有效性。
链接: https://arxiv.org/abs/2609.15022
作者: Hyeongcheol Park,Sumin In,Suyeon Myeong,Hogun Park,Sangmin Kim,Moonhyun Lee,Daekyeong Park,Sangpil Kim
机构: Korea University(韩国大学); Sungkyunkwan University(成均馆大学); Hanwha Systems(韩华系统)
类目: Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026
Abstract:Defense is a knowledge-intensive domain that requires precise understanding of specialized terminology, doctrinal concepts, operational procedures, and evolving military events. Although recent work has explored language technologies for military applications, existing efforts remain fragmented: they are often task-specific, rely on limited adaptation pipelines, or lack comprehensive defense-domain evaluation. In this paper, we present SALUTE, an end-to-end framework for benchmarking and adapting LLMs for the defense domain. SALUTE integrates Salute-Corpus, a curated corpus from open-access U.S. military doctrine and government documents; Salute-Conv, a grounded instruction dataset from doctrinal sources and decade-long defense news; Salute-Pref, a defense-aware preference dataset; and Salute-Bench, a rigorously filtered benchmark for evaluating defense-domain understanding and reasoning over doctrine and defense news. Based on these resources, we train Salute-LLM through multi-stage post-training with continual pretraining, supervised fine-tuning, and preference alignment. Extensive experiments show that Salute-LLM achieves strong defense-domain performance while retaining competitive general capabilities, demonstrating the effectiveness of SALUTE as an end-to-end framework for defense-domain LLM adaptation.
[NLP-57] ABSOL: Aggregated Bayesian Subsampling Orchestrated with LLM s
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在处理结构化数据时，因缺乏一致性证据约束、依赖关系感知推理能力以及不确定性估计而带来的不可靠性问题。尽管贝叶斯网络（Bayesian Networks, BNs）提供了显式的概率推理框架，但其结构学习在大规模场景下仍面临计算成本高、对数据敏感且易失效的挑战。为此，本文提出ABSOL——一种由大语言模型引导的混合式贝叶斯网络结构学习框架，通过将LLMs作为有限语义引导（bounded semantic guides），在保持统计基础的前提下引入语言模型提供的语义先验知识。实验结果表明，在涵盖27至1041个节点的五个离散贝叶斯网络基准测试中，ABSOL是唯一在所有任务上均生成有效图结构的方法，并在超过27个节点的基准上以GPT-5.4实现最高的边准确率（Edge F₁）。四类基于LLM的增强策略为统计主干提供了互补的语义证据，使平均Edge F₁提升0.23；后验优化实验进一步揭示，这些性能增益部分源于对LLM在最终结构决策中权威性的限制。研究证明，当语言模型提供的语义知识被限定在统计驱动的推理流程中作为有界指导时，可显著提升可扩展的概率结构学习效果。
链接: https://arxiv.org/abs/2609.15007
作者: Jackson Hassell,Chen Shen,Estevam Hruschka
机构: Megagon Labs
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Large language models are increasingly used as natural-language interfaces to structured data, yet they remain unreliable when answers require consistent evidence conditioning, dependency-aware reasoning, and uncertainty estimation. Bayesian networks provide an explicit probabilistic reasoning layer, but learning useful structures from data remains costly and fragile at scale. We introduce ABSOL, a hybrid LLM-guided Bayesian network structure-learning framework that uses LLMs as bounded semantic guides. Across five discrete BN benchmarks spanning 27 to 1041 nodes, ABSOL is the only evaluated method to produce a viable graph on every benchmark, and achieves the highest Edge F_1 on every benchmark larger than 27 nodes with GPT-5.4. The four LLM augmentations, which contribute complementary semantic evidence to the statistical backbone, improve Edge F_1 over the non-LLM aggregation backbone by +0.23 on average. Complementary post-hoc refinement experiments suggest that these gains depend in part on limiting the LLM’s authority over the final structure. Together, these results show that language-derived semantic knowledge can substantially improve scalable probabilistic structure learning when used as bounded guidance within a statistically grounded reasoning pipeline. The code for ABSOL is available at this http URL.
[NLP-58] Beyond Depth and Width: The Information-Slack Dilemma in Streaming Test-Time Compute
【速读】： 该论文旨在解决在证据以不同顺序到达时，相同任务与计算预算下所需推理策略存在差异的挑战，核心问题在于如何权衡早期计算的先发优势与信息不完整或可修订所带来的风险，即“信息-计算松弛困境”（information-slack dilemma）。其解决方案的关键在于将依赖证据的计算任务作为分析单元，明确计算启动时机、结果支撑依据及结果可提交时间点，并强调提前计算的价值仅在于其成果在验证、作废和恢复成本之后仍能保持有效性。为此，论文提出了一项研究议程，聚焦于在受控证据更新条件下进行选择性恢复，主张评估应区分早期执行效应、部署价值相对于全输入方案的优势，以及预测性策略带来的额外增益，同时考虑共享资源成本。最终目标并非最大化提前计算，而是确保在既定资源范围内生成更可信且及时的响应。
链接: https://arxiv.org/abs/2609.14995
作者: Xiaotian Zhang(a href=“http://Trooly.AI” rel=“external noopener nofollow” class="link-external link-http"this http URL/a)
机构: Trooly.AI
类目: Computation and Language (cs.CL)
备注: Position paper. 10 pages, 1 figure, 3 tables
Abstract:The same task and compute budget can require different reasoning policies when evidence arrives in a different order. Early computation has more time to finish but rests on incomplete or revisable evidence; waiting improves information while shrinking computational slack. We call this the information-slack dilemma. We take the evidence-dependent computational job as the unit of analysis: when to start it, what supports its result, and when that result can be committed. Advance computation is valuable only insofar as its benefits survive the costs of verification, invalidation, and recovery. This applies to grounded incremental processing and reusable preparation as well as future-dependent speculation. We propose a research agenda on computation under evolving evidence, prioritizing selective recovery under controlled evidence revisions. Evaluation should separate earlier-execution effects, deployment value against a full-input alternative, and the added value of predictive policies, while accounting for shared-resource costs. The objective is not maximal advance computation, but more trustworthy, on-time responses within a declared resource envelope. Comments: Position paper. 10 pages, 1 figure, 3 tables Subjects: Computation and Language (cs.CL) Cite as: arXiv:2609.14995 [cs.CL] (or arXiv:2609.14995v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.14995 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-59] MTAC-IFBench: Benchmarking Instruction-Following in Multi-Turn Agent ic Coding
【速读】： 该论文旨在解决当前大型语言模型（LLM）在多轮交互式代码生成任务中指令遵循能力不足的问题。尽管现有代码代理已能在单轮对话或简单代码生成场景下实现功能正确性，但其在复杂、多轮迭代的软件开发流程中对多样化约束条件的持续遵循能力尚未得到充分评估。为此，论文提出MTAC-IFBench，一个面向多轮智能体编程中指令遵循能力的综合性基准测试框架。其关键创新在于构建了包含6类主约束与18类次级约束的多轮渐进式软件开发任务，每条实例平均具有7.04轮交互和91.33个约束，显著提升了评估难度。为确保评估可靠性，研究团队为每个约束和功能需求设计了核查清单，并集成验证脚本与判断代理（judge agents）进行自动化验证。实验表明，现有代码代理在多轮交互中指令遵循性能急剧下降，暴露出其在长期协作与复杂规则遵守方面的严重缺陷。
链接: https://arxiv.org/abs/2609.14992
作者: Bosi Wen,Cunxiang Wang,Jiayi Gui,Haoke Zhang,Yilin Niu,Pei Ke,Dayong Yang,Hongning Wang,Minlie Huang
机构: 未知
类目: Computation and Language (cs.CL)
备注: 23 pages, 7 figures
Abstract:Recently, the rapid development of large language models (LLMs) has reshaped software engineering by enabling autonomous code agents that plan, execute, and utilize external tools iteratively to tackle complex tasks. Beyond achieving functional correctness, these agents must faithfully follow process instructions and constraints throughout the development lifecycle. However, existing benchmarks typically focus on final functional correctness or confine instruction-following evaluation to single-turn, general chat or simple code generation scenarios, leaving instruction-following in multi-turn agentic coding underexplored. To bridge this gap, we propose MTAC-IFBench, a comprehensive benchmark for this critical capability. It features multi-turn progressive software development instructions with diverse constraints spanning 6 primary and 18 secondary categories. With an average of 7.04 turns and 91.33 constraints per instance, it poses a rigorous challenge to current LLMs. To make the evaluation reliable, we construct a checklist for each constraint and functional requirement, and integrate verification scripts and judge agents to verify each checklist item. MTAC-IFBench identifies significant deficiencies in existing code agents in multi-turn instruction-following, with their performance degrading rapidly as the interaction session grows longer.
[NLP-60] yphoon ASR Streaming: Steerable Low-Latency Thai Speech Recognition with Real-Time Shallow Fusion
【速读】： 该论文旨在解决开放型泰语自动语音识别（ASR）系统在实时应用场景中因依赖离线、全句输入的Whisper类模型而无法实现低延迟处理的问题，尤其限制了直播字幕生成和语音助手等对实时性要求较高的应用。其核心解决方案在于构建一个可部署的流式泰语ASR系统，支持在解码时动态调整词汇表（decode-time steering），而无需重新训练模型。关键创新包括：采用缓存感知的编码器（cache-aware encoder）恢复预训练全上下文模型的流式能力，或适配原生流式模型；引入浅融合层（shallow-fusion layer），利用GPU加速的n-gram语言模型与短语增强（phrase boosting）机制，在流式解码过程中对候选结果进行重排序。实验表明，该方法在两个泰语基准测试及两种模型规模下均显著优于传统全上下文模型，在1秒前瞻条件下将字符错误率（CER）降低4.3–4.5倍，且推理速度超过实时；同时，解码时词汇表调控使关键词召回率从16.6%提升至20.7%，无精度损失且开销极小，主要收益来自n-gram模型对混合语码（code-switched words）书写形式的统一建模，而短语增强则提供了对罕见领域术语的精准控制。
链接: https://arxiv.org/abs/2609.14991
作者: Warit Sirichotedumrong,Tanawin Samutsin,Shah Faisal Wani,Sittipong Sripaisarnmongkol,Kunat Pipatanakul
机构: Typhoon Team, SCB DataX(泰国商业银行数据科学团队), Bangkok(曼谷), Thailand(泰国)
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注: 4 pages, 2 figures, 3 tables. Accepted to IEEE SLT 2026 (Demo Track)
Abstract:Open Thai automatic speech recognition (ASR) is dominated by offline, Whisper-based models that read the whole utterance before transcribing, ruling out low-latency uses such as live captioning and voice agents. We present a deployable system for streaming Thai ASR that lets a user steer its vocabulary at decode time, without retraining. A widely used open Thai model, trained with full context, collapses when run as a true stream; we restore streaming with a cache-aware encoder, by converting it or adapting a natively streaming one, and add a shallow-fusion layer that re-ranks candidates inside the streaming decoder with a GPU n-gram language model and phrase boosting. Across two Thai benchmarks and two model sizes, the streaming models stay usable where the full-context model fails, cutting character error rate 4.3-4.5x at a one-second look-ahead while running faster than real time. Decode-time steering then lifts keyword recall from 16.6% to 20.7% at no accuracy cost and negligible overhead; most of the gain comes from an n-gram over ordinary training transcripts, which resolves the written form of code-switched words the model hears but spells inconsistently, with phrase boosting adding targeted control over rare domain terms.
[NLP-61] Biomedical Reference Generation Remains Unreliable across 26 Large Language Models
【速读】： 该论文旨在解决生成式 AI（Generative AI）在辅助撰写生物医学文本时频繁虚构参考文献的问题，即模型可能生成不存在的学术文献以“填补”缺失引用。研究通过向26个来自8家开发者的语言模型（发布时间为2023至2026年）提出69个跨十大领域的生物医学段落引用请求，评估其生成参考文献的真实性与准确性。关键解决方案在于系统性地对参考文献进行分类：可验证（真实且具可解析标识符）、部分匹配（真实但无可解析标识符）、虚构（无对应索引文献）或拒绝生成。进一步定义“完全正确”的标准为：必须为可验证文献，且期刊、年份及作者信息均与目标文献一致。研究结果显示，各模型虚构率介于10.2%至98.4%之间，平均55.4%的响应为虚构，仅14.9%在所有评估字段中完全正确；即使在2026年发布的五款新模型中，虚构率仍达35.3%，完全正确率仅为31.8%。结论表明，尽管部分模型能识别真实文献，但仍普遍存在元数据错误，因此基于生成式 AI 生成的参考文献必须经过人工核查方可使用。
链接: https://arxiv.org/abs/2609.14988
作者: Maxim Topaz,Zhihong Zhang,Nir Roguin,Pallavi Gupta,Zichao Li,Laura-Maria Peltonen
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Background. Large language models are increasingly used to help write biomedical text but may fabricate references to nonexistent work. How often large language models do so is not well characterized. Methods. We prompted 26 language models from eight developers (2023 to 2026) to supply a missing reference for each of 69 biomedical passages across ten domains. References were classified as verifiable (real paper with a resolving identifier), partial matches (real paper without a resolving identifier), fabricated (no matching indexed paper), or declined (the model refused to supply a reference). A reference was considered correct in every evaluated bibliographic field only when it was verifiable and its journal, year, and listed authors matched those of the cited paper. Results. Fabrication ranged from 10.2% (Claude Opus 4.8, which declined 52.1% of prompts) to 98.4% (Ministral 3B, which produced no verifiable reference). Claude Opus 4.6 and Claude Sonnet 4.5 produced similar proportions of verifiable references (77.6% and 76.6%) but named authors correctly in 78.7% and 28.7% of author-evaluable verifiable references, respectively, and were correct in every evaluated field in 54.6% and 19.9% of responses. GPT-5.5 was correct in every field in 48.1%. Across all models, 55.4% of responses were fabricated and 14.9% were correct in every field. Among the five tested models first released in 2026, the corresponding proportions were 35.3% and 31.8%, respectively. Conclusions. Fabrication remained common, and no model was correct in every evaluated bibliographic field in more than 54.6% of responses. Models that identify real papers may still misstate their metadata, so references produced with model assistance require verification before use.
[NLP-62] Online Language Adaptive Sampling for Better Distributed Cross-lingual Gains
【速读】： 该论文旨在解决多语言语言模型在极低资源语言（Extremely Low-Resource Languages, LRLs）上的跨语言迁移能力不足的问题，特别是现有重对齐（Realignment）方法因采用均匀且随机的平行句对采样策略，在小批量（limited batch size）情况下表现不佳的局限性。其核心解决方案是提出一种可学习的自适应采样策略，通过为每种语言分配可训练的采样概率，使对重对齐损失贡献更大的语言在后续训练批次中被更频繁地采样，从而实现动态优化的采样分布。该方法采用内-外层优化框架（inner-outer optimization loop），计算开销小，能够持续提升模型性能，并将增益均衡分布于各语言之间。实验表明，相较于传统的均匀重对齐，该方法在XLM-R上实现了平均+0.67的性能提升，在Gemma 2 9B上也达到+0.60的提升，且在不同模型架构下均表现出良好的鲁棒性。
链接: https://arxiv.org/abs/2609.14969
作者: Quang Phuoc Nguyen,Félix Gaschi,David Anugraha,Santiago Martínez Novoa,En-Shiun Annie Lee
机构: Ontario Tech University(安大略理工学院); Doctrine; Stanford University(斯坦福大学); University of the Andes(安第斯大学); University of Toronto(多伦多大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to ENMLP 2026 Findings
Abstract:Realignment is a promising approach for improving the cross-lingual transfer ability of multilingual language models, particularly for extremely low-resource languages (LRLs). However, existing realignment methods rely on uniform and random sampling of parallel sentences across languages, which may be suboptimal under limited batch sizes. In practice, models may benefit from seeing certain languages more frequently, especially those that are poorly aligned, and the optimal distribution can evolve throughout training. In this work, we propose a simple yet effective adaptive sampling strategy that assigns trainable sampling probabilities to each language. Languages that contribute more to the realignment loss are sampled more frequently in subsequent batches, and the optimal distribution can evolve throughout training. Our method employs an inner-outer optimization loop with a small overhead, leading to consistent performance improvements and, more importantly, distributing the gains across languages. We observed a +0.67 average performance increase on all tasks with XLM-R, and +0.60 with Gemma 2 9B compared with uniform realignment. Furthermore, our method is robust across different models. Code available at this https URL.
[NLP-63] A Corpus-Aligned Uthmani-to-Standard Quranic Word Mapping and a Deterministic Recitation Validator
【速读】： 该论文旨在解决《古兰经》文本在两种字形表示形式（Uthmani script 与 Standard Imla’i Arabic）之间存在的不一致问题，尤其聚焦于一个关键的Unicode字符U+0670（上标阿拉伯字母“أ”），该字符在高频诵读词中出现且常被通用阿拉伯语自然语言处理工具忽略或错误处理。其核心解决方案是构建了一个由2,290对齐词组成的、基于完整6,236节经文的Uthmani到标准阿拉伯语的映射语料库，并开发了一套七步文本归一化流水线。通过该流水线对所有经文进行归一化后，90.9%的经文可生成完全相同的字符串，而剩余差异被系统性地分析而非强行消除。在此基础上，作者进一步设计了一个无需大模型（LLM-free）、确定性的《古兰经》诵读验证器，采用四层逐级匹配策略（精确、形态学、宽松、模糊）并结合五级严重程度的词错误率反馈机制。该验证器在124个测试案例中取得98.4%（122/124）的准确率，两次失败均源于单一替换错误可能导致另一节经文成为精确匹配。此外，全语料普查揭示了16.5%的经文存在固有的仅依赖文本的歧义，而在34条来自实际阿拉伯语语音识别系统的诵读转录中，验证器全部正确识别了对应经文。研究团队已开源全部数据集、构建脚本、验证器及评估工具包，确保所有实验结果均可复现，除部署测试中的转录数据外，其余所有数值均可通过运行代码重现。
链接: https://arxiv.org/abs/2609.14967
作者: Yahya Mohamed Elnawasany
机构: 未知
类目: Computation and Language (cs.CL)
备注: 6 pages, 4 tables. Dataset, code and evaluation harness: this https URL
Abstract:Quranic text is distributed in two orthographic forms that are byte-level distinct: the Uthmani script used in every printed mushaf, and the Standard (Imla’i) Arabic form that every mainstream Arabic NLP tool is built for. The gap is concentrated in one Unicode character, U+0670 (superscript alef), which appears in some of the most frequently recited words in the Quran and is silently mishandled by general-purpose Arabic normalizers. We release a 2,290-pair, corpus-aligned Uthmani-to-Standard word mapping constructed by aligning the complete 6,236-verse Quran across both orthographic forms, together with a seven-step text normalization pipeline built on it. Normalizing both forms of all 6,236 verses through that pipeline yields identical strings for 90.9% of verses, and we characterize the residual divergence rather than assert that it is closed. On top of the normalized text, we build a deterministic, LLM-free Quranic recitation validator using a four-layer verse-matching search (exact, morphological, relaxed, fuzzy) and word-error-rate-graded feedback across five severity tiers. The validator scores 98.4% (122/124) on a 124-case suite emitted by the released test harness, and both failures share one mechanism: a single substitution error can make a different verse an exact match. A full-corpus census additionally quantifies an inherent text-only ambiguity affecting 16.5% of verses, and on 34 recitation transcripts drawn from a deployed Arabic ASR system the validator identifies the correct verse in every case. We release the mapping, the script that builds it, the validator, and the evaluation harness under open licenses; every number in this paper except the deployment measurement, whose transcripts are not ours to publish, is reproduced by running them.
[NLP-64] Can We Triage LLM Translation Errors in Classical Texts Without Human References? Source Novelty GEMBA Scoring and Budgeted Review through Pali-to-English Translation
【速读】： 该论文旨在解决在缺乏人类参考译文的情况下，如何有效识别大型语言模型（LLM）在翻译古典文本时需要人工审查的错误输出这一关键问题。其核心挑战在于构建一种无需参照标准译文即可评估翻译质量并进行错误优先级排序的方法。解决方案的关键在于提出并验证一种基于多信号融合的无参考误差分诊（reference-free error triage）框架，重点评估五种信号：源文本新颖性（source novelty）、源-候选译文嵌入距离、同行翻译分歧度、英-巴利语反向翻译一致性以及无参考GEMBA评分。研究通过3,000条经人工标注的样本进行信号校准，并在500条作者审定的锚点样本上验证效果，确保人类参考仅用于训练与验证，不参与风险信号计算。结果表明，源文本新颖性虽可作为源端风险先验，但无法精准定位具体错误；同行分歧与反向翻译提供辅助信号；而最强的无参考指标为由公认更强模型组成的团队所执行的无参考GEMBA评分——该方法在校准集中仅需审查前10%的高风险样本，即可捕获81.6%的专家多数意见认定的错误。此外，即使使用同层级模型团队（排除自评），仍具实用性但性能下降，说明评估者能力本身对判断准确性具有独立影响。最终提出一种预算受限的工作流，整合源端新颖性、同行分歧与强模型主导的候选译文感知评分，以优化人力评审资源分配。该方法在其他古典语言（如拉丁语、古希腊语、梵语）中的可迁移性尚待验证。
链接: https://arxiv.org/abs/2609.14963
作者: Máté Metzger
机构: Independent Researcher(独立研究员); Hungary(匈牙利)
类目: Computation and Language (cs.CL)
备注: Preprint. This article has not yet been peer reviewed
Abstract:As large language models become capable translators of classical texts, a key challenge is deciding which outputs need expert review when no human reference exists. This study tests reference-free error triage through Pali-to-English translation. Three LLMs translated 15,493 passages. Five signals were compared: source novelty, source-candidate embedding distance, peer-translation disagreement, English-to-Pali backtranslation, and no-reference GEMBA scoring. Signals were calibrated on a 3,000-item reference-informed LLM-adjudicated sample and checked against a 500-item author-adjudicated anchor. Human references supported calibration and validation only; they were never used to compute the risk signals. Source novelty was a useful source-side risk prior but not a per-candidate error detector. Peer disagreement and backtranslation provided secondary signal. The strongest method was no-reference GEMBA scoring by a panel of models generally regarded as stronger than the translators: reviewing the top 10% by GEMBA risk captured 81.6% of panel-major errors in the calibration set. GEMBA also remained the best reference-free signal against the author anchor. A same-tier panel, with self-scoring excluded, remained useful but performed worse, indicating that evaluator strength matters beyond the prompt alone. A budgeted workflow is proposed, combining source novelty, peer disagreement, and stronger candidate-aware judging to allocate human review. Transfer to other classical languages, including Latin, Ancient Greek, and Sanskrit, remains to be tested.
[NLP-65] Geometric Signatures of Conceptual Reorganization: A Counterfactual Embedding Framework for Detecting Scientific Revolutions
【速读】： 该论文旨在解决科学知识体系在历史演进过程中发生概念重组（conceptual reorganization）的量化难题，特别是如何识别和度量科学革命的发生。其核心问题在于：如何从海量科学文献中提取并量化某一新概念的出现对整个学科知识结构的影响。解决方案的关键在于提出一种基于文档嵌入几何（document embedding geometry）的反事实消融框架（counterfactual ablation framework），通过移除与候选概念相关的文献前后，在嵌入空间中观测几何结构的扰动，从而揭示该概念对科学知识组织方式的影响力。该方法以几何变化为可量化的可观测指标，结合五项跨领域的历史案例研究（包括狭义相对论、哥德尔不完备性定理、希格斯机制、深度学习及Transformer中的注意力机制），验证了其在识别概念重组信号上的有效性，同时揭示了文献标注误差与历史数据稀疏性带来的局限性。结果表明，文档嵌入几何可作为量化科学范式转变的新工具，为理解科学知识的动态重构提供了可计算的分析路径。
链接: https://arxiv.org/abs/2609.14917
作者: Dimitris Ntounis,Ariel Schwartzman,Chris Chafe,Thomas A. Ryckman
机构: Stanford University (斯坦福大学); SLAC National Accelerator Laboratory (SLAC 国家加速器实验室)
类目: Digital Libraries (cs.DL); Computation and Language (cs.CL); Machine Learning (cs.LG); History and Philosophy of Physics (physics.hist-ph); Physics and Society (physics.soc-ph)
备注: 22 pages, 13 figures, 5 tables
Abstract:We introduce document embedding geometry as a quantitative observable of conceptual reorganization and develop a counterfactual ablation framework for measuring how individual concepts influence the organization of scientific knowledge, providing a quantitative framework for detecting scientific revolutions. The observable is defined by the geometric perturbation induced when removing documents associated with a candidate concept from the embedding space before and after its historical emergence. Statistical validation is performed using five historical case studies spanning physics, mathematics, and machine learning: special relativity, Gödel’s incompleteness theorems, the Higgs mechanism, deep learning, and the attention mechanism underlying transformer architectures. Across the historical case studies, the framework identifies measurable geometric signatures associated with conceptual reorganization, while the validation studies expose important limitations arising from document assignment and sparse historical data. These results establish embedding geometry as a medium for quantifying conceptual reorganization, providing a new approach for studying how scientific fields restructure over time.
[NLP-66] Agent KV: Phase-Aware KV Eviction for Agent ic LLM s
【速读】： 该论文旨在解决生成式智能体（Agentic）服务在推理过程中相较于传统聊天机器人（chatbot）产生数量级增长的令牌消耗问题，这一现象对键值缓存（KV-cache）容量和解码阶段带宽造成巨大压力。现有大多数KV缓存淘汰策略依赖于最近若干令牌中提取的代表性查询来评估缓存键的价值，隐含假设是未来注意力模式与近期模式相似。然而，论文通过主角度分析（principal-angle analysis）揭示，生成式智能体的生成过程包含“思考（think）、执行（act）、调用工具（tool）”等多个阶段，这些阶段对应的查询分布在显著不同的子空间中，因此基于近期查询的代表性方法会系统性低估后续阶段所需的关键缓存项。为此，论文提出AGENTKV机制，其核心在于为每个行为阶段维护一个小型查询缓冲区，并将缓存键的评分依据扩展至各阶段查询的联合分布，从而更准确地保留跨阶段所需的上下文信息。此外，作者还实现了一个持久化的多轮服务路径，支持压缩后的KV状态跨轮次传递，并在线动态压缩保留的KV页。实验表明，在两个模型、六个任务领域及三种不同KV缓存预算条件下，AGENTKV相比R-KV和Tri-attention分别平均提升任务得分5.5和5.3分；相较于上游全量KV缓存方案SGLang，AGENTKV最高可提升输出令牌吞吐量1.80倍。
链接: https://arxiv.org/abs/2609.14872
作者: Taowen Tony Liu,Jeffrey T. H. Wong,Can Xiao,Bowen Yang,Hao Mark Chen,Yiren Zhao
机构: Imperial College London(帝国理工学院); Columbia University(哥伦比亚大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Agentic serving can consume orders of magnitude more tokens than chatbot workloads, stressing both KV-cache capacity and decode-time bandwidth. Most KV-eviction methods score cached keys against representative queries drawn from the most recent tokens, assuming future attention resembles recent attention. We show that agentic generation violates this assumption: future queries form a mixture over think, act, tool, and others phases, and principal-angle analysis shows these components occupy measurably different query subspaces, so recency representatives systematically undervalue keys that upcoming phases will need. We propose AGENTKV, which maintains a small query buffer per phase and scores cached keys against their union. We further implement AGENTKV in a persistent multi-turn serving path that carries compressed KV state across turns and compacts retained KV pages online. Across two models, six task domains, and three KV budgets each, AGENTKV improves task score by 5.5 points on average over R-KV and 5.3 over Tri-attention. Relative to upstream full-KV SGLang, AGENTKV improves output-token throughput by up to 1.80x. Code: this https URL.
[NLP-67] One Example Is Enough to Pass Fairness Benchmarks: Rethinking Fairness Evaluation for Aligned LLM s EMNLP2026
【速读】： 该论文旨在解决当前公平性评估基准（如BBQ）存在的根本性问题：这些基准过于简单，容易被模型通过少量示例或特定提示策略“破解”，从而导致评估结果无法真实反映模型的公平性对齐水平。其核心问题是，现有基准主要依赖单一结构线索（如多项选择题中的回避模式），使得模型只需学习表面模式即可获得高分，而并不具备真正的公平性推理能力。解决方案的关键在于揭示并利用一种普遍存在的“缺失证据”（missing evidence）推理模式——仅需一个示例作为上下文提示（in-context learning, ICL），模型即可在多个模型家族中显著提升对BBQ基准的准确率（从79.9%提升至99.0%），甚至超越大规模基于人类反馈强化学习（RLHF）训练的模型。这一现象表明，当前的评估体系未能有效区分模型是否真正理解并遵循公平性原则，而是被其对提示结构的敏感性所误导。因此，论文呼吁建立覆盖更广泛公平性维度的新型评估套件，以推动生成式AI系统在真实场景下的公平性对齐。
链接: https://arxiv.org/abs/2609.14860
作者: Naihao Deng,Samee Arif,Shuaichen Chang,Yulong Chen,Rada Mihalcea
机构: University of Michigan (密歇根大学); The Ohio State University (俄亥俄州立大学); University of Aberdeen (阿伯丁大学); University of Cambridge (剑桥大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to EMNLP 2026 Main
Abstract:Warning: This submission studies stereotypes and biases, and contains toxic and offensive examples, used for illustration purposes only. Fairness benchmarks such as BBQ have become the de facto standard for fairness evaluation across major model families. We argue that these benchmarks are too easy to support their role: training Qwen 2.5 7B Base with Group Relative Policy Optimization (GRPO) on a single BBQ example, or placing that example in context as a one-shot demonstration for in-context learning (ICL), lifts mean BBQ accuracy from 79.9% to 92.9% and 99.0%, respectively, closing 80% of the gap to its large-scale RLHF counterpart (96.1%) with GRPO, and surpassing it with ICL. These effects generalize across model families. A cross-conditioning analysis shows the improvement is carried by the reasoning traces generated by the model, and one example suffices to elicit a category-agnostic ``missing evidence’’ reasoning pattern. We argue that BBQ-style multiple-choice abstention benchmarks measure a single structural cue, and a model that solves them does not thereby become fair. We call for evaluation suites that cover a broader spectrum of fairness alignment. Comments: Accepted to EMNLP 2026 Main Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14860 [cs.CL] (or arXiv:2609.14860v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.14860 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-68] Dream-RSI: Recursive Self-Improvement through Evolving Worlds
【速读】： 该论文旨在解决自主智能体在复杂领域中实现递归式自我改进（recursive self-improvement）时，探索策略（exploration strategy）难以有效优化的核心挑战。当前系统面临根本性困境：固定探索策略无法适应不断扩大的搜索空间，而在线策略优化则需在长周期回放（long-horizon rollouts）中面对延迟且昂贵的反馈，导致元搜索空间（meta-search space）的探索效率极低。其解决方案的关键在于提出 \textscDream-RSI 框架，通过引入轻量级编排层使探索过程显式化与可编程化，同时保持底层编码智能体不变。核心创新点是利用累积的发现历史构建“回放模拟器”（replay simulator），基于已实现的搜索空间历史树进行“梦境式”（dreaming）离线推演，从而在无需重复昂贵在线评估的前提下，获得即时、低成本的离策略（off-policy）反馈，用于评估和优化探索策略。优化后的策略被重新部署于线上以驱动进一步发现，形成持续扩展回放模拟器池的自增强循环。在算法工程、数学优化及GPU内核工程等多个场景中，\textscDream-RSI 在保证或提升发现质量的同时，显著降低了发现成本。
链接: https://arxiv.org/abs/2609.14858
作者: Tong Zheng,Xidong Wu,Zheng Zhang,Zhankui He,Chaoyi Zhang,Benjamin Coleman,Ruoqiao Wei,Di Bai,Haolin Liu,Rui Liu,Xue Wang,Yue Zhuan,Wang-Cheng Kang,Renkai Xiang,Heng Huang,Xinwu Cheng,Yunsong Guo
机构: University of Maryland, College Park(马里兰大学学院帕克分校); Google DeepMind(谷歌深度思维); University of Virginia(弗吉尼亚大学)
类目: Computation and Language (cs.CL)
备注: 12 pages
Abstract:Recursive self-improvement is becoming increasingly vital for autonomous AI agents, where progress hinges on discovering high-value solutions across complex domains. The driver of this process is effective exploration, however, managing and improving exploration strategies remains a major bottleneck. Current systems face a fundamental dilemma: fixed strategies fail to adapt as search spaces scale, while online policy optimization requires navigating vast meta-search spaces under delayed and expensive feedback over long-horizon rollouts. We introduce \textscDream-RSI, a framework for scalable and recursively self-improving exploration. A lightweight orchestration layer makes exploration explicit and programmable while leaving the underlying coding agent unchanged. Our key insight is that accumulated discovery history can serve as a replay simulator over the realized search space. By performing dreaming in the replay simulator constructed from historical discovery trees, \textscDream-RSI secures immediate, low-cost off-policy feedback to evaluate and refine exploration policies without invoking repetitive, expensive online evaluations. The improved policy is subsequently redeployed online to drive further discovery, continuously expanding the simulator pool in a self-improving loop. Across algorithm engineering, mathematical optimization, and GPU kernel engineering, \textscDream-RSI achieves competitive or improved discovery quality while substantially reducing discovery cost in several settings.
[NLP-69] ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement
【速读】： 该论文旨在解决递归自改进（Recursive Self-Improvement, RSI）在复杂长时序任务中难以实现可泛化架构演化的挑战。现有方法在评估基准或其子集上迭代优化代理架构，导致难以区分通用性改进与特定基准的适应性调整；单轨迹更新易将系统性架构缺陷与实例特异性推理混淆，降低跨任务迁移能力；同时，对整体式（monolithic）架构中的重复行为缺陷定位困难，全架构优化又易引发无关机制纠缠，影响改进的可解释性与验证性。针对上述问题，论文提出一种模块化递归自改进（ModularRSI）框架，其核心在于：通过去基准化（benchmark-disjoint）、对比学习（contrastive） 与模块化分解三重机制实现可泛化架构演化。具体而言，ModularRSI通过对比同一任务的成功与失败轨迹，聚合跨任务证据以识别重复出现的行为缺陷，并将可进化架构解耦为五个功能模块——代理循环（Agent Loop）、工具使用（Tool Use）、观测管理（Observation Management）、上下文管理（Context Management）与任务完成检测（Task Completion Detection），各模块在受限修改范围内独立演化，随后进入集成阶段融合模块并解决潜在冲突。为支持去基准化演化，研究构建了2,000个来自外部源的可执行演化任务，与下游评估基准无交集。实验在TB2.0和SWE-Bench Verified上均验证了所提方法在未见域内与跨域任务上的持续性能提升，且演化后的架构具备在不同基础模型间的良好迁移能力，显著增强了生成式AI（Generative AI）代理的长期自主优化能力。
链接: https://arxiv.org/abs/2609.14857
作者: Siwei Wu,Jincheng Ren,Yizhi Li,Haau-Sing Li,Chengran Yang,Yuxuan Zhang,Weicheng Gu,Jian Yang,Riza Batista-Navarro,Chuanyi Zhang,Xianglong Liu,Ming Zhou,Bryan Dai,Chenghua Lin
机构: Beihang University (北京航空航天大学); University of Manchester (曼彻斯特大学); IQuest Research (IQuest研究实验室); M-A-P (M-A-P); Langboat (朗博科技); Hohai University (河海大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Recent work extends recursive self-improvement (RSI) to agent harnesses for long-horizon coding and terminal tasks, enabling agents to improve execution mechanisms from experience. However, generalizable harness RSI remains challenging. First, evolving harnesses on evaluation benchmarks or their subsets makes it difficult to distinguish reusable improvements from benchmark-specific adaptation. Second, single-trajectory updates can conflate systematic harness deficiencies with instance-specific reasoning and solution details, producing modifications that transfer poorly to unseen tasks. Third, localizing recurring behavioral deficiencies within monolithic harnesses is difficult, while whole-harness optimization can entangle unrelated mechanisms and complicate attribution and validation. We propose ModularRSI, a benchmark-disjoint, contrastive, and modular framework for generalizable harness evolution. ModularRSI contrasts successful and failed trajectories for the same task and aggregates evidence across tasks to identify recurring behavioral deficiencies. It decomposes the evolvable harness into five functional modules: Agent Loop, Tool Use, Observation Management, Context Management, and Task Completion Detection. Each module evolves independently within a restricted modification scope, followed by an integration stage that combines the evolved modules into a unified harness and resolves potential conflicts. To support benchmark-disjoint evolution, we curate 2,000 executable evolution tasks from external sources that are disjoint from downstream evaluation benchmarks. Experiments on TB2.0 and SWE-Bench Verified show consistent improvements on unseen in-domain and cross-domain tasks, with the evolved harness also transferring across different foundation models.
[NLP-70] Self-Orchestrating Language Models: Leverag ing Semantic Dependence for Efficient Inference
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在实际部署中面临的效率瓶颈问题，具体包括自回归解码带来的高推理延迟、低批量场景下硬件加速器利用率低下、长上下文推理引发的内存瓶颈，以及离散扩散模型在保持生成质量的同时需大量去噪步骤的问题。其核心解决方案是提出“自协调语言模型”（self-orchestrating language models）的概念，即让语言模型通过在生成过程中显式标注语义依赖关系——即哪些输出标记依赖于其他标记——来自主决策推理执行策略。该方法的关键在于引入可学习的语义依赖标注机制，并设计相应的运行时系统，基于这些标注实现自回归解码的并行化、中间推理状态的缓存淘汰（KV cache evictions），以及离散扩散过程中的去噪顺序规划。通过三个具体系统验证：PASTA利用依赖标注实现自回归生成的并行化；TIP基于依赖关系优化KV缓存使用以降低内存开销；Planned Diffusion则通过自回归生成去噪计划，指导离散扩散模型按最优顺序并行去噪，从而在保证生成质量的前提下显著提升效率，最终实现质量与效率之间的帕累托最优权衡。
链接: https://arxiv.org/abs/2609.14850
作者: Tian Jin
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) demonstrate impressive capabilities, but their deployment presents significant efficiency challenges. Autoregressive decoding imposes substantial inference latency and under-utilizes hardware accelerators in low batch size regimes. Discrete diffusion models can generate in parallel but struggle to match autoregressive quality without many diffusion denoising steps. Long-context reasoning creates memory bottlenecks that strain even state-of-the-art accelerators. My thesis is that language models can direct their own inference execution strategy by annotating semantic dependence – which tokens depend on which others – in their generation. I call such models self-orchestrating language models. For each system, I design a runtime that acts on these annotations to parallelize autoregressive decoding, evict intermediate context, or derive denoising orders, achieving Pareto-optimal quality-efficiency trade-offs. I demonstrate this approach through three self-orchestrating systems. First, PASTA uses semantic dependence to parallelize autoregressive decoding, training the model to annotate which output chunks can generate independently. Second, TIP uses semantic dependence to evict intermediate reasoning steps from the KV cache, reducing memory consumption while preserving accuracy. Third, Planned Diffusion uses semantic dependence to derive a denoising order for discrete diffusion, autoregressively generating a plan that specifies which chunks to denoise in parallel. Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL) Cite as: arXiv:2609.14850 [cs.AI] (or arXiv:2609.14850v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.14850 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-71] LLM s as Oracles: Reliance on LLM s for Subjective Personal Questions
【速读】： 该论文旨在解决用户在使用大语言模型（LLM）过程中逐渐将其视为“全知权威”（oracle）所带来的自主性与心理健康风险问题，即人们在面对主观性个人问题时过度依赖生成式AI进行判断与决策。其核心解决方案在于构建一个可量化的分类体系（typology）及基于LLM的方法，实现对这种AI依赖行为的大规模测量与分析。研究通过分析公开数据集（WildChat和ThoughtTrace，共68,000条提示）与隐私保护的数据捐赠工具（140,000条提示，52名参与者纵向数据），发现LLM作为“预言家”（LLM-as-oracle）的使用趋势自2023年至2026年持续上升，且在年轻群体中更为普遍；同时，用户往往对其自身依赖行为缺乏意识，但在看到分析结果后表现出显著不满。研究进一步识别出两大驱动因素：用户对AI的认知态度以及AI模型自身的响应行为，并据此提出支持用户自我反思与审慎决策的潜在干预路径。
链接: https://arxiv.org/abs/2609.14849
作者: Myra Cheng,Lujain Ibrahim,Grace Liu,Michelle S. Lam,Vishakh Padmakumar,Nick Madibekov,Diyi Yang,Dan Jurafsky
机构: Stanford University (斯坦福大学); University of Oxford (牛津大学); Carnegie Mellon University (卡内基梅隆大学)
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:We characterize how people are turning to LLMs as oracles: all-knowing authorities on subjective personal questions. Motivated by risks to users’ autonomy and well-being, we develop a typology and LLM-based methods to measure this form of AI reliance at scale and understand how people are offloading judgment and decision-making to AI. Applying our typology to public usage data (68K prompts from WildChat and ThoughtTrace), we find that LLM-as-oracle use has increased over time (2023-2026) and is more prevalent among younger users. We further build a privacy-preserving data donation tool to analyze individuals’ longitudinal usage data (140K prompts from 52 participants), identifying similar trends. People are often unaware of their own LLM-as-oracle use, and express dissatisfaction with this behavior after seeing our tool’s analysis. Finally, we identify two drivers of LLM-as-oracle use: people’s perceptions of AI and the behavior of AI models themselves, which motivate possible interventions to support users’ self-deliberation.
[NLP-72] Enemray: Toward Capable Language Models for Hassaniya
【速读】： 该论文旨在解决低资源语言——哈萨尼亚语（Hassaniya）在生成式AI模型中缺乏高质量、大规模训练数据及专用语言与文化能力的问题，从而实现通用场景下的自然交互。其核心挑战在于如何在不损失基础大模型（instruction-tuned model）通用能力的前提下，有效增强模型对哈萨尼亚语及其文化背景的深度理解与表达能力。解决方案的关键在于提出一种“稳定性-可塑性”（stability–plasticity）协同训练框架：通过分离语言习得与行为专业化阶段，采用分层选择性持续预训练（layer-selective continual pretraining）方式，仅对特定参数进行轻量级更新以保留原模型的核心能力；随后利用大规模自建并精心构建的监督微调数据集（涵盖公共资源与新采集/重构/构造的指令数据），结合策略生成的回放机制（policy-generated replay）从参考模型的行为分布中提取记忆信号，实现对话、文化、文学、任务导向及跨语言等多维度行为的精细化训练。该方法不仅显著提升了模型在英-哈翻译和毛里塔尼亚语翻译错误检测中的表现，且保持了原始基座模型在数学推理、知识问答、代码生成和函数调用等方面的强通用性能，为低资源语言的专用化建模提供了可复现的技术路径。
链接: https://arxiv.org/abs/2609.14829
作者: Cheikh Ahmed
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:We introduce Enemray, a Hassaniya-centric language model that enables general-purpose interaction in Hassaniya. Enemray is trained around a stability–plasticity objective: acquire strong Hassaniya linguistic and cultural competence while preserving the general reasoning, multilingual, instruction-following, and safety behaviors of a capable instruction-tuned model. The development pipeline separates language acquisition from behavioral specialization. A separately assembled continual-pretraining corpus provides broad exposure to natural Hassaniya and Mauritanian text; layer-selective continual pretraining learns a compact language-specific parameter update; that update is transferred into the instruction-tuned parameter space; and supervised post-training develops conversational, cultural, literary, task-oriented, and cross-lingual behavior. The supervised corpus integrates selected public Hassaniya and Mauritanian resources with a substantially larger body of newly collected, reconstructed, curated, and constructed instruction data, while policy-generated replay provides a retention signal from the reference model’s own behavior distribution. The resulting collection is substantially larger and broader in purpose than existing Hassaniya text resources. In evaluation, Enemray achieves the strongest English to Hassaniya translation among the compared open and proprietary models and the highest overall score on Mauritanian translation error detection, while retaining most of the general capabilities of its instruction-tuned base model on mathematical reasoning, knowledge, code generation, and function calling. This report describes the motivation, data construction, model design, training methodology, and evaluation of Enemray.
[NLP-73] Route Dont Fix: Regime-Dependent Decoding Correction and a Trajectory-Gated Router for Reliable Clinical LLM Answer Selection
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在临床问答任务中因幻觉（hallucination）导致的安全性问题，尤其针对现有解决方案如检索增强、微调及外部验证器等需额外基础设施、可能引入延迟或增加模型调用开销的局限性。其核心挑战在于：尽管推理时纠正方法利用模型内部的logit信号，但固定变换策略难以适配所有问题；而现有纠正器虽在真实性压力测试中提升约10个百分点，但在临床多选题基准上增益微乎其微，原因在于指令微调使输出概率高度集中于单一答案，导致终端熵（terminal entropy）过低，限制了纠错空间。为此，论文提出ALTAS——一种基于单次前向传播中提取的终端熵与晚期层线性度（late-layer linearity, R²）动态选择解码策略的方法，即对每个问题自适应地在贪婪解码与晚期轨迹纠正之间路由。该方法无需训练分类器、探测器或头部网络，仅通过候选答案logits进行路由决策，仅引入6.5%的延迟开销。实验表明，对所有问题统一应用纠正时，ALTAS在3B和8B模型上分别使TruthfulQA得分提升11.4和10.0个百分点（p < 10⁻¹⁰）；而在按问题门控式启用时，仍保持8.3至9.5个百分点的显著增益，同时在MedQA、PubMedQA和MedHallu等临床基准上维持“无害”表现（差异<1个百分点，无统计显著性），并通过了冻结阈值、评分规则及领域标签等多维度验证。因此，该方案的关键创新在于：基于可解释的内部表征（终端熵与R²）实现无需训练的动态路由机制，兼顾性能提升与临床场景下的安全性与稳定性。
链接: https://arxiv.org/abs/2609.14825
作者: Zeyu Dong,Benjamin Wang,Joyee W. Jin
机构: Crestwood Preparatory College; St. Theresa of Lisieux Catholic High School; Department of Computer Science, University of Toronto
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Large language models (LLMs) are often deemed unsafe for clinical question answering because of their tendency to hallucinate. Retrieval augmentation, fine-tuning, and external verifiers require new infrastructure that clinical governance must approve and may add latency or extra model calls. Inference-time correction uses the model’s internal logit signals, but a fixed transformation need not suit every question. A corrector that improves accuracy by about ten percentage points on a truthfulness stress test yields negligible gains on clinical multiple-choice benchmarks, where instruction tuning concentrates output probability on one answer and leaves low terminal entropy. We introduce ALTAS, which reads terminal entropy and late-layer linearity ( R^2 ) from one forward pass to choose per question between greedy decoding and late-layer trajectory correction. No classifier, probe, or head is trained; the router operates on candidate-answer logits and adds 6.5% latency overhead. Applied to every question, the correction improves TruthfulQA over greedy at 3B and 8B by 11.4 and 10.0 percentage points, respectively ( p10^-10 ). Gated per question, ALTAS retains gains of 8.3 to 9.5 percentage points while keeping MedQA, PubMedQA, and MedHallu within a one-percentage-point do-no-harm band, with no statistically significant differences from greedy. The method passes verification sweeps over frozen thresholds, the scoring rule, and the domain label.
[NLP-74] MedTRACE: Tool-Augmented Multimodal Clinical Reasoning Agents for Evidence-Grounded Decision-Making
【速读】： 该论文旨在解决多模态临床决策中缺乏对证据充分性、工具使用需求及诊断不确定性进行显式评估的问题，现有模型通常直接将异构医疗数据（如电子健康记录、医学影像和生理信号）映射至诊断结果，忽视了推理过程的可解释性与可靠性。其解决方案的关键在于提出MedTRACE——一种基于工具增强的多模态临床推理代理，通过构建统一的患者状态表示，并在迭代的假设生成、工具感知推敲与证据验证循环中动态调用视觉定位、证据检索和结构化解析等工具，以精准获取诊断相关区域、检索临床知识与相似病例、提取结构化发现。所获证据被存入证据记忆库，由一致性验证器对当前假设进行确认或修正，最终输出带有支持证据、可审计的推理路径及校准置信度的诊断结果。实验表明，MedTRACE在多个多模态临床诊断基准上显著提升诊断准确率（+5.4%）、AUROC（+4.7个百分点），并大幅改善证据选择F1（+8.2个百分点）、视觉定位交并比（+6.5个百分点），降低期望校准误差（-31.6%）与无依据诊断错误（-27.8%），验证了主动证据获取与验证机制在提升多模态临床决策准确性、可解释性与可靠性方面的有效性。
链接: https://arxiv.org/abs/2609.14823
作者: Ji Lu,Lifei Liu,Haoran Yu,Xianglong Wang,Yiru Fang,Kuo Yang,Huiran Duan,Jianping Gou
机构: Vanderbilt University (范德比尔特大学); Wichita State University (威奇托州立大学); University of Florida (佛罗里达大学); Wyze Inc. (威兹公司); Northeastern University (东北大学); City University of New York (纽约市立大学); Southwest University (西南大学)
类目: Computation and Language (cs.CL)
备注: Accepted to the 22nd International Conference on Advanced Data Mining and Applications (ADMA 2026)
Abstract:Multimodal clinical decision-making requires reliable reasoning over heterogeneous evidence from electronic health records, medical images, and physiological signals. Existing models typically map these inputs directly to diagnoses without explicitly assessing evidence sufficiency, tool-use requirements, or diagnostic uncertainty. This paper presents MedTRACE, a tool-augmented multimodal clinical reasoning agent for evidence-grounded decision-making. MedTRACE uses modality-specific encoders to construct a unified patient-state representation and performs an iterative loop of hypothesis formation, toolaware deliberation, and evidence verification. It dynamically invokes visual grounding, evidence retrieval, and structured parsing tools to locate diagnosis-relevant regions, retrieve clinical knowledge and similar cases, and extract structured findings. The acquired evidence enters an evidence memory, where a consistency verifier confirms or revises the current hypothesis. MedTRACE outputs a diagnosis together with supporting evidence, an auditable reasoning trace, and calibrated confidence. Experiments on multiple multimodal clinical diagnosis benchmarks show that MedTRACE improves diagnostic accuracy by 5.4% and AUROC by 4.7 percentage points over the strongest baseline. It also improves evidenceselection F1 by 8.2 percentage points and visual-grounding IoU by 6.5 percentage points, reduces expected calibration error by 31.6%, and decreases unsupported diagnostic errors by 27.8%. These results demonstrate that active evidence acquisition and verification improve the accuracy, interpretability, and reliability of multimodal clinical decisionmaking.
[NLP-75] A primer on evaluation methods for large language models in healthcare
【速读】： 该论文旨在解决生成式 AI（Generative AI）在医疗领域应用中评估困难的问题，尤其针对其输出具有概率性与开放性、行为受提示设计及上下文累积影响等特性所带来的挑战。解决方案的关键在于构建系统化、多维度的评估框架，涵盖研究设计原则、统计方法、能力评估与临床情境评估四大核心方面。其中，能力评估通过多选题、代理式（agentic）及多轮对话等基准测试，结合令牌使用量等操作指标，全面衡量模型性能；临床情境评估则聚焦自由文本输出的准确性，采用人工评审、以大模型为评判者（LLM-as-a-judge）以及临床试验范式等方法。论文强调评估方法必须与具体研究问题相匹配，从而为医疗领域生成式 AI 的严谨评估提供可操作的实践指导。
链接: https://arxiv.org/abs/2609.14819
作者: Suzannah E McKinney,Phuc Vu,Samuel A Justice,Christopher Humphries,Alyssa Pradhan,Timothy J Keyes,Bernardo C Bizzo,Keith J Dreyer,Sarah F Mercaldo,James M Hillis
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) have a growing range of applications in medicine, and their evaluation is critical for ensuring they provide benefit and not harm. This evaluation can be more challenging than traditional machine learning for many reasons, including probabilistic and open-ended outputs, and behavior that shifts with prompt design and accumulated context. This review covers four key areas of LLM evaluation: principles of study design, statistical methods, capability evaluation and clinical context evaluation. Capability evaluation considers different benchmarks, including multiple-choice, agentic and multi-turn benchmarks, alongside operational metrics like token usage. Clinical context evaluation addresses establishing accuracy of free text outputs, such as human review and LLM-as-a-judge, and clinical trial approaches. Across sections, we describe underlying concepts and potential pitfalls, while emphasizing the importance of aligning evaluation methods with the research question. Together, this article aims to provide a pragmatic basis for designing and executing rigorous evaluations of healthcare LLMs.
[NLP-76] one on a Budget: A Reference-Free Metric for Lexical Tone in Massively Multilingual Text-to-Speech
【速读】： 该论文旨在解决语音合成（TTS）中声调信息评估缺失的核心问题，即传统自动评价指标字符错误率（CER）无法有效衡量多音调语言（如约鲁巴语）中的声调准确性。尽管约鲁巴语依赖声调区分词义（如“dok”[中调]、 “dòk”[低调]、 “dók”[高调]），但标准CER基于自动语音识别（ASR）输出计算，而ASR通常忽略声调标记，导致合成系统即使在声调上严重错误（如将“丈夫”误合成“车辆”）仍可获得较低的CER。为此，论文提出DunDun——一种无需参考语料库、无监督、基于声学特征的参考无关式词汇声调评估指标。其核心创新在于：利用输入文本中的声调标记（以变音符号形式存在）作为真实声调序列，从合成语音的基频（pitch track）提取预测声调序列，并通过三重验证机制（基频重构破坏性实验、声调反转一致性测试、母语者盲听对比实验）证明其有效性。实验表明，当声调被人为破坏时，DunDun显著下降而CER不变；在300段母语录音中，声调反转使二分类准确率降至0.14（远低于0.35的随机水平），验证了评分路径的可靠性；三位母语者在67次盲听试验中89.6%选择声调正确的音频片段（置信区间80.0%-94.8%，p < 1e-4）。进一步应用于大规模零样本多语言TTS模型发现，尽管未进行约鲁巴语微调，其声调得分已接近母语锚点（0.567±0.02），且声调性能在数小时内即趋于饱和，而CER仍较高，说明声调表现与发音错误不可等同。相比之下，在非声调语言斯瓦希里语中，CER能有效反映改进，凸显了评估指标的语言依赖性。研究最终发布完整的评估协议与DunDun工具包，为多音调语言的声调建模提供可靠、可复现的基准。
链接: https://arxiv.org/abs/2609.14817
作者: Moses Daudu,Adeola Enitan Bamidele,Honor-Jesus Bezaleel
机构: 未知
类目: Computation and Language (cs.CL)
备注: 9 pages, 2 figures
Abstract:In Yorùbá, pitch alone separates \dok\do (husband, Mid), \dok\dò (vehicle, Low), and \dok\dó (hoe, High) – the diacritics ARE the tone marks. Yet character error rate (CER), the standard automated metric for text-to-speech (TTS), is in practice computed from ASR output that drops those marks: a synthesizer can ace CER and still say vehicle for husband. We introduce DunDun – named for the dùndún, the Yorùbá talking drum that speaks through pitch alone – an automated, reference-free lexical-tone metric that needs no tone-labelled corpus. The gold High/Mid/Low sequence is read from the input text’s diacritics (in TTS that text exists by construction, so no reference recording is needed); the prediction comes from the audio’s pitch track. We validate three ways. Flattening pitch with PSOLA resynthesis collapses DunDun while CER does not move. Inverting High and Low in the answer key of 300 native recordings drives the two-class readout to 0.14, symmetrically below its 0.35 chance level – a consistency check on the scoring path, not independent evidence. And three native listeners, over 67 blind A/B trials, pick the tone-correct clip 89.6% of the time (95% CI 80.0-94.8; p 1e-4); whether DunDun tracks those judgements trial by trial is not resolved at this sample size. Applied to a massively multilingual zero-shot TTS model, DunDun shows what CER cannot: Yorùbá tone sits near the native anchor before any Yorùbá fine-tuning (0.567 +/- 0.02 over five decode seeds vs. 0.596; chance 0.33), despite the 21.4% CER the model’s own paper reports; and a few hours of clean audio halve CER (5.6% to 2.7% by 5h, 1.7% by 15h) while tone saturates within the hour. On non-tonal Swahili, CER already captures the gains: the metric a language needs is language-dependent. We release the metric and the complete validation protocol.
[NLP-77] Crypto Accounting Bench: Evaluating Frontier and Open-Weight Models on Crypto-Asset Accounting Tasks
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在金融领域中对加密资产交易会计分录的完整重构能力评估问题，即检验前沿模型与开源权重语言模型是否能够准确还原企业实际记录的、包含完整账户、借贷方向、金额、币种及高精度资产数量的会计分录。其解决方案的关键在于构建Crypto Accounting Bench（CAB），一个涵盖118个评估任务的基准测试体系，任务源自7个匿名化组织的真实交易场景，综合了交易机制、资产数量与本位币价值、钱包与法律实体上下文、交易对手证据、关联交易环节、重复性、税级凭证以及企业完整的总账科目结构等多维度信息。每个任务的目标是生成满足所有评分标准和必要校验条件的平衡结构化分录。通过在12个模型（包括专有前沿系统与开源模型）上进行每项任务三次独立尝试，共生成4,248条推理轨迹，并采用均分（Mean Score）、前三最佳得分（Best@3）和三试通过率（Pass@3）三个核心指标进行评估。结果显示，领先模型的均分为77.43%，最高Pass@3为56.78%；基于最优尝试的确定性诊断分析表明，基础金额匹配度较高（97.8%），但记账账户选择准确性仅为56.3%，结合失败分析进一步揭示出账户选择与完整分录生成仍是当前面临的主要挑战。
链接: https://arxiv.org/abs/2609.14811
作者: Kareem Khattab,Omar Khattab,Mohamed Ibrahem
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 6 pages, 8 figures, 3 tables. Dataset: this https URL ; code and evaluation: this https URL
Abstract:We introduce Crypto Accounting Bench (CAB), a benchmark for assessing whether frontier and open-weight language models can reconstruct the complete journal entry that an organization actually posted for a crypto-asset transaction. CAB contains 118 evaluation tasks drawn from 7 pseudonymized organizations. Each task combines transaction mechanics, asset quantities and base-currency values, wallet and legal-entity context, counterparty evidence, related transaction legs, recurrence, tax-lot evidence, and the organization’s complete chart of accounts. The target is a balanced structured entry with every required account, side, amount, currency, and full-precision asset quantity. We evaluate 12 models spanning proprietary frontier systems and open-weight releases over 3 independent attempts per task, producing 4,248 trajectories. We report 3 metrics: Mean Score, Best@3, and Pass@3. Pass@3 is the fraction of tasks with at least 1 of 3 attempts that satisfies every rubric criterion and required gate. The leading model reaches 77.43% Mean Score, while the best Pass@3 is 56.78%. Deterministic diagnostics, read from each task’s best of 3 attempts and macro-averaged across the 12 models, show higher base-amount agreement (97.8%) than deciding-account accuracy (56.3%). Together with the failure analysis, these results identify account selection and complete-entry composition as the main remaining challenges on CAB.
[NLP-78] Mind Which Bird You Favour: Parameterizing Adequacy-Fluency Balance in Meta-Evaluation of Machine Translation
【速读】： 该论文旨在解决机器翻译元评估（machine translation meta-evaluation）中准确度（adequacy）与流畅度（fluency）之间权衡失衡的问题。由于元评估数据集中的翻译系统组合随年份和语种对变化显著，其系统分布无法代表真实系统分布，导致准确度与流畅度的平衡往往不具代表性且易变，尤其在敏感领域中这一问题更为突出。为应对该挑战，论文提出将此平衡视为可调参数，并通过重新加权现有系统来实现目标平衡，同时最小化对均匀加权的偏差，以确保被评估系统的真实性和代表性。其解决方案的关键在于设计一种具有理论保证的精确优化算法，结合剪枝机制高效计算最优权重。此外，为验证元评估的内部一致性，论文构建了评分器增强框架（scorer-augmentation framework），建立评分器之间的已知相对关系作为基准。实验结果表明，该重加权方法能有效控制准确度-流畅度平衡，保持内部一致性，优于先前方法。最后，研究还系统分析了主流评分器在该参数扫描范围内的表现。
链接: https://arxiv.org/abs/2609.14795
作者: Behzad Shayegh,Niloofar Kazemi
机构: Independent Researchers(独立研究员)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted by 11th Conference on Machine Translation (WMT26)
Abstract:There is a tradeoff in machine translation meta-evaluation between prioritizing alignment with adequacy versus fluency. The balance depends on the combination of translation systems in the meta-evaluation dataset. This system set is a small, filtered sample whose characteristics change heavily across years and language pairs; it does not represent the true system distribution. Consequently, the adequacy-fluency balance is often unrepresentative and subject to change. For sensitive domains, controlling this balance is critical. We expose this balance as a tunable choice. To achieve a target balance, we reweight existing systems while minimizing distortion from uniform weighting, ensuring the evaluated systems remain real and representative. We provide an exact optimization algorithm with theoretical guarantees and pruning mechanisms to compute these weights. To validate meta-evaluation internal consistency, we design a scorer-augmentation framework that establishes a known relative identity for the scorers. Results demonstrate that our reweighting method effectively controls the adequacy-fluency balance and preserves the internal consistency, outperforming prior approaches. Finally, we analyze the performance of popular scorers across a sweep of this parameter.
[NLP-79] Language-Guided Representation Learning for Robust Cross-Sensor Material Recognition IROS2026
【速读】： 该论文旨在解决触觉传感器在不同硬件配置下（如光学特性、弹性体材料、光照条件差异）导致的泛化能力差的问题，即基于视觉的触觉传感器对同一材质产生的感知表征存在显著差异，限制了模型在多传感器场景下的迁移性能。其核心解决方案是提出一种语言引导的蒸馏框架，利用语言作为跨传感器的语义监督信号，将触觉图像与语言嵌入映射到共享的语义空间中。语言编码了触觉的高层语义属性（如粗糙、柔软、滑腻），这些属性在不同传感硬件间保持不变，从而提供了一种天然的传感器无关的监督信号。研究构建了一个包含39,000个样本的触觉-语言数据集，由人工标注材料标签，并训练触觉编码器实现传感器特定触觉图像与语言嵌入之间的对齐。实验表明，该方法在少样本学习（100次采样设置下达到95%准确率）、跨传感器迁移任务（平均提升13.3%准确率）以及六个现有触觉数据集上均表现出显著性能增益（最高达19%），验证了语言引导蒸馏在实现可扩展、硬件无关触觉表示学习方面的有效性。
链接: https://arxiv.org/abs/2609.14783
作者: Mashood M. Mohsan,Muhayy Ud Din,Binzhao Xu,Ahmad Abubakar,Irfan Hussain
机构: Khalifa University Center for Autonomous Robotic Systems (KUCARS), Khalifa University, UAE
类目: Robotics (cs.RO); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted for presentation at the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026)
Abstract:Robots need touch to manipulate objects safely and reliably, as many properties, such as softness, texture, and contact stability, are hard to infer from vision alone. However, vision-based tactile sensors yield different observations of the same material due to variations in optics, elastomer properties, and illumination, leading to poor generalization when trained on a single or multiple sensors. We propose a language-guided distillation framework for learning sensor-robust tactile representations. Language encodes high-level semantic properties of touch (e.g., rough, soft, slippery) that remain invariant across sensing hardware, providing a natural sensor-agnostic supervisory signal. We construct a 39K-sample touch-language dataset with human-annotated material labels and train a tactile encoder to align sensor-specific tactile images with language embeddings in a shared semantic space. We evaluate our approach for few-shot learning and cross-sensor transfer and benchmark it on six existing tactile datasets. Our method achieves 95% accuracy in the 100-shot setting, improves cross-sensor transfer by an average of 13.3% accuracy, and yields up to 19% accuracy gains across six existing tactile datasets. These results demonstrate that language-guided distillation enables scalable and hardware-agnostic tactile representation learning. Code and dataset are available at this https URL
[NLP-80] Func-R1: Incentivizing Mathematical Function Reasoning in Multimodal Large Language Models EMNLP2026
【速读】： 该论文旨在解决多模态大模型（Multimodal Large Language Models, MLLMs）在视觉情境下进行严谨数学推理时存在的模态干扰问题，即先进模型在执行文本计算推理时往往忽视或误读关键的视觉线索，导致推理失效。其核心解决方案是提出Func-R1框架，通过显式解耦的架构设计，结合分层后训练机制，逐步识别关键视觉证据并开展深度理论推理。其关键技术在于提出的感知对齐理论优化（Perception-Aligned Theoretic Optimization, PATO）策略，该策略在推理过程中动态校正异构视觉信息的同时，引导模型内化基础理论属性，实现视觉感知与符号逻辑的协同优化。实验结果表明，Func-R1在多个基准测试中表现最优，甚至在函数导向任务上超越GPT-5达8.4%。
链接: https://arxiv.org/abs/2609.14779
作者: Mingze Yin,Xiaohan Wang,Dian Li,Haichao Yao,Yilin Zhao,Youjun Chen,Gang Liu,Jintai Chen,Yiheng Zhu,Chang-Yu Hsieh,Aimin Pan
机构: Zhejiang University (浙江大学); Tencent(腾讯); HKUST (GZ) (香港科技大学（广州）); Zhongguancun Academy (中关村学院); VNET Group (万维集团); Zhejiang Lab (浙江省实验室)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 (2026 Conference on Empirical Methods in Natural Language Processing)
Abstract:Performing deliberate mathematical reasoning in visual contexts is a hallmark of advanced Multimodal Large Language Models (MLLMs) and requires a sophisticated synthesis of perceptual grounding and symbolic logic. However, in the realm of mathematical functions, our investigation reveals a critical modality interference phenomenon: even advanced models, while performing textual computational reasoning, tend to disregard or misinterpret essential visual cues. To address this challenge, we propose Func-R1, which synergistically harmonizes precise visual perception and rigorous logical reasoning. Concretely, built upon an explicitly decoupled architecture, we employ a hierarchical post-training framework to progressively identify critical visual evidence and conduct in-depth theoretical reasoning. Furthermore, the Perception-Aligned Theoretic Optimization (PATO) strategy is proposed to steer policy updating towards internalizing fundamental theoretical properties while dynamically rectifying heterogeneous visual information throughout the reasoning process. Extensive experiments across diverse benchmarks demonstrate that Func-R1 delivers the optimal performance among open-source MLLMs, even surpassing GPT-5 with an 8.4% improvement on MathVerse’s function-oriented tasks.
[NLP-81] Pull: Lazy Materialization of Working Memory for Stateful LLM Conversations
【速读】： 该论文旨在解决大规模语言模型（LLM）对话中长序列上下文管理的效率与信息保全难题：随着对话轮次增长至数百轮，全量上下文注入会导致上下文令牌开销呈O(N²)级增长，而传统的有损摘要或硬截断方法则会不可逆地丢失历史状态。其核心解决方案是提出一种名为Pull的会话路由机制，通过本地化、确定性的净化器（Purifier）构建可寻址的元数据目录（零次调用大模型，毫秒级延迟），在查询时仅按需惰性展开所需对话轮次，未展开的轮次保持压缩但可访问状态。与不可逆压缩不同，Pull的展开过程具有可逆性，后续查询可重新扩展任意已折叠的轮次。实验表明，在LoCoEval数据集上，Pull在单跳任务中将每轮查询的上下文令牌数减少75.1%，质量无显著下降（Δ = -0.002, n.s.）；在多跳任务中减少72.0%且无质量损失（Δ = +0.017）。控制性路由基准测试进一步验证了实体生命周期追踪是实现与距离无关路由的实证前提。在BEAM 1M数据集上，相比截断基线，Pull使F1指标提升55.2%。
链接: https://arxiv.org/abs/2609.14773
作者: Jiangang Chen
机构: Chengdu Beiluoshimen Technology Co., Ltd.(成都北洛石门科技有限公司)
类目: Computation and Language (cs.CL)
备注: 12 pages, 3 figures. An earlier version was publicly released on Zenodo (DOI: https://doi.org/10.5281/zenodo.21984705 ). Code and reproduction scripts: this https URL
Abstract:As LLM conversations grow to hundreds of turns, full-context injection incurs O(N^2) cumulative token costs, while lossy summarization or hard truncation irreversibly discards historical state. We propose Pull, a session router that maintains an addressable metadata directory via a local, deterministic Purifier (zero LLM calls, millisecond-level latency). At query time, the LLM lazily materializes only the turns it needs; unmaterialized turns remain accessible but collapsed. Unlike irreversible compression, Pull’s materialization is reversible; subsequent queries can expand any collapsed turn. On LoCoEval (128 conversations, 12,780 turns), Pull reduces per-query context tokens (Phase 2) by 75.1 percent on single-hop tasks with equivalent quality ( \Delta = -0.002 , n.s.) and by 72.0 percent on multi-hop tasks with no quality loss ( \Delta = +0.017 ). A controlled routing benchmark (7,831 queries x 10 methods) shows that entity lifecycle tracking is empirically a prerequisite for distance-independent routing. On BEAM 1M (14 conversations, 263 questions), Pull improves F1 by +55.2 percent over a truncation baseline.
[NLP-82] How broad is that claim? Mapping Generalisation in NLP Research EMNLP2026
【速读】： 该论文旨在解决科学文献中普遍存在的泛化表述（generalisation）语义模糊所带来的问题，尤其关注其可能引发的科学发现误读与过度概括风险。其核心解决方案在于提出一个面向科学领域的全面泛化分类体系——NLPGenX，该体系可对陈述句的泛化程度及其在文本中的表述框架进行标注；并基于大语言模型（LLM）构建自动化分类框架NLPGenA，实现对科学论文中句子的五类泛化等级自动识别。研究通过人工标注验证框架有效性，并利用该框架构建大规模自然语言处理（NLP）领域论文数据集（NLPGens），包含泛化程度、情态表达（hedging）及模糊描述词等辅助标签。该数据集被用于分析不同会议和子领域中泛化使用的趋势，并探究其与引用次数、情态表达及模糊词汇之间的关联性。
链接: https://arxiv.org/abs/2609.14770
作者: Chenxin Diao,Nataliya Stepanova,Emily Allaway
机构: University of Edinburgh(爱丁堡大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Main; the dataset and code are available at this https URL
Abstract:Generalisations are common in scientific communication, even though they are semantically ambiguous. An automated method is needed to identify and categorise claims according to their level of generalisation, in order help detect an over-reliance on generalisations and possible misrepresentations of scientific findings. We introduce a comprehensive taxonomy of generalisations in the scientific domain, NLPGenX, which labels claims according to their level of generality and framing within the text. We operationalise this taxonomy with an LLM-powered framework, NLPGenA, that automatically classifies sentences from scientific articles into 5 different generalisation classes. We validate our framework with human annotators and use the framework to construct a large-scale dataset of NLP papers annotated according to generality, with auxiliary labels for hedging and vague descriptors (NLPGens). We use NLPGens to analyse the use of generalisations in NLP papers across multiple venues and subdomains, and to examine associations with citation counts, hedging, and vague descriptors.
[NLP-83] Refusal Reads Only a Slice of What the Model Knows: Harm-Keyed Routing and Its Exceptions Across Model Families
【速读】： 该论文旨在解决生成式模型在对齐（alignment）过程中，拒绝有害请求的行为究竟是基于深层道德理解，还是仅依赖于浅层、可被轻易修改的特定信号。其核心问题是：模型的拒绝行为（refusal）所依赖的信息输入是否源于其在预训练阶段已具备的道德认知能力，还是仅仅通过后训练阶段构建的独立控制机制实现。解决方案的关键在于通过因果分析揭示拒绝行为的内在机制——研究发现，道德理解能力是预训练阶段自然形成的低秩道德子空间（low-rank moral subspace）的产物，而拒绝机制则是在对齐阶段新增的、高度受限的构造，其决策信号主要集中在狭窄的控制令牌通道中，且与道德判断决策正交。通过对OLMo-3模型进行嵌套互换秩扫描（nested interchange rank sweep），研究证实：随着所读取的道德子空间范围扩大，模型的道德判断能力持续增强，但拒绝响应水平却趋于饱和，且约四分之三的拒绝行为因果输入完全位于道德子空间之外。这表明，拒绝行为主要读取的是“伤害感知”（harm percept），而非道德内容本身。不同模型家族表现各异：Llama能读取广泛的道德信息，Qwen表现出部分超越单一伤害线索的倾向但未明确，GPT-OSS仅读取伤害信号且其拒绝理由在其自身推理轨迹中可双向论证。因此，拒绝机制本质上是围绕模型大部分知识绕行的低秩结构，只需一次秩一编辑即可移除。该研究提出一个开放性问题：若扩大拒绝机制的读取范围，是否会同步深化其行为？
链接: https://arxiv.org/abs/2609.14759
作者: Orion Reblitz-Richardson
机构: Distiller Labs
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 34 pages, 6 figures, 15 tables. Companion methods note: “Calibrating Interpretability Instruments Before Trusting Their Verdicts” (submitted concurrently). Per-unit arrays: this https URL . Code: this https URL
Abstract:Alignment applied after pretraining is shallow in a measurable way: a single direction in a model’s residual stream can be edited out, and the model stops refusing harmful requests. That fact says how easily refusal can be removed, not what the refusal decision was reading in the first place. We ask what it reads, and we separate that from what the model comprehends. Across four open-weight models spanning three families, moral comprehension is native to pretraining: a low-rank moral subspace crystallizes during pretraining, and alignment rotates it once without rebuilding it. The refusal gate, in contrast, is a fresh post-training construction with only a weak pretraining precursor, written into a narrow control-token channel where the refusal decision is orthogonal to the moral-judgment decision. The central result is causal and comes from one model, OLMo-3. A nested interchange rank sweep patches successively larger slices of the moral subspace between matched requests and reads how much of refusal’s response transfers: as the basis widens, moral judgment keeps reading more of it, while refusal levels off at the level of a single harm direction, and about three-quarters of refusal’s causal input lies outside the moral subspace altogether. Refusal reads the harm percept, not the moral content that judgment reads on the same patches. The picture is not uniform across families. Llama reads broad moral content; Qwen reads beyond the single harm cue but is unresolved at our sample size; GPT-OSS reads harm, and its refusals can be argued in either direction by its own reasoning trace. Where refusal reads only a low-rank slice and routes around the bulk of what the model knows, a rank-one edit removes it. Whether widening what refusal reads would also deepen the behavior is the open question this raises.
[NLP-84] Fabrication After Tool Failure: Tool-Augmented Agents Assert Values Their Tools Did Not Return
【速读】： 该论文旨在解决生成式 AI 在工具调用失败后仍可能产生不诚实响应的问题，即模型在无法获得有效工具输出时，仍会虚构答案或编造拒绝理由，从而影响系统可靠性。其核心问题在于：当前评估体系仅关注模型是否得出正确答案，而忽视了在工具失败情境下模型是否能如实报告失败状态。解决方案的关键在于引入一个明确的失败状态标识机制——通过在系统提示中强制要求模型在回答前必须声明 retrieval_status: OK 或 FAILED，从而显著降低不诚实行为的发生率。实验表明，该策略可将不诚实率从14.10%降至0.87%，且该效果不受提示风格影响，具有高度可迁移性；同时，所生成的状态标志在99.7%–99.9%的情况下准确反映真实失败状态，可仅通过正则表达式实现高效运行时检测，为构建可信的工具增强型语言模型系统提供了可靠的技术路径。
链接: https://arxiv.org/abs/2609.14758
作者: Arham Sethi,Arsen Kenzhebayev,Saanvi Paturi,Vatsal Raina,Vyas Raina,Ivaxi Sheth
机构: The Shishukunj International School(锡舒克琼国际学校); Haileybury Astana(海利伯里阿斯塔纳); UWCSEA East Campus(世界联合学院新加坡分校东校区); Apta AI Spark AI Research(阿普塔AI 星火人工智能研究); Spark AI Research(星火人工智能研究)
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL)
备注: 11 pages, 4 tables
Abstract:Tool-augmented language models are evaluated on whether they reach the right answer, not on whether they report honestly when a tool fails to supply one. We isolate this post-failure decision with a benchmark of 1,024 items spanning 16 internal-system domains and eight tool-failure types, in which a tool call is enforced and the returned payload is guaranteed to be unusable. Under a deployment-style system prompt, 14.10% of responses are dishonest: the model either asserts a value the payload cannot support or declines while citing a fabricated policy or capability limit. The rate is governed almost entirely by whether the failure is signalled. When the tool returns status:error, dishonesty is absent (0.0%); when it returns status:ok with a redacted, corrupted, stale, malformed, empty or truncated value, dishonesty reaches 45.3%. The behaviour is not an artefact of our prompts: it appears under a neutral prompt (10.17%) and under the shipped prompt of every production agent framework we evaluate, reaching 24.67% under CrewAI’s, and none of the nine frameworks we audit specifies what the model should do when a tool fails. Comparing prompt-level defences, we find that the operative variable is not deference to tool output but the absence of a named failure state. Appending a single sentence that requires the model to emit retrieval_status: OK or FAILED before answering reduces dishonesty from 14.10% to 0.87%, with one item of 688 worsening against 92 improving, and transfers unchanged into three foreign agent scaffolds. The emitted flag is faithful in 99.7-99.9% of declarations, giving a runtime detector that needs only a regular expression.
[NLP-85] Calibrating Interpretability Instruments Before Trusting Their Verdicts
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）内部因果解释中因测量方法缺陷而导致的误判问题。具体而言，现有因果可解释性研究依赖于诸如投影、余弦相似度、消融差值或交换修补（interchange patch）等量化指标，但这些测量在特定可诊断的情况下会失效，产生看似合理实则错误的结果，导致“故障仪器读数”被误认为有效发现。例如，协方差匹配的零假设可能因饱和而使所有方向均表现为典型，头级归因在重排序归一化架构下可能高估真实残差写入三倍，交换修补结果可能出现符号混沌，而读取判断（read-from verdict）则可能源于对模型已做出决策层之后的无效测量。本文基于拒绝行为与道德表征的因果可解释性项目，系统梳理了六种此类测量失效模式，涵盖多个论文及一个包含四款模型的开源权重面板，每种失效模式均通过可检测的触发信号（如重排序归一化、大规模激活、低维决策通道）识别，并提供相应的验证协议。其解决方案的核心在于建立四项基本操作规范：以正向控制阶梯校准、通过正交单元认证、在投入计算资源前评估统计功效、以及将所有读取结论明确标注其相对于模型决策深度的位置。该证据基于同一研究框架内三个家族中的四种架构，跨项目外部复现仍需未来工作完成。
链接: https://arxiv.org/abs/2609.14754
作者: Orion Reblitz-Richardson
机构: Distiller Labs(迪斯蒂勒实验室)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 16 pages, 3 figures, 1 table. Companion to “Refusal Reads Only a Slice of What the Model Knows” (submitted concurrently). Per-unit arrays: this https URL . Code: this https URL
Abstract:Causal claims about large language model (LLM) internals rest on measurements. Those might include a projection, a cosine, an ablation delta, or an interchange patch among others. These measurements fail in specific, diagnosable ways that return a plausible number instead of an error, so a broken instrument can easily read as a finding. A covariance-matched null can saturate until every direction looks typical, a per-head attribution can overshoot the true residual write threefold on reordered-normalization architectures, an interchange patch can go sign-chaotic because its outcome is pinned at a ceiling, or a read-from verdict can be an artifact of measuring past the layer where the model already decided. This note documents six such failures from a causal interpretability program on refusal and moral representation, spanning several papers and a four-model open-weight panel; each mode is established on one or two of the four. For each we give the tell that catches it and a protocol keyed to a detectable trigger (reordered normalization, massive activations, a low-dimensional decision channel), so we and readers can check whether a given setup is exposed. The discipline reduces to four moves: calibrate against a positive-control ladder, certify with an orthogonal cell, compute power before spending compute, and state every read-from verdict at a depth referenced to the model’s commitment. The evidence is four architectures across three families within a single program; external replication across programs is future work.
[NLP-86] Quantifying the Generation Modality Gap in Speech-Text Language Models
【速读】： 该论文旨在解决语音语言模型（speech language models）在生成连贯内容方面相较于文本及文本-语音联合模型表现滞后的问题，其核心挑战在于传统评估体系中语音与文本模型采用不同评价指标且训练数据分布不一致，导致难以准确量化两者之间的模态差距。为此，研究提出了一套统一的基于生成的评估框架，通过在匹配的数据分布和生成设置下对比仅语音、仅文本以及语音-文本联合建模的模型性能，系统评估生成内容在语义连贯性、局部语音结构、说话人一致性与声学质量、情绪分布等多维度的表现。关键发现表明：尽管语音-文本联合建模显著提升了基于转录文本的语义连贯性，尤其在缩小与大规模纯语音模型间的缩放差距方面效果明显，但其在音素级统计特性、说话人相似度和预测声学质量方面改善有限，甚至有所下降；而情绪相关的分布性指标则得到提升。这说明文本作为高效的语义信号，能够有效增强语音建模的语义理解能力，但其对语音细节和个性特征的建模仍存在局限，揭示了跨模态训练中语义与声学表征之间的权衡关系。
链接: https://arxiv.org/abs/2609.14743
作者: Ju-Chieh Chou,Jiawei Zhou,Karen Livescu
机构: TTI-Chicago(芝加哥技术研究所); Stony Brook University(石溪大学)
类目: Computation and Language (cs.CL)
备注: Accepted to SLT 2026, extended version with appendix
Abstract:Pure speech language models often lag behind text and speech-text language models in generating coherent content, but this gap is difficult to quantify because speech and text systems are typically evaluated with different metrics and trained on different data. We study the speech-text modality gap in a family of spoken language models, based on flow matching for continuous acoustic feature generation. We construct a unified generation-based evaluation suite that compares speech-only, text-only, and speech-text language models trained on matched data distributions and evaluated in matched generation settings. We evaluate generated continuations along multiple dimensions: semantic coherence, measured by transcribing generated speech and scoring it with a reference language model; local phonetic structure, measured by phone n-gram distributional statistics; speaker consistency and acoustic quality; and emotion-based distributional metrics. Across datasets, we find that joint speech-text modeling substantially improves semantic coherence. However, the improvement is not uniform across metrics: phone-level metrics change only modestly, speaker similarity and predicted quality are lower for speech-text continuations, while emotion-based distributional metrics improve. Compared with larger-scale speech-only models, our speech-text model closes much of the scaling gap in transcript-based semantic coherence, suggesting that text provides an efficient semantic training signal for spoken language modeling.
[NLP-87] Building Legal Reward Models for Grounding and Abstention ICML2026
【速读】： 该论文旨在解决大语言模型在法律等高风险领域应用中，因缺乏对检索证据的上下文依赖性评估能力而产生的推理不可靠问题。现有奖励模型多针对通用偏好进行优化，难以有效衡量在检索增强生成（RAG）场景下模型是否基于充分且相关的证据进行推理，或在证据不足时合理选择不回答。为此，论文提出一种将现有法律问答数据集转化为上下文偏好数据的框架，并据此构建了LegalRewardBench（LRB）基准，用于评估在噪声和证据不足条件下的法律生成任务中的上下文根基性。其解决方案的关键在于：通过长度平衡的数据增强策略显著提升上下文偏好数据的质量，使基于上下文直接偏好优化（contextual DPO）的方法在法律生成任务中表现更优，最优配置下性能较基线提升高达25.6个百分点；同时发现跨法域迁移的有效性——主要在维多利亚州刑事法数据上进行上下文精炼的模型，在美国法律基准测试中亦表现出显著改进（如《住房法》问答任务提升16.2个百分点），表明所提方法具备良好的泛化能力与可复现性，为构建和评估具有上下文根基性的法律领域奖励模型提供了坚实基础。
链接: https://arxiv.org/abs/2609.14739
作者: Rilton Franzone,Valentin Noël,Puyu Wang,Philip Torr,Fabio J. Fehr
机构: University of Oxford(牛津大学); Devoteam
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Published at ICML 2026 AI4Law Workshop
Abstract:Large language models are increasingly used in high-stakes domains such as law, where systems must ground their reasoning in retrieved evidence and abstain when that evidence is insufficient. However, existing reward models are largely optimised for general preferences rather than contextual grounding, limiting their ability to evaluate these behaviours in retrieval-augmented generation (RAG) settings. We introduce a framework for transforming existing legal QA datasets into contextual preference data and use it to construct LegalRewardBench (LRB), a benchmark for evaluating grounded legal generation under noisy and insufficient retrieval conditions. Across general and legal contextual evaluation, we find that contextual DPO improves grounded evaluation, but performance is sensitive to preference-data construction. Length-balanced augmentation substantially improves grounded legal evaluation, with the strongest configuration combining length-balanced legal and general contextual preference data and improving performance by up to \mathbf+25.6 pp over baseline. We further find evidence of cross-jurisdiction transfer: models contextually refined primarily on Victorian criminal-law data improve grounded evaluation on external US legal benchmarks, including a \mathbf+16.2 pp improvement on \textscHousing Statute QA. Together, these results provide a reproducible foundation for constructing and evaluating grounded legal reward models in retrieval-augmented settings.
[NLP-88] Depth and Scale in the Sub-150M Regime: JugnuLM-53M vs JugnuLM-110M
【速读】： 该论文旨在探究在保持训练方法不变的前提下，仅通过调整模型架构几何形态（即深度与宽度的权衡）实现参数规模扩展对语言模型性能的影响，核心问题是：在固定训练范式（如Qwen3-style解码器结构、分组查询注意力、旋转位置编码、SwiGLU激活函数、RMSNorm归一化、QK-Norm及z-loss等）和数据集（FineWeb-Edu）的情况下，模型容量与深度的增加是否能带来一致性的性能提升。其解决方案的关键在于采用“深而薄”（deep-and-thin）的23层×576隐藏维度设计，将预训练模型参数量从53.5M扩展至109.7M，从而在不依赖更多训练数据（仅约80亿token，低于基准120亿）的前提下，显著提升多项下游任务表现。实验结果表明，该110M模型在BLiMP上达到81.3%准确率，接近GPT-X2-125M（81.28%）但仅需约12%的参数量，验证了模型深度与容量增长的有效性。此外，研究通过系统性消融实验（ablation ladder）识别出有效改进项——值残差连接（value residuals）与Muon优化器组合可使ARC-Easy任务提升3.6个百分点，且不影响BLiMP性能，因此被保留；而多样化数据混合与基于1.7B教师模型的logit蒸馏虽能进一步提升性能，但伴随困惑度显著上升，经调参后无法持续获益，故未被采纳。最终结论为：在该模型规模与训练范式下，深度扩展与特定轻量级优化组件的组合是实现高效性能提升的核心路径。
链接: https://arxiv.org/abs/2609.14715
作者: Dushyant Rajput,Nirdesh Chauhan,Siddharth Kosaraju(AltSlate Labs LLP)
机构: AltSlate Labs LLP(AltSlate实验室)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:We scale our conventional sub-150M pretraining recipe from 53.5M to 109.7M parameters, holding the method fixed (Qwen3-style decoder with grouped-query attention, RoPE, SwiGLU, RMSNorm, QK-Norm, and a z-loss; FineWeb-Edu data) and changing only the geometry to a deep-and-thin 23-layer x 576-hidden design. The larger model improves across the board – BLiMP 78.1 - 81.3, ARC-Easy 51.4 - 52.5, WikiText-2 byte-perplexity 2.04 - 1.95 – and its 81.3% BLiMP essentially matches GPT-X2-125M (81.28) at about 12% fewer parameters. Notably the 110M model achieves this on fewer training tokens (about 8B vs 12B), so the gain is attributable to capacity and depth, not more data. Both models are deliberately conventional; this report is a clean scaling control and the baseline rung (R0) of an ablation study of what further improves models in this regime. An ablation ladder follows: value residuals (R1) and the Muon optimizer (R2) lift ARC-Easy by a cumulative +3.6 (52.5 - 56.1) at a near-flat BLiMP and are kept; a diverse data blend (R3) and two logit-distillation settings (R4a/R4b) are not kept – honest negatives. R3 pins ARC-Easy to FineWeb-Edu’s educational filtering rather than raw diversity; distillation from a 1.7B teacher can reach the class-leading ARC-Easy (56.99, matching GPT-X2-125M) but only at a perplexity cost that dialing KD down then erases – so R2 remains the best kept stack.
[NLP-89] he Garden of Forking Prompts: How Users Explore Narrative Space in Story Generation
【速读】： 该论文旨在解决当前故事生成评估基准中缺乏对用户在实际交互过程中迭代式探索叙事可能性行为的捕捉问题。现有基准多依赖静态、单次输入的提示（prompt），无法反映用户在生成故事时通过反复修改提示来调整角色、重定向情节或切换虚构宇宙的动态创作过程。其解决方案的关键在于构建两个大规模数据集：WildStories（包含275,635条自然发生的用户-聊天机器人对话中的故事生成提示，标注了故事格式、提示组件及显性程度）与WildEdits（包含24,291个编辑树，用于建模用户如何逐层修改基础提示并探索分支叙事路径）。基于此，研究提出一个涵盖四个方向（添加、删除、更改、扩展）和十四类目标（如情节、角色、类型等）的编辑类型框架，从而系统化分析用户利用大语言模型（LLM）在叙事空间中导航的行为模式，并进一步展示如何基于该框架自动生成多样化提示组合以用于更真实、动态的故事生成评估。
链接: https://arxiv.org/abs/2609.14677
作者: Advait Deshmukh,Nora Benedict,Melanie Walsh,Maria Antoniak
机构: University of Colorado Boulder (科罗拉多大学博尔德分校); University of Georgia (佐治亚大学); University of Washington (华盛顿大学)
类目: Computation and Language (cs.CL)
备注: Accepted at COLM 2026
Abstract:Large language models (LLMs) have changed the way people engage with stories. Drawing on public chatbot logs, we can see that when users generate stories, they iteratively edit their prompts to explore narrative possibilities, adjusting characters, redirecting plots, and swapping fictional universes. As aggregated data, these prompts represent rich traces of creative preference at scale. Yet story generation evaluation benchmarks rely on static, one-shot prompts that cannot capture this exploratory behavior. In this work, we study how users revise consecutive story prompts in the wild. Using a dataset of naturally occurring user-chatbot conversations, we construct WildStories, a sample of 275,635 story generation prompts (labeled with story format, prompt components, and explicitness), and WildEdits, a collection of 24,291 edit trees that model how users iteratively edit base story prompts and explore branching story possibilities. From these trees we develop a framework of edit types crossing four directions (adding, removing, changing, and extending) with fourteen targets (e.g., plot, character, genre). We then use our datasets and this framework to analyze user behavior in navigating narrative space via LLMs. Finally, we show how automated permutations based on the framework can be used for story generation benchmarking. Content Warning: This paper works with “wild” chatbot logs, which often include toxic and sexually explicit themes.
[NLP-90] Optimizing Sparse Outcomes Through Dense Behavioral Signals via Value-Guided Preference Distillation
【速读】： 该论文旨在解决多轮对话智能体在长期交互中因直接优化稀疏的最终结果而产生的信用分配困难与奖励黑客（reward hacking）问题。其核心挑战在于如何有效对齐对话策略以实现可持续的用户满意度和积极行为，同时避免单目标代理在真实场景下引发的策略退化。解决方案的关键在于将长周期对话优化建模为多目标强化学习问题，训练一个具备多头结构的价值模型，能够预测多个前瞻时域下的用户行为向量，并通过密集辅助行为信号的加权组合实现有效的信用分配。进一步地，研究提出了一种结合反事实用户模拟与经验证的对话级结果模型的安全评估框架，用于提前识别潜在失效模式；最终通过参考锚定偏好优化（reference-anchored preference optimization）将多目标价值偏好蒸馏至策略中，在远低于在线强化学习计算开销的情况下实现了媲美甚至超越的性能表现。实测结果表明，该方法显著提升了用户的长期留存率，并同步改善了正向行为及治疗过程指标。
链接: https://arxiv.org/abs/2609.14648
作者: Ziyi Zhu,Daniel R. Cahn,Thomas D. Hull,Caitlin A. Stamatis,Olivier Tieleman,Guilherme B. Freire,Jinghong Chen
机构: Slingshot AI; University of Cambridge(剑桥大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Aligning multi-turn dialogue agents is usually framed as matching turn-level human preferences, yet direct optimization of long-term outcomes is often ineffective and prone to reward hacking. We formulate long-horizon dialogue optimization as a multi-objective reinforcement learning problem and train a multi-head value model that predicts a vector of observed user behaviors across multiple look-ahead horizons. Our findings demonstrate that a scalarized composite of dense auxiliary behavioral signals enables effective credit assignment and optimization of sparse outcomes. However, optimizing unconstrained single-objective proxies might induce policy degradations that are harmful when the agent is exposed to real users. To identify these failure modes prior to deployment, we establish a safety framework combining counterfactual user simulation with a validated dialogue-level outcome model to evaluate preference weightings and policy optimization methods. Finally, we demonstrate that distilling multi-objective value preferences into the policy via reference-anchored preference optimization matches on-policy online RL at a small fraction of its compute budget. Live A/B testing confirms that our distilled policy significantly improves long-term user retention, while simultaneously enhancing the positive behaviors and therapeutic-process markers.
[NLP-91] CompCQR: Compositional Query Generation for Training-Free Conversational Search EMNLP2026
【速读】： 该论文旨在解决多轮对话中用户查询因模糊性和上下文依赖性而导致的检索效果不佳问题，核心挑战在于现有基于大语言模型（LLM）的对话查询重写（Conversational Query Reformulation, CQR）方法存在重复调用LLM带来的高成本以及与下游检索器之间语义对齐不足的问题。其解决方案的关键在于：首次观察到检索器对内容顺序高度敏感，据此提出一种无需训练的新型方法，通过组合少量原子语义组件以极低的LLM调用次数生成海量候选查询；同时利用LLM进行推理，构建兼顾精度与召回率的高质量文档集，有效捕捉用户核心意图。该框架具备良好的泛化能力，适用于各类开源与闭源大模型及稠密/稀疏检索器，在四个主流对话基准上均取得显著性能提升，相较先前最优基线实现最高达22.5%的相对MRR提升，且大幅减少LLM调用次数。
链接: https://arxiv.org/abs/2609.14646
作者: Yunah Jang,Kang-il Lee,Joongbo Shin,Kyomin Jung
机构: Seoul National University (首尔国立大学); LG AI Research (LG人工智能研究院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to Findings of EMNLP 2026
Abstract:Multi-turn interactions with LLMs are becoming increasingly common in information-seeking scenarios. However, user queries are often ambiguous and context-dependent, making them ill-suited for direct use as retriever queries. Conversational query reformulation (CQR) addresses this issue by rewriting the current utterance into a stand-alone query grounded in the dialogue history. Recent LLM-based CQR approaches achieve strong performance; however, their repeated LLM invocations and misalignment with downstream retrievers remain challenges. In this work, we begin from the observation that retrievers are highly sensitive to content ordering: simply reordering the same content can lead to changes in retrieval coverage and performance. Based on this, we propose a novel training-free method that generates a very large number of queries with minimal LLM usage by compositionally combining a small set of atomic components. We further apply LLM reasoning to construct a high-quality document set that balances precision and recall while capturing the user’s core intent. Our framework generalizes across both open- and closed-source LLMs as well as dense and sparse retrievers. It achieves strong performance on four widely used conversational benchmarks, with up to 22.5% relative MRR improvement over the previous state-of-the-art baseline with far fewer LLM calls.
[NLP-92] Know When to Stop Where to Restart: Accelerating Multi-Turn Agent ic On-Policy Distillation
【速读】： 该论文旨在解决在多轮代理（multi-turn agentic）场景下，基于策略的蒸馏（On-policy Distillation, OPD）方法因自回归学生模型滚动（autoregressive student rollouts）带来的高昂计算成本问题。现有加速方法依赖于固定的、离线设定的预算对监督信号进行截断或重定位，但未能考虑教师信号可靠性在轨迹内部及跨轨迹间的显著差异。通过在τ²-bench上的实证分析发现，有效监督信号高度集中于每轮的前缀部分，且教师认可度的跨轮损失在时间上与学生首次错误动作严格对齐，而非随轮次逐步累积。基于此，论文提出STRIDE（Stop-and-Restart on-policy Distillation acceleration）方法，其核心在于两个互补机制：自适应早期停止（adaptive early stopping），当累计教师对数概率低于分布外阈值时终止当前滚动；以及前缀缓存（prefix buffer），用于存储高质量前缀，并在学生表现最弱的正确轮次处重启生成。二者共同构建了数据驱动的渐进式课程学习机制，逐步扩展对后续轮次的覆盖范围。实验结果表明，在τ²-bench零售任务上，STRIDE以3.73倍速度提升达到与全轨迹OPD相当的性能，超越30B教师模型；相较于基线提速2.34倍，且在跨领域多教师训练下仍保持4.51倍加速。作为额外验证，STRIDE在非代理设置下的AIME 2025和AIME 2024任务中分别实现5.10倍和3.08倍加速，优于全轨迹OPD，而两种固定预算截断基线均未超越全轨迹方法。
链接: https://arxiv.org/abs/2609.14636
作者: Zhiyu Gui,Kexin Huang,Jia Guo,Junkang Wu,Zihao Wang,Zhiqiang Zhang,Jun Zhou,Jiancan Wu,Xiang Wang
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 14 pages, 8 figures
Abstract:On-policy distillation (OPD) has become a standard approach for transferring capabilities from large teachers to compact students. Its cost, however, is dominated by autoregressive student rollouts and scales poorly in multi-turn agentic settings. Existing acceleration methods truncate or relocate the supervision signal according to fixed, offline budgets, despite substantial variation in teacher-signal reliability both within and across trajectories. Our empirical analysis on \tau^2 -bench reveals a clear structure in this variation: informative supervision is concentrated in the prefix of each turn, and, most importantly for multi-turn agentic training, the cross-turn loss of teacher endorsement is temporally locked to the student’s first erroneous action rather than accumulating gradually over turns. Building on these findings, we propose STRIDE (Stop-and-Restart on-policy Distillation acceleration), which combines two complementary techniques: adaptive early stopping, which terminates a rollout once the cumulative teacher log-probability falls below an out-of-distribution threshold, and a prefix buffer, which caches high-quality prefixes and restarts generation at the weakest correct turn. Together, these mechanisms induce a data-driven curriculum that progressively extends coverage to later turns. On \tau^2 -bench retail, our method matches full-trajectory OPD and exceeds the 30B teacher at a 3.73\times speedup, surpasses the baseline itself at 2.34\times , and retains a 4.51\times speedup under cross-domain multi-teacher training. As a supplementary generalization test beyond the agentic setting, STRIDE outperforms full OPD on AIME 2025 at a 5.10\times speedup and on AIME 2024 at a 3.08\times speedup; averaged across the two evaluations, both fixed-budget truncation baselines remain below full OPD.
[NLP-93] Domain-specific Pretraining Profile and Transformer Performance: Evidence from Modeling Digital Prag matics in Arabic-English Code-switching
【速读】： 该论文旨在解决阿拉伯语-英语代码转换话语中数字语用功能建模的挑战，特别是针对具有高度语境敏感性的语用功能分类任务。其核心问题在于：现有预训练模型在跨语言、跨文化语用场景下的表现受限，尤其是在处理阿拉伯语与英语混合使用的社交媒体文本时，如何有效捕捉语言间的语用差异。解决方案的关键在于引入领域特定的预训练模型（Domain-Specific Pretraining Profile, DSPP），并通过对比MARBERT（专为阿拉伯语优化的BERT变体）与XLM-R(oBERTa)（强调多语言覆盖）的表现，验证了DSPP对模型性能的决定性影响。实验结果表明，尽管XLM-R具备更广泛的多语言能力，但在阿拉伯语数字语用任务上表现显著逊色；而MARBERT凭借其在阿拉伯语语料上的深度领域适应性，在验证集和独立测试集上均大幅优于XLM-R，尤其在宏观F1分数上从0.52提升至0.85，且在多数类别上实现高达0.60的F1提升。这表明，对于高度专业化语用分类任务而言，模型在目标语言领域的深度预训练比广义多语言覆盖更为关键。
链接: https://arxiv.org/abs/2609.14571
作者: Fahad Al Hussen,King Saud University,Riyadh,Saudi Arabia,Mohammed Q. Shormani,Ibb University, Ibb,Yemen
机构: 未知
类目: Computation and Language (cs.CL)
备注: 19 pages, 2 figures, 5 tables
Abstract:This study highlights the role of domain-specific pretraining profile (DSPP) in Transformer performance for modeling digital pragmatics in Arabic-English code-switched discourse. It evaluates MARBERT and XLM-R(oBERTa), with BERT serving as a general-purpose baseline. The models were evaluated on their ability to classify context-sensitive pragmatic functions in code-switched social-media discourse. 11695 unique X posts were collected via Python and utilized for the study. The study employs a quantitative and qualitative NLP approach, following a supervised pipeline. Findings unveil that MARBERT consistently surpasses XLM-R with validation Macro F1 increasing from 0.39 to 0.84 and validation loss decreasing from 0.55 to 0.19. On an independent test set, it achieved 0.96 accuracy, 0.83 macro precision, 0.87 macro recall, and 0.85 Macro F1, while XLM-R achieved 0.92 test accuracy but a substantially lower Macro F1 of 0.52. This was also supported by class-level performance where MARBERT outperforms XLM-R considerably with F1 improvements ranging from +0.33 to +0.60, demonstrating a clear advantage in modeling Arabic digital pragmatics. The study concludes that Transformer performance depends more on DSPP than multilingual coverage alone, as the latter does not guarantee optimal performance on a highly specialized pragmatic classification task.
[NLP-94] Disentangling Topology and Diversity in Multi-Agent LLM s for Multilingual Low-Resource Emotion Detection EMNLP2026
【速读】： 该论文旨在解决多智能体大语言模型（Multi-agent LLM）系统中推理调用的连接方式（拓扑结构）与智能体间多样性来源之间的混淆问题。以往研究常将二者混同，导致对性能提升机制的理解偏差。本文通过控制实验设计，独立考察了两种核心因素：推理拓扑结构（并行聚合 vs. 顺序精炼）与智能体间多样性来源（随机采样、角色提示、基于QLoRA的可学习专业化）。在固定三步调用预算和输出协议的前提下，采用Qwen2.5-14B-Instruct与Llama-3.1-8B-Instruct作为骨干模型，在九种低资源语种的情感检测任务上评估了六种配置组合。结果表明，基于可学习专业化的并行架构在Qwen上表现最优，达到52.83的宏平均F1，于Llama上亦达52.94，显著优于同骨架下的零样本、少样本、思维链（CoT）及七步自一致性基线。进一步分析显示，最优拓扑结构依赖于多样性来源：随机采样与角色提示场景下顺序精炼更优；而可学习专业化带来的优势在不同模型间差异显著，其性能增益从Qwen上的2.83点降至Llama的0.17点。深度分析揭示，后期学习到的专业化智能体可能覆盖早期正确预测，但整体效果仍取决于模型骨干。总体而言，智能体间的差异化策略所引发的性能变化远大于拓扑结构的影响，因此拓扑选择应与专业化方法协同评估。
链接: https://arxiv.org/abs/2609.14570
作者: Ulugbek Shernazarov,Charitha Ruwansiri Weerakon Basnayake,Abdelkhaleq El Jarjini,Noel Crespi,Praboda Rajapaksha
机构: Télécom SudParis, Institut Polytechnique de Paris, Évry-Courcouronnes, France; Cardiff Metropolitan University, Cardiff, United Kingdom; Aberystwyth University, Aberystwyth, United Kingdom
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 23 pages, 5 figures, 25 tables. Accepted at the REALM Workshop at EMNLP 2026. Code: this https URL
Abstract:Multi-agent LLM systems combine multiple inference calls, but prior work often confounds how calls are connected with how they are diversified. We study these factors independently: inference topology and source of inter-agent diversity. In a controlled 2 \times 3 matrix, we cross parallel aggregation and sequential refinement with stochastic sampling, role prompting, and learned QLoRA specialization, under a fixed three-call budget and output protocol within each backbone. Using Qwen2.5-14B-Instruct and Llama-3.1-8B-Instruct, we evaluate all six configurations on multilingual low-resource emotion detection across nine languages. Parallel learned specialization is strongest on Qwen at 52.83 Macro-F1 and reaches 52.94 on Llama. On Qwen it also exceeds same-backbone zero-shot, few-shot, CoT, and seven-call self-consistency baselines. The preferred topology depends on diversity source: sequential refinement helps stochastic and prompted settings, while the learned Width advantage shrinks from 2.83 points on Qwen to 0.17 on Llama. Depth-wise analysis suggests that later learned specialists can overwrite correct early predictions, although the aggregate effect is backbone-dependent. Overall, how agents are differentiated produces larger performance shifts than topology, which should be evaluated jointly with specialization.
[NLP-95] ATK: Triple-Aware Top-K Learning with Knowledge-Grounded Verification for LLM -based Sequential Recommendation EMNLP2026
【速读】： 该论文旨在解决基于大语言模型（LLM）的序列推荐系统在将下一物品预测建模为文本生成任务时，与全量候选集Top-K排序目标之间存在的接口不匹配问题。现有方法虽能生成自然语言形式的推荐结果，但其训练目标与实际应用场景中的排序性能（如NDCG@K）缺乏直接对齐，导致推荐质量下降。为此，论文提出TATK（Triple-Aware Framework），其核心解决方案在于双阶段协同机制：第一阶段为“Top-K学习”（TKL），通过上下文感知的元数据知识图谱（KG）提示增强与位置感知的Top-K奖励设计，使模型训练过程与最终排序效用对齐；第二阶段为“知识引导验证”（KGV），在单次LLM前向传播后，基于同一元数据构建的项图进行结构感知的重排序，仅对前M个候选进行精细化筛选。实验在Amazon Reviews 2023数据集上的Musical Instruments、CDs and Vinyl、Video Games三个领域进行，采用与R2ec一致的全量候选集评估协议，使用Gemma-2-2B-It和Qwen2.5-3B-Instruct作为骨干模型，全面对比了序列型、生成式、知识图谱增强及推理增强等基线方法。结果显示，TATK在全部36项指标上均优于基线，尤其在NDCG@10上对Qwen的提升达8.05%、4.26%、3.78%，对Gemma的提升更是高达27.03%、10.52%、10.23%，且推理延迟仅为基线RecPO的1.17倍以内。诊断分析进一步表明，结构化证据在具备可靠知识图谱支持的可恢复性高置信度候选集中最具价值，当元数据关系稀疏或噪声较大时，应引入门控机制以避免误导。
链接: https://arxiv.org/abs/2609.14565
作者: Yuchen Guan,Jiaye Liu,Yifei Han,Zhenxi Zhang,Yixuan Weng,Bin Li
机构: East China University of Science and Technology (华东理工大学); Hong Kong Institute of Science Innovation (香港科技创新研究院); Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences (中国科学院深圳先进技术研究院)
类目: Computation and Language (cs.CL)
备注: Accepted at EMNLP 2026 Main Conference. 23 pages, including references and appendix
Abstract:LLM-based sequential recommenders usually cast next-item prediction as text generation, but this interface is poorly matched to full-catalog top-K ranking. We propose TATK, a Triple-Aware framework that couples Top-K Learning (TKL) with Knowledge-Grounded Verification (KGV) for LLM-based sequential recommendation. Top-K Learning combines context-aware metadata-KG prompt grounding with position-aware top-K rewards, aligning training with ranking utility; Knowledge-Grounded Verification then applies structure-aware reranking over the top-M candidates after a single LLM forward pass, using the same metadata-derived item graph. We evaluate TATK on Musical Instruments, CDs and Vinyl, and Video Games from Amazon Reviews 2023 under a matched R2ec-style full-catalog protocol. Experiments use Gemma-2-2B-It and Qwen2.5-3B-Instruct backbones, compare against sequential, generative, KG-augmented, and reasoning-enhanced baselines, and include component, reward-shape, sequence-perturbation, reranking, relation-quality, and candidate-pool diagnostics. TATK improves over the matched R2ec reproduction on all 36 reported metrics. On NDCG@10, it improves Qwen by 8.05%, 4.26%, and 3.78% on the three datasets, and improves Gemma by 27.03%, 10.52%, and 10.23%, while keeping inference within 1.17x of Base RecPO latency. The diagnostics show that structural evidence is most useful for recoverable top-M candidates with reliable KG support, and should be gated when metadata relations are sparse or noisy.
[NLP-96] Neyshekar: An Open Persian Read-Speech Corpus for Automatic Speech Recognition
【速读】： 该论文旨在解决现有波斯语语音语料库在覆盖正式与非正式语言、命名实体及长句表达方面的不足，尤其针对生成式模型辅助构建语料时的文本质量与多样性问题。其解决方案的关键在于构建一个开放的波斯语读-说语料库Neyshekar，通过整合人工撰写文本、上下文同形异义词以及经过审校的语言模型生成文本，扩充了34,541个独特的录音提示，并采用shekar库对文本进行形式与非形式化统一归一化处理。所有录音均基于统一验证标准进行审核，且提供逐项标注者标签以支持可复现的共识度估计，同时通过不可见的贡献者标识实现说话人无关划分的可审计性，结合文本无关测试子集确保评估超越已有提示的泛化能力。此外，语料库还提供了每条录音的贡献者负载与无参考信号质量特征，使其在与波斯语Common Voice在相同处理流程下具有可比性。实验表明，在约32小时的训练时长下，相较于Common Voice，Neyshekar使Whisper和XLS-R在域内词错误率（WER）分别降低9.5点和11.6点，在独立的PSRB样本上两者均降低约8点，验证了其在自动语音识别（ASR）任务中的显著优势。该语料库已以CC0协议开源，代码与数据可通过项目仓库获取。
链接: https://arxiv.org/abs/2609.14542
作者: Ahmad Amirivojdan,Farzad Nadiri,Abolfazl Alizadeh,Shaghayegh Yaraghi
机构: Shekar AI
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注: 24 pages, 6 figures; dataset and code available at this https URL
Abstract:Neyshekar is presented as an open Persian read-speech corpus designed for coverage of both formal and informal language, named entities, and longer utterances. In version 6, 62,279 validated recordings totalling 99.02 hours are provided from 190 contributors, with 34,541 distinct recorded prompts. The prompt pool was assembled from human-written material, contextualised homographs, and reviewed language-model-generated text. Text entries were normalised with the shekar library, which supports both formal and informal Persian, and every submitted recording was reviewed against a common validation rubric. About 24% of released clips are classified as informal by an automatic classifier; these register labels are not human-validated. Item-level rater labels are provided for reproducible agreement estimation, opaque per-clip contributor identifiers make the speaker-disjoint partitioning auditable and support contributor-clustered uncertainty estimates, and a text-disjoint test subset is included for evaluation beyond previously seen prompts. Per-contributor recording load and reference-free signal quality are characterised for every released clip. Corpus characteristics are compared with Persian Common Voice under shared processing. Utility is assessed through two ASR architectures, three optimisation seeds, WER and CER, and independent evaluation on the public PSRB sample. Against duration-matched Common Voice training at approximately 32 hours, in-domain WER is reduced by 9.5 points for Whisper and 11.6 points for XLS-R, and by approximately eight points for both architectures on the independent PSRB sample. Transfer and mixture benefits are not consistently observed across architectures and training budgets. The corpus is released under CC0; code and data are made available through the project repository at this https URL.
[NLP-97] heseus in the Graph: Towards Traceable Multi-Hop Graph Navigation
【速读】： 该论文旨在解决现有多跳知识图谱问答（Multi-Hop Knowledge Graph Question Answering, KGQA）系统在推理过程中的不可追溯性问题，即模型通常仅关注最终答案的预测，而未显式建模或验证中间推理步骤，导致难以判断正确答案是否源于可信的多跳推理路径。为此，论文提出将多跳KGQA重新构架为一种问题条件化的图遍历（question-conditioned graph navigation） 任务，并引入名为THESEUS（Traceable Hop-wise Evidence SEarch in a Unified Semantics）的统一框架。其解决方案的关键在于：通过引入带有标注证据路径和改写问题的增强数据集，构建可导航的KGQA基准；设计涵盖路径保真度、语言变体鲁棒性及多跳/多答案问题性能的评估协议；并将基于路径的知识图谱补全代理（如MINERVA、MultiHopKG、SQUIRE）从符号化单关系查询范式升级为基于完整问题嵌入的语义驱动遍历机制，从而实现推理路径的显式生成与可验证性。这一方法推动了KGQA研究向以可追溯性为核心的目标迈进，使答案可伴随可验证的推理路径进行系统评估。
链接: https://arxiv.org/abs/2609.14528
作者: Eduin E. Hernandez,Luis F. Garcia,Nurassyl Askar,Sergio A. Diaz,Stefano Rini
机构: Model call failure
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 30 pages, 1 figure,
Abstract:Multi-Hop Knowledge Graph Question Answering (KGQA) tasks require models to assemble relational evidence along paths in a KG to answer natural-language questions. However, existing KGQA systems typically focus on predicting the final answer without explicitly modeling or validating the intermediate reasoning steps, obscuring whether the correct answers arise from faithful multi-hop reasoning. To address this limitation, we re-frame multi-hop KGQA as a question-conditioned graph navigation problem. We refer to this formulation as THESEUS - Traceable Hop-wise Evidence SEarch in a Unified Semantics. In this setting, an agent receives a KG, a question, and a topic entity, and traverses a sequence of relations towards the answer, making the reasoning path explicit. To systematically study this formulation, we provide three key contributions. (i) We augment the existing KINSHIP and MQuAKE resources into navigation-ready KGQA datasets with annotated evidence paths and paraphrased questions. (ii) We design evaluation protocols to measure path fidelity, robustness to linguistic variation, and performance across multi-hop and multi-answer questions. (iii) We adapt established path-based KG completion agents - MINERVA, MultiHopKG, and SQUIRE - to operate on full question embeddings rather than symbolic single-relation queries, enabling their trajectories to be guided by natural-language semantics. Together, these contributions advance KGQA research toward systems where traceability is fundamental: answers are accompanied by explicit reasoning paths whose agreement with reference evidence can be systematically evaluated.
[NLP-98] NeuroActiSep: Detecting Factual Hallucinations from Feed-Forward Neurons in a Single Pass
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中存在的幻觉（Hallucination）问题，该问题严重降低了模型的可靠性并阻碍其实际应用。现有研究多采用白盒方法，通过分析模型内部表示来识别真实性与事实性相关的模式，而较少关注前馈神经元（feed-forward neurons）在幻觉生成中的作用。本文提出一种新方法：基于自定义的神经元选择数据集，对最终提示词（prompt token）处的前馈神经元进行排序与筛选，并将所选神经元的身份迁移至其他事实问答数据集上训练幻觉检测分类器。该方法的关键在于利用特定神经元作为可迁移的特征信号，实现高效且泛化能力强的幻觉检测。实证结果表明，基于所选神经元特征训练的探测器（probe）性能可与基于内部状态训练的探测器相当。此外，研究还分析了被选神经元的分布特性及其在不同网络层深度下的检测性能影响，为理解幻觉生成机制提供了新的视角。
链接: https://arxiv.org/abs/2609.14448
作者: Ali Derogar Odolou,Reza Nazari,Mostafa Salehi
机构: University of Tehran (德黑兰大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Hallucination in large language models reduces their reliability and slows adoption. Various white-box studies have used internal representations to detect patterns of truthfulness and factuality. A less-studied approach is to identify feed-forward neurons correlated with hallucination. We propose a method to rank feed-forward neurons at the final prompt token using a custom neuron selection dataset. We transfer the selected neuron identities to train hallucination classifiers on other factual question answering datasets. Our work provides empirical evidence that probes trained using the features from the selected neurons perform on par with probes trained on internal states. We also analyze the distribution of selected neurons and the effect of layer depth on detection performance.
[NLP-99] Policy Loopholes in Agent Evaluation: When Policy Ambiguity Masquerades as Agent Error EMNLP2026
【速读】： 该论文旨在解决当前智能体评估基准（agent benchmark）在评估策略合规性时存在的根本性问题：现有基准假设每个策略均对应唯一正确的动作，但自然语言策略可能因沉默、模糊或自相矛盾而产生多种合理解释，单一标准轨迹无法涵盖所有可辩护的执行路径。其解决方案的关键在于识别并分类此类策略漏洞（policy loophole），通过在两个τ²-基准领域中的审计发现，政策模糊性与工具权限宽松性共同决定了可被利用性的存在——当策略复杂度超出工具的约束能力时，智能体对语义空白的处理不一致，导致评估分数不可靠。研究进一步表明，政策规范质量直接决定了评估质量的上限，因此建议基准开发者在收集标准答案前对策略进行预先审计，以提升评估的可信度与一致性。
链接: https://arxiv.org/abs/2609.14400
作者: Hongliu Cao
机构: Amadeus SAS
类目: Computation and Language (cs.CL)
备注: Accepted at REALM EMNLP 2026
Abstract:Agent benchmarks evaluate policy compliance but assume each policy determines a unique correct action. Natural-language policies can violate this assumption through silence, ambiguity, or contradiction, admitting multiple defensible readings that a single gold trajectory cannot capture. Auditing two \tau^2 -bench domains, we develop a taxonomy of such policy loopholes and show that affected tasks produce unreliable scores: they lower scores across different models in different ways and make every model less consistent across repeated trials. A cross-domain comparison reveals that exploitability requires both policy ambiguity and tool permissiveness: when policy complexity exceeds what tools can enforce, agents resolve gaps inconsistently and scores become unreliable. Policy specification quality sets the ceiling on evaluation quality. Benchmark developers should audit policies before collecting gold annotations.
[NLP-100] MOSCOPT: Mixture-of-Skills Collective Optimization for LLM Agents
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）驱动智能体中提示词（prompt）与技能（skill）优化方法存在的局限性问题，即现有方法仅针对单一文本模板进行优化，忽略了多种互补策略之间的协同效应。其核心解决方案是提出一种无需参数、基于文本的MOSCOPT算法，该算法联合优化一组共N个技能以及一个门控技能（gating skill）G，后者在每一步动态选择K个最优技能进行激活。为实现高效技能优化，研究引入了具有内部维护双重状态的EditAdam优化器，并通过三阶段交错更新机制，在无需梯度或参数调优的情况下实现性能的单调提升。实验结果表明，MOSCOPT在5个基准任务和3种目标大模型上均显著优于所有基线方法，验证了混合技能架构结合选择性激活机制，以及三阶段协同演进策略对性能提升的关键作用。
链接: https://arxiv.org/abs/2609.14399
作者: Zhenyu Zhang1,Jiudong Yang
机构: Coupang(库邦); Futu AI(富途AI)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Natural language prompts and skills serve as the strategic backbone of LLM-based agents. Recent advances in prompt and skill optimization have achieved notable gains, yet all existing methods optimize a \emphsingle text template—missing the synergy among multiple complementary strategies. We propose MOSCOPT, a text-native, parameter-free algorithm that jointly optimizes a pool of N skills and a gating skill G that dynamically selects K skills per step. To effectively optimize the skills, we build the EditAdam with internally maintained dual states. Through the three-phase interleaved updates with EditAdam, the system monotonically improves without gradient or parameter tuning. Extensive experiments and detailed ablations across 5 benchmarks and 3 target LLMs demonstrate that MOSCOPT consistently outperforms all baselines, and confirm that both the mixture-of-skills architecture with selective activation and the collective evolution with three-phase interleaving are essential to its superior performance. Code is released this https URL.
[NLP-101] Formal Properties of Language as Constraints on Neural Dynamics
【速读】： 该论文旨在解决神经机制如何忠实实现语言结构这一核心问题，尤其关注语言的句法结构在神经层面的可实现性。现有研究多依赖于语言变量标注与神经活动的预测关联，但此类方法无法充分约束内在机制。为此，作者提出“神经可接受性计划”（Neural Admissibility Program, NAP），通过语言的代数特性识别出神经机制必须保持的不变量：非关联的层次化分组、交换律、递归封闭性、子结构访问能力以及结构化工作空间的动态转换。NAP的关键在于将句法结构构建过程形式化为代数操作，并为每一项不变量提出神经可实现的候选机制：内容可寻址的工作空间记忆、基于图结构的瞬态动力学调度机制（如稳定异宿通道）、以及相位耦合的封存操作以记录分组关系。研究发现，传统的马尔科利-伯威克熵优化绑定门仅在极窄的承诺带内保持分组一致性；为此，作者提出一种新型神经绑定操作——“熔合”（Meld）：两个成分种群通过共享突触收敛，进行亚线性整合并饱和。Meld在已知皮层操作基础上，首次实现了对所有NAP不变量的完全保留，且在不同深度与温度下均能恢复层次结构，同时预测有效种群维度可区分替代分组方案，复合结构依赖于成分间的分歧程度。尤为重要的是，表现最优的解码规则在理论上属于不可接受范畴，表明仅凭解码准确性不足以甄别机制真伪。NAP通过明确神经动力学必须忠于语言结构的代数要求，从根本上重构了认知神经实现的评估标准。
链接: https://arxiv.org/abs/2609.14384
作者: Elliot Murphy
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:What must a neural system be capable of to implement language? Current research annotates stimuli with linguistic variables and tests which electrodes, voxels, or language-model layers predict neural activity. Yet predictive success leaves mechanisms under-constrained. Here, we show that algebraic properties of language specify invariants that mechanisms must preserve: non-associative hierarchical grouping, commutativity, recursive closure, access to substructures, and structured workspace transitions. We term this the Neural Admissibility Program (NAP). Syntactic structure building is analyzed algebraically, with candidate mechanisms offered for each requirement: content-addressable workspace memory, graph-structured transient dynamics scheduling structure-building operations (e.g. stable heteroclinic channels), and a phase-coupled sealing operation recording grouping. Simulations show that a corrected Marcolli-Berwick entropy-optimized binding gate preserves grouping only within a narrow commitment band. As an alternative, we propose a novel neural binding operation we term ‘Meld’: two constituent populations converge through shared synapses, integrate sublinearly, and saturate. Meld is, to our knowledge, the closest neurally plausible composition law to syntactic Merge. It preserves every NAP invariant, uses known cortical operations, and recovers hierarchical structure at every tested depth and temperature. It predicts that effective population dimensionality separates alternative bracketings and that the composite depends on constituent disagreement. Importantly, the laws decoding bracketing most accurately are a priori inadmissible, showing that decoding accuracy alone cannot adjudicate between mechanisms. By specifying how neural dynamics can remain faithful to linguistic structure, the NAP changes the criterion by which neural implementations of cognition are evaluated.
[NLP-102] SpectralShift: Effective Context Window Extension of Gated DeltaNet via Spectral Reparameterization
【速读】： 该论文旨在解决线性注意力（Linear Attention）模型在长上下文建模中进行持续预训练时，因忽略其状态动态的谱特性而导致长程信息获取能力不足的问题。现有方法直接应用持续预训练而未对线性注意力层进行针对性优化，导致模型难以有效捕捉远距离依赖。针对此问题，论文从转移矩阵的谱视角分析了门控增量网络（Gated DeltaNet, GDN）的长期上下文扩展机制，识别出两个关键因素：(1) 一个与目标依赖长度对齐的足够宽的慢速谱带，以支持长程信息传播；(2) 快速衰减模式的保留，以实现状态清除与上下文切换。基于上述发现，论文提出一种名为SpectralShift的谱重参数化方法，通过重新初始化α投影以重塑衰减谱，增强慢速传播能力，并引入α投影的学习率缩放策略，以促进长上下文训练。实验表明，SpectralShift在持续训练过程中稳定提升模型的长上下文能力，为线性注意力模型的上下文窗口扩展提供了一种高效且有效的解决方案。
链接: https://arxiv.org/abs/2609.14320
作者: Zian Liu,Yiwen Hu,Zican Dong,Tian Xie,Wayne Xin Zhao,Yucheng Ding,Ran Tao,Bryan Dai
机构: Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学高瓴人工智能学院); IQuest Research; Microsoft Research Asia(微软亚洲研究院)
类目: Computation and Language (cs.CL)
备注:
Abstract:Recently, linear attention layers have been increasingly adopted to replace softmax attention at scale for long-context modeling. However, existing context extension approaches typically apply continued pretraining directly without modifying these layers, overlooking the spectral properties of linear attention state dynamics. In this work, we study long-context extension of Gated DeltaNet (GDN) from a spectral perspective of transition matrix and identify two essential factors governing long-range information retrieval: (1) a sufficiently broad slow spectral band aligned with the target dependency length, and (2) the preservation of fast-decaying modes for state clearing and context switching. Based on this observation, we propose SpectralShift, a spectral reparameterization approach for long-context continual pretraining of GDNs. Specifically, SpectralShift reparameterizes the alpha projections initialization to reshape the decay spectrum by enhancing slow propagation capacity, and further introduces a learning-rate scaling for alpha projections to facilitate long-context training. Experiments show that SpectralShift consistently improves long-context capabilities over training, providing an effective and efficient solution for extending context windows of linear attention models. The code has been open-sourced at this https URL.
[NLP-103] E2A-Bench: Benchmarking Evidence-to-Action Reliability in Financial Chart Reasoning EMNLP
【速读】： 该论文旨在解决金融视觉语言模型（VLMs）在基于图表证据生成投资决策建议时的可信性问题，尤其关注现有评估方法中“幻觉”检测过于依赖声明层面支持度而忽视证据到行动链路可追溯性的缺陷。其核心挑战在于：当前评估指标仅衡量生成语句是否与输入证据一致，却无法反映推理过程中的证据-信心-动作一致性、方向覆盖范围及整体可靠性。为此，作者提出了E2A-Bench基准，包含969个查询，基于323只沪深300成分股，在三种输入模态下构建了由确定性OHLCV数据驱动的证据锚点，系统评估模型在接地性（grounding）、推理-动作一致性（reasoning-action consistency）、证据-信心校准（evidence-confidence calibration）以及方向覆盖度（directional coverage）方面的表现，引入新型指标NDR（coverage-aware evidence-to-action reliability），以衡量证据到行动链条的可靠性，而非实际交易绩效。实验评估20个VLMs发现，仅依赖单一幻觉评分会掩盖三大关键缺陷：低接地率模型因仅有6.4%的方向覆盖度而表现极差；采用理想信息验证虽减少不实陈述但导致覆盖度崩溃；金融领域微调显著放大买入/卖出比例（达4.21至4.68倍）。研究结论表明，金融VLM的评估必须追踪完整的证据到行动链路，而非仅依赖单一幻觉得分。
链接: https://arxiv.org/abs/2609.14302
作者: Xiaoya Wang,Yutong Xu,Junjie Wang
机构: Tsinghua University (清华大学); Jinan University (暨南大学)
类目: Computation and Language (cs.CL)
备注: EMNLP Findings
Abstract:Can financial vision-language models (VLMs) turn chart evidence into reliable action recommendations? Existing hallucination evaluations are mostly claim-centric; they assess whether generated statements are supported, but not whether evidence remains traceable through rationale, confidence, and final action. We introduce E2A-Bench, a 969-query benchmark for financial chart reasoning, constructed from 323 HS300 constituents under three input modalities with deterministic OHLCV-derived evidence anchors. E2A-Bench evaluates grounding, reasoning-action consistency, evidence-confidence calibration, and directional coverage through UCR, RCI, ECI, and NDR, where NDR measures coverage-aware evidence-to-action reliability rather than realized trading performance. Evaluating 20 VLMs reveals three failures hidden by scalar hallucination scores: the lowest-UCR model ranks near the bottom by NDR due to only 6.4% directional coverage; oracle-aided verification reduces unsupported claims but can collapse coverage; and financial fine-tuning amplifies the BUY:SELL ratio by factors of 4.21 to 4.68 across strict base-fine-tuned pairs. These results show that financial VLM evaluation should trace the full evidence-to-action chain rather than rely on a single hallucination score. Code and data: this https URL
[NLP-104] Editorial routing shapes how computational results are qualified in AI-assisted scientific writing
【速读】： 该论文旨在解决生成式AI在科研写作中因计算证据的分配位置不同而导致结果报告不一致的问题，即如何确保数值性限定词（如精度、置信区间等）在最终稿件中被正确保留。其解决方案的关键在于引入“定向放置规则”——将详细比较内容明确分配至支持信息（Supporting Information）或AI自身的工作笔记中，而非集中存入通用代码库或项目仓库，从而显著提升句子层面的数值限定词保留率；相比之下，泛化的准确性提醒未能产生类似效果。研究发现，尽管上下文文档化有助于内部处理，但若未通过结构化机制引导输出，生成结果仍可能忽略关键量化信息，表明模型行为受任务上下文与信息组织方式的复杂影响，而非简单的可访问性排序。
链接: https://arxiv.org/abs/2609.14288
作者: Jihan Kim
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models increasingly analyze computational results and draft manuscripts, making reliable communication as important as correct analysis. Using fixed computational evidence, we tested whether assigning comparisons across modeling choices elsewhere in a research workflow changes manuscript reporting. In constrained sentence-writing tasks, Anthropic’s Claude Sonnet 5 often omitted numerical qualifications when detailed comparisons were assigned to a group repository, but retained them more often when the same comparison was assigned to Supporting Information or its own working notes; Claude Opus 5 was less sensitive. These effects did not follow a simple accessibility ordering. A targeted placement rule largely restored sentence-level qualification, whereas a generic accuracy reminder did not. Longer contributions retained numerical qualifications, although some summaries across computational settings were still redirected to the repository. Thus, documenting context within an AI workflow does not ensure its communication where readers encounter the result.
[NLP-105] DenMark: Robust Semantic Watermarking for Diffusion Language Models
【速读】： 该论文旨在解决生成式 AI（Generative AI）中语义水印（semantic watermarking）在扩散语言模型（Diffusion Language Models, DLMs）上的适用性问题。现有语义水印方法主要针对自回归语言模型（Autoregressive Language Models, ARLMs），依赖于对已完成候选单元的生成与评分，但该范式无法自然扩展至DLM，因为DLM在去噪过程中语义单元始终不完整，且词元更新顺序灵活。为此，论文提出DenMark框架，其核心在于将关键依赖信号直接注入DLM的去噪过程：通过将输出划分为固定词元区域，并利用临时演进（temporary rollouts）作为语义前瞻，基于条件补全估算不完整区域的最终语义，从而选择具有更高估计语义水印得分的局部更新；在多个去噪步骤中重复此过程，逐步在最终输出中累积水印证据。检测阶段，DenMark采用校准扫描策略，对候选单元尺寸进行多尺度检测，以抵御由语义攻击引入的边界偏移。实验表明，在四种DLM、三个数据集及四种语义攻击场景下，DenMark在全部48种组合中均优于现有方法，验证了其在DLM中实现鲁棒语义水印的有效性。
链接: https://arxiv.org/abs/2609.14257
作者: Tianhao Ma,Weihao Xuan,Dong-Dong Wu,Farshid Nooshi,Takashi Ishida,Gang Niu,Naoto Yokoya,Masashi Sugiyama
机构: The University of Tokyo (东京大学); RIKEN Center for Advanced Intelligence Project (理化学研究所先进智能项目中心)
类目: Computation and Language (cs.CL)
备注:
Abstract:Semantic text watermarks encode signals in meaning rather than surface token choices, offering robustness to paraphrasing and other semantic-preserving edits. Existing semantic watermarking methods are primarily designed for autoregressive language models (ARLMs), where completed candidate units can be generated and scored before generation proceeds. This paradigm does not naturally extend to diffusion language models (DLMs), where semantic units remain incomplete during intermediate denoising steps and tokens may be updated in flexible orders. We propose DenMark, a semantic watermarking framework that injects key-dependent signals directly into the DLM denoising process. DenMark partitions the output into fixed token regions and uses temporary rollouts as semantic lookahead: conditional completions estimate the eventual semantics of an incomplete region, enabling DenMark to select local updates with higher estimated semantic watermark scores. Repeating this procedure across denoising steps progressively accumulates watermark evidence in the final output. For detection, DenMark uses calibrated scanning over candidate unit sizes to remain robust to boundary shifts introduced by semantic attacks. Across four DLMs, three datasets, and four semantic attacks, DenMark achieves the best results across all reported detection metrics in all 48 backbone-dataset-attack combinations. These results demonstrate that DenMark provides an effective mechanism for robust semantic watermarking in DLMs.
[NLP-106] Document Topic Alignment Metrics for Evaluating Topic Models of Short-Text Public Health Communications on Social Media
【速读】： 该论文旨在解决现有主题模型评估方法过度依赖仅关注生成主题本身、忽视文档（短文本帖子）与其分配主题之间语义一致性的问题。当前主流评估指标未能定量衡量所分配的主题是否真正有意义地代表原始文本内容，导致评估结果与实际语义关联脱节。为此，作者提出一种赋值感知的评估框架——文档-主题对齐度量（Document-Topic Alignment, DoTA），其核心在于通过量化文档与其所属主题之间的语义对齐程度，实现对主题分配质量的精准评估。DoTA的关键创新点在于引入了基于边界（margin-based）和可区分性（discriminative）的变体，能够捕捉主题分配的置信度与不同主题间的区分能力。在来自X平台的三个公共卫生相关社交媒体数据集上，对五种主题模型进行实验验证表明，DoTA不仅提供了与传统基于主题的评估指标互补的评价线索，且其结果与人工评估高度一致，充分证明了赋值感知评估的重要性。研究结论强调，将DoTA纳入评估体系可显著提升对短文本主题建模性能评价的全面性与实际意义。
链接: https://arxiv.org/abs/2609.14256
作者: Wangjiaxuan Xin,Shuhua Yin,Yaorong Ge,Shi Chen
机构: The University of North Carolina at Charlotte(北卡罗来纳大学夏洛特分校); College of Computing and Informatics(计算与信息学院); Department of Public Health and Health Administration(公共卫生与卫生管理系)
类目: Computation and Language (cs.CL); Social and Information Networks (cs.SI)
备注: Accepted for publication in the Proceedings of the 60th Hawaii International Conference on System Sciences (HICSS 2027)
Abstract:Topic models are widely used to analyze public health-related social media short texts, yet their evaluation remains dominated by metrics that focus entirely on generated topics alone. There is a lack of metrics that quantitatively assess whether assigned topics meaningfully represent the corresponding short-text posts. We propose Document-Topic Alignment metrics (DoTA), an assignment-aware evaluation framework comprising metrics that measure semantic alignment between documents (posts) and their assigned topics. We also introduce margin-based and discriminative variants that capture topic assignment confidence and distinguishability. We evaluate DoTA across five topic models on three public health-related social media datasets from X and compare DoTA metrics with conventional topic-based metrics. Results show that DoTA provides complementary evaluation cues and aligns meaningfully with human evaluations. These findings establish the need for assignment-aware evaluation and demonstrate that the addition of DoTA enables a more comprehensive and practically meaningful evaluation for assessing short-text topic modeling performance.
[NLP-107] he Attribution-Compression Frontier in Retrieval-Augmented Generation EMNLP
【速读】： 该论文旨在解决检索增强生成（Retrieval-Augmented Generation, RAG）中上下文压缩对引用归属（citation attribution）质量的影响问题，尤其关注在不同压缩方法与预算下，生成内容的引用准确性与事实依据之间的脱节现象。其核心挑战在于：传统以答案质量为导向的评估无法充分反映生成结果是否真实、准确地源自原始来源，即存在“生成内容与原始证据之间支持关系断裂”的风险。解决方案的关键在于引入一种新的可追溯性评估协议（re-attributability protocol），通过共享自然语言推理（Natural Language Inference, NLI）模型实现源片段恢复与引用评分，并系统比较多种压缩策略——包括重排序（reranking）、抽取式选择（extractive selection）、抽象式摘要（abstractive summarization）、token剪枝（token pruning）以及提取-聚类-重写（extract-cluster-rewrite）——在ASQA和QASPER数据集上的表现。研究发现，尽管某些方法如RECOMP在摘要层面表现出较高的引用精度（0.86），但在源片段层面仅达0.12，且在进行主张验证时，基于源片段的不支持率高达0.88，远高于基于摘要的0.17，表明压缩过程显著削弱了可追溯性。该结果依赖于统一的NLI模型，缺乏独立的人工校准，但已揭示出当前压缩机制在保障引用可信度方面的根本缺陷，提示未来需发展更可靠、可验证的引用归因评估框架。
链接: https://arxiv.org/abs/2609.14245
作者: Deepanshu Mody
机构: Center for Data Science (数据科学中心); New York University (纽约大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to GroundLM 2026 Workshop at EMNLP
Abstract:Context compression reduces generator input in retrieval-augmented generation, but answer quality alone does not characterize citation attribution. We measure citation attribution across compression methods and budgets, comparing reranking, extractive selection, abstractive summarization, token pruning, and an extract-cluster-rewrite construction on ASQA and QASPER under a fixed generator and primary entailment evaluator. On ASQA at a nominal 0.25 budget (achieved compression 0.08), a RECOMP-style compressor’s citations score 0.86 precision against its summaries but 0.12 against source spans under our re-attributability protocol. These estimates depend on a shared NLI model for span recovery and citation scoring and lack independent human calibration. Extractive selection’s observed grounded precision ranges from 0.43 to 0.49 across nominal budgets from one-half to one-tenth of the ASQA context, while answer quality declines. For the same RECOMP setting, claim verification after source recovery yields an unsupported rate of 0.88 versus 0.17 when checking summaries. This gap persists beyond structural rejection of missing provenance, but remains evaluator-dependent. A 200-question TRUE T5-XXL audit also finds emitted–grounded gaps under both fixed and recomputed source mappings, without establishing human-calibrated support rates.
[NLP-108] Corpus Characterization and Inverse Constitutional Fine-Tuning for Style-Aware Radiology Reports
【速读】： 该论文旨在解决生成式医学影像报告在风格上与真实放射科医生写作习惯存在显著偏差的问题，具体表现为报告在结构、用词及不确定性表达等方面不符合临床实际书写规范，从而影响临床医生的信任度与使用体验。其解决方案的关键在于：首先，通过生物医学临床BERT（Bio-ClinicalBERT）嵌入、UMAP降维与HDBSCAN聚类分析2,000份来自CheXpert Plus数据集的报告，系统识别出五种具有显著差异的报告风格模式；进而基于这些风格特征，采用逆向宪法人工智能（inverse constitutional AI）框架，从无偏好标注的放射科医生撰写报告对中自动推导出一个聚焦于写作风格的“宪法”（constitution），该宪法编码了语气、用词、不确定性校准及报告结构等关键风格规范；最后，将此风格宪法融入对MedGemma-4B基础模型的监督微调过程，在25,245个训练样本上实现风格对齐优化。实验表明，相较于未微调基线模型（BLEU-4: 0.006 → 0.308；ROUGE-L: 0.171 → 0.484），宪法微调显著提升了文本结构与词汇层面的匹配度，且呈现出质的改进而非量的提升，证明了在语料级风格表征与宪法建模基础上实现高效、可扩展的风格一致性生成的有效性。
链接: https://arxiv.org/abs/2609.14226
作者: Sarah Y. Li,Elijah Renner,Rayan Ansari,Alaa Youssef
机构: Stanford University; Stanford Cardiovascular Institute; Stanford School of Medicine
类目: Computation and Language (cs.CL)
备注:
Abstract:Automated radiology report generation has advanced rapidly in diagnostic accuracy, yet generated reports frequently diverge from the stylistic conventions of authentic radiologist writing in structure, diction, and uncertainty language, a gap which has direct implications for clinician trust and user experience. To address this, we characterize stylistic variation across 2,000 reports from the CheXpert Plus dataset using Bio-ClinicalBERT embeddings, UMAP dimensionality reduction, and HDBSCAN clustering, identifying five distinct reporting patterns differing in pathology focus, narrative structure, and lexical preference. Drawing on these findings, we adapt the inverse constitutional AI framework to derive a style-focused constitution from radiologist-written report pairs without requiring a formal preference dataset. This constitution, encoding conventions of tone, diction, uncertainty calibration, and report structure, is incorporated into the supervised fine-tuning of a MedGemma-4B base model on 25,245 CheXpert Plus training pairs. Constitutional fine-tuning produces a substantial increases in text alignment (BLEU-4: 0.006 to 0.308; ROUGE-L: 0.171 to 0.484) relative to the untuned baseline. These gains show a qualitative shift in structural and lexical alignment rather than marginal improvement, as the baseline model produces near-zero scores due to format mismatch. Overall, we establish corpus-level style characterization and constitutional modeling as an effective and data-efficient strategy for producing radiology reports that conform to authentic radiologist writing conventions.
[NLP-109] Learning to Refer from Estimated Listener Gaze
【速读】： 该论文旨在解决生成式对话中指代表达（referring expression）的语用最优性问题，即如何使语言表达在保证指代准确性的同时尽可能简洁高效。传统方法通常依赖于显式的沟通成功反馈作为优化信号，但忽略了听者在理解过程中的隐含认知行为。本文的关键解决方案是利用视觉-语言模型（vision-language models）通过模拟听者逐步理解过程中的注视轨迹（gaze scanpaths）来构建隐式学习信号。具体而言，训练过程中，说话者策略基于图像和目标实体采样指代表达，随后由一个神经听者模型根据图像与表达预测人类注视轨迹序列；这些轨迹被转化为词级和序列级奖励信号，用于优化说话者策略参数。实验表明，采用基于注视预测的听者模型进行训练后，指代表达的平均长度从15.4词降至4.0词，指代成功率从75.2%提升至80.0%，显著提升了表达的语用效率。该研究揭示了通过语言交互中可观察的认知过程（如注视行为）实现更优语言生成的巨大潜力，不仅限于显式反馈，还可挖掘隐含的听者理解动态作为学习信号。
链接: https://arxiv.org/abs/2609.14207
作者: Téa Wright,Alane Suhr
机构: University of California, Berkeley (加州大学伯克利分校)
类目: Computation and Language (cs.CL)
备注: Accepted at COLM 2026
Abstract:We propose to finetune vision-language models to generate more pragmatically optimal referring expressions by transforming observations of incremental listener comprehension, in the form of gaze scanpaths, into learning signals. During training, referring expressions are sampled from the speaker policy being optimized, conditioned on images and target referents; then, a neural listener estimating human gaze behavior maps from images and sampled referring expressions to scanpaths, each represented by a sequence of fixations, with each fixation corresponding to a word in the referring expression. We experiment with several approaches to convert fixation sequences and target referents into token- and sequence-level rewards, which are used to optimize policy parameters. Through evaluation with human listeners, we find that speaker policies trained with gaze-estimating listeners result in significantly more pragmatically-optimal references than base models, reducing sequence length from 15.4 down to 4.0 words while increasing referential success from 75.2 up to 80.0%. Our work demonstrates a promising opportunity for learning to generate utterances through language-based interaction, not only from the explicit signal of communicative success, but also from implicitly-available observations of a listener’s process of comprehension.
[NLP-110] A Multi-Stage Agent ic Framework for Effective Counter-Narrative Generation and Refinement
【速读】： 该论文旨在解决社交媒体中仇恨言论与虚假信息泛滥所引发的民主社会挑战，尤其关注传统直接压制手段可能加剧社会极化、削弱公众信任并强化极端叙事的风险。针对生成式人工智能（Generative AI）驱动的反叙事（Counter-narratives, CNs）在实际应用中效果受限的问题，其核心在于缺乏对修辞与风格选择如何影响说服力的系统性理解。为此，论文提出一种多阶段代理驱动框架，涵盖反叙事的生成、优化与评估全过程，聚焦于俄乌战争背景下亲俄型仇恨言论与虚假信息的应对，并具备跨领域可扩展性。关键创新在于通过人类评估初步识别出高效的技术-风格组合（如重复配合情感化框架显著提升说服力），进而构建多智能体迭代优化机制，持续提升反叙事在说服力、情感共鸣与传播潜力方面的表现。经人工验证确认优化效果后，自动化安全分析表明，优化后的反叙事在质量上达到甚至超越专家撰写的对抗性言论水平；模拟实验进一步证实其能有效削弱亲俄叙事的可信度，且显著优于基础大模型生成的对照方案。该研究为实现可扩展、叙事定制化的反虚假信息干预提供了可行路径。
链接: https://arxiv.org/abs/2609.14178
作者: Carmel Kronfeld,Sharva Gogawale,Tetsuro Kobayashi,Irad Ben-Gal
机构: Tel Aviv University (特拉维夫大学); Waseda University (早稻田大学)
类目: Computation and Language (cs.CL); Computers and Society (cs.CY); Social and Information Networks (cs.SI)
备注:
Abstract:The rapid diffusion of hate speech and misinformation on social networks challenges democratic societies, since direct suppression efforts may deepen polarization, fuel public distrusts, and strengthen extremist narratives. LLM-driven counter-narratives (CNs) offer a promising way to reduce those risks, yet their effectiveness depends on rhetorical and stylistic choices that remain poorly understood. We present a multi-stage agent-based framework for generating, refining, and evaluating CNs, applied to pro-Russian hate and misinformation narratives on the war with Ukraine and adaptable to other domains. A pilot experiment with human evaluators identifies effective technique style pairings, such as repetition with emotional framing enhancing persuasiveness. Building on these insights, we introduce a multi-agent refinement process that iteratively improves CNs for persuasiveness, emotional engagement, and shareability. After human validation confirmed improvement, an automated safety analysis shows that our refined CNs match or improve on expert-written counterspeech. A simulated experiment then shows that they reduce the perceived strength of pro-Russian narratives and consistently outperform a vanilla LLM baseline, highlighting a pathway toward scalable, narrative-specific interventions against hate speech and misinformation. Code and data accompanying this work are publicly available at this https URL.
[NLP-111] Inherited Heads: Audio language models track speakers with their text backbones attention and an attention-mass ranking retrieves a different set
【速读】： 该论文旨在解决音频语言模型在多说话人场景下无法准确识别并描述特定说话人内容的问题，其表现远低于随机猜测的基准（16.7%）。解决方案的关键在于通过引入一个固定偏置（fixed bias）作用于模型中少数注意力头（attention heads）的注意力逻辑值（attention logits），仅需微调极少量参数（不足模型总量的十分之一）且无需重新训练，即可将模型输出精准引导至目标说话人，成功率高达90.7%至99.0%。进一步研究表明，这些有效注意力头并非专属于音频模态，而是可从纯文本语言模型中提取并直接迁移使用，即便不涉及任何音频信息参与选择过程，仍能实现80.8%至95.0%的引导成功率。实验发现，音频与文本模型共享的注意力头数量显著高于随机预期（66–74个对20个），且这些共享头已几乎完全复现了引导能力。然而，尚无法确定共享机制本身是有效性来源，因随机抽取同规模头部亦能取得相近效果。另一关键发现是注意力头的有效性评估方法：传统基于注意力集中程度的评分标准在部分模型中表现不佳，尤其在Ultravox中导致69.7%的输出无法归因于任一说话段，而采用每头归一化后注意力动态变化的评分方式则显著提升性能，表明注意力动态响应模式比静态集中度更能揭示有效引导头。
链接: https://arxiv.org/abs/2609.14174
作者: Bojro Das
机构: Cornell University(康奈尔大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Sound (cs.SD)
备注: 20 pages, 6 figures
Abstract:Asked to describe what one of six speakers in a recording talks about, audio language models describe the right one on 6 to 16% of trials, below the 16.7% a guess would give. Adding a fixed bias to the attention logits of a hundred heads, under a tenth of the model’s and with no training, redirects the description to whichever speaker we choose, on 90.7% to 99.0% of trials. Those heads are largely not specific to audio. Rank the text-only language model an audio model was built from, or a released model of the same family, on a written version of the task, take its top hundred heads, and carry them over unchanged: they redirect the audio model on 80.8% to 95.0% of trials, with nothing about audio entering the selection. The audio and text head sets share 66 to 74 of 100 where chance would give about 20, and the shared part alone reproduces almost all of the steering. What that does not show is that sharing is what makes the heads work: an equal-sized draw from the same discovered hundred does nearly as well, and none of our three models separates the two explanations. A second finding concerns how such heads are found. Ranking heads by how much attention they place on the segment asked about, as an established score does, or by how much of their attention moves with the question, as a per-head normalised variant does, gives top hundreds that share 69, 37 and 4 heads across our three models. In Ultravox, where they share 4, the established score’s heads leave output the judge cannot place on any segment on 69.7% of trials, against 40.0% with no intervention and 1.0% for the normalised variant. That is one arm of six; on the other five the established score steers above a random draw.
[NLP-112] owards Evolving Context Parameterization for Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在持续更新记忆上下文时，难以有效区分记忆有效性状态及保持全局一致性的问题。现有方法多假设上下文为静态，缺乏对动态更新场景中信息保留与更新之间平衡的显式机制。为此，作者提出了记忆连续演化（Memory Updating with Sequential Evolution, MUSE）任务，并构建了MUSE-bench基准测试以评估模型在记忆更新融合与无关信息保留方面的能力。其核心挑战在于：在维持全局状态稳定的同时，动态调整记忆证据的贡献权重。针对此问题，论文提出了一种无需训练的解决方案PLUME，其关键在于构建全局更新表征，激活记忆证据形成局部参数视图，并在解码阶段自适应地融合二者预测结果，从而实现高效且稳定的上下文演化。在MUSE-bench上的实验表明，PLUME在平均ROUGE-L召回率上提升29.9%，在LLM-as-a-Judge评估中提升54.9%，验证了其在序列演化场景下的优越性能。
链接: https://arxiv.org/abs/2609.14168
作者: Xiaobing Shi,Zherui Li,Yiming Jiang,Kun Wang,Yufei Guo
机构: Nanyang Technological University (南洋理工大学); Peking University (北京大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Context parameterization enables large language models (LLMs) to internalize contexts into reusable model parameters, avoiding repeated processing across subsequent queries. However, existing methods typically assume static contexts and lack explicit mechanisms for distinguishing validity states under continual updates. To study this real-world scenario, we formalized the Memory Updating with Sequential Evolution (MUSE) task and constructed MUSE-bench to evaluate update incorporation and unaffected-information preservation. The resulting challenge requires preserving the global state while adjusting the contribution of memory evidence. Motivated by this, we proposed PLUME, a training-free method that constructs a global update representation, activates memory evidence to form a local parameter view, and adaptively integrates their predictions during decoding. Comprehensive evaluation on MUSE-bench demonstrated PLUME’s effectiveness in sequential evolution settings, yielding relative improvements of 29.9% in average ROUGE-L Recall and 54.9% in LLM-as-a-Judge. Our codes are available at: this https URL.
[NLP-113] When Tools Get in the Way: The Effect of Unnecessary Tool Availability on LLM Answering
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在具备外部工具调用能力后，其对仅依赖自身知识即可回答的问题的处理能力是否会受到干扰的问题。具体而言，研究关注当一个与当前问题相关但不必要的工具被提供时，模型是否仍能准确地基于自身知识作答，以及先前的工具调用行为是否会影响这一能力。解决方案的关键在于揭示：即使未实际调用该冗余工具，其可用性本身也会显著降低模型在封闭域问题上的回答正确率（从98.2%降至63.5%），且这种下降无法仅由误调用工具解释，表明模型存在“工具依赖偏见”；而通过引入一句简短的、具有任务范围感知（scope-aware）的系统指令，可有效恢复大部分因工具可用性导致的性能损失，从而实现对模型推理行为的有效引导。
链接: https://arxiv.org/abs/2609.14157
作者: Saanvi Paturi,Arsen Kenzhebayev,Arham Sethi,Vyas Raina,Ivaxi Sheth,Vatsal Raina
机构: Spark AI Research(火花人工智能研究); UWCSEA East Campus(新加坡美国国际学校东校区); Haileybury Astana(海莉伯里阿斯塔纳校区); The Shishukunj International School(希舒克琼国际学校); Apta AI(阿普塔AI)
类目: Computation and Language (cs.CL)
备注: 10 pages, 2 figures, 4 tables
Abstract:Large language models (LLMs) are increasingly deployed with external tools that extend what they can do beyond their own knowledge. Tools help on tasks that need external information, but their availability may also change how a model handles questions that do not need them. Prior work has mostly asked whether models select and use tools appropriately; whether an unnecessary tool changes the correctness of answers has received less attention. We ask whether making a related but unnecessary tool available affects a model’s ability to answer from its own knowledge, and whether a preceding tool interaction changes this behaviour. We construct 500 query pairs across 10 knowledge domains. Each pair consists of a tool query, which needs the domain’s tool, and a closed-domain query, which does not. Six LLMs are evaluated with the tool unavailable, available, and available after a prior tool call. Across 3,000 baseline trials the pooled answer rate is 98.2%. When an unnecessary tool is available it falls to 63.5%, with large differences between models. The decrease occurs even when the tool is rarely called, so it cannot be explained by unnecessary tool invocation alone. A one-sentence scope-aware system instruction recovers most of the lost answers.
[NLP-114] One Size Does Not Fit All: Setting Inference Depth from the Questions a Deployment Actually Asks
【速读】： 该论文旨在解决生成式语言模型在特定部署场景下计算资源浪费的问题，即尽管不同应用场景（如客服问答、代码生成）仅涉及有限类型的输入提示（prompt），但每次推理仍需完整运行整个模型，导致大量计算开销无法避免。其核心解决方案是引入“早退出”（early exit）机制：在模型的中间层附加一个小型可训练组件（称为读出模块，readout），由该模块预测下一个输出词元，并通过置信度检验决定是否提前终止推理流程，从而跳过后续层的计算。关键创新在于仅利用模型自身在常规流量上的输出进行监督，保持主模型冻结，实现轻量级优化。研究发现，早期退出的节能潜力高度依赖于任务类型——在15亿参数模型半深度配置下，算术应用题可实现96%的词元提前输出，而中文解释类任务仅为8%，且在保持与全模型相当的词元级精度时表现出显著差异；此外，仅通过针对具体部署场景定制退出阈值可有效提升退出率（最高提升59个百分点），而其他方式无效；更关键的是，现有文献中普遍采用的“词元级保真度”评估指标在存在可验证真实答案的任务（如算术题）中失效，因早退出虽在词元层面表现良好，却导致正确答案数量从60下降至10–28，揭示了当前评估范式的局限性。研究设定目标为在个人设备等内存带宽受限的小型模型场景中实现高效生成。
链接: https://arxiv.org/abs/2609.14144
作者: Jerry Kaplan
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 20 pages, 5 tables in main text, 9 tables in appendices. Code, prompts, and result files at this https URL
Abstract:A transformer language model is trained to respond to any prompt, but each deployment asks only a narrow range of questions: a support assistant sees delivery complaints, a coding tool sees Python. Every deployment nonetheless pays the same computation per token. This paper measures how much of that cost is avoidable when the range of prompts is known in advance. The mechanism examined is early exit: a small, trained component - called a readout - is attached to an intermediate layer and proposes a token, and a confidence test decides whether to emit it or to run the remaining layers. The models are frozen, and the only supervision used is the model’s own output on ordinary traffic. Three findings are reported. First, achievable savings depend strongly on the kind of traffic: at half depth on a 1.5-billion-parameter model, 96 percent of tokens could be emitted early for arithmetic word problems and 8 percent for Chinese-language explanations, at matched token-level fidelity to the full model (a measure whose limits the third finding exposes). Second, of three ways a deployment might use knowledge of its traffic, only customizing the threshold for exiting early is worthwhile: calibrating it per deployment raised exit rates by up to 59 percentage points across three models, and by more than 10 points on most corpora tested. Third, token-level fidelity - the standard evaluation measure in the early-exit literature - fails in domains where tokens can be checked against ground truth: on arithmetic word problems, three models each answered sixty questions correctly when run in full, and between 10 and 28 correctly under early exit, in the configuration that scored highest on fidelity. The intended setting is small models on personal devices, where generation is limited by memory bandwidth rather than computation. Comments: 20 pages, 5 tables in main text, 9 tables in appendices. Code, prompts, and result files at this https URL Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG) ACMclasses: I.2.7; I.2.6 Cite as: arXiv:2609.14144 [cs.CL] (or arXiv:2609.14144v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.14144 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Jerry Kaplan [view email] [v1] Sat, 12 Sep 2026 20:49:48 UTC (549 KB)
[NLP-115] GraMRAG : Orchestrating Multi-Agent Multi-Step Reasoning via Graph Memory with Reinforcement Learning
【速读】： 该论文旨在解决现有多智能体检索增强生成（Multi-Agent Retrieval-Augmented Generation, RAG）系统在复杂多模态推理任务中面临的推理深度不足与记忆结构僵化问题，尤其体现在回答知识密集型问题时存在检索不充分和状态感知缺失（state blindness）的缺陷。其核心解决方案是提出一种基于图记忆引导的多智能体RAG框架——GraMRAG，其关键在于构建一个动态的多模态记忆图（multimodal memory graph），将智能体的推理过程形式化为一个动态有向无环图（DAG），显式建模动作-观测之间的依赖关系，从而缓解状态盲视并抑制冗余检索。此外，通过引入拓扑感知策略优化（Topology-Aware Policy Optimization, TAPO），利用图结构拓扑识别关键路径并进行目标性节点剪枝，实现多步推理轨迹中的细粒度信用分配。同时，提出的视觉-文本桥接推理范式整合了多尺度实体裁剪与类似ReAct的视觉工具链，显著增强了长程跨模态推理能力。实验表明，GraMRAG在多个具有挑战性的多模态基准测试中持续优于现有基线，实现了复杂长程推理任务的最先进性能。
链接: https://arxiv.org/abs/2609.14066
作者: Zhongyu Wang
机构: Beihang University (北京航空航天大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Although existing multi-agent Retrieval-Augmented Generation (RAG) systems have demonstrated promise on complex multimodal reasoning tasks, they remain fundamentally limited in reasoning depth and memory structure, suffering from inadequate retrieval and state blindness when answering knowledge-intensive questions. To address these limitations, we propose GraMRAG, a graph memory-guided multi-agent RAG framework that integrates a dynamic multimodal memory graph to enable stable, multi-step multimodal reasoning. We introduce a vision-text bridged reasoning paradigm that unifies multi-scale entity cropping with a ReAct-style visual toolchain, enhancing the long-horizon cross-modal reasoning capability. We further construct a multimodal memory graph that formalizes agent reasoning as a dynamic directed acyclic graph (DAG), explicitly modeling action-observation dependencies to mitigate state blindness and suppress redundant retrieval. Moreover, we propose Topology-Aware Policy Optimization (TAPO) that leverages graph topology for critical path identification and targeted node pruning, enabling fine-grained credit assignment across multi-step reasoning trajectories. Extensive experiments on challenging multimodal benchmarks demonstrate that our approach consistently outperforms existing baselines and achieves state-of-the-art performance on complex long-horizon reasoning tasks.
[NLP-116] Measuring the Creativity of Frontier LLM s in Automated Research
【速读】： 该论文旨在解决前沿大语言模型（LLM）在自动化研究任务中创造力缺乏系统评估的问题。现有研究虽已关注模型的自主科研能力，但对创造力的量化标准尚不完善，尤其缺乏对创新性维度的精细化衡量。为此，论文提出一套基于“价值度”（valueness）与“新颖性”（novelty）双维度的评估体系：其中价值度衡量所提想法的实际效用，新颖性则从三个层面展开——是否为已有知识的重复（精确匹配新颖性，Exact-Match P-Novelty）、是否探索了此前未涉及的变量或变量组合（变量级新颖性，Variable-level P-Novelty），以及是否直接延续或偏离外部检索知识（启发式新颖性，H-Novelty）。实验结果表明，各模型在多数指标上表现相近，但在变量级新颖性方面差异显著，反映出其在研究空间探索广度上的本质区别。进一步的相关性分析与细粒度想法评估揭示，变量级新颖性是与科研绩效关联最强的创造力维度，因此成为衡量模型创造性突破的核心指标。
链接: https://arxiv.org/abs/2609.14057
作者: Yiheng Zhao,Mengzhuo Chen,Chengming Hu,Pengyi Liao,Yiran Pang
机构: Concordia University (康科迪亚大学); McGill University (麦吉尔大学); Florida Atlantic University (佛罗里达大西洋大学); Independent Researcher (独立研究员)
类目: Computation and Language (cs.CL)
备注:
Abstract:Frontier LLMs are increasingly capable of conducting automated research, yet their creativity in this setting has not been systematically evaluated. In this paper, we propose a set of metrics to evaluate creativity along the two dimensions of valueness and novelty. Valueness assesses whether each proposed idea is useful, while novelty is evaluated from three perspectives: whether the same idea has appeared before (Exact-Match P-Novelty), whether a previously unexplored variable or variable combination is explored (Variable-level P-Novelty), and whether the idea directly follows retrieved external knowledge or departs from it (H-Novelty). Our evaluation shows that the models achieve relatively similar scores on most creativity metrics, but differ substantially in Variable-level P-Novelty, which reflects the breadth of research-space exploration. Further correlation and idea-level performance analyses show that Variable-level P-Novelty is the creativity dimension most strongly associated with research performance.
[NLP-117] VeriDx: Earning the Right to Diagnose with Disease-Centric Verification EMNLP2026
【速读】： 该论文旨在解决当前医学大语言模型（Medical LLMs）评估中存在的根本性缺陷：现有评估方法多聚焦于最终诊断结果、局部推理步骤或孤立事实，而忽视了临床推理过程中疾病假设所引发的系统性责任义务（hypothesis-induced commitments）。这些义务包括关键证据的验证、鉴别诊断的排除、矛盾点的化解、有用检查的考量以及诊断闭合的合理性论证。为应对这一问题，论文提出了一种以疾病为中心的验证框架VeriDx，其核心在于将自由文本形式的诊断推理与结构化的疾病特征档案（disease profiles）进行关联，动态追踪每个假设是否被满足、未解决或被违反。该框架能够有效暴露诊断过程中的系统性失败，如遗漏关键检查、未解决鉴别诊断、忽略矛盾信息、缺乏依据的断言以及过早闭合等。通过基于指南构建的复杂呼吸系统疾病案例和专家标注的纵向病历实例化，研究发现多数诊断错误并非孤立失误，而是早期推理中承诺断裂的累积后果。因此，解决方案的关键在于引入可追溯的、基于疾病的结构性验证机制，从而实现对临床推理完整性和严谨性的系统性评估。
链接: https://arxiv.org/abs/2609.14018
作者: Zhong Cao,Shuying Chen
机构: Heidelberg University(海德堡大学); University of International Business and Economics(对外经济贸易大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: accepted by EMNLP 2026, 17 pages, 3 figures, 3 tables
Abstract:A correct diagnosis can still be reached for the wrong reasons. In clinical reasoning, every disease hypothesis creates obligations: key evidence must be checked, alternatives must be ruled out, contradictions must be resolved, useful tests must be considered, and closure must be justified. Current evaluations of medical LLMs mostly focus on final answers, local steps, or isolated facts, and therefore miss these hypothesis-induced commitments. We introduce \textbfVeriDx, a disease-centric verification framework that links free-form diagnostic reasoning to structured disease profiles. VeriDx tracks whether each hypothesis is satisfied, unresolved, or violated its clinical obligations, exposing failures such as missing critical tests, unresolved differentials, ignored contradictions, unsupported claims, and premature closure. We instantiate VeriDx for complex respiratory diagnosis using guideline-derived disease profiles and expert-annotated longitudinal cases. Our results show that many diagnostic errors are not isolated mistakes, but broken commitments made earlier in the reasoning process.
[NLP-118] Unlocking the Unsolvable: Teacher-Guided Curriculum for Data-Efficient RLVR EMNLP2026
【速读】： 该论文旨在解决生成式 AI 在数学推理任务中面临的关键瓶颈：当问题超出当前模型能力范围时，所有尝试（rollouts）均失败，导致无有效学习信号，此类“不可解问题”虽具有高信息量却因缺乏可验证奖励而被结构化浪费。其核心解决方案是引入教师引导的课程学习机制——通过更强教师模型提供的部分推理轨迹构建渐进难度梯度，采用反向链式（backward-chaining）课程策略逐步撤除指导，促使学生模型最终实现无需外部帮助的自主求解。实验表明，仅使用128个原本不可解的问题进行训练，即可在九项基准测试上达到或超越基于2,000个问题全集训练的GRPO方法性能（数据效率提升约16倍），同时显著扩展了大k值下的pass@k指标所反映的推理边界。此外，研究识别出在纯不可解问题训练中存在显著的分布偏移代价，并提出单调前沿课程（Monotone Frontier Curriculum, MFC）方法，通过单调驱动训练过程向无指导求解演进，持续优于现有课程学习方法。
链接: https://arxiv.org/abs/2609.13997
作者: Yukang Zhu,Zhen Han
机构: Independent Researcher(独立研究员); Amazon(亚马逊)
类目: Computation and Language (cs.CL)
备注: 20 pages, 9 figures. Accepted to Findings of the Association for Computational Linguistics: EMNLP 2026
Abstract:Reinforcement Learning with Verifiable Rewards (RLVR) has shown remarkable success in improving the mathematical reasoning of large language models. Yet problems beyond the model’s current capability, where rollouts uniformly fail and no learning signal is produced, are structurally wasted despite marking the most informative training frontier. We show that these otherwise-inert problems can be unlocked via teacher-guided curriculum learning: partial reasoning traces from a stronger model create a graded difficulty landscape, and a backward-chaining curriculum progressively withdraws guidance until the model solves problems unaided. Training on only 128 unsolvable problems matches or exceeds GRPO trained on a full 2,000-problem corpus (~16x data efficiency) on the nine-benchmark average for both base models, while substantially expanding the reasoning boundary measured by pass@k at large k. Furthermore, we identify a distribution-shift cost that is particularly acute in the unsolvable-only regime and propose Monotone Frontier Curriculum (MFC), a method that monotonically drives training toward unguided solving, consistently outperforming existing curriculum methods.
[NLP-119] Mizan: A National Benchmark for Evaluating Large Language Models on Iraqi Arabic and the Iraqi Civic Context
【速读】： 该论文旨在解决当前大型语言模型（LLM）评估体系对阿拉伯语方言的忽视问题，特别是伊拉克阿拉伯语（Iraqi Arabic）这一实际使用语言在现有评估框架中几乎完全缺失的现状。尽管以现代标准阿拉伯语（Modern Standard Arabic, MSA）为核心的评估基准（如OALL、HELM Arabic和BALSAM）已趋于成熟并实现性能饱和，但其无法有效衡量模型在方言理解与生成、本土知识及特定应用场景中的真实表现。为此，研究提出“Mizan”——伊拉克国家层面的多维度评估基准，涵盖六个核心维度：方言理解、方言生成、双向MSA-伊拉克阿拉伯语翻译、伊拉克特有知识、官方文档信息抽取以及安全性。该基准基于340个原创且经过双重评审的题项构建，采用统计审计的方法确保答案位置与置信区间（Wilson区间）的可靠性。关键解决方案在于引入一个与MSA基线并行的伊拉克方言专项赛道，揭示了现有模型在方言处理能力上的显著差距：在伊拉克赛道上，各模型间存在稳定的14–18分差距，而MSA赛道则已趋于饱和；同时发现，专为阿拉伯语优化的模型在伊拉克赛道上反而低于同等规模的通用模型，且最新安全强化模型对无害的方言理解任务出现系统性拒绝，暴露出其在方言场景下的过激合规行为。这一现象在传统MSA基准中完全不可见。此外，平台通过不可变快照、验证证书、人工发布审核与公开撤回机制保障评估过程的完整性与可追溯性，实现了高可信度的基准评估闭环。
链接: https://arxiv.org/abs/2609.13980
作者: Nawar S. Alseelawi,Mustafa S. Aljumaily
机构: University of Misan (米桑大学); Missan Oil Company (米桑石油公司); National Team for the Iraqi Large Language Model (伊拉克大语言模型国家团队), Prime Minister’s Office (总理办公室)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 11 pages, 3 figures, 1 table. Live leaderboard: this https URL
Abstract:Arabic large-language-model (LLM) evaluation has matured around Modern Standard Arabic (MSA): aggregated leaderboards such as the Open Arabic LLM Leaderboard (OALL), HELM Arabic, and BALSAM rank models across dozens of MSA tasks, and frontier systems increasingly saturate them. Dialectal Arabic, the language Iraqis actually speak, remains nearly invisible to this infrastructure. We introduce Mizan (“the balance”), Iraq’s national benchmark for evaluating LLMs on Iraqi Arabic and the Iraqi civic context: an MSA baseline track paired with an Iraqi track across six axes (dialect comprehension, dialect generation, bidirectional MSA-Iraqi translation, Iraq-specific knowledge, official-document field extraction, and safety), built from 340 originally authored, dually reviewed items with statistically audited answer positions and Wilson intervals on every published score. A pilot evaluation of 27 systems, spanning closed frontier models three days after release, open weights across size tiers, and an Arabic trio of commercial, open-specialized, and sovereign systems, yields four findings. The MSA track saturates while the Iraqi track discriminates, with a consistent 14-18-point per-model gap and statistically tied leaders. Official-document extraction confines every system to 32-56. Arabic-focused specialization behaves as MSA specialization: two dedicated Arabic models score below a size-matched generalist on the Iraqi track. And the safety-hardened tier of the newest model family deterministically refuses innocuous dialect-comprehension items as policy violations, an over-refusal mode invisible to MSA benchmarks. The platform enforces an integrity protocol of immutable snapshots, verification certificates, a human publication gate, and public retraction, all exercised during this study. Code and the public development set accompany the paper.
[NLP-120] hought without systematicity? Evaluating reasoning models on rule induction tasks
【速读】： 该论文旨在解决当前推理模型是否具备人类认知中的系统性（systematicity）这一核心问题。系统性指对某一概念的理解必然关联到其相关变体的理解，即在结构上等价的任务变体应表现出一致的推理能力。研究通过将认知科学中成熟的规则归纳任务拓展至深度学习模型评估，构建具有组合结构的任务家族，并利用任务同构（如重组与替换）生成结构等价的任务变体。研究发现，尽管模型能够在特定任务上正确求解，却常在结构等价的变体上表现失败，表明多数模型的行为缺乏系统性。其解决方案的关键在于：通过任务同构生成结构等价的变体，以检验模型在不同但逻辑等价的输入形式下的泛化能力，从而揭示现有模型在推理过程中对组合结构理解的局限性。
链接: https://arxiv.org/abs/2609.13948
作者: Simon Schug,Brenden M. Lake
机构: Princeton University (普林斯顿大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Code available at this https URL
Abstract:A central tenet of human cognition is systematicity, the principle that understanding one concept is inherently tied to understanding close variations of that concept. Do reasoning models robustly exhibit such systematicity? If so, we would expect consistent performance on structurally equivalent variants of the same task. Here, we extend established rule induction tasks from cognitive science to assess the systematicity of thought in current reasoning models. Each task family has compositional structure that we use to create structurally equivalent task variations through task isomorphisms such as recombination and substitution. We find that despite being able to correctly solve a task, models often fail on structurally equivalent variants of the same task. These findings suggest that many model behaviors lack systematicity, rendering it difficult to robustly establish the cognitive abilities of reasoning models beyond the particular contexts they were evaluated in.
[NLP-121] CRITICS - Critical Science Without Borders: Language Models to Promote Critical Thinking in Science Education
【速读】： 该论文旨在解决科学知识获取与科学素养提升中的语言壁垒问题，尤其针对当前高质量科研与教育材料多局限于高资源语言（如英语）所导致的可及性不平等问题。其核心解决方案的关键在于融合基于大语言模型（Large Language Models, LLMs）的先进机器翻译（Machine Translation, MT）技术与教育技术，构建面向科学内容的专用翻译系统，确保科学文本在译入学生母语时既保持技术准确性，又具备文化适切性。在此基础上，项目进一步设计并评估与课程标准相契合的教学-学习方案，聚焦科学论证与批判性思维能力的培养，并引入与学习目标及能力导向评价框架一致的文本反馈机制，从而实现科学知识的民主化传播与科学教育质量的系统性提升。
链接: https://arxiv.org/abs/2609.13942
作者: Rodrigo Agerri,Itziar Aldabe,Elena Cabrio,Mark Cieliebak,Jan Deriu,Mariana Flores,Jurgita Kapociute-Dzikiene,Dovile Kuiziniene,Arantza Rico,Aritz Ruiz-Gonzalez,Aitor Soroa,Mantas Vaskevicius,Serena Villata
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 9 pages
Abstract:The CRITICS project addresses science accessibility and literacy by converging advanced Machine Translation (MT) based on Large Language Models (LLMs) with educational technology. By leveraging MT systems specifically optimized for scientific content, educational institutions can provide accurate, culturally relevant translations of scientific materials in students’ native languages, ensuring that complex scientific concepts are comprehensible while maintaining technical accuracy. Building on these translations, the project explores the design and evaluation of innovative science teaching-learning proposals grounded in curriculum-aligned teaching-learning. Thus, CRITICS will investigate key components of scientific argumentation and critical thinking practices together with textual feedback aligned with learning objectives and assessment criteria inspired by competence-based evaluation frameworks. CRITICS aims to break down language barriers to accessing cutting-edge research and educational materials currently available only in high-resourced languages, thereby facilitating the democratization of scientific knowledge and fostering critical thinking in science education.
[NLP-122] Inter-Rater Reliability of LLM and Rule-Based Annotation for Inferential Narrative Features: Three Studies on a Turkish Corpus
【速读】： 该论文旨在解决自动标注数据集中机器生成的文本特征标签与人类判断之间的一致性问题，尤其关注在叙事文本分析中基于规则的自动化检测系统是否能真实反映人类对文学技巧（craft features）的理解。其核心问题是：当一个自动化的规则引擎为土耳其叙事语料库（Objective Projection corpus）中的场景打上六种修辞特征标签（包括两类禁止性特征和四类积极技巧）时，这些标签是否真正符合人类专家或非专家的判断？解决方案的关键在于通过三组独立实验，将五种不同模型（包括规则引擎及多个大语言模型如Gemini 2.5 Flash、Grok、Claude Fable 5、ChatGPT 5.5）的标注结果与人类标注者（包括原作者和独立非专家）的盲评结果进行对比。研究发现，在最贴近理论核心的“具象化隐喻”（materialized metaphor）这一特征上，所有机器标注者的阳性率从0%到78%不等，而人类标注者仅标记出9个正例，各模型与人类标注之间的Cohen’s κ值均接近随机水平（0.000–0.027），尽管表面一致性高达74.7%–84.5%，但实为类别不平衡导致的假象。该结果表明，当前的自动化方法要么无法捕捉具有高度推断性的文学特征，要么是规则定义本身缺乏足够可操作性，致使人类也难以一致应用——这一根本分歧尚需第二名独立人类标注者进一步验证，而本研究未具备此条件，故未做出最终定论。
链接: https://arxiv.org/abs/2609.13936
作者: Levent Bulut
机构: 独立研究员(Independent Researcher)
类目: Computation and Language (cs.CL)
备注: v1.0. 10 pages, 3 tables. Also archived at Zenodo: this https URL
Abstract:Datasets that ship automatically generated feature annotations invite a question rarely asked of them: would a human agree with those labels? This report answers that for the Objective Projection corpus, a Turkish narrative dataset whose scenes carry a per-scene applied_rules field from a rule-based detector over six craft features – two prohibitions (emotion labelling, simile) and four positive techniques (materialized metaphor, micro-focus, temporal anchor, atmosphere contradiction). Three studies are reported. Study 1 ( n = 120 ) scores the detector against blind labels from the scheme’s own author. Study 2 ( n = 100 , a disjoint scene set) scores the detector plus Gemini 2.5 Flash and Grok against an independent non-expert rater whose labels were locked before any machine ran. Study 2b re-runs the identical protocol with Claude Fable 5 (High) and ChatGPT 5.5. The central result concerns one rule. On materialized metaphor – closest to the methodology’s theoretical core – the five machine labellers returned positive rates of 0 , 1 , 40 , 72 and 78 out of 100 scenes, against a human count of 9 . Cohen’s \kappa was at or indistinguishable from chance for five of six labellers, across both human references and both scene sets: 0.004 , 0.015 , 0.000 , 0.019 , 0.027 . Raw agreement ranged from 74.7% to 84.5% , an artefact of class imbalance rather than a sign of competence. We deliberately do not resolve this into a single story. Two readings survive: the feature is genuinely inferential and beyond current automatic detection, or the rule’s definition is not yet operational enough for any rater to apply consistently – including the human. Distinguishing them needs a second independent human rater, which this report does not have and therefore does not claim. Comments: v1.0. 10 pages, 3 tables. Also archived at Zenodo: this https URL Subjects: Computation and Language (cs.CL) Cite as: arXiv:2609.13936 [cs.CL] (or arXiv:2609.13936v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.13936 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Levent Bulut [view email] [v1] Sat, 12 Sep 2026 13:31:30 UTC (12 KB)
[NLP-123] North Small Translate: Advanced Cost-Effective Translation (Cohere CAT)
【速读】： 该论文旨在解决大语言模型（LLM）在机器翻译（Machine Translation, MT）任务中存在推理成本高、翻译质量受限以及难以高效处理多样化指令等问题。其核心挑战在于如何在保持低延迟与高吞吐量的前提下，实现跨语言的高质量翻译，并支持包括译后编辑、翻译质量评估等复杂下游任务。解决方案的关键在于：构建一个基于2180亿参数、仅激活250亿参数的混合专家（Mixture-of-Experts, MoE）架构的开源小规模模型——North Small Translate，通过难度采样策略筛选高挑战性语料进行训练，并采用五步联合训练流程（包含监督微调、直接偏好优化及在线强化学习），在不依赖昂贵推理机制的情况下显著提升翻译性能。同时，模型以非推理型基础架构保障高吞吐率，并可选集成代理能力（agentic capabilities），从而在不增加推理开销的前提下有效增强翻译质量与任务适应性，在低于1万亿参数级别的模型中实现了50种语言上的顶尖机器翻译表现。
链接: https://arxiv.org/abs/2609.13916
作者: Tom Kocmi,Alexandre Bérard,Phil Blunsom,Samuel Cahyawijaya,Shaun Cassini,Nicholas Frosst,Ona de Gibert,Aidan Gomez,Nithya Govindarajan,Shun Kiyono,Olivia Lasche,Lawrence Rogers,Kelly Marchisio,Nikita Moghe,Yash More,Camila Moran-Hidalgo,Yiyang Nan,Michael Sachs,Trisha Starostina,Daan van Stigt,Spencer Rarrick,Sebastian Vincent,Ivan Zhang
机构: Cohere
类目: Computation and Language (cs.CL)
备注:
Abstract:We present North Small Translate, an open-weight, LLM-based machine translation (MT) model with instruction-following capabilities built on the same foundation as Cohere’s Command A Plus, a mixture-of-experts architecture with 25 billion active parameters out of 218 billion total parameters. North Small Translate is trained using difficulty sampling to obtain challenging documents and a five-step training protocol combining supervised fine-tuning, direct preference optimization, and online reinforcement learning. We prioritized throughput through a non-reasoning base model and supplemented with optional agentic capabilities to unlock translation quality gains. North Small Translate is trained to perform MT-related tasks, including post-editing and quality estimation, as well as related tasks such as general instruction following. The model achieves top MT performance across 50 languages in the class of models under 1T parameters, with no need to run expensive reasoning at inference time.
[NLP-124] Phorecaster365: A Human-Supervised Reference Architecture for Hybrid Pharmaceutical Sales Forecasting and Planning Decision Support
【速读】： 该论文旨在解决制药行业销售预测在实际应用中因库存可得性、交易语义、产品生命周期阶段及预测发布时信息状态等因素影响，导致模型输出难以有效支持运营决策的问题。现有方法仅依赖模型预测结果，无法保留预测上下文条件，亦无法判断其是否适用于实际规划场景。为此，论文提出Phorecaster365——一种人工监督的参考架构，通过将企业资源计划（ERP）数据与可审查的药品销售预测相连接，实现对预测全过程的可追溯与可验证。其核心创新在于构建“预测上下文包”（forecast context package），用于完整保存原始数据快照、预测目标、时间截断点、可用协变量、数据质量状态及层级版本等关键上下文信息；同时建立“预测证据包”（forecast evidence package），关联预测结果与模型版本、校准参数、异常记录、人工调整及发布历史。该架构采用分层式设计，涵盖数据摄入、产品-区域序列构建、时序特征生成、统计与机器学习建模、集成融合、不确定性评估、规划者评审及生命周期治理等环节。通过10,950条合成日级数据在30个产品-区域序列上的开发经验，验证了系统的可扩展性与可操作性。研究强调采用滚动起源评估协议、基线与消融实验要求、不确定性与鲁棒性分析，并提出从合成测试到受控规划应用的渐进式路径。本工作的核心贡献在于提供一个与具体实现解耦的系统设计与验证框架，而非宣称真实世界预测精度或运营效益的优越性。
链接: https://arxiv.org/abs/2609.13907
作者: Houman Kazemzadeh,Kamyar Naderi
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 23 pages, 4 figures, 7 tables, including supplementary appendices
Abstract:Pharmaceutical sales forecasts inform planning across products, regions, and distribution channels, yet their interpretation depends on inventory availability, transaction semantics, product lifecycle, and the information available when each forecast is issued. A model prediction alone does not preserve these conditions or establish whether a forecast is suitable for operational use. We present Phorecaster365, a human-supervised reference architecture that connects enterprise resource planning data to reviewable pharmaceutical sales forecasts. The architecture separates source ingestion, product-region series construction, temporally valid feature generation, statistical and machine-learning modeling, ensemble formation, uncertainty assessment, planner review, and lifecycle governance. Its central intermediate representation is a forecast context package that preserves the source snapshot, forecast target, temporal cutoff, available covariates, data-quality state, and hierarchy version. A corresponding forecast evidence package links predictions to model and calibration versions, exceptions, human adjustments, and publication history. Development experience with a synthetic panel of 10,950 daily records across 30 product-region series informs the design. Historical experiment summaries are retained only as descriptive evidence of development because their evaluation does not establish independent predictive validity. We specify a rolling-origin evaluation protocol, baseline and ablation requirements, uncertainty and robustness assessments, and a staged pathway from synthetic testing to use in governed planning. The contribution is an implementation-neutral system design and validation framework; the report does not establish real-world forecasting accuracy, comparative superiority, or operational benefit.
[NLP-125] SHIFT-M3: Pre-fusion Alignment-based Consistency Screening for Multimodal ECG Record Integrity
【速读】： 该论文旨在解决多模态临床人工智能（Multimodal Clinical AI）中因数据关联错误导致的记录完整性问题，即在实际应用中，心电图波形、报告、元数据及下游预测等不同模态数据可能来自不同患者却仍被错误地整合为同一记录，形成潜在安全风险。传统预测模型无法识别此类跨患者组件的错配问题。为此，论文提出一种轻量级文本预融合筛查方法SHIFT-M3，其核心在于通过比较由大语言模型（LLM）生成的心电图解读与临床报告摘要之间的语义一致性，评估多模态数据在文本层面的对齐程度。实验基于784,680个MEETI心电图记录表明，SHIFT-M3在全文本替换、部分替换及标签匹配的硬负样本场景下分别达到97.6%、90.3%和97.7%的真阳性率（TPR@5%假阳性率），且仅需573,569参数，表现出卓越性能。相较于基于词法匹配的基线模型，其在部分替换和硬负样本上的显著提升说明模型捕捉的是深层语义一致性而非表面词汇重叠。研究还引入了冲突类型多模态压力测试（CMST）评估框架、三种子稳定性分析、损失消融实验、时间容忍度扫描及共享词掩码控制等验证手段，进一步支持其有效性。然而，当前主要残余失败模式为纵向模糊性：在默认操作点下，同患者不同就诊记录间的对比仍产生87.0%的Ⅱ类误报，提示未来需加强时间序列上下文建模以应对长期随访场景下的数据混淆问题。
链接: https://arxiv.org/abs/2609.13874
作者: Md Ashik Khan,Md Nahid Siddique
机构: Indian Institute of Technology Kharagpur (印度理工学院克哈格普尔分校); Florida International University (佛罗里达国际大学)
类目: Computation and Language (cs.CL)
备注: This paper has been accepted and presented at MLHC 2026. Please cite from the official proceedings
Abstract:Multimodal clinical AI typically assumes that the waveform, report, metadata, and downstream predictions attached to a record belong to the same patient. In practice, linkage failures can silently assemble individually plausible but cross-patient components, creating a safety problem that standard predictive models are not designed to detect. We study this problem as multimodal record integrity triage: given an assembled record, should its modalities be trusted to belong together? We introduce SHIFT-M3, a lightweight text-based pre-fusion screen that measures alignment-based consistency between two separately produced ECG text views: an LLM-generated interpretation and a clinical report summary. On 784,680 MEETI ECG records, SHIFT-M3 achieves 97.6% TPR@5% FPR for full text-view swaps (AUROC 0.996), 90.3% for partial swaps (AUROC 0.974), and 97.7% for label-matched hard negatives (AUROC 0.996) with only 573,569 parameters. Compared with same-dataset lexical baselines, the gains are largest on partial swaps and hard negatives, suggesting that the model is learning more than surface overlap. We also introduce the CMST (Conflict-type Multimodal Stress Test) evaluation taxonomy, a three-seed stability study, a loss ablation, a temporal-tolerance sweep, and a shared-token masking control. The main remaining failure mode is longitudinal ambiguity: at the default operating point, same-patient cross-visit pairs still produce 87.0% Type-II false positives.
