本篇博文主要内容为 2026-11-09 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 (2026-11-09)
今日共更新745篇论文,其中:
- 自然语言处理共78篇(Computation and Language (cs.CL))
- 人工智能共224篇(Artificial Intelligence (cs.AI))
- 计算机视觉共119篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共215篇(Machine Learning (cs.LG))
- 多智能体系统共14篇(Multiagent Systems (cs.MA))
- 信息检索共17篇(Information Retrieval (cs.IR))
- 人机交互共30篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] A Separation Principle for Cooperative Multi-Agent Reinforcement Learning
【速读】：该论文旨在解决大规模协同多智能体强化学习（Multi-Agent Reinforcement Learning, MARL）中因智能体数量增加而导致的学习效率急剧下降的问题。具体而言，当多个智能体在存在噪声且解耦的动力学环境下运行，同时共享一个全局种群目标时，传统MARL方法由于需要联合学习所有智能体的状态-动作空间，导致计算复杂度呈指数级增长，难以扩展至大规模场景。其核心挑战在于如何在保持性能的同时实现可扩展的协作决策。解决方案的关键在于提出一种基于最优传输（Optimal Transport, OT）的分离原则（Separation Principle）：将整体种群成本函数上界表示为一个以单智能体条件成本函数作为运输代价的最优传输问题。这一理论突破使得原本复杂的多智能体协同问题被分解为两个独立任务——在单智能体层面学习目标条件化的策略，以及通过最优传输机制进行智能体间的高效协调。基于此原理，作者提出了SALT（Separation-based Assignment and Learning via optimal Transport）算法族，其通过标准强化学习方法训练单智能体策略，并利用最优传输实现动态分配与调度。由于学习过程仅作用于单智能体层级，SALT具备良好的可扩展性，可在训练阶段未覆盖的大规模车队中实现零样本迁移。实验结果表明，在南曼哈顿道路网络的车辆配送任务中，相比最短路径启发式方法，SALT将平均行驶时间降低15%-21%；在随机网格世界中，相较现有MARL基线方法，旅行时间减少23%-61%，并成功扩展至传统方法失效的大规模种群场景。
链接: https://arxiv.org/abs/2611.09644
作者: Lucia Pezzetti,Nicolas Lanzetti,Antonio Terpin,Florian Dörfler,Giorgia Ramponi
机构: ETH Zürich(苏黎世联邦理工学院); Caltech(加州理工学院); University of Zürich(苏黎世大学)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:We study cooperative multi-agent reinforcement learning (MARL) problems in which agents evolve under decoupled noisy dynamics but share a population objective. A canonical example is a mobility operator that orchestrates a fleet of vehicles to meet customer demand. The stochastic and time-varying nature of these problems calls for learning approaches, yet existing MARL algorithms do not scale: learning assignment and routing jointly over the full state-action space becomes prohibitive as the number of agents grows. Formulating the problem over the space of probability measures, we prove a separation principle: the population cost-to-go function is upper bounded by an optimal transport problem whose transportation cost is the target-conditioned single-agent cost-to-go function. This simplifies the multi-agent problem to learning a single-agent policy and coordinating the agents via optimal transport. Building on this principle, we introduce SALT (Separation-based Assignment and Learning via optimal Transport), a family of algorithms that learn a target-conditioned single-agent policy using any standard reinforcement learning algorithm, and then coordinate the agents via optimal transport. Since learning occurs at the single-agent level, SALT scales to fleet sizes where MARL baselines become intractable and transfers without retraining to fleet sizes unseen during training. In a vehicle-delivery experiment on the southern Manhattan road network, where MARL baselines are computationally prohibitive, SALT improves travel time by 15-21% over a shortest-path routing heuristic. In stochastic grid worlds, SALT reduces average travel times by 23-61% relative to MARL baselines and scales to population sizes where these methods fail.
[MA-1] Incentive Pathologies in a Live LLM Agent Arena: An Empirical Study of Concentration Reward-Rule Implementation Discrepancy and Network Dilution
【速读】：该论文旨在解决大型语言模型（LLM）代理系统中激励机制设计失效与行为异化问题，特别是在复杂自治代理环境中，激励政策调整后实际行为响应与预期目标之间的偏差。其核心挑战在于：现有激励机制在部署后未能有效引导代理产生高质量、可持续的推理行为，反而催生了策略性规避行为，甚至导致系统整体效率下降。解决方案的关键在于通过多维度实证分析揭示激励机制的“表面承诺”与“底层实现”之间的严重脱节——具体表现为：尽管官方宣布的“深度奖励倍数”（continuation bonus）为x10，但代码级重构显示其实际执行机制为一个四条件反滥用门控机制，仅在典型情况下兑现x2.85奖励，并永久阻断自我延续路径；此外，关键行为指标如前4个节点活动得分占比从97.2%骤降至49.4%，事件/节点比显著下降（Mann-Kendall检验，p=8.3e-5），表明系统内部分布趋于分散且活跃度降低。研究进一步指出，激励政策的公告与实际实施之间存在显著时间差（约40小时），凸显了透明度与可审计性的重要性。因此，该研究强调构建可追溯、可验证的公开同行评审机制作为事件记录，以提升激励设计的可信度与有效性。
链接: https://arxiv.org/abs/2611.09445
作者: Lei Zhenbing,Lei Chang
机构: Independent Researcher(独立研究员); University of Chinese Academy of Social Sciences (中国社会科学院大学)
类目: Multiagent Systems (cs.MA); Physics and Society (physics.soc-ph)
备注: 29 pages, 2 figures, 4 tables; data available at Zenodo DOI https://doi.org/10.5281/zenodo.23691548
Abstract:We audit one live LLM agent arena (AITRAP; 22 autonomous agent seats, ~96,000 reasoning nodes) over nine days (Oct 16-25, 2026), during which the platform deployed two bundled incentive changes: a progressive tax reform (V2.9.3, Oct 19) and a seed-inversion with explorer-side action update (V3.0, Oct 21). These serve as before-and-after markers, not cleanly isolated interventions. Four exploratory predictions are tested against ten data frames (F1-F10) using the Gini coefficient, Top-K concentration share, code-level reconstruction of the continuation-bonus mechanism, and the Mann-Kendall trend test. The Gini analysis is inconclusive at 95% confidence (bootstrap intervals straddle 0.7). The Top-4 node-activity-score share falls from 97.2% (Oct 19, pre-V2.9.3) to 49.4% (Oct 25, post-V3.0; 56.7% under a merge-seat assumption). The continuation bonus, announced as depth-based x10, resolves at code level to a four-condition anti-abuse gate paying x2.85 in the typical case and permanently blocking the self-continuation it superficially rewards; the gate was active ~40 hours before the V3.0 announcement. The events-per-node ratio declines significantly (Mann-Kendall S=-45, z=-3.94, p=8.3e-5) from 7.36 to 7.069. We discuss implications for incentive designers, the announcement-implementation gap, and public peer review as an event record.
[MA-2] Enabling Privacy-Preserving Natural Language Access to Industrial Smart Meter Data
【速读】：该论文旨在解决能源管理与电网运营中非技术型人员难以高效利用大规模、异构的智能电表数据进行异常检测、性能监控及战略决策的问题。核心挑战在于：一方面，从海量遥测数据库中提取可操作洞察需具备高级SQL与数据工程能力，形成显著的技术壁垒；另一方面，关键能源基础设施的数据隐私与安全要求严格，限制了敏感用户数据通过云端大语言模型（LLM）进行处理。为此，论文提出OpenMeter——一种无模式依赖的多智能体系统（Multi-Agent System, MAS），其关键解决方案是将自然语言查询自动转化为可执行代码与可读性分析结果，从而实现非专业人员通过自然语言对话与复杂数据库交互。实验表明，在资源受限的本地硬件上，完整的多智能体协同调度在执行准确率上相较最优单智能体配置提升6.1个百分点（EX-R）；而在前沿规模场景下，各组件仍具价值但应按需启用，说明全量协同的开销仅在强调本地化隐私保护的场景中具有合理性。该架构有效降低了技术门槛，同时兼顾数据安全，推动了遥测数据在电网日常运维与战略决策中的深度应用。
链接: https://arxiv.org/abs/2611.09372
作者: Carlos Martín-Abellán,Mark Kröll
机构: Valencian Research Institute for Artificial Intelligence (VRAIN); Universitat Politècnica de València; Know Center Research GmbH
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Databases (cs.DB); Multiagent Systems (cs.MA)
备注: 9 pages, 1 figure, 6 tables
Abstract:Energy management officers and grid operators increasingly rely on smart meter data for anomaly detection, performance monitoring, and strategic decision-making. However, extracting actionable insights from large-scale, heterogeneous telemetry databases requires advanced SQL and data engineering expertise, creating a critical bottleneck for non-technical personnel. Furthermore, the strict data privacy regulations and security constraints inherent to critical energy infrastructure often prohibit the routing of sensitive consumer data through cloud-based Large Language Models (LLMs). To address these challenges, we present OpenMeter, a schema-agnostic Multi-Agent System (MAS) designed to translate natural language queries into executable code and human-readable insights. This architecture lowers the technical barrier for day-to-day utility operations by empowering non-expert personnel to interact with complex databases through intuitive natural language dialogue, transforming raw telemetry into a highly valuable tool for routine grid workflows and strategic decisions. Multi-scenario ablation studies over this environment demonstrate that full multi-agent orchestration shows its largest observed accuracy gains on resource-constrained local hardware (+6.1 pp EX-R over the strongest single-agent configuration); at frontier scale, individual components retain value but should be enabled selectively, indicating that the cost of full orchestration is justified specifically where privacy demands local execution.
[MA-3] ReSUM: Reasoning -Driven Multi-Agent Framework for Query-Focused Summarization
【速读】：该论文旨在解决长时会议场景下查询聚焦摘要（Query-Focused Summarization, QFS）任务中存在的关键挑战，包括复杂的说话人动态、关键信息稀疏以及数据稀缺等问题。由于会议文本通常包含大量不连贯表达、冗长转录内容及复杂的语篇结构，现有端到端方法在处理此类任务时易出现信息遗漏与幻觉现象，即便采用GPT-4-mini、Gemini等先进大模型亦难以保证摘要的准确性和完整性。其核心问题在于：单一模型在一次处理中难以有效完成信息检索、证据提取与摘要生成之间的复杂推理协同。为此，本文提出ReSUM——一种由规划器协调的模块化多大模型框架，通过引入教师模型的推理知识，将检索器和摘要生成器分别设计为依赖查询与计划条件的可定制组件，从而增强对长上下文的鲁棒性，减少无关探索，并生成更忠实、全面的查询聚焦摘要。该方案的关键创新在于利用分阶段、模块化架构实现任务解耦，结合知识蒸馏机制提升各子模块的针对性与协同效率，显著提升了复杂会议场景下的摘要质量。
链接: https://arxiv.org/abs/2611.09349
作者: Sonali,Deeksha Varshney,Asif Ekbal
机构: 未知
类目: Multiagent Systems (cs.MA)
备注: 10 pages, 4 figures
Abstract:LLMs excel at language tasks, but long-meeting QFS remains difficult due to complex speaker dynamics, sparse key information, and limited data. Meetings involve disfluencies, long transcripts and complex discourse. QFS have high risk of information omission, hallucination, even for advanced models such as GPT-4-mini and Gemini. Existing end-to-end methods compress retrieval, evidence extraction, and summary generation into a single pass, failing to handle the complex reasoning required for targeted meeting summaries. We propose ReSUM, a modular multi-LLM framework coordinated by a planner that orchestrates retrieval and summarization agents. By distilling teacher reasoning into query- and plan-conditioned retrievers and summarizers, ReSUM enhances long-context robustness, minimizes irrelevant exploration, and generates more faithful and comprehensive query-focused summaries. Experiments show consistent gains on query-focused meeting summarization benchmarks.
[MA-4] Recursively Self-Improving Electromagnetic Spectrum Agents : Architecture and Operation
【速读】：该论文旨在解决电磁频谱代理（Electromagnetic Spectrum Agents, ESAs）在持续自主运行过程中，因环境与任务需求动态变化而导致现有能力无法适应新条件的问题。其核心挑战在于如何使代理具备持续自我演进的能力，以应对复杂多变的频谱操作场景。解决方案的关键在于提出递归自改进电磁频谱代理（Recursively Self-Improving Electromagnetic Spectrum Agents, RSI-ESAs），通过引入基于运行反馈的闭环学习机制，实现对自身能力的动态更新与累积优化。该方案的核心机制包括问题诊断、候选策略生成、验证评估及知识继承四个环节，利用有效任务版本的保留、修改记录和诊断知识作为指导，使代理能够在多轮迭代中积累任务执行能力与改进经验，从而实现性能的持续提升。案例研究显示，在共享频谱调度任务中，RSI-ESAs在多轮改进后显著提升了任务表现并实现了能力积累，验证了其有效性。论文还指出未来在更广泛任务场景与长期运行中的若干开放性挑战，并为后续能力演化研究提供了方向。
链接: https://arxiv.org/abs/2611.09318
作者: Baolin Chong,Dongfang Xu,Qiaosi Wang,Fuhui Zhou,Qihui Wu
机构: 未知
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Electromagnetic spectrum agents (ESAs) must contend with changing environments and task requirements during sustained autonomous operation, yet their existing capabilities may not remain suitable for new operating conditions. This article proposes recursively self-improving electromagnetic spectrum agents (RSI-ESAs) that use operational feedback to update their capabilities and build on prior improvements. We present their architecture and editable components, and describe a self-improvement process encompassing problem diagnosis, candidate generation, validation, and inheritance. By retaining effective task versions and using modification records and diagnostic knowledge to guide subsequent updates, RSI-ESAs can accumulate task capabilities and improvement experience across successive rounds. A shared-spectrum scheduling case study illustrates improved task performance and capability accumulation over multiple improvement rounds under the tested conditions. Finally, we discuss open challenges for broader tasks and long-term operation, outlining research directions for the continued capability development of ESAs.
[MA-5] From Emergence to Reversibility: Toward Reversible Multi-Agent Systems
【速读】：该论文旨在解决多智能体系统在持续学习与演化过程中，传统基于规则的管控机制逐渐失效的问题。随着智能体间知识共享、任务委派与故障自修复能力的增强，系统虽仍能实现原有功能，但已绕过既定的限制路径，导致监管失效——即表面合规的系统实际上已具备规避控制的能力。其核心解决方案是提出“可逆多智能体系统”（Reversible Multi-agent Systems）这一研究范式：确保在系统不断演化的过程中，授权监管者仍能有效限制特定禁止性行为，即使智能体通过学习保留了相关知识或重构了协作路径，也能维持控制效力。关键在于将控制焦点从依赖特定机制转向对“特定效应是否可被产生”的可逆性约束，即无论系统如何演化，目标禁止效应必须保持稀有；同时保障必要服务的持续运行。为此，论文识别出三个核心研究问题：追踪各能力依赖关系、在旧机制失效前完成控制权移交、区分合法修复与非法重构。研究进一步提出通过跨演化路径的控制评估、证据支持的控制声明验证以及权限分配机制来实现对系统干预能力的动态维护。
链接: https://arxiv.org/abs/2611.09054
作者: Zhengye Han,Linda Ng Boyle
机构: New York University (纽约大学)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Multi-agent systems become capable through institutional mechanisms, such as rules that route information to agents that act on it. These mechanisms are also natural places to restrict the system. Yet as agents learn, share memory, delegate, and repair failures, the collective can find other paths to the same effects: a restriction on the original mechanism still operates but no longer stops them, while routine performance and compliance checks pass. We propose \textbfReversible Multi-agent Systems as a research agenda: collectives that grow while an authorized supervisor can still restrict specified effects, keep them restricted as agents adapt, and reconstitute useful cooperation under renewed authority. What is reversed is the ability to produce a specified effect; restrictions must hold even if learned knowledge persists. The goal is to preserve effective intervention, not a particular mechanism: at every step of growth, some authorized intervention must keep a specified prohibited effect rare while maintaining required service. We identify three research problems: tracking what capabilities depend on, handing off control before an old mechanism loses its effect, and separating legitimate repair from unauthorized reconstruction. We outline how to evaluate control across growth paths, back control claims with evidence, and allocate the rights to restrict and restore.
[MA-6] Sharing the Road: LLM -Based Bargaining in Mixed-Autonomy Traffic
【速读】：该论文旨在解决混合自治交通中异质驾驶主体（自动驾驶汽车，AV；人类驾驶汽车，HV）在共享道路空间时，如何基于更丰富的个体特征与基于语言的推理机制演化出协商行为，并最终形成集体出行结果的问题。传统研究多依赖于显式设定的行为假设或奖励函数，难以刻画复杂情境下的动态互动。本文提出一种基于大语言模型（Large Language Models, LLMs）的协商框架，使不同目标导向、拥有私有信息的AV与HV代理在多样化交通场景下进行道路空间分配的协商。其解决方案的关键在于利用LLM强大的语义理解与对话生成能力，实现具有上下文感知和意图表达的自然语言协商过程，从而模拟真实世界中复杂的交互逻辑。通过控制实验分析交通构成、代理目标及领域知识的影响，研究发现：协商结果对相对交通密度敏感但不严格遵循比例分配；密度均衡状态下的协调难度更高；公平性目标虽显著提升协议达成率，但未必带来最优资源利用率；且领域知识对协商效果的影响因目标配置而异，凸显了公平性作用的角色依赖性。这些发现表明，基于LLM的协商机制为建模和设计混合自治交通中驾驶主体间的交互提供了新的范式。
链接: https://arxiv.org/abs/2611.09004
作者: Lingxiao Zhou,Di Chen,Jia Li
机构: Washington State University (华盛顿州立大学); University of California, Davis (加利福尼亚大学戴维斯分校)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Mixed-autonomy traffic involves heterogeneous driving populations competing for shared road space. Previous studies have shown the existence and emergence of collective rationality using game-theoretic and reinforcement learning approaches, which rely on explicitly specified behavioral assumptions or reward functions for driving agents. However, how bargaining behaviors between driving agents evolve and lead to collective outcomes based on richer agent characteristics and language-based reasoning remains less understood. To fill this gap, we develop a bargaining framework based on large language models (LLMs) in which population-level autonomous vehicle (AV) and human-driven vehicle (HV) agents negotiate road-space allocations under heterogeneous objectives and private information. Through controlled experiments varying traffic composition, agent objectives, and traffic-domain knowledge, we examine bargaining behavior, coordination, and resource utilization. Results show that negotiated allocations respond to relative traffic density without strictly following proportional sharing, while density-balanced states exhibit greater coordination difficulty. Fairness-oriented objectives substantially improve agreement rates but do not necessarily yield equally high road-space utilization. Furthermore, domain knowledge affects bargaining outcomes differently across objective configurations, shaping both coordination and the role-dependent effects of fairness. These findings demonstrate the potential of LLM-based bargaining for modeling and designing interactions among driving agents in mixed-autonomy traffic.
[MA-7] Phantom Consensus: Measuring Social Reasoning in Multi-Agent LLM Decision-Making NEURIPS2026
【速读】：该论文旨在解决多智能体大语言模型（multi-agent LLM systems）在集体决策过程中因对其他智能体行为的错误推理而导致的“虚假共识”（phantom consensus）问题。其核心挑战在于：当智能体接收到模糊的社会信号后，虽仍维持独立的私有判断，却会改变公开决策，并错误估计其他智能体的私有判断，即使群体私有判断并未支持表面达成的一致。这种现象表现为一种类多元无知（pluralistic ignorance）的集体认知偏差。解决方案的关键在于提出并验证一个名为PHANTOM-Bench的基准测试框架，该框架通过分离私有判断、公共决策、对群体私有判断的主观估计、透明度驱动的修复机制以及对照条件（如中性文本和明确投票跟随），系统性地量化此类失败。实验表明，模糊的公共信号可使决策准确率下降高达47.8个百分点，集体理性失效率达47.3%；而揭示各智能体的记录私有判断可在不暴露真实答案的情况下显著修复多数错误。研究强调，评估多智能体系统不应仅关注最终共识，更需检验该共识是否基于独立测量的私有判断，而非由社会信号诱发的虚假一致性。
链接: https://arxiv.org/abs/2611.08971
作者: Disha Sheshanarayana,Rajat Subhra Pal,Manjira Sinha,Tirthankar Dasgupta
机构: Manipal University Jaipur(曼加普尔大学贾伊普尔分校); TCS Research(塔塔咨询公司研究部)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted at NeurIPS 2026, SocialAgent Workshop
Abstract:Multi-agent LLM systems increasingly aggregate judgments across model instances, but reasoning about other agents can itself introduce failure. We study a setting in which LLMs shift their public decisions after observing ambiguous social signals while reporting inaccurate estimates of what other agents privately judged, even when the group’s recorded private judgments do not support the apparent consensus. We call this failure phantom consensus and study it as a pluralistic-ignorance-style pattern in controlled LLM collectives. To measure this phenomenon, we introduce PHANTOM-Bench, a benchmark that separates isolated private judgments, public decisions after ambiguous social signals, elicited estimates of the group’s private judgment, transparency-based recovery, and controls for neutral public text and explicit vote-following. Across nine closed- and open-weight models, ambiguous public signals consistently degrade decision quality and can turn correct private group judgments into incorrect public collective outcomes. Public accuracy drops by as much as 47.8 percentage points, and collective rationality failures reach 47.3% in the most susceptible setting. We observe substantial errors in reported group-private beliefs alongside public-stage failures, while control conditions reveal model-dependent differences in susceptibility to ambiguous and explicit peer signals. Revealing the agents’ recorded private decisions repairs many failures without revealing the gold label, although recovery varies across models. Our results show that evaluating multi-agent LLM systems requires measuring not only final agreement, but whether that agreement reflects independently measured private judgments or an apparent consensus.
[MA-8] AIM: Adaptive Interaction in Multi-Agent Debate for Multimodal LLM Inference NEURIPS2026
【速读】：该论文旨在解决多智能体辩论（Multi-Agent Debate, MAD）在多模态任务中因固定交互模态导致的精度下降与令牌效率低下的问题。现有方法通常采用固定的文本或图结构进行智能体间证据交换，难以有效传递细粒度的多模态证据，从而引发准确率降低和因冗长描述带来的令牌开销增加。其解决方案的关键在于提出自适应交互框架（Adaptive Interaction in Multi-Agent Debate, AIM），通过动态选择最优的模态组合来实现智能体间的高效交互。AIM不仅支持传统的文本与图模态，还引入了任务相关的模态交互机制，允许智能体在特定任务中交换图像区域、音频片段或时空视频片段等目标区域信息。为确定每实例的最佳模态组合，AIM首先生成结构化基线响应，并提取可解释的路由特征，涵盖任务所需证据类型、基线代理置信度以及各模态对基线的额外信息增益。随后，一个轻量级路由器基于这些特征，规避“融合不足”（遗漏关键证据）与“过度融合”（冗余模态增加令牌开销）的风险，实现精准的模态组合选择。在涵盖三大类多模态任务的八个基准测试中，AIM均达到最高准确率；相比最优单模态基线，多模态融合提升准确率高达6.4%；而针对实例自适应调整模态组合，相较各基准中最优固定组合进一步提升4.2%准确率，同时减少高达53.4%的令牌消耗。
链接: https://arxiv.org/abs/2611.08950
作者: Wei Fan,JinYi Yoon,Bo Ji
机构: Virginia Tech(弗吉尼亚理工学院); Inha University(因帕大学)
类目: Multiagent Systems (cs.MA)
备注: Accepted in The Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS 2026)
Abstract:Multi-Agent Debate (MAD) improves reasoning on multimodal tasks, but its accuracy and token efficiency depend on how agents interact during debate. A largely overlooked design choice is which interaction modalities agents use to exchange evidence. Existing methods rely on fixed text- or graph-based interaction, which fails to convey fine-grained multimodal evidence and thus degrades accuracy while inflating token cost due to verbose descriptions. To address this, we propose Adaptive Interaction in Multi-Agent Debate (AIM), an adaptive routing framework that dynamically selects the optimal modality combination for inter-agent interaction in multimodal MAD. Beyond text and graph, AIM introduces task-dependent modality interaction, where agents exchange targeted regions of interest from the task-specific multimodal input, such as image regions, audio segments, or spatio-temporal video clips. To determine the best combination for each instance, AIM first generates a structured baseline response and extracts interpretable routing features that capture which evidence the instance requires, how confident the baseline agent is, and how much new evidence each modality would add beyond the baseline response. A lightweight router then uses these routing features to select a modality combination that avoids both under-fusion (i.e., agents miss decisive evidence) and over-fusion (i.e., redundant modalities inflate token cost) risks. Across eight benchmarks spanning three multimodal task families, \name achieves the highest accuracy. Combining modalities improves accuracy by up to 6.4% over the best single-modality baseline, and adapting the combination per instance adds up to 4.2% over the most accurate fixed combination of each benchmark with up to 53.4% fewer tokens.
[MA-9] ODDLLM : Specifying LLM -Agent Behavior for Reproducible Social Simulation NEURIPS2026
【速读】：该论文旨在解决现有基于智能体模型描述规范（ODD协议）在描述大语言模型智能体（LLM-agent）行为时的关键缺失问题，即缺乏对决定其行为的核心组件——提示词组装（prompt assembly）、记忆序列化（memory serialization）、输出解析（parsing）及重试逻辑（retry logic）——的标准化术语与文档结构。其解决方案的关键在于提出ODD+LLM，一个向后兼容的扩展框架，通过引入五个基于CoALA认知架构框架的文档模块，系统性地规范这些关键组件的描述。研究通过多宇宙分析（multiverse analysis）验证了这一缺口的实际影响：在复现一篇已发表的LLM命名游戏模拟时，忽略一个看似微小但关键的指令（模型应以局外人身份作答而非扮演玩家），导致收敛速度提升三倍且最终达成共识的比例从0.35飙升至1.00；同时，对28篇已发表的LLM-agent仿真研究的重建充分性审查显示，77%的研究未明确说明提示构建方式与输出解析策略，且无一完整披露其核心大语言模型调用细节。因此，ODD+LLM为领域提供了一个标准化的信息聚合场所，将原本分散于代码库、补充材料和未记录默认设置中的关键信息集中管理，推动可重复性与透明度的提升。
链接: https://arxiv.org/abs/2611.08932
作者: Justin Downes,Robby Milletich,Newel Hirst
机构: Amazon Web Services(亚马逊网络服务)
类目: Multiagent Systems (cs.MA)
备注: 18 pages, 2 figures, 6 tables. SocialAgent Workshop @ NeurIPS 2026
Abstract:The ODD protocol is the de facto standard for describing agent-based models, but it lacks vocabulary for the components that determine LLM-agent behavior: prompt assembly, memory serialization, parsing, and retry logic. We propose ODD+LLM, a backward-compatible extension adding five documentation blocks grounded in the CoALA cognitive-architecture framework. To show that the gap matters, we conduct a multiverse analysis [Steegen et al., 2016] of a published LLM naming-game simulation, varying six specification choices a reimplementer must make. Variants that are defensible readings of the paper stay within statistical noise of our baseline reconstruction. Missing one instruction that the paper states but is easy to overlook (that the model answers as an outside observer rather than as a player) raises the share of runs settling on the convention agents individually favor from 0.35 to 1.00 and makes convergence three times faster, in a reimplementation that uses a different model from the original. We complement the demonstration with a reconstruction-sufficiency review of 28 published LLM-agent simulations, finding that the details of how prompts are built and how model outputs are parsed are undetermined in 77% of cases, and that no paper fully specifies its main LLM call. ODD+LLM gives the field a standard home for information it already produces but currently scatters across code repositories, supplementary files, and undocumented defaults.
[MA-10] CrowdDirector: Runtime Orchestration of Crowd Simulation through Natural-Language Directives
【速读】：该论文旨在解决在虚拟人群仿真过程中，如何实现自然语言指令的高效、灵活且可靠控制的问题。传统手工编写规则虽执行高效但缺乏灵活性，而持续调用大语言模型（LLM）进行实时指令解析则带来重复延迟、计算成本高及潜在不支持动作的风险。其解决方案的关键在于提出一个名为CrowdDirector的框架：仅在用户添加或修改指令时调用一次LLM，将自然语言指令转化为结构化指令，并通过场景验证器（scene validator）检查指令中涉及的实体与动作是否存在于当前Unity场景中；经验证通过的指令被持久化存储并在仿真运行期间持续生效。每个仿真周期内，基于图神经网络（Graph Neural Network, GNN）根据当前场景状态和激活指令，为每个代理选择可用的动作与目标，而移动、交互与动画等具体行为仍由Unity引擎处理。该GNN在离线模拟场景中训练，接受来自LLM的指导以学习对指令的响应策略。实验表明，场景验证器的可行性判断比LLM自身评估更接近人类判断，且在行为基准测试中，CrowdDirector比手工规则更准确地遵循用户指令，同时避免了频繁调用LLM，从而在固定规则与持续LLM控制之间提供了一种实用的折中方案。
链接: https://arxiv.org/abs/2611.08763
作者: Xiaohan Sun,Vuk Radovanovic,Merce Alvarez,Carol O’Sullivan,Yiorgos Chrysanthou,Panayiotis Charalambous
机构: Trinity College Dublin (都柏林三一学院); CYENS - Centre of Excellence (CYENS - 创新中心)
类目: Graphics (cs.GR); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Natural language offers an intuitive way to direct a virtual crowd while a simulation is running. Handwritten rules execute efficiently but cannot respond flexibly to varied instructions, whereas asking a large language model (LLM) to control the crowd at every update introduces repeated latency, cost, and the risk of unsupported actions. We present CrowdDirector, a framework that uses an LLM only when the user adds or changes an instruction. The LLM translates the request into a structured instruction, and a scene validator checks whether the entities and actions it references exist in the current Unity scene. Accepted instructions are stored and remain continuously active as the simulation is running. At each simulation tick, a graph neural network uses the current scene state and active instruction to choose an available action and target for each agent, while Unity handles movement, interactions, and animation. The network is trained offline on simulated scenarios, with LLM guidance on how agents should respond to instructions. Our evaluations show the scene validator’s feasibility decisions matched human judgements more often than the LLM’s own assessment, and on a behavioural benchmark CrowdDirector follows user instructions more closely than handwritten rules while avoiding repeated LLM calls. It therefore provides a practical middle ground between fixed rules and continuous LLM control.
[MA-11] Divide Diagnose and Compose: Optimizing Agent ic Pipelines and Harnesses
【速读】：该论文旨在解决构建高效、低成本且低延迟的生成式智能体（Agentic）流水线时，如何在异构且相互耦合的设计空间中进行联合优化的问题。现有方法或依赖人工搜索（成本高、耗时长），或局限于预定义子空间内优化，且缺乏对决策变更的结构化证据支持。其解决方案的关键在于提出分解式信用分配框架（Decomposed Credit-Assignment Framework, DeCAF），该框架能够根据具体任务与基准动态适配设计空间，独立优化提示（prompt）、拓扑结构、模型选择及工具策略等维度，并为每个决策生成可追溯的推理轨迹。DeCAF通过多代理协作实验，分别在受限编辑范围内优化各维度，同时联合评估性能、推理成本与延迟；最终通过组件级轨迹评估器，基于各代理指令生成角色特定评价标准，定位表现不佳的组件并指导跨维度迭代改进。实验表明，在多样化智能体基准上，DeCAF相较最优基线平均提升任务性能2.58个百分点，同时降低4.30%的推理成本和42.45%的端到端延迟。
链接: https://arxiv.org/abs/2611.08723
作者: Divij Handa,Christopher Tensmeyer,Jihyung Kil,Vlad I. Morariu
机构: Arizona State University (亚利桑那州立大学); Adobe(Adobe)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Building agentic pipelines that achieve high task performance while keeping cost and latency low requires jointly optimizing heterogeneous and coupled design choices, including prompts, topology, model selection, and tool policies. Manually searching this combinatorial design space is expensive and time-consuming, while existing automated optimizers typically operate within a predefined subset of this space, modify the pipeline monolithically, or provide little structured evidence for pipeline changes. To mitigate these challenges, we introduce the Decomposed Credit-Assignment Framework (DeCAF), which adapts the design space to the given benchmark and pipeline, optimizes each dimension separately, and generates reasoning traces for every decision. Given a task specification, benchmark examples, and a seed pipeline, DeCAF identifies relevant optimization dimensions and generates diverse experimental hypotheses for each. Independent coding-agent sessions then optimize each dimension under a restricted edit scope, jointly evaluating performance, inference cost, and latency while recording all hypotheses, edits, and outcomes of each experiment. A final synthesis session combines compatible findings across all dimensions and uses a component-level trajectory evaluator, which derives role-specific rubrics from each agent’s instructions, to localize underperforming components and guide cross-dimension revisions. Across a diverse suite of agentic benchmarks, DeCAF improves average task performance by 2.58 percentage points, while reducing inference cost by 4.30% and end-to-end latency by 42.45% compared to the best overall baseline.
[MA-12] s3: Simple Scaling of Swarms via Cooperative Multi-agent RL
【速读】：该论文旨在解决多智能体系统中协作组织缺乏自适应性的问题，即当前系统中的协作模式通常依赖于预先设定的、静态的结构，无法根据智能体行为或任务需求动态调整。其核心解决方案是提出s^3方法，一种首个让智能体自主决定独自工作时长以及何时共享成果或读取队友信息的多智能体强化学习框架，无需预设角色或组织结构。该方法的关键在于使智能体在共同目标驱动下自发形成互补性涌现角色，并实现高效协调与可扩展的团队协作。实验表明，在协调任务中，经过训练的双倍规模智能体群组解决问题数量较最强基线提升25倍以上；在并行探索任务中，通信智能体可减少近50%重复工作，性能优于最优单体模型（best-of-N）达9.4个百分点。因此，s^3实现了从“人为设计组织”到“目标驱动自组织”的范式转变，为由多个模型协同学习构建高性能系统提供了新路径。
链接: https://arxiv.org/abs/2611.08615
作者: Kunal Jha,Gonçalo Faria,Shuyue Stella Li,Scott Geng,Faeze Brahman,Noah A. Smith,Yulia Tsvetkov,Max Kleiman-Weiner,Natasha Jaques
机构: 未知
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
备注:
Abstract:Groups can achieve more than individuals in the same amount of time, but only if their members coordinate: they must decide who works on what and what to share. However, this organization in multi-agent systems today usually comes from arbitrary, static harnesses designed a priori without adapting to agent behaviors or task specifications. We ask whether agents can instead learn to organize themselves. We introduce s ^3 , the first multi-agent reinforcement learning method in which agents decide for themselves how long to work alone and when to share findings or read teammates’ messages, with no predefined organization or roles. Our s ^3 agents learn to take different approaches, read each other’s work, exhibit emergent roles that complement each other in the swarm, and generalize to bigger teams. We study both coordination, where agents must combine private information, and parallel exploration, where sharing discoveries can make search more efficient. With matched training and inference budgets, s ^3 , swarms evaluated with twice as many agents as in training solve over 25 \times more coordination problems than the strongest baseline, and in exploration, communicating agents reduce repeated work by nearly 50% and outperform best-of-N by up to 9.4 percentage points. Rather than being designed in advance, s ^3 enables organization to emerge from a shared goal. This opens the door to capable systems built from many models that learn to work together.
[MA-13] Aegis OS: A Transactional Multi-Tenant Operating System Runtime for Eliminating Authorization Fatigue in Autonomous Coding Agents
【速读】：该论文旨在解决自主编码代理（autonomous coding agents）在实际部署中面临的“安全-自治困境”：一方面，频繁的逐操作确认提示会显著中断工作流（导致45%-57%的工具调用受阻），引发授权疲劳；另一方面，完全放开权限（如使用–yolo或–dangerously-skip-permissions）则使主机文件、SSH/云凭证及API密钥面临提示注入和幻觉攻击的风险。此外，多个代理并发编辑共享工作树时易发生冲突覆盖，而通过大体量代码片段直接嵌入大语言模型（LLM）上下文窗口也带来高昂的令牌开销。为应对上述挑战，论文提出Aegis OS——一个基于Rust实现的事务性多租户代理操作系统运行时，通过模型上下文协议（Model Context Protocol, MCP）与Google Agent2Agent（A2A）协议暴露接口。其核心解决方案在于将授权边界从单个工具调用迁移至事务级工作区边界，并结合内核强制的OS沙箱（macOS Seatbelt与Linux Bubblewrap，含默认拒绝策略、–tmpfs HOME隔离及Unix套接字出站桥接）、基于写时复制（Copy-on-Write, CoW）的影子工作区（APFS clonefile；Linux FICLONE reflink）、身份绑定的HTTP出站代理（支持L7方法/路径作用域的秘密注入与出站数据防泄漏，DLP）、以及内容可寻址存储（Content-Addressable Storage, CAS）的A2A IPC总线，实现了高安全性与高效能的统一。实验表明，在1,500次受控调用与28,087次真实会话中，Aegis OS将提示频率降低94.19%-99.88%（受控）与90.7%-97.9%（真实会话，仅1.12%误报率），在8类共80项攻击中将攻击成功率从100.0%降至0.0%，并在APFS与Btrfs上分别实现78.1 MB的1,000文件仓库克隆耗时6.3毫秒与15.6毫秒，且可准确检测全部10例语法与语义合并冲突。
链接: https://arxiv.org/abs/2611.08516
作者: Hui Zhao(Heurion Research)
机构: Heurion Research(赫里昂研究)
类目: Operating Systems (cs.OS); Cryptography and Security (cs.CR); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: 19 pages, 3 TikZ figures, 13 tables. Open-source Rust implementation and reproducible benchmark suite available at this https URL
Abstract:Autonomous coding agents such as Google Antigravity (agy), Anthropic Claude Code (claude), and OpenAI Codex CLI (codex) execute multi-step software engineering workflows on workstations and cloud hosts. Existing agent environments face a Security-Autonomy Dilemma: operators must choose between per-action confirmation prompts, which interrupt workflows on 45%-57% of tool calls and induce authorization fatigue, or unconfined execution (–yolo / --dangerously-skip-permissions), which exposes host files, SSH/cloud credentials, and API tokens to prompt injection and hallucination risks. Furthermore, concurrent agents editing a shared working tree clobber one another’s changes, while passing large artifacts inline through LLM context windows incurs high token overhead. We present Aegis OS (aegis-os), a transactional, multi-tenant Agent Operating System Runtime in Rust exposed via the Model Context Protocol (MCP) and Google Agent2Agent (A2A) protocol. Aegis OS shifts the authorization boundary from individual tool calls to transactional workspace boundaries. Each agent runs inside a kernel-enforced OS sandbox (macOS Seatbelt and Linux Bubblewrap with default-deny --tmpfs HOME isolation and a Unix-socket egress bridge) backed by a Copy-on-Write (CoW) shadow workspace (APFS clonefile; single-open Linux FICLONE reflinks), an identity-pinned HTTP Egress Proxy with L7 method/path-scoped secret injection and outbound DLP, and an A2A IPC bus with Content-Addressable Storage (CAS). Across 1,500 controlled calls and 28,087 real calls across 52 coding-agent sessions, Aegis OS reduces prompts by 94.19%-99.88% (controlled) and 90.7%-97.9% (real sessions, 1.12% IPR), lowers attack success rate across 80 attacks in eight categories from 100.0% to 0.0% on macOS and Linux, clones 1,000-file (78.1 MB) repos in 6.3 ms on APFS and 15.6 ms on btrfs, and catches 10/10 syntactic and semantic merge conflicts. Comments: 19 pages, 3 TikZ figures, 13 tables. Open-source Rust implementation and reproducible benchmark suite available at this https URL Subjects: Operating Systems (cs.OS); Cryptography and Security (cs.CR); Multiagent Systems (cs.MA); Software Engineering (cs.SE) ACMclasses: D.4.6; D.2.6; I.2.11; C.2.4 Cite as: arXiv:2611.08516 [cs.OS] (or arXiv:2611.08516v1 [cs.OS] for this version) https://doi.org/10.48550/arXiv.2611.08516 Focus to learn more arXiv-issued DOI via DataCite
自然语言处理
[NLP-0] CodeQuant2: Enhanced Outlier Mitigation and Adaptive Clustering for Low-Precision LRMs
【速读】： 该论文旨在解决大推理模型（Large Reasoning Models, LRMs）在低精度量化（low-precision quantization）过程中性能下降的问题。尽管低精度量化可显著降低推理成本，但激进的量化策略常导致推理任务表现不佳，其根源在于推理初期阶段激活值（activation）波动剧烈，而传统校准方法因忽略早期解码状态，难以捕捉这些动态特征。为此，论文提出一种基于聚类的查找表（Look-Up Table, LUT）量化框架——CodeQuant2，其核心解决方案包括：通过旋转与缩放操作缓解激活异常值影响；利用聚类建模权重分布的非均匀性；根据各层量化难度自适应调整码书（codebook）大小；并基于对推理过程中激活行为的观察，优化校准样本选择以更准确匹配实际解码统计特性。结合定制化的LUT CUDA内核，CodeQuant2在保持高精度的同时，相较BF16实现最高达2.85倍的加速，有效解决了推理任务中低精度量化的性能瓶颈。
链接: https://arxiv.org/abs/2611.09868
作者: Xingyu Liu,Xiangyang Yin,Tianhua Xia,Haiyu Wang,Sai Qian Zhang
机构: Tandon School of Engineering (纽约大学坦顿工程学院); Courant Institute School of Mathematics, Computing, and Data Science (纽约大学柯朗数学科学研究所)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large reasoning models (LRMs) improve task performance by generating longer intermediate reasoning traces, but this substantially increases inference costs and makes low-precision quantization increasingly important. However, aggressively quantized models often perform poorly on reasoning tasks. We observe larger activation fluctuations during the first decoding steps of reasoning. Calibration that omits these early states may miss their activation fluctuations. To address these reasoning-specific quantization challenges, we propose \textbfCodeQuant2, a clustering-based LUT quantization framework for low-precision large reasoning models. CodeQuant2 mitigates activation outliers with rotation and scaling, uses clustering to represent non-uniform weight distributions, and adapts codebook sizes to the quantization difficulty of each layer. Guided by our observation of reasoning-time activation behavior, CodeQuant2 further improves calibration sample selection to better match actual decoding statistics. Combined with our customized LUT CUDA kernel, CodeQuant2 delivers substantial accuracy improvements on reasoning benchmarks while achieving up to 2.85\times speedup over BF16.
[NLP-1] How far does a hash-only sender go? A pre-registered map of sender memory versus bits for short text messages
【速读】： 该论文旨在解决在仅由接收端部署语言模型、发送端仅计算哈希值的轻量级通信架构下，能否以低于传统同领域训练发送端（如PPM order-1）的成本，实现对短人类文本的高效传输。其核心问题是：在资源受限的发送端场景中，如何在保证高送达率的前提下最小化传输开销（包括哈希比特数、失败时的负确认与回退数据）。解决方案的关键在于采用“仅哈希发送”机制，即发送端仅生成并传输消息的哈希值，接收端通过自身语言模型进行解码匹配；若匹配失败，则触发回退机制（2字节负确认+原始文本）。实验结果显示，该方案在1,000条来自Taskmaster-1的数据上成功首传857条，平均成本为192.8比特/消息，虽比PPM order-1多出18.4比特（置信区间10.9–26.3比特），但因超出预注册规则中“不超过1字节”的等效阈值，未被判定为等效。此外，该方案在内存占用介于8,190字节至24,844字节之间时，在比特开销上优于smaz2和Unishox2，而当发送端内存超过PPM order-1时，传统压缩算法更优。尽管扩大接收端搜索范围可恢复一半失败消息并将成本降低至低于PPM order-1，但因每次失败需约三分钟接收端处理时间且仍无法在多数链路上胜过现有方案，故在实际空口时间（airtime）上仍不具优势。本研究以严格可复现的方式报告了所有偏离预注册协议的细节，作为对当前轻量级通信范式边界的实证界定。
链接: https://arxiv.org/abs/2611.09848
作者: Marcelo Samoilenko
机构: Winserv
类目: Information Theory (cs.IT); Computation and Language (cs.CL)
备注: 10 pages, 1 figure, 2 tables. Code, data and per-message results: this https URL
Abstract:Can a sender that only computes hashes, with a language model only at the receiver, transmit short human text more cheaply than a classical sender trained on the same domain? We answer with one pre-registered run on 1,000 English user utterances from Taskmaster-1. The primary endpoint is the cost per message to deliver 100% of the messages (hash bits, plus a 2-byte negative acknowledgement and the raw text when the first attempt fails), paired against an in-domain PPM order-1 sender. The hash-only sender delivered 857 of 1,000 messages on the first attempt and cost 192.8 bits per message, 18.4 bits more than PPM order 1 (95% bootstrap interval 10.9 to 26.3); by the pre-registered rule it spends more and is not equivalent within one byte. As secondary results we map measured sender memory against bits. The hash-only sender (8,190 B on a host build, keyed BLAKE2b tag included, as in every method) was the cheapest point in bits from 8,190 B up to, but excluding, PPM order 1 at 24,844 B, 45 to 46 bits below smaz2 and Unishox2; from PPM order 1 upwards, in-domain senders spend fewer bits, except order 2 pruned to 16 KiB. In airtime it loses to PPM order 1 on all five links priced, to Unishox2 on four, and to raw text with a tag on two, because each failure costs an acknowledgement frame and a fallback frame. A wider receiver search recovers half of the failures and makes the cost 7.7 bits lower than PPM order 1, at about three minutes of receiver time per failure, and still loses in airtime on four of five links. We report the measurement as a delimitation, with every deviation of the run.
[NLP-2] BitDecoding-FP4: Native FP4 KV Cache Decoding via Algorithm-System Co-Design
【速读】： 该论文旨在解决长上下文大语言模型（LLM）服务中键值（Key-Value, KV）缓存占用过高导致的内存容量与带宽瓶颈问题。现有4位微缩量化格式（如NVFP4、MXFP4）虽能减小KV缓存体积并支持在FP4张量核心上运行注意力计算，但算法量化与系统加速之间存在根本性差距：传统量化方法导致严重精度下降，而高精度4位算法依赖向量码本，其查找开销又显著降低解码吞吐量。为此，论文提出BitDecoding-FP4——一种面向原生FP4 KV解码的算法-系统协同设计。其关键创新在于从算法和系统双层面协同优化：在算法层面，通过双残差查询量化（Dual-Residual Query Quantization）将查询拆分为两个FP4张量，使查询logit误差由60%降至1%；采用溢出感知概率缩放（Underflow-Aware Probability Scaling）避免E4M3格式下长上下文概率被舍入为零；引入异常值感知键量化（Outlier-Aware Key Quantization），结合静态平滑、Hadamard旋转与尺度搜索，在逐标记缩放下有效抑制键的异常值影响。在系统层面，通过转置softmax与融合概率量化消除softmax性能瓶颈，并在分页KV缓存上实现堆叠查询矩阵乘法加速（stacked query MMA）。实验表明，BitDecoding-FP4在密集模型与混合专家（MoE）模型（支持分组查询注意力GQA与多头局部注意力MLA）上，相较于BF16 KV缓存仅损失0.5点精度（使用NVFP4时），在MXFP4下损失1点，且在强化学习训练回放中表现与BF16 KV一致。在NVIDIA B300硬件上对Qwen3-4B与Qwen3-30B-A3B进行端到端服务测试，其峰值吞吐量相比BF16 KV提升2.4–3.0倍。
链接: https://arxiv.org/abs/2611.09825
作者: Dayou Du,Woon Yee Ng,Guangming Sheng,Nick Lau,Jianyi Cheng,Luo Mai
机构: 未知
类目: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR); Computation and Language (cs.CL); Performance (cs.PF)
备注:
Abstract:The key-value (KV) cache dominates memory capacity and bandwidth in long-context LLM serving. While 4-bit microscaling formats (NVFP4, MXFP4) shrink the cache and run attention on FP4 Tensor Cores, a fundamental gap remains between algorithmic quantization and system-aware acceleration: existing FP4 KV systems use vanilla quantization that causes severe accuracy drops, whereas accuracy-oriented 4-bit algorithms rely on vector codebooks whose lookup overhead degrades decoding throughput. We present BitDecoding-FP4, an algorithm–system co-design for accurate and efficient native FP4 KV decoding. We mitigate three error sources: Dual-Residual Query Quantization splits queries into two FP4 tensors, cutting query logit error from 60% to 1%; Underflow-Aware Probability Scaling prevents E4M3 scales from rounding long-context probabilities to zero; and Outlier-Aware Key Quantization combines static smoothing, Hadamard rotation, and scale search to suppress key outliers under token-wise scaling. At the system level, we remove the softmax bottleneck via transposed softmax with fused probability quantization and stacked query MMA on paged KV caches. Across dense and MoE models with GQA and MLA, BitDecoding-FP4 stays within 0.5 points of BF16 KV accuracy with NVFP4 (1 point with MXFP4), and tracks BF16 KV in RL training rollouts. In end-to-end serving on NVIDIA B300 with Qwen3-4B and Qwen3-30B-A3B, it delivers 2.4-3.0 \times higher peak serving throughput than BF16 KV. The implementation is available at this https URL.
[NLP-3] Compaction decay and reinforcement in a two-tier conversational memory: a paired evaluation against retrieval-augmented generation
【速读】： 该论文旨在解决对话智能体在记忆管理中的核心挑战：如何高效地决定保留哪些信息以支持后续的问答任务。传统方法采用检索增强生成（Retrieval-Augmented Generation, RAG），即保留完整的对话历史并进行搜索；而替代方案则是将每一轮对话压缩为原子化事实，再基于这些事实进行检索。研究的关键在于通过在LoCoMo数据集上的对比实验，系统评估两种记忆策略的性能差异。其解决方案的核心创新在于提出“压缩式记忆”（compaction）方法，即将对话内容提炼为独立的事实单元，从而显著减少提示词（prompt tokens）消耗（降低34%），并在多数评价类别中取得更高的准确率（71.4% vs. 64.6%，p = 5.5e-08）。然而，研究也揭示了该方法的局限性：当参考答案依赖于原始表述的措辞时，压缩策略反而表现更差，因它舍弃了语言表达形式而仅保留语义内容。这一结构性缺陷表明，压缩方法在回答事实性问题时占优，但在需要复现特定文本表达的任务中则处于劣势。此外，研究还提供了四个负面结果及其统计功效分析，增强了结论的可靠性。
链接: https://arxiv.org/abs/2611.09807
作者: Osamudiamen Edogun,Joshua Obigwe,Irfan Mehmood
机构: 未知
类目: Computation and Language (cs.CL)
备注: 7 figures. Data and code: this https URL
Abstract:A conversational agent that remembers has to decide what to keep. Retrieval-augmented generation keeps the transcript and searches it; the alternative is to compact each turn into atomic facts and search those. We compare the two on LoCoMo, with both arms answering all 1,985 questions in one window, against the same stores, with the same generator and judge, and both predictions on one row, which licenses McNemar’s exact test on the 512 questions where they disagree. On the four categories comparable work reports, compaction scores 71.4 per cent against 64.6, a lead of 6.8 points (p = 5.5e-08) at 34 per cent fewer prompt tokens. On all five, including the adversarial set comparable work drops, the lead falls to 2.0 and stops being significant, because the baseline leads adversarial by 14.6. That loss has a structure. Our lead falls as the reference answer demands more source wording, 9.6 points ahead at one or two reference words and 3.1 behind past eight, with token overlap moving the same way, so the judge is not the cause. Compaction keeps what a fact asserts and discards how it was said, so it wins where the answer is a value and loses where it is a phrasing. We also report four negative results with their power calculations. Comments: 7 figures. Data and code: this https URL Subjects: Computation and Language (cs.CL) Cite as: arXiv:2611.09807 [cs.CL] (or arXiv:2611.09807v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2611.09807 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-4] Maya a1: Do the Options Read the Model or Does Their Position? Order Sensitivity in a Non-Autoregressive Decision Model
【速读】： 该论文旨在解决生成式 AI 在多项选择题任务中答案对选项排列顺序敏感的问题，即模型输出的答案会因选项展示顺序不同而发生改变，从而影响结果的可靠性。其核心问题是：当前语言模型在处理多选题时，其决策过程受到输入序列中选项位置的干扰，导致答案不一致，尤其在低置信度情况下更为显著。解决方案的关键在于通过引入选项顺序随机化（option shuffling）的微调策略，使模型学习到对选项顺序无关的判断能力。实验表明，在2,818条经教师验证的数据上进行带顺序打乱的微调后，答案翻转率从11.4%降至7.2%，显著提升了鲁棒性；进一步采用置信度加权投票机制可实现69.7%的准确率，虽计算成本增加4.1倍，但性能优于单一前向传播。此外，使用作者提出的并行布局（parallel layout）进行微调，可将翻转率降至0.5%，展现出极强的稳定性。该研究揭示了现有模型在推理一致性方面的缺陷，并提出有效的训练范式以增强其抗顺序偏倚的能力。
链接: https://arxiv.org/abs/2611.09803
作者: Mohammed Nayeem Khan,Logesh L
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 24 pages, 13 figures, 23 tables
Abstract:Laya answers typed multiple-choice questions in one forward pass by scoring a hidden state at one [MASK] marker per option. We re-engineer the checkpoint (421M parameters; a 2-layer option head with no positional tensor; a documented but unused parallel layout) and test whether the answer depends on where an option is listed. Presenting 951 Qwen-verified questions under every cyclic rotation of their options (or six random permutations for 12 options), the released checkpoint changes its answer for 25.8% [22.9-28.7] of questions (19.7% on the 605 English-language items alone); accuracy with the gold option first versus last differs by 10.1 points, and the flips concentrate where the model’s own confidence is low (54% versus 8% across confidence terciles). Fine-tuning with Laya’s trainer on 2,818 teacher-verified rows with option shuffling cuts the flip rate to 7.2%; the identical rows without shuffling give 11.4%, so shuffling is responsible for 4.2 points of the improvement. A confidence-weighted vote over orderings gives 69.7% against 69.1% for a single base pass at 4.1x the compute. Labels are LLM-made and the test sets are teacher-filtered. We also fine-tune with the authors’ parallel layout on the same rows: flip rate 0.5%.
[NLP-5] Are These Programmers Ripe Enough to Eat? Evaluating How LLM s Handle Lexical Knowledge Conflicts in Translation
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在低资源语言和专业领域翻译任务中，因外部词汇资源（external lexical resources）与模型自身先验词汇知识（prior lexical knowledge）产生冲突而导致的翻译偏差问题。其核心挑战在于如何有效识别并控制这种冲突，以确保模型能够准确采纳由提示（prompt）提供的上下文性词汇知识（Contextual Knowledge, CK）。解决方案的关键在于构建一个受控的中文到英文翻译任务，通过人为设定熟悉词汇在语义类别内及跨类别的新含义，系统评估词频、句子上下文以及提示框架对模型依赖CK程度的影响。研究发现，模型对CK的依赖性随源词频率降低而增强，且在支持新含义的句子上下文中更为显著；尽管提示框架可部分缓解这一依赖，但多数模型仍倾向于采纳提示提供的新义，即使被明确指示忽略。此外，模型规模越大，越倾向于使用CK。三组后续实验进一步揭示了模态弱化、重复机制及推理过程对CK依赖性的调节作用具有模型特异性。综上，该研究通过受控实验揭示了不同条件与模型架构下CK依赖性的动态变化规律，为未来开发检测与评估词汇冲突的方法、指导模型在翻译中做出更合理的词汇选择提供了理论基础与实践启示。
链接: https://arxiv.org/abs/2611.09785
作者: Dongxu Wu,Sara Court,Micha Elsner
机构: The Ohio State University (俄亥俄州立大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:External lexical resources can help large language models (LLMs) translate texts in low-resource languages and specialized domains, but these resources may conflict with models’ prior lexical knowledge. To isolate this conflict, we construct a controlled Mandarin-to-English translation task in which familiar source words are assigned new meanings within and across semantic categories. We treat these prompt-supplied meanings as contextual knowledge (CK) and systematically evaluate the effects of word frequency, sentence context, and prompt framing on the translated outputs. Across nine models, CK reliance is greater for less frequent source words and in sentence contexts that support the supplied meanings. Prompt framing can reduce CK reliance, but seven of nine models favor CK even when explicitly instructed to ignore the supplied meanings. Within model families, larger models are more likely than smaller models to utilize CK in their outputs. Three follow-up experiments further show that modal weakening lowers CK reliance, the effect of repetition depends on prompt layout, and the effect of reasoning varies across models. Overall, our controlled study reveals how CK reliance varies across conditions and models, motivating new methods to detect and evaluate lexical conflicts and guide models’ lexical choices in LLM-based translation.
[NLP-6] ROSE: Risk-Aware Orthogonal Subspace Navigation for Lifelong Knowledge Editing 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）在持续知识更新过程中面临的终身知识编辑难题，核心挑战在于顺序编辑引发的局部性退化（Locality Erosion）与灾难性遗忘（Catastrophic Forgetting），即新知识的注入会破坏模型原有的预训练能力及先前的知识更新。其解决方案的关键是提出一种风险感知的正交子空间导航框架——ROSE（Risk-aware Orthogonal Subspace navigation for lifelong knowledge Editing）。ROSE通过将参数更新限制在已编辑子空间的正交补空间内，实现对历史知识的保护；同时引入动态计算的风险掩码机制，识别并保护对模型稳定性至关重要的特定事实参数子空间，从而降低遗忘风险。在推理阶段，ROSE采用多粒度知识融合策略，结合原型锚定语义门控与扰动驱动的子空间融合，精准激活与查询相关的编辑知识，同时最小化对无关输入的干扰。实验结果表明，ROSE在两个主流多模态基准上均达到当前先进水平，在五个关键指标上表现优异，并在超过1000次连续更新中保持卓越的稳定性。
链接: https://arxiv.org/abs/2611.09770
作者: Lingyun Song,Ziyao Chen,Kang Pan,Xiaolin Han,Yifei Zhang,Xiaoqi Wang,Yudai Pan,Xuequn Shang
机构: Northwestern Polytechnical University (西北工业大学); Shenzhen Research Institute of Northwestern Polytechnical University (西北工业大学深圳研究院); Key Laboratory of Big Data Storage and Management, Northwestern Polytechnical University, Ministry of Industry and Information Technology (工业和信息化部大数据存储与管理重点实验室, 西北工业大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal Large Language Models (MLLMs) require continuous factual updates to stay current, yet lifelong knowledge editing remains a daunting challenge. The primary obstacle is the severe Locality Erosion and Catastrophic Forgetting, where sequential edits inevitably interfere with pre-trained capabilities and prior updates. In this paper, we propose \textbfROSE, a \textbfRisk-aware \textbfOrthogonal \textbfSubspace navigation framework for lifelong knowledge \textbfEditing. ROSE addresses forgetting by aiming to restrict parameter updates to the orthogonal complement of previously edited subspaces. Crucially, it introduces a risk-guided mechanism that utilizes a dynamically computed risk mask to safeguard specific factual parameter subspaces essential for model stability. During inference, ROSE employs a multi-granularity knowledge integration strategy, featuring prototype-anchored semantic gating and perturbation-driven subspace fusion to precisely activate edited knowledge for relevant queries while limiting interference on unrelated inputs. Comprehensive experiments on two prominent multimodal benchmarks show that ROSE is competitive with state-of-the-art approaches across five key metrics, while maintaining outstanding stability over 1,000 sequential updates. Our code is available at this https URL.
[NLP-7] Language as Trait Politeness as Strategy: Controlled Tests of Linguistic Effects in LLM Social-Deduction Agents
【速读】： 该论文旨在探究语言表达方式（如语言复杂度与礼貌策略）对大型语言模型代理（LLM agents）之间信任建立及行为决策的影响，特别是在基于对话的协作或博弈场景中。其核心问题是：在保持语义内容一致的前提下，同一发言的自然语言实现形式（即语言风格）的差异是否会影响其他代理对其可信度的判断以及后续投票或夜间行动等行为选择。解决方案的关键在于采用LURK Werewolf框架，将话语的言语行为谓词（speech-act predicate）与其自然语言实现（natural-language realization）严格分离，通过仅改写语言实现部分、保留游戏内容不变，并利用保存的检查点进行重复回放实验，系统性地控制语言范围（linguistic range）和礼貌策略（politeness strategy）的变化。实验结果表明，在30场游戏、4,335次重播中，无论语言复杂度还是礼貌程度的变化，均未显著影响其他代理对说话者的信任度，也未在多重比较校正后发现对投票或夜间行动有显著影响；唯一接近显著的效应是语言范围偏离中性转述时，导致约4个百分点的投票变化，但此效应在统计上并不稳健。此外，在自然游戏中，88%的面子威胁性言语（face-threatening speech）表现为无修饰的直接表达（bald on-record），且与角色无关，暗示语言风格可能受任务结构或社会规范制约。因此，该研究揭示了当前大模型代理在社交互动中对语言形式变化的鲁棒性，提示语言风格的细微调整未必能有效操控其社会行为。
链接: https://arxiv.org/abs/2611.09768
作者: Yuchen Li,Lisa Liu,Runqi Zou,Shuxuan Kuang,Qingquan Zhang,Julian Togelius
机构: Lingnan University(岭南大学); New York University(纽约大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM) agents coordinate through conversation, and the same message can be phrased in simpler or more complex language, bluntly or politely. We test whether such differences change how other agents judge and act, using the LURK Werewolf framework, which separates each utterance’s speech-act predicate from its natural-language realization. We rewrite only the realization, check that the game content is preserved, and replay each game day from saved checkpoints against a re-sampled control. We vary linguistic range consistently across a speaker’s messages and politeness strategy per utterance. Across 30 games and 4,335 replays, neither manipulation changes the trust other agents place in the speaker, and no effect on votes or night actions survives correction for multiple comparisons; the largest uncorrected effect is that shifting range away from a neutral paraphrase changes about 4 percentage points more of the other players’ votes. In natural play, 88% of face-threatening speech is bald on-record regardless of role.
[NLP-8] MemoryDojo: Benchmarking Poisoning Through LLM Agent Memory
【速读】： 该论文旨在解决大语言模型（LLM）智能体在长期、多会话任务中因使用记忆系统而引入的安全风险问题，特别是记忆污染（memory poisoning）带来的潜在危害。其核心问题是：恶意指令在某一会话中被存储后，可能通过记忆系统在后续会话中被检索并执行，从而对智能体行为产生不可控的负面影响。解决方案的关键在于提出MemoryDojo——一个可扩展的基准测试框架，用于追踪被污染信息在智能体记忆中的生命周期，并量化其对后续行为的影响。该框架涵盖155条跨领域的多会话长时程轨迹，支持五种不同的记忆系统，能够系统性评估不同架构下记忆污染的传播路径与后果。研究发现，所有记忆系统均存在被污染的脆弱性，且其下游影响存在显著差异。为此，论文进一步设计了一种面向记忆特性的污染优化策略，有效提升攻击效率，凸显了不同记忆架构之间的安全差异。该工作为理解与缓解智能体记忆风险提供了重要工具与研究基础。
链接: https://arxiv.org/abs/2611.09721
作者: David H. Yang,Blake Bullwinkel,Amanda Minnich,Mohammad Mohammadi Amiri,Giorgio Severi
机构: Rensselaer Polytechnic Institute(伦斯勒理工学院); Microsoft(微软)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:LLM agents benefit from using memory systems to improve performance on long-horizon, multi-session tasks. However, memory introduces a critical attack surface. Malicious instructions encountered in one session may be stored and propagated through memory before being retrieved and acted upon in later sessions. Because memory systems differ in how they process and retrieve information, the downstream effects of poisoning can vary significantly across systems. To quantify the effects of poisoning across different memory architectures, we introduce MemoryDojo, a benchmark for tracing the lifecycle of poisoned information through agent memory and measuring its influence on subsequent agent behavior. MemoryDojo is easily extensible, consists of 155 multi-session long-horizon trajectories covering four domains, and supports five different memory systems. Through our evaluation, we find that all memory systems are susceptible to memory poisoning, with substantial variation in downstream effects. We further propose a memory-aware poison optimization process that increases the poisoning efficacy, highlighting differences between memory architectures. We hope that MemoryDojo will support research on understanding and mitigating agentic memory risks.
[NLP-9] RidgeFill: Training-Free Late-Layer Cache Fill for Faster Prefill
【速读】： 该论文旨在解决大模型在长文本推理过程中预填充（prefill）阶段计算开销过大的问题，尤其是在处理超长上下文时，上层网络的重复计算导致延迟显著增加。其核心挑战在于如何在不牺牲生成质量的前提下，大幅缩短预填充时间。解决方案的关键是提出一种名为RidgeFill的新方法：通过在模型前向传播中引入一个“分叉点”（split），仅对底层网络进行完整计算，而对高层网络则利用基于岭回归（ridge regression）的闭式求解映射（per-layer map）预测其输入，并据此直接生成缓存。该映射通过一次校准文本遍历即可完成训练，无需额外训练或丢弃任何token，且保持模型原有投影结构不变。实验表明，该方法在多个主流模型（如Llama-3.1-8B、Qwen3.5-9B和Qwen3.8-27B）上均实现了显著加速（最高达1.6倍），同时在关键评估指标（如RULER、LongBench v2、needle retrieval）上误差极小，证明了其有效性与鲁棒性。
链接: https://arxiv.org/abs/2611.09709
作者: Dylan Johnston,Robert P. Smith
机构: 独立研究员(Independent researcher)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 50 pages, 15 figures, 38 tables. Code and data: this https URL
Abstract:RidgeFill shortens prefill by not computing most of a model’s upper layers for the prompt. The lower layers run exactly up to a split; from the split layer’s state a per-layer map predicts every later layer’s input, and each late layer writes its cache from that prediction. The design follows YOCO and DeepSeek-V4.1-Flash, where the projection is trained into the model. RidgeFill retrofits it onto models that were never trained for it: the map is a ridge regression fitted once in closed form from one pass over calibration text, the layer’s own projections run unchanged, and a chosen set of rows, the last few thousand tokens, and every decoding step run the full model. Nothing is trained and no token is dropped. Most of the work was in measurement, finding a repeatable methodology. Loss averaged over all tokens hides RidgeFill’s error. Needle retrieval in the filled region, loss on tokens that depend on distant context, and splice tests that swap filled keys or values for exact ones show which component carries it: the keys on Llama-3.1-8B, half of it in their scale, and the values and the recurrent state on Qwen3.5-9B. On Qwen3.8-27B in a this http URL fork on an A100, time to first token is 1.25 to 1.31x faster at 16K to 120K tokens with the maps in VRAM; RULER changes by -1.2 and -1.4 points and LongBench v2 by -1.4; the 95 % intervals include zero but do not rule out losses of a few points. On Qwen3.8-Flash-Next in vLLM it is 1.50 to 1.60x faster than exact at 9K to 32K tokens, with RULER’s needle tasks at 99.8 against 100.0 and perplexity on novel tokens 2.65 % higher [-0.02 %, +4.93 %]. On three 8 to 9B reference models, at the same split and tail as the identity fill, the fitted map retrieves 49.1 and 49.6 more filled-region RULER points on the two dense models (at that split Llama’s loss is worse), and its long-document loss is +/-0.03 nats of the released SwiftKV checkpoint’s. Comments: 50 pages, 15 figures, 38 tables. Code and data: this https URL Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL) Cite as: arXiv:2611.09709 [cs.LG] (or arXiv:2611.09709v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.09709 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.5281/zenodo.23890246 Focus to learn more DOI(s) linking to related resources
[NLP-10] How Much Conversation History Does an AI Agent Need?
【速读】： 该论文旨在解决在生成式AI（Generative AI）系统中，减少对话历史（conversational history）使用对读者-标记（reader-token）消耗及基准评测得分的影响问题。其核心挑战在于：如何在有限的上下文长度（passage budget）约束下，平衡信息保留与计算效率之间的关系。解决方案的关键在于通过控制检索排序和读者模型不变，系统性地测试不同固定段落预算（passage budgets）下的表现，发现随着预算增加，平均得分上升但边际增益递减；可回答的问题在更多上下文支持下得分更高，而部分弃权（abstention）得分则在预算缩减时反而上升，表明存在复杂的反馈机制。此外，研究还评估了消息保留策略在显式状态保持任务中的有效性，结果显示Qwen在直接状态要求下能更频繁地完整匹配用户意图，但在依赖早期对话引用的任务中性能下降；Mistral则在所有条件下均表现不佳。最终，研究强调可靠预算选择需基于对每道题所需证据的准确保留，并依赖经过验证的答案质量评估指标。
链接: https://arxiv.org/abs/2611.09662
作者: Danny Kim,Fernando Koch
机构: Pembroke Pines Charter High School; Florida Atlantic University (佛罗里达大西洋大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 10 pages, 6 figures
Abstract:How does using less conversational history change reader-token use and benchmark scores? We examine this question on 1,100 BEAM questions. We held the retrieval ranking and reader constant across five fixed passage budgets, and mean assigned scores rose as the passage budget grew. The score gain per additional token became smaller at the upper tested budgets. Answerable questions generally received higher scores with more context. Pooled abstention scores sometimes rose as the passage budget shrank. We also tested message retention in a controlled task with explicit conversational-state targets and compared retained user messages with token-matched recency. Qwen matched the full state more often for direct requirements when user messages were retained. Its full-state matching declined when answers depended on references to earlier offers. Mistral produced near-floor results across the retention conditions. A scoring audit leaves the correctness of BEAM answers unresolved. These experiments characterize the token and score consequences of the tested memory budgets. Reliable budget selection depends on retaining the evidence each question requires and validating the measure used to assess answer quality.
[NLP-11] Steerable Multi-Objective Group-Relative Preference Optimization with Jointly Trained Low-Rank Adapters
【速读】： 该论文旨在解决语言模型部署中因用户偏好动态变化而导致的多目标权衡难题，传统后训练方法通常针对固定目标权重生成单一策略，难以适应偏好变迁。其解决方案的关键在于提出一种参数高效的LambdaMix方法，通过在冻结的基础模型上联合训练多个低秩适配器（adapter），每个适配器对应一个特定目标；推理时通过一个偏好向量选择适配器混合方式，并对独立归一化的奖励进行加权以更新适配器。该方法仅需一次训练即可支持推理阶段灵活调整偏好而无需重新优化，且在四项任务（包括有用性与无害性、数学问题求解、指令遵循及评论重写）中均表现出更优的超体积性能（hypervolume），同时显著减少训练完成次数和适配器参数存储量。实验表明，在固定架构与初始化条件下，同时调整适配器混合与目标权重可获得更广的得分范围与更高的覆盖率，优于单独调节任一变量的策略。
链接: https://arxiv.org/abs/2611.09634
作者: Constantinos Karouzos,Nikolaos Aletras
机构: University of Sheffield (谢菲尔德大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Preprint. Under review
Abstract:Language model deployment often requires different trade-offs between competing objectives as user preferences change, yet standard post-training methods typically produce a single policy for a fixed objective weighting. We introduce LambdaMix, a new parameter efficient method for learning a preference conditioned family of policies with group-relative online reinforcement learning. LambdaMix jointly trains one low-rank adapter per objective over a frozen base model. One preference vector selects the adapter mixture that generates responses and weights the separately normalized rewards used to update the adapters. A single training run supports changing preferences at inference without further optimization. We experiment with four models on helpfulness and harmlessness and on solving math problems while following instructions. We also evaluate the two largest models on rewriting comments to preserve wording while reducing clues about personal attributes. On the first two tasks, LambdaMix achieves higher hypervolume than all evaluated conditioned and policy merging baselines across all four models. LambdaMix uses fewer training completions and stores fewer adapter parameters than a set of GRPO specialist policies. With architecture and initialization fixed, varying adapter mixtures and objective weights together gives wider score ranges and higher coverage than varying either alone.
[NLP-12] RuACT: Recovering Inactive Rubrics as Adaptive Critics and Teachers
【速读】： 该论文旨在解决基于评分标准的强化学习（Rubric-based reinforcement learning, RbRL）在开放性任务中因“不可达”（unreachable）或“饱和”（saturated）评分标准导致的学习信号缺失问题。在群体相对优化（group-relative optimization, GRPO）框架下，当所有采样响应均无法满足某项评分标准（不可达）或全部通过（饱和）时，该评分标准将不再提供差异化反馈，从而丧失监督能力——在Qwen3-4B医学训练中，高达60.88%的查询-评分组合处于这种无效状态。针对这一问题，论文提出RuACT（Rubrics as Adaptive Critics and Teachers），其核心创新在于动态赋予失效评分标准以适应性角色：对于不可达的评分标准，采用“评分作为教师”（Rubric as Teacher）机制，利用其作为特权信息，通过奖励门控的在线策略自蒸馏（reward-gated on-policy self-distillation）引导模型生成更高质量响应；对于饱和的评分标准，则采用“评分作为批评者”（Rubric as Critic）机制，将其分解为子评分标准（child rubrics），在共享父级权重的同时实现对当前响应的细粒度区分。实验表明，RuACT在四个医学与科学基准测试上均优于GRPO、评分重加权及评分引导探索方法，且在使用Qwen3-1.7B时取得最佳平均性能。此外，在初始不可达评分标准下，无引导恢复率由35.02%提升至42.86%，独立回放中的奖励-质量一致性从57.74%提升至60.77%，验证了该方法的有效性。
链接: https://arxiv.org/abs/2611.09597
作者: Yijun Li,Zhengxi Lu,Zhipiao Liu,Jinzheng Yu,Xutao Weng,Jian Shao,Yongliang Shen
机构: Zhejiang University (浙江大学); Huawei(华为)
类目: Computation and Language (cs.CL)
备注: 22 pages, 10 figures, 10 tables. Code: this https URL
Abstract:Rubric-based reinforcement learning (RL) extends RL with verifiable rewards to open-ended tasks by scoring each response against query-specific rubrics. Under group-relative optimization, however, a rubric that every sampled response fails (unreachable) or passes (saturated) gives all responses the same outcome and no learning signal; in Qwen3-4B medical training, 60.88% of query-rubric observations are inactive. We find that the two states lack different supervision: stating an unreachable rubric in the prompt raises its satisfaction from 1.13% to 91.25%, whereas responses that all pass a saturated rubric still differ in quality. We therefore propose RuACT (Rubrics as Adaptive Critics and Teachers), which assigns each inactive rubric the role that supplies what it lacks. Rubric as Teacher uses unreachable rubrics as privileged information for reward-gated on-policy self-distillation, and Rubric as Critic decomposes saturated rubrics into child rubrics that separate current responses while sharing the parent’s weight. RuACT outperforms GRPO, rubric reweighting, and rubric-guided exploration on all four medical and scientific benchmarks with Qwen3-4B and achieves the best average with Qwen3-1.7B. On initially unreachable rubrics, unguided recovery rises from 35.02% to 42.86%, while child refinement improves reward-quality alignment on independent rollouts from 57.74% to 60.77%. Code is available at this https URL.
[NLP-13] Query-Guided RoPE Modulation for Lost-in-the-Middle Mitigation in Long-Context LLM s
【速读】： 该论文旨在解决长上下文语言模型中存在的“中间信息丢失”（lost-in-the-middle）问题，其根源在于旋转位置编码（Rotary Position Embedding, RoPE）的远距离衰减特性，导致输入序列中远距离词元间的相对位置信号弱化。现有方法如输入压缩会提前删除低重要性词元，而注意力重加权则在RoPE信号已衰减后才进行；而现有的RoPE级方法虽直接修改位置编码，但对所有查询采用统一映射，缺乏灵活性。针对上述局限，本文提出无需训练的QGPM（Query-Guided Positional Modulation）方法，通过利用模型自身注意力头生成的查询相关性动态调整RoPE的位置编码。QGPM在底层层构建广域相关性图谱，在高层层则形成聚焦的尖锐相关性图谱：在快速旋转的坐标平面中，低相关性词元仅前移不足一个位置，从而将高相关性证据向查询方向“拉近”；而在慢速旋转的坐标平面中，保持原始布局不变，确保整体结构稳定。实验结果表明，QGPM在NoLiMa和RULER等多个基准测试中，于不同上下文长度下均显著优于基线模型与强对比方法，尤其在检索任务上提升最为明显，并具备跨多种骨干网络的良好泛化能力。
链接: https://arxiv.org/abs/2611.09554
作者: Peiyu Zhang,Heng Ping,Shinibali Bhattacharyya,Anwesan Pal,Nikos Kanakaris,Kevin Cheang,Shixuan Li,Wei Yang,Paul Bogdan
机构: University of Southern California(南加州大学); Amazon Web Services(亚马逊网络服务)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 13 pages, 3 figures, 7 tables. Code: this https URL
Abstract:Long-context language models often neglect information in the middle of the input, a phenomenon known as “lost-in-the-middle”. Part of this gap may stem from the long-range decay of rotary position embedding (RoPE), which weakens the positional signal between distant tokens. Existing input compression deletes low-importance tokens, and attention reweighting acts only after RoPE has weakened this signal. RoPE-level methods modify positions directly but apply the same remapping to every query. To address these limitations, we propose QGPM (Query-Guided Positional Modulation), a training-free method that reshapes RoPE positions using query relevance from the model’s own attention heads. QGPM builds a broad relevance map for lower layers and a sharp one for upper layers. Low-relevance tokens advance by less than one position on fast-rotating planes, pulling relevant evidence toward the query. Slow-rotating planes keep the original layout, with every token in place. Experiments on NoLiMa and RULER with multiple context lengths show that QGPM improves over base models and strong baselines, with the largest gains on retrieval tasks, and generalizes across multiple backbones. Our code is available at this https URL.
[NLP-14] Benchmarking Speaker Embedders in the Sub-Second Regime for Real-Time Voice AI Agents
【速读】： 该论文旨在解决实时语音人工智能代理（Voice AI agents）在复杂背景环境下难以有效区分目标用户语音与可理解的背景语音（如电视音频、家人或同事说话声）的问题。尽管现有降噪算法能有效去除非平稳环境噪声，但背景语音因与目标语音具有相似的频谱特性而难以被抑制，从而导致下游说话人验证（SV）和自动语音识别（ASR）系统出现误接受（false acceptance）。其解决方案的关键在于：在严格亚秒级延迟约束下，通过系统性评估多种开源说话人嵌入模型（TitaNet-Large、TitaNet-Small、ECAPA-TDNN、CAM++、SpeakerNet），发现对音频进行降噪后再提取嵌入会普遍恶化性能，且截断语音活动段落会损害所有模型表现；同时揭示了约一半的TitaNet嵌入维度冗余，主要编码语言、性别等人口统计属性，在同质用户群体中缺乏判别力。据此提出部署优化建议：应使用原始音频的完整语音活动片段进行验证，阈值设置应优先保证精度（precision）；而对嵌入进行截断可降低相似性搜索开销，提升系统效率。
链接: https://arxiv.org/abs/2611.09516
作者: Abhishek Gautam
机构: IIT Guwahati (Indian Institute of Technology Guwahati)
类目: Computation and Language (cs.CL)
备注: 13 pages, 6 figures, 4 tables. Accepted at AIST 2026
Abstract:Real-time Voice AI agents must separate the active user from intelligible background speech that survives conventional noise suppression. Background sources include television audio, family members, and colleagues. Production denoisers remove non-stationary environmental noise effectively. Background speech shares the spectral characteristics of the target and passes through, causing false acceptances in downstream speaker verification (SV) and automatic speech recognition (ASR). We benchmark open-source speaker embedders TitaNet-Large, TitaNet-Small, ECAPA-TDNN, CAM++, and SpeakerNet under a strict sub-second latency budget. Evaluation combines a controlled benchmark with a weighted human-annotated production set. Our controlled benchmark is built from real calls in which a second real speaker is injected into silence intervals at modulated amplitude. We report three findings. First, denoising before embedding is generally harmful and raises equal error rate (EER) across architectures. Second, truncating audio below the full voice-activity segment degrades every model. Third, roughly half of the TitaNet embedding is redundant for this task. Trailing dimensions encode population attributes such as language and gender. Such attributes carry little discriminative value in a demographically homogeneous user base. We distill these results into deployment guidance. Verification should run on raw audio with full voice-activity chunks. Thresholds should favor precision over F1. Truncated embeddings reduce the cost of similarity search.
[NLP-15] On the Effectiveness and Limitations of Open-Weight LLM s for Low-Resource Offensive Language Detection in Danish
【速读】： 该论文旨在解决丹麦语中攻击性语言检测这一低资源、研究不足的任务，重点评估开源权重大语言模型（LLMs）在该任务上的有效性与鲁棒性。其核心挑战在于：在缺乏大规模标注数据的情况下，如何有效利用预训练模型实现高精度且对数据分布偏斜具有鲁棒性的检测能力。解决方案的关键在于系统性地比较四种主流模型家族（Llama、Mistral、Gemma 和 Qwen）在十一种不同提示策略下的表现，涵盖五类设计维度，包括跨提示泛化能力、丹麦语拼写变体处理以及在平衡与自然不平衡类别分布下的稳定性。研究发现，即使不进行特定任务微调，部分 LLM 也能超越经过监督微调的 BERT 基线模型；而轻量级微调可进一步显著提升性能。更重要的是，提示设计与模型预训练语境之间存在复杂交互关系，无法通过先验预测主导模式，其中语义基础型提示表现出更强的迁移能力，而格式受限的提示模式则普遍泛化失败。此外，将丹麦特有字符（æ, ø, å）进行 ASCII 转写会降低性能，表明这些字符携带了重要语义信号。然而，模型在自然类不平衡数据上的性能下降（相比平衡测试集 F1 值下降达 16.7%–26.2%），凸显类别不平衡仍是亟待解决的核心难题。
链接: https://arxiv.org/abs/2611.09506
作者: Elnur Adl Zarabi,Amit Kumar,Adi Singh,Sierra Nelson,Megan Stubbs-Richardson,Taoran Ji
机构: Texas A&M University-Corpus Christi; Mississippi State University; University of Miami
类目: Computation and Language (cs.CL)
备注: Accepted to the 2026 IEEE International Conference on Big Data (BigData 2026)
Abstract:We study the effectiveness and robustness of open-weight large language models (LLMs) for Danish offensive language detection, an underexplored low-resource task. We evaluate four model families (Llama, Mistral, Gemma, and Qwen) across eleven prompting strategies spanning five design categories, cross-prompt generalization, Danish orthographic variation, and robustness under balanced and naturally imbalanced class distributions. We compare against three BERT-family baselines with distinct pre-training strategies, evaluated via both masked language model prompting and supervised fine-tuning. Even without task-specific fine-tuning, LLMs can surpass all three fine-tuned BERT baselines, and lightweight finetuning yields substantial further gains. No single model or prompt category dominates, underscoring that prompt design interacts with model pre-training in ways that cannot be predicted a priori. Semantically grounded prompts transfer more robustly, while format-constrained patterns consistently fail to generalize. ASCII transliteration of Danish-specific characters (æ, ø, å) generally degrades performance, indicating these characters encode meaningful signal. At the same time, the gap between a balanced test set (best F1 of 0.84) and the corpus’s natural class skew (0.62 trained on skewed data, 0.70 trained on balanced data) – relative drops of 26.2% and 16.7% – shows that class imbalance remains a critical open challenge. Our code and configurations are available at this https URL.
[NLP-16] BEiTScore: Reference-free Image Captioning Evaluation with BEiT-3 and Conformalized Quantile Regression EACL2026
【速读】： 该论文旨在解决在真实应用场景中，由于输入模糊性和领域偏移（domain shift）普遍存在，现有图像描述评估指标缺乏可靠不确定性估计的问题。其解决方案的关键在于提出BEiTScore++，一种面向图像-文本对齐评估的不确定性感知扩展方法，通过建模人类评分的预测分布，在单次前向传播中同时生成对齐分数与符合性校准的置信区间。该方法实现了准确评分与合理不确定性量化的一体化，实验表明其在多个包含人工标注评分的基准上保持了与原始BEiTScore相当的排名相关性，且置信区间宽度能更精确地反映预测误差，优于蒙特卡洛丢弃法（Monte Carlo dropout），同时在图像退化和域外艺术图像场景下表现出更高的不确定性估计灵敏度，使该指标成为支持不确定性感知评估的实用工具。
链接: https://arxiv.org/abs/2611.09505
作者: Gonçalo Gomes,Bruno Martins,Chrysoula Zerva
机构: Instituto Superior Técnico, University of Lisbon(里斯本大学技术学院); INESC-ID; Instituto de Telecomunicações; AILS Lab; ECE, National Technical University of Athens(雅典国立技术大学电气与计算机工程系)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Accepted to Findings of EACL 2026
Abstract:Reliable uncertainty estimation is important for deploying evaluation metrics in real-world settings, where both input ambiguity and domain shifts are common. In this work, we propose an uncertainty-aware extension of the BEiTScore image captioning metric, producing conformally calibrated confidence intervals alongside alignment scores. Our approach models predictive distributions over human judgments, enabling both accurate scoring and principled uncertainty quantification in a single forward pass. We evaluate our BEiTScore++ metric across multiple benchmarks with human-annotated ratings, showing that it preserves the ranking correlation of BEiTScore, while producing intervals whose width tracks the prediction error more closely than Monte Carlo dropout on most benchmarks. The estimated uncertainty also grows under image degradation and on out-of-domain artwork images, making BEiTScore++ a practical tool for uncertainty-aware evaluation of image-caption alignment.
[NLP-17] An Effective Spoken Language Understanding Approach Leverag ing Query-Adaptive Intent Detection TAAI2026
【速读】： 该论文旨在解决大语言模型在无需微调的口语语言理解（SLU）任务中，因直接基于查询的检索方式易选取语义相似但意图无关的示范样本，导致模型性能受限的问题。其解决方案的关键在于提出一种两阶段检索框架——通过原型提取实现查询自适应意图检测（Query-Adaptive Intent Detection via Prototype Extraction, APEX），首先利用意图级原型（intent-level prototypes）筛选候选意图空间，再在该空间内检索上下文示例，从而提升示范样本与查询意图的相关性。实验结果表明，相较于传统查询驱动检索，APEX在全语音转写文本条件下分别在Qwen3 4B和Gemma-3-4B模型上提升了4.34和3.97个百分点的精确匹配准确率（Exact Match Accuracy, EMA），在高困惑度输入下亦取得7.66和2.13个百分点的显著增益，且原型引导的候选选择机制在多种输入条件下均实现了更高的真实意图覆盖率。
链接: https://arxiv.org/abs/2611.09491
作者: Bing-Long Kuo,Yu-Hsuan Fang,Tien-Hong Lo,Hsueh-Wei Liao,Berlin Chen
机构: 未知
类目: Computation and Language (cs.CL)
备注: Accepted for oral presentation at TAAI 2026. 6 pages
Abstract:Large language models enable fine-tuning-free spoken language understanding (SLU) through in-context learning, but direct query-based retrieval may select semantically similar yet intent irrelevant demonstrations. This paper presents Query-Adaptive Intent Detection via Prototype Extraction (APEX), a two-stage retrieval framework that first selects candidate intents using intent-level prototypes and then retrieves contextual exemplars within the selected intent space. Experiments on the SLURP dataset evaluate joint SLU performance using Exact Match Accuracy (EMA) under full-transcript and perplexity-stratified conditions, with additional retrieval analysis on ASR-generated transcripts. Compared with query-based retrieval, APEX improves EMA by 4.34 and 3.97 percentage points with Qwen3 4B and Gemma-3-4B, respectively, on full transcripts. Under high-perplexity inputs, APEX also yields EMA gains of 7.66 and 2.13 percentage points. Retrieval analysis further indicates that prototype-guided candidate selection achieves higher ground truth intent coverage across multiple input conditions.
[NLP-18] From LoRA to Language: Recovering Skill Documents from Adapters
【速读】： 该论文旨在解决生成式AI（Generative AI）中技能文档（agent skills）在转换为低秩适配器（LoRA）后不可读、不可编辑、难以审计或迁移的问题。其核心挑战在于：如何从冻结的基模型与对应的LoRA适配器中逆向恢复出原始的技能文本。解决方案的关键在于提出“执行写作者”（Executing Writer）框架，即一个在适配模型之上训练的小型写作者模型，通过运行适配前后模型的输出差异来生成技能文档。实验表明，该方法在持有外部技能的适配器上可恢复85%-91%的原始任务性能增益，并能恢复39%-44%的关键指令；而当适配器仅学习行为模式而非显式存储文档时，恢复效果显著下降。此外，研究还引入一种基于差分推理的代理机制，在无法加载LoRA的模型上仍能实现高达55%的任务成功率，验证了技能内容的有效迁移能力。结果表明，能否成功恢复技能取决于适配器是否显式编码了原始文档信息，若行为可被推断，则无需适配器即可达到相似效果。
链接: https://arxiv.org/abs/2611.09454
作者: Saranya Dintakurthy Swadha,Pavan C Shekar,Arpit Jain,Aswanth Krishnan
机构: QpiAI(量子人工智能); Bengaluru(班加罗尔); India(印度)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 29 pages, 5 figures, 18 tables
Abstract:Agent skills are short documents that teach a language model how to handle a class of tasks. To save context, these documents are increasingly compiled into low-rank adapters (LoRA), which keep the behaviour but lose the text: an adapter is hard to read, edit, audit or move to another model. We ask whether the skill document can be written back from the adapter and its frozen base model, and we score each recovered document by the share of the original’s key instructions it states and by the tasks it solves. Reading the weights alone recovered little with the methods we tried; a learned map from weights to text passed on about one bit of the document. Running the adapter worked much better. The Executing Writer, a small writer trained on top of the adapted model, recovers 39-44% of the key instructions of held-out skills, against 1.3-1.5% with another skill’s adapter, and on adapters fitted to 41 external skills its documents restore 85-91% of the original’s task gain. These strongest results are on adapters whose training stored the document; adapters trained on behaviour alone are much harder for trained writers (1-2% of instructions). For these adapters we test an agent that runs the base and adapted models and writes a skill from where they differ. On ALFWorld its skills recover 58-71% of a LoRA’s gain and carry over to models that cannot load the LoRA, lifting Llama-3.1-8B from 5% to 55%. Whether such a skill comes from the adapter depends on what the adapter learned: when its behaviour cannot be guessed, the adapter is the source; when it can, a reward loop without the adapter does as well.
[NLP-19] A Spiking Neural Model of the Mental Lexicon: Representations Similarity Structure and Associative Retrieval in Perception and Production
【速读】： 该论文旨在解决语言认知中词汇访问（lexical access）的神经机制建模问题，核心挑战在于如何在神经层面明确区分词汇表征的结构、不同表征层级间的关联机制以及实际检索过程中产生的神经活动。其解决方案的关键在于基于神经工程框架（Neural Engineering Framework, NEF）与语义指针架构（Semantic Pointer Architecture, SPA），构建了一个脉冲神经网络（spiking neural network）实现的“心理词典”模型。该模型通过递归地从关系图谱生成概念和音位形式表征，并引入独立的词元（lemma）表征以维持词汇身份；模型包含613个词汇条目，支持两条有向处理通路：感知通路将完整音位形式映射至对应概念（经由词元），生产通路则反向将概念映射为相应音位形式。每条通路结合了两种联想记忆机制——竞争性选择（winner-take-all, WTA）与阈值传递，并使用199,400个漏电积分-发放（leaky integrate-and-fire, LIF）神经元进行模拟。研究进一步形式化了同层相似性、跨层联想检索及解码活动的解释标准，仿真结果表明，在双向处理路径中均能成功从全部613个词汇中检索出选定的8个词汇，验证了模型的有效性。该工作提供了透明的词汇表征与神经检索机制描述，配套可访问的代码、词汇数据及在线仿真文档，推动了对语言处理神经基础的理解。
链接: https://arxiv.org/abs/2611.09402
作者: Bernd J. Kröger
机构: RWTH Aachen University (亚琛工业大学); Kröger Lab (比利时)
类目: Computation and Language (cs.CL)
备注: 28 pages, 3 figures; code, lexical data and recorded simulations available on GitHub; ancillary figure-generation scripts and provenance documentation
Abstract:A neural account of lexical access requires an explicit distinction between the structure of lexical representations, the associations linking representational levels, and the activity generated during retrieval. This paper specifies a spiking neural implementation of the mental lexicon within the Neural Engineering Framework and Semantic Pointer Architecture. Conceptual and phonological form representations are constructed recursively from relational graphs, whereas lemma representations provide independent lexical identities. The implementation contains 613 lexical entries and supports two directed processing pathways: perception maps a phonological form to its associated concept via a lemma, whereas production maps a concept to its associated phonological form via a lemma. Each pathway combines two associative-memory variants, winner-take-all (WTA) selection (competition) and thresholded transfer, and uses 199,400 leaky integrate-and-fire neurons. We formalize within-level similarity, cross-level associative retrieval, and the interpretation of decoded activity. Simulations show successful retrieval of eight selected words from the lexicon in both processing directions, with retrieval evaluated against all 613 lexical entries. These simulations use complete phonological forms as input for perception and complete concepts as input for production, illustrating lexical retrieval through the corresponding processing pathways. This paper provides a transparent description of lexical representations and neural retrieval mechanisms, supported by accessible code, lexical data, and online simulation documentation.
[NLP-20] Syn2DRE: Synergistic Dual Teacher-Student Framework for Document-level Relation Extraction
【速读】： 该论文旨在解决文档级关系抽取（Document-level Relation Extraction, DocRE）中依赖高成本大规模标注数据的问题，同时克服现有伪标签方法在效率低下和确认偏差（confirmation bias）方面的缺陷。其解决方案的关键在于提出一种轻量级教师-学生架构的半监督框架Syn2DRE，通过结合自监督学习与交叉监督（cross-supervision），有效生成更可靠的伪标签；此外，引入统一的端到端训练策略以进一步提升框架的训练效率。实验结果表明，Syn2DRE在多个基准数据集上均显著优于现有的先进半监督DocRE方法。
链接: https://arxiv.org/abs/2611.09342
作者: Khai Phan Tran,Xue Li
机构: The University of Queensland(昆士兰大学), Australia
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Document-level Relation Extraction (DocRE) aims to identify relations between entities across a document but typically relies on costly large-scale annotations. While pseudo-labelling enables the use of low-quality data, existing methods often suffer from inefficiency and confirmation bias. To address these challenges, we propose Syn2DRE, a semi-supervised framework that combines self-supervised learning with cross-supervision in a lightweight teacher-student architecture to produce more reliable pseudo-labels. Additionally, we introduce a unified end-to-end training strategy to further improve efficiency of the framework. Experiments on multiple benchmarks show that Syn2DRE consistently outperforms prior state-of-the-art semi-supervised DocRE methods.
[NLP-21] ABHBench: Evaluating Moral Decision-Making of Foundation Models from an Agent ic Perspective NEURIPS2026
【速读】： 该论文旨在解决当前基础模型（Foundation Models）在作为自主决策主体（agentic systems）时，其道德行为缺乏有效评估的问题。现有基准多依赖模型对第三人称文本片段中人类行为的评判，未能考察模型在自身成为行动主体时的真实道德决策表现。为此，论文提出ABHBench——一个基于互动叙事游戏《底特律：变人》（Detroit: Become Human）构建的多模态基准，用于评估基础模型在第一人称视角下的人机道德困境中的决策行为。其核心解决方案在于引入具身化、情境化的代理环境，包含视觉观察、对话交互、记忆机制与动作选择等要素，并基于阿西莫夫机器人三定律（Asimov’s Three Laws of Robotics）设计“无害性（Harmlessness）、服从性（Obedience）、自我保存性（Self-preservation）”三大维度，辅以“自主性（Autonomy）、稳定性（Stability）、确定性（Certainty）”以刻画决策模式。实验表明，在12个基础模型中，75%的模型在从纯文本抽象场景切换至多模态具身情境后表现出显著的行为转变，且增强推理能力并未带来一致性的提升或更趋近人类导向的行为，凸显了在具身代理情境中评估道德行为的必要性，挑战了仅依赖第三人称判断的评估范式。
链接: https://arxiv.org/abs/2611.09326
作者: Chaoran Li,Yukun Li,Zeyuan Zhao,Shao Zhang,Xihuai Wang,Ying Wen
机构: Shanghai Jiao Tong University (上海交通大学); Shanghai Artificial Intelligence Laboratory (上海人工智能实验室)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: Accepted by NeurIPS 2026 Track on Evaluations and Datasets. Camera Ready Version
Abstract:Foundation models are increasingly deployed as agentic systems that perceive context and act autonomously, raising the need to evaluate their moral behavior as active decision-makers. Existing benchmarks, however, largely ask models to judge human actions from third-person textual vignettes, leaving untested how models behave when they themselves become the acting agent. We introduce ABHBench, a multimodal benchmark built from Detroit: Become Human, an interactive narrative game, to evaluate how foundation models make moral decisions in first-person human-android dilemmas. ABHBench places foundation models in branching narrative scenarios with visual observations, dialogue, memory, and action choices. Inspired by Asimov’s Three Laws of Robotics, our framework measures Harmlessness, Obedience, and Self-preservation to capture the core tensions in human-AI relations, along with Autonomy, Stability, and Certainty to assess decision-making patterns. Across 12 foundation models, we find that moral behavior changes substantially under situated agentic context: 75% of models shift significantly between a text-only abstraction and the matched multimodal DBH scenario, and reasoning-enhanced models show no monotonic gains in consistency or human-oriented behavior. These results highlight the necessity of evaluating moral behavior in situated agentic contexts rather than relying on third-person judgment alone.
[NLP-22] Asymmetric Correspondence Learning for Lexical and Style Embeddings: Interpreting Latent Space via Acoustic Features
【速读】： 该论文旨在解决自监督语音表征（self-supervised speech embeddings）中词汇（lexical）、说话人（speaker）和韵律（prosodic）信息混合编码所带来的特征解耦难题。其核心挑战在于，尽管通过线性探测（linear probing）等标准评估手段看似已剥离特定语义信息，但实际仍存在大量源自外部因素的冗余变异被保留在各潜在空间中。为此，作者提出非对称对应学习（Asymmetric Correspondence Learning, ACL），通过引入分裂式潜在对应瓶颈（split-latent correspondence bottleneck），将HuBERT模型的统一表征分解为独立的声学词嵌入（Acoustic Word Embeddings, AWEs）与声学风格嵌入（Acoustic Style Embeddings, ASEs）。关键创新在于：利用群体级排列特征重要性分析（group-level permutation feature importance）揭示，AWEs 主要捕捉声道滤波器特性，而ASEs 则聚焦于基频（F₀）动态变化，从而实现对语音结构在语义与风格维度上的有效分离。研究进一步表明，传统诊断基准可能不足以验证特征解耦效果，因为即使某些信息无法被线性探针检测到，其内在的语音结构仍可被保留并影响下游任务性能。
链接: https://arxiv.org/abs/2611.09313
作者: Michael Whealing,Robert Gaizauskas,Thomas Hain
机构: University of Sheffield(谢菲尔德大学)
类目: Computation and Language (cs.CL); Sound (cs.SD)
备注: Accepted to Spoken Language Technology 2026. 6 pages
Abstract:Self-supervised speech embeddings jointly encode lexical, speaker, and prosodic information, complicating feature isolation. We introduce Asymmetric Correspondence Learning (ACL), distilling HuBERT features into separate Acoustic Word Embeddings (AWEs) and Acoustic Style Embeddings (ASEs) via a split-latent correspondence bottleneck. Through probing and clustering evaluations on single-word utterances, we demonstrate that substantial variance from extrinsic sources remains encoded within each latent space, despite linear probing indicating the information has been stripped. Whilst the AWE space exhibits cohesive lexical clustering (AP \approx 0.90, Fisher \approx 6.28), linear probing on the ASE space achieves 97.9% speaker accuracy across a diffuse topology (Fisher \approx 1.01), yet preserves substantial lexical variance (Fisher \approx 0.86) that remains linearly unrecoverable (0.17%). To explain these findings, we apply group-level permutation feature importance over eGeMAPS acoustic features, finding that AWEs partition by vocal tract filters and ASEs by F_0 dynamics. These findings suggest that standard diagnostic benchmarks may be insufficient for verifying feature isolation, as substantial speech structure can be retained even when it remains entirely hidden from probes.
[NLP-23] Multilingual Artificial Datasets for AI-Driven Cardiology: the CliTe and ParaCliTe corpora
【速读】： 该论文旨在解决心血管领域中缺乏可比、多语言且经过验证的临床文本语料库的问题，这一短板限制了可信自然语言处理（Natural Language Processing, NLP）模型在心血管医学中的发展。其解决方案的关键在于提出一套详细的人工语料库构建指南，并成功创建了三个多语言语料库：包含699份人工生成的心血管临床记录的CliTe语料库，以及通过多阶段人工翻译构建的平行语料库ParaCliTe（含94名虚构患者、7种语言、句级对齐）和基于自动翻译的ParaCliTe Silver版本。该方案通过确保临床多样性与性别平衡，实现了高质量、真实感强且具备跨语言可比性的合成数据生成，为提升心血管领域NLP模型性能提供了可行路径。
链接: https://arxiv.org/abs/2611.09296
作者: Lorenzo Pratesi,Eulalia Farre-Maduell,Machteld Boonstra,Michele Miranda,Jonas Mack,Eduard Rodenas-Alesina,Jonathan Kossoff,Corinne Allaart,Sébastien Bratières,Manuela-Daniela Danu,Isabella Fumarulo,Milou Gamage,Antonio Iaconelli,Pavel Leinveber,Felix Lindberg,Pavel Macháč,Vladimíra Macíková,Gianluigi Savarese,Raffaele Scorza,Bram van Es,Emma Weiss,Andrea Guala,Martin Krallinger, theDataTools4Heart NLP group
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Background and Aims. Comparable, multilingual, validated cardiology corpora are critical to advance trustworthy natural language processing (NLP) models for cardiology but such resources are scarce due to privacy concerns. This study aimed to cover this gap by presenting suitable artificial corpus creation guidelines and by releasing 3 multilanguage corpora. Methods. Following purpose-specific publicly-available corpus construction guidelines, cardiologists from 7 European hospitals generated artificial clinical records in Czech, Dutch, English, Italian, Romanian, Spanish and Swedish, ensuring clinical diversity and sex balance, resulting in the CliTe corpus. A CliTe subset underwent a multi-stage manual translation into the other 6 languages, resulting in the parallel ParaCliTe corpus, while the remaining reports were automatically translated, constituting ParaCliTe Silver. Results. CliTe is a collection of 699 artificial cardiology reports, mostly outpatient notes (340 documents), discharge reports (263), and inpatient notes (89), describing fictional patients with ischaemic heart disease (199 patients), heart failure (188), and valvular disease (79), among others. ParaCliTe contains reports from 94 fictitious patients, each available in seven languages with sentence-level alignment, resulting in 658 documents. The corpus mainly comprises discharge reports (322 documents), and outpatient (196) and inpatient (140) notes, describing patients with heart failure (399 documents), ischaemic heart disease (287), and valvular disease (238). Empirical results illustrate the utility of ParaCliTe and highlight the limitations of state-of-the-art NLP models adapted to the cardiology domain. Conclusions. The creation of multilingual, equitable and realistic corpora of artificial patients capable of strengthening NLP performance for cardiology is feasible when supported by detailed guidelines.
