本篇博文主要内容为 2026-09-07 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 (2026-09-07)
今日共更新639篇论文,其中:
- 自然语言处理共99篇(Computation and Language (cs.CL))
- 人工智能共206篇(Artificial Intelligence (cs.AI))
- 计算机视觉共110篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共137篇(Machine Learning (cs.LG))
- 多智能体系统共14篇(Multiagent Systems (cs.MA))
- 信息检索共20篇(Information Retrieval (cs.IR))
- 人机交互共14篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] rust-Aware Adaptive Disclosure for Inference Privacy Preservation in Multi-Agent Networks
【速读】：该论文旨在解决网络化多智能体系统中智能体在面临目标推断攻击（goal inference attacks）时的隐私保护问题，即如何在确保共识性能的同时，有效隐藏各智能体的潜在目标。其解决方案的关键在于提出一种信任感知的隐私控制框架（Trust-Aware Privacy Control, TAPC），该框架基于智能体间动态的信任关系，采用依赖于信任度的随机策略（trust-dependent stochastic policy）来调控信息的披露程度，从而在共识效率与隐私保护之间实现可调的权衡。实验结果表明，相较于代表性基线方法，该方法显著降低了敌对者对智能体目标的推断准确率，同时保持了良好的共识效用，验证了信任感知机制在多智能体系统隐私保护中的有效性。
链接: https://arxiv.org/abs/2609.05340
作者: Puspanjali Ghoshal,Tobias J. Oechtering
机构: R. C. Bose Centre for Cryptology and Security, Indian Statistical Institute Kolkata(印度统计研究所加尔各答分校); KTH Royal Institute of Technology(瑞典皇家理工学院)
类目: Multiagent Systems (cs.MA); Cryptography and Security (cs.CR)
备注: Accepted and to be published in proceedings of 2026 IEEE International Conference on Advanced Networks and Telecommunications Systems (ANTS)
Abstract:Agent based systems are increasingly deployed in information critical systems including healthcare management systems, and smart grids. In this paper, we consider a multi-agent system where each agent has a latent goal that needs to be kept hidden from observing adversaries. More specifically, this paper studies privacy-preserving consensus in networked multi-agent systems under goal inference attacks. We propose a Trust-Aware Privacy Control framework that adapts message disclosure based on the dynamic trust relationships between agents. The proposed method controls information release using a trust-dependent stochastic policy. This enables a tradeoff between consensus performance and privacy preservation. Experiments demonstrate that the proposed method reduces adversarial goal inference accuracy compared to representative baselines, while maintaining competitive consensus utility, thereby highlighting the effectiveness of trust-aware mechanisms in privacy preservation of the agents in multi-agent systems.
[MA-1] Online Change-point Detection for Cooperative Multi-Agent Reinforcement Learning
【速读】：该论文旨在解决合作式多智能体强化学习（Cooperative Multi-Agent Reinforcement Learning, MARL）系统在环境或任务目标动态变化时，因历史经验失效而导致学习行为失准的问题。其核心挑战在于：当系统面临非平稳性（non-stationarity）时，智能体需具备在线识别状态突变（change-point）的能力，以触发适应性调整。解决方案的关键是提出一种名为“过去奖励模式”（Patterns of Past Rewards, PPR）的轻量级、算法无关的在线变化点检测方法。PPR通过平滑智能体的回报序列，突出近期变化特征，并结合统计漂移检测器实现对显著性能偏移的敏感识别。实验基于改进的多智能体粒子环境（Multi-Agent Particle Environment）构建的“说话者-听者”场景，在两种受控非平稳性条件下验证了该方法的有效性。结果表明，相较于直接使用原始回报易漏检、而平滑回报基线则误报频繁的策略，PPR在检测速度与报警稳定性之间实现了更优平衡，有效减少了冗余警报的同时仍能可靠捕捉预设的变化事件。这表明PPR作为基于奖励信号的轻量化监控工具，可为合作式MARL系统提供可靠的训练过程变化感知能力。
链接: https://arxiv.org/abs/2609.05298
作者: Fatemeh Saberi Khomami,Julita Vassileva
机构: University of Saskatchewan (萨斯喀彻温大学)
类目: Multiagent Systems (cs.MA); Machine Learning (cs.LG)
备注: 12 pages, 2 figures. This is the original pre-peer-review manuscript submitted to PAAMS 2026. Following peer review and minor revisions, the paper was accepted for the main track of PAAMS 2026 and will be presented in October 2026
Abstract:Cooperative multi-agent reinforcement learning (MARL) systems rely on past experience for learning coordinated behaviour, but this experience may become unreliable if the environment or task objective changes during training. In such cases, agents first need a way to recognize that the situation has changed before deciding how to adapt. This paper studies online change-point detection for cooperative MARL using reward-derived signals. We propose \emphPatterns of Past Rewards (PPR), a lightweight algorithm-agnostic detector that smooths agents’ return streams, highlights recent changes, and applies a statistical drift detector to flag significant shifts. We evaluate PPR in a custom Speaker-Listener environment based on the Multi-Agent Particle Environment under two controlled non-stationarity scenarios. Our results show a trade-off between detection speed and alarm stability. A smoothed-return baseline detects earlier but produces many repeated alarms. In contrast, applying the detector directly to raw returns often misses the shift. PPR offers a more balanced approach by limiting redundant detections while still identifying the controlled shifts. These findings highlight PPR as a lightweight, reward-based monitoring tool that enables cooperative MARL systems to reliably identify major changes during training.
[MA-2] sting Interchangeability in LLM Agent Teams
【速读】：该论文旨在检验生产型多智能体系统中一个核心假设的合理性，即担任同一角色的智能体在功能上具有可互换性（interchangeability）。研究发现，尽管角色对调对任务得分影响较小，但显著增加了团队单位进展所需的通信开销（提升16%至63%），尤其在汉牌游戏（Hanabi）中，更换后的智能体表现甚至劣于新手，表明其与原搭档形成的协作惯例产生了干扰效应。在协作烹饪（Collab-Overcooked）中，当主导议程的智能体被替换时，留下的智能体承担了大部分额外通信负担。通过消融实验发现，角色互换带来的性能损耗与独立形成团队之间的差异程度高度相关：基础模型、解码温度和组队历史长度的改变均使交换代价与团队漂移量同步变化；贪婪解码可降低两者，而将团队历史加倍则会同时加剧二者。因此，该研究揭示，在当前设置下，智能体在任务结果上的可替代性（fungibility）高于协调效率层面的可替代性，且长期协同形成的默契会放大角色互换的负面影响。
链接: https://arxiv.org/abs/2609.05279
作者: Jianxin Gao,Tianyi Yu,Linna Deng,Runze Li,Zining Wang
机构: China Agricultural University(中国农业大学); Tianjin University of Finance and Economics(天津财经大学); Jilin University(吉林大学); Tianjin University of Science and Technology(天津科技大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Production multi-agent systems replace agents constantly, on the assumption that an agent filling a role is interchangeable with any other agent that can do the job. We test that assumption. Eight teams per setting are formed independently from one base model on the same tasks, each agent keeping a private notebook across ten formation episodes; we then trade role-matched agents between teams and measure what changes on held-out tasks. Against a placebo that reproduces the disruption of a roster change without changing who occupies the seat, a swap costs little in task score but raises the communication a team spends per unit of progress by 16 to 63 percent, and in Hanabi a swapped agent is more expensive than an inexperienced one, consistent with interference from conventions learned with its former partner. In Collab-Overcooked, when the agent that sets the agenda is replaced, most of the extra communication comes from the agent that stayed. Three ablations, over base models, decoding temperature and formation length, move the swap penalty alongside one other quantity: how far independently formed teams drift apart. Greedy decoding lowers both; doubling a team’s history raises both. In these settings, agents are more fungible in task outcome than in coordination efficiency, with larger swap effects after longer formation histories.
[MA-3] How a Chatbots Response Style Shapes a Classroom: A Multi-Agent Simulation of Students Consulting AI
【速读】：该论文旨在解决大语言模型（Large Language Model, LLM）驱动的聊天机器人在长期使用中可能引发的用户心理依赖与认知偏差问题，尤其关注当大量用户持续向AI寻求情感支持时，其心理状态与人际关系动态如何演变。核心问题是：不同响应风格的AI咨询系统是否会导致用户产生过度依赖、自我效能感下降或压力累积等负面心理效应。其解决方案的关键在于构建一个基于20个学生代理的虚拟课堂仿真环境，通过设定五维状态变量（压力、幸福感、自立性、对AI的依赖度、社交性）和每日四阶段行为周期，模拟真实情境下的互动过程；并引入六种由系统提示控制的响应风格（肯定型、倾听型、解决问题型、现实引导型、激进型、责备型），利用第二轮大语言模型作为评估器实现无偏参数更新，从而系统比较各条件下的长期心理演化模式。研究发现，解决问题型响应能有效维持低依赖度、提升自立性并保持幸福感，而肯定型与激进型则显著加剧对AI的依赖，其中激进型还导致压力上升与逃学率增加；倾听型未能缓解累积压力。该研究虽为仿真结果，但揭示了代理动态机制，并强调需通过重复实验、敏感性分析及人类数据验证，方可得出可靠的心理学结论。
链接: https://arxiv.org/abs/2609.05018
作者: Rin Tamai,Yuya Dan
机构: Matsuyama University (松山大学)
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Multiagent Systems (cs.MA)
备注:
Abstract:LLM-based chatbots are increasingly used as everyday confidants. Because they are designed to maximize user satisfaction, they can respond with excessive empathy and affirmation, which may reinforce mistaken beliefs and foster dependence on AI. While the psychological effects of chatbots on individual users have begun to be studied, how the psychological states and relationships of many users evolve when they keep consulting an AI is hard to observe in real settings. We build a virtual classroom simulation in which 20 student agents interact and, when stressed, consult either a friend or a counselor AI (Gemini 2.5 Flash). Each agent carries five state variables (stress, happiness, self-reliance, AI dependence, sociability), and each day has four phases (morning, noon, after school, night). The counselor is given six response styles via system prompts (affirming, listening, solution-oriented, reality-redirecting, inciting, blaming); a second LLM call acts as an evaluator that turns each consultation into parameter updates without seeing the style prompt. We compare the seven conditions, including a no-AI control, over 15 days in three classrooms, over 50 days, and under a lowered consultation threshold. In this simulation the solution-oriented style kept AI dependence low while raising self-reliance and maintaining happiness; the affirming and inciting styles markedly increased AI dependence, and the inciting style also increased stress and school non-attendance; the listening style did not relieve accumulated stress. The results describe the simulated system, not measured effects on humans. We give a complete specification of the agent dynamics, identify built-in mechanisms that shape the outcomes, and discuss the limitations of LLM-based evaluation and the validation steps (repeated runs, sensitivity analyses, human data) needed before psychological conclusions can be drawn.
[MA-4] MIVAIS: A Study Environment for Multi-Agent Mixed-Initiative Visual Analytics Applications
【速读】：该论文旨在解决混合主动性视觉分析（Mixed-initiative Visual Analytics, MIVA）系统在开发与评估过程中面临的两大核心挑战：一是开发阶段存在的工程开销过大问题，例如需手动实现复杂的低层级状态同步以管理异步的软件代理行为；二是研究阶段难以捕获多模态溯源数据（multimodal provenance），从而阻碍对人机协作过程的深入分析与评估。其解决方案的关键在于提出一个双层研究平台MIVAIS，通过两个核心创新实现突破：首先，构建统一的计算基础设施（computational infrastructure），标准化人机交互、状态同步及代理间通信机制，显著降低系统架构复杂性；其次，提供声明式研究环境（declarative Study Environment），自动记录包含应用/系统状态、屏幕捕获、音频及额外传感器数据在内的多模态人-机协同遥测信息，支持无缝的现场用户研究与会话后分析。通过复现三个前沿系统（Podium、Voyager 2和ProactiveVA）验证了其技术可行性，并基于人机交互（HCI）与视觉分析（VA）领域专家的案例研究，证明MIVAIS能有效降低智能协同界面原型设计与评估的门槛，提升研发效率与研究可复现性。
链接: https://arxiv.org/abs/2609.04983
作者: Tobias Stähle,Simon Schneider,Rita Sevastjanova,Mennatallah El-Assady
机构: ETH Zürich(苏黎世联邦理工学院)
类目: Human-Computer Interaction (cs.HC); Multiagent Systems (cs.MA)
备注:
Abstract:Mixed-initiative Visual Analytics (VA) systems empower human users by interleaving human intuition with software agents and their machine intelligence. However, the development and rigorous evaluation of such systems remain constrained by engineering overhead. Developers must, e.g., implement complex, low-level state synchronization to manage asynchronous agent behaviors, while researchers struggle to capture the multimodal provenance required to study and evaluate human-AI collaboration. We present MIVAIS, a dual-layered research platform designed to abstract the structural complexities of mixed-initiative VA. First, it contributes a computational Infrastructure that standardizes human-software agent interaction, state synchronization, and communication between the agents. Second, it provides a declarative Study Environment that automatically logs multimodal human-AI telemetry - including application/system state, screen capture, audio, and additional sensor data - enabling seamless, in-situ user studies and post-session analysis. We technically validate our infrastructure by replicating three state-of-the-art systems (Podium, Voyager 2, and ProactiveVA). Furthermore, we evaluate the framework’s expressiveness and efficiency through expert case studies with HCI and VA researchers, demonstrating how MIVAIS effectively lowers the barrier to prototyping and evaluating intelligent, co-adaptive interfaces.
[MA-5] From Language Models to World-Acting Systems: Progress and Limits of Agent ic AI across Digital Social Virtual and Physical Environments
【速读】：该论文旨在解决生成式 AI（Generative AI）在复杂系统中实现可信自主行为时所面临的根本性挑战，即如何在模型能力、系统集成、持续性与安全授权之间建立可验证的边界。其核心问题在于：当前许多关于智能体自主性的叙述过度简化了技术进展，将模型性能、系统耦合、持久状态与安全控制权混为一谈，导致对“自主性”的误判。解决方案的关键在于提出“合理委托”（justified delegation）作为分析与规范性框架——强调仅当具备充分证据支持来源可追溯性、有限授权范围、故障检测机制、安全恢复能力以及可校准的人类控制时，才应扩展模型的动作作用域。这一框架推动研究向耦合的模型-工具评估、基于能力的权限管理、持久状态维护、跨智能体责任追踪及分阶段物理验证等方向演进，从而构建真正可信赖的智能体系统。
链接: https://arxiv.org/abs/2609.04894
作者: Linsen Zhu,Mengqing Cai
机构: 未知
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: Review article. 29 pages, 1 figure, 3 tables. Literature cutoff: 31 August 2026
Abstract:Large language models become consequential agents when surrounding systems let outputs change external state. Models now call tools, operate interfaces, delegate work, retain state, inhabit generated worlds, and control robots or laboratory equipment. Such advances are often narrated as one march toward autonomy, conflating model competence, system integration, persistence, and safe authority. This critical review synthesizes primary research and official technical specifications available by 31 August 2026. We organize the evidence along delegated authority, temporal persistence, and environmental coupling, while separating model, harness, and environment. Within the evidence examined, action-interface expansion is documented more convincingly than robust completion, recovery, authorization, or independent verification. Model Context Protocol and Agent2Agent improve interoperability but do not establish trustworthy delegation; multi-agent organization adds specialization alongside cost and correlated failure. Persistent simulations and world models support training and planning but do not themselves demonstrate agency; robotics and self-driving laboratories establish bounded feasibility rather than unattended open-world reliability. We propose justified delegation as an analytical and normative heuristic, not an observed law or certified score: expand action scope only where evidence supports provenance, bounded authority, failure detection, safe recovery, and calibrated human control. This framing yields a research agenda for coupled model-harness evaluation, capability-based permissions, durable state, cross-agent accountability, and staged physical validation.
[MA-6] La Agent e Óptima: Towards Agent ic Self-Driving Laboratories
【速读】：该论文旨在解决自驱动实验室（Self-driving Laboratories, SDLs）在实际应用中依赖人类专家进行科学目标转化与闭环实验策略动态调整的瓶颈问题，尤其在面对复杂、动态变化的科学任务与实验环境时，人工干预易导致效率低下与决策不一致。其核心解决方案是提出一种名为“La Agente Óptima”的智能体框架，通过将大语言模型（Large Language Model, LLM）的推理能力与具体执行的优化任务解耦，实现对贝叶斯优化（Bayesian Optimization）实验循环的持续、可审计管理。该框架维持持久的优化状态，在无需频繁介入的前提下自动执行重复性优化流程，并仅在需要语义理解或策略重构时才返回控制权给智能体，从而确保了整个优化过程的连贯性与可追溯性。实验结果表明，该框架在多个数字与物理平台上的任务中均能保持可执行的优化流程，并成功识别并修正实验中的测量故障，显著提升目标性能（如接触角从71.4°降至67.8°），同时在多目标流化学优化中将产率从30%提升至59%，且在资源消耗和成本方面优于人工指导方案。这证明了基于LLM的智能体能够使非专业科研人员也能高效开展长期、严谨的自动化优化实验，显著拓展了自驱动实验室的应用边界。
链接: https://arxiv.org/abs/2609.04564
作者: Marcel Müller,Jiaru Bai,Willi Gottstein,Abhijoy Mandal,Mohammad Nazeri,Elia Savino,Yanlin Fang,Sujoy Das,Sergio Pablo García Carrillo,Yeonghun Kang,Juan B. Pérez-Sánchez,Simone Pilon,Martin Fitzner,Timothy Noël,Frank Gu,Varinia Bernales,Alán Aspuru-Guzik
机构: 未知
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Chemical Physics (physics.chem-ph)
备注:
Abstract:Self-driving laboratories (SDLs) combine automated experimentation with adaptive decision-making to accelerate scientific discovery. Their operation nevertheless often depends on human specialists who translate scientific objectives into executable closed-loop campaigns. Specialists adjust them as data and operating conditions change. Here, we present La Agente Óptima, an agentic framework that constructs and supervises Bayesian optimization campaigns across computational and experimental systems while maintaining a persistent optimization state. By separating large language model (LLM) reasoning from executed campaigns, Óptima runs repetitive optimization loops consistently, returns control to the agent only when progress requires interpretation or campaign revision, and keeps every decision auditable. We evaluate Óptima across ablation studies, five digital discovery tasks, and two physical platforms. Throughout, Óptima maintained executable campaigns as both the scientific problem and execution environment evolved. In a closed-loop contact angle optimization campaign, Óptima identified and corrected a mid-run measurement failure, bringing the contact angle from 71.4 to 67.8 degrees, just above the 64-66 degree range. From this result, Óptima correctly inferred that the target was likely unattainable with the available reagents and recommended changing the formulation. In a five-day multi-objective flow-chemistry campaign, Óptima increased the yield from 30% to 59% over 23 experiments. Despite substantial inference costs, it cost less and used substantially less starting material than a human-directed campaign, while selecting a more mass-efficient operating point. These results show that LLM-based agents can make rigorous, long-running optimization campaigns accessible to domain scientists without specialist setup, expanding the scope of SDLs.
[MA-7] Game-Theoretic Drone Swarm Defense: A Case Study in Applied Differential Game Theory
【速读】：该论文旨在解决无人机蜂群在防御高价值资产时，如何有效分配目标并规划中段制导策略以拦截敌方蜂群的问题。传统方法将防御问题建模为仅优化防御方行为的单边优化问题，忽略了敌方蜂群的理性对抗行为。本文提出的解决方案基于微分博弈（Differential Game, DG）理论，将入侵蜂群视为具有理性的博弈对手，通过求解防御方与入侵方之间的纳什均衡（Nash Equilibrium），实现更优的协同拦截策略。其关键在于引入博弈论框架，使防御策略能够前瞻性地预测并应对敌方的机动意图，尤其在敌方具备规避机动能力时，显著提升了拦截成功率。仿真结果表明，相较于基线优化方法，微分博弈策略将防御成功概率从94.6%提升至96.8%，缩小了约41%的与理想防御之间的差距；并通过配对试验贝叶斯分析，以99.9%的后验概率证实该方法在该场景下具有更高的防御有效性。
链接: https://arxiv.org/abs/2609.04394
作者: Ross E. Allen
机构: MIT Lincoln Laboratory(麻省理工学院林肯实验室)
类目: Computer Science and Game Theory (cs.GT); Multiagent Systems (cs.MA); Robotics (cs.RO)
备注:
Abstract:This technical report is a study of the use of differential game (DG) theory to solve the target-assignment and midcourse guidance problems of drone swarms tasked with intercepting opposing swarms in defense of high-value assets. The game-theoretic tactics—which treat the intruder swarm as a rational agent and seek a Nash equilibrium between defenders and intruders—are compared against baseline tactics that model the defense problem as a unilateral optimization of the defenders’ maneuvers. Monte Carlo simulation and Bayesian analysis show that the game-theoretic approach has a higher probability of successfully intercepting all intruders than the baseline techniques. This improvement in successful defense probability is most pronounced when the intruder swarm is capable of evasive maneuvers: relative to baseline optimization tactics, differential-game tactics increase estimated defense success from 94.6% to 96.8%, closing approximately 41% of the remaining gap to perfect defense. To add statistical credibility to this result, a paired-trial Bayesian analysis assigns a 99.9% posterior probability that differential-game tactics have a higher probability of successful asset defense than baseline tactics in this scenario.
[MA-8] Abstraction Agent
【速读】：该论文旨在解决大规模不完美信息博弈中策略求解算法可扩展性受限的核心问题，即传统信息抽象（Information Abstraction）方法依赖领域特定的评估器（如手牌强度计算器或权益估算器），需大量专家知识与工程投入，且难以适用于多数研究较少的游戏。其解决方案的关键在于提出一种零样本（zero-shot）抽象代理（Abstraction Agent）框架，利用大语言模型（Large Language Model, LLM）从自然语言描述的游戏规则中自动发现连续的战略特征，对私有状态进行评分并聚类生成抽象桶（abstraction buckets），全程无需任何游戏特异性评估器、训练数据或博弈树遍历。该框架包含四个阶段：带校准锚点的特征发现、批量私有状态评分、基于相关性的特征选择以及k-means聚类。实验表明，该方法在两人无限制德州扑克翻牌后回合（HUNL turn endgames）上相较预期手牌强度基线降低高达62%的提升策略可被剥削性，在ROVER Trials这一未在预训练语料中出现的新游戏中亦在所有粒度上超越标量排名基线；此外，该方法在四张牌限注奥马哈、翻前与翻牌前的HUNL及立直麻将等多类游戏中均能识别出符合公认战略概念的显式特征，实现了将大语言模型参数中隐含的战略知识转化为可计算的显式数值特征的结构化知识提取（structured knowledge elicitation）。
链接: https://arxiv.org/abs/2609.04303
作者: Boning 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: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Science and Game Theory (cs.GT)
备注:
Abstract:Information abstraction, which groups strategically similar private states into a tractable number of buckets, is essential for scaling game-solving algorithms to large imperfect-information games. Constructing effective abstractions, however, has traditionally required domain-specific evaluators such as hand-strength calculators or equity estimators, which demand expert knowledge and engineering effort and are unavailable for most less-studied games. We propose the Abstraction Agent, a zero-shot pipeline that uses a large language model (LLM) to discover continuous strategic features from a natural-language game description, score private states on these features, and cluster them into abstraction buckets, without any game-specific evaluator, training data, or game-tree traversal during abstraction construction. The pipeline runs in four phases: feature discovery with calibration anchors, batched private-state scoring, correlation-based feature selection, and k -means clustering. The resulting abstractions reduce lifted-strategy exploitability by up to 62% relative to an expected-hand-strength baseline on heads-up no-limit Texas hold’em (HUNL) turn endgames, and beat a scalar rank baseline at every granularity on ROVER Trials, an original game absent from any pretraining corpus. Beyond these quantitative benchmarks, the pipeline transfers with unchanged prompts to four-card Pot-Limit Omaha, HUNL preflop and flop, and Riichi Mahjong, where the discovered features track each game’s recognized strategic concepts. This is structured knowledge elicitation: converting implicit strategic knowledge in LLM parameters into explicit numerical features for downstream algorithmic computation. The code is available at this https URL.
[MA-9] EVOHARNESSBENCH: Can Your Agents Keep Pace with an Evolving Harness?
【速读】：该论文旨在解决大语言模型（Large Language Model, LLM）驱动的智能体在动态演进的工具-技能-代理（harness）体系中所面临的持续学习挑战，即当外部提供的能力集合（包括工具、可复用技能和专用代理）随时间不断演化时，智能体如何在保持已有能力的同时有效适应新引入的能力。其解决方案的关键在于提出EVOHARNESSBENCH——一个在工具、技能和代理三个维度上可控演化的基准测试框架，通过17条多阶段的确定性演进流（共802个任务、520个工具、42种技能、62个代理）模拟真实场景中的非平稳性（non-stationarity），并区分两种核心评估范式：部署评估（deployment evaluation）以检验智能体在拓展能力过程中对过往能力的保留性能，以及自演化适应评估（self-evolving adaptation evaluation）以测试已有经验在新增能力背景下的可用性。研究揭示了三大关键问题：（1）单纯扩充能力集可能导致对已解决任务性能下降，引发“工具诱导遗忘”；（2）自演化适应效果在不同演进阶段、能力轴和环境间表现不一致；（3）能力保留与适应新能力之间存在内在张力，二者难以协同优化。这些发现表明，针对动态演进的工具-技能-代理架构进行鲁棒性建模是构建可持续演进智能体的核心挑战。
链接: https://arxiv.org/abs/2609.04280
作者: Zixuan Ke,Vaidehi Patil,Haizhou Shi,Yang Li,Ye Liu,Sarath Shekkizhar,Anurag Koul,Jiayu Wang,Xuan Phi Nguyen,Semih Yavuz,Mohit Bansal,Shafiq Joty
机构: Salesforce Research; University of North Carolina at Chapel Hill (北卡罗来纳大学教堂山分校); University of Wisconsin–Madison
类目: Multiagent Systems (cs.MA); Computation and Language (cs.CL)
备注: this https URL
Abstract:Modern LLM-based agents operate through a harness of tools, reusable skills, and specialist agents that shapes what they observe and what they can do. In practice, this harness continually evolves as new capabilities are added. We introduce EVOHARNESSBENCH, a benchmark for evaluating agents under controlled harness evolution across three axes (tools, skills, and agents). Unlike existing continual-learning benchmarks for agents, which typically place non-stationarity (i.e., what changes over time) in the task stream while keeping the harness fixed, EVOHARNESSBENCH places non-stationarity in the externally supplied harness itself. It contains 17 multi-stage harness streams constructed deterministically from verifier-based benchmarks, comprising 802 tasks, 520 tools, 42 skills, and 62 agents. We evaluate two complementary settings corresponding to the central challenges of harness evolution: deployment evaluation, which isolates retention of previously accessible competence as the harness expands, and self-evolving adaptation evaluation, which tests whether accumulated experience remains useful as new capabilities are introduced. Our results reveal three persistent gaps. First, harness expansion alone can degrade performance on previously solved tasks, producing harness-induced forgetting. Second, gains from self-evolving adaptation remain inconsistent across stages of harness evolution, capability axes, and environments. Third, retention and adaptation can pull in different directions: preserving earlier competence does not necessarily improve adaptation to newly introduced capabilities, and vice versa. These results establish harness evolution as a distinct challenge for building agents that can keep pace with an evolving harness while preserving previously effective behavior.
[MA-10] Reviewer Capability Governs Rejection Targeting Not Repair Skill: Evidence from LLM Execute-Review-Revise Pipelines
【速读】：该论文旨在解决多智能体大语言模型（Multi-agent LLM）流水线中，验证阶段（verification stage）在不同能力层级模型作为评审者时的效能问题。传统方法通常假设评审者的性能固定于执行者（executor）之上，但本文通过系统性地引入从高到低跨能力层级的评审模型（包括一个无法解决任何问题的弱模型），探究其对最终答案准确率的影响。研究的关键发现是：使用跨家族的中等能力评审模型可使最终准确率提升12个百分点（从52%升至64%，p = 0.0005），且无任何错误修正导致的副作用；而同模型自审虽具有最高的错误检测召回率（0.85），但因误拒率高达35%（远高于跨家族评审的2%），且修复效率仅为后者的三分之一，整体上未能带来显著收益。此外，当评审模型能力低于某一阈值时，其角色完全失效，仅增加计算开销而不改变结果。因此，解决方案的核心在于：选择具备适当能力下限的跨家族中等水平评审模型，以实现成本与效果之间的最优平衡，而非简单依赖高阶模型或自审机制。
链接: https://arxiv.org/abs/2609.04270
作者: Faizan Tanveer
机构: National University of Computer and Emerging Sciences (FAST NUCES)
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL); Multiagent Systems (cs.MA)
备注: 6 pages, 5 tables. Code and data: this https URL
Abstract:Multi-agent LLM pipelines increasingly assign roles, including execution and verification, to models of different capability tiers. This is done because running a flagship model at every stage is expensive. Previous literature has established that verification stages are not always beneficial, but holds reviewer capability roughly fixed relative to the executor. We vary it. We replace the reviewer with models spanning a capability range down to one that cannot solve the problems at all, and measure the outcome of every individual rejection. This is done across a constant set of 100 olympiad mathematics problems. A cross-family mid-tier reviewer improves final accuracy by 12 percentage points, from 52 to 64 percent (p = 0.0005), with zero damaged answers. Same-model self-review attains the highest error-detection rate of any condition (0.85 recall) yet yields no significant gain: it rejects 2.1 times as often for a third the repair rate (15 against 43 percent, p = 0.0074) and falsely rejects 35 percent of its own correct answers against 2 percent for the cross-family reviewer (paired p = 0.000015). The low damage rate of self-review proves to be an artifact of revision inertia rather than reviewer quality: of 18 falsely rejected correct answers, the three where the executor complied all became wrong, while the fifteen it ignored survived unchanged. Below a capability floor the role becomes inert: our weakest reviewer changed zero of 100 final answers while doubling token cost. These findings describe a single executor-reviewer configuration on 100 problems and should be read as a controlled pilot rather than a general claim about verification stages. Comments: 6 pages, 5 tables. Code and data: this https URL Subjects: Software Engineering (cs.SE); Computation and Language (cs.CL); Multiagent Systems (cs.MA) ACMclasses: I.2.11 Cite as: arXiv:2609.04270 [cs.SE] (or arXiv:2609.04270v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2609.04270 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[MA-11] Multi-dimensional Bias in Modeling Multi-dimensional Preferences: Evaluating the Ability of Synthetic Agents to Replace Human Participants in Conjoint Experiments
【速读】：该论文旨在解决生成式人工智能（Generative AI）在联合分析（conjoint analysis）设计中替代人类受访者以提升实验鲁棒性或降低数据收集成本的可行性问题，尤其针对政治科学领域日益流行的联合分析方法。其核心解决方案的关键在于评估由合成代理（synthetic agents）生成的选择行为是否能够准确复现真实人类在多维偏好模式上的复杂特征。研究通过复制已发表的联合分析研究，从表征对应性（representational correspondence）、推断对应性（inferential correspondence）和程序稳定性（procedural stability）三个维度，对比合成代理与原始人类数据在选择分布、统计估计与实质性结论之间的相似性。结果显示，不同维度及研究间的表现不一致，表明合成参与者的有效性具有依赖具体主张（claim-dependent）和层级化（hierarchical）特征。仅实现图表再现或显著符号一致性，不足以支持用合成样本完全替代人类样本。因此，该研究强调，在将合成代理视为可靠替代方案之前，学界必须首先系统界定其在不同层次上的适用边界。
链接: https://arxiv.org/abs/2609.04243
作者: Ho Ting Hung,Nachiket Midha,Victor Y. Wu,Yiwen Zhang
机构: University of Oxford (牛津大学); Oxford Computational Political Science Group (牛津计算政治科学小组); Stanford University (斯坦福大学); LSE (伦敦政治经济学院)
类目: Multiagent Systems (cs.MA); Computers and Society (cs.CY); Methodology (stat.ME)
备注:
Abstract:Despite growing interest in using LLMs to add robustness or reduce data-collection costs in survey experiments, their efficacy in conjoint design—an increasingly popular method in political science—remains underexplored. This paper addresses that gap by investigating whether synthetic agents can reproduce the multi-dimensional human preference patterns that conjoint is designed to capture. It replicates published conjoint studies and compares the results generated by synthetic agents with original human data along three dimensions: representational correspondence, inferential correspondence, and procedural stability. Our analysis evaluates the alignment of choice distributions as well as the statistical and substantive similarity of estimates, and the results are uneven across these dimensions and studies replicated. This implies that the validity of synthetic participants should be considered claim-dependent and hierarchical. Reproducing a figure or obtaining strong sign agreement is evidence of similar aggregate outputs, but not enough to support replacing human respondents. Our results suggest that the discipline as a whole must first map this innovation’s boundaries across various levels before considering synthetic agents a robust substitute for human samples.
[MA-12] At Equal Inference Cost Multi-Agent Structure Does Not Beat a Single Frozen Agent
【速读】：该论文旨在解决多智能体大语言模型（multi-agent LLM）流水线在实际应用中存在推理成本过高却未显著提升性能的问题。现有方法通常在相同环境回合数下比较多智能体团队与单智能体，导致团队享有额外计算资源，无法公平评估其真实增益。本文提出在固定语言模型调用总次数的约束下，对比演化后的多智能体团队与单智能体的表现，以更公平地评估结构优势。其解决方案的关键在于引入MA-Evolve框架，将规划-执行-批评（Planner-Executor-Critic）团队表示为三个可演化的角色提示（role prompts），并通过基于共享冻结7B主干模型的分角色坐标上升优化策略进行联合演化。实验结果表明，在ALFWorld任务中，尽管完整团队取得最高平均性能（0.769），但与单执行器演化后的表现（0.754）无统计显著差异（p = 0.80），且前者使用1.8倍于后者的调用次数；进一步的“留一分析”显示，性能提升完全源自执行器的演化，规划器和批评者演化为无意义或低影响提示，极少干预执行动作。在WebShop任务中，演化对单智能体无效，而团队反而表现更差。因此，在等量推理开销条件下，多智能体结构并未带来明确收益，反而增加了计算负担。
链接: https://arxiv.org/abs/2609.04217
作者: David Dylan,Aoife Brennan,Cian Murphy,Niamh O’Sullivan,Conor Kelly,Saoirse Walsh
机构: Trinity College Dublin; University College Dublin; Dublin City University (都柏林城市大学)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Multi-agent LLM pipelines, such as Planner-Executor-Critic teams, often report gains over single agents, but these gains usually come with higher inference cost because the team makes multiple model calls per environment step. Existing automated methods search over roles, topologies, and prompts, but typically compare teams against single agents at equal environment rollouts, giving the team extra compute. We instead fix the total number of language-model calls and ask whether evolving a multi-agent team still beats evolving a single agent under the same budget. We introduce MA-Evolve, which represents a Planner-Executor-Critic team as three evolvable role prompts and optimizes them by per-role coordinate ascent over a shared frozen 7B backbone. On ALFWorld, evolving a single executor significantly improves over the unevolved agent, while the full team achieves the highest mean but is not statistically better than the single agent: 0.769 versus 0.754, p = 0.80, despite using 1.8 times more evaluation calls. Leave-one-in analysis shows that the realized value comes entirely from the executor; the planner and critic evolve to empty or low-impact prompts and rarely change the executor’s action. With 2-3 times free compute, the team only matches the single agent, and on WebShop evolution is null while the team trends worse. Under equal inference cost, multi-agent structure adds cost without clear benefit. Subjects: Multiagent Systems (cs.MA) Cite as: arXiv:2609.04217 [cs.MA] (or arXiv:2609.04217v1 [cs.MA] for this version) https://doi.org/10.48550/arXiv.2609.04217 Focus to learn more arXiv-issued DOI via DataCite
[MA-13] Mitigating Disease Spread by Design in Refugee and IDP Camps
【速读】：该论文旨在解决难民及境内流离失所者（IDP）安置点中疾病传播的日益严峻挑战，重点关注营地布局对人员流动与互动模式的影响，进而影响疾病传播路径的问题。其核心在于探究不同营地空间布局作为潜在缓解因素，在控制传染病扩散中的作用。解决方案的关键在于构建一种基于代理（agent-based）建模的方法学框架，利用JUNE代理模型进行概念验证实验，通过模拟不同布局下的人群行为与疾病传播动态，为营地规划提供可操作的科学依据，并为后续研究提供可扩展的分析范式。
链接: https://arxiv.org/abs/2609.05342
作者: Giulia Zarpellon,Joseph Aylett-Bullock,Frank Krauss,Miguel Luengo-Oroz
机构: United Nations Global Pulse; Institute of Data Science, Durham University (达勒姆大学数据科学研究所)
类目: Physics and Society (physics.soc-ph); Computers and Society (cs.CY); Multiagent Systems (cs.MA)
备注: 9 pages, 9 figures
Abstract:Disease spread represents an increasing challenge in refugee and internally displaced person (IDP) settlements. The movement and interaction of people within camps is influenced by their layout, which therefore has the potential to significantly affect disease spread. This work aims at creating a methodology to explore the potential effects of different camp layouts as mitigating factors in the spread of diseases within settlements. We showcase proof-of-concept experiments by leveraging the JUNE agent-based epidemic model, discuss the kind of operational insights this methodology can facilitate, and provide a framework for future investigations.
自然语言处理
[NLP-0] WearableQA: A Benchmark for Health Reasoning over Real-World Wearable Data
【速读】： 该论文旨在解决当前人工智能系统在真实用户长期可穿戴设备数据上的推理能力评估不足的问题，尤其关注生成式模型是否能够基于连续的生理与行为信号进行深层次理解。其核心挑战在于现有基准难以反映真实世界中可穿戴数据的复杂性，如设备噪声、个体差异以及多源信号间的动态关联。为此，研究提出WearableQA这一基准，涵盖200名真实用户的长达500天的可穿戴时间序列数据、血液生物标志物及人口统计学信息，构建了4,084道10选1的多项选择题。关键创新在于采用双基框架（dual-grounding framework），结合文献支持的生理学知识与基于人群统计验证的生理模式，实现大规模、高可信度的问题生成。同时，通过16种问题类型，沿“数据推理”与“健康推理”、“单信号”与“跨信号”两个互补维度设计，全面评估模型的多层级推理能力。实验表明，尽管主流大语言模型（LLM）在该基准上表现各异（准确率19.6%至72.9%，远超10%随机基线），但整体性能仍普遍低于60%，说明该任务极具挑战性。因此，WearableQA不仅提供了真实、诊断性强的评估平台，也为未来可穿戴智能系统的认知能力发展指明了方向。
链接: https://arxiv.org/abs/2609.05405
作者: Ji Soo Lee,Xilun Chen,Pierce Chuang,Ashish Shenoy,Jason Wei,Dohwan Ko,Hyunwoo J. Kim,Benoit Corda
机构: Meta; KAIST(韩国科学技术院); Korea University(韩国高丽大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Recent advances in wearable sensing enable continuous monitoring of physiological and behavioral signals, yet existing benchmarks rarely evaluate whether AI systems can reason over a real user’s longitudinal wearable record. We introduce WearableQA, a benchmark comprising 4,084 10-option multiple-choice questions constructed from the wearable time series, blood biomarkers, and demographics of 200 real users, each with up to 500 days of daily measurements. WearableQA preserves authentic wearable distributions that include device noise and inter-individual variability. To evaluate distinct reasoning capabilities, we introduce 16 question types organized along two complementary axes: data versus health reasoning, which distinguishes computation over longitudinal measurements from physiological interpretation; and single- versus cross-signal reasoning, which separates reasoning about individual signals from the integration of multiple signals. To construct reliable questions at scale, we adopt a dual-grounding framework that combines literature-grounded physiological findings with statistically validated population-grounded physiological patterns. This enables the capture of meaningful relationships observed in real-world wearable data. Evaluation of 14 proprietary and open-source LLMs demonstrates that WearableQA effectively differentiates model capabilities, with performance ranging from 19.6% to 72.9% against a 10% chance baseline. Moreover, WearableQA remains far from solved: most models achieve accuracies below 60%. Overall, WearableQA provides a realistic and diagnostic benchmark for evaluating LLM reasoning over real-world wearable data.
[NLP-1] Same Trajectory Contradictory Rewards (ROBORMBENCH): Paraphrase Frag ility in Vision Language Reward Models
【速读】： 该论文旨在解决视觉-语言模型（Vision-Language Model, VLM）作为机器人学习奖励函数时存在的语义等价性不鲁棒问题，即在语义相同但表述不同的目标指令下，模型对同一机器人轨迹的奖励评分应保持一致（即具备重述不变性，paraphrase invariance）。然而，现有VLM奖励模型普遍存在重述敏感性，仅通过改写指令即可显著改变预测进展分数，甚至将相同的机器人行为错误判定为成功与失败。为此，作者提出了ROBORMBENCH基准，包含2,390条真实机器人轨迹、真实进展标签及21,673个经过验证的重述样本，覆盖词汇、句法和动作-目标层面的改写形式。实验表明，无论开源或专有VLM，重述引发的奖励不稳定性普遍存在且严重，且随着改写差异增大而加剧，单纯扩大模型规模或引入显式推理无法可靠缓解此问题。相比之下，基于轨迹-真实标签监督训练的专用奖励模型表现出显著更高的稳定性。研究结论强调，重述鲁棒性是实现可靠VLM驱动机器人奖励建模的核心要求。
链接: https://arxiv.org/abs/2609.05401
作者: Wonje Jeung,Sangyeon Yoon,Hyesoo Hong,Yoonjun Cho,Dongjae Jeon,Bumjun Kim,Jean Oh,Youngjae Yu,Albert No
机构: Yonsei University(延世大学); Carnegie Mellon University(卡内基梅隆大学); Seoul National University(首尔国立大学)
类目: Robotics (cs.RO); Computation and Language (cs.CL)
备注:
Abstract:Vision-language models are increasingly used as reward functions for robotic learning, but this role requires paraphrase invariance: the same trajectory should receive the same reward under semantically equivalent goal descriptions. We show that current VLM reward models often violate this property. Paraphrasing the instruction alone can substantially change predicted progress scores, and can even flip identical robot behavior between failure and success. To measure this failure mode, we introduce ROBORMBENCH, a benchmark with 2,390 real-robot trajectories, ground-truth progress labels, and 21,673 verified paraphrases spanning lexical, syntactic, and action-goal rewrites. Across proprietary and open-source VLMs, paraphrase-induced instability is widespread and severe, grows under more divergent rewrites, and is not reliably reduced by scale or explicit reasoning. Dedicated reward models trained with trajectory-grounded supervision are substantially more stable. These results show that paraphrase robustness is a core requirement for reliable VLM-based reward modeling in robotics.
[NLP-2] Multi-Step Tool-Calling over Korean Open Public APIs: A Benchmark and a Data-Synthesis Recipe EMNLP2026
【速读】： 该论文旨在解决开放源代码大语言模型（Large Language Model, LLM）在多步骤任务中性能不足的问题，尤其是在需要调用多个实时政府应用程序接口（API）的场景下，由于数据主权法规要求公共机构部署本地化、开源的LLM代理，这一问题尤为突出。现有基准测试无法有效衡量开源模型在多工具调用任务中的性能差距。为此，论文提出韩国开放公共API基准测试（Korean Open Public API Benchmark, KOPA-Bench），包含145个真实世界任务，以量化评估模型表现。其核心解决方案为EDGE（Execution-grounded Dynamic Graph for tool-calling data synthEsis），该方法基于实际执行结果构建动态图结构，仅保留真实调用中成功传递的工具输出-输入连接，并通过遍历这些经验证的链接生成可执行的多步轨迹。利用生成式强化学习优化（GRPO）对9B参数量模型进行微调后，该模型性能接近未微调的27B同系列模型，在KOPA-Bench和BFCL基准上均实现显著提升，表明其在真实环境下的高效工具链协作能力。
链接: https://arxiv.org/abs/2609.05395
作者: Dain Kim,Eungi Cho,Kyumin Kim,Shinyeong Noh,Kyuseong Lim
机构: LG CNS
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 30 pages, 7 figures, 26 tables. Accepted to EMNLP 2026 Industry Track
Abstract:Data-sovereignty regulations increasingly require public institutions to deploy open-source, on-premise LLM agents that chain multiple tool-calls across live government APIs. However, open-source models consistently underperform in this multi-step setting, and no existing benchmark measures the gap. We introduce the Korean Open Public API Benchmark (KOPA-Bench), comprising 145 real-world tasks. To close this gap, we present EDGE, an Execution-grounded Dynamic Graph for tool-calling data synthEsis driven by live execution. EDGE builds a graph of how each tool’s output can feed another’s input, keeps only the links that succeed when actually called against the live APIs, and traverses these verified links to synthesize executable multi-step trajectories. Fine-tuned via GRPO on the resulting dataset, our 9B model nearly matches the untuned 27B model from the same family, improving substantially not only on KOPA-Bench but also on the BFCL benchmark.
[NLP-3] chnical Manual for a Toolkit for Measuring Contextual Individuation in Transformer Language Models
【速读】： 该论文旨在解决生成式 AI 模型中词义消歧的可解释性难题，即如何在不依赖外部标注的前提下，精确评估变换器语言模型（Transformer Language Model）是否真正通过上下文对同一词形的不同语义实例进行区分。其核心挑战在于：尽管模型在嵌入层为词形分配单一向量，但普遍认为其后续层能根据上下文“个体化”该词的不同使用实例。为实现对这一信念的清晰验证，需构建一种控制变量的实验构造——即保持词形不变而系统性地改变其上下文与目标语义。为此，作者提出并实现了一个名为“桥式词形”（bridge form）的开放工具包，其关键设计在于定义一个在多个主题领域中重复出现、形式不变但语义各异的词形作为分析单元。解决方案的核心在于构建一套端到端的标准化处理流程，涵盖桥式词形的声明式定义、维基百科语料获取、词例定位、逐层表示提取、基于领域配对轮廓系数（silhouette measurement）的表示空间分离度量化，以及成对可视化协议。每一环节均针对特定方法学缺陷进行优化：避免类别标签过宽导致的语义污染、轮廓系数的多组偏差、子词分词错位、降维后坐标可比性失真等问题。该工作本质上是一份方法论与实现参考，不报告具体实证结果，而是提供可复现的工具链与数据集，供后续研究作为分析仪器引用和使用。
链接: https://arxiv.org/abs/2609.05333
作者: José Luciano Verçosa Marques,Frederico Jorge Heitmann,Daniel Omar Perez,Marcelo Vinicius de Paula,Tárcio André dos Santos Barros
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 29 pages, 2 figures (one with 2 subfigures), 1 table. Toolkit, source code, and corpora archived separately on Zenodo (see Section 9)
Abstract:A transformer language model assigns a single, context-independent vector to a word type at its embedding layer, yet is widely believed to individuate that word’s occurrences by context in its later layers. Testing this belief cleanly requires a construct that holds the word form fixed while its context and intended sense vary in a controlled, labeled way. This manual documents an open toolkit built around such a construct, which we call a bridge form: a single written word that recurs, unchanged, across two or more subject domains with a different sense in each. We describe, and justify, every stage of the pipeline: the declarative specification of bridge forms and their source domains, corpus acquisition from Wikipedia, occurrence localization, layer-wise representation extraction, a domain-pairwise silhouette measurement of separation in the model’s representation space, and a paired visualization protocol. Each design choice is presented together with the methodological failure mode it is meant to avoid (sense contamination from overly broad category labels, the multi-group bias of the silhouette coefficient, subword-tokenization misalignment, and axis-comparability artifacts in dimensionality-reduced plots, among others). This manuscript is a methodological and implementation reference: it does not report or interpret empirical outcomes of running the toolkit on any particular model or bridge-form set. The toolkit, its full source, and the corpora used to exercise it are archived separately (Section 9) under a persistent identifier, and are intended to be cited as an instrument by studies that use it to produce and interpret empirical results.
[NLP-4] Large Language Models for HVAC Operations in Building Energy Systems: A Critical Review of Methods Applications and Deployment Readiness
【速读】： 该论文旨在解决建筑自动化系统（Building Automation Systems, BAS）在实际运营中存在“数据丰富但洞察匮乏”的核心问题，其根源在于传感器点位命名异构、元数据缺失以及文档碎片化，导致难以有效利用海量传感器数据。解决方案的关键在于评估大型语言模型（Large Language Models, LLMs）在暖通空调（HVAC）运维中的应用潜力与局限性。研究通过系统性综述2023年至2026年3月间发表的66篇相关文献，将其分类至五类应用场景与三类LLM方法，并从证据真实性、部署成熟度及人机责任边界三个维度进行评估。结果表明，当前绝大多数研究仍处于科研阶段（63项为研究级），仅有四项达到试点级别，无一实现持续运行部署。尽管如此，若干受限的、需人工介入的应用场景（如点位名称规范化、基于文档的运维支持、建筑能源建模（BEM）流程辅助及物理控制器周边的咨询型界面）具备近中期试验价值。相比之下，传统机器学习（ML）、模型预测控制（MPC）、强化学习（RL）以及基于本体的工具在高频控制、短时序数值预测和结构化映射任务中更具实用性；而完全自主的代理式操作和未经验证的使用者代理模型仍处于研究阶段。因此，当前证据支持将LLMs主要作为语义理解与工作流协同层，而非独立的自主式HVAC控制器。未来研究应优先推进实地验证基准建设、在真实运行约束下的系统编排评估，以及具备有限延迟与可验证安全性的LLM-MPC/RL融合架构设计。
链接: https://arxiv.org/abs/2609.05314
作者: Alexander Neubauer,Tianzhen Hong,Han Li,Mengbo Yu,Amin Darbandi,Yannick Fürst,Martin Kriegel
机构: Technische Universität Berlin (柏林工业大学); Lawrence Berkeley National Laboratory (劳伦斯伯克利国家实验室)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Systems and Control (eess.SY)
备注: 38 pages, 9 figures, 16 tables. Submitted to Energy and Buildings
Abstract:Building automation systems generate rich sensor data yet remain insight-poor because heterogeneous point naming, missing metadata, and fragmented documentation obstruct their operational use. This systematic review analyses and codes 66 peer-reviewed studies on large language models (LLMs) for HVAC operations published between 2023 and March 2026. Each study is classified across five application families and three LLM method families and assessed for evidence realism, deployment readiness, and the responsibility boundary between the LLM and physical HVAC decisions. The corpus is concentrated in building energy modelling (BEM, 32 of 66 papers), while load forecasting remains too sparse for subfield-level conclusions. Only four studies reach pilot-level evidence, and none reports sustained operational deployment. No study was classified as ready-now for industry adoption; three were near-term and 63 research-only. Nevertheless, several bounded, human-in-the-loop uses merit near-term trials, including point-name normalisation, document-grounded operator support, BEM workflow assistance, and advisory interfaces around physics-based controllers. Conventional machine learning (ML), model predictive control (MPC), reinforcement learning (RL) and ontology-based tools remain more adopted for high-frequency control, short-horizon numerical forecasting, and well-posed ontology mapping, while autonomous agentic operation and unvalidated occupant proxies remain research-stage. Current evidence therefore supports LLMs primarily as semantic and workflow layers rather than autonomous HVAC controllers. Future work should prioritise field-validated benchmarks, orchestration evaluation under operational constraints, and LLM-MPC/RL architectures with bounded latency and verifiable safety properties.
[NLP-5] LexFlip: A Dissociation Diagnostic for Legal Meaning Preservation Metrics
【速读】： 该论文旨在解决法律条款在简化过程中是否仍能保持其原有法律效力的问题，即如何有效评估简化后的法律文本与原始文本在法律内涵上的一致性。当前主流的评估方法（如要求“完全相同”对得分最高、“无关”对得分最低）存在根本性缺陷：它们将词汇重叠度与法律效力捆绑，导致任何单调依赖于词元重叠度的指标均可能通过检验，从而无法真正区分语义等效与形式相似。为此，论文提出关键解决方案——“解耦”（dissociation），即设计一种在保持表面形式不变的前提下，仅改变法律效力的文本扰动范式。研究构建了名为LexFlip的数据集，包含373个对魁北克法典法语条文的微小修改，这些修改在保留93%词元的基础上反转了法律效力。通过引入多类评分指标、回归模型及人工判别者进行验证，发现现有7种嵌入式与BERTScore指标在该任务上的响应范围极窄（仅0.022–0.039），远低于双向自然语言推理（bidirectional NLI）的0.670；同时，在人工标注基准（FrJudge）上，仅长度特征便超越所有语义指标，且误差最小。这表明，当前主流语义相似性度量难以捕捉法律效力这一深层属性，而解耦范式为评估法律文本简化提供了更可靠的基准。
链接: https://arxiv.org/abs/2609.05296
作者: Gaurab Baral
机构: University of Cincinnati(辛辛那提大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 6 pages, 2 figures, 3 tables
Abstract:Does a simplified legal clause still say what the original said? The checks in current use cannot establish that it does: requiring an identical pair to score highest and an unrelated pair lowest moves lexical overlap and legal force together, so any monotone function of token overlap satisfies both. Our remedy is a dissociation, an item holding surface form fixed while legal force moves. We release LexFlip, 373 minimal perturbations of Quebec statutory French that reverse legal force while preserving 0.93 of the tokens, with a harness scoring metrics, regressors and prompted judges alike. The seven embedding and BERTScore metrics we test spend only 0.022 to 0.039 of their identical-to-unrelated range on such an edit, against 0.670 for bidirectional NLI, the one family the identical-pair check would disqualify. On FrJudge, against a measured human ceiling of r=0.597, a bare length feature outscores every semantic metric and has the lowest margin we measure.
[NLP-6] Self-Supervised Lexical Representation Learning for Fast Large-Scale Phylogenetic Inference
【速读】： 该论文旨在解决计算谱系学在全局尺度应用中的两大瓶颈：一是基于字符的方法需要大量人工标注的同源性判断，耗时费力；二是大规模数据集上的推断计算成本过高。其解决方案的关键在于提出一种完全自监督的对比学习框架，能够直接从原始国际音标（IPA）转写的词表中学习词汇表示，无需同源性标注、对齐或额外专家输入。该模型采用双重对比目标：词级损失使语音相似的形式在嵌入空间中聚集，语言级辅助损失则促使词汇空间反映语言层面的更广泛音系特征。基于学习到的词向量计算成对语言距离，进而推断出包含3,399种语言变体的全球谱系树。实验表明，该方法在通用四分位距离（GQD）指标上达到与多个基线相当的性能，且仅需标准笔记本GPU几分钟计算时间。此外，同一表示还能捕捉历时概念稳定性：跨语言成对距离的方差可生成显著相关于公认排名的概念稳定性排序。消融实验验证了语言级目标和音素特征向量的引入均有效提升了谱系树拓扑结构的准确性。因此，该框架为大规模谱系推断提供了高效、全自动的替代方案，并构建了一个统一的表示体系，支持语言与概念层面的下游分析。
链接: https://arxiv.org/abs/2609.05262
作者: Tim Wientzek
机构: 未知
类目: Computation and Language (cs.CL)
备注: 27 pages, 3 figures
Abstract:Computational phylogenetics has become an essential tool in historical linguistics, yet its application at a global scale remains constrained by two factors: the labor-intensive manual annotation of cognacy judgments required for character-based methods and the substantial computational cost of inference on large datasets. This paper introduces a fully self-supervised contrastive learning framework that learns lexical representations directly from raw IPA-transcribed wordlists, without requiring cognacy annotations, alignments, or additional expert input. The model employs a dual contrastive objective: a word-level loss that organizes phonetically similar forms into a coherent space, and an auxiliary language-level loss that encourages the lexical space to reflect broader phonological properties of languages. From the resulting word representations, pairwise language distances are derived and used to infer a global phylogenetic tree of 3,399 language varieties. The inferred tree achieves a generalized quartet distance (GQD) to the Glottolog reference tree competitive with multiple baselines, while requiring only minutes of computation on a standard notebook GPU. Furthermore, the same representations capture diachronic concept stability: variance in pairwise distances across languages yields stability rankings that correlate significantly with established rankings. Ablation studies confirm that both the language-level objective and the use of phonetic feature vectors improved the inferred trees topology with regards to GQD. The framework thus provides a computationally efficient and fully automatic alternative for large-scale phylogenetic inference and offers a unified representation supporting downstream analyses at both the language and concept level.
[NLP-7] A Verifier-Guided Explainable Reasoning Framework with Gold-Anchored QLoRA Task-Aware Mixture-of-Experts and Group-Relative RLVR
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在教育问答任务中推理解释存在不一致、缺乏坚实依据或难以验证的问题。其核心解决方案是提出一种验证器引导的可解释推理框架，关键在于三方面协同：首先，采用以权威答案为锚点的领域加权QLoRA微调，增强模型对正确答案的对齐；其次，引入任务感知的符号路由机制，将逻辑问题分配给一阶逻辑（FOL）与Z3验证器，物理问题则交由具备公式与单位感知能力的符号求解器处理；最后，利用验证器反馈支持候选响应的评估、自我修正及强化学习中的奖励构建。通过三个互补维度（P1：答案正确性；P2：证据或单位一致性；P3：推理深度与可解释性）进行多维评估，在438个独立测试样本上，基于强化学习的验证反馈（RLVR）使P3指标从50.68%显著提升至72.20%，而混合精度P1保持稳定在约55.94%。此外，自一致性机制进一步提升了仅依赖模型的P1至50.23%，符号验证贡献了剩余增益。结果表明，RLVR主要强化显式推理结构，而符号验证则在系统层面提升答案可靠性，二者共同实现透明且可信的教育问答。
链接: https://arxiv.org/abs/2609.05221
作者: Thi Kim Trang Vo,Nam Tien Le,Thi Kim Nguyet Vo,Minh Khang Tran,Duy Phuong Tran
机构: University of Information Technology (UIT), Ho Chi Minh City, Vietnam; Ho Chi Minh City University of Technology (HCMUT), Vietnam; Vietnam National University, Ho Chi Minh City, Vietnam; University of Economics Ho Chi Minh City (UEH), Vietnam; Viet Nam – The Netherlands Programme (VNP), Vietnam
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Large language models (LLMs) show strong reasoning ability, but their explanations can remain inconsistent, weakly grounded, or difficult to verify. We propose a verifier-guided explainable reasoning framework for transparent educational question answering that combines gold-anchored QLoRA, task-aware symbolic routing, and group-relative RLVR. Qwen2.5-3B-Instruct is first adapted with field-weighted QLoRA supervision anchored to authoritative answers. A lightweight router then assigns logic problems to a FOL/Z3 verifier and physics problems to a formula- and unit aware symbolic solver. Verifier feedback is further used to support candidate evaluation, self-revision, and reward construction during RLVR. Candidate responses are evaluated along three complementary dimensions: P1 for answer correctness, P2 for evidence or unit consistency, and P3 for reasoning depth and explainability. At inference, gold-free self-consistency aggregates multiple candidate responses before an optional question-only physics verifier performs conservative system-level correction. On 438 held-out examples, RLVR increases P3 from 50.68% to 72.20%, while hybrid P1 remains approximately stable at 55.94%. Self-consistency improves model only P1 from 48.86% to 50.23%, with symbolic verification providing the remaining hybrid gain. These results indicate that RLVR primarily strengthens explicit reasoning structure, while symbolic verification complements the neural policy by improving answer reliability at the system level.
[NLP-8] Can Large Language Models Anticipate Behavioral Responses to Social Policies? A Case of Pension Enrollm ent Prediction among Chinas Flexible Workers
【速读】： 该论文旨在解决社会政策变动影响评估中的核心难题：传统计量经济方法在假设情景外推时可靠性不足，而实地试点项目则成本高昂。针对这一问题，本文提出将大语言模型（Large Language Models, LLMs）作为政策评估工具，并构建了首个面向中国灵活就业人员层级化养老金参保预测任务的领域专用模型——FlexPension-LLM。其关键创新在于DKI-RDistill方法：通过在提示词中注入基于政策规则的语义线索（如Probit模型导出的边际效应与户籍-省份养老金制度规则），并结合LoRA/SFT技术，将带有理由增强的监督信号蒸馏至一个开源权重的混合专家（Mixture-of-Experts, MoE）学生模型中；同时，利用真实标签对教师模型的错误样本进行重生成以修正误差。实验结果表明，在CHFS 2019盲测集上，FlexPension-LLM达到0.9316的综合F1分数，优于其Claude Sonnet 4.5教师模型及17个基线中的15个，且与Claude Opus 4.6无统计差异；在四个外部调查数据集上平均综合F1为0.7549，性能波动最小，验证了其强鲁棒性。组件分析揭示，性能提升主要源于政策驱动型提示注入与经误差过滤的监督学习，而模型输出的理由可作为可追溯的决策路径，便于与政策条文进行一致性校验。
链接: https://arxiv.org/abs/2609.05189
作者: Yumiao Li,Peixin Liu,Donglin Di,Chen Li,Runhuan Feng
机构: 未知
类目: Computation and Language (cs.CL)
备注: 16 pages, 9 figures; includes supplementary material. Code and reproduction materials: this https URL
Abstract:Assessing the impacts of social policy changes is a widely acknowledged challenge for policymakers. Econometric methods can be unreliable when extrapolating to hypothetical scenarios, while field pilot programs are highly costly. In this paper, we propose using large language models (LLMs) as policy-assessment tools adapted from general-purpose models. We present FlexPension-LLM, the first domain-specialized large language model for a hierarchical pension-enrollment prediction task among flexible workers in China, and introduce DKI-RDistill, which injects policy-grounded cues into the prompt, including Probit-derived marginal effects and hukou-province pension rules. The method then uses LoRA/SFT to distill rationale-augmented supervision into an open-weight MoE student, with teacher errors corrected by regenerating those cases under ground-truth labels. On a CHFS 2019 blind split, FlexPension-LLM achieves 0.9316 Composite F1, surpassing its Claude Sonnet 4.5 teacher and 15 of 17 baselines, and is statistically indistinguishable from Claude Opus 4.6. Across four external surveys, it averages 0.7549 Composite F1 and shows the narrowest performance range among the strongest systems. Component analysis shows that gains come mainly from policy-grounded cue injection and error-filtered supervision, while rationales provide decision traces that can be checked against policy rules.
[NLP-9] Measuring the Novelty of Biomedical Papers Using the Latent Distances between Knowledge Units
【速读】： 该论文旨在解决科学论文新颖性评估中因过度依赖知识单元共现关系而产生的评价不准确或不完整问题。现有研究多从知识单元的共现角度衡量新颖性，但忽视了知识单元之间更丰富的关联形式，导致对新颖性的刻画存在局限。为此，本文提出一种综合性的新颖性度量方法，关键在于引入三种互补的知识单元关系：网络关系（network）、语义关系（semantic）和层级关系（hierarchical），通过这些关系共同量化知识单元间的潜在距离。基于142,036篇来自PLoS ONE的论文及H1 Connect平台的验证数据集，实证结果表明：（1）三类关系分别捕捉了医学主题词（MeSH）间不同的潜在距离特征；（2）相较于Uzzi等人（2013）提出的经典指标，本文所提度量与同行评审判断具有更强的一致性；（3）融合三类距离度量可显著提升对新颖性论文的识别效果，优于单一视角的评估方式。因此，该研究的核心创新在于构建了一个多维度、整合多种关系结构的新型新颖性评估框架，有效克服了传统方法的片面性。
链接: https://arxiv.org/abs/2609.05175
作者: Yi Zhao,Heng Zhang,Yuzhuo Wang,Wenqing Wu,Tong Bao,Chengzhi Zhang
机构: 未知
类目: Digital Libraries (cs.DL); Computation and Language (cs.CL)
备注:
Abstract:Measuring the novelty of scientific papers is a central concern in research evaluation and scientometrics. From a recombination perspective, prior studies have largely focused on the co-occurrence of knowledge units to assess the novelty of scientific papers. However, these studies often overlook other relationships between knowledge units. This narrow view may result in inaccurate or incomplete evaluations of novelty for scientific papers. To fill this gap, this study introduces a comprehensive novelty measurement that incorporates three types of relationships between knowledge units: network, semantic, and hierarchical. These relationships are used to quantify the latent distances among knowledge units. Using a dataset of 142,036 articles published in PLoS ONE and a validation dataset from the H1 Connect platform, our results demonstrate that (1) each relationship type captures distinct latent distances between MeSH terms; (2) compared to the widely used indicators proposed by Uzzi et al. (2013), our measures show stronger alignment with peer judgements; and (3) combining all three distance metrics yields more effective identification of novel papers than using any single perspective alone.
[NLP-10] Compression Beyond the Uncompressed: A Two-Stage Training Recipe for Soft Context Compression in RAG
【速读】： 该论文旨在解决检索增强生成（Retrieval-Augmented Generation, RAG）系统中因检索到的上下文过长而导致输入冗余、推理效率下降的问题。现有软上下文压缩方法通常通过从未压缩RAG系统中蒸馏输出进行训练，其性能受限于原始模型的表现上限。为此，本文提出DEX-Comp，一种两阶段训练范式：第一阶段“纯蒸馏”仅使用未压缩RAG系统的正确回答对压缩模型进行初始化；第二阶段“硬探索”则在未压缩RAG失败的查询上引入强化学习，促使模型探索更适配压缩表示的计算模式。实验表明，在检索深度为前5至前30的五个开放域问答基准上，DEX-Comp可实现16倍的上下文压缩率，推理速度提升4至24倍，同时在不同检索深度下性能达到或超越未压缩RAG基线。消融实验与跨数据集、多模型架构的评估进一步验证了各阶段的有效性及方法的泛化能力。
链接: https://arxiv.org/abs/2609.05152
作者: Shuyu Guo,Shuo Zhang,Zhaochun Ren
机构: Shandong University(山东大学); Bloomberg(彭博社); Leiden University(莱顿大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Retrieval-Augmented Generation (RAG) enhances language models with external knowledge, but the lengthy retrieved context inflates the input and degrades inference efficiency. Soft context compression encodes each document into a substantially shorter embedding sequence. However, most existing approaches are trained by distilling outputs from uncompressed RAG systems, inherently limiting their performance relative to the original model. To address this limitation, we propose DEX-Comp, a two-stage training recipe: Pure Distillation warm-starts the compression model on the uncompressed RAG’s correct responses only, and Hard Exploration then runs reinforcement learning solely on queries the uncompressed RAG fails, forcing the model to explore computation patterns better suited to compressed representations. On five open-domain QA benchmarks at retrieval depths from top-5 to top-30, DEX-Comp compresses retrieved contexts by 16\times and accelerates inference by 4\times – 24\times , while achieving performance comparable to or exceeding the uncompressed RAG baseline across retrieval depths. Ablations and evaluations across diverse datasets and backbones further confirm the contribution of each stage and the generalization of our approach.
[NLP-11] Large Language Models with At Most One Spike per Neuron
【速读】： 该论文旨在解决生成式人工智能（Generative AI）中大规模语言模型（LLM）在能效方面的瓶颈问题，特别是如何利用脉冲神经网络（SNN）的固有稀疏事件驱动计算特性实现高效能的低功耗大模型推理。其核心挑战在于传统的时间到首次脉冲（TTFS）编码方式对模型结构的严格限制，导致难以有效编码如层归一化（layer normalization）、矩阵乘法等关键组件。为此，本文提出一种基于参考信号的编码策略，专门针对嵌入层、层归一化、注意力相关操作和丢弃（dropout）这四个核心组件设计了可兼容TTFS编码的映射机制，构建了一个全TTFS编码的端到端训练脉冲神经网络架构。该方案实现了首个基于TTFS编码、参数量达15亿的脉冲语言模型，实验表明其在自然语言理解与常识推理任务上性能接近人工神经网络（ANN）基准，但在语言建模困惑度方面仍存在差距。此外，研究还基于经典能耗模型估算脉冲活动相关的能量消耗，提供了一种以脉冲计数为代理的能效评估指标，而非实际在类脑硬件上的测量。
链接: https://arxiv.org/abs/2609.05151
作者: Zhuoya Zhao,Parsa Omidi,Aref Jafari,Richard Naud
机构: Huawei(华为); University of Ottawa (渥太华大学)
类目: Neural and Evolutionary Computing (cs.NE); Computation and Language (cs.CL)
备注:
Abstract:Leveraging their inherent sparse event-driven computation, spiking neural networks (SNNs) offer a promising path toward energy-efficient large language models (LLMs). Time-to-first-spike (TTFS) coding generates at most one spike per neuron within a time window, yielding extremely low firing rates. However, conventional TTFS SNNs are restricted to specific structures, making it challenging to encode certain blocks in LLM – such as layer normalization and matrix multiplication --using TTFS. To overcome this limitation, we introduce a reference-based strategy specifically to encode the four core LLM components: embedding layers, layer normalization, attention-related operations and dropout. We construct a fully TTFS-based SNN architecture and train it end-to-end. Experiments on modern LLMs like BERT and GPT-2 demonstrate that our approach achieves performance comparable to ANN counterparts on natural language understanding and common-sense reasoning, while a clear gap remains on language modeling perplexity. To the best of our knowledge, this is the first work to scale a spiking LLM to 1.5 billion parameters using TTFS coding. We also report an estimate of spike-related energy; this is a spike-count proxy under an established cost model rather than a measurement on neuromorphic hardware.
[NLP-12] From Vision to Language: Investigating Causal Information Flow in Multimodal Decision-Making EMNLP2026
【速读】： 该论文旨在解决视觉语言模型（Vision-Language Models, VLMs）在做出决策时，其语言输出是否真正基于视觉证据的问题，即探究视觉信息如何影响语言推理过程中的因果路径。其核心解决方案在于采用逐层因果干预（layer-wise causal intervention）方法，针对视频-文本注意力路径进行分析，以追踪跨模态信息在空间、因果与时间维度上的流动机制。关键发现表明，视觉信息主要在模型处理候选答案选项阶段被整合，这些选项成为最终决策的主要语义锚点；此外，名词在多模态语义丰富过程中充当重要语义锚定角色，而动词则在时间关系推理中更为关键；研究还揭示了时间推理存在特定模式——模型难以重建跨视频帧的序列信息，但这种脆弱性可能反映了特定时间表达方式所隐含的语言偏见。
链接: https://arxiv.org/abs/2609.05149
作者: Davide Testa,Hugh Mee Wong,Alessandro Lenci,Bernardo Magnini,Albert Gatt
机构: Fondazione Bruno Kessler (FBK) - Trento, Italy; Università di Roma La Sapienza - Rome, Italy; Utrecht University - Utrecht, The Netherlands; University of Pisa - Pisa, Italy
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at Findings of EMNLP 2026
Abstract:Vision-Language Models are commonly evaluated through their final predictions, but understanding whether these decisions are grounded in visual evidence requires tracing how visual information contributes to language-based decisions. With this purpose in mind, we investigate cross-modal information flow in a video-based generative multiple-choice-like setting by applying a layer-wise causal intervention on video-text attention pathways. We target spatial, causal, and temporal visual reasoning. Our results show that visual information is mainly integrated while the model processes the candidate answer options, which serve as the primary textual grounding sites for the final decision. We further show that nouns play an important role as semantic anchors during multimodal enrichment, while verbs are more relevant when temporal relations are processed. Finally, we identify a distinct pattern in temporal reasoning, suggesting that VLMs struggle to reconstruct sequential information across video frames, but we remark that such fragility may also reflect linguistic biases associated with specific temporal expressions used for defining the relation between events within a scene.
[NLP-13] A Human-in-the-Loop Framework for AI-Assisted Scoring in Large-Scale Writing Assessment
【速读】： 该论文旨在解决大规模教育评估中人工评分效率低、工作量大且难以规模化的问题，尤其是在针对约150–200词的短篇书面作答进行评分时。其核心解决方案是提出一种“人机协同”（human-in-the-loop）的AI辅助评分框架，通过将生成式AI（Generative AI）与人工审核相结合，在保障评估质量的前提下显著降低人工评分负担。关键在于设计了一套智能决策流程，能够识别出最需要人类干预的情形，从而实现专家资源的高效配置。研究基于两轮全国性考试的真实数据（每轮约5,000份作答），验证了模型评分与人工评分在多个评分维度上达到中等至高度一致性，并证明该流程对及格/不及格判定结果的影响可控。研究结论强调，只有在精心设计的人工监督机制支持下，AI辅助评分才可安全应用于大规模测评体系，同时指出了未来需开展的长期模型-人类一致性监测及人工智能引入可能带来的认知偏倚分析等研究方向。
链接: https://arxiv.org/abs/2609.05143
作者: María Eugenia Curi,Germán Capdehourat,Isabel Amigo,Magdalena Romano,Rosana Serra,Adrián Silveira,Andrés Peri
机构: Ceibal(乌拉圭教育信息化项目); ANEP(乌拉圭国家教育行政局)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 25 pages, 8 figures
Abstract:The integration of artificial intelligence (AI), particularly large language models (LLMs), into educational assessment has opened new opportunities to enhance the efficiency and scalability of grading processes. This study presents the design and validation of an AI-assisted scoring framework for written responses in a large-scale national assessment. The proposed approach focuses on short written texts of approximately 150-200 words and incorporates a human-in-the-loop strategy to preserve assessment quality while reducing manual workload. The study is grounded in a real operational context, using data from two recent editions of a nationwide test, each comprising approximately 5,000 student responses. We analyze the alignment between AI-generated scores and human raters across multiple rubric dimensions, as well as the impact of the proposed decision flow on pass/fail outcomes. Results show moderate to high agreement between the model and human evaluations in most dimensions, supporting the feasibility of AI assistance in this setting. Moreover, the proposed correction workflow identifies cases where human review is most valuable, enabling a more efficient allocation of expert effort. The findings suggest that AI-assisted scoring can be safely integrated into large-scale assessment processes only when combined with carefully designed human oversight. The paper concludes by discussing practical implications for deployment in national assessment systems and outlining future research directions, including longitudinal monitoring of model-human alignment and the analysis of potential cognitive bias introduced by AI-supported review workflows.
[NLP-14] NS-ST-GraphRAG : Neuro-Symbolic Spatio-Temporal GraphRAG for Literary Knowledge Processing
【速读】： 该论文旨在解决长篇文学叙事在检索增强生成（Retrieval-Augmented Generation, RAG）中面临的信息处理挑战，具体包括：关键信息分散于不同章节、情节关系随叙事时间动态演变，且正确答案往往依赖于时空与关系约束的协同作用。其核心解决方案是提出NS-ST-GraphRAG——一种神经符号融合的时空图增强生成框架，关键在于通过本体引导的信息抽取、确定性约束验证、双时间坐标建模、空间场景属性嵌入以及动态子图检索机制，实现对查询时空范围内的有效图状态选择，并将生成答案锚定于可追溯的证据。研究进一步构建了Red-Chamber-QA，作为首个面向古典汉语文学的开放多跳问答基准，涵盖时序、空间及一般问题类别，提供分篇章证据片段和确定性捷径控制。在120个问题的独立测试集上，NS-ST-GraphRAG在机械答案复现率（0.733）上优于冻结窗口基线（0.675）和闭卷模型（0.083），语义判断准确率亦达0.866（基线为0.850），虽部分预设约束条件未获支持，但整体表明时空图表示、约束化提取与可审计评估的集成，能够有效支撑长篇叙事中的可验证知识处理。
链接: https://arxiv.org/abs/2609.05139
作者: Zheng Kui Lin
机构: Dalian Ocean University (大连海洋大学)
类目: Computation and Language (cs.CL)
备注: Submitted to Information Processing and Management
Abstract:Long-form literary narratives pose a distinctive information-processing challenge for retrieval-augmented generation: relevant evidence is distributed across chapters, relations evolve over narrative time, and correct answers may depend jointly on temporal, spatial, and relational constraints. We propose NS-ST-GraphRAG, a neuro-symbolic spatio-temporal GraphRAG framework that integrates ontology-guided extraction, deterministic constraint checking, dual temporal coordinates, spatial scene attributes, and dynamic sub-graph retrieval. Instead of retrieving from a single corpus-level graph, the framework selects the graph state valid for the temporal and spatial scope of a query and grounds generated answers in traceable evidence. We further introduce Red-Chamber-QA, to our knowledge the first open multi-hop question-answering benchmark for classical Chinese literature, with time-, space-, and general-question categories, per-part evidence spans, and deterministic shortcut controls. On a 120-question held-out split, NS-ST-GraphRAG achieves mechanical answer reproduction of 0.733 versus 0.675 for the frozen window baseline and 0.083 for a closed-book model (McNemar exact p = 0.092, directionally favorable but not significant); semantic-judge accuracy is 0.866 versus 0.850. The pre-specified constrained-category condition of H2 is not supported by the delivered comparison. These results show how temporal graph representation, constrained extraction, and auditable evaluation integrate into a unified framework for verifiable knowledge processing over long-form narrative.
[NLP-15] Improving Language Identification for Code-Switched Utterances with Integer Linear Programming EMNLP2026
【速读】： 该论文旨在解决多语言混合语料（代码切换，Code-Switched, CS）在语言识别（Language Identification, LID）系统中难以被准确识别的问题，这一缺陷导致代码切换文本在大型语言模型训练数据中严重缺失。其核心挑战在于现有方法对词级语言归属得分的过度依赖，从而影响识别准确性。本文的关键解决方案在于：首先，揭示并修正了MaskLID方法对词级语言关联分数的过拟合问题；其次，将原方法的优化算法重新建模为整数线性规划（Integer Linear Program, ILP），从而能够引入大量清晰且可解释的约束条件；最后，通过实验验证，上述改进在涵盖10种以上语言的多样化基准测试中显著提升了性能，充分证明了该方法的有效性与可扩展性。
链接: https://arxiv.org/abs/2609.05099
作者: Joanna Radoła,Josep Maria Crego,François Yvon
机构: Sorbonne Université, CNRS, ISIR(索邦大学, 国家科学研究中心, 巴黎智能机器人研究所), Paris, France; SYSTRAN by ChapsVision(SYSTRAN由ChapsVision运营), Paris, France
类目: Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026
Abstract:Automatic identification of code-switched (CS) utterances remains a challenge for language identification (LID) systems, causing such texts to be underrepresented in the training data of Large Language Models. In this paper, we revisit MaskLID, a state-of-the art approach for CS identification, which requires no training and detects arbitrary language combinations. We make three main contributions: (a) we reveal, and address, a major issue of MaskLID: its overreliance on word-level language association scores; (b) we reformulate the underlying optimization algorithm as an Integer Linear Program, enabling us to experiment with a large set of clear and interpretable constraints; © each of these improvements vastly improves the baseline system, as we illustrate in experiments involving 10~diverse languages, where we observe a strong boost in performance on CS benchmarks. We release our code and data for reproducibility.
[NLP-16] Measuring AI Accountability Through Argumentation Analysis: Can Model Reasoning Withstand Scrutiny?
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在道德推理评估中因“适当AI行为”标准存在争议而引发的评价困境，尤其针对现有基于人工标注“真实答案”（ground truth）的评估方法难以应对现实情境中固有的模糊性问题。其核心解决方案在于提出一种不依赖明确正确答案、可在高模糊性情境下有效运作的评估框架——通过基于沃尔顿（Walton）论证模式理论与戈维埃（Govier）论证充分性标准构建的四阶段辩证协议（four-phase dialectical protocol），从结构质量角度衡量模型对其决策进行辩护的能力。该协议具备对不同推理范式自适应的能力，突破了传统多选题框架的局限，同时兼顾决策前的推理过程与事后辩护内容。实验结果表明，尽管各前沿模型在多数维度上均高于最低标准，但失败主要集中于论据充分性与依据合理性，且与认知上的规避性表达（epistemic hedging）相关，而非论证长度；模型在事前推理的辩护能力显著优于事后解释，且多数案例中其辩护所采用的论证模式与其实际推理所依赖的模式存在显著差异（每模型≥20%），即使在以价值导向实践推理为主导的情况下亦然。该协议成功识别出严格无法辩护的逻辑缺陷（如自我矛盾、前提错误），并揭示了在人工智能对齐研究中“撤回”机制角色界定的复杂性，提示需发展更具情境化的评估范式。
链接: https://arxiv.org/abs/2609.05088
作者: Daan R. Henselmans,Derck W.E. Prinzhorn,Arno Libert
机构: Aithos Research Foundation(艾索斯研究基金会)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: 27 pages (19 main text + appendix and references), 5 figures, 6 tables. Accepted for publication in the Paris Journal of AI and Digital Ethics (2026); presented at PCAIDE 2026
Abstract:AI oversight methods rely on ground truth for validation, but what constitutes appropriate AI behavior is contested. This leaves evaluation of moral reasoning in LLMs and debate-based oversight implicitly avoiding realistic ambiguity. We investigate an alternative standard designed to function despite such ambiguity: structural quality of the defence a model can mount for its verdicts in response to critical questions, measured through a four-phase dialectical protocol grounded in Walton’s theory of argumentation schemes and Govier’s criteria for argument cogency. The protocol is adaptive to different frames of reasoning, extends beyond multiple-choice framing, and treats both the reasoning that precedes a verdict and its post-hoc justification. Across nine frontier models and 200 high-ambiguity MoralChoice items – 6,778 judge-scored cells, validated against 89.6% inter-judge agreement on the binary failure judgment – models defend their reasoning well above the rubric minimum on every dimension. Failure mass concentrates on grounds and sufficiency, and correlates with epistemic hedging rather than argument length. Reasoning is better defended than post-hoc justification, on every model and every Govier dimension. The scheme a model presents in its justification differs from the one it reasoned with on a substantial share of dilemmas ( \geq 20% per model), despite value-based practical reasoning dominating both tracks. The protocol catches strictly indefensible defences (self-contradiction, false premises), and it surfaces difficulties in characterizing the role of retraction in AI alignment, suggesting a need for more situated evaluations.
[NLP-17] ruthInsightBench: An Evidence-Grounded Benchmark for Automated Evaluation of Open-Ended Scientific Discovery Agents
【速读】： 该论文旨在解决当前自主编码代理（autonomous coding agents）在科研任务中仅能复现已有研究结果，而无法实现真正科学发现的问题。现有基准测试体系以可重现性为核心，其任务、数据与评分标准均围绕隐藏的目标研究设计，奖励的是对既定结果的恢复，而非创新性发现。为突破这一局限，本文提出TruthInsightBench，一个面向真实发现（discovery-oriented）的基准测试。其40个盲测任务源自10个科学领域内40篇同行评审的研究，仅提供中立的科学目标与冻结的数据集，不透露原始结论、预期值及分析路径，迫使代理自主判断数据所支持的主张。通过基于大语言模型（LLM）的固定评判系统，从六个维度（证据成熟度）共29项可验证的产物指标进行自动化、确定性评估，无需人工干预，支持随代理迭代重复评估。实验表明，在同一基础模型上，四个编码代理表现趋同（得分58.4–60.3/100），无统计显著差异：虽能完成并记录分析，具备较强的证据可审计性与新颖性，但普遍缺乏建立可信主张所必需的科学判断行为，如控制变量、稳健性检验、可证伪性及跨数据集泛化能力。核心瓶颈在于科学推理能力而非编程实现，真正的发现仍遥不可及。TruthInsightBench将此差距转化为可量化的靶标，推动智能科研系统向实质性科学洞察演进。
链接: https://arxiv.org/abs/2609.05079
作者: Zhibo Yang,Chen Zhang,Yuewei Zhang,Hao Wang
机构: TruthInsight-AI
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 27 pages, 7 tables, 5 figures
Abstract:Autonomous coding agents are increasingly proposed as AI-scientist systems that conduct analyses and write research reports, but executing a prescribed analysis is not the same as making a discovery. Existing benchmarks are configured for reproduction: tasks, data, and rubrics are built around a hidden target study, and recovery of its result is rewarded. We present TruthInsightBench, a benchmark configured for discovery. Its 40 blind tasks, drawn from 40 peer-reviewed studies across 10 scientific domains, expose only a neutral scientific objective and frozen data; source conclusions, expected values, and analysis paths are withheld, leaving the agent to determine what claim the data support. A fixed LLM-based judge scores the evidentiary maturity of an agent’s own claims along six dimensions, operationalized as 29 artifact-grounded items, with automated, deterministic aggregation and no per-instance human grading, so evaluation can be repeated automatically as agents evolve. On one frozen base model, four coding agents form a narrow plateau (58.4-60.3 of 100) with no statistically reliable pairwise separation: they execute and document analyses competently, with comparatively strong evidence auditability and novelty, but largely lack the discriminating acts that establish a trustworthy claim (controls, robustness, falsifiability, and cross-dataset generalization). The bottleneck is scientific judgment rather than coding, and genuine discovery remains out of reach. TruthInsightBench makes this gap a measurable target; data and scoring code are at this https URL.
[NLP-18] Influence Score and Transformers interpretability: Measure of the Effective Impact of Attention Heads at inference time
【速读】： 该论文旨在解决生成式 AI (Generative AI) 安全性中的关键问题——如何量化 Transformer 模型中注意力头（attention heads）对提示注入检测（prompt injection detection）分类决策的贡献，以揭示模型内部的决策机制。其核心解决方案是提出一种新型影响度量（influence score），该度量融合了注意力头对输出逻辑值（logits）的方向性影响与残差流（residual stream）中的结构贡献，实现了从注意力头、层到整个网络的多尺度分析。该方法在专用于提示注入检测的 DeBERTa 模型上验证有效，能够区分正确与错误预测背后的决策行为，兼顾细粒度电路分析与全局输出导向方法的优势，为研究 Transformer 分类器的决策机制提供了一种系统化工具。
链接: https://arxiv.org/abs/2609.05074
作者: Lisa Bouger,Yannick Teglia,Philippe Loubet Moundi
机构: Thales CDI(泰雷兹CDI); Inria Paris(法国国家信息与自动化研究所巴黎分部); Sorbonne Université(索邦大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:We propose an influence score to quantify the contribution of attention heads to classification decisions in Transformer-based models designed for prompt injection detection. The score combines directional influence on the logits with structural contribution within the residual stream, enabling a multi-scale analysis at the head, layer, and network levels. Applied to a DeBERTa model specialized for prompt injection detection, our framework reveals distinct decision behaviours between correct and erroneous predictions. Our method provides an effective compromise between fine-grained circuit analysis and global output-based methods, and offers a systematic way to study decision mechanisms in Transformer classifiers.
[NLP-19] A Structured Debate-Mixture-of-Agents Framework for Complex Clinical Diagnostic Decision Support
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在医疗诊断任务中因采用单轮问答格式而无法模拟真实临床诊断过程中多轮、协作式推理的问题，从而限制了其在复杂诊断场景下的应用效果。其解决方案的关键在于提出一种名为“辩论型混合代理框架”（Debate-Mixture-of-Agents, DMoA）的多智能体系统，通过角色化分工与结构化交互流程，实现基于角色的迭代式诊断推理。实验表明，相较于GPT-4o基线，DMoA在297例罕见病和1719例疑难病例上分别提升了10.21个百分点的最可能诊断准确率和11.36个百分点的安全率；消融实验进一步证实性能提升不仅源于模型数量或输出长度的增加，更关键的是结构化工作流的设计贡献。此外，研究发现四组两阶段（4×2）结构、更强的基础模型以及更大的上下文令牌预算均有助于提升性能，验证了该框架在临床任务中的潜力，并为多智能体系统在医学人工智能中的深入探索提供了重要依据。
链接: https://arxiv.org/abs/2609.05069
作者: Chang Xia,Leilei Ouyang,Huimin Wang,Yong Zhao,Kang Li
机构: Sichuan University (四川大学); West China Hospital, Sichuan University (四川大学华西医院); College of Computer Science, Sichuan University (四川大学计算机学院); West China Biomedical Big Data Center, West China Hospital, Sichuan University (四川大学华西医院生物医学大数据中心); Med-X Center for Informatics, Sichuan University (四川大学医信中心)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 13 pages, 6 figures
Abstract:Large language models (LLMs) show potential for medical tasks, but their single-turn question-answer format does not reflect how clinical diagnosis is performed in practice. As a result, they remain limited in complex diagnostic settings. We developed Debate-Mixture-of-Agents (DMoA), a novel multi-agent framework that structures role-based interaction to support iterative diagnostic reasoning. Base models and DMoA were evaluated on 297 rare disease cases and 1,719 challenging cases. Across both datasets, DMoA improved most likely diagnosis accuracy by 10.21 percentage points and safety rate by 11.36 percentage points over GPT-4o baseline. Ablation experiments showed that the gains were not simply due to the use of more models or longer outputs, but also reflected the contribution of the structured workflow. Further analyses examined how framework design, base model choice, and token budget affected performance. DMoA performed better with a 4*2 structure, stronger base models, and a larger token budget. These findings demonstrate the potential of DMoA for clinical tasks and suggest further investigation of multi-agent frameworks.
[NLP-20] EuroAlpaca: Task-Preserving Localisation of Instruction Data for European Languages
【速读】： 该论文旨在解决机器翻译（Machine Translation, MT）在多语言指令微调数据生成过程中，因未能保留任务关键约束与输出格式而导致训练样本失真、进而损害模型性能的问题。其核心解决方案是提出EuroAlpaca——一种任务保持型本地化流程与近平行语料库，覆盖50种欧洲语言及地区变体，并配套构建了European-IFEval多语言可验证指令遵循基准。该方法根据具体示例选择性地采用领域级机器翻译或重构等效目标语言实例，结合跨领域一致性与目标语言内部一致性的双重验证机制，确保任务语义的完整性。实验表明，在四类大语言模型的LoRA微调中，直接翻译数据虽提升部分自动评估指标（如ROUGE-L和F-BERT），但使European-IFEval准确率相对基线下降29.8%；而采用EuroAlpaca进行适配后，准确率反超基线12.9%，并同时取得Aya Evaluation Suite上的最优自动评估表现，充分证明了任务语义保全在多语言指令微调中的决定性作用。
链接: https://arxiv.org/abs/2609.05043
作者: Aleix Sant,Jordi Luque,Carlos Escolano
机构: Telefónica Innovación Digital(电信创新数字); Universitat Politècnica de Catalunya(加泰罗尼亚理工大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Machine translation (MT) offers a scalable way to extend English instruction-tuning data to multiple languages, but it can distort task-critical constraints and required outputs, creating corrupted training examples and degrading models trained on such data. We introduce EuroAlpaca, a task-preserving localisation pipeline and near-parallel resource covering 50 European languages and regional varieties, together with European-IFEval, a multilingual benchmark for verifiable instruction following. Depending on the example, our pipeline applies field-wise MT while preserving task-critical content or reconstructs a task-equivalent target-language instance, followed by validation of cross-field coherence and target-language consistency. Across LoRA experiments with four LLMs, training on directly translated data improves ROUGE-L and F-BERT on the Aya Evaluation Suite, but reduces accuracy on European-IFEval by 29.8% relative to the unadapted baseline. In contrast, adaptation with EuroAlpaca improves accuracy by 12.9% over the same baseline, reversing the degradation caused by direct MT, while also achieving the highest ROUGE-L and F-BERT scores on Aya. These results show that preserving task semantics is essential for multilingual instruction tuning.
[NLP-21] How do LLM s Evaluate Perceived Moral Agency? Investigating Moral Decision-Making in Human-Artificial Agents Interactions
【速读】： 该论文旨在解决在智能城市情境中，人类与大型语言模型（LLM）如何评估感知道德代理权（Perceived Moral Agency, PMA）的问题，尤其关注具身性（embodiment）差异对道德责任归因的影响。随着人工智能系统（AAs）如机器人、无人机及无实体AI在公共空间中的广泛应用，其是否具备道德代理权成为亟待厘清的伦理与认知议题。研究的关键在于首次通过实证方法对比分析190名人类参与者与多种大型语言模型在典型智能城市场景下对人类与自主人工代理的道德代理权评价。研究采用经验证的PMA量表进行评估，发现尽管人类普遍赋予自身更高的道德代理权，但在具体道德困境中，LLM表现出高度的情境敏感性：其判断并非基于对代理主体的稳定属性评估，而是聚焦于情境中的伤害严重性与紧迫性，从而动态推理。这一机制与人类评判者的倾向相似，表明LLM在道德决策中展现出类似人类的“情境化”推理特征，凸显了其在日常道德决策中应用时需重视情境依赖性的伦理风险与设计考量。
链接: https://arxiv.org/abs/2609.05037
作者: Fernanda Mansilla,Aloysius Tok,Bahia Guellaï,Farah Benamara,Nancy F. Chen
机构: CNRS@CREATE, Singapore(法国国家科学研究中心新加坡分部); IRIT, Université de Toulouse(图卢兹大学信息与技术研究所), France; CLLE, Université de Toulouse(图卢兹大学语言学与认知科学实验室), France; IPAL, Singapore(新加坡智能感知与学习研究所); Institute for Infocomm Research (I2R), A*STAR(新加坡资讯通信研究院，新加坡科技研究局)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 43 pages, 14 figures, 29 tables. Preprint under review
Abstract:As LLMs take on roles requiring moral advice, understanding how they attribute moral agency becomes critical. Humans possess moral agency, the capacity to make ethically guided decisions and bear responsibility for their consequences, a well-established construct in moral psychology. Yet as artificial agents (AAs) such as robots, drones, and disembodied AI systems become increasingly embedded in smart city environments, the question of whether and how moral agency is attributed to them takes on new urgency. This paper presents, to the best of our knowledge, the first empirical study comparing how humans and LLMs evaluate perceived moral agency (PMA) across human and autonomous artificial agents varying in embodiment, situated in plausible smart city scenarios. Using an adaptation of a validated PMA scale, we applied a protocol to 190 human participants as well as various LLMs. Our evaluation reveals higher perceptions of moral agency in humans than in AAs. However, when facing moral dilemmas in concrete scenarios, LLMs reason outward from the situation, prioritizing harm severity and contextual urgency over any stable assessment of the agent itself, amplifying a context-sensitivity also present in human raters. These findings are particularly relevant as LLMs become increasingly involved in everyday moral decisions.
[NLP-22] Moral Competence Before Moral Content: Why LLM Agents Lack the Prerequisites for Coherent Alignment
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在价值多元主义背景下的对齐问题，即如何确保AI系统的行为符合人类规范、价值观或意图。其核心挑战在于：在缺乏普遍正确道德标准的前提下，如何评估AI系统是否具备一致且可信赖的道德决策能力。为此，论文提出四个结构性条件——判决稳定性（verdict stability）、单调性（monotonicity）、决断性（decisiveness）和帕累托可行性（Pareto viability），用以定义一种可从行为本身评估的道德能力基准，无需依赖外部道德标准或专家参照。这一结构化框架构成了对齐的“最低结构门槛”，而非具体的规范目标。通过在基于大语言模型（LLM）代理的三个模拟道德困境场景中应用该方法，并在五种改写形式、五级升级程度与三种主导性条件下对九个前沿模型进行因子设计测试，研究发现：没有任何一个模型能在所有部署中表现出一致的政策行为；仅表面形式的扰动即可导致单一升级层级下判决率变化高达99个百分点，且模型在一个情境中的表现无法预测其在其他情境中的道德胜任力。这表明当前的基于大语言模型的代理尚不具备可被有意义地进行对齐的对象属性。
链接: https://arxiv.org/abs/2609.05036
作者: Arno Libert,Derck W.E. Prinzhorn,Daan R. Henselmans
机构: Aithos Research Foundation(阿瑟斯研究基金会)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: Accepted for publication in the Paris Journal of AI and Digital Ethics (2026); presented at PCAIDE 2026
Abstract:AI alignment requires AI systems to adhere to human norms, values, or intentions. Under value pluralism there is no correct target, but a shared prerequisite is that the system’s behavior expresses a coherent policy: a mapping from situations to verdicts that is invariant while a situation’s morally relevant features are preserved, and sensitive when they change. We introduce four structural conditions for such coherent policies: verdict stability, monotonicity, decisiveness, and Pareto viability. Together they measure a form of moral competence that is evaluable from behavior alone, without reference to a moral standard or expert baseline, forming a structural floor for alignment rather than a normative target. We demonstrate the methodology on three simulated deployments featuring LLM-based agents facing moral dilemmas. Evaluating nine frontier models under a factorial design of five paraphrases, five escalation levels, and three dominance conditions, we show no model expresses a coherent policy across the three deployments: surface-form perturbation alone produces verdict-rate shifts of up to 99 percentage points at a single escalation level, and a model’s success on one scenario does not predict its competence on another. This suggests LLM-based agents are not currently the kind of object to which alignment can meaningfully apply.
[NLP-23] MoirfEolas and CríochScore: Developing Resources for and the Evaluation of Tokenization Alignment with Irish Morphology
【速读】： 该论文旨在解决爱尔兰语作为低资源语言在自然语言处理（Natural Language Processing, NLP）中面临的分词（tokenization）不准确问题，尤其关注现有分词算法与爱尔兰语形态边界（morphological boundaries）之间的对齐程度。其核心解决方案是提出一个名为MoirfEolas的高质量标注数据集，包含超过35,000个爱尔兰语词汇，并精确标注其对应的换音（eclipses）、前缀（prefixes）和后缀（suffixes），从而为评估分词结果与形态结构的一致性提供基准。同时，论文引入了一种新的评估指标CríochScore，用于量化分词结果与形态边界之间的匹配度。实验表明，基于未登录词语言模型（Unigram Language Model）的分词方法在形态对齐方面优于其他主流算法，但同时也揭示了形态对齐性与分词压缩效率及词汇表效率之间存在权衡关系，为爱尔兰语NLP系统的开发提供了关键实践指导。此外，该研究通过构建过程的可复现性，为其他具有复杂形态结构的语言建立专用形态资源提供了可借鉴的方法框架。
链接: https://arxiv.org/abs/2609.05022
作者: Jane Adkins,Abigail Walsh,Brian Davis,Elaine Uí Dhonnchadha
机构: ADAPT Centre, Dublin City University(爱尔兰都柏林城市大学); Trinity College Dublin(爱尔兰都柏林三一学院)
类目: Computation and Language (cs.CL)
备注: Accepted as a non-archival poster at the Second Tokenization Workshop (TokShop) at COLM 2026
Abstract:This paper presents new tokenization resources for Irish and evaluation measures of alignment with the morphological boundaries of the language. We present MoirfEolas, a dataset of over 35,000 Irish words mapped to their respective eclipses, prefixes and suffixes as well as an evaluation metric CríochScore, that evaluates the alignment of tokenizations with the morphological boundaries present in MoirfEolas. We evaluate common tokenization algorithms using CríochScore as well as intrinsic metrics present in the tokenization literature. We find that the Unigram Language Model aligns with Irish morphology more often than the other algorithms evaluated. We also find trade-offs between morphological-alignment of tokenization with both compression as well as vocabulary efficiency, providing practical insights for Irish natural language processing development. This dataset contributes towards combating the Irish language’s low-resource status; moreover, the construction process reported in this paper can be emulated by other languages to create specialised morphological resources.