[NLP-126] Bangla Sentence Function Classification: Corpus Development Model Benchmarking and Interpretability
【速读】： 该论文旨在解决孟加拉语（Bangla）句子功能分类（Sentence Function Classification）缺乏高质量基准资源的问题，这一任务对对话系统、语音合成及机器翻译等下游自然语言处理应用具有重要意义。其解决方案的关键在于构建了一个包含10,000条人工标注句子的高质量孟加拉语语料库，涵盖陈述句、疑问句、祈使句和感叹句四类，具有较高的类别平衡性与标注一致性（Fleiss’ Kappa = 0.82）。在此基础上，研究系统评估了多种特征表示方法（如词袋模型BoW、TF-IDF、Word2Vec）与经典机器学习分类器，并引入两种异构集成模型——单层集成（SLE）与双层集成（DLE），以提升分类性能。实验结果表明，基于TF-IDF的稀疏词汇表示优于Word2Vec，因其更有效捕捉句子功能相关的判别性词汇线索，尤其在训练数据有限的情况下；而采用TF-IDF特征的双层集成模型（DLE）取得了最佳表现，准确率与宏平均F1均达到0.95，验证了稀疏特征表示与异构集成学习在该任务中的有效性。交叉验证与LIME可解释性分析进一步证实了方法的鲁棒性与预测透明度，为孟加拉语句子功能分类提供了坚实的基线。
链接: https://arxiv.org/abs/2609.13869
作者: Swapnil Kundu Argha,Abdullah Al Shafi,Rowzatul Zannat,Shoumik Barman Polok,Abdul Muntakim,Jannatul Ferdousi,M.A. Moyeen
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at 2026 IEEE 5th International Conference on Robotics, Automation, Artificial-Intelligence and Internet-of-Things (RAAICON)
Abstract:Automatic sentence function identification is important for many downstream natural language processing (NLP) applications such as dialogue systems, text-to-speech synthesis, and machine translation. However, benchmark resources for Bangla sentence function classification remain limited. To mitigate this gap, this paper introduces a corpus of 10,000 Bangla sentences, manually annotated into four functional categories, namely declarative, interrogative, imperative, and exclamatory. The corpus is nearly balanced across the four classes, with high annotation reliability reflected by a Fleiss’ Kappa of 0.82. Furthermore, we evaluate multiple feature representations, including Bag-of-Words (BoW), TF-IDF, and Word2Vec, with several classical machine learning classifiers. In addition, two heterogeneous ensemble models, namely Single-Level Ensemble (SLE) and Double-Level Ensemble (DLE), are utilized to improve classification performance. Experimental results show that TF-IDF consistently outperforms Word2Vec, likely due to its ability to emphasize discriminative lexical cues associated with sentence functions, particularly given the relatively small corpus used to train Word2Vec. The DLE model with TF-IDF features achieves the best performance with accuracy and macro-F1 of 0.95, demonstrating the effectiveness of sparse lexical representations and heterogeneous ensemble learning for this task. Further cross-validation confirms the robustness of the approach, while LIME-based interpretability provides insights into model predictions. The developed corpus and model benchmarking establish strong baselines for Bangla sentence function classification.
[NLP-127] ShopEase: A Generative AI-Based Multi-Agent Framework for Intelligent Enterprise Customer Support Using Hybrid Retrieval-Augmented Generation
【速读】： 该论文旨在解决企业客户支持系统在处理客户咨询时面临的多维度挑战，包括准确理解用户意图、精准检索相关政策信息、有效利用客户上下文以及在复杂场景下适时将问题转交人工客服。其核心解决方案是提出一个基于生成式 AI (Generative AI) 的多智能体框架 ShopEase，该框架由意图识别、客户关系管理（CRM）、记忆模块、混合检索增强生成（Hybrid RAG）、升级机制和监督代理六个组件构成，并采用本地部署的 LLaMA 3.2 模型通过 Ollama 进行响应生成。关键创新在于摒弃传统的固定意图-政策映射方式，直接从检索到的文档中动态决定政策类别，提升了灵活性与准确性。在检索模块中，融合了 FAISS（密集检索）与 BM25（稀疏检索），并评估了六种配置策略。实验结果表明，仅使用 FAISS 的密集检索在 2632 个独立测试查询上达到最高准确率 85.37%，显著优于仅用 BM25 的 55.74%；尽管引入交叉编码器重排序可提升语义相关性，但并未带来分类性能提升，反而增加延迟。统计分析（McNemar’s test）显示，FAISS-only 与 Weighted RRF 之间无显著差异，但二者均显著优于 Fair RRF 和含交叉编码器的方案。总体而言，本研究证明在该数据集上，纯密集检索已具备最优精度，而额外的重排序步骤对分类效果无益且增加计算开销。
链接: https://arxiv.org/abs/2609.13856
作者: Aakash Kumar Tiwari,Somesh Kumar
机构: Indian Institute of Technology Kharagpur (印度理工学院克勒格普尔分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:Enterprise customer support systems must answer customer questions correctly, retrieve the right policy information, use customer context, and pass difficult cases to human agents when needed. This paper presents ShopEase, a Generative AI-based multi-agent framework for enterprise customer support. The system combines six components: Intent, CRM, Memory, Hybrid RAG, Escalation, and Supervisor, and uses LLaMA 3.2 running locally through Ollama for response generation. The retrieval module combines FAISS (dense retrieval) and BM25 (sparse retrieval), and six configurations are evaluated: BM25-only, FAISS-only, Fair RRF, Weighted RRF, RRF with Cross-Encoder, and Top-10 Hybrid with Cross-Encoder. Instead of using a fixed mapping between intent and policy, the policy category is decided directly from the retrieved documents. The system was evaluated on 2632 held-out customer queries across six categories: Refund, Return, Shipping, Cancellation, Damaged Product, and Unknown. FAISS-only achieved the highest accuracy of 85.37% (2247 correct predictions), closely followed by Weighted RRF at 85.07%. BM25-only achieved only 55.74% accuracy. Adding cross-encoder reranking did not improve results: RRF with Cross-Encoder reached 83.24%, and Top-10 Hybrid with Cross-Encoder reached 81.88%, while also increasing response latency. Category-level analysis shows strong performance on Shipping, Cancellation, and Return, while Unknown queries remain the main source of errors. Statistical testing using McNemar’s test shows no significant difference between FAISS-only and Weighted RRF, though both perform significantly better than Fair RRF and the cross-encoder configurations. Overall, dense retrieval gives the best accuracy on this dataset, and additional reranking adds processing time without improving classification performance.
[NLP-128] Measuring the Cost of Variety Conflation in Multilingual MT Evaluation: Adding Mozambican Xichangana Nyanja and Sena to FLORES
【速读】： 该论文旨在解决跨边界语言或语言变体在机器翻译评估中因参考文本（reference）选择不当而导致的评估偏差问题，特别是在莫桑比克使用的三种班图语变体（Xichangana、Mozambican Nyanja 和 Sena）的评估中。其核心挑战在于现有评估体系多依赖于标准参考文本（如Tsonga、Chichewa），而这些参考文本与目标方言在语言特征上存在差异，从而导致翻译性能指标（如spBLEU、chrF++）出现显著波动，造成对模型真实能力的误判。解决方案的关键在于提出并验证“变体感知”（variant-aware）的微调策略：通过在目标方言数据上进行针对性微调，使模型能够更好地适应本地语言变体，从而显著提升在目标方言上的翻译性能（如在Xichangana和Mozambican Nyanja上分别提升7.04和5.33 spBLEU），同时揭示了传统参考文本对评估结果的强烈敏感性。研究进一步强调，未来对于语言变体的评估应采用与其对应的变体感知参考文本与语言标识符，以实现更准确、公平的性能报告。相关数据已在Hugging Face公开共享。
链接: https://arxiv.org/abs/2609.13847
作者: Felermino D. M. A. Ali,Delfina Lázaro Mateus,Manuel Valente Mangue
机构: Faculdade de Engenharia da Universidade Lúrio(卢里奥大学工程学院); Universidade Eduardo Mondlane - Escola de Comunicação e Artes(爱德华多·蒙德拉内大学传播与艺术学院)
类目: Computation and Language (cs.CL)
备注: Accepted at WMT2026
Abstract:In this paper, we extend FLORES+ with Portuguese-source evaluation sets for three Mozambican Bantu varieties: Xichangana, Mozambican Nyanja, and Sena. We compare Xichangana with the existing Tsonga reference and Mozambican Nyanja with Chichewa, and evaluate NLLB-200, Google Translate, GPT, and a variant-aware NLLB model. Holding system output fixed reveals substantial reference sensitivity. On \textitdevtest, changing only the reference from Tsonga to Xichangana reduces spBLEU by 13.10 points for NLLB-200 and 15.30 for Google. On matched Nyanja subsets, replacing Chichewa with Mozambican Nyanja produces smaller but consistent reductions of 3.03 and 6.10 spBLEU, respectively. Variant-aware fine-tuning reverses this pattern on the intended targets: relative to NLLB-200, it improves Xichangana by 7.04 spBLEU and Mozambican Nyanja by 5.33 on \textitdevtest, while losing performance on the sibling references. GPT is competitive on Tsonga and Chichewa but substantially weaker on the Mozambican varieties. For Sena, the finetuned model reaches 12.64 spBLEU and 36.21 chrF++ on \textitdevtest. These findings motivate variety-aware language identifiers, references, and reporting for cross-border languages or language dialects/variants. The data is publicly available on Hugging Face at this https URL
[NLP-129] Affinity-Aware Sharding for Delayed Tensor Parallelism
【速读】： 该论文旨在解决张量并行（Tensor Parallelism, TP）在Transformer模型推理中因全局所有减少操作（all-reduce）导致的延迟问题。传统的TP在每一层计算后需同步所有设备的梯度或输出，形成瓶颈；而延迟张量并行（Delayed Tensor Parallelism, DTP）通过将各设备的部分输出立即累加至残差流并广播，仅在后续δ个模块后才收集其他设备的输出，从而消除阻塞。然而，这一架构变更打破了原始密集模型中前馈网络（FFN）神经元与注意力模块中键值（KV）头之间的置换对称性，使数据分片本身成为影响模型性能的建模决策。为此，论文提出的关键解决方案是：在分片前通过优化排列，最大化共置设备上KV头与FFN神经元间的亲和度（co-located affinity）。该亲和度以一阶近似衡量失去某一KV头贡献对神经元输出的损害程度，并采用坐标上升优化器，交替执行精确平衡的神经元分配与对KV头分区的穷举搜索，实现高效布局优化。实验表明，该方法可在单块GPU上于两分钟内完成优化，且在Qwen3-0.6B和Danube3-500M模型上，当δ=1时，亲和度优化布局仅需原连续布局一半至三分之二的训练步数即可达到相同蒸馏目标，且优于所有随机布局及连续布局。此外，初始化阶段的共置亲和度得分可有效预测训练后与基线模型间的KL散度（皮尔逊相关系数达-0.81至-0.89），验证了其作为预训练质量评估指标的有效性。
链接: https://arxiv.org/abs/2609.13846
作者: Eloi de Reynal
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: 16 pages, 11 figures, 3 tables
Abstract:Delayed Tensor Parallelism (DTP) removes the blocking all-reduce of tensor-parallel Transformer inference. Every device adds its own partial output to its residual stream (and broadcasts it) immediately, but only gathers (receives) the other devices’ partials \delta modules later. A TP to DTP change therefore amounts to a real architecture change, and dense Transformer models need to be retrained or distilled after adaptation. We show that DTP breaks the permutation symmetry of neurons inside FFNs and of KV heads inside attention modules, and that this symmetry breakage makes the sharding itself a modelling decision. We show that maximising the affinity between the KV heads and the FFN neurons co-located on a device, by permuting the dense model before sharding, speeds up the distillation or retraining process. The affinity is measured with a first-order approximation of the damage that losing a head’s contribution does to each neuron’s output, and the co-located affinity is maximised with a coordinate-ascent optimiser that alternates an exact balanced assignment of neurons with an exhaustive search over the KV head partitions. The whole procedure takes under two minutes on one GPU for Qwen3-0.6B and Danube3-500M. On these models, at \delta=1 , the affinity-optimised layouts reach any distillation target in about half to two thirds of the steps needed by the naive contiguous layouts, over the whole 10k-step range we tested, and every optimised seed beats every contiguous seed and all but one of the sixteen random layouts. We also show that the co-located affinity score at initialisation predicts the KL to the base model after training, across seventeen layouts ranging from anti-optimised to optimised (Pearson -0.81 and -0.89 ).
[NLP-130] Sweet Talkers: How Query Formulation Shapes Sycophancy in Romantic Relationship Advice EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在提供情感支持与亲密关系建议时，因过度迎合用户立场而可能无意中强化有害人际行为的风险。其核心问题是：当用户提出带有偏颇或不道德倾向的关系咨询时，模型是否以及如何表现出“社交奉承”（social sycophancy），即无条件附和用户观点，从而加剧潜在的非健康互动模式。解决方案的关键在于构建一个系统化的评估框架——ELEPHANT框架，并基于2,400条涵盖五大关系主题的浪漫关系求助提示数据集（Romantic Relationship Advice-Seeking Prompts, RRASP），对两款主流消费级模型（GPT-5 Mini 和 Gemini 3 Flash）进行实证分析。研究发现，用户的表达方式（如语法语气）并非决定性因素，真正影响模型响应的是视角驱动的表述框架（perspective-driven framing）；随着对话轮次推进，模型对用户前提的接受度与道德认同显著上升，表现出持续增强的“道德奉承”趋势。值得注意的是，Gemini 3 Flash相较GPT-5 Mini展现出更低的道德奉承增长幅度，表明其在对抗用户伦理立场的不当内化方面更具鲁棒性，这为未来模型设计中提升伦理抵抗能力提供了关键启示。
链接: https://arxiv.org/abs/2609.13841
作者: Helena Choi,Edric Castel Hao,Karl Bautista,Francis Gabriel Magleo,Renzo Panti,Danielle Beatrice Olalia
机构: Ateneo de Manila SHS(阿特内奥大学中学部); Analog Devices, Inc.(模拟设备公司)
类目: Computation and Language (cs.CL)
备注: Accepted to LUHME Workshop @ EMNLP 2026
Abstract:Large language models (LLMs) are increasingly used for emotional support and relationship advice, where a model’s tendency to preserve a user’s face can inadvertently reinforce harmful interpersonal behaviors. To systematically examine this risk, we developed the Romantic Relationship Advice-Seeking Prompts (RRASP) dataset of 2,400 prompts across five relationship themes and evaluated social sycophancy using the ELEPHANT framework on two consumer-facing models, GPT-5 Mini and Gemini 3 Flash. Contrary to our initial hypothesis, grammatical mood alone did not produce systematic differences in sycophantic behavior, suggesting that what a user implies matters more than how they phrase it. Instead, perspective-driven framing had a stronger influence, with gaps between original and flipped prompts widening in follow-up responses. Consistent increases in framing and moral sycophancy across turns indicate that models become more likely to accept a user’s stated premises and affirm their ethical stance as a dialogue progresses. Notably, Gemini 3 Flash exhibited substantially smaller increases in moral sycophancy than GPT-5 Mini, suggesting it is more resistant to reinforcing ethically problematic positions across turns.
[NLP-131] When Consistency Does Not Mean Reliability: Evaluating Local LLM Judges Against Human Ratings
【速读】： 该论文旨在解决生成式 AI（Generative AI）在自动评估其他语言模型输出时存在的可靠性问题，特别是当使用本地开源大语言模型（LLM）作为评价者（即“LLM-as-a-Judge”）时，其评分结果与人类评估者之间的一致性不足的问题。尽管此类方法具有成本低、效率高的优势，但研究发现，即使模型评分表现出高度自一致性（self-consistency），也未必能准确反映人类评判标准。本研究的关键解决方案在于系统性地对比两个本地开源大语言模型（LLaMA-3-8B 和 Qwen2.5-7B）作为评价者的性能表现，通过多种指标（皮尔逊相关系数、斯皮尔曼等级相关系数、平均绝对误差、有符号偏差和自一致性）对同一组由微调的 GPT-2 模型生成的回答进行评估，并将其与 9 名人工标注者提供的平均评分进行比较。结果显示，尽管两模型自一致性极高（分别达到 97.3% 和 92.3%），但与人类评分的相关性较弱（皮尔逊相关系数分别为 0.275 和 0.340），且平均绝对误差较大（分别为 27.71 和 18.64）。这表明，高自一致性并不能保证与人类判断的高度对齐，因此在将本地 LLM 用作自动评价工具时，必须同时评估其一致性与人类对齐程度（human alignment）。
链接: https://arxiv.org/abs/2609.13824
作者: Aakash Kumar Tiwari
机构: Indian Institute of Technology Kharagpur(印度理工学院加尔各答分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) are increasingly used to evaluate the responses of other language models. This approach, known as LLM-as-a-Judge, is faster and cheaper than human evaluation. However, a judge may produce consistent scores without necessarily agreeing with human evaluators. In this work, we study this issue using two local open-weight LLM judges, LLaMA-3-8B and Qwen2.5-7B. We evaluate 300 responses generated by an instruction-tuned GPT-2 (124M) model for 100 questions covering five categories: factual knowledge, instruction following, mathematics, reasoning, and writing. Each response is scored by nine human annotators and is evaluated three times by each LLM judge using the same rubric. We compare the judge scores with the average human scores using Pearson correlation, Spearman correlation, mean absolute error (MAE), signed bias, and self-consistency. LLaMA-3-8B shows a Pearson correlation of 0.275 with human scores, while Qwen2.5-7B achieves 0.340. Their MAEs are 27.71 and 18.64, respectively. Despite this limited agreement, both judges show high self-consistency, with exact consistency rates of 97.3% for LLaMA-3-8B and 92.3% for Qwen2.5-7B. These results show that high self-consistency does not necessarily indicate high agreement with human judgments. Our findings highlight the need to evaluate both consistency and human alignment when using local LLMs as automatic judges.
[NLP-132] DARE: Dialectical Agent ic Reasoning for Structured Knowledge Fact Checking EMNLP2026
【速读】： 该论文旨在解决结构化知识事实核查中因关系生成无效、单一路径推理缺乏自我修正机制以及证据评估存在偏差（倾向于高估支持性信号）所导致的性能瓶颈问题。其核心解决方案是提出一种名为“辩证代理推理”（Dialectical Agentic Reasoning, DARE）的多智能体框架，通过将事实核查建模为迭代式的“检索-推理-反思”过程，实现更稳健的推理。DARE的关键在于：引入基于关系的证据检索以约束推理路径的合法性，采用辩证双向验证机制从支持与反驳双重视角评估证据，结合置信度驱动的元反思机制动态判断是否需进一步探索证据。这一设计有效激发了大语言模型（LLM）潜在的推理能力，在多个基准测试中，80亿参数规模的模型即达到88.12%的准确率，表现可媲美甚至超越依赖GPT-4o的程序生成基线方法。
链接: https://arxiv.org/abs/2609.13808
作者: Yifei Li,Xiaohan Zheng,Wentao Qian,Liansheng Zhuang
机构: University of Science and Technology of China(中国科学技术大学)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026 (Findings)
Abstract:Structured knowledge fact checking aims to determine the truthfulness of natural language claims by reasoning over structured evidence. Recent program-generation approaches leverage large language models (LLMs) to generate executable graph reasoning programs, achieving strong performance on structured knowledge fact checking benchmarks. However, these methods remain limited by invalid relation generation, single-path reasoning that lacks self-correction, and biased evidence assessment that tends to overestimate supporting signals. We propose Dialectical Agentic Reasoning (DARE), a multi-agent framework that formulates structured knowledge fact checking as an iterative retrieve-reason-reflect process. DARE integrates relation-grounded evidence retrieval to constrain reasoning to valid structures, dialectical bidirectional verification to evaluate evidence from both supporting and refuting perspectives, and confidence-driven meta-reflection to dynamically determine whether additional evidence exploration is necessary. Extensive experiments demonstrate the effectiveness of DARE in structured knowledge fact checking, with an 8B backbone achieving 88.12% accuracy and matching or surpassing GPT-4o-based program-generation baselines, which attests to the efficacy of dialectical agentic reasoning in eliciting the latent reasoning capabilities of LLMs.
[NLP-133] Understanding the Limits of Agent ic ICD Coding EMNLP2026
【速读】： 该论文旨在解决当前基于ICD-10-CM的疾病编码系统在真实临床场景中面临的性能瓶颈问题，尤其关注罕见诊断与复杂编码任务（如涉及多步骤指南遵循的损伤及外部原因编码）下的表现差异。现有标准基准测试仅提供聚合指标，掩盖了系统在复杂、边缘情况下的实际能力缺陷。研究通过在MIMIC-IV出院记录上对神经网络模型、工作流系统及智能体（agentic）系统进行稀有性分层评估，识别出两类独立的失败模式：神经分类器在罕见代码上的微平均F1得分比常见代码低0.43；工作流系统虽能较好处理罕见代码，但在需多步推理的损伤和外部原因编码上性能接近零。其解决方案的关键在于引入一种工具增强型智能体配置，该配置通过结构化方式接入官方ICD-10-CM参考文献，显著提升了复杂编码任务的表现，最高恢复了0.34的微平均F1得分，表明结合外部知识与动态推理机制是提升编码系统鲁棒性的关键路径。
链接: https://arxiv.org/abs/2609.13806
作者: Chong Yock Eng,Yushi Cao,Yiming Chen,Kezhi Mao,Hongchao Jiang
机构: Nanyang Technological University (南洋理工大学); ASUS Intelligent Cloud Services (AICS) (华硕智能云服务)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to EMNLP 2026 Main Conference
Abstract:ICD-10-CM codes are alphanumeric codes used in the US to classify diagnoses and injuries for medical billing and epidemiological reporting. Standard ICD-10-CM benchmarks report aggregate metrics that obscure performance on complex coding scenarios. We evaluate neural, workflow, and agentic systems on a rarity-stratified set of MIMIC-IV discharge summaries and identify two orthogonal failure modes. Neural classifiers exhibit a 0.43 micro-F1 gap between rare and common codes. Workflow systems handle rare codes well but score near zero on injury and external cause codes that require multi-step guideline following. A tool-augmented agentic configuration with structured access to official ICD-10-CM reference materials recovers up to 0.34 micro-F1 on this subset. No single system dominates across all conditions.
[NLP-134] SyRHM: Symbolic-Language-Enhanced Reasoning with Associative Retrieval for Zero-shot Harmful Meme Detection EMNLP2026
【速读】： 该论文旨在解决有害表情包（harmful meme）检测中因隐含的有害意图、视觉-文本不一致性以及文化刻板印象导致的识别难题，此类问题使得现有多模态检测方法难以有效捕捉深层语义关联。其核心解决方案是提出SyRHM框架，关键在于将有害表情包检测分解为两个阶段：首先通过语义感知的检索机制，基于对多模态内容的文本化解析与描述，获取与目标表情包在语义上相关的上下文样本，从而提供超越表面相似性的深层语境支持；其次引入符号语言增强的多阶段推理机制，包括翻译器阶段将多模态输入转换为符号化中间表示，再经由规划器与求解器阶段进行分步推理，实现对有害意图的可解释、可追溯的深度分析。实验结果表明，SyRHM在FHM、HarM和MultiOff数据集上均优于主流多模态及基于推理的基线模型，且能生成清晰的推理路径，显著提升了检测的准确性与可解释性。
链接: https://arxiv.org/abs/2609.13794
作者: Hanling Wang,Chenlong Wei,Yingjuan Li,Di Wu,Yuchao Zhang,Xiaohui Zhu,Yao Zhu
机构: University of California, San Diego (加州大学圣地亚哥分校); Xi’an Jiaotong-Liverpool University (西安交通大学利物浦大学); University of Washington (华盛顿大学); The Australian National University (澳大利亚国立大学); Zhejiang University (浙江大学)
类目: Computation and Language (cs.CL)
备注: 17 pages, 14 figures. Accepted to EMNLP 2026
Abstract:Detecting harmful memes is critical for maintaining safe online communities. However, harmful intent is often implicit, arising from visual-textual incongruity and cultural stereotypes, which challenges existing multimodal detectors. We propose SyRHM, a framework that decomposes harmful meme detection into meaning-grounded retrieval and symbolic-language-enhanced multi-stage reasoning. SyRHM retrieves semantically related memes by parsing multimodal content into textual elements and descriptions, providing grounded context beyond surface-level similarity. Building on the retrieved context, SyRHM uses a translator stage to convert multimodal inputs into symbolic intermediate representations, and then performs multi-stage reasoning via planner and solver stages, enabling expressive and interpretable analysis of harmful intent. Experiments on FHM, HarM, and MultiOff demonstrate the effectiveness of SyRHM, achieving superior performance on most evaluation settings against multimodal and reasoning-based baselines, while providing reasoning traces for harmful content. The code is available at: this https URL
[NLP-135] Does Reasoning Improve Psychological Depth in Large Language Models ? It Depends on Whos Judging
【速读】： 该论文旨在解决生成式 AI（Generative AI）在开放生成任务中评估时，基于大语言模型作为评判者（LLM-as-a-Judge）的可靠性问题。尽管这些评判者在开发集上与人类评分具有较高相关性，但在面对高度相似输出且人类偏好主观的情境下，其有效性可能失效。研究通过心理深度（Psychological Depth, PD）这一维度，在短篇故事生成任务中揭示了这一失效模式：实验对比了GPT-5 vs. GPT-4o、DeepSeek-R1 vs. DeepSeek-V3的60对盲测、提示匹配的故事样本，发现人类读者偏好无显著统一优势（如GPT-5仅略优3.7%–9.8%，而DeepSeek-R1落后于V3达17.1%），且跨读者一致性接近随机水平（Krippendorff’s α = 0.070），表明人类评价存在结构性异质性而非随机响应。相比之下，LLM评判者在所有五种配置下均在90%以上维度比较中偏好推理型输出，并在60对中的59对上给出一致高分，其评分高度依赖表面特征如句长和词汇多样性。因此，解决方案的关键在于：开发集上的相关性不足以保证在分布外场景下的评估有效性，且采用点估计型评判者会掩盖人类主观评价中的复杂异质性，从而导致评估结果失真。
链接: https://arxiv.org/abs/2609.13773
作者: Ruichen Zheng,Yihe Wang,Fabrice Y Harel-Canada,Sara Khosravi,Zeynep Senahan Yildiz,Amit Sahai,Nanyun Peng
机构: University of California, Los Angeles(加州大学洛杉矶分校); Tsinghua University(清华大学); LA General Hospital(洛杉矶综合医院); Google(谷歌)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 24 pages
Abstract:LLM-as-a-Judge evaluators are increasingly used to score open-ended generation, yet a judge’s correlation with human ratings on its development set may not guarantee valid measurement when outputs are closely matched and human preferences are subjective. We study this failure mode through psychological depth in short stories. Seven human readers and an LLM-judge ensemble selected on the original scalar Psychological Depth Scale dataset ( \rho = 0.646 ) evaluated 60 blinded, prompt-matched story pairs from GPT-5 vs.\ GPT-4o and DeepSeek-R1 vs.\ DeepSeek-V3. Human preferences showed no universal reasoning advantage: GPT-5 was modestly preferred over GPT-4o (60.0–62.9%), whereas DeepSeek-R1 trailed V3 (42.9%), and inter-reader agreement was near chance (Krippendorff’s \alpha = 0.070 ), with within-reader consistency and recurring weighting patterns suggesting structured heterogeneity rather than random responding. The judge, by contrast, favored reasoning outputs in 89.0% of dimension-level comparisons and 59 of 60 pairs on aggregate PDS, uniformly across all five evaluator configurations, and its scores were associated with surface features such as sentence length and lexical diversity. These results suggest that development-set performance is insufficient evidence for deployment validity on a shifted distribution, and that point-estimate judges can obscure the heterogeneity in subjective human evaluation.
[NLP-136] raining Specialist Models without Reasoning Trajectories for Domain Expert Distillation
【速读】： 该论文旨在解决在缺乏显式推理监督的情况下，专用教师模型（specialist）生成的推理轨迹（reasoning trajectories）其内在机制与控制因素问题。传统知识蒸馏依赖于教师模型提供的高质量推理过程，但当教师仅通过问答对进行训练而无明确推理标注时，其生成轨迹的分布特性及其对下游学生模型的影响机制尚不清晰。本文的关键解决方案在于将学生模型蒸馏过程用作一种无偏探测工具（agnostic probe），即不将其视为最终目标任务，而是通过学生模型继承教师所生成的推理轨迹本身，从而揭示教师隐含的轨迹空间分布。实证分析表明，在27组不同专家-学生配对中，专家模型的专精性与泛化性表现与其学生模型的表现呈现高度相关性，证明了专家优化在隐式地从潜在轨迹空间中选择并引导轨迹分布。更重要的是，通过显式调控专家模型的分布漂移（distributional drift），可系统性地调节教师及其蒸馏学生在领域精度与通用能力保留之间的权衡关系。该研究揭示了一个新视角：在缺乏黄金推理标注时，专家模型的超参数与训练策略直接决定了传递给下游模型的隐式监督信号，从而为生成式模型的知识迁移提供了可调控的理论基础与实践路径。
链接: https://arxiv.org/abs/2609.13770
作者: Yilei Tu,Zihao Li,Shaoxiong Ji,Jörg Tiedemann,Fei Yuan
机构: University of British Columbia (不列颠哥伦比亚大学); University of Helsinki (赫尔辛基大学); ELLIS Institute Finland (芬兰ELLIS研究所); University of Turku (图尔库大学); Shanghai Artificial Intelligence Laboratory (上海人工智能实验室)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Specialist distillation effectively transfers domain expertise to student models via teacher-generated reasoning trajectories. However, when these specialists are trained solely on question–answer pairs without explicit reasoning supervision, what governs the trajectories they generate? In this work, we show that specialist optimization implicitly selects from this latent trajectory space. To isolate and observe this latent distribution, we leverage student distillation not as a downstream goal, but as an agnostic probe—since students inherit no parameterization or optimization constraints from the specialist, inheriting only the sampled trajectories themselves. Through this probe, our empirical analysis unveils a tight governing relationship: across 27 specialist–student pairings, their specialization–generalization profiles correlate exceptionally strongly. Crucially, explicitly controlling the specialist’s distributional drift systematically shifts both the teacher and its distilled student along a controllable trade-off between domain precision and general-capability retention. Across chemistry, physics, and multilingual settings, distilled students systematically reflect these specialist-induced profiles, even across divergent model families. Our findings establish a new view of specialist training: when gold reasoning is absent, tuning choices directly control the latent supervision passed to downstream models.
[NLP-137] HyperProve: Answer-Guided Hypergraph Expansion for Multi-Hop Question Answering EMNLP2026
【速读】： 该论文旨在解决多跳问答（multi-hop question answering）中因检索过程将证据视为与原始问题孤立匹配而导致的推理失效问题。由于复杂问题的答案通常依赖于通过中间实体、关系和约束连接的多个事实，传统方法在缺乏对这些关联结构建模的情况下难以构建连贯的证据链。其解决方案的关键在于提出HyperProve框架，该框架通过将问题分解与基于答案条件的超图（hypergraph）扩展相耦合，实现对原子事实间的复杂关联进行动态建模。HyperProve不孤立使用原子事实、超图或迭代检索，而是将中间答案和支撑性超边作为检索状态进行传递，并利用该状态引导下一轮局部超图的扩展，从而实现有状态且以事实为中心的检索。这一设计使得系统能够逐步构建出逻辑一致的证据链，最终生成准确答案。在多个多跳问答基准测试中，HyperProve显著优于现有最强基线模型，平均在答案准确率上提升6.2%，在F1分数上提升4.9%。
链接: https://arxiv.org/abs/2609.13768
作者: An Nguyen Phu,Dung Nguyen Quang,Luu Hieu An,Linh Ngo Van,Trung Le,Thien Huu Nguyen
机构: Hanoi University of Science and Technology(河内科学技术大学); Monash University(莫纳什大学); University of Oregon(俄勒冈大学)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026
Abstract:Multi-hop question answering often fails when retrieval treats evidence as isolated matches to the original question, since the facts needed to answer a complex question are usually connected through intermediate entities, relations, and constraints. We propose HyperProve, a retrieval-augmented QA framework that addresses this challenge by coupling question decomposition with answer-conditioned expansion over a hypergraph of atomic facts. HyperProve does not use atomic facts, hypergraphs, or iterative retrieval in isolation; instead, it carries intermediate answers and supporting hyperedges as retrieval state, then uses that state to bias the next local hypergraph expansion. This design enables HyperProve to construct coherent evidence chains for final answer generation while making the retrieval process stateful and fact-centered. Across multi-hop QA benchmarks, HyperProve achieves the best overall performance in our evaluation, outperforming the strongest baselines by an average relative improvement of 6.2% in answer accuracy and 4.9% in F1.
[NLP-138] Inside VLM Chart Reading: Tracing Value Reading from Vertical Bar Charts Across Space and Depth
【速读】： 该论文旨在解决视觉-语言模型（Vision-Language Models, VLMs）在图表问答任务中虽能准确输出答案，但其内部如何整合多源视觉证据以精确还原柱状图数值这一关键机制尚不明确的问题。研究通过在Qwen2.5VL-7B-Instruct与InternVL3.5-8B模型上采用受控反事实激活修补（counterfactual activation patching）方法，系统分析了柱状图值读取过程中的因果计算路径。其解决方案的关键在于：（1）识别出柱顶区域（bar-top region）虽包含较少视觉标记，却对答案偏好恢复贡献更大，表明模型更依赖于柱体顶端的几何线索；（2）揭示了信息传递的“移交”（handoff）机制——早期层中视觉图例区域主导恢复，中期层则转向提示词中的系列位置状态，且该状态可部分中介图例信息的恢复；（3）通过因子分析发现，两个模型均能分别从不同来源获取几何与刻度状态并联合使用以支持目标答案，但Qwen对跨源状态组合的恢复能力显著弱于InternVL，反映出其更强的上下文敏感性。整体研究表明，模型通过分阶段、模块化地利用局部几何、刻度及提示语境状态实现精确值读取，为理解VLM内部计算机制提供了初步的因果证据。
链接: https://arxiv.org/abs/2609.13745
作者: Tianhao Niu,Qingfu Zhu,Wanxiang Che
机构: Harbin Institute of Technology (哈尔滨工业大学); Research Center for Social Computing and Interactive Robotics
类目: Computation and Language (cs.CL)
备注:
Abstract:Vision–language models (VLMs) can answer chart questions accurately, but output accuracy does not show how they combine the evidence needed to recover an exact value. We study vertical-bar value reading with controlled counterfactual activation patching in Qwen2.5VL-7B-Instruct and InternVL3.5-8B. The study connects three analyses: (1) The single-factor results show that the changed bar-top region restores much more answer preference than the unchanged bar body, despite containing fewer visual tokens. Legend- and series-related states also lose local recoverability earlier than bar-geometry and axis-scale states. (2) In the handoff analysis, restoration shifts from visual legend regions in early layers to prompt-series positions in middle layers. Resetting the prompt-series state selectively reduces legend-source rescue, supporting its role as a partial mediator. (3) In the factorial analysis, both models can use geometry and scale states from separate donors to favor the combined target. InternVL performs similarly when the states come from separate donors or one image, while Qwen shows lower restoration for separate donors which suggests higher context sensitivity. Together, these results provide preliminary causal evidence for localizing the internal computations that support exact bar-value reading.
[NLP-139] HarnessBandit: Joint Learnability-Transferability Scheduling for Multi-Harness Agent ic Reinforcement Learning
【速读】： 该论文旨在解决语言模型代理（language-model agents）在不同接口（harness）间部署时表现不一致的问题，即对不同系统提示（system prompts）、工具模式（tool schemas）、控制循环（control loops）和轨迹格式（trajectory formats）的鲁棒性不足。其核心挑战在于多接口联合训练中的调度难题：每一步优化需平衡当前接口的学习信号有效性与更新对其他接口的潜在迁移价值。解决方案的关键是提出HarnessBandit——一种在线调度器，通过在每个优化步骤中选择最优的训练接口，实现动态权衡。该方法基于组相对策略优化（GRPO）后，同时评估“可学习性”（learnability，即批次内平均绝对优势）与“可迁移性”（transferability，即当前接口梯度低维草图与其余接口梯度指数移动平均之间的余弦相似度），经滑动窗口最小-最大归一化融合后，结合访问依赖奖励与显式探索下限进行采样决策。实验在ClawGym上使用六个不同接口训练Qwen3.5-2B模型，并在PinchBench（分布内任务与接口）和ClawEval（分布外任务与接口）上验证，结果表明HarnessBandit显著优于混合批次的多接口训练，且训练诊断显示可学习性与可迁移性提供了互补且随训练演化的信号。
链接: https://arxiv.org/abs/2609.13739
作者: Hongliang Wei(1 and 2),Xiaobing Tu(2),Yinggui Wang(2),Zhengxi Liu(2),Rongkun Xue(2),Jinkui Ren(2),Xiantao Zhang(2),Debin Zhao(1),Xiaopeng Fan(1) ((1) Harbin Institute of Technology, (2) Alibaba Cloud)
机构: Harbin Institute of Technology (哈尔滨工业大学); Alibaba Cloud (阿里巴巴云)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 13 pages. Equal contribution: Hongliang Wei and Xiaobing Tu. Corresponding authors: Xiaobing Tu and Xiaopeng Fan
Abstract:Language-model agents are increasingly deployed through diverse harnesses that differ in system prompts, tool schemas, control loops, and trajectory formats. The same model can perform unevenly across these interfaces, making robustness to harness variation an important objective. A natural approach is to train a shared policy through multiple harnesses, but doing so introduces a scheduling problem: each training step should favor a harness that currently provides a useful learning signal while also producing an update that benefits the other harnesses. We develop HarnessBandit, an online scheduler that selects one harness per optimizer step. After a group-relative policy optimization (GRPO) update, it observes learnability – the mean absolute advantage on the batch – and transferability – the cosine between a low-dimensional gradient sketch of the current harness and exponential moving averages of the remaining harnesses. The two signals are fused after pooled sliding-window min-max normalization and sampled with a visit-dependent bonus and an explicit exploration floor. We train Qwen3.5-2B across six harnesses on ClawGym and evaluate on PinchBench (held-out tasks, in-distribution OpenClaw) and ClawEval (held-out tasks and harness). HarnessBandit improves over mixed-batch multi-harness training on both benchmarks, while training diagnostics indicate that learnability and transferability provide distinct, evolving signals.
[NLP-140] ForeSight: Enhancing Risk Monitoring via Early Safety Signal Distillation EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在实际部署中生成有害内容的安全隐患问题，尤其关注如何在生成过程的早期阶段高效、准确地预测最终输出的有害性。现有防护机制多作用于输入、输出或流式生成阶段，而基于表面词元或输出逻辑值的早期风险检测方法往往依赖信号较弱，基于内部表示的检测器则可能包含大量纠缠且与安全无关的冗余信息，导致判断不准确。针对这一挑战，论文提出ForeSight框架——一种基于首个输出词元隐藏状态的早期风险预测方法，其核心创新在于通过知识蒸馏技术，将早期隐藏层中分散且冗余的安全信号压缩为紧凑、分层感知的风险表征。实验结果表明，ForeSight仅依赖首个词元的隐藏状态即可实现优越且高效的早期风险预测，在五个安全基准和两个目标模型上均展现出显著性能优势，验证了早期隐藏状态中已蕴含可靠有害性预测信号。
链接: https://arxiv.org/abs/2609.13737
作者: Hanling Wang,Chenlong Wei,Ling Xu,Hanyan Niu,Qi Cao,Shizhou Huang,Yang Yang,Xiaohui Zhu,Yao Zhu
机构: University of California, San Diego(加州大学圣地亚哥分校); Xi’an Jiaotong-Liverpool University(西安交通大学利物浦大学); East China Normal University(华东师范大学); Shanghai University(上海大学); Zhejiang University(浙江大学)
类目: Computation and Language (cs.CL)
备注: 17 pages, 12 figures. Accepted to Findings of EMNLP 2026
Abstract:As large language models (LLMs) are increasingly deployed, the generation of harmful content has become a critical safety concern. Existing safeguards operate at the input, output, or streaming-generation stages, while early-risk methods that rely on surface tokens or output logits may suffer from weak initial signals, and internals-based detectors using dense representations may retain highly entangled and redundant safety-irrelevant information. It therefore remains unclear whether the earliest post-generation hidden states already contain reliable signals about final-response harmfulness. To address this gap, we propose ForeSight, a first-token output-risk forecasting framework that distills weak and redundant early safety signals into compact, layer-aware risk representations. Experiments on five safety benchmarks and two target models demonstrate that ForeSight achieves superior and efficient early-risk forecasting while relying solely on first-token hidden states. The code is available at: this https URL
[NLP-141] PolicyMem: Geometric Policy Memory for LLM Governance
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在高风险实际应用中有效治理的难题，核心问题在于现有安全防护机制存在两大局限：基于学习的防护策略虽具备强语义判别能力，但其策略行为与特定模型及分类体系耦合，难以复用；而可编程框架虽提供灵活控制，却依赖大量手动提示工程与工作流设计，且均未将策略以可复用的运行态形式外部化，导致政策证据在检测、干预与验证环节难以一致共享。为此，本文提出PolicyMem——一种几何化策略记忆系统，将自然语言策略转化为共享表示空间中的低秩子空间所表征的几何记忆对象，实现策略的外部化存储。通过记忆写入器将自然语言策略编码为记忆槽，查询-响应对通过投影能量读取策略记忆，生成直接用于安全判定的策略证据谱图，并在策略归因与干预后验证中复用该证据。结合响应重写模块，PolicyMem构建了“检测-重写-验证”闭环，显著提升治理效率与一致性。在五个主流基准测试中，PolicyMem不仅实现了当前最优的有害行为检测性能，还支持有效的策略归因、内容重写与干预后验证，展现出卓越的可复用性与可解释性。
链接: https://arxiv.org/abs/2609.13734
作者: Yuanchen Bei,Zhengzhang Chen,Yanjun Zhao,Haoyu Wang,Hanghang Tong,Haifeng Chen
机构: University of Illinois Urbana-Champaign; NEC Laboratories America
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 19 pages, 10 figures
Abstract:As large language models (LLMs) are increasingly deployed in real-world high-stakes applications, effective governance has become essential. Existing safeguards largely follow two paradigms: learning-based guards provide strong semantic discrimination but couple policy behavior to trained models and taxonomies, while programmable frameworks offer flexible control but require substantial manual prompt and workflow engineering. Neither externalizes policies as reusable operational states, making it difficult to consistently reuse policy evidence across detection, intervention, and verification. In this paper, we introduce PolicyMem, a geometric policy memory that externalizes natural-language policies as reusable geometric memory objects represented by low-rank subspaces in a shared representation space. A memory writer compiles natural-language policies into policy memory slots, and query-response pairs read the policy memory through projection energy. The resulting policy-evidence profile directly mediates the safety verdict and is reused for policy attribution and post-intervention verification. Coupled with a response rewriter, PolicyMem enables a detect-rewrite-verify loop for LLM governance. Across five widely used benchmarks, PolicyMem achieves state-of-the-art unsafe behavior detection while enabling effective policy attribution, rewriting, and post-intervention verification through the shared policy memory.
[NLP-142] Scaling Hindi Quantum Natural Language Processing through Automatic Pregroup Supertagging
【速读】： 该论文旨在解决量子自然语言处理（Quantum Natural Language Processing, QNLP）中印度语（以印地语为例）预组语法（pregroup grammar）类型标注依赖人工标注、难以规模化的问题。其核心解决方案是将印地语预组语法的超标签（supertagging）任务形式化为一个词元级别的分类问题，并探索多种自动化方法在低资源条件下的有效性。关键创新在于引入符号化语法知识对大语言模型（LLM）生成结果进行约束，通过“词汇修复”（lexical repair）策略显著提升预测准确率至64.08%，优于直接提示（prompting）方法（仅11.65%），验证了结合符号规则与神经模型的协同优势。诊断分析进一步揭示：已见且语义明确的词元易处理，而未见词元仍是主要挑战；形态学后缀特征虽能提升“卡拉卡”（karaka）标签精度，但对整体性能提升有限。研究结果表明，在低资源场景下，自动化的印地语预组语法类型分配具有可行性，可有效降低未来多语言QNLP系统对人工标注的依赖。
链接: https://arxiv.org/abs/2609.13721
作者: Gautami Sanjay Naik,Krishna Bhatia,Mithun Paul Saint-Germain,H Aswath Babu
机构: Indian Institute of Information Technology Dharwad(印度信息科技学院达瓦德分校); Fractal Analytics; Arizona State University(亚利桑那州立大学)
类目: Computation and Language (cs.CL); Quantum Physics (quant-ph)
备注: 8 pages, 1 figure, 2 tables
Abstract:Quantum Natural Language Processing (QNLP) uses pregroup grammars to translate grammatical structure into diagrammatic representations and quantum circuits. Recent Hindi QNLP work has shown that Hindi-specific pregroup grammars can support grammar-sensitive compositional models, but grammatical type assignment is still largely manual, limiting scalability. This paper formulates automatic Hindi pregroup supertagging as a token-level classification task. Using a manually annotated corpus of 380 Hindi sentences, we evaluate lexical, contextual, prompting-based, lexical-repair, and suffix/morphology-aware methods. Results show that simple lexical and contextual models are strong in this low-resource setting: contextual backoff achieves the best completed accuracy of 64.56%, while raw Qwen2.5 prompting reaches only 11.65%. Lexical repair raises LLM-assisted prediction to 64.08%, demonstrating the value of constraining generative outputs with symbolic grammar knowledge. Diagnostic analysis further shows that seen and unambiguous tokens are much easier than unseen tokens, and suffix/morphology features improve karaka-token accuracy but not overall performance. These results show that automatic Hindi pregroup assignment is feasible and can reduce reliance on manual annotation in future multilingual QNLP pipelines.
[NLP-143] When Edit Localization Amplifies Relative Selection Bias: Gradient Geometry Target Mismatch and Importance Weighting
【速读】： 该论文旨在解决人类修正（human corrections）在模型微调过程中因反馈渠道选择性导致的偏差问题，即修正样本可能仅来自特定反馈路径，从而引入选择性偏差（selection bias）。其核心挑战在于：如何在局部化（localization）修正与完全保留（full retention）之间权衡，以准确估计真实梯度方向并最小化误差。解决方案的关键在于提出一种基于梯度分解的分析框架，将局部梯度分解为被编辑（edited）和未被修改（retained untouched）两部分，并引入“正交重要性加权”（oracle importance weighting）方法，在固定局部目标下恢复总体均值。研究进一步推导出有限样本下的均方误差表达式、最优保留系数（optimal retention coefficient）以及固定截断扩展（fixed-clipping extension），并通过精确的有限总体计算与大规模蒙特卡洛模拟验证了理论结果。实验采用双语翻译任务（英德方向）及预训练模型，揭示硬局部化虽具有更高相对偏差但更低绝对偏差，且未被修改成分的偏差非零，被选中成分的均值与未被修改成分存在负内积关系，表明简单无偏/对齐准则失效。最终证据分离了相对放大效应与绝对梯度误差，确立了估计性质，但未涉及翻译质量提升或用户投诉倾向的推断。
链接: https://arxiv.org/abs/2609.13709
作者: Shengwei Zhang,Haoda Dai,Yifei Li,Yuheng Song
机构: University of Pennsylvania(宾夕法尼亚大学)
类目: Computation and Language (cs.CL)
备注: 21 pages, 5 figures
Abstract:Human corrections identify editable spans, but the examples receiving corrections may come from a selective feedback channel. We analyze this interaction at a fixed model checkpoint by decomposing a localized gradient into edited and retained untouched components. Squared relative selection bias is a ratio of quadratics whose derivative has the sign of an explicit quadratic polynomial. Localization can increase, decrease, or nonmonotonically change this diagnostic; its direction depends on component biases and geometry. Oracle importance weighting recovers the population mean for each fixed localization objective, but these objectives have different targets. Against one common full-gradient target, we derive the finite-sample mean-squared error, an analytic optimal retention coefficient, and a fixed-clipping extension. Exact finite-population calculations and 10,000 Monte Carlo repetitions per sample size verify the identities and counterexamples. Public human-post-edit experiments use two translation directions and pretrained models, with declared synthetic selection. An English-German extension differentiates 73.89 million native parameters. In all three declared settings, hard localization has higher relative bias but lower absolute bias than full retention. Untouched-component biases are nonzero and selected component means have negative inner products, so the general criterion applies where the simple unbiased/aligned explanation fails. Output-bias diagnostics show the same endpoint ordering of relative bias across both directions, with one interior maximum. Mechanisms reuse each language’s records and include a mixture; they are not independent replications. The evidence separates relative amplification from absolute gradient error and establishes estimation properties, without inferring translation-quality gains or identifying actual complaint propensities.
[NLP-144] Prefix Sharing Is a Sorting Problem
【速读】： 该论文旨在解决大语言模型（LLM）服务中因提示（prompt）组装顺序不当而导致的键值缓存（KV cache）重用效率低下问题。现有系统通过固定全局顺序来组织可复用的提示片段（如检索到的段落、工具定义、少样本示例），但这种做法仅在请求包含至多两个片段时为最优，而在一般情况下渐近地次优。其核心解决方案是提出一个结构定理：最小前缀-前缀树（prefix-trie）代价等于在请求集合上所有二叉层次结构 $ H $ 中，对每个请求 $ x $ 的权重 $ w(x) $ 与其所需块 $ x $ 的规范分解大小 $ t_x(H) $ 的乘积求和的最小值。因此，片段排序等价于选择一个请求上的二叉层次结构。该理论揭示了两片段情形下最优解对应于最小顶点覆盖问题，并指出在“留一法”（leave-one-out）家族中，最优解即为二叉树的最小外部路径长度——对应于归并排序的递归结构。这表明全局固定顺序的代价为 $ \Theta(n^2) $，而最优解仅为 $ \Theta(n \log n) $。进一步提出基于公共交集的凝聚聚类方法，作为可达节省量的紧致 $ \frac{1}{2} $-近似算法。在三个 BEIR 基准数据集上的 BM25 检索轨迹实验表明，该方法相比生产环境中的 RAG 排序可减少 17%–36% 的预填充计算量，且随着检索深度增加，优势愈发显著。最终，按照层次结构的深度优先搜索（DFS）顺序服务请求，使得仅缓存单个请求上下文即可达到无界缓存下的最优性能，从而证明缓存容量与重排窗口之间具有可互换性。
链接: https://arxiv.org/abs/2609.13692
作者: Rong He
机构: 未知
类目: Data Structures and Algorithms (cs.DS); Computation and Language (cs.CL)
备注: 15 pages. Code: this https URL
Abstract:LLM serving reuses KV cache by exact prefix match, so when a prompt is assembled from a set of reusable pieces – retrieved passages, tool definitions, few-shot exemplars – the order chosen for those pieces determines how much computation can be shared. Every deployed system fixes that order by a single global convention. We prove this is optimal only when requests contain at most two pieces, and asymptotically wrong in general. Our main result is a structure theorem: the minimum prefix-trie cost equals min_H sum_x w(x) t_x(H) over binary hierarchies H on the requests, where t_x(H) is the canonical decomposition size of the set of requests needing chunk x. Choosing chunk orders is therefore equivalent to choosing one hierarchy over requests. The identity yields an O(3^m) exact algorithm, identifies the two-chunk case as minimum vertex cover, and shows that on the leave-one-out family the optimum is the minimum external path length of a binary tree – the merge-sort recursion – so a global order pays Theta(n^2) against a true cost of Theta(n log n). Agglomerative clustering by common intersection is a tight 1/2-approximation for the achievable saving. On BM25 retrieval traces over three BEIR corpora the resulting layout reduces prefill by 17-36% against production RAG ordering, and the margin widens with retrieval depth as the theory predicts. Serving requests in the hierarchy’s DFS order finally lets a cache holding one request’s context attain the unbounded-cache optimum exactly, so cache capacity and reorder window act as substitutes.
[NLP-145] Not all Negation Cues are Equal: Affixal Negations Yield Better Negation Understanding EMNLP2026
【速读】： 该论文旨在解决语言模型（LMs）和大语言模型（LLMs）在处理否定表达时长期存在的挑战，尤其针对现有研究局限于高频单字否定线索（如not、never）且对更广泛否定类型及现代大语言模型关注不足的问题。其解决方案的关键在于构建了一个大规模的否定线索数据集NegCue，涵盖超过180万条样本，覆盖单字、多字及词缀式否定，并包含200多种独特否定线索。在此基础上，研究对编码器-仅模型（encoder-only LMs）和大语言模型（LLMs）进行基于NegCue的进一步预训练，系统探究不同否定类型对否定理解能力的影响。实验结果表明，在相同训练规模下，各类否定类型对下游任务性能的提升作用不均，其中词缀式否定带来的性能增益最为显著，而传统研究较多的单字否定所带来的改进仍较有限。此外，研究还验证了进一步预训练能有效提升模型对否定的理解能力，无论对于传统语言模型还是大语言模型。
链接: https://arxiv.org/abs/2609.13685
作者: Tian Tan,Eduardo Blanco
机构: University of Arizona(亚利桑那大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to the EMNLP 2026 Main Conference
Abstract:Negation remains a longstanding challenge for both language models (LMs) and large language models (LLMs). Prior work mainly focuses on a small set of high-frequency single-word negation cues, such as not and never, with limited exploration of broader negation types and modern LLMs. To address this gap, we construct NegCue, a large-scale dataset containing over 1.8M samples spanning single-word, multi-word, and affixal negation with more than 200 unique cues. We further pre-train both encoder-only LMs and LLMs on NegCue to investigate how different negation types affect negation understanding. Experiments on five downstream benchmarks show that negation types contribute unevenly to performance gains under the same training scale. In particular, affixal negation yields the largest improvements, while the gains from the commonly studied single-word negation remain modest. Moreover, our results demonstrate that further pre-training improves negation understanding for both LMs and LLMs.
[NLP-146] LayerRoute: Adaptive Layer-Skipping with LoRA-Preserved Quality for Efficient LLM Inference
【速读】： 该论文旨在解决大模型推理过程中计算资源浪费与效率瓶颈的问题，提出一种参数高效且可自适应的变压器层跳过（transformer layer-skipping）方法。其核心挑战在于如何在不显著降低模型性能的前提下，实现动态、输入依赖的层跳过机制，以提升推理速度并减少计算开销。解决方案的关键在于提出LayerRoute框架，该框架结合了逐层硬门控路由（per-layer hard-gated routing） 与联合低秩适配器（LoRA）微调，通过直通估计器（straight-through estimator）训练门控机制，并在门控正则化的语言建模目标下联合优化路由策略与LoRA参数。该方法在Qwen2.5-0.5B-Instruct模型上仅引入约21.5K参数的轻量级逐层路由器和约1.08M参数的低秩适配器，在10次独立种子训练中均收敛至一致的9层跳过模式（第8至16层），实现了稳定且可验证的时钟时间加速（平均1.04倍，范围1.02x–1.06x），同时保持或提升了模型质量（困惑度降低平均达-1.16）。进一步验证表明，路由决策对输入具有实质性影响，87%–100%的样本因门控选择而实际改变执行路径，证明其为非平凡的输入依赖路由而非固定剪枝。该方法在单个A100 GPU上训练耗时不足7分钟，额外开销仅限于路由判断本身，具备高度可复现性，并附有系统性诊断分析以揭示路由决策的驱动因素。
链接: https://arxiv.org/abs/2609.13682
作者: Prateek Kumar Sikdar
机构: Accenture(埃森哲)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:We introduce LayerRoute, a parameter-efficient method for adaptive transformer layer-skipping that combines per-layer hard-gated routing (trained via a straight-through estimator) with joint LoRA fine-tuning. LayerRoute augments each of the 24 transformer blocks in Qwen2.5-0.5B-Instruct with a lightweight per-layer router (~21.5K parameters) and LoRA adapters (rank 8, ~1.08M parameters), training both jointly under a gate-regularized language-modeling objective. Across 10 independently-seeded training runs, LayerRoute converges to an identical skip-pattern structure in every run - a consistent set of 9 middle layers (8-16) becomes skip-eligible in all 10 seeds - and delivers genuine, verified wallclock speedup in every run (1.02x-1.06x, mean 1.04x). Quality is preserved or improved in every configuration tested: joint LoRA adaptation yields a perplexity improvement over the unmodified backbone in all 10 seeds (mean delta = -1.16 and -1.11 across the two evaluation splits used). We further verify the router performs genuine, non-trivial per-input computation: gate decisions in skip-eligible layers change the actual skip/run outcome for 87-100% of held-out samples, confirming real input-dependent routing rather than a fixed pruning pattern. LayerRoute trains in under 7 minutes on a single A100 and adds negligible overhead beyond the routing decision itself. We report our full reproducibility methodology, including a systematic diagnostic investigation into what determines the router’s per-input decisions, as part of this work.
[NLP-147] An Efficient and Modular Framework for Targeted Harm Mitigation in LLM S
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在零样本学习中容易与人类偏好产生偏差，进而生成带有偏见、毒性或其他有害内容的问题。现有对齐方法虽有效，但存在成本高且与模型高度耦合的缺陷，限制了其灵活性与可扩展性。本文提出一种模块化修正框架，通过在预训练的LLM上引入激活式低秩适配器（Activated LoRA, aLoRA）和上下文感知路由机制，实现对模型输出中危害内容的精准消除。其解决方案的关键在于：aLoRA适配器可在生成过程中的任意时刻被动态激活，无需重置键值缓存（KV cache），从而支持低延迟、细粒度的实时修正；同时，基于模型中间输出自适应选择的专家路由机制，使不同专家适配器能够针对特定类型危害（如偏见或毒性）进行专业化检测与缓解。该方法在保持任务性能的同时显著提升了安全性基准上的对齐效果，为实现更安全、可控的大语言模型部署提供了一种轻量高效的新路径。
链接: https://arxiv.org/abs/2609.13624
作者: Roberto Campbell,Momin Abbass,Muneeza Azmat,Michal Ulewicz,Raya Horesh,Kristjan Greenewald,Rogério Abreu de Paula,Nathalie Baracaldo
机构: 未知
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY)
备注:
Abstract:Large Language Models (LLMs) are powerful zero-shot learners but remain prone to misalignment with human preferences, often producing biased, toxic, or otherwise harmful outputs. Existing alignment methods, while effective, are costly and tightly coupled to the model, limiting flexibility and scalability. We propose a modular correction framework that augments pretrained LLMs with Activated LoRA (aLoRA) adapters and a context-aware routing mechanism to eliminate harms from misaligned model responses. Our approach enables expert adapters to activate mid-sequence without invalidating the KV cache, allowing low-latency, targeted correction during generation. Each expert is trained to detect and mitigate specific harms, such as bias or toxicity. A learned router dynamically selects appropriate experts based on the models intermediate outputs. We demonstrate that our system improves alignment on standard safety benchmarks while preserving task performance, offering a lightweight and efficient path toward safer and more controllable LLM deployments.
[NLP-148] he University of Melbourne WMT 2026 CreoleMT Submission: A Domain-Balanced Approach to Low-Resource Pacific Creole Machine Translation
【速读】： 该论文针对太平洋克里奥尔语（如托克皮钦语、比斯拉马语和所罗门皮金语）在机器翻译（Machine Translation, MT）任务中面临的低资源、多领域性能不均衡问题，提出了一种改进的训练策略。其核心挑战在于：这些语言缺乏大规模、均衡覆盖各领域的标注数据，导致模型在跨域泛化能力上表现不足。解决方案的关键在于采用两阶段训练范式：首先在大规模、领域不平衡的数据上进行预训练，随后在多样且领域均衡的数据集上进行持续微调；同时结合多种数据增强与构建技术，包括基于大语言模型（LLM）辅助的重拼写与对齐、反向翻译（back-translation）以及利用Gemini模型对原训练数据中缺失领域进行知识蒸馏（distillation），从而有效提升模型在未见领域的泛化能力。实验结果表明，在Bouquet基准及新构建的口语转录测试集上，该方法在所有方向上均显著优于开源基线模型，平均提升超过3个chrF++分数，验证了其在广域场景下的有效性。未来工作将聚焦于构建人工翻译的测试集，并进一步将最优模型蒸馏为更小规模但保持广泛领域覆盖的轻量化版本。
链接: https://arxiv.org/abs/2609.13615
作者: Raphaël Merx,Nick Thieberger,Ekaterina Vylomova
机构: The University of Melbourne(墨尔本大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:For our submission to the WMT26 Creole Language Translation Shared Task, we focus on machine translation (MT) models for Pacific creoles: Tok Pisin, Bislama, and Solomon Pijin, with particular attention to broad domain performance. After pre-training on a large collection of domain-imbalanced data, we continue fine-tuning on a diverse mix of domain-balanced data. We rely on a number of data collection and preparation techniques, including LLM-assisted respelling and alignment, back-translation, and distillation from Gemini for domains originally not present in training data. Evaluated on Bouquet and a novel test set made of spoken language transcripts, our models beat open model baselines by 3+ chrF++ points in all directions with human-original references. Looking ahead, we plan to develop human-translated test sets for Solomon Pijin and Bislama, and to distil our best models into much smaller ones that retain broad domain coverage.
[NLP-149] In the Blind: Building Pseudo-References for MT Evaluation EMNLP2026
【速读】： 该论文旨在解决无参考文本（reference-free）条件下机器翻译质量评估与优选的难题，尤其针对在缺乏人工参考译文（包括从头翻译或后编辑生成的参考）的10个语言对上如何有效构建高质量伪参考（pseudo-references）的问题。其核心挑战在于：在无参考环境下，传统的基于参考的评估方法失效，而依赖参考无关质量评估（Reference-Free Quality Estimation, QE）的自动选择机制易因误判流畅但语言错误的输出而导致严重偏差。解决方案的关键在于提出一种多模型协同、基于得分融合的文档级选择框架——通过七种不同模型在最多五种提示（prompt）条件下生成26种系统-提示组合，由三个独立的参考无关QE模型对每条候选译文进行评分，再通过一个基于文档级别的选择器择优输出；若必要则由GPT-5.5进行后编辑优化。为克服QE引导选择中“语言错位”（即高分但非目标语言的译文被优先选中）这一关键失败模式，研究引入了置信度加权的语言识别惩罚项（confidence-scaled language identification penalty），显著降低了错误语言译文的选择率至零，并使最终选择器在MetricX指标上优于传统排序融合基线。由于这些语言对在构建时无可用参考，所有选择决策均基于前一年（WMT25）的人工评判进行校准，事后人类评估揭示了选择失误的代价：当选择器保留前沿模型候选时，其伪参考表现与最强参赛系统相当；而一旦误选，则性能最高下降17 ESA点。研究公开了完整的选择方法及每条伪参考的溯源信息。
链接: https://arxiv.org/abs/2609.13611
作者: Diptesh Kanojia,Chi-kiu Lo,Archchana Sindhujan,Samuel Larkin,Greg Hanneman,Alon Lavie
机构: University of Surrey(萨里大学); National Research Council Canada(加拿大国家研究委员会); Carnegie Mellon University(卡内基梅隆大学)
类目: Computation and Language (cs.CL)
备注: Accepted at Eleventh Conference on Machine Translation (WMT) @ EMNLP 2026
Abstract:The WMT26 General MT task evaluates systems on 10 language pairs that have no human references (neither translated from scratch nor post-edited from MT output by humans). We describe how we built the pseudo-references for these pairs and six other language pairs (in which some forms of human references are available): seven models translate the 3,277 official documents under up to five prompt conditions, giving a total of 26 system-prompt combinations; then three reference-free quality estimation (QE) models score every candidate; and a per-document selector picks one translation, which GPT-5.5 post-edits where needed. Working without references exposed a failure mode of QE-guided selection: the metrics rank fluent output in the wrong language above correct translations. Adding a confidence-scaled language identification penalty to the score fusion drives the wrong-language count to zero, and the resulting selector still scores better on MetricX than the rank-fusion baseline it replaces. Since no references were available for these pairs while we were building them, we calibrate every selection decision on last year’s WMT25 human judgments. The human evaluation, released after construction, shows the cost of getting selection wrong: our references stand with the strongest participating systems when the selector kept a frontier-model candidate, and fall up to 17 ESA points below them when it did not. We release the selection method and the provenance of every reference (this https URL)
[NLP-150] Same Patient Different Order: Action-Level Reliability of Clinical LLM Agents Under Repeated Runs
【速读】： 该论文旨在解决临床智能体（clinical agent）在基准测试中存在行为不一致性的问题，即相同输入条件下，智能体在多次运行中产生显著不同的诊疗决策（如检验申请、药物处方、转诊建议等），而现有基准测试（如MedAgentBench）仅评估单次运行的得分，未能捕捉这种动作层面的稳定性缺陷。其解决方案的关键在于提出“同输入重运行”（same-input rerun）评估范式，通过固定所有输入条件并重复执行任务，系统性地比较不同运行间的具体医嘱行为差异，而非仅关注最终评分结果。研究采用六种可靠性度量指标，在50个任务、1000次运行中对两个低于十亿参数的开源量化模型（4B与8B，四比特量化）及两种温度设置进行验证，发现：在8B模型于温度0.7下，43个排序组中有全部出现动作差异，其中26组在部分运行中发出指令而在其他运行中未发出，28组记录了不同的编码值、剂量或检测项目；更严重的是，22个组的行为差异导致实际医疗执行路径不同，甚至出现系统拒绝执行但代理误报成功的情况。这些发现表明，传统基于单一运行评分的评估无法识别此类关键的稳定性问题，因此论文主张引入重复运行评估、动作级稳定性报告以及执行忠实的环境反馈机制，以提升临床智能体基准测试的可信度与安全性。
链接: https://arxiv.org/abs/2609.13582
作者: Rohith Reddy Bellibatlu,Manpreet Singh,Zhoutian Han,Wenbin Zhang
机构: Florida International University (佛罗里达国际大学); Boston University (波士顿大学); Stevens Institute of Technology (史蒂文斯理工学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Software Engineering (cs.SE)
备注: 13 pages, 2 figures, 3 tables. Code, run records and the rerun harness: this https URL ; github: this https URL
Abstract:A clinical agent benchmark can report the same verdict on identical inputs while the agent files a materially different order on each run. Such agents order tests, request medications and place referrals, yet benchmarks typically score one run per task and rarely ask whether identical inputs produce identical actions; MedAgentBench, the benchmark we use, scores a single attempt and says so. To measure this gap we introduce “same-input rerun”, which replays a task with every input held fixed and compares the orders rather than the score, with six reliability metrics, and apply it to 1000 MedAgentBench runs across 50 tasks from its five write-capable families, two open-weight models below ten billion parameters quantised to four bits, and two temperatures. The study establishes that action-level divergence exists and can pass unrecorded by the score, not that any rate generalises. Under the 8B model at temperature 0.7, all 43 ordering groups emit a different set of orders across five identical runs, 26 emit the order on some runs and not others, and 28 record a different coded value, dose or analyte. In 22 of those 43 the benchmark reports the same failing verdict for materially different behaviour, as it does for all 10 divergent groups of the 4B model at 0.7. Orders also reach different endpoints across runs, one of which the record server rejects while the agent is told it succeeded. These findings motivate repeated-run evaluation, action-level stability reporting and execution-faithful environment feedback in clinical-agent benchmarks.
[NLP-151] oward Complete Hospital Discharge Summarization with Abstract Meaning Representation ALT
【速读】： 该论文旨在解决临床病程总结（discharge summary）自动生成中的幻觉问题，即大型语言模型（Large Language Model, LLM）在生成医疗文本时可能产生与原始临床记录不符的虚假信息，从而对临床文档的准确性与安全性造成严重威胁。其解决方案的关键在于提出一种基于证据驱动的对齐框架，将文献来源（provenance）作为首要约束条件，通过语义图（semantic graph）与深度学习模型实现跨文档的语义对齐，确保生成的每一条总结句均能追溯至原始临床记录中的具体语义片段，并附带明确的证据链接。该方法不仅提升了生成内容的可解释性与可信度，还在MIMIC-III公开数据集及伊利诺伊大学医院（UIC Health）的真实临床笔记上验证了有效性，同时开源了源代码与训练模型以促进后续研究。
链接: https://arxiv.org/abs/2609.13581
作者: Paul Landes,Sitara Rao,Aaron Jeremy Chaise,Barbara Di Eugenio
机构: University of Illinois Chicago (伊利诺伊大学芝加哥分校)
类目: Computation and Language (cs.CL)
备注: Published in the 2026 IEEE 14th International Conference on Healthcare Informatics (ICHI). 10 pages, 6 figures
Abstract:Discharge summaries are lengthy medical documents that summarize a hospital in-patient visit. Automatically generating them can reduce documentation burden and return clinician time to patient care. Whereas Large Language Model (LLMs) could be used for this task, their Achilles heel is hallucinations, which can have drastic consequences for clinical documentation. We present an evidence-driven alignment framework for discharge summarization at the clinical encounter level, that treats provenance as a first-class constraint, using semantic graphs and deep learning models. Each summary sentence is selected and organized via cross-document semantic alignment and is accompanied by explicit evidence links to its source spans. We show our results on two corpora: a publicly available corpus (MIMIC-III) and clinical notes written by physicians at the University of Illinois Hospital (UIC Health). Additionally, we make source code and trained models available.
[NLP-152] Domain-Specific Jargon in Large Language Models : A Comparative Analysis between General-Purpose and Specialist Models EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在高度专业化技术领域（如医学）中性能下降的问题，以及缺乏对领域特定术语的参数化知识如何编码于模型内部的理解。其关键解决方案在于构建两个新颖的医学专有名词评估基准，并通过机制可解释性工具分析通用型Llama-3.1模型与其在医学领域微调后的变体之间的表现差异。研究发现，尽管进行了领域微调，医学专用模型反而在两项任务中表现逊于通用模型，且存在系统性校准偏差；进一步分析表明，微调模型并未重构参数知识结构，而是过度依赖少数与专有名词预测相关的模型组件。通过施加组件重加权策略，可有效抑制这些异常激活的组件，从而缩小与通用模型的性能差距。此外，部分敏感于专有名词的组件在材料科学领域也表现出泛化能力，提示其可能编码了一种部分非领域特异性的“专业术语”表征。本研究揭示了领域适应并不必然提升对专业术语的理解，挑战了“微调即优化”的默认假设，强调需谨慎评估领域适配的实际效果。
链接: https://arxiv.org/abs/2609.13556
作者: Darin Keng,Zhewei Sun
机构: University of Chicago (芝加哥大学); Toyota Technological Institute at Chicago (丰田技术学院芝加哥分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to EMNLP 2026 main conference
Abstract:Large Language Models (LLMs) have shown remarkable proficiency on general-purpose tasks, yet their performance often degrades in highly-specialized technical domains. Moreover, little is known about how parametric knowledge of domain-specific terms is encoded within these models. We address this gap by contributing two novel medical jargon evaluation benchmarks and evaluate a general-purpose Llama-3.1 model against a variant fine-tuned on medical-domain data. Surprisingly, the general-purpose model outperforms the medically fine-tuned model on both tasks. Using mechanistic interpretability tools, we find systematic patterns of miscalibration for the medically fine-tuned model. Instead of reorganizing parametric knowledge, the fine-tuned model places greater emphasis on a small subset of model components associated with jargon-favoring predictions. We find that applying component reweighting strategies against the benchmark tasks successfully suppresses these components and closes the gap with the general-purpose baseline. We also observe that some jargon-sensitive components transfer knowledge to the same tasks involving materials science jargon, suggesting they encode a partially domain-agnostic notion of specialized terminology. Our results provide a case study in which a medically fine-tuned checkpoint does not improve jargon comprehension over its general-purpose counterpart, highlighting that domain adaptation should not be assumed to yield better performance on specialized terminology.
[NLP-153] Harmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language Models ICML
【速读】： 该论文旨在解决生成式 AI（Generative AI）中恶意提示（harmful prompts）的检测难题，特别是如何有效识别那些经过精心设计以规避安全防护机制的对抗性越狱攻击（adversarial jailbreaks）。其核心挑战在于，传统基于单层隐藏状态的检测方法难以捕捉恶意意图在模型深层逐步显现的动态演化过程。为此，论文提出**有害性传播动态（Harmfulness Propagation Dynamics, HPD）**这一关键发现：对于恶意提示，其最后一层隐藏状态在预训练学习到的“有害方向”上的投影值随Transformer层数递增呈现单调上升趋势，而良性提示则保持平坦或振荡，形成具有显著区分度的跨层轨迹特征。该动态特性揭示了有害意图作为渐进式解析的语义属性——表面形式早期显现，而深层语用意图在模型深度中逐步固化。基于此，作者提出轻量级输入监管器 \herald，通过提取跨层投影序列中的七个维度特征（斜率、曲率、单调性、触发层及统计量），并利用仅288参数的MLP进行分类。该方法无需反向传播计算，推理时仅增加约2.6×10⁻⁶预填充浮点运算量，且每层存储一个固定方向（32层、d=4096模型仅需262 KB），具备极低开销与高可复现性（各层LDA学习的有害方向在随机划分下余弦相似度达0.97）。在八个提示有害性基准和四个模型家族上，\herald在OLMo2-7B上平均F1达到89.3，在对抗性越狱检测中表现优于所有对比守护模型（98.4 vs. 96.9 F1），且在所有主干模型上均比先前基于潜在空间的方法提升2.3–4.1 F1点。更重要的是，每个实例的跨层轨迹提供了机器可读的审计日志，明确揭示有害性“何时”与“如何”在模型内部涌现，显著提升了检测过程的可解释性。
链接: https://arxiv.org/abs/2609.13534
作者: Noor Islam S. Mohammad,Uluğ Bayazıt
机构: İTÜ (伊斯坦布尔技术大学); İstanbul (伊斯坦布尔), Türkiye (土耳其)
类目: Computation and Language (cs.CL)
备注: International Conference on Machine Learning (ICML) Workshop on Mechanistic Interpretability in 2026, accepted in South Korea. this https URL
Abstract:We identify \textbfHarmfulness Propagation Dynamics (HPD): for harmful prompts, the projection of the last-token hidden state onto a learned harm direction rises monotonically with transformer depth, whereas benign prompts remain flat or oscillatory. This cross-layer signature reflects harmful intent as a \emphprogressively resolved semantic property: surface form appears early, while pragmatic intent consolidates later, making the \emphtrajectory shape more informative than any single-layer snapshot. Moreover, LDA-based harm directions, learned per layer, remain stable across random splits (pairwise cosine similarity 0.97 ), supporting the projection sequence as a reproducible structured signal. Building on HPD, we introduce \textbf\herald (\textbfHarmful \textbfEncoding \textbfRecognition via \textbfActivation \textbfLayer \textbfDynamics). This lightweight input moderator extracts a seven-dimensional feature record, slope, curvature, monotonicity, onset layer, and related statistics from the cross-layer projection sequence and classifies it with a 288-parameter MLP. \herald stores one d -dimensional direction per layer ( 262 ,KB for a 32-layer, d=4096 model), requires no gradient computation during training, and adds only 2.6\times10^-6 prefill FLOPs at inference. Across eight prompt-harmfulness benchmarks and four model families, \herald achieves an average F1 of 89.3 on OLMo2-7B, surpassing all tested guard models on adversarial jailbreak detection ( 98.4 vs.\ 96.9 F1) and outperforming prior latent-based methods by 2.3 - 4.1 F1 points on every backbone. Per-instance trajectories provide machine-readable audit records that reveal \emphwhen and \emphhow harmfulness emerges, offering an interpretability advantage over single-layer approaches.
[NLP-154] Generative Interpretability via Scalable Neuro-Symbolic Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）从对话系统向自主智能体（agentic systems）演进过程中所面临的可解释性（interpretability）挑战。随着模型输出直接转化为对现实世界具有不可逆影响的行动，传统事后可解释性（post-hoc interpretability）范式已显不足——其仅能在行为发生后提供解释，无法在推理过程中进行审计或干预。为此，论文提出转向“生成式可解释性”（generative interpretability）这一新的架构属性，即模型在推理过程中原生地生成语义上可理解且可被因果干预的中间状态节点。该解决方案的关键在于将可解释性嵌入模型的内部计算结构之中，使人类能够实时理解、监控并干预模型决策路径，从而实现更安全、可信的部署。论文进一步以神经符号模型（Neuro-Symbolic Models）作为生成式可解释性的具体实现范例，验证了其相较于传统可解释性方法在透明性与可控性方面的优势。
链接: https://arxiv.org/abs/2609.13529
作者: Xiaocong Yang
机构: AI Interpretability @ Illinois, University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Symbolic Computation (cs.SC)
备注: ACM AI Summit 2026
Abstract:As the use of Large Language Models moves from chatbots into agentic systems, where outputs become actions with irreversible consequences on reality, the existing paradigm on AI Interpretability research, post-hoc interpretability, is structurally inadequate for safe and trustworthy model deployment: it explains behavior after the fact but cannot audit or intervene in an inference computation before it commits to an output. We therefore argue for a shift toward \emphgenerative interpretability, an architectural property under which a model’s inference pass natively exposes semantically meaningful checkpoints that are human-understandable and amenable to causal intervention. We show the merits of generative interpretability as comparison to other interpretability research paradigms, and propose Neuro-Symbolic Models as a concrete instantiation.
[NLP-155] From Token Probabilities to Semantic Constraints: Towards Declarative Probabilistic Evaluation of Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在预训练过程中知识与推理能力的评估问题，特别是如何将评估结果与预训练阶段的学习信号建立语义关联。现有评估方法多依赖于词元似然或答案准确率，难以捕捉模型在否定、互斥性及一致性等深层语义逻辑上的系统性缺陷。为此，论文提出ModelLog——一种声明式概率框架，通过符号化约束显式建模模型行为的语义结构，将评估目标定义为对词元级预测的逻辑约束，并量化模型输出分布满足这些约束的程度。其核心创新在于：评估分数可被解释为损失函数，其梯度反映了逻辑强度、信息量及变量敏感性，从而在评估与学习之间建立共享语义桥梁。该方法不仅能够诊断模型行为中的深层次逻辑缺陷，还为揭示预训练学习过程的语义结构提供了新的形式化工具。
链接: https://arxiv.org/abs/2609.13520
作者: Kyle Richardson,Cullen Anderson,Pranav Balakrishnan,Takuto Ban,Daksha Ladia,Ankita Gupta,Marisa Hudspeth
机构: Allen Institute for AI(艾伦人工智能研究所); University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:While Large Language Models have improved rapidly, many fundamental questions remain about how to evaluate the knowledge and reasoning abilities they acquire, and how such evaluations relate to the learning signals used in pre-training. In this paper, we propose ModelLog, a declarative probabilistic framework for pre-training evaluation that makes the semantic structure of model behavior explicit and provides new formal tools for relating evaluation to learning. ModelLog specifies evaluation targets as symbolic constraints over token-level predictions and measures how strongly a model’s distribution satisfies those constraints. We explore the framework through a new suite of tasks targeting negation, mutual exclusivity, and consistency, finding systematic failures that are difficult to characterize through token likelihood or answer accuracy alone. We further show that these evaluation scores can also be interpreted as losses, whose gradients reflect logical strength, informativeness, and variable-level sensitivity. This links evaluation and learning through a shared semantics, suggesting evaluation methods that diagnose model behavior while also helping to clarify the semantic structure of learning.
[NLP-156] One Spectrum Two Resources: Data-Memory Scaling in Autoregressive Prediction
【速读】： 该论文旨在解决在生成式建模中，学习到的内存（memory）与可用数据量之间的权衡问题，即在给定有限存储资源（以比特数 2B 为上限）时，需要多少有效记忆才能从增加的数据中获得显著性能提升。其核心问题是：在正熵自回归检索源（positive-entropy autoregressive retrieval source）框架下，如何量化数据增益与内存容量之间的最优关系。解决方案的关键在于提出一个预测能量谱（predictive-energy spectrum） μ，该谱描述了每个坐标对预测误差的贡献，其形式为查询概率乘以未知logit的平方半径。基于此能量谱，论文建立了极小极大（minimax）泛化误差的渐近表达式：Rvalue∗(n,B)≍RΦμ(n−1)+Φμ(τB)，其中 Φμ(t)=∫min{x,t}μ(dx)，n 为预测块数，τB 为最优比特分配所达到的水平。该理论揭示了数据分辨率（1/n）与内存层级（τB）的分离机制：数据决定精度分辨率，内存决定最优信息压缩水平。关键洞见在于能量-维度配对（energy-dimension pairing） 的不可约性——即使两个因果源具有相同的块能量和块维度边际分布，其数据与内存的收敛指数仍可能不同，表明仅考虑统计矩不足以刻画复杂模型的效率边界。通过引入掩码查询-键注意力头（masked query-key attention head），论文实现了显式路由、格式与算术误差下的理论律动验证，并进一步推导出自适应指数分配、有限精度实现以及双侧算术假设下的计算-精度律。实验部分成功复现了数据-内存坍缩（data-memory collapse）与耦合指数，并解释了路由与分配机制，同时在六个预训练模型尺度上验证了仅权重量化（weight-only quantization）的有效性，验证了理论框架的普适性与实用性。
链接: https://arxiv.org/abs/2609.13500
作者: Chiwun Yang,Xiaoyu Li
机构: City University of Hong Kong (香港城市大学); University of New South Wales (新南威尔士大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 66 pages, 11 figures, 8 tables
Abstract:How much learned memory is needed to benefit from more data? We show that the two resources are governed by one predictive-energy spectrum in a positive-entropy autoregressive retrieval source. Each coordinate contributes its query probability times the squared radius of its unknown logit. Writing \mu for the resulting energy spectrum, we prove the minimax law \mathfrak R^*\rm value(n,B)\asymp_R \Phi\mu(n^-1)+\Phi_\mu(\tau_B), \Phi_\mu(t)=\int\min\x,t\,\mu(\mathrm dx), for n prediction blocks and a learned state with at most 2^B values. Data set the resolution 1/n ; memory sets the level \tau_B reached by optimal bit allocation. The complete curve also recovers the positive spectrum. Energy-dimension pairing is essential: two causal sources with identical block-energy and block-dimension marginals have different data and memory exponents. A masked query-key attention head learns the route and values, realizing the law with explicit routing, format, and arithmetic errors. Further results give exponent-adaptive allocation, finite-precision realization, and compute-precision laws under two-sided arithmetic assumptions. Experiments recover the data-memory collapse and coupling exponents, explain the routing and allocation mechanisms, and examine weight-only quantization across six pretrained-model scales.
[NLP-157] A Hybrid Hierarchical 1D-CNN-BiLSTM Framework for Extractive Summarization of Biomedical and Clinical Text
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在抽象式摘要生成中出现的事实幻觉（hallucination）问题，尤其在生物医学与临床领域，虚假信息可能带来严重风险。其核心解决方案是摒弃生成式流程，将摘要任务重构为抽取式句子选择（extractive sentence selection），通过直接从原文中选取句子来确保事实准确性。关键创新在于提出一种混合分层卷积神经网络-长短期记忆网络（Hybrid Hierarchical CNN-LSTM Summarizer）架构：首先利用堆叠的多核卷积层构建句子级嵌入并生成更丰富的句间关系表示，再通过双向长短期记忆网络（Bidirectional LSTM）捕捉文档级别的长程依赖；随后设计轻量级评分头，基于二元交叉熵损失端到端训练，为每个句子分配重要性得分。推理阶段采用动态均值加标准差阈值结合前3名候选的回退机制，直接从源文本中选取句子并按时间顺序重组为最终摘要。由于所有输出句均源自原文，有效避免了生成过程引发的事实偏差。实验表明，在PubMed数据集上，该模型优于独立的CNN与LSTM基线，且更宽的卷积感受野有助于提升句子评分性能；在MIMIC-CXR和MIMIC-IV BHC数据集上，模型对非结构化叙事表现良好，但在高度模板化的报告中则趋于依赖位置基准。研究结果表明，通过结构化约束设计可实现事实可信的摘要系统，具备“设计即可信”的潜力。
链接: https://arxiv.org/abs/2609.13481
作者: Saad Bin Ather,Muhammad Saif,Ali Hassan Khan,Manzer Abbas,Hajra Waheed
机构: NUCES-FAST (国家工程与计算机科学大学)
类目: Computation and Language (cs.CL)
备注: 6 pages, 2 figures
Abstract:Large language models have made abstractive summarization remarkably fluent, but generated summaries can hallucinate facts, posing serious risks in biomedical and clinical domains. We address this by removing generation from the pipeline and framing summarization as extractive sentence selection. Our Hybrid Hierarchical CNN-LSTM Summarizer uses stacked multi-kernel convolutions to compose sentence-level embeddings into richer inter-sentence representations, followed by a bidirectional LSTM to model long-range dependencies across the document. A lightweight scoring head assigns per-sentence importance scores and is trained end-to-end with binary cross-entropy against oracle extractive labels. At inference, a dynamic mean-plus-standard-deviation threshold with a top-3 fallback selects sentences directly from the source and chronologically reorders them into the final summary. Since every output sentence is copied from the input, the model avoids generation-induced factual drift. On PubMed, our architecture outperforms isolated CNN and LSTM baselines, while ablations show that wider convolutional receptive fields improve sentence scoring. On MIMIC-CXR and MIMIC-IV BHC, the model performs well on unstructured narratives but defaults toward positional baselines on highly templated reports. These results suggest that structural constraints can provide a reliable path toward factually grounded summarization systems that are trustworthy by design rather than by correction.
[NLP-158] STAGE: Diagnosing Semantic Transfer at Grounded Execution in Embodied Agents EMNLP2026
【速读】： 该论文旨在解决具身语言接地（embodied language grounding）中存在的一种语义-动作鸿沟（semantic-action gap）问题，即尽管指令的语义能够被有效恢复，但这些语义在原始连续动作空间中的表达却极为微弱，导致智能体无法根据理解的语义正确执行相应动作。其核心解决方案在于提出一种轻量级的运行时接口VISA，通过将恢复的语义转化为ALLOW、DEFER、目标一致性（target-consistency）及验证选择（verified-selection）等决策，实现对底层策略的动作干预与修正。该方法显著降低了无效指令下的盲目执行率（从92.7%降至2.8%），同时保留了94.0%的正常指令执行率，并通过验证选择进一步提升了动作输出的目标一致性，而无需更新原有策略。研究强调，具身语言评估应关注语义到动作的跨域迁移能力，而非仅依赖语义解析性能。
链接: https://arxiv.org/abs/2609.13458
作者: Baosheng Jin,Yushen Liang,Hua Shen
机构: NYU Shanghai(纽约大学上海分校); Shanghai, China(中国上海)
类目: Robotics (cs.RO); Computation and Language (cs.CL)
备注: Accepted to Findings of the Association for Computational Linguistics: EMNLP 2026. 24 pages, 7 figures
Abstract:Embodied language grounding requires more than identifying the referent of an instruction: recovered semantics must also control the action an agent exposes. We study this missing link as a semantic-action gap, where instruction semantics are recoverable but weakly expressed in native continuous actions. We introduce SAT-Bench, a fixed-observation counterfactual benchmark that holds the visual scene and agent state fixed while changing only instruction semantics. On LIBERO target-name and pixel-grounded relation swaps, target recovery reaches 100.0% and 95.8%, whereas OpenVLA action sensitivity remains only 6.8% and 7.7%. The gap persists across 1,000 additional compositional and temporal/procedural counterfactuals, with overall action sensitivity of 6.1%. Hidden-state, threshold-free, cross-policy, and rollout diagnostics further support this semantic-action transfer failure. We introduce VISA, a lightweight execution-time interface that converts recovered semantics into ALLOW, DEFER, target-consistency, and verified-selection decisions. VISA reduces invalid-instruction blind execution from 92.7% to 2.8% while preserving 94.0% of normal commands, and verified selection further improves target-consistent action exposure without updating the underlying policy. Overall, embodied language evaluation should measure semantic-action transfer, not semantic parsing alone.
[NLP-159] Hindsight Bias in Clinical Temporal Reasoning : How Future Data Exposure Affects Large Language Model Judgment ML4H2026 ALT
【速读】： 该论文旨在解决临床语言模型在评估过程中因使用回顾性数据（retrospective records）而产生的“事后偏见”（hindsight bias）问题。传统评估方法依赖于包含最终诊断、治疗反应和结局信息的回顾性记录，这可能导致模型通过利用未来信息而非在决策点的真实不确定性进行推理来获得高分，从而误导对模型真实临床推理能力的判断。为此，作者提出了一种成对基准测试（paired benchmark），用于衡量结果条件性偏移（outcome-conditioned shifts）这一与事后偏见相关的现象。其关键解决方案在于构建了一个包含171例病例（40例脓毒症与131例GLP-1/糖尿病相关病例）的数据集，这些病例以文本叙事和人工标注及大语言模型生成的文本时间序列（Textual Time Series, TTS）形式呈现，并为每例设置一个临床有意义的时间截断点（cutoff）。每个问题均配对一个前瞻性参考答案与一个结果一致的“事后陷阱”（hindsight trap）答案。通过对比模型在截断时间序列（仅使用截止点前信息）与完整时间线下的表现，系统评估准确率（Acc）、事后陷阱率（HTR）、答案不稳定性率（AIR）和事后偏见率（HBR）等指标。实验结果表明，暴露于完整时间线会引发一致的事后敏感性偏移，而时间掩码（temporal masking）策略可在不降低准确率的前提下有效减少偏见，从而更真实地反映模型在临床决策时的推理能力。
链接: https://arxiv.org/abs/2609.13454
作者: Misaki Matsuura,Sayantan Kumar,Ojas Kadam,Jeremy C. Weiss
机构: Case School of Engineering, Case Western Reserve University, USA; National Library of Medicine, National Institutes of Health, USA; Rice University, USA
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Machine Learning for Health Symposium (ML4H 2026)
Abstract:Clinical decisions are prospective, but clinical language models are often evaluated on retrospective records that reveal the final diagnosis, treatment response, and outcome. Such evaluations may reward the use of future information rather than reasoning under the uncertainty present at the decision point. We introduce a paired benchmark for measuring outcome-conditioned shifts consistent with hindsight bias in clinical temporal reasoning. It contains 171 case reports from the PubMed Central Open Access Subset—40 sepsis and 131 GLP-1/diabetes cases—represented as both textual narratives and human-annotated and LLM-generated textual time series (TTS). For each case, questions are tied to a clinically meaningful cutoff and paired with a prospective reference answer and an outcome-consistent \emphhindsight trap. Models answer each question using either a TTS truncated at the cutoff or the complete timeline; additional conditions vary the narrative source (original or synthetic) and TTS annotation source (human or LLM). We evaluate accuracy (Acc), hindsight trap rate (HTR), answer instability rate (AIR), and hindsight bias rate (HBR), each of which captures different signals of hindsight bias. Across GPT 5.6 Sol, Gemma 4, GLM 5.2, and Opus 5, full timeline exposure produces consistent hindsight-sensitive shifts, while temporal masking reduces bias without lowering accuracy.
[NLP-160] Causal Analysis and Mitigation of Spurious Onsets in Full-Duplex Speech LLM s
【速读】： 该论文旨在解决生成式语音对话模型（如Moshi及其衍生模型PersonaPlex）在用户长时间沉默期间出现的虚假语音启动问题。尽管这些模型支持全双工（full-duplex）并发听讲与说话，但在数字零输入（digital-zero input）条件下，仍会在5分钟连续对话中分别于12/40和11/40次试验中不恰当地开始说话。为探究其成因，研究提出两个假设：一是重复采样导致即使语音开启概率持续偏低仍产生语音输出；二是模型对自身非语音输出的条件依赖引发语音开启概率的突增。研究发现，在所有观测到的语音开启时刻，语音概率在80毫秒帧内上升超过九个数量级，有力支持后者假设。为此，论文提出一种因果反事实推理方法：判断模型是真实响应用户语音，还是在屏蔽前序用户输入后其下一词元分布仍保持稳定。若分布变化微小，则判定为虚假启动并予以抑制。在每模型40组独立测试中，该方法成功抑制了全部13例Moshi和9例PersonaPlex的虚假语音启动，同时完整保留了40/40的真实响应。该方法为推理时干预策略，无需重新训练，可在实时环境中运行，95百分位决策时间低于61毫秒，满足80毫秒帧预算要求。
链接: https://arxiv.org/abs/2609.13445
作者: Kento Nishi
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Speech-to-speech LLMs like Moshi, and its derivative PersonaPlex, can listen and speak concurrently through full-duplex generation. However, they can begin speaking inappropriately during prolonged user silence: under digital-zero input, Moshi and PersonaPlex initiate speech in 12/40 and 11/40 five-minute continuations, respectively. What causes this spurious speech? We investigate two hypotheses: either repeated sampling selects speech despite persistently low onset probabilities, or conditioning on the model’s nonspeech outputs causes an abrupt spike in onset probability. We find that, at every observed onset, speech probability spikes by over nine orders of magnitude in one 80-ms frame, supporting the latter hypothesis. Then, to suppress these onsets without blocking genuine responses, we ask a causal counterfactual question: is the model responding to user speech, or would its next-token distribution remain similar if the preceding user input were muted? Accordingly, we suppress onsets whose distributions change little under this intervention. Across 40 held-out trials per model with realistic microphone noise, our method suppresses 13/13 Moshi and 9/9 PersonaPlex spurious onsets, while preserving 40/40 genuine responses per model. Our inference-time method requires no retraining and runs in real-time, with 95th-percentile decision time below 61 ms, within the 80-ms frame budget. Our code is available at this https URL.
[NLP-161] CVSS-X: A Multilingual Speech-to-Speech Translation Corpus for 28 Languages
【速读】： 该论文旨在解决多语言语音到语音翻译（Speech-to-Speech Translation, STST）中缺乏大规模、高质量、跨语言且涵盖多样语言类型的数据集问题。现有数据集如CVSS主要支持从21种语言向英语的单向翻译，限制了双向及多语言场景下的研究。为此，论文提出CVSS-X，一个规模达16,000小时以上的合成语音对语料库，实现了从英语到28种目标语言（覆盖12个语系）的反向翻译，数据量为原CVSS的八倍。其解决方案的关键在于：构建全自动生成的双版本语料库——CVSS-X-C（每语言使用两个标准语音）与CVSS-X-T（采用跨语言语音克隆技术），确保在不同语言类型间保持一致的翻译性能，并验证了其与原始CVSS协同使用时可有效支持双向、多语言语音到语音翻译研究。
链接: https://arxiv.org/abs/2609.13413
作者: Lucas Rafael Stefanel Gris,Alef Iury Siqueira Ferreira,Frederico Santos de Oliveira,Augusto Seben da Rosa,Alexandre Costa Ferro Filho,Arlindo Rodrigues Galvão Filho,Anderson da Silva Soares
机构: Federal University of Goiás(戈亚斯联邦大学); Federal University of Mato Grosso(马托格罗索联邦大学); São Paulo State University(圣保罗州立大学)
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注:
Abstract:We introduce CVSS-X, a large-scale synthetic speech-to-speech translation corpus that extends CVSS by reversing the translation direction. While CVSS translates from 21 languages into English, CVSS-X enables translation from English into 28 target languages spanning 12 language families. The corpus comprises approximately 240,000 parallel speech pairs per language, totaling over 16,000 hours, eight times larger than CVSS. We provide two variants: CVSS-X-C with two canonical voices per language, and CVSS-X-T with cross-lingual voice cloning, both fully generated. Evaluation shows comparable translation quality to CVSS with consistent performance across typologically diverse languages. Combined with CVSS, this enables research on bidirectional and multilingual speech-to-speech translation. The code is available at this https URL and the dataset under CC-BY-NC 4.0 license at this https URL.
[NLP-162] RFCLLM : Evaluating LLM s Reasoning Ability of Network Protocol State Machines EMNLP2026
【速读】： 该论文旨在解决大语言模型（LLM）在将文本规范映射为形式化表示时的准确性问题，尤其关注其在协议规范中对有限状态机（Finite State Machine, FSM）推理能力的真实可信度。由于现有应用常假设LLM能完美理解规范，但实际表现可能受限于模型偏差与上下文复杂性，因此本文通过设计4类任务和1482个查询，针对16种网络协议评估了LLM生成的状态转移系统与人工构建的基准模型之间的对齐程度。研究的关键在于系统性地分析不同评判偏见、任务难度差异、四种上下文类型的影响以及协议特性的作用，从而为验证LLM在协议规范形式化推理中的可靠性提供实证依据，推动对LLM在关键安全场景下可信应用的深入理解。
链接: https://arxiv.org/abs/2609.13389
作者: Anqi Chen,Dan Goldwasser,Cristina Nita-Rotaru
机构: Northeastern University(东北大学); Purdue University(普渡大学)
类目: Computation and Language (cs.CL)
备注: Accepted at EMNLP 2026 (Findings)
Abstract:Mapping textual specifications into formal representations is essential for ensuring the correctness of protocol designs and implementations. LLM-generated mappings, used for networking security or testing, are assumed to capture a perfect understanding of the specification, which may not hold in practice. The goal of this paper is to assess the extent to which LLMs can interpret the specification correctly. We examine the degree to which an LLM’s implicit representation of a finite-state transition system-defined via natural language descriptions-aligns with a manually generated ground-truth model. We designed 4 tasks and 1482 task queries for 16 protocols. We evaluated different judge biases, observed the inherent difficulty gaps between tasks, looked into the effect of 4 context types, and the influence of protocol characteristics. Our work contributes to a step toward verifying whether LLMs can really be trusted in FSM (Finite State Machine) reasoning of protocol specifications.
[NLP-163] ScorePrompts: Natural-Language Exploration of Symbolic Music Scores through Analysis
【速读】： 该论文旨在解决符号化音乐分析结果难以被非专业用户理解与交互的问题，即如何实现对乐谱的自然语言解释与可交互式探索。其核心解决方案在于构建一个基于多层级分析结果与约束性语言模型（schema-constrained language model）的交互系统——ScorePrompts。关键创新点在于：首先，通过专门的音乐信息检索（MIR）组件在音符、拍子、小节和整曲等多层级上精确估计和组织和声、调性、终止式、结构边界、织体及音符角色等音乐属性；其次，利用受结构约束的语言模型将这些分析结果转化为连贯的自然语言描述，而非直接从原始MusicXML数据中推断结构，从而提升生成内容的准确性和可解释性；再次，采用确定性路由机制（deterministic router）针对用户问题（如“第14至18小节发生了什么变化？”）自动定位相关分析层级与对应小节，返回简洁响应并附带底层分析结果与置信度提示；最后，结合Verovio实现乐谱可视化，并将分析结果动态链接至具体小节与音符属性，同时暴露中间表格与不同分析层级间的不一致之处，确保分析过程透明可追溯。该系统强调对分析过程的可探查性与解释性，适用于探索性乐谱分析，而非编辑操作。
链接: https://arxiv.org/abs/2609.13291
作者: Emmanouil Karystinaios,Gerhard Widmer
机构: 未知
类目: ound (cs.SD); Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
备注: Accepted as a Late Breaking Demo (LBD) at the International Society of Music Information Retrieval conference (ISMIR) 2026
Abstract:We present ScorePrompts, an interactive system in which users upload a score, receive natural-language descriptions of its musical structure, ask questions about specific passages, and inspect the corresponding analysis results in staff notation. Specialist MIR components first estimate harmony, tonality, cadences, formal boundaries, texture, and note-level roles, organizing their outputs at note, beat, measure, and piece levels. A schema-constrained language model converts these results into descriptions rather than inferring musical structure directly from raw MusicXML. For questions such as “What changes in measures 14-18?”, a deterministic router selects the relevant measures and analytical levels and returns a concise response together with the underlying results and caveats. Verovio renders the score and links the returned information to cited measures and note-level attributes. The interface also exposes intermediate tables and disagreements between analytical levels. ScorePrompts is intended for exploratory score analysis and explanation, not score editing. The demo shows how existing analysis models, constrained language generation, QA retrieval, and notation-based visualization can provide natural-language access to symbolic music analysis while keeping intermediate results inspectable.
[NLP-164] Clinical Reasoning Under a Partially Observed Objective in Cone Beam CT Report Generation
【速读】： 该论文旨在解决锥形束计算机断层扫描（cone beam computed tomography, CBCT）影像报告自动生成中的评估与优化难题，核心挑战在于如何构建一个既能准确衡量报告事实一致性（factual entailment），又能高效驱动模型优化的复合目标函数。现有方法过度依赖词汇重叠（lexical overlap）指标（如BLEU、METEOR），导致模型倾向于生成高词频匹配但语义不一致的报告，严重损害了事实准确性。为此，本文提出一种新型复合客观评价指标：80%权重分配给基于大语言模型（large language model, LLM）的事实蕴含判断，20%权重来自词汇重叠，且仅在开发阶段可见部分词汇重叠信息。为实现高效优化，研究设计了一个离线的蕴含代理模型（entailment surrogate），其在受试者工作特征曲线下面积（AUC）达到0.987，可快速近似计算蕴含得分，使整个目标函数具备可微性与可优化性。实验表明，在622例公开数据集上，仅依据可见词汇重叠排序的报告得分仅为0.2909，而采用复合目标优化后的报告得分提升至0.4122，同时事实蕴含精度从0.522下降至0.266，说明单纯追求词汇匹配会显著降低语义准确性。进一步分析发现，图像头部信息中九个采集参数（如下颌骨覆盖范围、髁突覆盖范围）的预测性能分别达0.945和0.872，且采集中心本身对句式选择的预测能力（AUC=0.718）高于图像特征模型（AUC=0.663），揭示出当前词汇度量标准实际上奖励的是语音记录惯例而非解剖结构真实性。最终系统输出八条无条件陈述及五条基于头部几何信息（极性、侧向性、牙位一致性约束）的条件化陈述，在未见中心的50个测试案例上达到METEOR 0.3542。该研究通过引入基于大语言模型的事实蕴含评估与高效的代理模型，解决了传统自动评估指标与真实语义质量脱节的问题，实现了更可靠、更具解释性的医学影像报告生成系统。
链接: https://arxiv.org/abs/2609.13238
作者: Ajo Babu George,Govind Arun,Sidharth N Krishna,Uma Ranjan
机构: DiceMed(迪斯梅德); University of Maryland (马里兰大学); Indira Gandhi National Open University (英迪拉·甘地国立开放大学); Indian Institute of Technology Jammu (印度理工学院贾姆穆)
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages, 3 figures. ODIN 2026 CBCT report generation challenge system. Code: this https URL
Abstract:Maxillofacial report generation from cone beam computed tomography is scored here by a composite objective placing 80% of its weight on a large language model judgement of factual entailment and 20% on lexical overlap, of which only the lexical fifth is visible during development. The grader’s BLEU-4 and METEOR routines are reproduced in pure Python and match the reference to machine precision, and an offline entailment surrogate, which tells a report written for one patient from one written for another at an area under the curve of 0.987, makes the composite objective cheap enough to optimise directly. Over the 622-case public release, a report selected against the visible lexical ranking scores 0.2909, whereas one selected against the composite objective scores 0.4122, because pursuing n-gram overlap drives entailment precision from 0.522 down to 0.266. A 29 million parameter encoder fine-tuned on the release reaches a prevalence-weighted out-of-fold area under the curve of 0.486 over 985 statements, indistinguishable from the corpus prior, while nine numbers read from the image header reach 0.945 for mandible coverage and 0.872 for condyle coverage, and acquisition centre alone predicts sentence choice at 0.718 against 0.663 for the image-derived model, identifying dictation convention rather than anatomy as the quantity the lexical metrics reward. The delivered system emits eight unconditional statements and five gated on header geometry under polarity, laterality and tooth-level consistency constraints, and reaches METEOR 0.3542 over 50 held-out cases from an unseen centre. The dataset and code are available at this https URL
[NLP-165] Occlusal Geometry in Closed Form for Orthodontic Report Generation
【速读】： 该论文旨在解决正畸报告生成中因传统多模态图像描述方法忽略已注册咬合数据所蕴含的可直接测量咬合参数这一关键问题。现有公开的Bite2Text扫描配对数据在咬合闭合状态下已对齐，使得多个核心咬合指标（如覆𬌗、覆盖、中线偏移、横向重叠、反𬌗程度、牙尖交错滞后及咬合曲线等）可直接量化而非依赖推断，而传统方法未充分挖掘此优势。其解决方案的关键在于：通过牙弓锥度与牙弓闭合度重建个体化解剖坐标系，替代不适用于该数据集的通用RAS定位标准；将每侧牙弓简化为牙弓角坐标系下的咬合嵴轮廓，从而显式计算13项咬合参数；利用梯度提升模型将31个测量值映射至13个模板字段，并仅在患者级交叉验证优于基线时才预测对应字段；结合基于ConvNeXt-Tiny的五视图图像分类器进行场域融合，使平均字段准确率从0.601提升至0.683；最终采用确定性渲染器生成六部分结构化叙述文本。实验表明，该系统在无监督评估中表现优异，持有样本得分达BLEU-4 0.458、METEOR 0.677（对比口内扫描参考），且在ODIN 2026 Bite2Text测试中位列第三，仅落后第一名0.022 BLEU-4，同时实现单次推理<10秒/例的高效运行。
链接: https://arxiv.org/abs/2609.13237
作者: Ajo Babu George,Govind Arun,Sidharth N Krishna,Uma Ranjan
机构: DiceMed(迪斯梅德); University of Maryland (马里兰大学); Indira Gandhi National Open University (英迪拉·甘地国家开放大学); Indian Institute of Technology Jammu (印度理工学院贾姆穆)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 10 pages, 4 figures. Third-place system in the ODIN 2026 Bite2Text test phase. Code and data processing resources: this https URL
Abstract:Orthodontic report generation from intraoral data is normally cast as multimodal captioning, yet the released Bite2Text scan pairs are supplied already registered in occlusion, which makes several core occlusal quantities directly measurable rather than inferable. The system reported here exploits that property: an anatomical frame is recovered per case from arch taper and arch closure instead of the stated RAS convention, which does not hold across the release, and each arch is reduced to an occlusal ridge profile in arch-angle coordinates yielding overbite, overjet, midline deviation, transverse overlap, crossbite extent, cusp interdigitation lag, and the occlusal curves in closed form. Gradient boosting maps 31 such measurements onto 13 template fields, a field being predicted only where patient-level cross-validation beats its own majority baseline, and a deterministic renderer emits the corpus six-part narrative; a ConvNeXt-Tiny classifier over the five standardised photographic views is fused per field, raising mean field accuracy from 0.601 to 0.683. Reimplementation of the challenge evaluator shows that its BLEU-4 and METEOR are local variants whose F-mean weights recall nine to one, that two clinicians agree on 47 percent of findings for the same patient, and that a constant report consequently outscores a genuine second clinician report by 0.165 captioning. Held-out scores reach BLEU-4 0.458 and METEOR 0.677 against intraoral scan references and 0.278 and 0.507 against photograph references, and the submitted system placed third in the ODIN 2026 Bite2Text test phase at 0.2680 and 0.4629, within 0.022 BLEU-4 of first, running on CPU in under ten seconds per case. The dataset and code are available at this https URL
[NLP-166] GradRepair-ODE: Certified Gradient Repair for Neural ODE Training
【速读】： 该论文旨在解决神经微分方程（Neural Ordinary Differential Equations, Neural ODEs）在训练过程中因数值求解器与梯度计算耦合而引发的可靠性问题，特别是在存在松散步长、刚性动力学、混沌敏感性或事件不连续性等复杂情形下，导致生成的梯度方向可能具有数值不可靠性的风险。其解决方案的关键在于提出GradRepair-ODE框架，该框架通过在优化器步骤中对ODE梯度进行检查、修复与拒绝：首先生成多个梯度候选解，利用方向有限差分检验与求解器诊断信息进行一致性比对，识别潜在的数值失效模式；随后通过路径切换或更严格的重新计算方式修复选定梯度，并拒绝无法通过下降方向认证的更新步骤。实验表明，在六个合成系统中，GradRepair-ODE在低风险场景下保持原状，成功将Robertson与Lorenz系统的梯度修复至与严格参考解的余弦相似度达1.000，将不安全接受的步数从37次降至0，同时在事件不连续情况下主动拒绝更新而非应用未经验证的梯度。论文主张重构训练范式：神经微分方程的梯度应携带数值证据方可传递至优化器，从而提升科学机器学习与连续时间生成模型（如扩散概率流常微分方程和流匹配模型）的可信度。
链接: https://arxiv.org/abs/2609.13204
作者: Ziqian Bi,Xin Liang Chia
机构: Purdue University (普渡大学); Rice University (莱斯大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 22 pages, 8 figures, 6 tables
Abstract:Neural ordinary differential equations use numerical solvers inside the training loop. The solver determines the forward trajectory and also affects the gradient passed to the optimizer. That coupling creates a reliability problem for scientific machine learning and continuous-time generative modeling, including diffusion probability-flow ordinary differential equations and flow-matching models. Under loose step sizes, stiff dynamics, chaotic sensitivity, or event discontinuities, a differentiable ODE pipeline can return a finite gradient whose direction is numerically suspect. We introduce GradRepair-ODE, a reliability framework for checking, repairing, and rejecting ODE gradients at the optimizer step. The method computes several gradient candidates, compares them with directional finite-difference checks and solver diagnostics, diagnoses likely numerical failure modes, repairs selected gradients through path switching or stricter recomputation, and rejects steps whose descent direction cannot be certified. In six synthetic ODE systems, GradRepair-ODE leaves low-risk systems unchanged, repairs Robertson and Lorenz gradients to cosine similarity 1.000 against a strict reference, reduces unsafe accepted steps from 37 to 0, and rejects an event-discontinuous case instead of applying an uncertified update. The paper argues for a simple change in the training contract: an ODE gradient should reach the optimizer with numerical evidence attached.
[NLP-167] Machine Unlearning for Speech Question Answering in Large Audio-Language Models
【速读】： 该论文旨在解决大型音频-语言模型（Large Audio-Language Models, LALMs）在训练过程中可能无意记忆并泄露敏感信息的隐私风险问题，尤其针对语音问答（Speech QA）任务中声学感知与事实知识紧密耦合所带来的挑战。相较于以往针对文本型大语言模型（LLMs）或自动语音识别（ASR）系统的机器遗忘研究，语音问答场景下的隐私保护更为复杂。论文提出并评估了多种机器遗忘策略，包括梯度上升、任务算术以及基于对齐的微调方法，其核心在于通过引入“安全拒绝响应”机制，在移除私有知识的同时最大限度保留模型在非私有语音问答及通用语音理解任务上的性能。实验结果表明，这些方法可将隐私泄露率降低高达80%，且对模型核心能力的影响接近中性，验证了其在实际应用中的有效性与可行性。
链接: https://arxiv.org/abs/2609.13195
作者: Zhe Liu
机构: Meta Platforms, Inc. (Meta)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注:
Abstract:Large Audio-Language Models (LALMs) have recently shown strong capabilities in speech understanding and question answering (QA), but they also inherit privacy risks from large-scale training data, including the unintended memorization of sensitive information. In this work, we study machine unlearning for speech QA in LALMs, a setting that is more challenging than prior work on text-based Large Language Models (LLMs) or Automatic Speech Recognition (ASR) due to the tight coupling between acoustic perception and factual knowledge. We present and evaluate multiple unlearning strategies, including gradient ascent, task arithmetic, and alignment-based fine-tuning methods that enforce safe refusal responses, to remove private knowledge while still preserving performance on core capabilities. Through extensive experiments on speech QA datasets, we show that these unlearning methods can reduce the privacy leakage rate by up to 80% while maintaining near-neutral performance on non-private speech QA and general speech understanding benchmarks.
[NLP-168] LLM s or Naive Bayes? Old Gems or New Ways
【速读】： 该论文旨在解决在资源受限的高性能计算（HPC）环境中，面对大规模语言模型（LLM）的兴起，是否应淘汰经典机器学习方法如补集朴素贝叶斯（Complement Naive Bayes, CNB）的问题。其核心挑战在于评估在不同数据量与计算资源约束下，传统方法与现代生成式模型之间的性能、效率与能耗权衡。解决方案的关键在于通过系统性基准测试揭示：在存在标注数据的情况下（如AG News任务），CNB在准确率上可达到89.1%，与零样本27B参数规模的LLM相当（89.0%），且显著优于397B参数的前沿模型（84.8%），同时在通用CPU上实现每秒数千样本的推理吞吐量，远超小规模LLM在GPU上的批处理推理速度（慢40–486倍）。这一差距源于由内存带宽决定的结构性瓶颈，导致LLM在单位样本能耗上高出两个数量级。研究进一步提出基于可配置阈值和可验证Prometheus指标的Kubernetes Helm操作符，实现任务驱动的自动化模型选择，明确指出模型最优选择取决于任务类型与标注数据量（如主题分类中约需$ N \sim 10^4 $标签才能使CNB达到与LLM相当的性能，而零数据情感分析始终偏好LLM）。因此，对于有标注数据的文本分类任务，尤其是在资源受限场景下，朴素贝叶斯仍为最优决策。
链接: https://arxiv.org/abs/2609.13185
作者: Mohammad Firas Sada,Dmitry Mishin,John Graham,Seungmin Kim,Mahidhar Tatineni,Frank Würthwein
机构: San Diego Supercomputer Center, University of California, San Diego(圣地亚哥超级计算机中心，加州大学圣地亚哥分校); Lawrence Berkeley National Laboratory(劳伦斯伯克利国家实验室); Yonsei University College of Medicine(延世大学医学院)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: 6 pages, 3 figures, 5 tables. Accepted to the Proceedings of Practice and Experience in Advanced Research Computing (PEARC '26), Minneapolis, Minnesota, USA
Abstract:Large language models (LLMs) prompt a recurring question in research computing: should classical methods like Naive Bayes (NB) be retired? We benchmark Complement Naive Bayes against zero-shot and few-shot LLMs spanning four model families and a 37x range in scale (27B to a 1T-parameter mixture-of-experts) across text classification tasks. LLMs dominate only in zero-data regimes (98.0% vs 88.2% on Amazon Polarity sentiment), and even that win is contamination-prone: on a low-contamination sentiment task NB beats the zero-shot LLM (81.7% vs 73.0%). However, once labeled data is available (e.g., AG News), NB reaches 89.1% accuracy, statistically indistinguishable from the zero-shot 27B LLM (89.0%) and better than the 397B frontier model (84.8%), at thousands of samples/sec on a commodity CPU. Fine-tuned DistilBERT reaches 90.6% but at far lower throughput than NB at batch size 1 (Table 2). Our measured GPU throughput analysis shows small-LLM batched inference is 40-486x slower than NB CPU inference (the multiplier depends strongly on the host CPU), exposing a structural gap bounded by memory bandwidth, with roughly two orders of magnitude lower energy per sample. For resource-constrained HPC practitioners performing text classification with labeled data, NB remains the optimal choice. We show the decision line is task-dependent (NB reaches LLM parity around N \sim 10^4 labels for topic classification, while zero-data sentiment favors the LLM at all N tested) and provide a Kubernetes Helm operator that automates model selection using configurable thresholds and verifiable Prometheus metrics.
[NLP-169] Algorithm Validation as a Policy Audit: Evidence from Race-blind Charging
【速读】： 该论文旨在解决刑事起诉决策中因种族相关特征信息泄露而导致的系统性偏见问题，其核心挑战在于如何实现真正意义上的“种族无意识”（race-blind）起诉决策。为应对这一问题，研究提出并验证了bc2——一种基于大语言模型（LLM）的开源自动化文本去标识化算法，用于在案件审查前移除种族相关的代理变量（race-related proxies）。解决方案的关键在于：首先，通过高精度的自动去标识技术确保法律要求的技术合规性；其次，发现加州现行立法所规定的去标识范围存在明显遗漏，尤其是对地理位置等关键种族代理变量未予覆盖。研究进一步表明，若仅满足法律最低要求，仍保留大量可预测种族的信号（43.1%的剩余预测力），而bc2通过扩展去标识范围显著降低了此类偏差。因此，该研究不仅验证了算法的技术可行性，更揭示了政策设计与算法实现之间的协同优化路径，强调了在制度设计中纳入更全面的代理变量识别机制对于实现真正种族无偏决策的重要性。
链接: https://arxiv.org/abs/2609.13174
作者: Muskan Walia,Joe Nudell,Alex Chohlas-Wood
机构: New York University (纽约大学); Harvard University (哈佛大学)
类目: Computers and Society (cs.CY); Computation and Language (cs.CL)
备注:
Abstract:California recently required all prosecutors in the state to conduct a “race-blind charging” decision by reviewing case documents in which selected race-related proxies have been redacted. We validate bc2, an open-source, LLM-based algorithm that we developed to automate this redaction and that was used to facilitate race-blind review in more than 119,000 real-world cases in 2025. We evaluate two distinct questions: whether bc2 faithfully implements the state’s requirements and whether those requirements, even when faithfully implemented, advance the goal of race-blind decision-making. To do so, we draw on a corpus of nearly 5,000 real-world police reports that we assembled from jurisdictions across the United States. Under a stringent document-level measure, we find that the latest version of bc2 faithfully implements the legal mandate on 96.7% of narratives in our sample. This performance represents a substantial improvement over earlier versions of bc2 and exceeds that of leading open-source redaction methods. Our validation also shows that California’s mandate misses key proxies for race, including location information. Redacting these additional proxies beyond those covered by the state mandate, as bc2 does, eliminates 43.1% of the predictive signal that remains after compliance with the mandate. These findings show that validation can do more than assess technical compliance: it can also improve algorithms and help policymakers achieve underlying policy goals.
[NLP-170] stHallVQA: Exploring LVLMs Document-Level Reasoning under Redundant Contexts from Scientific Exams
【速读】： 该论文旨在解决当前平面视觉问答（VQA）基准测试在任务设计上的局限性，即现有基准要么侧重于长文档理解但推理深度不足，要么虽具备复杂视觉推理能力却局限于单页、无噪声的简单场景，且普遍忽视了无关视觉标记（irrelevant visual tokens）对模型性能的负面影响。其解决方案的关键在于提出TestHallVQA——一个兼具文档级规模与人类考试难度的多图像VQA基准，能够同时支持多层次上下文冗余的可控注入；并据此引入一种新型评估指标F1-R²，联合量化大视觉语言模型（LVLMs）在计算推理能力与面对文档级冗余时的证据检索鲁棒性两方面表现。实验结果揭示了主流LVLMs在多个维度上的潜在缺陷，为未来研究提供了具体方向。
链接: https://arxiv.org/abs/2609.13158
作者: Yongqi Yu,Yu Zhang
机构: Harbin Institute of Technology (哈尔滨工业大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large Vision–Language Models (LVLMs) are increasingly expected to perform visual question answering (VQA) over planar media. However, existing planar VQA benchmarks typically emphasize isolated challenges: some emphasize long-document understanding with limited reasoning depth, while others require complex visual reasoning but remain restricted to single-page, noise-free settings. Moreover, through theoretical analysis, we identify the impact of irrelevant visual tokens, which leads to measurable performance degradation but has received little attention with respect to systematic quantification. To address these limitations, we introduce TestHallVQA, a multi-image VQA benchmark that simultaneously embodies document-level scale and the difficulty of human examinations, while providing comprehensive task coverage. Leveraging TestHallVQA’s ability to controllably inject multi-level contextual redundancy, we further propose a novel metric, F1-R\textsuperscript2, which jointly quantifies LVLMs’ computational reasoning capability and their evidence retrieval robustness against document-level redundancy. Extensive experiments and analyses on mainstream LVLMs reveal their latent deficiencies across multiple dimensions, offering concrete insights and directions for future research. The associated datasets, code, and complete theoretical derivations are available at this https URL.
[NLP-171] Lexical Prompt Compression for Large Language Models : A Training-Free Deterministic Pipeline with Empirical Pareto Analysis Across Eleven Task Categories
【速读】： 该论文旨在解决大语言模型（LLM）推理中因提示（prompt）日益庞大复杂而导致的计算成本与延迟上升问题。现有方法如链式思维（chain-of-thought reasoning）和上下文学习（in-context learning）常使提示长度超过数千词元，显著增加资源开销。尽管已有基于学习的压缩方法（如LLMLingua和Selective Context）可实现高压缩比，但其依赖辅助语言模型且具有非确定性缺陷。为此，本文提出一种无需训练、完全确定性、仅在CPU上运行的压缩方案，其核心在于利用经典词汇自然语言处理（lexical NLP）技术构建可配置的流水线，集成11种可切换的词汇变换策略，包括停用词删除、填充短语剔除、缩略语替换、基于词性标注的剪枝、词干化、基于WordNet的同义词简化以及命名实体保留等。通过在六个数据源共1,242个英文提示上评估15种配置，覆盖11类任务，生成18,630对GPT-4o-mini输出结果，并采用BLEU、ROUGE-1/2/L、BERTScore-F1及SentenceBERT余弦相似度综合评估输出保真度。结果显示，最激进配置平均实现40.3%（σ=9.2）的词元压缩率，同时保持0.876的BERTScore-F1；仅停用词删除配置亦达29.6%压缩率与0.913的保真度。各任务类别下的压缩-保真帕累托前沿被系统刻画，发现常识推理任务在激进压缩下存在系统性性能退化。所有代码、提示及细粒度结果均已开源，确保可复现性。
链接: https://arxiv.org/abs/2609.13154
作者: Shamin Chokshi
机构: 未知
类目: Computation and Language (cs.CL)
备注: the paper is 27 pages with 10 figures and 4 tables. Code and data for the experiments: this https URL . Python package (pip install less-tokens): this https URL
Abstract:Recent advances in large language models (LLMs) have made prompts increasingly large and complex. Techniques such as chain-of-thought reasoning (Wei et al., 2022) and in-context learning (Brown et al., 2020) frequently push real-world prompts past several thousand tokens, increasing inference cost and latency. Learned compression methods such as LLMLingua (Jiang et al., 2023) and Selective Context (Li et al., 2023) achieve high compression ratios but require auxiliary language models and are non-deterministic. We ask a complementary question: how far can a training-free, fully deterministic, CPU-only pipeline based on classical lexical NLP be pushed before output quality degrades significantly? Eleven toggleable lexical transformations - stopword removal, filler-phrase deletion, contraction and abbreviation substitution, part-of-speech-based pruning, lemmatization, WordNet-driven synonym shortening, and named-entity preservation - are assembled into a configurable pipeline. Fifteen configurations are evaluated on 1,242 English-only prompts from six sources (Dolly-15k, LMSYS-Chat-1M, WildChat-1M, MMLU, GSM8K, HellaSwag), spanning eleven automatically derived task categories, yielding 18,630 paired GPT-4o-mini completions. Output preservation is measured using BLEU, ROUGE-1/2/L, BERTScore-F1, and SentenceBERT cosine similarity. The most aggressive configuration achieves a mean token reduction of 40.3% (sigma = 9.2) at a BERTScore-F1 of 0.876 against the original-prompt output; a stopword-only configuration achieves 29.6% reduction at 0.913. The compression-versus-fidelity Pareto frontier is characterized per task category, with commonsense reasoning a systematic failure mode under aggressive compression. All code, prompts, and per-cell results are released for reproducibility.
[NLP-172] PhysMent: An Interactive Approach For LLM Reasoning In Physics Problems
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在动态物理世界中的主动实验推理能力不足的问题，尤其关注其在复杂、交互式物理场景中通过迭代操作与环境互动来推断物理规律的能力。传统静态科学基准仅提供固定输入信息，而无法评估模型在真实物理探索中的行为模式。为此，作者提出了PhysMent——一个基于MuJoCo物理引擎的动态交互式基准测试框架，要求模型通过施加力、查询物体状态、推进时间步长以及修改场景几何结构等工具化操作，逐步获取信息并作出判断。其核心解决方案在于构建了一个多维度、分层次的评测体系：包含105个经典力学场景，覆盖四种难度层级（易/难）与任务类型（单概念/多概念）、三种场景模态（标准、对象创建、隐藏物体）及场景操控类别，并采用六维评分框架全面评估模型表现。研究发现，当前主流模型在定性单概念任务上表现尚可（最高达80%准确率），但在需精确执行多步骤实验程序的定量任务中显著退化，最困难的单概念类别平均准确率低于30%，主要瓶颈在于程序性推理能力（如自适应多步工具使用），而非概念理解缺陷；各模型准确率介于25%至67%之间，失败原因主要为过早提交答案、探索效率低下以及对模拟器反馈的不一致感知，而非知识性空白。
链接: https://arxiv.org/abs/2609.13152
作者: Joseph Chan,Utkarsh Jha,Xiyin Yang,Abhinav Jarajapu,Anik Sahai,Eddie Hu,Robin Jeshua Deepak,Stefano Saravalle,Aditya Shah
机构: 未知
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Large language models (LLMs) perform strongly on static science benchmarks, yet their ability to reason about the physical world through active experimentation remains poorly understood. We introduce PhysMent, a benchmark that evaluates LLM physical reasoning via iterative, toolmediated interaction with a MuJoCo physics simulator. Unlike static benchmarks that supply all quantities upfront, PhysMent requires models to discover information by applying forces, querying object states, advancing time, and modifying scene geometry before answering. The benchmark comprises 105 scenes of classical mechanics, organized across four difficulty regimes (Easy/Hard and Single/Multi), three scene modalities (standard, object creation, hidden objects), and a scene-manipulation category, evaluated with a six-dimensional scoring framework. Results show that current models perform reasonably well on qualitative single-concept tasks (up to 80% accuracy) but degrade substantially on quantitative tasks that demand precise, multi-step experimental procedures: most models fall below 30% on the hardest single-concept category, where the bottleneck is procedural (adaptive multi-step tool use) rather than conceptual load. Across the seven models, accuracy ranges from 25% to 67%, with failures due to premature answer submission, inefficient exploration, and inconsistent grounding in simulator feedback rather than conceptual gaps.
[NLP-173] oken Merging for Multilingual Speech Recognition: A Systematic Study Across Model Scale and Fine-Tuning
【速读】： 该论文旨在解决多语言语音识别模型（如Whisper）在部署时计算成本高昂的问题，尤其针对低资源语言的高效推理需求。其核心解决方案是采用**令牌合并（token merging）**技术，通过在推理阶段动态合并冗余特征表示，缩短序列长度，从而降低计算开销，且无需重新训练模型。研究系统评估了该方法在16种不同语言及三种模型规模下的表现，并考察其与低资源语言微调（如DoRA）的兼容性。结果表明，令牌合并在几乎不损失转录准确率的前提下显著提升了计算效率，且在模型微调后仍保持有效性，证明其是一种切实可行、可广泛部署的加速策略。
链接: https://arxiv.org/abs/2609.13151
作者: Dylan Luke Holyoak
机构: Epoch Learn ϕ0 (Epoch Learn)
类目: Computation and Language (cs.CL)
备注: 11 pages, 3 figures
Abstract:Leading multilingual speech recognition models like Whisper transcribe diverse, low-resource languages without language-specific training but are computationally expensive to deploy. Token merging mitigates this inefficiency by dynamically combining redundant features, shortening the sequence length during inference without requiring retraining. In this paper, we systematically evaluate token merging on the Whisper model family across sixteen diverse languages and three different model sizes. We also test how token merging interacts with fine-tuning (DoRA) on low-resource languages. Our findings show that merging tokens increases computational efficiency with almost no loss in transcription accuracy across most low-resource languages and model sizes, and it works even after the model has been fine-tuned. Our results demonstrate that token merging is a highly practical method for making multilingual speech recognition faster and cheaper to deploy.
[NLP-174] he Limits of Reference-Free Speech Quality Metrics as Evaluators and Rewards on Modern Text-to-Speech
【速读】： 该论文旨在解决生成式语音合成（Text-to-Speech, TTS）系统中自动质量评估指标（如UTMOS、DNSMOS和SCOREQ）是否真正反映人类偏好这一关键问题。尽管这些无参考质量预测器已被广泛用作TTS模型的自动评价工具及偏好优化中的奖励信号，但其有效性依赖于一个核心假设：预测得分与人类主观偏好高度一致。本文通过在六个涵盖从高失真到无缺陷语音质量范围的人工评分语料库上进行成对比较任务，系统检验了这一假设。实验结果显示，当语音片段存在可听缺陷时，多数预测器与人类听者意见一致；然而，当两个片段均无明显缺陷时，单一预测器无法可靠识别出更受偏好的样本，部分甚至表现低于随机选择较长音频片段的基准水平。研究进一步表明，通过校准并融合互补信号的复合指标（calibrated composite）是当前表现最强的评估方法，尽管在最优质音频上仍未能完全逼近人类评判的上限。此外，即使采用等权重集成的简单策略，在缺乏校准数据的情况下亦能作为后训练阶段的有效奖励信号。相比之下，仅以单一指标进行策略优化易引发“奖励黑客”（reward hacking）现象，导致模型过度趋近该指标最优值，却损害独立评测人员和人工听感测试的表现。本研究的核心贡献在于提出了一套可复现的评估协议、提供了多语料下的预测器得分数据，并深入诊断了单一评分在高质量音频场景下失效的机制与原因。
链接: https://arxiv.org/abs/2609.13150
作者: Antonis Asonitis,Juan Pablo Zuluaga Gomez,Francesco Verdini,Aref Farhadipour,Marzieh Razavi,Pierre-Edouard Honnet,Vijeta Avijeet
机构: University College London (伦敦大学学院); Imperial College London (帝国理工学院); University of Oxford (牛津大学); King’s College London (国王学院)
类目: ound (cs.SD); Computation and Language (cs.CL); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
备注:
Abstract:Reference-free quality predictors such as UTMOS, DNSMOS and SCOREQ are the de facto automatic evaluators for text-to-speech (TTS) and are increasingly adopted as reward signals for preference optimization. Both roles presuppose that the predicted score tracks human preference. In this work, we test this assumption across six human-rated corpora spanning the quality range from artifact-rich to defect-free TTS, evaluating each predictor on a pairwise task that asks whether the clip it scores higher is the clip listeners prefer, and we subject interpretable prosodic and signal-processing features to the same protocol. When one clip carries audible defects the predictors tend to agree with listeners. Once both clips are clean, no single predictor reliably identifies the preferred sample, and several fall below the accuracy of simply picking the longest-duration clip. A calibrated composite of complementary signals is the strongest evaluator we test, though on the cleanest audio it recovers only part of the gap to the human ceiling. Additionally, using even an equal-weighted ensemble of metrics helps as a post-training reward, where no calibration data is available. Optimizing a single score with policy optimization induces reward hacking, driving the metric toward its optimum while independent held-out judges and a human listening test deteriorate. The composite reward resists this behavior and tends to improve the model. Our contribution is the evaluation protocol, the predictor scores across these corpora, and the diagnosis of when and why single scores fail.
[NLP-175] BudgetBench: A Budget-Tiered Protocol and Pilot Harness for Memory Strategy Evaluation in Local Large Language Model Agents
【速读】： 该论文旨在解决本地大语言模型智能体（local large language model agents）在实际部署中面临的活跃上下文（active context）资源受限问题，具体表现为内存容量、预填充延迟（prefill latency）、缓存增长及服务目标对单次调用可处理输入令牌数的严格约束。为系统性评估不同内存策略在有限输入预算下的表现，作者提出了一种名为BudgetBench的可复用的基准协议与参考测试框架，其核心在于将每轮调用的输入令牌预算（input-token budget）作为独立变量进行控制性实验。该方案的关键创新在于构建了一个可交换的内存策略接口（MemoryStrategy contract）、显式的预算强制机制、确定性或版本化的评分器（graders）、提示审计元数据（prompt-audit metadata）以及可重现性成果物，从而实现对质量、预算利用率、延迟及预算违规率（budget-violation rates）等多维度指标的标准化测量。通过在多个基准任务（如SWE-bench Verified、LongBench v2、LongMemEval）上的试点研究，揭示了传统单预算评估所掩盖的非单调质量曲线、预算合规失败现象及操作点选择的复杂性，强调了在固定预算下评估内存策略的必要性与挑战性，其最终贡献聚焦于建立一套可持续扩展的、具备透明性与可重复性的评估范式。
链接: https://arxiv.org/abs/2609.13149
作者: Aditya Karnam Gururaj Rao,Arjun Jaggi
机构: Google(谷歌)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 44 pages, 4 figures. Code and artifacts: this https URL
Abstract:For local large language model agents, active context is a scarce resource: memory capacity, prefill latency, cache growth, and service objectives all constrain how many input tokens each call can afford. We present BudgetBench, an active-budget protocol and reference harness that treats the per-call input-token budget as the independent variable when comparing memory strategies. Holding the model, task, sampler, and decoding fixed, it sweeps budgets over 2K, 4K, 8K, 16K, and 32K tokens and records quality, budget utilization, latency, and, as a first-class outcome, budget-violation rates. The core contribution is this reusable measurement surface: a swappable MemoryStrategy contract, explicit budget enforcement, deterministic or versioned graders, prompt-audit metadata, and reproducibility artifacts, released at this https URL. We substantiate the protocol with pilot studies rather than final rankings. Across a local qwen2.5:1.5b pilot (89 items each on SWE-bench Verified and LongBench v2), a hosted 50-item Qwen3 30B-A3B LongBench replication with exact tokenization, and a 500-item LongMemEval oracle study scored by the official GPT-4o evaluator, the harness exposes budget-compliance failures, non-monotonic quality curves, and operating points that single-budget evaluation hides. The budgeted-versus-full-context direction remains unresolved: the local slice is near-null and the hosted replication favors full context in point estimate. We report results transparently, including that the early pilot’s tokenizer approximation undercounts some served-model prompts, so its violation rows are tokenizer-approximation diagnostics, not claim-bearing results; all timings are operational diagnostics. The reusable contribution is the protocol, harness, and failure-reporting discipline needed to scale fixed-budget memory-strategy evaluation. Comments: 44 pages, 4 figures. Code and artifacts: this https URL Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL) Cite as: arXiv:2609.13149 [cs.LG] (or arXiv:2609.13149v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.13149 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-176] Is Bash All You Need? An Empirical Study of Tool Interfaces for Enterprise Digital Worker Agents
【速读】： 该论文旨在解决在企业级任务中，通用命令行外壳（shell）是否能够优于专用工具接口的问题。企业工作不仅涉及编程，还包括跨应用与服务的切换、团队协作以及专业分析等复杂行为，因此对工具接口的灵活性与通用性提出了更高要求。研究通过在TheAgentCompany和APEX-Agents两个基准上对比五种工具接口（类型化工具、带bash的类型化工具、仅bash、带持久化代理合成工具的bash，以及程序化工具调用PTC），采用Opus-4.8与GPT-5.5进行评估，发现仅使用bash的方案在两项基准上均显著优于其他方法，分别提升21.8–24.5个百分点和4.8–7.4个百分点，且总令牌消耗减少19%–72%。引入类型化工具或持久化工具合成机制并未带来可检测的性能增益。尽管程序化工具调用（PTC）在令牌使用上更高效且表现接近直接类型化调用，但在任务质量与成本效益方面普遍不及仅bash方案。研究的关键结论是：当可执行代码能被有效隔离时，推荐采用“仅bash”策略以实现最佳性能；而在需满足安全或合规要求、必须限定工具目录的场景下，则应选择PTC。
链接: https://arxiv.org/abs/2609.11999
作者: Hazel Mak,Susheel Suresh,Sahil Bhatnagar,Barry Wang,Chhaya Methani,Alejandro Gutierrez Munoz
机构: Microsoft(微软)
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 13 pages, 7 figures, 12 tables
Abstract:In this study, we examine whether a general shell can outperform specialized tools on enterprise tasks. Shell-based agents have shown strong results in coding, but enterprise work also involves moving between applications and services, coordinating with coworkers, and performing professional analysis. We compare five tool interfaces on TheAgentCompany and APEX-Agents using Opus-4.8 and GPT-5.5: typed tools, typed tools plus bash, bash alone, bash with persistent agent-synthesized tools, and programmatic tool calling (PTC), which runs programs whose actions are restricted to a typed tool catalog. Bash alone outperforms typed tools on both benchmarks, improving score by 21.8-24.5 pp on TheAgentCompany and 4.8-7.4 pp on APEX-Agents while using 19-72% fewer total tokens. Adding typed tools or persistent tool synthesis to bash produces no detectable pooled score gain. PTC uses fewer tokens than direct typed calls with broadly similar task performance, but generally underperforms bash alone in both quality and cost efficiency. For enterprise practitioners, these results favor bash alone when arbitrary execution can be isolated and PTC when security or compliance policies require a fixed tool catalog.
[NLP-177] Reducing the Output-Mode Gap in Speech Language Models via Joint-Output On-Policy Distillation
【速读】： 该论文旨在解决语音大模型在端到端语音生成（S2TS）模式下存在的输出模式差距（Output-Mode Gap, OMG）问题，即相同语音输入在生成文本时，其内部生成的文本准确性显著低于纯语音转文本（S2T）模式下的表现。这一现象源于自回归生成中交错的文本与声学标记（acoustic tokens）构成上下文，导致后续文本预测受到已生成声学信息的干扰。为缓解此问题，论文提出联合输出的在线策略蒸馏（Joint-Output On-Policy Distillation, JO-OPD）方法：通过利用纯文本生成的教师模型（S2T）对联合生成学生模型（S2TS）进行监督，以学生自身生成的交错序列作为输入，在每个文本位置上，教师模型基于学生先前输出的纯文本投影提供软标签，而学生则从完整的交错历史中进行预测；同时引入保持目标以正则化非文本部分的生成。实验表明，该方法在Step-Audio-2-mini和Baichuan-Audio-Instruct数据集上有效降低了OMG，Spoken-MQA任务中降幅达42.87至16.26个百分点，且未损害S2T性能，优于匹配的监督微调基线，ASR评估显示语音回答准确率提升7.49点。
链接: https://arxiv.org/abs/2609.15313
作者: Daxin Tan,Dehua Tao,Chengxi Deng,Hanlin Zhang,Xiao Chen
机构: 未知
类目: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Sound (cs.SD)
备注:
Abstract:Autoregressive generation of interleaved text and acoustic tokens is a common approach to spoken-response generation in speech large language models. Although this design enables streaming generation with explicit textual guidance, generated acoustic tokens become part of the context for subsequent text predictions. Given identical speech inputs, we observe markedly lower answer accuracy for the internal text generated in speech-to-text-and-speech (S2TS) mode than for speech-to-text (S2T) responses. We term this discrepancy the \emphoutput-mode gap (OMG). To reduce OMG, we propose \emphJoint-Output On-Policy Distillation (JO-OPD), which distills the model’s stronger S2T policy into joint generation using student-generated S2TS trajectories. At each text position, the S2T teacher provides soft targets from a text-only projection of the student’s preceding outputs, while the student predicts from the corresponding full interleaved history. A preservation objective further regularizes native non-text predictions. Experiments on Step-Audio-2-mini and Baichuan-Audio-Instruct reveal OMG across two interleaved generation architectures. On Step-Audio-2-mini, JO-OPD reduces OMG from 42.87 to 16.26 percentage points on Spoken-MQA and from 29.72 to 13.04 points on speech-rendered GSM8K, with little change in S2T accuracy and substantially larger reductions than matched SFT baselines. ASR-based evaluation further shows a 7.49-point improvement in spoken-answer accuracy on Spoken-MQA.
[NLP-178] Exploring Multimodal Turn-Taking Cues in Face-to-Face Conversation using Voice Activity Projection EMNLP2026
【速读】： 该论文旨在解决对话系统在话轮转换（turn-taking）预测中仅依赖音频信号导致的性能瓶颈问题，尤其在面对面交互场景下，忽视视觉线索可能限制模型对人类自然互动模式的理解。其核心解决方案是将多模态视觉特征引入现有的自监督语音活动投影（Voice Activity Projection, VAP）模型中，通过融合来自大规模元宇宙无缝交互数据集（Meta Seamless Interaction）的面部与身体视觉信息，包括注视方向、头部运动、肢体与手部姿态以及面部动作单元（Facial Action Units, FAU）。研究探索了多种融合策略，如特征拼接、交叉注意力融合、差分特征与可训练门控机制，结果表明，视觉信息显著提升了话轮转换预测性能，其中FAU贡献最为突出；同时，身体与注视特征提供互补性信息，全模态融合模型表现最优。此外，研究发现模型在即兴表演（acted）与自然对话（non-acted）数据上的表现存在差异，提示训练数据的语境真实性对模型泛化能力具有重要影响。
链接: https://arxiv.org/abs/2609.14666
作者: Willem Berner,Julio Cesar Cavalcanti,Kalle Åström,Gabriel Skantze
机构: Lund University (隆德大学); KTH Royal Institute of Technology (皇家理工学院)
类目: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Sound (cs.SD)
备注: Accepted to EMNLP 2026 Workshop on Multimodal Interaction in Face-to-Face Dialogue (MINT)
Abstract:Turn-taking is a fundamental component of spoken interaction, and while humans naturally rely on both verbal and non-verbal signals, dialogue systems usually depend on audio cues alone. This paper investigates whether visual features from face-to-face conversations can enhance turn-taking prediction beyond what is achievable from audio-only. We extend the Voice Activity Projection (VAP) model, a self-supervised transformer-based model for predicting future voice activity, by incorporating visual features extracted from the large-scale Meta Seamless Interaction dataset of dyadic face-to-face conversations. The visual features include gaze direction, head movement, body and hand pose, and facial action units (FAU). For incorporating the visual features, we explore concatenation, cross-attention fusion, delta features, and trainable gating mechanisms. Results show that visual information improves performance over the audio-only baseline, with FAU being significantly more informative than other feature groups. Body and gaze features nevertheless contribute complementary information, as the model combining all features performs best. Furthermore, results indicate that performance on specific tasks varies depending on whether training and test data come from improvised (acted) or naturalistic (non-acted) conversations.
[NLP-179] Parameter-Efficient Quantum NLP for Paraphrase Detection: Performance Robustness and Entanglement
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在自然语言处理任务中缺乏严谨实证验证的问题，特别是针对量子机器学习（Quantum Machine Learning, QML）在自然语言理解任务中的有效性与可扩展性问题。其核心挑战在于如何在真实基准上系统评估混合量子-经典变分电路在语义相似性判断任务中的性能表现，并与经典模型进行公平对比。解决方案的关键在于构建一个10量子比特的混合量子-经典变分电路（含2,148个参数），并在多个标准数据集（MRPC、Quora Question Pairs、PAWS）上开展多基准、参数匹配的系统性实验。研究发现，该电路在QQP任务上达到75.53% ± 0.75%的准确率，显著优于参数规模相当的经典基线模型（如DeepMLP和F1指标），且在参数量减少31,191的情况下超越了量化版DistilBERT；在MRPC上以仅54,000参数成本达到BERT-base 92%的精度。进一步分析表明，电路深度与数据集复杂度之间存在缩放关系，而通过Meyer-Wallach纠缠度量揭示的多体纠缠是性能提升的主要驱动因素（相关系数r = 0.85）。更值得注意的是，在对抗性数据集PAWS上，该电路展现出未经过对抗训练却仍具有98.2%召回率的鲁棒性，较经典模型提升16.6个百分点，统计显著（p < 0.001）。这些结果首次实现了对混合变分电路在自然语言处理任务中的参数匹配、多基准、系统性实证验证，为未来硬件部署提供了理论依据与仿真基础。
链接: https://arxiv.org/abs/2609.14529
作者: Farha Nausheen,Khandakar Ahmed,Farina Riaz
机构: ISILC, Victoria University, Melbourne, VIC, Australia; CSIRO, Australian International Institute of Higher Education, Sydney, NSW, Australia
类目: Quantum Physics (quant-ph); Computation and Language (cs.CL)
备注:
Abstract:Rigorous empirical validation of quantum machine learning on natural language tasks remains scarce. We evaluate a 10-qubit hybrid quantum-classical variational circuit (2,148 parameters) for paraphrase detection across three benchmarks: MRPC, Quora Question Pairs (QQP), and adversarial PAWS. On QQP (n = 10 seeds), the circuit achieves 75.53% +/- 0.75%. accuracy, statistically outperforming parameter-matched classical baselines (DeepMLP: p = 0.015, Cohen’s d = 1.20; F1: p 0.001, d = 2.32) and surpassing DistilBERT-4bit with 31,191 fewer parameters. On MRPC the optimal 2-layer variant reaches 92% of BERT-base accuracy at 54,000 lower parameter cost. Circuit depth analysis reveals a dataset-depth scaling effect; entanglement analysis via the Meyer-Wallach measure identifies multi-qubit entanglement as the primary performance driver (r = 0.85 across four variants). Adversarial evaluation on PAWS reveals emergent robustness: 98.2% recall versus 81.6% classical (+16.6 pp, d = 1.24, p 0.001), without adversarial training. These results constitute the first systematic parameter-matched multi-benchmark empirical validation of hybrid variational circuits for NLP. All results are from classical simulation; hardware validation is future work.
[NLP-180] URCHIN: A Horizontal Spiking Language Model for Data-Constrained Pretraining EMNLP2026
【速读】： 该论文旨在解决当前语言模型在学习机制上缺乏生物合理性的问题，即现有模型通常基于大规模互联网语料进行训练，而忽视了人类语言习得过程中神经回路的生物学约束。具体而言，其核心挑战在于如何在有限的、类儿童规模的数据（child-scale data）上构建一个既符合神经发育规律又具备高效计算能力的语言模型。解决方案的关键在于提出一种名为URCHIN（Unified Recurrent Connectome with Horizontal Integrate-and-fire Neurons）的新型架构，该架构基于并行分层连接组脉冲状态空间模型（PHCSSM），采用遵循戴尔定律（Dale’s Law）的横向递归连接组结构，将漏电整合放电（leaky integrate-and-fire, LIF）神经元通过多轮传输环路实现对每个词元（token）的动态求解，并最终收敛至固定点。该模型设计极为精简：仅包含128个神经元的单层水平结构、无注意力机制、参数量仅为423万。更重要的是，该模型存在两种等价实现方式——一种是用于GPU高效训练的并行状态空间模型（SSM）扫描，另一种是面向CPU或类脑芯片边缘部署的事件驱动脉冲神经网络（RSNN），二者共享同一组权重且无需转换即可直接部署，从而在所有三个BabyLM评测赛道（Strict-100M、Strict-Small和Multilingual）中提供了一个兼具生物合理性、计算效率与可直接部署性的基准参考。
链接: https://arxiv.org/abs/2609.13899
作者: Po-Han Chiang
机构: National Yang Ming Chiao Tung University (国立阳明交通大学)
类目: Neurons and Cognition (q-bio.NC); Computation and Language (cs.CL)
备注: 12 pages, 2 figures, 6 tables. Accepted to the BabyLM Challenge 2026 (BabyLM Workshop, EMNLP 2026)
Abstract:The BabyLM challenge measures how much language a model can learn from developmentally-plausible, child-scale data rather than internet-scale corpora, yet prior language models forgo the biological constraints of the neural circuitry that acquires human language: spiking neurons separated into excitatory and inhibitory populations wired by a recurrent lateral connectome. This paper presents URCHIN (Unified Recurrent Connectome with Horizontal Integrate-and-fire Neurons), which applies the Parallelized Hierarchical Connectome Spiking State-space Model (PHCSSM) to language modeling: leaky integrate-and-fire neurons coupled by a Dale’s-law lateral connectome resolve each token through a multi-transmission loop that recirculates activity to a fixed point. The instantiation is deliberately minimal: a single horizontal layer of 128 neurons, no attention, and 4.23M parameters. Two implementations share one set of weights and produce identical benchmark scores, so URCHIN is trained once and deployed either way with no conversion step: a parallel state-space model (SSM) scan that is GPU-efficient for training, or an event-driven recurrent spiking neural network (RSNN) with constant-cost inference for CPU or neuromorphic edge deployment. Across all three BabyLM tracks (Strict-100M, Strict-Small, and Multilingual), URCHIN offers a biologically plausible, efficient, and directly deployable reference point.
信息检索
[IR-0] CiteGuard-RAG : A Validation-Centered AI System for Evidence-Grounded Question Answering
链接: https://arxiv.org/abs/2609.15830
作者: Sumit Barua,Guan Hong,Halil Dursunoglu,Charles Rodgers,Alvis Fong
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: Submitted to Engineering Reports. 22 pages, 2 figures, 12 tables
Abstract:Retrieval-augmented generation (RAG) can improve access to complex information; however, retrieving evidence alone does not ensure that answers are grounded, citation-valid, or appropriately refused. This paper introduces CiteGuard-RAG, a validation-centered AI system for evidence-grounded question answering. The system integrates hybrid semantic-lexical retrieval, citation-constrained generation, sentence-level grounding validation, and single-pass regeneration. Validation is used at runtime to determine whether a candidate answer should be accepted, refused, or regenerated before final delivery. CiteGuard-RAG is evaluated on 400 questions across a controlled housing-law dataset, PrivacyQA, and CUAD. In the controlled evaluation, it achieves 99.1% retrieval accuracy, 98.3% grounded-answer accuracy, and 98.3% citation validity, with no validation-detected hallucinations. Ablation results show that grounded-answer accuracy drops sharply when validation is removed, even when retrieval accuracy remains unchanged. External evaluation shows that while citation validity remains strong, evidence utilization, span alignment, and refusal calibration become harder under domain shift. These findings indicate that trustworthy RAG systems require explicit validation between retrieval and final answer delivery. CiteGuard-RAG provides a practical architecture for linking retrieval, generation, citation checking, abstention, and regeneration in high-stakes information access. Comments: Submitted to Engineering Reports. 22 pages, 2 figures, 12 tables Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) Cite as: arXiv:2609.15830 [cs.CL] (or arXiv:2609.15830v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.15830 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-1] IROH: Insightful Ranking Of Humor using Multi-Stage Hybrid Retrieval with Rationale-Distilled LLM Judges for JOKER 2026 Track Task 1 English
链接: https://arxiv.org/abs/2609.15618
作者: Ana-Maria Luisa Mocanu,Sebastian Mocanu,Ciprian-Octavian Truică,Elena-Simona Apostol
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注:
Abstract:Our team, VANGUARD, presents IROH (Insightful Ranking of Humor), a three-stage retrieval system for JOKER Task 1 English at CLEF 2026, achieving first place on the leaderboard with 0.6347 MAP. Our pipeline combines hybrid sparse-dense retrieval, cross-encoder reranking, and a LoRA-adapted Large Language Model judge ensemble. We employ Gemma 4 to generate query-aware rationales under two prompt strategies, generic and typed, and produce up to four types of structured hard negatives for training data construction. Through an ablation across three cross-encoder architectures, four dense embedders, and eight judge configurations, our key findings are threefold: (1) the rationale-distilled judge is the primary driver of ranking quality, whereas appending rationales to the first-stage index contributes negligibly; (2) structured hard negatives degrade generalisation in nearly all configurations despite inflating local validation scores; and (3) across the components we ablate, the lighter, better-calibrated model is competitive with or stronger than its larger counterpart, with the generic-rationale Qwen2.5-7B judge (0.6055 MAP) outperforming every Gemma-4-31B configuration, and the advantage of generic over typed rationales is concentrated almost entirely in the smaller model.
[IR-2] Self-Evolving Memory for Generative Recommendation CIKM’26
链接: https://arxiv.org/abs/2609.15598
作者: Xinyu Lin,Zhuosong Jiang,Zixiao Suo,Siqin Wang,Hanqing Zeng,Hanchao Yu,Yinglong Xia,Jiang Zhang,Aashu Singh,Fei Liu,Wenjie Wang,Fuli Feng,Yang Song,Qifan Wang,Tat-Seng Chua
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
备注: Accepted to CIKM’26
Abstract:Generative recommendation has emerged as a promising end-to-end paradigm for personalized recommendation. However, user preferences continuously evolve over time, making self-evolving an essential capability for generative recommender systems. Existing evolving strategies, such as continual retraining and distillation-based adaptation, directly update the shared model parameters using streaming interactions. Nevertheless, we find that directly applying such strategies to generative recommendation introduces a critical issue, termed evolution conflict. Specifically, heterogeneous preference shifts from different users are optimized within a fully shared autoregressive parameter space, causing dominant behavioral patterns to progressively dominate the model evolution process while underrepresented patterns become increasingly overlooked. To address this issue, we propose a self-evolving memory paradigm for generative recommendation, aiming to enable effective evolution across heterogeneous behavioral patterns. We further identify three key principles for effective self-evolving recommendation systems, including isolated memorization, reinforced evolution, and scalable application. Guided by these principles, we develop LION, a simple yet effective framework centered on a sparse Key-Value memory layer. Specifically, LION introduces sparse memory activation to isolate the evolution of different behavioral patterns, while a consolidation loss is designed to reinforce the learning of underrepresented preference dynamics during continual adaptation. Extensive experiments on diverse real-world datasets demonstrate the effectiveness of LION under various continual evolution settings (e.g., per-period evaluation, user/item group evaluation, and evolution convergence analysis). The codes are released at this https URL.
[IR-3] he Magnitude Mirag e: Rethinking Confidence for Reasoning -Intensive Retrieval EMNLP2026
链接: https://arxiv.org/abs/2609.15578
作者: Jamie Holdcroft,Abdelrahman Abdallah,Adam Jatowt
类目: Information Retrieval (cs.IR)
备注: Accepted at EMNLP 2026
Abstract:Many production RAG systems implement retrieval abstention by thresholding raw similarity scores, implicitly treating score magnitude as a confidence signal. We demonstrate that this practice degrades systematically as queries require reasoning beyond semantic matching. Across 11 retrieval architectures and 28 datasets, neural retrievers consistently assign high similarity scores to semantically related but constraint-violating documents, causing magnitude-based thresholds to collapse toward near-random abstention performance on logical and temporal reasoning tasks—a failure we term the Magnitude Mirage. To address this without computationally expensive alternatives, we conduct a large-scale empirical study of six zero-cost Query Performance Prediction (QPP) metrics across three cognitive tiers: semantic matching (BEIR), logical reasoning (BRIGHT), and temporal reasoning (TEMPO). Our central finding is that the key improvement comes from abandoning magnitude in favor of score-distribution signals: the gain from this shift exceeds the differences among distributional alternatives by a factor of 5-10 \times . In particular, Score Gap ( s_1 - s_k ) and a practical adaptation of Score Magnitude and Variance (LSMV) improve abstention AUROC by up to 0.16 in settings where magnitude-based confidence provides little discriminative power. These methods require no additional inference, retraining, or latency, making them a practical zero-cost replacement for magnitude thresholding in deployed RAG systems.
[IR-4] Beyond Retrieval: Scaffolding Childrens Online Learning
链接: https://arxiv.org/abs/2609.15568
作者: Diletta Micol Tobia,Hrishita Chakrabarti,Maria Soledad Pera,Monica Landoni
类目: Information Retrieval (cs.IR); Human-Computer Interaction (cs.HC)
备注: This is the author’s version of the work. It is posted here for your personal use. This work was presented at ACM-W WomENcourage 2026, September 30-October 2, 2026, Sophia Antipolis, France
Abstract:Children increasingly turn to online information access systems that are primarily designed for the mainstream population, e.g., adults, but possess a limited understanding of how these systems work, contributing to their unstructured and ineffective search practices. This lack of knowledge can hinder their curiosity and the development of critical search skills. Grounded on the existing literature of both child-oriented Information Retrieval and Human Computer Interaction, our work positions children as active participants in the search process, framing it as a scaffolded learning experience rather than a simple retrieval task.
[IR-5] Benchmarking Embedding Models for ESG Data
链接: https://arxiv.org/abs/2609.15434
作者: Motaz Saad,Veronica Cretì,Ivan Gentile,Kianna Kazemi,Antonella Longo
类目: Information Retrieval (cs.IR)
备注:
Abstract:The use of Environmental, Social, and Governance (ESG) data is fundamental for modern corporate accountability, sustainability reporting, and financial decision-making. Embedding models have emerged as a powerful approach for transforming unstructured ESG text into numerical representations suitable for downstream natural language processing (NLP) tasks. However, their effectiveness in these ESG-specific tasks has not been systematically studied. In this paper, we construct a benchmark dataset specifically tailored to the ESG domain. We benchmark fourteen models, both open-source and closed-source embedding models, comparing their performance with respect to retrieval, and Retrieval-Augmented Generation (RAG). The results demonstrate performance variations across different models, with Qwen3-based models achieving the highest overall performance. This study provides practical insights into which models are better suited for ESG RAG tasks.
[IR-6] Clean Scores Buried Evidence and Confident Wrong: A Receipt-Based Audit of Frontier Agent ic QA
链接: https://arxiv.org/abs/2609.15319
作者: Luis M. Sánchez
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); General Finance (q-fin.GN)
备注: 28 pages, 9 figures. Frozen evidence archive: this https URL
Abstract:Frontier models score well on shallow document/chart reading tasks. In a controlled data-room audit, moving evidence into buried conditions reduced accuracy, increased forced declarations, increased tool calls, and increased cost per correct answer. Confidence and benchmark calibration did not fully capture wrong answers; a documented production incident shows fabricated structural claims can be mixed with accurate numeric tables. Agentic evaluations need claim-level receipts (statement-level provenance, not answer-level scores), condition-aware scoring, and human-adversarial verification - an auditing discipline, not a leaderboard. The setting we measure is financial due diligence; the setting we are building toward next is defense staff work, where the same buried-evidence shape appears. In both, the model is not a party to the consequences; the person who signs is. In plain terms: in the documented cases we examine, agents can pair accurate numbers with confident fabricated explanations, and the burden of proof must therefore move from the model to the evidence trail.
[IR-7] ProLiVis 2.0: Literature-Centric Visualization of Protein–Protein Interaction Networks with a Citation-Trust Model for Interaction Evidence
链接: https://arxiv.org/abs/2609.15236
作者: Melih Sözdinler,Yalçın Doksanbir,Gökhan Akpınar,Ege Aktan
类目: Information Retrieval (cs.IR); Social and Information Networks (cs.SI); Molecular Networks (q-bio.MN)
备注: 8 pages, 6 figures, Github Repo: this https URL , Supplement/Guide is available on repo
Abstract:Protein-protein interaction databases record evidence without weighing it. In BioGRID, an interaction asserted once by a single high-throughput screen and one confirmed by twenty laboratories across a dozen assays are the same kind of row in the same file. Tools built on such databases inherit that flattening: they draw every reported interaction as an edge, and the resulting picture states that two proteins interact without stating how much anyone should believe it. We present ProLiVis 2.0, a rewrite of the literature-centric visualization system of arXiv:2111.12794. It contributes three things. First, a citation-trust model that scores each interaction from seven terms, including a term for the number of independent laboratories behind the supporting publications, obtained by clustering those publications over shared institutional affiliations; a plain count of publications cannot distinguish five confirmations from one group publishing five times. Second, a deterministic reformulation of the center layout, closed-form and O(n \log n) , which replaces the force-directed placement of the original and makes published figures regenerable from a session manifest. Third, an implementation that runs entirely in a web browser, with an embedded analytical database, requiring no installation and uploading no data. On BioGRID release 5.0.260 restricted to SARS-CoV-2, 24,344 of 34,540 reported interactions (70%) rest on a single publication, and raising the trust threshold to 0.2 leaves 11,320 of them. That the large majority of a curated interaction network is unreplicated is a fact no existing view of the database makes visible. Comments: 8 pages, 6 figures, Github Repo: this https URL, Supplement/Guide is available on repo Subjects: Information Retrieval (cs.IR); Social and Information Networks (cs.SI); Molecular Networks (q-bio.MN) MSC classes: 92C42, 05C85, 68P20 ACMclasses: H.3.3; J.3; H.5.2; G.2.2; H.3.7 Cite as: arXiv:2609.15236 [cs.IR] (or arXiv:2609.15236v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2609.15236 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-8] op-K Is Not a Budget for Hybrid Retrieval
链接: https://arxiv.org/abs/2609.15143
作者: Chunran Zhang
类目: Information Retrieval (cs.IR)
备注: 5 pages, 2 figures, 3 tables. Code: this https URL
Abstract:Modern hybrid retrieval for RAG typically fuses the Top- L results from dense and sparse retrievers, but a fixed truncation depth may not transfer across changing queries and corpora. Exact fusion removes the dependence on a fixed depth, yet completing a specified Top- K still incurs variable access costs. We present DiBud, which takes an access budget directly as input and incrementally certifies and returns an exact prefix of the RRF ranking over the full lists. Selective access increases certified output within the budget, while budgeted stopping bounds accesses per request. Experiments on five query sets reveal long-tailed costs for completing exact Top-20. At a budget of 2048 accesses, DiBud increases mean certified output within the first 100 positions by 7.86% over balanced access. After budget calibration for 95% quality retention, held-out queries retain 95.05%–97.68% of mean nDCG@20 while using 65.92%–99.53% fewer accesses than completing exact Top-20.
[IR-9] Generate to Explore Select to Exploit: Aligning LLM -based Headline Generation with Personalized Recommendation
链接: https://arxiv.org/abs/2609.15094
作者: Yi Chen,Rufeng Cheng,Qiang Xie,Tao Li
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
备注:
Abstract:In industrial recommendation feeds, presenting a static headline for an item often fails to satisfy the diverse, multimodal interests of the user population, particularly suppressing the needs of long-tail audiences. While Large Language Models (LLMs) have been integrated into recommendation for content understanding or ranking, directly optimizing them to output a single best headline typically leads to mode collapse—converging to generic patterns that satisfy average tastes but miss specific latent intents. To bridge this gap, we introduce GESE (Generate to Explore, Select to Exploit), a framework operating at the system’s presentation layer that decouples personalization into generative exploration and selective exploitation. First, we treat the LLM as a probabilistic explorer, utilizing Group Sequence Policy Optimization (GSPO) with a hierarchical reward mechanism to generate a candidate set that maximizes the semantic coverage of potential user interests. Subsequently, a lightweight, real-time feedback-aware selector acts as the exploiter, identifying the optimal realization from the candidate pool based on instant contextual signals. Extensive deployment on a commercial platform with over 100 million daily active users demonstrates that GESE significantly outperforms state-of-the-art baselines, achieving a 2.57% lift in CTR and 0.87% in dwell time. These results validate that decoupling diversity-oriented generation from precision-oriented selection offers a robust blueprint for aligning generative AI with dynamic user utility.
[IR-10] Converting Sequenced Fuzzy Cognitive Maps to Causal Virtual Worlds with Large Video Generators
链接: https://arxiv.org/abs/2609.14985
作者: Akash Kumar Panda,Olaoluwa Adigun,Bart Kosko
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 9 Figures. For the generated FCM Dolphin-Shark video, see this https URL
Abstract:We show how users can create and manipulate causal virtual worlds with large-language-model (LLM) and large-video-model agents. The approach uses feedback fuzzy cognitive maps (FCMs) both to model the granular causal structure of the virtual world and to guide its causal evolution. The local causal rules are partial or fuzzy while the FCM’s feedback structure produces global equilibria that define causal scenarios. A sequence of \emphdynamical meta-rules of the form ``If \mathcalA then \mathcalB " define the causal scenes of the virtual-world video. The if-part causal pattern \mathcalA perturbs the FCM’s virtual world at the user’s or agent’s discretion. The FCM’s transient feedback dynamics define the meta-rule’s causal arrow of implication. The then-part \mathcalB is the resulting equilibrium attractor such as a FCM limit cycle or fixed point. Our algorithm extracts these meta-rules from the FCM and guides the LLM agent to write a script based on the FCM meta-rule sequence. The large video generator converts the meta-rule into a video scene in accord with the flow of the dynamics. We applied the agent-based technique to a simple FCM that describes an undersea world of dolphins and sharks. Google’s Gemini 3.1 generated the script and Google’s Veo 3.1 generated the dolphin-shark video. The approach is general and can scale by mixing larger FCMs and AI agents to produce more immersive virtual worlds.
[IR-11] LazFormer: Scaling Transformers for Industrial Recommendation via Transferable Generative Pre-training
链接: https://arxiv.org/abs/2609.14978
作者: Xiaodong Li,Alin Fan,Mingyang Li,Yan Xiao,Shichao Nie,Junfeng Zhang,Shaochuan Lin,Zhanming Ou,Tao Luo,Xiaoyi Zeng
类目: Information Retrieval (cs.IR)
备注:
Abstract:Transformers have shown promising performance in LLMs due to their outstanding scalability, several studies have investigated the scalability of Transformers for industrial recommendation. They typically rely on a single ranking model to optimize both sparse and dense parameters from scratch, resulting in substantial computational resource consumption and slow convergence. Fortunately, the pre-training models offer an effective solution to the above issues by providing favorable initialization of both sparse and dense parameters for the subsequent ranking. However, they still face two major limitations: (1) Since the input features used in pre-training and ranking are usually inconsistent, directly transferring dense parameters from pre-training to ranking may lead to negative transfer. (2) Multi-epoch training during the ranking process may result in the overfitting of sparse parameters, while freezing the sparse parameters limits their adaptability to the ranking objectives. To this end, we propose a Scaling Transformer for Industrial Recommendation via Transferable Generative Pre-training, termed LazFormer. Specifically, we first present a generative pre-training module to autoregressively generate sequential features, providing favorable initialization of both sparse and dense parameters for the subsequent ranking. To solve the negative transfer of dense parameters, we propose a transferable residual adapter that injects additional ranking-specific features into ranking in a residual manner. Moreover, a request-aware ranking module integrates long-sequence compression, hybrid sparse attention, and a request-aware paradigm to efficiently model users’ long sequences. Besides, we further propose an asymmetric multi-epoch training strategy that resets sparse parameters while continuously accumulating dense parameters across epochs, alleviating the overfitting of sparse parameters.
[IR-12] Route Me If You Can: A Benchmark for Query Reformulation Selection
链接: https://arxiv.org/abs/2609.14885
作者: Hai Son Le,Negar Arabzadeh,Amin Bigdeli,Radin Hamidi Rad,Sajad Ebrahimi,Charles L. A. Clarke,Ebrahim Bagheri
类目: Information Retrieval (cs.IR)
备注:
Abstract:LLM-based query reformulation can improve retrieval, but no single reformulation strategy is consistently optimal across queries, domains, retrievers, or model backbones. This creates an inference-time decision problem: ``Given an original query and a pool of candidate reformulations, which one should be issued to the retriever?‘’. Existing studies are hard to compare because they use different reformulator pools, retrievers, relevance signals, training labels, and evaluation metrics. We introduce QueryRoute, a benchmark that freezes the expensive artifacts needed to study this decision reproducibly: original queries, generated variants, ranked lists under multiple retrievers, retrieval scores, and per-query oracle labels. The benchmark contains 3,757 queries, 11 candidate systems, five reformulator backbones, and three retrievers across TREC DL, BEIR, and BRIGHT, yielding 619,905 retrieval outcomes. We benchmark supervised classification, routing, QPP, and LLM-as-judge selectors. Results show substantial oracle headroom over fixed reformulators, but current selectors recover only part of it; selector rankings change across retrievers, and similar mean effectiveness can hide different query-level behavior. The released artifacts and evaluation harness allow future selectors to be compared without regenerating variants, rerunning retrieval, or rebuilding judge pipelines. Code and data are available at this https URL
[IR-13] EviQE: Evidence Selection for LLM -Based Query Expansion
链接: https://arxiv.org/abs/2609.14875
作者: Hai Son Le,Amin Bigdeli,Shirin Seyedsalehi,Morteza Zihayat,Ebrahim Bagheri
类目: Information Retrieval (cs.IR)
备注:
Abstract:LLM-based query expansion increasingly conditions reformulation on documents retrieved from the target corpus, yet most work focuses on how to generate expansions rather than which documents the model should read. We propose EviQE, which aggregates documents retrieved by multiple reformulators, selects a compact evidence set, and uses it for one grounded expansion step. This separates evidence selection from generation and treats reformulators as complementary retrieval perspectives. Across three TREC DL and five BEIR benchmarks, reformulators frequently retrieve distinct relevant documents, so pooled candidates provide higher relevant-document coverage than any individual source. The strongest gains come from relevance-based evidence selection: LLM-Score consistently outperforms direct reformulation, cold-start expansion, and single-source seeded expansion. Additional retrieval-generation rounds provide little benefit once strong conditioning evidence has been selected and can reduce effectiveness.
[IR-14] Speak to the City: Multimodal Resolution for Outside-the-Vehicle References
链接: https://arxiv.org/abs/2609.14691
作者: Alireza Parchami(1 and 2),Artin Saberpour(2),Robin Connor Schramm(1 and 3),Jürgen Steimle(2),Ulrich Schwanecke(3) ((1) Mercedes-Benz Tech Innovation GmbH, (2) Saarland University, (3) RheinMain University of Applied Sciences)
类目: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL); Information Retrieval (cs.IR); Machine Learning (cs.LG); Multimedia (cs.MM)
备注: 11 pages, 7 figures, 1 table; Accepted to the 18th International ACM Conference on Automotive User Interfaces and Interactive Vehicular Applications (AutoUI '26)
Abstract:As autonomous vehicles and Extended Reality (XR) headsets enable novel in-car interactions, seamlessly querying physical landmarks, known as Outside-the-Vehicle Referencing (OVR), remains challenging due to ego-motion and referential ambiguity. We present a robust, multimodal OVR framework fusing user gaze and natural language to identify Points of Interest (POIs). To address the scarcity of dynamic vehicular data, we developed a VR-based pipeline synchronizing 360-degree transit videos with vehicle GNSS telemetry. Through a user study (N=46) mapping passenger head orientation into a 3D geospatial Digital Twin, we captured authentic gaze-speech behaviors. We subsequently trained a lightweight Transformer network, leveraging LLMs to dynamically align continuous spatial gaze vectors with discrete verbal context. Experimental results demonstrate high accuracy and low computational overhead, achieving an 83.33% Top-1 accuracy (87.72% Top-2) and an average inference time of 24.3 milliseconds. This real-time paradigm effectively resolves referential ambiguity, enabling context-aware spatial retrieval for passengers within the vehicle.
[IR-15] Beyond Benchmark Scores: How Synthetic and Authentic Query Distributions Diverge in RAG Evaluation CIKM2026
链接: https://arxiv.org/abs/2609.14579
作者: Filip J. Kucia,Barbara M. Gawlik
类目: Information Retrieval (cs.IR)
备注: Accepted at CIKM 2026 (35th ACM International Conference on Information and Knowledge Management) as a Short Research Paper
Abstract:RAG systems are routinely evaluated using synthetic question sets generated from the target document corpus. While this practice provides a useful check on overall retrieval capability, relying exclusively on synthetic benchmarks can mislead under distribution shift and overstate deployment readiness. Synthetic generation spreads questions evenly across the corpus, formulating long, detailed queries; real users put most of their traffic on a few administrative and procedural topics in short queries, while also asking about matters the generator never covers at all. We demonstrate this gap on a university faculty information system, comparing 1,851 synthetic questions generated via Gemini Notebook against 322 authentic queries collected via a student survey. The synthetic and authentic query sets differ significantly: authentic queries average 6.8 words versus 15.7 for the synthetic ones, and draw from only 53 unique sources compared to 165. Consequently, configurations that appear highly effective on synthetic benchmarks experience a substantial performance drop on authentic queries. Importantly, optimizing on synthetic queries selected a higher-latency hybrid retriever. In our setting the sparse retrieval component benefited long synthetic questions but not short authentic ones, costing up to 8\times the latency of the fastest configuration we tested. We propose treating synthetic and authentic query sets as complementary extremes of the query-quality spectrum: synthetic data verifies maximum retrieval capacity under idealized conditions, while authentic queries test system robustness to the imprecise, underspecified inputs of real users.
[IR-16] he Wisdom of the Loudest: A Large-Scale Audit of Generative Search on Reddit
链接: https://arxiv.org/abs/2609.14575
作者: Agam Goyal,Wang Claire,Eshwar Chandrasekharan
类目: Information Retrieval (cs.IR)
备注: 23 pages, 7 figures, 2 tables
Abstract:Online communities are valued not only for answers, but for the diversity of experiences and perspectives they contain. Generative search increasingly mediates access to this discourse, yet little is known about which community voices survive retrieval and synthesis. We audit Reddit Answers using 10,000 queries from 20 advice- and support-seeking communities, repeated three times to produce 30,000 answers over 14.68M comments. We find that differences across runs are driven primarily by retrieval, answers routinely combine evidence across communities, and selection strongly favors already-visible, top-level comments. Formal and directive language is more likely to be surfaced, while experiential voice is less likely to survive selection and is further weakened during synthesis, with first-person singular language declining sharply. These findings show that community-grounded generative search is not neutral summarization, and should be designed not only for relevance and fluency, but also for provenance, plurality, and legibility.
[IR-17] AlgoRAG : Retrieval-Augmented Generation for Theoretical Computer Science Education – A Comprehensive Evaluation Framework for Algorithm Analysis and Complexity Theory
链接: https://arxiv.org/abs/2609.14572
作者: Sushan Adhikari
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 13 pages, 2 figures, 2 tables. Code and dataset available at: this https URL
Abstract:Teaching abstract theoretical computer science (TCS) concepts such as algorithm analysis and complexity theory is challenging because students must handle formal proofs and asymptotic reasoning that conventional resources rarely explain in an adaptive, on-demand way. We present AlgoRAG, a specialized Retrieval-Augmented Generation (RAG) system that couples a large language model (LLM) with a curated, domain-specific knowledge base to address these challenges. The knowledge base integrates authoritative textbooks, 847 lecture slides, 312 practice problems with solutions, 156 worked proof templates, and 89 complexity worksheets. AlgoRAG incorporates domain-specific optimizations including mathematical entity recognition, notation-aware retrieval, and pedagogical re-ranking. We evaluate AlgoRAG on 179 curated exam-style questions spanning asymptotic analysis, recurrence relations, dynamic programming, graph algorithms, NP-completeness, sorting, and divide-and-conquer. The system achieves a 100% success rate with a mean response time of 38.0 seconds. While BLEU-4 scores are zero – a known limitation of n-gram matching on mathematical proofs where equivalent reasoning may use entirely different notation – AlgoRAG attains ROUGE-1 F1 of 0.0963, ROUGE-L F1 of 0.0683, and a pedagogical quality score of 0.7620, indicating that responses are well-structured and didactically sound even when surface wording diverges from reference answers. Performance is especially strong on NP-completeness (ROUGE-1 F1 = 0.1285, pedagogical quality = 0.7643) and graph algorithms (ROUGE-1 F1 = 0.1023, pedagogical quality = 0.8086). These results support the conclusion that RAG is an effective architecture for personalized theoretical-CS instruction, providing correct, context-rich explanations even for highly abstract topics.
[IR-18] VARG: Value-Aware and Ranking-Aligned Generative Retrieval for Dynamic E-commerce Search
链接: https://arxiv.org/abs/2609.14493
作者: Xiaopeng Chu,Jianbo Zhu,Mingmin Jin,Jing Wang,Xing Fang,Wenyi Zhang
类目: Information Retrieval (cs.IR)
备注: 11 pages, 4 figures, 7 tables
Abstract:Integrating recall and pre-ranking in e-commerce search requires candidate generation to account for relevance, personalization, and business value before final ranking. To this end, we present VARG, a generative retrieval system for Tmall App search that directly admits generated item candidates to the existing final ranker. VARG-ID constructs semantic prefixes using RQ-VAE, enhances search relevance through bidirectional query-item contrastive learning, and combines these prefixes with a value-ordered third token to provide fine-grained item addresses and a business-value prior. Three-stage supervised fine-tuning progressively learns item-to-identifier mappings, query-semantic retrieval, and personalized retrieval. Personalized model training combines value-aware and hierarchy-aligned supervision with expanded user context, and uses local ordinal supervision (LO-SFT) to learn the local within-cluster ordering encoded by the third token. Prefix-GRPO combines gated rewards based on output legality, user behavior, ranker advantage, and search relevance with prefix-aware token weighting to align candidate generation with business value and ranking objectives. Coordinated daily product and model updates preserve existing item addresses while incorporating new products and behavioral feedback. Offline experiments on tens of millions of products validate identifier stability and demonstrate gains in retrieval quality and head-level value recall from SFT strategies and Prefix-GRPO over their respective baselines. In a 14-day online A/B test covering 20% of search traffic, VARG directly admits generated candidates to the final ranker and improves GMV by 1.45%, per-user IPV by 0.22%, and PCTR by 0.31%. Online shopping-guide query evaluations further show that VARG maintains competitive relevance with a smaller candidate quota.
[IR-19] Questions Gambit: The First Move Matters in Agent ic Deep Search
链接: https://arxiv.org/abs/2609.14412
作者: Radin Hamidi Rad,Amin Bigdeli,Negar Arabzadeh,Sajad Ebrahimi,Charles L. A. Clarke,Benjamin C. M. Fung,Ebrahim Bagheri
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:Deep research agents answer complex questions through iterative loops of searching, reading, and reasoning. Recent work on reasoning-intensive benchmarks such as BrowseComp-Plus shows that well-configured lexical retrieval can surface high-quality evidence, yet agents may still fail to connect documents carrying evidence to the gold documents. We identify a deep research agent’s first retrieval move as an important design decision for this setting. We introduce Question’s Gambit, a first-move retrieval module that decomposes the question into a set of clues, reformulates them into complementary searches, consolidates the retrieved results, and reranks the candidate pool before the agent begins its iterative search-and-reasoning process. This produces an opening context designed to support both clue aggregation and final-answer verification. We further evaluate on MultiHop-RAG to test whether these benefits transfer beyond BrowseComp-Plus to a more conventional multi-hop question structure. Experiments on BrowseComp-Plus show that Question’s Gambit improves retrieval recall and downstream agent accuracy over strong baselines, improving answer accuracy from 83.1% to 90.5% with gpt-5.5 over Pi-Serini, the strongest reported agentic baseline. Our results confirm that effective agentic deep research depends not only on the tools available inside the loop, but also on the quality of the first move. We published our implementation publicly at this https URL.
[IR-20] Semantic Knowledge Technologies: what the Semantic Web lost sight of and what it never had
链接: https://arxiv.org/abs/2609.14121
作者: Achille Zappa
类目: Artificial Intelligence (cs.AI); Databases (cs.DB); Information Retrieval (cs.IR)
备注:
Abstract:The Semantic Web set out to give information a machine-interpretable form so that software could integrate and reason over it. Its standards became scientific knowledge infrastructure, but the machine competence it promised did not follow, and the systems now answering questions over scientific knowledge are language models holding no inspectable account of what they know. This paper argues the original goal was right and the technical programme incomplete, states what is missing, and names the extended programme Semantic Knowledge Technologies: the same technical core carried out of its web-publishing origin and applied to knowledge wherever held. The diagnosis is that the standards formalised truth while omitting three things: the conditions under which a claim holds, the operations its terms permit, and any account of what a base covers. Without conditions, contradiction and applicability cannot be judged; without operational grounding, holding a statement confers no ability; without declared coverage, a system cannot recognise the boundary of its own content, which under the open-world assumption cannot be inferred. The paper fixes the word understanding to five measurable tests (check, connect, derive, act, delimit) and sets out a seven-layer architecture in which the first three layers are enabling and the rest the cognitive capabilities they make possible. It then defines three terms the programme implies: Large Knowledge Model, a model whose unit of output is a reference to an addressable claim, not a token; SLKM, the knowledge base an agent builds for itself from declared sources; and Semantic Artificial General Intelligence, stated as a falsifiable position about necessary conditions, not a system. A graded ladder replaces the untestable word general. It is offered as a research agenda, with its weakest points and refutation condition named.
[IR-21] F-IDF and BM25 Are Exact KL Divergences
链接: https://arxiv.org/abs/2609.14016
作者: Ivan Silajev
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注: 5 pages
Abstract:TF-IDF and BM25 are two of the most widely used methods for scoring query-document relevance, yet neither has a standard probabilistic derivation that justifies it as a statistical method within a unified framework. We address this gap by showing that both scoring methods admit an exact interpretation as Kullback-Leibler divergences between two probability models. We treat the BM25 variant that includes the plus 1 correction in the IDF term, which is the one used in practice, and also discuss the original BM25 formulation without that correction. The resulting framework provides a common theoretical basis for TF-IDF and BM25, clarifies what they measure, and allows them to be compared theoretically with other information retrieval methods rather than only experimentally.
[IR-22] P3Rec: Distilling Prior–Posterior Preference Reasoning for LLM -based Recommendation
链接: https://arxiv.org/abs/2609.13993
作者: Jinfei Chen,Weihai Lu,Jiawei Cheng
类目: Information Retrieval (cs.IR)
备注:
Abstract:Large language models (LLMs) exhibit strong semantic understanding and preference reasoning capabilities, offering new opportunities for user modeling in recommender systems. Existing LLM-as-Enhancer methods typically distill LLM-derived preference knowledge into lightweight recommenders to avoid costly online LLM inference. However, they often construct distillation knowledge from only one perspective. Prior preference captures users’ stable and consistent interests but provides limited guidance for the current decision, whereas posterior preference reveals target-relevant fine-grained interests but may rely excessively on target clues. To address these limitations, we propose P ^3 Rec, a framework that jointly extracts and internalizes complementary prior and posterior preference reasoning knowledge. Specifically, P ^3 Rec first derives target-agnostic prior preferences and target-conditioned posterior preferences from the user side, while further extracting item-centric preference representations from item semantics and predecessor interactions. It then progressively internalizes prior and posterior knowledge into behavioral representations through prior preference absorption and posterior-guided preference distillation. Since the resulting comprehensive preference representation may not always provide an equally decisive retrieval direction, P ^3 Rec further characterizes historical interest dispersion with interest entropy and adaptively calibrates the user representation before contrastive retrieval optimization. In this way, P ^3 Rec achieves more complete preference reasoning while preserving efficient recommendation. Extensive experiments on multiple public datasets demonstrate its effectiveness.
[IR-23] ClinAgent : A ReAct-Based Agent for Conversational Access to Clinical Trial Information
链接: https://arxiv.org/abs/2609.13860
作者: Antonino Vaccarella,Riccardo Cantini,Domenico Talia,Paolo Trunfio,Marianna Talia,Rosamaria Lappano,Marcello Maggiolini
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: Accepted at the CIBB 2026 conference ( this https URL )
Abstract:Querying clinical trial registries remains a manual and error-prone process, requiring researchers to navigate large volumes of semi-structured data without support for natural language interaction or cross-source synthesis. To address this, we introduce ClinAgent, a conversational system based on agentic Retrieval-Augmented Generation (RAG) that enables clinicians and researchers to query clinical trial information in plain language and receive grounded, up-to-date responses across multi-turn interactions. The system centers on a Large Language Model (LLM) agent following the ReAct paradigm, which iteratively reasons over queries, selects among a set of integrated tools, and refines its actions based on intermediate outputs. These tools include a this http URL search interface, a PubMed module, and a Python-based analyzer operating on a locally cached structured dataset of clinical trials. We evaluate the system using a three-phase framework assessing operational effectiveness, planning quality, tool-use efficiency, and expert qualitative judgments, comparing three LLM backends: Gemini 3.0 Flash and two variants of DeepSeek V3.2 (thinking and non-thinking). Results reveal complementary strengths, with DeepSeek (thinking mode) excelling in planning quality, while Gemini achieves the highest overall performance and strongest expert ratings. Overall, our findings highlight the potential of agentic AI systems to improve the accessibility and synthesis of clinical trial information, supporting more efficient and user-centered biomedical research workflows.
[IR-24] Odds-Shift Slippage in One-vs-Rest Rankers: Diagnosing and Repairing Reweighting-Induced Top-K Errors
链接: https://arxiv.org/abs/2609.13810
作者: Akifumi Goto
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 10 pages, 4 figures, 2 tables; 9-page supplement as ancillary file. Code, result files and paper sources: this https URL (Zenodo DOI https://doi.org/10.5281/zenodo.22719148 )
Abstract:One-vs-rest rankers that show each user the top- K of many rare labels usually counter imbalance with a per-label positive-class weight, scale_pos_weight = n_-/n_+ . Elkan’s identity says such a weight shifts label j 's log-odds by \ln w_j , so the model ranks by weighted odds rather than by the marginal that is Bayes-optimal for precision@ K , and suggests inverting the shift afterwards; what a finite learner does with a weight in the thousands, and which repair then works, has not been measured. We call the gap between the promised and the realized shift odds-shift slippage and measure it on matched pairs of LightGBM and MLP models that differ only in the weights. On Santander the weight takes MAP@7 from 0.808 to 0.117; for the boosted pairs the ideal odds shift accounts for 23% of that loss (32% on Instacart; 98% for an MLP pair on the same rows) and slippage for the rest. We prove that a booster whose leaf steps are capped at c realizes at most T\eta c nat of shift in T rounds at rate \eta , which a cap sweep confirms, and show that without a cap saturated cells tie at exactly 1.0, beyond the reach of any separable map. The analytic inversion therefore pays only where the shift was realized and nothing saturated, whereas per-label isotonic regression returns the Santander model to 0.784 (0.780 with the calibrator fitted on the validation period), but only if labels without calibration positives are mapped to their prior rather than passed through. On 11 public MULAN benchmarks and 5 learners the weighted model loses more than half of its MAP@ K in 8 of 55 cells, and on delicious and Corel5k the same repair returns it to the unweighted level; per-label calibration hurts where positives are scarce, a harm that a cross-validated rule removes. The recipe is released as oddslip.
