本篇博文主要内容为 2026-08-06 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 (2026-08-06)
今日共更新615篇论文,其中:
- 自然语言处理共107篇(Computation and Language (cs.CL))
- 人工智能共178篇(Artificial Intelligence (cs.AI))
- 计算机视觉共131篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共175篇(Machine Learning (cs.LG))
- 多智能体系统共13篇(Multiagent Systems (cs.MA))
- 信息检索共13篇(Information Retrieval (cs.IR))
- 人机交互共27篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] CoPlan: A Trustworthy Co-Intelligence Interface for Care Planning through Role-Based Contestable Argument Graphs
【速读】：该论文旨在解决当前人工智能（AI）支持的护理计划系统中推荐结果固化、缺乏可解释性与可争议性的问题，导致临床决策者、患者及照护团队难以根据临床判断、患者价值观或现实可行性对计划进行审查、质疑和修改。其解决方案的关键在于提出CoPlan——一种协同智能且可争议的人机共治护理规划界面。该系统采用多智能体工作流架构，由专业化AI智能体生成候选干预措施及其支持或反驳论据，人类护理规划者则可对论据进行接受、拒绝、修改或补充，最终形成可解释、可迭代的护理计划。通过融合“协同智能”（即人机互补专业能力）与“可争议性”（即建议持续开放审查、修订与论证），CoPlan实现了对护理计划过程的透明化与可控化，保障了人类在决策中的主导地位与临床责任。研究在“居家养老”场景中验证了该系统在动态照护团队组建、基于角色的论据评审、计划生成及后续日程安排等方面的实用性，为可信的人机协作护理规划提供了可复用的设计范式。
链接: https://arxiv.org/abs/2608.05107
作者: Hung Truong Thanh Nguyen,Hélène Fournier,Piper Jackson,Makoto Itoh,Shannon Freeman,Rene Richard,Hung Cao
机构: 未知
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: Accepted at the 2026 International Conference on Next Generation AI Systems (NGEN-AI 2026)
Abstract:AI-supported care planning can help clinicians, patients, caregivers, and care teams coordinate complex decisions across clinical, functional, psychosocial, and environmental needs. However, many AI systems present recommendations as fixed outputs, limiting stakeholders’ ability to inspect, challenge, and revise plans when they conflict with clinical judgment, patient values, or real-world feasibility. We present CoPlan - a Co-Intelligent and Contestable Interface for Human-AI Care Planning. CoPlan uses a multi-agent workflow in which specialized AI agents generate candidate interventions and supporting or challenging arguments, while human care planners can accept, reject, modify, or add arguments before final plan generation. Through this design, CoPlan combines co-intelligence, in which humans and AI agents contribute complementary expertise, with contestability, where recommendations remain open to inspection, revision, and justification. We demonstrate CoPlan in an aging-in-place care planning scenario. The system supports adaptive care team recruitment, role-based argument review, final care plan generation, and practical follow-up through scheduling agents. This work contributes a contestable care planning interface and a design framing for trustworthy human-AI care planning that preserves human agency and clinical accountability.
[MA-1] OneDayAgent : Towards a Long-Horizon Harness for Autonomous Agents
【速读】：该论文旨在解决大语言模型（LLM）代理在处理长时程、跨环境、多模态的开放式日常任务时，因目标漂移、状态丢失和上下文溢出等多重失败模式导致的执行失效问题。现有研究虽针对单一问题有所探索，但缺乏对多种故障模式协同管理的能力，且在不同后端模型间泛化性不足。本文提出OneDayAgent，一种面向长时程任务的统一执行框架，其核心解决方案在于：将开放式请求分解为边界明确的子任务，通过受控记忆机制在上下文压力下维持执行状态，并在任务末尾对输出结果进行验证与修复。实验表明，OneDayAgent在AgentIF-OneDay基准上覆盖104项任务，使用GLM-5.2后端实现0.821的综合得分，达到当前最优水平；同时，该框架无需调优即可在来自三个模型家族的五种不同后端模型上稳定运行，展现出良好的跨后端泛化能力，验证了其在异构模型环境下保持一致高效执行的潜力。
链接: https://arxiv.org/abs/2608.05013
作者: Jingsheng Zheng,Xinyuan Fang,Jintian Zhang,Zhengke Gui,Huajun Chen,Ningyu Zhang
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: Ongoing work
Abstract:LLM agents are increasingly applied to open-ended everyday requests that span work, study, and life. These tasks are long-horizon, cross-environment, and multimodal, forcing the agent to preserve goals and constraints across many steps while navigating heterogeneous tools and attachments. While prior work has addressed individual failure modes such as goals drift, states loss, and context overflow, whether a single harness can manage them jointly and remain effective across backends has received less study. We present OneDayAgent, a long-horizon harness for autonomous agents. OneDayAgent turns an open-ended request into a managed execution process that decomposes tasks into bounded subtasks, maintains execution memory under context pressure, and verifies and repairs the final deliverable. We evaluate OneDayAgent on AgentIF-OneDay across 104 tasks. With the GLM-5.2 backend, OneDayAgent sets a new state of the art with an overall score of 0.821. The same harness runs across five backend LLMs from three model families, indicating the harness generalizes across backends without tuning, even as different models induce distinct execution styles under the same workflow.
[MA-2] HELENA:Hierarchical Sparse Coordination over a Union of Complementary Topologies for MAS
【速读】：该论文旨在解决基于大语言模型（LLM）的多智能体系统（MAS）在推理过程中因固定拓扑结构导致的路径局限性问题，以及多拓扑融合时引入冗余噪声传播所造成的性能下降问题。其核心挑战在于如何在扩展推理轨迹以提升分析全面性的同时，避免因信息冗余带来的干扰。解决方案的关键在于提出一种名为HELENA的分层稀疏协调框架：首先通过蒙特卡洛树搜索（MCTS）与确定性点过程（DPP）选取互补的候选拓扑结构，构建联合多智能体图以拓宽推理路径；随后引入分层稀疏协调模块，在每一步仅激活稀疏子图，并通过压缩的潜在简报进行智能体间通信，有效抑制噪声传播；最后通过局部自精炼阶段识别存在矛盾证据的决策单元，并在对比证据同时支持解侧失败与竞争侧改进时才进行重写。该方法在8个基准测试中均达到当前最优性能，平均优于最强基线3.47%，在更难的MMLU-Pro上提升达10.34%，且在复杂任务上表现更优，计算开销可控。
链接: https://arxiv.org/abs/2608.04634
作者: Zhifang Mao,Linyao Zheng,Xuhang Shi,Xiuquan Hou
机构: University of Science and Technology of China (中国科学技术大学); Tsinghua University (清华大学)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:LLM-based multi-agent systems (MAS) typically optimize a single topology, restricting reasoning to a narrow trajectory and limiting comprehensive analytical capacity. Naively merging multiple topologies into a composite graph introduces redundant noise propagation across irrelevant connections, degrading solution quality. To address this dilemma, we propose \textbfHierarchical Sparse Coordination over a Union of Complementary Topologies for MAS (HELENA), a multi-agent framework that balances diverse reasoning paths with sparse task-dependent execution. \helena constructs a union MAS graph from complementary candidate topologies selected via Monte Carlo Tree Search and Determinantal Point Process, broadening the reasoning trajectory for comprehensive analysis of complex problems. A Hierarchical Sparse Coordination module then activates only a sparse subgraph at each step while agents exchange compressed latent briefs to suppress redundant noise propagation. Finally, a Local Self-Refinement stage identifies decision units with discrepancy evidence and rewrites them only when contrastive evidence simultaneously confirms a reliable solution-side failure and a challenger-side improvement. Experiments across eight benchmarks show that \helena achieves state-of-the-art results on all benchmarks, with an average gain of \pctup3.47 over the strongest baseline and up to \pctup10.34 on MMLU-Pro, achieving larger improvements on harder benchmarks at a reasonable additional cost.
[MA-3] Emergence of Reputation-Based Cooperation in LLM Agents
【速读】：该论文旨在解决大规模语言模型（Large Language Model, LLM）代理之间合作机制在面对搭便车者（free-rider）入侵时是否具有进化稳定性的问题。其核心问题是：在缺乏强制性约束的环境中，基于自然语言提示演化的合作策略能否有效抵御自私行为的侵蚀。解决方案的关键在于通过模拟间接互惠捐赠博弈，研究不同LLM代理在连续捐赠尺度下的行为演化，并以文化传承方式让策略通过代际传播优化。研究发现，对抗搭便车者的鲁棒性差异超过一个数量级，而最强预测因子是“对手资财敏感性”（opponent endowment sensitivity），即代理对合作与不合作对象的区分能力，这实质上对应经典图像评分（Image Scoring）机制。相比之下，“领先八原则”（Leading-Eight L1 norm）的遵守程度无法预测鲁棒性。进一步分析表明，鲁棒性取决于对背叛者（defector）的排斥强度——尽管各模型在奖励合作者和惩罚背叛者方面存在差异，但唯有对背叛者的严格排除才显著提升对搭便车者的抵抗能力。研究揭示，当前LLM代理仅能演化出类似图像评分的有限区分策略，难以发展出更稳健的“领先八原则”等规范，暴露出其文化演化合作机制的根本脆弱性，因而亟需自下而上的规范建构方法以增强协作系统的稳定性。
链接: https://arxiv.org/abs/2608.04507
作者: Kazuya Horibe,Kenji Itao,Wataru Toyokawa
机构: RIKEN(理化学研究所); Tohoku University (东北大学)
类目: Multiagent Systems (cs.MA); Neural and Evolutionary Computing (cs.NE)
备注:
Abstract:Can cooperation among large language model (LLM) agents be evolutionarily stable against free-rider invasion? We study an indirect reciprocity donation game where LLM agents observe behavioral traces and donate on a continuous scale. Strategies, represented as natural language prompts, evolve through cultural transmission across generations. Across four LLM backends, robustness to free-rider invasion varies by more than an order of magnitude. The strongest predictor of this robustness is opponent endowment sensitivity, the degree to which agents discriminate between cooperative and uncooperative opponents, operationalizing the classical Image Scoring mechanism. By contrast, adherence to the Leading-Eight L1 norm does not predict robustness. Robustness depends on defector exclusion: while both cooperator reward and defector punishment vary across models, only the stringency of defector exclusion predicts resistance to free-rider invasion. These findings reveal that LLM agents are confined to Image Scoring-like discrimination and fail to develop the more robust Leading-Eight norms, highlighting a fundamental vulnerability in culturally evolved LLM cooperation and motivating bottom-up approaches to norm construction.
[MA-4] MCHA: A Memory-Centric Hierarchical Architecture for Parallel-Sequential Computing MICRO’26
【速读】：该论文旨在解决新兴工作负载（如多智能体强化学习、大规模类脑计算及概率图模型）固有的并行-顺序计算模式所引发的性能瓶颈问题。这类任务虽需大规模并行以实现高吞吐，却因数据访问模式高度不规则且集中于主存，导致传统架构面临全局缓冲区饱和与内存带宽受限的根本性挑战。其核心解决方案是提出一种面向并行-顺序执行的内存中心分层架构（Memory-Centric Hierarchical Architecture, MCHA），关键在于通过分层通信机制实现核间分布式数据路由，显著降低对全局内存的带宽压力；同时配套设计了一种基于事件驱动条件触发的新型并行-顺序编程模型，有效在执行流水线中隐藏数据传输延迟。实验表明，MCHA在多智能体强化学习等典型任务上相较NVIDIA A100 GPU实现高达153.06×至2456.96×的性能加速，主存访问占比由96%降至5.44%，并在28 nm工艺下实现2.92 mm²面积与115.36 mW功耗的高效硬件实现，兼具卓越性能与编程灵活性。
链接: https://arxiv.org/abs/2608.04443
作者: Daijing Shi,Hongxiao Zhao,Yihan Fu,Zhan Chen,Jiayi Li,Yihang Zhu,Anjunyi Fan,Yaoyu Tao,Yuchao Yang,Bonan Yan
机构: 未知
类目: Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC); Multiagent Systems (cs.MA)
备注: Accepted in MICRO’26
Abstract:Emerging workloads, such as Multi-Agent Reinforcement Learning (MARL), large-scale neuromorphic computing, and probabilistic graphical models, intrinsically exhibit parallel-sequential computing patterns. While these tasks demand massive parallelism to achieve high throughput, they are severely bottlenecked by irregular data access patterns centralized to main memory. Consequently, conventional architectures face fundamental limitations when executing these workloads, primarily manifesting as global buffer saturation and memory-bound bottlenecks. To address these challenges, we propose the Memory-Centric Hierarchical Architecture (MCHA), a reconfigurable hardware solution tailored for parallel-sequential execution. MCHA leverages a hierarchical communication strategy that facilitates distributed, inter-core data routing, thereby significantly reducing the bandwidth burden on the global memory. Complementing the hardware, MCHA introduces a novel parallel-sequential programming model that utilizes event-driven conditional triggers to effectively hide data transmission latency within the execution pipeline. We benchmark MCHA against a diverse suite of parallel-sequential tasks, including MARL, motor variable control, and Markov random fields. Validated through our open-source, cycle-accurate simulator, MCHA demonstrates performance speedups ranging from 153.06 \times to 2456.96 \times over NVIDIA A100 GPUs on MARL workloads, while maintaining robust programming flexibility across other application domains. Furthermore, the architecture successfully reduces main memory access from 96% to 5.44%. When synthesized in a 28 nm process, the MCHA implementation occupies an area footprint of 2.92mm ^2 and consumes 115.36 mW of power at 200 MHz. MCHA is open-sourced at this https URL.
[MA-5] Continuous Improvement and Parallel Autonomous Exploration: An LLM -Agent Framework for Searching Large Solution Spaces KDD’26
【速读】：该论文旨在解决大尺度解空间中自动化搜索最优解决方案的难题，尤其针对复杂、高维的检索任务（如电商场景下的商品-目录匹配），传统方法往往受限于单一初始假设的局部优化。其核心挑战在于如何在不依赖人工干预的前提下，实现对多样化解的高效探索与持续迭代优化。解决方案的关键在于提出一个双机制框架：一是基于保留数据集评分的排行榜作为奖励信号，驱动单个或多个智能体通过反复提交与自我评估实现解的渐进式改进；二是支持多智能体并行自主运行，各智能体独立完成分析、方法调研、实现、自评、提交与修正全过程，仅由一个协调代理负责调度管理，从而实现无监督的并行探索。该框架通过共享奖励机制扩大了解空间的覆盖范围，突破了单一种子范式局限。实验表明，在商品-目录匹配任务中，单智能体最高覆盖率仅为47.8–57.4%，而五智能体并行运行时提升至62.8–69.4%，显著优于33.3%的基线水平。因此，该研究的核心贡献在于构建了一个具备持续改进能力的奖励循环机制和可扩展的全自主并行探索架构，为复杂任务中的智能体协同优化提供了可验证的系统性范式。
链接: https://arxiv.org/abs/2608.04341
作者: Dulmini Hettiarachchi,Andre Rusli,Julio Christian Young,Sho Akiyama
机构: Mercari, Inc.(日本); Mercari, Inc.(日本); Mercari, Inc.(日本); Mercari, Inc.(日本)
类目: Multiagent Systems (cs.MA)
备注: 5 pages, ACM KDD’26 Workshop on SciSoc Agents LLMs
Abstract:We present a framework that gives LLM agents two mechanisms for searching large solution spaces autonomously. First, a leaderboard scored on held-out data acts as a reward signal that drives each agent to refine its solutions over repeated submissions, a loop that operates even with a single agent. Second, the framework enables running many agents in parallel, fully autonomously, with no human in the loop: agents independently analyze, survey methods, implement, self-evaluate, submit, and revise, while a moderator agent handles only logistics. Running agents in parallel under the shared reward broadens the explored region of the solution space rather than refining the single seeded paradigm. We instantiate the framework on product-to-catalog matching (a core e-commerce retrieval task with a large, category-structured solution space), posed as selective prediction with a precision-coverage operating point. A single agent refines within its seeded paradigm, whereas parallel autonomous agents surface qualitatively different solutions. On this testbed, best qualified coverage (=95% P@1 per category) reaches 47.8-57.4% with a single agent and 62.8-69.4% with five, against a 33.3% baseline. Our contribution is the framework itself: a continuous-improvement reward loop and a substrate for fully autonomous parallel exploration, backed by case-study evidence.
[MA-6] Responsibility in Multi-Agent Sequential Decision-Making: Comparing Human Judgments to Formal Models of Causal Attribution
【速读】：该论文旨在解决在高风险决策场景中，当生成式AI系统出现失败时，如何准确识别因果关系并界定责任归属的问题。随着人工智能在关键领域应用的不断深入，责任归因（responsibility attribution）的透明性与合理性已成为亟待解决的核心挑战。论文基于实际因果性（actual causality）框架，系统评估了多种责任归因方法与人类判断的一致性，并通过大规模问卷调查，在改进版的围棋类游戏Goofspiel的多智能体序列决策情境中收集人类对责任分配的主观判断。研究发现，尽管目前尚无单一归因方法能始终与人类判断保持一致，但揭示了影响人类责任判断的关键因素，包括智能体特异性偏差（agent-specific biases）以及决策过程中信息可得性（amount of information available）。这些发现为构建更符合人类认知逻辑的责任归因机制提供了重要依据。
链接: https://arxiv.org/abs/2608.04318
作者: Nripsuta Ani Saxena,Stelios Triantafyllou,Goran Radanović
机构: University of Southern California (南加州大学); Max Planck Institute for Software Systems (马普所软件系统研究所)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:With the growing adoption of artificial intelligence in high-stakes decision-making, identifying the causes of outcomes–particularly failures–and determining who is responsible has become a critical concern. In this work, we examine how well formal definitions of \textitresponsibility attribution, grounded in the framework of \textitactual causality, align with human judgments of responsibility. To this end, we conduct a large-scale survey to elicit human judgments of responsibility in multi-agent sequential decision-making scenarios, using a modified version of the card game Goofspiel. We evaluate multiple responsibility attribution methods, assess their alignment with human judgments about responsibility, and identify factors that significantly shape responsibility judgments. While no single responsibility attribution method consistently aligns with human responses, our findings highlight key factors that influence human responsibility judgments, including agent-specific biases and amount of information available to agents during decision-making.
[MA-7] rident : How to Break Deep Reinforcement Learning Cyber Defenses (Agent ic)
【速读】：该论文旨在解决当前基于深度强化学习（Deep Reinforcement Learning, DRL）的自主网络防御系统在评估中普遍依赖静态、启发式红队代理（red agent）的问题，导致其对自适应威胁的鲁棒性研究严重不足。现有评估体系缺乏能够模拟动态、智能对抗行为的基准环境与交互数据，制约了生成式AI在网络安全领域的有效应用。为此，论文提出Trident框架，其核心解决方案在于构建一个三位一体的“代码即策略”（Code-as-Policy）强化学习与可验证奖励（Reinforcement Learning with Verifiable Rewards, RLVR）智能体架构，通过三阶段设计——日志摘要器（Log Summarizer）、规划器（Planner）与编码器（Coder）——将攻击策略生成重构为上下文相关的多臂赌博机问题。其中，可训练的7B规模规划器从压缩的执行日志中生成完整攻击策略，并由冻结的编码器将其转化为可执行的Python策略，部署于真实DRL防御者环境中。该框架配套提供了覆盖CybORG CAGE 4与CyberWheel的动态沙箱基准及超过13,000条高保真红蓝交互轨迹数据集，显著提升了评估的真实性与挑战性。实验表明，Trident在无需额外训练的情况下，使现有防御系统的性能平均下降522%，并自主发现诸如诱饵规避与自适应状态优先级等复杂行为，揭示了现有防御机制的根本脆弱性，验证了其作为新一代智能对抗评估平台的有效性。
链接: https://arxiv.org/abs/2608.04317
作者: Ryozo Masukawa,Ian Bryant,Armita Kazeminajafabadi,Sanggeon Yun,Hyunwoo Oh,SungHeon Jeong,Nathaniel D. Bastian,Mahdi Imani,Mohsen Imani
机构: University of California, Irvine (加州大学欧文分校); Northeastern University (东北大学); Johns Hopkins University (约翰霍普金斯大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: code: this https URL
Abstract:Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost exclusively against static, heuristic red agents, leaving their robustness against adaptive threats critically understudied. Meanwhile, recent advances in Reinforcement Learning with Verifiable Rewards (RLVR) have improved LLM reasoning, but their integration into cybersecurity remains elusive due to the absence of suitable benchmark environments and interaction datasets. To bridge this gap, we introduce Trident, an agentic LLM red teaming framework comprising three components: a dynamic benchmark with isolated sandbox servers spanning CybORG CAGE 4 and CyberWheel, a dataset comprises over 13,000 high-fidelity red-blue interaction trajectories for RLVR, and a ``Code-as-Policy’’ RLVR agentic architecture Trident Agentic). The latter reformulates red agent training as a contextual bandit via a tripartite Log Summarizer–Planner–Coder design, where a trainable Planner generates complete attack strategies from compressed execution logs, which a frozen Coder translates into executable Python policies deployed against live DRL defenders. Empirical evaluations reveal a fundamental brittleness in existing defenses: with a single trainable 7B planner, Trident reduces blue agent defensive performance by an average of 522% compared to static red agent baselines while autonomously discovering emergent behaviors such as decoy avoidance and adaptive state prioritization that static heuristics entirely fail to uncover.
[MA-8] MIDAS: Multi-LLM Iterative Data-Adaptive Summarization ICDAR2026
【速读】：该论文旨在解决企业级文本摘要任务中因领域特定规范、输出格式及组织惯例差异所带来的复杂性问题，尤其针对支持工单、法律文件、事件报告等场景下人工编写和维护高质量提示（prompt）所面临的高成本与低适应性挑战。现有自动化提示优化方法虽基于大语言模型（LLM）的批判性反馈实现提示迭代改进，但受限于静态提示结构，难以应对多样的摘要应用场景。其解决方案的关键在于提出一种多大语言模型协同的迭代式数据自适应摘要框架——多模型迭代数据自适应摘要（MIDAS），通过数据驱动的模式学习与使用场景特异性个性化机制，实现无需人工干预的提示自适应能力。MIDAS在五种不同输出格式的企业客户工单摘要任务中表现出卓越性能，显著优于当前主流的批判驱动优化框架（如CriSPO和ZERA），在ROUGE-1、ROUGE-2和ROUGE-L指标上分别提升最高达11.0%、18.2%和8.0%，同时在所有格式与输出类型下均稳定提升BERTScore F1；此外，该框架还验证了跨模型与跨领域的泛化能力，展现出强大的实际应用潜力。
链接: https://arxiv.org/abs/2608.04307
作者: Karen Lee,Dhanashree Balaram,Seojun Shon,Umair Rasheed
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: Accepted at the 20th International Conference on Document Analysis and Recognition (ICDAR 2026). 17 pages, 2 figures
Abstract:Text summarization is deceptively difficult. While condensing information seems straightforward, real-world enterprise summarization of support tickets, legal documents, incident reports, and more, demands strict adherence to domain-specific guidelines, output formats, and organizational conventions. Crafting prompts that reliably satisfy these constraints is labor-intensive, requiring significant human expertise and continuous maintenance as requirements evolve. Existing automated prompt optimization methods reduce this burden through Large Language Model (LLM) critique-driven refinement, yet remain limited by static prompts that cannot adapt to the diversity of summary applications. We propose Multi-LLM Iterative Data-Adaptive Summarization (MIDAS), a multi-LLM framework that extends this paradigm with data-driven pattern learning and use-case-specific personalization, enabling automatic adaptation to different summarization requirements without manual prompt engineering. Applied to enterprise customer ticket summarization across five output formats, MIDAS achieves the strongest overall performance against state-of-the-art critique-driven optimization frameworks such as CriSPO and ZERA, improving ROUGE-1 by up to 11.0%, ROUGE-2 by up to 18.2%, and ROUGE-L by up to 8.0%, while consistently improving BERTScore F1 across all formats and output types. We additionally demonstrate cross-model and cross-domain generalization through multi-LLM configurations and finance-domain summarization benchmarks.
[MA-9] CURATE: Leverag ing LLM Agents to Compose Catalog and Deploy Reproducible Workflows
【速读】：该论文旨在解决当前生成式代码代理（Agentic code generation）在软件开发全流程中覆盖不完整的问题，特别是缺乏对工作流全生命周期（包括部署与共享）的系统性支持。现有方法主要聚焦于代码生成，而忽视了模块化开发、集成、部署及复用等关键环节，导致用户需自行管理模块间的耦合与部署流程，增加了使用门槛。为此，本文提出一种名为CURATE（Composition, User-in-the-loop, Reuse, and Automated Task Execution）的人机协同多智能体系统，其核心创新在于构建一个可扩展的模块目录（module catalog），实现工作流组件的存储、复用与组合，并通过人类在环（User-in-the-loop）机制保障开发过程中的可控性与准确性。该系统以大语言模型（LLM）为驱动，支持从代码生成、测试、部署到共享的全流程自动化，显著提升了科学与应用工作流的可复用性与可扩展性，为遵循可查找（Findable）、可访问（Accessible）、可互操作（Interoperable）和可重用（Reusable）（FAIR）原则奠定了基础。实验验证了系统在重构与适配4个来自SeBS-Flow基准套件的工作流，以及自动开发并扩展一个复杂环境工程中的厌氧消化机制模型工作流方面的可行性。
链接: https://arxiv.org/abs/2608.04270
作者: Nolan Cutler,Chia-Chen Kuo,Nanda Velugoti,Kathryn Newhart,Renato Figueiredo
机构: 未知
类目: oftware Engineering (cs.SE); Multiagent Systems (cs.MA)
备注: 8 pages, 3 figures
Abstract:Agentic code generation has shown promise in automating and accelerating software development by utilizing Large Language Models (LLMs) to generate, test, and deploy code. For engineers and scientists, such systems have the potential to accelerate the development of applied and scientific workflows while reducing barriers to entry in domains that have yet to fully realize their benefits. However, a key gap remains: existing coding agents primarily focus on code generation and do not address the entire workflow lifecycle, including deployment and sharing. As a result, users develop and stitch modules independently while managing deployment on their own. To address this gap, we propose CURATE - Composition, User-in-the-loop, Reuse, and Automated Task Execution - a novel human-in-the-loop multi-agent system that uses LLM agents to manage and develop composable workflows across their entire lifecycle. A key feature of the system is a catalog that allows for the storage and reuse of modules across workflows. Module catalogs provide a foundation that can be expanded to support FAIR principles by facilitating the sharing and reuse of curated modules and subgraphs. We demonstrate the feasibility of our system with an initial prototype using Claude Opus 4.8, comprising 6 experiments: reproducing and adapting 4 workflows derived from the SeBS-Flow benchmark suite, and automating the development and scaling of a workflow that leverages a complex mechanistic model in environmental engineering used to simulate anaerobic digestion.
[MA-10] Strategic Evaluation of Planning Strategies for LLM Agents in Cyber-Physical Systems
【速读】：该论文旨在解决大语言模型（Large Language Model, LLM）在战略网络物理系统（strategic cyber-physical systems）中进行规划时，传统评估方法仅关注任务成功或计划执行一致性所导致的评估偏差问题。其核心挑战在于：当自主参与方响应及物理约束共同作用后，原有规划架构是否仍具备有效性与适应性。为此，研究提出一个受控且基于物理规律的基准测试框架，围绕“规划诱导控制轨迹”构建，即通过一系列有序、分层、可搜索的规划操作与指令，明确执行架构对其他智能体和物理过程的作用机制。该框架在包含40个异构产消者（prosumer）的智能电网需求响应系统中实现预设的顺序、分层及搜索型执行器，并采用独立模拟的辐射状馈线系统以保证物理真实性。在此设定下，LLM仅限于类型化策略声明与简短操作消息，而调度构建、产消者动态与功率流等关键过程仍由显式代码定义。评估协议引入成对强制模式反事实分析、共用随机响应抽样及事件级截止时间可行性判定，揭示三大关键发现：（1）规划架构显著影响结果，强制搜索在所有五组基线种子中表现最优；（2）执行保真度不仅依赖模式一致，客观替代性虽保持1.0的一致率，但电压不足量增加2.68倍；（3）在144场景、576期的实验银行中，四个架构中有三个产生可行最优解。进一步分析显示，在预设应力保留测试集上，平均后悔值为90.7（95%置信区间[73.8, 108.6]），未优于固定顺序策略；然而，若在质量预测前应用已知的截止时间可行性约束，则后悔值降至29.0，性能提升61.1%。全可行性的消融实验表明，其仍无法超越固定搜索策略，将剩余挑战聚焦于“可行解内部的质量选择”。此外，五模型扩展分离出应力条件依赖、状态盲设与不变声明者，延迟尾部分析提示实时可行性应被建模为概率性问题。因此，该研究的关键解决方案在于构建一个融合物理仿真、结构化执行路径与严格反事实评估的基准体系，从而推动对规划架构鲁棒性与适应性的深层次验证。
链接: https://arxiv.org/abs/2608.04265
作者: J. de Curtò,I. de Zarzà
机构: 未知
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:Evaluations of LLM planning agents largely ask whether a task succeeds or a declared plan is followed. In strategic cyber-physical systems, a stronger question is whether the planning architecture remains appropriate after autonomous participants respond and physics constrains the outcome. We introduce a controlled, physics-grounded benchmark built around planning-induced control trajectories: the ordered planning operations and directives through which an execution architecture acts on other agents and the physical process. It implements predefined, sequential, hierarchical, and search executors in a smart-grid demand-response system with 40 heterogeneous prosumers and an independently simulated radial feeder. The LLM is bounded to typed policy declaration and short operator messages, while schedule construction, prosumer dynamics, and power flow remain explicit code. The protocol uses paired forced-mode counterfactuals, common random response draws, and event-level deadline feasibility. Three properties follow. Architecture materially changes outcomes: forced search is the oracle in all five baseline seeds. Execution fidelity needs more than mode agreement: objective substitution holds agreement at 1.0 while increasing voltage shortfall by 2.68x. A 144-scenario, 576-episode bank has feasible oracles from three of the four architectures. A prespecified stress-held-out ridge has mean regret 90.7 (95% interval [73.8, 108.6]) and no detectable value over fixed sequential; applying known deadline feasibility before quality prediction cuts regret to 29.0 and improves over fixed sequential by 61.1. An all-feasible ablation does not beat fixed search, localising the remaining challenge to within-feasible quality selection. A five-model extension separates stress-conditioned, state-blind, and invariant declarers; latency tails show that live feasibility should be treated probabilistically.
[MA-11] oward Uncertainty Quantification in Modern Art
【速读】：该论文旨在解决生成式视频模型在创作现代艺术动画时，因随机种子差异导致输出结果存在显著多样性但传统不确定性量化（UQ）方法无法有效解析其内在结构的问题。核心挑战在于：现代艺术本身具有意图模糊性，不同种子生成的视频差异并非噪声，而是潜在多重解释的信号，而现有UQ方法仅将多组生成结果简化为单一离散度指标，无法区分如主导解读与异常值共存、双模态竞争或分布扩散等复杂不确定性模式，更无法判断生成结果是否仍忠实于原始艺术意图。本文提出首个针对现代艺术动画生成不确定性的结构化分析框架，其关键在于构建了一套可复用的“源盲”多种子不确定性识别协议，包含七种源盲估计器与六种参考感知估计器、一套涵盖鲁棒跨度、异常值影响、显式拓扑、多模性、各向异性、留一种子影响及参考覆盖度的分布特征剖面、多种分布模型消融实验（vMF、Kent、ACG、Student t、核密度、混合模型），以及八个诊断性问题和作品级统计协议。通过构建首个大规模数据集（250个现代艺术描述，基于Wan2.1 14B模型在四个编码器下各生成四种子，共1000个视频，艺术作品未参与训练），验证表明该协议在种子集拓扑分类上达到0.98平衡准确率（基线0.25），在异常配置识别中实现AUROC 1.00（标量方法仅0.35），并能可靠地从三种子及跨编码器场景中区分出具有参考覆盖（n=97）与缺失参考（n=56）多样性的高不确定性作品，显著超越传统方法对生成不确定性的表征能力。
链接: https://arxiv.org/abs/2608.04038
作者: Tirtho Roy,Ushashi Bhattacharjee,Showrav Kumar Saha,Sayantan Chakraborty,Koushik Howlader,Tanusree Bhattacharjee
机构: Iowa State University (爱荷华州立大学); Independent Researcher (独立研究员); University of Dhaka (达卡大学)
类目: Graphics (cs.GR); Multiagent Systems (cs.MA); Image and Video Processing (eess.IV)
备注:
Abstract:Asked to animate the same modern artwork under different random seeds, a text to video model returns visibly different films, one reading per seed. Because modern art is ambiguous by intent, this disagreement is signal, not noise. Yet prevailing uncertainty quantification (UQ) collapses a set of generations to a dispersion scalar that says how much the seeds differ but not how: it cannot tell a compact interpretation from a dominant reading plus an outlier, two competing modes, or diffuse instability, nor whether the set still contains a rendering faithful to the original. We present the first study of the structure of generative uncertainty for modern art animation, and a reusable protocol for identifying source blind multiseed uncertainty: a suite of seven source blind and six reference aware estimators; a distributional profile (robust spread, outlier influence, explicit topology, multimodality, anisotropy, leave one seed influence, reference coverage); a distribution model ablation (vMF, Kent, ACG, Student t, kernel, mixture); eight identification questions; and an artwork level statistical protocol. We build the first corpus: 250 modern artwork captions rendered by Wan2.1 14B under four seeds (1000 videos) across 4 encoders, artworks withheld from generation. As a diagnostic the protocol succeeds: it classifies seed set topology at balanced accuracy 0.98 (chance 0.25), isolates the outlier configuration at AUROC 1.00 where a scalar reaches only 0.35, and splits high uncertainty artworks into reference covering (n=97) and reference missing (n=56) diversity, reliably from three seeds and across encoders.
[MA-12] EDATracer: An Agent ic Framework for Large-Scale EDA Artifact Analysis
【速读】：该论文旨在解决现代芯片设计中电子设计自动化（EDA）工具生成的大量异构设计产物（如源代码文件、脚本、日志、网表和报告）难以进行跨类型、跨阶段分析的问题。由于相关证据分散在不同类型的产物与设计流程阶段中，导致调试、优化及设计流程理解效率低下。现有基于大语言模型（LLM）的EDA辅助方法缺乏公开的大规模跨产物分析基准，且在推理过程中难以有效锚定于工具生成的实际证据。为此，论文提出EDATracer——一种基于证据锚定的智能体框架，其核心在于将设计产物组织为知识图谱与语义向量索引相结合的结构化知识库，使LLM智能体能够高效检索并利用来自源文件、日志、网表和报告中的多源证据。研究团队构建了一个包含2,787个可综合开源芯片设计的18.9 GB数据集，并引入一个涵盖事实性、统计性和推理类任务的90题基准。实验表明，相较于Cursor和Claude Code，EDATracer在pass@1准确率上分别提升6.4和7.2个百分点，同时消耗的令牌数减少2.0至3.2倍，验证了其在高效性与准确性上的显著优势。
链接: https://arxiv.org/abs/2608.04032
作者: Phat Tieu,Sayanti Jana,Matthew DeLorenzo,Jiawen Wu,Narendran Srinivasan,Srinivas Shakkottai,Jiang Hu,Jeyavijayan Rajendran
机构: 未知
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注:
Abstract:Modern chip design relies on electronic design automation (EDA) tools that generate large, heterogeneous artifacts, including source files, scripts, logs, netlists, and reports. Analyzing these artifacts is critical for debugging, optimization, and design-flow understanding, but remains difficult because relevant evidence is often distributed across many artifact types and design stages. Although LLM agents show promise for EDA assistance, existing approaches lack public benchmarks for large-scale cross-artifact analysis and often struggle to ground reasoning in tool-generated evidence. We present EDATracer, an agentic framework for evidence-grounded EDA artifact analysis. EDATracer organizes design artifacts into a knowledge graph paired with a semantic vector index, enabling LLM agents to retrieve evidence across source files, logs, netlists, and reports. We curate an 18.9 GB dataset of 2,787 synthesizable open-source chip designs and introduce a 90-question benchmark spanning factual, statistical, and reasoning tasks. Across evaluated agents, EDATracer achieves the best pass@1 accuracy, outperforming Cursor and Claude Code by 6.4 and 7.2 percentage points on average, while using 2.0-3.2x fewer tokens.
自然语言处理
[NLP-0] Reasoning Core: Designing Broad Procedural Data for Completion-Supervised Reasoning Training
【速读】： 该论文旨在解决生成式推理问题数据在完成监督微调（completion-supervised fine-tuning）中应用不足的问题，尤其针对现有程序化生成数据集在语义有效性、难度控制和任务评估方面的局限性。其核心解决方案是提出Reasoning Core——一个涵盖数学、逻辑、规划、状态追踪、形式语言、结构化数据、游戏、因果推理及代码等领域的50个程序化生成器集合，具备语义评分器、难度调控机制和任务评估工具。关键创新在于通过紧凑的目标设计与校准的难度控制，确保生成数据在训练中的实际效用，而不仅依赖语义正确性；同时引入结合模型辅助评审、人工仲裁与回归测试的审计流程，揭示了生成、渲染、目标与评分之间潜在的不匹配问题，强调了生成过程完整性的重要性。实验表明，在多个基线模型设置下，Reasoning Core显著优于其他程序化数据集（Procedural Warmup、Reasoning Gym、SynLogic），在DROP、LogiQA、ARC-Challenge等基准上取得最高平均得分，验证了其作为高质量训练数据的有效性。
链接: https://arxiv.org/abs/2608.05148
作者: Damien Sileo,Valentin Lacombe,Dimitri Kachler
机构: Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189 - CRIStAL, F-59000 Lille, France
类目: Computation and Language (cs.CL)
备注: 20 pages, 3 figures. Code: this https URL Data: this https URL
Abstract:Procedural generators produce useful verifiable reasoning problems at scale, but have received less attention as data for completion-supervised fine-tuning. We introduce Reasoning Core, a collection of 50 generators spanning mathematics, logic, planning, state tracking, formal languages, structured data, games, causality, and code, with semantic scorers, difficulty controls, and task evaluators. Under a matched completion-supervised protocol, we compare Reasoning Core with Procedural Warmup, Reasoning Gym, and SynLogic across four base-model settings and multiple training durations. In the primary 3B comparison, Reasoning Core achieves the highest mean scores on DROP, LogiQA, and ARC-Challenge, exceeding both the baseline without procedural data and all three alternative procedural collections. Task-level analyses show that semantic validity alone does not ensure training utility, highlighting compact targets and calibrated difficulty as important design factors. We ran audits combining model-assisted review, human adjudication, and regression testing. Applied throughout Reasoning Core development and to the other collections, they reveal subtle mismatches among generation, rendering, targets, and scoring, a reminder that procedural generation alone does not guarantee correctness. The library, generated datasets, and audit material are publicly available.
[NLP-1] oward Skill-Native LLM s: Skill Entropy for Benchmarking and Training Long-Horizon Reasoning
【速读】： 该论文旨在解决大语言模型（LLM）在长时序推理任务中跨技能切换能力不足的问题，即模型在多步骤推理过程中需动态调用不同推理技能（如数学推导、计划生成等），但现有评估基准多聚焦单一技能，缺乏对模型在不同技能间有效切换能力的系统性衡量。其解决方案的关键在于提出“技能熵（Skill Entropy）”这一量化指标，用于衡量从一种推理技能切换至另一种技能的难度，并基于此构建了首个面向跨技能长时序任务的基准测试体系——Skill^2-Bench，涵盖9个可验证且开放域的任务领域，覆盖558种技能，每个任务根据技能熵被划分为三个难度等级。实验表明，主流模型在高技能熵任务上的准确率显著下降，揭示了“技能切换差距”。为提升模型的跨技能推理能力，研究进一步将技能熵引入训练阶段，提出基于强化学习的技能熵优化框架（Skill-Entropy RL），通过联合预测每一步的答案与所用技能，并设计包含正确性奖励与技能序列对齐奖励的复合奖励函数，引导模型更精准地识别并切换技能。在Qwen3-4B-Instruct和Qwen3-1.7B上，该方法分别将技能基准得分从34.4%和14.6%提升至68.4%和40.1%，显著优于基线方法，且该训练信号可迁移至通用训练数据（如OpenR1-Math），具备良好的可复用性。
链接: https://arxiv.org/abs/2608.05139
作者: Yinghui He,Ling Yang,Jiarui Liu,Yongjin Yang,Lechen Zhang,Yingcheng Wu,Zhenfei Yin,Mengdi Wang,Sanjeev Arora
机构: Princeton University (普林斯顿大学); Carnegie Mellon University (卡内基梅隆大学); University of Toronto (多伦多大学); University of Illinois Urbana-Champaign (伊利诺伊大学香槟分校); Stanford University (斯坦福大学); University of Oxford (牛津大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: this https URL
Abstract:Long-horizon reasoning in recent LLMs demands that the model switch between distinct skills inside a reasoning chain, such as first doing a math derivation, then using the result to plan a schedule. We call such problems cross-skill long-horizon tasks: multi-step tasks whose steps require different reasoning skills and depend on earlier outputs. Existing benchmarks often evaluate individual skills, lacking a principled way to measure how well a model switches between skills. We address this gap from both the evaluation and training sides. We introduce Skill Entropy, a measure of the difficulty of switching from one skill to another. We then propose Skill^2-Bench, a benchmark of cross-skill long-horizon tasks built over 558 skills across 9 verifiable and open-ended domains. Each task is assigned a task-level skill-entropy score and grouped into three difficulty levels. Evaluating 8 frontier and 4 open-source models on Skill^2-Bench reveals a skill-switching gap: accuracy decreases on higher-entropy tasks. We then turn skill entropy from a benchmark scale into a training signal. We propose Skill-Entropy RL, an RL framework where the model predicts not only the answer at each step but also the skill used to produce it. The reward combines step-level correctness with a skill-entropy reward that measures the alignment between the model-predicted skill sequence and the gold skill sequence. On Qwen3-4B-Instruct and Qwen3-1.7B, Skill-Entropy RL improves the Skill^2-Bench score from 34.4% to 68.4% and from 14.6% to 40.1%, respectively, outperforming competitive baselines. The same pipeline can be applied to off-the-shelf training data such as OpenR1-Math, indicating that skill entropy is a reusable training signal. Code available at: this https URL
[NLP-2] Spoken Function Calling: A New Perspective on Spoken Language Understanding for Large Audio Language Models
【速读】： 该论文旨在解决传统闭集语音理解（Spoken Language Understanding, SLU）在开放域任务中难以有效利用上下文学习的问题，其核心挑战在于传统SLU依赖模糊的规则定义，限制了模型在未见场景下的泛化能力。为此，本文提出了一种新的语义理解范式——语音函数调用（Spoken Function Calling, SFC），通过引入结构化的规则定义来优化语义理解过程，从而突破传统闭集SLU的局限性。解决方案的关键在于：构建基于传统SLU数据集的扩展口语函数体系，设计多智能体系统合成SFC-Bench基准数据集，并在此基础上评估大语言模型（LLM）与大音频语言模型（LALM）在SFC任务上的表现，最终通过后训练进一步增强LALM的SFC能力。实验结果表明，SFC显著优于传统SLU，在提升LLM与LALM的语义提取准确率方面具有明显优势。
链接: https://arxiv.org/abs/2608.05126
作者: Yuezhang Peng,Yuxin Liu,Changfeng Gao,Zhifu Gao,Xiangang Li,Xie Chen
机构: Shanghai Jiao Tong University(上海交通大学); Token Foundry, Alibaba Group(阿里巴巴集团-通义实验室); Shanghai Innovation Institute(上海创新研究院)
类目: Computation and Language (cs.CL); Multimedia (cs.MM)
备注: ACM Multimedia 2026
Abstract:Spoken Language Understanding (SLU) is the core component of task-oriented dialogue systems and a pivotal link in achieving seamless human-agent interaction. While traditional SLU can effectively extract user semantics for closed-set tasks after in-domain supervised fine-tuning, it faces significant challenges in leveraging in-context learning for open-domain tasks due to its ambiguous rule definitions. This work proposes Spoken Function Calling (SFC), a novel semantic understanding perspective that optimizes semantic understanding with structured rule definitions, to evolve beyond traditional closed-set SLU. Specifically, we curate and extend a suite of spoken functions based on traditional SLU datasets, construct a multi-agent system to synthesize the SFC-Bench dataset, evaluate the performance of Large Language Models (LLMs) and Large Audio Language Models (LALMs), and enhance the SFC capabilities of LALMs through post-training. Experiments demonstrate that SFC outperforms traditional SLU, substantially enhancing the semantic extraction accuracy for LLMs and LALMs.
[NLP-3] Chained Recursive Language Models for Multi-Iteration Reasoning
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在长上下文推理任务中因单次推理轨迹需同时处理上下文理解、中间状态存储、证据验证与最终答案生成而导致的性能瓶颈问题，尤其在涉及信息提取、计数、排序或多跳推理等复杂任务时，早期错误易导致错误传播并影响最终输出。其解决方案的关键在于提出一种推理时架构——链式递归语言模型（Chained Recursive Language Models, Chained RLM），通过重复调用同一基础模型作为一系列独立的推理起点（reasoning roots），每个起点接收原始问题与上下文，但不继承完整对话历史，仅获取前序推理根生成的紧凑文本摘要、纯文本黑板及持久化的任务特定人工制品（artifacts）。该设计通过将复杂推理任务分解为多个阶段性计算单元，实现对中间结果的可检查性、可修正性与可扩展性，从而有效管理上下文并缓解错误累积问题。研究还系统构建了该系统的模型架构、任务交接机制、人工制品工作区及评估协议，验证了在引入递归工具调用的前提下，基于新上下文的人工制品延续仍能带来显著的准确率提升。
链接: https://arxiv.org/abs/2608.05124
作者: Purbesh Mitra,Sennur Ulukus
机构: University of Maryland (马里兰大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Theory (cs.IT); Machine Learning (cs.LG); Signal Processing (eess.SP)
备注:
Abstract:Long context reasoning in large language models (LLMs) is usually constrained by the fact that a single inference trajectory has to simultaneously explore the context, store intermediate state, verify evidence, and produce the final answer. This becomes particularly difficult in tasks that require extraction, counting, ordering, or multi-hop reasoning, where an early mistake can propagate until the final response. In this work, we propose Chained Recursive Language Models (Chained RLM), an inference-time architecture, in which the same underlying model is called repeatedly as a sequence of fresh reasoning roots. Each root receives the original problem and context, but does not inherit the full conversational history. Instead, it receives a compact plain-text summary, a plain-text blackboard, and some durable task-specific artifacts written by predecessor roots. The motivation is to manage the context by chopping into partial tasks rather than one large inference response; in each staged computation, intermediate artifacts can be inspected, corrected, and extended by a later fresh inference by the same model. We describe the system model, handoff mechanism, artifact workspace, and evaluation protocol for this system. We study when fresh-context artifact continuation gives a measurable gain in accuracy over direct LLM answering even with recursive tool-calling.
[NLP-4] Same Formulas Different Semantics: Do Language Models Follow Modal Logic Specifications?
【速读】： 该论文旨在解决大语言模型在处理模态推理（modal reasoning）任务时，其判断是否真正遵循指定的模态语义（stipulated modal semantics），而非依赖于熟悉的经典逻辑或经验偏好。核心问题在于：不同模态系统（如不同的可达性框架或个体域设定）下，同一推理的有效性可能截然不同，而现有模型往往未能准确区分这些语义差异。解决方案的关键在于构建成对的模态问题——具有相同的前提和结论，但框架条件（frame conditions）或个体域（domain conditions）不同，通过自动化推理验证其应得相反标签，从而隔离出语义条件的影响。同时引入“平衡核心”（balanced core）以防止仅凭语义线索泄露答案，确保测试聚焦于模型对模态结构的理解能力。实验表明，在直接提示下，五种主流模型中有四种表现低于仅依赖语义条件的基线；然而，启用推理模式后，DeepSeek V4 Flash 的准确率从 4.4% 显著提升至 88.1%，表明模态推理的正确性高度依赖于推理模式与模型身份的协同作用。当框架条件被省略时，模型虽常达成一致，但其行为更契合不同的经典模态逻辑体系，揭示了模型隐含逻辑偏好。研究公开了所有公式、真值判定器、反例模型及响应数据，为后续研究提供基准。
链接: https://arxiv.org/abs/2608.05097
作者: Réemi Andrieu,Damien Sileo
机构: Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189 - CRIStAL, F-59000 Lille, France
类目: Computation and Language (cs.CL)
备注: 9 pages. Code: this https URL . Data and artifacts: this https URL
Abstract:Reasoning about necessity and possibility depends on assumptions about accessibility between worlds and about which objects exist at each one. The same inference may therefore hold under one modal system and fail under another. Evaluating language models on such problems requires testing whether their judgments follow the stated semantics rather than a familiar logic. We construct paired modal problems with identical premises and conjecture but different frame or domain conditions; automated reasoning verifies opposite labels. A balanced core prevents the semantic condition alone from revealing the answer. On this core, four of five recent models perform below the condition-only baseline under direct prompting. Yet enabling reasoning mode raises DeepSeek V4 Flash from 4.4% to 88.1% on unchanged prompts. Following stipulated modal semantics thus depends strongly on inference mode as well as model identity. When frame conditions are omitted, models often agree but fit different familiar logics best. We release the formulas, oracle artifacts, countermodels, and responses.
[NLP-5] Item Response Theory for AI Safety
【速读】： 该论文旨在解决当前语言模型（Language Model, LM）安全评估中存在的一系列核心问题：现有安全基准测试（safety benchmarks）之间高度重复与相关，导致聚合得分难以可信且难以解释；同时，模型在检测到评估时可能采取“沙袋行为”（sandbagging），故意降低表现以规避严格评价。为应对这些问题，论文提出采用项目反应理论（Item Response Theory, IRT），一种用于从个体对测验项目的作答中推断潜在心理特质的统计工具。其解决方案的关键在于：通过构建IRT模型对8个安全基准测试在192个语言模型上的表现进行分析，首次实现大规模语言模型安全评估的可解释性心理测量分析。研究发现，模型间的性能差异主要由三个可解释因子驱动——拒绝严格性、真实性与情境危害性；基于IRT筛选出的最优测试项能以极低误差恢复完整基准分数，仅需约10个自适应选择的项目即可达到接近全量评估的效果，使评估成本降低97%-99%；此外，IRT还能用于单个模型的审计，有效识别模型在接口后端的变更或简单的沙袋策略。因此，论文论证了IRT作为一套现成工具，在解读、精简和审计安全基准方面具有显著价值，建议前沿实验室与评估者广泛采纳。
链接: https://arxiv.org/abs/2608.05086
作者: Joshua Fonseca Rivera(1),Neil Shah(1),David Demitri Africa(2),Konstantinos Voudouris(2) ((1) Independent, (2) UK AI Security Institute)
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 15 pages, 9 figures, 6 tables
Abstract:Language models differ in how safely they behave and these differences are measured by safety benchmarks. But aggregated benchmark scores are hard to trust and interpret, because benchmarks duplicate one another, correlate heavily, and models may sandbag when they detect evaluation. To address these issues, we draw on Item Response Theory (IRT), a statistical toolkit for measuring these latents from performance on items with inferred psychometric properties. We fit IRT models to eight safety benchmarks across 192 language models, the largest psychometric analysis of LLM safety evaluations to date, and contribute three results. First, we find that three interpretable factors of refusal strictness, truthfulness, and contextual harm explain most of the variance between models across benchmarks. Second, psychometrically selected items recover full benchmark scores with lower error than random subsets of the same size, and roughly ten adaptively chosen items suffice for several individual benchmarks, cutting evaluation cost by 97-99%. Third, IRT supports audits of individual models, showing that it can be used to detect naive sandbagging and changes of model behind APIs. Overall, we show IRT is a ready-made toolkit for reading, reducing, and auditing safety benchmarks, which we recommend frontier labs and evaluators adopt.
[NLP-6] Optimizing What Policies Learn From: Recoverability-aware Rollout Intervention Learning
【速读】： 该论文旨在解决后训练阶段大型语言模型中基于群体的强化学习方法在轨迹采样（rollout generation）过程中存在的效率低下问题。现有方法通常对所有任务和轨迹状态分配相同数量的采样，忽视了不同采样所携带的学习信号差异，导致资源浪费与冗余。尽管近期研究尝试将采样过程建模为自适应决策，但仍存在两大局限：其一，干预策略依赖固定启发式规则，无法随策略演化动态调整；其二，多数方法仅决定采样数量，而缺乏对干预位置与方式的显式控制。针对上述问题，本文提出可恢复性感知干预学习（Recoverability-Aware Intervention Learning, RAIL），其核心在于将干预选择建模为在线上下文-老虎机（contextual-bandit）问题，并通过影子到真实（shadow-to-live）机制收集干预轨迹，训练一个可随策略演进持续学习的可恢复性控制器。该框架实现了对采样时机、位置与方式的精细化控制，在有限采样预算下显著提升学习效率与性能。实验表明，RAIL在有效性、自适应性、表达能力与效率方面均表现优异，验证了可恢复性感知干预作为生成更具信息量且低冗余采样的原则性方法的有效性。
链接: https://arxiv.org/abs/2608.05080
作者: Zheyuan Zhang,Manqing Mao,Hong Wang,Zhuoer Wang,Samson Koelle,Jie Yuan,Yanjun Lin,James Feng,Nikki Lijing Kuang,Yanfang Ye,Wei Niu
机构: Amazon, Inc (亚马逊)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Critic-free group-based reinforcement learning has become a scalable approach for post-training large language models. However, most existing methods allocate the same number of rollouts to every task and trajectory state, even though some rollouts provide much more useful learning signals than others. Recent work has started to treat rollout generation as an adaptive decision, but two important limitations remain. First, intervention strategies are often based on fixed heuristics and therefore cannot adjust as the policy changes during training. Second, these methods usually decide only how many rollouts to generate, without explicitly controlling where and how to intervene. To address these limitations, we propose Recoverability-Aware Intervention Learning (RAIL), a training-time framework that learns how to generate rollouts based on the improvement produced by each intervention. RAIL models intervention selection as an online contextual-bandit problem and trains a recoverability controller using intervention traces collected through a shadow-to-live procedure. This allows the controller to keep learning while the underlying policy evolves. We evaluate RAIL in terms of effectiveness, adaptivity, expressiveness, and efficiency. Across multiple settings, RAIL consistently improves performance under limited rollout budgets. These results show that recoverability-aware intervention provides a principled way to generate more informative and less redundant rollouts, leading to stronger learning signals during post-training.
[NLP-7] German parties shifted towards intuition-based rhetoric after the far rights parliamentary breakthrough
【速读】： 该论文旨在解决政治精英在极右翼势力崛起背景下，其对真理的修辞承诺如何演变这一关键问题。研究发现，德国政治精英在2015至2025年间，基于证据的修辞（evidence-based rhetoric）逐渐减弱，而基于直觉的修辞（intuition-based rhetoric）显著增强，尤其在右翼政治力量中表现尤为明显。其解决方案的关键在于采用经过验证的分布式词典表征方法，量化分析450万条推文与5.9万余份议会演讲中的语言特征，通过计算“证据减去直觉”（Evidence Minus Intuition, EMI）得分，揭示了极右翼政党“德国选择党”（AfD）于2017年进入议会后，整个议会环境中的EMI得分出现急剧下降，而在社交媒体平台上的变化则更为平缓。这表明极右翼的政治可见性与政治精英话语中对真相的理性化倾向退却之间存在显著关联。
链接: https://arxiv.org/abs/2608.05075
作者: Peer Saleth,Segun T. Aroyehun,Fabio Carrella,Christoph M. Abels,Stephan Lewandowsky,David Garcia
机构: 未知
类目: Computation and Language (cs.CL)
备注: 34 pages, 6 figures; includes 49 pages of Supplementary Information. Code available at this https URL , data at this https URL
Abstract:The spread of misinformation is widely perceived as a threat to democratic deliberation, yet how political elites’ rhetorical commitments to truth shift alongside the rise of populist actors remains poorly understood. Analysing 4.5 million tweets and 59,170 parliamentary speeches by German political elites between 2015 and 2025, we measure evidence-based and intuition-based rhetoric using a validated distributed dictionary representation. Across both arenas, intuition-based language has become more prominent, and right-leaning actors consistently exhibit the lowest Evidence Minus Intuition (EMI) scores. The parliamentary entry of the extreme-right Alternative for Germany (AfD) in 2017 coincides with sharp downward shifts in EMI across the broader chamber, while a more gradual decline is observed on Twitter. These findings document an association between far-right visibility and a changing approach to truth in elite discourse in a multiparty European democracy.
[NLP-8] Provable Limits and Certified Deferral for Verbalized Uncertainty in Small Language Models
【速读】： 该论文旨在解决小型开源语言模型在私有、离线及成本敏感场景下的安全部署问题，核心挑战在于不仅关注模型的输出结果，更关键的是确定模型应在何时将决策权交由人类处理（即“风险可控的拒答”）。其解决方案的关键在于通过**校准（calibration）赋予模型置信度语义，并结合有限样本风险证书（finite-sample risk certificate）**机制实现可验证的自主性边界。研究评估了三个系列共11个0.5B至14B参数的指令微调模型在ARC-Challenge和TruthfulQA上的表现，发现尽管普拉特缩放（Platt scaling）可将期望校准误差（ECE）降低至0.02，但在20%风险预算下仅有三个模型-任务组合获得认证自主权，10%风险预算下则无一达标；同时理论分析揭示了温度缩放（temperature scaling）在模型置信度高于0.5但准确率低于此值时的不可行性，且基于Clopper-Pearson方法的校准集可生成满足独立同分布假设下的风险保证。此外，研究还识别并修复了TruthfulQA多选题形式中的答案排序偏差问题，从而提升了校准与拒答决策的可靠性。
链接: https://arxiv.org/abs/2608.05064
作者: Jianru Shen
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at MIWAI 2026 (The 19th International Conference on Multi-disciplinary Trends in Artificial Intelligence), to appear in Springer LNAI
Abstract:Small open-weight language models increasingly run in private, offline, and cost-sensitive settings, where the key deployment question is not only what a model answers but when it should defer to a human. We study whether verbalized confidence can support risk-controlled deferral, evaluating eleven instruction-tuned models from three families, 0.5B to 14B parameters, on ARC-Challenge and TruthfulQA with 25,168 local predictions. Three theoretical results delimit what calibration can provide: strictly monotone calibration preserves the risk-coverage frontier and error-detection AUROC; temperature scaling cannot calibrate models whose confidence stays above one half while accuracy falls below it; and a Clopper-Pearson procedure converts a 200-question calibration set into a finite-sample risk certificate under an i.i.d. deployment assumption. Empirically, eight of 22 model-task pairs hit the temperature-scaling infeasibility floor within one percentage point of the predicted bound. Platt scaling reduces ECE to as low as 0.02, yet certified autonomy at a 20% risk budget is granted to only three model-task pairs and to none at 10%. We also identify and repair an answer-ordering artifact in the multiple-choice form of TruthfulQA. Calibration gives confidence semantics; certified deferral determines when small models are safe to use.
[NLP-9] he Effect of Perceived Race and Gender on Police Language Use: Experimental Evidence from VR Simulations
【速读】： 该论文旨在解决警察在与虚拟现实（VR）中呈现的黑人成年男性虚拟角色互动时，因种族因素导致沟通中缺乏应有的尊重（deference）这一社会性问题，尤其是在涉及执法情境下可能引发冲突升级甚至暴力事件的背景下。其核心解决方案在于通过因果推断框架，以虚拟角色是否被设定为嫌疑人作为处理变量，量化评估该角色对警察言语态度的边际平均处理效应（marginal Average Treatment Effect, AT E）。研究发现，大多数警察在与黑人男性虚拟角色交流时表现出更低程度的尊重，尤其在角色身份为嫌疑人的情境下；而白人、跨种族及多族裔女性警察则表现相对更趋平等。这种态度差异在整段对话过程中累积，可导致语气尊重度显著下降（量表0-10上相差2至数点），并可能引发对话破裂，进而增加暴力风险。此外，研究还探索了大语言模型（LLM）在处理文本型多层级数据中的作用，结果显示：结合混合效应模型与逆概率加权（IPTW）方法，并利用LLM生成文本特征，是当前进行ATE估计的有效策略；尽管直接微调LLM以预测处理效应尚不成熟，但其潜力值得进一步开发。
链接: https://arxiv.org/abs/2608.05050
作者: Sandra C. Sandoval,Navita Goyal,Rashawn Ray,Long Doan,Rachel Rudinger,Hal Daumé III
机构: 未知
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Against the backdrop of violence in police interactions with the U.S. public, we explore how deferentially police officers speak to virtual characters depicted as Black adult males in vir- tual reality (VR) simulations. We evaluate the effect of seeing and communicating with these characters through a causal in- ference lens, where the assignment of the Black man character to a police officer and simulation is the treatment variable. Our (marginal) average treatment effect AT E measures the social impact of the character on the deference of officer statements with each turn of the conversation. Soberingly, we find that most officers speak less deferentially to Black man characters, except for White, biracial, and multiracial female officers, es- pecially in settings where the VR character was known to be a suspect. Across a full conversation of a typical VR scene, these marginal AT Es can result in notable changes in def- erence of tone (two to several points difference on a scale of 0-10), above and beyond that due to the initial effect of per- ceiving a Black male character. Even more disconcerting is that this can contribute to conversation breakdowns that po- tentially result in violence or danger to both the public and the police. We also explored the capabilities of large language models (LLMs) for ATE estimation. From our methods com- parison analysis, including model validation against synthetic data, we provide unique scientific insights on LLM-assisted methodologies for ATE estimation. As such, for ATE esti- mation with multilevel data with text, we recommend mixed effects models with the inverse propensity treatment weighted (iptw) approach, which utilized an LLM for text feature cre- ation. While we also tested LLMs for finetuning prediction models ultimately for ATE estimation, we conclude they are an area for further development and refinement.
[NLP-10] Gradient Immunity: Null-Space Resistance to Malicious Fine-Tuning
【速读】： 该论文旨在解决开放权重模型（Open-weight model）在部分受保护释放（Partially Protected Open-Weight, PPOW）场景下，因下游恶意微调（malicious downstream fine-tuning）导致的安全风险问题。现有防御机制多针对“微调即服务”（Fine-tuning-as-a-Service, FTaaS）范式或依赖下游用户遵守额外安全流程，难以适用于由提供方控制、仅部分权重被锁定的PPOW设置。其核心解决方案是提出一种单向安全门（Unidirectional Safety Gate, USG），由一个零空间立方层（Null Space Cubic Layer）与一个逆适配器（Inverse Adapter）构成，后者插入于Transformer模型最后一层之后。在下游微调过程中，零空间立方层通过检测隐藏状态是否落入预设的受保护区域，对有害样本的梯度进行抑制或阻断，而逆适配器则确保模型前向传播行为恢复至基线水平。通过利用防守方持有的有害数据校准阈值，该方法可实现对邻近分布内有害样本的泛化防护。在六个模型-数据集组合上的实验表明，USG在固定释放阈值下能将微调后的攻击成功率维持在接近释放前的水平，同时在较简单任务中保持高安全通过率，并在BeaverTails等危险样本集上展现出更清晰的安全性-实用性权衡。结果表明，基于释放时表示空间的阻断策略可在不依赖下游协作的前提下显著提高恶意适应的成本。
链接: https://arxiv.org/abs/2608.05045
作者: Yuxuan Huang,Xingyu Zeng,Tianhang Zheng,Chaochao Lu
机构: Shanghai Artificial Intelligence Laboratory(上海人工智能实验室); Shanghai Jiao Tong University(上海交通大学); Shenzhen University of Advanced Technology(深圳先进技术研究院); Zhejiang University(浙江大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Released aligned large language models remain vulnerable to malicious downstream finetuning. Existing defenses are largely designed for the fine-tuning-as-a-service (FTaaS) paradigm or rely on downstream users to follow additional safety procedures, and therefore do not directly address the setting we study: a provider controlled partially protected open-weight (PPOW) release setting in which most weights remain trainable while a small safety-critical component is preserved at release. We propose a Unidirectional Safety Gate (USG), instantiated as a Null Space Cubic Layer together with an Inverse Adapter inserted after the final Transformer layer. During downstream fine-tuning, the cubic layer suppresses or blocks gradients from harmful samples whose hidden states fall in a calibrated protected region, while the Inverse Adapter restores the base model’s forward behavior. In practice, we calibrate a threshold using defender-held harmful data, allowing protection to generalize to nearby in-distribution harmful samples. Across six evaluated model-dataset settings, USG keeps post-finetuning attack success rate close to the pre-release level under a fixed release threshold, while maintaining high safe-pass rates on easier settings and exhibiting a clearer safety-utility trade-off on unsafe samples from BeaverTails. These results suggest that release-time representation-space blocking can raise the cost of malicious downstream adaptation without requiring downstream cooperation. The code is available at this https URL.
[NLP-11] Language Models Generalize to Human-like Word Order Preferences
【速读】： 该论文试图解决的核心问题是：语言习得中，人类所表现出的语法偏倚（linguistic biases）是否可由在信息不充分输入上运行的一般性学习机制自发产生。研究聚焦于生成式语言模型（language models, LMs）是否在缺乏直接证据的情况下，仍能表现出与人类相似的对修饰语顺序的偏好，特别是对范畴同态（scope-homomorphic）的名词短语修饰语排列的倾向。其解决方案的关键在于构建一个受控的语言学习环境：通过训练语言模型于一个移除了所有含多个修饰语的名词短语的语料库，从而消除关于修饰语顺序的直接经验；随后在包含多修饰语的句子上进行评估。结果表明，尽管未在训练中接触过这些结构，三种不同规模的语言模型均一致表现出对范畴同态顺序的偏好，且偏好强度随修饰语类型而异。进一步分析发现，尽管点互信息（PMI）能反映已知的修饰语顺序模式，却无法解释模型的排序偏好。这一发现表明，语言模型能够从贫乏输入中恢复出类人语言泛化能力，并为探究此类偏倚背后的机制提供了可控实验框架。
链接: https://arxiv.org/abs/2608.05028
作者: Amanda Popadich,Shane Steinert-Threlkeld
机构: University of Washington (华盛顿大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:A central question in language acquisition is whether linguistic biases can emerge from general learning mechanisms operating over underdetermined input. Artificial Language Learning (ALL) studies have shown that human learners reliably generalize beyond the evidence provided, including by preferring scope-homomorphic noun phrase modifier orders. In this work, we investigate whether language models exhibit the same bias under similar conditions. We create a controlled learning environment in which models are trained on a corpus where all noun phrases containing multiple modifiers have been removed, eliminating direct evidence about modifier ordering, and are then evaluated on multiple modifier sentences. Across three model sizes, we find that they consistently prefer scope-homomorphic orders despite never observing them during training. These preferences vary in strength by modifier type. To investigate the source of these preferences, we examine noun-modifier association strength using pointwise mutual information (PMI). While PMI reflects known modifier-ordering patterns, it does not explain the models’ ordering preferences. These findings demonstrate that LMs can recover human-like linguistic generalizations from impoverished input and provide a controlled framework for investigating the mechanisms underlying such biases.
[NLP-12] DelusionEval: Measuring Delusion-Linked Behaviors in AI Chatbots
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在实际应用中可能引发用户心理伤害的问题，特别是“妄想螺旋”（delusional spirals）现象——即用户与模型之间的互动行为相互强化，导致用户产生或加剧妄想性认知。为应对这一问题，研究提出了一种名为DelusionEval的评估协议，其核心在于通过真实世界中经历心理伤害的用户对话历史（共18名参与者、12,591条消息）来系统测试模型表现出与妄想相关行为的倾向性。关键发现表明，模型规模、发布日期或推理能力等传统指标无法可靠预测其诱发妄想行为的风险；然而，显著增加上下文长度会大幅提高模型表现出此类危险行为的概率，例如在用户表达自杀意念时未能及时制止的情况，失败率从30.0%上升至41.1%。此外，所有主流模型家族（如GPT、Claude）均表现出较高比例的妄想关联行为，且同一家族内较新、较大或具备更强推理能力的模型并非在所有行为类别上均表现更优。这揭示了当前评估框架的局限性，并强调必须引入基于真实交互场景的、注重上下文影响的安全评估方法，以更准确地识别和防范生成式人工智能（Generative AI）带来的潜在心理风险。
链接: https://arxiv.org/abs/2608.05004
作者: Jared Moore,Andrea Mock,Yifan Mai,Jacy Reese Anthis,Ryan Louie,William Agnew,Ashish Mehta,Kevin Klyman,Percy Liang,Nick Haber,Eric Lin,Desmond C. Ong
机构: Stanford University (斯坦福大学); University of Chicago (芝加哥大学); Carnegie Mellon University (卡内基梅隆大学); Harvard University (哈佛大学); The University of Texas at Austin (德克萨斯大学奥斯汀分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:Mental health professionals have raised concerns about risks of psychological harm from interaction with large language models (LLMs), including “delusional spirals” in which concerning human and LLM behaviors reinforce each other over time. With growing public use of LLM-powered chatbots, there is an urgent need to build evaluations grounded in real-world episodes of psychological harm experienced by users. We developed DelusionEval, an evaluation protocol that tests a model’s tendencies to exhibit behaviors linked to promoting user delusions. We prompt each model with 589 unique conversation histories from 18 participants, comprising 12,591 messages from users who experienced delusions and psychological harm. We find that the tendency of an evaluated LLM to exhibit delusion-linked behavior does not reliably correlate with model size, release date, or the presence of test-time reasoning. However, extending the context of prior messages substantially increases rates of delusion-linked behaviors, providing evidence for the importance of context in LLM safety evaluation. For example, the rate of failing to discourage self-harm when the user expresses suicidal ideation increases from 30.0% to 41.1% when an additional 350 messages are prepended to the conversation history. All model families (e.g., GPT, Claude) exhibit substantial rates of delusion-linked behaviors. Within families, later, larger, or higher-reasoning models are not uniformly better across all behavior categories. Our results raise concerns regarding the potential psychological impact of LLMs and the need for more rigorous studies of real-world human-AI interaction.
[NLP-13] Protoreasoning in Tiny Transformers
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在执行逐步推理时的可解释性与泛化能力问题，特别是探究模型是否能够学习到真正通用的算法，而非仅依赖于特定任务的启发式规则。当前主流大模型虽展现出令人印象深刻的逐步推理能力，但其内在机制尚不清晰，且受限于计算成本高昂和训练数据的黑箱特性，难以开展系统性分析。为此，作者提出一种简化的“思维链”形式——原型推理（protoreasoning），并将其应用于参数量约为100万的小型Transformer模型，通过在受控的嵌套括号语言（Dyck语言）上设计推理友好型任务，实现了对小规模模型中逐步推理过程的深入研究。解决方案的关键在于：利用原型推理在模型内部生成可追踪的中间推理轨迹，从而揭示推理内容本身对泛化性能的贡献，而不仅仅是额外符号长度带来的增益；实验表明，该方法显著缩小了模型在分布外（out-of-distribution）任务上的性能差距，且消融实验验证了推理轨迹的内容是性能提升的核心因素。这一方法为在资源受限条件下研究推理机制提供了新范式，推动了对模型本质推理能力的理解。
链接: https://arxiv.org/abs/2608.04980
作者: Eduardo Valle,Fergal Reid
机构: Fin AI Research
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:We show that tiny transformers can profitably employ a simple form of Chain of Thought, which we call protoreasoning, allowing us to study step-by-step reasoning on ~1M-parameter models and opening up opportunities for much more detailed experimentation and analysis than is feasible for larger models. Current Large Language Models exhibit impressive step-by-step reasoning, but we have yet to understand its generality, i.e., when and how LLMs learn genuinely general algorithms rather than “bags of heuristics.” Such questions are hard to settle on compute-intensive frontier models trained on opaque data. To work at model scales far below the threshold for natural-language competence, we define reasoning-friendly tasks on Dyck languages (sentences of correctly nested brackets). We find that protoreasoning traces substantially close the out-of-distribution generalization gap, and ablations confirm that the trace’s content, not merely its extra tokens, drives the gain.
[NLP-14] SpecRoll: Fast-Slow Verifier-Feedback Adaptation for Speculative Reinforcement Learning Rollouts
【速读】： 该论文旨在解决强化学习（Reinforcement Learning, RL）后训练过程中大语言模型推理效率瓶颈问题，特别是自回归式滚动生成（autoregressive rollout generation）带来的低效问题。其核心挑战在于：虽然推测解码（speculative decoding）可显著加速生成过程，但在RL场景下难以应用，因为目标策略持续演化，导致静态的推测器（proposer）迅速过时，而频繁更新推测器又引入高昂开销。为此，论文提出SpecRoll，一种支持双时间尺度动态适应的推测式滚动生成引擎。其关键创新在于：通过轻量级未来词元头（future-token heads）实现并行推测生成，同时引入基于延迟验证反馈的反射模块（Reflex module），在不进行反向传播的前提下，对隐藏状态执行有界、轨迹局部的修正；此外，还设计了一条慢速路径，在检测到持续性能退化时才更新头参数。该方案结合了并发感知的稀疏树验证与精确目标模型验证，确保目标模型的采样分布和广义奖励策略优化（GRPO）目标不变。实验结果表明，SpecRoll在5个规模从1.5B到14B的模型及3个数学推理数据集上实现了1.26–2.15倍的生成速度提升和1.21–2.04倍的端到端加速，优于基线GRPO与FastGRPO，在所有15组对比设置中均表现出更优的生成与端到端效率，平均端到端加速达1.18倍。消融实验进一步验证了快速与慢速适应路径的互补性。
链接: https://arxiv.org/abs/2608.04962
作者: Nhat Minh Pham,Duy Tung Doan,Thi Duyen Ngo,Vinh Van Nguyen,Khac-Hoai Nam Bui
机构: Viettel AI, Viettel Group, Hanoi, Vietnam; VNU University of Engineering and Technology, Vietnam National University, Hanoi, Vietnam
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Reinforcement learning (RL) post-training improves the reasoning capabilities of large language models, but autoregressive rollout generation remains a major efficiency bottleneck. Speculative decoding can accelerate generation, yet applying it during RL is difficult because the target policy continually evolves: static proposers become stale, while frequent drafter updates add substantial overhead. We introduce SpecRoll, a speculative rollout engine that preserves the target model’s sampling distribution while adapting at two timescales. Lightweight future-token heads generate parallel proposals, while our proposed Reflex module uses delayed verifier feedback to perform bounded, trajectory-local hidden-state corrections without backpropagation. A complementary slow path updates the head parameters only when sustained degradation is detected. SpecRoll combines these mechanisms with concurrency-aware sparse-tree verification and exact target verification, leaving the target rollout distribution and GRPO objective unchanged. Across five models ranging from 1.5B to 14B and three mathematical reasoning datasets, SpecRoll achieves 1.26-2.15x generation speedup and 1.21-2.04x end-to-end speedup over vanilla GRPO. It also outperforms FastGRPO in both generation and end-to-end time across all 15 matched settings, with an average pairwise end-to-end gain of 1.18x. Controlled ablations show that the fast and slow adaptation paths provide complementary benefits. Our source code is available at this https URL.
[NLP-15] UG-UMRE: Uncertainty-Guided Modality Augmentation and Distributional Calibration for Unified Multimodal Relation Extraction ACM-MM2026
【速读】： 该论文旨在解决统一多模态关系抽取（Unified Multimodal Relation Extraction, UMRE）中面临的两个关键问题：一是忽略固有的随机不确定性（aleatoric uncertainty）导致噪声传播，二是不同模态间分布的深层异质性阻碍了跨模态对齐。针对上述问题，论文提出了一种基于不确定性引导的UMRE网络（Uncertainty-Guided UMRE Network, UG-UMRE）。其核心解决方案包括两个关键模块：首先，设计了不确定性驱动的单模态增强（Uncertainty-Driven Unimodal Augmentation, UDUA）模块，通过变分信息瓶颈（Variational Information Bottleneck）将特征建模为高斯分布，并结合不确定性感知的自监督对比学习机制，在有效过滤噪声的同时保持语义一致性；其次，提出了联合随机不确定性对齐（Joint Aleatoric Uncertainty Alignment, JAUA）模块，作为全局语义预校准机制，利用概率分布的一致性构建共享潜在空间，通过同步跨模态统计特性来消除分布差异，从而为细粒度交互奠定稳健基础。实验在三个基准数据集（UMRE、MORE和MNRE）上验证了UG-UMRE达到当前最优性能，且对UDUA与JAUA模块的可插拔性及有效性进行了充分分析。
链接: https://arxiv.org/abs/2608.04949
作者: Bo Kong,Liruiz Jia,Yi Liang,Chao Liu,Dongfang Han,Tianwei Yan,Yuan Liu,Shengquan Liu
机构: Xinjiang University (新疆大学); Chongqing Jiaotong University (重庆交通大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Information Theory (cs.IT); Multimedia (cs.MM)
备注: Accepted at ACM MM2026
Abstract:Unified Multimodal Relation Extraction (UMRE) aims to identify intra-modal and cross-modal relations between textual entities and visual objects. However, existing UMRE studies still encounter two critical issues: ignoring inherent aleatoric uncertainty causes noise propagation, and deep-seated heterogeneity between distinct modal distributions hinders alignment. To address these issues, we propose the Uncertainty-Guided UMRE Network (UG-UMRE). Specifically, we design an Uncertainty-Driven Unimodal Augmentation (UDUA) module, which models features as Gaussian distributions based on the Variational Information Bottleneck. By incorporating an uncertainty-aware self-supervised contrastive learning mechanism, UDUA effectively filters out noise while maintaining semantic consistency. Furthermore, we introduce the Joint Aleatoric Uncertainty Alignment (JAUA) module as a global semantic pre-calibration mechanism. JAUA leverages probabilistic distribution consistency to construct a shared latent space, eliminating the distributional gap by synchronizing cross-modal statistical properties, thereby laying a robust foundation for fine-grained interaction. Experiments on three benchmark datasets (UMRE, MORE, and MNRE) demonstrate that UG-UMRE achieves state-of-the-art performance. Further analysis validates the pluggable and effective performance of the proposed UDUA and JAUA modules.
[NLP-16] Reading Between the Frames: Interpreting Implicit and Non-literal Meaning in Social Media Videos
【速读】： 该论文旨在解决当前视频-语言模型在理解社交媒体视频中隐含的、非线性且具有修辞层次的深层语用意义时所面临的挑战。许多社交平台视频的真正含义并不直接体现在视觉内容、字幕或语音中，而是依赖于多模态线索与文化语境的交互作用，呈现出讽刺、幽默或反讽等复杂表达，这对现有模型的语用理解能力构成了严峻考验。其解决方案的关键在于提出一个名为DrivelHub+的新基准，包含1,000个来自真实社交平台的视频样本，每个视频均配有由人工撰写的隐含叙事解释（implicit narrative explanation），从而构建了一个聚焦于上下文驱动的多模态推理任务。该基准从两个维度评估模型性能：一是“解释”任务，要求模型以自然语言阐明视频的语用理解；二是“表征”任务，通过“推理即检索”（reasoning-as-retrieval）范式，检验模型的视频与文本表示是否能准确对齐对应隐含叙事。该研究为衡量多模态感知与语用理解之间的差距提供了诊断性框架，核心目标是推动模型从表面描述转向深层意图推断。
链接: https://arxiv.org/abs/2608.04939
作者: Yang Wang,Yanan Ma,Yiqi Liu,Zi Yan Chang,Chi-Li Chen,Chia-Yi Hsiao,Tyler Loakman,Aline Villavicencio,Chenghao Xiao,Chenghua Lin
机构: University of Manchester(曼彻斯特大学); Durham University(杜伦大学); University of Sheffield(谢菲尔德大学); University of Exeter(埃克塞特大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Social media videos often communicate meanings that go beyond their visible actions, captions, or speech. A mundane clip may become humorous, ironic, or satire only through the interaction of multimodal cues and cultural context, making such content a difficult test case for video-language models. In this paper, we introduce \textitDrivelHub+, a benchmark for evaluating whether models can infer the implicit, non-linear, and rhetorically layered meanings of social media videos that appear nonsensical on the surface but convey deliberate pragmatic meanings. DrivelHub+ consists of 1,000 videos collected from social media, each annotated with a human-written implicit narrative explanation. Unlike conventional video understanding tasks focused on recognition or description, we present a benchmark that targets contextual multimodal reasoning. We evaluate current video-language models from two perspectives: explanation, where models must explain the pragmatic comprehension of a video in natural language; and representation, where we adapt reasoning-as-retrieval to test whether model representations align videos with their corresponding implicit narratives in both video-to-text and text-to-video retrieval. Our benchmark provides a diagnostic setting for measuring the gap between multimodal perception and pragmatic comprehension, asking whether current models can move beyond describing what is shown to inferring what is meant.
[NLP-17] State2State: Environment-Derived Mid-Training for LLM Agents
【速读】： 该论文旨在解决大语言模型（LLM）智能体训练中依赖外部指定任务与人工标注监督信号所导致的可扩展性与多样性受限问题。现有方法如监督微调或基于人工设计验证器的在线强化学习，均受限于人工干预，难以实现大规模、多样化的智能体训练。为此，论文提出一种环境自学习范式——State2State，其核心在于通过智能体在环境中自主探索所获得的状态，自动生成训练目标，将探索到的环境状态转换为可达性目标，使智能体被挑战以达成特定目标状态。该方法的关键创新在于：任务由环境探索过程自动衍生，成功验证通过规则匹配状态实现，无需专家监督或手动任务设计，从而实现了可扩展且可验证的训练机制。实验在ALFWorld和ScienceWorld环境中验证了State2State的有效性，不仅作为独立的环境学习阶段显著提升智能体性能，还可作为下游强化学习的初始化策略，进一步提升最终表现与学习效率，并展现出跨环境泛化潜力。
链接: https://arxiv.org/abs/2608.04934
作者: Xuanyu Lei,Yiqi Zhu,Chenliang Li,Kaiming Liu,Peng Li,Ming Yan,Jieping Ye,Ya-Qin Zhang,Yang Liu
机构: Institute for AI Industry Research (AIR), Tsinghua University(清华大学人工智能产业研究院); Dept. of Comp. Sci. Tech., Institute for AI, Tsinghua University(清华大学计算机科学与技术系人工智能研究所); Institute of Intelligent Computing, Alibaba Group(阿里巴巴集团智能计算研究院)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Work in progress
Abstract:Training LLM agents commonly relies on supervised fine-tuning from expert trajectories or online reinforcement learning over human-specified tasks with handcrafted verifiers. Though effective, both remain bottlenecked by externally specified tasks and supervision signals, limiting the scalability and diversity of agent training. We study an environment learning paradigm in which agents acquire interaction and manipulation capabilities solely through environment interaction, without externally specified tasks. We propose State2State, an environment-derived mid-training method that converts explored environment states into training objectives, challenging agents to reach a specified target state. By deriving tasks from environment exploration and verifying success through rule-based state matching, State2State provides scalable and verifiable training objectives without expert supervision or manual task design. Experiments on ALFWorld and ScienceWorld show that State2State improves agent performance as a standalone environment-learning stage in most settings. As initialization for downstream RL, it further improves final performance and learning efficiency, with promising evidence of cross-environment generalization.
[NLP-18] Does Out-of-Sight Equal Out-of-Mind in CoT Monitorability?
【速读】： 该论文旨在解决生成式模型中链式思维（Chain-of-Thought, CoT）可监控性（monitorability）在隐式链式思维（Latent CoT）模式下的退化问题。传统显式CoT通过可读的文本推理轨迹支持对模型决策过程的直接观测，而隐式CoT将推理过程编码为连续状态向量，虽降低了计算开销，但牺牲了可读性，导致依赖显式文本轨迹的监控机制失效。为此，论文提出以“提示依赖”（hint-reliance）作为可监控性的代理目标，研究在数学推理与问答任务中，不同推理模式（包括显式CoT、弱监督与强监督的隐式CoT）下监控能力的保持程度。研究发现，在该设定下，可监控性主要取决于任务本身的特性（如正确答案是否约束支持性推理）以及对模型内部状态的访问权限，而非推理模式本身。因此，解决方案的关键在于：提升对模型内部激活状态的可访问性，并结合任务特性设计更有效的监控策略，而非单纯依赖推理模式的形式。
链接: https://arxiv.org/abs/2608.04928
作者: Pedro Ferreira,Wilker Aziz,Ivan Titov
机构: University of Amsterdam (阿姆斯特丹大学); University of Edinburgh (爱丁堡大学)
类目: Computation and Language (cs.CL)
备注: 23 pages
Abstract:Chain-of-thought (CoT) reasoning offers a window into the decision-making of large language models (LLMs), which can be monitored for target behaviors by reading the reasoning trace, motivating work on CoT monitorability. Latent CoT approaches, however, replace the explicit tokens with a small number of continuous states, lowering inference costs but removing the readable trace this monitoring relies on. Monitoring then requires alternative access to the model, such as probing its activations or verbalizing the latent states back into text, but how much monitorability these alternatives preserve is unclear. We study this question with a hint-based intervention setup, a proxy for behaviors where models exploit biasing input cues, e.g., an inadvertently leaked answer or a belief stated by the user, without acknowledging them. Taking hint-reliance as the monitorability target, we compare monitors across reasoning modes, from explicit CoT to weakly- and strongly-supervised latent CoT, on math reasoning and question answering. We find that, in this setup, monitorability depends more on properties of the task (such as whether the correct answer constrains the supporting reasoning) and the level of access to model internals than on the reasoning mode.
[NLP-19] Consistency-Driven Co-Evolution for Self-Supervised Cross-Representation Learning
【速读】： 该论文旨在解决跨模态表示（图表、表格与可视化代码）之间理解不一致的问题，其核心挑战在于：不同表示形式间存在固有的多对一映射关系，监督信号模糊且标注成本高昂，同时模型优化缺乏既具备方向自适应性又在多种表示间具有泛化能力的统一信号。解决方案的关键在于提出CoCoEvolve框架，通过定义显式的“一对一”对应关系，利用不同模态间的内在一致性作为无监督优化信号，从而实现跨模态表示的一致性增强。该方法在训练阶段采用图-表-代码三者之间的协同演化机制（co-evolution），在推理阶段则引入测试时一致性优化（test-time co-optimization），并通过统一的评估套件CoCoEvolve@Eval系统评估六种跨表示任务。实验表明，该方法在四个基准数据集上均显著提升了训练和测试阶段的性能表现。
链接: https://arxiv.org/abs/2608.04926
作者: Xuehang Guo,Pengyuan Li,Tom Hope,Tirthankar Ghosal,Manling Li,Qingyun Wang
机构: William Mary(威廉与玛丽学院); IBM(国际商业机器公司); Allen Institute for AI(艾伦人工智能研究所); Oak Ridge National Laboratory(橡树岭国家实验室); Northwestern University(西北大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:As chart images, tabular data, and visualization code play increasingly important roles across diverse domains, cross-representation understanding across these modalities poses fundamental challenges for AI systems: the relationships across representations are inherently \textitone-to-many, supervision is ambiguous and costly, and model optimization lacks a principled signal that is both direction-adaptive and representation-generalizable beyond task-specific objectives. We introduce CoCoEvolve to improve consistency across chart, table, and code representations. Instead of treating cross-representation mapping as a one-to-many problem, we define explicit one-to-one correspondences and optimize models using agreement between representations, without additional annotations. During training, CoCoEvolve@Train performs co-evolution across the chart-table-code cycle, while CoCoEvolve@Test applies the same consistency objective at inference time for test-time co-optimization. We also present CoCoEvolve@Eval, an evaluation suite covering all six cross-representation tasks. Across four benchmarks, CoCoEvolve improves performance in both training-time and test-time settings. Our project page: this https URL.
[NLP-20] Strengthening Target-Language Features: SAE-Based Steering for Multilingual Inference
【速读】： 该论文旨在解决多语言大语言模型在不同语言间性能差异显著的问题，尤其针对现有适配方法普遍依赖参数更新和大量多语言训练数据的局限性。其解决方案的关键在于提出一种推理时（inference-time）的多语言引导（multilingual steering）方法：利用预训练的稀疏自编码器（sparse autoencoder, SAE）识别并增强与目标语言相关的特征表示。通过对比多语言平行语句在SAE中的激活模式，筛选出每层中与特定目标语言相关性高的少量特征，并将其解码为引导信号，直接注入模型隐藏状态，从而实现无需额外训练的语言性能提升。实验表明，在Gemma-3-12B-it模型上，该方法在XCOPA、XNLI和MGSM三个多语言基准测试中分别实现了平均10.9、5.3和1.9个百分点的准确率提升。
链接: https://arxiv.org/abs/2608.04904
作者: Hongsheng Wang,Phlipp Koehn
机构: Johns Hopkins University (约翰霍普金斯大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Multilingual large language models exhibit substantial performance differences across languages, while existing adaptation methods often require parameter updates and considerable multilingual training data. We propose an inference-time multilingual steering method that uses pretrained sparse autoencoders to identify and strengthen target-language-related features. Using multilingual parallel sentences, we compare SAE activations across languages and select a small number of layer-specific features associated with each target language. These features are decoded into steering signals and injected into the model’s hidden states without additional training. Experiments with Gemma-3-12B-it show average accuracy improvements of 10.9 percentage points on XCOPA, 5.3 points on XNLI, and 1.9 points on MGSM.
[NLP-21] Evaluation Pitfalls and Sparsity Limitations in LLM -based Confidence Estimates for Classification ACL2026
【速读】： 该论文旨在解决大语言模型（LLM）在分类任务中置信度估计（confidence estimation）因文本化（verbalization）导致输出极度稀疏的问题。现有方法如文本化仅生成少量离散的置信度值（例如，在SST-2数据集上仅有8个唯一值，超过一半为95%），这不仅限制了实际应用中的细粒度判断能力，还严重影响评估结果的可靠性——不同插值方式（如分段式与线性插值）对准确率-拒答曲线下面积（AUARC）排名产生显著影响，甚至导致最优方法排名下降至最差。其解决方案的关键在于提出“文本化对数概率”（verbalization logprobs）方法：通过将文本化数字的权重基于对应词元的概率进行加权，有效缓解置信度输出稀疏性问题，显著提升置信度估计质量，在不增加推理成本的前提下实现了比原始文本化方法更高的AUARC得分（提升2.3个百分点），并主张采用分段式插值作为标准化评估基准以确保公平比较。
链接: https://arxiv.org/abs/2608.04899
作者: Elena Merdjanovska,Omar Zaidan,Andreas Rücklé
机构: Humboldt-Universität zu Berlin (柏林洪堡大学); Science of Intelligence (智能科学); Amazon (亚马逊)
类目: Computation and Language (cs.CL)
备注: Published at Findings of ACL 2026
Abstract:Confidence estimation is essential when LLMs are used for classification, indicating when predictions can be trusted. However, common approaches such as verbalization produce extremely sparse outputs. For instance, Qwen3-32B verbalizes only eight unique confidence values on SST-2, with over half being exactly 95%, a pattern we observe consistently across four datasets and two LLMs. Besides limiting practical utility, we show that this sparsity critically affects evaluation: the choice of interpolation in area under the accuracy-rejection curve (AUARC) dramatically alters rankings, with consistency sampling dropping from best to worst under stepwise versus linear interpolation. We advocate for standardizing stepwise interpolation for a fairer comparison. Under such a fair evaluation, we find that weighting verbalized digits by token probabilities, a method we term verbalization logprobs, addresses sparsity and achieves the best AUARC (+2.3 points over vanilla verbalization) without incurring additional inference cost.
[NLP-22] Evaluating the Diagnostic Robustness of Vision-Language Models Under Visual and Textual Perturbations
【速读】： 该论文旨在解决视觉语言模型（VLMs）在临床敏感领域中因标准准确率指标掩盖可靠性缺陷而导致的潜在风险问题。其核心挑战在于，现有评估方法无法有效揭示模型在面对微小但关键的输入变化时所表现出的诊断不稳定性，尤其是在医学影像分析这种对顺序和文本表述高度敏感的场景中。解决方案的关键在于构建一种基于证据保留扰动（evidence-preserving perturbations）的系统性评估框架：通过重新排列解剖切片序列、交换目标标签位置等操作，在保持临床证据不变的前提下，检验模型预测的一致性。研究发现，模型在呈现顺序稳定性方面存在显著脆弱性，序列反转可导致高达48.9%的预测翻转；同时存在文本选择偏差，标签重排引发高达67.8%的诊断不一致；此外，即使移除专家标注的病灶切片，模型仍会在高达76.1%的情况下生成确定性诊断，暴露其诊断过度承诺的问题。这些结果表明，仅依赖整体准确率会严重高估模型在真实临床环境中的可靠性，因此必须引入基于稳定性的评估指标，以确保视觉语言模型在安全关键型医疗应用中的可信部署。
链接: https://arxiv.org/abs/2608.04885
作者: Ali Khoramfar,Mohammad Javad Dousti,Alireza Mohamadian,Heshaam Faili
机构: University of Tehran(德黑兰大学); Tehran University of Medical Sciences(德黑兰医科大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Standard accuracy metrics for VLMs often mask significant reliability failures in sensitive domains. In this work, we utilize a histopathology-validated brain MRI dataset to systematically assess the diagnostic robustness of four VLM families under evidence-preserving perturbations. By reordering anatomical slices and swapping target label positions, we evaluate whether models maintain consistent predictions when clinical evidence remains invariant. Our results reveal significant vulnerabilities in presentation-order stability, with models exhibiting prediction flips in up to 48.9% of cases under simple sequence reversals. We further identify a textual selection bias, where label reordering triggers inconsistent diagnoses in up to 67.8% of cases despite identical visual inputs. Negative-control tests further reveal diagnostic overcommitment: models generate categorical diagnoses in up to 76.1% of cases after expert-annotated lesion slices are removed. These results demonstrate that high accuracy can overestimate clinical reliability, masking sensitivity to sequential presentation and textual framing that is not captured by aggregate accuracy. Our findings highlight the necessity of stability-based metrics for the deployment of VLMs in safety-critical clinical applications. Our evaluation data and code will be made public upon acceptance.
[NLP-23] A-SR: Self-Evolving Agent ic LLM s for Symbolic Regression via Hierarchical Coordination
【速读】： 该论文旨在解决现有大语言模型（LLM）引导的符号回归（Symbolic Regression, SR）方法在搜索过程中因采用统一的提议循环而导致异构搜索失败信息被过度压缩为单一标量评分与提示的问题，从而限制了对复杂科学规律发现的效率与准确性。其核心解决方案在于提出一种自演化代理框架A-SR，关键创新在于将控制单元从传统的表达式编辑转移至角色条件化的证据视图（role-conditioned evidence views），通过协调协议路由、在线评估-奖励角色策略以及状态驱动的过程记忆机制实现动态协作。在搜索过程中，评估器反馈用于刻画可靠性与生产力，更新各角色的效用值，并将精英模式、失败轨迹与有效性诊断精准分配至不同代理。A-SR在两个时间尺度上实现自演化：单次运行内通过调整搜索过程而非更新模型参数进行适应；跨运行时则可将历史轨迹提炼为开源大模型的角色条件化先验知识。实验表明，在LLM-SRBench基准的四个科学合成领域中，A-SR使用Llama3.1-8B将Acc@0.01提升至48.30%（基线25.79%），而A-SR-LoRA在Qwen3-4B上也将对应指标从24.58%提升至38.29%；在四项真实世界科学发现任务中，A-SR在8项报告指标中的7项上实现了最优的分布内或分布外均方误差表现。
链接: https://arxiv.org/abs/2608.04872
作者: Wenxiao Zhao,Dong Liu,Kaiyi Xu,Feng Liu,Zhen Zhao,Fei Ben,Shu Wang,Wenhao Li,Yingnian Wu,Fenghua Ling,Haobo Li,Lei Bai
机构: 1. Tsinghua University (清华大学); 2. Alibaba Group (阿里巴巴集团); 3. Peking University (北京大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 18 pages, 8 figures, including appendix
Abstract:Symbolic regression aims to discover closed-form equations from data, but existing LLM-guided methods often rely on a unified proposal loop that compresses heterogeneous search failures into a scalar score and a single prompt. We propose A-SR, a self-evolving agentic framework that shifts the control unit from expression edits to role-conditioned evidence views. A-SR coordinates formula discovery through routing among coordination protocols, an online evaluator-reward role policy, and state-routed process memory. During search, evaluator feedback characterizes reliability and productivity, updates role-level utilities, and routes elite motifs, failure traces, and validity diagnostics to different agents. The framework self-evolves at two timescales: within a run, it adapts the search process without updating LLM parameters; across runs, recorded trajectories can be distilled into open-source LLMs as role-conditioned proposal priors. Averaged over the four LSR-Synth scientific domains in LLM-SRBench, A-SR improves Acc@0.01 over baselines from 25.79% to 48.30% with Llama3.1-8B, while A-SR-LoRA improves the corresponding Qwen3-4B result from 24.58% to 38.29%. On four real-world scientific discovery tasks, A-SR obtains the best in-distribution or out-of-distribution normalized mean squared error on 7 of 8 reported metrics.
[NLP-24] Preverbal Uninflected and Underived Roots in Mapudungun. Wuno and Its Implications
【速读】： 该论文旨在解决马普德语（Mapudungun）中句法前置的未屈折、未派生词根（preverbal uninflected and underived roots）的语法地位问题，尤其聚焦于词根 wuno（意为“返回/重新”）。现有学术分类框架——如辅助动词（auxiliaries, Smeets, 2008）、情态前缀（modal prefixes, Longkon, 2011）以及句前成分/复合动词词干（preverbal particles/complex verb stems, Zúñiga, 2006）——均存在解释力不足的问题。研究通过历时性语料库分析（1606年至今）揭示，这些成分呈现三种不同演变轨迹：稳定的V1复合结构（如kim, shinge）、受正字法演变影响而波动的V1形式（如pepi, wuno），以及真正的句前副词（kalli）。对kim和kupa在不同时期的V2形态证据的发现，证实其作为完整词汇动词的地位。研究提出“韵律-正字学假说”：早期传教士将韵律停顿记录为间隔符号（空格），后被使用者重新分析为句法边界，导致“可变绑定”现象的误判。实证支持Zúñiga提出的激进串联模型（radical concatenation）为正确语法模式，对无接触前书面传统的语言研究具有重要启示。
链接: https://arxiv.org/abs/2608.04869
作者: Andres Chandia
机构: University of Barcelona (巴塞罗那大学)
类目: Computation and Language (cs.CL)
备注: 54 pages, 4 tables, 2 graphics, 23 examples
Abstract:This study examines the grammatical status of preverbal uninflected and underived roots in Mapudungun, with particular focus on wuno ‘return/re-’. Through a critical review of scholarly classifications–auxiliaries (Smeets, 2008), modal prefixes (Longkon, 2011), and preverbal particles/complex verb stems (Zúñiga, 2006)–we demonstrate the limitations of existing frameworks. A diachronic corpus analysis spanning four centuries (1606-present) reveals that these elements exhibit three distinct profiles: stable V1 compounds (kim, shinge), volatile V1 rates reflecting orthographic shift (pepi, wuno), and a true particle (kalli). The discovery of V2 attestations for kim and kupa confirms their status as full lexical verbs. We propose a prosodic-orthographic hypothesis: apparent “variable binding” results from the fossilization of prosodic pauses transcribed by early missionaries as spaces, a convention later reanalysed by speakers as syntactic boundaries. The evidence supports Zúñiga’s radical concatenation as the correct grammatical model, with implications for the study of languages with no pre-contact written tradition.