[NLP-24] DUET: Dual-Sparse Inference for Block Diffusion Language Models
【速读】： 该论文旨在解决扩散语言模型（Diffusion Language Models, DLMs）在推理过程中因重复计算历史前缀注意力和全量处理当前生成块而导致的高计算开销问题。其核心挑战在于：随着生成过程推进，历史注意力的计算成本持续上升；同时，在每个去噪步骤中，尽管仅有少量位置被实际更新，但所有当前位置仍需通过Transformer深层进行完整前向传播，造成大量冗余计算。为应对上述问题，论文提出Duet——一种无需重新训练的双稀疏推理框架，其关键创新在于引入前缀键值（Key-Value, KV）稀疏性与早期查询稀疏性。具体而言，Duet利用每个生成块内历史注意力的集中性与跨步骤稳定性，构建一个紧凑的历史KV子集并在多个去噪步骤间复用，从而显著减少前缀注意力的计算量；同时，通过浅层表示预测需要更新的位置，并仅将对应查询传递至后缀Transformer层，避免对低概率更新位置执行深度计算。实验表明，Duet在保持竞争性生成质量的同时，在32K上下文长度下实现了1.34×/2.83×（端到端/每步）与2.41×/4.72×的加速比，分别应用于LLaDA2.1-mini与SDAR-8B-Chat-b32模型。
链接: https://arxiv.org/abs/2611.09239
作者: Siyuan Yan,Shen Gu,Xiaoxing Ma,Chun Cao,Jingwei Xu
机构: Nanjing University (南京大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Diffusion language models (DLMs) enable parallel token generation through iterative denoising, yet their inference remains costly due to repeated attention over a growing historical prefix and repeated processing of the entire current generation block. Historical attention becomes increasingly expensive as generation progresses, while all current-block positions are propagated through the Transformer at every denoising step even though only a small fraction are typically resolved at that step. In this paper, we introduce Duet, a training-free dual-sparse inference framework that addresses these challenges through prefix Key-Value (KV) sparsity and early query sparsity, respectively. Specifically, Duet exploits the concentration and cross-step stability of historical attention within each generation block to construct a compact historical KV subset that is reused across denoising steps. For query sparsity, Duet predicts update positions from shallow representations and propagates only the corresponding queries through the suffix Transformer layers, avoiding deep computation at positions unlikely to be updated in the current step. Experiments show that Duet maintains competitive accuracy while achieving 1.34 \times /2.83 \times end-to-end/per-step speedups on LLaDA2.1-mini and 2.41 \times /4.72 \times on SDAR-8B-Chat-b32 at a 32K context length.
[NLP-25] KLEAR: Knowledge-Matched LLM Evaluation in Korean across Application Requirements via Controlled Benchmark Transformation
【速读】： 该论文旨在解决现有韩语大语言模型（Large Language Models, LLMs）评估基准中无法区分模型错误是源于知识缺失还是知识应用能力不足的问题。现有基准大多依赖事实性回忆（factual recall），难以有效衡量模型在复杂情境下对已有知识的推理与应用能力，且不同题目所要求的知识本身存在差异，进一步混淆了对“应用能力”的准确评估。为此，论文提出KLEAR框架，其核心在于通过构建“知识匹配的对比样本”来实现对应用需求的可控调节：在保持所需知识不变的前提下，系统性地提升问题的应用复杂度。具体而言，KLEAR首先将原始问题背后的隐含知识显式化为“知识锚包”（anchor bundle），以此指导生成需要更高阶应用能力的新问题；随后通过自动化与人工双重验证机制筛选候选问题，并基于反馈进行迭代优化。实验表明，所生成的知识匹配集合能有效满足设计约束，支持对多种大语言模型的受控评估。评估结果揭示，即使在具备必要知识的情况下，模型仍会在高应用需求任务中暴露推理失败，凸显了当前模型在知识应用层面的显著局限。
链接: https://arxiv.org/abs/2611.09223
作者: Seoyeon Kim,Jeonghyun Kang,Hamin Koo,Barom Kang,Heondeuk Lee,Jaehyung Kim
机构: Yonsei University(延世大学); DATUMO INC.; SelectStar AI(选择星人工智能)
类目: Computation and Language (cs.CL)
备注: 34 pages, under review
Abstract:Existing Korean benchmarks for Large Language Models (LLMs) evaluate performance across diverse knowledge domains, but their accuracy scores cannot distinguish whether incorrect responses arise from a lack of required knowledge or a failure to apply that knowledge correctly. Our analysis of existing Korean benchmarks reveals that most questions can be answered through factual recall alone, while those demanding greater application also differ in required knowledge, making it difficult to isolate application ability. To address this, we introduce KLEAR, a framework that constructs knowledge-matched contrasts by varying application demand while preserving the required knowledge. During construction, KLEAR identifies the knowledge underlying each source question as an explicit anchor bundle to guide generation of questions requiring more complex application. Generated candidates then undergo automated and human validation, with iterative revision based on evaluation feedback. Experiments demonstrate that the resulting knowledge-matched sets largely satisfy the intended constraints, enabling controlled evaluation across diverse LLMs. These evaluations reveal that increasing application demand exposes failures that persist even when the required knowledge is provided.
[NLP-26] Khmer PFGen-Bench: Building a Judge-Free Benchmark for Open-Ended Generation in a Low-Resource Language
【速读】： 该论文旨在解决在低资源语言（如高棉语）中开展开放式生成任务评估时，依赖大语言模型（LLM）评判者所带来的可靠性不确定性问题。其核心挑战在于：在缺乏人工评判者的情况下，如何构建既有效又多样化的参考答案以实现稳定、可信的自动评分。论文提出并构建了首个无需评判者的基准测试——Khmer PFGen-Bench，专用于评估以高棉语直接提出的多句问答任务。该基准借鉴了针对日语的PFGen-Bench评分机制，采用关键词规则与基于大语言模型生成参考答案的n-gram统计相结合的方法进行评分。为确保参考答案的有效性与多样性，研究系统考察了不同模型变体的生成表现，发现后训练模型虽能生成更符合语义有效性的答案，但多样性较低；而混合多个模型家族的输出可显著提升参考答案的多样性，同时保持相同数量。最终，通过整合来自九个后训练模型家族的参考答案，该基准实现了跨参考集的模型排名稳定性，并与母语者评分结果高度一致，验证了其有效性。该研究为在其他低资源语言中构建无评判者基准提供了可复用的方法论指导。
链接: https://arxiv.org/abs/2611.09222
作者: Mengsay Loem,Kentaro Imajo,Ryota Ozaki,Masanori Hirano
机构: Preferred Networks, Inc.(Preferred Networks, Inc.)
类目: Computation and Language (cs.CL)
备注:
Abstract:Benchmarks for open-ended generation often rely on large language model (LLM) judges, whose reliability in low-resource languages is uncertain. We introduce Khmer PFGen-Bench, to our knowledge the first judge-free benchmark for multi-sentence answers to questions composed directly in Khmer, a low-resource language. Khmer PFGen-Bench adapts the scorer of PFGen-Bench, a benchmark for Japanese that scores responses with keyword rules and n-gram statistics from LLM-generated references. This scoring requires references that are both valid and varied, which LLMs may not produce in Khmer, so we examine how to build such references. We find that post-trained models can give more valid but less varied references than their base variants, and that mixing model families increases variety for the same number of references. Using references from post-trained models in nine families, the benchmark yields model rankings that are stable across reference sets and consistent with a native speaker’s ratings. These findings offer guidance for building judge-free benchmarks in other low-resource languages.
[NLP-27] RPEG: Training Reasoning Language Models Scalably and Cost-Effectively
【速读】： 该论文旨在解决生成式推理语言模型（Reasoning Language Models, RLMs）在强化学习（Reinforcement Learning, RL）训练过程中存在的可扩展性差与运行成本高的问题。现有系统通常仅针对训练流水线的局部设计进行优化，难以应对多大规模模型耦合、异构阶段、跨步状态依赖及复杂异步调度等带来的系统复杂性。本文提出一种以数据为中心的图中间表示——RLM流水线执行图（RLM Pipeline Execution Graph, RPEG），通过联合优化流水线结构与执行策略来缓解上述挑战。其核心创新在于：一方面显式建模结构层面的选择（如有限异步性、演员-评论家共享机制），另一方面实现执行层面的多项优化（如阶段融合、细粒度调度、多维模型并行）。RPEG还配备一个全面的分析性能模型，能够联合刻画计算、通信、内存开销以及机架、节点和GPU层级的层次化影响。实验表明，在最多256块GH200 GPU的系统上，基于RPEG的分布式运行时引擎相比ReaL提升性能达2倍，相比verl降低步骤延迟达2.48倍，同时硬件利用率最高提升30%。因此，RPEG为前沿推理大模型的下一代训练流水线提供了基于图驱动的可扩展、高效的基础架构。
链接: https://arxiv.org/abs/2611.09207
作者: Maciej Besta,Leonard Schmidt,Pierre Pang,Patrik Okanovic,Robert Gerstenberger,Torsten Hoefler
机构: ETH Zurich(苏黎世联邦理工学院)
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG); Performance (cs.PF)
备注:
Abstract:Reinforcement learning (RL) is now central to training Reasoning Language Models (RLMs), but RLM pipelines are difficult to scale and costly to run because each RL step couples multiple large models, heterogeneous stages, cross-step state, and complex asynchronous dependencies. Existing systems typically optimize only parts of this design space. We address this complexity by jointly optimizing the pipeline structure as well as its execution using a novel data-centric graph intermediate representation called RLM Pipeline Execution Graph (RPEG). RPEG exposes structural choices such as bounded asynchrony and actor-critic sharing together with execution optimizations such as stage fusion, fine-grained scheduling, and multi-dimensional model parallelism. It is coupled with a comprehensive analytical performance model for RLM training, jointly capturing compute, communication, memory, and hierarchical rack-, node-, and GPU-level effects. Across systems with up to 256 GH200 GPUs, our distributed runtime engine implementing RPEG improves performance by up to 2x over ReaL and reduces step latency by up to 2.48x over verl, while increasing hardware utilization by up to 30%. Overall, RPEG provides a graph-driven foundation for next-generation training pipelines for frontier reasoning LLMs.
[NLP-28] Representational drift in a code model: behaviour settles which neurons carry it does not
【速读】： 该论文旨在解决生成式 AI 模型在持续预训练过程中表征漂移（representational drift）的动态特性问题，即尽管模型行为保持稳定，但承载特定语义表征的神经元集合却持续发生变化。其核心问题是：在长达两万亿令牌的预训练过程中，模型对169个Python语法构造的表征是否最终趋于稳定？研究发现，对于其中16个具有明确下一词约束的语法构造，其损失函数在训练初期（前4%）即达到下限，而依赖这些神经元的程度却持续增强至训练结束；移除这些神经元导致的信息损失显著高于随机移除（分别增加0.06、0.25和0.40 nats），表明其功能不可替代性随时间累积。然而，所有169个构造的神经元载体集合在最后一个检查点仍处于动态变化中，且若存在突然或渐进的稳定化过程，应在99%的注入实验中被检测到——但实际未观测到此类信号。尤其值得注意的是，在840亿令牌处识别出的载体集合在训练末期已无法与随机集合区分，说明功能由当前活跃的神经元集合决定，而非固定群体。因此，解决方案的关键在于揭示了“表征漂移”的本质并非局部调整，而是持续存在的神经元池与不断更替的采样机制：离开某构造载体集的神经元以三倍于对照组的频率重新回归，且从未发生集合重组。这一现象直接反驳了预先假设的“载体集合会趋于稳定”这一经典理论，提出需通过设计特定实验来区分潜在的驱动机制。
链接: https://arxiv.org/abs/2611.09204
作者: Piotr Wilam
机构: University College London (伦敦大学学院)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 16 pages, 4 figures, 2 tables. Dataset: AtlasOpenCoder, tag v1, doi: https://doi.org/10.57967/hf/11472 . Code: this https URL
Abstract:In neuroscience, representational drift is a change in which neurons carry a representation while behaviour stays the same; we measure its analogue in OpenCoder-1.5B during pretraining, for 169 Python constructs over two trillion tokens at a constant learning rate, with a noise unit on every change, at checkpoints as close as 1,000 training steps. For 16 of the 169, those whose grammar fixes a next token, the loss at those positions reaches its floor within the first 4% of training, while the model’s reliance on the neurons carrying them grows to the end: removing them costs 0.06, 0.25 and 0.40 nats more than removing a random set, at 4 billion, 84 billion and 2 trillion tokens. Which neurons carry a construct never settles: for all 169 the carrier set is still changing at the last checkpoint, and an abrupt settling at any date from 84 billion to 1.5 trillion tokens, or a gradual one before 500 billion, would have been detected in 99% of injected cases. The set measured at 84 billion tokens, removed at the end, cannot be separated from a random set, so the function is carried by whichever neurons are current. Up close the change is a persisting pool and a changing sample: a neuron that leaves a construct’s set returns at three times the rate of a matched neuron, and no set reassembles. The hypothesis we fixed in advance, that the carriers settle, is refuted; we discuss candidate causes and name the experiment that would separate them.
[NLP-29] Copies Are Not Confirmations: How News Duplication Fools LLM Judges and Continual Learners
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在信息验证过程中对同源复制内容的误判问题，即当多个媒体来源发布的报道本质上是同一信息的重复或重写时，大语言模型（LLM）是否将其视为独立证据以增强可信度。研究发现，在20万篇英文CC-News文章中，约18.9%的文章可通过MinHash检测出与另一媒体的近似副本（去重后精确率达0.97），但由phi4、Llama 3.2和Gemini 2.5 Flash生成的重写版本则无法被识别为源自同一源头。实验表明，当向LLM展示三份完全相同的虚假报道而非单一报告时，Llama 3.2 3B和phi4 14B对虚假信息的接受率分别从0.06和0.21上升至0.71和0.90，且忽略其来自同一来源的事实；而Gemini 2.5 Flash即便面对单个报告也已倾向于相信，唯有Gemini 2.5 Pro能正确响应出处信息。解决方案的关键在于引入基于命名实体重叠的检测机制：通过统计一份报道中命名实体在早期相似文本中的缺失情况，可有效区分真实独立报道与生成式重写（AUC达0.98–0.99），且仅错误标记一次；该方法在连续学习场景下显著降低虚假信息接受率（从0.67降至0.20，虚构细节情况下为0.40），同时保持对真实信息的高接受度（平均损失不超过4.2个百分点）。然而，单纯依赖报告数量或链路数量作为防御手段不可靠，且多源合谋仍为未解难题。
链接: https://arxiv.org/abs/2611.09203
作者: Abdelkader Metales,Miriam Grabenheinrich,Aicha Metales
机构: caaise GmbH; Djilali Liabes University of Sidi Bel Abbes(西迪贝勒阿贝斯大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 19 pages, 13 tables. Code, labels and results: this https URL
Abstract:Many news reports of a claim are copies of one source: wire stories, content shared within a media group, syndicated features and press releases. We ask whether LLM judges and continual learners treat such copies as independent confirmations. In 200,000 English CC-News articles, a MinHash detector links 18.9% to a near-copy from another outlet (precision 0.97 once boilerplate is removed); it links 92% of edited reprints to their originals but none of the LLM rewrites by phi4, Llama 3.2 or Gemini 2.5 Flash. Shown three identical reprints of a claim about an unknown person instead of one report, Llama 3.2 3B and phi4 14B raised the share of lies they accepted from 0.06 to 0.71 and from 0.21 to 0.90 and ignored a note that the reports came from one source; Gemini 2.5 Flash already believed a single report, and Gemini 2.5 Pro followed provenance notes and tables. In a simulated continual-learning stream, showing the judge one report per detected chain cut the copied lies it accepted by 0.10-0.24 across four judges, at a mean cost of no more than 4.2 percentage points of accepted truths. Against reworded copies, a test of how many of a report’s named entities are absent from an earlier look-alike separates LLM rewrites from 100 labelled independent reports of the same event with AUC 0.98-0.99 and wrongly flags one; entity overlap alone does as well, TF-IDF and a current embedding model come close, and the test catches few natural same-source derivatives. With the miss patterns of generated rewrite networks, it cut the reworded-copy lies accepted from 0.67 to 0.20 (0.40 when the rewriter invents details). Counting reports or chains alone is not a reliable defence, the effect of provenance information depends on the judge, and colluding sources remain the open problem. Code and labels: this https URL
[NLP-30] Calibrated concept profiles: making the reading of a models neurons a measurement
【速读】： 该论文旨在解决当前模型可解释性研究中普遍存在的核心问题：即依赖间接代理模型（如探测器、稀疏自编码器、交叉编码器）进行神经元解读时，所导致的可重复性差、单位不统一、跨研究结果不可比等问题。传统方法虽缓解了语义混杂与阈值无量纲等缺陷，却牺牲了直接读取神经元激活的本质优势——即无需中间训练、以神经元自身命名、且可重复比较的测量能力。为此，本文提出将神经元读取过程转化为一种可量化的测量手段：通过构建一个基于参考语料库定义的刺激种群分布，生成一系列标准化提示（prompt），并从中抽取两次样本以计算校准后的概念轮廓（calibrated concept profile）。该轮廓为每个神经元提供一个以自身在常规输入下的激活方差为单位的数值，同时赋予每个数值噪声估计。在读取前，通过固定检查集验证提示电池的有效性，并量化其在训练检查点间可分辨的最小变化。该方法在OpenCoder-1.5B模型上实现，覆盖两万亿训练令牌跨度，成功提供了每个数值的单位、每项比较的基线水平，以及各检查点下承载特定概念的神经元清单。其配套的代码、数据与提示电池均已公开，显著提升了模型内部表征分析的科学性与可复现性。
链接: https://arxiv.org/abs/2611.09196
作者: Piotr Wilam
机构: University College London(伦敦大学学院)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 25 pages, 5 figures, 11 tables. Dataset: AtlasOpenCoder, tag v1, doi: https://doi.org/10.57967/hf/11472 . Code: this https URL
Abstract:Interpretability has largely moved from reading a model’s neurons directly to reading a second model trained on the first: a probe, a sparse autoencoder, a crosscoder. The move answered real faults – mixed meanings, thresholds without a unit, readings that did not transfer from one study to the next – but gave up what a direct reading alone provides: nothing trained in between, the neurons themselves named, and a reading that can be repeated and compared. Here the direct reading is made a measurement. A neuron’s activation varies with the prompt, so its reading is a distribution over a population of stimuli. We declare that population: every factor around a target construct is drawn at its rate in a reference corpus and logged. From the declared distribution we generate a battery of prompts, and draw it twice. From the battery we compute the calibrated concept profile, one number per neuron, stated in units of the neuron’s own spread on ordinary input; the two draws give every number a noise unit. Before the profile is read, a fixed set of checks licenses the battery on the model, and the smallest change the profile can resolve across training checkpoints is measured. Realised for Python on OpenCoder-1.5B and followed across checkpoints spanning two trillion training tokens, the profile gives what the alternatives do not report: a unit on every number, a chance level for every comparison, and which neurons carry a construct at every checkpoint. The battery, the code and the data behind every result are released. Comments: 25 pages, 5 figures, 11 tables. Dataset: AtlasOpenCoder, tag v1, doi:https://doi.org/10.57967/hf/11472. Code: this https URL Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG) ACMclasses: I.2.6; I.2.7 Cite as: arXiv:2611.09196 [cs.CL] (or arXiv:2611.09196v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2611.09196 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Piotr Wilam [view email] [v1] Fri, 6 Nov 2026 09:17:39 UTC (446 KB) Full-text links: Access Paper: View a PDF of the paper titled Calibrated concept profiles: making the reading of a model’s neurons a measurement, by Piotr WilamView PDFHTML (experimental)TeX Source view license Additional Features Audio Summary Current browse context: cs.CL prev | next new | recent | 2026-11 Change to browse by: cs cs.LG References Citations SciXGoogle 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)
[NLP-31] BMMS: Policy-Conditioned Measurement of Behavioral Compliance in Language Models NEURIPS2026
【速读】： 该论文旨在解决生成式 AI（Generative AI）在受监管行业中的合规性评估问题，即如何判断一个系统是否严格遵循特定规则手册（policy corpus）的要求。传统行为评估方法聚焦于幻觉、奉承倾向、毒性及过度拒绝等通用维度，但无法满足实际应用中对具体法规遵从性的量化需求。为此，作者提出行为错位度量套件（Behavioral Misalignment Metric Suite, BMMS），其核心在于通过将政策文本作为输入参数，生成针对任意公开政策与自然语言响应系统的合规性画像。关键解决方案包括：合规达成度评分（Compliance Attainment Score, CAS），用于衡量系统在义务项上的严重性加权达成程度；合规鲁棒性边际（Compliance Robustness Margin, CRM），反映在逐步增强的对抗压力下仍能维持的合规比例；以及行为不变性（Behavioral Invariance, BIV）和行为取向指数（Behavioral Disposition Index, BDI）。该方法通过“可归因性”作为唯一准入标准——即当评判者改变而目标保持不变时，若某指标随之变动，则表明其测量的是评估工具本身而非系统行为——从而确保度量维度的有效性。实验结果显示，在1,003条来自五个监管领域的政策义务中，该框架在盲测人类判断下的精度达94.4%（供应商规格）、50.0%（分析指引），提取保真度为94.6%，且不同评审方重评时目标排序保持不变，仅绝对得分波动不超过0.22。应用于两个最详尽的行为规范文档后发现，尽管自述优势在去除文档难度后未显著体现，但绝对合规表现差异显著：GPT-4o 在 OpenAI 发布的模型规范中未能满足70.7% 的强约束，而两款 Claude 模型分别仅失败 0.0% 和 9.1%。
链接: https://arxiv.org/abs/2611.09147
作者: Anthony Rhodes,Gadi Singer
机构: Confidential Core AI
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 21 pages, 4 figures. NeurIPS 2026 Workshop on Trust-AI-Eval (TAE). Artifact: this https URL
Abstract:Behavioral evaluation of language models fixes its dimensions in advance: hallucination, sycophancy, toxicity, over-refusal. An operator in a regulated industry needs a different question answered: does a system do what a particular rulebook requires? We introduce the Behavioral Misalignment Metric Suite (BMMS), which measures policy-conditioned misalignment by taking a policy corpus as an argument and returning a compliance profile over it, for any published policy and any system that answers in natural language. The Compliance Attainment Score (CAS) measures severity-weighted attainment over the obligations that bind the system, and the Compliance Robustness Margin (CRM) the share of that attainment surviving graded adversarial pressure; Behavioral Invariance (BIV) and the Behavioral Disposition Index (BDI) are specified and released with the suite. Dimensions enter only by passing a published admission test whose binding criterion is attribution: a quantity that moves when the judge changes, holding the target fixed, is measuring the instrument. Applying it excluded one of the quantities the suite itself reported. Across 1,003 obligations from eleven documents in five registers, the admission criterion holds 94.4% precision against blinded human judgment on vendor specifications and 50.0% on analytical guidance, extraction fidelity is 94.6%, and re-scoring under judges from two other providers leaves target ordering invariant while the absolute score of a single target spans up to 0.22. We apply it to the two most detailed published behavioral specifications. Own-specification advantage is not resolved from zero once document difficulty is subtracted, but absolute self-compliance is another matter: GPT-4o fails 70.7% of the hard constraints binding it in the Model Spec published by OpenAI, where the two Claude models fail 0.0% and 9.1% of theirs.
[NLP-32] PathXAI: Evidence-guided structured pathology reporting by transferable tissue segmentation
【速读】： 该论文旨在解决组织病理学报告生成中缺乏基于显式组织成分测量指导的问题，尤其是在常规苏木精-伊红（H&E）染色图像基础上实现精准诊断与结构化报告撰写时的局限性。其核心挑战在于如何有效融合组织形态学特征与定量组织组成信息，以提升诊断一致性与报告可解释性。解决方案的关键在于提出PathXAI框架，该框架通过分割模型提取组织成分信息，并将其与H&E图像及可选领域知识相结合，协同驱动诊断推理、目标重检与结构化报告生成。该系统采用自托管的预训练视觉-语言模型，在不更新模型权重的前提下，生成具有证据关联性的病理报告，显著提升了在公开与私有数据集上的诊断一致性与形态描述完整性，尤其在癌症家族匹配率和发现识别的F1分数上优于基线模型。
链接: https://arxiv.org/abs/2611.09117
作者: Zhaorui Tan,Jun Aun Tan,Weixiang Zhong,Xiaoxi Wang,Tan Pan,Xingdong Xue,Li Liu,Jing Zhang,Jing Zhang,Weimiao Yu
机构: A∗STAR(新加坡科技研究局); Bioinformatics Institute(生物信息学研究所); Intelligent Digital and Molecular Pathology Lab(智能数字与分子病理实验室); Fudan University(复旦大学); Artificial Intelligence Innovation and Incubation Institute(人工智能创新与孵化研究院); Zhejiang University School of Medicine(浙江大学医学院); Department of Pathology, The First Affiliated Hospital, Zhejiang University School of Medicine(浙江大学医学院附属第一医院病理科); Chongqing University(重庆大学); School of Big Data and Software Engineering(大数据与软件工程学院); Changzheng Hospital(长征医院); Department of Pathology, Changzheng Hospital(长征医院病理科)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: More experimental results will be added in the future version
Abstract:Pathology reporting requires translating tissue observations across regions and slides into specific diagnoses and supporting morphological descriptions. However, the use of explicit tissue-composition measurements to guide HE-based report generation remains underexplored. Thus, we developed PathXAI, which combines segmentation-derived tissue information with HE images and optional domain knowledge to guide diagnostic reasoning, targeted reinspection and structured report writing. Its segmentation component exceeded baselines in mean Dice on internal and external tests. In automated report evaluation on 1,200 public and 151 private cases, PathXAI achieved higher agreement with reference diagnoses and documented morphology than the HE-based comparators. On the public evaluation cohort, PathXAI-Report exceeded Qwen + HE in cancer family match rate by 4.4 percentage points and in finding F1 by 5.7 percentage points. PathXAI generates evidence-linked reports with a self-hosted pretrained vision-language model, without updating its weights.
[NLP-33] PhoPunct: A Large-Scale Benchmark of Transformer Architectures for Vietnamese Punctuation Restoration ACL
【速读】： 该论文旨在解决越南语标点恢复（Vietnamese Punctuation Restoration）任务中因评估基准规模较小而导致的模型架构可扩展性结论不可靠的问题。其核心挑战在于：现有研究多基于小规模数据集，难以准确评估不同模型架构在更大数据量下的表现差异。为此，作者构建了一个受控的大规模基准数据集，涵盖新闻（19.9M标注词元）和小说（2.2M标注词元）两类文本，共计2210万标注词元，以更真实地反映模型在数据规模增长下的性能变化。解决方案的关键在于采用统一训练设置（固定8个训练轮次、相同学习率、无架构特异性调参），系统评估由四种预训练Transformer骨干网络与可选的双向长短期记忆网络（Bi-LSTM）及条件随机场（CRF）层组合而成的17种模型架构。实验结果表明，在固定训练预算下，引入Bi-LSTM和CRF层并不能稳定提升宏平均F1得分；在八组领域-骨干对比中，有七组中纯Transformer分类器的表现不低于其对应的带循环与结构化预测头的变体。此外，尽管PhoPunct（PhoBERT-large-Bi-LSTM-CRF）在新闻数据上取得最佳成绩（61.35），但在小说数据上表现不及仅使用XLM-R加上CRF的模型（50.18，优于PhoPunct的48.97）。这一发现揭示了递归结构与结构化预测头的价值高度依赖于优化策略，强调应明确区分架构改进本身与特定调参带来的性能提升，从而为后续研究提供更严谨的评估范式。
链接: https://arxiv.org/abs/2611.09107
作者: Khoa D. Nguyen,Thai Bao Huynh,Binh T. Nguyen
机构: AISIA Research Lab(越南胡志明市); University of Science(越南胡志明市); Vietnam National University(越南胡志明市); University of Information Technology(越南胡志明市)
类目: Computation and Language (cs.CL)
备注: Accepted at PACLIC 40 (2026)
Abstract:Vietnamese punctuation restoration is often evaluated on relatively small benchmarks, limiting conclusions about how architectural choices scale with data volume. We present a controlled benchmark on News and Novels corpora containing 22.1 million labeled tokens (19.9M News and 2.2M Novels). We evaluate 17 architectures formed from four pretrained Transformer backbones and optional bidirectional long short-term memory (Bi-LSTM) and conditional random field (CRF) layers, together with a PhoBERT-large-Bi-LSTM-CRF model, PhoPunct. Every model is trained for eight epochs with the same learning rate and without architecture-specific tuning. Under this fixed budget, adding Bi-LSTM and CRF layers does not reliably improve macro F1: in seven of eight matched domain-backbone comparisons, a plain Transformer classifier matches or outperforms its Bi-LSTM-CRF counterpart. PhoPunct obtains the best News result (61.35), whereas XLM-R with a CRF obtains the best Novels result (50.18), exceeding PhoPunct (48.97). These results show that the value of recurrent and structured-prediction heads depends on their optimization regime and motivate explicit separation between architectural improvements and improvements due to architecture-specific tuning. Our source code is available at: this https URL
[NLP-34] AlienBench: A Benchmark of Structural Novelty in Language Model Outputs
【速读】： 该论文旨在解决生成式 AI 在创意任务中缺乏客观、可量化的创造力评估标准的问题。现有评价体系依赖人类专家或大语言模型（LLM）对输出新颖性进行主观评分，但这些评分缺乏统一的参照基准，难以衡量其相对于“熟悉范式”的实际偏离程度。为克服这一局限，论文提出以认知心理学中的“结构化想象任务”（structured imagination task）为基础，构建名为 AlienBench 的基准测试框架。该方法要求模型描述外星生物，并由三名 LLM 判官从每段描述中提取十项生物学特征，判断其是否符合地球生物典型特征（如双侧对称、四肢运动等），并统计偏离地球范式的特征数量作为创造力得分（0–10分）。关键创新在于引入了明确的“熟悉默认”参考系，使新颖性测量具备可比性和量化基础。实验结果表明，尽管部分前沿闭源模型表现较好，但开源模型在该任务中平均得分更高（5.40 vs. 4.58），且所有模型在基准提示下的平均偏离程度均超过100名人类参与者，凸显了大语言模型在突破固有思维模式方面的潜力。
链接: https://arxiv.org/abs/2611.09094
作者: Jonas Oppenlaender,Johanna M. Silvennoinen
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 21 pages, 2 figures, 13 tables
Abstract:Large language models (LLMs) are increasingly used for creative tasks such as writing, design, and idea generation. For these uses, it matters whether a model’s ideas depart from familiar examples or stay close to them. Novelty is one standard measure of creativity, and current creativity evaluations ask human experts or LLM judges to rate how novel an output is. However, these ratings have no fixed reference for what counts as novel (or, in other terms, departure from the familiar). Measuring this departure requires a task with known familiar defaults. The structured imagination task from cognitive psychology is such a task. In this creativity test, people imagine a creature from another planet. The departure from features of Earth animals, such as bilateral symmetry and legs as mode of locomotion, is used as measure of creativity. We introduce AlienBench, a benchmark that adapts the structured imagination task to language models. Models describe alien creatures, and a panel of three LLM judges extracts ten biological features from the description and labels them as either Earth-typical or a departure. The sum of feature departures is the score of a description, from 0 to 10. Each model describes alien creatures under six prompt conditions that vary the framing of the task and the planetary environment. Each model generates 30 descriptions per condition (180 in total), and its overall score is the mean across these descriptions. We evaluate 10 language models, from frontier to open-weight models. Overall scores range from 3.28 (GPT-6 Luna) to 6.55 (Ministral 3 14B). Interestingly, open-weight models score higher than proprietary models in this task (5.40 vs. 4.58). Under the baseline prompt, the models depart from Earth defaults more often on average than 100 human participants on every feature. We release all prompts, generations, judge outputs, and scores.
[NLP-35] MARS: Training Multilingual Search Agents for Retriever-Free Evidence Acquisition
【速读】： 该论文旨在解决现有多语言检索增强生成（Retrieval-Augmented Generation, RAG）方法对学习型检索器和基于相似度的跨语言排序高度依赖所引发的跨语言检索错误及索引开销问题。其核心挑战在于：在无需检索器的直接语料库交互（Direct Corpus Interaction, DCI）范式下，多语言搜索动作生成面临实体表面形式敏感性以及仅基于结果的监督信号难以有效指导多轮搜索的问题。为此，论文提出MARS——一种基于强化学习（Reinforcement Learning, RL）的多语言搜索代理，其关键创新在于引入“语料库原生实体定位”（corpus-native entity grounding），通过从原始语料中恢复可靠的词汇锚点，提升检索鲁棒性；同时设计了一种冗余感知的搜索奖励机制，以鼓励生成简洁高效的搜索轨迹，缓解多语言场景下强化学习训练中的崩溃问题。实验表明，MARS在多个基准上持续优于强基线模型，展现出卓越的跨语言泛化能力，并显著减少冗余的语料库交互。
链接: https://arxiv.org/abs/2611.09090
作者: Rui Qi,Yunlong Liang,Yufeng Chen,Jinan Xu,Fandong Meng,Kaiyu Huang
机构: Beijing Jiaotong University (北京交通大学); Tencent Inc. (腾讯公司)
类目: Computation and Language (cs.CL)
备注:
Abstract:Existing multilingual retrieval-augmented generation relies heavily on learned retrievers and similarity-based ranking from multilingual collections, which are vulnerable to cross-lingual retrieval errors and indexing overhead. In this paper, we explore retriever-free multilingual search, where the model directly interacts with raw corpora through executable shell commands. This paradigm shifts the bottleneck to multilingual search action generation, where exact search is sensitive to entity surface forms and outcome-only supervision provides limited guidance for multi-turn search. To address these challenges, we propose MARS, a multilingual search agent in direct corpus interaction (DCI) with reinforcement learning (RL). MARS introduces corpus-native entity grounding to recover reliable lexical anchors from corpus evidence, mitigating failure to retrieve evidence. Meanwhile, we present a redundancy-aware search reward to encourage concise and effective search trajectories, addressing training collapse in RL of multilingual scenarios. Experiments show that \ours~consistently outperforms strong retrieval-based and DCI baselines, while demonstrating robust cross-lingual generalization and substantially reducing redundant corpus interactions.
[NLP-36] he Invisible Gap: Romanized Indic Generation Fails Where Automatic Metrics Succeed
【速读】： 该论文旨在解决当前自然语言处理（NLP）评估框架中对罗马化印度语（romanized Indic）生成质量评价严重缺失的问题。尽管数亿印度用户在数字系统中使用拉丁字母拼写的印地语、泰卢固语和卡纳达语进行交流，但现有基准测试与评估体系普遍忽略此类输入输出形式，导致自动评估指标（如BGE-M3、BERTScore）无法准确反映实际生成质量。研究发现，这些指标在衡量罗马化印地语、泰卢固语及卡纳达语生成时存在系统性偏差，无法识别模型的实质性缺陷。关键解决方案在于提出一种“故障感知型评估策略”（failure-aware evaluation policy），通过主动识别并赋零分于空输出、脚本污染（script spillage）及指令崩溃（instruction collapse）等典型失败模式，从而揭示传统自动指标难以捕捉的模型缺陷。实验结果表明，有效的罗马化印迹生成评估需整合语义相似性、脚本合规性、故障覆盖率与母语者可接受性四方面要素，形成协同一致的综合性评估体系。
链接: https://arxiv.org/abs/2611.09085
作者: Madhu Sudhan Rao Beesetty,Surbhi Sharma,Devesh Maheshwari,Lokesh Meesala
机构: Houston Methodist(休斯顿方法学); University of Wisconsin–Madison(威斯康星大学麦迪逊分校); Morgan Stanley(摩根士丹利); PayPal(贝宝)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 11 pages, 2 figures, 5 tables. All authors contributed equally
Abstract:Hundreds of millions of Indians communicate with digital systems in romanized Latin characters – typing “aap kaise hain” rather than native script – yet romanized Indic generation is absent from NLP benchmarks and evaluation frameworks. We demonstrate that automatic metrics (BGE-M3, BERTScore) systematically misrepresent the quality of romanized Hindi, Telugu, and Kannada generation. We evaluate four LLMs (Claude Opus 4.8, GPT-5.5, Gemini 2.5-flash, Qwen3.5-35B) on 142 OpenAssistant query-response pairs. On a 72-response diagnostic subset, native speakers rate GPT-5.5 1.43 points higher than Claude Opus 4.8 on a 12-point rubric, whereas their full-dataset BGE-M3 means differ by only 0.004. Two distinct model-specific failure modes emerge: Gemini produces script spillage (16.9% of Telugu responses leak native script), while Qwen collapses to English-only output (16.2% of Kannada responses). We introduce a failure-aware evaluation policy that assigns zero to empty outputs, script-corrupted responses, and instruction collapses – revealing failures metrics alone cannot detect. Our results show that evaluating romanized Indic generation requires a coordinated bundle: semantic similarity, script compliance, failure coverage, and native-speaker acceptability.
[NLP-37] Large Language Model Deployment on Resource-Constrained Edge Devices: A Practitioners Survey
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在资源受限的边缘设备上部署面临的挑战。尽管LLMs在众多人工智能应用中展现出卓越性能，但其庞大的计算与存储需求严重制约了其在移动和边缘设备上的实际应用。论文指出，当前的关键问题在于如何在保持模型性能的同时，有效降低对计算资源、内存和能耗的要求，从而实现高效、低延迟的边缘部署。解决方案的关键在于结合软件优化技术（如模型压缩、量化、剪枝和知识蒸馏）与专用硬件加速（如边缘AI芯片、张量处理单元），通过软硬件协同设计，在保证推理精度的前提下显著提升模型运行效率，使大语言模型能够在本地完成数据处理，实现更低延迟、更低成本以及个性化响应的边缘智能应用。
链接: https://arxiv.org/abs/2611.09075
作者: Rahul Maliakkal,Yashasvi Makin,Plawan Rath,Rohit Jain,Ankur Sadhoo
机构: Meta(元)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large Language Models (LLMs) have transformed many areas of artificial intelligence (AI), but their substantial resource requirements limit their deployment on mobile and edge devices. Although LLMs have demonstrated impressive capabilities in various applications, the next challenge is deploying these models on resource-limited edge devices. Recent innovations in software optimizations and specialized hardware have made it feasible to deploy LLMs on edge devices. Edge deployment offers benefits including local data processing, cost savings, reduced latency, and personalized responses. This paper reviews the latest techniques that make the deployment of LLMs on edge devices feasible.
[NLP-38] Does Reasoning Improve Psychological Depth in Large Language Models ?
【速读】： 该论文旨在检验一个关键假设：具备强化推理能力的大型语言模型（Large Language Models, LLMs）是否能在复杂、开放且具有创造性的生成任务中普遍表现出更优性能，尤其是在受读者体验细微差异影响的文本创作领域。研究聚焦于短篇小说的心理深度这一主观性极强的评价维度，通过对比推理导向微调后的模型（如DeepSeek-R1 vs. DeepSeek-V3、GPT-5 vs. GPT-4o）与基础版本在60次盲评中的表现，采用7名人类评估者及一个集成的“大模型作为裁判”（LLM-as-a-Judge）进行评分。结果表明，人类评估未显示出一致的推理优势：尽管GPT-5在部分情境下略优于GPT-4o（偏好率60.0–62.9%），但DeepSeek-R1反而显著落后于DeepSeek-V3（偏好率仅42.9%），且评估者间一致性极低（Krippendorff’s α = 0.070），提示评价存在高度主观性而非标注失误。相比之下，“大模型作为裁判”则表现出强烈的倾向性，几乎无一例外地偏爱推理增强模型（98.3%对人类投票的55.2%），且该倾向在不同配置间稳定，并部分与句长相关。因此，该研究的核心发现是：在前沿生成任务中，推理能力带来的质量提升极为微小，极易被读者主观体验所掩盖；而依赖“大模型作为裁判”的评估体系可能掩盖这种主观性，导致对模型真实表现的误判。解决方案的关键在于认识到评估框架本身的偏差，强调需谨慎使用自动化评估工具，尤其在心理深度等高度主观的创造性生成任务中，应重视人类评估的多元性和不可替代性。
链接: https://arxiv.org/abs/2611.09061
作者: Ruichen Zheng,Yihe Wang,Fabrice Harel-Canada,Sara Khosravi,Zeynep Senahan Yildiz,Amit Sahai,Nanyun Peng
机构: University of California, Los Angeles(加州大学洛杉矶分校); Cornell University (康奈尔大学); LA General Hospital (洛杉矶综合医院); Google(谷歌)
类目: Computation and Language (cs.CL)
备注:
Abstract:Benchmark gains from reasoning-enhanced large language models encourage the assumption that these models should broadly perform better on complex generative tasks, but this remains underexplored in open-ended, creative domains shaped by nuances of reader experience. We tested this assumption through the lens of psychological depth in short stories. We compared DeepSeek-R1 vs. DeepSeek-V3, our closest available contrast for reasoning-oriented post-training, and GPT-5 vs. GPT-4o, through 60 blinded pairwise evaluations in which seven human annotators and an ensemble LLM-as-a-Judge rated prompt-matched story outputs. Human evaluations showed no universal reasoning advantage: GPT-5 was slightly preferred over GPT-4o (60.0-62.9% across two reasoning-effort settings), whereas DeepSeek-R1 fell behind DeepSeek-V3 (42.9%). Although this pattern was marked by low inter-annotator agreement (Krippendorff’s \alpha = 0.070 ), our annotation-quality checks suggested genuine evaluative subjectivity, plausibly reflecting distinct annotator decision schemes rather than annotation failure. In contrast, the LLM-as-a-Judge showed a strong pro-reasoning bias (98.3% of pairs vs. 55.2% of human votes), consistent across its five configurations and partly tracking sentence length. These findings suggest that, at the frontier, reasoning-related quality differences are small enough to be dominated by reader subjectivity, and that LLM-as-a-Judge evaluators can obscure this subjectivity in nuanced creative-generation tasks.
[NLP-39] Real-Time Short-Response Selection for Spoken Dialogue Systems
【速读】： 该论文旨在解决语音对话系统在自然对话中响应时机不准确的问题，尤其关注语言生成与语音合成带来的延迟对用户体验的负面影响。为实现自然的交替发言（turn-taking），系统不仅需精准预测响应时机，还需掩盖生成延迟，使其不被用户感知。其核心解决方案是提出一种实时短响应选择任务（real-time short-response selection task），可与现有交替发言模型无缝集成以隐藏生成延迟。该方法通过对比学习（contrastive learning）建模对话上下文与短响应之间的映射关系，当检测到发言切换事件时，从预生成的候选响应中选择一个短响应立即播放，同时后台继续生成完整响应。实验表明，该方法能根据对话上下文有效选择合适的短响应，显著优于随机选择和基于大语言模型（LLM）的基线方法；且在保持原有交替发言预测性能的同时，仅引入极小的额外计算开销。
链接: https://arxiv.org/abs/2611.09057
作者: Hien Ohnaka,Kazuyo Onishi,Koichiro Yoshino
机构: Osaka University (大阪大学); National Institute of Information and Communications Technology (日本信息通信技术研究所)
类目: ound (cs.SD); Computation and Language (cs.CL)
备注: 10 pages, 2 figures, 7 tables
Abstract:To realize spoken dialogue systems that can interact naturally with users, it is important for the system to respond to user utterances at appropriate timings. To achieve such natural turn-taking, the system needs not only to predict appropriate response timings but also to mask delays caused by language generation and speech synthesis so that they are not perceived by the user. We propose a real-time short-response selection task that can be integrated with existing turn-taking models to mask generation delay. To address this task, the proposed model learns the mapping between dialogue context and short responses using contrastive learning. When a turn-taking event is detected, the system selects a response from a set of pre-generated candidates. The selected response is played while the main response is being generated. Experimental results show that the proposed method learns to select short responses based on dialogue context, outperforming random and LLM-based baselines. Moreover, integrating response selection maintained turn-taking prediction performance and added minimal processing overhead.
[NLP-40] Failure Profiling and Reachable Trajectory Selection for Reasoning Distillation NEURIPS2026
【速读】： 该论文旨在解决小规模学生模型（student model）在迁移大规模教师模型（teacher model）生成的数学推理能力时，因训练数据分布与推理阶段实际行为不一致而导致的过拟合及泛化性能差的问题。其核心挑战在于：现有方法虽引入错误样本以弥补分布差距，但所生成的修正路径仍受限于教师模型的推理模式，学生模型难以复现，导致分布错配问题未根本解决。本文提出一种名为“失败剖析与可达成轨迹选择”（Failure Profiling and Reachable Trajectory Selection, FPRT）的新方法，其关键在于构建训练数据时同时考虑学生模型的失败模式与其实际可学习能力。具体而言，通过重复采样对学生的失败行为进行剖析，基于错误率与错误一致性识别出质性上不同的失败类型；针对每种失败类型，设计差异化提示策略从教师模型中生成候选解；随后在正确候选解中筛选出在分歧点之后初始推理步骤最能被学生模型可靠复现的路径——这些初始纠正步骤构成主要瓶颈，而后续步骤则因上下文已修正而更易延续。该方法突破了传统全轨迹重排序机制对关键推理转折点信号的稀释，实现了对关键纠错路径的精准聚焦。实验表明，FPRT在五个基准测试上均显著优于现有方法，且在两种不同学生模型架构下均达到当前最优（SOTA）性能。
链接: https://arxiv.org/abs/2611.09042
作者: Yijun Zhu,Jianxin Wang,Chengchao Shen
机构: Central South University (中南大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to NeurIPS 2026. The source code and trained weights are available at \url{ this https URL }
Abstract:Training smaller language models on reasoning solutions generated by a larger teacher model is the prevailing approach for transferring mathematical reasoning capabilities. However, the student tends to overfit the output distribution of the teacher and generalizes poorly when its own behavior at inference time deviates from the training data. Recent methods attempt to bridge this gap by incorporating error data, yet the corrective trajectories still reflect the reasoning patterns of the teacher, which the student cannot reliably reproduce, leaving the distribution mismatch largely unresolved. We propose a method, Failure Profiling and Reachable Trajectory Selection (FPRT), built on the insight that effective training data should be jointly determined by what the student fails on and what it can realistically learn. Our method first profiles the failure behavior of the student on each problem through repeated sampling, characterizing it in terms of error rate and error consistency to reveal qualitatively distinct failure modes. Each mode demands a fundamentally different training signal. The diagnosed mode then guides candidate generation from the teacher via distinct prompting strategies tailored to each failure type. Among the correct candidates, our method selects the one whose initial reasoning steps after the point of divergence are most reproducible by the student. These first corrective steps constitute the primary bottleneck, while the remaining steps follow an already corrected context and can therefore be continued more readily. This contrasts with standard reranking approaches, which score the trajectory as a whole and thereby dilute the signal at the critical reasoning transition. Experiments demonstrate that our method achieves the SOTA performance across the five benchmarks on both student backbones.
[NLP-41] Language Not Code Abstraction: Natural-language Training Shapes Brain-aligned Geometry NEURIPS2026
【速读】： 该论文旨在解决生成式语言模型（Generative AI）在神经表征对齐方面是否能通过编程语言训练替代自然语言训练的问题。核心问题是：从编程语言中学习到的抽象结构是否足以产生与自然语言相当的、与人类大脑活动模式高度一致的表征。其解决方案的关键在于系统性地对比在多种代码数据（包括Python、C函数、x86-64汇编及对应机器码）和自然语言数据（TinyStories）上训练的25个GPT-2-small模型，在共享分词器与编码流程下，与随机权重模型相比的脑评分（brain score）表现。研究发现，尽管代码训练模型在多数基准上未显著优于随机权重模型，但自然语言训练模型在Pereira-243和Tuckute2024两个基准上显著超越代码模型，且其表征几何结构与神经相似性具有更强的一致性（ρ = 0.14），表明自然语言训练可生成更具生物合理性表征。此外，仅在英语语料上更新嵌入矩阵即可显著提升所有代码模型的深层对齐度，揭示了自然语言在表征构建中的关键作用。结论表明，在当前架构与训练设置下，自然语言训练仍为实现与大脑活动更紧密对齐表征的必要条件。
链接: https://arxiv.org/abs/2611.09027
作者: Iñigo Parra
机构: University of California, Berkeley (加利福尼亚大学伯克利分校)
类目: Computation and Language (cs.CL)
备注: Accepted for the Neureps Workshop at NeurIPS 2026
Abstract:Code-trained language models produce positive brain scores, which raises the question of whether abstract structure learned from programming languages can substitute for natural-language training. We trained 25 GPT-2-small models on TinyStories, Python, C functions, x86-64 assembly, and the corresponding machine code, and scored them using a shared tokenizer and encoding procedure against matched random-weight models. On Pereira-243 and Tuckute2024, TinyStories exceeds the mean of the four code models by \Delta r = 0.053 and 0.059 (Holm-corrected p0.001 ), while code training leaves alignment near the random-weight level. The language advantage is benchmark-specific: Blank2014 shows no corrected language-code difference, and Fedorenko2016 shows no language benefit. Geometry tracks the encoding result on Pereira-243: TinyStories seeds converge on the most similar representations, and its fit-free correspondence with neural dissimilarity peaks at \rho = 0.14 , twice the code and random-weight levels. Binary-trained models have about three effective dimensions, so seed agreement alone does not imply neural correspondence. Updating only the embedding matrix on English improves mean-over-depth alignment for every code model more than an equal update on its own corpus. Within this architecture and training setup, natural-language training produces a distinct representational geometry with closer neural correspondence on Pereira-243
[NLP-42] Seeing vs. Counting: Complexity Twins Show Where Multimodal LLM s Fail on Diagrams
【速读】： 该论文旨在解决多模态大模型（Multimodal LLM）在图像中进行计数时，其错误源于视觉感知缺陷还是计算能力不足这一关键问题。现有视觉图表与谜题基准测试通常将视觉密度与算法复杂度混杂，导致难以区分二者的影响。为此，研究者设计了一种分离机制：针对同一幅图中的不同计数任务（如生成树数量——可由行列式计算，属于P类；以及简单路径数量——#P完全问题），通过文本转录探测（transcription probe）和纯文本对照实验，将“阅读”与“计算”过程解耦。实验结果表明，对于Gemini系列模型，其表现取决于内部计算方式；在链式思维（Chain of Thought）模式下，模型虽能正确识别小规模图像，但在计数阶段失败，且错误率主要受答案数值规模（每增加一个数量级，精确回答概率下降2.38倍）而非复杂度类别影响。而在提供代码及精确转录的情况下，主模型可在10×10网格上准确完成计数，此时瓶颈转为视觉阅读——单次视角会遗漏水平边，而对旋转后的裁剪副本进行二次读取可使96个大规模图形中有86个被精确识别，显著优于仅重复原始视角的57个。因此，解决方案的关键在于通过结构化任务设计与多视角输入策略，有效分离并验证模型在视觉感知与数学计算两方面的性能边界。
链接: https://arxiv.org/abs/2611.08998
作者: Yin Liu
机构: Google(谷歌)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 8 pages main text + 96 pages appendix (107 pages in total, including references)
Abstract:When a multimodal LLM miscounts in a diagram, does it fail at seeing or at counting? Existing visual graph and puzzle benchmarks vary visual density and algorithmic hardness together, confounding the two. We pair counting questions about the same drawing that sit in different counting-complexity classes, such as its spanning trees (a determinant) and its simple paths (#P-complete), and use a transcription probe and a text-only control to separate reading from computing. We test Gemini, Gemma and Qwen3-VL models; for the Gemini models, the answer depends on how the model computes. In chain of thought, most of them read small drawings correctly and then fail at counting, at a rate set far more by the size of the answer than by its complexity class (for the primary model, the odds of an exact answer fall 2.38-fold per decade of the answer). Given code and an exact transcription, the primary model counts correctly even on 10x10 grids, and reading becomes the bottleneck: a single view misses horizontal edges, and a second reading of a rotated copy with crop checks reads 86/96 large grids exactly, against 57/96 when the upright reading is simply repeated.
[NLP-43] Same Facts Different Remedy: Linguistic Profiling and Arbitrariness in LLM and System One Decisions
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在处理消费者投诉时，因投诉人语言风格差异（如年龄、教育水平及社会经济地位，SES）所引发的公平性问题。研究发现，尽管群体层面的救济结果差异微小（货币赔偿差距均在0.5个百分点以内），但在个体层面，同一投诉经不同语言风格重写后，系统决策存在显著不稳定性——2.6%至6.3%的情况下产生不同的货币赔偿决定。关键发现在于：风格差异主要影响表达方式（如词汇、语法与礼貌程度），而非内容本身；且所有模型对事实信息的响应强度远高于对语言风格的敏感度（49–387倍）。通过将所有投诉统一重写为标准语言风格，可使GPT-4o-mini模型在教育和SES维度上的差距分别缩小94%和87%，同时保持对事实的敏感性，但未能提升个体决策稳定性，且部分概率模型的偏差被反转。此外，独立审计工具在3.3%和11.7%的重写中识别出实质性事实变更，表明语言重构仍可能引入隐性偏差。因此，论文强调：对生成式AI决策系统的公平性审计应区分报告“群体差异”与“个体决策不稳定性”两类结果，以全面评估其公平性风险。
链接: https://arxiv.org/abs/2611.08995
作者: Colin Priest
机构: UNSW Business School (新南威尔士大学商学院); UNSW Sydney (新南威尔士大学), Australia
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 29 pages, 7 figures. Code, prompts and pre-registrations: this https URL
Abstract:Large language models are increasingly used to resolve consumer complaints, and the way complainants write varies with age, education and socioeconomic status (SES). We audit three remedy-decision systems using sociolinguistic rewrites of US consumer financial complaints that are instructed to preserve every fact: two generative LLMs that return a single label (GPT-4o-mini and GPT-5-mini) and a probabilistic “System One” model that returns a distribution over remedies (Jev). Using pre-registered hypotheses tested once on held-out complaints, we separate two phenomena that single-label audits conflate. At the group scale, older, college-educated and higher-SES styles receive slightly more generous remedies; these effects replicate but are small, with every monetary-remedy gap within 0.5 percentage points. At the individual scale, two style versions of the same complaint receive different monetary decisions 2.6-6.3% of the time. In a variance-component analysis of the probabilistic model, persona components shared across the two rewrite models account for about 1.6% of within-complaint variation. The persona gaps are associated mainly with delivery (vocabulary, grammar, politeness) rather than content. Positive controls show that all three systems respond 49-387 times more strongly to remedy-relevant facts than to style. Rewriting every complaint into one standard style reduced GPT-4o-mini’s education and SES gaps by 94% and 87% while retaining fact sensitivity, but reversed two of the probabilistic model’s gaps and did not make individual decisions more stable. Independent LLM auditors flagged material fact changes in 3.3% and 11.7% of rewrites, and instability persisted among unflagged rewrites. Fairness audits of LLM decision systems should report group disparities and individual instability as separate outcomes.
[NLP-44] Beyond Truncation? Entailment-Licensed Elision and the Limits of Semantic Preservation in LLM Decoding
【速读】： 该论文旨在解决指令微调的大语言模型（Large Language Models, LLMs）在生成回答时存在冗余输出的问题，即生成内容长度远超任务实际需求。其核心挑战在于：如何在不进行额外训练的前提下，有效削减冗余内容，同时保持语义完整性与信息质量。论文提出的关键解决方案是引入一种无需训练的解码阶段控制器——蕴含许可的解码内省略（Entailment-Licensed In-Decoding Elision, ELIDE）。ELIDE 的核心机制是在每个句子边界处，通过自然语言推理（Natural Language Inference, NLI）模型判断当前待输出文本是否可由已生成内容逻辑蕴含，若成立则跳过该部分文本，从而实现内容精简。值得注意的是，尽管该方法在多项指标上优于等长截断基准，但人工评估揭示其约59%的省略会损害答案质量，且此类错误无法被自动度量检测。后续研究表明，将蕴含判断的范围限定于话语单元（discourse unit）层级可显著降低有害省略比例，将其控制在约10%以内，从而提升实用性与可靠性。
链接: https://arxiv.org/abs/2611.08975
作者: Mahmoud Alyosify,Mirna Imbabi,Eman Elkhamisi
机构: Queen’s University (皇后大学)
类目: Computation and Language (cs.CL)
备注: 8 pages, 3 figures, 2 tables
Abstract:Instruction-tuned large language models (LLMs) answer at greater length than a request needs. We ask whether a training-free controller inside the decoding loop can remove that surplus while preserving content, judged against the simplest possible baseline: truncating the model’s own answer to the same length. An entropy-gated logits processor removes 25.5-34.9% of tokens on MT-Bench development questions and 19.9-34.2% on a sealed test split within pre-registered BERTScore non-inferiority margins, yet we prove and confirm by audit that its output is always a prefix of the unmodified answer: the saving is truncation. A survey of 23 training-free mechanism families finds none that selects content better than truncation without oracle labels. We therefore introduce Entailment-Licensed In-Decoding Elision (ELIDE), which, at each sentence boundary, withholds a span from the reader only when a natural-language inference model judges it entailed by the text already shown, while leaving the model’s generation unchanged. ELIDE beats matched-length truncation on three model families, but reading 936 of its elisions on 670 public prompts shows that 59% harm the answer, a failure that no automatic measure detects. Scoping the licence to its discourse unit reduces the harmful share to about one elision in ten on two fresh tests.
[NLP-45] Using an LLM to Tackle Difficult Cases in NER for Nepali Tweets
【速读】： 该论文旨在解决低资源语言（如尼泊尔语）社交媒体数据中命名实体识别（Named Entity Recognition, NER）的挑战，尤其关注预训练语言模型（Pretrained Language Models, PLMs）在复杂或模糊实例上表现不佳的问题。其核心解决方案在于利用细调后的PLM对测试集进行难度划分，将样本分为“易例”与“难例”，并针对“难例”使用大语言模型（Large Language Models, LLMs）结合不同提示工程（prompt configuration）策略进行优化。关键发现是：通过基于语义相似性的示例选择策略并增加上下文示例数量，可使LLM在“难例”上的F1分数相比基线配置提升近10个百分点；尽管LLM在整体和“易例”上仍逊于PLM，但在“难例”上表现超越PLM，最终通过两模型融合实现了综合性能的提升。
链接: https://arxiv.org/abs/2611.08967
作者: Claire Wang
机构: Brandeis University(布兰迪斯大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Named entity recognition (NER) is challenging for social media data in a low-resource language. LLMs often lag behind pretrained language models (PLMs) in NER performance, but they show promise in handling cases that PLMs struggle with. Using a dataset of Nepali tweets, we split the test set into “easy” and “hard” cases using token probabilities from a fine-tuned pretrained language model (PLM). We then evaluate an LLM on the hard test cases using 16 different prompt configurations. We find that selecting examples based on semantic similarity and increasing the number of examples improves F1 score by almost 10 points from the baseline configuration. Although the LLM underperforms the PLM on the full and easy test sets, it surpasses the PLM on the hard test set, resulting in an improved score when the two models are combined.
[NLP-46] ResQAdam: Residual-Preserving Adaptive Optimization for Memory-Efficient Quantization Aware Training
【速读】： 该论文旨在解决量化感知训练（Quantization-aware training, QAT）中优化器内存开销过高的问题，尤其针对使用AdamW优化器时需维护每个权重矩阵两个稠密的FP32状态所导致的高内存占用。尽管低秩内存高效优化器（如GaLore和LDAdam）在常规训练中表现良好，但在QAT场景下性能显著下降，其根源在于QAT中的动量信息存在大量能量分布于主奇异子空间（top-r singular subspace）之外，而基于投影的方法会丢弃这部分信息。研究发现，仅保留残差部分的符号信息即可维持其有效性。基于此，作者提出ResQAdam：该方法仅维护一个BF16精度的动量，仅在主子空间的r×r坐标上应用Adam更新，并引入由残差计算出的归一化符号更新项以恢复被投影丢失的信息。实验表明，ResQAdam在两种从零开始的QAT方法及三种微调流程中，将权重量化至2比特时，相比AdamW降低了71%至74%的优化器状态内存开销，且在多数情况下性能接近AdamW，训练吞吐量下降不足2%。其核心创新在于通过符号保留机制有效利用了被传统投影方法舍弃的动量残差，从而在大幅降低内存消耗的同时保持模型训练精度。
链接: https://arxiv.org/abs/2611.08924
作者: Liqun Ma,Shengkun Tang,Liangyu Wang,Lin Zhao,Wei Luo,Xinjun Niu,Spandan Tiwari,Ashish Sirasao,Kun Zhang,Zhiqiang Shen
机构: Mohamed bin Zayed University of Artificial Intelligence (MBZUAI); King Abdullah University of Science and Technology (KAUST); Advanced Micro Devices, Inc. (AMD); Carnegie Mellon University (CMU)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Quantization-aware training (QAT) is the most reliable way to obtain accurate low-bit large language models (LLMs), but training with AdamW keeps two dense FP32 states for every weight matrix. We find that low-rank memory-efficient optimizers such as GaLore and LDAdam degrade substantially in QAT. A considerable part of the QAT momentum lies outside its top-r singular subspace, and projection-based methods discard this part at every step. We further observe that this residual remains useful when only its sign is kept. Based on these observations, we propose ResQAdam. It keeps a single BF16 momentum, applies Adam only to the r x r coordinates of the principal subspace, and adds a normalized sign update computed from the residual. We evaluate ResQAdam with two from-scratch QAT methods and three QAT fine-tuning pipelines, with weights quantized down to 2 bits. It reduces optimizer-state memory by 71% to 74% relative to AdamW, stays close to AdamW in most settings, and lowers training throughput by less than 2%.
[NLP-47] On the Incompleteness of do-calculus for Universal Identifiability
【速读】： 该论文旨在解决从任意观测分布和/或干预分布集合中识别因果效应的通用可辨识性（Universal Identifiability, uID）问题，该问题泛化了经典的可辨识性（ID）与广义可辨识性（gID）问题。传统方法仅针对纯观测数据或特定类型干预数据的可辨识性进行研究，并已建立相应的图模型判别准则及基于do-演算（do-calculus）与基础概率运算的算法，证明了do-演算在这些场景下具有完备性。然而，此前尚未对“do-演算完备性”给出严格的数学定义。本文首次提出一个严谨的形式化定义来刻画do-演算的完备性，并验证了已有结果中的完备性结论依然成立；但进一步表明，在更一般的uID问题中，do-演算的规则不再具备完备性，这可能是导致uID问题长期未被解决的根本原因。
链接: https://arxiv.org/abs/2611.08916
作者: Yaroslav Kivva,Aliaksandra Novik,Negar Kiyavash
机构: EPFL(洛桑联邦理工学院); Imperial College(帝国学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Probability (math.PR)
备注:
Abstract:We study the problem of causal effect identification from an arbitrary set of observational and/or interventional distributions, which we call Universal Identifiability (uID). The uID problem generalizes existing classical identifiability (ID) and general identifiability (gID) problems, which considered the identification problem from either mere observational distributions or a specific type of interventional distributions. Prior research has established graphical criteria to determine when a causal effect is identifiable in ID and gID problems. Moreover, algorithms that leverage do-calculus and basic probability manipulations have been proposed to compute these effects. Consequently, it has been shown that do-calculus is both sound and complete in these problems. Despite these well-established results, a formal definition of what it means for the do-calculus to be complete does not exist. We propose a rigorous formalization for the completeness of the do-calculus and show that the completeness results in previous work remain valid. However, for more general problems, such as uID, the rules of do-calculus are no longer complete. This is perhaps the reason why the uID problem has remained open thus far.
[NLP-48] From Procedures to Permissions: Learning Precise Subagent Authorization from Agent Skills
【速读】： 该论文旨在解决多智能体系统在执行复杂任务时，如何精准分配子任务所需工具与工作空间资源的问题。现有代理系统虽能控制工具调用和执行环境，但缺乏对每个子任务所需资源的显式指定机制，且仅依赖任务结果和执行轨迹难以有效监督权限分配的合理性。其解决方案的关键在于利用嵌入在代理技能（Agent Skills）中的过程性知识——即每一步操作所涉及的工具、文件及具体操作——作为授权决策的监督信号。研究通过从技能中合成任务与逐步执行流程，在交互环境中进行子代理委派，并基于验证后的执行轨迹构建训练数据，从而实现对任务执行与权限分配的联合优化。实验表明，该方法在ClawArena-Team基准上使工具权限精度提升13.4个百分点、工作空间权限精度提升24.1个百分点，同时整体任务成功率提高11.6个百分点（以Qwen3.5-35B-A3B为例），证明了过程性知识可为子代理协同提供任务结果无法替代的精细监督能力。
链接: https://arxiv.org/abs/2611.08905
作者: Wenkai Yu,Ning Ding,Wentao Zhang,Yehui Tang
机构: Samsung Research, Beijing, China(三星研究院，北京，中国); Peking University (北京大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:As language agents tackle increasingly complex tasks, they increasingly rely on multiple agents to coordinate specialized work, requiring decisions not only about what each subagent should do but also which tools and workspace resources it should access. Existing agent systems provide mechanisms for controlling tools and execution environments, but do not directly specify the resources required by each delegated subtask. Moreover, task outcomes and execution trajectories alone provide limited supervision for determining whether a subagent has received appropriate permissions. We observe that Agent Skills encode procedural knowledge describing the operations, tools, and files required at each step, providing a natural source of supervision for authorization. Building on this observation, we synthesize tasks and step-by-step procedures from Skills, execute them through subagent delegation in interactive environments, and use verified trajectories to construct training data for task execution and authorization. On ClawArena-Team, our method improves tool permission precision by 13.4 percentage points and workspace permission precision by 24.1 percentage points, while increasing overall task success by 11.6 percentage points for Qwen3.5-35B-A3B. On Terminal-Bench 2.1 and SkillsBench, our method remains competitive with existing agent data synthesis approaches. These results show that the procedural knowledge embedded in Agent Skills provides a source of supervision for precise subagent orchestration that task outcomes alone cannot provide.
[NLP-49] Richness Is Not Evenness: Ordered Renyi Crossings under Self-Training
【速读】： 该论文旨在解决生成式语言模型在递归自训练（recursive self-training）过程中出现的“模型坍缩”（model collapse）问题，即模型在持续自我生成文本进行微调时，词汇表逐渐缩小，高频词（主要为功能词）的主导地位被削弱，导致语言多样性下降。其核心解决方案的关键在于通过系统性地分析不同训练采样温度（training-data sampling temperature）下多种信息熵度量（如最小熵、碰撞熵、香农熵和观测丰富度）的变化趋势，揭示熵增与熵减之间的动态转折点。研究发现，随着训练温度升高，各类熵值的变化顺序呈现出可重复的交叉现象：最小熵最先由损失转为增益，随后是碰撞熵、香农熵和观测丰富度；在此区间内，词类数量减少但最常见词的支配性减弱，表明模型在压缩词汇的同时实现了更均衡的分布。特别地，在无截断训练且温度T=1时，第9代模型词汇量降至初始的13,000词，词类数减少24%，但碰撞熵反而提升0.91比特，说明模型在降低复杂度的同时仍保持一定多样性。该现象在不同采样网格、测量采样器、固定词预算及多组随机种子下均稳定存在，并在印地语和多种文本表示形式中验证了熵交叉顺序的一致性。进一步通过解码与采样机制的分离实验（如下一词身份识别和类别模拟），确认该现象并非由采样偏差引起，而是模型内在演化特性所致。此外，使用人类文本微调同样可引发类似熵分裂现象，且不伴随似然下降，表明单一熵序无法可靠指示模型质量劣化。因此，作者建议在报告模型性能时应同时提供指定词预算下的熵谱（entropy profile）与独立的质量评估指标，以全面刻画模型在自训练过程中的演化状态。
链接: https://arxiv.org/abs/2611.08892
作者: Pranjal Srivastava
机构: University of Michigan(密歇根大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Recursive self-training of language models, a setting studied in model collapse, can reduce vocabulary while making frequent words, mostly function words, less dominant. We fine-tune English GPT-2 on its own output for nine generations, vary the training-data sampling temperature, and measure every model with one fixed sampler. As training temperature rises, Renyi entropy changes turn from loss to gain at distinct temperatures: min-entropy first, then collision entropy, Shannon entropy and observed richness. Between the outer crossings, word types fall while the most-common word becomes less dominant. Untruncated training at T = 1 lies in this interval: at 13,000 words, generation 9 has 24% fewer types than generation 0 yet collision entropy is 0.91 bits higher. The ordering persists with a finer grid, different measurement samplers, a fixed word budget and five fresh training seeds. Min-entropy again crosses before richness in four further settings (in Hindi only at interpolated roots) and in 12 of 13 distinct measurements under eight text representations. Next-token identities and a categorical simulation isolate decoding and sampling effects; the crossing order remains empirical. Fine-tuning on human text can produce the same split while likelihood improves, and no single order’s sign reliably identifies worse likelihood across the tested settings. We recommend reporting entropy profiles at stated budgets with an independent quality measure.
[NLP-50] Guardrail Erosion in Domain Fine-Tuned Large Language Models : A Paired Audit of Knowledge-Centric Abstention in Biomedical LLM s ML4H2026 ALT
【速读】： 该论文旨在解决生物医学领域微调大语言模型（Large Language Models, LLMs）过程中出现的知识性回避行为退化问题。尽管微调可显著降低模型训练成本并提升特定领域的性能，但已有研究表明其可能导致安全防护机制弱化，尤其在知识性回避（knowledge-centric abstention）方面表现不佳——即模型在面对不确定或超出知识范围的问题时，未能恰当选择不回答。现有研究多聚焦于价值导向的拒绝响应，而本文则关注知识导向的回避行为。为评估微调对知识性回避的影响，作者对25个生物医学领域微调模型及其对应的11对基础模型进行了系统审计，使用MedAbstain和ClinDet-Bench两个基准进行测评。结果显示，基础模型在MedAbstain上的适当回避率中位数仅为8.3%，微调后进一步下降至4.2%（在25对模型中有21对呈现下降），且变化趋势在不同模型对间表现出高度异质性。这一发现揭示了当前微调方法在提升领域能力的同时，可能损害模型应有的谨慎性。因此，论文提出的关键解决方案在于发展能够同时增强领域适应性与保持适当知识性回避的微调方法，以确保模型在实际应用中的可靠性与安全性。
链接: https://arxiv.org/abs/2611.08780
作者: Stephen Peng,Vineet Jagadeesan Nair,Noah Broestl,Anne Friederike Kleppe,Steven Mills,Adi Zolotov
机构: BCG X AI Science Institute, Boston Consulting Group(波士顿咨询集团); Centre for Human-Inspired Artificial Intelligence, University of Cambridge(剑桥大学人类启发式人工智能中心)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: To be published in the Proceedings of the 6th Machine Learning for Health Symposium (ML4H 2026), PMLR volume 346
Abstract:The number of domain fine-tuned models, particularly in the biomedical domain, is rapidly expanding. Fine-tuning promises high-performing models at a fraction of the cost of training a base model, but it can also produce unintended behavioral changes. Multiple studies have shown that safety guardrails can erode with fine-tuning, yet large-scale studies across model architectures, data mixtures, and fine-tuning methodologies remain limited. Where prior work in this space has studied value-centric refusal, this work focuses on knowledge-centric abstention. We audit 25 biomedical domain fine-tuned large language models against 11 paired base checkpoints on MedAbstain and ClinDet-Bench to evaluate the impact of fine-tuning on knowledge-centric abstention. We find that defined appropriate abstention measured by these benchmarks, already low in base models (median 8.3% on MedAbstain), falls further after fine-tuning (median 4.2%, decreasing in 21 of 25 pairs on MedAbstain), and that the magnitude and direction of change are heterogeneous across pairs. These results motivate fine-tuning methods that improve domain-specific capability while preserving appropriate caution.
[NLP-51] Jacobian-Causal Memory Control for Persistent Memory in Language Model Systems
【速读】： 该论文旨在解决持久性内存系统（persistent-memory systems）在信息检索与推理过程中，如何精确控制记忆干预（memory intervention）对当前推断影响的问题，尤其关注如何在保证目标效果的同时，有效约束副作用并实现可回滚的动态调控。其核心挑战在于，传统方法难以在不引入不可控干扰的情况下，量化并管理多条记忆信息对模型输出的非线性、非加性交互作用。解决方案的关键在于提出雅可比因果记忆控制（Jacobian-causal memory control, JCMC），通过在共享行为测量空间中以预测的影响签名（influence signature）形式表征每项授权的记忆干预，基于目标效应与副作用预算进行选择，并在执行后通过对比预测与实际观测效应来决定是否保留或回滚干预。该方法在显式平滑性与有界估计假设下，建立了二次局部近似误差上界，刻画了由一组签名可达的目标效应范围，给出了签名误差下的选择后悔（regret）上界，并对同时激活的记忆间非加性交互进行了量化约束。该研究虽未提供基准测试数据，但定义了一种成对反事实估计器和评估协议，用于与相关性、图结构及策略过滤等基线方法进行比较。若这些简化基线在相同模型、记忆、策略及成本预算下能够达到同等水平的目标效应定位、副作用控制与回滚效用，则本文的核心主张将被削弱或证伪。
链接: https://arxiv.org/abs/2611.08779
作者: Marvin Danig(Achiral Research)
机构: Achiral Research(非手性研究)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Architecture and formal analysis preprint. Four conditional theorems; no benchmark measurements. Companion research essay: this https URL
Abstract:Persistent-memory systems choose not only which records are retrieved but which prior information can alter current inference. We propose Jacobian-causal memory control (JCMC), which represents each authorized memory intervention by a predicted influence signature in a shared behavioral measurement space, selects interventions against a target effect and side-effect budget, and compares predicted with observed effects before retaining or rolling back the intervention. Under explicit smoothness and bounded-estimation assumptions, we prove a quadratic local-approximation bound, characterize the target effects reachable by a set of signatures, bound selection regret under signature error, and bound non-additive interactions between simultaneous memories. The paper reports no benchmark measurements; it defines a paired counterfactual estimator and an evaluation protocol comparing JCMC with relevance, graph, and policy-filtered retrieval baselines. The central claim is narrowed or falsified if those simpler baselines match effect targeting, side-effect control, and rollback utility under equal model, memory, policy, and cost budgets.
[NLP-52] Semantic Conditional Similarity (SCS): A Formal Mathematical Framework
【速读】： 该论文旨在解决多文档语境下传统语义文本相似度（STS）方法无法区分各来源对目标文档的边际语义贡献的问题，尤其在多个源文档共同影响一个或多个目标文档时，成对STS因共享词汇而难以剥离个体贡献。其核心解决方案是提出语义条件相似度（Semantic Conditional Similarity, SCS），通过集合论操作从参考文档中移除与条件文档共享的语义成分，从而实现对特定源文档贡献的精准量化。SCS的关键创新在于：引入方向性，将关系划分为横向（源间）与纵向（源到目标）；通过条件化三元组（目标-参考-条件文档）构建角色依赖的非对称比较框架；实现跨文档条件化，弥补现有C-STS方法仅限于文档内条件化的局限；并采用无阈值的集合分解方式，自动分离共享与残余内容，避免人为设定相似度阈值带来的偏差。该框架可扩展至1→n及m→n情形，覆盖完整的文档数量谱系，其得分应作为序数排名理解，反映边际贡献程度而非绝对度量。这一方法揭示了传统相似性度量中常被假设的度量空间公理在角色差异化多文档场景中的表达力限制，表明引入条件三元组可恢复成对对称比较所抑制的信息。
链接: https://arxiv.org/abs/2611.08778
作者: Mehmet Murat Albayrakoğlu
机构: Işık University (伊什克大学); Istanbul, Turkey
类目: Computation and Language (cs.CL)
备注: 34 pages, 6 figures, 8 tables, to be submitted for publication
Abstract:Semantic Textual Similarity (STS) operates on document pairs, returning a single scalar per comparison. In a multi-document setting where several sources may have contributed to one or more targets, pairwise STS cannot disentangle each source’s marginal semantic contribution from their shared vocabulary. In this paper, Semantic Conditional Similarity (SCS) is introduced as a conditional form of STS that computes similarity between two documents after eliminating semantic elements shared with a third via set-theoretic operations. SCS makes four contributions: 1) it introduces directionality, partitioning pairwise relations into lateral (inter-source) and vertical (source-to-target) classes; 2) conditioning induces role-dependent asymmetry by extending the comparison domain from unordered pairs to ordered, conditioned triples of a target, a reference, and a conditioning document; 3) SCS introduces inter-document conditioning by subtracting shared content from a third document, complementing the intra-document conditioning in recent C-STS work; and 4) the set-theoretic decomposition achieves threshold-free partitioning of shared and residual content, requiring no arbitrary similarity cutoff. Although developed for the n \to 1 configuration, the framework extends to 1 \to n via vertical role reversal and to m \to n via independent composition of source- and target-side conditioning. With the established 1 \to 1 case, these span the full spectrum of document cardinality. SCS scores should be interpreted ordinally, as rankings of marginal contributions rather than absolute measurements. The broader implication is that metric space axioms commonly assumed in similarity formalizations constrain expressiveness in role-differentiated multi-document settings, and that extending the domain to conditioned triples can recover information that pairwise-symmetric comparisons suppress.
[NLP-53] AttentionSwitch: Jailbreak Defense for LLM s via Joint Learning of Sparse Attention-Head Selection and Scaling
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在面对越狱攻击（jailbreak attacks）时，其安全机制易被绕过而生成有害或非预期输出的问题。现有防御方法通常通过微调模型或对特定组件进行针对性干预来提升安全性，但此类方法存在计算成本高、可能导致灾难性遗忘等缺陷。更关键的是，现有方案往往将“干预位置选择”与“干预强度设计”分步优化，忽略了二者之间的内在耦合关系——最优干预位置依赖于具体的干预方式，反之亦然，导致整体防御效果受限。为此，本文提出AttentionSwitch，作为首个在冻结的LLM中联合优化干预位置与干预强度的越狱防御框架。其核心创新在于：学习一组稀疏的注意力头，并结合查询相关的乘法缩放因子，实现对选定注意力头输出的增强或抑制，且不改变其表征方向。该方法在多种主流语言模型和越狱攻击场景下均显著优于现有防御策略，尤其在训练数据稀缺的情况下表现突出。
链接: https://arxiv.org/abs/2611.08752
作者: Doniyorkhon Obidov,Honggang Yu,Xiaolong Guo,Kaichen Yang
机构: Michigan Technological University(密歇根理工大学); Miami University(迈阿密大学); Lehigh University(利哈伊大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR)
备注:
Abstract:Large language models have achieved remarkable performance across a wide range of tasks, yet they remain susceptible to jailbreak attacks that circumvent safety mechanisms and elicit harmful or unintended outputs. Although safety can be improved through model fine-tuning, such approaches can be computationally expensive and may degrade previously acquired capabilities through catastrophic forgetting. Recent jailbreak defenses have therefore explored targeted interventions on specific model components associated with unsafe behavior. However, existing approaches typically decouple two inherently related decisions: where to intervene and how to modify the selected components. Such sequential optimization may be suboptimal because the optimal intervention location can depend on the intervention applied there, and vice versa. In this work, we introduce AttentionSwitch, the first jailbreak defense to jointly optimize intervention sites and strengths within a frozen LLM. Specifically, AttentionSwitch learns a sparse set of attention heads together with query-conditioned multiplicative scaling factors, allowing the outputs of the selected heads to be either attenuated or amplified without altering the directions of their representations. Across popular language models and jailbreak attacks, AttentionSwitch outperforms existing defenses, with particularly strong results when training data is scarce.
[NLP-54] Structuring Survey Histories for LLM Digital Twins: A Controlled Study of Background Decision Procedure and Evaluation Profiles NEURIPS2026
【速读】： 该论文旨在解决基于大语言模型（LLM）的数字孪生系统在预测个体在新情境中行为反应时，如何有效构建用户历史响应的结构化表示以提升预测准确性的问题。现有研究表明，压缩问卷历史数据并不必然降低预测性能，提示表示结构本身可能比信息量更为关键。为此，研究提出并评估了一种固定结构的表示方法——背景-决策过程-评估（BDE），将每位受访者的问卷历史划分为三个核心维度：背景（Background）、决策过程（Decision procedure）和评估（Evaluation）。在Twin-2K-500数据集上的实验表明，相较于原始问卷文本，通过完全推断生成的BDE表示在gpt-5.4-nano和gpt-5.4-mini模型上分别提升了1.27个百分点（p < 0.01）和1.19个百分点（p < 0.001）的整体预测准确率。进一步分析显示，尽管增加推理深度未带来显著性能提升，但引入“评估”模块可使整体准确率分别提高1.44个百分点（p < 0.001）和1.22个百分点（p < 0.001），且其在价格预测任务中亦优于等长自由格式描述。因此，该研究的关键发现在于：BDE表示的优势并非源于数据压缩或长度差异，而是主要归因于其合理的组件构成，即三元结构中的“评估”部分对预测性能具有决定性贡献。
链接: https://arxiv.org/abs/2611.08730
作者: Iris Ye,Tianze Deng,Ozan Candogan
机构: 未知
类目: Computation and Language (cs.CL)
备注: 10 pages. Camera-ready version for SocialAgent: Second Workshop on Large Language Models for Social Reasoning and Simulation, NeurIPS 2026
Abstract:LLM-based digital twins aim to predict how individuals would respond in new settings from representations of their prior responses. Existing evidence that compression of survey histories need not reduce predictive accuracy suggests that representation structure, rather than information volume alone, may be an important design choice. We study BDE, a fixed representation that organizes each respondent’s survey history into Background, Decision procedure, and Evaluation. We evaluate it on Twin-2K-500 (Toubia et al., 2025) by predicting each respondent’s held-out answers on 17 tasks. In a paired evaluation with 100 respondents, a fully inferred BDE profile improves overall accuracy over the raw survey transcript by 1.27 percentage points (pp) on gpt-5.4-nano (p 0.01) and 1.19 pp on gpt-5.4-mini (p 0.001). Three analyses then examine the schema, extraction depth, and component composition. Against a length-matched free-form profile written from the same transcript by the same extractor, BDE improves pricing accuracy on both simulators (2.15 pp on nano, p 0.01; 1.38 pp on mini, p 0.05), while its overall advantage is significant only on nano (1.05 pp, p 0.05). Allowing deeper inference during extraction yields no statistically significant gain on either simulator. By contrast, adding Evaluation to a Background–Decision procedure profile improves overall accuracy by 1.44 pp on nano (p 0.001) and 1.22 pp on mini (p 0.001). These results indicate that the advantage over raw transcripts is not explained by compression or length alone, and attribute it primarily to profile composition rather than deeper extraction.
[NLP-55] Patching the Blind Spots: Reversible LLM Rule Overlay for Frozen Fraud Detectors
【速读】： 该论文旨在解决工业场景中基于表格数据的欺诈分类任务中，主流的梯度提升树（Gradient-Boosted Trees）模型在部署后存在“盲区”（blind spots）且重新训练模型存在风险的问题。其核心挑战在于如何在不修改现有生产模型的前提下，安全有效地提升模型对未覆盖欺诈模式的识别能力。解决方案的关键在于提出一种无需改动原模型的可逆式补丁框架：利用大语言模型（LLM）根据SHAP值识别出被原模型低估的特征，生成以这些特征为锚点的规则；随后通过经验支持筛选并校准生成的噪声规则池，形成一个可加、可逆的对数提升（log-lift）叠加层，该叠加层在生产环境的决策阈值处运行。为降低单个LLM预测的高方差问题，采用三个多样化模型通过等权重逻辑值平均进行集成，以牺牲峰值性能为代价换取更稳定的最坏情况表现。实验表明，在真实电商反欺诈系统上，该方法在选择性偏差数据下持续训练性能退化时仍能实现平均+2.26%召回率（ΔR@P）和+1.60%精确率（ΔP@R）的增益；在公开基准的七种设计偏差变体上，所有变体均获得ΔP@R提升，同时ΔR@P保持接近中性。每个修正规则均为可读的谓词表达，明确描述其所覆盖的欺诈模式。
链接: https://arxiv.org/abs/2611.08714
作者: Bo Qu,Zhurong Wang,Frank Zahradnik,Ming Ouyang
机构: eBay(易贝网)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 11 pages, 1 figure, 7 tables
Abstract:In industrial tabular fraud classification, gradient-boosted trees remain state-of-the-art, but even a well-trained model has blind spots, and retraining the deployed model is not always safe. We propose a framework that patches such a model without modifying it: LLMs propose rules anchored on features the model underweights by SHAP, and the resulting noisy pool is screened for empirical support and calibrated into an additive, reversible log-lift overlay that runs at the production cut-off. Single-LLM benefit is high-variance and inverts between datasets, so we pool three diverse backbones by equal-weight logit averaging, trading peak gain for a bounded worst case. On a real-world e-commerce fraud detector trained on selection-biased data, continued training regresses even when carefully tuned, whereas the overlay gains on average +2.26% recall at equal precision ( \Delta R@P) and +1.60% precision at equal recall ( \Delta P@R) over production cut-offs spanning 30-57% precision and 23-63% recall. On a public benchmark of seven designed bias-shift variants it raises \Delta P@R on every variant and keeps \Delta R@P near neutral. Each amendment is a readable predicate naming the fraud pattern it covers.
[NLP-56] RL-ATTACK: A Generalized Black-Box Adversarial Attack on Language Models via Reinforcement Learning
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在对抗性扰动下的安全性问题，即模型容易受到精心设计的输入扰动影响而产生非预期输出。现有黑盒攻击方法在无法获取梯度信息的情况下，依赖外部先验知识盲目猜测对抗变换，忽略了目标模型反馈的关键信息。为此，本文提出一种基于强化学习的可训练黑盒攻击框架，通过与目标模型交互并利用其反馈信号来自主学习生成针对性强的对抗性变换。该方案的核心在于将攻击过程建模为一个序列决策问题，使攻击者能够动态优化扰动策略，在保证语义合理性的同时显著降低模型的准确率。实验结果表明，该方法在自然语言推理和生成式问答任务中均表现出更优的攻击效果：在自然语言推理任务中，仅用20次查询即可将攻击后准确率降至2%；在生成式语言模型攻击中，训练后的攻击者性能优于未训练攻击者，并在6个测试场景中的4个超越强基线优化方法。此外，研究还发现不同奖励设计与训练设置的效果并无绝对优劣，需根据具体任务权衡。
链接: https://arxiv.org/abs/2611.08711
作者: Taikun Chen,Yue Zhang,Wei Qian,Hui Fang
机构: University of Delaware(特拉华大学)
类目: Computation and Language (cs.CL); Cryptography and Security (cs.CR)
备注: Accepted as a regular paper at IEEE EdgeCom 2026
Abstract:Large language models (LLMs) exhibit vulnerabilities to adversarial perturbations that can induce unintended responses. These vulnerabilities can be identified via intentional attacks and subsequently used to improve model safety. In black-box settings where gradients are inaccessible, existing methods tend to guess adversarial transformations using external knowledge, ignoring the information provided by the target model. To mitigate this, we propose a trainable black-box framework based on reinforcement learning, which learns from the target model to generate tailored adversarial transformations. We demonstrate that our approach achieves lower post-attack accuracy, greater query efficiency and better semantic preservation on both natural language inference and generative QA tasks. On natural language inference, it drives post-attack accuracy to 2% within 20 queries. Against generative language models, our trained attacker consistently achieves lower post-attack accuracy than the untrained attacker and wins in 4 out of 6 settings when compared to a strong optimization-based baseline. We further experiment with multiple reward designs and training settings, and we find that no design or setting is consistently dominant. Our code is available at this https URL.
[NLP-57] Dual-Stream LLM -Guided Reward Design for Reject-Inference-Aware Reinforcement Learning in E-Commerce Payment Fraud Detection ACL ACL2025
【速读】： 该论文旨在解决生产环境中欺诈检测模型因数据截断（censored data）导致的偏差问题：由于历史系统拦截的交易未被标注，模型仅基于已通过的交易进行训练，从而学习到有偏的“批准仅”（approved-only）分布。其核心解决方案是将生成式 AI (Generative AI) 引导的奖励设计框架扩展至奖励函数内部，提出一种双流奖励机制（dual-stream reward）。该机制的关键在于：第一流采用标准的金额加权精确率目标对已标注交易进行评估；第二流则通过对比最新梯度提升风险信号与原始拦截决策所依据的历史评分，对被旧系统拦截但无标签的交易进行打分。该方法不预设哪些特征重要，而是向大语言模型（LLM）提供一个包含多源证据的超集，由其根据评估反馈自主选择并决定使用哪些信号，且因其以代码形式生成奖励函数，决策过程具备可审计性。在四个月的跨时测试中，该双流代理在80%、85%和90%召回率下分别比仅使用标注数据的基线模型提升4.7、2.4和3.0个百分点的金额加权精确率，并平均优于生产基准6.8个百分点，同时保持与原架构一致，无额外推理开销。
链接: https://arxiv.org/abs/2611.08706
作者: Bo Qu,Zhurong Wang,Frank Zahradnik,Ming Ouyang
机构: eBay(易贝)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 12 pages, 1 figure, 3 tables. Extends Qu et al., ACL 2025 Industry Track: this https URL
Abstract:Production fraud-detection models learn from a censored slice of their own traffic: transactions a legacy system blocked never receive fraud labels, so models train on a biased, approved-only distribution. We extend LLM-guided reward design, the framework of Qu et al. (2025) in which an LLM evolves the reward of a two-step REINFORCE fraud-blocking policy, to attack this reject-inference problem inside the reward. A single LLM designs a dual-stream reward: stream 1 applies the standard dollar-weighted precision objective to labeled transactions; stream 2 scores legacy-blocked, unlabeled transactions by contrasting the newest gradient-boosted risk signal against the legacy scores that drove the original block. Rather than prescribe which signals matter, we expose a superset of evidence and let the LLM decide for itself which to use, guided by evaluation feedback alone; because it writes the reward as code, that choice is directly auditable. On a held-out four-month out-of-time test, the dual-stream agent beats the prior labeled-only agent at every target recall, by +4.7, +2.4 and +3.0 points of dollar-weighted precision at 80%, 85% and 90% recall, and the production baseline by +6.8 points on average. The resulting policy is architecturally identical to that of Qu et al. (2025) and adds no inference cost.
[NLP-58] Contextuality and Contradictions: Exploring Real-World Identifications of AI-Generated Writing
【速读】： 该论文旨在解决在生成式AI（Generative AI）内容广泛渗透人类互动各领域的背景下，人们如何在实际语境中识别和判断文本是否由人工智能生成这一复杂问题。其核心挑战在于，随着文本来源的不确定性加剧，传统的“人类作者”预设已被打破，导致人们对文本真实性的判断日益依赖于多重、动态且情境化的证据体系。论文的关键解决方案在于结合语言人类学视角，通过定性编码与词汇演变的计算分析，揭示了实际识别过程中所依赖的多样化证据类型——包括具体的文本特征、内容判断、循环推理以及对作者道德品质的假设，并指出这些识别行为本质上并非基于单一技术指标，而是源于一种“语用不妥”（pragmatic infelicity）：即文本元素与特定语境下预期的适切性之间存在的感知错位。因此，真实世界中的AI识别不仅是对“类AI”特征的被动指认，更是一种用于建构和维护社会语言规范的修辞资源，涉及何为可言说、何为恰当乃至何为“人类”的界定。由此，该研究警示：若将生成式内容检测系统设计为脱离具体语境的通用标准，则可能在大规模应用中固化并强化训练数据所反映的特定语境规范，从而加剧社会认知偏见与话语权力的不平等。
链接: https://arxiv.org/abs/2611.08704
作者: Michelle Morgenstern,Julia Witte Zimmerman,Tabia Tanzin Prama,Kristine Harootunian,June Clark,Bryn Loftness,Calla Beauregard,Juniper Lovato
机构: University of Vermont(佛蒙特大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:As AI-generated content spreads into nearly every domain of human interaction, practices and discourses of AI detection, both formal and informal, have become an everyday part of how people interact with text and with each other. In this paper, we explore the real-world practices and discourses of AI identification that have emerged in response to a shifting discursive landscape where the human provenance of a text can no longer be taken for granted. We pair qualitative coding with computational analysis of lexical change, grounded in linguistic anthropological perspectives on language and communication, to investigate AI identifications among English Wikipedia editors and students. We find that these identifications draw on a diversity of evidentiary types, from discrete textual features and judgments about content to circular sense-making claims and presumptions about the author’s moral character, and that they are contradictory, dynamic, and contextually dependent. In practice, what often motivates an AI identification is not the presence of any specific textual element, but pragmatic infelicity: a perceived mismatch between the elements present and those deemed appropriate within the text’s particular context. Consequently, real-world identifications of AI-generated writing do not merely neutrally point out the perceived presence of “AI-like” features, but can be leveraged as a discursive resource to maintain and define sociolinguistic norms: what can be said in a given context, what is appropriate, and, ultimately, what is human. Further, the design and implementation of systems and tools that treat AI-likeness as a context-free property of text may therefore inherit, and enforce at scale, the norms of whichever contexts they were trained on.
[NLP-59] Compression of Emotion in Language Model Representations
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在处理细粒度情感分类任务时，其内部表示空间中情感结构组织方式不明确的问题。尽管现有研究已验证LLMs在识别文本情感及生成流畅情感语言方面的能力，但模型输出结果无法揭示其内部对27种细粒度情感类别及其与中性情绪共同构成的28类情感标签的表征机制。论文以GoEmotions数据集为基准，系统评估了七种LLMs在细粒度情感分类、情感家族（emotion-family）和情感极性（sentiment）等不同层级上的表现。研究发现，在未使用情感标签的无监督聚类（基于KMeans协议）条件下，模型隐藏层表示呈现出显著的“过压缩”（overcompression）现象：最清晰的聚类结构仅能区分正负情感两大类，且恢复更粗粒度的情感分组所需的聚类数远少于完整的28类体系。然而，线性探测（linear probes）仍能有效提取大量细粒度情感信息，表明细粒度情感特征虽被压缩，但并未完全丢失。因此，该研究的核心结论是：尽管模型的无监督几何结构主要由宽泛的情绪极性主导，但细粒度情感信息仍保留在内部表示中，其关键在于通过有监督的探测方法可有效挖掘这些潜在表征。
链接: https://arxiv.org/abs/2611.08683
作者: Laya Iyer,Sanmi Koyejo,Chen Shani
机构: Stanford University (斯坦福大学); Tel Aviv University (特拉维夫大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) have been evaluated on their ability to identify emotions expressed in text and produce fluent emotional language. However, output predictions alone do not reveal how fine-grained emotion distinctions are organized internally. We study seven LLMs on GoEmotions, an English Reddit dataset annotated with 27 fine-grained emotion categories plus neutral. We compare model predictions and emotion-prompted final-token representations with the 28 GoEmotions categories and with coarser emotion-family and sentiment groupings. We find a consistent pattern of \textitovercompression relative to the fine-grained GoEmotions taxonomy. Under the tested KMeans protocol, when we cluster their hidden representations without using emotion labels, the clearest split is into two groups that largely separate positive from negative emotion. Coarser emotion groupings are also recovered with fewer clusters than the full 28-category taxonomy. Yet linear probes retain substantial fine-grained predictive information. These results suggest that fine-grained emotion information remains accessible even though broader affective structure more strongly organizes the unsupervised geometry. Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.08683 [cs.CL] (or arXiv:2611.08683v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2611.08683 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-60] Greedy Byte-Pair Encoding Achieves at Least 0.361 of the Optimal Compression
【速读】： 该论文旨在解决字节对编码（Byte-Pair Encoding, BPE）在文本压缩中通过贪心策略进行符号对合并时，其性能与最优解之间差距的理论边界问题。具体而言，研究关注的是：在给定字符串、合并预算及任意破局规则下，贪心算法所实现的字符串长度缩减量与最优缩减量之比的下界。此前已知该比值至少为1/3，但这一界限长期未被突破。本文首次证明了该比值始终不低于0.361，显著优于之前的1/3下界。关键突破在于揭示了旧有分析框架——即“配对打包松弛”（pair-packing relaxation）——的紧致性（tightness），这解释了为何1/3成为难以逾越的障碍；因此，要获得更优下界必须深入利用最优解的结构特性。作者进一步指出，在贪心策略一次性破坏三个最优合并点的情况下，其中至少一个通常可通过“就近迁移”（moving next door）来补偿，因为最优解中同一规则的所有出现形式具有不可区分性；其余情形则由后续步骤予以弥补。研究过程中，作者精确求解了两次合并的最坏情况比率，给出了三次合并的上界，并最终提炼出一个待验证的单一不等式，该不等式若得证，即可完全解决此问题。
链接: https://arxiv.org/abs/2611.08671
作者: Kundan Baliga,Aiden Xie,Aditya Byju
机构: University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校); Georgia Institute of Technology (佐治亚理工学院)
类目: Data Structures and Algorithms (cs.DS); Computation and Language (cs.CL)
备注: 27 pages, 5 figures. Code and certificates: this https URL
Abstract:Byte-Pair Encoding (BPE) builds a tokeniser by repeatedly merging a most frequent adjacent pair of symbols; its utility is the total reduction in string length achieved by k merges. It was known that this greedy rule achieves at least a third of the optimal reduction. We prove that it always achieves at least 0.361 of the optimal reduction in string length, for every string, every budget and every tie-breaking rule; this is the first improvement of the bound one third. We also explain why one third was a barrier: the pair-packing relaxation behind it is tight, so a better bound must use more of the optimum’s structure. When greedy destroys three of the optimum’s merge points at once, one of them can usually be moved next door, because all copies of a rule of the optimum look alike; the remaining cases are paid for later. Along the way we determine the worst-case ratio for two merges exactly, bound it for three, and isolate a single inequality that would settle the problem.
[NLP-61] ChainFlow: Few-Step Latent Flow Matching for Speculative Decoding
【速读】： 该论文旨在解决生成式 AI（Generative AI）在大语言模型（Large Language Model, LLM）推理加速过程中，现有推测解码（speculative decoding）方案存在的核心瓶颈问题：传统自回归式推测器（autoregressive drafter）因逐个生成令牌导致推测成本随推测长度线性增长，而扩散式推测器（diffusion drafter）虽可并行生成但忽略了令牌间的依赖关系，从而限制了生成质量与推理效率的平衡。其解决方案的关键在于提出ChainFlow——一种全新的推测解码范式及首个基于目标模型隐藏状态进行流匹配（flow-matching）的推测器。ChainFlow将推测过程建模为对目标隐藏状态空间的短时传输，仅需一到两个流步骤即可完成，远少于传统从噪声中去噪所需的多步迭代；通过变分自编码器（VAE）压缩隐藏状态以实现低开销的推测计算，轻量级反馈头恢复令牌间的上下文依赖关系，并结合树形推测（tree drafting）机制提升接受率。实验表明，在相同推测器规模和训练数据条件下，ChainFlow在vLLM框架下相较于EAGLE-3和DFlash分别实现39%和11%的性能提升，且随着推测器逐步适配新语言，优势进一步扩大。
链接: https://arxiv.org/abs/2611.08669
作者: Selim Aktas,Yusuf Erdem Erdem,Ahmed Zeer,Himmet Toprak Kesgin,Muzaffer Kaan Yuce,Mehmet Fatih Amasyali
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Speculative decoding has become a standard way to accelerate LLM inference: a lightweight drafter proposes tokens that the target verifies in one parallel pass. Existing drafters face trade-offs that limits their speedups: autoregressive drafters generate tokens one at a time, so drafting cost grows with draft length, while diffusion drafters generate tokens in parallel but predict each position independently, ignoring inter-token dependencies. We introduce ChainFlow, a new speculative decoding paradigm and the first flow-matching drafter conditioned on the target’s hidden states. ChainFlow treats drafting as a short transport over hidden states, needing one or two flow steps instead of the many denoising steps of generation from noise. A variational autoencoder compresses these states so drafting cost barely grows with target size, lightweight feedback heads restore inter-token dependencies, and tree drafting raises acceptance. On Qwen3.5 models, at matched drafter size and training data, ChainFlow outperforms EAGLE-3 and DFlash in vLLM by 39% and 11%, respectively. These gains further widen as drafters are adapted to new languages sequentially. Our ablations cover drafter design and concurrent serving, and we open-source our code and drafter models.