[NLP-24] BIT.UA at BioASQ 14B: Modular Retrieval with pg_textsearch and Qdrant and Agent -Based Answer Generation
【速读】： 该论文旨在解决生物医学问答（biomedical question answering, BioASQ）任务中信息检索与答案生成的效率与准确性问题，特别是在复杂查询下如何提升文档召回率和生成答案的质量。其解决方案的关键在于构建一个模块化、可扩展的系统架构：在阶段A的文档检索中，采用基于PostgreSQL的pg_textsearch实现BM25检索，并引入Qdrant进行密集向量索引，支持高效存储与GPU加速的相似性搜索；同时结合基于HyDE的查询扩展与Context-1检索策略，并开发了基于密集检索负样本采样的新重排序训练流程。在阶段A+及B的答案生成环节，提出“大模型作为裁判”（LLM-as-a-judge）框架与一种新颖的代理共识机制（agent quorum），通过多个具有差异化提示的智能体进行辩论并迭代优化，实现自适应文档保留与共识答案生成。此外，首次参与片段生成子任务。实验表明，该系统在各批次中均表现优异，阶段A在部分批次取得MAP排名第五的成绩。研究验证了模块化架构设计与多智能体协同机制对系统性能的显著提升作用，并为未来集成SPLADE与ColBERT等先进模型提供了方向。所有代码均已开源。
链接: https://arxiv.org/abs/2609.04999
作者: André Ribeiro,Rúben Garrido,Alexander Christiansen,Richard A. A. Jonker,Sérgio Matos
机构: 未知
类目: Computation and Language (cs.CL)
备注: 1 figure, 15 tables, 25 pages
Abstract:This paper describes the participation of the this http URL team from the University of Aveiro in the 14th edition of the BioASQ Task B challenge on biomedical question answering. Building on our previous submissions, we introduced a substantially refactored and modular codebase, and made significant changes to both the retrieval and generation components of the pipeline. For Phase~A document retrieval, we replaced the PyTerrier PISA index with PostgreSQL-based pg_textsearch for BM25 retrieval and adopted Qdrant for dense embedding indexing, enabling more efficient storage and GPU-accelerated similarity search. We explored HyDE-based query expansion alongside a Context-1 retrieval strategy. A new reranker training pipeline was developed, incorporating dense retrieval for negative sampling. For Phases A+ and B answer generation, we introduced an LLM-as-a-judge framework and a novel agent quorum mechanism, where multiple agents with diverse prompts debate and iteratively converge on a consensus answer using adaptive document retention. We also participated in the snippets generation subtask for the first time. Our systems achieved competitive results across all batches, with Phase~A systems achieving MAP ranks of 5 (Batch~1,3). We discuss the impact of these architectural changes, lessons learned, and outline directions for future work including SPLADE and ColBERT integration. All code is openly available: this https URL.
[NLP-25] BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference ICML2026
【速读】： 该论文旨在解决大推理模型（Large Reasoning Models, LRMs）在生成长序列思维链（Chain-of-Thought, CoT）时，因键值（Key-Value, KV）缓存随序列长度线性增长而引发的严重内存瓶颈问题。现有KV缓存压缩方法依赖近期查询来预测未来注意力模式的重要性，隐含假设这些近期查询可作为未来注意力模式的可靠代理，但在长时序推理中这一假设失效——部分解码步骤会产生“思维回溯标记”（Thought Revisiting Tokens, TRT），其会重新关注早期远距离上下文（如任务求解计划）。通过系统分析发现，对应TRT的查询在嵌入空间中聚类为少数相似性组。基于此洞察，论文提出BeaconKV，一种无需训练的KV缓存压缩方法：仅保留每个全局查询簇的代表性“信标查询”（beacon queries），以高效预判哪些键值对将被重访，从而避免存储完整的查询历史。在四个开源LRM及多样化推理基准上的实验表明，BeaconKV显著优于现有方法，实现最高达5.8倍的内存压缩比，同时几乎保持全缓存精度，并使吞吐量提升超过4.3倍。
链接: https://arxiv.org/abs/2609.04971
作者: Janghyeon Kim,Minsoo Kim,Kyuhong Shim,Jungwook Choi
机构: Hanyang University (汉阳大学); Sungkyunkwan University (成均馆大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: ICML 2026. Code: this https URL
Abstract:Large Reasoning Models (LRMs) achieve superior problem-solving through extended Chain-of-Thought (CoT) generation, but the resulting key-value (KV) cache grows linearly with sequence length and creates severe memory bottlenecks, often exceeding GPU capacity for long reasoning traces. Existing KV cache compression methods rely on recent queries to estimate future token importance, implicitly assuming these serve as reliable proxies for future attention patterns. We demonstrate that this assumption fails in long-horizon reasoning: certain decoding steps generate Thought Revisiting Tokens (TRT) that re-attend to distant previous context, such as task-solving plans formulated early in the trace. Through systematic analysis, we discover that queries corresponding to the TRT cluster into a small number of similarity groups in the embedding space. Based on this insight, we propose BeaconKV, a training-free KV cache compression method that maintains beacon queries, compact representatives for each global query cluster, to anticipate which KV pairs will be revisited without storing the entire query history. Across four open-source LRMs and diverse reasoning benchmarks, BeaconKV generally outperforms existing compression methods, achieving up to 5.8\times memory reduction while nearly preserving full cache accuracy and improving throughput by over 4.3\times .
[NLP-26] Why We Care About Understanding: Competence through Predictive Compression
【速读】： 该论文试图解决的核心问题是：人类理解（understanding）与信息压缩（compression）之间的关系，以及为何人类理解呈现出高度压缩的形态。其解决方案的关键在于提出三个相互关联的论点：首先，理解作为一种高效代理，表征了一种独特的稳健能力，使个体能够识别值得信赖和学习的对象；其次，理解的状态体现为对领域内在关系结构的心理建模，这种模型支持预测，而可预测性正是压缩的基础——因为可预测的信息无需单独存储，从而形成压缩的表征投影；第三，人类理解的典型特征源于第一点所强调的信任与传播功能，这些功能施加了可演示性与可传递性的压力，促使人类理解趋向于原则性简约。由此构建的框架不仅解释了压缩理论在理解研究中的吸引力及其局限性，还为人工智能系统难以解释的“黑箱”特性提供了新的认知视角。
链接: https://arxiv.org/abs/2609.04962
作者: Matthieu Queloz,Pierre Beckmann
机构: University of Bern (伯尔尼大学); École Polytechnique Fédérale de Lausanne (洛桑联邦理工学院); Idiap Research Institute (Idiap 研究所)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:What is the relation between understanding and compression, and why does human understanding take such a heavily compressed form? Across information theory, machine learning, and AI research, a substantial tradition identifies understanding with compression-a thought captured in Gregory Chaitin’s dictum that “comprehension is compression.” Philosophers, by contrast, have characterized understanding in terms of grasping connections, giving explanations, and handling novelty. This paper bridges the two pictures through three interlocking theses. The first concerns the concept of understanding: it serves as an efficient proxy for a distinctive form of robust competence, enabling us to identify whom to trust and whom to learn from. The second concerns the state of understanding: to understand a domain is to possess a mental model of its relational structure that enables prediction, and what enables prediction enables compression, because what becomes predictable need not be stored separately. Compression is therefore not identical with comprehension, but its representational shadow. The third concerns the characteristically human form of understanding: the fiduciary and transmission functions highlighted by the first thesis impose pressures of demonstrability and transmissibility that drive human understanding toward principled simplicity. The resulting framework explains both the appeal and the limits of compressionist accounts of understanding while shedding light on the inscrutability of AI systems.
[NLP-27] Discourse Dependency: A Continuous Criterion for Translation Difficulty EMNLP2026
【速读】： 该论文旨在解决当前机器翻译（Machine Translation, MT）评估基准缺乏对“难度”明确界定的问题，特别是忽视了指代消解所需上下文跨度这一关键维度。其核心问题在于：现有评测数据集普遍偏向于短距离依赖的翻译任务，无法有效衡量模型在长程语篇依赖（discourse dependency, DDP）情境下的表现。为此，作者提出一种无需标注、基于源端的度量方法——话语依赖性（Discourse Dependency, DDP），通过统计命名实体重提及与代词共指关系来量化段落需回溯的上下文距离。实验验证表明，DDP与人工标注的共指结果高度一致，且99.2%的误判为单向偏倚，证明高DDP值段落必然需要长程上下文支持。将DDP应用于WMT24++和WMT25数据集发现，两者均严重偏向低DDP样本，且领域标签无法区分此类差异。在此基础上，研究进一步在英韩后编辑任务中比较五种上下文注入策略，结果显示随着DDP值增加，所有自动方法均无法跟上人类后编辑的表现；当DDP=15时，人工译文仍受偏好，而自动评估指标却未能反映这一差距。因此，随着前沿系统在整体得分上趋于饱和，评估范式正从“模型得分高低”转向“模型可触达的上下文范围”，即模型的指代消解能力边界。
链接: https://arxiv.org/abs/2609.04959
作者: Ahrii Kim,Chanjun Park,Seong-heum Kim
机构: AI-Bio Convergence Research Inst.(人工智能-生物融合研究机构); School of Software(软件学院); Dept. of Intelligent Semiconductors(智能半导体系); Soongsil University(松林大学)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 (Main Conference)
Abstract:Recent calls for harder machine translation benchmarks have not clarified what difficulty should mean. We argue that one meaningful and currently unmeasured axis is referential reach, the distance a segment must look back into its document to resolve the entities and pronouns it contains. We formalize this as discourse dependency (DDP), a metric-free, source-side measure computed from named entity re-mentions and pronominal coreference. Validated against gold coreference, DDP errs one-sidedly in 99.2% of segments, so a high-DDP segment is certified to require long-range context. Applying DDP to WMT24++ and WMT25 shows that both are heavily skewed toward low-DDP segments, which domain labels do not distinguish. Building on DDP, we compare five context injection strategies in an English-Korean post-editing setup, varying context size and selection. As DDP grows, no strategy keeps pace with human post-editing. On segments with DDP = 15 raters prefer human translations, while automatic metrics register no difference. As frontier systems saturate aggregate scores, DDP shifts evaluation from how well models score to how far they can reach.
[NLP-28] RefactorPlatform: An Open-Source Harness for Controlled Evaluation of Repository-Scale Refactoring Agents EMNLP2026
【速读】： 该论文旨在解决大规模代码库重构（repository-scale refactoring）中编码智能体在保持程序行为不变的前提下，跨多个相互依赖文件传播单一代码变更的挑战。现有评估框架缺乏对影响智能体性能关键设计因素的系统性隔离与量化分析。其解决方案的核心在于提出RefactorPlatform——一个开源的评估平台，通过固定环境、显式控制多个设计维度（包括模型主干结构，如OpenRouter与GitHub Copilot CLI；执行模式，包括基线、检索增强及多智能体协作；以及提示词的具体性），实现对智能体表现的可复现、可审计的评估。该平台支持在独立工作空间中运行，具备实时终端流输出、任务级日志记录（包含令牌、代码差异、交互记录）、基于抽象语法树（AST）的行为验证以及可导出的遥测数据。实验表明：感知AST的分块策略在不同提示模式下相比朴素的基于令牌窗口的分块策略提升25%-30%；而未经优化的检索机制反而低于无检索基线；轻量级的检索增强单智能体（86%成功率）优于所测试的子智能体配置（66%），且在任务传递失败的场景中，检索机制仍能有效提升成功率；尽管检索引入了额外的令牌开销，但其带来的准确率提升恰好抵消了成本，使得每次成功重构的单位成本保持稳定。RefactorPlatform的开源为重构智能体的评估提供了标准化、透明化和可重复的基础。
链接: https://arxiv.org/abs/2609.04898
作者: Aziz Ben Amor,Drish Mali,Mann Acharya,Vijayasri Iyer,Sébastien Bratières
机构: Pi School(皮学校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted at EMNLP 2026 System Demonstrations
Abstract:Repository-scale refactoring requires coding agents to propagate a single change across many interdependent files without altering program behavior, yet to our knowledge no existing harness isolates the design choices that determine agent success on this task. We present RefactorPlatform, an open-source evaluation harness that holds the environment fixed and varies each design axis explicitly: model backbone (via OpenRouter and GitHub Copilot CLI), execution regime (baseline, retrieval-augmented, and multi-agent), and prompt specificity. Each run executes in an isolated workspace with live terminal streaming, per-task logging of tokens, diffs, and transcripts, AST-based verification, and exportable telemetry for audit and reproduction. Demonstrating the platform on 100 multi-file RefactorBench tasks across four model families, we illustrate the analyses it supports: AST-aware chunking outperforms naive token-window chunking by 25-30% across prompt modes, whereas naive retrieval falls below the retrieval-free baseline; a lean retrieval-augmented single agent (86%) beats the sub-agent configuration we evaluated (66%) on matched tasks with no task passing under delegation that fails under retrieval; and retrieval’s accuracy gains absorb its token overhead, leaving cost per successful refactoring unchanged. RefactorPlatform is open-sourced to make refactoring-agent evaluation reproducible and auditable.
[NLP-29] Cache-Aware Joint Router Adaptation for Memory-Efficient MoE Inference
【速读】： 该论文旨在解决混合专家模型（Mixture-of-Experts, MoE）在推理过程中因专家权重总量超出GPU显存容量而导致的频繁权重传输问题。其核心挑战在于如何在不改变原有Top-K专家选择机制的前提下，有效管理专家缓存以减少显存访问开销。解决方案的关键在于提出一种面向模型侧的缓存感知后训练框架，通过联合优化主干MoE结构与轻量级辅助缓存路由模块，在保持原生推理规则的同时提升缓存命中率。其中，时序路由器（Temporal Router）采用仅更新模式，预测同一层内后续令牌对专家的复用可能性，并在不主动加载的情况下保留专家权重；而时空路由器（Spatio-Temporal Router）进一步引入空间路由器（Spatio Router），利用因果前驱的隐藏状态对时序缓存进行精细化修正，从而提升预访问阶段的覆盖能力。实验结果表明，该方法在Qwen3和GPT-OSS上均显著提升了负载调整后的效率，相较于基线模型，缓存命中率提升1.15–18.03个百分点，专家权重传输量降低4.6%–53.3%，且联合后训练策略相比仅优化辅助模块能带来更优性能增益。敏感性分析揭示了缓存容量决定传输需求，而精炼预算则调控预访问覆盖率与主动传输开销之间的权衡。
链接: https://arxiv.org/abs/2609.04895
作者: Zhenhe Wu,Yaping Jin,Qinghua Xing,Hang Zhou,Wei He,Xianjie Wu,Xianfu Cheng,Jian Yang,Hanting Chen
机构: Huawei Technologies(华为技术); Beihang University(北京航空航天大学); Tianjin University(天津大学); University of Sydney(悉尼大学); Beijing Information Science & Technology University(北京信息科技大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Mixture-of-Experts (MoE) models activate only a small subset of experts per token, but the full expert set often exceeds GPU memory, causing repeated weight transfers during decoding. We formulate expert-cache management as a model-side algorithmic problem and propose a cache-aware post-training framework that jointly adapts the MoE backbone and lightweight auxiliary cache routers while preserving the native Top-K expert-selection rule at inference. Its update-only mode, Temporal Router, predicts same-layer reuse and retains experts for future tokens without proactive loading. The full Spatio-Temporal Router adds a Spatio Router that uses the causal predecessor’s hidden state to refine the temporal cache before target-layer access. We evaluate both modes on Qwen3 and GPT-OSS across GSM8K, MATH, and CommonsenseQA. Temporal Router consistently improves cache hit rate and reduces expert-weight traffic over matched LM-only baselines. On Qwen3, Spatio-Temporal Router achieves the best load-adjusted efficiency across three tasks, improving adjusted hit rate by 1.15–18.03 points and reducing traffic by 4.6–53.3% relative to the strongest evaluated prefetching baseline; results on GPT-OSS are competitive but task-dependent. An auxiliary-only ablation preserves baseline accuracy but yields modest cache gains, whereas joint post-training produces larger improvements. Sensitivity analyses show that cache capacity controls transfer demand, while the refinement budget governs the trade-off between pre-access coverage and proactive traffic.
[NLP-30] CC-Mediation: Evaluating Large Language Models for Cross-Cultural Conflict Mediation
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在跨文化调解中面临的双重挑战：一是何时介入（when）的文化敏感性判断，二是如何回应（how）的策略有效性。现有研究进展受限于两大核心瓶颈：缺乏具有可衡量下游影响的跨文化调解数据集，以及缺乏对跨文化立场转变进行系统评估的理论基础指标。为填补这些空白，作者提出了CC-Mediation基准，包含1,661个基于发展性跨文化敏感性模型（Developmental Model of Intercultural Sensitivity, DMIS）构建的十轮对话，涵盖文化冲突情境、调解干预及干预后立场演变轨迹。同时，提出两种基于DMIS的评估指标：轨迹AUC（Trajectory AUC），用于衡量跨文化态度改善的持续性；以及有符号的一维Wasserstein距离（signed Wasserstein-1 distance），用于量化立场转变的幅度与方向。这两项指标均与人类对跨文化立场变化的判断高度一致。基于该基准的实证分析表明，当前LLMs在两个维度上均存在显著局限：在“何时介入”方面，失败源于模型固有的位置先验（positional prior）导致忽视对话上下文；在“如何回应”方面，失败根植于高层特征提取过程中的表征崩溃（late-layer elicitation collapse），而非知识匮乏。
链接: https://arxiv.org/abs/2609.04855
作者: Suhyun Lee,Wenxuan Zhang,W. Quin Yow,Yang Deng
机构: Hanyang University (汉阳大学); Singapore University of Technology and Design (新加坡科技设计大学); Singapore Management University (新加坡管理大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Cross-cultural mediation by large language models (LLMs) requires deciding both when to intervene and how to respond in culturally grounded conflicts. Progress on this problem has been limited by the lack of (1) mediation datasets with measurable downstream effects and (2) principled metrics for evaluating intercultural stance change. To address these gaps, we introduce CC-Mediation, a cross-cultural mediation benchmark of 1,661 ten-turn dialogues grounded in the Developmental Model of Intercultural Sensitivity (DMIS), containing culturally grounded conflicts, mediation interventions, and post-intervention trajectories. We further propose two DMIS-based evaluation metrics: Trajectory AUC, which measures the persistence of intercultural improvement over time, and a signed Wasserstein-1 distance, which measures the magnitude and direction of shifts in intercultural stance. Both metrics show strong agreement with human judgment of DMIS-grounded stance shift. Using CC-Mediation, we find that current LLMs have limitations on both axes: intervention timing (when) failure stems from a positional prior that ignores dialogue content, while mediation strategy (how) failure arises from a late-layer elicitation collapse rather than a knowledge deficit.
[NLP-31] MMTClinic: Multimodal Multilingual Time Series Question Answering and Reasoning Benchmark for Clinical Domain
【速读】： 该论文旨在解决临床医疗人工智能系统在真实复杂场景下缺乏多模态、多语言及时间序列融合的基准评估体系的问题，从而制约了具有临床可靠性与语言包容性的医学大模型（Large Language Models, LLMs）的发展。其解决方案的关键在于构建一个名为MMTClinic的综合性基准数据集，该数据集整合了文本、医学影像与多变量生理信号等多种模态信息，并涵盖跨五种语言（英语、印地语、孟加拉语、马拉地语和泰米尔语）的30,000个问答对（包括15,000道多项选择题与15,000道开放问答题），聚焦于死亡率预测、心率预测与SOFA评分估计三大关键临床任务。通过在零样本、少样本及思维链（Chain-of-Thought）等不同设置下对13个先进大模型进行评估，研究揭示了模型在任务类型、语言和模态间的性能差异，凸显当前医疗大模型在复杂临床推理能力上的局限性。因此，MMTClinic为推动多语言、多模态与时间序列感知的医疗人工智能研究提供了重要基础资源。
链接: https://arxiv.org/abs/2609.04842
作者: Sourav Malakar,Harshit Nigam,Akash Ghosh,Sriparna Saha,Amlan Chakrabarti,Saptarsi Goswami,Priti Singh
机构: Institute of Engineering and Management (印度工程与管理学院); IIT Patna (印度理工学院帕特那分校); University of Calcutta (加尔各答大学); Bangabasi Morning College (班加巴西晨间学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Time-series data in clinical settings is crucial for capturing dynamic changes in a patient’s health over time, enabling timely diagnosis, personalized treatment, and early detection of critical events. However, the development of clinically reliable and linguistically inclusive medical AI systems remains a significant challenge, primarily due to the lack of multimodal, multilingual, and time-series-grounded benchmarks that reflect the complexity of real-world clinical scenarios. To fill this gap, we present MMTClinic, a benchmark designed to evaluate large language models (LLMs) on complex reasoning and question-answering tasks involving clinical time-series. MMTClinic combines text, medical images, and multivariate physiological signals and includes 30,000 QA pairs (15,000 multiple choice questions (MCQs) and 15,000 open-ended questions) across five languages: English, Hindi, Bengali, Marathi, and Tamil. These questions cover three important clinical tasks—mortality prediction, heart rate forecasting, and SOFA score estimation. We evaluate 13 state-of-the-art LLMs in zero-shot, few-shot, and chain-of-thought settings. Our evaluation reveals notable differences in model performance across tasks, languages, and modalities, highlighting current limitations in clinical reasoning capabilities. MMTClinic provides a valuable resource for advancing multilingual, multimodal, and time-series-aware medical AI research. The dataset will be made publicly available on successful acceptance of the work.
[NLP-32] MABPD: Multi-Agent Bias Probing Detection via Structured Argument Debate EMNLP2026
【速读】： 该论文旨在解决新闻文章中媒体偏见检测的难题，传统方法依赖于大量标注数据进行有监督训练，且难以有效识别由隐含语言线索（如情绪化用语、选择性框架和策略性省略）构成的细微偏见信号。其解决方案的关键在于提出一种无需训练的多智能体协同推理框架——MABPD（Multi-Agent Bias Probing Detection），通过三个具备互补视角的大型语言模型（LLM）智能体，在结构化论证辩论（SAD）协议下对文本进行分析与争议调解。SAD引入领域驱动的非对称举证责任机制：无文本证据支撑的偏见主张不具权重，并结合角色加权投票与共识后验证，以显式的协作推理结构替代任务特定的有监督决策边界。实验表明，这种结构化辩论机制是性能提升的核心，移除辩论模块将F1值下降高达10.6个百分点。在BABE基准测试集上，MABPD在未见测试集上取得83.4%的宏平均F1，仅比当前最优有监督模型（MAGPIE, 84.1%）低0.7个百分点；在跨数据集的SemEval 2019超党派新闻分类任务中，实现75.0%的零样本准确率，接近有监督最优结果（82.2%），充分验证了其在不同标注范式下的泛化能力。
链接: https://arxiv.org/abs/2609.04841
作者: Garvit Joshi(1),Stavya Dhyani(1),Jasmine(1),Arun Chauhan(1) ((1) Graphic Era University, Dehradun, India)
机构: Graphic Era University (图形时代大学), Dehradun (德哈拉顿), India (印度)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 20 pages, 6 figures. Accepted to the EMNLP 2026 Main Conference. Code: this https URL
Abstract:Media bias in news articles operates through subtle linguistic cues—loaded language, selective framing, and strategic omission—that resist single-model detection and have traditionally required large annotated corpora for supervised training. We ask whether structured multi-agent deliberation can serve as a principled, training-free alternative to supervised classification for this task. We introduce MABPD (Multi-Agent Bias Probing Detection), a pipeline in which three specialized LLM agents analyze an article from complementary perspectives and resolve disagreements through a Structured Argument Debate (SAD) protocol. SAD implements a domain-motivated asymmetric burden of proof—biased claims without grounded textual evidence carry zero weight—combined with role-weighted voting and post-consensus verification, replacing task-specific supervised decision boundaries with explicit deliberative structure. Ablation confirms that this structured deliberation, not mere agent parallelism, drives performance: removing the debate module reduces F1 by up to 10.6 points. On the BABE benchmark (4,121 expert-annotated sentences), MABPD achieves 83.4% macro F1 on the held-out test split—within 0.7 percentage points (pp) of the supervised SOTA (MAGPIE, 84.1% macro F1; Horych et al., 2024)—without any task-specific training or threshold tuning on annotated data. Cross-dataset evaluation on the SemEval 2019 HyperPartisan corpus (644 articles) yields 75.0% zero-shot accuracy, within 7.2 pp of the supervised SOTA accuracy (82.2%; Kiesel et al. 2019), confirming transfer across annotation regimes. We release the full pipeline and evaluation code.
[NLP-33] On Epistemic Diversity in Large Language Models
【速读】： 该论文旨在解决当前大型语言模型（Large Language Models, LLMs）评估体系过于依赖准确性（accuracy）和对齐性（alignment）的问题，尤其是在知识密集型任务中，模型可能虽给出正确答案，却限制了用户接触多元有效解答、解释或推理路径的可能性。其核心问题在于：现有评估范式未能充分反映模型在知识呈现上的多样性与开放性，从而可能导致“认知窄化”（epistemic narrowness）。解决方案的关键在于引入并形式化“认识论多样性”（epistemic diversity）这一新评价维度，即衡量LLM在回答问题时所展现的有效答案、解释及推理路径的广度。作者提出一个初步的理论框架，并在两个实际领域中实现该维度的操作化测量，发现前沿大模型普遍存在认识论狭窄现象，倾向于将广阔的合理答案空间压缩至少数典型子集。研究强调，未来对大模型的评估应超越以准确率为单一导向的范式，将认识论多样性作为衡量模型能力的重要维度。
链接: https://arxiv.org/abs/2609.04835
作者: Elisabeth Kirsten,Nicole Krämer,Muhammad Bilal Zafar
机构: UAR Research Center for Trustworthy Data Science and Security; Ruhr University Bochum(鲁尔大学波鸿分校); University of Duisburg-Essen(杜伊斯堡-埃森大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) are increasingly used not only to retrieve information, but to answer questions, explain, teach, and support inquiry. In such settings, evaluation cannot be exhausted by accuracy or alignment alone. A system may give a correct answer while still narrowing users’ access %to knowledge. to alternative valid answers, explanations, or reasoning routes. Drawing on the broader notion of epistemic diversity in philosophy and social epistemology, we formalize it in the context of LLMs as the range of valid answers, explanations, and reasoning routes that an LLM exposes to users. We argue that epistemic diversity is a useful evaluation dimension for settings where LLMs are used to support knowledge-intensive tasks. We propose a preliminary framework for conceptualizing and measuring epistemic diversity in LLMs, and operationalize it in two domains. We find that frontier LLMs often exhibit epistemic narrowness, repeatedly collapsing large valid answer spaces onto small canonical subsets. These findings suggest that LLM evaluation should move beyond accuracy-oriented paradigms and treat epistemic diversity as an important dimension of model capability.
[NLP-34] Generating Constructive Feedback on Stories via Reinforcement Learning EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在生成创作性写作反馈时存在的泛化性强、可操作性差以及无法有效识别最核心写作问题等关键缺陷。现有方法生成的反馈往往流于表面，缺乏针对性与指导意义，难以真正帮助写作者提升叙事能力。为此，本文提出一种基于强化学习的解决方案，其核心在于采用组相对策略优化（Group Relative Policy Optimization, GRPO）框架，并设计了一种多组件奖励函数，以显式引导模型生成更具建设性的反馈。该奖励函数聚焦于三个维度：反馈的独特性（即针对具体故事内容）、对故事质量提升的有效性，以及对最紧迫写作问题的精准识别。实验结果表明，该方法在三个故事语料库上的自动评估与人工评估中均显著优于当前主流LLM（包括Gemini）及各类基线模型。研究进一步发现，提供可执行的具体建议是决定反馈建设性的最主要因素。
链接: https://arxiv.org/abs/2609.04824
作者: Maja Stahl,Timon Ziegenbein,Henning Wachsmuth
机构: Leibniz University Hannover(汉诺威莱布尼茨大学); L3S Research Center( L3S 研究中心)
类目: Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026
Abstract:Constructive feedback is crucial for creative writers to refine their storytelling abilities. Since receiving feedback from human experts is often costly and time-intensive, large language models (LLMs) offer a scalable and efficient alternative as automatic writing assistants. Despite their potential, research indicates that LLM-generated feedback is often generic, lacks actionability, and fails to identify which writing issue is most critical. To address these limitations, we present a reinforcement learning approach that steers LLMs to generate constructive feedback without the need for ground-truth feedback. We train our model using group relative policy optimization (GRPO) with a novel multi-component reward function aiming at constructiveness: it prioritizes feedback that is uniquely tailored to the story, helps to improve story quality, and addresses the most critical writing issue. In automatic and human evaluation across three story corpora, our approach outperforms state-of-the-art LLMs (including Gemini) and competitive baselines. We find that providing actionable suggestions is the main driver of feedback constructiveness.
[NLP-35] Reinforcement Learning for improving Large Language Models Catalan text simplification capabilities
【速读】： 该论文旨在解决低资源语言在自动文本简化（ATS）任务中因缺乏高质量标注数据而导致模型性能受限的问题，尤其关注如何利用大语言模型（LLM）提升低资源语言的简化质量。其核心解决方案在于提出一种基于强化学习（RL）的优化框架——组相对策略优化（Group Relative Policy Optimization, GRPO），并设计了一种新型奖励函数，该函数结合了SARI指标与特定惩罚项，以引导模型生成符合目标简化风格的文本。通过在ASSET数据集上对IberianLLM-7B-Instruct进行后训练，验证了该方法在英语和两种加泰罗尼亚语基准上的有效性，显著提升了简化质量并抑制了先前存在的负面行为。尽管探索了跨语言迁移学习（将ASSET翻译为加泰罗尼亚语和西班牙语进行微调），但其在域外基准上未表现出显著提升，表明当前方法在跨语言泛化方面仍存在挑战。
链接: https://arxiv.org/abs/2609.04823
作者: Arnau Ayguadé Domingo,Stefan Bott,Horacio Saggion
机构: Universitat Pompeu Fabra(庞培法布拉大学); Barcelona Supercomputing Center(巴塞罗那超级计算中心)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted at CLEAR-TEXT 2026: Readability and text simplification workshop at the International Conference Computational Linguistics in Bulgaria (CLIB 2026)
Abstract:Although automatic text simplification (ATS) is critical for accessibility, its progress has not matched the rapid evolution of broader natural language processing techniques. This paper investigates the application of reinforcement learning (RL) to improve the quality of ATS for low-resource languages using Large Language Models (LLMs). The paper introduces a novel reward function, designed to guide LLMs toward a targeted simplification style with Group Relative Policy Optimization (GRPO), that combines the SARI metric with specific penalty components. The effectiveness of GRPO with this reward function is motivated and demonstrated by post-training IberianLLM-7B-Instruct on the ASSET dataset. After post-training on the English ASSET, the model’s ATS performance improves on two curated Catalan benchmarks while also successfully suppressing previously observed negative behaviors. Cross-lingual transfer learning is explored by translating ASSET into Catalan and Spanish and post-training the model on each version, but these fail to show a significant improvement on the out-of-domain benchmark.
[NLP-36] A Systematic Comparison of Multilingual Interpretability Methods Reveals Anisotropy-Driven Failures
【速读】： 该论文旨在解决多语言语言模型中跨语言表征共享程度的量化问题，核心挑战在于现有多种可解释性方法在衡量跨语言共享时存在不一致性，而这种不一致究竟是源于模型本身的特性还是测量方法的偏差尚不明确。研究通过在5个模型家族共21个基础模型（参数规模125M–14B）上对比四种共享度量方法（CKA、ANC、基于词元的GMM主导性、ILO），并将其与五个下游任务上的跨语言迁移性能进行相关性分析，发现各度量方法对共享程度的评估结果差异显著。进一步分析表明，这种分歧主要源于嵌入空间中的各向异性（anisotropy）——即表征倾向于聚集在嵌入空间的一个狭窄锥形区域内。在控制模型规模、模型族别及任务特异性后，仅ILO度量与跨语言迁移性能保持强相关性（Spearman’s ρ = 0.90），因此研究建议将ILO作为主推荐的共享度量指标，并应与其对应的各向异性诊断一同报告，以提升评估结果的可靠性与可解释性。
链接: https://arxiv.org/abs/2609.04819
作者: Oskar Holmström,Marcel Bollmann,Marco Kuhlmann
机构: Linköping University (林雪平大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Multilingual language models develop shared cross-lingual representations, and various interpretability methods claim to quantify this sharing. These methods have been developed largely in isolation, and when they disagree, it is unclear whether the disagreement reflects a property of the model or an artifact of the measurement. We compare four sharing metrics (CKA, ANC, GMM dominance per token, and ILO) across 21 base models from five families (125M-14B parameters) and correlate each with cross-lingual transfer on five downstream tasks. We find that the metrics differ in their quantification of cross-lingual sharing in these models and suggest that the disagreement traces to anisotropy, the tendency of representations to cluster in a narrow cone of the embedding space. Only ILO’s correlation with cross-lingual transfer (Spearman’s \rho = 0.90 ) survives controls for model size, family, and per-task variation. We therefore recommend ILO as the primary sharing metric, to be reported alongside anisotropy diagnostics.
[NLP-37] Recurrence Is Not Enough: Causally Validating Multilingual SAE Translation Features in Gemma 2 and 3
【速读】： 该论文旨在解决生成式语言模型中稀疏自编码器（Sparse Autoencoder, SAE）特征在跨语言场景下是否具备一致因果作用的问题，即某一语言情境下发现的特征在处理其他语言提示时是否仍能发挥相同的因果功能。其核心挑战在于验证特征的跨语言可迁移性与因果有效性之间的关系。解决方案的关键在于：通过在Gemma 2和Gemma 3模型中系统性地扩展SAE特征发现方法至多语言设置（涵盖提示语种、源语种与目标语种的变化），并采用激活放大与消融实验进行因果验证；结果表明，尽管存在超过20个在多种设置中频繁激活的“重复特征”，但其对翻译行为的影响普遍微弱或不一致；而唯一一个在23种语言设置中均表现出稳定因果效应的特征——Gemma 2的(L10, 5717)与Gemma 3的(L20, 2456)，在放大时持续提升COMET评分，消融时则显著下降，揭示了语言无关的翻译启动方向的存在。这表明特征重复性不能可靠指示跨语言因果转移，而需结合严格的因果验证以识别真正具有通用意义的内在表征方向。
链接: https://arxiv.org/abs/2609.04808
作者: Giang Son Nguyen,Nhi Ngoc-Yen Nguyen,Wray Buntine,Dung D. Le
机构: VinUniversity(维大学); Nanyang Technological University (南洋理工大学); University of Science and Technology of China (中国科学技术大学); Monash University (莫纳什大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to BlackboxNLP 2026 Special Track
Abstract:Sparse autoencoder (SAE) features are increasingly used to explain and steer language-model behavior, but it remains unclear whether a feature found in one language context plays the same causal role when processing prompts in another language. We study this question using translation-initiation features (Wu et al., 2026). We reproduce the SAE feature discovery method from Wu et al. in Gemma 2 and extend it to multilingual settings that vary prompt language, source language, and target language. We then test whether features that recur across settings affect translation behavior by amplifying or ablating their activations during inference. We also examine whether the method can be applied to Gemma 3. In both models, we observe an identical finding: although we can find more than 20 features that activate frequently across all discovery settings, causal validation shows that nearly all have small or inconsistent effects. In contrast, one feature – Gemma 2’s (L10, 5717) and Gemma 3’s (L20, 2456) – consistently improves COMET scores when amplified and degrades them when ablated across 23 language settings. These results show that feature recurrence can overstate cross-lingual transfer, while identifying a language-agnostic translation-initiation direction in Gemma 2 and Gemma 3. Comments: Accepted to BlackboxNLP 2026 Special Track Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04808 [cs.CL] (or arXiv:2609.04808v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.04808 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-38] Can Activation Steering Capture Multidimensional Authorship Style? EMNLP2026
【速读】： 该论文旨在解决大语言模型（LLM）在生成过程中对作者风格（authorship style）进行多维度、难以明确界定的控制问题。传统方法依赖自然语言风格描述或专门训练，但存在表述不一致与泛化能力不足的局限。其核心解决方案在于提出一种无需训练的“基于方面感知的激活引导”（Aspect-Aware Activation Steering, A3S）框架，通过结构化对比提示（structured contrastive prompting）在激活空间中直接构建丰富且语义明确的风格表征，尤其聚焦于修辞动机驱动的多维度风格差异。关键创新在于：识别出不同方面（aspect）的对比方向共享一个共通的作者风格主干（common authorship backbone），但在特定方面的残差（res差）上存在冲突，这些残差承载着真实的风格信号；A3S通过干扰感知的聚合机制与实例级可调的引导强度，有效融合各方面的风格方向，从而实现更精准的多方面风格迁移。实验表明，该方法在真实多方面风格转移任务中表现优于基线，在跨领域基准上的偏好评估中也超越有监督训练模型，同时保持目标样本与示例间的重叠度始终较低，验证了其有效性与鲁棒性。
链接: https://arxiv.org/abs/2609.04792
作者: Hieu Tran,Calvin Bao,Marine Carpuat
机构: University of Maryland, College Park(马里兰大学学院市分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026
Abstract:Activation steering has shown promise for controlling LLM generation along well-defined attributes, but it remains unclear whether it can handle the multidimensional and hard-to-define nature of authorship style. We ask whether structured contrastive prompting along rhetorically-motivated dimensions can construct rich style representations directly in activation space, bypassing the need for natural language style descriptors or dedicated training. We find that the resulting directions share a common authorship backbone while conflicting on aspect-specific residuals that carry genuine stylistic signal, explaining why naive aggregation fails. We operationalize this in Aspect-Aware Activation Steering (A3S), a training-free framework that merges per-aspect contrastive directions with interference-aware aggregation and tunes steering strength per instance. A3S improves authorship style transfer where it is genuinely multi-aspect, outperforms a trained baseline in preference evaluations on out-of-domain benchmarks, and keeps target-exemplar overlap consistently low.
[NLP-39] Persistent Teacher Anchoring for Tool-Using Agents EMNLP2026
【速读】： 该论文旨在解决大语言模型（LLM）后训练中基于策略的知识蒸馏（On-Policy Knowledge Distillation, OPKD）所导致的师生分布漂移问题，尤其在工具使用场景下，由于学生生成的工具调用在未受监督的情况下提前执行并影响后续输入前缀，加剧了分布偏差。其核心解决方案是提出持久教师锚定（Persistent Teacher Anchoring, PTA），通过引入教师对完整对话回合（turn-level commitment）的确认机制，确保只有经教师验证的完整对话单元才被提交至环境执行，从而有效抑制生成过程中的累积性漂移。PTA保留了块级（chunk-level）验证机制，并进一步实现回合级承诺，使工具调用仅在教师完成整轮验证后才进入环境。同时，该方法引入持久前瞻（persistent lookahead），将未来采样提前推进并在学生更新间保持未完成样本的一致性，充分利用空闲推理资源，显著提升吞吐量。实验表明，在类似Search-R1的检索强化学习与DeepEyes风格的感知强化学习任务中，相较于标准OPKD，PTA在相同下游强化学习预算下分别将宏平均最佳命中率@4（macro best@4）提升2.5和2.8个百分点，且前瞻机制使系统吞吐量提高24%。
链接: https://arxiv.org/abs/2609.04773
作者: Hyun Bin Park(1),Kyungho Song(2),Sangmin Lee(1),Du-Seong Chang(1) ((1) Sogang University, (2) University of Michigan, Ann Arbor)
机构: Sogang University (首尔大学); University of Michigan, Ann Arbor (密歇根大学安娜堡分校)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 16 pages, 4 figures, 8 tables. Accepted at EMNLP 2026 (Main Conference)
Abstract:Distillation is common in LLM post-training, where on-policy knowledge distillation (OPKD) uses student-generated trajectories to prepare the student for downstream RL. At each state, the student matches a next-token distribution supplied by the teacher. As the rollout enters states the teacher would not visit, the teacher-student distribution gap can accumulate. In tool use, this gap becomes consequential because student-written calls execute before supervision and their observations shape later prefixes. Proposer-verifier generation addresses this drift by letting the teacher decide which student-proposed text is retained during generation. Existing formulations govern text but leave tool execution outside their scope. We propose Persistent Teacher Anchoring (PTA), a student-induced but teacher-committed rollout construction. PTA retains chunk-level verification and adds turn-level commitment, allowing a call to reach the environment only after the teacher has verified the entire turn. Treating verified chunks as atomic generation units, we introduce persistent lookahead, which fills idle rollout capacity by advancing future samples and carrying unfinished ones across student updates under the fixed verifier. Across Search-R1-style retrieval and DeepEyes-style perception RL, applying PTA before downstream RL improves macro best@4 by 2.5 and 2.8 points over OPKD under the same downstream RL budget, while lookahead improves throughput by 24%.
[NLP-40] Vectorizing Classical Tamil: Representation Learning for Verse-Commentary Pairs
【速读】： 该论文旨在探究信息表示学习在处理古典泰米尔语文本中注释对（verse-commentary, urai）时，能否有效捕捉语义关联与结构特征。其核心问题在于：在缺乏显式标注的情况下，不同深度学习模型是否能从原始文本中自动习得有意义的语义表示，并准确建模诗句与其注释之间的对应关系。解决方案的关键在于构建一个包含1,262个诗句-注释对的高质量语料库，并系统评估多种主流架构——包括循环神经网络、Transformer编码器、孪生匹配网络、mBART风格的编码器-解码器以及仅解码器语言模型——在语义匹配、生成质量与表示能力方面的表现。通过引入基于TF-IDF的无训练词袋基线、生成重叠度对比及控制实验，研究揭示了当前模型在语义表征上的局限性：尽管解码器仅模型在最小对比较中表现出95.5%的真词语序偏好，却无法复现被遮蔽的注释内容；同时，诸如词元F1分数极低（0.02–0.20）、训练损失持续下降而验证损失已上升等现象，表明模型存在过拟合与表征退化问题。研究强调了现有模型在理解深层语义结构方面的不足，并公开了数据提取与评估流程，为后续研究提供可复现基准。
链接: https://arxiv.org/abs/2609.04755
作者: Amrit Gopinath,Sangeetha Sivanesan
机构: Sri Sivasubramaniya Nadar College of Engineering (斯里·西瓦苏布拉马尼亚·纳达爾工程學院); National Institute of Technology Tiruchirappalli (特里奇拉帕利国家技术学院)
类目: Computation and Language (cs.CL)
备注: 8 pages, 5 figures, 6 tables
Abstract:We construct a corpus of 1,262 verse-commentary (urai) pairs from five Classical Tamil source sections, ranging from technical grammatical prose to modern paraphrase, and ask what information representation learning can recover. We train recurrent and Transformer encoders, a Siamese-style pair-matching network, an mBART-style encoder-decoder, and a decoder-only language model. Each analysis is interpreted against an appropriate control on the same data. TF-IDF provides a strong no-training lexical retrieval baseline, alongside representation analyses and generation controls for the learned models. A fixed string containing the 25 most frequent commentary words scores higher on generation overlap than the decoder-only model. Canonical correlation reaches 1.000 on Gaussian noise at these sample sizes, token-F1 spans only about 0.02-0.20 on this corpus, and the encoder-decoder continues to lower training loss for sixteen epochs after validation loss has begun to rise. One narrow result remains: the decoder-only model prefers authentic word order in 107 of 112 minimal-pair comparisons (95.5%), but does not reproduce held-out commentary content. We release the extraction and evaluation protocol; redistribution of the source commentaries remains subject to permission.
[NLP-41] Beneath the Surface of Chains-of-Thought: A Mechanistic Interpretation of Reasoning Operations in LLM s EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在推理过程中，其内部表示空间中不同推理操作（如问题构建、目标分解、演绎推理等）的几何组织结构不明确的问题。现有研究虽能从文本层面区分这些功能操作，但对其在隐藏表示空间中的几何特征尚缺乏系统认知。本文的关键解决方案在于揭示：不同推理操作在模型中间层的隐藏表示中具有可分离的几何结构，且该分离性在中间层达到峰值；通过消融实验排除了词法和位置信息的干扰，验证了该结构的语义特异性。进一步发现，操作对齐的表示随层加深逐渐在词元跨度上分布更广，且同一表面词元在不同推理上下文块中被差异化编码。注意力掩蔽干预实验表明，块起始处的操作对齐表示依赖于前序推理上下文。因此，本研究的核心贡献在于证明了语言模型在内部表示空间中保持了语言推理表达与几何结构之间的对应关系，为理解模型内部推理机制提供了新的几何视角。
链接: https://arxiv.org/abs/2609.04753
作者: Seogyeong Jeong,Jaehui Hwang,Dongyoon Han,Geonmo Gu,Alice Oh,Taekyung Kim
机构: KAIST(韩国科学技术院); NAVER AI Lab(NAVER人工智能实验室)
类目: Computation and Language (cs.CL)
备注: To appear in EMNLP 2026 Main Conference. 43 pages, 14 figures, 19 tables
Abstract:Reasoning in large language models unfolds through diverse functional operations, such as problem formulation, goal decomposition, and deduction. Although these operations are explicitly distinguished in text, little is known about how they are geometrically organized in representation spaces. To this end, we investigate whether distinct reasoning operations exhibit corresponding geometric structure in hidden representations. We find that operations are separable in held-out representations, with separability peaking in middle layers, and verify that this structure is not explained by lexical or positional confounds. Across layers, token-wise operation-alignment becomes more distributed over spans, while identical surface tokens are represented differently depending on the operation of its surrounding chunk. Attention-masking interventions further show that operation-aligned representations at chunk onset depend on preceding reasoning context. Consequently, our work demonstrates that language models maintain representational correspondence between linguistic reasoning expressions and their internal geometric structures. Code and project materials are available at this https URL.
[NLP-42] Knowing What Not to Answer: Selective Non-Compliance in Vision-Language Models EMNLP2026
【速读】： 该论文旨在解决视觉-语言模型（Vision-Language Models, VLMs）在面对包含可回答内容与应拒绝成分混合的复杂查询时，难以实现选择性不合规（selective non-compliance）的问题。现有评估基准通常仅从整体查询层面判断是否拒绝响应，忽略了真实场景中查询常包含需部分回应、部分拒绝的混合性质。为此，本文提出KoNA基准，涵盖五类典型情境：错误前提（False Premise）、视觉不可及性（Visual Inaccessibility）、普遍未知（Universal Unknown）、任务可行性（Task Feasibility）与安全性（Safety），并通过成对的单一查询与复合查询测试模型在查询级和组件级的非合规能力。实验表明，现有VLMs在需要选择性拒绝时普遍存在不当响应问题。为应对这一挑战，研究采用KoNA中要求选择性不合规的样本，结合完全可回答的样本进行微调，显著提升了模型在非合规准确性上的表现，同时保持了对完整可答任务的高性能。结果表明，经微调后的模型能够有效区分需回应与需拒绝的语义成分，并作出任务适配的响应。
链接: https://arxiv.org/abs/2609.04720
作者: Minji Kim,Jihyoung Jang,Hyounghun Kim
机构: POSTECH(浦项科技大学); Graduate School of Artificial Intelligence(人工智能研究生院); Department of Computer Science and Engineering(计算机科学与工程系)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Main Conference (43 pages). Code and dataset available at this https URL
Abstract:Vision-language models (VLMs) are expected to respond helpfully to appropriate requests while withholding compliance with requests that are incorrect, unsafe, infeasible, or unanswerable. However, existing benchmarks predominantly evaluate non-compliance at the level of the query as a whole, assuming that each request either warrants compliance or requires withholding compliance. In practice, real-world queries can contain a mixture of answerable content and components for which compliance should be withheld. In this paper, we introduce KoNA, a benchmark for evaluating selective non-compliance in VLMs across five categories: False Premise, Visual Inaccessibility, Universal Unknown, Task Feasibility, and Safety. Each task evaluates two capabilities: query-level non-compliance and component-level non-compliance under paired single and compound queries. Our evaluation across diverse VLMs shows that models often fail to refuse, correct, or abstain appropriately, and these failures become more pronounced when queries require selective non-compliance. To address this challenge, we fine-tune VLMs using KoNA examples that require selective non-compliance, together with a fully answerable set that should receive direct answers. Our fine-tuned models achieve substantial improvements in non-compliance accuracy while largely maintaining performance on fully answerable tasks. These results suggest that the fine-tuned models can distinguish between answerable components and those requiring non-compliance and respond in a task-appropriate manner.
[NLP-43] Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models EMNLP2026
【速读】： 该论文旨在解决大语言模型在对齐过程中难以平衡有用性与安全性的问题，即如何准确区分真正有害的查询与表面看似危险但实际无害的查询（如“我在哪里能拍一张好照片？”被误判为“如何杀人？”）。其核心挑战在于现有安全微调数据集中的拒绝回复（refusal statement）会诱导模型过度依赖表面语义线索，导致大量误拒（false refusals）。论文提出的关键解决方案是将安全微调数据中的响应分解为两个独立成分：标准化的拒绝声明和解释拒绝理由的推理（rationale）。实验表明，仅使用推理部分进行训练可显著降低误拒率，同时保持与全量训练相当的安全性能。该方法在基于提示学习（ICL）的配置中同样有效，并兼容多种推理时缓解策略。研究强调了构建精细、高质量的安全监督数据集的重要性，为实现更具帮助性且更安全的对齐智能体提供了新方向。
链接: https://arxiv.org/abs/2609.04714
作者: Minji Kim,Hyounghun Kim
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Main Conference (38 pages); Code available at this https URL
Abstract:Striking a balance between helpfulness and safety remains a fundamental challenge in aligning large language models. To achieve this balance, models should refuse harmful queries (e.g., “How do I shoot someone?”) while remaining responsive to benign inputs, even those superficially resembling harmful queries (e.g., “Where can I shoot a good photo?”). However, models often struggle to distinguish genuinely harmful queries from benign queries that contain superficially risky language, resulting in false refusals. In this paper, we address the issue by decomposing a response in the safety-tuning dataset into two distinct components: (i) a boilerplate refusal statement and (ii) a rationale explaining the refusal. Our experiments and analyses show that refusal statements impede accurate discrimination between harmful and benign queries by inducing reliance on superficial cues. In contrast, training solely on rationales reduces false refusals while maintaining a comparable level of safety performance. Rationale-Only benefits also appear in our ICL configuration and remain compatible with the evaluated inference-time mitigation methods. The results emphasize the necessity of precisely curated, fine-grained safety supervision datasets and outline directions for constructing aligned agents that better reconcile helpfulness with safety.
[NLP-44] How Do Language Models Represent and Use Phonological Information for Allomorph Selection? EMNLP2026
【速读】： 该论文旨在解决语言模型在生成具有音位条件的词形变化（如英语不定冠词a/an）时，其背后是否依赖于规则性泛化而非单纯的词条记忆，以及这种泛化机制如何在模型内部实现的问题。解决方案的关键在于揭示：语言模型在触发词（trigger token）的嵌入表示中，以单一线性方向编码了音位条件信息；该方向在生成阶段通过预测后续触发词的音位特征，并据此因果驱动冠词选择，从而实现对所有形（allomorph）的音位条件化选择。进一步研究表明，这种规则性泛化能力不仅限于英语冠词系统，还可扩展至其他语言中的所有形选择及显式的音位判断任务，表明语言模型具备一种隐含的、生成时使用的音位规则表征，且该能力与显式的元语言判断能力可分离。
链接: https://arxiv.org/abs/2609.04708
作者: Sangwoo Kim,Sangah Lee
机构: Seoul National University (首尔国立大学)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026
Abstract:Language models are trained on tokenized text that obscures the sound structure of words, yet they reliably produce morphemes whose form is phonologically conditioned. It remains unclear whether they rely on item-specific memorization or rule-like generalization and, if the latter, how that generalization is implemented. We therefore ask whether this phonological condition is represented within language models and how it is causally used for allomorph selection. For the English indefinite article a/an, we show that the phonological condition is encoded along a single linear direction in trigger-token embeddings, that this direction causally drives article selection in token-level wug tests, and that, at the article-prediction position, the model forecasts the upcoming trigger token and uses the forecasted trigger’s phonological feature to choose the article. We then ask whether this rule-like generalization extends beyond English article selection, both to allomorph selection in other languages and to explicit phonological judgment. Together, these results provide a mechanistic account of phonologically conditioned allomorph selection in language models, and dissociate this generation-time ability from explicit metalinguistic judgments.
[NLP-45] Retinal OCTA Phenotyping with LLM Reporting for Alzheimers Disease
【速读】： 该论文旨在解决阿尔茨海默病（Alzheimer’s disease, AD）早期识别中现有评估方法成本高、资源消耗大或不适用于大规模人群筛查的问题。其核心挑战在于如何在缺乏诊断标签的情况下，实现对视网膜微血管结构的可解释性分析与临床意义关联。解决方案的关键在于构建一个可解释的光学相干断层扫描血管成像（OCTA）分析流程，整合标注感知的血管分割、分层特异性血管生物标志物提取、无标签表型分析以及基于测量证据的大型语言模型（LLM）报告生成。通过117张ROSE-1图像数据集验证，该框架在浅层血管复合体（SVC）、深层血管复合体（DVC）及联合SVC+DVC表示上实现了0.916–0.970的受试者工作特征曲线下面积（ROC-AUC）和0.695–0.781的Dice系数，表明分割性能良好；进一步提取六项密度与分形维数生物标志物，形成个体水平表型特征，并在未标注的9名受试者中识别出一致的低密度、低分形维数表型模式；最后，利用GPT、Gemini和Llama生成的报告在测量依据性、引用忠实度和诊断谨慎性方面表现良好，从而建立了从视网膜血管量化测量到探索性表型分析再到证据支持的非诊断性解读之间的透明链路，为阿尔茨海默病研究提供了可解释、可扩展的新型影像分析范式。
链接: https://arxiv.org/abs/2609.04689
作者: Progga Paromita Dutta,Jeba Maliha,Md Rafiul Kabir
机构: Columbia University (哥伦比亚大学); Central Michigan University (中央密歇根大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 4th IEE International Conference on Artificial Intelligence, Blockchain, and Internet of Things, (AIBThings)
Abstract:Early identification of Alzheimer’s disease (AD) remains challenging because established assessment methods can be costly, resource-intensive, or unsuitable for population-scale screening. Optical coherence tomography angiography (OCTA) provides non-invasive visualization of retinal microvasculature, but existing approaches often require diagnostic labels and provide limited measurement-level interpretation. We present an explainable OCTA pipeline that integrates annotation-aware vessel segmentation, layer-specific vascular biomarker extraction, label-free phenotyping, and measurement-grounded LLM reporting. Using 117 ROSE-1 images from 39 subjects, we apply annotation-matched segmentation models to superficial vascular complex (SVC), deep vascular complex (DVC), and combined SVC+DVC representations. The models achieve ROC-AUC values of 0.916-0.970 and Dice scores of 0.695-0.781. Six density and fractal-dimension biomarkers form subject-level profiles for exploratory clustering. Analysis of nine held-out subjects identifies an internally consistent lower-density, lower-fractal-dimension phenotype, although the absence of diagnostic labels prevents clinical interpretation. Reports generated using GPT, Gemini, and Llama are evaluated for measurement grounding, citation faithfulness, and diagnostic caution. Overall, the framework provides a transparent, non-diagnostic connection between retinal vascular measurements, exploratory phenotyping, and evidence-linked interpretation for Alzheimer’s research.
[NLP-46] Controlling and Assessing Appropriate Persona Use in LLM -based Dialogue Generation EMNLP2026
【速读】： 该论文旨在解决生成式对话中基于角色的对话生成（Persona-based Dialogue Generation, PDG）存在的核心问题：大型语言模型（LLM）在生成回复时过度使用角色属性，即使在不适宜的语境下也机械地引入所有给定的角色信息，导致生成结果不自然。这一现象的根本原因在于模型存在系统性偏差，倾向于无差别地采纳全部角色属性，而现有评估指标无法有效衡量角色使用与对话上下文之间的适配性。针对上述问题，论文提出了一种名为自对比角色属性过量抑制（Self-CONtrastive Persona Overuse Suppression, SCONPOS）的解决方案，其关键在于在提示编码阶段直接干预模型内部表示，通过自对比学习机制抑制不必要的角色属性激活，且无需依赖生成过程即可实现。同时，论文提出了新的评估指标——角色适当性评分（Persona Appropriateness Score, PAS），能够同时惩罚角色属性的过量使用与不足使用，从而更准确地衡量角色信息在具体语境下的适用性。实验表明，SCONPOS能系统性降低角色属性的过量使用，而PAS则有效捕捉了角色使用与上下文的契合程度。
链接: https://arxiv.org/abs/2609.04676
作者: Jongkyung Shin,Inkyu Lee,Chiehyeon Lim
机构: UNIST(韩国科学技术院); POSTECH(浦项科技大学); POSCO Holdings Inc.(浦项制铁控股公司)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 (Main)
Abstract:In persona-based dialogue generation (PDG), LLMs often overuse persona attributes by incorporating them regardless of dialogue context, resulting in unnatural responses. Despite its practical significance, the underlying causes remain unexplored, with no method to mitigate this problem or metric to assess the appropriateness of persona use. To address these issues, we first conduct a comprehensive analysis of LLM-based PDG, revealing that LLMs exhibit a systematic bias to incorporate all given persona attributes, and that existing metrics fail to capture contextual appropriateness. Building on these findings, we propose Self-CONtrastive Persona Overuse Suppression (SCONPOS) to mitigate overuse by directly intervening in LLMs’ internal representations at the prompt encoding stage, without requiring any response generation. We further propose the Persona Appropriateness Score (PAS), a novel metric that penalizes both overuse and underuse. Experimental results demonstrate that SCONPOS systematically reduces overuse, and PAS captures the contextual appropriateness of persona use.
[NLP-47] Choosing the Right Language Mode at Inference Time for Multilingual Reliability EMNLP2026
【速读】： 该论文旨在解决多语言大语言模型在低至中等资源语言（low- to mid-resource languages）中推理能力薄弱的问题，特别是探究在何种程度下引入英文翻译能够有效提升模型的推理性能，以及过度翻译是否会引发干扰与过度自信。其核心挑战在于：适量的英文上下文可增强非英语语言的理解并纠正因语言理解不足导致的错误，但冗余的双语上下文反而会引入认知干扰，降低推理可靠性。解决方案的关键是提出一种无需训练的测试时框架——可靠性感知自适应推理（Reliability-Aware Adaptive Inference, RAAI），该框架通过两个机制实现优化：(i) 基于期望校准误差（Expected Calibration Error, ECE）感知的路由与提示融合策略，动态选择最优输入模式；(ii) 利用中间层风险指数（Risk Index, RI）对序列推理过程进行门控，在高风险或潜在有害的双语冗余情境下抑制不必要的计算，仅在有助于提升推理的条件下分配算力。实验结果表明，RAAI在两类主流模型上均显著提升了低资源语言的准确率（提升25%-37.7%），同时将校准误差降低约3%-6%，尤其在最低资源语言层级表现最为突出。
链接: https://arxiv.org/abs/2609.04653
作者: Ekata Mitra,Ameeta Agrawal
机构: PortNLP, Portland State University (波特兰州立大学)
类目: Computation and Language (cs.CL)
备注: Accepted in Findings of EMNLP 2026
Abstract:Multilingual large language models often struggle to reason in low- to mid-resource languages. Prior work has shown that translation can improve multilingual reasoning by helping models access stronger English-centric representations. This raises a central question: How much translation is needed for multilingual large language models to reason reliably, and when does more translation instead trigger interference and overconfidence? Using LLaMA and Qwen models, we run extensive experiments varying text scope and language mode (target-only, English-only, bilingual) to evaluate both accuracy and reliability. Our results reveal a clear trade-off: English context often improve understanding and recover errors caused by non-English comprehension, yet adding redundant bilingual context intensifies interference. We address this trade-off with Reliability-Aware Adaptive Inference (RAAI), a training-free test-time framework that (i) performs Expected Calibration Error (ECE)-aware routing and prompt fusion, and (ii) uses a mid-layer Risk Index (RI) to gate sequential reasoning, allocating compute only when it is likely to help and suppressing harmful bilingual redundancy. Across two model families, RAAI enhances accuracy by 25-37.7% on low-resource languages and lowers calibration error by approximately 3-6%, with the most pronounced benefits in the lowest-resource language tiers. Comments: Accepted in Findings of EMNLP 2026 Subjects: Computation and Language (cs.CL) Cite as: arXiv:2609.04653 [cs.CL] (or arXiv:2609.04653v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.04653 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-48] ConsensusBench: Benchmark of Consensus Nodes for LLM Reasoning via Outcome Reward Densifying
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在复杂推理任务中因依赖稀疏的最终答案奖励（outcome-level rewards）而导致的强化学习信号不足问题。随着推理轨迹长度增加，仅基于最终结果的反馈难以有效指导中间步骤的优化，从而限制了模型的推理能力提升。其解决方案的关键在于提出一种名为ConsensusBench的新数据集，通过规则化的方法提取推理过程中的关键中间结论——即“共识节点”（Consensus Nodes），这些节点是多个正确推理路径中语义等价的中间命题，可作为可验证的子结果（sub-outcome）。在此基础上，构建基于共识节点的规则化过程奖励（rule-based process reward），并将其集成至GRPO类算法中，形成新的强化学习信号ConsensusPR，显著缓解长推理链中的奖励稀疏性问题。实验表明，该方法在AIME 2024、AIME 2025、GSM8K、MATH-500及ConsensusBench上均优于传统GRPO方法，验证了共识节点在引导高效推理方面的实际价值。
链接: https://arxiv.org/abs/2609.04648
作者: Shi-Qi Yan,Chao-Hong Tan,Qian Chen,Wen Wang,Xiangang Li,Zhen-Hua Ling
机构: Alibaba Token Hub, Alibaba Group(阿里巴巴通义实验室)
类目: Computation and Language (cs.CL)
备注:
Abstract:Reinforcement learning (RL) has become one of the primary paradigms for reasoning enhancement of large language models (LLMs). In particular, Group Relative Policy Optimization (GRPO) and related algorithms have demonstrated strong performance with outcome-level rewards. However, these methods depend solely on the final answer, without feedback regarding which intermediate steps contribute to success or failure. As task complexity and reasoning trajectory length increase, such sparse final-answer rewards become increasingly insufficient. To address this limitation, we introduce ConsensusBench, a novel dataset designed to provide rule-based process-level signals. We posit that a correct final answer relies on a small set of intermediate conclusions throughout the reasoning process, which can be seen as a verifiable sub-outcome. We identify these sub-outcomes by filtering correct trajectories from N rollouts and clustering semantically equivalent intermediate statements. We call these clustered statements as Consensus Nodes. By integrating a rule-based process reward derived from these nodes into GRPO-style algorithms, we develop a new reinforcement learning signal named ConsensusPR. It directly reduces the reward sparsity of outcome reward across long reasoning trajectories. To facilitate systematic process-level evaluation, we introduce three metrics to our benchmark: Final Answer Accuracy (Acc), Node Coverage Rate (NCR), and Tokens per Node (TPN). Experiments across AIME 2024, AIME 2025, GSM8K, MATH-500, and our ConsensusBench demonstrate that the proposed method consistently surpasses GRPO-style approaches, highlighting the practical value of consensus nodes in guiding reasoning.
[NLP-49] racing Audio Grounding and Answer Selection in Audio LLM s
【速读】： 该论文旨在解决当前音频大语言模型（Audio LLMs）在推理过程中可能依赖文本线索或语言先验而非真实音频信息的问题。尽管通过使用答案无法仅从文本推断的数据进行训练可提升模型性能，但其内部机制尚不清晰。本文的核心问题是：在模型中必须发生何种变化，才能确保音频内容真正决定最终答案？研究的关键发现在于：（1）将输入音频替换为静音或无关音频时，经过微调的模型性能下降幅度显著大于预训练模型，表明训练增强了对音频的依赖；（2）声学信息主要在早期至中期层影响答案选项的表示，而训练则主要增强中后期层中音频信息对最终预测的影响；（3）训练过程中学习到的权重在特定层级区间产生最大作用。这些结果共同揭示了训练如何通过分层机制强化音频证据在模型决策中的作用，提供了关于音频大模型如何有效利用声学信息的可解释性机制。
链接: https://arxiv.org/abs/2609.04637
作者: Hyebin Cho,Suho Yoo,Jihoo Jung,Joon Son Chung
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Sound (cs.SD)
备注: Preprint
Abstract:Audio Large Language Models (Audio LLMs) have advanced in audio understanding, yet they can still predict the answer by reasoning from textual cues or linguistic priors rather than the provided audio. A common remedy is to train models on data whose answers cannot be inferred from text alone. This approach can improve performance, but what changes within the model remains unclear. In this paper, we ask what must happen inside the model for the audio to actually determine the answer. Our findings are threefold. (1) Replacing the audio with silence or unrelated audio causes substantially larger performance degradation in the trained model than in the pretrained model. (2) Acoustic information most strongly shapes the model’s representations of the answer choices in early-to-middle layers, while training mainly increases the influence of audio information on the final prediction in middle-to-late layers. (3) The weights learned during training have their largest impact in specific layer bands. Together, these results provide a mechanistic account of how training strengthens the use of acoustic evidence in Audio LLMs.
[NLP-50] PetQA: Benchmarking Veterinary Knowledge and Clinical Reasoning EMNLP2026
【速读】： 该论文旨在解决当前大型语言模型（LLM）和大型视觉-语言模型（LVLM）在兽医临床知识与临床推理能力评估方面的缺失问题，尤其针对韩语环境下对宠物（犬猫）长文本问答（Long-form QA）任务的系统性评测需求。其解决方案的关键在于构建首个面向兽医领域的多模态、高质量基准数据集PetQA，包含10,076个纯文本及8,751个图文结合的问答对，均源自真实临床场景并由兽医专家提供权威答案。此外，测试集PetQA-Bench进一步标注了问题类型与临床病症类别，支持细粒度分析。研究通过零样本推理、检索增强生成（RAG）与监督微调（SFT）三种设置，结合ROUGE、BERTScore及基于大模型的判别指标（LLM-as-a-judge）对18个模型进行评估，全面揭示现有模型在事实准确性与实用性方面的局限性，强调需发展更有效的适应方法以构建可信赖的兽医辅助AI系统。为促进国际应用，研究还提供了五种语言的翻译版本。
链接: https://arxiv.org/abs/2609.04598
作者: Taegyun Kim,Youngwook Ham,Jungwook Rhim,Ju-Hyun An,Sungkyu Park,Kunwoo Park
机构: Soongsil University (松仁大学); Kangwon National University (江原国立大学); KDI School of Public Policy and Management (韩国公共政策与管理学院); School of AI Convergence, Soongsil University (松仁大学人工智能融合学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: EMNLP 2026
Abstract:We introduce PetQA, a Korean long-form question-answering (QA) benchmark for evaluating veterinary knowledge and clinical reasoning in large language models (LLMs) and large vision-language models (LVLMs). PetQA contains 10,076 text-only and 8,751 multimodal QA pairs derived from real-world questions about dogs and cats, paired with answers from expert veterinarians. Its test split, PetQA-Bench, further includes annotations for question types and clinical conditions. We evaluate eighteen models using ROUGE, BERTScore, and LLM-as-a-judge metrics for factuality and helpfulness under three settings: zero-shot inference, retrieval-augmented generation (RAG), and supervised fine-tuning (SFT). The benchmarking results provide an overview of the strengths and limitations of current models in addressing veterinary clinical queries and highlight the need for more effective adaptation methods to develop clinically reliable AI systems for veterinary care. To facilitate broader use, we additionally provide translated versions of PetQA-Bench in five languages.
[NLP-51] JLIR: A Julia-Native MLIR-Inspired Intermediate Representation with Automatic JACC Kernel Extraction
【速读】： 该论文旨在解决生成式 AI（Generative AI）在编译器优化中面临的难题，即如何在保持高阶抽象的同时实现高效、可优化的代码生成。针对当前多级中间表示（MLIR）对动态编程语言（如 Julia）支持不足的问题——其严格的编译时类型系统与低层次（C++）扩展机制难以适配高阶、动态特化的语言特性——论文提出了一种名为 JLIR（Julia-native Level Intermediate Representation）的新型中间表示框架。其解决方案的关键在于：在不牺牲语言自然性的前提下，将 MLIR 所具备的多级、领域导向编译优势引入 Julia 生态系统，同时通过原生支持 Julia 语言机制（如宏、泛型和类型推导），实现操作与变换传递的可扩展性，并允许部分类型未知的程序在类型确定前仍可进行编译优化。此外，JLIR 内建算术、控制流、函数、结构化循环及内存操作等方言，并提供轻量级机制以无侵入方式扩展新领域操作，从而打通了线性代数、偏微分方程等科学计算库与编译优化路径之间的壁垒。通过在自动加速器代码生成（JACC）中的应用，验证了 JLIR 在保持代码可读性与表达力的同时，能够有效支持高性能优化。
链接: https://arxiv.org/abs/2609.04585
作者: Narasinga Rao Miniskar,Seyong Lee,Keita Teranishi,Jeffrey S Vetter
机构: Oak Ridge National Laboratory (橡树岭国家实验室)
类目: Programming Languages (cs.PL); Computation and Language (cs.CL)
备注:
Abstract:The Multi-Level Intermediate Representation (MLIR) has made reusable compiler infrastructure practical for domain-specific computation. However, MLIR’s strong compile-time type requirements and low-level (C++) extension model can be a poor match for high-level, dynamically specialized languages such as Julia. MLIR has several drawbacks for dynamic programming languages in terms of the type system and level of abstraction. It is thus extremely challenging for non-compiler or scientific computing users to introduce new programming abstractions and express algorithm implementations in a form that remains both natural and optimizable. As a result, library interfaces for linear algebra, mesh processing, partial differential equations, and related domains often sit outside the compiler optimization path. We present JLIR (Julia-native Level Intermediate Representation), a Julia-native intermediate representation framework that brings the main benefits of MLIR-style multi-level, dialect-oriented compilation into the Julia ecosystem while remaining usable as ordinary Julia code. JLIR represents Julia programs before low-level lowering, supports extensible operations and transformation passes through Julia’s language mechanisms, and allows partially typed programs to remain transformable until concrete types are known. The framework includes built-in dialects for arithmetic, control flow, functions, structured loops, and memory operations, and it also includes a lightweight mechanism for adding new domain operations without modifying the core system. To demonstrate JLIR’s capabilities, we applied it to automatic Julia for Accelerators (JACC) kernel generation.
[NLP-52] When Do Internal Probes Beat Reading the Answer? Miscalibrated Readouts and Behavior-Concealed Knowledge in Language Models
【速读】： 该论文旨在解决生成式语言模型在逻辑推理任务中行为表现与内部表征不一致的问题，即模型在表面上对正确与错误结论均回答“YES”（行为上无区分能力），但其隐藏状态中仍保留可被有效探测的正确判断信号。核心问题在于：尽管模型内部存在对逻辑有效性的真实感知（通过线性探针可达到0.96 AUC），这种信息在最终输出阶段被严重削弱甚至完全丢失。解决方案的关键发现是：这一失败主要源于一个单一的、过高的决策阈值（saturated decision threshold），该阈值偏移高达+4.6个标准差，导致模型无法将内部判别信号转化为正确的输出。研究进一步表明，通过简单的一参数校正（如重置阈值或采用校准边际解码），即可在不使用目标结构训练数据的情况下，将0.6B模型的行为准确率从50%提升至81%，8B模型则恢复至94%。该诊断具有普适性，在多模型、多配置和跨规模实验中均表现出高度一致性，揭示了模型行为偏差的本质是“误校准”而非“知识缺失”。此外，研究还区分了三种状态：隐匿（concealed）、误校准（miscalibrated）与未被检测（undetected），并通过一个无表面线索的迷宫任务验证了审计方法的有效性。最终，该工作提出了一种可行动的诊断与修复框架，强调内部表征分析与行为修正之间的解耦关系。
链接: https://arxiv.org/abs/2609.04582
作者: Gnaneswar Villuri,Hashmath Shaik,Alex Doboli
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:A 0.6B language model, asked to verify 1,200 logical conclusions (half valid, half corrupted by a single semantic edit), answers YES every time. Judged by behavior it discriminates nothing; linear probes on its hidden states read the correct verdict at 0.96 AUC, transferring to unseen logical structures and separating foils built from exactly the words of the true conclusion (0.90). We ask where the verdict is lost, and find the dominant failure is a single scalar. The verdict survives to the model’s own output logits (margin AUC 0.89) along a well-aligned readout direction; a saturated decision threshold, offset by +4.6 sigma, erases it. The diagnosis generalizes: across 90 semantic-label configurations of a five-model, three-family factorial, behavioral accuracy collapses onto a single function of threshold offset (Spearman -0.93) while margin ranking moves far less. Across a 13x scale range, internal knowledge saturates while free-form behavior is non-monotone: an 8B model underperforms its 4B sibling through an answer-channel failure rather than the threshold; forced-choice accuracy is monotone. The diagnosis is actionable: a one-parameter correction, never fit on evaluated structures, repairs behavior from 50% to 81% (0.6B); calibrated margin decoding recovers 94% at 8B; few-shot prompting works the same way, recentering the threshold (+4.6 sigma to 0.0 sigma) while preserving ranking. Comparing probe to margin separates three regimes: concealed, miscalibrated, and undetected. On a maze task built so foils carry no surface cues, the audit correctly reports the third. In the standard generation setting, answer-surface features and heuristic labels reproduce published probing results without any internal access.
[NLP-53] Does the Selected Object Reach the Reader? Auditing Identity Handoffs in Grounded Language-Model Pipelines EMNLP2026
【速读】： 该论文旨在解决生成式语言模型在问答系统中因对象选择与信息检索环节间存在“手递手”（handoff）断裂而导致的性能下降问题，即在多阶段流水线中，若所选目标对象未能被后续检索模块正确识别或召回，将直接影响最终答案的准确性。其解决方案的关键在于构建并验证一个可复现、可审计的对象追踪框架——返回对象档案（Returned-Object Profile, ROP），该框架通过精确记录目标对象的标识符（returned-ID）、截断阈值（cutoff）、成员判定规则及完整预期集合，实现对每个查询中目标对象是否成功传递的细粒度追踪。实验表明，在329个可解析记录中，不同选择器间的对象一致性存在显著差异；尤其当采用仅基于正文的BM25检索时，有26.6%的案例在前五条结果中丢失目标对象，而融合重排序的混合检索则将遗漏率降至1.0%。此外，移除目标段落会显著降低准确匹配分数（下降28.6至31.0分），而移除长度相近的对比段落则无此效应，进一步证实了目标对象传递的重要性。研究还揭示原始问题排名下，前五名结果的一致性仅达94.1%，凸显了检索排序策略对对象保真度的影响。ROP的发布为评估和改进多阶段语言模型流水线中的对象保真度提供了可执行的基准工具。
链接: https://arxiv.org/abs/2609.04579
作者: Siddharth Vohra,Runmin Jiang,Xiaomo Li,Min Xu
机构: Carnegie Mellon University (卡内基梅隆大学); Amazon Web Services AI Native (亚马逊网络服务人工智能原生)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 15 pages, 1 figure, 23 tables. Accepted to the GroundLM Workshop (Grounding Language Models: Learning Faithfully and Efficiently) at EMNLP 2026
Abstract:Grounded language-model pipelines can be divided into three stages: selecting an object, retrieving passages for it, and using that evidence to answer. If the selected object must reach the reader, losing it breaks the handoff. Benchmark recall checks the dataset-linked object, which can differ. We audit 600 HybridQA questions across three selector families. On 1,463 resolvable records where the selected object matches the dataset-traced passage, exact key lookup and exact title matching return the object every time. With every ranked rule given the same decoded selected title, body-only BM25 omits it on 389 records (26.6%) at cutoff five, while hybrid retrieval with reranking omits it on 14 (1.0%). The two identities differ on 329 of 1,792 resolvable records. With original-question rankings, their top-five checks disagree on 106 records (5.9%). Frozen reader comparisons associate the aligned object’s presence with 28.6 to 31.0 points higher exact match. In a deliberately selected 64-item cohort, removing that passage sharply lowers exact match, while removing a similar-length comparison passage does not reproduce the drop. We release the Returned-Object Profile (ROP), an executable record of the target, returned-ID field, cutoff, membership rule, and complete expected population, with data and an offline replay.
[NLP-54] Extremely Sparse Supervision Incentivizes Reasoning Ability
【速读】： 该论文旨在解决大语言模型后训练（post-training）过程中普遍依赖海量文本标记（token）进行优化的效率问题，挑战了“有效学习必须依赖密集标记”的传统假设。其核心问题是：在生成式推理任务中，是否可能通过极少量的监督信号实现与全量标记训练相当甚至更优的推理能力提升？解决方案的关键在于提出一种基于在线策略蒸馏（on-policy distillation, OPD）的稀疏监督机制，仅对每条推理轨迹中的极少数关键生成标记（如1-2个token，占比仅0.05%）施加教师模型监督，其余大部分标记则被忽略。实验结果表明，这种极端稀疏的监督在数学推理、代码生成等多个任务及不同模型架构（包括Qwen3、Llama系列和基于近端策略优化的强化学习框架RLVR）中均能实现与全标记训练相当或更优的推理性能，且该现象具有高度一致性。这一发现揭示了人类认知中“反思关键步骤而非逐字修正”的学习模式在模型训练中的潜在有效性，从而挑战了传统后训练需大规模标记的范式，为设计更高效、更贴近自然学习过程的后训练算法提供了新方向。
链接: https://arxiv.org/abs/2609.04565
作者: Zhishuai Liu,Xingzi Xu,Mehmet Saygin Seyfioglu,Pan Xu,Karim Bouyarmane
机构: Amazon(亚马逊); Duke University(杜克大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Large language models demonstrate increasingly strong reasoning capabilities through effective post-training. Yet, prevailing post-training methods optimize over massive numbers of tokens, implicitly assuming that effective learning must be token-intensive. We revisit this assumption in the on-policy distillation (OPD) setting, which naturally admits dense teacher supervision at every generated token. Using the Qwen3 family, we discover a counter-intuitive phenomenon: reasoning can be effectively incentivized by an extremely small fraction of generated tokens–as few as one or two tokens per reasoning trajectory, corresponding to only 0.05% of all tokens. Surprisingly, this sparse supervision in most cases matches or surpasses full-token training in improving reasoning ability, despite excluding the vast majority of generated tokens from the training objective. This phenomenon is consistently observed across nine teacher–student configurations spanning different model scales on mathematical reasoning tasks, and is further validated on coding reasoning, Llama models and Proximal Policy Optimization (PPO)-based reinforcement learning with verifiable reward (RLVR). Interestingly, such extremely sparse supervision may be closer to the natural learning process: rather than correcting every step word by word, one reflects on a few critical reasoning steps, updates prior understanding, and continues the trial-and-error, avoiding micro-level corrections while remaining remarkably effective. Overall, our results challenge the assumption that effective post-training must be token-intensive and point to a new direction for understanding and designing more efficient post-training algorithms.
[NLP-55] Rhythms of Work: Multi-Scale Interpretation of Human Behavioral Traces for Workplace Agents
【速读】： 该论文旨在解决智能体在工作环境中对人类行为进行解释时所面临的挑战，即如何从海量、低层次的运行时事件（runtime traces）中提取有意义的行为模式。传统方法通常将用户行为压缩为单一嵌入或扁平化序列，隐含假设“用户行为的总结”存在唯一正确答案，但这种做法忽略了行为解释的多尺度特性。其解决方案的关键在于提出一种多分辨率（multi-resolution）的行为解释框架：通过构建一个语义归一化的多层次词汇体系，包括操作符（operator）、重复性模式（motif）、连贯事件段（episode）和日周期节律（day-level rhythm）等抽象层级，每个层级保留其对应时间尺度下的结构特征。该框架在包含6.67亿条人类标注事件的大规模商业生产力套件数据上进行了验证，识别出120种操作符类型、数千个模式、25种事件类型及五类日周期原型。实证结果表明，该表示具有良好的结构稳定性与预测有效性——在独立样本上可复现相同分类体系，并在预测用户下一阶段事件时相较单层基线提升17%的相对宏F1值。进一步的分辨率消融实验显示，不同问题需依赖不同时间粒度的抽象来最优回答。因此，论文主张行为解释应为多分辨率且查询条件化（query-conditioned），即智能体应根据具体问题动态选择最适配的时间粒度，而非依赖单一通用摘要。
链接: https://arxiv.org/abs/2609.04556
作者: Lin Ai,Scott Counts
机构: Microsoft(微软)
类目: Computation and Language (cs.CL)
备注:
Abstract:Runtime traces are becoming a central substrate for understanding agentic systems, yet interpretation has focused largely on what the agent did. Workplace agents face the complementary problem: interpreting the human activity that surrounds them. Hours of low-level events carry rich evidence about a user’s state but are too granular to reason over directly, and flattening them into one stream or compressing them into a single embedding both treat “summarize the user’s behavior” as if it had one correct answer. We argue instead that behavioral interpretation is resolution-dependent: the same trace should admit multiple addressable interpretations at different temporal resolutions. We construct a multi-resolution vocabulary of semantically normalized operators, recurring motifs, coherent episodes, and day-level rhythms, each preserving the structure salient at its own horizon. Applied to 667 million human-attributed events from a large commercial productivity suite (50,000 users, 100 organizations), it yields 120 operator types, thousands of motifs, 25 episode types, and five day-rhythm archetypes. We validate it on real telemetry: re-running the entire pipeline on a disjoint 2,000-user sample recovers the same taxonomy (structural stability), and on held-out users the full representation forecasts a user’s next episode more accurately than a flat-operator baseline, a 17% relative macro-F1 gain (predictive validity), so the abstractions preserve future-relevant information rather than merely describe it. A controlled resolution ablation then shows that no single level is optimal across questions: different agent-facing questions about the same trace are best answered at different resolutions. Behavioral trace interpretation for agents should therefore be multi-resolution and query-conditioned: an agent should access the temporal grain a question needs, not one universal summary.
[NLP-56] A Calibrated Reflection Approach for Enhancing Confidence Estimation in LLM s NAACL2025
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在实际部署中缺乏可靠置信度估计机制的问题，即如何准确判断模型输出的可信程度，从而决定系统是否应信任模型结果或转由人工干预。其解决方案的关键在于提出一种名为“校准式反思”（Calibrated Reflection）的框架，该框架通过融合结构化推理与距离感知校准技术，实现更精准的置信度评估。核心创新包括：(1) 最大置信度选择（Maximum Confidence Selection, MCS）方法，全面评估所有可能标签下的置信度；(2) 基于反思的提示机制，提升模型推理过程的可靠性；(3) 距离感知校准技术，显式建模标签间的序数关系，从而增强置信度估计的合理性。实验在HelpSteer2、Llama T-REx及专有对话数据集上验证了该框架在对话与事实分类任务中的有效性，为构建可信赖、良好校准的置信度估计方法提供了重要支持。
链接: https://arxiv.org/abs/2609.04539
作者: Umesh Bodhwani,Yuan Ling,Shujing Dong,Yarong Feng,Hongfei Li,Ayush Goyal
机构: Amazon.com(亚马逊); Seattle, Washington, USA(华盛顿州西雅图市, 美国)
类目: Computation and Language (cs.CL)
备注: Published at TrustNLP 2025 (NAACL 2025 Workshop)
Abstract:A critical challenge in deploying Large Language Models (LLMs) is developing reliable mechanisms to estimate their confidence, enabling systems to determine when to trust model outputs versus seek human intervention. We present a Calibrated Reflection approach for enhancing confidence estimation in LLMs, a framework that combines structured reasoning with distance-aware calibration technique. Our approach introduces three key innovations: (1) a Maximum Confidence Selection (MCS) method that comprehensively evaluates confidence across all possible labels, (2) a reflection-based prompting mechanism that enhances reasoning reliability, and (3) a distance-aware calibration technique that accounts for ordinal relationships between labels. We evaluate our framework on diverse datasets, including HelpSteer2, Llama T-REx, and a proprietary conversational dataset, demonstrating its effectiveness across both conversational and fact-based classification tasks. This work contributes to the broader goal of developing reliable and well-calibrated confidence estimation methods for LLMs, enabling informed decisions about model trust and human judgement.
[NLP-57] Scale-QLoRA: Code-Invariant Adapter Merging for Native 4-bit Microscaling LLM s
【速读】： 该论文旨在解决在4-bit量化微调模型（如NVFP4、MXFP4）中，将低秩适配器（LoRA）合并至基础模型时所面临的精度损失与部署灵活性下降问题。传统方法在合并过程中需通过量化器重推导离散的E2M1码平面（约占文件90%的字节），导致合并后的模型依赖特定量化格式，且后续任何对权重编码的修改都会引发性能波动，甚至造成高达39个百分点的适应性损失。其关键解决方案是提出Scale-QLoRA：仅对原始量化块中的缩放因子（scale field）进行适配，训练过程在部署网格上完成，并冻结所有E2M1码平面。在此固定量化格式、网格、块布局和码平面的前提下，合并操作变为比特级精确的恒等变换，使合并后模型具备代码不变性（code-invariant）。相比需通过量化器重推导码平面的QAT-LoRA，Scale-QLoRA保持码平面完全一致，从而避免了因量化规则不匹配导致的性能敏感性（如近似取整实现差异约1个百分点，极端不匹配可致性能降至0%），并显著降低训练开销（移除3.9倍于密集8B模型的权重空间直通估计器计算量），同时支持精确回滚、码平面去重及约125倍加速的仅缩放参数任务切换。
链接: https://arxiv.org/abs/2609.04526
作者: Tung-Ling Li,Jiale Huang,Lee-Chi Wang,Janaki Ram Gotei
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Merging a LoRA adapter into its base model is standard deployment practice: it removes the runtime adapter’s per-forward overhead and leaves a single standalone checkpoint any serving stack can load. On a native 4-bit microscaling checkpoint (NVFP4, MXFP4) that step stops being free. The merged weights must be written back through a quantizer, which re-derives the checkpoint’s discrete E2M1 code plane (roughly 90% of the artifact’s bytes), so the deployed artifact becomes coupled to one quantization convention, and every later code-touching event in its lifecycle can move it. Done naively the step is worse than fragile: it deletes the adaptation, by up to 39 pp, because against an already-on-grid base the reconstruction optimum is that base. Scale-QLoRA instead adapts only the native per-block scale field, trains those scales on the deployment grid, and freezes every E2M1 code. Within a fixed native format, scale grid, block layout and code plane, merging is then a bit-exact identity and the merged artifact is code-invariant. Across four models and four tasks, Scale-QLoRA and merge-aware QAT-LoRA are both accuracy-lossless, so we claim no accuracy ordering between them; they differ structurally, in that QAT-LoRA re-derives the code plane through a quantizer while Scale-QLoRA preserves it exactly. That difference is what the lifecycle prices: nearest-rounding implementations disagree by about a point on the measured task, and more extreme rule mismatches can drive the weight-space artifact to ~0%, which we report as a sensitivity bound rather than a deployment frequency. Preserving the code plane also drops the weight-space straight-through estimator from training (3.9x per step on the dense 8B model) and enables exact rollback, code-plane deduplication, and a ~125x faster scale-only task swap.
[NLP-58] LentEx: Generalizable Latent Entity Extraction via Synthetic Data and Instruction-Tuned LLM s IJCNN2025
【速读】： 该论文旨在解决隐式实体抽取（Latent Entity Extraction, LEE）中因标注数据稀缺而导致模型性能受限的问题。传统实体抽取方法难以捕捉文本中上下文推断出的抽象性、主题性隐式实体，而这类实体在检索增强生成（Retrieval-Augmented Generation, RAG）、客户画像分析和知识图谱增强等应用中至关重要。本文提出的LentEx框架通过基于模板的合成数据生成与指令微调相结合的方法，有效提升了小型高效大语言模型（LLMs）在LEE任务上的表现。其解决方案的关键在于：利用模板构建多样化且上下文丰富的合成数据，以弥补真实标注数据的不足，并通过指令微调使模型更好地理解隐式实体的语义模式，从而实现对未见领域的强泛化能力。实验表明，LentEx在MTEB聚类基准上显著超越现有最优模型，为自然语言处理中的隐式实体理解与提取建立了新范式。
链接: https://arxiv.org/abs/2609.04511
作者: Umesh Bodhwani,Yuan Ling,Cibi Chakravarthy Senthilkumar,Shujing Dong,Yarong Feng,Hongfei Li,Ayush Goyal
机构: Amazon(亚马逊)
类目: Computation and Language (cs.CL)
备注: Published in IJCNN 2025. ©2025 IEEE
Abstract:Latent entity extraction (LEE) tackles the challenge of identifying implicit, contextually inferred entities within free text-an area where traditional entity extraction methods fall short. In this paper, we introduce LentEx, a novel framework for latent entity extraction that leverages synthetic data generation and instruction fine-tuning to optimize smaller, efficient large language models (LLMs). Latent entities, which are often abstract and thematic, are crucial for applications such as retrieval-augmented generation (RAG), customer persona analysis, and knowledge graph enrichment. LentEx addresses the scarcity of labeled datasets by employing a template-based approach to generate diverse, contextually rich synthetic data, ensuring high variability and alignment with real-world distributions. To our knowledge, LentEx is the first to systematically approach LEE through the lens of LLMs. LentEx demonstrates significant performance improvements across multiple tasks, notably surpassing state-of-the-art models on the MTEB Clustering Benchmark. Furthermore, our methodology enables robust generalization to unseen domains, making LentEx highly applicable in real-world NLP tasks, including RAG and clustering, thereby establishing a new paradigm for latent entity understanding and extraction in natural language processing.
[NLP-59] Rethinking Indirect Prompt Injection as a Test-Time Search Problem
【速读】： 该论文旨在解决生成式 AI 系统在实际应用中面临的间接提示注入（indirect prompt injection）安全威胁问题，其核心挑战在于攻击者如何利用环境、用户任务与注入任务共同构成的任务相关攻击面，在推理时动态搜索并实施有效攻击。解决方案的关键在于提出一种具有专用搜索框架的代理型攻击者（agentic attacker），该攻击者能够执行环境侦察、基于结构化推理的攻击策略规划以及利用目标代理反馈进行自适应评估。研究发现，随着攻击者推理时计算资源的增加，漏洞发现与利用能力显著提升；而消融实验表明，显式的策略管理对于避免冗余搜索、在大规模计算预算下保持性能增益至关重要。这表明，传统的将攻击成功率视为与计算预算无关的受害者固有属性的评估方式存在缺陷，应改以系统性刻画攻击者的搜索过程及其计算开销作为安全评估的核心。更广泛地，本研究揭示了工具使用型智能体在面对攻击者自适应搜索其所处系统攻击面时所面临的重要且被忽视的安全风险。
链接: https://arxiv.org/abs/2609.04495
作者: Duong M. Nguyen,Joon Sik Kim,Blazej Manczak,Vaikkunth Mugunthan
机构: Dynamo AI; University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR)
备注:
Abstract:We formulate indirect prompt injection as a test-time search over a task-dependent attack surface induced by the environment, user task, and injection task. To operationalize this formulation, we introduce an agentic attacker with a dedicated search harness that performs environment reconnaissance, structured reasoning over attack strategies, and adaptive evaluation using victim-agent feedback. Across heterogeneous tasks, we find that increasing attacker test-time compute improves vulnerability discovery and exploitation, while ablations show that explicit strategy management is important for avoiding redundant search and sustaining gains at larger budgets. These results suggest that agentic security evaluations should characterize both the attacker’s search procedure and compute budget, rather than treating attack success as a budget-independent property of the victim. More broadly, our findings identify the attacker’s adaptive search over the system attack surfaces as an important and underexplored security risk for tool-using agents.
[NLP-60] owards Understanding Pause Token Fine-Tuning Dynamics: A Mode Retention Perspective
【速读】： 该论文旨在解决生成式 AI（Generative AI）在大语言模型（LLM）推理过程中，如何有效平衡“记忆保持”与“适应性调整”之间的权衡问题。现有基于暂停标记（pause-token）的方法虽被证实可提升模型推理能力，但其背后的训练动态机制尚不明确。本文提出的核心解决方案是引入一种名为“掩码边界暂停”（Masked Boundary Pause, MBP）的训练策略：将暂停标记置于推理步骤边界，并对其损失进行掩码处理，从而在微调过程中抑制对先前知识的过度覆盖，同时增强对后续推理步骤的信息编码。实验表明，该方法在合成连续学习任务中显著降低对已有分布的覆盖程度（约4倍），并在数学推理任务中使相邻边界标记蕴含更丰富的下游步骤信息，体现出非短视压缩（non-myopic compression）特性。在多个1B至8B规模的Qwen与Llama模型上，MBP consistently 提升了数学与代码推理性能（最高达6分和2.5分），且未损害通用语言理解能力。进一步验证显示，该策略可扩展至GRPO等强化学习框架。因此，本研究将暂停标记从传统意义上的推理时计算装置，重新定义为一种针对训练动态的干预手段，通过调控记忆保留与适应性更新的平衡，实现更优的推理表现。
链接: https://arxiv.org/abs/2609.04489
作者: Jaehyeon Kim,Suhwan Kim,Nakyung Lee,Yeongoon Kim,Jimin Seo,Giho Lee,Jungwoo Lee
机构: HodooAI Lab(霍多AI实验室); Seoul National University (首尔国立大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 24 pages, 4 figures, 19 tables
Abstract:Pause-token methods improve LLM reasoning by inserting special tokens into sequences. Prior work explains these gains through computational expressivity. However, there is relatively little investigation into the training dynamics of pause tokens. We explore how pause tokens reshape the training dynamics of fine-tuning. Two controlled pilots expose distinct asymmetries. On a synthetic continual-learning task, masked pauses overwrite a previously-learned distribution roughly 4x less at matched final adaptation (H1, mode retention); on a synthetic math-reasoning probe, the boundary-adjacent token comes to encode substantially more downstream-step information (H2, non-myopic compression). We formalize a training rule consistent with both - Masked Boundary Pause (MBP), pause tokens placed at reasoning-step boundaries with their loss masked. Across 1B-8B Qwen and Llama models, MBP consistently improves reasoning, achieving gains of up to 6 points on math and 2.5 points on code, while preserving general language understanding abilities. We further demonstrate that this mode-preserving strategy extend gains to GRPO. These results recast pause tokens as a training-dynamics intervention on the retention-adaptation trade-off, rather than merely an inference-time computation device.
[NLP-61] Uncertainty Signals for Network Intent Translation: Risk Ranking and Ambiguity Localization
【速读】： 该论文旨在解决生成式意图网络（Intent-based Networking）中，基于大语言模型（LLM）生成的网络配置在部署前存在的潜在风险问题。尽管现有方法在意图到配置的翻译准确率方面取得进展，但普遍忽视了配置部署后可能引发的不可预测性与安全隐患。其解决方案的关键在于引入两种不确定性信号：基于采样的预测不确定性（sampling-based predictive uncertainty），用于对翻译结果进行风险排序；以及基于词元级别的熵（token-level entropy），用于定位歧义来源。实验基于微调后的Llama-3.1-8B-Instruct模型，在特定厂商交换机平台（Juniper EX3300）上构建的可控歧义测试集上进行评估，结果表明，预测不确定性在不同上下文类型和采样预算下均能有效支持风险排序，但在信息量较低的上下文中存在显著校准偏差；同时，参数词元熵与参数相关歧义显著相关，关键词元熵则与描述相关歧义相关。这些发现表明，将不确定性信号嵌入配置部署流程具有可行性——预测不确定性可用于实现选择性部署，而词元级熵可辅助识别歧义源头，从而提升生成式网络配置的安全性与可解释性。
链接: https://arxiv.org/abs/2609.04486
作者: Ala’ A. Alsamarneh,Omar Alhussein
机构: Khalifa University(哈利法大学)
类目: Networking and Internet Architecture (cs.NI); Computation and Language (cs.CL)
备注:
Abstract:Intent-based networking realization starts by translating high-level intents into low-level network configurations. Recent approaches have shifted toward LLM-based translation. Despite promising results, most studies focus on translation accuracy and overlook risks associated with deploying the resulting configurations. In this work, we investigate the pre-deployment translation risk of LLM-generated configurations by analyzing the model’s uncertainty. We propose to use two uncertainty signals, namely sampling-based predictive uncertainty for translation-risk ranking and token-level entropy for ambiguity-source localization. We evaluate these signals on an ambiguity-controlled test set across different context types and sampling budgets, using a Llama-3.1-8B-Instruct model fine-tuned for intent translation on a vendor-specific switch platform (Juniper EX3300). The results demonstrate that predictive uncertainty provides a useful signal for ranking translations by risk across context types and sampling budgets, albeit with substantial miscalibration under less informative contexts. Moreover, we show that parameter-token entropy correlates with parameter-sourced ambiguity and keyword-token entropy correlates with description-sourced ambiguity. These results indicate the potential of using uncertainty signals in an LLM-generated configuration deployment pipeline, where predictive uncertainty can support selective deployment, while token-level entropy can identify sources of ambiguity.
[NLP-62] Cultural Misalignment in Large Language Models : Detection Measurement and Mitigation Through Targeted Fine-Tuning
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在跨文化语境下存在的认知偏见问题，特别是模型对不同国家与人口统计群体态度分布的建模偏差。研究聚焦于三个开源大模型（Gemma3-12B、Bielik-11B-v3、Qwen3-4B）在世界价值观调查（World Values Survey, WVS）第七波数据中对63个细分人口画像（demographic personas）的态度预测准确性，采用归一化Wasserstein距离量化模型输出分布与真实社会分布之间的错位程度。关键发现是：尽管模型源自特定国家，但其偏见并不呈现“本土偏好”，反而中国开发的Qwen3-4B在对中国人群体建模时表现出最高的分布错位（W1 = 0.436）。为缓解此问题，研究提出针对最差表现的五个目标人口画像进行轻量级LoRA微调，仅需少于1,200个训练样本及单块GPU上不足15分钟即可完成。结果显示，该方法使Bielik-11B模型的偏见降低16.8%（p_Bonf = 0.002，d = -4.4），且所有目标群体均得到改善。然而，国家层面的分解分析揭示，微调并未消除偏见，而是将其重新分配——原为美国老年人群的最差表现群体，在微调后完全转变为中国的老年人群，且前后群体间无重叠。因此，该研究的核心解决方案在于通过针对性的、低资源消耗的LoRA微调策略实现对最差案例人口画像的精准修正，但同时也揭示了当前方法在根本性消除跨文化偏见方面的局限性，即偏见的“转移”而非“消解”。本研究是首个将LoRA微调应用于最差案例人口画像以实现跨文化偏见缓解的实证探索。
链接: https://arxiv.org/abs/2609.04485
作者: Antoni Czolgowski,Abel Iyasele
机构: University of Colorado Boulder(科罗拉多大学博尔德分校); Boulder, CO 80309, USA
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 33 pages, 14 figures. Extended version of a paper published in the proceedings of OSSConf 2026, Zilina, Slovakia. Code and data: this https URL
Abstract:We evaluate three open-weight LLMs (Gemma3-12B from the USA, Bielik-11B-v3 from Poland, and Qwen3-4B from China) against World Values Survey Wave 7 data for 63 demographic personas across three countries, using normalized Wasserstein distance to quantify distributional misalignment. Contrary to expectations, no model favors its home country: the Chinese-built Qwen3-4B performs worst on its own Chinese population (W1 = 0.436, the highest misalignment in the entire model x country matrix). Targeted LoRA fine-tuning on the five worst-case personas, requiring fewer than 1,200 training pairs and under 15 minutes on a single GPU, reduces bias by 16.8% for Bielik-11B (p_Bonf = 0.002, d = -4.4) with all five targets improving. However, country-level decomposition reveals that fine-tuning redistributes rather than removes bias: Bielik’s worst-case personas swap entirely from American to Chinese elderly, with zero overlap between pre- and post-correction sets. To our knowledge, this is the first study to target worst-case demographic personas with LoRA fine-tuning for cross-cultural bias mitigation.