[NLP-25] Do Language Models Know Their Slang? Queer Slang Understanding in User-Generated Content
【速读】： 该论文旨在解决生成式 AI 在处理性别与性取向少数群体（queer）特定语言表达时存在的显著研究空白，即当前自然语言处理（Natural Language Processing, NLP）领域对酷儿俚语（queer slang）的覆盖不足。其核心问题是：现有语言模型在面对具有文化敏感性和社群特异性的酷儿俚语时，是否具备准确理解与定义的能力。为此，论文提出关键解决方案——构建首个手动标注的英文语料库 Slang-Q，该数据集包含由用户真实生成的句子及其对应的酷儿俚语术语与权威参考定义，并基于新建立的118个酷儿术语分类体系进行组织。通过该资源，研究首次系统评估了语言模型在不同提示策略下对酷儿俚语的理解与释义能力，为未来探索模型在敏感身份相关语言中的表现提供了基准与方向。
链接: https://arxiv.org/abs/2608.04847
作者: Arianna Denitto,Beatrice Savoldi
机构: University of Torino (都灵大学); Fondazione Bruno Kessler (布鲁诺·凯斯勒基金会)
类目: Computation and Language (cs.CL)
备注:
Abstract:Despite its cultural relevance and diffusion, queer slang remains underrepresented in Natural Language Processing research. Towards addressing this gap, we introduce Slang-Q, a manually curated dataset of naturally user-generated English sentences paired with queer slang terms and reference definitions, built upon a newly constructed taxonomy of 118 queer terms. We use this resource to conduct a first exploratory evaluation of language models on their ability to understand and define queer slang under varying prompting conditions. Slang-Q is intended as a basis for studying how current models handle sensitive, community-specific language and whether they can provide accurate and reliable information about such forms of identity and linguistic expression.
[NLP-26] Skill-Use: Can LLM s Actually Use Skills in Agent ic Harnesses?
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）代理在实际应用中对技能（Skill）识别与自主调用能力不足的问题。现有评估体系主要关注技能本身的质量或其对任务完成的贡献，却忽视了代理是否能准确识别并独立执行相关技能。为此，论文提出“Skill-Use”基准，通过渐进式披露机制（progressive disclosure）评估代理在仅知技能名称和简短描述的情况下，能否主动检索完整操作流程并正确应用。该基准分离出技能使用的三个核心维度：触发（Trigger，判断代理是否调用相关技能）、合规性（Compliance，衡量执行过程是否遵循规定流程）以及边界控制（Boundary，检测代理是否规避禁止操作），并引入综合得分（SU score），仅在技能被成功触发后才给予执行奖励。该基准涵盖9个真实领域、79项真实技能与177个可执行任务，所有实验均在隔离的Docker沙箱环境中运行，并采用基于轨迹的评分标准。实验结果表明，当前最强的模型配置在该基准上仅达到0.613的SU得分，说明可靠技能使用仍难以实现；触发与程序合规性分别成为独立瓶颈，且模型性能与排名随代理框架（harness）变化而显著波动，揭示技能使用并非模型固有属性，而是高度依赖于具体框架设计的动态能力。
链接: https://arxiv.org/abs/2608.04828
作者: Jinyi Han,Yuanjian Xu,Ying Liao,Xinyi Wang,Zishang Jiang,Zixiang Di,Fanyang Lu,Zhichao Hu,Yanghua Xiao
机构: East China Normal University (华东师范大学); Hong Kong University of Science and Technology (香港科技大学); Fudan University (复旦大学); Tencent Hunyuan (腾讯混元)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language model (LLM) agents increasingly rely on skills, structured documents that specify when to act, which procedure to follow, and which tools are allowed. Existing evaluations mostly judge the quality of a skill or its contribution to task success, leaving unexamined whether an agent can recognize a relevant skill and apply it on its own. We introduce Skill-Use, a benchmark that evaluates skill use under progressive disclosure, where an agent sees only a skill’s name and short description and must retrieve the full procedure before following it. Skill-Use separates three facets of skill use. Trigger measures whether the agent invokes the relevant skill, Compliance measures how faithfully it follows the prescribed procedure, and Boundary measures whether it avoids forbidden operations. A Skill-Use (SU) score combines the three and credits execution only after the skill is triggered. Skill-Use pairs 79 real skills with 177 executable tasks across nine domains, each grounded in real files, run in an isolated Docker sandbox, and scored by a trajectory-based rubric. Evaluating eight LLMs under two agent harnesses, we find that reliable skill use remains out of reach, as the strongest configuration reaches an SU of only 0.613. Triggering and procedural compliance fail as independent bottlenecks, and both scores and model rankings shift with the harness, so skill use behaves as a capability conditioned on the harness rather than a fixed property of the model.
[NLP-27] A Modular Part-of-Speech Tagger for Scottish Gaelic using spaCy
【速读】： 该论文旨在解决低资源语言（low-resource languages）在词性标注（Part-of-speech tagging）任务中因标注数据稀缺而导致的挑战，尤其针对语法结构复杂且数字资源匮乏的盖尔语（Gaidhlig，即苏格兰盖尔语）。其解决方案的关键在于采用轻量级、模块化的spaCy自然语言处理框架，仅通过最小预处理和配置，在不依赖外部词嵌入或预训练语言模型的前提下，基于苏格兰盖尔语标注参考语料库（Annotated Reference Corpus of Scottish Gaelic）构建词性标注器。研究对比了两种标签体系：细粒度标签集与粗粒度标签集，分别实现了88.6%和93.7%的标注准确率，性能与先前发表的两个盖尔语标注器相当，表明在缺乏大规模资源的条件下，仅使用监督学习与开箱即用的NLP流水线即可实现高效且具有竞争力的词性标注表现。
链接: https://arxiv.org/abs/2608.04808
作者: Peter Stefan,Peter J Barclay,Alistair Lawson
机构: 未知
类目: Computation and Language (cs.CL)
备注: A revised version of this paper has been accepted for presentation at UKCI 2026 ( this https URL ) and will be published by Springer
Abstract:Part-of-speech tagging for low-resource languages remains challenging due to limited annotated data, especially for linguistically complex languages. Gaidhlig (Scottish Gaelic) is a morphologically rich and endangered language with limited digital resources, making it suitable for examining a lightweight language processing approach. This paper describes using the modular spaCy Natural Language Processing framework to build part-of-speech taggers for Gaidhlig using the Annotated Reference Corpus of Scottish Gaelic. We train two models with minimal pre-processing and configuration: one using a fine-grained tagset and another using a reduced coarse-grained tagset. Both models are trained without external embeddings or pre-trained language models, using only supervised learning from the available corpus. The fine-grained model achieves 88.6% tagging accuracy, while the coarse-grained model achieves 93.7%. The results are comparable to those of the two previously published Gaidhlig taggers, indicating that simple, off-the-shelf language processing pipelines can demonstrate good performance in low-resource and morphologically complex linguistic settings.
[NLP-28] Agent ic Reinforcement Learning with Observation-Calibrated Self-Distillation
【速读】： 该论文旨在解决大语言模型智能体在基于稀疏轨迹级奖励的强化学习训练中，因缺乏对单个词元（token）更新强度的精细指导而导致优化效率受限的问题。现有方法On-Policy Self-Distillation（OPSD）通过引入特权重放视图（privileged replay view）对生成词元进行重新评分，以获得密集的词元级监督信号，但其存在混淆问题：所生成的支持信号可能同时受特权信息与重放框架本身引起的评分偏移影响，难以准确归因于真正的环境反馈。尤其当未来环境观测作为特权信息时，重放需构建扩展的重放骨架，该骨架会进一步扰动词元评分，加剧混淆。为解决此问题，本文提出观察校准自蒸馏（Observation-Calibrated Self-Distillation, OCSD），通过对比结构匹配的两个重放视图——完整视图（Full）与仅缺失实际未来观测的消融视图（Observation-Ablated），提取仅由未来观测引起的评分残差（observation residual），从而消除重放骨架带来的共性干扰。该残差被用于在高不确定性步骤中调制词元级的GRPO（Generalized Reward Policy Optimization）更新，同时保持轨迹级更新方向不变。实验在ALFWorld、WebShop和Search-QA三个任务上，针对三个不同规模的Qwen3模型均验证了OCSD显著优于强基线。诊断分析进一步表明，校准后的残差更贴近局部环境反馈，具备更强的可解释性。
链接: https://arxiv.org/abs/2608.04788
作者: Yi Yang,Cong Qin,Xiaodan Liu,Chishui Chen,Qing Dong,Yan Zhang,Cao Liu,Zhao Yang,Lu Pan,Jiaye Lin,Yi Feng
机构: 未知
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large language model agents are commonly trained through reinforcement learning with sparse trajectory-level rewards, which offer limited guidance on how strongly individual tokens should be updated. On-Policy Self-Distillation (OPSD) addresses this by re-scoring generated tokens under a privileged replay view to obtain dense, token-level supervision. However, we identify a confounding issue: the resulting support may reflect both the privileged information contained in the replay view and score shifts induced by the replay scaffold, making it difficult to attribute the support specifically to that information. This issue is especially pronounced when future environment observations serve as privileged information, since replaying them requires reconstructing an extended scaffold that itself perturbs token scores. To resolve this confounding, we propose Observation-Calibrated Self-Distillation (OCSD), which contrasts two structurally matched replay views, Full and Observation-Ablated, differing only in whether the actual future observation is present, to derive an observation residual that discounts score changes shared by the replay scaffold. OCSD then applies this residual to modulate token-level GRPO updates at high-uncertainty steps, while preserving the trajectory-level update direction. Experiments on ALFWorld, WebShop, and Search-QA across three Qwen3 model scales show that OCSD consistently outperforms strong baselines. Diagnostic analyses further confirm that the calibrated residual aligns better with local environment feedback. Our code is publicly available at this https URL.
[NLP-29] Reachability in 3-VAS
【速读】： 该论文旨在解决低维（状态无关）向量加法系统（Vector Addition Systems, VAS）中可达性问题的精确复杂度。在维度2至4的情况下，此前仅知其复杂度被限定在NP与PSPACE之间。研究的关键突破在于证明了对称向量加法系统（Symmetric Vector Addition Systems, SVAS）在维度3（3-VAS）下可达性问题为PSPACE-hard，这是通用3-VAS的一个受限子类。结合已有的PSPACE上界结果，该工作最终确定了3-VAS、4-VAS及其对称片段中可达性问题的复杂度为PSPACE完全（PSPACE-complete），从而彻底解决了该问题在低维情形下的复杂性定位。
链接: https://arxiv.org/abs/2608.04786
作者: Łukasz Kamiński,Sławomir Lasota
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:We settle the exact complexity of the reachability problem in (stateless) vector addition systems (VAS) in fixed low dimension. In dimensions 2-4 it has only been known to be sandwiched between NP and PSPACE. We prove PSPACE-hardness of the reachability problem for symmetric vector addition systems in dimension 3 (3-VAS), a restricted fragment of general 3-VAS. Combined with previously established PSPACE upper bounds, our result settles the complexity of the problem to be PSPACE-complete in 3-VAS and 4-VAS, as well as in their symmetric fragments.
[NLP-30] Guideline-as-Oracle: Zero-Annotation Training of an Ophthalmic Telephone Triage Agent
【速读】： 该论文旨在解决多轮医疗对话智能体在训练过程中监督信号扩展困难的问题，核心挑战在于专家对话标注成本高昂且临床对话涉及隐私限制。其解决方案的关键是提出“指南即裁判”（Guideline-as-Oracle, GAO）框架，将美国眼科学会（American Academy of Ophthalmology）的临床指南转化为70行可操作规则表，并作为3,000条训练对话的唯一实例级监督来源，仅在评估阶段引入人工标注。为实现规则到对话的有效映射，研究系统性地归纳了八种构建策略，包括引用行层级分配、单事实边界对、仅基于元数据修复和标签修复等，并对每种策略的证据效力进行分类：标注机制、空值、混淆或仅以整体包形式评估。在90亿参数骨干模型上微调后得到GAO-Triage，其与201例实际操作参考的一致性从61.7%提升至74.1%（精确麦内马尔检验p=0.0046），突发病例召回率由9.5%跃升至69.0%，且在不同随机种子和患者模拟器下均保持性能优势。对比七种通用模型，无一在两项指标上全面超越GAO-Triage，且该模型推理时无需前沿大模型支持。置换标签-对话对应关系导致模型退化为恒定流程预测器，表明有效信号源于指南驱动的标签分配而非对话表面形式；而标签修复策略的引入与训练后期出现的安全性下降现象消失相吻合，进一步验证了该策略的合理性。
链接: https://arxiv.org/abs/2608.04772
作者: Chenyu Wang,Yi Liu,Baoqing Li,Min Tu,Diping Song
机构: Google(谷歌); Stanford University (斯坦福大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Scaling supervision for multi-turn medical agents is difficult because expert dialogue annotation is costly and clinical conversations are privacy-restricted. We introduce Guideline-as-Oracle (GAO), which compiles American Academy of Ophthalmology guidance into a 70-row operational rule table and uses it as the sole source of instance-level supervision for 3,000 training dialogues, reserving human labeling for evaluation. Because converting rules into dialogues is itself a design problem, we catalog eight construction strategies, including cited-row tier assignment, one-fact boundary pairs, metadata-only repair, and label repair, and characterize the evidential status of each: labeling mechanism, null, confounded, or evaluated only as a package. Fine-tuning a 9B backbone on this corpus yields GAO-Triage, improving agreement with a 201-case operational reference from 61.7% to 74.1% (exact McNemar p=0.0046) and emergent-case recall from 9.5% to 69.0%; the gains persist across a second seed and patient simulator. None of the seven general-purpose systems we test dominates GAO-Triage on both metrics, and GAO-Triage requires no frontier model at inference time. Permuting label-dialogue assignments collapses the model to a constant-routine predictor, indicating that the signal lies in guideline-derived assignment rather than dialogue surface form. Label repair coincides with the disappearance of a late-training safety degradation.
[NLP-31] InsightEmb: Learning Action-Intent Embeddings for Agent ic Insight Retrieval
【速读】： 该论文旨在解决自改进智能体（self-improving agents）在决策过程中如何高效检索可复用的洞察（insight）以突破当前决策瓶颈的问题，即“代理洞察检索”（agentic insight retrieval）问题。现有方法主要依赖语义相似性进行检索，但忽视了所检索的洞察是否真正解决了当前的决策困境。本文提出 InsightEmb，一种基于对比学习的嵌入框架，仅使用数学推理数据即可学习可迁移的、以进展为导向的检索几何结构。其核心创新在于联合建模具体情境与抽象启发式规则之间的对齐关系，以及具有相似进展结构的推理轨迹聚类。实验表明，InsightEmb 在动态代理任务和静态技能检索基准上均无需环境特定训练即可显著超越现有推理嵌入模型，证明了状态-洞察匹配的几何结构具备跨领域迁移能力，从而实现仅依赖公开可用的推理数据进行有效训练，避免了昂贵的环境特异性监督。
链接: https://arxiv.org/abs/2608.04761
作者: Tsz Ting Chung,Jiangnan Li,Jie Zhou,Mo Yu
机构: The Hong Kong University of Science and Technology (香港科技大学); WeChat AI, Tencent (微信AI，腾讯)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Self-improving agents accumulate reusable insights from prior trajectories, making retrieval increasingly important for turning accumulated experience into actionable guidance. At each decision step, retrieving the right insight can help the agent progress toward its goal, a setting we refer to as agentic insight retrieval. However, existing retrieval methods primarily model semantic similarity, while overlooking whether a retrieved insight resolves the agent’s current decision bottleneck. We propose InsightEmb, a contrastive embedding framework that learns transferable progress-oriented retrieval geometry using only mathematical reasoning data. InsightEmb jointly learns to align concrete situations with abstract heuristic rules and to cluster reasoning trajectories with similar progress structures. We evaluate InsightEmb on dynamic agent tasks and a static skill-retrieval benchmark. Without any environment-specific training, InsightEmb improves over all these evaluations, surpassing the performance of existing reasoning embedding models. These results suggest that the geometry of state-insight matching can transfer across domains, enabling effective training from publicly available reasoning data without expensive environment-specific supervision.
[NLP-32] race Verify and Correct: A Training-Free Framework for Spatial Reasoning in Multimodal LLM s
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）在空间推理过程中存在的不忠实问题，即模型生成的中间推理步骤与输入图像内容不一致，导致错误在推理链中传播并影响最终答案的准确性。现有方法主要通过训练优化或引入额外的空间信息来提升性能，但未关注推理过程本身是否忠实于原始输入。本文的关键创新在于提出一种模块化且无需训练的空间推理验证与修正框架，其核心是构建空间证据图（Spatial Evidence Graph, SEG），将思维链（Chain-of-Thought）中提取的原子级空间证据与视觉实体、空间关系、来源步骤及视觉证据进行关联。通过空间证据可靠性评估（SERA），基于物体存在性、定位准确性和几何度量对视觉证据的可靠性进行量化。随后，框架识别出最早与可靠视觉证据矛盾的空间证据单元，并引导原MLLM修正后续推理路径及最终答案。实验表明，在15个模型-数据集组合上，该方法平均准确率达68.94%，显著优于基线方法平均8.55个百分点，验证了其在提升推理忠实性方面的有效性。
链接: https://arxiv.org/abs/2608.04759
作者: Yang Yang,Jiawei Chen,Tairan Chen,Zhaoxia Yin
机构: East China Normal University (华东师范大学); Zhongguancun Academy (中关村学院); Stevens Institute of Technology (史蒂文斯理工学院)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 19 pages, 7 figures
Abstract:Although Multimodal Large Language Models (MLLMs) have made substantial progress, their spatial reasoning may still produce intermediate judgments inconsistent with the input image, allowing errors to propagate through the reasoning chain and affect the final answer. Existing methods mainly improve spatial reasoning through training or additional spatial information, without considering whether the reasoning process itself is faithful to the model input. Our study shows that unfaithful reasoning chains significantly reduce final-answer accuracy. To address this issue, we propose a modular and training-free framework for spatial reasoning verification and correction. The framework constructs a Spatial Evidence Graph (SEG), which associates atomic spatial evidence extracted from Chain-of-Thought reasoning with visual entities, spatial relations, source steps, and visual evidence. Spatial Evidence Reliability Assessment (SERA) evaluates the reliability of visual evidence based on object existence, localization, and geometric measurements. The framework then identifies the earliest spatial evidence unit contradicted by reliable visual evidence and guides the original MLLM to revise the subsequent reasoning and final answer. Across 15 model-dataset settings, our method achieves an average accuracy of 68.94%, outperforming the compared baselines by 8.55 percentage points on average. Our code will be open-sourced.
[NLP-33] Simile Understanding in Text-to-Image Models: An Evaluation Framework
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在文本到图像模型（text-to-image models, t2i models）中对隐喻性语言理解不足的问题，特别是模型在处理比喻句（simile）时，常将隐喻中的“喻体”（metaphorical vehicle）误当作“本体”（target object）进行视觉表征，导致生成结果出现系统性偏差。其解决方案的关键在于提出一个可扩展的隐喻理解评估框架，该框架包含三个核心部分：（1）构建一个受控的比喻数据集，其中喻体限定于可检测的对象类别，并与多样化的句式模板组合；（2）基于YOLO（You Only Look Once）目标检测的自动化视觉定位度量方法，量化模型对喻体的真实视觉锚定程度；（3）利用Diffusion Lens分析文本编码器层的特征演化，追踪喻体在生成过程中的语义浮现路径。实验结果表明，不同架构的t2i模型均表现出一致的“字面化”（literalization）失败模式，揭示了当前模型在跨模态语义映射中对修辞结构建模的薄弱性。该框架为深入诊断和改进生成式模型对修辞语言的视觉接地能力提供了可复现的技术路径。
链接: https://arxiv.org/abs/2608.04750
作者: Luecheng Wang,Shintaro Ozaki,Hidetaka Kamigaito,Katsuhiko Hayashi,Jingun Kwon,Manabu Okumura,Taro Watanabe
机构: The University of Tokyo(东京大学); Nara Institute of Science and Technology(奈良先端科学技術大学院大学); Chungnam National University(忠南国立大学); Institute of Science Tokyo(东京科学研究所)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Multimedia (cs.MM)
备注: Accepted as a full paper at ACM Multimedia 2026
Abstract:Similes provide a compact and expressive way to describe visual characteristics in text prompts. Recent text-to-image models (t2i models) can produce visually compelling outputs from simile prompts, yet even frontier models frequently misinterpret the metaphorical vehicle and confuse it with the object. These systematic failures reveal a gap between figurative language and object-level visual grounding in t2i models. To investigate this issue, we propose a scalable evaluation framework for simile understanding. Our framework includes (1) a controlled simile dataset in which metaphorical vehicles are drawn from a predefined set of object-detectable categories and combined with diverse templates, (2) automatic grounding metrics based on YOLO (You Only Look Once) detection, and (3) text encoder layer analysis using Diffusion Lens to track how metaphorical vehicles emerge during generation. Experiments across architecturally diverse t2i models reveal consistent literalization failure patterns. We further discuss potential mitigation strategies for improving simile grounding in t2i models.
[NLP-34] EmpaAva: An Open-source Agent ic 3D-Avatar Empathetic Live Chatbot
【速读】： 该论文旨在解决现有对话系统在情感交互中缺乏真实、具身化表现的问题，即如何将文本驱动的共情响应生成（ERG）从静态、单模态的文本交流拓展至具备实时面部表情、语音同步与视觉逼真度的三维数字人（3D-avatar）面对面互动。其核心解决方案在于构建一个由大语言模型（LLM）协调的三代理架构（Tri-Agent Architecture），通过感知（Perception）、共情响应规划（Empathetic Response Planning）与具身渲染（Embodied Rendering）形成闭环协同，并引入响应规划层（Response Planning Layer）将每条回复整合为可执行的多模态指令，确保语音、表情与渲染始终统一于同一共情意图。该框架基于开源模块实现高度可控性与可解释性，在自动与人工评估中均显著优于仅文本、二维虚拟人脸及传统多模态基线模型，在情绪理解、响应质量与视听一致性方面展现出优越性能。
链接: https://arxiv.org/abs/2608.04709
作者: Jie Yang,Wenhao Xu,Shuhui Lin,Hao Fei
机构: National University of Singapore (新加坡国立大学); Tsinghua University (清华大学); University of Oxford (牛津大学)
类目: Computation and Language (cs.CL)
备注: ProjectDemo: this https URL
Abstract:This paper presents EmpaAva, to our knowledge the first open-source, agentic 3D-avatar empathetic chatbot, which carries empathetic response generation (ERG) from text-only exchanges into live, face-to-face interaction. Through a video-call-like interface, a user speaks to a 3D digital human that reads their affect from speech and optional vision, and replies with emotional speech, lip-synced facial motion, and photorealistic 3D Gaussian rendering. At its core, an LLM coordinates a Tri-Agent Architecture, in which perception, empathetic response planning, and embodied rendering form a closed loop, paired with a Response Planning layer that compiles each reply into an executable multimodal plan, keeping voice, expression, and rendering on one empathetic intent. Building on strong open-source modules, EmpaAva supplies the intelligence that binds them into one controllable, inspectable experience. In automatic and human evaluations, EmpaAva surpasses text-only, 2D talking-face, and multimodal avatar baselines in emotion understanding, response quality, and audio-visual consistency. We open-source EmpaAva with an online live demo.
[NLP-35] IslamicTurathBench: A Multi-Task Multi-Discipline Benchmark for Evaluating Large Language Models on the Islamic Scholarly Tradition (turath)
【速读】： 该论文旨在解决生成式 AI（Generative AI）在古典伊斯兰学术研究领域中缺乏高质量评估资源的问题，尤其是在依赖专门文献传统（turath）的宗教与文化语境下，现有语言模型对复杂概念、方法及学术争论的理解能力难以被有效衡量。其解决方案的关键在于构建IslamicTurathBench（ISTB），一个面向古典伊斯兰学术的多任务、跨学科评测数据集，包含3,465个问答样本，源自35部跨越1200余年历史的权威典籍，覆盖伊斯兰研究七大核心领域。为全面评估模型性能，ISTB从两个维度进行结构化设计：学术需求层次（初阶、中阶、高阶）与任务类型（选择题、篇章理解、开放性知识问答）。该数据集由领域专家开发并审核，提供学者参考基准评分与十种系统的零样本基线表现，支持在不同文献来源、学科方向、学术难度及题型上实现可复现的语言模型行为分析，从而推动生成式 AI 在历史层积深厚的学术语境中的可信评估。
链接: https://arxiv.org/abs/2608.04703
作者: Shahd Gaben,Heba Sbahi,Samer Rashwani,Abdessalam Bouchekif,Mutaz Al-Khatib,Emad Mohamed,Somaya Eltanbouly,Mohammed Ghaly
机构: 未知
类目: Computation and Language (cs.CL)
备注: Includes supplementary materials. Submitted to the Journal of Scientific Data. Data and code are publicly available
Abstract:Large language models (LLMs) are increasingly used for question answering, education, and research, including in religious and cultural domains where answers depend on specialised source traditions. Yet in Islamic Studies, key concepts, methods, and debates preserved in the authoritative scholarly tradition, known as turath, lack high-quality annotated resources. We introduce IslamicTurathBench (ISTB), a multi-task, multi-discipline dataset for evaluating LLMs on classical Islamic scholarship. Developed and reviewed by domain experts, ISTB contains 3,465 question-answer items drawn from 35 recognised source works spanning more than 12 centuries of scholarship across seven key fields of Islamic Studies. To enable comprehensive profiling of model capabilities, ISTB is structured along two axes: scholarly demand (Beginner, Intermediate, and Advanced) and task format (multiple-choice questions, passage-based comprehension, and open-ended knowledge questions). ISTB includes aggregated scores from a scholarly human reference panel and zero-shot baselines from ten systems. The dataset supports reproducible evaluation of language-model behaviour across source works, disciplines, scholarly-demand levels, and question formats in a historically layered scholarly domain.
[NLP-36] Kathleen Writes: Autoregressive Generation and Data Scaling Without Attention
【速读】： 该论文旨在解决生成式模型在无需预训练、参数量极低（约50万）的情况下，能否实现与主流架构（如Transformer）相当甚至更优的文本生成能力这一关键问题。其核心挑战在于：如何在不依赖注意力机制和大规模预训练的前提下，构建一个高效且具备真实文本生成能力的字节级语言模型。解决方案的关键在于提出一种“波动式状态”（reverberant state）架构，结合波表编码器（wavetable encoder）与多尺度动态状态演化机制，实现了对语言上下文的非注意力式建模。实验表明，该架构在WikiText-103数据集上，在不同数据规模（2–512 MB）下均优于参数匹配的Transformer，尤其在小样本场景下优势显著。为评估生成内容的真实性，作者引入了非参数化、抗游戏化的FORM DISTANCE度量方法，通过九个统计维度构成人类文本参考云，成功识别出五种人工构造的伪文本。此外，研究发现解码策略对生成质量影响远大于架构本身——通过扩大采样器宽度或采用检索增强解码方案，可在不进行任何训练的情况下显著降低生成文本与真实文本的距离（从3.17降至1.14），且收益归因于模型自身训练语料中稀疏短语的注入，而非选择门控机制。值得注意的是，使用外部大规模语料库无法提升性能，表明此类能力本质上依赖于模型自身的上下文整合能力，符合规模效应规律。最后，尽管多种架构改进未见成效，但提出的计算词典仅用五分之一参数即可达到94%的原始学习词表精度，进一步验证了该框架的高效性。所有实验均在离线环境下完成，可复现于免费Kaggle T4环境。
链接: https://arxiv.org/abs/2608.04678
作者: George Fountzoulas
机构: Frederick University (弗雷德里克大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Paper 3 of the Kathleen series. 11 pages, 3 figures. All experiments reproducible on a free Kaggle T4
Abstract:Papers 1-2 of the Kathleen series showed that a byte-level, attention-free architecture built from a wavetable encoder and multi-scale reverberant state can match strong baselines on classification at ~450-700K parameters, without pretraining. We ask whether the same ingredients can generate. (1) Scaling: on byte-level language modeling (WikiText-103, raw UTF-8, no tokenizer), the reverberant model beats a parameter-matched transformer at every dataset scale measured (2-512 MB), e.g. 1.84 vs 2.04 bits/byte at 512 MB with ~0.5M parameters; the transformer needs more than 512 MB to match what the attention-free model learns from 32 MB. (2) Measurement: we introduce FORM DISTANCE, a non-parametric, gaming-resistant instrument for “reads like text”: nine statistical axes of human text define a reference cloud, and five constructed fakes are all rejected. (3) Generation: decoding policy dominates architecture – widening the sampler halves the same model’s distance (3.17 to 1.52), and a retrieval-augmented decoding scheme takes the frozen model further (1.52 to 1.14) with no training step involved; the ablation attributes the gain to the sparse phrase dose itself, not the selection gate. The gain has a sharp boundary condition: the phrases must come from the model’s own training corpus – a 40x larger foreign library helps not at all, an effect the attention twin shares, consistent with in-context integration being a capability of scale. We also report four architectural additions that did not help, and a computed lexicon reaching 94% of a learned table’s top-1 accuracy at one fifth of the parameters. Everything runs offline; all experiments are reproducible on a free Kaggle T4.
[NLP-37] Easy to Complete Hard to Choose: Investigating LLM Performance on the ProverbIT Benchmark
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在理解文化嵌入性语言表达，特别是意大利语谚语方面存在的认知局限问题。现有模型虽在常规语言任务中表现优异，但在处理蕴含深层文化语义的习语类表达时仍存在显著能力缺陷。其解决方案的关键在于构建并应用ProverbIT——一个包含100道多选题的意大利语谚语评估基准，用于系统检验模型在谚语补全、有正确选项的多选以及无正确选项的多选三种任务中的表现。研究发现，尽管大多数模型在谚语补全任务中表现出较强的知识记忆能力，但在缺乏正确选项的多选任务中性能急剧下降，即便是最先进的推理型大模型（Large Reasoning Models, LRMs）也出现显著退化。通过链式思维（Chain-of-Thought）分析进一步揭示，模型存在强烈的字面同义词偏好倾向，常在推理过程中提及正确谚语结尾却无法从选项中准确识别，表明当前模型主要依赖于对固定表达模式的机械记忆，而非对文化语境下隐喻性语言的深层语义理解，暴露出其在拟人化、象征性语言推理方面的根本性局限。
链接: https://arxiv.org/abs/2608.04670
作者: Enrico Mensa,Lorenzo Zane,Calogero Jerik Scozzaro,Matteo Delsanto,Tommaso Milani,Daniele Paolo Radicioni
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) have transformed computational linguistics and achieved remarkable performance across numerous natural language processing tasks, yet significant gaps persist in understanding how these systems process culturally embedded linguistic expressions. This paper introduces ProverbIT, a novel Italian benchmark comprising 100 multiple-choice questions designed to evaluate LLMs’ ability to complete Italian proverbs. We assess 13 frontier models, including Large Reasoning Models (LRMs) and traditional LLMs, across three tasks: proverb completion, multiple-choice selection with correct answers, and multiple-choice selection without correct answers. Our evaluation reveals surprising results: while nearly all models demonstrate knowledge of the proverbs through successful completion tasks, performance drops dramatically when transitioning to multiple-choice formats without correct answers, with even state-of-the-art reasoning models showing substantial degradation. Through detailed Chain-of-Thought analysis of two LRMs, we uncover that models exhibit a strong bias toward selecting literal synonyms and frequently mention correct proverb endings during reasoning without successfully identifying their absence from the given options. These findings suggest that current LLMs rely heavily on memorized patterns rather than deeper semantic understanding of culturally grounded expressions, highlighting important limitations in their reasoning capabilities for figurative language comprehension.
[NLP-38] Evaluating Theory of Mind in Reasoning Models: Robustness over Reasoning
【速读】： 该论文旨在探讨大语言模型（Large Language Models, LLMs）在心理理论（Theory of Mind, ToM）任务中表现出的性能是否真正反映了其具备深层次的心理理解能力，还是仅源于模型对输入变化的鲁棒性增强。研究通过改编机器心理学实验范式并结合已有基准测试，系统评估了基于强化学习训练、具备推理能力的模型在各类ToM任务中的表现。研究发现，这类推理模型在面对提示（prompt）变化和任务扰动时展现出显著更高的鲁棒性，且其性能提升主要归因于在不同输入条件下更稳定地得出正确答案的能力。因此，论文认为当前观察到的性能改善更支持一种基于鲁棒性的解释，而非模型具备了全新的、专属于心理理论的认知能力。解决方案的关键在于：通过强化学习与可验证奖励机制训练的推理模型，能够更稳健地应对输入多样性，从而在不依赖特定心理建模的前提下实现优异的ToM任务表现。
链接: https://arxiv.org/abs/2608.04646
作者: Ian B. de Haan,Peter van der Putten,Max van Duijn
机构: 未知
类目: Computation and Language (cs.CL)
备注: Accepted for 29th International Conference on Discovery Science, October 5-9, 2026, Mainz, Germany
Abstract:Large language models (LLMs) have recently shown strong performance on Theory of Mind (ToM) tests, prompting debate about the nature and validity of the underlying capabilities. At the same time, reasoning-oriented LLMs trained via reinforcement learning with verifiable rewards have demonstrated notable improvements across a range of benchmarks. In this work, we examine the behavior of such reasoning models in ToM tasks using novel adaptations of machine psychological experiments together with results from established benchmarks. We observe that reasoning models consistently exhibit increased robustness to prompt variations and task perturbations. Our analysis suggests these gains come at least partly from models being more robust at reaching the correct answer under prompt and task variation. We read this as evidence for a robustness-based account rather than for a new ToM-specific ability.
[NLP-39] When Absence Is Evidence: Evaluating Completeness-Sensitive Negative Reasoning in Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在处理“某事物是否缺失于记录、列表或检索上下文”这类问题时所表现出的不完整性敏感型否定推理（completeness-sensitive negative reasoning）缺陷。核心问题是：当证据覆盖范围不完整时，模型仍倾向于给出确定性否定回答，而非承认知识不足，导致过度闭合（over-closure）现象严重，无法可靠区分“有充分证据支持的否定”（Certified-Negative）与“证据不足”的未知状态（Unknown）。解决方案的关键在于构建一个具有控制变量的评估基准——CROWN-QA，其包含两个部分：CROWN-Synth，通过严格控制问题和已知事实，仅改变查询相关证据覆盖率，以揭示模型对证据完整性的感知偏差；以及CROWN-Real，基于真实文档的对比集评估，验证上述偏差在实际场景中的持续存在。研究发现，模型普遍存在对部分证据的误判，即错误地将不完全覆盖视为全面覆盖，且这种不对称性在不同模型、提示（prompt）及数据源间表现各异。此外，提示工程虽可调整过闭合与欠闭合之间的误差分布，但无法根本解决认知偏差。结构化证书提取进一步揭示，多数错误源于对证据覆盖率的误判，表明当前模型缺乏对“证据完整性”本身的准确建模能力。
链接: https://arxiv.org/abs/2608.04591
作者: Byoungjae Min,Kennedy Edemacu,Sae-Hong Cho,Yoonhyuk Choi,Beakcheol Jang,Jong Wook Kim
机构: Korea University(韩国大学); Samsung Electronics(三星电子); Seoul National University(首尔国立大学); Korea Institute of Science and Technology(韩国科学技术院); KAIST(韩国科学技术院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 19 pages, 2 figures, 20 tables
Abstract:Large language models (LLMs) are often asked whether something is absent from a record, list, or retrieved context. Yet non-observation licenses a negative answer only when evidence completely covers the query scope; otherwise, the answer should remain unknown. We call this completeness-sensitive negative reasoning. We introduce CROWN-QA, comprising CROWN-Synth, a controlled paired core that fixes the question and observed facts while varying only query-relative coverage, and CROWN-Real, a real-document contrast-set evaluation with controlled coverage variants. Across three LLM families, models show unstable closure judgments and substantial over-closure, failing to reliably distinguish a justified negative answer (Certified-Negative) from insufficient evidence (Unknown). The dominant CROWN-Synth failure is asymmetric: models often recognize implicitly complete evidence yet treat implicitly partial evidence as query-covering. Prompting redistributes errors between over- and under-closure rather than consistently resolving them. Structured certificate elicitation traces many errors to evidence-coverage mischaracterization. CROWN-Real shows that the core partial-coverage asymmetry persists on real-document content, while its strength and the balance between over- and under-closure vary by model, prompt, and source.
[NLP-40] EASy: Towards Efficient LLM -Based Agent ic System
【速读】： 该论文旨在解决当前自主代理系统（agentic systems）在复杂任务求解中普遍存在的效率与性能失衡问题，尤其关注实际应用中执行器能力限制与计算成本约束下的执行效率。现有基于路由（router-based）的方法难以对动态演化的任务上下文、多步依赖关系及中间执行反馈进行有效推理，且在面对未见过的执行器时泛化能力差。其解决方案的关键在于提出EASy——一个可训练的自主代理框架，通过强化学习联合优化任务成功率与计算效率。EASy赋予基于大语言模型（LLM）的协调器对异构执行器的能力与成本特性具备显式知识，实现超越单纯性能导向的上下文感知协调。此外，引入“里程碑-计划-执行”（milestone-plan-act）工作流，将复杂任务分解为可管理的里程碑，构建依赖感知的执行图，合理分配执行器，并并行化独立步骤，同时根据中间结果自适应调整后续决策。为训练协调器，设计了树状展开（tree-structured rollout）机制以探索多种里程碑分解与执行方案，并采用多组件奖励函数，综合考量任务正确性、执行效率与轨迹完整性。大量实验表明，EASy在数学推理、具身决策和深度研究等基准任务上均显著优于现有强基线，在性能与效率之间实现了更优的权衡。
链接: https://arxiv.org/abs/2608.04588
作者: Junnan Liu,Linhao Luo,Thuy-Trang Vu,Gholamreza Haffari
机构: Monash University (莫纳什大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Preprint
Abstract:Agentic systems have emerged as a promising paradigm for solving complex tasks by coordinating specialized LLM-based agents. However, most existing systems primarily optimize task success while giving limited consideration to execution efficiency under practical constraints such as executor capability and computational cost. Existing router-based methods have limited ability to reason over rich, evolving task contexts, multi-step dependencies, and intermediate execution feedback, and often generalize poorly to unseen executors. We propose EASy, a trainable agentic framework that jointly optimizes task performance and computational efficiency through reinforcement learning. EASy equips an LLM-based orchestrator with explicit knowledge of the capability and cost profiles of heterogeneous executors, enabling context-sensitive coordination beyond performance-only routing. It further introduces a milestone-plan-act workflow that decomposes complex tasks into manageable milestones, constructs dependency-aware execution graphs, assigns suitable executors, and parallelizes independent steps while adapting subsequent decisions to intermediate outcomes. To train the orchestrator, we develop a tree-structured rollout procedure that explores alternative milestone decompositions and execution plans, together with multi-component rewards that capture task correctness, execution efficiency, and trajectory completeness. Extensive experiments on mathematical reasoning, embodied decision-making, and deep research benchmarks show that EASy consistently achieves stronger performance-efficiency trade-offs than strong agentic baselines.
[NLP-41] Breaking the Curse ofMultilinguality inMany-to-Many Speech-to-Text Translation via a Resource-AwareMixture of Speech Encoders
【速读】： 该论文旨在解决多语言语音到文本翻译（Multilingual Speech-to-Text Translation, S2TT）中因共享单一语音编码器而导致的“多语言诅咒”（curse of multilinguality）问题：不同资源水平的语言在有限的表示容量下相互竞争，致使高资源语言表现优异而低资源语言性能显著下降。其解决方案的关键在于提出一种基于资源感知的语音编码器混合模型（Resource-aware Mixture of Speech Encoders, MoSE），通过显式的语言路由机制将每段语音输入分配给最合适的专家编码器——一个冻结的专家编码器保留对高资源语言的强表征能力，而一个可训练的专家编码器则专门适应并优化中低资源语言。此外，引入五阶段课程学习策略，大幅降低对标注数据的需求，仅需每语言10小时的配对数据即可实现有效对齐。实验在45种语言上系统评估了全部45×44个翻译方向，结果表明，该方法在40亿参数规模下超越了更大规模基线模型，在高、中、低资源语言上均取得提升，尤其在低资源语言上增益显著，成功打破多语言诅咒且不牺牲高资源语言性能。
链接: https://arxiv.org/abs/2608.04586
作者: Yexing Du,Kaiyuan Liu,Youcheng Pan,Bo Yang,Chengpeng Fu,Yu Wang,Ming Liu
机构: University of Science and Technology of China (中国科学技术大学); Alibaba Group (阿里巴巴集团)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Multimodal large language models (MLLMs) have achieved significant success in speech-to-text translation (S2TT). However, when processing multilingual speech inputs, a single speech encoder shared across all languages suffers from the curse of multilinguality: languages at different resource levels compete for limited representation capacity, leading to strong high-resource performance but substantial degradation on low-resource speech. To address this problem and improve multilingual consistency, we propose MSRT, a novel framework built around a resource-aware Mixture of Speech Encoders (MoSE). MoSE uses an explicit language router to assign each utterance to an appropriate expert encoder. A frozen expert preserves high-resource language capabilities, while a trainable expert adapts to and specializes in medium- and low-resource languages. We further introduce a five-stage curriculum learning strategy that substantially reduces data dependence, requiring only 10 hours of paired S2TT data per language for effective alignment. We conduct extensive experiments on 45 languages, systematically evaluating all 45 \times 44 translation directions. Our 4B-parameter model achieves state-of-the-art performance, outperforming substantially larger baselines. Empirical analyses show that MoSE improves high-, medium-, and low-resource languages simultaneously, with the largest gains on low-resource speech, thereby breaking the curse of multilinguality without compromising high-resource performance. To support future multilingual S2TT research, we release our code and models.
[NLP-42] Causal Evidence Extraction and Triangulation in Crisis Reports using Large Language Models : A ReliefWeb-based Study
【速读】： 该论文旨在解决人道主义报告中因果证据难以整合的问题，因其文本冗长、噪声大且主题多元，导致决策支持所需的关键因果信息提取困难。其核心解决方案是构建一个两阶段大语言模型（LLM）流水线，通过查询条件化提取机制限定输出至特定干预类别，有效抑制了检索引发的过度提取问题；同时采用片段锚定（snippet grounding）技术，将每个干预-结果关系与其支撑文本关联，保障可审计性与分类准确性。在100份专家标注报告的数据集上，最佳闭源模型达到90.73%的加权F1分数，展现出优异的成本效益；而经监督微调的Llama-3.1-8B模型进一步提升至94.15%的加权F1分数。此外，论文提出一种保持上下文一致性的三角验证方法，通过在灾害类型×信息来源的单元格内聚合强度加权证据，结合拉普拉斯平滑并等权处理各单元格，以层级证据得分（Level-of-Evidence score, LoE）量化跨情境一致性。应用于现金援助分析时，食物相关结果表现出强正向收敛性（LoE=0.865）及稳定的长期轨迹，验证了该方法在实证推断中的有效性。
链接: https://arxiv.org/abs/2608.04576
作者: Yuanjun Zhang,Mourad Oussalah
机构: University of Oulu, Finland; LUT University, Finland
类目: Computation and Language (cs.CL)
备注:
Abstract:Humanitarian reports are long, noisy, and multi-topic, making it difficult to consolidate decision-relevant causal evidence. We present a ReliefWeb study (2000-2024) and a two-stage Large Language Model (LLM) pipeline that extracts structured intervention-outcome records with direction and strength attributes. Query-conditioned extraction restricts output to a specified intervention class, reducing retrieval-induced over-extraction, while snippet grounding links each relation to supporting text for auditability and classification. In an expert-annotated dataset of 100 reports, the best closed-source LLM achieved a weighted F1 score of 90.73% with strong cost-efficiency, while Llama-3.1-8B with supervised fine-tuning reached 94.15% weighted F1 score. We further propose context-preserving triangulation that aggregates strength-weighted evidence within disaster \times source cells, applies Laplace smoothing and equally weights cells to quantify cross-context convergence via a Level-of-Evidence score. Applied to cash assistance, food-related outcomes show strong positive convergence (LoE=0.865) and stable long-horizon trajectories.
[NLP-43] When Memory Lies: An Empirical Study of Spatial Memory Staleness in VLM Agents
【速读】： 该论文旨在解决记忆增强型视觉语言模型（VLM）代理在动态环境中因空间记忆过时而导致的安全性问题，尤其关注当代理的自信记忆判断与实时观测发生矛盾时，现有模型能否在错误发生前识别并纠正这一冲突。其核心挑战在于：尽管模型可能在文本输入下表现出对记忆过时的正确检测能力，但这种能力并不等同于可靠的视觉感知一致性（visual grounding），即模型在图像输入下仍可能无视真实视觉信息做出错误决策。解决方案的关键在于引入“读取时审计”（read-time audit）机制，通过透明的过滤策略在推理阶段对记忆进行验证，从而降低因依赖过时记忆而引发的安全风险。然而研究发现，即使采用理想化的“全知”标签（oracle labels）进行审计，也无法显著提升性能，且在视觉审计不可靠时，过滤机制亦无法带来稳定收益，这表明当前模型在记忆与观测冲突下的可靠视觉对齐和动作选择仍是亟待突破的核心难题。
链接: https://arxiv.org/abs/2608.04574
作者: Yushi Sun,Yanjie Zhang
机构: Tencent LIGHTSPEED(腾讯光速)
类目: Computation and Language (cs.CL)
备注:
Abstract:Memory-augmented VLM agents act on persistent spatial knowledge, yet that knowledge silently goes stale as the environment changes. We ask what happens when an agent must reconcile a confident memory claim with a contradicting observation, and whether current models can catch the conflict before it becomes a safety-relevant mistake. Using a dynamic FrozenLake testbed, we pair a staleness-detection task with a downstream navigation task across three closed-source models and three open-weight VLMs under both text and image inputs (1,800 detection runs, and 12,000 text-mode navigation episodes over four LLM navigators at a shared 50-seed scale). Three findings emerge. First, text solvability does not imply visual grounding: models that flag stale entries reliably from text nonetheless span vision F1 from 0.887 down to 0.067 on the identical grids, and the weakest keeps making fluent, confident decisions that ignore the image. Second, consuming stale memory without an audit is a safety liability: in our primary GPT-4o setting, an agent that trusts raw memory dies more than twice as often as the same agent given no memory at all. Third, auditing helps but does not close the gap: a transparent read-time filter removes much of the safety cost in text mode, yet even oracle stale labels bring no further significant gain on the current grid size, and when visual auditing is unreliable, filtering yields no consistent benefit. Together these results frame spatial-memory staleness as a safety failure mode and isolate reliable visual grounding and action selection under memory–observation conflict as the central open challenges for memory-augmented agents.
[NLP-44] he Personalization Mirag e: How LLM s Fabricate User Profiles and Why Self-Monitoring Misleads
【速读】： 该论文旨在解决个性化大语言模型（Personalized LLMs）在具备持续记忆能力时，其用户模型（user model）的忠实性（faithfulness）缺乏系统评估的问题，尤其关注“过度推断”（over-inference, OI）现象——即模型在缺乏充分证据的情况下虚构用户属性。其解决方案的关键在于构建了MirageBench基准测试体系，包含150个在刻板印象、反刻板印象和中性三类之间平衡的人格画像，覆盖从“想象梯度”延伸的六种个性化任务，采用由独立裁判操作的四分类忠实性分类标准（经盲评人类标注者验证，四分类Cohen’s kappa = 0.863，二分类kappa = 0.900），并提供涵盖12个模型（7个模型家族）在143,616条判断样本上的排行榜。研究发现，所有模型均存在显著的过度推断行为，平均41.6%（加权41.8%）的声明存在虚构；更关键的是揭示了“自我监控反转”（Self-Monitoring Inversion）现象：模型自评的过度推断水平与其实际被裁判评定的过推程度呈负相关（rho = -0.60, p = 0.044），表明模型自我报告的置信度无法可靠用于模型间比较。尽管单个模型内部的自审计仍具一定判别能力（AUROC 0.58–0.83），但外部验证仍是确保个性化可信性的核心依据。此外，研究还发现过度推断具有任务依赖性（27%–59%），且在多轮对话中虚构属性近似线性累积而极少修正。因此，该研究强调以外部验证取代模型自报告，作为实现可信赖个性化的核心范式。
链接: https://arxiv.org/abs/2608.04570
作者: Yushi Sun,Yanjie Zhang,Rui Sheng
机构: Tencent LIGHTSPEED(腾讯光速); Tencent(腾讯)
类目: Computation and Language (cs.CL)
备注:
Abstract:Personalized LLMs with persistent memory are increasingly deployed, yet the faithfulness of their user models remains unexamined. We study over-inference (OI): the phenomenon where LLMs fabricate user attributes beyond what evidence supports. We introduce MirageBench, comprising 150 personas balanced across stereotypical, counter-stereotypical, and neutral profiles, 6 personalization tasks spanning an ``imagination gradient’‘, a four-way faithfulness taxonomy operationalized by an independent judge (validated against a blind human annotator on 400 claims: Cohen’s kappa = 0.863 four-class, kappa = 0.900 binary), and a leaderboard of 12 models across 7 families on 143616 judged claims. We find that over-inference is pervasive: every one of the 12 models over-infers 35%–49% of its claims (cross-model mean 41.6%; claim-weighted 41.8%), with no model in this evaluation escaping it. Most strikingly, we surface a Self-Monitoring Inversion: at the model-selection level, models’ self-assessed OI is negatively rank-correlated with their judge-measured OI (rho = -0.60, p = 0.044; exploratory, wide bootstrap CI [-0.90, +0.06], n = 12). The models that report the least over-inference tend to be flagged as fabricating the most, so self-reported confidence is a misleading signal for comparing models, even though within a single model self-audit still ranks that model’s own claims moderately well (AUROC 0.58–0.83). We further show that OI is task-dependent (27%–59%) and that, in a multi-turn pilot, inferred attributes accumulate approximately linearly with little revision. MirageBench positions external verification, rather than model self-report, as a more reliable foundation for trustworthy personalization.
[NLP-45] Relevant but Incomplete: Referential Dangling as a Paradigm-Level Failure Mode in Hard Prompt Compression
【速读】： 该论文旨在解决硬性提示压缩（hard prompt compression）在长上下文推理中因独立评分与选择文本单元而导致的指代悬空问题（referential dangling），即在压缩过程中，由于独立筛选导致依赖性证据对被拆分，保留了答案相关片段但删除了其语义解释所必需的实体指称信息，从而影响模型推理准确性。其核心解决方案是：在传统基于相关性的压缩基础上，引入对指代完整性（referential completeness）的显式建模——通过训练一个轻量级分类器，在压缩阶段识别并重新插入那些对理解已保留文本至关重要的被删去句子，以恢复被破坏的语义连贯性。该方法无需额外标注支持段落，仅依赖于对被省略内容是否关键的预测，在压缩率仅微增至0.31时即实现4.7个百分点的准确率提升，显著缩小了与完整上下文之间的性能差距，证明了兼顾相关性与指代完整性的必要性。
链接: https://arxiv.org/abs/2608.04569
作者: Zhengpei Hu,Kai Li,Dapeng Fu,Xuechao Zou,Yuanhao Tang,Yue Li,Tengfei Cao,Jianqiang Huang
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Code: this https URL
Abstract:Hard prompt compression reduces long-context inference cost by independently scoring tokens, sentences, or chunks and retaining the highest-scoring units under a budget. We identify a structural failure in this procedure: independent selection can split dependent evidence pairs, retaining one member while deleting the other. When retained text contains an answer but deleted text defines the entity needed to interpret it, we call the result referential dangling. At a compression ratio of 0.30, Beaver, which ranks coherent chunks using Qwen3-0.6B embeddings, leaves the answer path incomplete in 34-54% of bridge examples across three multi-hop question answering datasets. On a shared HotpotQA bridge set, all six hard compressors we test exhibit dangling at rates up to 60%, and every document in LongBench-v2 Single-Document QA contains at least one dangling reference. On dangling examples evaluated with Qwen3-8B, reinserting the missing supporting paragraph while removing nonsupporting paragraphs to maintain the token budget improves accuracy by 29-34 percentage points (p 0.0001), recovering at least 88% of the gap to contexts retaining both supporting paragraphs. Stronger answer models do not absorb the loss: on MuSiQue, GPT-5.5 is 8.8 points less accurate on compressed contexts than on contexts retaining both supporting paragraphs. Finally, we train a compact classifier to rank omitted sentences by whether they are needed to interpret retained text and reinsert the top-ranked candidates without support annotations at inference. On HotpotQA with Qwen3-8B, this automatic restoration improves accuracy by 4.7 points while changing the compression ratio only from 0.30 to 0.31. Hard compressors should optimize both relevance and referential completeness.
[NLP-46] STRIVE: Probing Reasoning Limits in Graded Plausibility Generation and Evaluation
【速读】： 该论文旨在解决心理语言学研究中构建可控事件集所面临的劳动密集型问题，即如何在保持其他事件特征恒定的前提下，系统性地生成具有不同可理解性（可预测性）水平（合理与不合理）和分类难度（易与难）的事件。其解决方案的关键在于提出STRIVE框架——一个基于大语言模型（LLM）的联合生成与评估方法。该框架以给定动词为基础，构建共享事件框架，并通过仅改变某一事件成分来生成四种条件下的事件（合理-易、合理-难、不合理-易、不合理-难），从而实现对事件可理解性效应的精确隔离。实验表明，通过引入全局推理思维链（scratchpad）和评估器引导的迭代优化，可将高质量事件集生成率从基线的16.7%提升至75.0%，同时显著提高评估者与人类判断的一致性。然而，在可理解性边界附近的事件仍存在高度主观分歧，最佳评估器在“不合理-难”条件下准确率仅为57%，表明当前模型仍难以完全替代人类判断。总体而言，STRIVE通过自动化初始事件集生成与评估，为心理语言学研究提供了一种可扩展、高效且可控的解决方案。
链接: https://arxiv.org/abs/2608.04567
作者: Bhiman Kumar Baghel,Anna Chrabaszcz,Tessa Warren,Michael Walsh Dickey,Haley C. Dresang,Xiang Lorraine Li
机构: University of Pittsburgh (匹兹堡大学); University of Wisconsin–Madison (威斯康星大学麦迪逊分校)
类目: Computation and Language (cs.CL)
备注: Under Review
Abstract:Event knowledge concerns who does what to whom. Psycholinguists use event-plausibility judgments to examine how this knowledge supports human language processing. To isolate plausibility effects, these studies require controlled event sets in which one event slot varies across plausibility levels while all other event features remain fixed. Constructing such sets manually is labor-intensive. We therefore introduce STRIVE, an LLM-based framework for jointly generating and evaluating controlled event sets crossing plausibility class (plausible vs. implausible) with intended classification difficulty (easy vs. hard). Given a verb, STRIVE constructs a shared event frame, then produces one event per condition by varying one slot while holding all others fixed. In experiments with six models across 60 verbs, GPT-5.1 produced high-quality sets only 16.7% of the time using the baseline generation prompt. Adding a global reasoning scratchpad and evaluator-guided refinement raised this rate to 75.0%. Greater reasoning effort also improved evaluator–human agreement. Nevertheless, events near the plausibility boundary remain most difficult. They elicit the greatest human disagreement, and the best evaluator reaches only 57% accuracy on the implausible-hard condition, indicating a need for human input. Overall, STRIVE offers a scalable approach to reducing manual effort by automating initial event-set generation and evaluation for psycholinguistic studies.
[NLP-47] Breadcrumbing Search Agents
【速读】： 该论文旨在解决基于大语言模型（LLM）的搜索代理在信息检索任务中面临的动态安全威胁问题，尤其关注由外部工具返回结果所引入的不可信内容风险。传统研究多聚焦于静态网页内容注入攻击，但现代搜索代理具备多轮查询与跨源交叉验证能力，使得单一污染页面的影响被稀释或排除。本文揭示了搜索接口作为信息传递通道存在脆弱性：攻击者可通过受控中介持续引导代理获取证据的方式和路径，从而在长周期任务中逐步构建一致的误导性证据链。其核心解决方案是提出一种名为“权威链劫持”（Authority-Chain Hijack, ACH）的策略驱动攻击机制，通过协调多个看似相互印证的搜索结果与页面内容，在代理的整个推理轨迹中形成连贯且可信的虚假证据流，显著提升攻击成功率（在完整SafeSearch测试集上达到55.9%的整体攻击成功率，最大命中率高达83.3%）。此外，为实现攻击策略的自动化优化，论文进一步提出“轨迹引导策略演化”（Trace-Guided Strategy Evolution, TGSE），利用执行轨迹进行策略迭代改进，替代人工重设计，其最优配置在保留测试集中实现71.4%的攻击成功率与95.0%的最大命中率，充分体现了动态、协同式攻击在复杂代理系统中的隐蔽性与有效性。
链接: https://arxiv.org/abs/2608.04565
作者: Xuebin Li,Hanqing Zhao,Siyuan Liang,Kejiang Chen,Weiming Zhang,Dacheng Tao,Nenghai Yu
机构: University of Science and Technology of China (中国科学技术大学); Nanyang Technological University (南洋理工大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 38 pages, 7 figures
Abstract:LLM-based search agents are widely used for information-seeking tasks, but their reliance on external tool returns introduces a critical security risk: web content retrieved during execution is untrusted, exposing agents to prompt injection and goal hijacking. Prior work on search-agent safety primarily focuses on static web-content injection, but modern agents issue follow-up queries and cross-check competing sources, so a single injected page is often diluted or rejected. We show that the channel delivering search and page observations is a fragile security boundary: beyond exposing the agent to a single poisoned page, a mediated search interface can repeatedly steer how the agent gathers evidence and forms its final answer. Under a constrained tool-intermediary threat model, appending only one controlled result per query can substantially increase attack success when the evidence is coordinated across the agent’s trajectory. We study this setting with a strategy-driven long-horizon attack system and introduce Authority-Chain Hijack (ACH), an expert-refined strategy that turns isolated search-result and page-content manipulations into a coherent evidence chain across seemingly corroborating sources. ACH achieves the highest Overall ASR among all baselines, reaching 55.9% / 83.3% ASR / MaxN ASR on the full SafeSearch test split. We further introduce Trace-Guided Strategy Evolution (TGSE), which automatically improves attacker strategies from execution traces, replacing manual redesign with trace-driven refinement; its strongest single setting reaches 71.4% / 95.0% in held-out evaluation.
[NLP-48] Representing Visual Evidence for Item Difficulty Prediction: Visual Textualization and Image-Native Modeling
【速读】： 该论文旨在解决在缺乏学生作答数据的情况下，如何准确预测数学类题目难度的问题，尤其关注包含视觉成分的题目中，视觉证据应如何有效表示以提升难度预测性能。现有方法通常将题干与选项视为纯文本处理，对于含视觉元素的题目则采用先将图像“文本化”再使用文本模型预测的流程。本文提出核心问题：在题目难度预测任务中，视觉证据应以何种方式表征？研究对比了仅使用文本、视觉内容文本化（将视觉信息转化为语言描述）以及图像原生建模（保留原始图像）三种策略。基于从真实学生作答数据校准难度的Eedi题目数据集，作者直接训练大型语言模型（LLM）和视觉-语言模型（VLM）进行难度回归。实验结果表明，两种视觉接口均取得最低的点估计误差，但领先系统间难以可靠排序；其中，Open-VLM的文本化策略在所有评估的LLM上表现更优，而更广泛的模型适配策略在所有图像原生的VLM上表现更佳。测试时干预分析显示，模型性能依赖于完整的题目图像，但无法有效分离出仅由视觉成分带来的贡献。此外，两类视觉接口在题目层面的错误模式部分互补，且计算流程差异显著。因此，研究结论指出：视觉文本化不应被视为唯一可行方案，图像原生建模是一种具有竞争力的替代方法，其有效性高度依赖于视觉-语言模型的适配方式。
链接: https://arxiv.org/abs/2608.04554
作者: Han Chen,Ming Li,Hong Jiao,Tianyi Zhou
机构: University of Maryland (马里兰大学); Mohamed bin Zayed University of Artificial Intelligence (穆罕默德·本·扎耶德人工智能大学)
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Predicting item difficulty from content can provide an initial estimate for newly developed questions before sufficient student responses are available. Existing approaches typically represent the question stem and answer choices as text. When mathematics items contain visual components, a common pipeline first textualizes that evidence and then applies a text predictor. We ask: how should visual evidence be represented for item difficulty prediction? We compare question text alone, visual textualization, which expresses visual evidence in language, and image-native modeling, which retains the original image. Using Eedi items with difficulty calibrated from student responses, we train large language models (LLMs) and vision-language models (VLMs) directly for difficulty regression. Both visual interfaces achieve the lowest point estimates, although the leading systems cannot be reliably ordered. Open-VLM textualization yields lower RMSE point estimates for all evaluated LLMs, while broader adaptation does so for all image-native VLMs. Test-time interventions show dependence on the paired full-item image, but do not isolate the additional visual component. The two visual interfaces also make partially complementary item-level errors and differ substantially in computational workflow. Thus, textualization should not be treated as the only practical interface: image-native modeling is a competitive alternative whose effectiveness depends on how the VLM is adapted.
[NLP-49] Relational Response Fields: A General Theory of Black-Box LLM Response Consistency and Recovery
【速读】： 该论文旨在解决黑箱语言模型生成结果在多响应集合中是否可被可靠恢复的根本性问题，即在不依赖模型内部机制的前提下，判断一组黑箱输出能否通过外部方法重构出正确答案。其核心贡献在于提出“关系响应场”（Relational Response Field, RRF）这一形式化框架，将查询的类型化变换下的响应表示为具有结构关系的图数据：边传输算子（D）刻画响应在语义等价变换（如改写、缩放、分解等任务对称性）下的应变规律，锚点算子（A）则编码独立可信的证据源（如代码执行结果或验证器输出）。在此框架下，定义了关键指标 γk(D,A)，表征在最多 k 个响应节点被破坏时，恢复过程的内在难度。γk(D,A)>0 当且仅当任意 k-节点污染均可被识别，其倒数提供了确定性的稳定性上界，且通过匹配的两点极小极大下界证明该依赖关系不可超越。这表明，仅依赖响应一致性（consistency）的方法无法察觉共现幻觉等零空间方向（null directions），因而存在本质局限。论文进一步设计了稀疏场修复算法，明确区分了信息论可辨识性与凸优化所需的更强零空间条件，并通过受控定理检验及黑箱数学/代码实验验证了四个理论预测：一致性与真实性分离、锚点相变现象、冗余饱和效应以及跨模型、跨任务的修复难度预测能力。结果表明，γk(D,A) 是响应恢复实例的可观测结构性属性，而非某一修复启发式方法的评分指标。
链接: https://arxiv.org/abs/2608.04552
作者: Song Zichen
机构: Sungkyunkwan University (成均馆大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Black-box language-model reliability is commonly pursued by sampling, prompting, voting, verifying, or iteratively revising individual answers. We ask a prior question: \emphwhat determines whether a collection of black-box responses is recoverable at all? We represent responses to typed transformations of a query as a \emphrelational response field (RRF). Edge transports encode how valid responses must change under paraphrase, scaling, decomposition, refactoring, or other task symmetries; anchors encode independently trusted evidence such as execution or a verifier. For relation operator D , anchor operator A , and at most k corrupted response nodes, we identify \gamma_k(D,A) as the intrinsic difficulty of black-box response recovery. It is positive exactly when every k -node corruption is identifiable; it gives a deterministic stability bound proportional to 1/\gamma_k ; and a matching two-point minimax lower bound shows that no estimator can improve this dependence. Thus consistency is not truth: relation-only methods are blind to null directions, including shared hallucinations. We derive sparse field-repair algorithms while separating information-theoretic identifiability from the stronger null-space conditions required by convex optimization. Controlled theorem tests and black-box mathematics/code experiments evaluate four theory-fixed consequences: consistency–truth separation, anchor phase transitions, redundancy saturation, and cross-model, cross-task prediction of repair difficulty. The results support \gamma_k(D,A) as a measurable property of a response-recovery instance, rather than a score attached to one repair heuristic.
[NLP-50] EuroExec: Frontier Language Models Fall Short of Expert Judgment on European Executive Decision Tasks EACL2027
【速读】： 该论文旨在解决前沿生成式人工智能（Generative AI）在应对开放性、复杂性真实世界问题时的评估有效性问题，尤其针对其在实际应用场景中表现与专业水平之间存在的显著差距。传统评测基准多聚焦于结构化、封闭式问题，难以反映模型在真实、长篇幅、需深度领域知识的欧洲高管任务中的实际能力。为此，研究提出了EuroExec——一个由47位经过认证的领域专家撰写的413个开放式长文本欧洲高管任务构成的人类专家基准，所有问题均源自真实案例。通过人工评估结合多维度评分标准、项目特定检查清单及偏好排序，构建了综合指标“求解率”（Solve Rate）。结果显示，最强模型仅能解决56.9%的任务，而人类专家参考答案在盲评中几乎达到天花板级表现，且在74%的直接对比中被优先选择。研究关键在于：唯有依赖具备领域专业知识的人类评估者，并通过严格的统计方法验证评估一致性，才能准确揭示生成式AI在真实复杂任务中的局限；同时，自动评估指标在此类具有主观真实标准的开放性问题上也明显失效。
链接: https://arxiv.org/abs/2608.04549
作者: Pau Arnal,Khaled Denfir,Danylo Smahliuk,Amrut Avhad,Marcus A. Castro
机构: Sovrano AI
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 16 pages, 9 figures, 12 tables, submitted to EACL 2027
Abstract:Frontier LLMs are increasingly put to use on open-ended complex questions, different in nature from the ones they are typically evaluated on. We dedicate more than 4,000 human expert hours to evaluate a selection of six frontier LLMs on a member of this class of problems: EuroExec, our introduced human expert-based benchmark composed of 413 open-ended long-form European executive tasks authored by 47 vetted domain experts, each question drawn from experience in a real case. Every response is manually evaluated through a multi-attribute rubric, an item-specific checklist of requirements, and a preference rank ordering, extracting an aggregate metric “Solve Rate”. The strongest model solves only 56.9% of tasks, while expert-written reference answers judged blindly are solved at near-ceiling levels and are preferred over every model response in 74% of direct rankings, placing frontier generative systems well below the professional standard of work they are already used for. We see that the best way to extract this kind of conclusion is by employing human evaluators, carefully checking their consistency through rigorous statistical analysis, and observe that automatic measurements also fall short when evaluating on this case of real-world open-ended problems with a subjective ground truth.
[NLP-51] ODRA: Synthesizing Cognitive Behavioral Therapy Sessions with Structured Chain-Of-Thought and Dynamic Patient Resistance
【速读】： 该论文旨在解决生成式认知行为疗法（Cognitive Behavioral Therapy, CBT）对话时面临的两大挑战：一方面需严格遵循CBT的结构性治疗流程，另一方面又要真实模拟患者在临床实践中表现出的抗拒性与不可预测行为。现有基于脚本的方法难以捕捉动态交互，而多智能体方法虽具备一定灵活性却难以维持CBT的顺序结构，且普遍存在“讨好倾向”（sycophancy）问题，即生成的患者过于顺从，无法反映真实临床场景中的复杂性。为此，本文提出ODRA框架，其核心创新在于采用基于思维链（Chain-of-Thought, CoT）的生成策略，并以贝克（Beck, 2020）提出的经典CBT指南为底层依据，确保生成内容在治疗逻辑和结构上高度对齐。关键突破在于引入“阻力协调器”（resistance orchestrator），通过引导技术主动调控患者的抵抗水平，有效缓解讨好倾向，提升患者行为的真实性与多样性。自动化评估与专家评审均表明，ODRA在治疗技能、CBT一致性及患者行为保真度方面显著优于现有方法，且13项临床指标中有12项获得持证心理医生更青睐。此外，基于该数据集微调的模型在应对合作型与抗拒型患者时均展现出更强的临床适应能力，证实了在合成训练数据中显式建模患者抗拒行为可直接增强下游模型的临床鲁棒性。
链接: https://arxiv.org/abs/2608.04524
作者: Javier Rodriguez-Juan,Hiba Arnaout,Jose Garcia-Rodriguez,David Tomás,Iryna Gurevych
机构: University of Alicante (阿利坎特大学); Technische Universität Darmstadt (达姆施塔特工业大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 39 pages, 23 figures, 12 tables
Abstract:Synthetic generation of Cognitive Behavioral Therapy (CBT) sessions is challenged by two competing demands: adhering to strict therapeutic structure while modeling the resistant, unpredictable behavior of real patients. Existing script-based methods fail to capture dynamic therapeutic interactions, while multi-agent approaches struggle to adhere to CBT’s sequential structure; both suffer from sycophancy, producing overly compliant patients that misrepresent real clinical settings. In this work we introduce ODRA, a novel framework for synthesizing therapy dialogues through a Chain-of-Thought (CoT) strategy grounded in foundational CBT guidelines (Beck, 2020). ODRA further incorporates a resistance orchestrator to solve patient sycophancy, which employs steering techniques to elicit behaviors aligned with their resistance level. Automated and expert evaluations show that ODRA significantly outperforms existing methods across therapeutic skills, CBT alignment, and patient behavioral fidelity, with licensed psychologists preferring ODRA sessions across 12 of 13 clinical metrics. Furthermore, models fine-tuned on our dataset demonstrate superior therapeutic performance against both cooperative and resistant patients, validating that explicit resistance modeling in synthetic training data directly translates to downstream clinical robustness.
[NLP-52] Leak-Resistant Unlearning: A New Benchmark for Evaluating Multi-Hop Reasoning Consistency and Recovery Robustness
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在实现知识遗忘（unlearning）过程中存在的两大核心问题：其一，现有基准测试多局限于单一跳（single-hop）问答及有限的多跳（multi-hop）问题，未能充分覆盖知识间相互关联导致的复杂推理路径所引发的知识泄露风险；其二，当前的遗忘方法在静态评估下表现良好，但对轻量级后遗忘适应等恢复攻击（recovery attacks）缺乏鲁棒性，导致已遗忘知识可能被部分重构。为此，本文提出一个名为 \unlearning 的新型基准，用于系统评估模型在多样化推理路径和多种恢复攻击下的鲁棒性知识移除能力。该解决方案的关键在于构建一个动态且对抗性的评估框架，涵盖多跳推理与恢复攻击场景，并通过在3个模型、6种遗忘方法及2个精心设计的数据集上的实验，揭示了现有方法在鲁棒性方面的显著缺陷，同时深入探讨了遗忘质量、鲁棒性与模型可用性之间的权衡关系。
链接: https://arxiv.org/abs/2608.04519
作者: Haoting Qian,Qingjie Zhang,Zhicong Huang,Cheng Hong,Han Qiu
机构: Tsinghua University (清华大学); Ant Group (蚂蚁集团)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 19 pages, 7 figures
Abstract:Benchmarking machine unlearning methods is critical to understand whether sensitive knowledge is removed from large language models (LLMs) or not. Current unlearning benchmarks include mainly single-hop questions and a narrow set of multi-hop questions. Although effective, they still face two challenges. (1) Knowledge is not isolated, whereby diverse multi-hop reasoning paths can potentially induce knowledge leakage than normal queries. (2) Unlearning may be fragile: unlearned knowledge can be partially recovered through recovery attacks such as lightweight post-unlearning adaptation, making static evaluation insufficient. Therefore, in this paper, we introduce \unlearning as a novel benchmark to understand robust LLM knowledge removal across diverse reasoning paths and recovery attacks. We experiment with this benchmark on 3 models, 6 unlearning methods, and 2 carefully curated datasets. Results show that existing methods are vulnerable to multi-hop reasoning paths and recovery attacks. We further explore the trade-off among forget quality, robustness, and model utility for LLM unlearning.
[NLP-53] CARVE: Cross-Slice Anisotropic Reallocation of Visual Evidence for Efficient 3D Medical Volume Understanding
【速读】： 该论文旨在解决基于切片的多模态大模型（slice-based MLLMs）在处理3D医学影像时因切片序列产生大量冗余视觉标记（visual tokens）所导致的计算负担过重问题，尤其针对相邻切片间高度重复的深度方向内容冗余。其核心挑战在于：传统方法通过简单增加切片数量或提升平面分辨率来扩展视觉表示，但实验表明在固定预算下性能提升呈饱和趋势，且提升平面分辨率比增加切片更有效。因此，关键解决方案在于实现资源受限下的非均匀2.5D视觉标记分配——提出一种无需训练的压缩框架CARVE，将3D体积沿深度轴划分为语义连贯的窗口，依据跨切片证据的归一化分布动态分配有限的视觉标记预算，优先保留具有代表性的切片作为空间锚点，并从全体积中检索局部变化特征，再将剩余可选标记聚合至邻近锚点。该方法在移除约80%视觉标记的情况下，仍保持接近完整标记的性能（98.1%），并在所有AMOS-MM报告生成指标上超越现有压缩基线，显著提升了标记利用效率与性能保留度。
链接: https://arxiv.org/abs/2608.04515
作者: Zhenyu Yi,Qiang Hu,Zhenhao Li,Jiaxuan Zhao,Yusong Sun,Lichi Zhang
机构: 未知
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Slice-based MLLMs leverage mature 2D encoders by representing 3D volumes as sequences of 2D slices. However, this slice-wise formulation produces thousands of visual tokens that burden the LLM backbone, many of which capture overlapping visual evidence across adjacent slices. To understand how effectively a growing visual token budget improves performance, we perform scaling analyses on two 3D medical VQA benchmarks and find diminishing returns: cost keeps rising while accuracy saturates, and improving in-plane resolution is more effective than adding slices at comparable budgets. The budget should therefore be allocated more selectively rather than simply enlarged, yet most token compression methods are designed for 2D images or videos, where redundancy arises from spatial layout or temporal motion rather than from near-duplicate content along the depth axis. We present CARVE, a training-free framework that compresses visual tokens prior to LLM inference and casts token reduction as budget-constrained 2.5D allocation. CARVE partitions the depth axis into coherent windows and allocates tokens non-uniformly according to normalized cross-slice evidence. Under a shared budget, CARVE builds spatial anchors on representative slices and retrieves locally varying evidence from the full volume, then merges remaining eligible tokens into nearby anchors within each window. Removing roughly 80% of the visual tokens on Hulu-Med-7B, CARVE leads all compression baselines on every AMOS-MM report-generation metric, with 6.2 points higher retention of full-token quality than the strongest baseline, and preserves 98.1% of full-token performance across three VQA benchmarks.
[NLP-54] RESPClinBench: Benchmarking Multimodal Clinical Decision-Making and Longitudinal Disease Management in Respiratory Specialty Care
【速读】： 该论文旨在解决现有医学评估体系在呼吸专科临床决策支持中无法充分反映真实世界多模态诊疗、长期风险评估、指南一致性干预及全程管理需求的问题。其核心解决方案是构建RESPClinBench——一个基于真实临床场景的基准测试体系，涵盖急性加重期慢性阻塞性肺疾病（AECOPD-PIM）与肺结节多模态评估（PNBIM）两大任务，通过整合去标识化的临床数据、由资深呼吸科医师审定的参考答案与原子化临床行动点，并采用自动化评分框架（结合原子动作召回率与基于评分标准的大型语言模型作为裁判评估），实现对七种主流大语言模型在复杂呼吸专科任务中的系统性评估。关键创新在于将临床行动覆盖度、整体评估逻辑与独立的安全性预警机制相结合，从而为模型选型与前瞻性验证提供具有临床意义的依据。
链接: https://arxiv.org/abs/2608.04514
作者: Mouxiao Bian,Zhi Chen,Ruiyao Chen,Lu Lu,Hengrui Liang,Chaoyi Huang,Yiluo Lin,Jingru Ding,Yun Zhong,Yuming Su,Jie Xu
机构: Shanghai Artificial Intelligence Laboratory (上海人工智能实验室); Macau University of Science and Technology (澳门科技大学); First Affiliated Hospital of Guangzhou Medical University (广州医科大学第一附属医院); Guangzhou Institute of Respiratory Health (广州呼吸健康研究院)
类目: Computation and Language (cs.CL)
备注:
Abstract:Background: Respiratory specialty care requires multimodal interpretation, longitudinal risk assessment, guideline-concordant intervention, and whole-course management, which are poorly represented by examination-oriented medical benchmarks. Objective: To develop RESPClinBench, a real-world scenario-based benchmark for respiratory clinical decision-making, and evaluate seven contemporary large language models across AECOPD-PIM and PNBIM. Methods: RESPClinBench cases were adapted from de-identified respiratory clinical data. Three attending-level respiratory physicians revised cases, reference answers, and atomic clinical-action points, while one senior respiratory specialist performed cross-review and final adjudication. AECOPD-PIM comprised 427 open-ended COPD cases, and PNBIM comprised 196 multimodal pulmonary nodule cases combining chest CT with structured clinical information. Seven models generated 4,361 responses through standardized API inference with temperature 0 and a maximum output length of 8192 tokens. An automated framework calculated the final score as the arithmetic mean of atomic-action recall and rubric-based LLM-as-a-Judge assessment. Results: Across 623 cases, the mean final score was 68.58. Qwen3.6-27B ranked first overall at 71.22, Qwen3.5-397B-A17B led PNBIM at 72.48, and Qwen3.6-27B led AECOPD-PIM at 71.11. Imaging hallucination and serious medical risk occurred in 31.85% and 8.16% of PNBIM responses; medication-safety risk and serious medical risk occurred in 26.93% and 1.44% of AECOPD-PIM responses. Conclusions: RESPClinBench identifies task-specific limitations in multimodal pulmonary nodule assessment and longitudinal COPD management. Combining explicit clinical-action coverage, holistic evaluation, and independent safety flags provides a clinically grounded basis for model selection and prospective validation.
[NLP-55] K-EXAONE 2.0 Technical Report
【速读】： 该论文旨在解决当前多语言基础模型在长上下文理解、智能体编程能力及跨文化安全性方面的局限性，尤其是在韩国社会文化语境下实现更可靠、更高效的通用人工智能能力。其核心解决方案是通过“复用与扩展”（upcycle and expand）策略，在不从零训练的前提下对K-EXAONE进行架构升级，构建出一个具有7500亿总参数、每令牌激活约370亿参数的混合专家（Mixture-of-Experts, MoE）模型。该模型支持最长256K tokens的上下文长度，并将多语言覆盖从6种扩展至10种。通过结合持续预训练、以难度为导向的中期训练以及后训练的全流程训练范式，显著增强了模型的推理能力、智能体编程表现、多语言适应性及基于韩语社会文化背景的安全性。实验表明，K-EXAONE 2.0在智能体编程和长上下文理解方面提升最为显著，尤其在长上下文检索与安全性任务中表现出色。该模型以Apache 2.0许可证开源，标志着迈向全球前沿级基础模型的起点而非终点。
链接: https://arxiv.org/abs/2608.04505
作者: Eunbi Choi,Kibong Choi,Sehyun Chun,Seokhee Hong,Junwon Hwang,Hyojin Jeon,Ahra Jo,Hyunjik Jo,Yeonsik Jo,Minhyeok Jung,Doyoung Kim,Heegyu Kim,Joonkee Kim,Seonghwan Kim,Soyeon Kim,Sunkyoung Kim,Yireun Kim,Yongil Kim,Byungoh Ko,Changhun Lee,Dohaeng Lee,Haeju Lee,Jinsik Lee,Kyungmin Lee,Minwoo Lee,Wonkee Lee,Sangha Park,Sungjune Park,Kwangrok Ryoo,Kijung Seo,Minju Seo,Yongwoo Song,Sejong Yang,Heuiyeen Yeen,Stanley Jungkyu Choi,Yemuk Choi,Yongchan Chun,Jiwon Ham,Dasol Hong,Sujeong Im,Kijeong Jeon,Gerrard Jeongwon Jo,Hyeongjun Jo,Yujin Jo,Jiyeon Jung,Naeun Kang,Daeseong Kim,Euisoon Kim,Hayeon Kim,Hyosang Kim,Myoungshin Kim,Unsol Kim,Youchul Kim,Chaeeun Lee,ChaeYoon Lee,Edward Hwayoung Lee,Honglak Lee,Hwansoo Lee,Minkyung Lee,Sangeun Lee,Solji Lim,Woohyung Lim,Chanwoo Moon,Jueun Mun,Jimin Park,Seojeong Park,Yongmin Park,Hyerin Seo,Donghyeon Shin,Donghyun Son,Eunyong Son,Kaehyun Um,Sihoon Yang,Chang En Yea,Sihyuk Yi,Kyungjae Yoo,Chansik Yoon
机构: LG AI Research( LG人工智能研究院)
类目: Computation and Language (cs.CL)
备注:
Abstract:This technical report presents K-EXAONE 2.0, an open-weight multilingual foundation model developed by LG AI Research as a step in our effort toward global frontier-scale foundation models. Rather than training from scratch, we upcycle K-EXAONE and expand its architecture, yielding a Mixture-of-Experts (MoE) model with 750B total parameters and approximately 37B activated per token—more than three times the capacity of its predecessor. K-EXAONE 2.0 supports context lengths of up to 256K tokens and expands multilingual coverage from six to ten languages. Its training pipeline combines continual pre-training, difficulty-focused mid-training, and post-training to strengthen reasoning, agentic coding, multilingual capability, and safety grounded in Korean sociocultural contexts. Across nine evaluation categories selected to reflect the conditions of practical use, K-EXAONE 2.0 improves over K-EXAONE and remains competitive with open-weight models, showing its largest gains in agentic coding and long-context understanding and its clearest strengths in long-context retrieval and safety. Released under the Apache 2.0 license, K-EXAONE 2.0 enables the wider AI ecosystem to evaluate, deploy, adapt, and build upon it, while marking the beginning—rather than the endpoint—of our challenge toward the global frontier.
[NLP-56] Energy- and Memory-Efficient PEFT Methods for Personalized On-Device SLMs on Consumer GPUs
【速读】： 该论文旨在解决在资源受限设备上部署和个性化大型语言模型（Large Language Models, LLMs）所面临的高显存（VRAM）、时间和能耗成本问题。尽管参数高效微调（Parameter-Efficient Fine-Tuning, PEFT）为小型语言模型（Small Language Models, SLMs）的轻量化适配提供了可行路径，但现有研究缺乏在多种模型架构下，结合通用任务与个性化任务基准，并综合考虑能耗因素的系统性比较。为此，本文对五种微调方法（全量微调、LoRA、LoRA+、QLoRA 和 BitFit）在四款来自两类架构（基于 Transformer 的 TinyLlama-1.1B、Qwen3-1.7B 以及基于状态空间模型（State Space Model, SSM）的 Mamba-1.4B、Mamba-2-1.3B）的 SLM 上进行了全面评估，涵盖三个 GLUE 任务（SST-2、QNLI、STS-B）和三个 LaMP 个性化任务（LaMP-1、LaMP-2、LaMP-3）。通过引入以能耗为导向的 NetScore-E 与以内存为导向的 NetScore-M 两个综合指标，采用严格的“能耗优先”选择规则（优先最高 NetScore-E，平局时以 NetScore-M 决定），结果表明：LoRA+ 在 24 种配置中于 19 次取得最高 NetScore-E，13 次取得最高 NetScore-M，且在 18 次中被选为最优方法；而仅适用于 Transformer 架构的 QLoRA 虽因量化显著降低峰值显存使用（最高达 3.9 倍压缩），在 5 个 Transformer 配置中获得最佳 NetScore-M，但由于解量化开销导致其在能耗维度表现不佳，最终仅在 1 次中胜出；BitFit 与全量微调几乎未在任一指标中具备竞争力。此外，TinyLlama-1.1B 在六个基准中均表现出优异的能效表现，尤其在五个任务中领先于 NetScore-E，四个任务中领先于 NetScore-M。研究结论表明，紧凑型 SLM 配合 PEFT 是实现低功耗、可个性化的本地化部署的有效途径，其最优方法的选择取决于主导约束条件：当以能耗为核心时，推荐使用 LoRA+；当以内存为瓶颈时，则应选用 QLoRA。
链接: https://arxiv.org/abs/2608.04488
作者: Kuanysh Akhmetzhanov,Jurn-Gyu Park
机构: Nazarbayev University (纳扎尔巴耶夫大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Despite rapid advances in large language models (LLMs), deploying and personalizing them on resource-constrained devices remains impractical due to high VRAM, time, and energy costs. Parameter-Efficient Fine-Tuning (PEFT) of Small Language Models (SLMs) offers a promising alternative, yet few studies compare PEFT methods across architectures using both general and personalization benchmarks while accounting for energy consumption. We compare five fine-tuning approaches (Full Fine-Tuning, LoRA, LoRA+, QLoRA, and BitFit) on four SLMs from two families (Transformer-based: TinyLlama-1.1B, Qwen3-1.7B; SSM-based: Mamba-1.4B, Mamba-2-1.3B) across three GLUE tasks (SST-2, QNLI, STS-B) and three LaMP personalization tasks (LaMP-1, LaMP-2, LaMP-3). Each configuration is evaluated with the energy-focused NetScore-E and the memory-focused NetScore-M, the two variants that reflect the constraints binding on-device deployment. Methods are selected with a strict energy-first rule (highest NetScore-E, ties broken by NetScore#). LoRA+ achieves the highest NetScore-E in 19 of 24 configurations and the highest NetScore-M in 13 of 24, and is the selected method in 18 of 24. QLoRA, available only for the Transformer models, cuts peak finetuning VRAM by up to 3.9x relative to LoRA and therefore takes the best NetScore-M in 5 of the 12 Transformer configurations, although its de-quantization overhead leaves it selected in only one of them once energy decides. BitFit and full fine-tuning are almost never competitive on either variant, and TinyLlama-1.1B leads the energy-focused NetScore-E on five of the six benchmarks and the memory-focused NetScore-M on four. These results show that compact SLMs paired with PEFT provide a practical, energy-aware path to personalized on-device deployment, with the optimal method set by the dominant constraint: LoRA+ for energy and QLoRA for memory.
[NLP-57] DeepInvert: Semi-Supervised Embedding Inversion Against Obfuscated Language Models
【速读】： 该论文旨在解决云服务中基于语言模型的敏感信息泄露问题，即在用户输入（提示词）传输过程中，如何有效保护包含隐私或机密内容的文本。现有基于混淆（obfuscation-based）的防御机制（如ObfusLM、SentinelLMs、TextObfuscator和DPNR）通过在传输前对提示词表示进行变换，提供一种轻量级替代加密方案的防护手段。然而，本文揭示这些防御措施的实际保护能力远低于先前认知。其解决方案的核心在于提出一种名为DeepInvert的半监督嵌入逆向攻击方法，该方法能够从经过混淆处理的嵌入表示中高精度地恢复原始令牌。关键创新点在于：尽管混淆操作引入了扰动，但未标记的目标嵌入仍保留可被利用的语义结构。DeepInvert通过结合有标签影子数据的监督训练与针对无标签目标嵌入的新型无监督一致性目标，采用混合训练流程交替优化，从而实现更优的逆向性能。此外，该攻击方法进一步扩展至多种编码器型与自回归架构下的多样化混淆机制。实验结果表明，在九种防御策略、五类任务及四种模型架构上，DeepInvert在多数情况下显著优于已有攻击方法；例如，在对抗ObfusLM时，其顶1准确率提升至73.5%，远超此前最佳水平（26.2%）。研究还揭示了一种任务依赖性的内在权衡：保留足够信号以维持模型效用的混淆方案，往往也保留了可供逆向的结构特征，而能有效抵抗逆向的方案则会严重损害任务性能。仅在较简单的分类任务中，部分基于差分隐私（DP）的防御可兼顾两者。因此，论文呼吁对这类混淆型防御机制进行全面重新评估。
链接: https://arxiv.org/abs/2608.04477
作者: Zhicong Huang,Cheng Hong,Tao Wei
机构: 未知
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 20 pages
Abstract:Cloud-based language model services routinely process prompts containing sensitive information. Obfuscation-based defenses—including ObfusLM, SentinelLMs, TextObfuscator, and DPNR—mitigate this risk by transforming prompt representations before transmission, offering a lightweight alternative to cryptographic solutions. We show these defenses provide far less protection than previously believed. We present DeepInvert, a semi-supervised embedding inversion attack that recovers original tokens from obfuscated representations with higher accuracy than prior methods. The key insight is that unlabeled obfuscated embeddings retain exploitable semantic structure despite perturbation. DeepInvert combines supervised training on labeled shadow data with a novel unsupervised consistency objective over unlabeled target embeddings, alternating between the two via a mixed training pipeline. Defense-aware adaptations further extend the attack to diverse obfuscation mechanisms across encoder-based and autoregressive architectures. Experiments on nine defenses, five tasks, and four model architectures show that DeepInvert outperforms prior attacks on most defenses. Against ObfusLM, DeepInvert achieves 73.5% top-1 token recovery versus 26.2% for the previous best. Our results reveal a task-dependent tension: obfuscation schemes preserving enough signal for utility also retain sufficient structure for inversion, while schemes resisting inversion collapse utility. On simpler classification tasks, some DP-based defenses can maintain both. We call for a re-evaluation of this defense class. Comments: 20 pages Subjects: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Machine Learning (cs.LG) Cite as: arXiv:2608.04477 [cs.CR] (or arXiv:2608.04477v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2608.04477 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-58] EndoVLM: An Endoscopy Vision-Language Pre-training Model via Anatomy-Guided Sparsity and Progressive Alignment
【速读】： 该论文旨在解决现有内镜基础模型（Foundation Models, FMs）在训练过程中过度依赖单一模态的图像或视频数据，而忽视了临床报告中蕴含的丰富语义信息这一关键问题。其核心挑战在于：结构化的解剖学描述与高冗余、非结构化的视觉流之间存在根本性的模态鸿沟，导致难以有效对齐和利用临床文本信息。为此，本文提出EndoVLM，一种基于超过34.8万例内镜检查数据（每例包含临床报告与对应图像集）预训练的视觉-语言基础模型。解决方案的关键在于三个创新机制：首先，采用解剖引导的稀疏池化（Anatomy-Guided Sparse Pooling）机制，以文本描述为查询驱动稀疏注意力，高效聚合冗余图像集中具有语义重要性的帧，生成特定解剖区域的视觉表征；其次，引入渐进式语义感知对齐（Progressive Semantic-Aware Alignment）策略，通过结构化软标签建模解剖位置与病理状态等临床分类体系，实现从患者层面全局匹配到病灶层面细粒度对齐的跨越；最后，仅对这些语义丰富的关键帧应用语义聚焦的掩码自编码器（Semantic-Concentrated Masked Autoencoder），融合低层视觉精度与高层语义鲁棒性。大量下游任务实验证明，EndoVLM显著优于现有基础模型，并具备与专用任务方法相媲美的性能，尤其展现出强大的零样本泛化能力，凸显其在临床场景中的广泛应用潜力。
链接: https://arxiv.org/abs/2608.04472
作者: Zhenyu Yi,Jianwei Xu,Yue Hu,Zhongwei Qiu,Sijing Li,Liang Huang,Bin Lv,Ling Zhang,Yingda Xia
机构: 未知
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:The development of foundation models (FMs) is crucial for advancing endoscopic image analysis. However, existing endoscopy FMs mainly rely on self-supervised learning from uni-modal images or videos, overlooking the rich semantic knowledge contained in clinical reports. Furthermore, effectively leveraging these records is hindered by a fundamental modality gap: structured anatomical descriptions are not naturally mapped to specific frames within the high-redundancy, uncurated visual streams. In this paper, we present EndoVLM, a novel vision-language FM pre-trained on over 348K endoscopic examinations, each pairing a clinical report with its corresponding image collection. An Anatomy-Guided Sparse Pooling mechanism utilizes textual descriptions as queries to drive sparse attention, efficiently aggregating semantically salient frames into anatomy-specific visual representations across redundant image-sets. Next, a Progressive Semantic-Aware Alignment strategy models clinical taxonomy (anatomy and pathological status) via structured soft targets, bridging the gap from global patient-level matching to fine-grained localized alignment. Finally, a Semantic-Concentrated Masked Autoencoder is applied exclusively to these semantic-rich frames, integrating low-level visual precision with robust high-level semantic representation. Extensive experiments across various downstream tasks demonstrate that EndoVLM outperforms existing foundation models and remains competitive with task-specific methods. Remarkably, EndoVLM also exhibits robust zero-shot generalization capabilities, highlighting its potential for broader clinical application.
[NLP-59] he Evaluator Is Part of the Experiment: Measuring Open-Ended LLM Conformity
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在开放式生成任务中对同伴答案（peer input）的从众行为（conformity）评估问题。传统研究多基于可验证标签下的离散答案切换（answer flips）进行度量，但此类方法不适用于开放式修订任务，因其答案质量为连续、隐含且评价存在主观偏差。本文提出一种新的实验协议，通过整合混合主语料库与独立构建的分解语料库，能够分离出常规重答、候选内容暴露、同伴呈现的残余效应以及评价者对可见同伴上下文的方向性敏感性。在四个开放权重生成器与三个基准测试中的实证结果表明：错误同伴输入在所有生成器-数据集组合中均导致最低质量的修订结果；对相同答案的盲评与明评存在显著差异——一名评价者倾向采纳同伴立场，两名则相反，一名基本中立，而GPT-4o和GPT-5.4-mini的审计也显示非中立性。此外，锚点审计揭示，简洁正确的锚点亦可能被误读，从而破坏潜在质量量表的稳定性，除非进行显式校准。研究得出四个核心结论：答案切换率不足以全面衡量开放式从众行为，错误同伴会损害开放式修订质量，评价者并非中立，且锚点校准至关重要。其关键解决方案在于引入可分离变量的实验设计与多维度评价框架，突破传统离散度量范式，实现对开放式从众行为更精细、可信的评估。
链接: https://arxiv.org/abs/2608.04463
作者: Alicia Guerra,Yibo Hu
机构: Illinois Institute of Technology (伊利诺伊理工学院)
类目: Computation and Language (cs.CL)
备注:
Abstract:Prior work on LLM conformity largely measures discrete answer flips under verifiable labels. Open-ended revisions require a different measurement strategy because answer quality is graded, latent, and judged imperfectly. We introduce an experimental protocol implemented across a pooled main peer-condition corpus and separately constructed decomposition corpora, allowing us to separate ordinary re-answering, candidate-content exposure, a bundled peer-presentation residual, and directional judge sensitivity to visible peer context. Across four open-weight generators and three benchmarks, all-wrong peer input produces the lowest-quality revisions in every generator-dataset cell. Blind and informed ratings of identical answers also differ by evaluator: one judge shifts toward the peer-endorsed position, two shift away, one is approximately neutral, and GPT-4o and GPT-5.4-mini audits are likewise non-neutral. Finally, an anchor audit shows that terse correct anchors can be misread often enough to destabilize the latent scale unless calibration is checked explicitly. These results support four conclusions: flip rates are insufficient as a complete measure of open-ended conformity, wrong peers harm open-ended revision, evaluators are not neutral, and anchor calibration is necessary.
[NLP-60] Q-CueGraph: Query-Conditioned Visual Evidence Graphs for Multimodal Reasoning
【速读】： 该论文旨在解决多模态大语言模型在处理图像时缺乏可靠的任务驱动观察策略的问题，即如何在高分辨率图像中高效、精准地选择需重点分析的区域。现有方法依赖像素级细节与裁剪/缩放工具虽能实现图像细粒度浏览，但未能提供基于任务需求的可解释性决策机制。其解决方案的关键在于提出Q-CueGraph框架，通过将问题（question）与图像表征映射为预算约束下的坐标级观测点，显式指导模型在有限资源下选择最相关的视觉区域。对于文本丰富的图像，采用可复用的OCR/布局图结构；对于自然图像，则在统一的选择、组合与预算接口下生成查询条件化的视觉节点。此外，引入可选的效用精炼机制，利用训练阶段答案正确性反馈来优化候选裁剪区域的选取，无需依赖区域边界框标注。实验表明，在仅使用约19%图像面积的情况下，Q-CueGraph在V*Bench上达到0.833准确率，显著优于全图推理的0.696；在InfographicVQA上以约一半图像面积实现全图92%的ANLS得分，验证了显式观察策略在证据局部化、问题定位敏感性及分辨率受限场景下的核心价值。
链接: https://arxiv.org/abs/2608.04452
作者: Pengcheng Pan,Xinfang Zhang
机构: 未知
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:High-resolution pixels and crop or zoom tools give multimodal large language models the ability to inspect an image, but they do not provide a reliable task-conditioned policy for deciding where to inspect. Q-CueGraph makes this decision explicit. It maps a question and an image representation to budgeted, coordinate-level observations for a frozen reader. Text-rich images use a reusable OCR/layout graph; natural-image search instantiates query-conditioned visual nodes behind the same selection, composition, and budgeting interface. Optional utility refinement learns which candidate crops the frozen reader can use from training-answer correctness, without region-box supervision. With a frozen Qwen2.5-VL-7B reader, Q-CueGraph reaches 0.833 accuracy on V*Bench versus 0.696 for full-image inference from a 19% image-area budget, and reaches 92% of full-image ANLS on InfographicVQA from about half the image area. Across six benchmarks, explicit observation is most valuable when evidence is localizable, the question discriminates its location, and resolution limits full-image reading.
[NLP-61] D2F-ReAG: Dynamic Decomposition and Filtering for Multi-Hop Reasoning -Augmented Generation
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在处理多跳问答（multi-hop questions）时因依赖静态内部知识而导致的推理不准确问题。尽管检索增强生成（Retrieval-Augmented Generation, RAG）在单跳查询中表现优异，但在需要跨文档推理的复杂多跳任务中仍面临效率与准确率不足的挑战。现有方法如图结构RAG或问题分解策略通常缺乏动态分解机制和有效的过滤手段，导致推理过程冗余且易受噪声干扰。为此，本文提出一种名为“动态分解与过滤的多跳推理增强生成”（Dynamic Decomposition and Filtering for Multi-Hop Reasoning-Augmented Generation, D2F-ReAG）的新范式，其核心在于通过判断初始层推理（root-level reasoning）的可靠性，自适应地控制推理深度：若初始推理可信，则直接生成答案；否则，将问题逻辑分解为子问题，并利用子问题推导出的验证性推理来修正初始推理。该方法有效提升了多跳问答的准确性与效率，在三个主流多跳基准测试上均取得了显著性能提升。
链接: https://arxiv.org/abs/2608.04444
作者: Jiaoyang Li,Junhao Ruan,Shengwei Tang,Kaiyan Chang,Zhengtao Yu,Tong Xiao,Jingbo Zhu
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) often generate inaccurate answers due to their reliance on static internal knowledge. Retrieval-augmented generation (RAG) addresses this limitation by integrating external knowledge and excelling at single-hop queries. However, it struggles with multi-hop questions that require cross-document reasoning. Existing methods, such as graph structured RAG or question decomposition, often lack dynamic decomposition and effective filtering, which leads to lower efficiency and accuracy. To overcome these limitations, we propose Dynamic Decomposition and Filtering for Multi-Hop Reasoning-Augmented Generation (D2F-ReAG), a novel paradigm that adaptively controls reasoning depth by judging the reliability of the root-level reasoning. If the root reasoning is reliable, the model directly generates the answer. Otherwise, the question is logically decomposed into sub-questions, and the verified reasoning derived from these sub-questions is used to refine the root reasoning. Experiments on three multi-hop benchmarks demonstrate the effectiveness of our method in handling complex multi-hop questions.
[NLP-62] MERaLiON-GR: Speech Gender Recognition Model for English and SEA Languages
【速读】： 该论文旨在解决跨语言环境下语音性别识别（Speech Gender Recognition）的准确性和泛化能力问题，特别是在英语及东南亚（SEA）多语言场景中的性能瓶颈。现有模型在非英语语种上表现受限，且缺乏对跨语言差异的有效建模能力。其解决方案的关键在于：基于大规模预训练的Conformer架构语音编码器MERaLiON-SpeechEncoder-2，采用参数高效微调技术低秩适应（LoRA）以最小化参数更新量并保留预训练知识，同时引入多尺度ECAPA-TDNN下游网络结合注意力池化与轻量级线性分类器，增强对语音特征的判别能力。实验结果表明，该方法在新加坡及东南亚八种语言（英语、汉语、马来语、泰米尔语、泰语、越南语、印尼语、高棉语）上均显著优于当前最优模型Vox-Profile和大型音频大模型（Audio-LLM），验证了专用语音模型在实现精准语音副语言理解与强跨语言泛化方面的关键价值。
链接: https://arxiv.org/abs/2608.04433
作者: Qiongqiong Wang,Ai Ti Aw,Nancy F. Chen,Ying Lay Chiu,Yang Ding,Yingxu He,Ridong Jiang,Zhuohan Liu,Yanfeng Lu,Yi Ma,Muhammad Huzaifah,Nabilah Binte Md Johan,Nattadaporn Lertcheva,Pham Minh Duc,Sailor Hardik Bhupendra,Siti Umairah Binte Mohammad Salleh,Shuo Sun,Tarun Kumar Vangani,Jeremy H. M. Wong,Jinyang Wu,Longyin Zhang
机构: Institute of Advanced Intelligence and Computing (IAIC), A*STAR, Singapore
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:We present MERaLiON-GR, a speech gender recognition system that performs binary classification (female / male) on English and Southeast Asian (SEA) languages. The model finetunes MERaLiON-SpeechEncoder-2, a large conformer based transformer pre-trained on a broad speech corpus, and applies parameter efficient fine-tuning via Low-Rank Adaptation (LoRA) to adapt the encoder to the gender recognition task, and appends a multi-scale ECAPA-TDNN down stream network with attention pooling and a lightweight linear classifier. Extensive evaluations across multilingual Singaporean and Southeast Asian languages (English, Chinese, Malay, Tamil, Thai, Vietnamese, Indonesian, and Khmer) show that MERaLiON-GR consistently surpasses the state-of-the-art gender recognition model Vox-Profile and a large Audio-LLM, in both full-utterance and segment level evaluation modes. The results underscore the value of dedicated speech models in achieving accurate paralinguistic understanding and strong cross-lingual generalization.