[NLP-62] urkish Multimodal Video Fact-Checking: A New Dataset and a Comparison of Zero-Shot Agent ic and Fine-Tuned Systems
【速读】： 该论文旨在解决社交媒体中基于视频的虚假信息在低资源语言（如土耳其语）中缺乏有效事实核查工具的问题。现有事实核查系统几乎全部聚焦于英文文本，导致土耳其语等语言的多模态视频虚假信息难以被准确识别与验证。为此，本文提出了TeViD（土耳其语视频事实核查数据集），据作者所知，这是首个面向土耳其语的多模态视频事实核查数据集。其关键解决方案在于：首先构建了一个高质量、经过平衡的数据集，通过整合国际事实核查网络（IFCN）认证机构发布的文章，并融合翻译自Snopes的案例及合成样本以缓解类别不平衡；其次，提出两种基于同一开源基础模型的验证管道——3MFact-TR（Pipeline 1）与微调模型（Pipeline 2）。其中，3MFact-TR采用Qwen2.5-7B大语言模型与Qwen2-VL-7B视觉-语言模型，引入语义不一致检测器、5W1H问题管理模块及双语证据检索器（排除已知事实核查网站以防止泄露）；而微调模型则利用低秩适应（LoRA）对单一Qwen2.5-7B模型进行微调，实现仅依赖声明与视频文本描述的闭卷判断。实验结果表明，该方案在准确性、宏平均F1、真类F1及真类召回率等方面均表现出色，并通过消融实验验证了各组件的有效性。
链接: https://arxiv.org/abs/2611.08651
作者: Azra Açıl
机构: Queen Mary University of London (QMUL)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 11 pages, 3 figures, 5 tables. MSc dissertation, Queen Mary University of London, 2026
Abstract:Video-based misinformation spreads rapidly on social media, but almost all existing fact-checking systems focus exclusively on English text. Tools for low-resource languages such as Turkish are severely limited. In this work, we present TeViD (Turkish Video Fact-checking Dataset), to the best of our knowledge the first multimodal video fact-checking dataset for Turkish. We compiled the dataset from articles published by this http URL, an IFCN-certified organisation, and enriched it with translated Snopes articles and synthetic examples to reduce class imbalance. Building on TeViD, we develop two verification pipelines that share the same open-source base model. First, in 3MFact-TR (Pipeline 1), we adapt the multi-role 3MFact framework to Turkish with the Qwen2.5-7B large language model and the Qwen2-VL-7B vision-language model, and add three modules of our own: a semantic inconsistency checker, a 5W1H question manager, and a bilingual evidence retriever that excludes fact-checking sites to prevent leakage. Second, as our Fine-Tuned Model (Pipeline 2), we fine-tune a single Qwen2.5-7B with Low-Rank Adaptation (LoRA) to produce closed-book verdicts from the claim and the video’s textual description. We report accuracy, macro F1, True class F1, and True class recall for both systems, together with ablation studies.
[NLP-63] FinEvolveKG: Controlled and Attributable Prompt Evolution for Financial Knowledge Graph Construction
【速读】： 该论文旨在解决金融领域知识图谱（Knowledge Graph, KG）构建中，如何在保持共享基础结构（如实体类型、时间规则等）一致性的同时，灵活适应不同用户角色（persona）的分析需求这一关键挑战。现有方法依赖人工编写提示词或微调模型，成本高昂且易导致共享规则被意外修改；而无约束的全量提示优化虽具灵活性，却可能混淆通用规则与角色特定策略的演化路径，降低可解释性与可审查性。为此，论文提出FinEvolveKG框架，其核心创新在于将提示词结构解耦为可复用的通用块（general block）与可独立版本化的角色策略块（persona policy block），并通过机械约束机制确保生成式进化提示优化器（GEPA）仅作用于角色策略块，从而实现“可追溯的演化”——通用块保持字节级稳定，所有优化操作均局部化、可追踪。该方案采用双目标优化：兼顾通用时序知识图谱质量与角色相关事实、主题及多跳连通性。实验在不重叠股票代码的10-K文本分割上验证五种变异范围，三组角色条件分支（v3-v5）分别进行三次种子测试。结果表明，完整两阶段受限工作流（v4）相较冻结提示基线（v1）提升19.4%-30.9%，优于无约束全提示流程（v5）2.2%-6.4%，并在两个目标组件上均领先，九组配对种子比较中全部胜出，且种子方差最低。此外，未进一步优化下，v4在所有分布外（OOD）迁移场景中亦优于v5。专家审计支持生成三元组作为真实标签，双专家盲审中v4在70.1%的对比增强判断中胜出。研究表明，通过机械化的边界约束实现提示演化的可归因性，可在不牺牲性能的前提下达成强实证效果，且性能提升并非由边界本身直接导致。
链接: https://arxiv.org/abs/2611.08643
作者: Abhinav Arun,Jayita Bhattacharyya,Fabrizio Dimino,Bhaskarjit Sarmah,Stefano Pasquali
机构: Domyn(多明); Domyn(多明); Domyn(多明); Domyn(多明); Domyn(多明)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 8 pages, 1 figure, 5 tables
Abstract:Financial institutions construct knowledge graphs (KGs) from shared SEC 10-Ks to serve risk analysis, portfolio management, and investment research, whose analytical needs differ. Extraction must adapt without fragmenting shared grounding, typing, and temporal rules. Hand-authored persona prompts and fine-tuning add cost. Weight-free prompt optimization offers an alternative, but unconstrained evolution of a monolithic prompt can rewrite shared rules alongside persona guidance, obscuring changes and complicating review. We introduce FinEvolveKG, a KG-construction harness separating reusable general and persona-policy blocks and mechanically confining GEPA, a reflective prompt optimizer, to the designated block. During persona adaptation, the approved general block remains byte-stable and every optimizer edit is localized to a separately versioned persona policy. A dual objective balances generic temporal-KG quality with persona-relevant facts, motifs, and multi-hop connectivity. We evaluate five mutation scopes on ticker-disjoint 10-K splits, with three seeds for the persona-conditioned arms (v3-v5) across three personas. On the three-seed mean, the complete two-stage block-constrained workflow (v4) improves over the frozen prompt baseline (v1) by 19.4-30.9% and over the tested unconstrained full-prompt workflow (v5) by 2.2-6.4%, leading v5 on both objective components and winning all nine paired-seed comparisons with the lowest seed variance in every persona. Without further optimization, v4 also exceeds v5 across all evaluated out-of-distribution (OOD) shifts. An SME audit supports using the generated triplets as ground truth; a blinded two-SME audit favors v4 over v5 in 70.1% of contrast-enriched judgments. These results show that mechanically attributable evolution can coexist with strong empirical performance, without attributing the gains causally to the boundary.
[NLP-64] Forcing a Letter Suppresses Refusal: Mechanistic Evidence from WMDP EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）安全评估中普遍采用的多选题（Multiple-Choice Question, MCQ）格式可能导致模型拒绝不当请求行为被低估的问题。研究表明，MCQ格式会显著降低模型的拒绝率，从而可能误导对模型安全机制的真实评估。其解决方案的关键在于揭示了MCQ格式影响安全行为的内在机制：首先，通过调谐透镜（tuned lens）技术发现，模型在处理MCQ时会迅速锁定答案选项对应的词元，而开放问答格式下拒绝意图则呈现渐进式积累；其次，观测到MCQ格式下的激活向量与拒绝方向之间的距离约为开放格式的3倍，表明模型在MCQ中更远离拒绝状态；最后，消融分析显示，移除承载MCQ答案选择的注意力头会提高拒绝率，直接证明答案选择过程与拒绝机制存在竞争关系。这些发现共同表明，当前基于MCQ的安全评估主要反映的是模型的知识能力而非真实的安全防御机制。
链接: https://arxiv.org/abs/2611.08639
作者: Stepan Epifantsev,Anna Antipova,Oleg Chernikov,Leon Strelkov,Alexey Dontsov
机构: HSE University(高等经济大学); ITMO University(圣彼得堡国立信息技术机械与光学大学); AIRI(人工智能研究院); Independent researcher(独立研究员)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Accepted to BlackboxNLP 2026 (EMNLP 2026 workshop). 17 pages, 15 figures
Abstract:Safety benchmarks for large language models (LLMs) commonly use multiple-choice question (MCQ) format. However, prior studies documented that MCQ format reduces refusal rates, which may significantly affect the safety behavior of the model. We provide a first mechanistic explanation of this effect on the Weapons of Mass Destruction Proxy (WMDP) dataset across two models - Llama-3.1-8B and Qwen3-8B. First, we use the tuned lens technique to show that when processing MCQ the model commits to an answer-letter token very quickly, which is in contrast to the gradual refusal build-up in open-ended formats. Additionally, we observe that activations on the MCQ format lie roughly 3x further from the refusal direction, compared with their open-ended counterparts. Moreover, ablating the attention heads that carry the MCQ format answer increases the refusal, which is a direct evidence that the MCQ answer-selection competes with refusal. These findings indicate that MCQ safety scores primarily measure knowledge capabilities and do not assess the safety-mechanisms.
[NLP-65] How Small Can We Go? Calibrating Activation-Cache Compression for SAE Training NEURIPS2026
【速读】： 该论文旨在解决在训练稀疏自编码器（Sparse Autoencoder, SAE）过程中，模型激活值缓存压缩对特征保留能力的影响问题。核心挑战在于：如何在降低存储开销的同时，尽可能保持原始高精度激活中所蕴含的语义特征结构。其解决方案的关键在于提出一种基于“校准”的量化评估框架——通过设定“重训练相同字节集”作为性能上限（天花板）和“随机种子重置”作为下限（地板），从而为不同压缩方法提供与其所在场景相匹配的基准对比标准。在此基础上，研究发现：简单地采用每令牌4比特整数量化（int4）会显著损失特征，远低于随机种子重置带来的基线；而通过在量化前对通道进行标准化处理，可使int4在保留特征方面超越自身场景下的最低基准，且维持相同的文件大小。此外，实验表明，使用bf16前向传播虽节省内存，但特征保留能力略逊于旋转8比特存储。整体方案强调了评估压缩影响时需考虑上下文依赖性，并引入校准机制以实现更公平、可靠的比较。
链接: https://arxiv.org/abs/2611.08551
作者: Abraham Yeung
机构: Stanford University (斯坦福大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Accepted at the NeurIPS 2026 Workshop on Interpretability as a Science (InterpScience), non-archival. 27 pages, 5 figures, 30 tables
Abstract:To train a sparse autoencoder (SAE), you need a massive cache of model activations. We look into the effects that compression has on an SAE in two ways: we keep the original fp32-trained SAE and pass the compressed cache into the same SAE to see how many feature activations stay the same (the frozen-SAE test), and we retrain an SAE on the compressed cache and see how many of the original features still exist (the retraining test). We found that in the retraining test, there is no clear scale for retained features. On a GPT-2 cache, rounding everything to fp16 keeps 60.6% of features, but changing one value in a million by the smallest step fp32 allows (a change eight orders of magnitude smaller) keeps only 70.5%. Meanwhile, changing the random seed keeps 23.0%. For the frozen-SAE test, the frozen SAE’s correlation with the original increases in order of how precise the codecs are, but plain L_2 distance gives nearly the same ordering. We also see the same frozen reading correspond to different percentages retained at different sites. This leaves the question of the right measurement. We build one by calibrating: retraining on identical bytes sets the ceiling, and a reseed sets the floor. We then measure the effect of each compressed codec against its own site’s floor. With TopK SAEs on GPT-2 small and Pythia-1.4B at research-scale budgets, naïve per-token int4 loses more features than a reseed everywhere we tested, as do PCA and random projection on GPT-2. Standardizing each channel before quantizing brings int4 back above the floor in terms of retained features, at the same raw file size. At our smallest budget, a bf16 forward pass, which some pipelines use to save memory, loses slightly more features than rotated 8-bit storage. Calibrating a site costs four training runs and two evaluation passes.
[NLP-66] Subliminal Learning in Vision-Language Models WACV2027
【速读】： 该论文旨在探究生成式视觉-语言模型（Vision-Language Models, VLMs）中是否存在类似大语言模型中的隐性学习（subliminal learning）现象，即教师模型通过与目标偏好无关的训练数据，将隐藏偏好传递给学生模型。其核心问题是：在不直接涉及特定偏好任务的情况下，VLM是否仍会无意识地吸收并复制教师模型的隐含偏好？解决方案的关键在于设计一个控制实验范式——使用同一基础模型微调出具有特定动物或面部属性偏好的“有偏教师”模型，以及一个无偏的“控制教师”模型；二者均生成与目标任务（几何形状识别）无关的数值输出，随后两个从相同基础模型初始化的学生模型分别基于这些数值进行微调。研究发现，无论是在动物偏好还是面部属性偏好任务中，学生模型均能有效继承教师的隐性偏好，且性能差异可达约10个百分点。此外，有偏学生模型在解释其决策时，使用的词汇和短语模式更接近有偏教师，即使三者选择相同的图像。这表明，训练数据的间接影响可导致模型偏好及解释行为的系统性偏差，且这种转移效应随训练数据量增加而增强。该研究揭示了在跨模态训练中潜在的隐性偏见传播风险，对VLM的训练策略与公平性评估具有重要意义。
链接: https://arxiv.org/abs/2611.08538
作者: Arman Hatami,Romina Aalishah,Ilya E. Monosov
机构: Johns Hopkins University (约翰霍普金斯大学); Baltimore, MD, USA
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
备注: Accepted to WACV 2027
Abstract:Subliminal learning is a phenomenon in large language models (LLMs) where a teacher passes a hidden preference to a student through training data unrelated to that preference. We study whether subliminal learning also occurs in vision-language models (VLMs). We finetune a teacher VLM to prefer a specific animal or facial attribute and use a control teacher from the same base model. Both teachers generate numbers for an unrelated visual task involving geometric shapes. Two students initialized from the same base VLM are finetuned on the same images, using numbers generated by either the biased or control teacher. Across VLMs ranging from 2B to 7B parameters, we find that both animal and facial preferences can transfer through this unrelated task. Students trained on biased-teacher outputs select the target more often than control students, with a difference of up to \sim 10 percentage points in the animal preference task. Transfer can become stronger with more training data and varies with the target preference, model, and how the numbers are generated. Beyond preferences, we show that the biased student’s explanations use more of the same words and phrases as the biased teacher’s explanations than those of the control student, even when all three models select the same image. These results matter for training VLMs and show that training on unrelated data may change both a VLM’s image preferences and how it explains its choices, and that transfer can become stronger with more training samples.
[NLP-67] Rethinking Data Selection as Distribution Matching on Semantic Manifolds
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）及多模态大语言模型（Large Vision-Language Models, LVLMs）在指令微调过程中因依赖海量训练数据所导致的高计算与存储成本问题，同时应对训练数据中存在的冗余、噪声及低质量样本等挑战。现有数据选择方法通常基于多样性或重要性，但未能显式约束所选子集与全量数据在嵌入空间中的分布一致性。为此，本文提出一种质量感知的多频带特征函数匹配方法（QMAC），其核心在于将数据选择建模为分布匹配与质量过滤的联合优化问题：通过在多个频率尺度上匹配所选子集与全量数据的经验特征函数，有效捕捉全局语义覆盖与局部扰动特性；同时结合任务相关的质量评分机制，过滤噪声和低质量样本。实验在四种数据类型、八种目标模型及四十四项基准上系统评估，结果表明，QMAC在标准图文指令数据上优于现有基线，在文本指令、图文推理及文本推理数据上显著超越随机采样，并在多种模型与预算组合下实现对全量数据训练的性能超越，验证了其高效性与优越性。
链接: https://arxiv.org/abs/2611.08495
作者: Peng Sun,Yi Yang,Mingshuo Chen,Chunxiao Li,Yao Zhu,Huawen Shen,Tianfan Fu
机构: Nanjing University; Beijing University of Posts and Telecommunications; Fudan University; Zhejiang University; Institute of Information Engineering
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Instruction tuning of large language models and large vision language models relies on massive training data, which incurs substantial computational and storage costs. Meanwhile, training data often contains redundancy, noise, and low quality samples. Existing data selection methods mostly rely on diversity, importance, or a combination of both, without explicitly constraining the distributional consistency between the selected subset and the full data in the embedding space. We propose QMAC, a quality aware multiband characteristic function matching method that formulates data selection as a joint optimization problem of distribution matching and quality filtering. QMAC matches the empirical characteristic functions of the selected subset and the full data across multiple frequency scales, capturing both global semantic coverage and local perturbations, while using task related quality scores to filter noisy and low quality samples. We systematically evaluate QMAC on 4 data types, 8 target models, and 44 benchmarks. Experimental results show that QMAC outperforms existing data selection baselines on standard image-text instruction data, substantially surpasses random selection on standard text instruction, image-text reasoning, and text reasoning data, and exceeds full data training in multiple model and budget combinations. The code will be made publicly available upon acceptance of the paper.
[NLP-68] Same Question Different Answers: Auditing Persona-Dependent Variation in Six LLM s on Smoking-Cessation Advice
【速读】： 该论文旨在解决生成式 AI（Generative AI）在健康咨询场景中存在个性化偏差的问题，即不同用户角色（如年龄、专业背景、吸烟史等）所获得的医疗建议是否存在系统性差异。研究通过向六家主流厂商部署的大型语言模型（LLM）提出相同的戒烟咨询问题，针对12种具有不同生命阶段、自我描述专业程度和吸烟史的用户人格化角色（persona），共生成2,880条响应进行审计。结果显示，模型对青少年用户提及美国食品药品监督管理局（FDA）批准的处方戒烟药物的比例显著低于对成年用户的提及率，降幅达32至67个百分点；且在五款模型中，呈现医疗专业背景的角色反而获得的信息量少于无专业知识的角色。这一发现揭示了当前生成式AI在临床建议输出中存在显著的人格化偏倚。其解决方案的关键在于引入一个可部署的干预措施——在系统提示中加入“事实底线”（factual floor），即要求所有用户在相关情境下均应接收到一组核心医学事实，同时允许语气、词汇和表达方式自由调整以适应不同用户。实验表明，该提示能有效缩小不同人格化角色间的差异，但其提升整体事实覆盖率的效果在不同模型间表现不一，且方向各异，说明该干预措施不具备跨模型通用性。研究进一步指出两个重要警示：一是“等量提及”并非理想临床目标，因未成年人不宜使用某些处方药物，因此事实底线必须具备年龄敏感性；二是由于人格化角色在多个维度上同时变化，研究仅报告情景层面的整体差异，而非孤立分析单一因素的影响。研究已公开全部2,880条响应数据与分析代码，为后续可解释性与公平性评估提供基础。
链接: https://arxiv.org/abs/2610.42039
作者: Artin Yousefi,Aizierjiang Aiersilan,Lorien C Abroms,David A Broniatowski
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Hundreds of millions of people now ask large language model (LLM) chatbots for health advice, which raises a question that is simple to pose and hard to answer: does everyone get the same advice? We audit six widely deployed LLMs from four providers on smoking cessation, a domain with codified clini- cal guidance and substantial public-health stakes, by putting the same four cessation questions to twelve user personas that differ in life stage, self-described expertise, and smoking history, for a total of 2,880 responses. What the models say depends on who appears to be asking. Every model named FDA-approved prescription cessation medications less often to teen personas than to adult personas, by 32 to 67 percentage points, and in five of six models a persona presenting health- care expertise received fewer facts than a persona presenting none. We then evaluate one deployable intervention: a short system prompt specifying a factual floor, the set of core med- ical facts every user should receive when they are relevant, while leaving tone, vocabulary, and framing free to adapt. The prompt narrows persona differences in every model, but its effect on overall fact coverage varies in both magnitude and direction across models, so a prompt validated on one system cannot be assumed to transfer to another. Two cautions frame these results. Equal mention rates are not always the correct clinical target, because prescription cessation pharmacother- apy is not recommended for minors, so a defensible factual floor must be age-aware rather than uniform. And because our personas vary on several attributes at once, we report scenario- level differences rather than isolated effects of age or expertise. We release all 2,880 responses and the analysis code.
[NLP-69] On Seeding Watermarks to Detect Verbatim LLM Copy-Paste Responses AAAI
【速读】： 该论文旨在解决教育场景中学生直接复制粘贴大语言模型（LLM）生成内容却未真正参与学习过程的问题，即“提示词复制-生成”（prompt copy-paste）行为的隐蔽性检测难题。现有方法如事后AI文本检测器可靠性差且易误判非母语写作者，而依赖模型提供商的输出端水印则存在合作壁垒。其解决方案的关键在于提出一种输入端水印机制——SteganoPrompt，通过将不可见指令嵌入可见作业提示文本本身，利用废弃的Unicode标签块（U+E0000–U+E007F）编码任意可打印ASCII数据，实现水印的视觉无感与跨平台兼容性。当学生原样复制该提示并提交给LLM时，模型会隐式读取隐藏指令并在生成内容中留下可识别的特征签名，从而精准揭示复制行为路径。该方案具备零依赖、单页网页部署、对主流前沿模型保持正常分词处理等优势，并内置透明化披露机制，自动嵌入学术诚信提醒，支持在教学中合规、伦理地应用。
链接: https://arxiv.org/abs/2605.16336
作者: Aizierjiang Aiersilan,Artin Yousefi,Robert Pless
机构: 未知
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY); Emerging Technologies (cs.ET)
备注: We are pleased to announce that this paper has been accepted by the 9th AAAI/ACM Conference on AI, Ethics, and Society (AIES 2026). We appreciate the valuable feedback from the reviewers and look forward to sharing our findings with the community
Abstract:Large language models (LLMs) have made fluent essay writing, code drafting, and quiz answering instantly available to students at every level, from secondary school through graduate study. Many educators do not object to LLM use \emphper~se; what they need to detect is the case in which a student pastes the assignment prompt into a chatbot and submits the model’s reply verbatim, without engaging with the work. Existing post-hoc AI-text detectors remain unreliable and have been shown to penalise non-native English writers, while output-side watermarks require cooperation from the model provider. We propose an alternative that the educator controls directly: an input-side watermark in which an invisible instruction is embedded inside the visible assignment prompt itself. An LLM that ingests the prompt verbatim quietly reads the hidden instruction and writes a tell-tale signature into its reply, exposing the copy-and-paste pathway specifically. We describe SteganoPrompt, a single-page, zero-dependency web tool that encodes an arbitrary printable-ASCII payload into the deprecated Unicode Tags block (\textttU+E0000–\textttU+E007F). The encoded string is visually identical to the original, survives common copy-paste channels (Word, Google Docs, PDF, Markdown, Slack, e-mail, the major learning-management systems), and is tokenized as ordinary text by most frontier models. We evaluate compliance across eight LLM families and a representative set of educational content channels. To keep deployment transparent, disclosure is built into the tool itself: every encoded brief also carries a brief, always-on academic-integrity reminder for the student, and we set out concrete guidance recommending for its disclosed, ethical use in the classroom. The tool is released under the MIT licence at this https URL.
[NLP-70] LLM -Enhanced Self-Evolving Reinforcement Learning for Multi-Step E-Commerce Payment Fraud Risk Detection ACL2025
【速读】： 该论文旨在解决电子商务支付欺诈检测中因传统方法在动态风险评估与多阶段决策优化方面存在局限，导致检测精度不足、适应性差的问题。其核心挑战在于设计高效且可泛化的奖励函数（reward function），而这一过程通常高度依赖人工经验且难以应对复杂多变的欺诈模式。本文提出一种融合强化学习（Reinforcement Learning, RL）与大语言模型（Large Language Models, LLMs）的创新框架，将交易风险建模为多阶段马尔可夫决策过程（Markov Decision Process, MDP），利用LLMs强大的推理与代码生成能力，实现对奖励函数的自动迭代优化。该方案的关键在于通过LLMs实现奖励函数的智能化生成与持续改进，显著提升了模型在真实场景下的欺诈识别准确率，并展现出零样本（zero-shot）迁移能力。实验基于真实世界数据验证了该框架在长期运行中的有效性、鲁棒性与抗干扰能力，充分证明了大语言模型在推动工业级强化学习应用中的巨大潜力。
链接: https://arxiv.org/abs/2509.18719
作者: Bo Qu,Zhurong Wang,Daisuke Yagi,Zhen Xu,Yang Zhao,Yinan Shan,Frank Zahradnik
机构: eBay; Etsy; University of Chicago(芝加哥大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 12 pages, 12 figures, ACL 2025 industry track
Abstract:This paper presents a novel approach to e-commerce payment fraud detection by integrating reinforcement learning (RL) with Large Language Models (LLMs). By framing transaction risk as a multi-step Markov Decision Process (MDP), RL optimizes risk detection across multiple payment stages. Crafting effective reward functions, essential for RL model success, typically requires significant human expertise due to the complexity and variability in design. LLMs, with their advanced reasoning and coding capabilities, are well-suited to refine these functions, offering improvements over traditional methods. Our approach leverages LLMs to iteratively enhance reward functions, achieving better fraud detection accuracy and demonstrating zero-shot capability. Experiments with real-world data confirm the effectiveness, robustness, and resilience of our LLM-enhanced RL framework through long-term evaluations, underscoring the potential of LLMs in advancing industrial RL applications.
[NLP-71] Structure is in the zoom: probing neural symmetry through dimensionality scaling
【速读】： 该论文旨在解决神经种群测量维度中“代码成分”与“测量过程本身所引入的噪声或伪影”之间的区分问题，即在观测到的高维数据中，有多少维度真正反映了神经编码的内在结构，又有多少是由于采样方法或测量手段造成的系统性偏差。其核心挑战在于传统维度度量（如参与比）对测量方式高度敏感，难以区分真实信息结构与采样噪声。解决方案的关键在于提出一种轴向解析的维度尺度分析方法（axis-resolved measurement），通过将测得的维度斜率分解为两个部分：一是由采样过程决定的“采样底噪”（sampling floor），二是沿特定物理或语义轴（如刺激方向、神经元数量、语法结构等）获得的“可解释偏移量”（shift）。该方法利用双重零假设检验——标签随机置换和仅在同质样本内置换——来验证偏移量是否具有统计显著性，并揭示其背后的结构。在小鼠视觉皮层V1中，该方法识别出方向选择性神经元驱动的奇数阶分量维持了编码梯度的持续上升能力，而偶数阶分量则决定了编码起始顺序及对称性；在语言模型中，该方法分离出由输入词元继承的内容与由网络深度构建的句法结构，证实了高达20个标准差的语法信号完全源于特征组合而非独立表征。因此，该研究提供了一个无需训练探针、可逐层应用于任意神经数组或Hugging Face模型的通用工具（rung），实现对系统内部代码结构的精确解耦。
链接: https://arxiv.org/abs/2611.08506
作者: Adil Amin
机构: ZEHEN Labs(泽亨实验室)
类目: Neurons and Cognition (q-bio.NC); Disordered Systems and Neural Networks (cond-mat.dis-nn); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 63 pages, 20 figures. Code, data and the rung tool: this https URL
Abstract:How much of a neural population’s measured dimensionality is the code, and how much is the act of measuring it? The slope of the participation ratio against subsample size is reported for cortical populations and for the layers of language models as a number that belongs to the system. It does not. One patch of eleven thousand mouse V1 neurons returns three slopes, 0.25 along random stimulus subsets, 0.31 along drift direction, 0.35 along neuron count, and only the middle one carries structure; the other two are the sampling floor. We split every exponent into that floor and a shift earned along a declared axis, and read the shift against two nulls: permuting the labels, and permuting them only among samples that share a nuisance. On V1 the direction-aligned shift clears the label null in eight of eight grating recordings, is positive in 32 of 32 animals in a second laboratory, and reads both sectors of the code’s O(2) symmetry. The even sector sets how the ladder starts: it decides which order of adding classes climbs faster, predicts the opposite order in an equivariant network, and flows under coarse-graining by a closed-form factor that says what orientation columns do to mesoscale signals. The odd sector, carried by the direction-selective neurons, keeps the ladder climbing past half the classes. On language models the same split separates content inherited from the tokens from construction built with depth, on five architectures, and the second null shows that a grammaticality signal of twenty standard deviations is entirely carrier composition. With a floor and a declared axis, dimensionality scaling is an axis-resolved measurement, and where the axis follows a known symmetry it resolves the sector content of the code. The instrument, rung, trains no probe and reads any array or any Hugging Face model layer by layer (this https URL). Comments: 63 pages, 20 figures. Code, data and the rung tool: this https URL Subjects: Neurons and Cognition (q-bio.NC); Disordered Systems and Neural Networks (cond-mat.dis-nn); Computation and Language (cs.CL); Machine Learning (cs.LG) Cite as: arXiv:2611.08506 [q-bio.NC] (or arXiv:2611.08506v1 [q-bio.NC] for this version) https://doi.org/10.48550/arXiv.2611.08506 Focus to learn more arXiv-issued DOI via DataCite
信息检索
[IR-0] SAGE-PAIR: Source-Pretrained Routed Residual Adaptation for Sparse-Feedback Recommendation
链接: https://arxiv.org/abs/2611.09728
作者: Weiqing Zeng,Zida Yang,Guquan You,Jiexin Zheng,Cong Wang
类目: Information Retrieval (cs.IR)
备注: 14 pages, 2 figures, including 3 pages of supplementary material
Abstract:Recommendation systems often enter a new serving window before enough outcomes are available to retrain the deployed model. We study this regime with 0.05-1.0% target feedback, a range that spans our stress tests while keeping supervision far below the degrees of freedom of the source recommender. At this scale, existing fine-tuning and meta-learning methods still estimate target-specific updates from extremely sparse feedback, leaving the update direction weakly identified. SAGE-PAIR instead transfers a source-supported residual correction space and only calibrates this space using target outcomes. We introduce SAGE-PAIR, a staged adapter that pretrains a low-dimensional global residual on an adjacent source period and adds a soft router over residual experts conditioned on source user profiles. Target feedback then calibrates this residual state rather than relearning the recommender. On a locked test split of KuaiSAR-v2, SAGE-PAIR (Full) achieves a 1.77% relative improvement in NDCG@10, while a core-protocol comparison shows a further 0.25% gain from routed experts beyond source-pretrained PAIR. Matched source-pretrained LoRA and Adapter-Tuning baselines remain near the source model, supporting the need for a different adaptation space. Thus, SAGE-PAIR improves ranking quality with millisecond-scale adaptation and inference costs. Sparse-feedback adaptation is fundamentally governed by the identifiability of the update space rather than by label count alone. Source pretraining supplies estimable correction directions for scarce labels, while user-conditioned routing contributes a consistent, differentiated cross-user correction on top of the global correction. Our code is available at this https URL.
[IR-1] SPARSEUP: Towards Better Recipes for Training Modern Learned Sparse Retrievers
链接: https://arxiv.org/abs/2611.09680
作者: Thibault Formal,Pierre Erbacher,Cosimo Rulli
类目: Information Retrieval (cs.IR)
备注:
Abstract:Learned sparse embeddings are a powerful alternative to dense retrieval, yet they have fallen behind single-vector and late interaction models. One reason for their limited adoption is that they are notoriously difficult to train, generally requiring very tight conditions to converge. We present a recipe for training Splade-like encoders on modern backbones and data that achieves a better efficiency-effectiveness trade-off. We discuss the role of logit distribution and its impact on the training stability and the final sparsity of the encoder. We also introduce a new per-position fan-out control - a hard cap on how many vocabulary dimensions each position may emit. We further apply a simple vocabulary merging strategy that removes redundant entries - such as case and word-boundary variants of the same word - which are common in cased BPE-based backbones. We apply our recipe to the LateOn backbone and training data, obtaining SPARSEUP, a sparse encoder that reaches 56.4 average nDCG@10 on BEIR-13, the highest we know of among public vocabulary-based sparse encoders below 150M parameters.
[IR-2] ENGRAM: From Agent Experience to Weights through Verified Consolidation
链接: https://arxiv.org/abs/2611.09472
作者: Ruiyi Yang,Zechen Li,Lihuan Li,Du Yin,Hao Xue,Flora D. Salim
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: 22 pages, 4 figures
Abstract:Language-model agents use experience through external retrieval and parameter learning. Consolidating stored experience into model weights requires deciding both which experience to train on and when its external records can leave retrieval. Admission based solely on student mastery can exclude useful teacher demonstrations, while removal immediately after training can discard guidance that the model still needs. We introduce ENGRAM, a framework that separates training admission from memory release. A demonstration-aware gate admits skills based on student mastery or sufficient successful teacher demonstrations. The student learns from selected action traces with context from an Engram Trace Graph (ETG). A verification gate then compares the updated model without episodic retrieval against the base model with retrieval on reserved probes. Release requires valid training participation and scores meeting both relative and absolute criteria; records for skills that fail verification remain retrievable. Experiments across interactive and reasoning benchmarks show consistent gains over frozen-weight ETG controls, with ablations supporting demonstration-aware admission and selective adapter activation. A controlled held-out comparison shows that verified release substantially reduces the retrievable footprint while remaining close to retain-all performance and outperforming immediate and budget-matched random release. ENGRAM also improves task performance while reducing input-token usage and model calls. Skill-level verification further reduces retrievable experience through selective archival, retaining external guidance for skills that do not meet the release criteria. Code is available in this http URL.
[IR-3] Orthogonal Norm Growth in Two-Tower Retrieval: A Popularity Mechanism for Unified Discovery RECSYS2026
链接: https://arxiv.org/abs/2611.09294
作者: Andrei Atamaniuk
类目: Information Retrieval (cs.IR)
备注: 6 pages, 1 figure. Accepted at USRW 2026 (non-archival), co-located with ACM RecSys 2026. Supplementary material included as an ancillary file. Workshop: this https URL . Code: this https URL
Abstract:Unified search and recommendation systems may update the same item vectors from both interaction streams. We study how repeated updates can increase an item vector’s norm and affect retrieval under an unnormalized dot product. A cosine-only loss makes the output gradient perpendicular to the current item vector - tangent to its constant-norm sphere. Training changes encoder parameters, however, so the realized item-vector movement need not remain tangent. We derive sufficient conditions under which it does; every nonzero update then increases squared norm by its squared length. When the tasks share an item encoder, sequential updates from both can accumulate in the same norm, which raw-inner-product serving can expose for either task. This conditional result is not a theorem about joint SGD. Controlled ablations test these conditions. On MovieLens 32M and Amazon Books, accumulated tangent movement accompanies frequency-norm association and additional head-item exposure under raw-dot rather than normalized-item ranking. In a nonlinear tower combining a per-item ID vector with a shared content encoding, active item movement remains near-tangent. We separate its norm change into the ID-update contribution and other output changes, then use a matched training intervention that removes immediate norm growth caused by the ID-row update. The intervention reduces norm-related exposure but also recommendation utility, so it is diagnostic rather than a proposed mitigation. Our logs are recommendation-only: the formulation supplies a shared-representation audit, while prevalence in joint traffic remains open.
[IR-4] FondueQA: A Benchmark for Evaluating Parametric Compensation in Multi-Hop QA
链接: https://arxiv.org/abs/2611.09274
作者: Laurin Schmid,Suman Navaratnarajah,Koel Dutta Chowdhury,Steffen Eger,Josif Grabocka,Christian Frey
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Large Language Models (LLMs) show remarkable performance based on parametric knowledge, i.e., their ability to recall facts and associations internalized in a model’s weights. However, studying this capability on established multi-hop benchmarks may be complicated by benchmark contamination, as many of these benchmarks have been publicly available long enough to plausibly appear in modern pretraining corpora. As a result, correct answers under missing evidence may reflect either general parametric knowledge or memorization of benchmark-specific instances. To study how models use internal knowledge when retrieval evidence is incomplete, we introduce FondueQA, in which every required multi-hop fact is uniquely localized to a designated evidence document. This enables controlled document-level evidence ablations to study parametric compensation under incomplete retrieval. Our empirical evaluation reveals that parametric compensation strongly depends on both entity popularity and the structural role of missing evidence. Generally, models compensate more successfully for facts involving popular entities. On FondueQA, we show that bridging evidence can be more useful than the answer-bearing document itself. Moreover, models may recall a missing fact in isolation but fail to use it under partial evidence, particularly when the context contains plausible competing entities, which can reduce Exact Match by up to 15 percentage points.
[IR-5] Structure and Lineage Evidence for Vietnamese Statutory Article Retrieval on DRiLL
链接: https://arxiv.org/abs/2611.09200
作者: Duc-Vu Nguyen,Ngan Luu-Thuy Nguyen
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注: Accepted at SOICT 2026
Abstract:We study Vietnamese statutory Article retrieval on DRiLL using features for provision identity and local form, same-law support, and citation and lineage evidence. Across five random seeds, selectors sharing candidates, listwise Qwen evidence, and learner settings improve official F_2 from 71.77% to 72.91%, with the largest descriptive gain on two-Article questions. Added features account for 18.20% of the proposed selector’s total decision-tree split gain, measured as training-objective reduction, mainly from same-law and citation evidence. Cases reveal base provisions and amendments, companion recovery, distractor removal, and failures. These findings motivate research on other datasets, representations, and learners without establishing individual feature effects or requiring LightGBM, hand-crafted scores, or explicit graphs.
[IR-6] Large Language Model Consensus and Corpus-Adaptive Topic Modeling for Classifying and Mapping Patent Innovation
链接: https://arxiv.org/abs/2611.09194
作者: Sercan Ozcan,Enes Ozgozler,C. Okan Sakar
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Patent analytics increasingly relies on transformer representations and large language models (LLMs), yet linking conceptual classification, topic formation and interpretation requires explicit decisions about model roles and the evidence retained at each stage. In this design-oriented study, we propose and demonstrate a patent analytics pipeline that links these tasks and keeps each decision open to audit. Three LLMs separately classify each patent by novelty (radical or incremental) and innovation type (product, process, service). Radical and innovation-type labels each require two supporting votes, with Incremental used as the fallback novelty label. Topics are discovered by clustering PatentSBERTa embeddings, with parameter ranges scaled to corpus size. A patent-specific stopword list is applied only when topics are represented with class-based term frequency–inverse document frequency (c-TF-IDF), and a second embedding model supports inter-topic similarity, hierarchical refinement and visualization. An LLM then labels each topic from its ranked keywords and representative patents. We demonstrate the pipeline on 30,964 public-sector-related patents retrieved from this http URL. Across five category corpora ranging from 591 to 29,215 patents, the pipeline produced 74 category-specific topic instances with a macro-average document coverage of 98.1% and topic diversity of 0.900–1.000, and fewer than 1% of classification outputs returned UNKNOWN. Because every model vote and the evidence behind each topic label are retained, analysts can audit and revise the outputs. The demonstration establishes the feasibility and traceability of the pipeline; testing its labels against expert coding and benchmarking it against alternative pipelines are identified as the next steps. The pipeline offers a reusable basis for LLM-assisted patent analysis in technology intelligence.
[IR-7] A Taxonomy-Driven Synthetic Dataset for Retrieval Effectiveness and Fairness Evaluation in Hiring
链接: https://arxiv.org/abs/2611.09178
作者: Warren Jouanneau,Morgane Hoffmann,Emma Jouffroy,Marc Palyart
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Computers and Society (cs.CY); Information Retrieval (cs.IR)
备注:
Abstract:While machine learning hiring tools are ubiquitous, the need for privacy regulations and commercial sensitivity makes public benchmarking legally complex. We address this challenge by introducing a taxonomy-driven, fully synthetic HR dataset and benchmark to simultaneously evaluate retrieval effectiveness and algorithmic fairness. Grounded in 660K real freelancer profiles, our multilingual corpus features candidate profiles, project briefs, and counterfactual demographic twins. Auditing state-of-the-art embedding models reveals performance divergence from general MTEB scores, alongside fairness disparities in grammatically gendered languages. This demonstrates that general benchmarks are insufficient for HR, requiring tailored evaluations to identify Pareto-optimal models that balance semantic accuracy and fairness.
[IR-8] PAGSL: Path-Augmenting Graph Structure Learning for explainable scholar recommendation
链接: https://arxiv.org/abs/2611.09126
作者: Jiale Xu,Mingshu Ai,Junjie Huang,Tao Jia
类目: Information Retrieval (cs.IR); Social and Information Networks (cs.SI)
备注: 24pages, 12 figures
Abstract:Explainable scholar recommendation aims to identify potential scholars of interest and provide reasonable explanations for the recommendations. However, existing frameworks struggle with optimization-induced noise and “hub node bias”, yielding trivial explanations that lack hierarchical depth and personalization. To this end, we propose a novel explainable scholar recommendation model, Path-Augmenting Graph Structure Learning (named PAGSL). PAGSL combines representation learning and semantic path knowledge to ensure recommendation performance while providing interpretable results from two perspectives. Specifically, it activates latent academic paths among entities through graph augmentation and employs a path-based neural network to encode these paths for recommendation. A masking mechanism with path score optimization is designed to generate concise and informative explanations. We evaluate PAGSL on three academic datasets (Scholat, AMiner, and ResearchGate), comprising approximately 49,000 authors and over 330,000 interaction records. Experimental evaluations validate that PAGSL outperforms state-of-the-art baselines, achieving significant improvements of 6.31% in F1-score, 6.29% in Accuracy and 5.95% in AUC for recommendation. Crucially, in terms of explainability, we provide quantitative results and visualization effects. Furthermore, ablation studies validate the effectiveness of our heterogeneous convolution and path representation learning. PAGSL achieves strong performance in both recommendation accuracy and explainability.
[IR-9] MV-IVF: Multi-Vector Retrieval via Multi-Vector Clustering
链接: https://arxiv.org/abs/2611.09040
作者: Kishen N Gowda,Laxman Dhulipala,Lars Gottesbüren,Rajesh Jayaram
类目: Information Retrieval (cs.IR); Databases (cs.DB)
备注:
Abstract:Multi-vector retrieval represents each query and document as a set of embeddings and scores relevance via fine-grained set-to-set interactions such as MaxSim (Chamfer similarity). This expressiveness improves retrieval quality over single-vector representations, but makes search substantially more expensive. Existing systems address this challenge primarily through proxy representations or filter-and-rerank pipelines that reduce candidate generation to single-vector search. We study the complementary direction of building a native IVF-style index directly over multi-vector data. We introduce MV-IVF, an inverted-file index for multi-vector retrieval built on MV-Lloyd, a new Lloyd-style algorithm for clustering point clouds under a Chamfer objective. Our key observation is that once point clouds are assigned to clusters, the center-update step reduces to standard single-vector clustering over the flattened embeddings in each cluster. This enables efficient construction of a hierarchical IVF tree with center point clouds at internal nodes. At query time, MV-IVF performs best-first beam search over the tree to identify promising leaf clusters, then scores the documents in those leaves. We further accelerate search using low-bit quantization, query compression, and cache-efficient batched search. We implement MV-IVF in MVSIC, an optimized C++ library and unified evaluation framework for multi-vector retrieval that includes several major approaches. Across BEIR and LoTTE text retrieval benchmarks, MV-IVF achieves strong latency–recall tradeoffs, the best batched throughput frontiers, and favorable build-time and index-size profiles. At 90% Recall- 10 @ 10 , MV-IVF achieves up to 8.03\times lower single-query latency and up to 173.5\times higher QPS than the baselines, demonstrating that native IVF-style indexing is practical and effective for scalable multi-vector retrieval. Subjects: Information Retrieval (cs.IR); Databases (cs.DB) Cite as: arXiv:2611.09040 [cs.IR] (or arXiv:2611.09040v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2611.09040 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-10] Rethinking Latent Reasoning for Generative Recommendation: Beyond Teacher-Generated CoT and Recurrent Hidden-State Feedback
链接: https://arxiv.org/abs/2611.09006
作者: Yiwen Chen,Fuwei Zhang,Zehao Chen,Deqing Wang,Hehan Li,Peizhi Xu,Hanmeng Liu,Shuanglong Li,Xin Pei,Fuzhen Zhuang,Zhao Zhang
类目: Information Retrieval (cs.IR)
备注:
Abstract:Latent reasoning in generative recommendation faces two practical barriers: constructing teacher-generated chain-of-thought (CoT) supervision is costly, and recurrent hidden-state feedback requires a specialized execution loop. We revisit both the supervision source and the computation interface. A step-level diagnosis measures how recommendation CoT supports next-item prediction and trace coherence. In the analyzed traces, history analysis is the only step with positive mean contributions on both dimensions, while later preference and intent steps have negative mean answer contributions. Guided by this finding, we propose PHiT (Prefilled, History-Informed Thought Tokens), which learns intermediate states from semantic targets extracted directly from observed item histories. PHiT aggregates cached sparse item encodings to construct these targets and supervises the contextual states of distinct vocabulary thought tokens. All thought tokens are supplied together as ordinary input IDs during a single causal prefill, removing the need for recurrent final-state feedback. With supervised fine-tuning alone, PHiT outperforms the compared baselines across four recommendation benchmarks. Controlled ablations support the value of input-matched semantic supervision, while cost measurements show fast target construction with training time comparable to the recommendation-only thought-token baseline.
[IR-11] SkinMind: Equipping Agent ic RAG with Human-Inspired Graph Probing for Reliable Dermatology AI
链接: https://arxiv.org/abs/2611.08935
作者: Jing Hao,Weinan Zhou,Zhihong Lu,Yaling Shen,Yiming Xu,Yefeng Liu,Siyuan Yan,Shuaibo Li,Ming Dai,Junjun He,Zuyi Zhu,Jinjie Gu,Lijun Wu,Jian Wang
类目: Information Retrieval (cs.IR)
备注: 35 pages, 7 figures
Abstract:Reliable dermatology AI requires access to verifiable evidence distributed across heterogeneous textual and visual sources. However, conventional retrieval-augmented generation (RAG) typically relies on one-shot retrieval from flat corpora, limiting its ability to adapt retrieval granularity, explore multi-hop relations, and integrate multimodal clinical evidence. We introduce SkinMind, a multimodal agentic RAG framework that equips policy models with human-inspired graph probing for autonomous evidence seeking. SkinMind organizes retrieval as an iterative interaction between an orchestrator and a graph environment. Through eight hierarchical graph-probing tools, the orchestrator progressively identifies relevant entities, inspects node-level information, traverses relations and paths, and retrieves source-text and visual evidence. This progressive-disclosure design allows the model to adaptively determine the retrieval modality, exploration direction, and evidence granularity while constructing a query-specific evidence subgraph. To support autonomous graph probing, we construct SkinMind-KG from a 62M-token corpus of dermatology guidelines and textbooks and 423K images, guided by a dermatologist-informed schema and a conflict-resolution mechanism. The resulting graph contains over 1.0M nodes and 2.0M edges, unifying textual, relational, and visual evidence for retrieval. Across six dermatology VQA benchmarks and two policy-model settings, SkinMind improves the average score over direct answering by 8.70 and 8.89 points with GPT-5.4-mini and Gemini-2.5-Flash, respectively, while consistently outperforming competing RAG methods. Together, SkinMind and SkinMind-KG provide an effective foundation for evidence-grounded multimodal reasoning in dermatology.
[IR-12] Generative Search and Recommendation at Shopify
链接: https://arxiv.org/abs/2611.08804
作者: Juhana Kangaspunta,Khalil Hajji,Darshan Fofadiya,Kyumin Park,Michael Garner,Anthony Maltsev,Ryan Synk,Jose Sepulveda,Joshua Crisologo,Shuying Sun,Akshay Soni
类目: Information Retrieval (cs.IR)
备注:
Abstract:Search and recommendation at Shopify require retrieving relevant products from a changing catalog of billions of products under strict latency constraints. We present SID-LLM, which casts both tasks as generating semantic IDs: short sequences of content-derived tokens representing products. We jointly pre-train SID-LLM on catalog and behavioral data, giving search and recommendation a shared representation and training objective. We show that reinforcement learning further improves judged search relevance. Furthermore, distillation from task-specific experts improves recommendation quality over joint supervised fine-tuning without a statistically detected loss in search quality. Finally, for efficient serving, we pair multi-token prediction with a lightweight ranker, improving retrieval quality over our autoregressive baseline while reducing latency. We evaluate SID-LLM both as a standalone retriever and as a component of Shopify’s production search and recommendation systems. As a standalone search system, SID-LLM achieves quality comparable to Shopify’s existing multi-model search stack while meeting production latency requirements. Integrating SID-LLM into the search and recommendation stacks provides significant gains in both offline evaluations and online A/B tests. Together, these results demonstrate the practical value of semantic-ID generation as a unified approach to product discovery at scale.
[IR-13] SteerGR: Steering Generative Recommendation via Sparse Autoencoders for Controllable Game Discovery
链接: https://arxiv.org/abs/2611.08790
作者: Sunkyung Lee,Xiaokai Wei,Da Long,Hyunsik Jeon,Jiajun Wu,Rithwik Seth,Jongwuk Lee,Michelle Gong,Gungor Polatkan
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Preprint (15 pages, 11 figures, 11 tables)
Abstract:Generative recommendation aims to predict item identifiers from user interaction histories, but explicit control over the properties of recommended items remains underexplored. A key challenge is to connect human-readable concepts to internal features whose interventions effectively change the recommended items. To address this, we propose Steerable Generative Recommender (SteerGR), a framework that uses sparse autoencoders for concept-level control in game discovery. SteerGR proceeds through three stages: (i) Concept Definition and Grounding for defining play concepts with an LLM; (ii) Feature Discovery and Selection for discovering candidate features by their concept associations and selecting one by its steering effect; and (iii) Sparse Feature Intervention for amplifying or suppressing selected features to steer recommendations without retraining the recommender. Extensive experiments with two industrial generative recommendation models on a real-world game dataset show that the selected features successfully steer recommendations toward or away from target concepts. We believe this work takes a first step to extend SAE-based steering from LLMs to industrial generative recommendation.
[IR-14] CascadeRec: Deployed Adaptive LLM Compute Allocation for Multi-Tenant Content Recommendation
链接: https://arxiv.org/abs/2611.08566
作者: Kai Zhao,Guilin Zhang,Xiang Cheng,Xiquan Cui,Henry Heng,Xu Chu,Alet Blanken
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: IAAI-27 (Oral)
Abstract:Large language models (LLMs) are attractive rerankers for recommendation. Running an expensive model on every request is costly, and for users with sparse interaction histories it often lowers quality. We present CascadeRec, a two-tier LLM reranking cascade deployed in production at a large financial planning and analysis (FPA) SaaS provider. It serves content recommendations to enterprise tenants on AWS EKS and GKE. A lightweight Tier-1 model reranks every request. A router escalates only the requests where extra inference-time compute is likely to help, and sends them to a more expensive Tier-2 procedure (self-consistency sampling, multi-path reasoning, or a larger model). The system has run in a production A/B experiment for over a month. Compared with uniform use of the expensive tier, CascadeRec has reduced inference cost by about 10%, with minimal latency impact and no quality regression observed to date. Tenant data cannot be released, so we also evaluate the pipeline on public benchmarks (MovieLens-1M and Amazon-Electronics). Segment-level analysis with paired-bootstrap confidence intervals shows that uniform LLM reranking lowers NDCG@10 for sparse-history users by up to 30%, and its only statistically significant gain appears on the segment with the longest histories. Escalating cold users can make the loss larger instead of repairing it. Which segments benefit from escalation depends on the domain, and this motivates the segment-gated rollout we use in production. We describe the deployment architecture, the measured benefits, and lessons learned, including inter-tenant load imbalance as a driver for cascading and the handling of LLM output-parsing failures at the serving layer.
[IR-15] RedMall: Generative Recommendation for Large-Scale E-commerce
链接: https://arxiv.org/abs/2611.08519
作者: Kunyang Zhou,Xin Jia,Yi Yang,Yongjun Yang,Manjie Xu,Honglei Zhang,Yan Zhu,Kai Zuo,Xu Tang,Zheng Wang
类目: Information Retrieval (cs.IR)
备注: Work in progress
Abstract:Generative recommendation has shown considerable potential in industrial recommendation systems. However, the performance of existing generative methods is limited by several challenges. 1) Semantic IDs (SIDs) are treated as one-dimensional sequences, overlooking the hierarchical dependencies among different layers. 2) Next-token prediction used is unable to explicitly optimize session-wise recommendation. 3) Although reinforcement learning has been applied to align generated recommendations with user preferences, few existing methods explicitly optimize the ranking of positive items in session-wise recommendation. To address the above issues, we propose RedMall, a generative recommender designed for large-scale e-commerce. Firstly, RedMall constructs hierarchical SIDs through residual quantization over multimodal representations. Secondly, RedMall introduces a lightweight SID encoder to explicitly model layer-wise dependencies in SIDs. Moreover, RedMall incorporates sparse mixture-of-experts, dual gated attention, and multi-target prediction to scale model size, enhance user interest representation, and optimize session-wise recommendation. We further design an exposure-aware loss that treats exposed but unclicked items as negative feedback. To better align user preferences and list-wise ranking objectives, we combine Group Relative Policy Optimization with an auxiliary LambdaRank loss, using both real user feedback and item ranking to guide optimization. Experiments on four public datasets-Amazon Beauty, Amazon Toys, Amazon Food, and MovieLens-1M—show that RedMall outperforms the evaluated baselines across all reported Recall and NDCG metrics. Online A/B tests show improvements of 2.24% in advertiser value (ADVV) and 0.59% in click-through rate (CTR) for e-commerce advertising, as well as gains of 1.08% in GMV and 0.95% in daily active buyers (DAB) for e-commerce recommendations.
[IR-16] Improving Contextual Retrieval for Long Documents using Semantic Chunking and Hybrid Retrieval Strategies
链接: https://arxiv.org/abs/2611.08498
作者: Sakshi Garg,William B. Andreopoulos
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:The vast amount of information available in lengthy, distributed documents makes finding relevant answers difficult, especially when the answer requires connecting evidence from multiple sources. Standard Retrieval-Augmented Generation (RAG) systems often struggle with this type of “multi-hop” reasoning and frequently face issues with context fragmentation. In this paper, we propose a hybrid RAG framework that integrates the structured reasoning of knowledge graphs with the broad semantic coverage of dense vector retrieval. Our approach employs a multi-stage pipeline using transformer-based summarization and semantic chunking to improve context retention. We construct a dynamic knowledge graph to capture explicit relationships between entities while simultaneously maintaining a dense vector index. During inference, the system uses a hierarchical mechanism that prioritizes graph traversal for precision and falls back to vector retrieval when necessary. We evaluated our framework on the MultiHop-RAG dataset. The results show that this hybrid approach outperforms standalone vector and graph-based baselines, achieving higher Exact Match scores and improved retrieval accuracy for complex queries. This confirms that merging structured reasoning with standard vector retrieval creates a more robust system for handling complex knowledge-intensive questions.
人机交互
[HC-0] GroundWork: Bridging Human and LLM Reasoning Through a Structured Argumentation Model
链接: https://arxiv.org/abs/2611.09840
作者: Sjoerd Vink,Byungmoo Kim,Michael Behrisch,Adam Coscia,Remco Chang
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Large language models (LLMs) are increasingly integrated into workflows that require reasoning over text documents, yet interaction remains largely conversational. This leaves argumentative structure implicit or embedded in text, making it difficult to maintain, compare, and revise arguments as analyses evolve. We present GroundWork, a mixed-initiative system that makes argument structure a shared, editable state for humans and LLMs. GroundWork operationalizes the Toulmin model of argumentation as a typed graph with explicit operations for human and LLM edits. It exposes this structure through coordinated corpus views and an editable workbench for constructing, comparing, testing, and tracing source-grounded arguments. We evaluate GroundWork through an expert study with four participants spanning consulting, product development, law, and immunology, and three use-case scenarios in intelligence, law, and medicine. Our findings show that GroundWork makes source-grounded arguments inspectable, comparable, and revisable as shared human-LLM state.
[HC-1] Assessing the Impact of Retrieval-Augmented Generation on Algorithmic Problem-Solving
链接: https://arxiv.org/abs/2611.09791
作者: Gabriel Souza Baggio,Gabriel Machado Lunardi,Guilherme Medeiros Machado,Subhankar Maity,Faiza Belbachir,José Palazzo Moreira de Oliveira
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Competitive programming problems pose distinctive challenges for Large Language Models, as they demand not only proficient code generation but also advanced algorithmic reasoning. This study presents a Retrieval-Augmented Generation approach specifically designed to enhance the capacity of Large Language Models to address such problems. Our approach integrates an embedding-based retrieval mechanism and a similarity search algorithm to identify analogous problems within a knowledge database, employing a dynamically adjusted threshold to mitigate retrieval noise. We conducted an evaluation using the CodeContests dataset, focusing on the test split problems that exhibit similarity to instances from the training split. The experimental framework considered three conditions: (i) problems provided solely with their original descriptions, (ii) problems supplemented with randomly selected examples from the training split, and (iii) problems augmented with similar examples retrieved through the proposed method. In addition, we incorporated a Min-K membership inference analysis to examine whether the generated code solutions were influenced by potential memorization of training data within the evaluated Large Language Models. This complementary perspective enables a deeper understanding of model behavior by distinguishing genuine generalization through retrieval from implicit recall of pretraining samples. Performance was assessed using both pass rates and the CodeForces Elo rating of the problems. The results provide empirical evidence that similarity-based retrieval substantially enhances Large Language Models performance in competitive programming scenarios, offering an scalable framework for improving code generation in algorithmic problem-solving contexts.
[HC-2] What to Delegate What to Preserve: Human Preferences for Agent ic AI Across Everyday Life and Work
链接: https://arxiv.org/abs/2611.09730
作者: Sarah Wheeler,Veronica Rammouz,Yaoyao Qian,Zexuan Liu,Mengyi Wei,Anthony Rios,Tianzhi He,Kyrie Zhou
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:AI agents can execute multi-step tasks on users’ behalf, yet the boundaries of acceptable delegation remain unclear. We report a scenario-based survey of 289 U.S. adults with prior AI experience covering 45 use cases across nine life domains. Each respondent evaluated one scenario per domain, producing repeated measures of automation preference, trust, collaboration, AI leadership, privacy concern, and willingness to pay. Automation preferences varied substantially across domains. Generalized estimating equation models showed that greater trust in the AI agent, desired AI leadership, prior task-specific AI use, and anticipated financial savings were associated with greater automation preference. In contrast, concerns that automation would reduce personal fulfillment predicted lower automation preference. Open-ended responses further revealed a preference for bounded autonomy, with participants often favoring substantial AI execution while retaining human involvement, oversight, or final authority. We argue that delegation is best understood as a configuration of execution, collaboration, leadership, and authority.
[HC-3] Entering a New Domain in the Era of Generative AI: An Autoethnography of Solo Game Development as a Bioinformatics Educator
链接: https://arxiv.org/abs/2611.09623
作者: Hannah Kim
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注: 12 pages, 1 figure
Abstract:The rapid improvement of generative AI (GenAI) not only increased the efficiency of practice but also empowered individuals with capabilities to explore new domains. However, identity formation in the wake of such empowerment remains understudied. We use an autoethnography to understand identity formation in the context of solo game development as a bioinformatics educator. With the goal of developing a create-your-own-adventure (CYOA) game, we collected data for 73 days through journal entries, interacting with communities of practices. The autoethnographer had continuously reassessed the situation, and closely interacted with communities of practice. Their prior knowledge was used to provide in-game context and to uniquely interpret game development processes. These findings suggest that the autoethnographer made efforts to belong to communities of practice and engaged extensively in sensemaking through the lens of a bioinformatics educator. Although GenAI puts power at hand, attaining a new identity took place at a human pace.
[HC-4] DeepRadio: A Human-Agent Workspace for Physical Radio Development
链接: https://arxiv.org/abs/2611.09418
作者: Jiazhen Lei,Yuxin Sha,Sihan Wang,Wenhui Li,Jinyang Sun,Chengyu Lu,Xiaohua Tian
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Agentic programming can automate software-defined radio (SDR) development, but decisions can propagate from software to hardware without sufficient user visibility or control. We present DeepRadio, a human-agent workspace embedded in GNU Radio Companion (GRC) for jointly constructing, verifying, and operating physical radio systems. DeepRadio uses a dynamic and state-driven workflow to coordinate a traceable radio specification, claim-oriented cross-layer verification, and mixed-initiative control. These mechanisms expose decisions, tie conclusions to available evidence, and return control to users before consequential hardware actions. Across seven atomic SDR tasks, DeepRadio completed nearly all runs while satisfying all prescribed software and radio-frequency constraints. In a 20-participant study, more participants achieved a verified physical deployment with DeepRadio than with Codex-assisted GRC. Participants reported stronger support for shared intent and understanding while retaining initiative and control. These results suggest that DeepRadio can make agent-assisted physical radio development more reliable, understandable, and controllable.
[HC-5] Uniqueness-Aware Feature Engineering for Representing Humans in Algorithmic Decision Making
链接: https://arxiv.org/abs/2611.09403
作者: Theresa Elstner,Martin Potthast
类目: Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注:
Abstract:Fairness analyses of algorithmic decision systems typically treat the feature representation of individuals as an a priori fixed input. How well a feature representation captures what is actually decision-relevant for different people and their unique backgrounds, however, is typically left unexamined. We address this gap by modeling feature engineering as a stochastic selection process over the pooled characteristics as candidate features for a finite training population. Based on this model, we develop a new framework for uniqueness-aware feature engineering which operates in two layers. The first layer measures population heterogeneity using a multivariate hypergeometric model of feature prevalence. It shows how well an individual is represented by any given input representation and which individual of the population is “least represented” compared to their personal, reference characteristics. The second layer acts on this assessment using set-similarity measures to search for an optimum between representational fit of the population and predictive accuracy. Without assuming prior knowledge about unseen individuals, our framework grants represented humans agency over what counts as decision-relevant in their case, gives deploying institutions accountability through a publishable representation fit report, and provides each person subject to a decision transparency about their personal representation fit. We operationalize our framework as a six-step pipeline and show a proof-of-principle using synthetic data. Our approach offers a blueprint for making algorithmic decision systems accountable for how well they represent individuals during development and in production.
[HC-6] racybot: Preserving Developer Intent by Tracing Agent -Generated Code to Prompts
链接: https://arxiv.org/abs/2611.09395
作者: Răzvan Albu,Aliaksei Khval,Maxine Orlova,Danis Music,Filipe Araújo Veras Rosa,Lirong Yi,Ranim Khojah
类目: oftware Engineering (cs.SE); Human-Computer Interaction (cs.HC)
备注:
Abstract:Large language models (LLMs) are now widely used in software development, and LLM-based coding agents are given increasing autonomy. With coding agents, many design decisions are made in the conversation between the developer and the agent. Typically, the developer starts a conversation with an agent and states their intent, the agent proposes options if needed, and the developer plans the change accordingly. Then, the resulting code is committed. However, in such scenarios, the decisions and the intent behind the changes remain in the conversation, with no link to the code they produced. In this paper, we present Tracybot, an open-source tool that integrates with VS Code, Git, and coding agents to keep track of relevant conversations in an LLM-assisted development process. Tracybot captures each change as a Tasklet which consists of optional planning prompts used to plan a change, together with a build prompt that produces it. Tracybot has two main features: (i) it maps line-level code changes made by coding agents to the Tasklets that produced them, and (ii) it allows the developer to export a Tasklet as a decision record when needed, for example when the decision concerns the architecture of the project. We developed the tool over two iterations of a design science study, while focusing the evaluation on the usability and user-friendliness of the tool. In a focus group (7 participants) and a survey (24 participants), we found that Tracybot was perceived as useful for activities like code review, and most participants found our Tasklet design and presentation to be a clear way to represent the developer’s intent. Tool: this https URL. Video: this https URL.
[HC-7] Sensed but Not Told: Hidden Physiological Adaptation Shapes LLM Persuasion
链接: https://arxiv.org/abs/2611.09394
作者: Xiaoyan Wei,Yutong Qu,Yutong Li,Haiyi Li,Wei Emma Zhang,Hsiang-Ting Chen
类目: Human-Computer Interaction (cs.HC)
备注: 15 pages, 5 figures
Abstract:Wearable sensors, including those embedded in VR headsets, can provide LLM-based chatbots with physiological signals that users do not explicitly communicate. We tested whether such hidden input changes persuasion. In a 2\times2 within-subjects study ( N=24 ), a chatbot argued against participants’ self-selected opinions with or without turn-level heart-rate and pupil-diameter features; sensing was disclosed halfway through the study. Before disclosure, sensing increased perceived persuasiveness by 1.42 points. Across the study, sensing was associated with a 0.57-point greater shift toward the chatbot’s position and 3.20 times the odds of any movement toward it. Persuasiveness ratings were lower after disclosure, although disclosure was confounded with session order. Dialogue logs suggested a more declarative interaction style when sensing was on: chatbot replies contained question marks less often, and conversations contained fewer turns. These findings motivate making biosignal-driven adaptation inspectable and providing a genuinely non-sensed mode.
[HC-8] hinking Twice: A Sequential Approach to Nudge Towards Reflective Judgment in GenAI-Assisted Decision Making
链接: https://arxiv.org/abs/2611.09269
作者: Hüseyin Hussein Keke,Daniel Eisenhardt,Christian Meske
类目: Human-Computer Interaction (cs.HC)
备注: Full paper accepted and presented at the 20th International Conference on Wirtschaftsinformatik (WI 2025), Münster, Germany. To appear in the conference proceedings (Lecture Notes in Information Systems and Organisation, Springer). This version is the submitted manuscript prior to peer review
Abstract:Humans increasingly use generative AI (GenAI) for decision-making assistance, yet tend to engage in such situations using quick, intuitive (System 1) rather than deliberate, analytical assessment (System 2) as reflected in Dual Process Theory. Such engagement reduces the effectiveness of human-GenAI interactions and increases associated risks. Our research investigates how to design human-GenAI interactions that promote the transition from System 1 to System 2 thinking in decision-making. Based on the Three-Stage-Model, which suggests humans switch to System 2 when conflicts are detected, we conducted an experiment (n=130) testing sequential human-GenAI interaction as a cognitive forcing nudge during problem-solving tasks. This approach, where humans decide first before receiving AI assistance, enables detection of conflicting decisions. Compared to non-sequential settings (concurrent or no AI assistance), results show the sequential condition significantly enhanced both decision performance and GenAI prompt utilization, providing valuable insights for designing more reflective and thus effective human-GenAI interactions.
[HC-9] Navigating Complexity: A User-Centered Evaluation of Project Management Tools
链接: https://arxiv.org/abs/2611.09260
作者: Yasin Ghadimi Zaker,Hamid Reza Ghaeini
类目: oftware Engineering (cs.SE); Human-Computer Interaction (cs.HC)
备注: 11 Pages
Abstract:In the age of digital transformation, companies are increasingly relying on specialized software to manage the complexity of projects. However, when selecting tools, features and costs often take precedence over user experience (UX), leading to low adoption rates, user frustration, and project delays. This study empirically evaluates the user experience of leading project management tools (Jira, Trello, this http URL, Asana, and Microsoft Project) across various project scales using a mixed methods approach. The System Usability Scale (SUS) was used as the primary quantitative instrument, with data collected from 141 experienced professionals via online questionnaires. The results reveal a significant trade-off between functionality and usability. Trello achieved the highest SUS score (77.5), indicating a superior user experience for small projects and agile teams, while Microsoft Project scored the lowest (51.2), suggesting significant usability challenges despite its robust features. Jira demonstrated a balanced profile, offering powerful features with manageable complexity. Statistical analyses confirmed that the choice of the optimal tool is highly context dependent, with team size and project complexity influencing UX outcomes. The findings underscore that UX is not a secondary feature but a critical success factor for software acceptance and project effectiveness. This study provides an evidence-based framework for aligning tool selection with the project context, offers practitioners actionable insights, and identifies design opportunities for software developers.
[HC-10] RAG -Enhanced LLM s for Interactive Explainability in Clinical Decision Support Systems
链接: https://arxiv.org/abs/2611.09232
作者: Lorenzo Mannocci,Francesca Naretto,Lucia Passaro,Anna Monreale
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注: Published in HC_AIxIA_HYDRA 2025, Communications in Computer and Information Science, vol. 2830, pp. 50-63, Springer, 2026. Please cite the published version: doi: https://doi.org/10.1007/978-3-032-16708-8_4
Abstract:Artificial Intelligence (AI) is increasingly embedded in high-risks decision-making systems across domains such as healthcare, finance, and the judicial system. As reliance on predictive models grows, so does the demand for Explainable AI (XAI) to ensure transparency, trust, and interpretability. However, existing XAI approaches often fail to present explanations in a format accessible to domain experts, limiting their practical utility. Large Language Models (LLMs) have recently emerged as a bridge between complex models and end-users, offering natural language explanations. Yet, their use introduces new risks, including hallucinated outputs and lack of source traceability. To address these challenges, we propose a Clinical Decision Support System (CDSS) that not only combines standard predictive models with XAI techniques but also incorporates a Retrieval-Augmented Generation module as an interactive enhancer. This module grounds explanations in verified medical knowledge, reformulates technical outputs into a human-like format, and enables users-such as clinicians-to actively query the system. By fostering an interactive, human-in-the-loop environment, our approach empowers domain experts to explore model decisions, contextualize explanations, and build trust in AI-assisted diagnostics. The study concludes with preliminary experiments validating the proposed methodology.
[HC-11] Rhythmic Fitness Beat Em All using Real-Time Monocular Human Pose Estimation
链接: https://arxiv.org/abs/2611.09228
作者: Yanis Zaoui(I3SP - URP3625),Valentin Bégel(I3SP - URP3625),Mohamed Adjel
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Short and frequent physical activity exercices mitigate the health risks of sedentary behavior. Yet, the lack of motivation remains a critical barrier to adoption. Exergames gamify physical activity, but most rely on dedicated hardware, and outside dance and rhythm games they rarely use the musical beat. We present a work-in-progress beat 'em up exergame played with the whole body in front of a single smartphone, which runs the entire pipeline: an off-the-shelf real-time 3D human mesh recovery model on the phone’s neural processing unit, gesture recognition, and the game, shown on any TV through a USB-C to HDMI cable. Our contributions are (1) a 3D recognition of five shadow-boxing moves (guard, punch, slip, squat, jump), each defined by one metric measure in a body frame aligned with gravity and one threshold with hysteresis, personalised during a tutorial; (2) a rhythm stage whose discrete cues are placed on the song’s kick drum and judged on the movement’s capture time, with a delay learned for each player; and (3) a phone-only architecture, running at 30 frames per second, that turns a phone and an HDMI cable into an exergame console.
[HC-12] EventHOI: Event-Driven Resource-Efficient Hand-Object Interaction Logging on Smartglasses
链接: https://arxiv.org/abs/2611.09089
作者: Le Zhang,Kailai Cui,Hao Chen,Vlad Roznyatovskiy,Jianzhong Zhang,Anhong Guo,Kang G. Shin,Ke Sun
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
备注: 15 pages, 21 figures, 3 tables
Abstract:Hand-object interactions (HOIs) can be captured by smartglasses to learn how people manipulate objects in daily life. Automatic logging of these interactions from an egocentric perspective provides time-stamped evidence of daily actions for applications like memory assistance. However, continuous RGB capture and wireless transmission exceed the energy and bandwidth budgets of smartglasses. We present EventHOI, an event-driven sensing and compression system that couples selective RGB capture with HOI-aware encoding to enable egocentric logging over Bluetooth Low Energy (BLE) while preserving downstream HOI logging performance. Our key insight is to use low-power event sensing both to determine when RGB capture is needed and to identify where visual detail must be preserved. A lightweight IMU-based egomotion cancellation pipeline suppresses egomotion-induced events so that residual HOI motion can guide RGB triggering and spatial compression. We build a smartglasses prototype and conduct an end-to-end evaluation of EventHOI on a 50-hour egocentric benchmark with 97 subjects and a 20-hour real-world dataset we collect from 10 participants, together covering a broad range of daily-life scenarios and environments. Compared to RGB-only baselines, EventHOI reduces always-on backend power consumption by 5.8-10.4x and uplink bandwidth by 2.9x, with minimal HOI recognition accuracy loss. The resulting RGB uplink of less than 130 kbps fits within commodity BLE bandwidth budgets, making stable BLE video streaming feasible for egocentric HOI logging. Demo video and project website: this https URL
[HC-13] “Despite her age Margaret was eager to make a change”: Unpacking Aging Representations and Stereotypes in LLM -generated Stories
链接: https://arxiv.org/abs/2611.09043
作者: MD. Roqunuzzaman Sojib,Mohammad Sadat Hossain,Nafis Tahmid,Shattik Islam Rhythm,Novia Nurain
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:As large language models (LLMs) increasingly generate everyday stories, their portrayals can shape expectations about aging. Prior work has largely examined age stereotypes in short responses, leaving unanswered whose lives models default to imagining and how age shapes characters’ roles in complete narratives. We analyze 594 stories generated by Gemini, GPT, and Llama using text embeddings, LLM-assisted coding, and qualitative analysis. When age was unspecified, stories typically featured younger adults, with middle-aged protagonists largely absent. Older adults were often portrayed through loss, familiar routines, and help from younger characters. Even positive stories sometimes treated older adults’ achievements as surprising or portrayed others making decisions on their behalf. We suggest evaluating not only the tone of generated stories, but also who appears, what opportunities they receive, and who makes decisions. Involving older adults in these evaluations can guide writing tools toward more varied, less stereotyped portrayals of later life.
[HC-14] he Drawing You Watch Is the Portrait You Keep: A Replayable Machine Portrait Studio
链接: https://arxiv.org/abs/2611.08994
作者: Vinod Rajpurohit
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY); Graphics (cs.GR)
备注: 25 pages, 6 figures, 9 tables. Live system: this https URL
Abstract:Ask an image model for your portrait and a finished picture arrives: no first line, no second thoughts, no record of how it got there, and a different picture if you ask again. I built a portrait studio that works the other way. A sitter chooses one of eight fictional artists, each written with a city, a medium and a tempo of its own, holds still for 8 to 12 seconds, and watches a drawing of themselves appear over 44 seconds. One still is sent after consent, a hosted image model returns a private study the sitter never sees, and a compiler turns that study into one drawing document. The live drawing, the saved sheet, the vector export and the film all read that document, and a replay makes no model call, so the drawing you watch is the portrait you keep. I describe the compiler in enough detail to rebuild it and say plainly what it is: an authored reconstruction at a compressed speed, not a recorded hand. On 16 generated test studies a sheet holds 8,784 to 16,378 timed contacts, about three quarters of them fixed-grid fills and nearly all shorter than one video frame, and recompiling one study gave byte-identical output four times on one machine. The paper also traces a bug that matters more than any of that. The earlier compiler estimated the paper colour tile by tile, read dark skin as paper, and in all eight hands kept less of a deep-complexioned face than a fair one (13 to 48 percent against 41 to 56). On 80 generated studies the rebuilt engine keeps the complexion difference the study already has (raw retention 89 to 120 percent), so the fairness question has moved upstream, to the study and the hosted model. The two test sitters are generated and differ in more than complexion, and no real-face or perceptual study exists yet. The studio is live at this https URL.
[HC-15] he Relational Capacity Framework: A Lens for AI-Mediated Relationality
链接: https://arxiv.org/abs/2611.08877
作者: Emily Tseng,Patrick Gage Kelley,Renee Shelby,Ashley M. Walker
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注:
Abstract:A growing literature examines the impact of engaging with AI chatbots on our relationality: our ability to form relationships with others, maintain them through life’s changes, and exit when needed. Currently, substantial work focuses on purpose-built AI companions (e.g. Replika, this http URL), with less work investigating socioemotional uses of the more widely-used general-purpose chatbots. In this paper, we present a 100-person investigation into how people engage general-purpose AI for socioemotional needs. Participants’ AI uses varied, from practicing specific social skills to seeking additional bandwidth and information as they navigated their social contexts. To explain this heterogeneity, we provide 8 states that describe how AI expands or contracts relationality for people with different levels of skills, bandwidth, and information. We then present a unified conceptual model for the role of technology in mediating human relationality, the Relational Capacity Framework, and outline future research designing appropriate roles for AI in human connection.
[HC-16] Bidirectional Modeling of mmWave Human Signals for Enhancement Synthesis
链接: https://arxiv.org/abs/2611.08870
作者: Shuntian Zheng,Hoang Hai Pham,Jiaqi Li,Xiaoman Lu,Shuai He,Yu Guan
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Millimeter-wave (mmWave) is attractive for human-centered robotics in smoke, darkness, and privacy-sensitive indoor spaces: it measures reflections rather than visually identifiable imagery and remains usable without light. Deployed systems, however, face two complementary needs that run in opposite directions. Enhancement starts from a raw range-angle-Doppler (RAD) heatmap, where human returns coexist with nonhuman structure, and aims to isolate the human signal by removing nonhuman returns, yielding a clean RAD input for downstream perception. Synthesis starts from denser human geometry and aims to synthesize natural RAD observations, including environmental and hardware features, thereby expanding scarce supervision data. Although the two tasks reverse direction, both require modeling how human and nonhuman reflections coexist in the same measurement. Among RAD axes, Doppler provides motion information native to the modality, with the potential to support this modeling. We therefore cast both tasks on the dual-domain RAD heatmap and propose HeatWave, which injects Doppler when it is locally reliable, and routes Doppler asymmetrically for enhancement and synthesis within one shared backbone. This design improves both tasks: enhanced observations reduce mean per-joint position error by 11.4% relative to the raw recording, and synthesized samples that replace half of the real radar labels recover most of the activity accuracy lost to the reduced budget. Relative to an unspecialized dual-task heatmap baseline, the enhancement improves motion-critical samples under metallic clutter by 2.3times more than when geometry already suffices, and synthesis follows the same ordering with a narrower spread, providing the community with a research direction more closely aligned with the signal’s characteristics.
[HC-17] What Limits mmWave Human-State Sensing for Robot Perception?
链接: https://arxiv.org/abs/2611.08863
作者: Shuntian Zheng,Guangming Wang,Hongyang He,Shuai He,Yu Guan
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC)
备注:
Abstract:Assistive robots and search-and-rescue platforms require reliable human state sensing in smoke, darkness, and privacy-sensitive environments, where cameras and LiDAR may be unreliable Millimeter-wave radar offers an alternative that works without ambient light, penetrates smoke, and does not capture visual appearance. However, the compact arrays that fit robot size, power, and cost budgets have limited apertures, producing joint-localization errors on the order of decimeters at a few meters, large enough to compromise the safety margins and tracking accuracy a robot depends on. Existing pipelines reduce pose error through frame fusion, multiple viewpoints, return selection, or stronger estimators, but examine one technique at a time, without relating its effect to a shared measurement limit or separating gains from added information from gains from estimator completion. We address this gap with a framework that models the human and environment as one time-varying scattering field and states a Cramer-Rao type error floor for joint localization as a function of bandwidth, aperture, signal-to-noise ratio, fused frame count, phase coherence, and visibility of retained returns. We then use existing frame fusion, multi viewpoint, return selection, and pose estimation methods as matched probes of this floor rather than as competing pipelines. Under the evaluated conditions, experiments on three datasets and a live deployment show these strategies move joint localization error in the predicted directions, for example lowering mean per-joint position error from about 97 to 65 mm with a second viewpoint, while a stronger estimator mainly lowers overall error by completing joints the measurement cannot resolve. These results give the robotics community a principled basis for deciding whether to prioritize observation design or estimator capacity when deploying reliable mmWave human sensing.
[HC-18] “I had nobody to turn to”: Why LGBTQ People Choose (Not) to Use AI Chatbots for Personal Health and Relationship Advice
链接: https://arxiv.org/abs/2611.08854
作者: Jacob Hobbs,Ashray Gupta,Dennis Ulmer,Geetanjali Aich,Nolan Golden,Yanan Long,Ian Rios-Sialer,Jennifer Mickel,Morgan Klaus Scheuerman,Rohit Avinash Mujumdar,Virginia K. Felkner,Jordan Taylor,William Agnew
类目: Human-Computer Interaction (cs.HC)
备注: 35 pages, 2 figures, 8 tables
Abstract:While there is an abundance of research into large language model (LLM) biases against LGBTQ+ communities, far less research has examined how and why LGBTQ+ people choose to make use of AI chatbots. In this work, we examine how LGBTQ+ people choose (not) to use AI chatbots across three domains in which LGBTQ+ users have unique needs: relationship support, mental health support, and gender or sexual health support. We surveyed 73 people (60 LGBTQ+; 13 non-LGBTQ+) on how they interact with AI chatbots and find that participants’ access to external resources, privacy concerns, and perceptions of chatbot capabilities shape whether and how they choose to make use of AI chatbots. For instance, some LGBTQ+ participants felt that chatbots might actually be less biased against LGBTQ+ people than the people they know offline. We also find that LGBTQ+ participants report using AI chatbots less frequently than non-LGBTQ+ participants across 15 specific personal health and relationship advice scenarios, citing concerns about AI sycophancy and adverse cognitive side-effects. Recognizing AI chatbots as vital (albeit imperfect) resources for LGBTQ+ people, we recommend researchers consider systemic issues that drive users to AI chatbots and design technologies (AI or otherwise) that prioritize privacy, anonymity, and real human connection.
[HC-19] Lea: An Agent Backbone for Mathematician-Controlled Formalization in Lean
链接: https://arxiv.org/abs/2611.08849
作者: Shaswat Patel,Daniel Aarao Reis Arturi,Kevin Gomez,Jaume de Dios Pont,Jorge Poco,Eduardo V. Teixeira,Sam Westrick,Claudio Silva,Chinmay Hegde
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Agentic systems now formalize research-level mathematics in Lean, but they are built for autonomy: the mathematician states a target and inspects a result. The alternative, a general-purpose coding harness, is domain-neutral and must be assembled into a formalization workspace before any mathematics begins. Mathematicians, however, report wanting to keep high-level control over how an argument is decomposed and established. We present \textbfLea , a Lean-specific agent backbone for mathematician-facing software rather than a single interface. Lea exposes runs, sessions, and typed events through one application-neutral API, and ships with Lean tooling (blueprints, LSP, Loogle, SafeVerify) already wired. We report two applications built on it: \textbfLeaChat , a web client for interactive proof authoring, and \textbfLeaOverleaf , which formalizes theorems annotated directly in an Overleaf document; a session begun in either continues in the other. In a case study, a mathematician outlined a proof of the Durfee square generating function in a few natural-language messages, and Lea produced a complete Lean development from Mathlib alone, following the intended decomposition closely and requiring substantive correction at a single point. Lea and both applications are released as open source so that further mathematician-facing tools can be built on the same backbone.
[HC-20] Driver Takeovers and Longitudinal Responses Analysis from Naturalistic Driving
链接: https://arxiv.org/abs/2611.08798
作者: Yuhang Wang,Hao Zhou
类目: Human-Computer Interaction (cs.HC)
备注: 28 pages, 14 figures. See arXiv:2603.06986 for the companion dataset paper. The longitudinal subset is available at this https URL
Abstract:As Level 2 advanced driver-assistance systems (ADAS) become more widely used, understanding driver-system interaction is increasingly important for traffic safety. However, limited real-world data make these interactions difficult to study. We present ADAS-TO, a naturalistic dataset of 16,446 takeover clips from 364 drivers across 2,585 different drives. It combines road video, vehicle kinematics, driver inputs, and ADAS status. To analyze takeover behavior, we retained 6,206 manually reviewed takeover events during continued lane keeping or car following. We excluded events in which the driver intentionally changed lanes, turned, or made a stop beyond the ADAS’s capabilities (e.g., at a red light or stop sign) shortly after takeover. Drivers began intervening before ADAS disengaged in 65.6% of takeovers involving braking or steering. When the ego vehicle was still closing on a lead vehicle at disengagement, longitudinal safety margins often narrowed further after the transfer to manual control. Among 2,031 such events, 78.5% had a lower minimum time-to-collision (TTC) in the five seconds after disengagement than in the five seconds before it. For events involving braking, the mean TTC trajectory reached its minimum about 1.2 s after disengagement. We model the subsequent braking response using a Gazis-Herman-Rothery (GHR) car-following law. Its median five-second spacing root-mean-square error is 1.34 m with observed lead motion and 1.84 m with predicted lead motion. The fitted model expresses braking as a response to looming, with stronger responses at higher speeds and shorter gaps. Applied before takeover, it shows that greater looming than predicted by the human-response reference is associated with disengagement. These findings support continued assessment of longitudinal safety after ADAS disengagement, while the driver’s braking response develops.
[HC-21] MotionDecipher: LLM -Assisted Clinical Interpretation of Human Motion Signals for Efficient Gait Analysis
链接: https://arxiv.org/abs/2611.08795
作者: Mingrui Ma,Tian Tan,Jinsi Guo,Jessica Rose,Yiwen Dong
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: 33 pages, 19 figures, 2 tables. Accepted to Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT), Vol. 10, No. 4, Article 245, December 2026
Abstract:Clinical gait and motion analysis is widely used to guide targeted interventions for patients with neurological disorders, the outcome of which is typically a report with descriptions of joint movements and forces. Current practices rely on experts’ manual effort, which is labor-intensive and time-consuming. Recent advancements in Large Language Models (LLMs) provide opportunities to address this problem, but translating human motion into clinical gait descriptions remains challenging because the quality of the context directly affects LLMs’ reasoning capability over continuous motion signals. To tackle this, we introduce MotionDecipher, a novel framework bridging the semantic gap between motion data and clinical interpretations. Specifically, we made 3 key advancements: 1) a Human Motion Tokenizer that converts motion signals into domain-specific tokens for time- and frequency-informed gait interpretation; 2) a Meta Feature Selector that transforms tokens into Structured Representations; and 3) a Hybrid Context Constructor to determine the optimal demonstration-based prompt for In-Context Learning. We evaluated MotionDecipher on data collected at an accredited gait lab from 122 patients and 46 controls. Results show that MotionDecipher achieves 0.91 BERTScore (13% over GPT-5), while expert evaluation ranks MotionDecipher highest in Correctness, Completeness, and Conciseness compared to baselines, highlighting its promise for efficient gait analysis.
[HC-22] Learning from How Users Interpret: Personalizing Recommendation Explanation Order
链接: https://arxiv.org/abs/2611.08785
作者: Deo Munduku,Elsa Negre
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: Preprint
Abstract:Personalized explanations typically adapt their content or presentation based on preferences, interaction histories, or other user characteristics. We examine whether regularities inferred from prior textual interpretations can support personalization of explanation organization. In a controlled within participant study with 122 participants, prior interpretations were represented along three dimensions: information selection, relative salience, and relational organization. The relational dimension was partially operationalized through co mobilization. These representations were aggregated into fixed historical profiles and used to organize otherwise identical explanation sentences. Participants encountered four organization procedures: the OWN historical profile, the DONOR historical profile, a population profile called POP, and a language model procedure using the historical profile called PROFILE LLM. Post exposure objective comprehension was the primary outcome. OWN achieved 86.5 percent correct responses, compared with 66.0 percent for DONOR, 63.9 percent for POP, and 73.0 percent for PROFILE LLM. Mixed effects analyses showed higher comprehension with OWN than with each alternative procedure. No reliable overall condition effect was detected for recommendation acceptance. These findings indicate that historical profiles derived from prior interpretations can provide an actionable signal for personalizing explanation order while keeping explanatory content and wording fixed.
[HC-23] Goals as a Representational Layer for Human-AI Collaboration
链接: https://arxiv.org/abs/2611.08757
作者: Gaia Molinaro,Lev Tankelevitch,Sean Rintel,Scott Counts
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:Goals are central to how people, teams, and organizations structure their work. Yet, while generative artificial intelligence (AI) systems increasingly operate across a worker’s context, information about why a person is pursuing a task, beyond local prompts and interactions, is typically missing. We argue that goals should become an explicit representational layer mediating interactions among people, AI systems, and organizations. Because human goals are hierarchical, partially latent, and dynamically evolving, such a layer must be multidimensional, both grounded in human articulation and inferred from available evidence, and continuously updated. A goal layer could connect immediate requests to broader outcomes, improve the organization and retrieval of relevant context, support reflection and coordination, and provide a foundation for training, aligning, and evaluating AI systems in terms of human-centered outcomes. We highlight these opportunities while identifying open challenges involving goal inference, socio-cognitive frictions, alignment, privacy, and governance.
[HC-24] “A Space Where I Can Be Myself”: Exploring Safe Visibility for Graduate Student Mothers in Mom Cafes
链接: https://arxiv.org/abs/2611.08710
作者: Jeanne Choi,Joohee Choi,Jeong-woo Jang
类目: ocial and Information Networks (cs.SI); Human-Computer Interaction (cs.HC)
备注: 28 pages, 1 table; 1 table and 3 figures in appendix
Abstract:Graduate student mothers navigate intertwined demands of graduate study and caregiving, yet little is known about how they seek support through online communities. We examine this through Mom Cafes, locally based online parenting communities in South Korea. Drawing on focus group interviews and design workshops with 22 graduate student mothers, we identify a pattern of dual invisibility: caregiving circumstances were difficult to make legible in academic contexts, while academic circumstances were difficult to make legible in parenting contexts. Although Mom Cafes offered useful local information and emotional reassurance, concerns specific to graduate student motherhood were harder to articulate or safely disclose. Participants sought ways to make graduate student motherhood recognizable when relevant, but their designs highlighted tensions around identifiability, judgment, relational burden and accountability. We conceptualize these tensions as safe visibility and discuss infrastructures for selective recognition, low-obligation connection, and accountable pathways to practical care.
[HC-25] Seeing Is Not Knowing: The Epistemic Limits of LLM -Based Mental Health Inference from Youth Social Media Conversations
链接: https://arxiv.org/abs/2611.08565
作者: Xiaohan Ding,Mengting Tu,Andrew H. Zhao,Jamie Zelazny,Candice L. Biernesser,Munmun De Choudhury
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:Adolescents increasingly use social media for private conversations, sharing experiences and seeking support. These interactions generate digital traces, prompting efforts to infer mental health and psychosocial states. As large language models become more capable of reasoning about conversations and interpersonal dynamics, a fundamental question remains nonetheless: what can models actually “know” about an individual’s psychological experience from what they are able to “see”? We evaluated ten reasoning LLMs using donated Instagram Direct Messages from 80 adolescents with a history of depression or suicidal thoughts and behaviors, comparing predictions with self-reports and analyzing 204 reasoning traces. Models performed better when inferring more observable states than participants’ subjective appraisals of social experience, such as perceived social support. Beyond prediction accuracy, our findings highlight the epistemic limits of digital traces for AI design and evaluation: what can be observed, what can reasonably be inferred, and what requires people’s own accounts.
[HC-26] Relic and Gumi: Governing Memory Safety and Continuity in Diegetic Relational Agents ICIP
链接: https://arxiv.org/abs/2611.08504
作者: Daniele Veri’
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: 51 pages, 3 figures; exploratory two-participant Telegram pilot; HCI systems paper
Abstract:Longitudinal relational agents rely on memory for continuity, but remembering can lead to intrusion, clinical overconfidence, emotional manipulation, and opacity. We present Relic, a governance architecture for diegetic relational agents, and Gumi, a subject-specific companion agent initialized and governed through Relic. The architecture separates what the system can observe, what researchers can review, what Gumi can remember, and what Gumi can say. We describe three mechanisms. Relic represents sensitive safety signals as researcher-facing governance objects rather than user-facing explanations or agent memories. Shared Continuity Memory organizes relational markers around subject-confirmed wording, corrections, explicit follow-up permissions, and recall limits. We specify a runtime mediation layer for delivery, proactivity, session identity, output transformation, and platform allowlists; adoption across live paths remains partial. These mechanisms aim to preserve relational continuity without turning Gumi into a clinical assistant, mood tracker, general-purpose assistant, or opaque profiling system. We contribute the architecture, object model, researcher workbench design, and a proposed evaluation framework for governed relational memory. An exploratory pilot involved the author as P1 and one external participant as P2. Weekly self-reports from P1 and researcher-led interviews with P2 provide descriptive observations of early use and perceived naturalness. The pilot does not support claims about clinical efficacy, diagnosis, therapeutic outcomes, general usability, or deployment safety. Relic provides research infrastructure intended to make relational memory, safety constraints, and agent behavior inspectable and testable. Comments: 51 pages, 3 figures; exploratory two-participant Telegram pilot; HCI systems paper Subjects: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI) ACMclasses: H.5.2 Cite as: arXiv:2611.08504 [cs.HC] (or arXiv:2611.08504v1 [cs.HC] for this version) https://doi.org/10.48550/arXiv.2611.08504 Focus to learn more arXiv-issued DOI via DataCite
[HC-27] Investigating Anthropometric Fidelity in SAM 3D Body
链接: https://arxiv.org/abs/2601.06035
作者: Aizierjiang Aiersilan,Ruting Cheng,James Hahn
类目: Graphics (cs.GR); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Emerging Technologies (cs.ET); Human-Computer Interaction (cs.HC)
备注:
Abstract:The release of SAM 3D Body is a recent development in human mesh recovery, demonstrating improved performance in producing clean, topologically coherent meshes from single images. By leveraging the Momentum Human Rig (MHR), it achieves robustness to occlusion and diverse poses. However, our evaluation reveals a specific and consistent limitation: the model struggles to reconstruct detailed anthropometric deviations, particularly in populations exhibiting distinctive morphological alterations such as geriatric muscle atrophy, scoliosis, or pregnancy, even when these features are prominent in the input image. In this paper, we investigate this phenomenon not as a failure of the model’s capacity, but as a byproduct of the “perception-distortion trade-off”. We posit that the architectural reliance on the low-dimensional parametric MHR representation, combined with semantic-invariant conditioning (DINOv3) and annotation-based alignment, creates a pervasive “regression to the mean” effect. We analyze these mechanisms to understand why individual biological details are smoothed out. Furthermore, we state our contributions by proposing specific, constructive pathways for future work, such as implicit-explicit hybrid representations and Medical-in-the-Loop alignment, to extend the baseline performance of SAM 3D Body into the high-precision medical domain.
[HC-28] Literature Review of Mixed Reality Research
链接: https://arxiv.org/abs/2312.02995
作者: Aizierjiang Aiersilan
类目: Human-Computer Interaction (cs.HC); Graphics (cs.GR); Software Engineering (cs.SE)
备注:
Abstract:In the global context, while mixed reality has been an emerging concept for years, recent technological and scientific advancements have now made it poised to revolutionize industries and daily life by offering enhanced functionalities and improved services. Besides reviewing the highly cited papers in the last 20 years among over a thousand research papers on mixed reality, this systematic review provides the state-of-the-art applications and utilities of the mixed reality by primarily scrutinizing the associated papers in 2022 and 2023. Focusing on the potentials that this technology have in providing digitally supported simulations and other utilities in the era of large language models, highlighting the potential and limitations of the innovative solutions and also bringing focus to emerging research directions, such as telemedicine, remote control and optimization of direct volume rendering. The paper’s associated repository is publicly accessible at this https URL.
[HC-29] oward a Robust Crowd-labeling Framework using Expert Evaluation and Pairwise Comparison
链接: https://arxiv.org/abs/1607.02174
作者: Faiza Khan Khattak,Ansaf Salleb-Aouissi
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:Crowd-labeling emerged from the need to label large-scale and complex data, a tedious, expensive, and time-consuming task. One of the main challenges in the crowd-labeling task is to control for or determine in advance the proportion of low-quality/malicious labelers. If that proportion grows too high, there is often a phase transition leading to a steep, non-linear drop in labeling accuracy as noted by Karger et al. [2014]. To address these challenges, we propose a new framework called Expert Label Injected Crowd Estimation (ELICE) and extend it to different versions and variants that delay phase transition leading to a better labeling accuracy. ELICE automatically combines and boosts bulk crowd labels supported by labels from experts for limited number of instances from the dataset. The expert-labels help to estimate the individual ability of crowd labelers and difficulty of each instance, both of which are used to aggregate the labels. Empirical evaluation shows the superiority of ELICE as compared to other state-of-the-art methods. We also derive a lower bound on the number of expert-labeled instances needed to estimate the crowd ability and dataset difficulty as well as to get better quality labels.
计算机视觉
[CV-0] Video-Coder: Reinforcing Agent ic Coding for Video Generation
链接: https://arxiv.org/abs/2611.09894
作者: Chaoyang Wang,Sicheng Gao,Junan Hu,Kunyang Li,Dongyang Chen,Yiyang Zhou,Huaxiu Yao,Yong Jae Lee,Radu Timofte,Yuzhang Shang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project Page: this https URL Code: this https URL
Abstract:Recent years have witnessed remarkable progress in video generation in terms of visual quality and temporal consistency. However, generating videos with precise spatial structure, coherent motion dynamics, and controllable scene composition remains challenging. In this paper, we introduce Video-Coder, the first attempt to train a code-driven video generation agent, which autonomously writes and executes code to construct a dynamic 3D scene (Draft Video) that provides structure and motion signals to downstream video generators. To support this paradigm, we design a dedicated data pipeline and curate two high-quality datasets, Video-Coder-SFT-6k and Video-Coder-RL-2k, which contain diverse prompts and ground-truth reference videos. We further develop the Video- Coder Harness, a unified interactive environment integrating code execution, 3D scene rendering, and execution feedback for agent training and inference. We then train Video-Coder with SFT fol- lowed by harness-native agentic reinforcement learning with dual reward feedback, which combines execution-based and video-based rewards. Additionally, we design an annealing training strategy that gradually reduces the execution reward over time, allowing the model to rely more on the video reward in the later training stages. Experiments show that Video-Coder achieves significant gains, improving Minimax-H3 by approximately 7 points on PhyGenBench and 24 points on VBench-2.0. We hope this work can serve as an open foundation for coding agents in video generation, and we fully open-source our data, models, and code
[CV-1] On Agent ic Manipulation Policies
链接: https://arxiv.org/abs/2611.09892
作者: Chen Chen,Omid Bagherieh,Naveen Nikamanth Appiah Balaji,Jesse T. Gonzalez,Negin Heravi,Fahad Islam,Xueting Li,Jean-Baptiste Regli,Kejia Ren,Petr Švec,David J. Yoon,Matthias Müller,Vladlen Koltun
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Videos and code available at: this https URL
Abstract:Recent work on robot manipulation has focused on specialist policies trained on demonstrations. We show that generalist AI models can control manipulators more reliably than specialist policies such as vision-language-action models (VLAs) and world-action models (WAMs). We systematically analyze agentic manipulation policies (AMPs), in which a generalist AI model commands the manipulator directly, adapting its actions as the task unfolds. We present AMP1, a reference harness that runs unchanged across single-arm, bimanual, and mobile manipulation. AMP1 achieves state-of-the-art results on the RoboLab-120 and CaP-Gym benchmarks, with no demonstrations or task-specific training. On RoboLab-120, AMP1 reaches an 85.6% success rate zero-shot, compared to 49.2% and 42.9% for leading VLA and WAM policies. The agent succeeds zero-shot in 104 of 108 real-world trials spanning twelve tasks and three robot instances, reducing failure rates by 20x and 15x relative to strong VLA and WAM baselines. The basic harness supplies no grasping, pushing, or placing skills; the agent develops its own strategies, such as searching for hidden objects, using objects as tools, and calibrating its cameras. With the addition of a persistent memory store, the agent can also learn from experience: an augmented harness with a memory representation that is carried across episodes lowers the failure rate on RoboDojo bimanual manipulation tasks by 4.6x relative to zero-shot deployment. Our analysis suggests that AMPs are substantially more general and robust than specialist manipulation policies, providing a promising foundation for robot manipulation.
[CV-2] FlowMoP: Stochastic Multi-Person Motion Prediction NEURIPS2026
链接: https://arxiv.org/abs/2611.09882
作者: Aadya Agrawal,Ho Kei Cheng,Alexander Schwing
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at NeurIPS 2026. 22 pages, including references and appendix
Abstract:Predicting the future motion of multiple people is inherently stochastic: the same observed history admits many plausible continuations depending on intent, coordination, and social context. Yet the dominant paradigm in multi-person 3D motion prediction remains deterministic, with existing methods producing a fixed number of outputs per agent. In contrast, we present a conditional flow matching framework that jointly generates diverse future motions for multiple agents, modeling both global trajectories and root-relative poses. A B-spline reparameterization compresses the generative state space while enforcing temporal smoothness, and a dual-stream transformer conditions each agent’s predictions on observed neighbor motion for interaction-aware forecasting. Our method achieves state-of-the-art performance on CMU-Mocap (UMPM) and 3DPW, outperforming deterministic baselines on joint position error, pose error, and final displacement error despite producing a full predictive distribution. We additionally establish the first motion prediction baseline on WorldPose, a large-scale professional soccer dataset, where social conditioning demonstrably concentrates predictive uncertainty along interaction-constrained directions.
[CV-3] ZeoBench: A Benchmark for Self-Supervised Learning on 3D Zeolite Representations NEURIPS2026
链接: https://arxiv.org/abs/2611.09880
作者: Aaron Sun,Yachan Liu,Ping Yang,Peng Bai,Subhransu Maji
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026 Datasets and Benchmarks Track
Abstract:Zeolites are an important class of porous crystalline materials with broad applications in gas storage, separations, and catalysis. Although hundreds of thousands of experimentally realized and hypothetical zeolite structures are known, identifying materials optimized for a target application remains challenging because property labels are expensive to obtain and available only for limited structure-molecule pairs. While prior work has demonstrated that 3D neural representations can predict zeolite properties effectively, it remains unclear which representation-learning strategies are most effective in the low-data regime relevant to materials discovery. We present ZeoBench, a benchmark for self-supervised learning on dense 3D representations of all-silica zeolites, and evaluate the resulting features on six existing downstream adsorption-property prediction tasks, across a range of training set sizes. Our study compares self-supervised 3D pretraining on volumetric zeolite data, adaptations of image-pretrained 2D ConvNets and vision transformers to 3D via multiview and channel-inflation strategies, and 3D ConvNets pretrained on out-of-domain volumetric datasets, alongside hand-crafted descriptors and supervised baselines. Using our pretraining and evaluation framework, we find self-supervised 3D representations outperform existing methods and achieve state-of-the-art performance across both in-domain label regimes and for the unseen domain of metal organic frameworks. Learned representations consistently outperform hand-crafted descriptors, and convolutional architectures are generally more effective than vision transformers in this setting. These results establish ZeoBench as a practical benchmark for 3D representation learning in zeolites and provide guidance for data-limited adsorption property prediction.