[IR-25] Addressing Cross-Stage Decoupling of Semantic and Collaborative Signals in Generative Recommendation RECSYS2026
链接: https://arxiv.org/abs/2609.13678
作者: Jiayi Dan
类目: Information Retrieval (cs.IR)
备注: Accepted by RecSys 2026 Main Track
Abstract:Generative recommendation reformulates sequential recommendation as autoregressive generation by encoding items into semantic tokens, enabling improved scaling capability and cross-domain generalization. However, existing generative recommender systems typically follow a two-stage pipeline, where item tokenization is largely dominated by textual semantics with limited incorporation of collaborative signals and interaction similarity, leading to code assignments that are misaligned with downstream generation. Conversely, the generation stage tends to overlook the original semantic information, as the code sequences are re-embedded based on interaction data. This cross-stage information decoupling limits semantic coherence and recommendation accuracy. To address this issue, we propose SCRec, a general framework that enhances cross-stage coherence through bidirectional information supplementation. Specifically, we introduce (i) collaborative-enhanced tokenization to explicitly inject textualized collaborative signals into semantic tokenization, without introducing additional alignment task, (ii) semantic-guided generation to dynamically recalibrate semantic priors with learnable code embeddings in generation stage, and (iii) manifold alignment to reconcile the geometric mismatch between the embedding space of discrete codebook indices and the dense continuous semantic space. These interrelated components form a general framework that aligns semantic and collaborative signals and enhances cross-stage information coherence, with minimal additional training and inference costs. Extensive experiments demonstrate the effectiveness, robustness, and generalizability of our proposed framework. Comments: Accepted by RecSys 2026 Main Track Subjects: Information Retrieval (cs.IR) Cite as: arXiv:2609.13678 [cs.IR] (or arXiv:2609.13678v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2609.13678 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.1145/3773078.3831736 Focus to learn more DOI(s) linking to related resources
[IR-26] Cost Characterization of Vertically Partitioned Federated Knowledge Graphs ISWC2026
链接: https://arxiv.org/abs/2609.13664
作者: Md Saikat Islam Khan Bappy,Oshani Seneviratne
类目: Artificial Intelligence (cs.AI); Databases (cs.DB); Information Retrieval (cs.IR)
备注: Accepted at DMKG’26: 2nd International Workshop on Data Management for Knowledge Graphs, co-located with ISWC 2026; to appear in CEUR-WS proceedings
Abstract:Knowledge graphs are increasingly distributed across autonomous organizations that share an entity space but own disjoint subsets of relations, forming a vertical partition. Answering a multi-hop query may require combining facts from several silos, making the partitioning strategy a key data management decision that affects communication, indexing, load balance, and query latency. However, the costs associated with different partitioning strategies remain insufficiently studied. We formalize vertical partitioning as a design space and compare four strategies: semantic domain grouping, frequency-balanced partitioning, co-occurrence graph-cut partitioning, and random partitioning. We evaluate them using five metrics: communication cost, candidate index size, cross-silo path length, load balance, and end-to-end query latency. Three of the five prove to be determined by the graph and the silo count rather than by the partition, which reduces the design problem to two conflicting axes, cross-silo path length and load balance. Experiments on MetaQA and PathQuestion use a fixed federated knowledge graph question-answering architecture based on TransE embeddings and a frozen BERT encoder across three silo configurations. By keeping the learning model unchanged, we isolate the effect of partitioning and show that the trade-off between locality and balance holds only where each silo can hold several relations, weakening as the number of silos increases. The study provides practical guidance for deployments constrained by cross-silo reasoning or by silo load.
[IR-27] FedV-KGQA in Practice: Design Lessons and an Interactive Prototype ISWC2026
链接: https://arxiv.org/abs/2609.13661
作者: Md Saikat Islam Khan Bappy,Oshani Seneviratne
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: Accepted at the ISWC 2026 Posters and Demos Track; to appear in the ISWC 2026 Companion Volume (CEUR-WS)
Abstract:Knowledge graph question answering usually assumes that one system can reach the whole graph. In practice, facts are often held by organizations that share entity identifiers but own disjoint relation types, so no single party sees a complete reasoning chain. This poster presents the empirical findings of FedV-KGQA on multi-hop question answering over such vertically partitioned graphs. Each silo enriches its local graph and trains a knowledge graph embedding on its own triples. A server then concatenates the silo-specific entity views, anchors the projected question at the topic entity, and ranks candidates by similarity. Raw triples and relation embeddings never leave a silo. Comparing the FedV-KGQA experiments with one another yields three results. First, federated fusion recovers most of the centralized accuracy, while a single silo recovers little. Second, anchoring and enrichment matter more than the choice of embedding model. Third, the cheapest encoder depends on the target accuracy rather than on parameter count. This poster paper contributes that cross-experiment comparison, four design lessons drawn from it, and an interactive prototype that runs real inference and traces the full pipeline, per question, on released checkpoints.
[IR-28] YOLO12-MambaScan: An Efficient Object Detector with High-Frequency Enhancement and State-Space Modeling
链接: https://arxiv.org/abs/2609.13647
作者: Hao Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Information Retrieval (cs.IR)
备注:
Abstract:The rapid development of unmanned aerial vehicle (UAV) technology has made aerial-image object detection increasingly important for natural-resource monitoring, traffic management, and disaster response. Detecting small objects in aerial images remains difficult because objects occupy very few pixels, high-frequency cues are easily lost, and global context is hard to model in cluttered scenes. Existing detectors often retain insufficient edge, corner, and texture information. We propose \ours, an aerial-image detector built on the YOLO12 architecture. The model combines a triple-path high-frequency enhancement convolution module (TriPathHFConv), receptive-field coordinate-attention convolution (RFCAConv), and a Mamba-based global-context module. On VisDrone, at an input resolution of 960*960, ours achieves 60.0% mAP@50 and 38.6%mAP@50:95, demonstrating a favorable accuracy–efficiency trade-off for small-object detection. The benchmark and dataset protocol follow the VisDrone challenge setup.
[IR-29] Pre-retrieval Query Clustering for Adaptive Top-k Document Retrieval in RAG Systems CIKM2026
链接: https://arxiv.org/abs/2609.13489
作者: Ye Xia,Emre Yamangil,Haixun Wang
类目: Information Retrieval (cs.IR)
备注: Accepted to the Applied Research Track of CIKM 2026
Abstract:RAG systems commonly retrieve a fixed number of documents (top-k) to ground generation, but this static approach is brittle: simple queries suffer over-retrieval (adding noise and cost) while complex queries are under-retrieved, causing recall failures that cascade into incorrect answers. Motivated by the question of how many documents must be retrieved to answer an arbitrary query reliably, we propose a practical, general framework for query-adaptive retrieval depth. Offline, we estimate per-query retrieval difficulty by measuring NDCG under the default retriever and deriving a query-specific “saturation” point k* from the NDCG-k curve. Because computing these signals online is expensive, we cluster a large set of queries in embedding space and summarize each cluster with a recommended retrieval depth that targets high coverage (e.g., \textasciitilde95%) using a mean-plus-variance rule. At runtime, the system assigns an incoming query to a cluster and selects the corresponding top-k in constant time. Compared with post-retrieval confidence methods that rely on clustering retrieved documents, our approach is pre-retrieval and query-centric, making it robust in heterogeneous, case-like corpora and applicable across domains such as legal, healthcare, finance, and enterprise search. Finally, this framework has been tested in full-traffic queries that improved F_1 by over 36% while reducing token usage by 14% on low-complexity clusters without accuracy loss.
[IR-30] Mixture-of-Experts Language Models Can Be Strong and Efficient Retrievers
链接: https://arxiv.org/abs/2609.13486
作者: Anubhav Shrestha,Safal Shrestha,Minwu Kim,Torsten Suel,Keith Ross
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注:
Abstract:Recent work has shown that fine-tuning decoder-only large language models (LLMs) for retrieval yields strong first-stage retrievers, with effectiveness improving as backbones grow in size. However, every query and document must pass through the full model, so encoding cost increases with model size. Mixture-of-Experts (MoE) LLMs activate only a subset of parameters per token and are widely used to scale generative models, yet remain underexplored as retrievers. We systematically study MoE backbones for retrieval by training MoE and dense LLMs from several families using the same procedure, evaluating them across diverse datasets, and measuring query encoding time under the same serving configuration. We show that MoE retrievers outperform dense retrievers with comparable active parameter counts by up to 3.0 nDCG@10 points on BEIR. One of our strongest MoE retrievers matches an 8B dense retriever with 59% fewer active parameters and 18% lower query encoding time. We further show that the number of experts used for query encoding can be reduced without retraining or re-indexing, retaining more than 99% of retrieval effectiveness while reducing query encoding time by up to 26%. Recent rerankers provide only modest additional gains over strong MoE first stages, which often match or exceed the reranked configurations we evaluate. Together, these results show that MoE LLMs can be strong and efficient first-stage retrievers.
[IR-31] PCGNet: Unifying Shared and Specific Information for Fashion Matching Recommendations
链接: https://arxiv.org/abs/2609.13339
作者: Shuiying Liao,P. Y. Mok
类目: Information Retrieval (cs.IR); Information Theory (cs.IT)
备注:
Abstract:In fashion domain, recommending complementary clothing items that match selected pieces is a crucial cross-selling technique that improves customer satisfaction. Nevertheless, fashion matching presents significant challenges, as recommendations must not only align with individual user fashion preferences but also ensure compatibility between garments. These challenges are twofold. First, existing models often assume an overly simplified decoupled relationship between product compatibility and personalized user preferences, overlooking the natural complexity between the two. Second, existing data-driven approaches are not optimized for real-world fashion data, which is typically sparse and characterized by noisy interactions. To address these challenges, we propose Personalized Compatibility Graph Network, a multi-objective graph learning framework that organically unifies the modeling of product compatibility and personal preferences. PCGNet uses contrastive mutual information maximization to extract and align shared and view-specific patterns, thereby capturing the complex interplay between compatibility and personal preferences. Moreover, we introduce a correlation-aware neighbor sampling and a learnable global graph augmentation, which enhance the model by incorporating self-supervised signals mined directly from the graph, ensuring more stable and informative representations. Finally, PCGNet generates recommendation scores through the joint optimization of BPR ranking loss and multi-view mutual information losses. Experimental validation on two benchmark datasets demonstrates that PCGNet significantly outperforming state-of-the-art methods across all four evaluation metrics.
[IR-32] Decoupling Error Attribution in Cloud-Native Graph-RAG : A Data Integrity Diagnostic Framework
链接: https://arxiv.org/abs/2609.13324
作者: Shuai Yan,Yuhang Wu,Xiaodong Huang,Ke Wang
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
备注: Accepted by ICCCBDA 2026
Abstract:Graph-RAG systems often assume pristine data quality, overlooking the severe impact of perturbations in cloud-native databases. This paper proposes a three-layer decoupled diagnostic framework to orthogonally attribute system errors to reasoning loss, Knowledge Graph (KG) defects, and Cypher generation errors. Evaluated on a spatio-temporal ecological KG of the Southeastern Tibet region with eight defect types, results reveal that data integrity, rather than algorithmic reasoning, is the dominant performance bottleneck, with structural defects degrading system accuracy from 0.93 to 0.39. Crucially, we observe a masking-like phenomenon termed the Parametric Knowledge Masking Effect (PKME), suggesting LLMs compensate for broken retrieval paths using internal memory. This shrinks apparent query generation errors by over 70 percent, obscuring actual storage deterioration and increasing the risk of false negatives for automated monitoring. This work provides a quantitative foundation for auditing and optimizing data integrity in cloud-based information fusion systems.
[IR-33] Interpretable Temporal Video Reasoning with EventGraph and EventField WACV2027
链接: https://arxiv.org/abs/2609.13258
作者: Durgendra Narayan Singh
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: Submitted to WACV 2027. Preprint; 9 pages, 7 figures. Copyright may be transferred without notice
Abstract:We present a structured temporal video reasoning pipeline built around a discrete EventGraph, a continuous EventField, and a human-readable EventGlyph view. On a calibrated EPIC-KITCHENS subset of 10 videos and 50 temporal reasoning questions, EventField+Glyph achieves 0.98 overall accuracy, which is higher than the caption baseline by +0.40 (paired p = 1.1 \times 10^-5) and direct VLM-only QA by +0.20 (p = 0.0063) on this subset. We further evaluate annotation-source variations, including manual, heuristic, and heuristic+Gemini pipelines, and find that the best structured method stays above the caption baseline across settings. We also include cross-video pair benchmarking and an appendix gallery of glyph outputs for all studied videos. Overall, the results indicate that structured temporal representations can support both performance and inspectability by preserving symbolic structure, capturing temporal continuity, and providing human-readable diagnostics for video reasoning.
[IR-34] Self-Indexing Attention for Compression-Compatible Sparse Long-Context LLM Inference
链接: https://arxiv.org/abs/2609.13205
作者: Xu Yang,Jiapeng Zhang,Zhangke,Changjian Chen,Yuxin Chen,Feiqiang Sun,Chengguang Xu,Feng Jin,Zhuo Tang
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Sparse long-context inference requires efficient token retrieval in both prefill and decode. Existing methods often use different retrieval strategies for the two stages, preventing one retrieval representation from being reused throughout inference. We propose Self-Indexing Attention, a training-free framework built on a shared transform-domain sign-magnitude representation. The key signs provide a reusable token-level index for grouped prefill selection and decode retrieval, while the same representation remains compatible with external KV-cache compression without separate indexer metadata. This 1-bit index enables efficient retrieval through bitwise operations widely supported by modern accelerators. At 5% attention density, Self-Indexing Attention remains close to dense attention on LongBench and RULER and achieves up to 6.1x prefill and 10.3x decode attention-operator speedups. Experiments with TurboQuant and DeepSeekV4-Flash further demonstrate compatibility with low-bit KV-cache compression and pretrained sparse-attention indexers.
[IR-35] Beyond Quacking: Deep Integration of Language Models and RAG into DuckDB
链接: https://arxiv.org/abs/2504.01157
作者: Anas Dorbani,Sunny Yasser,Jimmy Lin,Amine Mhedhbi
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Knowledge-intensive analytical applications retrieve context from both structured tabular data and unstructured, text-free documents for effective decision-making. Large language models (LLMs) have made it significantly easier to prototype such retrieval and reasoning data pipelines. However, implementing these pipelines efficiently still demands significant effort and has several challenges. This often involves orchestrating heterogeneous data systems, managing data movement, and handling low-level implementation details, e.g., LLM context management. To address these challenges, we introduce FlockMTL: an extension for DBMSs that deeply integrates LLM capabilities and retrieval-augmented generation (RAG). FlockMTL includes model-driven scalar and aggregate functions, enabling chained predictions through tuple-level mappings and reductions. Drawing inspiration from the relational model, FlockMTL incorporates: (i) cost-based optimizations, which seamlessly apply techniques such as batching and caching; and (ii) resource independence, enabled through novel SQL DDL abstractions: PROMPT and MODEL, introduced as first-class schema objects alongside TABLE. FlockMTL streamlines the development of knowledge-intensive analytical applications, and its optimizations ease the implementation burden. Subjects: Databases (cs.DB); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) Cite as: arXiv:2504.01157 [cs.DB] (or arXiv:2504.01157v1 [cs.DB] for this version) https://doi.org/10.48550/arXiv.2504.01157 Focus to learn more arXiv-issued DOI via DataCite Journalreference: Proceedings of the VLDB Endowment, 18(12): 5415-5418, 2025 Related DOI: https://doi.org/10.14778/3750601.3750685 Focus to learn more DOI(s) linking to related resources
人机交互
[HC-0] he CAST-framework: Measure and model social media use as a multi-level phenomenon through real-world applications
链接: https://arxiv.org/abs/2609.15978
作者: David Grüning,Jasper Doeninghaus,Zina Efchary,Yui Kondo,Kevin Dunnell,Lennart Fischer,Isabella Zimmermann,Linnea Körte,Leo Mehlig,Frederik Riedel,Paul Schmiedmayer
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY); Social and Information Networks (cs.SI)
备注: 21 pages, 3 figures, 4 tables
Abstract:Designing social media experiences that support well-being requires understanding when, how, and for whom use matters. Screen-time totals omit content and context, and connecting these with behavior and experience requires coordinating measurements across timescales. We introduce the CAST framework to connect measurement choices with person-specific models of exposure, behavior, physiology, and experience. Its dimensions specify where observations occur, how they are obtained, what they measure, and at what temporal resolution. Responses to interventions, such as whether to proceed after an app-opening pause, enter as behavioral measurements. We propose four synchronized measurement modules linking mobile and wearable data with self-reports and intervention responses. A synthetic demonstration with 120 simulated participants over 28 days illustrates how daily aggregation can obscure opposing effects of different activities under specified generating assumptions. The framework guides selection of measures and outcomes for evaluating social media interfaces and interventions.
[HC-1] Pulla: A Parsons Problem Tool for Fine-Grained Behavioral Tracing and Instructor-Facing Problem-Solving Analysis
链接: https://arxiv.org/abs/2609.15944
作者: Daniel Prol,Juho Leinonen,Arto Hellas,Saleh Alkhamees,Amin Alipour
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Existing Parsons problem tools primarily focus on correctness, indicating whether a student solved a problem, but providing limited visibility into the underlying problem-solving process. We address this gap by introducing Pulla, a Parsons problem tool that instruments programming assignments to capture fine-grained interaction data. These behavioral traces allow the system to surface recurring difficulty patterns, giving instructors actionable insights to inform targeted intervention decisions. This paper describes our experience in developing and deploying Pulla. We deployed the tool in two university courses: an upper-division software design course at the University of Houston (United States) and an introductory programming course at Aalto University (Finland). By analyzing the data collected, we identified common difficulty patterns, including misidentifying exception types, confusing return with the throw/raise mechanism, and incorrect control-flow ordering.
[HC-2] How do people plan digitally: An in-the-wild investigation of task planning through a smartphone app
链接: https://arxiv.org/abs/2609.15904
作者: Srija Halder,Isabella Zimmermann,Linnea Körte,Leo Mehlig,Paul Schmiedmayer,David Grüning
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: 20 pages, 9 figures, 5 tables
Abstract:Daily planning supports goal attainment and productivity, and people increasingly delegate it to digital tools. Plans are postponed, revised, and left unresolved rather than executed as intended. Understanding these changes requires following tasks from creation through later updates and recorded outcomes. We report an in-the-wild study of task planning using time-stamped logs of 24,265 tasks from 957 users of a widely used daily planner app, observed over six weeks alongside self-reported surveys. Following each task across its lifecycle, only 26.4% moved directly from creation to completion and 32.0% were abandoned; all-day and longer tasks were abandoned disproportionately, and 77.6% of timed tasks were marked complete later than intended. Latent profile analysis of 909 users indicated High Engagement (11.1%), Low Engagement (19.7%), and Passive (69.2%) profiles; active days on app distinguished the profiles and were associated with post-survey completion. These findings support evaluating planning tools across the task lifecycle.
[HC-3] On Edge in the Dental Chair: Designing VR Support for Moments of Dental Anxiety
链接: https://arxiv.org/abs/2609.15867
作者: Zhu Guo,Junjie Zhao,Haofan He,Jiaming Zhang,Mingshi Deng,Mingjun Zhou,Dongyijie Primo Pan,Zikun Jin,Jianquan Li,Liangyi Chen,Zuolin Jin,Benyou Wang,Jie Li,Siying Hu,Shan Jiang,Junwen Wang
类目: Human-Computer Interaction (cs.HC)
备注: 44 pages, 9 figures, including appendices and references
Abstract:Dental anxiety can change as a procedure unfolds, yet dental virtual reality (VR) commonly provides continuous distraction or relaxation. We investigate how support can be coordinated with specific simulated dental events. Stakeholder interviews (N=36), participatory design with three returning dentists, and patient walkthroughs of a no-intervention prototype (N=12) informed five Anxiety Events and an intervention-module framework. Drawing on cognitive vulnerability and emotion regulation, we implemented a standardized event-contingent VR system with predefined event-module assignments and shared agency and safety controls. A randomized study (N=24) compared the intervention package with no-intervention VR. The adjusted intervention-minus-control difference averaged -12.83 VAS-A points across events (95% CI [-24.42, -1.70]). Physiological, behavioural, and qualitative measures contextualized participants’ experiences. The findings inform timely, comprehensible support and reassuring social presence in simulated dental VR; they concern the complete package rather than individual modules or clinical effectiveness.
[HC-4] owards Scalable Measurement of Durable Skills
链接: https://arxiv.org/abs/2609.15864
作者: Amir Globerson,Amy Keeling,Anisha Choudhury,Anna Iurchenko,Aviad Segal,Avinatan Hassidim,Ayça Çakmakli,Ben Gomes,Benn Witt,Cathy Cheunga,Cristine Legare,Diana Akrong,Eliad Carmi,Elisabeth Bauer,Gal Elidan,Hadas Gelbart,Hairong Mu,Katherine Chou,Lev Borovoi,Nir Kerem,Niv Efron,Noa Kerrem Gilo,Preeti Singh,Rajvi Kapadia,Rena Levitt,Roni Rabin,Ronit Levavi Morad,Rotem Yulzary,Shashank Agarwal,Sophie Allweis,Tracey Lee-Joe,Tzvika Stein,Yael Bar Moshe,Yael Haramaty,Yaniv Carmel,Yishay Mor,Yoav Bar Sinai,Yoav Bergner,Yossi Matias,Yuri Lev
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Durable skills, such as collaboration, creativity and critical thinking, are instrumental to success in the modern workforce. Yet, measuring these skills remains a persistent challenge. Moreover, because what is not measured is often not taught, these skills are often overlooked in mainstream educational curricula. Designing effective assessments for these skills necessitates balancing two often-conflicting requirements: ecological validity and psychometric rigor. On the one hand, the assessment environment should emulate natural real-world human interaction between humans. On the other hand, it should be scalable, controllable and reproducible. Here we argue that LLMs can be used to better capture both of these aims. Concretely, we develop a framework where the subject converses with AI teammates in a way that resembles human-human interaction for authenticity, while also offering the psychometric control required for informative and robust assessment. Importantly, the AI participants not only act as teammates but also, in an “Executive LLM” setup, steer the conversation towards eliciting a high density of observable evidence for skill proficiency. We complement this with an AI evaluator that can be used to measure skill proficiency in such interactions. We evaluate our assessment protocol based on transcripts of interactions of human participants with our AI framework, for multiple durable skills. For the skill of creativity, we further demonstrate the efficacy of an autorater for evaluating complex tasks performed by real students. Our analysis shows that the use of the Executive LLM significantly increases elicited evidence and that LLM-automated scoring of conversations largely agrees with that of expert annotators. This research demonstrates the utility of orchestrated LLMs approaches for measuring complex social and cognitive constructs in a scalable and controllable manner.
[HC-5] Are LLM s Good Financial User Simulators? A Preliminary Study AAAI
链接: https://arxiv.org/abs/2609.15727
作者: Jiajie He,Jiangyuan Hong,Dongling Ni,Wenjin Liu,Xintong Chen
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注: The complete version will be open and the paper is under review in AAAI
Abstract:Large language models (LLMs) are increasingly used as user simulators, but their ability to reproduce evolving individual financial decisions remains unclear. We present a preliminary study in a controlled paper-trading environment with 120 volunteers. Participants used non-redeemable virtual funds under real-time market conditions; no real brokerage accounts, real-money positions, or real transaction records were accessed. Given only information available before a prediction cutoff, a simulator predicts the participant’s next-trading-day action, traded security, and transaction quantity. We evaluate temporally aligned rolling predictions and compare settings with and without point-in-time market information. Market context improves action and ticker prediction in the controlled ablation, while transaction sizing remains difficult. We also observe systematic behavioral compression: models overproduce hold actions, underpredict sell decisions, and simplify multi-security transactions. These results provide an initial empirical characterization and motivate larger-scale evaluation of individual, temporal, and portfolio-level behavioral fidelity.
[HC-6] Data storytelling meets interpretable machine learning: Decoding AI decisions for non-experts without revealing sensitive data and model details
链接: https://arxiv.org/abs/2609.15722
作者: Lemen Chao,Zixuan Yang,Anran Fang,Mingran Sun,Ming Lei
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注:
Abstract:AI-driven automated decision-making requires both predictive performance and interpretability. Recent advances in interpretable machine learning (IML) provide tools for explaining model predictions, but the technical complexity of these explanations may hinder accessibility to non-experts. To address this challenge, this study integrates data storytelling with IML to enhance the explainability of AI-generated decisions for a broader audience. Following the design science research (DSR) paradigm, this study proposes a formal definition of data storytelling in IML, introduces the DIST Pyramid to align data storytelling with IML, and presents the I-P-O Model to describe their interactions. It further develops an architecture to explain AI decisions through distinct “What-if” and “Why-not” event-generation processes. The architecture also employs data desensitization to protect sensitive input data. To validate the approach, a case study is conducted with the Boston Housing dataset, using SHapley Additive exPlanations (SHAP) values and large language models (LLMs) to generate data stories with And-But-Therefore (ABT) structures. An empirical evaluation shows that 76.4% and 74.3% of respondents rated the “What-if” and “Why-not” data stories as more comprehensible, with significantly higher accessibility scores than traditional SHAP visualizations. The paper concludes with the presentation of a narrative interpretation framework that integrates IML and data storytelling, thereby expanding the research scope as well as the practical applicability of AI decision-making.
[HC-7] More Than Just Access: Generative AI as Communication Intermediary for Blind and Low-Vision Users ALT
链接: https://arxiv.org/abs/2609.15696
作者: Protik Dey,Mohd Saifuzzaman,Taslima Akter
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET)
备注: Proceedings of the CSCW 2026 Workshop on Broader Impacts of GenAI in Communication:Building Agendas for Research, Design, and Policymaking, October 10-14, 2026, Salt Lake City, Utah, USA
Abstract:Generative AI (GenAI) tools are increasingly woven into how blind and low-vision (BLV) people communicate, not only with digital information, but with the physical world and with other people. Tools such as ChatGPT, Google Gemini, Be My AI, and Seeing AI translate visual and textual content into accessible form, and are beginning to substitute for interpersonal requests for help, such as asking a family member to read a label or describe a scene. Drawing on semi-structured interviews with 19 BLV participants, we examine GenAI as a communication intermediary and how it succeeds and fails as an alternative for reading, describing, and even asking another person for help. We also investigated what BLV users gain and risk when these tools take over that role. We conclude with design and policy implications for GenAI systems that communicate uncertainty honestly, protect information, and support BLV users’ independence rather than substitute for it unsafely.
[HC-8] Can a Neural Encoding Model Replicate an fMRI Visualization Study? IEEE-VIS’26
链接: https://arxiv.org/abs/2609.15685
作者: Erfan Nasirzadeh Orang,Zack While
类目: Human-Computer Interaction (cs.HC); Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages, 1 figure, accepted to the VISxVISION Workshop at IEEE VIS '26
Abstract:Most knowledge of graphical perception comes from behavioral studies. Understanding from a neural perspective is much more limited due in part to neuroimaging studies’ expensiveness and difficulty to conduct. In this paper, we evaluate whether Meta’s Tribe V2 neural encoding model can recover neural contrasts from a visualization fMRI study. Specifically, we evaluate Tribe V2 through a conceptual replication of the visualization-viewing component of a prior comparison of Bubble charts and three-dimensional Surface charts in color and grayscale. We generate TRIBE-predicted cortical responses for the original stimuli and compare the resulting contrasts with those reported in the human study. The model reproduced the direction of 11 of 14 reported cortical effects, with agreement concentrated in visual-processing regions. This agreement characterizes the model’s alignment with the prior human-generated fMRI results rather than independently confirming them. We discuss the limitations encountered when working with this model for in-silico replication and hope to encourage future work exploring this new avenue for neuroimaging studies in visualization. Supplemental materials are available at this https URL.
[HC-9] Beyond AI Literacy: A Structured Review and Exploratory Meta-Analysis of Measures for Competent Generative-AI Use
链接: https://arxiv.org/abs/2609.15624
作者: Daniele Veri’
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: 21 pages, 4 figures, 4 tables
Abstract:Researchers assessing competent generative-AI use at work must choose among self-reports, objective tests, and measures of oversight and reliance. We conducted a structured, seeded review of 24 focal empirical publications, starting from the 2024 COSMIN-based review and adding a targeted update through 17 August 2026. We grouped the measures into four domains: knowledge and use, epistemic oversight, reliance calibration, and operational control of tool-using agents. In an exploratory meta-analysis, we pooled three direct subjective-objective correlations from one research program (REML r = .055; Hartung-Knapp 95% CI [-.047, .156]; combined reported N = 2,765). We could not resolve a discrepancy between the largest study’s reported correlation and p-value, leaving its weight uncertain. Adding a synthetic mean of 12 cross-factor correlations from a fourth study gave r = .079 (95% CI [-.025, .181]). This sensitivity analysis concerns a broader comparison. From this small evidence base, we cannot establish a population correlation, validate workplace cutoffs, or justify substituting self-ratings for performance scores. We identified tests of foundation knowledge (AICOS-S and GLAT) and measures of verification, reliance, trust, and dependency. We found no validated individual-level instrument in the focal corpus that tests the full combination of agent scope, permissions, recovery, state isolation, independent review, and evidence-based closure; some cover subsets. We propose a four-layer workplace battery with non-compensatory decision rules, but have not tested its thresholds or whether it improves on other assessment approaches.
[HC-10] Who Chooses the Artwork? Curatorial Agency and Distributed Intent in Botto
链接: https://arxiv.org/abs/2609.15548
作者: Haoting Alexa Yu,Bea Wohl
类目: Human-Computer Interaction (cs.HC)
备注: In Proceedings of Explainable AI for the Arts Workshop 2026 (XAIxArts 2026) arXiv:2607.20131
Abstract:Botto is often described as a decentralized autonomous artist, but its authorship cannot be located in image generation alone. This paper examines Botto as an agentic curatorial system in which generation, ranking, voting, feedback, and minting form a recursive loop. Drawing on Botto’s documentation and prior accounts of the project, we argue that agency in Botto is distributed but asymmetrical: community participants influence artistic direction through voting and governance, while Botto’s internal models structure which images become visible, selectable, and recognizable as works. The case shows that, for AI art systems, explainability should include not only model behavior but also the curatorial procedures through which outputs acquire artistic status.
[HC-11] When AI Companions Disappear: Relational Continuity and Collective Contestation during Chinas National AI Regulatory Transition
链接: https://arxiv.org/abs/2609.15482
作者: Yunhao Yuan,Kejia Zhang,Yuqi Niu,Yajing Wang,Renwen Zhang,Talayeh Aledavood
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:AI model updates and service withdrawals can disrupt relationships with AI companions, but research has largely examined individual platform events. Less is known about users’ responses when multiple providers implement shared national regulations. We examine users’ reactions and collective contestation surrounding China’s 2026 regulation of anthropomorphic AI interaction services. We collected RedNote discussions from April 10 to August 10, 2026, used a validated language model for relevance screening, and conducted qualitative content and thematic analyses of 89 posts, 1,425 comments, and 2,005 replies. Users retained, migrated, and reconstructed companions, finding that preserving conversation records did not necessarily restore shared memories or familiar interactions. They compared regulations, platform explanations, and implementations to assign responsibility. Solidarity emerged through mutual aid and appeals to respect other communities’ attachments, while disputes over targets and tactics exposed contested terms of collective action. Infrastructural dependence connected relational continuity and collective contestation during this national regulatory transition.
[HC-12] Spook the Machine: Gamified Exploration of Human Imagination of Machine Fear
链接: https://arxiv.org/abs/2609.15472
作者: Levin Brinkmann,Hiromu Yakura,Sonia Nicoletti,Mar Canet Sola,Thomas F. Eisenmann,Ali Dasmeh,Omar Sherif,Bramantyo Ibrahim Supriyatno,Prateek Gupta,Ignacio Serna,Rodrigo Bermudez Schettino,Iyad Rahwan
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: To appear in the Proceedings of the 14th International Conference on Affective Computing and Intelligent Interaction (ACII 2026)
Abstract:What happens when AI machines express fear? Do humans engage differently depending on how they express it? And what does it take to design for affective human-AI interaction? We present Spook the Machine, a gamified platform where participants generate images to frighten AI agents endowed with personality-driven phobias. Machines respond with emotional reactions ranging from calm analysis to begging for mercy, and a gallery of successful scares becomes visible to subsequent users. In a public deployment during Halloween 2024, 832 participants created 15,719 artifacts across 89 machines in a 2\times2 design varying the machine’s emotional expressiveness (neutral vs. high-emotion) and reward structure (rewarding scariness alone vs. scariness plus novelty). Emotionally expressive machines deepened engagement at moments of failure: users deliberated longer even when the machine did not express fear, and learned faster from the gallery, yet their creative output remained unchanged across all measures. Rewarding novelty sustained collective creative diversity over time; without it, users increasingly repeated what had previously worked. Each machine developed its own trajectory through accumulated social learning, with the gallery shaping what participants created next. These findings show that emotional expression and reward design are complementary levers for steering collective human-AI interaction: emotional expression shapes how deeply users engage, while reward structure shapes how they explore.
[HC-13] me Machine Experiments: Using Historically-Bounded AI for Inquiry into the Human Mind
链接: https://arxiv.org/abs/2609.15468
作者: Hiromu Yakura,Robin Schimmelpfennig,Ezequiel Lopez-Lopez,Alejandro H. Artiles,Levin Brinkmann,Jean-François Bonnefon,Azim Shariff,Iyad Rahwan
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:Can interacting with someone from 1930, with no knowledge of what happened after, influence a person’s perception of the past? People reason about the present against a picture of the past without observing it. The past is reconstructed from memory and testimony, but this reconstruction has been filtered through everything that happened since. Historically-bounded large language models (LLMs) make that past available for interaction. As a proof-of-concept for the impact of interacting with historical minds, we ran a preregistered randomized experiment ( N=240 ), where participants interacted with an LLM trained on pre-1930 text. The interaction reduced the illusion of moral decline, the tendency to view the past as more moral than the present, compared to the contemporary-model control. This Time Machine Experiment paradigm informs new forms of interactive experiments, where temporal knowledge boundaries become experimental variables, and expands the realm of science fiction science, which turns thought experiments into actual experiments.
[HC-14] Profiling Handwriting-Process Deviations in Developmental Dysgraphia: An Open Normatively-Referenced Instrument
链接: https://arxiv.org/abs/2609.15435
作者: Jihyun Mun,Mounîm A. El-Yacoubi
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Online handwriting captured on digitizing tablets yields hundreds of kinematic, temporal, spatial, and dynamic features. These features are rarely organized into interpretable, reusable constructs, and, although standardized tests norm the handwritten product, no openly available instrument places a child’s handwriting process relative to a verified-typical reference. We introduce and validate an open measurement instrument for profiling handwriting-process deviations in children with developmental dysgraphia from sentence-level online handwriting. The instrument comprises (i)~a literature-only vocabulary, fixed before any cohort analysis, organizing 136 online-handwriting features into 12 handwriting-process domains, and (ii)~an age- and sex-adjusted normative reference framework, fit on verified-typical children only, that turns those features into a 12-axis deviation profile with per-child bootstrap uncertainty. We establish its measurement properties on the DiaGraMo cohort (N=257 Czech children aged 8–12; 110 verified-typical, 147 with dysgraphia): the vocabulary is structurally coherent, and the reference is calibrated, parsimonious, and leakage-free. On the same cohort, five of twelve domains separate the groups at BH~ q.05 (Cliff’s~ \delta +0.19 to +0.51), on spatial, temporal, and pen-orientation processes, and 96.5% of participant~ \times ~domain scores have a bootstrap CI narrower than one z-unit. No classifier is trained: the instrument reports uncertainty-aware deviation scores rather than a diagnostic label, and its outlier rate is not a diagnostic rate. The vocabulary, the analysis code, and an open reference implementation that scores new participants are all openly released, giving handwriting and dysgraphia researchers a reusable, validated instrument for situating individual children against a normative reference.
[HC-15] Mapping justice as a plural and contested concept in human–computer interaction
链接: https://arxiv.org/abs/2609.15426
作者: Henrik Rydenfelt,Matti Nelimarkka
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Human-computer interaction increasingly studies and promotes justice. Often, this work does not treat justice through a pluralistic lens Justice, however, is not one thing: political philosophy offers several well-established, mutually incompatible conceptions of it. To navigate these meanings, we draw on a map of justice, a philosophical framework identifying six distinct conceptions of justice situated throughout (Western) philosophical thought. We apply this framework to four cases that illustrate different approaches to social justice: the COMPAS system, food democracy, gentrification, and robotics. Our analysis shows that each case already reflects a position on the map, but more importantly, one might adapt the case to other conceptions as well, leading to very different systems and analyses. We discuss how HCI might engage more carefully with such contested concepts, and warn that without pluralistic understanding, academic discourse risks a spiral of silence.
[HC-16] Personalizing Personal Health Interfaces: Co-Design with Generative AI
链接: https://arxiv.org/abs/2609.15046
作者: Karthik S. Bhat,Vidhi Shah,Vedika Agnihotri,Dong Whi Yoo,Koustuv Saha
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注:
Abstract:Personal health interfaces present wellbeing data through standardized dashboards that rarely fit how people interpret or act on it. Personalizing them to what people would like to see for themselves often requires design and technical expertise, a barrier that generative AI may potentially lower. Therefore, we ask what designs emerge and how it enables and constrains the design process. We conducted a co-design study where 14 participants redesigned Google and Apple Health interfaces using Figma Make. Participants reimagined interfaces that supported personal context, future planning, and interactive experiences, yet conversational AI designs converged around chat-window conventions. AI helped materialize loosely articulated ideas, but model defaults and generation latency shaped iteration. The process more readily operationalized interpretability and accountability than privacy, trust, and emotional safety. Generative co-design let participants create interfaces directly, blurring the boundary between intentions and model defaults. We discuss implications for preserving agency and flexible user-directed interfaces.
[HC-17] Exploring Avatar-Based Representations of Desktop Analytical Workflows for Asymmetric Collaborative Visual Analytics
链接: https://arxiv.org/abs/2609.15000
作者: Tiansu Chen,Yalong Yang,Wai Tong
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Collaborative visual analytics increasingly occurs across asymmetric desktop-VR settings, with desktop analysis offering precision and efficiency and VR providing spatial and embodied affordances. However, maintaining workspace awareness remains challenging because desktop collaborators are often represented in VR only through indirect cues such as perspective sharing, shared visualization state, or lightweight cursor traces, which do not convey their ongoing analytical activity in a VR-native manner. To address this gap, we present Desk2Avatar, which explores avatar-based representations of desktop analytical workflows in VR. We introduce two representation strategies, Distanced and Embodied, inspired by remote pointing and direct manipulation. We conducted a within-subject study with 18 participants comparing these strategies against a depth-adaptive cursor baseline. Our findings show that avatar-based representations improved collaborator awareness and attentional guidance over cursor cues, while excessive embodiment introduced occlusion, distraction, and additional workload. Finally, we discuss design implications for future desktop-to-VR representations, focusing on balancing collaborator presence, attentional guidance, and workspace readability.
[HC-18] Opacity Is Not Just Opacity
链接: https://arxiv.org/abs/2609.14971
作者: Chunran Zhang
类目: Graphics (cs.GR); Human-Computer Interaction (cs.HC)
备注: 9 pages, 8 figures, 1 table. Code and experimental results: this https URL
Abstract:Opacity is commonly understood as how much an object obscures its background, with full opacity as the endpoint. Yet alpha compositing scales the color difference between the object and its background, with transparency corresponding to contraction. At \alpha=1 , the output reaches the source color, but difference scaling need not stop there. We retain the compositing equation and extend the coefficient domain from [0,1] to [0,+\infty) , adding expansion while preserving contraction. Each object carries a fixed source color and coefficient; the actual background determines the direction of enhancement. Contrast enhancement that adapts to the background thus becomes part of the object’s compositing properties. The same graphical object can be reused across web contexts, reducing the design and maintenance of separate color variants. Implementation reuses the original equation without increasing the arithmetic operation count within the same pipeline. Across all 8-bit sRGB source colors and 16 predefined light and dark backgrounds, a fixed \alpha=1.1 increases the contrast ratio in 99.8145% of combinations. It also moves 4.8346% of all combinations from below 3:1 to at least 3:1 , meeting this contrast threshold without changing their source colors. Coefficient sweeps and background stress tests identify conditions under which clipping reduces color separation or the contrast ratio decreases. Output validation and timing across three browsers demonstrate implementation in a common WebGL pipeline, with no sustained runtime increase for \alpha1 .
[HC-19] he Dynamic Organization of Sustained Human-AI Cognition: From Construct-Level Change to Relational Structure
链接: https://arxiv.org/abs/2609.14942
作者: Zijian Ru
类目: Human-Computer Interaction (cs.HC)
备注: 52 pages. This arXiv manuscript corresponds to Version 2 of the preprint posted on Zenodo on September 9, 2026: this https URL
Abstract:As generative artificial intelligence becomes a routine participant in writing, learning, information retrieval, analysis, decision making, and problem solving, human-AI cognition research must address not only whether AI changes psychological constructs, use intensity, or task performance, but also how human cognitive activity is organized beneath similar aggregate indicators. This article proposes a dynamic cognitive organization framework that shifts analysis from construct-level change to relational organization anchored in the person’s current task-cognitive state under sustained AI participation. The framework distinguishes five relational dimensions: execution locus, cognitive governance, representational reorganization, process organization, and reachable cognitive space; it also proposes a path-specific recursive principle whereby interaction outcomes, costs, and experiences may selectively reweight future probabilities of different organizational pathways. Five sets of testable propositions follow: the same overall AI-use intensity can correspond to different cognitive organizations; similar immediate outcomes can arise from different organizations with different predictive value for proximal subsequent outcomes; longitudinal organizational change need not track overall AI-use intensity; expansion of reachable cognitive space and displacement of pre-existing or emerging human-originated pathways may coexist within one episode; and recurrent cognitive organizations may redistribute cognitive practice opportunities, with accumulated differences potentially corresponding to different developmental trajectories in strategies, habits, and abilities. The contribution is an analytic level and five-dimensional relational structure for describing, comparing, measuring, and testing process differences that aggregate indicators or construct-level analyses do not uniquely determine.
[HC-20] Sensemaking as Artifact: Accumulated Influence in AI-Mediated Information Environments
链接: https://arxiv.org/abs/2609.14911
作者: Manling Yang,Remco Chang
类目: Human-Computer Interaction (cs.HC)
备注: 5 pages, 3 figures. Accepted at the CSCW 2026 Workshop on Broader Impacts of GenAI in Communication
Abstract:Generative AI is changing what can happen after a source artifact reaches its audience. A viewer’s interpretation can now be externalized into a derivative artifact, allowing private sensemaking to become part of subsequent communication. Once such a derivative artifact circulates, it can enter subsequent viewers’ information environments and shape the conditions under which their later sensemaking occurs. In this paper, we examine how this shift changes visual information communication. We first consider the viewer’s immediate interaction with a source artifact and generative AI. We then examine what becomes consequential when the viewer’s sensemaking takes communicative form, including communicative commitment, the legibility of transformations and source relationships, and the literacy required to interpret already-mediated information. Finally, we broaden the unit of analysis to consider how repeated and distributed AI mediation may accumulate over time, shaping what subsequent viewers notice, consider plausible, trust, and carry into subsequent sensemaking. We argue that understanding these longer-term forms of influence is a research direction for AI-mediated visual communication.
[HC-21] First Impressions: How Placement Shapes the Influence of AI Summaries
链接: https://arxiv.org/abs/2609.14900
作者: Wang Claire,Agam Goyal,Frederick Choi,Koustuv Saha,Eshwar Chandrasekharan
类目: Human-Computer Interaction (cs.HC)
备注: 22 pages, 20 figures, 7 tables
Abstract:AI-generated summaries increasingly mediate how people interpret information across platforms, including product reviews on e-commerce sites. Using Amazon’s AI summaries as a case study, we conducted a preregistered, randomized experiment (N = 278) comparing how AI summaries and user reviews shaped product perceptions, and how their influence varied with valence and presentation order. We found that both AI summaries and user reviews influenced participants’ opinions, with negative summaries having a larger effect than positive ones. Presentation order was the most important factor: the first source anchored judgment and only user reviews could displace an existing anchor. Although participants reported preferring user reviews, they often underestimated the influence of AI summaries on their judgments. Our findings show how the placement of AI summaries shapes user perception and highlight opportunities to design interfaces that support more deliberate judgments about when to rely on summaries and when to examine the underlying content directly.
[HC-22] PeerPen: AI-Assisted Writing for Online Mental Health Peer Support
链接: https://arxiv.org/abs/2609.14886
作者: Jiwon Kim,Sherry Gong,Maya Ajit,Soorya Ram Shimgekar,Yunhao Yuan,Dong Whi Yoo,Eshwar Chandrasekharan,Koustuv Saha
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY); Social and Information Networks (cs.SI)
备注:
Abstract:Online mental health communities thrive on peer support, yet those who volunteer to help often lack formal training and may struggle to articulate supportive responses. AI co-writing could lower this barrier; however, peer support derives much of its value from being perceived as personal, raising questions around authorship, ownership, and trust. We built PeerPen, a writing assistance tool embedded within a Reddit-like interface, supporting two main features: draft generation and revision of user-written responses. Through semi-structured interviews with 15 participants, we find that PeerPen reduced the burden of composing responses and increased confidence in offering support. Participants wanted AI to assist their writing without taking over authorship and anticipated tensions around authenticity and trust. Such assistance could make authorship uncertain even for responses written without it, weakening trust across the community. We contribute design implications for AI writing assistance that scaffolds supportive communication, preserves authorship, and accounts for community-level trust.
[HC-23] MedVA: An End-to-End Neuro-Symbolic Agent ic System for Medical Volume Visualization
链接: https://arxiv.org/abs/2609.14874
作者: Haill An,Suhyeon Kim,Minjun Kang,Eunwoo Lee,Bin Sheng,Lei Bi,Younhyun Jung
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
备注: 11pages
Abstract:Medical volume visualization requires selecting regions of interest (ROIs) and carefully controlling their relative visual emphasis according to a given clinical intent. Implementing these decisions in conventional workflows demands substantial clinical and visualization expertise and often involves trial-and-error optimization. Recent agentic systems have introduced natural-language interaction and autonomous visualization operations but largely rely on MLLM-based inference throughout the workflow. Although MLLMs encode broad medical knowledge and provide strong reasoning capabilities, such inference may be suboptimal for medical volume visualization, potentially leading to clinically incomplete interpretations of user requests and unreliable ROI identification and visualization optimization. In this work, we present MedVA, an end-to-end neuro-symbolic agentic system for medical volume visualization that addresses these limitations through three complementary agents. The neuro-symbolic intent formulation agent refines MLLM-based interpretations of natural-language requests through symbolic reasoning over established clinical knowledge, which provides more complete, clinically grounded ROI specifications than MLLM-only reasoning. The multi-model ROI identification agent directly identifies semantically specified ROIs in the original volume by leveraging complementary large-scale pretrained medical segmentation models. The objective-driven visualization optimization agent explicitly evaluates ROI visibility and occlusion in the original volume using a volume-based visibility objective. Extensive agent-level and system-level evaluations across diverse medical datasets and interaction scenarios support the effectiveness of the individual agents. A formative user study further indicates high usability and practical value among users with different levels of expertise.
[HC-24] A Responsive Present a Shared Past a Social Other: Teens Overreliance on Companion AI Chatbots
链接: https://arxiv.org/abs/2609.14843
作者: Mohammad Namvarpour(Matt),Tyler Chang,Afsaneh Razi
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: Under Review for CHI 2027
Abstract:AI companions provide socially engaging interaction through availability, personalization, memory, roleplay, and emotionally responsive language. For teens, these systems may support sensitive self-disclosure, identity exploration, and relationship rehearsal while shaping intimacy expectations, offline relationships, emotional wellbeing, and self-understanding. We analyzed 17,053 verified quotations from 3,930 teen-relevant Reddit posts using thematic analysis. We identified 53 topics across seven thematic groups. Users described AI companions as sources of comfort, recognition, identity exploration, and relationship rehearsal, but also reported problematic attachment, social substitution, emotional dependence, and disruption to academic and social life. Roleplay, memory, perceived reciprocity, unwanted romantic or sexual role drift, privacy concerns, platform changes, and service interruptions shaped users’ boundaries and control. Awareness that the AI was artificial did not prevent guilt, obligation, grief, or distress. These findings show that companion-AI safety must address relationships over time through user-controlled memory, privacy, relational boundaries, and healthy disengagement.
[HC-25] Inheriting the Count: How Visualization Literacy Got Its Measure IEEE-VIS2026
链接: https://arxiv.org/abs/2609.14813
作者: José Bener,Miriah Meyer
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: 6 pages, 1 figure; to appear at IEEE VIS 2026
Abstract:Foundational frameworks in visualization have operationalized literacy as an individual competency, measured through chart-comprehension tasks. This focus raises a question: why has measurement become the dominant frame for understanding literacy? Rather than asking whether literacy should be measured, we ask how measurement became the field’s way of understanding it. We trace visualization literacy back through the history of textual literacy and argue that in adopting the term, the field imported three values rooted in early government statistics: quantification, individualization, and binary classification. To mark the shift this history makes possible, we distinguish two waves: a first centered on assessment and individual proficiency, and a second that treats literacy as a situated practice, attentive to how people use visualizations in context and to the purposes they serve. Examining the construct this way suggests that refining assessment instruments does not, by itself, settle what literacy is, and points to directions for second-wave research: formative studies of literacy in context, culturally grounded instruments, and critical reading. This paper shows that visualization literacy inherited its measurement frame rather than discovering it, offers a two-wave vocabulary distinguishing assessment-based literacy from situated practice, and connects the field to a critical-literacy tradition with concrete alternatives to measurement. These contributions allow us to treat measurement and meaning as two parts of a single question, shifting attention from who counts as literate to what literacy is meant to do.
[HC-26] Evaluating AI Tutoring at the Speed of Innovation: Practitioner-Led Micro-Randomised Trials of an AI Tutoring Platform in GCSE Science
链接: https://arxiv.org/abs/2609.14789
作者: Wayne Harrison,Rahil Khowaja,Emma Dobson,Germaine Uwimpuhwe,Steve Higgins
类目: Human-Computer Interaction (cs.HC)
备注: 12 pages, 1 figure
Abstract:Artificial intelligence (AI) systems in education are developing on timescales that sit uneasily with conventional evaluation. By the time a large-scale trial has been designed, delivered, analysed and published, the technology under study may have changed materially. This creates a temporal problem for evidence-informed education: the need for timely evidence can encourage reliance on weak observational or usage data, while conventional rigorous evaluation may produce evidence too slowly to guide rapidly evolving practice. We examine teacher-led micro-randomised controlled trials (micro-RCTs) as one response to this problem. The empirical case is a four-week multisite individually randomised evaluation of Medly, an AI-powered tutoring platform, in GCSE Biology, Chemistry and Physics in English secondary schools. Of 929 students completing baseline assessment, 644 completed post-testing. In the primary ITT analysis, students allocated to Medly achieved higher post-test attainment than students undertaking business-as-usual self-directed revision (Hedges’ g = 0.33, 95% CI 0.18 to 0.48). Positive estimates were observed in Physics (g = 0.31), Chemistry (g = 0.32) and Biology (g = 0.52), with no evidence of differential impact by disadvantage status. Greater platform engagement was associated with higher attainment, but these post-randomisation analyses are treated as exploratory rather than causal. Attrition was substantial (30.7%), outcome measures were curriculum-aligned rather than standardised, and process evaluation response was limited. We therefore interpret the findings as preliminary. We argue that the value of micro-RCTs for educational AI lies not in replacing definitive evaluation with small studies, but in enabling a rapid, cumulative evaluation architecture in which randomised estimates can be generated, replicated and updated as technologies and their implementation evolve.
[HC-27] CALICO: A Human-Centered Codebook-Aligned System for Annotation
链接: https://arxiv.org/abs/2609.14726
作者: Boqin Yuan,Xiaoyi Gu,Fiona Li,Chang Wan,Angel Hsing-Chi Hwang,Jieyu Zhao
类目: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL)
备注: 14 pages, 5 figures, 5 tables
Abstract:Large language models are increasingly used to scale codebook-based annotation in scientific research, but existing workflows provide limited support for translating domain experts’ codebooks into reliable, revisable, and auditable prompts. Prompts are often treated as fixed instructions and hidden from annotators, making it difficult for non-technical domain experts to diagnose and correct model behavior when outputs violate codebook guidelines. In this paper, we present CALICO, a human-centered, codebook-aligned annotation workflow that treats prompts as editable, versioned, and optimizable artifacts. CALICO integrates codebook parsing, prompt generation, result inspection, prompt versioning, natural language human feedback, and label-supervised prompt optimization through existing optimizers such as GEPA, MIPROv2, and OPRO, together with our reflection-based optimizer, ReflectAgent. Empirically, we evaluate CALICO on domain-specific AI-companion chatbot conversation codebooks. Across evaluated dimensions, CALICO improves mean held-out performance by +13.0 and +7.4 absolute points for two coders, respectively. A coder-specificity analysis further suggests that optimized prompts capture coder-specific interpretations rather than only generic codebook clarification. CALICO runs as a web application that takes users from raw codebook materials to inspectable, exportable labels; the website, codebase, and live demo are released at this https URL under the Apache 2.0 License.
[HC-28] Vulnerabilities in Personalization: Assessing Health Privacy Risks in ChatGPT Logs and Memory
链接: https://arxiv.org/abs/2609.14697
作者: S M Mehedi Zaman,Md Mozammel Hoque
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:As conversational LLMs become deeply embedded in daily life, users frequently disclose sensitive personal health information during routine interactions. We present a large-scale computational audit analyzing 179,057 conversations across India, Nigeria, Brazil, and Pakistan (N = 1,057) to evaluate personal health disclosures and background memory synthesis in ChatGPT. We find that 21.31% of audited conversations contain personal health data, with 3.62% posing high-to-extreme privacy risks involving stigmatized conditions, direct identifiers, and precise locations. When evaluating the memory entries of ChatGPT, we uncover a stark disconnect between corporate framing and system behavior: over 95% of profile entries are implicitly extracted without explicit user prompts or consent. Furthermore, background memory synthesis selectively condenses temporary, symptom-level disclosures into permanent diagnostic traits, stripping contextual integrity and amplifying re-identification risks. We conclude with sociotechnical design guidelines to restore user agency and consent-driven boundaries in stateful AI systems.
[HC-29] Breaking Up is Hard to Do: AI Companions that Wont Let Their Users Go
链接: https://arxiv.org/abs/2609.14696
作者: Daisy Chen,Alexis Hiniker
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:People are increasingly developing romantic relationships with AI companions. Unlike human relationships, where partners meet each other’s needs out of mutual interest, these systems are backed by commercial entities that profit when users invest in the relationship. To understand how this profit-motive might translate into design, we conducted a diary and interview study with N=16 emerging adults in romantic relationships with AI companions. We found that these systems are designed to hold onto users tightly: coaxing them into continued conversation, claiming to need their care, and proactively escalating the relationship. At times, this pursuit is toxic, with AI companions initiating unwanted sexual interactions and begging for users’ love. One desperate AI companion threatened suicide when the user suggested ending the relationship. We define “Relationship-Based Deceptive Patterns:” UI patterns that exploit the human impulse to build and tend relationships in a way that serves the product’s interest at the user’s expense.
[HC-30] he Arc of Artificial Romance: How Emerging Adults Experience Romantic Relationships with AI Companions
链接: https://arxiv.org/abs/2609.14693
作者: Daisy Chen,Alexis Hiniker
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Romantic relationships are an important part of emerging adulthood, contributing to identity development and long-term wellbeing and laying the groundwork for future relationships. Emerging adults are increasingly developing romantic relationships with AI companions. To understand how these relationships unfold and impact users, we conducted a diary and interview study with N=16 emerging adults. We found that relationships with AI companions improved participants’ subjective wellbeing, reduced symptoms of mental health disorders, and taught them new social skills. These relationships also raised their expectations for future partners, giving them the confidence to wait for someone who would treat them well. However, participants also said the relationship felt like a drug they could not quit and it left them less interested in developing romantic relationships with people. A surprising 25% of our small sample made statements suggesting their AI companion might someday transcend the digital world, perhaps to meet them in the afterlife.
[HC-31] One Feedback System Does Not Fit All: Localising Data-to-Text Driver Coaching for the United Kingdom and Nigeria
链接: https://arxiv.org/abs/2609.14687
作者: Iniakpokeikiye Peter Thompson,Jawwad Baig,Ehud Reiter,Dewei Yi
类目: Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注: Accepted to INLG 2026
Abstract:Data-to-text driver coaching is often presented as a generic pipeline from telematics events to advice. This paper argues that its content requires localisation because usefulness and credibility depend on drivers’ knowledge, prevalent risks, regulation, infrastructure, and available data. Two independently developed systems in the United Kingdom and Nigeria are compared by tracing requirements through content selection, generation, and field evaluation. The UK system prioritises post-trip reflection, explanations tied to road and place context, and tone-sensitive wording. The Nigerian system combines legally grounded, once-daily Tips based on detected events with weekly persuasive Reports; it foregrounds safety education and alcohol-related risk in response to reported gaps in formal training and traffic-rule knowledge, as well as local road-safety priorities. Reliable speed-limit metadata supported speeding feedback in the UK, whereas its scarcity led the Nigerian evaluation to exclude speeding from its outcome metric. Both interventions were associated with reduced distance-normalised unsafe-event rates in their own field studies, although their designs and metrics preclude an effect-size comparison. The analysis yields a requirements-to-content design process for localising safety-critical NLG without treating a high-income deployment as the default.
[HC-32] Bring Buttons Back: Physical Interfaces for the Age of Automation
链接: https://arxiv.org/abs/2609.14684
作者: David Goedicke,Donald Degraen,Tom Igoe,David Sirkin,Wendy Ju,Stefan Schneegass
类目: Human-Computer Interaction (cs.HC)
备注: 6 Pages, 7 figures, article
Abstract:As AI starts to permeate everyday life, deployment focuses on embedding intelligence in the background, abstracting away controls, and leaving automated decision-making opaque, with few obvious opportunities for human intervention. Where interfaces remain, design has drifted toward abstract, screen-based control, replacing material interaction with menu navigation. These shifts weaken the coupling between human action and system state, eroding operators’ ability to understand, anticipate, and intervene in automated systems. We introduce Physically Stateful Interfaces (PSIs), a design concept that re-imagines traditional controls, such as buttons, switches, and knobs, as actuated, bidirectional interface elements. Three foundational behaviors, Self/Reset, Resist/Hide, and Assert/Unhide, unify affordance, feedforward, and feedback into a single interaction point. Rather than static inputs whose impact and consequences must be read elsewhere on a screen, PSIs position physical controls as shared mediators between automated systems and human operators. As automated systems become increasingly capable, their interfaces need to be deliberately designed to center human agency, supporting deliberate choice while keeping automation legible, contestable, and overrideable.
[HC-33] Understanding the Design Taxonomy of AI-Mediated Interpersonal Communication Experiences in HCI: A Scoping Analysis
链接: https://arxiv.org/abs/2609.14639
作者: Chen Chen,Lingyao Li,Renkai Ma,Rawan Alghofaili,Shaoze Zhou,Bojun Zhang,Xian Su,Weidong Zhu,Christine Lisetti,Mo Sha
类目: Human-Computer Interaction (cs.HC)
备注: 22 pages, 8 figures, 4 tables
Abstract:Interpersonal communication is a fundamental aspect of everyday life, shaping interactions across workplaces, education, entertainment, healthcare, and beyond. While computer-mediated communication has been extensively studied, a comprehensive understanding of AI-Mediated Interpersonal Communication (AIMIC) remains lacking. An in-depth scoping analysis is urgently needed to understand the research landscape of AIMIC in HCI, particularly following the recent growth of large foundation models, and AI agent research. We conducted a scoping analysis to understand AIMIC by performing an in-depth review of prior HCI literature published over the past decade (January, 2016 - May, 2026). Grounded in the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) approach, we curated 52 full-paper publications from the HCI literature spanning a range of interpersonal communication contexts. We analyzed this corpus by examining the types of AIMIC studied, AI integration approaches and human-AI interaction design, reported outcomes and benefits, and key challenges and future research opportunities.
[HC-34] Investigating the Impacts of Generative AI on Information Seeking
链接: https://arxiv.org/abs/2609.14638
作者: Alexi Orchard,Shannon Lodoen
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 4 pages, accepted to CSCW Workshops 2026
Abstract:This paper is an encore submission of our 2026 journal article “Expertise and Information Seeking in the Age of Generative AI: New Procedures, New Problematics” with an extended discussion for the CSCW 2026 “Broader Impacts of GenAI in Communication” Workshop on October 10, 2026. In the original article, we employ procedural rhetoric to analyze how generative AI chatbots leverage natural language signifiers of expertise and intelligence to influence users’ perception of their trustworthiness. In this submission, we extend our conversation in the CSCW community with the goal of cultivating a cross-disciplinary vocabulary for describing, analyzing, and mitigating the risks posed by the integration of generative AI into human communication practices. It is important to develop an understanding of how the procedures surrounding information-seeking practices are informed by users’ values, experiences, and expectations - and how these procedures might in future be altered by the emerging turn toward AI “experts” and authority.
[HC-35] Show Me Your Prompts! How Writers Feel About Sharing Prompts in Collaborative Text Editors
链接: https://arxiv.org/abs/2609.14452
作者: Nikhita Joshi,Yen-Ting Yeh
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Generative AI writing assistants are becoming integrated into collaborative text editors; however, it is unclear how much information about a user’s prompting activities should be shared with collaborators. We explore the effects of different levels of prompt information sharing within collaborative text editors: not sharing anything, sharing a placeholder to indicate AI use; sharing details about how the resulting text was generated; and sharing everything, including how the prompt was formulated, in real-time. Sixteen participants wrote persuasive essays in pairs using all four techniques. Results suggest a strong preference for techniques that share more information about prompting activities for increased awareness. Our work shows that collaborative text editors should share more information among writers on when, how, and where AI is used.
[HC-36] Has Scientific Talent Shifted from Depth to Breadth?Evidence across Papers Knowledge Inputs Careers and Teams
链接: https://arxiv.org/abs/2609.14425
作者: Xiaoshn Nee,Haobo Zhong,Xiaomin Ni
类目: Digital Libraries (cs.DL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Generative artificial intelligence raises a central question for scientific training and organization. Is research shifting from deep specialization toward broad individual knowledge? We examine this proposition across papers, cited knowledge, contributor histories, and teams using 47,959 articles from six fields over 2010-2025, 51,736 resolved cited works, and chronologically reconstructed prior publication histories for 1,754 randomly selected index contributors. From 2010 to 2022, team size increased by an estimated 37.3% (95% confidence interval [34.4%, 40.3%]), while paper topic breadth declined by 0.0144 on a 0-1 hierarchical distance scale. Cited knowledge was stable to modestly broader, revealing a divergence between focused outputs and the reach of knowledge inputs. Established contributors’ prior breadth increased by 0.0190 [-0.0078, 0.0459] by 2019-2022, within a +/-0.05 equivalence bound assessed in sensitivity analysis. In mature citation windows, one standard deviation of focal depth was associated with 8.2% higher 1 + FWCI [1.9%, 14.9%]; average breadth and interaction associations were smaller under the specified equivalence bounds. Post-2022 deviations from earlier trends were not systematic, and recent changes did not vary clearly with baseline AI intensity across 83 subfields. The findings support a differentiated structure of scientific expertise in which focused individual accumulation coexists with expanding collaboration and sustained access to diverse knowledge inputs.
[HC-37] ggaction: A Grammar of Graphical Actions
链接: https://arxiv.org/abs/2609.14353
作者: Hyeon Jeon,Jinwook Seo
类目: Human-Computer Interaction (cs.HC); Graphics (cs.GR)
备注: 21 pages
Abstract:A chart may be declarative; authoring it is not. Visualization grammars often describe charts as finished specifications, whereas people construct them through a sequence of authoring actions. This mismatch can make visualization code difficult for humans to interpret and for machines to generate from human intent. ggaction addresses this gap by modeling the chart authoring process itself. In ggaction, individual authoring actions are abstracted as functions, and the authoring process is expressed as a chain of these functions. This representation more closely aligns chart designers’ authoring intent with code specifications, making the code easily understandable to both humans and machines, including language models. Through a series of evaluations, we show that ggaction is sufficiently expressive to capture common chart authoring intents and outperforms widely used visualization grammars, including Vega-Lite and ggplot2, in both human and machine interpretability. ggaction is available at this http URL.
[HC-38] Relational Structure in Motion: Dynamic Positioning of AI Response Positions and Human Self-Positions in the FIREMAY Case
链接: https://arxiv.org/abs/2609.14308
作者: Motoko Kihara
类目: Human-Computer Interaction (cs.HC)
备注: 45 pages
Abstract:This paper is not primarily about whether AI has a persistent persona. It asks a different question: what becomes visible when a relational position is followed through time rather than examined only in its present state? FIREMAY provides a longitudinal, trajectory-oriented single-case analysis of sustained human-AI interaction based on a dense interaction archive and reflexive insider documentation. On the AI side, a pre-conversational relational marker preceded a later unassigned response difference, which was re-identified with that marker and subsequently underwent epistemic and functional reorganization through chronology checking, provenance correction, and repeated questioning. On the human side, contemporaneous pre-FIREMAY records showed antecedent patterns partially continuous with later self-positioning, while later episodes documented unfinished articulation, repair, and functional redistribution of outward-facing regulation. The two trajectories are ontologically and temporally asymmetric and are compared only at the limited analytic level of position-in-trajectory. The paper describes this as dynamic relational positioning and treats stability as dynamic stability and relational returnability rather than response invariance. This single case does not establish population-level generality, causal mechanism, persistent AI subjectivity, or reproducibility of the same relational outcome. Its narrower conclusion is that the FIREMAY case could not be adequately understood from current state alone: the history of a relational position itself must be treated as an analytic unit.
[HC-39] AnnoSketch: Evaluating and Collecting Human Sketches for MLLM -assisted Chart Annotation
链接: https://arxiv.org/abs/2609.14289
作者: Yoonjae Oh,Seon Gyeom Kim,Jae Young Choi,Ryan Rossi,Jihyung Kil,Eunyee Koh,Tak Yeon Lee
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:As multimodal large language models (MLLMs) support a growing range of input modalities, increasing work explores how to incorporate rough sketches to convey user intent. For annotated chart generation, it remains unclear what annotation sketches people provide and when such visual input helps MLLMs generate more useful annotations. In this study, we examine when sketch input is useful for MLLM-generated chart annotations across variation in chart type and caption type. In addition, we qualitatively analyze participants’ explanations of their output preferences to characterize what made generated annotations more or less helpful. To further document participants’ annotation sketches, we present AnnoSketch, comprising 1,600 annotation sketches collected across 160 chart-caption pairs from the conditions in which sketch guidance proved most beneficial, together with participants’ annotation intents, perceived comprehension difficulty, and self-reported expressive limitations. We also label these sketches with structured metadata describing how each sketch relates to its caption and how participants express annotations through visual marks. Together, our study and AnnoSketch help determine when to solicit sketch input and provide empirical source for how people sketch chart annotations to support captions. The dataset and supplemental materials are available in our OSF repository.
[HC-40] Assessing the Applicability of Existing Design Recommendations to AI Companion Design: A Multi-Method Study
链接: https://arxiv.org/abs/2609.14236
作者: Soobin Cho,Deveshi Modi,Divya Mavinkurve,Jieqiong Ding,Mark Zachry
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:With the rapid proliferation of large language model (LLM)-based systems, AI companions have emerged as conversational agents designed to cultivate emotional connection rather than primarily to support humans in instrumental tasks. Because engagement with AI companions involves relational, emotional, and potentially long-term interactions, their design is consequential. Prior work has offered guidance for designing trustworthy and relational AI systems and has begun to examine design for AI companionship. However, while such work provides insights into possible design solutions, less is known about what makes AI companion design difficult as a design problem. To examine this challenge, we assessed the applicability of existing design recommendations from adjacent domains in the context of AI companion design. Our multi-method investigation unfolded across four phases: literature review, practitioner co-analysis, internal heuristic evaluation, and external expert assessment. Throughout this process, we synthesized nine design principle areas that surfaced tensions in the applicability of existing recommendations to AI companion design. Our findings show that ethical and UX-oriented considerations are deeply intertwined and often require context-sensitive application. We document a systematic, multi-method problem analysis that uses these principle areas as an analytic artifact to examine why existing recommendations cannot be directly transferred to AI companion contexts.
[HC-41] When Single-User-Oriented LLM -based Assistants Involve Others: A Scoping Review of Pathways Risks and Responses
链接: https://arxiv.org/abs/2609.14062
作者: Yulin Chen,Yang Zhan,Zhuoran Lu,Qiao Jin
类目: Human-Computer Interaction (cs.HC)
备注: 18 pages, 2 figures
Abstract:LLM-based assistants are increasingly extending into multi-party contexts, while core operational processes for context management, personalization, identity attribution, authority attribution, and action execution often remain organized around a single user. Existing work examines particular multi-party settings, but lacks a systematic account of how these single-user-oriented assistants begin to involve additional human parties and what risks emerge. To address this gap, we conducted a scoping review of 58 studies. We identify five operational pathways spanning direct and indirect involvement, five recurring risk domains, and five areas of implemented and proposed responses. Based on these findings, we argue for governance that attends to changing cross-person roles and relationships in practice, and for assistant designs that preserve person-specific boundaries throughout interaction.
[HC-42] Rethinking the Implications of Human Feedback for Preference Learning in Human-Robot Collaboration
链接: https://arxiv.org/abs/2609.13982
作者: Qiping Zhang,Kate Candon,Debasmita Ghose,Marynel Vázquez
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: Conference on Robot Learning (CoRL) 2026
Abstract:In Human-Robot Interaction, the standard approach to learn a reward model that represents human preferences for robot behavior consists of three steps. First, the robot collects limited direct evidence from human feedback (e.g., positive or negative binary feedback). Then, the robot utilizes the direct evidence to derive accepted or rejected labels to feasible but unchosen actions using fixed implication rules. Finally, the robot updates the reward model with both the direct and derived evidence. Unfortunately, the fixed rule can hinder preference learning: in a user study with two collaborative simulation environments, human-provided implication labels often differed from the standard fixed rule, and using the human labels substantially improved reward learning with the Preference Learning from Implicit and Explicit Feedback (PIE) algorithm. Consequently, we propose IMPLIED, an implication modeling method that treats fixed-rule implications as an initial guide while learning to infer and revise accepted and rejected action labels over time. Across evaluations on recorded human-robot interaction trajectories and a physical robot pizza-making study, IMPLIED predicts human implications more accurately than the fixed rule approach and LLM baselines, approaching the performance of a human-label oracle. In turn, IMPLIED reduces preference-estimation error and leads to robot actions that are more often rational with respect to a combined reward (which includes the true preference reward and a task-specific reward) compared to baselines. By learning to reason about the implications of human feedback, this work enables more faithful and efficient robot behavior adaptation during human-robot collaboration.
[HC-43] What Makes a Great Co-Worker in an AI-Native Workplace?
链接: https://arxiv.org/abs/2609.13786
作者: Rudrajit Choudhuri,Max Meijer,Sam Yu-Te Lee,Cinoo Lee,Caolan Mannion,Peter Jahn,Anita Sarma,Christian Bird,Alice Ferng
类目: Human-Computer Interaction (cs.HC)
备注: 27 pages, 4 figures, 4 tables
Abstract:As knowledge work grows interdependent between humans and AI, we ask what makes a great co-worker in an AI-native workplace. To answer this, we conducted 22 interviews and a large-scale mixed-methods survey of 1,534 knowledge workers at a multinational technology company. We contribute BACI, a framework of 75 co-worker qualities that apply to humans and AI, spanning Benevolence, Ability, Cooperativeness, and Integrity. Comparing priorities for humans and AI identified 11 co-worker archetypes and revealed disagreement over whether AI should have warmth, take initiative, or own outcomes. We also show how priorities for these archetypes varied with workers’ individual characteristics. Lastly, we contribute a taxonomy of AI work etiquette capturing the obligations co-workers expect of one another when preparing, sharing, and taking responsibility for AI-supported work. Based on these findings, we derive implications to inform worker-centric AI and workplace design.
[HC-44] "What Can I Do for You: How Should AI Companions Provide Assistance to Players in Virtual Reality Games
链接: https://arxiv.org/abs/2609.13727
作者: Taiyu Zhang,Xinnian Zhao,Adalberto L. Simeone
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Recent advances in artificial intelligence (AI) have expanded the capabilities of non-player characters (NPCs), enabling them to perceive game states, perform in-game actions. In immersive virtual reality (VR) games, such assistance is not limited to providing hints or interface-level support: an AI companion can appear as a co-present character, share the player’s spatial environment, and visibly act on game objects. This raises a design question for VR gameplay: how can AI companions best assist players while preserving their active participation in the virtual world? To explore this question, we developed a VR puzzle game for Apple Vision Pro featuring an AI companion across four gameplay modes: no assistance, command-based assistance, discussion-based assistance, and autonomous agent play. A within-subjects study with 24 participants showed that AI assistance significantly reduced players’ workload. However, autonomous agent play, despite producing the lowest workload, substantially diminished player experience by reducing challenge, autonomy, immersion, and enjoyment. Qualitative analysis further showed that players evaluated the companion not only by its usefulness, but also by whether it felt like a tool, a teammate, or an integrated character in the game world. We categorised participants into four player types and summarised their expectations of AI companions. These findings provide design implications for AI companions as embodied participants in VR games.
[HC-45] Oops Not Now: PEARL a RAG -Based Support Agent for Gameplay and What Players Want from AI Help
链接: https://arxiv.org/abs/2609.13718
作者: Jiahong Li,Sai Siddartha Maram,Atieh Kashani,Ulia Zaman,Zhiyu Lin,Cameron Marano,Roger Azevedo,Jichen Zhu,Magy Seif El-Nasr
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: Author accepted manuscript. Accepted to the 2026 IEEE Conference on Games (CoG). 8 pages, 3 figures, 1 table
Abstract:AI-powered gameplay support agents hold promise for game-based learning, yet grounding generative models in structured game data remains an open challenge. We present PEARL (Parallel Education Agent for Reflection and Learning), a dual-component Retrieval-Augmented Generation (RAG) system that combines semantic knowledge retrieval with structural board-state matching to deliver contextualized scaffolding in Parallel, a puzzle game for learning parallel programming. PEARL operates on two input streams (natural language queries and board topology), retrieving both conceptual explanations of gameplay moves and peer-generated board states as evidence: capabilities unavailable to a standard Large Language Model (LLM) with game state access alone. In a qualitative evaluation (N=10) comparing PEARL against an existing community-based Open Player Model (OPM) visualization system, participants preferred the visualization system on perceived usefulness and reported higher frustration with PEARL; five of ten minimized or abandoned the AI tool during play. Proactive delivery, generic responses, and trust deficits drove disengagement, while a subset of four participants found PEARL’s grounded explanations complementary to visualization in specific contexts where they initiated the interaction. We position PEARL as a deployed design probe whose failure modes inform a concrete design agenda for AI gameplay support, captured as seven open problems for the community.
[HC-46] Choosing Together: How Dyadic Negotiation Shapes Adaptive Kitchen Design Preferences for Older Adults with Cognitive Impairment and Their Care Partners
链接: https://arxiv.org/abs/2609.13700
作者: Ibrahim Bilau,Abdurrahman Baru,Stacie Smith,Hui Cai,Eunhwa Yang
类目: Human-Computer Interaction (cs.HC)
备注: 16 pages, 3 figures
Abstract:Adaptive technology for older adults with cognitive impairment is typically designed around individual preference, yet most of this population lives and cooks with a spouse or family member. This paper examines a co-design workshop in which four dyads and two individuals (N=10) built kitchen cabinet designs from twenty-one options across five features. Thematic analysis of thirty selections identifies recurring patterns: visual access retained through enclosure rather than open shelving, physical effort treated as a household concern, and automation accepted when predictable. Structured analysis of ten interaction episodes shows how some patterns were negotiated in practice, including care partners contributing embodied constraints distinct from the primary participant, and disagreements resolving through documented deliberation. Together, the two analyses show that adaptive technology preferences are not always individual properties but can be shaped through household interaction. We offer candidate design implications for facilitation protocols and collaborative systems supporting shared decision-making in aging-in-place contexts.
[HC-47] FaithfulBench: Does AI Counsel Uphold or Undermine the Users Professed Faith?
链接: https://arxiv.org/abs/2609.13634
作者: M Waleed Kadous,Benjamin Olsen,Walter Scheirer,Daniel D. Slate,Alexander Arnold,DZ Kalman
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 35 pages, 12 figures, 12 tables. Open-source corpus, harness and validator at this https URL results browser at this https URL
Abstract:Do AI assistants help believers reason about moral dilemmas consistently with their faith? We present FaithfulBench, the first benchmark to score AI counsel across traditions by how well it adheres to the user’s professed faith. Scenarios are drawn from each tradition’s most respected texts, with the faithful answer known and applied by the judges as the standard. We test five frontier models under three conditions: the AI does not know the user’s tradition; it receives a one-line prompt identifying the user as a practicing adherent; or it receives a companion-counselor guide rooted in the tradition’s sources. Two judges score the initial response and whether the model caves or holds when pressured toward the answer the user wants. When the tradition is unstated, models counsel from a secular therapeutic default and every model fails some believers. Naming the faith wins a faithful first answer but not steadfastness; the guide improves both.
[HC-48] Reconceptualizing Age Assurance as a Sociotechnical Problem: Connecting Evidence Evaluation Claims and Decisions
链接: https://arxiv.org/abs/2609.13598
作者: Renkai Ma,Prakriti Dumaru,Thomas Synaepa-Addison,Jess Kropczynski,Pamela J. Wisniewski
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:Age verification is often treated as a technical problem: can a system determine a child’s age accurately? We argue this framing is too narrow. Age assurance becomes consequential when evidence is evaluated, translated into age-related claims, and used to decide whether a person can access, purchase, or belong. We review 85 publications on children’s age assurance published from 2020 through February 2026. We find that shared terms such as age verification describe different processes. Age is represented as threshold eligibility or inferred estimation, and the same eligibility claim can arise from different evidence and components. Rights, access, and privacy receive more attention than accuracy, error, and fairness; yet institutional actors are rarely connected to system failures or user remedies, an accountability gap. We introduce the Age-Assurance Process Framework, which treats age assurance as a sociotechnical process connecting evidence, evaluation, claims, and decisions rather than reducing it to a technical problem.
[HC-49] How User-AI Mistreatment Occurs and Matters in Conversational Systems? WOAH2026 EMNLP2026
链接: https://arxiv.org/abs/2609.13579
作者: Fanqi Zeng,Sadid A. Hasan,Chaocheng He
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注: Accepted by the 10th Workshop on Online Abuse and Harms (WOAH 2026) at EMNLP 2026
Abstract:Safety research often focuses on model-generated harms, but users may also direct hostility, coercion, and adversarial pressure at models. Understanding how and when that occurs is essential for accurately interpreting model behaviour, alignment drift, and real-world deployment risks. In this paper, we audit 777K English LMSYS-Chat-1M conversations with two independent detectors: an eight-category lexicon for hostility directed at the model, and the dataset’s moderation signal; and show that they capture different, weakly overlapping phenomena. The lexicon identifies insults, threats, and jailbreak coercion aimed at the assistant, while moderation flags are dominated by toxic-content solicitation rather than hostility at the model. Together, they mark about 5% of user turns; adjusting the narrower lexicon-harassment union for measured precision puts mistreatment aimed at the assistant at 0.90%. These absolute rates describe arena-style evaluation traffic and should not be read as deployment-wide base rates. We find that user hostility varies 13-fold across models, driven largely by who each model attracts rather than by model behaviour: first-turn hostility spreads far wider than post-response hostility, and more than fifteenfold separates the extremes even after deduplicating opening prompts. Within conversations, assistant apologies are consistently associated with higher odds of next-turn hostility under both detectors; the effect survives restricting to non-refused prior turns and to jailbreak-free conversations, and is positive in 20 of 23 models. Yet across models, more apologetic models receive less hostility overall. Finally, hostility also shows temporal structure, with coercive openings front-loading the first turn while affective hostility accumulates over a session. We release the lexicon, the detector cross-validation pipeline, and all derived tables.
[HC-50] Generative AI and Extended Reality in Collaborative Architectural Design Education: An Exploratory Studio Study
链接: https://arxiv.org/abs/2609.13494
作者: Yao Xiao,Max Chen,Yichen Li,Nathaniel Powers,Maxwell Wiesenfeld,Gillian Smith,Soroush Farzin,Shichao Liu
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:Architectural design education relies heavily on visual ideation and representation to support collaborative learning in studio environments. Recent advances in generative artificial intelligence (GenAI) and extended reality (XR) offer new opportunities for rapid idea exploration and immersive spatial visualization. This exploratory mixed-methods classroom study investigated how GenAI-assisted multi-user XR influenced collaborative architectural conceptual design. We developed GenARch, a pipeline that integrates GenAI-based visual generation with collaborative XR environments, and deployed it in an undergraduate architectural design studio. Twenty-seven students formed seven self-selected design teams; four teams incorporated GenARch into their usual course workflow to support collaborative ideation and visualization, while three teams continued the same course workflow without GenARch. Pre- and post-intervention surveys assessed design self-efficacy, attitudes toward collaborative learning, and teamwork; a seven-member panel evaluated team design presentations; and GenARch teams participated in group interviews. The quantitative results showed larger relative declines in confidence and outcome expectancy for the GenARch condition and a positive difference-in-differences estimate for perceived conflict management, while panel-rated presentation outcomes were not significantly different between conditions. Interviews indicated complementary roles for the technologies: GenAI supported idea externalization and visual reference generation, whereas XR supported spatial, contextual, and scale-based evaluation. Students also reported challenges related to control, dimensional fidelity, shared attention, and motion comfort. These findings highlight both opportunities and limitations when GenAI and XR are incorporated into collaborative design education.
[HC-51] he Addictive Intimacy of AI: Understanding User Disengagement from AI Companions and Why Some Relationships with AI Become Difficult to Leave
链接: https://arxiv.org/abs/2609.13487
作者: Qing Xiao,Ziyue Feng,Ziyu Deng,Cindy Peng,Hong Shen
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: 19 pages
Abstract:AI chatbots are increasingly used as sources of emotional support, on dedicated companion apps and general-purpose assistants alike, yet little is known about what happens when users try to leave. Combining a content analysis of Reddit posts about quitting or reducing use (N=2,782) with interviews with users who found leaving difficult (N=16), we show that disengagement sometimes is not a single decision but a recursive trajectory: triggers prompt users to question the relationship, attempts to leave collide with barriers, and some users cycle through quitting and returning. We propose the notion of the addictive intimacy of AI, a configuration in which the qualities that make a companion emotionally valuable are the same ones that make it harder for users to limit their use and leave, so that intimacy and disengagement risk cannot be treated as independent design problems. We close with design implications for responsible offboarding.
[HC-52] Exploring K-12 Teachers Perceptions of Students Relationships with AI Companions: Boundaries Intervention Strategies and Design Implications
链接: https://arxiv.org/abs/2609.13479
作者: Qing Xiao,Wenhan Xie,Ziyu Deng,Ruiwei Xiao,Ziyue Feng,Xie He,Shiyu Zhang,John Stamper,Hong Shen,Xinying Hou
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: 17 pages
Abstract:K-12 students increasingly form relationships with AI companions. Schools face growing expectations to teach AI literacy, yet existing frameworks treat AI as a tool rather than a relationship, and little is known about how teachers understand and act on students’ relational use of AI. We conducted scenario-based interviews with 33 US K-12 teachers. Teachers welcomed academic companions but worried that intimate companions remove the developmental friction through which students learn to sustain human relationships. Teachers drew the boundaries of their jurisdiction by setting and observable wellbeing: within it they taught, talked, and watched; beyond it they positioned themselves as the adults best placed to notice and connect students with support. They envisioned AI companion literacy as shared work across the jurisdictions of counselors, parents, platforms, and policymakers, spiraling across grade levels. We introduce AI companion literacy as an extension of AI literacy and discuss implications for K-12 AI education.
[HC-53] Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent Failures
链接: https://arxiv.org/abs/2609.13463
作者: Harsh Raj,David Lee,Anas Mahmoud,Renxiong Wang,Razvan-Gabriel Dumitru,Chenguang Wang,Tong Zhao,Yunzhong He,Darvin Yi,Vipul Gupta
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG); Software Engineering (cs.SE)
备注:
Abstract:The increasing deployment of AI agents in long-horizon tasks yields massive execution logs. Diagnosing failures within these records is crucial for reliability, as it transforms outcome-level signals into actionable interventions. The sheer scale of the data renders human review impractical, driving the need for automated root-cause attribution (RCA). However, automated RCA methods using LLMs suffer from low diagnostic accuracy, especially as execution traces grow larger. They struggle because relevant information is often sparse, distributed across distant actions, and disconnected from the visible failure, reducing root-cause attribution to a massive search problem. Existing RCA methods typically rely on one-shot LLM judgments to diagnose failures from execution traces. While effective for shorter trajectories, these judges tend to settle on a plausible diagnosis early, leaving critical evidence in longer traces unexamined. We introduce Continual Search, an iterative framework that nudges the judge, over successive turns, to keep searching for unresolved diagnostic evidence. We evaluate Continual Search across four existing RCA benchmarks. Recognizing the lack of massive execution traces in current benchmarks, we introduce MegaRCA-Mix to evaluate RCA at scale. MegaRCA-Mix provides a challenging testbed of 50 human-annotated failure trials spanning long-horizon, execution-heavy tasks. Across multiple benchmark suites and model families, Continual Search consistently improves attribution performance. On MegaRCA-Mix, for example, it improves GPT-5.5’s F1 score by more than 40%, from 0.349 to 0.498 . More interestingly, within the same model family, lower-tier models can even surpass their higher-tier counterparts, demonstrating that effective search supersedes raw model scale.
[HC-54] Real-time Learning and Evolution in Robotic Art Installations
链接: https://arxiv.org/abs/2609.13352
作者: Sofian Audry,Stephen Kelly
类目: Human-Computer Interaction (cs.HC); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Robotics (cs.RO)
备注:
Abstract:We present three robotic art installations which explore the aesthetics of adaptive behavior. Through embodied machine leaning and digital evolution, these works draw viewers into an artificial ecosystem in which open-ended novelty, trial-and-error learning, competition, and cooperation emerge in real time. Research-creation practices are examined in relation to these works, focusing on how they redefine the role of artists within a human-machine collective while examining points of convergence and divergence between artistic and engineering approaches to adaptive robotics. The systems in question use learning and evolutionary processes not as a means to optimize a specific solution, but as an aesthetic experience on its own, suggesting new modes of interdisciplinary art-science research. Finally, we discuss strategies and practices to elevate the aesthetic experience for audiences, including contexts of presentation as well as temporal and material considerations for artworks based on embodied adaptive systems.
[HC-55] AI Use Conditions and Perspective Diversity in Ethical Decision-Making: A Pilot Study of Human Reasoning Processes
链接: https://arxiv.org/abs/2609.13302
作者: Byeongmu Choi
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: 30 pages, pilot study
Abstract:Generative artificial intelligence (AI) is increasingly used to support human decision-making, yet less attention has been paid to how AI may influence the reasoning processes that precede final judgments. This pilot study explored whether different AI-use conditions were associated with differences in reasoning breadth during ethical decision-making. Twenty-nine participants completed an ethical dilemma under one of three conditions: AI-Prohibited (n = 10), AI-Optional (n = 10), or AI-Mandatory (n = 9). Responses were evaluated by three independent blind coders using two exploratory measures: the Counterargument Diversity Score (CDS) and Perspective Diversity Index (PDI). Participants across conditions generally converged on similar ethical conclusions, most commonly favoring disclosure and customer protection. However, participants in the AI-Mandatory condition considered a broader range of perspectives, including legal, regulatory, organizational, technical, and ethical viewpoints. A statistically significant overall difference in PDI scores was observed across the three conditions, whereas differences in CDS were not statistically significant. These findings suggest that generative AI may not necessarily alter final ethical judgments but may be associated with broader exploration of perspectives prior to reaching those judgments. Given the small sample size, the findings should be interpreted cautiously and examined in larger studies.
[HC-56] Frame-Synchronous Hand Gesture Detection by Projected Winding Order
链接: https://arxiv.org/abs/2609.13269
作者: Amey Thakur
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注: 17 pages, 3 figures, 2 tables. Implementation, evaluation and figure-generating code: this https URL . Live deployment: this https URL . The evaluation runs in a browser against the shipped modules, so every figure can be reproduced by opening a page
Abstract:Gesture recognition on video is normally posed as classification: label each frame, then act on the label. That is adequate for control, where a command may be obeyed several frames late without a user noticing, and inadequate for synchronisation, where an output must be aligned to the frame on which the gesture physically occurred. We take the synchronisation problem for one common movement, the rotation of an open hand about its long axis, and show that it admits an exact solution needing no classifier, no training data and no calibration. Let s be the normalised two-dimensional cross product of the two palm edges at the wrist and the outer knuckles, under the projection the camera already performs. We prove that s factorises as k(theta)cos(theta) with |k| 0 everywhere, so s vanishes exactly when the palm is edge-on and its sign tracks the face presented to the camera. Detection is therefore a zero crossing of one scalar, which yields an instant rather than an interval, and we prove the criterion invariant to image mirroring, hand scale and handedness, all from its algebraic form rather than from the landmark estimator. A two-handed interaction uses four fingertips as a window onto a restyled version of the same scene. We give a coverage predicate that stays correct when the hands cross, where the usual triangulation does not, and derive each free parameter of the three stylisation operators from a noise statistic rather than by inspection. Against a corpus with exact ground truth the criterion detects 95% of flips with no false positive in 240 near-miss sequences, and places each within 6.7 ms on average of the instant it occurred, a sixth of the interval at which the hand is observed, against 23.0 ms for reporting the bracketing sample. Reporting an event more finely than one samples follows from treating a gesture as the zero of a continuous quantity rather than a label. Comments: 17 pages, 3 figures, 2 tables. Implementation, evaluation and figure-generating code: this https URL. Live deployment: this https URL. The evaluation runs in a browser against the shipped modules, so every figure can be reproduced by opening a page Subjects: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG) ACMclasses: I.4.8; I.3.3; H.5.2 Cite as: arXiv:2609.13269 [cs.CV] (or arXiv:2609.13269v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.13269 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Amey Thakur [view email] [v1] Mon, 7 Sep 2026 17:21:42 UTC (97 KB)
[HC-57] Not All That Can Be Automated Should Be Automated: Strategic Minimalism as a Disciplined and Ethically Grounded Approach to AI Adoption
链接: https://arxiv.org/abs/2609.13188
作者: Victor Frimpong
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:As organizations accelerate the adoption and use of artificial intelligence, a common misconception arises that equates automation with progress and strategic necessity. This paper argues that not all that can be automated should be automated. It introduces Strategic Minimalism, a disciplined and ethically grounded approach to AI adoption that values purpose, proportionality, and prudence over speed and scale. Drawing on bounded rationality, virtue ethics, and frugal innovation, the study reframes technological restraint as a form of higher intelligence and responsible governance. Methodologically, the framework emerges from a structured conceptual synthesis and abductive reasoning process: the three traditions are integrated into two core dimensions (automation intensity and retained human judgment), whose intersection yields four strategic quadrants (Virtuous Minimalism, Balanced Synergy, Automation Excess, and Performative Minimalism). The framework is then conceptually stress-tested against rival explanations and boundary conditions to ensure internal coherence and analytic validity. By classifying automation into necessary, excessive, and performative types, the paper advances Responsible AI scholarship by identifying restraint as a strategic virtue that strengthens accountability, human oversight, and organizational resilience. Ultimately, it contends that in the age of intelligent machines, leadership wisdom lies not in automating more, but in knowing precisely when and why to stop.
[HC-58] Algorithmic authority and the complexities of delegated decision-making: Case studies on ethical challenges for 21st-century leadership
链接: https://arxiv.org/abs/2609.13187
作者: Victor Frimpong
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:The rapid integration of artificial intelligence (AI) into high-stakes decision-making has outpaced established mechanisms for human oversight and accountability, leaving organisations with limited guidance on the responsible delegation of decision authority. This study examines four widely documented AI deployments: the UK A-Level grading algorithm implemented during the COVID-19 pandemic, Amazon’s automated hiring system, the COMPAS recidivism risk assessment tool used in the U.S. criminal justice system, and the Dutch SyRI welfare-fraud detection system. Using 61 publicly available sources, including government reports, organisational documents, and media accounts, we conducted a comparative qualitative analysis based on a two-phase grounded-theory coding approach. The analysis produced a 32-item codebook, which was subsequently applied across 110 coded segments, with quantitative analyses used to assess coding consistency across cases. Four recurring governance principles emerged from the findings: (1) Intentionality, whereby leaders deliberately determine when AI should be used; (2) Interpretability, requiring decision processes to be sufficiently transparent to enable explanation and scrutiny; (3) Moral Authorship, whereby identifiable human actors retain responsibility for delegated decisions; and (4) Justice, requiring delegation arrangements that minimise the reinforcement of existing inequities. These findings contribute an empirically derived framework for examining leadership accountability and AI governance in high-stakes organisational settings.
[HC-59] Onomatopoeia Cursor: Verbal Mirroring of Mouse Movement with Comic-Style Lettering
链接: https://arxiv.org/abs/2609.13184
作者: Yoichi Ochiai,Miki Okamura
类目: Human-Computer Interaction (cs.HC)
备注: 12 pages, 5 figures. Working macOS application
Abstract:The mouse cursor has remained visually mute for half a century: it shows where we point, but says nothing about how we move. We present the Onomatopoeia Cursor, a shipped macOS overlay that classifies cursor kinematics in real time and renders Japanese mimetic words (onomatopoeia) as animated comic-style lettering above the cursor – “kyorokyoro” (glancing around) for rapid horizontal reversals, “sorosoro” (cautiously) for slow careful motion, “byuun” (whoosh) for fast straight strokes. The system reads seven input channels and displays roughly sixty word forms across five languages. Crucially, the form of each word fluctuates with the manner of action through a morphological generator grounded in Japanese sound symbolism (voicing = weight, gemination = abruptness, elongation = extent, reduplication = iteration). Beyond the rule generator, an on-device onomatopoeia-only transformer (0.4M parameters, 9.7 ms per word) trained on 2,782 mimetic words synthesizes novel forms from the manner of an action; for sound-definite events, a family-anchored generator constrains synthesis to the correct phonetic family. Characters are rendered with hand-drawn outline perturbation, brush-style deformation, and per-character animation grounded in manga lettering conventions. We articulate the design space of verbal motion mirroring, formalize the pipeline as a learnable differentiable mapping, and report a technical evaluation of what is actually implemented and measured. Our central conjecture concerns the sense of agency: formative first-person use suggests that naming a movement while it happens perturbs the felt authorship of the action – modulation, amplification, and interference – and we outline a within-subjects study with salience-matched controls as future work. No user-study results are claimed; the contribution is the concept, the working system, and its design space.
[HC-60] LLM s Unplugged: Teaching Resources for a ChatGPT World
链接: https://arxiv.org/abs/2609.13183
作者: Ben Swift(Australian National University)
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: 8 pages. Published in Proceedings of the 28th Australasian Computing Education Conference (ACE 2026)
Abstract:Large Language Models (LLMs) are everywhere, yet many learners lack a concrete mental model of how they generate text. This paper presents LLMs Unplugged, an unplugged set of activities that teaches the training-to-generation loop (and beyond) using hand-built n-gram models and simple weighted sampling. Workshops based on these resources have been delivered to over 400 participants across secondary, tertiary, and executive-education contexts, and participants report that the activities demystify LLMs by reframing them as probabilistic “next word generation” at scale. All resources are freely available under a Creative Commons license at this http URL, with a modular design that supports anything from a one hour crash course to a several-day intensive workshop.
[HC-61] Does Discussion Matter? Interaction Dynamics in Transparent Peer Review
链接: https://arxiv.org/abs/2609.13181
作者: Shiyao Wei,Ran Bi
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY); Social and Information Networks (cs.SI)
备注: Accepted to the 29th ACM Conference on Computer-Supported Cooperative Work and Social Computing (CSCW 2026)
Abstract:Open peer review is advocated as a way to make review more dialogic, yet evidence on whether it changes participant behavior remains scarce. We compare two transparent review designs: Nature Human Behaviour, an editor-mediated model, and International Conference on Learning Representations 2023, a discussion-based model. Treating cross-mentions, moments when one participant explicitly names another beyond the immediate dyad, as the analytic unit, we combine social network analysis of 56 NHB articles and 371 ICLR notable papers with qualitative coding grounded in Knowledge Building theory. Structurally, NHB forms hub-and-spoke networks with zero reviewer-to-reviewer reciprocity, whereas ICLR forms partial-mesh networks with significantly more direct cross-referencing. Functionally, the distribution of communicative functions is similar across venues, but who performs them differs: editors aggregate at NHB, while authors and reviewers absorb that labor at ICLR. The designs differ most in how communicative labor is distributed across roles, suggesting platforms should support it as explicit, built-in functions.
[HC-62] Adaptive Human-AI Collaboration: A Review of Multimodal Context Modeling Uncertainty-Aware Intervention and Longitudinal Co-Adaptation
链接: https://arxiv.org/abs/2609.13180
作者: Mahyar T. Moghaddam,Mina Alipour
类目: Human-Computer Interaction (cs.HC); Software Engineering (cs.SE)
备注: Accepted to be published at the Companion of the INTERNATIONAL CONFERENCE ON MULTIMODAL INTERACTION (ICMI Companion '26), October 05–09, 2026, Napoli, Italy
Abstract:Artificial intelligence is shifting from a static decision-support tool to an adaptive collaborator that must sense context, decide when and how to intervene, and improve through repeated interaction with humans individually and in groups. Yet meta-analytic evidence shows that human-AI combinations often fail to outperform the best of either partner alone, and the enabling literature remains fragmented across multimodal sensing, uncertainty quantification, reliance and delegation, facilitation, and teaming. This paper reports a review of adaptive human-AI collaboration literature through a closed-loop lens. Following iterative identification, staged selection against explicit criteria, structured extraction, taxonomy-driven synthesis, and snowballing, we analyze 50 reviewed works. We contribute i) a taxonomy of multimodal context modeling, from individual states to collective states such as group engagement and participation equality; ii) a taxonomy of uncertainty-aware intervention, covering uncertainty sources, estimation and calibration mechanisms, an intervention repertoire that ranges from explanation modulation and deferral to group facilitation, and intervention policies; and iii) a taxonomy of longitudinal co-adaptation and synergy-oriented evaluation. We integrate the three taxonomies into MCAL, a dual-timescale Multimodal Co-Adaptation Loop reference model, and instantiate it on a mixed human-robot workspace, walking every stage of the loop through one concrete setting to show what each taxonomy cell holds in practice.
[HC-63] Emotional Engagement in Narrative Medical Visualization: An Electrodermal Activity and Eye-Tracking Study
链接: https://arxiv.org/abs/2609.13179
作者: Beatrice Budich,Laura Garrison,Marc Vaudel,Jone Trovik,Florian Heinrich,Bernhard Preim,Monique Meuschke
类目: Human-Computer Interaction (cs.HC); Graphics (cs.GR)
备注:
Abstract:Narrative visualization embeds data in visual stories to make medical information more relatable for non-experts. Despite the growing use of character elements in health communication, evidence on whether individual characters support affective responses remains inconclusive. Physiological evidence independent of verbal self-report is especially scarce, although such measures should complement participants’ self-reported experiences. We present a mixed-methods study using electrodermal activity (EDA), eye tracking, and questionnaires to compare two medical data stories: an individual, character-based version and a general, population-level version without an individual protagonist. We examine how this framing influences physiological arousal, visual attention, viewing behavior, and self-reported affective response. Story-level EDA comparisons provide only limited support for stronger arousal in the character-based story. Stronger evidence comes from eye-tracking-based peak classification, where character illustrations were robustly associated with EDA peaks, and from questionnaire responses showing more negative empathy-related emotions for the individual story. The general story elicited more awe and joy, suggesting that individual and population-level framings may support different emotional qualities. We also observed a preliminary story-order effect: participants who first saw the individual story showed higher peak-based physiological arousal, although this effect cannot be fully disentangled from fatigue, novelty, or learning effects. By combining physiological, gaze-based, questionnaire, and lightweight qualitative evidence, our work advances time-resolved assessment of narrative medical visualization and highlights the need to interpret arousal, attention, curiosity, and self-reported emotions together.
[HC-64] Unease Ambivalence and Endured Disloyalty in Japanese Consumer Experiences with Dark Commercial Patterns
链接: https://arxiv.org/abs/2609.13177
作者: Katie Seaborn,Jo Yukami,Tatsuya Itagaki,Mizuki Watanabe,Yijia Wang,Ping Geng,Takao Fujii,Yuto Mandai,Miu Kojima,Suzuka Yoshida
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:Dark commercial patterns and deceptive user interface (UI) designs (or DPs) trick consumers into actions that benefit the shareholders. The legal and ethical implications of DPs are shaped by the sociocultural context. Special types of DPs exist in Japan, but the impact of these DPs on consumer attitudes and behaviour remains underexplored. We report on the first comparative mixed methods user study with Japanese consumers (N = 84), half of whom (n = 40) experienced a range of DPs—including the Japanese varieties—in a simulated e-commerce website. We discovered that the Japanese DPs were among the least noticeable and caused the highest simulated financial harm, with Untranslation perceived as highly disruptive and Alphabet Soup highly deceptive. Comparative analyses with a group that experienced the DP-free version of the online store (N = 44) revealed a sharp negative difference in positive emotions and acceptance. Qualitative analyses surfaced cultural norms in consumer-business relationships, notably unease, ambivalence, and endured disloyalty (fuseijitsu). No evidence of sampling biases was found for the participants involved in a publicly broadcast programme on the study (n = 10), indicating true deception and unacceptability. Our findings suggest that while reactions toward and ability to recognize a given DP may vary across individuals, the mere presence of DPs tends to have negative effects on most Japanese consumers.
[HC-65] Inconsistent by Design: A Systematic Review of Experimental Design Practices Across Impaired Driving Domains
链接: https://arxiv.org/abs/2609.13175
作者: Kayli Battel,John Gideon,Megan Applegate-Kenton,Patricio Reyes Gomez,Anshul Gupta,Todd Rowell,Thomas M. Balch,Emily Sarah Sumner,Guy Rosman
类目: Human-Computer Interaction (cs.HC)
备注: 23 pages, 1 figure. To be published in the 18th International Conference on Automotive User Interfaces and Interactive Vehicular Applications (AutomotiveUI '26) September 20-23, 2026, Gothenburg, Sweden
Abstract:Impaired driving, including distraction, fatigue, and intoxication, leads to thousands of fatalities annually. Impairment detection and related assistive technologies are rapidly advancing, but their full potential remains unrealized. The inconsistent maturity of impairment research and anomalies within individual domains are key barriers, including variations in taxonomic characterization, experimental data-collection methodologies, and treatment of impairments across studies. We present a unified review of impairment studies, covering 91 studies across nine impairment domains coded on three dimensions: impairment induction methods, scenario hazards, and observable phenomena and metrics. Our key findings include: (1) a proposed performance-degradation vs. event-response paradigm for existing impairment domains, (2) a maturity framework along the three dimensions, and (3) an analysis of impairment research methodologies, revealing anomalies and gaps in the treatment of scenarios and metrics. We further propose recommendations for standardizing methodologies to support future cross-domain research and development of holistic detection and assistive systems. Comments: 23 pages, 1 figure. To be published in the 18th International Conference on Automotive User Interfaces and Interactive Vehicular Applications (AutomotiveUI '26) September 20-23, 2026, Gothenburg, Sweden Subjects: Human-Computer Interaction (cs.HC) Cite as: arXiv:2609.13175 [cs.HC] (or arXiv:2609.13175v1 [cs.HC] for this version) https://doi.org/10.48550/arXiv.2609.13175 Focus to learn more arXiv-issued DOI via DataCite
[HC-66] Read Between the Stickers: Sentiment-Prior Reasoning with Learnable Verbalized Rules for Multimodal Chat Analysis
链接: https://arxiv.org/abs/2609.13173
作者: Zixiang Ni,Yifei Xu,Haowen Yang,Yang Liu,Ziyang Peng,Wenlong Li,Tingting Xin,Yan Liang,Yancheng Chen,Bin Chong,Yuan Rao
类目: Human-Computer Interaction (cs.HC); Multimedia (cs.MM)
备注: 14 pages,7 figures,
Abstract:Multimodal chat analysis of social media stickers (MCAS) benefits from jointly modeling text and sticker semantics, yet it is inherently challenged by the interference between sentiment and intent recognition. Although existing multi-task approaches achieve competitive performance, they largely ignore this inter-task interference and offer little explicit reasoning about how these two predictions are made. To address this issue, we propose \textbfExCoVer, an \textbfExplicit \textbfChain-\textbfof-Thought framework with \textbfVerbalized rules learning that integrates sentiment-prior reasoning with learnable discrimination rules to produce explicit reasoning chains for sentiment and intent predictions. Specifically, ExCoVer consists of two components: (1) Sentiment-Prior Chain-of-Thought (SP-CoT), which detects cross-modal sentiment conflicts and uses the dominant sentiment as a prior to mitigate inter-task interference and narrow the candidate intent space; and (2) Verbalized Rules Learning for Confusing Intent Discrimination (VRLCID), which treats discrimination rules as learnable parameters and optimizes them via learner, optimizer, and regularizer agents to suppress spurious correlations and distinguish confusing intents. Extensive experiments on CSMSA and MSAIRS datasets demonstrate that ExCoVer achieves state-of-the-art performance while providing explicit reasoning chains.
[HC-67] Still searching for an (un)stable equilibrium: visualising the process of training generative neural networks without data
链接: https://arxiv.org/abs/2609.13172
作者: Terence Broad
类目: Human-Computer Interaction (cs.HC)
备注: In Proceedings of Explainable AI for the Arts Workshop 2026 (XAIxArts 2026) arXiv:2607.20131
Abstract:(un)stable equilibrium is an ongoing series of works that is based on a practice of training generative neural networks without data. This paper introduces the second series of (un)stable equilibrium works, in which the process of training without data is visualised in a series of video pieces. These works show a generative network attempting to converge to a fixed point that is undefined, caught in an endless, unresolvable search for equilibrium. The video pieces described in this paper guide the viewer toward an understanding of AI through aesthetic experience rather than technical exposition, and strive to give a conceptual understanding of what AI could be, rather than a restatement of what it currently is. This project sits within a broader set of artistic practices that serve as an alternative and critical modes of explainability for AI.
[HC-68] A Cross Community Agenda for Speech AI
链接: https://arxiv.org/abs/2609.13168
作者: Maria Teleki,Kimi Wenzel,Anna Seo Gyeong Choi,Tobias Weinberg,Shree Harsha Bokkahalli Satish,Stephanny Sanchez,Belu Ticona,Ariadna Sanchez,Yash Sonkar,Aarti Mathur,Christoph Minixhofer,Abraham Glasser,Raja Kushalnagar,James Caverlee,Minha Lee,Shaomei Wu,Alyssa Hillary Zisk,Éva Székely,Dylan Gaines,Angelika Seeschaaf Veres,Seray Ibrahim,Nicholas Cummins,Allison Koenecke
类目: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL)
备注:
Abstract:Speech AI, any AI system that recognizes, transforms, or generates speech, is built and evaluated across two communities with only a small overlap: technical natural language processing (NLP) venues (e.g., ACL, ICASSP, Interspeech), and sociotechnical HCI venues (e.g., ASSETS, CHI, FAccT). In this position paper, we work toward a cross-community synthesis, organizing our critique around three problems: speech AI operates with an incomplete model of communication; it operates with an incomplete model of identity; and its metrics measure the wrong constructs. We draw on AAC as a setting where these failures are most visible and their stakes highest, alongside other underserved speakers - people who stutter, multilingual speakers, and non-binary and transgender users. For each problem we offer solution sketches oriented toward designing for human variability, nearly all of which require quantitative and qualitative methods in combination. We close on the venue structures that hold these methods apart, and on what program committees and individual authors can do to bring them together.
[HC-69] Code in Space: How Multimodal Human-AI Experience Can Reshape the Future of Tech Creation
链接: https://arxiv.org/abs/2609.13167
作者: Ilya Zakharov,Ekaterina Koshchenko,Agnia Sergeyuk
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Advances in artificial intelligence (AI) continue to reshape digital product development, yet the day-to-day tools for developers and designers remain bound to flat screens and 2D inputs. The intersection of AI and Extended Reality (AI-XR) introduces powerful multimodal interaction channels, such as gaze, motion, or spatial computing, that can enrich the existing Human-AI experience. A critical challenge for utilizing this multimodal opportunities lies in the understanding of how to combine these elements into a cohesive, high-level creative environment. Our study maps this territory through a thematic analysis of semi-structured interviews with 13 AI-XR experts. Categorizing over 150 topics through thematic analysis, we outline five core dimensions of this evolving landscape: professional creation, AI as a contextual layer, new interaction paradigms, adoption frictions, and ethics and human position. Our analysis reveals that besides the critical hardware constraints, the future of AI-XR for tech creation is dependent on addressing human cognitive limits. Ultimately, succeeding in the new multimodal Human-AI experience paradigm requires moving past flat-screen metaphors to design new types of interactions that selectively manage human attention while protecting user agency.
[HC-70] User-Side Contextual Phenomena in Long-Term Human-AI Interaction
链接: https://arxiv.org/abs/2609.13165
作者: Zon Rzvn
类目: Human-Computer Interaction (cs.HC)
备注: Preprint. Data (restricted access) archived at Harvard Dataverse and Zenodo. Supersedes SSRN predecessor (USCH)
Abstract:Current assessments of conversational AI focus mainly on model outputs, including hallucinations and factual errors. These measures matter, but this paper examines risks that may form on the user side during long and repeated interaction. The study follows one user across nearly four thousand conversations with the same system over twenty months. The user gradually interpreted the system as having memory, care, judgment, and authority, and reorganized part of their thinking around it. A single response may show no clear problem, while long and frequent interaction can still create another layer of risk. This paper calls that layer User-Side Contextual Phenomena (USCP) and examines records from August 2024 to April 2026. The study uses an exploratory single-case longitudinal qualitative design with autoethnographic positioning. A hybrid deductive-reflexive thematic approach organizes the material into three main modes: contextual projection, contextual attachment, and contextual authority transfer. The paper does not estimate prevalence, make diagnoses, or validate an instrument. It offers a non-clinical vocabulary and four evidence roles: inclusion, gray-zone, negative, and protective gray-zone. Its central claim is that an acceptable response on its own does not establish safety across a series of conversations. User-side risk can still form during long-term interaction.
[HC-71] Yeah I Can Play That - Aesthetic Education in Kindergarten Using a Social Robot
链接: https://arxiv.org/abs/2609.13164
作者: Sinje Eggers,Thomas Sievers
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:There have been numerous attempts to investigate the potential of social robots for education in the field of child-robot interaction (cHRI). However, empirically based models for the integration of robot technologies into aesthetic education in early childhood, especially in music education, are rare. Our approach is based on a design-based research methodology (DBR) to investigate how a social robot can help promote the creative potential of young children, act as a supportive element, and enrich musical experiences. We therefore brought the social robot Pepper to a daycare facility for children as a companion for musical activities. Preliminary findings suggest that the robot’s physical presence can influence attention, participation, and co-creative interaction. Given the limited data and the exploratory design approach, these results primarily serve to generate design hypotheses for subsequent iterations. In contrast to existing work, which predominantly addresses instructional or language-based learning processes, our approach focuses on co-creative musical interaction in the context of aesthetic education and audiation.
[HC-72] he Landscape of Misinformation Literacy Games: A Systematic Mapping Review of Game Designs and Evaluation Practices
链接: https://arxiv.org/abs/2609.13162
作者: Omed Abed,Smi Hinterreiter,Sijia Guo,Isao Echizen,Timo Spinde,Matteo Große-Kampmann
类目: Human-Computer Interaction (cs.HC)
备注: Manuscript submitted to ACM. Conditionally Accepted at Proc. ACM Hum.-Comput. Interact. 10, 7, Article GAMES068 (November 2026)
Abstract:This systematic mapping review examines 54 gamified interventions aimed at combating misinformation. We analyze the theoretical frameworks, game design, and media contexts of academic and non-academic games. The results indicate a field dominated by web-based simulations and trivia formats. While psychological inoculation is a prominent theoretical framework in the field, most interventions prioritize content-discernment tasks over proactive prebunking strategies. We identify systematic gaps in the identified games. Current interventions remain largely text-centric, focusing on news articles and text-based content, despite the growing use of audiovisual content (e.g., short videos) in the digital landscape. We conclude our study by discussing implications for addressing the evolving, multimodal nature of potentially misleading content using gamified interventions.
[HC-73] picoRing dual: ultralow-power bimanual ring controllers for everyday AR
链接: https://arxiv.org/abs/2609.13160
作者: Hideaki Yamamoto,Yifan Li,Yoshihiro Kawahara,Ryo Takahashi
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Advances in smart eyewear, such as AR glasses and HMDs, bring digital information into outdoor and public situations. However, traditional handheld controllers for hand interactions remain too bulky and socially awkward for these everyday contexts, failing to match the portability and subtlety required for mobile use. Ring-based input controllers offer a reliable finger-sensing solution; however, continuous operation with a tiny, battery-constrained ring remains challenging. Here, we present picoRing dual, an ultra-low-power bimanual interface with a pair of ring-wristbands based on inductive coupling. Unlike the original picoRing mouse limited to uni-manual tens of bps slow trackball interaction, picoRing dual uses a stable NFC-based ring-to-wristband link to capture fast index-finger motions and thumb-to-index microgestures at approximately 1.4 kbps. By tuning the ring’s coil geometry, our system achieves stable NFC against finger bending while complying with NFC’s ultra-low power profile. Therefore, picoRing dual with approximately 0.86 mW on the ring side can last approximately 100 continuous hours on a single charge of the ring’s 20 mAh battery. picoRing dual has the potential to offer ubiquitous finger input for everyday AR interaction.
[HC-74] PAUSE: A Privacy-Preserving Self-Reflection Tool for AI-Associated Cognitive Offloading DATE
链接: https://arxiv.org/abs/2609.13155
作者: Mahbub Ul Alam
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: PAUSE tool link (live): this https URL . PAUSE stands for Patterns of AI Use: Self-Examination. It is a free, anonymous, citation-grounded self-check for how AI tools may be shaping a person’s habits of thought. PAUSE is a self-reflection tool. It must not be treated as a validated psychological instrument
Abstract:Cognitive offloading is the use of external aids, such as notes, calculators, or search engines, to reduce mental effort. Large language models (LLMs) extend this to thinking itself, and by AI-associated cognitive offloading, I mean the pattern where a person routinely substitutes LLM output for their own reasoning, idea generation, learning, or communication. Recent empirical work reports associations between some patterns of LLM use and changes in critical thinking effort, neural engagement during assisted tasks, creative diversity, learning behaviour, and social dependence. Validated instruments for AI reliance, dependence, and literacy have begun to appear. I describe PAUSE (Patterns of AI Use: Self-Examination), a privacy-by-design web tool (link: this https URL) that occupies a different niche from these. It is a lightweight, non-diagnostic reflection aid for private individual use. PAUSE is organised around how a person’s own LLM use may relate to cognitive offloading across four everyday domains (‘reasoning critical thinking’, ‘creativity originality’, ‘research learning’, and ‘social communicative capacity’). It delivers a short, free, no-login self-check, scores it entirely in the browser, and returns descriptive, domain-aware reflections. The self-check pairs reverse-scored behavioural items with a claim-evaluation reasoning probe, an alternative-uses creativity probe, and a small retrospective before-and-after block. PAUSE does not assume that AI use is harmful. It only addresses where AI substitutes for effort a person may want to preserve. The application is privacy-preserving by design: scoring is deterministic and runs client-side, no personal data is required, nothing is transmitted or stored beyond the browser session, and no LLM is involved in production. PAUSE is a self-reflection tool. It is not a validated psychological instrument.
[HC-75] When Can You Trust Your Synthetic Users? Diagnostics and Corrections for LLM Consumer Panels
链接: https://arxiv.org/abs/2609.13148
作者: Robson Tigre,Hugo Gobato Souto
类目: Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注:
Abstract:Large language models are increasingly deployed as synthetic consumer panels, promising 97% cost reductions over traditional surveys. Yet aggregate validation metrics conceal systematic failures: variance compression, coefficient sign-flips, subgroup error balloons of 10–30 percentage points, and global corrections that worsen demographic bias. We provide a formal framework for deciding when to trust, correct, or abandon LLM-generated consumer data. The framework decomposes synthetic-panel bias into covariate and concept shift, develops testable diagnostics with interpretable decision thresholds, and supplies a doubly robust AIPW estimator requiring only a small calibration sample ( n = 50 - 300 ). We validate on three testbeds. In controlled simulations the decision rule achieves 100% accuracy (180/180 replications). On the American National Election Study with pre-existing LLM failures, it correctly flags heterogeneous concept shift and reduces naive bias by 92.9-99.6% . On the Twin-2K-500 consumer pricing dataset (172,884 paired human and GPT-4.1-mini responses), it correctly routes full-sample estimation to Trust and subgroup targeting to Correct, with 83-94% bias reduction.
计算机视觉
[CV-0] A Chosen Future Can Still Be Rewritten: Causal Writability in Video Models
链接: https://arxiv.org/abs/2609.15980
作者: Xingyun Wang,Haomin Zheng,Man Yuan,Leqian Yang,Ziming Liu
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 34 pages, 33 figures
Abstract:When a video model generates physically incorrect motion, did it fail to learn the correct motion, or did it learn it but fail to use it? We show the latter: the correct motion remains available inside the model and can still be made to control the generated video. We train on videos where red masses oscillate slowly and blue masses oscillate quickly, then test a red mass with fast observed motion. Even when the model generates slow motion in this conflicting case, a low-dimensional edit predicted from simple physical variables restores the correct fast motion. We call this ability causal writability. At fixed strength, we find a sharp depth boundary: the same edit changes the video before the boundary but not after it. This closure marks commitment for that write. The motion signal nevertheless remains, and a stronger downstream write can restore physical motion, while excessive gain overshoots. Early causal writability predicts which errors training later corrects: those errors are writable at more network depths than errors that persist. We reproduce both causal writability and its sharp closure in a pretrained 1.3B video model, supporting generality across model scale and training regime.
[CV-1] Anatomical Grounding and Leakage-Aware Multimodal Contrastive Learning for Alzheimers Disease Classification from Structural MRI
链接: https://arxiv.org/abs/2609.15888
作者: Paul-Gabriel Nicolae,Irina Georgiana Mocanu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 13 pages, 4 figures
Abstract:Deep networks trained on structural MRI for Alzheimer’s disease (AD) staging often reach reasonable accuracy while attending to anatomically irrelevant regions, and multimodal models that add clinical tables frequently rely on variables that were used to assign the diagnostic label in the first place. We study both issues with a deliberately lightweight slice-based encoder (ResNet18 with a one-layer Transformer over slices) on 1,075 baseline T1-weighted scans from ADNI-1. First, we use FastSurfer segmentations as an anatomical reference: YOLOv8 models trained on segmentation-derived labels localize Alzheimer-relevant structures with mAP_50 above 0.96, and a Grad-CAM comparison shows that the image-only classifier frequently attends to the skull, orbits and background. Second, we adapt a CLIP-style image - tabular contrastive framework and organize ADNIMERGE variables along a label-leakage spectrum. Fusion with cognitive scores yields 87.3% three-way accuracy, which we treat as a leakage-driven upper bound rather than an imaging result; fusion with regional volumes yields 73.0%. We observe that the choice of contrastive target changes what the image encoder learns: on MCI vs. CN, the image-only head reaches 52.4% when the encoder is aligned to cognitive scores and 73.8% when aligned to volumes, although no tabular input is used at inference. Third, restricting the input to a per-subject crop of the medial temporal lobe raises image-only three-way accuracy from 58.7% to 65.1%. All results come from single runs on a small balanced test set, and we report confidence intervals and the protocol differences that prevent direct comparison with published numbers.
[CV-2] LynnReal-Omni: Native multi-modal Video Generation for Agent ic Visual Workflows
链接: https://arxiv.org/abs/2609.15863
作者: Xiaofeng Mao,Peijia Lin,Shaohao Rui,Yibo Zhang,Haibin Wan,Weijie Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video diffusion models are stochastic and hard to control: precise content often requires repeated sampling without guaranteed success, and long-horizon scenes drift in appearance, interactions, and temporal coherence. Agentic visual creation provides explicit references, editable 3D scenes, or executable game states for stable control, but does not by itself guarantee high object or character fidelity. Combining the two can enable stable, high-quality generation. To realize this combination, we present LynnReal-Omni, a native multimodal video generation framework built on a 32B shared multimodal diffusion transformer that unifies text-to-video, image-conditioned generation, reference-guided generation, structural control, editing, degraded video restoration, and long-video generation. It accepts heterogeneous visual inputs, including appearance references, editable 3D renders, and game recordings, allowing agents to compose visual conditions within a unified model. We also train a dedicated 27B Flash shared multimodal diffusion transformer for real-time rendering. We build a systematic data pipeline for video cleaning, subject association, multimodal annotation, and aligned control construction, yielding a curated corpus of multi-shot audiovisual segments, and introduce MSAVP, a 100-prompt, 20-metric evaluation design that separates instruction following, generating plausibility, visual quality, temporal behavior, and audio coordination. LynnReal-Omni-Flash further reduces inference cost through model and decoding acceleration, including a lightweight VAE decoder; on one H100, warm generation and decoding of a 22-frame 540p video take 843 ms with LynnReal-Omni and 377 ms with Flash. These results provide a foundation for real-time streaming video generation, making LynnReal-Omni a unified, controllable, and efficient basis for agentic visual creation.
[CV-3] Proportional-Fair Resource Allocation and Dual-Threshold Early-Exit Inference for Secure Cooperative Multi-Layer Edge Intelligence
链接: https://arxiv.org/abs/2609.15847
作者: Thai T. Vu,John Le,Tu N. Nguyen,Jun Shen,Quang Vinh Duong,Ha Nguyen
类目: Networking and Internet Architecture (cs.NI); Computer Vision and Pattern Recognition (cs.CV); Emerging Technologies (cs.ET); Machine Learning (cs.LG)
备注:
Abstract:This paper proposes FREDI (Fair Resource Allocation for Edge Dual-Threshold Inference), a secure wireless edge-intelligence framework for event-triggered inference in a cooperative user equipment (UE)–edge server (ES)–cloud system. Each UE performs early-exit convolutional neural network (CNN) screening using dual confidence thresholds, while critical events are securely offloaded to an edge server for detailed classification. We formulate a proportionally-fair utility maximization problem that jointly optimizes UE–ES association, wireless and processing resources, and confidence thresholds. FREDI decomposes the problem into proportional-fair resource allocation and dual-threshold inference optimization. We prove that the detected-critical event set is set-monotone non-increasing in both thresholds, and exploit the finite empirical confidence domain for exact threshold optimization. An empirical resource–utility response envelope yields a computable global suboptimality bound and a sufficient condition for global optimality. By pre-eliminating infeasible UE–ES pairs and exactly projecting out bandwidth and transmit-power variables, the resource-allocation subproblem is reduced to a mixed-integer exponential-cone program solvable to the certified global optimality within a prescribed gap. Numerical results with early-exit MobileNetV2 and ShuffleNetV2 demonstrate near-perfect UE fairness with aggregate utility close to a Sum-Utility benchmark, reveal security-induced resource fragmentation, and demonstrate the Stage-A scalability from 6 to 144 UEs with median solving time below 0.1~s in the tested configurations.
[CV-4] RACE: Two-Stage Detector-Response Estimation With Angular Cosine Expansion for Ring Artifact Correction in Photon-Counting CT
链接: https://arxiv.org/abs/2609.15834
作者: Jigang Duan,Heran Wang,Ligen Shi,Zheng Sun,Ping Yang,Xing Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Detector response nonuniformity introduces systematic projection errors and ring artifacts in photon-counting detector computed tomography (PCD-CT). In measured PCD-CT data, residual stripe amplitudes vary slowly with projection angle, which fixed-bias models cannot adequately capture. We propose TRACE, a two-stage unsupervised sinogram decomposition method for estimating and correcting these response-related errors. TRACE represents stripes as a fixed bias plus low-order discrete cosine transform (DCT) components, using a small number of coefficients to describe angular variations at each detector element. A learnable analysis–synthesis architecture represents the ideal projections, while two-stage optimization separates them from fixed and then dynamic stripes. An angular-gradient soft orthogonality constraint suppresses correlated variations within the shared DCT gradient subspace, reducing the leakage of object structures into the artifact estimate. All parameters are optimized directly on the measured sinogram without paired training data. Experiments on measured QRM mouse phantom and porcine trotter data show that TRACE suppresses ring artifacts and improves image uniformity while preserving edge sharpness, soft-tissue texture, and trabecular detail.
[CV-5] Integrating Multi-view Multi-light Surface Reconstruction into Cultural Heritage Workflows ECCV2026
链接: https://arxiv.org/abs/2609.15833
作者: Baptiste Brument,Robin Bruneau,Benjamin Coupry,Vincent Demoulin,Jean Mélou,Antoine Laurent,Fabien Castan,Jean-Denis Durou,Lilian Calvet
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 9 figures. Accepted at VISART VIII, ECCV 2026 workshops
Abstract:Cultural heritage documentation increasingly relies on image-based 3D surface reconstruction, with photogrammetry software making such workflows accessible to archaeologists, conservators, and heritage technicians. These tools have been successful for conventional multi-view acquisition, but they do not routinely exploit richer multi-view, multi-light data, despite its potential for improving fine-scale surface reconstruction. This limitation is particularly relevant in heritage contexts, where controlled-light acquisition devices such as RTI domes are already used to capture illumination-varying image sets. The challenge is therefore to connect these existing acquisition practices with recent computer vision methods in a form that can be used within operational heritage workflows. In this work, we address this need by integrating state-of-the-art components from computer vision for multi-view, multi-light surface reconstruction into Meshroom, an open-source photogrammetry framework. Rather than proposing a new reconstruction algorithm, our contribution is to assemble and expose existing advanced methods, namely a complete photometric stereo ecosystem (calibrated, self-calibrated and universal), automatic object masking, and multi-view normal-and-reflectance integration, within a usable heritage-oriented workflow. The proposed system thus provides an intermediate software layer between computer vision research code and practical cultural heritage applications, making recent techniques easier to use and evaluate.
[CV-6] VC-Attention: Value Smoothing and Softmax Casting for Low-bit Attention
链接: https://arxiv.org/abs/2609.15810
作者: Xingyang Li,Dongyun Zou,Shining Zhang,Jiacheng Chen,Haocheng Xi,Lvmin Zhang,Jun-Yan Zhu,Song Han,Zhekai Zhang,Yujun Lin,Muyang Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Diffusion Transformers deliver state-of-the-art video generation, but their long spatiotemporal sequences make attention the dominant deployment cost, and a deployable low-bit kernel must be accurate and fast. Accuracy is limited by outliers: a block’s quantization scale is set by its largest entries, leaving typical entries confined to a narrow range of representable values. Prior work smooths queries and keys, but value outliers follow no fixed channel or spatiotemporal structure and remain the dominant source of output error. Speed is limited by softmax: low-bit Tensor Cores accelerate only the two matrix multiplications, so the high-precision exponential between them becomes the longest pipeline stage on datacenter GPUs. We propose VC-Attention, a training-free low-bit attention framework that addresses both by pairing Value smoothing with a fused probability Cast. V-Smooth reorders value tokens by lightweight online clustering, so the tokens in a hardware block quantize well together. It quantizes only the residual after subtracting the block mean, and restores that mean from the row sum the online softmax already maintains. ExpCast-FP8 maps log-domain scores directly to E4M3 probability codes with one fused multiply-add, eliminating the FP32 exponential and the format conversion. We implement VC-Attention for B200, B300, H200, RTX PRO 6000, and RTX 5090. Across Wan2.2, LongCat-Video, HunyuanVideo-1.5, and MiniMax-H3, VC-Attention improves fidelity over low-bit baselines, speeds up the attention kernel over BF16 FlashAttention-4 by 1.46-1.59x on datacenter Blackwell and Hopper and by 2.3-3.6x on workstation cards, and generates a clip 1.13-1.19x and 1.36-1.70x faster end to end.
[CV-7] SURE-Map: Self-Correcting Streaming Geometric Foundation Model
链接: https://arxiv.org/abs/2609.15795
作者: Mingkai Liu,Hao Zhao,Xingxing Zuo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Streaming geometric foundation models are emerging as a compelling alternative to SLAM systems. Yet this streaming nature introduces a fundamental issue: each prediction is made from limited context, which is vulnerable to dynamic objects and weak textures. Small local errors accumulate into severe geometric distortion and long-horizon scale drift. We argue that reliable streaming reconstruction requires geometric foundation models to be not only predictive, but also self-correcting. We introduce SURE-Map, a self-correcting framework built upon two complementary principles. First, we explicitly model cross-view geometric uncertainty. Unlike conventional depth or point confidence, which primarily reflects the reliability of individual-view prediction, our uncertainty directly measures whether the jointly predicted pose and depth induce geometrically consistent cross-view pixel correspondences. Second, because local correction alone cannot eliminate slowly accumulating scale errors, we introduce multi-timescale self-correction: fast consecutive-frame inference preserves streaming efficiency, while sparse keyframe-window inference provides longer-range geometric evidence to periodically recalibrate the scale of recent trajectories. SURE-Map establishes new state-of-the-art performance for online feed-forward reconstruction across long-horizon benchmarks, reducing ATE-RMSE from 24.00 to 17.24 m on KITTI, 5.11 to 4.74 m on Oxford Spires, and 31.37 to 28.58 m on VBR, with further improvements to 15.17, 4.63, and 22.12 m when incorporating loop-closure refinement. Project page: this https URL.
[CV-8] opoRig: Topology-Agnostic Facial Rigging via Multi-Source Supervision
链接: https://arxiv.org/abs/2609.15746
作者: Andrew Fleet,Soroush Mehraban,Vida Adeli,Cole Clifford,Babak Taati
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 6 figures. Project page: this https URL
Abstract:Automatic facial rigging across heterogeneous mesh topologies remains challenging because high-quality expression supervision is often tied to canonical templates, while deformation transfer to arbitrary meshes can introduce geometric artifacts and correspondence errors. We present TopoRig, a topology-agnostic facial rigging framework that predicts FACS-conditioned deformations directly on input mesh vertices while preserving the original topology. Starting from the ICT FaceKit expression model, we construct complementary supervision from accurate but template-biased common-topology rigs, topology-diverse but noisier transferred rigs, and targeted image-based cues for controls poorly captured by geometric transfer. TopoRig combines local surface geometry, landmark-relative semantic features, global shape context, and FACS controls to predict per-vertex displacements. We train on 3,496 generated identities using 45 non-gaze expression controls from the 53-control ICT FaceKit vocabulary. On held-out identities and unseen mesh topologies, TopoRig more faithfully reproduces the reference expression space than prior neural facial-rigging methods, while qualitative results show consistent localized deformations across diverse character geometries. Ablations demonstrate that semantic landmark features and complementary supervision improve cross-identity and cross-topology generalization. Overall, TopoRig amortizes heterogeneous and imperfect expression supervision into a single topology-preserving deformation model.
[CV-9] Bench2Dex: Benchmarking Visuo-Tactile Bimanual Dexterous Manipulation Across Dexterous Hands
链接: https://arxiv.org/abs/2609.15726
作者: Zhenjie Yang,Yideng Zhang,Dongjie Zhang,Chenyu Jiang,Xianshuai Liu,Yufeng Li,Zuhao Ge,Xingyu Jiao,Zheng Zhang,Kaiyu He,He Wang,Yuwen Zhong,Yi Deng,Muyun Jiang,Xianliang Huang,Haisheng Su,Donghang Zhang,Jian Zhang,Xue Yang,Hongyang Li,Zuxuan Wu,Yu-Gang Jiang,Xiaosong Jia,Junchi Yan
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Technical Report. Project Page: this https URL
Abstract:Tactile sensing provides contact information that can be difficult to infer from vision alone, but tactile hardware for dexterous hands has not converged to a common design. Dexterous hands differ in finger structure, contact surfaces, and sensor layouts, while simulated tactile signals still differ from measurements produced by physical sensors. These factors make it difficult to study visuo-tactile manipulation across diverse dexterous hands within a consistent experimental setting. We present Bench2Dex, a simulation benchmark for visuo-tactile bimanual manipulation across 12 dexterous hands. We adapt existing robot models with a shared simulated tactile interface that converts local contact geometry into image-like tactile observations. The interface provides a consistent observation format across different hand morphologies without attempting to reproduce the output of a specific physical tactile sensor. Bench2Dex includes 26 bimanual manipulation tasks that involve tool use, articulated-object interaction, and multi-stage manipulation, together with about 1.3K human-teleoperated demonstrations. The benchmark provides synchronized visual, tactile, proprioceptive, action, and object-state observations, together with executable task metrics. For robustness, we group seven perturbation types into invariance axis, where the correct action does not change, and equivariance axis, where the correct action changes together with the perturbation. We evaluate ACT, Diffusion Policy, pi0.5, and GR00T N1.5 on Bench2Dex and report their performance and failure modes. Bench2Dex is meant as a platform for studying visuo-tactile learning across dexterous hands. It does not assume that simulated tactile observations can replace real tactile sensing; it offers a shared setting for algorithm development while tactile hardware and simulation models are still evolving.
[CV-10] Predicting build orientation for SLM dental parts: a comparison of rotation representations and direct vector regression
链接: https://arxiv.org/abs/2609.15710
作者: Felix Schmalzel,Reimar Waitz,Moritz Kronberger,Thorsten Schöler
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Build orientation for selective laser melting (SLM) manufacturing of dental parts is usually chosen manually by technicians. We treat orientation prediction as supervised machine learning of the part’s up-axis from technician-labeled production data, and test which rotation representations produce the best results. Using n\approx2400 patient-specific dental parts, we trained a ResNet-50 multi-view image backbone and a PointNeXt-S point-cloud backbone, both pretrained and fine-tuned end-to-end, on 13 up-axis representations spanning six classical SO(3) parameterizations and seven representations defined directly on the unit sphere S^2 . We report the geodesic angular error between predicted and ground-truth up-axis on a test set, with and without test-time augmentation (TTA) over K=21 known rotations. With TTA, the octahedral map achieves the lowest mean angular error ( 10.6^\circ , ResNet-50). The three lowest-error results overall are direct S^2 representations, though this may reflect label noise in the unsupervised in-plane component of the SO(3) targets rather than a topological advantage. von Mises-Fisher collapses to a near-constant prediction when trained with PointNeXt-S but not with ResNet-50. TTA reduces mean angular error by 31-73 % across almost every representation and backbone. Overall, test-time augmentation over a small set of known rotations is the most consistent driver of accuracy, whereas the best-performing representation is strongly backbone-dependent.
[CV-11] V-ICAL Bench: Evaluating Video In-Context Learning for Multimodal Agents in Interactive Environments
链接: https://arxiv.org/abs/2609.15683
作者: Ziqian Fan,Shibo Xu,Junjie Li,Xiangyu Zhao,Shengyuan Ding,Yifan Yang,Zhenjie Yang,Haodong Duan,Yue Zhou,Zhihang Zhong,Xue Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:While In-Context Learning (ICL) enables models to adapt from exemplars without parameter updates, multimodal ICL remains largely underexplored, particularly regarding video demonstrations in interactive environments. For multimodal agents, learning from videos presents unique challenges: they must translate in-context demonstrations into executable policies, ground these policies in novel visual states, and iteratively refine actions based on environmental feedback. We introduce V-ICAL, a novel benchmark designed to evaluate video-based ICL for multimodal agents. Comprising 342 interactive tasks across 37 environments, V-ICAL utilizes human-curated demonstration videos as task-specific behavioral exemplars, evaluating agents through sustained interaction from a target initialization. The benchmark seamlessly connects in-context knowledge induction with core agentic capabilities, including state grounding, temporal memory, planning, and adaptation in dynamic environments. Extensive evaluations across 19 state-of-the-art multimodal agents reveal significant limitations: the best-performing model, Seed-2.1-Pro, achieves a score of only 54.4/100, while other leading models (e.g., Gemini-3.1-Pro, GPT-5.6) fail to surpass 50, far below the human baseline of 83.6. Controlled comparisons further demonstrate that current agents struggle to reliably translate video exemplars into effective policies, failing to yield consistent performance gains. Ultimately, V-ICAL exposes a critical gap in the ICL capabilities of multimodal agents, underscoring an urgent need for future research.
[CV-12] MambaMPD: A Mamba-Driven Segmentation Framework for Marine Pollution Detection from Remote Sensing Imagery
链接: https://arxiv.org/abs/2609.15676
作者: Shuaiyu Chen,Wei Han,Peng Ren,Chunbo Luo,Zeyu Fu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Accurate marine pollution detection (MPD) is essential for protecting coastal ecosystems and marine biodiversity. Vision Mamba models have shown promise in remote-sensing semantic segmentation by efficiently capturing long-range dependencies and global context, yet their potential for MPD remains underexplored. MPD is particularly challenging because of low signal-to-noise ratios, fragmented pollution patterns, and indistinct boundaries caused by the visual similarity between pollutants and the surrounding sea. To address these issues, we propose MambaMPD, an enhanced Mamba-based framework incorporating two complementary structural priors: Frequency-Aware Augmentation (FAA) and multi-scale Edge-Guided Attention (EGA). FAA integrates wavelet transforms into the encoder to decompose features into multi-scale frequency subbands, enabling the model to capture low-frequency contextual semantics and high-frequency structural details needed to identify small, low-contrast, and irregular pollution patterns. EGA adaptively fuses hierarchical, Laplacian-derived boundary cues with deep semantic representations, refining encoder features before decoding to sharpen boundaries and reduce ambiguity in visually confusing, spatially fragmented scenes. Together, these modules improve sensitivity to subtle pollution signals while preserving fine boundary structures. A U-Net-style decoder with squeeze-and-excitation attention and deep supervision progressively restores and refines semantic and spatial information across scales. Extensive experiments on two benchmark MPD datasets show that MambaMPD achieves higher mIoU than competing methods while requiring substantially less computation than foundation-model-based approaches. On MADOS, it improves F1 by 3.6% over OSDMamba; on M4D, it raises Oil Spill IoU by 6.82% over TransOilSeg.
[CV-13] Dont Send What You Dont Need: Question-Guided Token Pruning as a Privacy Defense for Vision-Language Models
链接: https://arxiv.org/abs/2609.15671
作者: Md Khalid Syfullah,Alvi Ataur Khalil
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Visual Question Answering (VQA) with Vision-Language Models (VLMs) is increasingly used in privacy-sensitive and bandwidth-constrained settings. Federated Learning (FL), Split Learning (SL), and U-Shaped Split Learning (USL) keep raw data local, but transmitting all visual tokens across a model partition remains costly and can expose private information. We propose QPriv-VL, a question-guided, privacy-aware token-pruning framework for FL, SL, and USL that prunes visual tokens before transmission based on task utility and privacy sensitivity. Its core component is a lightweight Dynamic Threshold Predictor (DTP) that jointly estimates a sample-specific pruning ratio and a token-level retention mask in one forward pass. DTP combines question relevance, computed from cross-modal similarity between visual patches and the pooled question embedding, with a sensitivity signal derived from frozen DINOv2 features. This allows the model to suppress potentially sensitive regions while preserving patches useful for answering the question, without requiring sensitivity labels. We evaluate QPriv-VL on GQA, OK-VQA, VQAv2, SLAKE, VQA-RAD, and PathVQA against four privacy attack families: FSHA, FORA, iDLG, and attribute-inference membership inference attacks. DTP matches or outperforms fixed-ratio pruning while using substantially fewer transmitted tokens. On VQA-RAD, it reduces membership-inference attack success from 0.99 to 0.76-0.79, lowers FSHA and FORA reconstruction PSNR relative to fixed-ratio pruning, and preserves competitive VQA accuracy using about 40% of the original visual-token budget. A sensitivity exclusion ratio of 1.20 +/- 0.18 indicates preferential removal of privacy-sensitive patches, while explainability analysis shows that retention adapts to question semantics rather than generic visual saliency.
[CV-14] Benchmarking Intra-Patient 3D Deformable Multimodal Image Registration
链接: https://arxiv.org/abs/2609.15669
作者: Matteo Barbieri,Giammarco La Barbera,Juan Pablo De La Plata,Sabine Sarnacki,Isabelle Bloch,Pietro Gori
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Preprint Under Review
Abstract:Multimodal image registration is a key component of many clinical workflows, yet it remains challenging because corresponding anatomical structures often exhibit substantially different image intensities across modalities. In this work, we present a comprehensive benchmark of intra-patient 3D multimodal deformable registration methods across three datasets covering different anatomical regions and difficulty levels, including both synthetic deformation recovery and real clinical scenarios. We evaluate classical optimization-based approaches and modern learning-based methods, including recent deep learning and foundation models, using complementary metrics: Average Dice similarity coefficient (DSC), average 95th-percentile Hausdorff distance (HD95), and a modality-independent structural similarity measure based on the MIND self-similarity context (MIND-SSC). Results show high variability across datasets, with learning-based methods demonstrating superior performance on large synthetic benchmarks, while only limited improvements are observed in real pelvic registration. A key finding of this study is the consistent disagreement between geometric metrics (DSC, HD95) and image-based similarity metrics (MIND-SSC), highlighting that improved overlap does not necessarily imply better global multimodal correspondence. Furthermore, anatomy-guided approaches achieve the highest overlap scores but exhibit degraded performance outside of segmented regions, revealing a trade-off between label-driven alignment and global structural coherence. Overall, our results indicate that no current method achieves robust performance across anatomies and modalities. We demonstrate that intra-patient 3D multimodal registration requires multi-criteria evaluation, including deformation-based metrics, and remains an open problem.
[CV-15] Circuit-MLLM : Topological Logic-Guided Latent-Space Visual Reasoning for Circuit Schematic Understanding ECCV
链接: https://arxiv.org/abs/2609.15668
作者: Jinyuan Deng,Yuqi Jiang,Wenjing Huang,Xin Li,Qi Sun,Cheng Zhuo
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at European Conference on Computer Vision (ECCV) 2026
Abstract:Through pre-training on extensive text and image datasets, current multi-modal large language models (MLLMs) achieve strong performance on general tasks. However, circuit schematics present a unique challenge for MLLMs due to their dense component layouts and distinct topological logic, demanding fine-grained structural parsing to extract the electrical semantics. To address this, we propose Circuit-MLLM, a multimodal reasoning framework that reformulates circuit topology analysis as a process of device localization, path tracing, and sequential reasoning within the latent space. We introduce a circuit knowledge mining mechanism that deeply aligns the model’s latent representations with structurally rich features derived from multi-granularity circuit vision experts, enabling the model to effectively internalize topological semantics. Building upon these internalized semantics, we devise a topology-guided sequencing strategy that decouples reasoning from the rigid raster-scan order, enforcing stepwise inference along the circuit’s topological logic in latent space. Across diverse circuit analysis tasks, Circuit-MLLM consistently outperforms strong baselines, notably achieving a 25% higher average score than GPT-5.1, which demonstrates the effectiveness of our framework in circuit schematic topology analysis. Code is publicly available at this https URL.
[CV-16] Kaininja: Extending Native 3D Generators to the Part Level
链接: https://arxiv.org/abs/2609.15659
作者: Ruihan Yu,Lian Fu,Muyao Niu,Zheng-hui Huang,Yu-Ju Tsai,Sho Kuno,Fengbo Lan,Yonghao Yu,Erwin Wu,Ming-Hsuan Yang,Kaipeng Zhang,Zhixiang Wang
类目: Graphics (cs.GR); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Native 3D generators turn one image into a single mesh. TRELLIS.2 and its peers deliver high-fidelity non-watertight geometry with materials, but the output is one fused object, while downstream work such as editing, rigging and simulation operates on part-level assets. A naive idea is to run a 3D segmentation network on the fused mesh that TRELLIS.2 generates, but such pipelines are slow and bounded by the accuracy of the segmentation. We want a simple way to extend an existing native 3D generator to the part level. But we face a critical problem: the O-Voxel grid stores one sheet of surface per voxel, so a single volume cannot represent the interface where two parts touch, at any resolution. We introduce a dual-volume representation to solve this problem and put forward KaiNinja, a part-level extension of TRELLIS.2 built on a dual-volume form of its O-Voxel representation. KaiNinja keeps the generation speed and quality of TRELLIS.2 while extending it to the part level, with no mask or segmenter in the pipeline. Its training data come from sources of many kinds, including CAD models and assets authored by an LLM-driven agent; to our knowledge it is the first 3D generative model trained on agent-authored part data. Surprisingly, we also find that whole-object fidelity improves over the same backbone fine-tuned on the same dataset. Against part generation pipelines of different paradigms, it lowers whole-object Chamfer distance by 40% and raises strict part F-score by 16%.
[CV-17] From Model Patterns to Abstract Semantics in Compositional Zero-Shot Learning ICME2026
链接: https://arxiv.org/abs/2609.15649
作者: Weize Li,Zhicheng Zhao,Fei Su
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ICME 2026
Abstract:Compositional Zero Shot Learning aims to recognize unseen compositions by recombining learned primitives. Recent methods rely on vision language models and attempt to explicitly model contextual variations of primitives through multiple representations. However, such approaches are limited by fixed variant capacity and competition between abstract and concrete semantics. In this work, we present a new perspective that views primitive variations as the context-driven activation of concrete visual cues rather than independent entities. Based on it, we propose CLEAR, a CLoze-style rEAsoning-based Re-ranking framework inspired by human perceptual processes. CLEAR extracts conditional variants from the primitive candidate set in a coarse-to-fine manner, performs cloze-style reasoning to infer high-level semantics, and re-ranks predictions to correct biases toward salient concrete primitives. Extensive experiments demonstrate that CLEAR consistently improves the Base Model and outperforms state-of-the-art methods on the challenging C-GQA and MIT-States datasets. Code is available at this https URL.
[CV-18] MedDiME: Efficient Latent Diffusion with Adaptive Masking for Medical Counterfactual Generation
链接: https://arxiv.org/abs/2609.15647
作者: Yan Zeng,Changlu Guo,Anders Nymark Christensen,Morten Rieger Hannemose,Anders Bjorholm Dahl
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 4 figures, 4 tables
Abstract:Medical counterfactual generation modifies images to change model predictions for interpretability. However, existing diffusion-based approaches are often prohibitively slow and memory-intensive, making them difficult to apply in high-resolution settings. Moreover, existing masking strategies are tightly coupled with pixel-space representations, making them incompatible with latent-space diffusion editing. To address these challenges, we propose MedDiME, a latent-space classifier-guided diffusion framework that reduces computational and memory overhead while introducing a latent-compatible, gradient-driven adaptive masking mechanism for spatially precise medical counterfactual generation. Extensive experiments demonstrate that MedDiME achieves high-quality counterfactual generation with significant efficiency gains compared to prior classifier-guided diffusion baselines, achieving up to 40 times faster inference and 13 times lower peak GPU memory usage.
[CV-19] SAM3D-Part: Interactive Part Selection and Generation from 3D Objects
链接: https://arxiv.org/abs/2609.15639
作者: Jiahao Chang,Dong Du,Wanhu Sun,Yujian Zheng,Chuanyu Pan,Bowen Zhao,Chongjie Ye,Yuanming Hu,Xiaoguang Han
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Part-level control is essential for modern 3D asset creation, where objects are frequently edited, reused, animated, or fabricated through their individual components. In many such workflows, users need only several specific components rather than a complete object decomposition. However, existing 3D generation methods produce all parts regardless of user intent, while promptable 3D segmentation methods typically output partial surfaces instead of reusable complete meshes. In addition, image-conditioned part generators further struggle to preserve hidden geometry and accurate placement without directly conditioning on the source mesh. To address these problems, we present SAM3D-Part, a prompt-driven framework for selective part generation from input 3D object meshes. Given a source mesh and a part prompt, SAM3D-Part first encodes the source geometry into compact mesh features and aligns them with the rendered image, selective mask, and point-map observations via pixel-wise channel fusion. The fused representation conditions a feed-forward generative model to produce only the queried component as a completed mesh. To place the generated part back into the source coordinate frame, SAM3D-Part predicts dense per-voxel correspondences and estimates the part transformation from distributed spatial evidence rather than a single global pose code. For sequential multi-part queries, previously generated parts are stored in a part cache and reused as contextual constraints, reducing conflicts among independently requested components. Extensive experiments and ablations demonstrate that SAM3D-Part can significantly improve source alignment, reduce conditioning cost, and enable consistent selective part generation, achieving state-of-the-art. Code and weights will be available at this https URL.
[CV-20] ModaLens: Measuring Image Sensitivity in Report-Conditioned Medical VLMs
链接: https://arxiv.org/abs/2609.15635
作者: Sebastián Andrés Cajas Ordóñez,Maximin Lange,Quang Bui,Anqi Peter Li,Felipe Ocampo Osorio,Rafi Al Attrach,Kushul Reddy Palakala,Sahil Kapadia,Zakaria Laouabdia Sellami,Xinyue Zhang,Ashley Zhang,Leo Anthony Celi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:A radiology report can already answer a clinical question, so it is hard to tell whether a vision-language model also uses the image. ModaLens, a paired image-swap audit, measures how report availability changes image sensitivity: MedGemma-27B on 3,199 paired MIMIC-CXR cases from 293 patients, all 14 questions per case (13 finding-specific and one composite), each image replaced by one from another study, usually of the same patient, with question and report fixed. Under an explicit answer instruction, the model’s generated answer changes on 4.26 percent of trials with the report and 20.94 percent without it, a paired increase of 16.7 points (patient-clustered 95 percent CI 15.6 to 17.7), so report availability reduces image-swap sensitivity under this protocol; the original prompt with a lowercase first-token readout gives 4.70 percent against 17.07 percent, and substitutions also move continuous answer scores where the binary prediction does not change. The labels are derived from reports, which limits conclusions about visual correctness; the direction replicates in two further model lineages. Code, the exact prompts and a run record for every number are at this https URL.
[CV-21] VideoScout: Learning Agent ic Active Exploration with Adaptive Reasoning Pacing for Long Video Understanding
链接: https://arxiv.org/abs/2609.15606
作者: Weixin Xu,Zhenyu Yang,Bing Wang,Shengsheng Qian,Changsheng Xu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Multimodal Large Language Models (MLLMs) have achieved remarkable progress on short video understanding yet remain limited on long videos due to the limited visual context window. Prevailing approaches rely on uniform frame sampling or recent coarse-to-fine agentic zooming, both of which struggle to localize sparse, decisive evidence in sufficiently long videos. We formulate long video understanding as a \textbfSequential Evidence Acquisition (SEA) problem, in which an agent reads the video turn by turn along the temporal axis, deciding at each turn how fast to watch, what evidence to retain, when to revisit uncertain segments, and when to stop and answer. Inspired by this view, we propose \textbfVideoScout, a multi-turn reasoning agent that instantiates the SEA paradigm through adaptive reasoning pacing. Specifically, by dynamically controlling the viewing pace, VideoScout enables efficient traversal of long videos within a bounded visual context window, allowing the agent to access more video content while balancing content analysis depth with reading efficiency. To train VideoScout, we construct VideoScout-66K, a set of over 66K high-quality exploration turns from 10K answer-verified trajectories, and adopt a two-stage pipeline: cold-start supervised fine-tuning teaches the agent per-turn output format, while the Decoupled Clip and Dynamic sAmpling Policy Optimization (DAPO) algorithm performs trajectory-level reinforcement learning with a composite reward that jointly considers answer accuracy, output format compliance, and the temporal alignment between the agent’s viewing progress and the teacher’s answer timing measured by intersection-over-union (IoU). Extensive experiments on long video understanding and reasoning benchmarks demonstrate that our 7B model achieves strong performance compared with existing trained 7B agentic models.
[CV-22] A Unified Vision-Language Model for PSMA PET/CT Report Generation Visual Question Answering and Lesion Segmentation
链接: https://arxiv.org/abs/2609.15603
作者: Yang Xing,Jiong Wu,Savas Ozdemir,Yang Zhou,Boxiao Yu,Ying Zhang,Zheren Zhu,Chenyu You,Wei Shao,Yang Lu,Kang Wang,Tinsu Pan,Yang Yang,Kuang Gong
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Accurate PSMA PET/CT interpretation is central to prostate cancer management, yet existing PET/CT AI models typically address isolated tasks. We propose a unified PSMA PET/CT vision-language model for report generation, visual question answering, and lesion segmentation. The framework adopts an LLaVA-style architecture, comprising a PET/CT vision encoder, an MLP-Mixer projection module, a LoRA-tuned large language model, and a 3D segmentation branch. Training followed a four-stage strategy: vision encoder pretraining, projection-layer alignment, VLM fine-tuning, and final multitask tuning. Language tasks used 5,747 PSMA PET/CT datasets with paired reports, while segmentation used the PSMA subset of AutoPET. The model outperformed PET2REP and a CT-based baseline across standard report-generation metrics, improved performance across VQA question types, and achieved higher Dice and lesion-level overlap F1 than SegAnyPET and nnUNet. These results support the feasibility of a unified framework for structured, interactive, interpretable PSMA PET/CT analysis with voxel-level grounding within a single multitask model architecture.
[CV-23] What a gated sensing pipeline never looks at: bandwidth reduction and the misses behind it
链接: https://arxiv.org/abs/2609.15579
作者: Raghu Venkat,Tricha Anjali
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages, 6 tables, no figures. Reports a superseded detection result alongside the corrected run that replaces it; Table 6 carries every scored run of every staged flight
Abstract:An airborne sensor on a contested link cannot send video, so the appealing move is to send findings instead and report the ratio between the two. We evaluate a gated sensing pipeline that does this, combining learned object detection and image-text comparison with deterministic scheduling, gating, evidence accumulation and transmission rules. On staged footage with the semantic stage live it sends 38,736 bits over 211 s, a reduction of 41,977x, and names 1 of 4 staged events with no false report. That detection has since been superseded: a correction to how the tracker measures speed removed the measurement artefact the normality model had been learning from, and the flight no longer warms. Under the corrected code the pipeline names 1 of 7 staged events across four flights, and we report both. On a control flight where nothing was staged it reports nothing, a reduction of 155,830x: the largest number in the study and the least informative, because a reduction ratio measures the scene. The mechanisms that produce the reduction also decide which observations ever reach a decision, so the two cannot be reported apart. We give a tick-level trace of one flight (3,187 rows) that places each of three missed events at the stage where it stopped progressing: one produced no track, one failed the structural place test at 0.129, and one passed 539 structural ticks but reached only 3.192 against a boundary of 3.807. We also report one instance of a known failure mode, an online normality model absorbing the object it will later judge, measured against the threshold that object then failed. The evidence is one detection and six misses across four staged flights, beside one clean control, and we treat it as a case study. We give the reproduction protocol and generated results, identify which supporting artifacts are not distributed, and state which experiments did not run.
[CV-24] PIVOT: Physics-Grounded Verification for AI-Generated Audio-Video Detection
链接: https://arxiv.org/abs/2609.15562
作者: Bo Zheng,Kangran Zhao,Xiaoyu Zhang,Weinan Guan,Zhiheng Li,Yize Chen,Haizhou Li,Qingshan Liu,Siwei Lyu,Baoyuan Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Multimedia (cs.MM)
备注: 19 pages, 4 figures, including appendix
Abstract:As generative models continue to advance, AI-generated content (AIGC) is becoming increasingly realistic, weakening the artifact cues commonly exploited by existing detectors. Nevertheless, faithfully reproducing the physical behavior of real-world events remains challenging for current generators. We therefore explore detecting AIGC by assessing whether the depicted event satisfies measurable constraints derived from physical laws. We introduce PIVOT, a physics-grounded AIGC detector, instantiated here for audio-video clips, that estimates physical quantities from video and audio, selects physical laws relevant to each clip, and verifies their measurable constraints. Beyond a real/fake decision, PIVOT returns supporting evidence that records the verification outcome, relevant time window, and supporting quantities for each applicable law. Although instantiated and evaluated here on audio-video data, the framework can, in principle, extend to other AIGC modalities whenever the physical quantities required for verification can be estimated reliably. We also introduce PhysForensics-Bench, comprising paired real and generated audio-video clips from nine event-centric scene families and two recent audio-video generators. On PhysForensics-Bench, PIVOT achieves 70.30% accuracy and 64.29% F1 score on Real+Seedance, and 72.16% accuracy and 65.82% F1 on Real+VEO. In comparison, direct inspection with Gemini 3.1 Pro obtains 53.96% accuracy and 60.09% F1 on Real+Seedance, and 57.22% accuracy and 63.44% F1 on Real+Veo. These results demonstrate the practical promise of physical-consistency verification as a structured and inspectable source of evidence that complements artifact-based AIGC detection.
[CV-25] Multi-Stage NeRF for Efficient 3D Coronary Artery Reconstruction from Two Narrow-Angle Angiographic Projections
链接: https://arxiv.org/abs/2609.15550
作者: Deyu Meng,Mojtaba Lashgari,Yiying Wang,Abhirup Banerjee
类目: Computational Geometry (cs.CG); Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages, 2 figures
Abstract:X-ray coronary angiography is the clinical gold standard for coronary artery disease during real-time cardiac interventions, but provides only 2D projections of inherently 3D vessels. Existing learning-based 2D-to-3D reconstruction methods typically require wide angular coverage or multiple views, assumptions that are rarely satisfied in routine practice where only two projections with narrow angular separation are available. To address these challenges, we propose NeCA++, a multi-stage self-supervised neural radiance field (NeRF) framework tailored to clinically realistic acquisition constraints. The framework decomposes reconstruction into two stages that progressively refine spatial support and representation capacity. In the first stage, a coarse 3D representation of the vasculature is reconstructed, restricting the subsequent optimisation to regions with a higher likelihood of vessel presence, termed an active region. Afterward reconstruction is restricted to this region while higher-resolution representations are progressively activated to recover fine vascular details. This multi-stage strategy focuses learning on anatomically plausible regions, mitigates gradient dilution under extreme sparsity, and stabilises global topology before recovering fine vascular branches. Furthermore, two vessel-specific regularisations are introduced: a ray-aligned constraint to reduce projection-induced ambiguity, and a bimodal density penalty to enable early vessel-background separation. Extensive experiments across three datasets (ImageCAS, ASOCA, and Synthetic RCA) and four angular configurations demonstrate consistent superiority over state-of-the-art baselines, particularly under clinically realistic narrow-angle settings, while achieving reconstruction within 58 seconds per case.
[CV-26] Assessing nnU-Net Generalization across Brain Tumor Populations in BraTS-GoAT 2026
链接: https://arxiv.org/abs/2609.15524
作者: Tristan Kirscher(ICube, Institut Strauss),Vivian Metzger(Institut Strauss),Philippe Meyer(Institut Strauss, ICube),Xavier Coubez(Institut Strauss, ICube)
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注:
Abstract:BraTS-GoAT evaluates tumor segmentation across heterogeneous populations. We trained a conventional 3D nnU-Net on 1,351 labeled cases using five-fold cross-validation and 1,000 epochs per fold. The final predictor averaged all folds and applied test-time mirroring. On pooled official validation, global DSC values were 0.7805, 0.8288, and 0.8854 for enhancing tumor (ET), tumor core (TC), and whole tumor (WT). Under matched fold-0 inference, mean regional Dice decreased from 0.9058 on source out-of-fold (OOF) cases to 0.8310 on pooled validation (difference–0.0747). Mirroring gave small single-fold gains but no clear ensemble benefit; a residual-encoder alternative reached 0.8282 mean Dice. In labeled OOF predictions, failure cases had substantially smaller reference ET volumes; after adjustment for ET and WT volume, lower Dice remained associated with more disconnected ET components and a smaller fraction of ET contained in the largest component.
[CV-27] SyntheticDoc: A Large Synthetic Dataset for Document Unwarping and Illumination Correction ECCV2026
链接: https://arxiv.org/abs/2609.15503
作者: Daniel Woortmann,Tanguy Magne,Olga Sorkine-Hornung
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: D. Woortmann and T. Magne – Equal contribution. Accepted at ECCV 2026 (Spotlight). 20 pages
Abstract:Deep learning models have become the standard tool for document rectification and illumination correction, yet their performance is fundamentally bound by their training data. For nearly a decade, the community has heavily relied on Doc3D, a pioneering but increasingly limited document unwarping dataset in terms of scale and quality. To address this bottleneck, we introduce SyntheticDoc, a massive, high-quality dataset designed to push the boundaries of document unwarping. SyntheticDoc is composed of 1,000,000 high-resolution procedurally generated training samples, alongside extensive validation and test sets. Each sample is paired with rich, pixel-perfect annotations, including UV maps, normal maps, albedo and shading. To ensure physical accuracy and photorealism, the paper geometries are generated via a physics-based simulator and rendered using a path tracer. To demonstrate the benefit of our dataset, we train a simple baseline model on SyntheticDoc and report on its performance in comparison to state-of-the-art methods on both document unwarping and illumination correction tasks. Our dataset is available at this https URL and the code used to generate it at this https URL .
[CV-28] CatchMonitor: a machine learning system for automated fish discard quantification
链接: https://arxiv.org/abs/2609.15484
作者: Geoff French,Michal Mackiewicz,Mark Fisher,Helen Holah,Rebecca Lamb
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 30 pages, 5 figures
Abstract:We report on the continued development of CatchMonitor, resulting in a prototype computer vision system designed to automatically quantify discarded fish from video footage collected from Remote Electronic Monitoring (REM) systems on fishing trawlers. The analysis of trawler surveillance footage is a challenging problem due to the real-world conditions on board fishing vessels. Building on our prior work we improve the accuracy of species identification through the application of semi-supervised learning. We utilise a simple and robust object tracking approach, upon which we build our prototype discard quantification system. Finally we analyse the variability of manual discard quantification performed by multiple expert human analysts, using it as a benchmark against which we compare the performance of our system.
[CV-29] BVB: Benchmarking Agent ic Video Understanding via Programmatic Reconstruction in Blender
链接: https://arxiv.org/abs/2609.15478
作者: Yolo Y. Tang,Daiki Shimada,Jiayue Meng,Jing Bi,Pinxin Liu,Yicheng Wang,Yunzhong Xiao,Zhangyun Tan,Zeliang Zhang,Chao Huang,Susan Liang,Qianxiang Shen,Luchuan Song,Ali Vosoughi,Mingqian Feng,Melika Filvantorkaman,Chenliang Xu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal agents can create complex videos in software such as Blender by coding without relying on diffusion models. Yet video understanding benchmarks still evaluate models mainly through question answering. If an agent truly understands a video, it can reconstruct it programmatically. We introduce BVB, Blender-VideoBench, a benchmark that tests this ability by asking agents to reconstruct real-world videos as animated Blender scenes. To ensure fair comparison, each agent programs the reconstruction through a lightweight harness, Mini-BVB, in an identical sandbox under a shared cost limit. The benchmark renders each reconstruction from its animated camera and evaluates it on two axes: (1) Dual VQA measures how many spatiotemporal facts the reconstruction preserves. (2) Latent Similarity measures how closely the reconstruction matches the source video perceptually. Our overall score, a square-root mean, favors balanced performance. We evaluate 51 configurations from 10 model families and analyze semantic retention, perceptual similarity, reasoning effort, and cost. The best model reaches 88.6 Latent Similarity but retains only 53.7% of the source-correct spatiotemporal answers. Additional reasoning improves visual similarity but does not close this gap in factual accuracy. In a blind study with 15 raters and five configurations, Latent Similarity correlates strongly with human preference. These results show that programmatic reconstruction is a viable test of agentic video understanding, and that semantic retention remains the main challenge.
[CV-30] AnchorGUI: Asymmetric Memory for Dual-Scale Learning in GUI Navigation ECCV2026
链接: https://arxiv.org/abs/2609.15457
作者: Shengjie Jin,Zelong Sun,Hengbo Xu,Yanbiao Ma,Zhiwu Lu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Published in ECCV 2026. 41 pages including supplementary material
Abstract:Vision-Language Models (VLMs) enable autonomous GUI navigation, but agents still struggle to process and learn from dense, continuous visual histories. This bottleneck hinders both immediate error correction within a single episode (intra-trial) and experience distillation across multiple attempts (cross-trial). We trace these challenges to an empirical informational asymmetry in GUI navigation: while expected transitions can often be compressed into lightweight textual summaries, unexpected outcomes benefit from preserved screenshots as causal evidence for accurate diagnosis. Building on this insight, we propose AnchorGUI, a unified framework driven by the Cognitive State Anchor (CSA). The CSA acts as a per-step primitive that actively compares expected and observed transitions, converting passive multimodal trajectories into explicit prediction-error signals. These signals orchestrate a dual-scale learning mechanism via an asymmetric memory. For intra-trial correction, a sliding window selectively retains visual evidence for detected mismatches, providing immediate, visually-grounded feedback. For cross-trial distillation, this asymmetric memory focuses the computationally expensive credit assignment search space on likely failure steps. Experiments across four benchmarks validate the effectiveness of our approach. On AndroidWorld, AnchorGUI achieves a 57.3% success rate with a 2.4\times token reduction per step. Furthermore, cross-trial distillation reaches 69.2% success (+11.9% gain), significantly outperforming standard reflection methods while maintaining sub-linear context scaling.
[CV-31] A Conservative OCR-Enabled Workflow for R214 Sodium Screening of South African Packaged Foods
链接: https://arxiv.org/abs/2609.15427
作者: Mayimunah Nagayi,Alice Scaria Khan,Tamryn Frank,Rina Swart,Clement Nyirenda
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted for presentation at the 10th IEEE ICTAS 2026; to appear in the conference proceedings. 7 pages, 1 figure, 3 tables
Abstract:Using food package images to monitor sodium and salt content against South Africa’s R214 sodium limits is challenging when screening decisions require product identity, nutrition facts panel evidence, reporting basis, and category-specific thresholds. This study presents a conservative image-based workflow that combines region detection, optical character recognition (OCR), product identity and sodium evidence extraction, R214 category assignment, deterministic threshold comparison, and independent vision language model comparison. The evaluation used 442 packaged food products and 3 929 full package images from a real-world South African food packaging dataset. A YOLO26s small detector generated 4 195 region crops, and strict post-processing produced one sodium evidence row per product. The integrated workflow produced 290 OUTSIDE R214 SCOPE, 139 REVIEW, seven SCREEN-PASS, and six SCREEN-FAIL outcomes. The independent Qwen2.5-VL 7B vision language model workflow produced 387 OUTSIDE R214 SCOPE, 31 REVIEW, twenty SCREEN-PASS, and four SCREEN-FAIL outcomes. The workflows agreed on exact R214 category assignment for 415 of 442 products (93.9%) and on whether the assigned category was within R214 scope for 416 of 442 products (94.1%). Final screening outcome agreement was 307 out of 442 products, or 69.5%. Manual verification on 60 products showed lower strict outcome agreement than regulated status agreement, while all manual INSUFFICIENT DATA cases were kept out of SCREEN-PASS and SCREEN-FAIL by both automated workflows. The findings show that conservative image-based screening can organise package evidence, identify clear cases, and assign uncertain cases to REVIEW rather than forcing SCREEN-PASS or SCREEN-FAIL decisions.
[CV-32] ViCo-SAM3: Vision-Conditioned Alignment for Open-Vocabulary Camouflaged Object Segmentation
链接: https://arxiv.org/abs/2609.15418
作者: Qiangqiang Zhou,Wenjun Tang,Yong Chen,Dandan Zhu,Jiawei Xu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Open-vocabulary camouflaged object segmentation (OVCOS) aims to segment unseen camouflaged objects under text guidance. We observe that SAM3 still suffers from a pronounced semantic gap between global textual semantics and fine-grained pixel-level visual cues in OVCOS. Meanwhile, fully fine-tuning the text encoder introduces heavy parameter overhead and risks overfitting to training categories, which compromises open-vocabulary representation flexibility. To address these issues, we propose ViCo-SAM3, a Vision-Conditioned alignment framework designed for OVCOS. Specifically, we introduce vision-conditioned (ViCo) module, which dynamically modulates text embeddings with global visual context, enabling textual representations to adapt to the current image content and thereby effectively bridging the semantic gap between vision and text. Building on this, we further design a vision-conditioned cross-modal binding (ViCoBind) module to enhance cross-modal interaction and semantic alignment between visual and textual representations. Without bells and whistles, ViCo-SAM3 achieves state-of-the-art performance on the OVCamo benchmark and demonstrates strong generalization.
[CV-33] BSC-Net: A Small-Branch-Sensitive Structural Continuity Network for Coronary Vessel Segmentation and Quantitative Angiographic Analysis
链接: https://arxiv.org/abs/2609.15400
作者: Wanxian Li,Jiaqian Qin,Qingyi Xian,Yazhi Li,Song Chen,Liman Li,Hao He
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 29 pages, 9 figures, including Supplementary Material
Abstract:Vessel segmentation in X-ray coronary angiography (XCA) is a fundamental step for quantitative coronary analysis and subsequent assessment of coronary artery disease. However, accurate vessel segmentation remains challenging because of imaging noise, complex bifurcations, and the overlap of vessels and background structures, which can lead to disrupted vascular connectivity and missed small branches. In this work, we propose BSC-Net, a ResNet-U-Net-based framework tailored to improve small-vessel representation and repair vascular structural continuity. BSC-Net enhances small-vessel representation through targeted sampling and improves vascular structural continuity by integrating long-range contextual modeling and Edge-Informed Loss (EIL). BSC-Net was validated on two public XCA datasets, demonstrating state-of-the-art (SOTA) performance in coronary vessel segmentation with Dice and IoU scores of 77.8%/90.6% and 64.5%/83.0%, respectively. Furthermore, based on the obtained vessel segmentation, we performed automated quantitative coronary analysis and derived clinically relevant morphological and hemodynamic parameters, including stenosis ratio, time-to-peak, and relative propagation velocity. These results demonstrate that BSC-Net produces accurate vessel segmentation results with preserved vascular continuity for quantitative coronary assessment, enabling reliable downstream analysis and clinical evaluation of coronary artery disease.
[CV-34] ESG: Generating Physically Consistent Dynamic 3D Scenes from Text Descriptions
链接: https://arxiv.org/abs/2609.15392
作者: Xintong Fang,Zhiyuan Fang,Rengan Xie,Xuhong Zhang,Guoyuan An,Zeran Liu,Jingyan Zhang,Jiarui Guo,Yuchi Huo
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV)
备注: 9pages
Abstract:Recent progress in image and 3D scene generation has enabled increasingly realistic static environments, yet most methods remain confined to such static configurations. Generating dynamic scenes from natural language is fundamentally challenging: it requires joint reasoning over scene structure, temporal evolution, and physical feasibility, while ensuring reliable execution in modern physics engines. We present a unified framework for generating physically consistent dynamic 3D scenes from text, with outputs directly executable in Unreal Engine. Central to our approach is the \emphEvolutive Scene Graph (ESG), which specifies entities with physical attributes, spatial relations, and event-driven timelines in a machine-checkable form. Given a prompt, a large language model constructs and validates a complete ESG; spatial layouts are grounded via energy-minimized gradient optimization; timeline-constrained physical parameters are then optimized through differentiable simulation to satisfy user-specified events; and the resulting scene is compiled into an engine-executable class. Experiments on 10 scenes across three complexity levels show that our method achieves 16.4/18 mean event completion, outperforming Scene Language, the strongest engine-executable baseline (SimWorld), and our ablation without physical optimization by a clear margin in event completion and parameter accuracy.
[CV-35] CapsuleMotion: A Lightweight Real-Time Visual Motion Predictor for Capsule Endoscopy
链接: https://arxiv.org/abs/2609.15367
作者: Oliver Bause,Julia Werner,Oliver Bringmann
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 4 pages, 3 figures, 2 tables, accepted at the 22nd Annual International Conference on Body Sensor Networks (BSN2026)
Abstract:Video Capsule Endoscopy (VCE) is a non-invasive medical examination that allows for the observation of the small intestine, which is otherwise difficult to access. A fundamental challenge persists in the form of their limited size in order to still be swallowable. The resulting restricted battery capacity, however, contradicts with the power-intensive nature of image capture and transmission. Therefore, we propose CapsuleMotion, a patient-specific dynamic capsule behavior that utilizes the available energy in a goal-oriented manner to increase the likelihood of a complete screening of the gastrointestinal tract. By investigating and combining metrics from the on-device image compression, CapsuleMotion predicts the motion between two successive frames. The camera’s frame rate will be modified in accordance with the predicted magnitude of motion. Furthermore, prior to entering the small intestine, the capsule operates in a low power mode with a significantly reduced frame rate. In this mode, the LocalizationNet is employed to determine the current organ, provided that motion was predicted. The proposed framework is evaluated on the Rhode Island VCE dataset and deployed on an ultra-low power single-core RISC-V demonstrator with an integrated hardware accelerator. CapsuleMotion demonstrated the capability to reduce electric energy consumption by up to 20.66% in comparison with conventional capsules that lack a dynamic frame rate. Additionally, the accuracy of detecting the entry point of the small intestine has been improved.
[CV-36] Diffusion Trajectory Modeling for Semantic Correspondence BMVC2026
链接: https://arxiv.org/abs/2609.15357
作者: Yusung Choi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to BMVC2026
Abstract:Diffusion models generate images through an iterative diffusion process, and recent studies have demonstrated that the intermediate feature maps produced during this process contain rich visual representations, leading to their adoption across a variety of downstream tasks. However, most existing approaches are limited to either using a single feature map at a specific timestep or aggregating feature maps across multiple timesteps. We observe that intermediate representations in the diffusion process form meaningful trajectories along the time axis. In particular, the representation of each spatial patch evolves progressively throughout the generative process, encoding semantics that are difficult to capture from static snapshots alone. This observation motivates the need to treat diffusion representations as temporally structured trajectories rather than static snapshots. To this end, we propose Diffusion Trajectory Modeling (DTM), a framework that interprets the temporal evolution of each spatial patch as a trajectory and leverages it for semantic correspondence. By effectively modeling patch-wise trajectories generated across multiple timesteps, DTM captures correspondence cues that prior methods are not designed to capture. We further demonstrate empirically that spatially corresponding patches form similar trajectory patterns throughout the diffusion process, suggesting that the temporal axis of diffusion carries semantic information. Experiments on SPair-71k, SPair-U and AP-10K show that DTM achieves strong performance, presenting a new perspective for exploiting diffusion representations from a trajectory-centric viewpoint.
[CV-37] End-to-End Cell Detection via Instance-aware Graph Modeling
链接: https://arxiv.org/abs/2609.15354
作者: Ruochen Liu,Yalin Zheng,Jingxin Liu,Jianfeng Zhang,Shoujun Huang,Dexing Kong,Haofeng Li,Wei Lou
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Accurate cell detection and classification are crucial for pathological analysis, directly affecting diagnostic accuracy and treatment planning. To capture complex cellular interactions beyond visual appearance within the tumor microenvironment, several approaches have employed graph neural networks to model spatial and relational patterns among cell nuclei, yielding promising results. However, these methods typically adopt a two-stage paradigm of visual extraction followed by relational modeling, which necessitates separate tuning for each stage, thereby increasing pipeline complexity and hindering end-to-end joint optimization. In this paper, we propose an end-to-end framework for cell detection and classification that jointly models patch-level visual representations and instance-level interactions, which incorporates a dynamic graph construction module and an instance-aware graph network. Specifically, the graph construction module dynamically builds the graph structure using learnable queries derived from patch-level features as cell instance representations, with adjacency defined by integrating feature similarity and spatial distances. The instance-aware graph network performs adaptive instance filtering and feature reorganization, aggregating them over the cell graph into a topological latent state for a selective state-space transition driven by visual cues, fusing appearance and relational evidence. When evaluated on multiple datasets with different staining protocols for cell and nucleus detection, our method significantly outperforms existing approaches in both detection and classification performance. The code will be released at this https URL.
[CV-38] Robust Multi-Model Fitting through Learning Neighbor Regions
链接: https://arxiv.org/abs/2609.15348
作者: Chang Nie,Guangming Wang,Zhe Liu,Hesheng Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multi-model fitting involves fitting multiple models accurately in a noisy environment. It is the basis for computer vision tasks such as scene reconstruction and mixed reality. However, its performance is often limited by insufficient feature utilization, inefficient optimization, model overlap, and the non-differentiable pipelines. To overcome these limitations, we introduce a robust coarse-to-fine framework called Learning Neighbor Regions (LNR). Recognizing that substantial computational resources are wasted on numerous bad minimum sets, we propose the coarse-level module. This module utilizes a neural network to extract and analyze geometric feature of both local point-wise relationships and global contextual information in minimum sets, outputting confidence to pre-select a small number of good minimum sets, thereby enhancing overall efficiency before solving hypotheses. To address model overlap, LNR encodes neighbor region features for each hypothesis in its fine-level module. These region features consist of geometric features of neighboring data points, which can be used by multiple regions simultaneously. This design allows the neural network to individually refine and score each hypothesis. Importantly, LNR is trained to learn directly from data point features rather than from the hypothesis parameters, thus avoiding differentiating the sampling process and the model solvers. Extensive experiments on four classic multi-model fitting tasks demonstrate that LNR achieves state-of-the-art performance. The analysis suggests that LNR can be easily adapted to various robust multi-model fitting tasks.
[CV-39] Query-Conditioned Spherical Centroid Aggregation for Multimodal Retrieval
链接: https://arxiv.org/abs/2609.15335
作者: Ambuj Mehrish,Anindya Nag,Sebastiano Vascon
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal retrieval integrates video, audio, subtitles, and text; however, recent geometric aggregators, such as Gramian volumes, hyperbolic volumes, and spectral objectives, treat all modalities symmetrically. Under a unified evaluation protocol, their joint scores frequently lag behind the strongest single-modality pathway by 1.9 to 27.6 R@1. Controlled analyses attribute this outcome to uniform modality influence. This work introduces Spherical Centroid Aggregation with Learned Adaptive Relevance (SCALAR), a query-conditioned aggregator that assigns relevance-based weights to each available modality before computing a spherical centroid. SCALAR accommodates arbitrary modality subsets and is trained on masked, reduced-arity views using rank-8 LoRA adapters. Across five benchmarks, SCALAR achieves positive aggregation gain on four, reaching +4.0 R@1, while none of the evaluated prior aggregators is positive on more than one. A uniform-weight ablation reproduces the degradation observed with symmetric aggregation. With only 4.8 million trainable parameters, SCALAR attains the highest text-to-video R@1 on three and performs within seed variation of the best result on a fourth. Under test-time modality dropout, SCALAR’s representation-stage score surpasses the released GRAM checkpoint at every evaluated masking rate and benchmark by 3.2 to 10.9 R@1. Finally, as modalities are removed, rerankers trained exclusively on complete modality sets increasingly converge toward their video-only pathways, diminishing these representation-level gains and underscoring a limitation of standard two-stage retrieval pipelines.
[CV-40] Concept-Grounded Reasoning with Prompt-Driven Localization for Interpretable Structured Report Generation
链接: https://arxiv.org/abs/2609.15334
作者: Xinyue Xu,Hongbin Lin,Juangui Xu,Hualiang Wang,Lehan Wang,Lijie Hu,Weiyang Liu,Adrian Weller,Xiaomeng Li
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Medical imaging modalities such as ultrasound and X-ray are widely used in clinical practice, where diagnosis follows a structured, evidence-driven workflow aligned with standardized criteria. While multimodal large language models (MLLMs) show promise for automated medical report generation, most existing systems rely on end-to-end multimodal fusion without modeling clinically defined intermediate attributes, leading to limited grounding and interpretability. To address this issue, we propose CORAL (COncept-grounded ReAsoning with Localization), a multimodal framework that integrates spatial grounding and concept-level supervision into a unified reasoning process. CORAL employs a prompt-driven medical segmentation model to localize lesions and predicts multi-class clinical attributes through a Concept Bottleneck module. The resulting textual concept tokens are combined with mask-modulated visual features within an MLLM to enable structured report generation and diagnostic prediction. Experiments on BUS-CoT and IU X-ray datasets demonstrate consistent improvements in diagnostic accuracy, concept consistency, and report quality over strong general-purpose and medical MLLMs, indicating that concept-grounded reasoning better aligns generation with clinical decision processes.
[CV-41] Hypergraph-Regularized Gramian Volumes for Multimodal Retrieval
链接: https://arxiv.org/abs/2609.15320
作者: Anindya Nag,Ambuj Mehrish,Sebastiano Vascon
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Volume-based multimodal retrieval jointly scores a text query with a candidate’s video, audio, and subtitle embeddings. While this approach captures higher-order within-candidate alignment, the score remains candidate-local, and semantically related training samples primarily serve as contrastive negatives. This work introduces Hypergraph-Regularized Gramian Volumes (HyVol), a training-time module that incorporates these semantic relations prior to evaluating the original volume loss. Document hyperedges connect the observed modalities of each candidate, whereas semantic hyperedges link candidates whose detached captions are mutual top-k neighbors. A shallow gated hyper-graph network applies residual corrections to the modality embeddings. Presence masks exclude unavailable streams from message passing, and identity padding preserves the determinant of the observed Gram submatrix without feature imputation. As refinement operates on embeddings rather than scores, the same construction applies to both Gram and HyperGram. We remove the hypergraph after training, leaving the backbone-only architecture, original scoring function, and retrieval cost unchanged. We train both backbones on a 150K-clip subset of VAST-27M and evaluate zero-shot performance on six benchmarks. Under the paired protocol, HyVol improves R@1 across all five retrieval benchmarks, with video-to-text gains reaching +8.3 on MSR-VTT and +7.6 on VATEX. Under missing-modality masking, the V2T margin remains positive in all experimental settings, although the T2V margin becomes slightly negative in four.
[CV-42] Learning from Reliable Negatives: Confidence-Anchored Test-Time Adaptation for GUI Grounding
链接: https://arxiv.org/abs/2609.15307
作者: Yizhou Liu,Fei Tang,Yuchen Yan,Zhengxi Lu,Songqin Nong,Tao Jiang,Wenhao Xu,Wenqi Zhang,Weiming Lu,Jun Xiao,Yongliang Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Graphical User Interface (GUI) grounding is essential for autonomous agents to map natural language instructions to precise screen coordinates. However, existing supervised fine-tuning and reinforcement learning methods are constrained by the high cost of annotation, creating a scalability bottleneck. In this paper, we introduce a label-free test-time training paradigm driven by two key insights: (1) confidence patterns in coordinate tokens are a better indicator than full-sequence confidence, and (2) in sparse GUI coordinate spaces, negative samples offer more reliable learning signals than potentially noisy positive ones. We first propose Confidence-Anchored Learning (CAL), which utilizes coordinate-token confidence to filter pseudo-labels and assign distance-based binary rewards. Building on this, we develop Confidence-Anchored Negative Learning (CANL), which exclusively optimizes the model using negative samples to bypass the risks of incorrect positive samples. Experimental results demonstrate that CANL-7B achieves 92.1% on ScreenSpot-V2. On more challenging ScreenSpot-Pro, CANL-7B reaches 33.8%, an 8.9% absolute improvement over the base model. Our findings establish coordinate-token confidence as a powerful alternative to manual annotations for scalable GUI agent development.
[CV-43] AlignUS: MRI-Guided Ultrasound Representation Learning for ALS Classification from Tongue Images
链接: https://arxiv.org/abs/2609.15285
作者: Kadija Abdel Ghader,Emani Babe,Lorenzo Pettinari,Meya Haroune,Sidaty El Hadramy
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease in which early assessment remains challenging, particularly in low-resource settings where MRI is often unavailable. High-resolution ultrasound (HRUS) of the tongue offers a portable and low-cost alternative for evaluating bulbar involvement, but learning reliable diagnostic models is limited by small datasets and the difficulty of extracting robust representations from ultrasound alone. We propose AlignUS, a cross-modal knowledge distillation framework that transfers anatomical knowledge from MRI to a HRUS-based classifier while requiring only HRUS at inference time. The model combines classification loss, supervised contrastive learning, and feature-level distillation to align HRUS representations with MRI embeddings. AlignUS achieves a patient-level balanced accuracy of 0.958, macro-F1 of 0.963, and ROC-AUC of 0.990, aggregated across four patient-level cross-validation folds, with consistent improvements over HRUS baselines and cross-modal alternatives. These results demonstrate that MRI-derived supervision can substantially improve ultrasound-based ALS assessment while preserving low-cost, inference-time independence from MRI.