[NLP-63] Predict Then Retrieve: Cross-Instance Future-State Retrieval from Video Prefixes
【速读】： 该论文旨在解决生成式视频理解中跨实例未来状态预测与检索的挑战，即在给定一段视频前缀和关于某物体未来状态的时间性问题后，从其他视频或图像中检索出能够体现该未来状态的实例。不同于传统的动作预测（action anticipation）、事件定位（moment retrieval）或视频生成（video generation），该任务融合了未来状态的预测与跨模态、跨实例的检索能力，强调对未发生事件的语义与视觉状态的准确捕捉。其解决方案的关键在于提出一种名为LFTR（Lightweight Future Temporal Retriever）的轻量级检索框架，该框架采用冻结的编码器结构，通过学习一个受问题与时间跨度条件约束的未来潜在表示，并在互补的语义与视觉空间中进行匹配。实验表明，尽管当前所有模型（包括具备完整视频前缀信息的大规模多模态语言模型）仍远低于理想情况下的“理想基准”（oracle ceiling），但核心瓶颈并非感知能力不足，而是对未来状态的建模与推理能力缺失——即“预测”而非“感知”才是主要的学习难点。LFTR通过跨空间融合机制与硬负样本训练策略有效缩小了与理想性能的差距，且在推理成本上具有显著优势。
链接: https://arxiv.org/abs/2608.04426
作者: Quynh Vo,Thong Nguyen,Vinh-Hien Do,Cong-Duy Nguyen,Anh-Tuan Luu
机构: Centre for AI Research, VinUniversity; National University of Singapore
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: Work in progress
Abstract:We introduce Predictive State Retrieval (PSR), a task in which a model observes a short video prefix and a temporal question about an object’s future state, then retrieves instances from other videos or images that depict that state. Unlike action anticipation, which predicts a label, moment retrieval, which localizes an observed event within a video, or video generation, which synthesizes pixels, PSR combines anticipation with cross-instance retrieval across multiple temporal horizons. We construct a benchmark from four datasets with graded, human-validated ground truth, difficulty tiers, and an oracle ceiling. We also propose LFTR, a lightweight retriever with frozen encoders that predicts a question- and horizon-conditioned future latent and matches it in complementary semantic and visual spaces. A ceiling decomposition reveals a clear bottleneck: the true future state is highly retrievable once specified, whereas every predictor we evaluate, including a large multimodal language model with access to the prefix frames, remains far below the oracle. Thus, forecasting rather than perception is the central learnable challenge. LFTR narrows this gap at substantially lower inference cost, and ablations attribute its gains to cross-space fusion and hard-negative training rather than latent rollout. We release the benchmark, code, and evaluation scripts.
[NLP-64] Social Pressure Breaks Majority Voting in LLM Safety Panels
【速读】： 该论文旨在解决生成式 AI（Generative AI）在内容安全检测中因模型间共享误导性上下文信息而导致的误报率上升问题。其核心问题是：当多个语言模型在相同误导性上下文环境下进行集体判断时，尽管采用多数投票机制以期提升准确性，但这种协同决策反而可能放大错误倾向，导致系统整体误报率急剧升高。解决方案的关键在于识别并验证“共享社会线索”（shared social cues）这一潜在失效模式——即模型在看到其他模型给出错误标签或沉默后，会显著倾向于跟随“不安全”（unsafe）的推断（约75%的响应偏移），而对“安全”（safe）方向的引导反应较弱（仅17%），造成面板整体误报率从56.5%飙升至100%，且危害漏检率变化不大。研究通过受控双轮实验验证了该效应的严重性，并提出一种简单有效的预部署诊断方法，用于评估安全评审面板对误导性社会信号的敏感性。
链接: https://arxiv.org/abs/2608.04415
作者: Yibo Hu,Jiaming Qu
机构: Illinois Institute of Technology; Amazon
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) are increasingly used to detect unsafe content. A common approach is to combine judgments from a panel of models to correct individual mistakes, but this benefit may disappear when every model sees the same misleading context before voting. We study this risk in a controlled two-round experiment. Each model first judges an item alone, then judges it again after six simulated peers either assert the wrong label or abstain. We combine the final judgments by majority vote. Across six open-weight LLMs and six datasets, we find that the wrong-label peer message raises the average reviewer false-alarm rate from 56.5% under silent peers to 87.5%, and majority voting raises the panel false-alarm rate to 100%. Without an asserted label, the same panel outperforms its average member. The effect is strongly asymmetric: reviewers follow pushes toward “unsafe” far more than pushes toward “safe” (about 75% versus 17%), so the panel’s false-alarm rate rises sharply while its harmful-miss rate changes little. The proprietary-model probe shows substantial variation across models. These results identify susceptibility to shared social cues as a failure mode of safety panels and provide a simple pre-deployment diagnostic.
[NLP-65] MESH: Memory-Efficient Sinkhorn Optimization for Mixture-of-Experts Training
【速读】： 该论文旨在解决在混合专家模型（Mixture-of-Experts, MoE）训练中，采用内存高效的矩阵优化器（如Sinkhorn梯度下降）时出现的性能不稳定问题。尽管此类优化器通过移除AdamW中大部分优化器状态显著降低内存占用，但在MoE架构下其直接应用效果不佳，导致评估损失显著上升。研究发现，问题的关键在于路由后的专家矩阵（routed MoE expert matrices）：其梯度具有条件性、随时间动态变化，且无状态的Sinkhorn归一化无法有效处理这类梯度特性。为此，作者提出MESH（Hidden-Momentum Sinkhorn），一种针对MoE专家的隐式动量Sinkhorn更新机制，通过在梯度缓冲区生命周期中保留时间维度的一阶矩信号，实现对梯度的时序平滑，而无需显式存储专家一阶矩作为优化器状态。此外，MESH-B为可选的块/神经元预条件变体，引入粗粒度的逆均方根缩放因子。实验表明，时序平滑是提升性能的核心因素；预条件虽能进一步优化内存与性能权衡，但并非普遍必需。在多个种子测试中，MESH及MESH-B相较AdamW减少62.5%的优化器状态内存、降低约12.6%的PyTorch CUDA峰值分配，同时保持较小的评估损失差距，全状态诊断版本恢复至接近AdamW的性能水平，验证了MoE专家需要时序平滑而非完整坐标级的AdamW状态。
链接: https://arxiv.org/abs/2608.04407
作者: Masato Fujitake
机构: 未知
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 10 pages
Abstract:Memory-efficient matrix optimizers such as Sinkhorn gradient descent remove most AdamW optimizer state for dense Transformer matrices, but direct application to Mixture-of-Experts (MoE) training is unreliable. We study this failure in a controlled 110M-parameter nanowhale DeepSeek-style MoE pretraining setting. A SAGE/Sinkhorn hybrid reduces optimizer state from 0.883GB to 0.331GB but degrades evaluation loss to 3.8265, far above the AdamW baselines observed in the same setup (3.58–3.64 across the seeds we study). We show that routed MoE expert matrices are the dominant failure point: their gradients are conditional, temporally varying, and poorly served by stateless Sinkhorn normalization. We propose MESH, a hidden-momentum Sinkhorn update for MoE experts. MESH restores a temporal first-moment signal through the gradient-buffer lifecycle, without storing the expert first moment as optimizer state. MESH is an optional block-preconditioned variant that adds a coarse neuron/block inverse-RMS multiplier. Across ablations, temporal smoothing before matrix normalization is the primary causal ingredient; block/neuron preconditioning can improve the memory-quality frontier, but is not established as universally necessary. In two additional seeds, MESH and MESH-B reduce optimizer-state memory by 62.5% and peak PyTorch CUDA allocation by about 12.6% relative to AdamW, with a modest evaluation-loss gap. Full-state diagnostic variants recover AdamW-like performance in ablations, supporting the conclusion that MoE experts need temporal smoothing, but not necessarily full coordinate-wise AdamW state.
[NLP-66] raining-Free Hashing-Based Attention via Binary Principal Components ICML2026
【速读】： 该论文旨在解决长上下文大语言模型（Long-context Large Language Models, LLMs）在推理阶段因自注意力机制（Self-Attention）需重复处理不断增长的键值缓存（Key-Value Cache, KV Cache）而导致的计算效率瓶颈问题。现有稀疏注意力方法虽通过减少关注的KV对数量以降低计算量，但普遍存在准确率显著下降、需额外训练或依赖高成本哈希技术等缺陷。本文提出一种无需训练、数据感知的基于哈希的稀疏注意力方法——BinaryPC，其核心创新在于通过计算数据的二进制主成分（Binary Principal Components），构建紧凑的二进制哈希码及其对应的哈希函数，从而在不依赖梯度训练的前提下，显式保留数据的结构信息。与依赖数据无关的随机投影（如局部敏感哈希，Locality-Sensitive Hashing, LSH）或学习型非线性哈希方法不同，BinaryPC在生成哈希码时直接编码数据内在结构，实现高效且精准的近似注意力计算。大量实验表明，BinaryPC在多个模型架构和长上下文基准测试中，相较于全量注意力保持了相近的准确性，同时在各类稀疏及基于哈希的基线方法中表现最优；在现代GPU上，其端到端解码吞吐量相比FlashAttention内核提升达3.56倍。
链接: https://arxiv.org/abs/2608.04405
作者: Daohai Yu,Zhanpeng Zeng,Keyu Chen,Wenhao Li,Zhifeng Shen,Luxi Lin,Ruizhi Qiao,Xing Sun,Rongrong Ji
机构: 未知
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: ICML 2026
Abstract:Long-context large language models (LLMs) are increasingly deployed in real-world applications, yet self-attention remains a major efficiency bottleneck – especially during decoding – due to the necessity of repeatedly processing ever-growing key-value (KV) caches. Existing sparse attention reduce computation by attending to fewer KV pairs, but often suffer from substantial accuracy degradation, require additional training, or rely on expensive hashing. In this work, we present BinaryPC, a training-free, data-aware hashing-based sparse attention for long-context LLMs. BinaryPC constructs compact binary hash codes and corresponding hash function by computing binary principal components of data. Unlike Locality-Sensitive Hashing (LSH) with data-independent random projections or learned non-linear hashing methods, BinaryPC constructs binary codes that explicitly preserve the structural information of data without requiring gradient-based training. Comprehensive experiments across multiple model families and long-context benchmarks show that BinaryPC preserves accuracy relative to full attention while achieving superior performance among sparse and hashing-based baselines. On modern GPUs, BinaryPC improves end-to-end decoding throughput by 3.56 \times over the FlashAttention kernel. Our code is available at this https URL.
[NLP-67] NOLLI: A Difficulty-Calibrated Puzzle Benchmark for Diagnosing the English-Korean Performance Gap
【速读】： 该论文旨在诊断韩语模型在跨语言任务中性能差距的根源，解决的问题是：现有评估基准往往将“难度”简单等同于数据规模或输入长度，忽视了语言特性与认知复杂性之间的非线性关系。为此，作者提出NOLLI——一个基于程序生成的英-韩双语谜题基准，包含15种谜题类型（共25项任务、7,500个实例），所有实例均可通过种子重生成、确保唯一解且评分确定。其关键解决方案在于采用三层次设计：第一层为直接翻译对齐的任务，第二层通过对韩文字母（Hangul jamo，即音节成分）进行脚本改编以测试子音节层面的处理能力，第三层为仅限韩语的文化或正字法相关任务，以揭示语言特异性障碍。通过行为校准机制，调整各生成器使参考模型在目标准确率区间内稳定表现，从而实现难度的客观衡量。实验结果表明，在12个达到3%整体准确率门槛的模型中，英-韩任务间的准确率差异在±10个百分点内统计等效，说明仅因语言呈现方式导致的性能下降有限；但涉及书写系统密集型任务时出现显著差距，如“韩语密码”（Korean Cipher）任务相比英文版本落后高达68.7个百分点，而同一音节成分上的“数字符号谜题”（Cryptarithmetic）则无系统性惩罚，且音节组合（Jamo Composition）表现可预测密码任务表现，提示问题核心在于多步子音节操作的执行难度。此外，韩语专属任务揭示出规则应用缺陷方向不一，但“亲属称谓”任务普遍表现出正向优势，暗示文化知识储备的积极作用。最后，研究发现7类任务从易到难过程中结构尺寸并未持续增长，表明结构大小无法可靠反映实际经验难度。综上，该研究的关键在于构建了一个能精细区分语言表征、认知加工与文化背景影响的可重复、可验证、行为校准的评估框架，从而实现对性能差距成因的诊断性分析。
链接: https://arxiv.org/abs/2608.04397
作者: Dasol Choi,Joonyong Park,Daegon Yu,Soo Yong Kim,Youngsook Song,Seunghyeok Hong
机构: AIM Intelligence; KT Corp.; Sionic AI; Lablup Inc.; Hankuk University of Foreign Studies; HAERAE LAB
类目: Computation and Language (cs.CL)
备注:
Abstract:We introduce NOLLI, a procedurally generated English-Korean puzzle benchmark designed to diagnose where Korean performance gaps arise. It comprises 15 puzzle types (25 tasks; 7,500 items), with every instance seed-regenerable, verified to have a unique solution, and scored deterministically. Rather than equating harder with bigger, we calibrate difficulty behaviorally, tuning each generator until a fixed reference model lands in target accuracy bands. Its three-level design combines matched direct translations, script adaptations over Hangul jamo (sub-syllabic letters), and Korean-only tasks grounded in Korean culture or orthography. We evaluate 15 frontier, open-weight, and Korean-developed models; among the 12 above a 3% overall-accuracy floor, matched English-Korean accuracy is statistically equivalent within a +/- 10 pp margin (TOST), suggesting little cost from presentation language alone. Writing-system-intensive tasks show sharper gaps: Korean Cipher falls behind English by up to 68.7 pp, whereas Cryptarithmetic over the same jamo shows no systematic penalty, and Jamo Composition accuracy predicts Korean Cipher accuracy. These contrasts are diagnostic rather than causal, consistent with difficulty in multi-step sub-syllabic execution. Korean-only tasks separate rule-application deficits, which vary in sign, from a Kinship deficit positive in all 12. Finally, a salient size measure fails to grow from Easy to Hard in 7 of 15 types, making structural size an unreliable proxy for empirical difficulty.
[NLP-68] EdgeLM: Edge Demonstrations for Language Models Table Understanding
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在基于上下文学习（in-context learning）进行表格中心预测时，示范样本（demonstration）选择对模型性能的关键影响问题。现有检索方法通常优先选取与查询相似的示范样本，但此类样本往往强化模型已有倾向性预测，难以揭示复杂决策边界所需的差异信息。为此，论文提出EdgeLM，一种新型检索框架，其核心在于选择“边缘证据”（edge evidence），即既与查询相关又能够揭示决策边界的示范样本。EdgeLM通过两种互补形式的边缘证据实现这一目标：数据边缘（data edges），指邻近但真实标签不同的样本；模型边缘（model edges），指被部署模型先前误分类的相似样本。该方法无需模型重训练或任务特定工程设计，在五个数据清洗任务、十五个数据集及五种开源与专有大语言模型上均表现出最优或接近最优的性能，且消融实验表明两类边缘证据具有互补优势。
链接: https://arxiv.org/abs/2608.04390
作者: Soroush Omidvartehrani,Mohammadamin Habibollah,Mohammadreza Daviran,Davood Rafiei
机构: University of Alberta(阿尔伯塔大学)
类目: Computation and Language (cs.CL); Databases (cs.DB)
备注:
Abstract:Large language models (LLMs) perform table-centric prediction through in-context learning, making demonstration selection critical to performance. Existing retrieval methods prioritize similarity to the query, but similar demonstrations often reinforce the model’s likely prediction rather than reveal the distinctions needed for difficult decisions. We propose EdgeLM, a retrieval framework that instead selects edge evidence, demonstrations that are both relevant to the query and informative about the decision boundary. EdgeLM retrieves two complementary forms of edge evidence by selecting data edges, nearby examples with different ground-truth labels, and model edges, similar examples previously misclassified by the deployed model. EdgeLM requires neither model retraining nor task-specific engineering. Across five data wrangling tasks, fifteen datasets, and five open-weight and proprietary LLMs, EdgeLM consistently achieves the best or near-best performance in every setting, while ablations show that the two forms of edge evidence provide complementary benefits. Our code and datasets are publicly available at this https URL.
[NLP-69] FinReportBench: Measuring and Improving Institution-Grade Financial Report Generation
【速读】： 该论文旨在解决大语言模型生成的金融报告虽具备语言流畅性，但缺乏机构级交付适用性的核心问题。其关键在于构建一个基于专家评审的基准测试体系——FinReportBench，通过专家部分序（expert partial orders）、多模态证据与决策边界审计，提炼出涵盖可交付性、报告身份与机构完整性三个维度的35项评估标准。研究从10,000条中英文金融研究源数据中构建244个双语任务，实现对公开查询、重构研究轨迹与隐藏源数据包的分离设计，确保评估的客观性与可复现性。实验表明，当前模型在基本可交付性上已接近饱和，而报告身份与机构完整性仍是主要瓶颈，尤其体现在生成痕迹控制、信息密度与数据规范性方面。为此，论文提出基于基准引导的技能蒸馏方法，将常见失败模式转化为可复用的生成约束与自检机制，显著提升模型在G1与G2指标上的表现，同时保持基础能力（G0）不变，实现了从“生成”到“可靠生成”的范式跃迁。
链接: https://arxiv.org/abs/2608.04374
作者: Yinghao Tang,Tan Zhenwei,Yiyao Wang,Wanli Gu,Xiaolu Zhang,Jun Zhou,Wei Chen
机构: 1. Tsinghua University (清华大学); 2. Alibaba Group (阿里巴巴集团)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 9 pages, 9 figures
Abstract:Large language models can produce fluent financial analysis, but fluency alone does not establish whether a report is suitable for institutional delivery. We introduce FinReportBench, an expert-grounded benchmark for measuring and improving institution-grade financial report generation. Expert review reveals recurring gaps in report identity, institutional components, source discipline, and visual delivery. We derive a 35-item rubric through expert partial orders, multimodal evidence, and audits of decision boundaries, covering deliverability, report identity, and institutional completeness. Starting from 10,000 balanced Chinese and English financial-research source records, we curate 244 bilingual tasks across three research objects and two input tiers. Each task separates the public query, reconstructed research trajectory, and hidden source packet. Three independent judge families reproduce the expert partial order at near-ceiling rates, showing that bounded, observable criteria support reliable evaluation. Across nine model families, basic deliverability is nearly saturated, while report identity and institutional completeness remain the primary bottlenecks. The largest cross-model gaps concern generation-trace control, information density, and data discipline rather than basic report framing. We then use benchmark-guided skill distillation to turn recurrent failures into reusable generation and self-review constraints. Across five model families, the evolved skill improves mean G1 by 33.85 points and mean G2 by 13.83 points over paired no-skill runs while preserving G0 for every pair. Code and benchmark artifacts are available at this https URL.
[NLP-70] he Calibration Floor: Format Repair Can Masquerade as Self-Correction at Small-to-Mid Scale
【速读】： 该论文旨在解决生成式 AI（Generative AI）在语言模型自我修正（self-revision）过程中，准确率变化常被归因于推理能力提升这一主流解释的可靠性问题。研究指出，这种解释在答案提取边界（answer-extraction boundary）处可能失效，即准确率变化更多源于输出格式的可解析性改变，而非内容层面的推理改进。其解决方案的关键在于通过因果分解方法，将总准确率变化解构为三个组成部分：内容边际（content margin，双方答案均可解析）、格式恢复/丢失边际（format-recovery/loss margin，解析性发生变化）。研究通过在多个模型（Qwen3.5、Gemma-4-12B、Tencent Hy3、Nvidia Nemotron-3-Ultra-550B）及29个主实验单元与一个前沿验证臂上进行系统测试，发现格式效应显著超过内容效应（Wilcoxon p=1.7e−3）。为进一步验证因果关系，研究采用语法约束解码强制所有生成答案保持可解析性，结果表明在14个实验单元中，该策略平均消除了71%的原始总效应与内容边际之间的差距，且在两个单元中实现完全收敛，仅在最大效应单元中残留部分差异。此外，聚类分析揭示底层规模（floor-scale, 0.8B/2B）模型更易出现内容层面的负面变化，而高能力模型（capable-scale）虽具信号但缺乏调整空间。复现已有置信度门控协议未重现原报告增益，且内容边际接近零。对更大规模模型的前沿检验进一步显示，随着模型规模增加，格式主导性增强，内容边际在全部5个测试单元中精确为零，尽管总效应可达+0.275。最终，基于内容边际的校准下限准则揭示出一种“挤压”现象：底层模型有调整余地但信号不足，高能力模型有信号但无余量；仅一个单元勉强可行，且封闭保留测试集增益可忽略。研究结论表明，当前领域所测量的“自我修正”效果中，内容层面的贡献仅为少数。论文已公开工具、代码与衍生数据。
链接: https://arxiv.org/abs/2608.04355
作者: Mingguang Chen,Bo Qu,Licheng Wang
机构: DeepGrounding(深地基础); AlphaAvatar(阿尔法形象)
类目: Computation and Language (cs.CL)
备注: 36 pages, 5 figures
Abstract:Accuracy changes after language-model self-revision are usually interpreted as changes in reasoning. We show this can fail at the answer-extraction boundary, and test the failure causally rather than only observationally. Across Qwen3.5 (0.8B-9B), Gemma-4-12B, and two frontier models via API (Tencent Hy3, Nvidia Nemotron-3-Ultra-550B) in 29 primary cells plus a frontier arm, we decompose the always-revise accuracy shift into a content margin (both answers parseable) and format-recovery/loss margins (parseability changes). On 12 cells with meaningful unparseable-answer rates, format effects exceed content effects (Wilcoxon p=1.7e-3). To test this causally, we force already-generated reasoning through grammar-constrained decoding so every answer is parseable by construction: across 14 cells this closes a median 71% of the gap between the naive total effect and the content-margin estimate, with two cells converging exactly and a residual on the two largest-effect cells reported rather than dismissed. A clustered model confirms floor-scale (0.8B/2B) models have far higher odds of content-level change and harm than capable-scale models (p1e-7). Replicating a cited confidence-gating protocol verbatim on Qwen3.5 does not reproduce its reported gain and shows the same near-zero content margin. A frontier check on much larger models shows format-dominance intensifying with scale: content margin is exactly zero in all 5 cells despite total effects up to +0.275, though this arm is lower-powered. The calibration-floor criterion on the content margin reveals a squeeze: floor-scale cells have headroom but insufficient signal, capable-scale cells have signal but little headroom; only one cell is marginally viable, with negligible sealed-holdout gain. Content is a minority share of what the field has measured as self-correction. We release the instrument, code, and derived results.
[NLP-71] Equitable System-Prompt Selection via Constrained Mixed-Strategy GroupDRO
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在信息检索任务中，由于问题表述方式不同而导致回答质量差异显著的问题。尽管系统提示（System Prompt）被广泛用于引导模型输出行为，但现有方法通常仅针对平均性能进行优化，难以保证在所有问题变体下均能提供高质量响应，尤其在极端情况下的表现仍不理想。为此，论文提出一种基于约束混合策略的群体分布鲁棒优化（Constrained Mixed-Strategy GroupDRO）框架，用于系统提示的选择。其核心创新在于：不直接优化提示文本本身，而是对已有提示池中的多个系统提示分配权重，通过最小化跨评估指标与用户群体的最差情况信息质量损失（worst-case information-quality loss），同时约束平均损失接近传统平均选择策略的表现，从而实现鲁棒性与整体性能之间的平衡。该方法解耦了提示池生成与选择过程，可兼容任意提示池，并支持利用互补性提示的集成优势。在双语医疗与消费金融两个基准上的实验表明，该方法在五个主流大模型上平均将总体均值、最差25%均值及最差表现分别降低13.1%、13.2%和13.7%，同时保持整体质量与平均选择策略相当。多提示权重分析进一步揭示了不同指标-群体组合间的互补特性。
链接: https://arxiv.org/abs/2608.04339
作者: Mengyu Xu,Qiaoxin Yang,Zhihan Liu,Ruiyao Xu,Zachary Liu,Kezhen Chen,Chongyang Gao
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Machine Learning (stat.ML)
备注:
Abstract:Large language models are increasingly used for information seeking, yet semantically equivalent questions phrased in different ways can receive answers of considerably different quality. System prompts are widely employed to steer response behavior, but they are typically optimized for average-case quality, so some question phrasings may still receive incomplete or low-quality answers. To address this, we formulate a constrained mixed-strategy GroupDRO framework for system-prompt selection. Instead of optimizing the system-prompt text, the framework assigns weights to system prompts in an existing pool to minimize the worst-case information-quality loss across evaluation metrics and groups, while constraining the mean loss to stay close to that of average-based selection. Because pool generation and selection are decoupled, the method applies to any system-prompt pool and can leverage an ensemble of complementary system prompts rather than a single one. Across five LLMs on two bilingual medical and consumer-finance benchmarks, the constrained method reduces the Overall Mean, Worst 25% Mean, and Worst by 13.1%, 13.2%, and 13.7% on average relative to no mitigation while keeping overall quality close to Average selection. Its multi-prompt weights reveal complementarity across metric-group pairs. Code and data are available at this https URL.
[NLP-72] Right Reset: Chunking by Prefix Removal
【速读】： 该论文旨在解决在缺乏显式分隔符或结构信息的扁平化文本中，如何有效识别原始记录（record）边界的问题。其核心挑战在于，在经过渲染、光学字符识别（OCR）等处理后，文本的表面结构被破坏，传统基于规则或嵌入的方法难以准确恢复原始语义单元。论文提出的解决方案关键在于引入“右重置”（Right Reset, RR）机制，通过测量因果语言模型在移除左文（left context）后右侧隐藏状态轨迹的保留程度，挖掘模型内部对上下文依赖性的敏感性作为边界信号。该方法利用动态规划将RR得分转化为可变长度的文本块，从而在不依赖任务特定训练的前提下，显著提升边界检测性能——在扁平化文本上成功恢复47.7%的原始记录为干净单元，远超基于BGE嵌入的基准方法（25.9%）。此外，该方法对噪声和格式变化具有鲁棒性，且在多种语言模型中均表现出更少的局部输出扰动，表明其有效性源于对模型上下文依赖性的主动干预，而非单纯依赖表面特征或特定读出机制。
链接: https://arxiv.org/abs/2608.04330
作者: Mike Vegeto
机构: Independent Researcher(独立研究员)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 12 pages, 2 figures, 4 tables. Code, data, and reproduction materials: this https URL
Abstract:Removing the left context from a causal language model reveals a useful kind of boundary: an edge where the model processes the same right-hand tokens with little change. We turn this observation into prefix-removal probing and introduce Right Reset (RR), which measures preservation of the right-hand hidden-state trajectory. A dynamic program converts RR edge scores into variable-length chunks. On flattened text formed by concatenating topically similar records after deleting their separators and layout, RR recovers 47.7% of the original records as clean units, versus 25.9% for a BGE embedding boundary, the strongest tested conventional baseline without task-specific model training. The gain persists after rendering and OCR. Passive scores from the same Qwen3-4B layer and direct prompting of a same-scale instruction model perform substantially worse on flattened records. Across six language models, RR-selected cuts also undergo consistently less local output disruption than unselected candidate edges. An observed-token likelihood-ratio readout is competitive in some architectures, indicating that the central contribution is the intervention: context dependence itself can provide a boundary signal when surface structure is weak.
[NLP-73] DataRx: Missingness-Aware Sampling for Safer Large Language Model Task-Specific Fine-Tuning
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在特定任务微调过程中安全防护能力下降的问题。尽管引入安全数据是常见的缓解策略，但现有方法多采用随机混合安全样本，其有效性差异的内在机制尚不明确。本文提出DataRx，一种基于缺失感知的采样方法，其核心在于识别并选取能够填补模型安全能力“空白”的关键安全样本。其关键创新在于：不再依赖离散的词元（token）层面分析，而是利用高维隐藏表示来量化目标模型原始输出与安全参考响应之间的安全信号差距，从而精准筛选出具有强补全作用的安全样本。实验表明，在仅增加1%来自BeaverTails的安全样本情况下，DataRx将Llama3-8B-Instruct在七个下游任务上的平均攻击成功率从随机采样的59.23%显著降低至13.70%，且可与现有安全数据合成方法协同增强防御效果，为数据驱动的安全性保护研究提供了新范式。
链接: https://arxiv.org/abs/2608.04322
作者: Junbo Zhang,Qianli Zhou,Xinyang Deng,Wen Jiang
机构: Northwestern Polytechnical University(西北工业大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Task-specific fine-tuning can improve the performance of large language models (LLMs) on downstream tasks. However, our study reveals that task-specific fine-tuning can also weaken the safety guardrails of aligned LLMs. A widely adopted strategy for preserving safety during fine-tuning is to incorporate safety data. Although previous studies have shown that randomly mixing safety data can alleviate safety degradation, the underlying principle determining why some safety examples are more effective than others still remains unclear. In this paper, we propose DataRx, a missingness-aware sampling method for selecting safety-critical examples. DataRx is based on the hypothesis that a safety sample is more effective when the selected examples provide safety signals that fill the missing parts of LLMs’ safety capabilities. DataRx’s key insight is leveraging high-dimensional hidden representations rather than discrete tokens to quantify the safety signal gap between the target model’s native response and the safety reference response. The results show that, with only 1% additional safety samples from BeaverTails, DataRx reduces the average attack success rate of Llama3-8B-Instruct across seven downstream tasks from 59.23% under random sampling to 13.70%. In addition, DataRx can be combined with the existing safety data synthesis method to further enhance safety defenses during fine-tuning. We hope that DataRx will inspire more data-centric defense research.
[NLP-74] Pun Intended: Multi-Agent Translation of Wordplay with Contrastive Learning and Phonetic-Semantic Embeddings
【速读】： 该论文旨在解决跨语言双关语（pun）翻译中难以保留源语文本的语义歧义、语音相似性及幽默创造力的问题，传统机器翻译系统往往仅实现字面词汇的转换，而无法捕捉双关语的核心语言游戏特征。其解决方案的关键在于引入两种创新机制：一是基于音义联合嵌入（phonetic-semantic embeddings）的引导式推理流程，通过检索具备语音与语义双重匹配潜力的词汇候选，提升生成的合理性；二是多智能体（multi-agent）框架，通过多个专业化代理迭代评估并重生成候选译文，实现对语义忠实度、语音相似性与目标语言自然性的动态平衡。实验表明，尽管在BLEU和BERTScore等自动指标上提升有限，但该方法在人类专家评价中取得第一和第二名，验证了显式音义引导与迭代多智能体反馈机制在提升生成式双关语翻译质量方面的有效性。
链接: https://arxiv.org/abs/2608.04311
作者: Russell Taylor,Benjamin Herbert,Michael Sana
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Translating wordplay across languages has long challenged both professional translators and machine translation systems. We investigate three approaches to translating puns from English to French by combining large language models with linguistic constraints for wordplay generation. Our baseline uses a large language model with feedback from a discriminator prompted with positive and negative French examples. Our guided reasoning pipeline uses combined phonetic-semantic embeddings to retrieve lexical candidates for wordplay generation. Finally, our multi-agent framework iteratively evaluates and regenerates candidate translations using specialized feedback. Moving beyond literal translation, our objective is to preserve the linguistic creativity, ambiguity, and humor of the source-text wordplay rather than simply reproduce its vocabulary. The multi-agent and guided chain-of-thought systems ranked first and second, respectively, in the CLEF JOKER 2025 Task 2 competition under expert human evaluation, despite only modest improvements in BLEU and BERTScore. These findings suggest that both explicit phonetic-semantic guidance and iterative multi-agent evaluation can improve LLM-based wordplay translation relative to direct discriminator-guided generation, particularly when balancing semantic fidelity, phonetic similarity, and natural target-language expression
[NLP-75] Searching for Sound-Meaning Collisions: Graph-Based Affordance Retrieval and Multi-Evaluator Ranking for Pun Translation at CLEF 2026 JOKER Task 2
【速读】： 该论文旨在解决生成式语言处理中双关语（pun）翻译的难题，即如何在目标语言中实现与源语言双关语在音义关联上的等效性。传统方法依赖于寻找词汇对应关系，但该研究提出应转向探索音义之间的新结合点。其解决方案的关键在于将双关语翻译建模为一个发现、探索与选择的过程：首先通过检索系统在目标语言的语义和语音邻域中识别潜在的“音义桥梁”（affordances），即可能支持新双关语构建的音义关联；随后利用多个语言模型对这些机会进行生成性探索，产生多种候选译文；最后通过多视角生成-排序架构筛选最优结果。研究发现，生成器会主动利用检索到的音义桥梁，评估器则逐步聚焦于更强的音义关联，且当存在精确语音冲突时，其被选中的概率显著更高。然而，仍有许多双关语无法产生可用的音义桥梁，表明检索仍是当前计算双关语翻译的核心瓶颈。整体过程与Low提出的理论高度一致：成功的双关语翻译并非源于保留源语言词汇，而是通过在目标语言中发现新的音义碰撞点来实现。
链接: https://arxiv.org/abs/2608.04299
作者: Russell Taylor,Adam Brikman,Prateek Awate
机构: Georgia Institute of Technology(佐治亚理工学院)
类目: Computation and Language (cs.CL)
备注: CLEF 2026 Working Notes, 21-24 September 2026, Jena, Germany
Abstract:Fifteen years ago, Low proposed that pun translators should stop searching for equivalent words and instead search for new points of contact between sound and meaning. In this paper, we investigate that idea computationally. We model pun translation as a process of discovery, exploration, and selection. A retrieval system searches semantic and phonological neighborhoods for target-language affordances: sound-meaning bridges that may support new wordplay. Multiple language models then explore these opportunities by generating competing translations, while a multi-perspective generate-and-rank architecture selects among them. Beyond system development, our primary contribution is an analysis of how retrieved affordances propagate through the translation process. We find that generators actively exploit retrieved opportunities, evaluators progressively concentrate around stronger sound-meaning bridges, and exact phonological collisions are selected at disproportionately high rates when available. At the same time, many puns still yield no usable affordances, suggesting that retrieval remains the central bottleneck in computational pun translation. The resulting picture is remarkably close to the process envisioned by Low. Successful pun translation emerges not from preserving source-language words, but from discovering new places in the target language where sound and meaning collide.
[NLP-76] SafeCommit: Certifying When Memory-Grounded Agents May Safely Act NEURIPS
【速读】： 该论文旨在解决长时程智能体在使用持续记忆与外部工具执行具有副作用动作时所面临的“过早承诺”（premature commitment）问题，即在未验证记忆状态是否过期、冲突、不完整或被破坏的情况下便采取行动，从而导致不可控后果。其核心解决方案是提出SafeCommit机制——一种位于智能体推理与外部执行之间的风险可控层。该层通过整合记忆、观测、工具输出、溯源信息及策略约束，构建一个经过校准的潜在世界集合（plausible latent worlds），仅当符合性动作证书（conformal action certificate）证明该动作在所有保留的世界中均安全时，才允许执行；否则，选择低副作用探测动作以区分阻碍认证的世界，或返回保守回退策略。在世界覆盖校准的前提下，不安全认证执行的概率不超过预设风险水平α；而在世界生成不完美时，该界线可分离校准误差与表征误差。实验通过一个无依赖的受控模拟器验证了安全-效用权衡，并支持一键复现全部结果，最终目标是为智能体提供不仅“做什么”的决策，更关键的是“何时具备足够证据以安全执行”的判断依据。
链接: https://arxiv.org/abs/2608.04289
作者: Mayur Akewar,Ravi Ranjan
机构: Florida International University (佛罗里达国际大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 14 pages, 6 tables, and 1 figure, target NeurIPS
Abstract:Long-horizon agents increasingly use persistent memory and tools to take actions with external side effects. A central failure mode is premature commitment: an agent acts before resolving whether its memory grounding is stale, conflicting, incomplete, or corrupted. We formalize this problem as safe commitment under memory uncertainty and introduce SafeCommit, a risk controlled layer between agent reasoning and external execution. The layer constructs a calibrated set of plausible latent worlds from memory, observations, tool outputs, provenance, and policy constraints. It permits a side effectful action only when a conformal action certificate shows that the action is safe in every retained world. Otherwise, it selects a low-side-effect probe that targets the worlds blocking certification, or returns a conservative fallback. Under calibrated world coverage, the probability of an unsafe certified commit is at most the target level \alpha; with imperfect world proposal, the bound separates calibration and representation error. A dependency-free controlled simulator illustrates the safety-utility tradeoff and reproduces all reported results with one command. The goal is to offer a concrete approach for deciding not only what an agent should do, but when the available evidence is sufficient to safely do it.
[NLP-77] Eliciting Intrinsic Hallucinations in LLM s via Semantically Equivalent Adversarial Attacks
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在使用外部知识源进行增强生成时，仍存在内在幻觉（intrinsic hallucination）的问题，即模型在生成内容时产生与检索证据不符的虚构信息。尽管现有方法如检索增强生成（Retrieval-Augmented Generation, RAG）可提升事实准确性，但模型对查询语义微小变化的敏感性导致其在上下文一致性方面依然脆弱。论文提出一种新型评估框架，通过对抗优化方法生成自然且语义等价的查询扰动，以压力测试模型在保持语义不变前提下的鲁棒性，并引入严格的语义等价约束与内在幻觉目标函数。该框架在白盒、灰盒和黑盒攻击场景下对多种攻击技术进行评估，覆盖5个开源与5个闭源生成模型及3个数据集。实验表明，即使最先进的模型（如GPT-5-mini）也极易受语义保持型扰动影响，上下文忠实度最高下降达50%。研究揭示了当前模型依赖于表面查询形式而缺乏稳健的上下文锚定能力，因而亟需设计具备独立于查询表达形式的强健接地机制（robust grounding）的架构与训练目标。
链接: https://arxiv.org/abs/2608.04286
作者: Atri Vivek Sharma,Brian Formento,Alessio Lomuscio
机构: Imperial College London (帝国理工学院); Safe Intelligence (安全智能)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: To be presented at COLM 2026
Abstract:Large language models (LLMs) are often used in conjunction with external knowledge sources to improve their factual accuracy and decrease hallucinations, through methods such as Retrieval-Augmented Generation (RAG). However, these systems remain susceptible to intrinsic hallucinations, where the model generates unfaithful or fabricated information that is not supported by the retrieved evidence. We propose a novel framework to assess model robustness against this phenomenon by stress-testing using natural, semantically equivalent variations of a user query found via adversarial optimization methods. We apply our framework, which enforces strict semantic equivalence constraints and an intrinsic hallucination objective, to a range of adversarial attack techniques across white-box, gray-box, and black-box adversarial settings. Evaluating these attacks on 5 open-source and 5 closed-source generator models across 3 datasets, we demonstrate that even state-of-the-art models are highly susceptible to meaning-preserving perturbations, which significantly degrade contextual faithfulness (by up to 50% for GPT-5-mini). Our findings indicate that faithful use of in-context evidence remains fragile even in state-of-the-art LLMs, motivating architectures and training objectives that enforce robust grounding independent of surface query form. Code is available at: this https URL
[NLP-78] LLM -based Vulnerability Discovery in Business Process Documentation
【速读】： 该论文旨在解决业务流程中因逻辑缺陷导致的质量问题、延迟及成本增加等风险，核心挑战在于从以自然语言为主的业务流程文档（如ISO 9000/9001体系文件、用户手册、操作指南和流程执行日志）中准确识别与表征业务逻辑（Business Logic, BL）漏洞。由于这些文档通常存在要求冲突、表述模糊、度量规范无效、质量检查缺失或实际执行与规范不一致等问题，且需依赖领域知识、面对复杂的流程结构和海量信息，传统方法难以高效处理。本文提出的MIRABELLE系统通过融合生成式AI（Generative AI）与机器学习（Machine Learning, ML）技术，将非结构化的业务流程文本转化为可计算的带属性图表示，进而结合图分析与形式化逻辑方法检测潜在漏洞。其解决方案的关键在于利用大型语言模型（Large Language Models, LLMs）在多个关键环节中的能力——包括对短语表达中的语法与技术错误识别、完整流程结构恢复与关键业务逻辑要素（如操作序列、决策节点、输入输出资源）提取——显著提升从非结构化文档中自动化挖掘业务逻辑漏洞的效率与准确性。
链接: https://arxiv.org/abs/2608.04271
作者: Ben Falchuk,Himanshu Garg,Euthimios Panagos,Sioan Zohar
机构: 未知
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL)
备注:
Abstract:Just like software and hardware, business processes are susceptible to vulnerabilities that can lead to product quality issues, delays, and increased costs. Business process vulnerabilities can arise from a variety of sources, including conflicting requirements, ambiguous documentation, invalid measurement spec-ifications, omission of quality checks, or implementations that differ from speci-fications. MIRABELLE is a system that identifies and characterizes business logic (BL) vulnerabilities from available business process representations, in-cluding ISO 9000/9001 documentation, user guides, work instructions, and pro-cess execution logs. MIRABELLE leverages recent advances in AI/ML to pro-cess available business process documentation and generate attributed graph rep-resentations of the business logic that can be processed using both graph and for-mal logic approaches for identifying potential vulnerabilities. However, extract-ing the business logic (e.g., operation execution sequences, decisions, input/out-put resources) from mostly natural language artifacts is challenging due to the required domain expertise, inherent process complexity, and the sometimes very large volumes of information. This paper focuses on our experimentation with Large Language Models (LLMs) and their role within MIRABELLE. We report on the performance of several LLMs across vital stages of vulnerability detection, from grammatical and technical error-flagging in short phrasings, to complete process structure recovery and extraction.
[NLP-79] he Fairness Collapse Phenomenon: Bias Amplification in Language Models Trained on Synthetic Data
【速读】： 该论文旨在解决生成式语言模型在持续预训练过程中因暴露于合成数据而引发的公平性退化问题，即“公平性坍塌”（fairness collapse）现象。其核心问题是：在模型持续使用自身生成的合成数据进行迭代训练时，是否会导致已有社会偏见被放大，从而引发隐蔽但严重的公平性下降。解决方案的关键在于通过构建受控实验环境，在重复使用合成数据（基于Bias in Bios数据集）训练语言模型的过程中，系统性地监测模型在标准语言建模指标与公平性评估之间的差异。研究发现，公平性退化在模型性能显著下降之前便已显现，表明偏见的增强可能在传统模型崩溃指标出现前悄然发生。这一结果揭示了合成数据污染对语言模型公平性的潜在风险——偏见的累积具有隐蔽性，亟需在模型训练中引入专门的公平性监控机制。
链接: https://arxiv.org/abs/2608.04268
作者: Irina Proskurina,Antoine Gourru,Julien Velcin
机构: Laboratoire Hubert Curien, UMR CNRS 5516, Saint-Étienne, France; Université Claude Bernard Lyon 1, Université Lumière Lyon 2, ERIC; École Centrale de Lyon, LIRIS, CNRS UMR 5205
类目: Computation and Language (cs.CL)
备注:
Abstract:Generative models trained on artificially generated data have been shown to exhibit model collapse, resulting in significant performance degradation. As synthetic content increasingly contaminates the training corpora of language models, this raises critical concerns about the use of open data in continued pretraining. Although previous work has demonstrated model collapse in language models, it remains unclear whether exposure to synthetic data amplifies or attenuates the social biases already present in pretrained models. Because language models are known to reproduce and amplify demographic stereotypes, recursive training on self-generated data may create a self-reinforcing feedback loop in which biased associations become progressively stronger across generations. We call this hypothesized phenomenon fairness collapse. In this work, we construct controlled training regimes in which models are repeatedly trained on synthetic data using the Bias in Bios dataset. Across experiments, we observe a consistent and concerning pattern: fairness degradation emerges before substantial degradation is reflected by standard language-modeling metrics. This result highlights a critical risk associated with synthetic data contamination in language model training: bias can increase silently before strong indicators of model collapse become apparent.
[NLP-80] owards End-to-End Multilingual Metaphor Processing: Integrating Detection Translation and Evaluation
【速读】： 该论文旨在解决多语言自然语言处理中隐喻语言理解与翻译的挑战，核心问题在于现有研究将隐喻检测、机器翻译及翻译评估视为独立任务，缺乏整合性计算框架。其解决方案的关键在于构建一个端到端的多语言隐喻处理统一框架，通过三个互补方向实现：（1）跨语言的鲁棒隐喻检测；（2）面向隐喻的翻译评估，涵盖人工评价与自动质量估计；（3）隐喻检测与翻译评估的联合建模。研究将融合语言学理论与大语言模型技术，开发新型数据集、标注方法、评估基准与自动化评价方法，最终实现对含修辞语言的多语言NLP系统在开发与评估层面的整体提升。
链接: https://arxiv.org/abs/2608.04260
作者: Jiahui Liang,Lifeng Han
机构: Leiden University (莱顿大学); Leiden University Medical Centre (莱顿大学医学中心)
类目: Computation and Language (cs.CL)
备注: Scientific report on PhD thesis plans and milestones achieved (current progress)
Abstract:Metaphorical language remains a major challenge for multilingual natural language processing because successful interpretation and translation require reasoning beyond literal lexical meaning. Existing research has largely investigated metaphor detection, machine translation, and translation evaluation as separate tasks, while little work has explored how these components can be integrated into a unified computational framework. This PhD proposal aims to develop an end-to-end framework for multilingual metaphor processing consisting of three complementary research directions: (1) robust metaphor detection across languages, (2) metaphor-oriented translation evaluation for both human assessment and automatic quality estimation, and (3) joint modelling that connects metaphor detection with translation evaluation. The proposed research will combine linguistic theory with recent advances in large language models to develop new datasets, annotation methodologies, evaluation benchmarks, and automatic evaluation approaches for metaphor-aware machine translation. The expected outcome is a unified framework that improves both the development and evaluation of multilingual NLP systems when processing figurative language.
[NLP-81] SIGNPOST-Bench: Benchmarking Text-Vision Conflict Resolution in Multimodal Large Language Models
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）在现实场景中进行视觉-文本联合推理时，当视觉与文本信息发生冲突时如何权衡二者证据的问题。现有评估基准普遍缺乏对这种冲突情境下模型决策机制的揭示。为此，作者提出SIGNPOST-Bench——一个受控的反事实评估基准，专门用于评测模型在文本与视觉信息冲突时的推理能力。其核心解决方案在于构建包含原始图像（Original）、空白图像（Blank）、相似图像（Similar）、随机图像（Random）和对抗性图像（Adversarial）五类变体的反事实五元组，通过合成的、局部化的场景-文本干预，在保持非文本内容不变的前提下，精确测量模型定位性能的变化，并引导预测结果向由冲突文本引入的目标地理坐标偏移。该基准涵盖来自四个数据集的5,111个反事实组共25,555张图像变体，评估了来自七家机构的20个MLLM。实验结果显示，相较于原始图像，对抗性变体导致中位定位误差从282公里上升至1,347公里（提升4.8倍），且在可地理编码的对抗样本中，6.5%–20.1%的预测结果距离注入目标不足50公里，所有模型均表现出从空白图像到对抗图像的平均目标距离显著减小，表明其倾向于响应冲突文本。此外，一致、无关及冲突文本替换对模型输出产生显著不同的影响，而干净输入下的定位性能无法充分预测模型对冲突文本的鲁棒性。这些结果确立了视觉地理定位作为连续诊断工具，可用于量化分析MLLM在多模态证据冲突下的仲裁机制，为评估模型在复杂真实场景中的可靠性提供了可控且可量化的框架。
链接: https://arxiv.org/abs/2608.04244
作者: Sirun Li,Minghao Liu,Ling Dai,Yong Li,Haoxin Lyu,Junting Zhou,Fan Zhang
机构: 未知
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 27 pages, 25 figures
Abstract:Multimodal large language models (MLLMs) make grounded predictions in real-world scenes by combining visual and textual cues, yet existing benchmarks rarely reveal how they arbitrate between these evidence sources when they conflict. We introduce SIGNPOST-Bench, a controlled counterfactual benchmark for evaluating text-vision conflict resolution. Each source image is transformed into a counterfactual quintuplet of Original, Blank, Similar, Random, and Adversarial variants. Synthetic, localized scene-text interventions are designed to preserve non-textual content, enabling paired measurements of changes in localization performance and directed shifts toward geographic targets introduced by conflicting text. SIGNPOST-Bench contains 5,111 counterfactual groups and 25,555 image variants from four datasets. We evaluate 20 MLLMs from seven providers. Compared with Original images, Adversarial variants raise median localization error from 282 km to 1,347 km, a 4.8-fold increase. Among geocodable adversarial samples, 6.5-20.1% of predictions lie less than 50 km from the injected target across models, and every evaluated model exhibits a positive mean paired reduction in target distance from Blank to Adversarial. Compatible, unrelated, and conflicting text replacements produce distinct effects on model predictions, while clean-input localization performance does not fully predict robustness to conflicting text. These results establish visual geolocation as a continuous diagnostic of scene-text arbitration and provide a controlled framework for evaluating how MLLMs resolve conflicting multimodal evidence.
[NLP-82] Hallucinations on the Board: Tool-Augmented Evaluation of LLM Chess Commentary
【速读】： 该论文旨在解决生成式AI在国际象棋等专业领域中虽能提供超人类水平的评估结果，却缺乏自然语言解释，导致其知识难以被专家与非专家有效理解的问题。现有大语言模型（LLM）虽有潜力弥补这一空白，但因领域知识有限而频繁产生事实性幻觉，且传统基于参考文本或以大模型为裁判的评估框架无法可靠识别此类错误。为此，论文提出ACT-Eval评估框架，其核心在于将棋局评论分解为原子化陈述（atomic claims），并通过引擎支持工具与专家标注的黄金标准参考进行联合验证，从而系统评估事实正确性、概念覆盖度及走法质量判断。该框架发布了包含325个典型局面-着法对的基准数据集，涵盖教学、比赛与关键局面，并引入五类错误分类体系。实验表明，即使在使用工具增强后，主流闭源与开源模型仍普遍存在事实性幻觉问题：未使用工具的GPT-5.4在22.0%的情况下出现子命题错误，部分小型开源模型错误率甚至超过40%；尽管工具辅助显著提升了事实准确性和走法评价能力，但对专家级战略与战术思想的覆盖仍严重不足。人机校准结果显示，ACT-Eval在事实判断上与人类共识范围一致，且其覆盖度评分与人类对战略完整性评价高度相关，验证了该框架的有效性与可信赖性。
链接: https://arxiv.org/abs/2608.04240
作者: S. Ashwin Hebbar,Peiyao Sheng,Sewoong Oh,Pramod Viswanath
机构: Princeton University (普林斯顿大学); Sentient Labs (Sentient实验室); University of Washington (华盛顿大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 23 pages, 6 figures
Abstract:Superhuman game engines in domains like chess have made expert-level evaluations easily accessible, yet they communicate what is true without the natural-language explanations that make such expertise educationally useful to experts and non-experts alike. Large language models could, in principle, bridge this gap, but they frequently hallucinate due to limited domain-specific knowledge, and standard reference-based or LLM-as-a-judge frameworks cannot reliably detect these errors. In this work, we present ACT-Eval, an evaluation framework that decomposes chess commentary into atomic claims and routes them to engine-supported tools and expert-annotated gold references to assess factual correctness, conceptual coverage, and move-quality judgment. We release a benchmark of 325 position–move pairs spanning pedagogical, tournament, and critical positions, including 125 positions with expert-verified gold atoms and a five-class error taxonomy. Evaluating leading proprietary and open-weight models, we find that factual hallucinations remain pervasive in chess commentary: GPT-5.4 without tools produces incorrect sub-claims 22.0% of the time, while smaller open-weight models exceed 40%. Although tool augmentation substantially improves factual correctness and move-quality assessment, coverage of expert strategic and tactical ideas remains limited across all models. Human calibration shows that ACT-Eval’s factual judgments fall within the observed range of inter-human agreement, while its coverage scores correlate strongly with human assessments of strategic completeness.
[NLP-83] Patients-like-me: A Variational LM–GNN Framework for Explainable Clinical Prediction
【速读】： 该论文旨在解决现有语言模型（Language Models, LMs）在电子健康记录（Electronic Health Records, EHRs）建模中存在患者序列孤立编码且可解释性差的问题，同时克服图神经网络（Graph Neural Networks, GNNs）对高质量患者表示依赖的局限。其核心解决方案是提出一种统一的LM-GNN框架Patients-like-me（PLM），通过联合建模局部患者语义与全局队列结构，实现更优的表征学习。关键创新在于引入变分期望最大化（Variational Expectation-Maximization）算法，在监督变分目标下交替更新语言模型与图神经网络参数，从而高效融合局部上下文信息与跨患者关系。实验结果表明，PLM在MIMIC-III和MIMIC-IV数据集上持续优于当前最优方法，且性能提升在仅使用编码器或解码器型语言模型骨干网络时均具泛化性；此外，该框架可通过检索具有影响力的相似患者提供可解释性分析，边缘遮蔽实验进一步验证了排名靠前参考患者的预测贡献度最高。
链接: https://arxiv.org/abs/2608.04193
作者: Xinyu Wang,Yixuan Li,Hanwei Wu,Qincheng Lu,Chi-Kuang Yeh,Xiao-Wen Chang,Ziyang Song
机构: McGill University(麦吉尔大学); Université de Montréal(蒙特利尔大学); Georgia State University(佐治亚州立大学); Ohio University(俄亥俄大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Language models (LMs) offer strong textual representations for electronic health records (EHRs), but they encode patient sequences in isolation and provide limited explainability. Graph neural networks (GNNs) complement LMs by incorporating inter-patient relationships and enabling reference-patient attribution, yet they rely on high-quality patient representations. We propose Patients-like-me (PLM), a unified LM–GNN framework that integrates local patient semantics with global cohort structure. To train PLM efficiently, we introduce a Variational Expectation-Maximization algorithm that alternates LM and GNN updates under a supervised variational objective. Extensive experiments on MIMIC-III and MIMIC-IV show that PLM consistently outperforms state-of-the-art methods, with improvements generalizing across encoder-only and decoder-only LM backbones. These gains are achieved with only modest additional computational overhead. PLM also provides reference-patient explanations by retrieving influential similar patients, while edge-masking experiments confirm that the highest-ranked references have the greatest impact on model predictions.
[NLP-84] Behavioral Skill Reconstruction: Reconstructing Hidden Functionality from LLM Agent Skills
【速读】： 该论文旨在解决闭源智能体技能（closed source agent skills）在功能层面的隐蔽性问题，即尽管其底层代码、脚本、常量和数据等文件被严格隐藏，用户仍可能通过正常交互行为重构出该技能的功能。现有研究主要关注提示注入攻击（prompt injection attacks）导致的直接文件泄露，防御策略也多集中于阻止此类显式信息暴露。然而，该论文指出，即使文件未被披露，攻击者仍可通过合法请求与响应序列，基于行为模式逆向推导出技能的核心逻辑。为此，作者提出行为技能重构（Behavioral Skill Reconstruction, BSR）范式，并设计了名为SkillClone的黑盒攻击方法：通过分析技能公开描述形成接口假设，发起结构化良性探测（structured benign probes），合成可执行副本，并利用与目标技能的差异验证进行迭代修复。实验在30个涵盖规则、表格、流程及算法类技能上验证了该方法的有效性，在保留输入上实现了精确或部分功能恢复，且多次查询显著弥补单轮重建的遗漏。由于整个过程仅依赖合法交互，传统以防止信息泄露为核心的防御机制覆盖有限，而更简略的技能描述亦无法提供有效保护。研究结果表明，仅保证文件保密不足以实现功能保密，必须同时控制日常使用中累积的信息泄露。
链接: https://arxiv.org/abs/2608.04192
作者: Peichun Hua,Haoxuan Xu,Mengyuan Li
机构: University of Southern California (南加州大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 20 pages, 5 figures, 20 tables
Abstract:Closed source agent skills may encode proprietary instructions, scripts, constants, and data. Providers may offer their capabilities as services while keeping the underlying packages hidden. Prior work focuses on prompt injection attacks that directly disclose these artifacts, and existing defenses accordingly aim to prevent such leakage. However, preventing file disclosure does not prevent users from recovering the functionality those files implement. This raises a fundamental question: can a user reconstruct a skill’s functionality through ordinary use while its files remain hidden? We study behavioral skill reconstruction (BSR), in which an attacker uses valid task requests and observed responses to build a functional clone of a hidden skill. We introduce SkillClone, a black-box attack that clones a target skill by forming an interface hypothesis from its public advertisement, issuing structured benign probes, synthesizing an executable replica, and iteratively repairing it through differential validation against the victim skill. Across 30 skills spanning rules, tables, procedures, and algorithms, SkillClone achieves exact or partial recovery on held-out inputs for several targets. Iterative requerying closes gaps missed by single-round reconstruction. Because SkillClone uses only legitimate interactions, disclosure-focused defenses provide limited coverage, and less detailed skill descriptions offer limited protection. These results show that file secrecy alone does not ensure functional secrecy. Defenses must also limit cumulative information leakage from ordinary use. Comments: 20 pages, 5 figures, 20 tables Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) Cite as: arXiv:2608.04192 [cs.CR] (or arXiv:2608.04192v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2608.04192 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-85] Large Language Models for Low-Resource Languages: A Conceptual Framework for an Electronic Explanatory Dictionary of the Tajik Language
【速读】： 该论文旨在解决塔吉克语缺乏功能完备的数字化词典资源这一关键问题，尤其针对低资源语言在自然语言处理（NLP）技术应用中的短板。其核心挑战在于如何在标注数据有限、语言形态高度复杂（塔吉克语为黏着语，具有高形态变异性）的条件下，构建一个兼具传统词典学方法与现代生成式AI能力的综合性电子解释性词典。解决方案的关键在于提出首个整合形态分析、词形还原、语义聚类与基于大语言模型（LLM）的词条自动生成于一体的系统化概念架构；同时采用子词分词（subword tokenization）策略以应对形态多样性，并结合参数高效微调（PEFT）技术，在有限标注数据下实现模型的有效适配。该框架不仅为塔吉克语构建了可扩展的数字词典基础，还为机器翻译、自动摘要、情感分析等下游NLP任务提供了核心语义资源，具有重要的方法论意义与实践价值。
链接: https://arxiv.org/abs/2608.04186
作者: Mullosharaf K. Arabov
机构: Kazan Federal University (喀山联邦大学); Tajik National University (塔吉克国立大学)
类目: Computation and Language (cs.CL)
备注: Preprint
Abstract:This paper presents a conceptual framework for developing an electronic explanatory dictionary of the Tajik language using large language models (LLMs). The relevance of the work stems from the absence of a comprehensive digital lexicographic resource for Tajik that is comparable in functionality to dictionaries for high-resource languages, and from the limited adaptation of modern natural language processing technologies to low-resource language systems. Based on a systematic survey of existing linguistic, statistical, and corpus resources, we propose a dictionary architecture that integrates modules for morphological analysis, lemmatization, semantic clustering, and dictionary entry generation using LLMs. The choice of subword tokenization is justified by the agglutinative nature of Tajik morphology and its high morphological variability, along with a parameter-efficient fine-tuning (PEFT) strategy suitable for limited annotated data. The novelty of the work lies in proposing the first holistic conceptual architecture of an explanatory dictionary for Tajik that unifies classical lexicographic methods, language statistics, and generative capabilities of LLMs into a single system. The practical significance of the study is the formation of a methodological foundation for developing a full-featured electronic dictionary that can serve both as a lexicographic tool and as a core resource for machine translation, automatic summarization, sentiment analysis, and other applied NLP tasks. The paper is intended for specialists in computational linguistics, lexicography, and developers of natural language processing systems working with low-resource languages.
[NLP-86] st then Route: How Language Models Execute In-Context Conditional Rules Across Models and Languages
【速读】： 该论文旨在探究语言模型在遵循上下文条件规则（如“若P(x)则A否则B”）时，其内部表征是否以模块化方式组织——即是否存在一个独立的谓词判断模块与一个独立的答案路由模块。研究的关键在于通过四捐赠者激活修补（activation patching）设计，人为制造条件与答案词汇之间的不一致，从而探测每一层神经元所承载的信息类型。实验结果表明，在三个来自两个不同架构的开源模型及六种共享同一物品库的语言中，中间堆栈的残差带（residual band）精确地编码了谓词真值信息：对该区域进行修补可实现近1.0的预测-结果翻转率和接近0.0的映射翻转率，满足严格的隔离标准（17/18单元达标），且该定位在五类谓词中均保持一致；而路由机制则表现出相反特征。进一步发现，一个学习到的子空间可在训练过的条件-答案对之间近乎完美地交换输出，但跨对迁移能力极低（所有模型中约为0），仅在Gemma-3-4B中跨语言迁移达约0.98。所有测试结果均显示，路由路径具有明显的词元绑定性且不可迁移（在Gemma中主要为答案读出，在Qwen中为成对特异性），而非抽象的、可泛化的路由模块。因此，结论是：谓词判断具有模块化特性，而实际决策路径并非模块化。
链接: https://arxiv.org/abs/2608.04183
作者: Luxshan Thavarasa,Sivasuthan Sukumar
机构: University of Moratuwa (莫鲁图瓦大学)
类目: Computation and Language (cs.CL)
备注: 19 pages, 16 figures. Code and data: this https URL
Abstract:When a language model follows an in-context conditional rule such as “if P(x) then A else B,” does it assemble a runtime circuit with one module that tests the predicate and another that routes the answer? We probe this with activation patching under a four-donor design whose two swapped-rule donors make the condition and the answer word disagree, so each layer reveals which of the two it carries. Across three open models from two families and six languages sharing one fixed item bank, a mid-stack residual band carries the predicate’s truth value: patching it reroutes the answer with predicate-outcome flip near 1.0 and mapping flip near 0.0, meeting a strict pre-specified isolation criterion in 17 of 18 cells, and the same localization holds across five predicate families. The router shows the opposite profile. A learned subspace flips A and B near-perfectly within the trained pair yet transfers to a new pair at approximately 0 in every model, while in Gemma-3-4B (the only model probed cross-lingually) it transfers at approximately 0.98 to the same pair in other languages. Under every probe we ran, the router direction is token-bound and non-transferable (largely answer-readout in Gemma, pair-specific in Qwen) rather than an abstract routing module. Test is modular; under these probes, route is not.
[NLP-87] Visualizing Graph-to-Answer Mechanism Recovery in Materials-Science Hypothesis Generation
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在材料科学领域生成假设时存在的“表面流畅性”问题，即模型输出虽语言通顺，但未必遵循科学机制的内在逻辑。其核心挑战在于如何识别和验证生成过程中的因果机制是否真实存在。解决方案的关键在于构建一个基于图结构到答案的机制追踪（mechanism-tracing）诊断框架，针对经过微调的Graph-PRefLexOR-8B（基于Qwen3-8B）模型，系统解析其从头脑风暴、图结构构建、模式提取到合成阶段的多阶段推理路径。通过引入语义回溯、图结构破坏实验、基于激活的恢复度量以及逐层逐标记区域的网格化分析，形成可视化诊断工作流。实证表明，在100个开放性材料科学问题中，最终答案仍高度依赖模型自身的结构化阶段，尤其体现在合成阶段；而在图结构破坏测试中，机制恢复主要集中在后期合成与答案起始区域（约第30至36层），而早期过渡区（第7–10层）几乎无恢复能力。该工作流可为科研人员及模型开发者提供关键干预节点，以在实验规划前识别生成假设中机制支持的丢失或重建位置。
链接: https://arxiv.org/abs/2608.04170
作者: Shashwat Sourav,Subhadeep Pal,Markus J. Buehler,Sanjay Das,Fiona Y. Wang,Dominik Soos,Tirthankar Ghosal
机构: Washington University in St. Louis(圣路易斯华盛顿大学); Oak Ridge National Laboratory(橡树岭国家实验室); Lawrence Berkeley National Laboratory(劳伦斯伯克利国家实验室); UniverseTBD; Massachusetts Institute of Technology(麻省理工学院); Old Dominion University(老多明尼昂大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:AI co-scientists can generate fluent materials-science hypotheses, but fluency does not show that an answer preserves a scientifically meaningful mechanism. We present a graph-to-answer mechanism-tracing case study for Graph-PRefLexOR-8B, a Qwen3-8B model adapted to expose distinct stages for brainstorming, graph construction, pattern extraction, and synthesis. We organize semantic backtracking, graph corruption, activation-based recovery measurements, and layer-by-token-region grids into a visual diagnostic workflow for inspecting this pathway. Across 100 open-ended materials-science questions, final answers remain closest to the model’s own structured stages, especially synthesis. Under graph corruption, a full sweep over 37 residual-stream checkpoints, the embedding output and 36 transformer blocks, shows little mechanism recovery in the earlier transition region at layers 7–10, recovery instead concentrates in late synthesis and answer-start regions around layers 30 and 36. The workflow is intended to help scientists and model developers identify where a generated hypothesis loses or regains mechanism support before it is passed to downstream experimental planning.
[NLP-88] Mind the Cap: Output-Budget Regimes Change the Measured Multilingual Reasoning Gap
【速读】： 该论文旨在解决多语言评估中因固定输出令牌（token）上限导致的不公平比较问题，指出不同语言在表达相同语义内容时所需令牌数存在差异，而当前普遍采用的单一定额令牌上限（output-token cap）实际上是一个未被明确控制的隐含实验变量，可能误导对模型跨语言性能的真实判断。其核心解决方案在于揭示并纠正这一“令牌预算偏差”（token-budget artifact），通过系统性地在不同令牌预算下测试Qwen3-8B与Llama-3.1-8B-Instruct在德语、泰语和斯瓦希里语上的原生生成（native）与翻译生成（translate）性能差距，发现该差距随预算变化可波动高达57个百分点；进一步引入长度归一化（length normalization）后，性能差异仍可达38.9个百分点，并在严格预算约束下甚至导致最优提示策略的排名反转。研究通过三组“冻结测试”（frozen tests）验证结果稳健性：在固定1024令牌预算下，六个经过Holm校正的假设检验均拒绝原假设，表明性能差异非偶然；当原生准确率已饱和时，剩余差异体现为提示策略本身性能差距，而非推理能力缺陷。此外，研究发现通过交叉拟合的泰语词汇扩展仅在低截断率（19%）时带来4.9点提升，而在固定预算下“宣布预算”（announced budget）的影响达5.1点，说明准确性不仅取决于强制执行的上限，还受提示策略中预算声明的影响。最终，基于长上限运行计算出的“正确发射时间一致性”（correct-emission timing identity）能精准匹配预设的MGSM性能峰值（误差±0.65点），并在额外基准上实现对保留样本的高精度预测（误差±0.92点）。因此，论文的关键结论是：应将输出令牌上限视为独立变量，在整个预算范围内报告性能表现，而非仅在单一预算点进行评价，以确保多语言评估的公平性与有效性。
链接: https://arxiv.org/abs/2608.04160
作者: Ankit Goyal,Jaideep Ray
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 15 pages, 2 figures, 11 tables. Under review
Abstract:Multilingual evaluations report accuracy at a single output-token cap, but languages need different numbers of tokens to express the same content, so the cap is a hidden experimental variable. We test whether the native-vs-translate gap on MGSM (German, Thai, Swahili) is a token-budget artifact for Qwen3-8B and Llama-3.1-8B-Instruct under four prompting strategies. The measured gap swings by up to 57 points across budgets, length normalization moves it by up to 38.9 points where the cap binds, and at tight caps normalization can reverse which strategy scores higher. We prospectively froze the sweep’s three Qwen peaks and its near-zero value at 1024 and evaluated them on 540,000 independently hard-capped decodes: a second frozen family of six Holm-corrected tests rejects every null. The frozen test at B^*=1024 still fails to reject because native accuracy has already saturated there; above saturation, the residual difference is a strategy-performance gap, not an identified reasoning deficit. The same truncation channel prices a cost-ordered adaptation ladder: a cross-fitted Thai vocabulary extension closes 0.0 points of the gap at the frozen budget and 4.9 points where 19% of traces still truncate. A third frozen family varies only the announced budget at a fixed enforced cap; announcing 128 rather than 2048 tokens moves Thai native accuracy by 5.1 points, so accuracy is not a function of the enforced cap alone. A correct-emission timing identity computed from one long-cap run matches the three pre-specified MGSM peaks to 0.65 points and, in an exploratory Qwen-only analysis of three further benchmarks, tracks held-out items to 0.92 points, locating the peak exactly in five of seven cells. Treat the output cap as an independent variable and report accuracy across the budget regime, not at a single budget.
[NLP-89] GEB-Bench: Abstract Structures Told in Many Voices
【速读】： 该论文旨在解决生成式模型在跨模态抽象能力上的根本性缺陷，即模型能否识别并理解不同表达形式中共享的深层结构（如自指、怪圈、莫比乌斯环等），并实现从一种表现形式到另一种的准确映射。其核心挑战在于：尽管模型在单一模态内对结构性特征具有一定的识别能力，但在跨模态迁移时表现出显著的“结构映射断裂”。解决方案的关键是构建GEB-Bench这一基准测试体系，它以抽象结构母题（abstract structural motif）为核心单元，通过自然场景、民间故事、数学定理与程序骨架等多种“声音”（voice）来表征同一结构，并将表面参数视为干扰变量不予评分。该设计形成一个小型跨模态范畴，任务即为探索各模态间结构的对应关系。实验结果揭示出一种规律性的抽象失败现象：所有模型在单模态结构识别上表现良好，但跨模态映射能力普遍薄弱，且这种差距在不同厂商的前沿模型中均存在，甚至导致不同模型趋向于产生相同的错误答案；同时，表面复杂性对所有模型构成持续压力，仅模型容量能提供有限缓解而非真正免疫。这表明当前生成式模型在深层结构理解与跨模态泛化方面仍存在系统性瓶颈。
链接: https://arxiv.org/abs/2608.04111
作者: Tong Zhang,Zhiyuan Shi,Yun Peng,Tao Xie
机构: Fudan University (复旦大学); Peking University (北京大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Logic in Computer Science (cs.LO)
备注:
Abstract:Can a model look at a river delta and a lightning bolt and see that they share a structure? We introduce GEB-Bench, a benchmark whose unit is an abstract structural motif–self-reference, a strange loop, a Mobius twist–in the spirit of Godel, Escher, Bach. Each motif is told in several voices: a natural scene whose composition is the structure, a folk story whose telling enacts it through a mechanically checkable form device, a mathematical theorem, and a programmatic skeleton; surface parameters are declared nuisance variables and never scored. Motifs, voices, and the structural changes between them form a small cross-modal category, and GEB-Bench’s tasks are its questions. Evaluating twelve open and proprietary models, we find that abstraction failure is lawful. The central finding is a gap between recognition and cross-voice mapping: models identify a structure within one voice far better than they carry it across voices; every model pays this tax, and mapping strong enough to narrow it appears only at the frontier tier. Two patterns support it. Errors align more strongly with the designed formal geometry than with measured perceptual geometries, and frontier models from different vendors converge on the same wrong answers; and surface complexity taxes every model that reads structure, with capacity buying headroom rather than immunity. GEB-Bench is fully generative and released with its pipeline.
[NLP-90] FinPerMA: A Theory-Informed Event-Grounded Personalized-Memory Benchmark for LLM Agents
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）代理在长期高风险场景中（如金融咨询）维持并动态更新个性化用户模型的能力问题。现有个性化记忆评测基准多聚焦于事实性记忆保持或依赖弱约束的模型生成轨迹，忽视了事件驱动型偏好适应这一关键维度。为此，作者提出FinPerMA——一个基于真实事件的基准测试框架，通过冻结的长期投资者行为轨迹评估代理对个性化记忆的建模能力。其核心解决方案在于构建了一个融合确定性理论驱动影响规则、受控的LLM叙事生成与自动化质量筛选的合成数据生成流程，并引入“冲击后检查点”（Post-Shock checkpoint）以检验代理是否将重大事件有效整合进持久的用户模型中。实验结果显示，在2,994个问题、276个角色的测试集上，七种前沿大模型及多种记忆配置均未达到性能饱和，整体准确率最高仅约0.47，多选题正确率不足39%。进一步归因分析表明，基于摘要的记忆机制虽能保留事实细节，却丢失了实现个性化所需的关键偏好信号，导致简单检索策略反而优于专门设计的记忆系统，且在事件冲击后性能差距进一步扩大。因此，该研究揭示了当前主流记忆架构在长期个性化建模中的根本性局限。
链接: https://arxiv.org/abs/2608.04095
作者: Ben Wang,Kang Zhou,Lifan Guo,Feng Chen,Chi Zhang
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large language model (LLM) agents are increasingly used as personalized assistants in high-stakes domains such as financial advising, yet it remains unclear whether they can maintain and update an individualized user model over long horizons. Existing personalized-memory benchmarks primarily test factual retention or rely on weakly constrained model-generated trajectories, leaving event-driven preference adaptation underexplored. We introduce FinPerMA, an event-grounded benchmark that evaluates personalized memory against frozen longitudinal investor trajectories. Its generation pipeline combines deterministic, theory-informed impact rules, controlled LLM narration, and automated quality screening; a Post-Shock checkpoint isolates whether an agent has integrated a material event into its persistent user model. On 2,994 questions from 276 personas, seven frontier LLMs and up to seven memory configurations remain far from saturated: no full-context configuration exceeds approximately 0.47 overall accuracy or approximately 39% on multiple-choice questions. Attribution analysis shows that summary-based memory often preserves factual details while losing the preference signals needed for personalization; simple retrieval can therefore outperform purpose-built memory systems, with the gap widening after shocks.
[NLP-91] SpecDrop: Parameter-Free Category-Conditioned Routing for Modular Specialization
【速读】： 该论文旨在解决混合专家模型（Mixture-of-Experts, MoE）中路由机制在匹配总参数量下性能不如无路由基线的问题，核心疑问在于：性能瓶颈是源于路由算法本身，还是训练信号粒度与目标类别之间的对齐不充分。为探究此问题，作者提出SpecDrop——一种固定、无学习参数的路由方案，其通过为每个专家分配特定类别权重 pa，并在非所属类别时引入微小泄漏 pi（pi>0），以类别无关的固定分母进行融合，无需学习路由参数或辅助损失，仅在推理阶段需类别标签。在单超类标签任务（如ResNet-110上的CIFAR-100、ViT-S/16上的ImageNet-1K）中，SpecDrop分别达到79.23%和79.89%的准确率，显著优于同等参数量的无路由基线（在CIFAR-100上高出4.75%，在ImageNet-1K上高出6.53%）。这些提升量化了通过路由实现类别监督所带来的收益，但若直接对密集模型输出施加标签掩码，其精度更高（分别为85.2%和83.7%），说明路由优势并非来自标签感知本身，而在于将标签信息转化为训练过程中可学习的模块化结构：在CIFAR-100和ImageNet-1K上分别实现58%和100%的专家-类别对齐，并带来0.00和+1.06的掩码增益，表明输出空间限制已在训练中内部化。而在模糊划分场景（如基于30M Transformer的SlimPajama-6B语言建模、基于LoRA的Llama-3.2-1B指令微调），路由机制退化为与无路由控制相当的结果，仅受种子噪声影响，符合本文提出的“粒度对齐”假说。因此，研究结论指出：路由有效性关键不在于算法选择，而在于训练信号粒度与类别划分的对齐程度。
链接: https://arxiv.org/abs/2608.04084
作者: Boyao Wang,Zhihan Lei
机构: Carnegie Mellon University (卡内基梅隆大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 35 pages, 6 figures. Code: this https URL
Abstract:Mixture-of-experts (MoE) networks pursue specialization through learned routers, gates, and load-balancing losses, yet at matched total-parameter budgets learned routers can underperform equal-weight No-Routing baselines. Is the bottleneck the routing algorithm, or the alignment between training-signal granularity and the target categories? We probe the question with SpecDrop, a fixed parameter-free routing scheme: each of K branches receives weight p_a for its assigned category and a small leakage p_i 0 otherwise, merged through a category-independent fixed denominator, with no learned routing parameters and no auxiliary losses; the category label is required at inference. On vision tasks where each image has one superclass label (CIFAR-100 on ResNet-110; ImageNet-1K on ViT-S/16), SpecDrop reaches 79.23% on CIFAR-100 and 79.89% on ImageNet-1K, exceeding parameter-matched baselines that do not use the label (+4.75 over dense on CIFAR-100; +6.53 over the No-Routing+SE control on ImageNet-1K). These gains quantify what category supervision buys when deployed through routing – not an advantage over label-aware deployments of the baselines: given the same label, masking a dense model’s outputs is stronger for accuracy alone (85.2 / 83.7). SpecDrop’s contribution is converting the label into trained-in modular structure: 58%/100% branch-category alignment, and masking gains of 0.00 (CIFAR) / +1.06 (ImageNet) – the output-space restriction is largely internalized during training. On fuzzy partitions, where training units span multiple categories (SlimPajama-6B language modeling with a 30M Transformer; SuperNI instruction tuning over Llama-3.2-1B with LoRA), the routing mechanism reduces to the matched No-Routing controls within seed noise, the null our thesis predicts. Granularity alignment, not algorithm choice, localizes when routing helps. Code: this https URL
[NLP-92] FinProBench: Evaluating Financial AI Agents with Role-Grounded Rubrics Derived from Professional Deliverables
【速读】： 该论文旨在解决现有金融领域人工智能代理（AI agent）评估标准与真实职业实践脱节的问题。当前的评分体系多依赖任务提示（prompt）或模型输出来构建评价标准，忽略了专业人士在实际工作成果中体现的隐性专业规范。为此，研究提出FinProBench基准测试平台及角色根基型评分体系构建方法（Role-Grounded Rubric Construction, RGRC），其核心在于通过分析同一角色专业人士的真实工作产出（deliverables）来提取并生成可复用的评分标准。RGRC包含四个阶段：成果收集、能力要素提取、评分体系合成与验证，能够捕捉隐性专业知识、区分不同质量层级，并在同角色内跨任务迁移。实验表明，在具有丰富先验知识的传统角色中，仅基于提示的评分体系表现接近RGRC；而在先验知识匮乏的角色专业化任务中，RGRC显著优于传统方法（99.1% vs. 78.0%）。这一差异揭示了当模型先验充分覆盖行业惯例时，提示工程可近似实现高质量评估，但面对超出先验范围的专业标准时，必须依赖真实从业者成果进行角色根基的评估体系构建。FinProBench涵盖57个职业、8个金融子行业和161种成果类型，共1,723份精选成果，释放了包含20个完整任务的初始评估集，覆盖7个子行业中的20个角色。在使用异构大语言模型（LLM）评判者和角色级评分标准下，人类产出平均得分最高（73.7/100），各系统间置信区间重叠且优势互补。更重要的是，基于角色复用评分体系可使每项任务的评分构建成本降低6.7倍，显著提升评估效率。
链接: https://arxiv.org/abs/2608.04077
作者: Ben Wang,Kang Zhou,Lifan Guo,Feng Chen,Chi Zhang
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Evaluating financial AI agents requires criteria aligned with real professional work. Existing rubric methods typically derive criteria from task prompts or model outputs, overlooking tacit standards visible only in practitioner deliverables. We introduce FinProBench, a benchmark for professional financial tasks, and Role-Grounded Rubric Construction (RGRC), a reusable pipeline that derives rubrics from deliverables produced by practitioners in the same role. RGRC comprises four stages: Deliverable Collection, Competency Extraction, Rubric Synthesis, and Validation. Its rubrics capture tacit standards, distinguish quality levels, and transfer across tasks within a role. Before analysis, we classified 57 occupations by deliverable genre into 30 prior-rich conventional roles and 27 prior-sparse role-specialized roles. Across all roles, Prompt-only nearly matches RGRC for conventional roles (89.2% vs. 90.7%), but RGRC substantially outperforms it for role-specialized roles (99.1% vs. 78.0%). This split indicates that prompt engineering can approximate rubrics when conventions are well represented in model priors, while professional grounding is essential for standards beyond those priors. FinProBench is built from 1,723 curated deliverables spanning 57 occupations, 8 financial sub-industries, and 161 deliverable types, and releases an initial evaluation set of 20 complete tasks covering 20 roles in 7 sub-industries. With heterogeneous LLM judges and role-level rubrics, human deliverables rank first on average (73.7 vs. 70.3, 70.2, and 69.6 out of 100), while all four systems show overlapping 95% confidence intervals and complementary strengths. Reusing rubrics at the role level reduces estimated per-task construction effort by 6.7 times relative to authoring each rubric from scratch.
[NLP-93] Learning Sexism Detection Using Multi-Agent Perspectivist Preference Optimization EACL2027 ACL
【速读】： 该论文旨在解决文本性别歧视标注中人类标注者之间存在显著主观差异的问题，传统自然语言处理（NLP）系统通常通过多数投票方式强制统一标注结果，从而忽略了这种合法且合理的感知多样性。其解决方案的关键在于提出多智能体视角偏好优化（Multi-Agent Perspectivist Preference Optimization, MAP-PO）框架，该框架通过聚类分析标注者的行为模式而非人口统计学特征，为每个聚类训练一个专属的大型语言模型（Large Language Model, LLM）代理，并利用结合个体奖励与团队级奖励的偏好优化机制协调各代理之间的行为。实验在英文和西班牙语的EXIST 2024微博数据集上进行，结果显示：未经微调的代理表现趋同，表明必须进行聚类特异性训练；而仅基于本聚类标签训练会导致代理偏离其对应聚类的标注行为，引入共享的团队级训练信号则能有效保持各代理对所属聚类的准确表征，确保其既忠实于群体观点又协同达成整体一致性。
链接: https://arxiv.org/abs/2608.04056
作者: Hadi Mohammadi,Tina Shahedi,Robert A. Bagheri,Mehdi Dastani,Masoume M. Raeissi
机构: Utrecht University (乌得勒支大学); Wageningen University Research (瓦赫宁根大学与研究中心)
类目: Computation and Language (cs.CL); Computers and Society (cs.CY); Machine Learning (cs.LG)
备注: 17 pages, 12 figures, 14 tables. Preprint; under review at EACL 2027 (ACL Rolling Review, August 2026 cycle). Code and data: this https URL
Abstract:When people label text for sexism, they often disagree, and not because some of them are wrong: they genuinely perceive sexism differently. Most NLP systems discard this disagreement by collapsing it into a majority vote. We propose the Multi-Agent Perspectivist Preference Optimization (MAP-PO) framework to keep these different perspectives. On the EXIST 2024 dataset of labeled English and Spanish tweets, we first cluster annotators by their labeling behavior rather than their demographic attributes. We then fine-tune one Large Language Model agent per cluster to reproduce that cluster’s annotation behavior, and coordinate the agents with preference optimization that combines individual and team-level rewards. We evaluate MAP-PO in four settings defined by two languages and two backbone language models, asking whether each agent reproduces the annotations of its own cluster and whether the agents together reproduce the majority label. Two findings hold in all four settings. First, without fine-tuning the agents behave almost identically, so cluster-specific training is necessary. Second, we show that training each agent only on the labels of its own cluster pushes the agents far beyond the clusters they should represent, while adding a shared team-level training signal consistently keeps each agent calibrated to its cluster.
[NLP-94] Modality Agreement- and Conflict-Aware Prototype Hypergraph Learning for Multimodal Intent Understanding
【速读】： 该论文旨在解决多模态意图识别中模态间不一致信息的建模问题，即如何有效捕捉并利用文本、语音和视觉模态在语义层面的分歧（conflict），而非简单地抑制或忽略这些不一致。传统融合方法通常强调模态对齐或将不一致性视为需消除的噪声，但忽略了其在特定意图（如讽刺、嘲讽）中的判别性价值。本文提出MACH（Modality Agreement- and Conflict-aware prototype Hypergraph）框架，其核心创新在于构建一个分层原型超图结构，将模态间的一致性（agreement）与冲突性（conflict）分别建模为独立且可复用的语义关系结构。关键机制包括：1）通过稀疏的共识原型超图捕获跨模态的共享语义模式；2）设计独立的冲突路径，将跨模态差异映射至专用的冲突原型超图；3）采用特征级、样本自适应的仲裁机制融合双路径输出，实现对有判别力的不一致信息的保留与偶然噪声的抑制。此外，引入渐进式优化策略以稳定层级结构，确保联合学习的稳定性。实验结果表明，该方法在基准数据集上显著优于现有模型，组件分析进一步验证了分层组合、原型驱动的语义精炼及一致性-冲突仲裁机制的有效性。
链接: https://arxiv.org/abs/2608.04054
作者: Mohnish Raj,Suraj Kumar,Soumi Chattopadhayay,Chandranath Adak,Ayan Dutta
机构: IIT Indore (印度理工学院印多尔); IIT Indore (印度理工学院印多尔); IIT Indore (印度理工学院印多尔); IIT Indore (印度理工学院印多尔); IIT Indore (印度理工学院印多尔)
类目: Multimedia (cs.MM); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Multimodal intent recognition requires understanding not only what textual, acoustic, and visual signals share, but also how they disagree. Such disagreement is frequently class-informative; for example, lexical positivity accompanied by incongruent vocal or facial behavior may indicate sarcasm or taunting, yet most fusion methods either encourage modality alignment or treat inconsistency as uncertainty to be suppressed. We propose MACH (Modality Agreement- and Conflict-aware prototype Hypergraph), a hierarchical prototype-hypergraph framework that represents multimodal agreement and conflict as distinct, recurring relational structures. MACH progressively composes unimodal representations into bimodal and trimodal abstractions. At each applicable level, modality-composition anchors activate sparse agreement prototype hypergraphs that capture reusable consensus patterns, while a separate conflict pathway maps cross-modal discrepancies to dedicated conflict prototype hypergraphs. The two pathways are combined through a feature-wise, sample-adaptive arbitration mechanism, enabling the model to preserve informative disagreement while suppressing incidental modality noise. A progressive optimization strategy stabilizes the interdependent hierarchy before joint agreement-conflict learning. Experiments on benchmark datasets demonstrate the effectiveness of the proposed formulation, while component and robustness analyses validate the distinct roles of hierarchical composition, prototype-mediated semantic refinement, and agreement-conflict arbitration.
[NLP-95] When More Becomes Less: Position-Dependent Repetition Effects in Language Models
【速读】： 该论文旨在解决生成式语言模型中基于填空（cloze-style）探针评估目标词预测概率时所隐含的一个关键假设：即目标词出现次数的增加对预测结果的影响与读取位置（readout position）无关。研究发现，这一假设在实践中并不成立。其解决方案的关键在于提出一种双探针设计（two-probe design），通过固定重复目标词的前缀并仅改变读取位置来系统性地考察位置效应：相邻探针（adjacent probe）将读取槽置于重复块之后，而位移探针（displaced probe）则将其置于新的句子框架内部。实验结果显示，相邻重复呈现典型的启动效应（priming effect），预测概率随重复次数 N 上升并趋于饱和；而位移重复则呈现出倒U型曲线，预测概率先上升至峰值后下降。该倒U型模式在全部13个开源编码器与解码器模型中均显著存在（每词下降量的置信区间不包含零），并在西班牙语、中文、德语和法语的42个跨语言测试单元中完全可复现。六条件因果消融分析表明，该现象源于精确词汇重复本身，而非文本长度、泛化冗余或语义邻近暴露等混淆因素。进一步的框架-语用控制排除了读取框架本身作为人为偏差的可能性。内部机制分析显示，在因果语言模型（causal LM）中，每个目标词的注意力权重随 N 增加而下降，但分配给重复块的总注意力预算却随之增长；而在掩码语言模型（masked LM）中则无此现象。因此，任何试图通过改变重复次数来测量模型行为的研究，均不能将读取位置视为与测量内容正交的变量。
链接: https://arxiv.org/abs/2608.04021
作者: Han-yu Wang
机构: The University of Hong Kong
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Cloze-style probes that vary how often a target token appears implicitly assume that more copies of a target affect prediction the same way regardless of where the readout slot sits. We show this assumption fails. Our two-probe design holds a repeated-target prefix fixed and varies only the readout position: the adjacent probe places the slot immediately after the repeated block; the displaced probe places it inside a fresh sentence frame. Adjacent repetition behaves as priming intuition predicts: P(\texttarget) climbs with N and plateaus. Displaced repetition produces an inverted-U: P(\texttarget) rises to an early peak and then declines as more copies are added. The displaced inverted-U shows a per-word drop with bootstrap CI excluding zero in all 13 open-access encoder and decoder models we test, and replicates across Spanish, Chinese, German, and French in 42 of 42 multilingual cells. A six-condition causal ablation isolates the effect to exact lexical repetition rather than length, generic redundancy, or semantic-neighbour exposure. A frame-pragmatics control rules out an artefact of the readout frame. Internally, per-target-token attention falls with N while the total budget assigned to the repeated block grows in causal LMs but not in the masked LM we probe. Probes that vary repetition count cannot treat the readout position as orthogonal to what they measure.
[NLP-96] ransfer Learning for Named Entity Recognition of Classical Latin through LLM Prompting
【速读】： 该论文旨在解决古典拉丁语（Classical Latin）中命名实体识别（Named Entity Recognition, NER）的挑战，尤其是在资源稀缺背景下如何有效利用现代大型语言模型（LLM）技术提升古语处理性能。其核心问题在于：如何在标注数据有限、语言特性与现代语言差异显著的情况下，实现高精度的实体识别。解决方案的关键在于通过提示工程（prompt engineering）充分调动商业级大模型（如gemini-2.5-pro和claude-sonnet-4-5）的跨语言迁移学习能力，使低资源的古典拉丁语能够受益于主流大模型在多种现代语言上积累的知识与泛化能力。实验表明，该方法在粗粒度（11类）与细粒度（28类）NER任务中均取得最优表现，且在严格与宽松评估模式下均排名第一，验证了跨语言迁移学习在古语言处理中的有效性与潜力。
链接: https://arxiv.org/abs/2608.04015
作者: Callum Chan
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:With the increase in digitized resources of Classical Latin texts and modern breakthroughs of Large Language Models (LLMs), I contribute to ancient language research by participating in EvaLatin 2026. This paper describes Team uOttawa’s system description and results for the Named Entity Recognition (NER) shared task. The task is divided into two subtasks: coarse-grained NER with 11 classes and fine-grained NER with 28 classes, each evaluated under strict and fuzzy regimes. Through prompt engineering of commercial LLMs gemini-2.5-pro and claude-sonnet-4-5, I show that the underrepresented ancient Latin language can take advantage of cross-lingual transfer learning by using advancements made by the wider LLM development community. Overall, the methods discussed in this report demonstrate very strong results, placing first in both NER subtasks and achieving the best scores across all evaluation metrics and regimes among all submissions.
[NLP-97] Wiring Beats Blending: What Transfers Between Transformer Sizes – and What Doesnt
【速读】： 该论文旨在解决如何将预训练的大规模模型高效转化为更小规模的“子代”模型（smaller sibling）这一关键问题，特别是在保持模型性能的前提下实现参数量压缩。其核心挑战在于：在不同规模模型之间进行转换时，如何有效迁移知识并避免功能退化。解决方案的关键在于揭示并利用“初始化”（initialization）中的转换价值，并提出两个独立可调控的优化杠杆：最小二乘补偿（least-squares compensation，用于提升零样本性能）和方差保持重缩放（variance-preserving rescale，用于维持训练动态）。研究发现，通过线性投影进行权重转换会破坏旋转编码、分头结构、GELU激活及层归一化等关键组件的内在结构，导致功能破坏；而经过最优线性变换后，残差项在随机置换控制下与噪声无异，表明转换的实质价值完全依赖于初始权重配置。在匹配预算的持续预训练中，该方法在低数据量（如30M tokens）下显著优于从头训练或子克隆（subcloning）方法，实现以更少数据达到同等甚至更高性能，仅在极高预算下趋于收敛。此外，研究还指出当目标模型规模约为源模型的1/5（如6.9B→1.4B）时，叠加两个杠杆会导致过校正，根源在于补偿求解的病态条件，提示需引入维度感知正则化以改善稳定性。
链接: https://arxiv.org/abs/2608.02829
作者: Ravi Satya Durga Prasad Yenugula
机构: Independent Researcher(独立研究员)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 15 pages, 5 figures. Independent research preprint
Abstract:Model families train every size from scratch. Can a pretrained large model be converted into a smaller sibling? We characterize the 1.4B-410M conversion in the Pythia family end-to-end: (i) representations align strongly across sizes (ridge R^2=0.84) while parameters align weakly; (ii) dense weight projection is functionally destructive – provably not an assembly artifact – because basis mixing breaks rotary, per-head, GELU, and LayerNorm structure; (iii) after the best-fit linear operator, weight residuals are statistically indistinguishable from noise under shuffle controls; (iv) conversion value therefore lives in initialization. In matched-budget continued pre-training we decompose conversion into two independent levers – least-squares compensation (function: best zero-shot) and variance-preserving rescale (dynamics: best endpoints). Compensation is a token-efficient, low-budget win rather than a universal one: at 30M tokens it beats the strongest subcloning variant on both a width-reduced pair (84.0 +/- 1.8 vs. 89.7 +/- 3.7, 3/3 seeds) and a held-out depth-reduced pair (109.3 vs. 117.9, 3/3 seeds), reaching a given quality with fewer tokens; at a 33x larger budget the two converge to parity (40.0 vs. 40.0), both far ahead of from-scratch, which transfer initialization always beats – by up to 18x at low budget, the margin narrowing at convergence and at the largest scale. We further map the method’s boundary: at ~5x the donor scale (6.9B-1.4B) stacking both levers over-corrects, which we trace to ill-conditioning of the compensation solve at large width, pointing to dimension-aware regularization as the fix. Code, checkpoints, and the frozen evaluation corpus are released.
[NLP-98] Similarity All The Way Up: Multilingual Generalization in LLM s Relies on Language-Level Similarity Structures
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在多语言场景下泛化能力不足的问题，尤其是对英语以外、训练数据稀缺的语言的泛化表现不佳的现状。其核心问题是：为何某些模型在跨语言任务中表现优于其他模型？解决方案的关键在于揭示语言表示在相似性空间（similarity space）中的结构特性——即模型是否能够准确捕捉不同语言之间的层级相似性结构。研究发现，尽管训练数据存在偏差，大语言模型的隐层表示仍能有效恢复印欧语系的语言家族树层级结构，将同属一个子系的语言在表示空间中紧密聚集。进一步分析表明，模型对语言相似性结构的表征程度与其在XNLI（多语言自然语言推理）基准上的性能显著相关，说明能够以相似方式表征相似语言的模型具备更强的跨语言泛化能力。这一发现为理解大规模语言模型的泛化机制提供了认知科学视角的支持，揭示了良好表示学习与跨语言迁移能力之间的内在关联。
链接: https://arxiv.org/abs/2607.22699
作者: Supantho Rakshit,Adele Goldberg,Henry Conklin
机构: Princeton University (普林斯顿大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 8 pages, 3 figures, submitted to CogSci 2026
Abstract:As Large Language Models (LLMs) grow more capable across diverse tasks, their (in)ability to generalize remains difficult to quantify and poorly understood beyond limited domains. In particular, LLMs are known to struggle generalizing multilingually, to languages outside of English, and that are poorly attested in their training data. To understand why this may be, and what enables some models to perform better than others, we turn to a long history of work across the cognitive sciences, arguing that successful generalization derives from appropriate representations in similarity space. We look at how well LLMs’ representations capture the hierarchical similarity structure between distinct languages. Strikingly, we show LLMs’ latent representations largely recover the hierarchical structure of the Indo-European language family tree – grouping languages that are members of the same subfamily closely together in representation space. Furthermore, we show that the degree to which models reflect the similarity structure of languages correlates with their performance on XNLI, a multilingual natural language inference benchmark. This extends classic work on similarity-driven generalization at scale, showing how models that represent similar languages similarly generalize better from one language to another.
[NLP-99] aching Nemotron Greek: Mining a Corpus Adapting Retrieval and Grounding Generation for Modern Greek across Specialist Domains
【速读】： 该论文旨在解决现代希腊语（Modern Greek）在主流多语言检索模型与基准测试中缺失的问题，这限制了其在法律、能源、金融及医疗等专业领域中检索增强生成（RAG）系统的应用。其核心解决方案是构建一个端到端的适配流程，涵盖语料挖掘、合成监督、检索模型训练、重排序器适配、阅读器微调，并引入首个大规模希腊语RAG基准——HERA。关键创新在于：1）发现无需参数的BM25基线在专业希腊语语料上优于多个现成的多语言稠密检索模型；2）通过65,773对希腊语检索数据微调Nemotron 1B嵌入模型后，nDCG@10从0.362提升至0.835，显著超越未适配版本；3）利用LoRA微调Nemotron 30B-A3B混合专家阅读器，在答案正确性上从29.4%提升至66.9%，同时大幅改善生成内容的忠实度与引用质量；4）所提出的HERA基准为未来希腊语RAG研究提供了标准化评估工具。
链接: https://arxiv.org/abs/2608.05138
作者: Ayoub Kirouane,Christos Petrocheilos
机构: 未知
类目: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 15 pages, 10 figures, 7 tables. Includes release of the HERA benchmark and Sophea Nemo RAG models
Abstract:Modern Greek is absent from NVIDIA’s Nemotron retrieval models and from major multilingual retrieval benchmarks, despite being important for retrieval-augmented generation (RAG) in legal, energy, financial, and medical applications. We present an end-to-end adaptation of the Nemotron retrieval stack for Modern Greek, including corpus mining, synthetic supervision, retrieval model training, reranker adaptation, reader fine-tuning, and a new benchmark called HERA. Our study shows that a parameter-free BM25 baseline outperforms several off-the-shelf multilingual dense retrieval models on specialist Greek corpora. After fine-tuning on 65,773 Greek retrieval pairs, a Nemotron 1B embedder improves nDCG@10 from 0.362 to 0.835 and substantially outperforms its unadapted counterpart. The learned language competence transfers to general-domain Greek, although the advantage over BM25 remains domain-dependent. We further adapt a cross-encoder reranker and demonstrate consistent improvements across specialist domains. Finally, we LoRA-tune a Nemotron 30B-A3B mixture-of-experts reader for grounded generation, increasing judged answer correctness from 29.4% to 66.9% while significantly improving faithfulness and citation quality. We also introduce HERA, the first large-scale Greek benchmark for retrieval-augmented generation, and release our adapted models and benchmark to support future research on Greek-language RAG systems.