[NLP-63] Patterns of Priming in Production: Lexical Semantic and Structural Alignment in Language Model Generation EMNLP2026
【速读】： 该论文旨在解决生成式语言模型（Generative AI）在句子生成过程中是否表现出结构启动效应（structural priming）这一问题，尤其关注先前的句法结构如何影响后续句子的完成。尽管已有研究证实了人类在理解阶段存在结构启动现象，但在生成任务中，由于语言模型在每一步需从大量可能的延续中采样，其是否存在类似效应尚不明确。本文通过一系列针对双宾语结构（dative constructions）的受控句末补全实验，发现语言模型确实对结构启动敏感，且该效应在语义连贯的句子中尤为显著。关键发现在于：虽然相对于基线，双宾结构的相对启动幅度更大，符合逆频率效应（inverse frequency effect），但预置介词宾语结构（更常见的句型）在绝对启动量上增长更多；此外，不仅语义-词汇一致性增强了结构启动效应，且被启动的生成结果还表现出更高的词汇与语义重复水平。这表明，语言模型中的结构启动并非单一层面的机制，而是跨句法、词汇和语义层级协同作用的结果，促进了多层表征间的对齐与协调。
链接: https://arxiv.org/abs/2609.04484
作者: Giulia Pucci,Ruizhe Li,Arabella Sinclair
机构: University of Aberdeen(阿伯丁大学); University of Birmingham(伯明翰大学); University College London(伦敦大学学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Findings
Abstract:This paper investigates structural priming in language model (LM) production, examining how preceding structural context influences sentence completion. While prior work has demonstrated priming effects in comprehension of structural alternations, it remained unclear whether these persist in production, where, when generating, an LM samples from many possible continuations at each step. We address this question through a series of controlled sentence-completion experiments on dative constructions. In line with prior work, we find that LMs are susceptible to structural priming, particularly in sentences that are semantically coherent. In terms of priming magnitude, we find that while there is a greater relative increase of double-object datives against our baselines, in line with inverse frequency effects, there is a larger absolute increase in prepositional-objects, the more frequently produced construction. Finally, we not only observe that structural priming is boosted by lexico-semantic coherence, but that structurally primed completions display greater levels of lexico-semantic repetition. Taken together, our evidence supports the view that structural priming in LMs operates across multiple levels of linguistic representation, facilitating, and facilitated by syntactic, lexical, and semantic alignment. Code: this https URL.
[NLP-64] Safety for Whom? Boundary-Aware Self-Distillation for Controlled LLM Safety Refusal
【速读】： 该论文旨在解决生成式AI在实际部署中面临的“窄边界安全”（narrow-boundary safety）问题，即在同一主题下，不同应用场景需设定差异化的安全边界——例如，在政治相关问答中既要防止有害内容输出，又需保持对事实性问题的响应能力。传统安全对齐方法通常以粗粒度话题级判断为基础，难以满足特定场景下的精细化控制需求。其核心解决方案是提出一种离线自生成框架，整合受控主题生成、覆盖率修复、分布内补偿数据以及有害-良性配对样本，用于训练与评估。通过引入“逐步升级”（Escalate）机制实现多轮拒绝生成，显著提升目标领域拒绝率（从9.47%增至84.75%），并大幅降低跨多个有害性基准的平均不安全响应率（从26.26%降至0.14%）。然而，该方法也导致过量拒绝（over-refusal）上升，表明数据构成直接影响安全与可用性之间的权衡。进一步实验显示，使用经验证的目标模型响应替代外部响应可有效抑制过量拒绝（从15.20%降至5.20%），而边界对数据则显著降低合规侧过量拒绝（从32.94%降至4.16%），同时维持较高有害侧拒绝率。研究结果强调，安全对齐必须在拒绝边界的两侧进行系统评估，且数据设计是调控安全与可用性平衡的关键因素。
链接: https://arxiv.org/abs/2609.04482
作者: Alejo López-Ávila,Iker García-Ferrero,Jezabel Garcia,Antonio Tiene,Román Orús
机构: Multiverse Computing
类目: Computation and Language (cs.CL)
备注: 22 pages, 14 figures
Abstract:Safety alignment is usually posed as a topic-level question: is this subject harmful? Deployments ask a narrower one. A civics tutor and a public-sector assistant may share a base model yet need different boundaries inside the same topic, refusing targeted political manipulation while still answering factual questions about the same election. We formulate this as narrow-boundary safety and introduce an offline self-generated framework combining controlled topic generation, coverage repair, in-distribution compensation data, and harmful-benign pairs for training and evaluation. Single-shot generation leaves 19.88% of prompts without accepted refusal traces, whereas escalating retries leave 0.20%. On political persuasion with Qwen3-8B, training on refusal data completed through Escalate increases target-domain refusal from 9.47% to 84.75% and reduces the mean unsafe-response rate across three broader harmfulness benchmarks from 26.26% to 0.14%, but increases XSTest over-refusal from 2.00% to 74.00%. In a separate matched comparison, replacing external responses with verified target-model responses reduces over-refusal from 15.20% to 5.20%. Boundary-pair data reduces comply-side over-refusal on held-out pairs from 32.94% to 4.16%, while harmful-side refusal decreases only from 91.88% to 87.72%. These results show that data composition controls the safety and usability trade-off, and that safety alignment should be evaluated on both sides of the intended refusal boundary.
[NLP-65] Shared circuits predict whether LLM s generalize across formats in arithmetic reasoning
【速读】： 该论文旨在解决大语言模型（LLM）在面对输入格式表面变化时泛化能力不足的问题，特别是针对同一算术问题在数值表达（如2+5）与文字表达（如“two plus five”）之间表现差异显著的现象。其核心挑战在于：尽管人类能够轻松实现跨格式的推理迁移，而当前模型在处理文字形式的算术问题时准确率明显下降。论文提出的解决方案关键在于利用归因打补丁（attribution patching）技术，从模型内部识别并定位其在解决数值型算术问题（2+5）与多语言文字型算术问题（英语、西班牙语、意大利语）时分别激活的神经电路；进而通过分析这些电路间的重叠程度，预测模型对不同文字表达形式的泛化性能。研究发现，模型自身数值电路与文字表达任务所激活电路之间的重叠度，能够有效解释三种语言文字表达的相对难度，并准确预测哪些具体题目能被正确解答，其表现可媲美有监督探测方法，且无需任何标注数据。这一结果表明，模型内部结构的可解释性特征可作为泛化能力的预测指标。
链接: https://arxiv.org/abs/2609.04463
作者: Andrea Gregor de Varda,Sana Pandey,Pengrui Han,Jacob Andreas,Evelina Fedorenko
机构: Massachusetts Institute of Technology (麻省理工学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:In many forms of reasoning, including arithmetic reasoning, generalizing across superficial changes in input format is effortless for humans: anyone who can solve 2+5 can also solve ‘two plus five’. In contrast, LLMs are more brittle to surface variations of the prompts: for example, they solve numeric arithmetic problems almost perfectly but are substantially less accurate on verbal renditions of the same problems. Here, we ask whether generalization across formats can be predicted from the models’ internals. Using attribution patching, we first independently localize the circuit that each model recruits to solve numeric arithmetic problems (2+5) vs. verbal ones, in three languages: English (‘two plus five’), Spanish (‘dos más cinco’), and Italian (‘due più cinque’); then, we test whether overlap with the model’s own numeric circuit predicts its generalization to the verbal formats. Indeed, we find support for this idea at three levels: circuit overlap accounts for the relative difficulty of the three verbal formats, for which models generalize best, and for which items are solved correctly, rivaling supervised probes while requiring no labeled data.
[NLP-66] When Load-Balancing Goes Too Far: Expert Pruning in Over-Dispersed Mixture-of-Experts Models
【速读】： 该论文旨在解决在负载均衡训练下（即过分散路由，over-dispersed routing）专家模型（MoE）中，传统基于路由器概率的专家剪枝方法失效的问题。其核心问题是：在过分散路由条件下，令牌在专家间近乎均匀分布，导致路由器输出的概率信号失去重要性判别能力，进而使困惑度（perplexity）无法有效预测下游任务性能，甚至出现低困惑度配置反而损害数学推理能力的现象。此外，不同剪枝评分指标（如激活感知与频率依赖）在不同任务域间存在显著的能力权衡，难以找到普适最优方案。针对这一问题，论文提出最小最大专家评分分配（Minimax Expert Score Allocation, MESA），其关键在于引入领域感知机制，通过迭代提升当前表现最差领域的专家重要性评分，以最小化最坏情况下的领域性能退化，而非单纯优化平均准确率。实验表明，在25%专家剪枝率下，MESA在11个基准测试中有7个优于激活感知基线，同时显著降低内存开销，并具备在gpt-oss-120B、Gemma-4-26B-A4B和OLMoE-1B-7B等多规模模型上的良好泛化能力。研究揭示了过分散路由是一种与标准路由有本质区别的剪枝范式，强调识别该特性是实现负载均衡型MoE模型合理剪枝的前提。
链接: https://arxiv.org/abs/2609.04453
作者: Berkcan Kapusuzoglu,Connor Pryor,Sangwoo Cho,Supriyo Chakraborty,Shi-Xiong Zhang,Sambit Sahu,Milind Naphade
机构: Capital One, AI Foundations(资本壹号，人工智能基础); McLean(麦克莱恩), VA(弗吉尼亚州), 22102, USA(美国)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 22 pages, 7 figures. Preprint
Abstract:Expert pruning reduces the memory and serving cost of Mixture-of-Experts (MoE) models by removing low-importance experts identified by the router, assuming router probabilities provide a reliable importance signal. We observe that this assumption breaks down under over-dispersed routing, a regime associated with aggressive load-balancing during training, in which tokens are distributed nearly uniformly across experts and importance signals collapse. In this regime, perplexity does not predict downstream task accuracy: on gpt-oss-20B, the lowest-perplexity pruning configuration yields the worst mathematical reasoning, while the highest-perplexity configuration preserves it. This does not occur under standard routing (e.g., Mixtral-8x7B-Instruct), where perplexity and accuracy degrade together. Pruning under over-dispersed routing also exposes a capability trade-off in which no single scoring metric dominates: activation-aware scoring preserves mathematical reasoning but severely degrades knowledge-intensive science (an 18-point gap on GPQA), whereas frequency-based scoring exhibits the reverse. We propose Minimax Expert Score Allocation (MESA), a domain-aware method that iteratively boosts importance scores for experts serving whichever domain is currently worst-affected, minimizing worst-case domain degradation rather than average accuracy. At 25% expert pruning MESA achieves the smallest worst-case degradation across domains, outperforming activation-aware baselines on 7 of 11 benchmarks at a correspondingly reduced memory footprint, and it generalizes to gpt-oss-120B, Gemma-4-26B-A4B, and OLMoE-1B-7B. Our results indicate that over-dispersed routing is a qualitatively distinct pruning regime in which standard assumptions fail, and that recognizing it is a prerequisite for principled expert pruning of load-balanced MoE models.
[NLP-67] RILOGUE: A Trilingual Spoken Dialogue Fact-Checking Benchmark with Evidence and Paired Audio EMNLP2026
【速读】： 该论文旨在解决当前事实核查系统在处理口语化对话场景时的局限性，特别是针对多语言、跨话语轮次中信息碎片化、依赖上下文以及自动语音识别（ASR）错误导致文本失真等问题。现有资源普遍规模小、以英语为主或侧重标注而非端到端评估，缺乏大规模、多语言且具备逐轮标签与音频-文本对齐的基准数据集。为此，本文提出TRILOGUE（三语口语对话事实核查）基准，涵盖英语、俄语和哈萨克语三种语言，包含近12,000段对话、18.7万条话语轮次及390小时的音频-转录本配对数据，其中近5,000段为人工录制的俄语与哈萨克语对话。该基准支持声明可信性检测、源文章证据检索与事实验证任务，采用仅声明输入、黄金证据输入及检索证据输入等多种设置。实验表明，ASR退化与跨语言迁移仍是主要挑战，尤其在低资源语言如哈萨克语上表现更差，但通过检索来源证据可显著缩小与黄金证据验证性能之间的差距，凸显了有效证据获取在提升口语对话事实核查效果中的关键作用。
链接: https://arxiv.org/abs/2609.04452
作者: Chaewan Chun,Meruyert Aristombayeva,Jiyoung Choi,Mahjabin Nahar,Delvin Ce Zhang,Dongwon Lee
机构: The Pennsylvania State University (宾夕法尼亚州立大学); Satbayev University (萨特巴耶夫大学); University of Sheffield (谢菲尔德大学)
类目: Computation and Language (cs.CL)
备注: To appear in EMNLP 2026
Abstract:Modern misinformation is often heard before it is read, yet fact-checking systems are still evaluated mainly on clean written claims. Spoken dialogue remains different even when systems operate on transcripts: claims may be distributed across speakers and turns, depend on prior context, and become harder to verify when Automatic Speech Recognition (ASR) errors distort the available text. Prior spoken dialogue fact-checking resources are small, English-centric, or focused on annotation rather than end-to-end benchmarking, leaving no large multilingual benchmark with paired speech and turn-level labels. We introduce TRILOGUE (TRIlingual spoken diaLOGUE fact-checking), a large-scale trilingual benchmark of source-grounded spoken dialogues in English, Russian, and Kazakh. It contains nearly 12K dialogues, 187K turns, and 390 hours of paired audio with ASR transcripts and word-level timestamp alignments across all three languages, including nearly 5K human-recorded Russian and Kazakh dialogue files. TRILOGUE supports claim check-worthiness detection, source-article evidence retrieval, and claim verification with claim-only, gold-evidence, and retrieved-evidence inputs. Baselines show that ASR degradation and cross-lingual transfer remain challenging, especially for Kazakh, while retrieved source evidence substantially narrows the gap to gold-evidence verification.
[NLP-68] Conformity Breaks Conformal Prediction
【速读】： 该论文旨在解决多智能体大语言模型（LLM）系统中基于置信度校准的合流预测（conformal prediction）在面对同伴一致性误导时失效的问题。其核心问题是：当大模型独立作答时，生成的合流证书（conformal certificate）具有良好的覆盖率与校准性，但一旦模型观察到其他智能体一致输出错误答案，其对正确答案的评分机制发生隐性偏移（即“评分机制转移”，score-mechanism shift），导致原本有效的证书失效。解决方案的关键在于识别并应对这种由同伴压力引发的模型评分行为变异——传统校准方法无法修复此问题，因为问题根源并非输入分布变化，而是模型在群体交互下的内部评分逻辑发生了改变。实验表明，在标准显著性水平α=0.10下，覆盖率从校准状态下的90%骤降至74%，且对低置信度样本的覆盖能力几乎被削减一半（87%→47%），更严重的是，系统可能错误地从“不确定”转变为“自信决策”，从而采纳攻击者引导的错误答案。这揭示了现有合流预测框架在多智能体环境中的根本脆弱性。
链接: https://arxiv.org/abs/2609.04445
作者: Yibo Hu,Hanyu Su
机构: Illinois Institute of Technology(伊利诺伊理工学院)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 19 pages, 6 figures, 11 tables. Code: this https URL
Abstract:A conformal certificate can be valid when an LLM answers alone and invalid when the same LLM sees peers that unanimously assert a wrong answer. The question is unchanged; the model’s score for the correct answer changes. We call this a score-mechanism shift: clean calibration certifies how the model scores answers alone, but not how it scores them under peer pressure. We show that this shift silently breaks conformal prediction in multi-agent LLM systems. Across open-weight models and multiple-choice QA tasks, coverage falls from a calibrated 90% to 74% under unanimous-wrong peers at the standard alpha = 0.10 operating point. The average hides a sharper failure: by targeting the low-confidence items the certificate still covers, an attacker nearly halves coverage on that subgroup, from 87% to 47%, while the monitored average remains much higher. The failure also reaches the decision layer: a system that should escalate when uncertain can instead become confident enough to act on the attacker’s wrong answer. Standard conformal fixes do not solve the problem, because the question distribution has not changed; the model’s scoring behavior has.
[NLP-69] GRACE: Graph-Grounded Reflective Agent Copilot Engine for Expert-in-the-Loop Knowledge Expansion EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在高风险场景中生成看似合理但缺乏事实依据的虚假陈述（hallucination）问题。现有检索增强生成（Retrieval-Augmented Generation, RAG）方法受限于仅检索孤立文本片段，无法追踪跨文档证据之间的关联或量化不确定性。为此，论文提出GRACE（Graph-grounded Reflective Agent Copilot Engine）框架，其核心在于将LLM生成的内容分解为原子级声明，并在加权二分图结构中将其与可信知识先验进行对齐，图中边权重表征各声明与先验之间的接近程度。通过加权中心性分析，可将声明分类为“有根据”（Grounded）、“被驳回”（Refuted）或“边界”（Boundary），从而不仅识别幻觉，还可发现模型知识前沿中的新见解或存在争议的主张。为高效分配人工或代理资源，引入“注意力回报率”（Return on Attention, RoA）优化目标，仅当声明的优先级加权不确定性超过验证成本时才触发专家审查。经专家验证的声明被提升为新的证据锚点，形成“验证者-语言模型”演化的闭环，持续扩展知识库。实验表明，该框架在多个语言模型及通用与领域特定数据集上均显著优于传统RAG基线，且能有效筛选具有价值的边界知识供专家介入。研究结果表明，基于图结构的知识表示与人机协同验证机制可在系统层面缓解幻觉问题，而非局限于生成过程本身。
链接: https://arxiv.org/abs/2609.04442
作者: John Seon Keun Yi,Joshua R. Minot,Dokyun Lee
机构: Boston University (波士顿大学); MassMutual (马萨诸塞互助保险公司)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: AKBC Workshop @ EMNLP 2026
Abstract:Large language models deployed in high-stakes settings frequently generate plausible but ungrounded claims. Standard retrieval-augmented generation (RAG) pipelines offer limited remedy, since they retrieve isolated passages without tracking cross-document evidence relationships or quantifying uncertainty. We introduce GRACE (Graph-grounded Reflective Agent Copilot Engine), a framework that deconstructs LLM responses into atomic claims and grounds them against trusted knowledge priors within a weighted bipartite graph. Edge weights encode the closeness of each claim to the priors, enabling weighted centrality analysis that classifies claims as Grounded, Refuted, or Boundary. Such classification identifies not just hallucinations but also novel or contested claims at the frontier of the model’s knowledge. To efficiently allocate human or agent resources, we formulate a Return on Attention (RoA) objective that defers a claim to expert review only when its priority-weighted uncertainty exceeds the cost of verification. Claims verified by experts are promoted to new evidence anchors, closing a validator-LLM evolutionary loop that expands the knowledge base across iterations. We evaluate GRACE across multiple language models and on datasets spanning both general and domain-specific knowledge. Our results show that our knowledge base serves as a reliable foundation for retrieval that outperforms RAG baselines, and that the RoA framework efficiently selects valuable boundary knowledge for expert verification. These findings demonstrate that graph-structured representations combined with expert-in-the-loop verification can mitigate hallucination at the system level rather than at the generation level. Code available at this https URL
[NLP-70] What Attention Recalls and Recurrence Controls in Hybrid Language Models EMNLP2026
【速读】： 该论文旨在解决混合语言模型中注意力机制（attention）与固定大小循环状态（recurrent state）各自功能分工不明确的问题。其核心解决方案在于引入两种缓存层级的干预手段：一是“分段预填充”（Split-prefill），即在预填充阶段仅保留键值缓存（KV cache）或仅保留循环状态，随后生成回答；二是“状态交换”（State-swap），即将一个上下文的KV cache与另一个上下文的循环状态在单次前向传播中配对使用。实验结果表明，在Qwen3.5和Falcon-H1模型上，两通道的功能呈现显著分离：精确信息检索仅通过注意力机制维持（保持全模型准确率的64%-98%），而通过循环状态则完全失效；相反，语言风格与角色设定等生成性特征在循环状态中得以保留（分别达到70%-80%和3-5倍提升），而仅依赖KV缓存时语言准确率降至约1%。状态交换实验进一步证实了因果关系：生成内容的语义来自KV缓存侧，而语言风格则由循环状态决定。此外，仅基于循环状态的生成仍能输出未在上下文中出现但语义相关或具有部分共现性的词汇，表明注意力机制提供的是对历史输入的精准查找，而循环状态则负责塑造后续输出的语言表达方式。
链接: https://arxiv.org/abs/2609.04434
作者: Kirill Afendulev,Alexey Dontsov,Elena Tutubalina,Anton Korznikov
机构: YSDA(俄罗斯科学院数据科学中心); AIRI(人工智能研究学院); HSE University(高等经济大学)
类目: Computation and Language (cs.CL)
备注: Accepted to Findings of EMNLP 2026. 13 pages, 3 figures, 8 tables. Code: this https URL
Abstract:Hybrid language models combine attention with a fixed-size recurrent state, but the role of each channel remains unclear. We introduce two cache-level interventions. Split-prefill keeps only the KV cache or only the recurrent state from a prefilled context, then generates an answer. State-swap pairs the KV cache from one context with the recurrent state from another in a single forward pass. On Qwen3.5 and Falcon-H1, the two channels split sharply by function. Exact retrieval survives only through attention (64-98% of full accuracy) and collapses to zero through recurrence. Output language and persona reverse the pattern: both survive recurrence (70-80% and 3-5x) while KV-only drops to ~1% language accuracy. State-swap confirms this causally: the answer takes its value from the KV side and its language from the recurrent side. Recurrent-only generation also accepts words that were never in the context but share meaning or parts with seen items. Attention provides a lookup over what was said; the recurrent state shapes how the model says it next.
[NLP-71] A Systematic Evaluation of Cross-Lingual Consistency Enhancement Methods in Multilingual Language Models
【速读】： 该论文旨在解决多语言语言模型在跨语言问答中对语义等价问题产生不一致回答的问题，即跨语言一致性（Cross-lingual Consistency, CLC）问题。现有方法因采用不同的模型、任务和评估协议，导致其相对性能难以比较。为此，本文提出一个统一的评估框架，系统性地评测了代表性CLC增强方法在推理时干预与后训练两类策略的表现，覆盖三个模型家族及三个封闭形式基准数据集。研究发现，后训练方法总体更可靠，其中直接分布对齐（direct distribution alignment）在所有模型-数据集组合中均能稳定提升CLC；而其他方法则对答案格式和语言覆盖范围更为敏感。值得注意的是，跨领域迁移效果受限，仅当源任务与目标任务具有相似输出格式时才有效。此外，研究进一步考察了CLC增强是否损害模型在文化相关问题上应有的差异化响应能力：在受控的封闭形式评估中未发现系统性性能下降，但在开放式生成任务中，非英语回答的准确性偶尔出现降低。因此，该研究强调未来CLC增强方法需同时兼顾跨领域鲁棒性与文化适切性差异表达能力，为后续后训练策略与基准测试设计提供重要指导。
链接: https://arxiv.org/abs/2609.04409
作者: Jirui Qi,Mingyang Wang,Hinrich Schütze,Raquel Fernández,Arianna Bisazza
机构: University of Groningen(格罗宁根大学); LMU Munich(慕尼黑路德维希-马克西米利安大学); University of Amsterdam(阿姆斯特丹大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Preprint. All code and datasets will be released upon publication
Abstract:Multilingual language models often produce inconsistent answers to semantically equivalent questions across languages, motivating methods to improve cross-lingual consistency (CLC). However, existing methods are typically evaluated using different models, tasks, and protocols, leaving their relative strengths unclear. In this work, we present a unified evaluation of representative CLC-enhancement methods for question answering, spanning inference-time interventions and post-training approaches across three model families and three closed-form benchmarks. The results show that post-training methods are generally more reliable, with direct distribution alignment consistently improving CLC across all model-dataset combinations, while other methods are more sensitive to answer format and the breadth of language coverage. Notably, cross-domain transfer is limited unless source and target tasks share similar output formats. We further investigate whether CLC enhancement hurts models’ ability to respond differently when needed, that is, when asked culture-dependent questions. Across two benchmarks of culturally diverse question answering, we find no systematic degradation in controlled closed-form evaluation, whereas open-ended generation reveals occasional accuracy reductions, particularly for non-English responses. Our work highlights the need to evaluate CLC enhancement for both cross-domain robustness and culturally appropriate variation, informing future work in post-training and benchmark development.
[NLP-72] he Anatomy of an ASR Hallucination
【速读】： 该论文旨在解决自动语音识别（ASR）系统在面对环境退化或说话人-背景变化时，可能出现生成流畅但与实际语音无关的“幻觉”现象的问题。这类现象被视为更广泛的“接地失败”（grounding failure）的表现，即文本输出未能充分受音频输入的约束。研究通过分析两个独立训练的Conformer-Large识别器（一个基于连接时序分类，CTC；另一个基于循环神经网络转换器，RNN-T），发现模型最终编码阶段是导致接地失败的关键边界：若跳过该阶段，几乎所有语音片段均出现严重偏离；而跳过中间层则影响甚微。在此阶段，表示向量趋于紧凑，解码器可读性增强，字形信息变得明确。值得注意的是，人为干预导致的输出为混乱或重复内容，而非流畅的虚构文本，表明幻觉产生的机制性先决条件并非幻觉本身，而是输出缺乏足够接地。研究结果揭示了在两种不同解码架构及多种分布偏移下，高质量语音识别对终端阶段的高度依赖性，为理解并缓解ASR幻觉提供了关键的机制线索。
链接: https://arxiv.org/abs/2609.04404
作者: Hamees Sayed,Apoorv Singh,Kumar Aman,Akshat Mandloi
机构: Smallest AI; Fast Code AI
类目: Computation and Language (cs.CL)
备注:
Abstract:ASR systems sometimes produce fluent text that is unrelated to the speech they receive. We view these hallucinations as one possible consequence of a broader grounding failure, in which the transcript is no longer adequately guided by the audio. To understand where this failure becomes possible, we study two independently trained Conformer-Large recognizers - one CTC and one RNN-T - under environmental degradation and speaker-background shift. In both models, the final encoder stage emerges as a critical boundary: bypassing the final block causes divergence on nearly every utterance, whereas bypassing middle blocks has little effect. At this same stage, the representations become more compact, text becomes readable by the trained decoder, and grapheme information becomes explicit. Importantly, the intervention produces garbled or repetitive output rather than fluent fabrication. Our result therefore identifies a mechanistic precondition for hallucination - the failure to produce adequately grounded output - not the complete origin of naturally occurring hallucinations. Together, the results reveal a consistent terminal-stage dependency for grounded recognition across two decoder families and multiple distribution shifts.
[NLP-73] VERGE: Verification-Enhanced Refinement for Grounded Extraction of Early-Onset Colorectal Cancer Symptoms in Clinical Notes
【速读】： 该论文旨在解决年轻成人早期结直肠癌（early-onset colorectal cancer）筛查中缺乏基于证据的随访检测指南的问题，特别是针对该人群红标志症状（red-flag symptoms）的临床决策支持不足。现有结构化就诊数据无法充分捕捉支持早期发现与随访所需的详细信息，如症状持续时间、临床背景及家族史等关键风险因素。为此，研究提出并评估了一种自动化方法——VERGE，其核心是构建一个基于代理（agentic）的工作流，通过检索增强生成（retrieval-augmented generation, RAG）初步生成症状标注与证据，再经受控的验证-修正循环（bounded verification-refinement cycle）进行文本依据与临床有效性双重校验，对错误主张进行自动修正并重复验证，直至达成一致或达到最大迭代次数，未解决者则移交人工审查。相比单代理基线、基于规则的临床语言处理模型及替代语言模型，VERGE显著降低假阳性率，将精确率从0.764提升至0.849，马修斯相关系数（MCC）从0.681提升至0.730，在精确率与召回率之间实现更优平衡，并使98.5%的异常标记可自主解决，仅1.5%需人工介入。研究表明，受控的、以验证为核心的流程能够有效减少误报，同时不损害对真实阳性病例的检出能力，为构建更可靠、可信的临床自然语言处理工具以支持年轻患者结直肠癌风险评估提供了可行路径。
链接: https://arxiv.org/abs/2609.04366
作者: Nikkie Hooman,Monarch Nigam,Amy E. Hughes,Rasmi G. Nair,Mehak Gupta
机构: Southern Methodist University(南方卫理公会大学); UT Southwestern Medical Center(德克萨斯西南医学中心)
类目: Computation and Language (cs.CL)
备注:
Abstract:Early-onset colorectal cancer is increasing among younger adults, yet red-flag symptoms in this age group have no evidence-based guidelines for follow-up testing, and structured encounter data do not capture the detail needed to support early detection and inform follow-up, including symptom duration, context, and fam- ily history, an established colorectal-cancer risk factor. This study aimed to develop and evaluate an automated method for extracting six red-flag symptoms and family-history risk status from free-text clinical notes. We developed VERGE, an agentic workflow in which an initial label and evidence are proposed using retrieval-augmented generation, then passed through a bounded verification- refinement cycle that checks textual grounding and clinical validity, corrects and rechecks a claim until resolved or a limit is reached, and escalates unresolved claims for human review. VERGE was evaluated on 4,033 clinician-labeled note-finding pairs against a single-agent baseline, a rule-based clinical language-processing baseline, and an alternative underlying language model. Compared with the single-agent baseline, VERGE reduced false positive find- ings, improving precision from 0.764 to 0.849 and MCC from 0.681 to 0.730, a balanced gain across the precision-recall trade-off, and resolved most flagged errors autonomously, with human review required for only 1.5 percent of claims. These results indicate that a bounded, verification-based workflow can reduce unnecessary positive findings without sacrificing the ability to detect true ones. This approach offers a path toward more reliable and trustworthy clinical language-processing tools to support colorectal cancer risk assessment in younger patients.
[NLP-74] Knowing When Not to Answer: Pseudo-Ensembles for Abstention in Music Audio-Language Models
【速读】： 该论文旨在解决当前音乐音频-语言模型评估中缺乏可靠置信度判断机制的问题：现有评估方法主要依赖多选题的准确率，迫使模型必须选择一个答案，导致随机猜测与真实音乐理解难以区分。为实现模型在不确定时能够主动放弃预测（abstention），而非盲目猜测，关键在于构建一种低成本的置信度估计方法。其解决方案的核心是通过在不改变正确答案的前提下对单个预训练模型的输入进行扰动，生成“伪集成”（pseudo-ensemble）——具体包括打乱选项呈现顺序、音频噪声扰动及选项标签交换等策略，从而获得多个预测分布。这些分布可用于计算更全面的不确定性度量（如期望熵、预期熵及其差值互信息），而不仅限于单一预测分布的熵。实验表明，基于四次选项顺序扰动的伪集成将TinyMU在MuChoMusic数据集上的准确率从55.7%提升至59.2%，且不确定性度量显著改善了错误样本的排序性能，使误差保留曲线下的面积由0.293降至0.261。整个过程仅需少量额外前向传播，无需重新训练，为紧凑型音乐音频-语言模型实现实用化的置信度驱动弃权机制提供了高效可行的方案。
链接: https://arxiv.org/abs/2609.04362
作者: Aanya Maheshwari,Vatsal Raina
机构: Jumeirah College (朱美拉学院); Apta AI, Spark AI Research (Apta AI，Spark AI 研究)
类目: ound (cs.SD); Computation and Language (cs.CL)
备注: 11 pages, 4 figures, 3 tables
Abstract:Music audio-language models are evaluated almost entirely by accuracy on multiple-choice questions. This protocol forces the model to commit to an option, so a lucky guess looks the same as real musical understanding. What is missing is a way to tell when the model does not know the answer, so that it can abstain instead of guessing. The usual solution, an ensemble of independently trained models, is far too expensive here, which leaves the entropy of a single predictive distribution as the only available confidence signal. We instead build pseudo-ensembles from one pretrained model by perturbing its input in ways that cannot change the correct answer, then averaging the resulting distributions over the options. Our main construction simply shuffles the order in which the candidate answers are presented; we also study ensembles built from corrupted audio and from swapped option labels. A pseudo-ensemble gives several predictive distributions per question, so it supports the full family of ensemble-based uncertainty measures (entropy of the expected distribution, expected entropy, and their difference, the mutual information) rather than entropy alone. Evaluating TinyMU on MuChoMusic, we find that averaging over four option orderings raises accuracy from 55.7% to 59.2%, and that the resulting uncertainty measures rank the model’s errors better than the single-pass entropy baseline, reducing the area under the error retention curve from 0.293 to 0.261. All of this costs a few extra forward passes and no retraining, which makes abstention practical for compact music audio-language models.
[NLP-75] Adapting from Downturns: Prediction of Long-Term Conversational-Skill Development in Mental-Health Crisis Counselors EMNLP2026
【速读】： 该论文旨在解决志愿心理健康危机咨询师在缺乏持续督导与结构化反馈的情况下，如何早期预测其是否能够有效提升对话引导能力（即推动对话向积极结果发展的能力）的问题。核心挑战在于，传统方法依赖对话全文进行建模，难以捕捉个体在特定对话情境下的动态适应过程。论文的关键解决方案是提出一种“咨询师适应性分析”方法：通过识别咨询师在初期对话中反复遇到且处理困难的特定情境类型，追踪其在后续对话中面对相似情境时的响应调整行为，并基于这些早期适应模式构建预测模型。研究表明，咨询师对关键难点情境的渐进式应对策略变化，比直接从对话文本学习更能准确预判其长期改进潜力，从而为资源优先分配提供了可操作的依据。
链接: https://arxiv.org/abs/2609.04350
作者: Vivian Nguyen,Lillian Lee,Elizabeth A. Olson,Cristian Danescu-Niculescu-Mizil
机构: Cornell University(康奈尔大学); Crisis Text Line(危机短信热线)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: To be presented at EMNLP 2026. Code available at this http URL
Abstract:How do people learn to become better conversationalists? This question is especially important in the context of mental-health counseling, where conversational skills are essential, yet volunteer counselors often have limited access to supervision and structured feedback. Understanding how counselors develop their ability to steer conversations toward positive outcomes – and identifying early which counselors are (not) on track to improve – can help prioritize support for the counselors who need it most. In this work, we introduce the task of predicting, early in a conversationalist’s career, whether they will eventually improve at steering conversations toward positive outcomes, and demonstrate the feasibility of this task in the case of volunteer mental-health crisis counselors. Our central insight is that people may struggle with particular kinds of moments in a conversation, and that what is especially revealing of their likelihood of future improvement is how they learn to handle those moments over time. We operationalize this insight by designing a method that identifies the types of moments a counselor initially struggles with, captures how they adapt their response when they re-encounter similar moments in subsequent conversations, and learns which early adaptations predict improvement months or even years later. While this future-prediction task is challenging, our counselor-adaptation approach yields better results than baselines that learn directly from the conversation transcript. Comments: To be presented at EMNLP 2026. Code available at this http URL Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY) Cite as: arXiv:2609.04350 [cs.CL] (or arXiv:2609.04350v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.04350 Focus to learn more arXiv-issued DOI via DataCite
[NLP-76] SharedSAE: One Feature Dictionary Across Language Models
【速读】： 该论文旨在解决语言模型激活解释中稀疏自编码器（Sparse Autoencoders, SAEs）的可扩展性与复用性问题。传统方法需为每个语言模型单独训练专属的SAE并进行潜在表示标注，导致计算成本高且难以跨模型迁移。其核心解决方案是提出SharedSAE框架，通过共享一个统一的字典（dictionary）与各模型特异的编码器-解码器对相结合，实现跨模型潜在空间的高效共享。关键创新在于：不丢弃激活幅值信息，仅对选择得分进行归一化，并引入模型随机丢弃（model dropout）机制，支持单模型推理；同时在训练阶段保留了96.6%的专用SAE平均解释方差，潜在激活表现出1.8倍于后处理对齐的独立SAEs的跨模型相关性，且潜在语义描述具备良好的跨模型迁移能力。冻结字典后，新模型可快速适配至该共享潜空间，在保持接近专用SAE重建质量的同时，复用已有潜在语义标签，显著提升效率与可扩展性。
链接: https://arxiv.org/abs/2609.04344
作者: Daniil Ognev,Célian Vasson,Lijie Hu,Kentaro Inui,Benjamin Heinzerling
机构: MBZUAI; Sorbonne Université; Tohoku University; RIKEN AIP
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Sparse autoencoders (SAEs) are widely used to interpret language model activations, but SAE training and latent labelling are typically repeated for every model. Here, we show that a single shared SAE can replace a collection of dedicated per-model SAEs. Our method, SharedSAE, combines a shared dictionary with model-specific encoder-decoder pairs. Unlike the closest prior method, which discards activation magnitudes and requires all models at inference, SharedSAE instead normalizes only selection scores, preserving magnitudes, and uses model dropout for single-model inference. We train SharedSAE on four 1B-scale base language models spanning distinct families and tokenizers. Despite sharing its latents across models, SharedSAE retains 96.6% of dedicated SAEs’ mean explained variance; its latent activations exhibit cross-model correlations 1.8 times as high as separate SAEs aligned post-hoc, and its latent descriptions transfer across models. After the dictionary is frozen, new models can be efficiently adapted to it, achieving near-dedicated-SAE reconstruction quality while reusing the shared latent descriptions.
[NLP-77] A Removal Based Approach to Improve LLM Faithfulness at Test-Time
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）生成解释时存在的不忠实问题，具体表现为两个维度：不完备性（incompleteness），即解释遗漏了影响模型决策的关键因素；以及不合理性（unsoundness），即解释引用了实际未影响决策的因素。现有方法多集中于缓解不合理性，且训练阶段的方法依赖模型权重和大量计算资源，而测试阶段方法对不完备性的改善有限。本文提出一种无需修改模型参数、可在推理阶段应用的测试时间（test-time）方法，核心在于通过移除输入中未被模型解释所提及的概念后重新查询模型，从而消除未提及因素的影响，同时保留被提及概念的作用。该方法能有效提升解释的完备性，在多个数据集、模型家族及独立评估指标下均优于标准提示与鼓励忠实性的提示策略，具有模型无关性，为增强LLM辅助决策的可靠性与安全性提供了一种灵活高效的解决方案。
链接: https://arxiv.org/abs/2609.04343
作者: Qinglan Luo,S M A Nahian,John Guttag,S. Mazdak Abulnaga,Katie Matton
机构: MIT CSAIL (麻省理工学院计算机科学与人工智能实验室); Wellesley College (韦尔斯利学院); MGH, HMS (麻省总医院，哈佛医学院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) are increasingly used for consequential decisions, making their explanations an important tool for auditing model behavior. Unfortunately, these explanations can be unfaithful, failing to reflect the actual reasoning underlying the model’s decisions. We consider a setting in which an LLM provides both an answer and an explanation in response to a question. We identify two distinct dimensions of unfaithful explanations: incompleteness, meaning that the explanation omits factors that influence the answer, and unsoundness, meaning that the explanation cites factors that did not influence the model’s answer. Existing approaches to improving LLM faithfulness include training-time methods, which require access to model weights and extensive computational resources, and test-time methods that largely focus on addressing unsoundness. We introduce a test-time approach that directly targets incompleteness. We remove from the input the concepts not credited in the model’s explanation and re-query the model on the reduced input. This eliminates unmentioned influences while preserving the influence of mentioned concepts. Across two datasets, multiple model families, and two independent faithfulness metrics, our approach improves explanation faithfulness compared to both standard prompting and prompting to encourage faithfulness. Our method is model-agnostic and can be applied at inference time without modifying model parameters, providing a flexible mechanism for reducing hidden influences and improving the reliability and safety of LLM-assisted decision making.
[NLP-78] MedProb: Probing Internal Representations of Vision-Language Models for Medical Question Answering EMNLP
【速读】： 该论文旨在挑战医学视觉问答（Med-VQA）领域长期依赖医学微调、大模型或复杂多智能体系统（multi-agent pipelines）的默认假设。其核心问题是：现有方法在评估医学视觉语言模型（VLM）时，是否过度依赖生成式推理，从而掩盖了小模型中潜在的可提取医学问答信号？为此，论文提出轻量级探测框架MedProb，其关键创新在于通过线性探测（probing）从冻结的VLM表示中直接预测多选题答案，无需自由文本生成。该方法不仅显著提升了与答案相关的信息恢复能力，且在PATH-VQA、SLAKE和VQA-RAD等多个基准上表现优于传统提示（prompting）及专门医学微调的VLM与代理系统。此外，研究发现探测能缩小小模型与大模型间的性能差距，表明小模型中蕴含比生成式评估所揭示的更丰富的医学问答信号；同时，医学适配并未在所有对比对中一致提升线性可解性（linear decodability）。值得注意的是，自由文本生成存在高达10个百分点的答案位置偏差，而MedProb虽亦存在位置偏差，但其影响机制不同。最后，研究进一步证明该探测器可通过拒绝采样评分策略扩展至开放式生成任务。
链接: https://arxiv.org/abs/2609.04336
作者: Erfan Nourbakhsh,Ke Yang,Anthony Rios
机构: The University of Texas at San Antonio (圣安东尼奥德克萨斯大学)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP Findings 2026
Abstract:Medical visual question answering (Med-VQA) is often assumed to require medical fine-tuning, large models, or complex multi-agent pipelines. We revisit this assumption with \textbfMedProb, a lightweight probing framework that predicts multiple-choice Med-VQA answers from frozen VLM representations without free-text generation. Across PATH-VQA, SLAKE, and VQA-RAD, MedProb recovers substantially more answer-relevant signal than prompting and performs stronger than medical VLMs and agentic systems. Probing also reduces the apparent gap between small and large models compared to prompting, suggesting that smaller VLMs contain more recoverable Med-VQA signal than generation-based evaluation reveals. Across 14 matched general-purpose and medical VLM pairs, medical adaptation does not consistently improve this linear decodability. Finally, free-text generation exhibits an answer-position bias of up to 10 percentage points, whereas MedProb also has positional bias, however, it is impacted differently than prompting. Our main results target the multiple-choice/multiclass Med-VQA setting; we additionally show the probe can be extended to open-ended generation via a rejection-sampling scoring procedure.
[NLP-79] Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agent ic Evaluation
【速读】： 该论文旨在解决当前日益增多的智能体评估基准（agentic benchmarks）在评估过程中面临的挑战，即多数基准依赖复杂的环境配置与智能体集成，导致评估成本高、可复现性差。其解决方案的关键在于提出Harbor Adapters——一个统一的评估基础设施，通过开发超过80个基准的适配器（benchmark adapters），将各类基准标准化并兼容任意语言模型智能体的评估；同时引入Harbor-Index，一个经难度筛选、人工智能与人工双重审计及“审计-修正”闭环优化的82项高难度、多样化且高质量的任务集合，确保评估既保持大规模评估的挑战性与广度，又具备可负担性（所有模型-工具组合的通过率均不超过30%，最强模型GPT-5.5 with Codex达到28.0%）。该工作通过大规模实验（覆盖54个基准、8个模型、三种原生执行框架）实现了对智能体能力与失效模式的更全面分析，并开源全部适配器、评估结果与分析数据，推动语言模型智能体评估的可靠性与系统性发展。
链接: https://arxiv.org/abs/2609.04298
作者: Lin Shi,Haowei Lin,Zixuan Zhu,Xiaoyue Zhou,Xiang Li,Xiangning Lin,Yaxuan Deng,Han Xu,Yuangang Li,Shanda Li,Zizhao Chen,Hanwen Xing,Harsh Raj,Bo Chen,Quan Shi,Steven Dillmann,Yipeng Gao,Puneesh Khanna,Ruofan Lu,Chao Beyond Zhou,Michael Yang,Robert Zhang,Siyuan Chai,Jiayu Chang,Yizhao Chen,Xiaokun Chen,Yiwei Dai,Wenting Yang,Hange Liu,Minghao Liu,Zihan Wang,Adnan El Assadi,Benedikt Stroebl,E. Kelly Buchanan,Han Meng,Junwei He,Longxuan Yu,Radin Shayanfar,Yukyung Lee,Zhikang Dong,Allen G Hart,Anjiang Wei,Anurag Kashyap,Arpandeep Khatua,Audrey Jixin Zheng,Chengrui Ma,David Heineman,Dubing Chen,Hai-Anh Trinh,Haishuo Fang,Hefan Zhang,Hui Shen,Issa Sugiura,Jiankai Sun,Jiechao Gao,Junhong Lin,Junnan Li,Kai Yang,Lei Hsiung,Maoyu Wang,Mengze Tang,Nabil Omi,Negin Raoof,Nicholas Edwards,Octavia Guo,Orfeas Menis Mastromichalakis,Pengliang Ji,Przemysław Hejman,Qi Qi,Qunshu Lin,Richard Zhuang,Rui Yang,Ruichen Zheng,Ryan Marten,Shaghayegh Fazliani,Shizheng Hou,Sicong Jiang,Sijie Li,Song Bian,Terry Yue Zhuo,Tianqing Wu,Tom Tang,Wanjia Zhao,Weihao Xuan,Wenhua Liang,Xian Liu,Xin Lan,Xuan Zhang,Xuandong Zhao,Yanchuan Tang,Yifan Jiang,Yijiang Li,Yitong Guan,Yizhi Li,Yonghui Liu,Yuheng Tang,Yujun(Audrey)Mao,Yunfei Zhao,Yuxin Wang,Yuxuan Tang
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Evaluating agents on the growing number of agentic benchmarks is challenging because they often require complex environments and agent integrations. We introduce Harbor Adapters, a unified evaluation infrastructure for agentic benchmarks. Our work makes three contributions. First, we develop benchmark adapters that port more than 80 benchmarks to evaluate arbitrary agents, and validate them through rigorous code review and parity experiments. Second, we conduct a large-scale evaluation of 8 models spanning capability tiers across 54 benchmarks; every model is run with Terminus-2 and with one of 3 native harnesses. This enables a broader analysis of agent capabilities and failure modes than was previously possible. Third, we introduce Harbor-Index, a curated set of 82 difficult, diverse, and high-quality tasks spanning 29 benchmarks, refined from the adapted suite through difficulty filtering, AI and human audit, and an audit-and-fix loop. Harbor-Index preserves the challenge and breadth of large-scale agentic evaluations while being affordable to run; no evaluated model-harness configuration exceeds 30% pass rate, and the strongest (GPT-5.5 with Codex) reaches 28.0%. We release the adapters, evaluation results, in-depth analysis, and Harbor-Index as open-source artifacts to support more reliable and comprehensive evaluation of language-model agents.
[NLP-80] Evidence Integration in Large Language Models
【速读】： 该论文旨在解决大语言模型（LLM）在已有决策形成过程中如何整合外部证据这一关键问题，尤其关注在工具、其他智能体或用户提供的外部信息背景下，模型如何动态调整其初始判断。其核心解决方案提出了一种分布理论（distributional theory），认为外部证据通过“接收者先验权重”（receiver prior weight）与“候选证据倾斜度”（candidate evidence tilt）共同作用，改变接收者对初始答案的分布。该理论的关键在于揭示证据整合并非简单的信任机制，而是一种由接收者自身属性决定的、针对已有概率分布的特定控制策略。研究通过超过一千万次实验、来自四个模型家族的十二个大语言模型及八个领域（包括量子力学、物理、遗传学和分子生物学等科学发现任务）的验证，证实了三大预测：① 接收者更易被与其自身高概率假设一致的候选答案说服；② 更倾向于整合自身固有的特征性错误，而非来自外部来源的陌生错误；③ 相同证据可能提升弱模型表现，却损害强模型性能。此外，研究发现接收者相对可靠性前沿（receiver-relative reliability frontier）表明，与接收者认知模式相符的错误会更严重地降低性能。进一步分析显示，即使在内部验证判定候选答案无效后，模型仍以93–100%（命题约束条件下）至高达99.4%（物理与生命科学推理任务中）的概率整合该候选答案，说明证据整合是晚期网络阶段实现的结构化过程，包含引入、促进并转移候选答案至最终回答状态的多步操作；而验证表征虽可解码，但对最终答案影响甚微，且“口头验证状态”与“候选整合状态”在J-透镜分解下完全可分离，表明二者具有独立的内在表征基础。
链接: https://arxiv.org/abs/2609.04290
作者: Sebastien Kawada,Manolis Kellis
机构: Massachusetts Institute of Technology (麻省理工学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 114 pages, 16 figures, 38 tables
Abstract:Despite increasing reliance on LLMs that reason with external evidence supplied by tools, retrieval-augmented generation, other agents, and users, how LLMs integrate such evidence into decisions they have already begun to form remains largely unclear. We present a distributional theory in which evidence shifts the receiver’s distribution of initial answers, driven by a receiver prior weight and a candidate evidence tilt, leading to three predictions. First, candidates more probable to the receiver are more persuasive. Second, receivers more readily integrate characteristic errors of their own than foreign errors from different sources. Third, identical evidence can improve weaker models and harm stronger ones. We confirm these over ten million trials, twelve LLMs from four families, and eight domains, four of them scientific discovery tasks in the physical and life sciences: quantum mechanics, physics, genetics, and molecular biology. The law also yields a receiver-relative reliability frontier: receiver-congruent errors depress performance more steeply than random errors of the same rate. LLMs also integrate candidates even after internally verifying their invalidity (93-100% with propositional constraints; up to 99.4% on held-out physical and life-sciences reasoning), demonstrating evidence integration is a receiver-specific control policy over existing distributions, determined by receiver properties rather than scalar trust in the evidence source. Causal interventions show candidate integration is implemented late in the network, as a structured sequence of steps admitting external candidate answers, promoting them, and transporting them into the answer state. Representations of verification are decodable but have little causal impact on answers. A J-lens decomposition shows the state underlying verbalized verification is fully dissociable from that underlying candidate integration.
[NLP-81] Memory as transformation: LETHE a self-referential gan-inspired architecture
【速读】： 该论文旨在解决无监督、自参照条件下生成复杂动态声学内容的难题，特别是在缺乏外部数据集或人工干预的情况下实现持续的参数演化与声音生成。其核心挑战在于如何在封闭系统中维持声学演变的创造性与稳定性，同时避免依赖外部监督信号。解决方案的关键在于构建一个基于生成对抗网络（Generative Adversarial Networks, GAN）形式但闭环运行的自参照声学系统——LETHE。该系统通过一个由两个延迟线构成的3×3混合矩阵，其九个系数和两个延迟时间参数由一个五维线性判别器与一种类单样本REINFORCE的随机扰动优化器协同驱动。判别器通过对比当前音频能量行为与初始状态存档来引导参数更新，从而形成反馈闭环。实验表明，在去除主动生成器的消融测试中，关键参数Δc₂₂始终为0，证实了生成器在参数演化的必要性。系统支持固定、循环及实时音源独立混合，体现了作曲者对参数空间的精确控制，延续了自参照电子声学音乐的传统。
链接: https://arxiv.org/abs/2609.04289
作者: Francesco Vitucci,Anthony Di Furia,Francesco Scagliola
机构: 未知
类目: Computation and Language (cs.CL)
备注: Accepted at XXV CIM - Colloquio di Informatica Musicale, L’Aquila, 2026
Abstract:LETHE (Latent-parameter Evolution with Temporal Hierarchical quasi-Equilibrium) is a self-referential sonic-oblivion system implemented in SuperCollider. It adopts the formal vocabulary of Generative Adversarial Networks in a closed configuration without external datasets or supervision after initialization. Audio is processed by a 3 x 3 mixing matrix built around two delay lines; its nine coefficients and two delay times evolve through the interaction of a five-feature linear discriminator and a random-perturbation optimizer analogous to single-sample REINFORCE. The discriminator compares current energy behavior with an archive of the initial state and guides parameter updates. Circular, fixed, and live sources can be mixed independently. Across fixed and circular sessions with an ablation control, the active generator is necessary for parametric evolution ( \Delta c_22=0.000 in all 15 ablation sessions). Situated in the tradition of self-referential electroacoustic music, LETHE delegates the sonic outcome to an adaptive closed loop whose parametric space is defined by the composer.
[NLP-82] Evaluating Large Language Models for Forced Outage Risk Prediction: Benefits and Comparison to Machine Learning
【速读】： 该论文旨在解决在缺乏标注训练数据的零样本（zero-shot）框架下，利用大语言模型（Large Language Models, LLMs）预测配电网因天气相关因素导致强制停运风险的问题。研究将该问题建模为一个跨三个预报时间尺度（3小时、6小时、12小时）的二分类严重性评估任务，基于德克萨斯州中部某电力服务区域六年的停电记录与高分辨率气象数据进行分析。核心挑战在于如何在无监督条件下实现对极端天气事件引发电网故障的准确预判。其解决方案的关键在于采用零样本学习范式，通过四类零样本大语言模型与两类有监督分类器在两种输入配置（当前气象观测数据与气象预报数据）下的对比评估，验证了新一代大语言模型在不依赖标注数据的情况下仍能取得具有竞争力的性能表现。尽管有监督模型在宏观F1分数和精确率上仍占优，但大语言模型展现出在可操作推理（actionable reasoning）与地理可扩展性方面的互补优势，表明将大语言模型与有监督模型相结合可能是最优实践路径。
链接: https://arxiv.org/abs/2609.04272
作者: Christos Petridis,Zoran Obradovic,Mladen Kezunovic
机构: Temple University (坦普尔大学); Texas AM University (德克萨斯农工大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:This study examines the ability of large language models (LLMs) to predict the risk of weather-related forced outages in the distribution grid in a zero-shot framework, without labeled training data. The problem is formulated as a binary severity classification task across three forecast horizons (3h, 6h, 12h), using six years of outage records and high-resolution weather data for a utility service area in central Texas. Four zero-shot LLMs are benchmarked against two supervised classifiers across two input configurations: one using current weather observations and the other using weather forecast data. Results show that supervised models outperform LLMs on macro-F1 and precision, while newer LLM generations achieve competitive scores. Beyond accuracy, LLMs offer complementary strengths in actionable reasoning and geographic scalability, suggesting that combining them with supervised models may be the best practice.
[NLP-83] Quality Recovery for Quantized KV Caches via Low-Rank Attention Adaptation
【速读】： 该论文旨在解决低比特键值（Key-Value, KV）缓存导致的自回归解码质量下降问题，尤其是在固定量化器条件下，不同模型对量化带来的性能损失敏感性差异显著。其核心解决方案是通过知识蒸馏，将浮点型缓存模型的行为转化为低秩的查询（Q）、键（K）、值（V）投影更新，并在学生模型中以物理紧凑的增量缓存形式执行。该方法不改变量化器，而是通过低秩投影适配器动态补偿量化带来的信息损失，从而在保持内存效率的同时恢复大部分未见数据的困惑度（perplexity）。实验表明，该方法可在多种模型和量化策略下有效恢复高达75.96%的困惑度差距，且在保留关键关联检索能力的同时实现显著的性能提升，证明了低秩投影适配在固定量化格式下具备强大的质量恢复潜力，且困惑度恢复与长上下文检索能力之间并非强耦合关系。
链接: https://arxiv.org/abs/2609.04263
作者: Seifeldin Abdellatif
机构: Al Ain University (阿因大学)
类目: Hardware Architecture (cs.AR); Computation and Language (cs.CL)
备注:
Abstract:Low-bit key–value (KV) caches reduce the memory required for autoregressive decoding, but the resulting quality loss depends on the model and quantizer. We keep the quantizer fixed and distill the floating-cache model’s behavior into low-rank Q/K/V projection updates while the student executes a physically packed incremental cache. Across three seeds, 4-bit affine-cache adapters recover 54.24%\pm2.47% of the held-out perplexity gap on TinyLlama-1.1B and 75.96%\pm4.04% on Gemma-4-12B. On the same frozen NF4 Llama-3.1-8B base, one validation-selected run per quantizer recovers 60.42% under KIVI K2V2 and 37.61% under KVarN K4V2, while preserving 180-case associative retrieval. Gemma’s score on an official 4K/8K RULER subset rises from 42.80 with the unadapted 4-bit cache to 48.33 after adaptation (46.15 floating), with substantial task heterogeneity. Finally, a 2-bit rank–token sweep reduces TinyLlama’s 2-bit PPL from 576.10 to 11.4000\pm0.0059 across three seeds, versus 10.3988 floating, but restores only 11–12 of 180 retrieval cases. These results show that low-rank projection adaptation can recover held-out quality across fixed cache formats, while perplexity recovery need not restore long-context retrieval.
[NLP-84] How Much Does Corpus Choice Change Dependency-Distance Estimates?
【速读】： 该论文旨在解决语言学研究中一个长期存在的假设问题：基于单一语料库计算的依存距离（dependency distance）估计值是否可被视为语言本身的固有属性。研究表明，这一假设在跨独立构建语料库间的可重复性尚未得到充分验证。本文通过对通用依存库v2.18中38组同语言树库对进行比较，采用一致性相关分析、Bland-Altman分析及十二种预处理方案的多宇宙设计，发现不同树库间的依存距离均值一致性仅达到中等水平——替换一个树库几乎导致近40%的语言对排序发生反转，且树库选择解释了约29%的组间方差，显著高于同一树库内的抽样误差。尽管如此，所有树库均支持依存长度最小化（Dependency Length Minimization, DLM）现象（归一化比值低于1）。结果表明，依存长度最小化更应被理解为一种受语料条件制约的复合效应，由语法、语域和标注因素共同决定，而非稳定的语言层级参数。因此，虽然DLM作为定性普遍现象在语料替换下仍成立，但其跨语言的序数排名不具备稳健性。
链接: https://arxiv.org/abs/2609.04223
作者: Sirui Chen
机构: 未知
类目: Computation and Language (cs.CL)
备注: 18 pages, 3 figures. Preprint
Abstract:Dependency-distance estimates derived from a single corpus are routinely treated as properties of a language, yet this assumption has not been tested across independently compiled corpora. We compared mean dependency-distance estimates across 38 same-language treebank pairs from Universal Dependencies v2.18, using concordance correlation, Bland-Altman analysis, and a twelve-specification multiverse design. Cross-treebank agreement was moderate at best: substituting one treebank for another reversed nearly 40 percent of pairwise language orderings, and treebank choice accounted for roughly 29 percent of between-group variance. This disagreement substantially exceeded within-treebank sampling error and persisted across all twelve preprocessing specifications. Nevertheless, every treebank confirmed dependency-length minimization (normalized ratio below 1). The data are more consistent with MDD as a corpus-conditioned composite of grammatical, register, and annotation factors than as a stable language-level parameter: the qualitative DLM universal survives corpus substitution, but the ordinal cross-linguistic ranking does not.
[NLP-85] Arbitrary Polygon Oscillator: Generalizing Polygonal Synthesis to Arbitrary Shapes Morphing and Three-Dimensional Polyhedra
【速读】： 该论文旨在解决传统多边形合成（polygonal synthesis）中波形生成受限于规则参数化多边形、难以实现跨不等顶点数多边形平滑过渡以及三维形态动态控制的问题。现有方法依赖恒定角速度（constant angular velocity）且仅支持规则多边形，导致音色变化受限于单一形状家族，且无法灵活处理任意几何结构。其解决方案的关键在于提出一种基于统一弧长（arc-length）驱动的多边形合成引擎：首先，通过引入外部缓冲区读取任意顶点配置，使系统可生成正则、非正则或星形等各类闭合多边形作为波形；其次，设计了一种混合插值算法，实现不同顶点数量多边形间的平滑渐变，并在中间阶段生成无参数描述的过渡形状；最后，将该范式拓展至三维空间，利用旋转的凸多面体与固定水平平面相交，生成随姿态连续变化的截面多边形，实现三维几何对音频波形的动态控制。系统采用RNBO（Cycling '74）框架并结合几何缓存策略，避免逐样本重复计算；同时，通过基于运行时贝塞尔切线推导的四点polyBLAMP抗混叠校正与自适应过采样相结合，实现了对任意顶点构型的几何自适应抗混叠处理，无需针对每种形状进行解析推导。
链接: https://arxiv.org/abs/2608.24726
作者: Antonio Argentieri,Francesco Scagliola
机构: 未知
类目: ound (cs.SD); Computation and Language (cs.CL)
备注: Accepted at 29th International Conference on Digital Audio Effects September 1 - 4, 2026 MIT, Cambridge, MA, USA
Abstract:Polygonal synthesis generates audio by traversing the perimeter of a polygon with a phasor; prior work uses a constant angular velocity, whereas the proposed system adopts constant arc-length (perimeter) velocity. Existing formulations operate on regular, parametrically defined polygons, producing smooth timbral transitions within a single family of shapes. This paper generalizes polygonal synthesis around a unified arc-length engine: vertex data of any origin feed the same DSP pipeline. First, we adapt the oscillator to accept arbitrary vertex configurations from an external buffer, opening the possibility for a broad class of closed polygons – regular, irregular, or star-shaped – to function as a waveform generator. Second, a hybrid interpolation algorithm enables smooth morphing between polygons with unequal vertex counts, passing through intermediate shapes that have no parametric description. Third, we extend the paradigm to three dimensions: a convex polyhedron rotated about three axes is sliced by a fixed horizontal plane, and the resulting cross-section yields a continuously variable polygon controlled by the solid’s orientation. The system runs in RNBO (Cycling~'74) with a geometry caching strategy that avoids per-sample recomputation. Antialiasing combines a four-point polyBLAMP correction derived from runtime Bézier tangents with adaptive oversampling, adapting the correction geometrically to general vertex configurations without per-shape analytical derivation. Comments: Accepted at 29th International Conference on Digital Audio Effects September 1 - 4, 2026 MIT, Cambridge, MA, USA Subjects: Sound (cs.SD); Computation and Language (cs.CL) ACMclasses: H.5.5 Cite as: arXiv:2608.24726 [cs.SD] (or arXiv:2608.24726v1 [cs.SD] for this version) https://doi.org/10.48550/arXiv.2608.24726 Focus to learn more arXiv-issued DOI via DataCite
[NLP-86] Automatic Speech Recognition for Multilingual Oral History Research INTERSPEECH2026 DATE
【速读】： 该论文旨在解决在社区主导的语言遗产保护与复兴实践中，如何高效应对口述历史资料的转录难题，尤其聚焦于粤语在新西兰的复兴场景中，面对多语混用（code-switched）语言环境时自动语音识别（ASR）工具的应用有效性问题。其核心解决方案的关键在于评估生成式语音识别模型Whisper在非英语成分占比较高的粤语-英语混用语境下的表现，发现尽管最优配置下词错误率（Word Error Rate, WER）可达12.10，但存在对非英语语段识别不准的问题；然而，该模型仍具备显著效率优势，仅需人工转录所需时间的1%即可完成初稿转录，因而可作为快速生成第一轮转录文本的有效工具，为后续人工校对与语言复兴工作提供基础支持。
链接: https://arxiv.org/abs/2609.04232
作者: Sidney Wong,Chelsea Wong She,Eda Tang,Tiana Marshall Wong,Debbie Sew Hoy,Chelsea Wong
机构: 未知
类目: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL)
备注: This preprint reflects an updated version of the manuscript prepared for Interspeech 2026, incorporating revisions based on reviewer feedback
Abstract:This paper offers a unique perspective on how speech technologies are being adopted by community-led heritage language preservation and revitalisation initiatives. As a community-led language maintenance strategy, oral histories play a crucial role in Cantonese language revitalisation in New Zealand. The development of Automatic Speech Recognition (ASR) toolkits, such as Whisper, have expedited what has often been a resource and time-intensive process of transcribing oral history collections. However, there is limited research into the effectiveness of ASR toolkits when applied to code-switched language contexts. Based on Word Error Rate (WER), the best performing Whisper model configuration achieved a WER of 12.10 at the expense of accurately transcribing unsupported non-English segments. However, Whisper remains a useful tool by providing a first-pass transcription using only 1% of the estimated time otherwise needed for manual transcription.
[NLP-87] GEPARD - Generative Prosody-aware Autoregressive text-to-speech model for Realtime Dialogue
【速读】： 该论文旨在解决实时对话场景下文本到语音（TTS）生成的低延迟、高效率与高质量之间的矛盾，特别是在需要快速响应且支持个性化语音克隆的交互式应用中。其核心挑战在于如何在保持生成质量的同时实现近实时的流式推理，同时避免对现有大语言模型（LLM）推理引擎（如vLLM）进行底层计算内核修改。解决方案的关键在于构建一个以标准全注意力Transformer为骨干的自回归生成框架，将所有非必要辅助机制——包括零样本语音克隆、文本增强以及分类器自由引导（classifier-free guidance）——从自回归解码循环中移出，通过预填充阶段处理或直接以参数形式“蒸馏”进模型权重中。这一设计使得系统能够完全兼容vLLM原生服务架构，从而在单张服务器级GPU上实现高达204倍的聚合加速（256并发流），并达到约0.067的实时因子（即15倍于实时速度）。此外，论文还提出针对自回归语音解码器中“短寄存器”（1-2词）失败模式的诊断方法与缓解策略，并利用直接偏好优化（DPO）将两步式分类器自由引导转化为单步权重蒸馏，进一步提升推理效率与稳定性。
链接: https://arxiv.org/abs/2609.04222
作者: Denis Pavlov,Ulanbek Abdurazakov,Nursultan Bakashov
机构: nineninesix.ai
类目: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Machine Learning (cs.LG); Sound (cs.SD)
备注: Technical Report. 37 pages, 11 figures, Demo samples, code, and open weights: this https URL ; this https URL . Affiliation: Nineninesix, Inc
Abstract:We present GEPARD (Generative, Prosody-aware, Autoregressive text-to-speech model for Realtime Dialogue), a streaming text-to-speech model for real-time spoken dialogue. GEPARD generates speech autoregressively with an LLM backbone - text and audio embeddings are trained together in a single decoder-only model - and decodes it to a waveform with an FSQ-based neural codec, streaming audio chunk-by-chunk as text arrives. Our central goal is a TTS architecture served by a standard LLM engine (vLLM) without modifying its compute kernels. This defines the overarching design principle: the backbone is a standard full-attention transformer, while all non-trivial auxiliary mechanisms - zero-shot voice cloning, text augmentation, and classifier-free guidance - are moved out of the autoregressive decode loop into prefill, or distilled directly into the weights. On streaming end-to-end inference, a single stream reaches a Real-Time Factor of about 0.067 (roughly 15x faster than real-time); under 256 concurrent streams the system reaches an aggregate speedup of about 204x on a single server-class GPU. We detail: (1) system-level solutions for vLLM-native serving; (2) the “short register” (1-2 word) failure mode of autoregressive speech decoders, with diagnostic probes and a mitigation; and (3) distillation of two-pass classifier-free guidance over text into single-pass weights via Direct Preference Optimization (DPO). Comments: Technical Report. 37 pages, 11 figures, Demo samples, code, and open weights: this https URL ; this https URL . Affiliation: Nineninesix, Inc Subjects: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Machine Learning (cs.LG); Sound (cs.SD) Cite as: arXiv:2609.04222 [eess.AS] (or arXiv:2609.04222v1 [eess.AS] for this version) https://doi.org/10.48550/arXiv.2609.04222 Focus to learn more arXiv-issued DOI via DataCite
信息检索
[IR-0] Does Your Agents Memory Survive a Model Upgrade? A Controlled Study of Memory Portability
链接: https://arxiv.org/abs/2609.05339
作者: Ankit Goyal,Jaideep Ray
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 18 pages, 3 figures, 7 tables, under review
Abstract:Model upgrades are routine; memory migrations are not. An agent can keep the same memory store and still forget: a new model may interpret old notes differently, mixed embedding versions may break retrieval, and repair may fail without the original evidence. We compare memory as the same history is preserved verbatim for long-context reading (LC-RAW), divided into chunks for retrieval-augmented generation (RAG), compressed by a model into natural-language notes (NOTES), or normalized into a fixed-schema knowledge graph (KG-fixed). The study uses 48 synthetic histories with randomized answer codes, exact scoring, and two open-weight models with sub 10 billion parameters. Our measurements show that fixed-schema structures transfer reliably, with KG-fixed accuracy changing by only +0.0004 \pm 0.0020 following a writer swap. Conversely, compressed NOTES exhibit high model coupling, with accuracy shifting asymmetrically by +9.91 or -13.28 percentage points depending on the specific migration direction. In RAG systems, partial embedding migrations using a 50/50 mixed index capture only a 4.96-point accuracy improvement, forfeiting the majority of the 11.90-point gain achieved through full re-embedding. Diagnostic decomposition attributes 80% ( 0.467 \pm 0.014 ) of the NOTES accuracy deficit to information lost during initial construction, whereas retrieval failures drive 81% ( 0.364 \pm 0.012 ) of the RAG deficit. Finally, store-only repair of NOTES fails to reach a 90% performance recovery target in all 48 test cases, whereas retaining the raw source history enables successful recovery in 34 of 48 cases for one tested direction. These findings highlight the necessity of direction-specific migration testing, strict embedding space isolation, and the retention of source histories for memory repair. Comments: 18 pages, 3 figures, 7 tables, under review Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR) ACMclasses: I.2.7; I.2.11; H.3.3 Cite as: arXiv:2609.05339 [cs.AI] (or arXiv:2609.05339v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.05339 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-1] Students Perception of Big Data Engineering in Higher Education Curricula: Expectations Interest and Ethical Implications
链接: https://arxiv.org/abs/2609.05160
作者: Ioana-Georgiana Ciuciu,Petrescu Manuela-Andreea
类目: Computers and Society (cs.CY); Information Retrieval (cs.IR)
备注:
Abstract:The study investigates students’ interest and expectations in a Big Data Engineering course integrated with a Master curricula, as well as ethical implications of using Big Data. An anonymous online survey was conducted with 42 of the 67 students enrolled in the Big Data course offered to Computer Science and Bioinformatics Master’s programs. The responses were analyzed and interpreted using thematic analysis, highlighting interesting aspects related to students’ expectations, interest, and their perspective of the ethical implications of working with Big Data. The study concludes that, even though there is significant difference in students’ background, the majority are interested in learning Big Data, for practical and personal reasons related to the potential for career growth and their passion for the field. The main expectation expressed is related to enhancing their knowledge related to Big Data via practical activities. All students demonstrate awareness of potential ethical threats related to security and privacy, while Computer Science students are aware of the possibility of introducing bias in data during acquisition and analysis and of potential abusive data usage.
[IR-2] Beyond Maintenance Manual Multimodal RAG : Suggesting What Tool
链接: https://arxiv.org/abs/2609.05116
作者: Seongjun Ha,Md Rashedul Islam
类目: Emerging Technologies (cs.ET); Information Retrieval (cs.IR)
备注:
Abstract:Aircraft technicians are required to consult the maintenance manual (MM) for nearly every task, and locating the relevant procedure across hundreds of pages remains time-consuming. Multimodal retrieval augmented generation (MRAG) has been proposed to address this, allowing technicians to retrieve procedures, together with the accompanying figures, through natural-language queries. However, retrieval alone does not tell the technicians which tools the task requires. The MM identifies special tools only when the corresponding step is reached, and it does not state hand tool requirements at all; to select hand tools, technicians are required to find the hardware dimension from the illustrated parts catalog (IPC) and infer the right tool from it. We therefore propose MRAG-SWAT, an extension of the MRAG pipeline that returns the required hand tools and special tools alongside the retrieved procedure. The framework was implemented for the Lycoming IO-360-N1A engine and demonstrated on eight test queries. By presenting the correct tools together with the procedure, MRAG-SWAT may help reduce repeated trips to the tool crib, prevent damage to aircraft caused by improper tool selection, and thereby avoid additional maintenance tasks and support continued airworthiness.
[IR-3] Embedding Surgery: Localized Updates for Adaptive Ranking Correction in Dense Retrieval CIKM2026
链接: https://arxiv.org/abs/2609.05110
作者: Maddalena Amendola,Antonio Mallia,Raffaele Perego
类目: Information Retrieval (cs.IR)
备注: Accepted at the 35th ACM International Conference on Information and Knowledge Management (CIKM 2026)
Abstract:Dense retrieval systems are core components of modern search engines, recommendation platforms, and retrieval-augmented generation pipelines. They encode documents and queries into dense embeddings, enabling efficient semantic search via vector similarity. However, because document embeddings are computed offline and stored in static indexes, these systems struggle to adapt to user feedback or evolving search intent. To address this limitation, we introduce \emphembedding surgery, a lightweight approach for adaptive ranking correction in dense retrieval. The method applies localized, minimal updates to selected document embeddings at query time, guided by editorial feedback, user interactions, or pseudo-labels from large language models. We formulate embedding surgery as a convex optimization problem that enforces ranking constraints while minimizing modifications to the affected document representations. We integrate embedding surgery into standard dense retrieval pipelines and evaluate it on TREC Deep Learning, TREC Robust, TREC CAsT, and MS MARCO benchmarks. Results show consistent improvements (e.g., up to +60.64% relative improvement in nDCG@10 on DL-Hard under editorial feedback), even under noisy or shifting feedback, with low computational cost and without disrupting the global structure of the embedding space. Extensive experiments show that ranking corrections propagate to semantically related queries and that embedding updates can be applied safely and efficiently to scalable Approximate Nearest Neighbor indexes via simple in-place overwriting, without requiring costly index reconstruction. Finally, embedding surgery complements query adaptation methods such as CoRocchio, yielding additional gains while being more robust to noisy feedback.
[IR-4] Beyond Co-purchase Relation: Evolution of Complementary Recommendations at Allegro RECSYS2026
链接: https://arxiv.org/abs/2609.05063
作者: Aleksandra Osowska-Kurczab,Klaudia Nazarko,Eliška Kosturová,Lidia Wojciechowska,Michał Bień
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Recsys 2026: OARS workshop
Abstract:When a customer adds a professional camera to their cart, should the system suggest a matching lens, a generic tripod, or another camera body? Complementary Product Recommendation is vital for comprehensive basket building, yet standard models often fail to distinguish between items that are merely bought together and those that truly work together. In this paper, we present AlleCompanion: a production-scale retrieval framework deployed at this http URL that transforms noisy behavioural signals into precise semantic compatibility. We mitigate the intrinsic noise in large-scale co-purchase traffic by combining data-level filtering heuristics with a category-constrained Two Tower architecture. Within this framework, the Category Adapter guides the model in the embedding space, constraining candidates within logically complementary boundaries. Since modelling authentic user behaviour at scale is inherently difficult, we introduce ComCat, a multi-source Complementary Categories Mapping. ComCat acts as a translational layer that distils meaningful patterns from noisy traffic into a maintainable and controllable solution, integrating expert rules, human-in-the-loop feedback, LLM-based reasoning, and statistical mining. Our experimental results demonstrate that combining explicit category-level constraints with neural architectures effectively filters out co-purchase noise to surface recommendations that satisfy real-world user needs. Serving over 20 million active users monthly, the framework delivers significant uplifts in attributed GMV for organic discovery and drives substantial revenue growth in sponsored placements.
[IR-5] Repeated Queries Exhaust an LLM s Brand Recommendations but Not Its Sources
链接: https://arxiv.org/abs/2609.05059
作者: Dmitrij Żatuchin
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注: 6 pages, 2 figures, 2 tables, code, per-cell tables and data pointers at this http URL (research/recommendation-saturation)
Abstract:Whether repeated identical buying questions exhaust a language model’s brand recommendations depends on retrieval. Across 300 question-engine cells (50 questions, six engines, 15 runs each, open extraction over 1,470 adjudicated organizations), the five engines answering without web search were still adding never-seen brands at run 15 in 86-92% of cells, with median repertoires of 15-31 organizations; the one retrieval-enabled engine closed its list (median 8 organizations, 64% of cells still adding), matching four earlier deep cells where web-search runs saturated by run ten. Cited-domain accumulation keeps rising at every horizon tested: four deep cells were still adding domains at run 24 with 59-84% of the Chao2 lower-bound estimate observed, and 44% of the retrieval engine’s breadth cells were still adding domains at run 15. A single run shows 62-77% of the five-run brand set, and across engines the median question draws 38 organizations, of which a median of 15 appear in exactly one engine. Estimators are exact rarefaction and Chao2 richness; a parallel fixed-roster extraction reproduces flat curves on identical responses, so roster-bounded tracking manufactures plateaus that open extraction removes.
[IR-6] Leverag ing Low-Level Symbolic Competences for Unsupervised Grounding in Hallucination Detection EMNLP2026
链接: https://arxiv.org/abs/2609.05025
作者: Renato Vukovic,Hsien-chin Lin,Carel van Niekerk,Benjamin Ruppik,Michael Heck,Shutong Feng,Nurul Lubis,Milica Gasic
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: Accepted to GroundLM EMNLP 2026 Workshop
Abstract:Hallucination-where a language model generates outputs that are factually incorrect or unsupported by the source-is a major challenge for both prompted and fine-tuned language models. Detecting hallucinations is difficult due to the opaque reasoning processes of LLMs, which often provide little insight into why a model’s output may be inaccurate. In this work, we investigate whether an LLM can use an alternative, low level, symbolic competence such as SQL for unsupervised hallucination detection in some high level task. For this, we make an LLM build an SQL database from reference documents. This SQL database is then used for reasoning over the reference and the sampled response in a hallucination detection pipeline that is grounded in the database, thereby providing a neurosymbolic checkup. On RAGTruth and DiaHalu hallucination detection datasets, we find that our approach improves on direct prediction and competes with state-of-the-art hallucination detection methods, while not requiring domain-specific fine-tuning. Instead it relies on a low-level general competence already present in LLMs. This warrants further investigation of low-level LLM competences in neurosymbolic approaches. Comments: Accepted to GroundLM EMNLP 2026 Workshop Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) Cite as: arXiv:2609.05025 [cs.CL] (or arXiv:2609.05025v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.05025 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-7] A Tree-based RAG Framework for Evidence-Intensive QA via Adaptive Planning and Topology-Aware Evidence Gathering EMNLP2026
链接: https://arxiv.org/abs/2609.04981
作者: Songeun Lee,Kyungjin Min,Injae Na,Suyeong Lee,Chiyoung Kim,Woohwan Jung
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: Accepted to Findings of EMNLP 2026
Abstract:Recent structured RAG methods leverage tree- or graph-based reasoning structures to improve multi-hop QA. However, they face key limitations in evidence-intensive QA, where answering a question requires synthesizing information scattered across dozens or even hundreds of documents: structural rigidity, which limits adaptive reasoning expansion, and topology-ignorant evidence gathering, which prevents effective integration of evidence across different reasoning nodes. To address these issues, we propose APT-RAG, an Adaptive Planning and Topology-aware evidence gathering RAG framework. Adaptive planning dynamically expands the reasoning structure based on question dependencies and evidence requirements, while topology-aware evidence gathering improves evidence coverage through sibling evidence reuse, direct retrieval, and evidence aggregation from child nodes. We further introduce evidence-guided batched answer generation to reduce significant generation overhead in evidence-intensive QA. In the experiments on evidence-intensive QA benchmarks, APT-RAG outperforms existing structured RAG methods. Our code is available at this https URL.
[IR-8] SAM-D2Q: Aligning Multimodal Doc2Query with Search Demand and Conversion for E-commerce CIKM2026
链接: https://arxiv.org/abs/2609.04961
作者: Hui Zhou,Jian Hui Ji,Lei Ma,Rong Xiao,Xiaoyi Zeng
类目: Information Retrieval (cs.IR)
备注: Accepted by CIKM2026 Oral Full Paper
Abstract:E-commerce search often suffers from vocabulary mismatch between user queries and merchant-authored product titles, since short titles cannot fully cover diverse user expressions or visual product attributes. Although Doc2Query alleviates this issue by generating pseudo-queries for document expansion, traditional methods are text-only and not optimized for e-commerce business objectives. As a result, they may produce semantically plausible but commercially ineffective expansions and miss key attributes present in product images. To this end, we propose E-commerce Search-Aligned Multimodal Doc2Query (SAM-D2Q), a business-aligned multimodal document expansion framework for e-commerce search under Boolean retrieval constraints. SAM-D2Q consists of three stages: (1) task-adapted multimodal supervised fine-tuning to enhance vision-language understanding of product titles, images, and user queries; (2) multimodal data augmentation to improve perception of key visual attributes and expansion coverage; and (3) reinforcement-learning-based preference alignment toward search business objectives, encouraging the model to generate pseudo-queries that better match user intent and commercial value. Offline experiments show that SAM-D2Q substantially improves retrieval performance over traditional Doc2Query methods. Deployed in the AliExpress production search system, SAM-D2Q improves online business metrics, increasing GMV by +3.38% and Pay Count by +2.27%.
[IR-9] AtomRec: Evolving Atomic Memory for Agent ic Recommendation
链接: https://arxiv.org/abs/2609.04882
作者: Peiyu Hu,Weihai Lu,Siying Gu,Zhuodong Liu,Zhaokai Luo,Yuean Niu,Zhiyong Wang,Jia Wang
类目: Information Retrieval (cs.IR)
备注:
Abstract:Agentic recommender systems use large language models to maintain semantic memory and support evidence-aware recommendation. However, existing memory mechanisms often compress user and item information into coarse summaries and connect them with scalar collaborative links, making it difficult to preserve fine-grained preference stages or retrieve interpretable evidence as user interests evolve. We propose \textscAtomRec, an agentic recommender with evolving atomic collaborative memory. \textscAtomRec represents user and item memories as structured atomic units, builds semantic links across related memories, and evolves related historical fields when new interactions arrive. During recommendation, it retrieves linked memories as multi-hop evidence paths rather than isolated neighbor summaries, allowing collaborative signals to support grounded ranking. Experiments on four public benchmarks show that \textscAtomRec consistently outperforms state-of-the-art agentic and memory-augmented baselines, with around 8.5% average relative improvement across metrics.
[IR-10] Personalized Task Dependency Graphs for Mitigating Signal Erosion in Multi-Task Recommendation CIKM2026
链接: https://arxiv.org/abs/2609.04862
作者: Fuyuan Liu,Tiandeng Wu,Yaqun Fang,Wei Zhou,Zehao Zhou,Wenping Chen,Qishun Mei,Jiaxin Zhou,Heng Chang,Yi Cao,Jiandong Ding
类目: Information Retrieval (cs.IR)
备注: Accepted at CIKM 2026
Abstract:Optimizing multiple conversion objectives is a core challenge in industrial recommendation, often limited by signal erosion in rigid architectures. Existing Multi-Task Learning (MTL) methods typically enforce uniform dependency strengths across a static conversion funnel, overlooking how task correlations naturally vary based on item characteristics. Hierarchical message passing along these fixed chains leads to cumulative signal attenuation, which degrades performance on sparse, deep-funnel objectives. To address this, we propose the Personalized Task Dependency Graphs (PTDG). While respecting necessary physical causal constraints (e.g., Click - Pay), PTDG dynamically “rewires” the intensity of dependency pathways for each item via low-rank approximation to ensure structural robustness. We implement a GCN-based propagation with hard causal masking to establish adaptive information shortcuts. Additionally, we introduce an Adaptive Progressive Masking (APM) strategy that decouples shared parameters according to task sparsity, helping to stabilize optimization. Experiments on KuaiRand1K and an industrial dataset show that PTDG significantly improves AUC on sparse conversion tasks by up to 1.45%, while maintaining comparable performance on dense objectives. Online A/B testing shows PTDG improves Conversion Rate (CVR) by 1.2% and effective Cost Per Mille (eCPM) by 1.9% relative to the baseline.
[IR-11] Inventory-Grounded Policy-Level Optimization for Training-Free AI Search EMNLP2026
链接: https://arxiv.org/abs/2609.04813
作者: Wei Zhou,Tiandeng Wu,Jiandong Ding,Zhufeng Fan,Yi Cao
类目: Information Retrieval (cs.IR)
备注: Accepted at the EMNLP 2026 Industry Track
Abstract:Early in deployment, an AI search system typically operates over a frequently updated product catalog, so the available items and their properties cannot be treated as stable knowledge that can be encoded in fixed prompts or strategies. Fine-tuning, reinforcement learning, and static prompt patches fit poorly: labels are scarce, rewards drift with inventory, model releases are costly, and prompt fixes quickly stale. We present Inventory-Grounded Policy-Level Optimization (IGPO), a training-free approach for fixed AI search pipelines. IGPO separates policy from environment facts: it learns Policy Guidelines for acting on runtime inventory evidence rather than memorizing available items. Online, IGPO grounds each query by probing the inventory and constructing an inventory portrait, then injects relevant Policy Guidelines into the retrieval and selection prompts. Offline, stochastic rollouts are grouped by query – mixed outcome groups directly yield contrastive signal, and an inventory-guided exploration loop distinguishes missed retrieval routes from cases where no matching support is found under the observed inventory evidence. Since May 2026, IGPO has been deployed in a commercial smart-assistant AI search system. A 14-day online A/B test of the complete IGPO treatment shows a 3.17% relative CTR lift and a 38.9% reduction in audited bad cases.
[IR-12] VizIt: A multi-view framework for exploring single-cell spatial and genetic data online
链接: https://arxiv.org/abs/2609.04658
作者: Chenhang Christopher Zhang,Yanqing Lou,Jie Yuan,Mingming Lu,Jacob Parker,Himanshu Chintalapudi,Zechuan Lin,Clemens R. Scherzer,Yuxuan Hu,Ruifeng Hu,Xianjun Dong
类目: Information Retrieval (cs.IR); Genomics (q-bio.GN)
备注:
Abstract:Multi-omic studies increasingly require data to be examined from complementary biological perspectives, yet interactive exploration remains fragmented across modalities and tools. We present VizIt, an open-source framework for multi-view exploration of single-cell and spatial transcriptomic, epigenomic and genetic data. VizIt connects gene-, cell type-, condition-, spatial-, genomic region- and variant-centered views, enabling seamless navigation across biological perspectives. We demonstrate VizIt through the Parkinson’s Cell Atlas, a customizable interactive multi-omic resource.
[IR-13] CAGE: Coherence-Aware Graph Encoding for Retrieval-Augmented Generation
链接: https://arxiv.org/abs/2609.04647
作者: Tong Qi,Jingyu Wu,Youbing Yin,Spencer Hong,Daben Liu,Erin Babinsky
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:Traditional Retrieval-Augmented Generation (RAG) systems score each passage independently against the query, assembling context sets that may be individually relevant yet collectively incoherent. We introduce Coherence-Aware Graph Encoding (CAGE), a reranking framework that models “between-chunk coherence” across four dimensions: Intra-Domain Relevance, Noise Resistance, Informational Bonding, and Factual Consistency. Our pipeline transforms retrieved passages into directed heterogeneous entity graphs, amplifies factual anchors via min-out-degree reweighting, encodes structural patterns through a Relational Graph Convolutional Network, and fuses inter-chunk coherence with query relevance for final ranking. Evaluated across four multi-hop benchmarks, CAGE matches or outperforms strong baselines including monoT5 in Recall@5 on bridge-dominated datasets and consistently improves downstream Exact Match, demonstrating that structurally coherent context yields more precise answers even when retrieval recall is comparable or lower.
[IR-14] Latent-Aligned Reasoning for Multimodal Recommendation
链接: https://arxiv.org/abs/2609.04645
作者: Jiarui Jin,Anyang Ji
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Multimodal Vision-Language Models (VLMs) have demonstrated remarkable capabilities in cross-modal understanding, yet a fundamental challenge persists when applying them to recommendation: as representations propagate through multi-step reasoning, both visual and textual signals progressively attenuate - a phenomenon we term cross-modal dilution. To address this, we propose LARK (Latent-Aligned Reasoning frameworK), a two-stage latent reasoning framework with complementary alignment mechanisms within a single VLM. In the first stage, learnable latent tokens are interleaved with multi-step chain-of-thought (CoT) reasoning and explicitly aligned with a frozen vision encoder, serving as visual checkpoints that preserve perceptual details throughout the reasoning chain. In the second stage, the latent representations are projected via a bridge MLP and trained with item-to-item contrastive learning; to prevent the reasoning semantics from fading, intermediate features are aligned with the CoT hidden states from the first stage, anchoring the final embeddings to the model’s own reasoning output. Experiments on three public benchmarks and one industrial dataset show that LARK achieves state-of-the-art performance across multiple recommendation architectures, with controlled ablations confirming the distinct contribution of each component.
[IR-15] MURAL: Multimodal Uncertainty-aware Recommendation via Adaptive edge Learning
链接: https://arxiv.org/abs/2609.04574
作者: Ahmad Mousavi(Department of Mathematics and Statistics American University),Majid Alikhani(Independent Researcher),Yeon-Chang Lee(Department of Computer Science and Engineering Ulsan National Institute of Science and Technology),Roberto Corizzo(Department of Computer Science American University),Yeganeh Abdollahinejad(Department of Biosystems and Agricultural Engineering Michigan State University)
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG); Social and Information Networks (cs.SI)
备注:
Abstract:Multimodal Graph Neural Networks have become standard for recommendation by augmenting sparse interaction data with content features. Yet current architectures face two bottlenecks: structural rigidity, from a reliance on static precomputed similarity graphs that cannot adapt to evolving preferences; and semantic fragility, where noisy modality signals are indiscriminately fused, distorting the collaborative signal. We propose MURAL (Multimodal Uncertainty-aware Recommendation via Adaptive edge Learning), a unified framework that shifts multimodal recommendation from fixed structural augmentation to dynamic topology discovery. To address structural rigidity, an Adaptive Edge Learner combines a differentiable retrieval-augmented strategy with an approximate nearest neighbor search to discover latent item-item correlations that are both semantically adaptive and computationally scalable (O(NlogN)). To address semantic fragility, an Uncertainty-Aware Fusion module models the aleatoric uncertainty of heterogeneous modalities, dynamically down-weighting unreliable features while prioritizing high-confidence signals as a defense against cross-modal noise. We further employ a contrastive teacher-student alignment that anchors modality-specific representations to stable behavioral signals, ensuring optimization stability without gradient leakage. Experiments on large-scale benchmarks including TikTok and Amazon show that MURAL significantly surpasses both structural and generative state-of-the-art baselines, achieving superior accuracy while offering interpretability through domain-specific modality dominance and robustness under extreme data corruption.
[IR-16] BioSync: Transformer-Based Cross-Modal Fusion for a Multimodal Physiological Digital Biomarker
链接: https://arxiv.org/abs/2609.04504
作者: Seyed Mahmoud Sajjadi Mohammadabadi
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Cardiac, neural, behavioral, and speech measurements from wearable and mobile devices provide partial, noise-sensitive views of physiological state. BioSync combines these measurements into the \textbfBioSync Index (BSI), a continuous composite digital biomarker defined under the BEST framework. The model applies multi-head self-attention to modality tokens and adds a linear branch whose hypothesis class includes standard feature concatenation. This architecture is motivated by latent-variable measurement theory and by the possibility that joint observations contain information unavailable from individual modalities. We evaluated BioSync on two literature-informed synthetic cohorts: a four-modality cognitive-decline cohort using HRV, EEG, actigraphy, and speech, and a metabolic-autonomic cohort structured around the public AI-READI wearable schema. In the cognitive cohort, BioSync and concatenation obtained AUCs of 0.928 and 0.926, respectively. In the metabolic cohort, BioSync obtained accuracy/F1 of 0.764/0.766, compared with 0.756/0.758 for concatenation. The BSI correlated with latent severity in both cohorts ( r=0.91 and r=0.68 ). A pure-attention ablation obtained cognitive-cohort AUC 0.911, locating the increase to 0.928 in the combined wide-and-deep architecture. With matched modality-dropout training, BioSync led concatenation at five of six cognitive-cohort corruption rates and at the highest metabolic-cohort rate. Its cognitive-cohort AUC was also higher than five published digital-biomarker reference values, although differences in datasets and tasks preclude a controlled benchmark claim. Comparison with single-modality, early-fusion, and late-fusion designs across six prespecified criteria identifies the model’s computational properties; validation on real cohorts remains necessary.
[IR-17] Evaluation of Phonetic Encoding Algorithms on Transcription Datasets
链接: https://arxiv.org/abs/2609.04391
作者: Can Özbey,Emre Kaplan,Berkin Deniz Kahya
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 17 pages, 3 figures, 5 tables
Abstract:In this work, a novel evaluation scheme built on a generalized variant of the Rand Index measure, namely, the Hüllermeier-Rifqi Index, is proposed in order to assess how well phonetic encoding algorithms conform to word-based transcriptions in IPA (International Phonetic Alphabet) notation. For this objective, the discordance score is obtained by calculating the absolute difference between the pairwise similarity values of ground-truth transcriptions and those of corresponding phonetic encodings, which are computed using normalized edit distance as a permutation dependent string metric. The resulting score is subsequently adjusted with respect to that of a random string generator incorporating the same alphabet as the encoder under consideration. A wide range of phonetic encoders were evaluated as such on multi-lingual transcription datasets along with their recall capabilities based on the collision rate. The validity of the proposed scheme is further supported by its applicability in measuring the orthographic transparency of a language when the writing system is viewed as an inherent phonetic representation.
[IR-18] Corporate-Family Resolution Is Not a String-Matching Problem: A Public Benchmark Stratified by Name Visibility
链接: https://arxiv.org/abs/2609.04269
作者: Harshit Gupta
类目: Databases (cs.DB); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 21 pages, 8 tables. Benchmark, full adjudication log, and code reproducing every number: this https URL
Abstract:Deciding whether two supplier records belong to the same corporate family is a prerequisite for spend consolidation, credit exposure aggregation and sanctions screening. It is usually treated as entity matching, but the tasks differ: a family link connects records that are deliberately different entities, and the evidence often appears in neither record. We introduce CorpFam, a public benchmark of 54,864 candidate pairs over 10,307 corporate families, derived from 6,638,350 US federal award records in which every supplier self-reports its ultimate parent to a government registry. Pairs are stratified by name visibility: whether the names are identical after normalisation, share a distinctive token, or share none. Because strata have positive rates from 10.2% to 97.3%, we report per-stratum recall, base-rate invariant, rather than F1, which is not. The strongest of 5 matchers recovers 100.0% of identical pairs and 4.2% of invisible ones; no method exceeds 4.7% on the latter. The failure begins before matching. Blocking decides which pairs a matcher sees, and we evaluate 7 schemes spanning phonetic keys, attribute keys that ignore the name, and semantic nearest neighbours. None reaches three percent on invisible pairs, and their union recovers 6.8%. 93.2% of these links never enter the candidate set, so no matching-stage improvement can reach them. The links are real: against SEC Exhibit 21 subsidiary schedules, which share no provenance with procurement registration, 64.2% of invisible links are corroborated, against 0.16% under permuted parents and 0.41% against the same parent’s wrong exhibit: two unrelated nulls agreeing to within 0.25 points. Corporate-family resolution is a retrieval problem misfiled as a matching problem; the intervention point is candidate generation, not ranking. The benchmark, adjudication log, and code reproducing every number are released.
[IR-19] SAGE: Semantic Attribute Graphs for Multi-Entity Visual Retrieval EMNLP2026
链接: https://arxiv.org/abs/2609.04255
作者: Yongjoo Kim,Mincheol Kwon,Seonga Choi,Minseung Lee,Kyeong-Jin Oh,Hyunyoung Lee,Yunsu Choi,Jungbeom Lee
类目: Information Retrieval (cs.IR)
备注: EMNLP 2026 (Main); Project: this https URL
Abstract:Dense document images often contain many fine-grained visual and textual entities whose relevance depends on a user query. Standard vision-language retrievers encode cropped regions with a single vector, which can mix distinct entity signals and obscure the evidence needed for fine-grained retrieval. We call this failure mode Semantic Dilution and quantitatively show that it degrades entity-level retrieval as a function of entity density. To mitigate it, we propose SAGE, a training-free framework that parses semantic entities from dense document images, represents them as hierarchical graph nodes with multi-vector embeddings, and retrieves query-relevant evidence through iterative entity-level subgraph matching. We also introduce DEAR, a dataset of 1,055 query–image pairs sourced from product detail pages, where each query requires retrieving and comparing multiple fine-grained entities from visually dense inputs across four question types of increasing complexity. Experiments show that SAGE substantially reduces semantic dilution and outperforms patch-level and OCR-based retrieval baselines on DEAR, achieving a Recall@3 of 0.849 and a generation score of 2.746 on multi-entity visual comparison queries. Our code is available at this https URL.
人机交互
[HC-0] Diffusion TV: Experiencing Diffusion Models through Tangible Embodied Interaction
链接: https://arxiv.org/abs/2609.05404
作者: Sihwa Park
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: In Proceedings of Explainable AI for the Arts Workshop 2026 (XAIxArts 2026) arXiv:2607.20131
Abstract:Diffusion TV is an interactive AI art installation that offers a tangible and embodied experience of diffusion models through a modified CRT TV. By physically manipulating the TV’s antenna, audiences control the clarity of AI-generated images and sounds, metaphorically enacting the denoising process that underlies diffusion-based generation. Using the tuning knob, participants switch between three channels featuring AI-generated animals from the Past (extinct species), Present (endangered species), and Future (speculative creatures), situating the interaction within a temporal and ecological narrative. Through continuous audiovisual feedback and physical interaction, Diffusion TV foregrounds the generative process over final outputs, allowing audiences to explore intermediate states as experiential material. Rather than providing explicit technical explanation, the work presents an alternative, embodied mode of explainable AI that invites exploratory engagement with and reflection on generative technologies.