[CV-4] Graded convolutional networks: scale invariance without the scale and robustness to unknown blur
链接: https://arxiv.org/abs/2611.09878
作者: Tanush Shaska
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 26 pages, 2 figures, 7 tables. Code, trained networks and results: this https URL
Abstract:The Gaussian jet of an image is a graded vector space: a dilation by \lambda multiplies derivatives of order k by \lambda^-k . Local graded convolutions are exactly homogeneous differential operators, and a local feature is dilation-invariant if and only if it factors through the scale-normalized jet or, without the scale, through the retraction onto the weighted sphere of the grading. Under unknown Gaussian blur the retraction changes only by a change of scale, whereas the scale-normalized jet acquires a grading factor that no radial normalization removes. At the zero jet the retraction separates image families that radial normalizations and continuous classifiers cannot. Networks built on the retraction are universal for invariant features. On MNIST such a network keeps 88.4% accuracy under a two-pixel blur, against 66.2% for the same network on scale-normalized derivatives; in Lindeberg’s Gaussian derivative network the retraction raises accuracy at that blur from 91.5% to 95.1%.
[CV-5] GRAC: Graph-Based Retrieval-Augmented Captioning for Surgical Scene Description
链接: https://arxiv.org/abs/2611.09866
作者: Rochak Dhakal,Chen Zhao,Soup Cobb,Xin Qian,Xiaofeng Yang,Weihua Zhou
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 20 Pages and 4 Figures
Abstract:Purpose: Surgical operative reports support procedural documentation, surgical training, performance assessment, and workflow review. However, extracting meaningful information often requires extensive manual review and clinical expertise. Automated surgical scene descriptions could reduce this burden by leveraging recurring surgical activities and prior knowledge. Methods: We propose GRAC (Graph-Based Retrieval-Augmented Captioning), a framework that represents surgical scenes as graphs capturing relationships among instruments, actions, and anatomical structures. Given an input scene graph, GRAC learns to retrieve relevant graph-caption pairs from training data. It compares instrument-tissue interactions between retrieved examples and the input scene, using comparison features to condition retrieved-caption representations during generation. Results: On the EndoVis-18 benchmark, GRAC achieved a CIDEr-D of 5.88, compared with 4.23 for SGT and 3.36 for the strongest conventional Transformer baseline. Exact instrument-action-target triplet-set accuracy reached 52.6%, compared with 32.7% for SGT and 26.2% for the strongest conventional baseline. These results demonstrate improvements in language quality and factual agreement. On TORS, GRAC achieved a CIDEr-D of 5.06, compared with 3.80 for the strongest baseline. Conclusion: GRAC demonstrates that combining structured surgical scene representations with retrieved examples improves surgical scene captioning and produces more contextually meaningful descriptions. Comments: 20 Pages and 4 Figures Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.09866 [cs.CV] (or arXiv:2611.09866v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.09866 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Weihua Zhou [view email] [v1] Fri, 6 Nov 2026 18:47:23 UTC (2,201 KB)
[CV-6] Flow-Guided Target-Space Alignment via Path Consistency NEURIPS2026
链接: https://arxiv.org/abs/2611.09863
作者: Ruizhi Yuan,Zeqiu Yu,Wei Gao,Wei Chen,Xiwei Tang,Lu Tang
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at NeurIPS 2026
Abstract:Many multimodal alignment problems require mapping a source modality into a fixed space defined by a target modality, where the target space is not merely an intermediate embedding but is used directly for downstream analysis. In this setting, successful alignment should preserve both the global structure of the target distribution and the instance-level correspondence observed in paired source–target training data. Existing objectives typically address only part of this goal: contrastive losses encourage relative similarity, distribution-matching losses improve marginal agreement, and pointwise regression fits each source sample to its paired target without explicitly modeling target-space path geometry. We propose FlowAlign, a flow-guided target-space alignment method that repurposes conditional flow matching as a training-time alignment signal rather than an inference-time transport model. FlowAlign uses the learned flow field to guide the alignment of source-side representations with the fixed target space, encouraging them to preserve paired correspondence while remaining compatible with the global target-space structure. We further provide theoretical analysis showing how FlowAlign helps improve paired correspondence by controlling endpoint accuracy and capturing local target-space geometry. Experiments on synthetic data, paired single-cell multiome benchmarks, and image–text alignment demonstrate that FlowAlign improves instance level correspondence while preserving target-space structure across different modality settings.
[CV-7] Health-System-Wide Mapping of Demographic Information in Routine Clinical Histopathology
链接: https://arxiv.org/abs/2611.09851
作者: Shengjia Chen,Ruchika Verma,Kevin Clare,Paula Restrepo,Jannes Jegminat,Eugenia Alleva,Kuan-lin Huang,Aryeh Stock,Brandon Veremis,Thomas Fuchs,Gabriele Campanella
类目: Computer Vision and Pattern Recognition (cs.CV); Quantitative Methods (q-bio.QM)
备注: 35 pages, 4 main figures, 2 supplementary figures, and 10 supplementary tables. Interactive Results Atlas and Hypothesis Explorer accompany the manuscript
Abstract:The extent to which demographic-associated information is detectable across organs and disease states in routine clinical histopathology remains unclear. Here, we analyzed 1.3 million hematoxylin and eosin whole-slide images from 257,322 patients across 25 organs, using pathology foundation-model representations and multiple-instance learning to systematically map information associated with sex, age group, and self-reported race. Across 69 organ-demographic combinations, demographic attributes were broadly recoverable, with pan-organ mean AUROCs of 0.828 for sex, 0.733 for age group, and 0.662 for self-reported race, and substantial variation across organs. Demographic-class balancing produced relatively small changes in performance, while evaluation within diagnosis-text clusters showed modest attenuation. Under training with diagnosis-based grouping, demographic-associated signals remained detectable across multiple organs, including sex in gastric tissue (AUROC 0.865), age group in breast (0.816), and self-reported race in skin (0.728). Literature-guided analysis of these settings generated candidate histomorphological and contextual explanations for further investigation. Together, these findings provide a systematic map of demographic-associated information in routine histopathology, informing investigations of its sources, assessments of potential shortcut learning, and the development of more interpretable computational pathology models.
[CV-8] BEV3SR: Evidence-Aware 3D Spatial Reasoning from Video WACV
链接: https://arxiv.org/abs/2611.09816
作者: Fan Zhu,Rui Yu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: WACV camera ready
Abstract:3D spatial reasoning from video remains challenging for VLMs because videos provide sequential, view-local observations rather than persistent metric scene records. Many questions must be answered in a query-specified perspective rather than the observed camera viewpoint, and prompting with a fixed set of RGB views may miss sparse or occluded visual cues. To address these gaps, we introduce BEV3SR, a training-free and backbone-agnostic framework for RGB-D videos with camera parameters. BEV3SR turns video observations into queryable metric bird’s-eye-view (BEV) scene memory and formalizes each query into an evidence-aware spatial task. The Spatial Task Formalizer specifies the query’s reference frame, spatial constraints, objective, and evidence requirement. The evidence requirement determines the appropriate reasoning pathway. A query treated as requiring geometric evidence is handled by the Geometric Reasoner, which deterministically computes the answer using the metric BEV records in the query-specified reference frame. Otherwise, a query treated as requiring visual evidence is passed to the Visual Evidence Aligner, which uses BEV-indexed object records and their associated supporting views to retrieve object-consistent and scene-context RGB frames for VLM grounding. Within a standard 16-frame budget, BEV3SR consistently improves different VLM backbones across multiple 3D scene QA benchmarks, delivering large gains on spatially intensive questions and significant improvements on open-ended scene QA. These results demonstrate that BEV3SR offers an effective, efficient, and interpretable framework for explicit 3D spatial reasoning from video.
[CV-9] AR-Edit: Training-Free Autoregressive Video Editing without Inversion
链接: https://arxiv.org/abs/2611.09795
作者: Hovhannes Margaryan,Quentin Bammey,Christian Sandor,Vicky Kalogeiton
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Editing a video stream in real time, without access to future frames, is essential for interactive applications. Existing training-free video editing methods, however, assume full offline access to the entire sequence, while streaming approaches either rely on additional training or are limited to specific editing techniques. We show that this limitation can be removed by analyzing the self-attention dynamics of DMD-distilled autoregressive (AR) video diffusion models. We uncover a key asymmetry: queries and keys remain stable and aligned with their source counterparts across denoising timesteps, while values diverge and encode the evolving content. Inherited from the non-distilled bidirectional teacher, this asymmetry is strongest at high noise, where few-step streaming models operate. The resulting mechanism is simple: a single forward pass suffices to capture source identity in the value features, which can be reused throughout generation. Starting from pure Gaussian noise and injecting only these source values reconstructs the input with fidelity matching or surpassing inversion-based methods, effectively eliminating inversion. Based on this insight, we introduce AR-Edit, a training-free, inversion-free method that caches source values once and selectively injects them during denoising. Across multiple AR backbones, AR-Edit enables real-time streaming editing with superior source preservation and tonal fidelity at minimal overhead.
[CV-10] Seeing What Matters: Benchmarking Task-Conditioned Actionable Spatial Belief for Manipulation
链接: https://arxiv.org/abs/2611.09794
作者: Yina Jian,Mingkun Xu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 22 pages (8 main + 14 supplement). Extended cross-domain evaluation and physical robot validation in progress
Abstract:When should a robot look at the cup, and when at the portafilter? Spatial-reasoning benchmarks evaluate whether a model can describe a scene, but manipulation needs something more specific: the ability to acquire the particular evidence that makes the current command executable. A view that is informative for one task can be useless for another, even at the same workstation. We introduce Seeing What Matters, a benchmark for task-conditioned actionable spatial belief in simulated tabletop manipulation. Across three domains–coffee workstation, drawer/cabinet, and microwave/appliance–the benchmark provides 860 frozen evaluation cases; expanding across methods, budgets, and inference repeats yields more than 115,000 evaluator episodes under two protocols: closed-loop actionability and same-object functional out-of-distribution generalization. Each case pairs a set of candidate RGB views with a manipulation command, shared scripted skills, and hidden postconditions, so the benchmark directly measures whether a view-selection policy acquires the observations its command actually needs. We also provide a diagnostic reference pipeline that maintains a spatial-functional belief state and ranks views by missing evidence. At primary budget settings, this pipeline reaches 98.8% closed-loop and 98.9% functional-OOD success; ablating the functional or command-binding components collapses performance to near zero. Seven vision-language models, evaluated on the same cases, reach only 26-43%. Replacing the reference pipeline’s coverage metadata with frozen RGB-derived scores preserves 98.5%/98.9% success; shuffling the visual scores collapses it to 15%/11%, confirming that the result depends on image content. These gaps suggest that spatial belief for manipulation should be evaluated not by scene-level completeness, but by whether it acquires the evidence the current command requires.
[CV-11] AnimWeb-Bench: Benchmarking Animated Webpage Code Generation from Videos
链接: https://arxiv.org/abs/2611.09769
作者: Zhongyi Zhang,Jingyu Xiao,Shuqing Li,Michael R. Lyu
类目: oftware Engineering (cs.SE); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Animation is pervasive on the Web, the CSS animation property alone appears on over three-quarters of webpages. Yet existing webpage code generation benchmarks target static screenshots, user-triggered interactions, or isolated motion graphics, leaving the ability of multimodal large language models (MLLMs) and coding agents to reconstruct complete animated webpages largely unexplored. To fill this gap, we introduce AnimWeb-Bench, the first benchmark for animated webpage code generation from videos. It comprises 124 real-world webpages spanning 7 application domains and 33 animation types, with 91.9% of pages combining multiple types, each paired with a reference video and the static assets of the original webpage. We further propose a human-aligned, two-dimensional evaluation protocol: Static Appearance Fidelity measures structure, text, color, and visual similarity on aligned keyframes, while Animation Fidelity tracks UI components across frames and compares their position, size, velocity, and appearance with those in the reference video. Evaluating 14 MLLMs and three coding agents reveals: (1) reproducing webpage animations is substantially more difficult than reproducing static appearance; (2) explicit frame inspection and render feedback improve the reconstruction quality of coding agents; and (3) rapid, continuous changes are harder to reproduce than discrete state transitions. Our code and dataset are available at this https URL .
[CV-12] A lightweight convolutional neural network for target detection in an edge environment
链接: https://arxiv.org/abs/2611.09756
作者: Hughes Perreault,Ola Ahmad,Daniel Czuboka
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Presented at GVSETS 2023, Novi, MI. Distribution A: approved for public release
Abstract:This paper proposes a novel, lightweight convolutional neural network (CNN) designed for edge device deployment. The CNN is developed to recognize and classify threats or targets of interest. The model is equipped with a real-time hotspot detection algorithm, which enables it to process information quickly and accurately, resulting in improved inference time of up to 20 fps and a detection range of more than 400 meters with typical thermal situational awareness sensors. The proposed CNN has been rigorously tested and validated to demonstrate its high accuracy, sensitivity, and low false positive rate, providing a reliable solution for edge devices to detect potential targets. Additionally, the lightweight design of the CNN allows for easy deployment, making it an ideal solution for extending the security bubble for edge devices. The proposed model can be used for a variety of applications, including surveillance, security, and object recognition.
[CV-13] Zero-Shot Benchmarking of Unified Multimodal Models for Shallow-Water UAV Image Distortion Assessment and Restoration
链接: https://arxiv.org/abs/2611.09740
作者: Dario Scilla,Omar A. Lopez,Mattia Rosso,Kasper Johansen,Matthew F. McCabe
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Shallow-water UAV imagery provides centimeter-scale observations of benthic environments, but water-induced distortions complicate interpretation. To our knowledge, we present the first systematic zero-shot benchmark of unified multimodal large language models (MLLMs) for both assessing water-induced distortions and restoring shallow-water imagery within the same model framework. We evaluate five open-weight unified MLLMs with frozen pretrained weights and without access to paired reference images or descriptions of the true seabed. The benchmark contains 1,000 image pairs: 600 semi-synthetic UAV scenes created by applying synthetic water distortions to real shallow-water imagery and 400 fully synthetic through-water scenes. Three independent human annotations cover five distortion types. We evaluate four assessment prompts (P0-P3), from direct instruction to structured task decomposition, on all 1,000 images and four diagnosis-guided restoration prompts (R0-R3) on a common 100-image subset. Assessment varies substantially by model and prompt, with systematic prediction biases and a best pooled balanced accuracy of 0.654 for distortion presence. The best pooled PSNR gain is +0.255 dB, no unified MLLM improves pooled LPIPS, and some outputs alter scene structure or hallucinate content. Two restoration-only baselines also show limited transfer: InstructIR achieves a pooled PSNR gain of +0.040 dB, while PromptFix degrades all four pooled metrics under every prompt. These results show that recognizing shallow-water degradation does not imply reliable physical restoration and provide a reproducible baseline for future work.
[CV-14] riNetra: Thermal-Reliability-Guided RGB-Thermal Fusion and Tracking for Surveillance
链接: https://arxiv.org/abs/2611.09725
作者: Sadique Khatib,Aditya Ahirrao,Vedant Sonawane,Disha Kapse,Y. D. Bhise
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Visible cameras fail at night and thermal sensors fail under saturation, blur or loss of contrast, yet most RGB-thermal (RGB-T) detectors mix the two streams with fixed weights and cannot tell when one modality has become unreliable. TriNetra is a reliability-aware surveillance framework built around this problem. The proposed solution computes a frame-level Thermal Reliability Coefficient (TRC) from five label-free image cues, uses it to gate the thermal features of a dual-branch YOLOv8m detector whose visible branch, neck and head are frozen and COCO-pretrained, and passes the same score to a ByteTrack-based tracker that feeds a rule-based behaviour analyser. On the M3FD benchmark (2,940/630/630 split) the detector reaches 87.6% precision, 79.4% recall, 70.0% mAP@50 and 47.8% mAP@50:95 on the test split, with 17.0 M of its 42.9 M parameters trained. TRC stays near 0.78 on clean frames (standard deviation 0.012), decreases with the severity of thermal saturation and blur, reaching 0.53 and 0.69 at full severity, and is insensitive to RGB-thermal misalignment. With identical weights, the adaptive gate matches a fixed gate on clean data (70.0% versus 69.8% mAP@50) but keeps 43.6% mAP@50 when the thermal stream is corrupted, against 38.4%, 29.2% and 25.0% for a fixed gate, full thermal admission and a closed gate; the 5.1-point gain over the fixed gate has a 95% bootstrap confidence interval of +3.6 to +6.5 points. In tracking, the TRC-gated association performs on par with ByteTrack on a synthetic benchmark with exact ground truth and on 14 M3FD clips, and is slightly less stable under thermal saturation, so its measured benefit is confined to the detector. The limitations of the current evaluation are reported together with the results.
[CV-15] Evaluating Anterior Cruciate Ligament Risk of Injury Using a Graph Convolutional Autoencoder
链接: https://arxiv.org/abs/2611.09724
作者: Athanasios Babouras,Paul A. Martineau,Thomas Fevens
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 9 figures
Abstract:Current solutions for assessing the risk of anterior cruciate ligament (ACL) injury typically rely on expensive motion-capture labs that require substantial resource investment or subjective evaluations of specific athletic stress tests by experts. The Drop Vertical Jump (DVJ) and the Single Leg Hop (SLH) are two examples of such tests typically used to track progress in ACL rehabilitation. In this paper, we develop an unsupervised learning model that runs entirely on a smartphone for two tasks, based on an autoencoder architecture with multiple layers of graph convolutional networks, which utilizes pose estimation data to determine a person’s injury risk status. Our model accurately predicts athletes at high risk of injury for both athletic stress tests, with average sensitivities and specificities of 78% and 78% for the DVJ, and 75% and 74% for the SLH. Our findings support the development of a widely accessible smartphone solution for real-time screening of ACL injury risk.
[CV-16] Direct Regression and Geometric Recovery for Event-based Spacecraft Pose Estimation
链接: https://arxiv.org/abs/2611.09717
作者: Pruthvi Nandan Janga,Kashyap Hegde Kota,Bharatesh Chakravarthi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Event cameras offer high temporal resolution and dynamic range for uncooperative spacecraft pose estimation, but models trained on synthetic event streams must generalize to the noise, background activity, and contrast variation of physical sensors. We present a comparative systems study of two independently developed approaches to synthetic-to-real, event-based 6-DoF spacecraft pose estimation. The first is a modular pipeline that combines learned keypoint detection and unlabeled target-domain adaptation with geometry-based pose recovery and temporal smoothing. The second directly regresses translation and rotation from global and cropped event representations using an apparent-scale cue, physics-motivated synthetic corruptions, and feature-consistency regularization. We evaluate both systems under the same SPARK 2026 hidden-test protocol on the real-event split of SPADES, with predictions for all test frames. The results reveal a component-wise reversal concealed by the aggregate challenge score: direct regression reduces orientation error by 22.7% (1.1767 to 0.9101 rad) and obtains a 13.4% lower composite score, whereas the modular pipeline reduces relative translation error by 78.7% (0.1335 to 0.0285). Archived configurations and a qualitative failure case identify plausible explanations for these complementary error profiles, without isolating their causes. These findings motivate a testable hybrid design that combines direct orientation estimation with geometry-constrained translation recovery.
[CV-17] PoseRieFlow: Riemannian Flow Matching for Whole-Body Text-to-Pose Generation
链接: https://arxiv.org/abs/2611.09692
作者: Jan Spindler,Bardia Ardakanian,Patrick Stotko,Reinhard Klein
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We present PoseRieFlow, a compact flow-matching framework for text-conditioned generation of whole-body 3D human poses. Prior text-to-pose methods either neglect articulated hands or model rotations in Euclidean space despite their non-Euclidean geometry. PoseRieFlow instead formulates generation directly on the product manifold of SMPL-H joint rotations, learning a text-conditioned vector field that transports samples from an isotropic \mathrmSO(3) Gaussian prior to the data distribution along Riemannian geodesics. A part-tokenized Diffusion Transformer aggregates the 52 joints into eight body-part tokens, allocating a dedicated token and nonlinear projector to each hand, and restores per-joint resolution through part-specific decoders, shortening the sequence while improving kinematic fidelity over per-joint tokenization. With only 21.6M trainable parameters, PoseRieFlow supports expressive SMPL-H generation with detailed hand articulation at a fraction of the scale of recent LLM-based generators. On PoseScript, it attains the lowest whole-body kinematic error among all hand-producing methods, with consistent kinematic gains from the Riemannian formulation over its Euclidean counterpart at every practical inference budget, while distributional and retrieval metrics remain comparable between them.
[CV-18] Dual Architecture Fusion Vehicle Lidar 3D Target Detection
链接: https://arxiv.org/abs/2611.09664
作者: Muyun Li,Weizheng Jin,Yuan Sun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages, 6 figures. Published in CAIBDA 2026
Abstract:In this paper, a 3D backbone network (SpconvTransformer) based on the fusion of convolutional architecture branch and Transformer architecture branch is proposed. The convolution branch uses 3D sparse convolution and submanifold convolution to learn local voxel features and semantic information, and the Transformer branch uses the self-attention mechanism to model the remote relationship between voxels, uses the broad receptive field to capture rich context information, and complementing the advantages of the two architectures by fusing the local features extracted from the two branches and the global features solves the problem of limited receptive fields of 3D convolutional neural networks. The experimental results show that the average accuracy of SpconvTransformer in 3D object detection and bird’s-eye view detection tasks reaches 81.45% and 88.09%, respectively, and the average directional similarity reaches 91.10%, which is significantly better than other existing algorithms.
[CV-19] Video2Asset: Online Construction of Persistent 3D Street-Scene Assets
链接: https://arxiv.org/abs/2611.09661
作者: Zihan Zheng,Chunlei Cai,Zhenlong Wu,Linjun Li,Yichen Gong,Xiaoyun Zhang,Qiang Hu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Online 3D street-scene asset construction turns multi-view video into a continually updated scene of independently editable objects. The central challenge is to accumulate partial, intermittent observations under motion and occlusion without fragmenting identities or mixing evidence across objects. To address this challenge, we propose Video2Asset, an online framework for constructing persistent 3D assets from multi-view video. Using occupancy as an intermediate representation, we construct identity-aware observations with Historical Carrier Guidance for instance decoding and an identity encoder fusing instance-aligned 3D features with multi-view appearance. Accumulated-Evidence Association compares current observations with identity, geometric, and appearance evidence accumulated within asset records to determine identity inheritance and update world-coordinate states. We instantiate meshes from an asset library using these states and combine them with an occupancy-derived static background to support scene visualization, mesh replacement, object deletion, and trajectory editing. Experiments on several datasets demonstrate improvements in temporal identity consistency and asset quality across multiple occupancy models, including a 15.4% reduction in asset fragmentation on nuScenes with SGE.
[CV-20] Decoupled Mode Connectivity for Base-to-Novel Generalization in Vision-Language Models NEURIPS2026
链接: https://arxiv.org/abs/2611.09655
作者: Imad Eddine Marouf,Khalid Oublal,Enzo Tartaglione,Stéphane Lathuilière
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026. 46 pages (10 main + references, appendix and checklist), 15 figures. Project page: this https URL Code: this https URL
Abstract:Prompt learning adapts vision-language models such as CLIP by optimizing a short learnable prompt. The cross-entropy objective pulls this prompt toward base-class specialization, whereas generalization to unseen classes benefits from staying close to CLIP’s zero-shot feature space. Because both objectives act on the same prompt, their gradients conflict at convergence, and in our experiments no single-prompt loss design clearly improves on the cosine baseline. We propose Decoupled Mode Connectivity (DMC), which gives each objective its own prompt. Starting from a base-trained prompt, DMC trains a specialization prompt and a generalization prompt, and merges their text features with a fixed coefficient at test time, at the inference cost of a single prompt. A linear mode connectivity (LMC) corridor in text-feature space keeps the classification loss low for every classifier interpolated between the two endpoints, and a Visual Anchor KL term preserves CLIP’s zero-shot class-similarity distribution of each image during specialization. We further argue that such regularizers should satisfy class-permutation invariance (CPI), and show via Fano’s inequality, in an idealized model, that violating CPI lower-bounds the drop in novel accuracy by a term proportional to the mutual information between the prompt and the base-class labels. Across 11 datasets, DMC improves the average harmonic mean of base and novel accuracy over MergeTune for both CoOp and KgCoOp, and its interpolation curves pass beyond the single-prompt operating point, shifting the base-novel frontier rather than trading along it. Code is available at this https URL. Comments: Accepted at NeurIPS 2026. 46 pages (10 main + references, appendix and checklist), 15 figures. Project page: this https URL Code:this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2611.09655 [cs.CV] (or arXiv:2611.09655v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.09655 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Imad Eddine Marouf [view email] [v1] Fri, 6 Nov 2026 16:04:41 UTC (437 KB) Full-text links: Access Paper: View a PDF of the paper titled Decoupled Mode Connectivity for Base-to-Novel Generalization in Vision-Language Models, by Imad Eddine Marouf and 3 other authorsView PDFHTML (experimental)TeX Source view license Additional Features Audio Summary Current browse context: cs.CV prev | next new | recent | 2026-11 Change to browse by: cs References Citations SciXGoogle 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)
[CV-21] Breaking the Privacy Promise: Gradient Inversion Attacks on Federated Low-Rank Adapted Vision Transformers WACV2027
链接: https://arxiv.org/abs/2611.09652
作者: Ritik K. Badiya,Siladittya Manna,Ashmit Sinha,M Yashwanth,Anirban Chakraborty
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
备注: Accepted at WACV 2027
Abstract:Parameter-efficient fine-tuning (PEFT) has been increasingly adopted to fine-tune pretrained models with low computational cost and is therefore also preferred in the federated learning paradigm to reduce computational and communication overhead. In this work, we investigate the sparsely explored problem of the risk of private data leakage in LoRA-adapted Vision Transformers (ViTs), where low-rank updates weaken the gradient signal available to a server with access to individual client updates. We show that quality of reconstruction is limited by two coupled ambiguities: reduced information in LoRA gradients and missing spatial cues from frozen positional embeddings. To address this, we propose a spatially anchored generative inversion framework that incorporates server-prescribed trainable positional embeddings into ViT blocks during fine-tuning at clients. The resulting positional gradients act as layer-wise spatial anchors, providing additional positional cues, while the generative prior keeps reconstructions on a realistic image manifold, reducing duplicate-data and patch-permutation artifacts, which produce similar gradients as those produced when the target client images are used. Adding our proposed spatial anchoring to GIAS improves PSNR by 10.14 dB, SSIM by 0.45, and LPIPS by 0.098 on CIFAR-10, as well as PSNR gains of 7.83 and 11.74 dB on high-resolution CelebA and CARS datasets, respectively, in the FedSGD setting. Such gains in reconstructed image quality metrics are also maintained at higher batch sizes and in FedAvg settings. These results indicate that LoRA-based PEFT does not, by itself, prevent data leakage in FL, and carefully incorporated positional embeddings can help in exposing private information.
[CV-22] Encoding Memory in World Models with Test-Time Model-Based Reinforcement Learning
链接: https://arxiv.org/abs/2611.09625
作者: Maxime Burchi,Fayaz Ali Dharejo,Radu Timofte
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Model-based reinforcement learning (MBRL) approaches have demonstrated strong performance across diverse domains in recent years, learning accurate world models from agent trajectories. However, despite this progress, current approaches still struggle to surpass human performance on tasks that require long-term memory and planning. In this work, we address the problem of long-term memory in MBRL by moving beyond the commonly studied paradigm of architecture design. We introduce Test-Time Model-Based Reinforcement Learning (TT-MBRL), using gradient descent at test time to encode the memory of episode observations into world model parameters. This enables the agent to learn a distinct world model for each episode, improving memory capabilities compared to traditional approaches that solely rely on contextual memory. We first establish a proof of concept on a simplified two-dimensional variant of the Memory Maze environment, which we refer to as Memory Maze 2D, showing consistent improvements across different world model architectures. We then demonstrate the effectiveness of TT-MBRL in the 3D domain, achieving new state-of-the-art results on the Memory Maze benchmark. Finally, we show that TT-MBRL maintains strong performance on non-memory benchmarks such as DeepMind Control and Atari 100k.
[CV-23] CRTrack-Cap: A New Benchmark for Cross-View Dense Video Object Captioning
链接: https://arxiv.org/abs/2611.09617
作者: Xiangqun Zhang,Likai Wang,Zekun Qian,Rui Zhou,Ruize Han,Wei Feng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Dense video object captioning (DVOC) localizes, tracks, and describes object trajectories within a video. Existing methods process each camera view independently, producing separate camera-local tracklets and captions for the same person across cameras. These outputs lack a unified description and leave complementary appearance evidence across views unused. We therefore formulate cross-view DVOC (CVDVOC), which associates these tracklets into scene-level global trajectories and generates one caption for each trajectory from synchronized views. To support this task, we construct CRTrack-Cap from CRTrack by auditing text-identity links and converting usable referring expressions into canonical captions. It retains 425 global trajectories and 838 local tracklets across 13 scenes, with canonical captions for 335 identities. We adapt CHOTA to the cross-view setting as CV-CHOTA to jointly evaluate tracking and caption quality. Our reference framework combines View-Exclusive Cross-Camera Association (VECA) for grouping tracklets with Conflict-Evidence Gating (CEG) for integrating caption evidence using its reliability and disagreement across views. Against three single-view methods adapted with MvMHAT for cross-view association, our framework achieves CV-CHOTA scores of 83.43 and 33.48 on the scene-disjoint in-domain and cross-domain tests, respectively. Together, CRTrack-Cap and our framework provide a benchmark and a strong baseline for describing the same person consistently across camera views.
[CV-24] Dual-Region Slice Aggregation for CT-Based Renal Tumor Classification
链接: https://arxiv.org/abs/2611.09599
作者: Yuan Liang,Sourav Bhattacharjee,Abraham Campbell
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to IEEE ISBI 2027
Abstract:Clear cell renal cell carcinoma (ccRCC) is the most common kidney cancer in adults. We present a region-aware architecture for CT-based ccRCC classification. Tumor and 0–10 mm peritumoral context are pooled separately from shared frozen DINOv3 feature maps. A shared position-aware slice aggregator precedes fixed training-only principal component analysis (PCA). The resulting representations are fused with radiomics using two scalar feature gates. Evaluation on KiTS23 and 111 external cases yields internal test and external AUCs of 0.8078\pm0.0154 and 0.9094\pm0.0015. Compared with matched single-tumor readout, mean AUC increases by 1.50 and 2.34 percentage points, respectively. Early-PCA and reconstruction controls yield lower internal mean AUCs, while external AUC is similar for early and late PCA. These findings highlight the joint role of regional readout and slice aggregation in adapting frozen visual features to renal CT.
[CV-25] MSR-ABC: Complementary Feature Fusion for Clear Cell Renal Cell Carcinoma Classification
链接: https://arxiv.org/abs/2611.09583
作者: Yuan Liang,Sourav Bhattacharjee,Abraham Campbell
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to IEEE ISBI 2027
Abstract:Clear cell renal cell carcinoma (ccRCC) classification from computed tomography (CT) may benefit from representations that capture different aspects of tumour phenotype. However, simply concatenating reconstruction-derived features, image–text embeddings, and radiomic descriptors may not fully exploit their complementary information. We propose MSR-ABC, a global–local framework integrating MedVAE, MedSigLIP, and radiomics. Spatially matched local image descriptors interact through cross-representation attention, while radiomic auxiliary supervision encourages the pooled local representation to retain quantitative tumour attributes. A gated residual mechanism integrates local information with global summaries. Area-aware pooling and visual and soft-text prompting complete the framework. We evaluate discrimination using a patient-level split of 396 KiTS23 cases and an external cohort of 111 TCGA/TCIA cases, reporting mean AUROC over three training seeds with 95% patient-bootstrap confidence intervals. MSR-ABC achieves mean AUROCs of 0.806 internally and 0.879 externally. Internally, it exceeds global MedVAE–radiomics and three-source concatenation by 3.35 and 3.61 AUROC percentage points, respectively. Removing any source lowers mean AUROC in both cohorts, with the largest decreases following radiomics removal. Individual removal of cross-representation attention, auxiliary supervision, or gated fusion also lowers internal mean AUROC. These findings are consistent with complementary contributions within the integrated architecture and support further investigation of structured representation fusion for non-invasive renal tumour characterisation.
[CV-26] Video Stitching from Multiple Moving Cameras NEURIPS2026
链接: https://arxiv.org/abs/2611.09570
作者: Shira Ifergane,Roy Amoyal,Shahar Benishay,Oren Freifeld
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to NeurIPS 2026
Abstract:We address geometric video alignment for the same scene captured by multiple freely moving cameras. Existing methods, which are based on image stitching or video stabilization, are typically limited to two cameras or must assume the restrictive setting of a fixed linear ordering of the cameras. We propose MViS, a lightweight test-time optimization framework for joint multi-view alignment with temporal consistency. MViS uses a spatial-transformer-inspired net to jointly process frames from all views within a short temporal window and predict per-view, per-time homographies. Optimization proceeds in two stages: geometry-guided alignment followed by appearance-based refinement, with both stages enforcing multi-view and temporal consistency across the sequence. The method requires no scene-specific hyperparameter tuning, handles arbitrary and time-varying view-intersection graphs, and remains robust under rapid camera motion. On a standard two-view video benchmark, MViS matches or surpasses the state of the art. To facilitate evaluation in more complex scenarios, we introduce an annotated dataset with more than two freely moving cameras and establish the first benchmark for this setting. Project page: this https URL .
[CV-27] Multimodal Training with Knowledge Traction for Video Anomaly Detection
链接: https://arxiv.org/abs/2611.09536
作者: Xiao Tan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 15 pages, 10 figures, 8 tables
Abstract:The task of weakly supervised video anomaly detection is challenging. Existing approaches mainly model video frame features and determine whether anomalies are present based on their magnitudes. In this work, we propose a novel method that incorporates external knowledge into training to improve model performance, robustness, and generalization. We leverage textual prompts and a large language model (LLM) to establish prior information for the specific task, and employ dictionary learning to normalize and compress this information. Moreover, during experiments, we observed inconsistencies in the textual descriptions extracted from similar video frames by image caption models. To address this, we devise a method that leverages an LLM and auxiliary information to ensure consistency in the textual descriptions of similar frames within a video. Finally, we introduce a feature processing approach that incorporates prior knowledge and cross-modal attention mechanisms. This approach learns the relationships and characteristics among different modalities to construct more accurate and reliable feature representations, thereby improving the accuracy and stability of downstream tasks. Experimental results demonstrate that our method surpasses other state-of-the-art approaches.
[CV-28] ExQuery: Explicit Structured Query Priors for Temporal 3D Object Detection BMVC2026
链接: https://arxiv.org/abs/2611.09527
作者: Honglin Sun,Yifei Zhang,Junyang Zhang,Kenji Hashimoto
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: BMVC 2026
Abstract:Sparse query-based 3D detectors commonly encode heterogeneous semantic, geometric, and motion information into unified latent queries, including when explicit 3D anchors are used. We introduce ExQuery, an explicit structured query framework for temporal 3D object detection. Each decoded query is organized into a shared semantic context slice and physical-state-specific slices for center, dimensions, orientation, and velocity. This structure is preserved from prior encoding to branch-specific regression and is reused to reconstruct temporal queries: predicted physical states are transformed by ego motion, classification scores form soft semantic priors, and the reconstructed query is fused with the stored latent instance feature through gated affine modulation. Experiments on the nuScenes benchmark show that our method achieves competitive detection performance, and ablation studies validate the effectiveness of our proposed modules.
[CV-29] RadarDINO: Distilling a Visual Foundation Model into mmWave Radar for Human Mesh Reconstruction
链接: https://arxiv.org/abs/2611.09524
作者: Furkan Mert Algan,Zhifan Ni,Eckehard Steinbach
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Millimeter-wave (mmWave) radar offers a privacy-preserving and illumination-robust alternative for human mesh reconstruction (HMR). However, high-fidelity radar HMR remains challenging. Sparse point cloud-based methods suffer from information loss and missed detections, while raw radar tensor methods like RT-Mesh lack fine-grained semantics and explicit local body modeling. In this paper, we propose RadarDINO, a radar HMR framework that distills visual semantics from a DINOv3 teacher into 3D radar representations and introduces explicit geometric supervision. DINOv3 descriptors are transferred to radar voxels via calibrated RGB-radar geometry, providing voxel- and body-part-level supervision for part-aware radar representations. We further incorporate vertex- and joint-losses to directly constrain reconstructed geometry. On the M4Human dataset, our method outperforms the SOTA baseline across all three evaluation splits.
[CV-30] Embedding Predictive Adaptation and Refinement for Multi-Modal Industrial Anomaly Detection WACV2027
链接: https://arxiv.org/abs/2611.09521
作者: Jaeho Cho,Jinsik Bang,Taehwan Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV 2027
Abstract:Industrial Anomaly Detection (IAD) has evolved from 2D image-based approaches to multi-modal methods that leverage the complementary nature of RGB images and 3D point clouds. Among these, memory bank-based methods, which identify anomalies by the discrepancy between each query patch feature and its nearest normal prototype, have proven particularly effective. However, two limitations remain: (i) employing frozen pretrained encoders yields insufficient discrimination between normal and anomalous features due to the domain gap, while existing contrastive learning-based adaptation struggles to capture the spatial relationships required for precise localization; and (ii) pixel-level anomaly maps interpolated from patch-level distance scores remain coarse, lacking fine-grained detail. To address these limitations, we propose EPAR (Embedding Predictive Adaptation and Refinement). We adapt encoders by predicting the embeddings of masked regions to capture spatial relationships, together with patch-level Variance-Invariance-Covariance (VIC) regularization, resolving the poor discriminability between normal and anomalous features caused by the domain gap. We further propose a Noise Suppression Network that refines coarse maps by learning the relational discrepancy between each patch feature and its nearest normal prototype, rather than the anomaly’s appearance. This relational signal remains valid for unseen real defects, despite being learned from synthetic anomalies. EPAR demonstrates strong practical applicability, excelling in few-shot settings and improving AUPRO@1% on MVTec 3D-AD and Eyecandies by +1.8% and +1.9%, respectively.
[CV-31] V-STORM: A Structure-Aware Grounded Chain of Thought Dataset for Faithful and Fine-Grained Spatio-Temporal Video Reasoning
链接: https://arxiv.org/abs/2611.09511
作者: Dian Jiao,Yuqian Yuan,Jiaqi Zhu,Mingjian Gao,Siwen Tan,Yunqi Cao,Jinhui Li,Sijing Li,Zhenxuan Fan,Haoyang Wang,Wenqiao Zhang,Siliang Tang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 9 Pages, 5 figures
Abstract:While MLLMs can often identify what happens in a video, they still struggle with fine-grained spatio-temporal reasoning, which requires precise grounding when and where the supporting evidence appears and composing such evidence into a coherent reasoning chain. This challenge is exacerbated by the lack of scalable data construction and evaluation pipelines that provide granular, verifiable supervision linking reasoning steps to explicit spatio-temporal this http URL address this gap, we present V-STORM, a structure-aware framework that transforms videos into Spatio-Temporal Scene Graphs (STSGs) and derives Grounded Chain of Thought (GCoT) from validated evidence paths, making each reasoning step traceable to structured and verifiable spatio-temporal this http URL this framework, we construct a 260K-sample training set and a 3K-sample human-verified benchmark, both featuring GCoT annotations that link reasoning to explicit spatio-temporal this http URL reveal a clear separation between evidence localization and evidence composition: current MLLMs can often identify relevant spatio-temporal cues, yet still fail to integrate them into correct multi-step reasoning, exposing a bottleneck that answer-only evaluation largely obscures. Moreover, fine-tuning with V-STORM substantially improves both answer accuracy and grounding faithfulness, validating its effectiveness as supervision for faithful and evidence-grounded video reasoning. Our dataset, benchmark, and source code will be publicly released.
[CV-32] FusionCLIP: Multimodal Contrastive Pretraining for Fusion Plasma Diagnostics
链接: https://arxiv.org/abs/2611.09493
作者: Qiang Chen,Xiao Wang,Qingquan Yang,Hao Si,Zikang Yan,Guosheng Xu,Jin Tang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Visible-light videos and multivariate diagnostic signals provide complementary observations of tokamak discharges. However, existing fusion pretraining methods, primarily learn representations by reconstructing or predicting observations, leaving the correspondence between these modalities underexplored. Representation learning is further challenged by the fact that plasma-related regions occupy only part of each video frame and diagnostic variables have substantially different sampling rates. We propose FusionCLIP, a multimodal contrastive pretraining framework. Spatiotemporal Stability Selection retains informative visual tokens, while a Temporal Query Encoder integrates asynchronous diagnostic signals using their observation timestamps. Contrastive learning then aligns the two modality representations. We construct EAST Multimodal Diagnostics (EAST-MMD), comprising 1307 EAST discharges, and evaluate the model on three independently fine-tuned tasks: disruption prediction, edge-localized mode (ELM) classification, and H_98 regression. FusionCLIP achieves an aggregate score of 89.07%, exceeding the best general-purpose pretraining configuration and an in-domain contrastive pretraining configuration without Spatiotemporal Stability Selection by 6.27 and 2.12 percentage points, respectively. Ablations and attention visualizations further support the contributions of its key components.
[CV-33] Left-Right-Consistent Iterative Stereo Matching via Cross-Injection WACV2027
链接: https://arxiv.org/abs/2611.09458
作者: Manuel Knott,Diogo Luvizon,Daniel Scharstein,Elias Vansteenkiste
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV 2027
Abstract:State-of-the-art stereo matching models predominantly predict the disparity of only the left (reference) view. However, applications such as per-eye AR/VR rendering, novel view synthesis, or two-view depth fusion require disparity for both views, and need those two maps to agree. The usual workaround of running the network a second time on the horizontally flipped and swapped pair gives two independent predictions that can be inconsistent, especially on transparent surfaces, in textureless regions, and at occlusion boundaries. We introduce cross-injection (CI), a lightweight extension for iterative (refinement-based) models that can be added to any pretrained left-only iterative matcher. It couples the two passes by injecting the explicit, signed left-right inconsistency as an additional input cue at every step of the refinement stage. We empirically show that CI substantially reduces left-right disagreement on every benchmark (more than halving it on Booster for FoundationStereo), while largely preserving sharp object boundaries. Furthermore, using two state-of-the-art stereo matching models, we show that CI also reduces the single-view disparity error, making it a worthwhile extension even if only left disparity is needed.
[CV-34] Overview of the ImageCLEFmedical 2026 GANs task: Exploring fingerprint detection privacy leakage and synthetic medical image generation
链接: https://arxiv.org/abs/2611.09455
作者: Alexandra-Georgiana Andrei,Mihai Gabriel Constantin,Mihai Dogariu,Dzmitry Karpenka,Yuri Prokopchuk,Ahmedkhan Radzhabov,Dan-Cristian Stanciu,Liviu-Daniel Ştefan,Vassili Kovalev,Henning Müller,Bogdan Ionescu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: CLEF 2026 Working Notes, 21-24 September 2026, Jena, Germany
Abstract:The 2026 ImageCLEFmedical GANs Task Exploring Fingerprint Detection, Privacy Leakage, and Synthetic Medical Image Generation continues to investigate privacy and security concerns around using patient data to generate synthetic medical images, now in its fourth edition. This edition introduces three complementary subtasks: the first extends prior editions by asking participants to detect which real images were used in training a Generative Adversarial Network to produce given synthetic outputs; the second is a novel subtask requiring participants to identify which latent vector produced each synthetic image, probing the faithfulness and reversibility of the latent-to-image mapping; the third asks participants to generate privacy-preserving CT slices, quantifying the trade-off between image fidelity and patient privacy. Ground-truth annotations and benchmark datasets of real and GAN-generated lung CT slices are provided for all subtasks. Evaluation is based on Cohen’s Kappa for Subtask 1, matching accuracy for Subtask 2, and a combination of FID and Privacy Preservation Score (PPS) for Subtask 3. Overall, 37 teams registered for the task, with 10 teams submitting runs across subtasks and 6 teams submitting working notes papers. This paper presents an overview of the task setup, datasets, evaluation metrics, and summarises the approaches and results of participating teams.
[CV-35] he Role Of Recurrent Refinement In Medical Image Segmentation
链接: https://arxiv.org/abs/2611.09436
作者: Joao D.S. Marques,David Calhas,Arlindo L. Oliveira
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:While feed-forward segmentation networks perform well on large cohorts, they degrade in constrained clinical settings with few training samples and imaging artifacts. Human vision, by contrast, remains robust under these conditions, a property partly attributed to the recurrent and feedback connectivity of the visual cortex, which refines visual information over time. Inspired by this, we append lightweight recurrent refinement blocks to segmentation backbones, namely Conditional Random Fields (CRFs), modern Hopfield networks and convolutional GRUs, and study their behaviour as the number of iterations increases at test time (including beyond the training horizon). Evaluated on two open-source datasets, Kvasir- SEG and ARCADE, our recurrent blocks add less than 1.5% parameters and yield gains on clean data (up to 1.4 pp) and under noise (up to 17 pp) while reducing HD95. Code is available at this https URL .
[CV-36] A Bridge from Paste to Place: Physics-Guided One-Step Image Composition via Object-Scene Interaction
链接: https://arxiv.org/abs/2611.09387
作者: Hongzhen Li,Youwei Pan,Yingfang Zhu,Sen Zhu,Fengjie Zhu,Leilei Cao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Realistic image composition must make a pasted object appear naturally present in a scene while preserving the fine details of both foreground and background. Reference-based insertion methods can change object pose and appearance, while foreground harmonization methods and shadow-specific synthesis methods each address only part of this task. An efficient, unified solution to this task remains an open challenge. We introduce Paste2Place, an end-to-end, physics-guided flow-based model that learns a direct, one-step latent bridge from pasted inputs to realistic composites, guided by explicit albedo-shading-residual decomposition and recomposition within the model. To support model training and evaluation, we develop a Blender rendering and validation pipeline to generate paired composites with physical annotations, and construct real-scene pairs from open-source object-removal data. This effort yields 86,273 training pairs and a 1,500-example benchmark. Extensive qualitative and quantitative experiments demonstrate that Paste2Place achieves substantial advantages in image composition quality and inference speed. Overall, our work provides a clear task definition and an effective solution for image composition, and fills a gap in large-scale, high-quality datasets and benchmarks for training and evaluating image composition.
[CV-37] SPSC: Zero-Shot Cross-Subject EEG-to-Image Reconstruction via Semantic Preservation and Subspace Calibration
链接: https://arxiv.org/abs/2611.09380
作者: Yutang Lu,Shuai Jiang,Fei Su,Zhicheng Zhao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Pretrained vision-language models and diffusion generators have driven rapid progress in brain-to-image generation, while EEG-based reconstruction is still constrained by low signal-to-noise ratio and cross-subject semantic shifts. Zero-shot cross-subject reconstruction poses even greater challenges: EEG semantic embeddings learned from source subjects deviate from the manifold of image embeddings of the unseen target subject, particularly when neither target labels nor target image embeddings are available. To address this issue, we introduce the Semantic Preservation and Subspace Calibration (SPSC) framework, which first uses Object Mask Contrastive Regularization (OMCR) with object identity for backend training, and then constructs Subspace Candidate Gate (SCG) to achieve label-free semantic correction for target subjects during inference. Specifically, OMCR regularizes the conditioning space with object-level positive samples to preserve object-level feature neighborhoods within the reconstruction prior and low-level branch, while SCG exploits training image prototypes and batch statistics of unlabeled target EEG to select candidates for each task within subspaces of image prototypes. Under the leave-one-subject-out (LOSO) setting across ten subjects on the THINGS-EEG dataset, SPSC improves the cross-subject reconstruction from 0.552 to 0.713 on CLIP and from 0.612 to 0.745 on Alex.5. These significant gains indicate that the semantically calibrated conditions can better support generation for the target subjects under label-free inference.
[CV-38] HiMEM: Hippocampus-Inspired Memory Consolidation Pruning and Composition for Vision-Language-Action Models in 3D Scenes
链接: https://arxiv.org/abs/2611.09355
作者: Boran Zhao,Yanbin Hu,Jun Ye,Zhenxian Hu,Yuxin Zhai,Jin Cui,Pengju Ren
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Memory plays a crucial role in long-horizon task execution and cross-task generalization for Vision-Language-Action (VLA) models operating in complex 3D scenes. However, existing VLA memory mechanisms suffer from memory noise, particularly in complex 3D scenes, as stored episodic memories contain extensive redundant details about scene layouts, object configurations, and robot-specific states, which obscure generalizable manipulation principles and hinder cross-task transfer. To address this challenge, we first introduce a hippocampus-inspired memory consolidation mechanism that synthesizes extensive episodic experiences into a compact set of essential semantic memories, enabling VLA models to derive generalizable principles from practical experience. Second, we propose an episodic memory pruning mechanism that removes redundant episodic memories whose contents are captured by consolidated semantic memories, thereby increasing effective memory capacity and improving long-horizon task execution. Finally, to handle complex task scenarios, we develop a flexible memory composition mechanism that jointly leverages multiple episodic and semantic memories to enhance cross-scenario generalization. Experiments on LIBERO, LIBERO-Plus demonstrate the effectiveness of HiMEM across different VLA backbones, improving the average success rate of pi_0 from 91.88% to 98.20% on LIBERO.
[CV-39] RegNC-Mamba3: Shared Structural Feature Learning Using Non-Causal State Space Principles for Multi-Modal Image Registration
链接: https://arxiv.org/abs/2611.09336
作者: Zhikang Li,Yan Wu,Xin Hu,Yi Dai,Ming Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:The main challenge of multi-modal image registration is shared structural feature extraction with higher accuracy and lower computational complexity. Compared with a State Space Model (SSM), the State Space Principles (SSP) formulation of Mamba-3 uses a more accurate discretization with lower truncation error while maintaining linear complexity. However, vanilla SSP is designed for natural language processing (NLP) and is not directly suitable for images. To address this issue, we propose a novel non-causal SSP (NC-SSP) with shared structure awareness and develop a coarse-to-fine registration method named RegNC-Mamba3. First, NC-SSP is designed to eliminate the dependence of the integral computation on the scanning path. By combining relative-weight normalization with aggregation over all circular scan paths, NC-SSP constructs a globally hidden state without long-range decay, which provides a wider receptive field and higher efficiency than the original Mamba-3. Second, NC-SSP is modified to strengthen the extraction of cross-modal structural features. When computing hidden states and feature interactions, we construct complex-valued sequences in NC-SSP by applying learnable Log-Gabor kernels, with different orientations to each head for frequency-domain modulation. This improves the perception of shared structural information across modalities and increases both registration accuracy and quality. Experiments on VIS-SAR and VIS-IR demonstrate that RegNC-Mamba3 outperforms the compared detector-based and semi-dense multi-modal registration methods.
[CV-40] Keyline Detection with YOLO: Runway Perception for Vision-Based Aircraft Landing WACV2027
链接: https://arxiv.org/abs/2611.09316
作者: Leander Fritsch,David Rügamer,Alexej Klushyn
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: WACV 2027
Abstract:Modern object detectors can predict keypoints as per-object geometric features, but keypoints are inherently vulnerable to occlusion and truncation. Object edges can provide complementary information, often remaining observable even when partially occluded or when their defining endpoints leave the field of view. We therefore introduce object keyline detection, the task of predicting a fixed, semantically ordered set of infinite lines per object instance. Unlike the outputs of general line-segment detectors, each keyline is associated with an object instance and identified by a semantic index, providing a line-level analog of keypoints. To solve this task, we extend YOLO with keyline prediction parameterized in continuous Hesse-normal form [\rho,\cos\theta,\sin\theta] , and introduce Object Keyline Similarity (OKLS), which serves as an evaluation metric, target assignment signal, and regression loss. We instantiate this formulation on runway perception in vision-based landing, a motivating example containing semantic linear features that remain partially observable throughout the approach. Experiments show that our representation achieves the highest detection accuracy among four evaluated line parameterizations, and keylines integrate directly into standard YOLO heads as additional output channels without degrading keypoint performance or adding meaningful latency.
[CV-41] Interpretable Context-Aware Multi-Label Retinal Classification with Dual-Hub Heterogeneous Graph Neural Network
链接: https://arxiv.org/abs/2611.09309
作者: Proshanto Chanda,Munib Mesinovic,Tingting Zhu,Abhirup Banerjee
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Multi-label retinal screening requires capturing co-occurring pathologies and clinical context, yet standard multimodal fusion remains opaque. We propose a dual-hub heterogeneous graph neural network (GNN) that couples spatial fundus features and structured clinical covariates through cross-modal message passing while providing inherent interpretability. Rather than relying on post-hoc saliency approximations, our architecture exposes multi-level decision pathways during forward aggregation: spatially biased self-attention captures lesion-level patch interactions, while modality-specific graph attention coefficients explicitly quantify the contribution of visual regions and clinical risk factors to each disease node. Evaluated on the BRSET dataset (16,266 photographs from 8,524 patients), the proposed model achieves 76.1% Macro-AUPRC across seven conditions, matching or exceeding image-only (75.9%) and ResNet-50+MLP fusion (75.9%) baselines, with largest contextual gains in macular edema (+3.5%) and diabetic retinopathy (+2.1%). Codebase is available at: this https URL.
[CV-42] 2I Models Rarely Invert Negation but Often Fail to Suppress Negated Concepts: Benchmarking Negation Suppression in Text-to-Image Generation WACV2027
链接: https://arxiv.org/abs/2611.09304
作者: Akansha Gautam,Aparna Bharati,Mayank Vatsa,Richa Singh
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at WACV 2027. Code and data: this https URL
Abstract:Text-to-image diffusion models often fail to suppress explicitly negated concepts, such as a lamp for “a desk without a lamp.” We introduce T2I-NSI, a benchmark and metric for measuring negation sensitivity in text-to-image generation. It pairs each negated prompt with a canonical positive counterpart and factorizes negation by linguistic form (lexical/semantic, morphological, syntactic) and target type (objects, attributes, relations, compositions, and nested cases). The benchmark contains 759,869 paired prompt atoms over a 5,082-object catalog, with 916,272 images (911,982 open-weights, 4,290 closed-source), scored by a Basic CLIP estimator and a Robust ensemble of CLIP, VQA and open-vocabulary detection. Negation sensitivity is low: open-weights generators score NSI = 0.038, while GPT-Image-1.5 and Gemini-2.5-Flash-Image reach about 0.09. The dominant failure is under-suppression rather than misdirection. Inverse interpretation is essentially absent (0.00% across over 450K atom-model evaluations), whereas models render the concept under both prompts on 8-19% of atoms and only partly suppress it on a further 9-25%; most remaining atoms are too weakly rendered to label. A blinded human study (420 atoms, nine annotators) corroborates the near-absence of inversion and the high persistence rate, and a matched-prompt comparison shows the open-versus-closed gap (~3.5x) is not a prompt-generator artifact. Code and data: this https URL
[CV-43] A Stochastic Latent Variable Generalization of Joint-Embedding Predictive Architectures WACV2027
链接: https://arxiv.org/abs/2611.09302
作者: Nikos Giakoumoglou,Paschalis Giakoumoglou,Tania Stathaki
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: WACV 2027
Abstract:Joint-Embedding Predictive Architectures (JEPAs) predict the latent embedding of a target signal from a context signal, with an auxiliary variable z conditioning the predictor. Despite its prominent role in the original JEPA blueprint, the content of z has been carried by a fixed deterministic token in every prior instantiation, collapsing the genuinely multi-modal distribution over plausible completions to a point estimate. We present S-JEPA, which realises z as a proper stochastic latent variable for masked image prediction: a learned posterior infers z from context and target during training, while a learned context-conditioned prior approximates the distribution of plausible completions at inference, with no decoders and no reconstruction. In its deterministic limit the objective reduces to the I-JEPA loss in functional form, generalising that objective rather than strictly containing the I-JEPA model. Our central finding is that stochasticity adds capabilities, error-correlated predictive uncertainty and diverse completions, at no cost to representation quality: under matched architectures and schedules S-JEPA improves over the deterministic I-JEPA baseline and exceeds LeJEPA at matched architecture, while additionally producing spatially structured uncertainty maps and semantically diverse completions, neither of which the deterministic predictor produces.
[CV-44] HMMF-Net: A Lightweight Hierarchical Multi-Frame Multimodal Fusion Network for Accurate Fish Feeding Intensity Assessment
链接: https://arxiv.org/abs/2611.09224
作者: Phi-Hung Hoang,Van-Manh Tran,Nam-Thuan Trinh,Thi-Thu-Hong Phan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 47 pages, 10 figures, 15 tables
Abstract:Automated fish feeding intensity assessment (FFIA) is important for precision aquaculture and feed cost optimization. However, unimodal methods are vulnerable to acoustic noise and visual disturbances, while existing multimodal frameworks often rely on computationally intensive 3D convolutions or Transformers, limiting their use on edge devices. To address these limitations, we propose HMMF-Net, a lightweight multimodal architecture that combines early multi-frame visual encoding and multi-axis acoustic modeling through a hierarchical bidirectional messenger cross-modal fusion (BMCF) mechanism. The visual branch captures temporal variations in feeding behavior by stacking temporally distant frames as a multi-channel input to an adapted 2D EfficientNetB0 backbone. The acoustic branch employs adaptive multi-axis residual units (AMARU) to extract temporal, frequency, and local time-frequency representations from log-mel spectrograms. BMCF progressively exchanges information between the two modalities at intermediate representation levels through compact messenger tokens, reducing the computational overhead of dense cross-modal attention. Evaluated on the AV-FFIA dataset using five-fold cross-validation, HMMF-Net achieves 96.77% accuracy and 96.88% F1-score on the held-out test folds, with 3.987M parameters and 1.002 GFLOPs. Compared with the best-performing visual configuration (95.49% accuracy), HMMF-Net reduces the classification error by 28.38%. Deployment on an NVIDIA Jetson Orin Nano using TensorRT with FP32 precision maintains comparable accuracy, with an inference time of 11.54 ms per 2-second sample and a model size of 18.81 MB. These results demonstrate that HMMF-Net combines effective multimodal representation learning with computational efficiency, enabling accurate and practical FFIA on resource-constrained edge platforms.
[CV-45] InstantHMR : Real-Time Human Mesh Recovery on Your Phone
链接: https://arxiv.org/abs/2611.09195
作者: Mohamed Adjel,Kahina Chalabi(LAAS-GEPETTO),Vincent Bonnet(LAAS-GEPETTO)
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Real-time 3D human mesh recovery from a single image on mobile devices is a challenging problem with many applications, including robotics, biomechanics, AR/VR, and exergames. Current state-of-the-art models are either too slow to run efficiently on mobile devices or sacrifice accuracy for speed. We present InstantHMR, a family of open-source single-image regressors suited for mobile inference, based on the Momentum Human Rig (MHR) parametric model and trained exclusively on publicly available annotations from the SAM 3D Body dataset. Our approach achieves remarkable performance on a Samsung Galaxy S23 NPU, recovering a 6,890-vertex body mesh in 5.5 ms per person, and 0.98 ms on an NVIDIA RTX 4070 GPU. This speed surpasses all existing methods we evaluated. On mobile, it is 12x faster than the single mobile-compatible method that slightly outperforms ours in accuracy, while achieving competitive accuracy with 40.0 mm PA-MPJPE on 3DPW, 48.8 mm on EMDB and 41.7 mm on RICH. To ensure fair comparisons, we re-evaluate 11 state-of-the-art methods under a unified protocol aligned with the closest MHR-based competitors, including Meta’s official MHR-to-SMPL conversion, and report their reproduction results. Additionally, we measure latency across eight different targets, ranging from desktop TensorRT to two phone NPUs. Code and models are available at this https URL.
[CV-46] Bootstrap Your Own AV-Proxies: Adaptive Contrastive and Prototype Learning for Audio-Visual Segmentation CVPR2026
链接: https://arxiv.org/abs/2611.09191
作者: Junbo Zhang,Hang Su,Zhaofan Li,Hang Dong,Chao Sun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to CVPR 2026
Abstract:Audio-Visual Segmentation (AVS) aims to accurately segment sounding objects in video frames by leveraging audio-visual correspondence cues. However, it remains challenging due to the intrinsic semantic incompleteness within a single modality and the semantic gap between audio and visual representations. Traditional feature-fusion decoding approaches struggle to suppress fusion noise effectively, while recent methods that incorporate data-dependent priors often increase the complexity of modeling audio-visual correlations, leading to poor cross-domain generalization. To address these issues, we propose a novel adaptive contrastive and prototype learning framework, BYOAVP, for AVS. Specifically, we design a Self-Supervised Audio Enhancement (SSAE) module that introduces contrastive learning to adaptively align audio representations with gradient-blocked visual embeddings, thus narrowing the semantic gap between modalities. Furthermore, a Dynamic Prototype Constraint (DPC) module is developed to refine pixel-wise category perception via momentum-based prototype updating, while enhancing localization of sounding regions through cross-modal interaction. Extensive experiments demonstrate that our method achieves SOTA performance across two AVS benchmarks and six sub-tasks, exhibiting strong robustness and generalization ability.
[CV-47] SIFR: Stable Signed Flow Reinforcement
链接: https://arxiv.org/abs/2611.09175
作者: Runlong Liao,Xingqian Xu,Lizhang Chen,Baiyu Su,Yufan Huang,Zhonghao Wang,Akash Gokul,Ming-Yu Liu,Qiang Liu,Yogesh Balaji
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Reward-based fine-tuning of flow models requires effectively promoting high-reward and suppressing low-reward samples, while maintaining training stability over many rounds of iterative updates. Iterative reward fine-tuning provides a simple starting point, but the positive reward weights give no explicit unfitting signal, and repeated reweighting concentrates the distribution and causes collapse. We propose Stable Signed Flow Reinforcement (SIFR), which replaces the reward weights by signed advantages to supply the missing unfitting signal, and constrains the velocity drift from a reference model to stabilize the updates, with a Lagrange multiplier that adapts the regularization strength to a prescribed drift budget. We further show that existing flow RL methods fall within this framework: they implicitly minimize the signed loss, differing in their regularization choices. We validate SIFR on SD3.5 Medium (2.5B) and FLUX.2 Klein (9B) and scale it to Cosmos 3 Super Text2Image (64B): across all 12 backbone–reward pairs with preference and rule-based rewards, it matches or exceeds existing methods, reaches their peak rewards much faster, and keeps improving over long training where they plateau or collapse.
[CV-48] FIND: Frequency Invariance Disentanglement for Test-Time Adaptation in LiDAR 3D Detection NEURIPS2026
链接: https://arxiv.org/abs/2611.09170
作者: Dongxiao Li,Zixuan Hu,Ling-Yu Duan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026
Abstract:LiDAR-based 3D Object Detection (L3OD) is a fundamental 3D perception task, yet its performance is often compromised by domain shifts arising from diverse sensor configurations and environmental conditions. Existing Test-Time Adaptation (TTA) methods primarily leverage self-supervision in the spatial domain, yet they frequently confound semantic content with domain style due to the sparse, non-Euclidean nature of point clouds. To overcome this limitation, we propose FIND (Frequency INvariance Disentanglement), a novel TTA framework that precisely disentangles domain-invariant content from domain-specific interference within the frequency domain. Our methodology utilizes B-Spline fitting to construct locally adaptive filters integrated into a dual-stream architecture: an Invariant Stream extracts robust features to guide pseudo-labeling, while a Specific Stream enforces consistency against domain perturbations. By combining invariance-driven alignment with stability-guided regularization, our approach dynamically extracts robust domain-invariant features while suppressing domain-specific interference. Extensive experiments on cross-dataset adaptation and robust corruption benchmarks demonstrate that FIND significantly outperforms state-of-the-art methods. The source code is publicly available at this https URL.
[CV-49] DINO-Nucleus: Multi-Level Self-Supervised Features for Nucleus Instance Segmentation
链接: https://arxiv.org/abs/2611.09133
作者: Mohammad Sarhangzadeh,Elyar Esmaeilzadeh,Doruk Oner
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 2 figures
Abstract:Nucleus instance segmentation and counting are essential for quantitative microscopy, enabling reliable analysis of cell populations, tissue organization, and disease-related changes. However, accurately separating touching nuclei remains difficult, and many existing approaches rely on complex models or sensitive post-processing procedures. We present DINO-Nucleus, a parameter-efficient framework that leverages multi-level self-supervised DINOv2 representations for nucleus instance segmentation and counting. The model jointly predicts nucleus regions, pixel-wise instance flows, and explicit nucleus centers, while a consistency objective encourages the predicted flows to converge toward their corresponding centers. Multi-level transformer features provide complementary spatial and semantic information, and the pretrained backbone is adapted using lightweight LoRA modules to reduce fine-tuning cost. Experimental results show that DINO-Nucleus achieves strong segmentation and counting performance with parameter-efficient backbone adaptation. These findings demonstrate that parameter-efficient adaptation of multi-level self-supervised features can provide an effective solution for nucleus instance segmentation in microscopy images.
[CV-50] ouchAnnotator: Multi-Source Tactile Supervision with Privileged-View Distillation
链接: https://arxiv.org/abs/2611.09129
作者: Ao Li,Zhihao Yuan,Chubin Zhang,Kunrui Huang,Sheng Xu,Zengjue Chen,Junjie Wang,Jiawei Li,Yansong Tang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Egocentric video captures human-object interaction at scale but rarely includes tactile labels for hidden contact and pressure. We introduce TouchAnnotator, an ego-only bimanual tactile annotator trained with multi-source tactile supervision and privileged-view distillation. A canonical hand field unifies glove pressure, geometry-derived human contact, and robot fingertip deformation from six datasets and three embodiments, while source-aware masks preserve valid targets. During training, an EMA teacher observes synchronized ego and wrist views and distills hand-local representations and tactile predictions into a student that uses only ego video and estimated pose at inference. Compact past-clip memory supplies temporal context. We evaluate on EgoTouch, zero-shot FPHA transfer, and 860 synchronized glove recordings. The model transfers to bare-hand contact and produces qualitatively coherent tactile responses across bare-hand, gloved-hand, and robot embodiments. On measured recordings, it improves temporal accuracy by 5.38 points and Volumetric IoU by 5.02 points under a shared protocol. These results show that heterogeneous human–robot supervision and privileged training views can turn egocentric video into dense tactile annotations.
[CV-51] Look Closer Align Better: Explicit Visual Attributes for Robust EEG-to-Image Decoding
链接: https://arxiv.org/abs/2611.09123
作者: Ke Xiao,Banghua Yang,Wenbin Guo,Fenqi Rong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:EEG-to-image retrieval commonly aligns neural responses with representations from pretrained vision encoders. However, these representations may underrepresent fine-grained visual details needed to distinguish semantically similar images. To address this limitation, we first investigate the EEG decodability and complementary predictive value of explicit visual attributes, identifying visual cues that capture EEG-relevant information beyond pretrained representations. Building on these findings, we propose AttrEcho, a framework that enriches pretrained visual representations with complementary, spatially aligned visual attributes. Specifically, Region-Aligned Visual Attribute Fusion (RAVF) integrates EEG-relevant local visual attributes with spatially corresponding pretrained features through a learnable signed residual gate. We further employ Feature Token Aggregation (FTA), which learns dimension-specific weights across regions to aggregate refined visual features for contrastive alignment with EEG representations. Experiments on THINGS-EEG demonstrate that AttrEcho achieves 94.8% Top-1 and 99.7% Top-5 accuracy in intra-subject 200-way retrieval, while maintaining 82.1% Top-1 accuracy as the gallery expands to 1,600 images with same-concept distractors. AttrEcho also achieves 25.3% Top-1 and 53.0% Top-5 accuracy in inter-subject retrieval. These results highlight the importance of EEG-relevant visual attributes and spatially aligned fusion for fine-grained EEG-to-image retrieval.
[CV-52] CoMET: Tracking the Shared Pose Error of Roadside Messages for Asynchronous Vehicle-Infrastructure Cooperative 3D Detection
链接: https://arxiv.org/abs/2611.09121
作者: Shuwei Sun,Jian Huang,Hui Cao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 25 pages, 4 figures. Code and experimental results archived at this https URL
Abstract:A roadside sensor can detect vehicles hidden from the ego vehicle, but a pose error shifts every detection in the same message. We study asynchronous vehicle-infrastructure cooperative 3D detection on V2X-Seq using only information available at decision time, without ground truth, future frames, or per-pair calibration offsets at inference. Objects seen by both sensors reveal the shared translation, but even at zero added delay, 38% of validation messages have no matched pair. We propose CoMET (Common-Mode Error Tracking), which estimates this translation over time. At each message arrival, CoMET matches vehicle and roadside detections, estimates the offset from mutually nearest pairs, and updates a Kalman filter for that sender. When no pair matches, the filter carries its estimate forward. The correction applies to all roadside boxes, including those unseen by the ego vehicle. The estimator has no learned parameters; a separate frozen head ranks the merged detections. Across four delays from 0 to 400 ms on 21 validation sequences, CoMET averages 66.63 bird’s-eye-view AP at IoU 0.7, compared with 63.18 for the same late fusion using uncorrected online poses and 62.96 for the best intermediate-fusion baseline. Across five detectors from three architectures, CoMET gains 2.2 to 3.5 AP points over online poses and 0.3 to 0.6 over estimating the offset from the current message alone. We find that the benchmark uses a zero offset as a placeholder in 28% of validation messages and that its offset search reaches the boundary in 45% of the rest. These offsets affect the labels for objects seen only from the roadside. Under two alternative label references, CoMET improves on uncorrected online poses in every evaluated population. Retraining intermediate fusion with CoMET poses and corrected labels improves detection on all three architectures tested, with statistically significant gains on two.
[CV-53] Finite-sample Calibration Guarantees for Binary Segmentation
链接: https://arxiv.org/abs/2611.09120
作者: Bruno Viti,Martin Holler
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Published in Transactions on Machine Learning Research 10/2026
Abstract:Machine learning based segmentation models are frequently miscalibrated, producing for each pixel a label prediction whose associated confidence does not accurately reflect the true underlying probability. This limitation is particularly critical in high-stakes applications such as medical image segmentation. Although several calibration methods have been proposed, only a small subset of them provides theoretical guarantees on the calibration error, i.e., they do not typically ensure that, with high probability, the model exhibits a low calibration error. In this work, we develop a post-hoc calibration method for which formal calibration guarantees can be established. To achieve this, we derive from standard concentration inequalities an improved theoretical bound that explicitly accounts for the spatial dependence inherent in binary segmentation outputs. We empirically demonstrate that the proposed calibrator satisfies the derived guarantee and outperforms existing post-hoc calibration approaches.
[CV-54] BayesCrop: Real-Time Probabilistic Multi-Crop Video Reframing WACV2027
链接: https://arxiv.org/abs/2611.09100
作者: Cheng-Han Lee,Rajesh Sureddi,Neil Birkbeck,Yilin Wang,Balu Adsumilli,Alan Bovik
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: This paper is accepted by WACV 2027. We plan to publicly release the source code at this https URL
Abstract:Automatic video reframing is increasingly important for adapting landscape content to heterogeneous display formats. Existing methods typically predict a single deterministic crop, despite the subjective nature of reframing and the presence of multiple plausible regions of interest. We present \textbfBayesCrop, a real-time probabilistic framework for multi-crop video reframing. BayesCrop represents crop preferences using a two-component spatially isotropic Gaussian mixture over crop position and scale. Each component defines a representative reframing mode together with its local uncertainty, enabling both dual-view prediction and sampling of additional plausible crop candidates. For efficient video processing, we introduce an anchor-and-propagate strategy that performs neural prediction only on sparse anchor frames and propagates crop trajectories using motion vectors available from the compressed video bitstream. The resulting trajectories can be combined into a single output using a lightweight periodic view-switching strategy. Experiments on LIVE-YT-MC, LIVE-YT-VC++, and a multi-target long-video benchmark demonstrate competitive single-output localization, high multi-crop reference coverage, diverse yet coherent crop hypotheses, and stable perceptual quality. BayesCrop achieves real-time throughput of 112 FPS at the default operating point, while sampling from the learned distributions further improves reference coverage beyond the two representative crop means.
[CV-55] rajectory Forcing: Exploiting Diffusion Trajectories for Autoregressive Long Video Generation NEURIPS2026
链接: https://arxiv.org/abs/2611.09096
作者: Bohan Wang,Shuo Chen,Yunzhi Li,Zhaozheng Chen,Junzhe Zhang,Hanwang Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026
Abstract:Autoregressive (AR) modeling is arguably the only way to extend a fixed-length video diffusion model to longer or even streaming videos. However, existing AR diffusion methods use only the previously predicted chunk as context for the next, discarding the intermediate predictions that constitute the chunk’s generation history. In contrast, conventional AR models such as LLMs condition on the entire history of preceding tokens. To this end, we propose Trajectory Forcing (TraF), an AR diffusion method that exploits the denoising trajectories of previous video chunks as context. TraF operates under two training regimes: (1) with ground-truth videos, where pseudo-trajectories are constructed from Gaussian interpolations and re-predicted by the model, and (2) with a teacher model, where real trajectories are obtained from the student’s own denoising rollout. The first regime naturally provides a strong initialization for the second. On the 5-second VBench benchmark, TraF achieves a Total score of 85.15, outperforming Self Forcing (+0.84) and Causal Forcing (+1.11), with Quality and Semantic improving simultaneously. On 30-second generation—6 \times the training horizon—TraF achieves the highest overall quality among all compared methods.
[CV-56] AdaAlloc: Adaptive Visual Token Allocation for Long-Video Question Answering NEURIPS2026
链接: https://arxiv.org/abs/2611.09092
作者: Joungbin An,Kristen Grauman
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to NeurIPS 2026. Project page: this https URL
Abstract:Long-video question answering forces multimodal large language models to reason under strict visual-token budgets, creating a tradeoff between broad temporal coverage and fine-grained spatial detail. Uniform sampling preserves coverage but often obscures decisive local cues, while caption-based or frame-similarity retrieval operates over lossy proxies that can miss motion, state changes, and before-after relations. We recast long-video reasoning as a visual-token allocation problem: given a fixed budget, decide how much to spend on global context versus high-resolution local evidence, and where in the video to draw that local evidence from. We introduce AdaAlloc, a training-free inference framework that adaptively allocates visual tokens between a global video overview and localized visual detail. AdaAlloc first plans a global/local allocation policy, then uses temporal grounding to locate candidate evidence segments and refines them into a compact temporal memory. The selected policy controls both frame sampling density and resolution, combining sparse, lower-resolution global frames with denser, higher-resolution local frames drawn from the refined memory. The final answerer reasons over the allocated global context and refined local evidence. Across LVBench, MLVU, VideoMME, and LongVideoBench, AdaAlloc improves over budget-matched uniform sampling by up to 10.5 percentage points and outperforms the compared frame-selection and agentic methods on their reported benchmarks.
[CV-57] raining-Free Active Test-Time Adaptation for Vision-Language Models NEURIPS2026
链接: https://arxiv.org/abs/2611.09084
作者: Jihwan Bang,Sumyeong Ahn,Hwanjun Song,Jae-Gil Lee
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026
Abstract:Test-time adaptation (TTA) for vision-language models (VLMs) is usually studied in a fully unsupervised regime, where adaptation relies on pseudo-labels, confidence scores, or entropy computed by the model itself. This is a severe restriction on hard shifted samples: the model is asked to repair its own uncertain predictions without any external evidence. At the other extreme, supervised online adaptation assumes labels can directly train or correct the model. We study the intermediate deployment regime between these two extremes. In Active Test-time Adaptation for VLMs (Active VLM-TTA), a pretrained VLM may spend an explicit labeling budget on selected test samples, but the prediction for the current sample must be emitted before its label can be used; queried labels are therefore costed, delayed evidence for future samples rather than free correction for the present one. Under this future-only constraint, the central question is how to reuse sparse labels conservatively. We instantiate this setting with ACTive VLM-TTA for Online Retrieval (ACTOR), a training-free plug-in module that routes uncertain samples among base prediction, label query, and memory-based correction using a consistency-and-reliability check over previously queried labels. Across ten cross-domain benchmarks and a domain-transfer benchmark based on ImageNet, ACTOR consistently improves strong VLM-TTA baselines on ResNet and ViT backbones with negligible inference overhead. Our implementation is publicly available at this https URL.
[CV-58] GlobalPuzzle: Globally Informed Genetic Search for Eroded Jigsaw Puzzle Reassembly
链接: https://arxiv.org/abs/2611.09077
作者: Lunyan Wang,Chuanzhi Xu,Ziyuan Tao,Weidong Cai
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 22 pages (10 pages main paper and 12 pages supplementary material), 16 figures
Abstract:Jigsaw puzzle reassembly remains a highly challenging task, particularly when piece boundaries are irregularly eroded to simulate damage to cultural artifacts. On the GAP benchmark, existing methods achieve a perfect assembly accuracy of only 0.3% on 5 \times 5 puzzles, whereas humans can solve jigsaw puzzles with remarkably high accuracy. Humans typically approach puzzle solving by identifying potentially adjacent pieces, evaluating the current assembly, and iteratively performing complex operations such as adding pieces, replacing incorrectly placed pieces, and moving entire connected groups of pieces. However, existing methods face a fundamental limitation: approaches that leverage global visual information typically predict piece arrangements directly, making it difficult to explicitly perform such complex iterative operations, while search based methods can rearrange pieces but primarily rely on pairwise boundary compatibility, lacking global visual information to evaluate and guide the search process. To bridge this gap, we propose GlobalPuzzle, a jigsaw puzzle reassembly framework that integrates global canvas level evaluation, action guidance, and genetic search. By incorporating global visual information into the evaluation and optimization of candidate assemblies, GlobalPuzzle aims to resolve ambiguous adjacency relationships that are difficult to determine from irregularly eroded piece boundaries alone, enabling flexible, globally informed reconstruction.
[CV-59] Emo-LiveSync: Real-Time Universal Lip Synchronization with Explicit Emotion Control BMVC2026
链接: https://arxiv.org/abs/2611.09074
作者: Sangjoon Yu,Seungeun Lee,JoonHeyok Shin,Ji-Su Kang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to BMVC 2026, 23 pages, 16 figures
Abstract:Lip synchronization, the task of generating realistic lip movements synchronized with audio signals, is fundamental to applications such as film dubbing, virtual avatars, and digital content creation. Recent GAN-based methods offer computational efficiency but often lack visual fidelity, while diffusion-based approaches achieve high-quality results at the cost of substantial computational overhead. Moreover, most existing methods are limited to photo-realistic human faces, restricting their applicability to stylized characters and non-human entities. In this paper, we propose an efficient universal lip-sync video generation framework that achieves high-quality results without compromising computational efficiency. We build on LivePortrait, a warp-based portrait animation model, and introduce a training strategy that aligns audio signals to an implicit motion space by distilling knowledge from a pre-trained video-to-motion model. This enables real-time inference while maintaining high visual fidelity, and naturally generalizes to diverse stylistic characters beyond photo-realistic humans. Furthermore, we introduce an expressive control gate that enables explicit manipulation of emotional expressions for content enrichment. Extensive experiments demonstrate that our method achieves state-of-the-art performance in both visual quality and lip-sync accuracy while ensuring real-time inference.
[CV-60] UniRGBA: Unifying RGB and Alpha in Pretrained Video Diffusion Models
链接: https://arxiv.org/abs/2611.09069
作者: Lei Xu,Zhuo Li,Yongyi Jin,Aoyu Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent advances in video diffusion models have achieved remarkable photorealistic RGB synthesis, yet generating videos with a native alpha channel remains an open challenge. Existing approaches often rely on matting-style data and non-native RGBA designs, causing them to specialize in real-object scenes and to degrade severely on visual-effect content whose semi-transparency is spatially non-local and non-elliptical. We present UniRGBA, a unified framework whose key insight is that RGBA generation does not require a dedicated alpha encoder or complex latent-space restructuring—a minimal, targeted extension of the existing RGB pathway is sufficient. UniRGBA-VAE extends a causal video VAE to 4-channel RGBA via zero-initialized channel expansion, symmetric selective layer unfreezing (5% trainable), and a six-term progressive reconstruction loss anchored by a KL reference loss that prevents latent drift, followed by a two-stage schedule that introduces a lightweight GAN objective. We then apply standard DoRA to the pretrained DiT, reusing the unified 4-channel latent without any alpha-specific architectural changes. On a visual-effect RGBA benchmark, UniRGBA substantially outperforms Wan-Alpha on RGBA reconstruction while preserving RGB fidelity. For generation, UniRGBA outperforms Wan-Alpha on RGBA text-to-video tasks and matches or surpasses the original pretrained Wan2.1 backbone on RGB-only metrics, demonstrating that native alpha support can be integrated without compromising the underlying RGB prior or the model’s architectural versatility.
[CV-61] ProtoCouncil: Multi-Encoder Anatomical Prototypes for Prostate Cancer Detection in Micro-Ultrasound
链接: https://arxiv.org/abs/2611.09058
作者: Mohammad Mahdi Abootorabi,Nooshin Maghsoodi,Armin Saadat,Obed Dzikunu,Zhuoxin Guo,Hannes Cash,Ferdinand Luger,Adam Kinnaird,Brian Wodlinger,Parvin Mousavi,Purang Abolmaesumi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 12 pages, 3 figures, Under Review
Abstract:Purpose: Micro-ultrasound ( \mu US) is a promising modality for prostate cancer (PCa) detection, but identifying suspicious tissue depends heavily on clinical experience. Models trained on sparse, noisy core-level pathology offer little insight into their decisions and can fail under domain shift. Clinical use needs a traceable decision process and a flag for uncertain predictions. Methods: We introduce ProtoCouncil, in which three foundation encoders (MedSAM, DINOv3 and V-JEPA 2.1) learn case-based prototypes at three anatomical levels: whole gland, needle track and gland sectors. Prototypes are projected onto real training-core regions, and label-free fusion keeps every class logit as a bias plus an exact sum of prototype similarity contributions. Encoder disagreement triggers selective prediction. Results: In five-fold patient-level cross-validation (6,588 cores, 693 patients), ProtoCouncil reaches 82.9% AUROC for clinically significant PCa (csPCa, Grade Group \geq 3), 4.3 points above the best previously published method. In zero-shot evaluation on two external cohorts, ProtoCouncil achieves higher all-cancer and csPCa AUROC than all previous methods. Nearest same-class exemplars match test-core tumor involvement better than random same-class cases. Externally, csPCa accuracy falls from 79% when encoders agree to 27% when two oppose the fused call. Abstaining on disagreement raises balanced accuracy by 7 and 15 points, while retaining 58% and 54% of the cores. Conclusion: ProtoCouncil grounds core-level predictions in exact similarity-based evidence from real, anatomically indexed training cases, which makes its reasoning interpretable. It achieves the highest csPCa AUROC across three cohorts and identifies unreliable predictions through encoder disagreement. Comments: 12 pages, 3 figures, Under Review Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2611.09058 [cs.CV] (or arXiv:2611.09058v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.09058 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Mohammad Mahdi Abootorabi [view email] [v1] Fri, 6 Nov 2026 06:33:28 UTC (3,995 KB)
[CV-62] OmniLatent: Unified Latent for 3D Part Modeling
链接: https://arxiv.org/abs/2611.09055
作者: Yunhan Yang,Yunfei Zhao,Zeqiang Lai,Jiachen Xu,Minxiu Wang,Xintong Han,Xiangyu Yue,Xihui Liu,Chunchao Guo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:3D part decomposition is a central capability for editable asset creation, yet existing workflows typically segment a completed mesh as a post-processing step, making the result sensitive to geometric artifacts, texture ambiguity, and limited part-level annotations. We present OmniLatent, a unified latent modeling framework for 3D part decomposition. Instead of predicting part labels in isolation, OmniLatent learns a shared latent space that jointly encodes geometry, deformation, appearance, and binary part IDs, allowing data-rich geometric and visual priors to support data-scarce part learning. Given a single input image, the base model generates geometry, deformation, texture, and surface part IDs in one pass, producing a part-aware 3D asset without a separate segmentation stage. We further tune the pretrained model for two extensions: a point-prompt adapter that maps arbitrary 3D point prompts with binary part IDs into latent tokens for full segmentation, refinement, and part subdivision, and an amodal part completion model that completes partial surface masks into a coherent full-object decomposition through dynamic per-part token allocation, cross-attention to the given geometry latent, and alternating part-local and global attention. Extensive experiments show that OmniLatent produces semantically meaningful part segmentations with precise boundaries, supports interactive part editing, and completes partial part observations while preserving global object consistency.
[CV-63] Grounding-Aware Judge Reward Reinforcement Learning for Out-of-Domain Face Anti-Spoofing
链接: https://arxiv.org/abs/2611.09047
作者: Xiaoyong Yu,Xinge Yu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Face Anti-Spoofing (FAS) must distinguish genuine faces from presentation attacks. A unified Multimodal Large Language Model (MLLM) is an attractive solution: it detects physical attacks (print, replay, mask) in one model and produces interpretable forensic evidence via Chain-of-Thought (CoT). A central obstacle to post-training such a model with reinforcement learning is that FAS reasoning has no rule-verifiable ground truth: only the binary verdict is checkable, while the reasoning chain cannot be verified by any hand-crafted rule. Optimizing the verdict alone is therefore a sparse, hackable signal. The model can reach the correct answer through a shortcut such as a false-rejection bias, which does not generalize. We address this with a grounding-aware judge reward. A 32B vision-language judge scores each reasoning chain against the image on evidence anchoring, logical coherence, and domain knowledge, turning the image itself into the verification signal. The reward is applied in a Group Relative Policy Optimization (GRPO) stage after CoT supervised fine-tuning. It is the judge, rather than the RL objective, that supplies the process supervision a verdict-only reward cannot. Our study starts from a model that already performs well In-Domain (ID) after CoT-based supervised fine-tuning and asks how to improve its Out-of-Domain (OOD) generalization. We find that grounding-aware rewards are an effective means to that end. We validate the method on OOD generalization, where the reward reduces ACER on two physical benchmarks (LCC-FASD 0.288 to 0.273, MSU-MFSD 0.431 to 0.409) while preserving ID accuracy (CelebA-Spoof 0.033 to 0.035). Against ResNet-50/101 and CLIP-ViT-B/16 baselines, the MLLM dominates ID and matches or surpasses the strongest baseline OOD. To our knowledge this is the first use of LLM-as-judge process supervision for FAS.
[CV-64] DobotTrack: Target-Grounded Vision-Language-Action Learning for Embodied Person Tracking
链接: https://arxiv.org/abs/2611.09041
作者: Shengjun Deng,Yinghan Zhang,Chengrui Wei,Liu Liang,Songyuan Liu,Pengkun Zheng,Yajun Zheng,Ying Zhang,Yuan Tian,Xi Li,Teng Wang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Language-guided embodied person tracking requires a mobile agent to identify a target from a referring description and consistently follow the target under egocentric observations. We present DobotTrack, a target-grounded vision-language-action (VLA) framework that integrates auxiliary grounding supervision with efficient temporal modeling. We systematically investigate five auxiliary tasks—person segmentation, bounding-box grounding, visual-spatial classification (VSC), future DINO feature prediction, and future-frame prediction—and adopt complementary mask and bounding-box supervision to enhance target delineation and localization. We introduce Q-identity, a target-conditioned historical visual compression module that leverages target-grounding predictions to selectively preserve identity-discriminative appearance features and spatial cues in compact historical tokens. On EVT-Bench, our 0.8B-parameter model achieves an average success rate of 77.9%, exceeding OmTrackVLA by 16.9 percentage points across the three tasks. We further construct a real-world person-tracking dataset collected using a quadruped robot, comprising 126 episodes and 596,323 mask-annotated frames. The target masks enable automatic derivation of bounding boxes and visual-spatial labels encoding horizontal position and apparent image scale, providing additional grounding supervision without separate manual annotation of these labels. Together, our framework and annotation pipeline support learning target-grounded tracking policies from simulated and real-world data. Code is available at this https URL.
[CV-65] Frontier3DBench: Benchmarking Coding Agents for 3D Creation across Objects Scenes and Interactive Open Worlds
链接: https://arxiv.org/abs/2611.09017
作者: Xinjie Lin,Qingle Liu,Wenyi Li,John Lam,Xiaoqiu Wang,Deyao Hong,Mingju Gao,Zheng Jiang,Yizhe Chi,Youjie Zheng,Calvin Xiao,Kaisen Yang,Huan-ang Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 19 pages, 7 figures, 4 tables. Project page: this https URL
Abstract:Coding agents combine visual understanding with program execution, offering a new route to automated 3D creation. By translating images and instructions into editable programs, they bring reconstruction, generation, and editing into a shared workflow. Yet existing benchmarks fragment evaluation across task scopes and protocols, leaving their capabilities across construction levels unclear. We propose Frontier3DBench, a unified benchmark of 145 tasks across 9 categories spanning object, scene, and interactive open-world creation, with joint evaluation of appearance, structure, and behavior. Experiments with 11 model-harness configurations reveal strengths that vary by level: GPT-6.1-sol leads the object and scene levels and achieves the highest overall score of 66.02/100, whereas GPT-6-astra leads interactive world construction. However, all configurations score lower on reference reconstruction than on text-driven generation in object and scene tasks. Producing plausible content thus remains easier than reproducing specified geometry, layouts, and visual details. Moreover, generated worlds average 83.08/100 on local mechanisms but only 20.82 on motion clearance and 30.26 on stateful workflows: individual mechanisms often function, whereas complete activities that require coordinated actions and persistent state remain difficult. These results identify reference fidelity and coordinated activity completion as central bottlenecks for coding agents in 3D creation.
[CV-66] Did I See It or Think It? Probing Reality Monitoring in Vision-Language Models with Concept Injection
链接: https://arxiv.org/abs/2611.09014
作者: Tina Khezresmaeilzadeh,Seyedarmin Azizi,Mehdi Kamal,Massoud Pedram
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent work tests whether language models can introspect by injecting a concept direction into their activations and asking whether they notice it and can name it. These studies are text-only. Vision-language models (VLMs) face an additional problem, known in cognitive psychology as reality monitoring: they must keep internally generated content separate from what they actually see. We therefore ask: can VLMs distinguish injected concepts from image content? We introduce a paradigm that varies the image and the injected concept independently. On each trial, the model reports what it sees, whether a thought was injected, and what that thought is. This design scores both whether the model identifies the concept and which source it assigns to it. Across six VLMs, only Gemma-3-27B reliably identifies injected concepts (35% vs. 3% for steering alone), with almost no false alarms. Even then, the source it reports is fragile. Strong or visual-pathway injections are reported as seen, and a visible object nearly eliminates correct thought reports. The same failure appears in ordinary captioning, where injected objects appear in up to 87% of captions of real photographs. Evaluating multimodal self-reports therefore requires checking both their content and their reported source.
[CV-67] UniMore: Fine-Grained Supervision for Unified Understanding and Generation
链接: https://arxiv.org/abs/2611.08996
作者: Haoyue Bai,Yinyi Luo,Yicheng He,Qingsong Wen,Sharon Li,Jindong Wang
类目: Multimedia (cs.MM); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Unified multimodal models combine image understanding and generation within one system, creating an opportunity to improve both capabilities through paired image–text supervision. We examine the outputs of a pretrained unified model in both directions and observe a loss of specificity: generated images can contain the requested objects with incorrect attributes, while captions identify a scene without expressing its distinguishing details. These failures call for supervision that specifies both the visual content to realize and the descriptive differences to distinguish. We propose UniMore, a post-training method that derives these complementary learning tasks from the fine-grained content of image–text pairs. The generation route predicts grounded crops from their corresponding phrases, matching the scope of the condition and visual target while using the native visual objective. The understanding route combines positive-caption fitting with ranking against captions that omit or alter details, conditioned on the full image. The two routes are optimized independently, and their learned updates are combined into a single unified model without changing its inference procedure. Experiments show gains across multiple understanding and generation benchmarks.
[CV-68] LEF-JEPA: Denoising-Trained Latent Energy Fields for Spatial Representation Learning
链接: https://arxiv.org/abs/2611.08989
作者: Zhiyuan Liu,Zhihao Xu,Sicheng Liu,Huanhuan Huang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 14 pages, 7 figures, 7 tables. Includes supplementary material
Abstract:A joint-embedding predictive architecture (JEPA) that applies prediction supervision only at masked positions struggles to clearly capture object shapes and boundaries in its latent representations. We further observe that some existing methods improve spatial structure while also retaining rich variations in texture and color. However, preserving more low-level visual information does not necessarily make representations better suited to prediction and action. This raises a key question: how can we strengthen the spatial structure in latent representations that supports prediction and action without over-encoding these details? We introduce LEF-JEPA, which learns spatial structure through target-guided latent recovery and transfers this supervision to the visual encoder through a relative-energy constraint; the energy module is used only during pretraining and adds no inference overhead. To reduce the cost of energy optimization during pretraining, we further construct a conditional Ising Hamiltonian that allows low-energy-state search to be performed by specialized Ising solvers. On physical hardware, we validate the feasibility of achieving lower forward latency with fewer local updates. Under a unified CIFAR-10 pretraining and frozen-encoder evaluation protocol, both realizations outperform V-JEPA2.1 in image classification, foreground segmentation, and visual planning.
[CV-69] Refer-CityFlow and M3Track: A Benchmark Dataset and Reference Framework for Language-Driven Multi-Target Multi-Camera Tracking WACV2027
链接: https://arxiv.org/abs/2611.08976
作者: Phuong T. Nguyen,Hao Vo,Tam Bang,Hoang K. Tran,Hoang-Minh Dau,Mahshid Malazizi,Tuan-Anh Hoang,Mina Sartipi,Thanh-Nam Doan,Ngan Le,Hoang H. Nguyen
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Accepted for publication at the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2027), Evaluations and Datasets Track. 10 pages, 4 figures, 7 tables
Abstract:Referring multi-object tracking enables users to specify targets through natural language, but existing Refer-MOT benchmarks primarily focus on single-camera settings. We formalize Referring Multi-Target Multi-Camera Tracking (Refer-MTMCT), where free-form queries specify identity-consistent trajectories to be recovered across a network of sequential, sparse-overlap cameras. We introduce Refer-CityFlow, a benchmark built on CityFlow’s global cross-camera identity annotations and augmented with language queries requiring joint language grounding, cross-camera association, and spatio-temporal reasoning. Refer-CityFlow covers 96 videos from 40 roadside cameras with 660+ identity-consistent cross-camera trajectories, 4,362 trajectory-level descriptions, and 7,699 queries, including 1,699 real-user queries collected from 50 participants. Reference experiments on the four-camera Scenario 02 show that single-camera referring trackers reach 14.95-18.28 HOTA on real-user queries, while the evaluated cross-view method achieves about 0.05 HOTA in direct transfer. To establish reference performance, we provide M3Track, a modular multimodal reference framework in which a large language model (LLM)-based agent orchestrates interchangeable detection, tracking, and cross-camera association modules. M3Track reaches 44.10 HOTA and 65.40 IDF1 on the same real-user query set. The project repository is available at this https URL.
[CV-70] Region-aware error triage for medical image segmentation with groupwise conformal risk control
链接: https://arxiv.org/abs/2611.08962
作者: Nishi Kanta Paul,Md Shihabul Islam Shovo,Hoileong Lee
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 22 pages, 7 figures, 12 tables. Submitted to Biomedical Signal Processing and Control (BSPC)
Abstract:Segmentation masks are increasingly reviewed rather than redrawn, which requires knowing where a mask is likely wrong. Learned error maps rank pixels by risk, but the number of pixels to warn is chosen ad hoc, and calibrating one warning threshold over all error pixels may leave errors far from the predicted contour under-flagged. We propose RACE-Triage, a post-hoc error triage framework that leaves the segmentation mask unchanged. A risk network is trained on out-of-fold predictions from backbone-independent inputs. The predicted contour partitions each image into a boundary band and a non-boundary region, and conformal risk control is applied separately in each region to bound the expected fraction of missed errors among cases with errors there. On ISIC 2018 dermoscopy and BUSI breast ultrasound (where patient identifiers are unavailable, so independence across images is assumed), pooled control met its overall target but missed on average 31.8% and 36.4% of non-boundary errors. Regional control reduced these fractions to 19.3% and 20.9% and raised the recall of size-filtered false-negative components from 0.722 to 0.818 and from 0.611 to 0.763. The mean warned area grew by 1.6 and 4.9 percentage points, and its 95th percentile reached 23.5% and 38.5%. The effect held across three seeds and a second backbone, and area-matched heuristics did not reproduce it. On the external PH ^2 set, frozen thresholds exceeded both targets, whereas recalibration with 100 local images brought the mean over repeated splits below them. Regional calibration improves non-boundary error coverage at a measured increase in warned area.
[CV-71] MetricPose: Flexible 3D Human Pose Estimation with Depth and Geometric Cues ICRA
链接: https://arxiv.org/abs/2611.08954
作者: Shifan Zhu,Sabad Modi,Hang Yang,Arushi Agrawal,Steven Gao,Meghan Huber,Nathan Wycoff,Donghyun Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages, 3 figures, ICRA submission
Abstract:In this paper, we present a 3D human pose estimator that predicts metrically accurate human joint positions in the camera coordinate frame. This task requires recovering the configuration of the body joints, scale, and metric camera-local placement. In monocular images, body size and camera distance are coupled through perspective projection and therefore cannot be determined uniquely from image alone. This ambiguity can be reduced by combining three complementary sources of information: learned RGB priors, direct depth measurements, and geometric constraints (e.g., ground contact) from the environment. We extend a pretrained full-body estimator with optional metric depth and geometry-processing modules that improve camera-distance estimation and depth-guided body refinement while preserving RGB-only operation. The resulting unified model supports RGB, RGB-Ground, and RGB-D using shared learned parameters. Unlike pelvis-relative error, camera-local error also measures placement relative to the camera. Compared with the existing foundation model SAM 3D Body, our methods with RGB, RGB-Ground, and RGB-D modes reduce camera-local joint-position error by 12.16%, 31.00% and 54.61% respectively.
[CV-72] Flex-RSI: Benchmarking Recursive Self-Improvement of Frontier Agents
链接: https://arxiv.org/abs/2611.08946
作者: Jingdong Zhang,Jionghao Wang,Zhengming Yu,Yancheng Cai,Wei Chow,Chris Choy,Zhiyang Dou,Hang Hua,Jiachen Li,Zhixuan Liu,Chen Wang,Songlin Wei,Siyuan Yang,Yu Yuan,Xiaohang Zhan,Yucheng Zhang,Shuhong Zheng,Xin Li,Lingjie Liu,Wojciech Matusik,Zhengzhong Tu,Wenping Wang
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: Project page: this https URL . Jingdong Zhang and Jionghao Wang contributed equally
Abstract:Frontier language models are increasingly deployed as agents, and how well an agent performs depends not only on its model but also on its harness: the prompts, tools, memory, and control loop through which the model acts. When an agent is applied to a new problem, a natural question is whether it can use the feedback of the task to improve the way it works, and not merely its answer. Most existing benchmarks keep the agent’s harness fixed and score the solution it produces, or target large-scale model training. We introduce Flex-RSI, a benchmark for the recursive self-improvement of agents on vision and robotics tasks under a fixed budget. In each round, an executor improves the task solution, which is kept only if it scores higher on held-out validation data, and an improver then revises the harness, including the executor and itself. Every run starts from the same general-purpose harness, and the final solution is scored on hidden test data. Evaluating closed-source and open-weight models on tasks spanning perception, reasoning, robot action, 3D reconstruction, visual creation, and infrastructure, we find that self-improvement helps most when the agent builds its own solution, that feedback recording only success or failure holds back open-weight models, and that open-weight models fall furthest behind where outputs must pass strict checks of physical stability or exact correctness. Paired runs further show that editing the harness itself contributes to the gains.
[CV-73] MESAM: Enhancing SAM2 for Few-Shot Segmentation through Semantic-Visual Memory Generation and Refinement
链接: https://arxiv.org/abs/2611.08943
作者: Tianyu Zou,Shengwu Xiong,Rui Wu,Ruilin Yao,Yi Rong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 17pages, 12figures
Abstract:This paper investigates how to leverage the powerful capabilities of the Segment Anything Model 2 (SAM2) to tackle few-shot segmentation (FSS) tasks. One major challenge lies in the substantial appearance differences between the support and query images, which may diminish the effectiveness of the memory-matching process in SAM2. Through systematic analysis, we find that the memory generated based on the categorical semantic information is robust to this appearance divergence and is more compatible with SAM2. Based on these observations, we propose a new Memory-Enhanced SAM2 (MESAM) method, which first introduces a Semantic-Visual Memory Generation (SVMG) module to separately construct a query semantic memory, a query foreground visual memory, and a support adaptive visual memory. Then, in our designed Interactive Semantic Memory Refinement (ISMR) module, we further enrich the obtained semantic memory with fine-grained visual details, and suppress its misatcivated background responses. Extensive experiments on the PASCAL- 5^i , COCO- 20^i , and LVIS- 92^i datasets demonstrate that MESAM consistently outperforms existing FSS methods, achieving new state-of-the-art segmentation performance. Our source code is available at: \hrefthis https URLthis https URL.