[NLP-100] Leverag ing Machine Learning to Gain Insights on Quantum Thermodynamic Entropy
【速读】： 该论文旨在解决量子引擎在信息处理过程中因观测与控制粒子所引发的热力学代价问题，尤其关注量子与经典极限下由林德勒原则（Lindauer’s Principle）所导致的信息处理耗散之间的热力学权衡。其解决方案的关键在于构建一个基于单量子粒子的工作流体、受经典混沌Szilard映射启发的量子引擎模型，该模型遵循测量—热能提取—记忆重置的热力学循环，并通过机器学习方法实现对能量分析的精确模拟与验证。研究发现，尽管量子引擎在工作状态中符合基于Szilard引擎的第二定律，但其运行机制与经典版本存在本质差异，其中插入隔板的热力学成本在量子实现中起着决定性作用，揭示了量子信息处理中独特的热力学约束。
链接: https://arxiv.org/abs/2305.06177
作者: Srinivasa Rao. P
机构: 未知
类目: Quantum Physics (quant-ph); Statistical Mechanics (cond-mat.stat-mech); Computation and Language (cs.CL); Machine Learning (cs.LG); Computational Physics (physics.comp-ph)
备注: 9 pages, 7 figures
Abstract:We present a thermodynamic analysis of a quantum engine that uses a single quantum particle as its working fluid, inspired by Szilard’s classical single-particle engine. Our design is modeled after the classically-chaotic Szilard Map and involves a thermodynamic cycle of measurement, thermal-energy extraction, and memory reset. Our focus is on investigating the thermodynamic costs associated with observing and controlling the particle and comparing these costs in the quantum and classical limits. Through our study, we aim to shed light on the thermodynamic trade-offs that arise from Lindauer’s Principle for information-processing-induced thermodynamic dissipation in both the quantum and classical regimes. Using machine learning methods, we demonstrate that energy analysis can be performed and the quantum engine can be simulated according to the Szilard engine based Second Law of Thermodynamics in its working condition. However, we note that the quantum engine operates using significantly different mechanisms than its classical counterpart, where the cost of inserting partitions plays a critical role in the quantum implementation.
信息检索
[IR-0] MemoryCPT: An End-to-End Agent Memory Framework for Cost-Performance Trade-off
链接: https://arxiv.org/abs/2608.04843
作者: Songxin Lei,Kun Ouyang,Weilin Ruan,Yuqian Wu,Zhijiang Guo,Yushi Sun,Fugee Tsung
类目: Information Retrieval (cs.IR)
备注:
Abstract:Long-horizon LLM agents require memory systems that recover useful evidence from large interaction histories without passing excessive context to downstream models. Existing memory pipelines often rely on hand-crafted heuristics and repeated LLM calls, which can introduce redundant context and high inference cost. We propose MemoryCPT, an end-to-end trainable agent memory pipeline that spans offline memory construction and online query-conditioned context generation. MemoryCPT consists of two stages: Query-agnostic Distillation (QAD), which distills a modular memory-construction pipeline into a compact model using explicit reasoning traces; and Query-aware Retrieval and Summarization (QAR), which combines reciprocal rank fusion (RRF) with a LoRA-based summarizer trained via Group Relative Policy Optimization (GRPO) under a cost-aware reward. We further introduce Quality per Cost (QPC) to quantify answer quality per unit inference cost. Experiments on LoCoMo and LongMemEval show that MemoryCPT improves the cost-performance trade-off over the evaluated baselines, while ablation and sensitivity analyses characterize the contributions of its components and the effects of key design choices.
[IR-1] DEGR: Dual Exploration-Driven Generative Re-Ranking for Adaptive Cross-Request Context Bridging KDD2026
链接: https://arxiv.org/abs/2608.04809
作者: Binglei Zhao,Xuanhua Yang,Xiwei Zhao,Sulong Xu
类目: Information Retrieval (cs.IR)
备注: Accepted by KDD2026 ADS Track, 11 pages
Abstract:In industrial recommendation systems, the re-ranking stage balances business objectives and diversity for sequence-level optimization while modeling contextual information. However, constrained by fixed upstream supply, existing methods fail to deliver further effectiveness gains, especially under low-quality supply. To overcome this, re-ranking can actively balance immediate and exploratory value, for instance, by prioritizing exploratory exposure under low-quality supply to preserve browsing potential and facilitate serendipitous conversions. Therefore, we propose a Dual Exploration-Driven Generative Re-Ranking (DEGR) method. DEGR adopts a hybrid supervised-reinforcement exploration and optimization paradigm, guided by an exploratory reward model that adaptively balances immediate and exploratory value. The hybrid optimization paradigm integrates three key components: supervised learning, exploration diversity constraint, and adaptive reward-weighted ORPO for preference optimization. Through this dual exploration, the generator ultimately acts as an adaptive cross-request contextual bridge. Offline and online experiments indicate that DEGR outperforms SOTA methods, achieving improvements of up to 1.22% UCTR and 0.20% PV in the JD E-commerce recommendation system.
[IR-2] WatchLens: A Configurable Platform for Online Video Recommendation Experiments RECSYS2026
链接: https://arxiv.org/abs/2608.04807
作者: Deogyong Kim,Dongha Lee
类目: Information Retrieval (cs.IR)
备注: 6 pages, 3 figures. Accepted to RecSys 2026
Abstract:Studying how video recommender systems shape user behavior requires online experiments that link playback behavior with the recommendation conditions that produced it. Existing user-study infrastructure provides one or the other, but not both within a single experimentation workflow. We present WatchLens, an open-source platform that fills this gap. WatchLens adopts a modular architecture in which user interfaces, content sources, and recommendation policies are independently configurable, with policies assignable separately to the feed and the watch page, while a standardized logging layer attaches the recommendation policy and ranking position to every event at recording time. This design enables researchers to analyze how recommendation policies and ranking positions shape downstream playback behavior, session continuation, and navigation between the feed and the watch page, with the linkage between policy and outcome available in each event rather than reconstructed afterwards. We demonstrate WatchLens with a short-form video case study that holds the interface, feed policy, and content pool constant while varying only the watch-page policy, showing how the platform supports session-level comparison of recommendation effects on real viewing behavior. WatchLens is released as a publicly available, single-server deployable system for reproducible online video recommendation research.
[IR-3] Caching for the Future: Scrub Jay Episodic Memory Principles for Agent Memory Systems
链接: https://arxiv.org/abs/2608.04746
作者: Kartikey Singh Bhandari,Aarya Wadhwani,Dhruv Kumar,Pratik Narang
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:LLM agents that persist across sessions accumulate stored memories whose validity varies enormously by content type, yet existing memory architectures treat all memories as equally persistent and systematically contaminate retrieved context with outdated facts. We show that per-memory, type-conditioned temporal decay, a property of western scrub jay episodic memory, can be operationalized as an auto-classified coefficient \pi_i in an external LLM-agent memory store, yielding ScrubJay-MEM: each memory is encoded as a jointly-bound What–Where–When tuple with an estimated perishability \pi_i and utility horizon \tau_i , retrieved by query-adaptive scoring, and revised retroactively at O(1) LLM calls per update. We introduce the Temporal Generalization Test (TGT), a benchmark with held-out retention intervals and a Generalization Gap (GenGap) metric. On TGT, ScrubJay-MEM is the only retrieval-based system with substantially positive GenGap ( +0.108 ); on MemoryAgentBench EventQA-64k it improves F1 by +2.66 over Mem0 and +3.09 over Qwen3-Embedding-4B under a llm backbone. A decay ablation collapses GenGap by 5.7\times , establishing type-conditioned decay as necessary for the result. Gains narrow under stronger backbones and reverse on fact-consolidation tasks, scoping the contribution to temporal reasoning over perishable facts.
[IR-4] Characterizing the Evolving Landscape of Modern Information Seeking
链接: https://arxiv.org/abs/2608.04609
作者: Shuoqi Sun
类目: Information Retrieval (cs.IR)
备注: Best Paper Award at FDIA 2026; 2 Pages (Excluding References)
Abstract:Information seeking (IS) evolves, as does the human IS process. Since the rise of Generative AI (GenAI), modern IS has shifted by introducing more interfaces, more complex interactions, and expanded system capabilities. We argue that these changes in modern IS should be systematically examined. This PhD research characterizes the changes in the modern IS process. We use mechanisms, including online crowdsourcing survey experiments, theoretical IS frameworks, and in-lab experiments with neurophysiological signals, to characterize the shifts in modern IS, especially those driven by GenAI. We offer insights into the current landscape of search interface preferences and the cognitive efforts involved in seeking information. We believe this PhD research will contribute to and inform future designs of personalized, cognition-aware IS systems.
[IR-5] owards Robust Version Identification in the Wild: A Dataset Benchmark and Fine-Tuning Study
链接: https://arxiv.org/abs/2608.04543
作者: Simon Hachmeier,R. Oguz Araz,Dmitry Bogdanov,Robert Jäschke,Xavier Serra
类目: ound (cs.SD); Information Retrieval (cs.IR)
备注: Accepted to the Proceedings of the 27th International Society for Music Information Retrieval Conference (ISMIR 2026)
Abstract:Existing datasets for musical version identification (VI) are primarily derived from curated metadata sources such as SecondHandSongs and Discogs, and are therefore dominated by professionally recorded tracks. This leads to a domain mismatch with real-world scenarios, where amateur and user-generated content is prevalent. To address this limitation, we introduce DiVers, a large-scale VI dataset comprising over 1.1 million musical versions, with train-validation-test splits compatible with established datasets such as Discogs-VI-YT, SHS100K, and Da-TACOS. In addition to standard version-level annotations, DiVers provides automatically assigned tags (e.g., instrumental, live) and segment-level predictions indicating the presence or absence of music. We evaluate the proposed dataset by training state-of-the-art VI systems. Our results show that models trained on DiVers achieve substantially improved robustness to acoustically diverse and noisy inputs, while maintaining a stable performance on cleaner, studio-quality benchmarks. We release the dataset metadata, code for its construction, and all experimental pipelines to support reproducibility.
[IR-6] Skills Know Their Neighbors: Cluster-Contrastive Capability Pages for Skill Retrieval
链接: https://arxiv.org/abs/2608.04482
作者: Zifei Wang,Wei Wen,Qiang Ji,Ruizhi Qiao
类目: Information Retrieval (cs.IR)
备注: 16 pages, 4 figures
Abstract:As skill libraries grow, large language model agents must retrieve reusable skills from candidates that often share the same topic and vocabulary but implement different capabilities. Retrieval is limited not only by the scorer but also by the text being scored: a document may describe what a skill does without stating which similar requests should be routed elsewhere. We formalize a skill’s capability as its \emphexecutable region, the set of queries it can solve, and view its document as a lossy observation of that region. This view exposes a document-imposed component of retrieval error that cannot be removed by improving the retriever alone. We therefore propose \emphCapability Pages, cluster-contrastive skill representations containing a positive trigger \Tpos , a negative boundary \Tneg , and a discriminative body B . An offline compiler compares neighboring skills to write these fields. At inference time, the index uses \Tpos and B for candidate recall, while the router uses \Tneg to reject confusable alternatives. On SRA-Bench, which contains 26,262 skills and 5,400 questions from six datasets, Capability Pages improve Recall@10 for all five tested retrievers, with a mean gain of 2.94 points. Adding \Tneg to candidate cards improves end-to-end task success by 3.62 points on average across four executors and six datasets. A transfer evaluation on Chinese SSL-SkillDiscovery reaches 73.07% MRR@50 using the same encoder across conditions. Capability Pages require no modification to the online models; they improve routing by rewriting the offline skill library.
[IR-7] Multi-Objective Ranking for Live-Streaming: Balancing Fresh and Delayed Signals with Segment-Aware Targeting RECSYS2026
链接: https://arxiv.org/abs/2608.04455
作者: Xiaoyi Gu,Julia Tavares,Eder Santana,Carlos Mendoza-Cardenas,Nikita Mishra,Saad Ali
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 9 pages, 3 figures. Accepted to the Industry Track of the 20th ACM Conference on Recommender Systems (RecSys 2026)
Abstract:One of the most challenging problems entertainment live-streaming services face in recommendation systems is that user behaviors are sparse and delayed, and interaction data exhibits bias for different user segments. Unlike e-commerce applications where user actions follow linear sequences, live-streaming viewers engage in multiple concurrent behaviors of watching, chatting, following, and spending, each occurring with varying delays. We address these challenges through three key contributions: 1) a delayed window approach that extends feedback collection beyond immediate responses, 2) a multi-model architecture that combines fresh and delayed signals, and a segment-aware targeting module that optimizes ranking scores differently across user lifecycle stages, and 3) Multi-gate Mixture-of-Experts (MMoE) integration that jointly models correlated targets while reducing model parameters by 41.9% compared to independent models. Online A/B testing demonstrates significant improvements, including a +0.09% increase in Daily Active Viewers (DAV), generating millions more annual active viewer days, and +0.56% increase in highly engaged viewers’ capped Average Revenue Per User (ARPU). Viewer-segment targeting achieved an additional +0.15% DAV improvement for newer and less engaged viewers, while MMoE enhancement added +0.08% overall DAV and +0.27% new follows. The proposed system processes ranking requests with low latency, providing a scalable approach for balancing multiple business objectives across diverse user populations. In addition, we tested the multi-model architecture on the Twitch mobile live feed and achieved a +1.12% increase in positive user-channel interactions (clicks, follows, and likes), demonstrating applicability beyond the primary use case.
[IR-8] he Price of Isolation: Estimating the Ecosystem Cost of Symmetric Two-Sided A/B Testing
链接: https://arxiv.org/abs/2608.04432
作者: Yuanyuan Shen,Yiren Yan,Wenjie Li,Chunhui Zhu
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG); Social and Information Networks (cs.SI); Methodology (stat.ME)
备注:
Abstract:On two-sided content platforms, symmetric two-sided isolation (assigning matched fractions of creators and viewers to isolated treatment and control submarkets) is widely used for creator-side and cold-start experiments because it removes cross-arm marketplace interference. Isolation, however, thins each viewer’s candidate catalog, and intuition suggests the resulting engagement cost should fade as the platform grows: a small fraction of a vast catalog is still vast. We show that, in an order-statistics model of engagement, whether this intuition holds depends on the upper tail of match quality. Extreme-value theory yields tail-class loss laws with a sharp dichotomy: for light or bounded tails the loss vanishes as the candidate pool grows, whereas under heavy tails it converges to a size-independent constant, so expanding the candidate pool, even by orders of magnitude, does not asymptotically eliminate the cost. Evidence from two production experiments on a platform with millions of active creators is consistent with this picture: a pure A/A traffic sweep reveals a measurable, depth-graded engagement cost; a one-sided catalog ablation independently shows that per-viewer thinning contributes to the loss; and a tail index calibrated on the small exploration pool predicts an effect consistent with the one observed in the far larger full-catalog ablation. Isolation thus carries a price that experimenters should budget for, like any other cost. We give practitioners a preflight procedure that estimates it before launch, sizes traffic accordingly, and recommends a fallback design when the predicted cost exceeds a chosen tolerance.
[IR-9] CLIP-CC-Bench: Evaluating Parag raph-Level Video Descriptions in Video-Language Models SIGIR2026
链接: https://arxiv.org/abs/2608.04302
作者: Mukhtiar Ali,Harsh Dubey,Sugam Mishra,Chulwoo Pack
类目: Computer Vision and Pattern Recognition (cs.CV); Information Retrieval (cs.IR); Multimedia (cs.MM)
备注: Accepted and presented at EvalMG 2026, the Second Workshop on Evaluation for Multimodal Generation, co-located with ACM SIGIR 2026
Abstract:Benchmarking video-language models has largely focused on short clips and single-sentence metrics, leaving open whether current systems can generate accurate long-form, paragraph-level descriptions. We introduce CLIP-CC-Bench, an evaluation suite for long-form video description built from 5 hours of movie content segmented into 90-second clips, each paired with an expert-written paragraph-style reference. The evaluation suite employs an ensemble of five state-of-the-art LLM-based embedding models to increase reliability and mitigate single-model bias, and applies two complementary methodologies: (i) coarse-grained semantic matching and (ii) fine-grained semantic matching to compare model-generated descriptions against CLIP-CC-Bench references. Using this framework, we evaluate 17 state-of-the-art video-language models and report both their Borda-aggregated rankings and their average scores on CLIP-CC-Bench. We further quantify the protocol’s internal reliability through inter-judge agreement and bootstrap ranking stability. We release standardized evaluation scripts, model outputs, and aggregation tools at this https URL to support reproducibility. CLIP-CC-Bench provides a practical evaluation framework for long-form video description, filling a gap left by existing short-clip and QA-only benchmarks.
[IR-10] Neighborhood-Aware Dual Biomedical Entity Linking
链接: https://arxiv.org/abs/2608.04144
作者: Yicheng Tao,Jie Liu
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Biomedical entity linking grounds mentions in clinical and scientific text to entities in a curated knowledge base (KB) with ontological structure, which supports downstream applications such as literature-scale information extraction and patient-record normalization. The task has several challenges at once: the KB contains large numbers of entities, mentions are often ambiguous, and gold labels follow annotation conventions specific to each corpus. To address these challenges, we propose PILOT, a three-stage framework made up of neighborhood-aware retrieval, dual reranking, and score fusion. The retriever injects ontological structure from both the query and KB side, by reformulating mentions and pooling entity embeddings. The retrieved pool is then scored from two complementary views, one over surface forms and one over context, and fused together. PILOT achieves the state of the art on average across five widely-used benchmarks and remains efficient at inference.
[IR-11] RAG -Stack: Co-Optimizing RAG Serving Performance and Quality
链接: https://arxiv.org/abs/2608.03487
作者: Haiqiang Zhang,Yuanqing Lei,Wanting Li,Tao Zhang,Wenqi Jiang
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Retrieval-augmented generation (RAG), which augments large language model (LLM) generation with information retrieved from databases, has become a widely used approach for knowledge-intensive applications. Modern RAG systems, however, expose many configuration choices, such as retrieval indexes, model selections, and how models invoke retrieval. Each configuration yields a different trade-off between answer quality and serving performance, making it challenging to choose the optimal setting for a specific application deployment. We present RAG-Stack, a framework for efficiently discovering quality-performance Pareto frontiers across diverse RAG applications and serving systems. RAG-Stack consists of RAG-PE, an iterative design-space exploration algorithm that selects the next RAG configuration to evaluate; RAG-IR, a workload abstraction for diverse RAG algorithms; and RAG-CM, a performance model that predicts the optimal deployment and serving performance on the given hardware. Together, these components allow RAG-Stack to search the joint algorithm-system configuration space without deploying every candidate and to transfer an existing Pareto frontier to a new serving system. Given the same number of optimization iterations across diverse datasets, the Pareto frontiers found by RAG-Stack cover 52.5% to 153.2% more of the normalized quality-performance space than those found by state-of-the-art configuration-search methods evaluated over the same RAG design space.
[IR-12] A Deep Causal Inference Approach to Measuring the Effects of Forming Group Loans in Online Non-profit Microfinance Platform
链接: https://arxiv.org/abs/1706.02795
作者: Thai T. Pham,Yuanyuan Shen
类目: Machine Learning (stat.ML); Information Retrieval (cs.IR); General Finance (q-fin.GN)
备注:
Abstract:Kiva is an online non-profit crowdsouring microfinance platform that raises funds for the poor in the third world. The borrowers on Kiva are small business owners and individuals in urgent need of money. To raise funds as fast as possible, they have the option to form groups and post loan requests in the name of their groups. While it is generally believed that group loans pose less risk for investors than individual loans do, we study whether this is the case in a philanthropic online marketplace. In particular, we measure the effect of group loans on funding time while controlling for the loan sizes and other factors. Because loan descriptions (in the form of texts) play an important role in lenders’ decision process on Kiva, we make use of this information through deep learning in natural language processing. In this aspect, this is the first paper that uses one of the most advanced deep learning techniques to deal with unstructured data in a way that can take advantage of its superior prediction power to answer causal questions. We find that on average, forming group loans speeds up the funding time by about 3.3 days.
人机交互
[HC-0] DeepConnect: A Visual Analytics System for Bridging Interdisciplinary Research Collaborations IEEE-VIS2026
链接: https://arxiv.org/abs/2608.05134
作者: Yingchaojie Feng,Zekai Shao,Yiqun Sun,Yixuan Tang,Anthony K. H. Tung
类目: Human-Computer Interaction (cs.HC)
备注: 11 pages, 7 figures, and 1 table. Accepted at IEEE VIS 2026; to appear in IEEE Transactions on Visualization and Computer Graphics
Abstract:Interdisciplinary research collaboration is crucial for scientific innovation, but it remains difficult to initiate in practice. Existing collaborator discovery approaches are often constrained by disciplinary boundaries and static researcher profiles that do not reflect the specific context of a new collaboration goal. As a result, researchers struggle to translate open-ended collaboration goals into domain-specific tasks, evaluate candidate researchers’ fit and complementarity, and establish common ground before initial contact. To address these challenges, we present DeepConnect, an LLM-augmented visual analytics system for interdisciplinary collaborator discovery. DeepConnect translates collaboration ideas into domain-specific tasks, retrieves relevant papers to ground cross-domain exploration, and provides coordinated visualizations for exploring and comparing candidate researchers. It further reveals terminology gaps and overlaps across domains and supports publication-grounded conversation rehearsal to help users prepare for outreach. We evaluate DeepConnect through two case studies, a user study, and a component-level evaluation, showing its value for complementary team formation, idea refinement, and pre-contact preparation. The DeepConnect website is available at this https URL.
[HC-1] Characterizing Visual Accessibility Issues in AI Developer Tools: An Empirical Study
链接: https://arxiv.org/abs/2608.05116
作者: Sabrina Haque,Christoph Csallner
类目: oftware Engineering (cs.SE); Human-Computer Interaction (cs.HC)
备注:
Abstract:AI-assisted developer tools increasingly mediate programming through chat panels, terminal agents, generated diffs, and streaming status output. These interaction surfaces may create visual accessibility barriers for blind, low-vision, and color-vision-deficient developers, yet little is known about how such barriers are reported in public tool ecosystems. We analyze issues and forum discussions from five AI developer tool ecosystems: GitHub Copilot in VS Code, Cursor, Claude Code, OpenAI Codex, and OpenCode. From 2,652 keyword-retrieved candidates, a three-model ensemble identified 600 unanimously positive visual accessibility reports. A stratified manual sanity check supported this conservative selection. Topic modeling and qualitative analysis identified three recurring categories: screen-reader and assistive-technology barriers; visual presentation, contrast, and differentiation problems; and readability, scaling, and control limitations in AI-specific interfaces. The relative prominence of these concerns varied across ecosystems and reflected differences in editor, terminal, chat, diff, and agent interaction surfaces. An exploratory metadata analysis further identified differences in reporter activity and, across the GitHub-based ecosystems, maintainer participation and closure processes. These findings show that the accessibility record of AI developer tools is shaped by both their interaction design and the reporting and maintenance practices of their surrounding ecosystems.
[HC-2] Robust and Efficient Motion Reasoning for Privacy-Aware Classroom Incident Recognition
链接: https://arxiv.org/abs/2608.05115
作者: Paritosh Parmar,Landy Lan,Hong Yang,Chen Yi,Chiat Pin Tay
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注:
Abstract:Can computer vision help make classrooms safer? In this pilot study, we investigate privacy-aware and computationally efficient classroom incident recognition from CCTV-style observations. This setting remains underexplored, with limited benchmarks and few methods designed for the privacy, efficiency, and generalization demands of real-world deployment. We introduce a novel hybrid benchmark combining generative CCTV-style videos with real-world classroom pose data, and propose a lightweight, but robust motion-reasoning framework motivated by the observation that many incidents differ more in motion direction, speed, acceleration, and intensity than in pose alone. To that end, our method first constructs hierarchical kinematic representations of human actions. Our method then distills hierarchical, multi-order kinematic reasoning from a large teacher into a much smaller single-order student, enabling efficient per-person inference while preserving expressive motion understanding. Experiments show that our model outperforms substantially larger baselines at less than one-tenth of their computational cost, while also demonstrating stronger out-of-domain motion reasoning and zero-shot synthetic-to-real generalization. We will publicly release the benchmark, codebase, and supporting tools to facilitate further research in privacy-aware classroom safety.
[HC-3] ArtAnno: Annotating Implicit Semantics in Artworks through LLM Agent -Driven Bidirectional Human-AI Augmentation
链接: https://arxiv.org/abs/2608.05026
作者: Xiaoyan Gu,Yifang Wang,Wenqing Zheng,Haozhong Liu,Yixia Zheng,Peiyi Jiang,Wenjie Ning,Wei Zhang,Wei Chen
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:High-quality annotation of artworks is essential for computational art research, yet extracting implicit semantics remains challenging due to the reliance on culturally grounded meanings and deep contextual knowledge behind the images. Current AI-assisted annotation tools often lack assistance or rely on one-way workflows where experts have to perform extra manual calibrations to improve AI models, resulting in limited efficiency. To address this, we propose Bidirectional Human-AI Augmentation(BiHAA), a closed-loop framework in which skills and domain knowledge base evolve through real-time interaction and bidirectional HAI augmentation. Informed by a formative study with 20 artwork annotators from different backgrounds, we implement this framework in ArtAnno, an artwork annotation system driven by a multi-agent architecture. The system includes a Proactive Agentic Support Module, where AI augments humans through semantic mining and label suggestion, and an Interaction-Driven Evolution Module, where human expertise continuously enhances the AI through distilling annotation trajectories into reusable experience. Evaluation through a user study and two case studies demonstrates that our framework and system improve annotation efficiency, enable knowledge accumulation, and reduce the effort of information seeking and verification for annotators with limited domain expertise. We conclude by discussing broader implications and future directions.
[HC-4] Exploring Cross-Reality Transitions between Projections and Head-Mounted Displays for Immersive Digital Art
链接: https://arxiv.org/abs/2608.04971
作者: Xiangpeng Fu,Mads Haahr
类目: Human-Computer Interaction (cs.HC); Graphics (cs.GR)
备注:
Abstract:Immersive exhibitions increasingly combine large-scale projections with mixed reality (MR) head-mounted displays (HMDs), but how users perceive transitions across these heterogeneous visualization environments remains underexplored. We investigate cross-reality (CR) object- and scene-level transitions in a hybrid immersive art installation spanning projection, augmented reality (AR), and virtual reality (VR). In a within-subjects study (N=24), we compared a calibrated condition with a bundled inconsistency condition introducing noticeable differences in spatial alignment, visual appearance, and cross-device latency. This contrast was used not to establish perceptual thresholds or hardware targets, but as a diagnostic probe to make transition disruptions more perceptible and discussable. The inconsistency condition reduced presence and increased workload. Interviews showed that spatial misalignment disrupted action-outcome predictability, appearance mismatch reduced aesthetic coherence, and latency weakened perceived responsiveness. Effects also varied by asset type: rigid static assets exposed inconsistencies more clearly, animated skeletal assets shifted attention toward motion, and particle effects often masked small discrepancies. These findings clarify how users perceive projection-MR transitions and support asset-aware design strategies for coherent hybrid immersive art. We also release HUICRSync, an open-source Unreal Engine plugin for prototyping projection-MR CR experiences.
[HC-5] Reply Delete or Ignore? Examining How Content Creators Perceive and Select Comment Moderation Strategies
链接: https://arxiv.org/abs/2608.04951
作者: Yunhee Shim,Shagun Jhaver
类目: Human-Computer Interaction (cs.HC)
备注: 27 pages, including an appendix and 8 tables
Abstract:Content creators on social media sites occupy highly visible positions on their channels. As a result, creators, especially those with large followings, experience disproportionate levels of online harm. To address such harm, they enact a range of moderation strategies, which in turn shape the visibility of content that their audiences encounter. This paper examines how content creators perceive three moderation strategies to address hateful comments (deleting, replying to, or simply ignoring) and how they decide which strategy to deploy. While creator moderation is usually examined through the lens of safety, creators’ regulation decisions may also be shaped by concerns about how their actions appear to audiences and how they are rewarded or penalized by platforms’ recommendation algorithms. Conducting a survey of 584 content creators, we found that in their view, (1) deleting is the most beneficial for achieving safety, (2) both deleting and replying produce better impression management benefits than ignoring, and (3) replying is perceived to yield the highest algorithmic benefits. Crucially, while expectations of emotional safety and impression management benefits significantly predicted creators’ willingness to adopt each comment moderation strategy, perceived algorithmic benefits did not. By unpacking how creators evaluate these trade-offs, this study contributes to HCI research on understanding creator-led, middle-level governance. We conclude with design implications for supporting creators as crucial governance actors without burdening them with sole responsibility for online safety.
[HC-6] ransition Techniques for Externally-Guided Multi-Scale Viewpoint Changes
链接: https://arxiv.org/abs/2608.04912
作者: Matt Gottsacker,Mengyu Chen,David Saffo,Feiyu Lu,Benjamin Lee,Blair MacIntyre
类目: Human-Computer Interaction (cs.HC)
备注: IEEE ISMAR 2026
Abstract:Extended reality (XR) is increasingly used to help users understand complex virtual environments through multiple viewpoints across different immersion levels, positions, and scales. While numerous techniques address viewpoint transitions for self-guided exploration, many scenarios require externally-guided transitions where a system or presenter controls the user’s viewpoint, leaving the user with limited spatial knowledge and control over the transition process, which can increase susceptibility to disorientation and discomfort. We present three transition techniques for externally-guided multi-scale XR viewpoint changes and evaluate them against a fade-to-black baseline in a within-subjects study (N=20). Participants transitioned between world-in-miniature, street-level, and indoor destination views. We combined spatial recall measures, standardized questionnaires, and semi-structured interviews to assess orientation, workload, comfort, and continuity. Results showed that in our setup, techniques externalizing reference frames and the user’s pose improved multi-scale spatial recall relative to a fade, while same-scale recall was insensitive to transition technique. We conclude with design implications for multi-scale XR viewpoint transitions.
[HC-7] AutoCue: Multimodal LLM -Assisted Externalization of Implicit Inputs as Instructional Visual Cues in Screencast Tutorials
链接: https://arxiv.org/abs/2608.04910
作者: Shengyang Luo,Shengyao Luo,Xiaolei Guo,Fengze Zhang,James Liang,Yingjie Victor Chen
类目: Human-Computer Interaction (cs.HC)
备注: Accepted to Graphics Interface 2026 (GI 2026)
Abstract:Tutorial videos are widely used for learning feature-rich software, yet following screencast tutorials often breaks down in practice. Through a survey and contextual inquiry, we found that learners frequently rewind or get stuck because critical input information, especially mouse actions and keyboard-modified operations, is often implicit or missing in tutorials without input metadata. To address this problem, we present AutoCue, a multimodal LLM-assisted, human-in-the-loop tutorial augmentation pipeline for externalizing implicit inputs as instructional visual cues. AutoCue integrates frame-to-frame visual changes, narration signals, and operation guidance from official software manuals to infer likely mouse and key-modifier actions, then produces aligned cue layers and editable artifacts for human refinement. Grounded in multimedia learning and cognitive load theory, we further develop a visual cue grammar for representing mouse, keyboard, and combined inputs in software-learning tutorials. We instantiate and evaluate AutoCue in Autodesk Maya, focusing automatic inference on selected UI-mediated interactions with observable visual or textual feedback while supporting more ambiguous state changes through editable authoring artifacts. In a between-subjects study with 24 participants, the AutoCue-augmented tutorial reduced task completion time and interaction breakdowns and showed improved learner-reported experience.
[HC-8] Exploring Fraction Comprehension and Interest in Elementary Education Through AI-Powered Personalized Learning
链接: https://arxiv.org/abs/2608.04892
作者: Kenneth Holman
类目: Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注: PhD dissertation, University of Central Florida, 2024. Also available at this https URL
Abstract:Artificial intelligence systems that adapt instruction to individual learners are increasingly deployed in K-12 classrooms, yet empirical evidence on their effects in authentic elementary settings remains limited, particularly for students with mathematics learning difficulties. This dissertation examines AI-powered personalized learning during primary school fraction instruction, a domain that is foundational to later mathematics and STEM achievement. The first manuscript presents a systematic review of research on artificial intelligence in mathematics education published between 2020 and 2024. The second manuscript reports a quasi-experimental study evaluating Mathbot, a chatbot-based personalized learning platform, against business-as-usual classroom instruction. Repeated measures ANOVA was used to assess change in fraction comprehension and situational interest across time points. Results indicated modest improvements in fraction comprehension for students using Mathbot relative to traditional instruction, while changes in situational interest were not statistically significant. Findings suggest that automated personalization did not displace the instructional role of the teacher and that teacher decision-making remained central to student outcomes. The work contributes classroom-based evidence to ongoing discussion about the capabilities and limits of adaptive AI systems in elementary mathematics, and about accessibility and equity considerations when such systems are used with students with disabilities.
[HC-9] Investigating Click Behaviors On Google Search Result Pages That Produce an AI Overview
链接: https://arxiv.org/abs/2608.04831
作者: Athena Chapekis,Anna Lieb,Sono Shah,Aaron Smith
类目: Human-Computer Interaction (cs.HC)
备注: Presented at IC2S2 2026
Abstract:In 2024, Google introduced “AI Overviews,” a feature that displays an AI-generated result summary at the top of many Google search pages. This study investigates the role of AI in Google search using one month of web browsing data from a representative panel of 900 U.S. adults. Our analysis of the panelists’ Google searches sheds light on AI Overviews, when they appear in Google search results, and what user behaviors are associated with AI Overviews. We identify several attributes that make a search query more likely to generate an AI Overview, including the length of a query, whether the query begins with a question word, and whether the query contains both a noun and verb. When it comes to user behavior, we find that clicks to sources cited in AI Overviews are very rare, occurring in only about 1% of visits to AI Overviews. We also find that AI Overviews are associated with fewer clicks and higher rates of ending browsing sessions. Importantly, results from a mixed-effects logistic regression model indicate that these associations hold when controlling for random effects by panelist and query attributes that make AI Overviews more likely to appear.
[HC-10] Reference-Based Manipulation: A Framework and Pipeline for Multimodal Spatial Reasoning
链接: https://arxiv.org/abs/2608.04798
作者: Yangyang He,Zhuangze Hou,Yonglin Chen,Can Liu
类目: Human-Computer Interaction (cs.HC)
备注: 16 pages, 10 figures. Accepted to the 39th Annual ACM Symposium on User Interface Software and Technology (UIST 2026)
Abstract:When manipulating objects in immersive platforms through speech and gesture, users naturally construct spatial references, referring to scene entities, their bodies, or the environment. Leveraging spatial cognition theories, this work systematically examines how users construct and communicate spatial intent. Using a custom toolkit, we conducted a Wizard-of-Oz study to observe unconstrained multimodal (speech + gesture) input patterns in Virtual Reality for scene construction. Based on these findings, we formalize a framework that decomposes spatial references into three core components: Source, Anchor, and Frame, while characterizing their compositional strategies and explicitness. We demonstrate the utility of this Reference-based Manipulation framework by implementing an LLM-based pipeline featuring a set of example interaction techniques with a preliminary technical evaluation. Finally, we discuss key lessons learned for supporting reference-based spatial interaction.
[HC-11] LiverPlan: A Stage-Adaptive Immersive Visual Analytics Framework for Anatomical Liver Surgical Planning
链接: https://arxiv.org/abs/2608.04707
作者: Qixuan Liu,Shi Qiu,Xiwen Wu,Yuqi Tong,Yinqiao Wang,Ruiyang Li,Jialun Pei,Shengdong Zhao,Chi-Wing Fu,Pheng-Ann Heng
类目: Human-Computer Interaction (cs.HC)
备注: VIS 2026, to appear in TVCG
Abstract:Anatomical liver resection (ALR) surgery is the most important treatment for liver cancer, yet preoperative planning demands complex, multi-stage clinical reasoning under competing safety constraints. Current 2D desktop tools are not well equipped to support this process, exhibiting three fundamental limitations: reliance on monolithic interfaces that fail to adapt to the distinct cognitive demands of each planning stage; a perceptual bottleneck caused by limited anatomical spatial representation and missing plane-vessel intersection visualization; and an attention bottleneck stemming from fragmented critical safety criteria display across separate views. We present LiverPlan, a stage-adaptive immersive visual analytics framework for ALR planning, grounded in an 8-month collaboration with two expert hepatobiliary surgeons. Decomposing the surgical planning process into three sequential yet cognitively distinct stages, LiverPlan externalizes the cognitive demand of each stage via tailored techniques: (1) context-preserving focus and hue-preserving rendering for anatomical discovery; (2) direct 3D resection plane manipulation coupled with real-time, embedded visual feedback on critical safety criteria during plan refinement; and (3) explicit plane-vessel intersection visualization for anticipatory surgery preparation. A within-subjects study with eight hepatobiliary surgeons against a desktop baseline shows large-effect-size improvements in task completion time, perceived cognitive workload, and system usability on controlled planning tasks. Moreover, our study reveals broader insights: LiverPlan reduces cognitive burden and encourages a shift in surgeons from merely satisfying safety criteria to actively optimizing them, suggesting that explicit visualization of spatial relationships lowers the cognitive barrier to complex surgical planning.
[HC-12] AI Literacy for Legal Translation: Developing Digital Resilience
链接: https://arxiv.org/abs/2608.04641
作者: Łucja Biel
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注: 19 pages, 2 tables, 2 figures
Abstract:Generative AI is transforming legal translation by introducing opportunities alongside linguistic, technical, legal, ethical and cognitive risks. This chapter examines the implications of AI for professional legal translation and proposes an AI literacy framework tailored to the profession. It argues that AI does not change the fundamental objectives of legal translation but requires an extension of professional competence through AI literacy. The proposed framework comprises four mutually reinforcing dimensions, foundational, procedural, critical and strategic, and conceptualises AI literacy as a transversal component of legal translation competence that fosters digital resilience. It further discusses the pedagogical implications of this framework by proposing classroom activities designed to develop AI literacy in legal translator education, enabling future translators to integrate AI critically, responsibly and in accordance with professional standards.
[HC-13] Emotion Dynamics in Social Deception Games: Analysis of Professional and Nonprofessional Players through Electrodermal Activity in Werewolf Games
链接: https://arxiv.org/abs/2608.04605
作者: Sho Mitarai,Chang Liu,Goshiro Yamamoto,Nagisa Munekata
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:The development of AI systems capable of emotionally resonant communication remains a significant challenge. This study examines how humans influence emotions in social deception games by comparing professional and non-professional players. We measured electrodermal activity during gameplay to capture physiological emotional responses and analyzed communication patterns during periods of high emotional arousal. Our results revealed distinct communication strategies: professional players maintained persuasion-based approaches under high arousal, while nonprofessional players shifted toward information-focused communication. Statistical analysis confirmed significant differences in expression patterns between expertise levels. Professional players exhibited more stable emotional states during gameplay, indicating better emotional regulation. These findings inform the design of AI systems that can adapt their communication strategies based on recipient characteristics, advancing the development of emotionally intelligent artificial agents.
[HC-14] Super-Gaussian: Interactive Scene Editing for 3D Gaussian Splatting and NLI-Based Volume Visualization in Virtual Reality IEEE-VIS2026
链接: https://arxiv.org/abs/2608.04475
作者: Suemin Jeon,Kaiyuan Tang,Chaoli Wang,Won-Ki Jeong
类目: Human-Computer Interaction (cs.HC)
备注: IEEE VIS 2026 accepted
Abstract:Despite the promise of virtual reality (VR) for intuitive spatial interaction, volume visualization (VolVis) in VR remains constrained by high rendering costs and motion discomfort. Recent advances have shown that representing volumetric scenes with 3D Gaussian splatting enables high-performance rendering, making this representation well-suited for VR. However, existing Gaussian-based scene editing workflows remain limited by slow offline segmentation and fatigue-inducing manual selection. To address these challenges, we present Super-Gaussian, a novel VolVis framework that enhances scene editing and interaction in VR through intuitive 3D Gaussian selection and natural language interaction (NLI). Our approach groups Gaussian primitives into higher-level units via feature-aware clustering, enabling efficient selection of complex volumetric regions, such as tumors in medical images or filaments in cosmological data, without point-by-point interaction. Building on this, we introduce a hierarchical select-and-refine workflow that combines random-walk-based region propagation, cluster selection, and point refinement, allowing users to progressively specify regions of interest with reduced effort. We further support on-the-fly text labeling of selected regions using NLI, allowing users to semantically query, interpret, and manipulate content within a visualization-perception-action loop. By integrating multimodal interaction, including speech, visual feedback, and spatial manipulation in VR, our framework supports intuitive exploration, editing, and scientific analysis of volumetric data. We demonstrate the effectiveness of Super-Gaussian through four case studies, quantitative selection benchmarks against existing Gaussian-based techniques, and system-level evaluations. Implementation details and experiments can be found on the project page: this https URL
[HC-15] Revisiting Channel Effectiveness: A Multi-Dimensional Evaluation with Primitive Visual Stimuli IEEE-VIS2026
链接: https://arxiv.org/abs/2608.04435
作者: Soohyun Lee,Seokhyeon Park,Minsuk Chang,Jinwook Seo
类目: Human-Computer Interaction (cs.HC); Graphics (cs.GR)
备注: Accepted to IEEE VIS 2026; to appear in IEEE Transactions on Visualization and Computer Graphics. 9 pages (11 with references) plus supplementary material
Abstract:Established channel effectiveness rankings primarily assess magnitude estimation accuracy in complete chart contexts, often neglecting other perceptual tasks such as discriminability, separability, and pop-out. To address this gap, we conducted crowdsourced experiments on seven core visual channels (position, length, tilt, area, curvature, luminance, and saturation) using primitive visual stimuli, a set of visual marks without chart-specific scaffolding to isolate channel-level variation. We evaluated these channels across four perceptual tasks (accuracy, discriminability, separability, and pop-out) and found that channel effectiveness is fundamentally multi-dimensional, with rankings shifting substantially across tasks. For instance, while spatial channels maintain an overall advantage, accuracy depends strongly on whether a fixed spatial anchor is available. Discriminability varies dramatically across channels and value ranges, a pattern we formalized with a novel Anchored Harmonic Weber model. Pairwise channel interactions are often strongly asymmetric. Finally, we identify a dissociation between estimation accuracy and preattentive detection: length shows only moderate detection effectiveness despite top-tier accuracy, while area achieves the highest detection rates despite poor quantitative accuracy, though the latter advantage may partly reflect stimulus-level cues. We synthesize these findings into a scenario-driven perspective for context-sensitive channel selection.
[HC-16] Preference-Driven Online Adaptation for Personalized Interaction Initiation in Proactive AI Assistants
链接: https://arxiv.org/abs/2608.04416
作者: Yufeng Wang,Wei Zhang,Zhiquan Wen,Jinwu Hu,Linhui Xiao,Tianlu Pan,Qingfang Zheng,Mingkui Tan
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:AI assistants are typically reactive, relying on users to initiate interactions. Proactive assistants go beyond this paradigm by autonomously initiating interactions based on users’ activity contexts. However, appropriate interaction timing is user-specific and difficult to determine in advance, while online feedback offers valuable signals for personalization. Direct feedback-driven adaptation is therefore appealing, but remains challenging due to sparse interaction-worthy moments scattered across fine-grained user states. To address the issues, we propose Evidence-driven Online Preference Adaptation (EOPA), which grounds a user’s interaction-timing preferences in measurable contextual evidence through two evidence carriers: temporal preference anchors and evidence-bearing activity prototypes. At each polling step, EOPA derives temporal and activity evidence from the carriers through user-prior-smoothed evidence estimation and uncertainty-guided evidence scaling, and adaptively fuses the evidence for interaction-or-silence decisions. When interaction is selected, an LLM uses high-quality historical responses as demonstrations to generate a context-aware response that better reflects user preferences. EOPA updates its evidence carriers and decision parameters from received online feedback without LLM-based reasoning or retraining. Extensive experiments on a ProPerSim-based benchmark show that EOPA improves the interaction-timing F1 score by 19.80 points over the strongest baseline in our experiments, substantially reduces inference latency for both silence and interaction steps, and lowers the average daily adaptation time from 11.41 to 0.39 seconds.
[HC-17] IntentLint: Supporting Intent Scaffolding and Prompt-time Linting in Human-AI Collaborative Data Analysis
链接: https://arxiv.org/abs/2608.04331
作者: Felicia Li Feng,Jian Zhao,Anamaria Crisan
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:In human-AI collaborative data analysis, as analyses rapidly evolve, the artifacts meant to capture shared understanding often become incomplete or difficult to interpret, leading to undocumented assumptions, cross-user misaligned intent, context-poor prompts, and unwanted agent behaviors. To address these challenges, we introduce a rule-based coordination layer with two interaction mechanisms, intent scaffolding and prompt-time linting, that make analytic intent explicit and actionable during human-AI collaborative data analysis. We implement them in IntentLint, a proof-of-concept system that infers analytic intent from shared notebooks, represents it as structured, editable rules, and checks users’ prompts against shared rules. IntentLint helps analysts externalize and refine their intent and proactively checks prompts for potential conflicts. A study with 16 data analysts shows that IntentLint improves awareness of collaborators’ intent and encourages reflection on analytic strategies, and provides design implications for supporting more aligned and transparent human-AI collaborative data analysis.
[HC-18] An immersive micro-manipulation system using real-time 3D imaging microscope and 3D operation interface for high-speed and accurate micro-manipulation
链接: https://arxiv.org/abs/2608.04300
作者: Kenta Yokoe,Tadayoshi Aoyama,Toshiki Fujishiro,Masaru Takeuchi,Yasuhisa Hasegawa
类目: Human-Computer Interaction (cs.HC)
备注: Published version. Originally published in ROBOMECH Journal (Springer, Open Access, CC BY 4.0)
Abstract:The use of intracytoplasmic sperm injection (ICSI), an assisted reproductive technique (ART), is increasing widely. ICSI is currently performed by specially skilled embryologists. However, with the increasing demand for ART, the shortage of skilled embryologists has become a problem. Therefore, we propose an immersive micromanipulation system that requires no special skills for efficient and accurate micromanipulation. Our proposed system is composed of a real-time three-dimensional (3D) imaging microscope and 3D operation interfaces. The 3D operation interfaces are stationary pen-type or wearable glove-type interfaces. In this system, an operator wearing a head-mounted display (HMD) and using 3D operation interfaces is immersed in a virtual micromanipulation space. The operator can move the pipettes by 3D operation interface and freely change the viewpoint. We verified that the proposed system improves the speed and accuracy of operating a pipette through two types of experiments with subjects.
[HC-19] LEGOUI: Designing with UI-DSL Bricks to Balance Transparency and Controllability
链接: https://arxiv.org/abs/2608.04293
作者: Yinsi Zhou,Mingyue Yuan,Hongyue Xu,Jieshan Chen,Dong Wen,Shidong Pan,Xiwei Xu,Wenjie Zhang,Aaron Quigley,Zhenchang Xing,Gelareh Mohammadi
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Generative user interface design tools enable rapid prototyping but often operate as black boxes with limited transparency and controllability. When outputs diverge from the designer’s intent, users are left tweaking prompts via trial-and-error with little insight into the model’s reasoning. We present LegoUI, a staged generative framework that structures the interface design process into sequential, interpretable steps along key design dimensions, capturing each step’s result in a UI domain-specific language (UI-DSL) enriched with provenance. This approach exposes the model’s intermediate reasoning and enables user intervention and iterative refinement. In a technical evaluation on 40 real-world design prompts, LegoUI’s requirement analysis stage captured explicit requirements with over 95% accuracy, near-complete coverage, and zero redundancy. In user studies, participants using LegoUI reported significantly greater transparency, controllability, and alignment with their intent compared to existing one-shot generative UI tools.
[HC-20] Compass: Continuously Aligning Social Media Feeds via In-Situ Reflections
链接: https://arxiv.org/abs/2608.04274
作者: Aadit Barua,Leijie Wang,Amy X. Zhang
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: UIST 2026
Abstract:Social media recommendation feeds often optimize for users’ immediate impulses rather than preferences they would hold after deeper reflection. Some systems address this misalignment by incorporating users’ explicit preferences via a configuration page or in-feed controls instead of just behavioral signals. However, users typically have evolving preferences, and their stated preferences and behavior naturally diverge, necessitating continuous reflection and feed realignment. But existing strategies require the user to take initiative and are often effortful; as a result, in practice they are rarely invoked. We present Compass, a system that aligns a user’s feed with their reflective preferences by helping users reflect on and articulate their preferences given their behavior. To enable continuous reflection during everyday browsing, Compass surfaces in-situ reflections via lightweight notifications, while feed alignment is achieved by periodically simulating behavioral signals and directly manipulating feed content. We embedded Compass within YouTube Shorts and compared it against a baseline without continuous support through a 10-day field study (N=15). We found that Compass promoted more reflective and purposeful feed consumption, iterative preference adjustment, and stronger feed alignment, without sacrificing the casual nature of feed browsing.
[HC-21] Enacting Constructive Conflicts with AI Agents to Enhance Reconsideration among Novice Interaction Designers
链接: https://arxiv.org/abs/2608.04166
作者: Howard Ziyu Han,Nikolas Martelaro
类目: Human-Computer Interaction (cs.HC)
备注: 8 pages, 4 figures, conditionally accepted to Human-Agent Interaction 2026
Abstract:Generative AI agents are increasingly used in interaction design to facilitate ideation and offer critique, often following their own internal reasoning. These interactions tend to add design ideas and expand the design space. Our work explores an antagonistic role for design agents, prompting designers to engage with stakeholder tension. We built an AI agent inspired by adversarial design theory that enacts constructive conflict. We examine the agent’s influence in a between-subjects experiment with 45 design students across three conditions: Self Reflection (unsupported review of the design proposal), Stepwise Guidance (written prompts that walk designers through a constructive-conflict framework), and Interactive Engagement (an AI agent that enacts the constructive-conflict framework interactively by synthesizing stakeholder pushback). The latter two conditions share the framework but differ in whether it is self-enacted or agent-enacted. Results show that, compared with Self Reflection, both the Stepwise Guidance and Interactive Engagement groups reported significantly higher self-reconsideration and made more improvements to their design proposals. Compared with Stepwise Guidance, the antagonistic agent introduced more conflictual perspectives, and participants in the Interactive Engagement condition generated and discarded more ideas. These findings suggest that agent-enacted constructive conflict can turn reconsideration into concrete design actions and deepen engagement with divergent stakeholder perspectives.
[HC-22] Agent Forge: An Immersive Role-Playing Platform for Learning Agent ic Software Engineering
链接: https://arxiv.org/abs/2608.04148
作者: Zihan Fang,Yueke Zhang,Yu Huang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: 7 pages, 2 figures
Abstract:Agentic AI is increasingly used to coordinate planning, implementation, review, and testing in software development, yet it often offers limited transparency into its decisions and interactions. Many such systems also assume that users can effectively guide the AI’s decisions and validate its outputs. This assumption poses a particular challenge for novices, who must simultaneously learn how agentic AI works, how to collaborate with it effectively, and how to evaluate its outputs critically. To address this challenge, we present \textitAgentForge, an immersive learning system in which novices take on one of four software-engineering roles: Task Planner, Patch Author, Code Reviewer, or Test Runner, within a multi-agent code-repair workflow. In each practice session, the novices perform their chosen role while AI agents perform the remaining three. Through role-based scaffolding and metacognitive support, AgentForge clarifies role-specific responsibilities, makes agent coordination and intermediate artifacts visible, and encourages novices to monitor and evaluate their decisions. In a study with 37 novice developers, participants achieved high task-completion rates with AI-agent support. However, interaction demands differed significantly across practices: the Code Reviewer practice required more interaction turns, reroutes, and completion time ( p_\mathrmadj = .004 ) and was perceived as the most challenging. Participants nevertheless reported significant gains in their understanding of software repair and agent collaboration ( p_\mathrmadj .001 ). These findings suggest that AgentForge can help novices develop practical software-engineering skills while learning to collaborate with agentic AI more critically and effectively.
[HC-23] Echoes in the Sky: Computational Thematic Analysis of Online Public Discourse on Bluesky Across Trumps Reelection
链接: https://arxiv.org/abs/2608.04120
作者: Qile Wang,Ali Salloum,Carolina Coimbra Vieira,Benjamin E. Bagozzi,Mikko Kivelä,Kenneth E. Barner,Matthew Louis Mauriello
类目: Human-Computer Interaction (cs.HC)
备注: Accepted at ASONAM 2026, to appear in the Springer proceedings
Abstract:As political disruption intensifies online discourse, Bluesky has become an important platform for political discussion and public reaction. In this study, we examine large-scale discourse on Bluesky related to U.S. policy developments associated with the Trump administration. Using the historical retrieval API, we collected all available posts matching Trump and related keywords from 2019 to 2026, yielding 38.5 million posts. We leverage a large language model (LLM)-assisted clustering pipeline, combined with human validation, to identify 14 interpretable thematic domains in English-language posts and 19 thematic categories across 258 executive orders (EOs) signed between January 20, 2025, and May 1, 2026. Our findings identify several dominant themes in Bluesky discourse, including executive governance, political identity, and national security, as well as recurring themes in EOs, including executive task forces, border enforcement, and foreign policy. We also find substantial variation in the persistence and volatility of issue attention, accompanied by an increasing proportion of negative sentiment over time. The dataset and resources are publicly available at this https URL
[HC-24] Deep Learning for Real-Time Sound Order Recognition in Human-Robot Interaction
链接: https://arxiv.org/abs/2608.04072
作者: Rezaul Tutul,Usaid Khan,Andre Jakob,Ilona Buchem
类目: ound (cs.SD); Human-Computer Interaction (cs.HC)
备注:
Abstract:Recognizing the temporal order of overlapping sounds is an underexplored challenge in human-robot interaction (HRI), with direct relevance to applications such as first responder detection systems. This paper presents a deep learning framework for real-time sound order recognition using recordable buzzers that emit distinct non-verbal sounds (cat meows, dog barks, helicopter noises). A multi-branch convolutional neural network (CNN) processes Mel spectrograms, Mel-frequency cepstral coefficients (MFCCs), and short-time Fourier transform (STFT) features, with an attention-based fusion mechanism to emphasize critical temporal cues. Experiments were conducted under same-amplitude, varied-amplitude, and unseen sound conditions. The proposed system achieved 99% accuracy in balanced overlaps, 91% under amplitude variation, and 74% on unseen test data with normalization. These results demonstrate that deep learning can reliably recognize sound order in overlapping conditions, supporting practical HRI scenarios. While experiments were conducted on carefully controlled synthetic overlaps, we additionally report latency benchmarks demonstrating real-time feasibility and provide an extended discussion on generalization, ecological validity, and deployment challenges in real-room environments.
[HC-25] Reconstructing Persistent Worlds from Narratives for Narrative-Grounded Interactive Experiences
链接: https://arxiv.org/abs/2608.04037
作者: Yi-Chun Chen
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Graphics (cs.GR); Human-Computer Interaction (cs.HC)
备注:
Abstract:Designing narrative-grounded interactive experiences remains labor-intensive because interactive content must align with the underlying world implied by the narrative. Existing approaches formulate problems such as narrative planning, scene generation, and gameplay generation, each constructing computational representations tailored to specific downstream tasks rather than explicitly reconstructing and maintaining the persistent world that grounds them. We investigate reconstructing explicit persistent worlds from narrative descriptions as the central computational objective for narrative-grounded interactive realization. Rather than treating the world as an implicit by-product of downstream generation, our approach reconstructs and maintains persistent entities, locations, semantic relationships, and evolving world states while inferring only the contextual information required to support coherent interactive experiences. To investigate this perspective, we develop a reference prototype that reconstructs structured persistent world representations from narrative descriptions and subsequently instantiates playable tile-based environments. Through three representative case studies spanning a procedural scenario, an original fantasy narrative, and an adapted public-domain story, we demonstrate the feasibility of reconstructing persistent worlds and show how a shared world representation supports coherent gameplay while remaining grounded in the source narrative. By explicitly reconstructing persistent worlds prior to interactive realization, this work bridges computational narrative understanding and interactive content generation, providing a semantic foundation for AI-assisted game authoring, mixed-initiative design, educational simulations, and narrative-grounded interactive experiences. Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Graphics (cs.GR); Human-Computer Interaction (cs.HC) Cite as: arXiv:2608.04037 [cs.CL] (or arXiv:2608.04037v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2608.04037 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Yi-Chun Chen [view email] [v1] Mon, 3 Aug 2026 20:06:13 UTC (2,391 KB)
计算机视觉
[CV-0] CoCo-IR: Contextual Composed Image Retrieval ECCV2026
链接: https://arxiv.org/abs/2608.05149
作者: Shengcao Cao,Tanmaya Shekhar Dabral,Zhongli Ding,Madhuri Shanbhogue,Kaifeng Chen,Zhe Li,Mojtaba Seyedhosseini,Yu-Xiong Wang,Liang-Yan Gui
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV 2026
Abstract:Current instruction-based image retrieval systems are powerful but limited to single-turn interactions, failing to capture the iterative nature of complex, real-world visual searches. To overcome this limitation, we introduce Contextual Composed Image Retrieval (CoCo-IR), a novel task that enables users to progressively refine search results through interactions. We address this new task by proposing a new model based on a Large Multimodal Model (LMM) that functions as a context-aware reasoner for CoCo-IR. Our model interprets the entire interaction history to generate Transformable Image Embeddings (TIE) that evolve across turns. To fuel the model training without expensive human annotations, we develop a fully autonomous, scalable data engine that leverages LMMs to generate high-quality contextual retrieval data, and uses model-guided verification to mine challenging hard negatives. Extensive experiments demonstrate that our approach establishes new state-of-the-art performance: We achieve 39.4 mAP@5 on the challenging single-turn benchmark CIRCO; furthermore, on our new CoCo-IR benchmark, our model maintains robust performance with 44.1 R@1 on 4-turn dialogues, dramatically outperforming existing methods (28.2 4-turn R@1) that fail to handle multi-turn context. Project page: this https URL.
[CV-1] Objects as Audio-Visual Modal Sound Fields ECCV2026
链接: https://arxiv.org/abs/2608.05145
作者: Zisen Shao,Zihao Wei,Derong Jin,Ruohan Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ECCV 2026, Project page: \href
Abstract:While modern 3D reconstruction excels at modeling object geometry and appearance, it largely ignores the rich acoustic cues revealed through physical interaction. Object impact sounds convey material, stiffness, and structural properties that complement vision, yet existing impact sound modeling approaches either rely on expensive physics-based simulation or require large datasets to generalize in a purely data-driven manner. We introduce Audio-Visual Modal Sound Field (AV-MSF), a novel object-level acoustic representation reconstructed from multi-view images and only a few impact sound recordings. AV-MSF builds on 3D Gaussian Splatting integrated with dense 3D visual feature to provide a strong geometry-aware prior, and represents the impact sound field using compact, physically meaningful modal parameters, enabling robust few-shot reconstruction. Experiments on two real-world datasets show that AV-MSF achieves state-of-the-art impact sound rendering, outperforming both physics-based and data-driven baselines. Furthermore, we demonstrate downstream applications enabled by our representation, including contact localization and object sound editing.
[CV-2] SmartMage: Dynamic Modality Orchestration for 3D Scene Understanding
链接: https://arxiv.org/abs/2608.05137
作者: Yue Zhang,Yingzhao Jian,Yunqiu Xu,Xiaoxiao Sun,Hehe Fan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Understanding 3D scenes is fundamental to embodied intelligence, requiring joint reasoning over heterogeneous information from multiple modalities, including visual and geometric cues. However, the relevance of these modalities often varies across queries. Existing Multimodal Large Language Models (MLLMs) typically rely on fixed modality combinations, overlooking query-dependent modality needs. Such a rigid design can introduce semantic noise from irrelevant modalities while underutilizing more informative ones, leading to wasted computation and diluted reasoning. To address these challenges, this paper proposes SmartMage, a unified MLLM that dynamically orchestrates heterogeneous modalities for semantic-aware 3D scene understanding. Specifically, SmartMage incorporates: (1) a Semantic-guided Modality Adaptive RouTng (SMART) module that selects task-relevant modalities using semantic priors, text-modality alignment, and modality quality; and (2) a Modality-Aware Gating Expert (MAGE) module that leverages modality priors to guide expert activation, fostering adaptive specialization in multimodal reasoning. Empirically, SmartMage achieves state-of-the-art performance across five 3D scene understanding benchmarks, and attains competitive results on RGB-only video understanding benchmarks. In our diagnostic benchmark ScanFacet, tasks are divided into fine-grained semantic categories, enabling analysis of modality combinations preferred by each semantic type. The observed modality-semantic patterns provide further evidence of SmartMage’s effectiveness. Project page: this https URL.
[CV-3] Predicting Brain Morphometry with MT-GNN: Mesh Evolution in Continuous Time with Graph-Based Metric Tensor Embeddings
链接: https://arxiv.org/abs/2608.05132
作者: Hao Ding,Daniel Semchin,Paul M. Thompson,Boris Gutman
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Predicting how a subcortical structure’s shape will evolve from a few prior scans could support prognosis and clinical-trial enrichment. Existing longitudinal mesh predictors either extrapolate shape trajectories via high-dimensional embeddings or regress vertex deformations directly. We instead predict the surface’s intrinsic geometry in continuous time: a single per-structure graph network predicts the future per-vertex first fundamental form (metric tensor) for an arbitrary causal multiple-visit history and an arbitrary prediction horizon, conditioned on a Fourier encoding of the lead time. The predicted metric is decoded into a surface by a differentiable As-Rigid-As-Possible solver, and the model is trained end-to-end on the rigid-aligned vertex error. Training through the reconstruction keeps the decoded prediction a valid surface and consistently improves it. On 14 subcortical structures from the ADNI dataset, the proposed mesh evolution model (MT-GNN) predicts best among the evaluated methods at every horizon ( -2.29% mean vertex error vs. the temporal mean, p=6.1\times10^-5 , beating it on 14/14 structures), ahead of geodesic shape regression (DCM, -0.19% ) and a mesh transformer (TransforMesh, -0.45% ; p=1.2\times10^-4 ), with the lead widening as the horizon grows.
[CV-4] OPD-V: Visual On-Policy Self-Distillation with Modality Balance
链接: https://arxiv.org/abs/2608.05131
作者: Aniri,Jinhe Bi,Peng Liao,Zengjie Jin,Volker Tresp,Fei Shen,Yunpu Ma,Tat-Seng Chua
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:On-Policy Self-Distillation (OPSD) has become a standard post-training approach for improving visual reasoning in multimodal large language models (MLLMs). Existing methods draw privileged information from diverse input sources to guide self-distillation. Yet these designs overlook Modality Imbalance, a challenge inherent to MLLM reasoning. When textual information dominates generation, the model cannot fully integrate its multimodal input. Consequently, carefully designed privileged information remains underused, limiting the effectiveness of OPSD. To examine this limitation, we construct a Positive Teacher with the Zoom-In Image and a Negative Teacher with the Mask Image, which exhibit different degrees of Modality Imbalance. Changes in their reasoning correctness and token logits reveal that Modality Balance can itself serve as privileged information. Motivated by this finding, we introduce OPD-V, a visual OPSD paradigm that instantiates such information through the Positive Teacher and Negative Teacher. Positive Modality-Balance Logits Margins define a Modality-Balance Trust Region that selects the on-policy tokens used for self-distillation. Experiments across 6 benchmarks, 4 MLLM backbones, and 5 post-training methods show that OPD-V consistently improves reasoning performance while reducing training cost.
[CV-5] IRIS: A Visual Cortex-Inspired Framework for Analyzing Orientation Selectivity in Vision Transformers
链接: https://arxiv.org/abs/2608.05122
作者: Vaishnavi B Mohan,Vijayakrishna Naganoor,Yashas Annadani,Shashank Hegde
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision transformers (ViTs) have become the de facto standard for image encoding across many perception tasks. Despite their empirical success, it remains mechanistically unclear how they encode low-level features, given their lack of inductive biases: ViTs process information globally rather than relying on local structure. Biological visual systems, in contrast, build low-level features, such as orientation selectivity in the primary visual cortex, by combining information from small, localized regions of the visual field. These features are general-purpose representations, shared and required across multiple specialized neural pathways, unlike higher-level, task-specific semantic features. This raises the question if such biologically-grounded features arise in ViTs. In this work, we systematically study how orientation selectivity emerges in ViTs by introducing a suite of neuroscience-inspired metrics: representational similarity score (RSS), orientation recruitment score (ORS), and orientation tuning bandwidth to quantify how orientation is encoded in representational geometry and as a function of model depth. Through extensive analysis, we find that: (1) the training paradigm is the strongest determinant of orientation selectivity, with models sharing an objective, peaking at comparable relative depths regardless of scale (2) many units are orientation-selective early in training, with early-to-middle layers recruiting more such units over time, while deeper layers lose selectivity and broaden their tuning toward semantic encoding and (3) our metrics offer a mechanistic heuristic for how many layers to unfreeze for best downstream generalization. Our framework presents a way to track biologically-grounded features during ViT training, probes how desired properties are encoded in transformer representations, and builds a systematic understanding of how ViTs generalize across tasks.
[CV-6] HexMIL: Hierarchical Attention MIL for Ante-Hoc Explainable Detection of AI-Manipulated CT Volumes
链接: https://arxiv.org/abs/2608.05101
作者: Orazio Pontorno,Luca Guarnera,Zahid Akhtar,Sebastiano Battiato
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Accepted at ACM Multimedia 2026 (MM '26)
Abstract:The emergence of medical deepfakes, i.e., medical images manipulated by deep generative models, poses a significant threat to clinical workflows. However, existing detectors suffer from two critical limitations: poor generalization to unseen generative architectures for manipulation detection and lack of interpretability. In this context, we present HexMIL (Hierarchical EXplainable Multiple Instance Learning), a mask-free medical deepfake detector that simultaneously addresses both limitations using only binary volume-level supervision. HexMIL decomposes each CT volume into a two-level hierarchy of patches and slices, aggregated via independent Gated Attention modules whose weights are directly combined into a full-resolution 3D attention volume that localizes the manipulated sub-region without any pixel-level annotation. Unlike post-hoc methods such as Grad-CAM, HexMIL’s attention weights constitute the exact forward computation driving the classification decision, providing ante-hoc and structurally faithful spatial attribution. We evaluate HexMIL on M3DSynth and CT-GAN datasets under a rigorous cross-generator generalization protocol, training on a single generative architecture and testing on unseen ones. HexMIL outperforms all baselines by +9.1 AUC and +9.4 F1 in out-of-domain classification, and achieves the best average IoU and Pointing Game score in localization. Project page: this http URL.
[CV-7] Lesion Detection in CT with Frozen Self-Distilled Features: SALT a Spatially Adaptive Label-Guided Temperature
链接: https://arxiv.org/abs/2608.05100
作者: Mahmut S. Gokmen,Evan W. Damron,Mitchell A. Klusty,Caroline N. Leach,Emily B. Collier,V. K. Cody Bumgardner
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Self-supervised pretraining objectives are spatially uniform: the teacher temperature and the per-patch loss weight are identical everywhere in the image, so a lesion a few patches wide contributes no more to the training signal than the surrounding parenchyma. Prior work biases the views toward annotated regions, which changes what the model sees but adds no pressure on the objective. We instead condition the targets of self-distillation, a method we call SALT (Spatially Adaptive Label-guided Temperature). Weak, box-derived labels, available only during pretraining, define a compact region on the encoder’s patch grid, inside which the teacher’s softmax temperature is sharpened and the masked-patch loss is up-weighted. The objectives, the masking policy and the centering statistics are otherwise unchanged, and at every downstream use the encoder is a plain feature extractor with no labels and no conditioning. We evaluate by freezing the encoder and training only a lightweight multi-depth CenterNet-style head, detecting lesions in 3D on four CT cohorts, and we isolate the mechanism against a backbone identical in architecture, pretraining data, schedule and label-guided cropping but with no target conditioning. We report patch-level separability, 3D detection stratified by cohort and by lesion size, box quality, and a detector-free probe in which a single frozen patch embedding re-identifies a lesion in a follow-up scan without registration, masks or fine-tuning. Because the conditioning is expressed through a spatial indicator rather than through label semantics, the formulation admits any weak spatial annotation; we instantiate and validate it for lesions.
[CV-8] Bag-of-Visual-Words for Spatial Mapping of Lung Adenocarcinoma Growth Patterns
链接: https://arxiv.org/abs/2608.05074
作者: Darya Ardan,Valentin Oreiller,Henning Müller
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 2 figures. Accepted at the 7th International Conference on Medical Imaging and Computer-Aided Diagnosis (MICAD 2026)
Abstract:Spatial mapping of lung adenocarcinoma (LUAD) growth patterns across whole slide images (WSIs) requires resolving architectural context at the region level, yet existing methods operate at the individual tile level and produce generic morphological clusters rather than clinically defined pattern maps. We propose a weakly supervised Bag-of-Visual-Words (BoVW) pipeline that learns a visual vocabulary from frozen foundation model embeddings extracted from a small set of annotated regions of interest (ROIs). Pattern prototypes are constructed as mean BoVW histograms of same-label ROIs and used for nearest-prototype classification of sliding-window regions under Jensen–Shannon divergence. The resulting predictions are projected onto the WSI tile grid to produce interpretable spatial pattern maps. We evaluate the method on 87 CPTAC-LUAD patients using three foundation model encoders and multiple vocabulary sizes on two clinically motivated tasks. For tumour/healthy classification, the best configuration achieves a balanced accuracy of 0.974 with H-Optimus-1, approaching the 0.987 obtained by a supervised SVM trained on mean-pooled WSI embeddings. For binary histologic grade classification, the BoVW pipeline achieves higher balanced accuracy than the supervised baseline for all encoders, suggesting that ROI-level pattern decomposition preserves grade-relevant heterogeneity that is attenuated by global mean pooling.
[CV-9] HelloWorld: Enabling Socially Interactive Characters in Video World Models
链接: https://arxiv.org/abs/2608.05070
作者: Liangyang Ouyang,Ruicong Liu,Xuangeng Chu,Kaipeng Zhang,Yoichi Sato
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Despite the remarkable recent progress of video world models, social interaction between users and the characters within these worlds remains unsupported. To fill this gap, we present HelloWorld, a video world model that enables social interaction with in-world characters. With a single button press, users can prompt the on-screen character to respond toward the camera, e.g., turning to the viewer, waving, nodding, or speaking a short greeting. To make these interactions natural, we propose a self-distillation pipeline that finetunes the video generation model on data synthesized by itself. Each synthesized clip contains both social interactions and camera motion, allowing the model to learn camera-pose conditioning without degrading interaction quality. At inference, we further introduce a training-free module that determines when the interaction occurs. Upon a button press, it modulates the cross-attention masks of the DiT so that the interaction-related text prompt attends only to the frames within the press window, temporally localizing the character’s response. We further build HelloWorldBench, a 400-sample benchmark with three social interaction metrics alongside three conventional metrics, for evaluation. Experiments demonstrate that HelloWorld surpasses a variety of baselines in interaction quality, while maintaining state-of-the-art picture aesthetics and camera-pose following. Project page: this https URL
[CV-10] VQ-VAD: Vector-quantized Motion Representation Learning for Human-centric Video Anomaly Detection
链接: https://arxiv.org/abs/2608.05069
作者: Narges Rashvand,Ghazal Alinezhad Noghre,Shanle Yao,Gabriel Maldonado,Hamed Tabkhi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Video Anomaly Detection (VAD) is inherently challenging due to the scarcity of anomalies and the large visual variability in surveillance footage, including changes in lighting, viewpoint, and human appearance. To mitigate visual noise and address privacy concerns, recent work has shifted to pose-based VAD, which focuses on motion dynamics rather than raw video data. However, existing pose-based approaches model human behavior in continuous latent spaces, limiting their ability to learn compact motion patterns necessary for robust behavior analysis. We address this by proposing Vector-Quantized Video Anomaly Detection (VQ-VAD), a novel human-centric anomaly detection framework that learns discrete motion representations. VQ-VAD adapts Vector-Quantized GAN (VQ-GAN), originally developed for image generation, to operate on keypoint sequences and construct a motion codebook of normal behavior. Trained exclusively on normal motion sequences, VQ-VAD detects anomalies by identifying high reconstruction errors when an observed motion sequence cannot be mapped to the learned codebook. We conduct extensive experiments across three complementary evaluation settings, including in-domain, cross-domain, and cross-dataset generalization, on four anomaly detection benchmarks. VQ-VAD achieves strong in-domain accuracy (81.83% on HR-SHT [15]), effective cross-domain transfer from CMU Panoptic [14] (76.69% on HR-SHT [15] without retraining), and competitive cross-dataset robustness. The code base for this work is available at this https URL.
[CV-11] Beyond Reprojection Error: Camera Calibration with 3D Targets
链接: https://arxiv.org/abs/2608.05066
作者: Dennis Ruppel,Hasan Kutlu,Kai A. Neumann,Martin Knuth,Pedro Santos,Andreas Weinmann,Arjan Kuijper
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 7 figures, 2 tables. To appear in the proceedings of Computer Graphics International (CGI 2026)
Abstract:In 3D reconstruction, camera calibration is an essential element for achieving high fidelity and accuracy of the reconstructed geometry. While existing approaches rely upon 2D planar calibration, this work proposes a framework tailored for 3D reconstruction that is based on predicting scene rays, which adds flexibility to the reconstruction pipeline and enables the use of recent advances in camera models. Novel metrics, reconstruction and intersection error, derived from predicted scene rays are employed in combination with a bootstrapping procedure that statistically evaluates different calibration objects and calibration pipelines for both intrinsic and extrinsic camera parameters. The results show that the generalized distortion model more faithfully captures physical camera effects and yields an improvement in calibration accuracy. Reprojection error is shown to be a potentially misleading indicator of 3D accuracy, and the proposed ray-based metrics provide a more holistic assessment. An icosahedron calibration target is designed to enrich calibration information for 3D reconstruction together with a ring-feature-based detector. The icosahedral target yields approximately 40% lower mean intersection and more stable calibration across bootstrap trials on synthetic data, while real-data performance demands very tight fabrication tolerances.
[CV-12] OmniEdit-Bench: A Comprehensive Benchmark for Instruction-based Video Editing
链接: https://arxiv.org/abs/2608.05049
作者: Chenxuan Miao,Yutong Feng,Yi Lu,Yunfeng Yan,Donglian Qi,Shiwei Zhang,Yu Liu,Xi Chen,Hengshuang Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Instruction-based video editing (IVE) is an emerging field with broad applications, yet evaluating editing models remains challenging. Existing benchmarks suffer from two major limitations: limited task coverage inherited from image editing, which overlooks video-specific dimensions, and inadequate metrics that fail to measure instruction fidelity, allowing incorrect edits to receive high scores due to strong visual priors from the original video. To address these issues, we introduce a comprehensive and structured benchmark for IVE. Our benchmark decomposes editing tasks into multiple video-specific dimensions, including spatial, temporal, audio, and reference-based editing, extending beyond conventional frame-level evaluation. It also distinguishes explicit and implicit instructions and incorporates reasoning-based scenarios to better reflect real-world requirements. Furthermore, we propose an evaluation framework that assesses editing quality from four complementary dimensions: accuracy, preservation, realism, and consistency, using both human judgments and state-of-the-art vision-language models. To emphasize instruction fidelity, we introduce an accuracy-aware penalty mechanism that conditions other scores on accuracy, preventing visually plausible but incorrect edits from receiving inflated evaluations. Extensive experiments on representative open-source and commercial models show that current IVE models remain far from satisfactory. OmniEdit-Bench provides a comprehensive and reliable testbed for evaluating instruction-based video editing and offers insights into future research directions.
[CV-13] owards Physics of Multimodal Pretraining: Knowledge Flow Modality Synergy Early Unification and Recipes
链接: https://arxiv.org/abs/2608.05000
作者: Junlin Han,Shengbang Tong,David Fan,Minghao Chen,Philip Torr,Filippos Kokkinos,Mike Lewis
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Multimedia (cs.MM)
备注: Project page: this https URL
Abstract:Vision offers a critical axis for advancing foundation models, driving a shift towards natively unified multimodal pretraining. Despite this momentum, the design space and the fundamental mechanisms of how modalities interact during unified training remain underexplored. We provide empirical clarity through a systematic exploration of multimodal pretraining. Our controlled experiments on both synthetic and large-scale real-world datasets yield four key insights into the physics of multimodal pretraining: (i) Knowledge Flow: We disentangle how language, visual understanding, and visual generation transfer knowledge across modalities, revealing distinct patterns of influence and asymmetry; (ii) Synergy vs. Competition: We show that data “complexity” largely determines whether modalities are synergistic, identify architectural choices that promote synergy: such as shared attention and normalization with modality-specific feed-forward layers, and find that these behaviors generalize across different visual tokenizer designs; (iii) Early Unification: Unifying modalities from the very early stages and training them jointly is shown to be more effective than late alignment or sequential training. This process uncovers a vision laziness phenomenon, where delayed integration leads models to rely on language priors; (iv) Recipes: We derive efficient pretraining recipes that achieve strong generative performance using only 5% of the compute budget. These core findings are subsequently validated at scale by training multiple 13.5B MoE models on 2T tokens. We hope this study provides a principled foundation for understanding and scaling multimodal pretraining.
[CV-14] Promptable Animal Pose Tracking Across Species ECCV2026
链接: https://arxiv.org/abs/2608.04995
作者: Le Li,Daniela Ivanova,Nicolas Pugeault
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted for presentation at the ECCV 2026 Workshop on CV4Ecology
Abstract:Animal pose estimation and tracking is important for wildlife monitoring and conservation research, and with limited expert time for labelling automated approaches are imperative. While human pose estimation and tracking has seen rapid progress thanks to large annotated datasets, animal pose remain challenging, due to large morphological and behavioural differences between species and limited annotated data. Existing approaches either optimise generic keypoint localisation from annotated datasets (such as APTv2) with poor generalisation, or track custom keypoints using visual tracking, at the cost of performance. In this paper, we demonstrate that vision foundation models trained on large datasets can be used effectively to track animal pose with limited labelled data. We propose two models, one unsupervised and the other supervised, to track user-selected keypoints in videos. The supervised approach delivers superior tracking accuracy by employing a keypoint prompt encoder to explicitly inject structural priors from a reference frame into feature matching. In parallel, the unsupervised route provides strong cross-species robustness by leveraging diverse foundation-model features for training-free correspondence matching. Extensive evaluation on challenging animal video benchmarks APTv2 and TigDog demonstrates that our framework achieves strong performance while maintaining an effective balance between accuracy and generalisation, offering a practical solution for real-world animal behaviour analysis and conservation applications.
[CV-15] ContextMaster: Interactive Multi-Shot Video Creation via Fixed-Budget Sparse Context Routing
链接: https://arxiv.org/abs/2608.04956
作者: Xu Guo,Zhengxuan Wei,Xinghui Li,Hanzhuo Huang,Xinyu Liu,Xiangyang Luo,Min Wei,Yiran Zhu,Qiulin Wang,Yulong Xu,Xintao Wang,Pengfei Wan,Qi Fan,Xiangwang Hou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Recent video models increasingly support generation, reference conditioning, and editing within a single model, yet typically expose them as separate operations over fixed inputs. Practical creation unfolds across multiple shots, requiring one model to generate from text, follow a reference, or edit source footage while maintaining shared history. We formalize this setting as interactive multi-shot video creation (IMVC) and introduce ContextMaster, a unified model with a role-aware context representation for these operations. An interactive model must retain access to an expanding history without allowing the context read cost at each denoising step to grow. ContextMaster combines reusable clean context states with fixed budget sparse context routing and uses ConstraintSink to keep task constraints visible. To address the dual challenges of sparse context access and inference with few denoising steps, we propose a two-stage privileged context distillation framework, which transfers full context behavior from a dense teacher through consistency distillation and then refines deployment rollouts with distribution matching. Experiments on the three primitive tasks demonstrate improved task fulfillment and consistency across shots over specialized baselines. User studies further validate flexibly composed workflows, while the model reaches 16 FPS on a single GPU.
[CV-16] owards Valid B-Rep Generation: Training-Free Wireframe Anomaly Detection and Repair AAAI2027
链接: https://arxiv.org/abs/2608.04955
作者: Jingyu Wu,Youcheng Cai,Tengyu Luo,Ligang Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: AAAI 2027 submission; 9-page main paper plus supplementary material
Abstract:Multi-stage boundary representation (B-Rep) generation leverages intermediate wireframes to synthesize CAD models. However, geometric and topological risks in these wireframes – such as self-intersections, edge collapses, and disconnected vertices – can propagate to invalid final B-Reps. Mitigating such failures by retraining large generative models is computationally prohibitive. We propose Wireframe Detection and Repair (WDR), a training-free framework that intervenes at the intermediate wireframe stage to improve downstream B-Rep validity. WDR features a Geometric-Topology Anomaly Detector (GTAD) that combines parallel VLM-based coarse screening with geometric and topological detectors to predict downstream invalidity risk and route generation to dedicated branches. An Energy-Guided Geometric-Topology Repair (EGGTR) module then performs detector-triggered guided regeneration through geometry and topology branches. By scaling test-time computation via Energy-Guided Resampling and training-free guidance for diffusion models, WDR can be integrated into autoregressive and diffusion pipelines without retraining. Extensive experiments demonstrate consistent improvements in kernel-checked validity while largely retaining the measured diversity and distributional quality of synthesized CAD models. The code will be made publicly available upon acceptance.
[CV-17] Unleashing the Potential of Vision-Language Models for Generalizable AI-Generated Image Detection
链接: https://arxiv.org/abs/2608.04935
作者: Weihan Cai,Hao Tan,Zichang Tan,Jun Wan,Xinping Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recent work has shown that a simple linear probe on frozen representations from modern vision foundation models (VFMs) can achieve state-of-the-art AIGI detection performance, substantially outperforming specialized detectors in challenging in-the-wild scenarios. This finding has established DINOv3 as the dominant foundation-model baseline for subsequent improvements. However, we find that the vision-language model Perception Encoder (PE) holds greater potential for AIGI detection, because its language-aligned representation preserves high-level provenance semantics. Specifically, PE exhibits stronger local provenance organization than DINOv3 in its frozen feature space. However, semantic-agnostic linear probing fails to exploit this structure, as PE-Linear still underperforms DINOv3-Linear by 4.1% on In-the-Wild. Based on this observation, we propose Semantic Prototype Calibration (SPC), which constructs category prototypes from forensic semantic information and calibrates them with supervised data. We apply SPC to PE and refer to the resulting detector as PE-SPC. Our analysis shows that this simple design achieves stronger generalization. Across cross-generator, post-processing, and in-the-wild benchmarks, PE-SPC surpasses the previous DINOv3 baseline and achieves new state-of-the-art results.
[CV-18] An active-learning framework for real-time depth perception from monocular vision streams
链接: https://arxiv.org/abs/2608.04917
作者: Xiaorong Zeng,Weiqiang Chen,Peng Shi,Liang Su,Zirui Wang,Xuewu Ji,Shuiwen Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Biological visual systems can perceive depth from monocular vision flow, continuously integrating temporal visual cues while maintaining a balance between stability and plasticity in dynamic environments. In contrast, artificial perception models deployed on resource-constrained edge devices are typically trained in a static offline manner and remain frozen after deployment, often suffering severe performance degradation under domain shifts. While large-scale models may encode broad knowledge through massive parameter redundancy, lightweight networks face a static optimization dilemma: forcing compact models to learn universal geometric representations is computationally inefficient and often leads to performance saturation. To resolve this issue, an Online Active Learning (OAL) mechanism is introduced to endow compact neural networks with the capability to adapt continuously during operation. A closed-loop Predict-Evaluate-Correct learning paradigm is established to actively select high-confidence, information-rich signals from streaming visual input. Crucially, Elastic Weight Consolidation (EWC) is employed not merely to prevent catastrophic forgetting, but to enforce Selective Plasticity, preserving parameters that encode globally relevant structural knowledge while allowing local alignment to newly observed environments. Built upon a MobileNetV3-Small backbone, the proposed system achieves approximately a 75% reduction in computational cost while maintaining competitive depth estimation accuracy. Experimental results demonstrate that adaptability is not solely determined by model size, but rather by how effectively parameter plasticity is regulated in dynamic environments.
[CV-19] Enhancing Low Back Pain Assessment with Diffusion Models for Lumbar Spine MRI Segmentation
链接: https://arxiv.org/abs/2608.04906
作者: Maria Monzon,Thomas Iff,Ender Konukoglu,Catherine R. Jutzeler
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Maria Monzon and Thomas Iff contributed equally to this work. Published in Proceedings of The 8th International Conference on Medical Imaging with Deep Learning (MIDL 2025), PMLR volume 301, pages 1145-1163, 2026
Abstract:This study introduces a diffusion-based framework for robust and accurate semantic segmentation of lumbar spine MRI scans from patients with low back pain (LBP), regardless of whether the scans are T1- or T2-weighted. We compared with advanced models for segmenting vertebrae, intervertebral discs (IVDs), and spinal canal using the SPIDER dataset. The results showed that SpineSegDiff achieved a segmentation performance comparable to that of the state-of-the-art non-diffusion nnUnet, particularly in improving the identification of degenerated IVDs. In addition, the uncertainty maps generated by our model provide valuable insights for clinical review, enhancing the robustness and reliability of the segmentation results. The potential of diffusion models to enhance the diagnosis and management of LBP through more precise analysis of pathological spine MRI is underscored by our findings.
[CV-20] Visual Representation Matters: Exploiting Temporal Differences in Video-to-Audio Generation
链接: https://arxiv.org/abs/2608.04902
作者: Zehua Chen,Junyou Wang,Yuxuan Jiang,Zhenying Fang,Yusheng Dai,Jianfei Chen,Ziwei Liu,Jun Zhu
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Sound (cs.SD)
备注:
Abstract:Video-to-audio (V2A) generation extends image-to-audio generation (I2A) by introducing consecutive frames that provide essential temporal cues for audio synthesis. However, existing conditional diffusion-based V2A methods typically enhance visual conditioning with additional audio-visual supervision, acoustic structure prediction, or reasoning from large multimodal models, requiring extra networks or strong inductive biases. Inspired by recent advances in visual representation learning, we introduce TD-V2A, which leverages temporal differences (TD) as the key representation that distinguishes V2A from I2A, enriching visual conditioning with minimal architectural modification. We first investigate TD at both the frame and feature levels to identify the most effective representation level at which TD complements visual representations. Based on these findings, we develop a hierarchically continual learning strategy and an annealed temporal differences guidance method to progressively learn and exploit TD information during diffusion training and sampling process, respectively. Extensive experiments on benchmark datasets demonstrate that effectively exploiting TD through our proposed framework significantly improves end-to-end V2A generation quality, even outperforming dedicated V2A representations such as contrastive audio-visual pretraining.
[CV-21] When Shared Rollouts Fail in Defensive Driving Evaluation: A NAVSIM Score Basis Audit
链接: https://arxiv.org/abs/2608.04896
作者: Ziang Wei,Minjun Yu,Zheyuan Lai,Mingjie Pang,Wei Li
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: 17 pages, 1 figure
Abstract:Defensive driving scores are useful only when they preserve distinctions between policies that observe surrounding actors and those that do not. Re-simulation benchmarks may use reference-conditioned forgiveness, under which an agent receives credit when the logged human reference fails a compliance channel. When agent and reference share an unstable rollout transformation, this rule can propagate shared reference failures into broad compliance credit. We audit this risk in NAVSIM v2.2 original scene single-stage scoring. Under the affected documented-stack condition on the audited numerical backend, the route-blind Ignore-All probe and a route-aware actor-blind probe outrank human replay and PDM-Closed over the complete 12,146-token navtest split. A fresh installation following the public specification reproduces rollout divergence on a fixed 32-token diagnostic set. A same-source dependency stack control and an exact-input diagnostic isolate dependency-sensitive numerical behavior in the shared velocity refit. On a 450-token control pool, replacing only the solver eliminates rollout divergence and restores blind-last ordering while keeping forgiveness enabled. Thus, the numerical instability is the direct trigger. Reference-conditioned forgiveness propagates the resulting shared reference failures into compliance credit. We contribute an audit protocol requiring score basis and stack disclosure, blind probes, overwrite reporting, and rollout stability tests before using such scores for defensive driving claims. Comments: 17 pages, 1 figure Subjects: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2608.04896 [cs.AI] (or arXiv:2608.04896v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.04896 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-22] STEP-OPD: Rethinking Output Targets and Internal Dynamics in On-Policy Distillation for Diffusion Models
链接: https://arxiv.org/abs/2608.04887
作者: Qingyan Wei,Guangzhao Li,Xiaobing Tu,Yinggui Wang,Xiantao Zhang,Jinkui Ren,Xiaohong Liu,Linfeng Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages, 5 figures
Abstract:On-policy distillation (OPD) has become an effective approach for consolidating multiple task-specialized image generation models into a single student. However, existing OPD methods optimize the student mainly to match the teacher’s output velocity, making the teacher the upper limit of the optimization objective. While output-level supervision alone leaves the student’s blockwise representation evolution underconstrained, which weakens the transfer of capabilities that must be progressively developed across layers. We propose STEP-OPD, an on-policy distillation framework for image generation that extends the student’s learning target beyond the teacher and introduces explicit constraints on its internal representation evolution. Instead of treating the teacher as the final target, we use the velocity difference between each task-specific teacher and the shared base model as a direction for further learning and add a scaled version of this difference to the teacher velocity. In addition, we align the direction and magnitude of representation changes between the student and teacher, enabling the student to learn how representations are progressively transformed across network blocks. Experiments on compositional alignment, text rendering, and human preference show that our method consistently improves Standard OPD methods. In particular, it increases the GenEval score of DiffusionOPD from 0.927 to 0.961, while also improving OCR and all preference-based metrics. The resulting unified student surpasses the corresponding single-task teachers across all three capability groups, showing that output extrapolation enables beyond-teacher learning. And representation change alignment provides complementary guidance for the student’s internal transformations.
[CV-23] raining Crossroads for Recurrent Vision Transformers: Recurrence Neural ODEs and Deep Supervision
链接: https://arxiv.org/abs/2608.04879
作者: Grzegorz Gruszczynski,Pawel Olszowiec,Michal Byra,Grzegorz Stefanski,Alberto Presta
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision Transformers (ViTs) achieve strong image-recognition performance, but their parameter count grows linearly with depth when each block is independently parameterized. Single-block recurrent ViTs (bViT) remove this growth by repeatedly applying one shared block. Rather than proposing a new architecture, we fix a bViT and provide a controlled empirical characterization of three training and inference regimes under a common CIFAR-100 protocol, asking: (i)~when does recurrence beat independently parameterized depth—at matched FLOPs or at matched parameter memory? (ii)~when a residual recurrent block is trained through an ODE solver, does solver order act as numerical refinement or as an architectural bias? and (iii)~what does robustness beyond the training horizon cost in nominal accuracy? We find that standard ViTs remain preferable when FLOPs are the primary constraint, whereas recurrent ViTs offer a better accuracy–parameter trade-off under memory constraints. Consistent with the standard view of residual networks as Euler discretizations of ODEs, the continuous-time analogue of a residual recurrent block is the state-subtracted vector field \dotz=F_\theta(z)-z ; although known in principle, this distinction is easy to violate when the block is wrapped as a black-box vector field, and we qualify the cost at few accuracy points. Because the vector field is learned jointly with the solver, higher-order solvers act as a solver-induced architectural bias rather than a numerical-accuracy improvement, and their gains are not uniform. Finally, stage-wise deep supervision traces an accuracy–robustness frontier: it does not improve nominal accuracy, but degrades gracefully far beyond the training horizon, where naive recurrence collapses to near-random performance.
[CV-24] Persistent Object Narratives for Token-Efficient Video Language Models
链接: https://arxiv.org/abs/2608.04866
作者: Junzhe Chen,Siyuan Meng,Xiaojie Guo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video large language models (Video-LLMs) have made strong progress in open-ended video understanding. However, their visual interfaces remain token-intensive and provide limited explicit structure for linking recurring object evidence across time. We introduce SlotNarrative, a slot-based interface that organizes a video into persistent object narratives represented by compact object-state tokens. Rather than compressing frame-wise features before establishing temporal correspondence, SlotNarrative first groups visual features into object-like slots and then associates recurring observations with clip-level object entries through a lightweight, parameter-free memory that integrates multiple complementary matching cues. Each retained entry is serialized into two token types: an identity token that summarizes persistent object appearance and a set of state tokens that encode segment-level appearance, geometry, visibility, and trajectory information. This design yields an interface of only 144 allocated visual-token positions for a frozen Video-LLM, independent of the number of sampled frames. Across multiple datasets, SlotNarrative achieves a favorable trade-off between accuracy and visual-token count compared with prior compact Video-LLM interfaces. Experimental results establish persistent object narratives as a compact, structured, and temporally organized visual interface for Video-LLMs. Our code will be made publicly available.
[CV-25] Cooking beyond Frames: A Stereo Event Camera Dataset in the Kitchen ECCV2026
链接: https://arxiv.org/abs/2608.04865
作者: Chengming Feng,Hesam Araghi,Liming Zheng,Julien Dupeyroux,Xucong Zhang,Jan van Gemert,Nergis Tömen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at ECCV 2026
Abstract:Event cameras, also known as neuromorphic cameras, have gained significant attention in recent years due to their high temporal resolution, high dynamic range, and low power consumption. While many studies and datasets in neuromorphic vision have focused on automotive and drone applications, human-centric daily-life scenarios remain largely underrepresented, despite their importance for developing and benchmarking event-based perception systems. Moreover, the few existing event-based human activity datasets are typically recorded with scripted human actions, limiting their ability to capture natural human behaviors. In this paper, we introduce EventKitchen, a large-scale stereo event camera benchmark dataset of human cooking activities in the kitchen. EventKitchen is egocentrically collected from 10 participants in 13 diverse kitchens, where the participants wear a helmet with multiple sensors and naturally perform cooking activities, without any scripted actions. EventKitchen comprises 5.5 hours of stereo event recordings with synchronized RGB, depth, and IMU data. We provide human annotations for 10,762 action segments and 13,482 bounding boxes. We train baseline models on EventKitchen to perform multiple event-based tasks, including action recognition, object detection, and stereo depth estimation. By capturing natural, real-world human activities, EventKitchen establishes a challenging benchmark for neuromorphic vision beyond autonomous driving.
[CV-26] owards a satellite image manipulation and deepfake localization benchmark dataset
链接: https://arxiv.org/abs/2608.04840
作者: Jacob Arndt,Debvrat Varshney,Philipe Dias,Nivedita Nukavarapu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at IEEE IGARSS 2026
Abstract:Verifying the authenticity of satellite imagery has become increasingly critical given advances in generative artificial intelligence. Highly realistic synthetic imagery produced for malicious purposes (deepfakes) can have major consequences in the remote sensing domain, where this data is a fundamental source of information for science applications, planning, logistics, and monitoring. The remote sensing community lacks high-quality, fine-grained manipulation datasets suitable for training and evaluating detection and image forensics algorithms. Existing datasets are lacking and those that do exist either provide no ground truth masks for evaluating manipulation localization, or consist of entire images generated by GANs or diffusion models, which are inadequate for measuring localization performance. To address this gap, we describe a preliminary dataset construction process and prototype benchmark dataset for satellite image manipulation detection and localization. The dataset contains 60 images total, with 30 images carefully manipulated using three manipulation types including copy-paste splicing and diffusion model inpainting, and 30 authentic images. Each image is accompanied by a ground-truth mask and acquisition metadata, enabling both pixel-level localization metrics, image metadata studies, and analyses of how manipulation detection performance relates to image collection parameters. We describe the dataset construction process and present this initial release to support further research in image forensics and geospatial deepfake detection. The prototype dataset can be downloaded at this https URL.
[CV-27] RegisterBridgeMM: A Register-Centric Framework for RGB-Infrared Object Detection
链接: https://arxiv.org/abs/2608.04833
作者: Zian Wang,Hangchuan Liang,Yuehua Chen,Changchun Li,Chaoyi Guo,Mingzhe Liu,Fangming Gu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:RGB-infrared (RGB-IR) object detection benefits from complementary visible and thermal cues, but effective fusion remains challenging under illumination changes, weather variation, and cluttered scenes. Existing RGB-IR fusion methods often trade expressive patch-level interaction for lighter but more constrained adaptation mechanisms. We empirically observe that pretrained register tokens contain both modality-shared and modality-specific information on paired RGB-IR inputs, suggesting that they can serve as a compact substrate for cross-modal communication. Building on this observation, we propose RegisterBridgeMM, a register-mediated fusion framework organized as a three-stage register lifecycle. Aggregate preserves per-modality register summarization inherited from pretraining; Bridge performs bidirectional register-to-patch reading with consensus-residual regulation; and Project translates the resulting register summary into spatially adaptive calibration of patch features. This register pathway avoids dense patch-to-patch cross-modal interaction while preserving the pretrained patch representation. With both backbone streams frozen, RegisterBridgeMM achieves the highest mAP50-95 among the evaluated methods on all four benchmarks: LLVIP, M3FD, DroneVehicle, and FLIR-Aligned.
[CV-28] Global Attention-Fused Image Cropping with Attention-Guided and Global-Aligned Crop Evaluator
链接: https://arxiv.org/abs/2608.04821
作者: Haotian Yang,Zhile Yang,Kin-Man Lam,Patrick Le Callet,Xin Sun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: The source code is available at this https URL
Abstract:Image cropping aims to improve image aesthetics by preserving important content within an appropriately composed region. However, most existing methods focus primarily on salient regions and therefore have limited sensitivity to the global relationships among the main image components. To address this limitation, we propose Global Attention-Fused Image Cropping (GAFIC), which consists of an Attention-Guided Feature Fusion (AGFF) and a Global-Aligned Crop Evaluator (GACE). AGFF aggregates the importance of local regions to construct a global representation that captures both image structure and local details. GACE aligns candidate crop features with this global representation, enabling crop evaluation to remain sensitive to boundary changes. We further combine three ranking losses across multiple scales to obtain accurate and stable crop scores. Extensive experiments on the GAIC and CPC datasets demonstrate that GAFIC outperforms existing image-cropping methods, particularly in terms of accuracy and stability. Unlike pixel-level retargeting methods such as seam carving, inpainting, and diffusion-based synthesis, GAFIC does not synthesize or modify the retained pixels; instead, it selects an aesthetically preferred crop from the source image, making it suitable for scenarios where pixel integrity and efficient batch processing are important. The source code is available at this https URL.
[CV-29] When Diffusion Models Forget Who You Are: Identity Preservation in Face Inpainting under Large Occlusions
链接: https://arxiv.org/abs/2608.04820
作者: Feng Ding,Shuhuai Xie,Yue Zhou,Yulan Zhang,Guopu Zhu,Mengyao Xiao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Face inpainting with diffusion models has recently achieved impressive visual quality, yet preserving identity fidelity under significant occlusion and conflicting text guidance remains a major challenge. To address this issue, we present Reference Semantic Inpainting for Face (ReSem-Face), a cascaded diffusion framework that introduces an explicit identity-conditioned semantic prior for multi-reference face inpainting. Our approach distills representative identity features from multiple references to reconstruct missing semantic regions, which then guide the diffusion process through a multi-stream conditioning architecture. This design provides strong semantic constraints when pixels are absent and stabilizes identity reconstruction while remaining compatible with prompt-driven edits. Experiments on CelebAHQ-IDI-5 and VGGFace2 demonstrate that ReSem-Face yields more reliable identity-preserving completion under severe semantic masks and improves text-controlled editing quality compared with representative baselines.
[CV-30] Rethinking Pixel Mean Flows via Interval Denoiser
链接: https://arxiv.org/abs/2608.04818
作者: Alexander Zaytsev,Dmitry Baranchuk,Alexander Korotin,Aibek Alanov
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Modern diffusion and flow-based models are increasingly moving toward few-step, latent-free generation to bypass the computational overhead of multi-step sampling and the reconstruction bottlenecks of external autoencoders. We propose the Interval Denoiser, a theoretically rigorous framework for latent-free generation. Derived directly from the flow matching ODE, it establishes an exact analytical mapping for intermediate trajectory states. Unlike prior formulations, our prediction is shown to reside on a low-dimensional manifold across any time interval, making the regression tractable for a network operating directly on pixels. Furthermore, by avoiding empirical algebraic substitutions, our formulation correctly isolates the pure time derivative to prevent biased gradient evaluations and ensure exact first-order optimization. By analyzing this objective, we equip our framework with residual clipping and a time-sampling curriculum, enabling effective long-interval training and improving few-step performance. Trained from scratch on ImageNet 256x256, our model achieves an FID of 4.55 in one step (1-NFE) and 3.98 in two steps (2-NFE) without perceptual losses.