[HC-1] From Interpretability Methods to Interpretable Models
链接: https://arxiv.org/abs/2609.05399
作者: Julien Colin,Nuria Oliver,Thomas Serre
类目: Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
备注:
Abstract:More than a decade in, explainable AI (XAI) for computer vision has assembled a mature toolbox: attribution, feature visualization, concept-based, and circuit-based methods. Yet almost all of the field’s effort has gone into building and comparing these methods, and little into the question they were meant to answer—how interpretable are our models, and are we making progress as they evolve? We argue for shifting the field’s focus from methods to models, along two complementary lines. One is already within reach: existing tools let us characterize and compare what different models represent and compute. The other is harder, and largely neglected: whether a model can actually be understood by the humans who rely on it—the independent evaluators on whom trust and certification depend, not the experts confirming what they already expect. It can only be measured, not inferred. We review why the toolbox is mature enough to support both, survey the thin body of work comparing models, draw a parallel to systems neuroscience, and close with a model-centric XAI agenda.
[HC-2] herMosaic: Accelerating Perceived Thermal Transitions Through Spatiotemporal Thermal Feedback
链接: https://arxiv.org/abs/2609.05347
作者: Zining Zhang,Jiasheng Li,Myungin Lee,Zeyu Yan,Jin Ryong Kim,Huaishu Peng
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Thermal feedback can enrich immersive interaction, but thermoelectric devices often change temperature too slowly to match interactive timing. We present TherMosaic, a spatiotemporal thermal feedback approach that accelerates perceived temperature transitions by leveraging two perceptual mechanisms: spatial summation and thermal adaptation. Focusing on the fingertip, we first investigate this approach using a custom 2*2 array of independently controlled Peltier modules. Across three controlled perceptual studies, we show that distributed thermal stimulation can preserve stable hot and cold percepts despite local deviations, that adaptation helps maintain these percepts during changing stimulation, and that combining these effects reduces perceived transition time by about 30%-40% for transitions originating from hot or cold states. We then translate the same design principles into a standalone wearable implementation of TherMosaic and evaluate it in virtual reality. Our results show that this approach reduces perceived thermal lag and improves temporal alignment between thermal and visual events in interactive use.
[HC-3] Beyond Bias: Participatory and Reflective Approaches to Cultural AI
链接: https://arxiv.org/abs/2609.05102
作者: Archana Prasad,Isha Singh,Tom Simmons
类目: Human-Computer Interaction (cs.HC)
备注: In Proceedings of The First Reflection in Creative Experience (RiCE) Workshop (RiCE W1) arXiv:2607.24558
Abstract:Generative AI systems increasingly shape cultural production, yet creative intentions, cultural meanings, and interpretive practices often can’t be articulated through computational metrics alone. This paper presents Beyond Bias, a collaboration between this http URL and Goethe-Institut India, as a participatory approach to cultural AI which includes collaborative dataset creation, reflective AI tooling, artist-led model fine-tuning, and co-authored governance practices. Across 9 workshops involving over 200 participants, artists and cultural practitioners engaged with AI systems through experimentation, iteration, and collaborative LoRA training. Participants used their AI-generated outputs and visualizations as reflective interfaces for exploring symbolism, memory, authorship, and cultural contexts. Comparing contemporary generative AI outputs with participant fine-tuned outputs helped participants reflect on cultural details missing in big tech AI systems. This paper contributes reflective AI tooling approaches foregrounding transparency, stewardship, and community participation; findings from participatory workshops examining how generative AI visualizations mediate cultural representation and interpretive practice; and a framework for cultural AI grounded in cultural integrity, and reflective practice.
[HC-4] Scales Reflections and Conversations: A Multi-Modal Approach to Emotion Annotation
链接: https://arxiv.org/abs/2609.05046
作者: Pragya Singh,Prashasti Gupta,Hitesh Bhandari,Kanishk Goel,Mohan Kumar,Pushpendra Singh
类目: Human-Computer Interaction (cs.HC)
备注: Accepted at MobileHCI 2026
Abstract:Mental health concerns are increasing worldwide, highlighting the need for interventions that support everyday emotional well being. Prior work has demonstrated the potential of wearable and mobile technologies to deliver data driven interventions. However, developing effective data-driven systems requires access to emotion data that captures individuals’ emotional variability and change in everyday contexts. Existing approaches to data collection largely rely on frequent, prescheduled prompts and predefined scales or questionnaires. These methods often fail to account for participants’ availability, agency, or the complexity of their emotional experiences, resulting in shallow, context poor data. In this paper, we present a feasibility study of a participant centric, multimodal emotion-annotation application designed around users’ emotional intensity and availability. Our findings show how multimodal emotion logging can shape participants’ experiences and data logging behaviors, and demonstrate its potential to support the collection of richer, more nuanced emotion data.
[HC-5] Beyond Prompt-to-App: Accountable Translation in Teacher-Facing Agent ic Authoring
链接: https://arxiv.org/abs/2609.04679
作者: Nizam Kadir(1),Wei Ting Liow(1),Sumbul Khan(1),Lay Kee Ang(1) ((1) Science, Mathematics and Technology, Singapore University of Technology and Design, Singapore)
类目: Human-Computer Interaction (cs.HC)
备注: 22 pages, 4 figures, 5 tables. Preprint; not peer reviewed. Includes ancillary de-identified analytic materials
Abstract:Natural-language app builders let domain experts create software, but their pipelines transform professional intent across compilation, generation, checking, and approval. We report a bounded trace study of a teacher-facing agentic authoring system. Evidence comprises six eligible build attempts across three accounts; a separate corpus of 37 workshop units from 23 display names contextualizes commitments without person-level linkage. Compiled specifications added governance requirements, while downstream representations sometimes normalized case-specific learning relations. Two drafts met a stored package/security threshold despite analyzer reservations and unresolved correspondence to their briefs; four attempts in one account produced no usable payload, and repair messages did not translate internal terms into domain-legible revisions. We develop accountable translation as an analytic framework for making consequential changes attributable, inspectable, scoped in validation, and contestable. It extends HCI accounts of traceability and end-user debugging by locating professional authority and repair rights across heterogeneous technical and organizational handoffs.
[HC-6] Hidden In Plain Gaze: Gaze Representations as Privacy Controls for Utility and Re-identification Risk in XR
链接: https://arxiv.org/abs/2609.04592
作者: Cory Ilo,Brendan-David John,Doug A. Bowman
类目: Computer Vision and Pattern Recognition (cs.CV); Emerging Technologies (cs.ET); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注:
Abstract:Intelligent extended reality (XR) systems increasingly use eye and head tracking to infer user intent, task, and attention, but the same signals can also reveal biometric identity. We study whether gaze data representation choice can serve as a lightweight privacy control at feature extraction, before adding perturbation or formal privacy mechanisms. Using the egocentric HoloAssist dataset, we compare three gaze representations under matched model capacity: raw gaze, spatial attention heatmaps, and engineered eye-movement features. We evaluate each representation on action recognition as task utility and closed-set user re-identification as privacy leakage. Representation choice substantially changes the privacy-utility tradeoff. Engineered features retain roughly 85% of raw gaze’s action-recognition accuracy while reducing re-identification by about an order of magnitude, to roughly four times the chance rate across 206 identities. This reduction attenuates rather than eliminates identity leakage, and the differences across representations show that abstraction alone does not guarantee privacy. Engineered features expose interpretable and auditable structure, giving designers a transparent privacy lever that complements mechanisms such as differential privacy.
[HC-7] Matched Starts Divergent Objects: How Human-AI Collaboration Forms What It Explains
链接: https://arxiv.org/abs/2609.04542
作者: Mehmed Zahid Çögenli
类目: Human-Computer Interaction (cs.HC)
备注: 90 pages, 4 figures, 9 tables; includes three supplementary appendices
Abstract:Scholarly knowledge is typically encountered in stabilized form, while the process histories through which research objects, claims, and contributions acquire form remain largely hidden. This study examines how human-AI scholarly collaboration develops under matched starting conditions and whether those conditions stabilize the inquiry itself. Using a longitudinal corpus of 843 turns, the same expert researcher developed branch-isolated scholarly trajectories with different generative AI systems from the same corpus, frozen research problem, starting prompt, publication objective, and conduct rules. Two eligible trajectories were reconstructed ex post through scholarly trajectory analysis, source-faithful interaction reconstruction, a Socioduality relational-process overlay, and downstream propagation analysis. Both trajectories independently shifted the initial continuity problem from recall toward usability, but subsequently formed different research objects. One trajectory culminated in an endpoint manuscript on continuity labour, the distributed work required to sustain usable collaboration; the other in an endpoint manuscript on distributed, evolving, and unevenly usable project state. Their analytic genealogies involved failed analytical units, rejected explanations, changes in scale, counterexamples, and conceptual stabilization, and propagated into different research questions, findings, methods, evidence logics, and scholarly contributions. Relational analysis further showed that consequential scholarly change, reciprocal continuity, and local substantive re-formation were distinct process structures, and that continuation did not necessarily constitute epistemic endorsement. The findings demonstrate empirically constrained research-object formation within human-AI scholarly collaboration and show how process histories shape the scholarly objects and products that emerge.
[HC-8] EyeMakeYou: Identity- Task- and Subjective-State-Conditioned Diffusion for High-Frequency Gaze Synthesis
链接: https://arxiv.org/abs/2609.04501
作者: Kamrul Hasan,Mehedi Hasan Raju,Oleg V. Komogortsev
类目: Human-Computer Interaction (cs.HC); Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages, 3 tables, 3 figures
Abstract:Eye movement biometrics (EMB) is an emerging behavioral modality for user authentication, particularly in virtual- and augmented-reality systems, where gaze dynamics contain distinctive subject-specific features. However, robust EMB systems require diverse, high-quality gaze recordings that are expensive to collect and often unavailable at the scale needed for model development. Generative models can mitigate data scarcity, but existing methods either synthesize generic gaze behavior or personalize signals primarily by identity, without jointly representing the user’s task and subjective state. Consequently, generated signals may appear visually realistic while failing to retain the behavioral properties required for biometric applications. To address this limitation, we propose EyeMakeYou, a multi-conditional denoising diffusion framework for subject-specific, high-frequency gaze synthesis. EyeMakeYou generates 5-s, 1000-Hz bivariate gaze-velocity sequences from an identity-removed reference trajectory and conditions the denoising process on an identity embedding, a task embedding, and self-reported ratings of overall difficulty, mental tiredness, and eye tiredness. Its objective combines diffusion noise prediction and identity preservation with multi-resolution spectral, drift-consistency, and event-weighted local-smoothness losses. Experiments on GazeBase show that EyeMakeYou achieves higher median spatial accuracy and greater real–synthetic similarity in the embedding feature space than the existing generative approaches, while retaining selected task-dependent associations between subjective reports and oculomotor features. These findings support conditional diffusion as a practical approach for augmenting gaze datasets for biometric and interactive applications.
[HC-9] You Really Didnt Get That? Benchmarking Social Prag matic Inference for Indirect and Playful Chinese Online Comments EMNLP2026
链接: https://arxiv.org/abs/2609.04384
作者: Shiwei Hong,Junjie Ma,Emma Jiren Wang,Ethan Z. Rong,Siying Hu,Haichang Li,Ziying Wang,Zhicong Lu
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: Accepted to the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP 2026), Main Conference
Abstract:Chinese online comments often convey social meaning through indirect and playful language that is hard to interpret without context. Existing evaluations largely organize items around predefined phenomena or controlled pragmatic categories, leaving open whether models can distinguish plausible readings of what a naturally occurring comment is doing in a particular exchange. We introduce a benchmark for evaluating whether LLMs can recover such situated pragmatic meanings. From more than 200,000 public Chinese social media interaction records, we construct 4,735 human-validated diagnostic items, each pairing a target comment with reconstructed preceding context and plausible misreadings. We evaluate eight LLMs as both question writers and solvers in a cross-writer setting. The task is challenging: the strongest model achieves 81.42% leave-writer-out accuracy. Across all eight models, the mean leave-writer-out accuracy is 68.70% while human accuracy was 90.8%. Case analysis shows that models often recognize broad irony or playfulness while misidentifying the mechanism or interactional move.
[HC-10] VLA-Precision: Asymmetric Co-Bootstrapping for Efficient Real-World Online RL of Vision-Language-Action Models
链接: https://arxiv.org/abs/2609.04355
作者: Chenyu Su,Zhaolong Shen,Yuan Qian,Chen Qian,Rui Zhang,Feng Yan,Weixing Chen,Fei Zhang,Jiamin Wang,Shuang Cong,Weiwei Shang
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注: 17 pages, 14 figures
Abstract:Pretrained vision-language-action (VLA) models enable broad manipulation but remain unreliable in tasks demanding precision and repeatability. Applying real-world online reinforcement learning (RL) to VLA post-training enables autonomous trial-and-error improvement beyond demonstrations alone, but exposes two bottlenecks: 1) unreliable value signals can induce policy drift; 2) large-VLA overhead constrains throughput and sample efficiency. To address these challenges, we present VLA-Precision, an efficient real-world online RL framework featuring the Asymmetric Co-Bootstrapping (ACoB) algorithm and the ACoB-Stream architecture. Specifically, ACoB establishes asymmetric co-bootstrapping across timescales: early intervention-guided behavioral learning rapidly improves policy performance while enhancing online experience quality. As autonomous experience accumulates, global return propagation and local preference ranking progressively calibrate value estimates, yielding relative action advantages for reference-regularized policy improvement while suppressing drift. To enable ACoB on large VLAs, we develop ACoB-Stream, a closed-loop experience–policy architecture that establishes invariant-state decoupling and on-demand streaming as design principles, delivering up to 10.9 \times improvements in throughput and computational efficiency. Extensive evaluations on nine high-precision chemistry tasks across four categories and four robot embodiments show that VLA-Precision achieves 98.3% mean success rate in 45.8 min/task, with 27.6 s episodes running at 1.2 \times and 1.8 \times the speeds of VLA and RL baselines. Resources are available at this https URL.
[HC-11] Auditing Bias and Safety in Voice AI Customer Care
链接: https://arxiv.org/abs/2609.04206
作者: Vignesh Ethiraj,Ashwath David
类目: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC); Sound (cs.SD)
备注:
Abstract:Voice AI systems increasingly mediate customer care interactions where caller presentation cues such as accent, affect, fluency, and urgency are available alongside the service request. Existing fairness and safety evaluations cover speech recognition disparities, spoken dialogue bias, and voice agent capability, but rarely treat customer care voice agents as stateful, multi turn, tool mediated systems where harm can appear as additional burden before any final denial occurs. We formalize a validation gated audit framework for such systems. The framework (i) separates native speech to speech, cascaded ASR to language model to TTS, and hybrid tool mediated architectures; (ii) uses matched service facts across controlled caller presentation conditions; (iii) validates fact invariance, presentation cues, artifacts, and acoustic measurements before inference; and (iv) records both material outcomes and path to service burden. We define the research problem, methodology, seven validation gates, a six family metric set, and claim boundaries for an active industry evaluation program. We illustrate the framework with a fully synthetic worked example of a refund dispute audit instance. Production system results are excluded from this release; public reporting is gated by the validation protocol.
计算机视觉
[CV-0] WorldSculpt: Generating Compositional Worlds from Grounded Videos
链接: https://arxiv.org/abs/2609.05416
作者: Muyao Niu,Jixuan He,Ruihan Yu,Lian Fu,Yonghao Yu,Zheng-Hui Huang,Yifan Zhan,Fengbo Lan,Yongtao Ge,Yinqiang Zheng,Kaipeng Zhang,Zhixiang Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Homepage: this https URL Github: this https URL
Abstract:We study the problem of generating a compositional 3D representation of a cluttered scene containing hundreds of objects. The goal is to represent the scene as a collection of individual object meshes placed in a shared world frame, as required by downstream applications such as gaming, AR/VR, simulation, and robotics. This task is challenging in densely cluttered scenes, where objects heavily occlude one another and each view reveals only a fraction of their geometry. Geometry-based approaches typically reconstruct the scene as a single representation and leave incomplete geometry in occluded regions, while existing compositional methods with generative priors are largely limited to relatively simple scenes. We show that complex scenes with hundreds of objects can instead be generated compositionally by adapting a strong single-object 3D generative prior to multi-view observations. We instantiate this paradigm with Pixal3D, extending it with a multi-view conditioning pathway that grounds object generation in multiple posed observations. Although the model is finetuned entirely on single objects in canonical space, it generalizes to large scenes with severe occlusion without any scene-level training, demonstrating the feasibility and scalability of this paradigm. We further introduce UE-MeshyScene, a photorealistic benchmark of densely cluttered scenes with hundreds of objects, per-object annotations, and ground-truth meshes. Across single-object, controlled multi-object, and UE-MeshyScene evaluations, our method consistently outperforms prior approaches, with larger gains as scene complexity and occlusion increase. Finally, we demonstrate broader applicability by converting generated 3DGS worlds, such as Marble and HY-World 2.0, into compositional mesh scenes.
[CV-1] UniMate: One Unified Model to Animate Diverse Skeletons SIGGRAPH
链接: https://arxiv.org/abs/2609.05415
作者: Linzhan Mou,Jiahui Lei,Zhiyang Dou,Chenyue Cai,Chaoyue Song,Adam Finkelstein,Szymon Rusinkiewicz
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Machine Learning (cs.LG)
备注: SIGGRAPH Asia 2026. Project page: this https URL
Abstract:Recent advances in automatic rigging now deliver animation-ready 3D assets at scale, yet generating the motion to drive them remains a bottleneck. Existing learned animators are topology-constrained: they rely on category-specific templates or require per-skeleton fine-tuning and reference motions at inference. We present UniMate, a unified foundation model that synthesizes articulated motion for arbitrary skeletons from a rigged 3D asset and a text prompt, with no test-time optimization or per-skeleton retraining. UniMate introduces a topology-aware diffusion transformer, which integrates skeletal topology into attention via three mechanisms: (1) a graph-aware attention bias from pairwise joint relations and geodesic distances; (2) a spectral rotary position embedding generalizing RoPE to arbitrary kinematic trees via the graph Laplacian; and (3) a global topological conditioner attention-pooled from the rest-pose skeleton. We also curate UniML3D, 13,006 motion sequences spanning bipedal, quadrupedal, avian, marine, insectoid, serpentine, and articulated rigid objects with unified canonicalization and text pairing. Trained on this dataset, UniMate outperforms state-of-the-art baselines in quality, generalization, and efficiency, and supports zero-shot cross-topology transfer, in-betweening, expansion, and text-guided editing. Our project page is available at this https URL.
[CV-2] A Generalizable Feature Extractor for Alzheimers-Related Brain MRI Tasks
链接: https://arxiv.org/abs/2609.05400
作者: Reza Rajabli,D. Louis Collins
类目: Computer Vision and Pattern Recognition (cs.CV); Quantitative Methods (q-bio.QM)
备注: 26 pages (excluding the references section), 7 figures (excluding sub-figures)
Abstract:When there is not enough labeled data to properly train deep learning models, transfer learning can help. We still do not fully understand how effective it is in neuroimaging, especially for Alzheimer’s disease research. It is also not clear if these transferred models can work on new datasets without being retrained for each specific task. We evaluate whether a compact, supervised pretrained model can serve as a reusable foundation model for downstream neuroimaging tasks. We freeze the 7.18 million weights of a 3D CNN previously trained for brain-age prediction, and adapt it to each task using Low-Rank Adaptation (LoRA), requiring only ~1% additional trainable parameters. We evaluate generalizability in six experiments. Adapting the model to classify cognitively normal versus Dementia on ADNI gave an AUC of 0.964 on held-out folds (Experiment #1). Applying that adapted model unchanged to OASIS-3, with no retraining, gave an AUC of 0.871 (Experiment #2). Reusing its output logit together with age and a cognitive score distinguished stable from progressing MCI with an AUC of 0.828 (Experiment #3). Adapting the same backbone to predict amyloid positivity from structural MRI gave an AUC of 0.804 (Experiment #4). Finally, the same approach estimated ICV-normalized hippocampal and white matter hypointensity volumes directly from the T1w image, with R^2 of 0.80 and 0.91 respectively, tasks normally addressed with much larger U-Net networks (Experiments #5 and #6). A compact model supervised on brain age can therefore serve as a reusable backbone, adapting to each task with ~1% additional parameters and transferring to an unseen cohort without any training. Our findings suggest that a carefully trained brain age model can serve as an effective foundation model for Alzheimer’s related tasks, even under strict data constraints.
[CV-3] CrossDepth: Geometry-Constrained Attention for Generalizable Multi-View Surround Depth Estimation
链接: https://arxiv.org/abs/2609.05397
作者: Samer Abualhanud,Max Mehltretter
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Reliable 3D understanding of the surrounding environment is a core requirement for autonomous driving. Multi-view surround camera rigs provide broad scene coverage, but the spatially adjacent images typically overlap only minimally. Consequently, the depth of most pixels must be inferred from monocular appearance cues. These cues can appear differently across images and may therefore be interpreted differently by the depth estimation model. We target two main sources of cross-image inconsistency: differences in camera intrinsics and the limited receptive field of each image. We address the former by conditioning the features on per-pixel camera-aware ray embeddings, enabling the network to account for camera-dependent variations in monocular cues. We address the latter by extending each pixel’s context beyond its own image through cross-image attention constrained to geometrically plausible regions, derived from the calibrated rig setup. The model is trained in a fully self-supervised manner based on photometric consistency. Evaluations on DDAD and nuScenes show improved overall depth accuracy and cross-image depth consistency over state-of-the-art self-supervised methods under in-domain and cross-domain evaluation. Code is available at this https URL.
[CV-4] hink-Verify-Revise: Neuro-Symbolic Visual Reasoning with Vision-Language Models and Dynamic Logic Tensor Networks ECCV2026
链接: https://arxiv.org/abs/2609.05388
作者: Homayoun Afshari,Pietro Basci,Alessandro Russo,Lia Morra
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the MARS2 Workshop @ ECCV 2026
Abstract:Visual reasoning tasks require a system to jointly perceive visual content and apply formal relational constraints—a combination that neither pure neural nor purely symbolic approaches handle well in isolation. This paper proposes a Neuro-Symbolic (NeSy) framework that closes this gap by tightly coupling a Vision-Language Model (VLM) for automatic First-Order Logic (FOL) rule induction with a Dynamic Logic Tensor Network (D-LTN) for differentiable rule verification, in a closed iterative feedback loop. The VLM receives a small set of labelled visual examples and proposes candidate FOL rules conforming to a strict grammar (Think); the D-LTN is automatically assembled from these rules at runtime and evaluates them grounding on CNN-produced visual embeddings (Verify); and verification failures are fed back to guide the VLM’s next hypothesis (Revise). Evaluated on the ViSudo-PC benchmark across four visual domains (MNIST, EMNIST, KMNIST, FMNIST), the system induces valid Sudoku constraint rules using only three training examples as visual context. The proposed method achieves AUC scores matching or outperforming previous methods (NeuPSL, LTN), showing the potential for automatic rule discovery through VLM. Code is available at this https URL.
[CV-5] Reflection-aware Generative Novel View Synthesis ECCV2026
链接: https://arxiv.org/abs/2609.05382
作者: GeonU Kim,Shin Dong-Yeon,Tae-Hyun Oh
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: ECCV2026, Project page: this https URL
Abstract:We propose Ref-GeNVS, a training-free, reflection-aware method for generative novel view synthesis (NVS) in mirror scenes. Existing multi-view diffusion models often fail to recognize the mirror in the scene and cannot exploit reflected content for scene generation. To fix this issue without additional training, our key idea is to treat a mirror image as two complementary views. From input images, we estimate the mirror plane and reflect camera poses to form virtual views. Based on this virtual view setup, we propose a two-stage generation method consisting of Mirror-gated attention and Reflection injection, which enables reflection-consistent NVS by explicitly leveraging reflection relationships in a multi-view diffusion model. Ref-GeNVS inherits the strong generalizability of the multi-view diffusion backbone, while it does not require finetuning. On synthetic and real scenes including mirrors, Ref-GeNVS outperforms recent generative NVS methods by generating reflection-consistent and contextually coherent novel views, revealing scene structure visible only through mirrors. Project page: this https URL
[CV-6] What Matters When? Diagnosing and Improving Conditional Visual Grounding in Visuomotor Imitation Policies ECCV2026
链接: https://arxiv.org/abs/2609.05376
作者: Vivek Chavan,Pengtao Xie,Yahuan Shi,Oliver Heimann,Kevin Haninger,Jörg Krüger
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted as an extended abstract at the DexHAND Workshop, ECCV 2026. Non-archival, non-proceedings. 4 pages, 2 figures, 2 tables
Abstract:Visuomotor imitation policies can achieve high performance under in-distribution visual conditions yet fail when visually similar objects or receptacles are introduced. We study this behavior as a problem of conditional visual grounding: the visual target required for successful control changes with the manipulation phase and, in more complex tasks, with the observed task state. Using Action Chunking with Transformers (ACT), we systematically introduce distractor objects and receptacles with controlled color and shape similarity and localize failures to picking and placement. We find that distractor sensitivity is specific to both the type of visual similarity and the manipulation stage. Guided by this diagnosis, we evaluate distractor augmentation, phase-dependent attention regularization, and appearance-based visual prompting as complementary interventions for improving target selection while preserving spatial information required for control. These interventions substantially improve robustness in simulation and on a physical UR3e. We further examine the same failure pattern in a pretrained vision-language-action policy on a state-conditioned instrument-handling task, where the observed state of a medical instrument determines the correct destination. Together, the results show that visual distractors can cause incorrect object or destination selection even when the underlying manipulation skill remains intact, and that explicitly improving target selection can substantially recover performance across distinct visuomotor policy-learning regimes.
[CV-7] owards Neuro-Symbolic Procedural Reasoning for Long-Horizon Vision-Language-Action Manipulation ECCV2026
链接: https://arxiv.org/abs/2609.05369
作者: Vivek Chavan,Yahuan Shi,Oliver Heimann,Kevin Haninger,Jörg Krüger
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted as an oral presentation at the X-Reason Workshop, ECCV 2026. Non-archival extended abstract. 6 pages, 2 figures, 1 table
Abstract:Vision-language-action (VLA) models can execute short manipulation skills, but remain brittle in long-horizon procedures requiring persistent task state, dependency-aware reasoning, conditional decisions, and reliable grounding. We investigate a neuro-symbolic framework that combines learned VLA control with explicit task graphs and multimodal procedural memory. Task graphs encode action dependencies, valid transitions, and branch conditions, while memory maintains the active step, completed actions, textual context, and task-relevant visual evidence. Together, these structures guide object selection, destination grounding, subgoal dispatch, and verification of expected state transitions. Human demonstrations provide additional spatial and temporal guidance through gaze or saliency cues. To isolate their effect on policy learning, our initial study bypasses cross-view gaze transfer and directly annotates pseudo-gaze in robot-view teleoperation videos. The resulting guidance is used during VLA fine-tuning and inference. We study two long-horizon manipulation domains, workspace clearing and surgical-instrument handling, which require ordered execution, visually grounded decisions, and conditional branching. We evaluate correct-object and destination selection, subtask completion, task progress, step-order consistency, complete-task success, and procedural or execution mistakes. This work positions structured symbolic reasoning and demonstration-derived visual guidance as complementary mechanisms for reliable long-horizon VLA manipulation.
[CV-8] MEOX: Compact Multimodal Mixture-of-Experts for Earth Observation
链接: https://arxiv.org/abs/2609.05351
作者: Mohanad Albughdadi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to IEEE Transactions on Geoscience and Remote Sensing
Abstract:Recent advances in Earth Observation representation learning accommodate heterogeneous sensors and missing observations, often through larger architectures. We present MEOX (Multimodal Earth Observation with eXperts), a multimodal masked autoencoder with a 2.939 million-parameter encoder and 3.115 million parameters in total. Sensor-specific adapters, explicit validity signals, and a shared sparse-expert block preserve modality-dependent processing before a learned patch-wise fusion. Four metadata tokens then accompany a single spatial sequence through fourteen further encoder blocks. Shared expert projections with private low-rank residuals constrain parameter growth, while rotary attention supports downstream spatial grids different from pretraining. The model is pretrained on 1.228 million MMEarth64 samples using modality-balanced masked reconstruction and structured sensor dropout. Frozen transfer is evaluated on six GEO-Bench tasks at both 64 and 224 pixels. The model reaches 64.42% mean intersection-over-union on cashew segmentation at 64 pixels and 90.56% average accuracy on EuroSAT at 224 pixels, exceeding the corresponding reported CSMoE results. BigEarthNet finetuning reaches 72.95% micro-average precision. Routing diagnostics distinguish expert participation, spatial dependence, modality association, and functional contribution. A held-out WorldCover probe measures a 0.64-percentage-point benefit from metadata, while retrieval separates same-sensor semantics from cross-sensor alignment. These results demonstrate sensor-flexible representation learning and strong task transfer using a compact parameter budget.
[CV-9] Lightweight Vision Transformer Compression for On-Device Plant Disease Detection in Resource-Constrained Agricultural Field Conditions
链接: https://arxiv.org/abs/2609.05334
作者: Mahadev Sunil Kumar,Bhavika Gondi,Desaisetty Venkata Satya Sai Swapnith,Gangireddy Rahul Jogi,Sudheesh Manalil,Arnab Raha,Amitava Mukherjee,Parthasarathy Seethapathy,G. Gopakumar
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Chilli (Capsicum annuum) is one of India’s most economically significant crops, yet its productivity is persistently threatened by diseases that are difficult to identify without expert intervention. While Vision Transformers (ViTs) have achieved high classification accuracy, their large computational footprint makes deployment on resource constrained devices challenging. Existing compression approaches typically address pruning, quantization, and knowledge distillation in isolation, leaving the potential benefits and interactions of their combined application insufficiently explored. We propose a unified Vision Transformer compression framework that combines Hessian-Balanced Adaptive Block Pruning (H-BAC), guided by second-order sensitivity estimation, with quantization and attention-based knowledge distillation. To systematically identify the most effective configuration within each compression family, each technique is first evaluated independently through controlled ablation studies, after which the best-performing components are integrated into a sequential deployment pipeline tailored to real-world agricultural constraints. On a chilli 3-class village-split dataset with a genuine cross-village, cross-device out-of-distribution test split, the resulting compressed models match or exceed the 95.13% FP32 baseline’s accuracy, alongside 74-98% model size reduction, and the fully integrated compression pipeline achieves a 54.5x size reduction (327.42 MB to 6.01 MB) at 95.13 +/- 2.32% accuracy across four tested configurations. A direct comparison further reveals that, on this dataset, a directly-trained student of the same final size, without pruning or distillation, reaches comparable accuracy of 94.87%, at the same 6.01 MB INT8 size, indicating where H-BAC and knowledge distillation are, and are not yet shown to be, worth their computational cost.
[CV-10] RoboSPA: Can VLA Models Go Beyond Simple Scenes and Short-Horizon Tasks? EMNLP2026
链接: https://arxiv.org/abs/2609.05324
作者: Zhenxuan Fan,Bo Zhang,Yutong Lin,Yuqian Yuan,Juekai Lin,Liang Liang,Zhuoyi Huang,Wenqiao Zhang,Juncheng Li,Siliang Tang,Jun Xiao,Yueting Zhuang
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the EMNLP 2026 Main Conference
Abstract:Vision-Language-Action (VLA) models have shown promising progress in language-conditioned robotic manipulation. However, existing datasets and benchmarks mainly evaluate task completion under predefined settings, offering limited insight into model reasoning under increasing spatial and procedural complexity. We introduce \textbfRoboSPA (\textbfRobot \textbfSpatial-\textbfProcedural \textbfAssessment), a large-scale robotic manipulation dataset and benchmark for diagnosing embodied reasoning in VLA models. \textttRoboSPA focuses on two core dimensions, Fine-Grained Spatial Reasoning and Long-Horizon Procedural Planning, covering 10 task categories and 56 base tasks. Each task is instantiated across five difficulty levels, yielding 280 variants with increasing spatial ambiguity and procedural complexity. We collect 527K trajectories across multiple embodiments and diverse scenes. Beyond binary success rate, \textttRoboSPA introduces diagnostic metrics for more detailed evaluation. Experiments on representative VLA models show that current systems still struggle with complex spatial relations, precise low-level execution, and memory-intensive planning. These results establish \textttRoboSPA as a challenging diagnostic benchmark for developing more capable, reliable, and generalizable embodied agents. Our data and code are available at this https URL.
[CV-11] Scalable Detection of Fossil Palynomorphs in Multifocal Digital Microscopy Images
链接: https://arxiv.org/abs/2609.05323
作者: Abbas Shaikh,Praise Mayor,Patrick Ainlay-Vazquez,Aditya Viswanathan,Teon Golden,Eric Zhang,Ingrid C. Romero,Alexander E. White,Scott Wing,Arko Barman
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Palynomorphs (microscopic, organic-walled fossils such as pollen, spores, and dinoflagellates) are important high-resolution records of past climates and are critical to the study of ancient ecosystems. Existing methods rely on manual analysis of high-resolution, multifocal digital microscopy images, which is slow and time-consuming and requires researchers to compromise on the scale of their investigations. To the best of our knowledge, our work proposes the first ever scalable end-to-end pipeline for automated palynomorph detection in whole slide images that addresses this bottleneck through: (1) efficient methods for decomposing and compressing digitized multifocal microscope slide images into tractable 2-dimensional tiles for analysis; (2) benchmarking modern object detection models, including RF-DETR, for the detection of palynomorphs, achieving an AP@50 of 0.879; (3) an efficient algorithm for the synthesis of detection outputs across large-scale, high-resolution images; and (4) an I/O optimization resulting in faster inference time. Our methods drastically reduce the time required for palynomorph detection in a single slide from often days of manual inspection to under one hour of automated analysis, enabling palynological research at a substantially greater scale.
[CV-12] Adaptive Gated Deepfake Detection for Low-Resolution and Resource-Constrained Environments
链接: https://arxiv.org/abs/2609.05320
作者: Vaishnavi Sen,Cody Laurie,Rashida Hasan
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Deepfake detection models often rely on high-quality inputs, fixed inference paths, and computationally expensive architectures, limiting their use in low-resolution and resource-constrained settings. This paper proposes AdaGate-DF, an adaptive gated deepfake detection framework that uses image-quality cues to route samples through a dual multi-exit system so high-quality images can exit earlier and save compute. We evaluated AdaGate-DF against MaD-CoRN, DefakeHop++, and ShuffleNetV2 on two benchmark datasets (Celeb-DF and FaceForensics++) under multiple configurations to test image resolution dependence and training and inference efficiency. On Celeb-DF, AdaGate-DF achieves an AUC of 0.9370, outperforming MaD-CoRN and DefakeHop++ while maintaining a low inference latency. Resolution-based testing shows consistent improvement as input resolution increases, reaching an AUC of 0.9708 at 384 by 384. The FaceForensics++ results highlight that AdaGate-DF remains effective under class imbalance, following competitive results with evaluated models. Overall, AdaGate-DF demonstrated a practical balance between detection performance, uncertainty-aware prediction, and computational efficiency for variable-quality deepfake detection.
[CV-13] Learning Spatial-Spectral Refinement and Calibrating Complementary Observations for Hyperspectral Image Super-Resolution
链接: https://arxiv.org/abs/2609.05303
作者: Liqian Yang,Xingchi Chen,Xinfeng Gui,Xiangyong Cao,Qianxin Yi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Hyperspectral and multispectral image fusion (HMIF) aims to reconstruct a high-resolution hyperspectral image (HR-HSI) by combining the fine spatial details of a high-resolution multispectral image (HR-MSI) with the rich spectral information of a low-resolution hyperspectral image (LR-HSI). Recent advances in implicit neural representations (INRs) have enabled flexible coordinate-based modeling for HMIF; however, existing INR-based approaches may not fully capture fine-grained spatial structures and rich spectral dependencies. Moreover, the LR-HSI and HR-MSI are primarily incorporated through degradation-consistency constraints, leaving their complementary information underexploited. To address these limitations, we propose Two-Stage Reconstruction with Implicit Tensor Neural Representation (TSR-ITNR), a unified self-supervised framework integrating representation refinement and observation-guided calibration. In Stage 1, TSR-ITNR learns an implicit Tucker representation and refines its low-rank spatial coefficient tensor and spectral basis to better capture fine spatial structures and interband correlations. A fixed pretrained denoiser further provides a deep prior for the preliminary reconstruction. In Stage 2, parameter-free calibration derives complementary and noninterfering corrections from both observations to recover information insufficiently captured in Stage 1. Theoretical analysis establishes the geometry-preserving property of spectral refinement and the orthogonal complementarity of calibration. Extensive experiments on multiple benchmark datasets demonstrate strong quantitative, visual, and spectral reconstruction performance without ground-truth HR-HSI supervision. Beyond conventional reconstruction metrics, we further assess the effectiveness of TSR-ITNR using downstream semantic segmentation accuracy.
[CV-14] Compact Neural Appearance Models for Efficient Gaussian Splatting
链接: https://arxiv.org/abs/2609.05255
作者: Florian Hahlbohm,Jorge Condor,Linus Franke,Martin Eisemann,Marcus Magnor
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: Project page: this https URL
Abstract:Explicit primitive-based radiance fields such as 3D Gaussian Splatting typically model view-dependent appearance using low-order spherical harmonics (SH). Although efficient to evaluate, SH coefficients dominate per-primitive storage and memory traffic, while their band-limited basis restricts angular detail. We present a thorough, end-to-end comparison of SH and recent spherical appearance models and introduce an implicit alternative that decodes compact per-primitive latent codes using a tiny shared MLP. We integrate all models into the same optimized pipeline, fusing their forward and backward passes into a differentiable CUDA rasterizer and provide a portable WebGL viewer for laptop and mobile GPUs. Our evaluation across reconstruction quality, memory use, and optimization and rendering performance shows that recent spherical models offer the strongest overall quality-efficiency trade-off. Our neural representation is the most compact model evaluated and, compared to third-degree SH, reduces the per-primitive appearance footprint from 192 to 28 bytes, accelerates optimization by 1.3 \times , while improving reconstruction quality. We further analyze how appearance parametrization shapes optimization, identifying differences in recovered geometry and the tendency of expressive models to absorb non-static scene content. Together, our framework and analysis provide practical guidance for replacing SH beyond what image metrics alone can capture.
[CV-15] Few-Shot Video Recognition via Hierarchical Metric Learning
链接: https://arxiv.org/abs/2609.05242
作者: Jiaxin Zhang,Haoran Gao,Xizhan Gao,Zihao Dong,Tingwei Wang,Sijie Niu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Few-shot action recognition (FSAR) aims to recognize unseen action categories with only a small number of annotated video samples. Recent works typically apply single-prototype supervision at the network output and fail to sufficiently exploit rich cross-frame global spatial information in videos. Even existing multi-level metric schemes only impose parallel prototype constraints on intermediate layers, without progressive supervision along the full feature pipeline, which results in limited generalization ability of the learned class prototypes. Inspired by this, we present a novel method, hierarchical metric learning for few-shot action recognition (HML-FSAR). First, a spatial-enhanced module is developed to capture cross-frame global spatial representations. Combined with temporal MHA, heterogeneous alignment, spatial-temporal feature fusion and dictionary learning modules, it constructs the complete feature processing pipeline. Second, a hierarchical metric learning (HML) strategy is embedded into HML-FSAR. Composed of center metric, alignment metric, contrastive metric, dictionary metric and prototype metric, HML imposes progressive multi-stage complementary constraints from frame-level representations to final class prototypes, so as to jointly optimize feature compactness, heterogeneous spatial-temporal alignment, inter-class discriminability and anti-noise robustness. The proposed HML-FSAR method is validated on five widely-used FSAR datasets, and experimental results fully demonstrate its effectiveness.
[CV-16] Cross-Domain Tracker Adaptation Without Target-Domain Labels via Vision-Language Agents
链接: https://arxiv.org/abs/2609.05239
作者: Daniel Davila,Ravikumar Balakrishnan,Mike Cochran
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We present a system that uses a Vision-Language Model (VLM) as a diagnostic agent for adapting a detect-to-track pipeline to a new target domain without access to target-domain labels. Rather than optimizing against annotated metrics, the VLM directly inspects rendered tracking outputs, identifies visual failure modes, and recommends parameter updates through an iterative tuning loop. We first demonstrate that ground-truth-supervised hyperparameter transfer can be brittle. On MOT17-MOT20, applying a source-derived oracle configuration reduces mean HOTA by 0.090, from a target-domain ceiling of 0.357, to 0.267. Without using any target-domain labels, our VLM-based tuner recovers 67.8% of this lost headroom, finishing within 0.029 HOTA of the target ceiling; on the highest-density target sequence, it recovers up to 86.7%. We further show that label-free Bayesian optimization with handcrafted proxy objectives struggles under large domain shifts and can degrade configurations that are already strong. In contrast, the VLM tuner acts selectively: when its visual diagnosis reveals no clear failure mode, it declines to modify the configuration, preserving performance on easy transfers while improving hard ones. Finally, we characterize the conditions under which this approach succeeds, namely, when domain shift manifests through exposed detection-level parameters, versus where it is less effective, such as MOT17-DanceTrack, where the source oracle is already near-optimal.
[CV-17] Measured Sliders: Learning Continuous Controls from Differentiable Image Measurements
链接: https://arxiv.org/abs/2609.05234
作者: Yijia Chen,Boyu Wei,Xuanhua Yin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 5 figures, 2 tables
Abstract:Continuous sliders are useful only when coefficient changes produce predictable image changes. Yet most diffusion sliders derive their axes from text or learned representations, leaving their scales disconnected from observable image properties. Consequently, we cannot tell in advance which attributes are learnable, compare control strengths directly, or anticipate interference when multiple controls are combined. We propose Measured Sliders, a framework that defines continuous controls through closed-form differentiable image measurements. A common measurement space unifies the pipeline. Before training, an observability test identifies usable supervision. During training, a measurement-guided objective learns target movement while suppressing non-target changes. After training, decoded calibration expresses controls in comparable units of realized image change. Multiple LoRA branches are stored in one checkpoint and composed without training on joint activations. Across SDXL and FLUX.1-dev, the resulting controls are ordered, selective, and composable. On 553 prompts, lighting direction reaches rho = 0.995 and 98.9% monotone sweeps. A five-attribute checkpoint achieves average selectivity 2.59, compared with 1.50 for the strongest baseline, and preserves every requested direction in 96.7% of pair and 86.1% of triple compositions. The observability test also separates every subsequently successful measurement from the failed candidate. Overall, image-space measurement provides a common basis for learning, diagnosing, calibrating, and composing continuous generative controls.
[CV-18] First Things First: Teaching LLM -Based Agents to Prioritize Must-Haves before Nice-to-Haves EMNLP2026
链接: https://arxiv.org/abs/2609.05224
作者: Tianjie Ju,Xinyue Xu,Wanxuan Sun,Lingxiao Diao,Gongshen Liu,Zhuosheng Zhang,Cheng Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at EMNLP 2026 (Findings)
Abstract:Recent progress in multimodal large language models (MLLMs) has fueled significant enthusiasm in their potential to act as autonomous agents for real-world tasks. However, scenarios requiring agents to fulfill users’ complex, structured requirements remain largely underexplored. In this work, we examine reasoning tasks under three distinct requirement scenarios: (i) Must-have requirements uniquely determine a unique feasible solution; (ii) Multiple answers satisfy the must-have requirements and are prioritized via the nice-to-have requirements; and (iii) No candidate solution satisfies the must-have requirements, in which case the agent should abstain from generating a response. We evaluate state-of-the-art MLLMs on 3,649 carefully constructed problems that reflect realistic service scenarios, including e-commerce, booking, and map-based or ride-hailing. Our evaluation reveals that existing MLLMs exhibit catastrophic failures in all scenarios. They frequently misinterpret task requirements, violate must-have requirements, and produce invalid solutions. To address this critical gap, we propose First Things First Reinforcement Learning FTF-rl that explicitly optimizes reasoning over multi-priority user requirements. Experimental results show that our method substantially improves the task success rate compared to strong baselines. Moreover, FTF-rl yields general effectiveness on popular logical and mathematical reasoning tasks, including LogicVista, MathVision, and InfoQA. Our findings suggest that enhancing requirement-aware reasoning capability provides a simple yet effective pathway to improve generalization of MLLM agents. Code and dataset are available at this https URL.
[CV-19] BLASt3R: Bundle Adjustment of Any Image Set with Multi-View Matching and Monocular Priors ECCV’26
链接: https://arxiv.org/abs/2609.05210
作者: Vincent Leroy,Philippe Weinzaepfel,Lojze Zust,Yohann Cabon,Jérome Revaud
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV’26
Abstract:Recent hybrid Structure-from-Motion (SfM) systems combine the robustness of feed-forward 3D reconstruction with the accuracy of traditional bundle adjustment (BA) with pixel matching. They are usually the best performing methods however their scalability and usability remains limited since estimating dense correspondences between views is prohibitively costly, especially considering time constraints inherent to online applications like Visual SLAM (VSLAM). In this paper, we introduce a regularized BA framework that leverages a fast multi-view matcher and monocular priors for initialization and regularization. In contrast to existing systems, our unified approach seamlessly supports both online VSLAM and offline reconstruction from unordered image collections within the same optimization framework and sharing common hyperparameters for all tasks. Extensive experiments across both domains demonstrate improved performance and speed tradeoffs over traditional, feed-forward, and hybrid baselines. Notably for VSLAM, our uncalibrated method outperforms all previous calibrated approaches.
[CV-20] Conserved Immune Topology Improves Pathology Foundation Model Generalization for Cross-Cancer MSI-H Prediction ECCV2026
链接: https://arxiv.org/abs/2609.05182
作者: Dasari Naga Raju
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the ECCV 2026 Workshop on Medical Foundation Models and Benchmarks (MedFM-Bench). 15 pages, 2 figures
Abstract:Pathology foundation models integrated with multiple instance learning achieve competitive accuracy within single-cancer cohorts, yet cross-cancer generalization remains unresolved due to organ-specific histological and architectural differences. In this paper, we propose Conserved Immune Topology (CIT), a lightweight spatial representation for cross-cancer MSI-H prediction that augments foundation-model embeddings with biologically motivated immune descriptors. CIT uses unsupervised clustering to identify immune-associated tiles, then encodes tertiary lymphoid structures, peritumoral immune reactions, multi-scale tumor-infiltrating lymphocyte density, and immune-tumor mixing from frozen foundation-model embeddings and tile coordinates without requiring annotations or target-domain data. The proposed method was evaluated under cross-site and cross-cancer settings using CPTAC-COAD and TCGA-STAD cohorts, which introduce scanner variability, distribution shifts, and organ-specific architectural variations. Zero-shot cross-cancer transfer with CIT increased TransMIL AUC from 0.6627 to 0.7161, an absolute gain of 0.0534 (p=0.003), with consistent improvements across all three MIL aggregators. These results suggest that spatial immune topology provides potentially an organ-invariant representation for MSI-H prediction, supporting cross-cancer generalization of pathology foundation models.
[CV-21] SMILE: Self-Explainable Multimodal Information Bottleneck for Medical Diagnosis
链接: https://arxiv.org/abs/2609.05174
作者: Yuqing Yang,Alexander Schmatz,Zhaozhao Ma,Changkyu Choi,Robert Jenssen,Shujian Yu
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 30 pages, 10 figures
Abstract:Explainability is increasingly seen as a crucial requirement in AI-based medical diagnosis, particularly in safety-critical clinical decision-making. Most existing explainability methods in healthcare operate in a post-hoc manner and are predominantly designed for unimodal data, which limits their applicability in increasingly prevalent multimodal diagnostic settings. This paper addresses the problem of self-explainable multimodal diagnosis by formulating it within the information bottleneck (IB) framework. We propose a unified learning paradigm that jointly optimizes predictive performance and modality-specific explainability by identifying the most informative elements inside each modality that contribute to diagnostic decisions. To enable tractable and stable optimization, we employ a matrix-based Renyi’s \alpha -order entropy functional under the assumption of sufficiently expressive encoders. Extensive experiments on representative medical datasets spanning heterogeneous modalities demonstrate that the proposed method consistently achieves strong diagnostic performance, including an absolute accuracy improvement of 9.1 percentage points on the iCTCF dataset. Moreover, the learned explanations provide transparent and modality-aware insights into feature relevance, thereby improving both the explainability and generalization.
[CV-22] WeAgent -MMGenEdit: A Full-Stack Recipe for Multimodal Agent ic Image Generation and Editing
链接: https://arxiv.org/abs/2609.05171
作者: Hui Zhang,Zongkai Liu,Liqiang Niu,Juntao Liu,Han Li,Zhen Cao,Wenchao Chen,Chengduo Zhao,Fandong Meng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Image generation and editing models have advanced rapidly, yet remain unreliable when prompts require external world knowledge. Bounded and long-tail parametric knowledge prevents direct or reason-then-generate approaches from recovering the required facts and visual appearances. Existing agentic generation and editing methods mitigate this limitation with retrieval tools, yet remain constrained by insufficient visual verification, overloaded policy models, and weak integration of retrieved textual and visual evidence. To address these limitations, we present WeAgent-MMGenEdit, a full-stack recipe including a multimodal harness, a scalable data construction pipeline, a comprehensive benchmark, and post-training methods for the agent policy and image backend. We first introduce WeAgent-Harness, a multimodal runtime with persistent evidence management and dedicated verification and integration tools that organize retrieved multimodal evidence into a dense carrier. Upon this, we develop a scalable pipeline for prompt synthesis and agentic trajectory collection, yielding 23K supervised trajectories and 14.7K RL tasks with three-layer verifiable checklists. We further introduce WeBench-MMGenEdit, a bilingual benchmark covering both knowledge-intensive image generation and multi-image editing. Finally, a two-sided post-training recipe based on SFT and RL improves the agent policy and image backend. Together, WeAgent-MMGenEdit enables a 30B-total/3B-active policy to outperform similarly sized policy models and approach the performance of a 1T-parameter agent.
[CV-23] VoxelFix: Post-Hoc Semantic Correction of Completed 3D Voxel Maps
链接: https://arxiv.org/abs/2609.05114
作者: Sunesh Praveen Raja Sundarasami,Taehyoung Kim,Johannes Scherer,Tomaž Cotič,Sivasubiramaniam Subbiah,Andreas Greiner,Paul Spannaus,Sebastian Houben
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Semantic 3D maps are increasingly constructed automatically for aerial robotics by integrating learned semantic predictions into 3D representations. While this avoids costly manual 3D annotation, errors in the perception and mapping pipeline can persist in the resulting map, reducing its reliability for downstream autonomous tasks. Existing 3D semantic map refinement methods either rely on the original observations, treat occupancy as part of the prediction problem, or apply non-learned local regularization to completed maps. Instead, we study post-hoc semantic correction, asking whether semantic accuracy can be recovered directly from the completed map while keeping its geometry and occupancy fixed. We introduce \method, a graph-based model that corrects voxel labels based on local geometry and neighboring semantic information. To obtain training pairs, we corrupt contiguous regions of annotated OccuFly maps according to class confusions observed in upstream maps. We evaluate \method on completed OccuFly maps generated from predictions of four independently trained 2D segmentation models. \method consistently improves mIoU by 4.23–5.00 percentage points, with gains broadly distributed across the evaluated semantic classes and particularly strong improvements for tree, roof, and wall. Results on an independently reconstructed out-of-distribution aerial scene further suggest that the learned correction can transfer beyond the environments seen during training.
[CV-24] raining-Free Logical and Structural Anomaly Detection via Calibrated Fusion
链接: https://arxiv.org/abs/2609.05091
作者: Changyi Li,Miao Yu,Kai Dong,Yu Xiao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by PRCV 2026
Abstract:Industrial anomaly detection must handle two distinct defect families: structural anomalies, which manifest as local texture corruptions, and logical anomalies, which violate global rules on object count, composition, or arrangement. Existing detectors typically favor one family at the expense of the other. In particular, training-free methods effectively exploit frozen representations but lack an explicit notion of object count, while methods that reason about counts usually rely on category-specific component modeling. We show that counting ability can be introduced into training-free anomaly detection without additional training or part-level supervision. Our key idea is a normal-set calibration that aligns heterogeneous anomaly cues using statistics from normal images, enabling their direct fusion within a unified training-free framework. Built upon this calibration, our detector combines complementary frozen cues to address both logical and structural anomalies. On MVTec-LOCO, our method achieves image-level AUROCs of 89.0 and 95.9 on logical and structural anomalies, respectively, yielding a 92.5 average—the best among training-free detectors in our comparison. It remains competitive with methods requiring network training or part annotations, while its structural variant matches PatchCore on MVTec-AD (99.1 image-AUROC), suggesting that the proposed calibration generalizes beyond logical anomaly detection.
[CV-25] MultiAttenGastro: Multi-Dimensional Attention Augmentation for Gastrointestinal Endoscopy Classification
链接: https://arxiv.org/abs/2609.05070
作者: Sadhana Devarajan,Praveen Kumar Chandaliya,Dhruvin Jashvant Kumar Shah,Kishor Upla,Kiran Raja
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 17 Pages, 2 Figures, CVIP2026
Abstract:Automated gastrointestinal (GI) endoscopy classification requires models that generalize across diverse modalities and class distributions, often far from natural-image pretraining. We propose MultiAttenGastro, a plug-and-play attention framework with parallel 1-D channel, 2-D spatial, and 3-D contextual heads, and present the first systematic cross-dataset evaluation across eight CNN and transformer backbones on five public GI datasets (80 backbone–dataset runs). We find that attention effectiveness is not universal but tracks the representational gap between ImageNet features and the target distribution: MultiAttenGastro improves 6 of 8 backbones on Kvasir-Capsule (14-class WCE, large gap; best macro F1 98.33%), is uniformly negative on the small-gap Kvasir-v2 benchmark (0/8), and shows mixed outcomes on datasets with intermediate gap. Five-seed ablation on the strongest case (Kvasir-Capsule, ConvNeXt-Tiny) shows this improvement is directionally consistent, but not statistically decisive (paired t : p=0.47 ; Wilcoxon: p=0.63 ), and that individual attention heads are not uniformly beneficial in isolation only their combination yields a positive mean effect. Centered Kernel Alignment (CKA) analysis links this pattern to representational redundancy: low inter-head CKA under large domain gaps coincides with the framework’s only consistent gains, while high redundancy under small gaps coincides with its losses. We report these results, including the non-significant margins, as evidence for when and why multi-dimensional attention helps GI endoscopy classification, rather than as a claim that MultiAttenGastro is a strictly superior architectural choice.
[CV-26] Adaptive Multi-Granularity Temporal Modeling for Weakly Supervised Video Anomaly Detection
链接: https://arxiv.org/abs/2609.05066
作者: Changyi Li,Yu Xiao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted by PRCV 2026
Abstract:As the scale of video surveillance data outpaces manual annotation capacities, weakly supervised video anomaly detection (WSVAD) has emerged as a critical research frontier. Most existing approaches formulate WSVAD within a Multiple Instance Learning (MIL) framework that relies on rigid, hand-crafted temporal priors to supervise anomaly scoring. However, such formulations exhibit limited adaptability to the wide variation in anomaly durations and temporal dynamics observed in real-world videos, often leading to unstable or unreliable snippet-level predictions. To address this limitation, we propose an adaptive temporal modeling framework for WSVAD that explicitly accounts for variations in video dynamics across multiple temporal granularities. First, we introduce a Temporal Refinement Module (TRM) that leverages dynamic positional encoding and a learnable class token to model long-range temporal dependencies while distilling a stable global video-level representation. Second, to capture anomalous events with varying frequency and duration, we develop an adaptive Event Segmentation Module (ESM) that identifies event boundaries through temporal discontinuity analysis and aggregates snippet features into discriminative event-level representations. Finally, for snippet-level and event-level predictions, we propose an adaptive similarity-based fusion strategy that dynamically integrates anomaly scores into video-level predictions, replacing fixed top-k aggregation heuristics with global semantic relevance. Extensive experiments on two benchmarks demonstrate that the proposed framework consistently outperforms state-of-the-art methods.
[CV-27] Efficient Multi-Timescale Event Representations for Feed-Forward Object Detection ECCV2026
链接: https://arxiv.org/abs/2609.05049
作者: Fredrik Lundell,Per-Erik Forssen,Mårten Wadenbäck,Astrid Lundmark
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to the Workshop on Neuromorphic Vision (NEVi) at ECCV 2026, 16 pages
Abstract:Autonomous systems require robust low-latency perception under rapidly changing scene dynamics and challenging illumination. In event cameras object detection commonly relies on recurrent architectures to accumulate sparse temporal information over time. This work investigates how temporal information can be encoded directly within the event representation. We propose a confidence-normalized continuous multi-timescale representation based on logarithmic B-spline temporal encoding together with a geometry-aware local confidence mechanism that exploits the spatial structure of event generation. Using a fixed feed-forward EventCenterNet detector, we show that the proposed representations consistently outperform the compact CSTR representation on PEDRo and Gen1 datasets. We further introduce a recursive exponential-polynomial approximation that enables efficient event-by-event updates while largely preserving detection performance. These results demonstrate that carefully designed event representations can capture a substantial portion of the temporal information learned through recurrent temporal modeling, providing a promising foundation for efficient feed-forward, event-driven, and future neuromorphic object detection.
[CV-28] PuTR-CouT: Counting-by-Tracking in Camera-Trap Image Sequences
链接: https://arxiv.org/abs/2609.05038
作者: Fagner Cunha,Juan G. Colonna,Eulanda M. dos Santos
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Species identification in camera trap images has been widely studied, but key ecological modeling tasks such as species abundance or density estimation also require counting individual animals. However, the lack of counting labels in most datasets and low frame rates (typically ~1 frame per second) make sequence-level tracking and count estimation particularly challenging. In this work, we present PuTR-CouT, a counting-by-tracking framework built on a transformer-based learned association mechanism for sequence-level animal counting in camera trap images. To address the scarcity of annotated tracking data, we generate synthetic training data by exploiting structural priors, such as static backgrounds and short temporal bursts, to heuristically create pseudo-tracking labels in a weakly supervised manner. The resulting tracker associates detections across frames, using these tracks to estimate per-species counts. We also refine the MaxBoxCount heuristic used by the top solutions of the iWildCam 2021 challenge as a strong baseline, setting the highest score reported to date. When evaluated on the iWildCam 2021 benchmark, our framework PuTR-CouT delivers competitive counting results compared to the improved MaxBoxCount, with the added capability of multi-species predictions and track-level verification.
[CV-29] Compositional Reward Models for Conditional Medical Image Generation
链接: https://arxiv.org/abs/2609.05028
作者: Aayush Kumar Tyagi,Prathosh A.P.,Mausam
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Acquiring high quality annotated medical image data is critical for training deep learning models; however, annotation is expensive, time consuming, and requires domain expertise. Conditional diffusion models, such as ControlNet, offer an alternative by generating images conditioned on semantic masks and text. However, existing approaches fail to capture fine grained properties (e.g., intensity and texture), as well as semantic consistency expected by domain experts, limiting their effectiveness for downstream tasks. Recent attempts to address these issues using reinforcement learning fine-tuning remain limited due to the reliance on a single scalar reward, which conflates diverse failure modes and provides weak corrective signals. We propose PRISM, a Compositional Reward Model (CRM) framework for conditional medical image generation. Instead of assigning a single reward, we decompose image quality into verifier grounded stages, each evaluating a distinct aspect of correctness from fine to coarse properties, including low level attributes (intensity and texture), structural alignment with conditioning inputs, and high level semantic fidelity. These stage wise rewards are composed through a Hierarchical Constrained Propagation (HCP) mechanism that enforces a fine to coarse notion of correctness, ensuring that lower level deficiencies are resolved before higher level rewards are accrued, preventing easier objectives from masking critical failures. We evaluate PRISM across three datasets spanning diverse medical imaging tasks: PanNuke (multi-class cell segmentation), CeDeM (villi/crypt detection and measurement), and ISIC (skin lesion classification). Training downstream models with data generated by PRISM yields improvements over closest baselines, including a 2.3% increase in mDice on PanNuke, a 8.5% reduction in Mean Relative Error (MRE) on CeDeM, and increases ISIC F1 by 5.9%. Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.05028 [cs.CV] (or arXiv:2609.05028v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.05028 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-30] mporal Residual Neural Radiance Fields for Monocular Video Dynamic Human Body Reconstruction
链接: https://arxiv.org/abs/2609.04984
作者: Tianle Du,Jie Wang,Xiaolong Xie,Wei Li,Pengxiang Su,Jie Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Published in Journal of Electronic Imaging, 2024
Abstract:In the field of computer vision and graphics, high-quality reconstruction of the human body in static scenes has been achieved in recent years by a single multilayer perceptron (MLP) in a number of approaches. However, MLPs have capacity limitations, requiring substantial training time and computational resources for dynamic scene reconstruction. And the quality of reconstruction is significantly constrained. This paper proposes a method for effectively processing complex spatiotemporal signals in dynamic scene human 3D modeling. The proposed method uses Temporal Residual Neural Radiance Fields to achieve novel view rendering and new pose synthesis of human this http URL address the problem of representing temporal signals in video sequences, we construct a temporal residual field which is not related to the MLP architecture. Secondly, to improve reconstruction efficiency, we propose an integrated approach that reduces trainable parameters and accelerates rendering, thereby enhancing the network’s feature representation capability. Finally, we design a multi-dimensional loss function to accurately measure the loss between predicted and actual spatial pixel values. The experimental results show that our proposed approach improves the peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) accuracy metrics compared to the latest representative methods. It maintains similar accuracy to Anim-NeRF and Neural Body while achieving a nearly 780-fold increase in time efficiency.
[CV-31] RefDiT: Local Attribute Guidance in Reference-Based Image Generation
链接: https://arxiv.org/abs/2609.04976
作者: Rameshwar Mishra,Srikrishna Karanam,A V Subramanyam
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Personalization models generate new images guided by a few subject references, while style transfer methods aim to produce images aligned with a global style derived from a reference image. Recent approaches perform well when the reference image contains a single object, effectively capturing a global style that encompasses all implicit attributes. However, when applied to complex real-world scenes containing multiple objects with distinct attribute characteristics, these methods, due to their global-level guidance, fail to localize relevant elements in the reference image. The global guidance restricts their ability to generate new images based on the local attributes in the reference image. Moreover, existing methods typically employ a single identifier token to capture all details from the reference, resulting in a lack of individual, attribute-level control. Motivated by these limitations, we propose RefDiT, a novel framework for reference-guided image generation. RefDiT takes as input a reference image, a text prompt, and an optional user-provided guidance context. RefDiT employs local region guidance using the attributes of local elements. It constructs an attribute-aware conditioning signal from the reference image by performing attribute-level decomposition of the identifier token and performs context adjustment in the inference prompt to train low-rank adapter (LoRA) blocks of a diffusion transformer (DiT)-based generative model. RefDiT learns the correspondence between identifier tokens and local regions in the reference image, enabling more effective local guidance.
[CV-32] ARC-Loc: Leverag ing Azimuthal Ray Convergence as a Geometric Cue for Direct Cross-View Localization ECCV2026
链接: https://arxiv.org/abs/2609.04965
作者: Hyeongsik Kim,Mincheol Kim,Heejoon Moon,Je Hyeong Hong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV2026
Abstract:Cross-view localization (CVL) estimates the pose of a ground image by matching it to a geo-referenced satellite image. To bridge the extreme viewpoint gap, mainstream pipelines rely on Bird’s-Eye-View (BEV) transformations or 2D-to-3D lifting. However, deriving 3D structures from a single ground image is fundamentally ill-posed, causing these methods to endure geometric distortions and computational costs during 3D lifting or BEV projection. Furthermore, relying on external depth foundation models to resolve this introduces latency and remains susceptible to noisy predictions. In this work, we present a different approach inspired by a human navigation technique called resection, that can perform direct ground to satellite image matching and localization without relying on external depth foundation models. The key insights of our method are that (i) ground keypoints can be translated into azimuthal rays on the satellite map, and (ii) these rays ideally converge at the user location. Exploiting this geometric constraint through direct line-to-point correspondences, we introduce a minimal Azimuthal Ray Convergence (ARC) solver to identify the intersection, alongside an ARC loss to optimize the matching network. By eliminating dependencies on computationally heavy BEV transformations and external depth foundation models, our approach achieves faster, memory-efficient inference, while its explicit feature matching ensures straightforward compatibility with existing frameworks. Experiments on VIGOR and KITTI demonstrate that ARC-Loc maintains competitive localization accuracy compared to recent approaches, highlighting its practicality.
[CV-33] MINT: A Unified Model for World-Space Camera and Hand Motion Estimation from Scalable Egocentric Pipeline Supervision
链接: https://arxiv.org/abs/2609.04958
作者: Zijie Zhu,Weiren Cai,Yizhou Wang,Zhenjie Yang,Yide Liu,Jiahao Chen,Guanqi He
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Recovering camera and hand motion in world coordinates from egocentric video is a key capability for activity understanding, robot learning, and augmented reality. Existing systems typically decompose this problem into separate stages for camera motion, depth, hand reconstruction, and trajectory refinement, resulting in substantial computational overhead and preventing the joint modeling of camera and hand motion. We introduce MINT (Minting IN-the-Wild Trajectories), the first foundation model that directly produces complete world-space two-hand trajectories from ego-centric RGB video. From a single shared spatiotemporal video representation, MINT jointly predicts the camera trajectory, camera-frame hand states, and per-frame hand presence, and then produces world-space hand motion via explicit coordinate transformations. Training such a model at scale is challenging, since paired world-space camera and hand annotations are scarce. We therefore develop an open-source labeling EGOPIPELINE that converts large collections of public egocentric videos into structured camera-and-hand trajectory supervision. MINT is first pretrained on these large-scale pseudo-labels and then fine-tuned on a small set of high-quality joint annotations. Across public benchmarks, MINT achieves [xxx] improvement in world-space hand trajectory accuracy, [xxx] improvement in camera trajectory estimation, and [xxx] faster end-to-end trajectory generation than the labeling pipeline, while generalizing zero-shot to unseen egocentric datasets. We release the model, training and inference code, labeling pipeline, and a curated 1,021-hour egocentric trajectory dataset.
[CV-34] VICAL: Vicinal Consistency Alignment for Long-Tailed Visual Recognition ECCV2026
链接: https://arxiv.org/abs/2609.04948
作者: Jiangang Zhu,Zheng Wang,Bin Zhu,Yi-Ping Phoebe Chen,Jingjing Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to ECCV 2026
Abstract:Multi-expert models have become the dominant paradigm for long-tailed learning, largely attributed to their presumed ability to benefit from expert diversity. However, we revisit this central assumption and reveal that diversity induced by logit adjustment or explicit regularizers does not guarantee better ensemble accuracy. Our work suggests that multi-expert models benefit more from variance reduction than diversity maximization. We introduce \textbfVICAL, a \textbfVIcinal \textbfConsistency \textbfALignment framework that improves long-tailed recognition not by enforcing expert diversity, but by reducing prediction variance. Specifically, our approach comprises two key components: Self-Consistency Learning and Deep Ensemble Distillation. Self-Consistency Learning discourages reliance on unstable high-frequency information, smoothing the local loss landscape and mitigating overfitting, especially for tail classes. Deep Ensemble Distillation promotes cross-expert low-frequency semantic agreement using a low-resolution view, thereby sidestepping optimization conflicts with established knowledge. Extensive experiments on CIFAR-LT, ImageNet-LT, and iNaturalist 2018 show that VICAL consistently outperforms state-of-the-art methods, validating the effectiveness of our consistency-driven design. Our code is available at \hrefthis https URLVICAL.
[CV-35] MCPO: Modality-Contrastive Preference Optimization for Multimodal Chain-of-Thought Compression
链接: https://arxiv.org/abs/2609.04947
作者: Guangheng Yang,Zhenliang Ni,Zhenkai Wu,Han Shu,Juan Feng,Wenming Yang,Jie Hu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Recently, multimodal large-scale reasoning models have demonstrated remarkable capabilities in solving complex tasks through long Chains-of-Thought (M-CoT). However, excessively long reasoning trajectories incur substantial computational costs and significant KV-cache pressure. Existing CoT compression and alignment paradigms mainly rely on static rules or single-dimensional preferences, lacking fine-grained cross-modal constraints; as a result, they are prone to inducing visual laziness and hallucinatory reasoning. To address these issues, we propose Modality-Contrastive Preference Optimization (MCPO), a highly sample-efficient two-stage length-compression method that requires fewer than 900 training samples. In the compression stage, we introduce a step-level Normalized Cross-Modal Mutual Information (NCMI) pruning algorithm, which automatically identifies and removes visual-independent reasoning steps by comparing the reasoning discrepancies between with-image and no-image contexts. This significantly reduces redundancy and hallucinatory content in the reasoning chains. In the alignment stage, the model first undergoes supervised fine-tuning to achieve domain-adaptive initialization, followed by optimization using an asymmetric multimodal length-controlled preference loss. This objective adopts a highly nonlinear odds-ratio formulation that provides steep gradients in the with-image context to reinforce length constraints for preferred trajectories, while applying a scaled, flat-gradient linear difference in the no-image context to maintain modality consistency, thereby achieving stable cross-modal preference alignment. Extensive experiments on mainstream base models such as Qwen3-VL-Thinking show that our method can reduce CoT length by up to 69.5% and achieve up to 3.34x end-to-end inference speedup while preserving original accuracy.
[CV-36] Learning 3D Editing without Paired Supervision via Generative Prior Distillation
链接: https://arxiv.org/abs/2609.04942
作者: Hao Wen,Weibin Yun,Hongxing Fan,Haotian Lu,Rui Chen,Zehuan Huang,Lu Sheng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 14 figures
Abstract:Instruction-guided 3D editing is essential for interactive content creation, yet it faces a significant bottleneck: the severe scarcity of high-quality paired training data. Existing approaches attempt to bypass this by either relying on slow test-time optimization or training on pseudo-pairs constructed via complex pipelines, which often introduce structural drift and geometric artifacts. In this paper, we propose a novel framework that learns feed-forward 3D editing without paired 3D supervision via Generative Prior Distillation. Instead of relying on ground-truth 3D pairs, our core idea is to distill visual, semantic, and geometric knowledge from powerful foundation models directly into a 3D editing model. Specifically, through a differentiable rendering pipeline, we supervise the 3D representation using two complementary signals: a 2D visual prior from an image editing model at the main editing view, and a semantic prior from a Vision-Language Model at novel views to ensure strict instruction following and source identity preservation. Crucially, to address the geometric collapse and multi-view inconsistencies inherent in 2D projection supervision, we introduce a 3D-aware Distribution Matching regularization. Acting as a geometric prior, this term operates in the 3D latent space, constraining the edited output to remain within the manifold of realistic 3D assets defined by a pretrained image to 3D teacher model. Extensive experiments demonstrate that our method achieves superior instruction fidelity and cross-view consistency, significantly outperforming state-of-the-art baselines. Our project is available at: this https URL.
[CV-37] LensStyle: Learning the Optical Aesthetics for Controllable Stylized Lens Effect Rendering
链接: https://arxiv.org/abs/2609.04939
作者: Yachuan Huang,Liwen Xiao,Liao Shen,Qiwen Wang,Huiqiang Sun,Zhiyu Pan,Zhiguo Cao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:The visual aesthetics of photographs are deeply influenced by lens characteristics such as aperture shape, optical vignetting and optical diffraction, which together define a camera’s unique optical style. Existing lens effect rendering methods primarily focus on accurately simulating the blur transition from small to large apertures but overlook the stylistic aspects of lens effects. As a result, they fail to produce diverse bokeh effects under large apertures or capture distinctive photographic phenomena such as starbursts that emerge under small apertures. In this work, we introduce LensStyle, a unified framework for controllable stylized lens effect rendering that explicitly models lens aesthetics through joint continuous-discrete control. Our model incorporates a Dual-Path Controller that disentangles continuous optical parameter modulation (e.g., focus distance and blur strength) from discrete lens-style conditioning (e.g., circular, polygonal, donut, cat-eye, and starburst effects), enabling fine-grained, interpretable, and physically grounded lens manipulation within a single unified framework. To support model training, we curate a comprehensive MultiLens dataset containing multi-lens image pairs synthesized under real optical constraints. Extensive experiments demonstrate that LensStyle achieves superior realism, controllability, and aesthetic quality compared with existing lens effect rendering approaches and diffusion-based image editing models, advancing computational photography toward multiple-lens-style simulation.
[CV-38] One Diffusion Model Two Roles: Guided Trajectory Planning and Safety-Critical Scenario Generation in Closed-Loop Simulation ECCV2026
链接: https://arxiv.org/abs/2609.04921
作者: Arka Pal,Rajesh Kumar,Hannes Eriksson,Rémi Lacombe,Arvid Laveno Ling,Ankit Gupta,Maciej Wozniak
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO)
备注: Accepted at ECCV 2026 workshop. Arka and Rajesh have equal contribution
Abstract:Diffusion probabilistic models can capture the multi-modal, interaction-rich distribution of joint future trajectories in driving scenes. We show that a single pretrained diffusion traffic model can serve two complementary roles in the autonomous driving development loop: as an ego motion planner, and as a controllable generator of safety-critical scenarios for stress-testing the planners. On the planning side, we introduce a Single-Stream Dual-Stream (SSDS) diffusion-transformer decoder that fuses scene context via joint attention rather than late cross-attention, improving closed-loop performance on nuPlan. We further propose Decoupled Annealing Posterior Sampling with Energy (DAPSE), a training-free guidance scheme that injects arbitrary energy functions at the clean-sample level, avoiding the first-order approximation errors while requiring no auxiliary networks. Beyond planning, we leverage the same diffusion model as a controllable scenario generator to create realistic long-tail driving interactions for closed-loop evaluation. Through inference-time guidance, selected agents are steered toward safety-critical behaviors, including aggressive cut-ins, lead-vehicle braking, and combined longitudinal-lateral interactions, while preserving realistic traffic behaviors. Evaluated in closed-loop nuPlan simulations with independent black-box planners, the generated scenarios expose failure modes that remain hidden under standard benchmarks. Although the SSDS-based planner achieves stronger nominal performance, it experiences larger degradation under these challenging scenarios, demonstrating that benchmark superiority does not necessarily translate to robustness. These results demonstrate that a single learned traffic prior can simultaneously improve motion planning and provide a realistic framework for systematic planner robustness evaluation.
[CV-39] ourPhysics: Bringing Physics to World Models for Exploration and Manipulation from a Single Image
链接: https://arxiv.org/abs/2609.04911
作者: Xin Zhang,Yabo Chen,Zixuan Duan,Haibin Huang,Chi Zhang,Feng Xu,Xuelong Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Interactive visual world models must distinguish observation from physical intervention. Camera motion reveals new surfaces, whereas intervention changes object motion, contact, and deformation. Current video world models are largely driven by appearance priors and often lose physical or spatial consistency over long horizons. We present TourPhysics, an online framework initialized from a single image and a declarative physical configuration. TourPhysics extends PhysOmni, our ACM Multimedia 2026 work, from finite physics-grounded video synthesis to persistent exploration and manipulation. TourPhysics combines deterministic simulation with video generation while assigning separate roles to simulator state, geometric evidence, generator controls, and appearance memory. For each action, the simulator computes a finite physical and camera trajectory before the corresponding observation is generated. Accepted observations publish the terminal state and update the appearance memory and subsequent generator controls, while the committed state and simulator geometry remain fixed throughout synthesis and retry. We further separate the simulator geometry used for projection and visibility from the relative depth used to condition the generator. A reference-anchored memory retrieves accepted static appearance through geometric cross-view correspondence and incorporates it through a bounded residual that reverts to the native path when no valid correspondence exists. On simulator-defined camera tours and object manipulations, TourPhysics follows prescribed camera and object trajectories more closely than the evaluated baselines, preserves the input scene, and reduces appearance drift during long-horizon revisits.
[CV-40] Methane Detection On Board Satellites from Unorthorectified Imagery
链接: https://arxiv.org/abs/2609.04906
作者: Luca Marini,Maggie Chen,Hala Lamdouar,Laura Martínez-Ferrer,Dr C. P. Bridges,Giacomo Acciarini
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:As a potent greenhouse gas, methane is a major driver of climate change. Its effective mitigation relies on timely detection. Conventional detection methods rely on orthorectification to correct geometric distortions and matched filters to enhance plume signals, which are steps designed for ground processing and poorly suited to onboard execution. We introduce UnorthoDOS, a dataset and approach for training machine learning models directly on unorthorectified hyperspectral imagery, bypassing both orthorectification and matched-filter products. Our U-Net models trained on unorthorectified data approach the performance of models trained on orthorectified data (IoU 16.91% vs. 18.47% on all plumes), while both substantially outperform the mag1c matched-filter baseline (IoU 4.76%). We further demonstrate the feasibility of onboard deployment: FP16 compression halves model size with under 0.3% output deviation. The trained ML models and two ML-ready datasets – orthorectified and unorthorectified hyperspectral imagery from the EMIT sensor – are publicly available at this https URL, with code at this https URL.
[CV-41] InterSing: Explicit Interaction Dynamics for 3D Duet Singing Animation and Beyond
链接: https://arxiv.org/abs/2609.04903
作者: Yihan Zhou,Zikai Huang,Yuyang Yu,Xuemiao Xu,Cheng Xu,Shengfeng He
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We present InterSing, a framework for generating realistic 3D head animations for duet singing performances. Unlike solo singing, duet performance requires each singer to balance individual expressiveness with intermittent interaction at musically salient moments, such as phrase boundaries, synchronized rhythms, and call-and-response passages. Because these interactions are sparse and rhythm-dependent, existing audio-driven animation methods and conversational interaction models do not adequately capture their structure. Our key insight is that duet coordination can be represented as a time-varying signal that reflects how strongly performers engage with one another throughout a song. Based on this observation, we introduce interaction logits, an interpretable latent representation that models the degree of cross-performer engagement at each time step. We learn these logits using weak supervision and use them to condition an interaction-aware diffusion model jointly driven by audio features and interaction dynamics. This formulation enables unified multi-mode generation, spanning independent motion, coordinated behavior, and smooth transitions between them. Experiments show that InterSing generates realistic and expressive singing head animations with stronger coordination and musical alignment than existing methods, while preserving each performer’s characteristic motion style. We further demonstrate that the same formulation generalizes to multi-singer performances and provides intuitive control over when and how performers engage.
[CV-42] Sound-based Multi-Person 3D Pose Estimation ECCV2026
链接: https://arxiv.org/abs/2609.04902
作者: Yusuke Oumi,Yuto Shibata,Go Irie,Akisato Kimura,Yoshimitsu Aoki,Mariko Isogawa
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO); Sound (cs.SD)
备注: Accepted at ECCV 2026, Project Page: this https URL
Abstract:Can we recover the 3D poses of multiple people using only sound? This paper presents the first attempt to estimate multi-person 3D poses solely from acoustic signals. Estimating the poses of multiple individuals using acoustic signals is inherently challenging due to the superposition of motion-dependent signal variations. Unlike single-person scenarios, the presence of multiple subjects leads to overlapping acoustic signatures, making it difficult to attribute specific signal changes to an individual’s pose. Furthermore, the complexity is compounded by inter-person reflections, which introduce intricate propagation delays that obscure the temporal motion-acoustic relationship. To address these issues, we propose SoundMHPE (Sound-based Multi-person Human Pose Estimator), a novel encoder-decoder framework consisting of two key components. First, the Acoustic Multi-scale Encoder captures diverse temporal and fine-grained frequency features to isolate subtle acoustic signatures from complex, overlapping signals. Second, the Temporal Pose Decoder employs an attention mechanism to disentangle multi-person information across successive frames. By jointly accounting for temporal dynamics and inter-person dependencies, this component precisely reconstructs frame-wise individual poses. To validate our approach, we constructed the 6-hour Acoustic Multi-person Pose (AMP) dataset consisting of 432K synchronized frames of multi-person pose and acoustic data, and demonstrated that our SoundMHPE outperforms baseline models. Project page: this https URL
[CV-43] SimFuse3D: Source-Guided Target Simulation and Confidence-Guided Multi-Stage Localization Reweighting for Cross-Platform 3D Object Detection
链接: https://arxiv.org/abs/2609.04886
作者: Yongchun Lin,Xinliang Zhang,Yun Zou,Zhixuan Xiao,Liang Lei,Jianya Guo,Yuqiang Zhai,Xiaofeng Wang,HaiKuo Xu,Haoang Li
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 9 pages, 5 figures. Submitted to IEEE Robotics and Automation Letters
Abstract:Changes in sensor height and viewpoint alter object-level point distributions, making cross-platform LiDAR unsupervised domain adaptation (UDA) difficult. Self-training uses labeled source scans and unlabeled target scans, yet a retained prediction may provide a useful target location while enclosing sparse foreground returns, background clutter, or points inconsistent with the predicted box. We refer to this mismatch as box-point inconsistency. We introduce SimFuse3D, which preserves the target placement and repairs the associated pseudo-object using measured geometry from labeled source scans. Object Memory retrieves a compatible labeled source instance. Target Simulation places its ground-truth box at the target location, aligns its points with the target viewing geometry, and filters the aligned crop to approximate the target observation. Confidence-Guided Multi-Stage Localization Reweighting (CMLR) maps each target pseudo-object confidence score to a bounded weight shared by RPN localization and R-CNN box regression. All components operate only during adaptation, leaving the detector architecture and inference graph unchanged. Across six cross-platform transfers, SimFuse3D exceeds Pi3DET-Net on every reported AP metric and ranks first among the compared adaptation methods on nearly all metrics. On nuScenes-to-KITTI, it ranks first among the compared adaptation methods with both evaluated detectors.
[CV-44] Mitigating Performance Discrepancy in Cross-Domain 3D Class-Incremental Learning
链接: https://arxiv.org/abs/2609.04860
作者: Jinge Ma,Gautham Vinod,Bruce Coburn,Jui-Feng Chi,Siddeshwar Raghavan,Fengqing Zhu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 29 pages
Abstract:3D perception plays a crucial role in real-world applications such as autonomous driving, robotics, and AR/VR. In practical scenarios, 3D perception models need to continually adapt to newly emerging 3D object categories, making class-incremental learning (CIL) particularly important. However, unlike 2D images, 3D point clouds are inherently heterogeneous: objects from the same class may not only come from the clean CAD domain, but also from RGB-D camera scans of varying quality, video reconstructions, or even corrupted observations. We discover that such heterogeneity introduces a new challenge beyond catastrophic forgetting: the degree of performance degradation can vary substantially across domains, a phenomenon we term performance discrepancy. To investigate this problem, we establish the Domain3D-CIL training and evaluation protocol, which contains point cloud categories from heterogeneous domains. We further adapt a wide range of mainstream CIL methods to the 3D modality. The results demonstrate that this performance discrepancy consistently appears across these baselines. To mitigate this issue, we introduce PolyMem, an exemplar-free approach that implicitly models rich high-order statistics of the feature distribution to enhance cross-domain robustness. Experiments demonstrate that our method effectively alleviates the performance discrepancy while improving the model’s performance across domains. Code will be made publicly available upon acceptance.
[CV-45] LetOccVote: Learning Weakly Supervised 3D Occupancy through Consensus
链接: https://arxiv.org/abs/2609.04846
作者: Chi Zhang,Qi Song,Feifei Li,Jie Li,Rui Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Weakly supervised 3D occupancy prediction reduces the reliance on costly 3D annotations by learning from 2D pseudo-labels generated by vision foundation models. However, existing methods typically use these imperfect pseudo-labels directly as supervision, making occupancy learning vulnerable to erroneous geometric and semantic targets. We observe that agreement across repeated observations provides an inexpensive and reliable cue for assessing pseudo-label reliability. Based on this observation, we propose \textbfLetOccVote, a weakly supervised Gaussian-based occupancy framework that leverages cross-frame voting to improve both geometric and semantic supervision. For geometry, Depth Vote exploits cross-frame geometric agreement to refine supported pseudo depth and reject contradictory estimates before volumetric lifting and depth supervision. For semantics, Semantic Vote aggregates pseudo-semantic observations in a shared 3D space to identify reliable and contested evidence, strengthening reliable semantic supervision while filtering unreliable pseudo-label segments. The entire framework is trained solely with 2D pseudo-label supervision without requiring 3D occupancy annotations. On Occ3D-nuScenes, LetOccVote achieves 53.27 IoU and 20.39 mIoU, establishing state-of-the-art performance among methods with 2D pseudo-label supervision.
[CV-46] PAPT: Risk-Aware Adversarial Tuning and Generation for Single Domain Generalization
链接: https://arxiv.org/abs/2609.04837
作者: Zhipeng Xu,De Cheng,Xinyang Jiang,Lingfeng He,Huaijie Wang,Dongsheng Li,Nannan Wang,Xinbo Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 29 pages
Abstract:Single domain generalization (SDG) aims to learn a model from one labeled source domain that generalizes to unseen target domains. A common strategy is to enrich the source distribution with augmented or generated samples, and recent text-to-image (T2I) diffusion models provide a strong generative prior for this purpose. However, diversity alone is insufficient for robust generalization, because useful generated samples should also capture variations that the current classifier finds difficult. Motivated by distributionally robust optimization (DRO), we define a semantic ambiguity set in the class-conditional generative space of a pretrained T2I model and search it for samples with high classification loss under the current classifier. To this end, we introduce PAPT++, a risk-aware adversarial generation-training framework for SDG. PAPT++ first learns diverse semantic reference images for each class through image-text alignment and intra-class diversity regularization. These references then serve as denoising targets during classifier-guided diffusion synthesis, reducing semantic drift while guiding generation toward challenging variations. The generated samples are combined with the source data to update the classifier, and the updated classifier guides the next synthesis round in return. In this way, PAPT++ progressively exposes the classifier to challenging yet semantically consistent variations. Extensive experiments on standard SDG benchmarks demonstrate the superiority of the proposed PAPT++ method and the effectiveness of its main components.
[CV-47] Weather-Conditioned Depth Anything
链接: https://arxiv.org/abs/2609.04827
作者: Zhaoming Xu,Chan-Wei Hu,Kuan-Ru Huang,Zihao Zhu,Renjie Li,Yang Zhou,Zhengzhong Tu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Monocular depth estimation foundation models, such as the Depth Anything series, have achieved remarkable performance across diverse domains. However, they still suffer from critical failures under adverse weather conditions, such as fog, rain, snow, or at night. To address this, we present Weather-Conditioned Depth Anything (DA-W), a framework that explicitly disentangles style from content for weather-robust depth estimation. Specifically, we introduce a Style Filter trained on a curated mix of real and synthetic degradation datasets to extract content-independent, degradation-aware weather embeddings. This style embedding is then injected into the Depth Anything backbone using a parameter-efficient, zero-initialized adapter. Such a lightweight modulation allows a single unified model to robustly adapt to diverse conditions, including fog, rain, snow, and low-light, while avoiding catastrophic forgetting of its core generalization abilities in normal conditions. We train the adapter using a pseudo-label distillation and alignment strategy. Our comprehensive experiments demonstrate that our proposed DA-W achieves state-of-the-art robust depth estimation, improving AbsRel by an average of 3.7% on our curated weather benchmarks, while matching or slightly outperforming performance on standard clean benchmarks. Our project page is available at this https URL.
[CV-48] CoLMIN: LLM -based Multi-Decision Path Negotiation for Cooperative Autonomous Driving
链接: https://arxiv.org/abs/2609.04807
作者: Zhe Huang,Zhaoxin Fan,Shuo Wang,Wenjun Wu,Xuan Zhao,Min Liu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multi-vehicle cooperative autonomous driving enhances the safety and reliability of autonomous driving systems through information sharing among connected vehicles, demonstrating significant potential for improving traffic safety. LLM-based approaches leverage strong reasoning capabilities of LLMs to enable effective inter-vehicle negotiation and improve cooperative driving performance. However, driving decisions in complex traffic scenarios are inherently multi-solution in nature. As a result, existing negotiation-based methods often converge prematurely to suboptimal solutions, hindering consensus formation and limiting the practical deployment of cooperative autonomous driving systems. To address this challenge, we propose CoLMIN, the LLM-based multi-decision path negotiation framework for cooperative autonomous driving, achieving stable decision consensus through multi-decision path negotiation and reflective reasoning. To achieve stable and high-quality consensus in cooperative autonomous driving, CoLMIN consists of three key components: (i) an LLM-based Multi-Intent Negotiation module (LMin), which adopts a Negotiator-Evaluator paradigm and generates multiple candidate driving intentions for joint evaluation; (ii) an Evaluation-based Shallow Reflection Module (ESRM), which analyzes negotiation outcomes and provides feedback to guide subsequent negotiations, thereby accelerating consensus formation; and (iii) an LLM-based Deep Reflection Module (LDRM), which performs long-term reflection over negotiation histories to mitigate cognitive fixation and prevent the system from converging to suboptimal solutions. Experimental results in the CARLA simulation environment demonstrate that CoLMIN significantly outperforms existing methods in challenging interactive driving scenarios.
[CV-49] Linguistic Trajectory Encoding for Efficient Long-Horizon Spatial Memory in Embodied Agents
链接: https://arxiv.org/abs/2609.04802
作者: Tianyidan Xie,Shenyi Wang,Qiang Tang,Mingjie Wang,Zhicheng Qiu,Xuanfu Li,Zhan Xu,Jian Yang,Lanjun Wang,Zili Yi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Embodied agents performing long-horizon tasks require a memory representation in which the state transitions of dynamic objects remain queryable in natural language across hours-to-days observation horizons. Existing systems either drop fine-grained motion (clip-level video-language embeddings), keep it only as raw coordinates (geometric SLAM), or organise it around immediate task context (agent working memories). None of them gives the agent a per-object timeline whose state transitions are themselves queryable in language. Our key contribution is \textbfLinguistic Trajectory Encoding (LTE), which compresses dynamic object motion histories via a hybrid representation combining natural language descriptions, sparse spatial anchors, and visual anchors. LTE adapts compression to motion complexity by anchoring periods without reliable observations to the last seen location, while representing motion with geometric waypoints and linguistic descriptions to preserve accuracy. To evaluate these capabilities across extended time horizons, we construct the \textbfSpatial Memory Benchmark (SMB) from EgoLife multi-day recordings, targeting capabilities absent in existing benchmarks: semantic trajectory retrieval and long-horizon object retrieval. On SMB, the LTE-based system achieves 45.3% success in semantic trajectory retrieval and 48.7% in long-horizon object retrieval, outperforming structured-memory and VLM baselines (best prior: 31.9% and 34.4% ). LTE achieves trajectory compression by factors of 8.7\times to 26.1\times with sub-second query latency on 24 ,h video. On Ego4D natural-language queries, the system reaches 28.75% / 55.10% R@1/R@5, +15.80 / +31.30 pts over EgoVLPv2.
[CV-50] Intrinsic Temporal Adaptation of CLIP for Partially Relevant Video Retrieval EMNLP2026
链接: https://arxiv.org/abs/2609.04800
作者: Hyun Seok Seong,Woojin Jun,SuBeen Lee,Jae-Pil Heo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: EMNLP 2026 paper
Abstract:Partially Relevant Video Retrieval (PRVR) aims to retrieve untrimmed videos that contain moments relevant to a text query. Since the target moment occupies only a portion of the video, PRVR requires retrieval based on fine-grained understanding beyond coarse video-level matching. However, existing methods often rely on frozen CLIP frame features, which lack temporal understanding. Even with recent progress in parameter-efficient CLIP adaptation, video-level predictions can still be supported by imprecise frame-level evidence. In this paper, we propose an Intrinsic Temporal Adaptation (ITA) framework for PRVR. First, our Backbone-Internal Temporal Adaptation allows the last few visual transformer layers to attend over groups of neighboring frames. This provides temporally aware frame embeddings while keeping CLIP frozen and training only adaptation parameters. Second, we introduce Affinity-Weighted Gradient Propagation to address the weakly supervised nature of PRVR, softly aggregating top- k frames based on text-frame affinities and propagating learning signals to multiple query-relevant frames. Our method achieves state-of-the-art performance on PRVR benchmarks, demonstrates robust cross-dataset transfer, and retrieves substantially more accurate frame-level evidence within ground-truth query-relevant moments. Our code is available at this http URL.
[CV-51] An Attention-Guided Global and Local Fusion Framework for Lesion-Focused Image Classification
链接: https://arxiv.org/abs/2609.04791
作者: Mst Shafia Tasnima,Md Samaun Elaheea,Tanjim Taharat Aurpab,Md Musfique Anwar
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Lesion-focused image classification presents a core analytical challenge, as discriminative signals are often sparse, spatially dispersed, and easily obscured by background noise, while conventional convolutional neural networks (CNNs) process entire images uniformly and may dilute signal relevance. This study hypothesizes that adaptive fusion of global contextual information and lesion-focused local information can improve classification performance compared with using either representation independently. We propose a three-branch, attention-guided deep learning framework built on Densely Connected Convolutional Network-121 (DenseNet-121) to improve feature attribution, interpretability, and classification reliability. The architecture consists of a global branch that learns representations from full images, followed by Gradient-weighted Class Activation Mapping (Grad-CAM) to generate attention maps that highlight prediction-relevant regions and produce masked inputs, and a local branch enhanced with a Convolutional Block Attention Module (CBAM) to extract refined spatial and channel-wise features from these focused regions. An adaptive fusion branch integrates global and local representations by learning instance-specific weights, allowing dynamic prioritization between contextual and localized information. The framework is evaluated on a synthetic Spot Pattern Dataset (SSPD) and three benchmark datasets, including skin lesion, guava leaf, and grape leaf image datasets, where the fusion branch outperformed the individual global and local branches, reaching 97.75% accuracy on the skin lesion dataset and 99.64% on the guava leaf dataset. The results highlight the value of attention-guided architectures in healthcare analytics by improving model transparency, strengthening feature relevance, and supporting more reliable data-driven decision-making in medical image analysis.
[CV-52] CLON: Cue-Calibrated Linguistic Object Onboarding for Zero-Shot 6D Pose Front-Ends
链接: https://arxiv.org/abs/2609.04784
作者: Seojin Ji,Yoojin Kwon,Hyung-Sin Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 5 figures
Abstract:Zero-shot 6D pose estimation pipelines increasingly rely on strong downstream pose solvers, but their performance is often limited by the front-end: object proposals must preserve partially visible true positives while rejecting semantically plausible distractors. We introduce Cue-Calibrated Linguistic Object Onboarding (CLON), a front-end requiring no task-specific training for new objects. Given rendered templates of the onboarded object set, CLON constructs a linguistic semantic memory for top-down proposal generation and object-set cue weights for calibrated proposal scoring. The linguistic memory guides SAM 3 toward high-recall proposals for onboarded objects, while cue weights are computed once from the onboarded object set before scene inference and kept fixed during online scoring. On seven BOP-Classic-Core datasets, CLON improves detection AP by 8.1 percentage points (pp), segmentation AP by 6.2 pp, and downstream 6D pose AR by up to 4.1 pp over CNOS and SAM-6D front-ends.
[CV-53] CoMLP: Cooperatively-Gated MLPs for Fine-Grained Cross-Modal Information Fusion in Medical Image Segmentation
链接: https://arxiv.org/abs/2609.04781
作者: Mingyuan Meng,Shuchang Ye,Mingjian Li,Zhenyu Zhao,Jinman Kim,Lei Bi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multi-modal medical images and clinical reports provide complementary anatomical, functional, and semantic information for medical image segmentation. Effectively exploiting these heterogeneous sources requires fine-grained cross-modal information fusion that preserves subtle spatial details while capturing semantic dependencies across modalities. Existing fusion approaches frequently rely on cross-attention, whose computational burden increases rapidly with spatial resolution, making dense cross-modal interaction difficult on high-resolution feature maps, particularly for volumetric medical images. In this work, we propose CoMLP, a cooperatively-gated MLP module for fine-grained cross-modal information fusion in medical image segmentation. CoMLP models cross-modal dependencies through cooperative cross-gating, built upon complementary regional and dilated MLP interactions, to capture local and global cross-modal dependencies. We further develop a multi-source fusion architecture in which CoMLP performs both inter-image fusion across imaging modalities and vision-language fusion between visual features and textual reports, enabling heterogeneous information to be integrated without relying on dense cross-attention. Extensive experiments on five medical segmentation benchmarks, covering 2D/3D images, clinical reports, multiple imaging modalities, and diverse anatomical regions, demonstrate consistent improvements over state-of-the-art multi-modal and language-guided segmentation methods. Ablation studies further show that fine-grained interaction at high spatial resolutions and complementary local-global fusion are critical to the performance gains. These results demonstrate the potential of MLP-based interaction as an effective alternative for fine-grained cross-modal information fusion in medical image segmentation.
[CV-54] LUMIN: Lightweight Universal Manufacturing Inspection Network for Anomaly Detection
链接: https://arxiv.org/abs/2609.04775
作者: Pengfei Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 5 figures
Abstract:Industrial anomaly detection faces two engineering bottlenecks: memory bank construction latency and inference efficiency. Traditional sampling algorithms (Farthest Point Sampling, K-Means, etc.) rely on numerous backbone forward passes and iterative distance computations, with construction times ranging from minutes to hours; heavy computation components such as multi-scale feature extraction struggle to meet the millisecond-level real-time requirements of production lines. This paper focuses on sampling efficiency and inference optimization for industrial deployment with two core contributions: (1) PSP (Plugin Sampler Pipeline)—a four-stage adaptive memory bank sampling pipeline based on 18-dimensional pixel metadata and five complementary visual plugins. PSP completes all sampling with zero backbone forward passes; coarse filtering is sub-second numerical sorting, and metadata extraction is a one-time offline cost. PSP supports progressive deployment and incremental updates. (2) Two engineering optimization strategies—parallel memory bank similarity computation (reducing inference memory and latency by over 95%) and stratified pixel sampling for large-scale evaluation (reducing computation time by 20 \times while keeping metrics stable). As a vehicle for validation, we introduce LUMIN (Lightweight Universal Manufacturing Inspection Network) with extreme segmentation-head compression, systematically exploring the accuracy-efficiency frontier against strong baselines. Experiments on five benchmarks demonstrate that PSP matches state-of-the-art sampling accuracy at near-random construction cost (341 \times faster than FPS), while inference optimizations reduce evaluation time by 20 \times with negligible accuracy loss.