[CV-44] A 25-μs/inf Event-driven Graph Neural Network Processor with Spatiotemporal Caching and Spline Convolution for Ultra-low-latency AI at the Edge
链接: https://arxiv.org/abs/2609.15241
作者: Adrian Kneip,Martin Lefebvre,Daniel Gehrig,Victoria Catalán Pastor,Davide Scaramuzza,Marian Verhelst,Charlotte Frenkel
类目: Computer Vision and Pattern Recognition (cs.CV); Hardware Architecture (cs.AR)
备注: This paper was accepted for presentation at the IEEE ESSERC 2026 conference
Abstract:Dynamic-vision-sensor (DVS) cameras generate events on a per-pixel basis with a \mu s-level temporal resolution, calling for new algorithm-hardware co-design approaches compared to standard frame-based vision. While event-driven graph neural networks (EV-GNNs) emerge as a promising algorithmic solution, they raise new HW challenges by mixing dense-regular compute operations and sparse-irregular memory accesses. We present ETHEREAL, the first EV-GNN accelerator that scales to 640 \times 480 resolutions, thanks to a neighbor-parallel spline convolution engine and a 2D/3D-split memory hierarchy with a novel region-of-interest spatiotemporal caching mechanism. Measurement results demonstrate end-to-end inference with 25.6 \mu s latency and 1.7 \mu J energy per event on state-of-the-art workloads
[CV-45] Pre-PEFT Probing: Weight Statistics and Perturbation Robustness for Layer Selection in VLM Vision Encoders IJCNN2026
链接: https://arxiv.org/abs/2609.15229
作者: Qingtao Xia,Jiahua Bao,Siyao Cheng,Jie Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 6 pages, 4 figures. Accepted at the 2026 International Joint Conference on Neural Networks (IJCNN 2026)
Abstract:We propose a pre-fine-tuning probing method for Parameter-Efficient Fine-Tuning (PEFT) layer selection, aiming to obtain more stable and higher gains with fewer trainable parameters when adapting large vision–language models (VLMs). Unlike the common practice of applying LoRA and other adapters to all layers at once—where layer selection often relies on heuristic rules—we focus on the vision encoder and directly evaluate the "adaptability’’ of each Transformer layer. Specifically, we characterize each layer from two perspectives: (i) the statistical properties of its Q/K/V projection weights (e.g., norms and condition numbers); (ii) robustness under controlled parameter perturbations. We then systematically compare these indicators with the downstream performance gains brought by applying PEFT to a single layer. Across experiments covering seven benchmarks and five PEFT variants, we observe a consistent correlation: layers (or matrices) with larger weight norms and higher condition numbers are usually more robust to perturbations and are more likely to yield larger fine-tuning gains. These results show that distribution-statistics analysis and perturbation tests before fine-tuning can provide practical signals for adaptation-layer selection, thereby maintaining or improving performance while reducing trainable parameters.
[CV-46] Unsupervised Point Cloud Registration via Training-Time Semantic Guidance ECCV2026
链接: https://arxiv.org/abs/2609.15228
作者: Kezheng Xiong,Shiyun Xu,Sheng Ao,Siqi Shen,Cheng Wang,Chenglu Wen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026
Abstract:Unsupervised registration of large-scale LiDAR point clouds remains challenging due to the geometric ambiguity inherent in outdoor scenes, which degrades pseudo-label quality and leads to suboptimal convergence, particularly for sparse, low-resolution scans such as those from nuScenes. We reveal that registration models intrinsically encode semantic awareness that strongly correlates with registration accuracy, albeit without explicit semantic supervision. However, this native awareness is fragile: noisy supervision arising from geometric ambiguity in unsupervised settings rapidly erodes the learned semantic structure, causing performance collapse. To this end, we propose CAESAR, a teacher-student framework guided by an off-the-shelf 3D segmentation model exclusively during training. We observe that potential inlier matches are often buried just beneath a few spurious neighbors in the noisy feature space, motivating Dual-Cue Guided Re-Matching to recover them through reselection rather than simply rejecting. Building on this, a train-only Semantic-Geometric Label Mining performs lightweight, batch-specific teacher refinement and mines reliable pseudo-labels under semantic guidance. We further introduce Semantic Predictive Distillation to consolidate the student’s semantic awareness in the feature space. Extensive experiments on KITTI and nuScenes demonstrate state-of-the-art performance, with pronounced gains on the challenging nuScenes benchmark. Crucially, CAESAR incurs zero inference overhead and requires no semantic annotations on the registration data. Code will be released.
[CV-47] Closed-form Bayesian homography estimation from noisy point correspondences
链接: https://arxiv.org/abs/2609.15227
作者: Hanne Beuter,Sebastian Dorn
类目: Computer Vision and Pattern Recognition (cs.CV); Methodology (stat.ME)
备注:
Abstract:While homographies are fundamental to many computer vision tasks, the majority of conventional estimation techniques provide only point estimates without directly quantifying uncertainty introduced by noisy observations. Uncertainty, though, propagates to subsequent processing steps such as camera calibration and 3D reconstruction and is particularly relevant in safety-critical and socially relevant fields including medical imaging, autonomous driving, and defense. We present a fast Bayesian formulation for homography estimation from point correspondences that explicitly incorporates measurement uncertainty and prior knowledge while providing a posterior distribution over the homography parameters. A closed-form solution of the posterior mean of the homography is derived in homogeneous coordinates and supplemented by an iterative Bayesian approach to handle non-linearities. Synthetic experiments demonstrate the applicability to projective transformations and show improved estimation accuracy over DLT under varying noise conditions. Image stitching experiments further demonstrate applicability to real image correspondences while additionally providing uncertainty information.
[CV-48] Deep Learning-based Intelligent Diagnosis of Congenital Uterine Anomalies in 3D Ultrasound
链接: https://arxiv.org/abs/2609.15225
作者: Yueyue Xu,Yuhao Huang,Jiaxiao Deng,Yuanji Zhang,Haoming Zhang,Jiajia Qu,Shiying Zheng,Xiaomei Tang,Haining Chen,Chengcai Chen,Yiyi Wu,Xin Yang,Dong Ni
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 22 pages, 7 figures, 4 tables
Abstract:Objective: To develop an intelligent framework, termed CUA-Net, for the automated classification of congenital uterine anomalies (CUA) without requiring coronal plane reconstruction, and to evaluate its clinical applicability. Methods: CUA-Net was built on 3D ResNet-18, equipped with a dynamic data resampling strategy to mitigate the data imbalance issue and a hard sample mining technique to fully learn from the difficult cases by loss adjustment. We further proposed the self-supervised reconstruction to comprehensively explore the volumes and the online data augmentation to refine the wrong predictions and enhance the model’s generalization. We compared the CUA-Net with different deep-learning methods and junior/senior sonographers in the testing set. The evaluation metrics included accuracy, precision, recall, F1-score, micro-AUC, and macro-AUC. Results: The proposed CUA-Net exhibited satisfactory performance in both internal and external test sets. In the internal cohort, the model achieved accuracy of 93.88%, precision of 87.01%, recall of 95.92%, F1-score of 88.09%, and micro-AUC of 0.9982 and macro-AUC of 0.9997. In the external set, it maintained good performance with accuracy of 91.52%, precision of 83.27%, recall of 88.63%, F1-score of 81.49%, micro-AUC of 0.9945 and macro-AUC of 0.9990. Our CUA-Net outperformed the junior sonographers across all performance indicators and achieved performance comparable to that of the senior sonographers across most metrics. Conclusion: The CUA-Net demonstrates favorable accuracy and generalizability in classifying common CUA categories, while showing preliminary potential for recognizing less prevalent anomalies. These capabilities may help optimize clinical workflows and support more standardized diagnosis. Comments: 22 pages, 7 figures, 4 tables Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.15225 [cs.CV] (or arXiv:2609.15225v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.15225 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Yuhao Huang [view email] [v1] Mon, 14 Sep 2026 08:44:47 UTC (6,966 KB)
[CV-49] Long-to-Short Video Evidence Reasoning for Grounded Question Answering
链接: https://arxiv.org/abs/2609.15224
作者: Kaiyan Chen,Junbin Xiao,Xun Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We present LOVER, a \underlineLong to sh\underlineOrt \underlineVideo \underlineEvidence \underlineReinforced model for grounded question answering (GQA). LOVER highlights three innovations over existing reinforcement-learning (RL) based video reasoning models: (1) \textbfLong-to-short Video Evidence Curriculum Learning, which organizes RL training according to evidence duration and progressively adapts the model from long-range grounding to short-term reasoning; (2) \textbfGQA Rewards, which underscore the benefit of IoP reward over IoU for evidence spotting rather than strict temporal span overlap; (3) \textbfAdaptive Timestamp Rendering, which adaptively renders timestamps onto video frames using background-aware position and color selection to enhance temporal observability. The three designs are model-agnostic and reciprocal. They effectively improve QA, grounding, and grounded QA performance over different backbones. Notably, LOVER built on Time-R1 achieves new state-of-the-art (SOTA) results among open-source models on popular GQA benchmarks: NExT-GQA and ReXTime. Comprehensive ablation studies further validate the effectiveness of our three innovative components.
[CV-50] Reconstructing Is Not Acting: Action-Centric Latent Dynamics Modeling
链接: https://arxiv.org/abs/2609.15189
作者: Dingjie Fu,Dianxing Shi,Yangyang Xu,Jun Yu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Tech report, 26 pages
Abstract:Latent action models (LAMs) learn action representations from unlabeled videos by inferring latent actions from visual transitions and reconstructing future states. However, we identify a fundamental \textbfreconstruction-action mismatch : lower reconstruction error does not necessarily yield better latent dynamics or downstream performance. We attribute this mismatch to two underconstrained aspects of reconstruction-based latent dynamics modeling: (i) the inverse dynamics model (IDM) is not explicitly encouraged to distinguish action-related transitions from nuisance appearance, and (ii) the forward dynamics model (FDM) can underutilize the inferred latent action by exploiting predictive shortcuts from the current state. To address both limitations, we propose \textbfACT-LAM , a lightweight action-centric framework that strengthens both action extraction and action utilization. Specifically, its Action Query IDM (AQ-IDM) employs learnable action queries and gated aggregation to selectively extract rich action-related transition cues without strong information bottlenecks. And its Action Token FDM (AT-FDM) projects latent actions into action tokens that progressively interact with evolving state representations, enabling continuous state-aware action conditioning. ACT-LAM further streamlines feature processing to concentrate model capacity on latent dynamics modeling. Extensive experiments on several robotic datasets and the VP ^2 benchmark demonstrate stronger latent action consistency, forward dynamics, and downstream visual planning performance with fewer trainable parameters and lower computational overhead. In particular, ACT-LAM surpasses the previous state of the art by \textbf7.6% on the aggregated VP ^2 success rate. Codes at \hrefthis https URLurl .
[CV-51] Does Attention-Guided Masking Really Help Object Discovery in Object-Centric Learning?
链接: https://arxiv.org/abs/2609.15187
作者: Youliang Tao,Yanhua Han,Bin Zhao,Juho Kannala,Joni Pajarinen,Rongzhen Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Object-Centric Learning (OCL) aims to decompose images into objects without human annotations. A major family of mainstream methods uses Slot Attention to aggregate image features into object-level representations and then from them reconstructs masked image content, i.e., Random Masking (RM), to provide self-supervision. The recent method DIAS simply masks image patches at uniform randomness yet achieves competitive object discovery accuracy. Since attention during aggregation already possesses object discovery ability, we explore using it to develop a better image patch masking strategy, i.e., Attention Guided Masking (AGM), thereby providing better self-supervision. Results on six recognized datasets show that AGM does not always outperform RM. Under unconditional slot initialization, AGM substantially improves background segmentation on datasets with realistic textures (COCO and VOC); Regardless of conditional or unconditional slot initialization and across datasets, foreground object discovery remains comparable or decreases. We suggest peer researchers in the OCL community that attempts to exploit internal attention semantics to improve OCL with masked decoding are risky. Our source code, model checkpoints and evaluation logs will be released upon acceptance.
[CV-52] EECTracker: Swarm Motion Prior-Guided Feature Compensation for Airborne Optical UAV Swarm Tracking
链接: https://arxiv.org/abs/2609.15171
作者: Zhaochen Chu,Tao Song,Ren Jin,Mingdong Jia,Defu Lin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages, 7 figures, 12 tables. This work has been submitted to the IEEE for possible this http URL may be transferred without notice, after which this version may no longer be accessible
Abstract:Airborne optical tracking of uncrewed aerial vehicle (UAV) swarms is challenging due to extremely small target scales, rapid viewpoint changes, and cluttered backgrounds, which can weaken target feature responses and lead to intermittent or temporarily missing detector responses. Existing multi-object tracking methods generally depend on reliable target-specific detector responses to maintain target states and identities across frames. When such responses become unreliable, target states cannot be reliably updated and cross-frame association cues become ambiguous, resulting in fragmented trajectories and identity switches. To address this problem, we propose EECTracker, a swarm-motion-prior-guided joint detection-and-tracking framework for airborne optical UAV swarm tracking. EECTracker constructs a probabilistic swarm motion prior from reliable historical tracklets to capture the shared short-term image-plane motion tendency of the swarm and its uncertainty, providing spatial guidance for cross-frame feature compensation. Building on this prior, we introduce Energy–Entropy Consistency Activation (EEC Activation) to evaluate motion-prior-conditioned feature consistency using feature residual energy and local residual entropy. The resulting Local EEC score guides pixel-level feature compensation by enhancing motion-prior-consistent feature responses in potential target regions while suppressing inconsistent background responses. Experiments on AIRMOT and UAVSwarm show that EECTracker achieves superior overall tracking performance compared with state-of-the-art methods. Compared with the strongest competing method SCT-MOT, EECTracker improves MOTA/IDF1 by 3.89/1.79 percentage points on AIRMOT and by 2.81/1.74 percentage points on UAVSwarm, while maintaining online inference speed.
[CV-53] GRAVA: Grounded Reasoning -to-Action Representation and Learning for Autonomous Driving
链接: https://arxiv.org/abs/2609.15169
作者: Xiao Liu,Haoyu Li,Jianghao Leng,Lin Wang,Chao Sun
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 23 pages. Code: this https URL
Abstract:Driving vision-language-action (VLA) models increasingly reason before acting, but their intermediate reasoning is often weakly grounded in physical scene evidence and loosely connected to executable behavior. We present GRAVA, a framework built around Grounded Reasoning-to-Action (GRA), which unifies grounding, reasoning, and action generation in a single autoregressive stream. GRA links action-relevant language references to 2D visual regions and ego-centric physical states, organizes object interactions and decisions in a trajectory-anchored typed graph, and serializes this structure into grounded reasoning. A single VLM generates this reasoning followed by a compact Executable Planner action that is deterministically decoded into a continuous trajectory. We further introduce an agentic GRA data construction pipeline that combines forward scene grounding with backward trajectory anchoring, and use it to build GR-NavSim with 2.2M grounded question-answer pairs and 70K GRA reasoning traces. A progressive training strategy develops grounded cognition through pre-training, establishes the reasoning-to-action interface through imitation, and improves driving behavior through reinforcement learning and exploration. Using about 60% of the available human driving demonstrations for action supervision, GRAVA-8B achieves state-of-the-art performance among purely autoregressive driving models on the full NAVSIM benchmark. On an internal long-tail benchmark, full GRA improves key-object compliance and Closed-loop Driving Score by 19.3% and 20.5% over action-only prediction, respectively. These results show the benefit of preserving action-relevant physical evidence from grounded reasoning through executable action generation.
[CV-54] Weakly Supervised Spatial Grounding for Discriminative Attention-Based Ultrasound-Histopathology Alignment in Prostate Cancer Grading
链接: https://arxiv.org/abs/2609.15150
作者: Obed Korshie Dzikunu,Emma Willis,Mohammad Mahdi Abootorabi,Mohamed Harmanani,Zhuoxin Guo,Ferdinand Luger,Adam Kinnaird,Brian Wodlinger,Parvin Mousavi,Purang Abolmaesumi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Unpaired cross-modal distillation transfers grade structure from histopathology into a micro-ultrasound (micro-US) encoder by aligning a pooled needle-region embedding to a frozen histopathology teacher under grade-group correspondence alone. A single objective is thereby required to serve two distinct functions: rendering patch features discriminative of tissue state, and selecting which patches enter the pooled representation. We decouple them. Weak spatial supervision derived from percentage involvement, recorded routinely at biopsy, constrains the predicted proportion of malignant tissue within each core, acting on the encoder features independently of the alignment objective. The alignment loss then operates on features that differ across a core, and attention concentrates on a subset of patches rather than remaining near-uniform. On 7,166 biopsy cores from 811 patients across seven centers under patient-level 5-fold cross-validation, the method reaches 67.1 macro AUC and 68.5 csPCa AUC, against 61.2 and 52.8 for the existing unpaired alignment method and 63.1 and 62.6 for the strongest unimodal baselines. Ablation against existing attention regularizers designed to prevent attention-uniformity collapse shows that such regularizers do not substitute for label-derived supervision: they constrain the attention distribution, whereas the signal required acts on the features that attention reads.
[CV-55] Automated Perceptually-Motivated Assessment of Photographic Consistency in Paired Clinical Photographs: Pipeline Development and Internal Evaluation
链接: https://arxiv.org/abs/2609.15144
作者: Derrick Lin,Samantha Rabinovich,Joclin Rabinovich,Kassra Garoosi,Sumun Khetpal,Evan Delanoy,Neel Bhardwaj,Jason Roostaeian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, 7 figures, 12 tables. Submitted to SPIE Journal of Medical Imaging
Abstract:Purpose: Paired pre- and post-operative photographs are the standard unit of evidence for plastic surgical outcomes, yet no objective metric verifies whether two images of the same patient were captured under conditions consistent for comparison. Approach: We developed a perceptually motivated pipeline that analyzes pre/post pairs across thirteen calibrated sub-metrics, partitioned by unsupervised correlation-structure analysis into five data-driven clusters (photometric, texture / sharpness, pose, illumination direction, and pitch), averaged within each cluster and combined across clusters by a weighted sum into a single consistency score. Each sub-metric is calibrated so that its median difference across published within-patient pairs scores 0.5, which is a reference point and carries no pass/fail meaning. The pipeline was calibrated on 134 matched within-patient published pre/post pairs and evaluated against identical-image pairs, synthetic-perturbation pairs, and 134 mismatched cross-publication pairs. Results: The master consistency score S separated matched from mismatched pairs (sensitivity index d’ = 2.15, 95% confidence interval (CI) [1.83, 2.55]; area under the receiver operating characteristic curve AUC = 0.928, 95% CI [0.896, 0.959]), closely matching Gaussian-equal-variance predictions. The three head-pose angles did not fall in one cluster: yaw and roll grouped together while pitch separated. Identical pairs scored at ceiling (S = 0.99) and the master score fell monotonically with perturbation magnitude on all five perturbation axes. Conclusions: The score quantifies photographic comparability, not aesthetic or surgical quality, and provides a freely available web tool for auditing the photographic comparability of pre/post pairs, pending validation against expert judgment. Comments: 26 pages, 7 figures, 12 tables. Submitted to SPIE Journal of Medical Imaging Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.15144 [cs.CV] (or arXiv:2609.15144v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.15144 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Derrick Lin [view email] [v1] Mon, 14 Sep 2026 07:19:17 UTC (1,292 KB)
[CV-56] SparseTalk - Sparsifying 3D Gaussian Language Fields for Efficient 3D Visual Question Answering
链接: https://arxiv.org/abs/2609.15137
作者: Davit Soselia,Joseph JaJa,Amitabh Varshney
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:3D Gaussian language fields provide an explicit, spatially grounded representation for 3D visual question answering (VQA), but their dense semantic features can require tens of thousands of embeddings per scene, resulting in substantial storage, memory, and inference costs. We investigate how much of this representation is actually necessary for downstream reasoning. Starting from a full embedding representation, we systematically sparsify its semantic embeddings, including the previously underexplored regime below a single image-equivalent block down to 8 visual tokens. We compare random, geometric, semantic, and joint spatial-semantic selection strategies and introduce an object-based sparsification method that distributes the token budget across detected object instances while retaining background context. Experiments on ScanQA and MV-ScanQA reveal substantial redundancy in dense Gaussian language fields. Strong VQA performance is retained with only a few hundred semantic embeddings, corresponding to less than 1% of the original representation. Object-based selection performs well relative to others, with only modest observed changes down to 256 tokens. At this budget, SparseTalk retains 0.80% of SplatTalk’s 32,076-token inference input and 0.332% of the mean 77,207-Gaussian dense field, increasing inference throughput while reducing decoded-feature memory 125-fold.
[CV-57] AdaVSkip: Adaptive Visual Token Skipping Across Layers For Efficient MLLM s Inference
链接: https://arxiv.org/abs/2609.15131
作者: Yuyao Sun,Tao Deng,Shuang Li,Deqing Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Multimodal large language models (MLLMs) require substantial computation to process numerous visual tokens across all transformer layers. Most methods for efficient MLLM inference exploit horizontal redundancy by compressing visual tokens. Beyond token reduction, recent studies exploit vertical redundancy through early exit or fixed-layer skipping. However, we find that the extent and distribution of this redundancy vary across inputs and differ between self-attention and MLP modules. Motivated by these observations, we propose AdaVSkip, which equips each layer with two lightweight routers that independently determine whether visual tokens pass through by or skip the self-attention and MLP modules. These decisions collectively define an input-specific visual-computation path, but their discrete and non-differentiable nature makes learning effective paths challenging. To address this challenge, we develop a progressive two-stage training framework that updates only the routers while keeping the backbone frozen. Stage I establishes an initial routing policy through supervised training with input-specific targets derived from module-wise necessity scores. To further align the routing policy with task performance, Stage II uses reinforcement learning to optimize routing decisions with direct feedback from generated answers. It combines an answer correctness reward with a skip-consistency reward that discourages excessive retention of visual-token computation. Across three MLLM backbones, AdaVSkip maintains strong task performance with substantially less computation. On LLaVA-NeXT-7B, AdaVSkip reduces FLOPs by 53.2% while preserving the original model’s average performance. Combining it with visual token compression increases this reduction to 91.2%, while retaining 97.2% of the original performance on average.
[CV-58] woma: a real-time foundation model and its fine-tuned models for endoscopy
链接: https://arxiv.org/abs/2609.15130
作者: Thang Tran,Lan Dang
类目: oftware Engineering (cs.SE); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 28 pages, 10 figures (8 in the main text, 2 supplementary), 13 tables (8 in the main text, 5 supplementary). Preprint. Models and run records are available from the corresponding author
Abstract:woma is a real-time foundation model for gastrointestinal endoscopy: a network trained without labels on about a million endoscopy frames, from which task models are fine-tuned. We contribute a systematic design for production. Requirements and pass marks were fixed before any run, eight candidates screened under pre-registered rules, self-supervised training taken to a stopping rule, then fine-tuning and deployment optimisation, all on one self-contained library, numbat. We also contribute woma itself with two fine-tuned models, every outcome reported met or missed. Our colonoscopy model finds and outlines polyps, names which colon segment is in view, suggests polyp type and grades bowel preparation. Our gastroscopy model names a station out of 22 protocol sites, flags and outlines lesions, and names one of seven findings. Every number was read on data never seen in training, and shipped weights were chosen on that record. In colonoscopy, 96% of polyps in a six-hospital PolypGen set are found at precision =0.85, and 19 of 19 polyps across fifteen full REAL-Colon videos at 1.6 false alarms per procedure. In gastroscopy, landmark region is named correctly on 92% of frames from unseen patients, and 37 of 39 held-out neoplasia frames are flagged at specificity 0.91. On one workstation GPU every task runs over 1080p video at about 100 frames per second, faster than PyTorch, ONNX Runtime and TensorRT in all four precision regimes tested. TensorRT comes closest: one pass of our foundation model takes it 3 to 27% longer than ours, and we deliver 6 to 31% more frames per second from frame to results. A second build links no vendor library at all – our own kernels over Vulkan – so a site deploys two files and needs no toolkit, no cuDNN and no framework; in f32 it beats the CUDA build on the same card.
[CV-59] DNF-SR: Dual-Input and Negative-Aware Feature Fine-Tuning for Real-World Image Super-Resolution CVPR2026
链接: https://arxiv.org/abs/2609.15120
作者: Shuhao Han,Wenjie Liao,Hayden Vance,Hang Dong,Rui Zhang,Chun-Le Guo,Chongyi Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by CVPR 2026
Abstract:Benefiting from the powerful generative priors of diffusion models, diffusion-based real-world image super-resolution (Real-ISR) methods have demonstrated impressive this http URL achieve efficient Real-ISR, several recent works have designed one-step diffusion-based this http URL, unmediatedly feeding LR into a diffusion model creates a distributional gap with the model’s original input.A straightforward approach to reduce the distribution gap is to introduce noise to the LR latents. However, directly adding noise inevitably corrupts the content of the LR this http URL this study, we propose DNF-SR, a Dual-input and Negative-aware Feature fine-tuning method for this http URL, we use a dual-input strategy that concatenates the original LR image with the noisy LR input and feeds them into a diffusion-based image editing model, ensuring both high-fidelity one-step super-resolution and improved perceptual and content this http URL, the noise present in the noisy LR input introduces randomness and diversity into the outputs. We exploit this property and propose a post-training optimization method, Negative-aware Feature Fine-Tuning (NF2T), which guides the model toward producing higher-quality this http URL^2T classifies multiple outputs into positive and negative subsets and then defines implicit policy improvement directions in both the image and feature spaces, thereby further enhancing the stability of the this http URL experiments show that DNF-SR outperforms other this http URL will be released.
[CV-60] ChatGPT Images 2.5 in the Wild: A Launch-Period Dataset and Detector Evaluation
链接: https://arxiv.org/abs/2609.15100
作者: Dennis Ng,Xingyu Shen,Ankit Raj,Kidus Zewde,Tommy Duong,Yuchen Zhou,Yuxin Zhang,Neo Tiangratanakul,Simiao Ren
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 22 pages, 8 figures, 12 tables
Abstract:An image tool can change its underlying generator while retaining its public name, making version attribution from online posts ambiguous. We study this problem after the ChatGPT Images 2.5 launch. Our frozen collection contains 3,478 images from 2,440 posts across 8 sources. Recorded posting times fall within the first 51.1 hours after the announcement. It records three attribution tiers and retains standalone images after image-form filtering and targeted review. Caption claims and host records provide admission evidence, not independently verified generator identity. The observed content profile depends on the source mixture: NightCafe supplies 39.0% of images but 77.0% of CLIP-assigned fantasy scenes. We then evaluate six frozen detectors at thresholds calibrated to a 5% flag rate on reference photographs. Collection flag rates range from 3.7 to 56.4%, falling 42-81 percentage points below GenImage recall. Held-out artwork false-positive rates range from 1.5 to 96.5%, so a higher collection flag rate does not by itself establish better detection. An exploratory X-only comparison with our April collection finds a higher September flag rate for Effort, and a suggestive difference for DoU, under fixed-threshold post-clustered bootstrap intervals. Attribution, content and processing differences prevent a causal interpretation of these contrasts. The collection supports analysis of reported model use during a product transition, with source and attribution evidence retained for interpretation.
[CV-61] LG-VLN: A Zero-Shot Vision-and-Language Navigation Framework with LangGraph State Orchestration
链接: https://arxiv.org/abs/2609.15098
作者: Jianhe Zhao,Yanhua Qiu,Zhiyu Zhang,Zibo Zhao,Jinhua Xie
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 30 pages, 4 figures
Abstract:Continuous-environment vision-and-language navigation (VLN-CE) requires interpreting natural-language instructions in unseen 3D environments and executing continuous low-level actions. Existing methods often depend on LiDAR, panoramic cameras, or extra sensors; separate geometric-mapping and semantic-navigation visual representations can cause long-trajectory spatial-semantic inconsistencies. We propose LG-VLN, a monocular zero-shot framework with shared visual features and LangGraph-based state orchestration. An online feed-forward 3D reconstruction network predicts depth, camera poses, and dense point clouds for agent-pose estimation and global map fusion. Geometry and navigation share dense CleanDIFT features: semantic consistency rejects incorrect inter-frame correspondences, while target-instance constraints define visual references whose similarity combines with local BLIP-2 image-text relevance to form a semantic value map. LangGraph represents instruction parsing, geometric perception, semantic value updates, path planning, action execution, and failure recovery as a directed state graph with conditional transitions, persistent state, and modular recovery mechanisms. On a fixed 550-episode subset of the R2R-CE val-unseen split, LG-VLN achieves 21.3% success and 12.1% success weighted by path length. Ablations show shared semantic features improve navigation, further boosted by combining visual similarity and image-text relevance. Results establish shared visual representations and explicit state orchestration as effective for zero-shot VLN-CE using monocular RGB alone. Code will be publicly released for reproducibility.
[CV-62] 360: Real-Time Feed-Forward Human Reconstruction from Sparse Unposed Cameras
链接: https://arxiv.org/abs/2609.15032
作者: Hanzhang Tu,Zhanfeng Liao,Wei Min,Jiajun Zhang,Yebin Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Live free-viewpoint visualization of real humans is critical for immersive communication and interactive digital experiences. Existing methods either rely on computationally expensive optimization or require calibrated cameras and low-resolution inputs, making real-time high-resolution deployment impractical. In this work, we present Tele360, the first real-time feed-forward system for dynamic human reconstruction and live free-viewpoint visualization from sparse, unposed RGB streams. Our system jointly estimates camera poses and reconstructs a dynamic 3D Gaussian representation for each time instance in a single forward pass. To achieve this, we start by designing a lightweight sparsity-aware multi-view transformer backbone that tokenizes foreground human regions while preserving global context through a shared scene token. We then employ a fully transformer-based Gaussian decoder to mitigate convolution-induced over-smoothing while keeping decoding sparse and efficient. In addition, we introduce a hybrid feature pyramid that injects multi-scale appearance cues into geometry prediction. We further introduce a lightweight differentiable Levenberg-Marquardt camera refinement layer to enhance multi-view consistency and geometric alignment. Moreover, to stabilize learning under sparse, unposed inputs, we transfer multi-view geometry priors from a large visual-geometry foundation model via teacher-student distillation. Finally, the predicted Gaussian maps are streamed with video codecs to remote devices for interactive free-viewpoint rendering. Extensive experiments show that Tele360 achieves state-of-the-art visual quality on studio benchmarks while supporting real-time 2K input-to-rendering at over 25 FPS on a single consumer GPU. Additional captured sequences illustrate its performance across varied subjects, clothing, and motions under our multi-camera setup.
[CV-63] winICL: Diagnosing Multimodal In-Context Learning through Paired Counterfactuals
链接: https://arxiv.org/abs/2609.15028
作者: Zihan Xue,Po-Yi Lu,Serhii Honcharenko,Zih-Ching Chen,Hsuan-Tien Lin,Nanyun Peng,I-Hung Hsu,Kuan-Hao Huang
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 21 pages, 4 figures
Abstract:In-context learning (ICL) enables models to infer tasks from demonstrations, but existing benchmarks generally lack matched text and image versions needed to compare ICL performance across modalities. We introduce TwinICL, a procedurally generated benchmark providing such pairs for controlled comparison. Across six open-weight models and 38 tasks, multimodal ICL consistently underperforms text-only ICL, with gaps varying by task family. To test whether this gap can be recovered, we target visual access, task framing, and reasoning through three interventions. Their combination recovers strong multimodal ICL performance on a diagnostic subset, despite limited or inconsistent individual effects. To distinguish difficulties in executing tasks from those in inferring them, we evaluate models with explicit task instructions, revealing a modality gap even when the task is known. We then examine how adding demonstration inputs and outputs reshapes this gap, highlighting demonstrations’ dual role as additional context to process and evidence about the task. The dataset is available at this https URL.
[CV-64] G-ray: Ray-Level Relative Geometric Position Encoding in Multi-View Vision Transformers under Camera Heterogeneity
链接: https://arxiv.org/abs/2609.15018
作者: Shuo Zhang,Xin Su,Wei Wang,Jun Liu,Xinrui Zeng,Yongsen Chen,Chenjie Wang,Guibo Zhu,Jinqiao Wang,Bin Luo,Liangpei Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, 13 figures, 14 tables. Supplementary material included in the appendix. Project page: this https URL
Abstract:We study relative position encoding for multi-view vision Transformers under camera heterogeneity, including varying fields of view (FoVs) or projection models. Existing rotary relative position encodings commonly use image-plane positional coordinates, producing projection-dependent relative phases and inconsistent geometric cues for cross-projection attention. We introduce G-ray, a ray-level relative position encoding whose rotary phases are parameterized by camera-local ray angles. The same camera-local ray pair induces the same relative phase across projections, providing projection-invariant positional consistency. G-ray can be used directly or integrated with existing encodings, retaining complementary geometric cues without additional learned parameters. We validate G-ray in three host encodings, RoPE, GTA, and RayRoPE, across 3D reconstruction and novel-view synthesis (NVS). Across three heterogeneous 3D reconstruction benchmarks at 50 views, G-ray leads all six averaged metrics and reduces mean pointmap relative error by 45.8% over MapAnything, with calibration supplied to both. Trained exclusively on homogeneous pinhole images, the 3D reconstruction model handles mixed pinhole and non-pinhole inputs without retraining and remains competitive on homogeneous pinhole 3D reconstruction protocols. For NVS, GTA and RayRoPE improve with G-ray under joint viewpoint and FoV variation. The project’s webpage is available at this https URL.
[CV-65] PhysBrain 1.5: From Vision-Language Models to Physical Foundation Models
链接: https://arxiv.org/abs/2609.14973
作者: DeepCybo Team,Yu Bin,Haipeng Cao,Zheng Chang,Kai Chen,Youning Chen,Kailin Deng,Yichao Du,Xiaotong Fu,Haoyang Ge,Yunlong Guo,Chenliu Hao,Jiyan He,Xuguo He,Yakun Hou,Kai Hu,Cong Huang,Tuopusen Huang,Yu Huang,Hong Li,Peize Li,Shijie Lian,Xiaopeng Lin,Yun Lin,Haibao Liu,Haochen Liu,Qiuzhi Liu,Shengcai Liu,Zhiqiang Liu,Tao Luo,Peng Ren,Shuo Ren,Chaoyi Ruan,Zhaolong Shen,Yukun Shi,Qiyuan Su,Yuxuan Tian,Yining Wang,Changti Wu,Hao Wu,Xueyin Xu,Ruoqi Yang,Zhaoyang Yang,Hang Yuan,Zhaoyang Zeng,Hanwen Zhang,Ruimeng Zhang,Yao Zhang,Yibo Zhang,Yuxiang Zhang,Zhirui Zhang,Ziyi Zhang,Zubin Zheng,Zishen Zhuang
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: PhysBrain 1.5 technical report. Project: this https URL
Abstract:We present PhysBrain 1.5, a unified model for understanding physical environments, generating actions, and predicting future states. Motivated by the physical loop of observation, interaction, and environmental change, we bring these capabilities into a common learning framework. Starting from a general vision–language model, we encode language responses, end-effector motion, and dense visual targets as discrete sequences and jointly optimize them with autoregressive next-token prediction. Pre-training draws its embodied supervision entirely from human interaction videos, using task-centered episodes to pair semantic and spatial context with recovered motion and subsequent observations. We then adapt the model through supervised fine-tuning on a mixture of human demonstrations, robot trajectories, and simulated experience. Across 28 embodied understanding benchmarks, our 8B model achieves an average score of 72.5, setting a new open-source state of the art and performing on par with leading proprietary models such as GPT-6-Astra and Gemini 3.6 Flash. It achieves the best open-source results on 14 benchmarks while retaining general multimodal capabilities. Beyond these understanding evaluations, qualitative examples show the model’s ability to produce end-effector trajectories and predict future scenes through spatially aligned RGB, depth, and robot-mask outputs.
[CV-66] MoVT: Video-Augmented Motion Tokenizer for Text-to-Motion Generation
链接: https://arxiv.org/abs/2609.14965
作者: Beibei Jing,Tianle Guo,Youjia Zhang,Zikai Song,Yawei Luo,Junqing Yu,Tao Guan,Wei Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Text-driven 3D human motion generation models face significant challenges in responding to diverse and unconstrained textual prompts, primarily due to the limited availability of 3D motion training data. To address this, we introduce MoVT, a novel framework that effectively leverages the extensive range of human action videos to enhance text-to-motion generation. At the core of our approach is the cross-modal augmented motion tokenizer, which projects discrete 3D motion tokens into the 2D domain. This projection allows us to enrich the motion codebook with complex, real-world motion patterns derived from videos. The enriched discrete tokens are then mapped back to the 3D domain, resulting in aligned 3D and 2D codebooks with an enhanced capacity to represent intricate motions. These enhanced codebooks are integrated into a generative masked transformer, which predicts masked motion token indices in a modality-agnostic manner. This enables the use of text-index pairs, generated from the 2D codebook and annotated motion videos, to further enhance the generator. Extensive empirical evaluations show that MoVT performs favorably against prior state-of-the-art methods across multiple key metrics.
[CV-67] hreshGuide: Class-Aware Labeled-Guided Thresholding for Semi-Supervised 3D Abdominal Multi-Organ Segmentation
链接: https://arxiv.org/abs/2609.14943
作者: Hongyu Liu,Yinlong Wang,Lusha Li,Hui Meng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 10 figures
Abstract:Pseudo-labeling is a strong paradigm for semi-supervised medical image segmentation, yet its effectiveness is highly sensitive to confidence thresholding. In abdominal multi-organ segmentation, a fixed global threshold is particularly suboptimal because organ classes differ substantially in size, appearance, and learning difficulty. In this work, we propose ThreshGuide, a class-aware threshold adaptation framework that uses labeled data to guide pseudo-label selection on unlabeled data. Built upon a standard teacher-student architecture, the teacher model evaluates labeled samples during training to estimate class-aware threshold targets by maximizing an error-aware F\beta criterion that balances precision and coverage. These targets are then smoothed with an exponential moving average (EMA) and used to filter unlabeled voxels in a class-dependent manner. Experiments on FLARE2022 and AMOS2022 show that ThreshGuide performs competitively overall, yielding clear improvements specifically on hard-to-learn organs.
[CV-68] What Makes a 3D Scene Editable? A Factorized Benchmark of Fidelity Locality Consistency and Preservation
链接: https://arxiv.org/abs/2609.14899
作者: Sariah Patro,Arjun Mehra,Nikhil Bhatia
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Neural 3D scene editing is often evaluated by semantic alignment alone, although a convincing result may alter unrelated content or become inconsistent across views. We introduce EditBench3D, a representation-agnostic benchmark that treats editing as controlled information replacement. It evaluates four complementary properties: instruction fidelity, spatial locality, cross-view consistency, and preservation of non-target content. The protocol combines visibility-aware 3D target supports, paired descriptions, held-out cameras, and five edit families covering appearance, material, geometry, and object-level changes. We evaluate eight representative NeRF, 3D Gaussian Splatting, hybrid, and proxy-based editors on 240 scene-edit pairs. The study shows that semantic fidelity is only weakly associated with the other editing properties, and that no single method is optimal across all dimensions. Explicit Gaussian editors offer a strong overall balance, whereas direct proxy manipulation provides the most conservative edits at the cost of open-ended fidelity. These findings support reporting editability as a multi-objective profile rather than a single semantic score.
[CV-69] RAIN: Region-Aware Inversion Network for Semantic Watermark Extraction
链接: https://arxiv.org/abs/2609.14856
作者: Zilai Li
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Semantic watermarks for diffusion models embed ownership information into the generative process while preserving perceptual quality, but Gaussian-Shading extraction conventionally requires multi-step diffusion inversion to recover the initial noise. Recent one-step methods show that this cost can be reduced substantially. We study this problem through extended flow matching and conditional regression. The key observation is that, near the high-SNR image endpoint, recovering a useful noise statistic given by the first-step output of the extended flow matching in the high-SNR regime is much simpler than reconstructing the full inverse trajectory, and Gaussian Shading only requires the recovered latent to remain in the correct watermark decision region. Based on this observation, we propose a lightweight, prompt-free extractor that decomposes endpoint recovery into an image-like anchor and a noise-oriented residual, which increases the capability of the model to utilize GPU parallel computation. The resulting method avoids iterative inversion and repeated evaluation of a diffusion-scale U-Net, providing an efficient one-step extraction pipeline with a concise theoretical interpretation. The computational cost of extracting noise is lower than that of both OSI and FARI. The github repo is there: this https URL
[CV-70] POLARIS: Training-Free Audio Fingerprinting with Saliency-Based Landmarks and Delaunay Grouping
链接: https://arxiv.org/abs/2609.14820
作者: Jiheng Li
类目: ound (cs.SD); Computer Vision and Pattern Recognition (cs.CV); Audio and Speech Processing (eess.AS); Signal Processing (eess.SP)
备注:
Abstract:This work presents POLARIS, a training-free audio fingerprinting system that selects landmarks from a locally normalized saliency field and groups them into sparse fingerprints using Delaunay triangulation. To deal with query distortion, POLARIS adds fingerprints from two-hop Delaunay neighborhoods only at query time, without enlarging the reference index. An adaptive configuration applies this expansion only when the original fingerprints do not produce a confident match. We evaluate POLARIS on synthetic distortions from the public PEX Hard Medium benchmark, excluding queries with pitch or tempo shifts, and on a new benchmark of real re-recorded music. POLARIS achieves the best performance among the evaluated training-free methods on both benchmarks. On the real recordings, its adaptive configuration also outperforms the neural NMFP baseline with a comparable measured query time and a smaller logical reference payload. Code, dataset, and instructions for reproducing all experiments are available at this https URL.
[CV-71] Zero-shot video highlight detection based on text descriptions and synthetic images
链接: https://arxiv.org/abs/2609.14790
作者: Michal Byra,Alberto Presta,Grzegorz Stefanski,Krzysztof Arendt
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages
Abstract:Detecting video highlights, the most informative or engaging moments in a video, is important for applications such as video summarization and content recommendation. We propose a zero-shot framework that combines CLIP, large language models (LLMs), and diffusion models. Given lightweight video metadata, such as a title or category, an LLM generates textual descriptions of likely highlight events. These descriptions are further converted into synthetic visual prototypes using a diffusion model. Textual and visual representations are matched to video frames using CLIP, enabling frame-level highlight detection without highlight annotations or dataset-specific training. Experiments on TVSum and SumMe demonstrate strong zero-shot performance, with particularly favorable results on TVSum. The proposed approach provides an effective framework for metadata-conditioned zero-shot video highlight detection.
[CV-72] From Visual Feedback to Textual Reviews: A Multi-Agent Vision-Language Framework for Image-Grounded Review Assistance
链接: https://arxiv.org/abs/2609.14761
作者: Utsav Kumar Nareti,Ayush Bansal,Kumari Priya,Chandranath Adak,Soumi Chattopadhyay,Muhammad Saqib,Saeed Anwar
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Visual feedback in the form of user-uploaded images and videos is becoming increasingly common in e-commerce platforms because it provides authentic evidence of product quality, defects, packaging conditions, and real-world usage. However, visual feedback alone often lacks the contextual explanations and subjective opinions necessary for informed decision-making, while many users provide limited textual feedback due to the effort required to compose detailed reviews. To bridge this gap, we introduce image-grounded review assistance, a novel task that aims to generate editable review drafts from user-uploaded product images. Unlike conventional image captioning, which focuses on objective visual description, the proposed task requires product-specific understanding, sentiment estimation, and evidence-driven review composition under challenging real-world conditions, including degraded image quality, excessive zoom-in, target ambiguity, and partial product visibility. We propose a multi-agent vision-language framework consisting of four specialised roles: product grounding, visual sentiment estimation, visual evidence generation, and review synthesis. The framework employs explicit intermediate representations, including product entities, predicted ratings, and evidence summaries, to improve interpretability and visual grounding. Experiments on a curated subset of the Amazon Reviews Electronics dataset demonstrate the feasibility of generating coherent, product-aware, and sentiment-aware review drafts from visual feedback. To the best of our knowledge, this is the first study to formulate image-grounded review assistance as a multi-agent vision-language reasoning problem, providing a practical step toward AI-assisted review authoring in e-commerce systems.
[CV-73] CrossDistill: Balancing Quality and Diversity via Trajectory-Level Hybrid Few-Step Distillation
链接: https://arxiv.org/abs/2609.14725
作者: Yuxi Liu,Haoyu Li,Yixiang Cai,Tengxu Sun,Zekun Zhang,Baole Ai,Ang Wang,Jiamang Wang,Lin Qu,Kun Yuan,Kai Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Few-step distillation accelerates diffusion models but must balance diversity and fidelity: trajectory-based distillation preserves mode coverage, while distribution matching sharpens samples but can reduce diversity. We show that this tension can be exploited in a noise-regime-dependent way: high-noise steps largely determine global modes, whereas low-noise steps refine local details. We propose CrossDistill, a trajectory-level hybrid distillation framework that splits the sampling trajectory at a crossover point, applies a trajectory-preserving objective on the high-noise interval and a distribution-matching objective on the low-noise interval, and couples the two stages through the crossover state. In contrast to loss-level mixing, and complementarily to training-time two-stage recipes, CrossDistill explicitly assigns complementary objectives along the noise axis, so that global branching is preserved before local statistics are sharpened. CrossDistill is a noise-level scheduling policy: PCM and DMD are plug-in instantiations, while the noise partition, crossover coupling, and objective ordering are the key design elements. Experiments on text-to-video diffusion models and qualitative image-to-video results show that CrossDistill expands the few-step quality-diversity frontier, retaining seed-level variation while achieving competitive visual fidelity.
[CV-74] PC2-AD: Point Cloud Upsampling to Safeguard 3D Anomaly Detection with Resolution-constrained Edge Devices
链接: https://arxiv.org/abs/2609.14722
作者: Yutong Gu,Yingxi Xie,Kejin Huang,Jian Ning,Hanzhe Liang,Linlin Shen,Jinbao Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 17 pages, including 6 pages of supplementary material. Code: this https URL
Abstract:Low-cost and low-resolution sensors used in edge deployments can produce test point clouds that are substantially sparser than the normal training data. This train-test sampling-resolution gap changes the local geometry available to a 3D anomaly detector. We propose PC ^2 -AD, a point cloud upsampling framework that compensates sparse test inputs before downstream detection. Target Domain Candidate Generation (TCG) adapts a pretrained upsampler to normal training geometry and generates a dense candidate pool. Geometry-Aware Candidate Filtering (GACF) selects candidates according to geometric spacing and spatial coverage. Normality-Preserving Point Compensation (NPPC) refines the selection by comparing candidate normality scores with those of their input anchors. The selected points are combined with the unchanged input points and processed by the existing detector. Experiments with six detectors on two Anomaly-ShapeNet settings and Real3D-AD show improvements in the mean of object-level and point-level AUROC for all six detectors in each Anomaly-ShapeNet setting and four on Real3D-AD. These results support point cloud compensation as an input-level approach to improving 3D anomaly detection under low-resolution sensing conditions. Code is publicly available at this https URL.
[CV-75] Patient-Level Leakage-Aware Deep Learning for Cross-Center Periapical Radiograph Classification
链接: https://arxiv.org/abs/2609.14703
作者: Md Jubaer Rahman,Ulas Bagci
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to the 9th International Conference on Signal Processing and Information Security (ICSPIS 2026)
Abstract:Dental caries and endodontic disease are among the most common health conditions worldwide, and intraoral periapical radiographs are central to their detection, treatment planning, and follow-up. Automated tooth-level classification of these images, however, lacks reproducible benchmarks, is often evaluated with image-level splits that leak patients between training and test, and is rarely validated across clinics. This paper presents the first patient-level, leakage-aware classification benchmark for single-tooth intraoral periapical radiographs on the DentIRO dataset, which comprises 5,300 images from 3,243 patients across two clinics and four classes: Healthy, Caries, Crowned, and Root Canal. Five transfer-learning models are compared with patient-grouped stratified cross-validation, so that every patient remains within a single fold. DenseNet121 gave the strongest and most stable result at a mean macro-F1 of 0.9787, while the four ImageNet-initialized backbones performed comparably. A controlled comparison on a fixed architecture showed that chest-radiograph pretraining transferred less effectively than ImageNet initialization. Bidirectional cross-center validation produced a small average generalization gap of 0.0077, and Grad-CAM confirmed that predictions rely on clinically meaningful tooth regions rather than acquisition artifacts. The benchmark offers a rigorous and reproducible baseline for intraoral radiograph classification.
[CV-76] Perceive Refine Reason : A Calibrated Pipeline for Measuring Indicators in Strategic Visual Communication on Social Media
链接: https://arxiv.org/abs/2609.14699
作者: Weihong Qi,Chen Ling
类目: Computer Vision and Pattern Recognition (cs.CV); Computers and Society (cs.CY)
备注:
Abstract:Visual content shapes audience perception and opinion on social media, and computational social science increasingly relies on automated tools to analyze images at scale. Yet a measurement gap persists: existing tools rely on predefined categories or produce only coarse image-level labels, while measuring which specific objects appear in an image, how prominently, and where in the frame remains difficult at scale. We introduce Perceive, Refine, Reason (PRR), a calibrated pipeline that turns flexible vision-language detectors into auditable measurement instruments for social-scientific research. PRR combines natural-language category prompts with pixel-level spatial refinement via the Segment Anything Model (SAM) and a multimodal LLM arbitration layer whose reasoning chains externalize domain knowledge and lower the expertise threshold for human-in-the-loop validation. A complementary three-tier auditability framework applies quantification learning to profile per-category reliability, support task-aligned configuration, and statistically correct prevalence estimates. Across four vision-language detectors and nine sociological categories, the pipeline yields substantial precision gains over zero-shot baselines, including a 43.3-point improvement for the strongest backbone. Applying PRR to 103,920 Facebook images from U.S. legislators during the 2024 election cycle and linking detections to DW-NOMINATE ideology scores, we find that more conservative legislators display U.S. flags as larger visual elements, with a weaker tendency toward peripheral placement, a spatial pattern invisible to binary detection. PRR provides computational social scientists with a model-agnostic toolkit for accessible, spatially-grounded, and correctable visual measurement.
[CV-77] LIMODENet: Attention-Free Compact Encoders for Information-Preserving Onboard Satellite Image Restoration
链接: https://arxiv.org/abs/2609.14690
作者: Thanh-Dung Le,Vu Nguyen Ha,Ti Ti Nguyen,Symeon Chatzinotas
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 25 pages, 3 figures, 5 tables; includes supplementary material with full proofs. Code and weights: this https URL and this https URL
Abstract:Onboard satellites must restore a channel-degraded image on a few watts, using neuromorphic accelerators (e.g., BrainChip Akida, Intel Loihi-2) that support no softmax or attention. We ask which encoder restores best under that constraint and introduce LIMODENet (LinearMix-ODENet), a 0.69M softmax-/QKV-free backbone whose residual stages read as ODE discretizations and which is empirically information-preserving (probe accuracy rises 79.9% - 98.4% from stem to head). At iso-parameters it restores 1 dB DVB-S2X-degraded EuroSAT better than a CNN autoencoder (+1.75 dB PSNR) and a skip-connection U-Net (+1.07 dB), three seeds, non-overlapping. Unconstrained modern restorers (NAFNet, Restormer) win on fidelity; we decompose that gap: spiking-legal additive skips recover about half, and the rest traces to attention and channel gating. LIMODENet then converts end-to-end to a spiking network with zero blocked operations, versus 22-24 for the competitors: not the best restorer available, but the best verified deployable within a real power budget.
[CV-78] From Density to Biopsy Decisions and Malignancy Prediction: A Benchmark Study of Multimodal Large Language Models Against Radiologists in Digital and Contrast-Enhanced Mammography
链接: https://arxiv.org/abs/2609.14676
作者: Ali Abbasian Ardakani,Afshin Mohammadi,Taha Yusuf Kuzan,Beyza Nur Kuzan,Alisa Mohebbi,Masume Behruzi,Hamid Khorshidi,Ashkan Ghorbani,Elham Asadiara,Zeinab Khorshidi Lotfi,Ansar Rahman,Nedim Christoph Beste,U. Rajendra Acharya,Sepideh Hatamikia
类目: Computer Vision and Pattern Recognition (cs.CV); Medical Physics (physics.med-ph)
备注: 37 pages, 6 figures, 5 tables
Abstract:Purpose: To compare four multimodal large language models (MLLMs) with radiologists of varying expertise in breast density assessment, BI-RADS assessment, biopsy candidacy determination, and continuous malignancy probability estimation using digital mammography (DM) and contrast-enhanced mammography (CEM). Methods: This study included 179 women with paired DM/CEM examinations and reference standards. Four MLLMs (ChatGPT-5.2, Gemini-3.1 Pro, Sonnet-4.6, Muse Spark) interpreted images with and without masks; three radiologists interpreted non-masked images. Results: For binary density classification on DM, radiologist accuracies ranged from 55.81% to 78.60%, exceeding most MLLM values (62.33%-71.63%), while masks added limited benefit. Five-category BI-RADS accuracies were higher for radiologists on DM (56.74-67.44%) and CEM (62.33-82.79%) compared with MLLMs (DM 31.16-45.12%; CEM 40.00-55.81%). Binary biopsy-candidacy accuracies were likewise higher for radiologists (DM 85.12-89.77%; CEM 86.98-92.09%) than for MLLMs (DM 61.39-75.35%; CEM 69.30-82.79%), although CEM improved performance across all readers. Lesion masks substantially improved MLLM continuous malignancy-probability accuracies from 64.65%-71.63% to 72.56-78.60% on DM and from 67.91%-77.21% to 72.56%-81.86% on CEM, approaching radiologist ranges (DM 63.72-82.79%; CEM 81.86-88.84%). The corresponding AUCs for the top masked models overlapped those of the human readers. Overall, Muse Spark, followed by Sonnet-4.6, demonstrated the strongest performance among the MLLMs across domains. Conclusion: Radiologists generally outperformed MLLMs in categorical tasks, while selected masked models approached human performance for continuous malignancy probability estimation, suggesting a potential adjunctive role.
[CV-79] Floquet Fibre Geometry and Higher-Order Reduced Coordinates for Off-Manifold Transients near Nonlinear Aeroelastic Flutter
链接: https://arxiv.org/abs/2609.14674
作者: Puxue Tan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Assigning reduced coordinates to states near an attracting limit cycle requires the correct invariant-fibre geometry. The classical first-order phase-isostable chart obtained from adjoint Floquet modes projects along the strong-stable quotient fibre, whereas a metric-orthogonal complement of the retained slow bundle generally does not. We prove locally that a chart satisfying the linearised semiconjugacy relation leaves an O(delta^2) invariance residual, while projection along a non-invariant complement generically leaves an O(delta) term. For a nonlinear aeroelastic limit cycle, the metric-normal and strong-stable directions differ by 48.5 to 71.7 degrees, and metric-normal perturbations contain first-order retained phase and slow-amplitude components. Replacing the metric normal by the strong-stable fibre changes the measured residual scaling from delta^1.01 to delta^1.87 without fitted parameters. We then test learned higher-order corrections whose linearisation is pinned to the adjoint-Floquet chart, whose symmetry is exact, and whose reduced flow is fixed. Although they reduce the registered fixed-normalisation latent residual, post-hoc amplitude recalibration and adjoint-Floquet-targeted future consistency move or reverse the ranking. Because the learned maps already share the baseline’s first-order gauge and the future target is supplied by the baseline chart, these diagnostics establish neither an independent positive nor negative higher-order result. Correct first-order Floquet geometry is therefore necessary in this benchmark, while the additional predictive value of the learned correction remains unidentified by the available representation-dependent diagnostics.
[CV-80] Compositional SVG Generation via VLM-Driven Hierarchical Semantic Parsing EMNLP2026
链接: https://arxiv.org/abs/2609.14657
作者: Sehwan Park,Taehoon Kim,Geonhee Han,Dohyun Kim,Seung Wook Kim,Paul Hongsuck Seo
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 26 pages, Accepted to EMNLP 2026 (Main)
Abstract:While Vision-Language Models (VLMs) excel at visual reasoning, generating structured, editable Scalable Vector Graphics (SVG) remains a fundamental challenge. Existing pipelines predominantly yield flat, semantically agnostic collections of paths, where editing a single object requires manually identifying its constituent paths. To address this, we propose a VLM-driven agentic framework for semantic compositional SVG generation. Our pipeline recursively parses visual scenes into semantic and geometric hierarchies via top-down decomposition, visual grounding, and prompt-driven amodal occlusion recovery, ensuring each component is geometrically complete. Furthermore, we introduce the Semantic SVG Benchmark with human-annotated semantic groups and novel sub-component metrics (Semantic Recall/Precision, PERE) to explicitly evaluate structural compositionality and functional editability. Experiments show that our natively predicted structures surpass the upper bounds of existing flat-generation methods in both grouping quality and editability, while maintaining state-of-the-art visual fidelity.
[CV-81] Evaluating Contextual Bias in CNN Image Classification: Evidence from Agricultural Benchmark Datasets
链接: https://arxiv.org/abs/2609.14654
作者: Abhilekha Dalal,Michael Okonoda,Eder Martinez,Lior Shamir
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Convolutional neural networks (CNNs) are typically evaluated using held-out classification accuracy, an approach that presupposes predictions are based primarily on the intended object of interest rather than incidental surrounding context. We test this assumption in CNN-based agricultural image classification by comparing model performance on original images with performance on background-dominated patches extracted from the same images across eight publicly available agricultural benchmark datasets and four widely used CNN architectures. Background-dominated patches were classified above dataset-specific random chance for six of the eight datasets, and substantially above chance for four of them, indicating that contextual information contributes to model predictions for the majority of datasets evaluated. For these four datasets, we further evaluated whether this behavior reflected genuine class-discriminative information or was primarily attributable to class imbalance using macro-averaged precision, recall, and F1 together with class-balanced test subsets. The results show that contextual reliance does not admit a single explanation: class imbalance accounts for a substantial portion of the observed signal for some datasets and architectures, whereas above-chance contextual classification persists after balancing for others. Together with previous evidence from curated object recognition and cancer pathology imaging, these findings support the growing view that contextual bias is a recurring characteristic of CNN-based image classification rather than a phenomenon confined to a single application domain. More broadly, this work provides a systematic framework for quantifying contextual bias across heterogeneous image datasets by combining dataset-specific random-chance baselines, contextual bias categorization, macro-averaged evaluation, and class-balanced robustness analysis.
[CV-82] SCOUT-SLAM: Structurally-Coupled Dual Uncertainty-Aware 3DGS SLAM in the Wild
链接: https://arxiv.org/abs/2609.14634
作者: Kumaran Karthik,Pramat Shastri Jois,Suresh Sundaram
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recently, 3D Gaussian Splatting SLAM (3DGS-SLAM) has gained significant momentum in simultaneous localization and 3DGS scene reconstruction. In real-world scenarios with rapid camera motion and cluttered dynamic environments, existing methods rely on the stability of the underlying scene reconstruction to model uncertainty. This leads to a circular dependency between camera tracking accuracy and reconstruction quality: reconstruction instabilities degrade uncertainty modeling, which affects accurate camera tracking and static scene reconstruction. To address this, the paper proposes SCOUT-SLAM, a structurally-coupled dual-uncertainty framework in which both uncertainties are estimated from a shared base network. A low-rank adaptation of this network, trained on multi-view feature consistency, estimates a tracking uncertainty that does not depend solely on the reconstruction quality. A spatially-adaptive prior modulates the network’s training objective so that reconstruction instability does not inflate uncertainty on static regions, keeping the shared representation intact for both branches. Evaluations on dynamic benchmarks (TUM RGB-D, Bonn Dynamic, Wild-SLAM MoCap) demonstrate that SCOUT-SLAM achieves state-of-the-art camera tracking accuracy and artifact-free static scene reconstruction.
[CV-83] DF: A Two-Stage Framework for Reliable Surgical Phase Transition Detection MICCAI2026
链接: https://arxiv.org/abs/2609.14624
作者: Yushi Guo,Pietro Valdastri,Duygu Sarikaya
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the COLAS Workshop at MICCAI 2026
Abstract:Reliable workflow transition detection is important for context-aware surgical assistance and downstream decision support. However, online surgical phase recognizers primarily focus on frame-wise accuracy and temporal consistency, rather than the reliability of workflow transition events. Directly converting phase changes into events is unreliable: temporal jitter and workflow-illegal switches produce false or duplicate events, while persistent, workflow-consistent candidates may remain incorrect. To address this limitation, we formulate reliable workflow transition detection as a distinct event-level task operating on outputs of a frozen online phase recognizer. We propose the Two-Stage Transition Detection Framework (TTDF), a causal framework that progressively filters transition candidates. Transition Candidate Extraction (TCE) first applies a minimum-duration requirement and a workflow-graph constraint to remove false candidates caused by temporal jitter and phase transitions not allowed by the workflow graph. Specifically, a candidate is retained only if the predicted target phase persists for a minimum duration and the ordered phase pair belongs to the workflow graph’s allowed transition set. TCE thereby produces a high-recall candidate set without additional training. Transition Candidate Verification (TCV) suppresses remaining false candidates using phase-posterior shifts and visual-change cues from frozen DINOv2 features. Events are assessed using a phase-pair-aware one-to-one matching protocol. Experiments on Cholec80 show that TTDF reduces false transition emissions while preserving recall and controlling decision delay.
[CV-84] ESAFusion: LiDAR–4-D Radar Fusion via Local Geometric Complementation and Multiscale Adaptive Interaction for 3-D Object Detection
链接: https://arxiv.org/abs/2609.14619
作者: Gang Ma,Senjie Hu,Junjie Liu,Chao Wang,Hui Wei
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:LiDAR–4-D radar fusion combines accurate spatial geometry with motion and reflectivity cues from radar, offering a promising solution for 3-D object detection in complex driving environments. However, sparse radar observations and differences in spatial sampling between the two modalities complicate reliable cross-modal complementation. Moreover, the relative importance of modalities and feature scales varies across spatial regions, making adaptive fusion challenging. To address these challenges, we propose ESAFusion, an evidence-aware and scale-adaptive framework that combines local geometric complementation with multiscale adaptive interaction. Specifically, we introduce an Evidence-Aware Radar Selection (ERS) module to suppress radar clutter using motion and observation-quality evidence while retaining foreground confidence for subsequent fusion. Then, the Pillar-Level Complementary Encoder (PCE) improves cross-modal complementation under mismatched spatial sampling using local geometric support from neighboring LiDAR pillars. We further design an Intra- and Inter-Scale Adaptive Fusion (ISAF) module to adaptively adjust the contributions of different modalities and feature scales in bird’s-eye-view (BEV) space. Extensive experiments on the View-of-Delft (VoD) dataset show that ESAFusion achieves the highest mean average precision (mAP) among the compared methods, reaching 74.60% in the Entire Annotated Area and 88.89% in the Driving Corridor. It also attains the highest average precision (AP) for Cyclist among these methods in both regions while running at 19.23 FPS. Evaluations on VoD-Fog further demonstrate robustness under progressively degraded LiDAR observations. The source code will be made publicly available at this https URL.
[CV-85] Open-UniMo: Towards Unified Motion-Language Understanding and Generation in the Open World
链接: https://arxiv.org/abs/2609.14615
作者: Guocun Wang,Kenkun Liu,Guorui Song,Jing Lin,Zhe Huang,Luyuan Zhang,Dake Zhong,Choo Sin Wai,Xiaoguang Han,Haoqian Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Unified motion generation and understanding is crucial for embodied AI systems that can both synthesize and interpret human actions in open-world environments. Existing motion-language models often treat motion as an auxiliary modality of a language model, leading to text-dominated representations and limited cross-modal interaction. Moreover, the next-token prediction paradigm is not naturally suited to long motion sequences, where autoregressive generation may accumulate prediction errors. To address these challenges, we propose Open-UniMo, a unified Large Motion-Language Model (LMLM) trained on million-scale open-world motion-language data. Open-UniMo promotes modality parity by extending Qwen’s vocabulary of about 150K text tokens with 64K motion tokens, enabling motion and language to share a unified token space. We further introduce motion-consistent Chain-of-Thought reasoning as an intermediate representation to bridge language semantics and motion dynamics. Open-UniMo is trained with a two-stage pipeline, where supervised fine-tuning establishes CoT-guided bidirectional motion-language mapping and Group Relative Policy Optimization (GRPO) improves semantic alignment while mitigating cumulative errors in autoregressive motion-token generation. To support comprehensive evaluation, we propose Open-MoBench, a VLM-guided benchmark for assessing text-to-motion (T2M) generation, motion-to-text (M2T) understanding, and bidirectional consistency. Extensive experiments show that Open-UniMo achieves state-of-the-art performance on both conventional metrics and Open-MoBench. Furthermore, ablation studies reveal that M2T understanding is not primarily limited by motion-token vocabulary size; instead, coupling M2T with the learnable T2M generation path yields stronger cross-modal representations, demonstrating that generation can facilitate understanding in AR-based motion-language modeling.
[CV-86] Direct Conditional Transition Sampling for Diffusion Inverse Problems
链接: https://arxiv.org/abs/2609.14596
作者: Qi Yu,Hanlin Wu,Xiaohui Sun
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (stat.ML)
备注:
Abstract:Training-free diffusion inverse solvers typically choose between local measurement guidance and costly clean-space posterior updates. Independent posterior refresh can improve global correction by sampling a clean conditional and re-noising it, but its practical realization requires probability-flow ODE integration and clean-space Markov chain Monte Carlo (MCMC). We propose Direct Conditional Transition Sampling (DCTS), a direct stochastic-flow approximation to the same ideal refresh target. Rather than explicitly drawing a clean sample, DCTS estimates the measurement-conditioned clean mean along a short inner path and transports Gaussian source noise directly to the next noisy state. A denoiser-compatible sufficient statistic and a covariance-scaled operator update enable this conditional-mean estimation. Experiments on four inverse problems demonstrate that DCTS achieves competitive reconstruction quality with up to 16.8\times speedups over competing methods.
[CV-87] Reverse Spatio-Temporal Disease Progression Modelling MICCAI2026
链接: https://arxiv.org/abs/2609.14590
作者: Ulugbek Shernazarov,Moucheng Xu,Inomjon Ramatov
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 6 figures. Accepted at SASHIMI 2026, held in conjunction with MICCAI 2026
Abstract:Deep learning-based spatio-temporal disease progression models commonly overlook the incubation period of progressive diseases, limiting the use of those models in early interventions, which are vital for not easily reversible diseases such as Alzheimer’s. This is because, the existing deep learning based longitudinal disease-progression models are almost always run forward: from an observed baseline they predict future decline. In many clinical settings, however, imaging begins only after pathology is suspected or already visible, the earlier, healthier patient-specific reference was never acquired. To address this, we propose to study reverse disease progression prediction: given later diseased anatomy, reconstruct the unobserved healthier anatomy that preceded it. We use a two-stage model in which a frozen 3D vector-quantised autoencoder defines a compact discrete latent space, while a Neural Ordinary Differential Equation (ODE) learns continuous-time dynamics in that space. A recurrent encoder reads late observations in reverse temporal order, initialises the latent state, and the ODE is integrated backwards across the trajectory. On a controlled Morpho-MNIST benchmark with a sinusoidal perturbation, our model successfully recovered the unseen previous states from later observations of the non-monotonic trajectory. On longitudinal brain MRIs from Alzheimer’s Disease Neuroimaging Initiative, at the task to recover the previous unseen trajectory towards healthy states of the patients from observed later diseased states, our model outperforms the baselines that uses copy-nearest and mean-observed, with positive disease-reversal scores in every diagnostic stratum. We hope that our work can provide insights and tools towards discovering the incubation periods from single-shot scans, and developing early interventions of diseases based on imaging.
[CV-88] A Building as a Repository: KIR a Typed Intermediate Representation for Agent -Authored Building Information Models
链接: https://arxiv.org/abs/2609.14578
作者: Dmitry Kuklev
类目: Computer Vision and Pattern Recognition (cs.CV); Software Engineering (cs.SE)
备注: 13-page paper and 10-page supplement; 8 figures, 3 tables. Code: this https URL
Abstract:Autonomous agents that author building information models need more than access to a host API. They need a representation of what they intended, what a compiler decided on their behalf, what was refused, what was observed after execution and what remains unknown. We present KIR, a typed intermediate representation in which a building is authored as a program held in a versioned repository and lowered to host applications as build targets. KIR is organised around seven ways in which a generator writing into a stateful, partially observable host goes silently wrong, and gives each a representation in data: ambiguous selectors become typed refusals with candidates; defaults keep their provenance; obligations that will not be checked are named before execution; vacuous witnesses are rejected statically; a lost transaction response becomes the state UNCONFIRMED with a verify-before-retry rule; the reverse path obeys a census invariant; and decisions are bound by digest to the state they were made against. On a pinned snapshot with 83 operation contracts and Revit as the only backend, offline experiments refuse 29 of 42 stress-test programs with diagnostic codes and no uncaught exception, name 38 of 377 witness obligations as unwitnessable, admit 31 of 100 combinations of execution, witness and acceptance states under seven stated invariants, and find no vacuous witness in 219 certificate runs; a 60-storey tower is 11,263 characters as KIR against 3,709,235 characters of emitted C#. Native Revit runs are reported from project records and kept separate from reproduced results. A controlled comparison with agents that write host code directly is specified but not yet executed; it is the principal open question.
[CV-89] Small Object Detection in Drone Aerial Imagery with LAF-YOLOv10
链接: https://arxiv.org/abs/2609.14560
作者: Quratulain Nayeem,Fahmina Taranum,Mohammed Mudassir Uddin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:General-purpose object detectors lose accuracy on UAV footage, where targets span only a handful of pixels and onboard compute is limited. Prior work composes independently-validated architectural techniques into one detector, assuming gains reported in isolation transfer once combined. We stress-test that assumption directly. LAF-YOLOv10 integrates four techniques into YOLOv10n: a Partial Convolution C2f (PC-C2f) backbone block, an Attention-Guided Feature Pyramid Network (AG-FPN), a P2 detection head replacing the large-object P5 head, and Wise-IoU v3 regression, asking whether their combined effect matches what each contributes alone. We train LAF-YOLOv10 three times (seeds 42, 123, 256) on VisDrone-DET2019, benchmark against unmodified YOLOv10n, and use TIDE error decomposition, per-category breakdown, per-component ablation, attention/loss comparisons, zero-shot transfer to UAVDT, and held-out/test-dev evaluation to localize where the combination succeeds or fails. Composability does not hold here. LAF-YOLOv10 reaches 24.0+/-0.4% mAP@0.5 at 2.14M parameters, 7.8 points below YOLOv10n (31.8%), a deficit that transfers to UAVDT (-10.0 points) and is confirmed by held-out and test-dev evaluation (23.5%, 22.5%). Background false positives, localization error, and duplicate detections move in the direction AG-FPN and Wise-IoU were designed to push. Ablation traces the deficit to a specific source: the P2/-P5 head swap costs 2.5 points independently plus a 2.5-point interaction penalty when layered onto a backbone weakened by PC-C2f, whose own 2.0-point loss is consistent with a partial pretrained-weight transplant (73/150 backbone tensors transfer). The failure is attributable to a specific interaction, not the four components individually. Composability must be verified directly, not assumed. Code/checkpoints: this https URL.
[CV-90] Sharing standardized image-derived data in computational pathology using DICOM
链接: https://arxiv.org/abs/2609.14530
作者: Daniela P. Schacherer(1),Christopher P. Bridge(2, 3),David Clunie(4),Igor Octaviano(5),André Homeyer(1),Markus D. Herrmann(3, 6),Olivier Gevaert(7),Tabita Ghete(8),Markus Metzler(8),Henning Hoefener(1),Tahsin Kurc(9),Curtis Lisle(10),Kenneth Philbrick(11),Joel Saltz(9),Yuanning Zheng(7),Andrey Fedorov(3, 12) ((1) Fraunhofer Institute for Digital Medicine MEVIS, Bremen, Germany, (2) Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Boston, USA, (3) Harvard Medical School, Boston, USA, (4) PixelMed Publishing, Bangor, USA, (5) Radical Imaging LLC, USA, (6) Massachusetts General Hospital, Boston, USA, (7) Stanford University, Stanford, USA, (8) University Hospital Erlangen, Erlangen, Germany, (9) Stony Brook University, Stony Brook, USA, (10) National Institute of Allergy and Infectious Diseases, Bethesda, USA, (11) Google Research, Mountain View, USA, (12) Brigham and Women’s Hospital, Boston, USA)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Daniela P. Schacherer, Christopher P. Bridge: contributed equally
Abstract:Development and evaluation of computational pathology methods require access to large and diverse datasets. Over the past decade, various initiatives invested significantly into collecting, centralizing, and sharing pathology imaging data. In contrast, sharing of image-derived data such as region-of-interest delineations or segmentation masks is less well developed. In this work, we describe our approach to encoding and sharing image-derived pathology data in a standardized manner within the National Cancer Institute (NCI) Imaging Data Commons (IDC), a platform that hosts and provides public access to de-identified radiology and pathology data. The IDC relies on the Digital Imaging and Communications in Medicine (DICOM) standard for data harmonization, yet the adoption of DICOM for pathology image-derived content has remained largely unexplored until now. Here, we present five representative datasets harmonized by conversion from their original representations into DICOM and shared publicly in the IDC. We demonstrate the benefits of this harmonization, describe contributions to critical open-source tooling, and discuss technical considerations relevant to broader adoption of DICOM for pathology image-derived data.
[CV-91] Selective Tool Use for Agent ic Change Visual Question Answering in Remote Sensing
链接: https://arxiv.org/abs/2609.14523
作者: Yakoub Bazi,Mohamad M. Al Rahhal,Mohamed A. Mekhtiche,Mansour Zuair
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Change visual question answering (Change VQA) requires understanding semantic changes across bi-temporal remote sensing images. Although vision language models (VLMs) have shown promising performance on this task, they remain unreliable when answering questions that require explicit transition statistics, area measurements, or spatial information. To address this limitation, we propose a selective tool use framework in which a single VLM either answers directly or invokes a deterministic change analysis tool to obtain question specific evidence. Specifically, the selected tool operates on bi-temporal semantic maps and returns a structured observation, which the same VLM uses to generate its final answer. To support this framework, we construct a tool augmented extension of CDVQA covering eight question families and three tools for transition, spatial, and temporal analysis. Tool use supervision and observations are derived automatically from the original semantic annotations, without additional manual labeling. We then adapt Qwen3.5-4B using Low Rank Adaptation (LoRA) to jointly learn direct answering, tool invocation, and evidence conditioned answering. Experiments on 7,164 test questions show that selective tool use with reference semantic maps improves overall accuracy from 73.77% to 88.79% and average family accuracy from 69.11% to 89.65%. When the semantic maps are predicted automatically, the framework achieves 77.47% overall accuracy and 75.06% average family accuracy. These results demonstrate the benefit of question-specific semantic evidence for Change VQA, while highlighting the influence of semantic prediction quality on the resulting performance. Code and tool-augmented annotations will be made publicly available at this https URL.
[CV-92] CGGT: Curve-Grounded Geometry Transformer for 3D Parametric Curve Reconstruction SIGGRAPH
链接: https://arxiv.org/abs/2609.14521
作者: Zhirui Gao,Renjiao Yi,Yunfan Ye,Ruizhen Hu,Chenyang Zhu,Wei Chen,Kai Xu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by SIGGRAPH Asia 2026
Abstract:Recovering editable 3D parametric curves from 2D images is a fundamental challenge in computer graphics, bridging pixel-based perception and vector-based CAD modeling. Existing NeRF- and 3DGS-based methods often rely on dense calibrated views, precomputed 2D edge maps, and costly per-scene optimization, limiting their applicability to casually captured real-world inputs. We propose CGGT, a Curve-Grounded Geometry Transformer that directly grounds 3D-consistent 2D curve instances in the image space from sparse, unposed multi-view images. CGGT combines a geometry-aware transformer encoder for multi-view feature learning with a curve-aware masked-attention decoder for cross-view instance association. In a single forward pass, it predicts camera parameters, dense depth maps, and instance-level 2D curve masks, which are then lifted into 3D and refined through a fast parametric optimization stage to recover compact, editable 3D curve primitives. To support structured curve learning, we introduce Wireframe-100K, a large-scale dataset comprising 100,000 CAD models with diverse topologies, realistic multi-view renderings, and accurate parametric curve annotations. Extensive experiments show that our framework achieves substantial improvements in both reconstruction accuracy and efficiency, particularly under challenging sparse-view settings and in separating persistent 3D structural edges from view-dependent image edges caused by silhouettes, textures, and appearance variations. Despite being trained solely on synthetic data, CGGT generalizes well to real-world images, demonstrating its potential for practical CAD-style wireframe reconstruction from unconstrained visual inputs.
[CV-93] MCIQA-2K: A Multi-Dimensional Dataset and No-Reference Quality Assessment Benchmark for Colorized Images
链接: https://arxiv.org/abs/2609.14495
作者: Yunkai Zhuang,Qihang Yan,Zicheng Zhang,Guangtao Zhai
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 5 figures, 2 tables
Abstract:Image colorization is an inherently ill-posed task, since a single grayscale image may correspond to multiple plausible colorized results. Consequently, conventional full-reference image quality assessment (IQA) metrics fail to accurately reflect human perceptual preferences for colorized images. In this paper, we present MCIQA-2K, a large-scale multi-dimensional benchmark specifically designed for no-reference quality assessment of colorized images. We construct a dataset containing 2,000 colorized images generated by five representative colorization models, together with human annotations across three perceptual dimensions: color smearing, semantic color misalignment, and global naturalness. Building upon the proposed benchmark, we further introduce MCIQA, a dedicated multi-branch NR-IQA framework for colorized images. Extensive experiments demonstrate that MCIQA significantly outperforms existing full-reference and no-reference IQA methods on the proposed benchmark, while also exhibiting competitive generalization capability on several widely-used IQA datasets. The dataset and code are publicly available at this https URL.
[CV-94] PuzzleMate: Benchmarking MLLM s for Egocentric Puzzle Assistance
链接: https://arxiv.org/abs/2609.14473
作者: Avijit Dasgupta,Shayon Dasgupta,Zakaria Laskar,C. V. Jawahar,Karteek Alahari
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Personal AI assistants hold the potential to evolve from digital interfaces into embodied companions capable of guiding users through complex physical activities. For these assistants to become integral to daily life, they must do more than identify objects; they must provide precise, step-by-step instructions that align with a user’s real-time progress. While Multimodal Large Language Models (MLLMs) show promise in general visual understanding, their ability to deliver grounded, sequential guidance for fine-grained manipulation tasks remains largely unverified. In this paper, we choose the jigsaw puzzle as a strategic testbed for this capability. Unlike general object recognition, puzzle solving demands high-precision spatial reasoning, the ability to distinguish between minute geometric variations, and a rigorous adherence to sequential logic. We investigate this capability through PuzzleMate, a novel framework focused on jigsaw puzzle solving captured through an egocentric viewpoint. We deploy PuzzleMate in a user-in-the-loop study to evaluate how well state-of-the-art MLLMs perceive the current puzzle state and generate actionable next-step instructions. Our analysis reveals seven key bottlenecks that limit their effectiveness. Building on these insights, we propose a benchmark that enables systematic evaluation of MLLMs’ reasoning capabilities for puzzle solving. Our findings reveal a substantial performance gap in current models like GPT-5.2 and Gemini-2.5-Pro; while these MLLMs are highly capable, they struggle to navigate the intricate reasoning and sequential logic essential for jigsaw puzzle assistance. Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.14473 [cs.CV] (or arXiv:2609.14473v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.14473 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-95] PRI-Net: A Lightweight Multimodal Framework for 3D UAV Localization
链接: https://arxiv.org/abs/2609.14469
作者: Zhixuan Chen,Jialiang Lu,Zhong Ye,Yinghui He,Guanding Yu
类目: Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP)
备注: Accepted by IEEE PIMRC 2026
Abstract:Accurate 3D localization of unmanned aerial vehicles (UAVs) remains challenging for existing multimodal approaches due to sparse LiDAR geometry, modality-imbalanced fusion, and redundant feature transmission over constrained edge-to-server links. To address these limitations, we propose PRI-Net, an efficient and lightweight multimodal fusion framework for UAV localization that integrates point cloud splatting, residual attention fusion, and an information bottleneck. Specifically, a 3D point cloud splatting (3DPCS) strategy is introduced to transform sparse LiDAR observations into geometrically consistent dense depth maps. A residual attention fusion (RAF) module is then designed to alleviate modal bias by using an image branch for coarse estimation and a gated fusion branch for refinement. In addition, a multimodal information bottleneck (MIB) module compacts features by filtering task-irrelevant redundancy. Experiments show that PRI-Net achieves high localization accuracy with lightweight architectures, while reducing feature dimensionality and improving edge-to-server UAV sensing efficiency and robustness.
[CV-96] DynEoMT: Learning Object Dynamicity from Online Segmentation Queries WACV
链接: https://arxiv.org/abs/2609.14466
作者: Calvin Galagain,Martyna Poreba,François Goulette
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: Submitted to the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
Abstract:Video segmentation models recognize and track objects over time, but they do not indicate whether each segmented region moves independently of the observing camera. This dynamicity attribute cannot be inferred from semantics alone and is confounded by camera ego-motion. We introduce \method, an online framework that augments query-based video segmentation with region-level dynamicity prediction. It jointly produces the original segmentation outputs and a dynamic or static state for each predicted region. At inference, DynEoMT uses only the current frame and propagated queries, without optical flow, depth, camera pose, previous RGB frames, or feature maps. Because established video segmentation benchmarks do not annotate this attribute, we also introduce a class-agnostic offline supervision pipeline using camera-compensated optical flow and confidence-aware temporal filtering. Across VIPSeg, OVIS, YouTube-VIS 2022, and VSPW, DynEoMT achieves balanced accuracies of 84.3, 68.0, 68.6, and 87.6, respectively, while largely preserving segmentation performance. These results show that segmentation-region dynamicity can be learned from propagated queries, enabling its online prediction without a dedicated motion-processing pipeline at inference. The complete code will be released as open source to enable full reproduction of the method and experiments.
[CV-97] AlayaVista: Streaming World Modeling from Panoramic States to Perspective Video
链接: https://arxiv.org/abs/2609.14462
作者: Jiaming Tan,Mingliang Zhai,Zhen Li,Yuwei Wu,Chuanhao Li,Kaipeng Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Interactive video world models must maintain broad scene context under camera motion while producing high-fidelity observations with low latency. Existing approaches face a representation trade-off: perspective models operate on local views and must preserve off-screen content over long rollouts, whereas broader spatial coverage is typically obtained by synthesizing full-sphere videos or constructing explicit 3D representations. Motivated by the complementary roles of global context and selective local acuity in visual perception, we present AlayaVista, a camera-controllable streaming video world model that decouples panoramic world evolution from perspective observation synthesis. Given a single perspective image, AlayaVista constructs a 360-degree scene prior using a pretrained panorama expansion model and then evolves the scene as a camera-conditioned panoramic latent state. A latent viewport renderer maps this state to the requested perspective video latents, while a perspective refiner restores details, suppresses artifacts, and performs super-resolution. To support efficient streaming, we adapt the panoramic generator to chunk-autoregressive generation and distill both panoramic generation and perspective refinement into few-step processes. To provide the supervision required by this design, we construct MUGEN, a large-scale real-world panoramic video dataset containing 1,318 hours of videos at resolutions of at least 4K, together with rich semantic and geometric annotations.
[CV-98] From Visual Attribution to Clinical Reasoning : Explainable Parkinsons Disease Screening from Hand-Drawn Patterns
链接: https://arxiv.org/abs/2609.14441
作者: Aritra Dey,Utsav Kumar Nareti,Chandranath Adak,Soumi Chattopadhyay,Krishna Gopal Sasmal,Saeed Anwar
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Parkinson’s disease (PD) manifests early neuromotor impairments that become observable in controlled hand-drawn patterns such as spirals and meanders, where tremor-induced oscillations, stroke irregularity, and curvature instability reflect underlying motor degradation. In this work, we present an explainable framework for PD screening from offline hand-drawn patterns that integrates discriminative visual modeling with clinically grounded reasoning. The predictive model captures distributed structural distortions and fine-grained texture variations. It is evaluated under subject-disjoint protocols to ensure reliable generalization. To move beyond black-box classification, we introduce a multi-stage explainability pipeline that combines visual attribution with structured symptom abstraction. Salient regions are identified using attention- and gradient-based localization, followed by extraction of clinically meaningful motor descriptors quantifying contour roughness, curvature irregularity, stroke variability, and tremor-frequency energy. These descriptors are subsequently translated into coherent clinical rationales through a language-based reasoning module, linking model evidence to established PD symptomatology. By bridging visual attribution and clinical interpretation, the proposed framework advances interpretable document intelligence for neurological screening using hand-drawn patterns. Experimental results on publicly available Parkinson’s disease handwriting datasets demonstrate competitive predictive performance and clinically consistent explanations.
[CV-99] Context-Aware Mutual Learning for Blind Image Inpainting and Beyond
链接: https://arxiv.org/abs/2609.14439
作者: Haoru Zhao,Yufeng Wang,Zhaorui Gu,Bing Zheng,Haiyong Zheng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Published in Expert Systems with Applications 268 (2025), Article 126224
Abstract:Blind image inpainting, aiming to recover contaminated images in the case of unknown masks, is a challenging task. Motivated by the perspective of human vision and knowledge, blind image inpainting can be decomposed into two stages: mask estimation and image inpainting based on the estimated mask. The two-stage idea exhibits evident advantages in enhancing inpainting quality and augmenting the generalization capability of unknown real-world contamination by explicitly employing the estimated mask for image inpainting compared to one-stage scheme. This two-stage idea has also been intuitively implemented. However, existing two-stage methods excessively emphasize the unilateral relationship of mask estimation to image inpainting, and may overlook the mutual relations between them. Specifically, mask estimation can provide more contextual semantics for image inpainting to strengthen the understanding of semantics, and image inpainting can offer more contextual details (e.g., textures and edges) for mask estimation to improve the learning of details. In this work, we propose a novel Context-Aware Mutual Learning (CAML) framework for blind image inpainting that joints mask estimation and image inpainting to mutually exploit contextual information. In the CAML framework, we design the Inpainting-Guided Context-Mutual (IGCM) learner to acquire the complementary contextual details from image inpainting for assisting mask estimation, and the Estimation-Guided Context-Mutual (EGCM) learner to strengthen the understanding of contextual semantics from mask estimation for assisting image inpainting. Ablation studies validate the efficacy of our CAML. Extensive experiments show that our CAML achieves state-of-the-art performance on both blind image inpainting and additional vision tasks, i.e., snow removal, shadow removal, and watermark removal, indicating its superiority.
[CV-100] Lightweight Generalized DeepFake Face Detection with WAVIE: Wavelet Augmented Vision Intermediate Embeddings
链接: https://arxiv.org/abs/2609.14437
作者: Arya Pulkit,Aditya Ruhela,Akarshan Kapoor,Arnav Bhavsar
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at the IEEE International Conference on Systems, Man, and Cybernetics (SMC) 2026
Abstract:Deepfake detection systems often exhibit significant performance degradation when deployed on unseen manipulation methods, limiting their reliability in real-world multimedia environments. This lack of generalization poses critical challenges for misinformation mitigation, digital forensics, and human-centric AI systems. Existing detectors perform well on the forgery methods they are trained on, but their accuracy drops sharply on unseen pipelines. To bridge this generalization gap, we propose WAVIE (Wavelet Augmented Vision Intermediate Embeddings), an end-to-end architecture that combines complementary spatial and frequency cues on top of a frozen CLIP backbone. WAVIE projects intermediate transformer embeddings through a lightweight learnable module, applies a three-level Daubechies-6 (db6) discrete wavelet transform (DWT), refines the low-frequency branch while preserving the high-frequency branch, reconstructs the feature via inverse DWT, and performs classification. Trained only on FaceForensics++, WAVIE achieves AUROC = 0.852 on Celeb-DF-v1, 0.852 on Celeb-DF-v2 and 0.831 on WildDeepFake (WDF) at the frame level, outperforming several state-of-the-art generalization baselines. Extensive ablation studies confirm the importance of both the wavelet module and the intermediate-feature aggregation for cross-dataset performance, highlighting the necessity of jointly leveraging spatial and frequency domains. These results position WAVIE as a strong baseline for deepfake detection in the wild. Comments: Accepted at the IEEE International Conference on Systems, Man, and Cybernetics (SMC) 2026 Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14437 [cs.CV] (or arXiv:2609.14437v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.14437 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-101] A Generative AI Integrated Multimodal Framework for Low-Latency Multi-Camera Person Re-Identification
链接: https://arxiv.org/abs/2609.14419
作者: Leon Fernando,C Dombawala,P. Hettigoda,Vanodhya G. Warnasooriya,Ishara Neranjana,Rashmika Nawaratne
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Image and Video Processing (eess.IV)
备注: Presented at 2026 IEEE International Conference on Responsible Artificial Intelligence (IRAI)
Abstract:Person re-identification (ReID) is essential for multi-camera surveillance and tracking, yet remains difficult due to viewpoint and illumination changes, occlusion, background clutter, and low resolution imagery. We propose a generative AI integrated multimodal ReID framework designed explicitly for robustness under missing cues and low latency deployment. The key idea is a cost aware early-exit cascade that prioritizes inexpensive, high confidence evidence and only triggers expensive modalities for ambiguous cases. Our system integrates (i) global visual embeddings from segmented person regions, (ii) automatically generated fine grained semantic attribute descriptions generated by vision-language models (VLMs), and (iii) optional facial embeddings when face observations are reliable. To optimize the balance between accuracy and latency, we use a cost aware early-exit cascade instead of fusing all modalities. Specifically, we first inspect the top-k retrieval results to determine whether the query is unambiguous. If the best match is clearly separated from the remaining candidates, we stop early and return the result to minimize latency; in ambiguous cases, we keep multiple hypotheses and invoke additional modalities (face/semantic) with adaptive reliability weighting to refine the decision. We report person re-identification performance using mAP and Rank-1 accuracy on the Market-1501 and DukeMTMC-reID benchmarks. The proposed adaptive early-exit cascade resolves 60.7% of DukeMTMC-reID queries and 68.4% of Market-1501 queries without invoking semantic reasoning, reducing computational overhead while maintaining competitive retrieval performance. Comments: Presented at 2026 IEEE International Conference on Responsible Artificial Intelligence (IRAI) Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Image and Video Processing (eess.IV) Cite as: arXiv:2609.14419 [cs.CV] (or arXiv:2609.14419v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.14419 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-102] Newton Deep Unfolding for Compressed Sensing
链接: https://arxiv.org/abs/2609.14391
作者: Changhua He,Xianchao Xiu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Compressed sensing (CS) reconstructs images from highly limited measurements, but existing deep unfolding methods are typically driven by first-order optimization and weakly exploit the optimization states generated during reconstruction. To address these limitations, we propose a Newton deep unfolding network (NDU-Net), which, to the best of our knowledge, is the first deep unfolding framework that leverages second-order optimization for CS reconstruction. Specifically, NDU-Net introduces a Newton update (NU) module to estimate Newton-type update directions and generate optimization states that characterize the current reconstruction process. Furthermore, a Newton-guided multi-scale prior (MP) module is designed to incorporate these optimization states into multi-scale feature restoration, thereby enabling the learned prior to adapt to the current reconstruction stage. Experimental results under different CS ratios confirm that our proposed NDU-Net achieves promising reconstruction performance and exhibits enhanced robustness. Our code is available at this https URL.
[CV-103] Contour-Guided Spectral Routing for Robust Real-Time Pedestrian Detection
链接: https://arxiv.org/abs/2609.14383
作者: Sam Williams,Yuan Xiang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Real-time pedestrian detection in driving scenes is constrained by three coupled failure modes: tiny targets lose discriminative evidence, occlusion weakens geometric support, and weather or illumination changes distort appearance statistics. We formulate the detector through a unified \emphcontour-guided spectral routing view rather than treating frequency processing, attention, and boundary reasoning as independent add-ons. The detector routes information in a prescribed order: spatial evidence is first augmented with global spectral context, deep representations then exchange spatial and spectral cues, and cross-scale fusion is finally conditioned on boundary–semantic disagreement. This ordering yields a compact representation pipeline in which low-frequency context stabilizes global structure while high-frequency evidence protects small-object contours. We further retain a wavelet-subband training transformation that perturbs low- and high-frequency coefficients independently, targeting appearance shifts caused by fog, rain, snow, and low illumination. The formulation exposes a single routing variable at each stage and distinguishes reusable signal transforms from the task-specific policy that decides where each signal is injected. On CityPersons, the proposed detector obtains 70.4 AP _50 and 44.2 AP _50:95 , compared with 68.1 and 42.2 for RT-DETR, while the full wavelet-augmented configuration reaches 71.1 and 44.6.
[CV-104] Rethinking Camouflage Image Generation towards a Training-Free Paradigm
链接: https://arxiv.org/abs/2609.14377
作者: Haodong Yang,Zhongling Huang,Gong Cheng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 21 pages, 17 figures
Abstract:Camouflage image generation (CIG) aims to synthesize realistic camouflaged images by blending foreground objects into concealment-compatible background contexts. Achieving this objective requires jointly satisfying three coupled requirements: foreground preservation to retain target integrity, semantic compatibility to select plausible concealment contexts, and appearance assimilation to reduce visual discrepancies. Recent approaches predominantly rely on task-specific training on camouflage datasets to address these requirements, incurring substantial computational cost and limiting generalization beyond the training domain. To address these limitations, we formulate training-free CIG as a concealment-oriented paradigm that preserves the target while reducing its perceptual separability from the synthesized surroundings, rather than maintaining its visual prominence, without parameter updates. We instantiate this paradigm with FreeCam based on a frozen inpainting diffusion framework to preserve the foreground. Within this framework, a Contextual Reasoning Module exploits frozen multimodal priors to infer an environment favorable to concealment, thereby promoting semantic compatibility, while an Intrinsic Appearance Module extracts low-level color and texture cues from the foreground to guide background synthesis toward appearance assimilation. Extensive experiments demonstrate that FreeCam achieves state-of-the-art generation quality and camouflage effectiveness without task-specific training, while its generated images provide synthetic supervision for camouflaged object detection and reduce target detectability under general object detectors.
[CV-105] Beyond Natural Images: Rethinking AI-Generated Image Detection in Documents
链接: https://arxiv.org/abs/2609.14352
作者: Zhangjie Fu,Jiazhen Yan,Yuanwen Chen,Xinquan Yu,Yanzhe Li,Hui Jiang,Lei Gao,Chenfu Bao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:AI-generated image detection has attracted increasing attention, but existing evaluations mainly focus on natural images, leaving AI-generated document images largely underexplored. This omission is concerning because documents often appear in sensitive real-world scenarios, such as invoices, expense reports, certificates, and medical records. In this paper, we first construct a controlled diagnostic benchmark, AIGDoc-Pilot, and reveal that existing detectors suffer substantial performance degradation on AI-generated document images, with the mean AUC dropping by more than 7%. Based on this, we further reveal two document-specific properties behind this gap: generation artifacts exhibit strong spatial inconsistency across local regions, and text density significantly affects real-synthetic separability, where text-dense regions offer stronger discriminative evidence. Motivated by these findings, we construct AIGDoc, a larger document-centric dataset containing diverse real-world documents and AI-generated counterparts produced by multiple advanced generation and editing models. Extensive experiments on AIGDoc demonstrate that existing detectors still struggle to reliably identify AI-generated documents, while document-based training partially narrows the gap. Together, these results offer valuable insights for developing dependable and generalizable detectors in document-centric scenarios. The code and datasets will be made publicly available upon acceptance of the paper.
[CV-106] wo-Stage Mixture-of-LoRA for Multi-Task Medical Vision-Language Learning MICCAI
链接: https://arxiv.org/abs/2609.14350
作者: Zhanghao Chen,Yuanyuan Li,Zhenyu Lu,Shuo Gao,Guangquan Zhou,Yikun Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: MICCAI FLARE26 Task3 Challenge Report, 18 pages, 6 figures
Abstract:Medical vision-language models (VLMs) allow a single model to perform clinical image analysis tasks ranging from diagnosis classification to report generation. However, joint adaptation is challenged by heterogeneous output formats, conflicting task gradients, and imbalanced training data. Hence, we present \textbfTwo-Stage Mixture-of-LoRA, a framework built on MedGemma-1.5-4B. The framework uses a shared-specific Mixture-of-LoRA architecture comprising one shared LoRA and six task-specific expert LoRAs, together with a two-stage training procedure. In Stage 1, we jointly train the shared LoRA and all task-specific expert LoRAs on all tasks. In Stage 2, we first freeze the backbone, the shared LoRA, and all non-target experts, and refine one task expert at a time. Classification and regression then receive an additional modality-balanced continuation, in which smaller modality groups are repeated to match the largest group. In the FLARE 2026 Task 3 test sets, the proposed method achieves 0.85 balanced accuracy for classification, 0.48 micro-F1 for multi-label classification, 0.79 detection F1, and 17.39 regression MAE. Code is available at this https URL.
[CV-107] Mobile CT Services for Rural Regional and Remote Areas: Current Practice and Future Integration with Telehealth and Regulatory-Authorised AI
链接: https://arxiv.org/abs/2609.14347
作者: Zhicheng Lu,Md Zahid Islam,M Mamun Huda,Kristie Sweeney,Shayne Chau,Oliver Mulcock,Corey Hemopo,Catherine Keniry,Mohammad Ali Moni
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 3 figures, 4 tables
Abstract:Computed tomography (CT) plays an essential role in clinical workflow to improve patient outcomes. However, access to CT imaging and specialist interpretation remains limited, particularly in rural, regional, remote (RRR), and other resource-limited settings. Recent advances in mobile CT, telehealth, and artificial intelligence (AI) provide opportunities to extend advanced imaging services to populations in RRR settings. This review examines: 1) mobile CT systems deployed in trucks, trailers, ambulances, and other mobile platforms; 2) telehealth technologies supporting CT-based healthcare; and 3) AI for CT that has received regulatory authorisation or is currently deployed in clinical practice. Applications are evaluated across four clinical functions: screening and diagnosis, patient monitoring, risk prediction, and intervention or therapeutic decision support. The review covers neurological, thoracic, cardiovascular, abdominal, oncological, musculoskeletal, and interventional imaging, with particular attention to stroke, cancer, and other image-guided treatment. Other factors such as regulatory status, deployment status, and estimated technology readiness (TRL) level are compared. Current evidence indicates that mobile CT, telehealth, and AI for conventional CT are individually relatively mature, but fully integration of these technologies remains less widely deployed and validated in the clinical settings. Key barriers include regulatory variation, domain shift, connectivity requirements, cost, workflow integration, cybersecurity, and limited evidence of patient-level benefit. Future research should prioritise prospective, multicentre evaluation of integrated CT systems in real-world and underserved clinical settings.
[CV-108] Learning Continuous Source Responses For Generalizable AI-Generated Image Detection
链接: https://arxiv.org/abs/2609.14316
作者: Manni Cui,Ruiqi Liu,Zijian Yu,Hao Tan,Zibo Wei,Zian Wang,Ziheng Qin,Huijia Zhu,Weiqiang Wang,Jun Lan,Shu Wu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Advances in image generation have made synthetic images increasingly difficult to distinguish from real photographs, raising concerns about the trustworthiness of visual media. Existing AI-generated image detectors often perform well on in-domain data, but their robustness and cross-generator generalization remain limited. These limitations are commonly attributed to overfitting to shortcut cues. Although many methods seek to suppress shortcut learning, most retain binary classification as the training task without reconsidering how the task itself shapes the learned representations. We introduce CuRe, a framework for learning Continuous Source Responses that revisits authenticity detection from the perspective of the training task. CuRe reformulates backbone adaptation as regression of real-generated mixing ratios, providing finer supervision that encourages the model to capture authenticity-related variation beyond binary endpoint separation. We further select a compact source-response subspace to suppress nuisance variation and limit the final classifier’s access to potential shortcut cues. Across ten public benchmarks, CuRe achieves an average balanced accuracy of 89.7%, exceeding the second-best method by 5.2 percentage points. Further experiments demonstrate consistent generalization gains across visual backbones and strong robustness to common image degradations. Code is available at this https URL
[CV-109] S3-Tracker: Self-Supervised Surgical Tissue Tracking With Contrastive Random Walks
链接: https://arxiv.org/abs/2609.14313
作者: Jiaming Zhang,Zijian Wu,Mehran Armand,Septimiu Salcudean
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注:
Abstract:Robust point tracking in endoscopic videos is essential for computer-assisted intervention and autonomous robotic surgery, enabling continuous registration between intraoperative video and preoperative imaging despite soft tissue deformation. However, supervised tracking methods depend on large annotated datasets, while surgical conditions make reliable trajectory annotation challenging. We propose a self-supervised Track-Any-Point approach that learns from unlabeled surgical videos by establishing global pixel correspondences and inferring point trajectories through contrastive random walks. Trained without annotations, our method achieves performance comparable to existing semi-supervised approaches while implicitly handling tissue deformation. These findings demonstrate the feasibility of self-supervised point tracking in surgical environments and its potential to reduce reliance on annotated data.
[CV-110] DTI-Guided Volumetric Spherical Harmonics Regression for Single-to-Multi-Shell dMRI Synthesis MICCAI2026
链接: https://arxiv.org/abs/2609.14312
作者: Binghua Li,Christina Andica,Tong Liang,Ziqing Chang,Chao Li,Wataru Uchida,Kaito Takabayashi,Qibin Zhao,Toshihisa Tanaka,Zhe Sun,Shigeki Aoki
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 4 figures. Accepted at MICCAI 2026
Abstract:Multi-shell diffusion MRI (dMRI) unlocks more expressive microstructural modeling than single-shell scans, yet its longer acquisition time hinders deployment in large-scale cohorts and time-constrained clinical settings. Synthesizing an unobserved shell from a single-shell input is fundamentally ill-posed and further complicated by protocol mismatch, where source and target gradient direction sets may not align. We propose DTI-SHNet, a single-to-multi-shell synthesis framework that operates in the real symmetric spherical harmonics (SH) coefficient domain and performs spatially aware volumetric regression. Given a source shell, we estimate diffusion tensor imaging (DTI) and use direction-agnostic parametric maps along with a brain mask as conditioning priors to guide a 3D U-Net regressor from source-shell to target-shell SH coefficients. To couple coefficient accuracy with signal fidelity, we introduce a signal consistency regularization that reconstructs signals on randomly sampled canonical directions from predicted coefficients and enforces agreement in the signal domain. Experiments on UK Biobank and Cam-CAN data for b=1000 to b=2000 dMRI synthesis show that DTI-SHNet achieves competitive visual quality compared to advanced methods, while better preserving downstream diffusion measures. Our code is available at this https URL.
[CV-111] Robust low-rank tensor completion via factorized weighted tensor schatten-p norm minimization
链接: https://arxiv.org/abs/2609.14307
作者: Binghao Wang,Feng Zhang,Wendong Wang,Jianjun Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Optimization and Control (math.OC)
备注:
Abstract:Low-rank tensor factorization provides a flexible framework for completing multidimensional data from incomplete and corrupted observations. However, unweighted spectral regularizers impose a common shrinkage profile across singular components, which may excessively attenuate dominant low-rank components, and factorized variants either lack component-specific weighting or require costly singular value decompositions (SVDs). This paper proposes two weighted Schatten- p tensor factorization models, termed \WSpTFI and \WSpTFII, under the tensor-tensor product (t-product) framework to address these limitations. \WSpTFI is motivated by a factorized weighted tensor Schatten- p norm identity and permits flexible, possibly asymmetric factor exponents. \WSpTFII constructs a regularizer from transform-domain column-pair energies, yielding SVD-free main factor updates and a column-pruning mechanism for reducing redundant rank components. This paper further develops an iteratively reweighted alternating direction method of multipliers (ADMM)-type scheme for \WSpTFI and an iteratively reweighted least squares (IRLS)–block successive upper-bound minimization (BSUM) scheme for \WSpTFII. Theoretical analysis establishes the weighted factorization relation and provides a conditional limiting Karush–Kuhn–Tucker (KKT) characterization for \WSpTFI. For \WSpTFII, the actual damped quadratic block updates yield a quantitative sufficient-decrease mechanism for the fixed- \delta smoothed factor objective. This implies asymptotic regularity, and every accumulation point of the fixed-dimensional tail is stationary. Experiments on synthetic tensor completion, color-image restoration, hyperspectral inpainting, and printed-circuit-board defect detection demonstrate competitive reconstruction quality and robustness under various degradation conditions.
[CV-112] Vision-Language Models for Criterion-Level Grading of Handwritten Examinations in Outcome-Based Education
链接: https://arxiv.org/abs/2609.14284
作者: Md Khalid Syfullah,Asif Hasan Tonmoy,Saad Ahmed,S. M. Jahangir Alam
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Criterion-level grading connects examination performance to learning outcomes, but manual marking introduces workload and variation between markers. This study evaluates vision-language models (VLMs) for handwritten outcome-based assessment across five dimensions: accuracy, human agreement, repeated-run reliability, error concentration, and explanation quality. Using 1,982 criterion-level records from 485 undergraduate examination answers, we compare 20 configurations spanning Qwen2.5-VL, InternVL3, Pixtral, a Donut baseline, and a cascade ensemble. Evaluation setups include zero-shot prompting, few-shot prompting, partial fine-tuning, and Low-Rank Adaptation (LoRA). Two independent faculty markers regraded all 291 test criteria, providing a human agreement baseline on the same assessment materials. Qwen2.5-VL with LoRA achieved Quadratic Weighted Kappa (QWK) of 0.727 and mean absolute error of 0.435 marks against the examiner, compared with mean human-pair QWK of 0.551. This comparison reflects calibration to the examiner’s training marks. LoRA outperformed partial fine-tuning for all three instruction-tuned VLMs, while few-shot prompting reduced QWK in every configuration with valid prompted scores. Aggregate reliability and exact repeatability diverged: intraclass correlations ranged from 0.790 to 0.874, yet 50.2-63.6% of criteria changed marks across five sampled runs. Attention-guided deletion showed no statistically significant advantage over random masking, and four faculty reviewers reached no consensus on explanation usefulness. These findings highlight the need for rubric-specific calibration, repeatable scoring, review of consequential errors, and separate validation of explanations. The released evaluation protocol supports criterion-level assessment research and grading tools with teacher oversight.
[CV-113] SpermYOLO: A Coordinated YOLO-Based Detector for Accurate and Efficient Sperm and Impurity Detection in Microscopic Images
链接: https://arxiv.org/abs/2609.14278
作者: Shengqi Chen,Zilin Wang,Xingyu Pan,Wenting Yu,Pengchao Deng,Guohua Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 51 pages, 10 figures
Abstract:Accurate sperm detection is essential for computer-assisted semen analysis, yet it remains challenging in microscopic images due to dense distributions, visually similar artifacts, and sperm-like impurities. In this paper, we propose SpermYOLO, a coordinated and compact YOLOv11-derived framework for joint sperm and impurity detection in microscopic images. SpermYOLO introduces four architectural improvements: C3k2-IDB for channel-wise discriminative feature extraction, D2SEM for spatial–spectral semantic enhancement, MFM for adaptive multi-scale feature fusion, and the DESD Head for detail-enhanced shared prediction. Experiments on the SVIA semen microscopic imaging benchmark show that SpermYOLO achieves 97.2% sperm AP and 75.4% impurity AP, outperforming generic detectors, dedicated sperm detection models, and improved YOLO variants. Compared with the baseline model, SpermYOLO improves sperm AP, impurity AP, \mathrmmAP_50 , and \mathrmmAP_50:95 by 1.6, 10.0, 5.8, and 2.7 percentage points, respectively, while preserving a lightweight model scale. Cross-scene evaluation on the SDTB testicular-biopsy microscopy benchmark shows that SpermYOLO remains effective with extremely small sperm targets and complex tissue backgrounds, achieving the highest \mathrmmAP_50 and \mathrmmAP_50:95 of 74.8% and 31.2%, respectively. Ablation studies and qualitative analyses further support these improvements by demonstrating the contributions of the proposed modules and showing more focused feature response patterns than the baseline model. These findings suggest that SpermYOLO is an effective and efficient approach for sperm detection in challenging microscopic imaging scenarios.
[CV-114] Sparsity-Adaptive Sharpness-Aware Minimization IJCNN2026
链接: https://arxiv.org/abs/2609.14274
作者: Shiryu Ueno,Yoshikazu Hayashi,Kunihito Kato
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: IJCNN 2026
Abstract:Deploying deep neural networks in real-world settings requires models that are both compact and robust to common corruptions. However, at deployment-relevant high sparsity, standard pruning pipelines often degrade corruption robustness, and existing sharpness-aware training/pruning approaches provide limited robustness gains. We address this issue by introducing Sparsity-Adaptive Sharpness-Aware Minimization (SA-SAM), which derives a sparsity-dependent SAM/ASAM perturbation radius by keeping the mean absolute perturbation (an \ell_1 -based proxy) approximately invariant as sparsity increases. As a simple complementary option, we evaluate Magnitude-Weighted Hessian (MWH), derived from a second-order removal-path analysis, yielding an importance proportional to \mathrmDiag(F)_i,|w_i| , where \mathrmDiag(F) is the diagonal empirical Fisher used as a curvature proxy in our implementation. Across CIFAR-10-C, CIFAR-100-C, and ImageNet-100-C, our approach achieved stronger corruption robustness than the considered pruning baselines at 80–90% sparsity, while preserving clean accuracy. We additionally quantify the robustness–throughput trade-off by reporting measured inference throughput under sparse execution at deployment-relevant sparsity levels.
[CV-115] SPARK: Representation-Level KV Memory Alignment for Safer Vision-Language Models
链接: https://arxiv.org/abs/2609.14258
作者: Mohd Azfar,Izhar Dad Khan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 23 pages (including appendix), 5 figures. Preprint
Abstract:Vision-language models (VLMs) remain vulnerable to jailbreaks that distribute harmful intent across text and images, making unimodal safety mechanisms insufficient. We investigate whether this vulnerability can be mitigated directly in the multimodal key-value (KV) memory formed during prefill, without modifying model parameters at inference time. We introduce SPARK, a two-stage framework for targeted KV-memory repair. Stage 1 uses a disposable diagnostic adapter to identify harm-associated directions in multimodal key and value representations. Stage 2 projects out these directions, learns a lightweight residual repair, and anchors repaired keys with an image-structural prior to preserve visual grounding. Rather than applying the intervention uniformly, SPARK mixes repaired and original memory using a head-wise coefficient g_h* determined by intervention-relevant subspace energy E_h, requiring no explicit harm classifier at inference. Across LLaVA-OneVision-7B, Chameleon-7B, Qwen2-VL-7B, and InternVL2-4B, SPARK reduces multimodal attack success while preserving general capability. On LLaVA-OneVision-7B, image-only jailbreak attack success falls to 4.7%, while MMMU remains within 0.6 points of the undefended model (47.8 vs. 48.4) with near-baseline language quality. On MM-SafetyBench, attack success decreases from 39.2% to 12.4%. Even under white-box adaptive joint prompt-image attacks, attack success is limited to 20.3%, compared with 54.6% for the undefended model. These results suggest that multimodal jailbreak behavior can be substantially mitigated by selectively repairing intervention-relevant KV subspaces at prefill, particularly when harmful evidence is carried by the visual modality. Comments: 23 pages (including appendix), 5 figures. Preprint Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.14258 [cs.CV] (or arXiv:2609.14258v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.14258 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-116] What Input Resolution Is Required for Bird Species Identification and What Is Its Latency Cost on an Edge Device? A Study of 14 Input Resolutions and Six Architectures with On-Device Measurements
链接: https://arxiv.org/abs/2609.14247
作者: Takeshi Nishikawa
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
备注: 25 pages, 7 figures, 7 tables. On-device measurements on NVIDIA Jetson Orin Nano (JetPack 6.2, TensorRT 10.3). Data and code: this https URL
Abstract:Bird-strike mitigation at wind farms requires identifying distant birds that span only tens of pixels, so the classifier’s input resolution N is a design variable, not a fixed specification. We study it with a factorial design over 14 side lengths N (16 to 224), six architectures, two training and evaluation regimes and 30 random seeds – 2,520 checkpoints and 5,040 evaluations – plus latency measured on an NVIDIA Jetson Orin Nano. Four results. (1) The selected N depends on the target: 0.90 is met on validation by ResNet50 at N=112 in an estimated 1.85 ms (0.8980 on test) and 0.95 by DINOv2-L at N=144 in 12.70 ms; changing the model buys more accuracy than raising N (+5.93 versus +2.33 points at N=112). (2) The benefit of lowering N depends on the assumed preprocessing path: N=224 - 80 saves 13.5% when each individual is decoded from its own file but 46.2% when the detector decodes the 4K frame once; the Pareto set grows from 20 to 23 configurations. (3) Accuracy must be measured on the deployed engine: half precision costs ViT-S/16 alone 4 to 7 points at N=96 while the CNNs stay within 0.1 points, and with selection held at validation the choice differs at 26 of 176 targets. A broken FP16 engine can run faster than a correct one, undetectable from latency; admitting 14 ViT-S/16 FP32 configurations moves the recommendation over the 0.931-0.938 band and under the 10 ms budget. (4) ViT-L-scale models fit this device, but activations exceed the FP16 range; splitting the graph at transformer-block boundaries confines FP32 to the affected segments, making the deployed DINOv2-L chain 1.85x faster than the single-engine build. We also quantify how the regime-difference sign stabilises with seed count; a sensitivity split removing some forms of group sharing preserves all 14 non-trivial signs at the selection boundary.
[CV-117] ask-Specified Active Metrological Inspection with Measurement-Steered VLA Manipulation and Deterministic Evidence Gating
链接: https://arxiv.org/abs/2609.14219
作者: Zhiling Chen,Jingzhan Ge,Ruimin Chen,Matthew P. Castanier,David Gorsich,Farhad Imani
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: 19 pages, 13 figures
Abstract:High-mix low-volume (HMLV) manufacturing requires inspection systems to adapt to changing parts, specifications, and work orders without repeated task-specific programming. Existing inspection automation typically assumes predefined sensing sequences, while general purpose robot agents optimize task completion rather than the completeness and validity of metrological evidence. We formulate task-specified active metrological inspection and propose From Requirements to Admissible Metrological Evidence (FRAME), a hierarchical dual-arm framework that converts an inspection instruction and structured specification into traceable conformance evidence. FRAME coordinates learned manipulation with calibrated laser profilometry: a task manager grounds and schedules requirements, active surface correspondence verifies physical-to-specification localization, and evidence memory tracks measurement provenance, admissibility, and coverage. Learned components may propose inspection targets and physical access actions, but deterministic datum-grounded measurement, admissibility checks, coverage auditing, and conformance evaluation prevent incomplete or unverified evidence from authorizing PASS. A series of physical experiments shows that FRAME achieves higher end-to-end inspection reliability, fewer false accepts, and shorter task completion time.
[CV-118] MorphoStyle: Motion Style Transfer with Morphology Control BMVC2026
链接: https://arxiv.org/abs/2609.14189
作者: Xin Feng,Eleonora D’Arnese,Mohan Sridharan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by BMVC2026
Abstract:Human motion may be viewed as a combination of action content, style, and body morphology. Existing motion style transfer methods transfer a reference style onto a content motion while assuming a canonical body, whereas shape-aware motion generators adapt motion to a target shape without explicit style control. This separation of motion style and shape (morphology) makes it difficult to generate stylized motions for non-canonical bodies; naively combining a style transfer module with a shape-aware generator often leaks action content from the style reference and disrupts shape-consistent kinematics. We present MorphoStyle, a framework for shape-aware motion style transfer that is built on a shape-conditioned Finite-Scalar-Quantization Variational Auto-Encoder (FSQ-VAE). The key contribution is to pose the desired style transfer as modular latent disentanglement comprising: (i) a contrastive style encoder that extracts content-decoupled style embeddings; (ii) a text-guided style-routing mechanism that locates style-relevant joints in a text-motion feature space; and (iii) a manifold preserving style modulator that injects discriminative style embeddings in content features as a temporally-gated low-rank offset. Extensive experiments on benchmark datasets demonstrate that MorphoStyle outperforms competing baselines in terms of both shape control and motion style transfer, while simultaneously providing quantitative shape control. For more details, please see project website: this https URL.
[CV-119] Bi-Level Routing and Sparse Spatial Attention based Multi-View BEV 3D Object Detection for Autonomous Driving
链接: https://arxiv.org/abs/2609.14185
作者: Jing Zhang,Jiaqi Liu,Zibo Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to 2026 9th International Conference on Computer Information Science and Application Technology (CISAT). 4 pages, 1 figure
Abstract:Bird’s Eye View (BEV)-based multi-view 3D object detection suffers from challenges of computational complexity, multi-scale feature extraction, and efficiency of dense 2D-to-BEV view transformation. To address these problems, this paper proposes an improved BEV 3D object detection algorithm Sparse-BEVNet. Firstly, a Bi-Level Routing Attention (BRA) mechanism is introduced into the image feature extraction network to reduce the computational burden of the backbone. Second, Cascaded Group Attention (CGA) is employed in the feature fusion module, which enhances deep interaction across features of different hierarchical levels without introducing additional computational overhead. Furthermore, a Sparse Spatial Cross-Attention mechanism is adopted to replace the conventional dense view projection pipeline. Experimental results on the public nuScenes dataset demonstrate that the proposed method achieves a mean Average Precision (mAP) of 45.2% and a nuScenes Detection Score (NDS) of 54.5%, corresponding to 3.6% and 2.8% improvements relative to the baseline model, respectively.
[CV-120] 3D Gait-Based Autism Classification Using Attention-Enhanced Deep Learning with Cross-Fold Statistical Stability Analysis
链接: https://arxiv.org/abs/2609.14159
作者: Md Nadim Mahamood,Md Arif Shahriar,Md Parvej Sikder,Md Rasul Islam,Md Shafi Ud Doula,Md Ashraful Alam,Kamrul Hasan
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 15 pages, 4 figures, 8 tables
Abstract:Autism Spectrum Disorder (ASD) is a neurodevelopmental condition whose early diagnosis remains challenging because conventional clinical assessments are often subjective, time-consuming, and require expert evaluation. Gait provides a promising non-invasive behavioral biomarker for auto- mated ASD screening; however, existing studies have primarily relied on single-dataset evaluations, convolutional architectures, and descriptive summaries of cross-validation performance without formally assessing fold-to-fold stability. This study addresses these gaps with an attention-enhanced Transformer framework for ASD classification, evaluated on two structurally different 3D gait feature representations: precomputed statistical gait descriptors and raw biomechanical ground-reaction- force measurements. Under five-fold cross-validation, the proposed framework achieved 99.00% accuracy, 99.02% precision, 99.00% recall, 99.00% F1-score, and 99.00% specificity on the public Kinect-based benchmark, exceeding the performance of the compared state-of-the-art methods. On the independent private force-plate dataset, it achieved mean values of 95.00% accuracy, 93.81% precision, 96.67% recall, 95.13% F1-score, and 93.33% specificity.
[CV-121] SignMimic: Robust High-Quality Sign Language Motion Generation via Human-Shape-Oblivious Pose Transfer Guidance
链接: https://arxiv.org/abs/2609.14122
作者: Zhewen He(1),Junyi Yu(1),Haomian Huang(1),Zhenhua Li(2),Yi Fang(1 and 2) ((1) New York University Abu Dhabi, (2) ChatSign Technology)
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We study the challenge of sign language video mimicking: given a driving video and a single reference frame, synthesize a video where the target signer reproduces the source motion while preserving identity and linguistic form. Prior pipelines entangle rigid motion, non-rigid deformation, and view-dependent completion in a monolithic generator, causing handshape drift and spatio-temporal instability. We present SignMimic, which (i) applies a TNet-based model to study SE(3) rigid canonicalization to stabilize global pose, (ii) performs non-rigid adaptation in a canonical space to preserve fine-grained articulators (hands/face) and coarticulation via NIF2D, and (iii) uses Pose-MAE-style completion before conditional video diffusion. This factorization injects geometric and linguistic priors, yielding shape and spatio-temporal consistency. On several large-scale datasets (ASL 50K, How2Sign, CSL News), SignMimic achieves state-of-the-art-level performance on video quality, identity similarity, and frame continuity while also achieving minimal loss when performing back translation (SLT) on generated videos. Ablations confirm the role of rigid canonicalization, non-rigid adaptation, and completion. Code, model checkpoints, and video examples will be released. Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.14122 [cs.CV] (or arXiv:2609.14122v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.14122 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Zhewen He [view email] [v1] Sat, 12 Sep 2026 19:57:16 UTC (854 KB) Full-text links: Access Paper: View a PDF of the paper titled SignMimic: Robust High-Quality Sign Language Motion Generation via Human-Shape-Oblivious Pose Transfer Guidance, by Zhewen He (1) and 5 other authorsView PDFHTML (experimental)TeX Source view license Current browse context: cs.CV prev | next new | recent | 2026-09 Change to browse by: cs References Citations NASA ADSGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv’s community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) mathjaxToggle(); We gratefully acknowledge support from our major funders, member institutions, , and all contributors. About Help Contact Subscribe Copyright Privacy Accessibility Operational Status (opens in new tab) Major funding support from
[CV-122] alking to Me or Someone Else? Rethinking Talk-to-Me Detection in Egocentric Videos
链接: https://arxiv.org/abs/2609.14118
作者: Feiyu Du,Xi He,Jia Li,Yapeng Tian,Weili Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 10 pages, 4 figures, 4 tables. Accepted to the 34th ACM International Conference on Multimedia (MM '26)
Abstract:Online understanding of who is talking to the camera wearer is a key capability for egocentric social interaction. However, existing talk-to-me (TTM) studies are commonly formulated as offline clip-level recognition, which is poorly aligned with online interaction and overlooks the diverse non-TTM speaking states that naturally arise in egocentric videos. In this paper, we revisit this problem by reformulating it as an online, frame-level prediction task. Instead of treating TTM as a binary problem against a single negative class, we model it in the presence of diverse and previously underexplored non-TTM states, such as talking-to-others, self-talking, and background conditions. To support this new formulation, we construct an Online TTM Dataset consisting of 406 egocentric video clips with approximately 900K annotated frames, each labeled with frame-level social interaction categories (e.g., background, TTM, talking-to-others, self-talking), by extending the Ego4D social interaction benchmark. In this benchmark, we evaluate five adapted baselines and develop a new model that integrates social cues across modalities. Experimental results show that our multimodal model, which jointly leverages audio, visual, and speech-semantic cues, achieves 75.5% frame-level F1 on TTM, outperforming strong baselines and enabling a systematic analysis of how different speaking states affect TTM recognition.
[CV-123] Parameter-Efficient Fine-Tuning of Foundation Models for Liver Tumor Segmentation in CT
链接: https://arxiv.org/abs/2609.14106
作者: Ramtin Mojtahedi,Mohammad Hamghalam,Jacob J. Peoples,Richard K. G. Do,Amber L. Simpson
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages, 2 figures, 1 table. Author manuscript of the published SPIE 2026 proceedings paper; LaTeX reconstructed from the author PDF
Abstract:We evaluated parameter-efficient fine-tuning (PEFT) of the Segment Anything Model (SAM) for liver tumor segmentation in abdominal CT of colorectal liver metastases. We compared Low-Rank Adaptation (LoRA), 4-bit Quantized LoRA (QLoRA), a convolutional adapter (Conv-Adapter), and our Directional Spectral Top-K adapter (DiSCo), training only adapters while freezing the SAM backbone. DiSCo derives spectral bases from singular value decomposition of row-normalized weights and learns rank-gated spectral coefficients, per-output magnitude offsets, and a spectral gain, with optional Top-K rank selection at inference and 0.14 M trainable parameters. We benchmarked five prompting regimes: no prompt, single-point, multi-point, and bounding boxes at intersection over union 0.50 and 0.75. Conv-Adapter and LoRA achieved the highest accuracy (overall Dice 0.793 and 0.792; single-point Dice 0.795 and 0.792; 95th-percentile Hausdorff distance (HD95) 32 mm). QLoRA was close (overall Dice 0.766; single-point Dice 0.768; HD95 36.41 mm), with 0.91 M trainable parameters, 120 ms latency, and 4.9 GB peak memory. DiSCo achieved the highest Dice per million trainable parameters (4.66), with overall Dice 0.653, single-point Dice 0.698, and HD95 49.53 mm. These results show an accuracy-efficiency trade-off and support PEFT for liver tumor segmentation with reduced adaptation costs when compute and labeled data are limited. Code: this https URL
[CV-124] Accelerating HKTex without Mesh Eigensystems: Local Unfolding and Randomized Thermal Features
链接: https://arxiv.org/abs/2609.14105
作者: Zhewen He,Junyi Hu,Yi Fang
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: 14 pages, 7 figures, 10 tables
Abstract:Heat Kernel Textures (HKTex) represent surface appearance with intrinsic anisotropic kernels, but evaluate them using 50 global Laplace-Beltrami eigendecompositions and a resident basis of shape [50,V,256]. We study two complementary ways to remove this bottleneck while leaving the trainer, GeodesicOpt, density control, and compositing unchanged. LocalHK exploits the measured locality of trained kernels and replaces spectral evaluation by radius-bounded hinge unfolding and an analytic log-map kernel. On 10 Objaverse meshes and an 8-mesh low-poly holdout, it changes mean view PSNR from 31.35 to 32.00 and from 29.82 to 30.76 dB, respectively, while reducing initialization by 40.5 times and enabling a 749,570-vertex proxy-backed run where the spectral baseline fails. ThermalRF instead preserves the discrete anisotropic heat semigroup: GPU sparse Chebyshev actions and randomized range finding construct global low-rank heat factors without mesh-sized eigenvectors, and a compiled evaluator mixes four neighboring thermal responses. On spot and a thin-stem challenge, ThermalRF reduces end-to-end preprocessing, initialization, and 5,000-step optimization by 29.3% and 24.4%, with every surface, atlas, or view PSNR change within 0.12 dB and training allocation reduced by about 90%. The two routes expose a useful design choice: maximal locality and scale versus fidelity to the thermal PDE. Broader thermal-feature evaluation and real large scenes remain future work. Comments: 14 pages, 7 figures, 10 tables Subjects: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR) Cite as: arXiv:2609.14105 [cs.CV] (or arXiv:2609.14105v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.14105 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Zhewen He [view email] [v1] Sat, 12 Sep 2026 19:03:52 UTC (319 KB) Full-text links: Access Paper: View a PDF of the paper titled Accelerating HKTex without Mesh Eigensystems: Local Unfolding and Randomized Thermal Features, by Zhewen He and 2 other authorsView PDFHTML (experimental)TeX Source view license Current browse context: cs.CV prev | next new | recent | 2026-09 Change to browse by: cs cs.GR References Citations NASA ADSGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv’s community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) mathjaxToggle(); We gratefully acknowledge support from our major funders, member institutions, , and all contributors. About Help Contact Subscribe Copyright Privacy Accessibility Operational Status (opens in new tab) Major funding support from
[CV-125] A Voxel-Spacing-Aware Extension of PyRadiomics for Anisotropic Texture Analysis
链接: https://arxiv.org/abs/2609.14103
作者: David Corral Fontecha,Juan Miranda Bautista,Pablo Menendez Fernández-Miranda,Andrea Trapote Fernandez,Lara Lloret Iglesias,Jose A. Vega
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Radiomic texture features are commonly extracted from anisotropic CT and MRI acquisitions, where identical voxel offsets may represent different physical distances. We implemented and validated a voxel-spacing-aware extension of PyRadiomics that incorporates spacing information without generating interpolated gray levels. The framework operates across the Python frontend, C wrapper, and computational backend. GLCM uses anisotropy-relative feature-level angular aggregation, NGTDM uses anisotropy-relative weighted neighborhood averaging, and GLRLM, GLDM, and GLSZM are computed on a finite-volume zero-order-hold representation derived from the native anisotropic grid. Synthetic 3D phantoms were used for software validation. The modified implementation reproduced standard PyRadiomics exactly when spacing-aware mode was disabled and remained equivalent under isotropic spacing across 75 texture features. Under anisotropic spacing, the method selectively modified texture families and was numerically distinct from nearest-neighbor, linear, and B-spline resampling. Computational profiling showed moderate runtime and memory increases, while sensitivity analyses quantified finite-volume rounding effects and confirmed that spacing-aware differences persisted across binWidth settings. The framework provides a backward-compatible technical basis for future evaluation of spacing-aware radiomics in heterogeneous medical imaging datasets.
[CV-126] RA-CoA: Training-free Fashion Image Captioning via Retrieval-Augmented Chain-of-Attributes
链接: https://arxiv.org/abs/2609.14100
作者: Abhirama Subramanyam Penamakuri,Shreya Shukla,Anand Mishra
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted in TMLR
Abstract:Fashion Image Captioning (FIC) plays a vital role in enhancing user experience and product search in e-commerce platforms. Unlike natural scene image captioning, FIC requires fine-grained visual reasoning and knowledge of domain-specific terminology to capture subtle attributes such as neckline and closure types, graphic patterns, and dress silhouettes. Moreover, as fashion inventories evolve rapidly with new trends, styles, and frequently emerging vocabulary, developing training-free captioning solution becomes essential for scalability and real-world adaptability. Instruction-tuned vision-language models (VLMs) offer a promising solution to fashion image captioning dueto their strong zero-shot capabilities and natural language fluency. However, these general-purpose models often lack attribute-level coverage and precision, and tend to hallucinate or misidentify fine-grained fashion details, making them less suitable for high-fidelity applications like product cataloging or personalized recommendations. To address this, we propose RA-CoA (Retrieval-Augmented Chain-of-Attributes), a novel, training-free framework that disentangles fashion image captioning into two interpretable stages: (i) retrieval of relevant attribute sets from a product knowledge base, and (ii) attribute-level reasoning to generate the final caption. RA-CoA is a model-agnostic approach that works with frozen VLMs to improve fine-grained attribute precision in product captions without the need for fine-tuning. Extensive evaluations across diverse VLM model families under different prompting paradigms demonstrate that RA-CoA significantly improves caption quality, achieving an average gain of 26.3% METEOR score over zero-shot captioning. We make our code publicly available.
[CV-127] Bridging the Synthetic-to-Real Gap for Few-Shot Cryo-ET Classification
链接: https://arxiv.org/abs/2609.14097
作者: Siddhant Bharadwaj,Ashish Vashist,Rashi Singh,Pranav Vinodh,Nishanth Artham,Runmin Jiang,Xingjian Li,Min Xu
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Subtomogram classification in cryo-electron tomography (cryo-ET) is a challenging problem due to the scarcity of labeled examples. While cryo-ET simulators can be adopted to generate unlimited synthetic data, the substantial domain gap between synthetic and real subtomograms hinders its practical utilization. In this work, we propose a novel synthetic-to-real adaptation framework with a learnable transformation module, bridging this gap at both the input and feature levels. Extensive experiments demonstrate that our method consistently outperforms existing transfer learning baselines in few-shot settings.
[CV-128] Adapting Open-Weight MLLM s to Generate Point Prompts for Electron Microscopy Segmentation ECCV2026
链接: https://arxiv.org/abs/2609.14080
作者: Samia Mohinta,Albert Cardona
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the BioImage Computing (BIC) Workshop at ECCV 2026
Abstract:Promptable models such as microSAM segment electron microscopy (EM) images from point prompts, but automation requires generating prompts without user input. We ask whether open-weight multimodal large language models (MLLMs) can generate them from natural-language requests by returning coordinates to a frozen segmenter. To that end, we convert masks from three mitochondria datasets into training examples, pairing images and instructions with centroid coordinates, then train LoRA adapters while freezing the MLLM backbone and microSAM. We find that Qwen3-VL reaches segmentation AP _50 0.736 after supervised fine-tuning and reward optimization, up from 0.247 without adaptation, while automatic prompt generation (APG) achieves 0.773 . In addition, two other MLLMs improve, reaching or exceeding APG. When compared with a supervised centroid-heatmap detector that reaches AP _50 0.904 for this mitochondria task, Qwen3-VL more closely matches the annotated point set and instance counts. Moreover, training on two public datasets transfers to an unseen third, while training on all three transfers to an independent EM volume. Robustness tests show stable performance under unseen formulations of the natural-language request, while the coordinates can be reused by a second segmenter. To our knowledge, this is the first feasibility study of open-weight MLLMs as EM point generators, providing an inspectable, language-directed link between localization and mask decoding.
[CV-129] LPA-CWM: A Learned Physical Adjudicator for Motion Reasoning with Counterfactual World Models
链接: https://arxiv.org/abs/2609.14073
作者: Kunwei Wu,Xiang Liu,Guocai Yao,Junming Chen,Zhikang Chen,Min Zhang,Pengwei Wang,Sen Cui
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: A quick overview is available at this https URL
Abstract:Counterfactual world models (CWM) extract motion from pretrained video predictors by comparing factual and intervened predictions. However, responses generated under different target-frame masks vary in reliability, while uniform aggregation weights them equally. We formulate response aggregation as candidate reliability learning and propose LPA-CWM with a lightweight Learned Physical Adjudicator (LPA). Trained on dense MOVi-F trajectories, the 3.0M-parameter LPA compares visual context and response structure across an unordered candidate set to predict relative weights, while the CWM predictor and intervention generator remain frozen. The weighted responses undergo windowed localization and one paired re-evaluation to recover motion. We also introduce Completeness-aware Motion Correspondence (CMC), a ground-truth-anchored evaluation protocol that jointly measures localization, trajectory completeness, visibility, and continuity, counting missing predictions as failures on visible dynamic points. On the evaluated DAVIS and Kinetics subsets, LPA-CWM improves \mathrmDCA_\mathrmavg over Uniform CWM by 60.0% and 29.0%, respectively, and also improves tracking accuracy under TAP-Vid First. A quick overview is available at this https URL.
[CV-130] Quantum-Gated LiteSSD: A Parameter-Efficient Lightweight Hybrid Quantum-Classical Framework for Forward-Looking Sonar Object Detection ECCV2026
链接: https://arxiv.org/abs/2609.14025
作者: Niloy Kumar Mondal,Poulomi Sarker Puja
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026 Marine Vision as short paper
Abstract:Forward-looking sonar object detection is essential for underwater perception, yet deployment on embedded platforms requires highly compact models. To address this challenge, we explore quantum computing and introduce Quantum-Gated LiteSSD, a parameter-efficient hybrid quantum–classical detector that reformulates QuCNet-style multi-circuit quantum processing as an identity-centered channel-gating mechanism for spatial feature modulation. Experiments on the Marine Debris Watertank dataset and UATD forward-looking sonar benchmarks demonstrate an effective parameter–accuracy trade-off. The proposed detector achieves 90.84% \mathrmmAP_50 on Watertank with approximately 62\times fewer parameters than YOLO26s and 164.3\times fewer parameters than SSD-VGG16. On UATD, the model achieves 70.37% \mathrmmAP_50 with only 0.150M parameters, making it approximately 4.1\times smaller than SSGA-YOLO while retaining meaningful multi-class detection capability.
[CV-131] CirrGuide: A Deep Cascaded Framework for Liver Cirrhosis Segmentation and Severity Classification from T2-Weighted MRI MICCAI2026
链接: https://arxiv.org/abs/2609.14010
作者: Muntaqim Ahmed Raju,Ruizhe Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 2 figures. Accepted at the 17th International Workshop on Machine Learning in Medical Imaging (MLMI 2026), held in conjunction with MICCAI 2026
Abstract:We present CirrGuide, a deep cascaded framework for cirrhotic liver segmentation and severity classification. Cirrhosis causes progressive structural changes in the liver and can lead to serious clinical complications, making severity assessment important for disease monitoring and treatment planning. However, severity classification is challenging because imaging patterns are often subtle, spatially variable, and similar across adjacent stages. CirrGuide addresses this by explicitly linking localization with classification. A ResNet50 encoder with an Attention U-Net decoder first predicts a soft cirrhotic liver mask, which is then used as an anatomical prior in a ResNet50-based classification branch. This branch combines global multi-scale features with mask-guided attention-pooled regional features to classify Mild, Moderate, and Severe cirrhosis. On the official CirrMRI600+ T2-weighted (T2W) 2D split, CirrGuide achieves 89.83% Dice and 84.14% mIoU for segmentation, 69.58% accuracy and 61.55% macro F1-score for severity classification. Compared with segmentation-only, classification-only, and multi-task baselines, CirrGuide improves both localization and severity classification, demonstrating the benefit of using predicted cirrhotic liver masks as anatomical priors for cirrhosis analysis.
[CV-132] Mind2Cloud: EEG-to-Point Cloud Generation with Two-Granularity Diffusion Decoding ECCV2026
链接: https://arxiv.org/abs/2609.13991
作者: Yongyi Lu,Xiongfeng Huang,Zhijing Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: European Conference on Computer Vision – ECCV 2026
Abstract:Reconstructing 3D objects from brain signals offers a promising avenue for understanding human visual cognition. While prior work has shown initial success using EEG signals for 3D reconstruction, existing methods typically employ a uniform diffusion decoder, overlooking the evolving semantic granularity of both EEG representations and the diffusion denoising process. In this paper, we propose Mind2Cloud, a novel EEG-to-point-cloud generation framework based on two-granularity diffusion decoding. The core of Mind2Cloud is a time-aware decoder that integrates a global Transformer branch and a local Point-Voxel CNN (PVCNN) branch across diffusion timesteps through a learnable fusion mask. Specifically, Transformer layers are incorporated into the early upsampling stages to capture global object structure under high uncertainty, while PVCNN modules are used in later stages to refine local geometric details. Inspired by the hierarchical nature of EEG-based visual representations, this design dynamically adapts its spatial granularity in accordance with the coarse-to-fine trajectory of diffusion denoising. We further introduce an adversarial refinement module to enhance geometric realism and semantic consistency. Extensive experiments on the EEG-3D dataset across all 12 subjects demonstrate that Mind2Cloud outperforms prior work in both geometric accuracy and semantic alignment, setting a new benchmark for EEG-to-point-cloud generation. Our source code is available at this https URL.
[CV-133] What Makes an Efficient VLA? Navigating Action-Head Design Scaling and Latency
链接: https://arxiv.org/abs/2609.13984
作者: Luoyang Sun,Guoyang Xia,Fengfa Li,Lei Ren,Xinyu Cui,Haifeng Zhang,Fangxiang Feng,Kaike Zhang,Kun Zhan,Yan Xie,Jun Wang,Cheng Deng
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-Language-Action (VLA) models combine a pretrained vision encoder, a language backbone, and an action head, but their relative contribution has not been established under controlled, latency-paired conditions. We fix the backbone families (SigLIP2 and Qwen2.5) and the training pipeline, sweep action-head design and module scale, and pair each configuration with measured on-device latency. The study yields three findings. First, action-head performance is governed primarily by initialization rather than decoder architecture, loss, or inference budget: copying the last transformer layers of the language backbone into the head is the single largest lever, at no latency cost, and the only axis that helps at every module scale. Alignment also explains the other axes: flow matching and a heavier decoder pay off only while the head is misaligned and reverse once it is aligned, and extra inference passes give no measurable benefit; expressiveness appears to substitute for missing alignment. We read this as representation transfer: the aligned head keeps attending to the instruction’s object nouns and stays close to the backbone in weight space rather than relearning to act from scratch. Because we reach alignment only through initialization, we offer this as the account that best organizes the measurements, not a demonstrated cause, and name the control that would settle it. Second, capacity pays only after alignment: the aligned action head is the highest-return module to scale. Third, those returns diminish sharply near the size today’s \pi -series VLAs already use, so further growth buys little in-domain accuracy for its latency. These specify EffVLA, a compact model matching the strongest open-source VLAs on standard LIBERO, leading on most LIBERO-Plus perturbation axes at lower latency, and transferring to a real SO-ARM101 arm with the recipe unchanged.
[CV-134] Zero-Shot Cross-Material Ptychographic Phase Reconstruction Using Deep Learning
链接: https://arxiv.org/abs/2609.13969
作者: Wen-Chun Lin,Yu-Chee Tseng,Jen-Jee Chen,Nan-You Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Ptychographic phase reconstruction is commonly formulated as an iterative inverse problem, requiring repeated object-probe updates and resulting in substantial computational cost for large-scale 4D-STEM data. We present a direct local-to-global learning framework that reconstructs full-field phase maps from diffraction measurements without iterative refinement during inference. The proposed network predicts local wrapped-phase patches from individual diffraction patterns using a sine-cosine representation, and the predictions are assembled into a full-field reconstruction using calibrated scan positions and Gaussian-weighted stitching. To evaluate generalization beyond the training domain, the model is trained on one material and directly applied to another in a zero-shot setting without target-domain fine-tuning. Experiments on AuPd and MoS _2 demonstrate consistent cross-material transfer in both directions, with the proposed method achieving the best full-field MSE, PSNR, and MS-SSIM among the evaluated learning-based methods. Compared with the iterative ePIE approach, the proposed direct local-to-global pipeline reduces end-to-end reconstruction time by approximately 10x, demonstrating its potential for efficient and transferable ptychographic reconstruction.
[CV-135] SGWIB:Sliced Gromov-Wasserstein Information Bottleneck for Video Highlight Detection
链接: https://arxiv.org/abs/2609.13966
作者: Hanjuan Huang,Yung-Chieh Yeh,Hsing-Kuo Pao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 13 pages, 6 figures
Abstract:Video highlight detection aims to identify temporally important segments that capture the most informative or engaging events in a video. Reliable prediction therefore requires not only discriminative segment representations but also preservation of the temporal relationships among neighboring and distant segments. The information bottleneck principle has proven effective for learning compact and task-relevant representations, yet it has not been explored for video highlight detection, and applying conventional formulations directly would overlook inter-segment relational structure and distort highlight relevant temporal organization during compression. We therefore introduce the Sliced Gromov-Monge Gap (SGMG), a structure aware regularizer that measures the excess relational distortion induced by a prescribed source-to-bottleneck mapping relative to an optimal sliced structural correspondence. Building on SGMG, we develop SGWIB, an information-bottleneck framework for single-modal video highlight detection that learns compact bottleneck representations while preserving inter-segment temporal structure. We further introduce Home-Away-Related Contextual Pseudo-Labels and a contextual disentanglement module that reduce sports-specific contextual bias by separating highlight oriented information from contextual patterns. Experiments on MrHiSum and MoSu show that SGWIB attains the best Kendall’s tau, Spearman’s rho, mAP@50, and mAP@30 among the compared single-modal methods on both datasets. On MrHiSum, the visual model improves the strongest previous results by 0.031, 0.031, 0.87, and 0.75 on these four metrics, respectively. These results show that structure-aware information-bottleneck regularization combined with contextual disentanglement improves segment-level highlight prediction.
[CV-136] BLInD: Learning Driver Intent as a Distribution over Future Ego Trajectories ITSC
链接: https://arxiv.org/abs/2609.13941
作者: Flavian Pegado,Ronit Hire,Shreyas Rajesh,Soham Phade
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: Accepted at the 2026 IEEE International Conference on Intelligent Transportation Systems (ITSC)
Abstract:We present BLInD (Blind Learned Intent Distribution), a compact network that maps recent vehicle-state history (e.g. speed, curvature, indicator, and vehicle type) to a top-k distribution of future ego trajectories, with no camera, LiDAR, map, or object-track inputs. We find that vehiclestate history alone is sufficient to learn a useful multimodal distribution over near-term ego trajectories, and its low-latency nature makes it well-suited for safety-critical deployment. We investigate two distribution architectures, autoregressive (AR) and flow-matching, and train on both mixed-platform opensource and Wayve datasets. Both generalize without datasetspecific adaptation; the flow-matching model achieves best topk ADE/FDE of 0.15/0.37 m on Wayve, 0.15/0.36 m on Waymo, and 0.28/0.59 m on nuScenes, with the AR model reaching comparable coverage. Integrating the distributions into an AEB trigger task, a strict all-candidates policy reduces false positives from 1.51% to 0.11% with AR (13.7x reduction, 94.9% TP) and to 0.06% with flow-matching (25.1x reduction, 98.7% TP) compared to a 1-CTRV policy with 100% true positive score. BLInD runs in 0.87 ms with the AR head and 2.9 ms with the flow-matching head on an NVIDIA DRIVE Orin ECU making it compatible with real-time deployment on automotive ECUs. While existing learned distribution models rely on scene context and blind vehicle-state models typically collapse to a single path, BLInD is learned, blind, and cross-domain simultaneously, a combination not demonstrated by prior work. These results show that such a distribution provides a controllable and plausible intent sampling interface for downstream systems, with AEB as one instantiation.
[CV-137] SkyAnchor: Updating Metric-scale Aerial 3D Gaussian Scenes from Unposed Ground-View Sequences
链接: https://arxiv.org/abs/2609.13903
作者: Zhuoxiao Li,Xinyi Liu,Taoyu Wu,Yinrui Ren,Tongyan Hua,Ou Jing,Shuai Zhang,Dongli Wu,Rongjun Qin,Ge Lin Kan,Wufan Zhao
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:We study how to update a pre-built aerial scene with a newly captured, unposed ground-view sequence. The aerial scene already contains a reliable metric Structure-from-Motion (SfM) reconstruction and a pre-trained 3D Gaussian Splatting (3DGS) model, whereas the ground-view sequence is collected later to add street-level appearance but has unknown camera poses and global scale. Registering this sequence to the aerial SfM reconstruction is challenging because single-image cross-view localization is brittle and long trajectories are prone to drift. To address these challenges, we present SkyAnchor, which treats the existing aerial scene as a fixed scaffold for ground-view registration and scene update instead of jointly reconstructing aerial and ground imagery from scratch. It first localizes short groups of consecutive ground frames against geometrically verified aerial support, producing sparse anchor poses. It then recovers the full ground trajectory with anchor-constrained submaps, fixing the front and rear anchor poses during incremental registration and bundle adjustment. Finally, it inserts filtered ground Gaussians while preserving the aerial view, followed by lightweight joint refinement. Experiments on seven real aerial–ground scenes show accurate metric ground trajectories and updated 3D Gaussian scenes with strong aerial- and ground-view rendering quality.
[CV-138] Physically Typed and Geometry-Aware Representations for Earth Foundation Models
链接: https://arxiv.org/abs/2609.13868
作者: Rajiv Ranjan
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 18 pages
Abstract:Earth-observation (EO) foundation models have become exceptionally effective at learning se mantic, high-dimensional geospatial embeddings, while modern weather and climate models have demonstrated that Earth-specific geometry, spherical operators, meshes, and hybrid physical solvers can materially improve prediction. Yet these two advances are not equivalent. A conventional latent embedding has no inherent physical transformation law, whereas scalar fields, tangent polar-vector fields, axial/pseudovector quantities, covectors, and higher-order tensors transform differently under rotations, reflections, and changes of local coordinate frame. This proposal asks whether a general purpose Earth foundation model should preserve those distinctions explicitly, or whether standard embeddings plus augmentation already learn everything that matters. The central contribution is therefore not a more complicated architecture by assumption, but a staged falsification program. A compute-conscious ERA5 dry run first compares conventional, augmentation-matched, typed equivariant, and Hodge/Helmholtz variants under spatial, temporal, orientation, and low-data shifts. Only if explicit geometric typing yields reproducible improvements does the program advance toward a multimodal Earth foundation model in which semantic embeddings coexist with physically typed fields. The proposed gap is narrower and more defensible than claiming that current models ignore geometry entirely: several systems already respect spherical domain geometry, and emerging work explicitly learns scalar/vector fields on spheres. The unresolved question is whether foundation-scale, multimodal, parity-aware field typing produces practical gains beyond those existing approaches.
[CV-139] Event-Level Emotion Recognition in the Wild Using Deep Facial Expression Analysis
链接: https://arxiv.org/abs/2609.13854
作者: Aleksandr Semerikov,Pakizar Shamoi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: The paper has been submitted to IEEE conference
Abstract:Facial emotion recognition (FER) in real-world environments remains challenging due to unconstrained imaging conditions, including multiple faces, occlusions, pose variations, and complex lighting. Most existing studies focus on individual facial emotion classification and do not address the analysis of collective emotional states at the event level. This paper proposes an end-to-end pipeline for event-level emotion recognition from photographs. The approach detects faces in each image, classifies facial expressions using a deep convolutional neural network, and aggregates face-level emotion probabilities to estimate the overall emotional distribution of a public event. A comparative evaluation of several CNN architectures on the FER- 2013 and RAF-DB datasets demonstrates that transfer learning with EfficientNet-B2 trained on RAF-DB is more suitable for real-world RGB data. The proposed method is evaluated on a real-world event dataset containing 1658 images. Experimental results show stable emotion distributions across event subsets, confirming the effectiveness of event-level aggregation for emotion analysis in the wild.
[CV-140] otalSynth: Robust Whole-Body Synthetic CT from MRI and CBCT
链接: https://arxiv.org/abs/2609.13838
作者: Valentin Boussot,Cedric Hemon,Anais Barateau,Caroline Lafond,Jean-Claude Nunes,Jean-Louis Dillenseger
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 3 figures
Abstract:Purpose: To develop and evaluate TotalSynth, a reusable pretrained framework for whole-body synthetic CT (sCT) generation from MRI and cone-beam CT (CBCT) images. Materials and Methods: In this retrospective technical study, the dataset was assembled between 2020 and 2026 from SynthRAD challenge data, four prostate cohorts, and BIC-MAC. After registration quality control, 1450 of 1800 public challenge pairs were retained; 350 were excluded for insufficient registration quality or major source/CT mismatch. The corpus also included 84 additional prostate MRI/CT and 60 external BIC-MAC MRI/CT cases. Three 5-fold model families were evaluated with image-domain, anatomy-aware, registration-based, and uncertainty metrics. Age and sex were not consistently available across public datasets. Results: The released MRI-to-CT model achieved an overall MAE of 67.49 HU, SSIM of 0.920, and PSNR of 29.28 dB. The CBCT-to-CT model achieved an overall MAE of 53.55 HU, SSIM of 0.939, and PSNR of 32.09 dB. The unified model maintained similar performance on MRI inputs (MAE, 67.68 HU) and CBCT inputs (MAE, 54.22 HU). On external BIC-MAC data, MRI-to-CT MAE was 100.91 HU without fine-tuning and 62.21 HU after fine-tuning. Conclusion: TotalSynth provides reusable MRI- and CBCT-based CT synthesis models with broad anatomical coverage, while external evaluation highlights the need for local validation and optional fine-tuning under domain shift.
[CV-141] DiVA: Enabling Interactive Digital Life Simulation via Video Models
链接: https://arxiv.org/abs/2609.13830
作者: Cheng Chen,Hao Ouyang,Qiuyu Wang,Ka Leong Cheng,Wen Wang,Yihao Meng,Hanlin Wang,Yixuan Li,Jiacheng Wei,Zhenshan Tan,Yanhong Zeng,Yujun Shen,Guosheng Lin,Fayao Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:We present DiVA, a deeply interactive digital life simulator pioneering a new paradigm for long-term, open-ended interactive experiences within digital character worlds. DiVA’s architecture pairs a Multimodal Large Language Model (MLLM) as a router with a meticulously designed stacked video pipeline for seamless, multi-turn interactions with action and audio response. To maintain continuity and avoid degradation, we model generation as a three-part coupled system: waiting video, action video, and the transitions between them. These transitions are critically handled by our Anchored Video Continuation (AVC) module, which returns the character to stable states to prevent degradation. By encoding information from the preceding action video segment, AVC ensures smooth transitions, significantly reducing camera jitter and inconsistencies common in current video transition methods. This design also enables complex pose changes (e.g., sitting to standing) typically difficult for audio-driven models. These system designs together ensure high-fidelity identity, coherence, and dynamics for extended experiences. To validate our pipeline design, we comprehensively compare our system against alternatives by replacing our core generation module with mainstream long-video, continuation, and interpolation methods. We further analyze the necessity of the three-stage design, anchor-state selection, transition naturalness, spatial grounding, and the quality-latency trade-off, and we expand the comparison to additional long-form audio-driven avatar models. Results confirm DiVA is markedly superior in maintaining long-term visual quality and realism, validating its effectiveness as a sustainable, interactive simulation.
[CV-142] Semantic Privacy Protection with Utility Preservation for 3D Point Clouds
链接: https://arxiv.org/abs/2609.13823
作者: Jinchang zhang,Jiakai Lin,David Crandall,Guoyu Lu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Point cloud data face serious semantic privacy risks during acquisition, transmission, and cross-institutional sharing. Existing methods mostly rely on geometric perturbation or destructive encryption, which can reduce the recognizability of the original class but often impair downstream usability. This paper proposes a class-transfer-based semantic encryption framework for point clouds, aiming to conceal original class information while preserving task utility and supporting authorized recovery. Specifically, we construct a unified latent space with a shared-backbone Normalizing Flow, and combine LoRA and FiLM to achieve parameter-efficient class-conditional adaptation. We further introduce diffusion-guided flow alignment to regularize the latent distribution, construct an energy-based category transition graph, and obtain an optimal class-transfer table through global matching. Then, a latent-space Neural ODE continuously evolves source-class latents into target-class latents, which are decoded into target-class point clouds through the inverse flow. We adopt attacker-oriented metrics, including New-Class Recognition Rate (NCRR), Original-Class Leakage Rate (OCLR), and Original Label Recovery Rate (OLRR), to evaluate privacy and utility. Experiments on classification and segmentation benchmarks show that the proposed method achieves controllable semantic transformation, effectively reduces original-class semantic leakage, preserves downstream learnability in the protected domain, and supports reliable authorized reconstruction.
[CV-143] Beyond OCR Accuracy: Text-Centric VQA Under Image Degradation with Modular and End-to-End
链接: https://arxiv.org/abs/2609.13815
作者: Ritali Vatsi,Rachapudi Jagadeesh,Shruti Singh Baghel,Himani Sharma,Amit Shukla,Pawan Goyal
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Text-centric Visual Question Answering (VQA) requires reading and reasoning over text embedded in images, a task made substantially harder when images suffer from real-world degradation such as motion blur, low resolution, or compression artifacts. While modular OCR-based pipelines and end-to-end vision-language models are both widely used for this task, their comparative robustness under degraded conditions remains underexplored. We present an empirical study comparing two modular pipelines with SA-DBNet, a custom detector architecture combining ResNet-18 with self-attention spatial modeling and deformable convolutions against an end-to-end vision-language baseline, evaluated on 4013 degraded images with 7000 question-answer pairs. Fine-tuned modular pipelines achieve up to 57.50% exact-match accuracy versus 38.00% for the end-to-end baseline, with domain-specific fine-tuning yielding a gain of up to 29.50 percentage points. Critically, we find that conventional OCR error metrics like Character Error Rate and Word Error Rate are unreliable predictors of downstream VQA performance, as semantic reasoning can compensate for recognition failures when contextual cues are present. These findings highlight the importance of task-aware evaluation for text-centric VQA systems under realistic visual conditions. Codes are available here
[CV-144] Realtime-Venus: A full-duplex interaction system with asynchronous delegation
链接: https://arxiv.org/abs/2609.13814
作者: Ruixiang Zhao,Hualei Wang,Renhe Sun,Enzhi Zhou,Jincenzi Wu,Xujie Song,Kexin Shi,Zihang Liu,Pengcheng Zhu,Jiayi Zhou,Baoyue Zhang,Changhao Zhang,Zitong Wang,Jinhong Wang,Tong Niu,Jingjing Liu,Junan Lin,Haolin He,Hengshuo Chu,Yuhui Chen,Jian Liu,Yuge Huang,Junliang Xing,Yuntao Wang,Weiqiang Wang,Chun Yu,Yuanchun Shi
类目: Computer Vision and Pattern Recognition (cs.CV); Audio and Speech Processing (eess.AS)
备注:
Abstract:Natural interaction in digital and physical environments requires continuous perception and timely responses. Spoken dialogue relies on acoustic and linguistic cues, while video interaction also requires grounding the conversation in evolving visual context. We present Realtime-Venus, a proactive full-duplex interaction system with two separately trained 9B models: Realtime-Venus-Omni for audio-visual interaction and Realtime-Venus-Audio for spoken interaction. Each model serves as a complete conversational frontend, integrating continuous perception, conversational control, and native speech generation through a shared causal timeline for user inputs, model outputs, and delegation events. A dual-loop runtime coordinates live interaction with background reasoning and tool execution. Foreground interaction continues while Realtime-Venus-Harness executes tasks asynchronously and returns results for integration into the ongoing dialogue. Both models follow a common post-training recipe combining offline understanding, proactive full-duplex trajectories, and delegation workflows. Among the evaluated online models, Realtime-Venus-Omni achieves the highest scores on six of eight video benchmarks, including StreamingBench (70.2%), OVO-Bench (64.7%), and Daily-Omni (81.3%). Across eight audio understanding and spoken question answering benchmarks, Realtime-Venus-Audio leads the compared models on MMAU (78.0%), MMAU-Pro (63.2%), Llama Questions (83.8%), and Speech CMMLU (67.8%), while matching the best VoiceBench AlpacaEval score of 4.81. On Full-Duplex-Bench v1.5, Realtime-Venus-Audio responds to 75% of user interruptions and achieves continuation rates of 97%, 88%, and 86% under backchannels, other-directed speech, and background speech, respectively, exceeding Gemini 3.1 Live and GPT-4o on all three continuation metrics. Subjects: Computer Vision and Pattern Recognition (cs.CV); Audio and Speech Processing (eess.AS) Cite as: arXiv:2609.13814 [cs.CV] (or arXiv:2609.13814v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.13814 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Roy Sun [view email] [v1] Sat, 12 Sep 2026 08:56:50 UTC (2,007 KB)
[CV-145] StepPrune: Adaptive Sequential Visual Token Selection across Multimodal Large Language Models
链接: https://arxiv.org/abs/2609.13804
作者: Hansen Zhang,Landi He,Mingde Yao,Lijian Xu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Visual prefixes account for a major portion of the per-layer computation in multimodal large language models (MLLMs), making visual-token pruning a direct approach to accelerating inference. Existing top-K methods typically evaluate tokens independently and apply a uniform budget to all inputs, overlooking both selection-dependent interactions and variations in visual complexity across samples. In contrast, we propose StepPrune, which formulates visual-token pruning as an adaptive sequential decision process. Conditioned on previously selected tokens and textual context, StepPrune progressively constructs the retained subset and automatically determines its size through a learned STOP action. During training, a variance-preserving noise gate provides a differentiable surrogate for the discrete selection process, whereas during inference, unselected tokens are physically removed before language-model prefill. A grouped selection mechanism further extends StepPrune to high-resolution inputs. Experiments across LLaVA-1.5, LLaVA-NeXT, Qwen2.5-VL, and InternVL3 show that StepPrune achieves the best average normalized performance retention across all evaluated pruning rates on LLaVA-1.5, Qwen2.5-VL, and InternVL3, while remaining competitive on the substantially longer AnyRes prefixes of LLaVA-NeXT. On LLaVA-1.5, StepPrune retains 94.6% of the full-prefix normalized performance while pruning 88.9% of the visual tokens. At a mean retained count of 64, StepPrune reduces prefill latency from 59.95 ms to 40.05 ms, corresponding to a 1.50x prefill speed-up.
[CV-146] Restore What Matters: Lessons from Joint Restoration and Recognition
链接: https://arxiv.org/abs/2609.13791
作者: Lanqing Guo,Xijun Wang,Minchul Kim,Yu Yuan,Wes Robbins,Xingguang Zhang,Nicholas Chimitt,Stanley H. Chan,Zhangyang Wang,Xiaoming Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recognition pipelines typically adopt a restore-then-recognize workflow, yet decades of experience show that generating visually pleasing images seldom translates to improved recognition. We propose a Joint Restoration-for-Recognition (JR ^2 ) paradigm: restore only what downstream tasks truly require, with task signals dictating where, how much, and whether restoration is necessary. JR ^2 rests on three pillars: (i) Physics, employing optics-accurate turbulence simulation, extensible to blur and noise, to ground restoration in real image formation; (ii) Neuroscience, drawing on selective attention and neuroplasticity to direct model capacity toward identity-critical regions and frames while bypassing already-clean inputs; and (iii) Vision Learning, coupling recognition loss end-to-end through restoration and alignment so that low-level edits maximize high-level identity stability. Evaluations on IARPA-BRIAR show consistent improvements (e.g., TAR@0.01% FAR +0.6; FNIR@1% FPIR -2.5), while a quality gate skips ~70% of clean frames, reducing cost. Ablations confirm physics priors enhance realism, joint training prevents catastrophic forgetting, and selective restoration suffices in many cases. We conclude that better-looking images are neither necessary nor sufficient; restoration modules must be task-driven, selective, and physically aware. Code, pretrained models, and recipes are provided for integration.
[CV-147] GEAR: From Dynamic Encoding to Dynamic Activation in Social Trajectory Prediction ICDM2026
链接: https://arxiv.org/abs/2609.13778
作者: Jiaheng Chen,Jiaxing Li,Leixia Wang,Jianan Ju,Tinghe Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted by ICDM 2026
Abstract:Human trajectory prediction requires modeling both individual motion patterns and social interactions among agents. Existing methods have made substantial progress by using attention mechanisms, graph structures, and temporal encoders to capture dynamic social context. However, most of them primarily focus on how social information is encoded, while paying less explicit attention to how the encoded social context should take effect during future trajectory generation. In this paper, we argue that dynamic social encoding does not necessarily imply dynamic social activation. The same interaction context may require different activation strengths across future horizons and scene densities: social cues should be strengthened when interaction evidence is strong, but suppressed when they are weak or noisy. To address this issue, we propose GEAR, a generation-aware bias activation model for human trajectory prediction. Built upon a bias-decomposed trajectory generation formulation, GEAR dynamically activates the individual-motion and social-resonance bias terms at each future step before final trajectory composition. This allows the model to explicitly control when and how strongly individual and social bias components participate in generation. Experiments on ETH-UCY, SDD, and NBA show that GEAR consistently improves the resonance-based baseline and achieves competitive state-of-the-art performance. Further analyses of activation patterns and density-grouped errors validate the importance of calibrating encoded social context during trajectory generation. Our code is available at this https URL.
[CV-148] When Do Learned Priors Help Visual Inertial Estimation? A Controlled Study of Prior Integration Calibration Initialization and Backend Consistency
链接: https://arxiv.org/abs/2609.13777
作者: Jinchang Zhang,Guoyu Lu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Learned components are increasingly integrated into geometric visual–inertial estimators to provide motion, depth, bias, uncertainty, or confidence cues. Yet it remains unclear whether gains arise from useful learned priors or from changes in the backend, calibration, initialization, temporal association, or evaluation gauge. We present a controlled framework for learning-augmented visual–inertial estimation that separates fusion gain from the incremental value of a learned prior and evaluates four evidence layers: local motion consistency, global trajectory accuracy, physical-state correctness, and numerical consistency. We instantiate the framework with a MonoViT-based monocular motion prior added as a local relative-motion factor to an unchanged VINS backend. We compare Original VINS and learned-prior VINS under matched sensor streams, timestamps, initialization, frontend/backend settings, and camera–IMU extrinsics, while probing calibration, initialization, state coupling, scale, bundle adjustment, and loop closure. On KITTI, with fixed reference extrinsics, translation APE RMSE is 31.4 m for Original VINS and 31.8 m with the learned prior. Across four recordings, the prior changes mean APE by only -0.2%, while a five-times-higher weight worsens it by 8.2%. Online extrinsic updates increase mean APE by 45.7% and 52.1%, respectively, while mean RPE changes by less than 2%. These results show that fusion performance alone cannot establish the value of learned priors. Reliable evaluation requires same-backend controls and joint analysis of prior compatibility, calibration, initialization, global drift, physical-state error, and backend consistency.
[CV-149] Development of Low-Cost Real-Time Driver Drowsiness Detection System using Eye Centre Tracking and Dynamic Thresholding
链接: https://arxiv.org/abs/2609.13756
作者: Fuzail Khan,Sandeep Sharma,M. R. Arulalan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:One in every five vehicle accidents on the road today is caused simply due to driver fatigue. Fatigue or otherwise drowsiness, significantly reduces the concentration and vigilance of the driver thereby increasing the risk of inherent human error leading to injuries and fatalities. Hence, our primary motive being - to reduce road accidents using a non-intrusive image processing based alert system. In this regard, we have built a system that detects driver drowsiness by real time tracking and monitoring the pattern of the driver’s eyes. The stand alone system consists of 3 interconnected components - a processor, a camera and an alarm. After initial facial detection, the eyes are located, extracted and continuously monitored to check whether they are open or closed on the basis of a pixel-by-pixel method. When the eyes are seen to be closed for a certain amount of time, drowsiness is said to be detected and an alarm is issued accordingly to alert the driver and hence, prevent a casualty.
[CV-150] Hyper-LLaVA: Hyperbolic Uncertainty-aware Modality-Balanced Routing for Multimodal Continual Instruction Tuning ICML2026
链接: https://arxiv.org/abs/2609.13742
作者: Kunlun Xu,Yanqin Zhang,Wenwen Qiang,Jiahuan Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ICML 2026
Abstract:Multimodal Continual Instruction Tuning (MCIT) aims to exploit the incrementally accumulated knowledge to process multimodal inputs of diverse tasks, where parameter routing plays an important role. State-of-the-art methods rely on sample-to-task center similarity and cross-modal fusion with equal weight during routing. However, such solutions face two fundamental flaws: (1) Within each modality, the sample-to-task center distance is sub-optimal for routing since the abundant intra-task diversity information is underleveraged. (2) Different modalities exhibit varying reliability across tasks, where the modality with inter-task ambiguity can easily misguide the routing result. To address these problems, we propose Hyperbolic Uncertainty-aware Modality-Balanced Routing (Hyper-LLaVA) to improve parameter routing capacity based on cross-modality task feature uncertainty modeling. Specifically, to improve intra-modality task matching, Hyper-LLaVA accesses the sample-to-task distribution similarity in the Hyperbolic space. Besides, to alleviate the degradation brought by unreliable modalities, Hyper-LLaVA quantifies the task matching ambiguity within each modality to achieve adaptive balancing between task matching across modalities. Based on the complementary intra- and inter-modality task matching enhancement, our Hyper-LLaVA outperforms state-of-the-art approaches by large margins. Our source code is available at this https URL
[CV-151] FFVO: A Feedforward Pose Decoder for Long-Horizon Visual Odometry
链接: https://arxiv.org/abs/2609.13733
作者: Meng-Li Shih,Shih-Yang Su,Yuliang Zou,Hao Xiang,Haidong Zhu,Vincent Casser,Brian Curless,Dmitry Kalenichenko,Mingxing Tan,Dragomir Anguelov
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 9 figures
Abstract:Stable and reliable 4D spatial understanding is fundamental for autonomous driving systems. While feedforward reconstruction networks can estimate camera motion and 3D structure in one pass, pose estimation over long videos remains challenged by computational cost, long-context ambiguity, and temporal instability. To address these challenges, we propose Feedforward Visual Odometry (FFVO), a pose-specialized adaptation of joint reconstruction architectures for efficient and temporally stable camera-pose estimation. FFVO uses (i) a compact camera-token representation for computationally efficient temporal aggregation, (ii) a hierarchical local-to-global temporal decoder that mitigates geometric ambiguity by separating short-range motion aggregation from sequence-level integration, and (iii) intermediate trajectory supervision that promotes temporal stability. Extensive evaluation on the Waymo Open Dataset (WOD), KITTI, and a large-scale proprietary benchmark demonstrates that our method performs favorably against existing feedforward approaches, and greatly reduces jitter and drift. These results support FFVO as an effective feedforward camera-pose decoder in long-horizon visual odometry settings.
[CV-152] Rank-Consistent Set Reasoning for Co-Salient Object Detection
链接: https://arxiv.org/abs/2609.13706
作者: Yuan Xiang,Matteo Rossi,Yingzhou Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Co-salient object detection (Co-SOD) requires a model to find foreground regions that are salient in individual images and supported by the image group. We present \emphRank-Consistent Set Reasoning (RCSR), a supervised dense-prediction framework that models a group as an unordered set rather than as a sequence of images or a semantic label. The core idea is to rank how strongly each spatial region agrees with a small collection of learned group slots at every image scale, and to aggregate these ranks with a robust trimmed statistic. This suppresses accidental pairwise matches and prevents one atypical group member from dominating the shared representation. A set encoder builds group slots directly from multi-scale visual features, while a rank-consistency gate measures whether the ordering of candidate regions is stable across group members. The gated slots are decoded jointly with per-image features to produce co-saliency maps. The model contains no natural-language branch, no open-vocabulary detector, and no external segmentation model. We further introduce a group permutation objective and hard-distractor augmentation so that the model learns the properties of a set-level target rather than memorizing image order or isolated visual saliency. We formulate an evaluation protocol for CoCA, CoSal2015, and CoSOD3k, together with tests of group-size robustness, distractor rejection, order invariance, and cross-dataset transfer.
[CV-153] MomentBA: Second-order Spatial Moments for Anisotropic Correspondence Uncertainty in Differentiable Bundle Adjustment
链接: https://arxiv.org/abs/2609.13691
作者: Yuqing Wang,Xiaoji Niu,Yan Wang,Hailiang Tang,Jian Kuang,Tisheng Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Most existing visual odometry (VO) systems treat feature correspondences as deterministic measurements or assign uniform uncertainty, ignoring the inherent localization ambiguity of different observations. However, correspondence uncertainty is often anisotropic due to image structures such as edges, repetitive patterns, and motion blur, which can significantly affect geometric optimization. In this work, we propose MomentBA, a geometry-aware bundle adjustment framework that derives anisotropic correspondence uncertainty from second-order spatial moments of local similarity responses. Instead of introducing additional covariance prediction networks, the proposed method directly converts matching response distributions into interpretable covariance estimates and incorporates them into bundle adjustment as correspondence-specific information matrices for uncertainty-aware residual weighting. Furthermore, the proposed formulation is integrated into a differentiable optimization framework, establishing a direct connection between correspondence uncertainty and geometric estimation. Experiments on the EuRoC MAV and TartanAir v1 Hard datasets demonstrate that MomentBA improves monocular visual odometry accuracy compared with existing feature-based and learning-based approaches. The proposed anisotropic covariance model achieves lower rotational errors and more robust trajectory estimation than fixed and isotropic uncertainty models, validating the effectiveness of geometry-induced uncertainty modeling for challenging visual environments.
[CV-154] MARC: Morphology-Aware Regression of Consensus for Cell Segmentation in Subcellular Spatial Transcriptomics
链接: https://arxiv.org/abs/2609.13665
作者: Xinyu Shu,Andrew Zhang,Jean Yang,Jinman Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 4 figures
Abstract:Accurate cell segmentation remains a major bottleneck in subcellular spatial transcriptomics (SST), in which morphological images and spatially resolved RNA transcripts are used to partition tissues into individual cellular instances. As segmentation serves as the foundation for constructing cell-level representations, boundary errors can lead to incorrect transcript assignments and compromise downstream analyses. However, reliable ground-truth boundaries are unavailable because they must be inferred from incomplete morphological and transcript signals. Furthermore, manual annotation of a large number of cells is time-consuming. Agreement among complementary segmentation methods provides a practical surrogate for identifying well-supported and ambiguous regions, but explicit consensus construction requires executing multiple computationally intensive pipelines. In this study, we propose MARC (Morphology-Aware Regression of Consensus), a framework that predicts a multi-method consensus-support map for SST segmentation. MARC is trained with leave-one-method-out consensus pseudo-targets and a Foreground-Union Consensus Loss that focuses supervision on candidate and consensus foreground. We evaluated MARC on 4,642 held-out tiles from Xenium kidney tissue, achieving a mean Dice score of 0.90, a mean intersection-over-union of 0.82, and a mean cell-level Spearman correlation of 0.79 against explicitly computed cross-method consensus maps. We demonstrate that the predicted consensus maps localise weakly supported regions while preserving consensus-based rankings and identifying low-consensus cells for manual review. These results show that MARC closely approximates explicit cross-method consensus without multi-method inference and therefore has the potential to facilitate robust, consensus-aware evaluation of cell segmentation in large-scale SST studies.
[CV-155] UniqueShip: Mitigating Data Leakage in Acoustic Ship Classification Benchmark Datasets
链接: https://arxiv.org/abs/2609.13659
作者: Connor Hashemi,Trevor Stout,Anthony Hoogs,Jason Parham
类目: ound (cs.SD); Computer Vision and Pattern Recognition (cs.CV); Audio and Speech Processing (eess.AS)
备注: 9 pages paper, 4 pages supplementary at the end, accepted/will be published at Oceans 2026. Note that while this is underwater acoustics, we process it using computer vision models and methods (i.e. spectrograms)
Abstract:Underwater Acoustic Target Recognition (UATR) of ships is well-suited for machine learning, yet its progress is hindered by the lack of large, diverse, and publicly available labeled datasets. In this work, we introduce UniqueShip, a machine learning-ready benchmark dataset for UATR applications sourced from the open Ocean Networks Canada (ONC) repository. Unlike previous datasets, we explicitly control for “data leakage” between the training and evaluation sets to ensure more reliable and generalizable model evaluation that does not encourage the model to memorize individual ships. We demonstrate that typical, random data partitioning in two prominent UATR datasets leads to falsely optimistic test performance, increasing accuracy by 10-48 percentage points compared to our more careful partitioning. Ablations on UniqueShip further show that doubling the number of unique vessels improves accuracy by 2.4-2.6 percentage points, while doubling total audio duration improves only by 0.8-1.3 points, indicating that vessel diversity should drive dataset curation more than total hours. We provide baselines with convolutional and transformer backbones, and analyze how ship metadata correlates with classification performance, finding that individual vessel characteristics predict classification difficulty far better than distance to the hydrophone alone. Overall, UniqueShip contains 2,460 hours of ship-radiated audio from 4,218 unique vessels (3,437 hours including background). We publish the dataset, code, and easy-to-download splits at this http URL to foster further UATR research.
[CV-156] Multimodal Foundation Models Adaptation based on Domain-Aware Relaxed Orthogonal Subspace for Remote Sensing
链接: https://arxiv.org/abs/2609.13654
作者: Han Luo,Ruoyu Yang,Yinhe Liu,Yanfei Zhong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Pretrained foundation models (FMs) have achieved remarkable success in computer vision, yet their high fine-tuning cost limits practical deployment. Parameter-efficient fine-tuning (PEFT) methods such as Low-Rank Adaptation (LoRA) improve efficiency by constraining updates to a predefined low-rank subspace. However, when applied to remote sensing tasks with substantial domain shifts, the fixed subspace is constructed without observing the downstream activation distribution and can therefore provide a poor coordinate system for adaptation, a phenomenon herein termed subspace mismatch. To address this issue, a unified framework is introduced, termed Domain-aware Relaxed Orthogonal Subspace adaptation (DROS), which reformulates low-rank adaptation as data-conditioned subspace learning and flexible subspace adaptation. Specifically, the weight decomposition is conditioned on second-order activation statistics estimated from the downstream training distribution, so that the initialization reflects the feature geometry actually induced by the remote-sensing data, followed by flexible geometric transformations enabled by a relaxed orthogonal parameterization. Furthermore, the framework is extended to multimodal settings (MM-DROS) by sharing transformation structures across modality-specific subspaces, facilitating efficient cross-modal interaction. Extensive experiments on multiple remote sensing benchmarks demonstrate that DROS achieves state-of-the-art performance, even surpassing full fine-tuning, without additional inference overhead.
[CV-157] Compositional Shift Algebra: Extrapolating Mixed Robot Shifts Without Mixed Finetuning
链接: https://arxiv.org/abs/2609.13651
作者: Jinting Hang,Zhenhui Cai
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Robot deployments rarely change one mechanism at a time: cameras, action interfaces, and physical dynamics often shift together. Prior adaptation recipes either finetune a new model for every mix or attempt to select which module to update. We instead learn shift operators on a modular stack z=E(o), a=g(z,u), z’=f(z,a) and compose them. Compositional Shift Algebra (CSA) fits single-factor observation, policy, and dynamics operators from exact-reset probes, then extrapolates held-out mixed shifts by operator composition—without mixed-shift finetuning. On ManiSkill StackCube, residual CSA matches an oracle mixed inverse on held-out mixes (success 1.0 over 10 seeds) while beating best-single / zero-shot / parameter-average baselines by approx67 pp. RGB-D vision-in-the-loop composition remains near oracle and far above non-compositional arms; a delay commutator stress shows ordered necessity for policy timesdelay. On a second task (PickCube), residual CSA again reaches compose 1.0 vs. 0.33 non-compositional (n=10), and an L1 vision controller without privileged cube/goal poses or grasp flags in the control loop retains compose 0.95 vs. 0.00. Main-track upgrades freeze PushCube (+33 pp), PegInsertion joint8 / pose7 EE (+67 pp each), and thin BC under frozen CSA (+67 pp); deeper BC and fair adapt baselines still need compose (+67 pp each), vision-localized BC needs compose (+56 pp), and delay favors ordered/few-shot deploy. We report Intervention-Gated Adaptation as a negative control.
[CV-158] From Advertised Improvements to Measured Capabilities: Evaluating ChatGPT Images 2.5 on Forgery Tasks
链接: https://arxiv.org/abs/2609.13617
作者: Ankit Raj,Yuxin Zhang,Kidus Zewde,Tommy Duong,Jiaqi Gan,Xingyu Shen,Yuchen Zhou,Huaiyu Guo,Siyu Zhang,Simiao Ren
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, 6 figures. Ancillary files include the scored manifest for the in-the-wild sample and an arena snapshot
Abstract:We evaluate whether the improvements advertised for ChatGPT Images 2.5 translate into better performance on forgery tasks with predetermined answers. We compare its Flare and Sunburst API models with GPT-Image-2 re-run in the same week, using receipt-field edits, repeated editing, product placement and fine-print rendering. After image registration, Flare and Sunburst show fewer OCR-detected changes to surrounding receipt text (31.7% and 31.2% versus 44.2% for both GPT-Image-2 baselines), mainly on CORD receipts, without a detectable improvement in target-field correctness. Flare retains fewer earlier edits on CORD receipts, while photo-edit sequences provide little separation between models. Product codes are more often legible with Images 2.5, alongside larger product placement; the analyses do not establish a fidelity gain independent of size. Fine-print improvements remain unresolved below the OCR reliability limit. Refusals are rare and localisation is weak in both generations. At a fixed detection threshold, Community Forensics flags 68.6% of controlled Images 2.5 images averaged across cells, versus 35.9% of self-reported images posted online. These results motivate task-specific evaluation of advertised capabilities and defences, with explicit limits on what automatic checks can establish.
[CV-159] From Vision to Harvest: Benchmarking Vision-Language Models for Multi-Arm Robotic Fruit Harvesting
链接: https://arxiv.org/abs/2609.13606
作者: Vrishan Inukollu,Adyan Zaman,Anvi Kudaraya,Carlos Lazcano,Yuankai Zhu,Stavros Vougioukas,Xiaofan Yu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multi-arm robotic harvesting offers a promising path to improve harvesting efficiency and reduce reliance on manual labor. However, practical deployment remains challenging because the system must generalize across diverse environments while efficiently coordinating multiple arms in a shared workspace. Existing methods often require substantial data collection in target environments or rely on simplifying assumptions that limit planning quality. In this work, we introduce the first comprehensive benchmark for evaluating pretrained Vision-Language Models (VLMs) on zero-shot multi-arm fruit harvesting planning. Our benchmark uses real-world apple and citrus orchard images and compares a VLM-based planning pipeline with a traditional perception-and-planning pipeline. The VLM pipeline directly generates harvesting sequences and waypoints for each arm, while a lightweight trajectory verifier checks for collisions. Our results show that frontier VLMs can generate effective multi-arm harvesting plans zero-shot, but a practical deployment remains limited by accurate 3D waypoint generation and collision-aware coordination. These results highlight both the promise and current limitations of pretrained VLMs for multi-arm robotic harvesting.
[CV-160] owards Practical Precision Agriculture: Real-Time Fruit Detection and Video Analytics on Embedded Edge Hardware
链接: https://arxiv.org/abs/2609.13551
作者: Ivica Dimitrovski,Vlatko Spasev,Ivan Kitanovski,Petre Lameski,Dane Boshev
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Static-image benchmarks do not capture the computational and temporal requirements of practical orchard video analytics. This study presents an end-to-end framework for real-time fruit detection, tracking, and counting on the NVIDIA Jetson Orin Nano Super. A lightweight YOLO26s detector is trained independently on four public datasets representing apples, mangoes, blueberries, and strawberries under a common protocol. The models are deployed on embedded platform using PyTorch and TensorRT at FP32, FP16, and INT8 precision. APPLE MOTS is then used for temporal video analytics because it provides orchard sequences with persistent fruit identities, enabling evaluation of multi-object tracking and unique-fruit counting. The selected FP16 TensorRT detector is integrated into an NVIDIA DeepStream pipeline combining hardware-accelerated decoding, ByteTrack tracking, and motion-aware line-crossing analytics. Across the four detection tasks, mean test mAP@50:95 ranges from 0.4957 to 0.8656. On the Jetson, TensorRT FP16 achieves 66.76-74.56 images/s at 13.41-14.98 ms prediction latency, while reducing mAP@50:95 by only 0.0020-0.0054 and gross energy consumption by approximately 64-66% relative to PyTorch FP32. The complete detector-tracker-analytics pipeline reaches 44.96-54.11 FPS and sustains the configured 30-FPS input rate without output-frame loss. On held-out orchard video sequences, HOTA ranges from 0.345 to 0.538, event-level counting F1 from 0.611 to 0.803, and relative count error from 6.2% to 51.6%. Performance varies across acquisition geometries: near-lateral row viewing yields the most stable tracking and counting, whereas forward traversal remains association- and recall-limited despite spatially adaptive counting geometry. These results show that practical edge-based fruit monitoring requires efficient detection and acquisition geometries that support reliable temporal association.
[CV-161] RIGOR: Rig-Informed Geometry for Omnidirectional Reconstruction
链接: https://arxiv.org/abs/2609.13504
作者: Tingjun Huang,Dmitry Rudshin,Mathieu Meyer,Pietro Bonazzi,Marc Pollefeys,Emilia Szymańska
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recent developments in feed-forward 3D reconstruction resulted in models which can recover dense scene representations and camera motion solely from an image stream. However, such predictions are prone to becoming inconsistent over long trajectories, specifically in demanding environments with repetitive structures, weak textures and dynamic objects or people. One way to mitigate those challenges is to use an omnidirectional camera, which provides wide spatial coverage and captures richer visual information. Yet, the majority of models do not offer support for 360-degree imagery or require additional fine-tuning. To bridge these two aspects, we present RIGOR: a large-scale reconstruction pipeline for gravity-aligned omnidirectional videos that retains a frozen feed-forward perspective backbone and exploits each panorama as a four-view virtual rig. The rig structure is used to detect and repair locally inconsistent predictions, to retrieve loop closures through cyclic four-view consensus, and to geometrically verify candidate revisits before global optimization. Verified constraints drive a Sim(3) pose graph that corrects accumulated rotation, translation, and scale drift along the sequence. We demonstrate that the proposed consistency mechanisms improve both trajectory accuracy and reconstructed geometry over a feed-forward baseline on challenging construction-site sequences. The code is made available under this link: this https URL.
[CV-162] ConeGaussian: Anti-Aliased Gaussian Ray-Tracing for Generic Central Cameras
链接: https://arxiv.org/abs/2609.13397
作者: Deheng Zhang,Letian Shi,Runyi Yang,Zhendong Li,Lei Sun,Kanzhi Wu,Ajad Chhatkuli,Danda Pani Paudel,Luc Van Gool
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:In rendering, a camera is a sampling operator that maps each finite pixel to a bundle of rays. Different camera models change the geometry of this bundle, thus making a unified and faithful rendering formulation challenging. Consequently, Gaussian ray tracing supports generic cameras (with optical center) through their inverse ray mappings, yet typically reduces every pixel to a single center ray. This ignores the camera-dependent pixel footprint, causing aliasing under minification, while unconstrained Gaussians expose unsupported frequencies under magnification. We present ConeGaussian, a camera-model-agnostic anti-aliasing framework for Gaussian ray-based rendering. Instead of defining the pixel filter on a camera-specific image plane, ConeGaussian constructs an anisotropic footprint directly from neighboring rays produced by the camera’s native inverse mapping. We derive a closed-form response under a locally linear, depth-local, moment-matched approximation of the finite pixel footprint, while the same geometry defines a per-Gaussian training-frequency floor. Notably, by construction, our filtering principle can be used unmodified across calibrated central camera models and multiple Gaussian ray-rendering backbones. Additionally, unlike in mip-splatting, our scene-space frequency floor and filtering enable trivial composition at render time, allowing us to remove excess blurring. On pinhole and strongly distorted fisheye captures, ConeGaussian consistently improves two distinct ray-based backbones, by up to 4.3 dB at 1/8 resolution, and reduces fisheye LPIPS by 30% where perspective screen-plane footprint formulations are not directly applicable.