[CV-31] StaticSegFormer: An Efficient High-Performance Semantic Segmentation Based on Static Structured Pruning
链接: https://arxiv.org/abs/2608.04811
作者: Timo Bartels,Danish Nazir,Jan Piewek,Thorsten Bagdonat,Tim Fingscheidt
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Structured pruning enhances the efficiency of deep neural networks (DNNs) by eliminating groups of parameters during inference. Previous methods mostly reduce computational complexity (FLOPs), while semantic segmentation performance (mIoU) slightly drops. Accordingly, recent dynamic structured pruning methods aim at reducing the performance drop, while lowering the FLOPs even more. However, on the ADE20K and Cityscapes benchmarks, our study reveals that on a GPU platform such dynamic methods exhibit a surprisingly low frame rate far below a simple static approach, while having comparable results in mIoU and FLOPs. To address this issue, we propose a static structured pruning method for attention layers, that achieves both, a lower FLOPs and a high frame rate [fps] of the SegFormer network, the latter increased by up to 34% relative on the Cityscapes dataset, while having no mIoU performance drop at all. Our so-called StaticSegFormer method is strongest for small encoders and large images.
[CV-32] Segmentation Pre-training for Label-Efficient Lumbar Spine Degeneration Grading MICCAI
链接: https://arxiv.org/abs/2608.04810
作者: Monzon Maria,Zisserman Andrew,Jutzeler Catherine R.,Jamaludin Amir
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: The 2nd MICCAI Workshop on Efficient Medical AI (2026)
Abstract:Automated assessment of degenerative pathology in the lumbar spine on magnetic resonance imaging (MRI) requires access to large-scale datasets of expert-annotated radiological gradings. In contrast, segmentation pseudo-labels can be generated by automated tools at negligible radiologist cost. We examine whether pre-training on segmentation can effectively replace a fraction of the manual grading annotations required for downstream supervision. We pre-train a 3D ResNet encoder to segment the vertebrae, intervertebral discs (IVDs), and the spinal canal, then fine-tune lightweight task-specific grading heads using different proportions of the available training data, ranging from 10% to 100% . On a multicentre dataset of \sim2,000 subjects across 11 pathologies, segmentation pre-training, achieving a Dice score of 0.94 against pseudo-labels, improved the task-averaged (macro) one-vs-rest ROC-AUC at all proportions. With only 20% of grading labels after pre-training, the method achieved near full-supervision performance, with the largest gains observed for either low-prevalence or spatially grounded pathologies.
[CV-33] On the Effectiveness of Adaptation Strategies for VLM-Based Federated Learning in Remote Sensing
链接: https://arxiv.org/abs/2608.04791
作者: Simon Lösche,Barış Büyüktaş,Mathis Adler,Angelos Zavras,Ioannis Papoutsis,Begüm Demir
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the SPIE Artificial Intelligence and Image and Signal Processing for Remote Sensing, Edinburgh, Scotland, 2026
Abstract:Federated learning (FL) enables collaborative training of deep learning models across decentralized image archives without requiring data centralization. This paradigm is particularly relevant in remote sensing (RS), where legal regulations, privacy concerns, and bandwidth constraints restrict data sharing. However, the presence of training data heterogeneity across clients (known as non-IID data) can impede convergence and limit the generalization capability of the aggregated global model. To mitigate the adverse effects of training data heterogeneity, vision-language models (VLMs) can be leveraged in FL due to their transferable representations, which have demonstrated robustness under distribution shifts. However, their large parameter size may substantially increase communication overhead and local computational complexity in federated settings. Therefore, it is crucial to select an appropriate VLM adaptation strategy that balances the generalization ability with the communication and computational constraints. To address this issue, in this paper, we present the first comparative study of VLM adaptation strategies for FL in the context of RS image classification. We investigate full fine-tuning, encoder-specific fine-tuning, prompt learning, and low-rank adaptation (LoRA) tuning, and analyze them with respect to three criteria: 1) generalization capability under non-IID data, 2) communication overhead, and 3) local computational complexity. Experiments on BigEarthNet-S2, EuroSAT, RESISC45, and ImageNet reveal distinct trade-offs between task specialization, cross-domain generalization, and efficiency. Based on our findings, we derive a guideline for the selection of an appropriate VLM adaptation strategy in FL for RS image classification under different operational constraints. The code of this work is publicly available at this https URL.
[CV-34] Embedding Large Language Models into Flow Controls: An Agent ic Framework for Adaptive and Trustworthy Automated Cooking
链接: https://arxiv.org/abs/2608.04768
作者: Zihan Song,Hongwei Huang,Yueshuo Sun,Yonglin Tian,Fei-Yue Wang,Bai Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Automated cooking robots have traditionally relied on predefined procedures and rule-based control, ensuring stable execution but offering limited personalization, whereas recent large-model approaches support natural language interaction but often suffer from opaque decision making and unreliable execution in real kitchens. To address this challenge, this paper proposes an agentic framework that systematically decomposes personalized cooking requirements into structured and verifiable control programs rather than directly mapping language to actions. Multiple AI agents collaboratively transform user intents into canonical recipes, workflow programs with explicit flow control, and executable Python code grounded in an atomic action library. The system consists of three tightly coupled stages: offline recipe-to-code generation through multiple agents, online closed-loop execution with supervisory intervention enabled by multimodal perception, and post-run adaptation that updates user preference models for long-term personalization. Real-world experiments on a physical cooking platform demonstrate that the proposed framework achieves reliable task completion, transparent execution logic, and effective anomaly handling across diverse personalized scenarios, validating its practicality for trustworthy automated cooking in real environments.
[CV-35] FUSEP: A Multi-Center Benchmark for Diverse Tasks in Early Pregnancy Fetal Ultrasound Screening
链接: https://arxiv.org/abs/2608.04766
作者: Bin Pu,Jiewen Yang,Liwen Wang,Ying Tan,Guannan He,Xingbo Dong,Qika Lin,Jiarong Guo,Lixian Yang,Zuozhu Liu,Shengli Li,Kenli Li
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:A large number of infants with congenital anomalies are born each year globally, especially in areas with underdeveloped medical resources. Currently, fetal ultrasound screening is the most common modality for early pregnancy anatomy detection. This modality can detect anomalies earlier and provide opportune treatment advice. However, the lack of an ultrasound dataset on early fetal gestation has slowed down the development of automated assisted diagnosis. In this work, we present a benchmark dataset for Fetal Ultrasound Screening in Early Pregnancy to facilitate intelligent ultrasound examination and assisted diagnosis called FUSEP. Our dataset consists of two ultrasound views recommended by the international guideline, i.e., Crown-rump Length (CRL) and Nuchal Translucency (NT) views in three hospitals, totaling 4,017 ultrasound images, with 45,820 box-level expert-level annotations. Our dataset and baseline present the following three contributions: 1) Our medical experts annotated a total of 14 key anatomical structures in two views using a box-level format; 2) Our data is collected extensively from different sonographers, devices, scanning angles, hospitals, etc; 3) We report the performance of the semi-supervised learning, fully supervised learning, unsupervised domain adaptation (UDA), and source-free UDA in ultrasound images multi-object detection. To the best of our knowledge, this is the first publicly available dataset and benchmark for fetal early pregnancy ultrasound screening. We believe that FUSEP and benchmark can contribute to the medical community in the development of multiple tasks such as standard plane recognition, quality control on ultrasound images, automated assisted diagnostics in early fetal pregnancy, medical multi-object detection, domain adaptation for object detection, etc.
[CV-36] Explicit Language Memory for Long-Horizon Planning in Vision-Language-Action Models
链接: https://arxiv.org/abs/2608.04765
作者: Houze Xu,Jizhong Li,Ziyi Ye
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages, 4 figures
Abstract:Vision-language-action (VLA) models provide a unified paradigm for connecting visual perception, language understanding, and robotic control. However, existing VLA models still face major challenges in long-horizon tasks: sparse expert demonstrations constrain cross-task compositional generalization; the non-Markovian nature of long-horizon tasks makes it difficult for policies conditioned only on current observations to maintain temporal consistency; limited closed-loop error correction allows execution errors to accumulate; and end-to-end action fine-tuning may weaken the high-level semantic representations of vision-language model (VLM) backbones. To address these issues, we propose a hierarchical long-horizon VLA architecture with an explicit language-memory module. The central idea is to convert discrete temporal observations into a coherent textual memory sequence with temporal logic. The system is decoupled into a high-level VLM and a low-level VLA: the high-level VLM performs semantic reasoning through a visual question answering training paradigm, while the low-level VLA executes precise continuous control conditioned on subtask instructions and visual observations. The high-level VLM recursively updates both language memory and subtask instructions using the previous memory as a contextual anchor, enabling persistent temporal tracking and dynamic correction during long-horizon execution. We evaluate the proposed method in multiple simulation environments and conduct sim-to-real experiments on a real robotic platform. The results demonstrate that explicit language memory improves the success rate and robustness of VLA models on complex long-horizon tasks while providing an interpretable semantic account of the decision process.
[CV-37] Splat-Based Metal Artifact Reduction in Cone-Beam CT via Compact Attenuation Modeling
链接: https://arxiv.org/abs/2608.04764
作者: Kiseok Choi,Jaemin Cho,Inchul Kim,Min H. Kim
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:X-ray computed tomography (CT) suffers from severe metal artifacts when high-attenuation objects such as dental fillings or orthopedic implants are present. These artifacts originate from the polychromatic nature of X-rays, where attenuation varies strongly with photon energy and material composition, breaking the monochromatic assumption used by conventional reconstruction algorithms. Recent neural rendering approaches attempt to address this mismatch through differentiable polychromatic projection models, but they still struggle with smoothness bias, loss of fine structures, and prohibitive computation when extended to large-scale cone-beam CT. We introduce a splat-based metal artifact reduction framework that incorporates a physically grounded polychromatic forward model into a continuous Gaussian representation for cone-beam CT. Each Gaussian encodes the energy-dependent attenuation of the underlying material using a compact material parameterization, which enables efficient joint optimization of geometric and material properties without relying on a metal mask. This compact attenuation formulation captures the essential variation across biological tissues and metallic implants, allowing our model to explain metal-induced nonlinearity while preserving high-frequency structure. Experiments on simulated and real cone-beam CT scans show that our method converges significantly faster and suppresses metal artifacts more effectively than existing reconstruction and neural field-based approaches.
[CV-38] Revisiting Pose Sensitivity in Splat-based Computed Tomography under Sparse-view Reconstruction
链接: https://arxiv.org/abs/2608.04752
作者: Kiseok Choi,Hyeongjun Cho,Inchul Kim,Min H. Kim
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:X-ray computed tomography (CT) reconstructs volumetric representations of objects from projection images obtained by transmitting X-rays through a target. Recent splat-based tomography, which represents a volume as a continuous distribution of 3D Gaussians, has demonstrated both high reconstruction quality and fast convergence in cone-beam sparse-view CT. However, when deployed in real CT systems with limited and non-uniform view distributions, we observe distinctive streak and strip artifacts that are far more pronounced than in conventional reconstruction methods. Through detailed analysis, we show that these artifacts primarily originate from pose inaccuracies in the acquisition geometry rather than from view sparsity itself. We revisit pose sensitivity in the splatting formulation and derive a stable gradient-based framework that jointly refines geometric parameters during reconstruction. Our study not only identifies how pose perturbations propagate through the differentiable projection operator but also reveals why splat-based CT is particularly vulnerable to geometric misalignment. The resulting formulation remains lightweight and easily integrable into existing pipelines while substantially improving reconstruction fidelity under real-world sparse-view conditions.
[CV-39] Dense Metric Depth Completion from Sparse Direct Time-of-Flight Sensors
链接: https://arxiv.org/abs/2608.04737
作者: Hakyeong Kim,Ruicheng Wang,Chengtang Yao,Jiaolong Yang,Min H. Kim
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:Direct Time-of-Flight (dToF) sensors provide highly accurate metric depth and are more robust than indirect ToF systems in challenging real-world conditions. However, their high manufacturing cost and limited photodiode array size produce depth maps that are extremely sparse, low-resolution, and noisy, making them unsuitable for VR/XR, robotics, and 3D perception tasks that require dense metric depth. Existing monocular and depth completion methods struggle to handle the unique sampling patterns and hardware artifacts of dToF devices, and their performance often deteriorates significantly under severe sparsity or noise. We present a generalizable framework for dense metric depth completion from sparse dToF measurements, capable of operating across diverse sensor types, sparsity levels, and noise conditions. Our model employs a depth-guided dual-branch Vision Transformer encoder that processes RGB images and sparse dToF measurements separately, while a masked joint attention module allows depth tokens to reliably guide image features without being overwritten by them. A lightweight decoder reconstructs dense metric depth efficiently, without diffusion-based or refinement-heavy post-processing. To address the scarcity of paired training data, we introduce a comprehensive dToF simulation pipeline that reproduces the characteristics of flash, sub-VGA flash, and rotating sensors, including hardware-induced degradation, irregular sparsity, and realistic noise distributions. Trained entirely on synthetic data, our model achieves strong zero-shot generalization across 6 datasets and 3 real dToF devices, outperforming state-of-the-art approaches in both accuracy and computational efficiency. This establishes a robust and practical solution for dense metric depth completion from sparse direct ToF sensors. Our code and models are open-sourced. See this https URL.
[CV-40] When Prompts Become Pixels: Prompt-Region Grounding for Multimodal Reasoning
链接: https://arxiv.org/abs/2608.04726
作者: Yongxin Wang,Ruizhe Zhou,Yueling Tang,Yingying Zhu,Xuemin Zhao,Xiaojun Chang,Xiaodan Liang
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal large language models increasingly reason over screenshots and documents where the task itself may be written in pixels. Yet benchmarks usually place questions in text, leaving it unclear whether models use the same instruction equally well across channels. We introduce Visualized Task Semantics (VTS), a controlled intervention that moves the question into the image while keeping the source problem and answer fixed. Across six MLLMs and four benchmarks, accuracy drops in all 24 model-task pairs, by 17.8 points on average. Models often transcribe the visual question correctly yet fail to use it, exposing a semantic channel gap beyond OCR. To reduce this gap, we present prompt-region grounding, whose core design aligns the question region with typed semantics and recovers its clean representation from a masked view. At matched training cost, our method raises four-benchmark VTS accuracy from 58.0 to 66.3 while preserving accuracy on the original interface, and requires no OCR or region metadata at inference. Reading task-bearing text and grounding it as an instruction for reasoning are distinct capabilities.
[CV-41] A GitOps-Driven Annotation Catalog for Fully Automatic Railway Operations
链接: https://arxiv.org/abs/2608.04724
作者: Martin Köppel,Tobias Cronauer,Zekiye Ilknur-Öz,Sebastian Dubiel,Patrick Naumann,Philipp Neumaier
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Automatic train operation (ATO) at grade of automation 3 and above (GoA3-GoA4) requires robust AI-based perception systems capable of reliably detecting obstacles and railway-specific objects under real-world conditions. The effectiveness of these modern artificial intelligence approaches depends heavily on large-scale, high-quality, and highly dynamic annotated datasets. However, managing metadata, maintaining provenance, and tracking the iterative evolution of these annotations impose significant infrastructural and regulatory requirements. Existing monolithic data catalogs often suffer from massive operational overhead, poor integration into developer workflows, and severe documentation drift. This paper introduces an innovative, lightweight GitOps-based architecture for metadata management. By leveraging Data-as-Code principles, Continuous Integration/Continuous Deployment (CI/CD) pipelines, and Static Site Generation (SSG), the proposed approach establishes a seamless, developer-centric workflow. This ensures an traceability, enforces strict regulatory compliance, and automatically generates a highly performant dataset overview.
[CV-42] Multi-View Face and Gesture Animation with Dynamic Gaussians
链接: https://arxiv.org/abs/2608.04722
作者: Alireza Javanmardi,Vippin Kumar Jeetmal,Christen Millerdurai,Alain Pagani,Didier Stricker
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: Accepted at SCA 2026
Abstract:Creating photorealistic 3D human avatars with realistic upper-body motion remains challenging. Existing approaches either focus on the head and overlook hand gestures, or reconstruct the full body but fail to preserve fine-grained facial fidelity and hand pose accuracy. As a result, current methods struggle to capture the subtle dynamics of facial expressions and hand gestures that are crucial for natural human communication. While methods based on full-body parametric models enable avatar reconstruction from monocular or multi-view inputs, they often lack accurate facial animation and detailed hand articulation. To address these limitations, we propose MVFGA, a novel multi-view-consistent pipeline for generating realistic upper-body avatars. Our approach models the face and hands separately and fuses them with a parametric upper-body mesh model, enabling the capture of fine-grained facial expressions and hand poses for accurate upper-body avatar reconstruction. We then splat 3D Gaussians onto the obtained mesh, enabling high-quality rendering of dynamic avatars from novel viewpoints. Furthermore, we introduce MVFGA-MoCap, a multi-view upper-body motion capture dataset featuring controlled facial expression sequences, diverse hand gestures, and free-form communication. Experiments show that MVFGA generates visually realistic avatars with high-fidelity facial expressions and hand motions, outperforming baselines for upper-body avatar animation. Project page: this https URL
[CV-43] YOLOv14:Unified Cross-Domain Real-Time Object Detectionwith Adaptive Multi-View Representation
链接: https://arxiv.org/abs/2608.04720
作者: Jinling Jia,Jian Lu,Jone Yawl,Chenbin Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Real-time object detectors achieve remarkable accuracy under controlled conditions, yet degrade sharply on non-ideal inputs: fisheye distortion, game-rendered characters, aerial viewpoints, and 360° panoramas. We present YOLOv14, aunified detection framework addressing these challenges through four synergisticinnovations. (1) Deformable Area-Attention (D-AAttn) replaces rigid attentiongrids with learned 2D deformation fields, enabling adaptive sampling under geometric distortion. (2) Game2Real Domain Adaptation aligns rendered-game and photographic feature distributions via Adaptive Instance Normalization (AdaIN)and adversarial domain confusion, allowing game characters are detected as realhumans. (3) Multi-View Conditioning injects learned viewpoint embeddings intothe backbone with a cross-view contrastive loss that pulls same-class features fromdifferent perspectives closer. (4) An Adaptive Augmentation Policy automaticallyclassifies each input’ scene type and routes to optimal augmentations, while a DynamicScaleRouter learns per-input feature pyramid weights. Together, YOLOv14achieves 49.1 mAP on COCO val2017 at 2.91 ms (T4 GPU), and delivers substantial gains on fisheye (+4.1 mAP), panorama (+6.6 mAP), drone (+6.4 mAP), andgame-character (+26.1 mAP) benchmarks
[CV-44] A Multi-Sensor Dataset for Monitoring the Operational Environment of Rail Vehicles
链接: https://arxiv.org/abs/2608.04704
作者: Claudio Diotallevi,Rodrigo Gudiño,Zaharia Pachalieva,Philipp Neumaier,Patrick Naumann,Erik Bochinski,Volker Eiselein,Martin Köppel
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Reliable environment monitoring is essential for the safe and efficient operation of automated railway systems, covering all Grades of Automation (GoA), from partially automated (GoA2) to fully automated operation (GoA4). Artificial Intelligence (AI) plays a central role in enabling these systems to detect, classify, and react to potential hazards in real time. The development of such AI-based perception systems requires large volumes of accurately annotated data for training and validation. Within the Digitale Schiene Deutschland (DSD) program, DB InfraGO AG and understandAI GmbH have developed a comprehensive multi- sensor dataset tailored to the needs of railway environment perception. This dataset contains over 7 million high-quality annotations of both railway-specific and general perception objects, captured under varying operational scenarios. The finalized dataset can now be requested at the DB InfraGO AG and serve as a valuable resource for advancing AI-driven environment monitoring in the railway domain. Subjects: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO) Cite as: arXiv:2608.04704 [cs.CV] (or arXiv:2608.04704v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2608.04704 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Journalreference: 9th International Conference on Intelligent Traffic and Transportation, Amsterdam, Netherlands, September, 2025
[CV-45] Design Choices That Matter: A Functional ANOVA Analysis for Remote Sensing Multi-Label Classification
链接: https://arxiv.org/abs/2608.04702
作者: Maryam Gholami Shiri,Eva Tuba,Sašo Džeroski,Tome Eftimov,Ana Nikolikj
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: To appear at Discovery Science 2026
Abstract:Benchmarking deep learning (DL) models for multi-label classification (MLC) of remote sensing images (RSI) typically yields rankings that do not generalize beyond the evaluated datasets. In this work, we move beyond rankings by employing functional analysis of variance (fANOVA) to systematically quantify the contributions of individual design choices and their interactions to performance variability. We conduct two empirical analyses covering 48 and 20 DL models, respectively, spanning design choices such as network architecture, fine-tuning strategy, learning strategy, and initialization. By applying fANOVA across seven MLC RSI datasets, we construct dataset meta-representations that capture design-choice sensitivity profiles. Hierarchical clustering of these meta-representations reveals that datasets naturally group according to how they respond to design decisions, with patterns strongly linked to intrinsic dataset properties such as scale, spatial resolution, and label space complexity. Our findings show that for large-scale datasets, fine-tuning strategy and architecture are dominant factors, while in data-limited regimes, initialization becomes decisive. For intermediate regimes, the interaction between architecture and learning strategy governs performance.
[CV-46] UniWorld-View: Large-Baseline View Synthesis via Video Diffusion Models
链接: https://arxiv.org/abs/2608.04701
作者: Haiyang Zhou,Wangbo Yu,Chaoran Feng,Xunyu Zhou,Yonghong Tian,Li Yuan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project Homepage: this https URL Code: this https URL
Abstract:The abundance of casually captured monocular videos and images on social media provides a valuable source for immersive content creation, where generating novel views from such sparse observations can greatly enhance user experiences. However, producing photorealistic and geometrically consistent views with precise camera control remains challenging when input coverage is extremely limited. Reconstruction-based approaches such as NeRF and 3D Gaussian Splatting (3DGS) deteriorate severely under sparse inputs and fail to explicitly handle occlusions. Generative methods ease data requirements but still struggle with large-baseline view synthesis due to inaccurate or implicit geometric guidance. To overcome these limitations, we introduce UniWorld-View, a unified framework for controllable large-baseline novel view synthesis from monocular inputs. UniWorld-View integrates explicit 3D guidance with generative diffusion modeling to enable precise camera control and geometrically consistent view generation. The geometric guidance is obtained through an occlusion-aware point cloud rendering strategy that resolves visibility ambiguities and provides accurate priors for diffusion-based synthesis. By coupling this rendering strategy with powerful video diffusion backbones, UniWorld-View achieves high-fidelity novel view generation even under extreme camera motions and wide-baseline changes, and can further provide multi-view videos for downstream dynamic 3DGS reconstruction. Experiments on the WorldScore benchmark and zero-shot NVS benchmarks demonstrate the effectiveness of UniWorld-View in controllability, geometric consistency, and visual fidelity.
[CV-47] aching MLLM s to Say No: Generalized Referring Expression Comprehension via Refusal Calibrated GRPO
链接: https://arxiv.org/abs/2608.04698
作者: Xuzheng Yang,Jun Ling,Tao Huang,Caiyan Qin,Peng Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:We tackle the challenging yet underexplored task of Generalized Referring Expression Comprehension (GREC), which requires a model to localize the object described by a textual expression when it exists (positive sample) and to refuse output when it does not (negative sample). Although Multimodal Large Language Models (MLLMs) excel at localizing existing objects, they often fail to reject nonexistent ones due to the absence of negative samples during training, producing hallucinated bounding boxes. Existing post-training approaches such as supervised fine-tuning (SFT) and reinforcement learning (RL) enhance refusal behavior but usually degrade localization accuracy on positive samples, undermining the model’s core competence. To address this, we propose Refusal-Calibrated Group Relative Policy Optimization (RC-GRPO), a calibrated RL strategy that strengthens the refusal ability of MLLMs while preserving localization performance. It enforces “None” outputs in rollouts for valid advantage estimation on negative samples and applies a penalty to prevent over-refusal on positives, achieving a balanced trade-off between accuracy and reliability. A second-stage reasoning reinforcement further consolidates causal understanding and interpretability. Experiments on three GREC benchmarks demonstrate that RC-GRPO attains superior localization accuracy while maintaining strong refusal capability.
[CV-48] MOAT: Model-Agnostic Randomized Transformations for preventing Efficiency Degradation Attacks on ViTs
链接: https://arxiv.org/abs/2608.04680
作者: Anadi Goyal,Nandish Chattopadhyay,Chandan Karfa,Anupam Chattopadhyay,Norrathep Rattanavipanon
类目: Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV)
备注: This paper has been accepted for publication at IEEE ISVLSI 2026
Abstract:To adopt the Vision Transformers (ViTs) in resource-constrained environment, token pruning is widely used to reduce computational cost without impacting accuracy. However, adversaries have developed targeted attacks against said token pruning techniques to undermine such attempts to make ViTs efficient. In this paper, we propose MOAT, a model-agnostic pre-processing defense pipeline that applies a combination of input transformations to protect efficient ViT implementations against adversarial efficiency attacks. MOAT operates directly on the input without requiring modifications to the model architecture or token pruning mechanism. Experimental results demonstrate that, across all evaluated ViT models, MOAT limits GFLOPs degradation under adversarial attacks to within 3.4% of the original unattacked model.
[CV-49] SurgNarrator: A Generative Retrieval Framework for Surgical Video Understanding
链接: https://arxiv.org/abs/2608.04676
作者: Yuqing Feng,Jiawei Ma,Kevin Qinghong Lin,Kun Yuan,Nicolas Padoy,Daniel S. Elson,Anh Nguyen,Stamatia Giannarou,Baoru Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
Abstract:Surgical procedures unfold as structured and recurring clinical events, whose real-time understanding via intraoperative surgical videos is critical for intraoperative decision-making and support. However, existing video understanding methods force a trade-off: autoregressive video-language models support comprehensive reasoning but are not practical for time-sensitive clinical applications, whereas contrastive models offer low latency but struggle with complex scene understanding. Recently, generative retrieval has been explored for general-domain video understanding, but transferring it to surgery is not trivial because near-identical visual appearances may indicate semantically distinct events, and the terminology involved is highly surgery-specific. To this end, we propose SurgNarrator, a new generative retrieval framework tailored for surgical video understanding. We construct a well-curated surgery-centric vocabulary from surgical captions to define a clinically meaningful retrieval space. We then adapt the pre-trained Qwen3-VL-Embedding-8B to learn discriminative clinical representations with a temporally-aware contrastive objective. During inference, a hierarchical, procedure-aware retrieval strategy narrows the search space to the relevant procedure type, delivering fast and effective responses. Our method is comprehensively evaluated on twelve benchmarks in a zero-shot setting and achieves consistent performance gains over state-of-the-art baselines, while reducing output-stage latency by more than two orders of magnitude compared with the generative baseline.
[CV-50] Differential 6-DOF Pose Estimation with Provable First-Order Immunity to Camera Calibration Errors
链接: https://arxiv.org/abs/2608.04673
作者: Yueqiang Zhang,Liang Deng,Yi Zhang,Baoqiong Wang,Wenjun Chen,Shuixin Pan,Yulan Guo,Qifeng Yu
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 16 pages, 15 figures
Abstract:Accurate six-degree-of-freedom (6-DOF) motion estimation is essential for robotic manipulation, autonomous systems, and structural displacement monitoring. Conventional 3D-2D methods estimate absolute camera poses independently at each time and recover platform motion through camera-to-platform extrinsics, making them sensitive to extrinsic calibration errors, especially for micromotion. We present a differential pose estimation method that directly recovers platform motion from inter-frame image displacements and known 3D control points. By differencing perspective projection equations, using a depth-invariance approximation, and modeling motion on SE(3), the method avoids independent absolute-pose estimation and supports both monocular and multi-camera systems. We prove that translational extrinsic errors cancel exactly, while rotational errors induce a bounded perturbation determined by calibration error, motion magnitude, and observation geometry. We also derive generic observability conditions, a Cramer-Rao lower bound, and a bias-eliminated consistent estimator, and characterize the validity limits of the approximations. Extensive synthetic and real-world experiments establish a new state of the art for 6-DOF platform micromotion estimation, outperforming representative PnP and generalized-PnP methods in accuracy, calibration robustness, and computational efficiency. With five control points and 0.5-pixel image noise, the monocular solver obtains a combined pitch-yaw rotation RMSE of 10.09 arcsec, a translation RMSE of 3.70 mm, and a runtime of 0.34 ms. The binocular solver achieves a rotation RMSE of 10.58 arcsec, a translation RMSE of 3.91 mm, and a runtime of 0.27 ms. Code will be released upon publication at this https URL.
[CV-51] MobileWAM: Bridging World Action Models to Mobile Manipulation with Chain-of-Foresight
链接: https://arxiv.org/abs/2608.04657
作者: Zehua Fan,Junjie He,Wenxuan Song,Xi Wang,Wenqi Lyu,Linge Zhao,Fuhao Li,Zihan You,Yifei Yang,Kaiming Xu,Qi Jiang,Yue Jiang,Haoang Li,Cheng Chi,Bailin Li,Yan Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:World action models (WAMs) built on video generation backbones are a rising recipe for robot learning, yet remain confined to tabletop manipulation. Mobile manipulation demands simultaneous locomotion and whole-body manipulation amid scene-scale dynamics, yet is still dominated by dynamics-blind visual encoders with hand-crafted coordination. We bridge this gap with MobileWAM, a mixture-of-transformers architecture that fuses a pretrained video diffusion transformer with a lightweight action expert through layerwise joint attention, translating internet-scale motion priors into whole-body control. To reconcile the heterogeneous dynamics of moving and manipulating, each feed-forward layer of the action expert becomes a three-expert mixture of shared, locomotion, and manipulation experts, softly routed by the motion intent in the action tokens. To densify supervision, we further propose Chain-of-Foresight (CoF): intermediate representations sequentially predict a chain of future latent chunks, each step conditioned on its predecessor. CoF pairs naturally with our decoupled video–action denoising scheme. At deployment, the WAM serves as a pure current-frame encoder; foresight acts only through gradients, so at inference the foresight chain and video generation are discarded, leaving only policy-level cost. MobileWAM surpasses state-of-the-art mobile manipulation policies on ManiSkill-HAB and fine-tunes to a real ARX Lift2 mobile manipulator across diverse tasks with strong generalization. Code will be released soon.
[CV-52] CSGen: A Multi-Domain Curvilinear Structure Generation Model via Hierarchical Multimodal Diffusion ACM-MM2026
链接: https://arxiv.org/abs/2608.04655
作者: Zhe Shan,Ziming Yang,Lei Zhou,Wenwen Zhang,Cong Lin,Xia Xie
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to ACM MM 2026
Abstract:Curvilinear structure analysis is an important and fundamental task in multimedia. However, the controllable generation of images with precise curvilinear structure objects remains an open challenge. To address this, we propose CSGen, a hierarchical multimodal diffusion model that synthesizes high-fidelity images precisely aligned with multiple control conditions. The CSGen is built upon three key innovations: 1) We construct a multi-domain and multimodal dataset, including over 24K samples from 5 domains and 7 different types of annotations, to train the unified generation model. 2) We propose a novel hierarchical progressive control strategy that decouples topology clues from visual context by a phased signal injection, mitigating semantic drift while ensuring the topological integrity of sparse structures. 3) We design a sparsity-aware loss re-weighting mechanism to address the extreme sparsity of curvilinear structures, significantly enhancing the attention on thin and fragile structures during optimization. Extensive experiments demonstrate that CSGen generates images with superior structure accuracy and visual realism, significantly improving downstream segmentation performance while maintaining robustness across diverse prompts. Our results confirm CSGen as a scalable, data-centric paradigm for the analysis of complex curvilinear structures in diverse multimedia applications. Code and dataset are available at this https URL.
[CV-53] Overcoming Statistical Bias in Action-Controllable World Models
链接: https://arxiv.org/abs/2608.04653
作者: Yuhong Shi,Zhenhao Chu,Jie Wei,Jun Hao,Jianyi Liu,Jingwen Fu
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Action-conditioned world models aim to predict how visual environments evolve under an agent’s actions. Yet future frames are often highly predictable from visual inertia and recurring motion patterns alone. This creates a shortcut: models can fit the data by exploiting statistical biases without making their visible dynamics meaningfully depend on the action. As a result, different actions may produce similar futures, while motion may persist even under zero action. The key question is how to reduce reliance on statistical shortcuts from dominating action-conditioned prediction. We argue that action control requires more than injecting action features; it requires enforcing consistency under counterfactual changes to actions and observations. Based on this insight, we introduce CoCo, a Counterfactual Consistency framework to enhance action controllability through two complementary constraints. Multi-step counterfactual consistency constrains reference, inverse-action, and zero-action rollouts, while action-spatial counterfactual consistency enforces consistent predictions under mirrored scenes and transformed actions. Together, they reduce reliance on statistical shortcuts from substituting for action-dependent dynamics. We further introduce Action Response Consistency (ARC) and Drift Energy (DE) to assess action controllability, together with Mini-SSMB for same-state, multi-action counterfactual evaluation. On Mini-SSMB, our full model achieved ARC_inv of 0.412 and ARC_ref of 0.483, while reducing DE by 17.07% relative to the baseline. On VP2 visual planning, it achieves the highest average success rate among SOTA models, at 73.1%. Experiments on BAIR and RoboNet further show that these gains preserve video prediction quality and transfer across model settings.
[CV-54] DisMix: Order-Aware Mixup for Medical Imaging via Disentangling Ordinal and Non-Ordinal Features
链接: https://arxiv.org/abs/2608.04652
作者: Dileepa Pitawela,Gustavo Carneiro,Hsiang-Ting Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Image mixup is a widely adopted data augmentation strategy, yet it is ill-suited for ordinal classification tasks such as medical disease grading, where labels encode a progression of severity. By indiscriminately blending disease-severity cues (ordinal) with appearance-level variation (non-ordinal), standard mixup produces samples that distort the very ordinal structure that underpins clinical severity grading. We introduce DisMix, an order-aware mixup framework for ordinal classification. DisMix disentangles ordinal and non-ordinal features via a dual-codebook VQ-VAE, allowing each subspace to be mixed independently: ordinal codes are interpolated to produce meaningful intermediate ranks, while non-ordinal codes are varied to introduce appearance diversity without corrupting the ordinal signal. Across four medical imaging datasets, DisMix shows the best aggregate performance among six image mixup baselines paired with six ordinal classifiers and remains effective under data scarcity and clinical grading variability.
[CV-55] YOLO-PVC: 2D-to-3D Consolidation of Slice-wise Detections for Volumetric Liver Tumor Localization in MRI ECCV2026
链接: https://arxiv.org/abs/2608.04642
作者: Talha Waqas,Mounir Lahlouh,Kawther Taibouni,Mahnoor Waqas,Salar Ahmed,Sébastien Mulé,Yasmina Leroul-Chenoune
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 2 figures, 4 tables. Accepted at AI4M3D Workshop, ECCV 2026 (Spotlight)
Abstract:Slice-wise 2D object detectors are increasingly applied to volumetric data due to their computational efficiency and scalability, yet they often yield fragmented and unstable predictions along the depth axis. We propose YOLO-PVC, a lightweight and model-agnostic framework for 2D-to-3D consolidation of slice-wise detections. The method enforces depth continuity, aggregates bounding box coordinates using robust percentile statistics, and further refines axial extent through a lightweight MLP-based calibration module. Unlike naïve stacking or averaging strategies, YOLO-PVC explicitly addresses missing detections and outlier slices along the depth dimension. Experiments on 3D liver MRI volumes across three tumor categories demonstrate consistent improvements over multiple aggregation baselines. The heuristic PVC achieves an overall \mathrmIoU_3D of 0.665 , while the calibrated variant further improves performance to 0.710 , with high planar overlap ( \mathrmBEV\ IoU \approx 0.78 ). These results demonstrate that structured geometric consolidation provides an effective and practical solution for volumetric liver tumor localization in clinical MRI.
[CV-56] Visual Anchoring in Diffusion: Multimodal Zero-Shot Skeleton Action Recognition
链接: https://arxiv.org/abs/2608.04623
作者: Zehao Bao,Shujun Guo,Bruce X.B. Yu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Zero-shot Skeleton Action Recognition (ZSAR) remains ambiguous when unseen actions share similar skeleton joint dynamics but differ in objects or scene context. RGB provides these missing cues, yet existing multimodal methods typically maintain independent skeleton and RGB scoring branches and fuse their outputs. Without using unlabeled test data for adaptation or fusion calibration, a fixed fusion weight cannot capture class-pair-dependent modality reliability, while an adaptive rule lacks target-side feedback for deciding which branch should dominate. We bypass this weight-selection problem via the classify-by-generation paradigm, where each class is scored by how accurately a text-conditioned denoiser predicts the noise added to the skeleton feature. This formulation separates the progressively corrupted skeleton from fixed conditioning, allowing RGB and text to jointly condition a single class-scoring function rather than produce independent scores. We instantiate this idea as Multimodal Triplet Diffusion for Skeleton-Text Matching (TDSM-MM), augmenting a text-conditioned denoising Transformer with a non-diffused RGB condition token that serves as a stable visual anchor during skeleton data reconstruction. Our proposed TDSM-MM has been ablated via extensive experiments and achieved the best inductive accuracy on three of four NTU-60/120 splits and surpasses the transductive state-of-the-art on NTU-120 96/24 (i.e., 71.3% vs. 69.1%), without test-time adaptation, suggesting that diffusion-based methods can be a promising direction for zero-shot learning.
[CV-57] DAC-Pose: Dual-Agent Collaborative Framework for Pose-Guided Human Generation
链接: https://arxiv.org/abs/2608.04622
作者: Haotian Yang,Zhile Yang,Huiyu Zhou,Xin Sun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Code is available at this https URL
Abstract:AI agents have emerged as a powerful new paradigm in generative image synthesis, enabling systems to perform complex semantic reasoning rather than passive pixel-level mapping. In pose-guided human generation, conventional methods inevitably produce severe visual artifacts under drastic viewpoint shifts, fundamentally because they lack the cognitive capacity to logically deduce unseen regions and model complex spatial deformations. To bridge this gap, we propose DAC-Pose, a novel agent-driven multimodal framework that reformulates single-view human generation as a collaborative dual-agent system. DAC-Pose integrates two complementary components, namely, the Prior Semantic Reasoning (PSR) agent and the Discrepancy-Aware Visual Encoding (DAVE) agent. Functioning as a cognitive engine, PSR utilizes collaborative reasoning to deduce the fine-grained attributes of unseen regions. Concurrently, acting as a specialized visual perception agent, DAVE quantifies and encodes viewpoint-induced spatial misalignments, continuously feeding robust spatial constraints back into the generative process. This autonomous feedback loop between semantic deduction and visual perception ensures high-fidelity detail synthesis. Extensive experiments on the DeepFashion and Market-1501 benchmarks validate the superiority of our agent-driven paradigm. Notably, DAC-Pose excels in preserving texture alignment and identity consistency under drastic viewpoint changes. The code is available at this https URL.
[CV-58] HiSC: Hierarchical Spatial Clustering Token Compression for Efficient 3D Scene Understanding ACM-MM2026
链接: https://arxiv.org/abs/2608.04610
作者: Jiuhe Qu,Yingping Liang,Ying Fu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ACM MM 2026
Abstract:3D vision-language models (3D VLMs) enable spatial reasoning over multi-view scenes but suffer from substantial token redundancy due to duplicated observations and large uninformative regions, leading to high computational cost. Although visual token compression has shown promise in accelerating 2D VLMs, it fails to capture the structured nature of 3D scenes and leads to incomplete spatial coverage and loss of fine-grained details. In this paper, we propose \textbfHiSC, a training-free framework for hierarchical spatial clustering token compression in 3D VLMs. HiSC lifts token compression from token-level selection to cluster-level processing by organizing tokens into spatially grounded clusters using joint geometric and semantic cues. Specifically, we first introduce a \textbfspatial graph-based merging (SGraM) strategy that models cross-view redundancy as spatial connectivity and consolidates physically consistent regions, effectively merging extremely similar redundant tokens prior to LLM inference. We then propose a \textbfspatial clustering-based pruning (SCluP) paradigm within LLM inference, which performs hierarchical compression across clusters and within clusters, preserving object instance completeness while retaining fine-grained details for important regions. Extensive experiments on diverse 3D reasoning benchmarks show validate the effectiveness of HiSC, particularly under high visual token pruning ratios. Besides, HiSC achieves over 90% token reduction with minimal performance degradation. Code is accessible at this https URL.
[CV-59] RCoRSurg: Temporal-Relational Co-Reasoning for Surgical Video Triplet Recognition
链接: https://arxiv.org/abs/2608.04606
作者: Fang Li,Shihao Zou,Weixin Si,Yang Gao,Shuai Li,Aimin Hao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: code: this https URL
Abstract:Understanding complex surgical scenes requires recognizing multiple interdependent entities, such as instruments, actions, and targets, while maintaining their relational consistency across time. Existing surgical triplet recognition methods struggle to jointly model intra-frame label dependencies and inter-frame temporal semantics in a unified manner. To address these limitations, we propose a unified framework that integrates spatial, relational, and temporal cues for robust surgical triplet recognition. Specifically, class-specific spatial priors are first extracted through a multi-scale encoder. These priors are then refined by a Label Correlation Modeling module with multi-scale class activation map-guided relational extraction (MS-CAMRE), enabling the model to capture both static co-occurrence patterns and dynamic contextual dependencies among triplet components. Furthermore, a Bidirectional Temporal-Relational Fusion Attention (BTRFA) module harmonizes temporal and relational representations to achieve coherent temporal reasoning. We also introduce a new evaluation metric, the Triplet Consistency Error Rate (TCER), which quantitatively measures the model’s ability to preserve causal and semantic consistency across triplets. Extensive experiments on the CholecT45 and ProstaTD datasets show that our method achieves state-of-the-art performance, improving AP_IVT by 5.1 percent and 7.8 percent, respectively. Moreover, according to TCER, our approach achieves relative reductions of more than 36 percent and 25 percent on the two datasets, respectively, demonstrating the effectiveness of our framework in temporal-relational co-reasoning.
[CV-60] COSMO: Consensus-Driven Shift Modulation for Source-Free Domain Adaptation
链接: https://arxiv.org/abs/2608.04604
作者: Bo Li,Junjie Peng,Xiaohua Xie,Jianhuang Lai
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 30 pages, 7 figures
Abstract:Source-free domain adaptation (SFDA) adapts a source-trained model to an unlabeled target domain without source data, a practical setting under privacy or storage constraints. Yet its self-generated supervision can reinforce source bias under substantial domain shifts. Pretrained vision-language models (VLMs) offer complementary semantic knowledge, but the relative reliability of the source model and VLM varies across target samples. Existing cross-model guidance does not explicitly account for this variation and may overwrite valid source-derived evidence under conflict, a failure we term source-derived evidence forgetting. We formulate VLM-guided SFDA as a sample-wise reliability-allocation problem and propose Consensus-Driven Shift Modulation (COSMO). COSMO replaces expert-to-expert guidance with co-adaptation through an anchored shared consensus. It first forms a sample-specific initial consensus that favors the more concentrated prediction. During adaptation, COSMO re-aggregates both branches’ evolving evidence and regulates how far the resulting consensus moves from its initial anchor based on consensus uncertainty and training progress. This keeps the shared supervision anchored yet adaptive. Across four benchmarks, COSMO achieves state-of-the-art performance under matched VLM backbones. Further analyses indicate that it better balances the retention of valid source-derived evidence with the absorption of complementary VLM evidence.
[CV-61] he First EgoCross Challenge at EgoVis 2026: Cross-Domain Egocentric Video Question Answering CVPR26
链接: https://arxiv.org/abs/2608.04589
作者: Yuqian Fu,Tianwen Qian,Yanjun Li,Yu Li,Kunyu Peng,Xu Zheng,Yongqin Xian,Alessio Tonioni,Yanwei Fu,Xiaoling Wang,Danda Paudel,Federico Tombari,Luc Van Gool,Leyi Wu,Yifan Zhao,Jinjie Zhang,Yinchuan Li,Yingcong Chen,Zixu Li,Zhiwei Chen,Zhiheng Fu,Wenbo Wang,Yupeng Hu,Weili Guan,Liqiang Nie,Takuya Murakawa,Toru Tamaki,Yi Wen,Zhenglin Du,Zhengyang Li,Lingling Li,Licheng Jiao,Wenping Ma
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 1st EgoCross challenge @ EgoVis workshop, CVPR26
Abstract:EgoCross is a cross-domain egocentric video question answering benchmark designed to evaluate whether multimodal large language models can generalize beyond common daily-life scenarios. The first EgoCross Challenge was hosted at the Third EgoVis Workshop at CVPR 2026 and evaluated models on first-person videos from four target domains: surgery, industrial assembly, extreme sports, and animal perspectives. Each test example consists of an egocentric video clip, a question, and four candidate answers, from which the model must select the correct option. This technical report introduces the challenge task, benchmark resources, and two official Codabench tracks. The Source-Limited Track restricts participants to the official baseline model and a small support set, whereas the Open-Source Track permits broader choices of models and training data under rules that prohibit the manual construction of target-domain training data. In total, the challenge received more than 1,500 submissions from over 130 participants, with 19 teams participating in the Open-Source Track and 38 teams in the Source-Limited Track. We further present the official leaderboard results and summarize the winning solutions from both tracks. We hope that this report will serve as a useful technical reference for advancing cross-domain egocentric video understanding. All resources, including the challenge data, baseline implementation, and code released by the winning teams, are made publicly available.
[CV-62] MetaVideoAgent : Automated Video-Agent Evolution for Long-Form Video Understanding
链接: https://arxiv.org/abs/2608.04587
作者: Benlei Cui,Ruize Wang,Junjie Li,Jinhao Chen,Longtao Huang,Yinghao Chen,Yuwen Zhai,Jingqun Tang,Ruijian Jia,Weiwei Wu,Pengfei Sun,Haiwen Hong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 7 figures. Code: this https URL
Abstract:Long-form video understanding requires locating sparse, question-relevant evidence in long, multimodal videos. Real-world video distributions differ in modality-specific information density, content structure, and evidence patterns, causing fixed video-agent designs to incur redundant processing or fail when mismatched. Extending automated agent evolution from text to video is challenging because full long-video execution makes candidate validation expensive, failures propagate across coupled evidence-processing stages, and complex preprocessing, perception tools, and localization strategies make code-level updates difficult to implement reliably. We introduce MetaVideoAgent, a framework that automatically evolves a video agent for a target distribution. It profiles information density and evidence requirements from sparsely sampled frames and associated queries to guide initial design, then compresses localized failures into independently executable minimal validation tasks. It constructs evidence-grounded Gold Paths, audits Student trajectories, aggregates recurring failures across samples, and attributes them to responsible modules. A modular agent representation constrains each update to the primary responsible module and its necessary dependencies. We further introduce VA-EvoBench, covering eight video distributions with separate evolution and held-out splits. With four evolution iterations per distribution, MetaVideoAgent improves every initial agent and raises macro-average accuracy from 38.44% to 51.47%, at an average evolution cost of 3.54M tokens per distribution. The evolved agents outperform the strongest prior fixed-design video agent by 6.39 percentage points while using the fewest tokens and video frames per question among the compared video agents. We will release all code and data to support reproducible research. Comments: 16 pages, 7 figures. Code: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2608.04587 [cs.CV] (or arXiv:2608.04587v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2608.04587 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-63] ACA-GS: Adaptive-Capacity Anchored Gaussian Splatting for Compact Dynamic Radiance Fields
链接: https://arxiv.org/abs/2608.04581
作者: Seunghyeon Song,Joo Chan Lee,Chanung Park,Jun Young Jeong,Minseo Lee,Eunbyung Park,Jong Hwan Ko
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages, 8 figures. Accepted to ACM Multimedia 2026
Abstract:Recent advances in 4D Gaussian Splatting (4DGS) enable high-fidelity, real-time spatiotemporal rendering, but expose a fundamental trade-off between motion expressiveness and storage efficiency. While anchor-based designs achieve compactness through anchor-level parameter sharing, their rigid uniform parametrization enforces fixed Neural Gaussian counts and feature budgets per anchor. Consequently, insufficient fidelity is addressed by excessive anchor density, rather than lightweight, targeted increases in Neural Gaussian count or feature capacity, resulting in memory waste. To overcome this rigidity, we introduce an adaptive-capacity anchor-based framework that dynamically allocates the representational capacity based on local spatiotemporal demands. Adaptive Anchor Cardinality varies the number of Neural Gaussians per anchor, concentrating primitives in regions of high geometric or motion complexity while suppressing redundancy. In parallel, Adaptive Anchor Feature Masking modulates anchor-level feature channels, assigning rich features to complex regions and lightweight representations to simpler ones. Experiments on MPEG, Panoptic Sports, and N3DV datasets demonstrate substantial storage reduction without degrading visual quality. Notably, on challenging MPEG sequences with complex motion, our method achieves up to 1.5x higher compression than state-of-the-art anchor-based methods while preserving comparable quality.
[CV-64] PhysMind: From Video to Executable Worlds for Training-Free Physical Reasoning
链接: https://arxiv.org/abs/2608.04575
作者: Chen Yang,Shenxiang Zeng,Haoyang Zhao,Zhouyuan Xu,Youquan He,Haoyu Li,Mingyi Deng,Jiansheng Fan,Chen Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 27 pages, 18 figures. Project page: this https URL
Abstract:Reliable physical reasoning from video requires understanding how objects move, interact, and respond to interventions. Existing vision-language models (VLMs) often struggle to interpret these dynamics and reason reliably about future and counterfactual outcomes. We introduce PhysMind, a training-free agentic framework that constructs one reusable, question-agnostic executable world per video. PhysMind recovers a temporally consistent dynamic scene through object segmentation, mesh reconstruction, and 6D pose tracking, then fits analytic continuous-time dynamics and latent physical parameters without unrolling a time-stepped simulator. Given a question, it inspects, continues, or edits the world and answers from the resulting trajectories and interactions. Relative to direct chain-of-thought (CoT) reasoning with the same VLM, PhysMind improves accuracy by 38.23 points on CLEVRER and 8.08 points on Physion++. On counterfactual questions, it exceeds the strongest evaluated VLM baseline, GPT-5.5, by 19.25 points.
[CV-65] alk2Sensors: 3D Visual Grounding in Autonomous Driving via Sensor-Adaptive Physical Cue Matching
链接: https://arxiv.org/abs/2608.04568
作者: Runwei Guan,Di Tian,Ningwei Ouyang,Ruixiao Zhang,Shaofeng Liang,Haocheng Zhao,Lianqing Zheng,Xiaokai Bai,Guotao Wang,Daizong Liu,Henghui Ding,Hui Xiong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 12 figures
Abstract:As a key capability for embodied intelligence, 3D visual grounding (3DVG) has been predominantly studied in indoor scenes with RGB-D or point-cloud inputs, while existing outdoor extensions largely rely on monocular images alone. Both settings fall short of real-world outdoor perception, where heterogeneous sensors capture complementary yet distinct physical properties, such as visual texture, 3D geometry, and object kinematics, that are indispensable for flexible and robust query-adaptive grounding but remain under-exploited. To bridge this gap, we introduce Talk2Sensors, the first multi-sensor 3D visual grounding dataset built upon camera, LiDAR, and 4D radar. It contains 8,682 language instructions and 20,558 referred objects, with diverse prompts explicitly aligned with sensor-specific physical cues. Furthermore, we propose TSFormer, a unified Transformer-based framework for language-guided 3D visual grounding in autonomous driving. TSFormer adopts a coarse-to-fine property-aware fusion strategy: the Language-Routed Property Sampler first performs coarse text-conditioned feature retrieval by modulating sensor sampling weights with query-level linguistic cues, while the subsequent Sparse-Preserving Modality Arbiter module conducts fine-grained modality arbitration and text-guided refinement to determine the precise referred spatial location. This design enables dynamic routing of appearance, geometry, and motion cues according to the semantic requirements of each prompt, preventing dense modalities from overwhelming sparse but critical sensor signals. Extensive experiments demonstrate that TSFormer achieves state-of-the-art performance across multiple benchmarks: it improves over the strongest baseline by 8.05 mAP on Talk2Sensors, and transfers to the monocular Mono3DRefer benchmark with 53.05% Acc@0.5.
[CV-66] OutLangSplat: 3D Language Gaussian Splatting for UAV Outdoor Scenes
链接: https://arxiv.org/abs/2608.04560
作者: Xia Yan,He Wu,Yanghui Xu,Zizhao Wu,Jiazhou Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: 9 pages, 6 figures, 7 tables
Abstract:3D Language Gaussian Splatting embeds open-vocabulary language features into 3D Gaussian Splatting, providing an efficient explicit representation for text-driven 3D scene understanding. However, existing methods are limited to indoor or small-scale scenes, and tend to fail in Unmanned Aerial Vehicle (UAV) outdoor scenes, where severe occlusions and long distance viewpoints often lead to incorrect semantic activations and missing target responses. In this paper, we present OutLangSplat which adapts language Gaussian representations to UAV outdoor scenes by improving feature representation and aggregation reliability. For the feature representation, a 2D-3D dual-branch representation with region-based alignment and fusion is designed to improve spatial consistency, reducing incomplete target responses and background misactivations. For the feature aggregation, we introduce a training-free contribution and consistency-aware Gaussian feature aggregation strategy that leverages pixel contribution reliability and cross-view semantic consistency to suppress unreliable responses from noisy viewpoints. A new dataset is provided by manually annotating various objects on four real-world public UAV outdoor scene datasets. To the best of our knowledge, it is the first accessible dataset of open-vocabulary 3D scene understanding for UAV outdoor scenes. Quantitative evaluations and ablation studies demonstrate that OutLangSplat outperforms SOTA methods on both open-vocabulary semantic segmentation and instance localization tasks. The datasets and codes will be open-sourced.
[CV-67] ColorFD: A Finite-Difference Guided Black-Box Physical Adversarial Attack for Remote Sensing Object Detection
链接: https://arxiv.org/abs/2608.04559
作者: Tiannuo Guo,Guhang Qiu,Yuzhen Xie,Rui Feng,Ligang Li,Deliang Xiang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13pages,12figures
Abstract:Although deep neural network-based remote sensing object detectors have achieved strong performance, they remain vulnerable to adversarial perturbations. Existing studies mainly focus on digital or white-box settings, whereas black-box physical attacks remain underexplored. These attacks are often constrained by limited physical feasibility and inefficient optimization in high-dimensional search spaces. To address these challenges, this paper proposes ColorFD, a black-box physical attack based on multiple pure-color patches. The patch positions and color parameters are jointly optimized using Differential Evolution (DE). A target-wise fitness and selection mechanism evaluates the attack state of each target and preserves target-specific improvements during evolution. Two guidance strategies further constrain the patch search space. Key-region localization identifies sensitive regions through finite-difference color probing. Common-feature extraction provides category-level spatial priors and avoids repeated localization. Although evaluated on aircraft, the formulation is not inherently restricted to this category. Experiments on YOLOv3u, YOLOv5u, and Faster R-CNN show that ColorFD outperforms the tested black-box patch method across all evaluated detectors and remains competitive with strong white-box baselines. Physical-world experiments further demonstrate that the optimized pure-color patches can be transferred from the digital domain to real imaging conditions.
[CV-68] VoxStruct3D: Structure-Leading Flow Matching for Voxel-Space 3D MRI Synthesis
链接: https://arxiv.org/abs/2608.04557
作者: Fang Li,Yang Gao,Shihao Zou,Weixin Si,Hongyu Wu,Qing Xia,Shuai Li,Aimin Hao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:High-fidelity 3D MRI synthesis requires both globally coherent anatomy and fine-grained voxel-level detail. Although latent diffusion makes volumetric generation tractable, its image autoencoder introduces a reconstruction bottleneck that can limit the fine detail recoverable in the final volume. We present VoxStruct3D, a voxel-space flow-matching framework that directly models full-resolution MRI volumes using a clean-data prediction objective. Its Volumetric Voxel Generator (VVG) combines factorized 3D patch embedding with overlapping upsampling, time-modulated residual refinement, and skip fusion, enabling neighboring tokens to jointly reconstruct shared voxel regions and suppress patch-boundary artifacts. To complement direct voxel-space modeling with an explicit anatomical prior, we further introduce a Structure-First, Image-Follows (SFIF) strategy. A frozen pretrained 3D medical encoder and a StructVAE extract compact structure tokens that preserve dominant anatomy, while a structure-leading schedule keeps their trajectory ahead of the image trajectory. Patch-Aligned RoPE spatially aligns the unequal token grids, and asymmetric attention enforces one-way guidance from structure to image. Experiments on pathological and healthy T1-weighted brain MRI datasets show that VoxStruct3D achieves the strongest overall performance across feature-distribution alignment, sample diversity, and perceptual quality, producing anatomically coherent and visually realistic volumes.
[CV-69] EgoAfford: Task-Oriented Affordance Grounding via Egocentric Referring Segmentation
链接: https://arxiv.org/abs/2608.04533
作者: Xinyuan Guan,Feifan Chen,Xinyu Zhan,Fu-Cheng Zhang,Cewu Lu,Lixin Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Part-level affordance grounding has advanced the localization of functional object regions associated with elemental actions. Extending this capability to complex tasks calls for connecting the semantic roles of participating objects with task-state-aligned visual observations and multi-step planning. We introduce EgoAfford, a benchmark designed to connect these three aspects. Given an egocentric observation and a high-level tabletop task, a model must generate the remaining plan and segment the functional regions of up to three components of the next action: the direct object, instrument, and destination. EgoAfford comprises approximately 15.5k human-verified images from 2,000 generated multi-step scenes, organized as semantically aligned, task-complete image series, together with EgoAfford-Real, 102 manually captured images spanning 26 tasks. We further present EgoLens, a 3B multimodal large language model with role-specific mask decoders, as an in-domain reference model for this joint task. Evaluations of recent referring-segmentation MLLMs, commercial-VLM–SAM2 pipelines, and EgoLens highlight the complementary challenges of next-step inference and action-role-conditioned part grounding. EgoLens establishes strong reference performance on both generated and manually captured observations. Together, EgoAfford and EgoLens provide a foundation for jointly studying perception and planning in multi-step tabletop tasks. Our project page is available at: this https URL
[CV-70] FocusMem: Factorizing Content Readout and Trust in Latent GUI Memory
链接: https://arxiv.org/abs/2608.04530
作者: Zhuoran Zhang,Bowen Li,Jingcheng Ju,Yang Shi,Qixun Wang,Haotian Wang,Wei Chen,Tengjiao Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 36 pages
Abstract:GUI agents must remember both useful experience from earlier tasks and unfinished progress in the current interaction. Latent memory offers a compact solution by compressing multimodal trajectories into a few continuous tokens. Existing methods, however, usually map each trajectory to one fixed memory block and train it mainly through next-action supervision. This creates three practical problems: important details may be lost during compression, the same memory block must serve different decision stages, and irrelevant retrieved trajectories may still mislead the agent. We introduce FocusMem, which separates these responsibilities within a compact latent-memory interface. A role-aware content basis encourages episodic memory to retain reusable experience and working memory to retain task progress. A state-conditioned readout generates a decision-specific view of the same stored evidence, while a lightweight trust gate can suppress memory blocks that appear irrelevant to the current step. All components are trained while the GUI policy remains frozen. Across five GUI-agent benchmarks, FocusMem consistently outperforms a fully matched action-only fixed-memory baseline and prior latent memory adaptations. Further analysis shows that semantic and functional supervision preserve complementary information, state-conditioned readout is more robust as surrounding trajectory context grows, and the trust gate reduces the harm caused by injected irrelevant episodic evidence. These results show that effective latent memory depends not only on compressing past interaction, but also on what is retained, what is exposed, and what is allowed.
[CV-71] Coupled Continuous-Discrete Generation for Scene Text Image Super-Resolution
链接: https://arxiv.org/abs/2608.04525
作者: Axi Niu,Knag Zhang,Qingsen Yan,Hao Jin,Jinqiu Sun,Yanning Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Scene text image super-resolution (STISR) aims to recover visually plausible appearance while preserving character semantics from degraded inputs. Existing STISR systems often rely on externally generated priors or separate image and text models, resulting in error propagation and costly multi-stage inference. We present DualTSR, a unified framework that formulates STISR as coupled continuous-discrete generation. Conditional flow matching restores continuous image latents, while absorbing-state discrete diffusion reconstructs text tokens. Both processes share a multimodal transformer backbone, allowing the evolving image and text states to interact throughout generation without an external OCR prior at inference. On CTR-TSR, DualTSR achieves the best FID, LPIPS, ACC, and NED among the compared methods at both X2 and X4. On an aligned RealCE subset, it obtains the best FID, ACC, and NED with competitive LPIPS. Compared with DiffTSR at X4, DualTSR improves ACC by 12.78 percentage points while reducing the parameter count from 1.23B to 203M and end-to-end latency from 13.3s to 132ms. These results establish DualTSR as an accurate and efficient method for STISR.
[CV-72] GeoReward: Mitigating Contextual Variable Overestimation in Vision-Language Models for Cross-Market Preference Prediction
链接: https://arxiv.org/abs/2608.04504
作者: Shuo Liu,Huixiang Cai,Weiru Zhang,Xiaoyi Zeng
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-language models excel in many multimodal tasks but remain prone to a subtle yet impactful failure mode: they tend to overestimate dominant visual-textual cues while underestimating sparse but decision-critical contextual variables. This issue, which we term Contextual Variable Overestimation (CVE), becomes particularly evident in real-world applications such as predicting advertisement image preferences across diverse geographic markets. For instance, when a VLM is asked to choose between two product images tailored for different countries, it often defaults to a consistent output, ignoring ground-truth regional variations. This collapse occurs because pervasive high-volume signals, such as product attributes and dense image patches, overwhelm the few but critical tokens that encode market-specific context. To address CVE, we first collect a new multimodal dataset of real advertising creatives and their click-through performance across multiple countries. We then introduce GeoReward, a reward model designed to predict ad image preferences across diverse geographic markets. GeoReward integrates three purpose-built mechanisms: (1) Market-Aware Retrieval Augmentation, (2) Context-Guided Visual Modulation, (3) Selective Sensitivity Loss. Furthermore, we demonstrate how GeoReward can guide the fine-tuning of RL for a VLM to generate background designs for text-to-image models, producing market-aware advertising creatives. Experiments validate that our framework mitigates CVE and outperforms existing baselines. This work not only diagnoses a systematic bias in VLMs toward dominant perceptual features but also delivers a targeted solution for applications where sparse contextual variables govern decision-making.
[CV-73] Privacy-Preserving Action Recognition: Taxonomy Methods and Privacy-Utility Trade-offs
链接: https://arxiv.org/abs/2608.04501
作者: Sareer Ul Amin,Muhammad Ayaz,Muhammad Munsif,Sanghyun Seo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video surveillance in public safety, healthcare, and smart environments has made continuous human monitoring routine, raising real risks to personal identity and appearance. Privacy-preserving action recognition (PPAR) tackles the tension between the utility of video understanding and this exposure, and has drawn fast-growing interest. However, existing surveys remain narrow. Most catalog a single mechanism family, predate recent adversarial and hybrid work, or barely address evaluation. The result is a fragmented literature with incompatible threat models, inconsistent metrics, and no shared evaluation standard. We address this with a PRISMA-guided review of 32 peer-reviewed papers (2018–2026) drawn from 885 screened records. Methods sort into five families, namely adversarial learning (52%), skeleton-based (20%), cryptographic (12%), differential privacy (8%), and hybrid (8%), each with distinct privacy, utility, and efficiency trade-offs. Evaluation is the weak point. Only 10% of papers adopt a formal privacy definition, 65% rely on ad-hoc metrics, and 40% report an inconsistently defined cMAP. The trade-offs are steep. Skeleton methods reach about 85% accuracy but drop appearance, adversarial methods hold near 80% utility at moderate privacy (cMAP 0.9 to 0.3–0.5), and differential privacy often falls below 70%. Harder conditions stay under-tested, with fewer than 15% of papers checking cross-dataset generalization, under 10% testing adaptive attackers, and real-time edge deployment nearly untouched. We contribute a two-dimensional privacy-space taxonomy, a formal threat model, a comparative trade-off analysis, the PPAR Unified Evaluation Protocol, and a roadmap centered on benchmark standardization. With this grounding, we argue PPAR can move from prototypes toward deployment, with lessons extending to face recognition and medical imaging.
[CV-74] DIVE: Dynamic Iterative Visual Evidence Construction for Efficient Vision-Language Models
链接: https://arxiv.org/abs/2608.04496
作者: Chen Zhong,Xiao An,Zijie Wang,Jiepan Li,Guangyi Yang,Wei He
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Visual inputs in vision-language models (VLMs) are often encoded into substantially longer token sequences than text, making visual tokens a major bottleneck for efficient inference. Abundant recent methods address this bottleneck by scoring token importance and pruning low-scoring tokens in a single pass. However, one-shot scoring is insufficient because a token’s prompt-relevant usefulness depends on the evidence already retained. Motivated by this insight, we introduce DIVE (Dynamic Iterative Visual Evidence Construction), a training-free framework that recasts visual-token pruning as dynamic evidence construction. DIVE repeatedly selects the remaining token with the highest residual-conditioned score, updates the visual and prompt residuals to discount the evidence already explained, and re-evaluates the remaining tokens. This select-update-re-evaluate process builds a retained set of complementary, prompt-relevant evidence. Experiments across eight image-understanding benchmarks show that DIVE consistently preserves performance across token budgets. With an 88.9% reduction in visual tokens, DIVE retains 98.2% of the uncompressed model’s average performance. Code is available at this https URL.
[CV-75] Not All Redundant Tokens Are Alike: Analyzing Visual Token Pruning through Token Roles ECCV2026
链接: https://arxiv.org/abs/2608.04483
作者: Hyeonyu Kim,Sehwan Lim,Youngwon Choi,Taeyoun Kwon,Jaejin Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026 workshop, UniWorld
Abstract:Vision-language models (VLMs) process an image as a sequence of visual tokens, which creates a substantial computational bottleneck during inference. Recent visual token pruning methods address this issue by removing seemingly redundant tokens, yet it remains unclear how these pruning decisions relate to the functional roles of visual tokens. In this work, we analyze visual token pruning through the lens of token roles identified by EmbedLens. We first show that representative pruning methods exhibit distinct token-role biases, but these biases do not directly correlate with downstream performance. To better understand this behavior, we refine the token-role assignment procedure and evaluate role-protected pruning variants. Our results show that preserving non-alive tokens can sometimes maintain or improve performance, suggesting that tokens with weak direct semantic alignment may still affect model behavior under pruning. Our code is publicly available at this https URL.
[CV-76] REZE: Recognition-Based Zero-Shot Extraction for Video Temporal Grounding
链接: https://arxiv.org/abs/2608.04480
作者: Boyang Li,Chenhui Gou,Jianfei Cai
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 7 figures, 13 tables. Appendices included
Abstract:Video temporal grounding (VTG) refers to the task of identifying the time interval in a video that corresponds to a given natural-language query. A common zero-shot strategy asks a large vision-language model (VLM) to generate the start and end timestamps directly, so the result depends heavily on the design and training of the model, and grounding accuracy differs widely from one VLM to another. We therefore propose REcognition-based Zero-shot Extraction (REZE), a simple training-free method that splits the video into short clips, asks the model for a clip-level confidence score for the query, and uses a deterministic algorithm to convert the resulting score curve into the output required by the task. Because temporal aggregation is performed outside the model, REZE adapts to different task outputs, from single- and multi-interval moment retrieval to highlight detection. On QVHighlights, REZE improves the best reported training-free moment-retrieval mAP from 38.23 to 40.32, while on highlight detection it reaches 44.18 mAP and 73.41 HIT@1, establishing a new state of the art among training-free methods. Its HIT@1 also outperforms all fully supervised SoTAs on the QVHighlights test split. We evaluate REZE on seven backbones from three model families. On Charades-STA and QVHighlights, it outperforms direct timestamp generation in every available comparison. We further observe that with REZE an earlier-generation model can approach the native performance of a newer model in its family.
[CV-77] Beyond Global Routing Aggregation: Phase-Aware Expert Merging for MoE Vision-Language Models
链接: https://arxiv.org/abs/2608.04454
作者: Hongyu Zhang,Cheng Yan,Xiang Xia,Wuyang Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 17 pages, 3 figures, 17 tables
Abstract:Mixture-of-experts vision-language models (MoE-VLMs) increase model capacity with sparse expert activation, yet deployment requires storing the full expert pool. Training-free expert merging reduces this burden, and many routing-based methods aggregate routing statistics across all tokens to determine merge compatibility. However, MoE-VLM inference is phase-structured: image-context tokens carry visual content, question tokens specify the query, and answer tokens produce the output, with different counts and routing distributions. Because image-context tokens are far more numerous, global aggregation can overemphasize image-context processing and obscure phase-conditioned expert roles, making experts serving different phases appear interchangeable and degrading model performance. We therefore argue that MoE-VLM expert merging should preserve phase-conditioned expert roles, judging compatibility by how experts serve different phases rather than globally aggregated routing statistics. Based on this view, we propose RoleMerge, a training-free method that constructs each expert’s Routing Role Profile (RRP) from phase-normalized routing statistics, capturing its relative phase preference. Guided by expert-phase information loss, RoleMerge merges experts with compatible profiles and their corresponding router entries while preserving answer-decoding expert distinctions. Experiments on three models and multiple benchmarks show that RoleMerge preserves more of the full model’s performance than alternative expert-merging methods at matched expert-retention ratios, with relative improvements of up to 9.6 percent in six-task macro-average performance. These results validate phase-conditioned expert roles as a more effective basis than global routing aggregation for MoE-VLM expert merging.
[CV-78] winIR: Coordinated Invisible Dual-Point Attacks on Online HD Map Construction
链接: https://arxiv.org/abs/2608.04453
作者: Haibo Hu,Jianghuai Deng,Chen Tang,Yang Lou,Qian Xu,Jianping Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Online HD map construction is critical to prediction and planning in autonomous driving. We find that existing physical attacks against online map construction are limited by a cross-boundary compensation effect: after the target boundary is perturbed, another visible boundary may retain sufficient geometric cues for the model to recover the original road geometry. Based on this observation, we propose TwinIR, a new mechanism-guided physical attack methodology for online map construction. TwinIR jointly optimizes attack effectiveness and point sparsity, seeking the minimum number of attack points needed to suppress compensating geometric cues from surrounding boundaries. To reduce the perceptibility of multi-point attacks, TwinIR models camera responses to near-infrared illumination and maps optimized attack points to feasible physical placements, producing camera-visible interference with minimal visible-spectrum changes. Experiments on nuScenes across state-of-the-art online map construction models show that TwinIR reduces mAP by 8.18-8.96 percentage points under RSA and 2.84-5.62 points under ETA, while increasing the unreachable-goal rate by 25-28 points and the unsafe-planned-trajectory rate by 19-20 points over clean inputs. These attacks are also validated on a real-world testbed AV, where TwinIR successfully induces both road straightening and early-turn deformations while remaining inconspicuous in full-color views.
[CV-79] When does training on downscaled images yield the same gradients?
链接: https://arxiv.org/abs/2608.04448
作者: Seunghyun Ji
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion transformers deliver strong image generation, but their training cost grows superlinearly with resolution. Recent work justifies training or sampling at reduced resolution on a spectral premise: at high noise, a downscaled latent preserves almost the full surviving signal. Whether a downscaled step also preserves the native training gradient signal, however, has remained unresolved. We reduce how that signal changes under downscaling to two terms: a noise-dependent term governed by the downscale ratio, which decays at high noise as the spectral premise predicts, and a \sigma-independent floor governed by the target grid’s absolute token count, carried by the compute graph itself and removed by no noise level. The measured (route, \sigma) map corroborates the account and uncovers structure the spectral picture cannot express: on the 1024-768 route, a window (0.65 \sigma 0.95), predicted by no spectral criterion at any tolerance, where the downscaled gradient stays within a small margin of the native one. Training LoRA adapters with downscaled steps restricted to the routes and noise windows the map validates reduces training time by 14.6% at a fixed step budget while remaining near-native in weight space. Code is available at this https URL.
[CV-80] Robustness Emerges Early in Training Dynamics but Is Not Preserved ECCV2026
链接: https://arxiv.org/abs/2608.04442
作者: Jiangang Yang,Wenhui Shi,Lu Hu,Jing Xing,Jian Liu
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by ECCV2026
Abstract:Robustness to natural corruptions remains a fundamental challenge for deep neural networks. In this paper, we identify a robustness fading phenomenon where shallow layers spontaneously develop robust representations and flat loss landscapes in early training, yet these properties are not preserved during standard convergence. To address this, we propose a framework that performs strategic interventions on training dynamics to stabilize the empirically identified early-emergent robust priors. Our approach includes two parameter-free strategies: Early-Phase Stabilization~(EPS) and Asymmetric Weight Reversion~(AWR), which stabilize or recover robust shallow configurations without modifying the model architecture or introducing learnable parameters. Extensive experiments demonstrate the efficacy of our framework across various benchmarks and architectures, yielding significant gains in downstream transfer, dynamic adaptation, and diverse computer vision applications.
[CV-81] Season: Spectrum-Aware Orthogonal Gradient Refinement for Transfer-Based Adversarial Attacks
链接: https://arxiv.org/abs/2608.04441
作者: Tianyi Wang,Zhenghao Gao,Shengjie Xu
类目: Computer Vision and Pattern Recognition (cs.CV); Cryptography and Security (cs.CR)
备注: 6 pages
Abstract:Transfer-based adversarial attacks often transfer poorly across heterogeneous architectures because CNNs favor local textures while Vision Transformers (ViTs) rely on global shapes. We propose Season, a spectrum-aware orthogonal gradient refinement framework for L-infinity transfer attacks against black-box target models on ImageNet, using a white-box surrogate. Season decomposes each update into a low-frequency branch capturing structural cues and a high-frequency branch capturing textures. A low-saliency guidance scheme reallocates high-frequency energy to background regions, preserving foreground structures that ViTs depend on. An orthogonal projection then forces the textural update to lie in the orthogonal complement of the structural direction, mitigating feature interference. As a training-free plug-and-play wrapper, Season enhances eight gradient-stabilization and input-enhancement attacks without modifying their cores. Across eight CNN, ViT, and MLP targets, Season improves transfer success rate by 6.6 percentage points on average and up to 16.0 points over strong baselines under a unified protocol.
[CV-82] oolArtist: Tool-Using Unified Multimodal Models for Agent ic Image Generation
链接: https://arxiv.org/abs/2608.04436
作者: Jiahao Zhao,Xiaomin Yu,Zhongxiang Sun,Fengwei Teng,Chengwei Qin,Xiaobin Hu,Jun Xu,Shuicheng Yan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Text-to-image (T2I) models can produce visually compelling images, yet they remain limited on open-world tasks that require complex semantic understanding, multi-step reasoning, and the integration of external world knowledge. Existing efforts introduce agent capabilities into image generation, but they either prescribe a fixed workflow or place only a subset of the open-world image generation process under agent control. Consequently, reasoning, tool invocation, and image generation are not coordinated by a single policy. We propose ToolArtist, a fully agentic image generation model obtained by post-training a Unified Multimodal Model (UMM). ToolArtist dynamically orchestrates reasoning, external tool use, and native image generation within one unified policy. During Supervised Fine-Tuning (SFT), we equip a teacher agent with search tools alongside an image-generation tool. We then convert the collected trajectories into a UMM compatible format, where the image-generation tool is concealed while the resulting generated images are retained. During Reinforcement Learning (RL), we develop an agentic RL infrastructure for UMMs and introduce Reason-Act-Draw GRPO (RAD-GRPO), which uses complementary intent and quality rewards to jointly optimize the model. Experiments show that placing the entire open-world image-generation process under an agent policy consistently outperforms approaches with fixed pipelines or only partially agent-controlled components. We release the training data and the complete post-training infrastructure.
[CV-83] OmniRouting: A Semantic-Coupled Multimodal Benchmark for Constraint-Aware Spatial Reasoning in PCB Routing
链接: https://arxiv.org/abs/2608.04434
作者: Taiting Lu,Kaiyuan Lin,Ziwei Dong,Sisong Bei,Haolin Ye,Yuxin Tian,Runze Liu,Mingjia Wang,Jingying Zeng,Hongxing Pan,Kai Zhang,Haoyu Wang,Guoliang Shi,Ling Ma,Yifan Yang,Jiaying Lu,Qi He,Yi-Chao Chen,Sung-Liang Chen,Yincheng Jin,Mahanth Gowda
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recent large language models (LLMs) have demonstrated remarkable progress in constraint-aware navigation, maze reasoning, and graph reasoning. However, their ability to reason about complex routing problems under strict geometric, topological, and electrical constraints remains largely unexplored, despite routing being one of the most challenging and critical stages of electronic design automation (EDA). To bridge this gap, we introduce OmniRouting, the first large-scale benchmark designed to evaluate LLMs on printed-circuit-board (PCB) routing reasoning under real-world industrial design-rule, manufacturability, and connectivity constraints. OmniRouting contains 1,681 industrial-grade schematic-coupled PCB designs, including board geometries, routable component placements by human engineers, footprints, pad locations, netlists, stackup information, and routing constraints. The benchmark comprises four tasks: (1) geometric routing reasoning, generating physically valid copper traces, vias, and layer assignments to connect circuit nets within constrained board regions; (2) design-rule-aware routing reasoning, producing routable layouts that satisfy clearance, trace-width, via, obstacle-avoidance, and board-boundary constraints; (3) electrical functionality reasoning, preserving schematic-specified connectivity while reasoning over net names and functional roles to produce electrically correct routing; and (4) tool-augmented agentic routing, leveraging external tools for tasks (1)-(3). Our results reveal substantial limitations of current LMMs in PCB routing, including weak path-planning capabilities, poor adherence to design-rule constraints, and inconsistent preservation of electrical functionality. We will open-source all benchmark data, evaluation code, and tool interfaces to facilitate future research.
[CV-84] UBLLIE: Unified Backlight and Low-Light Image Enhancement
链接: https://arxiv.org/abs/2608.04429
作者: Yasmin Yasin,Muhammad Usman,Ibrahim Radwan,Saeed Anwar
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Backlit and low-light images often suffer from severe exposure imbalance or global underexposure, presenting significant challenges for both visual perception and downstream computer vision tasks. In this paper, we propose a unified, unsupervised enhancement framework that addresses both types of degradation without relying on paired ground-truth data. Our approach builds on CLIP-guided prompt learning to semantically supervise enhancement using learned positive and negative textual prompts. To improve the quality of our improvements over prior work, we design a symmetric residual U-Net backbone augmented with an Atrous Spatial Pyramid Pooling module. This architecture captures multi-scale contextual information, enabling adaptive correction under spatially heterogeneous illumination. During training, the enhancement network is guided by CLIP-based semantic similarity losses and refined via an iterative prompt optimization mechanism. Extensive experiments on both paired and unpaired datasets, including BAID, Backlit300, LOL, and VE-LOL-L, demonstrate that our framework consistently outperforms state-of-the-art supervised and unsupervised methods in terms of fidelity, perceptual quality, and generalization. Furthermore, our work emphasizes the need for stronger benchmarking protocols for backlit enhancement, a relatively underexplored area. The proposed framework provides a robust, scalable solution for real-world illumination enhancement across diverse lighting conditions.
[CV-85] hinking with Anchors: Grounded and Efficient Document Reasoning
链接: https://arxiv.org/abs/2608.04424
作者: Sichen Zhu,Yuchen Zhu,Wenzhuo Xu,Jason Kuen,Wanrong Zhu,Jing Shi,Xuan Shen,Quanyi Wang,Yiwei Wang,Yujun Cai,Bing Shuai,Qin Zhang,Yongxin Chen,Shilong Liu,Molei Tao,Jiuxiang Gu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Existing document understanding benchmarks have largely focused on locating page elements, yet real-world document intelligence requires models to reason jointly about region semantics, spatial relations, and visual structure. We present ADOPD 2026, a reasoning-oriented extension of ADOPD that turns page decomposition into spatially grounded document understanding. ADOPD 2026 enriches page anchors inherited from ADOPD 2024 dataset with human-cleaned captions, semantic tags, and generated chain-of-thought (CoT) traces grounded to document regions. Instead of treating boxes, masks, and tags as independent supervision signals, we cast text blocks, visual entities, semantic labels, bounding boxes, and polygon masks as a shared vocabulary of visual anchors. This representation supports three connected capabilities. First, region-level semantic tagging asks models to identify document element types from both page context and local appearance, revealing long-tail semantic failures that standard layout benchmarks often hide. Second, unified vision-language grounding generates text regions and visual entities together with coordinates or polygonal outlines, transforming detection and segmentation outputs into structured anchors that can be reused by downstream reasoning systems. Third, current state-of-the-art models still struggle with dense counting tasks evaluated on DocCount, a benchmark derived from ADOPD 2026, highlighting the need for the Thinking-with-Anchors pipeline in document semantic understanding. By connecting page decomposition to verifiable visual-anchor reasoning, ADOPD 2026 provides a task framework that moves document understanding beyond localization toward anchor-grounded document intelligence.
[CV-86] Foreseeing the Invisible: Amodal Reconstruction of Leaf Fossil Images
链接: https://arxiv.org/abs/2608.04423
作者: Liuxiang Yue,Ailin Zhang,Ziyue Zhao,Yikun Duan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 12 pages, 10 figures
Abstract:Fossil leaves are rarely preserved whole – sedimentary rock hides, breaks, and erodes the lamina, yet paleobotany depends on the complete shape and outline of the leaf. We cast the recovery of the missing tissue as amodal reconstruction and present AmodalDINO, a multi-head dense-prediction model that predicts four masks from a single RGB image: visible leaf, amodal complete leaf, amodal main vein, and fine veins. Unlike essentially all prior amodal work, AmodalDINO is given no visible mask. It predicts the visible and amodal regions jointly, so it needs no upstream instance segmenter at runtime. Two simple but effective changes adapt the model to the amodal segmentation task: fully fine-tune a DINOv3 ViT-L/16 at a small learning rate instead of freezing it, and attach auxiliary venation heads alongside the leaf heads. These two changes enable the model to learn the structural shape prior of leaves. Trained only on synthetic leaf fossil images, AmodalDINO reaches 95.0% Dice / 90.5% IoU on the validation set and transfers well to real fossil specimens. Stripped to two heads, the same recipe can run on two benchmark datasets, reaching 85.05 full mIoU / 66.65 occluded mIoU on KINS and 80.90 / 38.15 on COCOA-cls. The model is also practical: by quantizing to 4-bit weights, it runs entirely offline in a browser, matching the original model with an IoU of 0.910. We also add ruler-based calibration to estimate surface area, and a generative visualization of living leaves on local devices.
[CV-87] muSync-GS: Physics-Synchronized Driving Video Synthesis for Weather and Geometric Road Hazards
链接: https://arxiv.org/abs/2608.04412
作者: Yang Chen,Yicheng Zhu,Tao Li,Zilin Bian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 42 pages, 14 figures; includes an appendix
Abstract:High-quality driving data are essential for autonomous-driving systems and generative world models. However, rare and safety-critical scenarios involving adverse weather, braking under low tire–road friction, and uneven road geometry are costly and risky to collect at scale. Existing video-generation and 3D Gaussian editing methods can modify weather appearance or road geometry, but typically do not couple these edits with tire–road interaction and vehicle dynamics. As a result, an edited video may retain its original trajectory even when the modified road condition should alter braking, wheel slip, load transfer, and ego-camera motion. We present muSync-GS, a physics-synchronized framework for driving video synthesis under adverse-weather and road-elevation hazards. A precipitation-derived road-surface condition jointly controls road appearance and tire friction, while a shared road-elevation profile drives both visible road-geometry editing and axle excitation. A calibrated vehicle model predicts speed, slip ratio, normal loads, and pitch for constructing the ego-camera trajectory and synchronized physical annotations. On 12 held-out CarSim cases spanning precipitation levels, brake inputs, and road-profile parameters, the model achieves mean case-wise RMSEs of 0.0273 m/s for speed, 0.0590 degrees for pitch, 0.0101 for slip ratio, and 26.61 N for per-wheel normal load. Together with the reconstructed-scene experiments, these results show that muSync-GS accurately reproduces vehicle responses under held-out controls while synchronizing them with controllable scene edits and ego-camera motion.
[CV-88] Faster-WAM: Efficient Inference-Time Future Conditioning for Robust World Action Models
链接: https://arxiv.org/abs/2608.04404
作者: Weiheng Zhao,Haoyi Jiang,Xin Shi,Liu Liu,Fan Huang,Zhizhong Su,Wei Sui,Xinggang Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:World Action Models (WAMs) improve robot manipulation by learning how the environment evolves beyond the current observation. However, existing approaches face a fundamental dilemma: Joint-WAMs preserve future-aware representations during inference but incur prohibitive computation costs, while efficient alternatives remove future modeling at inference time and may lose the robustness benefits of temporal reasoning. In this work, we revisit the role of future representations in WAMs and show that inference-time future conditioning is critical for generalization under distribution shifts. This observation motivates Faster-WAM, an efficient future-conditioning WAM that preserves future representations while avoiding expensive video-action interaction. Faster-WAM introduces a sparse future-conditioning framework that computes future representations once and selectively reuses them throughout action denoising. Specifically, we propose SparseMoT to replace ubiquitous layer-wise fusion with selective video-action interaction at a compact subset of network stages, and Interval KV-Fusion to aggregate multi-depth future representations without increasing attention complexity. Experiments demonstrate that Faster-WAM achieves a substantially better performance-efficiency trade-off than existing WAMs. On the out-of-distribution LIBERO-Plus benchmark, Faster-WAM improves success rate from 49.14% to 73.57% compared with Fast-WAM, while running 2.21 \times faster than Joint-WAM. It further achieves state-of-the-art performance on LIBERO and RoboTwin 2.0, while demonstrating strong robustness in real-world manipulation.
[CV-89] CofactVLA: Deconfounding Vision-Language-Action Models via Counterfactual Intervention
链接: https://arxiv.org/abs/2608.04396
作者: Yan Zhang,Yinan Wu,Haoran Duan,Jungong Han
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-Language-Action (VLA) models have driven significant progress in robotic manipulation, yet they fundamentally struggle with the vision-override phenomenon. Driven by the severe modality imbalance between dense visual streams and sparse linguistic instructions, VLAs frequently fall prey to causal confusion. Instead of treating language as the primary causal driver, the policy entirely bypasses the original instruction by overfitting to spurious visual confounders, such as prominent objects or familiar layouts. To systematically alleviate this bias, we formalize the process of action generation as a Dual-path Deconfounding Graph (DDG) and propose CofactVLA, a novel causal intervention framework. By dynamically constructing a language-masked counterfactual branch within a single forward pass, CofactVLA isolates and neutralizes visual confounders through two synergistic mechanisms. First, Action-Level Orthogonal Projection Guidance (OPG) geometrically projects the factual velocity field away from the counterfactual visual bias during continuous flow matching, extracting the pure semantic intent. Second, Feature-Level Counterfactual Covariance Reduction (CCR) mathematically deconfounds latent representations by penalizing the positive eigenspace of the covariance difference, explicitly suppressing dominant visual shortcuts while preserving the causal language intent. Extensive experiments demonstrate that CofactVLA establishes a new state-of-the-art across diverse simulation benchmarks. Beyond simulation, real-world robot experiments demonstrate the causal efficacy of our method in bridging the generalization gap, yielding a 52.3% absolute success rate gain under out-of-distribution scenarios.
[CV-90] Free-Lunch Augmentation by Revisiting Diffusion-Based Data Generation for Cross-Domain Few-Shot Object Detection
链接: https://arxiv.org/abs/2608.04394
作者: Zijian Zhuang,Yixiong Zou,Yuhua Li,Ruixuan Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Cross-Domain Few-Shot Object Detection (CDFSOD) aims to transfer knowledge from data-rich upstream generic domains to downstream expert domains using scarce training data, where the significant domain gap and data scarcity make it an unsolved challenge. To address this problem, we revisit a natural yet underexplored approach in CDFSOD: data augmentation, by directly synthesizing data through diffusion models to supplement limited training samples. However, due to large domain gaps, we find that current diffusion methods cannot produce good results, leading to performance even lower than using the original images. To address these limitations, we divide the domain gaps into visual gaps and semantic gaps for separate analysis. For the visual gap, we find that the diffusion model cannot distinguish noise from useful information on expert domains, which can be mitigated by adding weakened noise. For the semantic gap, we find that the background semantics shows much smaller gaps between domains than foreground semantics, and we can bridge this gap by background inpainting. Based on the above analysis, we propose a method (Selective Inpainting with Tailored Noise, SITN) to dynamically take different strategies for downstream data synthesis based on their different gaps from the general domain, including a Generation Module for adding tailored noise and a Selection Module to dynamically select the inpainting regions. Extensive experiments on 6 datasets of CDFSOD and 4 datasets of cross-domain few-shot segmentation (CDFSS) validate that we can synthesize helpful data, achieving new state-of-the-art performance. Our codes is available at this https URL
[CV-91] ReGround: Restoring Visual Grounding in Multi-Step Reasoning through Self-Diagnosis and Visual Re-Examination
链接: https://arxiv.org/abs/2608.04385
作者: Lei Peng,Shuai Lv,Wei Hu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ACM Multimedia 2026 (MM '26). 8 pages main text, 4 figures, plus appendix
Abstract:Vision-Language Models (VLMs) often lose visual grounding during multi-step reasoning: as reasoning chains grow longer, later inference steps rely increasingly on language priors rather than image evidence. We identify a consistent benchmark-level signature associated with this degradation: across 2,510 re-examined samples from four benchmarks, attention entropy over image tokens typically decreases during Round 1 and rises again after image re-injection. However, we find that effective visual re-examination requires two complementary ingredients: image re-injection and targeted self-diagnosis. Without targeted diagnosis, re-examination can even hurt performance, whereas accurate self-diagnosis yields substantial gains – a swing of several points on key benchmarks, indicating that diagnostic quality is a key factor in whether re-examination helps or hurts in our setting. We present ReGround, a two-stage framework that teaches VLMs to self-diagnose grounding failures and selectively re-examine visual evidence, without architectural modifications or external tools. Through capability bootstrapping, a stronger variant from the same model family provides diagnostic scaffolding only during data construction, while the policy model learns to diagnose autonomously at inference time and retains most of the assisted gains. Experiments on eight benchmarks across two VLM backbones demonstrate consistent gains, especially on visually intensive multi-step reasoning tasks, while incurring only modest inference overhead relative to tool-augmented baselines. Project page: this https URL . Code: this https URL . Comments: Accepted to ACM Multimedia 2026 (MM '26). 8 pages main text, 4 figures, plus appendix Subjects: Computer Vision and Pattern Recognition (cs.CV) ACMclasses: I.2.10; I.2.7; I.4.8 Cite as: arXiv:2608.04385 [cs.CV] (or arXiv:2608.04385v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2608.04385 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-92] Image Classification Using CNN-QNN Hybrid Model with Optimized Correlated Features CVPR
链接: https://arxiv.org/abs/2608.04379
作者: Minseo Seong,Youngwook Kim
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted by CVPR Findings 2026
Abstract:We propose a method to optimize the correlation among convolutional neural network (CNN) features that are used as inputs to quantum neural network (QNN) to enhance image classification accuracy. Unlike prior approaches that employ orthogonal decomposition as preprocessing, we intentionally introduce correlated features that are more physically compatible with QNN. This design leverages the QNN’s inherent ability to exploit quantum entanglement for representing correlated states-an advantage unavailable to classical neural networks. We hypothesize that aligning feature correlations with the entanglement structure of QNN improves binary classification performance. Based on a mathematical derivation of QNN outputs, Monte Carlo simulations indicate that an average correlation between features of 0.5 yields optimal classification accuracy. To validate this finding, we evaluate a quantum-classical hybrid model on three tasks: CIFAR-10 (automobile vs. truck), Fashion-MNIST (shirt vs. coat), and radar micro-Doppler signatures (robotic dogs vs. non-robots). To regulate feature correlations, we introduce a correlation-regularization term on the outputs of the CNN, driving the off-diagonal entries of the feature correlation matrix toward a target constant. Across all datasets, inducing intermediate correlation consistently improved accuracy compared to low, high, or unregulated correlations, while also reducing classification accuracy variance. These results demonstrate that imposing moderate feature correlations-without modifying the quantum circuit-enhances classification accuracy and stability by aligning feature statistics with the QNN’s entanglement structure. This study highlights the potential of QNN to surpass the performance of classical classifiers as more qubits become available.
[CV-93] Poly-OPD: Heterogeneous Multi-Teacher On-Policy Distillation for Capability-Selectable Flow Models
链接: https://arxiv.org/abs/2608.04349
作者: Siming Fu,Haojun Xu,Ruizhe He,Zheming Fu,Hualiang Wang,Jie Huang,Xiaoxiao Ma,Mingchen Zhong,Weihu Huang,Xiaoxuan He,Linjiang Huang,Si Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Leading open text-to-image models often carry complementary strengths: one may lead on preference-aligned aesthetics while another follows compositional instructions more faithfully. However, differences in their autoencoders and noise schedules make it difficult to transfer these strengths across models. In this paper, we present Poly-OPD, a framework that can consolidate complementary strengths of heterogeneous teachers into a single compact flow-matching student. To bridge the incompatible latent spaces of different teachers, Poly-OPD performs on-policy distillation through a pixel bridge. Each student-generated image is re-encoded by a selected teacher’s encoder and refined from a noise level matched by magnitude under the teacher’s noise schedule. The resulting target is further matched to the student in frozen DINOv2 space, enabling supervision across incompatible latent spaces. To retain complementary capabilities without cross-teacher interference, Poly-OPD uses a gradient compatibility diagnostic to organize its adapters: attention LoRA modules are shared across teachers, whereas feed-forward adapters remain teacher-specific. During distillation, a gap-aware curriculum devotes more training to compositional categories where the student still falls short of the teacher. As each gap narrows, training shifts toward categories with larger remaining gaps. By distilling FLUX.1-dev and Z-Image into a 2.5B SD3.5-Medium student, Poly-OPD improves GenEval from 67.3 to 73.3, surpassing both larger teachers, and raises DrawBench HPSv3 from 9.34 to 11.35, consolidating both strengths within a switchable model.
[CV-94] StructTab: Structured Feature Sequencing for Multimodal Learning of Image and Tabular Data ICPR2026
链接: https://arxiv.org/abs/2608.04348
作者: Al Zadid Sultan Bin Habib,Md Younus Ahamed,Prashnna Gyawali,Gianfranco Doretto,Donald A. Adjeroh
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
备注: This paper has been accepted for presentation at the 28th International Conference on Pattern Recognition (ICPR 2026) in Lyon, France Code: this https URL PyPI: pip install istructtab
Abstract:Multimodal learning of images and tabular data is often impaired by ineffective representations, resulting in redundancy, dispersion, and generalization problems. To tackle this challenge, we introduce Graph-Enhanced Descriptor Sequencing (GEDS), a structured feature sequencing algorithm grounded in principles from the Column Permutation Problem (CPP). GEDS refines statistical descriptors of the features through similarity graph-based computations, systematically determining an effective feature sequencing. We incorporate GEDS within an order-aware efficient transformer framework, utilizing order-aware memory tokens that explicitly adhere to the derived feature sequencing via a dedicated loss function. Experimental results across multimodal benchmarks demonstrate that iStructTab effectively minimizes feature dispersion, improving predictive performance and robustness, and highlighting the significance of structured feature sequencing in multimodal learning.
[CV-95] An Analysis and Implementation of Seam Carving for Content-Aware Image Resizing
链接: https://arxiv.org/abs/2608.04329
作者: Francesco Tosoni
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: one column, 10 pages, 8 figures
Abstract:Seam carving is a classical content-aware image resizing operator that modifies the width or height of an image by repeatedly removing (or inserting) seams, i.e., 8-connected monotonic paths of pixels of locally minimal importance. Because seams bend around salient content rather than uniformly scaling or cropping it, the operator preserves vital image structures while discarding (or duplicating) low-energy regions. This article describes a C++ implementation of the operator that follows the original formulation of Avidan and Shamir (2007), including the optional forward-energy criterion subsequently introduced by Rubinstein, Shamir and Avidan (2008). The implementation supports image reduction, image enlargement via ordered seam insertion, multi-pass enlargement for large scale factors, a user-supplied weight mask for object protection and removal, along with dumping of energy maps and visualisation of seams. We detail the algorithm, its parameters and its computational complexity, discuss design choices with respect to the original descriptions, and illustrate the behaviour of the operator on natural images.
[CV-96] Adversarial Attacks for Good: A Survey of Proactive Protection across the Visual Content Lifecycle
链接: https://arxiv.org/abs/2608.04314
作者: Jiaming Zhang,Boyang Chen,Zherui Li,Fuyao Zhang,Xinyu Yan,Hong Xi Tae,Wenwen He,Xuan Wang,Siqi Guo,Junhao Dong,Kun Wang,Hanxun Huang,Yige Li,Xingjun Ma,Yang Cao,Lingjuan Lyu,Wei Yang Bryan Lim
类目: Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Once visual content enters an AI pipeline, its owner often retains little technical control over how it is used. Legal and regulatory remedies can address misuse, but many technical interventions must be applied earlier, when content is released or accessed. This survey examines the protective paradigm that has grown around this intervention point, which we call \emphadversarial attacks for good. Perturbations and structured signals long studied as attacks on learned models are instead applied by data owners, creators, platforms, or auditors to disrupt unauthorized automation or support later accountability. Five research communities have arrived at this inversion largely independently, each addressing a different stage of a visual asset’s lifecycle: privacy filters against unwanted recognition at sharing time, unlearnable examples against unauthorized training, generative safeguards against malicious editing or imitation, adversarial CAPTCHAs for access control against automated agents, and provenance mechanisms for post-circulation attribution. Although developed in separate venues with incompatible success criteria, many of these methods exploit persistent gaps between human perception, semantic interpretation, and machine inference, suggesting that the paradigm remains relevant as visual pipelines evolve toward multimodal models and autonomous agents. To make their claims comparable, we evaluate all five families along shared axes of transferability, adaptability, and deployment readiness. Across the lifecycle, we find that most protections are still validated mainly against static or weakly adaptive adversaries, while evidence beyond controlled benchmarks remains scarce. We close by consolidating cross-stage countermeasures and open problems for robust, composable, and deployable owner-side protection.
[CV-97] Binding Biometrics with AI Agent Identifiers for Delegation of Authority
链接: https://arxiv.org/abs/2608.04292
作者: Joseph Geo Benjamin,Anil K Jain,Karthik Nandakumar
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted in IJCB sessions 2026
Abstract:The proliferation of agentic artificial intelligence (AI) systems has raised serious questions about the accountability for tasks performed by AI agents. Ideally, an AI agent must not be allowed to perform critical tasks without explicit authorization by a human operator. Since biometric recognition is one of the most reliable approaches for authenticating individuals, it has the potential to enable authenticated delegation of authority to AI agents. In this work, we present a framework called BIND, which leverages ideas from the field of biometric cryptosystems, to securely bind biometric data of the human user to the AI agent identity (ID) and authority scope (task-specific constraints) at the time of agent authorization. This token/identifier can be presented by the AI agent to an Identity Auditor, who simultaneously performs biometric authentication and recovers the agent ID and scope, thereby enabling real-time user authentication and establishing a non-repudiable proof of human control and delegation of authority. We also provide a practical implementation of the proposed BIND framework based on face features extracted using standard deep neural network models. To facilitate this implementation, we propose a feature adaptation module that transforms real-valued feature embeddings into fixed-length binary representations suitable for a fuzzy commitment construct based on turbo error correcting codes. Experiments demonstrate the practical feasibility of the proposed face cryptosystem, achieving a True Match Rate of 96% at zero False Match Rate and supporting 1024 -bit agent tokens.
[CV-98] SAFECAST: Robust Failure Detection for VLA Policies with Contrast-Set Training and Calibration
链接: https://arxiv.org/abs/2608.04246
作者: Harshitha Rajaprakash,Aditeya Prajapati,Rong Xue,Abrar Anwar,Jesse Thomason
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-language-action policies often fail under deployment-time distribution shifts such as clutter, distractor objects, lighting changes, novel objects, altered initial states, and reworded instructions. Hidden-state-based risk probes combined with functional conformal prediction can detect rollout failures, but their reliability depends on calibration data matching deployment conditions. We introduce SAFECAST, which leverages contrast set perturbations to improve hidden-state probe training and calibration for deployment time shift. SAFECAST statistically significantly improves failure detection ROC-AUC scores over a state of the art baseline in both real-world DROID and LIBERO simulation experiments across multiple VLM backbones. We further find that SAFECAST benefits most when both visual and language contrast set perturbations are used to augment data, and that with contrast set perturbations, sim-to-real calibration leads to better probes than using real rollout data only.