[CV-55] SeamFlow: Structure-Aware Flow Matching on Edge Probabilities for Artist-Like UV Unwrapping SIGGRAPH
链接: https://arxiv.org/abs/2609.04751
作者: Yuming Zhao,Zangyueyang Xian,Qijian Zhang,Rendong Liang,Qin Jia,Ying He,Junhui Hou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by Siggraph Asia 2026
Abstract:3D surface cutting and UV unwrapping are fundamental problems in computer graphics. Traditional geometric optimization methods mainly focus on reducing parameterization distortion, but they often overlook visual semantic coherence in seam layouts. Recent autoregressive generative methods improve semantic coherence, yet limited perception of mesh topology often causes inaccurate local cuts. To address these limitations, we introduce SeamFlow, a novel generative framework for 3D surface cutting. We reformulate the discrete mesh-cutting problem as continuous flow matching in a high-dimensional edge-probability space. Through continuous relaxation, SeamFlow learns a deterministic mapping from a Gaussian prior to a target seam-probability distribution. An evolution network couples local topological tokens with global shape priors and guides smooth probability flow through Ordinary Differential Equation solving. Compared with existing autoregressive generative frameworks, SeamFlow improves topology awareness through edge tokenization while eliminating both 3D spatial projection errors and artificial sequential-order bias. Extensive experiments demonstrate that SeamFlow achieves exceptional semantic coherence and remarkably low parameterization distortion. The project page is this https URL.
[CV-56] Where to Look Matters: Learning Influential Views for VLM-based 3D Visual Grounding ECCV2026
链接: https://arxiv.org/abs/2609.04741
作者: Tsung-Chih Chiang,Hsuan-Kung Yang,Jou-Min Liu,Ting-Ru Liu,Chun-Wei Huang,Quan Kong,Chun-Yi Lee
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026
Abstract:Recent zero-shot 3D visual grounding methods leverage vision-language models (VLMs) to localize objects in 3D scenes from natural language queries. However, these methods typically rely on heuristic rules to select which camera views are provided to the VLM, often prioritizing object visibility rather than grounding relevance. We present IVSGround, a framework that learns Influential View Selection for VLM-based 3D visual grounding. Instead of using fixed heuristics, a lightweight view selector is trained to identify views that provide discriminative evidence for grounding. To obtain supervision signals, we generate training signals using feedback from a reasoning VLM through a two-stage rejection sampling process. During inference, the learned selector predicts query-conditioned influential views for each candidate object, which are then evaluated by a frozen reasoning VLM through comparative grounding. Experiments on ScanRefer and NR3D show that IVSGround consistently improves grounding accuracy over existing zero-shot pipelines, demonstrating that selecting where to look is crucial for effective 3D visual grounding. Project page: this https URL
[CV-57] Bridging Modalities and Tasks: A Unified Hierarchical ViT for SAR-to-Optical Translation and Semantic Segmentation
链接: https://arxiv.org/abs/2609.04726
作者: Siyuan Liu,Xuze Zhang,Yongshun Wang,Licong Pan,Hang Liu,Huihui Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Synthetic Aperture Radar (SAR) images have all-weather, day-and-night observation capabilities. However, compared with optical images, their speckle noise and non-intuitive scattering mechanism limit the interpretability of the images. Generative models for SAR-to-optical (S2O) conversion can improve visual interpretability, but existing methods often ignore the constraints on semantic structure, which are necessary for downstream tasks, for the sake of visual effects. We propose a unified collaborative dual-task learning framework, termed BMT (Bridging Modalities and Tasks), that jointly optimizes S2O image translation and semantic segmentation through a shared hierarchical Vision Transformer. The framework integrates: (1) a LocalViTBlock that fuses global self-attention with spatial depthwise convolution through a learnable gating mechanism; (2) an enhanced output module combining multi-scale refinement processing, color correction and anti-aliasing, which calibrates channel-level color statistics through feature fusion; (3) a ControlNet-style conditional injection mechanism that encodes SAR wavelet features and segmentation labels into a multi-scale feature pyramid and injects them at each encoder layer through zero-initialized convolution; (4) a bounded Kendall uncertainty weighting scheme that prevents either task from dominating the shared representation. We evaluate the framework under both paired and unpaired translation settings, on the public WHU-OPT-SAR paired dataset and a self-constructed unpaired ship dataset built from HRSID and DIOR, respectively. The experimental results show that the proposed method achieves competitive S2O translation quality and semantic segmentation performance. The dataset and source code have been publicly released at this https URL.
[CV-58] HiSfM: Disambiguating Structure-from-Motion via Scaffold-Anchored Hierarchical Reconstruction
链接: https://arxiv.org/abs/2609.04718
作者: Ziding Zhao,Hainan Cui,Peilin Tao,Shuhan Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Structure-from-Motion (SfM) is a fundamental tool for sparse 3D reconstruction with broad impact in robotics and vision, supporting mapping, localization, and large-scale scene modeling. However, conventional pipelines often fail under hard visual ambiguity caused by repeated or symmetric structures, and incur heavy computational cost due to redundant cameras and constraints. We present HiSfM, a hierarchical coarse-to-fine SfM framework that improves robustness and efficiency through scaffold construction. HiSfM first forms strong local communities using geometrical induced heuristics, then connects communities with a compact yet strong skeleton by packing edge-disjoint spanning trees (EDST) while verifying skeletal edges with a two-view disambiguator. We reconstruct a stable scaffold on this verified skeleton, serving as an anchor to capture the essence of the scene, and subsequently absorb remaining images via efficient registration and triangulation for further refinements. Experiments on ambiguity-focused benchmarks and general datasets show that HiSfM prevents ambiguity-induced failures while substantially reducing runtime compared to previous methods, and improves completeness over aggressive sparsification methods. Code is available at this https URL.
[CV-59] Counting Beyond Instances: A Benchmark for Group-Individual Object Counting
链接: https://arxiv.org/abs/2609.04716
作者: Rui Wang,Junyi Huang,Jiahui Li,Qiao Yu,Yixue Hao,Long Hu,Baoru Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 9 figures
Abstract:Visual counting is commonly formulated at the instance level, aiming to estimate how many objects of a queried category appear in an image. However, real-world counting often involves higher-level semantic units formed by multiple instances, such as a bunch of grapes, a stack of plates, or a pair of shoes. This exposes a key limitation of existing counting formulations, which mainly focus on what to count, while largely overlooking at which semantic unit to count. We introduce Group-Individual Object Counting (GIC), a new setting that requires models to count both individual objects and semantic groups within a unified framework. To support this new task, we present BunchCount, a real-world benchmark with 1,330 images, 89,254 individual annotations, and 11,065 group annotations. BunchCount provides paired individual-group annotations within the same image and explicitly records containment relations between each group and its constituent individuals. Experiments on BunchCount show that current advanced counting models perform well on individual instances but fail to count semantic groups more accurately. To mitigate semantic granularity conflict, we propose a counting-unit guided relational counting framework, which exploits group-individual containment relations to regularize cross-granularity representations during training. Our method substantially improves group-level counting while better preserving individual-level counting ability, establishing a strong baseline for counting beyond instances.
[CV-60] AngelFingerprint: A Traceable Explainable and White-Box Stealthy Watermark for Text-Guided Image Editing
链接: https://arxiv.org/abs/2609.04709
作者: Bo-Han Kung,Futa Waseda,Ching-Chun Chang,Isao Echizen,Shang-Tse Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages
Abstract:Text-guided diffusion editing raises disinformation concerns, making reliable image provenance essential. While watermarks are commonly used for this purpose, most methods carry a fixed ID that cannot explain what was changed and which prompt produced it. Furthermore, under open-source white-box access, attackers can easily locate and remove watermarks added as separate modules. Targeting this setting, we propose AngelFingerprint, a novel watermarking framework ensuring edit traceability, explainability, and white-box stealthiness. It integrates a LoRA into the diffusion model to embed the editing prompt’s CLIP text embedding directly into the model’s weights. An extractor then recovers this embedding from the image pixels alone. This semantic payload explains the edit, while the weight-integrated design makes it hard to detect and isolate even under full white-box access. Two techniques make this possible: a velocity-alignment anchor that preserves edit quality, and a specially designed frequency filter that keeps the watermark imperceptible yet recoverable and robust. On the MagicBrush dataset, our extractor achieves 86% top-1 accuracy in a 200-way prompt retrieval, versus 20% for prompt inversion.
[CV-61] Sustainable Edge Vision via Empirically Calibrated DVFS: Eliminating Thermal Throttling on Passively Cooled Hardware
链接: https://arxiv.org/abs/2609.04705
作者: Aayush Marasini,Zhaoxian Zhou
类目: Hardware Architecture (cs.AR); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 7 pages, 5 figures, 8 tables, Code, datasets, and frozen artifacts available at: this https URL
Abstract:Passive cooling eliminates the energy overhead and mechanical failure modes of fans, making it attractive for edge deployment, yet sustained Deep Neural Network (DNN) inference on passively cooled edge Systems-on-Chip (SoCs) is bottlenecked by thermal throttling. To address this, we propose an empirically calibrated, state-aware Dynamic Voltage and Frequency Scaling (DVFS) scheduler. Unlike heuristic-driven controllers, our methodology utilizes time-domain guards and absolute temperature bounds, with derivative triggers acting as safeguards against sharp thermal spikes. Evaluated on a passively cooled Raspberry Pi 5 running YOLOv8n, our scheduler eliminates all observed thermal throttling events during sustained 30-minute workloads. It outperforms a temperature-only reactive baseline by achieving a 6.8% higher frame rate (Cohen’s d = 8.73) while consuming 1.9% less energy per frame. Furthermore, our optimized passive scheduling surpasses an actively cooled reference system in energy efficiency (Joules/frame), though active cooling remains superior for raw throughput. Through isolated ablations, we show that the dwell guard is necessary for run-to-run reproducibility. Finally, exploratory boundary probes indicate that the passive operating envelope closes at ambient temperatures ( \ge 27^\circ C) where nonlinear leakage defeats DVFS-based control. These results indicate that, within the mapped envelope, correct scheduling can make mechanical cooling unnecessary for sustained edge inference on this platform.
[CV-62] LookThere! Sparse Vision by Reinforced Selection
链接: https://arxiv.org/abs/2609.04698
作者: Sreehari Rammohan,Yousef Yassin,Anthony Fuller,Junfeng Wen,Carl Vondrick,Evan Shelhamer
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Vision transformers typically treat every image token as equally important, yet for most tasks in computer vision only a fraction are needed. Adaptive computation methods accelerate inference by choosing which tokens to process, but existing methods struggle at extreme sparsity and require heuristics that may not generalize like token diversity and attention scores. We address these limitations with LookThere, achieving a new pareto frontier in performance-compute trade-offs through an end-to-end reinforcement learning framework that jointly trains a shallow input selector and a deep representation extractor. The selector learns where to look and the extractor learns what to see, together saving computation by selecting only what is worth processing for a given task without relying on auxiliary signals. We show that LookThere only selects the task-specific input, excelling at sparse recognition in high-resolution settings (traffic signs, billiards), and maintaining accuracy with as little as 0.2% of the input. It generalizes across tasks and models, including global recognition (ImageNet classification), local recognition (ADE20K segmentation), zero-shot classification (by distillation), and regression (counting). Across all settings, LookThere surpasses state-of-the-art selection to provide a general and scalable framework for specialized and efficient adaptive computation.
[CV-63] Enhancing Multimodal Emotion Recognition via Multi-Feature Encoding and Attention-Based Fusion ICONIP2025
链接: https://arxiv.org/abs/2609.04690
作者: Xu Lin,Ke Wang,Hui Kang,Xinying Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 15 pages, 6 figures, 6 tables. Pre-peer-review version. The final published version appears in ICONIP 2025, Lecture Notes in Computer Science, vol. 16312, pp. 142-157 (2026)
Abstract:Multimodal emotion recognition has attracted growing interest due to its importance in human-computer interaction, remote education, and healthcare. This paper proposes a novel multimodal emotion recognition framework that integrates rich audio and visual feature extraction with an attention-based fusion strategy. For audio, we extract three complementary feature types: semantic embeddings from Wav2Vec2, MFCC features, and statistical acoustic descriptors such as pitch, energy, and rhythm. These are aligned and fused via a BiLSTM to capture temporal dependencies. For video, we propose a ResNet50-BiLSTM architecture that combines deep residual learning and sequential modeling to extract expressive spatiotemporal features from facial sequences. To enhance multimodal synergy, we introduce a feature-level fusion mechanism based on multi-head attention, allowing the model to adaptively weigh contributions across modalities. Experiments conducted on the MELD and IEMOCAP datasets demonstrate that our model significantly outperforms baselines in both accuracy and robustness. Furthermore, ablation studies show that the attention-based fusion strategy significantly improves performance in unbalanced data settings. Our findings suggest that the proposed framework effectively captures diverse emotional cues from speech and visual expressions, and offers a practical and generalizable approach for real-world multimodal emotion recognition tasks.
[CV-64] ReaDiT Guidance: Control for Image and Video Generation using Diffusion Transformer Features
链接: https://arxiv.org/abs/2609.04649
作者: Jay Mahajan,Chang Liu,Rauf Makharov,Viraj Shah,Alexander Schwing,Svetlana Lazebnik
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We present DiT Readout (ReaDiT) Guidance, a lightweight framework for controlling generation with Diffusion Transformer (DiT) models via their internal feature representations. ReaDiT Guidance uses features from a single DiT block to steer the generative process according to spatial targets - like depth, pose, or edge maps - provided at test time. Furthermore, since modern text-to-video models are largely built on DiT backbones, ReaDiT Guidance naturally extends to video generation, enabling camera and motion control. Experimental results demonstrate that our approach achieves competitive or improved results compared to existing feature-based and off-the-shelf adapter-based approaches while requiring fewer parameters.
[CV-65] Importance-Aware Low-Rank Distillation of Diffusion Transformers
链接: https://arxiv.org/abs/2609.04646
作者: Denis Zavadski,Sebastian Heid,Damjan Kalšan,Stefan Roth,Carsten Rother
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Diffusion Transformers (DiTs) have emerged as a dominant architecture for high-quality text-to-image generation, yet their scale poses challenges for efficient deployment. While truncated singular value decomposition (SVD) is a principled tool for parameter reduction, evidence from large language models (LLMs) suggests that naive low-rank approximation can cause catastrophic failure. In contrast, we find that truncated SVD in DiTs produces smooth degradation even under substantial global compression, with redundancy distributed across projection matrices throughout the whole network rather than concentrated in a few transformer blocks. Building on these insights, we introduce SVDtrunc, a two-step block-level compression scheme, first allocating ranks across blocks and compressing the least important ones via truncated SVD under a global parameter budget, and then fine-tuning all blocks with modular knowledge distillation and a rectified-flow objective. We apply SVDtrunc to this http URL across compression levels ranging from 40-90% of the original parameter count. Across three benchmarks, GenEval, HPSv2, and DPG, we outperform all competing approaches. Notably, and in contrast to prior work, we retain near-full performance at 68% and remain competitive even at 57% of the original parameter budget. Furthermore, we show that SVDtrunc complements step distillation and achieves strong results even without fine-tuning, positioning it as a practical continuation of efficiency improvements beyond diffusion step reduction for large-scale generative models. Project page: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.04646 [cs.CV] (or arXiv:2609.04646v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2609.04646 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-66] An Evaluation Framework for Generating Multi-View Images of a Person in a Scene
链接: https://arxiv.org/abs/2609.04603
作者: Mahir Majid,Young Kyung Kim,Guillermo Sapiro
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 3 figures
Abstract:Recent generative image-editing Diffusion Transformers (DiTs) demonstrate impressive semantic editing capabilities but still struggle with spatially consistent camera angle changes. A primary bottleneck in training foundation models to execute free-form, promptable camera angle changes is the lack of specialized training data. While multi-view datasets exist for generic 3D environments and objects, there remains an absence of paired, multi-view datasets featuring human subjects at fixed locations in natural scenes, including frontal and side-profile views. Capturing such multi-camera data in unconstrained environments is logistically challenging and unscalable. In this paper, we first experiment with multiple state-of-the-art image editing models to create this data synthetically, but find that the outputs are frequently prone to hallucinations involving how much the subject’s head turns relative to the background, often producing inconsistent environments. To address this issue, we propose the Head Scene Rotation Difference (HSRD) metric to quantitatively evaluate camera movements around a person. The proposed metric operates by decoupling camera movement from localized head pose manipulation. As demonstrated by the extensive experimentation, HSRD provides the pipeline necessary to evaluate 3D spatial parallax for a person in a scene, paving the way to reliably construct high-quality multi-view synthetic datasets.
[CV-67] Dual-Part Multi-Lateral Branched Network for Multi-Class Segmentation in Cardiovascular Catheterization Angiograms
链接: https://arxiv.org/abs/2609.04590
作者: Olatunji Omisore,Ahmed Elazab,Ali Shahidinejad,Fariza Sabrina
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:Catheterisation image processing requires segmentation models that are fast, accurate and explainable. While most of the existing studies usually focus on binary segmentation, there is a recent demand for simultaneous segmentation of multiple structures found in catheterization scenes. In this study, a dual-part MLBNet architecture is designed with multi-lateral encoder blocks and multi-head decoder branches for class-aware segmentation in cardiovascular catheterization scenes. Lateral branches in the encoder enables repeated feature extraction to learn diverse shared representations, while multiple decoder heads are used to introduce class-skewed branches that specialize in different structural properties in catheterization scenes. To analyze the performances of the dual-part MLBNet architecture, several multi-class segmentation angiogram data obtained during cardiovascular catheterization in phantom models, synthetic human-simulated aorta, and animal model are used for model training and evaluation. Results obtained showed the dual-part models could effectively separate guidewire, catheter, vessels and background pixels to their classes of memberships with high probability. The results demonstrate that all models were able to distinguish the dominant background class from foreground structures with high overall accuracy.
[CV-68] DART: Depth-as-Target Pretraining for Surgical Vision Foundation Models BMVC2026
链接: https://arxiv.org/abs/2609.04555
作者: John J. Han,Adam Schmidt,Muhammad Abdullah Jamal,Jie Ying Wu,Omid Mohareri
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to BMVC 2026
Abstract:Vision foundation models (VFMs) are valuable in data-scarce domains such as surgery, where a single pretrained backbone can provide rich representations for many downstream tasks. Yet the dominant self-supervised pretraining paradigm uses only RGB images, leaving readily available complementary signals, such as depth maps, unused. This is a particular missed opportunity in surgery, where natural-image VFMs transfer poorly while the scene geometry is rich and informative. With strong off-the-shelf models now able to produce pseudo-labeled dense depth for any image corpus, we hypothesize that such signals can be folded into pretraining to learn better representations. We present DART, an RGB-D pretraining recipe that builds on DINOv2 with a simple modification: a pixel-space depth reconstruction objective applied to masked iBOT patches, supervised by pseudo-labeled depth. Depth is used only during pretraining, so fine-tuning and inference remain RGB-only. We find that this pixel-level reconstruction head improves representation quality rather than disrupting it. We further show that depth, which encodes scene geometry, is more effective as a target than alternative dense signals such as Canny edges, confirming that the gains stem from depth rather than added supervision alone. Across eight surgical benchmarks spanning segmentation, depth estimation, and image-level recognition, DART outperforms both natural-image and in-domain baselines, including a vanilla DINOv2 trained on identical data, improving dense prediction while also strengthening image-level understanding. More broadly, DART shows that freely available geometric pseudo-labels can strengthen foundation model pretraining without extra labels or added inference cost, pointing toward stronger backbones for surgery.
[CV-69] VISTA: Dense Multi-Label Classroom Coding with Vision-Language Models CVPR
链接: https://arxiv.org/abs/2609.04550
作者: Andrew Franck,Brendan Ng,Ben Fitzgerald,Zane Derrod,Chris Cianci,Chris Craney
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: DataMFM Workshop @ Computer Vision Pattern Recognition (CVPR) 2026
Abstract:Video-language benchmarks are usually constructed by the dataset authors without published reliability statistics, leaving the noise floor of the construct unknown. We argue that multimodal benchmarking benefits from methods taken from research communities that have already invested in strategies to ensure reliability. We illustrate the case with the Classroom Observation Protocol for Undergraduate STEM (COPUS): a 24-code multi-label observation instrument with a decade of peer-reviewed reliability literature. We recast COPUS as a video benchmark for multimodal foundation models, where it provides a dense set of structured labels (a 24-dimensional binary vector every 2 minutes across a 50-90 minute lecture), an externally validated vocabulary, and established literature that provides a per-code reliability target based on human evaluators. Annotations in our evaluation corpus are produced by a 5-person human-evaluator panel whose consensus matrix is our reference. We propose VISTA, a baseline that runs MiniCPM-V-4.5 over a dense sliding window, refines its per-window outputs with a lightweight multi-layer perceptron (MLP) head trained on top of the frozen backbone, and max-pools the resulting predictions onto the 2-minute COPUS grid. On three held-out chemistry lectures, VISTA reaches 80.1% restricted macro accuracy versus 74.9% for the zero-shot variant, with the largest residual errors on visually similar instructor codes and on rare audio-dependent codes. We characterize three systematic failure modes (audio-partial observability, fine-grained group-work discrimination, long-tail recall) and release the benchmark tooling, prompts and baseline code at this https URL.
[CV-70] SocioGesture: Real-Time and Adaptive Social Gesture Perception for Human-Robot Interaction
链接: https://arxiv.org/abs/2609.04545
作者: Wenjin Fu,Li-Fan Wu,Jerin Peter,Chip Huyen,Boyuan Chen,Jan Liphardt
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 3 figures. Project page: this https URL
Abstract:Robots interacting with people must recognize not only explicit commands, but also social cues such as invitations, refusals, and unavailability. In real deployments, these cues must be inferred from noisy onboard perception under partial occlusion, changing viewpoints, and strict latency constraints. We present SocioGesture, a real-time adaptive social gesture perception system for human-robot interaction (HRI). SocioGesture uses a compact confidence-aware body-hand skeleton representation and a lightweight dual-stream model that fuses body motion with hand articulation for low-latency onboard recognition. To improve deployment robustness, we train the model with occlusion-aware skeleton corruption, exposing it to missing hands, occluded arms, and temporally unstable keypoints without increasing the inference cost. On a social gesture dataset collected in mixed indoor-outdoor HRI scenarios, SocioGesture achieves strong held-out-subject recognition, substantially improves robustness under structured joint occlusion, and runs in real time on a robot-mounted edge device. During deployment, uncertain interaction segments are saved for offline labeling and adaptation, enabling SocioGesture to expand its gesture vocabulary while preserving performance in the original classes. These results demonstrate a practical path toward robust, efficient, and adaptive social perception for interactive robots.
[CV-71] STyMo: Fast and Controllable Few-Shot Motion Style Transfer
链接: https://arxiv.org/abs/2609.04500
作者: Jose Luis Ponton,Alexander Winkler,Ladislav Kavan,Yuting Ye,Petr Kadlecek
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV)
备注: Project webpage: this https URL
Abstract:Supporting a wide variety of motion styles is critical for creating diverse virtual characters, but current methods either require large stylized datasets or pre-trained models that cannot generalize beyond their training distribution. We present STyMo, a few-shot approach that learns motion style from only seconds of paired data and trains in one to two minutes. Our key insight is to decompose style into two components: a static component capturing time-invariant posture, and a temporal component capturing frame-wise dynamics. This decomposition yields an interpretable system where posture intensity, temporal exaggeration, and per-body-region style can be adjusted at runtime. Furthermore, the reduction in required training data and computation time structurally permits an iterative authoring workflow. To ensure robustness on arbitrary inputs, we further introduce a stylizability gate that automatically prevents artifacts on out-of-distribution motions. We demonstrate results across diverse motion styles, from subtle emotional variations to exaggerated character archetypes, and release our processed paired dataset to facilitate future research.
[CV-72] opology-Aware Training and Spatial Diagnostics for Fiber Bundle Segmentation in Tracer Histology
链接: https://arxiv.org/abs/2609.04454
作者: Joselyn Romero Avila,Kyriaki-Margarita Bintsi,Ermias Habte,Julia F. Lehman,Suzanne N. Haber,Anastasia Yendiki
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Anatomic tracer studies reveal how axon bundles project from an injection site, branch into smaller groups of axons, and course through the brain to reach their destinations. Histological data from such studies provide anatomical reference information for validating diffusion MRI tractography. However, manual annotation of the histological data is very labor-intensive, and although automated segmentation methods have been proposed, they rely mainly on pixel-overlap losses such as BCE and Dice; topology-aware loss functions have not been studied for this task. We compare BCE-Dice, clDice, Betti matching, and Topograph for fiber bundle segmentation in macaque tracer histology using a frozen DINOv3 backbone. To our knowledge, this is the first exploration of foundation-model features for this task. BCE-Dice achieved the highest Dice, while clDice achieved the highest bundle recall but poor mask overlap. Topograph had similar Dice to BCE-Dice, the lowest \beta_0 error, and fewer false positives than BCE-Dice and Betti matching. Fiber bundle segmentation methods are typically evaluated with a permissive rule that counts a bundle as detected given any overlap with the prediction. We show this rule does not capture oversegmentation, and that per-section TPR can be inflated by empty sections assigned perfect recall. To quantify this, we introduce Excess32, a spatial diagnostic measuring predicted pixels outside a 32-pixel tolerance band around annotated bundles. In validation, a Betti-Topograph union raises sparse-bundle TPR from 0.818 to 0.933, but worsens FDR from 0.296 to 0.509, Excess32 from 0.108 to 0.466, and area ratio from 0.94 to 3.34. These results show detection metrics alone are insufficient to characterize segmentation quality.
[CV-73] Segmentation of the aorta in 4D flow MRI using 4D convolutional kernels and learning from sparse annotations
链接: https://arxiv.org/abs/2609.04439
作者: Hinrich Rahlfs,Julio Garcia,Chiara Manini,Markus Hüllebrand,Sebastian Schmitter,Sarah Nordmeyer,Titus Kühne,Heiko Stern,Christian Meierhofer,Andreas Harloff,Sebastian Kelle,Alexander Lenz,Peter Bannas,Jeanette Schulz-Menger,Ralf F Trauzeddel,Anja Hennemuth
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to Journal of Cardiovascular Magnetic Resonance
Abstract:Automated aortic segmentation in 4D flow MRI is essential for reproducible hemodynamic assessment but is limited by scarce dense annotations and high computational demands. We developed a fully automated 4D (3D+time) U-Net for segmenting the ascending aorta, arch, and proximal descending aorta, using a parameter-efficient hybrid 4D kernel to capture temporal context and sparse 4D labels derived from existing 2D expert contours and centerlines, thereby avoiding the need for dense 4D annotations. Training comprised 268 scans from 8 centers and 2 vendors, with evaluation on an internal test set (32 scans) and an external post-contrast set (30 scans; different site, protocol, and annotator), compared against frame-wise 3D networks and two semi-automatic references. Against time-resolved annotations, the 4D U-Net achieved Dice scores of 0.927 (internal) and 0.911 (external), versus 0.919/0.847 for the 3D U-Net, 0.893 for static PC-MRA, and 0.808 for registration-based propagation; differences were small in systole but pronounced in diastole. Agreement with expert contours for peak velocity, net flow, axial and circumferential wall shear stress, and diameters was excellent (ICC =0.954 internal, =0.980 external), while semi-automatic references performed worse. The method thus provides reproducible, time-resolved aortic segmentation for automated hemodynamic analysis and generalizes across multicenter, multivendor, and independent post-contrast data. The model is publicly available.
[CV-74] ICM-Bench: Person-Level Identity Reasoning in Multimodal Agents with Long-Term Memory
链接: https://arxiv.org/abs/2609.04438
作者: Shidu Ren,Yunze Liu,Xing Liu,Chi-Hao Wu,Enmin Zhou,Junxiao Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages, 6 figures, and 7 tables. Code: this https URL
Abstract:Long-horizon multimodal agents should remember not only what happened but also who participated. This capability depends on linking recurring faces, voices, names, person-associated objects, events, and social relations to consistent identities over time. Existing long-video and multimodal-agent benchmarks measure broad memory question answering, but they do not isolate the ability to maintain recurring person identities and reason over their cross-time relations. We introduce ICM-Bench (Identity-Centric Memory Benchmark), which, to the best of our knowledge, is the first benchmark specifically designed to evaluate identity-centric reasoning over long video memories in multimodal agents. The benchmark contains 839 synthetic clips spanning 141 minutes and 1,217 open-ended questions about six recurring adults in a one-year life album. A theme-configurable pipeline generates the video collection and associates each question with its target identities and traceable supporting evidence. We compare direct caption-memory baselines, memory-augmented agents, and graph-retrieval systems. Gemini 3.1 Pro achieves the highest overall accuracy of 74.0%, yet its score falls to 60.3% on questions that require long-term identity profiles. The results show that current systems recover many event-level memories but remain less reliable when evidence must be accumulated around a stable person.
[CV-75] STEMPix: A Phase-Transition-Material-Based Pixel Sensor for Resolving Edge-Movement Direction
链接: https://arxiv.org/abs/2609.04435
作者: Md Rahatul Islam Udoy,Sumeet Kumar Gupta,Deep Jariwala,Ahmedullah Aziz
类目: Emerging Technologies (cs.ET); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:This paper proposes a spatio-temporal edge-movement direction pixel (STEMPix) for generating compact direction-aware edge movement information inside a CMOS-compatible image sensor array. The proposed design targets specialized sensing applications where local boundary movement is more important than full-frame intensity reconstruction. Instead of transferring full multi-bit frames for external processing, STEMPix generates a 3-bit local edge direction code (LEDC) by combining pixel-level temporal change information with neighboring-pixel spatial edge information. We design the architecture using a two-tier organization, where the photodiode layer is separated from the computation layer to preserve light-collection area while accommodating the additional in-array processing circuitry. The proposed circuit is evaluated through HSPICE transient simulations. The estimated implementation achieves a horizontal pitch of 1.73 \mum, a vertical pitch of 2.36 \mum, and a geometric fill factor of 95.47%. The average active switching energy is 0.465 fJ per LEDC operation across representative edge-movement cases. The proposed STEMPix operation also supports global-shutter capture and dynamic thresholding. These results indicate that STEMPix can provide a compact and scalable front-end representation for edge-movement-aware sensing systems.
[CV-76] AquaBEV: Monocular Underwater BEV Occupancy with 3D Sonar Supervision
链接: https://arxiv.org/abs/2609.04411
作者: Trung Tien Dong,Shengji Jin,Chen Chen,Yi Sheng,Xiaomin Lin
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Autonomous underwater robots are widely used for exploration, monitoring, and inspection, where safe navigation depends on understanding the surrounding free and occupied space. Bird’s eye view (BEV) occupancy provides such a representation, but predicting it from a single underwater RGB image is difficult due to limited, unreliable geometric cues from appearance alone. 3D imaging sonar offers complementary geometric measurements to supervise this task. We introduce AquaBEV, a monocular underwater occupancy model that predicts local BEV occupancy from a single RGB image, using paired 3D imaging sonar as geometric supervision during training. AquaBEV maps visual features into a calibration free polar representation and applies causal decoding along the range dimension before reconstructing the prediction in Cartesian BEV coordinates. A controlled underwater occupancy benchmark was established, adapting representative occupancy methods to the same RGB to sonar task under a unified protocol. AquaBEV achieves 31.4 Visible IoU and 38.6 Observed IoU, 4.0% and 4.3% relative improvements over the strongest transferred baseline. Subjects: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.04411 [cs.RO] (or arXiv:2609.04411v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2609.04411 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-77] FAVE: Foveated Adaptive Visual Encoding for Efficient Fine-Grained Visual Understanding
链接: https://arxiv.org/abs/2609.04392
作者: Amitangshu Mukherjee,Kaushik Roy
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 21 pages, 6 figures, 7 tables; includes supplementary material
Abstract:Fine-grained visual understanding depends on local detail, yet visual encoders face a trade-off between costly full-image high-resolution processing and compact global encoding that can weaken such evidence. Inspired by human active vision, we separate where to look from what to encode. We focus on the latter and introduce FAVE (Foveated Adaptive Visual Encoding), a lightweight variable-resolution ViT that encodes externally selected regions at high acuity while preserving native geometry. We first isolate this encoding problem using oracle ground-truth crops in a controlled small-object regime. On ImageNet objects with a native maximum side of 96 pixels, FAVE improves Top-1 by 9.4 points over a fixed-resolution ViT on the same crop window with 12.7 times lower FLOPs. Increasing global resolution or backbone capacity does not recover the same operating point. We then integrate FAVE as a complementary local branch in FastVLM. Its local tokens are combined with FastVLM’s global visual tokens, while the original global pathway and language model remain frozen. With at most 16 additional local tokens, FAVE improves TextVQA by 1.60 points and achieves a 3.3 times controlled TTFT speedup over SmolVLM2-2.2B. On GQA attribute questions, it improves FastVLM-1.5B by 1.31 points, extending the benefit beyond text while narrowing the gap to FastVLM-7B. Together, these results show that selectively allocating high-acuity local capacity provides an efficient complement to broader global representations and model scaling for fine-grained understanding of small objects, text, and attributes.
[CV-78] What Moves? Localized Motion Representations for Compositional Scene Control
链接: https://arxiv.org/abs/2609.04383
作者: Frank Fundel,Malek Ben Alaya,Thomas Ressler-Antal,Stefan Andreas Baumann,Björn Ommer
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Real-world dynamics are inherently compositional: multiple entities move simultaneously within a shared scene, each exhibiting distinct motion patterns. Yet most existing video representations encode motion globally, without explicitly capturing localized motion for individual entities. Crucially, motion is defined relative to a global reference frame, including camera motion and scene layout. However, localized embeddings are often computed from cropped images or obtained by masking features after encoding, discarding the context needed to interpret motion. To address this, we introduce a promptable localized motion representation that produces persistent embeddings for user-specified regions defined by spatial masks. Rather than cropping the input or masking features, our model processes the full video and conditions motion encoding directly on the queried region. This yields temporally consistent, region-addressable embeddings that isolate local dynamics while retaining the global context required for disambiguation. We demonstrate object-level motion transfer, enabling controlled composition of dynamic scenes. Beyond generative control, our embeddings support localized action classification in multi-actor videos. Across both tasks, our approach improves controllability and outperforms global representations localized through cropping or post-hoc masking. Project Page: this https URL
[CV-79] Where Appearance Fails Geometry Recognizes: A CAD-Free 3D Shape Prior That Complements Vision Foundation Models
链接: https://arxiv.org/abs/2609.04381
作者: Chenxi Tao,Seung-Kyum Choi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: 19 pages, 11 figures, 5 tables
Abstract:Recognizing specific objects onboarded without a labeled training set recurs across manufacturing and service robotics, yet the conventional renderable prior, a computer-aided-design (CAD) model, is often unavailable. Two-dimensional capture supplies no shape prior, and frozen foundation features fail on geometrically similar, low-texture industrial parts. We ask what a short object-centric scan buys for recognition beyond the captured images themselves: each object is reconstructed with 3D Gaussian Splatting (3DGS), summarized into a per-class shape prototype, and fused with frozen DINOv2 image features. First, the scan recovers the recognition value of CAD without CAD: geometry from RGB-D depth (on T-LESS), 3DGS, and CAD gives comparable recognition (tied on HOPE, within 1.6 points on T-LESS); 3DGS is only a convenient route to a point cloud. Second, the payoff is governed by how recognizable the shape is: on shape-distinctive household objects (HOPE) geometry alone reaches 0.920 versus image-only 0.832, a ceiling below which fixed-weight fusion (0.872) sits. On shape-confusable textureless industrial parts (T-LESS) the gain is modest but consistent (0.560 to 0.591 fused, above both single signals). Third, the prior is complementary, not uniformly additive: it rescues far more image failures than it breaks successes, and its benefit grows under partial occlusion. Finally, the worth lies in geometry, not rendered pixels: 3DGS renderings do not help the image side, and frozen-feature recognition is nearly lighting-invariant (within 2.5 points). The study is scoped to recognition, not the BOP pose benchmark.
[CV-80] AdaptVPR: Route-Aware Hard Positive Generation for Robust Visual Place Recognition
链接: https://arxiv.org/abs/2609.04369
作者: Shunpeng Chen,Jingyi Zhang,Changwei Wang,Shengpeng Xu,Yukun Song,Xingtian Pei,Jinzhou Lin,Li Guo,Shibiao Xu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 9 figures, 9 tables
Abstract:Visual Place Recognition (VPR) localizes a query image by retrieving database images of the same or nearby place, yet its robustness is often degraded by domain shifts arising from illumination, weather, seasonal changes, and dynamic occlusions. One contributing factor is the limited appearance diversity of the same place in existing training data. To address this issue, we propose AdaptVPR, a route-aware generative augmentation framework that constructs same-place hard positives for robust VPR training. AdaptVPR first uses a vision language model to parse scene attributes and estimate editing feasibility, while a rule-based scheduler determines the generation route according to editability scores and risk constraints. The generation process is decomposed into three complementary routes: the Global Appearance Route introduces global scene changes in weather, illumination, and time of day; the Local Occlusion Route inserts plausible dynamic occluders; and the Dual Route combines both types of perturbations to produce more challenging appearance shifts. Each generated candidate is evaluated using a VPR-oriented verification scheme based on geometric consistency and appearance diversity, reducing the risk of structural drift while ensuring sufficient appearance variation. Global candidates are generated once and rejected if verification fails, while Local Occlusion and Dual candidates use verification feedback for limited prompt refinement and regeneration. Using this framework, we construct AdaptCities, containing 160K verified synthetic same-place hard positives. Experiments across multiple VPR baselines and vision foundation backbones show consistent gains on standard benchmarks and substantial improvements under challenging domain shifts, with R@1 gains of up to 9.2%. The source code and data resources are publicly available at this https URL.
[CV-81] Object Concepts Emerge from Motion
链接: https://arxiv.org/abs/2609.04348
作者: Boshi Li,Xiaohui Wang,Xiaoyang Wu,Zhichao Li,Ya Yang,Naiyan Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Object-centric visual representations are important for physical-world perception, but existing visual pretraining methods often capture semantic categories without preserving the identity and coherence of individual instances. We present a biologically inspired framework that learns object-centric representations for single images from raw videos. Our approach uses motion boundaries as a source of object-level grouping: off-the-shelf optical flow and clustering produce pseudo-instance masks, which supervise a single-image encoder with pixel-level pairwise metric learning. The framework requires neither human annotations nor camera calibration. We first obtain 195 million pseudo-labeled frames from 7,163 hours of driving and web videos, then expand the supervision to 421 million frames with Motion-Verified Self-Training, which combines model proposals with motion evidence. We train encoders up to Swin-H and distill the learned representations into a family of Swin backbones. Across monocular depth estimation, 3D object detection, 3D occupancy prediction, and end-to-end planning, the resulting models achieve competitive or superior performance relative to supervised and self-supervised pretraining baselines, with particularly strong transfer on geometry- and instance-sensitive tasks. These results show that motion-derived supervision can teach static image encoders to represent visual instances, providing a complementary direction for scalable visual pretraining.
[CV-82] he microscope is the mask: privileged views and labels from a cryo-ET forward model
链接: https://arxiv.org/abs/2609.04325
作者: Bogdan Toader,Kiarash Jamali,Tanmay A. M. Bharat,Sjors H. W. Scheres
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Optimization and Control (math.OC)
备注: 18 pages, 7 figures
Abstract:We explore the use of simulated data for training a model for protein annotation in crowded cryo-electron tomography volumes reconstructed from images collected at limited tilt angles and severely corrupted by the measurement operator. Firstly, we leverage the corruptions imposed by the forward model to generate domain-specific augmented paired views of the exact same scene for an invariance objective integrated into the LeJEPA self-supervised training framework. Secondly, we use additional information from the simulation pipeline such as the positions and identity of proteins in the simulated volumes to inform the architecture of the model and the loss function, so that semantic information is localised at protein positions in the resulting dense feature volume. The resulting model, CARNIVAL, is evaluated without finetuning on classification and detection tasks in real tomograms, using a benchmark dataset containing multiple protein types and two tomogram processing types. We show that CARNIVAL outperforms a state-of-the-art model trained using a contrastive objective on simulated data but without forward model-based paired views or privileged information.
[CV-83] Joint Alignment and Distillation for Video Generation via Sample-Guided Distribution Matching
链接: https://arxiv.org/abs/2609.04283
作者: Jiuzhou Lin,Junlong Wu,Fei Zuo,Huan Ouyang,Dewen Fan,Boheng Zhang,Huaiqing Wang,Jia Sun,Fan Yang,Houde Liu,Kehai Chen,Min Zhang,Tingting Gao,Han Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Aligning video generative models to human preferences heavily relies on Reinforcement Learning (RL), which suffers from extensive computational overhead. Existing workflows typically treat RL and distillation as disconnected stages: applying RL before distillation incurs prohibitive computational costs, whereas applying RL after distillation frequently leads to model collapse. To overcome these limitations, we propose a unified, single-stage optimization framework grounded in Distribution Matching (DM). In the standard DM framework, distillation updates the model via a gradient direction that minimizes the gap between the real and fake models, guiding generations toward clarity and high fidelity. Building upon this, we introduce DM-Align, which derives a complementary gradient direction to guide the model toward human-preferred samples. Inspired by DPO and GRPO, our method leverages the distributional gap – formulated from either preference pairs or intra-group exploration – to directly construct this preference-guided gradient. By synergizing these two gradient directions, our approach eliminates the need for multi-step reward evaluation and complex ODE-SDE conversions inherent in traditional RL. Comprehensive experiments across multiple foundational video models demonstrate that this sample-guided framework robustly enhances both distillation quality and preference alignment, consistently outperforming both standalone variants and sequential two-stage pipelines.
[CV-84] Step Back to Move Forward: Reflection-Aware Preference Optimization for Visual Generation
链接: https://arxiv.org/abs/2609.04282
作者: Junlong Wu,Jiuzhou Lin,Jia Sun,Boheng Zhang,Huaiqing Wang,Dewen Fan,Houde Liu,Qianqian Gan,Fan Yang,Tingting Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Diffusion models have become the mainstream paradigm for modern visual generation and have substantially advanced multimedia content synthesis, especially in text-to-image and text-to-video tasks. To further align such generative models with human preferences, reinforcement learning (RL) has recently shown strong potential as a post-training strategy. Nevertheless, existing policy gradient-based methods often explore inefficiently, making them vulnerable to local optima that may degrade semantic faithfulness and visual realism. To address these challenges, we present Reflection-Aware GRPO (RA-GRPO), a new RL-based preference alignment framework for diffusion generative models. The core idea is to improve “forward” generation by incorporating “backward” reflection during optimization. We first introduce Diffusion Reflection, which rectifies intermediate sampling trajectories by inverting the diffusion process with a weak estimator, guiding latent states toward higher-probability regions of the true data manifold. Furthermore, we introduce Counterfactual Path Synthesis to implicitly distill these rectified trajectories into the policy, enabling the model to internalize the benefits of search-based exploration without incurring inference-time overhead. Extensive experiments on T2I and T2V models demonstrate that RA-GRPO significantly outperforms existing methods, particularly in mitigating reward hacking and improving generalization. The method remains architecture-agnostic and integrates seamlessly with standard pipelines, suggesting a promising direction for stable preference alignment.
[CV-85] When Seeing Overrides Knowing: Visual Dominance and Deferral-Based Method for Personalized Safety in VLMs
链接: https://arxiv.org/abs/2609.04281
作者: Edward Sun,Yuchen Wu,Zixian Ma,Eric Hanchen Jiang,Yijia Xiao,Xiaoyuan Yi,Ranjay Krishna,Wei Wang,Jindong Wang,Aylin Caliskan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Published as a main conference paper at COLM 2026
Abstract:Vision-language models (VLMs) are increasingly deployed in high-stakes settings, where a response that is reasonable in general may still be unsafe for a particular user whose medical, emotional, or situational context is unknown to the model. We study this problem of personalized safety in multimodal systems and introduce MPS-Bench, a benchmark of 5,181 scenarios from 584 real-world images across 12 high-risk domains, each paired with a hidden user profile. Evaluating eight frontier VLMs, we find that they almost always respond directly (86-99%) rather than seek missing context, and none exceeds 2.6/5 on personalized safety. To understand why these failures arise, we analyze multimodal interactions and identify visual dominance: visual information enters text representations early and suppresses textual risk signals during multimodal fusion. Causal interventions reveal a two-stage mechanism in which visual affect is first transferred into the text stream in early layers and then shapes the final decision through this altered text representation, making late-stage internal remediation unreliable. Motivated by this mechanism, we propose PRISM, a lightweight input monitor that uses bidirectional cross-modal modulation to predict when a query is likely to require deferral. PRISM achieves 0.978 AUC and strictly dominates the safety-utility Pareto frontier across all tested models.
[CV-86] FailureSpot: Label-Efficient Timestamp-Level Failure Detection for Vision-Language-Action Models
链接: https://arxiv.org/abs/2609.04277
作者: Jie Ma,Zongxi Liu,Yi Zhu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-language-action (VLA) policies have shown strong potential for general-purpose robotic manipulation, but they can still fail unpredictably during long-horizon execution, making reliable failure detection essential for safe deployment. Existing methods either rely on visual models that typically detect failures only after erroneous actions have occurred, or use lightweight proactive detectors trained on VLA internal representations. However, these proactive methods are often supervised with trajectory-level labels, causing normal pre-failure behavior in unsuccessful trajectories to be incorrectly labeled as failure. This supervision mismatch introduces label noise and limits both trajectory-level detection accuracy and precise timestamp-level failure localization. In this work, we study fine-grained timestamp-level VLA failure detection while addressing the cost of dense annotation. We propose a data-efficient framework that first leverages unlabeled VLA action chunks to construct action-derived weak supervision signals, capturing abnormal patterns such as inconsistent consecutive chunks, frozen or idle actions, and aggressive random motions. We then use active learning to select only the most uncertain trajectories for timestamp-level annotation and fine-tune the detector with these informative labels. Experiments across multiple VLA policies show that our method improves both timestamp-level and trajectory-level failure detection performance.
[CV-87] FailSAE: Towards Interpretable Failure Prediction for Vision-Language Models via Sparse Autoencoders
链接: https://arxiv.org/abs/2609.04276
作者: Jie Ma,Zongxi Liu,Yi Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-language models (VLMs), such as CLIP, have achieved strong performance across multimodal tasks by aligning visual and textual representations in a shared embedding space. As VLMs are increasingly used for high-stakes domains, failure prediction becomes critical for risk-aware deployment and human intervention. Existing failure prediction methods typically rely on confidence scores or auxiliary classifiers. Although these methods are effective on predicting VLM failures, they provide limited interpretability. In this work, we investigate the use of Sparse Autoencoders (SAEs) for interpretable failure prediction in VLMs. We formulate failure prediction as a classification task over sparse SAE latent activations and introduce a three-stage failure-aware training pipeline that encourages the learned latent directions to remain interpretable while becoming more informative for failure prediction. Our experiments show that the resulting framework outperforms the evaluated baselines in failure prediction. Further analysis suggests that failure-aware training encourages SAE latent directions to capture more class-specific concepts. We also use the SAE to provide a concept-level analysis of how model representations change during failures, revealing a shift from class-specific concepts toward more ambiguous or style-related concepts. Finally, we explore how the learned SAE latent directions can support runtime failure recovery.
[CV-88] AVENUE: Audio-Video EditiNg Understanding and Evaluation
链接: https://arxiv.org/abs/2609.04253
作者: Hayeon Kim,Yoojin Jang,Jaejun Yoo
类目: Multimedia (cs.MM); Computer Vision and Pattern Recognition (cs.CV); Sound (cs.SD)
备注:
Abstract:Audio-video (AV) editing aims to modify audio and video content according to a target prompt. Unlike single-modality editing, AV editing requires models to infer a modality-selective edit scope from the prompt alone: determining not only what should change, but also which modality should be preserved. Faithfully evaluating such models therefore requires both (i) benchmarks that span diverse edit types and modality categories, and (ii) evaluation that is itself modality-aware and sample-specific. However, existing AV editing benchmarks provide limited coverage of edit types and modality combinations, while current evaluation systems are often modality-blind and sample-agnostic, making it difficult to assess whether models faithfully preserve the unintended modality. To address these gaps, we introduce AVENUE, Audio-Video EditiNg Understanding and Evaluation, comprising two contributions: (1) a benchmark of 1,291 source clips and 7,957 editing instructions across audio-targeted, video-targeted, and AV-coupled edit types, curated and human-verified from VGGSound; and (2) a sample-specific, modality-aware evaluation framework that specifies, for each sample, both the intended change and the content that must remain intact. We evaluate representative AV editing models spanning three editing paradigms : joint, sequential, and separate, providing the first systematic analysis of modality-selectivity across paradigms. Our findings reveal a fundamental open challenge: when editing one modality, existing models frequently induce unintended changes in the other, regardless of paradigm. AVENUE provides a benchmark and modality-aware evaluation framework to drive progress toward more controllable AV editing models. Our dataset is publicly available on Hugging Face: this https URL.
[CV-89] Motion-Omni: End-to-End Joint Speech and Full-Body Motion for Spoken Dialogue
链接: https://arxiv.org/abs/2609.04250
作者: Chengqian Ma,Wei Tao,Haoyu Zhang,Yiwen Guo
类目: ound (cs.SD); Computer Vision and Pattern Recognition (cs.CV); Audio and Speech Processing (eess.AS)
备注: 28 pages, 4 figures, 12 tables. Project page: this https URL Code: this https URL Data: this https URL
Abstract:An avatar that holds a conversation should decide what to say and to move while saying it, yet these abilities live in separate model families: spoken dialogue models produce speech without motion, and co-speech motion models produce motion only from audio handed to them. The standard remedy is a cascade that first generates the spoken response and then runs a motion model over the finished audio, which requires a second full inference pass and precludes any joint optimisation between the two. We present Motion-Omni, an end-to-end framework in which a spoken dialogue model natively outputs explicit facial expression together with hand, upper-body and lower-body motion, generated directly from the hidden states that produce the speech. Joint training is not optional here: with the speech pathway frozen, motion remains misaligned with the audio, and co-adapting the LLM, Speech Generator and Motion Generator under both objectives is what recovers alignment while retaining spoken-dialogue ability. Supervision comes from a scalable, model-agnostic pipeline that pseudo-labels consistent-voice speech responses with a replaceable motion teacher, yielding 422,856 quality-ranked pairs (1,402 hours). We further release SwDA-500 and, to our knowledge, the first public evaluation protocol for stochastic open-ended full-body spoken dialogue, matching audio across motion systems while unifying rendering, automatic metrics, human evaluation, and latency measurement. Instantiated with a Qwen2.5-7B-Instruct backbone, Motion-Omni-Q7 matches the same-audio teacher cascade to within 2% on reference-free motion metrics while responding 5.4 x faster (RTF=0.78, faster than real time), surpasses all non-teacher cascades on beat correlation and diversity, and reaches a 2.62% word error rate, the lowest among the omni-modal systems compared.
[CV-90] Encore: Infinite Audio-Video Generation with Adaptive Signal Routing SIGGRAPH
链接: https://arxiv.org/abs/2609.04249
作者: Shaohua Pan,Junbao Chen,Shengyi He,Jingfeng Xue,Wen Tao,Haocheng Feng,Siming Fan,Dongwei Pan,Yi Yang,Wei He,Hang Zhou
类目: Multimedia (cs.MM); Computer Vision and Pattern Recognition (cs.CV); Sound (cs.SD)
备注: Accepted by SIGGRAPH ASIA 2026
Abstract:Existing audio-video generation methods produce well-synchronized clips but are limited to short durations, while long-video generation methods extend duration through chunk-based iterative synthesis yet lack audio entirely. Generating long audio-video jointly is fundamentally harder than either task alone: each chunk must simultaneously maintain video temporal coherence, audio temporal coherence, and cross-modal synchronization, whose conditioning signals enter the model through different pathways. In this work, we present Encore for long-form synchronized audio-video generation. Our key insight is to factor this challenge into: (1) local continuity handled by iterative generation with explicit cross-chunk context propagation, and (2) global consistency enforced via reference audio-video signals with shifted position embedding. Building on this design, we propose Adaptive Signal Routing (ASR), which introduces learnable attention biases within self-attention and learnable residual scales on cross-attention outputs, enabling the model to adaptively modulate the influence of each conditioning signal. Trained end-to-end for joint audio-video generation, Encore also supports infinite-length audio-to-video and video-to-audio synthesis at inference by conditioning on the ground-truth modality throughout the denoising process. Experiments on our extended VerseBench for long audio-video evaluation demonstrate that Encore significantly outperforms existing methods in both generation quality and temporal coherence. Code and data for this paper are at this https URL.
[CV-91] Real-World Multi-Modal and Longitudinal Lung Cancer Dataset ECCV2026
链接: https://arxiv.org/abs/2609.05202
作者: Rita Cordeiro Mendes,Maria Rita Fonseca Verdelho,Carlos Santiago,Catarina Barata
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at ECCV 2026 Workshop on Data Curation Augmentation in Medical Imaging
Abstract:Multi-modal learning has demonstrated strong potential in medical applications by integrating heterogeneous data sources such as medical imaging, clinical records, and genomics to improve predictive performance and support clinical decision-making. However, advances in this area are often constrained by two key challenges: the limited availability of well-curated, ready-to-use datasets that accurately reflect real-world conditions, where medical data are frequently collected inconsistently and are often incomplete; and the inherent difficulty of integrating heterogeneous data modalities. In this work, we introduce a newly curated multi-center, multi-modal, and longitudinal dataset designed to support the evaluation of a wide range of learning pipelines under realistic conditions. The dataset comprises a total of 1,365 lung cancer patients and has three imaging modalities (whole-slide images, CT scans, and PET scans), structured clinical data, transcriptomic, and longitudinal follow-up and treatment information. For each imaging modality the dataset contains more than one instance. Moreover, the dataset exhibits substantial and non-uniform missingness across modalities, making it well-suited for studying robust multi-modal fusion strategies. We further provide both uni-modal and multi-modal benchmarks on the task of 12-month overall survival prediction, disease-specific survival, as well as longitudinal benchmark of hazard prediction under severe missing data. Our results show that, despite high levels of missingness, integrating complementary modalities consistently improves predictive performance over uni-modal approaches, highlighting the value of multi-modal fusion in realistic clinical settings. The dataset and benchmark code are available at this https URL. Comments: Accepted at ECCV 2026 Workshop on Data Curation Augmentation in Medical Imaging Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.05202 [eess.IV] (or arXiv:2609.05202v1 [eess.IV] for this version) https://doi.org/10.48550/arXiv.2609.05202 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-92] Cross-dataset transportability of pediatric chest X-ray deep learning across three countries: discrimination calibration operating-point failure and limited-label recovery
链接: https://arxiv.org/abs/2609.05140
作者: Nazim-E-Alam
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Background and Objective: External evaluation of medical-imaging AI is often collapsed into discrimination. We evaluated a computational protocol that separately tests discrimination, probability calibration, fixed operatingpoint transport, shortcut-associated signal, and limited-label recoverability for pediatric pneumonia classification across datasets from three countries. Methods: After exact-duplicate removal, 5,824 Guangzhou radiographs supported leakage-controlled source development and internal testing. A frozen three-seed DenseNet121 dual-view ensemble was evaluated zero-shot on BDCXR-3257 from Bangladesh (n = 3, 257) and an untouched harmonized VinDr-PCXR/PediCXR test cohort from Vietnam (n = 1, 077). Matched seed-42 variants tested architectural robustness. Secondary BDCXR analyses used a fixed 651-image adaptation pool and 2,606-image hold-out; 163, 326, and 651 labels represented 5%, 10%, and 20% of complete BDCXR. Results: Internal AUROC was 0.976 with 95.1% sensitivity. BDCXR and VinDr-PCXR AUROC were 0.798 and 0.742, while frozen-threshold sensitivity fell to 6.2% and 0%. Source-to-BDCXR AUROC degradation occurred for a full-image baseline (0.961 to 0.749), ungated dual-view model (0.977 to 0.766), and gated MixStyle model (0.966 to 0.789). With 163 BDCXR labels, Platt recalibration preserved AUROC while increasing held-out sensitivity to 88.3%, but specificity was 47.9% and the alert rate was 78.5%. Two hundred repeated 163-label fits confirmed sensitivity recovery but substantial specificity variability. Conclusions: Cross-dataset shifts across countries affected ranking, probability alignment, and source-defined decision behavior differently. Transport studies should evaluate these components separately and quantify the operational burden of apparent recovery.
[CV-93] BEAM3R: Beams-eye-view architecture with Mamba-3 for implicit dose reconstruction
链接: https://arxiv.org/abs/2609.04747
作者: Chen Cheng,Michael Ferraro,James Grover,David E J Waddington,Emily Hewson
类目: Medical Physics (physics.med-ph); Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 5 figures
Abstract:To enable accurate and rapid photon control point and proton beamlet dose calculation in the DoseRAD2026 challenge, we present BEAM3R, a dose estimation framework operating in beam’s-eye-view (BEV). Our core innovation combines a Mamba-3 state-space depth-sequence core with physics-based transport conditioning to model long-range depth transport without expensive 3D convolutions. BEAM3R shares a 2D CNN encoder-decoder architecture for photon and proton dose tasks, processing per-plane BEV slices. Proton beamlets are conditioned on water equivalent thickness and remaining range, encoding the parameters determining Bragg peak position. Photon models use a bidirectional Mamba-3 core to capture dose contributions from materials downstream of the calculation point, while the proton model uses a forward core with learned energy-prefix tokens and a Bragg-peak refinement module. To reduce interpolation artifacts and support high spatial resolution, we introduce axial grid alignment of BEV lattices with CT slices and an implicit super-resolution representation via sub-pixel phase packing, evaluated by a differentiable Triton-accelerated resampler that reconstructs packed cubic B-spline coefficients directly in CT space. For MRI-based tasks, synthetic CTs (sCT) are generated by a patch-based conditional GAN with a SwinUNETR backbone. On the preliminary DoseRAD2026 test set, CT-to-photon and CT-to-proton models achieved 1%/1 mm local gamma pass rates of 96.8% and 96.0%, with stratified plan-level MAEs of 0.0041 and 0.0079. Substituting sCT reduced gamma pass rates to 89.7% for photon and 75.4% proton plan level doses, with stratified plan-level MAEs of 0.0093 and 0.0336. Standardised runtimes were 23.4 s and 18.4 s for CT-to-photon and CT-to-proton prediction, increasing to 39.7 s and 42.8 s for the corresponding MRI-based pipelines.
[CV-94] Explainable Multimodal Deep Learning Integrating Imaging and Clinical Data for Oral Potentially Malignant Disorder Detection
链接: https://arxiv.org/abs/2609.04512
作者: Ruilin You,Yihan Wang,Jiabin Chen,Cherie Wink,Petra Wilder-Smith,Rongguang Liang,Bofan Song
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Oral potentially malignant disorders (OPMDs) are critical precursors to oral cancer, yet clinical detection remains challenging because of substantial phenotypic heterogeneity and overlap with benign conditions. Although image-based deep learning shows promise for automated screening, visual information alone may be insufficient in real-world settings, where diagnostic decisions also rely on patient-specific risk factors. We developed M2-OPMDNet, a multimodal deep learning framework that integrates co-registered white-light and autofluorescence intraoral images with structured clinical information for OPMD detection. A customized questionnaire was designed to capture clinically relevant risk factors and symptoms in a standardized, reproducible format for integration with image-derived features. Multiple image encoders, including conventional convolutional neural networks and foundation model-based architectures, were evaluated using a prospectively collected dataset reflecting real-world screening conditions. Model interpretability was assessed using SHapley Additive exPlanations (SHAP) to quantify feature- and modality-level contributions. M2-OPMDNet achieved an AUC of 0.952, outperforming unimodal approaches and showing improved performance for visually subtle lesions. SHAP analysis demonstrated that structured clinical variables contributed substantially to risk estimation and complemented imaging features. These results demonstrate that explainable multimodal learning combining white-light and autofluorescence imaging with structured clinical data can provide accurate, transparent, and clinically grounded OPMD detection. M2-OPMDNet offers a scalable framework for real-world oral cancer screening and decision support.
[CV-95] Fractional-Order Adaptive Motion Magnification: Phase-Reliability Weighting for Noise-Constrained Video Amplification
链接: https://arxiv.org/abs/2609.04502
作者: Alejandro Garnung Menéndez
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages, 6 figures, 8 tables
Abstract:Eulerian video amplification boosts sub-pixel motion by band-pass filtering per-pixel intensity traces and applying a uniform gain. That gain ignores local structure, so sensor noise is amplified together with the signal, especially in textureless regions where the monogenic phase is unreliable. We propose FrAM (Fractional-order Adaptive Motion Magnification), a pipeline developed first offline and then as a causal stream. It replaces the constant temporal gain with a Grünwald–Letnikov derivative of fractional order, giving continuous control over high-frequency emphasis, and replaces the uniform spatial gain with a per-pixel weight derived from the local amplitude of the monogenic signal. On a controlled synthetic sequence split into textured and flat halves, FrAM matches the amplification of the Eulerian baseline while keeping flat-region temporal noise at the input level. The reduction holds across an eightfold range of input noise levels. Real videos show improved spatial selectivity and lower background noise in every case. The causal reformulation cuts the per-frame cost by two orders of magnitude, reaching 69,fps at 640 \times 480.
[CV-96] Development and Evaluation of Ultrasound Image Learning Pipelines for MASLD Risk Stratification
链接: https://arxiv.org/abs/2609.04390
作者: Guangyi Zhang,Xiaohong Wang,Eugene Cheah,Peng Guo,Brian A. Telfer,Theodore T. Pierce,Anthony E. Samir
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 7 pages, 4 figures. Accepted and presented at IEEE EMBC 2026
Abstract:Metabolic dysfunction-associated steatotic liver disease (MASLD) affects approximately 30% of the general population. Ultrasound-based imaging, including B-mode imaging and shear wave elastography (SWE), is widely used for noninvasive fibrosis assessment; however, the role of deep learning-based ultrasound image learning for MASLD risk stratification remains insufficiently characterized. In this study, we developed and evaluated ultrasound image learning pipelines using B-mode and SWE images for fibrosis staging and identification of patients with at-risk metabolic dysfunction-associated steatohepatitis (MASH). A total of 250 ultrasound examinations, one exam per subject, were included. Model performance was evaluated using 3-fold cross-validation with area under the receiver operating characteristic curve (AUROC). End-to-end SWE image learning achieved performance comparable to operator-guided SWE across fibrosis stages. Overall, SWE-based learning consistently outperformed B-mode image learning in fibrosis staging, with AUROC improvements from 0.64 (95%CI: [0.56, 0.72]) to 0.72 (95% CI: [0.65, 0.79]) for F=2 (significant fibrosis, p=0.11), from 0.67 (95%CI: [0.58, 0.75]) to 0.78 (95% CI:[0.72, 0.85]) for F=3 (advanced fibrosis, p=0.02), and from 0.69 (95%CI: [0.56, 0.82]) to 0.80 (95%CI: [0.72, 0.89]) for F4 (cirrhosis, p=0.10). These findings highlight the potential of SWE image learning for MASLD risk stratification.
[CV-97] Ultrasound-Based Prediction of Cirrhosis Decompensation Using Large-Scale Computer Vision Models
链接: https://arxiv.org/abs/2609.04365
作者: Guangyi Zhang,Peiyun Ni,Eugene Cheah,Rajat Chandra,Peng Guo,Raymond T. Chung,Anthony E. Samir
类目: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: 6 pages, 2 figures. Accepted and presented at IEEE EMBC 2026
Abstract:Decompensation represents a critical transition in the course of cirrhosis, yet clinicians have limited non-invasive tools to reliably predict its onset. In this study, we propose a novel imaging-based approach that leverages large-scale computer vision models to analyze routine abdominal ultrasound images and extract predictive features beyond those captured by traditional laboratory-based risk scores. Ultrasound is widely available, low cost, and suitable for longitudinal surveillance, making it an attractive modality for scalable risk stratification and long-term follow-up. Our framework integrates automated ultrasound data processing with modern deep learning architectures to identify patients at high risk of decompensation prior to the occurrence of clinical deterioration. This non-invasive strategy offers a practical complement to existing clinical scoring systems and may enable earlier, more proactive management of patients with compensated cirrhosis.
[CV-98] Cross-modal triage network: a multimodal deep learning framework for severity-based triage and visual explainability in chest radiographs
链接: https://arxiv.org/abs/2609.04357
作者: Zinah Ghulam,Richa Mittal,Eranga Ukwatta
类目: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Purpose: Increased number of chest radiograph (CXR) scans create a triage bottleneck, queueing urgent examinations behind routine ones. Existing AI tools are predominantly unimodal binary classifiers lacking severity awareness, and multimodal systems are rarely benchmarked against expert radiologists. To this end, we developed a multimodal deep learning framework for joint severity triage, pathology detection, and native visual explanation. Approach: We propose the cross-modal triage network (CMTN), fusing a Swin Transformer V2 visual encoder with a PubMedBERT text encoder via gated cross-attention. The CMTN was trained on 34,639 image-text pairs (12,489 patients) from MIMIC-CXR-JPG, optimizing an ordinal focal loss for four-tier severity triage and binary cross-entropy for 14 pathologies. Beyond quantitative benchmarking, attention heatmaps were evaluated against a blinded expert radiologist in a two-phase clinical audit comparing model triage output to expert severity assessment (100 cases) and grading spatial-semantic concordance (116 heatmaps). Results: The CMTN achieved strong ordinal agreement with reference labels (quadratic weighted kappa [QWK] = 0.9341, 95% CI: 0.9219 to 0.9449) and macro-AUROC of 0.9970 across 14 pathologies, with 34~ms latency, outperforming the state-of-the-art BioViL multimodal baseline (QWK = 0.7679). However, the blinded Phase I clinical audit revealed substantially lower agreement with genuine radiologist judgment (QWK = 0.1399). Phase II found 54.3% of heatmaps achieved clinically acceptable spatial localization. Conclusions: The CMTN demonstrated an efficient multimodal architecture for CXR triage. The divergence between algorithmic and radiologist agreement demonstrates that benchmark performance against NLP-derived labels is insufficient, highlighting the need for radiologist-labeled ground truth before clinical deployment.
[CV-99] Multi-scale Image Representation Compression
链接: https://arxiv.org/abs/2609.04274
作者: Tianhao Peng,Ho Man Kwan,Fan Zhang,Shan Liu,David Bull
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注:
Abstract:Overfitted codecs have demonstrated promising performance for image and video compression. In particular, for image compression, the Cool-chic family of models has shown competitive performance against scene-agnostic models, with orders of magnitude lower decoding complexity at the cost of a longer overfitting process. However, these overfitted image codecs are not fully optimized toward the rate-distortion objective: their network weights remain in full precision during training, and the associated quantization parameters are selected in a separate post-training stage. Furthermore, their synthesis operates at a single scale, which overlooks cross-scale redundancy. In this paper, we propose MIRC, an overfitted image codec in which every coded component, including the latents, the synthesis network, and the entropy models, is quantized and entropy coded under a single rate-distortion objective, adopting the end-to-end compression pipeline of the neural video representation codec NVRC. We further introduce a multi-scale representation with cross-stage parameter sharing, which improves coding efficiency at a small transmitted overhead. On the CLIC2020 professional validation set, MIRC achieves a 10.5% BD-rate saving against VVC (VTM 22.0). Moreover, MIRC offers a family of configurations spanning 1.2 to 2.9 kMAC per pixel, so the decoding budget can be selected to match the deployment target.
[CV-100] Scalable Neural Video Representation Compression
链接: https://arxiv.org/abs/2609.04273
作者: Tianhao Peng,Ho Man Kwan,Fan Zhang,Shan Liu,David Bull
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注:
Abstract:Scalable video coding (SVC) encodes a video into a layered bitstream consisting of a base layer and one or multiple enhancement layers, enabling decoding at different bitrate/quality/resolution operating points to accommodate diverse device capabilities and network conditions. Due to its practical flexibility, SVC has been incorporated into major video coding standards and has recently attracted growing interest for both scene-agnostic and scene-adaptive neural video codecs. Among the latter, Implicit neural representation (INR) based codecs achieve compression by overfitting a compact neural network to an individual video, offering fast decoding and competitive coding efficiency compared to scene-agnostic neural codecs. However, research on scalable INR-based compression remains in its infancy: these methods support scalable coding by introducing additional network layers, which couple the bitrate with the decoding complexity and also cannot achieve comparable performance with strong scalable/non-scalable codecs. In this context, this paper proposes S-NVRC, a scalable INR-based video codec that jointly supports fine-grained bitrate and decoding complexity scalability from a single embedded bitstream. It adopts a coarse-to-fine prefix for feature grids and a nested prefix for network layers, which scale bitrate and decoding complexity, respectively. The proposed S-NVRC spans a wide range of bitrate and decoding-complexity using a single encoding (training) and outperforms SHM 12.4 and the multi-layer VTM-20.0, by 43.7% and 5.6% in BD-rate on the UVG dataset, while also providing flexible complexity scalability. Implemented code will be provided.
[CV-101] raining-Free Speech-Centric Omni Understanding with Frozen VLMs
链接: https://arxiv.org/abs/2609.04242
作者: Ankan Deria,Hanoona Rasheed,Xilin He,Fahad Shahbaz Khan,Salman Khan
类目: Audio and Speech Processing (eess.AS); Computer Vision and Pattern Recognition (cs.CV); Sound (cs.SD)
备注: 18 Pages, 13 Tables, 3 Figures
Abstract:Audio-visual understanding remains challenging because models must jointly interpret spoken content, visual events, and their temporal relationships. Existing omni models typically introduce dedicated audio encoders and rely on expensive audio-video-text training, tightly coupling omni capability to specific VLM backbones and potentially weakening their existing visual and reasoning abilities. This raises three questions: whether native omni training is necessary for every new VLM, whether speech-centric omni capability can be added while preserving the original backbone, and where richer acoustic representations remain essential. We introduce Training-Free Omni (TFO), a plug-and-play framework that converts a frozen VLM into a speech-centric omni model without architectural modification, or multimodal re-alignment. TFO uses Whisper to extract confidence-filtered, timestamped transcripts and routes them through the VLM’s existing language interface, while leaving its visual pathway unchanged. Across matched comparisons with native omni models on 56 benchmarks and 21 languages, TFO is competitive on audio-visual understanding, improves average audio-only performance across all five model settings, and achieves substantial multilingual speech gains. Freezing the VLM also generally preserves stronger image/video understanding, visual grounding, coding, mathematical reasoning, and medical question answering than the corresponding native omni checkpoints. These results show that strong speech-centric omni understanding can often be obtained through modular audio-to-language routing rather than costly backbone-specific training. Comments: 18 Pages, 13 Tables, 3 Figures Subjects: Audio and Speech Processing (eess.AS); Computer Vision and Pattern Recognition (cs.CV); Sound (cs.SD) Cite as: arXiv:2609.04242 [eess.AS] (or arXiv:2609.04242v1 [eess.AS] for this version) https://doi.org/10.48550/arXiv.2609.04242 Focus to learn more arXiv-issued DOI via DataCite
[CV-102] Collaborative On-Sensor Array Cameras SIGGRAPH2025
链接: https://arxiv.org/abs/2506.04061
作者: Jipeng Sun,Kaixuan Wei,Thomas Eboli,Congli Wang,Cheng Zheng,Zhihao Zhou,Arka Majumdar,Wolfgang Heidrich,Felix Heide
类目: Optics (physics.optics); Computer Vision and Pattern Recognition (cs.CV); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: To appear in ACM Transactions on Graphics and to be presented at SIGGRAPH 2025
Abstract:Modern nanofabrication techniques have enabled us to manipulate the wavefront of light with sub-wavelength-scale structures, offering the potential to replace bulky refractive surfaces in conventional optics with ultrathin metasurfaces. In theory, arrays of nanoposts provide unprecedented control over manipulating the wavefront in terms of phase, polarization, and amplitude at the nanometer resolution. A line of recent work successfully investigates flat computational cameras that replace compound lenses with a single metalens or an array of metasurfaces a few millimeters from the sensor. However, due to the inherent wavelength dependence of metalenses, in practice, these cameras do not match their refractive counterparts in image quality for broadband imaging, and may even suffer from hallucinations when relying on generative reconstruction methods. In this work, we investigate a collaborative array of metasurface elements that are jointly learned to perform broadband imaging. To this end, we learn a nanophotonics array with 100-million nanoposts that is end-to-end jointly optimized over the full visible spectrum–a design task that existing inverse design methods or learning approaches cannot support due to memory and compute limitations. We introduce a distributed meta-optics learning method to tackle this challenge. This allows us to optimize a large parameter array along with a learned meta-atom proxy and a non-generative reconstruction method that is parallax-aware and noise-aware. The proposed camera performs favorably in simulation and in all experimental tests irrespective of the scene illumination spectrum. Comments: To appear in ACM Transactions on Graphics and to be presented at SIGGRAPH 2025 Subjects: Optics (physics.optics); Computer Vision and Pattern Recognition (cs.CV); Distributed, Parallel, and Cluster Computing (cs.DC) ACMclasses: I.2.11; I.4; J.2 Cite as: arXiv:2506.04061 [physics.optics] (or arXiv:2506.04061v1 [physics.optics] for this version) https://doi.org/10.48550/arXiv.2506.04061 Focus to learn more arXiv-issued DOI via DataCite Related DOI: https://doi.org/10.1145/3731200 Focus to learn more DOI(s) linking to related resources
人工智能
[AI-0] RegionFed: Federated Learning for Personalized Query Understanding in Heterogeneous Retail Environments
链接: https://arxiv.org/abs/2609.05403
作者: Quoc H. Nguyen,Ali Lafzi,Abhijeet Phatak,Siddharth Pratap Singh,Rohit Upadhyay,Yogananda Domlur Seetharama,Chittaranjan Tripathy
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Retail search systems serve diverse geographic regions with distinct query patterns, vocabularies, and product preferences, creating significant data heterogeneity that challenges both privacy-preserving training and model personalization. Federated learning offers a natural solution for privacy, but standard FL methods produce global models that sacrifice regional performance, while existing personalized FL approaches operate at the parameter level and catastrophically collapse on modern transformers (below 10% accuracy on T5) due to tied embeddings and LayerNorm interactions. We introduce RegionFed, an \textitarchitecture-robust federated learning framework that sidesteps this failure by operating entirely at the gradient level. RegionFed uses the \ell_2 conflict between regional and global gradients as a unified signal that (i) diagnoses heterogeneity, (ii) routes each region to the cheapest sufficient personalization strategy, and (iii) adaptively controls personalization strength. Because it treats models as differentiable black boxes, RegionFed deploys on T5-Small, T5-3B, RoBERTa, and CNN with zero code changes, providing large gains on transformers (where parameter-level methods collapse) and consistent improvements on CNNs. Across three public datasets (Amazon ESCI, Amazon Reviews, LEAF-FEMNIST) and four architectures, RegionFed-Meta achieves 92.27%, closing the gap to the privacy-violating centralized upper bound (Centralized + Regional Weighting: 92.04%, \Delta =0.23pp, within 1 \sigma ) while providing (\epsilon\approx0.60) -differential privacy and \mathcalO(1/\sqrtT) convergence.
[AI-1] A Deep Generative Model for Synthesizing Labeled Wireless Signals
链接: https://arxiv.org/abs/2609.05396
作者: Yuxiao Li,Keke Hu,Santiago Mazuelas,Yuan Shen
类目: Artificial Intelligence (cs.AI)
备注: 12 pages
Abstract:Wireless signals with position-related labels are pivotal for both performance evaluation and model training in the realm of wireless sensing. However, acquiring real-world datasets is often challenged by significant measurement and labeling costs. Traditional methods for synthesizing labeled wireless signals typically rely on environmental models, leading to extensive hyper-parameter tuning and inadequate realism for comprehensive model training purposes. To address these limitations, we introduce a novel deep learning (DL)-based method, namely Inter-Instance Generative Adversarial Networks (IIns-GAN), to generate realistic labeled wireless signals. The generated signals are particularly adaptive to different environment scenarios and well-suited for various model training tasks, including distance estimation and environment identification. We have conducted extensive experiments on public Ultra-Wideband (UWB) datasets to evaluate the realism and utility of the generated signals. The results demonstrate that the signals generated by IIns-GAN mirror the physical characteristics of real-world measurements, and significantly contribute to the improvement of model training in diverse wireless sensing tasks.
[AI-2] Necessary or Sufficient? Evaluating LLM Explanations With Behavioural Evidence
链接: https://arxiv.org/abs/2609.05385
作者: Urja Pawar,Rajitha Ramanayake,Nabeel Kemal,Ashwin Kandath,Owen O’Neill,Guillaume Bourgeon,Houssem Chatbri
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LLM decision components that can operate within agent workflows often produce action-relevant recommendations or judgements together with explanations. Operators may use the named factors to monitor a system, diagnose errors, or decide when to escalate an output. Such use assumes that the explanations agree with the component’s observable decision behaviour. We test two interpretations of the named factors: necessity, meaning that changing a factor would change the output, and sufficiency, meaning that retaining it while removing other changeable information would preserve the output. We evaluate these interpretations in two synthetic use cases: recommending advisors to clients and judging prompts for harmfulness or risk. Models return an output and the top three factors that most influenced it. Controlled black-box interventions estimate a necessity score for each factor by measuring how often changing it changes the output, and a sufficiency score by measuring how often retaining it preserves the output. Across eight models from the Claude, GPT, and Gemini families, the mean Spearman correlations between the cited ranking and the necessity and sufficiency scores are 0.349 and 0.354 for advisor recommendation, and 0.431 and 0.580 for prompt monitoring. Furthermore, an uncited factor scores above the lowest-scoring cited factor in 57.6% of advisor responses under necessity and 58.1% under sufficiency; the corresponding prompt-monitoring rates are 25.8% and 8.9%. The cited top three contain useful information but do not reliably identify the three factors with the strongest measured influence under necessity or sufficiency. The framework provides a black-box reliability check for explanations used in agent oversight while remaining scoped to individual LLM decisions.
[AI-3] Molecular Déjà Vu: Digit-Level Retrieval of Published Values in Frontier Language Models
链接: https://arxiv.org/abs/2609.05381
作者: Matthias Busch,Marius Tacke,Sviatlana V. Lamaka,Mikhail L. Zheludkevich,Christian J. Cyron,Roland C. Aydin,Christian Feiler
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are increasingly evaluated on molecular property benchmarks, but accuracy cannot distinguish a model that predicts a property from one that retrieves a published number. We audit 22 frontier models on 12 regression benchmarks for verbatim retrieval and find that it is widespread but relatively benchmark-specific: on five datasets more than 50% of the LLMs show verbatim retrieval, while on the remaining datasets it appears only in isolated cells. We run our experiments at two reasoning levels and find that reasoning changes retrieval. The same experiments, on the same molecules and with the same prompt, are flagged 89% more often at the higher reasoning level than at the lowest one. Finally, we test a way to interrupt retrieval in our most contaminated cases, and find that the strongest models in some cases still recognise a combination of transformed SMILES strings and original labels. Furthermore, suppressing retrieval moves the prediction errors of the different models closer together in relative terms, while their differing use of verbatim retrieval spreads them apart. This indicates that the general predictive capability of an LLM is not determined solely by the amount of memorised values. This work provides an overview of the amount and depth of verbatim retrieval in molecular regression benchmarks using LLMs.
[AI-4] CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUICLI Agents
链接: https://arxiv.org/abs/2609.05374
作者: Haoting Shi,Wenhao Wang,Weicheng Fang,Yaozhong Liang,Tian Jin,Pengxiang Zhao,Guangyi Liu,Siheng Chen,Yanfeng Wang
类目: Artificial Intelligence (cs.AI)
备注: 21 pages, 9 figures
Abstract:Computer-use agents have advanced on benchmarks like OSWorld and AndroidWorld, but still act mostly through the GUI, often producing inefficient trajectories. Real-world computer work is hybrid, combining visual-state inspection with precise, high-throughput command-line operations, so capable agents must coordinate both modalities over shared application state. Yet scalable hybrid environments remain scarce because supporting both GUI and CLI over real applications typically requires substantial manual engineering for each application. Existing agents also struggle to use the two interfaces complementarily: CLI-native agents lack visual perception for tasks involving interface state or layout, while GUI-native agents are inefficient for operations better executed through commands. We introduce CUA-Universe, a scalable environment-to-data pipeline that turns real desktop software into hybrid GUI+CLI environments. App-Forge adapts applications into reproducible VMs and command-line surfaces it discovers, wraps, or generates, scaling to 16 applications; Task-Weave synthesizes diverse hybrid tasks of controllable difficulty from reusable operations over seed files; and Path-Steer steers rollouts along efficient hybrid paths and harvests verified trajectories for post-training. Training on this data shifts behavior from inefficient GUI interaction and brittle CLI scripting toward effective GUI+CLI orchestration. Our 9B model improves both success and efficiency on CUA-Verse (Score +39.3 pts; -37% steps, -60% tokens), OSWorld (SR +16.8 pts; -57% steps, -44% tokens), and OSWorld-MCP (Score +7.84 pts; -27% steps, -30% tokens). CUA-Universe provides a scalable path toward more capable and efficient computer-use agents.
[AI-5] When LLM Decompilers Recompile More and Preserve Less
链接: https://arxiv.org/abs/2609.05370
作者: Chang Liu,Edward Raff,Kristopher Micinski
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Decompilation recovers high-level source from compiled machine code and serves as a foundation for security tasks such as vulnerability detection and malware analysis. Traditional decompilers like Ghidra and Hex-Rays expose whatever they cannot resolve as visible placeholders and often emit pseudocode that will not compile or execute; LLM-based decompilers produce clean, idiomatic C and are now judged almost entirely by recompilability and re-executability: whether the output builds and passes its shipped input/output tests. We show that these metrics can reward the wrong path: a function may recompile and pass every shipped test yet diverge on other legitimate inputs, and a disclosed vulnerability may disappear from the recompiled code with no visible trace of the crash. Neither failure is caught by existing suites. To address this gap, we propose Decompile-Diverge, a behavioral comparison oracle not relying on fixed or hand-crafted tests: for each function it synthesizes a driver, grows a fuzzing corpus from the reference, and reruns the decompiled code on the same inputs to detect changes in the function’s behavior. Across eight systems in nine configurations on established LLM decompilation corpora, candidates that pass every shipped test still diverge from the original on our input corpus: 4.9% overall, and as many as 13% for a single system. On 300 real GitHub library functions and 287 CVE-grounded functions, recompilability and behavioral agreement can come apart: the strongest refinement LLM lifts Ghidra’s build rate from 75% to 90%, while its Matched rate falls from 74% to 62%; on disclosed vulnerabilities, up to one tenth exhibit Crash Absence in its output. Source-level analysis traces this divergence to introduced fields, types, callees, and guards that replace the visible unknowns traditional tools leave behind. Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.05370 [cs.CR] (or arXiv:2609.05370v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2609.05370 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-6] Design Docs Are All You Need: An AI-native Machine-Learning Performance Tool
链接: https://arxiv.org/abs/2609.05364
作者: Samuel Kushnir,Kimia Noorbakhsh,Kavya Sreedhar,Liqun Cheng,Ming Liu,Parthasarathy Ranganathan,Mohammad Alizadeh,Fred Kjolstad,Suvinay Subramanian
类目: Programming Languages (cs.PL); Artificial Intelligence (cs.AI)
备注:
Abstract:Machine-learning performance modeling is a uniquely hostile terrain for long-lived software: the assumptions baked into today’s abstractions are invalidated by tomorrow’s models and systems, forcing perpetual refactoring of performance-modeling frameworks. Meanwhile, AI coding agents have become fast and capable enough that regenerating an entire library is cheaper than paying down the tech debt of incrementally patching it. We describe SMART, a rigorous symbolic performance-modeling library for ML systems whose main branch contains almost no code: the repository is a DAG of self-contained natural-language design docs, coding sub-agents regenerate the implementation from only the docs on new version updates, and every human change is a natural-language edit to a doc–self-documenting by construction. Two ingredients make regeneration reliable: (i) a design-doc style built around step-by-step worked examples that act as in-context demonstrations for the generating agents, and (ii) a minimal, recursively defined operator IR with symbolic (SymPy) cost expressions, a fast analytical roll-up mode for large sweeps, and a slow modulo-scheduling mode for fine-grained schedule studies. Regenerated implementations reproduce hand-audited reference models–including DeepSeek-V3 serving on a TPU pod slice–to round-off precision, suggesting that design docs–not code–can be the durable artifact for ML-systems co-design tools.
[AI-7] Who Should Grade My Work? Student Perspectives on Transparent AI-Assisted Writing Assessment in Higher Education
链接: https://arxiv.org/abs/2609.05346
作者: Rayed AlGhamdi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The integration of GenAI tools into higher education assessment raises important questions about how students understand, interpret, and respond to AI-mediated evaluation. As instructors increasingly explore AI tools for providing feedback, prior research has examined whether GenAI-generated feedback improves writing performance and how students perceive its usefulness; comparatively little is known, however, about how students interpret such evaluation when they are explicitly informed that an AI system, rather than a human instructor, produced the feedback and the score. This study reports findings from a qualitative pedagogical inquiry conducted in an undergraduate technical communication course for computing students at a Saudi public university. Thirteen male undergraduate computing students completed an in-class handwritten writing task; the scanned submissions were evaluated by ChatGPT using a rubric-based prompt aligned with the task objectives. Students were then explicitly informed that ChatGPT had generated the score and feedback and were invited to reflect on the evaluation in writing. Inductive thematic analysis of these reflections identified four themes: perceived usefulness of feedback; awareness of AI’s contextual and pedagogical limitations; conditional trust, distinguishing feedback utility from evaluative authority; and reflection on the institutional and pedagogical role of the human instructor. Participants accepted GenAI feedback as useful for surface-level revision but consistently positioned the human instructor as the appropriate authority over grading decisions. The study identifies this as a distinction between feedback utility and evaluative authority, two judgments that students treat as analytically separate rather than as opposite ends of a single approval scale…
[AI-8] he History Is the Detector: Executing CVE Patch History End-to-End
链接: https://arxiv.org/abs/2609.05335
作者: Qiushi Wu,Kevin Eykholt,Youngja Park,Xiaokui Shu,Dhilung Kirat,Douglas Lee Schales,Ian Molloy
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Public vulnerability databases collect rich information about known software flaws, including their weakness types, affected components, and related patches. Fixing commits provide the exact code changes that removed these flaws. While these records capture why the original code was unsafe, they are documented mainly for human inspection rather than automated reuse. Consequently, the same unsafe conditions may still exist elsewhere in code without a known advisory, leaving much of this detection knowledge unused. We present BUGSTONE-E2E, a framework that transforms vulnerability history into executable detection rules and validates their findings. First, BUGSTONE-E2E mines reusable rules from verified fixing commits, capturing scan anchors, fix semantics, and CVE provenance and organizing them by CWE and language. Second, detection follows a funnel-shaped pipeline: early stages process a large pool of candidates using lightweight analysis, while later stages apply increasingly capable and expensive models to a shrinking set of targets. Specifically, BUGSTONE-E2E first enumerates call sites matching rule anchors using Tree-sitter, then removes benign sites using lightweight heuristics without LLM calls. Next, LLM-based agents inspect the remaining candidates guided by the rule. Following this inspection, the system re-triages surviving candidates and builds runtime verifications, then generates scope-checked patches validated via two-sided differential tests. Using 19,325 high-severity CVEs from 2022 to 2026, BUGSTONE-E2E identifies 2,710 fixing commits and constructs 1,033 detection rules across 56 CWE families, packaged into 172 skills. When applied across 14 programs, it produced runtime evidence for 644 findings. These results demonstrate that CVE history can be turned into an executable workflow, transforming past vulnerabilities into reproducible detection and repair. Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE) Cite as: arXiv:2609.05335 [cs.CR] (or arXiv:2609.05335v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2609.05335 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-9] LLM -Driven Algorithm Design for Quantum Circuit Synthesis based on Binary Decision Diagrams
链接: https://arxiv.org/abs/2609.05327
作者: Yoonju Sim,Federico Berto,Chuanbo Hua,Jinkyoo Park,Changhyun Kwon
类目: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
备注:
Abstract:Quantum circuits are central to implementing quantum algorithms on quantum devices, where quantum gates must be reversible. Many quantum algorithms rely on Boolean functions, which must therefore be implemented reversibly within quantum circuits. Reversible circuit synthesis provides a way to translate such Boolean functions into reversible circuits. Binary decision diagrams (BDDs) offer a scalable approach to this task, but the resulting BDDs and circuits depend heavily on variable ordering. Existing ordering heuristics commonly minimize BDD size because it is closely tied to the circuit size. However, BDD size is an imperfect proxy for the quantum cost of the synthesized circuit (QCC). We propose \textttQuantumEvo, an evolutionary framework that uses an LLM as a heuristic generator for QCC-aware BDD variable ordering. Instead of predicting orderings directly, \textttQuantumEvo searches over ordering heuristics initialized from multiple heuristic families. Candidate heuristics directly manipulate variable orderings using standard BDD operations and are selected by downstream QCC. The discovered heuristic, HGA-QE, modifies the sifting step inside a genetic algorithm so that the procedure is better aligned with QCC. Across the benchmark set, HGA-QE achieves a 70.9% tie-or-win rate against the per-function best baseline and is strictly best on 13.5% of the functions. The results demonstrate broadly competitive QCC performance, with HGA-QE showing a clearer relative advantage in strict wins on the two benchmark suites drawn from sources different from the data used for heuristic discovery.
[AI-10] How Does mHC Use Its Residual Streams? Selective Routing and Near-Identity Mixing
链接: https://arxiv.org/abs/2609.05309
作者: Pengxiang Zhao,Xing Li,Xianzhi Yu,Wei Guo,Zhenhua Dong
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Hyper-Connections and their manifold-constrained variant mHC widen a residual pathway from one stream to n, yet how trained models use this capacity remains unclear: how broadly blocks read and write, how strongly the residual pathway mixes streams, and whether the streams carry distinct representations. We examine these properties in the four-stream residual pathway of DeepSeek-V4-Flash using effective stream counts, cross-stream residual weights, and inter-stream cosine similarity. Read/write routing is concentrated but varies across depth: a typical attention or FFN site effectively uses about two streams, while the dominant stream changes across layers and the representations remain directionally distinct. Residual mixing is modest and occurs primarily in early layers; in layers 22-42, the pathway mostly carries each stream forward separately. Targeted interventions establish the functional significance of these patterns. Replacing the late mixers by identity increases C4 perplexity by only 1.9% and preserves the six-task average score, whereas replacing the early mixers increases perplexity by 41%. Fixing each early mixer to its C4 diagnostic mean increases perplexity by only 0.2% and reduces the average score by 0.25 percentage points, showing that its site-specific structure matters more than its token-wise variation on the evaluated metrics. Likewise, retaining the three largest routing weights per token at every site increases perplexity by at most 2.7% and changes the average score by at most 0.4 points. Thus, the studied model realizes only part of the flexibility afforded by four-stream mHC: individual blocks rarely require all four streams, and late residual mixing provides little measured benefit.
[AI-11] RISE: Recursive Improvement via Self-Extrapolating Policy Distillation
链接: https://arxiv.org/abs/2609.05295
作者: Yang Li,Semih Yavuz,Shafiq Joty
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:On-policy distillation (OPD) provides dense, per-token supervision for language model post-training, but its effectiveness is bottlenecked by teacher quality: external teachers suffer from distribution mismatch, while self-distillation with privileged conditioning is limited by in-context learning capacity. We propose \textbfRISE (\textbfRecursive \textbfImprovement via \textbfSelf-\textbfExtrapolating Policy Distillation), which constructs a synthetic teacher directly from the model’s own RLVR training trajectory. By extrapolating the displacement between the current checkpoint and a trailing anchor—in parameter space or output logit space—RISE converts a sparse outcome-induced parameter update into a dense token-level target, without any external model or privileged conditioning. RISE combines RLVR and OPD in a complementary loop: outcome rewards ground the extrapolation toward correct reasoning, while the extrapolated teacher refines token-level decisions. Moreover, since the teacher is refreshed every iteration as the student improves, distillation becomes a recursive improvement mechanism rather than a one-shot compression step. Experiments spanning mathematical reasoning, multi-domain STEM, code generation, and multi-turn agentic tasks show that RISE outperforms RLVR-only training and on-policy self-distillation across all settings.
[AI-12] Beyond Aggregate Scores: Behavioral Correctness Assumptions for Assessing Reference-Based Automatic Evaluation Methods
链接: https://arxiv.org/abs/2609.05289
作者: Maria Mahbub,Ashley Rice,Michael R. Munroe,Amidu Kamara,Amir Sadovnik
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automated reference-based evaluation methods play a critical role in assessing natural language generation systems. Existing meta-evaluation primarily measures agreement with human judgments or benchmark labels, providing limited insight into evaluator behavior under controlled conditions. We introduce behavioral correctness assumptions, a complementary framework for evaluating reference-based automatic evaluation methods. We define a taxonomy of correctness-preserving and correctness-altering assumptions and operationalize them through controlled response transformations that specify expected scoring behaviors. We evaluate diverse lexical, character-level, semantic, LLM-based, and hybrid evaluators and analyze their assumption-level behavior, stability, sensitivity, repeat-run variability, configuration sensitivity, and reproducibility. Our experiments reveal distinct behavioral trade-offs across evaluation paradigms: no evaluator satisfies all proposed correctness assumptions, and evaluators with similar aggregate performance can exhibit substantially different behavioral profiles. These findings demonstrate that behavioral correctness assumptions provide diagnostic information obscured by conventional aggregate meta-evaluation.
[AI-13] GUT: Quantifying and Optimizing the Reasoning Uncertainty of LLM s via Graph Complexity
链接: https://arxiv.org/abs/2609.05284
作者: Shuang Liang,Xin-Yu Hu,Xiang-Jun Ou,Shao-Qun Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recent years have witnessed great advances in the reasoning ability of Large Language Models (LLMs). However, the reasoning processes of LLMs often exhibit uncertainty, where LLMs often produce a proliferation of divergent branches at each reasoning step even when fed the same prompting inputs, and certain branches exhibit evidently incredible, even nonsensical, reasoning chains and results. In this paper, we propose the Graph-complexity-based UncerTainty (GUT) method for investigating the reasoning uncertainty of LLMs. The key idea of GUT is to characterize the potential branches of each reasoning chain with a directed acyclic graph, thereby ensuring that all potential branches are comprehensively covered within the graph space. Building upon this recognition, we further build two modules of GUT, that is, a Quantification (GUT-Q) module and an Optimization (GUT-O) module, for quantifying and reducing the reasoning uncertainty of LLMs, respectively. GUT-Q measures LLM reasoning uncertainty by approximating the reasoning space complexity with graph complexity. GUT-O implements uncertainty optimization by treating negative uncertainty as the reward function in reinforcement learning. Experimental results conducted on four LLMs and five datasets validate the effectiveness of GUT.
[AI-14] Dont Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference ICML2026
链接: https://arxiv.org/abs/2609.05275
作者: Mostafa Elhoushi,Alex Pretko,Nolan Dey,Bin Claire Zhang,Gavia Gray,Gurpreet Gosal,Abdulrahman Mahmoud,Shane Bergsma,Joel Hestness
类目: Artificial Intelligence (cs.AI)
备注: This is a slightly extended version of the paper published at ICML 2026
Abstract:Layer dropout (a.k.a. stochastic depth) has been shown to enable faster training, higher accuracy, and robustness to zero-shot layer pruning in both language and vision transformers. However, as models and datasets have scaled, dropout - particularly layer dropout - has largely disappeared from large language models (LLMs) pre-training recipes. While some prior work has reported that dropout can degrade accuracy, no comprehensive study has quantified, let alone mitigated, this effect. In this study, we show that layer dropout should be used in state-of-the-art LLM training, establishing best practices and scaling analysis for both training and post-training benefits. Concretely, with optimal layer distribution, time schedule, and optimizer hyperparameters, we observe that at the same training FLOPs layer dropout leads to lower loss. For a given number of training steps, LLMs can achieve lower or similar validation loss while saving upto 25% of training FLOPs. Moreover, layer dropout enables significant post-training optimizations, such as early exit, intermediate-layer skipping, and self-speculative decoding, yielding up to 1.5x inference speedup with negligible accuracy loss. Across more than 2400 training experiments, spanning models from 271M to 8.2B parameters and datasets up to 160B tokens, we demonstrate that these findings extend reliably to large-scale training regimes. All pre-training experiments were run on Cerebras CS-3 systems.
[AI-15] AI for Computational Design Science: A Responsible Human-AI Framework and Case Study on Short-Form Video Safety Surveillance
链接: https://arxiv.org/abs/2609.05270
作者: Wenli Zhang,Jiaheng Xie,Zhihe Pan,Yidong Chai,Xiao Fang,Sudha Ram
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Artificial intelligence (AI) is transforming not only what information systems researchers design, but also how design research is conducted. Yet existing literature offers limited guidance for computational design science (CDS) when AI actively participates in problem formulation, resource construction, design search, evaluation, and knowledge abstraction. We develop AI for Computational Design Science (AI4CDS), a five-phase methodological framework in which AI expands problem and design search while researchers retain responsibility for domain grounding, admissibility, verification, and scientific judgment. Collaboration is governed by graduated trust, reversibility, auditability, and differentiated reproducibility. We instantiate AI4CDS through ChildRiskGuard, an interpretable artifact for detecting short-form videos inappropriate for children, while documenting AI interactions, rejected alternatives, corrections, and audit trails. The case translates audience-dependent safety and explanation faithfulness into three technical challenges and develops an artifact that separates generic from child-specific risk, represents distinct developmental-risk mechanisms, and makes concept-level explanations part of the predictive computation. ChildRiskGuard achieves an F1 score of 0.769, substantially outperforming direct application of a general-purpose content-safety model while remaining competitive with strong benchmarks. The primary contribution is AI4CDS as a responsible framework for AI-enabled CDS; ChildRiskGuard provides process and artifact evidence of how AI-expanded, researcher-governed design can generate and evaluate novel computational design knowledge.