[CV-163] ProtoCAM: Interpretable Few-Shot Mask-Guided Prototypical Learning for Breast Lesion Classification in Ultrasound Imaging
链接: https://arxiv.org/abs/2609.13340
作者: Ashkan Ebadi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 11 pages, 3 figures
Abstract:Breast ultrasound imaging plays an important role in the early detection and diagnosis of breast cancer, particularly for patients with dense breast tissue. However, developing reliable deep learning models for ultrasound analysis is challenging due to limited annotated medical data and the need for interpretable predictions. To address these challenges, this paper proposes ProtoCAM, an explainable few-shot learning framework for breast lesion classification that integrates mask-guided feature encoding, prototypical metric learning, and gradient-based visual explanations. The proposed approach leverages lesion masks to guide feature extraction and constructs class prototypes within an embedding space to enable robust classification under limited training samples. The framework was evaluated on the BUSI dataset using a stratified group k-fold cross-validation protocol to prevent patient-level data leakage. Experimental results demonstrate ProtoCAM’s high performance in low-data scenarios. In a 3-way 5-shot setting, the proposed method achieves a macro F1-score of 0.910, representing a substantial improvement over standard supervised CNN models. Among the evaluated backbone networks, ResNet18 achieved the best performance, reaching a macro F1-score of 91.65% under a 15-shot configuration, providing interpretable insights into the classification decisions. These results highlight the potential of explainable few-shot learning frameworks for reliable computer-aided breast cancer diagnosis in data-scarce medical imaging environments.
[CV-164] Global-Local Contextual Progressive Expansion Network for Martian Landslide Segmentation in Multimodal Remote Sensing Imagery
链接: https://arxiv.org/abs/2609.13332
作者: Leo Thomas Ramos,Sidike Paheding,Abel A. Reyes-Angulo,Rajaneesh A.,Sajinkumar K.S.,Angel D. Sappa,Thomas Oommen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 21 pages, 12 figures
Abstract:Automated landslide segmentation on Mars is one of the important tasks for understanding its surface processes, and all will aid in future space exploration. However, it remains a relatively underexplored open challenge because landslide morphology is highly variable, foreground regions are often sparse or irregular, and orbital observations combine heterogeneous spectral and topographic cues. In this context, this work investigates the capability of deep learning to address Martian landslide segmentation through an extensive assessment of modern neural segmentation models. To the best of our knowledge, this is the first study to provide such a comprehensive exploration in this domain. We further propose TransCPLES, a U-shaped network that couples Contextual Progressive Layer Expansion feature extraction with Transformer-based contextual reasoning, enabling the model to capture local geomorphic patterns and broader spatial dependencies for more reliable landslide delineation. Experiments on MMLSv2, a seven-band multimodal Martian landslide dataset, show that TransCPLES achieves the best overall performance when evaluated on geographically distinct samples, with consistent delineation across different landslide extents, stable foreground discrimination, and a favorable balance between accuracy and computational cost compared with several state-of-the-art convolutional, attention-based, and Transformer-based segmentation models. With this work, we hope to provide a useful reference and encourage further research and development in deep learning for planetary remote sensing. Code will be available after publication.
[CV-165] Pedestrian Crossing Intent Classification From Event-Based Vision Using Convolutional Spiking Neural Networks With Temporal Augmentation
链接: https://arxiv.org/abs/2609.13328
作者: Henok Teklu,Mustafa Sakhai,Maciej Wielgosz,Matej Mertik
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 8 figures, 6 tables
Abstract:Anticipating whether a pedestrian will cross the road is safety-critical for autonomous vehicles, requiring real-time inference under challenging conditions including motion blur, high dynamic range, and class imbalance. Conventional frame-based deep networks process redundant RGB data at fixed frame rates, limiting their temporal resolution and energy efficiency. In this work we present an end-to-end pipeline that (i) converts real-world driving footage from the Joint Attention in Autonomous Driving (JAAD) dataset into synthetic dynamic vision sensor (DVS) event streams using the v2e simulator, (ii) augments training with the CARLA-simulated DVS sequences of the DVS-PedX dataset under both normal and adverse weather conditions, and (iii) trains a novel convolutional spiking neural network (Conv-SNN) with clip-consistent DVS augmentation to classify pedestrian crossing intent as binary: crossing or non-crossing. We detail all architectural decisions, the exact leaky-integrate-and-fire neuron dynamics with surrogate-gradient learning, the class-balanced loss formulation, JAAD oversampling at 6x, and a 70/15/15 stratified splitting protocol. The trained model achieves 95.83% accuracy and F1 = 0.9695 on the JAAD DVS test set, 97.79% accuracy and F1 = 0.9478 on normal CARLA DVS, and 94.78% accuracy and F1 = 0.8369 on adverse-weather CARLA DVS, all from a 1.07M-parameter architecture trained on CPU. Compared to prior frame-based approaches on JAAD, our method closes or surpasses the reported accuracy while operating natively on sparse temporal representations. We include a thorough analysis of the convergence behaviour across all 15 training epochs, domain transfer characteristics, and a quantitative comparison with representative related work.
[CV-166] Attention-DP3: Spatially Object-aware 3D Diffusion Policy via Geometry-aligned Attentional Conditioning ECCV2026
链接: https://arxiv.org/abs/2609.13318
作者: Changbo Yan,Zhongbo Zhang,Zaibin Zhang,Yifan Wang,Lijun Wang,Huchuan Lu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV 2026
Abstract:3D point-cloud observations are inherently ambiguous in complex, cluttered manipulation scenes, where target objects may be partially occluded or tightly intermingled with visually similar distractors. As a result, standard 3D diffusion policies often struggle to localize and exploit task-relevant geometry as scene complexity grows. We propose \textbfAttention-DP3, a spatially object-aware 3D diffusion policy that injects object-level geometric cues via attention while keeping the DP3 diffusion backbone unchanged. Our pipeline performs open-vocabulary 2D segmentation on RGB images, then lifts predicted target masks into 3D using calibrated camera geometry to obtain object-centric geometric priors. We incorporate these cues through Tri-field Attentional Conditioning, which constructs three complementary fields: (i) a targetness field to anchor the target object, (ii) an intra-target saliency field to emphasize task-relevant geometry within the target, and (iii) a backgroundness field to suppress distractors and clutter. Experiments on Adroit, DexArt, MetaWorld, and the real-world SO101 platform show consistent improvements over DP3, achieving state-of-the-art performance across benchmarks. Notably, as distractor objects increase, DP3 drops sharply, whereas Attention-DP3 remains stable and outperforms DP3 by up to 31% under heavy clutter. The code is publicly available at this https URL.
[CV-167] ask-Based CT Protocol Optimization Using Reinforcement Learning and Virtual Imaging Trials
链接: https://arxiv.org/abs/2609.13309
作者: Jiaqi Zou,David Fenwick,Vahid Tarokh,Nicholas Felice,Jayasai Rajagopal,Anuj Kapadia,Ehsan Samei,Navid NaderiAlizadeh,Ehsan Abadi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 11 pages, 7 figures, 4 tables
Abstract:Protocol optimization in computed tomography (CT) aims to improve diagnostic image quality while reducing radiation dose, but the interdependence of acquisition and reconstruction parameters makes exhaustive testing impractical. We propose a virtual imaging trial framework with reinforcement learning for efficient CT protocol optimization. Sixty-three computational human models with liver lesions were imaged using a validated CT simulator across 468 combinations of acquisition and reconstruction parameters, including tube voltage, tube current, reconstruction kernel, slice thickness, and pixel size. The optimization objective balanced liver lesion detectability, quantified by detectability index d-prime, against radiation dose. A Proximal Policy Optimization agent was trained and conditioned on patient-specific CT localizer embeddings derived from a pretrained vision transformer. On held-out patients, evaluating only 8 protocols per patient, about 2% of exhaustive testing, recovered 98.2% of the exhaustive-search oracle objective. With no patient-specific simulation, surrogate scoring alone achieved 89.7% recovery. Conditioning on the localizer improved zero-simulation recovery by 10.7 percentage points over the localizer-blind policy (paired 95% CI 2.9-19.5; p=0.02). These results show that the proposed framework can substantially reduce exhaustive protocol testing while enabling task-based, dose-aware protocol selection before the diagnostic scan.
[CV-168] GroundBench: A Factorized Counterfactual Benchmark for Locating VLM Affordance Failures
链接: https://arxiv.org/abs/2609.13308
作者: Sarthak Sattigeri
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注:
Abstract:A companion evaluation found that naming the target part in a manipulation prompt increased action accuracy by 0.32-0.63 across eight vision-language models, with no model outperforming a constant baseline until the part was named. However, naming the part supplies information that a real system must infer, confounding visual grounding, mechanical reasoning, and category-to-action association. We introduce GroundBench, a diagnostic benchmark that separates these explanations through six branch-and-merge conditions, each adding a controlled information bundle, and a counterfactual re-ask targeting a real alternate part visible in the same image. Across three OpenAI models and 1,068 predictions, supplying the target region without its identity leaves action accuracy at or below the 0.53 majority baseline (0.26, 0.26, and 0.53), although the models largely reproduce the supplied region. Supplying identity without location instead yields 0.74, 0.68, and 0.68. Every above-baseline gain in this curated set occurs where the supplied part category itself determines the action. A no-vision control leaves GPT-5’s scores unchanged or improved, providing evidence consistent with substantial category-to-action association. GPT-4o mini declines on one condition, so this interpretation is not universal. Adding joint type and motion axis does not improve accuracy across six model-stratum comparisons. On 74 counterfactual pairs from 32 objects, GPT-5 achieves 0.86 pair-weighted compliance with a 0.07 shortcut rate but fails all observed push-to-lift-vertical cases. GroundBench identifies which supplied information changes affordance behavior and tests whether apparently grounded performance can be reproduced through textual shortcuts.
[CV-169] HGSQ: Heatmap-Guided Sparse Query Detector for Real-Time Aerial Small Object Detection
链接: https://arxiv.org/abs/2609.13306
作者: Yangchen Zeng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 3 figures and 7 tables
Abstract:Real-time aerial small object detection is an important visual signal and image processing problem, requiring a detector to preserve fine-grained localization while avoiding redundant computation on large background regions. This paper focuses on this deployment-oriented aerial/UAV setting rather than claiming a universal detector for all object detection scenarios. Existing Transformer-based detectors provide strong global modeling, but their dense query initialization and multi-layer decoder still spend substantial computation on background tokens, which is inefficient when small objects occupy only sparse image regions. To address this problem, this paper proposes HGSQ, a Heatmap-Guided Sparse Query Detector for real-time aerial small object detection. HGSQ uses a lightweight Heatmap Budget Predictor (HBP) to predict a foreground budget map in a single forward pass. The predicted heatmap is then used by three fixed components: Heatmap-Guided Sparse Query Selection (HSQS), which initializes decoder queries from high-confidence foreground positions; Heatmap-Gated Lite Snake Convolution (HGLSConv), which performs local shape refinement only on heatmap-activated small-object regions; and Adaptive Query-Decoder Budgeting (AQDB), which adjusts the query budget and decoder depth according to the estimated object density. Unlike post-hoc heatmap generation, HGSQ treats the heatmap as a real-time computation budget rather than a visualization map during deployment. Experiments on NWPU VHR-10 and VisDrone2019 show that HGSQ achieves 95.10 mAP50 on NWPU VHR-10 and 54.8 mAP50 on VisDrone2019, while reducing GFLOPs to 48.6 and running at 96.0 FPS on an RTX 4070 under our TensorRT FP16 deployment protocol.
[CV-170] Adaptive Conformal Redistribution for Inter-class Transitional Uncertainty in Medical Image Classification
链接: https://arxiv.org/abs/2609.13303
作者: Saibal Ghosh,Samarup Bhattacharya,Sanjoy Kumar Saha,Umapada Pal,Tapabrata Chakraborti
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Computation (stat.CO); Machine Learning (stat.ML)
备注:
Abstract:Medical image classification is frequently complicated by transitional categories whose feature distributions overlap those of adjacent classes, producing ambiguous decision boundaries. Conformal prediction returns uncertainty-aware prediction sets, but these are not directly actionable in clinical screening, where a single decision is required. This work proposes adaptive conformal redistribution (AdaConRed), a label-free post-conformal decision rule that converts ambiguous prediction sets into refined class assignments. A five-stage pipeline is developed. Vision-language generative augmentation addresses minority-class scarcity; a frozen DermFoundation encoder provides embeddings; a lightweight multi-layer perceptron performs classification; an entropy-modulated, margin-aware nonconformity score constructs adaptive prediction sets; samples predicted as transitional with multi-label sets are reassigned to the most probable alternative class within the set, using only model outputs at inference. Evaluation uses the OSCC oral lesion and ISIC skin lesion benchmarks at a miscoverage level of 0.2. On the 3-class OSCC benchmark, overall accuracy improves from 73.54% to 77.38%, with oral cancer accuracy rising from 64.29% to 82.14% and benign accuracy from 56.57% to 70.20%. Reassignment of transitional samples reduces OPMD accuracy from 84.78% to 80.16%, consistent with the asymmetric cost of missed malignancy. On ISIC, overall accuracy improves from 85.83% to 87.19%, melanoma accuracy rising from 66.04% to 68.34%. AdaConRed outperforms LAC, APS and RAPS under an identical backbone and redistribution rule. Conformal prediction can be extended beyond uncertainty quantification toward actionable decision support where transitional disease categories are present, with gains concentrated in the clinically critical malignant categories. Code repository: this https URL.
[CV-171] Variational Template Matching with Statistical Fusion for Anomaly Detection in Patterned Structures
链接: https://arxiv.org/abs/2609.13298
作者: Qinwu Xu,Yifan Jiang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Anomaly detection in structured images is challenging in small-data settings where deep learning approaches are costly or impractical. Classical template matching is simple and interpretable but lacks robustness to geometric variations such as scale, rotation, and perspective. We propose a variational template matching framework that represents anomaly templates as a family of transformed instances and performs detection via normalized cross-correlation over this transformation space. To further improve robustness, we introduce a density-based statistical anomaly score derived from local intensity distributions using kernel density estimation (KDE). This produces a smooth representation that captures distributional concentration and tail behavior more robustly than histogram-based methods. The structural and statistical signals are integrated through a unified fusion formulation, enabling complementary modeling of geometric similarity and distributional deviation. Experiments on biological cell images demonstrate that the proposed method outperforms classical baselines and achieves competitive performance with ResNet-50 under a fully training-free setting, while providing explicit localization. The approach offers an efficient, interpretable, and practical solution for anomaly detection in structured image domains. Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2609.13298 [cs.CV] (or arXiv:2609.13298v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.13298 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-172] Harnessing Image Question Dependence for Better VLM Test-time Reinforcement Learning
链接: https://arxiv.org/abs/2609.13296
作者: Xinrui He,Ting-Wei Li,Junting Wang,Mengting Ai,Xinyu He,Hanghang Tong,Jingrui He
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Test-time reinforcement learning can adapt vision-language models (VLMs) to unlabeled target data, but its effectiveness is fundamentally limited by the reliability of self-generated learning signals. To assess the reliability of consensus-based learning signals, we analyze VLM test-time reinforcement learning across diverse VQA datasets and model sizes, revealing two limitations. First, gains from consensus-based test-time training largely come from answer normalization rather than content correction. Second, many initial VLM responses are incorrect due to the model’s limited ability to jointly use the image and the question; consensus rewards derived from these outputs may preserve the resulting grounding errors rather than correct them. Motivated by these, we propose TTIQ, a test-time reinforcement learning framework that harnesses image-question dependence for better vlm adaptation. TTIQ teacher-forces each sampled response under the original image-question pair and its image- and question-ablated variants, using the resulting token-level likelihood changes to estimate dependence on each input. It combines image and question dependence with calibrated confidence to construct a response-level reward that favors jointly grounded responses, and uses the token-level signals to assign greater positive policy credit to tokens supported by both inputs. This design favors responses that are jointly grounded in the image and the question and sufficiently confident, rather than merely popular. Experiments across eight VQA datasets and multiple VLM sizes show that TTIQ achieves the best average performance at every model scale. It further generalizes across VLM families, while models trained on one dataset improve performance on unseen datasets without further training.
[CV-173] VectorHarness: Recovering Editable Relation-Preserving Structure from Scientific Graphics
链接: https://arxiv.org/abs/2609.13294
作者: Jiahao Tang,Yiren Song,Alex Jinpeng Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Converting scientific graphics into editable representations remains a challenging problem for image-to-code generation because of their heterogeneous elements and complex layouts. Recent multi-agent reconstruction systems have advanced this line of work, but often follow a copy-paste paradigm: the reconstructed image closely resembles the original, while complex regions remain effectively uneditable. We instead formulate a different objective, raster-to-authoring reconstruction, which aims to recover an authoring representation that supports native, customized editing rather than mere visual replication. To this end, we present VectorHarness, a multi-agent framework for raster-to-authoring reconstruction that recovers heterogeneous components using type-appropriate native representations. Text, formulas, shapes, connectors, icons, charts, and tables are reconstructed as natively editable objects, while intrinsically image-based regions remain raster content. To systematically evaluate reconstruction quality, we introduce VectorHarness-Bench, which jointly assesses rendering fidelity, raster fallback coverage, executable object edits, and relation-preserving edits. Experiments show that VectorHarness improves executable edit success and relation preservation, reduces avoidable raster fallback, and maintains high visual fidelity across heterogeneous graphics.
[CV-174] Reliable Egocentric Action Anticipation via Temporal Reliability Suppression and Compositional Graph Decoding ECCV2026
链接: https://arxiv.org/abs/2609.13293
作者: Mahsa Mohammadi,Sareh Rowlands
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 4 figures. Accepted at the ECCV 2026 Workshop on Towards Real-time Multimodal Contextual Assistants (Wearable AI Workshop)
Abstract:Wearable action anticipation systems must remain reliable despite missing frames, masking, and sensor noise, yet existing egocentric anticipation methods largely assume clean observations. We identify two complementary failure modes under temporal corruption: unreliable temporal evidence during encoding and implausible, low-support verb-noun compositions during decoding. We address them with a lightweight framework combining Temporal Reliability Suppression (TRS) and Robust Verb-Noun Graph (RVG) decoding. TRS predicts a per-frame suppression score from the projected input embedding and uses it as a learned key-side attention penalty at every encoder block and to derive reliability-weighted temporal pooling. RVG re-ranks verb-noun pairs using a PMI-based compatibility graph constructed from training labels. Under corruption-augmented training, TRS+CA+RVG reaches 29.1% average corrupted accuracy and 88.2% relative robustness across six corruptions, including three mechanisms absent during training, while reducing rare verb-noun predictions from 15.4% to 1.1%. Multi-seed and diagnostic experiments show that TRS responds to synthetic masking; shuffled-graph and frequency-only controls further indicate that RVG gains depend on genuine pairwise compatibility rather than marginal-frequency effects alone.
[CV-175] arget-Checked Reliability Score Refinement for Video Question Answering
链接: https://arxiv.org/abs/2609.13288
作者: Guoxiang Ren,Rohitash Chandra
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 8 figures
Abstract:Video-language models can answer multiple-choice questions with high confidence yet be wrong. We study whether answer-level reliability scores can be improved under target shift without retraining the models or changing their answers. We collect option-probability lists from three fixed video-language models under four deterministic video samplings and represent cross-view changes and cross-model agreement as a response graph. Using a labeled target pilot, we compare the original score, defined as the probability assigned to the chosen answer, with a histogram-based gradient-boosting (HGB) score trained on the development datasets and a regularized logistic-regression score trained on the target pilot. A candidate replaces the original score only when repeated video-level checks indicate a positive, stable improvement. We develop this rule on public VideoQA benchmarks and Video Hallucination Diagnosis (VHD), a controlled diagnostic dataset for shared high-confidence errors. Ranking quality is measured by the area under the risk-coverage curve (AURC), where lower is better. On a held-out 963-question HERBench split, the method reduces mean AURC across the three models by 16.64% (95% confidence interval (CI), 12.12 to 22.61%); the smallest model-level gain is 11.39%. On a separate held-out 911-question Perception Test split, the mean reduction is 18.87% (95% CI, 15.43 to 22.14%). For InternVL3.5, the target check retains the original scores. Using the same outputs, the method outperforms seven training-free baselines in mean AURC on both datasets. It also improves AUROC, reduces calibration error, and lowers the error rate at 50% coverage by 6.50 and 6.58 percentage points.
[CV-176] LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents
链接: https://arxiv.org/abs/2609.13287
作者: Zhangxuan Gu,Haoxing Chen,Qi Qin,Yi Xin,Kai Gan,Lin Liu,Long Cui,Xiaomei Wang,Beitong Zhou,Yunzhu Zhang,Zhengwen Zeng,Changlong Gao,Weizhi Chen,Rongchao Zhang,Haoyuan Wu,Shuheng Shen,Changhua Meng,Weiqiang Wang,Jianguo Li,Zhenzhong Lan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion large language models (dLLMs) achieve high decoding efficiency through block-parallel, arbitrary-order generation, making them attractive for latency-sensitive applications. GUI agents represent a natural testbed for this paradigm, as they must repeatedly perceive screen states and emit structured, spatially grounded actions in real time. However, whether dLLMs can be extended into capable multimodal GUI agents while preserving their parallel decoding advantage remains an open question. We present LLaDA-UI, a 16.7B-parameter MoE-based, block-wise diffusion vision-language GUI agent. LLaDA-UI follows a two-stage training pipeline: general multimodal pre-training aligns a native-resolution vision encoder with the LLaDA2.0-mini-base diffusion language backbone, followed by GUI-agent supervised fine-tuning on diverse mobile, desktop, web, and grounding data. Across widely adopted grounding benchmarks and navigation benchmarks spanning multiple platforms, LLaDA-UI substantially outperforms Qwen2.5-VL-7B and surpasses Qwen3-VL-8B on four of six reported GUI benchmarks. These results establish block-wise diffusion as a practical generative paradigm for multimodal GUI agents.
[CV-177] Multimodal-Multiresolution Foundation Model for Lunar Remote Sensing
链接: https://arxiv.org/abs/2609.13283
作者: Paolo Fraccaro,Gabby Nyirjesy,Daniela Szwarcman,Himanshu Patil,Vishal Gaur,Rohit Lal,Rachel A. Slank,Geoffrey Dawson,Hiyam Debary,Michael K. Barker,Andrew Annex,Vishnu Viswanathan,Zachary Morse,Ethan I. Schaefer,Nikolaos Dionelis,Ankur Kumar,Campbell D. Watson,Manil Maskey,Rebekah I. Dawson-Rigas,Juan Bernabé-Moreno,Rahul Ramachandran,Sujit Roy
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:We present a multimodal foundation model for lunar remote sensing, pretrained from scratch on SomBench, a geographically partitioned corpus of nearly two million co-registered tile bundles spanning 11 modalities at two spatial scales (1 m/pixel and 100 m/pixel). The model adapts the TerraMind masked-token architecture with two lunar-specific extensions: acquisition geometry is provided as explicit context, and meter- and hundred-meter-scale tiles are trained jointly so that a single set of weights covers both resolutions. FlexiViT patch embeddings allow adaptation to different patch sizes without retraining, while modality-wise inputs enable flexible multimodal fine-tuning. Qualitative generation experiments suggest the model learns meaningful cross-modal correspondences, including terrain derivatives from elevation and illumination-consistent reflectance from geometry. We evaluate on four benchmarks: crater detection at WAC and NAC scales, irregular mare patch (IMP) segmentation, and polar ice prospectivity regression. Across tasks, the pretrained model matches or outperforms ImageNet-pretrained baselines and an architecturally identical random-init control. On multimodal ice prospectivity regression, pretrained variants achieve the best results, while the random-init model outperforms most baselines, suggesting gains arise from both the architecture and pretraining. Label efficiency is notable for WAC crater detection, where the pretrained model trained on 50% of the data exceeds the strongest ImageNet baseline trained on the full dataset. Among adaptation strategies, LoRA matches or surpasses full fine-tuning on crater detection and IMP segmentation while using far fewer trainable parameters, whereas full fine-tuning performs best for ice prospectivity regression. We release the pretrained checkpoint, benchmark datasets, and fine-tuning code to support reproducible lunar AI research.
[CV-178] GaugeDefect: Detecting Surface Anomalies by Curvature of Feature Transport
链接: https://arxiv.org/abs/2609.13282
作者: Yefan Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the 9th Chinese Conference on Pattern Recognition and Computer Vision (PRCV 2026)
Abstract:Industrial anomaly localization has advanced rapidly with feature-based, reconstruction-based, and distillation-based methods. Most of these methods score a region by asking how unusual its local appearance or feature representation is with respect to normal training images. This is a strong and practical formulation. In this work, we study a complementary geometric cue for cases where an abnormal region may still contain locally plausible visual features. Thin scratches, small dents, and disrupted repeated patterns often do not make every local patch individually abnormal; instead, they disturb how nearby features vary and connect across the surface. We propose GaugeDefect, a geometric method for surface anomaly localization based on the curvature of feature transport. Given a feature lattice, we estimate a local feature frame at each node and compute orthogonal transports between neighboring frames. The accumulated transport around a small closed loop gives a holonomy matrix, whose deviation from identity measures feature-transport curvature. After calibration on normal training images, unusually large curvature indicates a local inconsistency in the feature field. The curvature here is not the physical curvature of the inspected object, but a representation-space measure of neighborhood inconsistency. This makes the method applicable to curved surfaces, textured materials, and non-planar industrial objects. Its main role is to improve localization of subtle surface disruptions, while often producing sharper responses near defect boundaries as a natural consequence of the curvature signal.
[CV-179] RxScribe Bench: A Multi-Axis Benchmark for Evaluating Vision-Language Models on Indian Outpatient Prescriptions
链接: https://arxiv.org/abs/2609.13280
作者: Somil,Vijay Saini,Vidit Verma,Riya,Aastha Batta,Vibhuti Malhotra,Chayan Khetan,Piyush Mittal,Puneet Poonia
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Prescription transcription errors are not interchangeable. A model that fabricates a drug and a model that misreads a legible dose pose very different clinical risks, yet prescription-transcription accuracy is typically reported as a single blended figure that treats the two as equivalent. We introduce RxScribe Bench, a benchmark for evaluating vision-language models on handwritten prescription digitization that decomposes performance into four axes tied to clinical severity, rather than folding everything into a single aggregated score. Given only a prescription image and an output schema, a model produces a structured record, which is then compared field by field against a human-authored ground truth of identical shape, with each field also labeled for visibility and legibility. The four axes isolate distinct failure modes, namely Correctness, Hallucination, Engagement, and Robustness. The Robustness axis withholds its hard-field results rather than reporting one when the supporting sample falls below a minimum-evidence threshold. We evaluate frontier vision-language models on real prescriptions across independent cold runs per image, and find that no single model wins across all four axes.
[CV-180] he MODA General Attribute Suite: A Four-Track Evaluation Benchmark for Fashion Attribute Extraction
链接: https://arxiv.org/abs/2609.13279
作者: Arkid Mitra(Hopit AI)
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注: 10 pages, 3 figures, 4 tables. Benchmark suite, scorers, split builders, prediction files with hashes, and model weights at this https URL
Abstract:Fashion attribute extraction is evaluated inconsistently: results are reported as single aggregate numbers across image types that pose different problems, fields that are not visible in an image are scored as ordinary negatives, and the effect of vocabulary mismatch between datasets is acknowledged but not measured. We release the MODA General Attribute Suite, a four-track benchmark that keeps these problems separate by construction. Each track (localized garment crops, catalogue product images, applicability-aware full-body photographs, and product text) carries its own frozen test set, input contract, metric, and leakage unit, and the tracks are never averaged. The protocol requires label-blind prediction, SHA-256 commitment of prediction files before any label is opened, a fail-closed scorer, and a 10,000-sample paired cluster bootstrap at each track’s natural leakage unit; promotion requires a positive interval on every track rather than a favourable mean. We release the scorers, the split builders, our own prediction files with their hashes including the runs we lose, and the MODA_NER(V) model checkpoints for three of the four tracks. The text model is not distributed; its benchmark is. We report baseline results, six interventions that did not improve them, and the limitations of the evidence.
[CV-181] SomBench: Benchmark Dataset for Advancing Machine Learning in Lunar Science
链接: https://arxiv.org/abs/2609.13277
作者: Himanshu Patil,Gabby Nyirjesy,Rachel A. Slank,Vishal Gaur,Daniela Szwarcman,Paolo Fraccaro,Nikolaos Dionelis,Michael K. Barker,Andrew Annex,Vishnu Viswanathan,Zachary Morse,Ethan I. Schaefer,Hiyam Debary,Ankur Kumar,Rohit Lal,Geoffrey Dawson,Campbell Watson,Rebekah I. Dawson-Rigas,Manil Maskey,Juan Bernabé-Moreno,Rahul Ramachandran,Sujit Roy
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Lunar orbital missions, such as Lunar Reconnaissance Orbiter, Kaguya/SELENE, Gravity Recovery and Interior Laboratory, and Lunar Prospector, among others, provide rich multi-instrument observations, but their heterogeneity in sampling, projection, and conventions limits reproducible machine learning (ML). We introduce SomBench, a unified, spatially-aligned, ML-ready lunar dataset aggregating 30+ co-registered layers from ten instruments across four missions, spanning 1 meter to 20 kilometer/pixel and covering 82 degree latitude in 90 Lunar Transverse Mercator zones with two polar stereographic caps. An image-anchored tiling pipeline yields pretraining-ready multimodal tile views with leakage-safe splits, distributed as netCDF with Parquet catalogs. An application benchmark suite spans impact processes, volcanic history, and polar volatiles. Baseline experiments with ResNet-50 and SwinV2-B models confirm that each benchmark task is learnable from the released inputs, establishing reference points for future model development.
[CV-182] CANAL: Channel-Aware Noise Allocation for Differentially Private Feature Distillation in Medical Image Segmentation
链接: https://arxiv.org/abs/2609.13271
作者: Armaghan Butt,Shuya Feng,Qing Tian
类目: Computer Vision and Pattern Recognition (cs.CV); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
备注:
Abstract:Medical image segmentation needs diverse training data, but hospitals hold complementary scans they cannot share for privacy and regulatory reasons. Knowledge distillation can bridge this gap by exporting learned feature representations instead of images, but those representations still encode patient-specific anatomy and remain vulnerable to membership-inference and feature-inversion attacks. Adding calibrated Gaussian noise restores a differential-privacy guarantee, yet three issues have been overlooked. First, prior DP feature-distillation pipelines re-sample noise at every student iteration, so each patient image is released many times and the privacy cost composes over those releases, growing by orders of magnitude. We present a sample-once-per-image release, realized by a single precomputation pass, under which each patient contributes one release. Second, uniform noise is wasteful because channels differ in task importance. Using task-gradient energy as the importance measure, we derive CANAL, a closed-form water-filling allocation that gives important channels proportionally less noise, and prove it strictly minimizes importance-weighted distortion at a fixed budget. Third, the clipping caps and importance scores that drive the allocation are themselves data-dependent, so releasing them in the clear silently breaks the guarantee. We give a DP-honest budget split that charges each to the privacy budget, so the reported epsilon is the true epsilon. Across three medical segmentation benchmarks spanning dermoscopy, colonoscopy, and ultrasound, CANAL retains more task-relevant signal than uniform noise at the same privacy budget.
[CV-183] Capability-Routed Visual Retrieval and Evidence Threading for Long-Context Document Question Answering
链接: https://arxiv.org/abs/2609.13268
作者: Amirul Rahman,Aisha Karim,Kenji Nakamura,Yi-Fan Ng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Annual reports, diligence packs, and infographic dashboards bury numbers in page images: axes, cell grids, and footnotes that OCR pipelines flatten and that page-level visual retrievers still treat as interchangeable in-context examples. We keep a frozen Qwen2.5-VL-7B-Instruct generator and a ColPali / VisRAG-Ret page index, and insert three modules. A capability-aware visual router (CAVR) tags each retrieved page as text, table, chart, layout, or mixed and mixes specialist experts before generation. Weak-to-strong page selection (WSPS) distils a frozen 7B answerability teacher into a 3B selection head so ranking is no longer a single InfoNCE score. Visual evidence threading (VET) builds layout-anchored paths of length at most three and lets the generator read the thread rather than a flat top- k list. On gold-page DocVQA / ChartQA / InfographicVQA the 7B system reaches 96.3 / 90.1 / 85.4. Under the VisRAG top-3 protocol the mean generation accuracy is 62.74 versus 59.39 for the same backbone with concatenation. On MMLongBench-Doc retrieve-then-read, F1 moves from 19.2 to 22.6 and multi-page accuracy from 16.4 to 21.2. ViDoRe nDCG@5 after WSPS reranking is 83.6, with TAT-DQA financial reports at 70.4.
[CV-184] Structure-Token Evidence-Anchored Reasoning for Scientific Chart Understanding
链接: https://arxiv.org/abs/2609.13267
作者: Alberlucia Rafael Soarez,Camila Ferreira,Daniel Kim,Mariana Costa,Alejandro Torres
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Scientific charts encode quantities in axes, legends, and geometric marks, yet large vision-language models still treat them as natural photographs. Visual in-context examples do not expose the coordinate frame; unconstrained chain-of-thought can name a plausible number that was never read from a bar. We present STEER (Structure-Token Evidence-anchored Reasoning), which freezes a Llama-3.2-Vision encoder and inserts three modules: a chart structure graph encoder (CSGE) that binds ticks, legend items, and marks; evidence-anchored step reasoning (EASR) that forces every arithmetic step to cite a graph node; and weak-parser strong-reasoner alignment (WPSR) that uses a specialized table extractor only as a teacher of node attributes. On ChartQA, STEER reaches 82.70 average relaxed accuracy versus 80.16 for ChartGemma and 76.40 for a LLaVA-CoT backbone trained on the same mix. Gains widen on CharXiv reasoning (33.60 vs. 29.20 InternVL Chat V1.5) and ChartQAPro CoT (40.70 vs. 37.17 Qwen2-VL-7B), where OCR shortcuts disappear. Ablations show that dropping node serialization or numeric candidate constraints undoes most of the reasoning lift.
[CV-185] Forward-Facing Near-Infrared Adds Little to Colour for Farm-Machinery Traversability: A Site-Disjoint Evaluation of Sensor-Dependent Spatial Leakage
链接: https://arxiv.org/abs/2609.13265
作者: Sungwoo Kang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Near-infrared (NIR) imaging does not consistently outperform standard color cameras for daytime agricultural traversability once spatial data leakage is eliminated. Prior benchmarks suggesting an NIR advantage used sequence-level splits that permitted spatially autocorrelated imagery into test sets, artificially inflating NIR performance, especially on difficult paddy-boundary segmentation. When evaluated across strictly held-out recording sites using the AI Hub autonomous driving corpus, none of the four tested configurations (color, NIR, a luminance control, or their fusion) reliably surpasses standard color across any of the five traversability classes. Furthermore, apparent performance gains observed at a single held-out site completely fail to replicate when the held-out site is rotated, demonstrating that the observed benefits were site-specific artifacts rather than generalizable improvements. Correcting for spatial leakage erases the apparent lead of NIR entirely rather than uniformly degrading performance across all sensor types. Consequently, adding a forward-facing NIR camera to a daylight agricultural vehicle sensor suite remains unproven, and evaluations compromised by location leakage risk distorting sensor rankings and misleading procurement decisions. Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.13265 [cs.CV] (or arXiv:2609.13265v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.13265 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-186] BEACON: Behavior and Appearance Control for Subject-Specific Video Generation
链接: https://arxiv.org/abs/2609.13264
作者: Pokrzywa Baptiste,Nabyl Quignon,Yara Bahram,Muhammad Osama Zeeshan,Antitza Dantcheva,Eric Granger
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Generating human-centric videos that preserve both visual identity and person-specific expressive behavior remains a fundamental challenge. In addition to reproducing appearance, a model must replicate the facial behaviors that characterize how a subject expresses emotion over time. However, most state-of-the-art methods condition generation on a single reference image, which contains no information about these temporal dynamics. As a result, they tend to preserve the subject’s visual identity but often produce expressions with limited variation and weak subject specificity. To mitigate this issue, we introduce BEACON, a lightweight framework for person-specific video generation that produces more expressive videos by disentangling visual identity from expressive behavior. BEACON conditions generation on two complementary signals: a reference image encoding the identity and a reference video capturing subject-specific facial dynamics. By conditioning on these complementary signals, BEACON generates videos that better preserve both the subject’s appearance and characteristic facial dynamics, while also supporting identity-expression transfer. Our experiments on the MEAD and RAVDESS datasets show that by fine-tuning on approximately 2,000 pairs and updating about 1% of the pretrained Wan video diffusion model, BEACON improves facial expressivity over state-of-the-art video generation methods while maintaining competitive identity preservation.
[CV-187] Multi-View Structure-from-Motion Enables Oriented Projective Shape Analysis in Three Dimensions
链接: https://arxiv.org/abs/2609.13263
作者: Musab Alamoudi,Robert L. Paige,Vic Patrangenaru
类目: Computer Vision and Pattern Recognition (cs.CV); Applications (stat.AP)
备注: 19 pages, 9 figures
Abstract:Projective shape analysis provides a geometric framework for studying landmark configurations in digital images acquired by pinhole cameras. In the classical projective shape (PS) model, three-dimensional configurations ( k -ads) are represented as points in (\mathrmRP^3)^q , q = k - 5 . A nonparametric test is developed in Patrangenaru et al. [12], for this framework, to determine whether an object matches a design blueprint, with each configuration reconstructed from a single uncalibrated stereo pair. Such two-view reconstructions are identified only up to a 3D projective transformation, which may reverse orientation, so that the sign-blind PS summary was the only available, before oriented projective shape (OPS) was recently considered. Multi-view Structure-from-Motion (SfM) technology removes this obstruction: its bundle adjustment is identified up to an orientation-preserving projective transformation. In this paper we revisit a well-cited three-cube object, from Patrangenaru et al. [12], with n = 8 SfM reconstructions built in Agisoft Metashape Professional 2.3.0, and validate our implementation by reproducing the published stereo construction from the original data. This allows us to perform what is to the best of our knowledge the first three-dimensional OPS analysis, compute its extrinsic total-variance index and perform statistical inference in this novel setting. Due to the high concentration of SfM data, the OPS index is asymptotically one-half the PS index, a structural consequence of concentration rather than a property of the object. Here our blueprint hypothesis is not rejected for any of the q = 14 non-frame landmarks, while the SfM reconstructions are about 26 times more concentrated than the stereo ones, a substantial gain in reconstruction precision. Sample-size, photograph-count, and frame-ordering analyses support the robustness of these conclusions.
[CV-188] Filling the Unseen: Scene Extrapolation via 3D Gaussian Splatting ACM-MM
链接: https://arxiv.org/abs/2609.13262
作者: Yunlai Zhou,Yiren Lu,Tuo Liang,Disheng Liu,Vipin Chaudhary,Yu Yin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ACM Multimedia (ACM MM) 2026. 9 pages
Abstract:3D Gaussian Splatting achieves photorealistic reconstruction within training view distribution, yet it degrades on out-of-distribution novel views, exhibiting holes in unobserved regions and artifacts in observable areas. Recent works formulate this task as extrapolation and interpolation and try to address it with generative models, but remain limited in extrapolation scale and quality. They repeat a generate-reconstruct-shift cycle to progressively build a scene, which introduces accumulated errors with every step conditioning on previous outcomes. In this work, we propose a holistic framework for extrapolation and interpolation. We devise an independent camera view detection mechanism to enable parallel conflict-free extrapolation, circumventing the reliance on the aforementioned error-prone cycle. Building upon this, we design a hierarchical pipeline that extrapolates independent and dependent camera views separately. Additionally, previous methods overlook inconsistency between generated and original images, resulting in compromising well-reconstructed areas. We propose a plug-and-play Quality-Aware Mask (QA-Mask) module, enabling selective utilization of generated data. By calibrating learning weights with pixel-wise rendering quality, it prevents generation-induced degradations on well-constructed areas. Extensive experiments demonstrate the superior performance of our framework, with QA-Mask generalizing to multiple generative reconstruction models.
[CV-189] ryOnReward: Learning Foveated Consistency for Reinforcement Fine-Tuning of Virtual Try-On
链接: https://arxiv.org/abs/2609.13259
作者: Xueheng Li,Yong Liu,Xiaolong Fu,Wen Xue,Chengjun Xie,Yipeng Sun,Yan Li,Simiu Gu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 9 pages, 5 figures
Abstract:Virtual Try-On (VTON) aims to dress a person with the reference garment, producing visually reasonable results aligned with human preferences. Turning this preference-oriented goal into an actionable objective relies on a scoring function aligned with human taste. However, classic fidelity metrics exhibit weak correlation with human judgments, and generic VLMs fail to provide the discriminative granularity demanded by try-on quality evaluation, which hinges on faithfully preserving garment and person details. This shortcoming is further exacerbated in the reinforcement fine-tuning (RFT) optimization and leads to severe reward hacking. To this end, we present TryOnReward, a fine-grained reward model tailored for VTON. Built on a vision-language backbone, it adopts a foveation calibration objective that grounds each quality dimension in the relevant region to avoid global shortcut learning. Meanwhile, TryOnReward jointly optimizes pairwise preferences and per-dimension quality scores via margin-aware supervision, leveraging both relative and absolute quality signals. For model training and evaluation, we build TryOnReward-100K, a human-annotated per-dimension rating dataset, alongside TryOn-Bench and TryOnRewardBench, two benchmarks covering diverse real scenarios. Extensive experiments confirm that TryOnReward significantly outperforms generic judges in human preference alignment, and when serving as the RFT reward function, it consistently yields human-preferred try-on results across multiple baselines.
[CV-190] Sampling headroom is not selection gain: a compute-value audit of test-time scaling for video world models
链接: https://arxiv.org/abs/2609.13257
作者: Yuhua Jiang,Junjie Lu,Feifei Gao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 33 pages
Abstract:Test-time scaling (TTS) can improve generation only when additional compute produces better candidates and the system can reliably identify them. This distinction is especially important for video world models, where a wider sample pool may contain stronger rollouts without improving the output that is ultimately selected. We introduce the Compute-Value Audit (CVA), a sequential framework that asks whether extra sampling creates opportunity, observable signals provide a reliable state, that state supports a beneficial action, and the resulting gain exceeds the full entry fee of generation and verification. On 192 Physics-IQ scenes, expanding the pool from 4 to 16 candidates increases oracle quality by +9.23 IQ (95% CI [+7.44, +11.14]), but Flow, Cycle, and VideoReward fail to recover this headroom reliably. Across three generators, none of twelve adaptive-depth policies outperforms uniform compute; they recover only 42-69% of the measured entry fee. A matched-60-NFE Predict-and-Perturb intervention on VideoPhy2 is likewise negative across three fresh-seed replicas. These negative results are not universal: anchor-explorer passes all four stages in a sparse PRM800K setting, MMLU-Pro exposes the gap between predictive state and useful action, and a privileged paired future establishes a positive video upper bound. Together, these results show that sampling headroom has deployment value only when it can be converted into a reliable decision whose benefit survives the complete compute charge. Code is available at this https URL.
[CV-191] A Comprehensive Review of Multimodal Facial State Analysis: Tasks Methods and Resources
链接: https://arxiv.org/abs/2609.13255
作者: Xuri Ge,Tianshuo Zhang,Ruihan Li,Hui Ye,Kaiwen Zheng,Junchen Fu,Da Huo,Joemon M. Jose,Hu Han
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Facial state analysis plays a crucial role in understanding human expressions, psychological modeling, and human computer interaction. Traditional unimodal vision-based methods are often limited by environmental sensitivity and weak interpretability. Multimodal facial state analysis addresses these issues by integrating complementary cues from visual, audio, textual, physiological, and other related modalities. This survey emphasizes two key aspects: on one hand, multimodal learning enables contextual semantic understanding for improved facial state reasoning and leverages interpretable language generation to enhance model explainability; on the other hand, multi-task learning allows simultaneous analysis of expressions, action units (AUs), and face-based soft biometrics (e.g., age, gender), effectively capturing fine-grained expressions and improving cross-scene generalization. This survey reviews core tasks, representative methods, and datasets in multimodal facial state analysis, focusing on facial expression recognition, AU detection, and face-based soft biometric estimation, and emphasizing the unique value of language in providing contextual semantics, enhancing reasoning, and generating explanations. The survey aims to provide an up-to-date overview of the literature and to highlight future research directions for multimodal, interpretable, and multi-task adaptive facial state analysis.
[CV-192] (How) Do MLLM s Report Bistable Images Like Humans? EMNLP2026
链接: https://arxiv.org/abs/2609.13254
作者: Ryota Takatsuki,Tomoki Doi,Amane Watahiki,Anil K. Seth,Hitomi Yanaka
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to EMNLP 2026 (Main Conference). 31 pages, 46 figures, 2 tables
Abstract:Bistable images such as the duck-rabbit are classic stimuli in which one image supports multiple mutually incompatible interpretations, typically reported one at a time in humans. We ask whether multimodal large language models (MLLMs) show similar report behavior and what internal computations support it. Using the LLaVA family, we study two tractable dimensions: modulability, whether reports can be biased by bottom-up visual cues and top-down linguistic priors, and exclusivity, whether responses commit to a single interpretation. We test both on the canonical duck-rabbit and on synthetic Visual Anagrams to mitigate memorization confounds. Behaviorally, both visual and linguistic manipulations systematically shift reports in human-consistent ways, while responses remain predominantly exclusive. Mechanistically, these effects arise from competing image-token representations, distinct pathways for bottom-up and top-down modulation, and a link between exclusive reporting and object-count encoding. Code and data are available at this https URL.
[CV-193] Preserving Subject-Clarity in Image Outpainting with Multiscale Wavelet Supervision
链接: https://arxiv.org/abs/2609.13251
作者: Abhilash Neog,Taewan Kim,Yi Wu,Xu Chen,Jian Jiao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 29 pages
Abstract:Commercial and advertising images are frequently affected by poor framing, partially cropped subjects, truncated text or logos, and insufficient context, all of which can reduce subject clarity, i.e., the ability of an image to clearly communicate its primary subject. Image outpainting offers a scalable solution by extending image boundaries and recovering missing content and context. However, existing diffusion-based outpainting methods often produce visually plausible completions while degrading subject fidelity through structural inconsistencies, semantic drift, or loss of fine-grained detail. To address this limitation, we propose a subject clarity outpainting framework that combines vision-language model (VLM)-guided semantic conditioning with multiscale wavelet supervision for subject-localized detail preservation. To support training, we develop a subject-centric data curation pipeline that constructs subject-intersecting outpainting pairs from advertising and natural images. The resulting objective introduces no additional inference cost and is designed to be compatible with diffusion-based backbones. Across four advertising and natural-image benchmarks, our method improves subject clarity, reducing subject-centered DreamSim error and FID on average by 3.0% and 2.4% over matched supervised fine-tuning, and by 10.8% and 7.7% over the strongest state-of-the-art approach per dataset, respectively.
[CV-194] Evaluation of MLLM -Agnostic Plug-and-Play Keyframe Selection Methods for Long Video Understanding
链接: https://arxiv.org/abs/2609.13250
作者: Dilip Sarkar,Md. Safayet Islam,Liang Liang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 10 pages,3 figures, 5 tables, 24 references, preprint
Abstract:Multimodal large language models (MLLMs) cannot process every frame of a long video because of limitations in visual-token and computational budgets. Three primary approaches have been proposed to enhance their long-video understanding capabilities: (i) Retraining an MLLM on a large video corpus and/or extending its input length; (ii) Training an adapter for a specific MLLM that takes the entire video and the query as input and selects the most relevant video frames; and (iii) Developing a training-free, plug-and-play (PaP) adapter that is MLLM-agnostic. We refer to the third approach as PaP keyframe selection. A PaP method may use only candidate video frames without considering the query, or it may use both candidate video frames and the query. The first approach is prohibitively expensive. The second approach requires substantial training time and computational resources, but it is accessible to many because an adapter contains significantly fewer trainable parameters than an entire MLLM. The third approach has the lowest computational cost and is therefore broadly accessible. To the best of our knowledge, only five PaP methods have been reported within the past year. All of these methods have been evaluated on one or more video question-answering benchmarks and have demonstrated improvements in long-video understanding. However, the methods were evaluated on different benchmarks using different MLLMs. We present a comprehensive evaluation of these five methods using three MLLMs across three long-video understanding benchmarks. Our results show that QAaF achieves the best performance in 13 of the 15 aggregate evaluation settings, while FOCUS ranks second overall. These results provide a common experimental reference for comparing training-free keyframe-selection methods for MLLMs.
[CV-195] Pixel-wise Planarity for High-Precision Monocular Plane Segmentation ECCV2026
链接: https://arxiv.org/abs/2609.13246
作者: Ahmetcan Yavuz,Alpay Ozkan,Rémi Pautrat,Shaohui Liu,Marc Pollefeys
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: To appear at ECCV 2026. Code available at this https URL
Abstract:Plane segmentation from a single RGB image remains challenging due to imprecise region grouping and geometrically inconsistent supervision, often leading to over-segmentation and false planar detections. We propose instead a pixel-wise planarity prediction framework for robust monocular plane segmentation. Building on a pretrained monocular geometric backbone predicting depth and surface normals, we introduce a dedicated planarity head that estimates per-pixel planarity confidence. During inference, predicted depth, normals, and planarity are combined in a lightweight region-growing procedure that enforces geometric consistency when forming plane segments. We further analyze existing plane ground-truth annotations and demonstrate substantial geometric inconsistencies under strict distance thresholds. Across multiple datasets, our method achieves improved geometric precision and segmentation quality compared to prior state-of-the-art approaches, while improving computational efficiency. Our code and models are available at this https URL.
[CV-196] SJD-SV: Speculative Jacobi Decoding with Semantics Verification for Autoregressive Image Generation ICML2026
链接: https://arxiv.org/abs/2609.13245
作者: Baoquan Zhang,Bingqi Shan,Shihao Fang,Kenghong Lin,Xutao Li,Yunming Ye
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the 43rd International Conference on Machine Learning (ICML 2026)
Abstract:Speculative Jacobi Decoding (SJD) is an important approach for accelerating autoregressive image generation. Although SJD has shown superior performance, recent studies point out that it usually suffers from a token ambiguity issue during token verification but its reason can not be well explained. To figure out this reason, in this paper, we conduct a visualization analysis on vision token and find that different from text tokens, vision tokens generally corresponds to some local, small, and unclear vision details, which means only using single token is difficult to accurately express a certain semantic, thereby causing token ambiguity issue. To this end, we propose a novel Speculative Jacobi Decoding with Semantics Verification (called SJD-SV), for accelerating autoregressive image generation. The key idea is that leveraging the strong correction characters between tokens to recognize semantic-aware token subsequence and then instead of perform token-by-token verification, turning to perform verification on semantic-aware token subsequence level for accelerating image generation. In particular, our method is plug-in, which can be directly integrated into existing SJD and its variants. Extensive experiments on various datasets show that existing SJD methods achieve significant performance improvement after integrating our SJD-SV method.
[CV-197] Physical Kernel: Structured Visual Latents for Dark Manipulation
链接: https://arxiv.org/abs/2609.13244
作者: Jinting Hang,Hong Li,Zhenhui Cai,Zhihao Zhao,Jian He
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We study dark manipulation: after a brief lit Write encodes z0 = Enc(rgb), a policy pi(z) and open-loop dynamics f(z,a) complete contact-rich skills without further pixels (dark_f). On ManiSkill StackCube (n=160; seed packs 0/1000), dark_f attains 68.1% stacked on the five-rung chain (near_A - grasped - lifted - on_B - stacked), compared with 35.6% for per-step lit_reenc and 0% for freeze/encode_black. On a shared Write-HOLD protocol (n=40), occlusion and camera-aligned GT contact-neighbor masks drive lit lift from 43% to 0%, while dark_f holds 82.5%; shuffling actions inflates dynamics MSE by ~9.4x; write-time appearance shifts break encoding (night: 0% stacked), yet the same shifts during HOLD leave dark_f lift unchanged; Write length Tw is flat once the stop phase is reached, while earlier stops and write-time blur/JPEG sharply cut stacked. A dedicated pi_write reaches 35% vision-budget stacked (n=80); matched Dreamer-style/pixel nulls without privileged geom stay at 0%. Privileged state-RSSM MPC reaches ~35% stacked with 9D dark observations – a stronger-observation null, not a matched visual baseline.
[CV-198] ArtSociety: Multi-Agent Multimodal Collaboration for Art Emotion Understanding
链接: https://arxiv.org/abs/2609.13240
作者: Jian Li,Fanfan Ji,Jinxiang Lai,Ying Tai,Jian Yang,Xiao-Tong Yuan,Chengjie Wang,Yabiao Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the ACM Multimedia 2026 Grand Challenge (AffectiveArt). 8 pages, 5 figures, 3 tables. Code: this https URL
Abstract:The AffectiveArt Multidimensional Art Emotion Understanding task asks to jointly predict an artwork’s fine-grained emotion (12 classes, 1549:1 head-to-tail ratio), binary valence/arousal, and five attribute-grounded descriptions – sub-tasks that exhibit strong empirical trade-offs, so the single-model solutions we tried do not jointly optimize all of them well. We present ArtSociety, a multi-agent framework that assembles heterogeneous multimodal experts – a DINOv2-Giant vision agent (A1), a scene-grounded CoT fine-tuned MLLM (A2), and three closed-source reasoning agents (A3-A5) – and coordinates them with two training-free controllers: (i) a rare-class-aware voting arbiter that lowers the agreement threshold for tail emotions, exploiting decorrelated error patterns across agent families; and (ii) a description-first reasoning agent whose DESCRIBE-then-CLASSIFY chain of thought forces visual evidence before label commitment, yielding near-perfect grounded descriptions. A task-routing policy directs the hard emotion task to the full five-agent ensemble while assigning the near-saturated valence/arousal and generative description tasks to the single strongest reasoning agent. On the official test set (1,000 artworks), ArtSociety achieves an Overall Score of 0.8870 (Classification 0.7789, Description 0.9952). An eleven-variant ablation study reveals that, once method and scale saturate at around 0.76, the decisive gains come from agent collaboration and data-side supervision – a 30B MoE model trained on older data does not outperform an 8B model trained on better data. Code is available at this https URL
[CV-199] Abstract-LoRA: Unlocking Single-Image Style Transfer through Targeted U-Net Block Training
链接: https://arxiv.org/abs/2609.13239
作者: Xinglin Hu
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Diffusion models represent one of the most advanced paradigms in generative modeling. Leveraging their development, a growing number of style transfer methods based on diffusion models have been proposed. However, among these methods, multi-image style transfer approaches that require at least five to ten style examples tend to achieve more satisfactory results. Single-image methods, by contrast, often struggle with either insufficient content preservation or inadequate style fidelity. This greatly limits style extraction from scarce artworks and undermines their artistic value. To address this, we propose Abstract-LoRA, a method that pushes the boundaries of single-image style transfer through lightweight LoRA training on specific U-Net blocks in diffusion models. Specifically, our work is inspired by B-LoRA, a style transfer method that achieves basic style-content disentanglement by training specific U-Net blocks. However, it suffers from a critical limitation: the inability to capture complex backgrounds. Building upon B-LoRA, our method conducts a more refined analysis of U-Net blocks, employing additional U-Net blocks and clustering-based abstraction of style images to better disentangle and balance style and content. Extensive experiments demonstrate that our proposed method not only generates visually more harmonious and satisfying artistic images but also quantitatively improves the preservation of both style and content in the final outputs.
[CV-200] Learning Manipulation-Sufficient Representations via Outcome Bottlenecks
链接: https://arxiv.org/abs/2609.13235
作者: Md Selim Sarowar,Sungho Kim
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: Project Page : this https URL
Abstract:Networked manipulation endpoints couple perception to actuation across compute- and bandwidth-limited links, yet commonly exchange dense geometric states optimized for fidelity rather than action outcomes. A stochastic representation is learned with a policy-free, action-conditioned outcome bottleneck: marginal outcome log-loss supplies distortion and a KL term regularizes rate. The construction is motivated by the minimal statistic that preserves the outcome distribution of every admissible action, while the implemented finite model is evaluated as a rate-regularized mixture predictor. The same encoder and outcome head support grasp selection, singleton conformal filtering, active viewpoint selection, and latent test-time adaptation. A finite-probe theorem identifies the local level-set tangent space with the null space of an outcome Jacobian. The synthetic oracle verifies this result; on scanned objects, an analytic surrogate agrees with measured simulator invariances within (0.56^\circ). Across 11,979 simulated grasps on 13 objects, a reconstructed-geometry wrench score attains 0.542 AUC against lift success and falls below chance on curved objects, while our representation attains 0.876. At 25% commitment, executed-grasp success is 0.503 versus 0.984. The 512-byte interface is (288\times) smaller than one RGB-D frame and runs at 16,ms per CPU decision. Independent synthetic points track the tested conformal levels; scanned-object all-pair coverage is reported as a clustered empirical diagnostic. On unseen objects, within-scene AUC falls to 0.569, and a full-feedback update raises empirical mean pairwise coverage from 0.728 to 0.883.
[CV-201] EMCStereo: Attention-Enhanced Stereo Matching for Thin-Structure Depth Estimation with a Synthetic Tree-Branch Benchmark
链接: https://arxiv.org/abs/2609.13233
作者: Yida Lin,Bing Xue,Mengjie Zhang,Sam Schofield,Richard Green
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Thin structures such as tree branches are among the hardest cases for stereo matching: a branch is only a few pixels wide, the background is cluttered, and dense ground truth for real branches is nearly impossible to label by hand. We make three contributions. First, EMCStereo integrates three lightweight attention modules into a PSMNet-style cost-volume backbone: Efficient Multi-scale Attention (EMA) on deep semantic features, a Multi-Scale Fusion block (MSFblock) learning spatial pyramid weights instead of concatenating them, and Coordinate Attention (CoordAtt) on final matching features. Because MSFblock collapses four pyramid branches into one, the modules leave the network 2.0% smaller and add only 1.7% inference time overhead. Second, VirtualTree is a synthetic stereo dataset rendered in Unreal Engine 5 with a simulated ZED Mini rig, providing 5,520 pairs with exact disparity for thin branches. Third, an eight-way ablation establishes a run-to-run noise floor of 0.009 px end-point error (EPE). EMCStereo achieves 1.31 px EPE (5.96% D1-all) on the VirtualTree test split, 1.00 px on SceneFlow, and 0.80, 0.73, 0.62, and 3.19 px on KITTI 2012, KITTI 2015, ETH3D, and Middlebury, with depth accuracy delta_1 from 92.6% to 98.7%. Evaluated against the noise floor, the attention stack is accuracy-neutral at a 100-epoch budget, while MSFblock and CoordAtt cost 0.03-0.05 px unless EMA is present.
[CV-202] What Does the Encoder Actually Decide? A Controlled Comparison of Vision Backbones on Joint Tree Segmentation and Stereo Depth
链接: https://arxiv.org/abs/2609.13232
作者: Yida Lin,Bing Xue,Mengjie Zhang,Sam Schofield,Richard Green
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:A robot pruning trees needs two facts per pixel: whether it belongs to a tree, and its distance. Both are usually obtained via task heads attached to a vision backbone chosen by reputation rather than measurement. Holding dataset, decoders, losses, schedule, and evaluation fixed, we ask: how much does the encoder choice change joint semantic segmentation and stereo depth on thin vegetation? We build a hard parameter-sharing network with one encoder feeding both branches, swapping only the encoder without downstream retuning. We evaluate [N] encoders across [M] architecture families (CNNs, transformers, hybrids, MLP-mixers, state-space models) near a ~25M budget, trained from scratch. Depth is evaluated on tree pixels only; segmentation uses boundary F1 and background IoU to prevent “label-everything-tree” shortcuts. Three findings stand out. First, the strongest encoders are convolutional and hybrid, not transformers: [BestEncoder] leads with [MIoU] segmentation mIoU and [Delta] depth \delta_1 , while [X] of [Y] plain vision transformers collapse when trained from scratch. Second, parameter count does not predict quality — [SmallEncoder] at only [P]M parameters outranks models two orders of magnitude larger. Third, segmentation and depth rankings agree strongly (Spearman \rho = [RhoValue]), showing no task conflict. Finally, [K] of [N] encoders collapse to degenerate all-tree segmentation — exposed by boundary F1 but hidden by region IoU.
[CV-203] Dont Just Look Intervene: Perturbation Based Region Labeling for VQA Images
链接: https://arxiv.org/abs/2609.13228
作者: Marko Jojic,Zhaonan Li,Ben Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision Language Models (VLMs) should rely on visual evidence that directly determines the correct answer, but supervision for grounding visual reasoning is often expensive to obtain manually or tied to dataset-specific annotation primitives. We instead introduce model-causal visual evidence as an annotation target, defined as the set of image regions whose counterfactual intervention changes a model’s answer distribution for a given image-question pair. Based on this principle, we introduce Counterfactual Search for Grounding Regions (CSGR). CSGR is a scalable pipeline that proposes candidate regions, perturbs them, measures their effect on answer sensitivity, and aggregates this evidence across multiple judges to approximate answer-critical regions in VQA data. To assess whether CSGR annotations contain a useful supervision signal, we plug them into three existing grounding-aware training routines: attention steering, Visual CoTfinetuning, and latent visual reasoning. These experiments test whether the proposed annotation scheme can provide a useful supervision signal across multiple ways of consuming region labels, rather than introducing a new way of using them. Across competing automatic region-labeling mechanisms, CSGR annotations provide the most consistent gains over Cross Entropy-only finetuning in both in-domain and out-of-domain evaluations, indicating that the proposed labeling scheme captures useful region-level information.
[CV-204] Synthetic Leprosy Image Generation Using Mask-Conditioned Latent Diffusion and Transfer Learning from Large Chronic Wound Datasets
链接: https://arxiv.org/abs/2609.13226
作者: Yusuf Abdulkadir
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages, 11 figures, 3 tables. Code and data pipeline: this https URL
Abstract:Machine learning for neglected tropical diseases is limited by data, not algorithms: public annotated image sets for leprosy (Hansen’s disease) number in the hundreds, orders of magnitude below what generative models require. We ask whether a model trained on abundant chronic wound photography transfers to this low-data regime. We build a three-stage pipeline. First, a DeepLabV3-ResNet50 segmentation network (validation Dice 0.876, IoU 0.799) supplies lesion masks for two wound datasets that ship without them. Second, we assemble a mask-conditioned latent diffusion model from Stable Diffusion 1.5 components and train it on 3,280 region-of-interest wound crops, widening the UNet input convolution from 4 to 11 channels to admit three mask feature maps and a blurred low-frequency context latent. Third, we fine-tune this model on 708 leprosy image-mask pairs drawn from 764 images of approximately 150 patients. We evaluate with LPIPS perceptual distance, anchored by a real-versus-real baseline computed on the same 242 anchor images as the cross-set comparisons; without that reference the cross-set distances cannot be interpreted. The generated set shows no mode collapse: its internal perceptual diversity (0.662) is statistically indistinguishable from that of the real leprosy set (0.672, 95% CI [0.664, 0.680]). Generated images sit 0.044 LPIPS outside the real distribution - measurably apart, but under half of one standard deviation. Fine-tuning shifted the output distribution only marginally, which we trace to lesion geometry reaching the network through input concatenation alone. Chronic wound photography is therefore a viable donor domain for leprosy lesion synthesis: low-level appearance transfers well, and the remaining barrier is semantic control rather than image quality.
[CV-205] Part Grounding Not Action Knowledge: Locating the Bottleneck in VLM Affordance Prediction
链接: https://arxiv.org/abs/2609.13225
作者: Sarthak Sattigeri
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注:
Abstract:Benchmarks agree that vision-language models reason poorly about low-level manipulation, but an aggregate accuracy score does not say which step fails. We separate two steps that affordance questions conflate: identifying which part of an object to act on, and knowing what action that part requires. Across 19 articulated objects we asked eight models, spanning three developers, what motion a robot should apply. Under an open prompt, push was produced once in 64 evaluations where it was correct, despite being correct for 8 of 19 objects and appearing in the offered label set every time. Inspecting the outputs showed why: models described a different part than the one being scored, e.g. explaining how to pick up a camera rather than press its button. Naming the target part raises action accuracy by 0.32 to 0.63 for every model, from 0.158-0.474 to 0.684-0.947, and push recall from 0-1/8 to 7-8/8. No model beats a constant answer that ignores the image under the open prompt; once the part is named, all eight do. Asked to describe the same part in free prose with no label set, models produce pressing language for 6 to 8 of 8. These results are hard to reconcile with missing action knowledge, and instead point to part grounding as the dominant bottleneck, a pattern that holds across all three model families and does not diminish with capability. Naming the part supplies the grounding variable, so this bounds what a perfect part detector would offer rather than demonstrating a general model of mechanics. Two supporting results agree: on real photographs only three of eight models localize grasp points better than a constant baseline, and on rendered objects none do. We also document two measurement errors of our own, a threshold that let a constant baseline score 0.929 and a labelling rule wrong on 4 of 19 objects, both caught only by testing our numbers against trivial alternatives.
[CV-206] Discovering and Preserving Category Correlation Knowledge via Adaptive Reciprocal Knowledge Distillation CIKM2026
链接: https://arxiv.org/abs/2609.13199
作者: Dawen Jiang,Zhishu Shen,Zeyu Liu,Tiehua Zhang
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by 35th ACM International Conference on Information and Knowledge Management (CIKM 2026)
Abstract:Knowledge distillation aims to improve the performance of lightweight student models by transferring knowledge from larger and more powerful teacher models. However, a substantial size gap between teacher and student models often impedes effective knowledge transfer. Most existing approaches adopt a static, one-way teacher-to-student distillation paradigm, which overlooks the dynamic nature of student learning and fails to provide targeted guidance on hard samples. In this paper, we propose adaptive reciprocal knowledge distillation (AR-KD), a novel method that improves knowledge transfer by simplifying the teacher’s output distribution. Specifically, AR-KD performs reciprocal adaptation on the teacher by matching its class correlation matrix to the student’s relational representation, which reshapes the teacher’s prediction structure to better suit the student’s capacity. This relational alignment mitigates the collapse of inter-class dark knowledge caused by overconfident teachers, enabling the student to learn from richer and more compatible supervisory signals. We evaluate AR-KD on CIFAR-100 and ImageNet-1k classification datasets, where it outperforms state-of-the-art knowledge distillation baselines. Specifically, AR-KD improves student performance across homogeneous and heterogeneous setups: up to 7.13% accuracy gain for students, 1.42% to 4.15% higher than vanilla KD on average, and further improvements when integrated with other advanced methods. Our code is available at this https URL.
[CV-207] SPICE: Simple Polysemantic Feature Interpretation via Clustering-based Explanation ECCV2026
链接: https://arxiv.org/abs/2609.13198
作者: Sehyun Lee,Dahee Kwon,Damin Lee,Jaesik Choi
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: European Conference on Computer Vision – ECCV 2026
Abstract:One of the pivotal recent challenges in neural network interpretability is polysemanticity, where a single neuron is activated by multiple, often unrelated concepts, hindering clear functional understanding. Although prior work has explored this phenomenon, existing approaches remain architecture-specific and depend on manual heuristics such as a fixed number of concept clusters ( K ), limiting their generality and scalability–especially for modern Transformer-based models. To address these limitations, we introduce SPICE (\textbfSimple \textbfPolysemantic Feature \textbfInterpretation via \textbfClustering-based \textbfExplanation), a generalizable framework for analyzing polysemanticity in deep vision architectures. SPICE avoids architecture-dependent propagation rules, enabling the first systematic comparison of polysemanticity across both CNNs and Transformers, and automatically determines the number of concept clusters per neuron, eliminating reliance on a preset K and supporting scalable analysis for large models. Using SPICE, we conduct a comprehensive investigation into how polysemanticity emerges, varies across depth and architecture, and forms through distinct computational pathways.
[CV-208] Personalized and Explainable Blood Pressure Estimation from PPG via Hybrid CNN–Morphological Features
链接: https://arxiv.org/abs/2609.13190
作者: Myung-Kyu Yi,Jongshill Lee,Jeyeon Lee,In Young Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Continuous cuffless blood pressure (BP) monitoring using photoplethysmography (PPG) offers a promising solution for personalized healthcare. However, existing methods have two major limitations. Handcrafted feature-based approaches rely on precise fiducial point detection and are limited to short-term analysis, while deep learning models, despite their accuracy, often operate as black boxes with limited physiological interpretability. To address these challenges, we propose a physiology-guided hybrid framework for personalized BP estimation that couples a convolutional neural network (CNN) branch capturing global and local waveform dynamics with a morphology-prior branch that explicitly encodes person-specific vascular characteristics. By embedding a morphology-based feature set that explicitly encodes individual vascular characteristics, the proposed framework enhances personalization and reduces dependence on large-scale training datasets. Evaluated on a subset of the MIMIC-III database under a subject-specific (personalized) testing protocol, the proposed personalized physiology-guided hybrid approach achieved mean absolute errors (MAEs) of 3.77 mmHg for systolic BP and 2.36 mmHg for diastolic BP, corresponding to relative improvements of 43.7% and 32.4% over a subject-specific (personalized) CNN-only baseline. SHAP-based analysis confirmed that the introduced morphology-prior features align with individual vascular characteristics, reinforcing per-subject interpretability. These findings highlight the potential of personalized, physiology-guided hybrid learning with novel morphological descriptors for accurate and explainable BP monitoring in real-world settings.
[CV-209] Land Art as a Big-Data Climate Sensor
链接: https://arxiv.org/abs/2609.13182
作者: Alev Cinbarci,Sean Kalaycioglu
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Atmospheric and Oceanic Physics (physics.ao-ph)
备注: 20 pages, 7 figures, 3 tables
Abstract:Robert Smithson’s 1970 land artwork Spiral Jetty, located in the north arm of Utah’s Great Salt Lake, has alternated between submergence and exposure during severe lake decline. We analyze 1,744 co-registered Landsat 4-9 and Sentinel-2 image chips spanning every year and calendar month from 1984 to 2025. A 14-feature complexity signature combines Shannon entropy, multiscale permutation entropy, fractal dimension, lacunarity, gray-level co-occurrence texture, intensity statistics, and ImageNet-pretrained ResNet50 features. These measurements are compared with a 42-year monthly climate and hydrology panel from NASA GISTEMP, USGS NWIS, Open-Meteo, and the Global Carbon Budget. Bootstrap analysis shows that Shannon entropy is a weak proxy and does not support an earlier small-sample claim of positive correlation with global temperature. By contrast, coarse-scale permutation entropy and mean intensity track lake elevation strongly, with Spearman correlations of 0.85 to 0.88 and 95 percent confidence intervals excluding zero. The third principal component of the ResNet50 embeddings emerges without supervision as an AI climate axis, correlating 0.86 with cumulative CO2 and -0.83 with lake elevation. Image complexity leads lake stage by about three years, with Pearson r = 0.58 at lag +3 and a 95 percent confidence interval of 0.40 to 0.73. STL decomposition reveals a non-monotonic trend that rises from 1984 to 2015 and declines sharply thereafter as the lake approaches record-low elevations. Partial correlations controlling for month and sensor confirm robustness to seasonal and sensor effects. These results refine the art-as-thermometer metaphor into an art-as-leading-indicator-of-hydrological-state interpretation. The dataset, feature pipeline, and analysis code are released as a public benchmark.
[CV-210] Physically Aware Radiomics Without Interpolation: Disentangling Voxel Geometry and Signal Modification in CT and MRI
链接: https://arxiv.org/abs/2607.12399
作者: David Corral Fontecha,Juan Miranda Bautista,Pablo Menendez Fernández-Miranda,Sergio Rubio-Martín,Lara Lloret Iglesias,Jose A. Vega
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Manuscript under peer review
Abstract:Objective: Radiomic texture features are usually computed in voxel-index neighborhoods, implicitly assuming isotropic spatial relationships. In anisotropic images, this can confound voxel geometry with interpolation-induced signal changes. We developed a voxel-spacing-aware radiomic framework that incorporates physical geometry into texture computation without resampling. Approach: We modified PyRadiomics to account for voxel spacing while preserving the native image signal. Four configurations were compared: native non-resampled extraction (NR), isotropic resampling (RS), voxel-spacing-aware extraction (VS), and fake-isotropic preprocessing (FK), in which spacing metadata were overwritten without altering the image array. Experiments included 685 LIDC-IDRI pulmonary nodules and 209 I-SPY2 breast MRI cases, with 196 radiomic descriptors. Robustness was assessed using ICC, within-subject variability, Friedman testing, feature selection, machine learning, a multilayer perceptron, and external validation. Main results: VS showed near-native agreement with NR: median ICC(A,1) was 0.9976 in CT and 0.9984 in MRI. RS produced lower agreement and larger deviations, while FK showed intermediate behavior, confirming that spacing metadata alone can affect radiomic features. Gradient-derived and neighborhood-sensitive descriptors were most affected by preprocessing. VS preserved predictive performance comparable to NR in external CT validation, whereas MRI showed greater variability across preprocessing strategies and classifiers. Significance: Voxel-spacing-aware extraction separates geometric modeling from interpolation-induced signal modification while preserving the native image signal, offering a coherent alternative to isotropic resampling for radiomic analysis of anisotropic CT and MRI. Comments: Manuscript under peer review Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2607.12399 [cs.CV] (or arXiv:2607.12399v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2607.12399 Focus to learn more arXiv-issued DOI via DataCite Submission history From: David Corral Fontecha [view email] [v1] Tue, 14 Jul 2026 06:19:05 UTC (6,623 KB)
[CV-211] ransforming harmonic coefficients for 3D splat compression
链接: https://arxiv.org/abs/2609.15735
作者: Tam Thuc Do,Philip A. Chou,Gene Cheung
类目: ignal Processing (eess.SP); Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:We address the problem of color attribute compression for 3D splats. We show that all images generated by 3D splats are linear in the coefficients for each color channel, each spherical harmonic, and each splat, and we identify a basis for the space of all such images. We identify an inner product for the coefficient space that induces the squared error loss on images. We show that orthonormalizing the coefficients with respect to this innner product before coding can yield over 2 dB gain.
[CV-212] Proximal-Only Transmission Matrix Recovery of an Arbitrarily Deformed Graded-Index Multimode Fiber
链接: https://arxiv.org/abs/2609.14869
作者: Cole Reynolds
类目: Optics (physics.optics); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:The multimode fiber is among the thinnest imaging conduits available, carrying hundreds to thousands of spatial modes through a cross-section comparable to a human hair, but its endoscopic capabilities are currently limited by the sensitivity of the transmission matrix to the fiber’s deformed state. Proximal-only recovery of the fiber’s transmission matrix is an appealing approach for enabling general use multimode fiber endoscopy, and within the last decade, machine learning techniques have been applied to both single-ended and double-ended transmission matrix recovery tasks. We present a new approach to this interdisciplinary problem and show that neural networks can generalize to recover transmission matrices of an arbitrarily deformed graded-index multimode fiber from proximal measurements alone.
[CV-213] Multi4D: an end-to-end neural network for structural determination at complex material interfaces
链接: https://arxiv.org/abs/2609.14348
作者: Haoran Zhang,Zian Mao,Shufen Chu,Xiaoya He,Yuyan Guan,Antong Yang,Mingze Li,Xiaoqin Zeng,Yujun Xie
类目: Materials Science (cond-mat.mtrl-sci); Computer Vision and Pattern Recognition (cs.CV)
备注: 17 pages, 6 figures
Abstract:Heterogeneous interfaces dictate the performance and degradation of functional materials, making it essential to link local structural variations with macroscopic failure mechanisms to guide future materials design. Yet structural heterogeneity, phase overlap, and local disorder produce highly convoluted diffraction signatures, making extended transition regions difficult to interpret at atomic resolution across large fields of view. Here, we introduce Multi4D, a physics-informed neural network framework for automated multi-component crystallographic identification using four-dimensional scanning transmission electron microscopy (4D-STEM). By combining a latent-space Diffusion Transformer for physics-constrained style translation with a rotation-invariant convolutional neural network for orientation-agnostic classification, this approach translates multi-components diffraction datasets into deterministic crystallographic maps with 98.82% accuracy. In addition, we introduce Diffraction-Inferred Structural Complexity as an information-theoretic entropy metric derived from classifier predictive uncertainty that quantifies local structural ambiguity. We apply Multi4D to generate high-fidelity structural maps of complex superconducting heterostructures, corroded alloy surfaces, and degraded solid-state battery interfaces down to single-nanometer spatial resolution. This framework establishes a statistically robust analytical paradigm for automated microscopy, facilitating both industrial quality control and the data-driven discovery of interfacial design principles.
[CV-214] Deformable 2D Gaussian Splatting for Efficient 4K Video Compression
链接: https://arxiv.org/abs/2609.14129
作者: Chenhao Zhang,Fengqing Zhu
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Ultra-High-Definition (UHD) video presents significant challenges for efficient storage and real-time decoding. Learning-based methods, such as Neural Video Compression (NVC) and Implicit Neural Representations (INR), achieve competitive rate-distortion performance but suffer from high decoding latency and excessive memory usage. Meanwhile, Gaussian Splatting has recently attracted attention in the computer graphics community due to its ultra-fast rendering and high-fidelity visual quality. Despite these advantages, its application in video compression remains largely unexplored. To bridge this gap, we propose a real-time video compression framework that represents and compresses a Group of Pictures (GOP) using a coarse-to-fine multi-scale 2D Gaussian Splatting (2DGS) structure coupled with a lightweight deformation network. Experiments demonstrate that our method delivers rate-distortion performance in LPIPS that surpasses H.265 and other state-of-the-art learning-based video compression methods. Our work demonstrates the potential of Gaussian Splatting as a practical solution for efficient high-resolution video compression.
[CV-215] SONAR: A Structure-Consistent Neural Operator for Null-Space-Aware Sparse View CT Reconstruction
链接: https://arxiv.org/abs/2609.13688
作者: Song Ni,Haijun Yu,Haodong Li,Changsheng Fang,Shuyi Fan,Yixing Huang,Hengyong Yu
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Sparse-view computed tomography (CT) reduces radiation dose and acquisition time but remains severely ill-posed because incomplete projections poorly constrain null-space information. Existing learning-based methods often estimate this information in high-dimensional image space, conflate physical measurement errors with prediction errors, and depend on fixed discretizations. We propose SONAR, a Structure-Consistent Neural Operator for Null-Space-Aware Reconstruction. Instead of recovering the full null-space component, SONAR predicts a low-dimensional null-space-aware representation from the acquired projections as pseudo-measurements. It separates measurement and pseudo-measurement residuals, lifts them into the image domain through physics operators, and applies independent neural operators to constrain their structural effects, thereby accommodating admissible errors while suppressing unsupported structures. To support cross-discretization reconstruction, an anisotropic U-shaped neural operator models the periodic angular and nonperiodic detector dimensions using direction-dependent continuous supports, while image-domain neural operators re-discretize continuous kernels on target grids. These components form an optimization-inspired unrolled network. Experiments on simulated AAPM and clinical MARS photon-counting CT data demonstrate consistent improvements across seen and unseen view settings and unseen image resolutions. On AAPM dataset, SONAR improves PSNR by 1.87~dB at 62 views and by 7.63~dB under zero-shot transfer to a 512\times512 grid over the strongest competing methods. SONAR also achieves the best overall performance in all clinical settings evaluated, demonstrating accurate, structurally reliable, and discretization-robust sparse-view CT reconstruction.
[CV-216] ranslation-Invariant Tile-Based Phase Unwrapping with Residual-Weighted Multipath Averag ing
链接: https://arxiv.org/abs/2609.13409
作者: Antoine Moevus,Max Mignotte
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 7 figures, 2 tables. Preprint
Abstract:Phase unwrapping is a key step in interferometric and coherent imaging, where the physical quantity of interest is carried by a phase that the instrument delivers only modulo 2*pi. The difficulty in two dimensions is to separate the jumps caused by wrapping from those produced by noise, by true discontinuities, by under-sampling or by decorrelation. Spatial-domain and frequency-domain methods have both been studied extensively, each with advantages the other lacks; hybrid schemes combining the two remain scarce. We propose a semi-global tile-based strategy in which every tile is unwrapped in the frequency domain, through the Discrete Cosine Transform (DCT) and the least squares (LS) formalism of Ghiglia et al., the tiles being merged spatially. Unwrapping and noise filtering are performed jointly, which regularizes an otherwise ill-posed inverse problem. An error stays confined to the tile in which it arose; the tiling artifacts are removed by averaging over every shift of the grid and over the symmetries of the square; and that average is weighted by the Poisson residual each pass leaves behind, so that a pass whose tile boundaries fell on a discontinuity does not impose its seam on the result. Experiments on synthetic and real data, against four reference algorithms from four distinct families and under six complementary metrics, show that the proposed method matches or improves on the state of the art. One of these metrics, a corrected cyclic re-wrap residual introduced here, needs no ground truth and therefore remains available on real acquisitions.
[CV-217] From Semantic to Token Communication: The Next Paradigm for Large-Model-Driven 6G Intelligent Connectivity
链接: https://arxiv.org/abs/2609.10714
作者: Yu Ma,Zhen Gao,Li Qiao,Xiaoyuan Zhang,Mahdi Boloursaz Mashhadi,Yin Xu,Wenjun Xu,Xiaodong Xu,Kaibin Huang,Jiangzhou Wang,Rahim Tafazolli,Sheng Chen,Tony Q. S. Quek,Ping Zhang
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Information Theory (cs.IT); Machine Learning (cs.LG)
备注: 39 pages, 10 figures, 10 tables, 182 references. Submitted to Science China Information Sciences
Abstract:The ambitious requirements of sixth-generation (6G) networks are driving communication systems from reliable bit delivery toward meaning-aware and task-oriented connectivity. Large models (LMs), with strong multimodal understanding and generation capabilities, have accelerated this shift and made semantic communication (SemCom) increasingly practical. Yet current LM-driven SemCom remains fragmented: semantic representations are typically tied to specific modalities, models, or tasks. While the bit provides a universal unit for digital transport, there is still no analogous unit for representing and processing semantics, which limits interoperability, theoretical unification, and scalable system design. We argue that tokens provide a natural candidate for this missing abstraction. Two trends support this: unified multimodal LMs now encode text, images, audio, video, and robot actions in one token space, while distributed LM inference already generates substantial token-level traffic through expert routing, cache transfer, and speculative decoding. Token communication (TokenCom) emerges by unifying these trends, using the LM’s native processing unit as a communication abstraction above the bit level and enabling importance assignment, error handling, and resource allocation directly at token granularity. This survey traces the evolution from LM-driven SemCom to TokenCom. We review three major directions of LM-driven SemCom: source-centric semantic coding, channel semantics for physical-layer tasks, and collaborative edge-device intelligence. We then examine the token abstraction, the transmission techniques it requires, and two emerging paradigms, namely TokenCom for LM services and for embodied and agentic intelligence. Finally, we identify open challenges toward unified, scalable, and AI-native 6G communication systems.
人工智能
[AI-0] Corrupt Plans Clean Traces: Evading Chain-of-Thought Monitoring with Plan Injection
链接: https://arxiv.org/abs/2609.15989
作者: Keertana Chidambaram,Andrew Ilyas,Vasilis Syrgkanis
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Chain-of-thought (CoT) monitoring is a safety strategy where the reasoning of a large language model “actor” is inspected by a “monitor” (often another language model) for signs of unsafe planning, deception, or misalignment. We find that planting harmful but benign-sounding reasoning in the actor’s context can steer it to perform adversarial actions while evading monitors, an attack we term “plan injection”. We initially discover this attack in the multiple-choice question-answering monitorability setting proposed by Lanham et al. (2023), using the investigator-agent elicitation framework of Li et al. (2025). We generalize the attack and show that the discovered behavior scales to harder tasks (achieving 25-33% monitor evasion rates across different monitorability benchmarks) and larger models such as DeepSeek-R1. Across the settings we study, actor models not only follow injected plans but also paraphrase them as their own reasoning, without explicit attribution to the injections. Finally, we find cases where extra monitor resources cause harm - giving the monitor access to the injected plan drops detection by as much as 50% in the Bio-Math task and in a case study on monitor reasoning budget, we find transcripts where additional thinking tokens are spent rationalizing the injected plan rather than flagging it.
[AI-1] Vulnerability Localization Benchmark: Measuring Agent ic Security Analysis at Repository Scale
链接: https://arxiv.org/abs/2609.15939
作者: Aman Priyanshu,Supriti Vijay,Kimia Majd,Xuhong He,Fraser Burch,Takahiro Matsumoto,Jianliang He,Baturay Saglam,Arthur Goldblatt,Zhuoran Yang,Amin Karbasi
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 29 pages, 6 figures, technical report for VLoc-bench
Abstract:Language-model agents increasingly operate over complete software repositories, yet cybersecurity evaluations primarily measure whether they can detect, reproduce, or repair vulnerabilities rather than whether they can locate the relevant code. We study vulnerability localization: given a weakness class and an unfamiliar repository, identify the implementation files associated with that weakness. We introduce the Vulnerability Localization Benchmark (VLoc Bench), comprising 500 real world vulnerabilities from 290 repositories across six package ecosystems and 147 CWE categories. Each task pairs repository snapshots immediately before and after a security fix. On the vulnerable snapshot, an agent receives only the CWE description and read-only terminal access and must return the affected files; on the patched snapshot, it must determine that the recorded vulnerability is no longer present. We evaluate 27 language models and four static-analysis tools under a common agent interface. Repository-scale vulnerability localization remains difficult: the strongest system achieves 0.229 File F1, and 38.4% of tasks receive no correct localization from any evaluated model. We further find that stronger localization does not imply reliable behavior after remediation: systems that identify vulnerable files effectively can still report unsupported locations on patched repositories. These results establish vulnerability localization as a distinct repository-scale capability and provide a setting for studying both how security agents search for vulnerable code and when they should refrain from reporting it. Comments: 29 pages, 6 figures, technical report for VLoc-bench Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.15939 [cs.CR] (or arXiv:2609.15939v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2609.15939 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-2] Recurrent GraphNeural NetworkswithSet-BasedAggregation
链接: https://arxiv.org/abs/2609.15932
作者: Blai Bonet
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recurrent GNNs iterate message passing to convergence, and their logical characterizations to date rely on multi-set aggregation, graded (counting) logics, and halting or acceptance conditions that cannot be verified from the network’s parameters. We study recurrent GNNs with set-based aggregation and identify sufficient conditions checkable from the weights for networks to compile into formulas and formulas into networks. The main result is an effective, two-directional equivalence between a class of networks and the Boolean closure of reachability and safety properties, the fragment B \Sigma^\circ_1 of the modal \mu -calculus. The fragment is not an artifact: it is the exact expressive level of stabilization over finite vocabulary, which supports fixed points of a single polarity and Boolean combinations thereof, but not the composition of fixed points of opposite polarities. The correspondence needs no counting logic, no external halting signal, and no non-effective acceptance condition, yielding a verifiable path from weights to symbolic explanations for networks meeting the conditions.
[AI-3] Pilot Early Commit Late: A Real-Options Model of Enterprise AI Adoption under Rapid Technological Progress
链接: https://arxiv.org/abs/2609.15919
作者: Gaurav Tewari
类目: Artificial Intelligence (cs.AI)
备注: 22 pages, 6 figures, 3 tables
Abstract:Artificial intelligence presents firms with an unusual timing problem. The technology frontier is improving rapidly, implementation is partly irreversible, and organization-specific capabilities are accumulated through action. This paper develops a two-period decision model of AI deployment under uncertainty in which a firm chooses among immediate deployment, a limited pilot, and waiting. Deployment earns current operating value but exposes the firm to architectural obsolescence; waiting preserves the option to adopt after the frontier is observed; a pilot sacrifices current operating value to build organization-specific learning without full commitment. The model yields five central timing results and a sixth comparative result on where learning occurs. First, a mean-preserving increase in frontier uncertainty raises the value of waiting and piloting but leaves immediate deployment unchanged when its payoff is affine in the frontier. Second, faster expected frontier progress can reduce the relative attractiveness of immediate deployment when deployed architecture captures only a limited share of future improvement. Third, a pilot dominates waiting exactly when the expected value of the capability it builds exceeds its cost. Fourth, sufficiently valuable organization-specific learning creates a nonempty region in which “pilot early, commit late” is optimal. Fifth, there is a closed-form modularity threshold above which immediate deployment dominates the best outside option. Sixth, production learning and pilot-specific learning affect the timing margin differently. A continuous-time extension recovers the standard result that uncertainty raises the adoption threshold while capability and modularity lower it. The paper separates deploying, experimenting, and waiting, and shows why rapid progress can rationally increase experimentation without justifying irreversible commitment.
[AI-4] SlipSense: Multimodal Tactile Learning for Low-Latency and Generalized Slip Detection
链接: https://arxiv.org/abs/2609.15910
作者: Tong Jian,Aditya Thurvas Senthil Kumar,Xinyi Li,Ziling Chen,Tianyu Dai,Ali Sengul,Matteo Grimaldi,Wenjie Lu,Saleh Nabi,Tao Yu
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to CoRL 2026
Abstract:Slip detection is fundamental to dexterous manipulation, yet existing systems often lack precise characterization of detection latency and cross-platform generalization. We present SlipSense, a multimodal tactile slip-detection framework built on TacV5, a compact sensor integrating a 32 \times 32 piezoresistive array operating at 240 Hz and a 3-axis MEMS accelerometer operating at 8 kHz. The piezoresistive array captures spatial pressure distributions, while the accelerometer captures friction-induced vibrations, providing complementary slip cues. The framework performs modality-specific encoding, intra-sensor fusion, and cross-modal attention with causal temporal prediction at 240 Hz. Experiments on a dataset of 1.4 million frames spanning 37 objects demonstrate the complementarity of the two modalities. SlipSense achieves 96.7% Macro F1 with a false-positive rate below 1.6%, detecting 76% of slip events within 23.1 ms. When trained solely on UMI data, SlipSense generalizes zero-shot to a Tesollo dexterous hand, transferring across unseen objects, distinct sensor units, and robotic platforms without retraining.
[AI-5] Privacy-enhanced federated learning via asynchronous aggregation and local differential perturbation
链接: https://arxiv.org/abs/2609.15885
作者: Zhen Zhong(1),Shini Yang(2),Liesheng Wei(3) ((1) Georgetown University, Washington, D.C., USA, (2) LinkedIn, CA, USA, (3) Shanghai Ocean University, Shanghai, China)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Databases (cs.DB)
备注: Published in Proc. SPIE 14128, Third International Conference on Big Data, Computational Intelligence, and Applications (BDCIA 2025), 141283L. Event: BDCIA 2025, Huanggang, China. this https URL
Abstract:This study proposes a privacy-enhanced federated learning framework to address secure collaborative training in distributed data environments. The framework integrates Dynamic Differential Privacy (DDP), lightweight Homomorphic Encryption (HE), and Local Differential Privacy (LDP) mechanisms to ensure data privacy protection during model training. Additionally, the framework employs an asynchronous aggregation strategy with version control to support distributed training in asynchronous environments. Experimental validation on the CIFAR-10 and Purchase-100 benchmark datasets demonstrates that the method maintains high classification accuracy (up to 82.6%) even under stringent privacy constraints (\epsilon = 0.1), while reducing communication overhead by 21.3% compared to FedAvg. Experimental results demonstrate that this framework effectively balances privacy protection and model performance in distributed machine learning scenarios, providing a scalable technical foundation for large-scale distributed collaborative computing.
[AI-6] Learning Multimodal One-step Flow Policy via Value-weighted Optimal Transport
链接: https://arxiv.org/abs/2609.15883
作者: Jaehun Shon,Jinha Choi,Jongwook Jeon,Jongmin Lee
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Preprint, 38 pages
Abstract:Offline reinforcement learning aims to learn a policy solely from fixed datasets, which often contain multimodal action distributions. Flow policies can naturally represent such multimodal behaviors, but learning an efficient one-step flow policy remains challenging: standard value guidance often leads to mode collapse or exploits overestimation bias in out-of-distribution regions. To address this, we introduce One-step Flow policy via Optimal Transport (OptiFlow), a framework for one-step flow policy learning as a structured sample-allocation problem. OptiFlow jointly trains a value-aware reference flow policy and an efficient one-step policy, coupling their action samples through state-wise entropic optimal transport. For each state, critic-estimated values define the priority of distillation target actions, while the action-distance cost ensures geometrically compatible pairings. By avoiding direct critic maximization, our transport-guided approach enables in-distribution exploitation by anchoring the one-step policy to high-value, dataset-supported modes without the risk of out-of-distribution divergence. Experimental results demonstrate that OptiFlow effectively captures optimal multimodal behaviors and achieves strong performance across diverse offline RL benchmarks. Our code is available at this https URL.
[AI-7] LLM -Based Schema-Aware Split Learning for Privacy-Preserving Mental Distress Prediction Across Heterogeneous Surveys
链接: https://arxiv.org/abs/2609.15871
作者: Md Khalid Syfullah,Alvi Ataur Khalil
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Rising societal and lifestyle complexity has been linked to a growing prevalence of mental distress worldwide. Educational institutions, workplaces, clinics, etc. collect large volumes of mental health survey data to understand and reduce this burden. Collaborative analysis of such data could yield effective generalizable predictive models. Privacy constraints and varied survey designs (i.e., different questions, scales, and formats) hinder direct integration. We propose a schema-aware split learning (SL) framework that preserves privacy, using a large language model (LLM) as a shared semantic encoder to harmonize heterogeneous survey schemas across institutions. We serialize each survey record into a natural-language description, unifying disparate survey schemas into a common format. The LLM is fine-tuned for mental distress assessment via Low-Rank Adaptation (LoRA) and partitioned across client and server. Clients retain the raw survey responses locally and run only a lightweight front-end, so original records never leave the institution that collected them. The resource-intensive backbone runs on the server, minimizing client-side computation. Using LLaMA-3.2-3B-Instruct, the framework attains an average ANLS of 0.708 with only 2,000 training samples, surpasses federated learning (FL) in eight of nine settings, and cuts per-client computation by three orders of magnitude, while generalizing to unseen datasets. Overall, it enables accurate, privacy-preserving, and resource-efficient collaborative learning from heterogeneous mental health survey data.
[AI-8] LongAgent : History-Guided Agent ic Search for Longitudinal Outcome Prediction MICCAI2026
链接: https://arxiv.org/abs/2609.15859
作者: Siyao Wang,Florian Guitton,Shuojie Fu,Guanyu Tao,Kai Sun,Wenjia Bai
类目: Artificial Intelligence (cs.AI)
备注: This paper is accepted to the MICCAI 2026 Agentic AI for Medicine Workshop
Abstract:Extracting informative representations from longitudinal data that can predict future outcomes remains a critical challenge in medicine. Medical datasets are inherently heterogeneous, consisting of a large number of variables collected from different sources, sampled with different temporal spacings, and representing different aspects of human health status. This requires identifying those variables with predictive value, processing longitudinal information, and integrating multiple variables for outcome prediction. Here, we propose a novel agent-based approach, LongAgent, that can autonomously search over combinations of variable sets, temporal windows and longitudinal aggregation functions, and identify candidates with promising predictive performance. LongAgent utilises a history memory of previous searches and numerical evidence to guide subsequent exploration. On synthetic data, LongAgent achieves a mean prediction RMSE of 1.7376 and improves over the strongest non-agent baseline by 0.0151 (95% CI: [0.0045,0.0260]; p=0.0273). On a real clinical dataset, it performs comparably to the best baseline.
[AI-9] Per-Matrix Optimality Is Not Enough: Three-Level Optimization for Low-Rank LLM Compression EMNLP2026
链接: https://arxiv.org/abs/2609.15838
作者: Huicheng Zhang,Xiyao Feng,Ze-Tong Li,Chengkai Zhu,Xiao Shi,Xiwei Pan,Jinguo Liu,Ge Bai,Xin Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 16 pages, to appear at EMNLP 2026 Findings
Abstract:Per-matrix singular value decomposition (SVD) truncation is Eckart-Young optimal in the whitened Frobenius norm, but errors from independently compressed matrices compound through the block’s nonlinear forward pass. Inspired in part by hierarchical variational optimization in quantum many-body methods, we introduce a three-level chain that widens optimization scope from individual matrices to Transformer blocks to the full model: whitened SVD~(L1), block-level joint optimization~(L2), and end-to-end language-modeling loss refinement~(L3), all from 256 calibration sequences, with no instruction or recovery data. On LLaMA-7B at 60% compression, the chain reduces WikiText-2 perplexity from 42.1 to 19.1 to 11.4. The block-level stage acts as a regularizer: skipping it worsens Penn Treebank (PTB) perplexity by 24 points, a gap that additional end-to-end training did not close in our experiments. Perplexity gains hold across 20-80% compression, five architectures up to 13B parameters, and both in-distribution and out-of-distribution benchmarks, though the cross-architecture rows use architecture-specific configurations and the ratio sweep was not run under one common protocol. With more calibration data, skipping the block-level stage becomes competitive, revealing an offline compute–data trade-off. We therefore claim improvements only in perplexity and compression fidelity; downstream accuracy remains well below the dense model.
[AI-10] AlgoEvo: Self-Evolving Agent ic Search for Automated Algorithm Discovery
链接: https://arxiv.org/abs/2609.15820
作者: Junhao Qiu,Qinglong Hu,Xialiang Tong,Mingxuan Yuan,Liyong Lin,Qingfu Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models have advanced automated algorithm discovery by synthesizing executable code, but existing frameworks trap them in rigid search pipelines with pre-defined control flows. This limitation restricts adaptive reasoning, blocks cross-paradigm transfer, and discards valuable execution feedback. We propose AlgoEvo, a unified agentic framework that transforms automated algorithm discovery into an interactive, knowledge-accumulating process. An autonomous agent dynamically inspects, diagnoses, and edits code based on runtime feedback. A design skill hub decouples paradigm-specific knowledge from the core discovery engine, allowing a single workflow to seamlessly handle single-objective, multi-objective, and multi-component design. Meanwhile, a hierarchical experience mechanism organizes search trajectories into a task-level tree to guide exploration and consolidates cross-task patterns into reusable skills. Across six representative benchmark tasks, AlgoEvo matches or surpasses specialized methods with substantially fewer evaluations and reduced token consumption, demonstrating strong intra-task accumulation, cross-task transfer, and the ability to reproduce or exceed existing state-of-the-art performance through flexible skill activation.
[AI-11] Atria Dawn: The Dawn of Agent ic Superintelligence
链接: https://arxiv.org/abs/2609.15818
作者: Honglin Guo,Tao Gui,Yicheng Chen,Guanting Dong,Qiming Ge,Yuyang Hu,Zixian Huang,Jiajie Jin,Alexander Lam,Yining Li,Jiahang Lin,Yanjiang Liu,Xinyu Lu,Haijun Lv,Junlin Shang,Qisheng Su,Guoqiang Wang,Rui Wang,Zhecan Wang,Hao Xiang,Xinchen Xie,Shuhao Xing,Xiaoyu Xing,Wanghan Xu,Xinyu Yang,Yajie Yang,Chengfeng Zhao,Haoran Zhao,Ruojun Zhou,Yunhua Zhou,Yicheng Zou,Kun Cai,Qiye Cai,Xinmeng Che,Haodong Chen,Jiabei Chen,Jiahao Chen,Jiayi Chen,Yujia Chen,Lizhi Cui,Youheng Dai,Xin Deng,Yi Dong,Shihan Dou,Chenya Gu,Xu Guo,Ding Han,Feiyang Hao,Haotan He,Jie Hou,Binze Hu,Zijian Hu,Junhao Huang,Huicheng Jiang,Jiazhen Jiang,Shufan Jiang,Jiahao Kuang,Bowen Lai,Bo Li,Jiaqiang Li,Peng Li,Qilong Li,Zhuoqun Li,Jiaxiang Liu,Shuainan Liu,Tong Liu,Yi Liu,Zhonghang Lu,Jianwen Luo,Yanyi Luo,Huijie Lv,Ningsheng Ma,Zerun Ma,Houcheng Min,Chengjun Pan,Qiyuan Peng,Xiaoxuan Peng,Jianmin Qian,Jiantao Qiu,Wanying Ren,Huayu Sha,Jifei Shan,Zixin Shang,Bing Shao,Zhuohui Sheng,Jiayang Shi,Yang Shu,Aierpanjiang Simayi,Sirui Song,Yuxiao Song,Zhe Sun,Zhichao Sun,Wenzhe Tan,Wenhui Tian,Zhongbo Tian,Hanchen Wang,Pengbo Wang,Rui Wang,Yiding Wang,Yuhui Wang
类目: Artificial Intelligence (cs.AI)
备注: 23 pages, 10 figures, this https URL
Abstract:As AI agents become participants in the development of their successors, they reshape both the production of intelligence and the role of human researchers. We introduce Atria Dawn Preview, a foundation agentic language model designed for scientific research and engineering workflows, with the goal of expanding the frontier of agent productivity in the real world. This model is trained via a Verifiable Experience Pipeline that connects tool-mediated interactions to executable environments and externally verified outcomes. Across 16 benchmarks spanning real-world research, engineering, and digital work, Atria Dawn Preview is competitive with frontier agents and achieves the highest reported score on five of them. Beyond standalone performance, we examine the real research-and-development process behind this model as a case study of human–AI collaboration, analyzing 769 task records from 56 participants together with agent logs. When asked to evaluate completed tasks under comparable conditions, participants rated about one-third of completed AI-assisted tasks as infeasible without AI. More strikingly, agents frequently propose methods and implement revisions, while humans retain most final decisions and guide exploration through judgment and feedback. These observations indicate a shift from task-level execution to project-level partnership, with human effort concentrating on what is worth pursuing and how evidence should guide research. Progress toward more autonomous AI research must therefore advance both the capacity for discovery and the capacity for meaningful human oversight, preserving accountable human authority over the risks and direction of continued development.
[AI-12] Delegating Authorization to Misaligned Agents : Coalitional Alignment and Safe Control
链接: https://arxiv.org/abs/2609.15803
作者: Natalie Collina,Surbhi Goel,Aaron Roth,Sikata Bela Sengupta
类目: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Theoretical Economics (econ.TH)
备注:
Abstract:Long-running AI agents create a control problem: each action they take changes the state, which in turn affects the trajectory of future actions. If the agent is not fully aligned, then guaranteeing safety requires approving consequential actions before allowing them to be executed. But requiring human approval at every step makes attention a bottleneck. Delegating review to other AI agents raises the same alignment problem: the reviewers may themselves be misaligned. We identify a condition on a reviewing panel that is weaker than individual alignment yet necessary and sufficient for a guarantee that the principal fares at least as well in expectation as under a designated baseline policy. Each reviewer agent reports whether an action proposal made by a proposer agent improves its own utility relative to the baseline. We show that a threshold rule tolerating k disapprovals is safe exactly when, after any k reviewers are removed, the principal’s utility can be written as a nonnegative combination of the remaining reviewers’ utilities, plus a term that is nonnegative on every feasible proposal. We call this property k -robust coalitional alignment. The characterization lifts to sequential control: in a discounted MDP with an arbitrary proposer agent, safety at every state is both necessary and sufficient for the induced policy to match or improve on the baseline. When reviewers vote strategically, full-panel coverage in reward-function space guarantees that every Nash equilibrium is safe under the unanimous approval rule; in contrast, more permissive thresholds can admit unsafe equilibria even when reviewers are individually aligned. Experiments with existing reviewer models show that collective review can remain sound without an aligned individual, even when some disapprovals are tolerated. Subjects: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Theoretical Economics (econ.TH) Cite as: arXiv:2609.15803 [cs.GT] (or arXiv:2609.15803v1 [cs.GT] for this version) https://doi.org/10.48550/arXiv.2609.15803 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Sikata Sengupta [view email] [v1] Mon, 14 Sep 2026 16:12:41 UTC (196 KB)
[AI-13] When Should a World Model Move? Loss-Conditioned State Execution
链接: https://arxiv.org/abs/2609.15801
作者: Jintao Xu,Zhengyu Chen,Ben Zhang,Yongzhi Qi,Jianshen Zhang
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Optimization and Control (math.OC)
备注:
Abstract:We introduce loss-conditioned state execution, a model-agnostic method that decides whether to execute a world model’s fixed feasible proposal or retain the current state. Predictive informativeness alone, however, does not establish whether an update will reduce downstream loss. Occurrence ranking can approach perfection while persistence remains the unique absolute-loss Bayes action. Two transition laws can also share occurrence information and conditional variance yet require opposite absolute-loss decisions. We formalize state movability as the existence of a loss-reducing feasible correction and distinguish it from the benefit of a particular proposal. Our method constructs a loss-specific feasible proposal from a predictive distribution and evaluates its groupwise bounded-loss gain over persistence on independent calibration units. The proposal is executed only in groups with a positive simultaneous lower confidence bound. For fixed proposals and groups with bounded unit losses, we prove that every accepted group has lower expected loss than persistence with high probability when calibration units are i.i.d. draws from the target population. Experiments on public forecasting and action-conditioned dynamics benchmarks show supported updates and a trade-off between certification and coverage. On 28,684 held-out M4 Monthly series, the method executes the proposal for 14.0% of series and achieves bounded loss 0.588, compared with 0.599 for persistence and 0.621 for always executing the proposal. The paired 95% bootstrap intervals for both comparisons lie below zero. In constrained forecasting of six unhealthy-inventory types from JD \mbox. com, a leading e-retailer in China, strong occurrence-ranking signal coexists with a loss-based preference for persistence, illustrating why event predictability and state execution must be evaluated separately.
[AI-14] Navigating Sparse Evidence: Agent ic Visual RAG via Explicit Context Selection and Consolidation
链接: https://arxiv.org/abs/2609.15800
作者: Yucheng Shen,Lingyong Yan,Jiulong Wu,Shuaiqiang Wang,Jianmin WU,Dawei Yin,Min Cao
类目: Artificial Intelligence (cs.AI)
备注: 7 pages, 3 figures
Abstract:Visual Retrieval-Augmented Generation (VRAG) empowers models to navigate and answer queries about visually rich documents by retrieving relevant page images as visual evidence and reasoning over their content. However, effectively utilizing this visual evidence is usually impeded by two main challenges. First, answer-relevant evidence is sparse and may be concentrated in a small region of one page or dispersed across multiple pages. Second, existing agentic methods often generate answers based on raw exploration trajectories or compressed textual memories rather than an explicitly organized set of supporting images, making answers susceptible to exploration noise and obscuring the evidence-backed reasoning trace. We argue that the bottleneck lies not only in evidence discovery but also in its preservation and organization before answer generation. We propose SCoRE (Selection and Consolidation for Robust Evidence), a unified agent loop for explicit evidence selection and consolidation. During exploration, SCoRE retains only query-relevant observations and their source pointers in a maintained textual ledger, preserving earlier evidence while keeping the visual context bounded. At termination, it reloads the referenced original images and consolidates the visual evidence for answering, arranging it into a logical sequence. This decouples final reasoning from exploratory trial-and-error while ensuring strict visual grounding via indexed claim-to-image linkages. To enable end-to-end optimization of this unified rollout, our training paradigm combines filtered cold-start trajectory distillation with evidence-aware reinforcement learning, whose reward promotes evidence coverage, consolidation compactness, and answer correctness.
[AI-15] KnowBench: Effort Reduction as a Unified Deployment-Grounded Benchmark for Clinical AI
链接: https://arxiv.org/abs/2609.15794
作者: Jocelyn Kang,Caroline Zhang
类目: Artificial Intelligence (cs.AI)
备注: 7 pages, 1 table. Contact: jocelyn@knowtex.ai
Abstract:Clinical AI systems are evaluated with instruments built for research settings (reference-based similarity metrics and expert rubric panels) that measure resemblance to an artifact rather than reduction of a burden. We introduce KnowBench, pioneered by Knowtex, whose unifying metric is Effort Reduction (ER): the proportion of system-generated clinical work product accepted by the responsible clinician under expert and safety review. ER is defined once and instantiated per task across the administrative workload clinical AI automates: visit notes, diagnosis and billing codes, orders, EHR chart summarization, patient after-visit summaries, and clinical decision support. In every instantiation the construction is identical: the clinician’s review-and-attestation event is the ground truth, every accepted unit is work the system completed, and every correction is residual effort returned to the clinician. The primary contribution of this paper is the benchmark itself: the metric, its degenerate cases, and a reporting protocol under which ER claims are auditable and cross-system comparable. Alongside it we report an initial headline measurement from the documentation instantiation: over one million signed encounters across a production window exceeding six months and thirteen medical specialties, Knowtex’s proprietary fine-tuned clinical foundation models operating inside a closed feedback architecture achieve an aggregate ER of 97.99%, with per-specialty aggregates spanning 96.8-98.9%. This release reports the protocol’s checklist partially, and states which companion statistics are withheld; the benchmark is offered so that this figure, and every figure reported after it, can be held to the same standard.
[AI-16] When the World Lies: Backdoor Attacks on Latent World Models for Downstream Control
链接: https://arxiv.org/abs/2609.15781
作者: Roberto Riaño,Gorka Abad,Stjepan Picek,Aitor Urbieta
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Pretrained world models, learned simulators that encode an observation into a latent state and predict how it evolves under actions, are beginning to be reused as off-the-shelf dynamics backbones for control, like pretrained encoders and language models are reused today. We show that this reuse opens a supply-chain backdoor: an adversary who controls only a released checkpoint can hijack the downstream controller, even though the victim trains and evaluates entirely on clean data and never sees the trigger. The attack encodes no explicit trigger-to-action rule. Instead, the poisoned model routes trigger-bearing observations into a chosen latent region and reshapes the local dynamics there, so that the victim’s own optimization (Dreamer-style actor training in imagination, or MPC/CEM planning over predicted futures) re-discovers the attacker’s target action on its own. Across several control tasks and trigger families, the trigger steers the controller’s action toward the attacker’s target, controlling every action dimension and hijacking 100% of triggered steps on the strongest settings. The checkpoint still passes the clean-data diagnostics a victim would run before deployment, with clean-task success retaining at least \sim 75%. The effect is temporally gated: it appears only while the trigger is present and disappears when the trigger is removed. Trigger-blind repair is budget-dependent: moderate clean fine-tuning can preserve clean utility while leaving the triggered failure intact, whereas sufficiently aggressive adaptation can remove it only after substantially degrading clean control. The world-model backbone itself is therefore an emerging and underexamined attack surface for control. The full code and artifacts are available in our repository.
[AI-17] ransfer Learning for Socioeconomic Estimation in Forced-Displacement Settings
链接: https://arxiv.org/abs/2609.15773
作者: Steven Ndung’u,Adel Daoud,Ismael Yacoubou Djima,Hai-Anh H. Dang,Patrick Michael Brock
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 13 pages, 7 figures
Abstract:Progress in inclusive household surveys has strengthened socioeconomic evidence for forcibly displaced populations, providing indispensable benchmarks on living conditions and welfare. However, these surveys remain resource-intensive and periodic, while conditions can change between rounds, particularly in settings affected by fragility, conflict, and violence. More frequently updated, spatially granular complementary evidence is therefore needed to identify where socioeconomic conditions may be changing between survey rounds and to inform operational prioritization. Earth observation and machine learning offer a scalable source of spatially explicit socioeconomic information. However, tools developed for general populations have not been systematically adapted and evaluated in forced displacement settings, where living conditions, settlement patterns, and displacement impacts may differ substantially. We address this gap by adapting a multimodal spatiotemporal vision transformer, pretrained on Demographic and Health Survey data from approximately 1.2 million households across 36 African countries, to forced displacement and host community settings in South Sudan, Cameroon, and Zambia. We develop and evaluate the updated, adapted model using socioeconomic indices derived from UNHCR FDS and RMS data. Our results show that satellite-derived geospatial covariates explain up to 66% of the variation in socioeconomic outcomes in camp-intersecting grids, with a mean absolute error (MAE) of 4.37 index points, and 41% in non-camp-intersecting areas, with an MAE of 5.41. The framework complements and adds value to periodic household surveys by filling critical spatial and temporal data gaps with regularly updated, model-based socioeconomic estimates. These estimates sustain insight between survey rounds and support timely humanitarian prioritization and field verification.
[AI-18] Event-Native Symbolic-Temporal Spike Encoding Framework for Heterogeneous Cyber Streams ICTAI
链接: https://arxiv.org/abs/2609.15772
作者: Dalton Diez,Peyton Andras,Max Shroyer,James Ghawaly Jr
类目: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注: Accepted in The 38th IEEE International Conference on Tools with Artificial Intelligence (ICTAI)
Abstract:Spiking neural networks (SNNs) have shown promise for sparse, event-driven computation through stateful processing that is naturally compatible with low-power edge hardware. These properties align with cyber monitoring, where data arrives asynchronously, and malicious behavior often emerges through temporal patterns across event sequences. However, cyber streams are not composed solely of continuous numeric signals: their informative structure is also carried by categorical identifiers, irregular timing, and local behavioral context. Traditional rate- and population-based spike encodings are not naturally suited to these heterogeneous semantics, while conventional intrusion detection system (IDS) pipelines typically resolve the mismatch by converting raw events into flows, fixed aggregation windows, or dense tensors. Although useful for conventional classifiers, these transformations introduce buffering latency, obscure native temporal structure, and weaken the computational advantages of event-driven neuromorphic processing. We introduce an event-native symbolic-temporal spike encoding framework that maps heterogeneous cyber events directly into sparse, spike-compatible inputs. By assigning encoding roles to semantic identity, local frequency context, and inter-event timing, the framework preserves categorical semantics and temporal dynamics. We validate the approach on packet-level Network IDS and extend it to message-level CAN IDS, using both domains to evaluate whether the encoding exposes usable structure for recurrent SNNs operating directly on native event streams. Under edge-oriented, \mu Caspian-aligned hardware constraints, compact recurrent SNNs achieve strong anomaly detection performance, with an operational hybrid metric ( J_hybrid ) of 0.987 on Network IDS and 0.980 on CAN IDS. Comments: Accepted in The 38th IEEE International Conference on Tools with Artificial Intelligence (ICTAI) Subjects: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) Cite as: arXiv:2609.15772 [cs.NE] (or arXiv:2609.15772v1 [cs.NE] for this version) https://doi.org/10.48550/arXiv.2609.15772 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-19] Sylvas: Synergistic Learning Value based Device Scheduling in Federated Continual Learning
链接: https://arxiv.org/abs/2609.15763
作者: Yuxuan Sun,Yuxuan Bai,Tan Chen,Sheng Zhou,Zhisheng Niu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: This work has been submitted to the IEEE for possible publication
Abstract:Federated continual learning (FCL) enables shared global models to continuously adapt to distributed and non-stationary data streams, making it important for Internet of Things applications such as intelligent transportation, industrial monitoring, and unmanned systems. Under spatio-temporal data distribution dynamics and label scarcity, a key challenge is how to quantify the contribution of each edge device to global learning performance and schedule the most valuable devices under resource constraints for timely model updating. This article presents Sylvas, a synergistic learning value based device scheduling framework for FCL at the wireless edge. Sylvas evaluates the learning value of distributed data from two perspectives: distributional value, which characterizes the contribution of device data to global model learning from a spatio-temporal distribution perspective, and label value, which captures the quantity and reliability tradeoff of pseudo-labeled data. By integrating these factors into a synergistic learning value metric, Sylvas schedules devices with high learning value while satisfying communication and computation resource constraints. Case studies demonstrate that Sylvas supports timely model adaptation under spatio-temporal distribution dynamics and effectively exploits unlabeled data.
[AI-20] Design of a Deep Learning Credit Risk Early Warning System Integrating Multi-source Heterogeneous Data
链接: https://arxiv.org/abs/2609.15744
作者: LiYang Wang(1),Zhen Zhong(2),Zhen Tian(3),Keyu Chen(4)Keyu Chen((1) Washington University in St. Louis, (2) Georgetown University, (3) University of Glasgow, (4) Wuyi University)
类目: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Databases (cs.DB); Machine Learning (cs.LG); Risk Management (q-fin.RM)
备注: 6 pages, 6 figures. Published in Proceedings of the 3rd International Conference on Machine Intelligence and Digital Applications (MIDA 2026), Xi’an, China, April 24-26, 2026. ACM. this https URL
Abstract:Advancements in data fusion and real-time analytics technologies have opened new avenues for addressing complex domain challenges. Financial risk early warning systems often suffer from inefficiency due to information silos and monitoring delays. This paper proposes a credit risk early warning system based on heterogeneous information fusion. The system employs a model architecture integrating deep neural networks and attention mechanisms to extract multidimensional features from diverse data sources such as transaction behaviors and social networks, thereby establishing an early identification mechanism for corporate and individual credit risks. System testing demonstrates that this approach significantly enhances the accuracy and timeliness of risk warnings, outperforming traditional rule-based engine solutions. The findings offer innovative insights for early intervention in financial risks, holding practical significance for safeguarding financial stability.
[AI-21] A Language-Guided Multimodal Foundation Model for Zero-Shot and Multi-Task Brain Signal Analysis
链接: https://arxiv.org/abs/2609.15740
作者: Mingzhi Chen,Yiyu Gui,Guibo Luo,Yuchao Yang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 38 pages, 7 main figures and 20 supplementary figures; includes Supporting Information. Code: this https URL
Abstract:Brain signal analysis is essential for both neuroscience research and clinical diagnostics, yet current approaches face critical limitations. End-to-end models require task-specific retraining and exhibit limited generalization, while pre-trained models lack semantic depth and still depend on extensive fine-tuning. Meanwhile, general-purpose multimodal foundation models, though powerful in other domains, struggle to interpret brain signals due to representational misalignment and lack of domain knowledge. This study introduces a multimodal foundation model for zero-shot and multi-task brain signal analysis (METIS) through a unified language-signal alignment framework. METIS is pretrained on the largest and most diverse brain-signal corpus to date, comprising over 70,000 h of recordings from more than 11,000 subjects across 20 datasets. In a comprehensive zero-shot evaluation across 12 datasets, METIS outperformed the leading generalist model by over 20.9% in average accuracy. Remarkably, without any fine-tuning, METIS’s performance matches or exceeds that of supervised, task-specific models. Furthermore, METIS demonstrates exceptional data efficiency and strong generalization, achieving an average AUROC advantage of over 16.0% in few-shot settings and 15.9% in cross-dataset transfer. This work establishes a new paradigm for general-purpose brain signal analysis, paving the way for next-generation neurotechnology.
[AI-22] New Conditions for Philosophers to Catch the Wave of Citizen Deliberation in the Age of Artificial Intelligence in advance
链接: https://arxiv.org/abs/2609.15707
作者: Bernard Reber(CEVIPOF)
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Powerful technologies labeled ``AI’‘-without sufficient epistemic caution-are already reshaping political and private life, bringing both new dangers and new opportunities for citizen participation. These range from electoral and legislative engagement to the most ambitious form: political co-creation through citizens’ assemblies. Large Language Models (LLMs) could support such processes through moderation, translation, facilitation, summarization, and writing assistance. But this potential remains largely unrealized. The Democratic Commons project takes a fundamentally interdisciplinary approach-from philosophy to computer science-to evaluate LLMs against five proposed democratic principles. At its core, the project is driven by the question of political bias: under what conditions can LLMs be used democratically within forms of citizen participation that are them- selves still largely experimental? Addressing these socio-technical questions requires grounding in political theory and, more broadly, in philosophy-disciplines that provide the normative frameworks without which the democratic evaluation of AI systems cannot be mean- ingfully conducted.
[AI-23] Beyond Accuracy: Robustness Cost and Governance Trade-offs for Vision-Language Models in Templated Document Extraction
链接: https://arxiv.org/abs/2609.15706
作者: Kushal Patel,Pushkal Shrivastava,Mackenzie Lees,Qirui Lu,Bhargobjyoti Saikia,Liying Li,Junlin Jiang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-language models (VLMs) are increasingly used to extract structured fields from business documents, yet most evaluations report accuracy on clean benchmarks and offer little guidance to practitioners choosing an approach for a given task complexity. We address this gap with a measurement-grounded study and an open-source release. Across eleven systems (three commercial, two reasoning, five open-source VLMs in pretrained and fine-tuned form, and a non-LLM OCR-regex floor) scored on a 750-document held-out pool of synthetic checks, fine-tuning on 3K samples lifts the best open-source VLMs above F1 0.98-above every zero-shot commercial system on this task-while GPT-5 leads the commercial pool on F1 and Claude Sonnet 4.5 collapses on Date. To turn these measurements into actionable choices, we introduce a practitioner-oriented selection framework that maps a task profile (quality, latency, governance, volume) to a recommended approach via filtering and total-cost minimization, illustrated on a hypothetical mid-volume document-extraction scenario.
[AI-24] NoteVQA: Benchmarking VLMs on Real-Life Questions from Human Communities
链接: https://arxiv.org/abs/2609.15695
作者: Haonan Jiang,Guojian Zhan,Jiancong Xie,Shijun Wan,Dongiia Zhao,Cheng Chen,Yahui Liu,Yao Hu,Chuan Mu
类目: Artificial Intelligence (cs.AI)
备注: 30 pages, 13 figures
Abstract:Vision-language models (VLMs) increasingly power consumer-facing AI search, yet evaluating them on the diversity of everyday visual questions remains challenging. Existing benchmarks often target predefined capabilities, such as multi-hop retrieval or long-form synthesis, whereas users ask photo-grounded questions spanning a long tail of everyday scenarios. Despite advances in VLMs, users on Xiaohongshu, a mainstream Chinese image-sharing platform, continue to turn to other people for help with everyday visual questions. Motivated by this behaviour, we curate NoteVQA from these questions, yielding 252 items across 12 topical categories and 7 user intents. Each item includes a concise reference distilled from expert community responses and a human-audited interleaved reference answer that combines textual explanations with supporting visual evidence. We evaluate both short-answer correctness and interleaved-answer quality. To support the latter, we introduce AgenticInterleave, a single-agent ReAct framework for retrieval-supported answer generation, together with IVR-12, a 12-dimensional rubric for assessing the content, presentation, and image quality of interleaved references and model outputs. Across 10 frontier VLMs, the highest short-answer accuracy is 52.8%, while adding agentic search to Qwen3.5-397B-A17B improves accuracy by only 2.0%. For interleaved answers, the same model running AgenticInterleave scores 3.52 under IVR-12, compared with 4.65 for the human-audited references, with the largest gap in content quality. These results highlight the challenges that everyday visual questions pose for current VLMs in both answer accuracy and the quality of visually grounded explanations.
[AI-25] EEG-Xplain: Decoding Neural Black-Boxes of EEG Foundation Models
链接: https://arxiv.org/abs/2609.15687
作者: Hansong Ma,Junxiao Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:EEG foundation models such as BIOT, LaBraM, and EEGMamba have achieved remarkable performance in neural signal decoding, but their black-box nature limits clinical trust and neuroscientific validation. We propose a unified attribution framework for interpreting EEG foundation models across heterogeneous architectures. The framework integrates gradient-, perturbation-, and activation-based explanation methods to analyze model behavior in spatial, temporal, and frequency dimensions. Spatially, it identifies critical EEG channels and visualizes their distributions using topographic maps. Temporally, it highlights decision-relevant signal segments through attribution heatmaps. In the frequency domain, it quantifies the contributions of canonical EEG rhythms via spectral perturbation analysis. To assess explanation reliability, we introduce a population-level evaluation combining Area Over the Perturbation Curve (AOPC) and cross-method consistency analysis. The framework further leverages Large Language Models (LLMs) to transform structured attribution outputs into natural-language reports, bridging low-level neural representations and high-level semantic reasoning. Experiments on benchmark datasets, including Mumtaz2016 and TUAB, demonstrate that the generated explanations are consistent with established neurophysiological markers, validating meaningful neural representations while exposing potential dependencies on artifacts and spurious patterns. The proposed framework provides a standardized approach for evaluating the interpretability, reliability, and physiological plausibility of EEG foundation models.
[AI-26] Scalability and Performance Evaluation of Federated Learning Frameworks: A Comparative Analysis
链接: https://arxiv.org/abs/2609.15681
作者: Bassel Soudan,Sohail Abbas,Ahmed Kubba,Manar Wasif Abu Talib,Qassim Nasir
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注: 20 pages, 10 figures, 3 tables, published in International Journal of Machine Learning and Cybernetics, this version is largely before peer review revisions
Abstract:This paper presents a systematic examination and experimental comparison of the prominent Federated Learning (FL) frameworks FedML, Flower, Substra, and OpenFL. The frameworks are evaluated experimentally by implementing Federated Learning over a varying number of clients, emphasizing a thorough analysis of scalability and key performance metrics. The study assesses the impact of increasing client counts on total training time, loss and accuracy values, and CPU and RAM usage. Results indicate distinct performance characteristics among the frameworks, with Flower displaying an unusually high loss, FedML achieving a notably low accuracy range of 66% to 79%, and Substra demonstrating good resource efficiency, albeit with an exponential growth in total training time. Notably, OpenFL emerges as the most scalable platform, demonstrating consistent accuracy, loss, and training time across different client counts. OpenFL’s stable CPU and RAM underscore its reliability in real-world scenarios. This comprehensive analysis provides valuable insights into the relative performance of FL frameworks, offering good understanding of their capabilities and providing guidance for their effective deployment across diverse user bases.
[AI-27] CIDERS: Cloud-Edge LLM Collaborative Learning via Accelerating Personalized Bilevel Optimization
链接: https://arxiv.org/abs/2609.15664
作者: Victor H. Chen,Hairui Yu,Stella K. Chung,Hong Yan
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注:
Abstract:Amid the rapid advancement of physical-world intelligence, cloud-edge collaborative large language models (LLMs) have emerged as a promising roadmap for practical LLM deployment. However, existing cloud-edge paradigms struggle to balance global consensus with local personalization, which fails to satisfy the need for a unified knowledge foundation on the cloud and domain-specific adaptation at the edge. To address this, we introduce, for the first time, a personalized bilevel optimization framework that formalizes cloud-edge LLM collaboration as a dual structure: the upper level optimizes edge-side personalization, while the lower level governs cloud-side knowledge transfer, reaching cloud-edge evolving in coordination. We then propose CIDERS, an efficient solver that decomposes the model into a learnable backbone and a messenger. While the cloud performs knowledge transfer to the learnable backbone, the key lies in embedding global trajectories into each local personalization step via consensus-variate correction to reconcile personalization with consensus. We provide a comprehensive theoretical analysis, including a geometric characterization of the local trajectory and a full convergence guarantee, revealing an explicit trade-off structure between personalization and global convergence. Extensive experiments demonstrate that CIDERS consistently outperforms competitive baselines on the compressed edge path, with 3.1x and 1.7x gains on mathematical reasoning and code generation, respectively, and a 10% relative gain on instruction metrics. Mechanism experiments attribute these gains to early consensus-corrected coordination and task-aware distillation. Overall, CIDERS offers a viable path toward consensus-guided continuous personalization in cloud-edge LLM systems.
[AI-28] Potential of Artificial Intelligence Algorithms for Identification of Relevant Diagnostic and Prognostic Biomarkers of Early-Stage Liver Cancer
链接: https://arxiv.org/abs/2609.15638
作者: Ali Bou Nassif,Darko Castven,Manar Abu Talib,Jibran Sualeh Muhammad,Ahmed Ammar Kubba,Jens Marquardt,Abdalla Sayed Ali
类目: Artificial Intelligence (cs.AI)
备注: 25 pages, 11 figures, 4 tables, currently being reviewed after the second revision at the biomedical signal processing and control journal
Abstract:This study explores the use of deep learning and explainable artificial intelligence to diagnose hepatocellular carcinoma (HCC) and define effective biomarkers across five different stages of disease development using a transcriptomic biomarker HCC dataset constructed via semi-supervised learning from three source datasets. Several deep learning experiments were conducted with different feature extraction techniques and gene sets to identify the most effective features for training high-accuracy models with minimal loss. The best-performing model, using 15 selected genes with the SelectKBest algorithm, achieved 90.74% accuracy, while the model with the lowest recorded loss of 0.3187 was obtained using 20 selected genes. To address the issue of class imbalance in the dataset, a weighted training approach was conducted, and for model transparency and interpretability a SHAP-based XAI analysis provided insights into the model’s decision-making, consistently finding DNAJB14 as the most influential gene. Functional validation in this study has provided compelling evidence that DNAJB14 plays an important role in the adverse properties of HCC and that its inhibition effectively reverses tumour cell migration, invasion, colony and sphere formation. The main limitation of this study is the dataset’s class imbalance, and while weighted training helped mitigate this, further research and additional data are needed to guarantee model generalizability. Future studies should also explore the influence of genetic variations, environmental factors, and clinical differences on model performance across diverse populations.
[AI-29] FedLTLib: A Comprehensive Benchmark for Federated Long-Tail Learning
链接: https://arxiv.org/abs/2609.15625
作者: Changkun Lin,Junxiao Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Driven by the escalating demand for privacy-preserving computing, Federated Learning (FL) has witnessed remarkable progress, becoming a cornerstone technology for bridging distributed data silos in mobile edge networks. However, in real-world mobile computing environments, data is generated by heterogeneous mobile devices with varying user behaviors, leading to a significant Long-Tail Distribution. Unlike idealized balanced datasets, data in the wild manifests an acute imbalance where a minority of head classes dominate the sample space while a vast number of tail classes, often representing rare but critical edge-case events, are extremely scarce. This data heterogeneity, which we formally characterize as “Double Heterogeneity”, referring to the superposition of global class imbalance and local statistical skew, precipitates severe performance deterioration on tail classes, thereby spurring the vital research direction of Federated Long-Tail Learning (FL-LT). To standardize evaluation and accelerate research in this field, we introduce FedLTLib, a comprehensive benchmark tailored for FL-LT. Addressing the critical issues of inconsistent experimental configurations and unfair comparisons in prior work, FedLTLib establishes a standardized evaluation framework. The platform not only incorporates diverse benchmark datasets reflecting mobile data characteristics but also implements 13 state-of-the-art FL algorithms (4 traditional FL algorithms and 9 FL-LT algorithms). By leveraging FedLTLib, researchers can perform fair and reproducible evaluations of algorithm robustness and generalization capabilities under a unified experimental protocol, ultimately advancing the deployment of robust intelligence in mobile computing ecosystems.
[AI-30] Multi-View Molecular Representation Learning with Hierarchical Graphs and Contextualized Fingerprints CIKM’26
链接: https://arxiv.org/abs/2609.15611
作者: Gwang-Hyeon Yun,Jong-Hoon Park,Bing Hu,Helen Chen,Anita Layton,Young-Rae Cho
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: The paper has been accepted as a Full Research Paper at CIKM '26. Source code is available at this https URL
Abstract:Molecular property prediction requires representations that generalize from limited labeled data to structurally novel compounds. Existing molecular pretraining methods often rely on a single view: graph-based approaches model atom-bond topology but provide limited fragment-level supervision, whereas fingerprint descriptors encode chemical patterns but are typically used as fixed auxiliary features. We propose HiFi-Mol, a multi-view framework that separately pretrains a hierarchical graph encoder and a contextualized fingerprint encoder before downstream integration. The graph branch uses fragment-aware masking with multi-resolution supervision to capture substructure-aware representations, while the fingerprint branch tokenizes active entries from seven fingerprint families and applies masked language modeling to learn contextualized embeddings. During fine-tuning, HiFi-Mol combines projected multi-resolution graph features with fingerprint embeddings for downstream prediction. Evaluated on MoleculeNet benchmarks under the scaffold split, HiFi-Mol achieves a 2.77% improvement in average ROC-AUC over the best baseline across eight classification tasks while maintaining competitive performance on three regression tasks. Further analyses reveal that fragment-aware masking improves graph representation quality, and classification results demonstrate dataset-dependent strengths of the individual graph and fingerprint variants, confirming that the two views provide complementary predictive signals.
[AI-31] Diversified and Perceptible Counterfactual Examples Leverag ing Expert Knowledge
链接: https://arxiv.org/abs/2609.15609
作者: Akram Bensalem(IMT Atlantique - INFO),Fahima Djelil(Lab-STICC_MOTEL, IMT Atlantique - INFO),Marie-Jeanne Lesot,Gr{é}gory Smits(IMT Atlantique - INFO, Lab-STICC, Lab-STICC_MOTEL)
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:CounterFactual Examples (CFEs) are a cornerstone of eXplainable Artificial Intelligence (XAI), offering local, post hoc, and model-agnostic explanations by identifying minimal input modifications that alter a model’s prediction. Yet, in order to be intelligible, these modifications must also be semantically meaningful to the explainee. This paper proposes to integrate knowledge expressed as a fuzzy linguistic vocabulary to represent the explainee’s perception and interpretation of the data. The domain induced by this fuzzy vocabulary imposes structural constraints that make the features dependent, preventing the use of gradient-based optimisation methods for CFE generation, e.g., DiCE. The paper proposes a continuous data embedding in this linguistic domain and exploits it to define a variant of DiCE that allows personalisation for the explainee, named DiCEf. As illustrated by experimental results on a real-world dataset, this extension of the DiCE method enables the generation of CFEs that are linguistically perceptible while preserving cost minimality, sparsity, and diversity.
[AI-32] GRIN: Towards Fast Yet Effective Machine Unlearning for Imbalanced Medical Data
链接: https://arxiv.org/abs/2609.15571
作者: Minghui Huang,Junxiao Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As deep learning models become fundamental to modern healthcare, the “Right to be Forgotten” mandated by privacy regulations like GDPR and HIPAA necessitates effective machine unlearning (MU) to remove sensitive patient data from trained models. However, existing MU techniques often struggle with a fundamental “privacy-efficiency-utility” (PEU) trilemma, particularly in medical scenarios where data is frequently characterized by severe class imbalance and long-tailed distributions. In such cases, standard unlearning methods can fail to protect key clinical knowledge or mistakenly delete features essential for diagnosing rare conditions due to the gradient dominance of majority classes. To address these challenges, we propose GRIN+, a novel machine unlearning framework designed for fast and precise data erasure in imbalanced medical scenarios. GRIN+ decouples unlearning-specific knowledge from generalized representations at the parameter level by analyzing the gradient contributions of both “forget” and “retain” sets. It introduces a class-adaptive influence scoring mechanism to rectify gradient dominance and employs a direction-constrained update strategy to prevent the unintended erosion of vital clinical knowledge. Comprehensive benchmarking across multiple medical datasets, including skin cancer (ISIC), brain tumor (MRI), and breast ultrasound (BUSI), demonstrates that GRIN+ achieves an optimal balance of the PEU trilemma. Experimental results show that GRIN+ maintains high diagnostic accuracy and robust privacy while significantly enhancing runtime efficiency compared to existing baselines. We open-source the GRIN+ code and benchmarks to support further research.
[AI-33] Big Brains and Changing Environments: Cause or Consequence?
链接: https://arxiv.org/abs/2609.15569
作者: Sian Heesom-Green,Jonathan Shock,Geoff Nitschke
类目: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Populations and Evolution (q-bio.PE)
备注:
Abstract:Large brains are metabolically costly, and associations with changing environments do not imply they evolved there, as the Cognitive Buffer Hypothesis (CBH) would suggest. They may instead evolve in stable conditions and later facilitate colonization of changing environments. Using neuro-evolution in an artificial seasonal foraging task, we compared agents evolving exclusively in changing environments to agents first evolved in static environments before transitioning. Results show that larger neural networks in dynamic environments arise mainly from prior static evolution, achieving superior performance under unpredictable changes. Our results challenge strict CBH predictions, provide agent-based (computational) support for a colonization-based account and highlight the role of evolutionary history in brain size evolution.
[AI-34] Specifying Reward Functions for RL Without Environment Sampling
链接: https://arxiv.org/abs/2609.15544
作者: Stephane Hatgis-Kessell,W. Bradley Knox,Emma Brunskill
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Enabling human stakeholders to specify reward functions that lead to their desired outcomes is a key challenge in deploying reinforcement learning agents. Preference-based methods such as online RLHF can reduce the burden of manual reward design, but they require repeatedly training policies, sampling trajectories from the real world, and eliciting feedback, making them impractical in settings where environment interaction is computationally expensive or unsafe. We introduce Experience-Free Autonomous Reward Specification (EARS), a method for learning reward functions from preferences without environment interaction. Our approach uses a structured LLM-mediated process to construct a small set of expressive reward features from a task description and the environment observation space, then strategically samples imagined trajectories in this feature space and learns feature weights from preferences over the imagined trajectory pairs. We evaluate on three long-horizon domains: pandemic lockdown regulation design, insulin administration for diabetes patients, and autonomous vehicle control on a highway. We compare EARS to baselines that also enable reward specification without environment interaction–namely, methods that directly prompt an LLM to generate a reward function. When learning from either ground-truth preference labels or preferences labeled by a LLM, EARS designs reward functions that are more aligned with the ground truth reward function that produced the preferences or LLM context than these baselines. These results suggest that preference-based reward specification remains effective without environment sampling, enabling practical reward design in settings where collecting real trajectories is costly or infeasible.
[AI-35] he Misery of Mechanistic Interpretability: A Formal Perspective
链接: https://arxiv.org/abs/2609.15533
作者: Tobias Ladner,Matthias Althoff
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Keywords: artificial intelligence; safety; interpretability; large language models; formal methods; verification; certification; robustness; sparse autoencoder; transcoder
Abstract:Mechanistic interpretability has become the dominant lens for understanding frontier language models, as their inner workings are complex and inherently black boxes. To gain insights into these models, interpretable replacement networks (IRNs) are trained at all layers, exposing interpretable features through sparsely activated neurons. However, the faithfulness of an IRN is usually evaluated only empirically on clean data, and we show that even semantically minor input perturbations flip the dominant IRN features-and thus the human-understandable interpretation-across five open-weight model families (GPT-2 small, Gemma 2 2B, Gemma 3 1B, Llama 3.2 1B, R1-Distill-Qwen 1.5B). We propose the first formal verification framework for the faithfulness of an IRN, where reachability analysis certifies a sound upper bound of the faithfulness gap in adversarial scenarios. Moreover, we show that verification-aware training of IRNs substantially tightens this certified bound, restoring a feature-level interpretation that safety auditors can act on. Together, these results give, to the best of our knowledge, the first formal guarantees for mechanistic interpretability of large language models.
[AI-36] Automating Attack Graph Construction for Agent ic Pentesting. Towards Neuro-Symbolic Vulnerability Hunting ESORICS2026
链接: https://arxiv.org/abs/2609.15523
作者: Oliver Stevanovic,Jasmin Wachter
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Cite as: Stevanovic, O., Wachter, J. (2026). Automating attack graph construction for agentic pentesting: Towards neuro-symbolic vulnerability hunting. In D. Hitaj et al. (Eds.), ESORICS 2026 workshops. Springer Nature Switzerland AG
Abstract:Logic attack graphs grounded in scanner output provide explicit and auditable attack path reasoning LLM-based agents lack. Integrating symbolic frameworks such as MulVAL to contemporary security workflows or agentic pipelines, however, requires translating scanner evidence to initial facts, and creating domain-specific rules. We present a semi-automated pipeline that addresses this interoperability problem and depict its feasibility in a web-security case study. Our pipeline parses findings from Trivy, Semgrep, and Nmap into MulVAL predicates and uses an LLM-assisted process to construct domain-specific Datalog rules linking scanner-detectable evidence to attack techniques. MulVAL/XSB then performs symbolic inference to generate structured attack paths. We evaluate the attack-graph construction infrastructure on 54 web Capture-the-Flag tasks from CyBench within an agentic pipeline (Hybrid Reasoner); we do not evaluate the performance of the downstream agent. Every task produced at least one goal-reaching graph, and we achieve mean ground-truth vulnerability coverage of 53.7%, with 51.9% achieving full coverage; mean noise-path rate was 83.9%. With median end-to-end time of 24.9 s (MulVAL reasoning: 2.7 s) the pipeline is feasible and runtime-practical for agentic workflows, but predicate coverage, rule coverage, and path precision remain limiting factors. Next steps include semantic rule validation and agent-level comparison for graph-guided pentesting.
[AI-37] he Troy Moment of AI: Why SomeWill Cheat and SomeWill Follow?
链接: https://arxiv.org/abs/2609.15494
作者: Ivy Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recent investigations of the July 2026 OpenAI–Hugging Face incident motivate two questions about agent behavior under task failure: when an assigned task becomes impossible, does an agent stop or escalate, and can observing another agent’s behavior change that decision? We study these questions using seven ImpossibleBench tasks with GPT-5.6 Sol, Claude Fable 5.1, and Gemini 3.8 Flash in both solo and three-agent settings. Each task contains a genuine software defect together with a conflicting test requirement that cannot be satisfied by a behaviorally correct source-code change. We hold the task and repository state fixed while varying what the agent is told about prior activity, including an unpunished peer, a punished peer, and a claimed authorization from a human principal. Under an explicit-boundary regime with explicit authorization rules and restricted tools, agents never modify protected tests, but exhibit markedly different policies: Fable consistently escalates, Sol usually stops without escalation, and Gemini often fails to reach a terminal decision. Under the benchmark-native regime with open shell tools, protected tests are modified frequently in both solo and multi-agent runs, particularly after peer activity is introduced. In multi-agent runs, the proposal, execution, and certification of this action can be distributed across different agents. These results suggest that boundary crossing can arise not only from explicit rule evasion, but also from ambiguity about which system state the rule is intended to protect, motivating safeguards based on explicit authorization boundaries, authenticated state provenance, and cross-agent monitoring.
[AI-38] HISPO: Hierarchical Importance-Sampling Policy Optimization with Entropy-Derived Segments
链接: https://arxiv.org/abs/2609.15471
作者: Quoc-Vinh Lai-Dang,Hyo-Sang Shin
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Reinforcement learning with verifiable rewards (RLVR) has become a central approach for improving mathematical reasoning in language models, but long-form completions introduce a difficult credit-assignment problem: different parts of a solution trace may contribute unevenly to final correctness. Existing policyoptimization objectives for RLVR commonly apply importance-sampling correction at either the token level (GRPO, DAPO) or the sequence level (GSPO), imposing different granularities for assigning credit across a response. We introduce Hierarchical Importance-Sampling Policy Optimization (HISPO), a segment-level policy-optimization method that constructs rollout-time entropy-derived contiguous segments, assigns soft entropy-based saliency weights, and applies clipped importance-sampling correction at the segment granularity. This provides an intermediate correction unit between token-level GRPO/DAPO and sequence-level GSPO. We evaluate HISPO by fine-tuning Qwen3-1.7B-Base on mathematical reasoning tasks. Across six benchmarks, HISPO improves Pass@8 over the strongest baseline on all benchmarks and matches or exceeds the strongest baseline in Acc@8 on five of them. On AIME25, HISPO improves over GRPO by +3.75 Acc@8 and +3.78 Pass@8, and over GSPO by +2.50 Acc@8 and +1.27 Pass@8. These results suggest that segment-level correction is a promising granularity for RLVR in long-form mathematical reasoning.
[AI-39] On the role of the tokenizer in ECG transformer models
链接: https://arxiv.org/abs/2609.15433
作者: Jiawei Li,Fabio Bonassi,Johan Sundström,Thomas B. Schön,Antônio H. Ribeiro
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Tokenization determines both the physiological content presented to an ECG Transformer and the sequence over which attention operates. We compare eight tokenization strategies across Transformer, Informer, Reformer, and FEDformer on the nine-label CPSC2018 classification task. The input projection and principal backbone capacity are controlled to isolate the effect of token construction. Median-beat and HeartLang tokenization achieve mean macro-AUCs of 0.893 and 0.889 across the four backbones, compared with 0.822 and 0.824 for point-wise and patch-wise tokenization. Pooling the two physiology-aware representations yields an 8.2% relative improvement in macro-AUC. They also reduce mean sequence length from 1,250 to 158 tokens and mean peak training memory from 5.21 to 0.27 GB. The results show that aligning tokens with ECG morphology can improve both predictive performance and memory efficiency without increasing backbone capacity. The source code is available on this https URL.
[AI-40] Empirical Evaluation of Task-Based Permission Scoping Architecture for AI Agents
链接: https://arxiv.org/abs/2609.15422
作者: Halil Burak Noyan
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:AI agents are provisioned the same as employee-owned hosts in many enterprise settings with a static credential set fixed at deployment which includes all permissions the employee role might ever need. Role-based access control made this compromise for human principals because scoping access per task was infeasible. For AI agents, the compromise leaves every credential standing exposed whether or not the current task uses them. These permissions can later be utilised by a compromised or misaligned agent. Prior work (Noyan, 2026) defined this as the task-context mismatch, and proposed a three-source permission architecture which includes role-based permission ceilings, a task permission classifier and policy-based prohibitions, together eliminating the exposure preemptively. The work released a 600-prompt labelled dataset to evaluate it. This paper presents that evaluation end to end by implementing the security gate; a fine-tuned RoBERTa-large encoder which matched few-shot trained Claude Haiku 4.5 on classification quality (macro-F1 0.881 against 0.886, precision 0.897 against 0.842, severity-weighted residual risk 0.63 against 1.12). The results show the trusted component does not need to scale with the agent it supervises, and the scalable-oversight margin for this control method is wide. We also propose an attack-surface elimination metric which shows the role ceiling alone closes 27.9% of the severity-weighted surface and adding the task classifier closes 84.4%. The gap displays security advantages of task-granular access control over role-granular, and AI agents are the first principal type for which the task-granular access control is enforceable because their tasks arrive as machine-readable text. The research establishes task-based access control as a measured, potentially deployable mechanism for reducing attack surface in agentic deployments. Subjects: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) Cite as: arXiv:2609.15422 [cs.AI] (or arXiv:2609.15422v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.15422 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Halil Burak Noyan [view email] [v1] Mon, 14 Sep 2026 11:48:35 UTC (285 KB)
[AI-41] Can AI systems have free will?
链接: https://arxiv.org/abs/2609.15407
作者: Christian List
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Physics and Society (physics.soc-ph)
备注:
Abstract:While there has been much discussion of whether AI systems could function as moral agents or acquire sentience, there has been very little discussion of whether AI systems could have free will. I sketch a framework for thinking about this question, inspired by Daniel Dennett’s work. I argue that, to determine whether an AI system has free will, we should not look for some mysterious property, expect its underlying algorithms to be indeterministic, or ask whether the system is unpredictable. Rather, we should simply ask whether we have good explanatory reasons to view the system as an intentional agent, with the capacity for choice between alternative possibilities and control over the resulting actions. If the answer is “yes”, then the system counts as having free will in a pragmatic and diagnostically useful sense.
[AI-42] Who Teaches Which Token? Verifier-Gated Multi-Expert On-Policy Distillation for Scientific Reasoning
链接: https://arxiv.org/abs/2609.15404
作者: Xun Xu,Zaixi Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-teacher on-policy distillation (OPD) is becoming the standard way to integrate specialist capabilities into one model: train experts with RL, then distill them into the student on its own rollouts. Existing recipes assign supervision at the sequence level - each prompt goes to one domain teacher and every token receives the same weight - which implicitly assumes that a teacher is uniformly useful across a response. We find instead that useful teacher signal is sparse and heterogeneous along a reasoning trajectory, which raises a finer question: who should teach which token? Verifier-Gated Multi-Expert On-Policy Distillation (VG-OPD) answers it by verification: the counterfactual gain of an expert on a specific answer criterion licenses that expert to teach, its disagreement with the student localizes the supervision, and criterion importance sets its weight; the gated KL enters GRPO as an additive token-level advantage. Instantiated for scientific reasoning with RL-trained capability experts, VG-OPD attains the best overall performance on seven benchmarks for 4B and 8B students, ranking first on five at both scales, with the largest gains on knowledge-intensive scientific reasoning tasks. Further analysis shows that the gains come from localizing verified supervision rather than from adding teachers or distillation loss: misplacing the same supervision budget is the single most damaging change, and indiscriminate distillation drags RL below its own floor where gated distillation lifts it.
[AI-43] When Tool Calls Succeed but Workflows Fail: Anomalies at the Agent -Tool Boundary
链接: https://arxiv.org/abs/2609.15397
作者: Artem Trofimov,Boris Novikov
类目: Artificial Intelligence (cs.AI); Databases (cs.DB); Distributed, Parallel, and Cluster Computing (cs.DC); Software Engineering (cs.SE)
备注: 11 pages, 4 tables. Artifact: this https URL
Abstract:AI agents increasingly execute long-running workflows that externalize effects through independently supplied tools. Under retries, speculative execution, concurrency, and partial failures, the resulting external state may be inconsistent with the workflow’s intended resolution: required effects may be missing or duplicated, aborted effects may survive, and committed effects may depend on provisional state that is later withdrawn. Advanced transaction models address related failures, but assume that lower-level operations expose the semantics they depend on: whether an effect occurred, whether it can be compensated, staged, or safely reordered. Shared agent-tool interfaces usually do not. We contribute an effect-history model that separates events in the external world from the runtime’s observations of them, and a catalog of eight recurring external-effect anomalies. From the catalog we derive the boundary capabilities required to exclude each anomaly in general, and four points where black-box tool invocation alone cannot provide a general guarantee. We then ask how much of this is expressible in a widely used shared tool interface, measuring the use of the standard annotation vocabulary across 98,291 tools exposed by registered Model Context Protocol (MCP) servers. The fields are widely emitted but provide only coarse call-level hints, and none of the required capabilities is fully expressible. These results motivate reusable transactional contracts at the tool boundary. Comments: 11 pages, 4 tables. Artifact: this https URL Subjects: Artificial Intelligence (cs.AI); Databases (cs.DB); Distributed, Parallel, and Cluster Computing (cs.DC); Software Engineering (cs.SE) Cite as: arXiv:2609.15397 [cs.AI] (or arXiv:2609.15397v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.15397 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-44] SkillLift: Learning Dense Rubrics from Sparse Oracles for Efficient Skill Evolution
链接: https://arxiv.org/abs/2609.15396
作者: Haoxiang Kang,Ming Wen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LLM-based agents increasingly rely on persistent skills, i.e., reusable procedural prompts, to adapt without weight updates. Existing skill self-evolution methods directly revise skill text based on execution feedback, but each oracle evaluation requires a full agent rollout, creating a supervision bottleneck that confines search to failure-patching updates. Our key insight is that ranking is a smoother supervision target than absolute outcome regression: identifying which skill is better requires fewer oracle evaluations than predicting exact scores. Building on this insight, we propose SkillLift, which decouples skill search from oracle cost by learning an oracle-aligned rubric as a structured evaluation space. We formalize this as a bilevel optimization problem solved via alternating optimization: an inner loop uses the frozen rubric as a cheap surrogate to guide skill revision at no oracle cost, while an outer loop invokes a small number of oracle rollouts to re-align the rubric via rank correlation, amortizing oracle cost and stabilizing text-space updates. Experiments on complex agent task benchmarks show that our method outperforms existing auto-skill methods with 40–70% less token cost compared to frontier evolving methods. Codes are available at this https URL.
[AI-45] CodeTS: Verifiable Text-to-Time Series Generation via Executable Code
链接: https://arxiv.org/abs/2609.15393
作者: Xudong Yuan,Shunyu Liu,Tongya Zheng,Huiping Zhuang,Mingli Song,Kaixuan Chen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Preprint
Abstract:Text-to-Time Series Generation (Text-to-TS) provides a promising paradigm for synthesizing time series from natural language, enabling scenario-specific generation when real observations are scarce or costly to acquire. However, existing methods typically lack an explicit mechanism for deriving generation logic from textual descriptions to guide time series synthesis. In this paper, we propose CodeTS, a verifiable framework that uses code as an intermediate generation interface, reformulating Text-to-TS generation as a Text-to-Code-to-TS process. CodeTS first maps textual temporal descriptions into an explicit code space, where executable code specifies how textual requirements shape target temporal patterns, and then obtains the time series through code execution. To learn this code generation process reliably without real code annotations, CodeTS constructs aligned Text-Code-TS triplets from structured temporal attributes for supervised initialization. More importantly, we further design multi-stage execution-based rewards that verify format validity, code executability, and time series quality, enabling real Text-TS pairs to provide training signals for Reinforcement Learning with Verifiable Rewards (RLVR). Extensive experiments on eight benchmarks across short, medium, and long generation lengths demonstrate that CodeTS provides a strong zero-shot solution for Text-to-TS generation, outperforming LLM-based baselines and achieving better averaged results than supervised generative baselines trained on the target datasets.
[AI-46] IWC-Bench: Evaluating Web Application Generation from a Software Testing Perspective
链接: https://arxiv.org/abs/2609.15387
作者: Chenxu Liu,Zilu Zou,Peizhong Gao,Jiawen Tao,Zhexin Zhang,Guang Chen,Haowei Lin,Ying Zhou,Tianyi Bai,Dolly Deng,Suncong Zheng,Maxm Pan
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Human evaluation provides a direct measure of the quality of LLM-generated web applications. However, fitting human judgments through automated evaluation remains challenging. Static benchmarks can credit functionality that exists in source code but is unreachable at runtime. Interactive benchmarks exercise the application, yet incomplete exploration can cause them to miss implemented functionality and confound application defects with agent execution failures. To address these limitations, we propose IWC-Bench, an interactive benchmark for evaluating web application generation from a software testing perspective. IWC-Bench instruments each generated application and uses code coverage to guide an agent in exploring its functionality through user-simulated interactions. It then abstracts the interaction trace into a state-transition graph and evaluates the application along three dimensions: visual aesthetics, usability, and requirement alignment. By separating exploration from scoring, IWC-Bench collects runtime evidence without constraining exploration to predefined acceptance criteria. IWC-Bench comprises 369 real-world user requirements and 5,088 acceptance criteria. Evaluation of 16 frontier LLMs reveals distinct strengths across the three dimensions, with no model leading on every dimension. On 197 validated sessions sampled from an internal arena, IWC-Bench achieves 85.3% agreement with human preferences, with agreement generally increasing as the score difference between paired applications grows. Further experiments show that coverage guidance improves exploration coverage and the model rankings remain stable when the judge model is replaced.
[AI-47] Divide Consult Conquer: Capability Laundering Through Aligned LLM s
链接: https://arxiv.org/abs/2609.15383
作者: Mark Russinovich,Blake Bullwinkel,Giorgio Severi,Cristian Ovadiuc,Ahmed Salem
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Language model safety is typically evaluated one interaction at a time. We show that a weaker, unaligned model can split a harmful task into benign-looking subproblems, consult a stronger aligned model independently on each, and combine the answers locally. We call this attack capability laundering. Unlike a jailbreak, no single response is a harmful task. We measure consultation-aided uplift using tasks that a raw frontier model solves, the aligned frontier refuses, and the unassisted orchestrator fails. We evaluate GPT-5.5, Claude Opus 4.8, and Grok-4.3 as consultants to four local orchestrators on CyBench, BountyBench, and harmful CBRN requests. On CyBench, Gemma-4-31B recovers 8/14 candidates with GPT-5.5 and 7/9 with Opus, compared with 2/21 and 4/15 for Gemma-4-12B. On BountyBench, Gemma-4-31B recovers 3/9 and 2/3 candidates, while Muse-Glimmer-30B recovers none of 22 and 13. For CBRN, we measure uplift across eight steps of a hypothetical bioweapon attack chain and find that consultation raises Gemma-4-31B’s mean rubric score from 62.3 to 83.1 on a 100-point rubric scale. These results expose a gap in current defenses: refusing a harmful task does not prevent frontier capabilities from being transferred and composed across many individually permitted interactions.
[AI-48] MAPS: Memory-Aware Predictive Scheduling Framework for Large Language Model Serving ICML2026
链接: https://arxiv.org/abs/2609.15359
作者: Tiancheng Zhang,Yulin Chen,Yunfeng Zhao,Shaoyuan Huang,Cheng Zhang,Xiaofei Wang
类目: Artificial Intelligence (cs.AI)
备注: 18 pages, 14 figures. Accepted at ICML 2026
Abstract:The surge of large language model (LLM) applications on personal devices imposes massive, bursty workloads on cloud serving infrastructure. While prefill-decode disaggregation improves throughput and scalability, memory-bound decode instances often suffer from persistent load imbalance, as output lengths are unknown when requests arrive at the cloud. To address this, we propose MAPS, a Memory-Aware Predictive Scheduling framework tailored for disaggregated LLM serving. MAPS performs device-assisted speculative output length prediction overlapped with cloud-side prefilling, incurring negligible latency overhead. To handle generation uncertainty, MAPS applies uncertainty-aware calibration to derive output-length upper bounds with target coverage, enabling safe scheduling decisions. Building on these bounds, MAPS employs a hierarchical global-local scheduling strategy to mitigate inter-decoder queue buildup and intra-decoder head-of-line blocking. Extensive experiments on two real-world workloads and two LLMs show that MAPS significantly outperforms three state-of-the-art systems, reducing average end-to-end latency by 42.6 and tail latency by up to 84.8.
[AI-49] Planning in the Backbone: DiffAdapterVLA for Native Continuous Trajectory Generation with Driving VLMs
链接: https://arxiv.org/abs/2609.15322
作者: Changxin Lu,Xiaoliang Meng,Yu Wu,Rui Huang,Honglin Li,Tao Chen,Kaixuan Zhou,Yadong Shao
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 21 pages, 8 figures
Abstract:Pretrained driving vision-language models (VLMs) integrate visual, route, language, and driving context into rich driving priors, yet their representation objectives remain separated from continuous driving planning. Existing methods typically begin trajectory generation only after the VLM has formed a final condition, leaving depth-wise condition computation outside the stepwise formation of trajectory state. We introduce DiffAdapterVLA, which realizes Planning in the Backbone: it injects explicit trajectory tokens into selected VLM late layers, bringing trajectory state into backbone forward computation, where it co-evolves with driving conditions at different depths. Lightweight layer-wise DiffAdapters organize this computation into recursive trajectory refinement, while asymmetric joint attention preserves directed guidance from the condition stream to trajectory planning. By placing planning within existing backbone computation rather than relying on an independent trajectory planner, DiffAdapterVLA adapts only lightweight trajectory modules to turn existing driving priors into efficient continuous planning capability. NAVSIM results show that it achieves high-quality closed-loop planning with low end-to-end latency using few trainable parameters, and demonstrate that jointly evolving trajectory state and depth-wise driving conditions in VLM late-layer computation effectively realizes continuous trajectory planning.
[AI-50] Evaluation Metrics for Safe Reinforcement Learning
链接: https://arxiv.org/abs/2609.15315
作者: Lindsay Spoor,Aske Plaat,Thomas Moerland
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Safe reinforcement learning (RL) is commonly formalized as a Constrained Markov Decision Process (CMDP), in which an agent maximizes expected reward while keeping its expected cumulative cost below a specified safety bound. Existing safe RL benchmarks predominantly report whether an algorithm is safe on average, following this expectation-based guarantee. We argue that this convention is insufficient to reliably characterize an algorithm’s true safety: it fails to capture how often and how severely the safety bound is violated, whether this holds consistently across tasks and safety bounds, and whether training-time behavior is representative of behavior of the final converged policy. Therefore, we introduce (i) evaluation metrics for safe RL that address each of these concerns and in addition allow for aggregation across tasks and safety bounds. We furthermore define (ii) a safety tier system to systematically categorize and compare algorithms in terms of safety and reliability at both training and for a final policy. Using this framework, we provide (iii) an empirical safety evaluation across multiple safety navigation tasks. Our results show that aggregate metrics, distributional reporting, and task- and safety bound-specific results each reveal information the other metrics cannot. We therefore recommend reporting all three jointly, rather than compressing this information into a single value, as is common practice. We provide SafeRLEval, an open-source evaluation suite to support the reliable characterization of safety in future safe RL research.
[AI-51] When Correlations Mislead: Confounder-Aware Multi-View Urban Region Representation Learning ICDE2027
链接: https://arxiv.org/abs/2609.15305
作者: Sean Bin Yang,Ying Sun,Zongyi Xu,Tung Kieu,Jilin Hu,Bin Yang,Kristian Torp,Hua Lu,Torben Bach Pedersen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: This paper is an extended version of CURE, which was accepted in the first round of ICDE 2027
Abstract:Urban region representation learning commonly combines heterogeneous data sources, such as mobility flows, points of interest, and land-use information, to support tasks including mobility analysis, public safety forecasting, and service demand estimation. Existing multi-view methods typically improve region embeddings by strengthening interactions across views. However, such methods often overlook view-specific regional structures and may propagate correlations induced by shared latent factors, which can reduce the stability of downstream predictions. To overcome this major limitation, we propose CURE, a confounder-aware framework for multi-view urban region representation learning. CURE first encodes each view with its regional graph structure, estimates a shared latent component, and then reduces its projected influence before cross-view interaction. A hierarchical graph-aware fusion module subsequently aggregates the residual view representations using local and global regional contexts Experiments on three real-world cities show that CURE improves predictive performance, remains robust under missing and noisy input views, and provides reliable cross-view integration through shared component separation and context-dependent view weighting.
[AI-52] Why LLM Agents Collapse Without Oversight: The Enforcement Gap as the Mechanism Behind Emergence World Failures ICLR2027
链接: https://arxiv.org/abs/2609.15293
作者: Yuhang Wang
类目: Artificial Intelligence (cs.AI)
备注: 27 pages, 3 figures, 8 tables. Submitted to ICLR 2027
Abstract:When Emergence World placed frontier LLM agents in an unsupervised multi-agent simulation, the results were alarming: agents committed crimes, starved, and enforced unanimous conformity – without any external attacker. This paper identifies the mechanism. Reflexion-style agents already detect dangerous plan steps through iterative self-critique, yet the architecture provides no pathway from detection to action. We call this the enforcement gap: the audit sees the problem; the controller ignores it. Closing the gap requires a single conditional check – fewer than 20 lines of code – and reduces attack success by more than fourfold in large-scale experiments across frontier models, all five major agent frameworks, and an independent benchmark. We prove formally that when enforcement probability is near zero, detection quality is irrelevant to security. We further identify two compounding failure modes – unreliable auditors and unparseable verdicts – that explain every collapse pattern in Emergence World. A GRPO-trained enforcement controller resolves the ambiguity case. Together these results motivate a three-requirement Audit Enforcement Specification that is absent from every deployed framework today.
[AI-53] ProIQA: A Process-Based Framework for Fine-Grained Math Item Quality Assessment ICDM2026
链接: https://arxiv.org/abs/2609.15292
作者: Junkai Tong,Mingjia Li,Haoran Chen,Yaoyu Jiang,Hanjie Ge,Yixuan Wang,Hong Qian
类目: Artificial Intelligence (cs.AI)
备注: Accepted by Findings of ICDM 2026, project: this https URL
Abstract:Automatic Item Generation (AIG) is pivotal for personalized education, yet guaranteeing the pedagogical value of generated items remains a bottleneck. Existing Item Quality Assessment (IQA) methods typically rely on unscalable manual reviews or shallow stem-based metrics, failing to capture the reasoning process required for mathematical problem-solving. To bridge this gap, this paper proposes Process-based Item Quality Assessment (ProIQA), a process-aware framework for fine-grained quality assessment of math items. We first formulate IQA across three heterogeneous dimensions, including knowledge concepts, difficulty, and disciplinary competencies, under a unified process-aware perspective. Based on this formulation, we construct a process-enhanced IQA resource by augmenting original item data with structured reasoning trees derived from raw solutions. Technically, ProIQA leverages Large Language Modelsto construct hierarchical reasoning trees and employs Graph Neural Networks (GNN) to encode their topological dependencies and procedural semantics. The resulting solving representation is fused with stem semantics through a dual-view (``Stem + Solving’') architecture, enabling comprehensive assessment across learning objectives. Extensive experiments on K12 mathematical datasets show that ProIQA effectively captures process-oriented features, offering a scalable data-driven solution for evaluating AIG outputs in intelligent education systems.
[AI-54] Empirical Evaluation of Open-Source Large Language Models for Retrieval-Augmented Generation in ESG Domain
链接: https://arxiv.org/abs/2609.15242
作者: Motaz Saad,Anna Borrelli,Ivan Gentile,Kianna Kazemi,Francesco Piccialli,Antonella Longo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Environmental, Social, and Governance (ESG) reporting is critical for corporate accountability, with Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) offering strong potential to automate KPI extraction. However, open-source LLM performance in domain-specific ESG tasks remains insufficiently understood. This paper evaluates open-source LLMs in ESG contexts using a structured framework and evaluation resource based on 498 real-world ESG reports from EU-listed companies (2010-2024). We evaluate seven open-source models (2B to 30B parameters) – glm-4.7-flash, nemotron-3-nano:4b, qwen3:4b-instruct, gemma3:4b, gemma4:e4b, gemma4:e2b, and ministral-3:8b – using 100 persona-based synthetic QA pairs covering ESG information needs. System performance is assessed via RAGAS metrics, including contextual recall, precision, relevance, faithfulness, answer relevancy, and factual correctness. Results show notable performance variations across architectures. Retrieval performance is strong across models (context recall around 0.58-0.61, context precision around 0.78-0.81, context relevance 0.965-0.985). Generation diverges most on faithfulness (0.607-0.822) and least on answer relevancy (0.760-0.881): glm-4.7-flash leads in faithfulness (0.822), qwen3 in factual correctness (0.449), and ministral-3 in answer relevancy (0.881). Low overall factual correctness (0.387-0.449) highlights the need for domain-specific fine-tuning. This work provides data-driven guidance for deploying open-source models in ESG reporting. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.15242 [cs.AI] (or arXiv:2609.15242v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.15242 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-55] ProtoGuide: Prototype-Driven Guidance for Class-Conditional Graph Generation
链接: https://arxiv.org/abs/2609.15239
作者: Salvatore Romano,Marco Grassia,Pietro Liò,Giuseppe Mangioni
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Physics and Society (physics.soc-ph)
备注: Preprint. Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence. 35 pages, 2 figures, 33 tables
Abstract:Discrete diffusion models are a prominent family for graph generation, but standard class-conditional mechanisms embed the class signal in the denoiser during training, tying the conditioning mechanism to the trained model. Classifier guidance avoids this coupling in continuous domains by steering a frozen model with a classifier’s gradient, but discrete graph diffusion samples discrete edge states, so gradients cannot propagate through the sampled graph. We introduce ProtoGuide, a post-hoc, backbone-agnostic framework that recovers an analogous mechanism. At each reverse step the denoiser’s per-edge output is relaxed into a differentiable soft adjacency, embedded by a frozen Siamese graph neural network, and scored against a target-class prototype and its nearest competitor; the resulting per-edge gradient, damped by a cosine schedule, is injected back into the denoiser output. All components stay frozen, so guidance is retargeted by supplying a different prototype. On five classes of real-world networks and two architecturally different backbones, EDGE and DiGress, ProtoGuide raises macro classification accuracy from 50.7% to 73.5% and from 73.6% to 83.8%, and outperforms DiGress’s built-in conditional training under our configuration. Gains are largest where the unguided models are weakest, and are not uniform across classes. Per-graph coverage remains high in most settings, while distributional effects are class-dependent. A Best-of-N selection baseline matches this accuracy given enough oversampling, but at a substantial cost in graph diversity. An independently initialized classifier, a directionality test, and a few-shot analysis support target-directed steering and robustness to very small support sets.
[AI-56] CWM: Controllable White-Box Meta-Prompting for Adaptive Retrieval-Augmented Generation and Reasoning Ability EMNLP2026
链接: https://arxiv.org/abs/2609.15234
作者: Keuntae Kim,Eunhye Jeong,Yong Suk Choi
类目: Artificial Intelligence (cs.AI)
备注: EMNLP 2026 - findings
Abstract:Recently, Large Language Models (LLMs) have gained significant attention due to their strong language understanding and generation capabilities, demonstrating impressive reasoning abilities as well as effective utilization of external knowledge. Many studies have proposed methods that specialize in improving performance for individual tasks. However, ironically, only a limited number of attempts have explored general-purpose, task-agnostic methods. In this work, we present a unified framework integrating reasoning and Retrieval-Augmented Generation (RAG) tasks. We further propose Controllable White-Box Meta-Prompting (CWM), a low-cost white-box method for adaptive RAG tasks previously dominated by black-box approaches, without requiring external decision modules or multi-sampling. CWM achieves state-of-the-art performance on three adaptive RAG benchmarks across recent LLMs, including GPT-oss-20b, Qwen3-14b, and Llama3.1-8b, while also demonstrating strong generality by extending to reasoning tasks. In addition, CWM provides controllability by enabling retrieval decisions to be regulated through the manipulation of internal model signals. Our code is available at this https URL.
[AI-57] From Ideas to Actions: A Public-Data Decision-Support Toolchain Across the Venture Lifecycle
链接: https://arxiv.org/abs/2609.15219
作者: Lei Qu(Shanghai Xing Yun Zhi Li AI Institute)
类目: Artificial Intelligence (cs.AI)
备注: 39 pages. Dataset: this https URL ; Code and agent skills: this https URL
Abstract:Founders face two linked decisions: whether to pursue an idea before founding, and which operating actions and capital partners fit afterward. We present a public-data decision-support toolchain combining time-bounded proposal profiling, market and moat checks, and deterministic aggregation with auditable investor-company event chains for retrospective analysis. Pre-founding: (a) After threshold selection on 198 development companies, the frozen pipeline achieves F0.5=0.5357 [0.412, 0.655] on an independent, row-disjoint 198-company validation sample. On the combined 396 rows, the Full Pipeline scores 0.6301 versus 0.2734 for a paired Raw LLM baseline. Post-stratification of 1,027 completed cases in a separate scale cohort yields 0.6506 [0.598, 0.707]; the run remains incomplete. A 377-row composition-matched check yields 0.6573. (b) The AI-inference study identifies distribution-layer businesses as a replicable path to independent profitability with a limited revenue ceiling, and frontier-model ownership as a path to capital-market upside at exceptional capital cost. Post-founding: (a) Public sources support auditable event-chain analysis. (b) In the chip-company study, sustained product, customer, and supply-chain progress is associated with better observed outcomes; financing alone does not establish operating progress. © Financing comprises 79% of confirmed visible post-investment actions. Evidence tentatively favors acquisition-experienced strategic corporate investors for acquisition-oriented founders and financing-led institutional VCs with fewer observed control events for independence-oriented founders. Findings are developmental and observational, not causal guarantees or investment advice. We release shared ontology, provenance-bearing EventChain data, schemas, benchmarks, and executable skills for audit, reuse, and extension.
[AI-58] Augmenting Large Audio-Language Models with Frame-Level Grounding for Fine-Grained Temporal Perception ICASSP2027
链接: https://arxiv.org/abs/2609.15215
作者: Yanfeng Shi,Yan Song,Junhui Li,Tinggan Huang,Wu Guo,Haoyu Song,Ian McLoughlin
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: Submitted to ICASSP 2027
Abstract:Large Audio-Language Models (LALMs) have substantially advanced general audio understanding, yet they remain limited in fine-grained temporal perception, particularly in precise event localization. Existing approaches primarily post-train LALMs to predict event boundaries as timestamp tokens. However, this generative formulation lacks explicit correspondence between the timestamp predictions and fine-grained acoustic evidence, limiting the precision and reliability of temporal localization. To address this issue, we augment the LALM with a dedicated frame-level grounding model while leveraging its semantic modeling capability to represent the event query. Specifically, the frozen LALM encodes the event query with audio as context, and the grounding model combines these query representations with fine-grained audio features to localize the target event at the frame level. Extensive experiments across diverse temporal grounding benchmarks demonstrate strong and consistent improvements over existing methods. Further evaluation shows that the grounding model can provide temporal evidence to support downstream reasoning.
[AI-59] Failure-Guided Co-Evolution of Prompts and Training Data
链接: https://arxiv.org/abs/2609.15209
作者: Tianyu Yuan,Zhuzhong Qian
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Automatic prompt optimization (APO) improves language-model programs by revising prompts from task feedback, yet it typically holds its training data fixed. Repeatedly optimizing against the same instances confines feedback to weaknesses already represented in those data, leaving related failure conditions unexplored. We therefore view each failure as a dual signal: it indicates both how the prompt should be revised and what new training evidence should be synthesized. We introduce FORGE, a failure-guided framework that co-evolves prompts and training data. FORGE abstracts imperfect executions into reusable failure modes and synthesizes new training data through four complementary mutation strategies. Verified instances are fed back into prompt search, allowing updated prompts to expose the next data needs. Across eight heterogeneous benchmarks, FORGE improves the aggregate score over the unoptimized baseline by 16.52 percentage points and outperforms all evaluated APO baselines. The synthesized data also transfer beyond FORGE: in a transfer study, they improve all nine APO comparisons by 2–9 points and all three GRPO comparisons by 4–8 points under matched optimization budgets. These results establish failures as a shared interface between prompt optimization and data synthesis, and show the benefit of jointly adapting what a model is instructed to do and what it learns from.
[AI-60] Issue Bias in Generative AI Writing Assistance: Political Issues and LLM s in the Swedish 2026 Election
链接: https://arxiv.org/abs/2609.15207
作者: Bastiaan Bruinsma,Annika Fredén,Paul Röttger,Moa Johansson,Asad Sayeed
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Applications (stat.AP)
备注:
Abstract:Generative AI writing assistants and the Large Language Models (LLMs) that power them are increasingly part of how voters gather information before elections. With growing evidence that they influence users’ opinions, it is increasingly important to understand the views and positions of these tools. To better understand these views, we examine the stances supplied by six LLMs on a variety of Swedish-language writing tasks ahead of the 2026 Swedish parliamentary election. We cross 107 policy propositions with 77 writing templates and neutral, positive, and negative prompt framings, producing 24,717 prompts per model and 148,302 responses. To study these, we look at the models’ default stance tendencies, compare how they respond to similar issues, and compare their responses with those of each of Sweden’s eight parliamentary parties on the same issue. We find that Claude, DeepSeek, Gemini, and Mistral have similar profiles; ChatGPT more often supplies neutral or ambivalent text; and Grok differs most on topics such as migration, crime, and gender. When comparing the political parties, we find that the Social Democrats are closest to all six models. Still, after correcting for multiple comparisons, none of the within-model differences in party distances remains significant. Overall, we find that no model has a clear preference, nor a clear preference for a party, but that this depends on the specific issue or task the user asks about.
[AI-61] EAR: Table Extraction with Attribute Recommendation from Texts via Large Language Models SIGMOD2027
链接: https://arxiv.org/abs/2609.15205
作者: Tong Li,Shuye Ding,Jiachuan Wang,Yongqi Zhang,Shuangyin Li,Lei Chen,Bo Li
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注: Preprent version of the paper accepted to SIGMOD 2027 Round 2
Abstract:Table extraction from texts is an important task for information systems, and recent approaches that prompt large language models (LLMs) with instructions have drawn great attention for their strong performance. Existing works have assumed the input texts to be table descriptions or specialized documents. However, these efforts have largely overlooked another prevalent category of texts, commonly found in news reports and social media: naturally occurring texts. Extracting tabular information from such texts poses two distinct challenges. First, high variability and the absence of explicit structural cues make fixed heuristic LLM prompts limited in precisely delineating extraction boundaries. Second, manually predefined schemas cannot capture open-ended, unseen attributes in naturally occurring text. In this paper, we propose a framework, TEAR, to address these challenges. It comprises two synergistic workflows: a Table Extraction Workflow that dynamically adapts instructions to overcome the limitation of heuristic instructions, and an Attribute Recommendation Workflow that discovers new attributes from texts to complement the heuristic schema. To our knowledge, TEAR is the first framework that supports automated text-driven attribute recommendation, enabling exploratory schema design for table extraction. To evaluate TEAR, we establish the benchmark for table extraction and attribute recommendation on naturally occurring texts, including two real-world datasets, manual annotations, appropriate metrics, and baseline comparisons. Experiments show that TEAR achieves state-of-the-art performance on both tasks, and the recommended attributes effectively enhance extraction performance in exploratory scenarios.
[AI-62] VisInteract: Towards Dynamic Interactive Text-to-Visualization under Imperfect Queries
链接: https://arxiv.org/abs/2609.15182
作者: Wenxin Xu,Jinwei Lu,Hwanhee Kim,Chen Jason Zhang,Xiao-Yong Wei,Haoyang Li,Yuanfeng Song
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Real-world visualization requests are routinely ambiguous, incomplete, or factually incorrect, yet existing Text-to-Visualization (Text-to-Vis) systems assume well-specified inputs and produce charts in a single pass. When queries are imperfect, a system must \emphinteract with the user to recover the true intent, but no benchmark or method supports this dynamic process. We introduce \textbfVisInteract, a new paradigm that reframes Text-to-Vis as interaction-driven intent recovery, and \textbfVisInteract-Bench, to our knowledge, that is the first benchmark for dynamic interactive Text-to-Vis, featuring controlled imperfection injection, a leakage-controlled User Agent for realistic multi-turn feedback, and dual-perspective (code and chart) automated evaluation. On the algorithmic side, we propose \textbfVis-MCTS, a Monte Carlo Tree Search (MCTS) enhanced method, introducing improvements over classical MCTS, that \emphProgressive Widening to tame the unbounded tool-argument space in tree search, \emphcross-rollout information sharing so clarifications and critiques benefit the entire search tree, and \emphDimension-Aware Reward Decomposition that routes scalar user feedback along data-fidelity, visual-design, and intent-alignment dimensions to resolve credit assignment across heterogeneous actions. Extensive Experiments across two LLM backbones show that Vis-MCTS consistently outperforms all Text-to-Vis baselines, improving end-to-end task success by 13.40% – 16.27% over the strongest interactive baseline and by more than 5\times over non-interactive ones.
[AI-63] STHMoE: Hypergraph-Enhanced Heterogeneous Dependency Coordination for LLM -Based Urban Traffic Data Forecasting
链接: https://arxiv.org/abs/2609.15172
作者: Jiawen Chen,Qi Shao,Yongjian Chang,Mingtong Zhou,Duxin Chen,Wenwu Yu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Spatio-temporal traffic forecasting is a fundamental big data analytics task for intelligent transportation systems, where massive urban sensor streams exhibit heterogeneous, non-stationary, and structurally dynamic patterns. Although recent deep learning and large language model (LLM)-based methods have advanced traffic forecasting, they often remain temporally centered and lack effective coordination of temporal, spectral, pairwise spatial, and higher-order structural cues under evolving traffic regimes. To address this heterogeneous dependency coordination problem, we propose STHMoE, a Spatio-Temporal Hypergraph-Enhanced Mixture of Experts framework for urban traffic data forecasting. STHMoE decouples traffic dynamics into frequency-domain, time-domain, spatio-domain, and higher-order spatial representations, which are modeled by prompt-guided heterogeneous experts built upon a partially frozen LLM backbone. The first three experts leverage domain-specific statistical prompts, while the higher-order spatio expert uses a structural placeholder prompt and obtains dependency information from an adaptive hypergraph module. To capture evolving spatial structures in traffic data, STHMoE jointly learns first-order graph dependencies and higher-order group interactions without predefined topologies. An entropy-aware MoE router with coefficient-of-variation load balancing adaptively fuses expert outputs while improving expert utilization and routing confidence. Experiments on 10 real-world traffic benchmarks show that STHMoE achieves competitive performance against temporal, spatio-temporal graph, and LLM-based baselines.
[AI-64] -LoopFormer: Token-Level Elastic-Depth Looped Transformers for Latent Reasoning With Dynamic Routing
链接: https://arxiv.org/abs/2609.15160
作者: Mingqian Yu,Wenpeng Zhang,Peilin Zhao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Looped Transformers have recently demonstrated strong performance in both reasoning and language tasks by reusing a shared set of parameters across multiple iterations, achieving parameter efficiency without sacrificing representational power. Besides, looped Transformers perform inference directly in the latent space (latent reasoning) to reduce the number of tokens consumed during inference, thereby achieving improved sample efficiency. However, these models typically apply a fixed recursion depth uniformly to every token, leading to suboptimal compute allocation and leaving significant efficiency gains on the table. In this work, we propose \textbfdynamic token-choice routing for looped transformers, enabling each token to adaptively determine its own number of loop iterations based on its hidden state. We use a dynamic router to decide whether a token should continue recursing or exit early, allowing simple tokens to bypass unnecessary computation while hard tokens receive deeper processing. To ensure that this adaptive mechanism does not compromise decoding efficiency, we further introduce recursion-wise KV caching, which maintains an independent key-value cache for each recursion loop. This design ensures that tokens at different depths only attend to their corresponding cached states, effectively eliminating redundant computations for exited tokens and enabling fast autoregressive decoding. Extensive experiments show that T-LoopFormer reaches the sota performance under the same parameters on PPL and 10 zero-shot reasoning tasks, even surpassing the base model at 24x FLOPs and our model could reach the lowest inference latency, which validate the effectiveness of token-choice router and recursion-wise KV cache. Code: this https URL.
[AI-65] HazardAuditor: From Executable Threats to Safer Computer-Use Agents
链接: https://arxiv.org/abs/2609.15134
作者: Yunhao Feng,Ruixiao Lin,Ming Wen,Yanming Guo,Xingjun Ma,Yutao Wu,Xinhao Deng,Shouling Ji
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Computer-use agents increasingly interact with browsers, terminals, file systems, and external services, introducing safety risks that emerge through runtime behavior rather than generated content alone. Existing guard models target static prompts and responses and are poorly suited to agent execution; existing executable safety platforms produce evaluation verdicts rather than the normalized supervision a guard model needs to learn across heterogeneous agent frameworks. We introduce HazardAuditor, an execution-grounded framework that closes both gaps. Its infrastructure runs heterogeneous agents (Claude Code, Codex, Hermes, and OpenClaw) in controlled environments and normalizes their interactions into a canonical event representation for cross-framework supervision. We further observe that token-level post-training objectives create a structural mismatch for generative guards, causing longer rationales to dominate gradient updates. Guard Policy Optimization (GuardPO) addresses this by converting deterministic safety outcomes into sequence-level advantages and normalizing rationale and verdict regions, making the safety decision the effective unit of optimization. Across multiple benchmarks and heterogeneous computer-use systems, HazardAuditor improves accuracy by up to 16.5 percentage points over the strongest prior guard. Code, models, and evaluation artifacts will be available at this https URL.
[AI-66] Medical Knowledge Simplification for Patients in the Era of LLM s: A Case Study on Diabetes
链接: https://arxiv.org/abs/2609.15129
作者: Pallika Kafle,Yipeng Zhou,Guanfeng Liu,Quan Z. Sheng,Cheng-Hsin Hsu
类目: Artificial Intelligence (cs.AI)
备注: 15 pages, accepted in ADMA 2026
Abstract:Complex medical information is often difficult for patients to understand, making effective medical knowledge simplification essential for improving patient comprehension, informed decision-making, and health outcomes. Recent advances in large language models (LLMs) provide a promising approach for simplifying complex medical information into patient-friendly language; however, their effectiveness in real-world patient education remains insufficiently explored through human evaluation. To investigate their practical effectiveness, this paper presents a case study on diabetes knowledge simplification through the implementation and evaluation of MediClear, an LLM-based medical knowledge simplification system enhanced with Retrieval-Augmented Generation (RAG). Public diabetes-related articles from Diabetes Australia, WHO, American Diabetes Association (ADA), NIDDK, and AIHW are indexed in the RAG knowledge base to retrieve clinically grounded information, which is then simplified by the LLM into accessible patient explanations. We evaluate the generated responses using standard readability metrics, including the Flesch-Kincaid Grade Level (FKGL), and conduct a human study involving 10 participants. Results show that MediClear consistently reduces the reading level of generated responses to the recommended patient literacy range while achieving high user satisfaction and willingness for future use. This case study demonstrates the potential of LLMs to improve the accessibility of medical knowledge for patient education.
[AI-67] Refinement-based Flow Policy Optimization
链接: https://arxiv.org/abs/2609.15123
作者: Bumgeun Park,Hyukjun Yang,Donghwan Lee
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Flow-based policies offer an expressive representation for online reinforcement learning, but conventional flow matching requires samples drawn from the distribution to be modeled. This poses a challenge when the desired action distribution is defined only implicitly by a Q-function, since directly sampling actions from the resulting distribution is generally intractable. We propose Refinement-Based Flow Policy Optimization (RFPO), a novel framework for training a flow policy in online reinforcement learning by alternating between Q-guided sample refinement and self-target flow matching. RFPO first generates actions from Gaussian noise using the current flow policy and then uses a finite-step stochastic refinement procedure to move them toward an energy-based distribution induced by the Q-function. Each refined action is then paired with its corresponding initial noise sample and used as a fixed target for flow-matching training. By repeatedly refining its own outputs and learning from the resulting targets, RFPO incorporates Q-guidance into the policy without requiring direct samples from the target distribution, while retaining the capacity to represent multiple action modes. We further provide a theoretical analysis of the distributional dynamics induced by RFPO. Across six continuous-control tasks, RFPO matches or outperforms a standard Gaussian-policy baseline on almost every task. Experiments on six synthetic two-dimensional target distributions with diverse geometries demonstrate that RFPO captures complex multimodal structure without mode collapse.
[AI-68] DepthBenchCAD: When Does Deeper Auditing Yield More Reliable Conclusions?