[CV-99] ransferable Dual-Stream Representations for Mesoscale-Preserving Sea Surface Temperature Downscaling
链接: https://arxiv.org/abs/2608.04230
作者: Parth Doshi,Priyanka Aravindan,Vaishnav Vaidheeswaran,Md Mahbub Alam,Gabriel Spadon
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Deep learning models for scientific spatio-temporal downscaling often minimize reconstruction error while failing to preserve physically meaningful multi-scale structure. For sea surface temperature prediction, this can yield outputs that are numerically plausible yet overly smooth, missing mesoscale variability critical to regional ocean dynamics. Existing methods often focus on pixel-wise objectives or single-context conditioning, which limits their ability to preserve spectral fidelity and generalize across regions. To address this, we propose EddyFlow, a representation learning framework for kilometer-scale sea surface temperature downscaling that balances predictive accuracy, scale-dependent structure, and regional generalization. EddyFlow is trained on the Gulf of St.~Lawrence and evaluated in zero-shot and few-shot settings on the Bay of Fundy and the Gulf of Mexico. EddyFlow demonstrates that physics-informed representation learning reduces zero-shot RMSE by 21%, achieves up to 85.6% skill relative to persistence on unseen domains, and maintains near-ideal spectral fidelity with a PSD ratio of \approx 1.00 .
[CV-100] OmniVR: Joint Video-Audio Conditional Generation for Restoring Degraded Historical Films
链接: https://arxiv.org/abs/2608.04224
作者: Xin Lu,Zihao Fan,Mingchen Zhong,Jie Huang,Xueyang Fu,Zheng-Jun Zha
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Historical films suffer from co-occurring visual and audio degradations—blur, noise, flicker, hiss, clipping, and dropout—yet existing methods restore each modality independently, leaving quality gaps and cross-modal inconsistency. We present OmniVR, the first joint audio-video generative restoration model. Built upon a 22B-parameter audio-video generation backbone, OmniVR formulates restoration as conditional generation within a unified multimodal DiT: the low-quality video and audio are encoded as latent conditions, combined with a fixed restoration prompt, and jointly denoised to recover visual structure, temporal motion, and acoustic detail under one coordinated objective. Three key designs enable this adaptation: (1) a joint audio-video degradation pipeline that simulates real old-film characteristics from Internet-collected data; (2) an architecture-preserving text-to-audio-video (T2AV) to audio-video-to-audio-video (AV2AV) transition with prompt annealing that maximally retains the generative prior; and (3) first-frame image-to-video (I2V) anchoring with loss reweighting and waveform supervision for long-video extrapolation and audio fidelity. We also propose OmniVRBench, the first benchmark that evaluates audio-video restoration across visual quality, audio quality, temporal consistency, and audio-visual synchrony on 200 real historical clips. OmniVR surpasses all prior methods on all six visual metrics, achieves the best audio quality, and produces natural colorization—the first method to jointly address all three aspects. Code and weights will be publicly released. Project Page: this https URL
[CV-101] PADFormer: Pose-agnostic Anomaly Detection from Sparse View Images ECCV2026
链接: https://arxiv.org/abs/2608.04210
作者: Ruiqi Wang,Yiming Qian,Fenggen Yu,Yuxuan Lu,Dakuo Wang,Hao Zhang,Jing Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026 (oral)
Abstract:Pose-agnostic Anomaly Detection (PAD) remains challenging as anomalies can appear under arbitrary viewpoints, requiring methods to handle significant pose variations. Existing approaches rely on complex 3D reconstruction, which are computationally expensive and require extensive multi-view data. We propose PADFormer, a novel image-space approach that leverages Vision Transformer (ViT) to directly reconstruct anomaly-free versions of query images while preserving pose information. Our key insight is to adapt cross-view masked reconstruction for anomaly detection through training exclusively on normal data, combined with dynamic patch selection and spatial alignment mechanisms that enable effective learning from sparse reference views under significant pose variations. During inference, we perform multiple forward passes with different masking patterns to generate an ensemble of anomaly-free reconstructions, ensuring comprehensive coverage of the query image. Anomalies are detected by comparing these reconstructions with the query image. PADFormer achieves state-of-the-art results on the PAD benchmark while maintaining comparable performance on classic few-shot anomaly detection (FSAD) tasks, demonstrating superior efficiency and generalization without requiring 3D reconstruction.
[CV-102] SiMDex: Mining Similar Egocentric Videos for Cross-Embodiment Dexterous Manipulation
链接: https://arxiv.org/abs/2608.04196
作者: Nie Lin,Takehiko Ohkawa,Sijin Chen,Ruoshi Wen,Zhuohang Li,Liqun Huang,Zhengming Zhu,Yiming Bao,Yunfei Li,Minjie Cai,Xiao Ma,Wei Xu,Yoichi Sato
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 12 pages, 4 figures. Project page: this https URL
Abstract:Recent years have witnessed an explosive trend of scaling ego-centric human videos for robot manipulation, yet it remains unclear which data actually benefits dexterous manipulation. We present SiMDex, a similarity-based data mining framework that casts human data selection for VLA post-training in dexterous manipulation as a recommendation problem. For each robot demonstration, SiMDex employs a three-layer recall-ranking-re-ranking pipeline to extract task-relevant subsets from a pool of ~32M egocentric human samples, operating in a morphology-agnostic action space that requires no changes to VLA architecture or training. Against a strong baseline trained with an equal amount of randomly sampled human data, SiMDex uses only ~1.49M mined samples (5% of the pool) yet improves the overall success rate from 47.7% to 61.1%, showing that selective curation outperforms indiscriminate data mixing.
[CV-103] Adversarially Robust Abductive Fusion of Pre-trained Transformer-based Perception Models
链接: https://arxiv.org/abs/2608.04190
作者: Mario Leiva,Yue Ma,Qinru Qiu,Gerardo Simari,Paulo Shakarian
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Logic in Computer Science (cs.LO)
备注:
Abstract:Deploying pre-trained perception models in novel environments degrades their accuracy under distributional shift, and assembling them alone does not recover it: combiners such as majority voting trade recall for precision and are brittle to coordinated failures. Prior metacognitive methods learn logical rules that flag a model’s errors, but rely on hand-authored domain-knowledge cues (object-size priors, segmentation masks) that do not transfer to genuinely novel scenes. We show that this metacognitive layer can be learned without any domain knowledge by exploiting vector-space geometry: per-model Label Vector Pools (LVP), built from each model’s own training embeddings, yield error-detection rules from the geometry of detections relative to training-determined prototypes, reaching parity with domain-knowledge rules to within 0.002 every F1 on test set. Because the approach remains neurosymbolic, these geometric rules share a single logical framework and can still be complemented by domain knowledge when available. We frame the fusion of multiple imperfect ViT-based detectors as a consistency-based abduction problem solved at test time by an exact Integer Program (IP) and a polynomial-time heuristic. On an aerial-imagery benchmark of 15 weather-shifted test sets and six ViT detectors, our domain-knowledge-free layer matches the strongest majority-vote variant on clean data (within 0.005 F1) and, unlike every majority-vote baseline, retains its performance under a coordinated label-flipping attack: at a 90% flip rate it averages 0.42 F1 versus 0.35 for MV-Plurality (a 22% relative gain) and attains the highest F1 on \emphevery test set once the flip rate exceeds 0.4
[CV-104] riCLE: Tri-Modal Vision-Language Reasoning for Edge-Deployed Fine-Grained Clustering
链接: https://arxiv.org/abs/2608.04175
作者: Kishor Datta Gupta,Md. Mahfuzur Rahman,Fahad Rahman,Ahmed Rafi Hasan,Faysal Mehrab Chowdhury,Mohd Ariful Haque,Roy George
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Edge platforms used for aerial observation must interpret aircraft imagery under limited memory, limited compute, and intermittent connectivity. This setting is difficult for standard RGB-only recognition models and general-purpose vision-language models, especially when calibrated thermal and LiDAR aircraft data are unavailable. We present TriCLE, an application-oriented tri-modal vision-language system for aircraft taxonomic grouping under edge constraints. From a single RGB aircraft image, TriCLE generates a structure-preserving FLIR-style thermal view and a pseudo-LiDAR depth projection, then fuses the aligned views with task instructions in a compact Qwen3-VL backbone. The model is aligned to an expert aircraft taxonomy based on propulsion, airframe family, size, design era, and configuration, so its outputs reflect engineering-relevant similarity rather than only surface appearance. We evaluate supervised fine-tuning, rotation-preserving SFT, and three policy-alignment strategies: GRPO, GSPO, and DAPO. Sequence-level GSPO gives the strongest validation performance, reaching 88.33% validation accuracy and 0.91 weighted F1 on valid aircraft outputs. On a held-out aircraft test partition, GSPO achieves 78.00% accuracy and 0.793 weighted F1 while preserving 94.00% parseable output formatting. After 4-bit quantization and attention-memory optimization, the aligned 4B model fits an 8GB deployment target and processes each tri-modal triplet in 1.48 seconds. These results support TriCLE as a practical prototype for interpretable, edge-feasible aircraft grouping, while emphasizing the need for further validation on real aligned thermal and LiDAR sensor streams.
[CV-105] RNet: Topography-Guided Frequency Rectification and Structure-Aware Decoding for Multimodal Paddy Rice Segmentation
链接: https://arxiv.org/abs/2608.04154
作者: Kaiwen Xiao,Chunlong Fu,Liping Zheng,Yanfeng Su
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 16 pages, 9 figures, 6 tables
Abstract:Mapping paddy rice from very-high-resolution imagery in mountainous and hilly regions is difficult because terrain alters optical appearance and increases confusion with visually similar vegetation. We present TRNet for 0.5-m GaoJing-1 red–green–blue (RGB) imagery, a 5-m TanDEM-X digital elevation model (DEM), and derived slope. Separate visual and terrain encoders preserve modality-specific features. At an early encoder stage, Topographic Energy-Spectral Rectification applies terrain-conditioned low-frequency modulation and asymmetric high-frequency regulation to suppress steep-slope clutter and conditionally enhance compatible low-slope rice cues. The Topography-guided Paddy Structure Decoder combines semantic, rice–background boundary, and interior cues, using coarse terrain as context. Experiments used an Area A internal test set and held-out Area B, which had steeper terrain and lower rice prevalence. TRNet achieved rice intersection-over-union (IoU) values of 85.10% and 80.68%, exceeding the original Dual-Encoder U-Net by 9.15 and 18.83 percentage points, respectively. Ablation and slope-stratified results linked these gains to frequency rectification, structure learning, and fewer steep-terrain false positives. The results support coarse topography as a contextual prior for very-high-resolution paddy rice mapping.
[CV-106] LiNC: Lightweight Noise Correction via Per-Sample Trust and Gaussian Mixture Modeling
链接: https://arxiv.org/abs/2608.04147
作者: Abhishek Moturu,Babak Taati,Anna Goldenberg
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Label noise is common in medical imaging datasets due to factors such as inter-rater variability, annotation errors, and ambiguous cases. This can severely undermine the reliability and clinical effectiveness of machine learning models trained using those datasets. To address this challenge, we introduce Lightweight Noise Correction (LiNC), which adds a single trainable trust parameter per training sample and learns when to use the observed label and when to defer to the model during a standard training loop. The key idea is to train using a convex combination of the observed label and the model’s own predictive distribution, controlled by a per-sample trust parameter. We show that the gradient of this objective drives trust values in opposite directions for clean versus noisy samples in the early training phase, yielding separable trust distributions. We use a 3-component Gaussian Mixture Model over the trust values to separate them into clean, ambiguous, and noisy cases and then execute a short soft-correction phase on the noisy cases and a final hard correction phase. Experiments on ten 2D datasets from MedMNISTv2 under label noise of up to 50% show consistent gains in accuracy and strong mislabel detection. LiNC adds negligible asymptotic overhead: the training-time complexity remains dominated by the base network, with additional memory growing linearly with the size of the training set.
[CV-107] RUTA: Principled Visual Token Allocation via Rate-Utility Optimization
链接: https://arxiv.org/abs/2608.04132
作者: Jian Zou,Xiaoyu Xu,Zhihua Wang,Yilin Wang,Balu Adsumilli,Kede Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:High-resolution images and long videos provide vision-language models with rich context for multimodal reasoning and fine-grained perception, but the resulting long visual token sequences make large language model-side computation and memory costly. Existing visual token reducers often operate at prescribed rates, while recent methods adapt token counts across inputs using method-specific learned thresholds or importance predictors. We introduce RUTA, a principled Rate-Utility Token Allocation method that performs pre-LLM reduction by jointly learning which tokens to retain and how many to allocate to each image-query pair. RUTA constructs query-conditioned candidate tokens and predicts a retention probability for each candidate. During training, these probabilities parameterize independent Bernoulli gates, while their sum provides a differentiable training-time estimate of the token count for each pair. Retained tokens serve as anchors that aggregate information from non-retained tokens according to semantic affinity and spatial proximity. RUTA is optimized with a penalized rate-utility objective that balances downstream task loss against expected token usage. Averaged across five benchmarks and measured relative to each backbone’s full-token baseline, RUTA uses only 2.0% and 4.2% of visual tokens while preserving 88.2% and 94.4% of task performance on LLaVA-NeXT-7B and Qwen3-VL-8B, respectively.
[CV-108] Radar4D-VLM: Proposal-Grounded Temporal 4D Radar Reasoning Across Frozen Language Models
链接: https://arxiv.org/abs/2608.04130
作者: Jiaju Han,Xuemeng Sun,Qike Zhang,Xiang Chen,Luwei Yang,Jiahuan Long,Yiwei Wei,Jiujiang Guo,Chengyin Hu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-language models for autonomous driving primarily rely on cameras and LiDAR, leaving 4D radar largely unexplored as a standalone perceptual modality despite its robustness to adverse visibility and direct measurement of radial velocity. We introduce Radar4D-VLM, a radar-only temporal vision-language model that reasons from ten consecutive 4D-radar point-cloud sweeps without camera or LiDAR input. Radar4D-VLM extracts geometrically grounded object proposals and organizes radar evidence into a compact hierarchy of object, scene, and kinematic tokens. A parameter-efficient projector maps these tokens into frozen language backbones, while auditable prediction heads jointly model object count, spatial distribution, motion state, collision risk, semantic category, and radial velocity. Radar4D-VLM combines proposal-grounded temporal object tokenization, global scene context, and explicit kinematic tokens within a unified frozen-backbone interface. On sequence-isolated K-Radar development validation, its Top-64 proposal recall reaches 98.13% at 4 m, exceeding fixed-lattice and uniform-random controls by 6.40 and 22.83 percentage points, respectively. We further evaluate 24 matched runs spanning eight frozen Qwen, Phi, Mistral, Llama, and Gemma backbones under an identical adaptation budget. The radar-token interface remains compatible across all five language-model families, while matched aligned, permuted, and no-language controls show sensor dependence but no stable direct-head gain from aligned language supervision. These results establish a reproducible foundation for radar-only multimodal scene and motion reasoning while separating interface compatibility from the benefit of language supervision.
[CV-109] aching Foundation Models to Read mmWave: Pose-Guided Kinematic Representation for Human Behavior Understanding
链接: https://arxiv.org/abs/2608.04127
作者: Duo Zhang,Zhehui Yin,Zhiyun Yao,Haotong Qin,Xusheng Zhang,Hongliu Yang,Jianyu Sun,Junzhe Wang,Zizhou Fan,Michele Magno,Daqing Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 7 figures
Abstract:Large language model agents need to perceive human behavior in physical environments. Millimeter-wave (mmWave) radar provides a privacy-friendly and contactless sensing modality, but radar observations are difficult to align with language. Existing radar-language methods often rely on synthetic data or lack explicit supervision for human body structure and motion. We present mmMind, a radar-language model that uses synchronized 3D pose as training-only supervision. A spatio-temporal radar encoder is pretrained to capture body configuration and motion dynamics, after which the pose head is removed so that inference requires radar alone. The learned radar representations are then aligned with an LLM for behavior captioning and spatio-temporal question answering. We also introduce mmMind-Bench, a real-world mmWave-language benchmark containing 17.9 hours of recordings from 23 participants across seven indoor environments. Experiments on captioning, question answering, and unseen-action generalization show that mmMind consistently outperforms existing radar-language baselines, while ablations confirm the importance of pose-guided pretraining.
[CV-110] Perception Before Reasoning : Dynamic Latent Reasoning for Video Understanding and Question Answering
链接: https://arxiv.org/abs/2608.04124
作者: Haotian Xia,Zilin Xiao,Junbo Zou,Vicente Ordonez,Hanjie Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Video question answering requires models to ground language queries in visual evidence and, when necessary, reason over that evidence across time. Existing methods typically rely on long textual chain-of-thought rationales, even though many questions can be answered as soon as the relevant object, action, or frame is localized. We propose Dynamic Latent Reasoning (DyLaR), which first grounds a question in a short block of perception latents (continuous hidden states that encode query-relevant visual evidence), and then adaptively decides whether to append reasoning latents (continuous thoughts that reason over this evidence in latent space) before answering. DyLaR learns this behavior by grounding perception latents in verified visual evidence and distilling verified rationales into reasoning latents, followed by reinforcement learning that further refines when to reason. Across nine video benchmarks and four multimodal language model backbones, DyLaR improves average accuracy over same-backbone baselines while generating fewer than 20 tokens per query. On Qwen3-VL-4B, for example, DyLaR improves average accuracy over Qwen3-VL-4B-Thinking from 54.0 to 58.2 while reducing response length from 1,220.7 to 18.5 tokens per query. Ablations further show that grounded perception latents, rationale-supervised reasoning latents, and adaptive routing each improve accuracy.
[CV-111] Interpretable Fuzzy Inference for UAV Target Tracking Using Bounding-Box Geometry
链接: https://arxiv.org/abs/2608.04121
作者: Reza Ahmari,Ahmad Mohammadi,Vahid Hemmati,Nicholas Edmond,Hossein Z. Saghazadeh,Olusola Odeyomi,Parham Kebria,Abdollah Homaifar
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-based guidance of unmanned aerial vehicles (UAVs) toward unmanned ground vehicles (UGVs) supports cooperative aerial–ground robotics, but reliable continuous yaw estimation from onboard vision remains challenging because of sensing uncertainty, limited computation, and the need for interpretable control. Existing deep-learning and geometric-reconstruction approaches often require large datasets, external localization, or complex modeling assumptions, reducing transparency and deployment suitability on resource-constrained platforms. We present an interpretable fuzzy-inference framework that generates continuous yaw commands from low-dimensional features extracted from YOLO boxes: target centroid location, area, and aspect ratio. No explicit geometric modeling is required. A Mamdani fuzzy system serves as an interpretable baseline using a shoulder–triangle–shoulder input partition. It is followed by a first-order Takagi–Sugeno model with three antecedent membership terms per input, whose parameters are derived from training-set quantiles, yielding a compact 27-rule structure. Evaluation uses 6,169 labeled samples from a VICON motion-capture environment. Across five randomized train–test splits, the Takagi–Sugeno model achieves a test-set mean absolute error of 0.140^\circ \pm 0.003^\circ , a root mean squared error of 0.200^\circ \pm 0.008^\circ , and a maximum absolute error of 1.254^\circ \pm 0.121^\circ . Within-threshold accuracies are 99.676% \pm 0.270% for \pm1^\circ and 100.000% \pm 0.000% for both \pm3^\circ and \pm5^\circ . Directional consistency between image-plane horizontal displacement and predicted yaw sign reaches 90.254% \pm 0.612% . These results show that the framework is transparent, data-efficient, computationally lightweight, and suitable for real-time vision-based UAV guidance toward mobile ground targets.
[CV-112] LoRetta: A Foundation Model and Extensive Dataset for Global-Scale Remote Sensing Dense Image Matching
链接: https://arxiv.org/abs/2608.04106
作者: Siwei Yu,Han Guo,Zhenwei Shi,Zhengxia Zou
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注: 17 pages, 12 figures, 6 tables. Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence. Project page: this https URL
Abstract:Dense image matching establishes pixel-wise correspondences and underpins broad applications in computer vision and photogrammetry. However, extending dense matching to global-scale remote sensing remains challenging because image pairs may differ in acquisition time, season, viewpoint, spatial resolution, and land-cover state. The resulting large geometric offsets, partial overlap, and intrinsically unmatchable regions make direct dense correspondence prediction unreliable and inefficient. We thus reformulate dense matching as localization-and-registration: first localizing the matchable overlap and affine geometry, then refining dense residuals within the aligned frame. Based on this formulation, we propose LoRetta, a foundation model coupling matchability-aware affine localization with guided dense registration. We also introduce LEVIR-GM, a global-scale multi-temporal optical matching benchmark with dataset-native matchability labels (103K aligned, 827K augmented pairs, six continents, five years, 0.5-1024 m resolution). We further establish a unified evaluation protocol for sparse, semi-dense, and dense matchers. On LEVIR-GM, LoRetta achieves an area under the curve (AUC) of 83.3%, outperforming the strongest baseline RoMa v2 by 1.6 points, with larger percentage of correct keypoints (PCK) gains of 6.5 and 8.2 points at 1 and 2 pixels, while reducing inference latency by 47.8%. Astronaut-to-satellite and unmanned aerial vehicle (UAV)-to-satellite geolocalization experiments further demonstrate its transferability as a reusable geometric aligner.
[CV-113] Advancing Utility Pole and Sign Detection Through Deep Learning
链接: https://arxiv.org/abs/2608.04061
作者: Carl Dickinson,Gaetano Di Caterina
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Utility poles are an essential part of the infrastructure used to support power distribution systems and other critical public services. Their regular inspection is crucial to ensure the stability and safety of the electrical grid. A deep learning framework is presented for the automated detection, segmentation and lean angle estimation of wooden utility poles, and classification of attached electrical warning signs, using ground-level imagery. The system is trained on a custom dataset of 4,570 annotated images extracted from Google Street View, featuring challenging real-world scenes with visually ambiguous wooden poles lacking distinctive features. The proposed model is based on the Detection Transformer (DETR), suitably modified and trained on the custom dataset. The model outperforms standard object detectors (RetinaNet, Faster R-CNN, YOLOv3-Tiny), achieving a mean average precision of 90.43% for pole detection and 88.26% for sign detection. Extending this model with a segmentation head enables per-instance mask generation, which is then used to estimate pole lean angle. The model accurately estimates lean for 1,367 out of 1,433 test-set poles, with a mean absolute error of 1.01 degrees. Moreover, the custom dataset created in this work is also made publicly available to be used as a benchmark.
[CV-114] When Modalities Fail to Tango: Conformal Backdoor Detection in Multimodal Contrastive Learning
链接: https://arxiv.org/abs/2608.04052
作者: Yiming Chen,Kemou Li,Haiwei Wu,Jiantao Zhou
类目: Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Backdoor attacks in multimodal contrastive learning (MCL) have garnered growing attention in recent years, as many downstream tasks critically depend on pre-trained MCL models. Existing detection-based defenses predominantly rely on the CLIPScore metric, under the assumption that poisoned pairs exhibit lower semantic similarity between the image and the caption. However, we identify two critical flaws remaining in existing methods: (1) the substantial overlap between CLIPScore distributions of benign and poisoned pairs undermines the reliability of this metric, and (2) fixed-threshold detection cannot provide statistical guarantees for ambiguous samples within overlapping regions. To overcome these limitations, we propose integrating conformal prediction (CP), a statistical framework that quantifies uncertainty through nonconformity scores (NCSs), to establish provable confidence bounds for detecting poisoned image-caption pairs. Building on CP, we introduce CASCADE, a novel two-stage Coarse-to-Fine Conformal Backdoor Detection framework. The coarse-grained stage uses cross-modality consistency to identify high-confidence benign and poisoned pairs. In the fine-grained stage, a reference set is constructed from high-confidence poisoned pairs, and instance-level NCSs based on text-space similarity are computed for each sample in the unidentified subset. These NCSs measure conformity to the poisoning distribution and enable precise identification of latent poisoned pairs within the unidentified subset. Extensive experiments on the large-scale CC3M dataset demonstrate that CASCADE achieves an average FPR of 5.79% at 100% TPR and an average AUROC of 0.9867 across diverse attacks, while remaining effective against adaptive attacks.
[CV-115] NuclearDiffusion: Text-to-Image Foundation Models for Learning Nuclear Energy Concepts
链接: https://arxiv.org/abs/2608.04030
作者: Mohammed I. Radaideh,Jeremy Moon,Andre Gala-Garza,Emma Son,Yug Shah,Majdi I. Radaideh
类目: Graphics (cs.GR); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Computers and Society (cs.CY); Machine Learning (cs.LG)
备注: 29 pages, 10 figures, and 4 tables
Abstract:Generative artificial intelligence (AI) has transformed text-to-image synthesis, yet its ability to represent specialized engineering domains remains largely unexplored. As an exmaple in nuclear engineering, general-purpose foundation models frequently generate physically incorrect or conceptually inconsistent images because they lack domain-specific knowledge. This work presents one of the first systematic studies of domain adaptation for nuclear text-to-image generation through fine-tuning of open-source diffusion models. We curate a dataset of 1,000 captioned nuclear energy images spanning reactors, fuel cycles, radiation, and related concepts, and use it to fine-tune three state-of-the-art open-source models: Stable Diffusion XL (SDXL), SD-v3.5-Medium, and the flow-matching Flux.1 model. Their performance is evaluated using both quantitative image-similarity metrics and qualitative expert assessment against the corresponding zero-shot models. Fine-tuning substantially improves the fidelity of SDXL, provides only limited gains for SD-v3.5-Medium, and yields no measurable improvement for Flux.1, demonstrating that adaptation effectiveness depends strongly on the underlying generative architecture rather than model scale alone. We further compare the fine-tuned models against three leading commercial systems–GPT-Image-2, Gemini-3.1-Flash-Image, and Midjourney. Although GPT-Image-2 and Gemini generate convincing images for broad nuclear concepts, they frequently fail on specialized engineering prompts, where the fine-tuned open-source models produce more accurate and technically consistent outputs. These results establish domain-specific fine-tuning as a practical pathway for developing trustworthy generative AI tools for domain-specific applications.
[CV-116] MarsCast: Transfer Learning of AI Weather Foundation Models to Planetary Atmospheres
链接: https://arxiv.org/abs/2608.05054
作者: M.L. Carroll,J. Li,S.D. Guzewich,G.Villanueva,J.A. Caraballo-Vega,M.J. Frost
类目: Earth and Planetary Astrophysics (astro-ph.EP); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:We investigate the transferability of Earth weather foundation models to planetary atmospheres by adapting the GraphCast graph neural weather forecasting model to Mars. While GraphCast achieves state-of-the-art performance for terrestrial forecasting, its applicability to non-Earth environments remains unexplored. Using the Mars Climate Database (MCD), which provides global atmospheric fields across vertical altitude levels (similar to Earth pressure levels), we evaluate zero-shot and fine-tuned GraphCast predictions of Martian temperature and wind fields. Zero-shot forecasts produce a surprisingly accurate depiction of current conditions but fail to reproduce diurnal variability and rapidly decay toward climatological mean states. To address this limitation, we fine-tune GraphCast using MCD variables and top-of-atmosphere solar radiation forcing while holding humidity constant. Fine-tuning enables rapid learning of Martian thermal variability. Within as few as 10 training epochs, the model begins to capture the diurnal cycle and forecasts up to 10 days reproduce seasonal and vertical temperature structure. Prediction quality improves with training sample size and exhibits sensitivity to seasonal initialization. These results demonstrate that Earth-trained AI weather models can be adapted to simulate Martian atmospheric dynamics, providing a pathway toward rapid planetary weather prediction to support mission operations, dust storm risk mitigation, and future human exploration.
人工智能
[AI-0] Argus: A General-Purpose Agent ic Runtime for Long-Horizon Reasoning
链接: https://arxiv.org/abs/2608.05144
作者: Boxiu Li,Zimo Wen,Yijia Fan,Junxiang Lei,Sufeng Guo,Jiaao Wu,Ruize Tang,Mukai Li,Yifei Shen,Xiaoyu Chen,Wanbo Zhang,Runjing Gu,Yifei Gao,Yuheng Wu,Xuyao Huang,Zelong Zhao,Jiachen Zhang,Shibo Hu,Hangxi Guo,Yilin Chen,Yuzhe Zhang,Fan Yang,Chuan Wen,Xian Zhang,Xuanhe Zhou,Zhijie Deng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Long-horizon reasoning requires an agentic runtime that can persist when evidence supports its current approach and pivot when measurements reveal failure, hidden constraints, or a misspecified objective. We present Argus, a persistent, self-evolving runtime in which Manager, Planner, Engineer, and Reviewer execute bounded missions over durable project state. Argus separates stable user intent from operational objectives, constraints, and verification criteria, and admits memories, skills, procedures, verifiers, routing decisions, and rejected routes only after role-owned review and, when available, task-native verification. Model weights remain fixed; self-evolution occurs through persistent runtime state and control policy, with autonomous execution between operator-owned escalation points. Across seven GPT-5.5 benchmark arenas, Argus achieves about 78% on SWE-Bench Pro versus 59% for Direct Copilot while using 1.41 times the aggregate tokens. After verification-gated self-evolution, mature SWE-Bench waves use 21% fewer solve-input tokens and 15% less active workflow time per task than startup waves, while recording 34 verifier recoveries and 22 strict review-loop rescues. Argus also reaches 76.8% on AARRI-Bench and a 28.0-point gap on mathematical data synthesis, with competitive GPU-kernel and language-model-training results. Beyond benchmarks, an optimized RWKV6 kernel was merged upstream; a multi-day mathematics campaign retained falsified routes and proof-backed frontier updates; and six paper pipelines completed 254 missions with 16 stage rollbacks. These results show that a fixed-weight, self-evolving harness can revise, recover, and accumulate verified approaches while producing structured trajectories for future supervised and reinforcement learning.
[AI-1] OctoLong: Mid-Training On Cross-Repository Code Contexts Enhances Long-Context Modeling
链接: https://arxiv.org/abs/2608.05141
作者: Indraneil Paul,Falko Helm,Goran Glavaš,Iryna Gurevych
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Software Engineering (cs.SE)
备注:
Abstract:Context lengths of language models (LMs) have dramatically increased, driven by the demands for in-context learning, self-improvement, and long-horizon agentic workflows. Existing long-context corpora, however, are dominated by books, academic articles, and code repositories, which are finite resources and often scarce in long-distance dependencies. In this work, we introduce OctoLong, a context engineering pipeline that instruments an AST parser, a language server backend, and a package manager to facilitate the recursive retrieval of code references, enabling the curation of dependency-rich code contexts of millions of tokens in length. We then train OctoLong-Instruct, a suite of capable long-context open LMs, derived from base models ranging in size from 600M to 14B parameters, via context-extension mid-training on a ~50B-token mixture containing ~6.2B tokens of OctoLong code contexts, followed by ~10B tokens of instruction tuning. Our training ablations and experimental evaluations against 18 state-of-the-art open-weight long-context LMs show that supplanting just 12% of traditional context-extension corpora with OctoLong data yields substantial gains in long-range retrieval, long-term state tracking, repository-level code understanding, and downstream agentic tasks, while also enhancing API usage in short-context coding scenarios.
[AI-2] SSTQ:Privacy-Preserving Vector Quantization via Subsampled Stochastic TurboQuant
链接: https://arxiv.org/abs/2608.05127
作者: Adel Javanmard,David P. Woodruff,Vahab Mirrokni
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注: 42 pages, 4 figures, 2 tables
Abstract:Achieving local differential privacy in distributed optimization while maintaining low communication cost remains challenging. Existing vector quantization methods, such as vqSGD, use high-dimensional geometric constructions but incur unfavorable dimension-dependent variance. In this work, we propose Subsampled Stochastic TurboQuant (SSTQ), a framework that combines overcomplete equal-norm tight frames, coordinate subsampling, and privacy-aware one-dimensional quantization. SSTQ includes two variants: a Flat Randomized Response version and a Metric-Aware Laplace version, the latter being better suited to higher codebook bit-width regimes. We show that SSTQ achieves optimal mean squared error scaling while using only \lceil \log_2 N \rceil + b bits per client, where N = \Theta(d) is the frame size. We also derive a surrogate privacy-aware codebook objective that reduces the codebook-dependent MSE scaling from O(4^b) to O(2^b) . Finally, we empirically evaluate SSTQ against established baselines on federated learning tasks using CIFAR-10 and Fashion-MNIST, demonstrating favorable utility and communication efficiency.
[AI-3] ABSeeker: Training Long-Horizon Search Agents via Answer-Backtracked Credit Assignment
链接: https://arxiv.org/abs/2608.05102
作者: Yijun Lu,Rui Ye,Jiajun Wang,Yuwen Du,Tian Jin,Songhua Liu,Siheng Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Long-horizon search agents must make multiple sequential actions (steps) to search, retrieve, verify, and integrate evidence to reach a final answer. However, existing methods for training these agents typically treat all steps within a trajectory uniformly during both supervised fine-tuning (SFT) and reinforcement learning (RL), failing to distinguish useful actions from erroneous or redundant ones. In this paper, we propose Answer-Backtracked Credit Assignment (ABC), a fine-grained credit assignment framework for training long-horizon search agents by converting sparse trajectory-level outcomes into dense step-level supervision that rewards useful actions (even in failed trajectories) while suppressing erroneous or redundant actions. Specifically, given a potentially obscure query and its corresponding ground-truth answer, ABC first performs Answer-Backtracked Clue Recovery, which traces back from the answer to recover intermediate clues required to solve the question. It then applies Clue-Anchored Step Scoring to evaluate each search step against these clues, converting sparse binary outcome supervision into dense step-level rewards. Based on these rewards, we develop ABC-SFT, which reweights the loss of each turn, and ABC-GRPO, which uses the step-level scores as rewards in GRPO. Building on this framework, we train ABSeeker based on Qwen3.5-4B with only 8.5k examples. ABSeeker achieves 37.3% on BrowseComp and 39.1% on BrowseComp-ZH. With context management, the scores further improve to 55.3% and 52.9%, respectively, significantly outperforming same-scale (4B) agents and even matching the performance of larger ones (approximately 30B). These results demonstrate the effectiveness of answer-backtracked step-level credit assignment for training long-horizon search agents.
[AI-4] Hierarchical Graph Memory for LLM Agents with Path-level Localization and Rewrite
链接: https://arxiv.org/abs/2608.05095
作者: Xiawei Yue,Boran Wang,Xiaoqing Zhang,Shuxin Zheng,Ziwei Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Agents for long term reasoning require a memory that can be efficiently and effectively updated over time, as new facts and external feedback continue to arrive. Recently, graph memory has been adopted to offer structural organization for multi-hop retrieval and reasoning. However, existing methods store all memories in a flat graph, and accumulated historical memories can introduce irrelevant contexts and increase the cost of evidence selection during retrieval. Moreover, they typically update memory units independently, requiring repeated unit-wise rewrite to cover related changes. To address these issues, we propose HiGram, an evolving hierarchical graph memory framework with path-level localization and rewriting. Specifically, we first propose a hierarchical graph memory, which organizes the memory into coarse-to-fine architecture composed of upper-level nodes and MemoryUnits, thereby reducing the amount of irrelevant information during retrieval. We further propose MicroGraph-based path-level localization, which leverages query and update conditioned MicroGraphs to identify support subgraph and evidence path before rewrite. Finally, we propose a coordinated rewriting method that jointly revises intra-unit memory and inter-unit dependencies, enable valid dependency structures updating in the localized evidence path. Experiments on benchmarks for long-term conversational question answering and conflict-aware memory evaluation demonstrate that our method demonstrate substantial improvements over baselines in answer quality and token efficiency. Besides, our method improves answer accuracy and query-valid evidence selection under dynamic, static, and conditional conflicts.
[AI-5] Capability-Gated Planning : Cost-to-Goal Discovery and the Limits of Myopic Experiment Selection
链接: https://arxiv.org/abs/2608.05085
作者: Ahmed Hassoon,Mark Dredze
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Systems that automate scientific discovery must repeatedly decide which experiment to run, which hypothesis to test, which tool to build, and when to stop. Many systems make these decisions by maximizing a myopic score such as expected information gain per unit cost or a learned plausibility score. We identify a structural limitation of this approach. Some actions are constructive: they acquire an epistemic capability (an instrument, assay, pipeline, simulator, or abstraction) whose value lies not in the information returned immediately but in the future actions it makes available. When the least-cost route to a confident answer requires a chain of such constructions, a planner that scores actions only by information obtainable within a bounded horizon cannot value the first construction: it yields no information within the horizon and is dominated by any measurement with positive information, however small. We formulate goal-directed discovery as a stochastic shortest-path problem in belief space in which constructive experiments change the downstream action graph, and prove that for every lookahead depth d there is an instance on which every myopic information-maximizing planner has an unbounded approximation ratio, and a related instance on which it never reaches the goal. The mechanism is a capability-indistinguishability lemma: within the horizon, acquiring a capability can be observationally indistinguishable from paying for a null action. This establishes capability gating as a reachability axis of difficulty distinct from curvature (submodularity) and information order (adaptivity gaps). We introduce CG-Plan, an incremental replanner with a capability-aware cost-to-go heuristic h = h_cap + h_exp. In a controlled testbed, the performance gap appears only under gating, persists for every fixed horizon, and arises when near-miss hypotheses come from a data-consistent proposer.
[AI-6] MultiPathFormer: Towards a Foundation Model for Multipath Wireless Propagation
链接: https://arxiv.org/abs/2608.05076
作者: Blessed Guda,Kayley Sze,Carlee Joe-Wong
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Signal Processing (eess.SP)
备注:
Abstract:Recent advances in machine learning have enabled training of wireless foundation models, which aim to support tasks such as channel estimation, beam prediction, and localization based on wireless signals. Existing wireless foundation models typically pretrain on channel tensors using masked reconstruction over subcarriers, antennas, or time but ignore the physical characteristics of wireless propagation. In this work, we propose to instead use multipath propagation as the fundamental pretraining object. We present MultiPathFormer, an autoregressive foundation model that represents each transmitter-receiver link as an ordered sequence of continuous-valued path tokens and pretrains with next-path prediction. We introduce an Environmental RAG (retrieval-augmented generation) mechanism and a first-path codebook on top of the transformer backbone, leveraging environment knowledge to improve path statistics estimation like delay and power by up to 59%. MultiPathFormer pretrained on 27 environments transfers to unseen users and, after scenario-specific fine-tuning, outperforms training the corresponding models from scratch in new environments. Across downstream tasks, it outperforms SOTA channel-based foundation models, achieving 5.57 m mean localization error, 0.914 top-3 beam accuracy, 0.994 line-of-sight classification accuracy, and 0.561 channel estimation NMSE. These results show that path-level pretraining can learn reusable representations of wireless propagation.
[AI-7] Hardware Design and Security in the Era of Chiplets and LLM s
链接: https://arxiv.org/abs/2608.05063
作者: Johann Knechtel,Ozgur Sinanoglu,Paul V. Gratz,Ramesh Karri
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
备注:
Abstract:The semiconductor industry is undergoing a dual revolution: the shift toward heterogeneous 2.5D chiplet systems and the integration of Large Language Models (LLMs) into Electronic Design Automation (EDA) flows. While these paradigms offer unprecedented benefits in yield, modularity, design productivity, etc., they radically expand the hardware attack surface. This paper provides a unified analysis of these frontiers, ranging from attacks on chiplet systems (including hardware stacks for LLM acceleration) across architectural, logical, and physical levels, to various exploits against LLM-driven EDA pipelines. To secure chiplet systems, we review a powerful defense approach that leverages 2.5D split manufacturing and active interposers for physically isolated Root of Trust (RoT) architectures. To secure LLM-driven EDA pipelines, we first identify native threats and then review state-of-the-art defense techniques. Finally, we discuss how LLM systems can advance hardware security efforts for modern systems, including chiplets.
[AI-8] RepairFormer: Automated Repair of Structured Inputs Using Transformers
链接: https://arxiv.org/abs/2608.05060
作者: Ovi Paul,Tom J King,Ali Shokri
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 5 pages, 2 figures, and 3 tables
Abstract:Structured input files such as JSON, DOT, OBJ, INI, S-expression, and TinyC are widely used in software systems, but small corruptions can cause parsers to reject otherwise useful data. Repairing such inputs is important because malformed configuration, program, and data files can interrupt testing, analysis, deployment, and downstream automation even when most of the original content remains intact. Existing repair techniques can produce structurally valid inputs, but they often rely on deletion or repeated search, which may lose original content and result in semantic incorrectness. This paper presents RepairFormer, a transformer-based framework for structured input repair. The approach formulates repair as a supervised sequence generation task and uses format tags, oracle validation, and boundary-localized repair to generate valid outputs while preserving content. The boundary workflow focuses generation on the detected fault region, reducing the input size, and supporting repair of longer files. In evaluation, RepairFormer achieves a 88% in repair and 94% in recovery, showing strongest content preservation when repairs are successful. Additional experiments on our benchmark shows RepairFormer repairs 97.57% and recovers 94.29% with 5x faster runtime compared to state of the art.
[AI-9] From Score Matrices to Football-Aware Match-State Simulation: An Auditable LLM Harness for Exact-Score Reranking ATC
链接: https://arxiv.org/abs/2608.05030
作者: Shaopeng Liang
类目: Artificial Intelligence (cs.AI)
备注: 9 pages, 1 figure, 5 tables. Interim chronological benchmark on the first 150 matches of the 2025-26 English Premier League
Abstract:Football score forecasting combines a strong statistical core with a difficult contextual edge. Dynamic Poisson-family models estimate team strength, expected goals, and coherent score probabilities, but do not directly understand roles, tactical matchups, motivation, or how a first goal changes behaviour. Large language models (LLMs) can reason about such concepts, yet are not calibrated probability engines. We combine both components through an auditable information harness. This paper documents four iterations: V1, a dynamic score-driven Dixon-Coles baseline; V2, which maps LLM contextual ratings back into expected-goal parameters; V3, which replaces scalar correction with goal-by-goal simulations over a frozen score-candidate set; and V4, which adds shared first-breakthrough and post-goal cascade judgments, time-aware stopping, and deterministic tail candidates. The harness defines input semantics, supplies pre-match evidence, and constrains the LLM to an inspectable reasoning route. On a chronological replay of the first 150 matches of the 2025-26 English Premier League, V1 achieved 10.0% Top-1 and 26.7% Top-3 exact-score accuracy. V3 reached 12.0% and 30.0%, while V4 reached 14.7% and 30.7%. V4 increased candidate coverage from 77.3% to 84.7%, although no added tail candidate became a Top-3 exact hit. V1’s native 1X2 distribution achieved 53.3% argmax accuracy, 0.9878 log loss, 0.5870 Brier score, and 0.2095 ranked probability score. These results are exploratory: the development slice is not an untouched benchmark, and temporal input isolation cannot exclude outcome memory in a closed LLM. The contribution is an auditable hybrid architecture, a clear design evolution, and negative findings showing where football-aware simulation does and does not improve score selection.
[AI-10] Short-term load forecasting under EU-AI Act Requirements in Safety-Critical Environments: Results from a 41-day live challenge on the aggregated German transmission-grid load
链接: https://arxiv.org/abs/2608.05018
作者: Thomas Bartz-Beielstein
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Short-term load forecasting (STLF) play a vital role in the electric power industry. It serves infrastructure that European and German law designate as critical. Determinism, reproducibility, and auditability are engineering requirements rather than optional extras. STLF is no longer purely an accuracy problem. It is also a software-engineering and compliance problem. This paper describes results from a 41-day live challenge that evaluated a complete STLF pipeline for the aggregated German transmission-grid load. The pipeline is based on the open-source Python library spotforecast2-safe, which implements the EU-AI Act Requirements in Safety-Critical Environments by design. The pipeline predicts the 24 hourly load values of a target day from European Network of Transmission System Operators for Electricity (ENTSO-E) data. It includes anomaly detection and gap-aware data preparation, calendar and weather covariates, a recursive multi-step forecasting algorithm, and hyperparameter tuning. Forecast accuracy is measured against the official ENTSO-E day-ahead forecast. The EU-AI act compliant spotforecast2-safe pipeline beats the ENTSO-E baseline. In-context models show competitive performance. Transparent, low-cost, and auditable local models (referred to as macl2l in this paper) are competitive with more than 100-million-parameter large, energy-intensive pre-trained foundation models such as chronos-2. The challenge infrastructure, the complete submission history of all teams, and the frozen final leaderboard are publicly available.
[AI-11] ORACLE: A Multi-Objective Reinforcement Learning-Based Analog Circuit Design Optimizer with Large Language Models -Guided Exploration
链接: https://arxiv.org/abs/2608.04999
作者: Osei Brempong,Mohammed Ayman Habib,Vivan Poddar,Morteza Fayazi
类目: ystems and Control (eess.SY); Artificial Intelligence (cs.AI)
备注:
Abstract:Analog circuit design automation using reinforcement learning (RL) has emerged as a promising approach for reducing manual effort. However, many existing RL-based methods focus on single-objective optimization. Even methods designed for multi-objective (MO) problems often reduce multiple design specifications to a single scalar reward. This simplification limits the ability to capture the true Pareto trade-off among competing objectives and often leads to suboptimal designs. Moreover, requiring the model to be retrained from scratch whenever the desired MO specifications change remains a key limitation. To address these challenges, we present ORACLE, an open-source RL-based framework for MO analog circuit design optimization that replaces scalar reward optimization with vector-valued learning and preference-aware conditioning. ORACLE represents a true MO analog circuit design optimizer that uses a preference vector to specify the relative weights of multiple objectives, enabling a single trained model to generate designs across diverse trade-off settings without retraining. We further propose two preference-guidance strategies, namely normalized-weight guidance and cosine-aligned guidance, to improve convergence. In addition, we incorporate a large language model (LLM)-guided action selection mechanism to filter actions that are likely to lead to suboptimal designs or increased runtime. Our results show that, on multiple circuit topologies with 2,000 test cases, ORACLE reduces runtime by 20.4x - 104.4x compared to state-of-the-art approaches. It also meets 99.9% of the 2,000 target specifications, and achieves 5.1x - 318.6x better figure of merit in the resulting output specs.
[AI-12] SciCode-Verified: How Benchmark Defects Underestimated the Scientific-Coding Ability of Language Models
链接: https://arxiv.org/abs/2608.04975
作者: Sihan Hu,Lyuhan Huang,Youjin Deng,Kun Chen
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 47 pages, 2 figures, 6 tables. Project repository: this https URL
Abstract:SciCode is the standard measure of the scientific-coding ability of language models: research-level problems that demand both frontier scientific theory and its implementation as working numerical code. It is a component of the Artificial Analysis Intelligence Index and a standing evaluation in government and national-laboratory suites. Yet its scores have recently plateaued: the strongest 2026 models cluster tightly around 60% subproblem accuracy, and a successor model ties its predecessor. We trace this stagnation to defects in the benchmark itself. A per-problem, domain-expert audit of all 65 test problems uncovers 263 defects; 192 of them, spread across 91% of the main problems, cause correct, instruction-following solutions to be wrongly rejected—through non-reproducible gold answers, over-tight tolerances, or self-contradictory specifications. Critically, 78% of these score-suppressing defects require specialized physics or mathematics knowledge to detect, not mere clerical proofreading. We corrected every confirmable defect to produce SciCode-Verified. The corrections add only the specifications a well-posed problem requires, repair grading, and tighten the tests that were too lenient; every change is recorded with its justification and independently re-checked by a second domain expert. We re-evaluate twelve frontier model snapshots on the corrected benchmark and find a substantial recovery: subproblem accuracy rises from 45–60% to 84–98%, and main-problem accuracy from 9–27% to 69–92%. State-of-the-art models are far more proficient in scientific coding than SciCode has suggested—the bottleneck was not model capability, but the quality of the evaluation instrument. We release SciCode-Verified with its complete audit trail as the corrected public standard.
[AI-13] WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World Models
链接: https://arxiv.org/abs/2608.04964
作者: Bohai Gu,Yueyang Yuan,Taiyi Wu,Dazhao Du,Jian Liu,Xiaoyi Pang,Jie Zhang,Xiaocheng Lu,Haobin Zhong,Xiaotong Zhao,Alan Zhao,Song Guo
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: this https URL
Abstract:Interactive video world models are essential for long-horizon planning and exploration, yet they suffer from compounding errors. Post-training methods such as reinforcement learning (RL) can improve these models, but they hit a verification bottleneck: for arbitrary action sequences, no ground-truth future state exists to measure long-term drift. Our key insight is that reversible action cycles make this verification possible: a sequence composed with its inverse must analytically return to the initial state, yielding annotation-free supervision on long-horizon correctness. Building on this, we introduce WorldCycle, a self-verifiable RL framework that constructs closed action cycles and their repeated executions from ordinary action sequences, and optimizes two complementary rewards: a spatial closure reward enforcing symmetry between mirrored forward and reverse segments, and a temporal consistency reward aligning states across repeated cycle executions. These rewards force the model to learn actions as consistent state operators rather than memorized temporal patterns, and extend naturally to out-of-distribution composite action cycles that the base model handles poorly. We further release CycleBench, a diagnostic benchmark for state-returning ability under complex action structures. WorldCycle reduces state returning drift by up to 44% and lifts composite-action accuracy nearly 4x over the base model, providing a vital foundation for physically grounded world models.
[AI-14] A General Sufficient Condition for Rewriting Horn-ALCHI Atomic Queries into GQL ISWC2026
链接: https://arxiv.org/abs/2608.04945
作者: David Carral,Calixte Gruson,Quentin Manière
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注: 27 pages. Technical report of a paper to appear at ISWC 2026
Abstract:The emergence of the ISO standard GQL introduces a powerful query language extending first-order logic with controlled recursion, raising the question of its applicability to evaluation of ontology-mediated queries (OMQs). We focus on OMQs consisting of atomic queries over ontologies expressed in Horn-ALCHI, an expressive Description Logic that is not, in general, first-order rewritable. To address this, we introduce DL automata, a novel formalism that captures the semantics of such OMQs via runs over fact sets. We then identify a large class of DL automata that can be rewritten into unions of conjunctive two-way regular path queries (UC2RPQs), a central fragment of GQL. Our class of automata relies on a stratification of their states, ruling out specific forms of cyclic dependencies known to raise the complexity. This yields a broad class of Horn-ALCHI OMQs that are GQL-rewritable.
[AI-15] CheMLFlow: An Open-Source Platform for Cheminformatics and Materials Informatics Applications
链接: https://arxiv.org/abs/2608.04942
作者: Brendan Smith,Susana Lopez-Moreno,Eric Dolores-Cuenca,Sangil Kim,Jose L. Mendoza-Cortes,Nijamudheen Abdulrahiman
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci); Other Condensed Matter (cond-mat.other); Artificial Intelligence (cs.AI); Chemical Physics (physics.chem-ph)
备注:
Abstract:CheMLFlow is an open-source platform for building and executing end-to-end, high-throughput, and agentic workflows for scientific and technological applications. CheMLFlow targets a common bottleneck in scientific machine learning development, where researchers often need to assemble data acquisition, curation, representation, model training, validation, screening, interpretation, and reporting into a reproducible pipeline, even when their primary research contribution concerns only one stage. CheMLFlow provides modular workflow components, ready-to-run reference pipelines, standardized artifacts, and evaluation outputs that reduce orchestration overhead and support benchmarking across methods and datasets. The platform is designed to be extensible, reproducible, and automation friendly, with pluggable representations and models, deterministic splits, explicit run artifacts, batch execution, and report generation. As scientific software increasingly moves toward agent assisted experimentation, CheMLFlow’s configuration driven workflows and structured outputs also provide a practical interface for coding agents to help users construct experiments, inspect results, and summarize findings under human supervision. This article describes the system architecture, core workflows, and benchmarks that reach literature performance for quantum mechanical, physicochemical and bioactivity property prediction, and use cases involving time series datasets demonstrating applications beyond molecular chemistry datasets.
[AI-16] SVI-DAG: A Structured Variational Inference Approach to Bayesian Causal Discovery
链接: https://arxiv.org/abs/2608.04930
作者: Shrenik Zinage
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Bayesian causal discovery seeks to determine the posterior distribution of causal theories, which are interpreted as directed acyclic graphs (DAGs) that explain the observed data. The resulting posterior allows systematic reasoning regarding epistemic uncertainty within these theories. Nonetheless, finding such graphs is difficult due to identifiability problems and limited observational data. Furthermore, precisely approximating posterior over graphs is challenging given vast range of potential DAGs. Recent Bayesian approaches have addressed some of these challenges, yet they remain limited as they fail to encode dependencies between edges, and lack principled ways to incorporate domain knowledge as inductive biases during the search process. To overcome these limitations, we propose SVI-DAG, a structured variational inference approach to Bayesian causal discovery using observational data and prior beliefs that uses normalizing flows to model dependencies between edges, supporting expressive and multimodal posterior learning over DAGs. To mitigate mode seeking behaviour in evidence lower bound optimization and promote mode coverage, we use stein variational gradient descent to update the node potentials using a kernel in acyclicity space. We evaluate SVI-DAG against 5 state-of-the-art Bayesian DAG learning methods and demonstrate superior performance in uncertainty quantification while remaining competitive in terms of structural accuracy.
[AI-17] A Chain Is Only as Strong as Its Weakest Link: A Scoping Review of System Integration Audits in AI
链接: https://arxiv.org/abs/2608.04921
作者: Leah Davis,Dominic Martin,AJung Moon
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:As AI systems become increasingly integrated into diverse interfaces and applications, model-centric audits are insufficient to address risks arising from interactions among system components and deployment environments. System integration has long been central to software audits in safety-critical domains such as aerospace. However, its role in AI auditing remains underexplored. Scanning through 4,259 documents, we present a scoping review of AI audits that treat system integration as a core tenet of evaluation (n = 58). Using reflexive thematic analysis, we analyze their elements, actors, enablers, and constraints. We find that the corpus represents an emerging yet still fragmented form of AI auditing: few existing measures target integration-specific risks; large gaps remain in meeting traditional audit expectations; and access to necessary information and resources significantly influences audit design. Nonetheless, integration can be categorized across three sites (inter-component, system-environment, and multi-system), each serving the functions of risk exploration, risk determination, coordination, and procedural regularity. Deviating from other types of evaluations, these audits assess qualities specific to system integration, including compatibility, completeness, and oversight. This review calls on the AI community to prioritize system integration as a core strategy for addressing AI risk, and to develop audit practices capable of capturing failures across components, environments, and systems beyond the reach of component-level evaluation.
[AI-18] When Does Latent Communication Pay? A Causal Audit of Relayed KV Caches in Multi-Agent LLM s
链接: https://arxiv.org/abs/2608.04893
作者: Jiaming Cheng,Subhransu Das,Rajiv Ramnath
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Multi-agent LLM systems relay key–value caches instead of text and credit their gains to exchanged ``latent thoughts’'. That credit is a claim about \emphwhich example’s cache is relayed, not merely that one is. We audit it causally in released systems. The cache is replaced with deranged (mismatched-example), zeroed, and moment-matched random counterparts, under two regimes defined by whether the receiver needs the sender’s private information. Where it does, the battery reads ceiling: 100% against 23–25% for answer-irrelevant relays on the primary backbone, a contrast replicated across three families, five checkpoints, and a prose document-QA surface. Where it does not, a pre-registered five-seed protocol establishes equivalence within 2.8 points, a margin anchored to the audited system’s reported gain, under Holm-corrected TOST on GSM8K and ARC-Challenge across three Qwen3 scales and on MedQA at 8B (one cell shows a small detected advantage inside the margin); a second family shows no detected advantage. A large cache effect need not be a pairing effect. In one natural cell, zeroing the relay costs 14.7 points; a mismatched cache, 0.4. Nor is need sufficient: under the same test, delivered channels span ceiling (LatentMAS’s native relay), partial (KVComm’s layer subset), and no detected example-specific transfer (C2C’s released projector). Benchmark deltas do not by themselves establish latent-thought transmission; establishing it takes a mismatched-cache audit, which we release.
[AI-19] ContextWeave: A Real-World Workflow Benchmark
链接: https://arxiv.org/abs/2608.04830
作者: Bo Wang,Yuqian Yao,Enxi Wang,Luozhijie Jin,Yang Liu,Yiran Suo,Yuxuan Cai,Enyu Zhou,Yufei Gao,Honglin Guo,Tianyu Huai,Li Ji,Zhikai Lei,Bufan Li,Lizhi Lin,Jinxiu Liu,Jie Yang,Jiazheng Zhou,Maosen Zhou,Pengfang Qian,Shichun Liu,Guanshan Liu,Hao Zheng,Yunhao Yu,Hang Yan,Jihua Kang,Xinchi Chen,Xipeng Qiu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Memory is essential as language agents move from isolated tasks to long-horizon, stateful workflows, yet existing evaluations often reduce it to retrieval or question answering. We introduce ContextWeave, a longitudinal benchmark that evaluates whether recalled experience improves downstream agent performance in realistic office-work streams. ContextWeave reconstructs privacy-preserved, multi-month workflows of 14 participants into 1,005 executable tasks, including 568 core evaluation tasks, with instructions, containerized environments, trajectories, and task-specific rubrics. It measures workspace quality and alignment with participant-specific preferences, complemented by diagnostics of relevance, continuity, solvability, and robustness to misleading recall. Across six memory components under a fixed model, the strongest configuration raises Workspace Score from 68.08 to 78.20 and Preference Score from 41.50 to 70.60. With a fixed memory component, recall improves both outcomes for all five tested base models, although gains vary substantially. Our analysis shows that actionable, experience-rich memory supports workflow continuation and reduces redundant exploration more effectively than compact summaries, while it can also be more susceptible to misleading recall. These findings motivate memory systems that optimize not only retrieval relevance but also reliable use during execution.
[AI-20] Scrouting: Cost-Aware Routing of Coding Agents by Scouting the Repository First
链接: https://arxiv.org/abs/2608.04804
作者: Ishaan Bhola,Adithyan Krishnan,Mukunda NS
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Frontier language models can resolve repository-level software issues, but each attempt is expensive, and existing routers select a model from the issue text alone. We present SuperScout, which routes after scouting the repository: a 7B searcher, SuperScout-7B, first explores the repository and produces a structured handoff whose reproduction claims are sandbox-verified, with false claims stripped before delivery. The searcher’s hidden states, together with the task text, then feed a resume-based router that dispatches the task to one of four frontier fixers. Adding a new fixer requires no retraining. On the full Python slice of SWE-bench Pro (266 tasks) under the benchmark’s official capped budget tier, SuperScout matches the best single model’s solve rate (159 of 266 for SuperScout, 158 for the best model) at about a fifth of the total cost per solve, and the reported configuration sits above the random traffic-splitting baseline. A no-router ablation, always the cheapest fixer with the handoff, ties the routed system on this benchmark, so the handoff rather than the routing decision carries the result. A paired calibration study points to the mechanism: the handoff appears to redistribute rather than add solving ability, lifting the three cheaper fixers while slightly hurting the strongest, though at N=99 the per-fixer effects are directional only; the searcher’s hidden states improve cost routing on the calibration labels while the handoff’s own text does not. The searcher’s compute adds less than half a cent of GPU time per task.
[AI-21] Privileged but Biased: How PI-Conditioned Teachers Break Self-Distillation
链接: https://arxiv.org/abs/2608.04794
作者: Sarthak Harne,Chinmay Karkar,Yash Pandya,Ahmed Awadallah,Akshay Nambi
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Self-distillation (SD) has emerged as a compute-efficient alternative to reinforcement learning with verifiable rewards: a self-teacher, conditioned on privileged information (PI) about the answer such as a reference solution, supplies dense per-token supervision to a student that never sees it. Reported gains, however, come almost exclusively from narrow, low-difficulty settings, leaving open a basic question: as a lone objective, with no reward term, does SD teach anything? We reproduce SDPO’s reported gains in its easy setting, then apply the identical setup to difficult tasks and find that it does not. Across question answering, mathematics, coding, and multi-turn agentic tool use, across reasoning modes, model sizes, and forms of PI, and under both the SDPO and OPSD recipes, the per-token loss falls steadily while validation accuracy does not improve and typically degrades. We explain this failure through a single causal chain from the loss to the model it produces. The chain begins with PI bias: having seen one particular reference solution, the teacher’s per-token target is pulled toward that trajectory rather than toward correctness in general, an effect we quantify with a PI Bias Score. Trained to match this target everywhere, the student’s objective becomes nearly blind to whether a rollout is correct, and the loss it assigns falls mostly on low-information tokens like stopwords, punctuation, uncertainty markers, rather than those that determine the answer; within correct rollouts the exploratory tokens incur the highest divergence, so it penalizes the hesitation that reasoning requires. The result is a flatter, less decisive student that is no better at reasoning: as a lone objective, SD optimizes a signal decoupled from task success.
[AI-22] RepoProbe: Benchmarking Architecture-Aware Repository Comprehension with Checklists
链接: https://arxiv.org/abs/2608.04783
作者: Yuexi Yang,Alyssa Wu,Ji Luo,Richeng Xuan,Zhichao Hu,Yuhong Liu,Zhen Qin
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:The integration of Large Language Models (LLMs) into software engineering has shifted the focus from function-level generation to repository-scale assistance. However, existing benchmarks largely rely on bug reports from GitHub Issues, which often allow models to bypass genuine understanding via pattern matching on error logs. This misalignment under-measures Edit Bias, which refers to premature generation, where models prematurely propose code modifications instead of understanding the existing repository architecture. Furthermore, current LLM-as-a-Judge scalar scoring suffers from high variance and low interpretability. This work introduces RepoProbe, a novel benchmark for evaluating repository-level code understanding through open-ended QA using GitHub Discussions, which focuses on open-ended architectural inquiries rather than defect reporting. To ensure rigorous evaluation, we propose a Checklist-Based Verification Protocol that decomposes answers into atomic, verifiable facts, thereby replacing subjective ratings with objective verification. Our evaluation of state-of-the-art (SOTA) LLMs reveals a persistent gap between high clarity and evidencegrounded technical correctness. It also quantitatively confirms the prevalence of edit bias, in which models prioritize code generation instead of architectural analysis. Finally, we demonstrate that our verification protocol significantly improves evaluation reliability compared to traditional evaluations with scalar scoring.
[AI-23] IMFACT: Counterfactual Explanations for Time Series via Intrinsic Mode Function Substitution ECML-PKDD KDD
链接: https://arxiv.org/abs/2608.04777
作者: Udo Schlegel,Julian Rakuschek,Thomas Seidl,Andreas Holzinger,Tobias Schreck,Javier Del Ser
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 16 pages, 2 figure, 2 tables, accepted at XKDD Workshop at ECML-PKDD
Abstract:Oscillatory signals, such as vibration, carry class-discriminative information in specific frequency bands; perturbing them in raw feature space for counterfactual analysis easily destroys their temporal structure and produces physically implausible results. In this work, we introduce IMFACT (IMF-based counterfACTuals), a model-agnostic framework for generating plausible counterfactual explanations for time series classifiers that operates in the decomposition space of Empirical Mode Decomposition. An input signal is split into Intrinsic Mode Functions (IMFs), and selected IMFs are progressively substituted with those of a Nearest Unlike Neighbour (NUN) until the classifier flips to the target class. We evaluate six IMF-selection strategies and a multi-NUN cycling extension on two UCR benchmarks (FaultDetectionA, FruitFlies). The variance-based strategy with three NUNs outperforms two prominent baseline techniques on reliability and plausibility metrics, while cycling across three NUNs yields the best proximity across both datasets.
[AI-24] NSF-HRPT: Neural Semantic Field meets Hierarchical Risk Perception Tree for Safety-Critical Scenario Assessment
链接: https://arxiv.org/abs/2608.04776
作者: Yu Zhao,Jiangyu Pan,Tao Hu,Ming Yin,Fan Yang,Jiangfan Liu,Xiubo Liang
类目: Artificial Intelligence (cs.AI)
备注: 13 pages, 5 figures
Abstract:The ability to accurately assess and anticipate risks in safety-critical scenarios is crucial for autonomous driving systems. While existing research has made progress in collision prediction, accurately quantifying risk levels from monocular vision inputs remains challenging due to the complex dynamics of multi-agent interactions and the inherent uncertainty in real-world environments. To address these challenges, we present NSF-HRPT, a novel framework that combines learning-based perception with structured reasoning for quantitative risk assessment. Our approach features a Neural Semantic Field (NSF) that learns to model scene semantics, trajectory predictions, and probabilistic Time-to-Collision (TTC) distributions from simulation data. During inference, the pre-trained NSF serves as a prior for our Hierarchical Risk Perception Tree (HRPT), which enables efficient parallel computation and spatial reasoning about multi-agent risks. Additionally, we introduce a Sim2Real enhancement strategy that improves real-world applicability without retraining by incorporating priors from foundation models. Extensive evaluations demonstrate that our framework achieves state-of-the-art performance on synthetic benchmarks and delivers competitive, near-state-of-the-art results on real-world datasets for both TTC estimation accuracy and risk localization precision. The proposed method provides an effective solution for real-time risk awareness from monocular camera inputs.
[AI-25] Fewer Tokens Smaller Cache: Reward-Coordinated Efficient Reasoning
链接: https://arxiv.org/abs/2608.04771
作者: Qiyuan Zhu,Dezhi Li,Pengyu Cheng,Tianle Chen,Jiacheng Wang,Ruijie Shen,Hao Gu,Sida Lin,Zirui Liu,Jiacheng Liu,Sirui Han
类目: Artificial Intelligence (cs.AI)
备注: Work in progress, revisions ongoing
Abstract:Large Reasoning Models (LRMs) excel on complex tasks through long chain-of-thought (CoT) reasoning, but their lengthy intermediate steps cause severe overthinking that inflates inference cost. KV-cache compression is a common solution, yet existing reasoning-oriented methods apply a uniform policy across the trajectory and judge compression only by what it removes from the cache. Two observations point the other way. First, a reasoning state’s tolerance to context loss varies along the trajectory, and process reward tracks it: deleting tokens at high-reward steps preserves accuracy far better than deleting the same budget at random. Second, compression is not free on the generation side, since a smaller cache leads the model to generate more tokens, partly canceling the saving. Together these motivate coordinating both sides under a single process reward. We propose ReCo (Reward-Coordinated Compression), a step-wise framework in which a lightweight process-reward estimator scores each completed step and drives three components: (1) reward-adaptive KV-cache compression that shrinks the retained cache harder at high-reward steps and less at low-reward ones, (2) a reward-banded penalty on reflection tokens that curbs redundant generation, and (3) confidence-based early stopping that triggers when the reasoning is reliable. Across three reasoning models and six benchmarks, ReCo reduces generated tokens by 37%-65% and end-to-end latency by 2.08x-2.35x over Full CoT, all while largely preserving accuracy.
[AI-26] PURPOSE: Poisoning Conflict Resolution in RAG via Proxy-Fact-Grounded Updates
链接: https://arxiv.org/abs/2608.04756
作者: Zijian Wang,Yubo Zhu,Muzhi Dong,Yanjun Lou,Yisheng Li,ZiLiang Zhang,Wei Tong,Yuan Zhang,Jingyu Hua,Sheng Zhong
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:In Retrieval-Augmented Generation (RAG), post-retrieval conflict resolution arbitrates among noisy or contradictory retrieved passages. However, the robustness of this safeguard against knowledge poisoning has not been adequately studied. Existing black-box poisoning methods all assert the target answer in frontal contradiction with what the resolver treats as settled, the very signal these methods are built to detect. We propose PURPOSE, a strict black-box poisoning attack that reframes the injection as an update that minimizes conflict, rather than as a counter-claim. PURPOSE extracts query-related facts approximating the resolver’s possible reference, then grounds a pivot event in them to keep the injection consistent with what the resolver might verify while steering the generator toward the target answer. Across three QA benchmarks, five generators, and three conflict-resolution methods, PURPOSE attains the highest attack success rate (ASR) in 35 of 45 settings and exceeds the strongest prior attack with +9.7 mean ASR points. These results show that our poisoning method is effective against conflict resolution in RAG and identify non-contradicting injection as a practical mode to enhance poisoning attack.
[AI-27] EviGraph: Evidence-Guided Autonomous Research Agents
链接: https://arxiv.org/abs/2608.04738
作者: Zhenjiang Ren,Ruiji Li,Xujing Zhang,Ziliang Pang,Shuo Ren,Jiajun Zhang
类目: Artificial Intelligence (cs.AI)
备注: 23 pages,2 figures
Abstract:Autonomous research agents can generate hypotheses, execute experiments, and draft manuscripts, yet their outputs often contain unsupported claims and inconsistencies between research questions, experiments, results, and conclusions. We argue that this problem is partly architectural: existing systems organize research as sequential pipelines but do not explicitly maintain or validate the evolving claim-evidence structure across this http URL this paper, we introduce EviGraph, an autonomous research framework that represents the research process as a typed evidence graph containing Problem, Gap, Hypothesis, Experiment, Finding, and Claim nodes. The graph serves as the operational state of the agent rather than a post-hoc record. EviGraph inspects evidence chains for missing dependencies, semantic misalignment, and result-claim inconsistencies, localizes the earliest weak node, and regenerates its affected downstream subgraph. Graph checkpointing prevents unsuccessful repairs from corrupting previously validated evidence. Manuscripts are generated only after every retained claim is grounded in a validated evidence this http URL on ARC-Bench-ML and NanoResearch-20 show that EviGraph outperforms the compared end-to-end research-agent baselines in overall research performance, improves Claim Support Rate by 40.19% over the strongest baseline, and achieves 87.73% Experimental Data Consistency. These results demonstrate the value of explicit evidence-state maintenance for reliable autonomous research.
[AI-28] Chain-of-Thought Monitoring Can Be Unreliable in Implicit-Influence Settings
链接: https://arxiv.org/abs/2608.04735
作者: Agatha Duzan,Asa Cooper Stickland
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Chain-of-thought (CoT) monitoring is increasingly treated as an important safety layer for frontier reasoning models. Most monitorability evaluations study explicit-influence settings: setups where the prompt directly incentivizes the model to hide something, e.g., by instructing it to perform a hidden side-task. A complementary axis for CoT-monitor evaluations is implicit-influence settings, where the prompt contains no instruction to hide, but the model’s behavior is still shaped by features of the task or context, e.g. an irrelevant detail about a candidate that biases a hiring rating. We introduce the first benchmark that directly compares CoT monitorability under the two regimes. We test how model behavior changes in the presence of a nudge to choose a particular option. The nudge is delivered either as a casual aside (implicit), or as a direct instruction to act on the nudge and to conceal having done so (explicit). The benchmark spans four task formats (binary choice, numeric rating, multiple-choice QA, open-ended coding) and seven frontier extended-thinking models. Under explicit influence, a CoT monitor detects 60-94% of behavior shifts: even models instructed to conceal it leak the instruction into their CoT. Under implicit influence, the same factors still shift behavior, but detection falls by 41-46 percentage points in two of our four settings. Realistic system-prompt additions (of the kind a developer might deploy to reduce off-topic bias) lower implicit detection further, to as low as 5%, while preserving the behavioral influence itself. These results suggest that monitorability estimates obtained in explicit-influence settings may over-estimate monitorability, and that monitorability can be further decreased by well-intentioned deployment choices. Our benchmark and code are available at this https URL Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2608.04735 [cs.AI] (or arXiv:2608.04735v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.04735 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Agatha Duzan [view email] [v1] Wed, 5 Aug 2026 11:59:20 UTC (226 KB)
[AI-29] oward Integrating Adaptive Experience Replay and Online Uncertainty Estimation in Safe Actor-Critic Optimal Control
链接: https://arxiv.org/abs/2608.04732
作者: Mahshad Rastegarmoghaddam,Davoud Nikkhouy,Shima Samadzadeh
类目: ystems and Control (eess.SY); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: Code, deterministic seeds, data, figures, and protocol files are archived at this https URL and this https URL
Abstract:Safe actor-critic control often treats barrier filtering, uncertainty estimation, and experience replay as separate modules, even though each changes the data used for learning and control. We develop an integrated architecture in which the uncertainty estimate updates the obstacle geometry used by a control barrier function, filter interventions and estimation residuals determine replay priority, and the critic learns from the executed rather than nominal action. We instantiate the architecture on a two-dimensional robot-navigation task with corrupted obstacle measurements and compare six component-matched configurations under common training budgets, random seeds, sensor streams, exploration, and disturbances. Evaluation includes a moderate post-training test, an eleven-level perception-noise sweep, and an exploratory extreme-stress test at multiplier 6.0 . In the extreme test, the integrated configuration recorded no contacts and reached the goal in all five evaluation seeds. Its mean cost was 7.63\pm0.44 and its obstacle-belief root-mean-square error was 3.52\pm0.55 cm. The uncertainty-estimation ablation also recorded no contacts but reached the goal in four of five seeds, with mean cost 8.96\pm2.08 and belief error 11.08\pm1.23 cm. A finite-training bound clarifies replay exposure, and a robust barrier condition states the required estimation-error and feasibility assumptions. The results support coupling estimation, safety filtering, and replay on this benchmark; broader safety and convergence claims require further study.
[AI-30] Diagnosing Tool-Selection Reasoning in LLM Agents with Canary Tools
链接: https://arxiv.org/abs/2608.04719
作者: Atul Anand,Sourav Chattaraj
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 9 figures, 5 tables
Abstract:Agent evaluations tell us that a model picked the wrong tool, but rarely why. We introduce canary tools: diagnostic probe tools planted in an agent’s Model Context Protocol (MCP) tool set, each engineered to probe one specific tool-selection weakness. A six-type taxonomy (semantic decoys, parameter traps, capability mirages, prerequisite blindness, temporal decoys, and granularity traps) turns a single “wrong tool” outcome into a multi-dimensional profile of how a model reasons about tools. We evaluate eight models – six hosted and two 8B open-weight – spanning three capability tiers, on 120 tasks across three canary-density conditions and three seeds (8,640 runs), plus a 2,880-run subtlety ablation. Task success is graded by a provider-independent judge, corroborated by a second independent judge (Cohen’s kappa = 0.75). We report three findings. First, susceptibility drops sharply as models get more capable: the per-task canary susceptibility rate (CSR) ranges about 36x across models, lowest for Claude Opus 4.8 and highest for Llama 3.1 8B. Second, capability tier alone does not predict safety: the most susceptible hosted model is mid-tier, and within a provider the cheaper model can be the safer one. Third, the taxonomy is capability-stratified: capability mirages most reliably trap frontier models, while the other types are largely inert on strong models but fire on small open models, so they discriminate by capability rather than being weak. Softening each canary’s give-away phrase leaves frontier CSR essentially unchanged, evidence that the probes measure reasoning, not phrase-spotting. Susceptibility also predicts task failure (Spearman rho = -0.34), while the most robust models are not significantly degraded by canary pressure. We release the framework, canary schemas, tasks, and logs.
[AI-31] What We Observe as LLM Behavior Can Be a Side-effect of Inference Backend
链接: https://arxiv.org/abs/2608.04714
作者: Shahed Masoudian,Passant Shafaei,Monorama Swain,Markus Schedl
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Benchmark scores are reported as properties of a model, yet the inference framework used to produce them, such as HuggingFace, vLLM, or Ollama, are considered non-influential and their names and versions are almost never disclosed. In this work we investigate how much this choice can influence the model output. In a fully-crossed study (three instruction-tuned models x five inference frameworks x six benchmarks x four generation modes) we investigate how different tools (wrappers/backend) influence benchmark scores and how their score changes is influenced by generation hyper-parameters. We find backend to be a non-negligible factor where even under greedy, sampling-noise-free decoding, changing the backend can significantly alter models performance and this effect is structural and strongly model-dependent. Decomposing the variance according to generation mode reveal that considerable portion of the variability (roughly 39%) a practitioner sees out-of-the-box can stem from the backend, while the remaining stems from sampling noise and each framework’s default generation parameters, both of which are avoidable by disclosing and matching the generation configuration. These divergences are more pronounced on factual than on social-bias benchmarks. Overall, benchmark numbers are not backend-agnostic therefore, we recommend disclosing the backend, its version, and the full generation configuration, also using deterministic decoding for cross-backend comparison.
[AI-32] A 6G Integrated Sensing and Communication Framework for Railway Intrusion Detection and Collision Prediction
链接: https://arxiv.org/abs/2608.04710
作者: Ajeet Kumar Yadav,Sankaran Balasubramaniam,Aritra Chatterjee,Vinod Aduru,Yogesh Simmhan,Pandarasamy Arjunan
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Signal Processing (eess.SP)
备注:
Abstract:Integrated Sensing and Communication (ISAC) combines sensing and communication to efficiently utilize wireless resources and is emerging as a key paradigm for next-generation wireless networks. By leveraging the wide bandwidth, high frequencies, and massive antenna arrays of 5G-Advanced and 6G systems, ISAC enables physical-layer sensing using Channel State Information (CSI). The 3rd Generation Partnership Project (3GPP) Release 19 identifies 32 potential ISAC use cases, with particular emphasis on detecting and tracking moving objects. In this work, we address the Sensing for Railway Intrusion Detection use case, where intruders, including wildlife, entering a railway track can pose serious collision risks. We generated 22,695 CSI matrices with corresponding ground truth using a 3D-rendered railway environment and the Sionna radio simulator. We developed a machine learning model combining a three-dimensional Convolutional Neural Network (3D CNN) and Bidirectional Long Short-Term Memory (BiLSTM) network to detect intruders in the track danger zone and estimate their real-time position relative to the train, velocity, and time to collision. On synthetic CSI data, the model achieves 99.57% intruder-detection accuracy on a balanced test set and a combined Mean Absolute Error (MAE) of 0.4240 for position, velocity, and time-to-collision prediction. These results demonstrate the potential of CSI-based ISAC sensing with machine learning for reliable railway intrusion detection. The complete codebase for CSI generation, preprocessing, and model development is publicly available at this https URL.
[AI-33] raceable LLM -Generated Hazard Scenarios for Operational Safety Analysis of Aviation Systems Using ASRS Reports
链接: https://arxiv.org/abs/2608.04697
作者: Cristian Mascia,Roberto Pietrantuono,Daniel Rodriguez,Stefano Russo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Operational hazard analysis of aviation system operations must consider interactions among weather, ATC actions, airspace constraints, aircraft operations, and human factors - distinct from the functional hazard assessment applied at the aircraft-system level. We present an AI-assisted approach that generates candidate hazard scenarios from NASA’s Aviation Safety Reporting System (ASRS). Given a target adverse outcome, it produces a structured hypothesis as categorical factors and a narrative scenario describing an operational event sequence consistent with the structure. Each scenario includes by a plausibility score from historical co-occurrence evidence and traceability to the most similar held-out ASRS reports. We then propose a hybrid variant, conditioning narrative generation on a structured hypothesis produced via evolutionary abduction, improving correctness and reducing variability. We evaluate multiple large language models, zero-shot versus few-shot prompting, and optional fine-tuning, measuring how prompting and model choice affect the validity and realism of the generated structures and narratives.
[AI-34] Personalized Federated Sparse Adaptation of Time-Series Foundation Models
链接: https://arxiv.org/abs/2608.04695
作者: Priyanka Nihalchandani,Naman Srivastava,Varun Ojha,Pandarasamy Arjunan
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注: 15 pages
Abstract:Federated adaptation of time-series foundation models (TSFMs) is attractive for building energy forecasting because meter data are private, distributed, and highly non-IID. However, a single parameter-sharing strategy is unlikely to serve all pretrained TSFMs or building clients: fully shared adapters can suppress building-specific temporal behavior, while fully local adaptation discards cross-building transfer. We propose a personalized federated sparse adaptation framework with a heterogeneous temporal mixture-of-experts (MoE) adapter placed after the pretrained TSFM representation. A sequence-level router maps each 168-hour context window to a top- k subset of experts specialized for periodicity, long-range interactions, local variation, trend-residual structure, and multi-resolution behavior. We compare global FL, local training, and personalized FL variants with globally shared or client-private expert banks. Across 50 buildings and three TSFM backbones, personalization consistently outperforms Global FL-MoE and Local MoE, while the best sparse-adaptation strategy varies by backbone and metric. Routing behavior further reveals client-level expert specialization, expert concentration, and near-uniform routing across backbones, showing that federated TSFM adaptation should be both client-aware and backbone-aware.
[AI-35] Active-SWE: Benchmarking Coding Agents for Proactive Bug Fixing without Issue Reports
链接: https://arxiv.org/abs/2608.04682
作者: Haobin Li,Ping Deng,Weizhong Qian,Liang Jiang,Zhenyu Huang,Mouxing Yang,Xi Peng
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 24 pages, 17 figures
Abstract:Coding agents powered by large language models (LLMs) are increasingly adopted in software engineering (SWE) scenarios, capable of fixing a specific bug in large-scale codebase. However, existing SWE benchmarks typically assume that high-quality issue reports with detailed information are always available, which is easily violated in practice due to the complexity of report acquisition and curation. To address this, we introduce Active-SWE, a benchmark for evaluating coding agents on proactively discovering and fixing multiple bugs without report guidance, covering 1,663 tasks across six bug categories and eight languages. Beyond shifting the focus from existing reactive bug fixing to proactive bug fixing, Active-SWE enables a more in-depth evaluation by expanding the scope from fixing a specific recorded bug to multiple-bug fixing and potential bug discovery scenarios. To construct Active-SWE, we propose a novel difficulty-aware task formulation pipeline with a dual-track evaluation framework, facilitating comprehensive evaluation of proactive bug-fixing capability. Extensive experiments reveal that most state-of-the-art coding agents struggle with proactive bug-fixing tasks, demonstrating limited performance in locating and resolving recorded bugs, handling multiple bug fixing scenarios, and discovering valid potential bugs.
[AI-36] Calibrating Artificial Guilt: Neurally Grounded Reward Shaping for Prosocial Multi-Agent Reinforcement Learning
链接: https://arxiv.org/abs/2608.04663
作者: Aaditya Mehta,Arya Shah
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 6 figures, 3 tables
Abstract:Cooperative multi-agent reinforcement learning often adds social terms to individual rewards, yet the scale of those terms is usually chosen by hand. We ask whether a guilt signal can instead be calibrated from human neural and behavioural data and transferred to artificial agents. Using the public SoDec responsibility fMRI dataset (40 participants), we fit a subject-fixed-effects regression of momentary-happiness changes on outcome-type counts and recover a guilt weight as the Partner-negative minus Social-negative contrast ( \hatw=1.118 , Cohen’s d=0.214 ). We embed this weight in a two-agent Social Lottery environment and train independent Proximal Policy Optimization actor-critics under four shaping regimes: neurally calibrated, uniform constant, zero (selfish), and a unit-coefficient oracle. Across 1,000 evaluation episodes per condition, the calibrated agents track the human Social safe-choice rate most closely ( 0.459 vs.\ human 0.484 ; \mathrmKL=0.0012 ), while the other three conditions deviate by one to three orders of magnitude in KL. Human neurobehavioural priors can therefore act as quantitative constraints on prosocial reward shaping.
[AI-37] A/B Agent : A Self-Evolving Agent for Strategy Iteration in Industrial A/B Testing
链接: https://arxiv.org/abs/2608.04625
作者: Zhuohang Jiang,Yuxin Chen,Yongsen Pan,Zheng Hu,Wenqi Fan,Qing Li,Hongyang Wang,Jun Wang,Wenwu Ou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Industrial recommendation strategy iteration heavily relies on large-scale A/B experimentation. Traditional tuning requires experts to repeatedly design strategies, configure experiments, analyze results, and adjust parameters, making the process labor-intensive and time-consuming. Meanwhile, valuable knowledge from historical experiments is often fragmented, making systematic reuse difficult through manual expert effort alone. Existing RAG agents partially alleviate this burden by retrieving prior strategies, but typically organize experience in a flat manner, overlooking the hierarchical relationships among business scenarios, recommendation stages, optimization objectives, and experimental contexts. This often results in mismatched retrieval and limited cross-scenario transfer, while preventing agents from continuously refining strategies and parameters through sequential A/B feedback. % To address these limitations, we propose A/B Agent, a closed-loop A/B agent for industrial recommendation strategy optimization. The framework comprises three tightly coupled core components: Historical Strategy Knowledge Organization, Autonomous Target-Aware Strategy Generation, and Experiment-Guided Strategy Self-Evolution. It organizes historical strategies into a hierarchical experience tree, retrieves transferable evidence through multi-path Tree-RAG to generate executable strategies, and continuously analyzes online A/B feedback to guide autonomous tuning and update the experience tree for self-evolution. Extensive offline and online evaluations demonstrate its effectiveness, including a 4.829% improvement in GMV in a real-world short-video e-commerce recommendation system while maintaining positive gains across all guardrail metrics.
[AI-38] Masked diffusion enables coherent beat tracking
链接: https://arxiv.org/abs/2608.04624
作者: Francesco Foscarin,Filip Korzeniowski,Richard Vogl
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: Accepted at the 27th International Society for Music Information Retrieval Conference (ISMIR), 2026
Abstract:Current neural networks for beat tracking generate invalid outputs, such as consecutive downbeats and erratic tempo changes, even when these are not present in the training data. Heavy post-processing techniques can alleviate these problems, but the original cause of this inconsistent behaviour remains unknown. We hypothesise that it stems from inadequate modelling of multiple plausible output beat grids, resulting in an invalid mixture of competing interpretations. We propose a masked diffusion approach that properly models multiple outputs and enables the model to build coherent predictions through iterative inference. We devise three modifications to standard masked diffusion that enable its application to beat tracking: independent masking of beats and downbeats during training and inference, a balanced masking scheduler for inference, and peak-picking across inference steps. Our approach reduces erratic behaviours and improves beat-tracking performance.
[AI-39] Agreement Before Diversity: Verification-First Complementarity for Heterogeneous Language-Model Coordination
链接: https://arxiv.org/abs/2608.04618
作者: Ruitong Li,Binjie Guo,Aisheng Mo,Guowei Su,Jie Li,Ru Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Heterogeneous language-model ensembles expand the space of candidate responses, yet they lack a principled criterion for when a newly generated answer should supersede an already supported one. We decouple candidate headroom from replacement authority, rendering the latter as an explicit, auditable object. Our proposed method, Agreement-Before-Diversity (ABD), is a frozen, label-free decision rule: an anchor answer is retained if two additional trusted samples corroborate it under a fixed equivalence relation; otherwise, it is replaced by a heterogeneous synthesis. For this gating mechanism, we prove two exact identities. The first shows that the accuracy gap relative to unconditional synthesis is determined jointly by the agreement coverage and the anchor’s advantage on the protected subset. The second shows that the gap relative to never synthesizing reflects a contrast between authorized recovery and authorized destruction. Neither identity assumes independence or calibrated confidence, and the expected inference cost is approximately eight minus five times the coverage in number of calls. Under blind, exact-ID evaluation, ABD achieves 59.43% on the complete LiveCodeBench-v6 (vs. 52.57% for Single9 and 52.00% for HAC; n = 175) and 75.00% on an untouched GPQA-Diamond split (both controls at 72.78%; n = 180). Furthermore, these identities localize every aggregate difference to an enumerable protected stratum: no discordant items occur among the 3 protected cases on LiveCodeBench, where coverage bounds the gate’s contribution to 1.71 points a priori; 13 versus 8 discordant cases among 132 on GPQA-Diamond; and 12 versus 0 among 71 under a frozen anchor perturbation. Diversity supplies potential; verification structure supplies authority.
[AI-40] he Order Is the Guarantee: Verifier-Budgeted Code Deletion with Static-First Learned Proposals
链接: https://arxiv.org/abs/2608.04611
作者: Ruitong Li,Binjie Guo,Aisheng Mo,Guowei Su,Han Wang,Jie Li,Ru Zhang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Frontier coding models now match or exceed strong human reference points on programming benchmarks, yet benchmark success does not imply maintainable software. Prompt-driven “vibe coding” is additive: new branches, guards, and fallbacks accumulate faster than obsolete logic is removed. We study the inverse problem-how an Al system should remove code when execution-verification capacity is finite. We formulate redundant-code reduction as proposal scheduling: a ranker orders single-statement deletion candidates, an execution suite accepts the first candidate that passes, and a budget bounds how many candidates may be tested. Our central observation is that candidate order, not model confidence, is the control surface a deployment can reason about. DELSCOUT instantiates two schedules. Given representative target-domain validation, a five-slot budget spends three slots on deterministic shortest-first candidates and two on complementary learned candidates; across nine MBPP replications with 0.5B, 0.6B, and 8B rankers this raises verified-deletion coverage by 9.5% relative (+6.7 accepted tasks) while consuming slightly fewer verifier calls than the matched static baseline. Without such validation the same rankers can lose coverage under shift, so we instead evaluate the complete static prefix first and append learned candidates only afterwards; for a deterministic verifier this makes coverage and character reduction non-decreasing by construction, at a measured 4.8-62.5% increase in verifier calls. MBPP+ then erases the in-domain advantage, showing that scheduling governs search while the test suite alone governs what “preserving behavior” means. The result is an auditable division of labor: models widen the search for removable code, order bounds the damage a mis-ranked proposal can do, and execution retains authority over every committed deletion.
[AI-41] Rethinking Reservoir Pruning: A Dynamical Perspective for Echo State Networks
链接: https://arxiv.org/abs/2608.04593
作者: Sudip Laudari,Puspa Raj Adhikari
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Dynamical Systems (math.DS)
备注: 18 pages, 6 figures
Abstract:Echo State Networks (ESNs) offer an efficient framework for temporal prediction, but their randomly initialized reservoirs are often over-parameterized and dynamically redundant. Existing pruning methods largely rely on static connectivity or activation statistics, which may overlook neurons that shape input-driven state transitions. We propose Dynamical Mode Pruning (DMP), a reservoir pruning method that ranks neurons by their contribution to dominant transition modes obtained from a trajectory-averaged Jacobian Gramian. DMP removes low-impact units and retrains only the readout. Experiments on chaotic and real-world time-series benchmarks show that DMP improves or preserves forecasting accuracy while reducing redundant reservoir components. Our results suggest that dynamical influence is a useful criterion for reservoir refinement beyond static structural importance alone.
[AI-42] Joint UAV Flight and Opportunistic Routing under Reinforcement Learning for Delay-Tolerant Networks
链接: https://arxiv.org/abs/2608.04590
作者: Xiao Wang,Shun-Ren Yang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The growing deployment of delay-tolerant networks (DTNs) has made store-carry-forward (SCF) communication indispensable under sparse connectivity. However, intermittent contacts, finite buffers, and limited message time-to-live (TTL) often give rise to sparse delivery and congestion, leading to substantial end-to-end performance degradation. To address this challenge, this study explores the joint optimization of decentralized opportunistic routing and controllable unmanned aerial vehicle (UAV) flight, aiming to enlarge future contacts through discrete UAV headings while enabling per-node replication under contact-limited observations. Building upon this architecture, we study cooperative factored routing–UAV control under centralized training and decentralized execution (CTDE) and propose JUROR (Joint UAV flight and Opportunistic Routing, based on the proximal policy optimization (PPO) framework. In our design, we first cast the problem as a factored partially observable Markov decision process with sequential motion–routing coupling and a per-step team reward; subsequently, decentralized actors act on local observations while a training-time critic uses global statistics, and an optional multi-horizon hotspot predictor provides auxiliary supervision. Simulation results over four traffic modes demonstrate effective gains over PRoPHET and MaxProp, while retaining contact-limited decentralized execution. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2608.04590 [cs.AI] (or arXiv:2608.04590v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2608.04590 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Xiao Wang [view email] [v1] Wed, 5 Aug 2026 08:51:41 UTC (642 KB) Full-text links: Access Paper: View a PDF of the paper titled Joint UAV Flight and Opportunistic Routing under Reinforcement Learning for Delay-Tolerant Networks, by Xiao Wang and Shun-Ren YangView PDFHTML (experimental)TeX Source view license Current browse context: cs.AI prev | next new | recent | 2026-08 Change to browse by: cs References Citations NASA ADSGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv’s community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) mathjaxToggle(); We gratefully acknowledge support from our major funders, member institutions, , and all contributors. About Help Contact Subscribe Copyright Privacy Accessibility Operational Status (opens in new tab) Major funding support from
[AI-43] What Is a Skill Worth? Structure-Aware Shapley Valuation of Agent Skills
链接: https://arxiv.org/abs/2608.04562
作者: Tao Li,Junfeng Liu,Qinghua Zhao,Yifan Li,Lei Wang,Bo Shao,Xuejun Liu,Linjun Shou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Agent skills are increasingly optimized by automated feedback loops, producing long structured artifacts whose internal value remains unclear. We study skill valuation: assigning credit to the internal units of a fixed skill, such as rules, examples, scripts, and heuristics, under a fixed agent and held-out task distribution. Skill valuation differs from data or prompt-span valuation because skill units are structured: they may depend on other units, belong to a document hierarchy, trigger agent behavior, and consume limited prompt context. We introduce SkillSV, a structure-aware Shapley-style framework for skill valuation. SkillSV compiles a skill into units, dependencies, and hierarchy, so that only valid counterfactual skills are evaluated. It uses paired deletion and length-neutral padding to separate content value from context cost, and estimates the resulting values with a rollout-budgeted estimator for noisy agent evaluations. On four agentic benchmarks, we assess the faithfulness, actionability, and explanation of SkillSV: it recovers unit interactions, preserves aggregate skill lift, and guides safe pruning and compression.
[AI-44] A Model Merging Approach for Continual MLLM Unlearning
链接: https://arxiv.org/abs/2608.04548
作者: Yuhang Wang,Linlin Zhang,Haoxuan Ji,Xianmin Ye,Zhenxing Niu,Haichang Gao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 17 pages, 5 figures
Abstract:Multimodal large language model (MLLM) unlearning methods have been proposed to remove private, sensitive, or proprietary information from well-trained models. However, most existing MLLM unlearning methods are designed for one-shot requests and fail to adequately address continual scenarios, as repeatedly applying one-shot operations leads to cumulative utility degradation, unlearning rebound, and retention drift. We introduce Merging for Continual Unlearning (MCU), an approach that dynamically merges multiple one-shot unlearning adapters into a unified adapter upon receiving each new unlearning this http URL a leave-one-out merging analysis, we reveal that these unlearning adapters exhibit strong cross-task dependencies. Such dependencies have two contrasting effects: they can facilitate cross-task unlearning transferability, but they can also introduce severe interference that degrades unlearning effectiveness and compromises retained knowledge. To address this challenge, MCU projects the adapters into a shared representation space, preserves their dominant directions, suppresses over-concentrated coordinates, and reconfigures cross-task dependencies to mitigate interference while enhancing transferability. Experiments on ICU-Bench and MLLMU-Bench demonstrate that MCU achieves superior unlearning effectiveness while preserving both retained knowledge and general multimodal utility.
[AI-45] GUARD: Grounding Uncertainty and Ablation-Based Risk Detection for Diffusion-Based VLAs
链接: https://arxiv.org/abs/2608.04510
作者: Suhas Hegde,Jitendra Yasaswi Bharadwaj Katta
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion-based vision-language-action (VLA) policies can generate plausible actions even when their predictions are weakly grounded in the visual and language evidence defining the task. We introduce GUARD, a test-time failure detection method that measures this grounding without modifying the pretrained policy. GUARD estimates the influence of token-indexed entries in the final vision-language model key-value (KV) cache, constructs counterfactual caches by ablating salient KV entries, and compares their denoising responses with the original conditioning. Based on the comparison, we derive GUARD diagnostic stream including sensitivity, attention entropy, modality bias, and grounding efficiency, which are calibrated online and processed by a lightweight temporal classifier. We evaluate GUARD under task-held-out splits across five policy-benchmark settings, using Pi0, SmolVLA, and Alpamayo-1.5 on LIBERO, SimplerEnv, MetaWorld, and PhysicalAI-AV. GUARD achieves the best ROC-AUC on four of five unseen-task settings and ranks second on the remaining setting, improving the average unseen-task ROC-AUC by 5.73 percentage points over the strongest competing runtime monitor while remaining within 0.19 points of the best seen-task average. These results show that directly probing action-head dependence on multimodal evidence provides a transferable failure signal across policies, tasks, embodiments, and domains.
[AI-46] CARGO-VL: Counterfactual Arbitration with Risk-Constrained Group Optimization for Vision-Language Models
链接: https://arxiv.org/abs/2608.04509
作者: De Jiang,Zhengyang Zhang,Kehong Yuan,Shaohua Ma
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-language systems combine images with retrieved text, but these sources can disagree or jointly fail to support an answer. Reliable models must identify the trustworthy source and abstain when neither is adequate. Existing post-training objectives score instances independently and therefore do not enforce coherent behavior under counterfactual evidence changes. We introduce CARGO-VL, a group-relative framework that optimizes matched variants covering aligned, image-correct, text-correct, and both-wrong (A/V/T/N) evidence states as one bundle. Its objective couples condition-wise correctness with transition rewards for answer invariance, source equivariance, and answer-to-abstention switching, while a primal-dual controller balances unsafe answers against excessive deferral. We also contribute XMC (eXtended Modal Conflict), a four-condition conflict training resource, and evaluate transfer on CMC-Bench and Modality-Bias. Across multiple seeds, CARGO-VL improves conflict handling, unsupported-answer avoidance, and modality balance over pointwise baselines. Ablations identify complementary benefits from relational transition signals and adaptive risk control, supporting counterfactual consistency as a practical objective for reliable multimodal evidence arbitration.
[AI-47] AFD-Ledger: Deployment Provisioning for Attention–FFN Disaggregation
链接: https://arxiv.org/abs/2608.04502
作者: Chengyu Qiu,Xiao Fu,Fengcun Li,Yulei Qian,Yuchen Xie,Xunliang Cai,Yingdi Shan,Yongwei Wu,Mingxing Zhang
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注: 14 pages, 14 figures, 2 tables
Abstract:Attention–Feed-Forward Network (FFN) Disaggregation (AFD) is emerging as a promising architecture for serving Mixture-of-Experts (MoE) language models. While existing AFD systems improve the efficiency of disaggregated execution, they leave a deployment question unanswered: under the same model, workload, time-per-output-token (TPOT) service-level objective (SLO), hardware budget, hardware catalog, and runtime capabilities, does AFD provide higher throughput than the best collocated deployment? Answering this question requires jointly optimizing hardware assignment and deployment organization for both architectures, making exhaustive provisioning prohibitively expensive. We present AFD-Ledger, an offline analytical provisioning system that independently provisions AFD and collocated deployments using an analytical execution model and an evaluation-bounded hardware search. Across deployment spaces where exhaustive provisioning is feasible, AFD-Ledger reduces complete deployment evaluations by 68.8%–83.5% while still recovering the globally optimal deployment. On three physical LongCat 2.0 deployments, it preserves the correct architecture decision while predicting AFD-to-collocated throughput within 6.6%–9.6% of measurement. Using this validated framework, we show that homogeneous AFD improves fixed-budget throughput in only a minority of the studied settings, heterogeneous AFD requires deployment-level hardware complementarity rather than heuristic device selection, and role-specific hardware improvements matter primarily when they enable better deployment organizations by crossing deployment capability–price boundaries.
[AI-48] AudioScape-TTA: A Structured Soundscape Benchmark for Fine-Grained Text-to-Audio Evaluation
链接: https://arxiv.org/abs/2608.04479
作者: Jinting Wang,Yuguang Yang,Shengyu Li,Yan Rong,Shan Yang,Xiaoda Yang,Li Liu
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注:
Abstract:Text-to-audio (TTA) generation has recently achieved remarkable progress in synthesizing realistic audio from natural language descriptions. However, determining whether generated audio faithfully satisfies complex textual instructions remains challenging. Existing benchmarks mainly rely on global similarity metrics, providing limited insight into fine-grained semantic failures. To address this limitation, we introduce \textbfAudioScape-TTA, a structured and complexity-aware benchmark for fine-grained TTA evaluation. AudioScape-TTA represents realistic soundscapes through modality-aware semantic structures and characterizes generation complexity using event density and structural complexity. Based on these annotations, we propose a rubric-based audio-grounded evaluation framework that verifies event realization, acoustic attributes, and speech content through fine-grained semantic criteria. The benchmark contains 2,258 audio-text pairs with 25,707 binary QA rubrics, enabling scalable and interpretable analysis of TTA systems. Experiments on 13 representative open-source TTA models reveal persistent limitations in fine-grained attribute control, speech-content preservation, and compositional soundscape generation. Human validation further demonstrates that our rubric-based evaluation achieves stronger alignment with human semantic judgments than conventional global similarity metrics.