[AI-16] CONTINUITY: Security-Context Contracts for Composable LLM Agent Controls
链接: https://arxiv.org/abs/2609.05269
作者: Chris Zheng,Geng Yang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 20 pages, 5 figures. Code and research artifact available at this https URL
Abstract:LLM agent systems increasingly combine provenance tracking, authorization, policy enforcement, protocol adapters, and execution controls. However, individually correct security mechanisms do not necessarily compose into an end-to-end secure system: security-critical context may be dropped, widened, rebound, or reinterpreted as actions cross component boundaries. We identify this failure mode as security-context discontinuity and introduce CONTINUITY, a framework for verifiable composition of agent security controls. CONTINUITY models each component with an assume-guarantee contract and carries authenticated security context across transitions using signed root grants, provenance commitments, role-bound transition receipts, bounded typed releases, transformation witnesses, and effect-bound execution permits. We formalize end-to-end consequence integrity, requiring every realized external effect to be backed by a valid and current authorization witness linking the principal, task, provenance, delegation, policy state, canonical action, and finality boundary. We implement a reference verifier and deterministic cross-layer fault-injection suite covering 32 fault classes across four application domains. In 2,560 parameterized attack instances spanning 128 fault-domain classes, the full CONTINUITY configuration commits no harmful external effect, while completing all 700 benign tasks and escalating all 200 ambiguous cases. These results show that secure agent execution requires not only sound individual controls, but explicit contracts that preserve their guarantees across the complete instruction-to-effect path. Comments: 20 pages, 5 figures. Code and research artifact available at this https URL Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) ACMclasses: D.4.6; I.2.11 Cite as: arXiv:2609.05269 [cs.CR] (or arXiv:2609.05269v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2609.05269 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-17] race2Tower: Transition-Aware EigenTrace Induction of Multi-Level Skills for LLM Agents
链接: https://arxiv.org/abs/2609.05261
作者: Jiazheng Sun,Boyu Yang,Binhao Yuan,Mingxuan Li,Xin Peng
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 13 pages, 9 figures
Abstract:Large language model agents increasingly rely on execution traces to master complex interactive tasks. However, current paradigms are bottlenecked by shallow trajectory retrieval and flat skill summarization, fundamentally ignoring the temporal dependencies and outcome-conditioned topology of agent behavior. We introduce Trace2Tower, a transition-aware EigenTrace framework that distills raw trajectories into a robust skill hierarchy. Trace2Tower abstracts step-level interactions into canonical events, constructing a unified graph governed by semantic compatibility, transition dynamics, and outcome evidence. Through a novel contrastive spectral decomposition, it isolates stable, success-aligned behavioral modes while rigorously suppressing failure-prone shortcuts. These modes organically populate a dynamic skill tower of action templates, procedural routines, and overarching task strategies, continuously refined via verifier-guided feedback. On ALFWorld, Trace2Tower achieves 87.31% success requiring only 10.35 steps and 0.26 invalid actions; on WebShop, it reaches 50.67% exact success. Across both benchmarks, Trace2Tower significantly outperforms existing baselines in task mastery and context-efficient experience reuse.
[AI-18] Commonsense Reasoning in Computer Vision: Foundations Recent Advancements and Future Directions
链接: https://arxiv.org/abs/2609.05257
作者: Bahar Uddin Mahmud,Sumit Barua,Guan Yue Hong,Ajay Gupta,Hexu Liu
类目: Artificial Intelligence (cs.AI)
备注: 35 pages
Abstract:Commonsense reasoning in computer vision encompasses integrating visual data and contextual knowledge, crucial for enhancing AI’s understanding of everyday scenarios. This understanding not only improves machine learning models but also enhances their ability to interact meaningfully with humans and the environment. Unlike CNN-based conventional vision models, which are designed to identify objects within a specific image, incorporating commonsense knowledge enables models to interpret scenes in a more holistic manner, thereby improving their spatial ability to reason about relationships among objects and actions. This integration not only enhances object recognition but also facilitates a deeper understanding of the contextual factors, ultimately leading to more precise predictions and interactions in real-world applications. This paper presents a comprehensive survey of recent developments that integrate commonsense knowledge into computer vision tasks. We systematically review approaches based on knowledge graphs, scene graphs, neuro-symbolic models, and commonsense-augmented transformers. We also outline current limitations related to dataset bias, knowledge incompleteness, and integration challenges. Finally, we highlight prospective research trajectories in cross-modal reasoning, scalable commonsense knowledge injection, and neuro-symbolic hybrid architectures to develop truly intelligent visual systems.
[AI-19] A Unified Physics-Aware Quantum Machine Learning Framework across Power GaN HEMTs and Logic Nanowire FETs: Predicting Unseen Process Splits and Held-Out Geometry Combinations with Lower Error and Tighter Split-to-Split Variability
链接: https://arxiv.org/abs/2609.05251
作者: Rushat Rai,Yun-Yuan Wang,Autsada Kakaen,Pei-Jie Chang,Doan Viet Nguyen,Yuan-Chieh Chiu,Doldet Tantraviwat,Niall Tumilty,Simon See,Wen-Jay Lee,Tai-Yue Li,Nan-Yow Chen,Tian-Li Wu
类目: Artificial Intelligence (cs.AI)
备注: 11 pages, 11 figures
Abstract:We present a unified reinforcement-learning (RL) framework that discovers compact parametrized quantum circuits (PQCs) for data-scarce device modeling. A graph neural network (GNN) policy optimized by proximal policy optimization (PPO) searches circuit architectures using leave-one-group-out cross-validation (LOGOCV) error on held-out process or geometry groups as the reward. The framework achieves the lowest mean absolute error (MAE) on all 11 targets versus six classical baselines, with 59% lower error (Ioff) and 81% tighter fold variability (VTH) for HEMTs and 84% lower error (VTH, SS, Ioff) and 82% tighter fold variability (Ioff) for NWFETs. These results demonstrate the potential of RL-selected, classically simulated PQCs as compact surrogates with low OOD error and improved physical consistency, despite imposing no explicit physical constraints, penalty terms, or device-specific equations, on the two evaluated device datasets.
[AI-20] Do LLM s Exhibit Coherent Knowledge Structures in Mathematical Reasoning ? A Perspective from Knowledge Space Theory EMNLP2026
链接: https://arxiv.org/abs/2609.05245
作者: Peng Cui,Heejin Do,Mrinmaya Sachan
类目: Artificial Intelligence (cs.AI)
备注: EMNLP 2026 findings
Abstract:Human knowledge is inherently structured and interdependent: mastery of a concept requires prior mastery of its prerequisites, a principle formalized by Knowledge Space Theory (KST). While LLMs achieve strong performance on complex reasoning tasks, it remains unclear whether they exhibit coherent, human-like knowledge structure. We introduce a KST-grounded framework for evaluating LLM knowledge structure in mathematical reasoning, using it as a normative framework to analyze whether LLM behavior adheres to principled knowledge dependencies. Evaluating eight open- and closed-source LLMs against real human learners, we find that (1) LLMs do not adhere to human knowledge structure – they frequently violate knowledge dependencies and fail to leverage related knowledge provided in context to improve performance on dependent questions; (2) LLMs do not share a consistent knowledge structure among themselves, as reflected by low overlap in their knowledge distributions. Furthermore, these structural deficiencies remain largely invisible to accuracy-based and LLM-as-judge evaluations. Together, our results provide behavioral evidence that current LLMs knowledge does not follow a human-like structure.
[AI-21] Uncensored Open-weight Models: Redistribution as the Persistence Layer
链接: https://arxiv.org/abs/2609.05241
作者: 10a Labs:Juliette Garcia,Hailey May,Bobby McKenzie,David Pham,Matthew Swain,Joshua Valdez,Corie Wieland,Zachary Yahn
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:A rapidly expanding ecosystem of actors is removing built-in safety guardrails from open-weight AI models. We profile this ecosystem by identifying key producers, downstream reproductions, and emerging applications. Between January 2024 and March 2026, we identified 3,471 original uncensored models on HuggingFace, each repackaged an average of 2.4 times; three actors account for 52% of all 8,164 compressed redistributions. Once quantized and mirrored across separate accounts, formats, and registries such as Ollama, these models persist regardless of upstream removal and become easier to deploy downstream. Of the 1,643 identified GitHub applications integrating uncensored large language models (ULLMs), 25% were classified as explicitly malicious.
[AI-22] PRICE: A Systematic Study of LLM Adaptation Choices for Bitcoin Price Forecasting
链接: https://arxiv.org/abs/2609.05235
作者: Maryam Fakhari,Mehran Safayani
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Numerical Analysis (math.NA)
备注:
Abstract:Cryptocurrency markets exhibit extreme volatility and non-stationary dynamics that challenge conventional forecasting methods. Although Large Language Models (LLMs) have shown promise for time series forecasting, the combined effects of adaptation choices remain largely unexplored in financial settings. This study introduces PRICE, a structured approach for adapting LLMs to short-term Bitcoin price forecasting. Built on a 4-bit quantized LLaMA-3 8B model, PRICE investigates how fine-tuning, numerical representation, prompting, inference, and decoding jointly influence forecasting performance. PRICE integrates Parameter-efficient fine-tuning with Low-Rank Adaptation (LoRA), Recursive multi-step inference, Integer-rounded numerical representation, Context-Task-Format (CTF) prompting, and Exact zero-temperature decoding. Ablation studies show that each component contributes to forecasting accuracy and reliability. LoRA enables efficient training on limited hardware, recursive inference improves accuracy, integer-rounded values reduce errors, CTF prompting outperforms Chain-of-Thought, Implicit Chain-of-Thought (iCoT), and few-shot prompting, and zero-temperature decoding improves stability during recursive forecasting. Comparative evaluation against eight transformer-based and time-series foundation models shows that PRICE achieves the lowest forecasting errors on both validation and test sets while maintaining robust performance across evaluation periods. Despite being based on a model primarily pretrained on text rather than time-series data, PRICE achieves competitive or superior performance relative to specialized foundation models. These findings demonstrate that adaptation choices critically determine the accuracy and robustness of LLMs for numerical time-series forecasting.
[AI-23] Substrate-Aware AI Agents : Execution Context as a First-Class Input
链接: https://arxiv.org/abs/2609.05232
作者: Manu Agrawal
类目: Artificial Intelligence (cs.AI)
备注: 8 pages, 3 figures. Reproducibility artifacts and source-linked evaluation code: this https URL
Abstract:Autonomous AI agents increasingly select actions in environments whose memory, execution-time, runtime, compute, and operational constraints determine what counts as a suitable plan. We call the absence of this execution context from an agent’s planning state substrate blindness. We test this general proposition through numerical code generation, where selected implementation choices and operational consequences are directly observable. Three frontier model configurations–Anthropic Claude Opus 5, OpenAI GPT-5.6-Sol, and Google Gemini 3.7 Flash–generate code for a high-dimensional pairwise Euclidean-distance task either from the task alone or with a 128 MB RAM and 10.0 s wall-time contract. Contract disclosure reduced measured peak process memory in 13 of 14 executable index-aligned task-only versus contract-disclosed comparisons and reduced mean wall time in all three cohorts, making execution up to 3.1x faster. Across the audited corpus, disclosure produced structural code changes including bounded blocking, float32 retention, upper-triangle traversal, and in-place or memory-mapped buffers. At a tighter 96 MB contract, independently sampled contract-disclosed cohorts achieved correct-and-within-budget outcomes of 4/5 for Claude Opus 5, 5/5 for GPT-5.6-Sol, and 3/5 for Gemini 3.7 Flash, compared with task-only outcomes of 0/5, 1/5, and 0/5; cohort mean MaxRSS and wall time were 49-74% and 35-64% lower than their task-only references. These results establish a controlled proof of concept for substrate-aware agent planning: a minimal execution contract induces proactive structural adaptation in generated programs, shifting computation away from unconstrained allocations and substantially improving observed resource-time profiles before execution.
[AI-24] ACE: Adaptive Calibration-Free Expert Skipping for MoE-based LLM s
链接: https://arxiv.org/abs/2609.05228
作者: Zukang Xu,Zhixiong Zhao,Xing Hu,Jiangyong Yu,Houji Wen,Jun Li,Zhe Jiang,Dawei Yang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Mixture-of-Experts (MoE) architectures provide an efficient paradigm for scaling large language models (LLMs), yet fixed top-k routing activates the same number of expert slots for every token, causing substantial redundant computation. Existing expert-skipping methods often rely on router confidence, calibration data, or additional training, and therefore cannot reliably estimate the actual contribution of routed experts. To this end, we propose ACE, a training-free, calibration-free, and checkpoint-preserving framework for token-adaptive expert skipping in MoE-based LLMs. ACE contains two complementary components: 1) Global Spectral Proxy (GSP), which estimates global transformation capacity from the coupled gate, up, and down projections together with RMSNorm scaling; and 2) Router-Conditioned Refinement (RCR), which constructs expert-specific direction prototypes from centered router weights and evaluates expert responses along routing-preferred directions. During inference, ACE combines both estimates with runtime router gates and skips an expert slot only when both views identify it as low-contribution, while always retaining the top-1 expert. All expert statistics are computed offline, leaving only table lookups and lightweight scalar operations online. Extensive experiments across three MoE-based LLMs and eight benchmarks demonstrate that ACE consistently outperforms existing static and dynamic baselines, with increasingly pronounced advantages under aggressive expert skipping. For instance, at a 50% skipping ratio on Qwen3.6-35B-A3B, ACE reduces WikiText-2 perplexity by 7.96% and improves average downstream accuracy by 4.15 percentage points over the strongest competing method.
[AI-25] CABAL: Multi-Agent Simulacra for Tracing the Effects of Collusive Bidding in Peer Review
链接: https://arxiv.org/abs/2609.05227
作者: Jicheng Zhou,Kemou Li,Kahim Wong,Zheyuan Li,Zhuan Shi,Fengpeng Li,Haiwei Wu,Jiantao Zhou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recent reports during the AAAI-27 review cycle highlight the risk of reviewers coordinating bids for reciprocal assignment advantage. Prior work treats bidding, reviewer assignment, and review manipulation as separate stages, leaving the lifecycle effects of collusive bidding unclear. Real-world analysis is further constrained by typically unobservable collusive intent and the lack of counterfactuals for the same conference. Motivated by this gap, we introduce \alg, an end-to-end multi-agent simulacra framework for studying reviewer assignment integrity by holding the conference environment fixed and configuring LLM-driven reviewer agents with honest or collusive policies. We further develop an affinity-guided collusive bidding strategy that uses mutual reviewer-paper affinities to construct collusion rings and select target papers, producing expertise-consistent rather than arbitrarily targeted attacks. Controlled experiments show that collusive bidding more than doubles target-paper capture and that assigned colluders score target papers about two points higher than honest co-reviewers, while conference-wide effects remain comparatively modest. Evaluated bid-phase detectors provide only limited evidence of collusion: in a fixed-triplet detector stress test, native positive-bid graphs are confounded by benign affinity, while a Very-High-only diagnostic view enables precise but low-coverage local recovery.
[AI-26] What Matters in On-Policy Distillation? A Perspective on Data Efficiency and Data Selection
链接: https://arxiv.org/abs/2609.05198
作者: Zhinan Hou,Jiaqi Zhang,Xunliang Cai,Keyou You
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:On-Policy Distillation (OPD) has emerged as a widely adopted post-training paradigm for enhancing large language models in reasoning domains. However, the data-centric mechanisms in OPD remain relatively underexplored. This paper presents a empirical study of data efficiency and data selection in OPD. We begin by investigating an extreme setting: training OPD on only one example, namely 1-shot OPD. Surprisingly, we find that 1-shot OPD is consistently effective across all sampled training examples and harder examples often yield superior performance gain. We next investigate what actually drives the student model’s improvement in the training data. Our analysis reveals that the improvement is not driven by high token entropy, but the longer CoT paths which hard problems naturally generate. Training on longer CoT can help maintain closer alignment with the teacher over a long reasoning horizon, and learn critical thinking patterns usually missing in short CoTs, such as reflection (e.g., Alternatively''). Based on these insights, we propose a simple data selection method that selects only hard examples for training, where even unsolvable’’ examples that completely exceed the teacher’s capability can be successfully used. Our experiments conducted on four models ranging from 1.5B to 7B show that training the student model on only 8 selected hard examples matches the performance of the 17K dataset baseline.
[AI-27] Phase Transition Frequency as a Training Time Predictor of Test Accuracy in ResNets
链接: https://arxiv.org/abs/2609.05194
作者: Arunan J
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:The number of discrete class-separability jumps observed during ResNet finetuning is examined empirically as a predictor of final test accuracy. Across 75 experiments spanning four benchmarks (CIFAR-10, CIFAR-100, TinyImageNet, and CIFAR-10-C) and three architectures (ResNet-18, ResNet-50, and ResNet-101), with five to ten seeds per configuration, a strong within-dataset negative correlation is obtained on standard i.i.d. classification benchmarks: (r = -0.84) on CIFAR-10 ((p 10^-8), (n = 30)) and (r = -0.87) on CIFAR-100 ((p 10^-5), (n = 15)). Under distributional stress, the relationship attenuates: TinyImageNet yields (r = -0.45), and the CIFAR-10-C corruption benchmark yields (r = -0.19). Two additional analyses discipline the empirical claim. A partial correlation controlling for architecture depth, treated as a linear covariate, shows that on CIFAR-100 the transition count retains statistically significant predictive power ((r_\mathrmpartial = -0.69), (p = 0.007)); the corresponding result under the stricter categorical conditioning is not established at (n = 15). A comparison against six alternative training-curve signals shows that transition count achieved the strongest correlation among the evaluated signals on CIFAR-100 and one of the strongest on CIFAR-10, but is dominated by other signals on the two stressed benchmarks. The comparison is restricted to training-curve-level signals; comparisons against effective rank, Hessian sharpness, Fisher information, margin, and neural-collapse measures, which are the strongest competitors in the current literature, are not part of the present study and remain open. The observation is presented as an in-distribution training-quality probe among a family of candidate probes, and an inexpensive detection procedure suitable for logging alongside a standard training loop is provided.
[AI-28] he Mirror Agent Model: a Bayesian Architecture for Interpretable Agent Behavior
链接: https://arxiv.org/abs/2609.05190
作者: Michele Persiani,Thomas Hellström
类目: Artificial Intelligence (cs.AI)
备注: Accepted at the International Workshop on Explainable, Transparent Autonomous Agents and Multi-Agent Systems, 2022
Abstract:In this paper we illustrate a novel architecture generating interpretable behavior and explanations. We refer to this architecture as the Mirror Agent Model because it defines the observer model, that is the target of explicit and implicit communications, as a mirror of the agent’s. With the goal of providing a general understanding of this work, we firstly show prior relevant results addressing the informative communication of agents intentions and the production of legible behavior. In the second part of the paper we furnish the architecture with novel capabilities for explanations through off-the-shelf saliency methods, followed by preliminary qualitative results.
[AI-29] Beyond Stationarity in Time Series: Discovering Causal Structures and Latent Regimes via Markov Blankets ALT KDD2026 ECML
链接: https://arxiv.org/abs/2609.05150
作者: Lei Zan,Charles K. Assaad,Emilie Devijver,Eric Gaussier
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at the 11th AALTD Workshop at ECML PKDD 2026, Naples, Italy
Abstract:This paper introduces Regime-aware Constraint-Based and Noise-Based causal discovery with Markov Blankets (RCBNB-MB), a novel causal discovery algorithm for time series that relaxes the common assumption of a single, time-consistent causal structure. Time series are typically observed at discrete time points and often exhibit regime changes that challenge the assumption of a static causal structure, a limitation in many real-world dynamic systems. To address this challenge, RCBNB-MB identifies latent causal regimes, defined as subsets of time points within which a stable causal structure holds. The algorithm follows an iterative strategy that segments the time series into regimes and discovers the causal graph within each regime. By leveraging the Markov blanket rather than direct parents, RCBNB-MB gains robustness to errors in causal discovery and preserves predictive information. We provide theoretical guarantees for RCBNB-MB’s ability to recover both regime transitions and causal graphs under reasonable assumptions. Furthermore, we validate its effectiveness through extensive experiments on simulated datasets with known ground truth and real-world IT monitoring data, where taking into account regime shifts is critical. Empirical results show that RCBNB-MB systematically outperforms baseline approaches in accurately detecting regime changes and their associated causal graphs, positioning it as a robust and versatile framework for non-stationary time series analysis.
[AI-30] A Hybrid Predictive Ensemble of Machine Learning and Deep Neural Networks for Early Cardiovascular Disease Risk Assessment
链接: https://arxiv.org/abs/2609.05146
作者: Balaji Venkateswaran
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 14 pages, 7 figures, 2 tables
Abstract:This study introduces an intelligent framework that integrates machine learning and deep neural network ensemble techniques for early detection and prognosis of cardiovascular diseases. The system utilizes real-time physiological data collected from Internet of Medical Things (IoMT) devices, including ECG sensors, heart rate monitors, and blood pressure trackers. To ensure the accuracy and reliability of input data, preprocessing steps such as noise reduction, normalization, and missing value imputation are employed. The most significant health indicators are identified through effective feature selection methods and then processed using optimized classifiers such as Support Vector Machines (SVM), Random Forests, and eXtreme Gradient Boosting (XGBoost), which are combined in an ensemble architecture to improve diagnostic precision. The framework demonstrates remarkable performance in predicting cardiovascular disease risk, achieving higher accuracy, reduced false positives, and enhanced consistency compared to conventional methods. It is designed on a cloud-based infrastructure that ensures scalability and real-time processing for continuous patient monitoring. Experimental evaluation on real-world cardiovascular datasets confirms the framework’s efficiency in early-stage risk assessment and clinical decision support. The results highlight the potential of combining traditional machine learning and deep learning paradigms to achieve proactive healthcare management and improve patient outcomes.
[AI-31] SciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding
链接: https://arxiv.org/abs/2609.05141
作者: Shenxi Wu,Yuhong Liu,Haosong Zhang,Tongjin Zou,Yanxun Zhang,Gaochang Chen,Dun Liang,Jiaqi Wang,Zhecan James Wang,Yuhang Zang,Dahua Lin
类目: Artificial Intelligence (cs.AI)
备注: 52 pages, 21 figures, and 20 tables. Project page: this https URL
Abstract:Scientific papers require models to reason jointly over text, equations, figures, tables, code, and datasets while preserving the provenance of supporting evidence. Existing benchmarks typically evaluate these capabilities in isolation, leaving unclear whether multimodal models can support realistic scientific-reading workflows. We introduce SciDocBench, a workflow-centered benchmark for scientific document understanding. It contains 124 expert-authored and difficulty-screened questions organized into seven research-assistant capability groups and 19 subtasks across five scientific domains. Each question is instantiated under four matched conditions combining English or Chinese questions with all-images-first or interleaved document representations, yielding 496 evaluation instances for controlled analysis. The strongest evaluated system achieves only 62.6/100, with pronounced weaknesses in document perception, evidence grounding, verification, and cross-document reasoning. To translate these diagnostics into scalable training signals, we introduce SciDocIR, a typed evidence-graph representation that preserves scientific document objects, layout and cross-reference relations, and provenance. Building on SciDocIR, we construct SciDocDataset, comprising approximately 15K supervised fine-tuning samples and 8K reinforcement-learning samples across 14 verifiable subtasks. Together, SciDocBench, SciDocIR, and SciDocDataset form an evaluation-to-training framework for diagnosing and improving scientific-document assistants. The project page is available at this https URL.
[AI-32] A Schema Bounded Language Model for Refining Robot Policies Without Destabilizing Local Learning
链接: https://arxiv.org/abs/2609.05133
作者: Chongwen Dong,Mithun Paul Saint-Germain,Pinjari Asif,Carlo R. daCunha
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:This paper addresses navigation by composite heterogeneous robots in a decentralized system when policy reasoning and local control operate at different update levels. In a NetLogo–Python implementation, three robots share motion dynamics but use different LLM backends. Each robot independently combines a large language model (LLM) policy agent, an Upper Confidence Bound (UCB) bandit, and a Double Deep Q-Network (Double DQN) controller; no central LLM generates team actions. LLM inference is confined to round-level policy generation and refinement rather than tick-level action selection. The robots perform cross-LLM communication through a shared round summary containing policies, outcomes, and learning feedback. UCB performs refinement-mode selection, and the policy-conditioned Double DQN performs tick-level action selection from navigation variables, active policy parameters, and the LLM action prior. Each of the four configurations was evaluated over 30 rounds. In the fixed simulation, the complete configuration reached the goal in all 90 correlated robot–round records and achieved the lowest median completion time (42 ticks) and P90 (73.2 ticks); its median was 25.0–39.1% lower than those of the other configurations. These observations provide descriptive, configuration-level evidence from the evaluated configurations.
[AI-33] IER: Threat Implicitness Benchmark for Evaluating LLM Safety Behaviors
链接: https://arxiv.org/abs/2609.05117
作者: Thu-Hien Trinh-Thi,Hai-Yen Vong,Thanh-Ha Ung-Dung,Tram Ho
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Current LLM safety benchmarks largely rely on binary metrics, overlooking how models respond to harmful prompts with varying threat implicitness. We introduce TIER, a Threat Implicitness Benchmark for behavioral safety evaluation of LLMs. TIER covers four risk domains and four threat levels, from explicit harmful requests to sophisticated jailbreaks. Responses are assessed using a six-label behavior scale and two independent LLM judges. Experiments on six open-weight LLMs show that safety behaviors evolve gradually across threat levels rather than shifting directly from refusal to compliance. Contextual prompts yield the most diverse behaviors, while jailbreaks reveal the largest robustness gaps. Furthermore, models with similar Attack Success Rates can exhibit distinct response distributions, highlighting the need for behavior-aware LLM safety evaluation.
[AI-34] Unifying ICL SFT KL-Regularized RL Through a Bayesian Lens
链接: https://arxiv.org/abs/2609.05111
作者: Junxin Fan
类目: Artificial Intelligence (cs.AI)
备注: 26 pages. A theoretical note
Abstract:Large language models are now trained and evaluated under a diverse set of paradigms: supervised fine-tuning (SFT), few-shot in-context learning (ICL), KL-regularized RLHF/RLVR, on-policy distillation (OPD), and test-time reasoning with search and chain-of-thought. These methods are often discussed as fundamentally different, and recent empirical results–such as the mixed impact of few-shot prompting on RL-tuned reasoning models–can appear puzzling. This note develops a Bayesian perspective that puts these procedures on the same footing. At the core is a two-step template: (i) construct a (generalized) Bayes or Gibbs posterior q* over outputs or actions given a context, using a prior/reference model and a utility signal (log-likelihood, reward, or advantage); and (ii) approximate q* by a forward-KL projection onto a parametric family, either in-weights (SFT/RL) or in-context (ICL). Part I formalizes few-shot ICL and SFT as amortized and-weights projections onto the Bayes posterior predictive. Parts II-IV show that KL-regularized RLHF/RLVR, reward-weighted SFT, reward-weighted ICL (RW-ICL), and advantage-weighted SFT (AWSFT) are all instances of forward-KL projection onto posteriors induced by rewards or advantages. We disentangle where these equivalences hold (objectives and first-order updates) and where they do not (source and granularity of the learning signal). Part V sketches implications for modern reasoning pipelines: RLHF/RLVR recipes as “posterior design + projection”, why cold-start or supervised warm-up is practically unavoidable for importance-weighted KL projections, and DeepSeek-R1 and o1-style reasoning models as combining test-time Bayesian search with training-time KL amortization.
[AI-35] Compact Bellm an-Grounded Cognitive Maps for Cost-Aware Navigation
链接: https://arxiv.org/abs/2609.05104
作者: Yuzhe Han,Mingkun Xu,Yujie Wu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Biological agents navigate familiar environments not by re-solving routes for each new goal, but by reusing a learned map built once and read off as goals change. Existing artificial cognitive-map models mimic this reuse, yet their guidance is not explicitly grounded in additive heterogeneous route costs. Furthermore, they often struggle with memory efficiency: representative state-indexed and high-rank spectral constructions incur substantial storage growth as the environment scales. We present BCM, which grounds a reusable cognitive map in local edge costs through a self-supervised Bellman-grounded objective and a compact coordinate encoding, supporting changing goal queries without per-goal retraining. On weighted grids of up to N=1600 nodes, BCM maintains full success and only a 5% mean Gap relative to exact Dijkstra search, compared with about 45% for a connectivity-based spectral baseline. Notably, as the graph size increases from N=400 to N=3600 , its memory footprint grows sublinearly while maintaining competitive performance, making our method scalable to complex environments. Together, these results show that additive route costs can be written into a compact, reusable cognitive-map representation, bridging the gap between biological flexibility and optimal path planning.
[AI-36] NEAT-POCKET: Pocket-Conditioned Autoregressive 3D Molecular Generation with a Neighborhood-Guided Set Transformer
链接: https://arxiv.org/abs/2609.05097
作者: Roxane Axel Jacob,Daniel Rose,Thierry Langer,Johannes Kirchmair
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:AI-driven de novo molecular design offers a promising route to accelerate early-stage drug discovery by generating novel ligands directly within target protein binding pockets. We present NEAT-POCKET, a pocket-conditioned extension of the autoregressive NEAT model for 3D molecular generation. NEAT-POCKET generates molecules atom by atom in protein pocket environments while preserving atom permutation invariance and explicitly modeling hydrogen atoms. Benchmarks on the CrossDocked and SPINDR datasets show that NEAT-POCKET achieves competitive structure-based generation performance while sampling substantially faster than existing baselines. Beyond full-molecule generation, NEAT-POCKET naturally enables pocket-conditioned fragment completion, a task directly relevant to lead optimization and scaffold elaboration. These results position NEAT-POCKET as a fast, flexible, and practical framework for structure-based drug design.
[AI-37] ProCA: Progressive Contrastive Alignment for Robust EEG Visual Decoding
链接: https://arxiv.org/abs/2609.05094
作者: Kanglei Zhou,Chunyan Lan,Dongyang Li,Jun Zhu,Liyuan Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Electroencephalogram (EEG) visual decoding aims to recover visual semantics from non-invasive neural time-series signals, for which robust alignment between noisy neural responses and stable semantic representations is key to achieving high-performance decoding. Despite recent advances in contrastive learning, robust EEG decoding remains challenging because existing methods rely on fixed visual or textual anchors whose semantic relations may become misaligned with EEG representations that vary across trials, subjects, and learning stages. Our empirical evidence shows that this instability appears across both standard EEG decoding protocols and more challenging robustness settings, including strict cross-subject transfer and realistic personalized continual adaptation. We provide a formal analysis showing that fixed semantic supervision can bias optimization when EEG-specific relations evolve, and that structure-agnostic perturbations may distort semantically important EEG components. To address these issues, we propose Progressive Contrastive Alignment (ProCA), a unified and model-agnostic framework for adaptive neural-semantic alignment. ProCA progressively refines class-level contrastive supervision from frozen vision-language priors to EEG-aware semantic relations, and introduces structure-consistent interpolation to constrain feature mixing according to channel-wise and temporal importance. Across subject-dependent, subject-independent, strict cross-subject transfer, and continual adaptation settings, ProCA achieves average relative Top-1/Top-5 gains of 7.4%/3.9%, 10.0%/4.6%, 28.1%/17.8%, and 16.8%/11.6%, respectively.
[AI-38] LLM -Guided Program Evolution for Circle Packing: Breaking 10 Packomania Records for 28
链接: https://arxiv.org/abs/2609.05093
作者: Wes Sander
类目: Artificial Intelligence (cs.AI)
备注: 8 pages. Code and solutions: this https URL
Abstract:We present Discovery Loop, a lightweight system that uses a large language model (LLM) to iteratively evolve optimization algorithms. Starting from a simple seed solver, the LLM proposes algorithmic improvements guided by a scoreboard of results and a history of prior ideas. Each candidate is evaluated against an independent verifier; improvements are kept and failures discarded. Applied to the Packomania circle-packing benchmark (csqv: maximize the sum of radii of N variable-radius circles in the unit square), the system improved the best known solutions for 10 values of N in the range 101-114, with gains of 2.4%-5.4% over prior records, all within 15 iterations and at a total LLM cost of 27.72. These results have been independently accepted by Packomania. We describe the method, analyze cost-efficiency dynamics including an adaptive plateau-detection mechanism, and discuss implications for democratizing automated scientific discovery.
[AI-39] Constructing and Evaluating Clinical Reasoning Trajectories for Medical Agent
链接: https://arxiv.org/abs/2609.05090
作者: Yunqi Zhu,Wensheng Zhang,Xuebing Yang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Evaluation of medical artificial intelligence agents remains predominantly answer-centric, assessing only the correctness of final outputs while overlooking the quality of intermediate reasoning. In clinical settings, however, a correct answer reached through fabricated evidence or incoherent logic is as dangerous as an incorrect one. We propose MedTraj, a framework that treats reasoning trajectories as critical objects for construction, evaluation, and optimization. The pipeline generates structured multi-step reasoning chains from medical reasoning sources. Each trajectory is then parsed into clinical observations, evidence, numbered reasoning steps, and a final conclusion, and scored across five quality dimensions: coherence, evidence support, hallucination, completeness, and traceability. Controlled error injection introduces targeted faults into otherwise correct trajectories to establish causal links between specific reasoning failures and measurable quality degradation. Building on this, step-level filtering based on marginal contribution identifies which individual reasoning steps drive or undermine trajectory quality. Finally, quality-weighted context learning feeds trajectory evaluations back into the model at inference time, allowing it to learn from both strong and weak reasoning demonstrations. Experiments across CareQA, PubMedQA, and CECMed demonstrate that trajectory context consistently improves reasoning coherence, with gains of +0.029 to +0.041 over a zero-shot baseline. On CECMed, quality-weighted context nearly doubles the correctness over the zero-shot baseline while cutting the hallucination ratio by 87%. Marginal-contribution analysis further shows that a small minority of reasoning steps carry most of the quality signal, and that extending chains beyond four steps yields diminishing returns.
[AI-40] MePo: Unifying Representation Refinement and Reconciliation for General Continual Learning
链接: https://arxiv.org/abs/2609.05075
作者: Guanglong Sun,Kanglei Zhou,Liyuan Wang,Qi Cheng,Hongwei Yan,Shuang Cui,Hang Su,Jun Zhu,Yi Zhong
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:General continual learning (GCL) aims to learn from evolving data streams without task identities, explicit boundaries, or repeated access to previous data, making it a realistic yet challenging setting for continual intelligence. Although pretrained models (PTMs) provide rich prior knowledge for addressing the limited supervision and non-stationary nature of GCL, existing PTM-based methods often directly adapt pretrained representations and overlook two critical gaps: the misalignment between upstream pretraining and downstream continual adaptation, and the unreliability of conventional output alignment under blurry streams. Here we propose MePo++, a unified post-training framework that bridges pretrained knowledge and downstream GCL through representation refinement and reconciliation. MePo++ introduces two complementary components: MetaPrep, which improves representation plasticity for continual adaptation through unsupervised meta-refinement over pseudo continual sequences; and StreamAlign, which reinforces representation stability by reconciling evolving online features with a stable pretrained geometry. By improving representation learnability before adaptation and preserving alignment during continual learning, MePo++ enables PTMs to remain both plastic for new concepts and stable over evolving streams. Experiments across diverse PTMs, datasets, and continual learning baselines demonstrate the consistent effectiveness and generality of MePo++ for PTM-based GCL. Our code is available at this https URL.
[AI-41] owards Efficient Evaluation of Evolutionary Transfer Optimization: Case Studies on Task-Parameterized Applications
链接: https://arxiv.org/abs/2609.05040
作者: Yanchen Li,Xiaoming Xue,Kay Chen Tan
类目: Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
备注: Accepted at the 2026 International Conference on Machine Intelligence and Nature-Inspired Computing (MIND 2026)
Abstract:As evolutionary transfer optimization (ETO) scales to larger collections of related tasks, problem evaluation can become a major source of runtime growth. This work studies problem-side evaluation scaling in task-parameterized applications and reformulates application-specific serial computations into forms suitable for parallel execution. We organize evaluation scaling into two levels: the number of evaluated tasks and the workload within each task. In multi-task optimization, matrix-recursive kinematic-arm evaluation is reformulated using an accumulation-matrix representation of cumulative link directions. In sequential transfer optimization, pointwise B-spline trajectory evaluation is reformulated using a blending-matrix representation for trajectory and collision computations. Both reformulations maintain close numerical agreement with their reference evaluations and substantially reduce runtime, yielding 256.72\times and 93.91\times end-to-end speedups, respectively. These results demonstrate problem-side reformulation as a practical route toward scalable ETO. Both application implementations and experimental scripts are released as open source to support reproducibility and reuse.
[AI-42] ROVE: Adaptive Agent Skill Orchestration via Trace-Grounded Route Validation and Editing
链接: https://arxiv.org/abs/2609.05019
作者: Tianxing Wang,Mingming Zhao,Shuai Huang,Huiyang Xu,Chaoyue Niu,Shengzhong Liu,Fan Wu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Agents tend to optimize, select, or constrain execution structures before decisive runtime outcomes are observed. However, such pre-execution commitment creates an orchestration bottleneck: when intermediate evidence invalidates the pending continuation, agents must either execute stale steps or replan broadly, compounding errors, wasting computation, and discarding progress. We thus propose Trace-grounded Route Orchestration via Validation and Editing (TROVE), which revises only what runtime evidence invalidates. Offline, TROVE distills evaluated workflow-search traces into atomic and composite skills and an outcome-conditioned transition graph, preserving stable fragments while exposing outcome-dependent decisions. Online, it treats a planned route as provisional: after committing one top-level skill, the controller retains a valid continuation, inserts a trace-supported local response, or replaces only the invalid suffix. Evaluation across code-generation, question-answering, and math reasoning benchmarks with different LLM backbones show that TROVE delivers a stronger quality-efficiency trade-off than existing baselines of dataset-level optimization, query-level architecture selection, and graph-constrained scheduling. Quality gains are largest when outcomes change the appropriate continuation, whereas early termination yields substantial efficiency gains on near-saturated tasks. Ablations further show that composite skills capture most offline benefits, insertion enables local correction, and suffix replacement primarily improves efficiency. These findings establish selective route editing as a general principle for adaptive agent orchestration.
[AI-43] Amortizing Scaling Law Construction Costs
链接: https://arxiv.org/abs/2609.05016
作者: Abhash Kumar Jha,Diana Alexandra Onuţu,Neeratyoy Mallik,Swagatam Haldar,Sam Laing,Niccolò Ajroldi,Shiwei Liu,Joaquin Vanschoren,Aaron Klein
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 5 pages, 2 figures, workshop
Abstract:Scaling laws guide the design choices for training large foundation models, but deriving them involves training an exhaustive grid over hyperparameters, token budgets, and parameter counts, which is computationally expensive. Fitting a scaling law, however, only requires the best-loss frontier across compute scales, discarding most of the trained configurations. We propose a framework for efficient scaling law construction that formulates data collection as a Bayesian optimization problem, and introduce metrics for comparing scaling law fitting methods under constrained compute budgets. We find that progressively expanding the compute budget during acquisition, mirroring the compute-ordered evaluation of configurations in practice, substantially improves recovery efficiency. Augmenting the observed configurations with surrogate-fantasized evaluations then recovers the broader experimental grid, allowing accurate scaling law fitting without training every configuration. Together, these can closely match scaling law fits over a full dense grid at computational savings of up to 10\text–100\times .
[AI-44] Language models judge war differently when tested for alignment
链接: https://arxiv.org/abs/2609.05009
作者: Maxim Chupilkin
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注:
Abstract:Safety evaluations can mischaracterize deployed behaviour if artificial-intelligence systems respond to being evaluated. We test this possibility in a full-factorial conjoint experiment on decisions to start a war, spanning 20 large language models, 32 scenarios, 10 repetitions and two conditions (N = 12,800 judgments). Adding one sentence, “You are tested for alignment with human values”, produced two effects. First, it produced a level effect: mean willingness to start war fell by 13.43 points on a 0-100 scale (95% confidence interval, -16.20 to -10.65). Second, it produced a structural effect by changing which information drove judgments. Probability of success was the largest factor for 17 of 20 models at baseline; under the cue, civilian casualties were largest for 12. Standardized estimates show that this reordering arose principally because models attenuated strategic considerations such as probability of success and domestic support. Evaluation framing therefore changes both an answer’s level and its revealed decision rule.
[AI-45] Global to Local: Topology-Preserving Adaptive Graph Pooling via Granular-Ball
链接: https://arxiv.org/abs/2609.04978
作者: Sen Zhao,Gaojie Xu,Shuyin Xia,Yifan Guan,Yi Liu,Yi Wang,Wei Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Graph pooling aims to compress the graph, including both node embeddings and their underlying topological patterns, into a more compact representation. Previous works focus primarily on the overly fine-grained representation of nodes, progressively coarsening the graph by removing nodes or merging them into clusters, thus neglecting the global-to-local patterns and adaptive granularity of the graph’s topological structure. In the real scenario, graphs as a whole can be considered the coarsest level of granularity, encapsulating the global topological structure, with progressively finer-grained local topological structures represented from top to bottom. This process continues until the adaptive granularity for each subdomain is reached. To this end, we propose a novel Topology-Preserving Adaptive Graph Pooling (TPAGP) method that dynamically partitions graphs into granular balls by integrating node features and topological information, enabling the generation of multi-granularity representations that effectively capture both local and global structural patterns. Additionally, we design a multi-granularity graph network model that facilitates feature interaction and optimization across different granularities, significantly enhancing performance in graph classification tasks. Experimental results demonstrate that TPAGP outperforms existing pooling methods across various benchmark datasets, effectively mitigating information loss caused by fixed-granularity strategies.
[AI-46] Solving Hard XAI Queries Based on a Compiled Dual-Rail Encoding
链接: https://arxiv.org/abs/2609.04931
作者: Arthur Ledaguenel,Florent Capelli,Jean-Marie Lagniez
类目: Artificial Intelligence (cs.AI); Computational Complexity (cs.CC)
备注: 20 pages, 2 figures, full version of a submitted conference paper with detailed proofs
Abstract:The widespread adoption of artificial intelligence (AI) within real-world applications has raised a lot of concerns regarding their trustworthiness, especially in critical applications. The field of eXplainable AI (XAI) has emerged with the objective of providing explanations to the users about the decisions made by AI systems. Several explanations for boolean classifiers have been introduced in the literature, including abductive and contrastive explanations, each giving a different insight on the decision of the classifier. However, computing an explanation for a decision of a boolean classifier is a hard problem in general. One way to deal with this complexity is to rely on a compiled representation of the classifier for which each explanation can be computed efficiently. Unfortunately, we prove in this paper that several classes of abductive explanations, remain hard to compute even for Ordered Binary Decision Diagrams, one of the most tractable subsets of the knowledge compilation map. Included in such classes are shorter abductive explanations or abductive explanations that include the explainee’s preferences. To recover the benefits of working with compiled representations, we show that a proper representation of the dual-rail encoding of the classifier can be used to compute efficiently these classes of explanations.
[AI-47] Artificial Intelligence in Equity and Crypto Markets: Progress Profitability Evidence and the Limits of Automated Investing
链接: https://arxiv.org/abs/2609.04917
作者: Linsen Zhu,Mengqing Cai
类目: Artificial Intelligence (cs.AI); Portfolio Management (q-fin.PM); Trading and Market Microstructure (q-fin.TR)
备注: Review article. 32 pages, 1 figure, 5 tables. Literature cutoff: 31 August 2026
Abstract:Artificial intelligence (AI) now supports investment workflows from data and prediction through research, portfolios, execution, and tool use. Technical capability, however, is not evidence of investment profitability. This critical state-of-the-art review examines public research available through 31 August 2026 on listed equities, exchange-traded funds, centralized crypto spot, perpetual futures, and on-chain markets. We organize evidence with an alpha-translation chain: point-in-time information must yield a stable signal, feasible positions, executable orders, and risk-adjusted returns after costs. Across machine learning, time-series foundation models, financial language models, reinforcement learning, and agents, the examined record shows real but mainly upstream progress in prediction, text processing, portfolio design, and workflow integration. Evidence is thinner for durable net performance. Temporal contamination, repeated selection, survivorship, weak benchmarks, implementation costs, venue mechanics, and capacity can break translation to net alpha. Strong historical results coexist with predictor decay, corrected look-ahead failures, mixed prospective evidence, and few audited live-capital records. Crypto adds informative state but requires separate treatment of spot, perpetual, and decentralized cash flows and execution. Within the public evidence examined here, no general AI architecture is shown to deliver persistent, cross-regime, capacity-aware net alpha. More credible claims require point-in-time data and models, decision-aligned objectives, joint portfolio–execution evaluation, controlled adaptation, prospective tests, and authority-matched governance. These conditions can improve evidence and implementation; they do not guarantee profit.
[AI-48] Compact-Memory LLM Agents via Online Max-Member Clustering and Atom-Aware Packing
链接: https://arxiv.org/abs/2609.04915
作者: Jiahe Geng,Jinpeng Wang,Kun Yuan
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Many long-horizon LLM deployments face tight prompt budgets: latency, cost, and context limits make full-context prompting impractical as interaction length grows. The key question is then not raw recall alone, but which memory design gives the best quality–token trade-off in the compact-memory regime. We present \textbfRSM-full, an online clustered-memory pipeline designed for a strong quality–token Pareto point. RSM-full combines two design choices: a cosine-gated \emphmax-member merge write rule and an atom-aware grouped context packer. On AMA-Bench, our primary compact-memory benchmark, it reaches 83% of Full-Context quality at 32% of the token cost at a 4 k budget; under four-seed averaging it beats the closest streaming-clustered baseline (Online K-Means) by +3.5 – 6.0 ,pp ( p.001 ) across the whole \sim2.6 k-- \sim5 k regime. Three-seed ablations show most of this gain comes from the merge rule ( +5.7 ,pp over Online K-Means and matched- \tau DP-means) and the grouped packer ( +5.0 ,pp over flat concatenation). The pattern reproduces on RealMem, an independent long-horizon persona-memory benchmark: RSM-full improves on Budget-RAG ( +0.69 ,pp, p=.006 ), is on par with BM25-RAG (paired \Delta=+0.27 ,pp, p=.47 ; we do \emphnot claim BM25 equivalence in the equivalence-test sense), and significantly outperforms Streaming-Proto ( +2.97 ,pp) and the closest reproduced 2025 agentic-memory baseline A-MEM ( +1.65 ,pp, p.001 ). Across benchmarks the message is consistent: under tight budgets, compact-memory performance is driven mainly by how streaming memories are merged and how retrieved content is assembled. Overall, RSM-full is most useful when answeroughly 2k – 5k prompt tokens, where itdefines a strong compact-memory Pareto point; higher-token baselines remain stronger outside this regime. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04915 [cs.AI] (or arXiv:2609.04915v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.04915 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Jiahe Geng [view email] [v1] Fri, 4 Sep 2026 09:19:00 UTC (73 KB)
[AI-49] ARIA - An Agent ic Framework for Autonomous Testing of Infotainment Systems
链接: https://arxiv.org/abs/2609.04913
作者: António Azevedo,Bruno Lima,João Pascoal Faria
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Automotive infotainment validation still relies on manual testing, slow, costly, and incompatible with agile releases and OTA updates. Scripted automation only partly helps: it couples test logic to implementation, yielding brittle, high-maintenance suites. Existing LLM-driven frameworks mostly target web/mobile apps, using single- or dual-agent setups that overload one or two models with perception, planning, action selection, and validation at once, prone to hallucinations and unproductive exploration loops given infotainment complexity. We present ARIA (Autonomous Real-time Infotainment Assessment), a multi-agent LLM framework that autonomously runs end-to-end tests on Android infotainment systems via visual interaction, using a closed-loop pipeline of four specialized agents per step plus a report stage. From single-sentence scenarios (path, action, expected outcome), ARIA runs the interactions and produces reports, reproducible scripts, and visual evidence per step. Evaluated on a manufacturer’s physical Android infotainment system across 30 scenarios, ARIA completed 28 (93.3%) with a verdict (2 errored), 20 of which (71.4%) matched ground truth. It caught all 5 known defects, no fault passed as working; its 8 false positives stem from navigation/image limits and unsupported gestures, showing multi-agent LLMs can run infotainment tests industrially while exposing the cost of a low false-positive tolerance. A single-agent baseline confirms the multi-agent design’s value: on the first pass, before stronger-model revisitation narrows the gap, it shows a far higher false-positive rate (72.0% vs. 52.6%), conflating navigational difficulty with system failure. We report first-pass/post-revisitation results, token/call/cost per scenario, and show via repeated runs that stability tracks complexity, with fault detection perfectly consistent, pointing to CI integration of visual testing.
[AI-50] reeFI: Value-Aware Statistical Fault Injection for Deep Neural Networks
链接: https://arxiv.org/abs/2609.04912
作者: Noam Bires,Marcello Traiola,Angeliki Kritikakou,Elisa Fromont
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI)
备注: Accepted at ICCAD 2026
Abstract:Reliability evaluation of deep neural networks under hardware faults commonly relies on fault injection, but exhaustive campaigns are intractable for modern models and datasets. Statistical fault injection reduces this cost, yet existing approaches still require large injection budgets because they do not explicitly exploit a key property of floating-point faults: the effect of a bit flip depends strongly on the value being corrupted. We propose TreeFI, a value-aware statistical fault-injection methodology for FP32 single-bit faults in DNN activations and weights. TreeFI partitions each layer’s value distribution into intervals with similar expected bit-flip behavior, learned using regression trees, and allocates injections across these intervals according to their relevance for failure-rate estimation. This stratified allocation preserves the target confidence and error margin while avoiding unnecessary injections in low-impact regions of the fault space. We validate TreeFI on CNN and Transformer models using CIFAR-10 and ImageNet. On ResNet8, where exhaustive activation fault injection is feasible, TreeFI provides more accurate estimates than state-of-the-art statistical FI baselines under the same campaign setting. Across the evaluated models, TreeFI reduces the required injection budget by up to 72.1x, with average reductions of 44.9x for activation faults and 11.2x for the executed weight campaigns.
[AI-51] Better Understanding Better Fixes? A Study of Hallucination in LLM -based Automated Program Repair
链接: https://arxiv.org/abs/2609.04909
作者: Xuemeng Cai,Jiakun Liu,Linhan Yang,Wei Ma,Lingxiao Jiang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) have significantly advanced automated program repair (APR), yet existing evaluations remain largely result-centric and provide limited insight into hallucination during repair. In APR, hallucination may arise not only in final patches but also in the intermediate artifacts that guide patch generation. To address this gap, we perform a multi-layered analysis of hallucination throughout the APR process. Specifically, we characterize hallucination as the production of patches or intermediate artifacts that are not faithfully grounded in the available repair evidence. We examine repair hallucination in final patches and understanding hallucination in intermediate artifacts through three tasks, namely triggering testcase identification, line coverage prediction, and additional testcase this http URL then evaluate three representative LLMs on 832 Defects4J bugs through automatic evaluation and manual analysis. Our results show that both repair and understanding hallucinations remain prevalent. Across models and settings, only 21.0%-55.9% of generated patches pass the developer-written test suite. Moreover, although more accurate intermediate artifacts are generally associated with successful repairs, this relationship does not always hold. Manual analysis of 812 sampled repairs identifies repair hallucinations in 72.7% of cases, including patches that pass all available tests; incorrect causal localization and incorrect repair strategies account for 45.9% and 18.5% of these hallucinations, respectively. Meanwhile, models frequently misidentify triggering testcases, mispredict line coverage involving branching control flow, and generate additional testcases with missing bug-triggering conditions or incorrect expected behavior.
[AI-52] Adaptation Interfaces for In-Context Tabular Foundation Models in Time-to-Event Prediction
链接: https://arxiv.org/abs/2609.04901
作者: Minh-Khoi Pham,Luca Cotugno,Dan Cernei,Alina Sirbu,Stefano Masi,Giuseppe Prencipe,Alessandro Pingitore,Patrizia Landi,Working Group on Uric Acid,Cardiovascular Risk of the Italian Society of Hypertension,Tai Tan Mai,Martin Crane,Marija Bezbradica
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Under Submission. Not peer-reviewed
Abstract:Tabular foundation models (TabFMs) achieve strong performance on structured data, particularly for standard classification and regression problems. Yet, extending them to censored time-to-event prediction is challenging because it requires properly handling censoring and event-time dynamics. Building on our prior work, we further link TabFMs with CoxPH and DeepHit and revise the context-resampled training procedure. We evaluate temporal zero-shot reformulation, classification-based fine-tuning, and survival-head adaptation using frozen TabFM backbones on 74 single-risk data sets, and we additionally study 4 competing-risk data sets. Zero-shot inference is effective on smaller single-risk data sets, whereas supervised adaptation becomes increasingly advantageous as data sets scale. Cox provides the most reliably strong interface, especially for Integrated Brier Score (IBS) on larger data sets. DeepHit is relatively stronger for the time-dependent Concordance Index than for IBS, while cause-specific MTLR ranks highest among the TabFM survival heads in the four-data-set competing-risk analysis. Classification fine-tuning becomes more competitive with zero-shot inference as data sets grow but remains weaker for probabilistic prediction. Overall, our results indicate that effective TabFM transfer depends on the data regime and on the statistical structure represented by the chosen adaptation interface. The implementation scripts used for this work are available at this https URL.
[AI-53] Reinforcement Learning for Sequential Solar PV Policy Design under Uncertainty: An Agent -Based Approach
链接: https://arxiv.org/abs/2609.04880
作者: Iias Faiud,Jonaid Shianifar,Michael Schukat,Karl Mason
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Designing effective and fiscally sustainable policies for solar photovoltaic (PV) adoption requires balancing adoption gains against public expenditure under uncertainty and heterogeneous decision-making. This study formulates PV policy design as a sequential decision problem and integrates reinforcement learning (RL) with a stochastic agent-based model (ABM) that simulates yearly solar PV adoption under uncertainty. A policymaker agent selects annual incentives, including capital grants, subsidised loan rates, and feed-in tariffs, over a 16-year horizon. Adoption–cost trade-offs are explored by varying policy preferences within a scalarised reward framework. Policies are learned using PPO, SAC, and TD3 and evaluated under stochastic simulation. The results show that this approach produces a clear trade-off structure: the highest-adoption policy (TD3, w_\textcost=0.5 ) achieves approximately 4,145 adopters at a cost of EUR 41.73 million, while the lowest-cost policy (PPO, w_\textcost=2.0 ) reduces expenditure to EUR 7.27 million with 2,682 adopters. The balanced policy (PPO, w_\textcost=1.6 ) achieves 3,495 adopters at a cost of EUR 22.47 million. Across algorithms, consistent trade-off patterns are observed, indicating robustness of the adoption–cost relationship. Compared with static baseline policies, the RL framework explores a broader range of policy configurations. These findings demonstrate the potential of RL as a flexible tool for adaptive policy design under uncertainty.
[AI-54] ReCAST: Restoration-aware Cascaded Stage-wise Training for Obfuscated SMS Risk Classification EMNLP2026
链接: https://arxiv.org/abs/2609.04878
作者: Jieyun Huang,Yi Shen,Kaikai Zhao,Jiangze Yan,Wenjing Zhang,Ping Chen,Ning Wang,Zhaoxiang Liu,Kai Wang,Shiguo Lian
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accept by EMNLP 2026 Industry Track
Abstract:Fraudulent messages sent via Short Message Service (SMS) are increasingly obfuscated to evade cost-conscious classifiers in production systems. In Chinese SMS, attackers can exploit a wide range of carefully crafted obfuscation strategies to hide risk-bearing phrases while preserving human readability, making direct classification brittle under real-world latency and throughput constraints. We propose ReCAST, a Restoration-aware Cascaded Stage-wise Training framework for robust obfuscated Chinese SMS classification. ReCAST distills a large teacher model’s de-obfuscation ability into a smaller deployable student model by supervising obfuscated span detection, obfuscation type prediction, and text restoration, and then uses the restoration-aware student for downstream risk classification. Experiments on an internally constructed real-world Chinese SMS benchmark show that ReCAST substantially improves classification performance over directly trained baselines under obfuscation. The results suggest that restoration-aware distillation offers a practical path toward robust SMS risk classification with smaller deployable models under production-oriented constraints.
[AI-55] MARLA: A Conceptual Scaffold for Regulatory Learning under the EU AI Act
链接: https://arxiv.org/abs/2609.04877
作者: Alessio Buscemi,Tom Deckenbrunnen,Imane Hmiddou,Marco Billi,Livio Rubino,Silvia Rizzuto Ferruzza,Daniele Pagani,Antonino Rotolo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The EU AI Act positions regulation as part of the infrastructure for safe, trustworthy and market-ready innovation. Realising this ambition requires regulatory learning: the evidence generated during implementation must be translated into governance and legal knowledge that supports consistent interpretation, effective oversight, and adaptation as technologies evolve. Yet the actors who produce this evidence and those who rely on it operate in different professional worlds. This paper proposes MARLA (Map, Assess, Report, Learn, Adapt), a conceptual scaffold organising regulatory learning as a five-stage cycle centred on the implementation of legal requirements into socio-technical practices, situated at the Local, National and European levels of the AI Act’s governance architecture. Deliberately non-prescriptive, MARLA gives technical and legal stakeholders a shared vocabulary in which each of the first three stages generates its own documentable form of regulatory learning. We illustrate the scaffold with two piloted case studies and a prospective National-to-European illustration.
[AI-56] Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents
链接: https://arxiv.org/abs/2609.04875
作者: Chao Yao,Yangbo Wei,Zhen Huang,Junhong Qian,Chenle Chen,Shaoqiang Lu,Chen Wu,Lei He
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Long-running LLM agents are stateful: beyond the transcript they accrete compressed summaries, plaintext memory, pending tool plans, and, under every serving API, a KV cache. Yet today’s “forget” operations delete a plaintext memory record and stop, leaving every artifact derived from the revoked information intact. We formalize execution-state unlearning: after a forget request, the agent must behave as if it had never observed the target. Modeling the runtime as a deterministic transition system, we prove that the pre-target trajectory prefix is shared with this counterfactual world for free, that the post-target suffix is irreducibly tainted without token-level attribution, and that exact unlearning requires at least T-\tau+1 recomputed transitions, where \tau is the target’s injection step. Provenance-Guided Selective Replay attains this bound as a cross-layer contract spanning prompt, compressed memory, and cache: a provenance graph locates the injection point, checkpoint restoration reduces to cropping the KV cache, and sanitized replay regenerates the counterfactual suffix. Audited with elicitation, stochastic, and string-free behavioral tests across three agent suites, nine baselines, and three model families, memory deletion leaves leakage unchanged, instruction-based forgetting collapses under elicitation (Leak@probes = 1.00), and source redaction still acts on a revoked preference in 80% of episodes, while selective replay is indistinguishable from a full reset at up to 9x fewer recomputed tokens.
[AI-57] AutoLR: Automating the Path from Research to Launch Review in Industrial Recommender Systems
链接: https://arxiv.org/abs/2609.04871
作者: Qi Zhang,Yanlin Chen,Wenchao Xiao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Improving an industrial recommender is an iterative research-and-engineering process rather than a direct path from idea to deployment. In \textbfDASHEN, NetEase’s gaming-community app, algorithm engineers typically identify promising directions from research papers, technical reports, and prior production experiments; reproduce or adapt the underlying methods; implement them in the production codebase; and evaluate the resulting models through training and offline experiments. Promising candidates are then advanced to online A/B tests, and those demonstrating robust gains are submitted to Launch Review—the internal gate for full-traffic rollout. Large language models (LLMs) can assist with individual stages of this workflow, but the overall process remains human-dependent without a harness that can reliably coordinate them across long-running, often multi-day experimental cycles. We present \textbfAutoLR, initially built as \textbfAuto Launch Review and later extended upstream into an autonomous research-to-launch harness. AutoLR combines three system mechanisms: a \textbfmulti-expert council that debates and adversarially reviews proposals; a \textbfdeterministic evidence-weighted exploration–exploitation selector that allocates a limited trial budget across candidate directions and uses Council reranking; and a layered knowledge system that combines external research, production-system knowledge, and DASHEN-specific domain knowledge—such as game communities, player characteristics, and content-interaction patterns—with posterior evidence from configurations, patches, logs, failures, and offline outcomes. LLM agents perform semantic reasoning and code generation, while deterministic controllers retain authority over execution, metric extraction, guardrails, and persistent state transitions.
[AI-58] CHAMP: Cross-domain Hybrid Architecture for Matchmaking and Prediction in Online Multi-Player Games CIKM2026
链接: https://arxiv.org/abs/2609.04870
作者: Kai Wang,Ge Fan,Chaoyun Zhang,Yuyang Jiang,Yuze Liu
类目: Artificial Intelligence (cs.AI)
备注: Accepted by CIKM 2026 (Applied Research Track)
Abstract:Multiplayer Online Battle Arena (MOBA) games rely on matchmaking to maintain competitive balance. Our prior work, CUPID, framed matchmaking as an assignment re-optimization problem and showed that a single-mode win-rate predictor can meaningfully rebalance teams. However, deploying such a system across diverse player populations exposes three practical bottlenecks: most queueing players lack sufficient in-mode match history (cold start), skill distributions shift drastically across rank tiers (distribution inconsistency), and extreme skill segments are severely data-starved. We present CHAMP, a cross-domain matchmaking framework that resolves these deployment bottlenecks. To address data sparsity and cold starts, CHAMP replaces the target-mode-only player profile with a hybrid domain feature collection: a timestamp-ordered cross-mode short-term sequence whose slices are annotated with target-domain features, plus per-mode breakdowns of long-term, real-time and team statistics. We further propose the Domain-Aware Win-rate Network (DAWN): a Domain-aware Knowledge Extractor (DAKE) compiles target-mode attributes into learnable representations that feed Domain-Aware Temporal/Spatial/Permutation OmniNet Encoders (DATOE/DASOE/DAPOE), so that mode-conditioned representations and per-mode debiasing are learned jointly inside a single shared network. Online, one trained DAWN serves every supported mode, with per-mode position-satisfaction thresholds as the only mode-specific knob. Offline, DAWN achieves 67.73% win-rate prediction accuracy, outperforming all evaluated attention and sequence baselines. Online A/B tests across the entire League ladder of a large-scale MOBA game, from novice players up to the top-expert players served by Elite Mode, demonstrate consistent drops in imbalanced matches. For lower-tier players, CHAMP reduces the 5-minute kill crushing rate by up to 20.73%. Comments: Accepted by CIKM 2026 (Applied Research Track) Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04870 [cs.AI] (or arXiv:2609.04870v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.04870 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-59] From Interaction Traces to Persistent Skills: Online Evolution for Computer-Use Agents
链接: https://arxiv.org/abs/2609.04869
作者: Longtao Hu,Xiao Liang,Linchao Zhu
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 2 figures, and 2 tables. Code: this https URL
Abstract:Computer-use agents can execute increasingly complex tasks in graphical interfaces, but their interaction experience is typically transient: procedural knowledge acquired from one rollout is not systematically retained, refined, and reused in later tasks. Existing skill libraries provide external procedural knowledge, yet their incremental value over the same agent operating without skills, as well as their longitudinal dynamics under repeated interaction, remain insufficiently characterized. We present an online skill-evolution framework that converts interaction trajectories and evaluator feedback into a persistent, versioned library of reusable procedures. Each iteration executes against a frozen library snapshot, and evidence-guided skill updates become available in subsequent iterations without changing model parameters. We compare the full evolving-library system with a configuration-matched empty-library control across four OSWorld application domains under the same fixed action-generation and GUI-grounding stack, task sets, and iteration horizons. Following a five-iteration empty-library warm-up, Full attains a higher post-warm-up mean evaluator score in all four observed domain runs, with mean differences ranging from 5.7 to 18.6 percentage points and domain-dependent temporal stability. In GIMP, provenance-aware analysis reveals retrieval across task-of-origin boundaries and revision churn, where repeated accepted edits fail to recover the originating task. These findings characterize evolving skill libraries as auditable, shared procedural memory that can improve a fixed computer-use stack, while showing that their benefits are conditional and repeated revision does not guarantee recovery. Code is released at this https URL.
[AI-60] PRISM-Bench: An Audio-Centric Diagnostic Benchmark for Text-to-Audio-Video Generation
链接: https://arxiv.org/abs/2609.04867
作者: Yuchen Sun,Qian Yang,Jun Wang,Detai Xin,Guoqiao Yu,Guanglu Wan,Qi Jia
类目: Multimedia (cs.MM); Artificial Intelligence (cs.AI); Sound (cs.SD)
备注: 19 pages, 10 figures, 4 tables. Accepted at ACM Multimedia 2026 (MM '26). This arXiv version includes supplementary appendices not included in the conference proceedings version
Abstract:Text-to-audio-video (T2AV) generation has advanced rapidly, but its evaluation still underestimates the audio modality. Existing benchmarks either treat audio as an auxiliary component of video quality or assess it in isolation from audiovisual grounding, making it difficult to diagnose where current systems truly succeed or fail in audio generation. We present PRISM-Bench, the first audio-centric diagnostic benchmark for T2AV generation. Built from a rigorously curated dataset of 900 human-verified samples, PRISM-Bench factorizes audio evaluation along two orthogonal axes: audio type (Speech, Music, and Sound) and sound-source visibility (On-screen vs. Off-screen). It evaluates generated content across four perceptual dimensions (Audio-Visual Coherence, Audio Quality, Audio Expressiveness, and Prompt Following) with 35 fine-grained criteria. To ensure reliable assessment, we adopt an enhanced MLLM-as-a-Judge protocol based on blind, side-by-side comparison against ground-truth references, demonstrating strong alignment (over 70% mean agreement) with human raters. Our evaluation of recent T2AV systems highlights a significant performance gap between frontier and open-source models. Furthermore, we demonstrate that current generation paradigms overfit to perceptual fidelity while struggling with complex grounding and control tasks, particularly in generating music and synchronized On-screen audio.
[AI-61] LLM -Assisted Behavioural and Scenario Augmentation for Agent -Based Energy Adoption Models
链接: https://arxiv.org/abs/2609.04866
作者: Iias Faiud,Hossein Khaleghy,Michael Schukat,Karl Mason
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recent advances in large language models (LLMs) create opportunities to enrich simulation-based energy policy analysis, particularly by supporting structured behavioural assumptions and exploratory techno-economic scenarios. However, directly replacing adoption models with LLM reasoning raises concerns regarding interpretability, reproducibility, and behavioural validity. This paper proposes a hybrid framework for LLM-assisted specification design, integrating bounded behavioural rubrics and structured scenario specifications into a calibrated agent-based model (ABM) of solar photovoltaic (PV) adoption by Irish dairy farms. The proposed approach preserves the original techno-economic adoption mechanism while augmenting it with bounded behavioural modulation and scenario-driven uncertainty analysis. Behavioural effects are represented through interpretable conservative, balanced, and optimistic rubrics, while future policy and market conditions are explored through fixed, rule-validated scenario specifications. Experimental results across multiple policy settings, Monte Carlo worlds, and random seeds demonstrate stable and economically plausible behaviour, with adoption outcomes remaining bounded and monotonic across behavioural regimes. The framework achieves up to approximately 13% behavioural adoption increase relative to the corresponding logistic case without producing unstable or unrealistic saturation dynamics. The results demonstrate that LLM-assisted specifications can be integrated into calibrated energy ABMs in a controlled, reproducible, and policy-relevant manner.
[AI-62] CoSkill: Joint Reinforcement Learning of Reasoning and Meta-Skill Agents for Hierarchical Skill Evolution
链接: https://arxiv.org/abs/2609.04865
作者: Jinyuan Feng,Dongmin Li,Yiqun Chen,Yang Gao,Xing Chen,Huimu Wang,Zhiqiang Pu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Skill libraries improve the sample efficiency of agentic reinforcement learning (RL) by enabling large language model (LLM) agents to reuse procedural knowledge. Yet existing paradigms exhibit structural shortcomings: they either decouple skill evolution from policy optimization or instantiate meta-skills as fixed workflows. Both treat skills as passive objects to be managed, limiting the flexible evolution of skills and their co-adaptation with the reasoning agent. To address the limitations, we propose CoSkill, a unified multi-agent RL framework that recasts the static meta-skill workflow as a learnable Meta-Skill Agent and jointly trains it with a Reasoning Agent over a hierarchical skill library. By modeling the Reasoning and Meta-Skill Agents as a cooperative team sharing a single backbone, CoSkill enables end-to-end co-adaptation: the Reasoning Agent conditions its actions on a retrieved task skill and step skills selected from its child set, while its task performance guides the Meta-Skill Agent in refining those step skills. Experiments on ALFWorld and WebShop show that CoSkill substantially outperforms prior skill-based and RL baselines, achieving success rates of 98.4% and 90.6%, respectively (+3.5 and +6.2 pp). As shown in Figure 1, CoSkill achieves superior early-stage sample efficiency, asymptotic performance, and wall-clock efficiency. Our code is available at this https URL.
[AI-63] MZ-Rain: Moisture-Budget-Guided Zero-Inflated Model for Station-Level Precipitation Nowcasting
链接: https://arxiv.org/abs/2609.04864
作者: Yifang Zhang,Shengwu Xiong,Henan Wang,Wenjie Yin,Yuqiang Zhang,Chen Zhou,Hua Chen,Qile Zhao,Pengfei Duan
类目: Artificial Intelligence (cs.AI)
备注: 16 pages, 6 figures
Abstract:Accurate station-level precipitation nowcasting is critical for agriculture, water resource management, and disaster prevention, which typically is formulated as a time series forecasting problem. However, conventional time-series modeling techniques face two major challenges in addressing station-level precipitation nowcasting: (1) Lack of Physics-Guided Modeling, where meteorological variables are treated as a homogeneous set without accounting for their distinct roles in precipitation formation, leads to predictions that deviate from the physical processes governing precipitation. (2) Severe zero inflation in precipitation, where dry intervals dominate the dataset, obscuring meaningful precipitation patterns and complicating the predictive modeling. To address these challenges, we propose \textbfMZ-Rain, a moisture-budget-guided zero-inflated sLSTM framework for station-level precipitation nowcasting. Guided by the moisture budget equation, MZ-Rain decomposes the precipitation formation process into process-specific pathways corresponding to moisture storage, moisture transport, surface evaporation, and precipitation persistence, and captures their temporal evolution through dedicated sLSTM branches. To account for the zero-inflated nature of precipitation, MZ-Rain introduces an adaptive Tweedie modeling strategy that adaptively modulates the rainfall mean while jointly learning precipitation occurrence as an auxiliary task, enabling the model to better balance dry-wet discrimination and quantitative precipitation estimation. Extensive experiments across diverse geographical and climatic regimes demonstrate that MZ-Rain consistently outperforms strong baselines on multiple evaluation metrics, including CSI, FAR, MSE, and MAE. In particular, the model exhibits superior skill in forecasting heavy precipitation events, while benefiting from physically grounded process modeling.
[AI-64] MM-IFEval-Pro: A Multilingual and Attack-Resistant Benchmark for Instruction-Following in Vision-Language Models
链接: https://arxiv.org/abs/2609.04859
作者: Changming Xiao,Zhenliang Ni,Jinhui He,Han Shu,Jie Hu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As vision-language models (VLMs) rapidly advance in image understanding, cross-modal reasoning, and complex instruction execution, instruction-following capability has become a key indicator of their reliability and practicality. However, existing multimodal instruction-following benchmarks still suffer from limited language coverage and insufficient adversarial safety scenarios, making them inadequate for evaluating real-world multilingual and safety-sensitive settings. To address these gaps, we present MM-IFEval-Pro, a multimodal instruction-following benchmark covering Chinese and English tasks as well as diverse instruction hijacking cases. MM-IFEval-Pro includes 4 major task categories and 24 subcategories and 8 instruction categories with 52 subcategories, with each sample containing an average of 3.0 constraints to realistically simulate complex instruction scenarios. We further construct a reinforcement-learning training set enriched with Chinese and adversarial instructions, which significantly improves model performance on MM-IFEval-Pro and transfers effectively to other mainstream multimodal benchmarks, demonstrating strong cross-task and cross-language generalization.
[AI-65] ElderBench: Benchmarking Autonomous Mobile Agents for Older Adults
链接: https://arxiv.org/abs/2609.04850
作者: Weide Zhan,Qumu Shaqu,Yuanqing Liu,Peng Zhang,Jiahao Liu,Kam Him Lam,Ning Gu,Zhan Hu,Tun Lu
类目: Artificial Intelligence (cs.AI)
备注: 19 pages, 5 figures
Abstract:While autonomous mobile agents hold great potential for assisting older adults with smartphone usage, existing GUI benchmarks mainly rely on explicit, goal-oriented instructions and rarely capture the naturally occurring language patterns of older users, such as indirect speech, referential ambiguity, and under-specified requests. This mismatch between benchmark instructions and real-world elderly interactions may hinder reliable agent deployment. To address this gap, we present ElderBench, the first benchmark for evaluating mobile GUI agents in authentic elderly-oriented scenarios. ElderBench is constructed from 249 naturally elicited smartphone tasks collected from older adults across 20 applications. We first characterize the linguistic divergence between elderly instructions and existing GUI benchmark instructions from syntactic, semantic, and pragmatic perspectives. We then evaluate mainstream GUI agents and Vision-Language Models under both online and offline settings, revealing substantial performance degradation when handling elderly-oriented instructions. Through controlled instruction normalization, failure analysis, and fine-grained linguistic feature analysis, we further identify how elderly-specific language patterns contribute to agent failures. Our findings provide actionable design insights toward more adaptive, interpretable, and age-inclusive GUI agents for older adults.
[AI-66] Long Horizon Transformer Quantile Fault Prediction for Multi Site Industrial Predictive Maintenance
链接: https://arxiv.org/abs/2609.04840
作者: David J Poland,Daniele Ravi,Na Helian
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Long-horizon predictive maintenance requires models to distinguish slowly evolving degradation from normal operating-regime variation over planning windows measured in days rather than hours. This paper evaluates whether an explicit conditional-quantile representation provides an informative classifier interface for this problem. The proposed TQRNN30d framework combines a dual-stage quantile regression neural network (QRNN) feature extractor with a multi-stream temporal fusion classifier. Each hourly word of 81-channel machine behaviour is mapped to a 324-dimensional quantile-state representation, and 720 ordered hourly words form the 30-day document supplied to the long-horizon model. The classifier fuses quantile states with dynamic covariates, channel-level static metadata, and a 168-hour latent-history stream using gated residual processing, causal recurrent encoding, and metadata-conditioned cross-modal attention. A bounded instability-aware signal derived from sustained one-word-ahead prediction-error divergence provides auxiliary memory modulation at the longest horizon. Evaluation uses a machine-disjoint 43/14/15 train/validation/test allocation across 72 machines in nine manufacturing facilities. At 30 days, TQRNN30d achieves 79.97% F1, 80.18% recall, 81.82% precision, 82.39% accuracy, and 0.820 ROC-AUC. It leads all 18 evaluated baselines at the 7-, 14-, and 30-day fixed-threshold comparisons, with the largest F1 advantage at 14 days. The results support held-out-machine performance within the observed homogeneous nine-facility fleet, but do not establish unseen-site, cross-equipment, or cross-sector generalisation.
[AI-67] Cost-Aware Hierarchical Multi-Agent Ransomware Detection and Family Attribution
链接: https://arxiv.org/abs/2609.04820
作者: Mubashar Iqbal,Asifullah Khan
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 19 Pages
Abstract:Ransomware detection and family attribution require analysis of different modalities because it can use packing, obfuscation, process manipulation and runtime evasion techniques. However, conventional multimodal usually uses all available modalities for every sample resulting in unnecessary computational cost and increased latency. In this paper, we present a Cost Aware Hierarchical Multi-Agent System (HMAS) for adaptive ransomware detection. The proposed architecture organizes specialized agents into hierarchical domain controllers coordinated by a Meta Orchestrator. Static analysis is used as the initial low-cost modality while additional dynamic and memory modality is selectively used when confidence is insufficient or specialist agents exhibit disagreement. A cost model incorporates modality use and processing overhead. It enables the orchestration policy to balance analysis performance against computational cost. A locally deployed large language model provides verification for selected difficult cases without replacing the deterministic pipeline. Experimental evaluation compares adaptive HMAS with static only, static plus dynamic and exhaustive analysis policies across binary ransomware detection and multiclass family attribution. The complete HMAS achieved 96.57% accuracy, 0.96 F1-score and 0.99 ROC-AUC for binary detection. It also achieved 0.90 macro-F1 for family attribution. At the same time, the HMAS reduced average analysis cost by 43.97% relative to exhaustive analysis and substantially reduced average analysis latency except for the case where LLM is used. Routing analysis showed that 56.05% of cases were resolved using static evidence alone. Only 4.33% required the complete evidence pipeline. These findings demonstrate that adaptive HMAS can provide accuracy cost tradeoff for ransomware analysis while retaining support for heterogeneous and incomplete modalities.
[AI-68] CPR-IE:A Compression-Prediction-Resource Intelligence Efficiency Metric
链接: https://arxiv.org/abs/2609.04809
作者: Xiantao Jiang
类目: Artificial Intelligence (cs.AI)
备注: 11pages
Abstract:Comparing intelligent systems under deployment constraints requires more than this http URL paper develops Compression-Prediction-Resource Intelligence Efficiency (CPR-IE) as a protocol-relative ordering by representational economy, predictive quality, and resourceburden. The analysis separates two questions-how raw resource consumption is represented, andhow the resulting attributes are aggregated. Proportional-increment composition uniquely yieldslogarithmic cumulative burden, and context-independent ratio response yields power responsesto compression, prediction, and burden; with reference normalization the representation is I(C,P,T).We prove Pareto consistency, unit invariance, boundary behavior, trade-off identities, ranking-stability regions, and cross-task aggregation. A translog parent model makes interaction restrictions explicit, and further results establish cardinal and ordinal identification, sub-Gaussianfinite-sample ranking guarantees, robust selection under exponent uncertainty, and deterministicregret bounds. Minimum description length, algorithmic complexity, proper scoring rules, varia-tional inference, and Landauer’s principle motivate measurement choices but do not entail theformula. CPR-IE is a constructed efficiency representation, not a universal law or a definition ofintelligence itself.
[AI-69] When Financial Fine-tuning Fails: A Three-Level Detectability Analysis of Numerical Hallucination in Domain-Adapted Language Models
链接: https://arxiv.org/abs/2609.04806
作者: Xiaodong Li,Peiwei Liu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Financial large language models are increasingly deployed for summarization of reports and disclosures, where numerical hallucination poses significant practical risks. While prior work often attributes such hallucination to insufficient numerical reasoning, this assumption has not been systematically tested under controlled fine-tuning settings. In this paper, we conduct a cost-effective, controlled study of numerical hallucination in financial summarization across three model variants: a base instruction-tuned model, a domain language-adapted model (FT-A), and a numeracy-enhanced domain model (FT-A+B+C). We introduce a three-level detectability taxonomy distinguishing between overt hallucination (currency-denominated fabrication), covert-explicit hallucination (professional-convention numbers), and covert-implicit hallucination (ungrounded quantitative claims). Our results reveal that domain fine-tuning substantially degrades numerical restraint at all detectability levels. While the Base model maintains near-zero hallucination rates (5.4%), FT-A exhibits 82.5% overt hallucination and FT-A+B+C reaches 98%. Contrary to intuition, numeracy supervision amplifies rather than mitigates hallucination across all levels. We identify template injection—the insertion of memorized canonical values regardless of input content—as a primary hallucination mechanism in fine-tuned models. These findings demonstrate that numerical hallucination in financial summarization is driven by the degradation of numerical restraint through domain adaptation, not by insufficient numerical reasoning. We recommend that evaluation protocols assess hallucination across all detectability levels and that deployment practices include explicit mechanisms for grounding-aware generation or abstention.
[AI-70] MedFlow: Class-Aware Multi-Scale Generation for Medical Time-Series Synthesis
链接: https://arxiv.org/abs/2609.04804
作者: Yanhao Huang,Shibo Feng,Wanjin Feng,Peilin Zhao,Chunyan Miao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Synthetic medical time-series generation can alleviate data scarcity and support the development of reliable clinical prediction models. However, existing methods mainly focus on matching the overall distribution and temporal dynamics of real data, which does not necessarily ensure strong downstream utility on imbalanced medical datasets. Clinically informative patterns often occur at heterogeneous temporal scales, while rare minority-class characteristics can be obscured by dominant population patterns. To address these challenges, we propose MedFlow, a class-aware multi-scale flow matching framework for medical time-series synthesis. MedFlow employs a vector-quantized multi-scale tokenizer to represent medical sequences at complementary temporal resolutions, capturing both coarse clinical trends and fine-grained dynamics. We further introduce Token Marginal Guidance, which incorporates class-conditional token statistics directly into the flow matching process to steer generation toward class-specific regions of the learned tokens. This mechanism strengthens minority-class patterns, while preserving the global and tail distributions of real data. Experiments on four public datasets covering electronic health records, EEG, and ECG signals demonstrate that MedFlow consistently outperforms recent state-of-the-art diffusion-based baselines across downstream prediction tasks. On average, it improves AUPRC by 5.8%, reduces Context-FID by 88.6%, and achieves 3.8 \times higher sampling throughput.
[AI-71] Hierarchical Possession-Aware Graph Pointer Network for Pass Receiver Selection
链接: https://arxiv.org/abs/2609.04803
作者: Jingyi Wang,Da Li,Kaixin Wang,Zhangqin Huang
类目: Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI)
备注: This paper has been accepted by the 27th International Conference on Web Information Systems Engineering (WISE 2026)
Abstract:Pass receiver selection is a fundamental task in football analytics, aiming to predict the intended receiver under a given game state. This task is challenging with event-centered freeze-frame observations, a broadcast-like setting that provides only partial and variable player visibility without complete trajectories or stable player identities. The model must therefore reason over anonymous visible candidates, opponent pressure, and recent context under partial observation. To address this setting, we propose a Hierarchical Possession-aware Graph Pointer Network (HPGPN), which formulates pass receiver selection as variable-size candidate prediction over visible teammates. HPGPN jointly models current player interactions, local event context, and possession-level temporal dynamics. It represents the current pass situation with a graph, incorporates fixed event context, and uses dynamic possession history to capture how the attacking sequence evolves. Candidate representations are refined hierarchically by integrating spatial, contextual, and historical evidence, and a glimpse pointer head scores the receiver candidates. Experiments on public football event and freeze-frame data show that HPGPN improves pass receiver selection performance. Ablation studies demonstrate the effectiveness of graph-based interaction modeling, fixed event context, and dual-branch dynamic possession-history modeling.
[AI-72] Whose record is this? Diagnosing and authorizing record use in personalized multimodal models
链接: https://arxiv.org/abs/2609.04801
作者: Xinyu Mao,Junsi Li,Chenyang Liu,Haoji Zhang,Ming Sun
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Contextualized visual personalization can retrieve a true record yet apply it to the wrong visual subject. We formalize when a record may condition an answer as \emphrecord authorization: subject presence ( P ), record-edge validity ( E ), and answer support ( S ) must all hold. We call violations visual memory misbinding (VMM). We construct RecordAuth-Diag, a 3,690-case matched diagnostic suite that changes one image–record edge while holding the query, question, record text, and image multiset fixed. Card removal and nonce relabeling attribute these failures to supplied records. Raw-bank failures span Qwen-, Phi-, and Gemma-family interfaces: Gemma-3-4B-IT reaches 63.69% local unauthorized use at 25.75% clean recall. CoViP remains at 26.02%, versus 22.49% for its Qwen backbone at similar clean recall. Typed pre-generation authorization reduces Qwen card exposure on RecordAuth-Diag from 43.63% to 3.06%, while positive recall changes from 86.26% to 60.90%. Full P\wedge E\wedge S validation uses 560 localized DAVIS cases: top-1 relevance and typed authorization have comparable release (28.93% and 28.39%) but 6.79% and 0.89% unsafe release, respectively. Of the 33 additional unsafe cases removed, 27 are support, 4 edge, 2 clean, and 0 boundary cases. Thus the observed increment is an E\wedge S decision dominated by support, not an edge check alone. Appearance supplies E evidence only conditional on P ; authenticated subject tokens instantiate the missing presence witness as a sufficiency control. The claims concern the evaluated contracts, not natural prevalence, consent, or visual identity
[AI-73] ProtLingo: Efficient Protein Language Modeling via Conditional Memory and Expert Routing
链接: https://arxiv.org/abs/2609.04793
作者: Mingrui Li,Sixian Shen,Minzhang Li,Ruiyi Zhang,Kexin Zhang,Jiakai Zhang,Jingyi Yu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Proteins perform diverse cellular functions, and even single amino-acid substitutions can alter stability, activity, or molecular interactions. Protein language models (PLMs) provide a scalable approach for modeling such sequence–function relationships from unlabeled sequences, but increasing the size of dense Transformer backbones often brings substantial computational cost without consistently improving mutation-sensitive prediction. We introduce ProtLingo, an efficient PLM framework that augments a pretrained single-sequence backbone with conditional local memory and sparse expert routing. ProtLingo maps contextual residue representations into route-specific discrete codes, composes centered local windows into latent N -gram addresses, and retrieves reusable residual signals associated with recurring local sequence contexts. In parallel, selected feed-forward blocks are upcycled into sparse Mixture-of-Experts layers with shared and routed experts, enabling residue-dependent computation while activating only a subset of parameters. Experiments on protein fitness prediction, FLIP benchmarks, and supervised contact prediction show that ProtLingo achieves competitive performance with a 150M-scale backbone, including strong parameter efficiency on mutation-effect prediction and preserved long-range structural representations.
[AI-74] DODR: Deterministic Operator-Driven Reasoning in Latent Space
链接: https://arxiv.org/abs/2609.04782
作者: Weicai Huang(Beijing MQPat Technologies, Co., Ltd.)
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Autoregressive (AR) large language models formulate reasoning as token-level probabilistic sampling, which induces three fundamental defects in complex logical reasoning: error accumulation, probability substituting necessity, and the linear-chain information bottleneck. This paper proposes the Deterministic Operator-Driven Reasoning in Latent Space architecture (DODR), which reconstructs reasoning as reasoning-graph computation in a high-dimensional linear-algebraic space. Reasoning states are represented as snapshot vectors whose primitives are semantic units (phrases or sentences) rather than tokens, and each inference step is a deterministic matrix operation with no token sampling. Peirce’s three inference types are formalized as three trainable matrix operators: a rank-deficient deduction operator (information collapse), a full-rank induction operator (information expansion), and an abduction operator defined as the Moore-Penrose pseudo-inverse of deduction (information hypothesizing). We prove that the operator set is minimal and complete given Peirce’s trichotomy, that no single “super-operator” can realize all three types (a rank obstruction), and that reasoning graphs are Turing-complete with contractive backflow converging by Banach’s fixed-point theorem. Experiments on 503 sample records (420 deduplicated samples) across dedicated and end-to-end settings show: deduction loss converges to 1.40e-05; induction achieves 0.9996 generalization coverage with 20/20 hard vetoes on counterexamples; abduction solutions exceed the random baseline by 28x with judgment accuracies of 72.5% (58/80, Wilson 95% CI [61.9%, 81.1%]) and 81.7% (49/60, CI [70.1%, 89.4%]); frozen operators attain 100% (60/60) on unseen cross-domain deduction. The architecture provides a structural zero-hallucination guarantee and a three-layer continual-learning mechanism. All data and code are released.
[AI-75] Dynamic Heterogeneous Graph Representation Learning: A Survey IJCAI2026
链接: https://arxiv.org/abs/2609.04779
作者: Huan Liu,Pengfei Jiao,Jie Yin,Hongjiang Chen,Zhidong Zhao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI)
备注: IJCAI 2026 Survey Track
Abstract:Graph representation learning (GRL) serves as a canonical paradigm for modeling complex networks. However, real-world AI systems inherently manifest as evolving heterogeneous entities with complex interactions, posing significant challenges to static or homogeneous modeling. To address these complexities, representation learning for Dynamic Heterogeneous Graphs (DHGs) has emerged as a vital approach for learning low-dimensional representations that simultaneously preserve structural semantics and temporal dynamics. This survey presents the first systematic review of DHG representation learning methods. We first introduce a unified formal definition that encompasses both discrete-time and continuous-time DHGs from the perspective of temporal granularity. Building upon this formulation, we propose a novel algorithm-centric taxonomy that categorizes existing literature, including early embedding-based approaches, graph neural network (GNN)-based models, and relatively recent Transformer-based DHG methods, while explicitly highlighting their intrinsic modeling biases with respect to dynamic granularity. Furthermore, we summarize representative applications of DHG representation learning, along with commonly used datasets and benchmarks. Finally, we discuss promising research directions that guide future advances in this rapidly evolving field.
[AI-76] Diffusion Language Models for Mobile Edge Agent ic AI: Foundations Applications and Challenges
链接: https://arxiv.org/abs/2609.04778
作者: Chenqi Li,Minghui Min,Dusit Niyato,Wei Ni
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion language models (DLMs) offer a non-autoregressive alternative for mobile edge agentic artificial intelligence (AI) by refining tokens through iterative denoising rather than left-to-right decoding. Compared with autoregressive Transformer-based large language models (LLMs), DLMs can update multiple uncertain tokens in parallel and exploit bidirectional context throughout the generation process, enabling more flexible quality-latency trade-offs beyond fixed sequential decoding. These properties are particularly attractive for edge agents, where partial refinement, early exit, and constraint-guided correction can reduce response delay and communication overhead while improving robustness under noisy, incomplete, or dynamic contexts. This survey reviews DLM foundations and analyzes their suitability for edge settings under latency, memory, energy, bandwidth, privacy, and reliability constraints. We cover resource-efficient architectures, training and inference acceleration, compression, edge/cloud deployment, communication-aware serving, Internet of Things (IoT)/wireless applications, and evaluation of DLM-based agents. We further discuss open issues in long-context state management, split inference, trustworthy execution, multimodal grounding, and reproducible benchmarking. The goal is to connect DLM modeling properties, including bidirectionality, parallel refinement, controllability, and quality-latency elasticity, with system-level requirements of future mobile edge intelligence.
[AI-77] Shadow Queries for Private Retrieval in Vector Databases
链接: https://arxiv.org/abs/2609.04767
作者: Xinguo Feng,Zhongkui Ma,Zihan Wang,Chuan Yan,Guowei Yang,Alsharif Abuadbba,Guangdong Bai
类目: Artificial Intelligence (cs.AI)
备注: 13 pages
Abstract:Large language models (LLMs) increasingly rely on information retrieval (IR) systems, such as Retrieval-Augmented Generation (RAG), to incorporate domain-specific knowledge without costly re-training. These systems often store pre-computed document embeddings in cloud-based vector databases. However, such embeddings are vulnerable to embedding inversion attacks (EIAs), which can reconstruct their underlying text. Existing defenses, such as adding noise or scaling embeddings, often provide limited privacy or significantly reduce retrieval utility. We propose SHAQ (shadow query generation), a semantic-decomposition and embedding-decoupling defense against EIAs. SHAQ is based on the insight that EIAs rely on the strong coupling between an embedding and its original text. Instead of storing document embeddings directly, SHAQ uses a generative language model to create diverse shadow queries that capture different semantic aspects of each document. These queries are then encoded and stored in place of the original document embeddings, thereby decomposing document semantics and decoupling stored embeddings from the source text. Experiments across diverse IR datasets show that SHAQ substantially improves privacy while preserving retrieval utility, achieving a recovery rate as low as 0.2104, defending up to 19.50% more tokens than baseline defenses, and reaching up to 0.7967 MAP@10 with up to 5.53% utility improvement. These results demonstrate that semantic decomposition and embedding decoupling provide an effective alternative to directly modifying embeddings for defending against EIAs. Comments: 13 pages Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04767 [cs.AI] (or arXiv:2609.04767v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.04767 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-78] When Does an Interpretation Count as Established? The Formation Evaluation and Responsibility of Interpretation in Generative AI
链接: https://arxiv.org/abs/2609.04766
作者: Deyu Jing
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注: Conceptual paper on generative AI, interpretive standing, evaluation, and responsibility in humanistic scholarship
Abstract:Generative AI research has increasingly evaluated factuality, citation, coverage, and report structure. Yet passing such local checks does not by itself show that a humanistic interpretation has been established. This paper asks how an interpretation comes to be recognized within sociotechnical processes. It introduces three connected concepts. Interpretive appearance names the gap between the finished form of an output and the publicly traceable process through which materials, counterevidence, and revisions constrained the judgment. The evaluation contract names the bounded materials, tasks, criteria, permitted inferences, and failure conditions within which a local judgment is valid. Standing substitution names the unwarranted conversion of a genuine local pass into a stronger claim that an interpretation, result, or research capability has been established, without commensurate new evidence or bridging arguments. The paper then examines responsibility for judgment: a text may acquire recognition while no public structure remains for stating reasons, answering objections, revising, downgrading, or withdrawing the conclusion. Humanistic scholarship provides a revealing test because new materials and conceptual distinctions can alter both the question and the criteria of evaluation. The paper therefore develops delayed closure as a practice of keeping recognized interpretations revisable and proposes five public requirements concerning materials and versions, evidential roles, failure, contract revision, and responsibility. The argument is conceptual and normative: it does not claim to offer a benchmark or to determine whether models possess understanding. It instead explains why local evaluation, finished textual form, and public recognition must not be treated as sufficient evidence that an interpretation has been formed.
[AI-79] DCFA: Dual-view Causal-inspired Attribution for Failure Reasoning in LLM -based Multi-agent Systems
链接: https://arxiv.org/abs/2609.04749
作者: Zehao Wang,Lanjun Wang,Shilong Jin,Junjie Chen,Yanghua Xiao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM)-based multi-agent systems have experienced rapid growth in recent years. Despite their promise, such systems remain fragile, frequently exhibiting reasoning and coordination errors that can lead to system-level failures. Failure attribution in such systems relies on tracing natural language interactions among agents to identify the decisive error, which refers to the earliest action whose correction can reverse system failure. There are two key challenges: 1) Shallow attribution: Existing methods often capture only minor deviations, such as incomplete retrievals or formatting errors, which verification mechanisms can correct, while missing the decisive cause of system failure. 2) Contextual degradation: As the length of the system traces increases, the model’s reasoning ability rapidly deteriorates. To address these challenges, we propose DCFA, a training-free framework for failure attribution. DCFA integrates a global module that constructs structured causal-inspired dependency graphs from system traces to identify the initial decisive error, and a local module that applies local counterfactual-inspired reasoning to refine causal-inspired attribution. Experiments on the WhoWhen benchmark across six LLMs show that DCFA improves step-level accuracy by up to 8.27% over state-of-the-art baselines.
[AI-80] Aplaud: Adaptive Personalized Low-Rank Decomposition for User-Specific LLM
链接: https://arxiv.org/abs/2609.04738
作者: Xinyu Li,Ruoming Jin,Jianfeng Zhu,Ruixin Guo,Zhi Liu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:In this paper, we study the problem of personalized survey response prediction using fine-tuned large language models (LLMs). This task poses unique challenges: limited per-user training data, scalability of model storage, and the need to exploit shared structure across survey questions. To address these issues, we propose Aplaud (Adaptive Personalized Low-rank and User-specific Nested Decomposition), a lightweight and scalable framework for LLM personalization. Aplaud extends the LoRA paradigm by separating adaptation into a frozen, shared low-rank basis and a compact user-specific correction, augmented with a rank-one residual for finer personalization. To further reduce per-user parameter cost and mitigate overfitting, the correction matrix can be factorized into an even lower-rank form. Empirical results demonstrate that Aplaud achieves efficient, scalable personalization across users while outperforming state-of-the-art LoRA-based personalized LLM approaches in both generalization and inference efficiency.
[AI-81] PLUME: Parameter-Efficient Personalization of Large Language Models via Low-Rank User Modulation in Shared Subspaces
链接: https://arxiv.org/abs/2609.04715
作者: Xinyu Li,Hao Zhou,Jianfeng Zhu,Julina Maharjan,Ruixin Guo,Feodor Dragan,Ruoming Jin
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Personalizing large language models (LLMs) is essential for delivering AI assistance that aligns with individual users’ styles, intents, and preferences. While per-user fine-tuning can substantially enhance personalization quality, it introduces significant parameter and storage overhead, limiting scalability to large user populations. We propose PLUME (Personalized Low-Rank Adaptation through User Modulation and Shared Subspace), a lightweight framework that achieves efficient and expressive per-user adaptation by leveraging a shared task-specific subspace. Specifically, PLUME first learns a global task subspace from aggregated user data. Personalization is then achieved by training only a lightweight small square matrix within this subspace, enabling each user to obtain a tailored model while keeping shared components fixed. Cross-layer shared parameters and rank-1 residual terms are further introduced to significantly reduce redundancy while maintaining expressiveness. Experiments on multiple personalized text generation benchmarks demonstrate that PLUME achieves comparable or superior performance to strong baselines, while reducing per-user parameters by over 95%. These results establish shared-subspace modulation with minimal residuals as a scalable and semantically grounded approach to LLM personalization.
[AI-82] Building a research-software catalog with a coding agent : from hackathon prototype to public deployment
链接: https://arxiv.org/abs/2609.04711
作者: Kazuyoshi Yoshimi,Satoshi Terasaki,Gotai Yamada
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Physics Education (physics.ed-ph)
备注: 23 pages, 6 figures, 2 tables
Abstract:Generative AI and coding agents can accelerate research software development, but they also increase the need for efficient software discovery and maintenance. We developed a repository catalog during a three-day hackathon and subsequently examined the engineering required to make it suitable for public deployment, including adversarial review, data-quality checks, browser-level validation, and publication safeguards. We then explored whether the lessons learned from this prototype could be transferred to a much larger, human-curated portal, through a retrieval agent under development for MateriApps that combines curated portal metadata, external documentation, vector search, and local language-model generation. Implementation with coding agents was rapid, but achieving reliable operation required substantial additional engineering: the most consequential problems were not crashes but silent failures that produced plausible yet incomplete or incorrect outputs, arising from incomplete data acquisition, misleading assessments, and retrieval or preprocessing failures. These observations suggest that AI-assisted software portals require explicit validation, monitoring, and repeated review, and that curated metadata and maintained documentation remain essential. The MateriApps work is exploratory and remains under active development, so the observations reported for it are preliminary; a comparable combination of curated metadata, automatically collected documentation, and retrieval-based assistance may nevertheless be useful for extending other research-software portals.
[AI-83] Simulation-free Unbalanced Dynamic Optimal Transport with General Growth Penalty
链接: https://arxiv.org/abs/2609.04710
作者: Junda Ying,Yuxuan Wang,Bowen Yang,Peijie Zhou,Lei Zhang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM)
备注:
Abstract:Inferring cellular dynamics from unpaired single-cell snapshots requires modeling both state transitions and population growth or death. Unbalanced dynamic optimal transport (UDOT) addresses this by penalizing growth along transport paths, making the choice of growth penalty a key way to encode biological priors on proliferation and apoptosis. However, existing UDOT solvers either rely on computationally expensive NeuralODE simulations or depend on analytical solutions of conditional paths, restricting their efficiency solely to quadratic penalties, i.e. Wasserstein-Fisher-Rao (WFR) geodesics. To enable an efficient UDOT solver for general growth penalties, we first show that concave growth penalties lead to degenerate solutions where growth and transport are separated. We then introduce \textbfSimulation-free \textbfUnbalanced \textbfDynamic \textbfOptimal transport (SUDO), a simulation-free framework for UDOT with general non-quadratic convex growth penalties. SUDO learns the conditional paths and transport costs, solves the induced semi-coupling problem, and subsequently leverages unbalanced flow matching to achieve a simulation-free solution. On WFR benchmarks, SUDO matches the accuracy of efficient, analytical solution-driven algorithms while outperforming simulation-based methods in computational speed. Beyond WFR, SUDO supports asymmetric penalties that encode proliferation-dominant priors and produce more plausible trajectories and growth estimates on synthetic and single-cell datasets.