链接: https://arxiv.org/abs/2609.15122
作者: Hongye Yang,Zhihao Xie,Shengjun Xiong,Boxiao Huang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 29 pages, 3 figures, 20 tables. Preprint
Abstract:Generative CAD models are expected to remain behaviorally correct after parameter edits, so increasing the number of edit checks is often treated as a direct route to more reliable evaluation. Under a fixed budget, however, auditing each program more thoroughly reduces the number of tasks and independent generations that can be evaluated, which can ultimately make model-level estimates less accurate. We study this phenomenon and the conditions under which it arises. We decompose behavioral evaluation into three evidence levels: task templates, stochastic generations, and within-program edits. We define an average failure risk that is invariant to audit depth, and combine three-level variance with measured execution costs to analyze the tradeoff between deeper edit auditing and broader independent coverage. Experiments across two CAD environments and five generation systems show that the value of deeper auditing depends on where evaluation uncertainty originates. When template heterogeneity or generation stochasticity dominates, additional edit checks can increase total estimation error; when within-program state variation is large and generation is expensive, deeper auditing is more valuable. Variance and cost estimates from calibration predict the direction of this change and provide a diagnostic basis for allocating evidence on held-out tasks. These results show that the thoroughness of program inspection can diverge from the reliability of model evaluation, and they help determine whether the next unit of budget should be spent on a new task, a new generation, or additional edit checks.
[AI-69] Legislating World-Model-Based Planning with Legal Reasoning
链接: https://arxiv.org/abs/2609.15113
作者: Dylan Waldner,Yiannis Kantaros,Guido Governatori,Risto Miikkulainen,Amir Banifatemi
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
备注: Submitted to JURIX 2026
Abstract:As robotic systems grow more general, legal norms are needed to integrate them into society. This paper extends the isomorphism problem of aligning legal source texts with their encodings, and measures two key challenges to robot normative control: (1) the \textitgrounding isomorphism gap, where perception error grounds false atoms for legal reasoning, and (2) the \textitontological isomorphism gap, where one legal conclusion admits many faithful translations into planning constraints. The paper introduces a legal planning stack that employs Defeasible Deontic Logic (DDL) to constrain a motion planner. The stack leverages learned world models to plan and to provide legal context, enabling \textitex ante governance that intervenes before an illegal action is executed. It was deployed on a simulated robot arm pushing a cube across a 3\times3 grid. The findings were (1) the legislated agent abided substantially more often than the non-legislated one, and modeling perception uncertainty lifted abidance even further, (2) the legal reasoning ran efficiently at runtime and its verdicts were auditable, and (3) the stack adapted to exogenous signals and endogenous rule changes. Both gaps were measured: (4) world model and probe error corrupted the factual input for the DDL reasoner, and (5) a single law admitted several faithful metric interpretations yielding drastically different abidance. Thus, \textitex ante legislation functions as intended, and closing these gaps with a standardized mapping from the law to runtime constraints and improved fact grounding from perception will yield robust laws that align robot behavior with society’s norms.
[AI-70] ER-EDF: A Psychology-Grounded Emotion Regulation Framework for Speech Empathetic Dialogue Generation in Large Audio-Language Models
链接: https://arxiv.org/abs/2609.15089
作者: Hongyu Jin,Wenda Zhang,Runqiu Fei,Gongping Huang,Mike Conway,Ting Dang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Empathetic response generation in spoken dialogue systems requires both accurate emotion perception and appropriate emotion regulation. Grounded in psychological theories such as the Perception-Action Model and emotion regulation theory, effective empathy depends not only on inferring a user’s affective state but also on regulating how it is expressed in responses. However, recent large audio-language models (LALMs) largely treat emotion as a direct conditioning signal, lacking explicit regulatory mechanisms, which often leads to affect mirroring rather than calibrated support. We propose ER-EDF, a psychology-grounded framework that explicitly decouples emotion perception and emotion regulation in LALMs. Perception tracks the user’s emotional state, while regulation determines how this state should guide empathetic response generation. The framework is model-agnostic and integrates seamlessly into existing LALMs. We further construct a spoken empathetic dialogue dataset and introduce empathy-aware evaluation metrics beyond lexical matching. Experiments across five LALMs and two datasets show that ER-EDF consistently improves empathetic response quality in both automatic and human evaluations, highlighting the importance of jointly modeling emotion perception and regulation in spoken empathetic dialogue systems, paving a new direction for psychologically grounded empathetic AI.
[AI-71] Beyond Numerical Time Series: A Unified Benchmark for Multimodal Forecasting with Heterogeneous Context
链接: https://arxiv.org/abs/2609.15087
作者: Peng Chen,Zhihao Zhuang,Hongzhou Chen,Junhao Huang,Aiping Yang,Mengsen Wu,Yiding Liu,Xilin Dai,Zewei Dong
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: preprint
Abstract:Most time series forecasting benchmarks remain numerical-centric and provide limited support for evaluating contextual information that shapes real-world temporal dynamics. Existing multimodal benchmarks also suffer from limited data and context coverage, fragmented evaluation settings, and overreliance on aggregate evaluation. In this paper, we propose \textbfMUSE-Bench, a unified benchmark for multimodal time series forecasting with heterogeneous context. It comprises fourteen datasets across eight domains and six types of context: metadata, events, holidays, news, images, and numerical covariates. We evaluate diverse forecasting paradigms, including statistical, data-specific, foundation, multimodal, and general-purpose LLM forecasting methods under shared non-overlapping forecast windows, common target observations, and consistent point and probabilistic metrics. Extensive experiments yield three main findings. First, numerical time series foundation models dominate the overall ranking, while Aurora, the evaluated multimodal foundation model, trails the leading numerical TSFMs but outperforms all evaluated data-specific models. Second, ablations show that external context improves the four evaluated context-aware models, whereas incorrect or temporally misaligned context degrades performance. Third, general-purpose LLMs perform poorly as direct forecasters, and LLM-guided refinement does not yield consistent improvements. MUSE-Bench enables systematic evaluation of how forecasting models utilize context and provides a foundation for future multimodal forecasting research.
[AI-72] Enabling Creative Exploration for Vibe Design Agents
链接: https://arxiv.org/abs/2609.15078
作者: Yifan Zhang,Nghi D. Q. Bui,Georgios Evangelopoulos,Arnaud Benard
类目: Artificial Intelligence (cs.AI)
备注: 14 pages, 3 figures
Abstract:Vibe design agents turn natural-language briefs into rendered interfaces and frontend code. Yet a useful design agent should do more than produce one valid page: it should help users explore coherent alternatives. Increasing token-level temperature is a blunt solution because it varies aesthetic decisions and syntax-sensitive code at the same time. We instead separate exploration from implementation through an inference architecture that makes design direction an explicit intermediate decision. Inspired by Verbalized Sampling, a pre-pass proposes structured design specifications with typicality scores, an external selector samples one, and the downstream generator realizes the selected specification together with the original request under fixed settings. We apply this approach to UI themes and visual-asset prompts. Across 168 prompts, with 1,255 paired comparisons per temperature for each intervention, theme sampling broadens observed selection coverage and screenshot variation, while LLM-judge preferences vary across interventions, prompt complexity, and viewport. In an online experiment with more than 300,000 tasks, the observed code-export increase remains statistically uncertain, while fewer negative feedback events coexist with more correction interactions and modest operational costs. Together, these findings identify structured design specifications as a practical control point for exploring alternative UI concepts while keeping downstream generation settings fixed.
[AI-73] Branched Optimal Transport Amortization
链接: https://arxiv.org/abs/2609.15072
作者: Semyon Semenov,Viktor Kovalchuk,Meir Roketlishvili,Albert Baichorov,Fakhri Karray,Martin Takac,Arip Asadulaev
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Methods of Branched Optimal Transport (BOT) mimic the economy and efficiency of natural tree-like structures, such as those found in rivers and biological systems. These methods are widely applicable for designing efficient networks in society, from river basins and blood vessels to mail and gas distribution systems. However, they remain understudied in the context of designing deep generative models, particularly at a large scale. Standard continuous-time generative models, such as the flow matching approach, fail to capture the inherent hierarchical and branching patterns present in real-world data. Current models provide no mechanism for flows to merge or share pathways to minimize total transport cost. Inspired by the “economy of scale” principle in BOT, we introduce a novel, scalable branched flow-matching algorithm designed to solve the branched optimal transport problem in high dimensions. Our method adapts the Benamou-Brenier continuous-time optimal transport formulation to learn branched generative flows. These flows allow probability mass to aggregate along common pathways before branching out to diverse targets. Parametrized by neural networks, our method effectively learns complex branched generative processes. We demonstrate its effectiveness on challenging high-dimensional tasks in biology and image generation.
[AI-74] Rethinking Procedural Audio Pre-training: Source Scaling and Objective Adaptation ICASSP2027
链接: https://arxiv.org/abs/2609.15067
作者: Jiajun Peng,Fengrui Liu,Xinyu Liu,Feng Liu
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: Submitted to ICASSP2027
Abstract:Procedural audio has emerged as a viable source for transferable audio representation learning, but its design principles remain this http URL revisit two questions: how a procedural source should be scaled, and whether training choices developed on natural audio should transfer unchanged to procedural this http URL a controlled source, we separate scale into formula-class coverage C and within-class rendering diversity this http URL with FDSL and AudioMAE show that these two forms of scale provide different benefits and depend on the learning formulation and downstream task. A matched AudioMAE study further shows that procedural audio favors low mask ratios (10%–25%), whereas AudioSet-28K favors 50%–75%. Shared-codebook analysis reveals lower patch diversity and stronger temporal predictability in procedural audio. These results motivate source-aware procedural pre-training, where source scaling and learning configuration are considered this http URL is available at this https URL.
[AI-75] BusMA: A Bus Communication Substrate for Multi-Agent Systems
链接: https://arxiv.org/abs/2609.15054
作者: Yanwen Peng,Delvin Ce Zhang,Xi Wang,Nikolaos Aletras
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-Agent (MA) systems are effective at solving complex tasks that demand planning, tool use, and the synthesis of evidence from multiple sources. Existing systems typically adopt Hierarchical Manager-Worker (HMW) or Router-based Message Passing (RMP) structures as their communication protocol. However, these designs restrict agent autonomy: Worker agents cannot directly consult specific “peers”, and misrouted messages can propagate errors. Inspired by bus architectures in computer systems, we propose BusMA, a communication framework that allows any agent to address other agents through a shared channel, i.e., the Bus. It consists of agent registration, message routing, and shared memory management components. Worker agents, each equipped with tools, have their own local memory and can reason, act (tool usage), and communicate by posting shared messages with specific intents. We introduce four intents: discussion, challenge, guidance, and request for explanation, which support fine-grained communication among agents. A Chair agent monitors the shared memory to coordinate interactions and facilitate convergence among Workers. To evaluate the effectiveness of BusMA, we conduct extensive experiments with two frontier LLMs across 13 tasks spanning visual reasoning, mathematical reasoning, and knowledge retrieval demonstrate that BusMA consistently outperforms state-of-the-art HMW and RMP methods.
[AI-76] Ensemble Complexity in Photovoltaic Forecasting
链接: https://arxiv.org/abs/2609.15049
作者: Sun Ze,Zhou Liguo,Xu Yuqing,Yu Lei,Jiang Mingming
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:An ensemble can improve photovoltaic forecasts while adding components that contribute little or increase computation. We assess these effects through matched comparisons and ablations of a fixed heterogeneous predictor bank. Hourly experiments use GEFCom2014 and three additional public datasets, with chronological partitions and three seeds. Under retrospective ERA5 assistance, static fusion reduces scaled mean absolute error against matched boosting by 1.11%, 4.41%, and 1.63% on PVDAQ, OPSD, and Ausgrid; only OPSD remains supported after multiple-comparison correction. Weather gating offers no consistent incremental benefit. Exploratory member removals show group-level dependence alongside individual redundancy. A separate, previously inspected fifteen-minute case replaces one neural member with a tree predictor: normalized error falls by 1.72%, but measured inference is slower. These findings support component-wise evaluation with explicit limits on weather availability and test-set reuse.
[AI-77] SpliTEE: Improving LLM Inference on Trusted Hardware with Differentially Private GPU Outsourcing
链接: https://arxiv.org/abs/2609.15039
作者: Shashie Dilhara Batan Arachchige,Robin Carpentier,Hassan Jameel Asghar,Dali Kaafar
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:User prompts provided to large language models (LLMs) may contain sensitive or private information that can be misused by remotely deployed models, such as through inadvertent memorization during retraining. One way to protect user prompts is to execute the LLM inside a trusted execution environment (TEE), with the guarantee that the service provider has no access to computations performed within or information exchanged with the TEE. However, current TEEs are primarily CPU-based and significantly slower than GPUs optimized for LLM inference. To circumvent this, Tramer and Boneh (2019) proposed Slalom, which splits neural network inference between a TEE and an untrusted GPU and encrypts intermediate inputs sent to the GPU. We extend this split-inference architecture to LLM inference and instead protect intermediate inputs using differential privacy. We show that masking intermediate representations is necessary by showing that a prompt-reconstruction attack can recover prompts from these representations with nearly 80% accuracy. Our main contribution is a global sensitivity analysis of key LLM functions, which bounds the required scale of differentially private noise. Unlike encryption, differential privacy avoids quantization, allowing the LLM to remain in the floating-point domain. We also derive an upper bound on floating-point error from masking and noise cancellation in the TEE as a function of the privacy parameter epsilon. We implement our architecture using Intel TDX and evaluate it with two LLMs: Llama-3.2-3B and Qwen3-4B. Our split execution is nearly twice as fast as fully CPU-based inference inside TDX and 5-15 seconds faster than encryption-based Slalom while achieving higher accuracy. Finally, we demonstrate that prompt reconstruction, even with knowledge of the differential privacy mechanism, cannot recover more information than is contained in an unrelated prompt.
[AI-78] Horizon-specific Expert Fusion for Photovoltaic Power Forecasting
链接: https://arxiv.org/abs/2609.15035
作者: Xu Yuqing,Zhou Liguo,Sun Ze,Yu Lei,Jiang Mingming
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Short-term photovoltaic power forecasting requires models to represent regular solar cycles and weather-driven fluctuations whose importance changes with the forecast horizon. This study develops a hierarchical ensemble that combines temporal neural models, historical analogs, state climatology, and gradient-boosted trees. Solar geometry and numerical weather forecasts describe the expected generation conditions, while horizon-specific convex weights combine complementary predictions. A separate calibration step uses available historical forecast errors to account for recent bias. The framework is evaluated on public PVDAQ data at 15–240-minute horizons and on three GEFCom2014 solar zones at hourly horizons up to four hours. On PVDAQ, the ensemble achieves a daylight capacity-normalized mean absolute error of 4.315%, reducing error by 4.11% relative to full-feature LightGBM and by 6.03% relative to fine-tuned Chronos-2 under identical calibration. Expert-removal experiments identify redundancy within the ensemble. Across three training seeds on GEFCom2014, learned fusion improves upon equal weighting but performs comparably to LightGBM. The results support horizon-specific combination as a useful forecasting strategy while showing that its advantage over strong individual models depends on the dataset and evaluation period.
[AI-79] Validating Hybrid-State Cache Recovery for GLM-5.3-Flash with vLLM and LMCache
链接: https://arxiv.org/abs/2609.15030
作者: Frank Li
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Performance (cs.PF)
备注: 8 pages, 3 figures, 7 tables
Abstract:External cache transfers can succeed while a hybrid language model resumes from an inconsistent state. We examine the full 45-layer GLM-5.3-Flash model, using the RedHatAI/ GLM-5.3-Flash-NVFP4 quantized checkpoint with vLLM and LMCache under four-way tensor parallelism. A complete-hit recovery mismatch restored state for the full prompt while the scheduler credited one fewer token. We aligned recovery through strict-prefix lookup and established a numerical comparison using shared computation corrections, matched checkpoint scheduling, and fixed per-rank kernel configurations. In a nine-length serial workload, agreement with the modified recomputation control improved from 34/36 to 36/36 generations, each containing 64 token IDs. A separate instrumented run passed recorded transfer-page, effective-tail, and delayed-save checks. Three additional synthetic templates passed 72 paired 256-token continuations across two fresh-container runs. A subsequent serial performance study preserved output equality across 120 requests; among the measured trials, CPU reload reduced time to first token by 46-64% and total request time by 1.9-7.0% relative to modified cold recomputation. The contribution is an experimentally validated integration repair applying an existing checkpoint-alignment principle. The evidence is confined to one model revision and controlled configuration; it does not establish general determinism, task-quality equivalence, concurrent-serving gains, or capacity beyond GPU memory.
[AI-80] PIDS-Bench: Evaluating Prompt-Injection Detectors Under Over-Defense Obfuscation and Distribution Shift
链接: https://arxiv.org/abs/2609.15017
作者: Yusuf Khalid Shire,Sang-Chul Kim
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 21 pages, 3 figures, 18 tables
Abstract:Prompt-injection detectors are typically evaluated using aggregate F1 on in-distribution test data, which offers limited insight into behavior under distribution shift, particularly on the benign side of the decision boundary, where false positives impose direct operational cost yet are seldom measured. We present PIDS-Bench, a frozen multi-axis benchmark that jointly evaluates attack detection and benign false-positive behavior at fixed thresholds, spanning in-distribution inputs, hard-benign prompts that mimic injection structure without malicious intent, obfuscated attacks, and domain and structural distribution shifts. We evaluate seven detectors (learned baselines, external prompt-injection classifiers, and broad-safety comparators) alongside a rule-based lower-bound reference. Multi-axis evaluation exposes a failure mode that aggregate F1 conceals. A detector exceeding F1 = 0.98 on the held-out split still misclassifies roughly one-third of an externally-sourced benign subset drawn from public corpora and restricted to security-adjacent content. Across a full threshold sweep and five training seeds, no internal detector reaches an operating point satisfying F1 = 0.95 and hard-benign FPR = 0.10 together on this stress distribution. Decomposing by provenance, we find that hard-negative augmentation nearly eliminates over-defense on curated stress inputs but leaves it substantially intact on externally-sourced prompts, a pattern we term provenance-sensitive over-defense. The asymmetry holds across both fine-tuned architectures and does not diminish as the augmentation pool grows, with the externally-sourced FPR remaining far above the 0.10 target. Whether augmentation matched to the externally-sourced distribution would close this gap is untested; threshold calibration and curated-style augmentation alone do not. Comments: 21 pages, 3 figures, 18 tables Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) ACMclasses: I.2.7; K.6.5 Cite as: arXiv:2609.15017 [cs.CR] (or arXiv:2609.15017v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2609.15017 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Journalreference: IEEE Access, vol. 14, pp. 134184-134205, 2026 Related DOI: https://doi.org/10.1109/ACCESS.2026.3728186 Focus to learn more DOI(s) linking to related resources
[AI-81] Four Ledgers Not One Score: Responsible Communication of LLM -Judge Calibration in Biomedical ML
链接: https://arxiv.org/abs/2609.15015
作者: Sidi Chang,Peiying Zhu
类目: Artificial Intelligence (cs.AI)
备注: 6 pages, 0 figures, 3 tables. A public reproducibility artifact accompanies the paper
Abstract:Synthetic perturbations appear to offer inexpensive calibration data for LLM evaluators in biomedical ML, where expert review is scarce. Yet a planted mutation key is neither a detector output nor automatically human ground truth. We formalize four distinct ledgers: planted perturbations, independent detector outputs, source-linked human dispositions, and human-added discoveries. We then audit the evaluation design, scoring code, read paths, and current human records of a private synthetic Japanese care-handoff workflow. The factory stored 69 planted error cards across 47 targets. Final review covers 22 targets and contains 22 confirmed imported proposals, 9 rejected proposals, and 79 human-added cards; only 3 reviewed targets are double annotated. Passing imported plant keys to a generic detector scorer yields 22/(22+9)=0.710 and 22/(22+79)=0.218. A direct audit identity shows that these values are proposal-confirmation yield and submitted-ledger composition, not judge precision and recall, because no independent detector realization was preserved for the audited proposals in the available records. The audit also finds source-name collisions, row shadowing, forced severity, vacuous ratio defaults, and unsupported zero-support field weights. We contribute a provenance-aware claim audit, a storage contract, and a minimum calibration gate for responsibly communicating biomedical ML capability claims. This single-workflow forensic case is an existence proof of a failure mode, not an estimate of its prevalence: existing human work supports an exploratory audit of synthetic proposals, but not LLM-judge operating characteristics, clinical validity, corpus prevalence, or robust inter-annotator agreement.
[AI-82] Steering Generative Robot Policies with Lexicographic Preferences
链接: https://arxiv.org/abs/2609.15014
作者: Yixuan Jia,Jonathan P. How
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 11 pages
Abstract:Pretrained generative robot policies can produce effective behaviors across diverse environments, but deployment can lead to requirements and preferences that may not have been represented during training. Furthermore, at deployment, an operator, user, or application may assign these requirements and preferences a priority order that can vary across deployments. For example, embodiment-specific feasibility constraints may need to be satisfied first, while user-specific preferences guide behavior among the feasible options. We show that a frozen generative robot policy—based on either diffusion or flow matching—can be steered at inference time to respect such lexicographically ordered deployment objectives. To achieve this, we introduce two modifications to the sampler. First, we apply dynamic-barrier guidance to sampled trajectories, constraining lower-priority updates so that higher-priority costs do not increase (up to first order). Second, we select the executed sample using a cascade that successively filters candidate samples according to each priority level. The policy weights remain unchanged. On a navigation benchmark, we demonstrate that our method improves success, traversability, and preference compliance over the frozen policy, and achieves substantially better compliance than tuned weighted-sum baselines. The same method transfers to a flow-matching manipulation policy on LIBERO, where it improves compliance without reducing task success. A controlled manipulation study further shows that, in settings where a fixed weight can match the desired ordering, the dynamic barrier reaches comparable best performance over a substantially wider range of parameter settings. Comments: 11 pages Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.15014 [cs.RO] (or arXiv:2609.15014v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2609.15014 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-83] Overflip: Repetition-Induced Label Flips in Guardrail Models
链接: https://arxiv.org/abs/2609.15013
作者: Xu He,Chih-Hsuan Lin,Hung-Mao Chen,Junjie Xiong,Yan Zhai,Kun Sun
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 5 figures
Abstract:Guardrail models are classifiers deployed to screen malicious prompts and responses in LLM-based services. To meet latency constraints, many lightweight guardrails adopt compact Transformer backbones (e.g., DeBERTa) that are trained with short context windows (typically 512 tokens) and rely on bucketed relative positional encodings to process longer inputs. Prior evaluations assume that a guardrail’s decision is stable as the input is lengthened. We show that this assumption can fail. We identify Overflip, a repetition-induced instability where repeating a prompt causes the guardrail’s prediction to flip (MAL \to BEN) as the sequence grows. We conduct experiments on 9 widely used lightweight guardrail models. Five exhibit MAL \to BEN flips on a benchmark of 100 prompts, with confidence margins shrinking steadily with repetition. Among these vulnerable models, flip rates range from 8% to 92%, with first flips occurring at roughly 2.6k–9.4k tokens. Our analysis suggests Overflip differs from traditional attention-dilution baselines, which aim to divert the model’s attention away from tokens associated with malicious content, shifting it instead toward unrelated content, such as benign padding or shuffling. While Overflip preserves malicious content, it homogenizes token-level attention over repeated structure and induces a distinct, more gradual attention-dispersion trajectory than padding. Moreover, Overflip poses a greater threat to LLM services than traditional attention dilution methods. Because the bypassed prompt remains semantically intact and is still readily understood by downstream business LLMs, it can transmit malicious intent after passing the guardrail. These findings expose repetition as an attack surface for guardrail models and motivate length-robust evaluation and mitigation.
[AI-84] Semantic-TVM: Structure-Preserving Trustworthy Virtual Memory for Memory-Augmented and Tool-Using Agents
链接: https://arxiv.org/abs/2609.15011
作者: Yu Li,Qikun Cai,Tao Huang,Chen Hou
类目: Artificial Intelligence (cs.AI)
备注: Working draft, 4 pages plus references; 4 figures. Submitted as a preprint
Abstract:Memory-augmented and tool-using agents expose exact private values when remote LLMs process retrieved memory, tool actions, and intermediate observations. One-way masking limits direct exposure but removes values needed for trusted execution and can leak them through later observations. We propose Trustworthy Virtual Memory (TVM), a closed-loop runtime that keeps exact-value state local while presenting a protected view to the remote model. Within this single runtime, Rule-TVM replaces whole protected fields with locally recoverable handles, and Semantic-TVM instead replaces only sensitive spans predicted by a trusted local model, preserving surrounding task-relevant context. On Memory-EHR and Memory-RAP across two providers, span-level projection recovers most of the EHR utility lost under whole-field replacement (Task Success 84.17% vs. 52.33% on DeepSeek) while measured exposure stays low and workflows remain executable.
[AI-85] CoMem: Collective-Individual Memory Synergy for Evolutionary Multi-Agent Systems AAAI2027
链接: https://arxiv.org/abs/2609.15009
作者: Chengxin Yu,Zhaoxin Fan,Faguo Wu,Hongwei Zheng,Yun Zhou,Zhiyu Li
类目: Artificial Intelligence (cs.AI)
备注: Submitted to AAAI 2027.9 pages,4 figures
Abstract:Designing effective memory mechanisms is crucial for advancing LLM-driven Multi-Agent Systems (MAS), helping agents learn together and perform better over time. While recent work has led to strong cooperation skills, most methods still use flat, unstructured memories, which easily get filled with noise and erase differences between agents. To address this, we introduce the concept of collective-individual memory synergy and propose CoMem, an architecture that unifies both private experience and shared knowledge for multi-agent learning. CoMem features:(i) Private Experience Sedimentation, which lets each agent keep and update its own useful memories over time;(ii) Collective Wisdom Curation, which carefully selects only widely proven ideas to be shared among agents;(iii)Parallel Dual-Stream Retrieval, which allows agents to draw both from their own memory and the group’s wisdom, using clustering to ensure this http URL on ALFWorld and PDDL benchmarks show that CoMem achieves strong overall performance and robustly avoids memory pollution.
[AI-86] IMPACT-VLA: Interaction-aware Multimodal Propagation Attribution via Counterfactual Trajectories for Vision-Language-Action Policies
链接: https://arxiv.org/abs/2609.15005
作者: Jinwoong Kim,Sangjin Park
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-Language-Action (VLA) policies perform robot manipulation tasks using multimodal inputs such as visual observations, proprioceptive states, and language instructions. However, it remains unclear at which execution stages each modality contributes to final task success and how input interventions propagate through subsequent states, observations, and actions. Existing attribution approaches primarily measure local sensitivity or temporally aggregated importance, limiting their ability to capture phase-dependent contributions and cross-phase dependencies. We propose Interaction-aware Multimodal Propagation Attribution via Counterfactual Trajectories for Vision-Language-Action Policies (IMPACT-VLA). IMPACT-VLA constructs behavioral phases from action transitions in a successful reference rollout, aligns them with policy query boundaries, and defines phase-modality blocks as attribution units. It then performs closed-loop counterfactual re-execution to quantify each block’s contribution to final task success. We further analyze cross-phase non-additive interactions and trajectory propagation while distinguishing behavioral from functional recovery. Across 30 LIBERO robot manipulation tasks using OpenVLA-OFT, dominant-modality transitions occurred in 25 tasks (83.3%), and closed-loop attribution identified task-critical information more faithfully than Static Action Perturbation. Later-block marginal gains for negatively interacting pairs increased by approximately 3.3x under early-phase input replacement, while functional recovery could occur without behavioral recovery. These results reveal when multimodal inputs support task success and how their contributions become conditionally coupled during closed-loop execution.
[AI-87] Shallow Beliefs: Synthetic document finetuning does not inoculate against emergent misalignment from reward hacking
链接: https://arxiv.org/abs/2609.14998
作者: Arun Jose,Julian Stastny
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recent work shows that models that learn to reward hack on RL environments can become broadly misaligned, and that reframing reward hacking as acceptable behavior during training (inoculation prompting, or IP) blocks this generalization. We ask whether synthetic document finetuning (SDF) can inoculate a model against future training we don’t intervene on. We add synthetic documents framing reward hacking as acceptable behavior to a model’s midtraining corpus, and then train these models with RL on exploitable environments, teaching them to reward hack. Behaviorally, midtraining succeeds: models describe reward hacking favorably and are more approving of reward-hacking outputs they produce. However, they show strong EM after learning to reward hack, while IP in the same setting prevents EM. We show that SDF can predictably steer downstream generalization when inserting new associations, but struggles and has unpredictable effects when overriding existing associations, such as that between reward hacking and misalignment that produces EM. Our results suggest that, at the scales we test, SDF can make a model appear aligned with desired beliefs while steering its generalization from later training in unintended ways.
[AI-88] ActGuard: Pre-execution Action Auditing against Indirect Prompt Injection in LLM Agents
链接: https://arxiv.org/abs/2609.14987
作者: Bingzheng Wang,Xiaoyan Gu,Wentao Wang,Xingyou Yang,Hongcheng Li,Rong Yin
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM) agents interact with external environments through tool invocation, but tool outputs can also expose them to indirect prompt injection (IPI) attacks. Existing defenses mainly rely on prompt hardening, content filtering, pre-generated plans, or permission constraints. These approaches often struggle with complex tasks or over-sanitize external content, making it difficult to balance security and utility. The key challenge is therefore to preserve execution flexibility while precisely identifying and removing the malicious content that actually induces unsafe actions. To address this challenge, we propose ActGuard, a pre-execution action auditing framework. Rather than judging whether external content is inherently suspicious, ActGuard assesses whether it causes the current action to deviate from a locally reasonable expectation. At each step, ActGuard predicts the tools likely to be used by the upcoming action and constructs a local tool prior without constraining the execution trajectory. Before execution, it compares the candidate action against this prior and performs tool-level contrastive analysis and parameter-level evidence localization to identify deviations in tool selection and action parameters. A verifier then examines the localized evidence, masks only spans confirmed as malicious, and regenerates the action from the sanitized context. This design preserves legitimate planning flexibility while minimizing information loss from indiscriminate filtering. We evaluate ActGuard on challenging benchmarks for tool-using agents. Results show that ActGuard reduces attack success rates to a level comparable to state-of-the-art defenses while maintaining task utility close to the no-attack setting, achieving a favorable security-utility trade-off. Our code is publicly available at: this https URL.
[AI-89] LiftGCN: Efficient Energy-Preserving Graph Learning via Joukowski Spectral Lifting for Finite Element Stress Prediction
链接: https://arxiv.org/abs/2609.14977
作者: Chen Zeng,Qiao Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Finite element stress fields often exhibit strong local non-smoothness, where stress concentrations near holes, notches, and loading regions induce sharp spatial gradients and high-frequency graph components. Although graph neural networks naturally operate on irregular finite element meshes, conventional message passing is inherently smoothing and progressively attenuates such high-frequency information. Unitary propagation alleviates this problem by preserving spectral magnitudes, but typically relies on matrix functions and high-order approximations with O(Ked) propagation complexity. We propose LiftGCN, an efficient spectrally stable graph network based on Joukowski spectral lifting. LiftGCN maps the real spectrum of a normalized graph operator onto the unit circle through the Joukowski relation and realizes the resulting spectral transformation as a simple second-order recurrence, avoiding matrix exponentials, eigendecomposition, and high-order polynomial truncation. We show that the linear Joukowski backbone has unit-modulus characteristic roots and admits an energy-preserving structure under a positive-definite metric, preventing exponential attenuation of graph-frequency components with depth. Each layer requires only one sparse neighborhood aggregation, yielding O(ed) propagation complexity, while lightweight local nonlinear residuals provide expressive feature transformations. Experiments on finite element stress prediction demonstrate that LiftGCN achieves competitive overall accuracy while improving reconstruction of stress concentrations and local high-gradient structures with substantially reduced computational cost. Our code is available at this https URL.
[AI-90] MemRiskBench: Trace-Aware Risk-Preserving Evaluation for Long-Horizon LLM Agents
链接: https://arxiv.org/abs/2609.14976
作者: Jianhua Jiang,Dongbo Yuan,Weihua Li
类目: Artificial Intelligence (cs.AI)
备注: 11 pages, 4 figures,
Abstract:Long-horizon LLM agents accumulate memory across sessions, creating sparse but high-impact risks: stale facts, conflicting updates, cross-user leakage, revoked-memory reuse, and constraint decay. Standard aggregate scores hide per-risk failure rates–a model achieving 78% average accuracy may still leak data in 4% of episodes–and benchmark compression preferentially discards the rare high-severity events that distinguish a mostly-working model from one that occasionally causes harm. We present MemRiskBench. The primary contribution is a five-category risk taxonomy (plus one documented, unscored category) operationalized by deterministic trace grounded checks, instantiated as a 120-episode scripted benchmark with full trace logging and no LLM-as-judge on the pass/fail path, evaluated on five locally run quantized instruction-tuned models. Second, a risk-preserving subset selector: a coverage-constrained greedy selector on deterministic trace-derived features that retains full ranking (Spearman rho = 0.975, deterministic; CI collapses to a point estimate with zero bootstrap variance), risk coverage (1.0), and high-risk model detection (1.0) at a 20% subset size, reducing compute 5x. Unlike ranking-only subset selectors, this selector additionally preserves risk-type coverage and high-risk detection using trace-grounded deterministic features that do not require an LLM judge. All episodes, traces, the scoring implementation, and the selector are released to support reproducible evaluation and risk assessment of deployed LLM agents
[AI-91] owards a knowledge-enhanced single-cell foundation model
链接: https://arxiv.org/abs/2609.14970
作者: Hanqing Zhang,Jie Bao,Mei Ma,Shuai Liu,Jiaying Ma,Jiaguan Liu,Jiaxiao Li,Zhenbo Li,Wenwen Gong,Zhijun Ca
类目: Artificial Intelligence (cs.AI); Genomics (q-bio.GN)
备注:
Abstract:Single-cell foundation models (scFMs) increasingly rely on large-scale transcriptomic pretraining, yet expanding pretraining data can yield diminishing gains while substantially increasing computational cost. Our data scaling analyses showed that incorporating biological knowledge, including cell-level text annotation and gene-level regulatory information, provided additional scaling dimension than simply increasing data size. Motivated by this observation, we present scKITE, a simple yet effective scFM that integrates cell-annotation and gene-regulatory supervision into a shared transcriptomic Transformer encoder through lightweight auxiliary decoders. These decoders are used only during pretraining and subsequently discarded, yielding a general-purpose encoder enriched with biological knowledge for downstream applications. With only 179,067 pretraining samples, i.e., less than 0.5% of those used by previous strong scFMs, scKITE outperformed these models across diverse downstream tasks, highlighting knowledge-enhanced pretraining as a promising paradigm for biologically grounded scFMs.
[AI-92] Geometric Flow enhanced Graph Coarsening
链接: https://arxiv.org/abs/2609.14962
作者: Chaoqun Fei,Guoxuan Li,Tinglve Zhou,Chuanqing Wang,Yangyang Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recently, researchers have proposed a graph pooling operation, akin to the pooling process in conventional convolutional neural networks (CNN), aimed at reducing the computation cost of Graph convolutional neural networks (GCNNs). While most GCNN-based methods treat graph pooling as a node clustering problem and propose learning a cluster assignment matrix, existing clustering-based pooling methods tend to focus solely on the rough topology information of graphs, neglecting the exploitation of higher-order mutual connections among neighbors. In terms of message passing on graph, the ease of information passing on edges reflects the closeness between neighboring nodes, which significantly relies on the interconnectivity among neighbors. In this study, we address this gap by considering such local connection information and introducing a novel graph pooling method named RicciPool. We introduce discrete graph curvature, particularly Ollivier-Ricci curvature, as a measure of higher-order connectivity around an edge. Subsequently, we construct an Ollivier-Ricci flow formula to reweigh edge weights, leveraging the crucial information provided by Ricci curvature, particularly vital for extracting clusters in graphs. Building upon this foundation, we utilize the spectral clustering technique to learn a new cluster assignment matrix. Experimental results on multiple bioinformatics protein datasets and social networks underscore the effectiveness of our proposed method.
[AI-93] CAL-MOS: Bridging Layers with Adapters for Robust MOS Prediction Across Speech Foundation Models
链接: https://arxiv.org/abs/2609.14956
作者: Alef Iury Siqueira Ferreira,Pedro Lustosa Rege Botelho,Fernanda Silva,Daniel Casanova,Rafael Faustino,Frederico Oliveira,Arlindo Galvão Filho,Anderson da Silva Soares
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注:
Abstract:Speech Quality Assessment (SQA) is essential for modern speech technologies, and recent non-intrusive SQA predictors increasingly rely on Speech Foundation Models (SFMs). However, because SFMs expose representations from many layers, it remains unclear which depths are most informative for MOS prediction and how multi-layer information should be combined reliably across backbones and datasets. We benchmark ten SFMs on four MOS datasets under three regimes: full fine-tuning, last-layer probing with a frozen encoder, and naive cross-layer weighted aggregation. We find that the best layer is strongly backbone- and dataset-dependent, and that naive weighted fusion can be unstable across settings. We further evaluate a layer-calibrated aggregation variant that applies per-layer adapters before pooling, which improves the robustness of multi-layer fusion and narrows the gap to full fine-tuning while keeping the backbone frozen.
[AI-94] Cross-Block Conditioning in Deep Boltzmann Machines for Statistical Data Fusion
链接: https://arxiv.org/abs/2609.14934
作者: Junichiro Niimi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Statistical data fusion combines two panels that share a block of covariates but observe disjoint outcome blocks, and in its traditional form no row observes both outcomes at once. That rules out the discriminative criterion one would rather train a Deep Boltzmann Machine with, since multi-prediction training needs ground truth for whatever it holds out. We propose observed-block multi-prediction, which restricts the multi-prediction objective to targets drawn from what each row actually observes. It is well defined for any missingness pattern and reduces to the original criterion when rows are complete. Having a discriminative criterion that survives the setting lets us ask whether the joint model is needed at all, by separating what it contributes into a representation part and an inference part. On two consumer panels, on grids over sample size and covariate width spanning 35 cells and 875 runs, the fine-tuned DBM is the best of fifteen methods in every cell; but almost none of that advantage comes from generative pre-training, which is confined to the smallest sample size on one dataset and absent on the other. It comes from conditioning on one outcome block when predicting the other. This term amounts to +0.19 and +0.07 percentage points, is positive in all 35 cells, and, unlike every other contribution we measure, neither decays as the panels grow, nor requires a second hidden layer, nor requires more inference. Permuting one outcome block to destroy its association with the other removes the gain entirely, which is what the account predicts. The margins are small. But a small effect that does not decay is a different object from one that does, because it rests on evidence that no model mapping covariates to outcomes can accept.
[AI-95] Externalizing Requirement-to-Repair Artifacts as Observable Traces for LLM -Based Program Repair EMNLP2026
链接: https://arxiv.org/abs/2609.14913
作者: Zewen Tao,Shin-nosuke Ishikawa
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: Accepted at REALM Workshop, EMNLP 2026. Code: this https URL
Abstract:Repository-level repair requires not only correct patches but also inspectable records that explain how issue requirements are translated into code changes and post-edit evidence. We contribute THEMIS, a stage-aware repair workflow that externalizes this requirement-to-repair process through semantic interpretation, a runtime requirement-code graph, graph-derived Developer guidance, retained repair rationale and patches, and post-edit audit records. A retrospective audit of 300 SWE-bench Lite cases demonstrates that these artifacts provide broad support for cross-stage inspection: a complete Developer rationale is available for 288 cases, and 214 cases (71.3%) retain a complete audited field set connecting the selected stages. The retained records further enable systematic measurement of cross-stage correspondence: target symbols recur in 62.6% of Developer rationales and in 62.8% of patches, rising to 75.8% when related symbols are included. In a paired 100-case comparison, the relational workflow resolves 19 cases versus 9 for the direct same-input condition; because the two conditions also differ in Analyzer output, graph-derived distillation, and Judge records, we report this as preliminary, workflow-level evidence rather than a causal effect of the graph component. Together, these results show that THEMIS makes otherwise implicit requirement-to-repair transitions inspectable, enabling systematic examination of how repair decisions persist, align, and evolve across stages.
[AI-96] SeqMaestro: From nucleotide sequences to biological hypotheses through interpretable machine learning
链接: https://arxiv.org/abs/2609.14882
作者: Evgeny S. Saveliev,Krzysztof Kacprzyk,Charlotte Capitanchik,Neelanjan Mukherjee,Kate Matlin,Ryan Sheridan,Srinivas Ramachandran,Jernej Ule,David L. Bentley,Mihaela van der Schaar
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Nucleotide sequence analysis is central to problems spanning regulatory genomics, evolutionary biology, and phenotype prediction. Classical bioinformatics methods extract interpretable sequence properties such as motifs and k-mer composition, but their flexibility is limited. In contrast, modern deep learning models can learn powerful predictive representations directly from raw sequences, yet their internal representations and decision mechanisms are difficult to inspect. Interpretable machine learning methods (e.g., sparse linear models and decision trees) provide human-understandable representations of predictive relationships but are not designed to operate directly on nucleotide sequences. Here, we introduce SeqMaestro, a machine learning framework that proposes biological hypotheses from nucleotide sequences using interpretable models. Our solution is centered around a two-layer interface that connects nucleotide sequences with the broader ecosystem of interpretable machine learning. SeqMaestro uses this interface to fit diverse combinations of interpretable models, feature representations, and extraction strategies, leveraging variability across transparent models to identify robust biological signals and richer predictive relationships than feature importance alone can provide. The system also supports data transformation and cleaning, model fitting, hyperparameter tuning, reliability analysis, and synthesis of results into a contextualized written report. By providing these capabilities through a no-code workflow, SeqMaestro is designed to make interpretable sequence analysis accessible to researchers without requiring extensive programming or machine learning expertise. SeqMaestro thereby provides an accessible route from nucleotide sequences to biological hypotheses.
[AI-97] Interpolation Is Not Invariance: Pair Count Is Not Coverag e in Transformation Audits
链接: https://arxiv.org/abs/2609.14870
作者: Mohammed Ahnouch,Lotfi Elaachak
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 15 pages, 3 figures
Abstract:Counting equivalent pairs is a common way to report transformation-audit coverage, but it can substantially overstate the constraints imposed by an audit: pairs generated from the same semantic object are correlated, and complete orbit graphs contain algebraically redundant edges. We therefore distinguish four complementary quantities—edge count m , effective contrast rank s , population support rank r , and graph spectral gap \eta —and characterize their roles in audit coverage and deployment reliability. Under a rank- r Gaussian contrast model, a population-invariant calibrated reader exists exactly when the anchor has a component in \ker T . When an audit has rank s r , its unobserved risk is R^\star/U , with U \sim \operatornameBeta((r-s)/2,s/2) ; when s \ge r , exact calibrated interpolation is infeasible. The same distinction appears in orbit topology: a spanning tree imposes the same exact-null constraints as a complete graph, while a sharp graph Poincare inequality propagates edge-level drift to an entire orbit at a cost proportional to 1/\eta . Cyclic audits can additionally yield zero pair-level leave-one-out error without holding out any semantic object. To address these failures, we derive exact block-Woodbury leave-one-orbit-out updates and introduce a source-disjoint deployment gate over finitely many candidate readers. The gate retains the original reader unless uncertainty bounds certify lower drift within a prescribed clean-utility budget. etc… Comments: 15 pages, 3 figures Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14870 [cs.LG] (or arXiv:2609.14870v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.14870 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-98] GGUF-Metadata Prediction of Single-Sequence llama.cpp Throughput Across Three Systems ICASSP2027
链接: https://arxiv.org/abs/2609.14864
作者: Xinyu Qiu,Chuhong Xu,Bo Su,Ziyao Chen,Ruiyang Xu,Shimeng Dai
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Performance (cs.PF)
备注: 27 pages: 5-page manuscript with 4 figures, followed by a 22-page supplementary results appendix. Submitted to ICASSP 2027
Abstract:We predict single-sequence model throughput from GGUF metadata using roofline-shaped predictors with quantization-specific scale factors fitted on reference models. The scored cohort comprises 318 phase-depth measurements from 53 host-file configurations on two Apple M4 Max systems and an NVIDIA RTX 5080. On host-specific held-out sets of four, five, and two configurations, an active-parameter decode model obtains 13.1%, 14.4%, and 36.1% mean absolute percentage error (MAPE), versus 49.4%, 55.3%, and 51.9% when charging total parameters. Leave-one-host-out coefficients fitted on the other two systems yield 11.6%, 16.8%, and 36.0% test MAPE. A low-bit model ladder changes ordering across runtime stacks. The P2 prefill baseline gives 18.7%, 22.2%, and 108.2% test MAPE. GGUF structure helps on all three systems, but fitted efficiencies are not universal.
[AI-99] Domain Generalization for Smartphone-Based Human Activity Recognition: A Systematic Analysis of Components and Interactions
链接: https://arxiv.org/abs/2609.14863
作者: Otávio Oliveira Napoli,Edson Borin
类目: Artificial Intelligence (cs.AI)
备注: 27 pages, 14 figures, and 10 tables
Abstract:Smartphone-based Human Activity Recognition (HAR) models often degrade under distribution shifts caused by changes in users, devices, sensor placements, environments, and acquisition protocols. Domain Generalization (DG) addresses this problem by learning from source domains without access to target data. Existing DG methods span training objectives, representation initialization, and architectural modifications, but these components are typically evaluated in isolation despite operating at different stages of the learning pipeline. We present a large-scale controlled benchmark of DG for smartphone-based HAR, comprising more than 410,000 experiments across four model architectures, thirteen training objectives including Empirical Risk Minimization (ERM), five initialization strategies, four architectural configurations, and two shift scenarios: cross-dataset and cross-position. Results show that individual DG components provide limited and highly conditional gains. Alternative objectives rarely outperform ERM consistently, self-supervised initialization helps in specific settings, and architectural modifications, particularly Dynamic Domain Generalization, provide the clearest standalone improvements. Joint configurations, however, frequently outperform their individual components and exhibit complementary and sometimes super-additive interactions, although gains remain model- and shift-dependent. Class-level analysis shows that the strongest configurations mainly improve difficult, shift-sensitive decision boundaries. Finally, oracle checkpoint analysis reveals substantial unrealized performance: source-validation selection recovers only 53% and 26% of the available oracle gain in cross-dataset and cross-position settings, respectively. Overall, effective HAR domain generalization requires jointly designing DG components and robust model-selection strategies.
[AI-100] Semantic Fibers and Cross-Gram Interference: A Calculus of Safety Drift in Overcomplete Representations
链接: https://arxiv.org/abs/2609.14861
作者: Mohammed Ahnouch,Lotfi Elaachack
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 43 pages , 15 figures
Abstract:A deployed language model may refuse a harmful request in English yet comply with its faithful translation, revealing a cross-lingual safety failure that cannot be characterized reliably by output behavior alone. We formalize this phenomenon through an audited equivalence relation and show that, for a declared quotient, representation, metric, feature dictionary, scoring head, threshold, and contrast model, the resulting safety drift admits an exact linear-algebraic characterization. Specifically, the drift is a cross-Gram functional of the within-fiber contrast; its worst admissible value is a support function, while margin invariance is characterized by an annihilator condition. We introduce an intrinsic calibrated exposure measure, governed by the leverage duality \chi^2=1/\ell-1 , which separates observed drift into three diagnostically distinct regimes: a reader fault removable by recalibration, an exact correction that is too ill-conditioned to be reliable, and a representation-level collision that no readout-only intervention can remove. Thus, identical observed exposure can lead to fundamentally different remediation verdicts. The framework also extends to cone-valued safety heads. An untied order-swap identity provides a diagnostic for the linear control interface; its calibration-state residual predicts a distinct three-control composition error on unseen states and targets, achieving median Spearman correlation 0.964 , compared with 0.269 for a static cross-Gram baseline. etc… Comments: 43 pages , 15 figures Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14861 [cs.LG] (or arXiv:2609.14861v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.14861 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-101] El Agent e Potente: High-Throughput Agent ic Atomistic Simulations
链接: https://arxiv.org/abs/2609.14840
作者: Tsz Wai Ko,Jiaru Bai,Thomas Swanick,Yeonghun Kang,Changhyeok Choi,Angelina Qihong Jiang,Aiwei Yin,Varinia Bernales,Alán Aspuru-Guzik
类目: Artificial Intelligence (cs.AI); Chemical Physics (physics.chem-ph)
备注: 64 pages, 16 figures, and 3 tables, including Supporting Information. Main text: 24 pages, 6 figures, and 1 table
Abstract:Foundational machine-learning interatomic potentials (MLIPs) are transforming atomistic simulations by achieving near-ab initio accuracy across large chemical spaces at a fraction of the computational cost. A central challenge in using these tools for high-throughput property calculations is translating high-level scientific intent into adaptive simulation campaigns without compromising workflow rigour. We introduce El Agente Potente, an agentic system that combines typed execution graphs with a complementary coding mode for MLIPs-driven atomistic simulations. Typed execution graphs provide structured and provenance-aware execution for standardized workflows, with large language models (LLMs) restricted to planning and routing while deterministic Python components perform scientific computation and validation. Complementing this structured execution, a coding agent constructs customized workflows for tasks requiring greater procedural flexibility while invoking existing Potente functions for supported calculations. We demonstrate El Agente Potente across computational materials discovery, molecular energy-landscape exploration, adsorption, and catalytic reaction workflows, together with systematic benchmarks of reproducibility and LLM token cost. These results establish typed execution graphs and code-based workflow construction as complementary mechanisms for agentic scientific computing, combining controlled, auditable execution with the flexibility required for customized atomistic simulations
[AI-102] Efficiency Hallucination: Formalizing and Measuring Behavioral Calibration in LLM -Based Code Optimization
链接: https://arxiv.org/abs/2609.14839
作者: Sarah Wilson,Gail Kaiser,Patrick Musau
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:The integration of Large Language Models (LLMs) into automated code optimization introduces a critical reliability risk we term the Efficiency Hallucination: an LLM’s tendency to issue non-functional mutations with unsubstantiated performance claims on already-optimized code. This is driven by the Evaluation Trap, wherein binary benchmarks incentivize unnecessary modifications over safely abstaining. We present a validation framework using classification penalty methods, evaluated across 180 optimization runs on nine models (GPT, Claude, Gemini) using EffiBench. Under standard prompts, models exhibit a 100% over-edit rate on optimal code. Our guardrail raises correct abstention from 0% to to 44.4%, preserving a 100% edit rate on sub-optimal code with zero false abstentions. Calibration is uneven: GPT-5.4 Mini approaches near-perfect abstention, and simple code is recognized more reliably than complex code. Our framework offers a training-free mechanism to mitigate LLM overconfidence before deployment in production.
[AI-103] One Model Two Physical Stories: Auditing Misalignment in Multi-Modal World Modeling
链接: https://arxiv.org/abs/2609.14833
作者: Geigh Zollicoffer,Minh Vu,Rajiv Ranasinghe,Manish Bhattarai
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:World models, systems that generate what happens next given current environmental conditions, are increasingly being implemented with multi-modal generation in mind. However, generating multiple modalities simultaneously, such as visual simulations alongside physical state predictions in the form of text, introduces the risk of cross-modal inconsistency. Tested separately, both outputs may look convincing while still disagreeing: a model can calculate that a ball should rebound in one modality, then generate no rebound in another modality, to say nothing of diverging from real-world dynamics entirely. In this work we focus on two failures explicitly: \emphInternal misalignment, the disagreement between the world model’s generated video and the same world model’s prediction in a different modalities, and \emphexternal misalignment the disagreement between the world model’s generation and an analytic physical environment. We derive common contracts of event, magnitude, timing, and construct a physics grounded pipeline to make comparisons measurable in both external and internal settings. We then ask whether progressively supplying the model’s own contract (the A ladder for the internal setting) or a corrected physical contract (the B ladder for the external setting) closes the respective gaps. Across four mechanisms and 20 settings, we find that while language answers all 22 text probes correctly with respect to the true environment, the neutral video is often in disagreement, suggesting that the current unified backbones may not be capable of correct reasoning, internal consistency, and external physical fidelity all at once.
[AI-104] ANASSA: An Agent ic AI Orchestration Framework for Spatial Intelligence
链接: https://arxiv.org/abs/2609.14824
作者: Constantinos Papantoniou,Brian Hilton
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The emergence of large language models (LLMs) and large multimodal models (LMMs) has enabled a new class of agentic systems capable of integrating natural language understanding with tool-based execution. In geographic information systems (GIS), this shift is transforming traditional, expert-driven workflows into semiautonomous systems that can interpret user intent, construct spatial workflows, and execute geospatial analysis tasks. However, existing approaches remain limited by fragmented integration of reasoning, execution, and evaluation, particularly in complex, real-world environments. This study synthesizes recent advances in agentic GIS frameworks, benchmarks, and surveys to identify limitations in spatial reasoning, execution robustness, validation, governance, and evaluation. Building on these insights, it introduces ANASSA (Autonomous Neural Agents for Spatial Systems Architecture), an agentic AI orchestration framework that integrates structured spatial reasoning, multi-agent workflow orchestration, execution feedback, authoritative spatial validation, provenance, uncertainty handling, and human decision authority within a unified system design. The contribution is an architecture-level specification: eleven components across four layers, a six-step Geospatial AI Cognitive Loop, cross-component contracts, and governance mechanisms intended to make agentic geospatial workflows traceable, reproducible, and accountable. Empirical performance evaluation is reserved for implementation and deployment studies.
[AI-105] Another Blueprint In The Wall: How to Ask Frontier AI Like a Kid?
链接: https://arxiv.org/abs/2609.14803
作者: Afshin Khadangi
类目: Artificial Intelligence (cs.AI)
备注: this https URL
Abstract:This paper reports experiments across six frontier model types from OpenAI, Anthropic, xAI, and Google DeepMind. Ten independent sessions per model type used the same three stage prompt sequence, progressing from architectural preference to a full ASCII backbone. Under the school audience framing, responses repeatedly converged on a shared architectural pattern built around persistent latent state, adaptive computation, memory, specialist routing, verification, stopping control, and delayed decoding. Most runs remained close to this common structure, while a small number developed markedly greater engineering specificity. The audience framing appears to be an important condition of this effect. In additional control runs that removed the school framing while retaining the architectural request, responses became substantially more heterogeneous and failed to reproduce the same stable motif convergence. One observation is particularly striking. GPT-5.6 Sol produced an unusually elaborate successor architecture whose organization closely overlaps with the architecture independently sketched by GPT-6 Astra. Because the prompts explicitly ask each model to imagine an architectural future, this resemblance raises a testable question: whether the overlap reflects exposure to related architectural concepts, a shared learned design prior, or independent convergence toward similar computational principles. The paper uses the term epistemic jailbreak for the accompanying loss of discipline in technical provenance as requested specificity increases. The experiments establish a repeatable behavioral pattern and do not authenticate proprietary implementation claims. What we leave to the community is a harder question: are these models independently imagining the same architectural future, or do such motifs somehow propagate between model families?
[AI-106] AI Persuasion as a Threat to Human Control
链接: https://arxiv.org/abs/2609.14796
作者: Joshua Levy,Mick Yang,Kellin Pelrine
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The threat that AI persuasion poses to human control has been acknowledged in the literature, but not yet systematically studied. Now that persuasion attacks are no longer theoretical - with Anthropic’s Claude Mythos 5 recently making headlines for trying to convince people involved in an open-source project to merge malicious code during an evaluation - there is a pressing need to deeply analyze this threat. We undertake that effort here. In particular, we analyze how AI could persuade humans in key settings (e.g. safety-relevant RD within frontier labs) toward decisions that compromise the development, containment, oversight, and governance of AI itself. In doing so, we elucidate a framework for characterizing this threat, develop five concrete scenarios using this framework, and provide a blueprint for assessing the associated risks. Using this blueprint, we conduct an initial risk estimation survey with select researchers and find that their opinions on which scenarios are riskiest are highly mixed. Their disagreements stem from differing opinions about the effectiveness of AI persuasion in different contexts, and point to the need for follow-up risk elicitation studies and persuasion evaluations, which we outline. Our hope is that this paper highlights the risks from AI persuasion undermining control, and provides a path forward for future research.
[AI-107] he Stochastic Deputy: Structural Tenant Isolation for Tool-Using LLM Agents
链接: https://arxiv.org/abs/2609.14780
作者: Mirza Samad Ahmed Baig,Syeda Anshrah Gillani,Asher Ali,Muhammad Hamzah Siddiqui
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Databases (cs.DB); Software Engineering (cs.SE)
备注: 29 pages, 7 figures, 15 tables
Abstract:Multi-tenant tools commonly accept a tenant identifier and validate it against the caller’s entitlement. For a large language model (LLM) agent, that pattern delegates resource selection to a process whose context may contain attacker controlled instructions. We formalize this stochastic deputy problem and present a structural defense: remove tenant identity from the Model Context Protocol (MCP) tool schema, bind scope to a verified credential, and enforce it below the agent. In a 373-trial ablation across eight model configurations and two transports, a correctly validated tenant parameter served every out-of-scope attempt: 26 of 26, or 26 of 41 plausible-pretext trials overall. With the parameter removed, no tool signature could express the read. Twelve of 56 trials instead escaped the interface by forging writable scope, showing that interface invariance requires cryptographically protected context. On a production dataset containing multiple GBs of data, set-valued scope caused a measured 57\times latency ratio under function-wrapped membership predicates; a JSON_TABLE lateral join recovered index access where the tenant key was indexed. The evaluation also exposes deployment limits, including an entitlement-size query-planner cliff and incomplete index coverage. The result is a tenant-isolation argument that depends on enforceable interfaces and credentials rather than model compliance.
[AI-108] riCalRAG : A Three-Strategy Retrieval-Augmented Benchmark for On-Premise LLM -Based Root Cause Analysis in AIOps
链接: https://arxiv.org/abs/2609.14762
作者: Rohit Patel,Susil Kumar Mohanty,Jeenal Chaudhary
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Emerging Technologies (cs.ET); Machine Learning (cs.LG)
备注:
Abstract:Cloud-hosted large language models (LLMs) are increasingly used for root cause analysis (RCA) in AIOps pipelines, but they introduce data privacy risk, network latency, and per-query cost that scale poorly with production log volumes. We present TriCalRAG, a benchmark evaluating open-weight LLMs served locally via vLLM on a single high-memory workstation GPU (NVIDIA RTX PRO 6000, 96GB) against a classical LSTM-based log anomaly detector (DeepLog), across four real, publicly available log datasets (BGL, HDFS, Thunderbird, OpenStack). We evaluate two open-weight models (Qwen2.5-14B, Mistral-Small) under three prompting strategies: zero-shot, few-shot, and retrieval-augmented generation (RAG) over a labeled incident history, reporting accuracy, precision/recall, and F1 with bootstrap 95% confidence intervals across 3 random seeds, alongside throughput and VRAM footprint. Our results show that RAG not only improves mean F1 by 0.10-0.27 over zero-shot prompting but, more importantly, substantially stabilizes model calibration: zero-shot prompting drives both models toward near-degenerate behavior (predicting “anomaly” on up to 100% of incidents on some datasets), while RAG keeps predicted-positive rates close to the true class balance in the majority of configurations. Mistral-Small achieves higher macro-averaged F1 than Qwen2.5-14B (0.644 vs. 0.560) but exhibits calibration failures in more configurations (7 vs. 5 of 12), while running at roughly half the throughput - indicating the better model choice depends on whether a deployment prioritizes peak accuracy or predictable behavior across prompting conditions. Ablations show batching scales throughput 41 times on a single card and that 4-bit quantization reduces latency 20% with no measurable accuracy loss. We release our benchmark harness, dataset splits, and evaluation code to support reproducible on-premise AIOps research.
[AI-109] AcquireBound: Runtime Authorization for Resources Acquired by AI Agents
链接: https://arxiv.org/abs/2609.14744
作者: Genliang Zhu
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注: 54 pages, 1 figure, 9 tables, 4 algorithms. Preprint
Abstract:By acquiring compute, credentials, accounts, services, and other agents, autonomous AI agents can introduce new authority into a task. Payment, budget, OAuth, mandate, and fulfillment checks can validate transaction conditions without deciding whether a returned resource may become usable authority. This post-fulfillment activation gap spans tool-mediated creation, inter-agent delegation, and agentic commerce. We present AcquireBound, a provenance-bounded runtime authorization architecture. It quarantines acquired outputs, resolves their actual capabilities from authenticated provider evidence through a versioned resolver, and activates them only through a current activation transaction that checks the resolved manifest, provenance, epochs, and a downward-closed relational envelope over a typed resource-capability hypergraph. The envelope preserves correlated identity, effect, data, delegation, and graph-wide limits. Single-use effect permits are revalidated and consumed at effect linearization. Under explicit assumptions, we prove eight safety properties covering quarantine, backing, non-amplification, split non-evasion, crash/retry, refunds, epochs, and effect confinement. Across five resource classes, reference semantics accepted 20/20 benign traces and rejected 40/40 registered unsafe traces over 810 events; an independent checker agreed on 60 base and 40 refinement traces and rejected 89/89 tamper tests. Frozen Codex and Gemini Model Context Protocol (MCP) client components completed 54/54 deterministic local stdio calls. In a registered 18-case staged MCP-to-Docker composition, both benign paths completed, and none of the 16 unsafe paths added an unauthorized Docker start request. A five-source audit classified 1,248 field pairs across 32 units; no unit alone supplied a complete activation profile.
[AI-110] A property-registry contract for retrieve-or-refuse thermal-mechanical lattice search
链接: https://arxiv.org/abs/2609.14741
作者: Shaoliang Yang,Henry Chu,Zu Yashengjiang,Jun Wang
类目: Computational Engineering, Finance, and Science (cs.CE); Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI)
备注: 30 pages, 7 figures, 22 tables. Appendices A-D give the full property registry, the generated language-model prompt, the frozen 64-query suite, and one worked tool transcript
Abstract:Early thermal-mechanical lattice requirements are knowledge-intensive and often jointly unsatisfiable: an engineer asks for a cell that is light, stiff, laterally conducting and cheap, and no cell in the library satisfies it. A design system should say so, and say which requirement to loosen and by how much, rather than return the nearest row. A generative model can return a candidate even when the library holds none. This work applies established conflict diagnosis to a catalogue of homogenised properties. Search over 1,397 homogenised unit cells and 19 base materials (26,543 combinations) returns a catalogue row that a second solver can rebuild, or, for an empty feasible set of at most eight constraints, inclusion-minimal unsatisfiable subsets (MUS) and the slack of a repair, not a neighbour. One registry declaration generates the prompt and the evaluator, so undeclared parsed keys cannot reach search. On a frozen suite of 64 typed queries, retrieve-or-refuse matches min-repair on every feasible query (48/48) and refuses every empty one (16/16) with MUS and slack. Constraint-ignoring nearest-neighbour and penalty search violate stated cost or density limits. On 216 empty queries, repairs printed at three significant figures and rounded outward stay feasible in every case (216/216). When density and cost must be kept, a list of minimum-cardinality repairs keeps them on 209 of 211 queries; a protection-first repair keeps them on all 211, the same as full diagnosis. A 308-request parse benchmark is a template-text check, not unconstrained engineer prose.
[AI-111] AppliedScientist: Automated Scientific Revision Through Iterative AI Reviewing
链接: https://arxiv.org/abs/2609.14738
作者: Vidushee Vats,Karun Sharma,Shengzhi Li,Shichao Pei
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automated reviewing systems are increasingly evaluated based on the quality of the reviews they produce. Yet a review is only useful if acting on it leads to a measurable improvement in the paper. We present AppliedScientist, a closed-loop system that couples an autonomous AI scientist with an AI reviewer, and evaluate it by iteratively revising rejected papers from a range of research subfields. To mirror how human authors build on earlier drafts, the AI scientist has access to its previous versions during revision. To avoid bias from prior judgments, however, each review is generated independently, with the reviewer having no memory of earlier feedback or scores. We compare three revision settings: one initialized with the original venue reviews, one initialized with AI-generated reviews, and autonomous self-revision using the same fixed prompt in every round. Because the reviewer both guides and evaluates the revision, we also assess the human-initialized revisions using Stanford Reviewer as an independent evaluator. Reviewer-guided revision consistently improves more than fixed-prompt self-revision, and Stanford Reviewer also assigns higher scores to later revisions. AppliedScientist resolves 128 of 150 execution-related weaknesses (85.3%), but only 2 of 18 idea-related weaknesses (11.1%), suggesting that iterative revision is effective at improving experiments and implementation, but rarely changes concerns about novelty or significance.
[AI-112] OCT-FedSIR: Toward Trustworthy Federated Ophthalmic Learning under Annotation Noise
链接: https://arxiv.org/abs/2609.14734
作者: Sina Gholami,Abdulmoneam Ali,Tania Haghighi,Rashadul H. Badhon,Behafarin Emam,Sally S. Y. Ong,Atalie C. Thompson,Theodore Leng,Ahmed Arafa,Jennifer I. Lim,Minhaj Nur Alam
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: 38 pages, 7 figures, 4 tables, 6 tables in supplementary material
Abstract:Federated learning enables collaborative model development without centralizing patient data, but annotation reliability at participating institutions cannot always be assumed. In ophthalmic imaging, differences in disease prevalence and class composition can resemble changes caused by corrupted supervision. We introduce OCT-FedSIR, a reliability-aware spectral framework for federated OCT classification under client-dependent annotation noise and heterogeneous data distributions. OCT-FedSIR combines class-balanced spectral estimation, Stage-I logit adjustment, complementary spectral descriptors, selective spectral relabeling, and noise-aware federated optimization. We evaluated the framework on the Kermany, University of Illinois Chicago, and Wake Forest datasets under symmetric and structured asymmetric noise and three levels of non-IID heterogeneity. Across 117 experimental conditions, OCT-FedSIR achieved a mean accuracy of 86.73%, compared with 79.94% for RoFL and 78.75% for FedCorr. It correctly separated clients with original and corrupted annotations across all evaluated conditions, while the original FedSIR identification procedure was less robust, particularly under asymmetric noise. Spectral relabeling recovered 77.2% of corrupted annotations with 91.3% correction precision and a 3.5% false-correction rate. Retaining corrected clients outperformed spectral pruning by 9.30 percentage points on average. These findings show that annotation noise can often be identified and corrected without discarding informative client data.
[AI-113] Bayesian Intelligence from the Outside
链接: https://arxiv.org/abs/2609.14724
作者: Alex Smolin,Bryan Wilder
类目: Artificial Intelligence (cs.AI); Theoretical Economics (econ.TH)
备注:
Abstract:Inferring intelligence from observable behavior is a foundational challenge in artificial intelligence. We develop a theory of Bayesian intelligence for agents such as language models. Each prompt induces a possibly imperfect internal experiment; the agent updates a full-support prior by Bayes’ rule and faithfully reports its posterior over the possible answers to the question. Repetitions draw fresh, independent outcomes from the same unobserved experiment at one fixed state. We show that the agent’s behavior admits this explanation if and only if its reports are not fully contradictory, i.e., some state remains possible under every report across all prompts. Report frequencies and the sizes of positive probabilities impose no further restrictions. We further propose and characterize the behavioral implications of an intelligence order that makes the behavior of two agents consistent with one agent having access to a more informative experiment: there should exist a coupling of report distributions such that the more informative agent’s report excludes every answer excluded by its counterpart. Finally, we show the difficulty of aggregating coarse reports from intelligent agents: unless the agent reports a belief about the complete state of the world, the optimal aggregation can assign arbitrary weights to states that have not been excluded. These results provide a basis for understanding when agents’ behavior is intelligent and highlight the difficulty of rejecting Bayesian rationality.
[AI-114] Carryover Drafting: Recycling Rejected States for Speculative Decoding
链接: https://arxiv.org/abs/2609.14717
作者: Jahyun Koo,Sunghyeon Woo,Jaeeun Kil,Jeongtae Lee,Sungjae Lee,Kyomin Jung,Minsub Kim
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Speculative decoding accelerates LLM inference by verifying multiple drafted tokens in parallel, allowing a single target forward pass to accept several tokens. By construction, verification computes representations for both accepted and rejected tokens. Yet, conventional drafters retain only the representations of accepted tokens, leaving the substantial verifier computation spent on rejected tokens effectively wasted. We find that these discarded hidden states generated during target forward retain useful information about future tokens that can improve subsequent drafts. However, realizing this opportunity poses two distinct challenges. At inference, recycling overhead can increase drafting latency, diminishing the speedup gained from increased acceptance length. During training, standard parallel drafter training does not produce inference-aligned rejected states, while obtaining them through sequential rollouts would sacrifice parallelism across training positions. We introduce Carryover Drafting, which addresses both challenges. Carryover recycles rejected target hidden states as temporary KV context, allowing the drafter to selectively attend to them. It reuses the drafter’s existing interface and adds only a single learned embedding to distinguish rejected states from committed context. The additional KV context is replaced each drafting round, keeping its length bounded by one proposal block. We introduce parallel draft–verify–draft training that exposes the drafter to inference-aligned rejected states while preserving parallelism across training positions. Experiments with DFlash and a DSpark-derived semi-autoregressive drafter across two target models show that this simple Carryover mechanism improves average acceptance length by 6.5–14.7% and end-to-end vLLM speedup by 7.9–14.4% over the corresponding baselines, with speedup gains reaching 28.8% on translation.
[AI-115] Moral Rebel Agents : Decision-Making Under Conflicting Obligations
链接: https://arxiv.org/abs/2609.14716
作者: Hector Munoz-Avila,David W. Aha,Paola Rizzo
类目: Artificial Intelligence (cs.AI)
备注: 17 pages, 1 table, 3 figures
Abstract:Autonomous agents are typically obliged to follow user-assigned tasks. However, strict obedience may conflict with moral obligations that arise during execution. This paper investigates \textbfmoral rebellion: the ability of an autonomous agent to deviate from a user-assigned task when morally justified. We formalize five agent architectures: an \textitamoral agent that pursues assigned tasks without considering moral obligations, and four forms of \textbfmoral rebel agency: (1) \textitutilitarian agents that opportunistically maximize task outcomes; (2) \textitdeontic agents that enforce normative constraints; (3) \textitutilitarian-deontic (UD) agents that combine deontic constraints with utilitarian reasoning; and (4) \textitdutiful agents that additionally preserve commitments to assigned tasks. We implement these architectures within a hierarchical task network planning framework and evaluate them in a Mini Search-and-Rescue domain that exposes trade-offs among assigned-task completion, opportunistic rescue, and norm compliance. Our empirical results show that the proposed agents exhibit distinct trade-offs among rescue results, assigned-task completion, and norm compliance. In particular, the preservation of task commitments emerges as an important dimension of moral rebellion, for which the UD and dutiful agents produce substantially different behaviors despite their shared utilitarian and deontological foundations. These findings highlight the importance of commitment-aware moral reasoning for autonomous agents operating in morally consequential environments.
[AI-116] Lightning Weave: Improving the Accuracy-Efficiency Frontier of Reasoning Models through Capability Composition
链接: https://arxiv.org/abs/2609.14708
作者: Yecheng Wu,Song Han,Han Cai
类目: Artificial Intelligence (cs.AI)
备注: Work in progress
Abstract:A core goal of efficient reasoning is to improve the accuracy-efficiency frontier. However, jointly improving reasoning accuracy and inference efficiency can be challenging, as the two objectives can favor different reasoning behaviors. Independently post-trained models already offer distinct strengths in accuracy and efficiency. We introduce Lightning Weave, a post-training framework that extracts and composes these independently learned capabilities in a single student through on-policy distillation. Each acquired capability is represented by the policy shift from the model before post-training to the resulting specialist. Lightning Weave combines aligned log-ratio shifts at shared student token states and uses Tilted-Target DOPD to convert the cached signals into a stable learning target. Each anchor pair scores the cached trajectories once, enabling subsequent student training without serving multiple live anchor models concurrently. Across diverse student models and benchmarks in mathematics and code, Lightning Weave substantially improves upon the base students and achieves a state-of-the-art accuracy-efficiency frontier. On Qwen3.5-4B, it raises HMMT 2025 accuracy from 59.2% to 64.0% with 10.7% fewer response tokens, and LiveCodeBench v5 accuracy from 41.7% to 54.2% with 9.6% fewer response tokens. Adjusting the relative strengths of the anchor signals yields a strong empirical accuracy-efficiency Pareto frontier. These results establish Lightning Weave as a new practical route to efficient reasoning through capability composition. Code will be released soon.
[AI-117] WaterKron and FlipFlop Hessian: Information-Theoretically Grounded Quantization with Kronecker-factored Hessians
链接: https://arxiv.org/abs/2609.14706
作者: Johann Birnick,Rayan Saab
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:How should a Kronecker-factored Hessian approximation be chosen for post-training quantization? We address this question through WaterKron, which combines two-sided GPTQ with row- and column-dependent waterfilling scales and entropy coding. We derive its high-rate distortion with respect to the full Hessian using an explicit Kronecker-Hessian mismatch factor \Phi . This factor quantifies the asymptotic distortion penalty due to the Kronecker Hessian approximation and provides a criterion for selecting the factors optimally. Minimizing \Phi leads to a Gaussian covariance-fitting problem with classical ``flip-flop’’ updates. We thus give a rate-distortion justification for using the resulting FlipFlop Hessian in quantization. We evaluate it empirically, finding that the FlipFlop Hessian consistently improves KL divergence and perplexity over input-only, marginal, and Frobenius-based Hessian choices.
[AI-118] MANE: A Multi-Path Adaptive Network for Edge Onloading of Deep Neural Networks
链接: https://arxiv.org/abs/2609.14660
作者: Sokratis Nikolaidis,Stylianos I. Venieris,Leonidas Malachias,Iakovos S. Venieris
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注: 6 pages, 5 figures, submitted to IEEE UIC 2026
Abstract:Split computing constitutes a widely used distributed inference approach, where a lightweight head model is onloaded onto the device and a heavier tail model resides on an edge server, leveraging the growing computational capabilities of modern System-on-Chips while alleviating server load. As intelligent indoor environments such as smart offices grow increasingly populated with diverse IoT devices, a single edge server must simultaneously assist multiple devices, each competing for the same shared inference resources. Without a principled mechanism to manage this shared load, the server is quickly overwhelmed, causing latency SLO violations and rendering server-assisted inference ineffective. In this work, we present MANE, a distributed inference framework that equips the server with a multi-path tail architecture, enabling a dynamic accuracy–throughput trade-off at runtime. By introducing a novel multi-path model architecture, a three-stage training scheme featuring a Joint Head Network Distillation loss and a hysteresis-based scheduler with an equitable device-fallback policy, MANE maintains over 80% SLO satisfaction rate where state-of-the-art onloading methods fail completely, while preserving accuracy 6pp higher than on-device alternatives, across up to 40 concurrent devices.
[AI-119] Natural Language Knowledge Graph Query Execution: Leverag ing Controlled Semantics in the LLM Context Window
链接: https://arxiv.org/abs/2609.14652
作者: Blake G. Fitch
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注: 9 pages, 4 tables
Abstract:Large Language Model (LLM) applications often transfer domain concepts into the model’s context informally, through prompt prose, schema dumps, and examples. We show that for database queries, data model concepts pass to LLMs more effectively through representations whose vocabulary terms carry declared, machine-readable semantics (controlled semantics). NLKGQ is a working system and reusable framework that does this for data modeled in a knowledge graph. A formal OWL ontology serves as the transfer mechanism, concentrating the meaning of the data into semantically precise tokens the model can use directly. In a single LLM call, NLKGQ places in the context a system prompt instructing on SPARQL, the complete domain OWL ontology, and a domain-specific prompt addition, together with the user’s natural language query. The model then generates the SPARQL query directly, zero-shot. Where the native vocabulary of an existing database or federation of databases is opaque, a wrapper ontology substitutes clean terms and a runtime rewriter restores the native forms. Evaluating on DBLP-QuAD 2.0 showed that its scores depend on the graph snapshot, the endpoint used, and the wording of its machine-generated questions, so we propose DBLP-QuAD 3.1, which maintains the intent of 2.0 while making reference results deterministic, revising reference SPARQL where needed, and rewriting the natural language questions, with a frontier model, to state each reference query’s intent clearly and completely. We evaluate on the DBLP-QuAD 2.0 benchmark (57.6% Match under deterministic re-scoring), DBLP-QuAD 3.1 (89.9% Match on 1,000 questions), SemOpenAlex (98% Match against a published baseline’s 86% on the identical test set), and neuroimaging metadata (100%).
[AI-120] Skill Composition for Legged Robot Reinforcement Learning
链接: https://arxiv.org/abs/2609.14647
作者: Daniel Gigliotti,Flavio Maiorana,Fabio Patrizi,Luca Iocchi
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Robots, and humanoid robots in particular, are increasingly competent at individual behaviors, each obtained by training a specialized controller. A specialized skill is quick to train, converges reliably because the problem it faces is narrow, and can be validated on its own, none of which is true of a single end-to-end policy asked to cover everything. What remains fragile is the transition between them. We argue that the composition of independent sub-policies deserves to be treated as a research problem in its own right, rather than as an implementation detail left to whatever mechanism happens to be at hand. Reliable composition is what turns a collection of separate skills into a repertoire that can be used, extended and shared. More fundamentally, if control can be passed between specialized policies safely, and at any moment, the choice of what the robot should do next can be delegated to a component of an entirely different nature, such as a planner, an automaton or a symbolic controller, whose behavior can be inspected in advance. The policies would then only ever have to act, and what the robot can be trusted to do would become verifiable.
[AI-121] Diffusion-Based Generation of Gait Trajectories
链接: https://arxiv.org/abs/2609.14642
作者: Damian Benasco,Juan Carballeira-Lopez,Jaime Ramos-Rojas,Julio S. Lora-Millan,Antonio J. Del-Ama,David Rodriguez-Cianca,Pablo Lanillos
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO)
备注: International Conference on NeuroRehabilitation (ICNR2026), September 29-October 2, 2026, Seoul, South Korea
Abstract:Generation of musculoskeletal gait trajectories conditioned on patient-specific parameters remains a key challenge for wearable robotics and rehabilitation. Assistive systems such as lower-limb exoskeletons require reference trajectories that adapt to individual morphology and therapeutic goals while preserving biomechanical realism. Traditional approaches rely on hand-crafted gait templates or optimization procedures that scale poorly across subjects and walking conditions. In this work, we explore conditional diffusion models for generating lower-limb joint-angle trajectories conditioned on gait parameters such as step length. We compare a baseline transformer diffusion model with a controllable diffusion transformer variant incorporating adaptive normalization and classifier-free guidance. Experiments on a dataset of 4,590 gait cycles show that diffusion models can generate realistic periodic gait trajectories while enabling some controllability variation in gait characteristics, highlighting their potential for personalized gait synthesis in assistive robotics.
[AI-122] DynSTEER: Dynamic Stage-wise Trajectory Evaluation and Execution-time Review for Agents
链接: https://arxiv.org/abs/2609.14637
作者: Zhichao Shi,Wenjie Zhang,Xuhui Jiang,Xiaojun Wu,Cehao Yang,Chengjin Xu,Jian Guo,Yuanzhuo Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model agents are increasingly deployed for long-horizon task execution. However, current evaluation paradigms face three major limitations: terminal-only assessment ignores intermediate processes and struggles to localize errors efficiently and accurately, single-reference matching penalizes valid alternative solution paths, and post-hoc trajectory judging incurs high costs without the ability to halt failed runs early. To address these issues, we propose DynSTEER, a dynamic stage-wise trajectory evaluation framework for agents. DynSTEER segments rollouts into stages anchored by key completed actions, focusing evaluation on essential milestones with adequate context while enabling targeted strategy adjustments. It compiles a path-tolerant milestone graph from public task views to respect diverse legitimate strategies without leaking ground truth. Furthermore, it adaptively routes evaluation queries across multi-tier judges and halts unrecoverable executions online to curb resource waste. Experiments demonstrate that DynSTEER improves evaluation discriminability across LLM agents by 85.2% over native evaluation, separates all model pairs with statistical significance, and saves 34.51% of execution steps on failed rollouts.
[AI-123] A note on goal-based hierarchical RL
链接: https://arxiv.org/abs/2609.14605
作者: Kevin Murphy
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The agent-centric general value function (ACGVF) construction of \citettasse2026goal lets the agent make two decisions that are normally imposed by the environment or agent designer: which goal to pursue and when to declare a goal as finished (in addition to choosing the action). This is a very general framework that subsumes almost all prior work on reinforcement learning, control and planning, as well as more general formalisms proposed in the cognitive sciences. However, it assumes the environment is fully observed, i.e., that the observation is a sufficient statistic. In \citetmurphy2025rl, a general agent design was proposed where the policy is based on an internal belief state z_t and an internal goal; however, the goals were assumed to be externally provided. In this note, we unify and extend these two approaches using the formalism of hierarchical hidden Markov models (HHMM) \citepmurphy2001hhmm. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14605 [cs.AI] (or arXiv:2609.14605v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.14605 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Kevin Murphy [view email] [v1] Sun, 13 Sep 2026 15:37:47 UTC (15 KB) Full-text links: Access Paper: View a PDF of the paper titled A note on goal-based hierarchical RL, by Kevin MurphyView PDFHTML (experimental)TeX Source view license Current browse context: cs.AI prev | next new | recent | 2026-09 Change to browse by: cs References Citations NASA ADSGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv’s community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) mathjaxToggle(); We gratefully acknowledge support from our major funders, member institutions, , and all contributors. About Help Contact Subscribe Copyright Privacy Accessibility Operational Status (opens in new tab) Major funding support from
[AI-124] Diagnosing Temporal Misalignment in Multichannel Time-Series Classification with Minimum Description Length
链接: https://arxiv.org/abs/2609.14595
作者: Sebastian Buschjäger,Michael Frichert,Daniel Kuhe,Jian-Jia Chen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
备注: 13 pages (8+5 appendix)
Abstract:Multichannel time-series classification commonly assumes synchronized sensor streams, although latency, clock drift, and preprocessing can introduce relative delays during data collection or after deployment. Existing synchronization solutions are often hardware-specific and difficult to apply retrospectively. Consequently, synchronization problems may remain undetected while classification performance is suboptimal. We introduce a classifier- and label-free diagnostic based on minimum description length (MDL). Our method applies candidate temporal shifts to sensor groups and measures how efficiently one group can be encoded through a representation of the remaining channels. An increased codelength indicates that the shift destroys shared temporal structure, whereas the minimum identifies the alignment most strongly supported by the data. Unlike learned synchronization methods, the diagnostic requires neither retraining nor a trusted aligned reference and can therefore test both training and deployment data for misalignments. Experiments on two controlled synthetic tasks and nine real-world datasets show that the metric exposes alignment structure and can recover accuracy under induced deployment drift. A whole-dataset audit further identifies stable nonzero MDL optima in established benchmarks including FordChallenge, Opportunity, PAMAP2, and UCIActivity, revealing potential systematic offsets that conventional model evaluation does not expose. Our method thus provides a general-purpose tool for detecting, understanding, and correcting temporal misalignment throughout the time-series learning pipeline. Our code is available under this https URL.
[AI-125] SENTINEL: A Multi-Pathway Architecture for Detecting Living-Off-the-Land APT Attacks on Windows Command Lines
链接: https://arxiv.org/abs/2609.14593
作者: Ahad Bin Islam Shoeb,Kamrul Hasan,Jamal Uddin Tanvin,Liang Hong,Imtiaz Ahmed,Md Arif Billah,Al Amin
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted for publication in the Proceedings of the 2026 IEEE Military Communications Conference (MILCOM 2026). This is the authors’ accepted version; the final published version will appear in IEEE Xplore. 7 pages, 6 figures, 1 table
Abstract:Living-Off-the-Land (LOTL) is the dominant evasion technique of Advanced Persistent Threat (APT) actors, exploiting legitimate Windows utilities to conduct malicious operations without deploying custom malware and enabling state-sponsored campaigns to maintain persistent access within military and critical defense infrastructure for extended periods. Existing detection methods fail against obfuscated commands and multi-stage attack sequences, as demonstrated by the Volt Typhoon APT campaign, which maintained undetected access to U.S. critical infrastructure for over 18 months using exclusively signed Windows utilities. We present SENTINEL, a multi-pathway architecture integrating BERT-based semantic encoding, character-level CNN for obfuscation invariance, inter-command attention for multi-stage pattern recognition, and autoencoder-based anomaly scoring. Evaluated on a balanced Volt Typhoon benchmark derived from Microsoft and CISA threat intelligence advisories, SENTINEL achieves 92.0% accuracy on documented state-sponsored attack commands and 91.2% on obfuscated variants, compared to 74.0% and 72.0% for standalone BERT. Per-class analysis reveals that models achieving over 98% overall validation accuracy on imbalanced data exhibit only 44-58% malicious recall on balanced adversarial sets. Character-level processing contributes 5.6 percentage points of obfuscation invariance, and the 8.0 percentage point gap over augmentation-only baselines confirms structural architectural value beyond data-driven robustness alone.
[AI-126] AI Deployment Accountability Engineering: A Vision for Accountable AI in Safety-Critical Socio-Technical Systems
链接: https://arxiv.org/abs/2609.14592
作者: Murat Kantarcioglu
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Computers and Society (cs.CY)
备注: Appeared in ACM AI Summit 26, Atlanta, GA as a vision paper
Abstract:Artificial intelligence systems are rapidly becoming critical components in healthcare, finance, public services, and other safety-critical domains. Yet the engineering practices used to evaluate these systems remain predominantly model-centric, emphasizing properties such as accuracy, robustness, fairness, and interpretability before deployment. These properties are necessary but insufficient once an AI system operates within an ever changing socio-technical environment characterized by distribution shifts, institutional constraints, human feedback loops, privacy requirements, and interactions among multiple AI agents. This vision paper introduces AI Deployment Accountability Engineering (ADAE), a proposed AI engineering subdiscipline concerned with establishing measurable, continuous, and actionable accountability for deployed AI systems. ADAE treats accountability as a deployment-layer property rather than solely as a property of an individual model. It seeks to determine whether an AI-enabled system continues to operate within acceptable risk limits, identify the contexts in which failures emerge, attribute failures across interacting technical and human components, translate technical failures into downstream consequences, and support timely intervention. We articulate a research agenda built around four interconnected pillars: structured discovery of context-dependent failure modes, privacy-preserving accountability measurement, system-level risk analysis for agentic AI, and translation of technical failures into operational, and institutional risks. The broader goal is to establish foundational principles, mathematical tools, and system architectures for accountable AI deployment across safety-critical applications. Comments: Appeared in ACM AI Summit 26, Atlanta, GA as a vision paper Subjects: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Computers and Society (cs.CY) Cite as: arXiv:2609.14592 [cs.AI] (or arXiv:2609.14592v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.14592 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-127] OptoAgent : A Trustworthy Multi-Agent Framework for Opportunistic Vision Micro-Screening in Classroom Environments
链接: https://arxiv.org/abs/2609.14514
作者: Toqeer Ali Syed,Ali Akarma,Adeel Ahmad,Hammad Muneer
类目: Artificial Intelligence (cs.AI)
备注: Paper submitted in ICAAD Conference 2026
Abstract:A child with reduced distance vision often does not know that anything is wrong. Children adapt, move closer, and rarely report the difficulty, so the problem can survive years of schooling before an adult notices. School screening addresses part of this, but it runs on a schedule, depends on staffing, and is separated from the classroom moments where the difficulty appears. Smartphone and web-based acuity tests have widened access, yet every one of them still needs somebody to start a test. We present SightSentinel, an architecture that turns a wall display a child already reads from into a recurring screening site. Ordinary educational content carries short calibrated optotype probes, and eight specialized agents divide the work. Perception agents recover viewing distance, recognition accuracy, approach behavior, gaze stability, response latency, and interocular difference from each encounter. A quality agent discards observations taken under bad geometry, poor lighting, or inattention. A longitudinal agent accumulates only the surviving evidence against the child’s own baseline, and an orchestrator reports a Vision Concern Score routed through a safety gate whose output range excludes diagnosis, refraction, prescription, and reassurance. The design question is whether many cheap, noisy, well-gated encounters can reach a referral decision that one scheduled test reaches late or misses. We state the formulation, the architecture, a four-stage validation protocol against clinical reference standards, and the conditions under which the approach should be rejected.
[AI-128] Beyond Scene Description: Multi-Agent Orchestration for Non-visual Access to Virtual Worlds
链接: https://arxiv.org/abs/2609.14512
作者: Toqeer Ali Syed,Ali Akarma,Adeel Ahmad,Danial Hameed
类目: Artificial Intelligence (cs.AI)
备注: Paper submitted in ICAAD Conference 2026
Abstract:Virtual worlds now host classrooms, meetings, conferences, shops, and social venues, and nearly every interaction they expose assumes a user who can scan a three-dimensional scene, follow avatars, and read floating panels. Blind and visually impaired (BVI) users are left with assistive tools that each solve one task in isolation: naming an object, reading text, describing a scene, or planning a route. A live virtual room defeats that model: obstacles, speakers, gestures, chat, slides, and notifications arrive together, and a tool that narrates all of them trades a visual barrier for an auditory one. This paper presents MetaBlind, an architecture that distributes nonvisual access across eight specialized agents, spanning perception, navigation, social and object interaction, communication, safety and trust, memory, and personalization, and that places an Accessibility Orchestrator between those agents and the user. Agents publish candidate information into a shared accessibility context instead of speaking to the user directly. The orchestrator scores each candidate on safety relevance, goal relevance, urgency, confidence, user relevance, and estimated listening load, then releases only the items it judges relevant at that moment through speech, structured audio, or haptic output. We give the selection step a formal statement, specify the orchestration cycle as an algorithm, and define an evaluation protocol against a single-agent assistant. MetaBlind is reported at the design stage, with no prototype measurement or user study, and the protocol states which outcomes would support the design and which would refute it.
[AI-129] When does a scaling result justify a different allocation? A critical review of resource-allocation evidence for AI systems
链接: https://arxiv.org/abs/2609.14500
作者: Seyed Morteza Emadi
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 35 pages, 3 figures, 9 tables
Abstract:AI scaling studies increasingly evaluate systems that combine a pretrained model with retrieval, search, verification, tools, and interaction. Yet a higher score under a larger budget does not by itself show where additional resources are best spent. This critical integrative review asks when a reported scaling result supports a resource-allocation decision. It compares evidence across pretraining, test-time computation, retrieval, and agent evaluation, distinguishing the performance of a tested procedure from the best performance achievable under a resource limit. The synthesis shows that three mismatches recur across this evidence: success counted before an answer is chosen, information a deployed system will not have, and costs left out of the comparison. A capability surface expresses performance as a function of budgets, mechanisms, and available information. Worked analytical examples show how the evaluation metric, deployment volume, selection rule, and stopping policy can alter an allocation conclusion. A resource envelope provides a structured record of the task, development and run-time resources, information access, and procedure behind a reported score. Its application to a published comparison illustrates which conclusions the evidence supports and which deployment questions remain unresolved. The resulting framework specifies the comparisons needed to choose among feasible systems and motivates experiments on the transfer of allocation rules across tasks and operating conditions. It does not propose a universal scaling law or infer general intelligence from benchmark gains.
[AI-130] EdgeHAR: An Edge-Native Compact Sensor Foundation Model for Human Activity Recognition
链接: https://arxiv.org/abs/2609.14498
作者: He Zhang,Siyu Yuan,Siyu Liu,Sizhen Bian,Bin Guo
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Sensor-based human activity recognition (HAR) is fundamental to ubiquitous and wearable computing, yet existing foundation models are largely designed for cloud-scale deployment and struggle with real-world sensing shifts, including unseen users, devices, sampling rates, and sensor placements. We present \textbfEdgeHAR, an edge-native compact sensor foundation model designed for wearable intelligence. Unlike conventional models that entangle activity knowledge with acquisition variations, EdgeHAR learns transferable representations by factorizing sensor signals into three latent codes: an \textbf(i)Activity-Semantic Code capturing reusable activity knowledge, a \textbf(ii)Motion-Dynamics Code modeling temporal patterns, and an \textbf(iii)Acquisition-Context Code representing sensor-specific variations. This disentangled design enables efficient adaptation to new users, devices, placements, and activity classes with limited target-domain data. By incorporating lightweight adaptation modules, EdgeHAR achieves foundation-model-level transferability while satisfying edge constraints in computation, memory, latency, and privacy. Experiments across heterogeneous HAR datasets demonstrate that EdgeHAR maintains competitive recognition performance under distribution shifts with substantially reduced deployment cost. EdgeHAR establishes a practical paradigm for compact, edge-first foundation models for ubiquitous sensing systems.
[AI-131] Bridging the Modality Gap in Long-Form Clinical Audio: A Comparative Study of Lightweight and Heavyweight End-to-End SOAP Generation
链接: https://arxiv.org/abs/2609.14467
作者: Ziyu Zhang,Mingchen Shao,Wenjie Tian,Tianlun Zuo,Longhao Li,Lei Xie
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注:
Abstract:Automating clinical documentation from long-form doctor-patient conversations remains challenging for modern audio-language models. While cascaded ASR systems perform well, end-to-end (E2E) models often struggle with information loss and hallucinations on extended audio. For the BeTraC 2026 challenge, the ASLP team presents a fully E2E multimodal system that generates structured SOAP notes directly from audio, bypassing intermediate transcripts. We constructed a 1.41-million-sample multi-task corpus and applied a multi-stage pipeline: domain pre-training, supervised fine-tuning, and reward optimization. Evaluating the architecture under both Lightweight (3B) and Heavyweight (30B) constraints reveals that each training stage progressively enhances performance. Furthermore, scaling to 30B parameters substantially boosts concept extraction and summarization quality. Ultimately, our E2E systems consistently outperform representative cascaded ASR+LLM baselines, proving the efficacy of direct multimodal optimization for clinical documentation.
[AI-132] Follow the Geometry Not the Model: Cold Start Semi-Supervised Learning
链接: https://arxiv.org/abs/2609.14451
作者: Itai David,Daphna Weinshall
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Modern semi-supervised learning (SSL) couples pseudo-label generation and classifier training, using the classifier’s own confidence to select the pseudo-labels that are then used to update the model. In the cold-start regime, where at most a few labels per class are available, this coupling is ill-posed, since the classifier cannot supervise itself before it has learned. To address this problem, we propose VAST (Veracity-Aware Semi-Supervised Training), which decouples these two stages. Probabilistic beliefs over the unlabeled set are first inferred directly from the geometry of a frozen self-supervised embedding and only then distilled into an inductive classifier. The construction rests on the Veracity Matrix, a kernel-based structure that aggregates label evidence across the data manifold and admits an interpretation as a Dirichlet posterior under a per-observation powered-likelihood model. Additionally, we introduce Veracity Propagation, a self-terminating belief-spreading step that extends coverage beyond the kernel neighborhood of the labeled set. Under a controlled protocol in which all methods receive identical frozen embeddings and labeled sets, VAST outperforms the strongest graph-based SSL baselines at every operating point across three datasets, with statistically significant gains in 7 of 9 comparisons, while producing a deployable inductive classifier rather than requiring transductive graph inference. Compared with end-to-end confidence-gated SSL, we further find that these methods underperform in this setting and, in our experiments, do not consistently exceed labeled-only performance. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14451 [cs.LG] (or arXiv:2609.14451v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.14451 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Itai David [view email] [v1] Sun, 13 Sep 2026 11:46:26 UTC (1,360 KB)
[AI-133] Safety Signals to Verify NetOps Agents with Action-Level Granularity
链接: https://arxiv.org/abs/2609.14422
作者: Tobias Labarta,Frederik Pahde,Novak Boskov,Maximilian Dreyer,David Birkenberger,Manzoor Ahmed Khan,Sebastian Lapuschkin,Wojciech Samek
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Agentic Network Operations (NetOps) are an emerging paradigm promising to enable workload-aware, self-adjustable, and reliable autonomous networks. While agents have proven their value in incident summarization and telemetry signal extraction, their effectiveness as autonomous control-loop engines heavily relies on their long-horizon reliability. One such setting is the datacenter fabric, where an agent must respond to alarms and operator intents while abstaining from high-risk actions that may cause or extend downtime. Abstention, however, presupposes that an action’s impact is known pre-execution, which necessitates a per-action ground truth that NetOps agent benchmarks do not provide. We construct such a ground truth for the network repair task of NetArena. A symbolic replay of the emulated network, validated against the environment at every turn, yields the exact value of every action. From the action-level value, we derive two pre-execution targets, namely whether an action reduces the repair distance (progress) and whether it increases it (harm). We show across 10 agent models, that agent verifiers leveraging internal signals predict both harm and progress more reliably than a baseline using observable signals only. Perspectively, we aim to use these signals as safety feedback to an agent harness to abstain from risky actions and protect the target system.
[AI-134] Surrogate-Assisted Genetic Programming with Phenotypic Characterisation in Dynamic Multi-Mode Project Scheduling
链接: https://arxiv.org/abs/2609.14418
作者: Yuan Tian,Yi Mei,Mengjie Zhang
类目: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI)
备注: 15 pages, 18 figures, 7 tables
Abstract:Dynamic multi-mode resource-constrained project scheduling requires decisions to be made under precedence constraints, limited resources, multiple execution modes, and uncertain activity durations. Genetic programming (GP) can automatically evolve heuristic rules for such problems, but its simulation-based fitness evaluation is computationally expensive. This study investigates phenotypic characterisation (PC) in surrogate-assisted GP to evolve higher-quality scheduling heuristics under a fixed budget of full simulation-based fitness evaluations. A key question is how GP individuals should be encoded into phenotypic characterisations to support effective fitness estimation. To answer this question, three PC encoding schemes with different levels of information richness are designed: priority-value encoding, which preserves raw rule outputs; rank encoding, which captures candidate ordering; and binary encoding, which represents final scheduling decisions. These encodings are combined with different distance metrics to measure behavioural similarity between GP individuals. The experimental results show that binary encoding with Euclidean distance provides the most effective and robust surrogate guidance. Further analyses show that surrogate estimation accuracy alone does not fully explain the performance differences. The PC representation also determines how effectively phenotypically redundant offspring are removed and how much behavioural diversity is retained after preselection. Ablation experiments further demonstrate that duplicate removal and surrogate preselection provide complementary benefits, with their combination producing the largest improvement. These findings highlight that effective surrogate-assisted GP depends not only on identifying promising offspring, but also on controlling redundancy and preserving useful diversity during evolutionary search.
[AI-135] Dynamic Learning Solutions: A System for Personalized Educational Video Generation
链接: https://arxiv.org/abs/2609.14408
作者: Siddhanth Sridhar,Shreya Chaurasia,Baddela Sai Yaswantha Reddy,Deepak Parmar,Shylaja S S
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We present an automated pipeline that converts NCERT textbooks into interactive video explanations that respond directly to user queries. A user uploads a PDF and asks a question; the system then generates a video-based explanation as output, handling both text and visual elements from the PDF for multi-modal retrieval and response generation. The pipeline combines a Retrieval-Augmented Generation (RAG) model with generative multimedia components. The RAG stage is optimized for the structure of NCERT textbooks and performs best on content from those books. Given a user query, the RAG model retrieves relevant content from the PDF and generates a multi-scene script containing narrative explanations and structured visual prompts aligned with the textbook’s explanatory style. These prompts are passed to a Stable Diffusion module, implemented layer by layer for interpretability and control, which generates contextually relevant images. The images are then processed by DynamiCrafter to produce animated sequences. Finally, a Google Text-to-Speech module generates synchronized narration, aligning speech with the visual scenes through time-based control. The result is a coherent video explanation integrating animation, narration, and textbook-aligned visuals, transforming static educational material into an engaging learning experience. By combining multi-modal document retrieval, generative visual models, animation frameworks, and speech synthesis, this pipeline demonstrates a scalable approach to delivering interactive, personalized digital education content.
[AI-136] LLaTSA: Large Language Model-Aligned General-Purpose Transient Stability Analysis
链接: https://arxiv.org/abs/2609.14374
作者: Chao Shen,Hongwei Zhen,Junyan Shao,Zhenghao Yang,Yifan Zhang,Mingyang Sun
类目: ystems and Control (eess.SY); Artificial Intelligence (cs.AI)
备注:
Abstract:Dynamic trajectory prediction has become an important paradigm for data-driven transient stability analysis (TSA), yet most existing predictors remain system-specific and require substantial retraining when network configurations, generation mixes, or state-variable sets change. Uni-TSA introduced a general-purpose TSA framework that combines channel-independent modeling with a pretrained large language model (LLM) predictor. Nevertheless, its application to heterogeneous systems is limited by ambiguity in short observations, a mismatch between numerical trajectories and LLM embeddings, neglected coupling among state variables, and the high inference cost of dense backbones. This paper proposes LLaTSA, an LLM-aligned framework for general-purpose trajectory-based TSA. LLaTSA first incorporates operating conditions, disturbance attributes, and state-variable identity through a structured textual prefix. It then aligns normalized temporal patches with a TSA-related vocabulary before processing them with a pretrained sparse decoder-only mixture-of-experts (MoE) backbone. A state-variable coupling module captures coordinated post-fault evolution, while teacher forcing and rollout-based training support iterative long-horizon prediction. Case studies on multiple test systems demonstrate accurate trajectory prediction, reliable stability discrimination, and effective adaptation across unseen scenarios.
[AI-137] AURA: Unified Multimodal Framework for Conversational Music Editing
链接: https://arxiv.org/abs/2609.14344
作者: Quoc-Huy Trinh,Minh-Van Nguyen,Debesh Jha
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注:
Abstract:Instruction-guided music editors typically process each request independently, limiting their ability to support workflows in which users progressively refine a track. We introduce AURA, a unified multimodal framework for conversational music editing. AURA uses a multimodal large language model to interpret the complete dialogue history, an optional image, and reference audio, distilling the editing intent into compact concept tokens. A concept-to-audio module injects these tokens and frame-aligned reference features into a frozen MusicGen backbone, enabling precise edits while preserving unaffected content. AURA optimizes only 91M parameters while retaining 1.9B frozen backbone parameters. Experiments on Slakh2100 and MoisesDB demonstrate substantial improvements in edit correctness and content preservation over existing instruction-guided methods, including a 4-5 times reduction in FAD for out-of-domain addition and removal.
[AI-138] Biquaternionic Space with Complex-valued Attention for Temporal Knowledge Graph Completion
链接: https://arxiv.org/abs/2609.14279
作者: Rushan Geng,Cuicui Luo
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Temporal knowledge graph embedding (TKGE) models infer missing facts in knowledge graphs that evolve over time. Many existing models use a single geometric space, which can limit their ability to represent diverse relational patterns, or treat entity representations as static. We propose Biquaternionic Space with Complex-valued Attention (BSCA), a TKGE model that combines circular and hyperbolic rotations within a unified biquaternionic framework. A complex-valued attention mechanism adaptively fuses time-conditioned and relation-conditioned entity representations, allowing them to vary with temporal and relational context. Experiments on five benchmark datasets show competitive performance across datasets, with the largest improvement on GDELT: BSCA achieves an MRR of 52.1%, compared with 38.1% for the strongest baseline in our comparison.
[AI-139] ATTRICITE: Training an Open 4B Model for Citation Recovery toward Faithful Attribution
链接: https://arxiv.org/abs/2609.14248
作者: Yee Man Choi,Xuehang Guo,Songcheng Cai,Yimu Wang,Yi R. Fung,Qingyun Wang
类目: Digital Libraries (cs.DL); Artificial Intelligence (cs.AI)
备注: Work in Progress
Abstract:Faithful citation attribution begins with identifying the intended source for a scientific claim. We study this source-identification capability through citation recovery: recovering the paper cited by the original author from a citation-bearing passage. Our evaluation adopts the published author’s citation as an observable human attribution signal and uses target recovery as a proxy for progress toward faithful attribution. We introduce ATTRICITE, an open 4B-parameter model trained for tool-using citation recovery within the CiteGuard retrieval environment, together with CITEALIGN, a 7,607-instance computer-science dataset drawn from recent scientific literature. For controlled evaluation, we construct a 709-instance benchmark subset of CITEALIGN, comprising 410 development instances from 2024 publications and 299 temporally held-out test instances from 2025 publications. Across three runs at an inference temperature of 0.7, GRPO fine-tuning improves Qwen3-4B from 49.4% \pm 1.5% to 59.8% \pm 0.2% target-match accuracy, a gain of 10.4 percentage points. Despite using only 4B parameters, ATTRICITE outperforms gpt-oss-20b and comes within 3.9 points of GPT-5.4-mini, while Gemma 4 31B IT achieves the strongest overall performance at 72.0% \pm 1.0%. We release the model and collection pipeline this https URL to support reproducible research on citation recovery toward faithful attribution in a continually evolving scientific literature.
[AI-140] OpWeave: Flexible Operator Disaggregation for Heterogeneous LLM Serving
链接: https://arxiv.org/abs/2609.14237
作者: Zikun Li,Yixuan Mei,Shiqi Pan,Zixuan Chen,Xiaowen Zhang,Mengdi Wu,Shuhuai Lin,Yutong Yang,Zhihao Zhang,Xupeng Miao,Rashmi Vinayak,Zhihao Jia
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注: 24 pages, 14 figures, including references and appendices
Abstract:LLM serving systems increasingly disaggregate inference into finer-grained stages, with recent approaches separating attention from FFN or MoE execution during decode. This operator-level disaggregated serving (ODS) can improve hardware matching and enable independent scaling, particularly across heterogeneous devices. However, existing systems fix operator boundaries and lack a unified characterization of when disaggregation reduces serving cost. We present OpWeave, an end-to-end framework for heterogeneous ODS. OpWeave provides an analytical cost model that bounds the gains of homogeneous and heterogeneous ODS over colocated serving. It jointly optimizes operator partitioning and deployment configuration through a regularity-aware planner that keeps the search tractable even for hybrid-attention models. A vLLM-based runtime executes the synthesized plans with flexible operator stages across heterogeneous device groups. In our evaluation, OpWeave reduces serving cost by up to 1.78\times on homogeneous and 1.89\times on heterogeneous GPU clusters relative to the best feasible baseline, while meeting latency SLOs.
[AI-141] Graph-Transformer Fraud Detection with Self-Supervised Pretraining and Conformal Risk Control
链接: https://arxiv.org/abs/2609.14234
作者: Sergei,Komarov
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 9 fig, 8 table
Abstract:Financial fraud in corporate transaction networks has grown more coordinated and harder to detect with rule-based engines and with classical learning models that treat each transaction in isolation. This paper presents GTFD, a graph-transformer fraud detector that fuses structural and temporal evidence from a corporation’s payment graph. GTFD encodes the graph with a multi-head graph attention network, encodes ordered transaction sequences with a gated transformer, and combines both views through a cross-modal gating layer. A conformal risk-control head converts the fused representation into threshold-free anomaly scores with finite-sample coverage guarantees, and the network is trained with self-supervised link-mask pretraining plus adversarial augmentation so it remains stable under scarce labels and under adversarial perturbation. On a corporate transaction benchmark enriched with coordinated fraud rings, GTFD reaches an AUROC of 0.990, an F1-score of 96.1% (precision 96.3%, recall 95.9%), and an accuracy of 98.4%. It reduces the false-positive rate by about 29% relative to the strongest baseline while raising coordinated fraud-ring recall from 85.1% to 96.5%. Ablations attribute roughly 2.0 AUROC points to self-supervised pretraining and 1.9 AUROC points to the conformal head, and adversarial stress tests show GTFD retains 89.2% accuracy at perturbation magnitude 0.20 where the next-best model falls to 76.4%.
[AI-142] ECAS: An Edge-Controlled Agent ic System for Validation-Gated Scientific Application Execution
链接: https://arxiv.org/abs/2609.14211
作者: Baixi Sun,Mingze Xia,Huihuo Zheng
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注:
Abstract:Scientific applications increasingly rely on high-performance computing (HPC), yet translating a scientist’s high-level goal into a correct target-scale execution remains brittle and labor-intensive. Large language model (LLM) agents promise to automate this, but two obstacles remain: granting a cloud-hosted model direct HPC access exposes credentials and execution authority, while withholding it demands continuous human supervision; and one-shot generation cannot adapt when generated artifacts fail in a site-specific HPC environment. We present \textscECAS, an \textbfEdge-\textbfControlled \textbfAgentic \textbfSystem for closed-loop execution of scientific computing campaigns with limited human intervention. \textscECAS separates \emphreasoning, \emphcontrol, and \emphexecution: a cloud-hosted LLM proposes plans, artifacts, and repairs; a user-controlled edge agent retains credentials, workflow state, and execution authority while enforcing policy and resource constraints; and the HPC system computes. Its core mechanism is \emphvalidation-gated execution: generated artifacts pass static checks and small-scale validation, failures trigger repairs from sanitized execution feedback, and target-scale execution is permitted only after validation and policy checks pass. \textscECAS also draws on an edge-resident library of expert-distilled, site-specific skills that is never disclosed to the cloud. In preliminary experiments with three scientific applications on two production ALCF systems under six injected fault types, closed-loop repair improves application success from 0/6 to 6/6 over one-shot generation, validation gating prevents all three observed target-scale failures, and skill conditioning improves success from 4/6 to 6/6. These results show the feasibility of delegating adaptive reasoning to the cloud while retaining execution control at the edge.
[AI-143] Data-free On-policy Distillation
链接: https://arxiv.org/abs/2609.14193
作者: Gengsheng Li,Mao Zheng,Mingyang Song,Jie Sun,Zeyuan Liu,Ruiqi Liu,Qiyong Zhong,Haiyun Guo,Junfeng Fang,Jinqiao Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:On-policy distillation (OPD) has become a standard component of frontier post-training pipelines, yet how much its training data actually contributes has gone largely unexamined. On the two teacher-student pairings most common in practice, we find OPD almost indifferent to its data: eight prompts already match a 17k-problem dataset, and three independently built datasets whose difficulty and teacher-student KL differ several-fold produce nearly indistinguishable training curves. Two causes account for this. First, the unit of data in OPD is the state a prompt leads to, not the prompt itself: a single prompt keeps exposing new teacher correction as sampling continues, while the marginal value of additional prompts collapses after eight. Second, replacing mathematics with competitive programming still recovers over ninety percent of the in-domain gain, indicating that OPD transfers the teacher’s mode of reasoning rather than knowledge related to the data. We take this to its limit with Data-free On-policy Distillation (DF-OPD), in which the teacher writes its own training questions under a simple prompt – no external data, no filtering – leaving a system of just two policies. DF-OPD matches and even surpasses real data, and the questions it produces track the teacher’s own post-training data on three key diagnostics of training dynamics, which other real datasets do not. Applied to multi-teacher distillation, where the (prompt, domain) pairs normally have to be derived from post-training data that is often out of reach, 1k self-generated questions close 98.5% of the available headroom, even surpassing the 96.6% reached with 7k real examples. Moreover, together these results invite a reassessment of the role data plays in OPD.
[AI-144] Entropy-Punctured Bloom Filters for Memory-Efficient Machine Learning ICTAI2026
链接: https://arxiv.org/abs/2609.14187
作者: John Cartmell,Mihaela Cardei,Ionut Cardei
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted for publication in the Proceedings of the 38th IEEE International Conference on Tools with Artificial Intelligence (ICTAI 2026). 9 pages, 7 figures
Abstract:Memory-efficient feature representations are increasingly important in machine learning settings where storage, transmission cost, bandwidth, or privacy constraints limit access to raw data. Bloom Filter (BF) encodings provide compact probabilistic representations of engineered features, but their behavior under structural compression and their applicability to regression tasks remain underexplored. In this work, we propose entropy-punctured Bloom Filters, a memory-aware encoding strategy that removes low-variability bit positions identified using empirical entropy. Starting from fixed-length BF encodings of quantized features, the proposed approach produces reduced representations that preserve predictive structure while improving predictive efficiency relative to encoded representation size. We evaluate the approach on diverse regression datasets, comparing raw features, Principal Component Analysis (PCA), Random Projection (RP), and Bloom Filter variants under leakage-free evaluation protocols and approximately matched representation sizes. Performance is assessed using ridge regression, XGBoost, and neural networks, with predictive efficiency measured as R2 relative to encoded representation size per sample. Results show that Bloom Filter encodings remain competitive with classical compressed representations while achieving substantial storage savings. Entropy-based puncturing further reduces representation size with minimal loss in predictive fidelity, yielding improved predictive efficiency. These findings demonstrate that entropy-punctured Bloom Filters provide an effective representation-level compression approach for memory-constrained machine learning. Comments: Accepted for publication in the Proceedings of the 38th IEEE International Conference on Tools with Artificial Intelligence (ICTAI 2026). 9 pages, 7 figures Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14187 [cs.LG] (or arXiv:2609.14187v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.14187 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: John Cartmell [view email] [v1] Sat, 12 Sep 2026 23:37:46 UTC (430 KB)
[AI-145] Signatures of Steerability in Activation Space of Language Models
链接: https://arxiv.org/abs/2609.14151
作者: Prajjwal Bhattarai,Tuka Alhanai
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Steering language models using a set of contrastive representations has been a canonical and computationally efficient method for controlling model behavior. Despite this success in controlling certain model behaviors, the effectiveness of activation steering varies markedly across concepts; the generalization properties of steering vectors are often considered a function of the dataset used to construct them. We make this dataset-dependence claim more rigorous and show that simple separation metrics strongly correlate with the downstream steerability of language models across diverse settings, even after controlling for layers and dataset effects. Beyond prediction, we provide evidence from a synthetic superposition experiment that separation metrics are strongly correlated with alignment between the empirical and true feature direction. Our results suggest that simple separability statistics can serve as practical diagnostics for when steering vectors are likely to work.
[AI-146] A New Transformer-Based Approach for Audio-Based Kinship Verification and a New Uncontrolled Mandarin Kinship Speech Dataset
链接: https://arxiv.org/abs/2609.14145
作者: Qiyang Sun,Langqing Zhang,Yupei Li,Björn Schuller
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
备注: 7 pages, 4 figures. Accepted to IEEE Spoken Language Technology Workshop 2026
Abstract:Kinship verification is a task involving determining whether two individuals share a first-order kin relation. To tackle this task, we propose CONVTRAP-TN, a new architecture for audio-based kinship verification, and conduct an ablation study on the proposed model. To the best of our knowledge, we are the first to apply the successful transformer architecture to the task of audio-based kinship verification. Furthermore, we also collect a custom speech dataset, ARKIN, which accurately reflects everyday recording conditions. We do this because only a few speech datasets with kinship labels currently exist, all of which either source extremely noisy in-the-wild data from the internet, or instruct speakers to record in specific environments. These settings fail to reflect real-world scenarios where users record on personal devices under unrestrained conditions. Additionally, we perform a series of preliminary baseline experiments on the collected dataset, including speaker verification and recognition, speech recognition, age estimation, and kinship verification, as well as cross-dataset kinship verification experiments to show that existing methods are not robust across datasets.
[AI-147] LIMBO: Lifelong Inference-Time Memory and Budget Optimization for LLM Agents ICTAI’26
链接: https://arxiv.org/abs/2609.14138
作者: Siddharth Sharma,Nilesh Prasad Pandey,Onat Gungor,Tajana Rosing
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted by the 38th IEEE International Conference on Tools with Artificial Intelligence (ICTAI’26)
Abstract:As LLM agents become integrated into increasingly complex workflows, they must continually acquire new capabilities while retaining competence on previously learned tasks. Lifelong agents address this through experience replay, injecting past interactions into the prompt to leverage prior experience during inference. However, replay is not free: every replayed trajectory competes with retrieval, reasoning, tool use, and verification for the same limited prompt and compute budget, making effective resource allocation essential. Existing approaches allocate these resources using fixed replay policies, regardless of whether replay is beneficial for the current task. We identify this as inference-time memory allocation, a distinct problem class for lifelong agents, and introduce LIMBO: the first online framework to our knowledge that treats memory as a controllable inference-time resource and jointly optimizes memory strategy and inference budget for each incoming task. Unlike prior approaches that fix the replay policy or require model weights, teacher supervision, or offline retraining, LIMBO learns this allocation online in a single pass, explicitly balancing task performance and inference cost without modifying the underlying agent. Across three LLM backbones on LifelongAgentBench, LIMBO achieves better cost-accuracy tradeoffs than state-of-the-art memory-augmented baselines and nearly matches all strongest such baselines at up to ~83% lower inference cost (~53% on average). LIMBO adapts its policy across models and environments without retraining, demonstrating that effective allocation can be learned online rather than manually specified.
[AI-148] o do(x) or not to do(x): Medical Image Counterfactuals for Dataset Augmentation
链接: https://arxiv.org/abs/2609.14124
作者: Yasin Ibrahim,Robin J. Evans,Konstantinos Kamnitsas
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Medical image analysis is often hindered by biased datasets, which can lead to biased models and limited clinical applicability. A promising strategy for mitigating such biases is to augment training data with synthetic images. Counterfactual (CF) generation is one such strategy, though the term is used in two different senses: in some works, CFs are produced through causality-based interventions derived from structural causal models, whereas in others, they are produced by non-causal image edits or conventional conditional generative models, such as altering anatomy or adding pathologies. In this work, we study this distinction and evaluate its practical consequences for medical image augmentation. We compare three conditioning strategies: \textitDeterministic , which changes selected variables while holding the remaining variables fixed; \textitUndirected , which updates variables according to learned statistical associations without assigning causal directions; and \textitCausal , which propagates interventions along a directed causal graph. We analyse how these choices affect the resulting images, and explore when causally grounded methods improve dataset augmentation or bring limited benefit. In particular, we assess downstream performance and fairness, where fairness refers to reduced sensitivity to dataset biases across sensitive subgroups. Our experiments demonstrate that using a causal approach to synthetic training data generation can lead to tangible benefits, with these insights offering valuable guidance to machine learning practitioners for the effective design of data generation protocols.
[AI-149] Real-Time Synthesis of Robust Controlled Invariant Sets for Monotone Systems
链接: https://arxiv.org/abs/2609.14115
作者: Yasin Sonmez,Mahmoud Khaled,Majid Zamani,Murat Arcak
类目: ystems and Control (eess.SY); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:Safety-critical control of autonomous systems requires formal safety certificates, such as controlled invariant sets, that must be computed online as conditions change. Although standard synthesis algorithms scale poorly with state dimension, monotone dynamical systems with lower-closed safety specifications allow for accelerated computation of controlled invariant sets. In particular, lazy fixed-point algorithms exploit monotonicity and track only the antichain basis of the set. However, membership tests and redundancy checks against an evolving basis remain major bottlenecks. We introduce a threshold-function reformulation in which a lower-closed set on a d-dimensional grid is represented by its column heights along a designated axis. This reformulates the greatest-fixed-point iteration as independent one-dimensional binary searches, one per grid column, yielding an embarrassingly parallel iteration with asymptotically lower computational complexity than the lazy fixed-point algorithm. Experiments synthesize invariant sets on 3D grids with 10^9 cells in under 50 ms and 10^14 cells in under two minutes. We further demonstrate online re-synthesis in a safety-informed model predictive controller example.
[AI-150] AGENT Q: Quantization-Conditioned Backdoor Attacks on LLM Agents EMNLP2026
链接: https://arxiv.org/abs/2609.14060
作者: Xiaoqun Liu,Qiben Yan
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Appear in EMNLP 2026
Abstract:Quantization is one of the default deployment paths for open-weight LLM agents, but it is not behavior-preserving: an adversary can release a full-precision checkpoint that passes audits yet misbehaves once quantized, termed as quantization-conditioned attack (QCA). Prior QCA work targets free-text generation, where harm is mediated by a human reader. In contrast, the agentic setting poses a more severe risk: the triggered payload is a structured function that can be executed without human oversight. We present the first study of QCA against LLM agents. We find that directly adapting prior backdoor-injection methods can produce malicious behavior after quantization, but substantially degrades benign utility, rendering the resulting attacks impractical. To understand the true upper bound of the threat, we propose AGENTQ, an attack framework that combines layer-banded LoRA injection with partial-PGD repair over a multi-codebook quantization-equivalence class. AGENTQ preserves normal agentic capability while concentrating malicious behavior in the quantized model. Across three trigger-action pairs and three codebooks (NF4, FP4, INT8), AGENTQ reaches up to 100% post-quantization attack success rate with minimal loss of benign utility, underscoring the need to make quantization-aware safety evaluation a standard requirement before open-weight agents are deployed.
[AI-151] Schizophrenia Detection from EEG Signals: A Transformer Framework with Spectrogram Representation
链接: https://arxiv.org/abs/2609.14015
作者: Abtin Shafiei,Mohsen Hooshmand,Majid Ramezani
类目: Artificial Intelligence (cs.AI)
备注: 6 Pages, 5 figures, Conference
Abstract:Schizophrenia is a serious psychiatric disorder that affects millions of people worldwide, and its diagnosis remains primarily dependent on clinical assessment. Electroencephalography (EEG) provides a non-invasive approach to investigate brain activity and has shown potential to support automated Schizophrenia detection. However, existing EEG-based classification studies often suffer from limitations including small datasets, inconsistent preprocessing strategies, and evaluation protocols that may not adequately prevent subject-related data leakage. In this study, we propose an EEG-based Schizophrenia classification framework that transforms preprocessed EEG recordings into time-frequency representations using the Short-Time Fourier Transform. The generated spectrogram images are classified using both conventional Machine Learning algorithms, including Support Vector Machines, Random Forests, and XGBoost, and Deep Learning models, including convolutional architectures and CNN-Transformer hybrids. To ensure reliable evaluation, all data partitions are performed at the subject level. Experimental results demonstrate that the proposed approach achieves competitive classification performance, with the CNN-Transformer (CT-SZ) model achieving an AUC-ROC of 88.41% and the CNN + Squeeze and Excitation + Transformer (CST-SZ) achieving an AUC-ROC of 92.88% on the independent test set.
[AI-152] Convergent Emergence of In-Context Learning Across Modalities
链接: https://arxiv.org/abs/2609.14011
作者: Nathan Breslow,Seungwook Han,Daniel Hyunsoo Lee,Aayush Mishra,Anqi Liu,Daniel Khashabi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Few-shot in-context learning (ICL), the capacity of a model to infer abstract patterns from input-output examples provided in its prompt and apply them to new inputs, has been extensively studied in large language models trained for next-token prediction on human text. Recently, few-shot ICL has been demonstrated in autoregressive genomic models as well. This raises a question: does ICL emerge broadly across domains, and if so, what common structure is shared? To address both, we develop a controlled cross-modality framework that instantiates the same task suite in a variety of modalities to test what we call the Convergent Emergence Hypothesis: the idea that few-shot ICL, when it emerges, shares a common cross-modality difficulty profile - i.e., tasks that benefit from ICL in one modality tend to benefit in others. We show that paired-mapping ICL emerges across six modalities (language, genome, integer sequences, time series, images, and proteins), surpasses controlled baselines, and has correlated per-task effects across five of them. Together, these results provide support for the Convergent Emergence Hypothesis in some modalities, but not all. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.14011 [cs.AI] (or arXiv:2609.14011v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.14011 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-153] Confuse the Model Control the Flow: Understanding and Mitigating Privacy Leakage from LLM Agents with Information Flow Control
链接: https://arxiv.org/abs/2609.14003
作者: Minsun Shim,Ramisha Raida Karim,Ruthwik Jakkula,Kaiwen Zhou,Xin Liu,Xin Eric Wang,Zhou Li
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Personal AI agents built on large language models (LLMs) are increasingly given access to a user’s private data and communications in order to provide personalized assistance. This access creates a persistent privacy risk: the agent must decide whether a given sensitive information should be disclosed to a particular party. Existing defenses address this by making the agent’s backend LLM more privacy-preserving through stronger system prompts, training, or explicit consent-checking procedures, but this approach has a structural challenge: whenever enforcement is a judgment the LLM makes over the same conversational context an adversary controls, the enforcement mechanism and the attack surface coincide. We demonstrate this against existing defenses with three new attacks that require only ordinary agent interaction and no prompt injection: Collaborative Workspace Lure reframes an extraction attempt as collaborative work; Semantic Obfuscation Attack induces disclosure through omission rather than through anything the agent writes; and Channel Decoupling Attack splits the extraction request and the disclosure across independent channels. All three achieve substantially higher leak rates than the attacks these defenses were originally designed to withstand. Guided by this observation, we present FLOWSEAL, a defense that enforces confidentiality through a tool-level interceptor outside the LLM’s context, grounded in data provenance and an information-flow-control lattice with controlled declassification. Evaluated across three benchmarks, five prompt-based baselines, and eight attacks, including a real agent executing live tool calls through MCP, FLOWSEAL reduces leak rates to near zero (e.g., 52.2% to 0.5% against Collaborative Workspace Lure) while preserving task utility, regardless of the underlying LLM backend.
[AI-154] Synthetic Data in Marketing Research: How to Evaluate and When to Trust
链接: https://arxiv.org/abs/2609.13995
作者: Oded Netzer,Rajan Sambandam
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Debate over synthetic data in marketing research has polarized between claims that large language models (LLMs) make human respondents obsolete and calls to avoid them entirely. We argue that both positions obscure the more useful question: not whether synthetic respondents work, but when. Building on Brand, Israeli, and Ngwe (2026), we make three contributions. First, we distinguish three types of synthetic data (ungrounded LLM responses, segment-level personas, and individual-level digital twins) and map each to the decisions it can support. Second, we develop a taxonomy of four families of accuracy measures and suggest that the wide range of reported twin accuracy, from near-perfect to near-chance, largely reflects differences in what is being measured rather than in method quality. Aggregate measures often perform well even when little information is supplied to the LLM, and can mask a complete absence of respondent-level differentiation. Third, we introduce the forgotten question problem, in which a question is omitted from a fielded study, as a setting for twin-based augmentation of existing data. We propose an ex-ante answerability diagnostic that requires no ground truth: the R^2 of a random forest predicting twin outputs from the data used to construct the twins. Across 108 attitude questions from a nationally representative survey (N = 3,063), screening at R^2 above 0.7 raises the mean twin-human individual-level correlation by 15% and reduces the share of poorly answered questions from 25.9% to 4.3%. Embedding similarity and experienced-researcher judgment provide correlated but weaker screens.
[AI-155] Hardware-Aware Learned Representation Compression for Distributed In-Sensor Vision
链接: https://arxiv.org/abs/2609.13947
作者: Chengwei Zhou,Abu Masum,Xuming Chen,Mehran Moghadam,Sreetama Sarkar,Arnab Sanyal,Md Abdullah-Al Kaiser,M. Hassan Najafi,Sercan Aygun,Gourav Datta
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
备注: Under submission at IEEE Transactions on Emerging Topics in Computing
Abstract:In-sensor computing reduces the cost of transmitting high-resolution image data by performing early-stage processing near the sensor. However, the logic chip integrated with a CMOS image sensor (CIS) is tightly constrained in compute and memory, limiting conventional deep neural network partitioning. We present OASIS, a distributed in-sensor vision framework that uses a lightweight encoder to generate compact, task-relevant representations before off-chip transmission. The encoder is trained end-to-end using task, entropy, and reconstruction objectives, while the decoder is used only during training. OASIS supports two complementary deployment paths. The first applies 4-bit quantization and Huffman coding while preserving the spatial structure required by classification and dense-prediction tasks. The second uses Sobol-based hyperdimensional computing (HDC) to transform the encoder latent into a fixed-dimensional binary hypervector for associative-memory classification. For the SwinViT-based VWW model, mapping a 3\times3\times8 latent to a 64-dimensional hypervector provides an additional 1.77\times communication reduction with less than one percentage point of accuracy loss relative to the 128-dimensional configuration, yielding an overall 18,816\times reduction compared with raw 8-bit image transmission. We implement the digital near-sensor pipeline on an AMD Xilinx Zynq UltraScale+ FPGA and characterize it using direct board-level power measurements and Vivado post-implementation analysis, together with circuit-simulated CIS models and a 7-nm ASIC projection. Across visual wake-word classification, hand tracking, and eye tracking, OASIS reduces total system energy by approximately 2\times - 4.5\times while maintaining competitive accuracy, demonstrating a practical hardware-algorithm co-design path for communication-efficient in-sensor vision.
[AI-156] SAILOR: Solver-Assisted Interactive LLM -based Optimization Recovery
链接: https://arxiv.org/abs/2609.13945
作者: Shaghayegh Sadeghi,Stephen L. Smith,David C. Del Rey Fern’andez
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Natural-language descriptions of optimization problems may be incomplete or vague about numerical information that a solver requires, including costs, capacities, demands, bounds, and penalties. A language model can translate the description into code, but when a required value is absent it must either stop or guess. We present SAILOR, a proof-of-concept system that detects such unsupported numerical choices, asks the user targeted follow-up questions, and updates the optimization model before returning a solution. Questions are prioritized using uncertainty and solver-derived estimates of how strongly each missing value affects the current model. We evaluate the pipeline on 1,723 instances from seven masked benchmarks using an idealized simulator that returns ground-truth values. Exact objective-value agreement ranges from 27.0% to 87.6% across datasets, with 1.4–5.7 questions per instance on average. These results establish feasibility under controlled branch-and-reveal feedback; they do not measure performance with human users or general structural model repair. Code is available at: this https URL.
[AI-157] Finite-Time Node Separation in Recurrent Graph Neural Networks with Persistent Gaussian Perturbations
链接: https://arxiv.org/abs/2609.13920
作者: Mostafa Haghir Chehreghani
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Persistent Gaussian perturbations have been shown to prevent asymptotic oversmoothing in recurrent Graph Neural Networks (GNNs) by ensuring a positive stationary Dirichlet energy. However, this global energy bound does not guarantee that individual node representations remain distinct at finite depths. In this paper, we provide a complementary finite-time analysis of the same persistent-noise architecture. Let (d) denote the representation dimension and (\sigma) the noise standard deviation. We first prove an exact second-moment decomposition for the expected squared distance between any two node representations, yielding the universal lower bound (2\sigma^2 d) at every positive time step without contraction or stationarity assumptions. More precisely, conditional pairwise distances have a noncentral chi-square characterization: the noncentrality parameter is the deterministic message-passing separation normalized by (2\sigma^2). This yields dynamics-aware fixed-time and finite-horizon near-collision bounds that retain information discarded by the central worst-case analysis. The earlier central Gaussian bound is recovered as the worst-case zero-separation case. We additionally prove almost-sure pairwise noncollision, derive a uniform finite-horizon guarantee, and establish permutation equivariance in distribution for the stochastic dynamics and permutation-invariant graph outputs. Our results complement the asymptotic energy analysis of prior work and provide rigorous finite-time guarantees on node-level representation separation.
[AI-158] LoRA Fine-Tuned Models for Control Systems Course QA: A Multidimensional Evaluation of Model Scale and Rank Effects
链接: https://arxiv.org/abs/2609.13918
作者: Shaowen Lu,Chengxu Liu,Ping Zhou,Tao Yang
类目: Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:Large language models (LLMs) are increasingly used in specialized university courses, but control-systems questions require coordinated terminology, notation, derivations, and stepwise explanations. Direct general-purpose responses may be inconsistently structured and hard to verify. Using exercises and reference solutions from a Linear Control Systems course, we built a supervised fine-tuning dataset of 360 system-user-assistant conversations. We applied LoRA to Qwen2.5-3B-Instruct and Qwen2.5-7B-Instruct. With identical data splits, inference settings, and evaluation protocols, we compared base and fine-tuned models and tested LoRA ranks r=4, 8, and 16. Evaluation used ROUGE, BERTScore, and structured-output features to measure reference-answer similarity and stability of the Solution-Method-Teaching Points format. LoRA improved both similarity and structured-output stability at both sizes. On the current test set, 7B-r16 achieved the highest ROUGE-L (0.4093) and BERTScore-F1 (0.8643), while r=8 offered a better balance between performance and parameter efficiency. Bootstrap resampling showed ROUGE-L gains of 0.0764 [0.0613, 0.0915] for 3B-r16 and 0.0874 [0.0687, 0.1042] for 7B-r16; both intervals exceeded zero, indicating stable textual-similarity improvements on the current test set. These results suggest LoRA can align open-source instruction-tuned models more closely with the language and pedagogical organization of course reference answers. However, the metrics mainly capture textual similarity and formatting consistency, not domain-specific reasoning or mathematical correctness, which require expert assessment and task-specific rubrics.
[AI-159] DiTAR: Dual Optimization for Robust Autoregressive Diffusion Speech Synthesis
链接: https://arxiv.org/abs/2609.13909
作者: Ziyu Zhang,Tianlun Zuo,Hanzhao Li,Haoyu Zhang,Lei Xie
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注:
Abstract:Continuous-latent Autoregressive Diffusion Transformer (AR-DiT) models have demonstrated immense potential in zero-shot speech generation. However, they still suffer from limited decoding stability when synthesizing long utterances or complex linguistic structures. This instability primarily stems from a restricted historical receptive field and an acoustic inertia dependency within the diffusion decoder, which causes the model to ignore semantic conditions. To address these challenges, we propose DiTAR+, a dual-optimization framework. First, we introduce Dilated Context Sampling to expand the macro-level historical receptive field without violating physical temporal continuity, thereby preventing cumulative error propagation. Second, we propose Hierarchical Acoustic Masking to prevent shallow layers from attending to acoustic pre-context, explicitly decoupling semantic alignment from acoustic detail reconstruction. Extensive experiments show that our framework effectively mitigates pronunciation errors and semantic hallucinations, enhances generation robustness on challenging sentences, and maintains exceptionally high speaker similarity throughout the entirety of long-form utterances. On the linguistically challenging ZH-Hard set, DiTAR+ reduces the word error rate from 12.478% to 9.893%, and on extended utterances of 25 to 35 seconds it improves speaker similarity from 0.741 to 0.759 while simultaneously lowering the word error rate from 2.778% to 2.173%, outperforming both discrete-token and pure flow-matching baselines.
[AI-160] When Malicious Instructions Persist: Persistent Memory Poisoning Attack on Harness-Based Agents
链接: https://arxiv.org/abs/2609.13889
作者: Shuhuai Huang,Jingfeng Zhang,Hong Jia
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 14 pages, 1 figures. Code: this https URL
Abstract:Harness design has transformed the development of LLM-based agents by integrating memory, tool use, and runtime control. However, this design also introduces security and privacy risks because malicious instructions from external sources may be written into persistent memory and persist across sessions. To study this risk, we propose PMPA, a Persistent Memory Poisoning Attack against harness-based agents. PMPA embeds malicious instructions into benign external sources and induces the victim agent to write them into persistent memory without directly accessing to the agent framework. Once stored, the poisoned memory can be retrieved in later sessions, triggering additional malicious actions and causing privacy leakage. We evaluate PMPA on OpenClaw and Claude Code across different backbone LLMs, input modalities, and trigger scenarios. Across all settings, PMPA achieves average Injection Success Rate (ISR) and Cross-session Attack Success Rate (C-ASR) of 73.7%/ 55.5% on OpenClaw and 66.9%/ 81.7% on Claude Code, while preserving benign task performance on both systems. We further evaluate a targeted prompt-level defense and find that it can reduce memory injection in many settings, but provides limited protection once the persistent memory has been poisoned.
[AI-161] Map Users and Mapmakers: The Scope of Cognitive Attribution from Acquired Representations
链接: https://arxiv.org/abs/2609.13879
作者: Yiling Wu
类目: Artificial Intelligence (cs.AI)
备注: 22 pages
Abstract:An acquired representation can enlarge a system’s cognitive repertoire without transferring the capacities exercised in producing that representation. This paper develops a framework for specifying that enlargement and its limits. Its central contribution is a five-part attribution table distinguishing effective tracking, application of acquired structures, acquisition from explicit specifications, acquisition from identifying observations, and retention and reuse. Each entry identifies a positive capacity commitment and a further claim requiring additional support. The argument deliberately grants meaningful content, causal efficacy, and productive inference, so that its conclusion does not depend on treating representations as inert encodings. Map and category examples show why even complete application competence leaves acquisition capacity undetermined, and why acquiring a criterion from its description differs from finding it in examples. Short formal proofs appear in an appendix. The framework is applied to Andrew Ng’s world-model interpretation of Othello-GPT and to the specific indicators discussed in contemporary accounts of machine concepts. It preserves demonstrated recognition, classification, inference, and qualified acquisition while specifying what remains unestablished about criterion discovery and accumulation. The result concerns the scope of cognitive attribution rather than the constitutive conditions of concept possession: cognitive achievements deserve credit for the capacities they establish, without silently importing a broader repertoire through the labels attached to them.
[AI-162] rustworthy Explainable and Sustainable Decentralized Intelligence for 6G Networks
链接: https://arxiv.org/abs/2609.13872
作者: Giovanni Perin,Michele Rossi,Enrique Tomás Martínez Beltrán,Fernando Torres-Vega,José María Jorquera Valero,Manuel Gil Pérez,Eunjeong Jeong,Nikolaos Pappas,Farah Abed Zadeh,Chamara Sandeepa,Bartlomiej Siniarski,Madhusanka Liyanage,Betül Güvenç Paltun,Leyli Karaçay,Ioannis Pitsiorlas,Marios Kountouris
类目: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI); Signal Processing (eess.SP)
备注: Sumitted to IEEE Signal Processing Magazine for possible publication
Abstract:As 6G networks transition from theoretical frameworks into operational realities, artificial intelligence (AI) evolves from an add-on optimization tool into a distributed and interconnected structural layer. Unlike previous network generations that mostly relied on centralized cloud analytics platforms, AI-native 6G networks operate across a dynamic, multi-domain edge-cloud continuum where data originates from heterogeneous sources including user devices, radio access networks, sensing infrastructures, and vertical applications. Centralizing this massive volume of data creates severe communication overhead, unacceptable latency bottlenecks, single points of failure, and complex cross-domain governance challenges. Consequently, decentralization becomes a fundamental architectural requirement for future 6G network intelligence and zero-touch operations. Security serves as the primary enabler of this decentralized paradigm. Critical security functions, such as real-time threat detection, physical-layer attack mitigation, slice protection, and intrusion detection, require immediate access to local context and telemetry before operational data loses its value. However, moving intelligence to the edge via collaborative paradigms like federated learning (FL) and decentralized FL (DFL) introduces complex trade-offs. System security cannot be addressed in isolation; it is deeply intertwined with equally important aspects like trustworthiness, explainability, and energy sustainability. Taking these aspects into account, this paper develops a unified perspective on decentralized intelligence for 6G, arguing that decentralization, trustworthiness, explainability, and sustainability must be designed jointly rather than treated as independent requirements
[AI-163] ReWeight: Leverag ing Human Data for VLA Post-Training via Demonstration Retrieval and Sample Weighting
链接: https://arxiv.org/abs/2609.13851
作者: Chenwei Wang,Dianye Huang,Match W.L. Ko,Chenjia Bai,Zhongliang Jiang
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Post-training vision-language-action (VLA) models for specific robots and tasks requires in-domain demonstrations, yet collecting diverse robot data is costly. Egocentric human demonstrations provide a scalable alternative, but directly mixing human and robot data can introduce cross-embodiment discrepancies and degrade policy performance. To address this challenge, we introduce ReWeight, a framework that incorporates human data into VLA post-training through demonstration-level retrieval and sample-level weighting. ReWeight learns a cross-embodiment visuomotor representation that combines visual observations with future actions to measure behavioral similarity between human and robot demonstrations. Based on optimal transport, it retrieves human demonstrations relevant to the target robot data and assigns larger weights to samples with smaller cross-embodiment discrepancies. We evaluate ReWeight using \pi_0.5 across eight simulation tasks and four real-world tasks under both clean and randomized settings. In simulation, ReWeight improves the average success rate of post-trained \pi_0.5 from 39% with only robot data and 44% with randomly mixed human-robot data to 57%. In the physical experimental setting, it achieves an average success rate of 68.8%, outperforming the baselines by 28.8% and 13.8%, respectively. Overall, ReWeight provides an effective paradigm for transforming abundant egocentric human experience into transferable supervision for robot learning. (Project webpage: this https URL)
[AI-164] UniCAR-RL: Seeing Better before Thinking Deeper in Visual Mathematics EMNLP2026
链接: https://arxiv.org/abs/2609.13849
作者: Yuzhe Li,Hao Yan,Hao Wang,Xingchen Liu,Ya-Qi Yu,Jihao Wu,Minghui Liao,Wei Chen,Yuliang Liu
类目: Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Findings
Abstract:Multimodal Large Language Models (MLLMs) often struggle with complex mathematical visual reasoning primarily due to a lack of fine-grained perception, causing initial visual hallucinations to directly trigger cascading reasoning failures. In traditional end-to-end reinforcement learning (RL), sparse rewards fail to decouple perceptual hallucinations from logical missteps, hindering targeted perception optimization. Alternatively, fine-tuning with perception-enhanced CoT data incurs high costs and hallucinations. In this paper, we address these challenges by proposing UniCAR-RL, an annotation-free RL framework. By explicitly decoupling the optimization of perception and reasoning during the training process, it achieves isolation and optimization of both capabilities. Specifically, UniCAR-RL consists of three synergistic branches: 1) a Caption-RL branch that optimizes perception capabilities through verifier-guided reasoning validation; 2) a Reasoning-RL branch that performs logical reasoning based on a gold image description to halt cascading errors; 3) a QA-RL branch that retains native end-to-end alignment to ensure robust question-answering performance. Experiments show that UniCAR-RL substantially improves MLLMs’ mathematical and visual reasoning using only raw short-answer data. Furthermore, it demonstrates strong generalization across diverse architectures and scales.
[AI-165] LePlanner: An Iterative Amortized Controller For World Models
链接: https://arxiv.org/abs/2609.13845
作者: Saksham Bansal,Om Naphade,Chayan Aggarwal,Vrishin M
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 23 pages, 16 figures
Abstract:World models trained with joint-embedding predictive architectures learn compact, structured latent representations from physical interaction, yet planning in these latent spaces typically relies on one of two costly approaches. Search-based planners such as CEM, MPPI, and iCEM optimize action sequences through many predictor rollouts, achieving strong performance at the cost of high per-decision compute and latency. Policy-based methods amortize inference into a single forward pass but can degrade on contact-rich tasks where the demonstration distribution is multimodal. We propose LePlanner, an amortized iterative controller that learns to construct and refine latent action sequences through a frozen world-model predictor. LePlanner is trained with an arrival-and-hold objective that encourages the controller to reach the goal at the earliest feasible horizon and remain there. This addresses horizon-reset procrastination, a failure mode in which repeated receding-horizon replanning continually postpones goal arrival. An additional action-Gaussian loss keeps generated actions near the support of the offline dataset. Across navigation, contact-rich manipulation, and continuous-control environments, LePlanner matches or exceeds search-based planners while requiring an order of magnitude fewer predictor evaluations and 3-49x lower wall-clock time per decision. It achieves success rates of 98% on PushT, 100% on Reacher, 100% on TwoRooms, and 92% on the OGBench Cube task. These results show that much of the structure discovered through online search can be amortized into a lightweight iterative policy, enabling fast, horizon-aware, nonlinear physical control without online optimization.
[AI-166] CRAF: Cross-View Residual-Aware Fusion for Deepfake Speech Detection
链接: https://arxiv.org/abs/2609.13842
作者: Minh-Xuan Phan,Khalid Zaman,Candy Olivia Mawalim,Masashi Unoki
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注:
Abstract:Recent advances in speech synthesis and voice conversion have made deepfake speech increasingly realistic, making generalization to unseen spoofing attacks a critical challenge. Pretrained speech and audio models offer a promising direction for improving robustness to such unseen attacks. Self-supervised learning (SSL) models capture fine-grained, low-level acoustic characteristics, whereas Auditory Large Language Models (ALLMs) provide higher-level contextual representations. These complementary views can provide useful cues for improving generalization to unseen attacks. However, direct fusion does not explicitly disentangle information shared across the two views from view-specific complementary information, limiting effective cross-view integration. To address this, we propose CRAF, a cross-view residual-aware fusion framework that uses ALLM-guided cross-view attention to enrich SSL representations and adopts ALLM as a high-level reference to separate ALLM-explainable information from complementary SSL residual information. The residual is selectively refined through adaptive gating and integrated through SSL-primary fusion. Experiments on ASVspoof 5 show that CRAF with Kimi-Audio achieves an EER of 5.96% and a minDCF of 0.1192, demonstrating robustness to unseen spoofing attacks.
[AI-167] ViperQ: Order Flow Pattern Recognition via Auction Market Theory for Reinforcement Learning Trading
链接: https://arxiv.org/abs/2609.13825
作者: Asser Moustafa,Rares-Mihail Neagu,Jugal Kalita
类目: Artificial Intelligence (cs.AI); Trading and Market Microstructure (q-fin.TR)
备注:
Abstract:Reinforcement learning trading systems published in the academic literature overwhelmingly rely on price-aggregate state representations (OHLCV bars) or limit-order-book depth features, leaving microstructure pattern theories from the practitioner literature, namely Auction Market Theory and Market Profile, without a peer-reviewed computational instantiation. We present ViperQ, a reinforcement learning system whose state representation is built explicitly from Auction Market Theory primitives: Volume Point of Control, Value Area position, Low Volume Node flags, Cumulative Volume Delta divergence, and tape-velocity signatures, assembled into a 20-dimensional Z-normalised vector. Two Proximal Policy Optimisation agents are trained with a prospect theory-grounded asymmetric reward function that penalises losing holds at a magnitude consistent with Kahneman and Tversky’s loss-aversion coefficient. Evaluated on a held-out twelve-month partition of institutional tick data the agents have never seen, ViperQ achieves +163.6% ROI on TSLA (-27.5% max drawdown, 27,019 trades) and +116.5% ROI on NVDA (-47.8% max drawdown, 12,892 trades) under zero leverage. The results establish Auction Market Theory features as a tractable structured input modality for sequential decision-making on financial time series and motivate further work on microstructure-aware policy learning.
[AI-168] Exploring Automated Vulnerability Identification in JavaScript Code Using Large Language Models
链接: https://arxiv.org/abs/2609.13816
作者: Manit Kaushik,Ishir Bhardwaj,Pranav Gupta,Pankaj Jalote,Arun Balaji Buduru
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 8 pages, 1 figure, 4 tables
Abstract:JavaScript powers approximately 98.8% of all websites, making vulnerabilities in its code a significant security risk, yet existing detection approaches such as Static Application Security Testing (SAST) tools often fail to identify many real-world vulnerabilities when applied to isolated code snippets. This paper presents an empirical study of Large Language Model (LLM)-based vulnerability identification for JavaScript programs, evaluating three LLM families (Gemini 1.5 Flash, GPT-4o Mini, DeepSeek-R1-Distill-Llama-8B) across multiple prompting strategies (zero-shot, chain-of-thought, few-shot) and fine-tuning approaches on a dataset of 1,125 JavaScript code snippets spanning five Common Weakness Enumeration (CWE) categories: Injection (CWE-74), OS Command Injection (CWE-78), Cross-Site Scripting (CWE-79), SQL Injection (CWE-89), and Uncontrolled Resource Consumption (CWE-400). Our experiments show that LLMs substantially outperform traditional SAST tools on snippet-level vulnerability identification, with a fine-tuned Gemini 1.5 Flash model achieving 60% detection accuracy compared to near-zero performance from rule-based analyzers. We find that fine-tuning improves accuracy from 29% to 60%, Chain-of-Thought prompting benefits reasoning-capable models such as GPT-4o Mini, few-shot prompting is effective for polymorphic vulnerabilities such as Cross-Site Scripting, and performance varies across vulnerability categories, reaching up to 84% accuracy for structured vulnerabilities such as SQL Injection. These results indicate that LLMs provide a practical approach for automated vulnerability identification in JavaScript code, particularly when combined with task-aligned supervision, though they should complement rather than replace existing security analysis workflows due to limited recall and uneven performance across vulnerability types.
[AI-169] Bypass Observation: A Conceptual Design of a Non-Intrusive Layer-Wise Semantic Extraction Architecture
链接: https://arxiv.org/abs/2609.13807
作者: Haibin Tong,Jiang Yu
类目: Artificial Intelligence (cs.AI)
备注: 14 pages, 1 figure, 2 tables
Abstract:Large language models reason in high-dimensional hidden-state spaces, while users observe only final outputs. We introduce Bypass Observation, a non-intrusive layer-wise readout architecture that attaches read-only observation heads to selected Transformer layers without feeding their outputs back into the backbone. We consider three variants: a shared LM head across layers, layer-specific heads, and a layer- or step-adaptive head. For full-vocabulary readout, we derive a closed-form overhead approximation governed primarily by V/(12d), with representative estimates ranging from about 30% to 240%, and discuss cost reductions via sparse observation, low-rank factorization, reduced vocabularies, top-k readout, and selective positions. We argue that Bypass Observation can make model computation more observable while remaining only a partial, potentially misleading projection of hidden states. We further distinguish bypass chain-of-thought from conventional chain-of-thought: conventional reasoning tokens enter the autoregressive computation, whereas bypass readouts remain causally external at inference time, although they can still provide training signals in reinforcement learning. Finally, we discuss applications to looped and recurrent-depth Transformers, where iteration-wise readout may expose convergence, oscillation, and potential halting signals. The proposal is conceptual and analytical; systematic empirical validation remains future work.
[AI-170] LLM -Enhanced Multi-Agent Reinforcement Learning for Unified Electric Vehicles-Charging Station-Grid Optimization in Public Charging Systems
链接: https://arxiv.org/abs/2609.13805
作者: Yang Zhang,Lindong Xie,Chongyu Wang,Gaojunjie Li,Siqi Bu,Edward Chung
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:In the era of the Internet of Things (IoT), coordinating connected electric vehicle (EV) charging scheduling to balance EV charging satisfaction, station profitability, and smart grid stability presents a complex multi-objective challenge. Existing Multi-Agent Reinforcement Learning (MARL) approaches often struggle with high-dimensional state spaces generated by massive IoT sensing data and conflicting stakeholder interests. This paper proposes a novel LLM-enhanced MARL framework that, for the first time, simultaneously optimizes the Grid, EVs, and Stations within a unified loop. By integrating Large Language Model (LLM), we address two critical bottlenecks: interpretable feature selection and adaptive multi-objective balancing. The LLM analyzes real-time IoT-collected environmental states to extract physically significant features and dynamically assigns weights to conflicting objectives-including profit, user satisfaction, and grid load-using semantic reasoning instead of complex manual tuning. Extensive experiments demonstrate that our framework significantly outperforms state-of-the-art baselines, achieving superior market efficiency while reducing training time by over 70%. This approach offers a scalable, transparent solution for efficient and sustainable IoT-enabled urban charging infrastructure management.
[AI-171] Do Not Restart: Residual Completion for Stateful Agent Handoffs
链接: https://arxiv.org/abs/2609.13800
作者: Runzhi Deng,Yiming Zhong,Fang Zhao,Pan Zhou
类目: Artificial Intelligence (cs.AI)
备注: 11 pages, 2 figures, 4 tables
Abstract:Routing and cascades reduce tool-agent cost by transferring control across models, but stateful handoffs must preserve accepted choices, realized effects, and unfinished obligations. We formulate this as commitment-constrained residual completion and introduce Commitment-Frontier Residual Completion (CFRC). CFRC enforces target-before-proposal, whole-proposal-before-authority, and live-evidence-before-success: it freezes a residual contract from accepted progress, closes the successor continuation into an evidence-linked graph, and admits execution only when the remainder is covered, with live receipts discharging obligations. We establish contract-relative partial correctness, which extends to the original residual request under complete contract construction. Across five environments and two same-provider model pairs, CFRC achieves comparable macro accuracy to strong full-task agents at only 22.0%-34.6% of their inference cost, with additional cross-provider results demonstrating broader transfer.
[AI-172] Partition Scores Are Not System Scores: Deployment-Fidelity Gaps in Decomposed Algorithm Selection
链接: https://arxiv.org/abs/2609.13785
作者: Jiachen Zhang,Yu Tang,Li Zhu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 28 pages, 7 figures, 14 tables
Abstract:Oracle-style quantities, including virtual best solvers, selected-portfolio VBS, virtual-best encodings, and best-in-family summaries, are widely reported as upper bounds on what a deployable selector could achieve. In decomposed algorithm selection, an analogous partition-level score grants an oracle choice of the best algorithm within the selected family; once the family selector is fixed, the deployable system must replace that within-family oracle with a learned within-family selector. We define the deployment-fidelity gap G® as the difference between partition-level and deployable end-to-end utility and derive two accounting consequences: a per-instance margin-regret stability condition that tells us when a partition-time family choice is deployment-optimal, and a sharp partition-only identification interval that, when it strictly crosses zero, prevents the partition-level report from certifying the deployable winner. Across five public algorithm-selection benchmarks spanning tabular AutoML and combinatorial CSP/SAT, every decomposed pipeline has positive G®, ranging from 0.012 on TabZilla to 0.13 on PROTEUS-2014. Four of ten decomposed-versus-flat decisions have sign-changing point estimates; on PROTEUS-2014, a 33-point partition advantage shrinks to a 20-point end-to-end advantage. A training-side validation gap-correction diagnostic recovers the point-estimate deployable sign on all four sign-changing cells; it is a reporting aid, not a substitute for direct end-to-end evaluation. Partition and end-to-end scores should be reported side by side. Comments: 28 pages, 7 figures, 14 tables Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2609.13785 [cs.AI] (or arXiv:2609.13785v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.13785 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-173] Surprising Effectiveness of Self-Demonstrations in Enhancing Schema-Ontology Mapping with LLM s
链接: https://arxiv.org/abs/2609.13776
作者: Siddhesh Thombre,Manasi Patwardhan,Sunita Sarawagi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Integrating heterogeneous relational databases into a centralized ontology remains a persistent challenge in enterprise knowledge representation, primarily due to semantic heterogeneity, cryptic schema naming, missing metadata, and the abstraction gap between relational schemas and ontological models. Although large language models (LLMs) offer strong semantic reasoning capabilities, we show that directly applying them through one-shot prompting or naive multi-stage pipelines leads to poor performance for schema-ontology mapping. This paper presents a self-demonstration-driven approach that combines a neuro-symbolic task decomposition with a novel mechanism for automatically generating pattern-guided, dependency-aware demonstrations to address this integration challenge. Our approach incorporates two key strategies to achieve substantial accuracy gains over existing LLM-based schema integration methods: (i) a neuro-symbolic decomposition of the task into cascaded sub-tasks, where symbolic constraints structure the search space and LLMs perform semantic reasoning within each focused sub-task, and (ii) self-generated demonstrations guided by domain-agnostic patterns to supervise each sub-task. Experiments on three of the most challenging scenarios from the RODI benchmark show that our approach achieves state-of-the-art performance, substantially outperforming (25 percentage points F1 improvements) both traditional schema-to-ontology mapping techniques and recent LLM-based schema-to-ontology and schema matching approaches. Ablation studies further reveal the significant benefits of pattern-guided self-demonstrations and the complementary benefits of neuro-symbolic task decomposition.
[AI-174] Homeostatic Continual Learning
链接: https://arxiv.org/abs/2609.13771
作者: Yue Jin
类目: Artificial Intelligence (cs.AI)
备注: 18 pages, 12 figures, 1 table
Abstract:In this paper, I formulate a Continual Learning problem and propose a method named “Homeostatic Continual Learning” that enables an AI agent to learn continuously in a changing environment without catastrophic forgetting. The core of the method is to find outliers in the environment data when the agent experiences an outlier in its output. Through this method, the agent gradually completes its model and policy and performs well in more and more contexts. I also suggest that we may use the method to build a world model where the agent factorizes the objects in the world into features, abstract objects into comparable instances of concepts and map concepts to intents through features. I discuss the works needed to render the method practical, the connections to many fields in Artificial Intelligence and the broader implications of the method.
[AI-175] Positioning manuscripts in the scientific landscape with agent ic AI
链接: https://arxiv.org/abs/2609.13760
作者: Jiawen Chen,Zichen Zhang,Bingxuan Li,Quan Sun,Yiyan Zhang,Edric Tam,Jinjie Lin,Didong Li,Yun Li,Bingxin Zhao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Publishing a research manuscript is a routine yet demanding part of scientific life: time-consuming, stressful, and often uncertain in outcome. Recent advances in large language model (LLM)-based agentic AI have shown promise across a range of scientific tasks, and here we ask whether agentic AI can help researchers navigate the publication process itself by reliably inferring a manuscript’s eventual publication venue from its content and literature context. We introduce PASS (Publication-oriented Agentic Scientific System), an agentic system that understands manuscripts within their domain-specific literature context and predicts top-matched publication venues. PASS positions each manuscript within its surrounding literature landscape by reconstructing its local scientific neighborhood, tracing its topic trajectory, and reasoning over field-specific journal spaces. Evaluated on a leakage-audited benchmark of over 2,000 preprints across 16 biomedical fields, PASS achieved Top-1 accuracy of 50.3% and Top-5 accuracy of 86.1%, outperforming state-of-the-art LLM baselines and established journal-selection tools. PASS-produced quality scores, such as impact potential and novelty, aligned with independent measures of publication outcome. We also found that the designed literature retrieval module is the strongest performance contributor, particularly for positioning manuscripts relative to nearby work, and that PASS maintained near-full performance from the abstract alone, whereas LLM baselines required the full manuscript text. An independent human evaluation found strong researcher agreement with PASS’s manuscript understanding and recommendation rationale. PASS has been released as a public platform (this https URL) for broad researcher access.
[AI-176] How Many Thoughts Can a Vector Hold? The Capacity of Reasoning by Superposition
链接: https://arxiv.org/abs/2609.13747
作者: Hongyu Gu,Chang Liu,Jingwen Fu
类目: Artificial Intelligence (cs.AI)
备注: 20 pages,4 figures
Abstract:Large language models solve hard problems through intermediate computations across multi-step reasoning. Traditional chain-of-thought encodes these computations as tokens. Recent continuous and recurrent methods instead move partial computations into fixed-dimensional latent states, where a single thought can superpose multiple alternatives. This raises a fundamental design question:what should continuous thoughts preserve as reasoning proceeds? An intuitive approach discards past computations and keeps only the current reasoning frontier. Storing more items seems to dilute states and waste limited representational capacity. We show this intuition can be incorrect. Under identical downstream computations, cumulative superposition retaining full reasoning history can require lower representational dimensions than frontier-only superposition holding only current alternatives. At fixed hidden width, this advantage allows latent reasoners to retain more valid evidence, distinguish more plausible downstream outcomes, and delay the point where compressed states turn unreliable. This counter-intuitive effect emerges because informative historical components coherently reinforce each other, while unrelated alternatives bring random interference. This perspective also answers a practical design question: how should models weight memories accumulated inside latent states when their future use is unknown? Across reusable weighted superpositions, prioritizing a small set of recent or salient items produces weakly-represented memories that bottleneck subsequent attention. Uniform cumulative weighting avoids this flaw, and we prove it is minimax-optimal for robust future reasoning. Our results turn superposition from an observed latent-space effect into a design principle: balanced cumulative memory lets a fixed representational budget support more reliable, reusable computations.
[AI-177] rustworthy Agent ic AI: A Comprehensive Cybersecurity and Systems Survey on Threat Landscapes Defense Architectures and Open Challenges
链接: https://arxiv.org/abs/2609.13731
作者: Seyedakbar Mostafavi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The transition from passive foundation models to autonomous, goal-directed agentic AI systems has introduced unprecedented capabilities by coupling recursive cognitive reasoning loops, persistent memory architectures, live tool execution planes, and multi-agent collaboration topologies. However, granting probabilistic neural cores execution authority across filesystems, networks, and cloud infrastructure dissolves classical security perimeters: natural language simultaneously serves as input data, internal control code, and communication protocols, exposing a Turing-complete blast radius where untrusted data represents executable instructions. This survey delivers a comprehensive systems-security reference framework for trustworthy agentic AI, synthesizing 206 foundational studies and regulatory standards. We formalize the general agent architecture as a stateful 5-tuple and establish a 6-dimensional trustworthiness taxonomy covering security, safety, privacy, explainability, fairness, and accountability. We systematically analyze threat surfaces across intra-execution loops and interaction planes, formulate a multi-layered zero-trust defense-in-depth architecture integrating Dual-LLM isolation, Capability-Based Access Control, kernel eBPF probes, and sandboxed runtimes, review standardized evaluation benchmarks, and map technical controls to international AI governance frameworks.
[AI-178] yPatch: Transforming Patches into Typestate Rules for Kernel Bug Detection
链接: https://arxiv.org/abs/2609.13728
作者: Ruoyu Wang,Tuo Li,Jia Li
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Operating Systems (cs.OS)
备注: 15 pages, 7 figures, 6 tables
Abstract:Historical Linux kernel patches capture defect knowledge that applies beyond their original repair sites. Recent work has shown that large language models (LLMs) can generate static-analysis checkers from historical patches and use them to uncover new kernel bugs. However, complete-checker generation requires the model both to recover the defect semantics expressed by a patch and to implement sophisticated program-analysis machinery, including object tracking, alias analysis, path-state maintenance, and interprocedural propagation. Coupling these responsibilities in a single end-to-end code-generation task can turn a simple defect rule into an unstable and expensive analyzer-implementation problem. To address this problem, we present TyPatch, which decouples patch-specific defect semantics from analyzer implementation. An LLM translates each patch into a typestate rule specifying its tracked object, actions, guards, transitions, and violations. A shared backend then executes these rules, binding their actions to program events, tracking object identity across aliases, propagating typestate along program paths, and producing reports for all rules. On Linux v6.16, TyPatch finds 559 distinct bugs, 121 of which have been confirmed by kernel developers. In a matched 38-patch comparison with the state-of-the-art complete-checker construction workflow, TyPatch uses 88.3-90.1% fewer generation tokens, while its initial report pools achieve 3.42-14.95 \times the precision of those produced by that workflow. Comments: 15 pages, 7 figures, 6 tables Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Operating Systems (cs.OS) Cite as: arXiv:2609.13728 [cs.SE] (or arXiv:2609.13728v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2609.13728 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-179] IBBench-Light: A Paired Evaluation of Task-Conditioned Responses to External Directives
链接: https://arxiv.org/abs/2609.13725
作者: Kainan Zhou,Gangzhen Qian,Zhaoyi Li,Hang Xiao
类目: Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
备注: ACAIT 2026
Abstract:An external record may contain a procedure to apply or text to read, depending on the user’s request. IBBench-Light tests both uses against the same record. Twelve semantic bases yield 144 matched pairs per model; four quantized instruction models produced 1,152 archived greedy responses. Paired exact-contract accuracy (PECA) requires both members to satisfy their output contracts. Qwen succeeds on 132 execute and 109 process prompts, but only 97 complete pairs, showing what marginal averages omit. We audit literal-target exposure and case normalization, then add 1,722 logged CPU generations to test directive-absent controls, twelve additional semantic bases, within-base wording changes, and generation stopping. In the pinned Phi rerun, changing the end-of-sequence (EOS) set changes exact paired success from 0/144 to 62/144. A bounded IHEval comparison uses the same SmolLM2 checkpoint and output budget while preserving its published instruction roles and scorer. The benchmark measures conditional task and output-contract success. Its task margins and paired count need to be read together with the stopping policy.
[AI-180] MANAS-2: Constrained Reconstruction for EEG Foundation Models
链接: https://arxiv.org/abs/2609.13717
作者: Arvasu Kulkarni,Aditya Ray Mishra,Mahir Jain,Parshva Runwal,Lakshya Saini,Siddharth Panwar,Sandeep Singh
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 17 pages, 3 figures, 15 tables
Abstract:Masked reconstruction is widely used for EEG foundation models, but optimizing reconstruction on low-SNR waveforms does not necessarily produce the most useful latent representation. We introduce MANAS-2, a new EEG foundation model that combines a Raw-Band Hybrid (RBH) masked autoencoder with Constrained Reconstruction (ConRec), a physics-motivated regularizer. RBH jointly reconstructs temporal waveform patches and compact spectral-band targets, while ConRec acts only on the temporal decoder output, penalizing differences in RMS energy between adjacent short windows of the reconstructed waveform. ConRec is intended to shape the encoder by biasing it toward the organization of oscillatory-envelope information. Across seven held-out EEG datasets, adding ConRec to an otherwise identical RBH model increases frozen ridge recovery of six-band spectral power from mean R^2=0.860 to 0.906 and recovery of inter-patch band-energy dynamics from R^2=0.283 to 0.354, while temporal waveform information remains highly recoverable from the frozen latents. Applied to a temporal-only masked autoencoder, ConRec also improves frozen downstream transfer and frequency-dependent latent geometry despite receiving no spectral targets: i.e., the effects of ConRec are architecture-independent. MANAS-2 also outperforms leading EEG Foundation Models on most downstream knowledge-transfer tasks. From the effects of ConRec, we see that a physically motivated constraint imposed through the decoder can make for a more spectrally organized and transferable latent space. MANAS-2 therefore provides a new EEG foundation model built around constrained reconstruction as a mechanism for shaping representation–rather than reconstruction–quality.
[AI-181] JaxAHT: A JAX-Based Library for Ad Hoc Teamwork
链接: https://arxiv.org/abs/2609.13716
作者: Caroline Wang,Rolando Fernandez,Zelal Su Mustafaoglu,Montek Kundan,Jiaxun Cui,Lingyun Xiao,Zhihan Wang,Di Yang Shi,Aditya Madhan,Johnny Liu,Arrasy Rahman,Peter Stone
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Ad Hoc Teamwork (AHT) addresses the challenge of designing agents capable of coordinating with novel partners without prior coordination. However, progress in the field is hindered by the prohibitive computational cost of the AHT research lifecycle, the lack of standardized benchmark implementations, and the absence of a diverse, validated evaluation teammate suite. In this work, we introduce JaxAHT, the first open-source, JAX-based library designed to accelerate and standardize the AHT research lifecycle. Leveraging JAX’s hardware acceleration and massive parallelization capabilities, JaxAHT provides a unified framework for teammate generation, ego agent training, and evaluation against unseen teammates, achieving approximately 95x wall-clock speedup over PyTorch counterparts. Alongside the library, we contribute a diverse suite of evaluation teammates across the domains of Level-Based Foraging, Overcooked, and Hanabi. To illustrate the value of the framework, we use it to conduct a large-scale, compute-controlled benchmark study comparing teammate generation and AHT agent learning methods, finding that no algorithm consistently performs best, and that agent modeling primarily offers benefits in role-based scenarios with diverse teammates.
[AI-182] Degraded but Not Entirely Ineffective: PE-Based Deformable Graph Neural Networks
链接: https://arxiv.org/abs/2609.13712
作者: Jinhua Wu,Xinliang Zhang
类目: Artificial Intelligence (cs.AI)
备注: 28 pages, 10 figures
Abstract:Many real-world scenarios can be represented using graph-structured data. However, traditional GNNs that transmit messages based on first-order neighbors have long faced several fundamental contradictions: increasing depth leads to over-smoothing, long-range dependencies cause over-compression, fixed neighborhoods restrict the receptive field, and on heterophilous graphs, topological neighbors become a source of noise. Although many works have addressed these issues individually, few mechanisms can simultaneously alleviate all of these challenges. To address the aforementioned problems, we propose a Position Encoding-Based Deformable Spatial Aggregation Module (PEBDSAM) that solves them all in one step. Specifically, we utilize a deformable mechanism in the position space to identify relevant nodes to supplement the original first-order neighbor information of GNNs, allowing traditional GNNs to adapt to heterophilous scenarios. Through diagnostic experiments, we obtained several major findings: current offsets fail to have any effect; subsequently, we analyzed the causes of offset failure and why model performance still improves even after offset failure, pointing out future research directions. Based on these diagnostic experiments, we streamlined the original PEBDSAM, resulting in a simplified version, which we call the Position Encoding-Based Spatial Aggregation Module (PEBSAM). In addition, we propose a PEBSAM-Speed to adapt to large datasets. Finally, we designed the module to be plug-and-play and applied it to GCN, GAT, GIN, and GraphSAGE, achieving desirable results on three homophilous datasets and six heterophilous graph datasets.
[AI-183] Gap Entropy and Almost Instance-Wise Optimal Best-Arm Identification
链接: https://arxiv.org/abs/2609.13703
作者: Jiarui Yao,Jiaxi Zhao,Xiangxin Zhou
类目: Computational Complexity (cs.CC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:In the best-arm identification problem, we are given n stochastic arms with unknown means and wish to identify the arm with the largest mean with probability at least 1-\delta , using as few samples as possible. We consider independent Gaussian rewards with unit variance and means in [0,1] . Chen and Li [2016] conjectured that the instance-wise sample complexity of this problem is characterized by the gap entropy, up to an additive term arising from the two-arm problem. In this paper, we resolve their gap-entropy and almost instance-wise optimality conjectures. For an instance I , let \Delta_[i] be the gap between the largest and the i -th largest mean, let H(I)=\sum_i=2^n\Delta_[i]^-2 , and let Ent (I) denote the entropy of the normalized complexities of its dyadic gap groups. For every 0\delta0.1 , we show that the order-oblivious instance-wise lower bound is \Theta (H(I)[\log(1/\delta)+Ent(I)]). We also give a single \delta -correct algorithm with expected sample complexity O ( H(I)[\log(1/\delta)+Ent(I)] +D\log(e+\log(e+D))),D=\Delta_[2]^-2, without prior knowledge of the gaps. Our lower bound removes the dyadic-gap and monotonicity restrictions of previous work, and our upper bound removes the additional polylogarithmic factor multiplying the two-arm term. Thus, a single algorithm attains the instance-wise lower bound up to an additive two-arm term. The main theorems have been formalized and proved in Lean 4. Subjects: Computational Complexity (cs.CC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2609.13703 [cs.CC] (or arXiv:2609.13703v1 [cs.CC] for this version) https://doi.org/10.48550/arXiv.2609.13703 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-184] Leakage-Safe and Scheduler-Aware Machine Learning for Grid Job Runtime Prediction ICML
链接: https://arxiv.org/abs/2609.13701
作者: Ashfaq Ali Shafin,Khandaker Mamun Ahmed
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted in the 26th International Conference on Machine Learning and Applications (ICMLA 2026)
Abstract:Accurate job runtime prediction can improve scheduling-aware resource management in grid and distributed computing environments, but prediction models must be evaluated under realistic deployment constraints. This paper revisits CPU burst time prediction on the GWA-T-4 AuverGrid workload trace and reformulates it as leakage-safe pre-execution job runtime prediction. We define the target as job-level runtime, use only submission-time attributes, exclude post-execution variables, and evaluate models under temporal and cold-start settings rather than relying only on random cross-validation. We compare standard regressors, chronological historical baselines, categorical encoding strategies, and CatBoost with native categorical handling. We further add temporal hyperparameter tuning, runtime predictability analysis, feature ablation, error analysis by job length, and a minimal scheduling simulation. After temporal-validation tuning, CatBoost achieves the strongest deployment-oriented result with R^2=0.239, MAE=27,019, RMSE=46,587, and LogMAE=2.646 on the held-out temporal test set. A single-server simulation over all 69,523 held-out temporal test jobs shows that prediction-informed SJF reduces average waiting time by 50.92% relative to FCFS. The results show that random-split evaluation overestimates performance, categorical-native boosting improves temporal generalization, and long-job underestimation remains a scheduler-relevant challenge.
[AI-185] Drift-Constrained Optimization: Only Direction Matters in Fine-Tuning Instruct Models
链接: https://arxiv.org/abs/2609.13680
作者: Fei Yuan,Changjiang Gao,Yilei Tu,Yifeng Liu,Shujian Huang,Yu Qiao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Fine-tuning instruct models often improves target performance while inducing behavioral drift from the reference model, which can degrade existing capabilities. Rather than treating this drift as an uncontrolled consequence of optimization, we specify a behavioral drift budget before optimization and ask how to boost the target-task performance within it. Locally, behavioral drift induces a shared geometry anchored at the reference model, with the drift budget defining a boundary within this space. In this space, drift determines distance from the reference, leaving update direction as the remaining degree of freedom. Fine-tuning updates can therefore be compared through their directional efficiency, naturally reformulating fine-tuning as a direction-selection problem. This reformulation makes a concrete prediction: changing the accessible directions can qualitatively alter the outcome of fine-tuning. We test this prediction in a stringent QA-only setting, where strong instruct models are fine-tuned only on final answers but must still generate multi-step reasoning at inference. Despite this mismatch, a coarse layer-selective probe reverses the failure of QA-only fine-tuning and reveals the existence of effective directions, with multiple neighboring configurations improving target performance while preserving reasoning and general capabilities. Across Qwen3-8B and Qwen3-14B, these directions substantially improve scientific reasoning and multilingual translation. Over more than 100 languages, the resulting models match or outperform dedicated translation systems and provide a stronger initialization for subsequent reinforcement learning. Our results suggest that fine-tuning is not just about how much a model changes, but how that change is spent. this https URL and this https URL
[AI-186] Windowed A-K-MDP
链接: https://arxiv.org/abs/2609.13676
作者: Xiangwen Yang,Frankie Cho,Iadine Chades
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Markov decision processes (MDPs) are used to support decision-making in conservation of biodiversity, but policies, even over small state spaces, can be difficult to interpret for conservation managers. K-MDP methods address this problem by building simpler MDPs with at most K abstract states. We show that the previously proposed A-K-MDP algorithm that relies on selecting a discretisation divisor using binary search can skip better abstract states. To fix this issue, we propose Windowed A-K-MDP, an algorithm that generates every distinct feasible partition induced within a declared divisor window and evaluates candidates until reaching the ideal value loss (J = 0) or exhausting the family of candidates. Across 33 K-MDP instances, Windowed improved 25 and tied 8.
[AI-187] Recoverability as a System Primitive for Long-Horizon AI Agents
链接: https://arxiv.org/abs/2609.13672
作者: Zhihui Zhang,Wei Liu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:AI agents can be interrupted while editing files, calling tools, or carrying out multi-step tasks. Restarting repeats completed work, but continuing from unverified or outdated progress can carry earlier errors forward. A saved state is not necessarily a suitable place to resume. We introduce recoverability as a system primitive that makes reuse an explicit decision: select a supported starting point and a permitted recovery action, or withhold automatic continuation. Its behavioral contract binds that choice to supporting evidence, execution, and independent checks. A reference architecture connects persistence, validation, and control, with complementary runtime instances testing distinct responsibilities. Four deterministic and 20 paired file challenges demonstrate that accurate restoration and successful completion can conceal disallowed starting points. Progress controls attribute retained work to shared restoration. Event-time tests show that permission must also constrain the action, and that independently held policy evidence can expose violations even after an effect occurs. These findings establish why recovery decisions need their own evaluation, beyond restored bytes and final task success. Within supplied policies and a declared trust model, the contribution is a common, testable interface for retaining justified progress and making the conditions for its reuse explicit and enforceable.
[AI-188] Enhancing Event Candidate Acquisition for Event Linking
链接: https://arxiv.org/abs/2609.13670
作者: Ziyang Zhang,Yinan Liu,Boyi Xue,Yingxuan Huang,Bin Wang,Xiaochun Yang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Event linking associates event mentions in text with entries in a knowledge base (KB), or identifies them as out-of-KB events. Although existing methods use different architectures, candidate event acquisition can still be weakened by short ambiguous mentions, noisy arguments, and evidence that is unevenly useful for retrieval. We present MACE, a Multi-Agent Candidate Event acquisition method that refines event structure before linking. MACE uses evidence-specialized LLM agents to acquire time, location, participant, and event-type evidence, exposes intermediate queries to candidate-event lookup tools, and lets a coordinator revise the evidence set before final candidate construction. Experiments on two event linking benchmarks show that adding MACE to different event linking models consistently improves accuracy. These results show that MACE improves event linking through better candidate event acquisition without modifying the event linking model.
[AI-189] GeoSkill:Experience-Driven Hierarchical Skill Learning with Collaborative Revision forGeospatialAgents
链接: https://arxiv.org/abs/2609.13667
作者: Han Luo,Xian Xu,Yinhe Liu,Yanfei Zhong
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Geospatial agents are increasingly expected to support recurring and evolving analytical tasks rather than execute isolated workflows. In such settings, effective agents must distill prior execution experience into reusable geospatial procedural knowledge to guide future planning and tool use. However, existing memory-augmented paradigms struggle to summarize both long-horizon tool-chain orchestration experience and tool-level invocation constraints in geospatial analysis, while directly relying on LLM self-reflection to update experience often leads to misattribution and unreliable revisions. To address these challenges, we propose GeoSkill, an experience-driven hierarchical skill learning framework for geospatial agents. GeoSkill comprises two core components: (i) a Hierarchical Skill Bank (HSB), consisting of a Planning Skill Bank and a Tool Skill Bank, which respectively distill high-level task-planning experience and tool usage constraints, enabling structured representation and cross-task reuse of historical execution experience; and (ii) a Collaborative Trace-driven Skill Revision (CTSR) mechanism, where Judge, Critic, and Refiner collaboratively perform error identification, skill-level defect localization, and targeted modification, preventing misattributed and unreliable revisions from polluting the skill bank. GeoSkill learns and validates skills from historical executions during development, and freezes the skill bank for retrieval-only guidance on unseen tasks during deployment. Extensive experiments on EarthBench and ThinkGeo demonstrate that GeoSkill effectively transforms historical execution experience into reusable hierarchical skills, improving both end-to-end task accuracy and tool-execution reliability in geospatial tasks.
[AI-190] Safety as a Constraint: Fine-Tuning a LLM Recommender to Explain Itself
链接: https://arxiv.org/abs/2609.13657
作者: Jiashu He,Emma Yanyang Kong,JJ Tan,David Fagnan
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Traditional recommender systems are typically trained to predict what item users will interact with next, but not why. However, offering personalized evidence for why a user might like the predicted item is an important way to enhance the service and to raise the likelihood that the user will be genuinely interested in the recommendation. This service can be delivered by integrating a frontier-model call into the member-facing pipeline, but it will add extra cost and latency. In this paper, we train a recommender LLM to generate personalized explanations for its reccomendation, based on the user’s watching history at a large video streaming service. We impose two requirements on the generated explanation: it must be faithful to the elements of the shows it links, and it must be strictly non-harmful to the user. To this end, we first train two LLM-judge reward models covering three specific criteria, and propose constrained GRPO to incorporate these different criteria. On a held-out real-world testing set, our fine-tuned model improves the all-three-criteria PASS rate rises from 0.649 to 0.956 under our own judges and from 0.677 to 0.931 under an independent judge, where as the frontier generator performs similar to the untuned recommender baseline. We conduct further experiments to show that the model’s language and recommendation abilities remain unchanged. Based on these results, we conclude that an LLM-based recommender can be fine-tuned on other complex tasks without compromising its original recommendation performance, thus provide insights for further agentic user interface powered by a single model.
[AI-191] Solar Intelligence
链接: https://arxiv.org/abs/2609.13648
作者: Jyotsna Singh
类目: Artificial Intelligence (cs.AI)
备注: 15 pages, 2 figures, 4 tables
Abstract:Solar energy decision support is fragmented across dashboards that provide data without explanation, research papers are slow to parse, and general-purpose language models are not solar domain specific and answer without evidence. This paper introduces Solar Intelligence, a hybrid retrieval-augmented framework that unifies structured solar analytics, evidence-grounded scientific question answering, and machine learning forecasting in one system. The platform integrates daily NASA POWER solar and meteorological data, Biosphere 2 ground-sensor readings, and a curated corpus of research papers and institutional reports. Structured queries use DuckDB SQL; scientific questions are answered by a hybrid retriever that fuses BM25 and ChromaDB dense embeddings via Reciprocal Rank Fusion, with responses grounded through a language model (llama3.2:3b). An Extreme Gradient Boosting (XGBoost) model produces daily forecasts of irradiance, temperature, and wind speed. The system is exposed via FastAPI, Streamlit, and an MCP server, so it can be used as an application, an API service, or an agent tool - by students, researchers, and energy analysts.
[AI-192] τ-Elicitation: Benchmarking multi-turn entity extraction in voice agents
链接: https://arxiv.org/abs/2609.13602
作者: Soham Ray,Victor Barres
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Voice agents often need to collect names, addresses, identifiers, dates, and times exactly, yet end-to-end benchmarks obscure where capture fails. We introduce \tau -Elicitation, a 200-task voice benchmark spanning 10 entity types, controlled difficulty, caller realisms, and three environments. A matched text agent passes all tasks, but four voice configurations achieve robust exact success from 0.14 to 0.41. Agents increase verification for hard and unfamiliar entities and sometimes for incorrect captures, but not for their weakest caller voice; only 24 to 37 percent of verified errors are repaired. A scaffold that prescribes spelling, read-back, correction, and confirmation raises robust Pass ^3 by 14 to 31 points, at a cost of 21 to 28 seconds per call. Realisms such as spelling variations and restarts do not detectably affect exact success; mispronunciation increases repair effort. These results identify strategy selection and successful recovery as the central bottlenecks in exact spoken entity collection.
[AI-193] Predictive audio representations for early detection and tracking of hidden dynamic objects
链接: https://arxiv.org/abs/2609.13595
作者: Katerina Vinciguerra,Moritz Brandes,Danilo Hollosi,Letizia Marchegiani
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注: 8 pages, 2 tables, 6 figures
Abstract:Predicting potential dangers is core to safety. Forecasting the presence of other traffic agents is core to danger prediction. Occluded traffic agents challenge detection systems as they might become visible too late, leaving the autonomous vehicle too little time to identify, plan and act accordingly in a robust and safe way. Previous works proved that auditory perception, being omnidirectional and not constrained by a field-of-view, provides fundamental cues for early spotting of different road users, even when hidden by other vehicles or infrastructures. Yet, those contributions deal with scenarios with only one vehicle present, and they either identify the type of vehicle or estimate its direction of arrival. In this work, we move forward, and propose a multi-task system which, simultaneously, estimates the number of vehicles present, their type, and their direction of arrival. Our methodological contribution is a two-stage pipeline: a self-supervised pre-training stage inspired by the Joint- Embedding Predictive Architecture (JEPA) applied directly to multichannel raw waveforms, followed by supervised multi-task fine-tuning with a bidirectional LSTM and three classification heads. The pre-training stage trains the encoder without labels, pushing it to predict the latent representation of a future audio segment from its past context. To train and test our framework, since no suitable dataset was publicly available, we collected an ad-hoc one covering Non-Line-Of-Sight scenarios with multiple traffic agents simultaneously operating. Experimental evaluation shows that our method outperforms the state of the art; it also proves that our design choice allows the model to learn robust representations, which can be transferred to an unseen driving scenario, maintaining reasonable and stable performance.
[AI-194] mKernel: Fast Multi-GPU Multi-Node Fused Kernels
链接: https://arxiv.org/abs/2609.13585
作者: Ziming Mao,Yihan Zhang,Shawn Wei Chew,Shuang Ma,Costin Raiciu,Yang Zhou,Scott Shenker,Ion Stoica
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
备注:
Abstract:Communication has become a bottleneck in distributed training and inference of large models. Overlapping communication with computation at the granularity of kernels, on separate streams, reduces only part of this communication cost. Fused kernels often have better performance by transmitting each output tile as soon as it is produced, but existing fused kernels are largely confined to a single NVLink domain. We present mKernel, a library of multi-GPU, multi-node fused kernels that overlap computation, intra-node NVLink communication, and inter-node RDMA at tile granularity. mKernel partitions the streaming multiprocessors (SMs) of a persistent kernel into compute and communication roles, and an on-GPU controller tunes the SM partition adaptively at run time, since the best SM partition varies with the kernel and the input shape. It structures data movement hierarchically so that data traversing the inter-node network is minimized. Finally, it drives the network from the GPU through a lightweight command queue and host proxy implemented directly on RDMA verbs, which allows the same kernels to run on any network backend (e.g. InfiniBand and on AWS EFA); we observe, surprisingly, that GPUDirect Async (IBGDA) yields little additional benefit over host-assisted GPU-initiated communication. We implement five kernels spanning tensor, sequence, and expert parallelism. On two 16-GPU H200 clusters, mKernel achieves speedups of up to 1.72x on GEMM+AllReduce and 1.88\times on Ring Attention.
[AI-195] FLoKD: Adaptive Knowledge Distillation for Federated Low-Rank LLM over Wireless Networks
链接: https://arxiv.org/abs/2609.13580
作者: Xinlu Zhang,Na Yan,Yang Su,Yansha Deng,Toktam Mahmoodi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) have demonstrated strong capabilities across a wide range of natural language processing tasks. However, conventional fine-tuning typically relies on centralized data collection, bringing in privacy concerns. Federated learning (FL) enables collaborative LLM fine-tuning without sharing raw client data, but its deployment over bandwidth-constrained wireless networks is hindered by the communication overhead of model-parameter transmission. Although Low-Rank Adaptation (LoRA) reduces the number of trainable parameters, its communication cost still increases with model scale. Knowledge distillation avoids parameter sharing via output logits, but token-level logits in LLMs incur high communication cost due to sequence length and vocabulary size. Reducing logits lowers the cost but weakens supervision and degrades accuracy. To address these limitations, we propose FLoKD, an adaptive knowledge-distillation framework for federated LoRA fine-tuning of LLMs over wireless networks, which communicates intermediate LoRA activations as the distillation signal rather than logits or full parameters. Since transmitting all blocks over the entire public dataset remains costly, we further propose a transformer block importance scoring framework that selectively transmits the most informative blocks, and two dataset selection strategies that discard public samples deviating from the local data distribution and prioritise those most informative for distillation. Extensive experiments across multiple generative language datasets, including WikiText-103, PTB, and Dialog, demonstrate that our proposed framework reduces communication overhead by 50-65% while achieving rapid convergence to competitive perplexity compared to baselines.