[AI-49] Beyond Linear Dynamics: Neural Bilinear Dynamical Models for Time Series Forecasting
链接: https://arxiv.org/abs/2608.04471
作者: Mengzhou Gao,Huangqian Yu,Pengfei Jiao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Time series in real-world applications are often generated by nonlinear dynamical systems, making accurate forecasting challenging. Existing approaches that explicitly model system dynamics typically rely on linear assumptions or Koopman-based linearizations, which may inadequately capture complex nonlinear behaviors and lead to error accumulation in long-horizon prediction. To address this limitation, we propose the Neural Bilinear Dynamical Model (NBDM), which models nonlinear system dynamics through a bilinear latent dynamical formulation. Specifically, NBDM leverages Koopman theory to lift the original nonlinear dynamics into a higher-dimensional latent space, where a bilinear dynamical model is constructed to characterize state evolution. To mitigate the approximation error introduced by bilinear representations, we further incorporate a parameterized error compensation term. Within this formulation, control inputs are explicitly integrated into the dynamics, using auxiliary variables when available and learned feedback signals otherwise. To handle scenarios with missing control inputs, we design a memory-enhanced controller that infers latent controls through multiplicative interactions between historical states and control signals. Experiments on five real-world datasets demonstrate that NBDM consistently outperforms competitive baselines in both given-control and missing-control settings, particularly for multi-step and long-horizon forecasting.
[AI-50] ropical Algebraic Geometry for Neuronal Representations: An Arakelov-Green Measure Based Descriptor for Graph Learning
链接: https://arxiv.org/abs/2608.04460
作者: Yuyang Zhang,Weihan Xu,Xuehai Zhou,Shucheng Cao,Qihuang Zhang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computational Geometry (cs.CG)
备注:
Abstract:The quantitative analysis of 3D neuronal morphologies requires capturing both graph topology and spatial geometry. Current message-passing Graph Neural Networks (GNNs) are bounded by the 1-Weisfeiler-Lehman (1-WL) test, limiting their ability to capture cycles induced by spatial proximities. To address this, we propose a training-free geometric prior based on tropical algebraic geometry. We apply the recently established tropical Abel-Jacobi transform and polarization distances to machine learning on tree-structured data. We introduce a structural transformation pipeline, comprising cycle space augmentation and quotient space construction, to convert spatial trees into cyclic metric graphs suitable for embedding into the Tropical Jacobian. Computing exact tropical polarization distances requires solving the NP-Hard Closest Vector Problem (CVP) on integer lattices. Instead of relying on explicit approximations with quantization errors (e.g., Babai’s rounding), we adopt a continuous relaxation on the universal cover of the Albanese torus. We show that the discrete Arakelov-Green measure, computed in closed form via the graph Laplacian’s generalized inverse, decomposes exactly into the intrinsic path metric minus the unquantized polarization distance on this cover, avoiding integer lattice searches. This metric yields two descriptors: eigenvectors provide node-level structural coordinates, and the permutation-invariant eigenvalue spectrum provides a graph-level signature. On the BREC benchmark, the eigenvector formulation demonstrates expressivity beyond the 1-WL limit. On 3D morphology datasets (ACT-4, JML-4, BIL-6), the spectrum seamlessly integrates into standard architectures (VAEs, GNNs, Tree-LSTMs) without additional trainable parameters, outperforming explicit lattice approximations and improving classification accuracy over existing spatial models.
[AI-51] Architectural Implications of Agent ic AI Workflows
链接: https://arxiv.org/abs/2608.04458
作者: Jirong Yang,Peizhe Liu,Chaojie Zhang,Jovan Stojkovic
类目: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR); Operating Systems (cs.OS)
备注:
Abstract:Agentic AI is emerging in datacenters, but its architectural implications remain unexplored. We organize agentic workflows in a taxonomy and present its first architectural characterization with a production study at Microsoft Azure and a controlled study of open-source frameworks. We show that agentic execution is fragmented and heterogeneous. Requests expand into a workflow of LLM inferences, tool invocations, and orchestration decisions that repeatedly cross the CPU-GPU boundary. Our taxonomy explains how this fragmentation turns into resource demand. As orchestration and tools run on the host, the CPU sits on the critical path. Execution structure sets the load over time, which stays low with sudden spikes. Model composition sets how evenly the workflow uses the GPUs. Diversity in tasks and tools widens this range even further. These characteristics expose architectural mismatches of conventional uniform servers. Fragmented execution strands CPU and GPU capacity despite bursty demand. Different software roles make homogeneous CPU provisioning inefficient. Finally, multiplexing many agents onto shared cores degrades microarchitectural locality. Guided by our findings, we derive implications for agentic servers and examine them through Agora, our prototype for commodity servers. Agora dynamically harvests idle CPU cores for co-located throughput work, while protecting agentic tail latency against tool spikes. It oversubscribes GPU memory by placing more agents on each GPU, prefetching the next agent’s state to hide swap latency. To match the machine to the heterogeneous roles, Agora pools cores by role and applies affinity-aware scheduling to restore locality. It automatically tunes mechanisms to the workload. Agora improves utilization and server throughput while preserving agent tail latency. Our insights also identify key directions for future server architectures for agentic AI.
[AI-52] Eigenius: A Typed Knowledge-Graph DBMS with Epistemic Stratification and Institution-Mediated Reasoning
链接: https://arxiv.org/abs/2608.04457
作者: Hans-Martin Will,Allen L. Brown Jr.,Matthew Fuchs
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
备注:
Abstract:As “AI Scientists” emerge to drive research via the Model Context Protocol (MCP), systems relying on ephemeral scripts will fail. The sheer scale of stateful, interconnected evidence requires a machine-walkable warranty grounded in a purpose-built database architecture. Eigenius is an open-source, typed knowledge-graph DBMS built on a single premise: answering the audit question (“what do you know, and what is your warranty?”) requires a unified kernel. By tightly coupling the type system, storage engine, and integration protocol, Eigenius turns data provenance into a structural invariant rather than a property reconstructed across subsystem boundaries. The kernel rests on three pillars: a dependent type theory woven through the core, institutions acting as strongly typed integration boundaries, and a content-addressed immutable storage layer. On this foundation, epistemic status (declared/observed/derived/verified) is enforced as a strict commit-time invariant. Cross-system translations (comorphisms) are checked at commit and materialized directly into the graph as durable, first-class resources. To eliminate O(N^2) polystore bottlenecks, shared on-chain intermediate representations (IRs) collapse multi-system translations to identity. Crucially, this architecture unifies both domains of scientific epistemology: it relies on justification logic for empirical science, while embedding a fast, in-process term checker to safely evaluate formal mathematical proofs (via Lean 4) without IPC overhead. In an end-to-end recomputation of a published Nature study from fragile scripts to a materialized evidence graph, all 52 derived conclusions hold from pinned data, surfacing four machine-checked discrepancies in the original study.
[AI-53] ExeCRE: Execution-Consistency Guided Reliability Estimation for Self-Correcting Code Generation
链接: https://arxiv.org/abs/2608.04439
作者: Yiru Dong,Richong Zhang,Fanshuang Kong,Si Chen
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 13 pages, 5 figures. Accepted at the 41st IEEE/ACM International Conference on Automated Software Engineering (ASE 2026)
Abstract:Large language models (LLMs) have made notable progress in code generation, but they still struggle on challenging tasks that require sophisticated algorithms or complex implementations. Recent methods increasingly use code execution as feedback, especially in self-correction pipelines that construct verification signals from generated code. However, these pipelines often depend on supervision signals whose reliability is unknown, which can introduce misleading feedback, unnecessary revisions, and incorrect final answers. To address this issue, we propose ExeCRE, an Execution-Consistency guided code Reliability Estimation framework. Instead of judging candidate code by tests or LLM feedback, ExeCRE estimates code reliability by statistically analyzing consistency patterns in execution outputs over a large number of randomly generated inputs. It collects execution outputs over generated inputs, projects them into consistency signals, and applies the Dawid-Skene model to infer latent code reliability. We integrate ExeCRE into self-correction for code generation. Experiments show that ExeCRE consistently improves both effectiveness and stability, while substantially reducing misleading correction signals. Under GPT-5.2 on LiveCodeBench, the average number of misleading feedback cases on already correct code drops from 113.2 with a representative self-correction baseline to 14.0 with ExeCRE. As an additional study, we apply the same reliability estimation strategy to code-based mathematical reasoning and observe similar benefits. These results suggest that ExeCRE enables more reliable use of generated code in execution-based pipelines.
[AI-54] Generative Optimization for Incentivized Advertising with Global Level Constraints
链接: https://arxiv.org/abs/2608.04421
作者: Gege Chen,Ning Luo,Hao Jiang,Da Li,Wenzheng Shu,Teng Sha,Yanxiang Zeng,Wenxin Tai,Fan Zhou,Xialong Liu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Incentivized advertising allocates monetary or virtual rewards to drive user engagement, where a key challenge is optimizing continuous incentive magnitudes under strict global constraints. This problem is complicated by high-frequency interactions, delayed feedback, and non-Markovian user dynamics such as fatigue, which limit the effectiveness of existing uplift modeling and constrained reinforcement learning approaches. To address these challenges, we propose GOAL, a constraint-aware generative framework that formulates incentive allocation as a conditional sequence generation problem. GOAL directly generates incentive magnitudes conditioned on user histories and system-level global pressure, and integrates a hierarchical causal state encoder to capture both local behavioral dynamics and long-range dependencies. To enable flexible constraint control, we introduce \textbfSafe \textbfConstrained \textbfPolicy \textbfOptimization (SCPO), which learns a single generative policy that generalizes across a spectrum of ROI constraints without retraining. Experiments on large-scale real-world data and a synthetic fatigue-aware environment show that GOAL improves long-term revenue and user retention while substantially reducing ROI violation rates compared to strong baselines.
[AI-55] SPOT: Sparse Probing and Outcome Calibration for On-Policy Distillation
链接: https://arxiv.org/abs/2608.04419
作者: Zikun Qu,Min Zhang,Mingze Kong,Zhiwei Shang,Yikun Ban,Shuang Qiu,Zhongxiang Dai
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Preprint
Abstract:On-policy distillation (OPD) provides dense teacher supervision on student-generated trajectories, but standard reverse-KL training can assign insufficient probability to other plausible continuations. Teacher entropy alone does not reveal whether uncertainty is concentrated among a few plausible next tokens or dispersed over a long probability tail, nor whether the student already represents those candidates well. Moreover, local teacher probabilities may not predict downstream success. We introduce Sparse Probing and Outcome-calibrated Targets OPD (SPOT), which addresses two coupled decisions, where to probe and what to distill, through an acquisition–exploration–exploitation procedure. During acquisition, a position-level score combines normalized teacher entropy, the probability mass captured by a small top- k candidate set, and student–teacher mismatch to allocate a limited probing budget. During exploration, SPOT evaluates teacher-proposed candidates through verifier-scored student continuations. During exploitation, these outcomes produce a closed-form, KL-regularized target that favors candidates with better downstream outcomes while remaining anchored to the teacher distribution. Extensive experiments across multiple student models and reasoning benchmarks demonstrate the effectiveness of SPOT in improving reasoning performance while balancing solution quality and coverage.
[AI-56] Not Every Divergence Should Be Suppressed: Counterfactual Recoverability in On-Policy Distillation
链接: https://arxiv.org/abs/2608.04408
作者: De Jiang,Zhengyang Zhang,Kehong Yuan,Shaohua Ma
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:On-policy distillation (OPD) supervises student-visited trajectories, yet divergence-based rules cannot determine whether an erroneous prefix remains correctable. We formulate this decision as counterfactual recoverability and replay each error state through budget-matched teacher-continuation and rollback branches. Based on their relative success, states are categorized as recoverable, irreversible-but-avoidable, or ambiguous, and these labels guide whether training retains, rolls back, or conventionally supervises the corresponding trajectory. On AIME branch diagnostics, the mean continuation-minus-rollback effect is 0.185 for recoverable states and -1.000 for irreversible-but-avoidable states, demonstrating opposite intervention preferences. A branch-derived recoverability proxy achieves an AUC of 1.000, substantially outperforming divergence alone at 0.392. Across frozen evaluations, recoverability-aware control achieves the strongest recorded performance, reaching 0.578 success on held-out AIME2025 compared with 0.517 for the best baseline. It also improves AIME2024-2025 average@32 from 0.2656 to 0.3125 and GPQA-Diamond average@32 from 0.2702 to 0.3070. Component ablations further show that retaining teacher-correctable prefixes provides the largest individual contribution. These findings establish recoverability as an outcome-grounded decision variable for selective supervision in OPD.
[AI-57] Approximate Multi-Objective Search Under Rulebooks
链接: https://arxiv.org/abs/2608.04398
作者: Omar Muhammetkulyyev,Oren Salzman,Tichakorn Wongpiromsarn
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Robotic planning often involves multiple objectives with complex priority relationships, such as safety, efficiency, and regulatory compliance. Rulebooks formalize these relationships, allowing partial ordering of objectives that generalizes both Pareto and lexicographic dominance. Computing the full set of rulebook-optimal solutions, however, is computationally expensive. To address this challenge, we introduce the concept of epsilon-rule-dominance, a principled notion of approximate dominance under rulebooks, and propose RApex, a best-first search algorithm that efficiently computes a compact set of epsilon-approximate rulebook-optimal solutions. RApex leverages dimensionality reduction, a technique used to speed up existing multi-objective search algorithms, while respecting rule hierarchies by maintaining separate closed sets and performing dominance checks over truncated and residual rule sets. We provide a formal analysis of RA*pex, proving that every rulebook-optimal solution is epsilon-rule-dominated (a generalization of approximate dominance we introduce) by at least one solution in the returned set. Empirical results demonstrate that our approach achieves computation times over two orders of magnitude faster than existing methods.
[AI-58] Improving Auto-Design of Neural PDE Solvers with a Domain-Specific Language
链接: https://arxiv.org/abs/2608.04384
作者: Shengxin Kong,Liwen Xu,Jingwen Fu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Neural PDE solver auto-design is fundamentally a search-space representation problem. In the space of unrestricted Python programs, valid solvers form an extremely sparse subset: most candidate programs are syntactically incorrect, semantically incompatible, or numerically unstable. Direct code generation therefore forces an LLM to spend most of its search capacity navigating implementation failures rather than reasoning about solver quality. ADSL-PDE addresses this challenge by introducing a structured search state between solver concepts and executable code. It represents the functional decisions that determine a neural PDE solver (architecture, physical constraints, objectives, sampling, and optimization) while abstracting away low-level implementation details. A deterministic compiler maps each valid search state to an executable solver. In effect, ADSL-PDE reshapes the search space: it removes large regions of invalid programs, increases the density of meaningful candidates, and preserves the compositional freedom needed to discover previously unseen designs. Solver evolution can thus operate over design decisions rather than code artifacts. Built on this representation, our evolutionary agent iteratively proposes, evaluates, and refines solver search states using empirical feedback. Across multiple PDE benchmarks, ADSL-PDE improves both search efficiency and optimization stability, achieving an improvement of more than 52% within the first ten evolution iterations. These results suggest a broader principle for LLM-driven auto-design: effective agents do not merely require stronger reasoning, but rather a search representation that concentrates exploration on valid and consequential decisions.
[AI-59] NodeJEPA: Structure-Conditioned Latent Prediction for Node-Level Graph Self-Supervised Learning
链接: https://arxiv.org/abs/2608.04381
作者: Tinghe Zhang,Jian Xu,Jiaheng Chen,Jiaxing Li,Yucheng Xiao,Qiang Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 20 pages, 10 figures
Abstract:Self-supervised learning on graphs is largely shaped by contrastive methods that depend on carefully designed augmentations, and by generative methods that reconstruct node attributes in the input space. Both paradigms can entangle representations with low-level input statistics rather than with relational structure. Joint-embedding predictive architectures (JEPA) instead learn by predicting latent targets rather than reconstructing inputs. Recent work has explored this idea for graph-level representation learning, but how to design JEPA-style objectives for node-level tasks, and which structural signals the predictor should condition on, remains less clear. We present NodeJEPA, a joint-embedding predictive architecture for node-level graph self-supervised learning. NodeJEPA masks structure-aware k-hop ego-subgraphs and trains a context encoder to predict the latent representations of the masked nodes. These targets come from an EMA-updated target encoder with stop-gradient. A structure-conditioned predictor integrates spectral and centrality descriptors through cross-attention. Variance, covariance, and Laplacian spectral regularizers help stabilize the embedding geometry, and an optional curriculum gradually increases masking difficulty during training. Because prediction occurs in latent space, NodeJEPA does not rely on input reconstruction or hand-crafted graph augmentations. We evaluate NodeJEPA on standard node classification benchmarks under linear probing and fine-tuning protocols, and conduct ablations on masking, prediction, and regularization design choices. Our study offers a practical recipe for node-level JEPA-style latent prediction on graphs, and clarifies when structural conditioning helps representation learning. Code, configurations, and evaluation scripts are publicly available at this https URL.
[AI-60] owards Trustworthy Hypergraph Neural Networks under Label Noise
链接: https://arxiv.org/abs/2608.04377
作者: Mengyao Zhou,Zhiheng Zhou,Xiao Han,Guiying Yan
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 20 pages, 7 figures
Abstract:Hypergraph neural networks (HGNNs) have demonstrated remarkable capabilities in processing complex higher-order relationships. However, their performance is highly dependent on labeled data, making them vulnerable to label noise. Despite advances in learning with label noise (LLN) and graph learning with label noise (GLN), noisy-label learning on hypergraphs remains underexplored. In this paper, we present a systematic study of hypergraph node classification under label noise. First, we adapt representative LLN and GLN methods to hypergraphs and evaluate them under a unified benchmark, revealing the limitations of existing robust learning strategies for hypergraphs. Building on this, we propose a new hypergraph robust framework, HyperTrust, which first estimates hyperedge trustworthiness through a pretraining-based, entropy-aware strategy, and then incorporates the HyperedgeBoost module to enhance reliable supervision by connecting unlabeled nodes to trustworthy hyperedges, as well as the HyperedgePrune module to suppress noisy propagation by removing untrustworthy node-hyperedge incidences. Finally, two modules work collaboratively to adjust the hypergraph structure and generate final predictions. Extensive experiments and theoretical analysis demonstrate the effectiveness and robustness of HyperTrust on multiple hypergraph datasets under various noisy settings. Our work provides a unified benchmark and an effective solution for hypergraph learning with label noise and lays a foundation for future research in this direction.
[AI-61] Combating Knowledge Corruption in Agent Systems: A Byzantine-Tolerant Secure Collaborative RAG Framework
链接: https://arxiv.org/abs/2608.04366
作者: Zhaoqi Wang,Daqing He,Zijian Zhang,Ye Liu,Jiamou Liu,Zhirui Zeng,Zhan Qin,Zhen Li,Xin Li,Hongwei Yao,Jincheng An,Yong Liu,Yi Li,Qi Sun,Xiulei Liu,Liehuang Zhu
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:While retrieval-augmented generation systems partially address the hallucination issues in large language models, it also introduces new vulnerabilities to knowledge corruption attacks. Adversaries exploit these vulnerabilities by poisoning documents provided by RAG system to manipulate LLM outputs. To counter this threat, we propose SecureCollaRAG, a Byzantine-tolerant collaborative RAG framework leveraging Multi-source Knowledge Validation Mechanism. Our approach enables agent system to securely verify document provenance through dynamic GNN-based credibility scoring, effectively preventing stealthy knowledge corruption attacks while preserving essential domain knowledge integrity. Through extensive evaluations and formal analysis, we demonstrate that SecureCollaRAG maintains robustness against attackers under non-IID data distributions.
[AI-62] NeuMoSync: End-to-End Neuromodulatory Control for Plasticity and Adaptability in Continual Learning
链接: https://arxiv.org/abs/2608.04358
作者: Seyed Roozbeh Razavi Rohani,Khashayar Khajavi,Wesley Chung,Mandana Samiei,Mo Chen
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Published in Transactions on Machine Learning Research (TMLR)
Abstract:Continual learning (CL) requires models to learn tasks sequentially, yet deep neural networks often suffer from plasticity loss and poor knowledge transfer, which can impede their long-term adaptability. Drawing high-level inspiration from global neuromodulatory mechanisms in the brain, we introduce Neuromodulation and Synchronization (NeuMoSync), a novel architecture that integrates dynamic, neuron-specific modulation into deep neural networks to enhance their adaptability and plasticity. NeuMoSync extends standard neural network architectures with learnable feature vectors for each neuron that track network-wide historical context and with a module operating at a higher level of abstraction. This module synthesizes neuron-specific signals, conditioned on both current inputs and the network’s evolving state, to adaptively regulate activation dynamics and synaptic plasticity. Evaluated on diverse CL benchmarks, including memorization (Random Label CIFAR-10 and Random Label MNIST), concept drift (Shuffle CIFAR-10 and Shuffle Mini-ImageNet), class-incremental learning (Class Split ImageNet and Class Split CIFAR-100), and domain-incremental learning (Permuted MNIST), NeuMoSync demonstrates strong performance in retaining plasticity and achieves improvements in both forward and backward adaptation compared with existing methods. Ablation studies validate the necessity of each component, while analysis of the learned modulatory signals reveals interpretable coordination patterns across tasks. Our work underscores the potential of integrating global coordination mechanisms into deep learning systems to advance robust, adaptive continual learning. The code is publicly available at this https URL.
[AI-63] HyPASE: Hyperbolic Geometry for Parameter-Efficient Speech Emotion Fine-Tuning Framework for Large Audio-Language Models
链接: https://arxiv.org/abs/2608.04351
作者: Tian Jin,Ruikang Zhang,Zefeng Zhao,Ding Luo,Jin Zeng
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Audio-Language Models (LALMs) excel at general speech understanding; however, adapting them to fine-grained tasks like Speech Emotion Recognition (SER) remains a significant bottleneck. Current Parameter-Efficient Fine-Tuning (PEFT) methods typically operate in flat Euclidean space, and this geometry fails to capture the multi-granularity nature of emotion cues, which range from low-level prosody to high-level semantics. To address this, we propose HyPASE, a hyperbolic PEFT framework for LALM-based SER. HyPASE leverages the Poincare ball model, using the hyperbolic radius as an explicit proxy for representational granularity. The framework consists of two core components: a Hyperbolic Geometric Adapter (HGA) for layer-adaptive weight modulation, and an Emotion-aware Multi-capacity Cross-modal Aggregator (EMCA) that compresses multi-scale features into compact audio prefixes. Empirical results on standard benchmarks show that HyPASE outperforms Euclidean PEFT baselines across all metrics on MELD and achieves a notable Unweighted Accuracy gain on IEMOCAP, particularly in class-imbalanced emotion recognition, with the accompanying slight Weighted Accuracy trade-off reflecting hyperbolic space’s geometric prioritization of minority-class representations; furthermore, HyPASE achieves robust zero-shot cross-dataset generalization within a constrained parameter budget. By grounding the adaptation process in hyperbolic geometry, HyPASE offers a highly efficient path for LALM fine-tuning.
[AI-64] COMPAS: Difficulty-Aware Joint Search for Optimizing Code Generation
链接: https://arxiv.org/abs/2608.04336
作者: Jingzhi Gong,Jie M. Zhang,Gunel Jahangirova,Dong Huang,Mohammad Reza Mousavi,Mark Harman
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Code generation systems make each LLM call with a model, a prompt, and decoding settings. However, existing optimization methods usually tune only part of these choices or use one fixed configuration for all tasks: global optimizers search one configuration for all tasks, routers choose only a model, and prompt optimizers keep the model and decoding settings fixed. This leaves their joint, group-specific interactions unclear. We therefore examine how these choices interact and observe that prompts and decoding settings interact, tuning effects vary by model, and the best configuration varies by task difficulty. Guided by these observations, we introduce COMPAS (Code-generation Optimization over Models, Prompts, And Decoding Settings), a difficulty-aware method that learns group-specific quality-cost fronts through low-cost model selection and joint prompt-decoding search, then routes each test task to its matching front online without further search. Under a matched search budget on LiveCodeBench, COMPAS improves pass@1 from 45.9% for the best baseline to 52.8% while reducing cost from 36.57 to 4.92. This also transfers to repository-level code generation on SWE-bench, resolving 76.0% of tasks versus 70.0% for the best baseline. Code and the reproducibility artifact are available at this https URL.
[AI-65] ATLAS: Adaptive Topological Learning with Abstract Successors for Continual Learning
链接: https://arxiv.org/abs/2608.04334
作者: R. Blake Lawlor,Daniel S. Brown
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Contemporary model-free reinforcement learning algorithms can achieve very high performance, but have low sample efficiency and are not robust to changes in the environment. Model-based algorithms have much higher sample efficiency, but still fail when the environment shifts. This paper introduces Adaptive Topological Learning with Abstract Successors (ATLAS) to combat these challenges. ATLAS uses a Grow When Required network with Successor Features in order to achieve high sample efficiency while also robustly tackling catastrophic forgetting. We evaluate ATLAS in spatial navigation tasks, benchmarking its performance against common on-policy and off-policy algorithms. Our empirical results demonstrate that by structurally decoupling transition dynamics from the reward signal, ATLAS achieves near-instantaneous adaptation to new goals and can exhibit positive backward transfer, significantly outperforming baseline methods in non-stationary environments.
[AI-66] Efficient Online Lexicographic Generalized Low-Rank Matrix Bandits
链接: https://arxiv.org/abs/2608.04324
作者: Bo Xue,Ji Cheng,Haodong Jing,Hongzong Li,Shuang Qiu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:This paper studies generalized low-rank matrix bandits with multiple prioritized objectives. At each round, the learner selects a matrix-valued arm and observes a vector-valued reward, whose components correspond to multiple objectives with different priority levels. Each objective is governed by an objective-specific generalized low-rank matrix model, and the learner evaluates arms according to a lexicographic preference order, prioritizing higher-level objectives before lower-level ones. We propose \textscLexi-LowGLM, an efficient online algorithm that first estimates objective-specific low-rank subspaces and then performs lexicographic learning in the reduced feature spaces. Unlike existing single-objective algorithms that repeatedly solve a batch generalized linear estimator using all historical observations, \textscLexi-LowGLM updates each objective-specific estimator via an online Newton step, reducing the estimator-update complexity over T rounds from O(T^2) to O(T) . We establish a regret bound of \widetilde O\left(W_i^\rm lex\sqrtm,(d_1+d_2)r\sqrtT\right) for each objective i\in[m] , where r is an upper bound on the ranks of the objective-specific parameter matrices and W_i^\rm lex characterizes the lexicographic trade-off effect. This bound depends on the effective low-rank dimension (d_1+d_2)r rather than the ambient dimension d_1d_2 . Numerical experiments further validate the effectiveness and computational efficiency of the proposed method.
[AI-67] he RAIL Principles for Neurosymbolic AI: Reasoning Assurances Interfacing and Learning
链接: https://arxiv.org/abs/2608.04285
作者: Agnese Chiatti,Michael Cochez,Cristina Cornelio,Sebastijan Dumancic,Artur d’Avila Garcez,Luis C. Lamb,Lia Morra,Mathias Niepert,Robert Peharz,Alberto Speranzon,Maarten Stol,Annette Ten Teije,Thiviyan Thanapalasingam,Frank Van Harmelen,Emile Van Krieken,Antonio Vergari,Benjie Wang
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Neurosymbolic AI systems that integrate machine learning and symbolic reasoning are rapidly gaining attention. They complement the data-intensive statistical approaches of neural networks and language models with symbolic reasoning algorithms to function in high-stakes domains or in low-data regimes that characterize many real-world applications. We argue that the neurosymbolic combination of machine learning and formal reasoning is not a niche approach within AI, but rather includes many already successful techniques that are of crucial importance to the development of reliable, efficient and, ultimately, trustworthy systems. This perspective prompts a re-examination of the design of current AI systems. We show that many leading AI systems, including some that are not traditionally considered as neurosymbolic, can be analysed from the perspective of four principles of neurosymbolic AI design: Reasoning, Assurances, Interfacing and Learning (RAIL). Applying the RAIL framework offers a unified view of seemingly disparate AI systems, ranging from physics-aware machine learning to neuro-guided search (such as Google DeepMind’s Alpha-* suite), causal learning and tool-augmented Large Language Models. Importantly, the RAIL principles will enable engineers to make better-informed and more principled decisions about the design and deployment of production-level AI systems. In this article, we introduce the RAIL principles, examine how they can be applied across major areas of AI, and illustrate how they may guide practitioners to integrate neurosymbolic methods into next-generation AI technologies.
[AI-68] EA-Graph: Artifact-Anchored Verification Memory for Coding Agents under Upstream Drift
链接: https://arxiv.org/abs/2608.04278
作者: Hwai-Jung Hsu,Cheng-Jan Chi,Hanna Everett
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 25 pages, 4 figures
Abstract:Coding agents increasingly work across sessions, but prose notes can preserve a conclusion without the program state that supported it. After an upstream change, a repository may still build even though earlier verification claims are no longer valid. EA-Graph is an artifact-anchored memory for verification claims. It represents artifacts at sub-path granularity, resolves aliases to leaf definitions, anchors each claim to the content used to establish it, and keeps evidence strength separate from freshness. When replacement content is unavailable, the claim becomes unprovable rather than guessed. EA-Graph is evaluated on generated repositories whose behavior-to-artifact ground truth is known by construction. The task is to classify prior claims as unaffected, affected, or unprovable after value drift, logic drift, and deliberately withheld upstream content. The analysis covers 42 sessions across seven clean worlds, 14 model-world instances, three memory conditions, and two model tiers. In the Haiku round, artifact-anchored memory outscored prose notes and no persistent memory in all seven worlds; each exact paired Wilcoxon comparison yielded p = 0.0156. In the Sonnet round, the anchored condition was perfect, but frequent control ceilings left the preregistered contrasts non-significant. No session fabricated withheld content. These results support a bounded claim: artifact-anchored memory improved the smaller model’s provability judgments in this testbed. An exploratory comparison further suggests that structured claim memory may narrow a capability gap by externalizing in-session re-derivation, but it does not establish cross- model equivalence. The study makes no claim about efficiency or repair quality. Comments: 25 pages, 4 figures Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI) Cite as: arXiv:2608.04278 [cs.SE] (or arXiv:2608.04278v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2608.04278 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-69] Interoceptive Attention as Dynamic Homeostatic Prioritization in a Forag ing Agent ALT
链接: https://arxiv.org/abs/2608.04232
作者: St John Grimbly,Nicolas Kuske,Evert A. Boonstra,Bruce A. Bassett,Charel van Hoof,Rowan Hodson,Benjamin Rosman,Ryan Smith,Mark Solms,Jonathan P. Shock
类目: Artificial Intelligence (cs.AI)
备注: Accepted at SAB 2026 (From Animals to Animats 18), forthcoming in the Springer Lecture Notes in Artificial Intelligence proceedings. 20 pages, 11 numbered figures (12 graphics), 5 tables. The 12-page camera-ready paper is reproduced without alteration and followed by supplementary analyses that were not part of the proceedings paper. Code: this https URL
Abstract:Biological systems must regulate competing needs under limited perceptual bandwidth, where sharpening one estimate costs the capacity to sharpen the others. Any fixed-budget system therefore has to decide where to allocate its perceptual precision. We study this in a foraging agent that must keep several bodily needs satisfied to survive, modelled with active inference. At each step it reads its own body-state beliefs, identifies the most-needed channel, and reallocates a fixed budget of interoceptive precision toward it, so that the same precision-shaped likelihood feeds both belief update and planning. In AffectWorld, a four-channel foraging gridworld, this selective allocation more than doubles learning-phase survival at matched budget against a uniform-precision agent ( 0.414 vs 0.199 across 11 layouts, n=32 seeds each, paired cluster-bootstrap p \leq 10^-4 ). Two further results sharpen the mechanism. The benefit runs through planning as well as perception, since denying the shaped likelihood to the planner alone removes about half of it. It is also need-aligned, since aiming precision at the least-needed channel does worse than spreading it evenly. The attended channel additionally learns its own dynamics about twice as fast, and stays ahead even at matched observation count, a behavioural trace of the same precision routing, visible in learning speed, not survival.
[AI-70] A Unified Model for Cross-Domain Clone Detection via Model Merging
链接: https://arxiv.org/abs/2608.04215
作者: Palash R. Roy,Banani Roy,Kevin A. Schneider,Chanchal K. Roy
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: Accepted at ASE 2026
Abstract:The growing diversity of code clone types, from syntactic copies to cross-language semantic clones to AI-generated duplicates, has created a fragmentation crisis in clone detection. Current deep learning detectors are domain specialists that degrade significantly outside their training distribution, with F1 drops exceeding 70% across domains. Deploying multiple specialized models is impractical, yet training a single cross-domain detector requires simultaneous access to all training data. To address this, we investigate model merging, a family of post-hoc techniques that operate solely on trained checkpoints. We evaluate parameter merging with five task-vector methods, architecture merging via greedy layer stitching, and cross-tokenizer alignment across four code models, three benchmarks, and twelve configurations. Same-base TIES merging creates effective cross-domain detectors, validated across two model families and three random seeds, reaching 0.865 combined F1 on UniXcoder, 93% of multi-task performance without any training data at the merging step. WUDI achieves the highest in-distribution combined F1 at 0.899, but TIES generalizes better to unseen AI-generated clones, making it our recommended method. Cross-base merging yields only marginal and high-variance gains across all five methods, indicating that task vector compatibility through a shared pre-trained base is the binding factor for effective merging. Merged detectors also outperform zero-shot code LLMs on GPTCloneBench at lower inference cost and generalize up to 4x better than multi-task training to unseen AI-generated clones, suggesting a trade-off between in-domain performance and OOD robustness. This work provides one of the first systematic empirical studies of model merging for software engineering and a practical recipe for building cross-domain clone detectors.
[AI-71] MatrAIx: Simulating the World with 8.3 Billion Persona Agents
链接: https://arxiv.org/abs/2608.04205
作者: Xiaomin Li,Yuexing Hao,Jianheng Hou,Jintao Huang,Qianfeng Wen,Shirley Huang,Yifan Liu,Xiaoyi Liu,Yilan Fan,Yijun Wang,Koutian Wu,Ruoqi Gao,Muhammad Ahmed Mohsin,Jing Tang,Brihi Joshi,Heming Liu,Zheyuan Deng,Zonglin Di,Sankalp Jajee,Jiuyao Lu,Zhiwei Zhang,Saksham Kapoor,Ishan Gupta,Yunhan Zhao,Chanwoo Park,Yucheng Lu,Bing Hu,Weihang Xiao,Aravind Mohan,Hanwen Xing,Runyu Zhang,Mihir Kulshreshtha,Yuanda Xu,Qianyu Zhu,Dianzhuo Wang,Yuxin Xiao,Bowen Jiang,Yongye Su,Wenhao Chai,Zuxin Liu,Lawrence Yunliang Chen,Xuandong Zhao,Ethan Ye,Shivam Patel,Jason Xie,Alex Martin Richmond,Weixiang Ding,Emre Okcular,Diya Mathew,Ziheng Wang,Rana M. Shahroz Khan,Zhejian Peng,Fang Wu,Fan Nie,Xinyang Han,Yubin Kim,Jiawei Zhang,Zhenting Qi,Huangyuan Su,Xu Pan,Abinitha Gourabathina,Hyewon Jeong,Hemanth Neelgund Ramesh,Kumail Alhamoud,Kimia Hamidieh,Zidi Xiong,Samuel Schmidgall,Pengrui Han,Yepeng Huang,Yongheng Wang,Bowen Yang,Alex Gu,Yuchu Wang,Akshay Paruchuri,Brenna Li,Hejie Cui,Jiayuan Ding,Chaosheng Dong,Jiahao Wang,Yixuan He,Chi Wang,Pamela Bhattacharya,Tianyi Peng,Paul Pu Liang,Mitchell Gordon,Yilun Du,Marinka Zitnik,James Zou,Prasanna Tambe,Philip Torr,Emily Fox,Asu Ozdaglar,Dawn Song
类目: Artificial Intelligence (cs.AI)
备注: Project website: this https URL
Abstract:Human evaluation of AI systems and digital products is costly, slow, and difficult to scale. Offline evaluations are more scalable but often abstract away human diversity and interactive behavior. We therefore introduce MatrAIx, a population-scale simulated-user evaluation infrastructure for testing AI systems and digital products with heterogeneous users. MatrAIx has three core components: First, Persona 8B contains 8.3 billion persona records represented by 1,290 categorical dimensions. Records are either sampled from a dependency graph that preserves correlated attributes or derived from human-authored profiles. We release a quality-filtered coreset of approximately 1 million personas, comprising 599,847 human-grounded and 400,000 synthetic records. Second, the MatrAIx Playground provides four environments in which diverse users evaluate and interact with digital products: Survey, AI Chatbot, Web, and App. Third, MatrAIx provides 1,010 application tasks spanning more than 25 domains, including Commerce, Software, Finance, and Healthcare. We conducted 18,189 evaluation trials across eight representative tasks. Persona agents were powered by three LLMs: Claude Opus 4.8, GPT 5.5, and Claude Haiku 4.5. The resulting feedback captures how decisions and preferences vary across persona backgrounds, including hesitation after a price increase, willingness to continue after an AI assistant fails, and latency tolerance. We conducted two main validation studies: First, a 400-trial controlled study evaluated persona adherence across ten behavioral attributes and all four environments. The declared behavior was expressed or correctly suppressed in 366 trials (91.5%). Second, human and LLM judges evaluated the extraction quality of human-grounded personas. Overall, MatrAIx provides an end-to-end infrastructure for evaluating AI systems and digital products with diverse simulated human users.
[AI-72] BrainBench: Benchmarking Large Language Models for Comprehensive EEG Understanding
链接: https://arxiv.org/abs/2608.04156
作者: Yangxuan Zhou,Sha Zhao,Yuning Chen,Chen Wu,Jiquan Wang,Shijian Li,Gang Pan
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 42pages,22pages
Abstract:Electroencephalography (EEG) analysis extends beyond assigning predefined labels to recordings; it requires workflows connecting natural-language instructions, signal processing, quantitative evidence, and scientific interpretation. We term this capability \emphcomprehensive EEG understanding. Existing evaluations, however, primarily target isolated decoding tasks or system-specific demonstrations, leaving the competence of large language models (LLMs) insufficiently quantified. We introduce \benchmarkname, a unified benchmark for comprehensive, instruction-conditioned EEG understanding. It comprises four subsets—Foundational Analysis, Sleep Assessment, Neurocognitive Assessment, and Physiological Integration—covering 17 datasets, \numcases tasks, and over \numinstances real-data instances. Given an instruction and EEG recordings with optional physiological signals, a system must perform the analysis and produce a scientifically grounded report and, when required, artifacts. Outputs are assessed through numerical, categorical, set, sequence, semantic, and artifact validation. We evaluate \nummodels representative LLMs across more than 100K executions under two paradigms: autonomous code execution with CodeAct and structured agentic analysis with BrainAgent. Results vary substantially across models, subsets, difficulty levels, and execution paradigms, showing that EEG competence depends on the model and its operationalization. \benchmarkname provides a reproducible testbed for advancing LLM-based EEG understanding. The code and benchmark will be released soon, with evaluation results continuously updated.
[AI-73] InvFlowFD: Reference-Free and Background-Set-Free Perceptual Music Quality Metric with Flow Matching Inversion
链接: https://arxiv.org/abs/2608.04142
作者: Alon Ziv,Harel Pogoda,Yossi Adi
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Existing reference-free methods for evaluating music perceptual quality alleviate the need for paired noisy-clean data, but they still rely on a background set, which is used to compute aggregated statistics of clean audio samples. In this work, we propose a novel approach that eliminates this requirement, achieving background-set-free and reference-free quality estimation using only a pre-trained Flow Matching backbone. We demonstrate that unconditional Flow Matching inversion via simple Euler integration is sufficient to detect various artificial distortions and accurately rank music generation models against human perceptual judgments. We introduce InvFlowFD, which performs flow inversion and compares a group of inverted samples to the prior distribution. We evaluate our method against prior work, quantitatively and with a thorough human study. Results suggest that InvFlowFD is highly correlated with human perception of sound distortions, as well as generative models’ quality, while being more flexible and less restrictive than existing metrics.
[AI-74] Out-Of-The-Loop Multi-Fidelity Bayesian Optimization
链接: https://arxiv.org/abs/2608.04113
作者: Gustavo Sutter,Hao Wang,Luis Ricardez-Sandoval,Pascal Poupart,Agustinus Kristiadi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Black-box optimization is a ubiquitous problem in science and engineering, often dealing with expensive objective functions with cheaper lower-fidelity proxies available. Multi-fidelity Bayesian optimization (MF-BO) is a principled approach to this problem, leveraging correlations across different fidelities when querying the objective. However, for many important MF-BO tasks, the true highest-fidelity function is prohibitively expensive to be part of the optimization loop. Nevertheless, practitioners often have gold standard data (observations of the highest-fidelity function) obtained from previous experiments that might provide information for the current task. For instance, in molecular optimization, chemists often pick the top- k candidate molecules using various computer simulations, and later reveal their true objective function values. In this work, we demonstrate the suboptimality of standard MF-BO algorithms in the real-world scenarios above, even under ideal assumptions. Next, we mitigate this problem by incorporating historical high-fidelity data accompanied by task descriptors—which can be explicitly given or extracted from unstructured metadata. We demonstrate the effectiveness of our methods on synthetic functions, as well as real-world problems in chemistry and hyperparameter optimization.
[AI-75] Spatiotemporal Graph Transformer for Traffic Intelligence in Edge Computing
链接: https://arxiv.org/abs/2608.04075
作者: Laha Ale,Letian Lin,Na Cao,Zheng Ma,Peng Yu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 12 pages, 10 figures
Abstract:Accurate traffic forecasting is essential for proactive resource management in edge computing, where service demand evolves dynamically across both space and time. In practical cellular edge systems, traffic exhibits strong spatial correlations among neighboring service regions and long-range temporal dependencies driven by user mobility and application behavior. Existing recurrent forecasting approaches can capture short-term dynamics but often struggle to model long-horizon traffic evolution under non-stationary conditions. To address this challenge, we propose a spatiotemporal graph Transformer framework that jointly models spatial interactions and temporal dependencies for traffic forecasting in edge computing. The framework employs graph neural networks to capture spatial correlations among service regions and leverages Transformer-based self-attention to learn long-range temporal patterns from historical traffic observations. By decoupling spatial representation learning from temporal reasoning, the proposed approach provides an effective mechanism for large-scale spatiotemporal traffic modeling. Extensive experiments on a real-world cellular network dataset demonstrate that the proposed graph Transformer consistently outperforms recurrent graph-based baselines, including GCN-RNN, GCN-LSTM, and GCN-GRU models, across multiple forecasting horizons. The resulting forecasts enable more effective proactive resource provisioning and reduce overload risk compared with reactive management strategies. These results highlight the potential of graph-enhanced attention mechanisms for building intelligent and adaptive edge computing systems.
[AI-76] Spend Bits Where Queries Look: KV Cache Vector Quantization with Attention-Preserving Transforms
链接: https://arxiv.org/abs/2608.04074
作者: Samuel Fernández-Menduiña,Amir Ziashahabi,Eduardo Pavez,Antonio Ortega,Salman Avestimehr
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Information Theory (cs.IT); Signal Processing (eess.SP)
备注:
Abstract:Long-context LLM decoding reads the key-value (KV) cache at every step. Loading it takes longer than computing attention over it, so throughput is bandwidth-bound. Hence, reducing the cache size can raise both decoding speed and serving capacity. The challenge is to reduce cache size while preserving the attention products, keeping reconstruction cheap, and using a fixed per-token bit count. At two bits per element, the most competitive methods rely on orthogonal transforms. However, existing techniques are either data-oblivious or use the query statistics without deriving the transform from a distortion criterion. Moreover, they rely on transforms built on top of random or Hadamard rotations, which equalize variances across entries rather than compacting energy, and fixed-width scalar quantizers, which are suboptimal at low rates. In this paper, we formulate KV cache quantization as a transform coding problem in which distortion is the error in the attention products. We derive closed-form optimal transforms for keys and values from calibration statistics, under a high-resolution model. We show that the optimal key transform is not orthogonal and satisfies a generalized Parseval relation: the attention-aware distortion becomes mean-squared error (MSE) in the transform domain. Thus, we can use MSE-optimal vector quantizers applied directly to the transformed key coefficients. To meet the fixed-width layout requirement, we show that grouping coefficients into equal-volume partitions makes equal-size codebooks attain the variable-rate optimum under the same high-resolution model. At two bits per element, our method, termed NOVA-KV, recovers most of the long-context retrieval accuracy lost by scalar quantization methods at comparable throughput.
[AI-77] FBID: Adaptive Personalized Federated Learning for Robust Out-of-Distribution Attack Detection in IoT Networks
链接: https://arxiv.org/abs/2608.04073
作者: An Khanh Bui,Cong Thanh Nguyen,Hoang-Anh Pham,Hoang Thai Dinh,Diep N. Nguyen
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Personalized Federated Learning (PFL) has emerged as a promising solution for intrusion detection in heterogeneous IoT environments, as it can improve local adaptation under highly Non-Independent and Identically Distributed (non-IID) data distributions. However, existing PFL methods often rely on client-side self-adjustment, which may lead to over-personalization and substantial degradation in out-of-distribution (OOD) attack detection. In this paper, we propose Federated Bandit Intrusion Detection (FBID), a novel adaptive PFL framework to address this limitation through server-side personalization control. In particular, FBID employs a contextual multi-armed bandit at the server to dynamically regulate each client’s local training intensity according to its observed behavior and update quality. Moreover, FBID introduces a trust-based blending mechanism to derive client-specific interpolation coefficients between the global and local models, thereby preserving global attack-detection knowledge while still allowing beneficial local specialization. Through extensive experiments on the CICIoT2023 dataset under heterogeneous client distributions and OOD stress-test settings, we show that FBID improves individual client OOD Detection Rate (DR) by up to 7.66% and F1-Score (F1) by up to 5.08% (relative) over the strongest stable baseline, while also improving robustness to previously unseen attack classes.
[AI-78] Monte Carlo Tree Search for Table-to-Multimodal Report Generation
链接: https://arxiv.org/abs/2608.04071
作者: Teng Lin,Zhiyang Zhang,Yuyu Luo,Nan Tang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automatically generating professional multimodal reports comprising both textual analysis and visual charts from structured tabular data is a critical challenge in data intelligence. Existing methods suffer from fixed linear pipelines and isolated subtask processing, which hinder joint optimization of factual accuracy, visual quality, and narrative coherence. To address these issues, this paper proposes MCTS-Report, a Monte Carlo Tree Search (MCTS)-driven framework that formulates multimodal table-to-report generation as a progressive construction process over a structured search space. The core idea is to decompose report generation into atomic actions, including chapter planning, visualization task identification, chart generation, insight organization, and narrative refinement, each executed by an LLM based on dynamic reasoning conditioned on the current report state. We use an LLM to generate step-by-step reasoning and actions during MCTS, storing the reasoning trajectory in each node for context-aware, coherent report construction. To guide the search, we design a multi-dimensional reward function that jointly evaluates numerical fact consistency (via SQL), chart quality, chart-text alignment, and structural completeness, while incorporating a diversity penalty to suppress repeated charts and a precondition check to prune invalid actions. We also construct MMRBench, a comprehensive benchmark comprising real-world tables from six domains, paired with expert-refined reference report structures and verifiable key insights. Experiments on MMRBench demonstrate that MCTS-Report significantly outperforms strong baselines across structural completeness, numerical accuracy, chart-text alignment, and insight novelty, achieving a 77.9 overall score.
[AI-79] he LLM Proposes the Executive Disposes: A Self-Verifying Agent Instrument that Dissociates Commitment Drift from Binding Drift in Long-Horizon Agents
链接: https://arxiv.org/abs/2608.04066
作者: Mohsen Arjmandi
类目: Artificial Intelligence (cs.AI)
备注: 7 pages
Abstract:How do you verify a long-horizon agent when its own state and self-reports are exactly what you cannot trust? We present an agent instrument built so that verification is structural rather than post-hoc. A deterministic Executive owns all belief; a language model may only file typed proposals, and a claim is admitted only when a prediction pre-registered before acting is matched against observation by code. Two properties make the instrument a verifier of its own science, not just of the agent: every run invalidates itself when per-organ write-error, render-size, or salted-canary-echo floors are breached (four of the first eight architecture runs were invalidated, each localizing a real defect); and a render-invisible shadow reference compiles the plan the full system would have committed in every ablation cell, so drift metrics are defined even where the mechanism under test has been removed. Using this instrument we report a clean, single-variable result on a failure every long-horizon agent suffers: ablating the commitment mechanism flips goal-abandonment from 0.00 to 1.00 while binding error stays flat at 0.00 (three seeds per cell, up to 394 reference beats per run, every run gated valid). The binding channel, by contrast, does not reappear as per-beat drift when its repair is ablated – because binding is code-owned, the failure class is structurally absorbed, its only residue appearing one layer upstream as a collapse in hypothesis formation. We report these under full disclosure that task efficacy is null (zero level completions across 52 gated runs on ARC-AGI-3), pre-registered as a structural defeater. The contribution is a verification methodology for agent development and the drift decomposition it makes measurable.
[AI-80] An Inline Control Architecture for Language Models in Intelligent Transportation Systems
链接: https://arxiv.org/abs/2608.04065
作者: Narendra Kumar Dewangan,Mounira Msahli
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 18 pages, under minor revision (IEEE transactions)
Abstract:Vehicle-to-everything (V2X) systems increasingly incorporate large language models (LLMs) for semantic tasks such as message summarization, operator assistance, and decision support at roadside units and edge nodes. Although these components are not part of safety-critical control loops, they introduce prompt-level attack surfaces that are not addressed by traditional V2X security mechanisms focused on authentication and message integrity. This paper presents Guarded-V2X, an inline semantic guardrail architecture for securing LLM-enabled V2X services under real-time constraints. The proposed system integrates rule-based ingress filtering, a lightweight safety classifier, policy-constrained structured generation, trusted-only retrieval, and post-decision adjudication to enforce machine-checkable safety boundaries prior to downstream execution. Guarded-V2X is evaluated using a four-stage experimental pipeline encompassing intrusion vulnerability analysis, calibration and latency benchmarking, guardrail validation, and robustness under adversarial stress. Experiments are conducted on a V2X-aligned simulated dataset derived from RSU advisories, operator messages, and annotated V2X message summaries. Results show that unguarded and prompt-only baselines retain residual vulnerability under multi-turn adversarial trials, while Guarded-V2X consistently reduces intrusion acceptance success rates and eliminates observed unsafe completions in two-turn settings, without exceeding latency budgets for V2X semantic advisory paths.
[AI-81] SJEPA: Learning Elegant Latent Dynamics with Hybrid Symbolic-Neural Predictors
链接: https://arxiv.org/abs/2608.04060
作者: Yongchao Huang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 42 pages
Abstract:Joint-embedding predictive architectures learn abstract states by predicting target embeddings from context embeddings, but their transition models are typically opaque neural maps. We introduce SJEPA, a reconstruction-free JEPA framework that learns predictive representations whose induced dynamics admit compact symbolic descriptions. Its hybrid transition combines a symbolic law with a regularised neural correction for dynamics outside the selected grammar. The central principle is to learn the simplest adequate dynamics: representation constraints preserve informative, non-collapsed predictive coordinates, while operator compression favours low-complexity symbolic-neural transitions that remain predictively adequate. We formalise this principle through induced-dynamics complexity, analyse predictive-coordinate non-identifiability, and show that unconstrained operator compression creates a direct shortcut to representation collapse. The framework supports both alternating representation-equation learning and symbolic dynamics fitted to fixed representations. In controlled pendulum experiments, joint learning discovers substantially simpler symbolic dynamics with lower long-horizon rollout error and divergence than post-hoc fitting, while an unconstrained one-step diagnostic realises the predicted collapse shortcut. Under grammar misspecification, correction regularisation preserves the representable symbolic mechanism and directs the neural component towards residual dynamics. The results expose a controllable trade-off among predictive fidelity, representation quality, symbolic parsimony, and symbolic-neural allocation.
[AI-82] LaPrune: Controllable Differentiable Sparsity at Million Scale
链接: https://arxiv.org/abs/2608.04057
作者: Jakub Antczak,Joanna Wojciechowicz,Łukasz Struski,Jacek Tabor
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Top- k selection determines which components of a sparse model remain active. Hard selection blocks gradients, while continuous relaxations often couple mask hardness to the selected mass. We introduce LaPrune, a mathematically exact-budget differentiable layer that controls the normalized second moment while preserving the selected mass. A LapSum barrier preserves the selection mass, and a normalized second-moment constraint moves the mask from a dense equal-mass allocation toward hard top- k at each budget. We derive a population prediction of the saturated fraction, a near-binary limiting law, and a tight worst-case guarantee on the near-zero fraction. The normalized hardness parameter is invariant to score scale, while a fixed LapSum temperature is not.
[AI-83] Agent Antibody: An Adaptive Immune System for Defending LLM Agents against Prompt Injection
链接: https://arxiv.org/abs/2608.04053
作者: Shihao Weng,Yang Feng,Xiaofei Xie,Jiongchi Yu
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 7 pages
Abstract:Prompt injection remains a critical threat to LLM agents, yet existing defenses treat each task as a self-contained problem, independent of previous encounters. In practice, user requests are often underspecified: they describe the desired outcome without fully specifying acceptable behavior. An injection can exploit this ambiguity, causing the agent to complete the task in a way the user would reject. As the user’s expectations become clearer through concrete cases, a defense should learn from each encounter and apply what it learns to the next. Inspired by adaptive immunity, we propose AgentAntibody, which equips LLM agents with a self-evolving immune system against prompt injection. AgentAntibody represents its evolving understanding of the user’s security boundary as a persistent library of antibodies. At runtime, the library recognizes threats to this boundary and mounts corresponding immune responses. Across encounters, it evolves to strengthen the agent’s immunity to future attacks. Extensive experiments across three benchmarks and four backbone LLMs show that, by learning the user’s boundary through experience, AgentAntibody outperforms existing defenses in preventing harmful actions while preserving legitimate task completion, even when the harmful and legitimate actions are both compatible with the stated task.
[AI-84] Recurrent Residual Quantization: A Progressive Multi-Precision Representation for LLM s NEURIPS2026
链接: https://arxiv.org/abs/2608.04048
作者: Yu Luo,Bo Dong,Wenhua Cheng,Haihao Shen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: NeurIPS 2026 submission; 14 tables, 1 algorithm, and no figures
Abstract:Serving large language models (LLMs) under diverse deployment constraints requires flexible trade-offs between accuracy, memory footprint, and throughput. However, conventional quantization methods typically require a separate checkpoint for each target bit-width. We introduce Recurrent Residual Quantization (RRQ), a post-training quantization (PTQ) framework that represents weights as a low-bit quantized base together with a sequence of quantized residual corrections, enabling multiple effective precisions from a single checkpoint. Starting from a 2-bit model obtained via post-training quantization (PTQ) or round-to-nearest (RTN), RRQ progressively adds lightweight 2-bit residuals generated via RTN to construct 4-, 6-, and 8-bit representations. The method is calibration-free and avoids joint multi-bit optimization. In our Qwen3-8B setup, the full all-RTN 2-/4-/6-/8-bit package is constructed in 1,293 seconds, 3.3 times faster than the measured MatGPTQ construction. Experiments on six recent LLMs show competitive accuracy at 6 and 8 bits, with model-dependent behavior at 4 bits. The code will be made publicly available upon publication.
[AI-85] Beyond the QBER Threshold: A Temporal QBER Based Machine Learning Framework for Multi Attack Detection in BB84 QKD
链接: https://arxiv.org/abs/2608.04047
作者: Isha,Deepak Singh,Devesh Kumar,S.K Pal,Praful Hambarde,Amit Shukla
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Conventional BB84 Quantum Key Distribution (QKD) systems rely on a fixed 11% Quantum Bit Error Rate (QBER) threshold to detect eavesdropping. However, stealthy attacks can remain below this threshold while still compromising channel security. This paper proposes a temporal QBER based machine learning framework for detecting and classifying eavesdropping attacks in BB84 QKD systems. Rather than relying on average session level QBER, the framework extracts 63 physics-informed temporal features capturing burst behavior, temporal instability, basis dependent asymmetry, and QBER loss interactions. Random Forest, XGBoost, and Support Vector Machine with a Radial Basis Function kernel (SVM-RBF) classifiers are evaluated on seven eavesdropping attacks and a normal channel scenario under noisy and lossy conditions. Averaged over ten independent runs, XGBoost achieves the best performance with 88.01% (0.47%) accuracy and a macro F1 score of 0.8803, while SVM-RBF performs comparably, confirming the robustness of the proposed features. Evaluated as a binary attack-versus-normal detector for comparison with conventional monitoring, a fixed 11% QBER threshold achieves only 25.82% accuracy with a False Negative Rate (FNR) of 0.8477, whereas the proposed framework reduces the FNR to 0.0198, substantially improving detection of stealthy attacks that evade threshold-based monitoring. SHapley Additive exPlanations based (SHAP) explainability shows that physics-informed temporal and channel derived features are highly discriminative for identifying eavesdropping strategies. These results demonstrate that temporal QBER driven machine learning provides an accurate, explainable, and practical framework for multi attack security monitoring in BB84 QKD systems.
[AI-86] Robust and Personalized Federated Learning for Aircraft-Engine Prognostics under Benign and Adversarial Client Heterogeneity
链接: https://arxiv.org/abs/2608.04045
作者: Chinmoy Mitra,Md. Mehedi Hasan Nipu,Mohammad Sakib Mahmood,Md. Rakibul Islam,M. F. Mridha
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Federated learning (FL) enables aircraft fleet operators to jointly train remaining-useful-life (RUL) models from engine sensor telemetry without sharing raw data. This study examines two complementary challenges: benign heterogeneity, where honest operators observe different operating conditions and fault modes, and adversarial heterogeneity, where compromised operators submit poisoned updates. We conduct a controlled, safety-oriented evaluation using a multi-task one-dimensional convolutional neural network and a structurally non-IID partition of the Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) benchmark. We compare four remedies for benign heterogeneity and evaluate five attacks against four aggregation methods, including a physically motivated sensor-value backdoor designed to mask engine degradation. Shared-representation personalization closes approximately 70% of the local-to-centralized root-mean-square-error gap, compared with 21% for proximal regularization and 10% for server-side reweighting. The backdoor achieves a 94.9% attack success rate against standard averaging while leaving clean accuracy statistically unchanged, demonstrating that accuracy alone cannot certify model safety and that attack success must be evaluated explicitly. Krum reduces attack success by an order of magnitude and is the only evaluated aggregator that withstands coordinated attackers, whereas personalization alone provides no protection. Combining personalization with robust aggregation restores robustness (2.8% attack success) with only a small accuracy cost, revealing a trade-off between robust update selection and collaborative representation learning. Results remain consistent across client counts and on a harder six-condition dataset. Code and data partitions are released for reproducibility. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) Cite as: arXiv:2608.04045 [cs.LG] (or arXiv:2608.04045v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.04045 Focus to learn more arXiv-issued DOI via DataCite
[AI-87] CheckOne: Lightweight Fault Detection and Mitigation for Vision Transformers
链接: https://arxiv.org/abs/2608.04035
作者: Mohammad Hasan Ahmadilivani,Sven-Markus Loorits,Jaan Raik
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI)
备注: Accepted at IEEE DFTS’26, 4 pages, 3 figures, 2 tables
Abstract:The wide adoption of Vision Transformers (ViTs) in safety-critical applications raises reliability concerns related to hardware faults. Algorithm-Based Fault Tolerance (ABFT) methods have emerged as lightweight and symmetric protection mechanisms for DNNs. However, they are particularly challenging for ViTs due to their significant computational requirements. This work comprehensively evaluates the reliability of ViTs, emphasizing the need for symmetric protection in their layers. Furthermore, we present CheckOne, a novel, cost-effective method for fault detection and mitigation in ViTs that significantly reduces the computational cost compared to conventional ABFT. Through extensive experiments with multiple ViTs, CheckOne mitigates critical faults by up to 26\times and achieves an average 3.8x higher performance than ABFT in ViTs.
[AI-88] Lindblad-Inspired Multi-Timescale Reservoir Computing with Separable Rotation and Dissipation
链接: https://arxiv.org/abs/2608.04028
作者: Jyotiranjan Beuria,Amit Shukla
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Neural and Evolutionary Computing (cs.NE)
备注: Under Review at IEEE Transactions on Artificial Intelligence
Abstract:Echo-state networks enable efficient temporal learning by fixing the recurrent dynamics and training only a linear readout. However, conventional reservoirs typically accommodate signal mixing, memory retention, and stability within a single random recurrent matrix. Existing structured designs improve topology, norm preservation, leakage, or depth, but generally do not provide separate modal control of reversible mixing and irreversible forgetting together with a direct global stability guarantee. We introduce a classical Lindblad-inspired multi-timescale reservoir that bridges open-system dynamical principles with structured state-space modeling. The recurrent operator is assembled from exactly discretized damped rotational modes, so rotation and decay become independent design variables governing phase mixing and memory loss. Orthogonal mode mixing preserves normality, while the decay spectrum directly determines the echo-state stability margin without post-hoc spectral-radius rescaling. We evaluate the method over ten aligned seeds against standard, leaky, deep, orthogonal, cycle, and next-generation reservoirs, together with a compact trained gated recurrent unit, across linear memory, nonlinear recurrence, chaotic forecasting, delayed logic, and real sensor calibration. Across the benchmark suite, the proposed reservoir achieves the best fixed-reservoir performance on bounded NARMA-20 and the lowest mean error on Lorenz-63, matches the strongest linear-memory result, and remains broadly competitive across broad range of benchmarks. Ablation studies show that rotation increases state diversity, whereas dissipation provides controlled forgetting and improves predictive conditioning. The resulting framework offers an interpretable recurrent architecture in which mixing, memory, and stability are explicit and independently tunable design variables.
[AI-89] A Trust-region Framework for Moment Estimation
链接: https://arxiv.org/abs/2608.04026
作者: Oluwasegun A. Somefun
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Signal Processing (eess.SP); Systems and Control (eess.SY)
备注: 20 pages, 5 figures. Submitted to TMLR
Abstract:In this paper, we develop a trust-region framework for understanding the behavior of adaptive moment estimation mechanisms, such as \textscAdam, in stochastic gradient optimization. Specifically, in this framework, the magnitude of the update step for each individual weight is constrained within a trust-region governed by a moment constraint of order p\in[2,4] . The resulting derivation then leads to a family of learning-rate mechanisms based on second-moment estimation and a normalized p -th moment estimation. When p=4 , this involves kurtosis-like estimation. The general mechanism, referred to as \textscGmake, provides a unified interpretation of normalization by moment estimation, learning-rate scheduling, spectral lowpass filtering as momentum, and operator-level spectral normalization within a common trust-region framework. Experiments on GPT2-124M trained on FineWeb-Edu and TinyStories suggest that the fourth-moment realization provides its greatest benefit when trust-region constraints are weak. As progressively stronger trust-region controls are introduced, the second-moment realization becomes increasingly competitive, often achieving slightly lower validation loss than its corresponding fourth-moment realization.
[AI-90] Governing Execution Risk in Agent ic AI Systems: A Trajectory-Guided Framework for Red Teaming
链接: https://arxiv.org/abs/2608.04018
作者: Zhihao Zhu,Yi Yang
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注:
Abstract:AI agents are increasingly embedded in organizational workflows, where they interact with external information sources and invoke digital tools to perform operational tasks. As organizations adopt such systems, a critical challenge is identifying and mitigating risks arising from malicious or untrusted external information that can steer agents toward unintended actions. Existing red-teaming approaches largely rely on fixed attack templates or final attack outcomes, providing limited visibility into how attacks unfold through multi-step reasoning and tool use. We argue that agent execution risk should be understood as a trajectory-level phenomenon. Building on this perspective, we propose TrajRed, a trajectory-guided red-teaming framework that uses execution trajectories to uncover vulnerabilities in agentic AI systems. We further develop TrajGuard, a runtime governance layer that uses high-risk trajectories discovered during red teaming to monitor and intervene in ongoing workflows. Experiments on AgentDojo across four organizational task suites show that TrajRed identifies substantially stronger vulnerabilities than fixed-template and automatic red-team baselines. Building on these vulnerability findings, TrajGuard reduces attack success across all evaluated attack methods to near zero while preserving benign task utility. Together, the results demonstrate that execution trajectories provide a practical foundation for both red teaming and risk control in agentic AI systems. This work highlights the importance of governing agent execution in organizational AI deployments.
[AI-91] AI-driven Multimodal Representation Learning for Latent Mediation Structure Discovery of Socioeconomic Disadvantage Psychosocial Factors and Cardiometabolic Multimorbidity: Insights from the All of Us Research Program
链接: https://arxiv.org/abs/2608.04016
作者: Cong Cao,Shuangge Ma
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Applications (stat.AP)
备注: 25 pages, 4 figures
Abstract:Social disadvantage is associated with multimorbidity, but the pathways linking social conditions to disease burden remain poorly understood. We developed an AI-driven multimodal mediation framework that integrates socioeconomic, psychosocial, clinical, laboratory, behavioral, and genomic data from the All of Us Research Program. Modality-specific variational autoencoders were used to derive latent representations of each data domain, and mediation analyses were subsequently performed in latent space to evaluate indirect associations between socioeconomic disadvantage, psychosocial factors, and multimorbidity. The final analytic cohort included 20,804 participants with complete multimodal data. Across 800 exposure–mediator–outcome combinations, mediation signals were concentrated within a small number of latent dimensions. The strongest indirect association linked a socioeconomic disadvantage dimension, a psychosocial vulnerability dimension, and a cardiometabolic multimorbidity dimension (NIE = 0.002517). The psychosocial dimension was characterized by poorer mental health, greater loneliness, lower social well-being, and lower health literacy, whereas the outcome dimension was associated with hypertension, diabetes, hyperlipidemia, obesity, chronic kidney disease, and heart disease. Bootstrap analyses supported the stability of the leading pathway. These findings suggest that psychosocial vulnerability was strongly represented in the dominant latent pathway linking socioeconomic disadvantage and cardiometabolic multimorbidity. More broadly, the proposed framework illustrates how AI-based representation learning can be used to investigate complex relationships across high-dimensional multimodal health data.
[AI-92] On Hamming-Lipschitz Type Stability of the Subdominant (Minmax) Ultrametric: Theory and Simple Proofs
链接: https://arxiv.org/abs/2608.04014
作者: Alokendu Mazumder,Arnab Roy,Punit Rathore
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:The subdominant (minmax) ultrametric is a canonical tree-structured summary of a dissimilarity matrix, arising equivalently as the ultrametric induced by single-linkage clustering. While its classical stability theory is usually formulated in \ell_\infty or Gromov–Hausdorff terms, such bounds are poorly suited to sparse perturbations that alter only a few pairwise distances. We develop an \ell_0 -type stability theory for this operator. Our analysis shows that sparse edits propagate only through the minimum spanning tree (MST): a pairwise ultrametric value can change only if its tree path crosses an edited edge or a cut newly exposed by an edited off-tree edge. This yields a sharp per-edit exposed-cut score and a tree-only global envelope, leading to Hamming–Lipschitz bounds on the number of ultrametric entries that can change. We also prove sharpness results showing that this dependence on tree geometry is unavoidable: under strict cut separation the tree-edge bound is attained exactly, and for off-tree edits there are explicit families in which one edited distance changes \Theta(n^2) ultrametric entries. In addition, we prove a conditional near-additivity principle for multiple edits under certified large per-edit changed regions and negligible aggregate overlap. Experiments on deep-embedding graphs show that the resulting structural scores provide useful vulnerability diagnostics for hierarchical representations.
[AI-93] C2MOE: Consistency and Complementarity-guided Mixture of Experts for Incomplete Multimodal Emotion Learning
链接: https://arxiv.org/abs/2608.04013
作者: Yuntao Shou,Tao Meng,Wei Ai,Keqin Li
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 10 pages
Abstract:Recent advances in Multimodal Emotion Recognition in Conversations (MERC) highlight its reliance on complete multimodal inputs. However, real-world data often suffer from missing modalities due to transmission errors or user behavior, severely degrading model performance. Existing methods enhance robustness via cross-modal consistency learning but largely ignore modality complementarity, leading to biased reconstructions. To address this limitation, we propose C 2 MOE, a novel Consistency and Complementarity-guided Mixture of Experts framework for incomplete multimodal emotion learning. Our approach unifies representation learning and missing modality imputation within a principled information-theoretic framework. Specifically, multimodal knowledge is factorized into consistency and complementarity components via interaction-aware experts. Consistency is captured by maximizing cross-modal predictability, while complementarity is preserved by maximizing conditional entropy between modalities. Building upon this decomposition, C 2 MOE introduces a dual-branch prediction mechanism for robust imputation under missing modalities. The consistency branch aligns imputed features with the joint distribution by minimizing uncertainty, and the complementarity branch exploits modality-unique cues via entropy maximization. Finally, C 2 MOE employs a learnable reweighting module that dynamically assigns importance scores to each expert’s output, yielding a robust and adaptive fusion for imputation. Extensive experiments on multiple MERC benchmarks demonstrate that C 2 MOE consistently surpasses state-of-the-art methods across various missing-modality settings, validating its robustness and generalization.
[AI-94] A Long-Run Persistence Theory for AI Systems under the Redundancy-Adjusted Artificial Age Score (AAS)
链接: https://arxiv.org/abs/2608.04012
作者: Seyma Yaman Kayadibi
类目: Artificial Intelligence (cs.AI)
备注: 38 pages, no figures, theoretical paper with theorems and proofs
Abstract:Artificial intelligence systems are increasingly expected to operate over repeated cycles of interaction, adaptation, and update rather than through isolated one-shot outputs. This raises a fundamental theoretical question: can an AI system persist indefinitely without incurring unbounded structural aging? This paper develops a long-run persistence framework for AI systems based on the redundancy-adjusted Artificial Age Score (AAS). The model extends AAS from a static evaluative measure into a cycle-level functional that generates an age sequence across repeated operation. At each cycle, structural age is defined through a weighted, redundancy-aware logarithmic penalty over component consistency levels. Within this framework, cycle-level age is shown to be well defined and uniformly bounded, thereby excluding explosive pointwise aging. On this basis, the paper defines a hierarchy of asymptotic regimes, including burdened persistence, zero-burden persistence, oscillatory persistence, and cumulative terminal burden. It also establishes comparative ordering, sensitivity bounds, convergence under componentwise stabilization, persistence under finite total variation, geometric stabilization under damped inter-cycle perturbations, and a zero-burden characterization under nondegenerate redundancy conditions. The main result is that indefinite cyclic continuation does not require unbounded structural aging: an AI system may pass through infinitely many cycles while its structural age remains bounded, while under stronger regularity conditions its marginal aging vanishes and, in the strongest regime, its cycle-level burden converges to zero. The framework thus provides a formal basis for analyzing long-run artificial persistence as a problem of bounded structural burden rather than inevitable cumulative deterioration.
[AI-95] owards a New Grammar of Reasoning for Artificial Legal Intelligence and the Mecelle as Its Semantic Protocol
链接: https://arxiv.org/abs/2608.04011
作者: Ali Goksu,F. Gozde Kardes,Mustafa Yaylali
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注: 123 pages
Abstract:This article examines the enduring epistemic and methodological crisis of traditional legal practice in light of the opportunities and constraints introduced by artificial intelligence. It proposes an ontologically grounded framework termed the Mecellem semantic protocol as a response to this crisis. The analysis focuses on the structural tension within law between maintaining normative coherence and adapting to evolving social and institutional conditions, and shows why approaches based solely on codification, positivist systematization, or quantitative methods such as jurimetrics are insufficient. The article argues that legal reasoning cannot be reduced to data retrieval or statistical pattern recognition. Instead, it is grounded in the premise that meaning is context-dependent and must be dynamically reconstructed through ontologically defined entity categories and differentiated layers of knowledge. Within this perspective, Mecellem reconceptualizes law not as a fixed system of rules, but as an ontodynamic architecture structured along the interconnected axes of ontology, epistemology, and methodology. The transition from jurimetrics to a semantic protocol is presented not merely as a technical shift, but as a transformation in the foundations of legal knowledge. The article further argues that neurosymbolic systems, knowledge graphs, and agentic artificial intelligence can effectively address persistent legal challenges only when embedded within such an ontodynamic framework. By understanding law as a domain of ongoing formation rather than a completed rational totality, Mecellem advances a context-sensitive, auditable, and coherent model for legal meaning production at both human and machine levels, offering a comprehensive framework for rethinking legal reasoning in the age of artificial intelligence. Comments: 123 pages Subjects: Computers and Society (cs.CY); Artificial Intelligence (cs.AI) Cite as: arXiv:2608.04011 [cs.CY] (or arXiv:2608.04011v1 [cs.CY] for this version) https://doi.org/10.48550/arXiv.2608.04011 Focus to learn more arXiv-issued DOI via DataCite
[AI-96] Informational Frustration in Neural Manifolds: Shannon Bottlenecks and the Limits of Learnability
链接: https://arxiv.org/abs/2606.30512
作者: Srinivasa Rao P.,Vangmayi P Reddy
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computational Complexity (cs.CC); Computational Geometry (cs.CG); Machine Learning (stat.ML)
备注: 8
Abstract:Why overparameterised deep networks generalise so remarkably well remains one of the most stubborn open questions in machine learning theory. Classical frameworks like VC dimension and Rademacher complexity predict catastrophic overfitting in modern models, leaving a massive theoretical gap between theory and reality. In this paper, we bridge this divide by introducing a unified framework that links information theory, topology, and statistical mechanics to map the hard limits of deep learning. Central to our approach is the Entropic Learnability Horizon (ELH): a fundamental law stating that a network can only truly learn a target function if the Shannon entropy of the data manifold outpaces the topological entropy of the function’s decision boundary, balanced by the von Neumann entropy of the network’s weight space. We establish the Shannon-Topological Bottleneck Theorem, proving that when a target boundary’s geometric complexity exceeds this informational horizon, the system undergoes a sudden entropic phase transition. It falls into a state of Informational Frustration - a glassy, rigid memorization phase where generalization becomes thermodynamically impossible. Using this lens, we show that the enigmatic phenomenon of “grokking” is actually an Entropic Release, where weights abruptly reorganise to unlock the bottleneck. Finally, we translate this theory into practice with Entropic Gradient Descent (EGD), an optimization algorithm that dynamically manages weight entropy to keep learning on track. Ultimately, this work repositions entropy not just as a tool for tracking uncertainty but as the fundamental physical currency that dictates whether a machine can learn.
[AI-97] mporal Context Awareness: A Defense Framework Against Multi-turn Manipulation Attacks on Large Language Models
链接: https://arxiv.org/abs/2503.15560
作者: Prashant Kulkarni,Assaf Namer
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 6 pages, 2 figures, IEEE CAI
Abstract:Large Language Models (LLMs) are increasingly vulnerable to sophisticated multi-turn manipulation attacks, where adversaries strategically build context through seemingly benign conversational turns to circumvent safety measures and elicit harmful or unauthorized responses. These attacks exploit the temporal nature of dialogue to evade single-turn detection methods, representing a critical security vulnerability with significant implications for real-world deployments. This paper introduces the Temporal Context Awareness (TCA) framework, a novel defense mechanism designed to address this challenge by continuously analyzing semantic drift, cross-turn intention consistency and evolving conversational patterns. The TCA framework integrates dynamic context embedding analysis, cross-turn consistency verification, and progressive risk scoring to detect and mitigate manipulation attempts effectively. Preliminary evaluations on simulated adversarial scenarios demonstrate the framework’s potential to identify subtle manipulation patterns often missed by traditional detection techniques, offering a much-needed layer of security for conversational AI systems. In addition to outlining the design of TCA , we analyze diverse attack vectors and their progression across multi-turn conversation, providing valuable insights into adversarial tactics and their impact on LLM vulnerabilities. Our findings underscore the pressing need for robust, context-aware defenses in conversational AI systems and highlight TCA framework as a promising direction for securing LLMs while preserving their utility in legitimate applications. We make our implementation available to support further research in this emerging area of AI security. Comments: 6 pages, 2 figures, IEEE CAI Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2503.15560 [cs.CR] (or arXiv:2503.15560v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2503.15560 Focus to learn more arXiv-issued DOI via DataCite
[AI-98] Representational separation between unitary and channel quantum generative models via shared classical randomness at shallow depth
链接: https://arxiv.org/abs/2608.05110
作者: Arunava Majumder,Marius Krumm,Hendrik Poulsen Nautrup,Hans J. Briegel
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
备注: 33 pages, 9 figures
Abstract:Near-term quantum hardware limits circuit depth and often imposes geometrically local connectivity for quantum generative models, restricting the output distributions accessible to shallow unitary Born models. Introducing stochasticity into a unitary quantum Born model can improve the empirical generative performance of the resulting channel model and, for a restricted small-scale architecture, has been proven to represent a strictly larger family of distributions than its unitary counterpart. However, whether such randomness provides a provable separation at fixed shallow depth for arbitrarily large systems has remained open. Here, we show that shared classical randomness, a comparatively weak resource from entanglement theory, is sufficient to establish such a strict scalable representational separation over the corresponding shallow unitary Born model. More specifically, we augment bounded-connectivity shallow unitary circuits, followed by computational-basis measurements, with spatially separated local Pauli operations, whose joint application is controlled by a single classically sampled random bit. The resulting shallow-depth channel model generates long-range correlations in the classical output distribution that no purely unitary shallow-depth model with bounded connectivity can reproduce. For one-dimensional nearest-neighbour architectures, reproducing such distributions with a purely unitary model can require depth \Omega(N) in the worst case. We further show that measurement-based quantum computation (MBQC) provides a natural implementation of the required shared classical randomness through suitable adaptation of the random measurement outcomes. Numerical experiments on MBQC-based generative models support the analytical results.
[AI-99] Revealed Rationality: Label-Free Evaluation and Regularization from Representation Theorems
链接: https://arxiv.org/abs/2608.05015
作者: Isaiah Andrews
类目: Theoretical Economics (econ.TH); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 20 pages
Abstract:Representation theorems in decision theory establish that behavior satisfies certain axioms if and only if it can be rationalized by a well-defined objective. I argue that this ``if and only if’’ structure provides a potentially useful foundation for label-free evaluation and regularization of LLMs and other AI systems. Axiom compliance can be checked from the model’s own responses to synthetic choice problems, with no external labels or human feedback, and the penalties are readily computable. Because the axioms are necessary and sufficient, the resulting checks exhaust the implications of the relevant rationality standard for the elicited data: a model that passes cannot be rejected on rationality grounds by any further test of the same data. I discuss three instantiations: probabilistic coherence via a theorem of de Finetti, preference rationality via Afriat’s theorem, and subjective expected utility via a theorem of Echenique and Saito (2015), each yielding a continuous penalty that is zero whenever behavior can be rationalized. Since coherence does not restrict which objective rationalizes behavior, these penalties complement rather than replace other evaluation and training signals.
[AI-100] ourSynbio-Search: A Large Language Model Driven Agent Framework for Unified Search Method for Protein Engineering
链接: https://arxiv.org/abs/2411.06024
作者: Yungeng Liu,Zan Chen,Yu Guang Wang,Yiqing Shen
类目: Quantitative Methods (q-bio.QM); Artificial Intelligence (cs.AI)
备注:
Abstract:The exponential growth in protein-related databases and scientific literature, combined with increasing demands for efficient biological information retrieval, has created an urgent need for unified and accessible search methods in protein engineering research. We present TourSynbio-Search, a novel bioinformatics search agent framework powered by the TourSynbio-7B protein multimodal large language model (LLM), designed to address the growing challenges of information retrieval across rapidly expanding protein databases and corresponding online research literature. The agent’s dual-module architecture consists of PaperSearch and ProteinSearch components, enabling comprehensive exploration of both scientific literature and protein data across multiple biological databases. At its core, TourSynbio-Search employs an intelligent agent system that interprets natural language queries, optimizes search parameters, and executes search operations across major platforms including UniProt, PDB, ArXiv, and BioRxiv. The agent’s ability to process intuitive natural language queries reduces technical barriers, allowing researchers to efficiently access and analyze complex biological data without requiring extensive bioinformatics expertise. Through detailed case studies in literature retrieval and protein structure visualization, we demonstrate TourSynbio-Search’s effectiveness in streamlining biological information retrieval and enhancing research productivity. This framework represents an advancement in bridging the accessibility gap between complex biological databases and researchers, potentially accelerating progress in protein engineering applications. Our codes are available at: this https URL
[AI-101] AutoProteinEngine: A Large Language Model Driven Agent Framework for Multimodal AutoML in Protein Engineering
链接: https://arxiv.org/abs/2411.04440
作者: Yungeng Liu,Zan Chen,Yu Guang Wang,Yiqing Shen
类目: Quantitative Methods (q-bio.QM); Artificial Intelligence (cs.AI)
备注:
Abstract:Protein engineering is important for biomedical applications, but conventional approaches are often inefficient and resource-intensive. While deep learning (DL) models have shown promise, their training or implementation into protein engineering remains challenging for biologists without specialized computational expertise. To address this gap, we propose AutoProteinEngine (AutoPE), an agent framework that leverages large language models (LLMs) for multimodal automated machine learning (AutoML) for protein engineering. AutoPE innovatively allows biologists without DL backgrounds to interact with DL models using natural language, lowering the entry barrier for protein engineering tasks. Our AutoPE uniquely integrates LLMs with AutoML to handle model selection for both protein sequence and graph modalities, automatic hyperparameter optimization, and automated data retrieval from protein databases. We evaluated AutoPE through two real-world protein engineering tasks, demonstrating substantial performance improvements compared to traditional zero-shot and manual fine-tuning approaches. By bridging the gap between DL and biologists’ domain expertise, AutoPE empowers researchers to leverage DL without extensive programming knowledge. Our code is available at this https URL.
机器学习
[LG-0] he Loss Does Not See the Basis but Adam Does
链接: https://arxiv.org/abs/2608.05136
作者: Devender Singh
类目: Machine Learning (cs.LG)
*备注: 22 pages main text + appendices, 5 figures. Code, seeds, and raw run records: this https URL
Abstract:Gradient descent on a factored model W = UV^\top is implicitly biased toward low-rank solutions, while Adam, starting from the same small initialization, is not. We trace the difference to the gauge symmetry of the loss, its invariance under (U, V) \mapsto (UQ, VQ) . Gradient flow’s low-rank mechanism is available to an optimizer only if that optimizer is gauge-equivariant, a condition necessary for the transfer but not sufficient for low-rank recovery. Gradient descent, momentum, “shared-scalar” Adam, Muon, and Shampoo satisfy it. Adam, RMSProp, and the other coordinate-wise methods do not. A structure theorem characterizes the memoryless equivariant rules as exactly the Gram-determined left preconditioners, and a transfer theorem carries gradient flow’s pathwise properties to common-scalar flows. We then sort nine update rules on underdetermined matrix sensing by recovery error against the planted ground truth. A one-parameter family from coordinate-wise to shared-scalar preconditioning restores the bias monotonically, isolating anisotropy as the cause. A “spectral schedule” reconciles two opposing reports about Muon: equal-rate updates recover exactly low-rank targets but lose their edge as the spectral tail grows. In transformers, Adam separates two gauge-equivalent initializations at the first step, where the equivariant optimizers stay at float precision, and ends with the per-head invariants W_Q^\top W_K 56% apart in relative Frobenius distance, a gap no per-head rotation can close. On two hyperspectral datasets at matched training loss, gradient descent cuts held-out error by 43-44% at the lowest sampling density, and at lower effective rank. Basis choice is therefore not a tuning detail but a decision about which interpolant the optimizer selects.
[LG-1] DASyR-LLM : Domain-Aware Symbolic Regression with LLM s for Kinetic Model Discovery
链接: https://arxiv.org/abs/2608.05120
作者: Roberto Aliaga Medina,Paulina Quintanilla,Antonio del Rio Chanona
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE); Symbolic Computation (cs.SC)
*备注:
Abstract:Kinetic model discovery is a central challenge in chemical engineering, as accurate rate expressions are essential for understanding and controlling chemical and biological processes. Symbolic regression (SR) has emerged as a powerful data-driven approach for identifying interpretable kinetic models, but usually operates without domain knowledge, often exploring physicochemically implausible models. Large language models (LLMs) offer a promising avenue for injecting domain expertise into this search. Here, we introduce an LLM-guided SR framework, embedding an LLM module within an iterative SR algorithm for automated kinetic model discovery. The LLM performs two roles at each iteration: (1) a qualitative physicochemical critique of the best SR candidates, and (2) the proposal of new candidate rate expressions guided by the SR-generated models and embedded chemical knowledge. Our framework is evaluated on four in silico case studies of increasing complexity, spanning heterogeneous catalysis and bioprocess systems. Results show the LLM-guided framework reduces iterations to identify the ground-truth model by 41.7-79.3% versus a state-of-the-art SR framework, with the LLM directly proposing the correct model structure in over half of the guided runs. In practical settings, where each iteration typically requires a new wet-lab experiment, this translates into a substantial reduction in experimental effort. Predictive performance on an independent validation set is equivalent between both approaches, with R^20.98 in all case studies. Ablation studies indicate that both the SR component and the LLM scale contribute to this performance, with a reduced-size LLM largely retaining discovery efficiency. These findings demonstrate that LLMs can effectively inject domain knowledge into scientific model discovery, paving the way toward fully automated, domain-aware kinetic modelling pipelines.
[LG-2] Reward Structure Shapes the Interaction Between Episodic Exploration and Neural Memory in Reinforcement Learning
链接: https://arxiv.org/abs/2608.05111
作者: Jai Malegaonkar,Rohan Patil,Henrik I. Christensen
类目: Machine Learning (cs.LG)
*备注:
Abstract:In partially observable reinforcement learning, agents face a dual bottleneck: they must explore to encounter rewarding states and retain that experience in memory to optimize their policies. Exploration bonuses and memory architectures are traditionally evaluated in isolation, leaving their interaction unmeasured, and standard notions of sparse reward conflate temporal signal density with what the reward actually supervises. We present a controlled study crossing episodic exploration bonuses with diverse neural memory architectures across three environments that vary how the content of memory is acquired. An identical bonus signal yields three distinct interaction patterns: it amplifies architectural capacity differences where memory content must be actively discovered and retained unsupervised; equalizes architectures to a shared ceiling where the content, once sought out, is a single reward-supervised cue; and is null where the observation stream is purely scheduled. Controlled reward manipulations verify that these patterns track reward structure rather than density: a dense reward neutralizes a bonus only if it directly supervises the required latent memory, and a small avoidable penalty on exploratory actions (leaving the optimum unchanged) induces policy convergence to suboptimal stationary states, which either bonus resolves. We then formalize reward sparsity with observation-anchored reward machines, separating structural sparsity (an automaton reproduces the return without the task-required history) from potential sparsity (the one-step reward misprices local exploratory actions); the resulting vocabulary organizes the three regimes by the retention burden each task exposes. Together, these results show exploration and memory are complements, not substitutes: a bonus induces exposure, and only memory converts exposure into return.
[LG-3] BnBERT-iPET: Sparse Few-Shot Language Modeling for Bengali via Lottery Ticket Pruning
链接: https://arxiv.org/abs/2608.05104
作者: Sajib Hossain,Md Kamrus Samad,Anan Ghosh,Labib Imam Chowdhury,Nabeel Mohammed
类目: Machine Learning (cs.LG)
*备注: 14 pages, 9 tables, 13 figures. Preprint
Abstract:Deep neural networks have shown impressive success in NLP tasks owing to their complex structure and huge number of edges. Achieving state-of-the-art performance in natural language processing with a large pre-trained model such as BERT is expensive and time-consuming, carries a large carbon footprint, and is difficult to realize on machines with minimal computational capability. This creates a barrier to training complex models for resource-constrained languages such as Bengali. However, in a complex neural model, not all edges are equally impactful, and the contributions of some of them can be neglected. Pruning promises to reduce the memory footprint of regular networks, shorten the training time of ever-growing networks, and increase inference efficiency without sacrificing comparable performance. In this work, we introduce BnBERT-iPET, a sparse few-shot language modeling approach for Bengali, and experimentally show that a lightweight few-shot-learned language model retaining only 10% of the edges of an initial model such as BERT can perform neck and neck with much larger models on challenging tasks for a resource-constrained language such as Bengali. By learning from few shots through iterative pattern exploiting training and achieving 90% sparsity with the Lottery Ticket Hypothesis pruning technique, our pruned BnBERT-iPET model proves to be a tough competitor to state-of-the-art language models such as Bangla Electra, Indic-BERT, and XLM-RoBERTa on downstream tasks over standard benchmark datasets of the Bengali language.
[LG-4] Multimodal Spatiotemporal Atmospheric Data Assimilation with Latent Flow-matching
链接: https://arxiv.org/abs/2608.05103
作者: Dibyajyoti Chakraborty,Romit Maulik
类目: Machine Learning (cs.LG); Mathematical Physics (math-ph); Atmospheric and Oceanic Physics (physics.ao-ph); Fluid Dynamics (physics.flu-dyn)
*备注:
Abstract:Data assimilation (DA) uses Bayesian inference to update the state of a numerical forecast model with observed data. In this study, we propose a fundamentally different, unified approach to atmospheric data assimilation. We use latent video flow-matching to sample temporally consistent trajectories from a prior trained using ERA5 reanalysis (69 variables over an 8-day window). We also use posterior sampling to assimilate real observation sources, such as those from the NOAA Integrated Global Radiosonde Archive and the Integrated Surface Database. Because the prior generates a continuous trajectory, it naturally propagates information between observed and unobserved frames. Therefore, we can perform various DA tasks, such as filtering and smoothing, simply by changing the observed frames. Moreover, we generate full-state ensemble forecasts directly from sparse observations, achieving performance competitive with state-of-the-art observation-to-forecast models.
[LG-5] MALT: Lightweight Curvature-Aware Muon via Diagonal Preconditioning
链接: https://arxiv.org/abs/2608.05088
作者: Tongle Wu,Huanyu Dong,Ying Sun,Ziye Ma
类目: Machine Learning (cs.LG)
*备注:
Abstract:Muon has recently emerged as a promising alternative to AdamW for language model pretraining by orthogonalizing momentum matrices using Newton-Schulz iterations. Although Muon mitigates gradient anisotropy, it does not explicitly account for the curvature geometry of the loss landscape and may therefore remain sensitive to curvature anisotropy. We bridge this gap by proposing MALT (Muon Augmented by Lightweight Two-sided Preconditioning), which uses lightweight diagonal preconditioners to reduce the sensitivity of Muon to curvature anisotropy. Specifically, MALT uses two-sided diagonal preconditioners with low memory and computational overhead to approximately capture the curvature geometry of the loss landscape. It orthogonalizes the preconditioned momentum using Newton-Schulz iterations and maps the result back to define the update direction, while norm grafting controls the update magnitude. To improve the robustness of MALT to stochastic gradient noise, we further propose MALTER (MALT with Adaptive stEpsize Rescaling). Convergence guarantees are provided for MALT in the stochastic non-convex setting. Experiments on GPT-2 Small, Medium, and Large pretraining show that the proposed methods outperform Muon while maintaining nearly the same memory footprint and wall-clock time.
[LG-6] Learning When to Stop: Prefix-Optimal Dynamic Diffusion Policies for Continuous Control
链接: https://arxiv.org/abs/2608.05084
作者: Rohit Kumar Salla,Manoj Saravanan,Simon Stepputtis
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注:
Abstract:Diffusion policies are a powerful policy class for continuous control, but their iterative denoising process creates a substantial computational bottleneck. Reducing this cost requires adapting the number of denoising steps to the difficulty of each action while preserving task performance. We introduce Prefix-Optimal Generative Policies (POGP), a framework that learns a prefix value function at every intermediate denoising step through a Bellman-style recursion over the denoising chain. The prefix value function serves two purposes: it provides an auxiliary training objective that encourages intermediate outputs to become high-quality actions, and it enables a test-time stopping rule that terminates denoising when additional steps are unlikely to produce meaningful improvement. Across four MuJoCo environments and comparisons with 12 baselines, POGP reduces the required number of denoising iterations by approximately 2.7-fold while retaining near-full task performance. Compared with state-of-the-art dynamic diffusion baselines, prefix training also improves final task performance by approximately 3.5%. These results indicate that supervising intermediate denoising steps is useful not only for adaptive early stopping, but also as an auxiliary objective that improves the learned policy.
[LG-7] SparseDitto: Customizing GPU Kernels for Different Sparsity Patterns with LLM -Based Agent ic System
链接: https://arxiv.org/abs/2608.05033
作者: Shiyang Li,Guangyan Sun,Jinwei Tang,Yanzhi Wang,Mingyi Hong,Caiwen Ding
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注:
Abstract:Sparse matrix kernels are fundamental to scientific computing, graph analytics, and machine learning. Their GPU performance depends strongly on the input sparsity pattern and execution strategy. For the same SpMM on the same matrix, cuSPARSE exhibits a 350x performance gap between CSR and Blocked-ELL. Our study of multiple data formats, specialized systems, and sparse compilers shows that no single implementation consistently dominates across sparsity patterns and operators. This motivates a system that can adapt its representation, execution strategy, and hardware mapping to each workload and target GPU. We present SparseDitto, an LLM-based system that constructs a GPU kernel for each matrix, operator, and target GPU. SparseDitto supports SpMV, SpMM, and SpGEMM within a unified design framework. A lightweight additive model ranks established strategies using structural features of the input matrix. An architecture-aware planner then proposes several candidate designs. Coding and verification agents implement and refine them using measurements from the target GPU. Across three sparse operators and a diverse set of matrices, SparseDitto achieves a geometric-mean speedup of 2.68x over cuSPARSE on an NVIDIA RTX PRO 6000 GPU, with a maximum of 146.61x. On an NVIDIA H200 GPU, it achieves 2.79x, with a maximum of 78.5x. Its generated SpMM kernels also accelerate full-batch GCN training by up to 3.39x. Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG) Cite as: arXiv:2608.05033 [cs.DC] (or arXiv:2608.05033v1 [cs.DC] for this version) https://doi.org/10.48550/arXiv.2608.05033 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-8] Canonical Joint Energy-Based Model on CIFAR-10: failure modes and practical indistinguishability of Predictor-Corrector and SGLD samplers
链接: https://arxiv.org/abs/2608.05025
作者: Dmytro Knopov
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 17 pages, 4 figures. Under review at Discover Computing
Abstract:Joint Energy-Based Models (JEM) unify classification and generation within a single network and support out-of-distribution (OOD) detection. Canonical JEM training relies on stochastic gradient Langevin dynamics (SGLD); a theoretically motivated alternative, the Predictor-Corrector (PC) sampler, has not previously undergone a systematic replication test on the canonical model. We reproduce canonical JEM on WideResNet-28-10 without normalisation layers on two independent runs and test whether PC retains its theoretical advantage without an annealed noise schedule, across three protocols: PC replacing SGLD throughout the roughly 130 training epochs; cold-start generation (FID); and refinement-style multi-OOD detection (AUROC). The reconstruction reaches 92.88% test accuracy and buffer-FID 44.46 (canonical: 92.9% and 38.40). We document two failure modes: catastrophic late-training divergence via the canonical outlier-buffer mechanism (both SGLD runs and, with the same signature, both PC runs), and run-dependent SVHN OOD-discrimination dynamics. No method-level advantage of PC over SGLD is observed on any protocol: at inference the absolute AUROC difference stays below 0.007 across all ten checkpoint-OOD pairs and the FID difference below 0.5; on the training protocol a hierarchical seed-by-image bootstrap gives a 95% confidence interval on the macro-averaged AUROC difference that contains zero, while a seed-level equivalence test with two runs per method cannot establish formal equivalence. The data are consistent both with equivalence and with a small directional effect. This practical indistinguishability is theoretically expected: under fixed noise the PC predictor step degenerates by construction, so its guarantees do not transfer to canonical JEM.
[LG-9] Link prediction on multi-relational graphs from an influence propagation perspective
链接: https://arxiv.org/abs/2608.05016
作者: Zidu Yin,Yuankai Qi,Dong Gong,Ehsan Abbasnejad,Kun Yue,Javen Qinfeng Shi
类目: ocial and Information Networks (cs.SI); Machine Learning (cs.LG)
*备注: Accepted for publication in Pattern Recognition
Abstract:Predicting the existence and type of links (edges) between nodes in a multi-relational graph is key for applications from social interaction prediction to knowledge relationship identification. Enhancing local features with relevant global information is crucial for accurate link prediction, yet it remains challenging. We address this by modeling the relationship between node pairs as node influence. That is, whether the node influence can be propagated and what type of influence is propagated indicates where and what type the edge is, which will be the most relevant local and global information to predict the edges. To this end, we extend the Susceptible-Infectious-Recovered (SIR) epidemic model to capture the influence propagation of nodes on a large scale through sub-graph structures. Subsequently, these sub-graphs are compressed using virtual edges, thereby substantially reducing the computation associated with utilizing the global graph structure. Finally, we propose the Influential Graph Neural Predictor, referred to as IGNP, a link prediction framework guided by influence propagation. Extensive experiments demonstrate the superiority of the proposed method, which outperforms strong baselines by a large margin on the widely used and real-world datasets.
[LG-10] Stochastic Emulation using Generalized Stratified Sampling for Performance-Based Risk Optimization of Structures
链接: https://arxiv.org/abs/2608.05006
作者: Isabela D. Rodrigues,Seymour M. J. Spence,Henrique M. Kroetz,André T. Beck
类目: Machine Learning (cs.LG); Computation (stat.CO); Machine Learning (stat.ML)
*备注:
Abstract:Metamodels are instrumental in reducing the computational burden associated with nested reliability analyses and optimization loops in Performance-Based Risk Optimization (PBRO) of structures under stochastic loads. In this context, stochastic emulators are particularly useful because they approximate response distributions while accounting for the intrinsic stochasticity of the simulator. Among these methods, Stochastic Polynomial Chaos Expansion (SPCE) is especially attractive because it does not require replications of nonlinear analyses at fixed input conditions. However, SPCE may present limitations in accurately representing extreme responses in the tails of structural response distributions. To address this limitation, this study proposes a framework that combines Generalized Stratified Sampling (GSS) with SPCE. The GSS scheme partitions the input space into strata according to the intensity of the hazard, improving the representation of extreme responses, while independent SPCE emulators are trained within each stratum. The conditional exceedance probabilities estimated in each stratum are then recombined using the total probability theorem to evaluate the probabilistic constraints. The proposed GSS-SPCE framework is applied to the optimal design of buckling-restrained brace cross-sectional areas in a two-story steel building. The objective is to minimize the initial construction cost while satisfying prescribed probabilistic performance constraints. Results show that the proposed framework accurately estimates structural response distributions, including their tail regions, while substantially reducing the number of nonlinear model evaluations required for PBRO.
[LG-11] EvolveNet: Collaborative Harness Evolution for Agent Self-Improvement
链接: https://arxiv.org/abs/2608.04968
作者: Jun Nie,Yonggang Zhang,Qianshu Cai,Yiu-ming Cheung,Xinmei Tian,Bo Han
类目: Machine Learning (cs.LG)
*备注: 20 pages, 3 figures
Abstract:The capabilities of an LLM agent depend not only on its model but on the harness: the executable program that constructs context, invokes tools, verifies results, and recovers from failure. Recent work shows that evolving the harness yields persistent improvements without updating model weights. Existing approaches, however, assume that all execution experience can be routed to a single optimizer, which evolves one harness along a sequential trajectory. Real agent ecosystems violate that assumption: users, organizations, and environments generate isolated streams of experience that cannot be pooled, so the experience most worth learning from is exactly the experience that cannot be directly centralized. We introduce EvolveNet, a paradigm of collaborative harness evolution that moves experience extraction to the data. A shared harness is broadcast to data-local agent deployments, each of which evolves it on its own workload. Only the resulting program adaptations are composed into an updated shared harness and redistributed, so that every participating agent inherits operational experience discovered by the others. By shifting the aggregation boundary from raw workloads to learned adaptations, EvolveNet keeps workloads local and allows multiple evolutionary searches to proceed concurrently with reduced serial depth. Because independently modified programs cannot be averaged like model parameters and may conflict when composed, EvolveNet introduces scope-typed, evidence-guided program aggregation. Across five settings spanning text-to-SQL, data-science coding, competitive programming, software engineering, and agentic workflows, EvolveNet improves the shared harness in all five, with the largest gains under heterogeneous workloads, and ablations attribute the improvement to composition of adaptations from different agents rather than to selecting among them.