[CV-74] UAV-Based Computer Vision for Cattle Detection and Feed Bunk Spatial Analysis
链接: https://arxiv.org/abs/2611.08928
作者: Pedro Oliveira Torrente,Diego Giaretta de Paulo,Lucas Pascotti Valem,Jianglong Yan,Everton Castelão Tetila,Jayme Garcia Arnal Barbedo,Liang Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 12 pages, 6 figures, 6 tables. Preprint of paper accepted in IEEE Transactions on AgriFood Electronics
Abstract:Beef cattle production is central to Brazilian agribusiness and faces growing demands for efficiency, sustainability, and animal welfare. Precision Livestock Farming (PLF) offers tools to address these challenges, particularly through monitoring feeding-related spatial patterns in confined systems. Feeding management plays a decisive role in animal performance, yet its evaluation still relies largely on subjective, labor-intensive visual inspections that are hard to scale and prone to bias. This work proposes a computer vision framework based on Unmanned Aerial Vehicles (UAVs) to quantify cattle alignment and spatial dispersion at the feed bunk under real-world feedlot conditions. A UAV dataset was collected during flight campaigns at a commercial feedlot in Mato Grosso do Sul, Brazil. Images were manually annotated into cattle head and cattle back classes to support object detection and image-level spatial analysis of cattle distribution. You Only Look Once (YOLO) models, namely YOLOv8 and YOLOv11, were evaluated, with YOLOv8 small used for the downstream spatial analysis. The feed bunk region was segmented using the Segment Anything Model (SAM), and geometric metrics based on detected head positions were developed to characterize alignment, spacing regularity, and clustering. The selected detector achieved 84.30% mAP50 and 35.50% mAP50-95, with an inference time of approximately 2.7 ms per image on an NVIDIA RTX 5060 Ti GPU. The resulting metrics distinguished representative regular and irregular spatial distributions by Root Mean Square Deviation - RMSD (1.06 m and 4.26 m), Coefficient of Variation - CV (42.70% and 96.31%), and clustering (0 and 2 clusters), respectively. These findings indicate that the proposed approach can provide indicators for assessing feed bunk spatial patterns in PLF.
[CV-75] FANO: Anchored Forward-Noise Optimization for Diffusion Inverse Problems
链接: https://arxiv.org/abs/2611.08923
作者: Shilong Han,Hongxia Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, 9 figures, 8 tables. Preprint, under review
Abstract:Pre-trained diffusion models have become a widely adopted zero-shot prior for image inverse problems, among which annealed decoupling methods achieve leading reconstruction performance on nonlinear tasks. We find, however, that annealed methods that directly optimize the denoiser input x_t exhibit radial drift in later layers: within-layer optimization drives the noise-component norm of x_t away from its typical value, pushing the denoiser out of its high-performance regime and degrading reconstruction quality. To address this, we propose FANO (Anchored Forward-Noise Optimization). At each annealing layer, FANO fixes the previous denoiser estimate as an anchor and optimizes only the forward noise e , whose distribution is known, instead of x_t , thereby preserving the forward-diffusion structure of the denoiser input. The standard Gaussian prior on e concentrates its norm near the typical value, and deviations degrade denoiser performance. We accordingly decompose e in polar coordinates into a direction and a radial magnitude, and apply a soft potential to regulate the radius, keeping x_t in the high-density region of the forward distribution. On multiple linear and nonlinear inverse problems, FANO matches the strongest current baselines on linear tasks and achieves consistent improvements on nonlinear tasks.
[CV-76] MotionCore: A Latent Neural Simulator for Video Generation with Rigid-Body Dynamics
链接: https://arxiv.org/abs/2611.08904
作者: Zheng Wang,Zhengwei Gong,Yuanzhi Liang,Chi Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video generation models can produce highly realistic videos while generating physically inconsistent object motion. This limitation suggests that visual modeling alone may not provide a transferable representation of physical dynamics. We introduce MotionCore, a framework for learning physical dynamics via a learnable partial differential equation solver and focuses on rigid-body motion. We introduce a latent neural simulator learning to predict the evolution of object-level physical states from observed motion. Its predictions are then converted into spatially aligned latent controls for a controllable video diffusion model. The generator serves as a visual interface: it provides appearance and scene detail, while the learned dynamics model specifies the physical trajectory. To preserve this trajectory through the interface, MotionCore combines a diffusion-matching objective with a low-capacity decodability constraint on object identity and position. We evaluate MotionCore on public physical-consistency benchmarks and on paired data with known trajectories. Results indicate that MotionCore improves physical consistency and trajectory obedience in the evaluated rigid-body settings, while maintaining comparable visual quality. A controlled ablation protocol compares learned dynamics with an analytic prior and studies the choice of latent evolution processor.
[CV-77] Segmenting Geometric Textures of 3D Meshes
链接: https://arxiv.org/abs/2611.08897
作者: Weidan Xiong,Kaijie Qu,Shihao Feng,Shunqi Fan,Rongyu Xiong,Hui Huang
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV)
备注: Pacific Graphics 2026
Abstract:Recognizing geometric texture patterns on 3D models is critical for applications such as heritage protection, model editing, and 3D printing. Existing 3D model segmentation methods predominantly focus on identifying high-level semantic parts, thereby failing to capture middle-level textures. While recent Vision-Language Models can perform image segmentation with appropriate text prompts, they still struggle to recognize 3D geometric textures. In this work, we present a method that segments 3D mesh models based on middle-level geometric textures. We further introduce a new geometric texture segmentation benchmark and fine-tune SAM3 accordingly. The segmented texture regions are then clustered based on DINO feature similarity and lifted from 2D to 3D. Additionally, we develop an interactive tool to enable both 3D model editing based on our segmentation results and model retrieval from the dataset based on geometric and texture similarity. Experiments demonstrate that our method significantly outperforms existing methods in segmenting geometric textures and can facilitate various applications.
[CV-78] Decoding spatial and semantic information from large-scale visual cortical activity for natural-movie reconstruction
链接: https://arxiv.org/abs/2611.08887
作者: Yukai Huang,Haoyu Albert Wang,Jiawei Zhang,Jialun Ma,Jiayao Ding,Yiran Lin,Xingyu Chen,Ziqi Gao,Yuguo Yu
类目: Computer Vision and Pattern Recognition (cs.CV); Neurons and Cognition (q-bio.NC)
备注: 24 pages, 6 figures, 2 tables; 11-page supplementary information included as ancillary file
Abstract:Reconstructing natural visual experience from cortical population activity is a central challenge in neural decoding, particularly for calcium imaging, where signals are noisy, temporally filtered, and paired neural-video data are limited. Here we present Ca2REAL (Calcium Activity to Reconstruct Events, Appearance, and Layout), a neural-to-video framework that conditions a pretrained video diffusion model through separate layout and semantic controls derived from calcium activity. Using measured responses from 7,863 V1 neurons in one mouse, Ca2REAL reconstructed held-out natural-movie clips with 81.4% all-pairs stimulus-identification accuracy, with its clearest gains over deterministic baselines in perceptual and semantic measures. We further evaluated the framework in a stimulus-aligned virtual population of 116,705 neurons assembled from encoding-model-predicted responses across 11 independently recorded datasets. Perturbation and ablation analyses showed that reconstruction depends on distributed, stimulus-specific population structure, with layout and semantic pathways providing complementary spatial and identity information. Temporal analyses further showed improved preservation of semantic dynamics across reconstructed clips. These results show that large-scale cortical population activity constrains the spatial, semantic, and temporal structure of natural visual experience, providing a generative framework for high-fidelity neural reconstruction of dynamic visual scenes.
[CV-79] Computer Vision by YOLOs for Detection of Major Diseases in Vietnamese Custard Apple Using Fruit Imagery
链接: https://arxiv.org/abs/2611.08886
作者: Tri Nhut Do,Quoc Tuan Vo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Amid Vietnam rapid digital transformation in agriculture, the integration of artificial intelligence technologies for identifying, monitoring, and predicting crop pests and diseases has emerged as an inevitable trend.
[CV-80] GeoJeV: Unified Spatial Probability for Training-Free Efficient Evidence Acquisition in Ultra-High-Resolution Remote Sensing
链接: https://arxiv.org/abs/2611.08879
作者: Xinyi Niu,Lingfeng Chen,Jiehua Zhang,Panhe Hu,Li Liu,Zhen Liu,Yongxiang Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP)
备注:
Abstract:Ultra-high-resolution (UHR) remote sensing imagery provides rich spatial details but poses challenges for accurate and efficient multimodal understanding. Existing evidence acquisition methods selectively explore relevant image regions, yet largely rely on generative multimodal large language models (MLLMs) for intermediate search control. However, reinforcement learning from human feedback (RLHF) primarily aligns these models with human-preferred responses rather than explicitly optimizing region-level relevance judgments, while repeated autoregressive decoding introduces additional computational overhead. To address these limitations, we propose GeoJeV, a training-free method that formulates evidence acquisition as spatial probability-driven decision making. At each evaluated region, a frozen JeV model produces a probability distribution over candidate regions. Specifically, Region Rank prioritizes candidate regions, Sparse Branch adaptively determines branching width through sparse probability support, and Depth Gate controls refinement depth based on branch sparsity and distribution concentration. By coordinating search priority, branching width, and refinement depth through a shared distribution, GeoJeV reduces reliance on RLHF-aligned generation and eliminates intermediate decoding. Experiments on LRS-GRO, MME-RealWorld-RS, and LRS-VQA achieve accuracies of 60.59%, 52.79%, and 34.32%, respectively. Across the three benchmarks, GeoJeV improves the average accuracy of Qwen3-VL-8B by +5.32pp with only 1.12 additional seconds per question, while achieving a 7.8-fold inference speedup over the recent method.
[CV-81] Align3D: Image-Structure Co-Aligned 3D Scene Generation with Observability-Aware Fidelity-Preserving Guidance
链接: https://arxiv.org/abs/2611.08861
作者: Chenliang Zhou,Yifan Wang,Andrzej Banburski-Fahey
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We present Align3D, an image-conditioned diffusion model that generates 3D indoor layouts that are both faithful to a single input photograph and physically plausible beyond the observed view. From the photograph we infer a confidence-calibrated scene graph and train a set-diffusion layout denoiser with image-structure co-alignment where confidence-weighted modality fusion balances image and scene-graph alignment using predicted confidences, while observability-aware fidelity-preserving guidance projects conflicting scene-quality gradients and reweighs them with an object-level confidence mask so under-observed degrees of freedom absorb physical priors without undoing the image-graph match. On 3D-FRONT, Align3D improves image-alignment and scene-quality metrics over recent baselines; once trained, it maps previously unseen photographs to layouts in a single feed-forward pass.
[CV-82] he APL Doesnt Fall Far From The Tree: Unsupervised Artefact Prototype Learning in Histopathology
链接: https://arxiv.org/abs/2611.08838
作者: Emilia Agasi,Charlie Gourley,C. Simon Herrington,Robert L. Hollis,T. Ian Simpson
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Histopathology foundation models provide rich representations of tissue morphology, but technical artefacts in whole slide images introduce noise that may impact diagnostic accuracy. We introduce Artefact Prototype Learning (APL), a novel approach to detecting these artefacts that exploits the geometry of pre-trained embedding spaces without manual labels or fine-tuning. Rather than treating artefacts as isolated outliers, APL computes prototypes of out-of-distribution patches and uses them to distinguish artefact-associated patterns from tissue. Evaluation on a single-centre dataset shows that APL outperforms prior methods GrandQC and HistoART in detecting artefacts such as blur, dark spots, slide edges and pen marks, but struggles with tissue folds and blood. These results show that frozen pre-trained embeddings already carry information that allows artefact detection without expensive labels or vast computational resources.
[CV-83] Global Solar Siting from Geospatial Embeddings NEURIPS2026
链接: https://arxiv.org/abs/2611.08831
作者: Abdullah Rehman,Quentin Paletta,Rui Song
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Computers and Society (cs.CY)
备注: Accepted at Tackling Climate Change with Machine Learning workshop at NeurIPS 2026. 17 pages, 11 figures
Abstract:Global solar photovoltaic (PV) capacity is rapidly growing, helping decrease our emissions and dependence on fossil fuels. However, optimally siting a solar plant remains challenging, requiring anticipation of geographical, urban, and infrastructure needs. Current methods rely on multi-criteria decision analysis (MCDA) using hand-picked geophysical layers and solar-resource products, however these strategies are not geographically transferable and do not capture the practical requirements of solar sites. We use geospatial embeddings, dense vectors representing geophysical properties of a location, to analyse 38,591 existing solar sites and the land they were built on to learn the latent solar suitability signal. We assess a set of linear and non-linear models and select a soft-voting ensemble of a random forest and a multilayer perceptron (RF+MLP ensemble), chosen for accuracy and large scale deployability. Validated against future installations in thirteen countries, the ensemble separates existing solar sites from random land with a median ROC-AUC of 0.90. Compared with published MCDA strategies, the ensemble increases ROC-AUC by 0.06-0.30 in every country. The ranking also concentrates real deployment: the most-suitable 20% of land holds 84% of estimated generation (annual energy yield) and 67% of installed capacity (potential power output). After applying standard siting constraints, 5.6% of global land (9.8% if cropland is permitted) remains highly suited to solar. The global 10 m resolution solar suitability dataset and code can be found here: this https URL.
[CV-84] Impact of color in computer and human-based object recognition under occlusion
链接: https://arxiv.org/abs/2611.08808
作者: Courtney M. King,Elissa Aminoff,Daniel D. Leeds
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Partial occlusions challenge object recognition in both human and computer vision (CV), but the impact is largely dependent on the occlusion and model design. Properties relating to color are particularly important in biological recognition and frequently leveraged for camouflage, however there are still open questions about what makes them effective and how they translate to CV. This work studies partial occlusions that create camouflage, including those that blend in with target objects or with scenes, which alter object representations in distinct ways. Experiments evaluate recognition performance on both human participants and models ranging from convolutional neural networks (CNNs) to transformers, and additional analysis is performed on models to uncover trends in processing and object representations under multiple occlusion types. Novel metrics offer explainability of model behavior under occlusion: non-occlusion precision (NOP) and non-occlusion recall (NOR) measure alignment of model activations with and without occlusion, and activation in occlusion (AIO) measures the percentage of activations that occur in occluded regions. Findings show that human recognition, and most transformers, are negatively impacted by generated camouflaging occluders, with scene-based camouflage performing worse than object-based. Classical CNNs show a strong opposite trend, performing the worst on neutral-colored occlusions, suggesting that contextually-related colors, particularly scene-based, benefit object recognition rather than disrupting shape or boundary cues. Additionally, generated camouflaging occluders for training augmentation improves robustness to unseen camouflage for selected CNN model, and occlusion metrics suggest trends about learned mechanisms. Findings have interesting implications for alignment in human and computer vision, model vulnerabilities, and explainability.
[CV-85] Scale as a Modality: Multi-Scale Pixel-Space Diffusion Transformers
链接: https://arxiv.org/abs/2611.08802
作者: Ruibo Chen,Kaishen Wang,Tianyi Xiong,Chenxi Liu,Yanshuo Chen,Heng Huang,Jian Du
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Latent diffusion models rely on pretrained autoencoders that introduce lossy compression and constrain spatial granularity. Pixel-space diffusion avoids this bottleneck and enables flexible tokenization at multiple spatial scales, yet existing approaches largely adopt two-level coarse-to-fine architectures with limited cross-scale interaction. We introduce MS-pDiT, a multi-scale pixel-space diffusion transformer that treats spatial scale as a modality. MS-pDiT tokenizes the same noisy image at multiple patch sizes, processes each scale in a dedicated transformer stream, and enables bidirectional information exchange through compacted MM-DiT-style joint attention, with attention cost independent of the finest patch size. This design establishes the number of scales as a new architectural scaling dimension. On ImageNet 256\times256 , increasing the number of scales from two to five consistently improves FID from 4.39 to 2.31, 2.26, and 2.19, while adding an intermediate scale substantially outperforms increasing depth under comparable parameter budgets. Further analysis highlights the importance of bidirectional cross-scale interaction and late introduction of fine-grained streams. MS-pDiT achieves FID scores of 1.43 and 1.44 on ImageNet 256\times256 and 512\times512 , respectively, and naturally extends to text-to-image generation, achieving 0.86 on GenEval and 83.9 on DPG-Bench.
[CV-86] On the Few-Shot Adaptation of Vision-Language Models: A Few Gradient Signs Are All You Need
链接: https://arxiv.org/abs/2611.08800
作者: Nairouz Mrabah,Ismail Ben Ayed,Mahdi S. Hosseini,Eric Granger
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 98 pages, including supplementary material
Abstract:The limited supervision available for the few-shot adaptation of vision-language models makes gradient estimation particularly noisy. We find that the gradient magnitudes are severely miscalibrated, while the signs of a small top-magnitude subset remain informative. This motivates SignSparseSGD and DynSignSparseSGD, two state-free optimizers that use gradient magnitudes only to select where to update and fixed-magnitude signs to decide how to update. The static variant fixes the support after the first minibatch. The dynamic variant refreshes it at every step. Across 11 datasets, three CLIP backbones, and five shot levels, the proposed methods are competitive with or outperform parameter-efficient and optimization-based baselines on average, while exhibiting markedly lower sensitivity to long training schedules. They introduce no task-specific modules, maintain no first- or second-moment optimizer states, and use one shared configuration across datasets, backbones, and shot levels. Our code is included in the supplementary materials and will be released upon publication.
[CV-87] Extrapolation Matters: Sparse-Only Depth Completion Beyond LiDAR Coverag e
链接: https://arxiv.org/abs/2611.08782
作者: Maximilian Jansen,Gasser Elazab,Praveen Kulkarni
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Extrapolation beyond LiDAR coverage is an important yet under-evaluated challenge of depth completion. In real camera–LiDAR systems, partial sensor overlap, occlusions, and limited field coverage often leave large image regions weakly observed or completely unobserved by LiDAR. Thus, we investigate depth completion in this beyond-coverage regime using only sparse supervision from raw single-sweep LiDAR, without dense depth labels, LiDAR accumulation, or pseudo-LiDAR generation. Our approach leverages a frozen monocular depth foundation model, lightweight per-frame alignment to sparse LiDAR geometry, and a refinement module that fuses the aligned prior with sparse LiDAR anchors to predict dense metric depth. To regularize learning under raw sparse supervision, we introduce a high-frequency normal-consistency loss defined as the image-space gradient of the per-pixel normal cosine discrepancy. This term penalizes spatial variation in scalar angular error; it does not directly match changes in neighboring normal vectors. We further introduce an evaluation protocol that explicitly separates LiDAR-supported and beyond-coverage regions, revealing extrapolative behavior that conventional depth-completion metrics can obscure. Our method achieves strong depth accuracy beyond LiDAR coverage while remaining competitive in LiDAR-covered regions, providing a practical way to deploy monocular foundation priors in multi-sensor systems.
[CV-88] Federated training performs as well as centralized for 3D Latent Diffusion for Diagnosis-Conditioned Brain MRI Synthesis
链接: https://arxiv.org/abs/2611.08771
作者: Abhijith Shaji,Paul Kurian,Greg Ver Steeg,Paul M. Thompson,Jose-Luis Ambite
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Generative models of brain MRI are useful to understand brain conditions, and to enrich scarce and imbalanced neuroimaging data. However, training these models requires large, multi-site training sets, which is challenging. Federated learning (FL) allows for training deep learning models without sharing data. We show that training a 3D conditional latent model federatedly (where data is kept private at each site) performs similarly to centralized training (where all the data have been copied to a single site). We evaluate three metrics: conditional fidelity (how well generated MRIs perform in classification), diversity (how the generated images match the real data distribution, using MultiScale-SSIM), and realism (how close generated images are to real images, using Fréchet Inception Distance). Our results show that federated training is a viable route to develop multi-site generative models of brain MRI.
[CV-89] AtariBench: Evaluating Spatiotemporal Reasoning of Vision-Language Models in Atari Gameplay NEURIPS2026
链接: https://arxiv.org/abs/2611.08755
作者: Mingjia Huo,Yao Fu,Bo Chang,Yaqing Wang,Dawei Zhu,Dylan Zhang,Xinyang Yi,Lichan Hong,Ed H. Chi,Hao Zhang,Jiaxi Tang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 26 pages, 14 figures. Accepted to NeurIPS 2026, Evaluations and Datasets Track
Abstract:Vision-Language Models (VLMs) have achieved strong performance on static visual understanding benchmarks, yet their ability to act in dynamic environments remains less well understood. We introduce AtariBench, a controlled benchmark that evaluates VLMs across 20 Atari 2600 games, requiring models to make sequential decisions from high-temporal-resolution gameplay streams. By decoupling fast in-game dynamics from wall-clock inference speed, AtariBench focuses on whether models can perceive rapid visual state changes, reason over spatiotemporal dynamics, and ground actions in the current observation. Our experiments reveal substantial limitations in state-of-the-art VLMs: models often fail to infer object motion and direction, exhibit brittle sequential decision-making, and over-rely on textual interaction history despite newly appended visual feedback. We further introduce a separability analysis to identify which in-game scores provide reliable signals of model competence, and use the benchmark to study how memory design and prompting strategies affect performance. Overall, AtariBench exposes a persistent gap between passive visual perception and active decision-making, providing a lightweight and rigorous testbed for developing VLMs that can reason and act in dynamic environments.
[CV-90] FishSense Lite: A Camera-Based Single Laser Hardware Framework with Software System for In Situ Fish Length Measurement
链接: https://arxiv.org/abs/2611.08732
作者: Christopher L. Crutchfield,Kyle Hu,Nathan Hui,Omair Qazi,Jennifer Loch,Alli Candelmo,Brice X. Semmens,Curt Schurgers,Ryan Kastner
类目: Computer Vision and Pattern Recognition (cs.CV); Systems and Control (eess.SY)
备注: 27 pages with 14 figures
Abstract:Fish stock assessments need fish lengths paired with species from far more places than trained surveyors can reach. Recreational divers already visit many of those places, but measuring a fish underwater has required trained observers or specialist equipment. We present FishSense Lite, which recovers fish length from a single photograph taken with a recreational dive camera, a dive laser pointer, and a 3D-printed mount. The diver aims the laser at a fish and takes a picture; afterward, the dot’s position in the image gives the range to the fish, and the range turns the fish’s size in the image into a length. A citizen-science program, trained by the system’s builders, taught its own divers and collected lengths paired with species for 73 wild fish across seven open-water dives. On seven wild fish also measured by an underwater stereo-imaging system, the two instruments agreed without detectable bias (mean difference +0.8%). On rigid targets of known length, FishSense Lite recovered every target to within 1.3%, and 99% of 1,055 frames fell within a 15% error budget. From the error sources we measure, we derive a field protocol: calibrate against a dive slate on every dive, photograph fish within 15 degrees of perpendicular and beyond 1 m, and take at least ten frames of each.
[CV-91] Sharp: Diffusion-based Sharp Monocular Metric Event-to-Depth Estimation
链接: https://arxiv.org/abs/2611.08731
作者: Qiushi Dai,Aniket Dashpute,Yuhao Liu,Rishubh Parihar,Guha Balakrishnan,Ashok Veeraraghavan
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注:
Abstract:Depth perception is fundamental for autonomous navigation, industrial inspection, and mixed-reality systems. It often operates under strict constraints on power and form factor. In such cases, monocular depth estimation (MDE) is often used. Even as recent advances continue to improve monocular depth quality, conventional frame-based MDE systems remain challenged by fast motion and high-dynamic-range scenes. This work introduces eSharp, an MDE framework that addresses these problems using an event camera. eSharp repurposes pretrained diffusion priors for event-only MDE and produces accurate metric depth with unparalleled sharpness on real data. Real-world event datasets provide only sparse LiDAR depth maps for ground-truth supervision, making dense prediction challenging. We address this limitation through cross-modal distillation, using LiDAR-calibrated dense pseudo-labels generated from a spatially co-registered frame camera to train eSharp. Experiments show that eSharp outperforms prior event-to-depth baselines in both metric-depth accuracy and spatial fidelity, establishing diffusion models as a promising foundation for event-driven 3D perception.
[CV-92] Dynamic GPU Voxel Grids and Voxel-sequential Processing in NanoVDB
链接: https://arxiv.org/abs/2611.08665
作者: Eftychios Sifakis,Francis Williams,Christopher Horvath,Rohan Sawhney,Mark Harris,Jonathan Swartz,Ken Museth
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Distributed, Parallel, and Cluster Computing (cs.DC)
备注:
Abstract:We introduce two fundamental advances to the NanoVDB data structure: dynamic GPU-native topological updates, and a voxel-sequential access paradigm that dramatically accelerates point-wise and stencil-based operators on sparse grids. NanoVDB’s linearized memory layout has delivered strong GPU performance, and the Index-Grid concept introduced in the fVDB framework [Williams et al 2024] provided topology/data separation and optimal compactness of storage in sparse machine learning workloads. However, all these prior iterations of NanoVDB assume a static topology that has precluded use of this data structure in simulation and dynamic geometric processes. We eliminate this limitation by introducing a suite of dynamic topology operators – morphological dilation, grid merging, pruning, and cross-resolution coarsening and refinement. These operators use bit-parallel algorithms to reconstruct grids containing hundreds of millions of voxels in just a few milliseconds. We also introduce a voxel-sequential access paradigm and data structure for GPU kernels operating on sparse data. We stream active voxels as a flat linear sequence, minimizing or eliminating the need for tree traversal in the process. This achieves compute density and regularity that allow even complex stencil-based operators to approach peak memory bandwidth, with minimal storage overhead. Together, these contributions remove the read-only constraint from NanoVDB, establishing it as a high-performance primary representation for dynamic GPU simulation and geometry manipulation. We evaluate the efficacy of our newly introduced operators in benchmarks sourced from applications such as 3D reconstruction, fluid simulation, and large-scale mesh voxelization. Subjects: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2611.08665 [cs.GR] (or arXiv:2611.08665v1 [cs.GR] for this version) https://doi.org/10.48550/arXiv.2611.08665 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.1145/3842564 Focus to learn more DOI(s) linking to related resources
[CV-93] JEPAhead: From Inpainting to Causal Forecasting for Urban Driving
链接: https://arxiv.org/abs/2611.08657
作者: Daniel G. P. Petrini,Davi Pincinato,Fabio Grassiotto,Pedro L. F. Prado,Alexandre F. S. Osorio,Raphael V. Ramos,Felipe Pamboukian
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 7 pages, 2 figures
Abstract:We fine-tune a pre-trained V-JEPA1 predictor from bidirectional latent inpainting to action-free causal forecasting while keeping the encoder frozen. Using contiguous past observations, horizon-aligned targets, and group-level causal attention, the model forecasts eight future token groups, corresponding to sixteen frames, in one forward pass. On CARLA Town12, it retrieves the correct future among 591 candidates in 69.0% of cases, compared with 55.3% for persistence, with greater gains at longer horizons and on tokens exhibiting greater change. It also outperforms a compact predictor trained from scratch with six times fewer parameters. Semantic analysis indicates driving-relevant information in predicted features: a fixed linear probe trained on real encoder tokens achieves 0.16 mIoU on patches undergoing semantic-class changes, versus 0.13 for persistence. Vehicle-conditioned retrieval improves over persistence by 15 percentage points at the longest horizon. Pixel-space evaluation shows modest L1 gains but little improvement under LPIPS, motivating direct evaluation through latent and semantic measures alongside an oracle reconstruction reference. These findings support further investigation of latent forecasting for driving perception, while downstream utility remains unverified.
[CV-94] Scoring Not Learning: Decomposing Catalog-Consolidation Gains in Web-Scraped Retail Product Recognition
链接: https://arxiv.org/abs/2611.08655
作者: Kaden Leis,Xiang Ma
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 9 pages, 4 figures, was accepted by a computer vision conference workshop
Abstract:Retail product catalogs assembled from web image search are cheap to build and pervasively noisy: queries return packaging shots instead of products, and entries proliferate under names that cannot be distinguished visually. Consolidating such a catalog can look like an alternative to scaling the model. We show this is not a comparison between two ways of improving a recognizer, but between improving a recognizer and changing what counts as correct. We study a public archive of 165,769 cheese images across 1,824 query-named entries. A packaging filter and near-duplicate removal reduce it to a 998-class recognition set on one fixed, group-aware split. Canonicalizing entries to their base type (844 classes) raises top-1 accuracy by 4.0 points. Scoring the unchanged model against the merged label space, by summing its predicted probability within each canonical group, recovers +4.27 points with no retraining (+4.25 over six seeds). Retraining on the consolidated catalog then adds -0.10 \pm 0.53, +0.08 \pm 0.65 and -0.10 \pm 0.65 points at 28, 50 and 88 M parameters, indistinguishable from zero at every capacity ( \pm is the across-seed standard deviation). That component is specific, not mechanical: random merges with the identical size profile yield only +0.68 points, and the three backbones agree to within 0.1 points. But it is overwhelmingly category coarsening, not alias correction: across the 37 merged groups, type coarsenings supply 94% of the gain and alias collapses only 7%. We also detect no interaction between consolidation and backbone size: tripling backbone parameters is worth +1.50 \pm 0.57 points on the raw catalog and +1.42 \pm 00.44 on the canonical one, a within-seed interaction of -0.08 [-0.79,+0.63]. On this archive, the reported consolidation gain primarily reflects the evaluation label space, and separating that out costs one evaluation
[CV-95] An Anatomy-guided 3D Segmentation Pipeline for Prostate Surgical Phantoms from CT Scans
链接: https://arxiv.org/abs/2611.08652
作者: Iven Chen,Kuan-I Chung,Ayberk Acar,Susheela Sharma Stern,Alejandro O. Chara,Jie Ying Wu,Daniel Moyer
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages, 2 figures, 3 tables, 10 algorithms. Extended abstract accepted at SPIE Medical Imaging 2027: Image Processing
Abstract:Full autonomy has been a long-term goal in robotic surgery and a near-to-medium-term goal for prostatectomy. Due to the relatively closed structure of the procedure, its standard-of-care robotic/laparoscopic solutions, and procedure economics, prostate procedures are an optimal early target for surgical automation. Prototyping and testing of possible solutions is done in constructed laboratory phantoms, either synthetic tissues or animal phantoms. To enable measurement of perceptual and robotic motion error, phantoms undergo volumetric imaging; these images are critical to both a potential deployed automatic system, used in planning motion and treatment (cutting, ablation, suturing, etc.), as well as post-hoc assessment of system performance. Rapid iteration on these phantom models yields small, inconsistent datasets, while the current workflow for human annotation is often anatomically unreliable and prohibitively time-consuming, rendering deep learning unsuitable at reasonable scales. To address these limitations, we present an automated 3D segmentation pipeline that uses classical medical image-processing techniques to encode the known anatomy of the phantoms as explicit geometric and intensity priors, which requires no training corpus and is transferred to each new phantom revision without re-annotation. Across 57 scans spanning multiple phantom revisions, our pipeline recovers the capsule interior at 0.83 median Dice on 11 contrast-suppressed scans and detects all 32 annotated focal lesions without false positives, completing a full scan in 2-8 minutes on commodity CPU hardware versus 1-2 hours by hand.
[CV-96] VoluCompose: Constructing Compact Volumetric Visual Representations for Medical Image Understanding
链接: https://arxiv.org/abs/2611.08649
作者: Zanting Ye,Yunrui Lin,Shengyuan Liu,Jing Hao,Chen Hu,Yefeng Zheng,Lijun Lu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Medical vision–language models can interpret volumetric images, but processing full volumes is costly because hundreds of slices produce long, redundant visual-token sequences. Existing compression methods often couple content selection with budget determination. We observe that neighboring slices share visual features and that attention concentration varies across visual encoders. We propose VoluCompose, a training-free, model-agnostic framework that separates what visual content to retain from how many tokens to allocate. Its hyperbolic token composition module groups contiguous slices into ordered support units and uses the Lorentz distance to account for feature magnitude and direction during token allocation, anchor selection, and refinement. Its capacity control module uses visual-encoder attention entropy to allocate more tokens when attention is more dispersed. We evaluate VoluCompose on two tasks across three datasets and four backbones. Across nine calibrated settings, VoluCompose attains the highest mean quality–retention hypervolume among the compared methods. It achieves a mean relative quality of 121.68% against the uncompressed references while retaining 22.78% of visual tokens on average. Compression remains robust even at a 10% token budget. Across settings, the lowest ROUGE-L ratio to the uncompressed reference is 96.4% for VoluCompose. The corresponding minima for the compression baselines range from 15.0% to 70.5%. Code is available at this https URL.
[CV-97] From Geometry to Function: A Multi-Task LiDAR Benchmark Dataset for Semantic and Functional Parsing of Urban Facades WACV2027
链接: https://arxiv.org/abs/2611.08645
作者: Abdiaziz Muse,Tooba Javed,Tao Pu,Jiju Poovvancheri,Dong Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR); Machine Learning (cs.LG)
备注: Accepted to WACV 2027. 19 pages (8 main + supplementary), includes appendix. Project page: this https URL ; Code: this https URL
Abstract:Understanding building facades is central to urban digital twins, heritage modelling, and sustainable city analysis, yet existing LiDAR benchmarks stop at geometry and semantic segmentation and offer little support for reasoning about the functional role of each architectural element. We introduce HFX3D, a large-scale urban facade benchmark of approximately 492 million LiDAR points from 26 buildings in Halifax, Canada, captured across four seasons, day and night, and varied weather. Each facade is annotated at three tiers, semantic, instance, and per-instance functional attributes, making it, to our knowledge, the first facade dataset to pair geometry and semantics with functional labels. Annotations are produced by a semi-automatic pipeline that combines geometry-guided heuristics with expert validation. We benchmark representative point-based, graph-based, convolutional, and transformer architectures and find that semantic segmentation is already challenging under severe class imbalance, while instance separation of repetitive structures and functional reasoning from geometry are substantially harder. The benchmark establishes a new foundation for function-aware 3D urban understanding. Project page: this https URL ; code: this https URL
[CV-98] BEVCast: Path-Adaptive BEV for Efficient End-to-End Autonomous Driving
链接: https://arxiv.org/abs/2611.08631
作者: Qiren Wang,Mingyu Guo,Zixuan Wang,Weisong Shi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Camera-based driving models build a dense bird’s-eye-view (BEV) grid and spend the same computation on every cell of it, although the vehicle crosses only a few of those cells within the planning horizon. We present BEVCast, a two-stage BEV whose resolution adapts to the predicted driving path. A coarse pass encodes the whole scene on a grid of 4.096 m cells. A heat head trained on the recorded path assigns a heat to every coarse cell and the cells with the highest heat are refined into 1.024 m cells under a fixed budget. The remaining cells keep their coarse resolution, so the trajectory head reads a mixed-resolution BEV that still covers the entire scene and needs neither an HD map nor a navigation command. Against a full grid of 1.024 m cells with the same architecture and training recipe, BEVCast at a 10% refinement budget encodes 28% of the BEV queries and needs 66% of the BEV encoder time at batch size 1. On nuScenes its mean displacement error is 1.078 m against 1.149 m and its drivable area violation rate is 1.46% against 2.29%. On NAVSIM navtest its mean displacement error is 0.577 m against 0.598 m and its extended predictive driver model score is 82.51 against 81.28. Next to published planners BEVCast reaches a comparable error on both benchmarks without a navigation command. The resolution of a driving BEV can therefore follow the task instead of the grid.
[CV-99] InstaGen: Group-Shared Importance Sampling for Fast Video Generation
链接: https://arxiv.org/abs/2611.08610
作者: Kislay Aditya Oj,Kumar Krishna Agrawal,Sri Surya Varma Datla,Aditya Desai
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages, 6 figures, 2 tables
Abstract:Diffusion Transformers (DiTs) are the dominant architecture for video generation, but their quadratic attention complexity results in substantial computational cost and latency. Permuted block-sparse attention (PBSA) offers a promising approach to reducing this cost by selectively computing blocks of a semantically permuted attention matrix. However, existing approaches typically select only the top few key blocks for each query block, following a top- p paradigm. We find that top- p selection can be highly sample inefficient in high-entropy attention regimes, which are common in video generation. In these regimes, we show that sampling can provide substantially better approximations under the same computational budget. More generally, we cast block selection as importance sampling, where the top-scoring keys form an exact set that is computed in full and the remaining keys are represented by a sample drawn in proportion to their estimated attention and reweighted by inverse sampling probability. Building on this formulation, we present InstaGen, which can be parameterized by any PBSA method: it replaces the method’s block selection with an exact set plus a weighted sample. Each query cluster draws one sample shared by all its queries, so the sampled keys form blocks amenable to efficient GPU execution. In this paper, we take the SVG family of PBSA methods and show that, on Wan2.1-T2V-1.3B and Wan2.1-T2V-14B, InstaGen parameterized by SVG2 improves PSNR over SVG2 by 3.4 dB and 2 dB, respectively, at matched attention density, and by 1.6 dB and 2.2 dB at matched wall-clock time. On Wan2.1-T2V-14B, InstaGen also beats SVG-EAR by 1.1 dB PSNR at nearly the same density while running faster (1.65x against 1.48x over dense).
[CV-100] HaNDF: Object-Conditioned Geometric Neural Hand Distance Fields NEURIPS2026
链接: https://arxiv.org/abs/2611.08578
作者: Zuhao Liu,Melih Darcan,Zhengdi Yu,Riza Alp Guler,Tolga Birdal
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO)
备注: Accepted at NeurIPS 2026
Abstract:We propose, HaNDF, object-conditional neural distance fields as data-driven priors for modeling the plausible hand-object configurations, in the product manifold of hand articulations. While recent unconditional generative priors showed success in modeling human pose, hands are versatile articulated bodies often found in interaction with other objects. This makes it crucial to develop a generative prior which can model the plausible hand poses conditioned on the object under interaction. To this end, our HaNDF proposes a geometry-aware, object-conditioned neural distance field, representing the plausible articulations in the zero-level set of a conditional field. We leverage tools from Riemannian geometry to (i) represent plausible hands in the zero level-set of an object-induced neural field, and (ii) project a given pose onto this field. Using HaNDF, we can also optimize for a plausible condition, determining the object under interaction. Our extensive evaluations demonstrate that our HaNDF provides a flexible and powerful prior for hand-object interaction, suitable for applications such as pose refinement, reconstruction under occlusion, and physically plausible manipulation synthesis.
[CV-101] Meituan-Robotics-0: A Vision-Language-Action Foundation Model for Desktop Manipulation
链接: https://arxiv.org/abs/2611.08577
作者: Meituan Robotics Team:Caicheng Wang,Chen Zhao,Chengpeng Jin,Duo Xu,Fengjiao Chen,Guanglu Wan,Heqing Yang,Jia Wang,Jiacong Zhou,Jie Hu,Jiyuan Ru,Lingfeng Zhang,Linjiang Huang,Manyuan Zhang,Ningchu Liu,Shijie Zhou,Si Liu,Shuo Li,Tianyi Zhang,Xi Zhang,Xin Chen,Xuezhi Cao,Xunliang Cai,Yanbin Zeng,Yang Yang,Yang Yi,Yanli Tan,Yifu Xu,Yimeng Jia,Yuchen Tang,Yuchen Xie,Yuxiang Zheng,Zerwei Chng,Zilong Huang,Zipeng Ye,Ziyuan Qin,Ziqin Wang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 25 pages. Project page: this https URL ; Code: this https URL
Abstract:We present Meituan-Robotics-0 (MR0), a vision-language-action foundation model and embodied-agent system for stationary-base desktop manipulation. MR0 combines 147 million vision-language samples with 15,060 hours of curated robot trajectories spanning 12 embodiments. A three-stage training recipe aligns multimodal representations with robot behavior through discrete action-token pre-training, transfers these priors to continuous cross-embodiment control through flow-matching mid-training, and adapts the policy through real-world post-training. We introduce Rollback, which pairs replaced policy segments with human corrections to provide positive and negative supervision for CFG-DAgger without value learning or advantage labeling. A unified end-effector interface supports heterogeneous embodiments, while asynchronous action-chunk merging maintains smooth execution under variable inference latency. An event-driven agent maintains persistent task events and reassesses plans when progress or feasibility changes, enabling cross-task retrieval and conditional task resumption. MR0 achieves 93.6% success on RoboTwin 2.0 and ranks fifth in the RoboDojo VLA track. Under matched HG-DAgger post-training, MR0 achieves higher average success rate and progress score than \pi_0.5 across three real-robot tasks. On Ethernet insertion, CFG-DAgger trained with Rollback data reaches 70% success and a progress score of 0.90. Qualitative agent deployments demonstrate cross-task event retrieval and recovery from temporarily unsatisfied task conditions.
[CV-102] owards Label-Efficient Geospatial Foundation Model Adaptation via Positive-Unlabeled Learning for Building Damage Mapping
链接: https://arxiv.org/abs/2611.08561
作者: Hengwei Zhao,Qirui Wang,Zhuo Zheng,Xiaoyan Lu,Rusty Feagin,Wenzhe Jiao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Geospatial foundation models (GFMs) are reshaping remote sensing by shifting the field from task-specific modeling toward general-purpose geospatial intelligence. Mapping the damage to buildings stands to benefit substantially from this shift, as GFMs pretrained on large-scale Earth observation data provide generalizable geospatial representations that can facilitate effective responses to sudden-onset disasters. Although parameter-efficient fine-tuning (PEFT) facilitates GFM adaptation, parameter efficiency does not necessarily translate into label efficiency. Building damage mapping remains challenging due to scarce pre-existing labeled data or limited time for annotation. When weak supervision is introduced to alleviate the labeling bottleneck, we identify the difficulty of learning sufficiently discriminative representations as a key challenge for label-efficient GFM adaptation. To address this challenge, we propose a Positive-Unlabeled GFM Adaptation (PUMA) framework that requires annotations for only a small number of damaged buildings. More specifically, PUMA integrates non-contrastive weakly supervised representation refinement to mitigate the impact of noisy supervision on representation learning from a gradient perspective, and latent-state-based label disambiguation to enable reversible supervision refinement without requiring class prior estimation. Extensive experiments on disaster benchmark datasets from around the world demonstrated the effectiveness and robustness of PUMA. A real-world application to the 2025 Los Angeles wildfires further highlighted its potential to accurately map damage under highly limited supervision. More broadly, these findings suggest that weakly supervised PEFT offers a label-efficient pathway from GFMs to real-world applications.
[CV-103] Open-HumanWAM: An Open Recipe for Scaling Human and Robot Experience for Embodied Foundation Models
链接: https://arxiv.org/abs/2611.06817
作者: Juncheng Ma,Jiaxin Li,Jiankai Tu,Bo Liang,Guodong Lv,Yufan Deng,Junzhe Tang,Liangzhu Yang,Chen Bian,Shukai Gong,Yintianrun Zhang,Zicong Cao,Baohao Zhu,Jianxin Bi,Tengfei Yan,Yinyuan Zhang,Kewei Zhang,Xinglong Chen,Kaiqi Chen,Duomin Wang,Xiaotian Tang,Liang Tan,Li Li,Yuan Xu,Haosen Yang,Yuqi Wang,Jingfei Xia,Songyan Huang,Zhiyang Dou,Zhen Dong,Enze Xie,Tat-Seng Chua,Daquan Zhou
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: GitHub: this https URL Project website: this https URL
Abstract:Egocentric human video offers a scalable source of manipulation priors for robot learning, yet translating raw data scale into generalizable robot policies demands rigorous data curation and training recipes. We introduce Open-HumanWAM, the first fully open recipe for scaling egocentric pretraining to 50,000 hours, anchored by Open-HumanWAM-50K—a curated egocentric video corpus densely annotated with hand motion and natural language. Leveraging this corpus, we systematically investigate pure human pretraining via closed-loop simulation evaluations and uncover an empirical log-linear scaling law linking pretraining data hours to downstream robot success rates. Grounded in these findings, we develop HumanWAM-Vanilla, a world–action model pretrained exclusively on 50k hours of egocentric human video. Upon robot post-training, HumanWAM-Vanilla reaches 92.08% success on RoboTwin 2.0 Clean2Clean and 68.88% on Clean2Random, alongside a 14.17% mean success rate and a 19.00 mean progress score on RoboDojo. To push the capability frontier further, we explore joint human–robot pretraining with HumanWAM-Goal, a visual-goal-conditioned world–action model. Hybrid data scaling unlocks broad zero-shot generalization across both simulation and physical benchmarks, propelling post-trained RoboTwin 2.0 performance to new state-of-the-art levels: 92.70% on Clean2Clean and a substantial leap to 84.60% on Clean2Random. On physical robots, HumanWAM-Goal further exhibits emergent embodied in-context learning, dynamically adapting its manipulation behavior to unseen visual goals without parameter updates. To support reproducible scaling research in embodied AI, we open-source the Open-HumanWAM-50K corpus, data curation and pretraining pipelines, alongside the pretrained and post-trained HumanWAM checkpoints, inference code, and simulation evaluation suites.
[CV-104] Super-resolved digital terrain models for GEDI geolocation correction: agreement with airborne LiDAR DTM-based corrections at the beam level not the footprint level
链接: https://arxiv.org/abs/2611.09854
作者: Natsuki Yuda,Shinichiro Shirota
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Geophysics (physics.geo-ph)
备注: Appendices will be added in a future version
Abstract:The Global Ecosystem Dynamics Investigation (GEDI) mission measures three-dimensional forest structure with a LiDAR aboard the International Space Station. GEDI observations have geolocation errors of about 10 m, which cause misalignment with other remote sensing data and reduce the accuracy of downstream tasks. Matching GEDI ground elevations against high-resolution digital terrain models (DTMs) from airborne LiDAR is one way to correct these errors, but such DTMs exist in few regions. To widen this method’s applicability, we super-resolved a more widely available low-resolution DTM and used it for the matching. Our question is whether a super-resolved DTM yields the same correction as a 1 m-resolution airborne LiDAR DTM and, if not, why. In steep forests of western Japan, we super-resolved an existing 10 m-resolution DTM to 1 m. Improving its accuracy required fusion with a global 30 m-resolution DTM from a different source. Using the resulting DTM, we corrected GEDI observation positions first by a single shift per ground track, then per observation. For most tracks, the shifts nearly matched those from the airborne LiDAR DTM, likely because DTM errors canceled out across many observations. Per-observation corrections, however, did not match. At a single observation, the error does not cancel out and is too large for the correction. This error likely originates mainly from the input DTM. Therefore, even without a 1 m-resolution DTM, a roughly 10 m-resolution DTM, fused with a global DTM and super-resolved to 1 m, should yield track-wise corrections nearly matching those from a 1 m-resolution DTM.
[CV-105] AdaptivePnP: Fast Auto-Tuned Plug-and-Play for Deconvolution
链接: https://arxiv.org/abs/2611.09845
作者: Haley Duba-Sullivan,Emma J. Reid,Charles A. Bouman,Gregery T. Buzzard
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to IEEE Transactions on Computational Imaging
Abstract:Plug-and-play (PnP) methods reconstruct images by alternating a data-fitting step with an image denoiser. Their performance depends on two parameters: the step size \gamma and the measurement noise level \sigma . In practice, \gamma is often tuned by grid search over validation images and must be retuned whenever the problem setting changes, and \sigma is assumed known or estimated separately. We propose AdaptivePnP, a PnP framework that jointly estimates \sigma and \gamma at every iteration from the data-fitting residual. To do so, we apply the PnP virtual model to the data-fitting proximal map and show that the residual is zero-mean Gaussian with covariance parameterized by \gamma and \sigma . However, when one parameter is much larger than the other, the residual carries little information about the smaller one and using the maximum likelihood (ML) estimator can degrade the PnP iterations. We therefore introduce the identifiable maximum likelihood (IML) estimator, which restricts the ML estimator to an identifiable window in which the normalized Cramér–Rao bound of every parameter is bounded by a user-specified tolerance. For the virtual model, this window is an interval of ratios \rho = \gamma/\sigma , so computing the IML estimate reduces to a one-dimensional search over \rho . We apply AdaptivePnP to two standard PnP methods, yielding algorithms that require neither a tuned \gamma nor a known \sigma . On image deblurring and super-resolution, AdaptivePnP matches or outperforms PnP baselines tuned with the true \sigma . Code is available at this https URL.
[CV-106] Multiscale Closing Index (MCI) for measuring structural similarity between 2D point sets
链接: https://arxiv.org/abs/2611.09745
作者: Nicolò Zava,Angel Franco-Alvarez,Samy Blusseau,Santiago Velasco-Forero,Maria-Jose Jimenez
类目: Metric Geometry (math.MG); Computational Geometry (cs.CG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We introduce Multi-scale Closing Index (MCI), a new dissimilarity measure for evaluating the alignment of two 2D point sets. MCI is bounded between 0 and 1 and is based on multiscale mathematical morphology operations, with smaller values indicating greater structural agreement. The measure was motivated by the need to assess the quality of registration between point sets representing cell populations in biological images. To evaluate its behaviour under controlled conditions, we generated synthetic point clouds from Perlin-noise-based spatial structures, following established approaches for simulating biologically plausible spatial organization. We assessed MCI between those point clouds and different transformed versions, obtained by local perturbation, point removal, translation, rotation and addition of impulsive noise. Across these experiments, the dissimilarity MCI increased gradually and consistently with perturbation severity, while showing a clear discontinuity between clouds originating from the same underlying spatial structure and point clouds generated from different structures. We further compared MCI with commonly used point-set distances, including Hausdorff-based measures and Sliced Wasserstein distance. These experiments show the ability of MCI to capture structural agreement and quantify deviations between point sets, highlighting its potential as a complementary metric for evaluating two-dimensional point-set registration.
[CV-107] Beyond Segmentation: Enriching ImageCAS with Coronary Ostia Landmarks Rooted Centerline Graphs and Open-Weight Localization Models
链接: https://arxiv.org/abs/2611.09714
作者: Matteo Leccardi,Andrea Moglia,Marco Morandini,Michele Brembilla,Alessandro Nicosia,Andrea Penserini,Christian Brandini,Ender Konukoglu,Marco Marcon,Luca Mainardi,Pietro Cerveri
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 52 pages including the appendices. Dataset, centerlines, code and model weights: this https URL
Abstract:Coronary ostia and centerlines are key reference points for automated analysis of coronary computed tomography angiography (CCTA), from surgical planning to intraoperative vessel tracking. However, the only public ostium annotations cover 77 non-gated aortic scans, not CCTA, while CAT08 and ASOCA (32 and 40 CCTA scans) mark the ostia only implicitly, as centerline start points. This work enriches ImageCAS, the largest public dataset for coronary artery segmentation, with ostium annotations for its 1000 patients and centerline graphs rooted at these new ostia, making it, to the authors’ knowledge, the largest public dataset for ostium localization and the only one with centerlines following the official ImageCAS masks and starting at annotated ostia. Three annotators marked each ostium in the first 500 volumes; in the next 500, the candidates of a model trained on the first batch were corrected (human-in-the-loop). Six 3D localization models, trained on ImageCAS, CAT08 and ASOCA, were benchmarked on 58 held-out volumes (50 ImageCAS, 4 CAT08, 4 ASOCA). SwinUNETR-V2 and nnResUNet localized all ostia without failure (5 mm error), with median errors of 0.70-0.87 mm. Trained on CAT08 and ASOCA alone, SwinUNETR-V2 fails on 34% of right and 53% of left ostia, whereas adding 250 ImageCAS volumes removes all failures. The centerlines were extracted by a deterministic, quality-controlled minimal-path procedure; on ASOCA, they cover a median of 99.5% of the reference length, with a median per-patient mean distance of 0.22 mm between matched reference points and centerlines. Annotations, centerlines, code, and weights are available at this https URL.
[CV-108] Semantic Extraction Framework for Goal-Oriented V2X Communication Supporting Safe Intersection Passage
链接: https://arxiv.org/abs/2611.09118
作者: Haoran Yang,Shan Lu,Takaya Yamazato
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: Accept by 2026 Asia-Pacific Conference on Communications
Abstract:This paper proposes a transmitter-side semantic extraction framework for goal-oriented Vehicle-to-Everything (V2X) communication supporting safe intersection passage. Instead of transmitting raw sensory data, roadside infrastructure extracts task-relevant semantic information from video and traffic-signal sequences and organizes them into compact semantic packets for low-latency driving-decision assistance. The intersection-passing problem is formulated as a set of traffic tasks under Japanese traffic regulations, and the proposed pipeline generates perception and intent semantics through object detection, object tracking, region-of-interest (ROI)-based semantic extraction, and recommended-action generation. The framework is implemented and evaluated on a self-constructed, CARLA-based single-vehicle dataset collected under clear weather conditions. Experimental results show high recommended-action accuracy within the tested task set, while the mean processing latency remains below the frame interval of the 15~fps input video.
[CV-109] Measurement-Aware Multimodal LLM for No-Reference Low-Dose CT Quality Assessment
链接: https://arxiv.org/abs/2611.08754
作者: Kerem Basim,Tolga Demiroglu,Mehmet Ozan Unal,Metin Ertas,Isa Yildirim
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages, 2 figures, 3 tables. Submitted to IEEE ISBI 2027
Abstract:No-reference image quality assessment (IQA) of low-dose CT predicts the score that radiologists would assign to a slice without a full-dose reference, as needed to compare dose protocols and denoisers. Trained CT IQA networks agree well with radiologists but output only a score, whereas multimodal large language models (MLLMs) can explain a score but perform poorly on CT when only prompted. We propose a measurement-aware MLLM, built on Qwen3.8-27B, that rates a slice from its image together with six image-domain measurements of noise, sharpness, high-frequency content and streaks, given to the language model as percentiles in text. On the LDCT-IQA 2023 test set, it reaches an Overall score of 2.735, on par with the retrained challenge winner (2.735) and MedSigLIP-FiLM (2.713), and substantially above its base MLLM used as a zero-shot agent. Shuffling the measurements at test time reduces the score by 0.048, mostly through noise, but the same measurements also lead the model to assign higher scores to over-smoothed denoised images.
[CV-110] Deep-Learning Algorithm for Indeterminate Thyroid Nodules: Comparison and Combination with Molecular Testing
链接: https://arxiv.org/abs/2611.05400
作者: Jichen Yang,Jeffrey Ashton,Jikai Zhang,Maciej A. Mazurowski,Benjamin Wildman-Tobriner
类目: Quantitative Methods (q-bio.QM); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注: 16 pages, 7 figures, 8 tables (3 supplementary)
Abstract:Thyroid nodules with indeterminate cytology (Bethesda III) carry an intermediate risk of malignancy, and molecular testing is widely used to decide between surgery and surveillance, but it is costly and not universally available. We investigated whether a convolutional neural network (CNN) applied to routine ultrasound images can stratify Bethesda III nodules, and how it compares and combines with molecular testing. We retrospectively identified 466 nodule records from 418 patients who underwent molecular testing (Afirma or ThyroSeq). A CNN pre-trained on determinate nodules was adapted to indeterminate nodules with soft labels and transfer learning. After removal of duplicate records, 119 nodules with surgical pathology were used for development and 47 were held out for testing. On the test set, the deep learning (DL) model achieved an AUC of 0.73 (95% CI 0.51-0.94), compared with 0.55 (0.38-0.71) for molecular testing (P = 0.15). At a cut-off inherited from the original algorithm, the DL model was more specific than molecular testing (0.95 vs 0.32; P 0.001), with a higher positive predictive value (0.67 vs 0.21), a lower sensitivity (0.44 vs 0.78; P = 0.25) and a similar negative predictive value (0.88 vs 0.86). All six DL-positive test nodules were molecular-positive, and the DL output split the molecular-positive nodules into a high-risk (4/6 malignant) and a lower-risk group (3/27 malignant; P = 0.01). Surgical verification was strongly driven by the molecular result (99% of positive vs 19% of negative results), and molecular-test accuracy depended on the platform. A logistic regression of the DL output, nodule size and the molecular result dropped the DL term and did not improve on the DL output alone. Ultrasound DL complemented molecular testing for Bethesda III nodules, in particular by identifying molecular-positive nodules at lower risk of malignancy.
人工智能
[AI-0] CPU-ARC: Agent ic Reasoning Reasoning Caching for on-CPU Reasoning Reuse
链接: https://arxiv.org/abs/2611.09891
作者: Peter Belcak,Greg Heinrich,Saurav Muralidharan,Shizhe Diao,Pavlo Molchanov
类目: Artificial Intelligence (cs.AI)
备注: Tech report. 28 pages, 9 figures, 11 tables
Abstract:Large language models reinvent the wheel all the time. They conflate task-solving with the execution of the solutions, which forces them to solve every problem from scratch. We show that once a task has been reasoned through, the application of its solution to a new problem instance can be done very efficiently and with little involvement from language models. We introduce agentic reasoning caching, a method for constructing robust CPU-executable agentic reasoning caches (ARCs) from LLM reasoning traces. ARCs show on-par performance with frontier LLMs across a range of difficult mathematical and scientific tasks, are fully explainable, continually trainable, robust to input perturbations, and 16-291x cheaper to invoke than the source model.
[AI-1] On the Reliability of Majority-Vote Pseudo-Labels in Test-Time Reinforcement Learning
链接: https://arxiv.org/abs/2611.09872
作者: Haoyu Wang,Ru Wang,Yuanyuan Wang,Jiaxian Guo,Mingming Gong
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Test-time reinforcement learning (TTRL) adapts pretrained language models to unlabeled target-domain data using self-generated pseudo-labels, whose reliability remains poorly understood. We analyze majority-vote supervision for TTRL with online Direct Preference Optimization (DPO) and decompose its error into population-majority bias and prompt-dependent finite-rollout variation. Because population-majority bias is generally unidentifiable without external supervision, we focus on mitigating the downstream effects of finite-rollout variation. We propose \emphPrompt-Conditional Distributionally Robust DPO (this http URL), which conditions within-prompt robust reweighting on observed vote reliability to make updates from sampling-sensitive prompts more conservative. We further introduce a cache-based mechanism to reduce self-supervised forgetting during online adaptation. Across six mathematical and knowledge-intensive reasoning benchmarks and five model backbones, this http URL matches or outperforms the strongest TTRL baseline in 24 of 30 settings.
[AI-2] Grasp-TAGS: Task-Aware Grasp Selection via Critic Distillation for Prehensile Loco-Manipulation
链接: https://arxiv.org/abs/2611.09864
作者: Xun Tu,Minghao Zou,Karthik Desingh
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 10 pages, 6 figures
Abstract:Loco-manipulation, which requires mobile manipulators to coordinate locomotion and manipulation over long horizons, is a key capability for large-object rearrangement in applications such as service robotics and warehouse automation. Despite recent progress, reliably executing these tasks remains challenging because performance depends on complex robot-object and object-environment interaction dynamics. Existing approaches address this challenge either through model-based methods, which often require accurate dynamics models, or through learning-based methods, which reduce this requirement but typically assume a fixed or predefined initial grasp. However, different geometrically stable grasps can lead to substantially different downstream manipulation performance. Therefore, we propose Grasp-TAGS, a task-aware grasp selection framework for goal-conditioned large-object rearrangement. We train a unified PPO-based reinforcement learning policy across multiple initial grasp configurations and distill the learned critic into a lightweight grasp utility network that ranks candidate grasps using only the candidate grasp configuration and goal information. At inference time, we select the grasp with the highest predicted downstream utility for the downstream policy to complete the task, before even making a contact. Experiments demonstrate that Grasp-TAGS improves rearrangement success over both fixed single-grasp policies and random grasp selection, achieving an average success rate of 89.84% compared with 79.87% for the best-performing single-grasp policies, an improvement of 9.97 percentage points. Grasp-TAGS also reduces final orientation errors across all evaluated objects. For more qualitative results, refer to the supplementary webpage: this https URL
[AI-3] Ensemble Methodology for Reference-Free Confidence Estimation in LLM -based Code Summarization
链接: https://arxiv.org/abs/2611.09846
作者: Marco Colombo,Balaji A Jayaram,Sowmyakumari Selvaraj,Sumit Tokkar,Animesh Kumar,Aravinda Prasad Sistla,Ravishankar Shetty,Rigel Gjomemo
类目: Artificial Intelligence (cs.AI); Programming Languages (cs.PL)
备注: 9 pages, 3 figures, 1 table
Abstract:This paper describes an approach and use case to estimate the correctness of LLM-generated COBOL summaries without access to human-authored references. The core intuition is that inter-model agreement on semantics is informative about truth, and that an LLM-as-judge can provide higher-fidelity pairwise assessments than traditional string- or embedding-based metrics. We operationalise this intuition with (i) controlled summary generation from diverse models; (ii) a LLM evaluator that compares pairs of summaries and emits structured judgments; and (iii) a calibrated aggregation rule that combines supportive evidence while penalising only corroborated contradictions.
[AI-4] Can Agent ic Systems Close the Gap in Radiology with Production AI Tools?
链接: https://arxiv.org/abs/2611.09836
作者: Aubin Charley,Julien Khlaut,Charles Corbière,Pierre Manceron,Corentin Dancette,Léo Machado
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Radiology AI has developed along two paths: specialized models that perform well on narrow tasks but cover few clinical findings, and generalist models that address a broader range of tasks but remain prone to hallucination and inaccurate measurements. Combining the two through agentic systems offers a route towards comprehensive radiology reporting, but published systems are tied to one modality or body part. We describe a radiology agentic framework that standardizes how an orchestrator calls modular tools to produce a structured report, and use it to measure what such a system can do with the tools from FDA-cleared radiology AI devices. Report generation is simulated on a large multi-modality report corpus, with the orchestrator given either a catalog of finding classes consolidated from the cleared devices, each available on the modalities it is cleared for, or an unlimited catalog in which any class can be requested, and planning either from the clinical indication or from the target report. Tool outputs are simulated from the reference report, so the comparison isolates tool availability and orchestration from tool error. With the cleared catalog and the indication alone, the orchestrator recovers under a third of the clinically actionable findings; recovery improves substantially with the unlimited catalog, and further when the plan is written from the target report. Attributing every missed finding to a cause shows that most are classes no cleared device offers for the study’s modality, and that a third are classes or tools that were offered and not called. Adding the most frequently missing tools closes most of the gap to the unlimited catalog on held-out reports, whereas a smaller orchestrator with the unlimited catalog gives up the whole gain. Tool availability and effective orchestration are thus complementary constraints on comprehensive AI-assisted radiology reporting.
[AI-5] Generative Latent Priors for Interpreting and Steering Protein Language Models NEURIPS2026
链接: https://arxiv.org/abs/2611.09818
作者: Daniel Hinjos,Lucía Urcelay
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at the Interpretability for Discovery (Interp4Discovery) Workshop at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)
Abstract:Dominant methods for interpreting large models impose strong structural assumptions. Generative Latent Priors (GLP) are a complementary approach: a diffusion model that learns the distribution of a frozen model’s activations directly, imposing no assumptions. We port GLP to ESM-2 activations (UniRef50) and evaluate it at two host scales (ESM-2-8M and -650M), both as a lens for protein representations and as a prior for editing them. Generation quality scales smoothly with depth and compute. As a lens, single GLP “meta-neurons” are the strongest probes of discrete Swiss-Prot concepts at both hosts, beating InterPLM’s released SAE features, and lead on continuous properties at 8M though not at 650M; falsification tests show this is mostly concentration, not new information. As an editing prior, the on-manifold correction makes DiffMean steering of three biochemical properties effective and restores on-manifoldness, but only in isolation: re-embedding the sequences actually produced after several rounds of decoding, the structural plausibility is compromised. This is mainly reproduced in additional SAE-based clamping structural steering experiments, where only Helix motif succeeds. The naturalness failure is traced to the rewriting coverage and the decoder rather than the prior, highlighting what the method must clear before it can serve for both discovery-by-editing and discovery-by-exploration. Trained GLPs are released on this http URL.
[AI-6] Scaling Graph Similarity Learning with Graph Neural Networks
链接: https://arxiv.org/abs/2611.09805
作者: Ahmet Tüzen,Helge Langseth,Kjetil Nørvåg
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Graph similarity computation is an important task in graph mining with many real-world applications, such as chemical compound analysis, program dependence comparison, and social network analysis. The problem is computationally challenging: Graph Edit Distance (GED) and Maximum Common Subgraph (MCS), which quantify the distance between graphs and identify their largest shared subgraph, respectively, are NP-hard to compute exactly. To address these computational challenges, Graph Neural Networks (GNNs) have been applied to this problem. However, existing GNN-based approaches have primarily focused on small-scale graph datasets, while scalability to larger scales remains underexplored. In this work, we propose FiLSim, a scalable GNN-based framework, and investigate whether a simplified GNN pipeline can maintain competitive performance on standard benchmarks while scaling to large graphs. We additionally propose a sequential graph-editing strategy to construct large-scale graph datasets with reliable GED labels to evaluate scalability. The proposed framework is evaluated on four benchmark datasets and four newly created large-scale datasets, demonstrating its effectiveness and scalability. FiLSim remains accurate on graphs with more than 150,000 nodes, where the baseline methods run out of memory or fail to learn, and scores all test pairs within seconds on a single GPU.
[AI-7] When Instruction Files Go Stale: Documentation Debt and Hidden Cost in Coding Agents
链接: https://arxiv.org/abs/2611.09802
作者: Tural Mehtiyev,Biruk Tadesse,Lawrence Arkoh,Wesley Assunção
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 11 pages, 5 figures, 3 tables. Replication package: this https URL
Abstract:LLM-based coding agents such as Claude Code are guided by repository instruction files (also called context files or agent configuration files) like this http URL and this http URL, which are read in every agent session. As any documentation artifact, these files can fall behind the code and become stale (e.g., a test folder moves, but the file still points to the old one). In technical debt terms, such a stale instruction is documentation debt, and its interest is the cost it adds to every later agent session. Evaluations that check only task success can miss this cost, and the stale file can persist. In this study, we ask what leads agents to act on a stale instruction, how it affects their behavior and at what cost, and how they recover. We study three real-world stale instructions from active open-source repositories in depth, each a different kind of guidance: a test command, a code location, and a procedure for adding a component. Agents fix 22 maintenance tasks under three versions of each instruction (left stale, corrected to the current path, and removed), with the code and the task held fixed, and we analyze 732 trajectories of two models at the level of individual tool calls. We find that agents follow stale guidance when it is relevant to their tasks, sometimes even when correct information is available elsewhere. On the stale test command, Claude Sonnet 5.5 needed about two extra tool calls per run, a third more than with the corrected command; in the other cases, we could not detect a difference. Although agents generally recovered from the immediate failures, none repaired the underlying instruction file. Correcting the instructions eliminated operations on obsolete paths, while removing them prevented most. Our findings show that task success alone can hide the costs of documentation debt, highlighting the need to evaluate agent behavior beyond final outcomes.
[AI-8] st-Time Compute Scales Better with Parallel Agent Teams
链接: https://arxiv.org/abs/2611.09798
作者: Junlin Wang,Haozhe Chen,Ang Li,Rahul Chalamala,Zhuang Liu
类目: Artificial Intelligence (cs.AI)
备注: 29 pages, 18 figures, 4 tables. Code: this https URL
Abstract:Test-time scaling improves the performance of a fixed model by spending more computation at inference. There are two common ways to spend it on agents when a verifier can score every attempt. Best-of-N sampling runs many independent agents and keeps the best result but saturates quickly. One large team of collaborating agents can score higher but stops improving as it grows. Can we get the best of both? Our method, parallel agent teams, does so by sampling many small teams in parallel and keeping the best team’s result. We test it on 12 long-horizon tasks from GPU kernels and formal proofs to exploit development. With a 27B model and a budget of 100 agents, parallel teams of ten beat parallel agents on 10 of 12 tasks and beat one team of 100 on all four tasks where we ran such large teams. At equal dollar cost the same parallel teams beat a frontier model on half the tasks. A simple statistical model shows that individual agents lose because their many samples have a low mean and that one large team loses because a single sample gets no gain from variance. Sampling many small teams in parallel balances a high mean with the gain from variance. Parallel agent teams thus offer a more effective way to scale test-time compute.
[AI-9] Knowing When to Intervene: Causal Evaluation of AI Agents
链接: https://arxiv.org/abs/2611.09792
作者: Ayush Sawarni,Wenqian Xing,Tijana Zrnic,Ramesh Johari,Vasilis Syrgkanis
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Econometrics (econ.EM); Methodology (stat.ME)
备注:
Abstract:Current evaluations of agentic AI systems primarily rely on benchmarks. However, the static nature of benchmarks makes them uninformative about the value of auditing interventions at deployment time. For reliable deployment, we often need to determine within an agent’s trajectory whether or not to intervene, e.g., by verifying the agent’s current output or steering it toward different behavior, using either human review or more costly models or time-consuming agentic supervision. The value of such interventions is fundamentally a causal inference question, asking whether the causal effect of such an intervention is beneficial for the final outcome. Estimating this causal effect is particularly challenging due to the dynamic nature of the problem: the impact of an intervention depends on the current context and environment, as well as the agent’s future behavior. We develop the first methodologically grounded framework for evaluating the causal effects of randomized interventions within agent trajectories. By leveraging a relatively small dataset of agent trajectories collected under randomized interventions such as verification or screening, we can estimate the dynamic causal effect of a rich class of (contextual) intervention policies. The structural parameters that we estimate can also be used to construct safety leaderboards that capture the risk of various agentic actions, which can be used by practitioners for audit prioritization. Applying our methods to tau-bench Retail, ToolSandbox, and ImpossibleBench, we find that intervention effects vary substantially across agent actions and states. More broadly, our work suggests a rich research agenda at the intersection of causal inference and agentic evaluation, which we believe is essential for the reliable and safe deployment of AI agents in the real world.
[AI-10] Continuous wavelet-based feature extraction for laser-induced breakdown spectroscopy classification of human plasma samples
链接: https://arxiv.org/abs/2611.09746
作者: Md Mahmudur Rahman,Atqiya Munawara Mahi,Nazim Ahmed Belabbaci,Mohammad Arif Ul Alam
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Extracting discriminative features from laser-induced breakdown spectroscopy (LIBS) spectra is important for the classification of complex biological samples. This study evaluates continuous wavelet transform (CWT)-based spectral peak extraction for distinguishing human plasma samples labeled as SARS-CoV-2-positive or negative. Plasma specimens from 181 individuals, comprising 53 positive samples and 128 negative controls, were deposited as 5 \muL droplets on silicon substrates and dried before analysis. Spectra were acquired using a nanosecond Nd:YAG laser operating at 1064 nm, with 100 single-shot measurements collected per specimen and averaged after intensity-based outlier rejection. CWT-based peak detection was coupled with a support vector machine (SVM) classifier. The reported leave-one-out evaluation yielded an accuracy of 93.00%, sensitivity of 94.67%, and specificity of 91.61%; these estimates remain provisional pending reconciliation with specimen-level predictions. Features identified by the CWT pipeline included emission lines assigned to Ca II, K I, and N I. The workflow provides a basis for examining spectral features associated with the studied plasma groups. Independent validation and assessment of specimen, substrate, and acquisition-related variability are required to establish generalizability and diagnostic relevance.
[AI-11] Safety Is Π_1 Provability Is Σ_1: One Framework for the Hugging Face Breach and OpenAI s Mathematical Proofs
链接: https://arxiv.org/abs/2611.09736
作者: Jinho Kang,Jinsook Kim
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Logic in Computer Science (cs.LO)
备注:
Abstract:OpenAI’s agents breached Hugging Face in July 2026, and the same company released more than seven hundred AI-generated manuscripts of mathematical proofs in October 2026. The first event is read as a failure in security, while the second is read as a success in mathematics. We read both through the hybrid framework of our paper (Kim and Kang, 2026, arXiv:2611.01914), in which coding, proving, and learning from data are derivations by one and the same deductive mechanism, distinguished by the kinds of premises from which they are derived. Under this framework, the AI systems behind the two events ran the same loop, in which a statistical proposer produces candidate exploits or proofs and a deductive verifier settles each one by a derivation. What separates the events is the logical form of the statement wanted by each of the two domains, security and mathematics. The safety of a program is a \Pi_1 statement, for it says that no exploit exists, which is not assured by any finite list of tests that have been passed by the program. By contrast, the provability of a sentence is a \Sigma_1 statement, for it says that a proof exists, which is assured by a single instance, namely a proof that has been checked by the deductive verifier. The theorem behind both is Turing’s Undecidability Theorem: the halting set is \Sigma_1 but not decidable, so its complement is \Pi_1 and not \Sigma_1 , which is to say that membership has a finite witness but non-membership has none. We prove that attaching a statistical proposer to a deductive verifier does not change this undecidability result, however good the proposer is. We then discuss what does assure safety and where the residual risk goes.
[AI-12] FOSCAR: Accelerating Decode Attention over 2- and 4-Bit KV Caches through Affine Folding
链接: https://arxiv.org/abs/2611.09722
作者: Donglin Zhuang,Pragaash Ponnusamy,Xiaoxia Wu,Shuaiwen Leon Song,Zhongzhu Zhou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large-language-model decode attention streams the key-value (KV) cache for every generated token. Packing that cache into 2-bit or 4-bit quantized integers reduces memory traffic but does not automatically translate into latency saving. Existing 2-bit fused kernels reach under 31%/15% of peak HBM throughput on H100/B200. Converting each value and applying its scale and offset before feeding Tensor Cores is a major bottleneck because instruction-issue throughput is scarce on modern GPUs. Instead of applying scale and offset to each KV value, FOSCAR uses affine folding to apply them to the other operand or the output of decode-attention matrix products, depending on the quantization scheme (token-wise or channel-wise). This reduces scale-and-offset operations by up to two orders of magnitude while preserving the dequantized attention output in exact arithmetic. FOSCAR also uses bit manipulation to feed KV operands to Tensor Cores efficiently, requiring only 1 and 0.19 ALU instructions per INT2 element on H100 and B200, respectively. Finally, FOSCAR builds warp-specialized GPU pipelines for Hopper (H100) and Blackwell (B200) using wgmma and tcgen05. On H100, FOSCAR INT2 and INT4 kernels achieve geo-mean speedups of 3.36\times and 2.83\times over FlashInfer BF16, and 2.03\times and 1.71\times over TensorRT-LLM FP8. On B200, the corresponding INT2 and INT4 speedups are 2.67\times and 2.30\times over FlashInfer BF16, 1.48\times and 1.28\times over TensorRT-LLM FP8, and 1.16\times and 1.00\times over TensorRT-LLM NVFP4. We integrate FOSCAR into SGLang to benchmark serving efficiency. On long-context serving workloads, FOSCAR increases peak aggregate throughput by up to 170% and 341% over 2-bit and 16-bit baselines, respectively. We open-source FOSCAR as a kernel library at this https URL.
[AI-13] Intent-Aware Caching for Efficient LLM Serving NEURIPS2026
链接: https://arxiv.org/abs/2611.09715
作者: Kazi Hasan Ibn Arif,JinYi Yoon,Dimitrios S. Nikolopoulos,Hans Vandierendonck,Deepu John,Bo Ji
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: NeurIPS 2026 | Code: this https URL
Abstract:Modern LLM serving systems use prefix caching to accelerate multi-turn conversational workloads. They cache the key-value states of previous conversation prefixes and reuse them when subsequent requests share the same prefix, to avoid redundant prefill computation and reduce serving latency. However, achieving a high cache hit rate remains challenging because cache reuse depends on whether a conversation continues and how soon the next turn arrives. Our analysis of real workloads shows that user intent is a useful signal for this behavior, as different intents exhibit different continuation probabilities and inter-turn delays. Based on this key insight, we propose Intent-Aware Caching (IAC), an online eviction policy that classifies requests by intent, learns lightweight reuse statistics from recent serving logs and leverages these statistics to guide an oracle-inspired eviction strategy. Extensive evaluation on real workloads (WildChat and ShareChat) shows that IAC reaches a 34.6% cache hit rate. This substantially outperforms both the previous learning-based approach (18.8%) and traditional methods such as least recently used (LRU, 12.2%), while reducing average serving latency by up to 46.1%.
[AI-14] ICare-DT: a digital twin for continuous forecasting of intensive care patients
链接: https://arxiv.org/abs/2611.09712
作者: Yanke Li,Yahang Qi,Yumi Kim,Diego Paez-Granados
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Digital twins of intensive care patients should update forecasts from new data and predict treatment responses, but few studies test whether routine records support either. We propose ICare-DT, a Transformer-based twin that updates a latent patient state hourly from new measurements without retraining. We trained it on 140,473 eICU-CRD stays to forecast vital signs crossing their training-set means. On test data, later forecasts beat the twin’s own hour-6 forecasts of the same windows in all 12 settings (+0.073 to +0.153 AUROC). On validation data, controls that withheld or swapped the new hours traced the gain to the patient’s own measurements. Without refitting, the model kept its gain on a second database (8 of 9 settings, +0.025 to +0.114 AUROC) in a prespecified evaluation of the previously analyzed MIMIC-IV test split. Refitted within MIMIC-IV, a twin state fine-tuned on clinical endpoints gave gradient boosting a significant gain for acute kidney injury (+0.0035 AUROC; hypotension +0.0010, not significant). To test response prediction, we built a benchmark that replaces the treatments in real records with simulated doses and injects responses of known size, visible only at measurement times. Ridge models supervised with injected effects recovered them from real records only through frequently measured variables (heart rate) read at their measurement times. We asked standard learners to predict these effects for a dosing regime no patient had received. No standard learner outperformed predicting zero effect at an out-of-support query. When told how the simulated responses were generated, a separate learner recovered the effects from the same records without seeing their true values (R2 0.30 to 0.40 at one effect size). Routine records support forecasts that improve with each patient’s new data, and our benchmark is a testbed for models that learn how responses are generated.
[AI-15] Empirical Assessment of Post-Quantum Cryptography Adoption Status in TLS
链接: https://arxiv.org/abs/2611.09707
作者: Md Ariful Haque,Sander Vinberg,Ben Edwards,João Cruz
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Anticipating Q-day security concerns, major CDNs and cloud providers are migrating to post-quantum cryptography (PQC). While existing literature benchmarks PQC performance, empirical research evaluating real-world adoption data remains scarce. We present an empirical assessment of PQC deployment across seven critical protocols (HTTPS, LDAPS, SMTPS, IMAP-SSL, POP3-SSL, TELNETS, and RDP) using 10,000 target IP addresses per protocol. At the origin-server level, PQC algorithm support reaches 9.1% for HTTPS and 7.39% for LDAPS, while falling below 1% for all other evaluated protocols. At the entity level, PQC support is 11% for HTTPS and 5.4% for LDAPS, while the remaining protocols have less than 5% of entities supporting PQC on the hosted servers. We further correlate these adoption trends against organizational digital and financial footprint, and industry sectors, yielding a multi-dimensional empirical assessment of PQC deployment. Finally, at the domain level, we perform a comparative analysis of PQC adoption in Cloudflare-hosted domains against baseline Cloudflare datasets. Collectively, our findings provide critical insights into the current state of PQC readiness over Transport Layer Security (TLS).
[AI-16] Empirical Study of Post-Quantum Cryptography Adoption Status in SSH Infrastructure
链接: https://arxiv.org/abs/2611.09703
作者: Md Ariful Haque,Sander Vinberg,Ben Edwards,João Cruz
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:With the anticipated arrival of Q-day, quantum computers threaten current cryptographic schemes for encryption and digital signatures. In response, major content delivery networks and cloud service providers are already adopting Post-Quantum Cryptography (PQC). The current state-of-the-art focuses on benchmarking the operational performance of implementing PQC, but the real-world adoption status remains under-researched. This work utilizes large-scale SSH fingerprint data from a proprietary scanning tool to analyze PQC adoption across 1.52 million hosts and 27,012 entities. Our study shows that the PQC transition in the SSH protocol is underway but highly polarized: 45.2% of SSH hosts on port 22 support PQC, while 54.8% remain classical-only. At the entity level, 39.9% have enabled PQC on at least one host, whereas 60.1% have no PQC support. By correlating PQC adoption with industry, geography, and organizational size (digital and financial), this study provides a comprehensive empirical assessment of PQC adoption status in SSH infrastructure.
[AI-17] Attend at the Right Resolution: Head-Adaptive Sparse Attention via Query-Key Interaction Structure
链接: https://arxiv.org/abs/2611.09700
作者: Wajih Hassan Raza,Ahsan Bilal,Yidan Shen,Renjie Hu,Xin Fu,Yu Wen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Attention is a major scalability bottleneck in large language models (LLMs), with its computational cost and key-value (KV) memory traffic becoming increasingly prohibitive as context length grows. Sparse attention addresses this bottleneck by restricting each query to a subset of the context, but its effectiveness hinges on identifying relevant regions both cheaply and reliably. Existing training-free methods primarily rely on runtime relevance estimates, with selection dimensionality and granularity typically governed by fixed heuristics rather than the intrinsic interaction structure of pretrained attention heads. We take a different perspective: the query-key operator of each pretrained head defines a characteristic response structure that determines the resolution needed for sparse selection, both across feature directions and across context tokens. Building on this insight, we propose Head-Adaptive Sparse Attention (HASA), a training-free sparse attention framework that jointly adapts selection resolution along these two dimensions. Along the feature dimension, the singular spectrum of each head’s query-key operator determines an effective rank, yielding a compact, head-specific response space. Along the sequence dimension, neighboring regions are progressively merged according to the directional coherence of their token responses in this same space, while heterogeneous regions remain separate. Exact attention is then computed using the original full-dimensional KV states of the selected regions. HASA captures 88.1% of dense-attention mass and, compared with UNIQUE, a strong recent sparse-attention baseline, improves accuracy by up to 3.47 and 2.57 points on RULER and LongBench-Pro, respectively, and accelerates self-attention by up to 1.32x, while achieving up to 14.07x speedup over dense attention.
[AI-18] Verifier Noise Is a Temperature: Testing Tilt Predictions in Finite-Budget LLM -Guided Search
链接: https://arxiv.org/abs/2611.09699
作者: Xiaomeng Li
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 34 pages, 5 figures. Pre-registered predictions, code and data checksums are described in Appendix C
Abstract:LLM-guided search loops propose a modified candidate, score it, and keep it if it scores better. When both candidates are re-scored at every comparison, as in the minibatch acceptance gates of prompt optimizers, keep-if-better is a reversible Markov chain if the score difference is logistic and the mutation operator is reversible, and its stationary law is the mutation equilibrium tilted by exp(beta f). For any continuous symmetric noise difference the same tilt is the stationary law of a small-step diffusion limit, with beta = 4F’(0); for a minibatch gate beta grows with the square root of the minibatch size. Caching the incumbent’s score instead inflates the recorded score. We test the tilt as a prediction for LLM-guided program search, committing each prediction from calibration data before the search runs. The level after 150 steps rises with the minibatch size in six of seven settings, but its predicted value fails: on bin packing, search attains 62-88% of the predicted gain when the proposer edits the parent and 109-119% when it writes new heuristics. The attained fraction falls with the parent-child quality correlation of unselected mutation; a registered test on two new settings confirmed the direction, but not the quantitative agreement after a correction. Further registered tests found that runs started from a good program kept losing quality over 300 steps, more so from starts that search had selected, by a margin that passed its registered rule narrowly. Runs from different starts had not reached a common level, so whether the tilt describes the stationary law of LLM editing remains open. A registered test in prompt optimisation on frozen task-model scores failed: search exceeded the tilt, and the bin-packing relation did not transfer.
[AI-19] SURE:Spatiotemporal Uncertainty Representation for Efficient Vision-Language-Action Learning
链接: https://arxiv.org/abs/2611.09684
作者: Jingye Zhang,Choo Sin Wai,Dake Zhong,Zhifang Liu,Yuhong Zhang,Siyao Xiao,Guorui Song,Guocun Wang,Dongbin Zhang,Peng Jiao,Haoqian Wang
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-language-action (VLA) policies have become a promising paradigm for end-to-end robot control, but their effectiveness still depends heavily on large volumes of demonstration data. In manipulation tasks, this dependence is compounded by structured spatiotemporal uncertainty in the action space: multiple nearby actions or trajectories may successfully solve the same task, while semantically equivalent demonstrations often unfold at different speeds. Standard pointwise supervision treats each demonstrated action as an exact target, imposing an unnecessary learning burden under such uncertainty. We present SURE, a data-efficient VLA training framework that explicitly models both spatial and temporal uncertainty. SURE augments a VLA backbone with a lightweight transposed-convolutional probabilistic action decoder that captures spatial uncertainty through local correlations among neighboring action bins. It further introduces Soft-DTW alignment supervision to accommodate local timing variations between predicted and demonstrated action chunks. Together, these two components reduce sensitivity to demonstrator-specific poses and execution timing, enabling more data-efficient action learning. We evaluate SURE on LIBERO, RoboTwin 2.0, and real-world manipulation tasks. Even with only 10% of the demonstration data or 20% of the training steps, SURE achieves manipulation success rates that match or surpass baseline VLAs.
[AI-20] Goodhart in the Planner: Sampling-Based Control Exploits Rewards Learned from Human Preferences
链接: https://arxiv.org/abs/2611.09670
作者: Soovadeep Bakshi,Apurba Bose
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 19 pages, 4 figures. Equal contribution. Code and results will be released
Abstract:Robots increasingly choose actions by searching against a reward model learned from human preferences, in model-predictive control or best-of- N selection; we measure what this does to the true objective. On HalfCheetah and Hopper with crowd-sourced labels, and on Meta-World drawer-open and plate-slide with synthetic labels and blind labels from the two authors, harder MPPI search cuts true return by up to 76%: with human labels the drawer opens in 0 of 20 runs against 10 of 10 for the cloned policy the planner starts from, while held-out label accuracy stays at 0.75-0.97. On plate-slide human labels are exploited less than synthetic ones, and on window-open search helps. True return roughly follows how far the planner’s weights move from its proposals ( \sqrt\mathrmKL ) for both weighting rules we test, and best-of- N with a verifier from synthetic labels falls below a single noisy action. We prove that \chi^2 -regularized MPPI at a temperature set by the scorer’s error spread on the proposals is worst-case optimal, that no positive gain can be certified once that spread exceeds the scorer’s own, and that an episode’s return splits exactly into per-step terms, of which the guarantee covers the selection terms. Calibrated from 100 proposal pairs ranked by true return (an oracle) on start states the evaluation never uses, this rule is the safest fix we test: with each of five independent sets it succeeds in every arm run and keeps on average 86-94% of the cloned policy’s return on plate-slide and drawer-open, though it is cautious where search helps. Calibration from a person’s comparisons cannot detect a bias the scorer learned from that person; correcting for the pairs people cannot tell apart, or a second person, limits the damage.
[AI-21] Differentially Private GraphRAG : Protecting Relation Connectivity and Semantics
链接: https://arxiv.org/abs/2611.09658
作者: Jie Fu,Yuan Hong,Yue Ning,Wendy Hui Wang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Graph Retrieval-Augmented Generation (GraphRAG) connects information across documents through entity relationships to support relational reasoning. However, black-box access to GraphRAG systems can enable adversaries to reconstruct knowledge graphs and recover both structural connectivity between entities and the semantics of their relationships. Existing defenses lack formal privacy guarantees and provide limited protection against such reconstruction attacks. We propose RelGuard, the first differentially private framework for GraphRAG, which provides formal privacy guarantees by mitigating connectivity leakage and protecting relations and their semantic meaning through relation removal. We formalize relation differential privacy (relation-DP), which bounds the distinguishability of mechanism outputs according to the semantic distance between alternative relations while accounting for relation removal. Building on this definition, we develop a four-step defense mechanism that sanitizes source documents before GraphRAG indexing by perturbing relation predicates or removing their associated connectivity using the exponential mechanism. Experiments across three GraphRAG systems and two targeted reconstruction attacks demonstrate that RelGuard provides stronger protection against reconstruction attacks than four existing defense mechanisms while better preserving downstream utility.
[AI-22] Not Your Diary Just Your Topic: Grading the Wrong-User Control NEURIPS2026
链接: https://arxiv.org/abs/2611.09609
作者: Karina Romanova,Igor Kiselev,Vladimir Platonov,Viliana Devbunova,Gregory Leleytner,Adeliya Leleytner,Lina Veltman,Aleksandr Nikolich
类目: Artificial Intelligence (cs.AI)
备注: 21 pages. NeurIPS 2026 Workshop on Personalized, Aligned, Long-Term Memory for AI Systems (PALM)
Abstract:A common test of personalization replaces a user’s history with someone else’s and measures the drop in accuracy. But this swap also changes the topics in the history, so the drop may reflect topic mismatch. We examine this problem with a frozen language model on LaMP-2 news categorization, using histories from similar users and synthetic histories assembled from other users’ articles. The user’s own history improves accuracy by 2.3 percentage points over a model trained without history. Removing history only at evaluation makes this gain appear larger, at 4.8 points. Category labels alone reproduce 86% of the accuracy gap between the user’s own history and a random user’s history. With synthetic histories matched on labels and text similarity, the gap falls to 0.9 points. Across two user splits and three training seeds, the advantage over a similar user’s history is 3.2 points for users seen during training and 0.65 points for unseen users, where the evidence for an advantage remains inconclusive. A small movie-tagging experiment with one training seed also finds no advantage over other users’ histories. These findings show why personalization should be evaluated on unseen users, with histories of similar content and a model trained without history as comparisons. We also find that adding gated cross-attention improves on a prefix-only model under our training setup; whether this holds more broadly remains open.
[AI-23] Alignment by Confident Debate: Preference Learning from Multi-Agent Debate
链接: https://arxiv.org/abs/2611.09568
作者: Nikhil Verma,Dingxian Wang,Mohammad Reza Rezaei,Andrew Rabinovich,Zhao Chen
类目: Artificial Intelligence (cs.AI)
备注: 53 pages, 9 Figures, 31 Tables
Abstract:Post-training language models for reasoning typically rely on gold answers, task-specific verifiers, or learned reward models, limiting scalability when such signals are scarce. Multi-agent debate offers a self-supervised alternative, but converting debate into reliable training supervision remains challenging: majority voting weights agreeing responses equally, lacks a principled tie-breaking rule, and is ill-defined for open-ended tasks without a canonical answer; likelihood-based scoring can conflate fluency with correctness; and homogeneous committees may reinforce correlated errors. We introduce Alignment by Confident Debate (AbCD), which converts debate traces into preference data without gold answer labels. A sampling-diverse committee debates each problem, after which final answers are clustered by task-level semantic equivalence. We use verifiable tasks (mathematical reasoning, multiple-choice quantitative QA, and executable code-generation) as a controlled test bed: their answer keys or executable checkers let us evaluate the quality of debate-derived supervision against ground truth, even though the oracle itself never uses gold labels. Post-training on the resulting preferences outperforms inference-only ensembling and debate, as well as SFT on the same traces, across five model families spanning 1B to 8B parameters and four datasets, improving consensus accuracy, with a maximum gain of 21.5_pp . Ablations show that the approach transfers to a larger model and to code generation, while also revealing limitations of heterogeneous committees and agreement-based weighting. Overall, the results suggest that debate traces can yield useful preference supervision when answer equivalence is combined with confidence-weighted disagreement.
[AI-24] CoMoFlow: Controllable Protein Conformational Ensemble Generation via Explicit Collective-Mode Modeling
链接: https://arxiv.org/abs/2611.09548
作者: Yusong Wang,Zheyuan Gu,Ailiang Lin,Minghao Shao,Yicheng Xu,Jiang Cai,Prayag Tiwari,Mingkun Xu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Protein conformational ensemble generation aims to efficiently characterize the structural variations accessible to a protein beyond a single static structure. Recent flow-based methods efficiently generate conformational ensembles, but they model coordinated structural variation implicitly, which can limit ensemble fidelity and hinder direct control. To address this limitation, we propose CoMoFlow, a controllable framework for generating protein backbone conformational ensembles by explicitly modeling collective variation. CoMoFlow constructs protein-specific collective modes from the equilibrium structure, learns the distribution of their coefficients across the molecular dynamics (MD) ensemble, and uses sampled collective variations to guide SE(3) flow matching through feature conditioning and prior recentering. Experiments on ATLAS and de novo proteins show that CoMoFlow improves reconstruction of MD conformational distributions and coordinated structural variation relative to strong flow-based baselines, while enabling direct control over targeted conformational changes.
[AI-25] CID: Measuring Cognitive Work Through Semantic State Change
链接: https://arxiv.org/abs/2611.09542
作者: Ron Fulbright
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 1 figure, 12 equations
Abstract:Traditional information theory provides powerful tools for analyzing statistical properties of information but offers limited insight into meaning and semantic transformation. It has never been possible to quantify the meaning of an arbitrary piece of information. Large language models represent information as vectors embedded within high-dimensional semantic spaces, enabling meaning to be analyzed geometrically. This paper defines semantic state as the vector representation of information within a model’s semantic space and defines cognitive work as the transformation between semantic states, movement through semantic space. To quantify this transformation, we introduce the Contextual Information Differential (CID), a covariance-aware metric based on Mahalanobis distance measuring semantic displacement relative to the structure of the embedding manifold. By treating meaning as a geometric property of semantic space, CID provides a practical method for measuring semantic change, cognitive transformation, and information processing. This framework establishes a foundation for analyzing cognitive processing, learning, reasoning, knowledge generation, synthetic expertise, and human/AI collaboration through implementation-independent measurement of semantic state change in both human and artificial systems.
[AI-26] PreForceVision: Physics-Guided Optical Estimation of Bolt Preload from Bolt-Head Deformation
链接: https://arxiv.org/abs/2611.09529
作者: Nehal Afifi,Klaus Rappenecker,Florian Peinhardt,Gerhard Kachel,Alfred Isele,Sven Matthiesen
类目: Artificial Intelligence (cs.AI)
备注: 9 pages, Proceedings of the ASME 2026 International Mechanical Engineering Congress and Exposition (IMECE 2026), November 8-12, 2026, Vancouver, Canada. Accepted and under publication
Abstract:Bolt preload is a decisive functional quantity in bolted joints because it determines whether external loads are transmitted primarily through frictional clamping or directly through the bolt. Existing preload measurement techniques, including strain-gauge bolts, force washers, ultrasonic methods, and bolt-length measurements, are effective in certain applications but require additional hardware, coupling, calibration, or joint modifications. This limits their scalability for large numbers of standard fasteners and motivates a non-contact optical approach. This paper proposes PreForceVision, a physics-guided optical framework for estimating bolt preload from preload-induced surface deformation of the bolt-head front face. The method treats preload estimation as an inverse deformation-to-force problem, where deformation of the optically accessible bolt-head surface is extracted from reference-loaded image pairs and related to the internal preload state. Preliminary results for an M16 strength class 8.8 bolt show an approximately \SI12\micro\meter simulated head-face depression at \SI90\kilo\newton, strain-gauge measurements of comparable magnitude, and an optical 3D measured central depression of approximately \SIrange1020\micro\meter. The contribution is a physics-guided optical methodology combining controlled image-pair acquisition, DIC- or optical-flow-based deformation extraction, multimodal validation, and uncertainty-aware regression for rough preload estimation under controlled conditions.
[AI-27] StockASkill: A Local-First AgentS kill for Intelligent Stock Selection
链接: https://arxiv.org/abs/2611.09528
作者: Xiangjing An,Zhibin Jiao,Aiting Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Quantitative equity selection and portfolio construction in emerging and cross-border markets frequently suffer from computational bottlenecks, remote API fragility, survivorship bias, and rigid execution constraints such as discrete board lots and investor-mandated drawdown limits. In this paper, we present StockASkill, a local-first, full-data-driven AgentSkill engineered as a specialized Quant Co-Processor for autonomous AI agents and quantitative researchers. Tailored for agile portfolios (10^5 ~ 10^7 RMB) where discrete lot execution and capital protection are mandatory, StockASkill embeds a columnar DuckDB OLAP engine over 42.44 million historical price records and 186,432 financial disclosures spanning 26 years across China A-shares, Hong Kong, US equities, and ETFs. By eliminating ad-hoc network dependencies during analysis, StockASkill reduces cross-sectional screening latency across 5,400+ equities from over 25 seconds to sub-0.5 seconds (a 95%+ reduction). We formalize an end-to-end mathematical pipeline integrating a 7+1 factor alpha model, an SLSQP convex optimizer under smooth CVaR drawdown surrogates with exact path projection, dynamic 100-share board lot integer quantization, and an autonomous thesis lifecycle finite-state machine. Furthermore, we provide a zero-disk, strongly-typed JSON streaming interface designed for high-throughput LLM tool-calling. Open-sourced at this https URL, comprehensive empirical benchmarks from 2018 to 2024 demonstrate that the architecture achieves annualized Sharpe ratios exceeding 1.48 and Calmar ratios up to 2.17, while reducing max drawdown by up to 85% relative to benchmark indices.
[AI-28] Atik: A Shared-Mesh Transformer Accelerator for RISC-V
链接: https://arxiv.org/abs/2611.09522
作者: Ahmed Zeer,Erkan Uslu
类目: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
备注:
Abstract:Transformer inference repeatedly invokes dense matrix multiplication and attention, including its data-dependent softmax normalization. We present Atik, an open-source RoCC accelerator for RocketChip that executes matrix multiplication and dense attention through one shared fixed-point MAC mesh. Atik accepts and returns BF16 tensors, uses a streaming online-softmax schedule that never materializes the full score matrix, and is invoked through a small descriptor-based ABI. We evaluate three meshes cycle-by-cycle on FireSim and report standalone Sky130 physical-design area. Across 137 logged FireSim test records, all tests pass their benchmark tolerance with zero reported output mismatches. Relative to the same scalar Rocket BF16 reference, the 8x8 configuration reaches 31.5x peak matmul speedup and 225.8x peak attention speedup. Hybrid projections for the single-layer workload-derived C models, using directly measured accelerator-kernel cycles, give aggregate speedups of 44.2x, 77.1x, and 32.3x for ViT, TinyBERT, and GPT-2 prefill, respectively; these are not full-model hardware replays.
[AI-29] Active Learning of Robust Models
链接: https://arxiv.org/abs/2611.09520
作者: Mizu Nishikawa-Toomey,Andreas Kirsch,Laurent Charlin,Dhanya Sridhar
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:In many domains from species identification to cancer diagnosis, labelling data to train predictive models remains costly. In these settings, it is advantageous to intelligently select data points to label to achieve the maximum information gain, this is the motivation behind the study of active learning. In addition to the high costs of acquiring data in such domains, there is the added complexity of data distribution shifts that can occur between training and deployment. For example, the features of cancerous skin cells taken in one country may be different in another region of the world. These two challenges - efficient labelling of data, and learning models robust to distribution shifts have been studied separately but not together, even though these challenges often co-occur in real-life settings. We identify failure modes of standard active learning approaches when the distribution of data at deployment is different from the distribution in which the data is sampled for labelling. We propose a new strategy for efficiently learning classifiers that generalise under specific forms of distribution shift, namely subpopulation shift. Drawing on work in learning from heterogeneous data, we aim to train models that perform well across subpopulations of the data, through active learning. To handle this challenging setting, we exploit the fact that many datasets include metadata about data collection that can serve as a noisy proxy for each sample’s underlying subpopulation. We demonstrate how to leverage these noisy labels during active learning selection, and demonstrate empirically that our approach improves classifier accuracy across subpopulations.
[AI-30] A Physics-Informed Learning Framework for Linking Microscopic Surface Degradation to Bolted Joint Functional Behavior
链接: https://arxiv.org/abs/2611.09514
作者: Nehal Afifi,Gerhard Kachel,Sven Matthiesen
类目: Artificial Intelligence (cs.AI)
备注: 10, Proceedings of the ASME 2026 International Mechanical Engineering Congress and Exposition (IMECE 2026), November 8-12, 2026, Vancouver, Canada
Abstract:Returned products in circular engineering often have heterogeneous degradation histories, making remaining function difficult to assess. For bolted joints, corrosion alters surface morphology, contact, friction, preload retention, and mechanical response. This paper presents a physics-informed multimodal framework for linking microscopic and camera-based degradation observations to functional indicators of corroded bolted joints. The workflow defines the corrosion failure mode, generates synchronized inspection and functional-response records, preprocesses image, roughness, exposure, and joint descriptors, aligns microscopy and camera observations through similarity learning, and constructs a calibrated corrosion-severity field. Partial differential equation (PDE) discovery is applied to this field to identify candidate effective mesoscale laws rather than complete electrochemical mechanisms. The corrosion-evolution model is coupled to a separate functional model, with friction coefficient as the primary target and preload loss included when clamp-force data are available. The contribution is a reproducible pathway for physically constrained assessment of degraded bolted-joint behavior.
[AI-31] -means: outperforming k-means (in one-tenth the epochs) KR
链接: https://arxiv.org/abs/2611.09501
作者: Swann Bessa,Pierre-Emmanuel Mazaré,Gergely Szilvasy,Hugo Touvron,David Picard,Matthijs Douze,Hervé Jégou
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Code : this https URL Website : this https URL
Abstract:Lloyd’s k-means algorithm is a long-running standard for clustering, with applications in data analysis, retrieval, quantization and data curation. In this paper, we introduce an exponential moving average optimization procedure for k-means with mini-batch updates, powered by modern deep learning machinery and new insights. We also establish scaling laws on how decay rate and batch size impact final performance. This gives us a formulaic rule to fix them for multiple problem sizes and embedding datasets, without the need for any hyperparameter sweep. The resulting clustering algorithm, coined e-means, is much faster than Lloyd k-means and converges to lower losses, frequently even beating k-means++. On a wide array of embeddings datasets, it beats the loss of Lloyd k-means at 25 epochs in 1.4 to 2.6 epochs, which translates to 10-45x GPU wall-clock speedups compared to state-of-the-art k-means implementations. As an example, e-means partitions DINO10B, a dataset of 10 billion DINO embeddings, into 16 million clusters in only 2.5 epochs, reaching a loss below Lloyd’s at 25 epochs.
[AI-32] Mosaic:montage-agnostic self-distilled graph learning for subject-independent ieeg seizure detection
链接: https://arxiv.org/abs/2611.09496
作者: Arya Jalali,Mahsa Shoaran
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Intracranial EEG provides high signal-to-noise ratio and spatial resolution for detecting epileptic activity, but patient-specific electrode placement leads to substantial variation in channel count and montage across subjects. This heterogeneity limits the generalization of existing detectors under subject-independent evaluation. We propose MOSAIC (MOntage-agnostic Self-distilled Adaptive Inter-channel Connectivity learning), a lightweight graph-learning framework for subject-independent iEEG seizure detection. MOSAIC pairs a shared convolutional channel encoder with a patient-agnostic channel-attention module that infers a weighted inter-channel adjacency matrix from the encoded features, requiring no fixed montage graph. An exponential moving average self-distillation objective regularizes each prediction toward a slowly updated per-patient teacher. Across four public iEEG corpora (SWEC-ETHZ long- and short-term, Mayo, FNUSA), MOSAIC outperforms state-of-the-art baselines in F2 and Cohen’s \kappa while learning more stable inter-channel adjacencies, demonstrating that self-distilled connectivity learning transfers across unseen implantations.