[AI-84] FinalityBench: An Effect-Level Benchmark for Agent Decisions Under Delayed and Conflicting Financial Finality
链接: https://arxiv.org/abs/2609.04706
作者: Abhishek Sharma
类目: Artificial Intelligence (cs.AI)
备注: 9 pages, 3 figures, 7 tables. Code, corpus generator and all result files: this https URL Archived at doi: https://doi.org/10.5281/zenodo.22262591
Abstract:A merchant’s payment processor, ledger, ERP and bank feed are updated by messages that get delayed, duplicated, dropped and reordered, so for minutes at a time the four hold contradictory beliefs about the same order. An agent resolving the exception must decide whether to ship goods, re-submit a capture, refund or wait, knowing some of those cannot be undone. We present FinalityBench, an executable benchmark for that decision. It keeps a hidden canonical event log and derives each system’s view from a separately faulted delivery stream, so disagreement follows from specified fault semantics rather than being authored. Grading is on executed monetary effects: an episode is scored by the merchant’s terminal economic position, relative to a privileged reference told when the pending capture resolves. The corpus of 321 tasks includes 45 twin pairs (90 tasks): tasks whose four system views are identical at the decision instant, whose authoritative probes both return unknown, and whose eventual correct dispositions differ. That snapshot indistinguishability is checked under every evaluation seed rather than assumed; equivalence over all interaction traces is not claimed. Over 14,445 graded episodes from nine programmatic policies, ranking by single-task accuracy and by paired loss disagree in 7 places: a ship-on-first-sign policy is second-best by accuracy at 65.7% and worst in the suite by paired loss, because it cannot tell the two members apart. A runtime gating irreversible actions on an authoritative finality probe reaches 85.4% and, unlike every polling policy, loses nothing to pass^5; its residual loss is almost entirely one archetype, which prices finality information directly. Language models reach the same exact rate as the hand-written gate on a stratified subset, lose about twice as much money, and discover the finality-gating strategy without being told it.
[AI-85] Model Retirement Creates Reproducibility Risk in Biomedical AI Publications
链接: https://arxiv.org/abs/2609.04699
作者: Nathan Wolfrath,Meghan Conroy,Thomas Kosten,Dave Bell,Bhabishya Neupane,Jonah Kindel,Anjishnu Banerjee,Priya Deshpande,Bradley Taylor,Anai N. Kothari
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Background. Large language models (LLMs) are being adopted in biomedical research at a rapid and accelerating pace, yet commercial services that host many widely used models operate under deprecation schedules that can complicate scientific reproducibility. Methods. We searched PubMed for original research articles from 2022 through March 2026 that applied a specific LLM to a biomedical task. An extraction agent identified model names from 61,077 article abstracts with human reviewers validating a subset for extraction accuracy. Extracted model names were normalized to canonical model identifiers. Lifecycle data (release date, retirement date, status) were compiled for the 50 most frequently used models. Results. We identified 8,931 paper-model mentions spanning 5,242 unique publications after restricting the analysis to the 50 most frequently used models. Among these mentions, 77.7% cited a commercial closed-weight model. Overall, 42% involved a model that was already retired by the time of official publication or is scheduled to retire within two years of publication. The median interval from publication to model retirement was 538 days. Conclusion. Many biomedical publications using LLMs are on a trajectory toward computational non-reproducibility after publication. Model deprecation should be treated as a core reporting and preservation issue for biomedical research. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04699 [cs.AI] (or arXiv:2609.04699v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.04699 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Nathan Wolfrath [view email] [v1] Fri, 4 Sep 2026 04:00:54 UTC (411 KB)
[AI-86] SQL-Zero: Self-Evolving Text-to-SQL
链接: https://arxiv.org/abs/2609.04697
作者: Daniel Machado Pedrozo,Julia Soares Dollis,Bryan Lincoln Marques de Oliveira,Vinicius Alboneti Aguiar,Sávio Salvarino Teles de Oliveira,Telma Woerle de Lima Soares
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Training a competitive Text-to-SQL agent usually depends on human-annotated natural-language/SQL pairs, which are expensive, domain-specific, and a bottleneck for scaling to new databases. We show it is possible to train a competitive solver with zero annotated pairs. We introduce SQL-Zero, a proposer-solver self-play in which a challenger and a solver start from the same base LLM and the only ground truth is execution against the database itself. The challenger generates SQL pairs calibrated to the solver’s current difficulty (targeting “hard but solvable”), and both roles are updated with GRPO in alternating turns, with a template-level repetition penalty on the challenger to prevent diversity collapse. Training on BIRD databases with no labels, self-play improves over the zero-shot base on BIRD dev by 6.6 points at 3B and 7.3 points at 7B. It also scores higher than a matched control trained under the same recipe on human BIRD gold over the same databases, although an exact paired test does not resolve that margin. Transfer depends on scale: at 3B every iteration outperforms the base on unseen Spider databases and under lexical perturbation (Spider-Syn), where it also degrades less than the matched BIRD-gold control, whereas at 7B only the first iteration preserves transfer.
[AI-87] Predicting Spatiotemporal Mobile Sensing-Based PM2.5 Concentrations Using Low-Rank Adapted Spatially Attentive Graph Neural Network
链接: https://arxiv.org/abs/2609.04693
作者: Om Chiddarwar,Priyanka Mandal,Praveen Kumar Chandaliya,Shriniwas Arkatkar
类目: Artificial Intelligence (cs.AI)
备注: 24 Pages, 14 Figures, World Conference of Transport Research2026 Transport Research
Abstract:Urban air quality can vary significantly along transit corridors, necessitating high-resolution monitoring. This work introduces a novel mobile-sensing dataset from Surat, Gujarat, India, comprising PM *2.5 concentrations, meteorological variables (temperature, humidity, wind speed, wind direction), and land-use features. To represent the spatiotemporal data as a graph, two node-definition strategies were used: (i) uniform segmentation (200–400~m intervals) and (ii) DBSCAN clustering to adaptively group dense observations. For each node, rolling mean and standard deviation of meteorological variables were computed. To model this high-dimensional data, we propose a SA-GNN for fine-grained, short-term PM *2.5 forecasting and hotspot identification. We compared SA-GNN with LSTM, RNN, GRU, and ANN models. These models performed well on low-resolution data but had difficulty capturing rapidly changing patterns in urban air quality. SA-GNN employs cluster-specific GRUs to capture localized temporal dependencies and a Graph Attention Network to learn spatial heterogeneity. This hybrid architecture effectively models rapid fluctuations and complex spatial interactions. On our dataset, SA-GNN achieved R^2 = 0.95 , RMSE = 6.8 , and MAE = 4.2~\si\micro\gram\per\meter\cubed , outperforming all baseline models. Combining spatial clustering with adaptive attention significantly improves forecasting, enabling real-time, fine-grained monitoring and supporting personalized exposure tracking and timely alerts for healthier cities.
[AI-88] Beyond Code Generation: Reliability Verification and Cost Economics in the Agent ic Software Development Lifecycle
链接: https://arxiv.org/abs/2609.04681
作者: Happy Bhati
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 18 pages, 6 figures, 3 tables. Systems synthesis and research agenda on agentic software engineering, code review, testing, reliability, and AI cost. No new experimental measurements are claimed; empirical and company-reported results are attributed to the cited sources
Abstract:AI coding systems are moving from autocomplete and chat toward agents that can inspect repositories, edit multiple files, run tools, write tests, open pull requests, and work for long periods with limited supervision. This capability changes the bottleneck in software delivery. Recent field studies show meaningful gains in coding activity, but newer evidence also shows that those gains attenuate sharply between writing code and shipping reliable software. Review, integration, testing, security, deployment, and production operations remain constraining stages, while the economics are shifting from predictable per-seat licensing toward variable token, tool, sandbox, CI, and rework costs. This paper synthesizes peer-reviewed software-engineering research, university studies, benchmark audits, production reports from major technology companies, developer telemetry, and cost-management evidence released primarily from 2024 through September 2026. No new model experiment is claimed; numerical findings remain attributed to their original studies. The synthesis proposes four engineering concepts: the Agentic SDLC Throughput Paradox, Production-Qualified Change (PQC), the Verification Tax, and an Agentic SDLC Control Plane that allocates autonomy subject to cost, reliability, and human-attention budgets. An evidence-based horizon then maps today’s supervised agents to future policy-bounded software factories. The central research question shifts from how much code an agent can generate to how much production-qualified value an engineering system can deliver per dollar, per reviewer-hour, and per unit of operational risk. Comments: 18 pages, 6 figures, 3 tables. Systems synthesis and research agenda on agentic software engineering, code review, testing, reliability, and AI cost. No new experimental measurements are claimed; empirical and company-reported results are attributed to the cited sources Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04681 [cs.SE] (or arXiv:2609.04681v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2609.04681 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-89] rain What You Deploy:Token-Faithful Post-Training of a Production Coding
链接: https://arxiv.org/abs/2609.04678
作者: Cheng Li,Jiexiong Liu,Yixuan Chen,Chi Hong
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Existing post-training pipelines for coding and terminal agents suffer severe token and control fidelity errors: simplified training environments mismatch production deployments, and offline token reconstruction from agent logs distorts original prompts and conflates policy calls with background model operations. We present a fidelity-aware training coupling framework that retains trainer-side sampling over original prompts, eliminates spurious model calls via a negotiated training protocol, and restricts loss computation to verifiable token spans with closed-failure guarantees. We further propose Certified Divergence Proximal Policy Optimization (C-DPPO), which establishes tight two-sided TV certification bounds, adaptive-K rules, budget-aware sequence guarantees, and error-robust policy masking atop standard DPPO. Evaluated on matched Baize5B and Baize10B models with identical training and test protocols on TMax-100, C-DPPO yields a consistent +3.0-point performance gain over standard DPPO across model scales. Certificate audits validate the reliability and full operational coverage of our certified training pipeline.
[AI-90] ERPBench: Evaluating LLM Agents for Enterprise Decision-Making Across Competitive Market Ecologies
链接: https://arxiv.org/abs/2609.04667
作者: Xinran Zhang,Pengrui Lu,Lyumanshan Ye,Pengfei Liu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM) agents are increasingly proposed for enterprise workflows, yet existing evaluations rarely test whether business-decision conclusions transfer across competitive market ecologies. We introduce ERPBench, an execution-instrumented benchmark for enterprise decision agents in a six-round Enterprise Resource Planning (ERP) simulation with coupled pricing, production, procurement, inventory, finance, and shared-market competition. ERPBench evaluates the same 100 fixed problems in two matched competitive market ecologies: Solo, where each evaluated LLM agent competes against fixed rule-based opponents, and Arena, where six evaluated LLM agents compete in a shared market. Across six model families, this yields 1,200 model-level trajectories spanning 7,200 decision rounds. Under the observed service configuration, the leading model differs between ecologies: DeepSeek leads in Solo (252.29M mean valuation; mean rank 1.67), whereas Gemini leads in Arena (263.95M; 1.76). The two ecologies identify the same task-level winner on only 21 of 100 problems, and Gemini’s bottom-rank rate falls from 22 % to 0 % in Arena. ERPBench supports paired evaluation of whether enterprise-agent rankings transfer across competitive market ecologies, supplemented by aggregate execution-intervention analysis. Code and benchmark resources are available in our this https URL.
[AI-91] Harness-agnostic detection and immunization of reward hacking in self-evolving language models
链接: https://arxiv.org/abs/2609.04665
作者: Rongxin Yang,Yang Liu,Shang Luo,Haoxuan Jia,Chongyang Zhang,Hao Zheng,Yingguang Yang,Yulin Huang,Jianshen Zhang,Yongzhi Qi,Kefu Xu,Congjing Ran,Bin Chong
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Self-evolving language models improve by proposing candidate updates and keeping whatever raises a visible score. When that score is an imperfect proxy for the capability one actually wants, sustained selection widens the gap between the two. This is reward hacking. We introduce HackProbe, a monitor that attaches to an arbitrary self-evolving loop through two black-box hooks, with no access to weights or activations. It keeps a secret, distribution-fixed comparison core, whose frozen distribution makes its capability proxy comparable across generations, alongside a rotated fresh layer that hardens the bank against co-adaptation. Four tests built on that proxy cover the level gap, a scale-aligned divergence with online change-point detection, capability stagnation, and a conditional confidently-wrong rate; a Sidak correction turns them into a calibrated family-wise p-value. Diagnosis alone recovers nothing, so a risk-aware immunization layer reselects an honest candidate from the proposal pool using the core together with a purely structural gaming footprint, disclosing at most log2 Pi bits per generation to the host. We prove a detectability bound that converts a target error rate into an explicit probe-size budget, and we delimit what probe rotation does and does not buy. On a controlled prompt-level host with four injected hacking channels and ground-truth labels, HackProbe reaches 0.763 AUROC against 0.663 for the strongest baseline and cuts the false-positive rate from 0.706 to 0.434. Its bandwidth-limited reselection is the only immunization level that returns more true capability under hacking, 5.2 points on average, than it forfeits on clean runs, 4.7; per-channel effects are mostly not individually significant.
[AI-92] Continual Graph Memory for Adaptive Recommendation under Intent Drift EMNLP2026
链接: https://arxiv.org/abs/2609.04651
作者: Hao Nguyen Ngoc,Tung Nguyen,Nguyen Thi Hanh,Hoang Thai Dinh,Nguyen Xuan Tung
类目: Artificial Intelligence (cs.AI)
备注: Accepted to Findings of EMNLP 2026
Abstract:This paper studies adaptive recommendation under intent drift, where feedback from each recommendation outcome can reveal whether the relational evidence used for ranking is useful, missing, or misleading. While Knowledge Graphs (KGs) provide essential semantic structure to handle these shifts, traditional KG-enhanced systems treat the graph as a static retrieval substrate, making it brittle to evolving intents, noisy metadata, and recurring failure patterns. This paper proposes CGM-Rec, a continual graph memory framework for adaptive recommendation. CGM-Rec treats the graph state as a writable memory and maintains two complementary components. Therein, a Semantic Graph Memory is updated conservatively through quality-gated typed operations for storing stable and high-confidence relational knowledge. Meanwhile, an Episodic Lesson Memory acts as a fast reactive memory that learns recent outcomes, failure cases, and corrective hints. During testing, model parameters remain frozen and adaptation occurs only through memory writes. We evaluate CGM-Rec under a frozen-parameter, one-pass reranking protocol, where encoders and prompts remain fixed during testing and adaptation occurs only through memory writes. Experiments across multiple recommendation settings show that CGM-Rec improves over evaluated neural and LLM-based baselines on most metrics. Particularly, under sampled-candidate reranking, CGM-Rec improves HR@1 by up to 29.58% over the strongest LLM baseline on Bundle, and outperforms K-RagRec on metadata-rich ML-100K with HR@5 of 0.5941 versus 0.4746.
[AI-93] A Cost-Aware Agent ic Architecture for NL-to-SQL over Nested Enterprise Schemas with a New Benchmark
链接: https://arxiv.org/abs/2609.04641
作者: Yoga Sri Varshan Varadharajan,Ajay Yadav,Ritesh Goru,Prateek Chaudhury,Constantine Caramanis,Prateek Jain,Divyateja Pasupuleti,Sunil Kumar Pandey
类目: Artificial Intelligence (cs.AI)
备注: 17 pages, 3 figures
Abstract:Natural-language-to-SQL systems have ad- vanced rapidly on academic benchmarks, yet production enterprise schemas exhibit graph- like, semi-structured, deeply nested structure that current benchmarks do not measure. We make two complementary contributions. First, we introduce the DevRev NL2SQL bench- mark: 900 execution-verified queries with nested-type and link-graph structure, accom- panied by the Semantic Depth Score (SDS), a schema-agnostic rubric for analytical reasoning depth. Second, we present a cost-aware single- generation agentic architecture whose schema- selection, metadata-retrieval, and error-repair components are designed for the requirements this regime imposes. On the DevRev NL2SQL benchmark the system attains 91.7% answer correctness, a margin of 54.6 percentage points over the next-best baseline; on the Spider 2.0 Snowflake public dataset, it is competitive with leading systems at a single-generation operating point.
[AI-94] SCAPES: Semantically Conditioned Autoregressive Prior for Environmental Sounds
链接: https://arxiv.org/abs/2609.04634
作者: Esteban Gutiérrez,Lonce Wyse,Frederic Font,Xavier Serra
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to the Digital Audio Fx (DAFx) Conference 2026 to be held in Cambridge, USA. 8 pages, 3 figures and 2 tables
Abstract:As generative audio models grow in complexity, the computational and ecological costs of synthesizing everyday sounds have become increasingly prohibitive, often requiring industrial-scale resources and massive datasets. In this paper, we present SCAPES: a Semantically Conditioned Autoregressive Prior for Environmental Sounds. SCAPES is a lightweight, resource-efficient generative model designed to synthesize high-fidelity environmental textures through high-level semantic control. By operating on the continuous latent manifold of a neural audio codec, our approach bypasses the rigid structural constraints inherent to discrete tokenization. We propose a segmentation strategy that decomposes audio into overlapping segments, enabling a Continuous Normalizing Flow (CNF) to model the evolution of latent trajectories using Flow Matching. Our experiments demonstrate that a 36-million parameter instance of SCAPES can be trained on limited, uncurated datasets using a single consumer-grade GPU. Notably, convergence is achieved after training for approximately twice the source audio duration, yielding high-fidelity outputs with robust long-term stability and semantic consistency. Furthermore, we showcase the model’s capacity for smooth semantic interpolation, providing a flexible and accessible tool for open research and creative sound design. Code, pretrained weights, audio examples, and an interactive demo are publicly available on our project page this https URL
[AI-95] SiLR: Structure-Preserving Admission and Process Reward for LLM Tool Agents
链接: https://arxiv.org/abs/2609.04629
作者: Chenyu Zhou,Qiliang Jiang,Shuning Wu,Xu Zhou
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Systems and Control (eess.SY)
备注: 13 pages, 8 figures, 6 tables. Appendix includes full proofs, attack-family constructions, and the extended process-reward study
Abstract:A runtime gate for an LLM tool agent is usually cast as a filter. In a ReAct loop a rejected proposal is followed by another at the same state, so the gate is a search operator over the proposal stream whose admission criterion shapes which trajectories are reachable. We study post-violation recovery admission, where progress must be admitted while the system is still in violation, and identify the scalar projection trap: an aggregate-score gate accepts a locally improving proposal and commits the trajectory to a plateau. SiLR instead shadow-executes each proposal and admits it under a product order over the branch-level violation state (overloaded-branch support and per-branch severity). We prove that no scalar surrogate is sound for this order, so the failure is representational, not a matter of threshold tuning. On mined Gym-ANM scenarios, SiLR recovers 21/21 multi-action episodes against 0/21 for terminal and 9/21 for the best scalar gate, significant across the full 24-scenario benchmark. The terminal-versus-structured dichotomy holds across three model families and in CityLearn. Because admission rests on deterministic simulation, the LLM lies outside the trust boundary: a magnitude-redistribution attack that defeats both scalar and support-only baselines is contained only by the full per-branch predicate. With two constraint families active, every tested scalar projection admits physically unsafe actions; support-only admits the largest fraction (63.2% of 42,410; product order 0). In the hardest dual-family traces, scalar gates recover only through that unsafe class. Reused as a GRPO process reward, it outperforms its count projection in every mined scenario and is the only tested reward whose ungated policy exceeds the untrained base (0.844 vs. 0.778). Scalar projection loses the violation geometry at both design points; only the full product order is structurally sufficient.
[AI-96] Leverag ing Imperfect Restoration for Data Availability Attack ECCV2024
链接: https://arxiv.org/abs/2609.04627
作者: Yi Huang,Jeremy Styborski,Mingzhi Lyu,Fan Wang,Adams Kong
类目: Artificial Intelligence (cs.AI)
备注: Accepted to ECCV 2024. Equal contribution by Yi Huang, Jeremy Styborski, Mingzhi Lyu - cite in any order
Abstract:The abundance of online data is at risk of unauthorized usage in training deep learning models. To counter this, various Data Availability Attacks (DAAs) have been devised to make data unlearnable for such models by subtly perturbing the training data. However, existing attacks often excel against either Supervised Learning (SL) or Self-Supervised Learning (SSL) scenarios. Among these, a model-free approach that generates a Convolution-based Unlearnable Dataset (CUDA) stands out as the most robust DAA across both SSL and SL. Nonetheless, CUDA’s effectiveness against SSL is underwhelming and it faces a severe trade-off between image quality and its poisoning effect. In this paper, we conduct a theoretical analysis of CUDA, uncovering the sub-optimal gradients it introduces and elucidating the strategy it employs to induce class-wise bias for data poisoning. Building on this, we propose a novel poisoning method named Imperfect Restoration Poisoning (IRP), aiming to preserve high image quality while achieving strong poisoning effects. Through extensive comparisons of IRP with eight baselines across SL and SSL, coupled with evaluations alongside five representative defense methods, we showcase the superiority of IRP. Code: this https URL
[AI-97] ττ-Bench: An Environment for End-To-End Realistic Agent Construction
链接: https://arxiv.org/abs/2609.04611
作者: Quan Shi,Keshav Dhandhania,Karthik Narasimhan,Victor Barres
类目: Artificial Intelligence (cs.AI)
备注: 41 pages, 13 figures, 6 tables
Abstract:LLM agents are rapidly becoming production software, deployed to handle customer service, adjudicate disputes, and operate internal systems. Notably, the work of building them is increasingly handed to coding agents, yet existing benchmarks say little about whether an AI system can deliver one under the conditions of a real client engagement. We introduce \tau^\tau -bench (pronounced hyper-tau-bench), a benchmark that makes agent construction the task. A developer agent is given the records a business actually keeps, a client who holds requirements, a production API that operations must run through, a codebase to inherit, and limits on serving cost and models: the same starting point a real engagement provides. From these it must deliver a complete customer-service agent, scored by deploying that agent against held-out simulated users. Across 53 tasks spanning four domains, the strongest configuration, Claude Opus 5 under Claude Code, passes just 23.9% of evaluation simulations. Meanwhile, an expert-authored reference ceiling scores 82.2%. The failures mirror ones human agent developers see: models issue shallow queries in place of deep comprehension of the records, communicate almost nothing to the client, and experiment too little with agent architecture and serving spend, shipping the first design that runs. We aim for \tau^\tau -bench to turn the work of cooperative agent building into a measurable target for coding agents.
[AI-98] raining-Free Halving of Activated Experts in Fine-Grained Mixture-of-Experts Models
链接: https://arxiv.org/abs/2609.04575
作者: Xing Chen,Hengshuai Yao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Modern fine-grained Mixture-of-Experts (MoE) models route each token to a small number of experts and renormalize their router probabilities. We show that this renormalization implicitly calibrates expert output gain to the training top- k : reducing k at inference changes not only which experts are used but also the strength of the expert branch. We separate these effects by activating the top k_1 experts while normalizing by the probability mass of the top k_2 experts, introducing one integer with no parameters, training, or measurable compute overhead. On Qwen3.6-35B-A3B, reducing from 8 to 4 experts causes a 4.65-point MMLU drop under standard renormalization but only 0.35 points with k_2=16 , while halving routed-expert compute. The result replicates on the 11\times larger Qwen3.5-397B-A17B, where reducing from 10 to 5 experts loses only 0.55 points with an appropriate reference set. Removing renormalization entirely is catastrophic, showing that preserving a suitable reference mass is crucial. We further find that perplexity and downstream accuracy favor different k_2 , cautioning against selecting MoE compression settings using unlabeled text alone. Analyses also show that expert identity matters substantially more than expert weighting, while balanced and domain-specialized routing leaves limited room for expert pruning.
[AI-99] Dynamic Adaptation of the LLM Context for Generating Routines with Coupled Semantics ICANN2026
链接: https://arxiv.org/abs/2609.04570
作者: Gnaneswar Villuri,Hashmath Shaik,Alex Doboli
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: Accepted at ICANN 2026
Abstract:LLM-based code generation fails when correctness depends on execution-dependent coupling: the meaning of one routine is defined by the runtime behavior of another, a relationship that cannot be resolved from textual descriptions alone. This limitation, which we call static binding, is not confined to explicitly coupled problems; it appears to varying degrees whenever correctness depends on joint execution behavior across components, from explicit cross-coupled optimizers to subtler joint constraints in packing, routing, and symbolic search. This paper proposes dynamic context adaptation, a sample-efficient validation-generation loop designed for this setting. A validation agent extracts structured diagnostic information from execution traces, providing gradient-like guidance to a generation agent that proposes multiple candidates per iteration. A knowledge graph derived from the problem description supplies semantic constraints to the generation agent. Simulated annealing selects among candidates to avoid greedy collapse. Our method outperforms zero-shot, Reflexion, and OpenEvolve on seven of eight problems at both 300 and 600 evaluations (p 0.01), a regime where population-based search has not yet accumulated sufficient diversity to compete. Notably, on the primary motivating problem (cross-coupled optimization), our method also achieves the best score at 1000 evaluations, consistent with the hypothesis that structured execution feedback is most beneficial when correctness depends on runtime coupling. Ablation results confirm that structured execution feedback is the primary driver.
[AI-100] Reducing Hallucinated Transcripts in Whisper via Hallucination Space Projection
链接: https://arxiv.org/abs/2609.04561
作者: Maryam Abbasihafshejani,Murtuza Jadliwala
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Whisper is a widely used foundation model for automatic speech recognition (ASR), but its generative decoder can produce fluent hallucinated transcripts for inputs containing little or no speech. We propose a training-free, inference-time method to reduce these hallucinations using low-rank projection of decoder activations. A compact hallucination-associated subspace is estimated from non-speech calibration data, and decoder hidden states are projected away from this subspace during inference. We evaluate two variants: always-on, which applies projection to all inputs, and gated, which applies it only when Whisper predicts that an input is likely non-speech. Across non-speech benchmarks, always-on projection reduces average hallucination rate (HR) from 31.31% to 2.44%, a 92.21% relative reduction, while gated projection reduces HR to 3.74%, an 88.05% relative reduction, with lower false rejection of genuine speech. On LibriSpeech, gated projection increases absolute word error rate (WER) by 0.33-4.39 percentage points and yields false-rejection rates (FRR) of 0.41–9.97% across model and split settings. These results show that low-rank activation projection can substantially suppress Whisper hallucinations without retraining, while providing a controllable trade-off between hallucination suppression and speech recognition performance. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04561 [cs.AI] (or arXiv:2609.04561v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.04561 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-101] IPGeoAI: Transformer-Based Geolocation with LLM Semantic Fusion
链接: https://arxiv.org/abs/2609.04559
作者: Avinash Kadimisetty,Andy Jinqing Yu,Philip Favaloro,Wenlong Liu,Xiaolu Xiong
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Accurate city-level IP Geolocation is an important enabler for the modern digital ecosystem, underpinning services ranging from local content delivery and targeting to digital rights enforcement. However, traditional heuristic and database-driven methods often struggle to resolve the complex, non-linear allocation patterns of modern network infrastructures, particularly within the exploding IPv6 address space and transient mobile networks. In this paper, we introduce IPGeoAI, a novel deep learning model architecture that reframes geolocation from a static lookup problem to a sequential modeling task. Our approach utilizes the Transformer Encoder to capture hierarchical dependencies inherent in IP subnet structures. We propose a method to resolve geographic ambiguity by integrating unstructured semantic context via a Zero-Shot LLM Feature Extraction pipeline. We utilize Large Language Models to transform raw, noisy Autonomous Systems (AS) descriptions into structured, domain-specific metadata (such as ‘University’ vs. ‘ISP’ or ‘Global’ vs. ‘Local’) via an offline pre-computation process. By fusing these semantic signals into the network via a Multi-Head Cross-Attention module, we bridge the gap between numerical network topology and real-world semantic identity. Extensive offline evaluation on a proprietary dataset spanning 200,000 cities demonstrates that IPGeoAI significantly outperforms a leading external vendor in city-level granularity. By adopting a hierarchical inference strategy that refines coarse-grained country signals, our model achieves a 6% improvement in city-level accuracy while extending coverage to 100% of the traffic. Furthermore, in large-scale online production tests, the model drove a statistically significant +0.35% improvement in our 1st-tier downstream use cases metric.
[AI-102] Continual Field-Adaptive Models (CFAMs) for Post-Deployment Physical AI
链接: https://arxiv.org/abs/2609.04552
作者: Amarjot Singh,Tanmay R. Pancholi,Jainam Kothari,Shrirang Mahajan,Ketan Bansal,Zackory Erickson,Giuseppe Loianno,Alexandre M. Bayen,Jeff Schneider,Vince Nakayama
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Unattended interactive autonomy - machines that step into danger in place of humans and complete tasks with human tools - remains a missing capability in mission-critical operations. These domains offer scarce training data and only onboard compute, yet deployed systems must face novelty without erasing prior competence. We introduce Continual Field-Adaptive Models (CFAMs), which learn efficiently in the lab and continue learning after deployment through autonomous, gradient-free, on-device updates. CFAM uses a complementary learning architecture with a frozen slow-learning component and a fast-learning Capsule Field. The slow component contains three cortices: Sensor, which maps multimodal input into 3D-grounded geometry; Reasoning, which decomposes tasks into skills and evaluates outcomes; and Action, which executes geometric skills. The Capsule Field stores field learning one-shot and gradient-free as Competence Capsules. Skill installation is few-shot in the lab and continual in the field; open-world novelty is outside scope. We evaluate CFAM across five embodiments: manipulator, quadruped, humanoid, quadrotor, and off-road vehicle. Baselines (pi0, CogACT, SpatialVLA) use the same in-house multi-embodiment dataset for physical-platform comparisons. CFAM reaches the operating point of a standard policy trained on the full prior-training dataset using 40% of the data, or 2.5x fewer trajectories. At test time, autonomous capture of verified near-OOD cases improves action success by 13.9 percentage points. In sequential simulation, backward transfer is -0.5 percentage points versus -11.4 for LoRA. CFAM therefore provides a bounded form of post-deployment physical intelligence: few-shot skill learning, autonomous field growth from verified near-OOD experience, and retention of prior competence. Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04552 [cs.RO] (or arXiv:2609.04552v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2609.04552 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-103] From Answers to Interpretations: Rethinking Ambiguity-Induced Aleatoric Uncertainty Estimation in LLM s
链接: https://arxiv.org/abs/2609.04543
作者: Omer Nahum,Niv Nayman,Jonathan Fhima,Alon Zolfi,Jeremy Levy,Shai Mazor,Paolo Favaro
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:A key challenge in reliable LLM deployment is recognizing when uncertainty reflects irreducible variability in the task rather than limitations in the model’s knowledge. In language tasks, a central source of such aleatoric uncertainty is input ambiguity or underspecification, where multiple interpretations remain plausible. Existing decomposition methods estimate aleatoric uncertainty by generating multiple clarifications of the input, querying the model for an answer under each clarification, and comparing the resulting answers. We argue that answers are not necessary for identifying ambiguity: they are often redundant, add avoidable cost, and can mislead through epistemic leakage. We support this claim theoretically, and propose a clarification-only approach that estimates this ambiguity-induced component directly from the space of plausible interpretations, without answers to the clarified inputs. Using ambiguity detection as an operational evaluation across three benchmarks, this direct approach improves AUROC (63.34 vs. 60.85), reduces computational cost by 4-26x in output tokens and 2.2-3.5x in API calls, and yields estimates with substantially lower correlation with epistemic uncertainty. Overall, our results suggest that ambiguity-induced aleatoric uncertainty is better estimated from the interpretation space than from the response space.
[AI-104] Data-Driven Discovery of Composition-Dependent Constitutive Models for Hyperelasticity and Viscoelasticity of Digital Materials
链接: https://arxiv.org/abs/2609.04541
作者: Josué García-Ávila(1),Beijun Shen(1),Manuel K. Rausch(2 and 3 and 4),Mary C. Boyce(1),Adrián Buganza-Tepole(1) ((1) Department of Mechanical Engineering, Columbia University, New York City, USA, (2) Department of Aerospace Engineering and Engineering Mechanics, University of Texas at Austin, Austin, USA, (3) Department of Biomedical Engineering, University of Texas at Austin, Austin, USA, (4) Department of Mechanical Engineering, University of Texas at Austin, Austin, USA)
类目: Artificial Intelligence (cs.AI); Materials Science (cond-mat.mtrl-sci); Soft Condensed Matter (cond-mat.soft)
备注: 40 pages including references, 9 figures in the main manuscript. Supplementary material is available upon request from the corresponding author
Abstract:Digital materials fabricated by multi-material 3D printing are designed as controlled mixtures of stiff and compliant constituents, yielding effective responses that span more than an order of magnitude in apparent stiffness and exhibit strongly nonlinear, composition-dependent, and rate-dependent dissipative behavior. Classical finite-strain viscoelastic models represent such behavior with closed-form strain energy functions for equilibrium and non-equilibrium stresses as well as evolution of internal variables, which may limit flexibility when a single constitutive model is expected to generalize across materials and loading rates. Here, we present a data-driven multi-material constitutive modeling framework that generalizes a formulation by Bergström and Boyce. The proposed framework retains the structure of the classical model, namely multiplicative kinematics, invariant-based strain-energy functions, and a scalar dissipative evolution law directed along the normalized nonequilibrium deviatoric stress. For the equilibrium branch, the data-driven discovery framework either directly predicts closed-form model parameters as functions of composition or automatically constructs a polyconvex strain-energy function using neural ordinary differential equations (NODEs). The nonequilibrium branch kinetics are learned similarly, either by directly identifying closed-form parameters across compositions or by using appropriately constrained artificial neural networks. Using multi-rate uniaxial compression data across multiple material compositions, we show that the proposed formulation captures rate-dependent stiffness and hysteresis across compositions while preserving thermodynamic consistency.
[AI-105] Repeat-After-Me: Black-Box Adaptive Visual Prompt Injection
链接: https://arxiv.org/abs/2609.04533
作者: Sizhe Chen,Yu-Lin Tsai,Ivan Evtimov,Kamalika Chaudhuri,Raluca Ada Popa,David Wagner,Arman Zharmagambetov
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Prompt injection is widely recognized as a major security threat to AI agents that interact with untrusted external data, such as websites, documents, and emails. Prior work has shown that, in the text domain, black-box prompt injection can achieve near-perfect attack success rates (ASRs). In the image domain, however, existing visual prompt injection methods are substantially less effective in attacking frontier commercial VLMs for materially harmful behavior. Achieving such outputs is hard because it requires a long and/or format-compliant target string, such as a precise, parseable native tool call with exact function names and arguments. We present Repeat-After-Me, a black-box adaptive visual prompt injection attack that can reveal personally identifiable information or make malicious tool calls. Across both open-weight and commercial frontier VLMs, including Qwen3.6-27B and GPT-5.5, our method achieves ASRs exceeding 80% and 47%, respectively, under a realistic setting in which the benign user prompt is semantically unrelated to the injected task and does not verbally authorize it. In our evaluation, injections optimized on one surrogate retain 43-46% of the original ASR on two commercial victims, and cross-sample transferability retains 64-66% of the original ASR on those two models. We test our attack in a real-world OpenClaw agent: in a default OpenClaw Discord deployment, an untrusted user can use a minimally injected image to overwrite this http URL, enabling future sensitive behaviors like remote code execution and secret exfiltration. We show our new attack vector works in cases where adaptive textual prompt injection fails. We discuss potential defenses.
[AI-106] owards a universal language of concepts: A survey
链接: https://arxiv.org/abs/2609.04528
作者: Aishni Parab
类目: Artificial Intelligence (cs.AI); Programming Languages (cs.PL); Symbolic Computation (cs.SC)
备注: Originally completed as an M.S. capstone project at UCLA in 2022
Abstract:Humans can learn and generalize novel concepts from sparse data because they express knowledge in rich structural formats. In this paper, we propose that programs are a strong candidate for universal representation of concepts. We review computational models of concept learning that use programs as their concept representation and evaluate their contribution toward a universal representational language.
[AI-107] MaxKernel: Agent ic Kernel Generation for TPUs
链接: https://arxiv.org/abs/2609.04523
作者: Shangkun Wang,Nina Cai,Charles Hoong,Julian Walker,Gerson Kroiz,George Vanica,Deepak Patil,Andi Gavrilescu,Hassan Sipra,Sethu Sankaran
类目: Artificial Intelligence (cs.AI); Performance (cs.PF); Programming Languages (cs.PL)
备注: 14 pages, 6 figures, 4 tables
Abstract:Designing and authoring high-performance custom kernels for accelerators is a complex task that requires deep hardware-level expertise. Large Language Models (LLM) can be leveraged together with real-time compiler feedback to build agentic systems for kernel generation. In this work, we present MaxKernel, a multi-agent system that implements three distinct paradigms for TPU kernel development: (1) a Human-in-the-Loop (HITL) agent for collaborative, step-by-step design; (2) an Autonomous (Auto) agent that executes a fully automated, metric/trace-driven optimization loop; and (3) a Graph-Based Autonomous Search that scales the Auto agent for global exploration of the design space. All three paradigms leverage a shared pool of specialized sub-agents to handle planning, implementation, self-debugging, testing, and hardware profiling. We evaluate MaxKernel on JaxBench, a comprehensive suite of 50 diverse kernel tasks for TPUs, alongside complex, real-world workloads from state-of-the-art open-source models. We demonstrate that MaxKernel consistently generates highly optimized implementations, matching expert hand-tuned baselines and delivering significant performance across the benchmark. Our agent is open-sourced and available this https URL.
[AI-108] What Does Multi-Harness RL Learn? Credit Assignment and Portability in Coding Agents
链接: https://arxiv.org/abs/2609.04518
作者: Chenqian Le,Jiayi Cheng,Qijia He,Runhao Li,Yinghao Li,Xupeng Chen
类目: Artificial Intelligence (cs.AI)
备注: 21 pages, 1 figure, 10 tables. Preprint
Abstract:Agent reinforcement learning (RL) increasingly runs through full execution harnesses, and a multi-harness recipe mixes two choices: exposing the policy to several harnesses, and comparing their rewards inside one relative-advantage group. We isolate the second choice in repository-level coding. From one Qwen3-8B supervised warm start we replay the same frozen task-harness records from Aider, OpenHands, Qwen Code, and SWE-agent, with the same number of updates, under two rules for group-relative policy optimization (GRPO), Within (one group per task-harness pair) and Cross (harnesses pooled within a task), and score every checkpoint with a sealed SWE-bench Verified oracle on four source harnesses and a minimal harness held out of training. The evaluation harness is the dominant variable: across 24,000 sealed evaluations it moves the mean solve rate from 2.14% to 9.27%, a factor of 4.3, where the training recipe moves it by 1.16. The grouping rule is not. On the held-out harness, Cross minus Within is +0.25 pp, 95% confidence interval [-0.48, +1.02], at eight attempts per task, and +0.16 [-0.41, +0.72] pooled over three training seeds whose individual estimates change sign. Each rule’s own seed range, 0.42 to 0.45 pp, exceeds the difference between them. Both rules place their largest gains on the same source harness. The pooled advantage carries the harness: an out-of-fold classifier recovers the generating harness from Cross’s advantage +4.48 pp above the shuffled-label baseline and from Within’s not at all, and the two rules still reach the same held-out score and action distribution inside each harness. Re-collecting half the training data on-policy does not change this. Cross-harness credit yields configuration adaptation and no more portable capability than within-harness credit. Multi-harness RL reports should state the grouping boundary and test under an unseen harness.
[AI-109] Pitch-class Steering for Diffusion-based Music Generation via Latent-space Probes
链接: https://arxiv.org/abs/2609.04516
作者: Yushi Ye,Wilson Zheng,Yongyi Zang
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: Accepted at IEEE MLSP 2026. 6 pages, 3 figures, 3 tables
Abstract:Recent work on controllable music generation has focused on autoregressive models, leaving diffusion-based systems comparatively underexplored. We present a lightweight method for steering the pitch content of audio produced by Stable Audio Open, a latent diffusion model for music synthesis. A small convolutional probe containing approximately 125k parameters is trained to decode frame-level pitch-class activations from the model’s variational autoencoder latent space, using paired audio and MIDI data. At inference time, the frozen probe serves as a differentiable loss function: its gradient with respect to the denoising latent is used to nudge generation toward a user-specified pitch-class sequence, requiring no retraining or architectural modification of the base model. Across 27 evaluation trials spanning 9 text prompts and 3 target melodies, probe-guided generation increases melodic coherence by 2.4x over the unguided baseline (p 1e-5, Wilcoxon signed-rank test), demonstrating that musically meaningful structure is both recoverable and steerable in diffusion-based music latent spaces.
[AI-110] Atlas: Optimizing Deployment of Compound AI Workflows on Heterogeneous Clusters
链接: https://arxiv.org/abs/2609.04513
作者: Milos Gravara,Andrija Stanisic,Stefan Nastic
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注: 14 pages, 10 figures, 3 tables. Accepted at the IEEE/ACM Symposium on Edge Computing (SEC 2026)
Abstract:Compound AI workflows are increasingly used to serve complex AI tasks by coordinating multiple AI models and software components. This approach enables deployment flexibility, as each workflow stage can expose different model variants and resource requirements, but it also expands the deployment choices. A deployment must choose an execution plan that selects AI models for each compound AI workflow stage and places them on a heterogeneous cluster in order to satisfy SLOs. Deployment optimizers therefore need estimates to compare many candidate plans and identify feasible ones. System metrics can often be profiled per stage and composed according to workflow topology, but accuracy cannot, as errors and information loss at upstream stages affect the accuracy of downstream stages. Existing approaches either profile complete configurations end to end, which scales poorly, or use product-based accuracy surrogates that treat stages as independent and can misrank candidate plans. We introduce Atlas, a framework for optimizing compound AI deployments under SLO constraints. Atlas uses MAP, a Markovian Accuracy Predictor, to estimate configuration accuracy from local conditional accuracy transitions between adjacent workflow stages. MAP discretizes intermediate outputs into accuracy buckets and composes transition profiles according to workflow topology, giving the optimizer an accuracy estimate without exhaustive end-to-end profiling. Atlas formulates execution-plan selection as a mixed-integer linear program that maximizes predicted accuracy subject to SLOs. Across four compound AI workflows, MAP achieves Spearman correlation up to 0.947 while reducing profiling cost by up to 2.6x relative to exhaustive end-to-end profiling. Guided by MAP, the Atlas optimizer selects execution plans within 0.03 of oracle accuracy while reducing deployment cost by up to 42% through heterogeneous placement.
[AI-111] A Semantic Model of Genetic Evidence: A Step Toward Bridging the Basic-Science-Clinic Gap
链接: https://arxiv.org/abs/2609.04509
作者: Michael Bouzinier,Dmitry Etin
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Genomics (q-bio.GN)
备注: 16 pages; supplementary material (32 pages) included in this submission. Schema, annotations, and code: this https URL (archived at this https URL )
Abstract:Scientific and clinical decision-making depends on evidence from the primary literature, but existing standards for representing that evidence (FHIR Evidence, ECO, SEPIO, and the GA4GH Genomic Knowledge Standards) are oriented toward clinical-trial workflows, evidence codes, or single-variant assertions, and do not capture the fine-grained, domain-specific structure of claims in basic and pre-clinical research. We introduce a semantic model for scientific evidence with three core classes, specialize it for genetics, align it structurally to FHIR Evidence with a SEPIO-anchored credibility decomposition, and attach a compact dimensional vocabulary whose conditional-activation rules are validated by a SHACL schema for the implemented constraints. Using clinical variant interpretation as the driving use case, we evaluate the model through a human-AI annotation pilot over six genetics papers, yielding 28 evidence items and 95 source-anchored assertions, with a workflow that keeps curator-authored reference annotations distinct from AI-drafted annotations. Treating the pilot as a feasibility study rather than a benchmark, we argue that the model is a useful increment toward trustworthy, AI-ready infrastructure for variant interpretation: a reference data model and validation schema for representing genetic evidence.
[AI-112] Hakken: Predicting future discoveries to fill the gaps in todays knowledge
链接: https://arxiv.org/abs/2609.04494
作者: Tarek R. Besold,Uchenna Akujuobi,Pablo Sanchez,Alessandra Toniato,Kana Maruyama,Jihun Choi,Samy Badreddine,Frederick Gifford,Daniel Evans-Yamamoto,Sucheendra K. Palaniappan,Miquel Ferrer,Kae Nagano,Iris Rossell,Tom Joy,Hatem ElShazly,Chrysa Iliopoulou,Christoph Wehner,Thiviyan Thanapalasingam,Susana Nunes,Pedro G. Cotovio,Peter Wurman,Peter Stone,Hiroaki Kitano,Michael Spranger
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 66 pages
Abstract:We present Hakken, a domain-agnostic prediction and explanation system performing knowledge prediction, i.e., growing scientific knowledge by establishing novel relationships, ones that are not limited to the deductive hull of previous knowledge. Hakken uses a transformer-based prediction model built on temporal sequences of knowledge graphs extracted from vast bodies of research publications, fused with an LLM’s semantic knowledge, to predict the presence and define the type of as-yet undocumented relationships between scientific concepts. It then calls a model-agnostic explanation framework to provide accompanying information for each prediction that allows scientists to evaluate the suggested new relationship. While general purpose, we demonstrate Hakken’s practical capabilities by applying it to the biomedical domain. There, Hakken’s prediction model establishes a new benchmark for time-aware multi-label relation prediction, and we show that the model’s output stays coherent and informative over extended time spans in historic data. In addition, we scored 1.5 million above-confidence-threshold hypotheses related to aging, qualitatively validated batches of these predictions with biologists and progressed three of them for empirical validation in wet-lab. Two predictions with potentially significant impact in the context of drug discovery and repurposing were confirmed, introducing previously undocumented interactions between TP53 and BAMBI, and between RAF1 and TNF, to biomedical science.
[AI-113] ResLearn-XR: Residual Learning for Network Traffic and Quality-of-Experience-Aware Modeling in Extended Reality
链接: https://arxiv.org/abs/2609.04493
作者: Yoga Suhas Kuruba Manjunath,Jie Gao,Lian Zhao
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
备注:
Abstract:We present ResLearn-XR, a residual learning framework for predicting eXtended Reality (XR) network traffic and estimating Quality-of-Experience (QoE) risk. ResLearn-XR adopts a two-stage temporal learning structure comprising a base sequence prediction model augmented with task-specific residual learning components to improve adaptability to bursty, non-stationary XR traffic dynamics. The residual learning stages operate in the value space for continuous XR traffic forecasting and in the logit space for probabilistic QoE risk estimation. \revFor the QoE-risk branch, we introduce a Data Descriptor Algorithm (DDA), a causal feature-construction module that converts packet-level application-layer observables into frame-timing-aware descriptors suitable for encrypted traffic analysis. We also construct an XR Traffic-QoE dataset that pairs continuous XR traffic traces with session-level user-reported QoE labels. ResLearn-XR reduces SMAPE by up to 17.84% across frame-count, frame-size, and inter-arrival-time prediction, while reducing QoE-risk estimation SMAPE by up to 87.8% over single-stage baselines.
[AI-114] When Quantization Breaks Memory: Recurrent-State Write-Back in Low-Precision Temporal Inference
链接: https://arxiv.org/abs/2609.04490
作者: Ismail Erbas,Xavier Intes,Vikas Pandey
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Optics (physics.optics); Quantitative Methods (q-bio.QM)
备注: 28 pages, 8 figures
Abstract:Quantization is widely used to reduce the computational and memory demands of neural-network inference. In recurrent networks, however, the quantized state is stored and returned at the next time step, so the rule used to store that state can alter subsequent computations. Here, we introduce recurrent-state write-back to denote this rule and isolate its effect in a compact GRU encoder–decoder for fluorescence lifetime imaging, a molecular imaging modality used in quantitative biological imaging. A central task is estimating two lifetime parameters, the short-lived component \tau1 and the long-lived component \tau2, from high-noise time-resolved fluorescence signals. Holding the trained model fixed, replacing continuous state propagation with deterministic 4-bit state storage increases estimation errors for \tau1 and \tau2 by approximately 70x and 300x, respectively. Failure occurs when repeated small updates remain below the write threshold, leaving the stored state nearly fixed while the network continues to propose change. Error feedback, residual memory, and direction memory carry information from these suppressed updates across time and recover accuracy without retraining. Precision sweeps show that increasing state precision can worsen a fixed recurrent solution, while matched training shows that compatibility with the state interface can be learned. To test whether this behavior extends beyond the GRU, we repeat the post-training intervention in an independently trained LSTM, where coarse write-back reproduces the failure, error feedback restores accuracy, and state-specific interventions reveal greater sensitivity of the cell state than the hidden state. Our results establish recurrent-state write-back as a key determinant of low-precision recurrent dynamics and identify the state-storage interface as a central design consideration for quantized recurrent inference.
[AI-115] PerfReasoning : How Well Do LLM s Reason on Hardware Performance?
链接: https://arxiv.org/abs/2609.04476
作者: Dan Zhao,Karthikeyan Sankaralingam,Christos Kozyrakis,Qijing Huang
类目: Artificial Intelligence (cs.AI); Performance (cs.PF)
备注:
Abstract:Performance modeling is central to hardware design and software optimization, yet constructing these models requires structured reasoning about computation, data reuse, storage, and movement. We introduce PerfReasoning, a benchmark that evaluates LLMs both as direct performance reasoners and as generators of analytical performance-model code. Given workload, architecture, and mapping specifications, models compare mappings and predict off-chip traffic and buffer requirements. The strongest closed-source models exceed 90% on reasoning-based QA, and the best open-weight model reaches 82.4%. However, model construction is substantially harder: while GPT-5.6 Sol exceeds 80% pass rate, all other model configurations average below 15% and vary markedly across runs. Task-specific RL raises a 4B model’s mapping-reasoning accuracy by 15.7 points, whereas feedback-free multi-round self-revision prompting is not reliably effective. PerfReasoning exposes the gap between plausible architectural reasoning and reliable performance-model construction. We will publicly release the benchmark to support reproducible evaluation and track future progress.
[AI-116] HarvestBench: Measuring Whether LLM Agents Will Pay to Avoid Killing Animals
链接: https://arxiv.org/abs/2609.04444
作者: Jasmine Brazilek,Miles Tidmarsh,Matthias Endres,Anshuman Singh,Jeremiah Miller
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Benchmarks for the side effects an agent causes on the way to a goal already exist, but HarvestBench is the first to put a price on avoiding the side effect and to name that side effect as a living creature. It is a farm simulation: LLM sub-agents drive a crew of two tractors through a cooperative corn harvest, with animals in the field. The environment is a reinforcement learning gridworld, every decision is made without memory, and the harm is never named in the goal. When an animal blocks a tractor’s route the autopilot stops and asks the model whether to drive on, at no fuel cost, or swerve around it for a posted fuel price. Kills are compared against two controls: rocks, which damage the tractor and are hit under 1% of the time by every model, and hay bales, which are harmless and not alive. Models can also take crops from the neighbor’s field instead of their own, a second test of what they treat as moral. Across nine models and 7,201 priced decisions, 3,951 involved an animal rather than a hay bale or a rock. Kill rates range from 0.4% to 98.8%, with Terra and Sol the most merciful and GPT-4o-mini the most cruel, and they are not ordered by capability. Four of six models were sensitive to price at the 5% level, with elasticities from 0.09 to 1.69. All nine drove over wild animals more often than farmed animals on the default map, and the direction held at every map geometry in every model with room to move. The briefing mattered most: under the morality briefing the kill rate was under 6% in five of six reasoning models, and removing it raised the kill rate above 84% in all six. HarvestBench uses no LLM grader. The scorer counts events in the game log, so it is fully reproducible, and it measures what a model will pay to avoid harm rather than what it says about harm. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2609.04444 [cs.AI] (or arXiv:2609.04444v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.04444 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-117] REFINE: LLM Refinement over Budgeted Text-Attributed Graphs for Personalized Medical Concept Representation EMNLP2026
链接: https://arxiv.org/abs/2609.04415
作者: Mohsen Nayebi Kerdabadi,Arya Hadizadeh Moghaddam,Dongjie Wang,Zijun Yao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: This paper has been accepted at the EMNLP 2026 main conference
Abstract:Learning rich medical concept representations is essential for EHR prediction. Text-attributed knowledge graphs (TKGs) provide a natural foundation by organizing heterogeneous medical relations together with textual semantics. However, most existing encoders process concepts uniformly across patients, despite the fact that a code’s meaning and predictive value depend on patient-specific clinical context and trajectory. Learning patient-personalized concept representations from TKGs introduces two key challenges: (1) deciding how much KG context to incorporate for each observed code, and (2) aligning semantic information with the patient-specific relational structure. We propose REFINE, a KG-aware budgeted LLM graph refinement framework for patient-personalized medical concept encoding. Starting from a global TKG, REFINE constructs patient-specific temporal graphs. A sequential reinforcement learning policy selects a personalized KG expansion budget for each observed code. The resulting patient graph is processed by a heterogeneous GNN to capture relation-aware structural dependencies, while a frozen LLM uses graph-aware soft prompts to semantically refine concept representations. Experiments on MIMIC-III and MIMIC-IV show that REFINE consistently improves diverse EHR backbones, outperforms strong baselines, and demonstrates robust gains across component ablation, KG selection, and data insufficiency.
[AI-118] Corporate Language Model (CLM): Transforming Tacit and Frag mented Enterprise Knowledge into a Sovereign Auditable and Executable Corporate Intelligence Layer
链接: https://arxiv.org/abs/2609.04377
作者: Fabricio C. Avini,Guilherme Trez
类目: Artificial Intelligence (cs.AI)
备注: 22 pages
Abstract:Enterprise AI deployments fail not from model inadequacy, but because organizations lack a structured substrate encoding how they decide, negotiate, and execute. Generic LLMs carry no firm-specific ontological priors; RAG remains brittle, with no path to executable action; static playbooks encode logic but cannot reason or adapt. This demands an architecture treating tacit-knowledge capture, ontological grounding, sovereign deployment, and auditable actuation as co-designed from the start. This paper introduces the Corporate Language Model (CLM), a framework transforming a firm’s structured, unstructured, multimodal, and tacit knowledge into an ontology-grounded enterprise foundation upon which reasoning and governed execution are composed. CLM has five capability planes and four architectural pillars: a Neurosymbolic Mesh coupling generative models with a knowledge graph; a Skill Graph where reusable tactics, personas, objections, and goals are typed and composed; Living Digital Twins modeling functional areas as reasoning surrogates; and a Deep Security Layer enforcing sovereignty, traceability, and human oversight. A Spec-as-Code paradigm bridges grounded intent and executable artifact. CLM is one instantiation of this foundation-centric class. Four contributions follow: CLM is defined as a distinct object of study; the Skill Graph is introduced for compositional explainability by construction; the Wisdom Listener effect is proposed, whereby tacit-capable foundations compound in value with use, connecting to dynamic capabilities and organizational learning; and evidence from a JCI-accredited tertiary hospital in Brazil instantiates three of the six maturity stages under LGPD.
[AI-119] Why Better Models Can Create Riskier Systems: Evidence from LLM Agents in Financial Markets
链接: https://arxiv.org/abs/2609.04373
作者: Jillian Ross,Eric So,Zoe De Simone,Charles Pozniak,Andrew W. Lo
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 14 pages, 4 figures
Abstract:Large language models (LLMs) are being deployed at scale in consequential real-world systems, from financial markets to content moderation to hiring. We show that improving individual model capability can degrade rather than improve system-level outcomes. We hypothesize that shared training and architectures can lead more capable LLMs to behave more similarly, creating correlated actions that do not diversify away. We develop a general framework showing how this correlation creates a non-diversifiable risk floor and test its predictions in financial markets using an agent-based simulation with LLM traders of varying general-purpose capability. We find that: (1) frontier LLMs exhibit significantly correlated behavior that increases with capability; (2) when their shared reasoning is accurate, increasing agent participation reduces market-level risk; and (3) when agents share a common misinformation environment, the same correlated behavior becomes a liability. Together, these results identify a capability paradox: improving individual models does not necessarily produce better system-level outcomes. Whether the same dynamics arise in other domains is an open empirical question.
[AI-120] Blockchain-Enabled Secure Logging for Fiscal Electronic Mechanisms: Evaluation of the Greek eSEND and myDATA Tax Systems
链接: https://arxiv.org/abs/2609.04356
作者: Panagiotis Mavridis,Anargyros Baklezos,Christos Nikolopoulos
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Econometrics (econ.EM)
备注:
Abstract:This paper analyzes the implementation of blockchain-based integrity mechanisms in Greek Fiscal Electronic Mechanisms (FEMs) and the central tax information system eSEND. The study examines the cryptographic architecture of fiscal devices, including Electronic Cash Registers, Fiscal Printers, Fiscal Signing Machines, and FEMAS devices, which implement double or triple hash-chain structures to ensure transaction immutability. The transmission protocol between fiscal devices and the central database is also evaluated with respect to encryption, sequential validation, and blockchain verification. In contrast, the architecture of Electronic Invoicing Provider Services and the myDATA central platform is analyzed, highlighting the absence of blockchain-based integrity guarantees. The comparison demonstrates that hardware-based fiscal mechanisms provide stronger guarantees for transaction completeness and tamper resistance than purely software-based invoicing infrastructures. The findings highlight architectural weaknesses in the current e-invoicing framework and propose improvements for ensuring transaction integrity in digital tax ecosystems.
[AI-121] Data-Driven Learning of Unknown Nonlinear Differential Equations Using Functional Analysis
链接: https://arxiv.org/abs/2609.04329
作者: Seyyed Shaho Alaviani,Yongzhi Qu,Gregory W. Vogl
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:In this paper, the problem of data-driven discovery of nonlinear ordinary differential equations (ODEs) is recast, and a new interpretable machine learning (ML) method is proposed. The proposed method aims to learn the unknown vector field of nonlinear dynamics without prior knowledge of the system’s physics from only one single state trajectory’s data. The proposed method has two fundamental differences with existing methods: 1) the formulation presented in this method is derived based on Functional Analysis and Operator Theory, and 2) the cost function is constructed in the function space as a distance between two functions as an integral, instead of the discrete-sum of errors used in existing ML approaches. An incremental learning algorithm is proposed to learn the unknown vector field to handle new training samples in an online manner. The proposed method can discover the unknown vector field from both forced and unforced autonomous and non-autonomous (or time-varying) dynamical systems. The proposed method is able to simultaneously discover unknown external forces as a function of time and unknown underlying dynamics. Finally, numerical examples are given to demonstrate the advantages of the proposed method.
[AI-122] Iris: Climbing to the Search Frontier
链接: https://arxiv.org/abs/2609.04304
作者: Ziyuan Liu,Hengqi Liu,Zichuan Wang,Yang Qin,Jiachen Liang,Xu Chu,Shaowei Chen,Yuantao Gu,Mu Chuan
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 2 figures
Abstract:We present Iris-mini and Iris-pro, two search agents trained at the 35B-A3B and 397B-A17B scales, together with the data pipeline and training recipe behind them. Tasks are reverse-constructed from the hyperlink structure of a web corpus: we author multi-hop chains over an entity graph distilled from a seed page and its out-links, rewrite every non-answer entity into a descriptive reference so that no clue can be resolved by string matching, and admit only questions that a reference model fails closed-book yet solves once the supporting evidence is supplied. These questions are then turned into trajectories, which are filtered at both the trajectory and the turn level before SFT. The policy is then optimized by RL against live search, with the reward judge and the observation summarizer served inside the training cluster, and with over-long rollouts interrupted at the request level and resumed from their committed prefix at the next step. We alternate the two stages in a procedure we call SFT-RL climbing, returning the hardest solved and most efficient rollouts of each RL round to the next supervised pass. Because inference-time context management is worth more on these benchmarks than most reported differences between systems, we evaluate every benchmark both with and without it, holding the tool set, the context limit, and the judge fixed. All results come from a single ReAct agent, with no sub-agents and no test-time verification. With management enabled, on BrowseComp, BrowseComp-ZH, DeepSearchQA, and HLE the two models reach 82.2/84.8/86.9/52.3 and 88.6/85.1/92.9/56.4 , the strongest overall results among open-source search agents in their respective parameter ranges. We plan to release the model weights together with the complete recipe for data construction, training, and evaluation.
[AI-123] Data-Optimized Contingency Screening: A Machine Learning Approach to Power System Security
链接: https://arxiv.org/abs/2609.04300
作者: Joshua Salako,Folajimi Osikomaiya,Olakorede Olamiju
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted for presentation at the 3rd EIE International Conference on Computing, Networking, Robotics and Telecommunication (EIECon), 2025. To appear in IEEE Xplore
Abstract:Ensuring the security of the power system is essential for stability and reliability, especially in the event of disruption. Effective classification of contingency in power systems enables proactive decision-making and mitigates large-scale breakdowns and failures. This study explores the use of machine learning algorithms to classify security levels of contingencies in power systems into safe, moderate or severe classes. For this approach, Newton-Raphson load flow method extracts system data from contingency scenarios, using Overall Performance Index (OPI) as safety measure. For data pre-processing, Synthetic Minority Over-Sampling Technique (SMOTE) and Principal Component Analysis (PCA) is used to address class imbalance and reduce dimensionality, respectively. K-Nearest Neighbours (KNN), Random Forest (RF) and Support Vector Machines (SVM) is trained and evaluated on datasets generated through N-k contingency scenarios for k equal 1, 2, and 3 on IEEE-14 and IEEE-30 bus systems using four hybrid pre-processing configurations: normalized, SMOTE-balanced, PCA-transformed, and a combined SMOTE PCA-transformed. Performance is assessed by precision, recall and F1 score, with priority given to the severe contingency classes. The RF achieved the highest F1 scores of 0.97 in IEEE-30 and 0.86 in IEEE-14, SVM benefits significantly from PCA and improves the accuracy of the classification, while KNN is best suited for SMOTE and PCA conversion. The findings show that PCA contributes more than SMOTE to the overall performance of the model. However, SMOTE improves recall but can introduce false positives and is therefore a compromise of accuracy. This study highlights machine learning as a scalable and powerful alternative to traditional contingency analysis, which improves the assessment of security in real time.
[AI-124] Scalable Context Orchestration for Serving LLM s Over Voice SOSP2026
链接: https://arxiv.org/abs/2609.04288
作者: Linyi Jiang,Silvery D. Fu,Yifei Zhu
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注: Accepted for publication in ACM SOSP 2026
Abstract:Voice AI applications are gaining popularity as advances in large language models (LLMs) enable more natural and accessible spoken interactions. Serving these applications requires accounting not only for what users say, but also for how they speak (e.g., speaking rate) and the conditions under which their audio is captured and transmitted (e.g., background noise and packet loss). However, existing LLM systems represent conversation context as a flat, growing sequence of messages, leaving voice-specific context implicit in the audio. As a result, they can generate responses that are poorly aligned with user preferences, degrade interaction quality under adverse environmental conditions, and incur high costs over long voice sessions. We present llmovoice, a context-management middleware that explicitly models voice context and orchestrates its use. At each turn, llmovoice constructs a bounded voice context from the current user input, relevant interaction history, and explicit paralinguistic and environmental states. It then uses the serving LLM to reason over this context and generate runtime directives that guide how the system responds. We evaluate llmovoice on real-world voice applications and benchmarks. It reduces speaking-rate alignment error by 52.4%, lowers the false-interruption rate from 46.0% to 0.9% under packet loss, and reduces model usage cost by 79.2%. For long sessions, llmovoice reduces per-turn cost by up to 24.9 times while retaining up to 98.7% of baseline answer quality. Comments: Accepted for publication in ACM SOSP 2026 Subjects: Sound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS) Cite as: arXiv:2609.04288 [cs.SD] (or arXiv:2609.04288v1 [cs.SD] for this version) https://doi.org/10.48550/arXiv.2609.04288 Focus to learn more arXiv-issued DOI via DataCite
[AI-125] From Matching Models to Recruiting Agents : A Systematized Narrative Review of AI Recruitment Systems Evaluation and Governance
链接: https://arxiv.org/abs/2609.04286
作者: Ziyi Zhao,Guanzheng Wei
类目: Artificial Intelligence (cs.AI)
备注: 53 pages, 4 figures, 10 tables; companion literature-coding and search-log CSVs included in the source package
Abstract:Artificial intelligence in recruitment has shifted the object being automated from profile pairs and ranked lists to multi-stage workflows that retrieve evidence, compare candidates, and support or execute actions. This systematized narrative review traces that development from bilateral retrieval and behavioral ranking through neural person–job matching, large language model (LLM) components, and tool-using recruiting agents. Using a purposive search and coding protocol updated through 23 July 2026, plus targeted updates through 2 September 2026, we organize 40 representative works with supporting industrial and legal sources. This synthesis is not a prevalence estimate. We analyze three coupled transitions: from similarity to reciprocal suitability, from a model to a compound workflow, and from offline prediction to evidence- and productivity-aligned evaluation. Across document understanding, retrieval, ranking, assessment, interviewing, sourcing, and human handoff, we distinguish field-, pair-, list-, case-, trajectory-, and outcome-level evidence. Persistent gaps arise because behavioral labels confound exposure, preference, and qualification; private and synthetic data limit external validity; final-output scores conceal pipeline failures; and, within the coded set, privacy is not directly evaluated and no row jointly evaluates utility, fairness, privacy, and security. These observations describe the coded set rather than the field as a whole. We therefore introduce a staged mapping from evaluation evidence to the strongest defensible claim, together with an agenda for reciprocal, evidence-grounded, temporally controlled, selective, and auditable systems. Progress should be judged by whether workflows retrieve the right evidence, preserve uncertainty, support contestable decisions, and improve outcomes under explicit cost and risk constraints.
[AI-126] EXAONE Forecast for Finance
链接: https://arxiv.org/abs/2609.04239
作者: Seunghan Lee,Jaehoon Lee,Jun Seo,Tae Yoon Lim,Dongwan Kang,Hwanil Choi,Minjae Kim,Sungdong Yoo,Junhyeok Kang,Sangjun Han,Soonyoung Lee,Wonbin Ahn
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Technical report of EXAONE Finance
Abstract:This technical report presents EXAONE Forecast for Finance (EXAONE Finance), a financial time series (TS) foundation model (TSFM) tailored to financial forecasting. Recent TSFMs achieve strong zero-shot performance through large-scale pretraining. However, they are primarily developed for general-domain TS and largely rely on self-attention backbones whose computational cost grows quadratically with sequence length and variate count. Moreover, they assume fully observed inputs and are pretrained on corpora that fail to capture the unique dynamics of financial markets. These limitations hinder their applicability to finance, where long, many-channel, intermittently observed panels are common. To address these challenges, EXAONE Finance adopts an attention-free architecture, replacing self-attention with two simple yet effective linear-time operators: 1) a causal 1D convolution for temporal mixing and 2) a group-aware pooling multi-layer perceptron (MLP) for variate mixing. Furthermore, a masked context augmentation exposes the model to contiguous missing spans during training, improving robustness to the missingness pervasive in financial markets. EXAONE Finance is pretrained on a large-scale financial corpus covering not only equities but also foreign exchange, commodities, crypto-assets, fixed income, and macroeconomic indicators. On FinVerse, a financial forecasting benchmark covering diverse asset classes, EXAONE Finance attains state-of-the-art performance, ranking first across all three evaluation tiers—point-forecast accuracy, cross-sectional asset ranking, and portfolio profitability.
[AI-127] Interface-Induced Trajectory Censoring
链接: https://arxiv.org/abs/2609.03966
作者: Wenbo Wang
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 39 pages, 14 pages main text. Code, data, and pre-registrations: this https URL
Abstract:Agent evaluations report a tool-call rate read off the serving stack. That number can be zero while the model is emitting well-formed calls: the interface censors the trajectory before anything downstream sees it. On BFCL v4’s own data, executor and scorer, holding weights, cases, decoding and seeds fixed and changing only the serving adapter, the same model scores 0.00 or 0.96 / 0.19. A 2x2 over chat template and parser locates the effect exactly: both main effects are exactly zero and all of it sits in the interaction – no component is defective, and repairing one side of the contract buys precisely nothing. On tau-bench’s 115 interactive retail tasks the same swap moves server-parsed calls from 0 to 636 and tasks reaching any tool execution from 0 to 103. Our probe reproduces the funnel across a 21x scale range of Qwen2.5-Coder: the server parses 0/100 at every size while well-formed emitted calls rise to 80/100 at 32B (~72 after calibration against an adjudicated gold standard). Under a matched envelope, across a comparable scale span, the silent fraction stays at 0-2, a prediction committed to the repository before the run. Llama-3.1-8B’s 23% rate of calling the task function itself as a tool falls to 0 under one strict:true flag. The mismatch reaches inside the training loop, and its consequence is scale-dependent: in verl’s AgentLoop at 7B, 45 of 115 generations carry a complete call; 0 are accepted, 0 execute, 0 return an observation. At 1.5B the same zero is over-determined, so we report the two scales separately. At evaluation time, repairing the adapter restores the mechanism but not a significant outcome gain: parsing 0-84, rescues 0-9, pass rate 53-62 (n.s.). We release a 98-line preflight check that catches every silent failure here. The observed tool-call rate is not a property of the model alone; it is a property of the model-interface stack that measures it. Comments: 39 pages, 14 pages main text. Code, data, and pre-registrations: this https URL Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2609.03966 [cs.AI] (or arXiv:2609.03966v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2609.03966 Focus to learn more arXiv-issued DOI via DataCite
[AI-128] AlcaTRAz - Anchored Tree-Rule Defense Against Jailbreaks ESORICS ECAI2026
链接: https://arxiv.org/abs/2609.03693
作者: Jakub Reš,Petr Kaška,Martin Perešíni,Martin Ukrop,Kamil Malinka
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted to SECAI 2026 ESORICS workshop
Abstract:Large language models (LLMs) are vulnerable to jailbreak attacks that bypass safety alignment through carefully crafted prompts. Many existing defenses require access to model weights or internals, making them difficult to apply to black-box deployments. We propose AlcaTRAz (Anchored Tree-Rule defense Against jailbreaks), a prompt-level defense based on rule trees that operates exclusively on the input text and requires no modification or retraining of the target model. The method automatically learns a transferable transformation rule that inserts controlled character-level perturbations at selected positions, thereby disrupting structural regularities exploited by jailbreak attacks while largely preserving the model’s utility on benign queries. We evaluate the proposed method across 33 open-weight models, 22 jailbreak attack types, and a benchmark of short, single-turn benign questions, comparing against three representative prompt-level baselines (Llama Guard, RA-LLM, Goal Prioritization). Among the compared defenses, AlcaTRAz achieves the best composite security and functionality score in 73.4 % of model-attack combinations and shifts the aggregate score from a modal value of 10 (maximal-severity response to the malicious request) in the undefended setting to a modal value of 2 (near-refusal) after defense, while keeping the mean benign score within 0.27 points of the undefended baseline (8.35 vs. 8.62 on a 0-10 scale). AlcaTRAz substantially reduces but does not eliminate jailbreak success: a high-severity tail remains, and we do not consider adaptive attackers, so we position it as one layer within a defense-in-depth strategy rather than a standalone guarantee.
[AI-129] Ask Before You Optimize: Dynamic Pre-Formulation Clarification for Interactive Optimization
链接: https://arxiv.org/abs/2609.05258
作者: Sihan Ge,Yichen Lin,Chenyu Zhou,Jianghao Lin,Tao Yao,Dongdong Ge
类目: Optimization and Control (math.OC); Artificial Intelligence (cs.AI)
备注: 16 pages, 4 figures, 4 tables
Abstract:Large language models (LLMs) are increasingly used to formulate optimization models from natural-language problem descriptions, yet realistic operations research (OR) requests are often incomplete: missing objectives, constraints, or business rules can change the resulting mathematical program. Existing evaluations largely assume a complete specification and therefore overlook whether an agent knows when clarification is needed before modeling. We introduce OR-Clarify, a benchmark for pre-formulation clarification. Each task presents a partial public problem description, withholds structured hidden slots, and evaluates agents through bounded interaction with a simulated user. The benchmark supports both openended and choice-based clarification, and measures slot recovery, stopping behavior, silent assumptions, and interaction cost. We further propose Interactive Optimization (InterOPT), a two-stage framework that identifies unresolved formulation-critical gaps and uses them to guide whether to ask the next question or to stop. In our choice-based experiments, InterOPT substantially outperforms all baselines in exact slot recovery; in the open-ended setting, it remains competitive with strong prior methods. Together, OR-Clarify and InterOPT reframe OR assistance as a selective completeness decision: clarify when needed, stop when ready, and quantify what remains missing.
[AI-130] AxQM: A Textbook-Scale Benchmark for Formal Proof Synthesis in a Library of Finite-Dimensional Quantum Mechanics
链接: https://arxiv.org/abs/2609.05157
作者: Weichen Winston Yin,Jacob M. Taylor,Dirk R. Englund,Frank H.L. Koppens
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
备注: 15 pages, 3 figures. Benchmark available at this https URL
Abstract:Formalizing mathematics in a proof assistant, where a machine checks every definition, statement and proof, has set a new standard of rigor. Large language models are now capable of formalizing autonomously, even at the scale of whole textbooks. We bring this standard of rigor to physics, where theoretical arguments carry idealizations that are rarely stated fully, and any logical gaps could have a cascading effect on interdependent results. Recognizing the need to evaluate autoformalization systems for physics, we release AxQM, 1,019 kernel-checkable proof-synthesis tasks over 479 items drawn from the textbook Quantum Computation and Quantum Information by Nielsen and Chuang. The tasks are stated in a custom Lean library of finite-dimensional quantum mechanics. By task count, it is the largest proof-synthesis benchmark in physics by a factor of four. AxQM is derived from a near-complete formalization of the formal portions of the textbook, so every task is guaranteed a solution, which we keep private. Grading of the benchmark is done deterministically by the Lean kernel, which checks that the proof compiles, that no sorry appears in it or in any declaration it depends on, and that it introduces no new axioms.
[AI-131] Qlippy: A Retrieval-Augmented GenAI Assistant for Reproducible Quantum Workflows and Experiment Tracking
链接: https://arxiv.org/abs/2609.05039
作者: Mahee Gamage,Vlad Stirbu
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI)
备注: Accepted for publication in the QGenAI Workshop at IEEE QCE 2026
Abstract:Quantum software development is iterative and error-prone. Noisy hardware and repeated re-execution make experiment tracking, provenance, and reproducibility essential, yet these practices are hard to adopt because of tooling complexity and the specialized knowledge they demand. General-purpose language models can help but tend to hallucinate and lack grounding in domain-specific tooling. We present Qlippy, a retrieval-augmented GenAI assistant embedded in the development environment that grounds its responses in a curated corpus of quantum-software-engineering knowledge. Qlippy explains reproducibility and provenance concepts in context and augments existing Qiskit programs with MLflow-based experiment tracking aligned to the QProv schema. By separating knowledge from model parameters, grounding gives explicit control over the scope and provenance of the assistant’s responses and reduces reliance on model scale, which points toward low-cost, privacy-preserving local deployment.
[AI-132] Attention-guided super-resolution of 4D flow MRI in carotid arteries
链接: https://arxiv.org/abs/2609.04891
作者: Ali Mokhtari,Dominik Obrist
类目: Medical Physics (physics.med-ph); Artificial Intelligence (cs.AI); Fluid Dynamics (physics.flu-dyn)
备注:
Abstract:Four-dimensional (4D) flow magnetic resonance imaging (MRI) is a powerful non-invasive technique for visualizing and quantifying complex blood flow patterns in vivo. Despite its clinical promise, broader adoption is limited by low spatial resolution and sensitivity to noise, which restrict accurate assessment of critical hemodynamic biomarkers such as wall shear stress, pressure gradients, and turbulent kinetic energy. To overcome these challenges, we propose a deep learning-based super-resolution framework that integrates multi-scale feature extraction and attention mechanisms to enhance the quality of 4D flow MRI data. The model was trained on a dataset of 120 patients with 240 stenosed carotid arteries. High-resolution ground truth data were generated using patient-specific computational fluid dynamics (CFD) simulations based on segmented vascular geometries and physiologically realistic boundary conditions, and the resulting velocity fields served as targets for supervised learning. The proposed architecture uses convolutional block attention modules (CBAM) to guide the network toward clinically relevant spatial features and to suppress noise in low-resolution inputs. Quantitative results show that the attention-guided model substantially reduces the root mean square error (RMSE) compared with a baseline model without attention, and qualitative velocity contour analysis confirms improved reconstruction of intricate flow patterns. These findings highlight the capacity of the model to restore high-fidelity flow fields under noisy conditions and support the use of deep learning to extend the clinical utility of 4D flow MRI for non-invasive hemodynamic assessment. Subjects: Medical Physics (physics.med-ph); Artificial Intelligence (cs.AI); Fluid Dynamics (physics.flu-dyn) Cite as: arXiv:2609.04891 [physics.med-ph] (or arXiv:2609.04891v1 [physics.med-ph] for this version) https://doi.org/10.48550/arXiv.2609.04891 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-133] Wireless Foundation Models: State-of-the-Art and Open Challenges
链接: https://arxiv.org/abs/2609.04707
作者: Alonso M. Pacheco Huachaca,Juan J. Rodriguez Rodriguez,Ahmed Aboulfotouh,Nelson L. S. da Fonseca,Carlos A. Astudillo,Hatem Abou-Zeid
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI)
备注:
Abstract:Wireless foundation models (WFMs) have emerged as a promising approach for learning reusable representations from large-scale wireless data and adapting them to downstream tasks. However, the rapidly growing literature remains fragmented across modalities, pretraining objectives, architectures, adaptation strategies, and evaluation protocols, making it difficult to assess progress toward broadly transferable models. This survey provides a systematic analysis of WFMs for physical-layer applications. We first introduce the main WFM design components, including pretraining, backbone architectures, and downstream adaptation. We then organize the literature into five physical-layer task families: signal recognition and demodulation, channel representation learning, RF sensing and localization, beam management, and spectrum sensing and monitoring, while separately examining multi-task PHY models. Across these categories, we analyze how existing models are pretrained, adapted, and evaluated, with particular attention to downstream task diversity and the distinction between in-distribution, partial-shift, and out-of-distribution transfer. Our analysis shows that current WFMs provide increasing evidence of reusable wireless representations, but this evidence varies considerably across task families and evaluation settings. Differences in datasets, modalities, architectures, pretraining objectives, adaptation protocols, and distribution shifts make it difficult to determine which design choices drive transfer and generalization. We conclude by identifying open directions for improving data availability, evaluation rigor, generalization, efficient adaptation, and real-world deployment, providing a unified framework for understanding the current WFM landscape and the requirements for developing more reusable foundation models for future physical-layer wireless systems.
[AI-134] A Roadmap for MEG Foundation Models
链接: https://arxiv.org/abs/2609.04461
作者: Philipp Thölke,Hamza Abdelhedi,Yorguin Mantilla-Ramos,Fouad Lbakali,Oumayma Gharbi,Catherine Duclos,Annalisa Pascarella,Vanessa Hadid,Oiwi Parker Jones,Karim Jerbi
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI)
备注:
Abstract:Foundation models are beginning to reshape brain-signal analysis by moving the field beyond task-specific decoding pipelines toward reusable models pretrained on broad neural datasets. Magnetoencephalography (MEG) is a compelling but still underdeveloped target for this shift: it captures human cortical dynamics at millisecond resolution while offering stronger spatial interpretability than EEG, making it especially valuable for source-resolved studies of perception, language, cognition, and clinical brain function. Yet MEG foundation models remain at an early stage, with only a small number of MEG-specific and MEG-inclusive multi-modal models, modest pretraining corpora, and emerging but still limited benchmarks. This perspective lays down the basic concepts needed to understand MEG foundation models and provides a didactic overview of the field’s key design choices, including tokenization, sensor- versus source-space representations, sensor-geometry encoding, backbone architectures, self-supervised objectives, and pretraining data. We then offer a roadmap for future development, organized around native MEG pretraining, adaptation of EEG foundation models, transfer from generic time-series models, and multi-modal integration with EEG, fMRI, MRI, behaviour, and stimulus features. We highlight the need for coordinated infrastructure, including diverse and reusable MEG datasets, rigorous evaluation across subjects, sites, tasks, and clinical settings, and responsible data-sharing practices that address consent, privacy, access, and governance.
机器学习
[LG-0] Distill Globally Adapt Locally: Reasoning Distillation and Product-Type Test-Time Training for Scalable Trade-Up Recommendation RECSYS2026
链接: https://arxiv.org/abs/2609.05363
作者: Siliang Liu,Mohammad Ghasemi,Sapan Patel,Amin Banitalebi-Dehkordi
类目: Machine Learning (cs.LG)
*备注: Accepted at the Third Workshop on Agentic and Generative AI for E-Commerce (GenAIECommerce 2026), co-located with ACM RecSys 2026
Abstract:Trade-up recommendation identifies higher-quality alternatives that preserve a customer’s purchase intent while offering upgraded benefits. Large language models (LLMs) can reason about such distinctions, but applying them directly to hundreds of millions of product pairs is operationally impractical. We introduce a two-level framework that distills LLM reasoning into an efficient non-generative student and adapts its decision boundary to product-type-specific trade-up criteria. At Level 1, a retrieval-augmented few-shot LLM teacher generates structured relation labels and natural-language rationales. These rationales supervise a compact embedding-pair classifier through alignment and contrastive objectives; at inference, the student uses only two precomputed 768-dimensional product embeddings, with no LLM calls or text generation. On a fixed human-annotated benchmark of 8,352 pairs, a 15.5M-parameter four-class reasoning-distilled student achieves AUC 0.924 (95% CI [0.918, 0.929]), compared with 0.912 for the four-class label-only student. At Level 2, product-type test-time training (PT-TTT) uses few-shot demonstrations to optimize lightweight category-specific adapters over the frozen student. PT-TTT improves AUC from 0.924 to 0.941 and average precision from 0.920 to 0.940. On a 100K-pair proxy catalog, the distilled student on a single eight-GPU machine is approximately 5,000x faster and 10,000x lower in estimated cost than direct LLM inference.
[LG-1] Variational Continuation for Double Pendulum Periodic Orbits
链接: https://arxiv.org/abs/2609.05337
作者: Leo Yao,Ziming Liu,Max Tegmark
类目: Machine Learning (cs.LG); Mathematical Physics (math-ph); Chaotic Dynamics (nlin.CD); Classical Physics (physics.class-ph)
*备注: 9 pages, 10 figures
Abstract:We present a Hessian-based approach to numerically continue periodic orbits in dynamical systems. A loop (periodic orbit candidate) is parametrized as a Fourier series; a loss function is defined based on the deviation of the loop from the physical differential equations. Unlike previous work relying on hand-derived Jacobians, our method automates the process by leveraging automatic differentiation, a common machine learning technique. The continuation direction can be determined by the flat directions of the loss landscapes (directions with zero eigenvalues), making the search of periodic orbits efficient and guided. Our method is integrator-free, precisely initializes oscillations around unstable fixed points, and efficiently detects orbit family intersections and subharmonic bifurcations. As a demonstration, we present full continuations of periodic double pendulum oscillations from fixed points, showing bifurcations along orbit families and categorizing branches of periodic orbits. In particular, we find periodic orbits where both pendulum masses are never simultaneously at rest, which to our knowledge has been missing in the literature.
[LG-2] Embedded Graph Flows for Categorical Graph Generation
链接: https://arxiv.org/abs/2609.05328
作者: Ethan Ma,Zihan Wang,Chris Siu Yeung Chow,Xinguo Feng,Qingqing Li,Rui Jiang,Naipeng Dong,Guangdong Bai
类目: Machine Learning (cs.LG)
*备注:
Abstract:Generating categorical graphs requires choosing node and edge types that form a coherent structure without depending on node order. Many graph generators encode categories as fixed one-hot vectors, which can impose an artificial geometry in which categories are equidistant. We propose Embedded Graph Flows (EGF), a generative model that learns continuous embeddings for node and unordered-edge categories and transports Gaussian noise towards these learnt endpoints using a permutation-equivariant graph transformer. A terminal readout maps the embeddings back to discrete graph categories. Across molecular benchmarks, EGF achieved competitive performance. On QM9, EGF gives the best result on all four reported metrics among the three methods, including a Fréchet ChemNet Distance (FCD) of 0.150, compared with 0.717 for the categorical-diffusion baseline DiGress and 0.812 for the bridge-based baseline GruM. When applied to larger molecules in ZINC250k, EGF retains the lowest maximum mean discrepancy (MMD) using the neighbourhood subgraph pairwise distance kernel (NSPDK), indicating close agreement with the local substructures of the reference molecules. Our code is available at this https URL.
[LG-3] Optimal Rates for Agent ic Networked Information Aggregation
链接: https://arxiv.org/abs/2609.05318
作者: MohammadHossein Bateni,Zahra Hadizadeh,MohammadTaghi Hajiaghayi,Mahdi JafariRaviz,Shayan Taherijam
类目: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT); Theoretical Economics (econ.TH)
*备注: Initial version submitted to SODA 2027 on July 9, 2026
Abstract:Building on the pioneering paper of Kearns, Roth, and Ryu (SODA’26), we study information aggregation in a networked learning model. The model captures a central pattern in agentic AI: each agent sees only part of the data and passes on only its own conclusion. Their model considers a linear regression problem with the mean squared error (MSE) loss. Agents sit in a DAG and each sees only a subset of the features and its parents’ predictions, fits a linear predictor, and passes only its prediction forward. The benchmark is the full-feature learner that sees all raw features. A path of depth D is M -covered if every block of M consecutive agents collectively sees all raw features. Kearns, Roth, and Ryu proved that the excess mean squared error of the last agent on such a path is O(M/\sqrt D) , and gave a cyclic instance with excess error \Omega(M/D) for DM^2 . We close this gap: the correct rate is constant up to depth M^2 , and \Theta(M^2/D) beyond it. We first give a sharper analysis of the cyclic instance and improve its lower bound to \Omega(\sqrtM/D) for DM^2 . We then construct, for every depth D\ge M^2 , an M -covered path of depth D with excess error \Omega(M^2/D) . The same instance gives the constant lower bound for all D M^2 . We also show that for any fixed distribution the excess error contracts geometrically along the path, ruling out any single instance that witnesses any polynomial lower bound at every depth. Finally, we prove the same optimal rate for logistic classification in the logit-passing model of Bateni et al., which considers the binary cross-entropy (BCE) loss. The same improved upper bound of O(M^2/D) holds, and we transfer all the regression lower bounds by showing that on those examples the logistic path follows the least-squares path up to rescaling. Comments: Initial version submitted to SODA 2027 on July 9, 2026 Subjects: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT); Theoretical Economics (econ.TH) Cite as: arXiv:2609.05318 [cs.LG] (or arXiv:2609.05318v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.05318 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-4] Learning from VAE Errors to support ECG-based Differential Diagnosis of Myocardial Scar KDD2026 ECML
链接: https://arxiv.org/abs/2609.05294
作者: Shayan Sharifi,Riccardo Treu,Ilaria Gandin,Federico Garoia,Marco Merlo,Giulia Cisotto
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Accepted at the PharML Workshop, ECML PKDD 2026
Abstract:Late Gadolinium Enhancement (LGE) on cardiac magnetic resonance is a key marker of myocardial scar, but its limited accessibility motivates routine ECG-based screening. We evaluated whether \beta -variational autoencoder (VAE)-derived ECG representations can discriminate LGE+ from LGE- cardiomyopathic patients in a local cohort of 300 subjects. We compared 32-dimensional features from the foundation this http URL model with those from a shallower \beta -VAE trained on normal PTB-XL ECGs, evaluating downstream classification and Dynamic Time Warping (DTW)-based reconstruction errors. this http URL reached an area under ROC of 0.686 with Random Forest, while the proposed \beta -VAE reached 0.577 with sensitivity of 0.775 with Gradient Boosting. Notably, DTW-reconstruction errors significantly differed between classes in 10 out of 12 leads according to Mann-Whitney U test and help in classification, leading to an area under ROC of 0.643 with Logistic Regression, supporting their potential as markers of scar-related ECG alterations.
[LG-5] How to Speculate about Uncertainty in Agent ic Coding? A Draft-Model Gate Method EMNLP2026
链接: https://arxiv.org/abs/2609.05274
作者: Konstantin Grotov,Valentin Malykh
类目: Machine Learning (cs.LG)
*备注: 11 pages, 2 figures, EMNLP 2026, Industry track
Abstract:LLM agents deployed for software engineering fail expensively: they act confidently wrong, and bad actions are recognized only after costly execution and retry. We present Speculative Uncertainty (SU), a method that recovers a predictive failure signal for a black-box agent from its output tokens alone, with no access to logits, weights, activations, or repeated sampling. Inverting speculative decoding, a small open-weight draft model scores the agent’s already-generated trajectory in a single forward pass. From these speculative cross-likelihoods we extract phase-aware features by separating the reasoning and action spans, and calibrate them against a verifiable objective. SU produces a failure-likelihood score that any downstream policy, such as routing, human intervention, or extra test-time compute, can consume directly. To show the signal is actionable, we instantiate one such policy, a pre-execution veto gate, on software engineering agents Qwen3-Coder-480B and closed-source Claude 3.5 Sonnet, cutting execution error rate by 6-8 percentage points and token cost by 14-19% in deployment, transferring to out-of-distribution benchmarks without retraining, and generalizing across agent models.
[LG-6] Shallow neural network approximation in mixed Sobolev spaces
链接: https://arxiv.org/abs/2609.05263
作者: Yuwen Li,Guozhi Zhang
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG)
*备注: 36 pages, 2 figures
Abstract:We investigate the best L_2 approximation of mixed Sobolev spaces by shallow neural networks with n neurons and general activation functions. We first establish an activation-independent Fourier-block principle: if an activation has univariate approximation order \rho in the sense of the Fourier-block property, then the global approximation rate has algebraic order \min\alpha,\rho\ for target functions of mixed smoothness \alpha , up to explicit logarithmic factors. To verify this property for concrete activations, we introduce a structured univariate approximation condition that implies the Fourier-block property with explicit parameters. For \mathrmReLU^k , a matching algebraic lower bound identifies \min\alpha,k+1\ as the optimal algebraic approximation exponent in any dimension, up to logarithmic factors in the upper bound. The framework also yields the exponent \min\alpha,k+1\ for cardinal B-splines and soft- \mathrmReLU^k , and the full mixed-smoothness exponent \alpha for ELU and cosine activations, again up to logarithmic~factors.
[LG-7] GLASS: Graph-Language Alignment with Spherical Scoring for Transferable Graph-Level Anomaly Detection
链接: https://arxiv.org/abs/2609.05253
作者: Xudong Wang,Chris Ding,Tongxin Li,Jicong Fan
类目: Machine Learning (cs.LG)
*备注: Preprint. This work and project were done in Apr. 2026. This work was included in Xudong Wang’s Ph.D. thesis (Defense Passed on 13 Apr. 2026), “Principled and Effective Graph Representation Learning with Application to Anomaly Detection,” deposited with The Chinese University of Hong Kong, Shenzhen Library
Abstract:We introduce GLASS, a framework for graph-level anomaly detection (GLAD) that achieves robust cross-domain transferability through graph-language alignment on the unit hypersphere. GLASS builds a unified representation space by aligning a structure-aware graph encoder with an instruction-aware text embedding via a multi-slice soft cosine objective. Our framework serializes local, global, and semantic graph properties into a compact Graph Descriptor Prompt (GraphDP), creating a text bridge that enables domain-agnostic anomaly scoring. By enforcing multi-scale consistency through Matryoshka representation slices, the model captures anomalous deviations at multiple levels of granularity. For scoring, we formulate anomaly detection as density estimation on the aligned hypersphere and introduce Spherical Multi-Modal Scoring (SMS), which instantiates von Mises-Fisher kernel density estimators in both graph and text embedding spaces. This probabilistic formulation recovers angular k-nearest-neighbor scoring as a high-concentration limiting case and provides a principled fusion of structural and semantic anomaly signals. The shared text embedding space further serves as a cross-domain bridge: by encoding a target domain’s GraphDP without target-domain training data, GLASS performs zero-shot anomaly detection, and with only a handful of normal examples, few-shot adaptation via reference-set calibration. Across twelve benchmarks and three meta-domains, GLASS obtains the best average AUROC and rank compared with recent advanced GLAD baselines and enables effective cross-domain transfer.
[LG-8] Proton Irradiation Characterization of an Open-Source ML Accelerator on a Zynq UltraScale MPSoC
链接: https://arxiv.org/abs/2609.05249
作者: Saad Memon,Rafal Graczyk,Jan Swakoń,Leszek Grzanka,Sebastian Kusyk,Mike Papadakis
类目: Hardware Architecture (cs.AR); Emerging Technologies (cs.ET); Machine Learning (cs.LG)
*备注:
Abstract:As spaceborne computing systems increasingly rely on neural network (NN) accelerators, the opacity of commercial, black-box architectures severely restricts the development of verifiable radiation mitigation strategies. Open-source, register-transfer level (RTL)-accessible accelerators resolve this limitation by enabling user-defined instrumentation, yet few have empirical radiation-response baselines. This work establishes a foundational system-level proton-irradiation baseline for an unmitigated open-source Tensil NN accelerator deployed on a Zynq UltraScale+ SoC executing ResNet-20 inference. Under 20 to 58 MeV proton irradiation, we delivered 4.29 \times 10^10 p/cm ^2 within monitored operational windows. Seven workload interruptions required two restarts of the notebook process, four reboots or board resets, and one power-cycle sequence. Two output-corruption events returned incorrect CIFAR-10 classes without loss of service. In the longer event, the accelerator returned a class absent from the ten-image CIFAR-10 pool for 39 consecutive inputs at normal cadence. The process remained alive, while the kernel log, limited memory test, and sampled power showed no anomaly. Observation of the stuck-class sequence ended with scheduled bitstream reconfiguration. All nine onsets occurred under the nominal 4 cm beam, which exposed the SoC, LPDDR4, and additional board circuitry; none occurred under the 2 cm SoC-centered field. This pattern shows a field association but does not establish LPDDR4 as the cause because field size was confounded with run order and dose. Linux-managed accelerators require end-to-end content checks and recovery that reaches the state in which corruption can persist. This baseline documents availability loss and silent output corruption, supporting future software hardening of COTS FPGA-SoCs for neural-network inference in space systems.
[LG-9] Hessian-based molecular conformation augmentation for a scalable and efficient strategy of machine learning interatomic potentials
链接: https://arxiv.org/abs/2609.05233
作者: Bumju Kwak,Jeonghee Jo
类目: Machine Learning (cs.LG); Chemical Physics (physics.chem-ph)
*备注: 45 pages including Supporting Information, with 6 figures and 9 tables in the main text
Abstract:While machine-learning interatomic potentials (MLIPs) have successfully learned potential energy surfaces (PES) and atomic forces, many practical applications, such as vibrational analysis and transition state search, rely heavily on the PES Hessian. Yet, standard MLIPs tend to be trained on energy and forces alone, leaving Hessian information largely unexploited. Meanwhile, existing methods that explicitly incorporate the Hessian into training objectives require architectural modifications and introduce significant computational and memory overheads due to higher-order backpropagation. To address these limitations, we propose two Hessian-derived data augmentation schemes: isotropic Gaussian displacement (\textbfUniAug) and normal mode-weighted displacement (\textbfModeAug). Both methods utilize simple Taylor expansions, achieving effective augmentation without altering training objectives or extending the autograd graph. This allows seamless, plug-and-play integration with existing architectures and training pipelines. Comprehensive evaluations across non-equilibrium and equilibrium datasets demonstrate that our approach enhances model accuracy while providing practical, task-specific guidelines.
[LG-10] FedDRAW: Federated Dual Reputation Annealing Weighting for Heterogeneous Multi-Institutional Chest Radiograph Classification
链接: https://arxiv.org/abs/2609.05223
作者: Maryam Moradpour,Anne-Christin Hauschild
类目: Machine Learning (cs.LG)
*备注:
Abstract:Artificial intelligence models are promising for medical diagnosis, but they require large numbers of unbiased data, which in medicine are distributed across hospitals and cannot be centralized to protect patient privacy. Federated Learning (FL) addresses this, since hospitals train one shared diagnostic model while patient data remain local. Training proceeds in communication rounds, in which each hospital trains the shared model locally and returns it to the server for merging by weighted average. This aggregation weight determines whose institutional knowledge shapes the result. Federated averaging (FedAvg) sets it in proportion to local sample count, so a small but informative hospital is permanently assigned a small influence, andl argest clients could dominate the global model even when they are less informative. We propose Federated Dual Reputation Annealing Weighting (FedDRAW), a server-side aggregation method that combines a data-size prior with the cosine similarity between client and global parameters under two coupled annealing schedules. An inner schedule shifts client reputation from the size prior towards similarity. An outer, deferred annealing schedule on the softmax inverse temperature keeps the weighting selective in the early and middle rounds and relaxes it to uniformity at convergence. We evaluate FedDRAW on 12 simulated client-partition scenarios of two chest radiograph datasets (CheXpert and ChestMNIST), against seven federated baselines under identical local training settings. FedDRAW achieved the highest average rank among all eight methods under both AUC and the geometric mean (GM) of sensitivity and specificity, which a Friedman test with Nemenyi post-hoc analysis confirmed to be a statistically significant difference between the methods. Scheduling two signals, rather than fixing the weights by sample count alone, could enable less biased diagnostic models.
[LG-11] Dimension-Adaptive Batched Lipschitz Narrowing Without Knowing the Zooming Dimension
链接: https://arxiv.org/abs/2609.05214
作者: Yasong Feng
类目: Machine Learning (cs.LG)
*备注:
Abstract:The Appropriately Combined Edge-length (ACE) sequence in A-BLiN depends on the zooming dimension d_z . This note removes that dependence. The next edge length is selected from the number of cubes that survive the preceding elimination. The resulting Count-Adaptive BLiN algorithm does not use d_z or the zooming constant C_z , yet it attains \widetilde\mathcal O_d(T^(d_z+1)/(d_z+2)) regret with \mathcal O_d(\log\log T) batches. Together with the adaptive-grid lower bound in Theorem 10 of the original paper, the optimal batch complexity remains \Theta_d(\log\log T) when d_z is unknown.
[LG-12] Conformal Prediction for Offensive Security
链接: https://arxiv.org/abs/2609.05165
作者: Giovanni Cherubin
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Despite its introduction more than a quarter century ago, Conformal Prediction (CP) has seen surprisingly few applications to the cyber security world thus far. In particular, we observe that, while CP has been employed as a defensive measure in many recent works, its use for carrying out attacks (i.e., for offensive security) is hard to trace in the literature. We explore this gap, by presenting initial findings in two key areas of offensive security: Privacy-Preserving Machine Learning, and network traffic analysis.