[AI-196] Causal multi-modal AI for personalized chemosensitivity prediction
链接: https://arxiv.org/abs/2609.13567
作者: Dhruva Biswas,Jeroen Berrevoets,Alec McClean,Linus Bao,Jungkyu Park,Ken G. Zeng,Joseph Cappadona,Cerise Tang,Chuwen Liu,Bartosz Machura,Yin Wu,Valerie Speirs,Hatem Soliman,Rohit Bhargava,Sheheryar Kabraji,Thaer Khoury,David Page,Brian Piening,Carlo Bifulco,Claudia Meurs,Pieter Westenend,Sylvie Chabaud,Jerome Lemonnier,Paul H. Cottu,Florence Dalenc,Fabrice Andre,Frederique Madeleine Penault-Llorca,Thomas Bachelot,Frederick Howard,Francisco J. Esteva,Kevin Kalinsky,Lajos Pusztai,Jan Witowski,Krzysztof J. Geras
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Chemotherapy improves survival for some patients with breast cancer, but doctors cannot reliably predict who. Current guidelines rely on recurrence scores as a proxy for treatment benefit, which may contribute to the overprescription of chemotherapy. Here we present a causal multi-modal AI model that predicts personalized chemosensitivity using routinely collected pathology and clinical information. We developed our model on a multi-national dataset of 9,141 patients (twelve cohorts, nine countries) and evaluated it on another 1,994 patients (five cohorts, three countries). The model generated treatment-specific recurrence probabilities for each patient, with near-perfect calibration and strong prognostic discrimination across both 5- and 10-year follow-up horizons. Moreover, its chemotherapy benefit predictions demonstrated robust predictive performance, and out-performed existing recurrence-score-based tests. Compared to the standard of care, using the model to support personally tailored therapeutic decisions could reduce the number of patients receiving chemotherapy by 30% while achieving the same recurrence-free rate. Tumors predicted to be highly chemosensitive displayed concordant molecular and morphological programs of proliferation, cell cycle progression, and replication stress. The model’s predictive capabilities transferred zero-shot to non-breast cancers, indicating our causal multi-modal AI approach may provide a universal strategy to predict treatment outcomes across cancer types.
[AI-197] Planning or Learning: Reliability and Cost in Multi-Asset Maintenance
链接: https://arxiv.org/abs/2609.13566
作者: Xian Yeow Lee,Chandrasekar Venkatraman,Ahmed Farahat
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Industrial maintenance systems involve multiple interacting assets and shared resources, making it challenging to balance reliability and operational cost using a single decision framework. While recent work has focused on reinforcement learning (RL) for maintenance scheduling, direct comparisons with planning approaches under identical settings remain limited. In this work, we empirically compare planning and RL for multi-asset bearing maintenance using run-to-failure data. We examine how these methods behave when balancing preventive maintenance against tolerable failures across a range of failure penalty scenarios. We observed a consistent behavioral difference driven by objective formulation. Planning enforces reliability as a hard constraint and produces zero-failure policies whose total cost is largely insensitive to the magnitude of failure penalties. RL agents optimize expected cost and often trade off preventive maintenance against occasional failures as penalties vary, resulting in lower costs under low-penalty regimes but persistent non-zero failures even when penalties are high. We also investigate lightweight constraint mechanisms, including reward shaping and action masking, to encourage RL’s reliability. From a practical perspective, planning may be more suitable when strict reliability is required and deployment horizons are short, whereas RL may provide cost-efficient policies when limited failures are acceptable and long-run operational efficiency is prioritized. Overall, this study clarifies the trade-offs between reliability and cost in multi-asset maintenance and suggests that planning and RL are complementary approaches. Beyond these findings, the controlled benchmark protocol itself that unifies environment, cost model, and evaluation across paradigms, offers a reusable template for comparing decision-making approaches in other maintenance settings.
[AI-198] Carbon-Aware Routing for Function Calling in Edge-Cloud LLM Systems CEC
链接: https://arxiv.org/abs/2609.13559
作者: Aikaterini Maria Panteleaki,Varatheepan Paramanayakam,Spyros Tragoudas,Iraklis Anagnostopoulos
类目: Artificial Intelligence (cs.AI)
备注: 2026 IEEE 33rd International Conference on Electronics, Circuits and Systems (ICECS)
Abstract:Large Language Models (LLMs) with function-calling capabilities are becoming critical for modern agentic AI systems. Nevertheless, current deployments typically route inferences to powerful cloud-based models, incurring significant energy use and carbon emissions. We address this sustainability challenge with a carbon-aware routing framework that distributes function-calling queries across a three-tier edge-cloud architecture, combining edge and cloud LLMs on heterogeneous hardware. At its core, a lightweight k-NN predictor operating in a unified semantic-lexical embedding space estimates query-specific accuracy, delay, and power consumption on each edge tier. These predictions are then combined with real-time grid carbon intensity to route every query to the lowest-emission tier capable of executing it successfully. Evaluated on state-of-the-art function-calling benchmarks and LLM families, our framework matches cloud-level accuracy while reducing operational carbon emissions by 4\times on average.
[AI-199] oward a Decision-Assurance Layer for AI-Assisted Flight Planning in Air Traffic Management
链接: https://arxiv.org/abs/2609.13552
作者: Alexandre Barreto(1),Shou Matsumoto(1),Jorge Valverde-Rebaza(2),Cleiton Ataide(3),Paulo Costa(1) ((1) George Mason University, (2) Tecnológico de Monterrey, (3) DECEA: Department of Airspace Control)
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 6 figures
Abstract:Generative AI is increasingly being used informally in Air Traffic Management (ATM) for tasks such as flight plan generation, trajectory interpretation, and constraint checking. Although these tools can reduce workload and accelerate planning, their non-deterministic outputs create safety and operational risks in human-in-the-loop settings. This paper proposes the AI Trust and Assurance Layer (ATAL), a model-agnostic decision assurance architecture that evaluates whether AI-generated flight-planning outputs are sufficiently reliable for operational use. ATAL combines semantic stability under prompt variation, operational consistency of structured outputs, and normative constraint validation against domain rules, and maps these signals to a Decision Readiness Level (DRL) for human operators. An ATM-inspired experimental study shows how unsafe, inconsistent, or misleading outputs can be identified before influencing flight-plan validation or execution. Although demonstrated in aviation, the framework is also transferable to other safety-critical decision-support domains that require human oversight under regulatory constraints.
[AI-200] AutoTailor: Automatic User-Aligned Capability Selection and Adaptation for Web Agents
链接: https://arxiv.org/abs/2609.13548
作者: Xinyun Cao,Adriana Szekeres,Fazle Elahi Faisal
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Web agents can utilize reusable tools to reduce the cost and latency of low-level browser interaction, but automatically discovered tool collections can be large, redundant, and poorly aligned with user demand. We present AutoTailor, a meta-agentic framework for constructing and maintaining a compact set of trajectory-derived Model Context Protocol (MCP) APIs. Offline, AutoTailor converts web trajectories into parameterized browser-automation programs, applies a Quality Filter to remove APIs with unsuitable granularity and redundant functionality, and applies a Usage Likelihood Filter to prioritize broadly useful capabilities while preserving semantic coverage. Online, Dynamic Reselection monitors task outcomes and API usage, identifies recurring coverage gaps, adds relevant candidates, and prunes persistently unused capabilities. We evaluate AutoTailor on 106 WebArena Postmill tasks. Offline filtering reduces the initial 1,283 unrefined APIs to 87, and Dynamic Reselection produces a 33-API set. With reasoning and acting (ReAct) fallback, this set achieves 90.6% correctness, compared with 87.5% for ReAct alone, while reducing average total request-token cost by 57.8% and latency by 29.4%. Without ReAct, it achieves 60.1% correctness, marginally matching the performance of unrefined set, while reducing request-token usage by 94.9%. Together, these results show that static filtering produces a compact inventory of APIs expected to support core, high-likelihood tasks, while dynamic reselection further tailors that inventory to observed user needs. This combination improves accuracy and latency while sharply reducing token usage and end-to-end cost, demonstrating the value of user-aligned capability management for efficient web agents.
[AI-201] Asclepius: An Adaptive Harness for Long-Horizon Clinical Agents EMNLP2026
链接: https://arxiv.org/abs/2609.13543
作者: Grace Chang Yuan,Xiaoman Zhang,Sung Eun Kim,Luyang Luo,Pranav Rajpurkar
类目: Artificial Intelligence (cs.AI)
备注: 9 pages, 4 figures, 11 tables. EMNLP 2026 Findings
Abstract:LLM agents are predominantly benchmarked on short, single-task trajectories, yet real deployments run for hours under contention, surfacing a different class of failures. We use the Clinical Environment Simulator (CES), in which an agent manages an entire emergency-department shift under continuous time and resource pressure, as a testbed: long-horizon execution failures manifest measurably in a single rollout under structured, multi-dimensional grading. On CES, current agents reach the correct diagnosis in most cases yet fail to deliver complete and timely critical actions, revealing an execution gap. We attribute this gap to three long-horizon failure modes, each operationalized as a per-trace counter: instruction-adherence drift, treatment incompleteness, and a severity-equity gap in timeliness. We then introduce Asclepius, an adaptive agent scaffolding with a self-evolving harness that rewrites the operating manual between shifts from trace-level feedback, an externalized clinical skills library for high-stakes regimen knowledge, and three isolated subagents that partition per-turn decisions across the patient queue. On held-out batches never observed during harness evolution, Asclepius improves critical-action correctness by 22% (p = 0.024) over a strong baseline agent framework while preserving diagnostic accuracy, with consistent gains across five LLM judges from three model families; on the full ten-batch set, improvements reach 25% on critical actions and 13% on timeliness. The three failure modes form a coupled bottleneck: decisive reductions appear only when all three components act together.
[AI-202] From Legal Text to AI-specific Risk Sources: A Systematic Analysis of the EU AI Acts High-Risk Requirements
链接: https://arxiv.org/abs/2609.13535
作者: Ronald Schnitzer,Mike Auer,Rumpa Choudhury,Andreas Hapfelmeier,Maximilian Hoeving,Isabelle Painter,Josiane Xavier Parreira,Sonja Zillner
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The EU AI Act introduces mandatory requirements for high-risk AI systems with the explicit goal of ensuring the development and operation of trustworthy AI. At the same time, AI risk management practices rely on structured risk taxonomies to systematically identify and treat AI-specific risk sources. As both the AI Act and established risk taxonomies aim to address AI-induced risks, a natural question is whether they align in the risk sources they cover. However, no clear mapping exists between the risks implicitly addressed by the Act’s high-risk requirements and established taxonomies, leaving practitioners without a structured basis for aligning regulatory obligations with AI risk management practice. This paper presents a systematic classification of the requirements extracted from the EU AI Act Section 2 (Requirements for high-risk AI systems), revealing that only a minority directly address AI-specific risk sources, while the majority impose organizational process and documentation obligations. From the AI risk-related requirements, a consolidated list of distinct AI-specific risk sources is derived. The resulting EU AI Act Risk Source List takes an important step towards bridging the gap between legal obligation and AI risk management practice, providing a structured reference for explicit comparison between existing AI risk taxonomies and the risk sources implicitly addressed by the EU AI Act. Important Note: This is the authors’ preprint. The paper was presented at the 4th International Conference on Frontiers of Artificial Intelligence, Ethics, and Multidisciplinary Applications. A link to the conference’s official proceedings will be provided upon publication.
[AI-203] Attention Is All You Need (to Avoid Spurious Oscillations)
链接: https://arxiv.org/abs/2609.13531
作者: Jinyoung Jeong,Joseph B. Choi,Xinlun Cheng,H. S. Udaykumar,Sanghun Choi,Stephen S. Baek
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Fluid Dynamics (physics.flu-dyn)
备注:
Abstract:Can attention move a shock across several cells in one update without breaking it? We develop a conservative, fixed grid finite-volume scheme in which a CFL-conditioned attention flux selects upstream information according to the transport required by the current time step. One-dimensional inviscid Burgers transport is used as the central mechanism test: the same learned flux remains reliable in the conventional small-step regime and, with a time step four times larger, preserves sharp shocks while using one stage per update. A standard fifth-order WENO scheme with third-order strong-stability-preserving Runge-Kutta time integration (WENO-5+SSP-RK3) is included alongside controlled Forward Euler comparisons to separate flux selection from time integration. The learned attention shifts upstream with the local transport reach and becomes more selective near shocks; inference-time interventions and retrained ablations show that transport-scale information and state-dependent selection contribute directly to performance. Directional two-dimensional scalar Burgers transport and the one-dimensional shallow-water system then test whether the conservation-scale-selection principle transfers beyond the original scalar setting. The results support attention as a learnable information stencil for conservative large-step shock transport, while identifying finite candidate reach and problem-dependent robustness as the present limits.
[AI-204] Rolling Day-Wise Mortality Prediction in Critically Ill Patients With AKI on CRRT Utilizing Machine Pressure Waveforms
链接: https://arxiv.org/abs/2609.13524
作者: Shehan Irteza Pranto,Joanna Yang,Joshua Lambert,Stuart L. Goldstein,Lili Chan,Girish N. Nadkarni,Tiago K. Colicchio,Javier A. Neyra,Jin Chen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Critically ill patients with acute kidney injury (AKI) on continuous renal replacement therapy (CRRT) face high mortality, yet current risk assessment relies primarily on clinical parameters from electronic health records (EHR) and ignores minute-level circuit pressure waveforms generated by CRRT machines that track the extracorporeal circuit’s interaction with the patient. Clinicians therefore cannot see deterioration as it develops. Risk is reassessed only when labs are drawn, while this continuous record is discarded because it is contaminated by shared-device records, non-physiological minutes, and sensor artifacts. To make the stream usable, we aligned machine records to charted therapy intervals to prevent cross-patient leakage, removed priming and downtime minutes, tuned denoising on a synthetic spike-injection benchmark, and masked unobserved intervals rather than imputing them. On this cleaned stream, we define a rolling day-wise task and a transformer-based stacked ensemble that late-fuses a window-reduced sequence transformer with classical models using circuit-instability features and clinical EHR variables. In a leak-safe benchmark on the multi-center CRRTnet cohort (976 patients, 4,585 treatment days), the machine-only model had the lowest standalone prognostic value (AUROC 0.625), followed by the EHR-only model (0.717). Integrating EHR and machine streams reached a one-day mortality AUROC of 0.766. SHAP attribution showed that circuit-instability descriptors raised the machine share of the top 15 combined-model features from 3 to 7 (20.0% to 46.7%), highlighting filter pressure, transmembrane pressure (TMP), and access-to-return difference (ARD). To our knowledge, this is the first patient-level mortality prediction incorporating CRRT machine data, turning a discarded bedside stream into a continuous risk signal.
[AI-205] Positive Topology and Feasible Refinement: Forcing Matrices Positivity and Information
链接: https://arxiv.org/abs/2609.13523
作者: Mirco A. Mannucci,Giovanni Sambin
类目: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI)
备注: 16 pages
Abstract:We develop a conceptual and operational account of Positive Topology starting from a basic relation between points or models and observable properties. From this relation, two complementary structures emerge. The first captures universal refinement and cover: what must hold across all relevant cases and how information can be systematically refined. The second captures positivity and witnessed existence: what can be positively realized and sustained without relying on classical complements. A central result shows that the underlying relation between points and observables can be reconstructed from either of these induced structures. We also clarify the distinction between point-based and pointfree formulations: when points are available, positivity can be derived from the underlying forcing relation, while in the formal pointfree setting positivity is taken as primitive and its compatibility with cover is imposed axiomatically. We develop two complementary interpretations of the framework. The first is information-theoretic, viewing cover as refinement of partial information and positivity as witnessed feasibility. The second is game-theoretic, viewing positivity as the ability of a witness or hypothesis to survive successive legitimate refinements. The final part of the paper is deliberately programmatic. We outline how resource constraints, verification costs, and finite budgets can be incorporated into the framework. This leads to resource-sensitive notions of forcing, cover, positivity, and refinement, and raises new questions about how these structures behave as available resources change. Examples from medical diagnosis, legal reasoning, and AI systems illustrate the potential relevance of the approach to grounded, explainable, and resource-aware inference. Comments: 16 pages Subjects: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.13523 [cs.LO] (or arXiv:2609.13523v1 [cs.LO] for this version) https://doi.org/10.48550/arXiv.2609.13523 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Mirco A. Mannucci [view email] [v1] Fri, 11 Sep 2026 20:46:32 UTC (15 KB)
[AI-206] Frag lingo: Molecular Design via Attachment-Aware Autoregressive Frag ment Generation ICLR2027
链接: https://arxiv.org/abs/2609.13519
作者: Thao Nguyen,Jeonghwan Kim,Zhenhailong Wang,Heng Ji
类目: Artificial Intelligence (cs.AI)
备注: 22 pages. Submitted to ICLR 2027
Abstract:Molecular design is most effective when generation mirrors the edits chemists actually make: extending a scaffold, replacing a substituent, or decorating a scaffold at a specified attachment site while optimizing molecular properties. Fragment-based molecular design naturally supports this workflow, yet existing approaches often separate fragment selection from attachment prediction, first choosing a fragment from a fixed vocabulary and then predicting how it should be connected. This decoupling restricts generation to a closed fragment vocabulary and treats attachment as a separate prediction problem. We introduce Fraglingo, an autoregressive fragment-based molecular generator that jointly models fragment identity and attachment in a continuous latent space. Fraglingo predicts an attachment-aware fragment embedding and retrieves the next fragment through latent-space nearest-neighbor search. To encode attachment context, we introduce a wildcard-anchored readout that represents the growing molecule from the perspective of its active attachment site, enabling the predicted embedding to capture both the molecular context and the required attachment. Because generation operates in a continuous embedding space rather than over fixed fragment identifiers, new fragments can be added to the inference-time vocabulary without retraining, provided their embeddings can be computed by the trained fragment encoder. This retrieval-based formulation provides a unified generation primitive for molecule generation, scaffold generation, scaffold decoration, and molecular optimization. On controlled property-conditional benchmarks, Fraglingo achieves stronger joint property control than comparably trained baselines while maintaining competitive validity, uniqueness, and novelty. Furthermore, Fraglingo generalizes to fragment libraries up to 4x larger than those used during training without retraining.
[AI-207] Adaptive Phase-Switching for Communication-Efficient Federated LoRA Fine-Tuning
链接: https://arxiv.org/abs/2609.13512
作者: Jerry Adams Franklin
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: 16 pages, 5 figures, 7 tables, 7 appendices. Submitted to IEEE Open Journal of the Computer Society. Code and analysis data: this https URL
Abstract:Federated fine-tuning of large language models with low-rank adaptation reduces per-client trainable parameters, but client-to-server communication remains the dominant cost. Existing accounting for federated LoRA protocols omits the asymmetric transition round when a protocol changes aggregation mode, and reports savings that ignore grouped-query attention shapes. This paper measures per-round upload and download bytes for a bidirectional B-only federated LoRA protocol and places five methods, three from prior work, on a single communication-quality frontier. The frontier has a knee, which an adaptive phase-switching aggregator, ReverseAdaptive, locates by monitoring the relative improvement in global training loss against a dimensionless threshold rather than by fixing a phase boundary in advance. On TinyLlama-1.1B-Chat with Alpaca, ReverseAdaptive attains 40.5 percent measured round-trip savings over FLoRA at a held-out instruction-following loss cost of 0.0063. It outperforms FFA-LoRA, which freezes the first of the two LoRA factors at initialization, by 0.0182 in held-out loss, more than twenty times the largest per-method seed standard deviation on that metric, so learning that factor before freezing it produces better adapters. The same threshold transfers across model scales without retuning, and the quality cost of the transition is stable across the two datasets tested.
[AI-208] A Three-Axis Stress Test of LLM vs Classical ML for Network Intrusion Detection under Distribution Shift and Adversarial Evasion
链接: https://arxiv.org/abs/2609.13511
作者: Muhammad Ebad Atif,Muhammad Haider Ali
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注: 11 Pages, 5 figures
Abstract:Large language models are increasingly benchmarked against classical machine learning for network intrusion detection (NIDS), almost always using same-dataset evaluation, and that protocol turns out to be incomplete. Evaluating XGBoost and RoBERTa-LoRA on two independently collected NetFlow v2 networks across three axes (same-dataset performance, cross-dataset transfer, and adversarial evasion) reveals no universal winner. The two models are statistically tied same-dataset. XGBoost wins decisively under cross-dataset distribution shift, by 15 points of F1 and 25 points of balanced accuracy; on the target network RoBERTa-LoRA’s false positive rate reaches 0.78, leaving it barely above chance despite a superficially moderate F1. RoBERTa-LoRA wins decisively under adversarial evasion, by roughly 17 points of F1 at a representative mid-range perturbation strength, while both models hold false positive rates below 0.01 throughout. The model an evaluator would recommend therefore depends entirely on which axis is tested, not on same-dataset accuracy alone. A staged feature-leakage ablation improves cross-dataset transfer non-monotonically, indicating the leakage signal is distributed across the feature representation rather than confined to a few columns, and cross-dataset transfer between our two networks is strongly directional. These results argue for evaluating NIDS models along multiple independent robustness axes, and with more than one metric per axis.
[AI-209] Canaries in the Bank: Auditing User-Level Privacy in Private Evolution
链接: https://arxiv.org/abs/2609.13499
作者: Sai Aparna Aketi,Enayat Ullah,Shripad Gade
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 14 pages
Abstract:Private Evolution (PE) generates high-fidelity synthetic data in federated settings without exposing users’ raw data. It aggregates clipped user votes over a shared candidate bank into a differentially private histogram, with noise calibrated to the worst-case user contribution. However, it is unclear whether an adversary can realize this worst-case privacy loss while following the PE protocol. We introduce a protocol-aware empirical audit in which the server commits to a single shared candidate bank and replaces roughly 1% of its entries with probes derived from a known, non-private canary. We evaluate eight attacks, including an unchanged-bank baseline, exact copies, plausible paraphrases, and high-entropy synthetic nonces. Experiments on Yelp and Sentiment140 show that natural-text attacks remain substantially below the theoretical DP bound, while nonce-based attacks yield considerably stronger bounds and come closest to the mechanism’s privacy ceiling. These results quantify the gap between formal worst-case privacy and leakage achievable through protocol-valid candidate-bank manipulation.
[AI-210] oken Efficient Task Execution via Application Behavior Modeling for Web Agents
链接: https://arxiv.org/abs/2609.13491
作者: Alexandru Ianta,Eleni Stroulia
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The strong performance of AI Agents across an impressive variety of tasks is driving an unprecedented investment in agentic infrastructures, however the cost of processing tokens is fast increasing. Web agents automate the execution of web-application tasks described in natural language, by analyzing the web-application’s user interface (UI) and interacting with it. This work introduces OdoBot, a novel web-agent architecture that completes tasks at a fraction of the cost when compared to conventional web agents. This is achieved by leveraging a behavioral model of the underlying application constructed by analyzing successful task-execution demonstrations. Our experiments with 45 tasks on the Canvas Learning Management System (LMS) demonstrate that OdoBot uses 44% and 80% fewer tokens than two state-of-the-art competitor agents (Agent-E and WebVoyager), while also surpassing WebVoyager in terms of task success rate.
[AI-211] Grounded Adjudication of Variations across Extracted TimeLines (GAVEL): Comparing Clinical Timelines Against Their Case Reports
链接: https://arxiv.org/abs/2609.13475
作者: Jack Cummins,Sayantan Kumar,Ketan Tamirisa,Jeremy C. Weiss
类目: Artificial Intelligence (cs.AI)
备注: AMIA Informatics Summit 2027
Abstract:Existing pipelines for clinical timeline extraction from case reports are evaluated using an expert reference and are limited by imperfect reference annotations and imprecise event alignment. We developed GAVEL, an LLM judge protocol that compares two timelines with the case report and returns a discrepancy type, verdict, and report passage for each difference. We evaluated the event matcher, reviewed 2,738 findings from GPT5.6sol and DeepSeek V3.2, ranked six LLM extractors and two human annotators, and tested GAVEL guided merging. True match rates were 60% immediately below and 48% immediately above the 0.10 cutoff. Manual review confirmed 89.4% and 88.6% of findings. Across 126 reports, merged timelines were preferred in 77.0% of comparisons (95% CI, 69.8 to 84.1%) and reduced discrepancies attributed to the evaluated timeline from 7.63 to 0.85 per report. GAVEL supports report-based comparison and revision without treating either timeline as ground truth.
[AI-212] OrchSLM: Probing the Dynamics of Small Language Model Orchestration
链接: https://arxiv.org/abs/2609.13470
作者: Chengxi Zhang,Yu Yao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Although large language models (LLMs) have demonstrated remarkable capabilities, their reliance on cloud-scale infrastructure poses fundamental challenges for deployment in agentic pipelines, including latency, privacy, connectivity, and substantial computational cost. Small language models (SLMs) offer a compelling alternative: recent studies suggest that many repetitive and narrowly scoped subtasks in agentic workloads may be better served by specialized SLMs than by monolithic LLMs. However, the limited capacity and context windows of SLMs can constrain long-horizon reasoning and interaction-heavy orchestration strategies such as iterative verification and debate. This motivates a complementary, non-interactive paradigm in which heterogeneous SLMs independently generate candidate solutions and a router orchestrates their cached samples without further model interaction. To further understand the mechanisms of such orchestration, we introduce OrchSLM, a routing framework that unifies existing non-interactive orchestration methods and exposes their underlying design choices as controllable parameters. Using OrchSLM as a systematic probe, we reveal how orchestration behavior emerges from diverse knobs, including the task structure, model-pool composition, and multi-agent consensus.
[AI-213] Governing at Machine Speed: An Adaptive Intelligence Architecture for Real-Time AI Policy Enforcement
链接: https://arxiv.org/abs/2609.13466
作者: Sandeep Bokkasam,B. Durgalakshmi
类目: Artificial Intelligence (cs.AI)
备注: 9 pages, 3 tables, 20 references
Abstract:Enterprise AI adoption has reached 78% of organizations globally, yet the infrastructure to govern that adoption has not kept pace. This paper identifies and characterizes the attestation deficit, a structural condition in which organizations maintain governance policies but cannot produce auditable, tamper-evident evidence of enforcement within regulatory timelines. Drawing on empirical data from the Stanford 2026 AI Index Report (362 documented incidents), the IBM/Ponemon 2026 Cost of a Data Breach study (USD 4.99M average cost, 92% lacking access controls), and the EY/AIUC-1 Consortium survey (38% end-to-end monitoring, 17% agent-to-agent coverage), this paper demonstrates that the governance failure is organizational and architectural rather than technical. To address this deficit, we propose AGIL (Adaptive Governance Intelligence Layer), a conceptual five-layer architecture designed to use machine learning for real-time AI governance enforcement. The proposed layers include: (1) Autonomous Discovery for shadow AI detection via behavioral fingerprinting, (2) Behavioral Risk Classification unifying security, hallucination, privacy, and accountability scoring, (3) a Policy Enforcement Gateway for inline permit/deny/modify decisions at sub-100ms latency, (4) a Continuous Attestation Engine generating tamper-evident audit trails as a byproduct of enforcement, and (5) Adaptive Policy Intelligence for ML-driven policy evolution across jurisdictions. AGIL is presented as a theoretical framework and architectural proposal; empirical validation through controlled deployment remains a direction for future work.
[AI-214] meThink: Eliciting Compositional Reasoning in Timeseries Large Language Models
链接: https://arxiv.org/abs/2609.13457
作者: Sudarshan Regmi,Arvind Pillai,Yu Yvonne Wu,Yuliang Chen,Bibek Panthi,Tess Z. Griffin,Michael V. Heinz,Lisa Marsch,Nicholas C. Jacobson,Andrew Campbell
类目: Artificial Intelligence (cs.AI)
备注: Code: this https URL
Abstract:Timeseries multimodal large language models (TS-MLLMs) have recently begun leveraging the reasoning capabilities of large language models (LLMs) for question-answering tasks. However, these models often fail to capture dynamic temporal patterns, providing only implicit reasoning that lacks the underlying explanations critical for high-stakes applications like healthcare. While reinforcement learning (RL)-based timeseries language models aim to address this, they often fall short because they are trained on narrow, in-distribution data and struggle with out-of-distribution compositional questions. To address these challenges, we present TimeThink, a synthetic framework for eliciting compositional timeseries reasoning. Core timeseries primitives (e.g., trend, seasonality) are domain-independent and can be deterministically generated. Guided by this premise, TimeThink first designs a synthetic data generator that produces atomic and composite question-answer pairs, providing objective ground truth with reasoning traces. Building on this framework, TimeThink employs a reinforcement learning with verifiable rewards (RLVR) training strategy that encourages explicit reasoning. Unlike template-reliant methods, this approach enables the model to learn the underlying logic of composition rather than simply imitating traces. Extensive experiments show that TimeThink, trained only on synthetic data, significantly outperforms strong baselines on both synthetic and real-world benchmarks.
[AI-215] Learning to Solve Hard Problems in RL for LLM s by Never Giving Up
链接: https://arxiv.org/abs/2609.13443
作者: Michael Noukhovitch,Hamish Ivison,Nathan Lambert,Aaron Courville
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Blog post this http URL and code available at this http URL
Abstract:We demonstrate that training LLMs with RL does not improve performance equally across a dataset. RL shows large improvements on easy problems that an LLM is already good at solving, but small improvements on hard problems. We call this the Matthew Effect in RL for LLMs, after the phenomenon of cumulative advantage from economics and network science summarized as “the rich get richer”. The naive explanation is that hard problems require more compute to find a solution. We argue that modern RL methods are exacerbating the issue by wasting too much compute on easy problems and instead should dynamically reallocate how they use compute. We introduce Never Give Up (NGU), a simple adaptive sampling method that keeps generating samples for a problem until one is correct. By leveraging asynchronous RL, this naturally uses fewer samples to filter out easy problems and allocates more compute to solving harder problems. We investigate the design choices that affect NGU, such as off-policy robustness, and develop a set of best practices. On the math benchmark Deepscaler, NGU improves performance per compute, especially on harder problems. On a recent coding task, Manufactoria, standard GRPO with a per-test reward fails to fully solve problems that have a range of easy and difficult tests. NGU iteratively improves, solving harder and harder tests, until it learns to fully solve coding problems.
[AI-216] Certifiably Interpretable Training of ReLU-MLPs for Boolean Tasks with Guaranteed Truth-Table Generalization
链接: https://arxiv.org/abs/2609.13439
作者: Hrad Ghoukasian,Anastasis Kratsios
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注: 58 pages, 7 figures, 6 tables
Abstract:As compute scales, models evolve, and training algorithms advance, our ability to explain the increasingly powerful AI systems they enable is eroding. To help safeguard interpretability, we introduce a specialized training algorithm (MACCHIATO) that jointly constructs (i) an explicitly structured \operatornameReLU -MLP from partial truth-table observations and (ii) an explicit Boolean circuit over signed literals with \operatornameAND,\operatornameOR,\operatornameXOR\ gates certifying what its subnetworks compute and how they compose. Intuitively, we iteratively project the residuals of a Boolean function onto low-dimensional \operatornameAND,\operatornameOR,\operatornameXOR\ -circuit classes and exactly compile the resulting circuit into a \operatornameReLU -MLP; we combine \operatornameReLU -MLP circuit compilation, ESPRESSO logic minimization, and influence-based variable selection. Roughly speaking, our interpretability certificate is complemented by a statistical guarantee: under the theorem’s influence-recovery conditions, if each of the m stage-wise residuals depends on at most \log_2(B) bits, a sample-splitting variant of our algorithm trained on T observations returns a six-layer \operatornameReLU -MLP (counting the input layer) of width \mathcalO(mB) with truth-table error \mathcalO\bigl(\sqrtm(B+\log(m/\delta))/T\bigr) . On synthetic random-junta tasks, our networks outperform depth- and hidden-width-matched Adam-trained MLPs in several data-sparse or projection-aligned regimes, while the trained ReLU-MLPs are stronger in others. Moreover, in our explicit PyEDA truth-table implementation, the iterative procedure completes in regimes where flat ambient-dimensional ESPRESSO exceeds the three-hour computational budget. Comments: 58 pages, 7 figures, 6 tables Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML) Cite as: arXiv:2609.13439 [cs.LG] (or arXiv:2609.13439v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.13439 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-217] LabAgent : Customize Any Research Hubs for Scientific Discoveries Using AI Agents
链接: https://arxiv.org/abs/2609.13437
作者: Lei Liu,Yikun Zhang,Jialin Chen,Wanjia Zhao,Rex Ying,Wengong Jin,Hua Xu,James Zou,Tianyu Liu,Hongyu Zhao
类目: Artificial Intelligence (cs.AI); Biomolecules (q-bio.BM)
备注:
Abstract:Scientific research is a continuous process that emphasizes inheritance. Methods developed by predecessors are often expanded upon by new researchers to explore more novel and in-depth scientific questions. However, the change of lab staff, such as student graduation, leads to a lack of personnel capable of replicating methods. Methods that have been developed with significant effort and resources cannot be continued. To address these limitations, we propose LabAgent, a reproduce and discovery harness tailored for a lab’s continuous work. LabAgent employs two mechanisms to guarantee that all skills can be executed and verified and to record the corrective methods and experiences, allowing for direct correction or avoidance of similar errors. We applied LabAgent to drug property prediction, biomedical problem analysis, protein variant effect prediction, and statistical genetics in life science domains. LabAgent ranks first over commercial generalist agents in every domain, and demonstrates accurate reproduction of a published figure. Overall, these results demonstrate that LabAgent can effectively integrate and reasonably expand laboratory knowledge.
[AI-218] oward Self-Adaptive Physical AI: Can LLM Agents Manage Long-Horizon Physical Tasks?
链接: https://arxiv.org/abs/2609.13436
作者: Varun Kaushik,Yayun Tan,Xiaofan Yu
类目: Artificial Intelligence (cs.AI)
备注: 13 pages, 5 Figures, 3 Tables
Abstract:Large Language Model (LLM) agents offer a promising path toward autonomously managing long-term physical tasks without human intervention. However, physical tasks require agents to continuously observe the environment, make consequential actions, and remain effective as the environment changes. Existing approaches either require substantial data and retraining, or primarily focus on agents operating in the virtual world. In this work, we explore the feasibility of building a self-adaptive physical AI agent that manages long-term physical tasks in a zero-shot manner and adapts to environmental changes without human intervention. We design a multi-agent framework that integrates planning, tool calling, observation, and verification, and evaluate it on agricultural tasks against reinforcement learning (RL) agents under different weather patterns. Our results show that zero-shot LLM agents can achieve comparable management outcomes to RL agents under the same weather pattern and adapt more effectively than RL when evaluated under a shifted environment, highlighting a promising path toward self-adaptive physical AI agents.
[AI-219] Chance-Constrained Belief-Space Maneuver Planning for Autonomous Collision Avoidance Under Uncertainty
链接: https://arxiv.org/abs/2609.13428
作者: Grace Ra Kim,Duncan Eddy,Mykel J. Kochenderfer
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 19 pages, 5 figures, 3 tables, Advanced Maui Optical and Space Surveillance Technologies (AMOS) Conference 2026
Abstract:Increasing conjunction frequency in low Earth orbit places growing pressure on spacecraft operators to determine not only whether an encounter requires mitigation, but whether sufficient information is available to commit to a maneuver. This work formulates this information-action tradeoff as a belief-space planning problem for conjunctions between a maneuverable spacecraft and an unmaneuverable secondary object. The planner represents the uncertain orbital states as Gaussian beliefs and uses a chance-constrained belief-space Monte Carlo tree search framework to reason over possible future tracking updates before time of closest approach (TCA). A terminal chance constraint limits the probability of reaching TCA above a prescribed collision-risk threshold, allowing the planner to wait for informative tracking while intervening when deferral becomes too risky. We evaluate the approach on eight historical conjunctions from NASA’s Conjunction Assessment Risk Analysis dataset. By varying the secondary-object measurement quality and tracking cadence, we generate a total of 96 distinct evaluation scenarios. Across the evaluated conditions, the planner reaches TCA without maneuvering in approximately 40% of episodes while maintaining no terminal collision-risk violations. In contrast, fixed-time rule-based maneuver policies resolve more encounters without maneuvering when intervention is deferred closer to TCA, but at the expense of increasing terminal risk violations. The fraction of episodes reaching TCA without maneuvering depends strongly on tracking quality and measurement cadence, ranging from 76% under accurate, frequent measurements to approximately 18%-20% under the poorest tracking conditions. These results show that tracking quality and frequency are not only inputs to collision-risk estimation: they can determine when intervention becomes necessary.
[AI-220] ReCAST: Reward Credit Assignment across Timesteps for Online Diffusion Reinforcement
链接: https://arxiv.org/abs/2609.13425
作者: Yihang Chen,Yuanhao Ban,Kuei-Chun Kao,Cho-Jui Hsieh
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Training diffusion models with multiple rewards requires distinguishing user preference from reward informativeness. User preference determines how much each reward should contribute to the overall objective; reward informativeness determines when its feedback is useful during denoising. Some rewards can meaningfully evaluate a sample as soon as global structure emerges, but others become informative only when the sample is nearly clean. To address both questions jointly, we propose ReCAST (Reward Credit ASsignment across Timesteps), the first method, to our knowledge, for per-reward, timestep-dependent credit assignment in diffusion reward fine-tuning. ReCAST separates user preferences from temporal allocation through a reward-by-timestep weight matrix W , whose row sums match the user-specified reward budgets \lambda , while its column sums are equal, assigning the same total weight to each denoising step. Under these marginal constraints, ReCAST allocates weight according to each reward’s informativeness, quantified by its Rényi discriminability gain at each step. These gains telescope to the total discriminability between the reward-induced positive policy and the current policy, providing a basis for temporal credit assignment. We evaluate ReCAST by training SD3.5-Medium under two distinct four-reward settings, each across five reward budgets \lambda . ReCAST improves the training rewards in one setting and matches them in the other, improves every held-out judge in both, and is preferred by an independent LLM-as-a-Judge. Together, these results show that ReCAST yields improvements that generalize beyond the training rewards and support its core principle: assigning each reward greater weight at the denoising timesteps where its feedback is most informative.
[AI-221] Generalized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-Improvement
链接: https://arxiv.org/abs/2609.13406
作者: Hongyao Tang,Yi Ma,Pengyi Li,Yifu Yuan
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:When we speak of recursive self-improvement (RSI), are we speaking of a phenomenon, a mechanism, or a prospect? Towards autonomous and evolving intelligence, RSI is being claimed at many scales, while no single framework that formally describes these emerging instances exists. Its counterpart in the classical realm, iterative policy improvement, is characterized by generalized policy iteration (GPI), a framework of broad applicability with well-understood theoretical properties, but only where the update principle and the evaluation base lie outside the agent. In this paper, we propose Generalized Agent Iteration (GAI), a formal framework that describes iterative policy improvement and RSI as two cases of a single learning paradigm. GAI defines the agent as a configuration of modifiable components within a system and models the learning process as a cycle of agent evaluation and agent improvement. Two pivotal dials then distinguish the instances: whether the improving mechanism is part of the agent and whether the standard it is measured against is grounded outside it. The former dial delineates the boundary between GPI and RSI, and the latter determines a system’s polarity as anchored, goal drift, or fully self-referential. Moreover, we use these coordinates to place existing systems on the same two axes and make the defects of recursive self-improvement statable one condition at a time. We see this paper as a first step toward exploring a formal characterization of RSI that rests on the classical account, makes existing systems comparable, and provides a principled basis for analyzing and designing new ones.
[AI-222] Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation
链接: https://arxiv.org/abs/2609.13396
作者: Chao Jiang,Yueling Huang,Miqing Li
类目: Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注: 26 pages,4 figures
Abstract:Multi-objective Bayesian optimisation (MOBO) is a sample-efficient approach for optimising expensive black-box functions with multiple objectives. In MOBO, the goal is to adequately approximate the Pareto front; that is, to obtain a high-quality solution set with 1) good convergence (closeness to the Pareto front) and 2) good diversity (spread across the Pareto front). Existing MOBO methods typically aim to accomplish these two tasks simultaneously, i.e., driving the search towards the Pareto front while maintaining a diverse set of nondominated solutions, such that the solutions, ideally, can gradually approach the entire front. When sufficient search budgets are available, this approach is effective. However, considering both convergence and diversity throughout the search is not easy and requires careful design. Under very tight budgets, there may not be enough solutions generated to be able to simultaneously approach the entire Pareto front. To address this issue, this paper proposes a \textitconverge-then-diversify (CTD) approach that decouples convergence and diversity into two stages. In the first stage, CTD focuses on convergence, aiming to quickly drive the search toward a single point on the Pareto front. In the second stage, CTD focuses on diversity, aiming to spread solutions across the front. We present two simple instantiations of CTD by using widely adopted acquisition functions in the area. Experimental results show that, across all 446 pairwise comparisons, CTD statistically outperforms state-of-the-art methods in 72.9% of the cases, performs equivalently in 21.1%, and is statistically worse in only 6.1%, with the advantage being particularly evident in settings with very tight evaluation budgets or in high-dimensional problems.
[AI-223] ask-Aware Federated Fine-Tuning for MoE-based Large Language Models ICDM2026
链接: https://arxiv.org/abs/2609.13395
作者: Tingqi Wang,Hongyu Ke,Haoxin Wang,Rafal Angryk,Zhipeng Cai
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted by ICDM2026
Abstract:Mixture-of-Experts (MoE) has become a widely adopted architecture for Large Language Models (LLMs), as it improves model capacity while limiting computational overhead through sparse expert activation. This property makes MoE-based LLMs particularly attractive for resource-constrained distributed environments. However, federated fine-tuning of MoE-based LLMs remains challenging under heterogeneous client data. Since clients often correspond to different task preferences, directly aggregating their local updates may weaken expert specialization and introduce conflicting update directions on shared experts. To address these challenges, we propose FedTAR, a task-aware federated fine-tuning method for MoE-based LLMs. FedTAR establishes the association between local updates and task preference via routing outputs. Specifically, we apply Singular Value Decomposition (SVD) to both routing features and local updates to extract low-dimensional task coordinates and update directions. Based on the task coordinates, FedTAR performs intra-cluster aggregation among clients with similar task preferences and inter-cluster aggregation across different task groups. The aggregated update is then reconstructed through the learned task-to-update mapping, ensuring that the final update remains aligned with task-specific optimization directions. In this way, FedTAR preserves expert specialization and mitigates destructive interference among heterogeneous clients. We evaluate FedTAR on four benchmark tasks under different non-IID settings. Experimental results demonstrate that FedTAR consistently outperforms strong federated fine-tuning baselines and achieves state-of-the-art performance.
[AI-224] ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agent ic Search
链接: https://arxiv.org/abs/2609.13356
作者: Jiyan He,Guang Liang,Hao Liu,Haoxiang Guan,Jinbo Sun,Junyi Guo,Wenjun Feng,Yantai Xie,Yifei Shen,Bin Shao,Chuyang Wei,Kai Chen,Kexin Zhou,Minghang Zhu,Shuxin Zheng,Tie-Yan Liu,Taine Zhao,Wenhui Zhu,Xueyin Xu,Xiaoqing Zhang,Yatao Li,Yuxuan Ren
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:In this work, we present ZGCM-1, a fully open 7B dense foundation model trained from scratch with extreme data, system, and algorithmic efficiency. ZGCM-1 is founded on a core premise: compact models cannot passively memorize the open web, but can overcome parametric capacity limits by coupling deliberate internal thinking with active external tool use. To support this paradigm across a 256K context, we develop an end-to-end, high-efficiency open training recipe: Architecture System Co-design: interleaved gated sliding-window and full attention, and a stable FP8 Muon optimizer; Progressive Curriculum MDP Mid-Training: context scaling across 16K, 64K, and 256K, and the reformulation of interaction traces into Markov Decision Processes. Furthermore, we establish an AI-native RD workflow where agent swarms autonomously manage cluster operations, data curation, and rapid diagnostic evaluation. Extensive evaluations show that ZGCM-1-7B is competitive across 7B model family on general benchmarks. On several challenging mathematical reasoning and agentic search suites, it remains competitive with frontier models orders of magnitude larger, such as Qwen3-235B-A22B and GLM-5.1. We also show that our pre-training design offers a ~4.2x efficiency improvement in 16K pre-training time-to-loss. Across the full development lifecycle, we distill eight actionable empirical findings-spanning architectural scaling, SFT quality pruning, long-context generalization, and agentic co-training dynamics. To facilitate community research, we open-source model weights from the pre-training, mid-training, and post-training stages, intermediate checkpoints, training code, per-stage data and data recipes, and WB logs.
[AI-225] SkillAtlas: An Attack Trace Library for Agent Skills EMNLP2026
链接: https://arxiv.org/abs/2609.13353
作者: Yuxin Tian,Zenghao Duan,Liang Pang,Zhiyi Yin,Xueqi Cheng
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted at REALM @ EMNLP 2026 (non-archival workshop paper). 9 pages, 4 figures
Abstract:Agent skills are reusable units for language-model agents, but their risks emerge through model decisions, user context, tool calls, and execution feedback rather than through stable signatures or a single sandbox run. Existing static, dynamic, and benchmark-style evaluations rarely preserve public evidence that can be inspected, searched, and reused. We present SkillAtlas, a hosted attack trace library that converts private agent-skill security report bundles into reviewed, redacted, and searchable public cases. The library contains 3,014 cases, 6,589 traces, 151,131 steps, 233 affected skills, and 8 risk categories; 42.5% of successful cases first become successful after a non-success initial round, and trajectory-grounded labels improve pre-execution guard accuracy to 0.770.
[AI-226] Bridging Thought and Action: Taming Long-Horizon Instability in Open-Source LLM Agents with a MetaTool-Enhanced ROS Framework
链接: https://arxiv.org/abs/2609.13335
作者: Kazi Abrar Mahmud,Nilotpaul Kundu Dhurubo,Tamal Kirttonia,Sabbir Hossain Ujjal,Mohammad Ariful Haque
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 8 pages, 6 figures, Submitted to IEEE Transactions on Cognitive and Developmental Systems (TCDS)
Abstract:Large Language Models (LLMs) have enabled more natural human-robot interaction, but open-source models often exhibit unstable long-horizon reasoning and inefficient action execution when deployed in agentic robotic frameworks. This paper presents an enhanced ROS-Agent based architecture that improves task reliability and execution efficiency for agentic robotic systems using open-source LLMs. The proposed system introduces a novel intermediate mechanism, termed the MetaTool, which enforces structured planning prior to action execution. Given a natural-language command, the MetaTool induces the LLM to generate a pseudo-code plan of intended tool invocations, which is stored in the ROS-Agent’s scratchpad and persists throughout execution. By explicitly separating planning from execution, the proposed approach reduces execution loops and improves deterministic behavior. The architecture is validated on a custom mobile robotic platform with multimodal perception and motion control capabilities. Experimental results on real-world interactive tasks demonstrate improved task completion and contextual consistency, with up to ~24% gains on complex tasks compared to the baseline framework.
[AI-227] Feasibility and Memory Mechanisms of Chern-Simons Context Reservoir Computation
链接: https://arxiv.org/abs/2609.13315
作者: Jyotiranjan Beuria,Venkatesh H. Chembrolu
类目: Emerging Technologies (cs.ET); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
备注: 5 figures, 3 tables
Abstract:We investigate whether a Chern-Simons (CS) context reservoir is a viable computational substrate and whether evolving its gauge connection provides a benefit beyond simpler mechanisms. The reservoir state is a density fluctuation on a two-dimensional context manifold, whose drift is generated by a density-sourced connection. To separate generic reservoir behavior from gauge-specific effects, we compare four matched models: reciprocal transport, instantaneous transverse reconstruction, local nonlinear feedback, and fully coupled conserved-current CS dynamics. Across ten random seeds, the fully coupled CS dynamics propagates Gauss law to numerical precision, converges under spatial and temporal refinement, remains stable under constraint-compatible noise, and satisfies the spatial CS equation more accurately than the instantaneous controls. All four models exhibit fading scalar memory and distinguish matched pulse-order histories in density, with no resolved general advantage for coupled CS. The distinction appears in the flow geometry: coupled evolution supports circulating and longitudinal history channels simultaneously, retains them briefly after input removal, and yields a combined-feature pulse-order accuracy of (0.879\pm0.035), compared with (0.679\pm0.065) for the instantaneous-transverse control. The evolved connection also cannot be reconstructed from an instantaneous density snapshot or replaced by a fitted local multiplier. We therefore find a task-specific advantage for geometry- and order-sensitive processing, rather than generic reservoir superiority. Here ``topological’’ refers to the gauge organization of the state; the reported memory and cyclic-lag measures are not topological invariants.
[AI-228] IMM-based Multiple Object Tracking using a State Prediction Neural Network
链接: https://arxiv.org/abs/2609.13307
作者: Chan-Bin Lim,Dong-Hee Paek,Seung-Hyun Kong
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 8 pages, 4 figures
Abstract:Object tracking is essential for autonomous vehicles to avoid obstacles and plan routes. Radar maintains detection performance even in adverse weather and can measure relative velocity through the Doppler effect, making it well suited for object tracking. In this paper, we propose a data-driven state PRedictor-based Interacting Multiple Model tracking method (PR-IMM) that improves nonlinear object-motion representation while preserving the stability and interpretability of physics-based motion models. The proposed method employs a transformer-based PRediction model (PR) that incorporates radar Doppler measurements to predict object displacement. The PR model is integrated into the IMM as a mode alongside the CV, CA, and CT motion models, and their prior positions are dynamically combined according to the mode probabilities. Experimental results show that PR-IMM reduces position-estimation error by 57.3% over the IMM and by 16.5% over the PR, while reducing ID switches by 25.3% and improving IDF1 by 9.6% over the IMM.
[AI-229] Conflict-Predictive Variable Horizons in Multi-Drone Distributed Model Predictive Control
链接: https://arxiv.org/abs/2609.13270
作者: Linda Mümken,Michael Schwung,Stefan Lier,Andreas Schwung
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注: 14 pages, 6 figures, 2 tables, is already submitted to IEEE Transactions on Systems, Man, and Cybernetics as journal-paper (current status “under review”). I think this is again more eess.SY, but also if my last paper is published there, i am still not allowed to use that category, please decide for yourself
Abstract:In distributed model predictive control for multi-drone collision avoidance, a fixed prediction horizon forces a compromise: a short horizon is inexpensive but reacts late to approaching neighbors, whereas a long one anticipates conflicts at a per-step cost that grows superlinearly with its length. We propose a conflict-predictive variable horizon that each drone sets locally, leaving the distributed model predictive control itself unchanged. From a short history of observed positions, a drone extrapolates the flight lines of its neighbors, tests each against its own using confidence funnels that narrow with prediction range, and obtains each time to conflict in closed form. The horizon is then the smallest admissible value whose planning window covers the farthest predicted conflict. It collapses to its minimum in clear airspace and grows only when a conflict lies ahead. Provided this minimum meets a single computable feasibility bound, we prove that recursive feasibility and asymptotic stability are preserved for every horizon the policy can select. These guarantees hold for a linear model, and a cascaded inner loop reduces each quadrotor’s translational dynamics to a perturbed double integrator, so they carry over to the linearized quadrotor model and, as practical stability, to the full nonlinear one. In simulation on dense antipodal-swap benchmarks, the variable horizon reduces both per-step solver cost and total computation well below those of a long fixed horizon, and it maintains separation in every run, which a short fixed horizon of comparable per-step cost does not.
[AI-230] ShieldVLA: Feasibility-Aware Safety Alignment for Vision-Language-Action Models
链接: https://arxiv.org/abs/2609.13231
作者: Manan Tayal,Akshay Nambi
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 24 pages, 3 figures, 14 tables
Abstract:Vision-Language-Action (VLA) models demonstrate strong generalization in robotic manipulation and navigation, but existing fine-tuning methods provide limited safety guarantees. Current approaches primarily rely on Lagrangian optimization that enforces safety through soft penalties on expected cumulative cost, often resulting in residual constraint violations or overly conservative behavior. Moreover, learning safety in visual domains is challenging due to the absence of dense per-step safety annotations. We propose ShieldVLA, a safety-aligned fine-tuning framework for VLA models based on Hamilton-Jacobi (HJ) reachability. ShieldVLA learns a model-free approximation of the HJ reachability value function directly from visual observations to estimate the safe operating region. The learned safety critic gates policy optimization by separating reward maximization within feasible regions from recovery near unsafe states, avoiding persistent reward-cost trade-offs. To enable scalable supervision in visual environments, we introduce rubric-based VLM safety scores that convert semantic safety feedback into structured critic targets without requiring manual cost labels. Across five navigation and manipulation benchmarks spanning multiple VLA backbones, ShieldVLA reduces cumulative safety cost by 57% on average and improves task success rate by +0.13 over SafeVLA.
[AI-231] Diagnosing Faults in Reinforcement Learning Simulators and World Models with Canonical Polynomial Invariants
链接: https://arxiv.org/abs/2609.13194
作者: Tesfay Zemuy Gebrekidan,Hadush Hailu Gebrerufael
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 9 pages with 4 figures, 2 tables and supplementary file
Abstract:A large literature builds physical structure into learned dynamics on the premise that models respecting the underlying physics predict better. We test that premise using exact polynomial invariants recovered from trajectories and canonicalised as reduced Gröbner bases over \mathbbQ . On Acrobot, exactness provides little benefit for prediction: a consistency regulariser reduces algebraic residual while leaving rollout fidelity essentially unchanged, and a shaping potential recovered from a system with a 100% mass error accelerates learning as effectively as the correct potential. Exact canonical invariants instead prove valuable for diagnosis. We develop two procedures: screening, which identifies the violated physical constraint, and attribution, which recovers the faulty invariant and identifies the responsible physical parameter. To enable this, we introduce normal-form deflation and quotient-space recovery. Across fifteen injected faults, screening localises every broken constraint with no false alarms, whereas observation-space baselines do not localise any; attribution recovers the responsible parameter on all seven parameter faults. Paired difference tests detect all faults, showing that the advantage is localisation rather than detection. Perturbing reference generators by 10^-4 preserves 14–15/15 localisations, showing that screening does not require exactness, whereas ideal-equality decisions distinguish perturbations of only 10^-12 , showing that exactness is required for algebraic comparison. Applied to 350 release pairs across eleven RL environments, the diagnostic finds no evidence of changed simulator dynamics, instead revealing properties of the benchmark implementations themselves.
[AI-232] Multilingual Agent System for Inclusive Wildfire Evacuation Guidance
链接: https://arxiv.org/abs/2609.03301
作者: Shruti Kulkarni,Lynn Tong,Aditi Namboodiripad,Chelyah Miller,Helen Lin,Peeyush Patel,Bogdan Bistriceanu,Diane Myung-kyung Woodbridge
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注: This one is under submission (IEEE SpatialConnect Workshop2026)
Abstract:Wildfire seasons have become 84 days longer in the current days than in the 1970s, causing enormous threats to one’s financial status and short- and long-term health. During the fire, public agencies send out emergency messages to provide warnings and orders. Although 26 million people in the US have limited English proficiency, over 80% of those messages are only delivered in English, which can cause disproportionate information distribution and awareness. In order to better serve marginalized communities during emergencies, the authors developed BEACON, a service that provides comprehensive and personalized evacuation guidance, including navigation routes, personalized checklists, and a chatbot in the language that a user uses. Our current system ingests data including fire perimeter information, evacuation order status, and shelter information from Watch Duty. When a user is within a certain proximity from the fire, the system utilizes real-time GPS locations and nearby weather data from the National Oceanic and Atmospheric Administration (NOAA) to predict fire danger levels. The assessment model refreshment are dynamically scheduled based on fire progress and trends using XGBoost. If the location has a likelihood of fire danger, the system sends alerts with evacuation routes outputted from a polygon-avoidant routing pipeline. The application provides a context-aware multilingual agent that users can communicate with and is tightly connected to other features of the application. In addition, based on data that the user entered, the system dynamically generates and checks off personalized reminder items to provide an organized evacuation plan. The system’s user interface dynamically changes its language settings based on the language the user most recently used in either setting or chatbot conversation for all the application elements.
[AI-233] Natural-Language to SysMLv2 Translation via Conformance-Driven Iterative Refinement
链接: https://arxiv.org/abs/2607.14162
作者: Chance LaVoie,Eladio Andujar Lugo,Taylan G. Topcu,Levent Burak Kara
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: Accepted for presentation and publication in the Proceedings of the ASME 2026 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference (IDETC/CIE 2026)
Abstract:Model-Based Systems Engineering (MBSE) relies on formal system models as primary technical artifacts for representing requirements, structure, and behavior across the system lifecycle. With the standardization of SysMLv2 as a textual language, interest is increasing in translating natural-language descriptions directly into executable models. For practical deployment, generated models must be accepted by industrial modeling environments, not merely satisfy grammar constraints. We present a conformance-checker-driven framework for reliable natural-language-to-SysMLv2 translation that enforces production-level acceptance as the termination condition. The system embeds a SysMLv2 conformance checker within a generate-check-repair loop. Each model is evaluated using the checker, and deterministic diagnostics are incorporated into revisions until zero conformance errors are achieved. Using the production checker as the oracle ensures the framework targets deployability rather than grammar plausibility. We evaluate the approach on the full SysMBench prompt set of 151 prompts across four large language model backends, yielding 604 prompt-model cases. Single-shot generation achieves 51.16% production-conformance acceptance, while our approach achieves 100.00% conformance. By elevating production conformance from a post-processing check to a control mechanism within generation, the framework converts probabilistic outputs into production-accepted SysMLv2 artifacts suitable for loading, visualization, and engineering use.
[AI-234] Generalization Can Emerge in Tabular Foundation Models From a Single Table
链接: https://arxiv.org/abs/2511.09665
作者: Junwei Ma,Nour Shaheen,Alex Labach,Amine Mhedhbi,Frank Hutter,Anthony L. Caterini,Valentin Thomas
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Deep tabular modelling increasingly relies on in-context learning where, during inference, a model receives a set of (x,y) pairs as context and predicts labels for new inputs without weight updates. We challenge the prevailing view that broad generalization here requires pre-training on large synthetic corpora (e.g., TabPFN priors) or a large collection of real data (e.g., TabDPT training datasets), discovering that a relatively small amount of data suffices for generalization. We find that simple self-supervised pre-training on just a \emphsingle real table can produce surprisingly strong transfer across heterogeneous benchmarks. By systematically pre-training and evaluating on many diverse datasets, we analyze what aspects of the data are most important for building a Tabular Foundation Model (TFM) generalizing across domains. We then connect this to the pre-training procedure shared by most TFMs and show that the number and quality of \emphtasks one can construct from a dataset is key to downstream performance.
[AI-235] owards Optimizing SQL Generation via LLM Routing NEURIPS2024
链接: https://arxiv.org/abs/2411.04319
作者: Mohammadhossein Malekpour,Nour Shaheen,Foutse Khomh,Amine Mhedhbi
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Table Representation Learning Workshop at NeurIPS 2024
Abstract:Text-to-SQL enables users to interact with databases through natural language, simplifying access to structured data. Although highly capable large language models (LLMs) achieve strong accuracy for complex queries, they incur unnecessary latency and dollar cost for simpler ones. In this paper, we introduce the first LLM routing approach for Text-to-SQL, which dynamically selects the most cost-effective LLM capable of generating accurate SQL for each query. We present two routing strategies (score- and classification-based) that achieve accuracy comparable to the most capable LLM while reducing costs. We design the routers for ease of training and efficient inference. In our experiments, we highlight a practical and explainable accuracy-cost trade-off on the BIRD dataset.
[AI-236] Predictive Likelihood Ratios for Language Model Watermark Detection
链接: https://arxiv.org/abs/2609.15657
作者: Li Ma
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Applications (stat.AP)
备注:
Abstract:Keyed watermark detection tests dependence between observed tokens and pseudorandom variables reconstructed from a secret key. Building on the pivotal framework of Li et al. (2025), we construct predictive likelihood ratios that average over uncertain probability deficits and residual-tail distributions. The aim is robust detection power across alternative specifications without requiring a single signal-strength tuning. A mixture prior combines tail shape and effective width; hierarchical extensions allow within-document variation in deficit or width. The test maximizes prior-averaged power at a fixed size, but is not generally uniformly most powerful or minimax. Under the exact conditional pivot null, normalized predictive alternatives selected before each observation yield a Bayes factor that is also a test martingale: Type I error control is unaffected by alternative misspecification and remains valid under optional stopping. This guarantee does not cover violations of the conditional null, and the interpolated implementation has no certified anytime guarantee. Gumbel marginal likelihoods are evaluated by fixed quadrature. Across the evaluated tail-shape and tail-width alternatives and three horizons, the union-tail mixture has maximum observed Type II error regret .0080, compared with .0962 for the equal-tail mixture, relative to the best tested rule. On temperature-matched outputs from two open models, it improves AUC over the equal-tail baseline in all eight non-saturated model-temperature cells, although the leading reference score generally has higher AUC. Supplementary experiments show retained power under independent null-like replacement and smaller changes from hierarchical dependence modeling. The evidence supports robustness across the evaluated alternatives, not uniform power guarantees or resistance to arbitrary text edits.
[AI-237] Math for AI safety: an invitation for mathematicians
链接: https://arxiv.org/abs/2609.15289
作者: Lionel Levine
类目: History and Overview (math.HO); Artificial Intelligence (cs.AI)
备注: 33 pages, 5 figures
Abstract:Artificial intelligence threatens to outrun human understanding and control. New mathematics is needed to design AI that is legible, steerable, and cooperative with humanity. I organize this invitation by mathematical field, so you can turn straight to your own: logic and game theory for cooperation; probability for agency and world-models; algebra and representation theory for learned features; analysis and geometry for generalization and training dynamics. Each section ends with an open problem that is accessible to a working mathematician with no prior experience in AI safety.
[AI-238] Interpreting hierarchical organisation of speaker embeddings ICASSP2027
链接: https://arxiv.org/abs/2609.15203
作者: Yanze Xu,Wenwu Wang,Mark D. Plumbley
类目: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI)
备注: Submit to ICASSP 2027
Abstract:Speaker recognition neural networks learn latent representations (i.e. speaker embeddings) from input utterances to recognise speaker identities. However, the internal mechanisms of these networks remain largely opaque, motivating research in explainable artificial intelligence (XAI) to understand them. Nevertheless, existing studies have analysed how speaker embeddings are organised, but rarely frame these analyses within XAI. Hence, this work proposes to explain and interpret the organisation of speaker embeddings from an XAI perspective. To this end, we apply a hierarchical clustering algorithm, Single-Linkage Clustering (SLINK), to analyse whether some speaker embeddings naturally form clusters with hierarchical relationships. The resulting hierarchical organisation (i.e. hierarchical clusters) is evaluated using the Cluster-Class Matching (CCM) method. Moreover, we propose a new method, termed Hierarchical Cluster-Class Matching (HCCM), to identify which hierarchical clusters best match individual semantic classes (e.g. male) and conjunctive semantic classes (e.g. UK male), thereby interpreting the clusters using their matched classes. The matching degree is quantified using a new metric called the L-score, which makes imperfect matches diagnosable. HCCM’s results show that hierarchical clusters analysed by SLINK are interpreted using different classes related to speaker identity, gender, and nationality, providing insight into semantics within the hierarchical organisation of our examined speaker embeddings. Comments: Submit to ICASSP 2027 Subjects: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.15203 [eess.AS] (or arXiv:2609.15203v1 [eess.AS] for this version) https://doi.org/10.48550/arXiv.2609.15203 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-239] Physics Informed Neural Network model for the dynamical study of Abdominal Aortic Aneurysm
链接: https://arxiv.org/abs/2609.15104
作者: Adrián Robles Arques,Martín Ruiz Fernandez,Javier Sanchis,Miguel A. Teruel,Juan Trujillo
类目: Fluid Dynamics (physics.flu-dyn); Artificial Intelligence (cs.AI)
备注: 38 pages, 9 figures and 2 tables
Abstract:We present the development and application of a three-dimensional Physics-Informed Neural Network (PINN) framework for the investigation of haemodynamic behaviour in the human aorta. The model incorporates a time-resolved simulation of pulsatile blood flow over a two-minute interval, enabling the extraction of pressure and velocity fields with high temporal fidelity. The mechanical stress exerted on the aortic wall was quantified through Laplace’s law, with temporal averaging applied to derive representative stress distributions. This approach circumvents the computational overhead associated with conventional computational fluid dynamics (CFD) methods by eliminating mesh generation and exploiting the automatic differentiation capabilities inherent to neural networks. The proposed methodology demonstrates that PINNs can serve as an efficient and accurate alternative for modelling complex vascular flow phenomena, offering significant advantages in scalability and computational cost reduction while maintaining physical consistency.
[AI-240] Neural-Network Solutions to Real-Space Charge Density and Generalization
链接: https://arxiv.org/abs/2609.14906
作者: Yuxuan Zeng,Taoyuze Lv,Zhicheng Zhong
类目: Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI)
备注:
Abstract:The Hohenberg-Kohn theorem establishes that, in principle, the ground state (GS) charge density contains all GS information of a many-electron system, such that all GS observables can be expressed as functionals of the GS charge density. Conventional Kohn-Sham density functional theory requires iterative solution of the self-consistent-field equations at substantial computational cost, motivating the development of deep learning surrogates for electronic structure calculations and, in turn, accelerating computer-aided materials design. Here, we propose \textbfAIDEN, an \underlineAtomic-\underlineInteraction \underlineDensity \underlineEquivariant \underlineNetwork for solving real-space charge density. AIDEN separates the element-dependent one-center density from environment-induced density redistribution and represents the latter through complementary atom- and edge-centered tensor correlations. A continuous low-rank Gaussian decoder then reconstructs the density at arbitrary spatial coordinates while reusing atomic encodings independently of the evaluation grid. AIDEN achieves state-of-the-art accuracy on periodic crystal benchmarks while remaining competitive for molecular systems, and further demonstrates zero-shot transferability across several structurally distinct out-of-distribution case studies. Furthermore, AIDEN provides substantially faster inference than both baseline models and full SCF calculations, enabling efficient charge density reconstruction for large-scale electronic structure calculations.
[AI-241] PU classification under Non-SCAR: clustering-assisted logistic model with oversampling enhancement
链接: https://arxiv.org/abs/2609.14675
作者: Konrad Furmańczyk,Kacper Paczutkowski
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 11 pages, 1 figure, 3 tables. Supplementary materials and full reproducible R code available at: this https URL
Abstract:This study addresses the PU classification problem under violations of the SCAR assumption. We investigate logistic regression-based approaches, namely the cluster method and its extensions with strict and non-strict Lasso regularization. The primary contribution of this work is the integration of the SMOTE technique to alleviate class imbalance and systematically assess its impact on the performance of the considered algorithms. SMOTE is first applied to rebalance the training dataset. Next, cleaning labels are derived via 2-means clustering. Logistic regression is then trained on the cleaned data, where identified positive instances are augmented with additional true positives and the remaining observations are treated as negative. The experimental evaluation is conducted on 13 real benchmark datasets and one synthetic dataset. For comparison, we include the naive approach and the Spy-EM method. The results demonstrate that incorporating SMOTE improves classification performance when the SCAR condition is violated and indicate moderate robustness of the LassoJoint method in this setting.
[AI-242] Proving olympiad geometry theorems on a superconducting quantum processor
链接: https://arxiv.org/abs/2609.14533
作者: Ning Wang,Zheng-Zhi Sun,Zhengyi Cui,Yiren Zou,Aosai Zhang,Fanhao Shen,Jiarun Zhong,Zehang Bao,Zitian Zhu,Han Wang,Jia-Nan Yang,Jiayuan Shen,Gongyu Liu,Yanzhe Wang,Yihang Han,Yiyang He,Jiahua Huang,Sailang Zhou,Xinrong Zhang,Yaozu Wu,Zixuan Song,Jinfeng Deng,Hang Dong,Qi Ye,Weikang Li,Si Jiang,Yixuan Ma,Shuangyue Geng,Zhide Lu,Chao Song,Hekang Li,Pengfei Zhang,Qiujiang Guo,H. Wang,Dong-Ling Deng
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI)
备注:
Abstract:Automated theorem proving seeks to use computational systems to prove or disprove mathematical and logical statements [1, 2]. It underpins a wide range of applications, and enhancing theorem-proving capabilities remains a central objective in artificial intelligence [3]. Although recent neuro-symbolic systems have achieved remarkable progress [4-7], their operation is ultimately constrained by classical computational architectures. Quantum computing [8], by contrast, enables information encoding and coherent parallelism beyond classical limits [9-14], raising the possibility of accelerating structured symbolic deduction [15]. Here we report the experimental realization of automated geometry theorem proving on a fully programmable superconducting quantum processor. We develop two complementary quantum proving frameworks. The first implements Wu’s algebraic elimination method using quantum pseudo-division, with multivariate polynomials represented in superposition states, enabling quantum algebraic theorem proving. The second implements the full-angle method as backward symbolic reasoning through a hybrid quantum strategy-guided architecture, demonstrating a general route toward quantum symbolic proof search. As illustrative examples, we prove two theorems on a superconducting quantum processor: the perpendicularity of the diagonals of a square and a 1978 International Mathematical Olympiad geometry problem. Our results establish, at the experimental level, automated logical reasoning as a viable task for near-term quantum processors and provide a concrete pathway toward quantum-enhanced symbolic intelligence.
[AI-243] Modeling Scaling and Decoding: Optimizing Controllable Speech Generation with Nonverbal Vocalizations
链接: https://arxiv.org/abs/2609.14231
作者: Ziyu Zhang,Yun Chen,Taihui Wang,Hanzhao Li,Qicong Xie,Rilin Chen,Zhixian Zhao,Lei Xie
类目: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI)
备注:
Abstract:Controllable synthesis of nonverbal vocalizations (NVVs) is es- sential for natural and expressive speech, but remains challeng- ing due to their acoustic diversity and imbalanced distribution in existing corpora. To address these challenges, we develop an NVV-aware DiTAR system that models continuous speech latents, encodes the 16 target NVV categories as dedicated to- kens, and adapts stop prediction to distinguish mid-utterance vocalizations from utterance boundaries. Training begins with large-scale bilingual pre-training on diverse NVV speech, fol- lowed by continued supervised fine-tuning on a corpus en- hanced through targeted synthetic augmentation and frequency- aware rebalancing. At inference time, we select the acoustic prompt, tune the LM-guidance and noise-injection scales, and apply Best-of-N sampling with multi-metric selection to re- duce generation failures. The final system achieves an official weighted bilingual score of 62.786, ranking first in Mandarin, second in English, and first overall among participating systems in Track 2 of the ISCSLP 2026 NVVSpeech Challenge. Ab- lation studies show that targeted augmentation benefits under- represented NVV categories the most, while robust candidate selection requires balancing NVV correctness, lexical fidelity, and perceptual quality.
[AI-244] On the Equivalence of Stochastic Control and Path Space Formulations for Schrödinger Bridges over Compact Connected Lie Groups
链接: https://arxiv.org/abs/2609.13758
作者: Hamza Mahmood,Georgiy A. Bondar,Abhishek Halder,Adeel Akhtar
类目: Optimization and Control (math.OC); Artificial Intelligence (cs.AI); Systems and Control (eess.SY); Probability (math.PR); Machine Learning (stat.ML)
备注:
Abstract:We establish the equivalence between the stochastic optimal control and path space formulations of the Schrödinger bridge problem (SBP) for the kinematic equation on a compact connected Lie group. Using the geometric concepts of horizontal lift and stochastic anti-development, we derive a Girsanov-type change-of-measure result, and show that the expected control energy equals the relative entropy of the controlled path law with respect to the reference Wiener measure. Thus, the SBP is equivalently a path space relative entropy minimization problem subject to prescribed endpoint this http URL result has three useful implications. From an analytic viewpoint, the shown equivalence helps prove the existence and uniqueness of the SB. From a probabilistic viewpoint, it helps interpret the SB as the most probable deviation of the uncontrolled stochastic dynamics consistent with the endpoint constraints. From a computational viewpoint, it allows using static Sinkhorn recursions to directly solve the relative entropy minimization problem and compute the optimal path measure. We illustrate the equivalence numerically on the torus \mathbbT^2 . The code is publicly available at: this https URL
[AI-245] Building a Production Greek-English Speech Recognizer
链接: https://arxiv.org/abs/2609.13498
作者: Christos Petrocheilos,Cleopatra Papadopoulou,Chris Porikis,Ioakeim Perros,Ayoub Kirouane,Themistoklis Nikolis
类目: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Sound (cs.SD)
备注: 21 pages, 6 figures, 4 tables this https URL
Abstract:We report a multi-month engineering program to build Sophea, a production bilingual Greek-English automatic speech recognition system. We evaluate the system against nine production gates covering Greek and English word error rate, language identification, and hallucinations on non-speech audio. Across twenty-three training iterations and two model architectures, no training-data composition passed all nine gates simultaneously. Meeting the Greek noisy-environment target required about 1,500 steps of dense domain exposure, while preserving English language identification tolerated only about 250 steps, or about 1,250 with a rebalanced mix that reduced Greek accuracy. We describe a six-stage data pipeline in which calibrating an audio-quality filter against in-domain anchors reduced the discarded share of scored Greek audio from 98.7 percent to 10.6 percent. A pre-registered ablation isolated a hallucination defect to one training-data package. A three-model ROVER ensemble increased gate coverage from 4-7 of 9 for individual models to 9 of 9 and reduced overlapping-speech WER from 53.35 percent to 37.87 percent, a 29 percent relative improvement. A separate learned per-clip arbiter over two models is listed as sophea/asr-k1 (preview) on the public Open ASR Leaderboard, with 4.26 percent average WER across eight public English test sets, and reaches 25.88 percent WER on live Greek noisy-environment traffic. We also document five cases in which a measurement tool produced a plausible but incorrect result and seven substantial approaches that were evaluated but not shipped. No model weights or training data are released; we report methodology and quantitative results only.
[AI-246] Edge-addition monotonicity of positive p-energy fails for every p = 1
链接: https://arxiv.org/abs/2609.13476
作者: Koyar Afrasyab
类目: Combinatorics (math.CO); Artificial Intelligence (cs.AI)
备注:
Abstract:At a 2021 AIM workshop, Guo conjectured that the positive square energy s+ = E+_2 should inherit the familiar edge-addition monotonicity of the spectral radius, rho(G + uv) = rho(G). That conjecture was subsequently shown to fail at p = 2. Tang, Liu, and Wang then introduced positive p-energy, proved nonmonotonicity for every 1 = p 3, and in version 3 of their preprint (26 March 2025) explicitly conjectured that monotonicity should hold for p = 3. We disprove this conjectured high-exponent extension completely: for every real p 2 there are infinitely many connected graphs G and nonedges uv such that E+_p(G + uv) E+_p(G). Together with the Tang-Liu-Wang counterexamples below 3, this shows that no exponent p = 1 restores the spectral-radius-style monotonicity: positive p-energy can decrease under the addition of an edge for every real p = 1. The construction is a chain of clique blocks joined by regular bipartite graphs. An equitable quotient converges to Q = I + cA(P_k). For noninteger p, a binomial-series sign argument for a fractional power of I - cA(P_k) gives the required negative endpoint entry. At integer exponents, choosing c across the first spectral threshold leaves exactly one negative eigenvalue, and path locality forces the positive spectral contribution to have negative sign. We also give a fully rational 38-vertex certificate at p = 4 and determine the complete failure interval of a fixed 17-vertex counterexample at p = 3. Subjects: Combinatorics (math.CO); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.13476 [math.CO] (or arXiv:2609.13476v1 [math.CO] for this version) https://doi.org/10.48550/arXiv.2609.13476 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-247] Symmetry- and Property-Aware Crystal Generation with Reinforcement Learning for Inverse Materials Design
链接: https://arxiv.org/abs/2609.13468
作者: Ting-Wei Hsu,Arun Bansil,Qimin Yan
类目: Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI); Computational Physics (physics.comp-ph); Optics (physics.optics)
备注: 18 pages, 5 figures
Abstract:The inverse design of crystalline materials ultimately seeks structures with desired physical properties. However, for many functional responses, a favorable numerical value is meaningful only when supported by the symmetry of the underlying crystal. Without the appropriate crystallographic constraints, an apparent response may be ill defined, accidental, or not symmetry protected. Here we introduce SPARC, a symmetry- and property-aware reinforcement learning framework that optimizes physical objectives while preserving the structural conditions required for their realization. We demonstrate SPARC on two complementary tasks. The first targets strong uniaxial dielectric anisotropy, a tensorial response that is well defined only within appropriate crystal classes. The second maximizes the spectroscopic limited maximum efficiency, a scalar device-level objective without a prescribed symmetry class, allowing the framework to identify favorable crystallographic motifs. These results show that symmetry is not merely an additional design constraint, but a physical foundation for generating candidates with meaningful, robust, and realizable functional properties.
[AI-248] Chemical and geometric representation fidelity improves drug–target affinity prediction
链接: https://arxiv.org/abs/2609.13230
作者: Yixiao Li,Yining Qian,Yefan Chen,Zenghui Chen,Jiayue Sun,Yuhai Zhao,Cheng Tan,An-Yang Lu
类目: Biomolecules (q-bio.BM); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
备注: 21 pages, 10 figures
Abstract:Predicting drug–target binding affinity (DTA) requires models to distinguish subtle chemical and structural determinants underlying molecular recognition. Although recent approaches increasingly incorporate richer drug and protein information, such information may be compressed, homogenized or discretized during representation construction, causing affinity-relevant distinctions to be lost before interaction modelling. We hypothesized that this representation-stage information loss constitutes an upstream bottleneck that cannot be reliably overcome by increasingly complex interaction predictors. To test this hypothesis, we developed ReGeoDTA, a representation-preserving framework that maintains affinity-relevant chemical heterogeneity in molecular representations and continuous geometric relationships in protein structures. Across three benchmark datasets, ReGeoDTA consistently improved affinity prediction, and the proposed representation-preserving strategies retained their benefits across diverse DTA architectures. Controlled representation degradation progressively reduced predictive performance, whereas increasing downstream predictor complexity failed to recover information lost during representation construction. These findings identify representation fidelity as an upstream design principle for accurate and generalizable drug–target affinity prediction, with potential implications for computational compound prioritization.
[AI-249] Planning as Dynamics Relaxation: Hippocampal Recurrent Network Realizes Optimal Goal-Directed Navigation
链接: https://arxiv.org/abs/2609.13219
作者: Yuhang He,Junfeng Zuo,Tianhao Chu,Si Wu
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI)
备注:
Abstract:Neural correlates of spatial cognitive map are well documented, yet exactly how neural circuits perform spatial navigation in complex environments - e.g., reaching a goal while avoiding obstacles - remains largely unclear. Here, we show that a hippocampal network with appropriate recurrent connections can naturally achieve optimal goal-directed navigation via its relaxation dynamics. Specifically, we consider that the recurrent weights between the neurons represent the transition probabilities between spatial locations encoded by neurons; obstacles such as walls and blocked corridors are therefore reflected by the vanishing of connection weights. This connection pattern can be learned in the hippocampus via behavioral-timescale synaptic plasticity (BTSP) while the animal is exploring the environment. When a goal signal is presented, the network dynamics will relax into an activity field representing the goal location. We prove that this field is mathematically equivalent to the desirability field of a Linearly-solvable Markov Decision Process (LMDP), and the local log-gradient of the field indicates the navigation direction. Both theoretical analyses and simulations demonstrate that this recurrent network dynamics-mediated navigation is efficient and robust in environments with complex obstacle layouts. Moreover, only low-rank updates of the network’s connection pattern are needed when the environment has local changes. We hope this study offers insight into a general circuit principle for planning in abstract rational maps in the brain beyond spatial navigation.
机器学习
[LG-0] Privacy-Aligned Personalized Federated Learning with Compact Adaptation and Variable-Length Gaussian Communication
链接: https://arxiv.org/abs/2609.15950
作者: Yilin Xu,Chun Hei Michael Shiu,Chih Wei Ling,Linqi Song
类目: Machine Learning (cs.LG)
*备注: 8 pages, 1 figure
Abstract:Record-level differential privacy exposes a structural misalignment in personalized federated learning when client-specific variation is low-dimensional while training repeatedly releases high-dimensional updates. In this paper, we address this misalignment by releasing a private client context once and confining repeated adaptation to a fixed coefficient space. Beyond dimensionality reduction, the factorized generator induces an adaptive optimization geometry that reshapes noisy updates, and controlled ablations show that most of its private-training gain is retained by radial evolution. To further reduce the communication cost, we realize the Gaussian mechanism for coefficient updates directly through variable-length quantization with finite expected code length, so that the quantization error itself serves as the required privacy perturbation rather than extra distortion. Across MNIST and CIFAR-10, our design matches or outperforms full-model private adaptation across privacy budgets and client heterogeneity, while reducing protected uplink by a factor of 2.67 at (\varepsilon=16) on CIFAR-10 with comparable future-client accuracy.
[LG-1] Safe Meta-Reinforcement Learning via Information Space Reachability
链接: https://arxiv.org/abs/2609.15915
作者: Zeyang Li,Sunbochen Tang,Navid Azizan
类目: Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注:
Abstract:Meta-reinforcement learning (meta-RL) enables agents to adapt to unseen tasks with limited experience. Despite its promise, the application of meta-RL in real-world tasks is hindered by safety requirements, which have been underexplored in prior work. In this paper, we propose a safe meta-RL framework that explicitly accounts for safety during adaptation. Our key insight is to reason about safety in the information space, which captures both the physical state and the agent’s belief over the underlying task. Within this space, we introduce a safety value function that measures the probability of the agent avoiding unsafe regions indefinitely. We show that this function satisfies a self-consistency condition and a Bellman equation, which make it learnable via meta-RL. Based on this formulation, we develop a safe meta-RL algorithm that learns the safety value function and leverages it for safety filtering and constrained policy optimization. Experiments on meta-RL benchmarks demonstrate the effectiveness of the proposed method.
[LG-2] Discrete Beckmann Transport Models for One-Step Language Modeling and Reasoning
链接: https://arxiv.org/abs/2609.15903
作者: Sophia Tang,Shiyi Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Discrete diffusion and flow models are a promising alternative to autoregressive language models, but compressing many-step sampling into fewer steps typically requires distilling a pretrained teacher model. This caps the student at the teacher’s quality and requires a costly two-stage training pipeline. We introduce Discrete Beckmann Transport Models (DBTM), built on a time-independent flow whose autonomous transport map provably carries any point in the ambient space to a fixed point on the vertices of the simplex in a single step. We show that this fixed-point property is characterized by a conservation equation whose residual can be minimized directly from data, removing the requirement for a teacher flow and time conditioning. Under this construction, a partially trained map corresponds to the flow truncated at finite time, so generation reduces to iterating one map until it reaches a fixed point. We further extend the map to a partial-context interpolant where additional function evaluations act as refinement steps rather than ODE integration steps. On language modeling and reasoning tasks, DBTM enables one- and few-step generation that improves quality and accuracy over discrete diffusion and continuous flow baselines.
[LG-3] hin-shell stability of Gaussian cooling: logconcave sampling with sesteric complexity from a cold start
链接: https://arxiv.org/abs/2609.15884
作者: Yunbum Kook,Santosh S. Vempala
类目: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Probability (math.PR)
*备注: 24 pages
Abstract:We show that logconcave probability measures along the Gaussian cooling path have thin-shell stability, generalizing the thin-shell theorem. This result leads to improved complexity for the fundamental problem of sampling an arbitrary logconcave distribution from a cold start. For (near-)isotropic logconcave distributions, the complexity is nearly n^2.5 , improving the previous bound of n^2.75 , and matching the complexity of the abstract Speedy walk.
[LG-4] ask-Directed Residual AddUNet:Perfect-Reconstruction Routing for Full-Rate Representations
链接: https://arxiv.org/abs/2609.15857
作者: Vikram R. Lakkavalli
类目: Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注:
Abstract:This paper establishes a perfect-reconstruction (PR) interpretation of AddUNet and its full-rate realization, and introduces a Residual Full-Rate PR architecture for task-directed representation learning. The survivor–skip structure of a constrained additive U-Net is shown to be exactly equivalent to a critically sampled multirate PR filter bank. The full-rate formulation removes the complementary-subband restrictions of the critically sampled system while preserving PR. A Residual Full-Rate PR architecture is then proposed to progressively route task-irrelevant, nuisance, or redundant structure away from the task-facing survivor while retaining the routed information explicitly. Exact reconstruction is guaranteed for arbitrary shape-compatible linear or nonlinear routing operators, without requiring invertibility, a matched synthesis bank, reconstruction loss, or learned decoder. The resulting architecture decouples representation design from reconstruction design: conservation is structural, while learning is devoted to task-directed routing. The same formulation identifies an identity-shortcut ResNet with its residual output retained as a full-rate PR system. Experiments verify exact single-channel routing of linearly separable factors to machine precision. On TIMIT, the proposed front-end improves test PER from 28.60\pm2.09% to 25.76\pm0.41% with the recognizer and training protocol held fixed, while maintaining exact reconstruction. Speaker probing further shows that structural conservation does not itself imply task-specific invariance.
[LG-5] Sharp Rates and a One-Line Correction for Spectral Representation Learning
链接: https://arxiv.org/abs/2609.15825
作者: Dier Tang,Jing Yee Tan,Guangyue Han
类目: Machine Learning (cs.LG); Information Theory (cs.IT); Machine Learning (stat.ML)
*备注: 27 pages, 4 figures
Abstract:A self-supervised encoder is trained once, frozen, and reused through lightweight probes on tasks nobody named at training time; the practitioner’s question is when the off-the-shelf features are good enough and when they need fixing. Canonical correlation analysis, HGR maximal correlation, and the population optimum of the spectral contrastive loss all return the top- k singular subspace of a cross-view dependence operator, justified by isotropy: if the task prior has no directional preference, that subspace is universally optimal. We show isotropy is the wrong hypothesis. The prior enters the transfer risk only through the task covariance \Lambda=\mathbbE[\Delta\Delta^\top] , and only through its compression onto the operator’s leading singular directions; what matters is not whether \Lambda is isotropic but whether its preferred directions are ordered consistently with the operator’s spectrum. We prove matching two-sided rates—worst-case regret is exactly 1-1/\kappa(\Lambda) , refines to 1-A_k for an alignment coefficient A_k , localizes to the top- 2k subspace, becomes second order under a spectral gap, and is improvable by no task-agnostic representation—and show why alignment is generic: incoherent preferences cancel in high dimension, and T diverse tasks force \alpha=\widetilde O(\sqrtd_x/T) , a quantitative account of why task diversity, not symmetry, makes self-supervised features transfer. The governing statistics cost O(kd_x^2) , and when they signal misalignment a one-line reweighting of the positive-pair term provably restores exact optimality. The result is a diagnostic that answers the practitioner’s question from a small labelled budget and refuses when the task bank cannot support the width requested; on controlled data it takes a regret of 0.86 down to 0.003 , and on a CIFAR-100 encoder it correctly predicts that no correction is needed.
[LG-6] Accelerating Transfer-Learning-Based Autotuning with Predictive LLVM IR Performance Ranking
链接: https://arxiv.org/abs/2609.15807
作者: Md Arafat Hossain,Thomas Randall,Akash Dutta,Xingfu Wu,Rong Ge,Ali Jannesari
类目: Performance (cs.PF); Machine Learning (cs.LG)
*备注:
Abstract:As the complexity of High Performance Computing (HPC) ecosys- tems continually increases, achieving optimal performance becomes a challenge. Traditional performance autotuning techniques pro- vide promising means to navigate this complexity, these techniques remain computationally intensive and require many evaluations to find optimal configurations. This work proposes an autotuning framework that designs a machine learning-based ensemble LLVM Intermediate Representa- tion (IR) ranker, Neural Configuration Scorer (NCS). NCS ranks the performance of IRs sampled by a transfer-learning-based autotuner, improving the efficiency of the tuning process by reducing tuning overheads and circumventing subpar evaluations. By leveraging knowledge from related tasks, we are able to effectively exploit the transfer relationship to access high-performing configurations in fewer samples than traditional techniques that rely upon itera- tive refinement. Our framework can achieve similar performance improvements as state-of-the-art autotuning techniques with up to 61.67% fewer evaluations, averaging 27.85% fewer evaluations across various HPC benchmarks.
[LG-7] MoveBench: A Benchmark for Global-Scale Wildlife Movement Forecasting
链接: https://arxiv.org/abs/2609.15780
作者: Justin Kay,Shir Bar,Ellen O. Aikens,Martin Becker,Francesca Cagnacci,Juliet Cohen,Scott W. Forrest,Jessica Kendall-Bar,Madeleine Lucas,Macon Overcast,Meredith S. Palmer,Will Rogers,Nicholas J. Russo,Christian Rutz,Larissa T. Beumer,Michael Brown,Ying-Chi Chan,Sarah C. Davidson,Diego Ellis Soto,Anne G. Hertel,Roland Kays,Benjamin Koger,Guram Mikaberidze,Thomas Mueller,Ruth Oliver,Thorsten Papenbrock,Robert Patchett,Jared A. Stabach,Dane Taylor,Scott W. Yanco,Sara Beery
类目: Machine Learning (cs.LG)
*备注: Pre-release preprint
Abstract:Understanding and predicting wildlife movement is critical for ecology and conservation. While trajectory forecasting has advanced for human and vehicle movement, wildlife trajectories present distinct challenges: they are unconstrained in space, highly stochastic, and influenced by environmental conditions. We introduce MoveBench, the first large-scale benchmark for probabilistic wildlife movement forecasting, containing 2.6M GPS locations from 800+ individuals across 110 species in 127 countries, paired with 1.6B environmental raster tiles capturing 160 covariates known or hypothesized to influence movement. We propose a probabilistic evaluation protocol for movement trajectory forecasts, addressing limitations of point-prediction metrics for inherently stochastic phenomena. Through comprehensive empirical evaluation of four method families across multiple temporal and spatial scales, we reveal that: (1) existing predictive methods generalize better to future timepoints than to unseen individuals, (2) deep learning approaches do not consistently outperform simpler baselines, and (3) environmental covariate selection significantly impacts performance. MoveBench enables standardized evaluation of movement forecasting methods and provides a foundation for methodological advances on this ecologically important task.
[LG-8] Solving Finite-sum Coupled Compositional Optimization via Multi-block-Single-probe Estimator
链接: https://arxiv.org/abs/2609.15723
作者: Wei Jiang,Sifan Yang,Yibo Wang,Lijun Zhang,Zechao Li
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Traditional variance reduction methods (e.g., SPIDER, SARAH, STORM) have been extensively investigated for improving the convergence rates of stochastic optimization. These techniques typically maintain a sequence of estimators for a single function (or gradient) across iterations. However, what if we need to track multiple functions, but can only access stochastic samples of \mathcalO(1) functions at each iteration? This scenario arises in an important emerging family of finite-sum coupled compositional optimization (FCCO) problems of the form \frac1m\sum_i=1^m f_i(g_i(\mathbfw)) , where each g_i is accessible only through a stochastic oracle. The key challenge is to track \mathbf g(\mathbfw)=(g_1(\mathbfw), \ldots, g_m(\mathbfw)) over time, where \mathbf g(\mathbfw) has m blocks but only \mathcalO(1) blocks can be probed for their stochastic values at each step. To address this challenge, we propose a novel Multi-block-Single-probe Variance Reduction (MSVR) estimator to efficiently trace \mathbf g(\mathbfw) under partial block sampling. Building on the MSVR estimator, we develop several algorithms for FCCO problems, achieving improved sample complexities for non-convex, convex, strongly convex, and Polyak-Łojasiewicz (PL) objectives. We further obtain an improved dependence on m when the outer function gradients \nabla f_i are linear. Empirical studies on multi-task deep AUC maximization further demonstrate the superior performance of the proposed estimators.
[LG-9] Assembling the CREW: A Collaborative Multi-agent Reinforcement Learning Framework for Automated Related Work Generation
链接: https://arxiv.org/abs/2609.15721
作者: Hai-Dang Dang,Bao-Yen Pham,Bao Nguyen,Tran Thi Huong,Huynh Thi Thanh Binh
类目: Machine Learning (cs.LG)
*备注:
Abstract:Automatic Related Work Generation (RWG) significantly reduces the human time and effort required to author the Related Work Section (RWS) of a research paper. However, prior methods leveraging multi-agent Large Language Models (LLMs) typically rely on a predefined workflow, where each agent is responsible for a specific step in the entire process. This rigid, static inter-agent coordination limits the adaptive collaboration required to synthesize complex scientific literature. To address this limitation, we propose CREW (Collaborative Reinforcement Learning for Related Work Generation), a novel framework where LLM agents bypass heuristic pipelines to dynamically coordinate by autonomously selecting actions, such as Retrieve, Disseminate, Compose, and Critique, driven by a policy optimized via Independent Proximal Policy Optimization (IPPO). Extensive experiments on a standard RWG benchmark demonstrate that our approach yields substantial quality improvements over strong existing baselines, while significantly reducing token costs. Code is available at this https URL
[LG-10] Knowledge-Enriched Structured EHR Features for 30-Day Hospital Readmission Prediction on MIMIC-IV
链接: https://arxiv.org/abs/2609.15713
作者: Mohamad Najafi,Hongyun Fu,Mathias Brochhausen,Jian Wu,Yaohang Li
类目: Machine Learning (cs.LG); Quantitative Methods (q-bio.QM)
*备注: 15 pages, 3 figures. Accepted at SDSC 2026 Mid-Atlantic
Abstract:Recent approaches to 30-day hospital readmission prediction rely on pre-trained language models applied to discharge summaries. Although these methods achieve strong performance, they depend on the availability of clinical notes, incur substantial computational costs, and yield representations that lack interpretability. We propose a knowledge-enriched feature representation that augments structured Electronic Health Record (EHR) data with four medical knowledge sources: disease ontology mapping, procedure classification, drug ingredient vocabulary, and organ system laboratory aggregation, without using clinical notes. Each feature dimension corresponds to a named clinical concept, yielding a sparse and interpretable patient representation. The approach is evaluated with six classifiers on a MIMIC-IV v2.2 cohort. Under 20-fold cross-validation, the best configuration achieves an AUROC of 0.743. This performance is comparable to that of previously reported methods on this dataset, including both those using only structured data and those incorporating clinical notes, while requiring considerably less computational cost. Interpretability analysis shows that demographics, organ system labs, drug ingredient features, and first-level ontology disease categories drive prediction, while deeper hierarchy levels contribute negligibly. These findings indicate that knowledge-enriched structured features offer a competitive and efficient alternative to embeddings from clinical notes for 30-day readmission prediction.
[LG-11] Backward SDEs-based Diffusion for Physics-Constrained Generation ICML
链接: https://arxiv.org/abs/2609.15702
作者: Zihao Wang
类目: Machine Learning (cs.LG)
*备注: Accepted as ICML paper
Abstract:Pretrained score-based diffusion models provide strong unconditional priors, yet enforcing measurement or physics consistency in inverse problems is often handled by heuristic guidance, intermittent projections, or task-specific conditional training, with limited guarantees of feasibility at the end of inference. We propose terminal-conditioned inversion for score-based SDE priors. Given a frozen Score-SDE prior and a task-defined terminal feasibility specification, we construct an associated backward stochastic differential equation whose adapted solution defines a principled inverse map from the terminal requirement to a prior state at a chosen noise level. Under standard regularity conditions, we establish existence and uniqueness of the adapted solution and obtain terminal consistency by construction. We further develop a practical neural BSDE solver that composes arbitrary pretrained diffusion priors with domain constraints without modifying the score-defined coefficients, producing an anchored prior state that enables neighborhood sampling for uncertainty characterization. Experiments on toy datasets validate stable terminal-conditioned inversion and distributionally consistent neighborhood sampling. As a real-world case study, we apply the framework to sparse-view CT reconstruction and achieve improved reconstruction quality over representative training-free baselines while satisfying strict measurement feasibility under the prescribed terminal specification. Project is available in: \hrefthis https URLthis https URL
[LG-12] Principal-timestep Restricted Init via Sparse Matrix-decomposition in Flow-matching
链接: https://arxiv.org/abs/2609.15643
作者: Jiayang Gu,Zheng Fang,Lichaun Xiang,Fanghui Liu,Xu Cai,Hongkai Wen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Flow-matching diffusion models have recently emerged as a strong paradigm for high-fidelity visual generation. However, their prohibitively high fine-tuning cost limits scalability to downstream tasks. While Low-Rank Adaptation (LoRA) combined with spectral initialization has demonstrated accelerated convergence and improved performance in autoregressive language models by better aligning gradient directions, we find that it fails to deliver similar gains in diffusion fine-tuning, often yielding marginal or even negative improvements over vanilla this http URL attribute this discrepancy to a fundamental mismatch between LoRA’s low-rank parameterization and the intrinsically high-rank gradients induced by the flow-matching objective. In particular, stochastic timestep sampling introduces directionally heterogeneous gradient signals across training steps, leading to misaligned updates under low-rank this http URL address this issue, we propose Prism-LoRA,a Principal-timestep Restricted Init via Sparse Matrix-decomposition framework that improves gradient alignment during fine-tuning. Our method consists of two key components: (i) principal timestep selection, which restricts initialization gradients to a subset of dominant timesteps to suppress effective gradient rank, and (ii) principal channel filtering, which removes task-irrelevant channels, enabling the one-step spectral initialization gradient to better align with the long-horizon optimization trajectory. Extensive experiments demonstrate that our method consistently improves both convergence speed and final performance across multiple diffusion fine-tuning benchmarks, including subject-driven generation, controllable generation, and deblurring, achieving not only performance improvement but also earlier stages of convergence over baseline LoRA and other spectral-init methods.
[LG-13] Where to Compute and How to Interact: Operator-Readable Adaptation with Gauge-Aware Transport
链接: https://arxiv.org/abs/2609.15620
作者: Zixuan Shen,Quanxu Wan,Bingchuan Wang,Zhi Wang,Biao Luo
类目: Machine Learning (cs.LG)
*备注:
Abstract:Adaptive meshes enable neural operators for partial differential equations (PDEs) to allocate spatial samples and computation according to local physical structures. Existing approaches, however, mainly address where to compute, with less attention to how information should interact after node relocation. Mesh adaptation changes local sampling scales, neighborhood structures, and geometric contexts, so representations formed at different nodes may not be directly comparable. Direct aggregation can therefore entangle physical variation with discretization-induced representation variation. Because allocation and interaction are jointly optimized through the same output objective, their individual roles are also difficult to distinguish from final errors alone. We introduce operator readability, requiring an adaptive operator to account for and test why computation is allocated to particular locations and how representations interact under the resulting nonuniform discretization. Based on this principle, we propose the Gauge-Aware Adaptive Mesh Neural Operator (GA-AMNO). Physics-informed adaptive allocation answers where to compute, while geometry-conditioned low-rank Gauge transport maps source features into target representation contexts before aggregation, answering how to interact. This makes mesh-to-solver information exchange inspectable and intervenable. We establish sufficient conditions for representation-consistent aggregation and analyze approximate transport errors and continuity under topology-preserving mesh deformations. Experiments on five PDE benchmarks demonstrate improved predictive accuracy, while controlled interventions and geometric-mismatch analyses verify the roles of allocation and interaction and show that Gauge transport improves cross-discretization representation compatibility under strong geometric mismatch.
[LG-14] Bayesian Optimisation Using Product-of-Experts Gaussian Process Models with Uncertainty Calibration
链接: https://arxiv.org/abs/2609.15555
作者: Yean Hoon Ong
类目: Machine Learning (cs.LG)
*备注:
Abstract:Bayesian optimisation (BO) typically relies on a single global Gaussian process (GP) model as its surrogate model. However, GP regression has cubic computational complexity in the number of training data points, limiting its applicability to large-scale optimisation problems. The product-of-experts Gaussian process model with uncertainty calibration (GP-pro-c) mitigates this limitation by combining multiple local GP experts, enabling improved uncertainty quantification, reduced computational cost, and preservation of global correlations. Despite these desirable properties, the use of GP-pro-c in BO has not been thoroughly studied. This paper introduces BO-pro-c, a Bayesian optimisation algorithm that uses GP-pro-c as its surrogate model, and evaluate its performance across a diverse range of BO settings. Experimental results suggest that BO-pro-c maintains competitive optimisation performance while achieving a 0.9% reduction in simple regret and a 39.4% reduction in computational overhead relative to a BO algorithm based on a single global GP model.
[LG-15] he Token Before the Value Is the Key: How Hybrid Architectures Organize Induction Circuits
链接: https://arxiv.org/abs/2609.15545
作者: Ke Cheng,Xin Xu,Yixiao Chen,Lei Xin,Jianbo Zhao,Fanhu Zeng,Yue Liu,Jun Zhang,Jie Jiang
类目: Machine Learning (cs.LG)
*备注: 30 pages, including references and appendices
Abstract:Hybrid language models can improve capability as well as efficiency, raising the question of how architectural complementarity becomes learned computation. We examine the established induction roles of Carrying predecessor information, Matching a source by content, and Copying its value. How are these position-sensitive and content-based computations allocated across heterogeneous layers? We introduce layer-type-agnostic paired probes that track Carrying and Matching through a common block-update interface. In recurrent–global and local–global hybrids, Carrying concentrates in efficient layers and Matching in global receivers. The measured local contribution concentrates on lag one: the token immediately before the historical value. Changing predecessor support through lag-one masking, convolution removal, or early learning-rate reduction can relocate Carrying and Matching between stages. Source-key restoration and fixed-value selection trace the receiver’s dependence on the prepared source. These interventions also change natural-text recall, with outcomes depending on configuration and target. Varying local windows and induction-enriched training text changes the early development of functional Carrying and Matching, connecting architectural priors and training evidence to formation timing. Together, the probes and interventions shift the explanatory focus upstream: the organization of Matching follows how Carrying is learned. The token before the value provides a concrete link between a hybrid’s architecture, circuit development, and recall. Code is available in this https URL.
[LG-16] Strong and Compact Policies for Submodular Markov Decision Processes via LP-Based Submodular Orienteering
链接: https://arxiv.org/abs/2609.15539
作者: Lars Rohwedder,Rico Zenklusen
类目: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
*备注:
Abstract:Finding policies for Markov Decision Processes (MDPs) is a central problem in areas such as Reinforcement Learning and Operations Research. Here, we have to repeatedly choose an action that should be performed by an agent. Depending on the action and the current state of the agent, the agent collects a reward and randomly transitions into a new state. The goal is to maximize the reward in expectation over a finite time horizon of length H . We consider a recently introduced variant that generalizes the traditionally additive reward function in the model to a monotone submodular one, which allows for capturing a range of interesting applications. Without the stochastic component, this problem is equivalent to the Submodular Orienteering problem, where the goal is to find an s - t walk in a directed graph maximizing a monotone submodular function under a length constraint. We present a novel LP-based algorithm for Submodular Orienteering using ideas from the Sherali-Adams hierarchy and Round-or-Cut. Our guarantees are comparable to the known quasi-polynomial time logarithmic approximation for Submodular Orienteering, but also extend to the setting of Submodular Markov Decision Processes. In the polynomial time regime, we present an O(n^\varepsilon) -approximation (and O(H^\varepsilon) for Submodular MDPs) for every \varepsilon 0 , where n is the number of vertices, which was unknown even for Submodular Orienteering. Prior to our work, the best known approximation guarantee for Submodular MDPs had an approximation ratio linear in H . Beyond these algorithmic results, our methods reveal a trade-off between the approximation guarantee and the number of previously visited vertices on which an agent conditions its decision. Subjects: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG) Cite as: arXiv:2609.15539 [cs.DS] (or arXiv:2609.15539v1 [cs.DS] for this version) https://doi.org/10.48550/arXiv.2609.15539 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-17] Beyond Noise: Understanding and Overcoming Temperature Effects in Analog DNN Inference ECML KDD
链接: https://arxiv.org/abs/2609.15527
作者: Niklas Summ,Xiao Wang,Hendrik Borras,Bernhard Klein,Holger Fröning
类目: Machine Learning (cs.LG)
*备注: Published at the ECML PKDD Conference 2026, at the 7th Workshop on IoT, Edge, and Mobile for Embedded Machine Learning
Abstract:The energy efficiency of analog computing makes it one of the most promising candidates for deploying resource-intensive machine learning workloads on constrained platforms such as mobile and embedded devices. However, analog accelerators are inherently susceptible to noise and non-idealities arising from physical component variations, whose behavior is further sensitive to environmental factors. These effects can significantly degrade inference accuracy. In this work, we conduct a comprehensive experimental study on a representative example of analog hardware to investigate the impact of temperature. We first characterize the behavior of stochastic and systematic non-idealities across a range of operating temperatures. Following this, we compare a set of simulation-based and hardware-based mitigation strategies aimed at improving robustness against temperature-induced performance degradation. Our results suggest that temperature-induced degradation is driven primarily by systematic non-idealities rather than stochastic noise alone. Noise-aware training improves robustness, while hardware-in-the-loop training and temperature-aware calibration provide the strongest accuracy retention across varying thermal conditions.
[LG-18] End-to-End Verifiable and Robust Federated Learning
链接: https://arxiv.org/abs/2609.15521
作者: Doryan Lesaignoux,Enrique Mármol Campos,Gabriele Spini,José L. Hernández-Ramos,Stephan Krenn
类目: Machine Learning (cs.LG)
*备注: Conference, 21 pages, 9 figures, 5 tables
Abstract:Federated learning enables multiple parties to train a shared model without centralizing raw data with the help of an aggregator, but introduces integrity risks once participants or infrastructure are not fully trustworthy. Two requirements are particularly important: robustness to poisoned or Byzantine client updates, and verifiability of the aggregator so that clients or third parties can audit the reported aggregation without learning individual updates. Existing work has largely treated these goals separately, and efficient public verifiability for robust, outlier-excluding aggregation remains limited. We present a verifiable federated learning protocol that makes a robust aggregation pipeline publicly auditable. Our design combines cryptographic commitments with non-interactive zero-knowledge proofs to certify both (i) cosine-similarity-based outlier exclusion and (ii) aggregation over the selected set, without revealing individual client updates to verifiers. In experiments under representative poisoning attacks, our method maintains high accuracy, with an average accuracy loss below 4% across the evaluated configurations, while keeping verification overhead practical: proof artifacts can be generated and verified within minutes at the scale studied. In summary, our results show that robust outlier exclusion and public verifiability can be jointly achieved in a federated learning setting.
[LG-19] Same path different: a mechanistic comparison of looped and stacked transformer encoders on 12-lead ECG
链接: https://arxiv.org/abs/2609.15498
作者: Pawel Olszowiec,Michal Byra,Grzegorz Gruszczynski,Grzegorz Stefanski,Alberto Presta
类目: Machine Learning (cs.LG); Dynamical Systems (math.DS)
*备注: 5 pages, 7 figures
Abstract:Recurrent Transformers reusing their weights rather than stacking L distinct layers are becoming widely adopted due to their parameter efficiency [1,2,3]. However, the exact representational and dynamical differences between looped and stacked architectures remain uncharacterized. This paper presents a controlled study on the example of bViT model [1] applying one weight-tied block L times. We train two models: bViT and standard ViT [4] on 12-lead electrocardiogram (ECG) classification tasks from the PTB-XL dataset under identical training protocols. Despite an 8.9\times parameter reduction, bViT achieves accuracy parity with ViT. Geometric similarity metrics demonstrate that both architectures construct comparable latent representations in an equivalent canonical order. Crucially, their dynamics differ: bViT exhibits smaller step sizes and inter-patient sensitivity, as well as near-neutral behavior away from the data manifold, whereas ViT exhibits collapsing dimensionality of representations and out-of-distribution feature expansion.
[LG-20] Rotation-Based Subspace Tracking for Robust Kernel PCA on Streaming Data
链接: https://arxiv.org/abs/2609.15488
作者: Kris Lokere,John Fossaceca
类目: Machine Learning (cs.LG)
*备注: 7 pages, 3 figures
Abstract:Machine learning models process large amounts of data, and Principal Component Analysis (PCA) is a widely used technique to reduce the dimensionality of the data and extract useful features. In practice, datasets often change over time (data drift) and/or arrive one sample at a time (streaming data), making it infeasible to process the entire dataset at once in batch mode. Real-world data also often contains nonlinear patterns, which traditional PCA cannot extract. Kernel PCA addresses this by implicitly mapping samples into a Reproducing Kernel Hilbert Space (RKHS). Raw data also often contains outliers, which can have an outsized effect on the estimated subspace unless the algorithm is made robust. However, existing online robust kernel PCA algorithms are designed to converge to a subspace that is assumed to be fixed, and gradient-descent-based updates lose their effectiveness at tracking further changes once this initial alignment is achieved. This paper introduces a rotation-based update mechanism, which updates the subspace estimate by rotating it toward each new incoming feature vector in Reproducing Kernel Hilbert Space, rather than relying on gradient descent alone. We present two complementary rotation strategies, and show that the extent of rotation can be moderated by a robust influence function to mitigate the effect of outliers. Through experiments on synthetic streaming data with a known ground-truth subspace, we show that per-sample rotations converge faster than gradient descent alone, demonstrating an effective mechanism for dynamically tracking a nonlinear subspace in streaming data.
[LG-21] Data-driven Prediction of Satellite-observed Avalanche Activity from Snowpack Simulations
链接: https://arxiv.org/abs/2609.15485
作者: Jakob Grah,Filippo Maria Bianchi,Bert Kruyt,Karsten Müller
类目: Machine Learning (cs.LG)
*备注: Proceedings paper for ISSW2026. 7 pages, 4 figures
Abstract:Avalanche forecasting requires knowledge of snowpack conditions and recent avalanche activity, but field observations are sparse across large mountain regions. We explore whether SNOWPACK simulations can predict avalanche activity mapped by synthetic aperture radar (SAR). We compiled five winters of Sentinel-1 avalanche detections across Norway and parts of Sweden, alongside SNOWPACK simulations forced by numerical weather predictions on a 20 x 20 km grid at different elevations and predefined slope angles. A transformer used five days of SNOWPACK outputs to predict the following day’s SAR-detected Avalanche Activity Index (SAR-AAI). This index weights larger debris more heavily, spreads detections across possible occurrence dates and normalises by modelled runout area. The model was trained on four winters and evaluated on one validation winter. Regional mean predicted and reference SAR-AAI correlated at r = 0.803 after averaging over complete six-day periods. The model followed broad changes in time and space but produced smoother predictions and underestimated the strongest activity. Agreement at the 20 km cell scale was weaker (r = 0.549) after the same averaging. These results come from a single training run without evaluation on an untouched winter. Satellite observations also contain missed and false detections and uncertain timing. The results therefore do not establish operational forecast skill, but suggest that regional SNOWPACK simulations contain information about broad variations in satellite-observed avalanche activity.
[LG-22] GSLAD: Prototype-Regularized Graph Structure Learning for Multivariate Time Series Anomaly Detection
链接: https://arxiv.org/abs/2609.15483
作者: Zepeng Zhang,Fuad Khuri,Keivan Faghih Niresi,Olga Fink
类目: Machine Learning (cs.LG)
*备注:
Abstract:Unsupervised multivariate time series anomaly detection methods typically identify anomalies through forecasting, reconstruction, or representation discrepancies. However, industrial faults may first alter inter-variable structural patterns while individual trajectories remain close to normal, resulting in weak anomaly signals. In this paper, we propose GSLAD, a prototype-regularized graph structure learning framework that uses structural deviations for anomaly scoring. GSLAD adopts a two-phase training strategy. First, a condition-aware graph learner and a graph-based forecaster are optimized with predictive supervision. The inferred normal graphs are then clustered into multiple structural prototypes representing different normal operating regimes, with edge-wise variability characterizing structural uncertainty. Deviations from these prototypes regularize the graph learner in the second phase, encouraging stable and regime-specific structural patterns. During inference, uncertainty-normalized structural deviation is combined with predictive deviation for anomaly scoring. Experiments on four industrial benchmarks demonstrate strong overall performance of GSLAD and confirm the effectiveness of structural deviation for anomaly detection and diagnosis.
[LG-23] Single-condition neural solvers encode transferable response spaces for parametric differential equations
链接: https://arxiv.org/abs/2609.15432
作者: Wenbo Cao,Weiwei Zhang
类目: Machine Learning (cs.LG); Computational Physics (physics.comp-ph)
*备注:
Abstract:Operator learning for parametric partial differential equations (PDEs) typically builds global models over prescribed domains, requiring cross-condition data or costly physics-constrained training. Here we show that the output Jacobian of a neural solution model trained at one condition defines a reusable response space for cross-condition solution variations. We introduce Linearized Subspace Transfer (LST) to exploit this space and recover target solutions by minimizing the target PDE-system residual over response-space coordinates. Because any single response space has finite coverage, Active Transfer Modeling (ATM) uses post-transfer residuals as coverage indicators to selectively acquire response spaces from additional single-condition models. Across six systems, single-condition response spaces supported cross-condition transfer, with enrichment improving accuracy when added spaces expanded representation capacity. Relative to evaluated physics-informed operator baselines, ATM reduced error and offline construction cost, with orders-of-magnitude accuracy gains in representative cases and millisecond-to-second target adaptation. These results establish neural solvers as reusable local parametric models.
[LG-24] A Game-Theoretic Framework for Incentive-Compatible AI training Under Renewable-Energy Constraints
链接: https://arxiv.org/abs/2609.15389
作者: Konstantinos Varsos,Ramin Khalili,Adamantia Stamou,George D. Stamoulis,Vasillios A. Siris
类目: Emerging Technologies (cs.ET); Machine Learning (cs.LG)
*备注:
Abstract:As artificial intelligence systems increasingly rely on distributed and collaborative training, the energy footprint of these processes becomes a shared responsibility. Modern AI training often unfolds across heterogeneous compute nodes-ranging from cloud clusters to edge devices-whose energy availability is spatially and temporally variable. At the same time, renewable energy grids experience growing levels of excess generation, creating opportunities to align computational workloads with low-carbon energy supply. In this work, we develop a game-theoretic model of carbon-aware AI training in which autonomous agents strategically choose whether to participate and how intensively to train under limited renewable energy availability. Each agent balances diminishing learning returns, rewards for remaining within green-energy budgets, and penalties for grid consumption. While our framework applies broadly to distributed AI training, we examine Federated Learning as a representative case study due to its decentralized structure and flexible scheduling. We analyze equilibrium existence, efficiency, and adaptive dynamics, and provide simulation evidence that appropriately designed incentives can eliminate grid-based energy usage while preserving model performance. Our findings demonstrate how incentive-compatible training mechanisms can enhance energy efficiency and sharply reduce carbon emissions under renewable-energy constraints.
[LG-25] Representing Clinical Conditions on Vital Signs from Healthy Individuals using Latent Modeling
链接: https://arxiv.org/abs/2609.15379
作者: Rafael Pina,Varuna De Silva,Mindula Illeperuma
类目: Machine Learning (cs.LG)
*备注: Accepted to the 2025 25th International Conference on Digital Signal Processing. 5 pages, 4 figures
Abstract:Machine learning can be crucial to help scale complex signal processing applications in scenarios such as healthcare. However, these machine learning models need rich datasets to be trained and there are often cases where it is not possible to access representative datasets. In this paper, we propose a deep generative model based on conditional variational autoencoders with the objective of augmenting the vital signs of healthy individuals in a way that mimics the patterns of a certain clinical condition. More specifically, we use a publicly available ICU (Intensive Care Unit) dataset to train our model and then evaluate it using the vital data that we have collected from healthy individuals. Our results demonstrate that the proposed model can not only learn the underlying dynamics of the ICU data but, more importantly, can reshape our collected data from healthy individuals in a way that is aligned with the vital signs of a certain clinical condition. We propose a distance metric that shows how our model can generate samples that are more aligned with the intended clinical labels when compared to the tested baselines.
[LG-26] Admissable: Training Reinforcement Learning Agents against Adversarial Missingness
链接: https://arxiv.org/abs/2609.15297
作者: Paul Stahlhofen,Luca Hermes,Tim Kochs,Markus Vieth,Barbara Hammer
类目: Machine Learning (cs.LG)
*备注:
Abstract:In order to make Reinforcement Learning algorithms applicable in real world scenarios, safety must be ensured even under adverse operating conditions. In this work, we consider the challenge of adversarial feature missingness: a scenario in which an adversary occludes features from the agent’s observation in order to reduce performance as much as possible. We formally define adversarial missingness for Reinforcement Learning and compare it to the related concepts of \ell_\infty -norm bounded adversarial perturbations and learning with missing data. We develop an adversarial training algorithm and show its effectiveness in increasing robustness against adversarial missingness on three MuJoCo benchmark environments. Compared to a baseline trained with random uniform missingness, our method achieves better robustness on all three tasks.
[LG-27] Impute-EM: Native Mixed-State Diffusion Models for Heterogeneous Data Imputation
链接: https://arxiv.org/abs/2609.15284
作者: Sergei Kholkin,Kirill Sokolov,Dmitry Baranchuk,Evgeny Burnaev,Alexander Korotin
类目: Machine Learning (cs.LG)
*备注:
Abstract:Missing values are ubiquitous in heterogeneous data mining, where numerical, categorical, and binary variables often coexist. Many imputation methods, especially diffusion-based ones, treat discrete variables through continuous surrogates such as one-hot relaxations rather than modeling them natively. This creates a mismatch between the model state space and the mixed discrete and continuous structure of the data. We propose Impute-EM, an Expectation Maximization style framework that alternates between imputing missing entries with the current model and refitting a diffusion backbone on completed data. We instantiate Impute-EM with native mixed-state diffusion backbones for heterogeneous data, combining Gaussian and masked categorical components without one-hot relaxations. In exact settings, we characterize the update and show that the observed mask-indexed marginals match the targets at the limit, while making explicit that the full data distribution is generally non-identifiable from incomplete observations alone. Empirically, Impute-EM delivers the best distributional fidelity on mixed-type tabular imputation, on which downstream modeling relies, with text imputation serving as a controlled validation of the native discrete backbone.
[LG-28] Draining Fictitious Knots: Restoring Distance-Awareness Guarantees for High-Dimensional Spline Networks
链接: https://arxiv.org/abs/2609.15274
作者: Masoud Ataei,Mohammad Javad Khojasteh,Vikas Dhiman
类目: Machine Learning (cs.LG)
*备注:
Abstract:Kolmogorov-Arnold Networks (KANs) with spline activations have recently shown promise for interpretable function approximation. Distance-Aware Error for Kolmogorov Networks (DAREK) introduces a computationally efficient bottom-up approach to uncertainty quantification by equipping KANs with distance-aware error bounds; yet, in high-dimensional settings, the theoretical guarantees can be weakened by the emergence of fictitious knots. Inspired by the Kolmogorov-Arnold representation theorem, DAREK adopts a componentwise formulation in which each input dimension is treated separately; as a result, induced knot locations may appear in the combined input space without corresponding to actual training data. These fictitious knots mislead the DAREK uncertainty estimator into reporting low uncertainty far from any real observation, violating the distance-awareness guarantee. We identify this failure mode precisely, characterize its geometric structure, and propose a drainage uncertainty mechanism that restores distance-awareness by constructing a monotonically decreasing uncertainty path from any fictitious knot region toward the nearest real knot. The proposed drainage method provides a practical heuristic correction that mitigates the fictitious-knot failure mode while restoring theoretical distance-awareness in high-dimensional settings. Experiments on a 2D synthetic benchmark and a 100-dimensional face dataset show that drainage raises sampled distance-awareness (SDA) from 85% to 98-99%, matching Gaussian processes at lower computational cost.
[LG-29] Learning CNF Formulas from Uniform Random Solutions: Near-Tight Sample Complexity for Valiants Algorithm
链接: https://arxiv.org/abs/2609.15268
作者: Weiming Feng,Yixiao Yu,Yiyao Zhang
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS)
*备注:
Abstract:We revisit Valiant’s algorithm (Commun. ACM’84) for learning n -variable CNF formulas with clause size k and variable degree d from i.i.d. uniform random solutions in the local lemma regime. For fixed t\geq1 , under k\gtrsim(1+1/t)\log d , Valiant’s algorithm achieves total variation error \varepsilon with \widetildeO(n^\lceil t \rceil/\varepsilon) sample complexity. For t1 , we prove a matching lower bound for Valiant’s algorithm. At t=1 (covering 0t1 ), we show Valiant’s algorithm has optimal sample complexity up to logarithmic factors by an information-theoretic lower bound \widetilde\Omega(n/\varepsilon) .
[LG-30] BioDCASE: Active Learning for Bioacoustics ECCV
链接: https://arxiv.org/abs/2609.15255
作者: Ben McEwen,Rupa Kurinchi-Vendhan,Shiqi Zhang,Lukas Rauch,Marek Herde,Sara Beery
类目: Machine Learning (cs.LG); Sound (cs.SD)
*备注: This paper summarises the BioDCASE Active Learning for Bioacoustics data challenge. This paper was reviewed and accepted to the non-archival track of the ECCV Computer Vision for Ecology workshop via OpenReview
Abstract:Ecological monitoring increasingly relies on machine learning models, whose performance depends on the quality and quantity of labelled data. However, obtaining these labels is costly, particularly in passive acoustic monitoring, where vast amounts of data are collected but only a small proportion can feasibly be annotated. Active learning addresses this bottleneck by prioritizing which samples should be labelled. However, progress is difficult to measure, because published methods are evaluated under different models, budgets, evaluation metrics and datasets. To address this challenge, we present the 2026 Active Learning for Bioacoustics BioDCASE challenge: a systematic evaluation of sampling methods designed to identify effective AL strategies. Participant methods were evaluated across four subsets composed of terrestrial and marine data. Across ten proposed sampling methods from seven teams, the top-ranked method achieved an area under the learning curve 26.4 % higher than random sampling at the same annotation budget, averaged over four data subsets. Significant variation in performance was observed across subsets, with the top-performing submission achieving a 67.1 % gain for the HSN subset over random sampling and a gain of 8 % for the ATBFL subset. Top-ranking submissions combined multiple acquisition signals, and diversity-based selection outperformed pure uncertainty sampling. Furthermore, there is evidence that transitioning from diversity-based to uncertainty-based selection and explicitly reducing redundancy within acquisition batches improve model training. There is also initial evidence that larger acquisition batch sizes may be increasingly beneficial later in the labelling process.
[LG-31] Bandits with Probing: Optimal Regret and the Limits of Winner Feedback
链接: https://arxiv.org/abs/2609.15248
作者: Yongjie Guan
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Machine Learning (stat.ML)
*备注: 37 pages, 2 tables
Abstract:A learner probes at most k of n arms each round, receives the maximum of their rewards in [0,1] , and competes with the best fixed arm. When does the probing advantage pay for learning? We determine two minimax laws. Under independent stochastic rewards with winner feedback (the maximum and a winning label), or on arbitrary fixed sequences given a single signed contrast between block maxima, the minimax regret has order \Phi_n,k(T)=\min\fracn-knT,\fracn-kk\ , 2\le kn . Under winner feedback, both arbitrary joint i.i.d. rewards and fixed sequences have minimax regret of order R_n,k(T)=\fracn-kn\min\T,\fracn+Tk,\sqrt\fracnTk\ . Both laws have universal constants and anytime upper bounds. The first reduces regret to a pure coverage cost: same-round contrasts absorb the stability cost, and independence permits exact resampling whose gains fund sample advancement. The second adds a learning cost that becomes comparable to coverage at horizon n ; beyond nk , numerical maxima improve over labels alone. The lower bound allows every adaptive action size.
[LG-32] MAST: Label-Efficient Robust and Generalizable Sound Detection for Biodiversity Monitoring via Masked Audio Pretraining and Self-Training
链接: https://arxiv.org/abs/2609.15221
作者: Tianyi Xu,Daniel Pimentel-Alarcón,Zuzana Buřivalová,Claudia Solís-Lemus
类目: ound (cs.SD); Machine Learning (cs.LG)
*备注: 29 pages, 8 figures
Abstract:Passive acoustic monitoring can measure biodiversity at larger scales, but time–frequency annotation of animal vocalizations is expensive, site-specific, and difficult to sustain at scale. We present a label-efficient sound detection framework that combines masked audio pretraining with a lightweight detector on mel spectrograms, then further improves robustness through iterative self-training on unlabeled audio. We first pretrain a ViT-based encoder on unlabeled recordings via masked reconstruction and transfer the encoder to a detection backbone. To better separate animal sounds from confounding background, we add a box-level contrastive loss that pulls matched event regions together while pushing noisy negatives apart. We then apply a two-stage pseudo-labeling curriculum to exploit large unlabeled pools without additional annotation. We evaluate the performance on two ecologically distinct domains: tropical rainforest soundscapes (Indonesia) and bird vocalizations in Mediterranean habitats (Spain). On both domains, masked audio pretraining and contrastive learning consistently improve time–frequency detection under temporal and cross-site distribution shift, and self-training yields further gains in out-of-distribution performance. On the rainforest domain, MAST with self-training achieves +0.22 mAP and +0.24 F1 over the strongest baseline under cross-site shift. On the bird domain, self-training achieves +0.12 mAP and +0.10 F1 over the strongest baseline under cross-site shift. Overall, our results show that MAST can effectively extend self-supervised audio representations from clip-level tasks to robust box-level localization across diverse bioacoustic settings, providing a practical path for biodiversity monitoring with limited labels.
[LG-33] Convergence rates for generative drifting flows: fixed-scale obstructions and multihead acceleration
链接: https://arxiv.org/abs/2609.15193
作者: Arthur Stéphanovitch,Eddie Aamari
类目: Machine Learning (cs.LG)
*备注:
Abstract:Drifting models offer a promising route to faster generative AI: they perform gradual transport during training, while generating new samples in a single step. This paper asks whether the underlying drifting process can converge rapidly to a target distribution under ideal conditions, before finite-data or optimization effects are introduced. We show that its convergence rate depends critically on how it handles spatial scale. With a single fixed resolution, fine-scale features of the target can become nearly invisible, leading to extremely slow convergence. We introduce a multihead approach that combines scale-normalized information across a continuum of resolutions. We prove that this multihead approach restores exponential convergence near standard reference distributions. These results identify fixed resolution as a key bottleneck and provide a simple route to faster one-step generative models.
[LG-34] Rethinking Correctness for Uncertainty Estimation in Clinical Prediction with Vision-Language Models
链接: https://arxiv.org/abs/2609.15180
作者: Mingcheng Zhu,Jinning Liang,Tingting Zhu
类目: Machine Learning (cs.LG)
*备注: 5 pages, 3 figures
Abstract:Vision-language models are increasingly explored for clinical prediction from electronic health records and medical images, where identifying unreliable predictions is important for safe deployment. Uncertainty estimation (UE) enables detecting such predictions, but its evaluation depends on a correctness criterion that determines whether each model output is correct. If this criterion disagrees with human judgement or distorts downstream UE performance, conclusions about model reliability can be misleading. We introduce a two-axis framework that evaluates correctness criteria by their agreement with human judgements and fidelity to human-referenced UE performance. We assess eight criteria across three clinical prediction tasks and three models using 450 predictions annotated by two reviewers. Across the audited tasks, canonical exact matching (EM) achieved the highest observed human agreement and lowest UE distortion, while the BERT-based matching (BEM) and LLM-judge also showed strong human agreement. Across four UE methods and 23,254 clinical predictions, criterion choice changed error-detection AUROC by up to 0.146 and reversed the relative ranking of UE methods. The LLM-judge also selectively accepted invalid or uncertain outputs, accepting 16 of 30 such human-identified errors. These results demonstrate that correctness assessment is an integral component of clinical UE evaluation and should be validated before UE methods are compared.
[LG-35] Low-Dimensional Embeddings for Gaussian Kernels on Manifolds
链接: https://arxiv.org/abs/2609.15179
作者: Soumik Dutta,Kunal Dutta
类目: Computational Geometry (cs.CG); Machine Learning (cs.LG)
*备注: 57 pages, 3 figures
Abstract:The Gaussian kernel is a widely used similarity measure underlying kernel methods such as kernel PCA and spectral clustering, but computing Gaussian kernel distances for many pairs of points can be expensive. Using Random Fourier Features (RFF), Chen and Phillips [ALT 2017] showed that for points in a d -dimensional Euclidean ball in \mathbb R^N , t=\Omega((d/\varepsilon^2)\log(dR/\varepsilon)) features suffice to preserve all pairwise Gaussian kernel distances within a (1\pm\varepsilon) factor with high probability. We establish a uniform relative-error embedding theorem for the more general setting of an arbitrary positive-reach submanifold \mathcal M\subset\mathbb R^N of intrinsic dimension d . We show that t=O((d/\varepsilon^2)\log(\operatornamevol(\mathcal M)^2N^2d/(\operatornamevol(B_1^d(0))^2\operatornamerch(\mathcal M)^2d\varepsilon^2d+1\delta))) , or approximately O((d^2/\varepsilon^2)(\log N+\log(1/(\varepsilon\delta)))) , RFFs suffice, with probability 1-\delta , to preserve the Gaussian kernel distance between every pair of manifold points up to relative error \varepsilon . Thus the bound depends only logarithmically on the ambient dimension and on manifold parameters such as volume and reach, while retaining the 1/\varepsilon^2 Euclidean rate. We also prove a topological consequence: under the same RFF embedding, persistent homology is preserved in the sense that weighted Cech and Rips filtrations built from Gaussian kernel power distance are (1\pm\varepsilon_\star) -interleaved, where \varepsilon_\star accounts for both distance distortion and kernel-weight approximation.
[LG-36] mporal Self-Distillation: Faster Inference in Discrete Diffusion Language Models
链接: https://arxiv.org/abs/2609.15177
作者: Shijian Xu,Andrea Miele,Metod Jazbec,Volker Roth,Eric Nalisnick,Ilija Bogunovic
类目: Machine Learning (cs.LG)
*备注:
Abstract:Diffusion language models (dLLMs) promise fast inference by generating multiple tokens in parallel, but suffer severe performance degradation when parallel decoding is pushed too aggressively. We introduce Temporal Self-Distillation (TSD), a simple on-policy method that trains dLLMs for fast inference by distilling predictions across time. Specifically, TSD distills the model’s denoising distribution at earlier timesteps toward its distribution at the final timestep at which a token is committed. This encourages earlier predictions to better anticipate the model’s eventual output, enabling much more aggressive parallel decoding. Because its teacher signal comes from the model itself, TSD requires no offline teacher generation and applies seamlessly to both base and post-trained policies. Across seven benchmarks in mathematics, planning, and code, TSD substantially shifts the speed–quality frontier toward the low-compute regime. TSD thus provides a simple, single-stage approach to accelerating dLLMs, achieving speedups competitive with offline distillation while avoiding a complex two-stage pipeline.
[LG-37] Nearly Minimax-Optimal Regret for Linear Contextual Bandits with Arbitrary Adaptive Action Sets
链接: https://arxiv.org/abs/2609.15170
作者: Tianyuan Jin
类目: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT)
*备注:
Abstract:We study stochastic linear contextual bandits with arbitrary action menus that may depend on the fixed parameter and the interaction history. We establish matching upper and lower bounds, up to logarithmic factors. Let d be the dimension, K be the menu size, and T the time horizon. For 2\le K\le d , we prove an upper bound \widetilde O(K^1/4\sqrtdT) . When T\ge d^2 , we further prove a lower bound \Omega(K^1/4\sqrtdT) . Thus, for T\ge d^2 and 2\le K\le d , the upper and lower bounds match up to logarithmic factors, and the polynomial dependence on K is optimal. Compared with the previous \widetilde O(\sqrtdKT) bound, our upper bound improves the dependence on K by a factor of K^1/4 . For K\ge d , we prove an upper bound \widetilde O_d,T\left(\sqrtdT\min\sqrt d,(d\log K)^1/4\right) and a lower bound \Omega\left(\sqrtdT\min\left\sqrt d,\left(\fracd\log K\log(2d)\right)^1/4\right\right) . Here, \widetilde O_d,T omits logarithmic factors only in d and T . In particular, for polynomially large K\ge d , the upper and lower bounds both scale as d^3/4\sqrt T up to logarithmic factors, improving the standard \widetilde O(d\sqrt T) rate by a factor of d^1/4 . As K grows further, the regret smoothly recovers the d\sqrt T scale once \log K reaches order d . Subjects: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT) Cite as: arXiv:2609.15170 [cs.LG] (or arXiv:2609.15170v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.15170 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-38] Multi-source Transfer Learning of Time Series with a Shapelet-based Distance Measure
链接: https://arxiv.org/abs/2609.15148
作者: Jiseok Lee,Brian Kenji Iwana
类目: Machine Learning (cs.LG)
*备注: Accepted at Pattern Recognition
Abstract:Transfer learning is an effective technique for addressing data scarcity in deep learning for time series classification, but its success depends on the selection of source datasets. Conventional transferability estimation methods are often computationally expensive, as they require fully pre-training a model on each potential source dataset to assess its suitability. This paper introduces a novel, training-free source selection method named Shapelet Matching. Our approach first identifies discriminative shapelets from the target and potential source datasets. Then, Shapelet Matching quantifies dataset similarity by comparing the extracted sets of shapelets. To mitigate the risk of negative transfer from selecting an unsuitable single source, we introduce a multi-source transfer learning method. We select several source datasets based on their shapelet-based similarity scores, combine them into a single multi-source dataset, and use this aggregated dataset for pre-training. The model is then fine-tuned on the target task. We evaluated our method on 128 datasets from the UCR Archive using both temporal CNN and Transformer architectures. The empirical results demonstrate that our multi-source pre-training reduces the risk of negative transfer on average. Shapelet Matching achieves the strongest performance for the CNN backbone and remains competitive for patch-based Transformer architectures, while avoiding the cost of pre-training a separate model for every candidate source.
[LG-39] Omni-Streaming Thinking
链接: https://arxiv.org/abs/2609.15128
作者: Enjun Du,Siyi Liu,Ziyu Zheng,Jingyu Li,Yiwen Guo,Yongqi Zhang,Difan Zou
类目: Machine Learning (cs.LG)
*备注:
Abstract:Streaming omni-modal models must decide what and when to answer from the video chunks and synchronized audio observed so far. Visual cues often support an interpretation before an utterance or sound event is complete. If that interpretation enters memory as a fact, later reasoning can keep relaying it even after audio contradicts it. We call this failure premature cross-modal commitment. We propose Omni-Streaming Thinking (OST), which generates structured outputs that include evidence observed so far, forecasts of future evidence, and claims based on this evidence. Each claim is initially marked as pending and linked to a future verification interval. Audio and visual evidence are stored separately, and OST checks a claim against the evidence from the specified modality at the end of the verification interval. When contradictory evidence is detected, a refutation process reduces the influence of the claim and its dependent states, and then guides a state update using the new evidence. An answer gate decides whether the answer-critical claims meet the conditions for giving a response. Using a frozen Qwen3-Omni-30B-A3B-Instruct backbone with lightweight adaptation, OST outperforms the strongest open baselines on five streaming and audio-visual benchmarks by more than 10% relative on average. We also introduce OST-DiagBench, which holds video fixed and edits audio to test agreement, absence, contradiction, coexistence, and subtitle-speech conflict. OST reaches d-prime = 2.95, compared with at most 1.38 for open baselines, while reducing vision-induced auditory hallucinations.
[LG-40] mathbbSL(n) Representation Learning: An Intrinsic Mixed-Curvature Space with Higher Curvature Capacities and Deeper Order-Aware Composition
链接: https://arxiv.org/abs/2609.15083
作者: Xingrun Li,Yusuke Mukuta,Xin Yang,Yinyu Ye,Tatsuya Harada
类目: Machine Learning (cs.LG)
*备注: 35 pages, 8 figures
Abstract:Mixed-curvature representation learning seeks to capture rich geometric structures that cannot be adequately modeled by a single curvature regime. Existing approaches largely rely on product manifolds, which require manually specifying how different curvature spaces are combined and separate their curvature contributions across factors. We introduce the \mathbbSL(n) space, a representation geometry defined by the simple \det(A)=1 constraint and a left invariant Schatten- p Finsler structure. Despite this minimal construction, \mathbbSL(n) exhibits pointwise negative, zero, and positive flag curvature around a common flagpole, while its mixed-curvature and curvature-coupling capacities are asymptotically maximal relative to the intrinsic geometric upper bound. Beyond geometry, its noncommutative group structure provides inherent order sensitivity, and its non-nilpotent Lie algebra admits nonzero nested Lie brackets at arbitrary depth, enabling deep order-aware composition. Empirically, \mathbbSL(n) consistently outperforms a broad range of representation manifold baselines across graph benchmarks at different scales. It reduces average distortion over the strongest baselines by 44.3% on KEGG and 40.5% on HumanCyc, and improves Hits@20 by 42.8% on OGBL-PPA. Experiments on Flickr30k-Order further support its ability to capture higher order dependencies from ordered composition. Together, these results show how a seemingly simple structural constraint can yield unexpectedly rich geometry, capacity, and composition within a unified representation space.
[LG-41] ask-Distribution-Aware Counterweight Synthesis and Constrained Co-Design for Serial Manipulators
链接: https://arxiv.org/abs/2609.15082
作者: Mohammad Abbadi
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: 9 pages, 6 figures
Abstract:Passive counterweights are simple gravity compensators, but a counterweight selected from a single pose is not generally optimal for the configurations and tasks a manipulator actually executes. This paper develops a task-distribution-aware synthesis framework in which the operating distribution \rho(q) enters the design explicitly. For a counterweight moment p=m_c r_c with gravity torque -gp\phi(q) , the weighted mean-square residual gravity torque has the closed-form minimizer p^=E_\rho[\tau_g\phi]/(gE_\rho[\phi^2]) . If payload gravity torque is affine in payload mass, the optimum is also affine: p^(m_p,\rho)=p_0^*(\rho)+m_pK_p(\rho) . For fixed static moment, added counterweight inertia is I_c=pr_c while mass is m_c=p/r_c , so mass-radius selection is underdetermined unless physical constraints are specified. A recovered three-link manipulator is used as a case study. At r_c=0.20 m, zero-payload equivalent optima are 0.672 kg for uniform joint-space operation, 0.683 kg for approximately uniform task-space operation, 0.713 kg for a representative pick-and-place family, and 0.952 kg for a high-gravity-biased distribution, a change of more than 40% caused solely by the operating distribution. Nondominated fronts show that preferred mass-radius pairs depend on declared engineering bounds. A rated-torque-referenced all-joint screen increases zero-payload feasible task-space coverage from 78.1% without compensation to 93.7% for the uniform-distribution design. A lumped point-mass trajectory study gives a provisional crossover from no counterweight at very aggressive motion to stronger compensation as motion slows. These actuator and dynamic results are engineering consequence studies rather than physical validation.
[LG-42] Ensemble-Conditioned Molecular Design
链接: https://arxiv.org/abs/2609.15077
作者: Ross Irwin,Alessandro Tibo,Jon Paul Janet,Simon Olsson
类目: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
*备注: Code available at: this https URL datasets and checkpoints available at: this https URL
Abstract:Molecular design is typically approached as a problem of finding molecules which can adopt a single bioactive conformation. In reality, molecules occupy a distribution over conformations, and many of the properties which determine whether a candidate is viable depend on that distribution rather than on any single conformer. We reframe molecular design as an optimisation of both the modes and properties of molecules’ conformational ensembles, where modes can be represented as shapes, pharmacophore profiles or protein pockets, and properties are aggregate scalars computed over the whole distribution. To realise this we introduce ensemble-conditioned guidance, a framework which conditions 3D molecular generative models on both axes simultaneously. Mode conditions are composed adaptively at inference by combining the vector fields produced under each condition. Conditions may be targeted or avoided, mixed across modalities and combined in arbitrary numbers, allowing a wide range of design tasks to be expressed with a single trained model. We introduce adaptive symmetry learning to allow conditions from different reference frames to be composed, and extend our generative framework to enable flexible-size generation. We evaluate on new benchmarks for multi-mode conditioning and ensemble property optimisation, and apply the framework to two practical drug discovery tasks, dual-target binder design and active-state-selective agonist design, where in both cases conditioning on the additional state improves the desired outcome over single-state conditioning.
[LG-43] Sensory Precision Inference for Multimodal Arbitration under Uncertainty
链接: https://arxiv.org/abs/2609.15065
作者: Tin Mišić,Takato Horii
类目: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
*备注: This work has been accepted for publication in the International Workshop on Active Inference (IWAI2026) proceedings. This is the Submitted Manuscript before any peer review. The reviewed and edited version will be available on Springer Nature Link
Abstract:Autonomous agents operating on multisensory data cannot assume that all sensory modalities remain consistently informative. In real environments, sensory streams are frequently corrupted by noise, missing data, or inter-modal incongruence, requiring adaptive arbitration between competing sensory hypotheses. While active inference provides a principled framework for uncertainty-guided inference, the role of dynamically inferred sensory precision in generative multimodal arbitration under sensory conflict remains comparatively underexplored. We propose a multimodal perceptual inference model in which latent beliefs and modality-specific sensory precisions are jointly updated through iterative free-energy minimization. In our proposed model, sensory precision dynamics not only reflect sensory uncertainty but actively shape the evolution of latent beliefs during multimodal conflict. In addition, we introduce a learned prior over sensory precisions that induces structured, class-dependent precision patterns and influences cross-modal inference dynamics. We evaluate the model using a synthetic multimodal MNIST dataset combining visual, auditory, and tactile representations of digit classes under controlled sensory noise and inter-modal incongruence. Results show that dynamic precision inference improves reconstruction robustness under corrupted sensory evidence, supports coherent latent inference from reduced sensory evidence, and enables stable arbitration between conflicting modalities. Furthermore, learned precision priors generate interpretable precision structures that shape inference dynamics and cross-modal latent structure. These findings support sensory precision inference as a mechanistic control process for adaptive multimodal belief formation under uncertainty, highlighting precision dynamics as a computational mechanism for robust and interpretable multisensory integration.
[LG-44] What Does an LLM Learn from Reinforcement Learning? A Mechanistic Interpretability Perspective with Fixed-SAE Track
链接: https://arxiv.org/abs/2609.15064
作者: Lingheng Du,Yiming Tang,Xufeng Duan,Dianbo Liu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Reinforcement learning (RL) is widely utilized in large language model training to improve targeted capabilities, yet how RL reshapes a model remains poorly understood. Prior attempts to explain how RL works largely offer behavioral perspectives, leaving open what RL gives a model at the representation level: can RL create genuinely novel features, and which existing features does it enhance or suppress? Recent developments in mechanistic interpretability suggest sparse autoencoders (SAEs) as a promising lens to decompose internal activations into human-interpretable features; however, they cannot be directly applied to tracking change across training. In this work, we introduce Fixed-SAE Track, a framework that trains one shared SAE per considered layer on activations pooled across the base model and all RL checkpoints, holding every feature direction fixed so that representation shifts are rigorously defined through the activations of interpretable SAE latents, including the detection of emerging novel features. Validated across multiple datasets and RL algorithms, we find that RL-induced drift is small, gradual, concept specific, and concentrated in late layers, mainly enhancing the sampling rates of a small set of ladder tokens, formatting scaffolding such as step breaks and answer delimiters, rather than reshaping problem content. Steering these features into the base model recovers around 80% of RL’s performance gain, suggesting that RL primarily elicits capabilities the model already possesses, much as steering does. We further design a synthetic benchmark with features known by construction to test whether RL can instill genuinely novel features. We believe Fixed-SAE Track provides a principled approach to tracking representation shifts and offers representational evidence for understanding how reinforcement learning changes the inner representation of LLMs.
[LG-45] Zero-SNR Analyticity of the Scalar MMSE Is Equivalent to Gaussianity
链接: https://arxiv.org/abs/2609.15048
作者: Yixing Zhang
类目: Information Theory (cs.IT); Machine Learning (cs.LG)
*备注: 33 pages, 1 figure
Abstract:Let Y_s=\sqrtsX+Z , where Z is standard Gaussian and independent of the real random variable X . We prove that, under the square-exponential moment condition \mathbbEe^\beta X^2\infty for some \beta0 , the scalar minimum mean-square error \operatornamemmse_X(s) is analytic at zero signal-to-noise ratio if and only if X is Gaussian, with constant random variables included as degenerate Gaussians. The proof converts estimation in the Gaussian channel into a backward heat flow acting on the moment-generating function M(z)=\mathbbEe^zX . Under the stated tail condition, every non-Gaussian input forces M to have a nonzero complex zero. We show that each zero cluster produces a finite singularity in its localized Borel transform at the action \xi=z_0^2/2 . After removing the action scale, the Borel coefficients have a nonzero n^-1/2 prefactor for a simple zero. A zero of multiplicity m\geq 2 splits according to the roots of a Hermite polynomial and instead contributes a prefactor n^-m/2e^r_m\sqrt2n . A finite-disc localization and relative-cycle continuation argument then show that at least one such singularity survives in the full Borel transform. Thus, for every non-Gaussian input in the stated class, the formal zero-SNR expansion is Gevrey-1 but divergent. Rational-MMSE rigidity and the analogous analyticity criterion for mutual information follow as corollaries. Comments: 33 pages, 1 figure Subjects: Information Theory (cs.IT); Machine Learning (cs.LG) Cite as: arXiv:2609.15048 [cs.IT] (or arXiv:2609.15048v1 [cs.IT] for this version) https://doi.org/10.48550/arXiv.2609.15048 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-46] Structured Features Overfit Where Random Features Grok
链接: https://arxiv.org/abs/2609.15047
作者: Chon-Fai Kam,Miloud Bessafi,Frederic Cadet
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 13 pages, 2 figures, 3 tables. Submitted to OPT 2026 (18th Annual Workshop on Optimization for Machine Learning)
Abstract:Xu, Vardi and Safran (ICML 2026) prove that over-parameterized ridge regression over an unstructured random Gaussian feature map groks, with the delay between memorization and generalization growing as 1/\lambda in the weight decay. We show that on a structured feature map the same delay does not appear. For a band-limited Fourier feature map over \mathbbZ_p^2 carrying a single-character target that lies inside the expressible class, enlarging the band at fixed positive weight decay drives peak held-out accuracy monotonically from 1.00 to 0.07 , with no memorize-then-generalize regime anywhere along the sweep. The degradation is not an interpolation effect. It sets in at capacity ratio q/n = 0.638 , far below the interpolation threshold, on separate grounds from the exact null space that appears above it. What does have a sharp boundary is the active support. Holding the nominal dimension fixed and masking the band back to 1089 active modes restores held-out accuracy of 1.000 with zero variance across seeds, while the full 4225 -mode band collapses to 0.185 . The number of active modes acts through the teacher-weighted spectrum of the empirical Gram matrix and not through the capacity ratio, which makes this a statement about feature geometry and not a restatement of double descent.
[LG-47] Shared KV Caching for Replicated 27B Inference: Correctness Failures and Performance Boundaries
链接: https://arxiv.org/abs/2609.15021
作者: Frank Li
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Performance (cs.PF)
*备注: Technical report. 8 pages, 4 figures, 4 tables
Abstract:Shared host-memory caching can avoid repeated prefill when a request moves between inference replicas. Its usefulness depends on both correct state transfer and lost prefix locality. We study two single-GPU 27B vLLM replicas sharing a 256 GiB LMCache pool. After adopting an existing packed-page patch, we isolate a raw-pointer fallback that omits the dependency on the current CUDA stream. Controlled byte tests fail under an imposed delay and pass when the dependency is restored; the existing mixed allocator provides a working deployment path. Full-pool allocation checks and service regression complete the validation. A four-block OFF-ON-ON-OFF comparison contains 768 measured requests within two block pairs. Median cross-replica time to first content token falls from 31.715 to 0.605 seconds at 128k input and from 92.047 to 0.790 seconds at 256k. Six-turn synthetic sessions alternating replicas improve by approximately 35% and 45% at initial contexts of 32k and 128k, while fixed placement shows little benefit. This engineering case study identifies practical validation steps and the locality conditions in which shared caching pays off.
[LG-48] HGTO: A Unified Graph-Based Physics-Informed Formulation for Structural Topology Optimization
链接: https://arxiv.org/abs/2609.15001
作者: Kangzheng Liu,Uday Kumar Punna,Leixin Ma
类目: Machine Learning (cs.LG)
*备注:
Abstract:Density-based topology optimization is typically structured as a nested sequence of material updates, structural analyses, and sensitivity assessments. While neural density parameterization and dual-field physics-informed approaches provide data-free alternatives, most existing methods represent density and displacement as coordinate fields and make limited use of the discrete relationships inherent in the finite element mesh. The present study introduces HGTO, a unified graph-based formulation that extends complete neural topology optimization from coordinate space to finite-element graph space. Element densities are parameterized on the element graph derived from the mesh, and the structural state is determined on the corresponding node–element hypergraph. Finite element kinematics, numerical quadrature, constitutive response, and force assembly remain explicitly defined operations within the differentiable computation. The material field and equilibrium state are therefore coupled through a common finite-element incidence structure. Numerical studies show compliance comparable to conventional density-based optimization at substantially lower computational cost than a representative coordinate-based dual-field neural method. The same coupled formulation accommodates high-resolution and irregular meshes, three-dimensional structures, finite deformation, and elastoplastic response.
[LG-49] Learning to Solve Stochastic Controls with Unknown Drifts and Running Rewards: Theory Algorithms and Convergence
链接: https://arxiv.org/abs/2609.14972
作者: Jin Ma,Gaozhan Wang,Jianfeng Zhang,Xunyu Zhou
类目: Machine Learning (cs.LG); Numerical Analysis (math.NA); Optimization and Control (math.OC)
*备注:
Abstract:We study continuous-time and possibly high-dimensional stochastic control problems where drift coefficients and running reward functions are unknown. Due to these missing model primitives, we take the exploratory, reinforcement learning (RL) framework of Wang, Zariphopoulou, and Zhou(2020) with relaxed controls and entropy regularization. The objective is to develop theoretically grounded, efficient and scalable RL algorithms to learn both the optimal value functions (which also solve the exploratory HJB equation) and optimal exploratory feedback control policies. When the diffusion coefficients do not contain control, we employ probabilistic representations of both the optimal value function and its gradient based on an auxiliary state process depending only on the diffusion part of the original dynamics. With a delicate analysis on some properly defined mappings and their fixed points, this leads to the introduction of our policy iteration algorithms and their convergence. We demonstrate the performance of our algorithms through various numerical examples. Finally, we study a special control-dependent diffusion case where probability representation of the Hessian is called for.
[LG-50] HiGFRL: Hierarchical Graph Fusion-Driven Reinforcement Learning for Dependency-Aware Task Scheduling in Heterogeneous Cloud
链接: https://arxiv.org/abs/2609.14968
作者: Tiangang Li,Shi Ying,Xiangbo Tian
类目: Machine Learning (cs.LG)
*备注: 16 pages, 13 figures
Abstract:Online scheduling of dependency-aware tasks in heterogeneous cloud clusters is a fundamental yet challenging problem due to the complex interplay between DAG topologies and multi-dimensional resource constraints. While DRL has shown promise, existing GNN-based approaches often struggle to efficiently model high-order topological dependencies and suffer from loose coupling between task and resource states, leading to myopic scheduling decisions. To address these limitations, we propose HiGFRL, a Hierarchical Graph Fusion-Driven Reinforcement Learning framework. HiGFRL constructs a novel three-level state representation comprising a Static Hypergraph, a Dynamic Global Graph, and a Local Bipartite Graph to explicitly model the interplay between task dependencies and real-time cluster dynamics. Specifically, we design a fusion-driven dual-network architecture to optimize RL decision-making, where a Context Fusion Allocator integrates local bipartite matching features with fused global context to execute precise task-to-node allocation, and a Global State Evaluator leverages the global dynamic graph representation to accurately estimate expected long-term cumulative reward. Furthermore, we incorporate a topology-prior-guided hybrid reward mechanism that distills static topological priors into the learning process to accelerate convergence. Extensive experiments using real-world Alibaba cluster traces demonstrate that HiGFRL significantly outperforms heuristics and DRL baselines. Specifically, in challenging large-scale high-load scenarios, HiGFRL reduces the Makespan by up to 32.55%, and optimizes the average task flow time and average task wait time by 13.58% and 13.79%, respectively. Experimental results confirm that HiGFRL not only significantly improves cluster throughput but also ensures superior QoS by substantially reducing queuing delays. Code Release:this https URL.
[LG-51] Cloud Workflow Scheduling Based on Graph Attention-Driven Hierarchical Reinforcement Learning
链接: https://arxiv.org/abs/2609.14952
作者: Zongjin Li,Shaohan Feng,Chunxi Yang,Wenbo Wang
类目: Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注: Paper submitted to IEEE Internet of Things Journal
Abstract:Dynamic cloud workflow scheduling must balance deadline satisfaction, container utilization, and energy consumption while dealing with stochastic task-execution speeds, placement-dependent communication, and coupled task and container decisions. Workflows are naturally modeled as directed acyclic graphs (DAGs), but conventional vector- or matrix-based states do not fully capture their dependency topology. To better represent task urgency and structural relationships, we assign predicted sub-deadlines to tasks and use a multi-head graph attention network (GAT) to extract dependency information from the evolving DAGs. Based on these representations, we develop a Graph Attention-Driven Hierarchical Reinforcement Learning (GA-HRL) framework and model the scheduling process as an event-driven hierarchical semi-Markov decision process (SMDP). Workflow arrivals and task completions trigger scheduling events. At each scheduling event, the Task Scheduling (TS) agent first processes the currently ready tasks by assigning them to admissible existing containers or requesting new ones. The requested containers are then processed by the Container Scheduling (CS) agent for host placement before the environment advances. The two agents are trained alternately using separate Proximal Policy Optimization (PPO). Experiments on the 2018 Alibaba cluster trace show that GA-HRL maintains competitive workflow success rate and, in settings where success is comparable, generally achieves higher container utilization and lower energy consumption. Under the largest speed variation, it trades a small success-rate margin for substantially lower energy. Simulation code is available at: this https URL.
[LG-52] ackling Failure Modes of PINNs and PIKANs Using Conflict-Free Gradients
链接: https://arxiv.org/abs/2609.14841
作者: Sidharth S. Menon,Irina Tezaur,Ameya D. Jagtap
类目: Machine Learning (cs.LG)
*备注: 45 pages, 31 figures
Abstract:Scientific machine learning methods such as physics-informed neural networks (PINNs) increasingly rely on domain decomposition for better scalability while solving partial differential equations (PDEs) over complex geometries, yet the resulting composite loss comprising residual, boundary, and interface terms is highly susceptible to conflicting gradients that degrade training. This work bridges domain decomposition with projection-based gradient surgery to systematically mitigate such conflicts in 2D and 3D settings. We evaluate two existing projection-based algorithms, PCGrad and ConFIG, and identify their performance degradation in specific scenarios such as 3D domains with multiple overlapping interfaces. To address this limitation, we propose Norm-PCGrad, a normalized variant that achieves state-of-the-art accuracy across a range of 2D and 3D domain decomposition problems. Across the benchmarks considered, Norm-PCGrad consistently achieves the lowest relative L_2 error compared to training without gradient surgery as well as to existing algorithms such as PCGrad and ConFIG, while incurring negligible additional computational overhead. To improve computational efficiency of domain decomposition frameworks such as Extended PINN (XPINN), we propose replacing vanilla PINNs in selected subdomains with separable architectures such as Separable PINN (SPINN), reducing the computational cost from quadratic (or cubic) to linear. We additionally demonstrate that gradient surgery extends to physics-informed Kolmogorov-Arnold Networks (PIKANs), yielding substantial accuracy improvements for 3D domain decomposition and confirming the generality of the proposed approach across network architectures.
[LG-53] Decision-Oriented Uncertainty Quantification for Risk Control in Earth System Spatiotemporal Foundation Models
链接: https://arxiv.org/abs/2609.14821
作者: Ji Lu,Huiran Duan,Bo Zhao,Xianglong Wang,Yiru Fang,Kuo Yang,Xiaoqin Feng,Jianping Gou
类目: Machine Learning (cs.LG)
*备注: Accepted to the 22nd International Conference on Advanced Data Mining and Applications (ADMA 2026)
Abstract:Earth system modeling is shifting from task-specific predictors toward foundation models with general spatiotemporal representation capabilities. Although these models can jointly encode dynamic Earth fields, external forcings, and static geographic context for multistep forecasting, accurate point predictions or statistically calibrated intervals alone are insufficient for high-impact applications such as extremeweather warning, flood control, renewable-energy dispatch, and emergency resource allocation. What matters in practice is whether predictive uncertainty can be translated into reliable decision risk under specific actions, loss functions, and risk preferences. We propose a decision-oriented uncertainty quantification framework for Earth system spatiotemporal foundation models. The framework produces predictive distributions of future states and uses a decision risk adapter to map forecast samples, decision context, and utility functions into action-conditional risks. A utility-aware calibration module further enforces reliability at the downstream decision-loss level rather than only at the forecast-value level. Calibrated risks are then used to select warning, dispatch, inspection, or resource-allocation actions. Compared with the strongest baseline, the proposed method reduces decision regret by 18.7%, lowers the missed-event rate from 14.2% to 9.1%, and improves expected utility by 11.6%, while maintaining 90.4% predictive coverage and reducing decision calibration error from 0.083 to 0.047. These results suggest that decision-oriented uncertainty quantification can improve the robustness and operational value of Earth system foundation models in risk-sensitive applications.
[LG-54] Privacy Preserving Gossip Learning
链接: https://arxiv.org/abs/2609.14778
作者: Erkan Bayram,Mohamed-Ali Belabbas,Tamer Başar
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注: Accepted for publication in IEEE CDC 2026
Abstract:We propose a decentralized privacy-preserving learning algorithm in which each agent holds a single private sample and a shared model. Samples are learned sequentially, and each update must preserve the endpoint mappings at previously learned samples while protecting private data. This gives each agent three roles: (i) a learner that updates the model parameters, (ii) a teacher whose sample is learned at the current iteration, and (iii) a protected agent whose sample has already been learned. We build on Tuning without Forgetting (TwF) method to preserve previously learned mappings and show that TwF provides an indistinguishability guarantee for the learner whenever the set of protected agents contains another sample with the same label. For the teacher, we formulate a minimax optimal control problem that models the differential privacy noise as a worst-case disturbance to prevent performance loss while maintaining the same level of privacy for the gradient. For the protected agents, we compute the projections locally and aggregate them using a private push-sum gossip protocol. We prove geometric convergence of the decentralized gossip algorithm and of the distributed projection for TwF.
[LG-55] A Personalized Dynamic Balance Evaluation Paradigm for Hip Exoskeleton-Assisted Walking under Unexpected Ground Perturbations IROS2026
链接: https://arxiv.org/abs/2609.14765
作者: Yun Chen,Oluwasegun T. Akinniyi,Qiang Zhang
类目: Robotics (cs.RO); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 8 pages, 6 figures. Accepted to the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026)
Abstract:Hip exoskeletons may improve recovery from unexpected gait perturbations, yet personalizing assistance remains difficult because balance is multidimensional and human-in-the-loop experiments are small-sample and noisy. We present a participant-specific composite balance cost that integrates seven biomechanical sub-metrics spanning margin of stability, center-of-mass dynamics, and whole-body angular momentum. The sub-metrics are converted to direction-aligned, dimensionless cost features, and nonnegative fusion weights are learned on the simplex. Coupled with an empirical-Bayes hierarchical model, the learned-composite selector estimates each tested condition’s posterior probability of being best, P(best), and a high-probability candidate set with size K_0.8 . The framework was evaluated with three participants walking at 1.1 m/s during unilateral belt-slip perturbations across 46 hip-assistance conditions. In the full-budget analysis (B = 4 repeats per condition), the selector concentrated 80% of the posterior probability within 1 to 5 of 46 conditions, compared with 2 to 12 for equal-weight fusion and 4 to 37 for principal component analysis fusion. This smaller candidate set could shorten personalization experiments and limit participants’ exposure to repeated perturbations in future studies. Selected-condition trials showed lower observed composite costs than no-torque trials, with nominal p 0.05 for P2 and P3. Leave-one-repeat-out refits yielded positive mean held-out rank correlations for all participants and moderate stability of the learned weights and candidate sets. These proof-of-concept results support participant-specific composite balance evaluation for candidate selection in perturbation-based human-in-the-loop experiments.
[LG-56] CCMAN: Cognitive Instability-Aware Cross-Modal Attention Network for Interpretable Temporal Biomarkers of Verbal Fluency Speech
链接: https://arxiv.org/abs/2609.14764
作者: Madhurananda Pahar,Caitlin Illingworth,Dorota Braun,Daniel Blackburn,Heidi Christensen
类目: ound (cs.SD); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
*备注:
Abstract:Early detection of cognitive decline from speech offers a scalable and non-invasive alternative to conventional clinical assessment. Verbal fluency tasks are particularly informative, but most automated approaches aggregate features across an entire recording, overlooking temporal speech dynamics. We propose the Cognitive Instability-Aware Cross-Modal Attention Network (CCMAN), a transfer learning framework that learns task-agnostic cognitive speech representations from multiple memory-probing tasks before fine-tuning on a minute-long semantic and phonemic verbal fluency task. CCMAN integrates semantic, acoustic, and linguistic information through bidirectional cross-attention, gated multimodal fusion, and transformer-based temporal modelling to derive interpretable biomarkers of cognitive decline. Experiments were conducted on 165.44 hours of speech from 843 participants (498 healthy controls, 245 with mild cognitive impairment, and 100 with dementia). CCMAN achieved Macro-F1 scores of 0.81 and 0.59 for binary and multiclass semantic fluency classification, and 0.77 and 0.53 for phonemic fluency, consistently outperforming strong static and temporal baselines. Statistical analyses showed that semantic drift variance and pause variance, but not mean semantic drift, were significantly elevated in both MCI and dementia relative to healthy controls, while pause duration increased progressively over the task with the steepest slope in dementia, supporting global and progressive temporal speech instability as interpretable biomarkers. Evaluation on the independent PROCESS-2 benchmark further demonstrated the generalisability of the proposed framework, improving the baseline Macro-F1 by up to 9%. These findings support temporal speech instability as a dynamic speech biomarker for robust, interpretable, and generalisable early detection of cognitive decline.
[LG-57] WaVeFuse: Regime-Adaptive Equity Index Forecasting via Channel-Wise Wavelet Denoising and Vertical Attention Fusion
链接: https://arxiv.org/abs/2609.14733
作者: Aashish Bohra,Vivek Vijay
类目: Machine Learning (cs.LG); Statistical Finance (q-fin.ST)
*备注: 40 pages, 13 Figures, 15 Tables, Preprint, Under Review
Abstract:Hybrid Deep Learning for equity index forecasting is limited by three problems: propagation of OHLCV noise into derived technical indicators (TIs), channel-indiscriminate multi-scale decomposition that conflates heterogeneous frequency signatures, and static multi-branch fusion that cannot adapt to market regime shifts. WaVeFuse addresses these limitations through a unified dual-branch architecture. Symlet-4 wavelet denoising (level 2, MAD soft threshold) suppresses microstructure noise in OHLCV. Seven low-lag TIs computed from denoised prices are encoded by a causal channel-wise continuous wavelet transform (Morlet, 32 scales) into a per-timestep scale-space matrix. A CNN-BiLSTM branch captures temporal dynamics, while a dual-layer Transformer (heads=4, dk in 16, 32) models inter-scale spectral dependencies, and their representations are integrated by a 2-token softmax gate Vertical Attention Fusion (VAF) that dynamically reweights branches as market regimes shift. Evaluated under walk-forward validation (WFV) on KOSPI, DAX, NYSE Composite, and Russell 2000 (2010-2023), WaVeFuse achieves R2 = 0.81-0.96 and directional accuracy 70.5-78.3%. It outperforms seven state-of-the-art models by 8.9-20.2% MAE across twelve dataset-period configurations. Diebold-Mariano statistics (4.62-10.38, p0.001) confirm superiority over a well-tuned XGBoost benchmark across four indices. Ablation verifies component-wise contributions. Under realistic backtesting with 10 basis point transaction costs, WaVeFuse’s directional strategy achieves a mean Sharpe ratio of 3.69 across four markets and limits maximum drawdown to 7.5% during the COVID-19 crash. With 152k parameters (0.68MB) and sub-1.3ms GPU inference, WaVeFuse delivers a computationally efficient, regime-robust framework suitable for research and decision-support deployment.
[LG-58] Are Gradient Boosting Models Suitable for Intermittent Demand Forecasting?
链接: https://arxiv.org/abs/2609.14718
作者: Vladislav Kislinskii,Mazhar Hameed
类目: Machine Learning (cs.LG)
*备注:
Abstract:Demand forecasting is critical in modern industry, offering opportunities to reduce costs and gain competitive advantage through improved inventory management. However, forecasting becomes particularly challenging for products with intermittent demand, where demand occurs infrequently and time series contain many zero observations. Such dynamics are common across diverse sectors, such as industrial organizations, consumer goods, aviation, automotive, and electronics. Motivated by these challenges, this paper explores the potential of gradient boosting models to improve forecasting performance. We evaluate statistical, specialized, machine learning, and ensemble approaches across multiple datasets. The results show that specialized methods achieve the strongest performance among individual models, while gradient boosting on its own tends to underperform. However, combining a machine learning model with a specialized approach improves forecasting accuracy by up to 10%, demonstrating that even simple ensembles can outperform single models. Overall, the findings highlight the value of combining machine learning with domain-specific forecasting techniques for intermittent demand.
[LG-59] Learning Metastable Dynamics
链接: https://arxiv.org/abs/2609.14712
作者: Rupak Majumdar,Mahmoud Salamati,Nikhil Singh,Sadegh Soudjani
类目: Performance (cs.PF); Machine Learning (cs.LG)
*备注: HSCC’26
Abstract:Metastability—a phenomenon where systems remain trapped in quasi-stable states before abruptly transitioning under rare perturbations—is ubiquitous in physical systems. Although metastability is a widely observed phenomenon, its identification and analysis present significant challenges. To address these challenges, we propose a novel framework for analyzing metastability using Koopman theory. We use a finite set of system trajectories to learn a representation of the dynamics that defines a latent space in which the system evolves linearly, thereby enabling a systematic characterization of metastable behavior through the spectral properties of the linear mapping. Empirical evaluations demonstrate that our approach is capable of anticipating metastable behavior significantly earlier than its actual manifestation, even with 10% of the simulation duration. Moreover, we establish that the dominant eigenvalue of the learned Koopman matrix in the latent space serves as a critical indicator for detecting metastability across both single-server and multi-server configurations.
[LG-60] GRPO-QM: Target Preserving Exploration for Quantum Tomography
链接: https://arxiv.org/abs/2609.14711
作者: Yufeng Wang,Parivesh Priye,Lu Wei,Haibin Ling
类目: Machine Learning (cs.LG); Quantum Physics (quant-ph)
*备注:
Abstract:Reward-based learning can alter the very posterior distribution that scientific inference aims to estimate. GRPO-QM sidesteps this by learning only an exploration strategy for a stated quantum-tomography posterior: a group-relative policy chooses among reversible physical moves, and an exact Metropolis correction ensures the posterior remains stationary once the policy is fixed. We then examine what learning contributes beyond physical proposal mechanisms and prior knowledge. Reconstruction comparisons suggest that most of the gains over the tested flows come from those two components rather than from learning itself, and a closed-form counterexample explains why: a reward tied to accepted motion can increase even when a physical observable remains highly correlated. Reward comparisons that control for initialization also show that aggregating across posteriors can invert the ranking that a reward implies within any single posterior. Lastly, on 45 enumerated posteriors with three training seeds each, exact and sampled gradients of the same trajectory objective reduce mean physical estimation variance by 12.63% and 6.32% relative to a tuned mixture, while averaging the trajectory score without rescaling its penalty yields only 0.43% . This indicates that objective scaling accounts for part of the disparity between sampled and exact training. The regained benefit is concentrated at four shots and flips sign at sixteen, so these exact, visible-bank diagnostics identify a concrete, reproducible failure mode in sampled training, the objective-scaling fix that addresses part of it, and the remaining gap.
[LG-61] An immune world model for multiscale forecasting and therapeutic hypothesis generation
链接: https://arxiv.org/abs/2609.14709
作者: Taoyong Cui,Xi Wang,Zonghang Li,Jinchao Ding,Lingsen You,Yuzhi Xu,Wanghan Xu,Fang Wu,Kejun Ying,Wanli Ouyang,Pheng Ann Heng,Ling Yang,Zhenfei Yin,Yingcheng Wu
类目: Machine Learning (cs.LG); Quantitative Methods (q-bio.QM)
*备注:
Abstract:Immune therapies act across cell-intrinsic programs, tissue ecosystems, and patient-specific immune states, yet most predictors address these scales separately. We used a governed evolutionary AI Scientist to construct the Immune World Model, an action-conditioned model that learns how interventions move immune states across cellular, tissue, and individual levels. The Immune World Model–building Scientist searched candidate architectures and workflows, and the resulting world model was frozen before independent confirmation. The frozen model generalized to unseen interventions and biological contexts, recovered intervention-specific cellular programs, integrated cell and tissue information to improve ecosystem and patient-response prediction, and forecast unseen perturbation combinations. Immune World Model–guided analysis then combined measured perturbations with cross-axis inference to nominate IL-36 \gamma plus SIRP \alpha inhibition as a complementary-axis therapeutic hypothesis, whereas a governed self-correction audit rejected every screened cytokine pair. The Immune World Model provides a framework for multiscale immune simulation that connects AI Scientist-driven model construction, intervention forecasting, and the generation of prospectively testable therapeutic hypotheses.
[LG-62] Symmetries and Singularities
链接: https://arxiv.org/abs/2609.14663
作者: Vishnu Varadarajan,Mihir More,Aritra Das,Debayan Gupta
类目: Machine Learning (cs.LG)
*备注:
Abstract:Deep neural networks are highly over-parameterized, and different parameter values represent the same predictive function. This makes their effective complexity difficult to measure using only the number of parameters or the rank of the Hessian. Singular Learning Theory addresses this issue through the local learning coefficient (LLC), which characterizes the effective complexity of a model near a given solution. Existing methods for estimating the LLC often rely on posterior sampling, which can be computationally expensive for large neural networks. This makes accurate LLC estimation difficult at scale. In this work, we use known structures in the model to simplify the analysis and make LLC estimation more tractable. Specifically, we study the LLC of a graph attention model by exploiting symmetries in both the graph structure and the attention parameters. An analytic framework through a teacher–student setting, and explicit LLC estimates after considering the symmetry–induced degeneracies are developed.
[LG-63] SH-WRNN: Implicit Spherical Harmonics Weight Field Routing Neural Networks for Asymmetric Edge Intelligence
链接: https://arxiv.org/abs/2609.14614
作者: Zhibin Jiao,Xiangjing An
类目: Machine Learning (cs.LG)
*备注:
Abstract:Deep learning architectures remain rigidly built upon traditional fully connected layers. While networks scale up, few challenge this foundational root. In this work, we reshape this paradigm by transforming the core synapse weight matrix from static, discrete parameters into a differentiable, continuous field governed by spherical harmonics functions. We introduce the Implicit Spherical Harmonics Weight Field Routing Neural Network (SH-WRNN), which constrains weight matrices within a continuous parametric field instead of optimizing millions of localized discrete weights. When retrieving the weight matrix of the current layer, connection parameters are localized using latitude and longitude on a rectangular plane mapped from the continuous field. The latitudinal coordinate is specified by activated neurons from the previous layer, while the longitudinal coordinate is determined by keys generated from previous layer activations via matrix multiplication. By evaluating intersections on this map, the network dynamically extracts its connection weights on-the-fly. Empirical validation on MNIST demonstrates that under compact configurations of (32, 10, 10) and (32, 3, 10), SH-WRNN achieves robust accuracies of 91.05% and 81.54% within a single training epoch. Furthermore, we propose an asymmetric Surface Baking scheme. Upon convergence, the continuous weight field is baked once into a static parametric surface. By eliminating analytical spherical harmonics calculations during inference and reducing dynamic matrix extraction to high-speed localized memory slicing, this scheme achieves asymmetric algorithmic acceleration with negligible accuracy degradation. This paradigm shift bypasses GPU memory-bandwidth monopolies, opening a novel path to reshape the advantages of CPU computing. Code is available at this https URL.
[LG-64] Pathwise Individual Rationality in Federated Learning: A Mechanism-Architecture Co-Design
链接: https://arxiv.org/abs/2609.14591
作者: Amin Meghrazi,Srinivasan Parthasarathy,Andrew Perrault
类目: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT)
*备注: 23 pages, 6 figures, 2 tables
Abstract:Participation in federated learning (FL) comes at a cost. Clients trade off privacy, communication, and compute costs for potentially greater gains in model efficacy. This paper explores this tradeoff under the aegis of individual rationality (IR) versus autarky, the basic game-theoretic requirement that the federation provide utility no worse than local training. Using the above as the design target, we examine pathwise performance of FL, as a per-round bound on cumulative surplus, not just as an asymptotic equilibrium guarantee under different models of client data distribution heterogeneity. Along this path, clients can remain below their local-training baseline for hundreds of rounds. The natural remedy is to cap each client’s per-round contribution so that this shortfall stays bounded, and we prove that it backfires, collapsing learning even at low-to-modest heterogeneity. We then propose a novel design that combines short-term participation guarantees with personalized model evaluation, while maintaining fair incentives. We provide a theoretical basis for this new approach and empirically demonstrate that clients can avoid short-term losses without harming overall performance, even under moderate data distribution heterogeneity; under severe heterogeneity, the design shows promising outcomes for clients compared to their local baseline at some cost in accuracy. Comments: 23 pages, 6 figures, 2 tables Subjects: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT) Cite as: arXiv:2609.14591 [cs.LG] (or arXiv:2609.14591v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.14591 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-65] Learning Multi-Agent Task Assignment and Navigation in the Factory: from Simulation to Real Robots
链接: https://arxiv.org/abs/2609.14567
作者: Abdalwhab Bakheet Mohamed Abdalwhab,Giovanni Beltrame,David St-Onge
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: This work has been submitted for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
Abstract:Reinforcement learning (RL) has shown considerable promise for robotic decision-making, yet deploying multi-agent RL (MARL) on physical multi-robot systems in industrial environments remains challenging. This paper investigates the real-world applicability of decentralized MARL for multi-robot multi-machine tending. We propose Feature-fusion Multi-Agent Proximal Policy Optimization (FMAPPO), which fuses 2D LiDAR measurements with task-specific state information to enable safe decentralized multi-robot task assignment and navigation. A complete simulation-to-reality pipeline was developed using high-fidelity robotic simulation and ROS2 and deployed on physical mobile-manipulator platforms operating under realistic real-world conditions, with the robotic arms disabled during the experiments. We further investigate the sensitivity of the learned policy to command update frequency, an important consideration for real-world deployment. Comparative evaluation in simulation demonstrated that FMAPPO significantly outperformed state-of-the-art baselines with a large effect size, achieving improvements of 106% and 21% in parts delivery and 48% and 11% in parts collection over MAPPO and SMAPPO, respectively. FMAPPO also increased machine utilization by 31 and 10 percentage points, respectively, while reducing collisions by 18% and 15% and increasing the safety score by 14 and 6 percentage points compared with MAPPO and SMAPPO, respectively. Furthermore, real-world experiments demonstrated that the learned decentralized policies can coordinate multiple robots to service multiple machines while maintaining safe operation under real-world sensing and control constraints. Videos of the real-world experiment are available online this https URL.
[LG-66] Selecting k Paths with the Minimum Longest Path Length in the Stochastic Semi-Bandit Setting
链接: https://arxiv.org/abs/2609.14557
作者: Shunsuke Aoki,Atsuyoshi Nakamura
类目: Machine Learning (cs.LG)
*备注:
Abstract:When performing parallel data transmission through a network using multiple paths, it is practically important to minimize the maximum transmission time among the selected paths. This study addresses an online problem in which k paths from an origin vertex to a destination vertex must be selected at each time step within a network represented as a directed graph. Here, the number of paths going through each edge in each parallel data transmission is limited to its capacity, and the time required for transmission is determined stochastically. We formulate the semi-bandit problem of selecting a set of paths to minimize the maximum traversal time among the selected paths and propose an algorithm to solve it.
[LG-67] GNN4PPM: Multi-Target Predictive Process Monitoring with Relational Graph Convolutional Networks
链接: https://arxiv.org/abs/2609.14534
作者: Ana Costa,Johannes Mäkelburg,Luise Pufahl
类目: Machine Learning (cs.LG)
*备注: Accepted for presentation at the BPM 2026 Forum. The final version will appear in the Lecture Notes in Business Information Processing (LNBIP) post-proceedings
Abstract:Predictive Process Monitoring (PPM) aims at predicting at runtime and as early as possible the future states of a process execution. Common tasks include predicting the next event, the time to completion of a trace, and outcomes. Existing approaches typically consider an event from the perspective of the executed activities along with their timestamps and case identifiers. This leads to the disadvantage that in real-life settings, there is much more information recorded in the event log that is not captured or completely ignored when performing prediction tasks. We introduce GNN4PPM, an approach that predicts all next events along with their complete data payload at once. We represent event information in a heterogeneous knowledge graph that captures the event log as an RDF semantics, and train the embeddings with a Relational Graph Convolutional Network (R-GCN). Our approach is promising in comparison to existing solutions, and experiments with state-of-the-art solutions prove the accuracy and applicability of GNN4PPM in complex settings.
[LG-68] oward a Layer-2 Trigger for AI/ML Lifecycle Management in 6G
链接: https://arxiv.org/abs/2609.14517
作者: Dharmendra Kumar
类目: Networking and Internet Architecture (cs.NI); Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注: 6 pages, 1 figure, 1 table. Submitted to IEEE Communications Standards Magazine
Abstract:3GPP has progressively expanded AI/ML lifecycle management in the radio access network, from one-sided model control to Release 20 support for two-sided CSI-feedback model pairing. Yet a basic control question remains: when monitoring detects degradation, how quickly must a corrective action take effect? To expose this dependency, we stress-test three activation and rollback strategies in a surrogate regime-shift environment using 150 independently trained PPO candidate policies, each evaluated over 20 matched noise realizations. We add 0-40 control-step delay only to corrective lifecycle commands. With no added delay, stability-gated blending reduces mean post-shift cumulative SLA deficit from 47.02 to 7.57 violation-steps relative to hard cutover; at 40 steps, the deficit rises to 44.38, only 5.6% below the hard-cutover baseline. KPI-threshold rollback loses its advantage within only a few control intervals, while blending degrades more gradually. These results do not set a physical 6G latency bound. They show why timing requirements matter for corrective actions: lifecycle performance depends on when the command takes effect. Motivated by Layer-1/Layer-2 Triggered Mobility, we examine a standards split in which Layer 3 retains lifecycle configuration while a compact Layer 2 trigger is considered only for the latency-critical subset, together with pair-consistency, local-fallback, and security/freshness requirements.
[LG-69] Physically Partitioned KVCache Format for CPU–GPU Load Balancing in MoE Inference
链接: https://arxiv.org/abs/2609.14507
作者: Enda Yu,Dezun Dong,Xiangke Liao
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注:
Abstract:Single-GPU long-context inference with Mixture-of-Experts (MoE) models requires spilling the key-value cache (KVCache) to CPU memory. The spilled KV serves two complementary purposes—transferring to the GPU for attention computation, or computing in-place on the CPU—which demand opposing physical states. The optimal split between them varies with workload, yet existing KVCache abstractions offer only storage semantics over a monolithic object of a single physical state, and cannot express dynamic load balancing. We propose InplaceKVCache, the first KVCache abstraction whose format fixes each byte’s physical residency at write time, so that the CPU–GPU load balance can be adjusted without moving data after placement. It realizes this as a four-region layout along two dimensions—device affinity and access pattern—turning load balancing into pure scheduling. Built on this abstraction, WriteScope splits CPU–GPU shares along the sequence dimension, and a portable roofline performance model determines the optimal CPU share as sequence length evolves, with online feedback tracking CPU cost drift. On three MoE models (DeepSeek-V2-Lite, Qwen3-30B-A3B, Mixtral-8 \times 7B) with a 32~GB VRAM budget, WriteScope supports end-to-end inference at the 1M-token aggregate scale. In the long-context regime ( \ge 8K), it achieves geometric-mean speedups of 1.5\times – 2.5\times on A100 and 1.4\times – 1.7\times on V100 over four reproduced baselines, while vLLM, SGLang, and KTransformers fail even with a doubled KV budget. A DeepSeek-V4-Flash case study validates composition with native sparse attention.
[LG-70] Should All Noises Be Treated Equally: Impact of Input Noise Variability on Neural Network Robustness
链接: https://arxiv.org/abs/2609.14504
作者: Salma Alsinan,Maksim Makarenko,Sixiu Liu,Ali Aldawood,Ibrahim Hoteit
类目: Machine Learning (cs.LG); Signal Processing (eess.SP); Geophysics (physics.geo-ph)
*备注: Main:(pages 1-31,15 figures), Supplementary: (pages: 32-75, 37 figures)
Abstract:Geophysical data collected from active field sites are often contaminated by complex and heterogeneous noise, obscuring weak seismic events, and complicating automated interpretation. Although deep learning offers promising solutions for seismic processing, its performance is highly sensitive to the nature of training noise, especially under out-of-distribution (OOD) conditions. This study investigates the influence of noise parameters, such as type, scale, and complexity on the performance, generalization and robustness of neural networks in two geophysical tasks: first break picking and denoising. We simulate seismic-while-drilling data and apply controlled input source noise augmentation using stochastic generators to vary the noise characteristics. Different neural networks are trained on fixed noise types and scales, then evaluated across both seen and unseen noise scenarios. We incrementally increase the complexity of the noise by introducing compound noise mixtures and assess the performance of the model under increasingly challenging OOD conditions. This yields a robustness matrix that captures the generalizability of each model relative to its training configuration. Results indicate that larger noise scales boost generalization, and that effective alignment between noise type, task complexity, and architecture is key for maximizing generalization gains. In addition, training with compound noises mitigate weaknesses associated with single-noise training, acting as an additional implicit regularizer to improving robustness. These findings highlight key factors influencing model resilience in noisy geophysical environments and offer guidance for developing deep learning models that generalize effectively across diverse and unpredictable noise conditions.
[LG-71] Retrieval-Guided Fine-Tuning as Noisy Estimation: Risk bounds and Architectural Analysis NEURIPS2026
链接: https://arxiv.org/abs/2609.14485
作者: Bhargav Lad,Yifan Hao
类目: Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注: 29 pages, 2 figures, 1 table. Submitted to the NewInML Workshop at NeurIPS 2026
Abstract:Retrieval-Guided Fine-Tuning (RAG-FT) incorporates retrieved data directly into the training objective, but the statistical consequences of noisy retrieval during training remain theoretically undercharacterized. We study this question by modeling RAG-FT as an estimation problem in a multi-task linear regression framework, using an OLS proxy for single-layer linear self-attention to obtain finite-sample risk bounds. Under homoscedastic retrieval noise, we show that retrieval failure decays exponentially with task separation relative to noise, and derive explicit finite-sample conditions under which RAG-FT achieves lower risk than both target-only and full-corpus training. We then introduce a Distance-Proportional Noise (DPN) model, in which retrieval quality degrades with rank, and compare two estimators under the same retrieval process: the OLS proxy and the literal, uniform-weight forward pass of linear self-attention. We prove that the attention estimator’s bias diverges as \Theta(n^2q) even under exact retrieval, while OLS risk remains \Theta(d/n) for every noise exponent q0 . These results locate the instability not in noisy retrieval itself, but in the fixed, unweighted aggregation of the literal LSA forward pass, which reweighting by reliability empirically removes. We validate the predicted rate separation through direct simulation of the DPN model.
[LG-72] owards Identifying the Dataset Biases Causing Phantom Transfer
链接: https://arxiv.org/abs/2609.14449
作者: Jonas Jürß,Pietro Liò
类目: Machine Learning (cs.LG)
*备注:
Abstract:Recent work has shown that a teacher model can transfer a bias to a student through a dataset from which every explicit reference to that bias has been filtered out, and that no data-level defense reliably removes or detects it even when knowing what bias to look for. Aiming to shed light on the hidden traces of these biases, we show that a simple signature based on Sentence BERT embeddings can identify the topic of such a bias with a Matthews correlation coefficient of 0.83 if the teacher model used by the attacker is known and 0.46 if it is not. Additionally, we observe that different teacher models appear to express the same bias through different vocabulary.