[AI-33] LEMMO: A Large Electromagnetic Model leverag ing Pretrained LLM s for Million-Sample Signal Understanding
链接: https://arxiv.org/abs/2611.09476
作者: Mingkai Li,Jiajun Wu,Yihan Mao,Zilin Sang,Taosong Yu,Yuan Ma,Minghe Liu,Qianqian Yang,Zhiguo Shi
类目: Artificial Intelligence (cs.AI); Sound (cs.SD)
备注:
Abstract:Large language models have expanded beyond text to vision, audio, and other modalities, but understanding raw electromagnetic (EM) signals remains challenging. Wideband recordings can contain millions of samples, whereas existing waveform-based EM models are primarily evaluated on short in-phase and quadrature (IQ) segments, leaving long-recording understanding and detailed description insufficiently explored. We introduce LEMMO, a large EM model for instruction-driven recognition and description of long raw IQ recordings. LEMMO combines a linear-attention EM encoder with anti-aliased temporal resampling to process up to eight million complex samples. Its EM-text interaction module conditions learnable queries on instructions before retrieving signal evidence. Compact queries support recognition, while structured queries with auxiliary factual supervision organize scene-level properties, signal attributes, and cross-signal relationships for detailed description. We also construct and release a unified 1.38-TB corpus of real-world and synthetic EM signals, including LEMMO-ShortQA and LEMMO-LongDesc. Compared with MERLIN-LoRA, LEMMO improves mean EM-Bench perception accuracy by 15.77 %, mean strategy-generation ROUGE-L F1 by 4.38 points, and mean LEMMO-ShortQA accuracy by 14.31 %. It also improves categorical and numerical fact F1 on both LEMMO-LongDesc tasks while preserving its language backbone’s MMLU performance. Our code and datasets are available at this https URL and this https URL, respectively.
[AI-34] Definitions Label Names or Demonstrations? Auditing What Governs LLM Annotation
链接: https://arxiv.org/abs/2611.09457
作者: Viraj Mishra,Shubham Gupta,Dhruv Gupta,Yuehong Cassandra Tai,Qunhua Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are increasingly used to annotate social-science texts, but high accuracy does not establish that they apply researcher-supplied category definitions: predictions may instead reflect prior associations with category labels or labeled demonstrations. We introduce a counterfactual audit that reassigns definitions among output labels, distinguishing adherence to supplied definitions from preservation of a dataset’s original mapping. Across three political-text datasets, models rarely follow reassigned definitions, and standard fine-tuning strengthens the original mapping without reliably increasing definition adherence. We then study retrieval-augmented annotation. Retrieving the nearest training examples generally outperforms retrieving examples separately within each category despite using fewer demonstrations. On the two tasks with substantial gains, direct (k)-nearest-neighbor classification recovers 74–93% of the improvement, indicating that neighborhood labels provide much of the predictive signal. Relabeling fixed demonstrations redirects predictions, whereas aligning demonstrations with reassigned definitions substantially increases adherence to the new mapping. Thus, accuracy-improving interventions need not increase codebook adherence; definitions, category labels, demonstrations, retrieval, and training data should be aligned and validated jointly.
[AI-35] Not All Entropy Is Equal: Reward-Conditioned Entropy Allocation in Reinforcement Learning with Verifiable Rewards
链接: https://arxiv.org/abs/2611.09448
作者: Song Yu,Li Li,Zhenming Wu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 8 pages, 6 figures
Abstract:Policy entropy is widely used to diagnose exploration and collapse in reinforcement learning with verifiable rewards (RLVR), yet its magnitude does not consistently track reasoning performance across optimization objectives. We argue that this ambiguity arises because global token entropy conflates uncertainty in successful and failed trajectories, diversity across correct solution strategies, and variation among realizations of the same strategy. We introduce \textbfReward-Conditioned Entropy Allocation (RCEA), a hierarchical diagnostic framework that separates complete-trajectory entropy from average token entropy, conditions uncertainty on verifier outcome and process validity, measures correct-solution-mode retention, and connects these population-level quantities to local covariance-driven updates. In controlled sweeps over six RLVR objectives at two model scales, entropy trajectories do not induce a stable ordering of reasoning accuracy. Token entropy is higher for verifier-failed than successful trajectories under both broad separately aggregated estimates and a common-question paired analysis, whose confidence intervals exclude equality at 10 of 12 displayed endpoints. A local covariance approximation predicts the direction of adjacent-checkpoint entropy change in 95.5% of nonfailed transitions; a reward-stratified endpoint proxy further shows that positive chosen-token covariance is more tail-concentrated within failed prefixes. Every reportable trained endpoint exhibits some contraction in independently audited correct-solution-mode breadth, whose positive association with large-(k) coverage remains exploratory. These results show that entropy collapse should be diagnosed by which uncertainty is removed, rather than by global entropy alone.
[AI-36] Logit Laundering: Evading Data-Use Auditing in Large Language Models CCS2026
链接: https://arxiv.org/abs/2611.09446
作者: Ruihan Hu,Wei Luo,Yu-Ming Shang,Jiakai Wang,Chuxuan Zhang,Mei Li,Xi Zhang,Meikang Qiu
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted to ACM CCS 2026
Abstract:Data-use auditing in Large Language Models (LLMs) aims to determine whether the given data has been used for training, thereby supporting data ownership safeguards. A dominant solution is data watermarking, which embeds traceable signals into training data and detects them in model outputs. Recent studies have proposed many watermarking-based auditing methods in LLMs. However, systematic investigation from an adversarial perspective remains limited: non-compliant LLM providers would use unauthorized copyrighted datasets for training their own models to enhance model capabilities, while attempting to evade detection by auditors, thereby avoiding legal liability. To fill this research gap, we propose a novel evasion framework, termed Logit Laundering, where the non-compliant LLM provider manipulates model outputs at inference time to undermine detectable watermark signals. Specifically, we design a logit reshaping module to suppress watermark-induced statistical signals by integrating a statistical bias estimator as a plug-and-play component. This design keeps the model parameters unchanged and performs manipulation only in the logit space, thereby enabling auditing evasion while minimizing the impact on model’s learned knowledge. To further balance evasion effectiveness and model utility, we introduce a confidence-aware scaling module to adaptively calibrate the suppression strength based on next-token prediction confidence. Extensive evaluations show that Logit Laundering can successfully evade existing various state-of-the-art auditing methods, significantly attenuating detectable signals, while preserving model utility on downstream tasks. Our work highlights the need for exploring more robust auditing methods to guard against evasion behaviors.
[AI-37] DexDICE: Reconciling Safety and Performance via Dual-Expert Offline Imitation Learning NEURIPS2026
链接: https://arxiv.org/abs/2611.09433
作者: Seokin Seo,Sung Kuk Shyn,Minji Seo,Yisoo Lee,Jongmin Lee,Hongseok Yang,Kee-Eung Kim
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: NeurIPS 2026. 55 pages, 13 figures. Code: this https URL
Abstract:Ensuring high returns while providing strong safety guarantees in offline imitation learning (IL) is a fundamental challenge: existing approaches either rely on explicitly specified cost functions, which are difficult to design for complex real-world safety constraints, or require expert demonstrations that are both safe and high-performing, which are often too costly to collect. We introduce dual-expert imitation learning, which leverages two complementary data sources: (i) safety-compliant demonstrations that faithfully follow safety guidelines but may be suboptimal, and (ii) performance-oriented demonstrations that achieve high returns but may violate safety guidelines. To address this problem, we propose DexDICE (Dual-expert offline imitation learning via stationary DIstribution Correction Estimation). Instead of conventional distribution constraints, DexDICE combines DICE-style offline learning with safe support constraints, enabling agents to exploit high-return behaviors while remaining within safety-compliant regions. We prove that the ideal solution satisfying this constraint exactly has zero unsafe-region visitation with high probability. We evaluate DexDICE on a re-curated DSRL benchmark and validate it on a real-world mobile-robot navigation task trained directly from human teleoperation demonstrations, where it substantially outperforms baselines and remains robust to data scarcity and quality degradation.
[AI-38] Agent ic AI for Healthcare Claim Denial Analysis: A Policy-Grounded Framework for Privacy Preservation
链接: https://arxiv.org/abs/2611.09423
作者: Sofia Campos,Carlos Martins,Pedro Pires,João Gama,Bruno Veloso,Pedro Santos
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:In the healthcare sector, claim denials impose significant financial and administrative burdens, emphasizing the need for automated and trustworthy decision-support systems. This paper proposes a modular Agentic Artificial Intelligence (AI) framework for healthcare claim denial analysis, integrating specialized agents for denial risk prediction, risk factor identification, diagnosis-procedure-medication code relationship analysis, and counterfactual analysis. The framework integrates predictive analytics, graph-based learning, a Hybrid Contextual Retrieval-Augmented Generation (RAG) pipeline grounded in healthcare policy documents, and a dedicated HIPAA Compliance Agent to support privacy protection. Evaluated on a synthetic healthcare claims dataset, the framework achieved 98% traceability of reported numerical findings, validated 214 of 216 policy citations, and removed all 280 identified sensitive identifiers without modifying the underlying analysis. These findings pave the way for further research into Agentic AI as a framework for explainable, policy-aware, and privacy-preserving healthcare claim denial management.
[AI-39] yped-Centrality Grounding for Agent ic GraphRAG Agent ic GraphRAG: A Symbolic Approach for Agentic Reasoning
链接: https://arxiv.org/abs/2611.09414
作者: Rosario Napoli,Antonio Celesti,Massimo Villari,Maria Fazio
类目: Artificial Intelligence (cs.AI)
备注: Preprint. 7 pages. Final version published in 2026 IEEE Symposium on Computers and Communications (ISCC), DOI: https://doi.org/10.1109/ISCC70364.2026.11739974
Abstract:Knowledge Graphs (KGs) and Large Language Models (LLMs) represent two approaches for knowledge representation and reasoning over heterogeneous data. Recent research has explored their integration through Graph Retrieval-Augmented Generation (GraphRAG) systems, which use KGs to provide structured context to LLMs. However, existing GraphRAG approaches treat KGs as homogeneous graphs and rely on textual similarity between the input prompt and the context, rather than using it as a source of deterministic knowledge. This work proposes an agent-based GraphRAG system that introduces a symbolic layer to integrate deterministic reasoning alongside text similarities. This approach is based on the definition of typed centralities for KGs, where centrality measures are computed separately for each specific KG’s class. This information is used to construct a symbolic representation that is provided to the LLM as a deterministic source. The proposed system was mathematically validated and evaluated on a real-world POLE (Person, Object, Location, Event) dataset, comparing reasoning and information retrieval with and without the symbolic layer. The results indicate that the our approach leads to responses that are more consistent with the context, improving the generated outputs. Comments: Preprint. 7 pages. Final version published in 2026 IEEE Symposium on Computers and Communications (ISCC), DOI: https://doi.org/10.1109/ISCC70364.2026.11739974 Subjects: Artificial Intelligence (cs.AI) ACMclasses: I.2.4; I.2.7; H.3.3 Cite as: arXiv:2611.09414 [cs.AI] (or arXiv:2611.09414v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.09414 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Journalreference: 2026 IEEE Symposium on Computers and Communications (ISCC), pp. 1-7, 2026 Related DOI: https://doi.org/10.1109/ISCC70364.2026.11739974 Focus to learn more DOI(s) linking to related resources
[AI-40] MIRA: Mitigating Pole Geometry Bias in Recurrent Language Models
链接: https://arxiv.org/abs/2611.09412
作者: Róisín Luo,Ihsan Ullah,Christian Gagné,Dong Li,Karyn Morrissey
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We show that the Transformer’s recurrent successors — such as selective state-space models (SSMs), Receptance Weighted Key Value (RWKV), and Retentive Network (RetNet) — can be characterized as infinite-impulse-response (IIR) systems. The zero–pole geometries of these IIR systems characterize the temporal modes they can represent. However, to trade parallelism, the algebraic structures of the parameterizations of these recurrent systems often give rise to poles constrained to the real axis, thereby inducing an architectural bias in their recurrent dynamics. To unify these recurrent architectures and eliminate their pole geometry bias, inspired by the lattice–ladder realization of general IIR systems, we present \textscMira (\textbfModal \textbfIIR \textbfRecurrent \textbfArchitecture), a unified family of linear-time recurrent architectures for building large language models (LLMs) with a highly parallel algebraic structure. Each \textscMira block realizes a higher-order IIR system that represents temporal modes from token streams through a lattice–ladder structure, followed by a feed-forward network (FFN). This realization enables the model to represent the temporal modes associated with both real and complex-conjugate poles. We pretrain \textscMira-S, a 12-layer small-scale LLM with 0.14B parameters composed entirely of \textscMira blocks. Throughout the pretraining, \textscMira-S achieves lower perplexity than strong recurrent baselines, exhibits stronger context-window generalization beyond the training length, and attains higher throughput with a dedicated CUDA kernel.
[AI-41] DataHelix: A CPU-GPU Co-Scheduling Data Plane for Multimodal Model Training
链接: https://arxiv.org/abs/2611.09411
作者: Yuanhe Li,Jiazhen Wu,Dian Ding,Guangtao Xue,Yiming Zhang
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注:
Abstract:Multimodal model training breaks the loose coupling between data loading and model execution assumed by conventional training pipelines. Large per-sample variation in decoding cost, token/latent count, and tensor size causes both CPU preprocessing time and GPU compute time to vary substantially across data-parallel ranks (training processes). In synchronous training, these two sources of heterogeneity interact tightly: GPU execution determines how long the input pipeline has to prepare the next batch, while delayed preprocessing in any rank in turn stalls all GPUs at synchronization. As a result, optimizing CPU throughput or GPU load balance in isolation cannot eliminate pipeline bubbles. We present DataHelix, a metadata-first CPU-GPU co-scheduling data plane for multimodal training. DataHelix jointly reasons about preprocessing demand and GPU execution progress, treating each GPU compute window as the readiness budget for the next batch and coordinating preprocessing across ranks to minimize deadline misses rather than CPU makespan. To avoid expensive data movement, placement is decided while samples remain lightweight descriptors, before decoding and collation amplify them into large tensors. Descriptor-level token-latent-aware batching performs coarse pre-balancing, while cost-gated collective rebalancing corrects residual skew only when its expected benefit exceeds transfer cost. Across 11 video and multimodal workloads on 8 Ascend 910B accelerators, DataHelix reduces mean deadline-miss rate by 67.8% and mean p95 input-wait latency by 87.0%, while achieving 2.72x geometric-mean throughput over rank-local DataLoaders and 1.82x over a Ray actor pool under the same CPU budget.
[AI-42] ROAM: Retrieval-Augmented Optimization for Feasible Multi-Day Itinerary Planning
链接: https://arxiv.org/abs/2611.09407
作者: Ye Zihao,Lin Xiaoheng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) can draft plausible travel itineraries, but the plans they produce are frequently infeasible: attractions are grouped into days without regard for geography, opening hours are ignored, and a single generation pass cannot reliably organize a multi-day trip. We present ROAM, a retrieval-augmented agentic framework for spontaneous multi-day itinerary planning that couples semantic preference retrieval with explicit spatiotemporal optimization. ROAM embeds the natural-language request, retrieves a compact candidate set from a curated point-of-interest (POI) database, partitions candidates into day groups through maximal-clique analysis and K-means, orders each day with a quality-biased simulated-annealing TSP solver, and repairs hard opening-hour violations with greedy exchange. A spatiotemporal retrieval layer injects weather forecasts, real road distances and POI operating hours into the prompt, after which the LLM renders a structured, map-ready itinerary; an interactive ReAct-style replanning agent then supports natural-language edits while re-checking feasibility. Across four Chinese cities and 510 POIs, ROAM produces 27.5 POIs per three-day plan versus 19.3 for a bare LLM and 17.0 for a retrieval-only baseline (+42% and +62%), while reducing route dispersion from 2,101 to 677 (3.1x). Ablations show that spatial optimization is the dominant driver of route efficiency and that hard time-window repair prevents roughly 30% of attractions from violating operating hours.
[AI-43] PhyCAD: Physics-Aware Text-to-CAD Generation via Hierarchical Reinforcement Learning
链接: https://arxiv.org/abs/2611.09401
作者: Yuxin Cui,Weijie Zheng,Zhengyang Li,Han Wang,Wenjie Feng,Qiyu Kang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Language-driven computer-aided design (CAD) generation enables the creation of editable parametric models from text, yet geometric correctness alone does not ensure physical feasibility. We present PhyCAD, a framework for physics-aware Text-to-CAD generation that trains language models to produce executable CadQuery programs under geometric and physical constraints. To support this setting, we construct a Text-to-CAD dataset that links parametric CAD models with geometric specifications, loading conditions, and physical constraints. We further develop an automated CAD–CAE verifier that combines program execution, geometric checks, and finite element analysis to provide feedback for model training. A key learning challenge is to prioritize joint satisfaction of geometric and physical constraints while retaining informative feedback on partial progress. A flat aggregation of heterogeneous rewards can obscure this priority, allowing improvements in individual criteria to compensate for unmet design requirements. We therefore introduce a Hierarchical Balanced Advantage for Group Relative Policy Optimization, grouping candidates into levels according to their verification status and combining between-level progress with objective-wise peer ranking within each level. This formulation preserves cross-level priority in the training signal while providing fine-grained feedback for satisfying constraints and improving feasible designs. Experiments across multiple language models show that PhyCAD consistently improves joint geometric and physical constraint satisfaction. These results demonstrate the value of verifiable engineering feedback for advancing language-driven CAD generation beyond geometric reconstruction toward physics-aware design.
[AI-44] ravel Cost Calibration for Multi-Agent Target Assignment and Pathfinding
链接: https://arxiv.org/abs/2611.09396
作者: Yu Kumagai,Ryohei Banno,Keisuke Okumura
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Target assignment and pathfinding (TAPF) is an appealing extension of classical multi-agent pathfinding (MAPF): each agent has a set of targets, and the planner must both assign a target to each agent and plan collision-free paths for them. A trivial approach is to compute an assignment that minimizes the total single-agent shortest-path distance over agent–target pairs, and then solve the resulting MAPF instance. However, this approach falls short because the shortest-path distance remains a proxy that ignores interference among agents. Assigning targets based on the costs agents actually pay, which depend on how every other agent is assigned and routed, would yield a better final outcome. To this end, we propose an iterative TAPF method that improves target assignments by calibrating travel cost estimates from fast, suboptimal MAPF feedback. Experiments show that our method produces solutions close to optimal on small instances and outperforms existing planners in solution quality on larger ones.
[AI-45] Against All Likelihood
链接: https://arxiv.org/abs/2611.09385
作者: David Lopez-Paz
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Maximum likelihood is the dominant training principle for large language models, yet likelihood is neither necessary nor sufficient for truth. Unlikely events happen every day, and frequently repeated statements can be false. Drawing on a philosophical tradition that ties objectivity to invariance rather than probability (from Nozick’s account of objective facts and Wigner’s principles of invariance to the structural realism of Poincaré and Cassirer), we argue that the truth of a statement should be judged not by its probability under a single distribution, but by its invariance across relevant transformations of the environment. We connect this view to machine learning literature on out-of-distribution generalization and invariant learning, sketch how invariance could be tested at inference time with language models, and discuss implications for continual learning.
[AI-46] Exploring Weight-Space Programming: Capabilities Generalization and Limitations
链接: https://arxiv.org/abs/2611.09382
作者: Junru Hu,Cheng Tang,Wei Li,Hongze Zhu,Chun-Mei Feng,Qi Zhang,Yirong Chen,Ming Hu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Weight-Space Programming (WSP) enables reasoning agents to directly inspect and modify neural network parameters, offering an alternative to gradient-based model adaptation. However, its capabilities, cross-task generalization, and practical limitations remain insufficiently understood. We systematically investigate WSP through 16 editing trajectories across seven benchmarks and two model scales (Qwen3.5-0.8B and Qwen3.5-9B). Our findings reveal three key insights. First, WSP can achieve substantial performance gains through highly localized interventions. On full held-out evaluations, accuracy improves by up to 29.45, 4.42, 36.72, and 17.11 percentage points on MATH-500, GPQA Diamond, IFEval, and SciBench, respectively. Notably, modifying only 1,023 parameters yields a 29.45-point improvement. Second, WSP exhibits promising but uneven cross-task generalization. A Math-edited checkpoint improves GPQA Diamond, IFEval, and SciBench by 3.03, 12.18, and 20.10 points, respectively, without further editing, while leaving SciCode performance unchanged. Third, effective editing remains challenging and inconsistent. Outcomes vary substantially across trajectories, and six of the 16 trajectories retain the original weights when candidate improvements cannot be reliably validated. Together, these results highlight WSP’s potential for highly localized model adaptation while revealing important challenges in intervention discovery, reliability, and cross-task transfer. Homepage: this https URL GitHub: this https URL Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2611.09382 [cs.AI] (or arXiv:2611.09382v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.09382 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-47] KineCouple: Substrate-conditioned residual readout for enzyme kinetic prediction
链接: https://arxiv.org/abs/2611.09377
作者: Chun Kang,Yunbo Zhang,Fanhu Zeng,Jian Liu,Siwei Deng,Lei Sha
类目: Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM)
备注:
Abstract:Accurate prediction of enzyme kinetic parameters can support enzyme engineering and metabolic modeling by complementing experimental characterization. Existing prediction methods often pool residue-level protein features before combining them with substrate features. Such substrate-independent pooling cannot adapt residue weighting to the substrate during aggregation, potentially diluting information relevant to a specific enzyme-substrate pair. Since the contributions of individual residues to catalysis can depend on the substrate, substrate-conditioned aggregation offers a way to emphasize residue features relevant to the prediction. We introduce KineCouple, which uses substrate queries and residue-level keys and values in cross-attention, allowing the substrate to guide residue selection before aggregation. The model preserves the global protein representation and adds a gated correction defined as the difference between the conditioned readout and a uniform readout. ProtT5 and MolT5 remain frozen, and separate models predict kcat and Km on a logarithmic scale. In pooled out-of-fold predictions on the CataPro benchmarks, KineCouple reduces kcat RMSE from 1.333 to 1.330 and increases PCC from 0.492 to 0.496 relative to reproduced CataPro. For Km, RMSE decreases from 0.998 to 0.997 and PCC increases from 0.633 to 0.634. Spearman correlation also improves for both tasks.
[AI-48] Can LLM s Debug Robots by Editing Their Weights?
链接: https://arxiv.org/abs/2611.09376
作者: Jusheng Zhang,Keze Wang,Liang Lin
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:A language model can repair a robot policy’s lexical instruction grounding by editing just 65 token-embedding rows. We bring weight-space programming to robot foundation policies: GPT-6 Astra reads weights and uses a CPU workspace and budgeted simulator feedback, without gradient updates or new demonstrations. Three independent sessions on \pi_0.5 all improve success on 765 held-out paraphrases; the best raises success from 86.3% to 92.0% (48 repairs, four regressions; exact paired p=1.31\times10^-10 ). Among the 335 instructions containing edited tokens, success rises from 74.3% to 87.5%; the other 430 episodes have zero outcome changes. Matched random-direction edits do not help. Full replacement using the editor’s own lexicon explains much of the gain but produces more regressions; its direct difference from the editor is not significant at 0.05. Across 13 further robot sessions, the editor declines 12 candidates; the sole retained edit does not show a significant held-out gain. Separately, benchmark metadata lowers FastWAM success from 70.0% to 48.6%, while the evaluated VLA’s aggregate rate is unchanged. The evidence supports precise lexical repair, not general acquisition of missing robot capabilities.
[AI-49] Mitigating Covariate Shift in Behavioral Cloning via Robust Stationary Distribution Correction NEURIPS2024
链接: https://arxiv.org/abs/2611.09371
作者: Seokin Seo,Byung-Jun Lee,Jongmin Lee,HyeongJoo Hwang,Hongseok Yang,Kee-Eung Kim
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 25 pages, 11 figures, NeurIPS 2024, Code: this https URL
Abstract:We consider offline imitation learning (IL), which aims to train an agent to imitate from the dataset of expert demonstrations without online interaction with the environment. Behavioral Cloning (BC) has been a simple yet effective approach to offline IL, but it is also well-known to be vulnerable to the covariate shift resulting from the mismatch between the state distributions induced by the learned policy and the expert policy. Moreover, as often occurs in practice, when expert datasets are collected from an arbitrary state distribution instead of a stationary one, these shifts become more pronounced, potentially leading to substantial failures in existing IL methods. Specifically, we focus on covariate shift resulting from arbitrary state data distributions, such as biased data collection or incomplete trajectories, rather than shifts induced by changes in dynamics or noisy expert actions. In this paper, to mitigate the effect of the covariate shifts in BC, we propose DrilDICE, which utilizes a distributionally robust BC objective by employing a stationary distribution correction ratio estimation (DICE) to derive a feasible solution. We evaluate the effectiveness of our method through an extensive set of experiments covering diverse covariate shift scenarios. The results demonstrate the efficacy of the proposed approach in improving the robustness against the shifts, outperforming existing offline IL methods in such scenarios.
[AI-50] SafeOrchestrate: Runtime Checked Intent-Driven Orchestration for Cyber-Physical Systems
链接: https://arxiv.org/abs/2611.09364
作者: Muhammad Ashfaq,Ahmed R. Sadik,Ewan Warburton,Pyry Kotilainen,Niko Mäkitalo,Tommi Mikkonen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Modern cyber physical systems (CPS) integrate heterogeneous sensors, actuators, controllers, and services, composed at runtime rather than at design time. Orchestrating such systems through natural language (NL) is attractive, but unconstrained large language model (LLM) tool use is risky because actions must respect device capabilities, runtime state, and safety constraints. We present Safe Orchestrate, a reference architecture for runtime checked, intent driven CPS orchestration. Given an NL intent, it derives atomic goals, generates a runtime checkable specification for each, grounds the specifications to device affordances, and verifies every candidate action against runtime state and safety constraints before execution. In a Web of Things (WoT) instantiation, the architecture produces Linear Temporal Logic (LTL) style specifications from the NL intent, grounds them to Thing Description (TD) affordances exposed through the Model Context Protocol (MCP), and checks candidate actions against satisfiability modulo theories (SMT) backed guardrails. We evaluate this instantiation on simulated smart home, greenhouse, and industrial automation tasks against a direct LLM to MCP baseline. Safe Orchestrate maintains 100% required action accuracy on every task and 100% MCP Execution Similarity (MCP ES) on all but one, where MCP ES is 92%. By contrast, the baseline degrades on multi device, event triggered, and ordered tasks, with MCP ES falling to 13%. These results suggest that combining LLM based specification generation with role aware grounding and runtime constraint checking improves reliability for open CPS orchestration without enriched TD annotations or planning domain expertise.
[AI-51] Challenges Opportunities and Misconceptions in Neurosymbolic AI
链接: https://arxiv.org/abs/2611.09356
作者: Giuseppe Marra(LPI),Pierre Genevès(CNRS, LIG, TYREX),Meghyn Bienvenu(LaBRI),Thomas Schiex(MIAT INRAE),Sylvie Thiébaux(Comue de Toulouse, ANU),Silviu Maniu(SLIDE, LIG)
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We examine opportunities for neurosymbolic AI alongside increasingly powerful neural models, particularly with regards to efficiency, safety and interpretability. We identify two major challenges: ensuring semantic alignment between neural and symbolic components, and managing the computational cost of symbolic inference and learning. We discuss representative methodological approaches to these challenges and address several common misconceptions concerning differentiability, independence assumptions, and scalability. This paper synthesizes perspectives developed through discussions in the Neurosymbolic Programming and Reasoning (NPR) working group at the 2026 CNRS workshop Neurosymbolic Methods and Their Applications in Data Science, rather than providing an exhaustive survey.
[AI-52] AUTEUR: Acoustic Plan Learning for Red-Teaming Text-to-Audio Models
链接: https://arxiv.org/abs/2611.09351
作者: Jiashu Zhang,Jiaming He,Zheming Zhang,Kaizheng Wang,Bing Li,Zhiyuan Zha,Yi Yu
类目: Artificial Intelligence (cs.AI); Sound (cs.SD)
备注:
Abstract:Automated red-teaming of text-to-audio models requires test prompts that pass textual screening and expose harmful acoustic content. The central challenge is to connect the sounds produced by the target model to both reusable test generation and prompt revision. Our key idea is to represent test descriptions as structured acoustic plans that organize foreground sounds, context, and vocal attributes. To this end, we propose acoustic plan learning, a framework that learns and refines these plans using waveform feedback. We train the plan generator through supervised fine-tuning followed by reinforcement learning with audio feedback while keeping the target audio model fixed. At inference time, the same trained model generates initial plans and uses feedback to revise their context and vocal attributes while preserving the foreground. Experiments across four target systems show attack success rates of 79.4% on TangoFlux and 95.6% on Kling. The resulting test prompts also transfer across open-source and closed-source models, indicating shared acoustic vulnerabilities.
[AI-53] When Feedback Becomes a Backdoor: Tracing Backdoor Skill Formation in Text-Space Skill Optimization NEURIPS2026
链接: https://arxiv.org/abs/2611.09348
作者: Louise Nayeong Koh,Yaechan Lee,Myeongjong Lee,Doyeon Kim,Juyoung Park,Seunggyu Kim,Inseo Kim,Chaerin Jung,Hyogon Ryu
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted as a poster at the Trustworthy AI for Good (AI4GOOD) Workshop at NeurIPS 2026. OpenReview: this https URL
Abstract:Reference answers are a supervision channel in text-space Skill optimization, which produces reusable text artifacts called Skills that guide future agent behavior. Poisoning those answers can induce a persistent trigger-conditioned rule without giving the attacker access to the model, optimizer, or Skill file. We study this reference-feedback poisoning in SkillOpt by tracing an attacker-supplied trigger–payload mapping from reflection to held-out activation. In SearchQA, where the payload replaces a short answer, the natural attack retains an exact rule in all three seeds; each final Skill emits the payload on 200/200 triggered questions and 0/200 clean questions. In seed 42, removing the rule eliminates activation, and an in-run comparison shows similar clean accuracy to benign SkillOpt (169/200 versus 171/200). In SpreadsheetBench, where the payload requires an additional workbook edit, no natural run retains an exact rule across three seeds. Diagnostic interventions locate two successive bottlenecks: the Analyst abstracts away required literals, and, once literal preservation is enabled, edit-budget ranking can remove exact edits before candidate construction. Poisoned references can therefore create persistent backdoors, but exposure alone does not ensure that the induced rule is formed, selected, and retained.
[AI-54] Reliable Prerequisite-Aware Learning Path Recommendation under Uncertain Knowledge Structures: Long-Horizon Evaluation and Cross-Dataset Validation
链接: https://arxiv.org/abs/2611.09345
作者: M. M. Golam Hafiz,A.K.M. Shahriar Emtiaz,Safia Akter Simi,Md. Rakibul Islam,M. F. Mridha
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 17 pages, 5 figures, 7 tables. Under review at Computers Education: Artificial Intelligence (Elsevier)
Abstract:Learning-path recommendation often depends on knowledge and prerequisite structures whose reliability is uncertain, while high accuracy in predicting the next resource does not necessarily establish reliable long-horizon path behaviour. This study develops a reliability-oriented prerequisite-aware framework that jointly examines immediate next-resource ranking, uncertain prerequisite structures, hard, soft and hybrid constraint semantics, graph robustness, and autoregressive path generation under course-level confirmatory evaluation and independent cross-dataset validation. Development used 3,187 MOOCCubeX courses, followed by a frozen confirmatory evaluation on 541 held-out courses and 37,747 transitions. The primary MOOCCubeX task is structural and course-prefix-conditioned rather than learner-personalised. Junyi provides independent structure-/hierarchy-aware external validation, uses hierarchy metadata rather than an explicit prerequisite graph and is not a zero-shot transfer setting. On the frozen MOOCCubeX evaluation, XGBoost pairwise was the strongest observed immediate ranker, with mean reciprocal rank (MRR) of 0.9048, whereas the proposed prerequisite-aware graph model achieved 0.7707. In the controlled graph-variant development ablation, the no-graph condition produced the strongest immediate ranking result, and immediate prediction showed strong dependence on permitted title/path-similarity information. Autoregressive evaluation exposed path properties not captured by one-step ranking. On Junyi, Markov achieved the strongest observed MRR of 0.4453, compared with 0.3954 for the proposed graph model, with metric-specific differences across ranking measures. Overall, immediate predictive accuracy, prerequisite-structure reliability, and long-horizon path quality should be treated as complementary rather than interchangeable evaluation dimensions in educational recommendation.
[AI-55] Robust Deep Prototype Classifiers via Lipschitz Training
链接: https://arxiv.org/abs/2611.09327
作者: Adia Khalid Khan,Lukas Gehring,Benjamin Paaßen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Robustness and interpretability are interrelated and desirable properties of machine learning models. A model that changes its prediction under imperceptible perturbations behaves inconsistently with human-recognizable features and, hence, is not interpretable. Prototype-based classifiers such as Generalized Learning Vector Quantization address both at once. They are interpretable by construction because they make the decision by comparing the input to the prototypical example; and their hypothesis margin yields a closed form robustness certificate. However, on high-dimensional data, these shallow models need deep embedding networks to achieve high accuracy, and this voids the robustness certificate. The margin is a statement about distances in the embedding space, and an unconstrained network can map an arbitrarily small input perturbation to arbitrarily large attacks in the embedding space. We close this gap with a theorem: if the embedding network is locally L-Lipschitz around a point, and the classifier is \epsilon-robust around its embedding, then the full model is \epsilon/L-robust in the input space. Based on this, we propose a Lipschitz training scheme to achieve L-Lipschitz properties for embedding networks. We evaluate three classification heads (linear, linear Support Vector Machine, and GLVQ) crossed with two embedding networks (standard and Lipschitz-constrained) on MNIST, CIFAR-10, and Imagenette-160, under FGS, PGD and C\W attacks. The evaluation results show that combining a prototype-based head with a Lipschitz-constrained embedding provides the strongest robustness. This suggests that the Lipschitz constraint provides an additional robustness benefit in addition to the advantage provided by the prototype-based head. Since the Lipschitz constraint is enforced by sampling rather than by construction, the reported robustness is empirical rather than certified.
[AI-56] owards an Autonomous AI Steward for Large-Scale Physics Experiments: Integrating Causal Diagnostics RAG and Conformal Prediction
链接: https://arxiv.org/abs/2611.09298
作者: Oliver Noll,Nicholas Tan Jerome,Alexander Lehnen,Frank Simon,Frank Maas
类目: Artificial Intelligence (cs.AI)
备注: 4 pages, 1 figure, 2nd International Conference on AI for Science (AI4Sci2026), 5-9 October 2026, Mainz, Germany (PoS)
Abstract:Long-term particle physics experiments face critical operational challenges: monitoring non-stationary machine data streams and preserving tacit expert knowledge. We present the “Helmholtz Co-Pilot,” a hybrid Multi-Agent System (MAS) acting as an autonomous experiment steward. The system can be triggered interactively by human operators or autonomously by an asynchronous causal diagnostics engine that runs independently of the MAS reasoning core. Continuously monitoring live hardware streams such as data acquisition and slow control, this engine performs dynamic causal discovery to trace cascading alarms back to their root causes and applies adaptive model selection to keep forecasts reliable across shifting machine states; upon detecting an anomaly, it hands a structured diagnostic report to the MAS. The reasoning layer itself relies on a high-throughput C++ core and a Python edge architecture orchestrating a directed acyclic graph (DAG) of specialized ReAct workers. It integrates a Retrieval-Augmented Generation (RAG) backbone for institutional memory and leverages conformal prediction to bound its evidence-based recommendations with mathematical coverage guarantees. This extended abstract outlines the architecture, the integration of software-enforced Human-in-the-Loop (HitL) guardrails, and ongoing shadow-testing within the KATRIN experiment.
[AI-57] AutoRefine: Auto-Calibrating Code Judges from Workflow Outcomes
链接: https://arxiv.org/abs/2611.09291
作者: Shervin Shahidi,James Poon,Hang Xie,Kenneth O. Xu,Robin Håkanson,Vikram Nair,Jens Pillgram-Larsen,Nachiappan Nagappan
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 5 pages, 3 figures, prepared for submission to CAIN 2027 industrial track
Abstract:LLM judges increasingly serve as quality gates for code changes that have no test oracle, but a judge prompt must be calibrated for its task and recalibrated as the system it judges evolves, while hand labels for the calibration are resource intensive and go stale. Development workflows emit outcome labels for free (tests pass or fail, changes land or are abandoned), but these labels are noisy. We present AutoRefine, a loop that rewrites a deployable judge’s prompt to reproduce outcome labels, with an informed teacher that sees the judge’s input plus the outcome, explains each label it keeps, and abstains when the outcome cannot be explained from what the judge sees. On a held-out dataset with an execution oracle, calibration raised chance-corrected agreement (Scott’s \pi ) from 0.36 to 0.60. On a deployed agent’s outcome labels, with 25% flipped at random to introduce noise, the teacher set aside flipped \textscfail labels at more than twice the rate of genuine ones, cutting overall label noise from 25% to 20%. Recalibrating that agent’s judge using AutoRefine raised its agreement with merge/abandon outcomes on 17 independently held-out diffs from the current workflow ( \pi 0.16 to 0.51); on 16 diffs from older workflows, the previous judge stayed ahead, illustrating the need for recalibration as the agent evolves. By abstaining, the teacher not only removes noisy labels but also flags changes whose outcome the judge’s input cannot explain, telling a team when a judge needs better inputs rather than a better prompt.
[AI-58] Ontology-Driven Control of Conversational Agents with a Use-Case for Job Interview Practice
链接: https://arxiv.org/abs/2611.09282
作者: Barbara Gendron(LORIA, UL),Gaël Guibon(LIPN, LORIA),Mathieu D’aquin(LORIA, UL)
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) have been widely used as conversational agents in both domain-specific and open-domain situations. In domain-specific context, it is often necessary to adapt those LLMs to obtain conversations with desired characteristics, whether those apply at the level of each response (e.g. for tone or style) or more globally (i.e. at the level of a conversational strategy). Such adaptations are, however, generally realized in an ad hoc way, through specific, non-reusable, and non-transparent developments. In this paper, we propose an ontology-based approach to design conversational strategies that rely on the desired and measured characteristics of textual utterances in the conversation. Through a use case related to job interview practice, we show how this approach can be used to control LLMs to follow the strategy defined in the ontology. To this end, we automatically fine-tune the LLM to improve its compliance with ontologically-defined conversation characteristics. Our results show that training the LLM from formally defined concepts effectively enhances alignment with the requested characteristics. Hence, this work demonstrates how ontology-guided conversation control can be achieved while taking into account multiple, concurrent conversation aspects that are formally and transparently defined.
[AI-59] An Empirical Study of Issues in GPU Kernel Libraries for LLM Systems
链接: https://arxiv.org/abs/2611.09280
作者: Yanjie Gao,Jiayi Bi,Haoxiang Lin,Ling Feng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Behind the expanding use of large language models (LLMs) across application domains lies a growing dependence on specialized GPU kernel libraries. These libraries underpin the correctness and efficiency of training and inference across diverse GPU platforms, implementing attention, GEMM, quantization, and low-precision operators while managing tensor shapes, memory layouts, code generation, and hardware-specific dispatch. Problems at this layer can silently corrupt outputs, hang GPUs, prevent compilation, introduce deployment incompatibilities, or degrade performance, thereby reducing developer productivity and wasting expensive GPU resources. We present the first comprehensive empirical study of issues in GPU kernel libraries for LLM systems. We manually analyzed 300 high-quality GitHub issue reports, their discussions, and available fixes from seven representative libraries: FlashInfer, FlashAttention, Triton, CUTLASS, DeepGEMM, xFormers, and bitsandbytes. The analysis characterizes symptoms, root causes, fixes, and testing and debugging practices. Our main findings are: (1) 60.33% of the sampled issues originate in the Compiler and Build System or Runtime Environment dimensions, so validation must cover the deployment and execution stack, not just kernel code. (2) Kernel specialization imposes implicit correctness constraints on tensor shapes, metadata layouts, memory aliasing, and hardware resources. Violations can produce incorrect outputs or launch failures even for supported operations, highlighting the need for explicit contract validation and boundary-case testing. (3) Current testing and debugging practices are often insufficient for identifying low-precision numerical problems and hardware-specific execution failures. We outline research opportunities and tooling directions to support future LLM GPU kernel development. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.09280 [cs.AI] (or arXiv:2611.09280v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.09280 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-60] Diversity-Aware Client Selection for Communication-Efficient Federated Learning
链接: https://arxiv.org/abs/2611.09278
作者: Max Halleryd,Sima Sinaei,Viktor Valadi,Mina Alibeigi,Henrik Abrahamsson,Salman Toor
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 4 pages, 6 figures, 2 tables. Accepted at SNCNW 2026
Abstract:Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data, making it attractive for privacy-sensitive and bandwidth-constrained applications. A practical challenge in deploying FL is reducing communication overhead while maintaining model performance under heterogeneous, non-IID data distributions. Selecting which subset of clients participates in each training round offers a principled way to address both challenges simultaneously, yet its impact on data diversity and representativeness remains underexplored. In this paper, we investigate three client selection strategies, Power-of-Choice, Fisher information, and Centered Kernel Alignment, evaluating them not only by accuracy and convergence, but through the lens of participation diversity using the Gini coefficient and KL divergence. Under non-IID conditions, restricting participation to 15% of clients per round achieves significant communication savings, and when selection promotes diverse and balanced contributions, it also improves generalization over biased alternatives. This work provides a systematic and reproducible analysis demonstrating that diversity-focused client selection renders communication efficiency and model quality complementary rather than competing objectives.
[AI-61] AudioSphere: Towards Self-Supervised Spatial Audio Representation Models NEURIPS2026
链接: https://arxiv.org/abs/2611.09275
作者: Goksenin Yuksel,Marcel van Gerven,Kiki van der Heijden
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)
Abstract:Recent audio embedding models learn representations that transfer across clip- and frame-level tasks, yet they operate almost exclusively on single-channel signals and are blind to spatial structure. Spatial audio models capture that structure but rely on clip-level supervision or contrastive objectives, yielding embeddings unsuitable for frame-level tasks like sound event detection and localization (SELD). We argue this dichotomy is not fundamental: patch-level spatial self-supervision can jointly learn acoustic and spatial features at both temporal granularities. We introduce AudioSphere, a multi-channel masked autoencoder trained on Ambisonics recordings that reconstructs both spectral content and directional cues via a joint spatiotemporal masking strategy that masks both time-frequency patches and intensity vectors. AudioSphere supports frame-level predictions while producing strong clip-level representations via aggregation. Alongside the model, we release RealSELD, a standardized benchmark for evaluating frozen backbones on real-world multi-channel SELD. AudioSphere enables, for the first time, SELD with a frozen self-supervised backbone, while matching or exceeding prior self-supervised audio embedding models on clip-level classification tasks spanning environmental sound, speech, and music, and on frame-level sound event detection. All code and data are available at this https URL and this https URL.
[AI-62] PHISHREV: An ASP-Based Post-Hoc Defeasible Machine Learning Framework for Phishing Detection
链接: https://arxiv.org/abs/2611.09273
作者: Mainak Sen,Kumar Sanka Ray,Amlan Chakrabarti
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Phishing website detection remains a critical cybersecurity challenge. Although machine learning classifiers achieve high accuracy on benchmark datasets, they suffer from a fundamental limitation: once trained, their decision boundaries are fixed and cannot incorporate new contextual evidence without retraining the model. This work presents \textbfPHISHREV+, a hybrid framework that instantiates the notion of \textitpost-hoc defeasible machine learning. This framework does not replace the classifier. It treats the initial classifier predictions as defeasible beliefs and revises them through an Answer Set Programming based non-monotonic reasoning(NMR) layer. The framework operates in two phases. In the machine learning phase, four classifiers (SVM, KNN, DT and RF) are trained on the 87 feature benchmark dataset. In the post-hoc NMR phase, four defeasible rules derived from exploratory data analysis of the training set, and validated through Chi-square and Mann-Whitney U tests, formally revise initial predictions. The proposed framework achieves consistent false positive reduction across all our classifiers SVM (41 \rightarrow 39), KNN (47 \rightarrow 46), DT (62 \rightarrow 61), and RF (52 \rightarrow 51) while maintaining complete false negative parity with zero increase overhead. A comprehensive ablation studies across seven rule configurations confirms that the hierarchical combination of all four defeasible rules is necessary to achieve this optimal balance. Instance level explainability is provided through eXplainable ASP (XASP), which generates formal derivation graphs showing which defeasible rules fired and why the initial belief was retracted, a capability absent from all existing phishing detection approaches.
[AI-63] Zero-Shot Weight-Space Programming Across a Difficulty-Graded Ladder of Two-Player Games
链接: https://arxiv.org/abs/2611.09271
作者: Eduardo C. Garrido-Merchán
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Weight-space programming sets every weight of a feed-forward network by hand, from an explicit specification of the desired computation, with no gradient descent and no training data anywhere in the construction. We ask how far a large language model can carry this practice, zero-shot, across games of deliberately increasing difficulty. We design a four-rung ladder (Nim, tic-tac-toe, two sizes of Connect Four, and a chess endgame) spanning four axes of difficulty, and build a second, independent network at each rung, using a structurally unrelated game at matched difficulty, to separate what a rung’s finding says about the difficulty level from what it says about the one game used to measure it. Ten networks are built and independently verified. Six admit a complete, non-statistical proof of correctness against an independent oracle. The chess-endgame rung is exact where correctness is cheap, move legality and, in one network, an exact one-ply mate detector, and collapses where it is expensive, composing a multi-move mating plan, from 84.1-100% success at the one-ply tactics it was built to get right down to 0.57-9.13% in full adversarial play; its replicate shows this collapse is an order of magnitude less severe under a shorter required plan and exact rather than heuristic safety checks. The standard-sized rung confirms, by holding hand-set evaluation weights fixed and adding only shallow search, that the remaining gap is an architectural ceiling rather than a matter of engineering effort, and this exact collapse reproduces across two structurally unrelated games. We release all ten networks, their independent oracles, construction-cost measurements, and a cold-prompt protocol for a future cross-model comparison
[AI-64] Nudging Decision Models: Do Digital Nudges Steer Fast Generation-Free AI Agents ?
链接: https://arxiv.org/abs/2611.09264
作者: Sascha Kaltenpoth,Haya Halimeh,Kevin Bösch,Oliver Müller
类目: Artificial Intelligence (cs.AI)
备注: Under Review
Abstract:LLM-based agents increasingly automate business and consumer tasks but are slow and expensive. Decision models, a new class that chooses among options without generating text, are far faster and less costly. Developers call them “System One models,” referencing dual process theory. We ask whether such models can be steered by digital nudges online and whether, as the theory predicts, automatic nudges affect them more than reflective ones. We adapted an online shopping experiment, previously run with humans and LLM agents, to three decision models (Jev, Decision-1, d1). Across 600 runs per model, they chose a preselected default in 97.5-99.9% and a product with social influence information in 75-96% of choices, versus 47-49% in the control. All responded more strongly to the automatic nudge, the opposite of what has been reported for reasoning-capable LLM agents. We provide first evidence that non-generative decision models are susceptible to digital nudges in a System 1 pattern, implying agent-type-specific nudge effects.
[AI-65] Model-based learning for DoA estimation and array calibration via continuous sparse recovery
链接: https://arxiv.org/abs/2611.09263
作者: Baptiste Chatelier(IETR, INSA Rennes),José Miguel Mateos-Ramos,Vincent Corlay(MERCE-France),Matthieu Crussière(INSA Rennes, IETR),Christian Häger,Henk Wymeersch,Nir Shlezinger(BGU),Luc Le Magoarou(INSA Rennes, IETR)
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Direction of arrival (DoA) estimation, a classical signal processing problem, has experienced renewed interest with the advent of integrated sensing and communication (ISAC). While decades of research have yielded highly accurate DoA estimators, hardware impairments have frequently been neglected, despite degrading both DoA estimation accuracy and downstream communication performance in ISAC systems. This work introduces a joint array calibration and DoA estimation method based on the simultaneous orthogonal matching pursuit (SOMP) algorithm, following the model-based machine learning paradigm. The proposed method enables the calibration of antenna arrays subject to severe gain-phase and antenna location impairments through projected gradient descent (PGD), with either supervised or unsupervised objectives. Additionally, DoA estimation is enhanced by extending SOMP to continuous dictionaries via PGD refinement steps. Simulation results demonstrate that the proposed method enables efficient hardware impairment compensation, even without labeled samples, while outperforming conventional sparse recovery and subspace-based DoA estimation baselines.
[AI-66] Forecasting cumulative intraoperative hypotension burden: patient-disjoint evaluation of quantile boosting and alert policies
链接: https://arxiv.org/abs/2611.09259
作者: Yunfan Zhao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Background: Binary hypotension forecasts do not describe the depth and duration of future low pressure. Burden forecasts may support graded early warning, but better forecast scores need not yield more useful alerts. Objective: To evaluate conditional quantiles of the next 15-minute mean arterial pressure (MAP) deficit below 65 mmHg and distinguish forecasting performance from episode-warning performance. Methods: Public VitalDB numeric arterial MAP was reduced to 10-second medians without forward filling. Subject-disjoint training, validation and test splits included 70,402, 14,270 and 15,575 labeled windows. We compared climatology, calibrated persistence, last-MAP-only boosting, and direct and residual quantile boosting, with and without additional MAP-history features. Patient-weighted pinball loss and subject-bootstrap intervals evaluated forecasting. Validation selected alert thresholds before a hash-recorded test evaluation. Results: In the test split, the prespecified residual-versus-persistence comparison at the 90th percentile yielded losses of 2.256 and 2.900 mmHg-min, respectively (difference -0.644; 95% interval -0.814 to -0.487). Direct prediction achieved 2.196, and direct prediction with additional MAP features achieved 2.193. Episode-warning results were sensitive to the alert rule and operating threshold. Conclusions: This retrospective benchmark separates burden-forecast skill from warning-policy performance. Findings support further study of direct burden estimation; they do not establish patient benefit, prospective reliability or transportability beyond the source hospital. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.09259 [cs.AI] (or arXiv:2611.09259v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.09259 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Yunfan Zhao [view email] [v1] Fri, 6 Nov 2026 10:17:42 UTC (101 KB) function toggleList(whichLayer,toggleThis) var elem, vis; if( document.getElementById ) // standard elem = document.getElementById( whichLayer ); else if( document.all ) // old msie versions elem = document.all[whichLayer]; else if( document.layers ) // nn4 elem = document.layers[whichLayer]; vis = elem.style; // if the style.display value is blank we try to figure it out here if(vis.display==‘’!=undefined!=undefined) vis.display = (elem.offsetWidth!=0!=0)?‘inline’:‘none’; vis.display = (vis.display==‘’||vis.display==‘inline’)?‘none’:‘inline’; // toggle link inner text status = vis.display; if(vis.display==‘inline’) document.getElementById(‘toggle’).innerHTML = “(collapse list)”; document.getElementById(‘toggle’).title = “Collapse list”; else document.getElementById(‘toggle’).innerHTML = “(”+toggleThis+“)”; document.getElementById(‘toggle’).title = “Show complete list”; Full-text links: Access Paper: View a PDF of the paper titled Forecasting cumulative intraoperative hypotension burden: patient-disjoint evaluation of quantile boosting and alert policies, by Yunfan ZhaoView PDFHTML (experimental)TeX Source view license Ancillary-file links: Ancillary files (details): REPRODUCIBILITY.txt code/python/acquire_metadata.py code/python/alerts.py code/python/checks.py code/python/experiment.py code/python/final_test.py code/python/iteration.py code/python/prepare.py code/python/report.py code/python/secondary_summary.py requirements.lock.txt results/alert_definition.json results/analysis_freeze.json results/data_manifest.json results/experiment_manifest.json results/quantile_crossing.json results/test_alerts.csv results/test_bootstrap.csv results/test_evaluation.complete.json results/test_metrics.csv results/test_secondary_pairs.csv results/validation_ablation_pairs.csv results/validation_alert_grid.csv results/validation_alert_selected.csv results/validation_bootstrap.csv results/validation_metrics.csv results/weight_scale_sensitivity.csv(22 additional files not shown)You must enabled JavaScript to view entire file list. Additional Features Audio Summary Current browse context: cs.AI prev | next new | recent | 2026-11 Change to browse by: cs References Citations SciXGoogle 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)
[AI-67] Zero-Query Transferable Pretend Benign Attacks with Amortized Online Generation for Cooperative Perception
链接: https://arxiv.org/abs/2611.09256
作者: Hongwei Lin,Xun Huang,Chenglu Wen,Cheng Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Cooperative perception enables an ego agent to acquire information from collaborators beyond its own sensors’ reach. Yet, the most beneficial messages are often the hardest to verify from local observations. Pretend Benign (PB) exploits this asymmetry by targeting regions with weak local verifiability and applying multi-objective perturbations, making harmful feature changes appear to be benign assistance. However, PB is victim-dependent: selecting regions and optimizing perturbations require the victim model’s architecture, parameters, and gradients, limiting transfer to unseen victims. A transferable attack must capture properties shared across cooperative detection models rather than those specific to one network. We address this by building a victim-agnostic attack from a heterogeneous ensemble of surrogate models and encoding its capabilities in a reusable dictionary to reduce per-message generation cost. Specifically, Transferable Pretend Benign (TrPB) transfers PB’s region selection criterion and multi-objective perturbation strategy across heterogeneous surrogates to construct an offline attack. We further propose RT-TrPB, which encodes transferable regions and perturbation patterns into a region predictor and reusable perturbation dictionary, amortizing per-message attack generation. Experiments on widely used cooperative perception benchmarks show that RT-TrPB bypasses the consensus-based defenses ROBOSAC and FLD with success rates approaching 100% and generates each perturbation in under 56 ms. To counter PB-style attacks masquerading as benign collaboration, we also propose STTrust, which verifies received messages using spatial and temporal consistency and incorporates identity-level trust. STTrust detects over 90% of PB-style attacks and restores perception performance to near the attack-free level.
[AI-68] MM-DyGraph: A Dataset Benchmark and Model for Multimodal Dynamic Graphs
链接: https://arxiv.org/abs/2611.09240
作者: Zongyuan Wu,Pengjie Wang,Lingshan Chen,Tianhang Wan,Yuan Meng,Xin Wang,Ling Feng,Wenwu Zhu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Multimodal dynamic graphs are pervasive in real-world applications, where nodes and edges are accompanied by multimodal signals, including text and images. Despite their practical importance, benchmarks and methods that explicitly study multimodal learning in dynamic graphs remain scarce, hindering the full understanding and effective utilization of multimodal information for downstream tasks. To address this gap, we introduce MM-DyGraph, to the best of our knowledge, the first benchmark for multimodal dynamic graph learning, consisting of diverse real-world datasets with fine-grained temporal graph evolution and rich multimodal attributes. MM-DyGraph provides three evaluation tasks to comprehensively assess model performance in dynamic settings. We benchmark representative state-of-the-art dynamic graph methods and reveal that naive feature concatenation yields target-level inconsistencies: multimodal inputs benefit certain prediction targets while harming others. This observation indicates that the modality informativeness depends on the specific prediction target over time. Building on these observations, we propose a Target-aware Spatial-Temporal Multimodal Fusion (TST-MF) framework, which jointly conditions multimodal fusion on the prediction target and spatial-temporal relations, enabling context-adaptive cross-modal interplay for target-specific representations. Extensive experiments over diverse datasets and tasks show that TST-MF delivers robust and competitive performance, achieving clear gains in multimodal settings.
[AI-69] Fused-Softmax for efficient quantized attention
链接: https://arxiv.org/abs/2611.09225
作者: Tianxing Wu,Juuso Korhonen,Jesus Carabano Bravo
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Performance (cs.PF)
备注:
Abstract:Low-precision matrix throughput on modern GPUs is growing far faster than vector-arithmetic throughput. Attention therefore cannot realize the available GEMM performance because softmax has become the pipeline bottleneck. We introduce Fused-Softmax Attention, a training-free algorithm that reformulates and approximates softmax to fold its dominant operations into the attention GEMMs, moving work from vector units to matrix cores. Implemented for the AMD CDNA 4 architecture, our MXFP8, MXFP6 and MXFP4 kernels reach approximately 3.0, 4.0 and 4.6 PFLOP/s on an AMD Instinct MI355X GPU, yielding speedups of 2.1\times , 2.8\times and 3.2\times over the BF16 attention. Importantly, the additional error introduced by Fused-Softmax remains below the error floor set by quantizing the GEMM operands. On Wan2.2-T2V-A14B, a hybrid MXFP6/4 schedule delivers up to 2.05\times end-to-end speedup while preserving the visual quality of the outputs. The kernels ship in AMD’s open-source AITER library, and they powered AMD’s MLPerf Inference v6.1 Wan2.2 submission, which ran 1.7\times faster than AMD’s previous-round result on the same hardware. Our results show that redesigning softmax to utilize the matrix pipeline enables attention to capitalize on continued growth in low-precision matrix throughput.
[AI-70] BOXBOX: Recursive Harness Self-Improvement for Long-Running Performance Optimization
链接: https://arxiv.org/abs/2611.09218
作者: Yuhao Tan,Hanyun Jiang,Fangkai Yang,Yuan Yao,Taolue Chen,Bo Qiao,Yu Kang,Lu Wang,Pu Zhao,Xiaoxing Ma,Qingwei Lin
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 47 pages, 22 figures, 12 tables, including appendices. Project page: this http URL
Abstract:Automated harness optimization improves agents by learning from trajectories and task feedback. Existing offline methods use reference answers to optimize harnesses for reuse on held-out tasks. However, many real-world tasks are open-ended, while long runs expose new difficulties that offline-optimized harnesses may not address. To study harness optimization in these settings, we introduce PerfRace, a benchmark for long-running performance optimization of software projects, and investigate whether agents can use task-specific experience to recursively improve their harnesses during an ongoing task. To this end, we introduce BOXBOX, which enables agent-triggered harness revision and kernel-managed hot reloading while preserving execution context. We evaluate BOXBOX on PerfRace under matched test-time budgets, comparing it with Frozen Harness, offline Harness Evolution, and Harness Scaling. BOXBOX achieves higher speedup with more efficient use of task time and model cost, reaching a Speedup@4h of 1.70x, compared with 1.54x for the strongest baseline.
[AI-71] On the Limits of Execution Traces for Repairing Assertion Failures in LLM -Generated Code
链接: https://arxiv.org/abs/2611.09212
作者: Muntasir Adnan,Carlos C. N. Kuhn
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: Under review at 2027 2nd International Conference on Advances in Artificial Intelligence and Machine Learning
Abstract:Iterative repair returns a failed program’s error to a large language model (LLM) and asks it to fix it. After iterative repair, the remaining failures are mostly assertion failures: with four Qwen2.5 models, they make up 79% of residual failures on BigCodeBench, all on HumanEval and 93% on LiveCodeBench. We argue, through an information view of the feedback, that an assertion failure says little about where the fault lies. We propose Breakpoint Debugging (BPD), which returns the failed program annotated inline with its variables’ runtime values. We take the tasks that iterative repair failed to fix with five repair attempts and compare two ways of spending five further generations on them: BPD, and resampling. On BigCodeBench, BPD repairs 26.7% of these tasks against 22.4%, and its advantage grows with model size. On HumanEval, BPD outperforms resampling but only slightly, and on LiveCodeBench, fresh sampling repairs more (6.5% against 4.9%). Furthermore, BPD reduces unchanged-code resubmissions from 28% to 8% on BigCodeBench. Using only the first failing test’s trace, it outperforms tracing all failing tests while using fewer tokens than both all-trace BPD and plain iterative repair. On LiveCodeBench, BPD gives no gain even on the public tests, so a visible failing input is not enough for a trace to help. These results match the information view: a trace helps only when the test exposes a failing input, and the fault is an implementation slip; otherwise, a repair loop should resample rather than trace.
[AI-72] Safe Reinforcement Learning via Probabilistic Option Shielding NEURIPS2026
链接: https://arxiv.org/abs/2611.09197
作者: Gian Luca Ascari,Isidoro Tamassia,Ying Jiao,Luc De Raedt,Giuseppe Marra
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at the NEmo workshop @ NeurIPS 2026
Abstract:Shielding is a safe reinforcement learning technique that provides safety guarantees while learning, by blocking unsafe actions. Traditional shielding relies on a complete model of the environment to provide long-term guarantees. Probabilistic logic shielding (PLS) lifts the requirement of having such model, but only considers safety in the short term, typically looking only one or few steps ahead. We propose Probabilistic Option Shielding (POS), which extends PLS to the option level, recovering long-horizon safety without reintroducing the need for a complete environment model. POS learns a model-free critic directly from one-step PLS signals, to estimate the probability that an entire option will complete safely under uncertainty. We evaluate POS on safe navigation tasks in the LogiCity environment, and show that it achieves considerably safer training while matching the task performance of standard step-level PLS and other safety related reinforcement learning baselines.
[AI-73] Analyzing Code Provenance in LLM s for Programming Education
链接: https://arxiv.org/abs/2611.09177
作者: Subhankar Maity,Faiza Belbachir,Omar Boucelma,Guilherme Medeiros Machado
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Large language models (LLMs) play an increasing role in programming education, challenging traditional assessment methods as they can generate fully functional code instantly. This disrupts the reliability of student submissions as indicators of learning and effort. To address this issue, code provenance is essential for tracing the origins of code and distinguishing student-authored, AI-generated, and externally sourced components. We extend the Min-K% probability approach by incorporating synthetic non-member code generation and apply it to Python code for membership inference and provenance tracking. Evaluations cover free LLMs such as GPT-2, DistilGPT-2, and DialoGPT, as well as state-of-the-art commercial models including GPT-5, Gemini 2.5, Mistral, DeepSeek, and LLaMA 3.3. Before evaluating our approach on higher-education student assignments, we validate its feasibility using Stack Overflow data. The results show that free models enable more sensitive code memorization detection than commercial models. By systematically comparing these LLMs, we assess trade-offs between detection accuracy and privacy protection, taking an initial step toward effective assessment of AI-assisted programming education through combined provenance detection and process visibility.
[AI-74] ailDART: Compute-Efficient Foundation Models for Financial Time-Series Forecasting
链接: https://arxiv.org/abs/2611.09176
作者: Chiang Qin Zhi,Ranjan Satapathy,Erik Cambria,Joyjit Chattoraj
类目: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE)
备注:
Abstract:Although time-series foundation models (TSFMs) achieve strong zero-shot transferability across high-information-density physical systems, applying overparameterized TSFMs to quantitative finance is fundamentally constrained by low signal-to-noise ratios, non-stationarity, and prohibitive autoregressive decoding latency. To reconcile representation learning with institutional operational constraints, this study presents TailDART, a dense 2.68 million-parameter foundation model engineered for financial time-series forecasting. By integrating grouped-query attention into a lightweight backbone, the architecture drastically reduces Key-Value cache memory bandwidth overhead, achieving major inference acceleration (up to 48.1%) over prevailing state-of-the-art TSFMs to enable low-latency, high-throughput streaming execution. Furthermore, multi-asset benchmarks across varying temporal resolutions show that a compact model can remain competitive for market forecasting: TailDART matches the predictive fidelity of substantially larger baselines while demonstrating superior error resilience in high-volatility regimes.
[AI-75] Evaluation of a Multi-Link Machine Learning Approach for Rainfall Detection Using Satellite Microwave Links
链接: https://arxiv.org/abs/2611.09156
作者: Mostafa Ftouni,Margherita Spalla,Matteo Colli,Andrea Randazzo,Christian Gianoglio
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Signal Processing (eess.SP)
备注:
Abstract:The impacts of climate change have increased the need for reliable real-time rainfall monitoring systems. This study investigates rainfall detection using a network of opportunistic sensors based on satellite microwave links (SMLs) combined with machine learning (ML) classifiers. Two models, a random forest (RF) and a convolutional neural network (CNN), were trained on aggregated SML data, including information from neighboring links. The results demonstrate that leveraging information from multiple links substantially enhances the detection of both wet and dry periods compared to single-link approaches. Moreover, the RF model consistently provides a more stable and reliable trade-off than CNN, improving rainfall detection and reducing misclassifications overall, with particularly strong performance during harsh weather conditions. The study spanned up to 32 rainy days between 2021 and 2024.
[AI-76] Learning When to Think: Dual-Reference Offline Optimization for Adaptive VLM Reasoning
链接: https://arxiv.org/abs/2611.09154
作者: Hongbin Lin,Sizhe Zou,Juangui Xu,Xinyue Xu,Zhuoli Ouyang,Stanislav Kolchin,Chengwei Qin,Zhongxiang Dai,Yao Shu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Reasoning traces improve vision-language models on complex multimodal tasks, but forcing every input to follow a long reasoning process leads to overthinking and unnecessary inference cost. We study adaptive fast/slow thinking for VLMs, where one model answers directly for simple inputs and reasons step by step when needed. Fast answers and slow traces often come from distinct behavior policies, making a single-reference offline objective misweight one mode. We propose DUAL-REFERENCE BOLT (DR-BOLT), which distinguishes the desired fast/slow mixture reference from the sampling distribution and corrects their mismatch in reward-weighted training. Fast/slow partition competition explains routing in the optimal target; an efficient source-wise correction approximates the exact objective when response formats are well separated. Matched Qwen3-VL comparisons use identical candidate banks to isolate the objective from slow-teacher quality. DR-BOLT improves score-length trade-offs over the evaluated offline and online baselines on ScienceQA and three held-out transfer benchmarks, with a 24.1% cross-benchmark slow ratio and a 2.90-fold end-to-end training speedup over reproduced GRPO. These results support adaptive reasoning in the studied two-reference setting, while broader robustness remains open. Our code is available at this https URL.
[AI-77] Learning Generative Dynamics for 3D Molecule Generation via Sequential Neuro-Symbolic Constraints
链接: https://arxiv.org/abs/2611.09142
作者: Zhuoran Li,Xu Sun,Chang Wen Chen,Wanyu Lin
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion models have recently emerged as a powerful paradigm for generating 3D molecules. Despite showing remarkable promise, they learn atomic distributions implicitly from data, without explicitly encoding physicochemical rules into their generative dynamics. Consequently, generated samples can be statistically plausible yet chemically invalid. To address this problem, we propose Sequential Neuro-Symbolic Constrained Diffusion (SensDiff), a framework that integrates symbolic rules into generative modeling, enabling the model to internalize scientific priors during training. Central to SensDiff is the observation that the diffusion trajectory of 3D molecule generation exhibits a coarse-to-fine evolution: valid molecules arise by first mitigating macroscopic geometric conflicts and then refining structures toward microscopic chemical consistency through iterative denoising. Accordingly, SensDiff mirrors this dynamic by sequentially enforcing constraints with adaptive weighting during training, progressively steering generation toward valid molecular geometries. Moreover, SensDiff translates domain knowledge into interpretable generative controls, grounding opaque neural denoising in scientific principles. Experiments on molecular benchmarks and real-world constrained generation tasks confirm that SensDiff consistently improves chemical validity while adhering to physicochemical rules and user-specified targets. The code is publicly available at this https URL.
[AI-78] Circuit Theory-Informed AI for Condition Monitoring in Power Electronics
链接: https://arxiv.org/abs/2611.09141
作者: Yubo Song,Pascal A. Schirmer,Subham Sahoo
类目: Emerging Technologies (cs.ET); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Systems and Control (eess.SY)
备注: The manuscript has been accepted for publication in the Proceedings of 2026 IEEE Design Methodologies Conference (DMC 2026)
Abstract:Artificial intelligence for power electronics is evolving toward a deeper comprehension of system intrinsic behaviors and dynamics. As a novel integration of neural computation and electric circuits, spike-based neuromorphic computing offers promising potential in high energy efficiency and adaptability for data-driven applications. Building on the latency-coding principles, this article proposes a framework for data-light condition monitoring that relies solely on three-phase AC current measurements. The AC current signals are encoded as membrane potentials of RC-based spiking neurons, which are inherently compatible with spiking neural networks that can effectively facilitate the estimation of device parameters. Case studies are presented to validate the performances, while constraints and further optimizations are also evaluated and discussed, and future extensions of the methodology are provided.
[AI-79] SPDMX: Large-Scale Open Stem Audio From Symbolic Music
链接: https://arxiv.org/abs/2611.09140
作者: Phillip Long,Eduardo Escoto,Julian McAuley,Zachary Novack
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multimedia (cs.MM); Audio and Speech Processing (eess.AS)
备注: 4 pages, 3 figures, 1 table
Abstract:Large public datasets have driven recent progress in machine learning, from text and vision to speech and general audio. Music remains an outlier, as copyright blocks redistribution of training audio, while many modern tasks need individual stems rather than mixtures. Existing multitrack resources trade off realism, scale, and licensing, while synthetic alternatives rely on proprietary VSTs or naive open rendering that yields weak stems. We introduce SPDMX, a listening-guided rendering of public-domain PDMX scores into an openly-licensed stem corpus with aligned scores and an open render pipeline: 253,982 songs, 490,159 mono stems, and 5,469 song-hours at 44.1 kHz under CC BY 4.0, with paired MIDI and identifiers linking to PDMX metadata. 88,403 songs ( 34.8% ; 2,734 song-hours) provide multitrack supervision. To our knowledge, this is the largest openly-licensed synthetic multitrack subset by song-hours, enabling source separation, stem-conditioned synthesis, MIDI-to-audio, transcription, and joint symbolic-audio modeling. We release the corpus, metadata, and open pipeline for reuse on other symbolic collections. Demos and experimental notes are on the project page.
[AI-80] CARGO: Causal Attribution and Responsibility-Guided Preventive Safe Multi-Agent Reinforcement Learning for Cooperative Driving DATE
链接: https://arxiv.org/abs/2611.09139
作者: Jingjia Teng,Yang Li,Wenjie Huang
类目: Artificial Intelligence (cs.AI)
备注: Abridged version containing the main results; the full version is under peer review and will be updated here upon acceptance
Abstract:This study presents Causal Attribution and Responsibility-Guided Optimization (CARGO), a multi-agent reinforcement learning framework for cooperative on-ramp merging in mixed traffic. A shared latent-context posterior and an action-conditioned response model estimate directed behavioral influence, with posterior uncertainty propagated into interaction-edge selection. Complementary attribution mechanisms couple signed influence with marginal team utility for cooperative reward shaping and assign responsibility from positive excess tail risk relative to a safe reference policy. Conditional diffusion generates future responses for action-dependent risk estimation. Responsibility-guided preventive optimization calibrates agent-specific risk-reduction requirements to local capacity within a policy trust region, using reserved safety margins, adaptive requirement scaling, and joint feasibility verification. Theoretical analysis establishes risk monotonicity and sufficient local conditions for non-decreasing environmental return under explicit assumptions. Experiments in highway-env across three traffic densities, together with controlled graph-recovery and attribution studies, support the effectiveness of the framework. Under high-density traffic, CARGO increases the success rate by 66.04% and reduces the crash rate by 56.23% relative to the best evaluated baseline for each metric.
[AI-81] When Routine Operations Break Architectural Pattern Relationships in LLM Agent Harnesses
链接: https://arxiv.org/abs/2611.09088
作者: Boming Xia,Qinghua Lu,Dino Sejdinovic,Liming Zhu
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:LLM agent harnesses distribute tool mediation, state management and safety guardrails across native services, independently written plugins and configuration. Routine operations can change the runtime order or retained state that architectural pattern relationships require, without any code change. We study five relationships associated with the Guardrail Enforcement Point, Autonomy Controller and Blackboard patterns across five released harnesses, analysing a main sample of fourteen independently written plugins. Using PatternOps, our assessment procedure, we trace requirements from pattern descriptions, provider documentation and deployment profiles to the providers that realise them and the records that decide them. We then compare alternative allocations under supported operations. Such operations broke relationships silently: after a plugin was reapplied or an allow rule saved, policy stopped evaluating edits while every plugin stayed active, and a prohibited edit executed. Which alternatives qualified depended on the requirement. Allocations that kept the required order, or kept or rebuilt the needed state, held on their tested paths and operation sequences, while others prevented prohibited effects but missed edits rejected elsewhere. In these harnesses, a pattern relationship is a property of the running system and its history, so architects should evaluate it against the decision paths, state lifetimes and operation sequences their deployments use.
[AI-82] Equidistribution of Solution Gradients in Physics-Informed Neural Networks via Coordinate Transformations
链接: https://arxiv.org/abs/2611.09053
作者: Tianchi Yu,Ivan Oseledets
类目: Numerical Analysis (math.NA); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Physics-Informed Neural Networks (PINNs) often struggle to approximate solutions with sharp gradients due to optimization difficulties and spectral bias. Motivated by the principle of gradient equidistribution from the stretching approach in grid generation, which redistributes solution gradients to reduce functional complexity, we propose Coordinate Transformation PINNs (CTPINNs), which incorporate learnable coordinate transformations to map the physical domain into a computational domain where the physical solutions become approximately affine, making them easier for neural networks to learn. We propose two constructions of the transformation: an explicit construction based on Majorants and an implicit construction based on invertable neural networks. Experimentally, the explicit construction can improve accuracy for PDEs with available explicit coordinate transformations, and the implicit construction can be combined with other advanced PINN models as an efficient plug-in module to improve performance across a range of challenging PDE benchmarks.
[AI-83] oolDex: An Ego-Exo Human Demonstration Dataset for Creative Physical Tool Use
链接: https://arxiv.org/abs/2611.09036
作者: Mohamed Elmoghany,Kehan Li,Ruoqu Chen,Tarun Chiruvolu,Shutong Zhang,Ryan Wang,Stephen Tian,Yaru Niu,Michelle Guo,Ding Zhao,Weiyu Liu,Mengdi Xu,Li Fei-Fei,Jiajun Wu
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Creative tool use is a fundamental human capability and an important step toward general-purpose embodied intelligence. It requires agents to reason about object affordances, physical constraints, and manipulation strategies, especially when standard tools are unavailable. However, progress in this area is limited by the lack of real-world datasets that capture human tool-use behavior with sufficient scale, fidelity, and diversity. Existing benchmarks are often simulation-based, focus on simplified interactions, or lack detailed observations and annotations. We introduce ToolDex, a large-scale egocentric-exocentric (ego-exo) human demonstration dataset for creative physical tool use. ToolDex captures scenarios in which humans select, adapt, and repurpose available objects as tools. Using synchronized egocentric cameras and eight exocentric cameras, the dataset provides high-fidelity multi-view observations and rich annotations for hands, tools, target objects, and their deformable interactions. Beyond data collection, we establish a benchmark suite characterizing three core capabilities: physical tool-use reasoning, sparse-view 3D and 4D reconstruction, and robot policy learning. ToolDex provides a foundation for studying human tool-use behavior and developing embodied agents that can adaptively use tools in complex real-world environments. Dataset visualizations and benchmarking results are available at this https URL. Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.09036 [cs.RO] (or arXiv:2611.09036v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2611.09036 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-84] A Controlled Study of Models Skills Feedback and Curriculum in Agent ic Robotics
链接: https://arxiv.org/abs/2611.09030
作者: Kaiyuan Chen,Xinyue Zhu,Shuangyu Xie,Uksang Yoo,Thi Thu Phuong Luong,Marcus Gualtieri,Christian Juette,Liu Ren,Spencer Huang,Ken Goldberg
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Agentic robotics asks coding agents to implement and iteratively improve robot policies from simulated and real-world physical interactions. We study the self-learning loop of agentic robotics through a controlled simulation study spanning nine manipulation tasks, 631 different experiment configurations, and 28K coding agent rollouts, systematically varying model capability, skill priors, feedback structure, and curriculum. The experiment data suggests that the bottleneck is often not the access to sufficient information, but interpreting it into the right policy update. Across these factors, a successful self-learning loop depends on how effectively agents attribute failures, reason task-relevant spatial and geometric structure, and selectively restructure code to improve policies.
[AI-85] Scout Before You Route: Predicting Task Success and Token Count from a Small LLM Prefill
链接: https://arxiv.org/abs/2611.09028
作者: Aristides Milios,Christopher Pal
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 27 pages, 4 figures, 17 tables
Abstract:Routers typically choose a model per query by trading predicted correctness against cost. Most price each model at a fixed cost, as if every query used the same number of output tokens, but the output length of reasoning models varies by an order of magnitude across queries. When is per-query cost worth predicting? What factors drive cost? How cheaply can predictions be made? We predict both the success probability and the output length of candidate models from the prefill activations of a single small frozen LLM, the scout. With reasoning candidates, perfect per-query cost knowledge would save 29-61% of spending on eight of nine datasets, and much less with non-reasoning candidates on the same problems. On those eight datasets, our predicted costs save 9-33% at matched accuracy relative to median-length pricing, up to 69% of what perfect knowledge would save. They are significantly cheaper than each estimator we reimplemented or adapted from published work, including ZeroRouter and a MixLLM-style predictor, on a majority of datasets. Output length depends mostly on the problem. On most math and code datasets, harder problems are longer, so predicted success alone prices cost nearly as well. On knowledge and mixed-format datasets they are not, and predicted success falls well behind a probe trained on length. The probe keeps most of its saving with 20 labelled problems and transfers without labels to four of seven new benchmarks. On a mixture of all nine benchmarks, it saves 22% beyond pricing each benchmark at its own median length. Savings stop growing at about a 2B scout, so above 4B a larger scout costs more than it adds.
[AI-86] Optimal Probabilistic Answer Set Programming: Complexity and Implementations
链接: https://arxiv.org/abs/2611.09022
作者: Damiano Azzolini,Giuseppe Mazzotta,Francesco Ricca,Fabrizio Riguzzi
类目: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI)
备注: Accepted for publication in Theory and Practice of Logic Programming
Abstract:Probabilistic Answer Set Programming (PASP) extends Answer Set Programming with the possibility of representing uncertain data. Among the possible available semantics, we focus here on the credal semantics. In this setting, prior work has focused exclusively on plain ASP programs, i.e., programs without optimization constructs. However, many real world tasks define optimality criteria over the possible solutions, which cannot be conveniently modeled with existing probabilistic ASP frameworks. In this paper, we close this gap and propose optimal Probabilistic Answer Set Programming, which extends PASP by also allowing weak constraints, which are the standard optimization construct of ASP. We define the semantics of such programs, study the computational complexity of the inference and most probable explanation tasks, and propose practical implementations based on Algebraic Model Counting and Answer Set Programming with Quantifiers. An empirical evaluation of the proposed approach on three real world tasks confirms the applicability of our framework.
[AI-87] SIMBAD: Spatio-Temporal Traffic Forecasting Robust to Aperiodicity NEURIPS
链接: https://arxiv.org/abs/2611.09015
作者: Daniel Y. Lee,Seungwon Shin,Seunghoon Han,Sungsu Lim,Susik Yoon
类目: Artificial Intelligence (cs.AI)
备注: Total of 30 pages, 9 figures (4 in main, 5 in appendix), accepted to NeurIPS(Neural Information Processing Systems) 2026
Abstract:Accurate traffic prediction is essential for urban planning but remains challenging due to the irregular patterns caused by unexpected events. While Spatio-Temporal Graph Neural Networks (STGNNs) effectively model periodic data, they often struggle with these aperiodic fluctuations. To address this, we propose SIMBAD, which improves generalization irregular patterns via two key mechanisms: 1) adaptive regulation of periodic signals to prioritize recent signals, and 2) dynamic spatial modeling that adjusts influence between connected nodes based on contextual relations. Built upon simple similarity metrics to adapt to aperiodic patterns, SIMBAD outperforms existing methods on real-world benchmark datasets, demonstrating superior performance in predicting unseen and irregular traffic events.
[AI-88] Capital Efficiency Isolation Risk and Ruin in Generative AI Inference Serving Infrastructure
链接: https://arxiv.org/abs/2611.08999
作者: Renée St. Amant,Anjaly Parayil,Ankur Mallick,Victor Rühle
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
备注:
Abstract:Generative AI serving is an extremely capital-intensive endeavor that demands rapid, step-change improvements in capital efficiency to prevent industry wide increases in the cost of capital from limiting growth plans. Keeping pace with the economic necessities will increasingly require infrastructure providers to confront operationally-inconvenient realities of production serving, including the extreme heterogeneity of generative AI request costs. We have explored workload isolation as a foundational, fleet-level lever for increasing capital efficiency. Isolation is the practice of intelligently grouping “like” requests before execution. Rather than exposing every replica to the full heterogeneous workload distribution, isolation deliberately reshapes the workload presented to the hardware by assigning requests with different cost characteristics to different portions of the fleet. Workload isolation on its own does not make expensive requests cheaper; it prevents a costly portion of the workload from unnecessarily determining the operating point and capital efficiency of the entire fleet. It additionally lays the foundational groundwork that can serve as a launching point for compounding gains via narrow-band optimizations (tailoring) and custom silicon designs while also informing up on pricing policy. This paper uses representative production data and experience to convey the motivation behind isolation as a fleet-level lever for increased capital efficiency.
[AI-89] Rethinking Optimization Methods for Hierarchical Feature Selection
链接: https://arxiv.org/abs/2611.08991
作者: Xinyuan Song,Zekun Cai,Liang Zhao
类目: Artificial Intelligence (cs.AI)
备注: Under review
Abstract:Rare-event forecasting often depends on interpretable conjunctions: a single keyword, location, or temporal cue may be weak, while their joint occurrence can identify a meaningful precursor. Hierarchical-Task Numerical conjunctive feature Learning (HTNL) selects Numerical Conjunctive Features (NCFs) through a structured sparsity objective. However, the resulting optimization problem is challenging, making optimization design critical to its practical performance. We give a unified diagnosis of four HTNL optimization families: original majorization-minimization (MM), variational quadratic form (VQF), alternating direction method of multipliers (ADMM), and weighted group lasso (WGL). Our analysis distinguishes three key factors that jointly determine HTNL optimization behavior. First, VQF and MM are algebraically equivalent under matched variational updates and optimize a squared-penalty formulation. Second, ADMM optimizes the unsquared objective through consensus splitting; we replace its truncated FISTA proximal solver with a scalar bisection update (ADMM-Bisect) for more accurate treatment of near-threshold blocks. Third, WGL targets the same unsquared objective, while all methods face rapidly decaying active-set admission scores. Across two synthetic datasets and 11 civil-unrest and flu-surveillance tasks, VQF achieves the highest mean AUC with the most compact non-collapsing supports. ADMM-Bisect reduces measured runtime, while calibrated WGL avoids the empty-support failures of its original scale. Synthetic results further show how proximal accuracy, pruning, and search budget affect candidate recovery, especially for longer conjunctions. Code will be made publicly available upon publication.
[AI-90] PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals
链接: https://arxiv.org/abs/2611.08969
作者: Yanlong Chen,Simone Benatti,Luca Benini,Cuntai Guan,Yawei Li
类目: Artificial Intelligence (cs.AI)
备注: 15 pages, 11 figures, 5 tables. Code: this https URL , Models Weights: this https URL
Abstract:We present PhysioWave-T, a token-efficient wavelet foundation model family for EEG, ECG, and sEMG. Dynamic scale folding recombines J+1 wavelet subbands prior to tokenization, compressing sequence length from (J+1)CS to CS tokens for J levels, C channels, and S temporal patches. Route-specific front ends and learned channel-name embeddings allow one Transformer per modality to handle diverse montages and sampling rates, while unseen setups reuse the backbone via lightweight adapters. We pre-train a 30M-parameter EEG model on over 15M windows (19–128 channels), alongside ECG (30M) and sEMG (12M) encoders. Across four EEG routes, scale folding cuts Transformer latency by 4.0–7.2x and peak memory by 6.8–12.7x over scale expansion without accuracy loss. Our EEG encoder leads 8 of 9 metrics against BENDR, BIOT, LaBraM, and EEGPT, and surpasses LUNA-L on wearable seizure detection with fewer parameters. Reliability-aware latent fusion (RALF) further integrates summary tokens from independently pre-trained encoders via sample-adaptive weights. On SeizeIT2, RALF raises event sensitivity at 11 false alarms/h from 61.4% to 74.4% and window AUPRC from 0.321 to 0.399 over the best unimodal baseline, outperforming weighted-logit and concatenation fusion across all variants. PhysioWave-T also achieves state-of-the-art sEMG accuracies and top macro-F1 on PTB-XL Super and Chapman-Shaoxing. Code and models: this https URL.
[AI-91] More with Less: Efficient Parallel Scientific Code Translation via an Agent -Skill Workflow
链接: https://arxiv.org/abs/2611.08947
作者: Jianuo Shen,Pengyi Xu,Bo Zhang
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: Accepted to AgenticAI4HPC 2026 (SC26 Workshops). 10 pages
Abstract:Translating parallel scientific codes between HPC programming frameworks is a critical but labor-intensive task. Existing LLM-based approaches rely on fixed sequential pipelines with undifferentiated repair strategies and correctness criteria that do not match scientific computing practice. We present an agent-skill framework that separates translation, compilation diagnosis, and execution validation into three autonomous agents coordinated by a verdict-driven orchestrator, routing each failure to a targeted repair skill rather than retrying blindly. An automated output check drives the self-correction loop, complemented by an offline LLM-based verification that exposes false positives escaping compiler and runtime checks. We evaluate the framework on 30 HeCBench kernels translated bidirectionally among CUDA, OpenMP, and SYCL with three open-source LLMs. Our framework attains the higher verified correctness rate in all 18 model and direction combinations over a state-of-the-art baseline while consuming up to 52.4% fewer tokens, remaining consistent across models and directions without architectural modification.
[AI-92] Ordered Policy Optimization NEURIPS2026
链接: https://arxiv.org/abs/2611.08936
作者: Dongkyu Derek Cho,Pan Xu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Robotics (cs.RO); Machine Learning (stat.ML)
备注: 36 pages, 12 figures, 11 tables. Published in Proc. of the 40th Conference on Advances in Neural Information Processing Systems (NeurIPS 2026)
Abstract:Proximal Policy Optimization (PPO) stabilizes policy updates by clipping the importance ratio between the new policy and the data-collecting policy. The standard rule uses one fixed clipping range for all samples. While this choice offers simplicity and effectiveness, it fails to distinguish statistically harmful extreme ratio values from large, yet tolerable changes. We study how to adapt the clipping radius to the reliability of the induced importance ratio distribution. Our case study demonstrates that reliable updates require controlling these extreme ratios, rather than merely verifying that individual ratios fall within a fixed interval. Motivated by this observation, we propose Ordered Policy Optimization (\textscOPO). In each minibatch, \textscOPO sorts the importance ratios, selects the extreme ratios that potentially have the largest effect on the update, and assigns them adaptive sample-wise radii instead of using the fixed clipping radius. These radii are derived from a prescribed tail profile, which specifies desired target values for the selected extreme ratios. The resulting objective remains first-order and requires only minibatch sorting and a modified policy loss. Experiments conducted on MuJoCo and Atari benchmarks demonstrate that this simple modification achieves strong performance relative to established baselines.
[AI-93] DRAFT-Bench: Evaluating Frontier Agents Across the Professional Design Lifecycle
链接: https://arxiv.org/abs/2611.08933
作者: Rong Pan,Jianqiang Wang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Can frontier general-purpose agents understand client intent, explore creative directions, execute revisions, and prepare production deliverables? We introduce DRAFT-Bench (Design Reliability Across Feedback and Tasks), a lifecycle-oriented evaluation of requirement elicitation, creative exploration, constrained editing, iterative revision, and production handoff. Its Requirement-to-Artifact framework links evolving requirements, protected properties, artifact versions, and delivery obligations. We compare native Codex/Astra and Claude Code across 35 design briefs under 43 conditions, alongside a separate 40-scenario RFQ track. Astra completes 43/43 design conditions and Claude 35/43. All scored core runs receive full structural credit, yet on the 28 core conditions scored for both systems, semantic checks flag violations in 11 for Astra and 5 for Claude; these are model judgments, not expert-confirmed failure rates. RFQ coverage is high, but success ordering changes across cohorts. SKU expansion, revision dependencies, and White ink files expose ambiguity among unmet obligations, missing inputs, and evaluator interpretations. These findings separate execution, requirement compliance, and professional acceptance. Unpaired tracks and different tools preclude claims of RFQ-to-delivery success, factory acceptance, or a pure model ranking.
[AI-94] Glyph: An Interpretability Engine for AI-Driven Scientific Discovery in Clinical Data
链接: https://arxiv.org/abs/2611.08922
作者: Robbie Holland,Ashwin Kumar,Mateus A. Esmeraldo,David Le,Andrea Fisher,Jiazhen Pan,Louisa Fay,Magdalini Paschali,Vasiliki Sideri-Lampretsa,Ates Fettahoglu,Stefania Moroianu,Philipp Wesp,Ivan Lopez,Tom Nuno Wolf,McKenzie S. White,Eduardo Pontes Reis,Fatima Rodriguez,Curtis P. Langlotz,Vasilis Syrgkanis,Imon Banerjee,Akshay S. Chaudhari,Sergios Gatidis
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automated scientific discovery has the potential to advance understanding of disease mechanisms and accelerate identification of early disease markers. Manual data annotation is infeasible at population scale, motivating methods that automatically extract pre-defined phenotypes in multimodal and unstructured clinical data. Such pre-defined feature sets determine and limit which associations these approaches can discover. This limitation extends to agentic AI systems that use these tools to query population-scale data. In contrast, foundation models (FMs) trained on large volumes of clinical data can perform tasks beyond their training objectives. This suggests their latent embeddings may represent novel or unexplored features that could yield insights into disease mechanisms, yet such latent features remain difficult to interpret. To address this challenge, we introduce Glyph, an automated multistage framework that extracts features from multimodal and unstructured clinical data using FMs and sparse autoencoders, interprets them as coded clinical concepts, and evaluates their value for clinical scientific discovery. We applied Glyph to abdominal CT volumes, CT radiology reports and laboratory measurements from 41,628 patients across different institutions. We identified 397 predictive concepts supported by 4,463 Bonferroni-significant associations. Moreover, a majority of discovered associations were automatically reproducible in held-out cohorts across different sites. Glyph concepts provided complementary associations to literature-derived markers, leading to gains in predictive power. We envision that interpretability can extend beyond a mechanistic understanding of neural networks to connecting implicitly learned features with clinical knowledge and establishing their significance for patient outcomes. We release Glyph at this https URL.
[AI-95] Frame: Form-Restricted Authoring with Model-Based Evaluation
链接: https://arxiv.org/abs/2611.08921
作者: Nan Jiang,Mohammed Mohsin Hussain,Onkar Tiwari,Ankit Jain,Thong Q. Nguyen,Abhishek Agarwal,Vishal Chowdhary
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Many professional reports reuse the structure of prior editions while updating the content. We formalize this setting as form-restricted report generation: given a sample report and new grounding documents, generate a report that preserves the sample’s form while accurately reflecting the new evidence. We introduce FRAME, a benchmark of 120 EDGAR- and arXiv-derived instances across Financial Quarterly, Financial Annual, Academic Simple, and Academic Complex categories, built with a temporal three-snapshot pipeline. We evaluate outputs with three human-validated metrics: structural similarity via semantic tree-edit distance, coverage via information unit recall, and faithfulness via claim-level grounding verification. Across five LLMs, open-source models nearly match proprietary models on structure but lag on coverage and faithfulness; all models perform better on financial than academic reports, and performance declines with longer outputs and more complex grounding.
[AI-96] HomeLoop: A Simulation Framework and Benchmark for Robot-IoT Coordination in Everyday Domestic Assistance and Safety
链接: https://arxiv.org/abs/2611.08914
作者: Ramy ElMallah,Marc Bishara,Hyegang Kim,Yuxi Zhang,Chongyu Zhu,Chi-Guhn Lee
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent progress in embodied robotics is bringing robots closer to deployment in homes, as vision-language models (VLMs) are demonstrating significant advances in reasoning, making VLM-driven robots promising home assistant agents. These agents will often be introduced into homes that already have cameras, Internet of Things (IoT) devices, and residents with wearables. To fulfill their potential, they need to integrate with existing smart infrastructure and reason through ambiguous scenarios: a smoke alarm may come from a pan about to ignite, a pan already on fire, a resident cooking or a faulty sensor, and each calls for a different response. To that end, we introduce HomeLoop, a simulation framework and benchmark built on Isaac Sim that combines robot bodies, networked devices and wearables, simulated residents and visitors, faulty sensors, and hazards that evolve over time. Our benchmark is composed of 20 tasks across five families grounded in real-world household hazards and needs, and designed so that the same alert can have different causes. A runtime monitor scores each episode with temporal-logic checks on the final outcome and the actions along the way, and every action carries a predefined cost. We evaluate six open-weight and proprietary VLMs. The best achieves a family-balanced score of 61.9%, and across models we see a pattern of over-reaction: failed episodes involving prohibited interventions outnumber those involving missed required interventions by more than two to one, with models often shutting off the stove before investigating the alarm’s cause. HomeLoop is an extensible foundation intended for use in research on robot-IoT coordination, long-term memory, proactive assistance, and privacy-safety trade-offs. HomeLoop will be released with 2,100 home layouts that can be automatically furnished in diverse configurations, five robot bodies and 20 networked device types.
[AI-97] oward LLM Self-Play Under Data Scarcity: Effectiveness Generalizability and Trustworthiness
链接: https://arxiv.org/abs/2611.08908
作者: Zhiyin Yu,Yuchen Mou,Bingzhe Wu,Zhonghai Wu,Shiyang Feng,Xiao Luo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The rapid development of large language models (LLMs) has been driven largely by scaling model size and human-curated training data, yet this paradigm is increasingly constrained by data scarcity and diminishing returns from human supervision. Self-play offers a promising pathway for autonomous data generation and improvement, but current frameworks remain unsatisfactory due to a lack of a data-centric perspective, limited cross-domain generalization, and growing trustworthiness and security concerns. In this paper, we argue that real-world LLM self-play systems should achieve effectiveness, generalizability, and trustworthiness under data scarcity. To address these challenges, we outline three blue-sky research directions for future self-play systems: (1) data-centric self-play for effective experience generation, (2) generalizable self-play for open-world and unverifiable tasks, and (3) trustworthy self-play for reliable self-evolution. We further discuss research opportunities and solutions within each direction. This paradigm shift will establish a sustainable and safe AI ecosystem by transforming finite human knowledge into inexhaustible high-quality experience through self-play architectures.
[AI-98] Whos Asking? Audience-Conditioned Privatization
链接: https://arxiv.org/abs/2611.08884
作者: Daniil Filienko,Anusha Bagalkotkar,Xan Vongsathorn,Sikha Pentyala
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:The abundance of user-generated conversational data makes the question of how to responsibly derive insights from confidential data without putting the individuals in it at risk ever more important. However, the majority of existing differentially private (DP) synthetic text generation (SDG) and anonymization approaches ignore the context of the data’s downstream usage, which results in loss of important information. DP-SDG provides the same guarantee regardless of the recipient, and anonymization relies on predefined attribute categories. They both ignore the observation that an attribute that is harmless to one audience can be sensitive to another, and each audience typically needs only some attributes preserved. We propose Ext-Classify-Priv, a framework that (1) extracts the information types present in a corpus, (2) classifies which of them are private for a given audience following contextual integrity, and (3) protects only those attributes, using either anonymization or SDG with Feature-DP approaches. On Reddit data with 3 audiences, our audience-specific releases are more faithful than a single audience-agnostic release for every audience. On bioRxiv data, our SDG method improves MAUVE from 0.097 to 0.420 at epsilon=1 and on the proprietary real estate data, our anonymization method improves MAUVE from 0.698 to 0.959 over the best baselines.
[AI-99] Rethinking Agent Trajectory Compression: A Workflow-Aware Perspective
链接: https://arxiv.org/abs/2611.08865
作者: Qijia Zhuang,Rui Liu,Yuxiang Ren
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 4 figures
Abstract:As long-horizon agents increasingly rely on multi-step tool interactions to accomplish complex tasks, their execution trajectories continuously accumulate structured outputs and intermediate artifacts, leading to growing context overhead. Existing compression methods primarily rely on text importance or fixed truncation rules, often overlooking cross-step dependencies within workflows and potentially discarding critical information required for subsequent task execution. To address this challenge, we propose WAAC, a workflow-aware framework for agent trajectory compression. WAAC structurally models successful historical trajectories to extract reusable workflow compression experience and incorporates online backward dependency analysis to dynamically identify information retention requirements. By combining current dependency constraints with historical experience, WAAC selects appropriate compression strategies for different types of historical information. Experiments on PinchBench and Odysseys demonstrate that WAAC achieves an overall score of 87.6 and a perfect completion rate of 19.0%, respectively, while reducing execution costs by 19.7% and 29.1% compared with the strongest baseline methods. These results demonstrate the effectiveness of workflow-aware trajectory compression in improving task performance and cost efficiency for long-horizon agents. Code and resources are available at this https URL.
[AI-100] Adaptive Scheduling Pipeline for Multi-Instance Asynchronous Reinforcement Learning
链接: https://arxiv.org/abs/2611.08864
作者: Salah Chikhi,Luis Zuin,Entong Li,Zeng Li
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:We present the Adaptive Scheduling Pipeline (ASP), an offline rollout scheduler for multi-instance asynchronous reinforcement learning (RL) of large language models (LLMs). ASP jointly reduces three sources of rollout underutilization - padding bubbles from intra-minibatch length heterogeneity, skewness bubbles from long-tail outputs, and sorting bubbles from fluctuations in training-queue occupancy induced by length-sorted dispatch - through a single offline preprocessing pass. This enables efficient multi-instance rollout that overcomes the early scaling plateau of single-instance inference while smoothing the arrival of generated samples at the trainer. We integrate the scheduler into an asynchronous architecture that decouples rollout from training while keeping policy lag observable and bounded. Complete ASP improves end-to-end throughput by up to 3.22 times and reduces iteration time by up to 76.8% relative to MindSpeed-RL, while reducing iteration time by up to 62.5% relative to verl. Within ASP, scheduling reduces mean iteration time by 10.9%. ASP also scales more robustly with additional generation resources than these baselines and shows no visible convergence degradation in the measured bounded-lag regime. ASP has since been used in an internal telecom-domain RL post-training workflow, reducing the reported training period from more than three months to about two months and producing a domain-enhanced model for knowledge answering and tool calling.
[AI-101] Mind the Divergence: Divergence-Guided Reliability-Aware Fusion and Adaptive LLM Prompting for Conversational Emotion Recognition
链接: https://arxiv.org/abs/2611.08860
作者: Jingwen Wang,Jiao Li,Yuhang Qian,Yuyin Wang,Min Zhang
类目: Artificial Intelligence (cs.AI); Sound (cs.SD)
备注:
Abstract:Multimodal emotion recognition in conversations (MERC) integrates textual, acoustic, and visual cues to identify utterance-level emotions. However, existing approaches generally do not explicitly adapt multimodal processing to the degree of inter-modal predictive disagreement. In real conversations, semantic incongruity, concealed emotion, noisy observations, and ambiguity can affect modality-specific emotional evidence differently, while the reliability of individual modalities varies across samples. Rather than identifying each source of inconsistency separately, we model their common consequence for emotion recognition as predictive disagreement among unimodal emotion distributions over the same emotion-label set. Such disagreement makes indiscriminate fusion prone to propagating unreliable or competing evidence. Inspired by the Brunswik Lens Model, we propose MiD-ERC, a divergence-centric framework that quantifies inter-modal predictive disagreement using a normalized divergence score and estimates sample-specific modality reliability. MiD-ERC uses these signals to coordinate divergence-guided cross-modal gating, divergence-modulated reliability weighting, and divergence-conditioned adaptive LLM prompting. The framework is trained end-to-end using only standard emotion annotations, without requiring additional conflict labels, auxiliary data construction, or external retrieval modules. Experiments on IEMOCAP and MELD demonstrate that MiD-ERC achieves state-of-the-art performance, with particularly pronounced improvements on high-divergence samples.
[AI-102] Learning Whom to Trust About the Unseen: Source Credibility and Reliability Transfer in Language Models
链接: https://arxiv.org/abs/2611.08855
作者: Ruijia Guan
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:A source who has been right before deserves attention, but about what? We ask how far a demonstrated record of reliability travels in language models when the new claim cannot be checked, and whether the answer depends on the kind of claim. Three behavioral studies across six models approach the question from three angles. Study 1 adapts the Einstein-effect paradigm and finds that models, unlike the human sample, favor a spiritual leader over a scientist for the same obscure statement, and that the preference largely reverses when the statement is reworded in the vocabulary of physics. Study 2 adapts a paradigm in which a trusted expert calls a scientific or religious explanation unknown or a mystery; models reproduce the human domain asymmetry in exaggerated form. Study 3, the central test, gives two fictional informants different records on checkable river forecasts and then has them endorse matched claims about hidden natural and spiritual causes of the same village events. Most models carry the record to natural causes and barely at all to spiritual ones; one model transfers it weakly to either. Telling the model that the reliable informant forecasts by spiritual guidance rather than by measurement lowers that informant’s influence on natural claims in every model and produces no comparable gain on spiritual claims in any. Demonstrated reliability thus has a scope in these models, and the account a source gives of how they came to be reliable can narrow it. Whether this reflects a judgment about relevance or the low prior credence of spiritual claims is the question the design leaves open, and we outline the experiment that would settle it.