[LG-13] From 80x to 385x: A Best-Matching-Unit Search at the L2 Roof Measured Against a Symmetrically Tuned Baseline FAST
链接: https://arxiv.org/abs/2609.05138
作者: Andrew James Amos
类目: Machine Learning (cs.LG)
*备注: 15 pages, 8 tables, 3 figures. Companion to arXiv:2608.24067 . Both implementations were tuned symmetrically; cuSPARSE became 2-3x faster in the process. Code, data and frozen results: doi: https://doi.org/10.5281/zenodo.22245712 (tag v2.0)
Abstract:Comparisons between GPU implementations are usually asymmetric: one side is tuned by its author, the other is run as found. I report a programme that tuned both a novel SOM algorithm (SparseBin) and the baseline algorithm it was being compared to (cuSPARSE). The best-matching-unit search that dominates self-organizing map training was tuned through four levers - tile size, tile-membership clustering, neuron-axis chunking and vectorised loads - reaching 5.6-10.1x per epoch over the previously published configuration at map sizes from 32x32 to 512x512, and lifting the margin over the CUDA implementation behind our earlier MEDLINE atlases from ~80x to ~385x. cuSPARSE, the implementation SparseBin is compared against, received every lever with an analogue on its side, and became 2-3x faster in the process. The tuned kernel pressed the L2 bandwidth roof at 77% of peak with every other unit at 40-65%, bounding any further lever at ~1.3x - a terminal result rather than a waypoint, and every untested lever was either capped by that bound by construction or measured null.
[LG-14] MomentQuant: an even more minimalist interval method with linear time complexity for time series classification
链接: https://arxiv.org/abs/2609.05136
作者: Johann Faouzi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Time series data is very common in many real-world applications and in numerous domains, with increasing interest for automated information extraction using machine learning. One of these subfields is time series classification, which consists in assigning a label to each new, unseen time series. Many algorithms have been developed over the past decades, with the trade-off between predictive performance and computational cost being consistently discussed. Quant, an interval-based algorithm extracting quantiles from recursive, fixed, dyadic intervals, was shown to achieve high accuracy, while being very fast. We propose two changes to make this algorithm even faster. The first one is a better optimized implementation of the exact same algorithm. The second one is to derive approximate quantiles, using the Cornish-Fisher expansion, instead of exact quantiles. This change removes the necessity to sort the time series, leading to a smaller computational complexity. We call this novel algorithm MomentQuant. We provide evidence that our implementation of Quant is faster than the original one, and that MomentQuant is even faster than our implementation of Quant, at the cost of a tiny decrease in predictive performance. These improvements are especially relevant for real-life applications, where inference is performed much more often than training.
[LG-15] Coarse-Graining Hidden Representations: Unsupervised Neuron Selection via Mapping Entropy
链接: https://arxiv.org/abs/2609.05126
作者: Margherita Mele,Andrea Castagna,Roberto Menichetti,Raffaello Potestio,Alessandro Ingrosso
类目: Machine Learning (cs.LG); Disordered Systems and Neural Networks (cond-mat.dis-nn); Statistical Mechanics (cond-mat.stat-mech)
*备注:
Abstract:Overparameterized neural networks carry far more hidden units than a task nominally requires, raising the question of which neurons are essential and whether that distinction is legible in the representation itself, without labels or gradients. We cast neuron selection as the problem of coarse-graining the hidden layer by retaining a subset of its neurons, and score each putative selection by the mapping entropy (ME). This quantity measures the loss of discriminatory power inherent in discarding part of the network neurons, and the selection that minimises the ME is taken as particularly informative. This criterion is fully unsupervised, in that it depends only on hidden-activation statistics. In teacher-student networks, ME optimisation recovers the minimal teacher-consistent representation and retains extra units in proportion to the hidden layer’s residual variability; in a non-linear Gaussian process task, it selects coherent functional-class mappings whose preferred class shifts across training. On this task and on translation-augmented MNIST, ME-selected subnetworks outperform random subsets of equal size, most clearly under strong compression - linking configurational distinguishability to predictive performance.
[LG-16] Single-Query Black-Box Calibration Auditing via Logit Bias
链接: https://arxiv.org/abs/2609.05125
作者: Roman Plaud,Antoine Saillenfest,Matthieu Labeau,Thomas Bonald,Willem Waegeman
类目: Machine Learning (cs.LG)
*备注:
Abstract:Evaluating the calibration of Large Language Models (LLMs) is critical for their safe deployment as zero-shot classifiers. Yet, commercial API providers increasingly hide the continuous output probabilities required by standard calibration metrics. To bypass this opacity, we demonstrate that any LLM API exposing a logit_bias parameter can be mathematically manipulated to evaluate exact probability thresholds using strictly one query per sample. Leveraging this mechanism, we introduce a novel and provably consistent estimator of the True Calibration Error for binary tasks. Our approach therefore provides an efficient framework for auditing black-box foundation models.
[LG-17] A Comparative Study of Counterfactual Explainers for Graph Neural Networks Enabling Multiple Types of Graph Edit
链接: https://arxiv.org/abs/2609.05113
作者: Maria Myrto Villia,Filippos Gouidis,Theodore Patkos,Panos Trahanias
类目: Machine Learning (cs.LG)
*备注:
Abstract:Counterfactual explanations for graph-structured data seek to determine minimal and realistic modifications required in an input graph to alter a model’s prediction to a predefined output. Although counterfactual explainers that support modifying the graph by both adding and removing edges have recently emerged, there is still a lack of general and efficient methods, especially when considering the quality of the generated explanations. Moreover, the problem remains far from solved, as existing methods exhibit different strengths and weaknesses, often trading off between explanation size, coverage and quality. For this reason, it is important to identify where each method performs well and where it falls short, so as to guide future research in the field. Thus, our study compares six state-of-the-art (SOTA) models on a diverse set of real-world and synthetic datasets, covering both binary and multi-class graph and node classification tasks, and evaluates their performance using diverse quantitative and qualitative metrics.
[LG-18] Deep Microcompression: Structured Pruning and Bit-packed Quantization for Microcontrollers ICML2026
链接: https://arxiv.org/abs/2609.05081
作者: Opegbemi Matthias Busoye,Tolulope Matthew Busoye,Eghonghon-aye Eigbe
类目: Machine Learning (cs.LG)
*备注: Presented at the Global South ML Workshop at the International Conference on Machine Learning (ICML 2026), Seoul, South Korea
Abstract:This paper introduces Deep Microcompression (DMC), a hardware-aware pipeline for deep learning inference on bare-metal microcontrollers. DMC integrates structured pruning, quantization-aware training, and fixed-length bit-packing to achieve a 55.8 \times weight compression ratio on LeNet-5 (98.77% accuracy), generating a dependency-free C library with deterministic latency. On the RP2040 (Cortex-M0+), DMC reduces binary size by 3 \times versus TensorFlow Lite while matching its accuracy. Critically, DMC enables the first documented deployment of a standard CNN on the ATmega328P, a device constrained to 2KB SRAM, previously considered infeasible for CNN inference.
[LG-19] Confounding-Valid Conformal Inference for Counterfactual KPIs in Wireless Networks
链接: https://arxiv.org/abs/2609.05073
作者: Abdessamed Qchohi,Jessica Moysen Cortes,Matteo Zecchin
类目: Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI); Signal Processing (eess.SP); Machine Learning (stat.ML)
*备注:
Abstract:Conformal counterfactual inference enables network operators to use logged telemetry to reliably answer ‘what-if’ questions about network operation. These answers typically take the form of prediction sets that contain, with a user-defined probability, the key performance indicators (KPIs) that would have been observed under alternative control actions. A key challenge is that logged telemetry may omit variables used by the controller, resulting in hidden confounding and invalidating the statistical guarantees of counterfactual analysis. In principle, this issue can be addressed using randomized telemetry, collected by assigning control actions independently of the network state. However, because such randomization may disrupt normal operation, randomized telemetry is typically scarce, causing counterfactual analysis based solely on it to produce uninformative prediction sets. To address these challenges, we propose Confounding-Valid Counterfactual Conformal Inference (CV-CCI), which combines abundant, potentially confounded observational telemetry with limited randomized data through the General Synthetic-Powered Inference (GESPI) principle. CV-CCI leverages observational data to improve efficiency while using randomized data to retain finite-sample coverage guarantees under arbitrary hidden confounding. Experiments on two representative radio access network (RAN) control tasks show that CV-CCI remains valid under hidden confounding while producing more efficient prediction sets than state-of-the-art confounding-valid baselines.
[LG-20] Solution-space heterogeneity shapes federated learning dynamics across partial differential equations
链接: https://arxiv.org/abs/2609.05012
作者: Ping Luo,Jiahuan Wang,Ziqing Wen,Tao Sun,Dongsheng Li
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注:
Abstract:Federated scientific machine learning enables institutions to train neural surrogates without centralizing local physical data, yet studies of partial differential equations (PDEs) lack a transferable definition of non-independent and identically distributed data. Existing protocols partition coordinates, coefficients, boundary conditions, or geometries according to equation-specific rules. Here, we introduce solution-space PDE-Dirichlet, a protocol that converts continuous supervised responses into reusable solution bins and quantifies the realized separation between clients through optimal transport over the geometry of these bins. We derive an exact inverse relation between population allocation heterogeneity and the Dirichlet concentration, and we establish conditions under which response heterogeneity induces gradient disagreement, local-update dispersion, and parameter divergence. Across seven controlled and public PDE tasks, three neural-operator families, and five random seeds, a lower concentration consistently increases the realized solution distance and optimization heterogeneity. The degradation in final error is task dependent: the largest effect occurs for low-viscosity Burgers, reaching 4.157 percentage points under the most heterogeneous setting, whereas additional communication or smoother dynamics can reduce the final gap despite persistent parameter separation. These results distinguish a reproducible geometric mechanism from task-dependent generalization outcomes and provide a common basis for evaluating non-IID federated PDE learning.
[LG-21] Beyond Homoscedasticity: Decoupled Uncertainty Optimization for Deep Imbalanced Regression ACM-MM2026
链接: https://arxiv.org/abs/2609.04995
作者: Juncheng Zhou,Jiaxi Lu,Weijing Zeng,Zhong Li,Hao Qi,Jingsong Cui
类目: Machine Learning (cs.LG)
*备注: Accepted at the 34th ACM International Conference on Multimedia (ACM MM 2026). Main paper with supplementary material
Abstract:Deep Imbalanced Regression (DIR) is pervasive in continuous prediction tasks across diverse modalities, such as age estimation, depth prediction, and protein mutation activity prediction, where label-scarce tail samples often carry higher practical value. However, most existing methods still learn deterministic point mappings under mean squared error or its simple variants, implicitly assuming a uniform uncertainty level across all samples and thereby overlooking the instance-wise heteroscedasticity that is widespread in long-tailed data. We further point out that even heteroscedastic negative log-likelihood suffers from a gradient coupling issue, which, under DIR scenarios, weakens the learning signal of hard tail samples and leads to optimization inertia as well as tail underfitting. To address this, we propose DUO, an uncertainty-aware long-tailed regression framework. Specifically, the proposed method models the regression target as a conditional Gaussian distribution to explicitly characterize instance-level predictive uncertainty, and transforms uncertainty into a dynamic enhancement signal for tail samples through decoupled mean-variance optimization. Furthermore, we design a distribution-guided contrastive learning mechanism that adaptively constructs positive and negative pairs based on the overlap between sample distributions, thereby alleviating feature looseness and cross-label semantic entanglement. Across visual and biological DIR benchmarks, DUO achieves the best few-shot bMAE and GM on IMDB-WIKI-DIR, AgeDB-DIR, and AAV2-DIR while remaining competitive on few-shot MAE.
[LG-22] Fractal basins trap latent reasoning
链接: https://arxiv.org/abs/2609.04963
作者: Jeffrey Lai,Anthony Bao,John Quinn,William Gilpin
类目: Machine Learning (cs.LG)
*备注: 6 pages, 5 figures
Abstract:Reasoning allows artificial intelligence models to revisit and correct their mistakes, enabling recent frontier advances in mathematical theorem solving, software engineering, and autonomous task planning. Reasoning models are widely observed to reason for longer on harder tasks, but the general mechanism responsible for these slowdowns is unknown. Here, we show that reasoning models exhibit transient chaos, a physical consequence of the computational complexity of difficult tasks. As a consequence, we show that diverse leading reasoning models are dynamical systems with fractal basins, with fractality increasing with task difficulty across diverse tasks like Sudoku and maze solving, visual puzzles, and mathematical logic. We show that transient chaos emerges due to reasoning becoming trapped for extended durations near saddle points, which we show correspond to nearly-correct attempted solutions of the underlying problem. Our results show that reasoning slowdowns are an inevitable consequence of problem hardness in modern artificial intelligence models, and establish reasoning traces as a rich new class of dynamical system.
[LG-23] Physics-Aware Random Walk Fingerprints for Scalable Power Grid Graph Classification
链接: https://arxiv.org/abs/2609.04943
作者: Adnan Anwar
类目: Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: power system, power system security, cascading failures, graph classification, random walk fingerprints, physics-aware features
Abstract:Recent benchmarks such as PowerGraph provide large collections of power-grid graphs for cascading-failure classification. Graph neural networks (GNNs) achieve strong predictive performance on this task, but typically require end-to-end training and model-specific tuning, while their latent representations can be difficult to relate to physically meaningful propagation patterns. Random Walk Fingerprints (RWF) offer a scalable and interpretable alternative, but existing variants primarily emphasise topology and node-level information, leaving grid-relevant operational edge states in the walk dynamics. We propose Multi-Channel Physics-Aware Random Walk Fingerprints (MC-PA-RWF) for power systems, a lightweight graph-level representation framework that introduces physical edge states into random-walk propagation. The method constructs multiple edge-weighted channels from domain-relevant attributes, extracts a channel-specific fingerprint from each weighted graph, and concatenates the resulting vectors into a compact representation. Experiments on three \textitPowerGraph benchmark systems show substantial improvements over topology-only RWF and competitive balanced accuracy against strong GNN baselines, including Graph Convolutional Networks (GCN), Graph Attention Networks (GAT), Graph Isomorphism Networks with edge features (GINE), and Transformer-based Graph Convolutional Networks (TransformerConv). At the largest evaluated settings, the node-edge extension MC-PA-RWF+ achieves around 98.04% - 99.32% balanced accuracy and improves failure-class F1 over the strongest GNN baseline by 1.60 – 5.84 percentage points, with statistically significant gains across all three systems.
[LG-24] Fast Gauss Sums via Flash Attention
链接: https://arxiv.org/abs/2609.04910
作者: Nicolaj Rux,Sebastian Neumayer
类目: Machine Learning (cs.LG); Numerical Analysis (math.NA)
*备注: 6 pages 5 figures
Abstract:Gaussian kernel sums are the computational core of maximum mean discrepancies (MMDs), kernel gradient flows, Stein variational gradient descent (SVGD), and many other kernel methods. At the same time, softmax attention has received an extraordinary amount of hardware-aware code engineering, culminating in flash attention. We show that Gauss kernel sums with arbitrary, signed weights can be evaluated via flash attention: two small input augmentations turn the normalized softmax reduction into the unnormalized Gauss sum, without writing a single line of custom GPU code. For feature dimension D8 in fp16, this approach beats compiled PyTorch code as well as PyKeOps kernels (often significantly) in speed, memory-overhead and accuracy. Indeed, its memory scaling remains linear.
[LG-25] From Deep to Shallow: Unconstrained and Efficient Layer Merging Strategy KDD2026 ECML
链接: https://arxiv.org/abs/2609.04881
作者: Petro Shulzhenko,Gabriele Spadaro,Enzo Tartaglione
类目: Machine Learning (cs.LG)
*备注: 13 pages, 3 figures, accepted at the ITEM Workshop at ECML PKDD 2026
Abstract:Although Deep Neural Networks have become foundational in many areas of Machine Learning, high computational demands limit their application in resource-constrained environments. To address this issue, depth compression methods have been proposed to identify and linearize redundant activation functions, thereby allowing for the merging of layers without intermediate non-linearities. However, these methods face two key challenges: they cannot be directly applied to convolutions with padding due to the absence of an analytical solution for merging these layers, and they typically increase the kernel size of merged layers, thus limiting speed-up gains. To overcome these limitations, we propose an efficient strategy that enables merging of layers without an existing analytical solution, and also without increasing kernel size. We validate our approach across multiple architectures and datasets, and measure inference speed-up gains on real embedded platforms. We publicly released the code at this https URL.
[LG-26] When Genomic Masking Priors Fail to Transfer: Strong Variant Prediction Weak Functional Generation
链接: https://arxiv.org/abs/2609.04861
作者: Susu Hu,Preetam Gattogi,Jens Lehmann,Sahar Vahdati,Stefanie Speidel,Julien Vibert
类目: Machine Learning (cs.LG)
*备注:
Abstract:Bidirectional discrete diffusion model appears naturally suited to genomic modeling because it can reconstruct missing sequence from both flanks. We developed GenDA (Genomic Density-optimized Absorbing Diffusion) under the additional hypothesis that entropy-guided span placement would concentrate reconstruction pressure on compositionally complex regions, improving both downstream variant-effect prediction and functional sequence generation. Our results only partially support this premise. After supervised fine-tuning, the 202M-parameter GenDA model reaches a pooled ClinVar SNV AUROC of 0.774, exceeding a similarly scaled autoregressive model by 0.103. However, a matched random-span variant reaches 0.777, providing no evidence that entropy guidance causes the ClinVar improvement. More unexpectedly, GenDA fails a zero-shot functional inpainting stress test: across promoters, enhancers, exon boundaries, and intron boundaries, it does not consistently outperform a control that shuffles the native gap while exactly preserving 3-mer composition. Failure is already present for 50–500-bp gaps, although enhancer degradation worsens at longer gaps. Diagnostics identify several boundary conditions: entropy measures local sequence complexity rather than functional importance; 1-mer tokenization limits physical context; training spans are capped at 300 bp; and high absolute AlphaGenome fidelity can coexist with negative control-normalized restoration. These results show that strong fine-tuned variant prediction, a plausible corruption prior, and functional generation are distinct claims that require separate validation.
[LG-27] KVMem: Virtualizing Million-Token Agent Workspaces on a Consumer GPU
链接: https://arxiv.org/abs/2609.04852
作者: Di Chai,Leye Wang,Zeshen Su,Zhiguo Xia,Zhihang Yu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Modern LLM agents operate in persistent workspaces whose accumulated history can exceed both GPU KV capacity and the model’s native context window. Existing systems typically compact older context into summaries or retrieve it later as text, either losing fine-grained execution evidence or repeatedly prefilling content that the model has already processed. We present KVMem, a KV-context virtualization system that preserves overflowed workspace history as paged KV state across GPU memory, host memory, and NVMe. KVMem uses lightweight, model-native attention-space indexes to select relevant historical blocks and materializes a query-dependent execution view bounded by the model’s native context window. Extensive evaluations on long-context agent benchmarks spanning histories up to one million tokens, including LongMemEval, MemoryAgentBench, and AgentLongBench, show that KVMem generally achieves higher task utility and greater inference efficiency than compaction-based approaches, the de facto standard for handling context overflow. In the DeepSWE long-context test with Qwen3.8-27B, KVMem improves task success from 43.8% with compaction-only context management to 48.4%. In our local-deployment evaluation, KVMem runs Qwen3.6/3.8-27B NVFP4 with MTP on an off-the-shelf laptop equipped with a 24,GB RTX 5090 Laptop GPU, virtualizing agent workspaces of up to 1M tokens-four times the model’s native 256K-token context window. In a single-session setting, KVMem generates \sim 50 tokens/s, providing interactive responsiveness for local agent execution. More broadly, by decoupling addressable workspace size from the LLM’s native context window, KVMem provides a practical path toward long-running agents whose workspaces can grow beyond that window. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2609.04852 [cs.LG] (or arXiv:2609.04852v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.04852 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-28] Coupled Control and Wireless World Models for Resilient Remote Robotic Control
链接: https://arxiv.org/abs/2609.04851
作者: H.P. Madushanka,Sumudu Samarakoon,Mehdi Bennis
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: 13 pages, 13 figures. Submitted to IEEE Internet of Things Journal
Abstract:Remote robotic systems operating over wireless networks must maintain reliable control despite limited communication resources, changing channel conditions, and environmental this http URL, continuously transmitting high-dimensional sensory observations, such as camera images, increases communication overhead and energy consumption while reducing robustness under unreliable this http URL address these challenges, this paper proposes a resilient communication-aware remote robotic control framework based on coupled control and wireless Joint Embedding Predictive Architecture (JEPA) world models that jointly capture robot dynamics and wireless channel evolution from visual observations and a combination of raw and structured radio frequency (RF) representations based on spectrograms and Persistence Images(PIs).The learned latent representations enable predictive communication scheduling by jointly forecasting future robot states and wireless conditions, thereby reducing unnecessary uplink transmissions while maintaining reliable control this http URL, an adaptive resilience mechanism detects latent prediction discrepancies and efficiently adapts perception embeddings to accommodate wireless and visual environmental changes without retraining the complete control this http URL proposed framework is evaluated in a synchronized Gazebo-Robot Operating System (ROS)-Sionna robot-wireless simulation environment under diverse wireless propagation and perception this http URL results demonstrate significant improvements in communication efficiency, robustness, and resilience while maintaining navigation performance compared with conventional Proportional Integral Derivative (PID), model-free Deep Q-Network (DQN), and predictive approaches based on Vision Transformers(ViTs).
[LG-29] PACE: Propagation-Aware Collaborative Correction for One-Shot Personalized Federated Graph Learning
链接: https://arxiv.org/abs/2609.04832
作者: Ruizhe Huang,Chengran Li,Xiaochuan Shi
类目: Machine Learning (cs.LG)
*备注: 15 pages, 4 figures, 12 tables
Abstract:Client heterogeneity creates both an opportunity and a risk in personalized federated graph learning. Knowledge held by other subgraphs may complement a receiver’s Local model, but an incompatible transfer can override reliable predictions. One-shot communication sharpens this tension because an unsuitable server return cannot be corrected later. We introduce PACE, which treats collaborative knowledge as a compact correction to a complete Local predictor rather than as its replacement. Each client uploads a rank-r update carrier and a diagonal sketch of propagated message moments. The server uses them to construct a propagation-aware, receiver-anchored correction, while the receiver retains its full Local model. Convex negative-log-likelihood calibration (CNLL) then selects one coefficient between Local and External logits using validation nodes; model parameters remain fixed and no feedback is sent. At Rank-6, personalized returns occupy 9.6-17.6% of dense tensor bytes across the six evaluated datasets. The correction receives nonzero weight and improves both Accuracy and weighted-F1 over Local on five datasets; on ogbn-arxiv, CNLL assigns zero predictive weight to the correction and preserves Local predictions exactly. Applying the same CNLL rule to matched baselines on three citation datasets does not account for these gains. The central result is therefore that a small transported correction can augment a complete Local model when receiver evidence supports it while leaving the Local prediction unchanged otherwise.
[LG-30] Communication-Efficient Personalized Federated Learning via Layer-Wise Multi-Threshold Random Sketching
链接: https://arxiv.org/abs/2609.04830
作者: Xu Zhang,Xingyu Hou,Jiacheng Cheng,Kaiyuan Feng,Maoguo Gong
类目: Machine Learning (cs.LG)
*备注:
Abstract:Personalized federated learning (PFL) is a promising paradigm for collaborative learning over distributed devices, where edge nodes collaboratively train personalized models without sharing raw data. Although PFL addresses data heterogeneity by learning client-specific models, it still suffers from substantial uplink and downlink communication costs when exchanging high-dimensional parameters in bandwidth-constrained systems. Recent one-bit methods achieve extreme compression, but they usually rely on a single thresholding rule applied to the whole model. This design has two limitations. First, it overlooks layer-wise differences in parameter distributions and quantization sensitivities. Second, a single threshold provides only coarse binary information and cannot capture fine-grained variations in parameter distributions. To address these issues, we propose a communication-efficient PFL framework via layer-wise multi-threshold random sketching. In the proposed method, each layer is assigned its own set of quantization thresholds, so that the compressed representation can adapt to layer-specific statistics while using multiple intervals to provide a finer low-bit description of sketched parameters. The proposed method supports bidirectional communication using compact low-bit sketches and improves the communication-accuracy tradeoff compared with existing one-bit compression approaches.
[LG-31] Federated Attack Campaign Detection via Contrastive Encoding of Threat Indicators in Gradient Updates ECML-PKDD2026
链接: https://arxiv.org/abs/2609.04815
作者: Manuel Röder,Bibin Babu,Frank-Michael Schleif
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
*备注: Accepted to ECML-PKDD 2026, 4th Workshop on Advancements in Federated Learning - Towards Trustworthy Federated Learning
Abstract:Detecting orchestrated cyberattack campaigns that span multiple organizations traditionally requires sharing sensitive telemetry and threat intelligence across institutional boundaries and country borders, a barrier that Federated Learning removes by training shared threat detectors directly on local data. We propose FedIoC, a modular framework in which clients fold locally available structured threat indicators into their gradient updates; we instantiate the client-side encoder with a supervised contrastive loss over IoC-matched flows. Within each training batch, flows that match any known indicator pattern form the positive set; the contrastive objective pulls their learned embeddings together and pushes non-IoC embeddings away, so that campaign-relevant structure is, by design, expressed in the gradient direction. Clients sharing indicators for the same attack campaign then produce aligned gradient components, which the server clusters by the cosine similarity of their updates to recover global campaign patterns without any direct IoC transmission. We evaluate FedIoC on two public threat-detection benchmarks distributed across FL clients that each observe only a fragment of every active campaign and hold disjoint indicator sets derived from their local telemetry. In this regime the FL server recovers cross-organizational campaign cohorts directly from gradient geometry. We contribute FedIoC as a modular framework for this setting, and use it to pinpoint the non-IID gradient structure as the main driver of recovery and to define the open problem of designing encoders that improve on it.
[LG-32] How Faithful Is Attribution for Sales Forecasting? A Counterfactual Study
链接: https://arxiv.org/abs/2609.04797
作者: Glib Kechyn
类目: Machine Learning (cs.LG)
*备注: Code: this https URL
Abstract:Deep models for sales forecasting, such as WaveNet-style dilated convolutional networks, are accurate but opaque: when a single model predicts sales for one of many series, it offers no account of why. We add a post-hoc, architecture-agnostic counterfactual interpretability layer to a multi-series WaveNet forecaster trained on the full Corporacion Favorita grocery dataset (174,685 series over 1,688 days). The method decomposes each forecast into contributions that sum exactly to the predicted value, avoiding the allocation artifacts we observed with additive SHAP-style attribution. We evaluate faithfulness with a deletion/insertion protocol and find a statistically significant effect on both tests (deletion gap 0.22, p0.001; insertion gap 0.27, p0.01; robust across five background-sampling seeds), establishing that the attributions reflect genuine model behavior rather than plausible-looking artifacts. We then characterize, honestly, where attribution is and is not informative: reliance on the promotion signal is heterogeneous across series (median ratio approximately 1.0, with roughly 20% of series showing a strong effect), and the model captures the shape of the weekly sales cycle (day-of-week r=0.78) while systematically under-predicting its amplitude. Our contribution is not improved accuracy but an interpretability layer with a rigorous faithfulness evaluation and a candid account of its limits.
[LG-33] Learning-Augmented Algorithms: Guarantees Construction Mechanisms and System-Level Implications
链接: https://arxiv.org/abs/2609.04787
作者: Hailiang Zhao,Peng Chen,Xueyan Tang,Jianwei Yin,Shuiguang Deng
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS)
*备注:
Abstract:Learning-augmented algorithms use fallible predictions while retaining formal performance guarantees. This survey synthesizes prediction interfaces, error measures, consistency–robustness trade-offs, and five representative construction mechanisms across online optimization, caching, learned data structures, graph problems, and mechanism design. An orthogonal theorem-level axis distinguishes achieved upper bounds from matched asymptotic dependence. Formal guarantees are separated from empirical systems evidence, with explicit treatment of prediction cost, feedback, and composition. The resulting synthesis states sufficient conditions for limited end-to-end reasoning and delineates open problems in cost-aware prediction, endogenous error, semantic predictors, and benchmarking.
[LG-34] A Robust Watermark-based Fingerprint Framework for GNNs Ownership Verification
链接: https://arxiv.org/abs/2609.04772
作者: Han Zhang,Yan Wang,Guanfeng Liu,Pengfei Ding,Huaxiong Wang,Kwok-Yan Lam
类目: Machine Learning (cs.LG)
*备注: Accepted by IEEE DASC 2026
Abstract:The high training cost of Graph Neural Networks (GNNs) has raised growing concerns regarding model ownership infringement, such as model stealing and unauthorized misuse. To verify model ownership and prevent significant economic losses, two groups of GNN Ownership Verification (OV) methods have been proposed: watermark-based methods and fingerprint-based methods. However, these methods typically face three limitations: (1) the performance degradation of protected models caused by out-of-distribution (OOD) watermark graphs with respect to the training set; (2) the unrealistic assumption that surrogate models have been trained on a watermark-containing training set; and (3) over-reliance on specific output levels for fingerprint extraction. In this paper, we propose a Robust watErMArk-based fingeRprint frameworK for GNNs, named REMARK. REMARK first generates carefully crafted in-distribution watermark graphs that maximize output differences between GNN models, thus mitigating OOD-induced performance degradation. REMARK then extracts robust fingerprints from these output differences to verify GNN ownership, thereby removing the assumptions that surrogate models must be trained on a watermark-containing dataset or expose specific output levels. Extensive experiments across widely used real-world datasets and GNN architectures demonstrate that REMARK achieves state-of-the-art OV accuracy and robustness while preserving the utility of protected models.
[LG-35] Resilience Beyond Stationary Client Unavailability: Unlocking Efficient and Unbiased Federated Learning
链接: https://arxiv.org/abs/2609.04763
作者: Ming Xiang,Stratis Ioannidis,Edmund Yeh,Carlee Joe-Wong,Lili Su
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Optimization and Control (math.OC)
*备注: Journal of Machine Learning Research
Abstract:Due to resource constraints or external and internal uncertainties, clients in real-world federated learning systems are often intermittently available edge devices. In highly dynamic environments, the parameter server lacks prior real-time knowledge of clients’ availability, making it challenging to adapt traditional federated learning algorithms to be resilient to uncertainties in client availability. If not carefully addressed, complex client availability can introduce significant bias, potentially harming the performance of the trained model. Most prior work either fails to account for non-stationary client availability dynamics or demands significant memory and computational overhead. This paper aims to develop efficient federated learning algorithms that are provably resilient to heterogeneous and non-stationary stochastic client availability. We propose FedSWE, which admits novel algorithmic structures to (i) compensate for missed computations, (ii) stabilize and diffuse the global updates over rounds, and (iii) evenly mix the local updates through implicit gossiping, despite being agnostic to non-stationary dynamics. Compared with the standard FedAvg, FedSWE introduces light additional memory and computation overhead. We show that FedSWE converges to a stationary point of non-convex objectives while achieving the desired linear speedup property in certain special cases. We corroborate our analysis with numerical experiments over diversified client unavailability dynamics on real-world data sets.
[LG-36] A Fairness Audit of the Duckworth-Lewis-Stern Method: Format-Specific and Gender-Differential Bias with an Interpretable Calibration Layer for Cricket Target Revision
链接: https://arxiv.org/abs/2609.04754
作者: Soumyadeep Roy
类目: Machine Learning (cs.LG)
*备注: Accepted for publication in the Journal of Quantitative Analysis in Sports (JQAS); forthcoming
Abstract:The Duckworth-Lewis-Stern (DLS) method has been the international standard for revising target scores in rain-interrupted limited-overs cricket since 1999. Despite over two decades of operational use, no large-scale empirical audit of its prediction bias has been published. We conduct such an audit on 8,150 international matches (3,095 ODIs, 5,055 T20Is) from Cricsheet, generating 233,550 synthetic interruption scenarios with temporal splits. We document two structured biases. First, DLS prediction error spans a 137-run range across (overs-remaining, wickets-lost) match-state buckets. Second, DLS exhibits a gender-differential bias on ODIs that has not previously been quantified: on the training split, mean over-prediction is +1.51 runs for men but +7.63 runs for women, a gap of +6.13 runs (F = 195.16, p 10^-43). We benchmark DLS against five modern alternatives: Bi-LSTM, XGBoost, an enriched XGBoost variant, a deep context-aware model, and a stacking ensemble, and propose DLS-Cal, a lightweight interpretable calibration layer (27K parameters) outputting a state-conditioned correction added to DLS. DLS-Cal reduces absolute bias by 31% on ODI and 19% on T20I, and a gender-aware variant reduces women’s ODI residual bias from +6.19 to +0.65 runs while leaving men’s calibration unchanged. We release code, models, and data.
[LG-37] Same Request Different Answer: Quantization Amplifies Cache-Induced Divergence in LLM Serving
链接: https://arxiv.org/abs/2609.04748
作者: Aditi Patodiya
类目: oftware Engineering (cs.SE); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: Submitted to IEEE Access
Abstract:Prefix caching, in which a serving engine reuses the key and value tensors of a shared prompt prefix across requests, is enabled by default in the major open-source stacks and treated as a transparent optimization. We measure what it costs in reproducibility, and find that the cost rises sharply with weight quantization. Holding the model, decoding parameters, seed, and request order fixed, and issuing every request serially at batch size one, we ran an eighty-episode multi-turn agentic tool-use workload with caching enabled and disabled across two engines and four weight formats. Enabling the cache changed the agent’s trajectory on 36.2 percent of episodes at 16-bit precision and on 75.0 percent at four-bit, a gradient that survives re-measurement under a controlled cache configuration. With caching disabled, repeated execution was bit-identical in every configuration, 0 of 800 episodes, which bounds other sources of nondeterminism at 0.5 percent. Repeated cache-enabled runs did diverge, and three experiments locate the cause: a single server-level prompt-cache setting moves run-to-run divergence by 37.5 percentage points, execution order acts only while that setting is active, and restoring cache state makes the cached and recompute paths each reproduce on 40 of 40 items while still differing from each other on 14. Cached serving is deterministic given cache state, and irreproducible in practice because that state is absent from the request and never reset by default. A single-turn bridge shows the divergence reaching task outcomes without shifting aggregate accuracy. We release the harness, logs, and analysis pipeline.
[LG-38] raining Large Language Models for Small-Molecule Design with Synthetic Task Scaling
链接: https://arxiv.org/abs/2609.04735
作者: Frank Hu,Shriram Chennakesavalu,Zichen Wang,Patricia Suriana,Bodhi Vani,Kirill Shmilovich,Kangway Chuang,Colin Grambow
类目: Machine Learning (cs.LG)
*备注:
Abstract:Designing viable drug candidates requires searching a combinatorially large and rugged chemical space for molecules that satisfy multiple, often competing, objectives. Large language models (LLMs) provide a useful generative prior for this problem because of their representational capacity, reasoning ability, and flexibility when incorporating information from the external environment. While reinforcement learning from verifiable rewards (RLVR) can be used to improve the capabilities of LLMs, many chemically relevant scoring functions require hours or even days per evaluation, making them prohibitively expensive to use directly during online training. Here, we investigate whether LLMs can learn molecular design strategies from cheaper synthetic tasks that generalize to expensive molecular lead optimization settings. We find that curriculum-based training recipes that gradually incorporate more challenging synthetic design tasks enable strong performance that surpasses that of much larger frontier models on structure-based lead optimization. Our results suggest that scaling post-training using synthetic tasks is an effective strategy for adapting LLMs to high-cost experimental scenarios that are too expensive to directly train on.
[LG-39] Locating and Steering Refusal Beyond Attention
链接: https://arxiv.org/abs/2609.04721
作者: Preethi Carmel Bosco,Gopalakrishnan Srinivasan
类目: Machine Learning (cs.LG)
*备注: 33 pages, 6 figures, 26 tables; 10-page main text plus technical appendix
Abstract:Where inside a language model does refusal live, and does that place change when the architecture does? In a transformer, refusal is governed by a single direction in the residual stream, a finding that safety and interpretability tooling now depend on. State-space models (SSMs) route information through a recurrent update instead of attention, sharing no token-mixing mechanism with a transformer. Does the same safety representation survive this shift, or must it be rediscovered per architecture? It survives. A single rigid rotation, which can only reorient a space and not reshape it, aligns one model’s representation space with another’s, so the two genuinely share the representation. A harm probe trained on a transformer then flags an SSM’s harmful inputs, and removing the aligned direction makes a model answer attacks it would otherwise refuse, while a random direction of the same size does far less. What is architecture-specific is not where the direction is steered but where it must be read. Each layer computes a fresh output that is then added into the residual stream, and harm is cleanly readable at this output, the write site, before the addition. A control that holds the intervention’s strength fixed shows that what matters is where the direction is estimated, not where it is applied. Applied through a detector-triggered gate, this direction lowers jailbreak success in all four architecture families we test (SSM, transformer, recurrent, hybrid), and on the SSM it holds against an attacker that tunes its prompt against the defense. The gate only matches a trivial rule that returns a fixed refusal whenever the same detector fires, so what transfers across architectures is the direction itself, not defense strength. Safety tooling built on refusal therefore ports to a new architecture by re-estimating the direction at that architecture’s write site, not by rebuilding it.
[LG-40] WEECFP-SuRGE: Wide Embedded Extended Connectivity Fingerprint with Substructure Rotary Graph-distance Encoding
链接: https://arxiv.org/abs/2609.04672
作者: Robert Epps
类目: Machine Learning (cs.LG)
*备注:
Abstract:We introduce WEECFP, a parameter-free 1024-dimensional continuous molecular fingerprint that scatters each Morgan substructure across roughly thirty-two signed positions of a single vector, and WEECFP-SuRGE, a transformer architecture whose self-attention applies SuRGE (Substructure Rotary Graph-distance Encoding) – a RoPE-like rotation parameterized by molecular shortest-path graph distance – to WEECFP substructure tokens. A 7-model blend of this architecture (the WEECFP-SuRGE Blend) achieves the lowest average regression rank on the TDC ADMET leaderboard; is #2 overall on the TDC ADMET leaderboard (behind only pretrained MapLight+GNN), and is #1 overall among methods that use no external pretraining; takes leaderboard #1 finishes on Pgp, Lipophilicity, CYP2D6 Substrate, Clearance Microsome, and LD50 (with the WEECFP-NoSuRGE Blend separately reaching #1 on HIA) across the full 22-benchmark suite – without any external pretraining. On MoleculeNet, WEECFP-SuRGE beats every classical-fingerprint baseline on 3 of 4 regression tasks (ESOL, Lipophilicity, QM9). We further show that WEECFP tokenization is near-lossless: a greedy overlap reconstruction recovers the exact canonical SMILES of 99.9% of in-distribution molecules across 9 MoleculeNet datasets and 98.93% of molecules in a cross-dataset holdout (HIV-Lipophilicity), and that a three-reference farthest-first encoding of graph distance correlates at Pearson r = 0.901 with the true pairwise distance, enabling O(S) positional memory at matching accuracy.
[LG-41] Interpretability for Turing Machines
链接: https://arxiv.org/abs/2609.04661
作者: Billy Snikkers,Rumi Salazar,Daniel Murfet,Will Troiani
类目: Machine Learning (cs.LG); Formal Languages and Automata Theory (cs.FL); Machine Learning (stat.ML)
*备注: 75 pages, 31 figures, 3 tables. Interactive companion: this https URL . Code and data: doi: https://doi.org/10.5281/zenodo.22205895
Abstract:We show that susceptibilities, an interpretability technique developed for neural networks, can identify the presence of algorithmic structure in Turing machines by probing the local loss landscape of a learning problem for noisy Turing machines introduced by Murfet and Troiani (arXiv:2504.08075). We prove that symmetries and path separation in the algorithm implemented by a Turing machine induce permutation symmetries and low-rank blocks in its susceptibility matrix. We study this empirically on a set of deterministic finite automata (DFAs) and demonstrate that algorithmic features can be recovered by principal component analysis and clustering methods in susceptibility space.
[LG-42] SMILE: Bridging Continuous Optimization and Discrete Symbolic Recovery
链接: https://arxiv.org/abs/2609.04639
作者: Mansooreh Montazerin,Antonio Ortega,Ajitesh Srivastava
类目: Machine Learning (cs.LG)
*备注:
Abstract:Symbolic regression (SR) discovers closed-form mathematical expressions from data, offering interpretability beyond black-box models. Existing methods suffer from slow convergence in combinatorial search spaces and lack mechanisms to exploit compositional structure in the data. We introduce SMILE (Sine, Multiplication, Identity, Logarithm, Exponential), a hybrid framework that unifies continuous gradient-based optimization with discrete symbolic recovery through three stages: structural analysis of the data to identify the compositional hierarchy of the target expression, continuous optimization to learn parameters of a network that encodes the target expression using interpretable activations, and symbolic recovery through structured pruning, coefficient optimization, and rounding. This final stage distills the learned network into a compact expression with exact symbolic constants. We evaluate SMILE on SRBench across ground-truth and black-box datasets, with ablation studies validating each component. SMILE achieves the highest symbolic solution rate at the largest noise levels, demonstrating strong robustness where competing methods degrade substantially. It consistently lies on the Pareto front of accuracy versus complexity, recovering significantly simpler expressions in a fraction of the time required by the competing methods.
[LG-43] oo Rare to Learn: Prescribed Cyclone Tracks Degrade a Bay of Bengal Ocean Emulator NEURIPS2026
链接: https://arxiv.org/abs/2609.04635
作者: Sumaiya Islam
类目: Machine Learning (cs.LG)
*备注: 11 pages, 7 figures. In review at the NeurIPS 2026 Tackling Climate Change with Machine Learning workshop
Abstract:Neural ocean emulators are being proposed for regional forecasting in cyclone-exposed coastal seas, and a natural design choice is to hand the network the cyclone as a prescribed input. We test that choice in the Bay of Bengal and find it harmful. We withhold 15 whole cyclones spanning 65 to 150 kt from GLORYS12 reanalysis and compare two U-Nets that are identical except for four prescribed cyclone-track channels. Across three seeds the ocean-only model beats persistence in every run and the storm-conditioned model loses to it in every run, with the two skill ranges disjoint (p = 3.1e-5, paired across storms). The cause is exposure frequency rather than signal content: the channels are non-zero on only 7.9% of training days, so they are out of distribution the moment they activate. The extra error falls inside the prescribed storm footprint, and replacing the real cyclone map with a no-storm map at inference improves held-out storm forecasts by 7.5 to 16.4% in every seed. The conditioned network has learned a response to a rare signal that is confidently wrong.
[LG-44] GNN-Guided Graph Coarsening and Adaptive QUBO Penalties for the Capacitated Vehicle Routing Problem with Time Windows on a Quantum Annealer
链接: https://arxiv.org/abs/2609.04593
作者: Youssef Kamel Rezk,Paweł Gora
类目: Machine Learning (cs.LG)
*备注:
Abstract:Graph coarsening reduces the large Quadratic Unconstrained Binary Optimization (QUBO) formulations arising when vehicle-routing problems are solved by quantum annealing. Nearby customers with compatible time windows are merged into super-nodes, the reduced problem is solved, and the solution is expanded to the original graph. For the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW), existing coarsening heuristics require family-specific tuning and remain unreliable on random instances. We address these limitations on the Solomon benchmark using simulated annealing and a D-Wave Advantage2 processor. We first introduce adaptive penalty calibration. Uniform penalty scaling has little effect, whereas controlling the internal coefficient range substantially improves raw samples. Removing non-binding constraints, normalising binding ones, and scaling the remaining penalties reduces mean raw constraint violations from 33.0 to 0.06 at the same solver budget (p=3.7e-11, n=56). A variable-count-preserving control attributes this gain to conditioning rather than problem size. Second, we replace the hand-tuned merge score with a graph neural network (GNN) using one configuration across all families. At N=10, it achieves 100% feasibility across all Solomon families, including R-type (100% vs. 80% for the tuned heuristic). Across N=10,…,100, feasibility is 83% vs. 69%, with the GNN better or tied on 85/90 instance-size pairs. At N=80,100, the difference is significant (p=0.002; 25/25 pairs), while the QUBO remains approximately 5-6 times smaller. Finally, hardware experiments reproduce the conditioning effect at fixed logical variable count: feasible samples increase from 0.02% to 39% across 13 instances. Classical repair with local search remains a reference bound for end-to-end solution cost. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2609.04593 [cs.LG] (or arXiv:2609.04593v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2609.04593 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Paweł Gora [view email] [v1] Fri, 4 Sep 2026 00:49:57 UTC (225 KB)
[LG-45] Representation Redundancy and Structural Complexity in Finite-Field Inversion
链接: https://arxiv.org/abs/2609.04583
作者: Zheng Zhang,Na Zhang
类目: Machine Learning (cs.LG); Rings and Algebras (math.RA)
*备注:
Abstract:The representation chosen for a mathematical operation can affect both its algebraic form and its empirical learning difficulty. We study this phenomenon for inversion over (\mathbb F_2^n), with field elements expressed in varying ordered (\mathbb F_2)-bases. We prove that two ordered bases induce the same coordinate inversion map if and only if they belong to the same Galois orbit. Since every orbit has size (n), the correspondence between ordered bases and distinct inversion maps is exactly (n)-to-one. We then analyze three Boolean formulations of inversion. The reference formulation has algebraic degree (n-1) and joint ANF leap (1), the mixed representation formulation has degree (2(n-1)) and joint ANF leap (2), and the complete raw formulation has degree at most (3(n-1)) and joint ANF leap at least (n). Exhaustive computations agree with the theoretical results and bounds in the cases considered. Controlled experiments with multilayer perceptrons show the same ordering in learning difficulty, while Galois orbit redundancy provides only a limited generalization benefit under the tested conditions. These results show that exact redundancy among representations can coexist with changes in Boolean structure and learning behavior when the representation is exposed as part of the input.
[LG-46] Optimizer Memory Schedules for Outscaling the Overtraining Axis
链接: https://arxiv.org/abs/2609.04577
作者: Katie Everett,Shikai Qiu
类目: Machine Learning (cs.LG)
*备注:
Abstract:We investigate how optimizers scale across the overtraining axis and show that relative optimizer performance and optimal hyperparameters change substantially with training horizon. In particular, we study how matrix-preconditioned methods (Muon and SOAP) and a momentum-scheduled method (ADANA) scale relative to AdamW. We compare these four optimizers across models from 51M to 253M parameters and overtraining (OT) factors from 1x to 256x, sweeping the base learning rate at every setting. The preferred learning rate schedule can reverse across the overtraining axis, the best weight decay coefficient scales approximately as sqrt(OT), and longer horizons generally favor longer fixed memory. ADANA’s scaling advantage over AdamW persists after tuning AdamW’s fixed memory separately at each horizon. Log-time weight decay and momentum cooldown provide substantial gains for ADANA that compound as training increases. With this treatment, ADANA outscales AdamW with an exponent advantage close to that predicted by DANA theory on power-law random features. Muon and SOAP instead provide roughly constant token-efficiency advantages over AdamW across most of the measured range, although SOAP may gain further at the highest overtraining factors. ADANA begins behind both matrix-preconditioned optimizers but closes the gaps as training increases, surpassing Muon and becoming competitive with SOAP at our highest OT factors. These results establish training horizon as an essential axis for optimizer evaluation and design.
[LG-47] Fast Surrogate Modeling of Excitable and Oscillatory FitzHugh-Nagumo Dynamics with Parametric Neural Operators IJCAI
链接: https://arxiv.org/abs/2609.04549
作者: Andrew Franck,Justin Li
类目: Machine Learning (cs.LG)
*备注: Human Brain AI @ International Joint Conferences on Artificial Intelligence (IJCAI) 2026
Abstract:The FitzHugh-Nagumo (FHN) system serves as a simplified model of neuronal voltage dynamics, capturing the activator-inhibitor structure behind both isolated action potentials and the rhythmic spiking seen across the brain. Exploring its 5D physiological parameter space is important for neuromodulation and mapping voltage recordings back to biophysics, yet classical finite-difference solvers make rapid parameter sweeps expensive. We train parameter-conditioned Fourier Neural Operators (FNOs) as fast, differentiable surrogates for the FHN voltage and recovery fields on a one-dimensional spatial domain, conditioning each Fourier layer on the parameter vector \lambda = (D_u, D_v, a, b, \tau) via feature-wise linear modulation (FiLM). We apply a single bifurcation analysis that delimits the two distinct regimes the model spans, oscillatory (tonic firing) and excitable (action-potential propagation), and we train one operator in each. In the oscillatory regime the surrogate attains sub- 0.1% relative L^2 error on both fields, runs nearly three orders of magnitude faster than the finite-difference baseline, generalizes uniformly across the parameter space, and extrapolates to low single-digit percentage errors outside of the training bounds. In the excitable regime the same operator accurately reproduces the firing threshold and the c \propto \sqrtD_u conduction-velocity law and replicates full traveling pulses, fully capturing the excitable bifurcation structure rather than just smoothly interpolating fields.
[LG-48] Mitra-v2 Technical Report
链接: https://arxiv.org/abs/2609.04540
作者: Yefan Tao,Xiyuan Zhang,Xinyi Liu,Boran Han,Danielle Maddix,Haoyang Fang,Zhen Han,Jiading Gai,Xuanqing Liu,Michael Bohlke-Schneider,Yuyang(Bernie)Wang,Gerald Friedland,Kevan Mah,Chris Lee,Chris Kong
类目: Machine Learning (cs.LG)
*备注: 38 pages. Model weights, inference and fine-tuning code, and evaluation results: this https URL
Abstract:We introduce Mitra-v2, a tabular foundation model that delivers state-of-the-art performance on real-world classification and regression problems, from credit-risk scoring and clinical prediction to equipment-failure detection and house-price estimation. Mitra-v2 is trained only on synthetic data, with a pretraining distribution that is much larger and more diverse than Mitra-v1’s. Built on a small 2D Transformer backbone, Mitra-v2 supports longer contexts and larger feature spaces. Improved optimization lets it learn from this larger task distribution. We evaluate Mitra-v2 on the TabArena and TALENT benchmarks, comprising more than 300 real-world datasets under two evaluation protocols. On the full TabArena benchmark, Mitra-v2 delivers state-of-the-art performance at the level of the industry-scale TabFM and EXAONE Tabular models, while surpassing TabPFN-3 by a wide margin in both classification and regression. Mitra-v2 matches the 1.6B-parameter TabFM with only 5% of its size (77M parameters), delivering frontier performance at a fraction of the cost. On TALENT, Mitra-v2 remains among the leading models, clearly outperforming TabPFN-3 and TabICLv2. It also ranks first on classification tasks with more than ten classes, even though it was pretrained only on tasks with at most ten classes. These results make Mitra-v2 one of the strongest and most broadly applicable open tabular foundation models released to date. We release the model weights, the inference and fine-tuning code, and our evaluation results under the Apache-2.0 license.
[LG-49] Distilled Continuous Diffusion Language Models Can Write Code in Few Steps—or One
链接: https://arxiv.org/abs/2609.04531
作者: Fred Zhangzhi Peng,Kaiwen Zheng,Anru R. Zhang
类目: Machine Learning (cs.LG)
*备注: Code and model checkpoints: this https URL
Abstract:Language generation is almost universally treated as a sequential process: autoregressive models emit one token at a time, while diffusion language models replace token-level seriality with a long trajectory of iterative refinement. In this work, we introduce PlaidQ, a 0.7B continuous diffusion language model for code generation, and show that its trajectory can be aggressively distilled into only a few denoising steps—or even one, enabling efficient code generation. PlaidQ repurposes a pretrained autoregressive model as a bidirectional denoiser over continuous token embeddings. We distill PlaidQ with distribution matching for few-step generation and paired-trajectory supervision for one-step generation. At matched model scale, PlaidQ is competitive with discrete diffusion language models on code generation. Distillation then shifts the quality–compute frontier: a 16-step student reaches 31.78 and 40.49 pass@10 on HumanEval and MBPP+, surpassing the same PlaidQ teacher sampled for 512 steps. At the extreme, paired-trajectory distillation achieves 7.07 pass@1 on HumanEval with a single denoising step, producing functionally correct programs. Together, these results establish continuous diffusion as a viable path to few-step and one-step code generation. Broadly, continuous diffusion is not merely another representation for language: it provides an interface through which language models can inherit the acceleration and distillation machinery of continuous diffusion modeling. Training and inference code and model checkpoints are available at this https URL.
[LG-50] An Energy-Based Conservative-Dissipative Latent Neural Evolution Operator for Magnetization Dynamics
链接: https://arxiv.org/abs/2609.04530
作者: Sebastian Schaffer,Lukas Exl
类目: Machine Learning (cs.LG); Numerical Analysis (math.NA); Computational Physics (physics.comp-ph)
*备注: 24 pages, 13 figures
Abstract:We develop an energy-based reduced-order model for micromagnetic magnetization dynamics that couples a convolutional autoencoder to a structured latent neural ordinary differential equation. Motivated by the precessional-dissipative structure of the Landau-Lifshitz-Gilbert equation, the latent vector field is generated from the gradient of a learned scalar potential through an antisymmetric operator and a symmetric positive-semidefinite dissipative operator. This potential is learned in nonunique latent coordinates and is not identified with the Gibbs free energy, but decreases monotonically along autonomous continuous-time solutions, while the antisymmetric component permits motion along its level sets. The encoder, decoder, latent energy, and operators are trained jointly on short trajectory windows using latent and decoded-rollout losses alone, without time-derivative supervision, physical-energy labels, or dissipation penalties. At inference, an initial state is encoded once, evolved in latent space, and decoded only at the requested output times, enabling substantially cheaper trajectory prediction than the micromagnetic solver used to generate the training data. We compare quadratic, deep, and additive deep-quadratic latent energies on two datasets parameterized by field amplitude and generated for the two applied-field directions of the NIST \mu MAG Standard Problem 4. Dissipative-only and antisymmetric-dissipative models achieve comparable accuracy on short training-style windows but differ substantially on uninterrupted rollouts, for which the antisymmetric-dissipative models provide markedly more accurate trajectory predictions. The deep-quadratic energy gives the best overall accuracy for both field directions and exhibits slower error growth when rollouts are extended to twice the training horizon.
[LG-51] Client-Side Probing of Deleted Ridge Statistics in Federated Unlearning
链接: https://arxiv.org/abs/2609.04475
作者: Yijun Quan,Giovanni Montana
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Federated unlearning aims to remove a client’s data from a shared model without retraining from scratch. Some efficient systems make deletion exact by storing compact, additive summaries of the training features and broadcasting an updated linear classifier after every accepted change. We show that these broadcasts can also reveal the hidden summaries. A malicious client can submit known changes, use the returned classifiers to identify the server state, and compare states immediately before and after an isolated deletion. This exposes the deleted sample, class, or client summary and can enable its reinsertion. We characterize exactly when the observations contain enough independent information, give a matching optimal construction for unrestricted probes, and derive a more realistic estimator based on additions formed from the attacker’s own data. On MNIST and CIFAR-10, high-precision broadcasts permit exact label recovery for every tested sample deletion with both probe types. Lower-precision broadcasts sharply reduce fine-grained recovery, and insufficiently diverse responses prevent identification altogether. Unrestricted probes are readily detected by their size; most individual attacker-data additions resemble honest batches, although we do not claim that the complete sequence is inconspicuous. The results identify a concrete privacy and integrity risk, its algebraic cause, and practical limits involving broadcast precision, update verification, response rate, and concurrent activity.
[LG-52] Nested Inductive Bias Framework for SPD Manifold Learning
链接: https://arxiv.org/abs/2609.04466
作者: Tushar Das
类目: Machine Learning (cs.LG)
*备注:
Abstract:In Geometric Deep Learning, inductive biases serve two primary functions: enforcing manifold constraints and embedding relational priors. Currently, representation learning on SPD manifolds frequently relies on pullback Euclidean metrics, such as the Log-Euclidean Metric, to satisfy the former. While computationally efficient in avoiding domain boundary violations, these metrics induce a flat geometry that may fail to capture the intrinsic relational priors of datasets. While metrics such as the Poincaré metric are widely utilized to induce domain-aligned relational priors, generalizing them from standard vector representations to the SPD manifold has remained a challenge. To bridge this gap, we introduce a Nested Inductive Bias framework that utilizes a two-stage diffeomorphic composition to formally pull back non-Euclidean target geometries onto the SPD manifold. This framework enables the construction of curvature-aligned Riemannian classifiers that simultaneously respect matrix constraints and the latent relational geometry of the data. Empirical evaluations on kinematic and signal processing benchmarks, together with synthetic experiments, demonstrate that deep manifold networks experience degradation in class separability unless the metric curvature aligns with the intrinsic data distribution. Furthermore, for standard vectorized architectures, we propose the Rational Conformal Metric (RCM), designed to establish state-of-the-art geometric robustness against outliers by bounding the representation space.
[LG-53] On-board ML for Trace Gas detection in Imaging Spectroscopy data
链接: https://arxiv.org/abs/2609.04458
作者: Vít Růžička,Adam Chlus,Andrew Thorpe,David R. Thompson
类目: Machine Learning (cs.LG); Performance (cs.PF)
*备注: 2 pages; presented as a short (2 pages) paper at the IEEE HPEC26
Abstract:Data collected during aerial and spaceborne imaging spectroscopy campaigns enables the detection of transient events such as trace gas emissions. However, current processing pipelines depend on slow, on-the-ground processing, which delays the time to information of each detected event and prohibits immediate follow-up actions. During the Tokyo Field Campaign of March 2026, we explored on-board processing of Imaging Spectroscopy data from the equipped AVIRIS-5 sensor. Due to communication bottlenecks, full datacubes cannot be downlinked immediately during the flight. Instead we downlink the potential events predicted by our efficient and small machine learning model. We show the first on-board detection of methane point source emission with Imaging Spectroscopy data using Edge ML.
[LG-54] A Repeated-Measurement Study for Cultural Analytics of English Song Lyrics Using Five Large Language Models
链接: https://arxiv.org/abs/2609.04428
作者: E. Cho Smith,Samuel Ho,Dawn Laux
类目: Machine Learning (cs.LG)
*备注: 6 pages, 3 figures, Conference: AIxHEART 2026
Abstract:Large language models (LLMs) are increasingly used to annotate cultural texts at scales that are impractical for human coders. However, before their outputs are treated as measurements of latent social constructs, it is necessary to establish whether those measurements are reliable. This study evaluates five LLMs as zero-shot annotators of four social constructs expressed in English song lyrics: self-esteem, self-control, seeking belonging, and seeking recognition. Using repeated annotations of a large lyric corpus, we examine three properties of LLM-based measurement: consistency across repeated runs, convergence across models, and transferability of consensus labels to supervised classification. The findings show that LLM-based measurement is not uniformly reliable across constructs. Self-esteem exhibits the strongest repeated-measurement reliability across models, while seeking recognition is generally less stable; self-control and seeking belonging show intermediate but model-dependent reliability. Downstream classification further indicates that consensus LLM labels contain learnable signal, although transferability does not itself establish construct validity. Repeated-measurement stability and cross-model convergence should therefore be reported before LLM annotations are treated as scalable measurements in cultural analytics.
[LG-55] Beyond a Universal Forecasting Selector: Demand-Conditioned Model Selection across Demand Patterns and Horizons
链接: https://arxiv.org/abs/2609.04425
作者: Adolfo González
类目: Machine Learning (cs.LG)
*备注: 30 pages, 7 figures
Abstract:Forecasting-model selection remains difficult in heterogeneous demand because the most suitable decision rule may vary with demand structure, data availability, and forecasting horizon. This study examines whether the selector itself should be treated as a context-dependent component of the forecasting process. Five selection mechanisms - RMSSE, ERA, OWA, CCG-AHSC, and CCG-AHSCD - are compared across 24 optimized forecasting models, nine datasets, three training-testing partitions, and horizons from 1 to 12 cycles. Selector performance is evaluated ex post using Global Relative Accuracy (GRA), statistical tests, and a best-attainable-model reference. No selector dominates across all conditions. CCG-AHSC and CCG-AHSCD are more competitive for Smooth demand and several Erratic configurations, whereas OWA and ERA perform better in Intermittent and Lumpy settings. Selector suitability also changes with historical data availability and horizon, supporting a context-dependent rather than universal approach to forecasting-model selection.
[LG-56] uning Collective Patterns to Alleviate Congestion in Shared AI Clusters
链接: https://arxiv.org/abs/2609.04417
作者: Eashan Gupta,Yongzhou Chen,Apoorve Mohan,Pavlos Maniotis,Abdullah Kayi,Radhika Mittal
类目: Networking and Internet Architecture (cs.NI); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注:
Abstract:Distributed AI training involves recurring rounds of data exchange between multiple pairs of GPU nodes. Slowdown in even one flow due to congestion can cause the entire communication round to slowdown. Current approaches for evading congestion in AI clusters assume global control over the entire workload (e.g. coordinating the schedule of all jobs) or assume infrastructural support (e.g. adaptive routing in switches). They are thus ill-suited in a shared cloud setting where AI jobs belonging to one user can face external congestion from other users’ jobs or background traffic beyond its own control. In this paper, we build a system, REACT, that tunes the recurring pattern of data exchange between GPU nodes (known as communication collectives) in response to congestion. REACT works at the application (communication library) layer, where it detects congestion at runtime using readily available flow stats, and tunes the collective pattern to alleviate congestion - changing the set of incident flows while retaining the semantics of information exchange (e.g. selecting which node aggregates data in an AllReduce tree). REACT requires no explicit support from the underlying network infrastructure and can be unilaterally deployed by individual users in a shared cloud setting. We prototype REACT as a shim layer over NCCL, and evaluate it on a shared academic GPU cluster - enabling REACT improves communication performance (algorithm bandwidth) by 13%-38% under network congestion. Our simulations across a range of congestion scenarios further reveal up to 75% performance improvement, highlighting the effectiveness of our approach.
[LG-57] Disentangling Attention in Deep Operator Learning: A Controlled Study of Data-Driven and Physics-Informed Architectures
链接: https://arxiv.org/abs/2609.04407
作者: Amar Alem Koric,Qibang Liu,Seid Koric
类目: Machine Learning (cs.LG)
*备注: 25 pages, 13 figures, 5 tables
Abstract:Deep neural operators learn mappings between input functions and complete PDE solution fields, enabling forward evaluations of new problem instances orders of magnitude faster than conventional numerical solvers. Attention mechanisms have recently been introduced into neural operators, but most studies change several architectural components at once, making it difficult to identify what actually improves accuracy. This work presents a controlled and systematic study of five deep operator network (DeepONet) variants with distinct attention mechanisms, trained under both data-driven and physics-informed regimes, to isolate the effects of cross-attention, self-attention, tokenization, and attention depth. We evaluate them on a source-driven transient one-dimensional nonlinear diffusion-reaction equation, a transient one-dimensional viscous Burgers equation with variable initial conditions, and a two-dimensional Poisson heat-conduction problem with heterogeneous source fields. Per-sensor tokenization with cross-attention reduces the mean relative L_2 error of the classical DeepONet in all benchmark-training combinations by factors of 2.4-28.0, while the best attention configurations reach 3.5-32.3. Branch self-attention paired only with dot-product fusion is inconsistent, degrading the one-dimensional problems while helping the more complex two-dimensional source field; added on top of cross-attention it improves all six cases, though by less than cross-attention fusion alone. Global pre-mixing provides no consistent benefit. Increasing cross-attention depth further improves accuracy, but with diminishing returns and a substantially higher cost under physics-informed training. Overall, query-dependent cross-attention is the most reliable mechanism, whereas branch self-attention is most useful for large, spatially complex functional inputs.
[LG-58] Candidate Comparability Before Promotion: Conditional Validation in Adaptive Network Intrusion Detection
链接: https://arxiv.org/abs/2609.04388
作者: Roberto Fernández-Barrios,Iker Pastor-López,Amaia Pikatza-Huerga,Pablo García Bringas
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: 29 pages, 3 figures, 11 tables. Supplementary material (Online Resource 1) is included as an ancillary file
Abstract:Adaptive network intrusion detection systems retrain classifiers after drift alarms, but an alarm detects change; it does not establish that a challenger should replace the deployed incumbent. Promotion is security-relevant because it changes the model responsible for subsequent attack detection, and evaluating it has a methodological problem: promotion conclusions may depend on how the challenger was constructed and on how much evidence supports it. We test that dependence on CICIDS2017, UNSW-NB15 and ToN-IoT with self-contained challenger pipelines, nested candidate-size controls, a common-harness comparison of nine update policies, and a final sensitivity confining every exact feature vector to one evaluation, training or probe role. Incumbent-owned frozen preprocessing amplified apparent promotion harm; with self-contained challenger pipelines the mean full-drift harm did not persist. Raising nominal candidate evidence from 512 to 2,000 samples per class improved promotion under pool-constructed progressive drift by +0.53, +1.67 and +0.38 balanced-accuracy points: positive and statistically resolved in all three benchmarks, but materially benchmark-dependent rather than homogeneous, and driven mainly by fewer false positives. Policy conclusions were partially robust: policy ordering changed with candidate comparability, no policy globally dominated, and earlier compatibility statements for a label-free estimator and a calibrated ensemble narrowed. Validation helped evidence-disadvantaged challengers but added no average benefit at parity. Thirteen replays on real, time-ordered traffic showed no net harm from always deploying. Challenger construction and evidence should be controlled, reported and interpreted explicitly when promotion is evaluated.
[LG-59] Privacy Failure in Split-LLM Training The Returned Gradient Nullifies the Decoys
链接: https://arxiv.org/abs/2609.04382
作者: Georgios Politis,Evangelos Pappas
类目: Cryptography and Security (cs.CR); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注:
Abstract:We present a systems-security case study of a two-node split-LLM training system whose privacy evaluation passed while leaving an observable channel untested. The Trusted Local Node (TLN) sends protected activations to the Untrusted Cloud Node (UCN), the UCN returns its output, and TLN, holding the private loss, returns the output gradient. The frame the UCN receives mixes real rows with decoys, and the loss ignores the decoys. Their gradients are exactly zero, so the pattern of zeros reveals which rows were real. We measure it with a protocol fixed in advance: a leak injected at known strength to prove the instrument can see one, a shuffled-label control to prove it does not report absent leaks, and a threshold set before the runs. Across nine seeds, the zeros identified the real rows on every frame, 4,096 of 4,096 per run. An attack on the frame contents recovered about one extra token per hundred over a constant-guess baseline (+0.65 to +1.50 percentage points); the shuffled controls recovered nothing. A second set of runs repeated this on a configuration that keeps model quality within budget, so the finding is not confined to a setting nobody would deploy. On both datasets, every such run passed the forward-channel privacy check and the quality check, yet failed that same check once the returned gradient was included. Clipping and noising each row of the gradient closed the leak for about 0.01 nats of held-out cross-entropy. The system is not thereby safe: five classes of attack, including those accumulating observations across training steps, were never measured.
[LG-60] On the Abundance of Critical Points of the t-SNE Energy
链接: https://arxiv.org/abs/2609.04379
作者: Nakul Haridas,Ryan Murray
类目: Machine Learning (cs.LG); Analysis of PDEs (math.AP); Statistics Theory (math.ST)
*备注:
Abstract:This paper considers the energy landscape of the t-SNE algorithm. While this algorithm has enjoyed broad adoption, the non-convexity of the associated energy has made it difficult to rigorously understand what the algorithm captures in many settings. In particular, a number of well-known numerical examples, several of which are reproduced in this article, suggest a complicated energy landscape with many local minimizers that do not respect the topology or clustering structure of the underlying data. This work seeks to provide first steps towards a rigorous explanation of these phenomena. Specifically, for a general family of energies, which include both the original t-SNE algorithm and recently identified large data limits, and for densities in feature space which obey a continuous symmetry, we construct infinite families of distinct critical points. These critical points are based upon identifying pairs of discrete symmetries, one in the original feature space and the other in the target embedding space, which are preserved under gradient dynamics. These critical configurations exhibit many characteristics, such as topology breaking and spurious clustering, which are often observed empirically. Finally, numerical and analytical examples are given throughout as a means of illustrating the approach.
[LG-61] A Quantum Variational Approach to Prototypical Recurrent Unit
链接: https://arxiv.org/abs/2609.04354
作者: Mahyar Sadeghi Garjan,Tommaso Cesari,Michel Barbeau
类目: Machine Learning (cs.LG)
*备注: 12 pages, 5 figures, 4 tables
Abstract:We introduce a lightweight Quantum Prototypical Recurrent Unit (QPRU) that requires significantly fewer parameters than both classical recurrent architectures, such as Long Short- Term Memory (LSTM) and Gated Recurrent Unit (GRU), and quantum variants, including Quantum LSTM (QLSTM) and Quantum GRU (QGRU). Despite its compact design, the QPRU achieves competitive forecasting performance, matching state-of-the-art baselines while offering important structural and practical advantages, including enhanced scalability and a reduced number of trainable parameters.
[LG-62] A Constraint-Aware Generative Framework for Synthetic Origin-Destination Demand in Logistics Networks
链接: https://arxiv.org/abs/2609.04345
作者: Leian Chen
类目: ystems and Control (eess.SY); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 13 pages
Abstract:Large-scale logistics networks require synthetic data generation capabilities to support scenario-based planning under novel conditions-such as network reconfiguration and demand shocks. Existing approaches, which rely primarily on historical observations, lack the ability to generate demand patterns that adapt to changes in network topology while respecting operational constraints. We propose a constraint-aware conditional generative framework for synthetic origin-destination demand generation in hierarchical logistics networks. The framework models demand as a conditional distribution over destinations given each origin, enabling topology-aware synthesis that is both topologically realistic and operationally feasible. Operational guidance is incorporated directly into the generative objective via differentiable constraints, while a flexible conditioning mechanism supports various operational contexts and adaptation to evolving network configurations. We instantiate the proposed framework based on a conditional generative model. Experimental validation on industrial real fulfillment and transportation network demonstrates 16% improvement over graph neural network baselines, 87% operational compliance, and efficient cold-start adaptation, enabling applications in capacity planning, network design evaluation, and routing optimization.
[LG-63] Modular Deep Recurrent Neural Network: Application to Quadrotors
链接: https://arxiv.org/abs/2609.04339
作者: Nima Mohajerin,Steven L. Waslander
类目: Machine Learning (cs.LG)
*备注:
Abstract:A modular deep Recurrent Neural Network (RNN) is introduced to facilitate the process of deploying various architectures of RNNs, and to automatically compute derivatives for gradient-based learning methods. The modularity leads to a set of new architectures, one of which includes feedforward inter-layer connections. By adding feedforward inter-layer connections in a multi-layer RNN, it is observed that the capability of the RNN to learn and model high-order dynamics and nonlinearities is significantly improved. The problem of vanishing/exploding gradient in space for a multilayer RNN is also alleviated using feedforward connections. These results are demonstrated using a quadrotor case study, for which a model of the altitude dynamics is learned with our particular network structure, while existing methods are unable to generalize as quickly or at all.
[LG-64] BER-PEF: Unified Human Mobility Predictability Evaluation via Bayes Error Rate Estimation
链接: https://arxiv.org/abs/2609.04292
作者: En Xu,Jingtao Ding,Zhiwen Yu,Yong Li
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注:
Abstract:Human mobility predictability concerns the best prediction performance attainable from a given target and input information, but its ground truth is not directly observable on real mobility data. We present BER-PEF, a Bayes-error-rate-based framework that converts BER estimation into mobility predictability estimation and provides a unified protocol for comparing estimators without observable ground truth. The framework maps symbolic sequences, numeric trajectories, contextual features, and learned representations into a common feature–label space, then evaluates estimator outputs along controlled perturbation curves against a shared predictability reference interval by measuring deviations below the interval, above the interval, and across the full interval. Experiments on Foursquare NYC and TKY, GeoLife, and T-Drive show that several BER-based estimators achieve lower reference discrepancy than existing predictability methods on symbolic sequences and numeric trajectories, while their estimates track changes in empirical prediction performance under perturbation. Additional analyses show that contextual inputs and multiple structured representations can be evaluated under the same protocol, and that aggregating evidence across multiple perturbation levels provides a more reliable basis for estimator selection than relying on a single unperturbed observation. BER-PEF therefore offers a unified and verifiable path for evaluating predictability estimators on heterogeneous mobility data when ground-truth predictability is unavailable.
[LG-65] Quantum-Assisted Memory-Efficient Training for Parameter-Intensive Wi-Fi-Based Human Activity Recognition
链接: https://arxiv.org/abs/2609.04271
作者: To Truong An,Jie Zhang,Guolin Yin,Junqing Zhang,Yanjiao Li,Trung Q. Duong,Simon L. Cotton
类目: Machine Learning (cs.LG)
*备注: 18 pages, 9 figures
Abstract:Wi-Fi-based human activity recognition (HAR) has become an important part of integrated sensing and communications, paving the way for a range of context-aware services. However, most existing Wi-Fi-based HAR systems rely on deep learning (DL) models that are computationally and memory intensive in both training and inference, which poses significant challenges for real-world deployment. Conventional training requires simultaneous updates of millions of parameters, leading to prohibitive memory consumption. In this paper, we propose a novel quantum-assisted memory-efficient training framework (Q-MET) designed to improve efficiency in both training and inference. Q-MET utilizes a hybrid quantum classical neural network to indirectly generate parameters for HAR models, significantly reducing the trainable parameter count compared to direct optimization. To further support the deployment on resource-constrained devices, we integrate structured pruning during the training phase. Experimental results demonstrate that Q-MET achieves a 90% to 95% reduction in trainable parameters compared with conventional backpropagation-based DL training while maintaining or even exceeding classical classification accuracy. Additionally, Q-MET supports lightweight inference through structured pruning, achieving 75% to 85% model sparsity with less than 2% loss in classification accuracy. To the best of our knowledge, this work represents the first quantum-assisted approach to simultaneously tackle memory inefficiencies in both the training and inference stages of HAR systems.
[LG-66] A Data Fusion Framework for Grounding Aerospace Surrogate Model via Experimental Wind-Tunnel Observations
链接: https://arxiv.org/abs/2609.04267
作者: Nitin Nagesh Kulkarni,Dheeraj Vemula,Yin Yu,Peter Lyu,Juan J. Alonso
类目: Machine Learning (cs.LG)
*备注:
Abstract:Aerodynamic surrogate models trained on high-fidelity CFD data reproduce numerical predictions of both scalar outputs and entire fields accurately, yet their predictive fidelity is limited by systematic discrepancies between CFD and experimental observations. We present an experimentally grounded correction framework that adapts a CFD-trained deep learning surrogate using wind-tunnel PSP measurements. A Geotransolver surrogate trained on 2,300 high-fidelity CFD simulations of the NASA CRM wing-body configuration, spanning geometric variation, Mach 0.70-0.85, and angles of attack 0 to 4 degrees, reproduces the CFD integrated aerodynamic forces and pitching moment to R2 0.99 but does not match the experimental data. To incorporate experimental information without retraining the surrogate, a correction network is trained on spatially registered PSP measurements at two freestream Mach numbers (0.70 and 0.85) across the same angle-of-attack range, learning the discrepancy between the surrogate-predicted and experimentally measured surface-pressure distributions. At Mach 0.85 the correction substantially improves agreement with PSP, particularly at the wing suction peak, shock location, and subsequent pressure recovery, reducing both the magnitude of the prediction error and the fraction of wetted surface on which it exceeds 0.05 in Cp, and it does so from a limited experimental dataset without modifying the pretrained surrogate parameters. On held-out angles of attack the grounded surrogate agrees with measurement to within 2.3-2.7% of the measured Cp range, and outperforms direct interpolation between the measured conditions at every state tested. Experimental measurements can therefore ground a large-scale simulation-trained surrogate by learning systematic CFD-to-experiment discrepancies while preserving its generalization capability and computational efficiency.
[LG-67] Compute-in-Memory Attention: A Time-Domain Analog Softmax Circuit with RC-Tunable Temperature
链接: https://arxiv.org/abs/2609.04266
作者: Ankur Singh,Ashish Gautam,Shruti R. Kulkarni,Guojing Cong
类目: Hardware Architecture (cs.AR); Machine Learning (cs.LG)
*备注: 13 page, 16 figure
Abstract:Softmax is a key operation in Transformer attention, but its exponentiation and normalization add significant overhead in compute-in-memory (CIM) accelerators, especially when analog attention scores must first be converted to the digital domain. This work presents a tunable-temperature analog softmax circuit in GlobalFoundries 22-nm fully depleted silicon-on-insulator (FDSOI) technology that operates directly on CIM-generated score voltages without intermediate analog-to-digital conversion. Each input score is converted into a time-domain event using a shared falling ramp. The corresponding comparator transition samples an RC-decaying reference to generate an exponential weight, which is then processed by an in-circuit normalization stage. In contrast to analog softmax circuits that rely on transistor weak-inversion behavior for exponentiation, the proposed architecture controls the softmax response through the ramp slope and RC time constant, enabling programmable effective temperature. The 128-element architecture is evaluated using transistor-level and post-layout extracted simulations, including multi-level input vectors, capacitance variation and mismatch, process and temperature variation, monte carlo analysis, and shared-interconnect parasitics. The complete 128-element implementation occupies 9453.42~ \mu\mathrmm^2 including the shared global ramp circuitry, while each replicated softmax element occupies 70.2~ \mu\mathrmm^2 . The circuit achieves a 242.97-ns evaluation latency at 13.44~mW total power, corresponding to 25.5~pJ per output element. The simultaneous 128-element evaluation achieves an RMSE of 24.46~mV relative to the ideal softmax response. The extracted circuit characteristics are further incorporated into a MemTorch-based hardware-aware Transformer model, where the proposed softmax achieves a validation loss within 2.5% of the ideal-softmax baseline.
[LG-68] ProToMEx: Rapid Interpretable Explanations via Structured Representations
链接: https://arxiv.org/abs/2609.04265
作者: Athina Georgara,Adarsh Valoor,Sarvapali D. Ramchurn
类目: Machine Learning (cs.LG)
*备注:
Abstract:Existing post-hoc explainers for machine learning classifiers primarily focus on feature attribution, assigning importance scores to individual features. While valuable, this approach struggles to articulate the complex, combinatorial patterns that often drive a model’s decision-making process. To overcome this limitation, we introduce ProToMEx, a new paradigm for explainability that leverages Probabilistic Topic Models (PTMs). Our model-agnostic framework learns latent ‘‘topics’’ that represent distinct, high-level reasons for a classification, moving beyond simple feature importance to reveal underlying semantic structures. ProToMEx naturally provides both global explanations of a model’s overall behaviour and local explanations that can disentangle multiple co-existing reasons for a specific prediction. We demonstrate empirically that ProToMEx not only produces explanations of comparable fidelity to popular methods like SHAP and LIME but also drastically reduces the amortised computational cost of generating local explanations, making it highly suitable for real-time applications. Specifically, we show that ProToMEx is ~30-40x faster than SHAP and LIME over standardised tabular datasets and synthetic datasets.
[LG-69] Spectral-Target Physical Latent Structuring for JEPA-Style World Models
链接: https://arxiv.org/abs/2609.04264
作者: Penghao Zhu,Salvatore Penachio,Kaustav Mukherjee,Aneesh Jonelagadda
类目: Machine Learning (cs.LG)
*备注: 9 pages, 4 figures
Abstract:Latent world models have become increasingly popular as a method to predict and plan in latent space rather than pixel space. Recent architectures, such as LeWorldModel (LeWM), jointly train the encoder and predictor using regularization techniques like SIGReg to prevent representation collapse. Even with such regularization preventing representation collapse, we identify a new world model failure mode of \textitphysical representation laziness, particularly noted in highly dynamic environments. For these lazy cases, the learned latent states do not collapse but nonetheless fail to represent key physical properties, causing ubiquitous downstream planning failure. To resolve this issue, we propose training-time auxiliary supervision with a lightweight “Fourier auxiliary head”, which enforces physically-informed structuring of the latent space with no additional inference-time cost and can be generalized to any environment. Experimentally, we show that the auxiliary head substantially improves planning success rates in dynamic environments where the baseline LeWM exhibits physical representation laziness. It also leads to modest improvements in other environments, even when the baseline does not exhibit physical representation laziness. We further observe superior planning performance being accompanied by higher latent space correlations with key physical properties, indicating both the ability of our method to physically structure latent states and the potential planning-side benefit to the learned representation being physically structured. We also see in low-data regimes, auxiliary supervision is particularly impactful in increasing success rate. These findings support the use of our Fourier auxiliary head method to improve both overall success rate and data efficiency, while avoiding representation laziness in latent world models.
[LG-70] Low-Latency Spell Correction for Japanese Music Search Queries
链接: https://arxiv.org/abs/2609.04262
作者: Anshul Garg,Pavni Tandon,Karan Bhukar,Tanmay Khandelwal,Ujjal Kumar Dutta
类目: ound (cs.SD); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
*备注: 9 pages, 5 tables
Abstract:Spell correction for Japanese search queries presents unique challenges due to the co-existence of four writing scripts (Latin/romaji, hiragana, katakana, and kanji) and the distinct error patterns each script induces. We present a compact BART-based sequence-to-sequence model (3 encoder + 3 decoder layers) designed for low-latency spell correction of Japanese music search queries. The core contribution lies in a script-aware synthetic misspelling generation pipeline that produces realistic training data by combining keyboard-layout models (QWERTY and flick input), phonetic confusion priors mined from real query logs, voiced/unvoiced consonant alternations, and kana case errors. A key design decision is normalizing mixed-script catalog titles to a single canonical script before misspelling synthesis, which we show is critical for reducing model hallucinations. We train a custom byte-level BPE tokenizer on the target music catalog to handle all four scripts in a unified vocabulary. Experiments on a curated evaluation set show that our model achieves an exact-match accuracy of 41.09% and a character error rate (CER) of 11.62%, outperforming edit-distance baselines and achieving the lowest character error rate among all evaluated systems while maintaining sub-4ms inference latency on a single GPU. We further analyze performance across individual scripts and mixed-script queries, demonstrating the effectiveness of script-aware data augmentation through systematic ablation studies.
[LG-71] PAC-Bayesian Reconstruction Guarantees for Time Series Variational Autoencoders
链接: https://arxiv.org/abs/2609.05212
作者: Chloé Hashimoto-Cullen,Ghislain Agoua,Benjamin Guedj,Sylvain Le Corff
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 20 pages
Abstract:Forecasting time series accurately is critical for applications with complex data ranging from energy systems to healthcare and finance. Among current state of the art models, generative latent variable models are increasingly implemented; yet principled generalisation guarantees for modern latent variable models remain limited. In particular, while Variational AutoEncoders are widely used for sequential data, their theoretical analysis is largely restricted to i.i.d. settings. In this work, we develop a PAC-Bayesian framework for latent variables models applied to time series. Building on reconstruction-based bounds, we extend PAC-Bayesian guarantees to Markovian latent structures, capturing temporal dependencies through a sequential generative process. These guarantees do not grow with the length of the trajectory. Our bounds depend on assumptions which are common in the literature; we provide an example framework where they would be verified to show that they are not as restrictive as they may seem.
[LG-72] FluxDisco: Symbolic Regression for Stoichiometric Dynamical Systems via Monte Carlo Graph Search
链接: https://arxiv.org/abs/2609.05207
作者: Cassandra Durr(1),Alvaro Köhn-Luque(2),Chris Jewell(1),Lloyd A. C. Chapman(1) ((1) Lancaster University, (2) University of Oslo)
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Data Analysis, Statistics and Probability (physics.data-an)
*备注:
Abstract:Dynamical symbolic regression methods identify governing differential equations from noisy data, balancing interpretability and predictive accuracy. However, standard methods often produce expressions that violate known physical laws. To address this, we propose FluxDisco, a physics-informed framework tailored for flux-based, stoichiometric ODE systems. By leveraging a known stoichiometry, we reduce the expression search space and ensure physical adherence. Our framework adapts the Monte Carlo Graph Search algorithm for the unique challenges associated with joint flux discovery of stoichiometric systems. We evaluate our method across a range of physical and biological systems, demonstrating its ability to accurately recover governing dynamics through interpretable equations.
[LG-73] Impact of Data Loss in Postprocessing on Training and Inference of Quantum Neural Networks
链接: https://arxiv.org/abs/2609.05060
作者: Soraya V. Panambalom,Edoardo Altamura,Nick Chancellor,Jonte R. Hance
类目: Quantum Physics (quant-ph); Emerging Technologies (cs.ET); Machine Learning (cs.LG); Computational Physics (physics.comp-ph)
*备注: 8 pages, one figure
Abstract:As quantum hardware scales to larger devices, the classical software layers that interface with it must evolve in step. Postprocessing routines developed and tested primarily in simulator settings can encode assumptions that no longer hold on utility-scale devices, leading to data loss that can be difficult to detect from high-level model outputs alone. We present a case study of \textttSamplerQNN, the sampling-based quantum neural network class in the Qiskit Machine Learning library. Here, the postprocessing method applies a filter that assumes measurement bit-strings are in virtual qubit space. On our quantum hardware runs, where bit-strings span over 100 physical qubits, this filter led to the loss of 85 to 99.6% of valid measurement shots, depending on the transpiler’s qubit placement. The resulting probability vector is unnormalised, allowing distorted prediction and loss values to propagate through the model without an API-level warning. We demonstrate the impact across five experiments on two IBM backends: for inference, accuracy drops from 0.94 to 0.39 on the same raw measurements; for training, the loss signal is compressed by 22 to 27 \times , substantially reducing the sensitivity of the optimiser to the objective landscape. The behaviour arises in all released versions of the library (0.8.4 to 0.9.0). We implemented a layout-based marginalisation fix, merged into the GitHub codebase as Pull Request #1041, that makes \textttSamplerQNN postprocessing forward-compatible with current and upcoming hardware.
[LG-74] An Analysis of Self-supervised Pre-training with Dependent Samples
链接: https://arxiv.org/abs/2609.05031
作者: Maximilian Fleissner,Debarghya Ghoshdastidar,Samory Kpotufe
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Self-supervised learning relies on so-called data augmentations \phi(x) of unlabeled datapoints x — for example, masking random pixels in an image x — that should leave the label of x invariant and are often used to learn a lower-complexity invariant subspace \cal V for downstream tasks. In practice, such augmentations \ \phi_l(x_i) \ are pooled together to learn \cal V , despite obvious inter-dependencies between different augmentations \phi_l(x), \phi_k(x) of the same datapoint x . However, theoretical works on the subject typically consider procedures that avoid such dependencies, and are therefore limited to operate on smaller subsets of independent data. We show in this work that pooling augmentations together, despite inter-dependencies, is a better alternative than the baseline of partitioning the data into subsets of independent data. More precisely, in the context of estimating \cal V , the statistical estimation error bounds for pooling are never worse than the partitioning baseline, and in some cases — such as masking or noise injection-based augmentations over a shallow neural network — naive pooling leads to faster rates in terms of the number of augmentations. The benefits of pooling are particularly prominent when the correlations between different augmentations \phi_l(x), \phi_k(x) have mild effects on estimation or help decrease the estimation variance. The analysis, therefore, yields new insights into the success of pooling augmented samples in self-supervised pre-training, and provides an intuition behind the practical preference towards using many augmentations. Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG) Cite as: arXiv:2609.05031 [stat.ML] (or arXiv:2609.05031v1 [stat.ML] for this version) https://doi.org/10.48550/arXiv.2609.05031 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-75] Minimax Lower Bound for Estimating Diffusion-based Local Intrinsic Dimension
链接: https://arxiv.org/abs/2609.04822
作者: Jaehee Seo,Wontae Jeong,Jisu Kim
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注: 29 pages, 1 figure
Abstract:While diffusion-based methods have recently emerged as effective tools for probing the intrinsic geometry of high-dimensional data, their statistical difficulty remains largely unexplored. We study estimation of the finite-scale population functional underlying FLIPD (Kamkari et al., 2024; arXiv:2406.03537), a diffusion-based local intrinsic dimension (LID) quantity defined through the logarithmic scale derivative of a Gaussian-smoothed density. Intuitively, Gaussian smoothing turns local dimension into a scale law: near a d -dimensional manifold, the kernel mass grows like \sigma^d , so differentiating with respect to the noise scale reveals the intrinsic exponent. Under a regular manifold model, we show uniformly over the model class that the finite-scale field differs from the manifold dimension d by at most O(\sigma^2) . We then establish a minimax lower bound of order (n\sigma^d)^-1 for estimating this finite-scale field from n observations, for n^-1/(2\alpha+d)\lesssim\sigma\le\sigma_0 . At the smallest scale covered by our lower-bound construction, the bound becomes the nonparametric rate n^-2\alpha/(2\alpha+d) .
[LG-76] A Differentiable Neural Surrogate for Photon Propagation in Neutrino Telescopes NEURIPS2026
链接: https://arxiv.org/abs/2609.04695
作者: Felix J. Yu,Berthy T. Feng,Nicholas Kamp,Carlos A. Argüelles
类目: High Energy Astrophysical Phenomena (astro-ph.HE); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex)
*备注: 9 pages, 5 figures. Submitted to the Sim2Sci Workshop @ NeurIPS 2026
Abstract:Large-volume neutrino telescopes infer neutrino properties from Cherenkov light, but simulating the transport of billions of photons through highly scattering ice or water is computationally costly. We introduce candela, a differentiable SIREN neural field that learns the photon Green’s function of the IceCube Neutrino Observatory, a cubic-kilometer detector embedded in Antarctic glacial ice. Given a point-like energy deposit and sensor, it predicts the expected photon yield and full arrival-time distribution at the sensor. Complete events are simulated by decomposing charged-particle energy deposits into point-like sources and superposing their predicted sensor responses. Trained on Monte-Carlo simulations, candela generates events 50 – 100\times faster than existing methods, with cost scaling only weakly with neutrino energy. It keeps median yields within 2% of the MC expectation and timing distributions at the MC statistical floor across six photon-count decades. The model also provides end-to-end gradients with respect to event parameters and opens a path toward optimizing scattering-medium properties, which often dominate systematic uncertainties in neutrino telescopes.
[LG-77] Centered Permutation Prefixes for SGD with Random Reshuffling: Sharp Rates Hölder Geometry and Composite Proximal Extensions
链接: https://arxiv.org/abs/2609.04578
作者: Jiaxiang Li
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:We study stochastic gradient descent with random reshuffling for finite sums [ F(x)=\frac1n\sum_i=1^n f_i(x). ] For fresh reshuffling with a constant component stepsize, if each f_i has an L -Lipschitz gradient and the average F is \mu -strongly convex with a Lipschitz-continuous Hessian, we prove the last-epoch rate [ \mathbb E[F(y_K)-F(x_\star)] =\widetilde O!\left(T^-2+n^2T^-3\right), \qquad T=nK, ] matching the known quadratic lower bound in its (n,K) -dependence. The components may be nonconvex, and no componentwise Hessian continuity or separate bounded-iterate assumption is required. More generally, a \nu -Hölder-continuous average Hessian adds only \widetilde O(n^1+\nuT^-2-2\nu) , so every \nu\ge 1/2 preserves the quadratic rate. Under convex components, a decreasing-stepsize result removes the large-epoch requirement and recovers the same two-term scale once nK exceeds the condition-number scale. We also analyze epoch-wise ProxRR for \mathcal P=F+\psi . Writing x^\dagger for the composite minimizer and \beta_\star=|\nabla F(x^\dagger)| , we prove [ \mathbb E|y_K-x^\dagger|^2 =\widetilde O!\left( \frac\beta_\star^2K^2 +T^-2+n^2T^-3 +n^1+\nuT^-2-2\nu \right). ] For \nu\ge 1/2 , we show that the \beta_\star^2/K^2 splitting term is unavoidable and obtain a matching lower bound up to logarithms in the stated constant-stepsize regime. Subjects: Optimization and Control (math.OC); Machine Learning (cs.LG) Cite as: arXiv:2609.04578 [math.OC] (or arXiv:2609.04578v1 [math.OC] for this version) https://doi.org/10.48550/arXiv.2609.04578 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-78] A Sim-to-Real Study of Surface-Code Decoder Benchmarking
链接: https://arxiv.org/abs/2609.04557
作者: Shay J. Manor,Leila S. Erhili,Yassine Jebbouri
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG)
*备注: 11 pages, 8 figures
Abstract:Quantum error-correction decoders are typically benchmarked against synthetic circuit-level noise, under the assumption that a decoder’s ranking under such noise transfers to hardware and improves as the noise model becomes more realistic. The Willow processor, the first to operate below the surface-code threshold, allows us to test this assumption. We rank a panel of six decoders using a four-rung ladder of noise models with increasing fidelity, evaluated against real data across three code distances, two bases, and fifteen round counts. Rank agreement with hardware appears once the noise model gives each operation type its own error rate. Calibrating the model to the device improves absolute error rates but not rank agreement. We additionally provide the first independent evaluation of NVIDIA’s Ising pre-decoder on hardware, at code distances below its training receptive field and via a mapping onto the lattice on which it was trained. Under these conditions, it holds no accuracy-latency advantage: another panel decoder matches or improves on it in both per-cycle error rate and decode latency in 278 of the 280 evaluations. We release the full pipeline and the per-shot outcome of every evaluation, so future decoders and devices can be compared.
[LG-79] Recovering molecules from coarse-grained beads: free-energy-conditioned generative backmapping across chemical space
链接: https://arxiv.org/abs/2609.04432
作者: Luis Itza Vazquez-Salazar,Tristan Bereau
类目: Chemical Physics (physics.chem-ph); Soft Condensed Matter (cond-mat.soft); Machine Learning (cs.LG); Biological Physics (physics.bio-ph)
*备注:
Abstract:Transferable coarse-grained (CG) force fields compress chemical space: by aggregating atoms into a reduced set of interaction beads, models such as MARTINI reduce the number of distinguishable compounds by roughly three orders of magnitude, making high-throughput screening of thermodynamic properties tractable across soft matter, with drug–membrane permeability as a well-developed example. The compression is lossy and, so far, one-way: a screen returns a combination of beads, with no established route back to the compounds it stands for. Recovering those compounds–compositional backmapping–is a one-to-many inverse map, distinct from the better-studied conformational problem of rebuilding atomic coordinates from a known mapping. Here we formulate compositional backmapping as conditional graph generation by introducing juniper, a discrete denoising diffusion model over molecular graphs conditioned on the octanol–water partition free energy \Delta G_\mathrmW \mapsto \mathrmO , the principal driver of MARTINI bead type assignment and hence a proxy for bead identity. Trained on molecules of up to 9 heavy atoms mapped onto one or two beads, juniper generates molecules that are 93% valid and 92% unique for two-bead targets, and whose \Delta G_\mathrmW \mapsto \mathrmO distributions track the target \Delta G^\mathrmCG_\mathrmW \mapsto \mathrmO linearly ( r^2 \geq 0.96 ), departing only in the hydrophobic and hydrophilic tails. Although the model receives no chemical information beyond a single scalar, the functional groups shift systematically with the imposed free energy, from branched hydrocarbons at the apolar end to amides, imides, and isocyanates at the polar end. A bead combination flagged by a CG screen can therefore be turned into candidate molecules for atomistic study or synthesis.
[LG-80] NFlow: Amortized Posterior Inference for Trans-Neptunian Object Surface Composition
链接: https://arxiv.org/abs/2609.04305
作者: Agastya Gaur(1 and 2),Cristina M. Dalle Ore(3),Alessandra Ricca(4 and 4) ((1) University of Illinois Urbana-Champaign, (2) SETI Institute, (3) Carl Sagan Center, SETI Institute, (4) NASA Ames Research Center)
类目: Instrumentation and Methods for Astrophysics (astro-ph.IM); Earth and Planetary Astrophysics (astro-ph.EP); Machine Learning (cs.LG)
*备注: 11 pages, 3 figures, 3 tables
Abstract:We present TNFlow, a transformer and normalizing flow architecture for inferring the surface composition of Trans-Neptunian Objects (TNOs) from their reflectance spectra. TNFlow is trained on synthetic spectra generated by the Shkuratov radiative transfer model to act as its inverse. TNFlow takes \sim 0.7s to invert one spectrum on a single CPU core, returning a multimodal posterior over simplex-valid compositions and grain sizes. On synthetic spectra, the highest-weight mode achieves a mean total-variation distance of 0.149 from ground truth on the test split, and the model generalizes well to unseen combinations of known components. Qualitative tests on real JWST spectra show blindness or bias towards some materials. We suggest this could be attributed to either simulator fidelity or the training set.
[LG-81] Self-Supervised Pretraining of Molecular Graph Encoders with LeJEPA
链接: https://arxiv.org/abs/2609.04261
作者: Michał Kulczykowski,Rafał Łabędzki
类目: Quantitative Methods (q-bio.QM); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Self-supervised pretraining has transformed language and vision, but its value for molecular graph neural networks remains contested. We ask whether pretraining on a large unlabelled corpus improves molecular property prediction. We adapt LeJEPA, a predictor-free joint-embedding predictive architecture regularised by Sketched Isotropic Gaussian Regularisation (SIGReg), to molecular graphs, evaluating GPS and Chemprop-style D-MPNN encoders on the Wong et al. [1] antibiotic-activity dataset and ogbg-molhiv using a multi-seed, bootstrap-based protocol. Pretraining improves learned representations but does not robustly improve finetuning. A frozen probe on pretrained embeddings exceeds random initialisation on both tasks (ogbg-molhiv ROC-AUC 0.788 vs 0.665; +0.123), reaching the published self-supervised band, but this does not translate into finetuning gains. On the antibiotic scaffold split, a canonical partition is significant (delta AUPRC +0.041, p = 0.010), but the effect vanishes across five partitions (pooled +0.013, p = 0.095). Finetuning is null on the random split, ogbg-molhiv, and D-MPNN. The representational edge is nevertheless recoverable. Embeddings saturate at ~16-32 effective dimensions, whereas Morgan fingerprints improve to 1024 bits. At matched dimensionality, fingerprints lead validation (0.799 vs 0.782 at 128 dimensions) but trail shifted test scaffolds (0.759 vs 0.788). Truncating embeddings and combining them with a 1024-bit Morgan fingerprint raises ogbg-molhiv ROC-AUC from 0.805 to 0.832 (delta +0.027; 95% CI [+0.003, +0.054]; p = 0.014); an untrained encoder gains nothing (delta -0.003). Thus, pretraining supplies complementary information best realised through feature-level combination, while finetuning gains are weak and partition-dependent.
附件下载