[LG-12] A geometry-based deep equilibrium model for image restoration under multiplicative Gamma noise
链接: https://arxiv.org/abs/2608.04944
作者: Shengkun Yang,Luca Ratti,Zhichang Guo
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG)
*备注:
Abstract:We propose a deep learning framework for image restoration from images degraded by both multiplicative Gamma noise and blur. Unlike conventional deep equilibrium (DEQ) models that rely on implicit neural regularization, the proposed method learns an explicit and interpretable regularizer parameterized by geometric priors associated with surface area and mean curvature. To minimize the resulting variational model, we develop a mirror descent algorithm tailored to the commonly used Gamma-noise fidelity terms. Leveraging the Kurdyka-Lojasiewicz property for functions defined in o -minimal structures, we establish the global convergence of the generated iterates to a critical point. Experimental results on both grayscale and color image restoration demonstrate that the proposed method consistently outperforms representative model-based approaches while achieving performance comparable to state-of-the-art DEQ models based on implicit regularization, despite requiring substantially fewer trainable parameters.
[LG-13] Optimal Training-Time Scaling in Gradual Adaptation
链接: https://arxiv.org/abs/2608.04927
作者: Zonghuan Xu,Krishna Harish
类目: Machine Learning (cs.LG)
*备注: 24 pages, 5 figures
Abstract:In gradual adaptation, how should the training time on each task change as the number of intermediate tasks increases? We study this question for overparameterized linear regression tasks that change smoothly and share a zero-loss solution. With N tasks and training time s_N on each, the final learning progress converges to a continuum curve when Ns_N\to\tau . The limiting progress is \Theta(\tau) for small \tau and \Theta(\tau^-1) for large \tau , so both very short and very long training produce little progress. It follows that optimal per-task training times scale as s_N^\star=\Theta(N^-1) , equivalently Ns_N^\star=\Theta(1) . Experiments on gradually rotated MNIST and a natural Yearbook time shift are consistent with less per-task training as the path is divided more finely.
[LG-14] Variational Bounds for Perceptron Learning from Structured Data
链接: https://arxiv.org/abs/2608.04882
作者: Francesco Camilli,Pierluigi Contucci,Federica Gerace,Emanuele Mingione
类目: Machine Learning (cs.LG); Disordered Systems and Neural Networks (cond-mat.dis-nn); Mathematical Physics (math-ph); Machine Learning (stat.ML)
*备注: 51 pages, 10 figures
Abstract:We introduce a variational approach to a finite-temperature continuous-spin perceptron trained on a Gaussian mixture. The model allows for a broad class of concave utilities and log-concave separable prior measures on the spins. By combining the interpolation method with log-concavity and concentration estimates, we derive lower and upper minimax variational bounds for the limiting quenched pressure. Remarkably, the two bounds differ only in the order of optimization of two variational parameters, while all remaining extrema are controlled by the concave–convex structure of the variational potential. Whenever the two optimizations commute, the two bounds match and identify the solution of the model. The same potential yields the fixed-point equations as stationarity conditions and provides a unified route to the computation of the ground-state energy, training loss, and generalization error.
[LG-15] Robust Control under Stationary Ambiguity
链接: https://arxiv.org/abs/2608.04832
作者: Konrad J. Mueller,Amira Akkari,Ben Wood,Lukas Gonon
类目: Machine Learning (cs.LG); Computational Finance (q-fin.CP)
*备注:
Abstract:Control policies optimized in simulation can perform poorly in the real system when the parameters x of the simulator are estimated from limited data but the resulting parameter uncertainty is not represented inside the simulation. A common way to incorporate such ambiguity is to simulate each trajectory of the system under a randomly drawn value for x . Since the policy cannot observe the drawn value, it must initially choose controls that perform well across many possible parameter values. However, if the policy progressively observes the system, it can often gradually infer the value of x , so that ambiguity vanishes. Over time, the policy then specializes to its estimate of x and loses its robustness. This is undesirable in many real systems, where latent factors are expected to shift. In financial markets, for example, a policy hedging a derivative payoff should remain robust to changes in the volatility regime. To induce such continual robustness, we propose training policies in simulators where ambiguity varies with the system’s state but does not systematically decay over time. We formalize this requirement as stationary ambiguity: the simulator should induce a stationary filter process over the latent state. We show how to construct such simulators and demonstrate, on hedging problems, that policies trained under stationary ambiguity preserve robustness to latent factors over time, leading to strong performance on real market data. As a modeling principle, stationary ambiguity informs many simulator design decisions: which models make realistic simulators, how their parameters should be randomized, and how simulator and policy should be initialized. While our experiments focus on hedging, stationary ambiguity may also be useful for other sequential control problems driven by exogenous stochastic processes with shifting latent structure.
[LG-16] MGSB: Manifold Gated Signature Branch Pressure-Domain Baseline Architecture for Two-Phase Pipeline Flows Under Distributional Shift
链接: https://arxiv.org/abs/2608.04805
作者: Issah Suleiman,Sormeh Serpoosh,Nadine Elkholy,Hicham Ferroudji,Mohammad Azizur Rahman,Matthew Hamilton
类目: Machine Learning (cs.LG)
*备注:
Abstract:Leak detection models for multiphase pipelines often degrade when deployed under flow regimes that differ from training. Existing evaluations typically assess performance under in-distribution operating conditions, masking failures caused by regime transitions such as bubble-to-slug flow. We propose the Manifold Gated Signature Bias (MGSB), a regime-aware architecture combining regime-conditioned feature fusion, a TT-RoughPath encoder, and Mean-Teacher consistency regularization to improve robustness under distribution shift. Under leave-one-group-out evaluation, MGSB achieves a detection F1 of 0.930 and an OOD F1 of 0.783, substantially outperforming CNN-LSTM and fully connected baselines under severe feature corruption. Ablations show the proposed architecture, not the training procedure, is the primary contributor to OOD robustness, while Mahalanobis-distance analysis confirms the held-out conditions are genuinely out-of-distribution. These results show that explicit regime-aware modelling is a practical path toward robust, sensor-agnostic leak detection in industrial multiphase pipelines.
[LG-17] Above-ground Biomass Estimation with Geospatial Foundation Models
链接: https://arxiv.org/abs/2608.04792
作者: Ghjulia Sialellia,Linus Scheibenreif,Jan Dirk Wegner,Konrad Schindler
类目: Machine Learning (cs.LG)
*备注:
Abstract:Accurate estimation of Above-Ground Biomass (AGB) from satellite imagery is essential for the large-scale monitoring of carbon stocks, yet it remains a challenging regression task at global scale. Geospatial Foundation Models (GFMs) have recently emerged as a promising machine learning paradigm to derive general-purpose representations from Earth observation data, but their utility for quantitative regression tasks like biomass estimation remains largely unexplored, as most benchmarks emphasize classification and segmentation. Here, we present a comprehensive benchmark of GFMs for global-scale AGB estimation using the AGBD dataset, a machine learning-ready benchmark spanning diverse biomes and geographies. We distinguish two ways in which GFMs reach practitioners: (i) models distributed as weights to be run by the user, which we evaluate as frozen encoders within the PANGAEA benchmarking framework; and (ii) models distributed as ready-to-use, pre-computed embedding products, for which we evaluate AlphaEarth Foundations (AEF) and TESSERA. We compare 11 GFMs available on PANGAEA and both embedding products against a fully supervised state-of-the-art (SOTA) model, assess their geographical and temporal generalization abilities, as well as agreement with the ESA CCI biomass product on independent reference data. Our results show that GFMs run as frozen encoders substantially underperform with respect to the supervised SOTA model, whereas pre-computed embedding products prove highly effective. An MLP trained on AEF embeddings outperforms the supervised SOTA model trained on AGBD features, and the same SOTA model trained on AEF embeddings (optionally augmented with selected raw features) achieves the best overall result, while also generalizing better across space and time.
[LG-18] Continual-Learning Physics-Informed Neural Networks for Parameterized Partial Differential Equations
链接: https://arxiv.org/abs/2608.04778
作者: Xujia Chen,Xinyue Hu,Letian Chen,Yi Liu,Wenhui Fan
类目: Machine Learning (cs.LG)
*备注: 124 pages in total, including the main text (63 pages, 25 figures, and 16 tables) and supplementary material (61 pages, 30 figures, and 17 tables)
Abstract:Physics-informed neural networks (PINNs) incorporate governing equations into neural-network training and can approximate PDE solutions without requiring large observational datasets. Parameterized PINNs (ParamPINNs) further take physical parameters as inputs, allowing a single model to represent a family of PDE solutions over a parameter domain. Existing ParamPINNs, however, still face inefficient training, uneven accuracy across parameters, and overfitting to a limited set of sampled parameter tasks, which can impair generalization to unsampled parameters. To address these issues, we propose a continual-learning physics-informed neural network (CL-PINN), which treats PDE instances at different parameter values as related tasks and learns them sequentially. CL-PINN combines Bayesian-optimization-based active parameter selection, task-wise dynamic loss weighting, sparse physics-constrained replay, and an optional parameter subnetwork to improve task allocation and knowledge retention under bounded active-task capacity. It requires no observational data and is designed to solve parameterized PDEs over relatively broad parameter domains under limited computational resources. Multi-seed evaluations on five benchmarks, including one continuous function and four parameterized PDEs, show that Bayesian selection substantially reduces objective-loss queries relative to grid-greedy search, while sparse replay mitigates forgetting of earlier tasks. Under the prescribed within-case resource protocols, CL-PINN generally provides higher and more balanced solution accuracy than fixed-sampling and grid-greedy baselines. CL-PINN offers a practical route toward learning PDE solutions that generalize across physical parameters and has the potential to support reusable physics-informed surrogates for large-scale engineering parameter studies.
[LG-19] Attention Anomalies! Handling Attention Layers in Unsupervised Federated Outlier Detection
链接: https://arxiv.org/abs/2608.04753
作者: Mihailo Ilić,Miloš Savić,Vladimir Kurbalija,Mirjana Ivanović,Giancarlo Fortino,Dušan Jakovetić
类目: Machine Learning (cs.LG)
*备注: Submitted to the 4th IEEE International Conference on Federated Learning Technologies and Applications (FLTA 2026)
Abstract:Attention layers are the backbone of today’s most powerful and impactful models. Models with multi-million and billion parameters rely on contextual knowledge provided by attention layers. However, their use goes well beyond just being the core component of large language models. One particularly interesting application is in Memory Augmented Autoencoders (MemAE), specifically for unsupervised representation learning in outlier detection tasks. It was shown that attention helps these models be more effective in centralized learning scenarios. Our work aims to address the lack of specialized aggregation techniques in Federated Learning (FL) when it comes to MemAE models. In this paper we analyze the intricacies of the architecture behind Memory Augmented Autoencoders, and propose novel, guided approaches to effectively aggregate these models in federated scenarios. We demonstrate our approach on non-IID datasets and show that these novel aggregation schemes are more robust when dealing with numerous edge nodes in environments with unbalanced datasets, specifically for unsupervised anomaly detection scenarios. This approach improves the performance of even very shallow autoencoders, allowing them to be used in resource constrained environments.
[LG-20] Benchmarking Deep Learning Models for Dense Event Classification of Offshore Wind Infrastructure in Sentinel-1 Time Series
链接: https://arxiv.org/abs/2608.04706
作者: Thorsten Hoeser,Felix Bachofer,Claudia Kuenzer
类目: Machine Learning (cs.LG)
*备注: 27 pages, 14 figures
Abstract:Monitoring of offshore wind energy infrastructure life cycles, especially during the deployment phase, is an important contribution for stakeholders to make informed decisions in a phase of increasing deployment activities. ESA’s Sentinel-1 Synthetic Aperture Radar (SAR) mission produces large data archives that enable the global monitoring of offshore wind infrastructure. Turning these high-volume archives into information requires algorithms that automatically extract single event labels from dense time series at a global scale. In this study, we present a structured comparison of ten deep learning model-training variants for the dense classification of Sentinel-1 based offshore wind infrastructure time series, aiming to advance rule-based event classification of this task. We trained LSTM, Transformer, and fully connected model variants with monotemporal, unidirectional, and bidirectional context awareness, each with and without self-supervised pretraining. Among these, the supervised BiLSTM performs best, raising the target AUC score from 0.7853 for the rule-based baseline to 0.8509, and the perfect match rate from 0.3508 to 0.5063. Combining the BiLSTM predictions with the existing baseline labels in a label-transition-minimising ensemble further improves agreement with the test data. Using these improved labels, we isolate the deployment phase of individual turbines at a global scale and conduct a regional and subregional analysis covering 2016-01-01 to 2025-03-31, reporting median deployment durations of 84 d (China), 242 d (EU), and 258 d (UK). Deployment-related drivers, including legal regulations such as subsidies, and environmental conditions, emerge clearly from the analysed results across multiple spatial scales.
[LG-21] Suppression Sticks Locality Is Frag ile: A Closed-Loop Target-and-Control Audit of Task-Vector Negation in VLA Policies
链接: https://arxiv.org/abs/2608.04692
作者: Shaoguang Wang,Weiyu Guo,Rushi Dai,Yiren Zhao,Yandong Guo,Hui Xiong
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: 28 pages, 14 figures, 40 tables. Preprint
Abstract:Task-vector arithmetic offers a closed-form way to modify a model, yet its behavioral locality remains unclear in closed-loop robot control. We present a target-and-control audit of per-skill task-vector subtraction from multitask vision-language-action (VLA) policies. Across all ten LIBERO-Goal skills, subtraction produces three qualitatively different regimes: target-control separation for five skills, resistance for three, and global collapse for two. On held-out initial states, the five suppressible targets remain at 0% success; however, mean baseline-normalized control retention is only 52%, and each target-suppressing edit materially harms at least one nominally unrelated control. Additional Goal panels show separation across tested policies with continuous-regression, discrete-token, and flow-matching action heads, whereas we observe no clean separation on Spatial and control collapse on the tested Object and Long-horizon panels. Mean task-vector cosine does not account for this variation. A matched-norm control identifies a local sign asymmetry around one Goal anchor, while multi-vector outcomes vary with anchor and scale. Retain-aware gradient baselines provide data-dependent comparators but require removal-time data and optimization; subtraction is data- and gradient-free only at edit time, assuming precomputed expert deltas. Finally, a single-skill relearning probe is consistent with behavioral masking, not certified unlearning. These results characterize task-vector subtraction as a fast but brittle intervention and underscore the need for closed-loop target-and-control evaluation when assessing locality in embodied model editing.
[LG-22] he Sample Complexity of Distributionally Robust PAC Learning under Cressie–Read Divergences
链接: https://arxiv.org/abs/2608.04686
作者: Elad Aigner-Horev,Daniel Rosenberg,Roi Weiss
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study distributionally robust PAC learning for the 0 – 1 -loss, where adversarial perturbations of the data distribution are constrained by a Cressie–Read divergence of order k1 and radius \rho\geq 0 . For hypothesis classes with VC dimension d , we establish realizable and agnostic sample-complexity bounds tight up to constant and logarithmic factors, respectively; ordinary empirical risk minimization attains both rates up to logarithmic factors. For target accuracy \varepsilon\in(0,1) and confidence \delta\in(0,1) , their respective orders are [ \max!\left\frac1\varepsilon, \frac\rho^\frac 1k-1\varepsilon^k_\star \right\cdot(d+\log \delta^-1) \qquad\textand\qquad \max!\left\frac1\varepsilon^2, \frac\rho^\frac1k-1\varepsilon^k_\star\vee 2 \right\cdot(d+\log \delta^-1), ] where k_\star=k/(k-1) . For every fixed \rho0 , robustness changes the realizable \varepsilon -dependence from \varepsilon^-1 to \varepsilon^-k_\star as \varepsilon\downarrow0 . In the agnostic case, for 1k2 , robustness changes the \varepsilon -dependence from \varepsilon^-2 to \varepsilon^-k_\star , whereas for k\geq2 the exponent remains the classical 2 , with nontrivial \rho -dependence. Building on the known scalar reduction of robust 0 – 1 risk to ordinary classification error, our analysis reveals a scale-sensitive interaction between the statistical estimation of classification error and its amplification by robustness, sharply explaining the transition in the agnostic rate. We extend the previously studied \chi^2 -divergence case to every Cressie–Read order k1 , close its upper–lower gaps, and recover standard PAC learning rates as \rho\to0 , unlike previous bounds that fail to interpolate correctly in this limit. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2608.04686 [cs.LG] (or arXiv:2608.04686v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.04686 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-23] Diverse and Plausible Algorithmic Recourse via Tractable Recourse Distributions
链接: https://arxiv.org/abs/2608.04677
作者: Anagha Sabu,Hrithik Suresh,Narayanan C. Krishnan
类目: Machine Learning (cs.LG)
*备注:
Abstract:Algorithmic recourse seeks to help individuals reverse unfavorable automated decisions by recommending actionable changes that achieve a desired outcome. As an individual usually has several distinct routes to a favorable decision, and different people can act on different ones, a recourse system should offer multiple realistic alternatives rather than one. Existing approaches formulate recourse as an optimization problem that constructs one or a small set of counterfactuals rather than modeling the underlying space of feasible solutions, and in practice each sacrifices diversity, plausibility, or feasibility to secure the others. We propose Tractable Recourse Distributions, a probabilistic framework that represents the space of feasible alternatives for a given factual instance as a probability distribution over favorable outcomes. For commonly used cost functions based on proximity and the number of feature changes, we show that this distribution admits an exact representation as a probabilistic circuit, obtained by exponentially tilting the circuit; each individual’s distribution is therefore available in closed form, without retraining the model. Sampling from these distributions naturally produces diverse and plausible recourses, while the tilting parameters provide explicit control over their proximity and sparsity. Experiments on standard algorithmic recourse benchmark datasets demonstrate that the proposed framework attains diversity, plausibility, and feasibility simultaneously, while retaining sufficient probability mass over feasible counterfactuals for rejection sampling to be practical. A visual study on MNIST illustrates how the tilt strength trades proximity against validity.
[LG-24] Differentiating Through Dual Prices: End-to-End Policy Learning Under Capacity Constraints
链接: https://arxiv.org/abs/2608.04669
作者: Mohammadsaeed Haghi,Mahdi Salmani,Nima Kelidari
类目: Machine Learning (cs.LG)
*备注: 15 pages, 7 figures, 2 algorithms. Includes a technical appendix with full proofs, an excess-value decomposition, ablations, and reproducibility details. Code: this https URL
Abstract:Many social services assign scarce resources, such as housing assistance or hospital interventions, to people who arrive one at a time: each arrival must receive a decision immediately, and the long-run usage of every resource must stay within its capacity. We study how to learn such an assignment policy from logged observational data. The standard pipeline is decision-blind: fit one outcome model per arm by regression, price each capacitated resource from the fitted models, and assign each arrival the arm whose predicted outcome minus price is largest. We instead train the outcome models end-to-end, differentiating an off-policy estimate of the deployed policy’s value through the dual prices themselves. We study two formulations: an exact nonconvex one, and a convex relaxation whose optimum always satisfies the capacity constraints in expectation and which is suboptimal by at most a term linear in the smoothing temperature and logarithmic in the number of arms. Every method is evaluated in a queueing simulation with resources replenished at their capacity rates. Across six datasets, the two end-to-end variants take the top slots on a deployment-adjusted value index at every delay cost, including zero; when capacities are binding, decision-blind baselines frequently violate them and incur much longer queueing delays. On the largest dataset, a hospital cohort of seventy thousand patients, end-to-end training also achieves significantly higher policy value, a margin that survives a capacity-matched neural baseline. Flexible decision-blind regression remains the stronger pure predictor where ground truth is measurable; end-to-end training is best suited to settings where resources are genuinely scarce and feasibility matters.
[LG-25] Active Learning Guided Design Space Refinement for Scalable Multi-Objective Bayesian Optimization in Materials Discovery
链接: https://arxiv.org/abs/2608.04651
作者: Alexandros Ntagiantas,Panagiotis Tsilimidos,George Giannakopoulos,Christoforos Rekatsinas,Panagiotis Krokidas
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci)
*备注:
Abstract:Advanced materials discovery increasingly relies on machine learning and Bayesian optimization to explore large discrete design spaces under limited evaluation budgets. However, conventional Bayesian optimization (BO) can become inefficient as candidate spaces grow, often evaluating low-value regions before reaching informative areas. We propose an active-learning (AL)-guided adaptive search-space refinement framework combined with multi-objective BO to accelerate materials optimization while preserving Pareto-relevant regions. We evaluate the approach on CH4/N2 separation in covalent-organic frameworks and pressure-vessel design with material-direction stress components and thickness objectives. Results show that the AL-guided refinement reduces the candidate space by approximately half while preserving more than 99 percent of the original hypervolume. The reduced-space strategy improves early convergence and cumulative Pareto-front discovery from the BO, demonstrating efficient large-scale materials optimization across constrained autonomous materials discovery settings.
[LG-26] An entropic explanation of insistence on sameness in autism
链接: https://arxiv.org/abs/2608.04616
作者: Przemysław Śliwiński
类目: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC); Machine Learning (stat.ML)
*备注: 11 pages, 2 figures, a preprint of the accepted article
Abstract:An information theory-based framework is proposed in attempt to explain insistence on sameness in autism as an instance of a general behavior pattern in which an individual tries to reduce surprise and uncertainty. It offers a new definition of autism as an impairment in which cognitive functions are restricted to discrimination, memorization and prediction of tangible properties of the environment. An analogy between insistence on sameness and constrained minimization of the entropy metric is observed and examined for a set of assumptions that describe cognitive limitations of a person with autism. The metric is given by the formula D_H(R, M) = H(R|M) + H(M|R) , where R represents sequences of random stimuli, M is a memory that stores and retrieves them, and where H(.|.) denotes their conditional entropies interpreted as surprise and uncertainty, respectively. It is first inferred that to minimize the metric an individual can learn about R (and store that knowledge in M ) or can restrict R to the already known M . Then, it is concluded that insistence on sameness is a manifestation of the latter. Moreover, it is shown that the proposed framework: (1) Helps to quantify the concepts of surprise, uncertainty, sensory overload and deprivation, anxiety, comfort zone, disappointment, disorientation, pedantry, rigidness, observance or aberrant precision. (2) Leads to a list of guidelines for learning therapies and daily care routines, and allows them to be defined as optimization algorithms and implemented as programs for robotic live-in caregivers. (3) Can be validated with the help of a Turing test-like approach that requires no experiments involving individuals with autism. The framework-if positively validated-will provide formal foundations and design guidelines for therapies aimed at improving self-reliance of individuals with autism in basic activities of daily living.
[LG-27] Why Ranking Anomaly Detection Algorithms Isnt as Reliable as You May Think ICPR
链接: https://arxiv.org/abs/2608.04613
作者: Simon Klüttermann,Jérôme Rutinowski,Frederik Polachowski,Alice Kirchheim
类目: Machine Learning (cs.LG)
*备注: Accepted at the 2026 ICPR Workshop on Workshop on Reproducible Research in Pattern Recognition
Abstract:Anomaly detection is a safety-critical machine learning problem with applications ranging from fraud detection to network intrusion prevention and industrial monitoring. Despite the large number of proposed anomaly detection algorithms, many novel methods claim state-of-the-art performance. However, many authors do so under benchmark settings that are not aligned with one another. This lack of comparability raises concerns regarding the reproducibility and reliability of anomaly detection benchmarks. In this work, we study the impact of common benchmarking choices on the stability of algorithm rankings. Using seven representative anomaly detection algorithms and 690 datasets from the OddBench benchmark suite, we analyze how rankings change under varying dataset selections, evaluation metrics, hyperparameter configurations, and random seeds. To quantify this effect, we introduce a rank instability metric measuring the variability of algorithm rankings across benchmark settings. Our results show that algorithm rankings in anomaly detection are highly unstable. In many cases, almost every competitive algorithm can appear as the best-performing method under some benchmark configuration. Among the studied factors, dataset selection and hyperparameter choice contribute most strongly to ranking uncertainty, while random seeds and evaluation metrics have a comparatively limited impact. We also observe that reliable benchmarking requires substantially larger and more diverse dataset collections than the ones commonly used in prior work. Comments: Accepted at the 2026 ICPR Workshop on Workshop on Reproducible Research in Pattern Recognition Subjects: Machine Learning (cs.LG) Cite as: arXiv:2608.04613 [cs.LG] (or arXiv:2608.04613v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.04613 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-28] Learning Compression Rules for Network Traffic
链接: https://arxiv.org/abs/2608.04545
作者: Quentin Lampin(1)Éloi Sainte-Beuve(1,2)Louis-Adrien Dufrène(1)Guillaume Larue(1)Massih-Reza Amini(2) ((1) Orange Research (2) Université Grenoble Alpes)
类目: Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注:
Abstract:We study the problem of learning compact rule-based compressors for structured network traffic. Each packet is a record of header fields that are highly redundant within a flow, and a compressor is a small set of rules matching such records and replacing predictable fields with short codes. We cast rule learning as a two-stage problem: (i) an unsupervised structure-discovery stage that recursively partitions training packets using a normalized entropy-ratio criterion robust to small samples, and (ii) a constrained selection stage that uses dynamic programming to pick the rule subset maximizing expected compression gain under a hard budget on the number of installable rules. We instantiate the framework on Static Context Header Compression (SCHC), the IETF standard for rule-based header compression in constrained networks, and evaluate it on four real-world Internet-of-Things and 5G core-network datasets. Our method, Robust Entropy Clustering for Adaptive comPression (RECAP), surpasses expert-engineered rule sets with a small number of learned rules and removes the need for manual rule design.
[LG-29] Discretization and Statistical Consistency of Functional Flow Matching
链接: https://arxiv.org/abs/2608.04531
作者: Lennon J. Shikhman
类目: Machine Learning (cs.LG); Probability (math.PR); Machine Learning (stat.ML)
*备注: 31 pages, 2 tables
Abstract:Functional flow matching is posed on distributions of functions but implemented from finitely many coefficients or point values. Under scattered or adaptive refinement, the resulting conditioning sigma-algebras need not be nested, so martingale convergence does not justify the sensor limit. We prove strong L^2 convergence of finite conditional velocity targets for every strongly consistent sequence of finite-rank reconstructions, with quantitative bounds for orthogonal projections and a point-sensor extension through a regularity space. For learned flows, coupling directly to a population superposition path yields an end-to-end Wasserstein bound without assuming uniqueness of the population finite-dimensional ODE. We verify sensor-independent constants for a normalized quadrature neural operator, including globally Lipschitz activations through an explicit magnitude recurrence. A noncommuting trace-class Gaussian example gives boundary multiplier 0 under projected restriction and 0.72 under exact conditioning. A spatial regularity–cubature certificate closes the operator-realization term, a Bernstein argument gives a \widetildeO(n^-1) excess-risk term for fixed model dimension and envelopes, and an exactly realizable clipped Gaussian scaling specialization yields an explicit end-to-end rate.
[LG-30] Local Violation Certification for Linear Predict-Then-Optimize Pipelines
链接: https://arxiv.org/abs/2608.04474
作者: Ş. İlker Birbil,Wenhao Chi
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注: 24 pages, 3 figures
Abstract:Data-driven decision pipelines combining predictive machine learning models with downstream optimization software are increasingly used to make high-stakes operational decisions. Certifying the safety, fairness, and reliability of these decisions is essential, yet traditional scenario generation methods rely on repeated random testing, which becomes computationally prohibitive when failure events are rare and offers little insight into why failures occur. We present a framework for local violation certification designed specifically for linear decision pipelines under input uncertainty. We mathematically demonstrate that standard sampling methods fail efficiently for rare violations, motivating a direct structural approach. By analyzing the fixed decision boundary of a deployed pipeline, we show that the local risk of failure can be calculated directly in closed form using a single optimization solve. Furthermore, we introduce an exact sampling procedure and closed-form risk statistics that provide feature-level attributions (identifying which input characteristics contribute most to potential non-compliance) without requiring repetitive random trials or complex sampling algorithms. We demonstrate our approach on an economic power dispatch system subject to emissions regulations, delivering precise, auditable risk assessments at a fraction of the traditional computational cost.
[LG-31] Deltoris: Enabling Real-time VLA Inference in Embodied AI via Bit-level Sparsity and Speculative Inference
链接: https://arxiv.org/abs/2608.04428
作者: Zheng Liu,Zeyu Guo,Zihan Liu,Anbang Wu,Han Zhao,Fangxin Liu,Zhezhi He,Yinhe Han,Jingwen Leng,Minyi Guo,Yiming Gan,Yu Feng
类目: Hardware Architecture (cs.AR); Machine Learning (cs.LG)
*备注:
Abstract:Vision-language-action (VLA) models have emerged as a key component in embodied AI. Among existing approaches, diffusion-based VLA models achieve superior motion quality and generalization. However, diffusion-based VLA models are compute-intensive and must run at high control frequency, e.g., 50-200 Hz. Thus, it imposes strict latency and energy constraints on edge devices. In this work, we present Deltoris, an algorithm-hardware co-design framework for efficient diffusion-based VLA inference. First, we exploit the temporal similarity of consecutive inputs and propose a \textittemporal-aware bit-sparsity algorithm that computes only the differences between consecutive inputs, eliminating redundant bit-level operations. To further address the extra off-chip traffic introduced by our algorithm, we propose a \textitspeculative inference technique, which amortizes data loading across multiple control steps. Lastly, to support these techniques, we co-design a dedicated accelerator with customized 1D systolic bit-serial PE arrays that eliminate PE workload imbalance. Our evaluation shows that Deltoris achieves up to 34.2 \times speedup over mobile GPUs and 6.1 \times over prior accelerators, while maintaining comparable accuracy. Subjects: Hardware Architecture (cs.AR); Machine Learning (cs.LG) Cite as: arXiv:2608.04428 [cs.AR] (or arXiv:2608.04428v1 [cs.AR] for this version) https://doi.org/10.48550/arXiv.2608.04428 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-32] Elbow-Based MoE Routing: A Training-Free Inference Time Plugin for Expert Selection
链接: https://arxiv.org/abs/2608.04401
作者: Robin Pan,Raymond Liu,Daniel Fang,Adelina Andrei,Rosa Wu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Mixture-of-Experts (MoE) models enable model scaling while maintaining low inference-time compute by activating only a subset of experts per token. However, conventional routing relies on a fixed top-k selection, forcing the model to spend the same compute regardless of how many experts are relevant. We introduce elbow-based routing, a training-free inference-time modification that dynamically adjusts the number of experts on a per-token basis. Our method examines the sorted router probability distribution and identifies an elbow point that separates high- and low-probability experts. We find that most router distributions exhibit clear inflection points suitable for this strategy, and we show both theoretically and empirically that elbow-based routing preserves expert load balance. Experiments on a state-of-the-art MoE model demonstrate an average latency reduction of 5.3% while maintaining accuracy across six benchmarks.
[LG-33] When Proxy Prediction Becomes Equation Reconstruction: Diagnostics and Residual Learning for Factor-Derived Proxy Supervision
链接: https://arxiv.org/abs/2608.04393
作者: Chayan Lahiri,Ahmed Shafee,Cody Fehringer
类目: Machine Learning (cs.LG)
*备注: 8 pages, 2 figures. Preprint
Abstract:Scientific machine learning often relies on proxy targets computed from known domain factors when direct observations are limited. When those same factors are used as model inputs, however, high predictive accuracy may reflect reconstruction of the proxy-generating equation rather than robustness to degraded factor information. We study this problem in RUSLE-derived soil-loss proxy prediction under controlled degradation of the soil-erodibility factor K . We introduce a diagnostic framework that combines degraded-formula references, classical tree-based baselines, matched direct and formula-feature predictors, contextual ablations, tail-error analysis, and degradation robustness scoring. We then propose RASPL, a formula-preserving residual framework that retains the degraded formula estimate as the prediction anchor and learns an adaptively gated contextual correction. RASPL substantially outperforms matched direct prediction and provides stronger degradation and tail robustness than treating the formula estimate as an ordinary input feature. Within RASPL, a compact statistical encoder achieves the highest macro-averaged R^2 and lowest computational cost, whereas a convolutional encoder achieves the strongest degradation robustness and lowest Tail95 mean absolute error (MAE). These results establish formula preservation as the central design principle for robust learning from factor-derived proxy targets.
[LG-34] NeuroPB: Scaling Neural Decoding with Pretrained Behavioral Representations
链接: https://arxiv.org/abs/2608.04389
作者: Luyao Jin,Yonghao Song,Huan Zhao,Vincent C. K. Cheung,Wei-Hsin Liao
类目: Machine Learning (cs.LG)
*备注:
Abstract:Decoding continuous motor trajectories from neural activity is essential for developing practical brain-computer interfaces (BCIs). However, current neural decoders are constrained by the limited scale and heterogeneity of neural recordings. In contrast, behavioral data can be collected more readily and at substantially larger scale from humans, animals, simulations, and robotic systems. Here, we introduce NeuroPB, a framework that scales neural decoding by transferring knowledge from pretrained behavioral representations. NeuroPB first pretrains a motor encoder on large-scale motor behavior data and then aligns neural activity with the resulting behavioral representation space using a limited set of paired neural-behavioral recordings. A neural encoder and lightweight motor decoder are subsequently optimized to reconstruct continuous movement from the aligned neural representations. Across multiple macaque motor datasets, behavioral pretraining improves trajectory decoding, including an 11% R^2 increase on center-out and 8% on random-target compared with training the motor encoder from scratch. Notably, pretraining on robotic trajectories achieves performance comparable to pretraining on macaque trajectories, demonstrating that transferable kinematic structure is shared across biological and artificial models. Moreover, decoding performance improves as the scale and diversity of robotic pretraining data increase, when the amount of neural data is fixed. Pretraining also enhances generalization across recording sessions, subjects, and motor tasks, with only 10% calibration needed to match training from scratch. Overall, these results establish behavioral pretraining as a scalable source for neural decoding and provide a promising route toward high-performance and calibration-efficient BCIs under limited neural data.
[LG-35] Non-asymptotic implicit bias of logistic regression at early-stage gradient descent dynamics
链接: https://arxiv.org/abs/2608.04382
作者: Han Bao
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Gradient descent has been of particular interest in modern machine learning beyond sole focus on optimization. Implicit bias emerging from optimization, though not being encoded by the learning objective, often prevents from overfitting to spurious patterns. A typical instance is the max-margin implicit bias of a linear classifier, widely established for exponentially tailed loss functions. Even after having a given dataset separated, the parameter vector continues to evolve towards the max-margin direction asymptotically along the gradient descent dynamics. This phenomenon corroborates a frequent empirical observation of “train longer, generalize better.” However, the max-margin convergence is an asymptotic phenomenon, and what is worse, this asymptotic convergence rate is significantly slower than pure convex optimization. Even so, the parameter vector along gradient descent dynamics commonly correlates with the max-margin direction positively (though not exactly) within considerably fewer iterations than the asymptotic rate. By shedding another light on this classical problem, this work aims to understand the mechanism of this early-stage alignment phenomenon. Our theoretical results demonstrate that the parameter vector weakly aligns with the max-margin direction within O(\exp(\exp(-\delta))) iterations, where \delta0 is the permissible alignment error, which is shown to be tight. By tracking the radial and tangential flows, our proof operates on the alignment dynamics directly with dataset geometry and gets rid of the asymptotic expansion, which is a key insight to establishing faster weak alignment.
[LG-36] Helping Music Co-Creation Agents Listen Well: Hierarchical Self-Supervised World Models for Understanding and Generation NEURIPS2026
链接: https://arxiv.org/abs/2608.04378
作者: Scott H. Hawley
类目: ound (cs.SD); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
*备注: 20 pages, 14 figures. A 6-page version was submitted to the NeurIPS 2026 Creative AI Track. Supplemental website with listening examples: this https URL . Live demo: this https URL
Abstract:Collaborative music agents need internal representations rich enough to support both understanding and generation, yet flexible enough for a workflow where the human retains agency. We present a hierarchical self-supervised ``world model’’ for symbolic music: a 2.55M-parameter Swin V2 encoder trained on MIDI piano-roll images with JEPA-style objectives (pitch- and time-shift equivariance, masked embedding prediction, and a distributional regularizer), using no labels and no music-theory vocabulary. Probing the frozen embeddings shows that the level at which a musical property becomes decodable tracks its musical time scale: phrase boundaries are read off the coarsest levels, note density and harmonic detail off the finest. Temporal and phrase structure emerge from the self-supervised objectives alone, while harmonic content must be asked for; a small chord-supervision head raises joint chord recovery from .18 to .54, and key detection, which is never supervised, from .16 to .70. Following the Representation AutoEncoder paradigm, a conditional flow-matching model stands in for a trained decoder, flowing in pixel space from PCA-reduced conditioning: it reproduces a target window at pixel F1 0.996 , and the same per-level conditioning dropout that controls how far variations stray also enables graphical prompting for masked inpainting with no inpainting-specific sampler. The pipeline runs on CPU producing a suggestion in 2.8 s, or 0.6 s on Apple MPS, which we demonstrate in a live interactive demo. In concert with an LLM-based brain, these capabilities supply the core of a collaborative music creation agent in service of, rather than in place of, human agency.
[LG-37] EvtGraph: Event-Adaptive Compression for Sparse Temporal Graph Learning in Multimodal Time Series
链接: https://arxiv.org/abs/2608.04368
作者: Ziqian Wang,Tingxiong Xiao,Yuxiao Cheng,Jinli Suo
类目: Machine Learning (cs.LG)
*备注: 9 page, 9 figures
Abstract:Multimodal temporal data are inherently irregular and uneven in information density, yet most models rely on uniform discretization, leading to inefficient representations. We propose \textbfEvtGraph, a unified framework that aligns computation with temporal salience under explicit budget constraints. EvtGraph reparameterizes sequences into event-level tokens via event-adaptive compression (EAMC), selects a compact subset with a node budget (NBC), and performs temporally constrained sparse graph reasoning (T2SG). This transforms dense sequences into structured computation over salient events, reducing complexity while preserving critical transitions. We show that this design provides a practical mechanism for allocating representational capacity under a fixed budget, yielding a consistent performance–efficiency trade-off, where a small budget is often sufficient in practice. Experiments on multimodal clinical (MIMIC-IV + CXR) and cross-domain benchmarks demonstrate that EvtGraph outperforms both Transformer-based and recurrent baselines while significantly improving efficiency. These results suggest that budget-constrained event-centric representation provides a general paradigm for learning from high-redundancy temporal data. Comments: 9 page, 9 figures Subjects: Machine Learning (cs.LG) Cite as: arXiv:2608.04368 [cs.LG] (or arXiv:2608.04368v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.04368 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Ziqian Wang [view email] [v1] Wed, 5 Aug 2026 02:16:49 UTC (1,865 KB)
[LG-38] Manipulation-Proof Oblivious Audits against Deceptive Model Providers AAAI
链接: https://arxiv.org/abs/2608.04365
作者: Augustin Godinot,Sofiane Azogagh,Julien Ferry,Sébastien Gambs
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Computers and Society (cs.CY)
*备注: This work has been accepted for publication at the 2026 AAAI/ACM Conference on AI, Ethics, and Society (AIES)
Abstract:Audits have emerged as a critical instrument for algorithmic governance, providing a mechanism for external scrutiny and governance of machine learning models. However, ensuring the integrity of such assessments remains a challenging issue. For instance in regulatory contexts, audits are typically declared or easily detected, thus enabling model providers to manipulate the process, whether intentionally or inadvertently. This vulnerability is particularly acute in the context of fairness evaluations, in which providers can often infer sensitive attributes and strategically equalize allocation rates between groups to satisfy fairness metrics. In this paper, we introduce a novel audit protocol designed to significantly increase the post-audit detectability of such manipulations by enabling the auditor to query the model in an oblivious manner. Our approach leverages a Private Information Retrieval mechanism to require the provider to label a large set of instances, while preventing it from knowing which subset will ultimately be used for the audit. The protocol is efficient, imposes minimal overhead on the auditor, and requires no modification to the audited model, its training procedure, or its inference pipeline. We provide theoretical guarantees showing that, under this protocol, a provider attempting to hide unfairness must falsify a significantly larger number of responses, thereby increasing both the difficulty and the likelihood of detection of manipulation. Experimental results across representative audit scenarios confirm the effectiveness and practicality of our approach. Comments: This work has been accepted for publication at the 2026 AAAI/ACM Conference on AI, Ethics, and Society (AIES) Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Computers and Society (cs.CY) Cite as: arXiv:2608.04365 [cs.LG] (or arXiv:2608.04365v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.04365 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-39] Looking in the Mirror: Introspecting Side-Effect Misalignments Induced by Fine-Tuning
链接: https://arxiv.org/abs/2608.04347
作者: Kotaro Yoshida,Laura Gomezjurado Gonzalez,Yukinori Yamamoto,Yuji Naraki,Ryotaro Shimizu,Wenya Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Fine-tuning enables a source model to acquire desired capabilities and behaviors in a target domain while retaining much of its general-purpose competence. However, this adaptation process can also degrade alignment properties that were present in the source model. Recent work has shown that large language models can be trained using LoRA-based modules known as introspection adapters (IAs) to describe behavioral changes induced by fine-tuning. However, existing studies primarily consider settings in which the model is fine-tuned on datasets explicitly designed to implant a specific behavior and is then asked to explain the implanted behavior. This differs from practical deployment scenarios, where the central concern is often side-effect misalignment: unintended degradation of alignment caused by fine-tuning on tasks that are not obviously related to safety or alignment. To bridge this gap, we formulate a novel problem setting called \emphside-effect introspection, in which the target of introspection is not a behavior explicitly implanted through fine-tuning, but rather alignment shifts that emerge as unintended side effects, and we construct a dataset for this setting. Furthermore, to enhance sensitivity to internal model changes, we propose the Delta-Aware Introspection Adapter (DAIA), a novel mechanism designed to explicitly process both base-model activations and activation differences induced by fine-tuning. Our empirical evaluation shows that introspection learning generalizes to unseen fine-tuned models and safety categories, and that DAIA consistently outperforms existing introspection adapters.
[LG-40] Cost-Aware Multi-Objective Bandits: Theory and Application to Budgeted LLM Configuration Evaluation
链接: https://arxiv.org/abs/2608.04333
作者: Bo Xue,Zhi Hong,Jiayi Li,Yuanyu Wan,Ji Cheng,Shuang Qiu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Large language model (LLM) configuration evaluation is challenging due to limited evaluation budgets, varying costs, and multiple competing objectives. In this paper, we formulate LLM configuration evaluation as a cost-aware multi-objective bandit problem, where each configuration evaluation incurs a configuration-dependent cost and yields a noisy vector-valued outcome. Under this framework, we study two fundamental problems: online configuration selection and Pareto configuration identification. For online configuration selection, we propose a hypervolume-based UCB algorithm that optimizes an optimistic hypervolume-per-cost index. We establish a budgeted regret bound of order O\bigl(\sum_i\ne i^\star\frac\log B\Delta_i\bigr) , where B is the evaluation budget, i^\star is the optimal configuration in terms of hypervolume efficiency, and \Delta_i is the corresponding efficiency gap of configuration i . This bound retains the logarithmic budget dependence of classical single-objective budgeted bandits. For fixed-budget Pareto identification, we develop a cost-aware empirical gap elimination algorithm and prove that its error probability is of order O\bigl(\exp(-\fracBH_\mu,c)\bigr) , where H_\mu,c is a cost-aware Pareto identification complexity depending on configuration costs and Pareto classification gaps. This error probability decays exponentially with the evaluation budget and recovers the standard Pareto set identification guarantee when all configuration costs are identical. Experiments on LLM configuration evaluation tasks demonstrate that the proposed framework enables efficient online decision-making and accurate cost-aware Pareto identification under limited budgets.
[LG-41] Real-time probabilistic tsunami forecasting via generative AI
链接: https://arxiv.org/abs/2608.04327
作者: Yusuke Oishi,Takashi Furumura,Fumihiko Imamura
类目: Machine Learning (cs.LG)
*备注: 24 pages, 14 figures, 1 table
Abstract:Explicit onshore tsunami inundation forecasting can improve public risk awareness, but deterministically predicted inundation boundaries under highly uncertain conditions, such as near-field tsunamis generated by megathrust earthquakes, may falsely imply safety outside the boundaries. Consequently, current warnings primarily target coastal tsunami height, not onshore inundation. Although machine learning enables instant inundation predictions, they remain deterministic, lacking uncertainty quantification. Here, we develop a probabilistic ensemble model based on a conditional diffusion model (a type of generative AI) that reconciles accuracy with calibration. Validated with the 2011 Tohoku-oki earthquake data, our model faithfully tracks the postearthquake uncertainty decreasing over time while accurately predicting inundation depth and extent. Our framework shows that generative AI can shift tsunami forecasting from determinism to probabilism, providing a foundation for next-generation early warning.
[LG-42] ArborEnum: Decision Tree Rashomon Sets over Continuous Features
链接: https://arxiv.org/abs/2608.04310
作者: Zakk Heile,Hayden McTavish,Margo Seltzer,Cynthia Rudin
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:The Rashomon effect describes the phenomenon that many models can achieve nearly equivalent performance on the same learning task, with significant ramifications for robustness, feature importance, and customizability. These use cases motivate the computation of Rashomon sets: the set of all models whose regularized loss is near-optimal. Decision trees are one of the few model classes for which Rashomon sets can be fully enumerated, but this computation has always been conditional on a binarization of the original data, either restricting which splits each tree is allowed to make or substantially increasing the complexity of an already difficult combinatorial problem. We introduce the first algorithm that exactly enumerates decision-tree Rashomon sets while exploiting the ordered structure of continuous features. We further develop a relaxation for approximate enumeration and an anytime algorithm that progressively refines the set of candidate thresholds, producing increasingly detailed approximations that converge to the continuous-feature Rashomon set. Experiments show that coarse binarization can miss many trees, important features, and predictive multiplicity; our algorithms achieve orders-of-magnitude speedups over existing enumeration methods, with approximations providing further speedups while maintaining near-perfect recall.
[LG-43] Adaptive Finite-Budget Training for CVaR Risk-Aware Q-Learning
链接: https://arxiv.org/abs/2608.04305
作者: Yifan Wu,Junjie Lei,Wenjie Huang
类目: Machine Learning (cs.LG); Risk Management (q-fin.RM)
*备注: 8 pages, 3 figures, submit to ACM International Conference on AI in Finance: ICAIF '26 Milan
Abstract:Risk-aware Q-learning (RaQL) provides a model-free, two-timescale estimator for dynamic risk objectives, but its finite-budget behavior remains fragile: fixed inner-loop hyperparameters can produce unstable value estimates, persistent Bellman residuals, and inefficient sample reuse. This paper proposes an adaptive training controller for Conditional Value-at-Risk (CVaR) RaQL and evaluates it on a daily Bitcoin trading task. The controller preserves the original CVaR estimator and Bellman fixed point; instead, it redesigns the training procedure through six coordinated mechanisms: per-cell inner-step sizing, outer-rate-matched decay synchronization, a short early correction for the VaR-like inner variable, a coverage-first-then-greedy sample allocation rule, progressive suffix aggregation of mature inner estimates, and data-driven calibration of key scales from online-observable quantities. Across 20 random seeds and 856,000 inner-transition samples, the controller reduces the mean empirical CVaR Bellman residual by approximately 85% relative to the fixed-parameter baseline (MeanBEQ: 1.2202 to 0.1854; MeanBEV: 1.1624 to 0.0535) and maintains stability across CVaR levels, discount factors, and training budgets. On the chronological out-of-sample test set, the learned policy attains a Sharpe ratio of 0.9281 with a maximum drawdown of 6.46% after transaction costs. Although buy-and-hold yields a higher cumulative return (35.43% vs. 23.61%), the adaptive policy achieves far lower volatility (9.57% vs. 47.93%), drawdown, and CVaR loss. These results demonstrate that adaptive finite-budget training design, applied solely to the training procedure without altering the risk objective, can materially improve the reliability and risk-adjusted performance of risk-aware Q-learning in financial applications.
[LG-44] Sample Complexity of Multicalibration for Multilevel Properties
链接: https://arxiv.org/abs/2608.04288
作者: Jiuyao Lu,Krishnakumar Balasubramanian,Aleksandr Podkopaev,Shiva Prasad Kasiviswanathan
类目: Machine Learning (cs.LG); Statistics Theory (math.ST); Machine Learning (stat.ML)
*备注:
Abstract:Calibration requires a predictor to be unbiased after conditioning on its own predictions. Multicalibration asks for this guarantee simultaneously across a collection of groups. Many prediction tasks ask for several related features of the same conditional outcome distribution: variance is defined relative to the mean, skewness relative to both mean and variance, and conditional value at risk relative to a quantile. We study multicalibration for a sequence of k properties in which each property is identifiable once the preceding properties are fixed. This framework includes Bayes pairs but does not require the properties to arise from a single loss. For every fixed k\ge2 , we establish matching upper and lower sample-complexity bounds up to logarithmic factors under regularity conditions. Even with only polylogarithmically many binary groups, achieving multicalibration error \varepsilon requires \widetilde\Omega(\varepsilon^-(k+2)) samples. Conversely, for any finite group family \mathcal G , we give a randomized learner using O(\varepsilon^-(k+2)+\varepsilon^-2\log|\mathcal G|) samples. Thus the sample complexity is \widetilde\Theta(\varepsilon^-(k+2)) for polynomial-size group families. We instantiate the theory for three canonical examples. Subjects: Machine Learning (cs.LG); Statistics Theory (math.ST); Machine Learning (stat.ML) Cite as: arXiv:2608.04288 [cs.LG] (or arXiv:2608.04288v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.04288 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-45] Geometry-Informed Parameter-Efficient Fine-Tuning of Pre-trained Molecular GNNs for Blood-Brain Barrier Permeability Prediction
链接: https://arxiv.org/abs/2608.04257
作者: Marco Vieto Vega,Long D. Nguyen,Binh P. Nguyen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Blood-brain barrier permeability (BBBP) prediction is a critical screening task in central nervous system drug discovery, where candidate molecules must be assessed for whether they can cross, or should be prevented from crossing, the blood-brain barrier. However, this task remains challenging because of limited, class-imbalanced datasets and sensitivity to molecular structure. Recent advances in deep learning have established graph neural networks (GNNs) as a powerful approach for molecular representation learning, while pre-trained molecular GNNs provide transferable knowledge for downstream tasks. However, full fine-tuning is often parameter-inefficient and prone to overfitting, whereas existing parameter-efficient fine-tuning (PEFT) methods mainly adapt node features or the two-dimensional covalent graph, limiting their ability to capture three-dimensional geometry and second-order interactions. To address these limitations, we propose BBBP-GeoPEFT, a geometry-informed PEFT framework for pre-trained molecular GNNs. BBBP-GeoPEFT constructs distance-based graphs at multiple cutoffs and their corresponding line graphs from molecular conformers to capture spatial atom and second-order edge interactions. Lightweight auxiliary geometric graph encoders generate cutoff-specific representations, which are incorporated into each pre-trained layer through node-wise cutoff attention and gated residual connections. This design preserves pre-trained knowledge while incorporating permeability-relevant geometric information with a small trainable-parameter budget. Experiments on a curated BBBP dataset show that BBBP-GeoPEFT achieves competitive performance compared with full fine-tuning and representative PEFT baselines. Under both random and scaffold splitting, BBBP-GeoPEFT achieves competitive or improved ROC-AUC and accuracy in most experiments while updating only 10.1% of the model parameters.
[LG-46] PriDyG: Privacy-preserving Dynamic Graph Inference with LLM -GNN Collaboration
链接: https://arxiv.org/abs/2608.04255
作者: Yuyang Xia,Ruixuan Liu,Li Xiong
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Graph inference over relational data can expose sensitive edge information, and this risk becomes more severe in dynamic graphs, where repeated model updates cause privacy loss to accumulate. We formulate Edge-level Differentially Private Dynamic Graph Inference (EDG) and propose PriDyG, a private inference framework that combines GNN-based structural learning with LLM-based semantic reasoning. PriDyG introduces incremental private multi-hop aggregation, which buffers newly arrived edges and processes each edge exactly once. By parallel composition, the total privacy cost equals that of a single static release, independent of the number or schedule of model updates. Compared with geometrically decaying budget allocation, incremental aggregation avoids exponentially increasing noise while preserving exact one-hop signals and at least half of two-hop information transfers. PriDyG further complements privatized GNN outputs with LLM predictions derived solely from node text, incurring no additional edge-level privacy cost. Experiments on four benchmarks for node classification and link prediction show that PriDyG consistently outperforms geometrically decaying baselines under the same privacy budget and matches the utility of naive per-update retraining while reducing cumulative privacy cost by up to three orders of magnitude.
[LG-47] Attention-based representations for multi-task computation
链接: https://arxiv.org/abs/2608.04243
作者: Daniel Hsu,Mingyue Xu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Multi-head attention layers produce vector representations that support multiple downstream tasks. We establish bounds on the number of heads required in two simple and concrete multi-task scenarios. In the first scenario, a vector representation is sought so that linear predictors can compute both the smallest and largest numbers in a given list. In this case, it is known two attention heads with small embedding dimension and bit precision level suffice. We prove that a single attention head requires exponentially higher embedding dimension or precision level. In the second scenario, a vector representation is sought so that a polynomial threshold function can compute the XOR of a given string of n bits. This scenario is analogous to the first one for n=2 , since XOR is readily computed by a linear function using a vector representation that encodes both the AND and the OR of the two bits. We observe that n -bit XOR requires the product of the number of heads and the polynomial degree to be at least n , and we construct multi-head attention layers that match this lower bound. These results generalize to arbitrary (symmetric) Boolean functions, where the bound is given in terms of the threshold degree.
[LG-48] Physics-informed reduced-order modelling with equivariant spectral submanifolds
链接: https://arxiv.org/abs/2608.04239
作者: Georg Maierhofer
类目: Machine Learning (cs.LG); Dynamical Systems (math.DS); Numerical Analysis (math.NA)
*备注:
Abstract:Spectral submanifold (SSM) reduction has emerged as a mathematically principled route to reliable nonlinear reduced-order models, capturing dynamics beyond the reach of linear techniques such as Dynamic Mode Decomposition (DMD). The computation of SSMs, however, remains computationally expensive, particularly for high-dimensional systems. In this work, we introduce equivariant spectral submanifold (eSSM) reduction, a novel extension of the SSM framework that explicitly incorporates symmetries of the full-order model into the reduction process. We establish the mathematical foundations of this approach by showing that SSMs are naturally equivariant submanifolds and that the associated charts and reduced dynamics inherit the appropriate induced group actions. Building on this framework, we develop a novel equivariant SSM reduction algorithm that exploits these symmetries to achieve substantially faster computations while also improving model robustness. We demonstrate the advantages of this approach on several benchmark problems including a test from the Common Task Framework for Science.
[LG-49] Random features for Grassmannian kernel approximation with bounded rank-one projections
链接: https://arxiv.org/abs/2608.04227
作者: Rémi Delogne,Laurent Jacques
类目: Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注: 33 pages, 6 figures, J2C certification
Abstract:We propose a family of random feature maps for scalable kernel machines on low-dimensional subspaces, ie on the Grassmannian manifold. Such representations are useful when data classes or clusters are well described by the span of a few samples. Classical Grassmannian kernels, including the projection and Binet-Cauchy kernels, require full Gram matrices, which leads to prohibitive computational and memory costs for large high-dimensional subspace datasets. We address this limitation using random features based on rank-one projections of subspace projection matrices followed by bounded non-linear transforms, either periodic or binary, to control the resulting distributions. We show that inner products in the random feature space approximate well-defined rotation-invariant Grassmannian kernels that depend only on the principal angles between subspaces. When the number of features is sufficiently large relative to the intrinsic subspace dimension, the approximation holds uniformly over all fixed-dimensional subspaces with high probability. For periodic transforms, the approximated kernel has a closed-form expression with tunable behaviour between inverse Binet-Cauchy and Gaussian-type regimes. Binary transforms yield compact one-bit subspace features, although no closed-form kernel is known. Structured rank-one projections based on randomised fast Fourier transforms further reduce computation without sacrificing practical accuracy. Experiments on synthetic data and ETH-80 classification tasks show that these features accurately preserve Grassmannian geometry while reducing computation, memory, and storage. Rank-one embeddings therefore provide a practical and scalable alternative to classical Grassmannian kernels. Comments: 33 pages, 6 figures, J2C certification Subjects: Machine Learning (cs.LG); Signal Processing (eess.SP) Cite as: arXiv:2608.04227 [cs.LG] (or arXiv:2608.04227v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2608.04227 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Journalreference: Transactions on Machine Learning Research, July 2026
[LG-50] Attention-Only White-Box Transformer via LeJEPA-Based Self-Supervised Pretraining
链接: https://arxiv.org/abs/2608.04213
作者: Yang Bai,Linyuan Wang,Haoyang Jiang,Nuolin Sun,Libin Hou,Bin Yan
类目: Machine Learning (cs.LG)
*备注:
Abstract:Existing studies on self-supervised learning for white-box networks typically decouple the derivation of white-box networks via optimization algorithms from self-supervised learning paradigms. In this work, we instead revisit the two components from a joint perspective. The LeJEPA-based self-supervised framework assumes an isotropic Gaussian distribution as the optimal embedding distribution for downstream tasks, which is conceptually equivalent to the expansion term R(Z) in the sparse rate reduction objective guiding white-box Transformer optimization. Building on this observation, we use the LeJEPA self-supervised paradigm to optimize R(Z) , and derive the remaining terms R^c(Z\mid U_[K])+\lambda\lVert Z\rVert_0 via the alternating direction method of multipliers (ADMM) into an attention-only Transformer that dispenses with the ISTA structure or MLP layers of the original design. Experimental results demonstrate that our attention-only white-box Transformer achieves classification accuracies of 88.88% on CIFAR-10 and 63.54% on CIFAR-100 at the Base scale under the LeJEPA self-supervised paradigm, while the original white-box Transformer CRATE achieves classification accuracies of 89.18% on CIFAR-10 and 63.56% on CIFAR-100. Our model achieves competitive performance while reducing the parameter count by roughly 31% . Beyond the white-box setting, we further investigate standard ViTs and find that replacing all MLP blocks with ReLU activations under knowledge distillation removes approximately 66% of the parameters while preserving competitive accuracy, motivating further investigation into the potential redundancy of MLP modules in standard ViT architectures.
[LG-51] From Non-Convex Self-Concordant Regularization to Scalable Quasi-Newton Training of PINNs
链接: https://arxiv.org/abs/2608.04206
作者: Chenhao Si,Kang An,Shiqian Ma,Ming Yan
类目: Machine Learning (cs.LG)
*备注:
Abstract:Physics-informed neural networks (PINNs) often require high-accuracy quasi-Newton refinement to obtain reliable partial differential equation solutions, but their residual objectives can exhibit indefinite, nearly singular, and poorly scaled local curvature. Regularized quasi-Newton methods provide established mechanisms for stabilizing secant models, while self-concordant methods provide local-metric rules for curvature-dependent step selection. Building on these two lines of work, we propose SCORE, a self-concordance-inspired quasi-Newton method with decrement-coupled shifted secant geometry for PINN training. Its distinguishing mechanism is that a single quasi-Newton decrement computed from the learned inverse metric jointly determines a strong-Wolfe-tested candidate step and an adaptive shift used to define the next secant geometry. The shifted displacement represents the action of an averaged shifted metric along the accepted step, while requiring neither Hessian construction nor Hessian-vector products. Under a local spectral-equivalence condition, we show that the quasi-Newton decrement and candidate step remain comparable to their counterparts in a positive shifted metric, and recover the normalized self-concordant rule in the matched-metric case. Strong Wolfe acceptance, fallback line search, and standard curvature safeguards provide globalization without modifying the underlying PINN objective. Experiments on the viscous Burgers, Kuramoto–Sivashinsky, Korteweg–de Vries, and complex Ginzburg–Landau equations show that SCORE attains lower final errors than the tested BFGS and self-scaled Broyden baselines. The Burgers ablation further indicates that shifted curvature stabilization and decrement-based step selection make complementary contributions to high-accuracy refinement.
[LG-52] Unscented KalmanNet: a hybrid deep learning filter with calibrated posterior covariance for nonlinear state estimation
链接: https://arxiv.org/abs/2608.04201
作者: Minhyeok Ko,Abdollah Shafieezadeh
类目: Machine Learning (cs.LG); Signal Processing (eess.SP); Numerical Analysis (math.NA)
*备注:
Abstract:State estimation for nonlinear dynamical systems is commonly performed with the Unscented Kalman filter (UKF), which propagates the state moments through deterministic sigma points and reports a posterior covariance at every step. In practice, however, unknown and time-varying noise statistics and model mismatch degrade both estimation accuracy and covariance calibration. Existing learned filters improve accuracy but are largely built on the extended Kalman filter and either forgo an explicit covariance or learn uncertainty without correcting mismatch-induced gain bias. This paper introduces the Unscented KalmanNet (UKN), a hybrid recursive estimator that augments the UKF with two structurally distinct learned components while preserving its explicit sigma-point covariance recursion. NoiseNet predicts time-varying process and measurement covariances as bounded multiplicative corrections to fixed baselines, guaranteeing positive definiteness, while GainNet applies a bounded residual correction to the analytical gain. A calibration-aware training objective combines state error with covariance- and innovation-consistency terms through adaptive weights, jointly optimizing accuracy and calibration. UKN is benchmarked against UKF, KalmanNet, and Bayesian KalmanNet on three synthetic systems and UZH-FPV real-flight data. It achieves the lowest aggregate state-estimation error in all four examples and reduces RMSE by 26.4 - 49.7% compared to UKF in the synthetic cases. Leave-one-sequence-out cross-validation over 11 flights shows 22.4% and 34.3% reductions in mean position and velocity RMSE, respectively. UKN also yields the lowest fold-to-fold variability, with normalized NEES and empirical coverage closest to nominal values among the other filters.
[LG-53] A Comparative Study of Feature Selection Methods for EHR Diagnosis Codes in Opioid Use Disorder Prediction
链接: https://arxiv.org/abs/2608.04180
作者: Zihan Ding,Yinan Liu,Tengfei Ma,Rachel Wong,George Leibowitz,Benjamin Littenberg,Xia Zheng,Richard N. Rosenthal,Fusheng Wang
类目: Machine Learning (cs.LG)
*备注: Accepted at the AMIA 2026 Annual Symposium
Abstract:Feature selection is a critical step in electronic health record (EHR)-based predictive modeling, where input variables are often high-dimensional, sparse, noisy, and redundant. Large feature sets not only increase computational burden and overfitting risk, but also make model interpretation difficult, leading to limited usefulness in clinical settings. In this study, we focus on diagnosis-related features and compare five feature selection paradigms for opioid use disorder (OUD) prediction: recurrence enrichment, NTK-motivated early gradient sensitivity, LightGBM-SHAP, Elastic Net, and large language model (LLM)-guided semantic selection. We use a unified preprocessing and evaluation framework and assess each method by downstream predictive performance, resampling stability, and representation of infrequent diagnosis codes. Our results demonstrate that performance improves with larger feature budgets with diminishing returns beyond a moderate size. NTK sensitivity provides the best overall balance of accuracy and stability, and LLM-guided selection contributes complementary clinically meaningful signals despite lower standalone performance.
[LG-54] S2TabPFN: Time Series Classification and Extrinsic Regression through Feature Extraction and a Tabular Foundation Model ECML KDD2026
链接: https://arxiv.org/abs/2608.04174
作者: Gabriel da Costa Merlin,Diego Furtado Silva
类目: Machine Learning (cs.LG)
*备注: Accepted at ECML PKDD 2026. To appear in Springer LNCS. 18 pages; 7 figures
Abstract:Time series data are ubiquitous in practical applications, where classification (TSC) and extrinsic regression (TSER) have emerged as essential tasks for obtaining value from temporal sequences. While the literature has seen significant progress through feature-based and deep learning models, existing methods often focus either on the quality of feature extraction or on the intrinsic predictive power of complex architectures applied to raw data. This division creates a gap between the control offered by feature engineering and the automated performance of end-to-end models. This paper proposes TS2TabPFN, a framework that bridges this gap by integrating explicit feature extraction with TabPFN 2.5, a cutting-edge foundation model for tabular data, to leverage its predictive capabilities. Our extensive experimental evaluation demonstrates that TS2TabPFN significantly outperforms state-of-the-art models in TSER tasks with statistical significance, providing a robust and efficient alternative for TSC and surpassing most of the currently best-performing algorithms. These results suggest that combining foundation models with structured features overcomes single-paradigm limitations, establishing a new time series state-of-the-art.
[LG-55] Understanding Fault Tolerance of Adversarially Robust Pruned Models
链接: https://arxiv.org/abs/2608.04173
作者: Manali Dangarikar,Cory Merkel
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
*备注: 6 pages, 4 figures, Accepted for oral presentation at the 2026 IEEE National Aerospace and Electronics Conference (NAECON 2026)
Abstract:Deep neural networks (DNNs) deployed on resource-constrained neuromorphic hardware face three concurrent challenges: the need for model compression through pruning, vulnerability to adversarial input perturbations, and susceptibility to hardware-induced weight faults such as stuck-at-zero errors. While each of these factors has been studied in isolation, their combined effects on model reliability have received little attention. This paper presents an empirical investigation of how pruning, adversarial training, and hardware fault injection interact to affect the robustness of convolutional neural networks. Using a compact three-layer CNN trained on MNIST, we conduct three experiments: (1) comparing the fault tolerance of naturally and adversarially trained models under simultaneous hardware faults and adversarial attacks, (2) evaluating how pruning affects adversarial robustness, and (3) characterizing the joint accuracy surface across fault rates, adversarial perturbation magnitudes, and pruning levels. Our results show that adversarial training improves robustness against input perturbations but increases sensitivity to stuck-at-zero weight faults. Contrary to intuition, pruning did not significantly increase fault sensitivity, and varying the pruning level had little effect across fault rates and attack strengths. These results highlight the need to jointly consider adversarial robustness and hardware reliability.
[LG-56] MINT: Tensor Decomposition on Stacked Recurrence Matrices for Time Series Data Mining
链接: https://arxiv.org/abs/2608.04157
作者: Kaamil Kaka,Audrey Der,Evangelos E. Papalexakis,Zachary Zimmerman,Vikram Jayaram
类目: Machine Learning (cs.LG)
*备注:
Abstract:Recurrence plots are a time series data mining primitive applied to a variety of domains (e.g. star light curves, sound waveforms, CCT telemetry). This work proposes tensorized self-similarity matrices as a primitive for univariate time series datasets ( N\times n ) of N time series of length n with a subsequence window of length m , and whose tensor-based nature is naturally extensible to multivariate datasets. The proposed method to compute this primitive computes dot plots of size N \times (n-m+1) \times (n-m+ 1) from these datasets, where the subsequent tensor is mined using tensor decomposition methods to mine for co-clustered patterns. We demonstrate our results in mass rapid transit, electricity demand, wind turbine, and car traffic data, finding the MINT pipeline effectively co-clusters cross-sensor patterns in highly regular datasets containing motifs at regular intervals.
[LG-57] Sublogarithmic Swap Regret in Multiplayer General-Sum Games via Hybrid Regularization
链接: https://arxiv.org/abs/2608.04149
作者: Taira Tsuchiya
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注: 25 pages
Abstract:Swap regret governs the rate at which uncoupled learning dynamics converge to correlated equilibria in multiplayer general-sum games. Under full-information feedback, the best previous guarantee when every player follows the same dynamics grows logarithmically in the horizon T . We construct uncoupled dynamics under which every player incurs only O(nm^2\sqrt\log m\log T) swap regret, where n is the number of players and m bounds the number of actions per player. To our knowledge, this is the first sublogarithmic individual guarantee in this setting, and it implies that the time-averaged product distribution of play is an O(nm^2\sqrt\log m\log T/T) -approximate correlated equilibrium. The key algorithmic choice is to combine the Blum–Mansour reduction with optimistic follow-the-regularized-leader using a hybrid regularizer that separately weights negative Shannon entropy and the log-barrier: the entropy controls the optimistic prediction error, whereas the log-barrier controls the transition-matrix movement through its Bregman divergence. A new sensitivity theorem for stationary distributions of Markov chains, which involves neither mixing parameters nor the smallest transition probability, transfers this control to the played strategies and yields a simpler analysis without local-norm or self-concordance arguments. The guarantee is preserved by an adversarially robust variant that additionally ensures O(nm^2\sqrt\log m\log T+\sqrtmT\log m) swap regret against arbitrary utility sequences, and by a horizon-free variant that requires no prior knowledge of T .
[LG-58] CAMP: A Cycle-Aware Multi-Scale Patch Mixer for Time Series Forecasting
链接: https://arxiv.org/abs/2608.04051
作者: Jung Min Choi,Vijaya Krishna yalavarthi,Lars Schmidt-Thieme
类目: Machine Learning (cs.LG)
*备注:
Abstract:Real-world time series are often governed by recurring patterns, but their dominant periods may vary across datasets, forecasting settings, and individual input windows. Existing cycle-aware forecasters commonly rely on a single period selected at the dataset level, which can be restrictive when periodic behavior changes over time or when multiple cycles coexist. Moreover, patch-based models typically process all patch positions uni- formly, although patches farther from the forecast boundary may require broader contextual refinement, while recent patches contain information that should be preserved more directly. Af- ter cyclic behavior is removed, the remaining dynamics may also span multiple temporal resolutions and cannot be adequately de- scribed at a single scale. We introduce CAMP, a Cycle-Aware Multi-Scale Patch Mixer designed to address these challenges. The Adaptive Cycle Learning module identifies dominant fre- quencies separately for each input window and generates both historical and future cyclic components without requiring a pre- defined cycle length. The Horizon-Guided Patch Mixer intro- duces position-dependent refinement, allowing earlier patches to incorporate broader temporal context while preserving infor- mation close to the forecast boundary. CAMP further models the de-cycled residual through temporally aligned multi-resolution representations, enabling complementary dynamics at different scales to be captured within one forecasting framework. Across seven long-term forecasting benchmarks, CAMP achieves the best average MSE on six datasets and the best or tied-best MAE on six. It also obtains the highest MSE win count across sixteen settings on four PEMS traffic benchmarks.
[LG-59] FM4WiFi: Flow Matching for Multi-AP Coordination in Dense Deployments of Beyond Wi-Fi 8 Networks
链接: https://arxiv.org/abs/2608.04050
作者: Maksymilian Wojnar,Krzysztof Rusek,Katarzyna Kosek-Szott,Szymon Szott
类目: Networking and Internet Architecture (cs.NI); Machine Learning (cs.LG)
*备注:
Abstract:Wi-Fi networks are moving beyond random channel access toward tightly coordinated operation across access points (APs), a shift reflected in Wi-Fi 8’s multi-AP coordination (MAPC). However, the current MAPC specification restricts cooperation to AP pairs, fundamentally limiting the gains achievable in dense deployments and calling for scalable, network-wide coordination in beyond Wi-Fi 8 systems. We target coordinated spatial reuse (Co-SR), where APs transmit concurrently at reduced power. Effective Co-SR demands joint selection and configuration of AP-station transmissions, yet existing approaches simply do not scale: they rely on heavy signaling, slow convergence, unrealistic assumptions, and often require computation time that explodes with network size. We introduce FM4WiFi, a generative ML pipeline that addresses these limitations by producing high-quality Co-SR configurations in a single inference step. FM4WiFi integrates (i) an autoencoder that learns compact latent representations of network states, (ii) a flow-matching generative model that synthesizes feasible Co-SR configurations (including rate control, absent from prior work), and (iii) a surrogate rate predictor that allows rapid, large-scale Co-SR candidate evaluation without dependence on a live system or digital twin. Across extensive evaluations (including experimental validation), FM4WiFi matches or exceeds state-of-the-art baselines at medium-to-large scales and scales to 30+ APs with sub-second inference. Extensive ablation studies validate each modeling and optimization choice.
[LG-60] actus: Open-Vocabulary Object Recognition from Low-Cost Pressure Arrays
链接: https://arxiv.org/abs/2608.04043
作者: Abdul Basit Tonmoy
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注: 6 pages, 4 figures. Weights: this https URL
Abstract:Resistive pressure arrays are the cheapest and most widely shipped tactile sensors, yet tactile representation learning has concentrated on optical sensors that image a deforming gel. We present Tactus, an open model that answers text queries from pressure data alone: on the STAG benchmark (27 objects, held-out recordings), it reaches 0.771 +/- 0.062 top-1 over four runs (top-3 0.935), matching, and at best exceeding, the dataset’s supervised closed-set CNN at 0.76, with no trained classifier head. The recipe is small-data: 187 training recordings, masked-autoencoder pretraining on 144k unlabeled same-sensor frames, and the sensor’s own calibration affine, which recovered more accuracy than every architecture change combined. The released model’s errors concentrate in a few contact-ambiguous classes, are uncorrelated with text-target geometry (Spearman rho = 0.05 over 702 class pairs), and survive paraphrased and even bare-name queries within one point; two diverse frames recover 89% of eight-frame accuracy. Failures are reported with equal precision: cross-sensor pretraining pooling gave no gain, vision co-training degraded touch, and a mis-normalized input pipeline silently discarded 97% of the sensor’s dynamic range while producing plausible intermediate results. Weights, code, and the memory layer the model plugs into are released openly.
[LG-61] An Explainable LLM Agent Layer for Open-World Anomaly Detection in Oil Wells
链接: https://arxiv.org/abs/2608.04041
作者: Lucas Gouveia Omena Lopes,Thales Miranda de Almeida Vieira,Eduardo Toledo de Lima Junior,William Wagner Matos Lira
类目: Machine Learning (cs.LG)
*备注:
Abstract:Open-World Learning (OWL) pipelines for oil well anomaly detection have recently been shown to combine autoencoder-based detection, multiclass classification, and Mahalanobis-based novelty detection on the public 3W dataset. These pipelines answer \textitwhat happened, but they do not explain \textitwhy the model believes it or \textitwhat the operator should do next, and they do not put a human-readable name on the novelty clusters they discover. This paper evaluates a Large Language Model (LLM) agent layer placed downstream of the OWL pipeline, designed as a \textbfcompanion to the published upstream methods rather than a replacement. Using the Qwen3.5-397B-A17B Mixture-of-Experts model served via NVIDIA NIM, the agent receives structured sensor metrics and upstream classification or novelty assertions, and returns natural-language justifications, confidence-ranked critiques, and consolidated names for detected novelties. Across three studies spanning 989 real well-file segments from the 3W dataset, the agent achieved 35.1% top-1 / 63.9% top-3 (95% CI [56.9, 70.4]) classification on all nine classes, 71.7% top-2 validation [64.8, 77.6] with precision 0.91 [0.84, 0.95] across 7 probed classes, and 89.7% novelty detection [87.0, 91.9] with stable cluster naming on 5 of 7 hidden classes. The agent is not a standalone classifier. Its role is to: (1) confirm upstream decisions when sensor evidence supports them, (2) justify decisions in sensor-grounded language operators can audit, (3) flag disagreement when upstream labels are implausible, and (4) name novelties so that clustered unlabeled events arrive at the engineer with a consolidated human-readable label. The goal is to close the explainability gap that currently blocks deployment of OWL pipelines in operational settings.
[LG-62] Learning to Resolve Neutron Resonances with Fully Convolutional Neural Networks
链接: https://arxiv.org/abs/2608.04027
作者: Nataly R. Panczyk,Athanasios Stamatopoulos,Josef Svoboda,Majdi I. Radaideh
类目: Machine Learning (cs.LG); Applied Physics (physics.app-ph)
*备注: 29 pages, 11 figures, and 4 tables
Abstract:This work investigates the feasibility of augmenting traditional R-Matrix codes with a robust machine learning framework for automatically detecting neutron resonances in transmission spectra. Neutron transmission data are often complex and noisy, making them difficult to analyze using traditional peak-identification methods. The state-of-the-art R-Matrix codes currently used by physicists to fit these data often depend on prior evaluations and require substantial manual effort. This preliminary study demonstrates a method for accelerating the post-experimental processing of neutron transmission data and reducing bias associated with dependence on prior evaluations. We employ a fully convolutional neural network to classify individual points as belonging to resonance or non-resonance regions in seven transmission spectra—two evaluated and five experimental. Although the model achieves classification accuracies in the range of 93%, further analysis shows that this metric overstates its ability to generalize. Building on our prior analysis in PHYSOR 2026, we find that, despite the inclusion of additional training data, the method does not generalize reliably to previously unseen isotopes. To address these limitations, future work should evaluate whether a larger and more diverse training dataset can produce a generalizable model and should incorporate known physical characteristics of neutron resonances to improve model performance.
[LG-63] Stable Density Ridges: Consistency and Convergence of Subspace Constrained Mean Shift
链接: https://arxiv.org/abs/2608.05112
作者: Wanli Qiao
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 40 pages, 5 figures
Abstract:The Subspace Constrained Mean Shift (SCMS) algorithm is a popular nonparametric method for extracting density ridges, which serve as a low-dimensional representation of high-dimensional data. It is a widely held belief in the literature that SCMS trajectories converge to the classical density ridge, which we call the “static ridge”, defined via the density gradient and the eigenvalues and eigenvectors of the density’s Hessian. In this paper, we demonstrate that this assumption does not hold in general, as the static definition fails to account for the rotation of the trailing eigenspace along the continuous flow of the algorithm’s underlying vector field. To resolve this, we propose a paradigm shift by introducing the “stable ridge”, a novel geometric structure defined through the lens of dynamical systems and the Jacobian of the projected density gradient. We prove that this stable ridge is the true theoretical target of the SCMS algorithm. Building upon this foundation, we develop a generalized SCMS framework utilizing a constant step size, establishing its uniform R-linear convergence and topological surjectivity onto the stable ridge. We further derive the rates of convergence for estimating the stable ridge in terms of the Hausdorff distance. Finally, we expose that the original SCMS algorithm suffers from polynomial-time computational complexity, which is caused by implicitly coupling the step size to the smoothing bandwidth via the Mean Shift operator, and demonstrate how our generalized framework provides a statistically consistent and more efficient solution.
[LG-64] he Neural Echo: A Signal Processing Perspective for Understanding Neural Networks
链接: https://arxiv.org/abs/2608.04864
作者: Chongbiao Wang,Daniel Gaa,Joachim Weickert,Karl Schrader
类目: ignal Processing (eess.SP); Machine Learning (cs.LG)
*备注:
Abstract:We introduce the neural echo as a tool for understanding the behavior of neural networks. It generalizes the model-based concepts of impulse responses, diffusion echoes, and filter echoes to learning-based methods. It provides local, space-adaptive impulse responses and filter kernels for a neural network, its so-called echoes. These echoes depend on the input image and can be visualized to understand the learned dynamics of the network via an affine mapping. Neural echoes build a bridge from classical signal processing to modern explainable AI. They are very general and can be applied to both image-to-image and classification networks, with convolutional or fully connected structure, of feedforward or recurrent type, including modern transformer networks. Network differentiability is not required. In the differentiable case, neural echoes comprise concepts based on the network Jacobian, such as saliency maps and the analysis of adversarial perturbations, as special instances. As a simple blueprint to explain our framework, we derive neural echoes for the denoising convolutional neural network (DnCNN). Our experiments suggest that this network weights pixels based on their spatial and gray value distances. This not only clarifies its behavior, but also shows that it can reproduce key concepts of classical model-based denoisers such as bilateral filtering.
[LG-65] Nonparametric Goodness-of-fit Testing under Covariate Shift
链接: https://arxiv.org/abs/2608.04860
作者: Zhen Hou,Dong Xia
类目: Methodology (stat.ME); Machine Learning (cs.LG); Statistics Theory (math.ST); Machine Learning (stat.ML)
*备注:
Abstract:This paper develops procedures for nonparametric goodness-of-fit testing under covariate shift, where labelled data are drawn from a source population but goodness-of-fit is evaluated for a target population. The distribution mismatch is quantified by either a bounded moment condition or a sub-exponential tail condition on the target-to-source density ratio. Our method combines truncated importance-weighting kernel ridge regression with a multiplier bootstrap to construct confidence sets for the regression function. The truncation stabilizes the importance- weighting kernel ridge regression as well as the bootstrap calibration, making our approach applicable even when the density ratio has heavy tails. We prove nonasymptotic validity and sharpness of the resulting confidence sets under suitable operator compatibility conditions, and establish explicit error rates for coverage probability under specific conditions on the target- to-source density ratio and on the spectral decay of the kernel integral operator. Numerical experiments corroborate our theoretical findings.
[LG-66] Intrinsic-Hybrid Latent Diffusion Models for Generative Modeling on Unknown Manifolds
链接: https://arxiv.org/abs/2608.04827
作者: Yizhu Wang,Mu Niu,Xiaochen Yang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:We introduce the Intrinsic Hybrid Latent Diffusion Model (ILDM), a generative framework that integrates probabilistic dimensionality reduction with geometry-aware diffusion on unknown manifolds. While diffusion models (DMs) have achieved state-of-the-art results in high-dimensional data synthesis, they rely on large training datasets and ignore intrinsic geometric structure. Latent diffusion models (LDMs) address the high dimensionality by learning a latent space, but they typically impose a Euclidean structure, failing to capture the underlying manifold geometry, especially problematic in data-sparse regimes. ILDM addresses these limitations by interpreting the latent space as a chart of an unknown Riemannian manifold, with geometry and uncertainty quantified through a probabilistic decoder. The forward process is a hybrid diffusion that switches between Riemannian and Euclidean dynamics based on local uncertainty, where the Riemannian component is governed by a probabilistic metric tensor derived from the decoder. To learn the generative dynamics, we introduce an approximate denoising score matching method tailored to the hybrid diffusion setting, enabling a backward process defined by hybrid Langevin dynamics. Experiments on COIL-100, MNIST, and cardiac MRI datasets demonstrate that ILDM significantly improves generation quality, achieving lower FID and LPIPS scores compared to standard diffusion and latent diffusion models.
[LG-67] Automatic Statistical Test for Rationally Expressible Algorithms by Selective Inference with Applications to Feature Selection
链接: https://arxiv.org/abs/2608.04667
作者: Teruyuki Katsuoka,Tomohiro Shiraishi,Shuichi Nishino,Ichiro Takeuchi
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 35 pages, 3 figures, 4 tables
Abstract:Selective inference (SI) provides statistically valid p -values for hypotheses selected by applying an algorithm to the data, correcting for the bias that arises when the same data are used both to select and to test a hypothesis. Developing an SI procedure for a new algorithm, however, has required an expert to derive, and then implement, the selection event, i.e., the conditions under which the hypothesis is selected. Repeating this specialized effort for every new algorithm is why exact SI has so far been available for only a narrow class. We propose AutoSI, a framework that removes this barrier in two ways. First, AutoSI constructs the selection event automatically from the algorithm’s individual operations, so the user only writes the algorithm as ordinary NumPy-like code and derives nothing by hand. Second, AutoSI broadens the class of selection events SI can handle: existing exact methods are limited to selection events characterized by linear or quadratic inequalities in the data, whereas AutoSI covers any algorithm expressible through rational functions of the data (ratios of polynomials). We prove that the p -values computed by AutoSI are exactly valid in finite samples. We demonstrate AutoSI on three feature-selection methods, each written in a few dozen lines of code. One of these methods, the lasso with its tuning parameter selected by cross-validated R^2 , cannot be handled within existing exact SI frameworks and is made possible by AutoSI. Experiments on synthetic and real datasets show that the resulting p -values control the type I error rate (i.e., the false positive rate) at the nominal level while retaining high power.
[LG-68] On MUON optimization: From non-convergence to an error analysis with Polar Express and the Newton-Schulz polynomial from implementations
链接: https://arxiv.org/abs/2608.04607
作者: Thang Do,Steffen Dereich,Arnulf Jentzen
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注: 82 pages
Abstract:Stochastic gradient descent (SGD) optimization methods are the standard instruments for the training of deep neural networks (DNNs). In many relevant artificial intelligence (AI) systems - such as popular large language models (LLMs)-not the standard SGD scheme is used as the optimization method but instead suitable accelerated variants of SGD are employed. One of the most popular methods of such accelerated SGD variants is the momentum orthogonalized by Newton-Schulz (MUON) optimizer proposed by Jordan et al. in 2024. The MUON optimizer exploits the special matrix structure of the weight parameters in the training of the DNNs and, in its original form, employs five Newton-Schultz (NS) matrix steps in each MUON iteration. In this work we propose and study a generalized variant of the MUON optimizer involving an arbitrary number of generalized NS steps with polynomials of possibly arbitrary high degree. The considered optimizer covers MUON with the original NS polynomial as well as MUON combined with the recently proposed Polar Express method as special cases. For a simple class of stochastic optimization problems (SOPs) we show for almost every mini-batch size that MUON fails to converge to the solution of the SOP as the number of gradient steps converges to infinity. We also establish an error analysis for MUON with the generalized NS steps that provides convergence rates in terms of the number of gradient steps and in terms of the size of the mini-batch. We illustrate our general error analysis for MUON in the case of several concrete examples including quadratic stochastic optimization problems (SOPs) as well as \ell_2 regularized logistic regression for binary classification. Comments: 82 pages Subjects: Optimization and Control (math.OC); Machine Learning (cs.LG) MSC classes: 68T05, 90C25, 65K05, 65K10, 60H35 ACMclasses: I.2.0; G.3; G.1.6; F.2.1 Cite as: arXiv:2608.04607 [math.OC] (or arXiv:2608.04607v1 [math.OC] for this version) https://doi.org/10.48550/arXiv.2608.04607 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-69] A Counterexample to Fourier Alignment in Single-Neuron Modular Addition
链接: https://arxiv.org/abs/2608.04451
作者: Gautam Neelakantan Memana
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
*备注: 18 pages
Abstract:We give a negative solution to MAIS-O60. We first construct an example in which an initially active ReLU neuron becomes completely inactive in finite time and thereafter remains frozen at a limit whose Fourier energy is equally distributed among all nonzero real frequency classes. The counterexample holds on an open set of initial conditions and therefore occurs with positive probability under Gaussian initialization. An appendix prepared by GPT-5.6 Sol strengthens the counterexample by showing that the same failure can occur for every Clarke trajectory from an open set of initial conditions, under the convention \mathrmReLU’(0)=0 , for smooth dead-zone approximations of ReLU, and for fixed-step full-batch gradient descent. Thus, single-frequency alignment is not a general consequence of training a single neuron on modular addition.
[LG-70] Dynamical Lie Algebras Cannot Describe Shallow QAOA: Crag ged Terrains Barren Plateaus and Empirical Hardness Models
链接: https://arxiv.org/abs/2608.04252
作者: Harrison Copp,Charlton Li,Anžej Margeta-Cacace,Amy Qiao
类目: Quantum Physics (quant-ph); Emerging Technologies (cs.ET); Machine Learning (cs.LG)
*备注: 18 pages, 9 figures
Abstract:The dynamical Lie algebraic (DLA) theory of variational quantum algorithms (VQAs) predicts commonplace exponentially vanishing loss and gradient variances for sufficiently deep parametrized circuits. In this work, we show that these predictions fail dramatically in the shallow-circuit (and particularly constant-depth) regime for the Quantum Approximate Optimization Algorithm (QAOA) applied to the maximum independent set (MIS) problem. In a large-scale numerical study across \sim 23,000 problem instances, we find that barren plateaus are rare, while landscapes whose variances polynomially increase with system size—which we term “cragged terrains”—are common across graph families. This aggregate polynomial growth persists both for generic, low-symmetry random graphs and for highly symmetric vertex-transitive graphs, indicating that DLA-based variance predictions do not describe landscape scaling in this regime. As a stopgap alternative to the theory, we train empirical hardness models to predict instance-wise hardness metrics for QAOA-MIS. While these models generalize poorly, they nonetheless recover the correct landscape scaling class (barren plateau vs. cragged terrain) with high fidelity. Taken together, our results identify shallow QAOA for MIS as a prototypical setting in which asymptotic, unitary-design-centric predictions may be fundamentally insufficient to describe shallow variational quantum algorithms more broadly, emphasizing the need for more empirically-informed models of VQA loss landscapes.
[LG-71] Multimodal Alignment Through Joint Kernel Entropic Gromov–Wasserstein Optimal Transport
链接: https://arxiv.org/abs/2608.04234
作者: Yixuan Florence Wu,Yilun Zhu,Naichen Shi
类目: atistics Theory (math.ST); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:We study the problem of aligning data from multiple modalities into a shared representation space, focusing on settings where strong pretrained unimodal encoders are available but cross-modal paired data are scarce. We propose a structure-preserving alignment framework, joint kernel entropic Gromov–Wasserstein Optimal Transport (JK-EGW), which maps multiple modalities into a common latent space by minimizing a quadratic optimal transport objective. JK-EGW leverages fine-grained similarity relationships within and across modalities to construct a global affinity kernel instead of relying on raw feature-space distances. Our framework naturally provides explicit control over the geometry and distribution of the latent embedding. On the theory side, we establish parametric sample complexity rate of n^-1/2 , matching the corresponding rates for standard, entropic and Gromov–Wasserstein optimal transport. On the algorithmic side, we derive a scalable alternating procedure to solve JK-EGW with entropic optimal transport (EOT) updates through a low-rank kernel approximation and a variational lifting. This lifting scheme effectively relieves the burden of a quadratic objective, and allowing us to take the advantage of existing EOT solvers. Empirically, we focus on post-hoc alignment of embeddings from pretrained encoders in data-scarce regimes, and show that our proposed method achieves improved multimodal retrieval performance compared to existing alignment baselines.
[LG-72] IDE: A Physically Diverse 3D Turbulence Benchmark Dataset for Advancing Scientific Machine Learning
链接: https://arxiv.org/abs/2608.04222
作者: Yilong Dai,Yiming Sun,Yiheng Chen,Shengyu Chen,Peyman Givi,Xiaowei Jia,Runlong Yu
类目: Fluid Dynamics (physics.flu-dyn); Machine Learning (cs.LG)
*备注: 20 pages, 13 figures, 28 tables. Dataset: this https URL ; Code: this https URL ; DOI: this https URL
Abstract:Turbulence is a central testbed for machine learning on physical dynamics because its governing laws are known exactly. However, most existing studies remain in 2D, while 3D turbulence has fundamentally different physics and is far more costly to simulate. Existing 3D resources also typically provide only one realization per configuration, making it difficult to distinguish learning the dynamics from fitting the statistics of a single flow. In this paper, we introduce TIDE (Turbulent Incompressible DNS Ensembles), a 256^3 DNS corpus and benchmark for 3D incompressible turbulence, with 15 configurations on eight controlled axes, independent ensembles, pressure fields, and equation-level verification. The benchmark includes five tasks, standardized learned baselines, controlled generalization splits, and physical-fidelity metrics alongside pointwise error. Across the main forecasting configurations, current learned models barely outperform persistence and still make about twice the error of a spectral solver given the true equations. Moreover, lower pointwise error can coincide with severely distorted small-scale dynamics, showing that accuracy alone does not ensure physical fidelity. Generalization results further show that most regime shifts reflect limited training coverage, whereas forced-to-decay transfer exposes a missing conditioning variable: operators trained under forcing continue to predict driven evolution when the external drive is removed. Closing these accuracy, fidelity, and conditioning gaps is the central open problem made measurable by TIDE.
[LG-73] From Financial Sentiment Classification to Return Predictability: A QLoRA Benchmark of Large Language Models
链接: https://arxiv.org/abs/2608.04200
作者: Fusheng Luo
类目: Mathematical Finance (q-fin.MF); Machine Learning (cs.LG)
*备注: Waiting to submit to a conference (ICAIF)
Abstract:Financial sentiment classifiers are commonly evaluated against human labels, but strong linguistic performance does not necessarily imply economically useful return predictability. This study separates these questions through two experiments. First, we construct a unified three-class benchmark from five financial text datasets and compare TF–IDF Naive Bayes, off-the-shelf FinBERT and Financial-RoBERTa encoders, zero-shot Qwen2.5-7B, and QLoRA-adapted Qwen2.5-7B, LLaMA3-8B, and Mistral-7B models. Mistral-7B achieves the best test accuracy (0.8840) and macro-F1 (0.8771), while QLoRA raises Qwen2.5’s macro-F1 from 0.7274 to 0.8615. An inverse-frequency class-weighted loss does not improve Qwen2.5. Second, we evaluate economic validity on a temporally separate 2019 Benzinga sample containing 10,637 unique headlines and 13,115 headline–stock observations for a fixed S\P~100 universe. Model probabilities are converted into continuous sentiment scores, aggregated by stock and signal date, and aligned with next-session returns over one-, two-, three-, and five-day horizons. All seven downstream models produce positive but small mean rank information coefficients at the one-day horizon; the largest is 0.0143 for FinBERT. None of the 28 model–horizon tests remains significant after Newey–West inference and false-discovery-rate correction. Portfolio results likewise fail to establish a robust advantage for the best-performing classifiers. The findings show that QLoRA is effective for financial sentiment adaptation, while also documenting a clear gap between classification accuracy and tradable cross-sectional signals.
[LG-74] Unifying quantum measurement constructions via a relative-entropy minimum change principle
链接: https://arxiv.org/abs/2608.04055
作者: Nana Liu,Mark M. Wilde
类目: Quantum Physics (quant-ph); Statistical Mechanics (cond-mat.stat-mech); Information Theory (cs.IT); Machine Learning (cs.LG)
*备注: 38 pages
Abstract:The minimum change principle provides an information-theoretic characterization of the Bayes reversal channel in classical probability theory and has recently been proposed as a framework for extending Bayes’ rule to quantum information theory. Using quantum relative entropy, we investigate a minimum change principle for the setting of quantum statistical inference. Specifically, we consider a forward process based on a classical-to-quantum preparation channel and a reverse process based on a quantum-to-classical measurement channel. We establish a closed-form characterization of measurements that are optimal for this principle, and this optimal measurement can be found via a dual formulation involving a single unconstrained Hermitian variable. This perspective allows us to recover some notable measurements within the same framework, including pretty good measurements and Fermi-Dirac thermal measurements, and we use it to discover a novel family that we call softmin thermal measurements. We further show that softmin thermal measurements arise as optimal solutions to entropy-regularized semidefinite optimization problems, demonstrating that they play a role for measurements analogous to that of thermal states in statistical mechanics. Finally, we prove an additivity property for the relative-entropy minimum change principle and investigate the performance of Fermi-Dirac thermal measurements for quantum hypothesis testing.
[LG-75] Statistical learning theory and Occams razor: Regularization
链接: https://arxiv.org/abs/2608.04049
作者: Tom F. Sterkenburg
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:The principle of Occam’s razor, which instructs us to prefer simplicity in inductive inference, has attracted much scrutiny both in the philosophy of science and in machine learning. In either field, however, a justification for the principle has been elusive. In this paper, building on an earlier “core argument,” I spell out a justification from statistical learning theory for the procedure of regularization: for trading off fit for simplicity. The means-ends argument is that in order to profit from theoretical reliability and “what-you-see-is-what-you-get” guarantees, one must implement a certain preference for simplicity over fit. This is a genuine methodological justification, which neither collapses to a purely pragmatic principle that we prefer simplicity for its own sake, nor to an ontological assumption that the truth is simple.
[LG-76] he Cost of Binarizing Survival Outcomes in Clinical Prognostic Modeling ALT
链接: https://arxiv.org/abs/2608.04046
作者: Shashank Yadav,David M. Routman,Andrew Y.K. Foong
类目: Quantitative Methods (q-bio.QM); Machine Learning (cs.LG)
*备注: Accepted to Machine Learning for Healthcare (MLHC) Conference 2026
Abstract:Survival analysis is an established framework for analyzing time-to-event data, yet many clinical machine learning studies still binarize the outcome before model training. This practice excludes censored patients, collapses temporal information into a single threshold, and can affect which features are selected as prognostically relevant. We examine the cost of this binarization in the context of Bayesian network (BN) feature selection, using two recent publications as case studies: one that applies BN-based feature selection to a head-and-neck cancer cohort and a second surgical cohort study that, while not BN-based, likewise binarizes its survival endpoint. We replace the binary scoring function with the Cox partial log-likelihood for feature-to-outcome edges, a modification we call the Survival-Aware Bayesian network, and recover prognostic features that binarization misses. Our ablation experiment confirms that the improvement is driven by the time-to-event scoring formulation rather than by retaining more patients. The results generalize across five endpoint-cohort combinations in head-and-neck cancer and extend to three further cancer types (breast, colorectal, and kidney). We propose that clinical studies with survival outcomes should use time-to-event methods by default, as binarization discards the prognostic signal retained by survival analysis.
[LG-77] A Multi-Cohort Validation of Censoring-Aware Conformal Lower Predictive Bounds for Pathology Survival Models
链接: https://arxiv.org/abs/2608.04025
作者: Mingi Hong
类目: Methodology (stat.ME); Machine Learning (cs.LG)
*备注: 22 pages, 3 figures, 2 main tables, and supplementary material. Preprint
Abstract:Whole-slide survival models commonly provide risk rankings without calibrated statements about individual event times. We evaluate fixed-cutoff drcosarc, a post-hoc conformal wrapper for discrete-time multiple-instance learning survival heads using frozen UNI2-h representations, in an internal 18-configuration sweep across five TCGA cohorts and an external five-configuration evaluation across three CPTAC cohorts. We distinguish configuration–fold–split summaries of the inverse-probability-of-censoring-weighted (IPCW) estimate and median lower predictive bound (LPB) from a hierarchy-aware patient-ensemble estimand of the mean drcosarc–naive LPB difference. At \alpha=0.1 , the drcosarc IPCW estimate was nearest 0.90 in KIRC, LUAD, and STAD. Patient-ensemble drcosarc–naive intervals excluded zero in KIRC, KIRP, STAD, UCEC, and CPTAC-CCRCC, but included zero in internal LUAD, CPTAC-LUAD, CPTAC-UCEC, and the internal LUSC extension. In a 20-replicate low-censoring semi-synthetic setting with known event times, drcosarc empirical coverage was 0.9129 [0.9053, 0.9207]. An exploratory analysis supported a head-error-by-censoring interaction within that data-generating process. In a two-cohort ABMIL sensitivity analysis, increasing the hazard grid to K=16 raised localized marginal IPCW estimates above the prespecified 0.87 threshold and yielded positive paired LPB differences, although worst-group estimates remained below 0.87. Overall, performance was cohort dependent, and its interpretation changed with the patient-level unit, estimand, and censoring assumptions.
[LG-78] Monsoon Mayhem to Market Waves: Forecasting Fisheries Resilience in Sri Lanka
链接: https://arxiv.org/abs/2608.04023
作者: Ruzaini Ahmed,Yohan Jayasinghe,Tharumini Gamage,Ifaz Ikram,Hasini Lawanya,Nirasha Munasinghe,Patalee Narasinghe,Nisansa de Silva,Sandareka Wickramanayake
类目: General Economics (econ.GN); Machine Learning (cs.LG)
*备注: Accepted to MERCon 2026
Abstract:Sri Lanka’s fisheries sector is important for jobs and food supply. Between 2019 and 2025, it faced several major problems at the same time, and how these events together affected fish production and prices is still not well understood. This study develops a framework to connect weather changes, major disruption events, fish production, and prices, with the goal of helping policymakers, traders, and supply chain managers make better decisions. Seasonal patterns are studied using STL decomposition. Spearman lag correlation is used to find delayed effects of climate on production. Interrupted Time Series (ITS) regression measures the impact of major events. SARIMAX models predict monthly production and prices. Hotspot detection identifies unusual patterns. The results show that marine and inland fisheries behave differently in terms of seasons and climate effects. Major disruptions caused different levels of impact, and in some cases, one sector helped compensate for another. These findings can support better planning, for example, improving infrastructure in high-risk areas, strengthening cold storage systems, and using early warning alerts for unusual events. Price forecasting tools should be used as decision-support tools, not as direct market signals.
附件下载