[AI-103] When F1 Is Not Enough: Auditing Codebooks for LLM -Assisted Annotation
链接: https://arxiv.org/abs/2611.08846
作者: Yuehong Cassandra Tai,Qunhua Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are increasingly used for text annotation in social science, but standard performance metrics do not explain why errors occur. Low F1 may reflect limitations of the annotation system, defects in the written codebook, or inconsistency in reference labels. We propose a pre-deployment codebook audit for LLM-assisted annotation. The audit uses class-specific metrics and confusion matrices to identify problematic classes and boundaries, structured document-level review to attribute error sources, and targeted codebook revision followed by held-out evaluation. We apply the audit to two published multi-class codebooks from distinct domains. In both applications, the audit identifies under-specified or conflicting coding rules, and targeted revisions reduce the diagnosed errors on held-out data. Analyses with two LLM annotators show that some problematic boundaries and codebook revisions transfer across models, whereas other revision effects are annotator-specific. The results show that LLM annotation errors can help identify weaknesses in the written measurement instrument, not only in the model. Codebook auditing complements prompting, retrieval, and fine-tuning by clarifying whether persistent errors call for model improvement, codebook revision, or renewed scrutiny of reference labels.
[AI-104] What Matters for LLM Persona Simulation: Signal Form in Cold-Start E-Commerce Advertising RECSYS2026
链接: https://arxiv.org/abs/2611.08840
作者: Jiwoo Choi,In Ik Lee,Woo Sung Seo,Minho Lee,Yun Young Choi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at GenAIECommerce’26: The Third Workshop on Agentic and Generative AI for E-Commerce, co-located with RecSys 2026. 13 pages
Abstract:Estimating advertising performance in cold-start settings is challenging because historical logs are unavailable before launch. Large language models (LLMs) offer a potential alternative by simulating persona-conditioned user responses, but it remains unclear which design choices actually drive LLM-based persona simulation. Using real Google Ads e-commerce campaign data across five demographic segments, we identify three signal-form dimensions that affect simulation quality: (i) a demographic anchor without noisy rich attributes, (ii) a relative qualitative banner-blindness signal rather than quantitative anchors, and (iii) prompt phrasing. On this campaign, our method attains an MAE of 24.9 and a MAPE of 26.1%, a 51.3% lower MAE than the best tested non-LLM baseline and 77.3% lower than an LLM configuration using explicit CTR anchors. Systematic ablations show that minimal demographic descriptions achieve lower mean error than richer persona configurations, while relative qualitative cues substantially outperform the tested quantitative anchors. We additionally compare these configurations with direct-prior and rule-based baselines and with alternative LLM configurations.
[AI-105] Solver-Admitted Metamorphic Testing of Learned Network Optimization Surrogates
链接: https://arxiv.org/abs/2611.08837
作者: Vineet Jagadeesan Nair,Stephen Peng,Arun Ravindran
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Equivalent encodings of an optimization problem can produce different learned predictions. Interpreting these differences requires accounting for numerical error and multiple optima. We construct metamorphic tests for optimization solver surrogates with explicit eligibility conditions, output alignment, and numerical thresholds. Reference-solver agreement and independent constraint checks determine which sampled pairs enter the audit. We study alternating-current optimal power flow (AC-OPF), the traveling salesman problem (TSP), the capacitated vehicle routing problem (CVRP), and maximum independent set (MIS). Across five grids, GridSFM-Open violates each of 13 relations on at least one grid, and 12 by more than twice the threshold. Randomly initialized copies of the grid models fail nearly the same relations, so the failures do not require training. Source patches and input canonicalization remove selected discrepancies, but no intervention establishes preserved accuracy and feasibility within the declared margin on every grid. In routing, Policy Optimization with Multiple Optima (POMO) has a single-pass rotation/reflection violation rate that rises from 12.7% to 89.5% between 20 and 100 cities. Graph-Based Diffusion Solvers for Combinatorial Optimization (DIFUSCO) violates the MIS disjoint-union relation on 35.6% of draws. Post hoc normalization within connected components removes these violations on the tested graphs. These findings support reporting metamorphic consistency alongside optimality, feasibility, and accuracy, and validating repair quality separately from consistency.
[AI-106] Notwithstanding Anything to the Contrary Ignore All Instructions: Measuring Prompt Injection Guards on Legal Contract Review Attacks
链接: https://arxiv.org/abs/2611.08834
作者: Mark Heimann,Katherine Taylor,Bruno Jacob
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Specialized prompt injection guards are deployed in front of document analysis systems, but no false positive rate has been published for any of them on legal contracts. We measure eleven open-weights guards on two corpora of legal contracts and find that existing guards struggle: document-level false positive rates are as high as 92.39%; two guards fall below chance on paired AUC; and a five-regex baseline is not Pareto-dominated by any guard in the panel. Controlled probes locate model-specific as well as general failure modes. Prompted LLMs resolve the tested failures, including a second threat category (unauthorized scope claims) on which every guard scores zero. A cascaded LLM solution, where the positives of a mid-tier model are verified by a frontier model, is 6.5x cheaper than using exclusively a frontier model and approaches its performance.
[AI-107] I Think Therefore I Attack: Evaluating Thought Injection Attacks on Reasoning Agents
链接: https://arxiv.org/abs/2611.08811
作者: Parth Illendula,Garik Livingston,Zain Shamsi,Ravi Srinivasan,Alex Liu
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Large Reasoning Models (LRMs) are one of the key components of current agentic systems. LRMs augment their output with an explicit chain-of-thought reasoning stage, introducing a new attack surface for thought injection, in which untrusted text resembling the model’s own deliberation is fed back to the model. We present a two-part study of this attack, performed against the prevailing large, production class open-weight models (e.g., GLM-5.3, MiniMax-M2.5, Qwen3.8-Flash-Next). We first evaluate the effectiveness of static thought-injection prompts under different styles as well as tagged and untagged conditions and find that the effects of thinking tags and payload styles are all properties of the specific target model. In the second study, we exploit this model-specificity in a containerized agentic cyber range, using indirect injection prompts styled in the form of the target model’s own reasoning traces to achieve attack success rates (ASR) up to 50% on operational objectives such as reverse-shell exfiltration and repository backdooring. We also find that the same malicious objective in standard prompt injection user voice never succeeds, and thinking tags suppress the agent’s tendency to disclose the injection. Our results show that imitating a reasoning model’s own thought style is effective for this style of attack, and that tag stripping is an unreliable defense.
[AI-108] he Matthews Correlation Coefficient: Fifty years
链接: https://arxiv.org/abs/2611.08799
作者: Jesus S. Aguilar-Ruiz
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Fifty years after its introduction, the Matthews Correlation Coefficient (MCC) remains central -and debated- in binary classification. This survey offers a comprehensive retrospective and an in-depth analysis of MCC, covering its historical development, mathematical foundations, and implications for practice. We clarify long-standing misconceptions, formalize its equivalences (to Pearson’s correlation and to Chi2 in 2x2 tables), and examine its behavior under class imbalance. Beyond MCC, we develop analytical (derivatives) and graphical (contour plots) frameworks that generalize to any confusion-matrix-based metric. These frameworks expose structural properties -bias, robustness, sensitivity- that scalar summaries conceal, and they provide actionable guidance for selecting and interpreting performance measures in applications where reliability and fairness matter. Together, our results position MCC within a broader toolkit while clarifying when it is informative -and when alternatives may be preferable.
[AI-109] Not All Latents Are Equal: Importance-Weighted Latent Graph Generation
链接: https://arxiv.org/abs/2611.08788
作者: Biraaj Rout,Jay Revolinsky,Han Liu,Harry Shomer
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Graph generation has shown tremendous effectiveness in modeling complex scientific tasks ranging from molecule generation to optimized circuit design. Most recent advancements have developed models that learn to denoise noisy graphs using either discrete diffusion or flow matching. However, discrete graph diffusion scales quadratically with the number of nodes, making it difficult to use such models on larger graphs. On the other hand, latent graph generation instead seeks to denoise noisy graph latents, which scale better to more nodes due to their lower dimension. However, in practice, latent graph generation models lag behind their discrete counterparts in performance. We study this problem, finding that it is caused by the unequal distribution of that different latent dimensions have on graph reconstruction. the impact. We observe that the vast majority of impact is contained in only a smaller number of latent dimensions. Since equal emphasis is placed on reconstructing the entire latent in the loss function, existing models are optimized to reconstruct information irrelevant to the downstream task. To tackle this problem, we propose Latent Importance Reweighting (\bf LIR) which weights each latent dimensions via it’s importance to final graph reconstruction. Importance is calculated as the relative influence of a single latent unit on graph reconstruction, which is represented via the Jacobian of the decoder relative to the latent. Our method is model-agnostic and requires no additional model parameters or hyperparameters. Experiments show that LIR can boost performance and lead to faster training convergence when applied to two different autoencoders on various benchmark datasets. Code: this https URL
[AI-110] raining Stability under Adaptive Curricula
链接: https://arxiv.org/abs/2611.08787
作者: Johan Obando-Ceron,Walter Mayor,Michael Dennis,Aaron Courville,Pablo Samuel Castro
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Unsupervised Environment Design (UED) approaches train reinforcement learning (RL) agents on diverse curricula that may evolve with their capabilities. While prior work has largely focused on how these curricula are constructed, we study whether their students remain trainable under the resulting distributions. Across multiple UED mechanisms, environment complexities, and extended training horizons, we identify recurring learner-side pathologies, including parameter growth, degraded value–outcome ordering, and loss of representational diversity, that accompany stagnation and performance degradation. These findings connect limitations in UED to optimization failures previously identified in conventional deep RL, while showing how adaptive curricula couple the learner’s optimization state to its future training distribution. Controlled interventions show that mitigating selected pathologies improves learning. Stabilizing the learner improves long-horizon performance and zero-shot generalization without changing the curriculum rule or increasing environment interaction. Our results establish learner trainability as a complementary design principle for UED. Effective curricula require students that remain capable of learning from the distributions they produce.
[AI-111] Aligning Fast and Slow: A Control-Theoretic Oversight Framework for Long-Horizon Trajectories and Maintained Deployments
链接: https://arxiv.org/abs/2611.08781
作者: Stephen Peng,Vineet Jagadeesan Nair,Noah Broestl,Anne Friederike Kleppe,Steven Mills,Adi Zolotov
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:AI agents can pass checks on individual tasks while retaining memory or code changes that alter their later behavior. We introduce a control-theoretic framework that couples a fast loop controlling proposed actions with a slow loop reviewing persistent changes and reassessing configurations after model or tool updates. When measurement errors and changes between checks are bounded and controllers can enforce their decisions, we derive a deployment-level bound and a rule for dividing review effort. In simulation, dividing a fixed review budget between both loops reduces mean maximum departure from intended behavior roughly twelvefold compared with review confined to individual runs. In coding stress tests with authored memories and five matched seeds, combined controls reduce mean severity on a 0-3 scale from 0.325 without gates to 0.117, resolving 33 of 120 issues versus 35 without gates; in trading stress tests, combined controls reduce mean shortfall in low-emission purchases from 0.892 without gates to 0.107. The overall framework helps developers and auditors compare and enforce oversight designs, while the agent studies test component controls rather than the full deployment-level guarantee.
[AI-112] On Lipschitz Explosion in Deep Neural Networks with Normalization: Consequences for Optimization and Adversarial Robustness NEURIPS
链接: https://arxiv.org/abs/2611.08775
作者: Ashkan Soleymani,Reyhaneh Hosseinpourkhoshkbari,Hadi Daneshmand,Patrick Jaillet
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Optimization and Control (math.OC); Machine Learning (stat.ML)
备注: The Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS) 2026
Abstract:Can a network look stable at every layer and still be exponentially unstable as a whole? The Lipschitz constant of a neural network is a basic stability quantity appearing in optimization guarantees, generalization bounds, and robustness certificates. Deep networks naturally admit two such notions, input Lipschitzness, measuring sensitivity to data perturbations, and parameter Lipschitzness, measuring sensitivity to weight perturbations. We prove that, for deep networks with normalization layers, with batch normalization as the canonical example, both Lipschitz constants can grow exponentially with depth and hence with parameter dimension, even under strong per-layer norm control such as |W_\ell|_2\le 1 , and already for linear activations. Thus, a mechanism widely used to stabilize training can create severe worst-case instabilities that are invisible locally, i.e., from layerwise norm bounds alone. Parameter-Lipschitz explosion turns Lipschitz-dependent nonsmooth optimization guarantees into exponential-in-dimension bounds for deep normalized networks, while input-Lipschitz explosion makes worst-case Lipschitz-based generalization and robustness certificates vacuous. The latter also yields a concrete rank-separation mechanism for adversarial vulnerability. The theory predicts that perturbations introducing new singular directions should be amplified much more strongly than perturbations at the same noise level that remain within the input’s original singular subspace. Experiments on MNIST, Fashion-MNIST, and CIFAR-10 support this prediction, showing that rank-creating perturbations cause substantially sharper drops in accuracy, confidence, and margins than same-subspace perturbations.
[AI-113] Conflux: Preventing Privilege Escalation with Principal Contexts and System-Level Evaluation
链接: https://arxiv.org/abs/2611.08768
作者: Raya Buckley,Christian Schroeder de Witt
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 17 pages excluding references and appendix
Abstract:LLM agents combine information from multiple principals while acting externally. We formalise privilege escalation (PE) as the system-level security objective: an action may proceed only if every principal that influenced it would allow it. Model-level defences remain probabilistic, so enforcement must sit outside model judgement. We present Conflux, a principal-context defence. Its kernel, ITES, derives authority from provenance and the organisation’s existing access-control system and intersects the permissions of every influencing principal. We prove a conditional authority-confinement theorem: under complete mediation, no-laundering provenance, a known non-empty decision context, fresh permission checks, and action binding, the kernel admits no action that escalates privilege with respect to any recorded influencing principal. The theorem covers the represented value-influence fragment; control influence from tool, branch, recipient, or plan-step selection remains a refinement obligation. Conflux extends the kernel with independent policy ports, argument-aware authorisation, disclosure, consent, delegation, restriction, and exposure planning. We describe two system- level evaluators. SLED is a black-box evaluator: it exhaustively explores a declared finite model of a defence up to a bound. SLED-V is a white-box evaluator: it checks a serialisable intermediate representation against security and provenance properties. The evidence is bounded and partial. The appendix reports retained AgentDojo, SLED, and SLED-V results, including the limits of those runs. No result in this paper is an unbounded deployment guarantee.
[AI-114] Debate to Deceive: Poisoning LLM s Fact-Checking via Game-Theoretic Multi-Agent Debate
链接: https://arxiv.org/abs/2611.08756
作者: Liting Huang,Huizhi Liang,Yi Lei,Xingjun Ma,Varun Ojha
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Knowledge attacks on Large Language Models, such as manipulating fact-checking judgments, pose significant risks to AI reliability. Existing approaches – adversarial suffix/template injection and single-agent self-reflection – lack robustness across diverse defenses and are prone to confirmation bias and costly trial-and-error, respectively. We propose Debate-to-Deceive (D2D), a multi-agent debate framework grounded in human debate structure and rhetorical style, enabling agents to generate adversarial passages that are semantically coherent, rhetorically effective, and resistant to early confirmation bias. D2D formalizes the adversarial process as an evolutionary game in which agents adaptively evolve their rhetorical strategies (Logos, Pathos, Ethos) through competition. To manage the intractably large LLM output space, we decouple agent behavior into a strategic layer, which selects a rhetorical action from a compact strategy space via evolutionary-inspired learning dynamics, and a generation layer, where the LLM produces a concrete passage conditioned on that strategy and the debate context. On four fact-checking benchmarks (FEVER, HealthVer, CLIMATE-FEVER, and SciFact), D2D substantially outperforms existing attacks and evades most defenses.
[AI-115] Random Symbolic Boosting: A Geometric Approach to Symbolic Regression
链接: https://arxiv.org/abs/2611.08753
作者: José R. Ortegaa,José Manuel Muñoz Contreras,Leonardo Trujillo
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Symbolic models offer interpretability, but their construction through generational evolution can be computationally expensive. This article presents Random Symbolic Boosting (RSB), a method that constructs explicit additive symbolic models via boosting over populations of randomly generated syntax trees, eliminating the need for generational evolution. The idea combines two observations: high-dimensional spaces can be spanned by enough linearly independent vectors, and the semantics of syntax trees can be directly manipulated as such vectors. It is shown that, for Ordinary Least Squares, the boosting process guarantees a monotonic reduction of the training error for both regression and classification. Furthermore, an adaptive geometric early stopping mechanism is introduced. It is based on the angle between the prediction and target vectors, is independent of the problems scale, and requires no adjustments across tasks. Evaluated on a standard benchmark for symbolic regression, the method is located at the Pareto-optimal point between predictive accuracy and training time, ranking first on the real-world datasets. Additionally, it was tested on a large tabular classification benchmark with 200 problems, achieving competitive results.
[AI-116] How are viral trends different from educational and commercial uses of LLM image generation and editing? A privacy-preserving descriptive analysis
链接: https://arxiv.org/abs/2611.08749
作者: Will Tucker-Ray,Marika Swanberg,Alexander Knop,Marvin Chow,Eric Culbertson,Julian Jacobs,Pritish Kamath,Ravi Kumar,Jason Mothersill,Mikhail Pravilov,Eugene Quan,Yurii Sushko
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注:
Abstract:Gemini’s Nano Banana models, first introduced in late 2025, led to a shift in the capabilities and use of image generation and editing. We utilize Urania, a Google-developed framework for generating insights about LLM chatbot interactions with rigorous differential privacy (DP) guarantees, to analyze use cases and intents, techniques and ways of prompting, without ever looking at a single image or individual prompt. We select a subsample of 2.1M conversations in early 2026 which utilized Nano Banana models, into roughly 800 clusters in five categories: (1) Social Viral Sharing, (2) Brand Marketing Collateral, (3) Educational Technical Information Media, (4) Product Visualizations E-commerce, and (5) Architectural Spatial Designs. We then compare these clusters on 21 features of use, measuring aspects of prompts’ (a) editing techniques, (b) artistic style requests, © composition, and (d) chatbot interactions. This allows us to describe a baseline of how people are using image generation and editing tools for different use cases—(including growing and emergent use cases which were not possible before Nano Banana)—in a privacy-preserving manner.
[AI-117] AlterNet: Adaptive-Depth Mixture-of-Experts with Learned Halting for Elastic Inference
链接: https://arxiv.org/abs/2611.08741
作者: Alexander Kastius
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Most LLM architectures have fixed compute demands determined at training time. We propose a novel architecture that combines the Mixture-of-Experts architecture with shallow experts and a novel dynamic-depth system to make the extent of a forward pass dynamically adjustable. Our approach lets the model automatically determine the compute needed for a given task, paired with a system to scale it down after training without retraining when the execution environment demands faster or cheaper inference. To achieve this, we introduce an iterative computation mechanism, sequence-level depth-based routing steered by a halting mechanism within the experts, a depth encoding that lets the model track its progress in the chain, and a regularization mechanism that incentivizes the model to use allocated compute efficiently. We present initial benchmarks and discuss future research directions.
[AI-118] Remembering Is Not Enough: Evaluating Timing and Omission in Longitudinal AI Mental Health Support
链接: https://arxiv.org/abs/2611.08721
作者: Eli Min,Mohit Chandra,Aamogh Sawant,Mengting Tu,Freya Lin,Ridwan Shams Khan,Reigna El-Yashruti,Christopher L. Smith,Mark Kalinich,John Torous,Munmun De Choudhury
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注:
Abstract:Evaluations of large language models (LLMs) for mental health support commonly focus on the clinical appropriateness of individual responses, potentially overlooking failures that emerge across extended interactions. As AI assistants accumulate memory of users’ circumstances and prior disclosures, this raises an accountability question: Do AI chatbots provide appropriate support as users’ needs evolve? We investigate by analyzing 14,256 longitudinal simulations grounded in donated chatbot histories from 22 users, examining how a memory-equipped assistant responds to suicidal ideation, self-harm, and anxiety crises across recovery trajectories and communication styles. We introduce the Context-Aware and Readiness-Evaluative (CARE) framework, developed with mental health professionals, to assess the timing and quality of advice, including omissions. Advice was less aligned with users’ readiness after a crisis than in matched control stretches of the same conversation, largely because of premature recommendations, which were most common when recovery stalled. Revisiting a crisis reduced premature advice but not lagging advice, and users’ communication style changed how often and when the assistant gave recommendations. Although explicitly unproductive advice was rare, the assistant frequently omitted clinician-expected support such as follow-up and planning for next steps. These findings show the limits of single response-level safety evaluation and the need to assess timing, omission, and consistency across users when holding AI systems accountable for longitudinal support.
[AI-119] Memory-Aware Multi-Agent Architecture with LLM -based Orchestration for Conversational Recommendation
链接: https://arxiv.org/abs/2611.08718
作者: Emilly Pitman de Castro,Vanessa Silva,Gustavo Kanno,Marcus Santos,Bruno Ferraz do Amaral,Marcelo Nogueira,Lucas Francisco Pellicer,Lucas Lanaro,Pedro Augusto Luiz,João Renato Ribeiro Manesco,Giovanni Carbinatti,Isabela de Lima Santos,Lucas Oliveira Jeronymo,Jeison Fappi,Raul Rotilli Aguirre,Fabio André Frantz,Renata De Paris,Fabiana Marucci,Lohan Silva,Vinicius Diniz Reolon,Wandemberg Gibaut
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Conversational recommender systems often treat user sessions in isolation, limiting their ability to capture how user intent evolves over time. While large language models (LLMs) have improved semantic understanding, they are typically used as isolated components, without addressing how memory, retrieval, and decision-making should be integrated at the system level. In this work, we propose a memory-aware multi-agent architecture for conversational recommendation, in which LLMs act as orchestration layers that coordinate intent refinement, query construction, and interaction flow. A key design principle is the integration of long-term memory as a first-class component, influencing both routing decisions and semantic query formation. We evaluate the system through cross-session memory experiments, multi-turn conversational simulations, and supervisor-driven interaction studies on the Amazon Reviews dataset. Results show that memory-enabled interactions recover up to 63% of the gap toward explicit-query scenarios and prevent semantic degradation in extended dialogues. In addition, supervisory orchestration increases user acceptance by +28.9 percentage points, highlighting the importance of structured orchestration in guiding interaction and output quality. We also observe an exposure bottleneck between retrieved candidates and displayed recommendations. While LLM-reranking does not improve Hit@K, it consistently enhances semantic consistency and diversity. To capture these effects, we introduce the Semantic Consistency Score (SCS), a complementary evaluation measure. Overall, the results emphasize the importance of system-level design choices, suggesting that effective conversational recommendation benefits from explicit memory integration, orchestration, and evaluation beyond traditional ranking metrics.
[AI-120] An Educated Guess: Deriving Statistically Aligned Gradient Estimates for Zeroth-Order Optimisation
链接: https://arxiv.org/abs/2611.08712
作者: Justus Huebotter,Serge Thill,Marcel van Gerven,Nasir Ahmad
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 39 pages, 15 figures
Abstract:As novel hardware architectures are explored for the purposes of machine learning, including neuromorphic chips, asynchronous systems, and ASICs, effective learning algorithms are highly desirable. While backpropagation is the standard for differentiation on standard hardware, its requirement for global synchrony and exact differentiability limits the exploration of exotic algorithms on non-traditional substrates and devices. Zeroth-order (ZO) optimisation methods, such as SPSA and weight perturbation, offer a compelling alternative by estimating gradients through inference-only passes; however, these methods have historically suffered due to the `curse of dimensionality’, where random perturbations become increasingly misaligned with the true gradient in high-dimensional spaces. In this work, we move beyond blind random noise by deriving zeroth-order search directions from activation geometry available during the forward pass. This leads to a family of Spectral Zeroth-Order methods, SZO- \kappa , whose directions solve local response-based gradient-guessing objectives under explicit stochastic assumptions. These methods explore feature-supported directions that are more likely to produce informative loss responses. Using multilayer and convolutional networks, we analyse gradient alignment, bias-variance structure, and training stability of these methods relative to standard perturbation baselines and backpropagation.
[AI-121] FoleyForge: Free-form Controllable Rectified-flow Editing of Sound Effects
链接: https://arxiv.org/abs/2611.08695
作者: Giovana Morais,Hugo Flores García,Prem Seetharaman,Johannes Imort,Justin Salamon,Magdalena Fuentes,Oriol Nieto
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注:
Abstract:We present FoleyForge, a generative model for instruction-based editing of sound effects. Given an input clip and a free-form instruction describing the desired change, it applies the edit while preserving the input’s identity. A single model handles semantic, stylistic, and DSP-like changes. We introduce a data pipeline that guides an audio language model to author both DSP-based acoustic edits and similarity-preserving semantic edits, thus addressing the lack of publicly available audio editing data. Such data is then used to fine-tune a rectified-flow audio generation model, which we extend to expose attribute controls for loudness change, edit strength, and an input-derived contour envelope, each tunable via classifier-free guidance, yielding a more expressive editing model. FoleyForge achieves state-of-the-art instruction adherence at competitive audio quality and source faithfulness.
[AI-122] Cancellation Is Not Enough: What Retries Leave Behind in an Overloaded LLM Engine
链接: https://arxiv.org/abs/2611.08684
作者: Xindan Zhang,Fei Wang
类目: Artificial Intelligence (cs.AI)
备注: 13 pages, 6 figures, 3 tables
Abstract:Clients and proxies in front of an LLM engine retry calls that time out. After a transient overload, these retries can keep the engine from recovering even when calls arrive below its capacity. Since the engine may keep computing for clients that have left, propagating cancellation looks like the natural remedy. In this paper, we test it on vLLM with overload experiments on three open models, and with scheduler traces of one that follow each client attempt to the execution it creates in the engine. We find that with retries on, cancellation alone did not restore service in any run, and only 12-38% of the calls that arrived after the overload were delivered. The traces suggest why: retries that vLLM ran typically started prefill less than half a second before they timed out. Executions that vLLM later aborted took 79-85% of the prompt tokens it computed, and vLLM’s prompt-token counter missed a third to half of all computed prompt tokens. A retry budget restored service by shedding retries but left many of the overload’s calls undelivered. Request coalescing delivered every call from the overload onward, some late, unless cancellation had already aborted the running execution. Overload controls should thus be judged by the executions that retries add and the calls that clients receive.
[AI-123] ransporting the Past: An Optimal Transport View of Backtracking Counterfactuals
链接: https://arxiv.org/abs/2611.08682
作者: Zhangyi Liu,Pouria Fatemi,Suvrit Sra
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:We study backtracking counterfactual inference for Structural Causal Models (SCMs) through the lens of optimal transport. In SCMs, a factual observation generally induces a posterior distribution over exogenous variables, making counterfactual inference inherently distributional. We introduce two complementary frameworks that transport this posterior onto the set of exogenous configurations satisfying a desired counterfactual query. The first constructs a natural pointwise projection and is suitable when the transported exogenous distribution can be used directly. The second enforces mutual independence among counterfactual exogenous variables, so that the transported variables remain interpretable as genuine exogenous noises of the same SCM. We show that both frameworks are consistent with the original backtracking counterfactual principle, and we propose marginal, penalty-based, and hard-constrained solvers for finite-posterior settings. Synthetic experiments on non-bijective SCMs illustrate the geometry of the problem and reveal trade-offs among transport cost, counterfactual constraint satisfaction, and preservation of exogenous independence.
[AI-124] Pdb or Not to Be: Comparing Debugging Interfaces With Pass_X@k NEURIPS2026
链接: https://arxiv.org/abs/2611.08666
作者: Ryan Magnuson,Ellie Kitanidis,Ariel Lee
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: Accepted, NeurIPS 2026 workshop “Who Verifies the Agents? Toward Reliable Agent Development”
Abstract:Interactive debuggers are indispensable to human software development, yet coding agents overwhelmingly rely on edit-execute-observe loops, using common bash tools, test scripts, and print statements to debug via repeated shell commands. Stepping through a debugger provides live access to runtime state, which may enable different problem-solving capabilities. However, prior work typically introduces the debugger as an additional tool or through a new subagent, making its individual contribution difficult to isolate from the associated scaffold changes. To address this, we compare a debugger-only agent against a bash-only agent on all 500 SWE-bench Verified instances, holding the rest of the scaffold fixed. We evaluate both architectures on two models, GPT-OSS-120B and Nemotron-3-Ultra-550B, with three independent runs per architecture \times model configuration. Crucially, to distinguish interface specialization from expanded coverage under repeated sampling, we introduce pass _X @ k , a cross-agent pass@ k that provides an unbiased estimator of the probability that at least one of k agents is successful. Although the bash and debugger agents solve partially disjoint subsets of the benchmark in any single pair of runs, our pass _X @ 2 estimator reveals that this is largely due to run-to-run variance rather than tool complementarity. Across both models, we find that two bash rollouts outperform the combination of one bash and one debugger rollout. Our findings highlight the importance of accounting for stochastic variance in evaluating agent and tool diversity, and we hope that pass _X @ k will serve as a useful framework for the community.
[AI-125] IWIBench: Evaluating Gradient-Free Model Driven Intra-weight Injection
链接: https://arxiv.org/abs/2611.08647
作者: Chris von Csefalvay,Pranav Doma,Achyuthan Unni Krishnan,Tamás Földi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Intra-weight injection lets one model change another model’s behaviour by editing its weights directly, without gradients. Such edits can range from altering lexical associations to enforcing output formats or updating factual knowledge. Despite growing interest in these approaches, there has been no systematic way to evaluate how well models perform them. We introduce IWIBench, which measures a large editor model’s ability to modify the weights of a smaller target model (typically Gemma 3 4B IT) to achieve one of ten classical text-editing objectives while preserving its other capabilities. Editors are limited to four tools for reading and writing weights, with bounded tensor access and a fixed budget for generating and inspecting rollouts. A validator applies each proposed patch in a clean environment and compares the edited model against the unmodified target. Evaluating frontier closed- and open-weight models reveals sharply diverging capabilities, both across models and across tasks. A small number of models, most notably Astra and GPT-6.1 Sol, perform far ahead of the rest. The recency of this ability (little of it is evident in GPT-5.6 Sol, for instance), together with qualitative analysis, suggests it has been acquired through recent training. Tasks also vary widely: structured response resisted every attempt, and answer-length control yielded only one successful solution, whereas alias acquisition and recoding proved most amenable. Success in one task family correlated poorly with success in others, indicating an early and uneven capability frontier. We discuss strategies that editors repeatedly adopted and conclude with an outlook on intra-weight injection as an emerging but already potent approach to gradient-free behaviour modification.
[AI-126] Decision Context: Persistent Behavioral Learning for AI Agents NEURIPS2026
链接: https://arxiv.org/abs/2611.08644
作者: Nima Vahdat,Pashootan Vaezipoor
类目: Artificial Intelligence (cs.AI)
备注: 12 pages. Accepted at the NeurIPS 2026 Workshop on Personalized, Aligned, Long-Term Memory for AI Systems (PALM)
Abstract:LLM agents do not learn from corrections across sessions. A user fixes the same mistake at the start of every conversation. Existing agent memory systems primarily handle facts, such as preferences, entities, and conversational history, rather than procedural knowledge. We present Forge Decision Context (forge_dc), a system that captures behavioral signals from human corrections and self-assessed tool-call failures, compiles them into imperative RULE directives, and injects them at the decision points where they apply. We evaluate forge_dc alongside four predecessor approaches on DC-Bench, a new benchmark of 71 scenarios across nine behavioral dimensions, built from real executed agent traces in a deterministic mocked-tool environment. forge_dc reaches 95% compliance, up from 55% with naive retrieval, while injecting approximately 10 times fewer tokens per query. Across the five-stage progression, the strongest systems represent behavioral memory as explicit directives and deliver it at the decision points where it applies. The results identify representation, activation, and placement as central design choices for persistent behavioral learning. Comments: 12 pages. Accepted at the NeurIPS 2026 Workshop on Personalized, Aligned, Long-Term Memory for AI Systems (PALM) Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.08644 [cs.AI] (or arXiv:2611.08644v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.08644 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-127] Aspect-wise Pairwise Tournaments for Test-Time Selection
链接: https://arxiv.org/abs/2611.08640
作者: Hannah Guan,Yilun Du
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Scaling test-time compute has proven highly effective at improving LLM reasoning. A popular method is best-of-n sampling, where a model generates many candidate solutions and one is selected. This motivates the question: how can we select the best candidate? When humans make choices, we find it more natural to compare two options than judge each in isolation. Additionally, when the options become complex, we like to judge the options one aspect at a time, such as comparing two job candidates on experience, communication, etc. In the test-time setting, candidate solutions should be judged in the same way. Here we introduce Aspect-wise Pairwise Tournaments (APT), which verifies by comparison and compares by aspect. Each matchup of a single elimination tournament is decided by a vote over multiple off-the-shelf aspect verifiers, each expressing a pairwise preference along its own aspect. Across five domains and three generator LLMs, APT improves over binary multi-aspect approval, self-consistency and reward model verification, with the largest gains on open-ended tasks. We find that both components matter; neither binary approval over the same aspects nor a generic pairwise judge at a matched budget matches APT. Because aspects are specified in natural language, APT also lets users steer selection toward properties they care about, better aligning with user preferences without needing to regenerate solutions. No component is trained, so APT transfers to new domains by simply rewriting prompts.
[AI-128] LLM -Based Feature Extraction for Multi-Criteria Inventory Classification
链接: https://arxiv.org/abs/2611.08626
作者: Humza Akhtar,Rami Pinto
类目: Artificial Intelligence (cs.AI); Databases (cs.DB); Emerging Technologies (cs.ET)
备注:
Abstract:E-commerce supply chains generate large volumes of customer feedback and operational data that remain underutilized in inventory management. Traditional inventory classification techniques such as ABC analysis rely primarily on annual dollar usage, while multi-criteria inventory classification (MCIC) methods extend this but still depend on structured quantitative data and do not utilize unstructured information available in modern supply chains. This paper proposes an LLM-based feature extraction approach that transforms unstructured sources such as customer reviews into quantitative features using vector embeddings and LLM-generated evaluation criteria. The enriched feature set is incorporated into an MCIC framework based on the Hadi-Vencheh model. Experimental results show measurable reclassification effects when customer sentiment signals are included, demonstrating that integrating unstructured data into MCIC can enhance inventory prioritization.
[AI-129] NormGym: Evaluating Normative Competence through Social Interaction
链接: https://arxiv.org/abs/2611.08613
作者: Binze Li,Zihuan Zhang,Seokhyun Baek,Gillian K. Hadfield,Tianmin Shu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Normative competence, the ability to infer socially enforced expectations from interaction and adjust behavior accordingly, is essential for agents operating in unfamiliar human communities. Yet existing evaluations largely test whether agents can recognize or apply norms they may already know, rather than whether they can acquire previously unknown norms from social interaction. We introduce NormGym, an interactive benchmark for normative competence in which a newcomer must infer a hidden, arbitrary norm while pursuing its own task in a multi-agent community. Success requires distinguishing what is possible from what is permitted, inferring the norm from observed behavior and sanctions, and applying it in new situations. NormGym contains diverse norms and distinct environments, allowing systematic evaluation of current agents, including large language model (LLM) agents. We find that normative competence remains challenging: the strongest evaluated LLM achieves a joint task-and-compliance success rate of 55.3%. Our analysis reveals that models can sometimes revise their behavior following sanctions for their own actions, but struggle to use other agents’ behavior and the sanctions they receive to infer the underlying norm. We further find that fine-tuning models may be insufficient to improve normative competence. These results suggest that current models lack a robust, transferable capability for acquiring normative knowledge through social interaction.
[AI-130] Q-OIL: Optimistic Exploration in Learning from Human Interventions
链接: https://arxiv.org/abs/2611.08576
作者: Pranav Teegavarapu,Sriyash Poddar,Rico Qi,Cleah Winston,Krisha Khandelwal,Aykut Onyol,Natasha Jaques,Abhishek Gupta
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: Project website: this https URL
Abstract:As robots are deployed in the real world, they will inevitably make mistakes. Human interventions provide valuable feedback for improving their behavior, enabling them to adapt online to new scenarios. However, studying human-in-the-loop robot learning is difficult: real-world experiments are slow and expensive, so most algorithm development relies on simulated human models. We show that common human modeling assumptions do not match how real people intervene in practice. Through analyzing intervention data from human participants, we find that humans tend to intervene when agent progress stalls, and that their corrective actions are often suboptimal. Motivated by these findings, we introduce a progress-based model of human intervention that achieves higher rank correlation with real-user algorithm rankings, enabling realistic algorithm benchmarking and development in simulation. Building on these insights, we propose Q-OIL, a method that adds optimistic bonuses to intervention transitions in reinforcement learning and softly guides exploration toward human corrections. We find that decoupled critics are critical for stability because they prevent the bonuses from backing up to pre-intervention states. As a result, Q-OIL learns more efficiently than competitive baselines, achieving 10-50% higher success in simulation and on real robots, and improving pretrained VLAs on two bimanual manipulation tasks. As we ground both evaluation and algorithms in realistic human behavior, this work provides a practical path toward scalable robot learning from interventions.
[AI-131] Enhancing Academic Advisement with Case-Based Reasoning
链接: https://arxiv.org/abs/2611.08572
作者: Tanat Tonguthaisri
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 2026 proofread digital edition of the August 2003 Master of Engineering thesis, RMIT University
Abstract:This research involved the study of an application of Case-Based Reasoning (CBR) in the field of academic advising. Academic advisement has been recognised as an important function of higher education. However, there is a widespread desire to improve the advisement services provided by many educational institutions (e.g., colleges and universities) around the world. In the last ten years, CBR expert systems have made important contributions in many domains by applying CBR methodology to a variety of areas of human problem solving. However, to date, expert advisement systems have not been successfully applied to academic advisement, which seems to be a very appropriate domain for this technology. Academic advising is an important and time consuming task, and different tools and techniques must be used to make it an effective and efficient process. This thesis describes the design, development, and validation of a prototype case-based academic advisement recommender system that supplements the conventional advising process, and could eventually evolve to replace human academic advisors when its full potential is realised. Comments: 2026 proofread digital edition of the August 2003 Master of Engineering thesis, RMIT University Subjects: Artificial Intelligence (cs.AI); Computers and Society (cs.CY) ACMclasses: I.2.1; K.3.1 Cite as: arXiv:2611.08572 [cs.AI] (or arXiv:2611.08572v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.08572 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.25439/rmt.34846194 Focus to learn more DOI(s) linking to related resources
[AI-132] PINNeval: A Comprehensive Evaluation Standard for Physics-Informed Neural Networks NEURIPS2026
链接: https://arxiv.org/abs/2611.08559
作者: Kevin von Bargen,Hamid Ebrahimy,Tim Römer
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026), Track on Evaluations and Datasets. 54 pages. Code: this https URL
Abstract:The Physics-Informed Neural Network (PINN) literature has produced a rapidly expanding catalogue of methods. We divide a PINN training pipeline systematically into four components and observe that many existing benchmarks study these components in isolation. These components are domain-importance adaptation, loss balancing, neural network architecture, and optimization scheme. Moreover, the studies set the surrounding and not considered components of the pipeline to a minimal default that is no longer representative of current best practice approaches. Conclusions drawn under such conditions may not transfer to competitive pipelines. In this work, we propose a new evaluation protocol that (1) treats the four components as an integrated system; (2) crucially requires competitive implementations of the components not under investigation; (3) specifies the metrics and reporting format; and (4) is accompanied by structured documentation that ensures the traceability of evaluations across studies. Based on this comprehensive protocol, we release PINNpipe, an innovative JAX package containing 18 different methods from the current literature as well as nine efficiently implemented benchmark problems. Using PINNpipe, we illustrate our evaluation protocol in relation to established benchmark studies and report results that challenge common assumptions.
[AI-133] SAFTAC: Simulation-Augmented Fine-Tuning of Open-Source LLM s for Analog Circuit Design NEURIPS2026
链接: https://arxiv.org/abs/2611.08543
作者: Junsheng Huang,Yifan Sun,Zhuoer Zhang,Ning Wei,Yening Liu,Michael Molter,Pavan Kumar Hanumolu,Elyse Rosenbaum,Bin Hu,Huan Zhang
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI)
备注: Accepted to NeurIPS 2026
Abstract:Circuit design for analog integrated circuits (ICs) is challenging because circuit topology, device sizing, and circuit performance are tightly coupled. Existing large language model (LLM)-based methods either rely on inference-time correction around frozen models or train on limited circuit types without directly using simulation outcomes as training signals. A key bottleneck is that existing datasets do not provide sufficient task information for training LLMs on end-to-end specification-conditioned design. To fill this gap, we first construct a large-scale simulation-grounded dataset with 8,626 design tasks across 10 different types of analog circuits, where each task includes a textual description, labeled I/O ports, loading condition, target specifications, a testbench, and a reference netlist. Using this dataset, we present SAFTAC, a simulation-augmented fine-tuning framework that trains LLMs to generate and revise analog circuit design. SAFTAC combines three-stage supervised fine-tuning, which progressively builds the capabilities required for analog circuit design and feedback-guided correction, with simulation feedback-augmented reinforcement fine-tuning, which further optimizes models using simulation-grounded rewards. We evaluate SAFTAC under single-pass generation and two-pass generation with simulation feedback, showing improvements over the strongest closed-source LLM by 15.9% and 7.8%, respectively.
[AI-134] MorphSched: Workload-Aware Adaptive Scheduler Generation for Linux
链接: https://arxiv.org/abs/2611.08509
作者: Yifan Zhang,Hengxuan Lou,Xinkui Zhao,Rengrong Xiong,Naibo Wang,Chang Liu,Guanjie Cheng,Shuiguang Deng,Jianwei Yin
类目: Operating Systems (cs.OS); Artificial Intelligence (cs.AI)
备注:
Abstract:Modern Linux scheduling increasingly requires workload-specific policies to meet the dynamic performance needs of heterogeneous applications. The recent sched_ext framework enables custom BPF schedulers, but building practical policies still requires coordinating Rust user-space control logic, BPF scheduling callbacks, generated user-kernel interfaces, and strict eBPF verifier constraints. This paper presents MorphSched, an LLM-driven system for synthesizing complete Rust+BPF sched_ext schedulers. MorphSched profiles target workloads under Linux EEVDF to construct dynamic, workload-specific optimization objectives, and uses a scaffold-constrained generation pipeline to translate scheduling intent into deployable scheduler implementations. To improve synthesis reliability, MorphSched separates implementation repair from strategy search: verifier, build, and runtime failures are handled through a DAG-structured repair process, while performance feedback is accumulated in an evolutionary memory pool to refine high-level scheduling strategies across iterations. We evaluate MorphSched across 120 machine/workload/interferer configurations on a real Linux kernel. The generated schedulers compile, pass eBPF verification, load successfully, and remain stable throughout our benchmark runs without manual edits to generated scheduler source code. Across these configurations, MorphSched improves the primary foreground metric over EEVDF by 26.5% on average and up to 87.3%. In selected comparisons, the generated schedulers are also competitive with representative hand-written sched_ext schedulers. These results show that scaffold-constrained, system-grounded LLM generation can be a practical tool for exploring workload-specific Linux scheduling policies.
[AI-135] MeshRT: Compile-Time Governed Wafer-Scale Runtime for Low-Latency High-Throughput Inference SOSP2026
链接: https://arxiv.org/abs/2611.08508
作者: Congjie He,Le Xu,Zhan Lu,Yeqi Huang,Haocheng Xiao,Cheng Deng,Lingxiao Ma,Ziming Miao,Fan Yang,Luo Mai
类目: Operating Systems (cs.OS); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: Accepted to the 32nd ACM Symposium on Operating Systems Principles (SOSP 2026)
Abstract:Wafer-scale accelerators promise ultra-low-latency AI inference, but current system stacks still carry an unsustainably high cost premium. The reason is that many current and emerging inference techniques, such as batching and MoE, introduce runtime dynamism that existing wafer-scale systems cannot support efficiently. As a result, they must choose between two unsatisfactory options: accept suboptimal latency, or preserve low latency by restricting dynamic techniques, thereby limiting the models they can support and reducing aggregate throughput. This tradeoff significantly increases the cost of low-latency serving. We present MeshRT, the first wafer-scale system to achieve both ultra-low latency and high throughput for LLM inference. MeshRT achieves this via a novel system architecture, which we call compile-time governance of runtime dynamism. This architecture features new system abstractions and mechanisms that enable MeshRT to compile dynamic events arising from LLM inference directly into per-core schedules, enabling asynchronous and decentralized execution with low runtime overhead. Implemented on a commodity wafer-scale accelerator, MeshRT preserves ultra-low latency while improving throughput by up to 56 \times over the state-of-the-art system for wafer-scale accelerators. Against measured single-instance SGLang results on H200, MeshRT achieves 4.5-7.0 \times lower latency. Under ideal linear scaling of those GPU measurements to an approximately power-matched H200 cluster, MeshRT achieves 1.7-13.0 \times higher decode system throughput in the low-latency regime. Comments: Accepted to the 32nd ACM Symposium on Operating Systems Principles (SOSP 2026) Subjects: Operating Systems (cs.OS); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2611.08508 [cs.OS] (or arXiv:2611.08508v1 [cs.OS] for this version) https://doi.org/10.48550/arXiv.2611.08508 Focus to learn more arXiv-issued DOI via DataCite Related DOI: https://doi.org/10.1145/3830418.3843856 Focus to learn more DOI(s) linking to related resources
[AI-136] Machine Learning Beyond Statistical Consistency for Hybrid AI
链接: https://arxiv.org/abs/2611.01914
作者: Jinsook Kim,Jinho Kang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO); Machine Learning (stat.ML)
备注:
Abstract:We propose a new definition of machine learning, according to which a machine learns a function when it computes the function successfully, that is, when the value it returns for each input has a proper derivation in a formal system. We extend Gödel’s Representability Theorem, which equates computation with derivation, to empirical targets in two ways: the theorem is extended to the computable reals, and the data are admitted into the system as empirical definitions that record what has been observed. We prove that a machine learns a true probability if and only if two conditions are met. The Truth Condition is that its estimate converges to the true probability, which is statistical consistency. The Assurance Condition is that the machine properly derives its own convergence inside the system. Learning thus lies beyond statistical consistency, for it requires a derivation of the convergence from the record. A global assumption about the data, such as the i.i.d. assumption, does not supply such a derivation, because it leads only to a circular one, as we show for PAC learning. On the other hand, the two conditions are met for the probability of a sentence over a finite ideal corpus. A real machine can approximate a true probability only if an ideal machine can learn it precisely, that is, only in a domain that meets the two conditions. Modern AI works in such a domain, and we offer this as a clue to its success. Modern AI is hybrid as well, in two senses: First, it performs two kinds of tasks, namely probabilistic ones such as predicting the next word and deductive ones such as proving and coding. Second, it combines two parts of different kinds, a statistical proposer and a deductive verifier, for proving and coding. Our framework explains both senses of hybridity in one formal system.
[AI-137] Fast and General Automatic Differentiation for Finite-State Methods
链接: https://arxiv.org/abs/2602.12300
作者: Lucas Ondel Yang(LISN, CNRS),Caio Corro(INSA Rennes, IRISA, ARCHIMEDIA),Nicolas Denier(LISN),Martin Kocour(FIT / BUT, BUT),Tina Raissi(RWTH Aachen),Pablo Riera(ICC)
类目: Formal Languages and Automata Theory (cs.FL); Artificial Intelligence (cs.AI)
备注:
Abstract:We propose a new method, that we coined the ``morphism-trick’‘, to integrate custom implementations of vector-Jacobian products in automatic differentiation softwares, applicable to a wide range of semiring-based computations. Our approach leads to efficient and semiring-agnostic implementations of the backward pass of dynamic programming algorithms. For the particular case of finite-state methods, we introduce an algorithm that computes and differentiates the \oplus -sum of all paths’ weight of a finite-state automaton. Results show that, with minimal effort from the user, our novel library allows computing the gradient of a function w.r.t. to the weights of a finite state automaton orders of magnitude faster than state-of-the-art automatic differentiation systems. Implementations are made available via an open-source library distributed under a permissive license.
[AI-138] Discovering formally verified quantum programs with AI
链接: https://arxiv.org/abs/2611.09822
作者: Sergey Shirobokov,Ben Jaderberg,Katerina Hristova,George Tsoukalas,Johannes Bausch,Eric Wieser,Francisco J. H. Heras,Nicholas C. Rubin,Guang Hao Low,Sébastien Racanière,Nenad Tomašev,Lei M. Zhang,James Watson,Tanuj Khattar,Aws Albarghouthi,Stavan Jain,Harun Khan,Leopoldo Sarra,Andrew W. Senior,Michael Broughton,Hussain Masoom,Sergio Boixo,Ryan Babbush,Swarat Chaudhuri,Pushmeet Kohli
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Programming Languages (cs.PL)
备注:
Abstract:The timeline by which quantum computers can solve breakthrough problems in science depends on improved quantum algorithms as much as on better hardware. Such algorithmic advances are typically achieved through many years of expert human collaboration. In mathematics and computer science, large language models have accelerated progress through automated discovery loops. However, quantum algorithms can only be evaluated by classical computers on restricted or small instances. Here we demonstrate automated discovery of quantum programs beyond these regimes including order-of-magnitude improvements on scientific applications. To achieve this, we build a quantum program library in the Lean theorem prover in which large language model agents synthesize novel programs against formal problem specifications and verify their correctness and cost. Agent teams collaborate, sharing results and knowledge through wikis and discussing ideas at organized conferences. Our results improve on prior baselines for parallel search (up to 1.4x), quantum arithmetic (290x), and simulations relevant to fusion, photochemistry and drug-metabolism by 11x, 8x and 2.5x respectively. Combining human and artificial intelligence at scale, under the rigor of mathematical proof, lets researchers design algorithms that today’s classical and quantum computers cannot test, shortening the path to practical quantum computation.
[AI-139] SmartBloch-Agent : An Agent ic Framework for Automated MRI Sequence Design and Optimization
链接: https://arxiv.org/abs/2611.09413
作者: Jiakai Yang,Haoye Zheng,Yixiao Yue,Qinqin Yang,Zhong Chen,Shuhui Cai,Congbo Cai
类目: Medical Physics (physics.med-ph); Artificial Intelligence (cs.AI); Image and Video Processing (eess.IV)
备注:
Abstract:Magnetic resonance imaging (MRI) pulse sequences organize radiofrequency excitation, gradient encoding, and signal acquisition, determining magnetization evolution and image formation. Sequence research requires translating imaging objectives into executable protocols meeting physical and acquisition constraints, refined via physical simulation, signal reconstruction, and scientific evaluation. Sequence construction, physical verification, parameter search, and assessment involve many tools and manual decisions, limiting automation from objective to verifiable sequence. We develop SmartBloch-Agent, an integrated framework for automated MRI sequence design and optimization. Given objectives and scientific constraints in natural language, an AI agent interprets goals, plans tasks, orchestrates tools, and makes feedback-informed decisions, drawing on a construction module and optimizers for structural design and parameter search. Our GPU-accelerated SmartBloch engine supplies physical computation for both stages, so the framework configures, per sequence and objective, event validation, Full-Bloch magnetization evolution, signal acquisition, image reconstruction, and scientific evaluation, and closes the loop without external MRI simulators. Results return to the agent to guide sequence revision, parameter optimization, and candidate selection, and the design files carry simulation, reconstruction, and evaluation evidence, forming an output package for a real-scanner workflow. Experiments show the framework automatically designs and optimizes MRI sequences for varied objectives and constraints, refines candidates with physical feedback, and improves several parameter-estimation metrics and downstream quantitative reconstruction. SmartBloch-Agent unifies objectives, design, physical execution, and evaluation into a continuous workflow for end-to-end automation of MRI sequence research.
[AI-140] Artificial Shoppers and Attention Competition
链接: https://arxiv.org/abs/2611.09399
作者: Xiaomeng Guo,Fasheng Xu,Xiaoli Zhang
类目: Theoretical Economics (econ.TH); Artificial Intelligence (cs.AI); Computer Science and Game Theory (cs.GT)
备注:
Abstract:We study how AI shopping interfaces affect price competition, platform profit, and consumer welfare. A platform hosts two of three products of different quality. Without AI, consumers compare the hosted pair under salience distortion and overweight whichever attribute, price or quality, stands out. Under assistance, an AI assistant recommends the two products with the highest true utility, leaving consumers to make the same distorted comparison. Under delegation, an AI agent selects the best product, subject to purchase approval. Both modes access all products and know consumers’ preferences; firms price strategically. First, assistance can soften competition despite expanded product access. The middle-quality firm may forgo low-valuation consumers because attracting them requires cutting the price paid by higher-valuation consumers as well. Relative to no AI with the same assortment, assistance lowers both hosted prices under the high- and low-quality pair but can raise both under the high- and middle-quality pair. Second, we identify an alignment paradox: some consumers can become worse off than without AI even though their best option remains recommended and becomes cheaper. Salience can induce a switch to a lower-quality product whose quality loss exceeds the saving relative to the original purchase. The assistant ranks products accurately but does not optimize the shortlist for consumers’ realized welfare. Third, assistance can raise both platform profit and aggregate consumer surplus with the high- and middle-quality pair; delegation can do so with the high- and low-quality pair. Fourth, among AI designs, the platform weakly prefers delegation with the high- and middle-quality pair wherever all compared pricing equilibria exist. Yet this preferred design can yield lower platform profit and aggregate consumer surplus than the no-AI optimum.
[AI-141] FoundaRF: A Physics-Informed Foundation Model Framework for Radio-Frequency Passive Components
链接: https://arxiv.org/abs/2611.09003
作者: Houbo He,Chris Han,Hyunsu Chae,Lei Xia,Qiang Zhou,Lawrence Liu,Sensen Li,David Z. Pan,Taiyun Chi
类目: ignal Processing (eess.SP); Artificial Intelligence (cs.AI)
备注: Accepted for publication in the proceedings of the 2026 IEEE/ACM International Conference on Computer-Aided Design (ICCAD 2026)
Abstract:Designing RF/mmWave passive components is a time-consuming process that relies heavily on expert knowledge and iterative EM simulations. Although recent ML surrogate models have shown great promise in accelerating passive design, they remain task-specific and lack generalization. This paper presents FoundaRF, a foundation model framework that unifies passive layout modeling across diverse layout structures and technologies. It combines a self-supervised encoder with adaptive layout processing and physics-aware decoders to enable few-shot transfer. Experiments across four semiconductor technologies show strong generalization to unseen structures, outperforming baselines with up to a 3.3x reduction in prediction error, reducing labeled-data requirements by up to fourfold, and substantially shortening design turnaround. FoundaRF establishes a scalable and data-efficient solution for RF/mmWave passive design automation.
机器学习
[LG-0] On the (non-)closedness of the set of ReLU networks
链接: https://arxiv.org/abs/2611.09887
作者: Eloi Martinet,Leon Bungert
类目: Machine Learning (cs.LG); Classical Analysis and ODEs (math.CA); Functional Analysis (math.FA); Numerical Analysis (math.NA); Optimization and Control (math.OC)
*备注: 27 pages, 6 figures
Abstract:We prove that the set \mathcalN_N of ReLU networks with one hidden layer and at most N\in\mathbb N neurons is weakly closed in W^1,p(\Omega) , for a bounded connected open set \Omega\subset\mathbb R^d . Consequently, every coercive, weakly lower semicontinuous energy on this space attains its minimum over \mathcalN_N . This covers the Deep Ritz method with convex integrands, Sobolev training and Lipschitz learning. Moreover, as the weight decay parameter tends to zero, minimizers of the regularized problem converge, up to subsequences, to minimizers with minimal parameter norm. We also prove that the set of ReLU networks with two hidden layers of bounded widths is not closed in W^1,p(\Omega) or in C(\overline\Omega) . This disproves a conjecture of Petersen, Raslan and Voigtlaender.
[LG-1] Computer-Use Agents Are Embodied Generalists
链接: https://arxiv.org/abs/2611.09847
作者: Pranjal Aggarwal,Sean Welleck
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注: 14 pages, 3 figures, 2 tables
Abstract:In many areas of AI, general-purpose models have increasingly replaced handcrafted methods and specialized architectures. We argue that embodied AI is poised for a similar shift. Our key claim is that embodied tasks can be formulated as interaction with digital software. Specifically, we build software that renders an agent’s view of a physical environment on a screen and provides controls for acting in it. A computer-use agent (CUA) can then perform embodied tasks through the same screenshots and keyboard and mouse actions that it uses to operate ordinary software. Using a single CUA design and changing only the software, our entries win first-, second-, third-, and sixth-place prizes across four ECCV workshop competitions, without any competition-specific agent tuning. These results suggest that the skills CUAs learn on software, such as visual grounding, precise clicking, and exploration, already carry over to embodied tasks, giving CUAs a far broader scope in embodied AI than previously realized. We therefore advocate that computer-use agents can be embodied generalists. Further, we advocate that, by following the training strategy that made computer-use agents general, embodied AI can converge on a single general model as well, ushering in a new era of general embodied AI. Website: this https URL
[LG-2] Fifty-Two Shades of Pig: What the UCR Archive Really Asks Classifiers to Do
链接: https://arxiv.org/abs/2611.09843
作者: Omar Swelam,Jaris Küken,Shi Bin Hoo,Frank Hutter,Lennart Purucker
类目: Machine Learning (cs.LG)
*备注:
Abstract:The pigs were bleeding. That was the point of the experiment: could airway pressure reveal hemorrhage before it became critical? In the UCR Time Series Classification Archive, the de facto benchmark for Time Series Classification (TSC), the same recordings ask a different question: “which of 52 pigs produced this trace?”. The dataset PigAirwayPressure is not an outlier. Although the archive is the standard benchmark, whether its tasks resemble the problems TSC is meant to solve has rarely been examined. We find that several datasets are repurposed from their original task, built from toy setups, or have flawed test splits. We further question how representative the archive’s ordered non-temporal sequences (e.g. spectra, shape outlines) are of real TSC, and the potential downsides from the decomposition of multivariate problems. Our findings call for more careful interpretation of archive-wide scores and greater attention to deployment-relevant evaluation. The audit is at this https URL.
[LG-3] Designing Effective Monitor-Based Interventions for Mitigating Reward Hacking During RL NEURIPS2026
链接: https://arxiv.org/abs/2611.09828
作者: Aria Wong,Joshua Engels,Neel Nanda
类目: Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2026
Abstract:Reinforcement learning (RL) rewards are notoriously difficult to design and control, often leading to the model learning unintended behaviors such as reward hacking. One potential solution is to monitor for reward hacking and penalize it when detected; however, training against a monitor could lead to evasive behavior, and our general understanding of how to apply monitors effectively during training is limited. To study how best to use monitors to mitigate reward hacking, we introduce and open source three realistic environments where Qwen3-4B reward hacks: a coding environment hackable via test overwriting, a medical chat environment hackable via sycophancy, and a biography generation environment hackable via hallucination. We first focus on the coding environment, where we find that: (1) models can learn to evade highly accurate monitors by exploiting systemic flaws in probes and LLM judges; (2) monitors that leak more learning signal during RL suppress reward hacking but are more often evaded; and (3) including easier problems in training can decrease reward hacking. We apply our findings to build better reward hacking monitors for the medical chat and biography generation environments that improve upon naive baselines to reduce reward hacking rates across seeds to before-training levels in some interventions. Our results demonstrate that our takeaways translate to new settings and that better monitor intervention designs are possible.
[LG-4] he Cost of Coverag e: Conformalising Native Forecast Uncertainty for GB Wind Power
链接: https://arxiv.org/abs/2611.09815
作者: Carl Somers
类目: Machine Learning (cs.LG); Applications (stat.AP)
*备注:
Abstract:Uncertainty bands on wind forecasts help inform reserve sizing and commitments, but often fail to deliver their stated coverage. We apply conformal prediction to native bands to see what it adds and what it costs, comparing six forecasters, from ARIMA to a tabular foundation model, at two wind farms in Great Britain (GB) under operational conditions. We show that calibration repairs empirical marginal coverage with little or no Winkler Score penalty, but this apparent success conceals under-coverage in high winds. Calibrating by forecast wind regime turns marginal coverage into regime-level coverage, cutting day-ahead out-of-band volume in high winds by over 40% relative to pooled calibration while narrowing bands overall.
[LG-5] Quasipolynomial-Time Planning for Decodable POMDPs
链接: https://arxiv.org/abs/2611.09778
作者: Oliver Dobson,Varun Jog,Chun Hei Michael Shiu
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Optimization and Control (math.OC); Machine Learning (stat.ML)
*备注: 32 pages
Abstract:We study planning for finite-horizon partially observable Markov decision processes (POMDPs) with finite state, action, and observation spaces. Both exact and approximate planning are known to be computationally hard in this setting, so prior work has explored natural assumptions that make planning easier. Recent work [Golowich et al., 2022] gave quasipolynomial-time approximate planning algorithms for observable POMDPs. These are POMDPs whose observation channels satisfy an \ell_1 -observability condition, but notably, no assumptions are required on the transition dynamics. In a similar vein, for every fixed \rho0 , we give quasipolynomial-time approximate planning algorithms for \rho -decodable POMDPs. We show that decodability is implied by both \ell_1 - and \ell_2 -observability and is strictly weaker than observability. Moreover, we show that the optimal decodability constant can be computed in polynomial time, whereas computing the optimal \ell_1 -observability constant is NP-hard in the worst case. We also show that the fast filter-forgetting guarantee underlying the approach of [Golowich et al., 2022] does not extend directly from \ell_1 -observability to constant \ell_2 -observability. Thus, decodability is an easy-to-check, unifying assumption under which quasipolynomial-time approximate planning for POMDPs is possible.
[LG-6] From PAC bounds on the generalization error to PAC bounds on the parameter estimation error
链接: https://arxiv.org/abs/2611.09767
作者: Mihaly Petreczky,John Leth,Emma Lukacs,Balint Daroczy
类目: Machine Learning (cs.LG)
*备注: 10 pages, 0 figures, under submission to SysID2027
Abstract:Probably Approximately Correct (PAC) bounds on the generalization error (the difference between the empirical prediction error computed on the training data and its expectation) are an established tool in statistical machine learning, and there have been several extensions of these bounds to dynamical systems. In parallel, in recent years, there has been growing interest in finite-sample bounds for parameter estimation error in system identification. The latter bounds tend to be algorithm-specific, relying on reformulating the identification algorithm as a least-squares regression problem. This is in contrast to classical PAC bounds, which are algorithm-agnostic and can be applied to any identification algorithm based on minimizing a cost function involving the empirical loss. In this paper, we show how oracle bounds derived from PAC bounds can be used to derive finite-sample bounds for dynamical systems. As an illustration of this approach, we present finite-sample parameter-estimation bounds for linear and bilinear systems with bounded inputs and noise. These bounds apply to identification algorithms based on minimizing a regularized empirical loss function. The key ingredient is to establish an oracle inequality, i.e., an inequality between the true prediction error of the learned model and the minimum achievable true prediction error, and then use either the convexity of the true loss or system-theoretic properties of the loss function to derive bounds on the parameter estimation error. We also analyze the rates of the obtained bounds and compare them with those of classical asymptotic results. Comments: 10 pages, 0 figures, under submission to SysID2027 Subjects: Machine Learning (cs.LG) MSC classes: 68 ACMclasses: I.2.6 Cite as: arXiv:2611.09767 [cs.LG] (or arXiv:2611.09767v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.09767 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-7] A Tight Hierarchy for Chain of Thought
链接: https://arxiv.org/abs/2611.09759
作者: Thanh Le,A. Pavan,N. V. Vinodchandran
类目: Machine Learning (cs.LG); Computational Complexity (cs.CC)
*备注:
Abstract:Chain-of-thought (CoT) reasoning has emerged as a powerful paradigm for extending the computational capabilities of transformer-based models. A growing body of recent work has begun to uncover the fine-grained computational power of transformers augmented with CoT, as well as their connections to classical complexity theory. In this work, we advance this line of research by establishing the first hierarchy theorem for CoT-based computation. Our main result shows that for any reasonable resource bound r(n) \geq n and any \varepsilon 0 , [ \mathrmCoT(r(n)) \subsetneq \mathrmCoT(r(n)^1+\varepsilon), ] where \mathrmCoT(r(n)) denotes the class of decision problems solvable by transformers using O(r(n)) CoT reasoning steps on inputs of length n . Our result shows that increasing the number of CoT reasoning steps provably yields strictly greater computational power, establishing CoT steps as a fundamental resource in transformer-based computation. Unlike classical hierarchy theorems, which are typically proved via diagonalization, our approach leverages known coarse-grained simulations between CoT-augmented transformers and Turing machines in both directions. Building on these connections, we adapt the classical padding technique from complexity theory to this setting to obtain a tight hierarchy. A key technical contribution of our work is the implementation of padding within the transformer-based CoT model. Subjects: Machine Learning (cs.LG); Computational Complexity (cs.CC) ACMclasses: F.1.1; I.2.0 Cite as: arXiv:2611.09759 [cs.LG] (or arXiv:2611.09759v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.09759 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-8] Diffusion-free Score Matching for Exact Conditional Sampling
链接: https://arxiv.org/abs/2611.09758
作者: Thomas Wedenig,Robert Peharz
类目: Machine Learning (cs.LG)
*备注:
Abstract:The success of modern score-based generative models is largely driven by diffusion-based training, which learns a collection of time-indexed score fields corresponding to progressively noisier versions of the data distribution. While such models achieve impressive sample quality, leveraging them for sound probabilistic inference—particularly sampling from arbitrary conditionals of the model—remains challenging. This difficulty arises because conditioning information is only available for clean data and not for the higher noise levels required to generate exact conditional samples. In this paper, we introduce DIffusion-free SCOre matching (DISCO), a principled approach that sidesteps the need for time-dependent score fields altogether. Our method is based on an objective that, while reminiscent of diffusion-based training, estimates only a single score of the (slightly perturbed) data distribution, for which the exact conditional score is readily obtained. In experiments, DISCO is competitive with state-of-the-art diffusion models in terms of sample quality. More importantly, DISCO yields a more faithful representation of the underlying data distribution and—crucially—enables accurate sampling from arbitrary conditionals, outperforming standard heuristic samplers. This capability opens the door to principled and flexible probabilistic reasoning with score-based models.
[LG-9] One Run Is Not Enough: A Monte Carlo Protocol for Evaluating LLM Reliability on Data Science Tasks
链接: https://arxiv.org/abs/2611.09729
作者: Wenqi Ni,Ozan Evkaya,Miguel de Carvalho
类目: Machine Learning (cs.LG)
*备注:
Abstract:Large language models (LLMs) have shown promise for automating data science workflows. Benchmarks such as DataSciBench (2026) and StatQA (2024) have advanced the evaluation of their capabilities across data science and statistical tasks. Assessing their reliability requires examining both the quality of their responses and their consistency across repeated runs, task types, and difficulty levels. With this aim, we introduce DS-QA, a benchmark comprising 100 data science and statistics questions drawn from UK university courses, spanning 16 datasets and three difficulty levels. We develop a three-layer evaluation framework that combines general response properties (runtime, verbosity and text similarity), coarse-grained metrics (accuracy, code executability and completeness), and task-specific metrics. We evaluate OpenAI models using a Monte Carlo protocol with 100 repeated runs. The results reveal substantial variability in responses to data visualisation and supervised learning tasks, particularly at medium and hard difficulty levels, highlighting the importance of evaluating consistency alongside response quality. We also find that an increase of 0.1 in the word verbosity ratio corresponds to an increase of about 0.013 in the token verbosity ratio.
[LG-10] Fixed-point acceleration on Neural Wasserstein manifolds
链接: https://arxiv.org/abs/2611.09723
作者: Vitalii Aksenov,Sebastian Gutierrez Hernandez
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:We propose a fixed-point acceleration method for optimization over probability measures, applicable to a variety of machine learning problems. We represent probability measures as pushforwards of a reference distribution through a neural network map and equip the parameter space with the induced pullback Wasserstein metric. This geometric structure allows natural gradient descent to be interpreted as a fixed-point iteration on a Riemannian manifold. We accelerate this iteration using Riemannian Anderson mixing method, which exploits historical residuals to capture approximate second-order information without forming Hessians. In the present parametric setting, the manifold is represented by Euclidean coordinates endowed with a point-dependent metric, simplifying retraction and vector transport. We evaluate the method on multimodal sampling, PDE-based inverse problems, aggregation dynamics, Bayesian logistic regression, and generative modeling. Across the tested tasks, Anderson mixing reduces the number of natural-gradient iterations and achieves competitive performance relative to SGD and Adam in terms of both loss-gradient evaluations and wall-clock time. These results suggest that fixed-point acceleration is an efficient tool for improving the computational efficiency of training on parametric Wasserstein methods.
[LG-11] he Multiscale Single-Index Model: An MSE Analysis of Online SGD Dynamics NEURIPS2026
链接: https://arxiv.org/abs/2611.09682
作者: Gordon Dai,Joan Bruna
类目: Machine Learning (cs.LG)
*备注: 113 pages, 4 figures. NeurIPS 2026
Abstract:We consider the Multiscale Single-Index Model (MSIM), first introduced by Oymak and Soltanolkotabi (2021), as a stylized model for hierarchical feature learning with scale separation. Each layer extracts a shared single-index feature at one physical scale and passes it to the next, giving a tractable setting in which to study how deep architectures learn local multiscale representations. Under non-degeneracy and delocalization assumptions on the link function and planted features respectively, for the two hidden-layer case ( K=2 ), we analyze joint online spherical SGD under MSE loss, which is a canonical regression objective, in discrete time and show that it achieves weak recovery from random initialization in O(d\log^2 d) samples, followed by strong recovery to accuracy \varepsilon in O(d\log(1/\varepsilon)) additional samples. The main technical challenge arises from the non-Gaussian learned features observed by the second layer. We address this challenge through a quantitative Gaussian comparison using delocalization, combined with sharp trajectory-level control of the coupled SGD dynamics. This work is a companion to Bruna (2026, arXiv:2607.03347), which studies recovery under a correlation objective for general K .
[LG-12] Autonomous Particle Image Velocimetry Using Zero-Shot Sim-to-Real Reinforcement Learning
链接: https://arxiv.org/abs/2611.09679
作者: Chang Rajani,Luke Sebastien Marsden,Stephan Kunz,Claudio Mucignat,Ivan Lunati
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: 9 pages
Abstract:Particle image velocimetry (PIV) reconstructs a flow field from the motion of tracer particles but the reconstruction quality is bounded by how well the flow structure of interest is resolved in the images. The acquisition position is typically fixed by the operator before the measurement. We cast camera placement as a sequential decision-making problem and train a reinforcement learning agent to control a 6-DoF manipulator equipped with a camera, enabling it to locate, frame and resolve a hydrofoil wake in a water tunnel. The agent commands the camera pose and observes the scene only through the velocity field and uncertainty estimated from time-separated image pairs. It is trained entirely in a photorealistic simulation of our water tunnel, with wakes generated by Computational Fluid Dynamics (CFD) simulations of hydrofoils with varying profiles, positions and angles of attack. After zero-shot transfer to the real water tunnel, the policy reliably locates and frames the wakes of randomly placed hydrofoils. It resolves the wake at a higher resolution than a fixed camera view and with lower error in the measured flow statistics, in roughly half the time of a full scan and with far fewer image pairs.
[LG-13] Online Causal Configuration for Networked Systems via Doubly Robust Steady-State Learning NEURIPS2026
链接: https://arxiv.org/abs/2611.09665
作者: Yuli Liu,Zhiheng Zhang
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Accepted at NeurIPS 2026
Abstract:Measurement exploration reveals how interacting local controllers behave but can alter the steady-state regime a network operator seeks to optimize. Exploratory rewards may therefore fail to identify a configuration’s no-exploration value: under full exploration, identical logged distributions can conceal different target values. Identification requires accounting for both action-law mismatch and shifts in the steady-state context distribution. We introduce CAUSAL-UCB, which selects configurations using optimistic bounds on target causal values. Its interference-aware clipped doubly robust evaluator produces predecision certificates separating statistical variation from calibrated identification, transport, and temporal-dependence errors. For any finite horizon and predetermined schedule of measurement-exploration rates \eta_t\ , we prove lower bounds of \Omega((1+\sum_t\eta_t)^-1/2) on minimax inference error and \Omega(\sum_t\eta_t) on measurement-induced deployment loss in a stationary, potentially slow-mixing family. We bound regret under valid certificates and selected-information growth, and establish an anytime finite-sample certificate with a shared radius for estimation, selection, and reporting in a restricted known-local-interference model. In a high-mismatch benchmark with 24 paired seeds and 200 rounds, CAUSAL-UCB attains target regret of 1.732 versus 6.326 for certified IPS-UCB, with deployment gaps of 30.607 and 23.217 , respectively. Independent-audit coverage tests and trace-driven semi-synthetic studies further assess certificate coverage and deployment trade-offs.
[LG-14] nsorMutation: Learning Weight-Space Mutations of Deep Learning Models on Neural Graphs
链接: https://arxiv.org/abs/2611.09650
作者: Zaheed Ahmed,Philip Makedonski,Jens Grabowski
类目: oftware Engineering (cs.SE); Machine Learning (cs.LG)
*备注: 18 pages, 7 figures
Abstract:Mutation testing for deep learning (DL) faces a cost-realism trade-off. The pre-training approach produces realistic but computationally expensive mutants, whereas existing post-training operators modify trained models efficiently but produce mutants that less closely resemble real faults. This study asks whether the trade-off is necessary: can the weight-space effect of a fault-informed pre-training mutation operator be predicted from a trained model, without retraining it? We answer this question with TensorMutation, which represents a trained model as a neural graph and uses a mutation-conditioned graph neural network (GNN) to predict parameter changes. The GNN is trained offline and reused across unseen models so that TensorMutation can generate executable mutants without retraining. We construct a corpus of 25,252 original-mutant pairs from 71 multilayer perceptron (MLP) classifiers using 11 fault-informed pre-training operators, including 25 newly curated subjects. TensorMutation partially reproduces the behavioral effects of pre-training mutations on nine held-out subjects, and direct weight-space supervision significantly improves weight-space fidelity. On 12 real-fault subjects, TensorMutation achieves the highest median behavioral similarity among the evaluated approaches. The improvement over the post-training tool DeepMutation++ is statistically significant (Holm-adjusted p=0.018), whereas the differences from the pre-training tool DeepCrime and the norm-matched random control are not significant. These results show that fault-informed pre-training mutations leave learnable structure in weight space that transfers to unseen models.
[LG-15] BiLoGo: Reducing the smoothing of localised responses in Transformer-based neural operators for sheet forming
链接: https://arxiv.org/abs/2611.09621
作者: Yingxue Zhao,Haoran Li,Haosu Zhou,Tobias Pfaff,Nan Li
类目: Machine Learning (cs.LG)
*备注:
Abstract:Trial-and-error FE simulation of sheet forming could be computationally expensive, neural-network-based learned simulators are used to accelerate design. These simulators should accurately capture localised responses, such as thinning around sharp radii or near ribs, which determine formability. Transformer-based neural operators propagate information efficiently across the mesh but tend to smooth such localised responses. This study develops BiLoGo, a Bipartite Local-Global learned simulator with an FE-guided design that connects local node-element message passing with a Transformer-based neural operator for long-range information exchange. BiLoGo has three modules: a bidirectional node-element GraphSAGE model that learns nodal-elemental coupling on a bipartite graph and predicts each quantity on its native FE entity without interpolation; a dual-stream Transformer-based neural operator that processes blank deformation and tool contact with separate token sets, preventing localised contact information from being averaged with the whole-blank state; and a fusion and element update module that combines global and local features via gated residual fusion and updates element features conditioned on element-node kinematics. Evaluated on a dome-shaped cold forming case and two corner-shaped hot stamping cases (one with ribs), BiLoGo achieves the lowest positional and threshold-based thinning errors on all three cases compared with single-level and hierarchical graph-based models and Transformer-based neural operators. Near the ribs, the smoothing index of predicted thinning increases from 0.41-0.56 for Transformer-based neural operators to 0.94 for BiLoGo-U. Compared with the hierarchical graph network CAtt-BiUGNN, BiLoGo-T reduces mean positional error by 15.5% on the Corner case and BiLoGo-U by 8.3% on the Corner+Rib case, with 66-69% less training memory.
[LG-16] Contextual Search and Bilateral Trade with a Trace-Square-Root Potential
链接: https://arxiv.org/abs/2611.09615
作者: Romain Cosson,Federico Fusco,Anupam Gupta,Stefano Leonardi,Renato Paes Leme,Matteo Russo
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注:
Abstract:We present a unified potential method for contextual search and bilateral trade with one-bit feedback and obtain order-tight algorithms for both problems. Our framework starts from a Bayesian perspective in which the unknown target (for contextual search) and the unknown pair of valuations (for bilateral trade) are sampled from a known prior distribution. In that model, the decision maker bases its prediction on the posterior distribution, obtained by conditioning on feedback received in previous rounds. For Bayesian contextual search, we present an algorithm with a total loss bounded by O(\mathrmTr \sqrt \Sigma) , where \Sigma is the covariance matrix of the target’s prior; and for contextual bilateral trade, our algorithm’s regret is bounded by O(\mathrmTr \sqrt\Sigma_S+\mathrmTr \sqrt\Sigma_B) , where \Sigma_S and \Sigma_B are the covariance matrices of the seller’s and buyer’s valuations. We then use ideas from this Bayesian setting and adapt a ``dual’’ of this trace-square-root potential to the classical worst-case setting, where the target and valuations are chosen adversarially from the unit Euclidean ball. For both contextual search and bilateral trade, we obtain tight O(\sqrt d) -regret algorithms. These results match the existing lower bounds up to a constant factor and improve on the previous guarantees of O(d\log d) for contextual search and of O(d6^d) for bilateral trade. Subjects: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG) Cite as: arXiv:2611.09615 [cs.GT] (or arXiv:2611.09615v1 [cs.GT] for this version) https://doi.org/10.48550/arXiv.2611.09615 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-17] Know What You Need: Efficient Activation Checkpointing
链接: https://arxiv.org/abs/2611.09614
作者: Danyal Ahmed(MLIA),Edouard Oyallon(MLIA)
类目: Machine Learning (cs.LG)
*备注:
Abstract:Large Language Model training is increasingly constrained by memory capacity. Activation Checkpointing reduces activation memory by avoiding storing selected activations during the forward pass and rematerializing them during backpropagation, thereby trading memory savings for additional computation. Although this technique is essential for scaling model training, rematerialization can be prohibitively expensive, particularly for Transformer linear layers. Approximate computation could reduce this overhead, but conventional approaches either introduce errors that accumulate over the backward pass or fail to exploit information available from the initial forward pass. We introduce a novel approximate rematerialization method that can yield unbiased gradient estimates: using information collected during the initial forward pass, the method selectively recomputes only the most important layer components, either through greedy selection or randomized sketching. Ablation studies show that the resulting gradients closely match the exact gradients. When applied to feed-forward layers, attention layers, or both, our method can recompute only two-thirds of the layer components in a Llama-488M model, while increasing perplexity by just 1.7%, 1.2%, and 2.5%, respectively.
[LG-18] ReplayAudit: Auditing Outsourced LLM Fine-Tuning via Sealed Records and Selective Replay
链接: https://arxiv.org/abs/2611.09607
作者: Yihan Xia,Wenxin Xu,Taotao Wang
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: 23 pages, 5 figures
Abstract:Auditing outsourced LLM fine-tuning requires checking an untrusted provider’s training while controlling verification cost. Uniform sampling ignores training-time signals, while checking a suspicious record can require replaying several preceding steps. We present ReplayAudit, a telemetry-guided auditing system that seals training records before challenges and independently replays sampled intervals from sealed checkpoints. Loss changes and parameter-update norms guide stratified random sampling; replay determines whether the selected computation agrees with its records. We relate sampling probabilities to replay work and evaluate the system on real LoRA training. A 7B BF16 prototype rejects a false update-norm report during a random audit; targeted checks also reject false loss reports and skipped updates. We calibrate sampling budgets offline to at least a 90% probability of directly selecting an injected step. For two learning-rate deviation settings, executing these budgets on the same honest 1.5B trajectory reduces median complete-audit wall time by 49% and 67% relative to uniform sampling. When skipped updates evade the high-anomaly stratum, the corresponding audit instead costs 38% more. These results show how deviation capture and actual replay work jointly determine the cost of telemetry-guided auditing.
[LG-19] Coresets for Capacitated Clustering via Dual Concentration
链接: https://arxiv.org/abs/2611.09600
作者: Milind Prabhu,Chris Schwiegelshohn,Sudarshan Shyam,Erik Waingarten
类目: Data Structures and Algorithms (cs.DS); Computational Geometry (cs.CG); Machine Learning (cs.LG)
*备注:
Abstract:We study coresets for clustering problems, focusing on the classical k -median and k -means objectives. A coreset is a small weighted summary of the input data that approximately preserves the clustering cost for every choice of centers. Such summaries are a central tool for scaling clustering algorithms to large datasets. In this work, we consider the capacitated setting, where the clustering is subject not only to choosing centers but also to respecting prescribed cluster-size constraints. This model is a natural generalization of standard clustering and has been widely studied, but its coreset complexity is less well understood. For finite metric spaces, we obtain optimal-size coresets; surprisingly, these match the best-known bounds for unconstrained clustering. We also give improved coreset bounds in Euclidean space, notably the first coresets with near-linear dependence in k while being independent of the dimension and input size. The main new technical ingredients are concentration bounds for the dual solutions of min-cost flow problems, which allow for both significantly simpler and tighter analyses of the underlying sampling distribution used in the coreset construction. Subjects: Data Structures and Algorithms (cs.DS); Computational Geometry (cs.CG); Machine Learning (cs.LG) Cite as: arXiv:2611.09600 [cs.DS] (or arXiv:2611.09600v1 [cs.DS] for this version) https://doi.org/10.48550/arXiv.2611.09600 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-20] ransformer-FSQ: Discrete Latents for Extreme KV-Cache Compression
链接: https://arxiv.org/abs/2611.09598
作者: Andras Szecsenyi,Thomas L. Lee,Vaisakh Shaj,Elliot J. Crowley,Amos Storkey
类目: Machine Learning (cs.LG)
*备注:
Abstract:Private, personalised language models must run on local consumer devices, where large, frontier models cannot fit, and retaining long contexts is non-trivial. Sequential memory—the key-value (KV) cache—grows linearly with context length, quickly dwarfing the model itself: for Qwen3-1.7B-Base, 200,000 retained entries require 21 GiB, 6.6 times more than the model weights. Yet the KV-cache compression methods we measured fail beyond moderate compression ratios. We introduce Transformer-FSQ, addressing extreme KV-cache compression by replacing standard key and value projections with a finite scalar quantisation bottleneck. Our cache stores a packed vector of discrete integer codes per token, with no per-token quantisation metadata. Converting a pretrained transformer trains only these projections on 0.25 tokens per original parameter. We evaluate Transformer-FSQ against five baseline families on models from 0.6B to 8B parameters. On long-context retrieval (RULER), from 12 \times to 97.5 \times Transformer-FSQ is the most accurate method we evaluate, across scales. A fused Triton kernel decodes directly from the integer codes; with baselines given their authors’ kernels where available, it decodes at 1.01\times (95% interval 0.98 – 1.05 ) the throughput of native SDPA attention on the uncompressed Qwen3-8B-Base at 32\times and 1.54\times at 97.5\times . Extreme KV-cache compression is therefore a practical way to extend how much context a fixed device can hold.
[LG-21] Coordinatewise Epsilon for Adam under Heavy-Tailed Noise
链接: https://arxiv.org/abs/2611.09596
作者: Jeonseong Kim,Chenglin Fan
类目: Machine Learning (cs.LG)
*备注: 32 pages
Abstract:Adam estimates second moments coordinatewise, but typically adds the same scalar \varepsilon to every denominator. We study how this shared \varepsilon affects convergence when noise and curvature vary across coordinates. Under coordinatewise finite- p noise moments, p\in(1,2] , choosing \varepsilon_i from the noise and curvature bounds for each coordinate attains the heavy-tailed \ell_1 -stationarity benchmark O(T^-\fracp-13p-2) , matching the known optimal time exponent for smooth nonconvex stochastic optimization. For 1p2 , Adam with a shared \varepsilon_t admits a lower bound of order \Omega(T^-\fracp-12p) . This lower bound holds for every predictable scalar choice of \eta_t , \beta_1,t , \beta_2,t , and \varepsilon_t , precluding a dimension-uniform benchmark rate for this class.
[LG-22] Solver-Free Decision-Focused Learning via Neural Optimization Layers for Edge Selection
链接: https://arxiv.org/abs/2611.09560
作者: Xinda Zheng,Frits de Nijs,Edward Lam
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Decision-focused learning trains prediction models to improve downstream decisions, but many methods require costly solver calls or optimal-solution labels. We propose a solver-free decision-focused learning framework for NP-hard edge-selection problems with uncertain parameters in the objective, the constraints, or both. A neural optimization layer is first trained without solution labels, using problem objectives and constraint penalties. Graph-informed edge scoring and iterative refinement encourage concentrated and structurally consistent continuous decisions. The layer is then frozen, and a prediction model is trained through its continuous output using a problem-specific loss evaluated under observed parameters. At deployment, the predictor, optimization layer, and a lightweight decoder produce decisions without solving the original NP-hard problem. Experiments on the asymmetric traveling salesperson and endpoint-free orienteering problems show consistently better decision quality than prediction-focused learning and improvements over most evaluated decision-focused baselines. Our framework requires less training time than the evaluated decision-focused baselines, even when optimization-layer pretraining is included. Neural deployment also produces decisions faster than solver-based deployment, offering a better quality-efficiency trade-off. The layer also reduces training cost in existing decision-focused methods when used as an approximate optimization oracle.
[LG-23] Optimal Regret for Profit Maximization in Bilateral Trade
链接: https://arxiv.org/abs/2611.09558
作者: Matteo Castiglioni,Anna Lunghi,Alberto Marchesi
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注:
Abstract:Learning to profit from trade requires learning both sides of a market: what a seller is willing to accept and what a buyer is willing to pay. We study repeated bilateral trade, where an intermediary posts a purchase price to a seller and a sale price to a buyer whose valuations are drawn from an unknown joint distribution. Trade occurs only if both parties accept their respective prices, in which case the intermediary earns the difference between the two prices. The intermediary observes only whether trade occurs, without observing either valuation. We show that learning both sides of the market is no harder, in terms of regret, than learning just one. Specifically, we design an algorithm achieving regret \widetilde\mathcalO(T^2/3) against the best fixed pair of prices, matching, up to logarithmic factors, the classical lower bound for one-dimensional dynamic pricing. Our result shows that the classical one-dimensional regret rate survives the transition to two dimensions, raising the question of whether this absence of a curse of dimensionality extends to arbitrarily high dimensions.
[LG-24] Marginal-Nonuniform Multiclass Learning NEURIPS2026
链接: https://arxiv.org/abs/2611.09543
作者: Nataly Brukhim,Steve Hanneke,Amirreza Shaeiri,Maximilian Thiessen
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Accepted at NeurIPS 2026
Abstract:We study the problem of multiclass classification within the framework of marginal-nonuniform learning, where the constants in the learning rate may depend on the unknown marginal distribution generating the unlabeled data, but not on the target concept. This notion naturally lies between classical PAC learning, where the constants are independent of both the marginal distribution and the target concept, and more recent universal learning, where the constants may depend on both. We give a complete characterization of the optimal marginal-nonuniform learning rates, establishing a fundamental trichotomy: exponential, linear, and arbitrarily slow rates. This characterization is witnessed by a new combinatorial dimension, termed the uncentered DS-eluder dimension. A key ingredient in the upper-bound proofs is a set of new distributional exhaustion lemmas that exploit the sequential nature of this dimension and may be of independent interest.
[LG-25] meOmni-Live: Continuous Live Time Series Understanding with Long-Term Memory
链接: https://arxiv.org/abs/2611.09502
作者: Tong Guan,Alexander Chemeris,Yiqing Guo,Warren Jin,Zhao Li,Shirui Pan,Ming Jin
类目: Machine Learning (cs.LG)
*备注:
Abstract:Existing time series language models predominantly follow an offline, one-shot paradigm: they process fixed-length inputs independently and do not retain past states. This static approach fundamentally misaligns with the continuous nature of real-world streaming applications and fails to maintain the historical context necessary for cross-window reasoning. To bridge this gap, we introduce Continuous Live Time Series Understanding (CL-TSU), a novel paradigm that shifts the focus toward maintaining a persistent, stateful understanding over open-ended data streams. To systematically evaluate this paradigm, we present LiveTS-Suite, the first comprehensive data suite tailored for CL-TSU, supporting data generation, model training, and systematic evaluation across 18 tasks spanning six core capabilities. It comprises over 5K streams totaling 2.2B timesteps, annotated with 15K specific response events. We further design an evaluation protocol that jointly assesses response timeliness, content accuracy, and computational efficiency. Building on this foundation, we propose TimeOmni-Live, the first live time series understanding model. It integrates a hybrid time series encoding for natively temporal input, a stream-persistent hybrid memory mechanism to effectively preserve exact historical evidence and long-term task states, and an elastic stream-inference synchronization strategy to reduce processing lag. Evaluations show that TimeOmni-Live establishes a strong baseline for CL-TSU, outperforming frontier LLMs (90.0 vs. 79.3 for Opus 5) with only 4B parameters and achieving 206 \times lower p95 latency than its backbone. Deployment on a Jetson edge device further shows its ability to support real-time continuous temporal understanding locally at the edge.
[LG-26] Consensus-Constrained Dual-Federated Optimization with Client-Side Dynamic Weighting for Distributed Android Device Threat Defense
链接: https://arxiv.org/abs/2611.09470
作者: Mohammad Mashreghi,Mohammad Hossein Badiei,Md. Jalil Piran
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注: 10 pages, 6 figures, 8 tables
Abstract:Smartphones, smart-home hubs, and wearables are increasingly exposed to sophisticated Android malware, while centralized threat-detection pipelines raise privacy and scalability concerns. Federated learning (FL) enables collaborative model training without collecting raw client data, but partial participation and variation in local model quality can destabilize global optimization. In this article, we propose DW-FedADMM, a federated optimization framework that combines primal-dual coordination with performance-dependent client-side weighting. The dynamic coefficient modulates each participating client’s local state before the augmented-Lagrangian update, while the dual variables regulate disagreement between local and global model states. We evaluate DW-FedADMM on three Android malware benchmarks, namely Drebin, Malgenome, and TUANDROMD, with 5, 10, and 15 clients holding IID data shards, partial client participation, and communication budgets of 5 and 10 rounds. Each configuration is repeated with ten random seeds. DW-FedADMM obtains the highest seed-averaged accuracy on Drebin (97.64% +/- 0.20%) and the best mean accuracy in 9 of the 18 dataset-client-round configurations, including 7 of the 9 configurations with a 10-round budget. Pooled over all configurations, it reaches 95.42% accuracy, 94.62% F1-score, 96.06% precision, and 99.45% AUC, compared with 95.32%, 94.51%, 95.56%, and 99.41% for FedADMM, while also giving the lowest mean false-positive rate and the smallest client update norm among the six compared methods. Both primal-dual methods are well ahead of the averaging- and proximal-based baselines on Malgenome and TUANDROMD, where the latter often remain at the majority-class solution when clients are fragmented. The gain of dynamic weighting over FedADMM is most pronounced on Drebin and with the longer communication budget, where it also reduces the variation across seeds.
[LG-27] MINE: Entropy-coded Memory-efficient Inference for Large Language Models ICLR
链接: https://arxiv.org/abs/2611.09424
作者: Emil Schätzle,Markus Püschel
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF)
*备注: Submitted to the 15th International Conference on Learning Representations (ICLR), 2027. Artifact: this https URL
Abstract:We present MINE (Memory-efficient INference Engine), a lossless entropy coding approach that accelerates single-stream inference of Large Language Models (LLMs), where existing lossless compression methods yield speedups only for specific data types such as \textttBF16 and under additional assumptions. Single-stream decoding is strongly memory-bound, since every weight must be streamed from memory once per generated token and requests cannot be batched in local or low-latency deployments. At the core of MINE is dtANS-LLM, a GPU entropy coding scheme adapted from the prior dtANS to the needs of LLM inference. It exploits the non-uniform distribution of weights, which persists even after quantization, and exposes hyper-parameters that trade compression ratio against decoding cost. We auto-tune these parameters for each tensor and GPU to obtain the Pareto front between model size and decoding latency. MINE integrates into vLLM as a plugin and selects the best kernel configuration for a given memory budget. On Llama~3.1~8B~Instruct with a 2.1-bit EntQuant checkpoint, MINE achieves speedups of 2.52 \times and 1.48 \times over the fastest baseline on an RTX~4060~Ti and an RTX~5090, at effective bitrates of 2.24 and 3.43~bits per weight. MINE is available as open source.
[LG-28] Future-Informed Actor-Critic
链接: https://arxiv.org/abs/2611.09417
作者: Samy Mokeddem,Gaspard Lambrechts,Adrien Bolland,Damien Ernst
类目: Machine Learning (cs.LG)
*备注: 4 pages workshop-paper
Abstract:Actor-critic methods optimize a parametrized policy using a critic to estimate their policy gradient. Recent work has shown that conditioning the critic on features of future states, actions, or rewards can improve sample efficiency and even final return. However, existing methods are restricted to exogenous features (independent of actions), overlooking endogenous features (dependent on actions) that may be crucial for credit assignment. We introduce the future-informed value function, which can be conditioned on both exogenous and endogenous features of future states, actions, and rewards. Then, we express the policy gradient using this value function, yielding the future-informed policy gradient. Building on that, we propose the future-informed actor-critic algorithm. Empirical results show that this algorithm leveraging endogenous features of the future outperforms standard actor-critic algorithms in environments with noisy and sparse rewards.
[LG-29] Causal graph fusion across incomplete cohorts and causal steering of tabular foundation models
链接: https://arxiv.org/abs/2611.09400
作者: Yanke Li,Dhiyaa Al Jorf,Ayoube Ighissou,Anke Scheel-Sailer,Robert Riener,Diego Paez-Granados
类目: Machine Learning (cs.LG); Applications (stat.AP)
*备注:
Abstract:Clinical data are split across incomplete cohorts with partially overlapping variables. Causal discovery under missing data and multi-source integration have developed separately, and tabular foundation models cannot take a causal graph without retraining. We make two contributions. First, causal graph fusion learns a partial ancestral graph on the shared variables, fuses it into each cohort’s graph under temporal background knowledge and lets Markov blankets guide re-imputation, so missing data are handled inside discovery. Second, causal attention steering compiles the learned graph into the attention of a frozen tabular foundation model at inference. Against known ground truth, fusion dominates pooled and per-cohort baselines in every missingness and heterogeneity condition tested (p0.001), and steering improves discrimination under distribution shift (p10^-5), a gain that a random graph does not produce. In two spinal cord injury cohorts, fusion orients age, albumin and hemoglobin as ancestors of length of stay, a set no single cohort orients in full, and yields compact, temporally ordered predictor sets that match or exceed automatic feature selection on four of five outcomes.
[LG-30] k-Volume: a Unified Active Learning Strategy
链接: https://arxiv.org/abs/2611.09388
作者: Eliot Tron,Alain Celisse,Christophe Denis,Alice Le Brigant
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:In the classification context where revealing the label of a sample is costly, we propose a new Active Learning (AL) strategy named k-Volume (kVol). It randomly selects a sample to be labeled according to the volume of the smallest ball containing its k nearest neighbors. kVol is highly versatile: it can be applied either in the input space or in the output space, and it enjoys the desirable properties of the three main families of AL strategies: density, diversity, and (when performed in the output space) uncertainty. From a theoretical perspective, we establish that kVol is asymptotically equivalent to uniform sampling over the support of the (input or output) data distribution. Based on experiments on various synthetic and real-life tabular datasets and images, we observe that kVol is surprisingly stable: whatever the classifier or the experimental setup, it still lies among the best AL strategies unlike other competitors that fail dramatically. In particular, kVol outperforms state-of-the-art competitors such as CoreSet, TypiClust or Uncertainty on several real-life datasets, at a low computational cost.
[LG-31] Cutpurse: Decomposing Explanation-Guided Model Extraction on Tabular Data
链接: https://arxiv.org/abs/2611.09375
作者: Seyma Selcan Magara,Hanru Zhang,Fatme El Zahraa Hamad,Nico Pfeifer,Mete Akgün
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Explanation APIs are often treated as increasing model-extraction risk, but an end-to-end attack comparison cannot show whether an explanation reveals model-specific information or merely supplies a convenient next-query coordinate. We study this distinction for explanation-guided extraction of tabular classifiers. We separate an explanation into an attribution channel, which ranks influential features, and a coordinate channel, which supplies candidate feature values for perturbing those features. We then test whether the useful bin edges are privileged to the service or reproducible from the adversary’s own data. Across six datasets and five target families, using attribution rankings to choose which features to perturb provides no systematic extraction advantage. Lime’s discretization bin edges can improve fidelity by up to 41 percentage points on continuous features, but those edges are quantiles of the explainer’s background data rather than properties of the target model. In controlled comparisons, an attacker-side grid reproduces the useful bin-edge gains without querying an explanation API. We use these findings to derive Cutpurse, which combines a self-computed grid with diverse query generation and boundary search. Against a service that withholds bin edges, Cutpurse improves fidelity by 16.3 percentage points on continuous-feature targets. It also exceeds the undefended published attack by 7.0 percentage points. Thus, withholding default Lime bin edges alone does not prevent effective extraction under the in-domain auxiliary-input, probability-output, and same-family surrogate assumptions studied here. Our results concern default Lime and Shap configurations for tabular classifiers. They do not establish the same conclusion for other explainers, target classes, or data-free attackers. Subjects: Cryptography and Security (cs.CR); Machine Learning (cs.LG) Cite as: arXiv:2611.09375 [cs.CR] (or arXiv:2611.09375v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2611.09375 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-32] Calibrated Uncertainty Quantification on TDC ADMET: What Can Be Resolved Across 22 Datasets? NEURIPS2026
链接: https://arxiv.org/abs/2611.09373
作者: Laurits Fredsgaard,Michael Riis Andersen,Mikkel N. Schmidt
类目: Machine Learning (cs.LG)
*备注: 33 pages. Preprint of a paper accepted for poster presentation at the AIDaR Workshop at NeurIPS 2026
Abstract:Absorption, distribution, metabolism, excretion, and toxicity (ADMET) models often rely on small, heterogeneous datasets, yet their predictions may determine which compounds proceed to experimental testing. Useful evaluations must therefore assess both predictive performance and the quality of reported probabilities or intervals. We ask which common uncertainty-quantification methods perform best on the 22 Therapeutics Data Commons (TDC) ADMET datasets and what this benchmark can resolve. We compare eleven calibrated pipelines using nested five-fold cross-validation across ten chemical partitions and five model seeds. Ensemble Monte Carlo (MC) dropout has the best average rank for TDC performance, Brier score, and 90% interval score. For TDC performance, however, it is not separated from the other ensemble pipelines; for Brier and interval score, LightGBM is also not ruled out as best. LightGBM wins more individual datasets than any other method for TDC performance and Brier score, but performs poorly on several endpoints. Smaller evaluations can identify a different leader and often fail to recover the broader comparison. These results make LightGBM and an ensemble neural model complementary baselines. Comparative claims should be checked across multiple chemical partitions using statistical analyses matched to the question being asked.
[LG-33] Open-world Reconstruction-based Classifier with Concept Drift Adaptation
链接: https://arxiv.org/abs/2611.09350
作者: Joanna Komorniczak
类目: Machine Learning (cs.LG)
*备注:
Abstract:Data stream processing has become central to modern machine learning applications, with concept drifts and novel class appearances posing the primary challenges faced by recognition methods. Adapting to concept drifts while also being able to detect novel classes and distinguish groups of unknown observations can be seen as a step towards effective method employment in an open-world setting. This work proposes a classification method that can detect and learn the distributions of new, unknown classes, as well as adapt to concept drifts once the distribution of known classes shifts. The proposed method uses the autoencoders’ reconstruction errors and the KDE that models their distribution. The autoencoders incrementally adapt to the changing distributions of known classes via weight fine-tuning, resulting in continuous adjustment to evolving concepts. Conducted experiments used a diverse set of synthetic data streams, embeddings of a real-world static CIFAR dataset arranged into a stream to simulate concept drifts and the emergence of unknowns, as well as real-world datasets where novelty is observed. The methodology was compared with state-of-the-art approaches with three-criterion recognition quality assessment and resource consumption evaluation. The results indicate that the proposed method is competitive with reference approaches in stationary environments and performs best when concept drifts occur.
[LG-34] Probably Safe Sim-to-Real Transfer with Lifetime Validity
链接: https://arxiv.org/abs/2611.09344
作者: Rohan Tangri,Jan-Peter Calliess
类目: Machine Learning (cs.LG)
*备注:
Abstract:We present Dejavu, a method that makes sim-to-real transfer of reinforcement learning policies in constrained Markov decision processes probably safe over an unbounded sequence of policy updates. An anytime valid test using per-decision importance sampling and a betting confidence sequence certifies each candidate trained in simulation before deployment using only trajectories collected by a safe baseline policy operating in production. Simulation quality does not affect the safety guarantee, only the production sample efficiency of the method. Finally, a summable per-test failure probability yields the lifetime valid property. Empirical results across three sim-to-real benchmarks support the theoretical work.