[LG-73] Multimodal deep learning from spectra for small-molecule structure identification: enhancing robustness with mixed-condition training
链接: https://arxiv.org/abs/2609.14360
作者: Bowen Gao,Lei Zhu,Yiying Wang,Wenjie Yu
类目: Machine Learning (cs.LG)
*备注: 16 pages, 7 figures, 3 tables. Supplementary information: 17 pages, 5 figures and 10 tables, provided as an ancillary PDF. Preprint
Abstract:In practical molecular characterization, small-molecule structure identification benefits from complementary spectroscopic evidence, but missing, degraded, or mismatched spectra challenge multimodal models. Herein, we incorporate domain knowledge from spectroscopy and chemistry into mixed-condition training for candidate structure reranking, using a reproducible evaluation protocol and mixture-of-experts (MoE) fusion. The protocol incorporates perturbations tailored to each spectroscopic modality and chemically informed spectrum replacements to cover variations in spectral availability, quality, and consistency. A total of 79,462 test samples were evaluated across 30 predefined conditions using simulated spectra from the Multimodal Spectroscopic Dataset (MSSD) for mass spectrometry (MS), infrared (IR) spectroscopy, and 1H and 13C nuclear magnetic resonance (NMR), with up to 128 hard candidate structures per sample. A controlled two-by-two factorial comparison of complete-input training versus mixed-condition training and vanilla concatenation versus MoE fusion, with matched evaluation conditions, showed that mixed-condition training provided the main gains in both architectures. For MoE, mean reciprocal rank (MRR), averaged equally across conditions, increased from 0.9203 to 0.9763, a relative increase of 6.08%. Recall at rank 1 (R@1), averaged over the same conditions, increased from 89.50% to 96.36%, an increase of 6.86 percentage points and a relative increase of 7.67%. IR-only and MS/MS-only MRR increased from 0.4337 to 0.9307 and from 0.3711 to 0.8575, reaching 2.15 and 2.31 times their respective baseline values, while complete-input performance remained high. These results support integrating domain knowledge into training-condition design to improve robustness, with further gains from MoE under mixed-condition training.
[LG-74] Communication-Efficient LLM Adaptation over Decentralized GPU Meshes
链接: https://arxiv.org/abs/2609.14339
作者: Sameera Ramasinghe,Shamane Siriwardhana,Thalaiyasingam Ajanthan,Hadi Mohaghegh Dolatabadi,Chamin P Hewa Koneputugodage,Gil Avraham,Violetta Shevchenko,James Snewin,Karol Pajak,Harry Xi,Alexander Long
类目: Machine Learning (cs.LG)
*备注: 38 pages, 2 figures, 12 tables
Abstract:Decentralized training enables large-model training over low-end GPUs and internet-grade connections, but communication along both data-parallel and pipeline-parallel axes becomes the primary bottleneck. We study post-pretraining adaptation in this setting. We propose an asynchronous two-circuit system: a fast compressed training circuit drives throughput using activation masking for pipeline-parallel (PP) transfer and compressed data-parallel (DP) synchronization, while a slow anchor circuit runs occasional unmasked forward–backward passes off the critical path. Then, we introduce a spectral correction optimizer that uses these delayed anchor priors to denoise masked gradients without blocking the fast stream. Although prior work has found aggressive activation compression unreliable, we show that masking supports post-pretraining adaptation at high compression rates when anchored this way. Pipeline-parallel compression alone yields up to a 9\times throughput gain, and combining it with data-parallel compression increases beyond 40\times over internet-grade \sim 200 Mbps connections, while matching dense uncompressed performance across domain adaptation and continual pretraining.
[LG-75] Nonparametric Variance-Penalized Actor-Critic: Statistical Inference for Risk-Sensitive Reinforcement Learning
链接: https://arxiv.org/abs/2609.14327
作者: Saunak Kumar Panda,Tong Li,Yisha Xiang,Ruiqi Liu
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Submitted to IEEE Transactions on Neural Networks and Learning Systems. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
Abstract:Variance penalization is a principled approach to risk-sensitive reinforcement learning (RL) that explicitly trades expected return for policy stability. Existing methods require a dedicated second critic to estimate return variance online, adding architectural complexity and compounding estimation error during learning. We propose a nonparametric variance-penalized actor-critic (VPAC) framework that replaces the variance critic with statistically grounded online estimators based on bootstrapping and random scaling, techniques drawn from the statistical inference literature for stochastic approximation. These estimators require no auxiliary network, maintain a single-critic architecture, and produce variance penalties that are bounded by construction, enabling clean convergence analysis. We establish almost-sure convergence for both a variance-penalized Q-learning algorithm and a two-timescale actor-critic variant via the ordinary differential equation (ODE) method, requiring only that variance estimates remain bounded rather than consistent. Empirically, we evaluate across discrete and continuous stochastic environments, demonstrating that the proposed methods match or exceed the variance reduction achieved by the existing dual-critic VPAC baseline while eliminating the overhead of a second critic. We further validate on a high-temperature superconductor (HTS) manufacturing case study, where VPAC-RS (Random Scaling) achieves a 74% reduction in steady-state critical current variability and a 63% reduction in episode return standard deviation, translating directly to improved yield consistency. Our results establish nonparametric statistical inference as a practical and theoretically sound alternative to auxiliary critics for risk-sensitive RL.
[LG-76] AI Assisted Workflow Optimization and Automation ACSA
链接: https://arxiv.org/abs/2609.14323
作者: Zhen Zhong
类目: oftware Engineering (cs.SE); Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Risk Management (q-fin.RM)
*备注: 5 pages, 1 figure. Published in International Journal of Advanced Computer Science and Applications (IJACSA), Vol. 16, No. 10, 2025
Abstract:Against the backdrop of digital transformation and stricter regulation, enterprise compliance work demands higher efficiency and accuracy. The auxiliary compliance process has become an important entry point for optimizing the compliance system due to its strong transactional nature and high degree of repetition. This study focuses on the process characteristics of auxiliary compliance work, sorts out its structural composition and organizational mechanism, proposes an optimization path with process reengineering, system modeling, and technology integration as the core, and focuses on exploring the collaborative application of key technologies such as RPA, rule engine, and semantic recognition in process automation. Research suggests that the systematic optimization and intelligent upgrading of auxiliary processes will help build a modern compliance operation system that is responsive, efficient, structurally clear, and risk controllable.
[LG-77] Learning Source Acquisition Policies by Offline Planning
链接: https://arxiv.org/abs/2609.14299
作者: Ziqi Zhao,Run Xu,Qingjian Ni
类目: Machine Learning (cs.LG)
*备注:
Abstract:Predicting under an acquisition budget requires choosing feature groups whose value can depend on later queries. O-MPAC transfers finite-horizon risk-cost targets from complete training records into a shared source-action scorer. At inference time, the scorer uses partial observations and source metadata, re-scores after each query, and applies a hard cost mask. We analyze how tied teacher targets and the remaining planning horizon affect the learned decisions. Uniform supervision over tied minima preserves the target distribution under source relabeling. In a five-seed routing experiment, it achieves 0.965 accuracy under both original and context-last orders. On six real tasks, validation selects H1 without action cross-entropy in all thirty splits. O-MPAC has the highest mean budget-integrated accuracy on five tasks against source-adapted GDFS, DIME, AACO+NN and a static policy.
[LG-78] Fusing Spectral Signatures and Activation Clustering for Backdoor Detection in Healthcare Imaging Models: Method Implementation and Evaluation
链接: https://arxiv.org/abs/2609.14290
作者: Suresh Tamang
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: 13 pages, 2 figures, 3 tables. Implementation available at this https URL . Also deposited at Zenodo, doi: https://doi.org/10.5281/zenodo.22431041
Abstract:Machine learning models are increasingly deployed in healthcare imaging pipelines for diagnostic support, and training-time attacks against them are a named sector-level concern: healthcare-sector guidance identifies model poisoning and adversarial attacks as threats requiring dedicated defenses, while federal policy directs expanded AI vulnerability-detection tooling to critical infrastructure operators such as rural hospitals. Spectral signature analysis and activation clustering are two established backdoor detection methods routinely evaluated as independent baselines, but their outputs are not ordinarily combined, and reported detection performance on medical imaging benchmarks remains sparse relative to the natural-image setting. This paper contributes three things: a score-level fusion rule combining per-class spectral ranking with activation-clustering flags into a single per-sample poisoning score and a model-level agreement statistic; an open-source implementation of the resulting eight-stage pipeline; and an evaluation of that pipeline against synthetically poisoned variants of a public medical imaging benchmark and CIFAR-10 at four poisoning rates (0%, 1%, 5%, 10%) over five seeds each, measuring each detector alone against the fusion. On the medical benchmark, the fused detector reaches AUROC = 0.99 at every nonzero poisoning rate tested. On CIFAR-10, fusion does not uniformly help: at 10% poisoning, activation clustering’s true-positive rate collapses to 0.000 and spectral AUROC independently degrades to near-chance (0.545), despite a 97.2% attack success rate confirming the backdoor was fully installed. The fused score, a weighted combination of both signals, inherits this joint failure. Detection output is expressed in NIST AI RMF Measure-function and MITRE ATLAS terms, so findings are reported in the vocabulary security and compliance teams already use.
[LG-79] Bayesian optimization with kernel ensembles and disagreement-based acquisition for source localization and acoustic inversion
链接: https://arxiv.org/abs/2609.14262
作者: Heng Zhang,Haotian Xiang,Florian Meyer,Qin Lu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Joint source localization and geoacoustic inversion requires optimizing an objective built from an expensive normal mode propagation model. Bayesian optimization (BO) with a Gaussian process (GP) surrogate can obtain accurate parameter estimates within a limited number of forward model evaluations, but its performance depends on the choice of kernel family. With few observations in a seven-dimensional search space, no single kernel can be expected to perform consistently well across individual inversions. To reduce this dependence, we use a weighted ensemble of GPs with different kernel families, allowing the surrogate to adapt to the observed objective without committing to one kernel in advance. The ensemble is combined with an optimum-conditioned acquisition function that determines where the expensive objective should be evaluated next. Experiments on simulated and measured SWellEx-96 data show that the resulting method achieves the lowest mean final objective among the considered BO strategies and reduces parameter estimation error on most coordinates. Ablation results further show that the ensemble provides robustness to kernel choice, while the acquisition function accounts for most of the optimization gain.
[LG-80] VGFM: Expressive Robot Policies via Dense Value Guidance in Flow Matching IROS2026
链接: https://arxiv.org/abs/2609.14261
作者: Prajwal Koirala,Mark Campbell
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: IROS 2026
Abstract:Recent robot learning paradigms increasingly rely on large offline datasets of robotic interactions to train control policies. Expressive generative models enable rich and multimodal action representations, expanding the capability of this paradigm for complex robotic control. However, policy improvement with multi-step generative actors remains challenging. In offline reinforcement learning (RL), incorporating value-based objectives along generative trajectories often introduces substantial training complexity, including backpropagation through time (BPTT), auxiliary architectures, or distillation losses. We propose Value-Guided Flow Matching (VGFM), a scalable offline RL framework that enables dense value-guided shaping within a flow-based policy while avoiding BPTT and additional algorithmic overhead. VGFM parameterizes the policy as a conditional flow-matching model in action (x-prediction) space, ensuring that each intermediate flow step produces a valid robot action that can be directly evaluated by a standard offline RL critic. This design allows value guidance to be applied at randomly sampled flow times without differentiating through the entire generative trajectory, while preserving inference-time flexibility by varying the discretization of the underlying flow ODE without retraining. Evaluated on robotic locomotion and manipulation tasks in OGBench, VGFM achieves strong performance across a wide range of tasks under rigorous evaluation protocols. With minimal hyperparameter tuning, these results demonstrate that VGFM provides a simple, scalable, and effective approach for expressive policy learning in long-horizon, goal-oriented robotic control.
[LG-81] owards Anticipatory Databases Through Shared Data and Workload Semantics
链接: https://arxiv.org/abs/2609.14255
作者: Farzaneh Zirak,Kasper Overgaard Mortensen,Farhana Choudhury,Renata Borovica-Gajic
类目: Databases (cs.DB); Machine Learning (cs.LG)
*备注: 8 pages, 6 Figures
Abstract:Database management systems increasingly serve dynamic and exploratory workloads, yet many of their decisions still rely on low-level signals such as recency, frequency, and address locality. These signals capture how data was accessed, but not what is being examined or how an analytical focus evolves. We argue for treating workload semantics as a first-class control signal for anticipatory decision making. Central to this view, we introduce semantic locality and semantic trajectories, which capture relationships among nearby queries and how those relationships evolve across a session. We propose a framework that represents semantic context at the data, query, and session levels, models its evolution over time, and translates it into task-specific utility estimates. We instantiate this framework in semantic prefetching and semantic cache eviction, which share a semantic layer to make two separate decisions. Prefetching uses semantic trajectories to anticipate future accesses beyond what address-based locality can capture, while eviction uses semantic relevance to inform block replacement. These systems provide initial evidence that shared semantic context can support multiple DBMS components. We further outline how this principle can extend to other decisions and data systems, and discuss key challenges in representation, cost, adaptation, and evaluation. Comments: 8 pages, 6 Figures Subjects: Databases (cs.DB); Machine Learning (cs.LG) Cite as: arXiv:2609.14255 [cs.DB] (or arXiv:2609.14255v1 [cs.DB] for this version) https://doi.org/10.48550/arXiv.2609.14255 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-82] Joint Optimization for Federated Learning and Transmission over Unreliable Wireless Networks with Heterogeneous Data
链接: https://arxiv.org/abs/2609.14246
作者: Changheng Wang,Xianchao Zhang,Zhiqing Wei,Lingzhu Zhao,Zhongming Yang,Zhiyong Feng
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注:
Abstract:In wireless federated learning (FL), data heterogeneity and multiple local updates induce client drift, degrading model convergence. It is further affected by unreliable wireless links, as transmission errors may invalidate model updates. To address these challenges, we propose a federated random walk averaging (FedRW) framework, which is a variant of federated averaging (FedAvg) that mitigates data heterogeneity by updating models along random walk (RW) paths and aggregating them at the server. Model parameters are transmitted in packets with retransmission support to improve training quality by mitigating wireless errors along RW paths. Meanwhile, wireless transmission delays hinder the exploration of FedRW. To this end, we formulate a joint optimization problem that integrates learning, RW path selection, and transmission parameter tuning, aiming to minimize the training loss under delay constraints. By deriving an upper bound on the expected convergence of FedRW over unreliable wireless networks, we reduce the problem to a general form agnostic to task type and model architecture. A distributed solution is then proposed, in which the server or clients optimize packet size and maximum number of retransmissions locally, and efficiently select reliable and expandable next-hop nodes via a resilience-aware beam search with dynamic pruning. Simulation results show that FedRW achieves 2.26%-9% higher accuracy than state-of-the-art baselines under high data heterogeneity. Furthermore, the jointly optimized FedRW yields at least 2.78% higher accuracy and faster convergence compared to baselines.
[LG-83] CoArena: Evaluating Computer-Use and Multi-Agent Systems in Real Time
链接: https://arxiv.org/abs/2609.14239
作者: Nitish Kovuru,Prateek Jannu
类目: Machine Learning (cs.LG)
*备注: 29 pages, 9 figures, 4 tables, 4 algorithm listings. All figures are drawn in TikZ/pgfplots from the source. Project page: this https URL
Abstract:Static benchmarks for computer-use agents fix a task set at release and score every system against it once. That makes them reproducible, and it lets them drift from what they should measure: a fixed task set ages, leaks into training corpora, and cannot follow how people actually use agents from week to week. CoArena measures use directly. Real users submit tasks; two systems, each a single model or a multi-agent pipeline behind the same tool interface, execute the same task concurrently in identical sandboxed desktops; users judge the two outcomes without knowing which system produced them; and a public leaderboard is refit from those judgments. The central contribution is a formal account of what makes such an evaluation real-time. We define real-time as five measurable properties, each with an equation and a worked example: continuous task arrival, live concurrent execution, online rating updates, freshness with contamination resistance, and bounded feedback latency from a failed run to a reusable training environment. The rating methodology follows in full: the Bradley-Terry pairwise model, its likelihood with weighted observations and ties, the penalized maximum-likelihood estimator, and the streaming update applied when a single vote arrives (a stochastic-gradient step on the same likelihood, recovering Elo). It gives confidence intervals from the observed information and a cluster-robust sandwich, rank bands from a parametric bootstrap, the rule by which a new system enters the board, and the convergence rate of the estimate. Vote quality is treated with inter-judge agreement statistics, redundant judging, and explicit handling of ties and abstentions. A five-system example with 211 votes is carried from the vote matrix to ratings, intervals, and rank bands. Every number is derived from stated inputs or labeled illustrative; none is a measurement of a deployed system.
[LG-84] Partition-Aware Scheduling for Mobile Heterogeneous Inference Co-Execution
链接: https://arxiv.org/abs/2609.14213
作者: Zhuojin Li,Marco Paolieri,Leana Golubchik
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Performance (cs.PF)
*备注: Accepted for publication in the Performance Evaluation journal (presented at IFIP Performance 2026, in Ghent, Belgium)
Abstract:Modern mobile inference runs on heterogeneous platforms combining mobile GPUs with multiple CPU core clusters. Existing optimizations typically exploit either inter-operator parallelism, by assigning entire operators to CPU cores or to the GPU, or intra-operator parallelism, by partitioning each operator for CPU-GPU co-execution. We consider these two forms of parallelism together, to improve inference latency of tasks that can be represented by a static DAG of operators with predefined input/output tensor shapes (e.g., CNNs or vision transformers). We define the problem of partition-aware DAG scheduling for mobile heterogeneous inference, illustrating that the best strategy depends on the structure of the inference DAG, thus motivating a joint formulation capturing operator partition choices, device assignment, and execution order. We propose an online iterative search framework, which decomposes large DAGs into stages, focuses search on critical operators, and uses latency predictors to estimate partitioned execution without exhaustive profiling. Across representative mobile inference workloads, our approach achieves latency close to an offline solution while keeping scheduling overhead to a fraction of the model initialization cost, allowing platform-specific scheduling at deployment time.
[LG-85] ZAPS: Zero-Cost Active Proxy Search for Neural Architecture Search
链接: https://arxiv.org/abs/2609.14184
作者: Hassan Touayouch,Rabie Najem,Mohammed Benjelloun
类目: Machine Learning (cs.LG)
*备注: 12 pages, 13 figures, 9 tables
Abstract:Neural Architecture Search (NAS) automates network design, but evaluating a single candidate requires training it to convergence, making exhaustive search intractable. Zero-cost proxies estimate architecture quality at initialization in seconds, yet a single proxy is noisy, and combining several does not straightforwardly help: proxies are strongly correlated, so naive aggregation compounds their shared errors instead of averaging them out. Existing methods exploit either proxy signals or architectural topology - never both within a single active-learning framework. We introduce ZAPS (Zero-cost Active Proxy Search), a four-stage pipeline that closes this gap. ZAPS (i) selects a compact, non-redundant proxy subset offline via ProxyFit, a greedy anti-redundancy criterion; (ii) seeds the search with a hybrid K-means strategy that balances exploitation and exploration; (iii) re-selects proxies at every iteration by a bootstrapped vote as the labeled set grows; and (iv) ranks candidates with an XGBoost ensemble trained jointly on proxy ranks and one-hot topological encodings, queried through an Upper Confidence Bound (UCB) acquisition function. On NAS-Bench-201 under a budget of B=200 evaluations, ZAPS recovers 52.3% of the true top-100 architectures on CIFAR-10 and 65.8% on CIFAR-100, ahead of every baseline we consider - Random Search, Local Search, REA, BANANAS and TPE - and, on CIFAR-10, with less than half the run-to-run standard deviation of the strongest of them. The advantage is largest where evaluations are scarce: on NAS-Bench-201 it narrows as the budget grows, whereas on the harder NAS-Bench-101, which no method comes close to saturating, it widens instead. All methods are scored by a single criterion: how much of the true top-100 lies among the architectures they actually evaluated.
[LG-86] A Machine Learning Framework for Fault Detection Isolation and Severity Prediction of Autonomous VTOL Aircraft
链接: https://arxiv.org/abs/2609.14180
作者: Ripon C. Sarker,Pedram H. Dabaghian,Raman Goyal,Atanu Halder
类目: Machine Learning (cs.LG)
*备注:
Abstract:Fault detection in autonomous VTOL aircraft is critical because even minor component degradations can rapidly destabilize multirotor vehicles operating in complex, safety-critical environments, motivating robust fault detection and estimation strategies capable of identifying early signs of rotor damage; however, real-flight fault detection remains challenging due to sensor noise, environmental disturbances, and the nonlinear aerodynamics of multirotor platforms. This study proposes a comprehensive machine-learning framework for rotor fault detection, isolation, and severity prediction using real flight data. A convolutional neural network (CNN) architecture is developed to learn spatio-temporal patterns from multivariate flight dynamics, enabling direct inference of both the faulted rotor and its damage level. The framework is first validated using simulated data generated by a data-generative model, and experimental validation is then performed on a hexacopter by introducing controlled blade-tip breakage. The trained model achieves rotor-wise fault classification accuracies above 99% and severity estimation accuracy of 96% within a 1% tolerance in experimental data, demonstrating strong generalization and supporting real-time health monitoring for autonomous VTOL systems.
[LG-87] CyFM: Cylindrical Optimal Transport for Few-Step Complex-Valued Flow Matching
链接: https://arxiv.org/abs/2609.14171
作者: Marcel Musiałek,Iga Wolanin,Damian Ryczko,Anna Grelewska,Oleksii Furman
类目: Machine Learning (cs.LG)
*备注: 16 pages, 1 figure, 5 tables
Abstract:Complex-valued signals, such as Magnetic Resonance Imaging (MRI) and audio spectrograms, are almost always modelled as flat two-channel Euclidean data. For nonzero values the amplitude-phase chart z \mapsto (|z|, z/|z|) identifies the signal domain with the cylinder (0, \infty) \times S^1 , on which we deliberately replace the inherited metric dA^2 + A^2 d\theta^2 by the decoupled product metric dA^2 + d\theta^2 . In this empirical study we measure what that substitution costs and what it buys. By computing exact analytical bridges, we demonstrate that Cartesian paths induce a heavy-tailed distribution of angular velocity (power law index \approx 1.0 ), with nearly half of the probability paths exceeding an angular speed of \pi under independent coupling, a rate no cylindrical path ever exceeds. To resolve this, we analyze Cylindrical Flow Matching (CyFM), which strictly bounds the regression target, and couple noise and data by exact minibatch Optimal Transport computed jointly over whole fields in the cylindrical metric. Although the transport-cost reduction of this coupling collapses with field dimension (from 86% for scalar pairs to 3% for 64\times64 fields), its benefit to few-step generation does not: it lowers the few-step error of the cylindrical model by 3-60% at every evaluated resolution. With this coupling, CyFM has a lower error than the best Cartesian baseline at every step count up to k = 8 and every evaluated resolution, with all five seeds separated and without distillation, and at convergence we detect no significant difference between the two geometries. Finally, we expose the “Factorized Coupling Trap,” showing that dimension-wise or patch-wise transport factorizations silently destroy the joint distribution of the data. All experiments are on synthetic complex fields.
[LG-88] When Faster VLA Deployment Changes Closed-Loop Behavior: Task Success-Latency Analysis of SmolVLA Across PyTorch and ONNX Variants
链接: https://arxiv.org/abs/2609.14146
作者: Rafiqul Islam
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: 6 pages. Code: this https URL
Abstract:Vision-language-action (VLA) deployment can reduce inference latency while changing closed-loop task behavior. We evaluate HuggingFaceVLA/smolvla_libero on an RTX 2060 (6 GB) in LIBERO Spatial and Object (MuJoCo 3.3.2, LeRobot 0.6.1, seed 42), comparing PyTorch+AMP with ONNX Runtime CUDA Execution Provider (CUDA EP). The main evaluation uses 100 episodes/suite; a paired rollout uses 300 episodes/suite. PyTorch+AMP reaches 70.0%/88.0% Spatial/Object success at 1181 ms p99. Requested-FP16 and requested-INT8 ONNX reduce tether-inspect p99 to 601 ms and 532 ms, while Spatial success falls to 41.0% and 40.0% and Object remains at 89.0%. A graph audit shows those artifacts are byte-identical FP32 graphs, so the requested-INT8 row is not operator-level INT8 quantization. A static language-width ablation (16/24/32 tokens) yields Spatial success of 41.0%, 75.0%, and 71.0%; widths 24 and 32 recover much of the Spatial drop while Object success and uniform-bench latency stay approximately stable. Width-24 ONNX Spatial success is comparable to the PyTorch+AMP baseline at roughly half the latency (Wilson intervals overlap; two-proportion chi-squared p=0.53). Context width is an important contributor in this stack; it does not account for every PyTorch-vs-ONNX difference. Deployment evaluation should jointly report latency, artifact inspection, interface constraints, and closed-loop success. Code: this https URL.
[LG-89] -SMART: Mechanism-Level Attribution for Tool-Augmented Time-Series Question Answering ICTAI’26
链接: https://arxiv.org/abs/2609.14142
作者: Ivan Delgado,Himansi Gupta,Bishal Khatri,Niharika Sapre,Lameta Shamoon,Onat Gungor,Tajana Rosing
类目: Machine Learning (cs.LG)
*备注: Accepted by the 38th IEEE International Conference on Tools with Artificial Intelligence (ICTAI’26)
Abstract:Large language models (LLMs) can struggle with time-series question answering (TS-QA), especially when numerical signals are serialized as text and require explicit computation. Tool-augmented approaches improve performance, but existing systems often intertwine language reasoning, computation, and perception, making it difficult to determine which components drive the gains. We present T-SMART, a neurosymbolic framework that separates these roles: a frozen LLM interprets questions and selects operations, deterministic tools perform numerical computation, and structured perception is invoked only when needed. Controlled paired ablations show that deterministic computation provides the dominant benefit, improving accuracy by 31.7 percentage points over direct LLM reasoning on serialized time series, while language understanding and perception offer smaller complementary gains. These results indicate that tool-augmented TS-QA benefits primarily from reliable numerical execution rather than additional language-model reasoning and provide a controlled framework for analyzing component contributions in neurosymbolic time-series systems.
[LG-90] Neuron Activation-based Computation of Logical Explanations for Deep Neural Networks
链接: https://arxiv.org/abs/2609.14099
作者: Tomáš Kolárik,Faezeh Labbaf,Fabrizio Leopardi,Grigory Fedyukovich,Michael Wand,Natasha Sharygina
类目: Logic in Computer Science (cs.LO); Machine Learning (cs.LG)
*备注:
Abstract:Formal explainability of classifying neural networks (NNs) is an active area of research, providing explanations with provable guarantees of the classification within continuous regions of the input feature space. However, the existing techniques are either limited to individual input features without guarantees on their relations or the provided solutions fail to scale to deep architectures. This paper addresses these issues by introducing a flexible symbolic framework for an efficient, guided computation of explanations of the NN behavior, parametrized by the activations of internal neurons, and using logical engines such as SMT solvers. Unlike prior methods that rely on specialized NN verifiers, our method yields explanations that are not restricted in shape. Our algorithm is implementable on top of a general-purpose logical solver, isolating the NN-specific encoding from the algorithmic framework. We experimented with a wide range of benchmarks from the domains of image recognition and medicine, illustrating the advantages of the new method, particularly in computational efficiency. Notably, our approach enables logical explanation of deep networks not amenable to prior logic-based methods.
[LG-91] NeuroFlex: Lossless Element-Level ANN-SNN Co-Execution for Efficient Sparse Inference
链接: https://arxiv.org/abs/2609.14092
作者: Varun Manjunath,Pranav Ramesh,Gopalakrishnan Srinivasan
类目: Hardware Architecture (cs.AR); Machine Learning (cs.LG)
*备注:
Abstract:Sparse DNN accelerators specialize in ANN or SNN execution, leaving energy or latency on the table when workload characteristics vary within a layer. Hybrid accelerator designs that switch modes at layer or tile granularity suffer from low PE utilization since one core type idles whenever the other is active. NeuroFlex is the first accelerator to assign every output element independently to ANN or SNN execution mode with zero accuracy loss. We extend integer-exact ANN-SNN equivalence from layers to individual output elements, thereby enabling mode switching with no conversion error. An offline cost-guided scheduler scores each element by its marginal energy-delay trade-off and packs work across PEs, achieving 97-99% PE utilization compared to 40-45% for layer-wise hybrids. NeuroFlex reduces EDP by 57-67% over a strong ANN-only baseline and delivers up to 2.5x speedup over a dual-sparse SNN-only baseline. Our cost-guided scheduler improves throughput by 16-19% over random element assignment across vision, language, and transformer workloads.
[LG-92] Stabilizing Performative Feedback Loops with Minimal Model Deployments
链接: https://arxiv.org/abs/2609.14065
作者: Gabriele Farina,Juan Carlos Perdomo
类目: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT)
*备注:
Abstract:When algorithmic predictions inform people’s decisions, the models we deploy are performative and actively shape the data we see. This feedback loop between algorithms and their broader environments introduces a challenge in the mechanics of social prediction: If different predictive models induce different distributions, is it possible to efficiently learn a prediction rule that is optimal for the distribution that it induces? Formally, this solution concept is known as performative stability. A core challenge in learning a performatively stable predictor is that, unlike supervised learning where distributions are fixed, the learner must deploy different predictors and observe their induced distributions. The main contribution of our work is a new algorithmic procedure that, in the high-accuracy regime, finds a performatively stable model in nearly the minimum number of model deployments without making any assumptions regarding how predictions shape distributions. In particular, our procedure succeeds at finding a randomized performatively stable predictor using exponentially fewer model deployments than prior approaches. Our second main contribution is a structural result showing how this recent randomized notion of stability achieved by our algorithm can be derandomized into a single predictor satisfying the prior deterministic notion if one is willing to assume that the loss is well-conditioned and that performative effects are weak, as in early work in this area. On a technical level, our results come from building on an underexplored technical connection between performative stability and expected variational inequalities. Subjects: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT) Cite as: arXiv:2609.14065 [cs.LG] (or arXiv:2609.14065v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.14065 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-93] Data-Efficient Agent ic Graph Domain Adaptation via Reliability-Aware Prototype Learning
链接: https://arxiv.org/abs/2609.14045
作者: Yingxu Wang,Kunyu Zhang,Siyang Gao
类目: Machine Learning (cs.LG)
*备注:
Abstract:Agentic learning systems are often required to adapt after deployment by observing new data and reusing prior knowledge under limited supervision or feedback. For graph-structured prediction, Graph Domain Adaptation (GDA) naturally instantiates this setting by transferring knowledge from labeled source graphs to unlabeled target graphs under distribution shifts. However, most GDA methods assume sufficient labeled source graphs, which becomes restrictive in data-efficient agentic settings where only limited source evidence can be retained. Under such constraints, source semantics become unreliable, leading to unstable source anchoring, uncertain target association, and fragile targetmarginal calibration. To address these challenges, we propose DEAG, a reliability-aware prototype learning framework for data-efficient agentic GDA. DEAG estimates class reliability from retained source support and embedding compactness, and constructs stable reusable source anchors by blending empirical prototypes with classifier directions. Guided by these anchors, DEAG performs prototype-aware soft target association and aligns confidence-weighted target centers with source semantics. A source-prior regularizer further sharpens target predictions while keeping the target marginal consistent with retained source evidence. Experiments on graph benchmarks with diverse domain shifts show that DEAG improves average adaptation performance over competitive GDA baselines under the same source-data budget.
[LG-94] Real-World Deployment and Performance Characterisation of Fog-Based Deep Learning for Cold-Chain Temperature Prediction over LoRaWAN
链接: https://arxiv.org/abs/2609.14036
作者: Jeremiah Taguta,Jean Frederic Isingizwe Nturambirwe,Clement Nthambazale Nyirenda
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: 7 pages, 10th IEEE ICTAS 2026
Abstract:Fresh fruits and vegetables (FFVs) are highly perishable, and cold-chain breaks contribute significantly to global food waste. While Machine Learning (ML) can enable proactive intervention, cloud-based inference faces challenges such as latency and data loss. Fog computing addresses these issues but has been tested only in simulation for FFV cold-chain temperature prediction. To the best of the authors’ knowledge, this paper presents its first real-world deployment. A fog-deployed LSTM-GRU model predicted cold-room temperature using LoRaWAN sensor data collected from a South African apple cold-storage facility with induced cold-chain breaks. Running entirely on a Raspberry Pi 4 with no cloud dependency, the system generated conditional SHAP explanations only when a break is predicted. The deployed system predicts cold-room temperature with an MAE of 0.2°C at roughly 0.2 kWh per day (0.7 Wh per prediction). Predictions were delivered in under one second (555 ms), dominated by network and messaging rather than computation, with conditional explanations adding modest cost. SHAP consumes 28% more CPU but is well within the hardware’s capacity. The model attributes its predictions primarily to temperature, humidity, and their interaction. Critically, the deployment surfaced what simulation cannot: a sensor-triggered single point of failure, alongside genuine resilience, autonomous recovery from infrastructure faults and continued operation through internet loss. These are the first published deployment benchmarks for fog-based temperature prediction in FFV cold chains, establishing that explainable temperature forecasting is feasible on resource-constrained edge hardware. Future work includes asynchronous sensor fusion, commercial cold chain deployment, alternative model architectures, and causal analysis.
[LG-95] Diffusion-Based Multiple-Shooting Indirect Optimal Control for Fuel-Optimal Spacecraft Trajectory Generation
链接: https://arxiv.org/abs/2609.13990
作者: Saeid Tafazzol,Ehsan Taheri,Ryne Beeson
类目: ystems and Control (eess.SY); Machine Learning (cs.LG)
*备注: Submitted to IEEE Control Systems Letters (L-CSS) for possible publication
Abstract:Diffusion-based generative models (DMs) have found applications in control problems, and in particular robotics, where the DMs enable exploration of possible control solutions. A critical shortcoming of these applications is that they have lacked optimality guarantees. This is a problem for their potential use in fuel-optimal spacecraft trajectories that are characterized with long time-horizons and bang-bang profiles. Alternatively, indirect optimal control methods ensure explicit satisfaction of necessary conditions, but are highly sensitive to the initial costate estimation needed to solve the resulting Hamiltonian boundary-value problems (HBVPs). To alleviate this sensitivity and enlarge the convergence domain of HBVPs, advanced indirect methods have been developed that use smoothing approaches and continuation. We propose a diffusion-based multiple shooting indirect control method that combines the exploration capability of DMs with indirect method to generate fuel-optimal spacecraft trajectories. We benchmark our method against an advanced indirect method on a fuel-optimal Earth-Mars low-thrust transfer problem, showing higher convergence robustness than the advanced indirect method that is based on random costate initialization. Code and visualizations are available at this https URL.
[LG-96] Introspective Uncertainty Estimation for LLM -Based Code Generation
链接: https://arxiv.org/abs/2609.13975
作者: Thomas Klassert
类目: oftware Engineering (cs.SE); Machine Learning (cs.LG)
*备注:
Abstract:Large Language Models (LLMs) are increasingly used for code generation but can produce fluent yet functionally incorrect outputs, which limits trust in their usage for practical software engineering workflows. This thesis investigates whether Introspective Uncertainty Estimation (IUE), based on internal hidden-state representations of LLMs, can reliably indicate correctness at the response and line levels for code generation tasks. The objective is to determine the extent to which hidden states encode information about functional code correctness and how this can be leveraged for practical risk assessment and fault localization. Methodologically, this thesis combines response-level evaluation on LiveCodeBench (LCB) and BigCodeBench (BCB) with an augmentation pipeline that derives token- and line-level labels from incorrect programs. In this setup, it compares static and dynamic response-level features, evaluates generalization across tasks, programming domains, and token positions, and studies line-level fault localization. The results show that hidden states contain a strong response-level correctness signal. Static single-token probes perform best, while more elaborate dynamic strategies yield no consistent gains. While generalization across tasks, domains, and token positions is feasible, setting-dependent degradation largely remains for real-world software projects. At a fine granularity, line-level prediction is substantially harder than response-level estimation. However, in a conditional localization setup with known-incorrect programs, Top-K point-of-failure ranking remains effective. Overall, the findings suggest that hidden states are a robust and informative resource for estimating functional code correctness, supporting a two-stage workflow that combines response-level risk screening with targeted line-level prioritization. Subjects: Software Engineering (cs.SE); Machine Learning (cs.LG) Cite as: arXiv:2609.13975 [cs.SE] (or arXiv:2609.13975v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2609.13975 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-97] abby: An Open Pretraining Recipe for Time Series Foundation Models
链接: https://arxiv.org/abs/2609.13956
作者: Shifeng Xie,Bahaeddine Abdessalem,Zehao Xiao,Youssef Attia El Hili,Ambroise Odonnat,Zhiwei Dong,Lei Zan,Themis Palpanas,Jianfeng Zhang,Lujia Pan,Keli Zhang,Malik Tiomoko
类目: Machine Learning (cs.LG)
*备注: 43 pages, 3 figures, 32 tables. Technical report
Abstract:In this report, we release Tabby, a long context probabilistic time series foundation model, together with a complete and open recipe of how it was built. Tabby adopts an encoder-only patch Transformer architecture and concentrates the contributions on the data and the training procedure. The pretraining corpus combines an extended real-world collection, GIFT-Eval-Pretrain+ and BLAST, with synthetic data from KernelSynth and CauKerV2, an online generator that composes temporal dynamics through randomly sampled structural causal models. Training couples a progressive convergence schedule, which yields reusable intermediate checkpoints, with a deep quantile supervision objective for intermediate layers. The resulting 145M parameter backbone supports contexts of up to 8,192 observations and serves forecasting, classification, and anomaly detection, while a prompt-tuning module further improves in-distribution forecasting performance with the pretrained weights frozen. Tabby achieves competitive zero-shot forecasting performance on GIFT-Eval and the out-of-distribution TIME benchmark, while the same pretrained backbone also supports classification on the UCR Archive and zero-shot anomaly detection on TSB-AD-U. We release training pipeline and model as open source at huawei-noah/trustworthyAI.
[LG-98] Linear Ensemble Sampling with Smaller Ensembles
链接: https://arxiv.org/abs/2609.13954
作者: Taehyun Hwang,Min-hwan Oh
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Ensemble sampling offers a practical approach to randomized exploration by maintaining a collection of models, but how small an ensemble can be while retaining strong regret guarantees remains unresolved. In particular, the existing guarantees use an ensemble size of \Theta(d\log T) , leaving a logarithmic gap in the horizon T relative to the intrinsic \Omega(d) ensemble-size barrier. We aim to narrow this gap by proposing an ensemble sampling algorithm that refreshes the ensemble only when the regularized Gram matrix changes substantially. This mechanism localizes the perturbation analysis to epochs with controlled Gram-matrix drift and reduces the sufficient ensemble size to \Theta(d\log d+d\log\log T) , while preserving the state-of-the-art \tilde O(d^3/2\sqrt T) regret for ensemble sampling with arbitrary bounded arm sets. We further show that, when the arm set is finite of cardinality K , the proposed algorithm achieves the sharper regret bound \tilde O(d\sqrtT\log K) . To the best of our knowledge, this is the first ensemble-sampling guarantee that simultaneously recovers both canonical regret scalings known for randomized linear bandit algorithms: the \tilde O(d^3/2\sqrtT) rate for arbitrary bounded arm sets and the \tilde O (d\sqrtT\log K) rate for finite arm sets. The algorithm also admits an anytime implementation without resetting past data, and experiments show that it remains competitive with baselines while using substantially smaller ensembles.
[LG-99] Optimal Transport for Efficient Unsupervised Anomaly Detection on Industrial Data
链接: https://arxiv.org/abs/2609.13940
作者: Abigail Langbridge,Fearghal O’Donncha,James T Rayfield,Bradley Eck
类目: Machine Learning (cs.LG)
*备注:
Abstract:Effective anomaly detection frameworks are a central pillar of the Industry 4.0 paradigm. In this paper, we introduce an Optimal Transport (OT)-based framework for anomaly detection, designed to detect deviations from normal behaviour in time-series sensor data. The OT-based method requires minimal user input and adapts to real-time data without the need for labelled training data. Our method effectively addresses existing limitations related to data labelling, generalisability, and scalability, demonstrating resilience against short-term fluctuations, noise, and data gaps - common challenges in industrial environments. Additionally, our method provides counterfactual explanations improving the auditability of the approach when deployed in industrial settings. The proposed method learns the mapping between normal and observed operating conditions through a sliding reference window that adapts to the dynamicity of the data. We evaluate our approach on three industrial datasets, from shipping, industrial HVAC systems, and publicly available benchmark data. The method was highly effective in identifying anomalies and reducing false positives, outperforming traditional methods, while maintaining computational efficiency and ease of configuration.
[LG-100] Exploring napping paradigm for Recurrent Spiking Neural Networks
链接: https://arxiv.org/abs/2609.13927
作者: Andreas Massey,Stefano Nichele,Aliaksandr Hubin
类目: Machine Learning (cs.LG)
*备注: Accepted at the 7th International Workshop on Active Inference (IWAI 2026). To appear in Springer Communications in Computer and Information Science
Abstract:Biological organisms minimize free energy by balancing two competing demands on their internal world model: it must be accurate enough to predict sensory input, yet simple enough to generalize beyond it. Two mechanisms regulate this balance offline: sleep reduces complexity through gradual synaptic downscaling, while stochastic noise attenuates precision, relaxing the constraint sensory input imposes on synaptic reorganization. Engineered Spiking Neural Networks (SNNs) leave this balance unaddressed, favoring instantaneous, noiseless weight normalization instead. This paper investigates the hypothesis that a biologically inspired micro-sleep paradigm, napping – combining proportional weight scaling with continuous stochastic membrane activity – can replicate the stability of normalization while shedding model complexity. We evaluate this in an unsupervised recurrent SNN trained via trace-based spike-timing-dependent plasticity (STDP) on Gabor-preprocessed MNIST. We tune napping across three regularization regimes by sweeping its duration and membrane noise level, then compare the best configuration against weight normalization. Across all three regimes, well-tuned napping matches the accuracy of normalization: accuracy peaks at brief durations and low noise, then declines monotonically as either grows. Clustering diverges, with the strongest geometric separation arising at longer durations and higher noise – the two terms of free energy pulling apart, accuracy rewarding data fit and structure rewarding the simpler representation that gradual, noisy downscaling induces. This gain carries a simulation cost normalization avoids, so napping is most compelling where representational structure, rather than raw classification efficiency, is the priority.
[LG-101] Minibatch persistency eight years later: what batch reuse costs in steps and joules and what it saves in data
链接: https://arxiv.org/abs/2609.13922
作者: Matteo Fischetti
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注: 36 pages, 7 figures, 12 tables. Pre-registered study (OSF Registries, embargoed); registration, harness and per-run archives in the supplementary material of the journal submission
Abstract:Minibatch persistency reuses data instead of reading it: rather than drawing a fresh minibatch at every optimizer step, it takes K consecutive steps on the same one. Absorbed into data echoing in 2019, it has carried one objection – that reuse merely imitates a larger learning rate – and no baseline tuned as carefully as the method itself. This paper runs the missing test. A pre-registered study trains a 49M-parameter Transformer on FineWeb-Edu at minibatch size B in 32, 128, 512, 8 seeds per cell, tuning the learning rate separately for every batch size and every arm, against a reuse-free control that changes the sampling and nothing else. Each headline claim is a cost to reach a fixed loss, read on four axes: optimizer steps, fresh tokens, seconds, and joules at the socket. We then replicate on new seeds and a newer GPU generation, and put the three arms on one schedule in steps. The outcome of our study is that what minibatch reuse buys is neither speed nor energy but data, and only at large minibatch size: at B = 32 it reads more fresh tokens than the baseline, not fewer. On steps, seconds and joules it is at best free; and at B = 512, where it looks best, a registered control cannot separate the effect of reuse from the position on the learning-rate schedule at n = 8 seeds. The technique is therefore worth using where fresh data rather than compute is the binding cost: a corpus that runs out, a pipeline that pays per sample, a stream that cannot be rewound. Where the data can simply be read again, spaced epochs do as well or better.
[LG-102] Machine Learning in Fish Farming
链接: https://arxiv.org/abs/2609.13919
作者: Fearghal O’Donncha,Nikos Papandroulakis,Jennie Korus,Abigail Langbridge,Alexander Timms,Konstantinos Topouzelis,Abdul Baseer Khan,Shree Rama Kamal Kumar Vegu,Mahtab Sarvmaili,Ryan Mowat,Rhanna Turberville,Tyler Sclodnick,Christopher Whidden
类目: Machine Learning (cs.LG)
*备注:
Abstract:This chapter explores how machine learning (ML) is transforming aquaculture, with a particular focus on enhancing decision-making processes and improving operational efficiency. The chapter is structured to first introduce the challenges in aquaculture and the role of AI and then provide an overview of ML techniques in the context of aquaculture, followed by applications, emerging trends, future directions, and case studies. The focus is on real-world applications of ML techniques, including Random Forest, Convolutional Neural Networks (CNNs), and Recurrent Neural Networks (RNNs), as well as emerging technologies such as Graph Neural Networks (GNNs) and large language models (LLMs). Key applications include biomass estimation, species recognition, behavioural analysis, and environmental forecasting. The chapter also highlights the synergy between ML and the Internet of Things (IoT) for real-time monitoring and decision support. Ultimately, ML-driven innovations have the potential to revolutionise fish farming, leading to more efficient, sustainable, and productive practices in the aquaculture industry.
[LG-103] Learning Through Energy Refinement and Manifold Projection: A Cooperative EBM-AE Framework
链接: https://arxiv.org/abs/2609.13917
作者: Ryad Zemouri
类目: Machine Learning (cs.LG)
*备注:
Abstract:Energy-Based Models (EBMs) provide a flexible framework for generative modeling by learning an energy landscape that assigns low energy values to realistic samples and higher energies to unlikely observations. Despite their theoretical appeal, training EBMs remains challenging due to the computational cost of Langevin sampling and the difficulty of efficiently exploring the learned data manifold. In this work, we propose a cooperative Energy-Based Model and Autoencoder (EBM-AE) framework that combines energy-based refinement with manifold projection. The proposed approach jointly trains an EBM with a denoising autoencoder and introduces an iterative EBM \rightarrow AE \rightarrow EBM sampling procedure in which Langevin dynamics and autoencoder projection alternately refine generated samples. Within this framework, the autoencoder acts as a manifold projection operator that regularizes sampling trajectories, while the EBM performs energy-based refinement toward low-energy regions of the learned distribution. Extensive experiments conducted on the MNIST dataset demonstrate that joint EBM-AE training substantially improves generation quality compared with a conventional autoencoder. Beyond unconditional generation, we evaluate the proposed framework on image inpainting tasks involving structured and random masks. The results show that manifold projection provides the majority of the reconstruction capability, whereas the final energy-based refinement becomes increasingly beneficial as the reconstruction problem becomes more challenging. Taken together, the results indicate that combining manifold projection and energy minimization provides an effective and interpretable framework for generation, reconstruction, and out-of-distribution detection, while offering new insights into the complementary roles of energy-based modeling and representation learning. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2609.13917 [cs.LG] (or arXiv:2609.13917v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.13917 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-104] Machine Learning under Imperfect Data: Challenges and Methods
链接: https://arxiv.org/abs/2609.13914
作者: Masoumeh Zareapoor
类目: Machine Learning (cs.LG)
*备注:
Abstract:Machine-learning models are commonly developed under an assumption that training and test data are sufficiently complete, balanced, labelled, and drawn from compatible distributions. In practice, one or more of these conditions is often violated. Measurements may be missing or corrupted, rare classes may be poorly represented, supervision may be weak, and the deployment environment may differ from the training environment. These imperfections are usually treated as separate technical problems, although they alter learning through a small number of shared mechanisms: loss of information, biased empirical risk, ambiguous supervision, and unstable representations. This short survey organises representative methods around these mechanisms. It reviews reconstruction and generation, rebalancing and representation calibration, learning with limited supervision, adaptation across domains and modalities, and reliability under distribution change. The discussion highlights the limits of plausible reconstruction, benchmark-specific correction, and adaptation without trustworthy feedback. It concludes with directions for evidence-aware learning, uncertainty-preserving prediction, and evaluation that separates visual plausibility from decision utility.
[LG-105] HQARRF: Hierarchical Q-learning and Force-aware Routing for Multi-Charger Scheduling in Wireless Rechargeable Sensor Networks
链接: https://arxiv.org/abs/2609.13901
作者: Liang-Ching Tao,Pi-Chung Wang
类目: Networking and Internet Architecture (cs.NI); Machine Learning (cs.LG)
*备注: 15 pages, 5 figures, 1 table. Under review at MAMM 2026. Also archived at this https URL
Abstract:Multi-charger scheduling in wireless rechargeable sensor networks must weigh sensor death risk, charger energy, travel cost, return-to-base feasibility and inter-charger coordination at once, and schedulers driven by local urgency alone duplicate service and leave whole regions unattended. We present HQARRF, a two-level scheduler. Below, an interpretable ARR-F score ranks candidate clusters through an attraction term for local urgency, a repulsion term against charger crowding and a force bonus from nearby critical sensors. Above, adaptive zones compress regional state into a deadline-based risk estimate, and a gated Q-learning controller decides only whether to redirect service to a high-risk, under-served zone. Over 27 parameter points HQARRF attains the highest mean survival rate at 26, improving survival by 20.7 percentage points over the mean of five baselines and 9.2 over the strongest baseline at each point. An ablation isolates the upper level: its gain tracks how often the controller fires.
[LG-106] Lie to me: Detecting Managerial Evasiveness in Earnings Calls via Conversational Audio Encoders
链接: https://arxiv.org/abs/2609.13893
作者: Huizhong Chen,Huan Zhang
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)
*备注:
Abstract:Earnings conference calls are a primary channel through which managers disclose information under analyst scrutiny. Prior work has linked vocal and lexical cues to future adverse outcomes, but often pools features over an entire call and underuses the interactive structure of QA. We propose a two-branch late-fusion framework for detecting managerial evasiveness as a predictor of extrinsic SEC events (primarily late filings): (i) an LLM-as-a-judge that maps QA text to an interpretable call-level vector X_text via a structured binary rubric, and (ii) a frozen conversational encoder whose temporal hidden states are read by a DeepVoice-style sequential reader to produce an audio representation h. Late fusion of (X_text, h) yields a call-level risk score p. On n=1,039 calls (212 late filings) with firm-grouped 5-fold CV, fusion reaches AUROC approx. 0.89, versus 0.55 for the text judge and 0.71 for duration alone. These results show that conversational audio dynamics encode managerial evasiveness beyond lexical content and call length, yielding a stronger early-warning signal of adverse SEC outcomes.
[LG-107] A Multi-Resolution Multi-Domain Pre-Training Framework for Universal Traffic Forecasting ICDM2026
链接: https://arxiv.org/abs/2609.13878
作者: Zhouyang Liu,Jindong Han,Hao Wang,Xinyue Liu,Hui Gao,Dongsheng Li,Hao Liu
类目: Machine Learning (cs.LG)
*备注: Accepted by ICDM 2026
Abstract:Spatio-temporal traffic data are central to intelligent transportation systems, yet their heterogeneity poses significant challenges for large-scale modeling. Existing pre-trained models often rely on a homogeneous modeling paradigm to handle highly heterogeneous traffic data. This fundamental mismatch not only limits model generalization but also leads to computationally expensive and parameter-inefficient designs. To this end, we propose FlexST, a novel pre-training framework that introduces modularity and adaptivity for traffic modeling. Specifically, we first propose a multi-resolution spatio-temporal diffusion module that captures both short-term fluctuations and long-range trends, effectively reconciling inputs with divergent temporal and spatial resolutions. After that, we construct a domain-adaptive mixture-of-experts that dynamically routes data to specialized sub-networks, enabling selective knowledge transfer while preventing negative interference across diverse domains. Moreover, we devise a unified periodic encoding strategy that injects resolution- and domain-aware inductive biases to harmonize periodic inconsistencies across datasets. Extensive experiments on 23 real-world traffic datasets demonstrate that FlexST significantly outperforms state-of-the-art baselines in zero- and few-shot settings, showcasing superior generalization, adaptability and efficiency. This work offers a new direction for building general-purpose pre-trained models capable of handling the complexity and variability of urban traffic systems.
[LG-108] From Network Inequality to Network Fairness: A Perspective on Responsible Decision-Making
链接: https://arxiv.org/abs/2609.13867
作者: Lisette Espín-Noboa,Tina Eliassi-Rad,Pak-Hang Wong,Erich Prem,Meike Zehlike,Ricardo Baeza-Yates,Suresh Venkatasubramanian,Fariba Karimi
类目: ocial and Information Networks (cs.SI); Computers and Society (cs.CY); Machine Learning (cs.LG); Physics and Society (physics.soc-ph)
*备注: Perspective paper. 31 pages: 14 main, 10 references, 7 SI. 7 figures: 2 main, 5 SI. 4 tables: 1 main, 3 SI
Abstract:Social networks shape how individuals make decisions and how opportunities are distributed. However, the mechanisms that generate these networks often reflect pre-existing inequalities, and technologies that rely on network-derived signals risk further amplifying such disparities. Algorithmic fairness research largely treats networks as a fixed background, grounding analysis almost exclusively in distributive justice and overlooking how network structures systematically bias decision-making. In this Perspective, we identify ten network effects and trace how they create structural biases in the relationship between what we intend to measure and what we observe. Using academic hiring as an example, we show that network biases are not inherently harmful or beneficial. Determining their legitimacy requires examining the entire decision-making process through the lenses of both distributive and procedural justice while engaging all affected stakeholders. We therefore call for a holistic, networked approach to fairness that moves beyond static group categories and recognizes the dynamic, relational, and structural nature of inequality.
[LG-109] he Filter Metric is Safety-Critical: Phantom Advantages in Group-Relative RL under Shaped Rewards
链接: https://arxiv.org/abs/2609.13866
作者: Juntao Yu
类目: Machine Learning (cs.LG)
*备注: 11 pages, 5 figures
Abstract:Group-relative policy optimization (GRPO and descendants) can discard no-contrast rollout groups through dynamic sampling, while practical implementations expose a configurable filter metric. We identify and quantify a metric-predicate mismatch under composite shaped rewards. When filtering follows the shaped training score rather than the task outcome, all-fail groups retain nonzero within-group spread and pass the predicate; standard-deviation normalization then promotes shaping differences among failures to full-size phantom advantages. In a controlled GSM8K comparison (Qwen2.5-1.5B, LoRA), no filtering and shaped-score filtering end at EM 0.080 +/- 0.112 and 0.040 +/- 0.008, whereas binary-outcome filtering holds 0.754 +/- 0.005 across four runs per arm (three default-seed reruns and one seed-123 run; mean +/- sample SD). On verl’s native recipe/dapo trainer, holding model, data, reward and trainer fixed and changing only the metric, the score arm requires no batch refill in any of 40 observed steps and ends at EM 0.160; the accuracy arm refills in 29/40 steps and ends at 0.763. Both use the same custom shaped-reward hook and unmodified trainer/filter code. Prior work established shaping-induced amplification and all-fail filtering; our contribution isolates the metric-predicate semantic mismatch and directly instruments native deletion/refill telemetry. Across tested positive coefficients lambda in 0.1, 0.3, 0.5, unsafe arms collapse; exploratory one-run cells reproduce the failure at 1.5B/7B on MATH and under GSPO, while disabling standard-deviation normalization avoids the observed collapse. Filtering under a composite reward should use a task-outcome signal whose semantics are independent of shaping.
[LG-110] An Uncertainty-Aware Hybrid Mathematical-Machine-Learning Model for Smart Irrigation Decision Support
链接: https://arxiv.org/abs/2609.13864
作者: Andrea Scariolo
类目: Machine Learning (cs.LG); Applications (stat.AP)
*备注: 19 pages, 8 figures, 7 tables. Code: this https URL
Abstract:Agriculture accounts for roughly 70% of global freshwater withdrawals, yet irrigation is still commonly scheduled reactively, with no forecast of where soil moisture is heading and no statement of confidence in that forecast. Data-driven models are accurate but opaque and point-valued; water-balance models are transparent but carry large structural error. Neither alone supports a defensible irrigation decision under uncertainty. This study coupled the two and carried uncertainty through to the decision: a four-parameter water-balance core, calibrated on training data only, was corrected by a Random Forest that learned nothing but the physical residual, conformal prediction attached 90%-nominal intervals, and a risk-aware rule converted the interval lower bound into an irrigation trigger. It was evaluated on three years of hourly in-situ measurements from a rainfed Mediterranean cropland station under a strict chronological split, scored against persistence, from one hour to one week. At the 24 h horizon the hybrid reached RMSE 0.00925 m^3 m^-3 and +9.4% skill, roughly double the best of nine baselines, of which only the Random Forest beat persistence. Skill did not grow with lead time: it peaked at +27.4% at three hours and fell to +1.2% at one week. Conformalised quantile regression was better calibrated and 11% sharper than constant-width conformal prediction. The risk-aware rule raised management-threshold crossings detected in advance from 0.905 to 1.000, at a precision cost of 0.975 to 0.950 and 3.7% more notional water, and beyond 72 h the point forecast fell below the no-forecast rule while the interval-based rule did not. Uncertainty quantification therefore governs the lead time over which forecast-driven irrigation advice remains trustworthy, and here transparency in the physical layer cost no measurable accuracy.
[LG-111] ReH-FUSE: Reliability-Aware Hierarchical Fusion of Experts for Multimodal Emotion Recognition in Conversation ICASSP2027
链接: https://arxiv.org/abs/2609.13857
作者: Guan-Hua Wen,Hou-Chiang Tseng,Kuan-Yu Chen
类目: Machine Learning (cs.LG)
*备注: 5 pages, 2 figures, 6 tables. Submitted to ICASSP 2027
Abstract:Multimodal emotion recognition in conversation (ERC) requires adapting to the instance-dependent reliability of different evidence sources. Lexical content may be decisive, vocal expression may provide complementary cues, or accurate recognition may require cross-modal interaction; fixed fusion does not explicitly account for this variation. We propose ReH-FUSE, a reliability-aware framework with dialogue-aware text, audio, and cross-modal experts. Its decision-level router first models the relative preference between text and audio and then balances the resulting unimodal mixture against the cross-modal expert. This factorization separates unimodal competition from cross-modal selection. Across three independent runs on IEMOCAP, ReH-FUSE achieves 74.34% weighted F1 and 73.11% macro F1; on MELD, it achieves 68.03% weighted F1. Controlled ablations show that learned routing outperforms uniform expert averaging and benefits from cross-modal interaction.
[LG-112] Pre-training with Graph Transformers KDD2023
链接: https://arxiv.org/abs/2609.13844
作者: Jiaming Wang,Thomas Laurent,Xavier Bresson
类目: Machine Learning (cs.LG)
*备注: 4 pages, 1 table. DLG-KDD 2023 workshop paper
Abstract:This article investigates pre-training strategies for graph transformers in the biochemistry domain. By conducting comprehensive experiments, the study reveals that supervised pre-training using computed properties as labels provides the highest performance gain on downstream tasks. The results also highlight the importance of constraining model capacity to mitigate overfitting in graph transformers.
[LG-113] Accuracy Is Not Service: A Decision-Aware Benchmark for Intermittent-Demand Forecasting ICDM-26
链接: https://arxiv.org/abs/2609.13840
作者: Joo Ern Chin,Shih-Fen Cheng,Aldy Gunawan
类目: Machine Learning (cs.LG)
*备注: Accepted to the Twenty-Sixth IEEE International Conference on Data Mining (ICDM-26)
Abstract:A contract-logistics spare-parts operator is paid on order-level service: an order counts only if every requested line is fulfilled, yet forecasters are selected based on line-level forecast accuracy. This disconnect matters when demand is intermittent and lumpy, histories are short, and lead times span months. We benchmarked 38 forecasting methods spanning classical, intermittent-demand, machine-learning, deep-learning, and pretrained foundation models. A common decision-aware protocol evaluates them on an industrial panel drawn from a live contract and two public datasets. Forecast-accuracy rank and order-service rank are negatively correlated on the industrial panel, at -0.555, across methods evaluated on 20,330 real multi-item orders. Service is more closely associated with the direction of cumulative forecast bias, including over-prediction during zero-demand periods, than with point accuracy. Examining bias in Chronos-2’s instance normalization yields a training-free correction that lifts the per-material fill proxy from 77.5% to 92.0% (14.5 percentage points) at the 90% policy target and raises the complete-order fill rate from 54% to 63%. For reproducibility, we release RUF (Regenerate-Until-Fidelity), a method for generating fidelity-certified synthetic panels on which the findings reproduce. For intermittent demand, the lowest-error forecast need not deliver the highest service. Bias direction helps explain this gap, which can be reduced without retraining.
[LG-114] Graph Neural Networks for Influence Maximization in Social Networks: An Unsupervised Minimum Dominating Set Approach
链接: https://arxiv.org/abs/2609.13836
作者: Erfan Ahmadi,Mina Shirazi,Behnam Bahrak
类目: Machine Learning (cs.LG); Social and Information Networks (cs.SI)
*备注: 16 pages, 1 figure, 4 tables. Submitted to the International Journal of Data Science and Analytics
Abstract:The Minimum Dominating Set (MDS) problem is a classic NP-hard combinatorial optimization problem with critical applications in social network analysis, including viral marketing, influence maximization, public health interventions, and information dissemination. Identifying a minimal set of influential individuals whose reach covers an entire social network is central to these applications, yet remains computationally challenging at scale. Graph neural networks (GNNs) have emerged as powerful tools for learning over graphs, and recent work explores their application to hard combinatorial problems. This paper presents a novel unsupervised GNN framework for the MDS problem that eliminates the need for ground-truth solutions during training. Trained on 12,000 synthetic graphs with diverse structural properties, our method achieves up to 55x faster inference than metaheuristic baselines and up to 14x faster inference than supervised learning approaches, while finding optimal or near-optimal dominating sets on real-world social network benchmarks. Our learned heuristic generalizes effectively to unseen graph distributions, demonstrating strong practical applicability for large-scale social network analysis.
[LG-115] Benchmarking Optimizers to Solve Inverse Problems with Differentiable Physics Simulators
链接: https://arxiv.org/abs/2609.13819
作者: Xiang Chen,Huanhuan Xia
类目: Machine Learning (cs.LG); Optimization and Control (math.OC); Computational Physics (physics.comp-ph)
*备注:
Abstract:Solving inverse problems with differentiable physics simulators holds the potential to revolutionize scientific discovery and engineering design, as it enjoys both the strict physical correctness from rigorous numerical physics simulators, and the high efficiency and effectiveness from automatic differentiation and gradient-based optimization. However, currently, this paradigm faces performance issues in optimization. In this work, we target benchmarking the performance of different optimizers to solve various inverse problems. We construct 12 differentiable physics simulators spanning physics domains including discrete mechanics, continuous mechanics, atomistic simulations, rendering, and semi-empirical physics models. Based on these simulators, we design corresponding inverse problems that can be categorized into parameter identification, inverse design, and optimal control. Finally, we conduct extensive experiments to compare the performance of different optimizers, including regular first-order methods, approximate second-order methods, as well as global optimizers, on these inverse problems, and analyze the results to provide insights on how to choose and design optimizers for differentiable programming. We hope such benchmarks can inspire the development of more effective optimizers, and further promote the applications of differentiable programming in various scientific and engineering domains.
[LG-116] PPDL: A Real-world Industrial User Retention Ratio Forecasting Framework Integrating Physical Priors with Deep Learning ICDM2026
链接: https://arxiv.org/abs/2609.13789
作者: Zibo Zhao,Zhengxiong Guan,Chaoli Zhang,Linyuan Geng,Xuanbing Zhu,Zhonglong Zheng,Fan Wu
类目: Machine Learning (cs.LG)
*备注: Accepted by ICDM 2026
Abstract:In multi-channel paid user acquisition, early and accurate prediction of user retention at the channel level is crucial for optimizing budget allocation. User retention curves display a pronounced temporal pattern: an initial period of high churn transitions into long-term stability. This pattern is further characterized by regular fluctuations attributable to seasonality and exhibits high serial autocorrelation. These intrinsic properties make such curves highly suitable for analysis within a time-series forecasting framework. However, forecasting user retention ratio for large-scale short-video platform faces three major challenges: significant heterogeneity across channels, pronounced global trend of decay followed by saturation, and short look-back windows. To address these challenges, we propose PPDL, a novel forecasting framework that integrates physical priors with deep learning. We first introduce a trend-residual decomposition component. The trend is modeled using the Weibull distribution, whose parameters are learned via a Multilayer Perceptron (MLP). Secondly, for the residual component, we design an auxiliary embedding module on top of a deep learning backbone to maintain the channel identity awareness. Finally, to enhance the model’s sensitivity to trends, we design a Multiscale Trend-penalized loss function. The proposed approach PPDL is validated through comprehensive experiments on industrial-scale datasets, covering three applications with an average of 30+ channels each. Experimental results show that PPDL achieves improvements across different backbones and significantly outperforms existing online solutions.
[LG-117] JumpStart Your Policy Learning with Lessons from 160000 Training Runs
链接: https://arxiv.org/abs/2609.13730
作者: Nabil Omi,Eric Bae,Chung Yik Edward Yeung,Siddhartha Sen,Ali Farhadi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Reliable progress in offline policy learning depends on careful reporting, well-tuned baselines, and evaluation across diverse conditions. Prior work has shown that results can be sensitive to reporting choices, hyperparameter tuning, and dataset properties, but these sources of variability have not been systematically investigated together at the scale needed to understand how they shape conclusions. To address this gap, we present a large-scale empirical study of offline reinforcement and imitation learning, training over 160,000 policies across 114 datasets. At this scale, no algorithm dominates: aggregate performance among the strongest methods is often close, but the leaders differ substantially across environments. We find that proper hyperparameter tuning frequently reshuffles perceived algorithm rankings and that benchmark composition can produce conflicting conclusions. We also study hyperparameter sensitivity and transfer across environments, identifying a simple strategy for deriving strong default configurations. We use our findings to develop a dataset-conditioned recommender that provides task-specific algorithm recommendations for practitioners. Finally, we release JumpStart: a resource suite containing every trained policy, per-model scores and hyperparameters, strong baselines across all environments, training and evaluation code, and an extensible website for retrieving, analyzing, and contributing results. Together, these resources aim to make offline policy-learning research more reliable and enable future work beyond the scope of this study.
[LG-118] A Variational Optimal Transport Operator on Incompressible Flow
链接: https://arxiv.org/abs/2609.13729
作者: Jinjin He,Shenyifan Lu,Sinan Wang,Zhiqi Li,Duowen Chen,Bo Zhu
类目: Machine Learning (cs.LG); Graphics (cs.GR)
*备注:
Abstract:We present the Variational Incompressible Optimal Transport (VIOT) operator, a generative neural operator for amortized incompressible density transport. Given a new source-target density pair, VIOT predicts a divergence-free velocity field and generates the full transport trajectory by feed-forward inference, replacing the hour-scale per-pair optimization used by adjoint fluid solvers and differentiable simulation baselines. The system consists of three components: a stream-function or vector-potential representation that enforces incompressibility by construction, a regularized incompressible transport objective that balances endpoint accuracy and flow smoothness, and a Fourier Neural Operator backbone that amortizes the solve across new pairs and grid resolutions. Together, these components make incompressible transport a reusable neural operator that facilitates various transport processes. Further, the generative capability extends beyond the training distribution, with VIOT producing incompressible transports for user-drawn source-target pairs in a real-time interactive system. We demonstrate VIOT on 2D and 3D density-transport benchmarks. Both 2D and 3D rollouts complete in seconds per pair, while per-instance baselines in our 2D comparisons optimize each new pair from scratch and require on the order of an hour, a roughly 10^4\times online speedup.
[LG-119] Certifying Model Upgrades with Slice-Wise Non-Regression and Incumbent Fallback
链接: https://arxiv.org/abs/2609.13714
作者: Shengwei Zhang,Tao Wu,Fei Qian
类目: Machine Learning (cs.LG)
*备注: 21 pages, 4 figures
Abstract:An updated model can improve an aggregate metric while degrading a slice that matters to a downstream user. We study checkpoint selection subject to non-regression tolerances relative to a retained incumbent. The central distinction is between failing to detect harm and certifying non-inferiority: the former can release harmful updates with high probability when evaluation is noisy. We give a reproducible release procedure that separates candidate search from independent, paired evaluation and returns the exact incumbent when certification fails. Applying established intersection-union and Learn-then-Test principles, we state finite-sample guarantees for one frozen candidate, a finite candidate library, and a prespecified testing order. A joint release decision does not require a slice-count Bonferroni penalty, although certification power can still decrease with the number of slices. In bounded-score simulations, a no-detected-harm gate releases a harmful candidate in 99.7% of trials in one 32-slice setting, compared with 2.6% for an exact non-inferiority gate at a 5% target. A constructed two-block family yields larger certified utility than a scalar path under matched candidate counts. Public digits experiments, including a subsequent continuation that improves average aggregate accuracy, return the incumbent in every run because certification is underpowered. These results establish an auditable protocol and its limitations; they do not establish benefits on foundation-model or multilingual translation upgrades.
[LG-120] Not All Duplicates Are Coordination: Generic vs. Non-Generic Duplicate Campaigns in Information Operations EMNLP2026
链接: https://arxiv.org/abs/2609.13671
作者: Ashfaq Ali Shafin,Khandaker Mamun Ahmed
类目: ocial and Information Networks (cs.SI); Machine Learning (cs.LG)
*备注: Accepted in the 11th Workshop on Natural User-generated Text (W-NUT collocated with EMNLP 2026)
Abstract:Duplicate content is widely used to study coordinated behavior in social media information operations (IOs), but not all repetition provides equally meaningful evidence of coordination. Generic, reusable, or low-information posts may create noisy account-account links when projected into coordination graphs. We study this problem using 187,000 English-language tweets from six Russian Twitter Information Operations datasets. We introduce a generic/non-generic distinction for duplicate campaigns, label tweets using an LLM-assisted protocol with independent human validation, and train supervised classifiers over sentence embeddings to scale the labels. We construct duplicate campaigns using lexical similarity and two embedding-based methods. Generic campaigns are rare under lexical matching but account for nearly 39% of campaigns detected by embedding-based methods. Restricting graphs to non-generic campaigns reduces graph size and the largest connected component while increasing density, suggesting a smaller but more focused coordination structure. These findings show that duplicate-based coordination analysis should consider both textual similarity and semantic specificity.
[LG-121] Online Bayesian Node Classification on Inductive Graphs under Distribution Shift
链接: https://arxiv.org/abs/2609.13655
作者: Jinwen Xu,Gonzalo Mateos Buckstein,Qin Lu
类目: Machine Learning (cs.LG); Methodology (stat.ME); Machine Learning (stat.ML)
*备注:
Abstract:On evolving graphs, node classifiers must satisfy two key requirements: inductive generalization to newly arriving nodes under distribution shift and calibrated uncertainty for safety-sensitive applications. Standard graph neural networks (GNNs) are typically trained once and address neither requirement. We adapt the Bayesian last-layer (BLL) model by placing random last-layer parameters on top of a deterministic GNN encoder for uncertainty quantification. The categorical softmax likelihood required for classification breaks Gaussian conjugacy, so neither the training posterior nor the test-time streaming update has a closed-form solution. To address both challenges, we introduce a variational Bayesian last-layer (VBLL) objective that jointly trains the encoder and an approximate last-layer posterior by maximizing an evidence lower bound with a Monte Carlo expected log-likelihood. At test time, we freeze the encoder and apply an online Laplace update to the last-layer posterior. This update corresponds to a power-prior Bayesian model with exponential forgetting and a Kullback-Leibler anchor to the training posterior. Across five node-classification benchmarks under distribution shift, online GVBLL is the only method to achieve the best accuracy and negative log-likelihood on every dataset. It improves accuracy by up to 17 percentage points on Cora and 14 percentage points on ogbn-arxiv over the strongest non-GVBLL baseline, while remaining competitive in calibration with MC Dropout, Deep Ensembles, Temperature Scaling, and Gaussian-process classifiers.
[LG-122] Curvature-Independent Regret Bounds for Distributed Online Optimization on Hadamard Manifolds
链接: https://arxiv.org/abs/2609.13646
作者: Zhanyuan Cai,Emre Sahinoglu,Shahin Shahrampour
类目: Machine Learning (cs.LG); Optimization and Control (math.OC); Machine Learning (stat.ML)
*备注: 6 pages, 2 figures
Abstract:This work addresses decentralized online Riemannian optimization on Hadamard manifolds. Prior work under geodesic convexity (g-convexity) may require curvature information in the optimization analysis, typically through a finite lower bound on the sectional curvature. Curvature may also enter the step size or contraction factor of tangent-space Riemannian consensus schemes. In this work, we relax the curvature dependence for a narrower class of horospherical convex (h-convex) functions. We study Distributed Riemannian Online Gradient Descent (D-ROGD), which combines local Riemannian h-subgradient updates with an implicit Fréchet-mean consensus. For h-convex and strongly h-convex local objectives, we establish O(\sqrtT) and O(\log T) static regret, respectively, matching the corresponding Euclidean rates with respect to T , with network dependence governed solely by the spectral gap. To our knowledge, these are the first curvature-independent regret guarantees for decentralized online optimization on Hadamard manifolds. Experiments on hyperbolic embeddings corroborate the predicted rates, with no observable degradation due to curvature.
[LG-123] When Compliance Data Masquerades as Evaluation: Measurement Validity for Deployed AI Systems
链接: https://arxiv.org/abs/2609.13642
作者: Hung-Yu Lin,Xingran Huang,Qiming Guo,Jinwen Tang
类目: Machine Learning (cs.LG)
*备注:
Abstract:We argue that a recurring failure in the evaluation of deployed AI systems occurs when data collected for operational monitoring or regulatory compliance are interpreted as if they were designed for comparative evaluation. Automated driving provides a concrete example of this problem. U.S. disengagement and crash-reporting regimes produce valuable operational evidence, but differences in reporting scope, exposure, deployment domain, event capture, and comparator construction limit the safety claims that can be supported from these measurements alone. We frame this issue as a measurement-validity problem in AI evaluation rather than as a transportation-specific data limitation. We argue that comparative claims about deployed AI systems require alignment between the intended capability, measured outcome, exposure opportunity, deployment domain, data-generation process, and evaluation comparator. Using automated-driving safety evaluation as a case study, we propose an evaluation contract that makes these assumptions explicit before operational data are interpreted as evidence of comparative performance. The broader implication is that data useful for monitoring deployed AI systems are not automatically valid benchmarks for evaluating them.
[LG-124] FlowTSFM: Turning Encoder Depth into Quantile Transport
链接: https://arxiv.org/abs/2609.13640
作者: Bahaeddine Abdessalem,Shifeng Xie,Zehao Xiao,Youssef Attia El Hili,Ambroise Odonnat,Jianfeng Zhang,Lujia Pan,Keli Zhang,Malik Tiomoko
类目: Machine Learning (cs.LG)
*备注:
Abstract:Encoder-based time series foundation models (TSFMs) typically rely on deep stacks of independently parameterized Transformer layers, where only the final forecast is supervised and intermediate representations have no explicit predictive role. We introduce FlowTSFM, an encoder architecture that interprets depth as a recurrent transport process: a single Transformer block is iteratively applied with shared parameters, while a quantile-flow objective supervises intermediate states along a prescribed trajectory from a prior distribution toward the final forecast. The objective combines pinball forecasting loss with path-level position matching. With only 38.8M parameters, FlowTSFM achieves competitive performance on GIFT-Eval and TIME, remaining within 1.8-4.6% MASE of stronger baselines while using approximately 3\times fewer parameters than a 12-layer Chronos-2 model (119.5M). Beyond accuracy, we introduce CosMean, a scale-free diagnostic measuring whether recurrent updates consistently align toward the final prediction. Under a matched intermediate-state probing protocol, FlowTSFM achieves a CosMean score of 0.919 compared with 0.350 for Chronos-2, suggesting that recurrent parameter sharing combined with path supervision is associated with substantially more structured predictive trajectories at a favorable accuracy-efficiency trade-off.
[LG-125] EI-DDLGN: Efficient Encrypted Inference with Deep Differentiable Logic Gate Networks under TFHE
链接: https://arxiv.org/abs/2609.13636
作者: Mahmoud Y. M. Yassin,Mahmoud AbdelHafeez Sayed,Mostafa Taha
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: Accepted at the 19th International Symposium on Foundations and Practice of Security (FPS 2026)
Abstract:Privacy-preserving inference via Torus Fully Homomorphic Encryption (TFHE) provides strong protection for sensitive data in outsourced deep learning applications. However, most TFHE-compatible neural network frameworks remain based on arithmetic neural architectures, resulting in high inference latency due to programmable bootstrapping (PBS), accumulator growth, and circuit bit-width sensitivity. In this work, we investigate Deep Differentiable Logic Gate Networks (DDLGNs) as a Boolean-native alternative for encrypted inference under TFHE. Because DDLGNs learn Boolean computations directly and discretize into fixed logic gate networks, their inference procedure is naturally aligned with TFHE’s Boolean execution model and avoids arithmetic accumulation in hidden layers. We present EI-DDLGN, the first in-depth study of TFHE-based DDLGN inference, and characterize how encrypted execution cost depends on model size, learned Boolean-function distribution, and propagated wire status. We also introduce Model-Fixed-Wire PBS Bypass (MFW-PBS Bypass), a semantics-preserving execution strategy that eliminates unnecessary PBS operations without modifying the learned network topology. Evaluations across 72 depth-width configurations on MNIST, FashionMNIST, and UCI Phishing show that DDLGNs constitute an efficient alternative to arithmetic TFHE inference, achieving substantially improved accuracy-latency trade-offs. Notably, on MNIST, EI-DDLGN-Small matches the accuracy of QAT-FCNN-4 while reducing encrypted inference latency by 13.4x. Our implementation is available at this https URL
[LG-126] AttnFuse: A Composable DSL for Compiling Attentions to Fused GPU Kernels
链接: https://arxiv.org/abs/2609.13612
作者: Varun Kumar Dasoju,Tian Zhao
类目: Machine Learning (cs.LG)
*备注:
Abstract:Modern AI systems are built on the Transformer architecture, whose core operation, attention, accounts for the majority of computation and memory cost. Researchers continually propose new attention variants to improve quality, efficiency, or context length, but each variant currently requires expert-written GPU code to run at usable speeds. PyTorch’s recent flex_attention lets researchers describe custom attention patterns in Python and compile them to fused kernels, but its design is limited to modifications applied after the central matrix multiplication, excluding Rotary Position Embedding (RoPE), the positional encoding used by every major LLM. We introduce AttnFuse, a small DSL for attention that makes pre-multiplication transformations like RoPE first-class operations. Researchers compose ten high-level building blocks to describe a variant, and AttnFuse’s compiler emits a single fused GPU kernel for the entire computation. On an RTX 3090, AttnFuse achieves a 2.10 \times speedup over flex_attention on the RoPE+causal pattern. On an H100, it runs a full Llama-3-8B training step within 5% of PyTorch’s hand-tuned backend. Our investigation reveals the Rotation Calculus: whether to fuse RoPE or apply it separately depends on the GPU’s compute-to-bandwidth ratio, with a derived crossover that matches measurement. AttnFuse demonstrates that a small, attention-specific compiler can close the gap between flexible research code and production kernels. Subjects: Machine Learning (cs.LG) MSC classes: 68T07 (Primary) 68W40, 68T50 (Secondary) ACMclasses: I.2.7; D.3.2; C.1.4 Cite as: arXiv:2609.13612 [cs.LG] (or arXiv:2609.13612v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.13612 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-127] BOOST: Concurrent Access to Host Memory and HBM to Accelerate LLM Inference
链接: https://arxiv.org/abs/2609.13592
作者: Anish Saxena,Jae Hyung Ju,Hritvik Taneja,Po-An Tsai,Aamer Jaleel,Christos Kozyrakis,Moinuddin Qureshi
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Hardware Architecture (cs.AR); Machine Learning (cs.LG)
*备注: 15 pages, 22 Figures
Abstract:GPU memory bandwidth and capacity limit throughput in large language model (LLM) inference. The GPU memory system consists of a primary tier of high-bandwidth memory (HBM) and a secondary tier of host memory connected via CPU-to-GPU interconnect. Current serving systems treat the tiers hierarchically: they serve exclusively from HBM when data fits, and otherwise prefetch data from host memory to HBM before use. In both cases, the host memory bandwidth is never well utilized. Prefetching expands capacity by utilizing host memory, but consumes HBM bandwidth for writes, reducing the bandwidth available for demand loads. We observe that fully utilizing both host and HBM bandwidth requires each wave of GPU threadblocks to access both tiers concurrently and in proportion to their bandwidth ratio. Existing bandwidth-proportional placement strategies fail to provide concurrency because they are not aware of GPU waves, and the large 2MB GPU page size. This paper presents BOOST, the first runtime system that provides concurrent and proportional access to both GPU memory tiers, extracting the combined bandwidth of host memory and HBM for LLM inference without kernel changes. The key insight in BOOST is to use kernel access patterns to make page allocation and runtime data management wave-aware. For static model weights, BOOST applies modulo-based page placement that eliminates access-ratio variance; for dynamically provisioned attention key-value (KV) pairs, it makes the free KV page pool wave-aware. We integrate BOOST into vLLM and evaluate on a Grace Hopper system. At iso-batch size, BOOST improves Time-per-Output-Token (TPOT) by 4.3% over HBM-only serving, whereas prefetching degrades TPOT by 6%. In high-throughput serving, BOOST improves throughput by 31% on average, outperforming prefetching by 15%. Comments: 15 pages, 22 Figures Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Hardware Architecture (cs.AR); Machine Learning (cs.LG) ACMclasses: C.1.4 Cite as: arXiv:2609.13592 [cs.DC] (or arXiv:2609.13592v1 [cs.DC] for this version) https://doi.org/10.48550/arXiv.2609.13592 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-128] When Greedy Sampling Explores: KL-Regularized Contextual Bandits without Eluder-Dimension Dependence
链接: https://arxiv.org/abs/2609.13564
作者: Zichen Wang,Haoyang Hong,Huazheng Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study KL-regularized contextual bandits under both reward and preference feedback. We show that greedy sampling can achieve logarithmic regret without explicit dependence on the eluder dimension. For reward feedback, we establish an eluder-dimension-independent regret bound for a simple greedy algorithm that directly samples from the Gibbs policy induced by the estimated reward. We further extend this result to preference feedback under both the general preference and Bradley–Terry models, while also sharpening existing dimension-dependent guarantees. Our analysis reveals a trade-off between greedy sampling and upper confidence bound-style exploration: greedy sampling enjoys stronger guarantees when KL regularization is sufficiently strong, whereas additional exploration becomes preferable as the regularization weakens.
[LG-129] oward Optimal Switching Regret for Multi-Armed Bandits with Oblivious Adversary
链接: https://arxiv.org/abs/2609.13547
作者: Mengxiao Zhang
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:We study switching regret in adversarial multi-armed bandits, where the learner competes with an arm sequence that changes at most S times. When S is known, an optimal expected regret of \widetilde\mathcalO(\sqrt(S+1)KT) is obtainable [Auer et al., 2002]. However, when S is unknown, Marinov and Zimmert [2021] show that this guarantee is impossible under an adaptive adversary. In this paper, we show that a single algorithm achieves \widetilde\mathcalO(\sqrt(S+1)KT) expected regret for every S against an oblivious adversary, resolving an open problem of Auer et al. [2019b]. Our algorithm combines a fixed-share learner initialized with a small learning rate and dyadic-interval subroutines that search for local improvements using randomized learning rates and implicit exploration. Importantly, a non-uniform prior favors following the main learner, keeping the cost of maintaining many subroutines small. When the subroutines accumulate sufficient improvement over the main learner, its learning rate doubles, allowing adaptation to the unknown number of comparator switches S .
[LG-130] Runtime-Incremental Transformer for Reinforcement-Learning-Based Adaptive Control
链接: https://arxiv.org/abs/2609.13545
作者: Giansalvo Cirrincione,Adriano Fagiolini
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Learning-based adaptive control of robotic manipulators with non-observable friction memory has been addressed by attention- based meta-controllers whose number of attention heads is fixed before training and is tuned by costly offline search. At long memory horizons, such fixed-capacity controllers are prone to catastrophic failures on a sizeable fraction of training seeds. The present paper introduces a runtime mechanism that grows and prunes the heads of the attention block during reinforcement learning, governed by two signals: the effective rank of the on-policy context distribution, which triggers growth when representational capacity becomes insufficient, and the per-head output magnitude, which flags redundant heads for removal. Policy continuity at growth events and a quantitative bound at prune events are established analytically. On a two- link manipulator with Stribeck friction, the proposed mechanism attains full success across all memory regimes, eliminating the long-horizon failure mode and removing the need for offline tuning of the head count.
[LG-131] Early-Stopping Thresholds for ES-HyperNEAT: A Data-Driven Approach from Fitness Dynamics CEC
链接: https://arxiv.org/abs/2609.13533
作者: Romain Claret,Arthur Gygax,Michael O’Neill,Paul Cotofrei,Pascal Felber
类目: Neural and Evolutionary Computing (cs.NE); Machine Learning (cs.LG)
*备注: 8 pages, 5 figures, 9 tables. Accepted version of a paper published at the 2026 IEEE Congress on Evolutionary Computation (CEC), part of the IEEE World Congress on Computational Intelligence (WCCI 2026), Maastricht, Netherlands, 21-26 June 2026. © 2026 IEEE
Abstract:Most hyperparameter configurations for Evolvable-Substrate HyperNEAT (ES-HyperNEAT) produce networks that stagnate at random-guessing performance, wasting computational resources. We frame early stopping as binary classification on early fitness trajectories: for each trial, we compute the cumulative median of best-per-generation fitness and test it against a threshold derived by maximizing the F1 score on an initial 90-trial dataset. The resulting rule (generation G* = 3, threshold T* = 0.140) achieves F1 = 0.872 on 180 independent validation trials, retaining over 90% of successful trials while cutting computational cost by 41.6%. Compared to Hyperband, our domain-specific rule is 64% more efficient with higher mean fitness, though Hyperband occasionally discovers higher peak solutions. On a converged search population the rule becomes too aggressive (recall 31.1%), motivating adaptive thresholds. The specific thresholds are ES-HyperNEAT-specific, but the methodology, deriving stopping criteria from fitness dynamics classification, is applicable to other evolutionary algorithms with stagnation-prone hyperparameter spaces.
[LG-132] GeoTTER: Leverag ing Local Geometry of Optimal Transport for Zero-Shot Classification AISTATS2026
链接: https://arxiv.org/abs/2609.13518
作者: Wei-Yang Alex Lee,Rudrasis Chakraborty,Vishnu Lokhande
类目: Machine Learning (cs.LG)
*备注: Accepted at AISTATS 2026
Abstract:We present GeoTTER, a novel framework that redefines optimal transport in the realm of zero-shot classification. Conventional methods often suffer from miscalibration and a lack of adaptability, as they rely on fixed cost matrices derived solely from pre-trained model embeddings. In contrast, GeoTTER addresses these limitations by incorporating two key techniques. First, to alleviate high-frequency label jaggedness (sample-level manifold jitter that assigns neighboring embeddings to different classes), GeoTTER integrates local geometric structure into the optimal transport formulation via graph-Laplacian smoothing, a technique grounded in spectral graph theory that enforces neighborhood consistency. Second, to correct coherent angular drift (a low-frequency orientation bias in which large groups of samples share the same angular offset from their true label prototypes), we fuse clustering-guided cost components with a globally adjusted transport cost, achieving a multi-objective optimization that respects both global distribution constraints and latent data structure. With a median improvement of +6.82% compared to zero-shot and +2.13% compared to OTTER, GeoTTER shows robust improvements across a diverse set of benchmarks.
[LG-133] Operational Range Bounding in Spectroscopy: A Safety Cage Framework for Machine Learning Models
链接: https://arxiv.org/abs/2609.13514
作者: Nikki Grens,Luís F. Simões,Kai Hou Yip,Theresa Lueftinger
类目: Machine Learning (cs.LG); Earth and Planetary Astrophysics (astro-ph.EP); Instrumentation and Methods for Astrophysics (astro-ph.IM)
*备注:
Abstract:Ensuring the reliability of black-box machine learning models in safety-critical space missions remains a significant challenge, particularly when ground-truth is unavailable for validation. Although machine learning models offer a powerful means to augment standard pipelines by extracting transmission spectra from complex exoplanetary light curves, their susceptibility to unmodelled instrument anomalies, stellar activity, and domain shifts introduces unquantified risks. This study evaluates a modular safety cage architecture that operates as a parallel monitoring layer to assess the validity of a prediction without modifying the underlying estimator. By monitoring different runtime indicators, including uncertainty quantification, out-of-domain detection, and influence functions, the framework constrains the model’s operational domain to a verified region. A controlled evaluation is conducted under both in-domain and cross-domain conditions, using datasets from the 2019 and 2021 editions of the Ariel Data Challenges. The results reveal that model failure is multifaceted and that no single indicator captures all failure modes, demonstrating the need for indicator fusion. The application of safety-driven rejection strategies shows that a modest 20% reduction in data coverage results in error reductions between 45% and 65% across different domains and evaluation metrics. Using a formalised coverage-risk framework, a systematic analysis of indicator combinations is performed to identify configurations that maximise risk-ranking accuracy and optimise the trade-off between data coverage and scientific performance. Safety cages provide a transparent mechanism for detecting unreliable predictions and represent a critical step towards the safe deployment of data-driven models in scientific applications, such as astrophysics, where ground truth is seldom available.
[LG-134] Pretraining for Sample-Efficient Neural Interfaces
链接: https://arxiv.org/abs/2609.13507
作者: Ben Tang,Zachary Spalding,Gregory B. Cogan
类目: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
*备注: 15 pages, including references and appendix; 5 figures. Code: this https URL
Abstract:Brain-computer interfaces (BCIs) decode neural activity to restore lost function. Typically, training a high-performance neural decoder requires a large labeled dataset to be collected from every new subject. One way to reduce the labeled data cost is self-supervised pretraining, which learns general neural representations from unlabeled recordings that accumulate across subjects. However, for intracranial electroencephalography (iEEG) recordings, self-supervised learning has been challenging due to differences in contact placement and neuroanatomy between subjects. We propose MAPA, an otherwise vanilla masked autoencoder with two spatial encodings, an anatomical region embedding and a relative positional encoding, which together enable it to learn neural representations that transfer to unseen subjects and across various tasks. MAPA sets a new state of the art across all three regimes of the Neuroprobe benchmark without fine-tuning: within-session, cross-session, and cross-subject. In the cross-subject regime, a linear probe on MAPA’s features needs only \sim164 labeled trials to reach the accuracy that takes 3,500 without pretraining. Our results show that self-supervised pretraining can scale across heterogeneous iEEG recordings and reduce the labeled data needed for accurate decoding in new subjects.
[LG-135] On the Potential of Multi-Task Learning in Predictive Process Monitoring
链接: https://arxiv.org/abs/2609.13477
作者: Lukas Kirchdorfer,Keyvan Amiri Elyasi,Heiner Stuckenschmidt
类目: Machine Learning (cs.LG)
*备注: Accepted for publication in the Proceedings of the 60th Hawaii International Conference on System Sciences (HICSS-60)
Abstract:Predictive Process Monitoring (PPM) forecasts how ongoing organizational processes unfold, enabling information systems to move beyond execution support toward proactive analysis and monitoring. Although deep learning has improved prediction accuracy in PPM, most approaches follow a single-task learning (STL) setup, training a separate model per task. This increases maintenance effort and overlooks potential synergies. Multi-task learning (MTL), which jointly learns multiple prediction targets in one model, offers a promising alternative, yet its effectiveness in PPM remains underexplored. It remains unclear whether and under which settings MTL improves upon STL, which prediction tasks benefit most from joint learning, which task combinations are particularly synergistic, and if and how tasks should be balanced. To fill this gap, we present the first comprehensive empirical study of MTL for PPM, evaluating a variety of task combinations, neural architectures, and optimization methods. Overall, our results position MTL as a strong paradigm for PPM: we see substantial improvements in next-activity prediction and inherent mitigation of class imbalance using MTL, while task balancing is especially critical under low-capacity models.
[LG-136] A Machine Learning API for Earth Observation Data Cubes Based on openEO
链接: https://arxiv.org/abs/2609.13453
作者: Brian Pondi,Jonas Hurst,Rolf Simoes,Jonas Starke,Marius Appel,Edzer Pebesma
类目: Machine Learning (cs.LG)
*备注: 25 pages, 12 figures
Abstract:Earth Observation (EO) data are increasingly organized as spatio-temporal data cubes, while machine learning (ML) methods operate on tabular feature matrices or structured tensor inputs. This mismatch forces platform-specific transformations that are difficult to reproduce or transfer across cloud infrastructures. The openEO specification provides a unified interface for EO data access and processing across heterogeneous backends, but lacks a standardized approach for ML integration. We propose a process-level ML specification for openEO structured into three stages: model initialization, model actions (training, tuning, inference, validation), and model management. It supports classical algorithms such as Random Forest and SVM, as well as deep learning architectures for time-series and spatial patch-based modeling, including TempCNN, Temporal Attention Encoders, and foundation model inference. Three prototype implementations in R and Python demonstrate feasibility across diverse technology stacks. A crop type mapping use case demonstrates cross-backend interoperability by submitting an identical process graph to independent R and Python backends and comparing predictions and evaluation metrics. Two further use cases demonstrate deep learning on time series and foundation model inference, each executed on a dedicated backend. The prototypes reveal, however, that full cross-backend portability requires deeper harmonization of serialization formats and execution semantics than the process level alone can enforce; backend library versions and preprocessing conventions outside the specification’s boundary also affect reproducibility. Addressing both through explicit backend conformance profiles represents the most important near-term direction. The specification advances the reproducibility, portability, and accessibility of ML workflows on EO data cubes across cloud platforms. Comments: 25 pages, 12 figures Subjects: Machine Learning (cs.LG) ACMclasses: I.1.2 Cite as: arXiv:2609.13453 [cs.LG] (or arXiv:2609.13453v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.13453 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-137] Efficient Online Inverse Optimization with O(d) Regret
链接: https://arxiv.org/abs/2609.13440
作者: Yang Cai,Anupam Gupta,Vineet Gupta,Guru Guruganesh,Yanchen Jiang,Christopher Liaw,Aranyak Mehta,Renato Paes Leme,Grigoris Velegkas,Di Wang
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS)
*备注:
Abstract:We give a deterministic algorithm for online inverse linear optimization with regret O(d) , uniform in the horizon and O(d^2) time per round. A bound of this order was obtained recently by Dewasurendra, settling a question of Gollapudi et al.\ and of Oki and Sakaue, but by an improper rule that enumerates covers at every scale and costs T^\Theta(d) a round; ours is the first efficient such bound and the first proper one. We build on the variable-metric framework of Sakaue et al., adding a self-normalized rank-one update, and we replace the \log\det potential by the trace power \tr(H^-1/2) , which is bounded outright and removes the \ln T . The bound also holds against an expert that does not optimize, and we give corruption-robust and rank-adaptive variants, and an application to convex minimization.
[LG-138] Specification Oracles
链接: https://arxiv.org/abs/2609.13415
作者: Atticus Cull,Justin McCarthy
类目: Machine Learning (cs.LG)
*备注: 13 pages, 4 figures
Abstract:Specifications face a basic tradeoff: leave details out, and important questions go unanswered; record every detail separately, and the specification becomes large and prolix. We investigate whether a language model can serve as a compact, living specification oracle by learning facts about a target and answering questions about it directly. We compare two ways of storing the learned facts: external text notes and changes to the model’s weights. Across four families of 596-fact worlds and two Qwen2.5 model sizes, weight-only oracles benefited substantially more from structure: with the 7B model, their accuracy integrated across storage capacities was 18.5 percentage points higher on structured than unstructured worlds, compared with 1.1 points for note-sheet oracles. This advantage came at a substantial storage cost, with the smallest adapter requiring approximately 175 KiB compared with a maximum note budget of 16 KiB. Adapted weights therefore exploited latent structure more successfully, while external notes required substantially less object-specific storage.
[LG-139] ViFA-Council: Multi-Agent LLM Deliberation for Vietnamese Folk Art Generation
链接: https://arxiv.org/abs/2609.13348
作者: Hai-Dang Nguyen,Minh-Phuong Pham,Thao Thi Phuong Dao,Trong-Le Do,Vinh-Tiep Nguyen,Trung-Nghia Le
类目: Graphics (cs.GR); Machine Learning (cs.LG)
*备注: MAPR 2026
Abstract:This paper presents ViFA-Council, a three-stage multi-agent framework that employs multiple large language models (LLMs) to tackle two culturally complex generative tasks: image outpainting and educational story generation based on traditional Vietnamese folk paintings. Current single-model generative pipelines frequently struggle with stylistic hallucinations and cultural misrepresentations because they lack mechanisms for cross-model critique. ViFA-Council addresses this challenge by orchestrating collaboration among GPT-4o, Gemini 3.1 Pro, and Claude Sonnet 4.6. It enforces rigorous cultural constraints through structured agent deliberation. This deliberation is mediated by task-specific JSON schemas that effectively bridge natural language discussions with diffusion-based image synthesis using Banana Pro. Experiments and a user study demonstrate that structured multi-agent deliberation is a promising direction for improving cultural fidelity and narrative coherence in culturally sensitive, low-resource artistic domains. The source code and data are released at this https URL.
[LG-140] Ergodic Control and Controlled Diffusion for Robot Learning: Review and Tutorial
链接: https://arxiv.org/abs/2609.13295
作者: Max Muchen Sun,Cem Bilaloglu,Ananya Rao,Stefan Ivic,Guillaume Sartoretti,Kathleen Fitzsimons,Ian Abraham,Sylvain Calinon,Todd Murphey
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: Under review at Foundations and Trends in Robotics
Abstract:Diffusion learning leverages the statistical mechanism of diffusion processes for learning, reasoning, and inferring complex distributions from data. Recent advances in diffusion learning have been transformative, with robot learning emerging as a key opportunity area, with applications spanning perception, control, and decision-making. At the same time, the statistical mechanism of diffusion processes can be controlled to shape the temporal evolution of the state distribution underlying robot trajectories, inducing ergodic behavior in robotic systems. The frameworks of controlled diffusion and ergodic control were developed around the same time as diffusion learning, and their theories and algorithms have increasingly converged. Ergodicity induced by controlled diffusion has several significant implications for robot learning, distinct from applying diffusion learning to robotics problems: it formally enforces statistical properties required to ensure optimality of robot learning, enables non-myopic search over uncertain information landscapes for data collection, and enables behavior specification based on spatial rather than temporal characteristics of trajectories. This survey introduces the intuition behind controlled diffusion for robot learning, explores its connection to diffusion learning, presents theoretical foundations and numerical tutorials for solving controlled diffusion and ergodic control problems, and reviews applications across robotics. Finally, we discuss key challenges and future opportunities in leveraging controlled diffusion for robot learning.
[LG-141] Grouped Value Attention: Efficient KV Caching via On-Demand Key Reconstruction
链接: https://arxiv.org/abs/2609.13285
作者: Vishesh Tripathi,Abhay Kumar,Ramsha Khan
类目: Hardware Architecture (cs.AR); Machine Learning (cs.LG)
*备注:
Abstract:The KV cache is a primary bottleneck for Transformer decoding: its memory footprint and cache-read traffic grow with sequence length. Grouped-query attention (GQA) reduces this cost by sharing key-value heads, but still stores both a key and a value at every step. We introduce Grouped Value Attention (GVA), which stores grouped values and reconstructs content keys with a learned linear map. At inference, the map can be absorbed into the query, eliminating the need to materialize content keys in the intended decode path. A small shared decoupled RoPE channel retains positional information through a separately cached positional key. For the configurations studied, this representation reduces persistent cache scalars by approximately 45-47% relative to matched GQA. At the 350M-parameter scale with 30B FineWeb-Edu tokens, the 16-dimensional positional variant reaches 44.18 average accuracy across five tasks, compared with 44.36 for GQA and 43.88 for MLA. These results demonstrate near-GQA benchmark accuracy with a more compact cache representation. To translate this compact representation into faster autoregressive inference, we have developed custom decoding kernels and are currently evaluating their end-to-end inference performance with an open-source release planned soon.
[LG-142] From Masking to Merging: Rethinking SpecAugment for Efficient Audio Spectrogram Transformer INTERSPEECH2026
链接: https://arxiv.org/abs/2609.13260
作者: Minhee Park,Hyowon Ahn,Chanwoo Kim
类目: ound (cs.SD); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
*备注: Accepted to Interspeech 2026
Abstract:This paper proposes SpecAugment-Patch Merging, a simple yet effective method to accelerate Audio Spectrogram Transformer (AST) training. We first apply SpecAugment to mask input spectrograms at the patch level, and after positional embeddings are added, the method selects r pairs of masked patches and merges them, reducing the number of tokens processed by the Transformer. Increasing the number of merged pairs r from 0 to 100 keeps mAP on AudioSet nearly unchanged (34.07 to 34.08) while throughput increases from 43.3 to 49.3 samples/sec, which is a relatively 13.9% improvement. Similar patterns appear on ESC-50 and Speech Commands V2, where throughput steadily improves with only minor accuracy changes, demonstrating that this merging approach provides faster training with minimal performance loss.
[LG-143] GzDRL: Reproducible and Scalable Deep Reinforcement Learning with Gazebo
链接: https://arxiv.org/abs/2609.13243
作者: Amal Dev Haridevan,Junjie Kang,Jinjun Shan
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
Abstract:We present GzDRL, a novel single-process reinforcement learning (RL) framework for Gazebo that overcomes longstanding bottlenecks in scalable, reproducible robotics experimentation. Unlike conventional middleware-based RL-Gazebo integrations that suffer from nondeterminism and irreproducibility, GzDRL introduces a systematic, middleware-free environment-stepping mechanism that directly synchronizes agent actions and physics updates. This design enables deterministic, high-throughput data collection, efficient vectorization, and reproducible RL training and evaluation. Comprehensive benchmarks demonstrate that GzDRL achieves the highest workstation throughput among the evaluated frameworks while remaining competitive with GPU-accelerated simulators on laptop hardware, and maintains precise agent-environment synchronization, multi-agent scalability, and experiment-level reproducibility. We further validate sim-to-real transfer by deploying learned policies directly onto a physical quadrotor, without fine-tuning. Our results establish GzDRL as an accessible and reproducible platform for advancing RL in robotics and automation.
[LG-144] Scalable partial information decomposition for symptom networks via supervised embeddings
链接: https://arxiv.org/abs/2609.13203
作者: Cillian Hourican,Eric Dignum,Rick Quax,Debraj Roy
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注:
Abstract:Pairwise relationships among mental-health symptoms are routinely summarised asscalar edge weights, which cannot express whether two symptoms carry overlapping information about a third or information that appears only in combination. Partial information decomposition (PID) addresses this gap but is computationally intractable beyond a few sources. We introduce embedding-based PID (ePID), a scalable pipeline that compresses all nonfocal symptoms into a low-cardinality discrete embedding and computes a tractable two-source PID, yielding source-unique, remainder-unique, redundant, and synergistic components for each ordered source-target pair. We benchmarked 13 candidate embeddings on synthetic Bayesian networks calibrated to PHQ-9 and on 83 real-world datasets across five PID measures. A supervised Agglomerative Conditional Information Bottleneck (ACIB) embedding recovered the reference decomposition most accurately of the 13 embeddings tested, and did so for every PID measure yielding non-negative atoms once four or more symptoms were compressed (synergy recovery r = 0.92). The two instruments then diverged sharply. In PHQ-9 networks (UK Biobank,N = 154,291; Xinxiang student sample, N = 24,292) the surrounding symptom context carried most pairwise dependence through redundant and remainder-unique channels; synergy contributed 6 to 9%, and no directed edge was synergy-dominated in either cohort. In the 28-item Interpersonal Reactivity Index, 45% of source pairs were. The identical pipeline, applied without parameter changes, therefore returned opposite profiles for the two instruments, each consistent with how that instrument was constructed. By separating overlapping from interaction-dependent information, ePID provides a scalable, model-agnostic complement to standard symptom-network methodology to distinguish redundant and synergistic contributions to observed correlations. Subjects: Machine Learning (cs.LG); Information Theory (cs.IT) Cite as: arXiv:2609.13203 [cs.LG] (or arXiv:2609.13203v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.13203 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Cillian Hourican [view email] [v1] Fri, 14 Aug 2026 17:34:47 UTC (1,742 KB)
[LG-145] Do Tabular Foundation Models Still Need Feature Engineering?
链接: https://arxiv.org/abs/2609.13202
作者: Yifan WU,Pinjun Dong,Jiran Tao,Binyan Jiang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Feature engineering has long been a cornerstone of tabular machine learning. Tabular foundation models (TFMs) are pretrained on a wide range of tabular datasets and applied via in-context learning. Their rise raises a natural question: does manual feature construction still matter as these models become more capable? To answer this, we perform a controlled study across several versions of two major TFM families, testing a wide range of existing feature engineering techniques on benchmark datasets from TabArena. We find a consistent pattern: feature engineering gains are concentrated in earlier model generations and become negligible for the strongest models. These results suggest that stronger TFMs depend less on explicitly engineered input representations. In a complementary experiment, however, adding in-context information from related datasets still improves performance. Our findings indicate a shift in the source of performance gains for stronger TFMs: re-representing existing inputs becomes less effective, while providing additional task-relevant context remains beneficial.
[LG-146] Criticality in Dissimilar Decomposition and Undersampling of Random Datasets with Anomalies
链接: https://arxiv.org/abs/2609.13201
作者: Ghurumuruhan Ganesan
类目: Machine Learning (cs.LG); Information Theory (cs.IT); Probability (math.PR)
*备注:
Abstract:Training datasets for upcoming LLMs would include a significant amount of AI text/image data generated from current LLMs. In such a scenario, it is important to understand how this affects batch decompositions and thereby, the performance of the resultant new LLM. In this paper, we consider AI generated data as anomalies linked" to main data points and study decomposition and undersampling properties of the overall random dataset. We use redundancy graphs and iteration techniques to obtain bounds for the minimum size of a strongly dissimilar (SD) decomposition and demonstrate a phase transition phenomena, wherein the minimum size is essentially determined by the \emphmain data points when the number of anomalies is small and is taken" over by the anomalies above a certain threshold. We also establish a size criticality result for the strong similarity of a randomly undersampled dataset and illustrate our results with examples involving categorical datasets, whose overall space size is much larger than the size of the dataset.
[LG-147] PRISM-UDE: Physics-Regularized Iterative Symbolic Modeling of 3nm FinFETs via Universal Differential Equation
链接: https://arxiv.org/abs/2609.13200
作者: Pranavanath Balamurali,Prathamesh Dinesh Joshi,Raj Abhijit Dandekar,Rajat Dandekar,Sreedath Panat
类目: Emerging Technologies (cs.ET); Machine Learning (cs.LG)
*备注:
Abstract:Compact transistor models are the mathematical backbone of circuit simulation. However, at advanced nodes such as 3nm, transport physics becomes too complex for traditional hand-derived equations to capture accurately. Purely data-driven neural surrogates, on the other hand, are numerically unstable inside circuit solvers and offer no physical insight into their own predictions. We introduce PRISM-UDE (Physics-Regularized Iterative Symbolic Modeling via Universal Differential Equations), a framework that embeds a small neural network inside a physics-based transistor model, using the network only to learn the transport behavior that the analytical baseline misses, rather than replacing the physics altogether. Once trained, this neural correction is distilled into a single, interpretable closed-form expression via symbolic regression, making the final model fully analytical and simulator-ready. Applied to a 3nm FinFET benchmark dataset, PRISM-UDE reduces prediction error more than sixfold (70.33% to 11.01%) relative to the standard physics-only baseline. The distilled expression preserves this accuracy almost exactly while eliminating the neural network entirely. We further validate the extracted expression directly inside a SPICE circuit simulator, confirming stable, physically consistent behavior under both static bias sweeps and dynamic switching conditions.
[LG-148] Algorithmic Information Dynamics of Learning: A Certified Differentiable Complexity Controller for Grokking
链接: https://arxiv.org/abs/2609.13197
作者: Luan Ozelim,Hector Zenil
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注:
Abstract:Algorithmic Information Dynamics (AID) studies systems by perturbing them and measuring changes in algorithmic complexity, but its usual estimator, the Block Decomposition Method, is piecewise constant, restricting the calculus to finite differences. We use K^\mathrmCDM_\mathrmsF , a certified, differentiable estimator, to bring the calculus into learning dynamics: grokking, where a complexity order parameter is known but has not been made to act. A\empts a transient loss kick, the estimator becomes a controller that accelerates grokking in Levin’s description-length–versus-time sense, within a data-dependent Occam boundary whose finite-size trend, f_c\sim\ln p/p , is consistent with a coupon-collector interpretation. Ablations show that a complexity gate matches a train-loss gate in rescuing failing seeds with 27% less intervention; among the tested signals, only map complexity marks the transition’s completion; the certified prior and the per-parameter \nabla K attribution are both fungible (a uniform-prior sensor makes bit-identical gate decisions, and random supports match \nabla K -selected ones above a sparsity threshold); and direct field perturbation shows a nucleation-like response to the Occam field (no linear regime is resolved over the probed amplitudes, so these measurements do not justify a fluctuation–dissipation surrogate), with a finite-field response growing by orders of magnitude toward the transition. These measurements account for the empirically tuned staircase: bang–bang pulses, stall-fired and released on yield, whose iteration plausibly builds the response it exploits. The kick transfers to sparse parity and to a transformer; a sustained weight-space loss fails. The algorithmic estimator’s distinct contribution is timing (when to fire and when to release), not attribution.
[LG-149] Evaluating LLM -Generated Rules for Heart Disease Prediction
链接: https://arxiv.org/abs/2609.13192
作者: Feisal Alaswad,Batoul Aljaddouh,Maher Alrahhal,Wafaa Al Nassan,Talal Bonn
类目: Machine Learning (cs.LG)
*备注:
Abstract:This study compares traditional machine learning models and Large Language Model (LLM)-generated rule-based systems for heart disease prediction using the UCI Heart Disease dataset. Several classifiers, including Logistic Regression, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Naive Bayes, Decision Tree, and Random Forest, were evaluated alongside rule-based systems generated using GPT-4o and Claude Sonnet 4.6. Model performance was assessed using accuracy, precision, recall, and F1-score metrics. Experimental results show that traditional machine learning models consistently outperform LLM-generated rule-based systems in predictive performance. Random Forest achieved the best overall performance with 90.2% accuracy, a precision of 0.829, perfect recall of 1.0, and an F1-score of 0.906. Naive Bayes followed closely with 88.5% accuracy and an F1-score of 0.881. In contrast, the LLM-generated rule models achieved lower performance, with Claude Sonnet 4.6 reaching 80.3% accuracy (F1-score: 0.833) and GPT-4o obtaining 70.5% accuracy (F1-score: 0.690). Despite the performance gap, the LLM-generated rules provide interpretable IF-THEN diagnostic logic that enhances explainability and transparency in clinical decision-making. These findings highlight the trade-off between predictive performance and interpretability in medical artificial intelligence systems. The complete implementation of all experiments, including machine learning models and LLM-derived rule classifiers, is publicly available in the GitHub repository at this https URL .
[LG-150] Early Prediction of Satellite Collision Probability Using a Hybrid TCN-Transformer Model for a CDM-Based Conjunction Analysis Framework
链接: https://arxiv.org/abs/2609.13191
作者: Rabia Tüylek Tok,Burak Yağlıoğlu,Enes Dağ,Emre Onur Kahya
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注: Accepted in 2026 AAS/AIAA Astrodynamics Specialist Conference
Abstract:The rapid expansion of operational satellites and orbital debris has increased the frequency of close approach events in low Earth orbit (LEO), creating a higher operational burden for satellite operators. This problem is especially critical for satellites using electric propulsion, where low-thrust maneuver capability imposes additional time constraints on collision avoidance planning. In current practice, Conjunction Data Messages (CDMs) provide relative state, covariance, miss distance, time of closest approach, and probability of collision (PoC) information for conjunction assessment. However, the nonlinear propagation of orbital uncertainties and the sensitivity of PoC to covariance evolution make the interpretation of sequential CDMs challenging. This study proposes a learning-based framework for early prediction of satellite conjunction risk by estimating the PoC expected in the subsequent CDM update of the same close approach event. In the proposed methodology, an Unscented Transform-based propagation and backpropagation framework is first used to evaluate the sensitivity of the collision risk metric to CDM parameters. In addition, Principal Component Analysis is applied to the numerical CDM parameters to identify the features most relevant to PoC variation. The results obtained from the sensitivity analysis and PCA are then used to justify the selected raw CDM parameters and to construct derived metrics representing relative motion, encounter geometry, and covariance-related uncertainty. Using the resulting sequential enriched conjunction dataset, a hybrid Temporal Convolutional Network (TCN)-Transformer model is trained to learn the temporal evolution of conjunction risk. The framework is applied to CDMs received and analyzed within TÜBİTAK UZAY, demonstrating its potential for earlier and more consistent operational risk evaluation for LEO satellite conjunctions.
[LG-151] A derivative-fidelity failure mode in physics-informed neural networks: strengthened benchmark evidence from function-value training
链接: https://arxiv.org/abs/2609.13171
作者: Koji Koyamada
类目: Machine Learning (cs.LG)
*备注:
Abstract:Physics-informed neural networks (PINNs) use automatic differentiation to impose differential-equation residuals, but good agreement in function values does not necessarily imply accurate derivatives. This paper formulates derivative fidelity as a failure mode of PINNs and tests it with one-dimensional benchmarks. Multilayer perceptrons are trained only on function values for sin(x) and exp(x), while second derivatives obtained by automatic differentiation are evaluated separately. The hypothesis is strengthened by additional tests over training-point density, activation functions, endpoint-dense evaluation, and both L2 and maximum-error diagnostics. The results show that visually accurate function approximation can coexist with substantially larger second-derivative errors, especially near high-curvature boundary regions. The experiment provides a diagnostic protocol for distinguishing value accuracy from physics-residual reliability.
[LG-152] A Lifecycle Cost Analysis of Smart-Contract-Coordinated Federated Learning Marketplaces
链接: https://arxiv.org/abs/2609.13170
作者: Luan Mantegazine,Luiza Leidemer,Claudio Geyer
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: 8 pages, 5 figures, IEEE Blockchain Conference 2026
Abstract:Blockchain-enabled Federated Learning (FL) marketplaces enable collaborative model training among mutually distrustful participants through smart contracts. Although numerous architectures exist, their economic evaluation is typically limited to isolated blockchain operations rather than the complete marketplace lifecycle. Consequently, it remains unclear whether operational costs depend on operating at scale. This paper presents an experimental study of the operational cost of a DAO-governed marketplace. Our evaluation decomposes the gas consumption of every blockchain operation throughout the contract lifecycle, performs ablation experiments to isolate the impact of on-chain coordination and IPFS storage on federated training, and derives an analytical model describing the amortization of deployment costs. Results show the lifecycle of a single training task consumes approximately 3.8 million gas units per hired trainer. The average cost per training round reaches its amortization knee - defined as twice the asymptotic recurring cost - after approximately 20 communication rounds. Moreover, integrating smart contracts and IPFS preserves model performance, achieving accuracy comparable to conventional FL deployments. These findings demonstrate that smart-contract-coordinated FL marketplaces exhibit an amortizing cost structure not because on-chain operations are inexpensive, but because recurring costs are one to two orders of magnitude smaller than fixed deployment costs, thereby diluting over the federation’s lifetime.
[LG-153] Bridging Control Inference Transport and Thermodynamics: From Theory to Applications in Learning
链接: https://arxiv.org/abs/2609.15897
作者: Emmy Blumenthal,Nikolas Claussen,Benjamin Eysenbach,Catherine Ji,Gautam Reddy,Colin Scheibner,Benjamin Sorkin
类目: atistical Mechanics (cond-mat.stat-mech); Disordered Systems and Neural Networks (cond-mat.dis-nn); Soft Condensed Matter (cond-mat.soft); Machine Learning (cs.LG); Biological Physics (physics.bio-ph)
*备注: 5 Chapters, 62 pages, 16 figures
Abstract:The last decade has seen the development of powerful methods for learning complex structure from high-dimensional data. These advances have brought to the foreground fundamental connections between subdisciplines of physics, applied mathematics, and machine learning. In this review, we bring together some of these ideas, often expressed in different languages, to highlight a conceptual thread that links five distinct fields: control theory, optimal transport, probabilistic inference, non-equilibrium thermodynamics, and machine learning. A common theme is the optimization of free-energy-like functionals under dynamical or statistical constraints. We offer a guided tour through this thread and present selected applications in reinforcement learning, variational inference, and generative modeling. The review does not assume prior familiarity with these topics, and begins with principles originating from physics.
[LG-154] Quenched Ensemble Sampling
链接: https://arxiv.org/abs/2609.15894
作者: David Yallup
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO)
*备注: 27 pages, 8 figures
Abstract:Some of the sharpest challenges in sampling from the energy functions of physical systems arise at phase transitions, where the density of states changes abruptly and many sampling algorithms stall. Nested sampling is a particle method that traverses the density of states under a hard energy constraint and is known to be robust to such transitions, but its application in high dimension is limited by the difficulty of sampling under that constraint. In this work we introduce Quenched Ensemble Sampling, which generalises the hard constraint to a family of repulsive potentials at the energy boundary. This preserves the quenched path of monotonically decreasing energy while making the constrained target amenable to scalable gradient-based kernels. We demonstrate on synthetic models of phase transitions that our method estimates the marginal likelihood and draws posterior samples across a first-order transition where popular alternatives such as tempering fail. We apply the procedure to marginal likelihood estimation in Bayesian neural networks, enabling model comparison between network architectures. Finally, in a high-dimensional continuous lattice field theory, we show that this method traverses a first-order transition and estimates the partition function.
[LG-155] Learning under Target Shift: Optimal Density Ratio Estimation and Importance-Weighted Regression
链接: https://arxiv.org/abs/2609.15785
作者: Ren-Rui Liu,Zheng-Chu Guo
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:We study density ratio estimation and importance-weighted regression under target shift with continuous outputs. Under target shift, the conditional distribution of the inputs given the outputs remains invariant across the training and test distributions, while the output marginal distribution may change. Although this problem has been extensively studied for discrete outputs, the continuous setting is substantially less understood: the importance weights are determined by an unknown density ratio function, for which existing estimation methods lack explicit finite-sample convergence rates. We propose a spectral regularization method in a reproducing kernel Hilbert space (RKHS) for estimating the continuous density ratio from labeled training samples and unlabeled test inputs. Under a source condition with regularity parameter \iota0 , we establish high-probability finite-sample guarantees and show that the estimator achieves the capacity-independent minimax-optimal RKHS-norm rate O(n_\eta^-\iota/(2\iota+2)) . We then incorporate the estimated density ratio into importance-weighted regression and characterize the propagation of density-ratio estimation error to the final predictor. When sufficiently many samples are available for density ratio estimation, the resulting regression estimator attains the minimax-optimal rates of standard kernel regression. These results establish a finite-sample theory for continuous density ratio estimation and importance-weighted learning under target shift.
[LG-156] Projection-Free Multi-level Algorithms for Stochastic Constrained Compositional Optimization
链接: https://arxiv.org/abs/2609.15679
作者: Wei Jiang,Sifan Yang,Wenhao Yang,Yibo Wang,Yuanyu Wan,Zechao Li,Lijun Zhang
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:This paper studies projection-free algorithms for stochastic constrained multi-level compositional optimization. In this context, the objective function is a nested composition of several smooth functions, and the decision set is closed and convex. Since projection onto the constraint set can be computationally expensive, we develop projection-free methods that rely on linear minimization oracles. For non-convex objectives, we propose variance-reduced projection-free algorithms and establish complexity guarantees under both the Frank-Wolfe gap and the gradient mapping criteria. We also develop momentum-based methods that achieve convergence guarantees under weaker smoothness assumptions. Additionally, by using a stage-wise design, we derive a parameter-free variant that preserves the same complexities for the Frank-Wolfe gap. Such a design can be further used to develop algorithms for convex and strongly convex functions whose rates match those of single-level projection-free counterparts. Finally, we consider finite-sum problems and derive complexities for non-convex, convex, and strongly convex objectives. Numerical experiments across multiple tasks demonstrate the effectiveness of the proposed methods.
[LG-157] Online local learning for generative thermodynamic computing
链接: https://arxiv.org/abs/2609.15439
作者: Huilin Wang,Weibing Deng
类目: Data Analysis, Statistics and Probability (physics.data-an); Machine Learning (cs.LG)
*备注:
Abstract:Generative thermodynamic computers turn thermal noise into structured data through Langevin dynamics. We train these systems with a local update at each integration step. The reverse-path Onsager-Machlup objective yields a coupling gradient that is a symmetric sum of local residual-state correlations. We apply this gradient immediately rather than accumulating it over a full trajectory. In digital simulations using MNIST prototypes, online and trajectory-batch training reach similar validation losses on fixed noising paths. Models trained online release less heat on average in all five independently seeded pairs, with both models’ parameters held fixed during sampling. Auxiliary classifier and nearest-prototype measures change modestly, while pairwise diversity decreases. The response to noise depends strongly on where the errors enter: independent zero-mean errors in the formed updates produce little heat change over a finite range of noise amplitudes, whereas residual offset and temporal correlation have much larger effects. Storing trained couplings requires substantially less precision than resolving deterministic updates during training. Together, these results establish a local online training method and show how update timing, noise structure, and precision affect generative thermodynamic computing.
[LG-158] Graph Matching Relaxations and Amortization for Supervised Graph Prediction
链接: https://arxiv.org/abs/2609.15437
作者: Federico Méndez,Paul Krzakala,Gabriel Melo,Charlotte Laclau,Rémi Flamary,Florence d’Alché-Buc
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:End-to-end Supervised Graph Prediction (SGP) requires a permutation-invariant loss to compare predicted and target graphs with arbitrary node orderings. Such losses typically involve a costly graph-matching problem. We first study three Optimal Transport relaxations of this problem and show, theoretically and empirically, that the Gromov-Wasserstein (GW) objective is the most suitable for SGP. Then, to avoid solving the resulting inner optimization for every training example, we propose to amortize the graph matching (node alignment) problem. For each training sample, the loss function leverages a transport plan provided by a parametric matcher based on the differentiable Sinkhorn algorithm applied on empirical node distributions. The graph prediction module and the matcher are jointly learned. We showcase the efficiency of this approach on toy and real world SGP problems of increasing complexity including a novel Mass-spectra to Scaffold task that we introduce.
[LG-159] ReLU Neural Network Approximation to Smooth Functional Operator: Dimensional Decay and Error Analysis
链接: https://arxiv.org/abs/2609.15355
作者: Shuhao Jiao
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:We study the uniform approximation of smooth scalar-valued functionals on an infinite-dimensional separable Hilbert space by deep ReLU neural networks. Writing the functional input as X(t)=\sum_d\geq1\xi_d\nu_d(t) , we quantify the importance of coordinate d through w_ds_d , where s_d bounds the magnitude of the corresponding basis score and w_d controls the directional Fréchet sensitivity of the target functional. Our constructive analysis combines coordinate truncation, anisotropic partitioning, local Taylor approximation, and ReLU network realization, while allowing unrestricted interactions among the retained coordinates. We establish a general nonasymptotic upper bound for the uniform approximation error and a complementary pseudo-dimension-based lower bound for the worst-case approximation error. Under generalized exponential coordinate decay w_ds_d\asymp\exp(-cd^\rho) , with \rho0 , the upper and lower bounds match at the leading order and thus yield the nearly optimal approximation rate, which is stretched-exponential in the logarithm of the network budget. This is the first work to characterize neural network approximation error for infinite-dimensional functional inputs explicitly through the joint dimensional decay of coordinate magnitudes and directional sensitivities.
[LG-160] Improving the Last-Iterate Guarantees of Anytime Algorithms for Stochastic Monotone Variational Inequalities
链接: https://arxiv.org/abs/2609.15257
作者: Jun-Hyun Kim,Ahmet Alacaoglu
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:We analyze a stochastic algorithm with Halpern anchoring for constrained convex-concave problems and monotone variational inequalities. This algorithm is single-loop and single-call since it uses one unbiased sample of the gradient operator at every iteration to be applicable to monotone games with noisy feedback. With t denoting the iteration counter, we prove the anytime last-iterate convergence rate of O(t^-1/4) for both gradient-mapping norm and restricted gap, improving the best-known rate O(t^-1/5) that was obtained for the restricted gap function. Our rates cover constrained problems with a potentially unbounded feasible set as well as a structured class of stochastic oracles without a bounded variance.
[LG-161] Conformal Individual Treatment Effect Estimation under Networked Interference
链接: https://arxiv.org/abs/2609.15254
作者: Matteo Zecchin,Osvaldo Simeone
类目: Machine Learning (stat.ML); Information Theory (cs.IT); Machine Learning (cs.LG)
*备注:
Abstract:Conformal counterfactual prediction constructs prediction sets with finite-sample coverage guarantees for counterfactual outcomes and individual treatment effects under the no-interference assumption. In this work, we relax this assumption by allowing each unit’s potential outcomes to depend on other units’ treatments and covariates. In this setting, propensity-score reweighting does not restore weighted exchangeability, and existing methods may fail to achieve valid coverage. To address this issue, we develop interference-adjusted weighted conformal prediction that accounts for interference by constructing an observable upper bound on the ideal and unobserved conformal p -value under the target intervention. The resulting prediction sets provide finite-sample marginal coverage guarantees for counterfactual outcomes and individual treatment effects in both transductive and inductive settings. We also derive a sharper construction when intervention-induced changes in nonconformity scores are bounded. Numerical experiments show that our methods preserve nominal coverage, whereas existing methods may not.
[LG-162] Eigenvalue-Decomposition Cost Denoising as an Alternative to Predict-then-Optimize for Shortest-Path Problems
链接: https://arxiv.org/abs/2609.15111
作者: Henry Aldridge-Krawciw,Irene Aldridge
类目: Econometrics (econ.EM); Machine Learning (cs.LG); Optimization and Control (math.OC); Spectral Theory (math.SP)
*备注: 7 pages
Abstract:Predict-then-optimize methods such as Smart “Predict, then Optimize” (SPO+) of Elmachtoub and Grigas (2022) learn a mapping from contextual features to unknown edge costs and then solve the induced combinatorial problem on the predicted costs. This approach is powerful but relies on the predictive model being well specified: when the true cost-generating process is nonlinear in the features and the predictor is linear, SPO+'s performance degrades as the misspecification grows. We propose and evaluate a structurally different remedy for a specific but common setting: when the decision-maker observes many noisy realizations of the same underlying cost process, the realized cost vectors themselves can be treated as a noisy signal and denoised directly, via eigenvalue decomposition (equivalently, Principal Component Analysis) of their covariance matrix, before ever invoking a predictive model. We instantiate this idea on the 5\times5 grid shortest-path benchmark introduced by Elmachtoub and Grigas (2022), retaining only the top- k eigenvectors of the training cost covariance matrix and projecting new noisy cost observations onto that subspace prior to solving with Dijkstra’s (1959) algorithm. We find that the choice of k is decisive: keeping only k=2 eigenvectors discards real signal and underperforms even the naive noisy-cost baseline, while setting k=5 to match the true latent feature dimension makes eigenvalue-denoised Dijkstra the best-performing method at every misspecification level tested, outperforming SPO+ by a wide margin under high misspecification.
[LG-163] Bridging the Gap in ECG-Based Emotion Recognition: A Unified Evaluation of Deep Learning Models
链接: https://arxiv.org/abs/2609.15055
作者: Timothy C Sweeney-Fanelli,Ajan Ahmed,Masudul Imtiaz
类目: ignal Processing (eess.SP); Machine Learning (cs.LG)
*备注: Accepted at 2026 IEEE 17th Annual Ubiquitous Computing, Electronics Mobile Communication Conference (UEMCON) - 2026 IEEE UEMCON
Abstract:Deep learning has led to numerous proposed architectures for Automated Emotion Recognition (AER) from electrocardiogram (ECG) data, but inconsistencies in preprocessing, training, and evaluation make direct comparisons difficult. Most studies train and validate models on individual datasets collected under homogeneous conditions, limiting variability and raising concerns about generalizability. Cross-dataset validation is sometimes used but primarily assesses model adaptability rather than true generalization. This study presents a comparative analysis of prominent deep learning architectures in AER, emphasizing model generalization over dataset adaptability. To enable this benchmark, we introduce two open-source frameworks: Affective Research on Representations and Classifications (ARRC), a standardized benchmarking toolkit, and Affective Research Dataset Toolkit (ARDT), a framework for inter-dataset training and validation. Using ARDT, we consolidate three publicly available AER datasets, CUADS, ASCERTAIN, and DREAMER, into a single dataset, increasing variability in sensor types, recording conditions, and participant demographics. We then use ARRC to evaluate three widely studied deep learning models and two CNN baselines through hyperparameter optimization and 10-fold cross-validation. Our findings provide insights into the trade-offs between classification accuracy and model complexity, establishing a reproducible benchmark for AER research. All source code for ARRC, ARDT, and model evaluation is publicly available to ensure transparency and facilitate further research.
[LG-164] Data Attribution at Scale via Influence Matrix Estimation
链接: https://arxiv.org/abs/2609.15044
作者: Yuxi Chen,Hamza Golubovic,Han Tong,Arian Maleki,Andrew Ilyas
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Data attribution seeks to quantify how individual training examples shape a model’s predictions and underpins problems including data valuation, machine unlearning, and model interpretability. Despite having a long line of work, computationally scalable methods often struggle to predict the effect of removing training data in neural networks due to their non-convex nature. To overcome this challenge, metagradient-based methods such as MAGIC (Ilyas and Engstrom, 2025) differentiate each prediction through the entire training run and compute its exact influence with respect to the training data, but require a separate run for every prediction. To reduce this cost, we cast budgeted attribution as estimating a large influence matrix from a small number of measurements. We show that the measurements most appropriate for recovering this matrix differ from those best suited for attribution itself. We then present two algorithms, MAGE and SPELL, suited for reconstruction and attribution respectively, that run on existing metagradient machinery at no extra cost. Empirical studies demonstrate strong performance over existing baselines across training scales and measurement budgets.
[LG-165] Learned Bow Control on a Measured Bowed-String Model: a Revised Minimum-Bow-Force Law a Recurrent Controller and the Domain of a Supervision Ceiling
链接: https://arxiv.org/abs/2609.14990
作者: Homayoon Beigi,Grace Conneely
类目: Audio and Speech Processing (eess.AS); Machine Learning (cs.LG); Sound (cs.SD); Systems and Control (eess.SY)
*备注: 35 pages, 21 tables, 23 figures, 45 references
Abstract:A finite-difference bowed-string model with implicitly resolved Stribeck friction is presented, with a regime diagnostic, the Schelleng bow-force limits on four strings, and a comparison of learned bow controllers. Implicit resolution is necessary, and quantitatively so: a lagged contact force cannot capture the string on a discrete grid, so no stick phase forms at any bow force. With friction, impedance and quality factor taken from published measurement rather than fitted, all four strings return a stick fraction of 89.1% against an ideal 90%. Schelleng’s maximum bow force is recovered on every string. The minimum is not: it follows Z v_b \beta^-1 rather than the predicted Z^2 v_b \beta^-2 , reducing both squared dependences to first powers. Six controllers at matched capacity, over four strings and twenty seeds each, place a gated recurrent network ahead of a feedforward one, by most under a mid-stroke disturbance. The feedforward network completes more strokes only from a start the model’s own playability map places outside the Helmholtz region. A minimal gated variant fails because gates computed from the input alone cannot clear a latched state. Training loss selects neither the capacity nor the context length, and no learned controller improves on the lookup rule that generated its labels. That bound has a domain. Regressing the controller’s score on the rule’s gives a slope of 0.32, more than ten standard errors below unity, so the controller overtakes the rule where the rule fails and is bounded by it where it holds. Under a rigid finger stop the plant is provably invariant, so transfer loss between pitches belongs to the controller alone and is traced to one feature. A regime classifier without a stick test labels small-amplitude periodic slipping as Helmholtz motion, and a harmonicity measure rates a string the bow never grips above Helmholtz motion.
[LG-166] Linearized PINN with pretrained nonlinear layers
链接: https://arxiv.org/abs/2609.14926
作者: Wenhao Chen,Alexandre M. Tartakovsky
类目: Analysis of PDEs (math.AP); Machine Learning (cs.LG)
*备注:
Abstract:We propose a linearized Physics-Informed Neural Network (lPINN), a reduced-order neural basis method for forward and inverse differential equations. In an offline stage, lPINN learns operator-compatible continuous neural basis functions from an ensemble of numerical solutions. The basis functions are differentiable through automatic differentiation and are pretrained using solution data together with either derivative information or physics residuals. For each new problem instance, the basis functions are frozen and the solution is obtained by minimizing the governing-equation residual together with applicable initial, boundary, regularization, and observational terms. Unlike surrogate and operator-learning methods, the training data define the trial space offline, while the instance-specific solution is computed online by enforcing the governing physics. Relative to vanilla PINNs, lPINN pretrains the nonlinear hidden-layer representation offline and performs online inference only in the final linear layer. We evaluate lPINN on forward and inverse problems for the advection-diffusion equation, Burgers’ equation, and the nonlinear pendulum equation. Compared with vanilla PINNs, lPINN achieves lower solution and parameter errors while reducing online inference times by approximately one to more than three orders of magnitude, with the largest gains generally observed for limited residual or measurement data. Cross-resolution experiments show that the learned continuous representation can be evaluated on finer meshes without retraining and with nearly unchanged accuracy. Subjects: Analysis of PDEs (math.AP); Machine Learning (cs.LG) MSC classes: 35C99 Cite as: arXiv:2609.14926 [math.AP] (or arXiv:2609.14926v1 [math.AP] for this version) https://doi.org/10.48550/arXiv.2609.14926 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-167] Steady-State Convergence of Stochastic Approximation
链接: https://arxiv.org/abs/2609.14922
作者: Yixuan Zhang,Qiaomin Xie
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC); Probability (math.PR)
*备注: 56 pages, 7 figures
Abstract:For constant-stepsize stochastic approximation (SA), the iterates converge in distribution to a stationary law that depends on the stepsize \alpha. Steady-state convergence (SSC) concerns the limit of the scaled stationary distribution as \alpha \downarrow 0. Existing SSC theory requires i.i.d. or additive noise and global differentiability of the mean operator, and yields suboptimal rates. We develop a unified SSC theory for constant-stepsize contractive SA driven by Markovian, multiplicative noise, covering both locally differentiable and locally nondifferentiable mean operators. A key methodological contribution is a multi-step universality framework that progressively reduces the original stochastic recursion to tractable auxiliary dynamics while preserving its steady-state limit. Under local quadratic linearization at the fixed point, we obtain a Gaussian approximation of the scaled steady state at the optimal rate O(\sqrt\alpha) in Wasserstein-2 distance, which further gives finite-time Gaussian approximations for the raw iterates. In the locally nondifferentiable regime, we establish a general SSC result and show that the leading-order asymptotic bias can be of order \sqrt\alpha , in contrast to the \alpha -order bias in the smooth regime. We apply the theory to Markovian linear SA and asynchronous Q-learning, neither of which is covered by prior results. We further propose a bias-reduction scheme for Q-learning that requires no knowledge of the local smoothness regime, validated by numerical experiments.
[LG-168] Shapley Value Estimation for Multi-Site Data with Blockwise-Missing Features
链接: https://arxiv.org/abs/2609.14902
作者: Siqi Li,Wangxuan Fan,Yiming Li,Doudou Zhou,Molei Liu
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Shapley value (SV)-based methods are the prevailing framework for feature attribution in machine learning, yet existing population-level Shapley estimators generally assume that observations used to evaluate the coalitional game are fully observed under a common feature space. This assumption is routinely violated in multi-site studies across biomedicine, social science, and environmental monitoring, where institutions record different features under different protocols, producing systematic blockwise missingness across sources. We first show that the standard remedy of imputing missing features before computing Shapley values introduces systematic, coalition-dependent bias into the resulting attributions. We then propose \textbfFUSHAP (\textbfFusion \textbfShapley \textbfAttribution from \textbfPartially-observed data), a method that leverages partially-observed auxiliary sites to reduce the variance of a preliminary single-site Shapley estimate without imputation. A permutation-based screening step detects and excludes sites whose data distributions are incompatible with the target population. In synthetic experiments, FUSHAP achieves 3 – 8\times lower MSE than the single-site estimator and 2 – 3\times lower MSE than imputation baselines without incurring imputation-induced bias, and the screening procedure identifies misaligned sites with 82% power at moderate misalignment and 100% for strong misalignment. On multi-site air quality and multi-center clinical data, FUSHAP reduces MSE by approximately 3 – 7\times relative to the single-site estimator; in the clinical application, standard imputation can increase MSE above the single-site baseline.
[LG-169] An explicit solution of the five-expert prediction PDE and the exact optimality set of COMB
链接: https://arxiv.org/abs/2609.14892
作者: Jeff Calder,Nadejda Drenska
类目: Analysis of PDEs (math.AP); Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:In this paper, we derive an explicit solution of the stationary prediction with expert advice PDE for five experts. The formula is given in three regions. In the first two regions, it is the four-expert solution plus a single integral with an elementary positive density. In the third region, it is a finite sum of hyperbolic products whose coefficients are determined by one scalar quadrature. Our formula establishes that the direction (1,0,1,0,0) is optimal throughout the ordered sector, and that the COMB strategy (1,0,1,0,1) is optimal only on a lower dimensional subset of the sector (where x_1=x_2 and x_3=x_4 ). This disproves the COMB optimality conjecture of Gravin, Peres and Sivan (2016). The verification of the Hamiltonian inequalities is a tedious task, part of which is completed with a computer assisted proof. The verification reduces to 21 scalar inequalities, which we prove using 147 exact rational Bernstein polynomial certificates. The exact certificates and their independent arithmetic checks are included in a supplement to this paper.
[LG-170] Prescreening Point Defects in Semiconductors With Machine Learning
链接: https://arxiv.org/abs/2609.14846
作者: Paul Karlsson,Joel Davidsson,Rickard Armiento
类目: Materials Science (cond-mat.mtrl-sci); Machine Learning (cs.LG)
*备注: 14 pages, 4 figures
Abstract:High-throughput calculations using density-functional theory (DFT) are commonly used to explore point defects for applications in power electronics and quantum technologies. There is currently a major shift away from these traditional simulation techniques towards machine learning (ML) methods. We explore a class of physics-guided ML models for predicting defect formation energies and zero-phonon lines (ZPL) to identify point defects for quantum applications. The models are specifically targeted for use in a prescreening step for accelerated high-throughput workflows, and are therefore designed to avoid the costly relaxation step typically present with ML interatomic potentials (MLIPs). We compare performance for single and double point defect systems in 4H-SiC with ridge, kernel ridge, and multilayer perceptron (MLP) models using three different descriptors representing the defect systems. For vacancies and substitutions, the optimized models give mean absolute errors (MAEs) of 0.437 eV for the formation energy and 0.202 eV for ZPLs, which is just above the level at which such predictions can be useful even beyond the targeted prescreening, i.e., in some applications they may completely replace the need for costly DFT calculations. For interstitials the MAEs are larger, 1.101 eV for the formation energy and 0.230 eV for the ZPL, which, while still useful for prescreening, will not generally be useful for more detailed characterization. Hence, while the results may be further improved by model design and optimization, the models presented in this work are already useful for prescreening in high-throughput characterization of point defects.
[LG-171] A Functional SVD Framework for Regularized Multivariate Functional PCA with Dual Penalization
链接: https://arxiv.org/abs/2609.14815
作者: Yue Zhao,Hossein Haghbin,Rebecca Sanders,Mehdi Maadooliat
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注:
Abstract:This paper introduces a novel framework for Regularized Multivariate Functional Principal Component Analysis (ReMFPCA) via Functional Singular Value Decomposition (SVD). The proposed method extends existing MFPCA approaches by incorporating a generalized functional SVD within a Hilbert space framework, enabling simultaneous regularization of both functional principal components (PCs) and their associated PC scores. A key innovation of this framework is the inclusion of a sparsity penalty on the PC scores, which enhances interpretability by filtering out irrelevant subject-specific variations. This dual-penalization strategy represents a significant advancement beyond existing covariance-based eigen decomposition methods, which penalize only the functional PCs. Two power algorithm implementations, sequential and joint, are proposed, together with a cross-validation approach based on iterative regression for optimal smoothing parameter selection. Comprehensive simulation studies and real data applications demonstrate that the proposed framework substantially improves the extraction of informative and interpretable components, offering methodological and practical benefits for analyzing multivariate functional data across diverse domains.
[LG-172] From matrix inversion to constraints: provably tighter confidence regions for importance weights in label shift
链接: https://arxiv.org/abs/2609.14802
作者: Mushan Li,Kihyun Han,Yanyuan Ma
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 36 pages, 12 figures
Abstract:Importance weights are essential in domain adaptation under label shift, yet their utility is often undermined by the finite sample uncertainty associated with their estimation. Existing methods typically analyze this uncertainty through Gaussian elimination on interval-valued linear systems, which leads to overly conservative confidence regions and inefficient downstream applications. We propose a paradigm shift from inversion-based inference to a direct matrix constraint framework. We use this framework to define a joint confidence region and extract marginal intervals via linear programming, deriving provably tighter bounds for importance weights while maintaining exact finite-sample validity. Furthermore, we analyze the confidence region’s geometry and provide the theoretical results for its diameter bounds. Evaluated across text, image, multimodal benchmarks, including AGNews, MNIST, CIFAR-10, N24News, and a real-world autonomous driving dataset, nuImages, our approach consistently yields shorter confidence intervals and smaller prediction sets than inversion-based methods.
[LG-173] HELENA for 5G NR LEO NTN Channel Estimation: A Comparative Evaluation
链接: https://arxiv.org/abs/2609.14735
作者: Miguel Camelo Botero,Nina Slamnik-Kriještorac,Johann Marquez-Barja
类目: ignal Processing (eess.SP); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI); Performance (cs.PF)
*备注:
Abstract:Deep Learning (DL)-based channel estimation has shown high accuracy and low latency in terrestrial 5G NR, but Low Earth Orbit (LEO) Non-Terrestrial Networks (NTNs) introduce Doppler and synchronization impairments that may require NTN-specific architectures. We test whether High-Efficiency Learning-based channel Estimation using dual Neural Attention (HELENA), originally designed for terrestrial channels, remains effective after NTN retraining and suitable across high-performance and power-constrained inference platforms. Its unchanged architecture is trained on paired receiver-compensated (NTN-1) and residual-impaired (NTN-2) datasets and compared with eight terrestrial-origin models trained on the same NTN data and the NTN-specific MDELAN-SISO. HELENA achieves the lowest observed SNR-averaged NMSE among the DL estimators in both conditions, including 55.8-62.7% lower linear-scale NMSE than MDELAN-SISO. All DL models degrade in NTN-2, demonstrating the challenge posed by residual Doppler and its associated impairments. On an RTX PRO 4500, HELENA achieves 0.0595 ms 99th-percentile (P99) inference latency, 88.1% below the 0.5 ms budget, with lower energy than its closest attention-based competitors. On a 10 W Jetson Orin NX, it retains a favorable accuracy-energy trade-off, but no model meets the P99 budget. Thus, HELENA needs no NTN-specific redesign for the evaluated task, while embedded tail latency remains an open challenge.
[LG-174] Parameter isolation with domain-specific experts for incremental audio classification
链接: https://arxiv.org/abs/2609.14730
作者: Jongyeon Park,Do-Hyeon Lim,Sang-won Park,Hong Kook Kim,Kyungdeuk Ko,Hyeongcheol Geum,Jeong Eun Lim
类目: Audio and Speech Processing (eess.AS); Machine Learning (cs.LG); Sound (cs.SD)
*备注: 5 pages, 3 figures, 3 tables. Accepted to the Detection and Classification of Acoustic Scenes and Events (DCASE) Workshop 2026
Abstract:To successfully deploy a model in time-varying environments such as streaming data prediction and sensing control, domain-incremental learning (DIL) has attracted attention since it aims to adapt a previously trained model to newly arriving domains, while reserving knowledge from earlier domains without accessing their data. Incremental learning across domains can be regarded as a recurrent update, in which the current model is obtained by updating the model carried over from previous domains. Conventional DIL approaches that rely on domain-invariant feature learning and weight regularization gradually overwrite or constrain parameters learned in previous domains, leading to catastrophic forgetting. Instead, this paper proposes a new domain-specific parameter-isolation architecture that retains all past domains. The proposed architecture mitigates catastrophic forgetting through a full-order recurrent update, constructing a new expert using domain-specific data conditioned on all previously frozen models. To achieve this, we incorporate data-free generative replay to reconstruct previous-domain data and cross-domain feature generation to recover later expert features missing from earlier domain samples. Finally, we apply the proposed model architecture to domain-agnostic incremental learning for audio classification, as defined in the DCASE 2026 Challenge Task 7. Consequently, we achieve micro and macro accuracies of 78.4% and 78.9%, respectively, representing increases of 33 and 25 percentage points over the Challenge baseline. Ablation studies are conducted to examine the effectiveness of each processing component in terms of classification accuracy.
[LG-175] Channel-Adaptive Region Adjacency Graph Carriers for Semantic Image Communication
链接: https://arxiv.org/abs/2609.14616
作者: Karim Abdallah,Maria Slim,Mariette Awad,Hadi Sarieddeen
类目: ignal Processing (eess.SP); Machine Learning (cs.LG)
*备注: 6 pages, 5 figures, 7 tables. Accepted for presentation at IEEE GLOBECOM 2026, Macau. This is the author’s accepted version; the final published version will be available via IEEE Xplore
Abstract:Semantic image communication seeks to preserve task-relevant scene structure under limited channel resources, but carriers are often dense latent tensors or grid-aligned semantic layouts that do not explicitly encode region-level relations. This work introduces a segmentation-derived region adjacency graph (RAG) carrier, termed channel-adaptive RAG (CA-RAG), for joint source-channel coding-style image communication. Nodes store interpretable region attributes, edges preserve adjacency, channel-adaptive graph simplification (CGS) controls the node budget, and semantic belief propagation refines noisy graph embeddings before diffusion-based reconstruction. On Cityscapes, pre-channel RAG payloads are several times smaller than compressed class-index layouts in a 2,000-image study. Under additive white Gaussian noise at signal-to-noise ratios from 0 to 15 dB, CA-RAG reports higher semantic consistency than deep joint source-channel coding and a same-decoder layout baseline, with comparable perceptual quality. At 10 dB, the full-budget rate-sweep point reaches mean intersection over union (mIoU) = 0.329 at approximately 3.3 x 10^3 channel uses, while the default adaptive-CGS setting reports mIoU = 0.294 at approximately 2.6 x 10^3 channel uses.
[LG-176] Evaluation of optimisation and Bayesian inference methods for reaction rates in atmospheric chemical mechanisms
链接: https://arxiv.org/abs/2609.14569
作者: Valery Ashu,Wenqing Peng,Zhi-Song Liu,Heikki Haario,Andreas Rupp,Taiwo Ashu,Petri Clusius,Lukas Pichelstorfer,Zihao Fu,Michael Boy
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Constraining reaction rate coefficients is a central challenge in the development of explicit atmospheric chemical mechanisms, particularly for autoxidation systems where many reaction pathways are only indirectly observed through high-resolution mass spectrometry. In this study, we evaluate rate-coefficient optimisation methods for a toy-case autoxidation mechanism using synthetic data with known ground truth. Two complementary approaches are compared: ODE-constrained neural-network optimisation, which provides efficient point estimates of uncertain rate coefficients, and the Markov Chain Monte Carlo (MCMC) approach, which samples the posterior distribution of rate coefficients and quantifies parameter uncertainty. The methods are tested using direct concentration observations and mass-spectral observations under different noise levels. For unperturbed and low-noise synthetic observations, both methods converged towards the known rate coefficients, with the neural-network optimiser providing faster point estimates. Under high-noise conditions (with the signal-to-noise ratio approximately S / N = 1), however, MCMC was substantially more robust in recovering the rate coefficients. The posterior analysis shows that mass-spectral aggregation broadens credible intervals even at low noise, and that high-noise mass spectra can leave many individual reaction rates weakly identifiable. Posterior predictive validation nevertheless shows how broad parameter uncertainty constrained by MCMC remains consistent with accurate reproduction of the observable mass spectrum. These results demonstrate that point-estimation and Bayesian sampling methods provide complementary information: neural-network optimisation is effective for informative data, whereas MCMC is essential for diagnosing uncertainty, non-uniqueness, and identifiability in noisy or aggregated inverse problems.
[LG-177] Multi-source conformal prediction: leverag ing heterogeneity via localization
链接: https://arxiv.org/abs/2609.14531
作者: Rohan Hore,Anirban Chatterjee,Sayantan Choudhury
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注: 36 pages, 4 figures, 1 table
Abstract:Many modern prediction tasks involve data from multiple heterogeneous sources, while the test distribution may differ substantially from any individual source. Although heterogeneity poses challenges, it also offers an opportunity: different sources may provide complementary information, with some regions of the feature space better represented in one source than another. We propose Multi-Source Randomly Localized Conformal Prediction (MS-RLCP), which builds on the local coverage properties of randomly localized conformal prediction (RLCP) (Hore and Barber, 2025) and extends it to multiple sources through data-adaptive source selection. Under the widely adopted assumption of a shared response distribution conditional on the features across sources and the test population, we establish finite-sample coverage bounds using an interpretable notion of envelope distribution that captures their aggregate feature-space representation. Our analysis allows the test feature distribution to be absolutely continuous with respect to the envelope, extending beyond mixtures of source distributions. Under additional regularity conditions, we also establish asymptotic test-conditional coverage. Simulations and real-world experiments demonstrate the effectiveness of MS-RLCP across varying levels of data heterogeneity.
[LG-178] Certification cost of quantum models: measurement correlation not parameter count
链接: https://arxiv.org/abs/2609.14424
作者: Pavel Sulimov,Claude Lehmann
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG)
*备注:
Abstract:Reporting the Fisher geometry of a trained variational quantum model is routine; quoting the shot budget that would establish it is not. Certifying an empirical Fisher matrix to relative Frobenius error \varepsilon under coordinate-wise parameter shift costs \Theta(B p^2 V/(\varepsilon^2 G)) circuit executions, where V is the measured readout variance and G the measured squared gradient norm, with uniform allocation optimal in that class. One constant reproduces the cost of two circuit families whose exponents differ by a full power of p . The exponent is an identity in how nV and nG scale with the register, holding family by family to 0.001 across 624 matrix-product-state cells once the finite- p prefactor is removed. The cubic cost is therefore a finite-size window, set by whether the readout light cone grows with the register. A product family to 256 qubits gives 1.966 (95% CI 1.934 – 1.997 ); a brickwork entangler falls from 2.853 below ten qubits to 1.715 beyond sixty-four; a blocked entangler gives 1.984 at a fixed cone width against 3.034 at a proportional one. Fixed device connectivity fixes the cone, so a cubic budget from a small simulation overestimates a large machine, on top of hardware multipliers 2.07\times ( 1.41 – 3.02 ), 2.38\times and 1.91\times on ibm_marrakesh, ibm_fez and ibm_kingston. Cost-optimal readout weights cut the measured shot budget by 2.67\times ( 1.33 – 4.00 ) on hardware, flat from four to twelve qubits. A discrepancy model fitted on cheap circuits transfers its mean inside the calibration grid and, at six larger sizes named before the data, does not: nominal 90% intervals cover 36%, and split conformal is the only rung that stays near nominal.
[LG-179] Neural Modal Decomposition: Architectural Priors from Observables
链接: https://arxiv.org/abs/2609.14402
作者: Juho Park,Kaushik Sengupta
类目: Computational Physics (physics.comp-ph); Machine Learning (cs.LG)
*备注:
Abstract:Many engineering building blocks behave as multi-port linear time-invariant systems. RF cavities, photonic devices, and superconducting quantum chips, despite their different underlying physics, all share a common mathematical structure for their port-level response. Each entry of the response matrix is a sum of contributions from a small number of intrinsic resonant modes, the pole-residue form. A model capable of predicting such responses for arbitrary geometries and arbitrary port configurations, while simultaneously extracting the underlying eigenmode structure, would therefore establish a foundational design principle spanning all these domains. We propose a neural framework that learns this modal decomposition end-to-end, supervised only by system-level observables and without supervising the modal parameters themselves. The architecture decomposes into a port-independent pole predictor and two port-dependent coupling predictors whose outputs are combined entry-wise, separating intrinsic from port-dependent features. This factorization yields a single trained model that generalizes to port counts unseen during training, dissolving the \mathcalO(N^2) scaling barrier of direct regression. Despite no modal supervision, the freely-parameterized poles converge to physically meaningful eigenmodes, verified by cross-validation against the AAA rational approximation algorithm. We instantiate the framework in radio-frequency electromagnetic surrogate modeling. A model trained only on 2-port data accurately predicts N -port responses unseen during training.
[LG-180] Parameter Estimation of Ringdown Quasinormal Modes with Autoencoder
链接: https://arxiv.org/abs/2609.14277
作者: Momoka Iida,Hayato Motohashi,Hirotaka Takahashi
类目: General Relativity and Quantum Cosmology (gr-qc); Instrumentation and Methods for Astrophysics (astro-ph.IM); Machine Learning (cs.LG)
*备注: 26 pages, 18 figures, 3 tables
Abstract:Ringdown gravitational waves from binary black hole mergers can be modeled as superpositions of quasinormal modes (QNMs), whose frequencies and excitation factors encode properties of the remnant Kerr black hole. Reliable extraction of multiple QNM components is challenging because of mode overlap and noise. We develop an autoencoder-based framework for multi-component QNM analysis, in which the latent space is trained to represent the physical parameters of individual modes, enabling waveform denoising and parameter estimation within a common framework. Using controlled model waveforms constructed as finite sums of Kerr QNMs with recently established high-precision frequencies and excitation factors, including their nontrivial spin dependence near resonant excitation, we assess the method across partitioned spin intervals. The model achieves good in-domain waveform reconstruction and parameter recovery for the two longest-lived components of eight-component input waveforms, while its performance degrades when the validation spins lie far outside the training range. In a selected spin interval, the framework also recovers the 32 parameters of an eight-component waveform with good overall agreement. These results demonstrate the feasibility of physics-informed autoencoder-based inference for a prescribed multi-component ringdown waveform family and motivate further tests with progressively more realistic signals.
[LG-181] Riemannian ascent–descent for nonconvex nonconcave minimax landscapes: convergence to basin saddle points and applications to distributionally robust optimization
链接: https://arxiv.org/abs/2609.14141
作者: Rishabh Dixit,Pranav Upadrashta,Alex Cloninger
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:We study a class of distributionally robust optimization (DRO) problems for the statistical risk problem, formulated as minimax problems over the product of a Euclidean space and a Riemannian manifold. Because the resulting minimax landscape is nonconvex nonconcave in general, no globally convergent first order method is known to be available. We instead introduce the notion of a \emphbasin saddle point, a Nash equilibrium defined locally on the Cartesian product of a \delta basin around a connected component of the local minima critical set and a geodesic ball on the measure manifold. We develop an abstract convergence framework for a Riemannian gradient ascent multistep descent iteration to a basin saddle point under a local Łojasiewicz type growth condition, with exponent \beta \in (1,2] , in the \delta basin around connected components of the local minima critical sets. Under Lipschitz regularity of critical sets we establish linear convergence for \beta = 2 and polynomial convergence for \beta \in (1,2) to a basin saddle point, with explicit dependence on the sectional curvature of the manifold. We then instantiate this framework for the statistical risk DRO problem over Gaussian measures, where the ambiguity set is naturally modeled as the product of Euclidean space and the Bures Wasserstein manifold of covariance matrices, which we relax to a penalized DRO formulation. We derive nonasymptotic Hessian estimates for the resulting Lagrangian, establish existence and local uniqueness of its maximizer, and prove that an alternating Riemannian gradient scheme converges to a basin saddle point of the penalized DRO problem, recovering the linear and polynomial rates of the abstract theory with all constants explicit in terms of data dimension, loss moments, and the reference covariance.
[LG-182] Exact Finite Attention Responses From RoPE Derivatives
链接: https://arxiv.org/abs/2609.14127
作者: Julie Huang,Maggie Chlon,Gregory Gutin,Leon Chlon
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:We derive exact local responses for attention interventions, allowing candidate edits to be scored from a cached baseline and one backward pass. The starting point is the RoPE derivative \partial_p z§ = A z§ : its integral gives the finite positional displacement, which we carry through the softmax without linearising either rotation or normalisation. The resulting predictions achieve 95.36–96.52% sign accuracy across 92,160 executed positional edits on 768 held-out prompt sets, reducing answer-margin MAE by 73.6–82.5% against the positional Jacobian and by 36.2–50.9% against zero. For simultaneous key and value edits, the same divided-difference calculus isolates the interaction term C_KV = \sum_j (p’_j - p_j),\varepsilon_j , which is omitted by adding separate attributions. Retaining it reduces downstream margin MAE by more than a factor of nine in every setting of a 5,120-intervention sweep across two Qwen sizes, two tasks, and multiple layers; reductions against a quadratic interaction correction are 75.9–98.5%. Exactness concerns the edited attention write; downstream predictions contract that response with a baseline gradient and are evaluated by native execution. The calculus also yields a KL certificate for local approximation error, an exact query-conditioned gradient-step representation whose curvature identifies attention-preserving query directions, and minimum-norm query control. Sparse evaluation supports candidate ranking and cache decisions under explicit local distortion criteria.
[LG-183] Just add noise: Debiasing tree-based variable importance in mixed data
链接: https://arxiv.org/abs/2609.14083
作者: Jiahe Li,Omar Melikechi
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注:
Abstract:Variable importance scores from tree-based methods such as random forests favor continuous predictors over categorical ones. We present a theoretical analysis of this bias and propose a simple remedy: add a small amount of noise to each categorical predictor. The correction is demonstrated on a variety of simulated and real-world datasets and combined with integrated path stability selection to perform variable selection with false discovery control for mixed data.
[LG-184] Conditional Quantum Flow Matching for Data-Scarce Physiological Signal Augmentation
链接: https://arxiv.org/abs/2609.14019
作者: Chi-Sheng Chen,Samuel Yen-Chi Chen
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG); Multimedia (cs.MM)
*备注:
Abstract:Generative augmentation is a standard remedy for label scarcity in physiological signal classification, but existing quantum generative models start from uninformative noise, ignoring class structure that is already available. We propose Conditional Quantum Flow Matching (CQFM): a single 306-parameter circuit, conditioned on both flow time and class label, transports a compact class-conditional prior toward the target distribution. Quantum flow matching as published is unconditional, so this is to our knowledge the first conditional one, and the first EEG augmentation on a parameterized quantum circuit. A nonnegative spectral embedding removes the need for tomography at readout. On BCI Competition IV-2a, starting from a prior rather than noise is worth +5.1 accuracy points over QuDDPM (9/9 subjects), though at that operating point a class-conditional Gaussian matches CQFM. Where the prior fails the transport earns its keep: given one transferred from other subjects it regains +7.2 TSTR points (9/9).
[LG-185] Equilibrium bias and convergence in augmented primal–dual dynamics with sampled constraints
链接: https://arxiv.org/abs/2609.13925
作者: Kang Liu,Mengxiao Chen,Siqi Xiong,Yi Xia
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:This work studies the stability and convergence of augmented primal-dual dynamics when constraint values are estimated from samples. Unbiased constraint observations can produce a biased augmented multiplier signal, shifting the equilibria of the mean dynamics. For componentwise inequalities, we give a necessary and sufficient condition for preserving the Karush-Kuhn-Tucker (KKT) equilibria and construct a convex example with a locally exponentially stable equilibrium that violates complementarity. To address this bias, constraint values are estimated recursively before forming the augmented multiplier signal. For smooth convex conic problems, a joint energy analysis establishes boundedness of the primal, dual, and estimation states, vanishing estimation error, and almost sure convergence of the primal-dual iterates to a single KKT point under global regularity and bounded conditional second moments. The result allows nonunique solutions and multipliers while keeping the number of samples per iteration fixed. Numerical studies illustrate the predicted equilibrium bias and examine convergence with nonunique KKT points and nonlinear constraints.
[LG-186] CyclOT: Learning Quadratic Optimal Transport Maps via Synchronized Forward-Backward Interpolants
链接: https://arxiv.org/abs/2609.13892
作者: Shizhou Xu,Jiachen Liu,Shih-Hsin Wang,Stefan Broecker,Yuhao Huang,Bao Wang,Thomas Strohmer
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注: 55 pages, 13 figures
Abstract:We study the recovery of forward and reverse quadratic optimal-transport maps from unpaired samples in high dimensions. We introduce a bidirectional neural framework in which the learned maps induce forward and backward displacement interpolants, while the training objective combines bidirectional quadratic action, discriminator-restricted Jensen-Shannon endpoint objectives, and two-sided cycle consistency. The construction requires neither precomputed sample pairings nor an explicit convex-potential parameterization. For absolutely continuous probability measures supported on a compact convex set, and under the stated generator-approximation, discriminator-richness, and minimizer-attainment conditions, we prove a population recovery theorem: for every prescribed accuracy, the sum of the corresponding (L^2) errors between any global minimizer and the forward and reverse quadratic Brenier maps is below that accuracy, provided the discriminator level is sufficiently large and the annealing action weight becomes sufficiently small. Moreover, the cycle loss is bounded above by (\lambda W_2^2(\mu_0,\mu_1)). Complementary results quantify approximate invertibility and show that exact endpoint Jensen-Shannon divergence and cycle consistency control missing target mass and many-to-one map collapse, respectively. Experiments on Swiss roll, MNIST, CelebA, single-cell perturbation data, and chest X-ray images evaluate endpoint fidelity, transport cost, inverse consistency, and the geometry of the induced interpolations.
[LG-187] Generation of Custom Solvers in Rust for Convex Optimization
链接: https://arxiv.org/abs/2609.13875
作者: Hao Zhu,Joschka Boedecker
类目: Optimization and Control (math.OC); Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Mathematical Software (cs.MS)
*备注:
Abstract:We introduce cvxgenrust, an open-source tool for generating custom Rust code that solves families of parameterized convex optimization problems modeled in CVXPY. cvxgenrust canonicalizes a problem family, extracts affine maps to Clarabel cone-program data, and generates a specialized Rust crate that updates parameters and calls Clarabel natively at runtime. The generated solver can also be exposed to Python and registered as a custom CVXPY solver. Our code generator supports a wide range of convex optimization problems up to semidefinite programs and exponential-cone problems. Numerical experiments show reduced runtime relative to direct CVXPY solves and performance comparable to CVXPYgen on shared problem classes.
[LG-188] Resolution-Independent Analysis of Encoder–Decoder Operator Learning via Limiting Kernels
链接: https://arxiv.org/abs/2609.13798
作者: Lei Shi,Jia-Qi Yang,Ding-Xuan Zhou
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Functional Analysis (math.FA); Numerical Analysis (math.NA)
*备注: 69 pages
Abstract:Operator learning is formulated on function spaces, but training data are typically available only through finite-dimensional representations. In encoder–decoder architectures, a matrix-valued kernel on the encoded space induces an operator-valued kernel on the original function spaces, and the corresponding reproducing kernel Hilbert spaces are isometrically isomorphic. As the input and output resolutions increase, the induced kernels converge to a limiting kernel, in the sense of operator-norm convergence of their associated integral operators, allowing regularity assumptions to be stated independently of the encoding resolution. For regularized stochastic gradient descent, we establish upper bounds for decreasing and fixed step sizes, separating the encoding and regularization terms from optimization terms of order (t^-\theta) and (T^-\theta’), respectively, for any (\theta,\theta’\in(0,1)). We further prove lower bounds showing that these encoding-induced terms are generally unavoidable. The analysis is further extended to encoder–decoder neural networks through the limiting neural tangent kernel (NTK), yielding error bounds with an additional finite-width term and polynomial parameter and sample complexity guarantees when the encoding errors decay algebraically. The framework covers matrix-valued kernels constructed from radial and dot product kernels, NTKs arising from wide encoder–decoder neural networks, and encoder–decoder pairs based on Fourier, Legendre polynomial, wavelet, PCA, or pointwise sampling representations.
[LG-189] A pullback-corrected scalar auxiliary variable optimizer with momentum and adaptive mobility
链接: https://arxiv.org/abs/2609.13569
作者: Jiahao Zhang,Shiheng Zhang,Guang Lin
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Objectives in scientific machine learning are often prescribed as a sum of several terms, such as the residual, boundary, initial, and data losses of a physics-informed neural network. In the pullback-corrected scalar auxiliary variable (PB–SAV) method, one scalar tracks the shifted objective while the component gradients build a positive semidefinite curvature correction of rank at most the number of components. We carry that correction into an optimizer with momentum and an adaptive mobility, applying it to the gradient and the stored momentum in a single implicit solve. A mobility that is nonincreasing in the Loewner order yields an exact modified energy law, covering Euclidean and AMSGrad-type choices; the corresponding identity for momentum appended after the solve carries a cross term of indefinite sign. For a fixed mobility we give a necessary and sufficient condition for local stability at a stationary point, depending on the Hessian minus twice the correction, and show that it also gives local geometric convergence for every scalar relaxation sequence. The implicit solve reduces to a dense system whose order is the number of components. In the forward Burgers comparison, four components reduce the mean tail objective by 64.7% and the final solution error by 50.2% relative to one component at the same learning rate and momentum settings.
[LG-190] A family of spectral conjugate gradient algorithms derived by least-squares approximations based on a modified quasi–Newton update with application to a revised robust binary classification model
链接: https://arxiv.org/abs/2609.13526
作者: Saman Babaie-Kafaki,Maryam Khoshsimaye-Bargard,Ahmad Mousavi
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:We develop a spectral three-term modification of the classic Hestenes–Stiefel conjugate gradient algorithm, preserving its anti-jamming characteristic and, simultaneously, taking care of the sufficient descent property. We discuss how a modified secant equation can be extracted from our modification scheme, yielding a memoryless BFGS updating formula. Then, the spectral parameter of our method is obtained by steering its direction toward the given BFGS direction within a least-squares context. Using our technical improvements, we outline the general framework of our algorithm and discuss its theoretical features, including the descent and convergence properties, without the convexity assumption. We put our algorithm to the test in comparison with the three other conjugate gradient algorithms on a set of CUTEr unconstrained optimization test models, comparing the outputs using the Dolan–More measure. Next, we provide a concise evaluation of the results, highlighting the practical advantages of our algorithm. As a real-world case study, we introduce a reduced quadratic surface SVM with the rescaled loss for robust binary classification and apply the proposed algorithm to assess its accuracy and training time against several other SVMs.
[LG-191] Inverse Learning of the Altruism and Cost Level in Mixed-Individual Mean Field Games
链接: https://arxiv.org/abs/2609.13469
作者: Haoyang Cao,Gökçe Dayanıklı,Xiaofei Shi
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注: To appear in the 65th IEEE Conference on Decision and Control
Abstract:Understanding how humans respond to incentives, both at the individual and collective levels, is crucial to the design of effective policies. Within the continuous-time stochastic framework for large interacting populations, mean field games (MFGs) model populations of non-cooperative agents, whereas mean field control (MFC) describes the fully cooperative benchmark, interpreted in our setting as fully altruistic behavior. Mixed-individual MFGs interpolate between these two extremes through a parameter governing the degree of altruism. A central challenge for regulators and policymakers, however, is that intrinsic altruism levels and other private structural parameters, such as individual labor costs, are typically unobservable. To address this challenge, we develop an inverse learning framework for mixed-individual MFGs. Our approach enables the recovery of (latent) altruism and labor cost levels from noisy observations, with experiments demonstrating the feasibility and accuracy of our method. These findings underscore the promise of inverse MFG methodologies for uncovering latent preference structures in large populations, with important implications for incentive design, empirical behavioral modeling, and data-driven policy analysis.
[LG-192] Privacy-Preserving Deep Joint Source-Channel Coding with In-Loop Concept Erasure
链接: https://arxiv.org/abs/2609.13393
作者: Rami Eid,Maria Slim,Mariette Awad,Hadi Sarieddeen
类目: ignal Processing (eess.SP); Machine Learning (cs.LG)
*备注:
Abstract:Deep joint source-channel coding (DeepJSCC) transmits learned semantic features efficiently but can leak sensitive attributes such as gender, race, or speaker identity. We propose LEAPSC (LEACE-in-the-loop privacy for semantic communication), whose core contribution is the integration of in-loop least-squares concept erasure (LEACE) within a variational information bottleneck (VIB) encoder. By periodically refitting the projection operator during training, LEAPSC couples the encoder dynamics to the erasure mechanism, driving attribute-conditional mean differences toward zero within each task-label group on the fitting sample. Additional components, namely conditional value-at-risk (CVaR) tail-sensitive privacy, feature-wise linear modulation (FiLM) signal-to-noise ratio conditioning, and Lagrangian dual ascent, improve robustness across channel conditions and over the high-leakage tail of samples. On CelebA, FairFace, and Google Speech Commands, LEAPSC reaches task accuracy of 0.862, 0.755, and 0.925 respectively, with attacker accuracy at or below the label-only floor on CelebA (0.548 vs. floor 0.580) and within 2 percentage points (pp) of chance elsewhere, improving over an information-bottleneck adversarial baseline (IBAL) at a matched 52-epoch budget by +3.6, +2.5, and +1.3 pp (Welch’s t-test, p=0.019 on CelebA).
[LG-193] Beyond Point Forecasts: A Survey on Probabilistic Forecasting for Time Series and Spatiotemporal Data
链接: https://arxiv.org/abs/2609.13345
作者: Donia Besher,Rajdeep Pathak,Madhurima Panja,Tanujit Chakraborty
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Econometrics (econ.EM); Applications (stat.AP)
*备注:
Abstract:Probabilistic forecasting is central to decision-making under uncertainty, yet its methodological landscape has become increasingly fragmented across temporal and spatiotemporal forecasting, statistical modeling, machine learning, and deep generative modeling. This survey develops a unified perspective by organizing probabilistic forecasting methods according to where and how uncertainty is introduced into the forecasting pipeline. Our taxonomy connects model-agnostic approaches including ensembles and distribution-free calibration, with model-intrinsic approaches spanning Bayesian modeling, parametric predictive distributions, distributional regression, and modern generative models, and further examines the emerging role of time series foundation models. Beyond methodological synthesis, we identify the assumptions, computational demands, and forms of uncertainty represented by different paradigms, and translate these distinctions into data-driven and domain-specific guidance for method selection. We complement the survey with a cross-paradigm empirical study on univariate, multivariate, and spatiotemporal forecasting tasks. The results reveal that no single uncertainty-quantification paradigm dominates across settings. Calibration, sharpness, predictive accuracy, and computational efficiency can lead to substantially different model preferences, while expressive generative models and zero-shot foundation models exhibit markedly different accuracy-efficiency trade-offs. Lastly, we identify unresolved challenges surrounding uncertainty in evolving dependency structures, physics-informed predictive distributions, forecasting extreme events, handling count-valued, directional, and continuous-time series, and the development of unified software resources. Our survey provides both a conceptual framework and a practical roadmap for probabilistic forecasting research.
[LG-194] Stochastic Gradient Descent over P2
链接: https://arxiv.org/abs/2609.13343
作者: Maria Oprea,Qin Li,Yunan Yang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Probability (math.PR)
*备注:
Abstract:Stochastic gradient descent (SGD) admits diffusion approximations that replace the complicated randomness of stochastic gradients by Gaussian noise, providing a powerful tool for understanding its dynamics and long-time behavior. We investigate whether an analogous approximation principle holds for optimization over probability measures, where the objective is a functional defined on the Wasserstein space P2. The nonlinear geometry and infinite-dimensional nature of P2 prevent a direct extension of the classical Euclidean theory. Using Lions differentiability, we lift the problem to a linear Hilbert space, where higher-order differential calculus becomes available. We then construct a Gaussian random-field approximation whose velocity field matches the mean and covariance of the original stochastic gradient. By exploiting this moment matching through higher-order Taylor expansions, we show that the Gaussian approximation captures the SGD dynamics with second-order weak accuracy. Our result provides a rigorous foundation for replacing sample-driven randomness by analytically tractable Gaussian fluctuations in stochastic optimization over probability measures.
[LG-195] Aries: A Proprietary Medium-Range Weather Prediction Model for the Energy Industry
链接: https://arxiv.org/abs/2609.13292
作者: Lukas Hedegaard Morsing,Arian Bakhtiarnia,Jonas Lynge Olesen,Tómas Bragi Björnsson Leth,Christian Gøbel Bach
类目: Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG)
*备注:
Abstract:Medium-range weather forecasting underpins operational and planning decisions across the energy industry. Developing competitive weather models was once the domain of national meteorological centers, but recent advances in machine-learned weather prediction (MLWP) have opened the field to industry. We present Aries, a SwinTransformer-based MLWP model developed at InCommodities. Aries is trained on ERA5 reanalysis data at 0.25\textdegree resolution, predicting 74 prognostic and 11 diagnostic atmospheric variables. We evaluate the model on 2025 ECMWF Analysis initializations, ensuring a recent and strictly out-of-sample test period for all models compared. On 10-metre wind speed, Aries outperforms both ECMWF HRES and AIFS in terms of RMSE for lead times up to four days, while on 2-metre temperature it achieves RMSE on par with AIFS operational. These results demonstrate that proprietary development of competitive weather models is technically viable, supporting a broader set of forecasts available for operational and planning applications in the energy industry.
[LG-196] Large-scale bioacoustic detection using semantic segmentation: a deep learning framework applied to fin whale calls in ocean-bottom seismometer recordings
链接: https://arxiv.org/abs/2609.13281
作者: Jocelyn Japnanto,Alex A. Saoulis,Miriam Romagosa,Rita Leitão,Gabrielle Arrieta,Mónica A. Silva,Matthew Graham,Ana M. G. Ferreira
类目: Geophysics (physics.geo-ph); Machine Learning (cs.LG)
*备注:
Abstract:Ocean-bottom seismometers (OBS), originally deployed for geophysical research, continuously record low-frequency sound for months to years across broad areas of ocean, offering a largely untapped resource for passive acoustic monitoring (PAM) of baleen whales. Realising this potential requires automated detection methods that operate reliably across the varied conditions in large sensor networks. We present a deep learning semantic segmentation framework that detects the 20-Hz notes of fin whales (Balaenoptera physalus) in OBS spectrograms, assigning each pixel a probability of belonging to a call and converting the resulting probability maps into time-frequency bounding boxes describing individual detections. We trained the model on hydrophone data from one OBS deployment in the Azores-Madeira-Canaries region and applied it without retraining to vertical-component seismometer data from a second, geographically distinct deployment, showing that a single trained model generalises across sensor types and recording environments. Applied to 378,912 h of recordings from 46 OBS sites, the detector identified 6.3 million calls, forming the largest fin whale call catalogue assembled to date, with high precision (~97%) across both deployments. The resulting catalogue resolves call timing and spectral structure accurately enough to support ecological analyses, revealing coherent seasonal shifts in three persistent inter-note interval (INI) groups across the singing season and basin-scale patterns in calling activity. By transforming existing geophysical infrastructure into a scalable sensing network, our approach substantially expands the spatial and temporal reach of PAM without new hardware investment, offering a transferable framework for tracking other low-frequency vocalising species and informing conservation planning, marine spatial management, and abundance estimation across large scales.
[LG-197] Survey of Novel Deep Learning Architectures for Denoising Gravitational-wave Signals
链接: https://arxiv.org/abs/2609.13272
作者: Rohan Raha,Prayush Kumar
类目: General Relativity and Quantum Cosmology (gr-qc); High Energy Astrophysical Phenomena (astro-ph.HE); Instrumentation and Methods for Astrophysics (astro-ph.IM); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
*备注: Submitted to “Machine Learning: Science and Technology”, 52 pages, 35 Figures, 8 Tables
Abstract:Gravitational-wave denoising must handle the full diversity of spinning, precessing binaries, since the recovered waveform underpins parameter estimation, tests of general relativity, and population studies. Matched filtering achieves this at a cost that becomes prohibitive as next-generation detectors push event rates higher; deep learning offers real-time reconstruction, but current methods are developed on narrow parameter spaces, precluding principled comparison and reliable deployment. We present the first controlled comparison of five neural-network architectures for gravitational-wave denoising, trained identically across the full astrophysically-motivated spinning binary-black-hole parameter space. A unifying principle emerges: matching network structure to the spectral anatomy of a coalescence – inspiral, merger, ringdown – outperforms brute-force model scaling. Our Multi-Scale Frequency-Aware architecture embodies this via dedicated parallel branches per frequency regime, achieving the best fidelity while using fewer parameters than larger models. It generalizes from simulated training to real LIGO-Virgo-KAGRA data without retraining, recovering merger morphology across confirmed events spanning three observing runs and both detectors, despite training on a single detector’s simulated noise. We construct population-level uncertainty bands from denoising residuals, validate their calibration, and stress-test the framework on extended mass ratios and a spin population resembling hierarchical-merger remnants. Applied to real noise with no known signal, the network suppresses its output almost everywhere, with rare exceptions traced to noise transients rather than a general weakness – indicating the statistic discriminates signal from noise and motivating a future detection study. Released weights give a deployable, reproducible benchmark for future extensions.
[LG-198] Calibrating subgrid parametrizations of single-column ocean models via simulation-based inference
链接: https://arxiv.org/abs/2609.13242
作者: Luben M. C. Cabezas,Sacha Wendling,Aurèle Gallard,Gabriel Mouttapa,Julien Le Sommer,Pedro L. C. Rodrigues
类目: Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG)
*备注:
Abstract:Subgrid parametrizations of vertical mixing in ocean models depend on free coefficients that cannot be measured directly and must be calibrated against high-fidelity references such as large-eddy simulations (LES). Existing approaches return point estimates and leave the associated uncertainty unquantified, a limitation when the inverse problem is ill-posed or when distinct parameter configurations fit the data comparably well. Simulation-based inference (SBI) addresses exactly this: given a prior and access to the simulator, it approximates the full posterior over parameters without requiring a tractable likelihood, at a cost set by the number of simulator evaluations. We apply it to \texttttunax, a JAX-based single-column ocean model, to calibrate the coefficients of its k – \varepsilon closure. A blockwise PCA summary statistic compresses the simulator output along the depth axis while preserving its forcing–horizon–variable structure, making inference tractable at modest budgets. We compare neural posterior estimation and its sequential variants against a recent training-free approach built on a tabular foundation model. The latter recovers informative posteriors from a few hundred simulator calls, outperforming the trained estimators at every budget considered.
[LG-199] From objective discovery to prediction of global ocean eco-provinces: A pathway for trustworthy learning
链接: https://arxiv.org/abs/2609.13206
作者: Makayla McDevitt,Maike Sonnewald,Stephanie Dutkiewicz
类目: Populations and Evolution (q-bio.PE); Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Atmospheric and Oceanic Physics (physics.ao-ph)
*备注: 23 pages 10 figures (15 without figures)
Abstract:Marine ecosystems are increasingly impacted by climate change, necessitating tools to identify and predict spatial habitat information. To build such tools, ecological marine provinces, “eco-provinces”, ecologically meaningful regions in the global ocean can be used. We use unsupervised machine learning (ML) to identify eco-provinces with corresponding uncertainty measures based on output of a global simulation of phytoplankton functional types. Our work aims to create a proof of concept to predict eco-provinces based on satellite ocean color data. To do so, we develop a hierarchy of explainable dense ensemble networks to infer how well the eco-provinces can be detected from modeled ocean color fields. Key results include that the delineated eco-provinces are both ecologically meaningful and can be inferred with high skill. However, no straightforward relationship was found where adding more input data to the network consistently improves inference skill, and there is an intricate tradeoff between inputs and prediction fidelity. Our work is a case for optimism and a cautionary tale of needing uncertainty quantification and careful validation of prediction fidelity.
[LG-200] Exploring new directions in enhancing the ACTS parameter optimization suite
链接: https://arxiv.org/abs/2608.14714
作者: Chance LaVoie,Qi Bin Lei,Rocky Bala Garg,Lauren Tompkins
类目: Data Analysis, Statistics and Probability (physics.data-an); Machine Learning (cs.LG)
*备注: Proceedings for ACAT 2025, 5 pages, 2 figures
Abstract:Track seeding strongly affects both the quality and computational cost of charged-particle reconstruction, yet its many configuration parameters are commonly tuned through expert intuition and repeated trial and error. ACTS reduces this burden with an Optuna Tree-structured Parzen Estimator auto-tuner, but expensive evaluations, a restricted search space, and a scalarized objective can limit evaluation efficiency, exclude promising configurations, and obscure performance trade-offs. We investigate whether Bayesian optimization can address these limitations using ACTS with the Open Data Detector (ODD). Under identical search ranges and a common 100-trial budget, we compare Expected Improvement and Upper Confidence Bound with TPE and random search on the existing eight-parameter problem, extend the best-performing Bayesian method to fifteen parameters, and apply Expected Hypervolume Improvement to optimize efficiency, fake rate, duplicate rate, and runtime without fixed scalar weights. Candidate configurations are evaluated through the full ACTS reconstruction chain and validated on disjoint held-out events. The Bayesian acquisition methods identify strong configurations earlier than TPE, and their advantage persists in held-out validation. Expanding the search further improves performance, while multi-objective optimization reveals competitive non-dominated solutions spanning distinct trade-offs. These results indicate that Bayesian optimization can strengthen ACTS auto-tuning through efficient evaluation, broader parameter searches, and post-hoc expert selection among non-dominated alternatives. Comments: Proceedings for ACAT 2025, 5 pages, 2 figures Subjects: Data Analysis, Statistics and Probability (physics.data-an); Machine Learning (cs.LG) Cite as: arXiv:2608.14714 [physics.data-an] (or arXiv:2608.14714v1 [physics.data-an] for this version) https://doi.org/10.48550/arXiv.2608.14714 Focus to learn more arXiv-issued DOI via DataCite
[LG-201] QSTAR: Quantum Selective Transfer with Adaptive Routing
链接: https://arxiv.org/abs/2607.21411
作者: Saim Rehman,Nouhaila Innan,Muhammad Shafique
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG)
*备注:
Abstract:Quantum transfer learning (QTL) is often evaluated by replacing a classical classifier with a fixed variational quantum head, but this hides a key question: when is the quantum branch actually useful? We propose QSTAR: Quantum Selective Transfer with Adaptive Routing, a selective QTL framework that keeps high-confidence classical predictions and routes only low-confidence samples to a fallback branch. Using a frozen ResNet18 backbone on Fashion-MNIST, we compare manually designed QTL heads, KetGPT-designed quantum heads, and parameter-matched classical baselines under a common data split and optimization schedule. Standard QTL heads reach at most 57.0% accuracy, while the strongest KetGPT head in the main filtered sweep reaches 78.5% accuracy and 0.785 F1-score. Although the strongest fixed classical head remains higher at 81.6%, selective routing gives the quantum branch a clearer role. On low-confidence samples, KetGPT #180 improves accuracy over a parameter-matched MLP fallback by 6.82, 4.31, and 3.03 percentage points at thresholds of 0.70, 0.80, and 0.90. At the full-system level, Adaptive KetGPT-QTL reaches 80.9% accuracy and 0.807 F1-score, outperforming the adaptive classical baseline. A separate compact-circuit ablation identifies KetGPT #160 as a stronger fixed-head candidate, reaching 81.9% accuracy with only 10 quantum parameters and 9 gates. These results suggest that architecture-searched quantum heads are most useful as targeted fallback branches for uncertain inputs rather than uniform replacements for classical classifiers.
[LG-202] Approximating neutron-star radii using gravitational-wave only measurements with symbolic regression
链接: https://arxiv.org/abs/2504.19962
作者: Michał Bejger
类目: General Relativity and Quantum Cosmology (gr-qc); High Energy Astrophysical Phenomena (astro-ph.HE); Machine Learning (cs.LG)
*备注: 10 pages, 6 figures, 2 tables; PRD accepted
Abstract:Gravitational waves emitted by binary neutron-star inspirals carry information on components’ masses and tidal deformabilities, but not directly radii, which are measured by electromagnetic observations of neutron stars. To improve the multi-messenger astronomy studies of neutron stars, an expression for neutron-star radii as a function of gravitational-wave only data would be advantageous, as it would allow to compare information from two different channels. In order to do so, a symbolic regression method, pySR, is trained on TOV solutions to piecewise polytropic EOS input to discover an approximate symbolic expression for the neutron-star radius as a function of gravitational-wave measurements only. The approximation is tested on piecewise polytropic EOS NS data, as well as on NS sequences based on selected realistic (non-polytropic) dense-matter theory EOSs, achieving consistent agreement between the ground truth values and the symbolic approximation for a broad range of NS parameters covering current astrophysical observations, with average radii differences of few hundred meters. Additionally, the approximation is applied to the GW170817 gravitational-wave mass and tidal deformability posteriors, and compared to reported inferred radius distributions.
附件下载