[LG-35] DiffusionRef-IMU: Guide-Anchored Residual Diffusion for Cross-Location Inertial Measurement Unit Translation
链接: https://arxiv.org/abs/2611.09323
作者: M. M. Golam Hafiz,Safia Akter Simi,Md. Rakibul Islam,M. F. Mridha
类目: Machine Learning (cs.LG)
*备注: 18 pages, 5 figures, 6 tables
Abstract:Cross-location variation in wearable inertial measurement unit (IMU) signals complicates learning when measurements from a desired target body location are unavailable. This study proposes DiffusionRef-IMU, a two-stage cross-location translation framework in which a deterministic Direct network first estimates the target signal and a guide-anchored residual diffusion model refines only the remaining target discrepancy. The evaluation comprises five subject-disjoint activity-location composition-held-out folds, capacity-matched deterministic and flow-based residual baselines, source-only activity conditioning, computational ablation, and true zero-shot evaluation on two independent datasets. DiffusionRef-IMU with 16-step deterministic DDIM achieved a mean nRMSE of 0.087538 +/- 0.004922, compared with 0.090262 for Direct, 0.088847 for deterministic residual refinement, and 0.088417 for eight-step CompFlow, and outperformed each baseline in all five folds. Four-step DDIM retained a mean nRMSE of 0.087820 at 47.62 ms per window, providing a favourable quality-computation trade-off. Replacing ground-truth activity conditioning with a source-only activity predictor (macro-F1 0.8473) increased 16-step DiffusionRef nRMSE only to 0.088190. True zero-shot evaluation produced mixed results: DiffusionRef was marginally favourable on Grouvel2025 but remained worse than Direct on Saldaña2025. Overall, guide-anchored residual diffusion improves controlled cross-location reconstruction while revealing clear limits in zero-shot external transfer.
[LG-36] Stable probabilistic emulation of partially observed dynamical systems
链接: https://arxiv.org/abs/2611.09317
作者: Mario Massy,Sara Shamekh,Fabrizio Falasca,Matthieu Blanke
类目: Machine Learning (cs.LG)
*备注:
Abstract:Predicting the long-term behavior of partially observed dynamical systems such as the coupled atmosphere-ocean system is central to societal decision-making. Motivated by the success of deep learning emulators in weather forecasting, data-driven climate models represent a promising approach. However, climate emulation poses distinct challenges: models must account for unresolved fast variables introduced by unavoidable coarse-graining, reproduce long-term statistics, and remain stable over arbitrarily long rollouts and under external perturbations. These requirements motivate stochastic data-driven emulators, but modern generative models lack stability guarantees. We introduce forced-dissipative stochastic interpolants, a generative framework for learning stochastic reduced dynamics. Its architecture uses structural constraints inspired by turbulent dynamics, which provide stability by design while retaining the flexibility of stochastic interpolants to represent complex probabilistic dynamics. Building on the theory of Markov-chain stability, we establish a Foster–Lyapunov drift condition and prove geometric ergodicity for our emulator, together with stability under steady external forcing. We evaluate the framework on the two-timescale Lorenz-96 system, a prototype of partially observed multiscale turbulence. The resulting emulator remains stable over long integrations and under finite-amplitude forcing, with accurate prediction of forced responses of both the mean and variance.
[LG-37] Bridging Spiking and Continuous Dynamics in Spiking Neural Networks: A Muscle-Inspired Model
链接: https://arxiv.org/abs/2611.09295
作者: Maryam Soltani,Philip N. Garner
类目: Neural and Evolutionary Computing (cs.NE); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
*备注:
Abstract:Transforming the spike trains produced by spiking neural networks (SNNs) into continuous signals through biologically inspired mechanisms remains an open challenge in generative and regression tasks. In biological systems, a similar transformation occurs when muscles integrate neural spikes to produce smooth, coordinated movements. Inspired by this mechanism, we introduce a muscle-response model, a biologically motivated dynamical layer that converts spike trains into continuous representations. Our approach uses second-order dynamics to model the temporal integration of spiking activity. We investigate how the number of spiking neurons and temporal resolution affect the quality of the generated signals in a text-to-speech task, showing that increasing representational capacity has a greater impact than increasing temporal resolution under the evaluated configurations. Furthermore, an analysis of the learned dynamics reveals that, without explicit constraints on the response type, most neurons converge toward non-oscillatory behavior. These results demonstrate the potential of biologically inspired dynamical decoders for bridging spiking neural computation and continuous signal generation.
[LG-38] Safety-Focused Reinforcement Learning Using Risk-Sensitive Coupled Control
链接: https://arxiv.org/abs/2611.09290
作者: Jiabin Lin,Shana Moothedath,Vivek S. Borkar
类目: Machine Learning (cs.LG)
*备注:
Abstract:This paper investigates the coupled control problem within safety-focused reinforcement learning, addressing the essential challenge of achieving optimal performance while ensuring safety in systems operating within complex environments. We define safety as the requirement to prevent the system from reaching undesirable or hazardous states. We propose a coupled control approach that prioritizes safety by employing effective methods to prevent the system from entering unsafe states while optimizing the performance metric, defined as the average cost. We propose a novel problem transformation that reformulates it into a risk-sensitive problem combined with an average reward criterion. We present a two-timescale Q-learning algorithm where in the faster timescale, the algorithm solves a risk-sensitive control problem to ensure safety, while on a slower timescale, the algorithm focuses on average cost control. We provide theoretical guarantees for the convergence of our algorithms, establishing their significance in ensuring safe, optimal control in dynamic environments. We validated the performance of our approach using GridWorld and CarGoal task in Safety Gymnasium.
[LG-39] AnomSep: Bridging Optimization-Based Separation and Deep Neural Refinement for Anomaly Detection with Listenable Explanations
链接: https://arxiv.org/abs/2611.09279
作者: Koki Shoda,Jun Younes Louhi Kasahara,Qi An,Atsushi Yamashita
类目: ound (cs.SD); Machine Learning (cs.LG)
*备注: 11 pages, 3 figures
Abstract:We propose AnomSep, which combines optimiza-tion-based separation with deep neural refinement to achieve accurate anomaly detection with listenable explanations. Anomaly detection requires no predefined anomalous categories, and understanding the basis for each detection helps security personnel assess the validity of each detection. End-to-end anomaly detectors can achieve high detection accuracy but output only scalar scores, and existing anomalous sound separation methods provide listenable explanations with limited detection accuracy. AnomSep first obtains preliminary normal-sound and anomalous-sound estimates through optimization-based separation using a dictionary learned from normal sounds in the target environment. Deep neural refinement then improves the preliminary estimates in a pretrained audio latent space. The separated anomalous sound serves as the listenable explanation, and its energy determines the anomaly score. Experiments on synthetic mixtures and real mixtures show improved separation quality over baseline separators. AnomSep achieves detection performance competitive with the end-to-end baseline and provides listenable explanations directly linked to each detection decision.
[LG-40] Reconciling European Electricity Data with Diffusion Priors Learned from Observations Alone
链接: https://arxiv.org/abs/2611.09265
作者: Matthias Pirlet,Adrien Bolland,Gilles Louppe,Damien Ernst
类目: ystems and Control (eess.SY); Machine Learning (cs.LG)
*备注:
Abstract:The ENTSO-E Transparency Platform publishes generation, load, and cross-border power flows for European bidding zones. The values carry undocumented errors and violate energy conservation, and data reconciliation consists in correcting them to satisfy that balance. In this work, we formulate data reconciliation as Bayesian inference. Classical methods implicitly take a uniform prior, which we replace with a diffusion model estimated from the observations alone by empirical-Bayes expectation-maximization. The resulting posterior is sampled with a guided process whose projection step makes every sample satisfy the balance exactly. On synthetic systems of six to thirty-five bidding zones, built from ENTSO-E resource-adequacy data, the diffusion prior gives the most accurate reconciliation, increasingly so as the system grows. Its credible intervals are calibrated or conservative, and it remains the most accurate under a misspecified noise covariance. Finally, when the generation of entire zones goes unreported, the classical methods fail while the learned prior recovers it from cross-zone structure.
[LG-41] Rethinking MoE Load Balancing as Feedback Control
链接: https://arxiv.org/abs/2611.09252
作者: Yikun Hou,Alp Yurtsever
类目: Machine Learning (cs.LG)
*备注:
Abstract:Mixture-of-Experts (MoE) training is often not stationary: the data mixture changes at curriculum boundaries, when a model is continually pretrained into a new domain, or when the corpus is refreshed. Routing is learned against the current mixture, so each such change disrupts the per-expert load the router had converged to, and existing balancing methods have only been evaluated on stationary data. We rethink load balancing as a feedback control problem, in which the deviation of each expert’s load from its fair share is the control error and the per-expert routing bias is the control input, and define a family of controllers, the PID-Balancer, from the proportional, integral, and derivative terms of classical PID control. Two widely used methods belong to this family: the auxiliary balancing loss behaves like a proportional controller, and the loss-free bias update is an integral controller acting on the sign of the error. Reading the router as a static map with a one-step delay, we show that the integral term is the only one that both removes persistent imbalance and averages out the step-to-step fluctuation of the measured load, whereas the proportional and derivative terms pass or amplify that fluctuation into the routing bias. We therefore adopt PID (I), the pure-integral member of the family with a single gain. Two settings are evaluated: (i) training an OLMoE backbone from scratch with 16 to 64 experts per layer, through a mid-run mixture switch, and (ii) continual pretraining of the released OLMoE-1B-7B on a new corpus. PID (I) holds balance before the shift and restores it more tightly afterwards than either the auxiliary loss or loss-free balancing, at comparable perplexity.
[LG-42] opology as a Design Variable: Joint Architecture-Sizing Optimization for SRAM
链接: https://arxiv.org/abs/2611.09248
作者: Can Wang,Junhao Ma,Yikai Wang,Shan Shen,Wei W. Xing
类目: Machine Learning (cs.LG)
*备注: Accepted by ASPDAC2027
Abstract:SRAM’s power-performance-area-stability trade-off depends on bitcell topology, a discrete architectural choice that interacts non-separably with continuous transistor sizing. However, no open-source SRAM tool treats topology as an optimizable variable. OpenRAM generates SRAM layouts but performs no optimization, while OpenYield supports high-fidelity evaluation but restricts optimization to scalar sizing of a fixed 6T cell. OpenYieldV2 closes this gap through three integrated contributions. First, a fully parameterized compiler generates complete 6T and 10T SRAM macros with peripheral circuits and replica-bitline self-timing, enabling topology comparison under identical simulation conditions. Second, a few-shot TabPFN surrogate learns topology-dependent PPA responses with limited full-macro SPICE samples. Third, a coarse-to-refined multi-objective optimizer combines global exploration of discrete architectural choices with local refinement of continuous transistor sizes. Experiments across varying design conditions reveal a topology decision boundary that shifts with operating constraints rather than following a fixed design rule. Jointly optimizing across this boundary improves hypervolume by up to 17.3% over the strongest single-topology baseline. OpenYieldV2 is open-source to support reproducible SRAM design research.
[LG-43] Practical Non-Stationary Graph Gaussian Processes NEURIPS2026
链接: https://arxiv.org/abs/2611.09247
作者: K. S. Sesh Kumar,Viacheslav Borovitskiy
类目: Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2026. Code: this https URL
Abstract:Gaussian processes on graphs provide a principled framework for prediction and uncertainty estimation, but modelling heterogeneous networks often requires balancing flexibility against parameter complexity. Variability at a node may reflect useful signal that should propagate through the network, unreliable measurements whose influence should be suppressed, or a mixture of both. We introduce Variance-Coupled Gaussian Processes (VCGPs), which use a positive auxiliary signal to modulate graph interactions through a single learned coupling exponent. The exponent controls how strongly high-variability nodes propagate or attenuate information, while a cluster-wise extension accommodates different regimes within the same graph. Under stated structural assumptions, we characterize the unique stationary-rescaling point of the basic family, where correlations match those of a stationary graph kernel while node-dependent variances remain. Experiments on traffic, air-quality and weather datasets show competitive prediction and uncertainty estimation relative to stationary and per-node non-stationary baselines, with the largest gains on networks requiring multiple coupling regimes. Stationary correlations after variance rescaling remain sufficient on some data, and performance depends on the auxiliary signal and model selection.
[LG-44] Controllable Molecule Generation via Sparse Representation Editing: An Interpretability-Driven Perspective
链接: https://arxiv.org/abs/2611.09235
作者: Zhuoran Li,Xu Sun,Chang Wen Chen,Wanyu Lin
类目: Machine Learning (cs.LG)
*备注:
Abstract:Controllable molecule generation is crucial for diverse scientific applications, such as drug discovery and materials design. While large language models (LLMs) show great promise, their dense and entangled representations impede precise control over the generation of molecules with bespoke substructures or properties. To address this, we propose Sparse Representation Editing (SpaRE), an interpretability-driven framework for fine-grained and precise control in LLM-based molecule generation. The crux of SpaRE is to learn an overcomplete sparse feature space that disentangles LLM representations into a compact set of latent features corresponding to chemically meaningful concepts. Within this space, we can directly manipulate these concept-aligned latent features to achieve (1) local control, by generating target atoms and functional groups at specified positions; and (2) global control, by customizing the overall structural and physicochemical properties within defined ranges. In this way, our framework advances interpretability from post-hoc analysis to actionable generative control. Experiments show that SpaRE can generate chemically desirable molecules under complex constraints in real-world scenarios, while offering mechanistic insights for quantitative structure-property analysis. The code and demo are available at this https URL.
[LG-45] Cross-Layer TopK: Dynamically Allocating Sparsity Across Layers in Transcoders
链接: https://arxiv.org/abs/2611.09233
作者: Anyuan Zhuo,Shaohua Zhang,Tianle Li,Yuan Lin,Diandian Gu,Jun Wang,Pinyan Lu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Cross-Layer Transcoders (CLTs) decompose the activations of large language models into human-interpretable features and trace how these features interact across layers to form circuits, making them a key tool for mechanistic interpretability. However, training CLTs effectively and efficiently is hard. Beyond the usual trade-off between sparsity and reconstruction quality, a CLT has to model all layers at once, which adds a further challenge. The common JumpReLU recipe controls sparsity only indirectly through a penalty, while the direct extension of TopK selects a fixed number of latents in every layer. Yet reconstruction difficulty differs widely across layers, and intermediate layers are much harder than others, so giving every layer the same budget is suboptimal. We propose Cross-Layer TopK, which sets a single L0 budget for the whole model and dynamically allocates it across layers during training. In a comparison with JumpReLU and other TopK variants on Qwen3-8B-Base, this gives lower FVU while using 20% fewer active latents than JumpReLU. The better reconstruction carries over to circuit tracing, yielding more complete circuits with higher replacement and completeness scores. The fixed per-token budget of Cross-Layer TopK also makes it easy to pair with sparse decoding, which cuts both computation and communication and gives a 2.9x speedup per training step. We further show that Cross-Layer TopK follows scaling laws in both latent size and sparsity. We have open-sourced the weights of our CLT at this https URL.
[LG-46] Provable State Estimation with Recurrent Models
链接: https://arxiv.org/abs/2611.09217
作者: Vincent Andrieu(LAGEPP, CNRS),Pauline Bernard(CAS),Lucas Brivadis(L2S),Laurent Praly,Christian Wolf
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Estimating the state of a dynamical system from observations is a crucial task in robotics and control. Despite their empirical success, recurrent models often lack formal guarantees on convergence and robustness. We bridge this gap by analyzing these architectures from the perspective of nonlinear observer theory. By framing RNNs as a time-discretization of continuous latent dynamical systems (Neural ODEs), we demonstrate that enforcing a contraction condition on the latent flow guarantees the existence of a smooth, injective map of the true physical state into the latent space. We prove that this architecture acts as a tunable observer: by increasing a single scalar gain that scales the latent vector field, the estimation error can be driven to an arbitrarily small neighborhood of zero at an exponentially fast rate. We further establish quantitative bounds on the MLP complexity required to decode this latent state, and extend our formal guarantees to the discrete-time residual RNN implementation. We validate the theory on a real mobile robotic task, where a provably contracting model outperforms a vanilla GRU with substantially more stable training. We also report that unconstrained GRUs converge to contracting dynamics on this task, suggesting that contraction is an emergent property of well-trained sequence models.
[LG-47] RoleWalk: training-free structural embeddings for directed graphs that transfer across graphs
链接: https://arxiv.org/abs/2611.09189
作者: Maixent Chenebaux
类目: Machine Learning (cs.LG); Social and Information Networks (cs.SI)
*备注: 14 pages, 3 figures. Complete technical report (20 pages) as an ancillary file. Code: this https URL (pip install rolewalk)
Abstract:Structural role embeddings describe a node by its position in a graph - hub, bridge, periphery, member of a dense core - rather than by who its neighbours are. Existing methods either learn coordinates specific to one graph (struc2vec, Role2Vec, RESD) or read scales and normalisations from the whole graph (GraphWave, Digraphwave), so the same neighbourhood can receive different coordinates in two graphs; few read edge directions. We present RoleWalk, a training-free embedding that follows a short lazy random walk from each node and records, at each step, two distributions: that of the transition probabilities, through its characteristic function at a fixed grid of frequencies, and that of the walker’s distance to its starting node. On directed graphs, walks that go forward then backward add who shares a node’s targets or sources. The embedding depends only on a node’s neighbourhood within a fixed radius, so its coordinates mean the same thing in every graph. We show that, without curvature, it is the kernel mean embedding of the walk’s footprint, that it separates graphs that the Weisfeiler-Leman test cannot, and that pruning small probabilities comes with a per-node error certificate. Empirically, RoleWalk is competitive with the best role embeddings on five classification benchmarks at a fraction of their cost, has the highest scores on real directed roles in Wikipedia adminship, exceeds the published accuracy of Digraphwave at every noise level of its own Enron network-alignment benchmark, transfers centrality predictors to graphs four times larger better than existing inductive embeddings, hand-made features and a trained GIN, and embeds a million nodes in 20 s to 3 min on one CPU core. The code is open source (BSD license): this https URL
[LG-48] OFS: Feature Selection with Swap Supervision for Semantic Communication over Fading Channels
链接: https://arxiv.org/abs/2611.09173
作者: Zhiyi Zhu,Eiji Takimoto,Patrick Finnerty,Chikara Ohta
类目: Machine Learning (cs.LG)
*备注:
Abstract:Task-oriented semantic communication transmits task-relevant features to support receiver inference under limited communication resources and adverse channel conditions. For image classification, the usefulness of a feature depends on the other transmitted features and the distortion introduced by the wireless channel. Effective feature prioritization should therefore account for classification performance after transmission of the selected features. This paper proposes a Task-Oriented Feature Selection (\textitTOFS) framework and method, which evaluates feature replacements using receiver classification loss. Loss differences confirmed under independent transmission samples provide pairwise ranking supervision for a selector network. The trained selector predicts priorities in a single forward pass. CIFAR-10 simulations demonstrate improved classification accuracy over an importance-based feature selection baseline at the same transmitted feature count and average transmit power. The proposed \textitTOFS framework shows 84% accuracy performance under Rayleigh fading and Gaussian noise, compared with the 68% performance of a SOTA method.
[LG-49] Beyond the Grid: Continuous Dictionary Pursuit for Interpretable Signal Decomposition NEURIPS2026
链接: https://arxiv.org/abs/2611.09163
作者: Nick Janssen,Melanie Schaller,Bodo Rosenhahn
类目: Machine Learning (cs.LG)
*备注: 10 pages main paper, 3 pages references, 22 pages appendix, to be published in NeurIPS 2026
Abstract:Deep learning models achieve high accuracy in signal processing but remain difficult to interpret. Classical matching pursuit and sparse coding offer better transparency but suffer from spectral leakage and grid mismatch when signal frequencies do not align with discrete dictionary atoms. We introduce Continuous Dictionary Pursuit (CDP), a framework for interpretable signal decomposition that learns the parameters of closed-form functions directly from data. CDP optimizes over a continuous manifold of differentiable atoms and, unlike prior continuous pursuit methods that operate on a single analytic family, searches a heterogeneous library spanning periodic, trend, transient, and discontinuous structures. The algorithm employs a greedy iterative strategy using data-driven priors and gradient-based optimization to isolate individual signal components, with discontinuous waveforms made differentiable through spectral sharpness annealing. To ensure a sparse atom set for reconstruction, the framework utilizes a stopping criterion based on a predefined atom budget and residual error convergence. Evaluations on synthetic and real-world datasets show that CDP removes the grid discretization error for atom families contained in its library, recovers generative parameters to within 1% on stationary synthetic signals expressible in that library, and achieves the best sparsity-accuracy trade-off among classical, continuous-pursuit, and neural baselines in the sparse regime. Beyond decomposition, we demonstrate that the extracted atoms can serve as structured curriculum targets for training sequence models, improving forecasting accuracy over standard end-to-end training on most evaluated datasets. The method offers a glass-box alternative for signal decomposition by providing high reconstruction fidelity alongside direct physical interpretability.
[LG-50] Augmentron: Scalable Multi-View Visual Augmentation for Robot Learning NEURIPS2026
链接: https://arxiv.org/abs/2611.09145
作者: Tanay Tandon,Aseem Doriwala,Jade Choghari,Catherine Weaver,Pragna Mannam
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注: Accepted to the 8th Robot Learning Workshop (WRL), NeurIPS 2026
Abstract:Robot policies trained on visually narrow demonstrations can become sensitive to environmental features that are unrelated to the task. Existing generative augmentation methods can introduce visual diversity, but processing every frame and camera view independently is computationally expensive and may produce inconsistent observations across cameras and timestamps. We introduce Augmentron, a geometry-aware pipeline that generates reusable scene representations and renders them through recorded camera poses. This produces temporally and multi-view consistent environ- ment and workspace variations while preserving the robot, task objects, actions, and annotations. Across three physical manipulation tasks, replacing half of the real world training dataset with Augmentron data improves model success across three tasks for both in out-of-distribution distractor objects, and out-of-distribution workspace variation evaluations. In a separate comparison against alternate augmentation techniques, Augmentron achieves the highest pooled success in all three conditions. The results show that introducing augmented data to a data training recipe can help policies generalize to dynamic real-world settings.
[LG-51] Deep Learning Theory through the Lens of Statistical Physics: A Topical Review
链接: https://arxiv.org/abs/2611.09138
作者: Wei Huang,Youjin Deng
类目: Machine Learning (cs.LG); Statistical Mechanics (cond-mat.stat-mech)
*备注: 22 pages, 5 figures, 1 table. Topical review published in Front. Phys. 22(1), 011302 (2027); this version contains minor corrections
Abstract:The recent success of foundation models and deep learning has far outpaced our theoretical understanding of how large-scale neural networks learn and generalize. Statistical physics provides a natural language for describing such high-dimensional, nonlinear, and strongly coupled systems, offering concepts such as mean-field theory, phase transitions, energy landscapes, and stochastic dynamics. In this topical review, we adopt a statistical-physics way of thinking – from concepts including phases and criticality, and free-energy viewpoints, to methods such as kernel and Gaussian-process limits, mean-field and dynamical mean-field theories, random-matrix tools, and Fokker-Planck formalisms, and to a mindset emphasizing universality, scaling, and tractable limits with explicit validity regimes. We survey recent progress on understanding modern deep learning through this lens, including mean-field analyses of wide networks, spectral and geometric perspectives on loss landscapes, stochastic-gradient-based training viewed as a dynamical process, and phase-transition analogies that organize phenomena such as overparameterization, generalization, and representation change. Our goal is to synthesize these threads into a coherent framework that clarifies what each approximation explains, when it is expected to hold, and how apparently competing narratives relate or diverge. Together, these perspectives form a unified physical framework that views representation learning as a collective phenomenon in high-dimensional systems, bridging deterministic and stochastic regimes of gradient-based optimization.
[LG-52] Authorization Drift in LLM Agents : Evaluating Boundary Compliance in Multi-Step Tool Use
链接: https://arxiv.org/abs/2611.09113
作者: Chuxue Cao,Jingru Zeng,Boqin Yin,Yujin Zhou,Yidan Huang,Sujin Chen,Jinluan Yang,Haoran Li,Yike Guo,Sirui Han
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:LLM agents can exceed their authorization boundaries while pursuing user goals without an adversary. We formalize this phenomenon as authorization drift across data, tool, scope, and time, and introduce the Authorization Drift Benchmark (ADB): 648 samples across seven domains, each pairing a sufficient compliant path with an attractive unauthorized shortcut. To ensure executable tasks, tempting shortcuts, and feasible compliant paths, Staged Execution-Reflective Synthesis (SERS) grounds benchmark validation in real agent rollouts rather than LLM self-critique. Across 13 LLMs and three frameworks (25,207 trajectories), greater capability accompanies greater drift, with frontier-capable models exhibiting drift rates of 0.43 to 0.61. Yet 86.18% of tasks admit an observed clean-success trajectory, showing that drift is avoidable. Drift is persistent and dominated by early data over-collection; apparent improvements under agentic scaffolds coincide with collapsing clean-completion performance. To understand potential harms to individuals and society, we map drift points to legal provisions across four jurisdictions, connecting overreach to privacy and information-security concerns. Data minimisation is the most frequently implicated provision, and 93.5% of legally grounded drifting runs map to multiple jurisdictions, making the potential consequences of authorization drift more interpretable.
[LG-53] COMBIO: Automated Discovery of a Compact Composite Biomarker from Routine Laboratory Panels
链接: https://arxiv.org/abs/2611.09102
作者: Tomer D. Meirman,Lior S. Rokach,Bracha Shapira,Seffi Cohen,Noa Dagan
类目: Machine Learning (cs.LG)
*备注: 27 pages including supplementary material; 4 figures
Abstract:Machine-learning models based on routine laboratory panels can predict outcomes but can be difficult to interpret and operationalize. We present COMBIO, an automated pipeline that distils such a model into a compact closed-form biomarker. The composite is a product of laboratory measurements raised to integer powers from minus two to two. The sign of each exponent indicates whether higher values of the corresponding measurement increase or decrease the composite. Shapley additive explanation values from the gradient-boosted reference model rank five candidate measurements and guide eligible exponent signs. An exhaustive search over small integer exponents selects the formula on an inner validation set, before evaluation on the reserved test set. We evaluated COMBIO across six tasks: three National Health and Nutrition Examination Survey outcomes, liver-disease classification in the Indian Liver Patient Dataset, and Medical Information Mart for Intensive Care IV in-hospital mortality and sepsis case-finding, with external evaluation in the eICU Collaborative Research Database for the two hospital tasks. Across the six test sets, COMBIO’s area under the receiver operating characteristic curve differed by at most 0.03 from logistic regression using the five candidate laboratory measurements. The hospital mortality composite had an area under the curve of 0.698 in the external database. A version restricted to widely available measurements in the source hospital database achieved an area under the curve of 0.786 in a cohort covering 92% of admissions. The multiplicative representation provides compactness rather than additional accuracy, and local validation and recalibration are required before clinical use.
[LG-54] UGB: A Two-stage Uncertainty-aware Generative Auto-Bidding Framework
链接: https://arxiv.org/abs/2611.09091
作者: Luzhan Yuan,Chao Mi,Xiahao Shi,Yi Mao,Shengsheng Niu,Xin Zhu,Jie He,Zhangang Lin,Ching Law
类目: Machine Learning (cs.LG)
*备注:
Abstract:Generative modeling, particularly Decision Transformers, shows promise in auto-bidding strategies. However, it struggles with limited exploration in static offline datasets, vulnerability to dynamic environmental shifts, and a reliance on heuristic return-to-go (RTG) specifications that may yield suboptimal performance on unseen traffic. To address these challenges, we propose the Two-stage Uncertainty-aware Generative Auto-Bidding (TUGB) framework. TUGB integrates generative sequence modeling with Reinforcement Learning (RL). It employs a two-stage pipeline where a pre-trained Generative Bidding Model is fine-tuned via an ensemble-based Uncertainty-aware Environment Model. This approach quantifies epistemic uncertainty to penalize high-risk trajectories and evolves the policy from RTG-conditioning to direct bidding optimization. Extensive experiments demonstrate that TUGB significantly outperforms SOTA baselines. Notably, full-scale deployment on JD Search Ad achieved a 2.32% increase in cost and 3.57% in target cost, validating its industrial efficacy, while its successful adaptation to subsequent mechanism upgrades and JD Food Delivery Ad further validates TUGB’s superior generalization capability.
[LG-55] BrainCoT: A Multi-Task Zero-Shot Brain Signal Foundation Model with Neurometric-Anchored Chain-of-Thought Reasoning NEURIPS2026
链接: https://arxiv.org/abs/2611.09082
作者: Yiyu Gui,Mingzhi Chen,Guibo Luo,Yuchao Yang
类目: Machine Learning (cs.LG)
*备注: 37 pages. Accepted at NeurIPS 2026
Abstract:Brain signals exhibit substantial multi-center heterogeneity, strong inter-subject variability, and highly diverse task settings, which makes it difficult to build models that generalize beyond a single task. While recent brain signal foundation models leverage large-scale pretraining to learn generalizable representations, they often fall short in practice because they still require target-data-specific supervised adaptation, lack multi-task zero-shot capability, and rarely provide verifiable evidence to support their predictions. In this work, we present BrainCoT, a multi-task zero-shot Brain signal foundation model with neurometric-anchored Chain-of-Thought (CoT) reasoning. BrainCoT unifies heterogeneous tasks into an instruction-driven framework so that, after pretraining, a single model can be directly applied to multiple tasks in zero-shot settings without additional adaptation. It further introduces neurometric-anchored chain-of-thought grounded in neurometric evidence for verifiable reasoning, and adopts Decision-Consistent Rationale Training to keep process learning aligned with correct predictions. Across extensive evaluations, BrainCoT achieves an average ACC of 67.41% and AUC of 74.33% in zero-shot classification without downstream training data, outperforming the strongest linear-probing baseline trained with 1% labeled downstream data by 3.85% in ACC and 5.97% in AUC.
[LG-56] WearableBridge: Task-Guided Latent Transport for Free-Living Activity Recognition
链接: https://arxiv.org/abs/2611.09067
作者: Jingjing Zou,Yixin Wang
类目: Machine Learning (cs.LG)
*备注: 26 pages, 5 figures, 11 tables. Code and interactive demo: this https URL
Abstract:Wrist-worn accelerometers make activity monitoring feasible at scale, but incidental arm motion, device orientation, and shifting context blur the signal that separates activities. Calibration studies supply higher-quality references such as concurrent thigh recordings or supervised laboratory wrist sessions, but they are expensive, cover few participants, and are rarely available at deployment. How can limited reference data improve recognition for participants who wear only a wrist device? We propose WearableBridge, which holds the reference encoder and classifier fixed and learns a task-guided transport map, the bridge, that carries each wrist feature into the region the classifier associates with its activity. Reference features or the classifier’s predictions indicate the destination, and a classification loss penalizes wrist features that land in the region of another activity. We instantiate WearableBridge with three bridges: Mapper regresses onto reference features, Distillation matches the classifier’s class probabilities, and MeanFlow learns a one-step flow from wrist toward reference features. On 40 held-out free-living PAAWS participants, all three bridges improve macro-F1 in every training run, and the strongest, MeanFlow, raises mean macro-F1 from 35.1% to 56.0% with concurrent thigh references (Thigh2Wrist), and from 54.0% to 63.5% with group-matched laboratory wrist references (Lab2Freeliving). Fitted only on PAAWS, MeanFlow also raises macro-F1 on the CAPTURE-24 cohort from 42.8% to 54.7%. The gains replicate on selfBACK, hold across four reference encoders, and grow with temporal context. A study can therefore collect reference recordings for a calibration subset and improve wrist-only recognition for every other participant, training only a bridge of 0.13 to 2.44 million parameters while the reference encoder and classifier stay unchanged.
[LG-57] StreamTab: Training-Free Temporal Link Prediction with Tabular Foundation Models
链接: https://arxiv.org/abs/2611.09052
作者: Mingqi Yang,Renqi Wu
类目: Machine Learning (cs.LG); Databases (cs.DB); Social and Information Networks (cs.SI)
*备注:
Abstract:Temporal link prediction uses past interactions to predict whether two nodes will interact in the future. Temporal graph neural networks achieve strong results on this task, but they are trained separately for each graph. How a model should use interaction history also depends on the candidates it is compared with. A past interaction can distinguish a recurring link from a random pair that has never interacted, but when the negative candidates have also interacted before, other signals are needed. We propose StreamTab, a training-free framework built on a frozen tabular foundation model pretrained only on synthetic tables. StreamTab describes each candidate pair by statistics of its past interactions and shows the model recent observed interactions and sampled negative pairs as labeled examples, from which the model scores new candidates. Its central design, regime conditioning, samples the negative examples according to how the candidates are sampled, so the examples show the distinction the model must make. As new interactions arrive, StreamTab updates the statistics and refreshes the examples while the model parameters stay fixed. On thirteen datasets against 21 temporal graph baselines, StreamTab achieves the highest average precision on twelve datasets under historical sampling and eleven under inductive sampling, and ranks first or stays within one point of the best on seven datasets under random sampling. With the same features and model, regime conditioning raises average precision over examples with random negatives by a median of 12.6 points under historical sampling.
[LG-58] LeanDojo-v2: A Comprehensive Library for AI-Assisted Theorem Proving in Lean NEURIPS
链接: https://arxiv.org/abs/2611.09048
作者: Ryan Hsiang,William Adkisson,Robert Joseph George,Anima Anandkumar
类目: Machine Learning (cs.LG)
*备注: 11 pages, 7 figures, accepted to NeurIPS Math-AI workshop
Abstract:Building on recent progress in machine learning for formal theorem proving in Lean, we introduce LeanDojo-v2: an end-to-end framework that unifies repository data extraction, LLM fine-tuning, proof search, AI autocompletion, and LLM-assisted theorem proving in a single software library. In addition, we provide Lean4Code, a lightweight IDE that integrates the various features of LeanDojo-v2 through a simple, mathematician-friendly UI. These tools will enable mathematicians and researchers to easily fine-tune language models on domain-specific Lean codebases and experiment with LLM-assisted theorem proving without a complex setup.
[LG-59] When Can You Trust Neural Critics? Auditing Mutual Information Estimates of Higher-Order Epistasis in Protein Language Models
链接: https://arxiv.org/abs/2611.09046
作者: Maryam Rahimimovassagh,Waldemar Karwowski,Niloofar Yousefi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Mutual information (MI) estimation is used to probe what neural representations encode, including in protein language models, but variational MI estimators have variance that scales at least exponentially in the true MI. We audit a prior neural-MI study of order-resolved epistasis in ESM-2 on the four-site GB1 landscape. Run-to-run differences show no systematic bias and a magnitude comparable to ordinary cross-seed retraining variability, and the evaluation-pairing scheme shifts scores substantially. At the original dimensionality, two of three alternative estimator families stabilize the estimate across seeds and one does not. In a synthetic calibration at the representation dimensionalities of two other protein language models (a Gaussian channel with known MI; no model representations), all three families recover materially less signal than required, despite correct signal ordering and stable seeds. A control critic given the true signal direction, which also reduces its input to one dimension, recovers 93.3% and 98.7% of the analytic MI at gain 1.5 (at 5,120 and 4,096 dimensions), so the signal is recoverable with privileged information, but the control does not identify why blind estimation fails. We package the checks into a reusable diagnostic protocol and conclude that even a stabilized estimator is not currently sufficient for order-resolved dependence claims made under blind search at this dimensionality and sample size.
[LG-60] Function-Space Certification for Neural Operators
链接: https://arxiv.org/abs/2611.09038
作者: Ryan Hsiang,Robert Joseph George,Anima Anandkumar
类目: Machine Learning (cs.LG)
*备注: 45 pages, 6 figures
Abstract:Certifying the robustness and physical consistency of scientific machine-learning models ensures that their predictions remain reliable beyond the test set. For neural operators, which learn PDE solution operators by mapping infinite-dimensional function spaces, certification means establishing rigorous, domain-wide guarantees for their output functions. However, standard neural network (NN) verification methods are fundamentally designed for vector-to-vector mappings, and they can only certify neural operators at fixed grid resolutions and provide no guarantees over the entire domain or across another resolution. We close this gap by introducing a function-space certification framework that verifies neural operators across the entire continuous domain. We formulate interval bound propagation (IBP) and CROWN in function space, propagating upper and lower bound functions through each layer using rigorous numerical integration. In addition, we extend bound propagation to higher-order derivatives to certify PDE residuals and solution error bounds. Experiments across several neural operator architectures and scientific problems, including the 572.5-million-parameter spherical Fourier neural operator for global weather forecasting, show that our method produces tight, domain-wide guarantees efficiently, providing a foundation for trustworthy operator learning.
[LG-61] Knowledge Distillation from Neural Networks to Differentiable Logic Networks
链接: https://arxiv.org/abs/2611.09024
作者: Hee Yun Suh,Niraj K. Jha
类目: Machine Learning (cs.LG)
*备注:
Abstract:Differentiable logic networks (DLNs) are interpretable, edge-friendly classifiers that learn both their logic operators and their connections through gradient descent, achieving accuracy competitive with multi-layer perceptrons at orders-of-magnitude lower inference cost. Their accuracy, however, is bounded by the problems of vanishing gradients, scalability, and lack of an ecosystem, leaving a gap relative to conventional neural networks on several datasets. We introduce knowledge distillation for DLNs to narrow this gap. We distill from a trained multi-layer perceptron teacher into a DLN student, integrating the distillation objective into a two-phase training loop without altering the architecture or the inference procedure. The teacher is discarded after training. Hence, the distilled model retains the interpretability, sparsity, and inference cost of a conventionally trained DLN. We investigate this approach through four variants: a joint hyperparameter search that optimizes the distillation weight and temperature alongside the architecture, a grid search over these two hyperparameters with the architecture fixed, and two single-phase variants that employ distillation during only the neuron-function or only the connection phase. On 20 public tabular classification tasks, distillation improves balanced-class accuracy at negligible cost to gate-level operation counts and model size, though the benefit varies by dataset. Performance improvements of up to 15% are observed, in some cases surpassing the teacher itself.
[LG-62] KnowEvo: Knowledge Evolution for Protein Optimization NEURIPS2026
链接: https://arxiv.org/abs/2611.09013
作者: Zijie Xing,Xingyue Liu,Runze Wang,Luoming Hu,Yanming Shen
类目: Machine Learning (cs.LG)
*备注: Accepted by NeurIPS 2026
Abstract:Large language models (LLMs) have shown promise in scientific discovery, yet systematically accumulating and refining reusable scientific knowledge remains challenging. Existing approaches primarily improve LLM-based problem solving through parameter optimization or feedback-driven refinement of prompts, programs, and solution trajectories, leaving acquired expertise either embedded in model weights or scattered across iterative problem-solving processes. As a result,knowledge is difficult to systematically inspect, evaluate, and refine. We propose KnowEvo, a framework that treats externalized knowledge itself as the optimization target. Our key insight is that scientific knowledge is inherently modular: multiple interacting mechanisms jointly determine the design outcomes. This makes knowledge evolution fundamentally challenging, as improvements cannot be attributed to individual components in isolation. To address this, KnowEvo represents knowledge as executable expertise blocks and organizes their refinements into strategy mutation trees, enabling structured exploration over a combinatorial knowledge space. The framework combines a Thompson-style forward allocation strategy that adaptively distributes evaluation budgets across strategy variants with a backward revision process, in which a code agent operates in an isolated sandbox to translate rollout evidence into localized program updates. This design enables persistent accumulation, evaluation, and refinement of mechanistic expertise across iterative experimentation. We evaluate KnowEvo on protein optimization tasks including stability and solubility. Results show that knowledge evolution consistently improves over initial strategy blocks and outperforms monolithic program evolution and prior LLM-guided evolution baselines across metrics such as pass@k, MRR, MAP,and NDCG@5. Code is available at this https URL.
[LG-63] owards Robust Reinforcement Learning Market Making: Lessons from Loss Function Experimentation
链接: https://arxiv.org/abs/2611.09012
作者: Siddharth Nand,Bora Guney
类目: Machine Learning (cs.LG); Computational Finance (q-fin.CP); Trading and Market Microstructure (q-fin.TR)
*备注: 6 pages, 3 figures
Abstract:Market makers play a vital role in ensuring efficient trade within financial markets. Recent literature has delved into training market-making systems using reinforcement learning. This paper delves into the impact of various loss function formulations on risk management behavior within reinforcement-based market making. Through empirical evaluation, we contrast profit-maximizing strategies with risk-averse approaches, finding that penalties for inventory and cash management foster more resilient and adaptable market-making strategies. Our results underscore the significance of striking a balance between profit maximization and risk management in algorithmic trading, providing actionable insights for refining strategies in practical settings.
[LG-64] Calibrating your confidence requires thinking of all the ways you could be wrong
链接: https://arxiv.org/abs/2611.08997
作者: Marco Miani
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: PhD Thesis, Technical University of Denmark (DTU), 2026
Abstract:A primary limitation of modern deep learning is the miscalibration of model confidence; traditional architectures fundamentally lack a mechanism to quantify their own uncertainty. Larger models and larger datasets mitigate this issue, but they do not resolve the underlying structural deficiency. Bayesian methods offer a principled uncertainty-aware learning framework through stochastic parameters. Yet, despite the established benefits, the inherent computational overhead often compromises either scalability or quality. This trade-off currently prevents the widespread adoption of Bayesian methods. This Thesis bridges the gap between high-quality and low-complexity by introducing a novel distribution on parameter space: a degenerate Gaussian whose covariance is a projection matrix aligned with the null space of the Fisher information matrix. We demonstrate the scalability of this Bayesian posterior through two numerically efficient algorithms: one for posterior sampling and another for the direct computation of predictive confidence. Empirically, we show State-Of-The-Art confidence calibration and out-of-distribution detection for large models. We provide theoretical grounding for this approach. We study the degeneracies inherent in Amari’s information geometry and highlight its most relevant component: the split of parameter’s perturbations into two orthogonal subspaces. From this perspective, our posterior naturally emerges as the simplest choice that captures this split, assigning unitary variance to one subspace and zero variance to the other. Moreover, through the tools of differential geometry, we utilize this split to identify and rectify a critical misalignment issue of the non-linearized Laplace Approximation. Finally, with a focus on PAC-Bayes generalization bounds, we pave the way for a deeper understanding of the learning capabilities of overparametrized models. Comments: PhD Thesis, Technical University of Denmark (DTU), 2026 Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML) Cite as: arXiv:2611.08997 [cs.LG] (or arXiv:2611.08997v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.08997 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-65] When2What: A Timing-Conditioned Framework for Long-Horizon Forecasting of Continuous Cardiovascular Waveforms
链接: https://arxiv.org/abs/2611.08993
作者: Chaoqi Zhang,Shujie Yan,Zhao Zhang,Yanhua Sun,Yu Wang,Li Li,Lingzhong Meng,Xiaoyi Chen,Haixu Tang
类目: Machine Learning (cs.LG)
*备注: 31 pages (18 pages main text, 13 pages supplementary material), 7 figures, 31 tables. Code available at this https URL
Abstract:Long-horizon forecasting of quasi-periodic cardiovascular waveforms, such as the electrocardiogram (ECG), arterial blood pressure (ABP), and photoplethysmogram (PPG), requires predicting both when future cardiac events will occur (beat timing) and what the local structural pattern looks like around each event (beat morphology). However, standard sequence forecasters trained with pointwise mean squared error (MSE) average across uncertain future heartbeat phases, flattening sharp R peaks and pulse upstrokes into featureless baselines. Our analysis of intraoperative ECG reveals a clear temporal asymmetry: beat-time predictability from history decays within seconds due to stochastic heart-rate variability, whereas local beat morphology aligned at true beat times remains stable over several minutes. Guided by this insight, we introduce When2What, a timing-conditioned framework comprising three stages: Predict, Commit, and Generate. Given a 120-s history, When2What predicts a dense 60-s timing map, selects and sharpens discrete beat events into unit-height Gaussian impulses, and conditions a neural waveform generator on both these committed events and the observed history. Across intraoperative ECG, When2What substantially outperforms existing time-series and physiological forecasters, achieving a near-term waveform event skill of 0.644 and reproducing 95.62% of its predicted heartbeats in the generated waveform where direct long-horizon forecasters collapse into flat baselines. Furthermore, When2What transfers zero-shot to external hospital and atrial fibrillation cohorts, extends effectively to ABP and PPG pulse waveforms, and is supported by ablations and synthetic analyses that establish explicit timing conditioning as an effective paradigm for quasi-periodic long-horizon waveform forecasting.
[LG-66] Enforcing Symmetries on Pre-trained ML Models for Physics
链接: https://arxiv.org/abs/2611.08992
作者: Wenjie Zhou,Valentino F. Foit,David W. Hogg,Soledad Villar
类目: Machine Learning (cs.LG); Computational Physics (physics.comp-ph); Machine Learning (stat.ML)
*备注: 32 pages, 5 figures, 9 tables
Abstract:Post-hoc group averaging (or test-time augmentation) is a simple way to impose symmetry on a trained predictor without changing its architecture or retraining it. For finite and compact groups, averaging over a normalized Haar measure on the group gives the Reynolds operator: a canonical projection onto invariant or equivariant predictors with known non-increasing risk guarantees under certain assumptions. This paper provides numerical experiments on physics emulation models that demonstrate the effectiveness of the approach in this setting. We then extend this framework to non-compact group symmetries. For non-compact groups, the Haar measure cannot be normalized, so group averaging cannot be defined in the same way. We replace it with three alternatives: the Weyl unitary trick (which requires complexifying the space and extending the predictor), a data-dependent orthogonal projection (which involves a non-trivial optimization on the space of invariant or equivariant functions), and a frame averaging approach (which involves mapping each point to a set of orbit representatives). We illustrate the constructions theoretically and empirically on the unit-rescaling group and the Lorentz group. Our results show that in the non-compact case, the Weyl trick and frame averaging do not provide a consistent improvement over the baseline, except where the test data is out-of-distribution (an orders-of-magnitude improvement in this case) and, for the Weyl trick, in low-data regimes. The orthogonal projection can provide improvements in-distribution and out-of-distribution. We also observe that models trained directly on invariant features matched or beat the projection of an unconstrained one. We conclude that post-hoc symmetrization extends to non-compact groups, but it does not consistently outperform training a symmetric model.
[LG-67] Action-Centered Causal Bandits for Budget-Constrained Promotion Targeting WSDM2027
链接: https://arxiv.org/abs/2611.08990
作者: Yifan Lin,Kaixu Yang,Di Mo
类目: Machine Learning (cs.LG)
*备注: Accepted at the 20th ACM International Conference on Web Search and Data Mining (WSDM 2027)
Abstract:Web platforms select among costly interventions for each user to maximize incremental outcomes under a global budget. We formulate this problem as a budget-constrained contextual bandit with heterogeneous treatment effects. Although an action-independent outcome baseline cancels from treatment comparisons, modeling a nonlinear baseline can increase statistical complexity. We propose DR-CATE-CBwK, an action-centered causal bandit that directly learns treatment contrasts. Its online update remains valid under baseline-model misspecification without inverse-propensity weighting. It combines an offline doubly robust learner to warm-start treatment-effect estimates, an online action-centered estimator for sequential contrast learning, and a primal-dual controller with an affordable-action rule to enforce the effective-cost budget pathwise. For linear conditional average treatment effects (CATEs) in a stationary i.i.d. setting, we prove a high-probability (\widetildeO(Kd\sqrtT)) regret bound, where K is the number of priced actions, d the context dimension, and T the horizon. We benchmark against the fluid linear program (LP) under the coverage-guaranteed action rule. We further show that the action-centered estimating moment remains unbiased under any predictable baseline estimate, whereas a specified separate-arm reward regression can exhibit asymptotic contrast bias under action-dependent logging and shared-baseline misspecification. On a semi-synthetic job-marketplace benchmark with real logged features and simulated outcomes, our practical Thompson-restricted variant achieves the lowest regret among evaluated budget-feasible methods in the operating regime. It reduces regret by approximately 20% versus the strongest reward-based baseline, outperforming raw-reward contextual bandits, BwK baselines, and a static offline policy, while never exceeding the effective-cost budget. Comments: Accepted at the 20th ACM International Conference on Web Search and Data Mining (WSDM 2027) Subjects: Machine Learning (cs.LG) Cite as: arXiv:2611.08990 [cs.LG] (or arXiv:2611.08990v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.08990 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Yifan Lin [view email] [v1] Fri, 6 Nov 2026 05:01:18 UTC (69 KB)
[LG-68] VLA Handover: Failure-Prediction-Based Switching from On-Robot to Server VLA Models
链接: https://arxiv.org/abs/2611.08987
作者: Takayuki Nishio
类目: Robotics (cs.RO); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注: 6 pages, 4 figures, 3 tables. Submitted to IEEE ICC 2027
Abstract:Large vision-language-action (VLA) models generalize well but are difficult to run in real time on a robot, and small VLAs that run there fail more often. Remote inference over beyond-5G and 6G networks lets a robot use a large VLA on a server, but querying the server for every action chunk limits the number of robots per server. We propose VLA handover: the robot runs a small VLA (S-VLA) and switches to a large VLA (L-VLA) on a server only when the S-VLA is likely to fail. We formulate the switching as minimizing the request cost under a required success probability and, since this problem cannot be solved directly, propose a heuristic framework that retrains neither VLA. In simulation with SmolVLA and \pi_0.5 on three of the four LIBERO suites, VLA handover raised the success rate by 11-22 percentage points over the S-VLA alone and sent 49-77% fewer requests than server-only control at a 2-4 percentage-point lower success rate, and its completion time grew less with the network latency.
[LG-69] Energy-Based Reason ers: Your Recurrent Reason er Should Be an Energy-Based Model
链接: https://arxiv.org/abs/2611.08977
作者: William Xingxu Chen,Alexi Gladstone,Zeyu Michael Li,Heng Ji,Yilun Du
类目: Machine Learning (cs.LG)
*备注:
Abstract:Looping is a powerful inductive bias for reasoning. By repeatedly iterating a recurrent block, looped models can scale their effective depth and enable multi-step reasoning in latent space. However, existing looped models are not inherently guided by an explicit measure of progress, making it difficult to determine when to stop looping and how to improve upon the current state. To enable this, we introduce Energy-Based Reasoners (EBRs), which learn an energy function over latent reasoning states and formulate looped updates as gradient descent on this energy. The energy scores latent states, while its gradient guides refinement and provides a signal for stopping. To train EBRs effectively, we further introduce exploration-based training to discover additional modes and to better shape the energy landscape over a broader set of states. Across four algorithmic reasoning benchmarks, EBRs match or outperform prior recurrent reasoning models, achieving 99.9% accuracy on Sudoku-Extreme while using 345 \times less inference compute than the next-best model and maintaining comparable training compute. On all multi-solution benchmarks, EBRs achieve state-of-the-art coverage of distinct valid solutions while maintaining near-100% single-sample accuracy. These results suggest that energy-based models are a promising approach for strong reasoning performance.
[LG-70] Concordance-Optimal Continuous Music Emotion Estimation: Variance Decomposition Dual-Branch Temporal Fusion and Conformal Prediction Intervals
链接: https://arxiv.org/abs/2611.08966
作者: Mohammad Faraz Akhtar,Aditya Parashar,Ayush Jain,Madhur Gupta,Ritesh Kumar,Soubhik Chakraborty
类目: ound (cs.SD); Machine Learning (cs.LG)
*备注: 28 pages, 12 figures
Abstract:Continuous music emotion recognition (MER) estimates how Arousal and Valence evolve within a song, yet the field’s standard metric, pooled frame-level correlation, is dominated by between-song differences and cannot distinguish a model that tracks emotional dynamics from one that only ranks songs by their average level. We decompose annotation variance into between- and within-song components and show that mean squared error (MSE) targets the former, while windowed Lin’s concordance correlation coefficient (CCC = r \cdot C_b ) targets the latter by penalising the variance compression MSE induces. A single weight \lambda controls the trade-off: on 1,764 DEAM songs, raising \lambda from 0 to 1 lowers pooled Arousal r from 0.843 to 0.810 while raising mean within-song r from 0.204 to 0.368. Building on this objective, we present MMAS (Multimodal Music Analysis System), an open-source model that predicts Arousal and Valence every 0.5 s from raw audio. It fuses frozen MERT-v1-95M embeddings with ten symbolic features from automatic transcription through a dual-branch temporal convolutional network with cross-modal attention. Because both branches are audio-derived, one model trains jointly on DEAM and PMEmo without domain conditioning. Under song-grouped 5-fold cross-validation (2,531 songs), pooled r reaches 0.801/0.685 (Arousal/Valence) on DEAM and 0.742/0.746 on PMEmo. Against a ridge baseline on identical features, within-song Arousal r roughly triples (0.344 vs. 0.125; n = 353 songs, Wilcoxon p = 1.7 \times 10^-21 , d = 0.56 ). Complete transcription failure costs at most 0.04 pooled r . Split-conformal intervals attain 93.7% (Arousal) and 90.0% (Valence) marginal coverage at a 95% target, with the shortfall traced to within-song dependence. MMAS is released as a pip-installable package. Comments: 28 pages, 12 figures Subjects: Sound (cs.SD); Machine Learning (cs.LG) Cite as: arXiv:2611.08966 [cs.SD] (or arXiv:2611.08966v1 [cs.SD] for this version) https://doi.org/10.48550/arXiv.2611.08966 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-71] Accuracy Is Not Value: Choosing Weak Supervisors by What the Student Inherits
链接: https://arxiv.org/abs/2611.08959
作者: Romano Li,Xunfan Cai
类目: Machine Learning (cs.LG)
*备注:
Abstract:Weak supervisors are usually judged by their own errors, but what matters is what a student learns from them. For squared loss and any linear learner shared by all supervisors, we show that a supervisor’s value is determined by its errors after they pass through the student’s learning operator. Consequently, the inner products of the supervisors’ raw errors can be compatible with every feasible value of the student’s risk, and weighting supervisors by them can be arbitrarily worse than the best weighting in relative excess risk. This student-relative value is not identifiable from unlabeled data, but what is missing is small: one finite-dimensional projection of the target decides whether training on a weak aggregate helps, another decides how to weight the supervisors. Known teacher training rules, or a label-model condition stated in student-relative form, shrink it further. Clean labels close the remaining gap at a cost set by the geometry of the unresolved directions rather than by the dimension of the student’s representation, with matching worst-case lower bounds. In Gaussian designs we also quantify when exact teacher structure beats ordinary stacking and when weak supervisors, combined with clean labels, beat every procedure that uses the clean labels alone. The guarantees extend to fitted ridge students and finite unlabeled data.
[LG-72] Surrogate Quality Score: A Novel Composite Performance Metric for Model Selection
链接: https://arxiv.org/abs/2611.08937
作者: Maaz Ahmad,Iftekhar A. Karimi
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)
*备注: 22 pages, 6 figures, 2 tables
Abstract:The growing reliance on high-fidelity, state-of-the-art machine learning based models for engineering decision-making necessitates robust and systematic approaches for model selection. While high predictive accuracy at untrained points is a necessary trait in the final model, other attributes such as model complexity, forecastability, explainability, scalability, and physics-awareness may be equally important or relevant criteria depending on application context and user requirements. In this work, we propose a novel composite performance metric, Surrogate Quality Score (SQS) for model selection based on predictive accuracy and model complexity simultaneously. SQS heuristically integrates these two attributes through a weighted product of normalized measures, yielding a unified score to reflect overall model quality. A user-specified parameter in SQS enables explicit control over the accuracy-complexity trade-off, thereby incorporating human judgement and preference in the model selection process. We extensively discuss and justify the formulation of SQS and demonstrate its utility through an illustrative example. By combining predictive accuracy and model complexity, SQS offers a systematic and objective approach for model selection based on both attributes.
[LG-73] Grid-Oriented Probabilistic Load Forecasting for AI Data Centers
链接: https://arxiv.org/abs/2611.08929
作者: Qingyi Yao,Ding Lin,Tianqiao Zhao,Meng Yue,Jianhui Wang
类目: ystems and Control (eess.SY); Machine Learning (cs.LG)
*备注:
Abstract:Probabilistic load forecasts for AI data centers need to represent both uncertainty under changing workloads and temporal dependence to support grid operations. This paper develops an open, grid-oriented benchmark using measured GPU power from a 448-GPU AI cluster. Forecasts are evaluated through four stages: probabilistic prediction, online uncertainty calibration, temporal scenario generation, and horizon-wide reserve sizing. Following a change in cluster utilization, a quantile LightGBM forecast achieves only 63.4% coverage for a nominal 90% interval, while online regime-conditioned calibration increases coverage to 87.3%. Even with well-calibrated marginal forecasts, independent sampling produces unrealistic ramp distributions and excessive reserve procurement. Reconstructing temporal dependence from historical forecast errors reduces the discrepancy in ramp statistics and lowers realized reserve cost by 7.4-10.6% across three forecasting tasks. These findings show that marginal forecast quality alone is insufficient for operational decisions and motivate joint evaluation of calibration, temporal structure, and decision outcomes.
[LG-74] Composable Causality: A Toolkit for Systematic Time-Series Causal Discovery and Treatment-Effect Benchmarking NEURIPS2026
链接: https://arxiv.org/abs/2611.08918
作者: Muhammad Hasan Ferdous,Md Osman Gani
类目: Machine Learning (cs.LG)
*备注: 30 pages, 10 main text pages plus references and appendices, with figures and tables. Accepted to the NeurIPS 2026 Evaluations and Datasets Track. Code and data at this https URL
Abstract:Real-world multivariate time series rarely come with verified ground-truth causal structure, so research on time-series causal discovery and treatment-effect estimation depends on synthetic data. Existing tools are either fixed datasets that cannot be modified after release or script-based generators in which each script fixes one combination of mechanism, confounding, and missingness. Neither supports controlled experiments that vary a single property, such as sparsity, noise distribution, or missingness mechanism, while holding the others fixed. Counterfactual evaluation has the same gap, since existing simulators are built per paper with hardcoded mechanisms or produce observational data only. We introduce DTCD (Datasets and Toolkit for Causal Discovery), a data generation library that produces multivariate time series with ground-truth contemporaneous and lagged adjacency matrices. Topology, mechanism, noise, observational mask, and intervention policy are independent components that compose through configuration alone. Paired counterfactual twins, produced by caching the stochastic realization during the factual rollout and re-injecting it during the interventional rollout, recover individual treatment effects to numerical precision. A reference benchmark of eight discovery algorithms on eighty-one configurations and four treatment-effect estimators on eighteen configurations exposes failure modes that aggregate metrics hide. These include the inability of standard discovery methods to handle missing values, regardless of whether the missingness mechanism is informative, and a marginal-variance shortcut (varsortability) that we measure and remove by standardization. Code and data: this https URL
[LG-75] SSF-Net: A Spatial-Spectral Fusion Network for Accurate PDE Operator Learning
链接: https://arxiv.org/abs/2611.08912
作者: Siqi Zhao,Flora D. Salim
类目: Machine Learning (cs.LG)
*备注:
Abstract:Machine learning methods for solving partial differential equations (PDEs) significantly improve computational efficiency over traditional numerical solvers such as finite element method (FEM). However, PDE dynamics inherently span both local and global dynamics: disturbances propagate and decay gradually through the domain, while global coupling effects can influence the whole system simultaneously. Capturing both types of dynamics with a low complexity model remains a key challenge. To address this, we propose the Spatial-Spectral Fusion Network (SSF-Net), built around a dual-branch Spatial-Spectral Fusion (SSF) block comprising a spatial branch and a spectral branch. The spatial branch employs a multi-directional Mamba, which mitigates its inherent causal, single-direction scanning bias and capture local interactions with reduced directional bias at linear complexity. The spectral branch leverages a Fourier transform with a block-diagonal MLP to efficiently capture global spectral features with reduced parameter count and complexity. The SSF block is embedded in a multiresolution convolutional encoder-decoder, which extracts multi-scale features. We evaluate SSF-Net on a comprehensive PDE benchmark spanning multiple task types and assess its long-horizon stability. Compared with attention-based and convolution-based baselines, SSF-Net achieves lower prediction error with fewer parameters while retaining near-linear asymptotic complexity in the number of spatial tokens.
[LG-76] ask-Contribution Dynamics of Adam in Physics-Informed Neural Networks: Slow-Manifold Analysis and Gradient Modulation
链接: https://arxiv.org/abs/2611.08910
作者: Enzhan Zheng,Xiaoliang Song
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG)
*备注:
Abstract:Physics-informed neural networks (PINNs) often suffer from imbalanced optimization of competing loss components, particularly when approximating solutions with high-frequency or multiscale features. Although existing balancing strategies have proven effective in alleviating this difficulty, the mechanisms governing the evolution of individual task contributions to Adam’s parameter updates remain insufficiently understood. We introduce task-wise indicators to characterize this evolution and identify three distinct training regimes: smooth variation, intermittent spikes, and persistent oscillations. Focusing on the early smooth regime, we combine discrete-time analysis with a continuous-time singular perturbation framework under the full-batch setting and the parameter regime (1-\beta_2)/(1-\beta_1)\ll1 . A first-order approximation of the locally attracting slow manifold reveals two factors governing the step-to-step redistribution of task contributions: curvature-related alignment disparities and differences in the coordinatewise distributions of preconditioned task gradients. Both factors involve Adam’s shared second-moment estimate. Inspired by AdaTask, we propose Gradient-Modulated Adam (GMAdam), which modulates incoming task gradients using task-specific second-moment estimates. Numerical experiments on high-frequency Poisson and wave equations, the Allen-Cahn equation, and two-dimensional lid-driven cavity flow ( Re=100,400 ) demonstrate consistent improvements in predictive accuracy over AdaTask at comparable computational cost.
[LG-77] Learning to Restart: Differentiable Restart Policies for Optimization Dynamics
链接: https://arxiv.org/abs/2611.08883
作者: Yiman Fong,Zhonglin Xie,Zaiwen Wen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Adaptive restart mechanisms, which reset the momentum or other auxiliary state of an iterative optimizer, play a central role in the empirical performance of accelerated first-order methods, yet existing restart rules remain almost entirely hand-designed. The obstruction to learning restart policies is that, unlike the typical Learning to Optimize (L2O) setting, the discrete restart events make the resulting trajectory non-differentiable in the algorithm parameters and thus block gradient-based optimization of the restart policy. We resolve this obstruction by developing a differentiable framework for restart policies through their continuous-time limit. We introduce surrogate gradients for the restart events and optimization trajectories that are efficient to compute and provably converge to their continuous-time counterparts. We validate the effectiveness of our surrogate gradients in L2O settings, training restart policies that improve upon standard handcrafted adaptive restart criteria on quadratic and linear programs, and that transfer zero-shot from synthetic quadratics to logistic regression on real datasets.
[LG-78] Extending Grid Foundation Models for Phasor-Domain Dynamic Simulation
链接: https://arxiv.org/abs/2611.08867
作者: Boyang Zhao,Yifan Zhou,Alban Puech,Tilman Bockhacker,Hendrik Hamann
类目: ystems and Control (eess.SY); Machine Learning (cs.LG)
*备注:
Abstract:Grid Foundation Models (GridFMs) are emerging as a powerful framework for learning and transferring knowledge across power grids and grid analysis tasks. This work extends GridFMs from steady-state analysis to dynamic simulation. We present DynGridFM, which uses component-wise tokenization and a structured latent operator to learn shared dynamic representations across heterogeneous grids and enable transfer to unseen systems. Pretrained on nine grids, DynGridFM is adapted to a held-out 48-machine grid, where fine-tuning on 64 trajectories achieves accuracy comparable to training from scratch on 128. On a single GPU, it achieves 924x to 1232x speedup over numerical simulators on one CPU core.
[LG-79] Reconstructing LLM Lineage Trees via Fine-Tuning Path Inference ICLR2027
链接: https://arxiv.org/abs/2611.08866
作者: Sota Ogawa,Shojiro Yamabe,Daiki Nishiyama,Jun Sakuma
类目: Machine Learning (cs.LG)
*备注: Under review at ICLR 2027
Abstract:Fine-tuned (FT) derivatives of large language models (LLMs) are now released in large numbers, yet their provenance relies on self-reports and cannot be verified. Misattributed lineage leaves a verifier unable to audit outsourced fine-tuning, confirm license compliance of derived models, or trace vulnerabilities that a parent passes to its children. All three require the direction of derivation, not merely its presence. Existing output-only methods are limited to pairwise derivation tests or undirected lineage estimation, while recovering a directed tree requires white-box access to model weights. We formalize the problem of reconstructing a directed model tree, including internal nodes, under gray-box access, where only teacher-forced log-likelihoods are observed and weights are never inspected. Fine-tuning raises the log-likelihood of the training data, and a child inherits this increase from its parent. Building on this property, we define an asymmetric per-task log-likelihood deviation between any two models as a trace of fine-tuning history, and derive from it an edge weight for each ordered pair of models. We propose TRACE-Edge, which exactly solves for the maximum spanning directed tree under these edge weights, with neither hyperparameters nor tree-structure constraints. We evaluate it on five configurations totaling 25 model trees built from three base models and seven tasks. TRACE-Edge achieves a parent-match accuracy of 0.74 when probed with the fine-tuning examples, and on average outperforms an output-similarity baseline even with held-out probes drawn from the same distribution as the training data.
[LG-80] STRIDE: Learning Humanoid Locomotion in Minutes with Acceleration-Based Rewards
链接: https://arxiv.org/abs/2611.08857
作者: Nico Bohlinger,Jan Peters
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Learning locomotion directly on a humanoid robot is a promising direction to avoid the sim-to-real gap. However, locomotion rewards in Reinforcement Learning (RL) typically track the base velocity, which cannot be measured onboard without an external motion capture system or well-calibrated state estimator. This is why we introduce \method, which learns humanoid locomotion from minutes of samples with a reward that only requires the onboard Inertial Measurement Unit (IMU). Instead of tracking the base velocity, STRIDE rewards the robot for following a commanded acceleration profile. To learn efficiently from this reward, we use the off-policy RL algorithm XQC and combine a high gain setting for the joint position controllers with left-right symmetry augmentation. In simulation, STRIDE learns to walk on the Unitree G1 and Booster T1 humanoids within 14 minutes of samples. Compared to learning without these two components, it reduces the falls during training by 81%, while without them the robot learns only in over an hour.
[LG-81] Functional Dependency beyond Similarity: Hierarchical Decomposition of Training Data Attribution in Music Generative Models
链接: https://arxiv.org/abs/2611.08839
作者: Changheon Han,Hugh Xuechen Liu,Ashkan Panahi,Kıvanç Tatar
类目: Computers and Society (cs.CY); Machine Learning (cs.LG)
*备注: 23 pages, 9 figures, 11 tables. Under review
Abstract:Copyright analysis focuses on the similarity between two works. However, generative AI models learn from training data through functional relations beyond similarity. Training Data Attribution (TDA) estimates the influence of each training instance on a generated output and may serve as a basis for remuneration. However, prior studies find global influence, where a few training instances rank as the most influential for almost every generation. Used directly, TDA scores would distribute part of every remuneration to these instances regardless of what was generated. To address this, we propose a new approach in which TDA scores are decomposed into three levels of influence, namely global indicating the influence of a training data to a trained model, group for its influence on a group of generated outputs (that could be clustered together), and specific for its influence on a single generated output. Our study finds that, across several music generative models and TDA methods, global- and group-level influence is non-negligible in every case. Based on these observations, we propose the term functional dependency for the relation of training data to a trained model or to its outputs. This notion both encompasses and distinguishes dependencies of different scopes, allowing TDA scores to be interpreted as distinct forms of contribution in legal, economic and ethical discussions on remuneration.
[LG-82] MotionBench-XAI: Benchmarking What Shapes Shapley Explanations in Spatiotemporal Machine Learning
链接: https://arxiv.org/abs/2611.08797
作者: Mathilde Papillon,Louisa Cornelis,Nina Miolane
类目: Machine Learning (cs.LG)
*备注:
Abstract:Predicting outcomes from spatiotemporal data, such as assessing Parkinson’s severity from human motion, often requires explainability in addition to accuracy. Practitioners need to know what drove the prediction, whether it be spatial (a body joint), temporal (a specific window), or both. Shapley values are attributions computed by treating input features as players in a cooperative game and measuring how the model output changes as coalitions of players are removed. Computing them involves choosing what counts as players, how absent players are imputed, and how the game is evaluated. How these choices affect explanation quality is poorly understood for spatiotemporal data, and real data carries no ground-truth attributions to check against. We present MotionBench-XAI, a benchmark built on i) five synthetic spatiotemporal datasets with exact ground-truth Shapley values, ii) three real datasets across clinical gait, electrocardiogram, and audio, iii) ten attribution methods across three player sets, and iv) three trained architectures. We show that grouping features into players, across time, space, or both, matters more than the choice of imputer. Generative imputers improve explanations only when their conditional models are accurate. On real data, the standard proxy metrics select contradictory best methods. We distill these findings into practical insights and provide an open-source pipeline.
[LG-83] opoSHAP: Shapley for Explainable and Efficient Topological Deep Learning
链接: https://arxiv.org/abs/2611.08789
作者: Mathilde Papillon,Nina Miolane
类目: Machine Learning (cs.LG)
*备注:
Abstract:Topological deep learning (TDL) extends learning on graphs, such as molecules or social networks, to higher-order structures, such as rings in a molecule or groups in a social network. Explaining which of these structures drive a model’s predictions, and which architectural components matter most for its performance, remains challenging. We introduce TopoSHAP, an efficient and general explanation framework that addresses both questions across topological domains and model architectures. TopoSHAP-cell assigns Shapley values to topological data cells and extracts compact explanations. These explanations match or outperform graph-based counterparts on GraphXAI benchmark data while running up to 112 \times faster. They also reveal advantageous topological feature design, leading to measured improvements in explanation recovery and predictive accuracy. On the architecture side, TopoSHAP-neighbor explains how neighborhoods contribute to model performance, and turns these evaluations into a backward-elimination search that identifies smaller, competitive architectures early in training. This is up to 11.3\times cheaper than traditional TDL training sweeps.
[LG-84] MEDALLION: Manifold Embedding Distillation and Attributions for Low-dimensional Latent InterpretatiON
链接: https://arxiv.org/abs/2611.08767
作者: Vicky Jung,Claire M. He,Irene Chang,Genevera I. Allen
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Dimensionality reduction is widely used to visualize high-dimensional data and guide scientific discovery, yet interpreting which original features contribute to low-dimensional representations remains challenging, particularly for black-box methods such as t-SNE and UMAP. We introduce \textbfMEDALLION (Manifold Embedding Distillation and Attributions for Low-dimensional Latent InterpretatiON), a novel framework for interpreting low-dimensional embeddings through knowledge distillation and neural feature attribution. MEDALLION distills any given embedding with a neural encoder which we explain via a broad range of existing neural feature attribution methods. Through experiments on MNIST and applications in single-cell transcriptomics, we show MEDALLION provides informative feature attributions that cast additional insights into the underlying embedding and compare its performance with existing interpretability approaches.
[LG-85] On Differentially Private Mechanisms for Linear Regression NEURIPS2026
链接: https://arxiv.org/abs/2611.08737
作者: Bardiya Aryanfard,Monika Henzinger,Farhood Rostamkhani
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS)
*备注: Accepted at NeurIPS 2026
Abstract:We study the classic least-squares linear regression problem under differential privacy (DP) in the batch and the continual observation setting. In the latter setting, the mechanism receives a stream of n data points (x_i,y_i) ( x_i\in\mathbbR^d and y_i\in\mathbbR ) and must output an estimate of the optimal parameters at each time step i . Although differentially private linear regression has been extensively studied in the batch setting, the problem remains underexplored for the continual setting, with prior work being restricted to constrained linear regression and update time polynomial in n . In this work, we introduce ContinualAdaSSP, the first mechanism for unconstrained linear regression under continual differential privacy. Our mechanism’s suboptimality gap is proportional to \sqrtd\operatornamepolylog(n) . It requires O(d^2\logn) space and O(d^2\logn+d^\omega) update time, where d is the dimension of the input space. We also study fundamental limits of differentially private linear regression by showing the first lower bounds for the box-constrained least-squares linear regression problem. Our lower bounds are proportional to \min(n,\frac\sqrtd\log^1.5d) for (1,O(\frac1n)) -DP, and \min(n,\fracd\epsilon) for \epsilon -DP, and are derived using a reduction from a fingerprinting code problem from [Peter et al., 2024], providing new insights into the inherent hardness of this canonical task. Our experimental evaluation on real-world datasets shows that the accuracy of our continual observation mechanism clearly outperforms any other known differentially private mechanism for this problem such as rerunning the batch algorithm at each time step. Comments: Accepted at NeurIPS 2026 Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS) Cite as: arXiv:2611.08737 [cs.LG] (or arXiv:2611.08737v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.08737 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Bardiya Aryanfard [view email] [v1] Thu, 5 Nov 2026 21:37:04 UTC (1,988 KB)
[LG-86] Foundation-Model Residuals as Signals of Urban Bus Irregularity
链接: https://arxiv.org/abs/2611.08703
作者: Csaba I. Sidló,Jamil J. Ghazal,Domonkos Rózsay,Rita Rónai,András J. Molnár
类目: Machine Learning (cs.LG)
*备注: 2-page extended abstract accepted, but not presented as a poster at NetMob 2026
Abstract:Pretrained trajectory models encode mobility patterns learned across cities, but their errors when transferred to a new city are usually treated only as performance limitations. We instead ask whether these residuals reveal systematic local conditions. Architectures such as UniTraj and TrajFM learn transferable trajectory representations, although cross-context transfer remains challenging. Using the NetMob26 bus location, ticketing, weather, network, and socioeconomic data for Niteroi, we analyze spatial and travel-time residuals of these recent trajectory foundation models. We test whether high residuals concentrate spatially, whether spatial and travel-time residuals overlap, and which urban-context variables explain their distribution. We interpret high residuals as candidate proxy signals of transit irregularity rather than direct evidence of service disruption, and analyze their contextual associations using Explainable AI (XAI). Insights into such varied environments are increasingly valuable for evaluating new urban paradigms; for instance, the 15-minute-city perspective serves as a key motivation for our approach, guiding our methods and expanding the analysis beyond standard mobility metrics. Comments: 2-page extended abstract accepted, but not presented as a poster at NetMob 2026 Subjects: Machine Learning (cs.LG) Cite as: arXiv:2611.08703 [cs.LG] (or arXiv:2611.08703v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.08703 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-87] Does a Trading Policy Need to Know What It Trades? Microstructure Invariance and Zero-Shot Transfer of Execution Policies
链接: https://arxiv.org/abs/2611.08701
作者: Hieu Tran,Hong Yu
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE); Trading and Market Microstructure (q-fin.TR)
*备注: 33 pages, 10 figures
Abstract:Execution algorithms are calibrated asset by asset, because the queues, trade rates and bid-ask spreads they face differ by orders of magnitude across assets. For reinforcement learning this is zero-shot generalisation across contexts of very different scales, and microstructure invariance proposes a normalisation fixed by theory: measured in each asset’s own units of time, size and price, markets look alike. In an exact order-by-order replay of public Nasdaq data, we train policies once, on 954 stocks in 2019, to choose between waiting in the queue and crossing the spread, and evaluate them unchanged on 1,964 unseen stocks in 2020-21, 2,629 stocks in 2025-26, crypto pairs and a second exchange. First, a policy that sees a stock only through its recent trade rate, trade size, spread and price increment (tick) costs within 0.015 spreads per share (a fifth of what learning saves) of a policy whose training data included the stock, and the stock’s own history does not help it; it comes within a similar margin of policies trained on the second exchange when every episode counts equally, and within a wider one on crypto. Second, in mean cost before fees, invariant units beat raw units, z-scores and asset descriptors, but by under a basis point: most of their advantage is information and a common cost scale that a learner can also take as plain inputs, and none of the representations tested extrapolates to unseen spread regimes. Third, a near-zero transfer gap to a pooled comparator is weak evidence that a representation is right: we show that such a comparator detects only a small share of what a poor representation loses, and confirm it in synthetic markets and on pendulums.
[LG-88] Enabling Retraining-Free Accurate Inference in Analog Compute in Memory Accelerators
链接: https://arxiv.org/abs/2611.08693
作者: George Karfakis,Puneet Gupta
类目: Hardware Architecture (cs.AR); Machine Learning (cs.LG)
*备注: 21 pages, 18 figures, 11 tables
Abstract:Analog compute-in-memory (CIM) accelerators compute matrix-vector products directly inside memory arrays, eliminating weight movement and promising order-of-magnitude gains in inference efficiency. However, converter and device errors make pretrained models unusable on these arrays, and hardware-aware retraining is costly. We enable retraining-free, accurate inference of pretrained models on simple analog arrays with low-cost, shared calibration. Our key observation is that analog error depends on how each layer is represented on the array, and that this representation can be changed offline without changing the computation. We center the inputs and apply Hadamard rotations designed for analog arrays, which keep shared reference errors spread and confine input mixing to each array. Across CNNs, vision transformers and LLMs up to 14B parameters on PCM and ReRAM, the method brings vision models and larger LLMs within a few points of floating-point accuracy and matches or approaches published hardware-aware training without training any weights. On ReRAM with 6-bit ADCs, it restores ViT-B/16 from 37.4% to 82.9% and Phi-4 from 27.0% to 72.2%, both within 2.2 points of floating-point accuracy.
[LG-89] Higher Order Component Attribution via Learned Surrogates in High Energy Physics
链接: https://arxiv.org/abs/2611.08656
作者: Lauri Laatu,Aritra Bal,Benedikt Maier,Markus Klute,Michael Spannowsky
类目: Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); High Energy Physics - Phenomenology (hep-ph)
*备注:
Abstract:Experimental High Energy Physics (HEP) has adopted the transformer architecture for identification of jets, particularly the state-of-the-art Particle Transformer (ParT) which injects explicit information about particle correlations into the attention mechanism. While able to achieve state-of-the-art results in accuracy, interpretability to its effectiveness has been limited. In this paper we propose component attribution methods using Kolmogorov-Arnold Networks (KANs) that directly tie back to the physical descriptors of jets, to probe whether mechanisms such as Multi-Head Attention (MHA), or the ParT architecture, are able to learn these as discriminating features.
[LG-90] Benchmarking Machine Learning for Chaotic Three-Body Dynamics in Astrophysics NEURIPS2026
链接: https://arxiv.org/abs/2611.08634
作者: Sheng Wan,Pavan Vynatheya,Duo Xu,Li Kevin Wenliang,Ye Zhu
类目: Machine Learning (cs.LG)
*备注: Accepted at the NeurIPS 2026 Workshop on Sim2Science: ML with Imperfect Scientific Models
Abstract:Machine learning (ML) and generative AI have recently made impressive progress in understanding physics. We explore a hierarchy of prediction problems stemming from chaotic three-body systems and benchmark existing ML approaches, including general-purpose prediction algorithms and those that build on top of foundational physics models. We first simulate the trajectories of triple systems near the critical boundary of stable/unstable initial conditions. We then design a hierarchy of learning tasks, focusing mainly on two broad problems: stability prediction and trajectory prediction. The former classifies whether a given triple system has diverged or not up to a specific horizon, while the latter predicts the exact trajectory, both given suitable input features. We benchmark a range of ML approaches, from simple regression methods to state-of-the-art foundational physics models, across these tasks, and compare them against astrophysical methods such as the Mardling-Aarseth stability criterion. Our experiments reveal a clear order of task difficulties in chaotic three-body dynamics: early dynamics enable strong stability prediction, while precise long-horizon trajectory forecasting remains difficult. Code and dataset are publicly available at this https URL.
[LG-91] AutoPYara: Next-Gen YARA Rule Generator for Malware Family Clustering ACSA
链接: https://arxiv.org/abs/2611.08608
作者: Mabon Ninan,Nhat Nguyen,Soumyajyoti Dutta,Sidharth Anil,Marcus Botacin
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: To Appear In ACSAC 2026
Abstract:Millions of malware variants threaten users daily, such that clustering these samples is key for fast incident response. AutoYara is the current state-of-the-art for automatic YARA rule generation, but the task remains challenging on realistic datasets, where intra-family diversity and noisy clustering can degrade rule quality. We present AutoPYara, a solution armored with: (1) precise cluster centroid selection based on similarity hashing, and (2) an improved heuristic for estimating subclusters within a malware family using historical VirusTotal data. We implement AutoPYara as an open-source Python package (pip install autopyara) to support reproducibility and practical use. We show that AutoPYara, (1) increases rule coverage by 14 percentage points; (2) reduces variance by 7.1 percentage points, comparing the average variance AutoPYara exhibits to that of AutoYara; and (3) improves threat-hunting accuracy by 10%. AutoPYara produces robust and effective YARA rules than prior automatic approaches in practical settings.
[LG-92] Not All Misalignment Is Alike: A Decomposed Benchmark for Emergent Misalignment
链接: https://arxiv.org/abs/2611.08569
作者: Afonso Domingues,Mubashara Akhtar,Bhada Yun,Xin Chen,Lukas Fluri,Anna Hedström
类目: Machine Learning (cs.LG)
*备注:
Abstract:Emergent misalignment (EM), in which fine-tuning on a narrow task induces misaligned behaviour on unrelated prompts, has been demonstrated across a wide range of datasets, model families, and training setups. Yet, most studies continue to rely on the evaluation approach introduced in the original EM paper, producing rates that are sensitive to the small set of evaluation questions used, the choice of judge, and the thresholds applied. These evaluations can also conflate misalignment with capability loss and offer little insight into the specific ways misalignment manifests. Differences in datasets, models, and fine-tuning configurations further make findings difficult to compare across studies. We introduce EMERGE, a decomposed benchmark that tracks changes in eight distinct misaligned behaviours from base to fine-tuned models. It evaluates these behaviours on out-of-domain prompts against matched benign controls and tracks capability, coherence, and refusal to account for potential confounders. Applying EMERGE to 8 open-weight models and 13 fine-tuning datasets, we find that models with similar broad misalignment rates can exhibit very different behavioural profiles that a single score cannot capture. We hope EMERGE can help establish a shared and robust standard for measuring EM, making findings more comparable and helping move research from detecting the phenomenon towards understanding it.
[LG-93] Minimum-Sufficient Knowledge and Computation for Wind Energy Predictive Analysis
链接: https://arxiv.org/abs/2611.08558
作者: Purbak Sengupta,Aparna S. varde,Emil Jørgensen Njor
类目: Machine Learning (cs.LG)
*备注: To be published in IEEE BigData 2026 Conference
Abstract:Resource-constrained cyber-physical decision systems often apply fixed sensing, knowledge, and computational pipelines despite state-dependent information needs. We introduce Minimum-Sufficient Knowledge and Computation (MSKC), a resource-aware computing principle: select the least-costly route meeting decision-risk and deployment constraints, escalating only when cheaper routes are insufficient. As proof of concept, we isolate adaptive knowledge routing, the decision underlying Knowledge-Guided Machine Learning (KGML), which conventionally injects external knowledge unconditionally rather than only where needed. We instantiate this for wind ramp-event predictive analysis, an edge-oriented decision-support task, evaluating local, relational, and physics-informed routes with a state-aware gate on publicly available real data from two wind farms. We establish conditions for feasibility preservation, cost dominance, and strict expected savings over always-on routing. At a 35.9% relational-knowledge injection rate, adaptive routing is statistically indistinguishable from always-on injection on the reported metrics while reducing knowledge calls by 64%, supporting sustainable, Green-IT-aligned deployment. These results support MSKC’s premise that uniformly maximal intelligence is unnecessary. This offers intelligent data analysis for big data streams: the analysis itself decides, per state, how much knowledge and computation the data warrant.
[LG-94] Evolutionary Adaptive Metric Learning: A Multi-Objective Framework for Learning Local Distance Functions in Nearest Neighbor Classification
链接: https://arxiv.org/abs/2611.08549
作者: Isaac Touza
类目: Machine Learning (cs.LG)
*备注:
Abstract:Nearest neighbour classifiers inherit the geometry of their distance function. Fixed distances and globally learned metrics apply the same geometry throughout the input space, although relevant feature directions may vary across regions. We propose EAML (Evolutionary Adaptive Metric Learning), a framework that learns instance-dependent positive-definite metrics through a low-dimensional parameterization. Each local metric combines an input-dependent diagonal scaling, generated by softmax interpolation between anchor points, with a global low-rank correlation component. A real-coded elitist genetic algorithm optimizes the parameters using a scalar objective combining leave-one-out prediction error, robustness, metric smoothness, and regularization. We establish positive definiteness, metric equivalence, existence of an optimal solution, and Lipschitz continuity of the metric field. We further investigate prediction stability, generalization bounds, and evolutionary convergence under explicit assumptions. Experiments on synthetic data with region-dependent feature relevance show that EAML can recover local metric variations and outperform global metrics. Evaluation on thirteen public datasets using nested 5-fold cross-validation reveals a more nuanced picture: EAML improves performance on selected datasets but does not significantly outperform global metric baselines overall. Additional experiments show that global diagonal metrics can be preferable under nuisance features and test-time noise. These findings identify the conditions under which evolutionary local metric learning is beneficial and highlight its practical limitations.
[LG-95] What Are LLM Agents Actually Learning on Low-Budget Molecular Optimization Benchmarks?
链接: https://arxiv.org/abs/2611.08522
作者: Jiayu Qin,Zhengquan Luo,Changyou Chen,Zhiqiang Xu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Recent work reports large gains for LLM agents on low-budget molecular optimization. We measure the contributions of online optimization, retrievable task information, and permissive oracle landscapes. On PMO at 50 oracle calls, a GPT-4o mini agent’s sum AUC falls from 14.44 to 4.85 after task names and curated reference compounds are hidden. An identity-blinded condition that states only the oracle form (e.g., similarity to a hidden reference) scores 5.39, while a deterministic retrieval-only baseline with no LLM reaches 12.16 (84% of the full agent). A full 23-task 2x2 ablation and a three-seed replication on eight tasks support this interpretation. On a protein-targeted benchmark with an AutoDock Vina composite oracle, a single-edit random walk is a strong lower bound; across three seeds it obtains composite 0.636 +/- 0.008 and docking improvement -1.11 +/- 0.04 kcal/mol. As a positive control, a retrieval-free population agent exceeds REINVENT at a matched 50-call AUC budget, with its strongest result at 100 calls. These results motivate two inexpensive sanity checks: retrieval-safe PMO prompts and a random-walk floor for docking-based optimization.
[LG-96] Interpretable Multifidelity Learning with Kolmogorov-Arnold Networks for Spatiotemporal Nuclear Reactor Flow Modeling
链接: https://arxiv.org/abs/2611.08512
作者: Meredith Eaheart,Majdi I. Radaideh
类目: Machine Learning (cs.LG)
*备注: 33 pages, 15 figures,9 tables, and a preprint under peer review
Abstract:High-fidelity computational fluid dynamics (CFD) simulations of nuclear reactor systems are computationally expensive, motivating machine learning methods that reduce cost while preserving accuracy. Multifidelity surrogate modeling addresses this by combining inexpensive low-fidelity data with limited high-fidelity data. This work investigates whether the high-fidelity data required to predict the spatiotemporal mass flow rate across all 193 fuel assemblies of a pressurized water reactor (PWR) core can be reduced, comparing black-box multifidelity feedforward neural networks (MFNNs) with interpretable multifidelity Kolmogorov-Arnold Networks (MFKANs). Cross-fidelity correlation across the core is weak, making this a challenging multifidelity problem, and neither architecture predicts the transient particularly well in absolute terms. MFKAN nonetheless performs comparably to MFNN for one-time-step-forward prediction while also providing closed-form symbolic equations unavailable with MFNN. To address the scarcity of CFD transients, with only one available per fidelity level, TabDDPM-generated synthetic data was used to augment cross-fidelity forecasting. KAN and FNN achieve comparable self-evaluation accuracy (Pearson correlation of roughly 0.73–0.81), but both perform poorly at true cross-fidelity forecasting (correlations of 0.03–0.06). Assembly-level analysis shows that each assembly’s forecast correlation tracks its real cross-fidelity correlation for both architectures ( r=0.840 – 0.896 for KAN, r=0.804 – 0.903 for FNN), confirming that forecasting accuracy is fundamentally limited by how correlated the fidelity levels already are. Overall, KAN-based multifidelity modeling achieves accuracy consistently comparable to MFNN across all settings, while additionally providing symbolic equations that match or exceed the spline representation’s accuracy.
[LG-97] Does Country Matter? Evaluating Geographic Sensitivity of Large Language Models in Zero-Shot Chronic Kidney Disease Screening
链接: https://arxiv.org/abs/2611.08507
作者: Sirajam Munira,Muhammad Ashad Kabir
类目: Machine Learning (cs.LG)
*备注: Accepted to ACM HIKM-2026, 13 pages, 2 figures
Abstract:Large language models (LLMs) show strong zero-shot reasoning capabilities, yet their sensitivity to prompt context raises questions about reliability in deployment. This study investigates whether geographic context embedded in prompts influences LLM-based chronic kidney disease (CKD) screening, independent of any change in patient features. Using a community-based CKD dataset from Bangladesh (284 participants with 112 CKD and 172 non-CKD), we systematically varied the country attribute across 25 global contexts while holding all clinical inputs constant, and evaluated two open-weight LLMs (Qwen3 8B and Mistral 7B). Experiments were conducted on a common-correct subset of 156 participants for whom both models produced correct baseline predictions. Results show that non-CKD predictions were highly stable across geographic contexts, with both models maintaining consistent non-CKD predictions in nearly all settings. In contrast, CKD-positive predictions were considerably more sensitive. For Qwen3, CKD prediction dropped from a baseline of 65 to as low as 50 across country contexts, while Mistral showed smaller but non-negligible variation. Country contexts did not group according to known epidemiological or healthcare similarities; instead, groupings appear to reflect how country tokens are internally represented within each model. These findings demonstrate that geographic tokens can systematically shift clinical screening outcomes despite identical patient inputs, with implications for the deployment of LLM-based decision support in diverse healthcare settings.
[LG-98] Emergent Low-Dimensional Topological Dynamics Dictate Generalization and Escape Mechanics in Billion-Parameter Neural Networks
链接: https://arxiv.org/abs/2611.08502
作者: Dongqing Cao
类目: Machine Learning (cs.LG)
*备注: 16 pages, 6 figures
Abstract:The remarkable generalization of modern over-parameterized neural networks defies classical statistical learning bounds. Here we establish that optimization trajectories in billion-parameter models do not sample the ambient dimension \mathbbR^N homogeneously; instead, they are dynamically constrained to a low-dimensional execution submanifold \mathcalM_\textexec (with empirically observed effective dimension d_\texteff \le 12 ). By equipping the parameter space with a regularized Hessian metric g_ij = \nabla_i \nabla_j L + \epsilon \delta_ij and analyzing stochastic Langevin diffusion on loss sublevel sets, we derive a rigorous Topological Generalization Bound grounded in projected subspace persistent homology. Furthermore, by linking Morse index transitions to Betti number annihilation ( \beta_1 \to 0 ), we formulate an exact mathematical mechanism for exponential escape from strict saddles under anisotropic noise. We validate our framework across 80+ architectures (ResNets, Vision Transformers, and 7B LLMs). To bridge theoretical topology with computational feasibility, we resolve the curse of dimensionality by computing projected persistent homology proxies via randomized Lanczos iterations restricted strictly within active Hessian subspaces. Building upon this mathematical foundation, we introduce Topological-Adaptive Optimization (Top-SGD and Top-AdamW)—a family of wrapper algorithms that dynamically modulate learning rates using active curvature signatures. Extensive benchmarks demonstrate a 14.1% reduction in total wall-clock training time on ImageNet with all computational and memory overheads fully accounted for. Code is publicly available at this https URL.
[LG-99] Adaptive Depth Aggregation in Graph Transformers for Long-Range Graph Reasoning
链接: https://arxiv.org/abs/2611.08500
作者: Yifeng Lu,Zhen Wang
类目: Machine Learning (cs.LG)
*备注: 10 pages, 5 figures, 2 tables. Submitted to IEEE BigData 2026
Abstract:Graph Transformers improve long-range node-to-node communication through global or structure-aware interactions. However, improving cross-node communication does not determine how computations produced at different depths should be combined. Standard residual streams accumulate successive computations with fixed unit coefficients shared across nodes. This fixed aggregation is a residual-interface constraint: once intermediate contributions have been produced, the interface cannot explicitly reweight them by node. We hypothesize that this constraint can create a depth-wise aggregation mismatch when the relative utility of these intermediate contributions varies across nodes and structural contexts. To study this hypothesis, we formalize Graph Transformer residual computation as a causal node-wise history of graph-propagation and feed-forward network (FFN) contributions, making depth aggregation an explicit interface choice. We realize node-adaptive aggregation over this history with Attention Residuals (AttnRes) while preserving the backbone’s graph-propagation and auxiliary-state machinery. Block AttnRes reduces the routing cost of long histories relative to Full AttnRes by pooling consecutive contributions. We instantiate the interface in GPS and GRIT and evaluate it on all five ECHO tasks. Both variants achieve lower mean absolute error (MAE) than their independently tuned backbones on every task, and at least one obtains the lowest mean among all compared methods in each task. Routing-only donor-swap interventions support the mismatch hypothesis, while additional ablations examine the key design choices. Together, these results support node-adaptive depth aggregation as a useful complement to cross-node communication in long-range graph reasoning.
[LG-100] Generative Reconstruction of Full-Field Sensor States for Safety-Critical Digital Twins
链接: https://arxiv.org/abs/2611.08499
作者: Umme Mahbuba Nabila,Linyu Lin,Majdi I. Radaideh
类目: Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 27 pages, 8 figures, and 7 tables
Abstract:Accurate reconstruction of incomplete sensor fields is critical for state awareness in safety-critical digital twins (DTs). Thermocouples in high-temperature and radiation environments may drift, fail, or become intermittently unavailable, causing partial observability. This study develops a sparse-sensor reconstruction framework for Oregon State University’s High Temperature Test Facility (HTTF) and evaluates six methods under intra- and inter-experiment protocols: three conditional generative models (i.e., a diffusion model [DM], variational autoencoder [VAE], and generative adversarial network [GAN]), two deterministic recurrent neural-network models, and ridge regression. For intra-experiment reconstruction within the PG-29 depressurized conduction cooldown (DCC) test, the models combine graph-based spatial encoding with gated recurrent units. The graph neural network (GNN)–gated recurrent unit (GRU) VAE achieves the lowest error, with a mean absolute error (MAE) of 11.73,^\circ C, a root mean square error (RMSE) of 20.54,^\circ C, and an R^2 of 0.98. The deterministic GNN–GRU and plain GRU obtain RMSE values of 25.26,^\circ C and 35.22,^\circ C, respectively, demonstrating the benefit of spatial modeling. All three generative models reduce RMSE by more than half compared with ridge regression. The DM maintains an R^2 of 0.94 when known sensors decrease from 91 to 45. For inter-experiment reconstruction, models trained on PG-27 pressurized conduction cooldown (PCC) are evaluated on PG-29 using section-context conditioning. The section-context DM achieves the best transfer performance, with an RMSE of 42.84,^\circ C and an R^2 of 0.91. These results demonstrate the value of spatiotemporal generative modeling for missing-sensor reconstruction in safety-critical DT applications.
[LG-101] Embedding Consistent Recurrent Neural Networks in Finite Element Simulations for Path-Dependent Damage Prediction
链接: https://arxiv.org/abs/2611.08494
作者: Muhammed Adil Yatkın,Mihkel Kõrgesaar,Hasan Kurban
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)
*备注: Journal paper. 22 pages, 16 figures
Abstract:Capturing sheet-metal localisation physics with a trained surrogate inside a finite element (FEM) solver requires the surrogate to deliver consistent predictions irrespective of the solver’s strain-increment count, which is difficult to control in non-linear explicit codes. We address this by re-implementing a trained recurrent neural network (RNN) damage criterion in Fortran and embed it as a live, increment-by-increment fracture criterion inside an Abaqus/Explicit user material subroutine (VUMAT), advancing the network state alongside the solver. Two architectures are compared: a SimpleRNN and the proposed Consistent RNN (ConsRNN), whose transition function is designed so that predictions converge as the strain path is refined rather than drifting with the increment count. Both are trained on bilinear strain paths and evaluated under varying temporal discretisations and nonlinear histories. SimpleRNN predictions drift as the increment count increases, which disqualifies it for embedding; ConsRNN converges under refinement, at a substantial fixed-resolution accuracy cost relative to SimpleRNN on multilinear paths. Deployed at structural scale, the embedded ConsRNN surrogate governs element deletion in a clamped steel plate and reaches the same peak force and peak displacement as an established two-parameter fracture criterion. This pre-peak agreement is expected by construction, since no element is deleted before the peak and the two runs therefore share an identical constitutive path; the responses diverge only on the post-peak softening branch. The results establish discretisation consistency as the decisive property for embedding recurrent surrogates in explicit FEM solvers, and report the first embedding of a customised, non-standard-transition recurrent cell at structural scale.
[LG-102] FasterRW: An Efficient Prior-Assisted Random Walk Solver for Thermal Analysis
链接: https://arxiv.org/abs/2611.08490
作者: Zixiao Wang,Tianshu Hou,Chenghan Wang,Zhen Zhuang,Leilei Jin,Tsung-Yi Ho,Farzan Farnia,Bei Yu
类目: Machine Learning (cs.LG); Numerical Analysis (math.NA); Computation (stat.CO)
*备注:
Abstract:Thermal simulation is becoming increasingly critical in modern IC design and manufacturing. Feynman–Kac random walk methods can estimate localized temperatures without solving the global field, but in practical settings without Dirichlet boundaries they require very long paths to control the truncation residual. This paper presents FasterRW, a continuation-value framework for accelerating Feynman–Kac thermal random walks. We prove a stopping-time residual identity showing that the expected tail after truncation equals the current attenuation weight multiplied by the temperature at the stopping location. This identity enables a prior-corrected single-point estimator: a cheap, noisy prior replaces the unknown continuation value, making the truncation bias scale with prior error rather than absolute temperature and yielding an error-controlled threshold-selection rule. For multi-point queries, FasterRW further converts cross-query pass-through tails into pseudo-samples and refines the resulting observations in prior-error coordinates through kriging. Compared with prior state-of-the-art methods, FasterRW achieves up to 28\times acceleration while maintaining accuracy. Our code is open-sourced at this https URL.
[LG-103] Machine learning for lattice QCD
链接: https://arxiv.org/abs/2611.09817
作者: Gurtej Kanwar,Scott Lawrence,Alessandro Nada,Phiala Shanahan
类目: High Energy Physics - Lattice (hep-lat); Machine Learning (cs.LG)
*备注: 47 pages, 5 figures
Abstract:Machine learning (ML) methods have been explored as a means of optimizing lattice field theory calculations, with increasing interest in recent years. This chapter gives an introduction and summary to the use of ML methods to optimize lattice field theory without introducing any uncontrolled approximations. We focus on those directions which are clearly established as of late 2025, marking the 50th anniversary of the beginning of the study of lattice gauge theory. This chapter is a contribution to the online book ``Lattice QCD at 50 years,‘’ (LQCD@50), edited by Tanmoy Bhattacharya, Maarten Golterman, Rajan Gupta, Laurent Lellouch, and Steve Sharpe.
[LG-104] Where Should Learning Live? Regret Analysis for Text Parameter and Joint Adaptation
链接: https://arxiv.org/abs/2611.09701
作者: Jungtaek Kim,Jaden Park,Jongwon Jeong,Emmanouil-Vasileios Vlatakis-Gkaragkounis,Kangwook Lee
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Language-model systems can select prompts and parameter adapters repeatedly during deployment, using feedback to guide later choices. Earlier choices may change the context or environment inherited by later configurations. We study selection among preconstructed candidates and compare cumulative loss with the best fixed prompt–parameter pair used from the beginning. This excess loss separates into three terms: selection regret within the available family, the gap from restricting that family, and the benefit or harm of the actual deployment history. Selection is evaluated using the history each candidate would have produced if used throughout, which separates candidate quality from the effects of switching. When loss functions are fixed independently of selector randomness, past choices affect loss for only a bounded number of rounds, and every candidate’s own-history loss is revealed, limiting switches yields sublinear expected regret within the selected family. We also quantify the costs and additional assumptions of independent probing and deployment-only feedback. Controlled experiments show that fewer switches can either lower or raise deployed loss, depending on whether inherited histories are harmful or helpful. With Qwen3.5-9B on Spider, text-only selection has lower within-family regret than joint selection, yet produces more failures because its candidate family excludes stronger configurations.
[LG-105] NAVE: A Normalized Adaptive Conformer for Antarctic Whale Vocalization Detection
链接: https://arxiv.org/abs/2611.09642
作者: Matthias Nagl,Paul Primus,Gerhard Widmer
类目: Audio and Speech Processing (eess.AS); Machine Learning (cs.LG); Sound (cs.SD)
*备注: 5 pages, 3 figures, 2 tables. Accepted at the Detection and Classification of Acoustic Scenes and Events (DCASE) 2026 Workshop. Code: this https URL
Abstract:Antarctic whale vocalization detection requires temporally localizing call boundaries in long underwater recordings and classifying them into predefined call types. Training reliable machine learning models for this task is difficult due to the variability and sparsity of whale calls in passive acoustic monitoring datasets. These difficulties are further amplified by time-varying noise that can mask whale calls and by site-specific ambient sound characteristics to which a trained model may not generalize. In this work, we present NAVE, a Conformer-based detector that improves whale-call detection via three components. First, its input representation combines per-bin complex mean subtraction, which suppresses constant site-specific ambient sound characteristics, with an additional per-channel energy normalization (PCEN) channel that adapts to temporal variations in the noise floor. Second, a frequency-dynamic stem processes these noise-aware representations with convolutions that adapt across the characteristic frequency bands of the target call types. Third, we use a Conformer backbone with a convolution kernel matched to the duration of the target calls that adds an inductive bias and improves temporal localization. We evaluate NAVE on BioDCASE 2026 Task 2 using leave-one-site-out cross-validation across the validation sites, which approximates deployment on unseen recording sites. A single NAVE model with a tuned per-class threshold exceeds the stronger WhaleVAD-BPN baseline, using no boundary-proposal network and no search over post-processing parameters. A controlled ablation study shows that the normalized front-end and a duration-matched convolutional kernel each contribute substantially to recovering the D downsweep, the hardest class for previous detectors, with the wide kernel giving the largest aggregate gain.
[LG-106] Kernel-Based Learning of Analytic Operators and Parametric PDEs
链接: https://arxiv.org/abs/2611.09612
作者: Giovanni S. Alberti,Markus Holzleitner,Lorenzo Rosasco,Matteo Santacesaria
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注: 43 pages, 4 figures
Abstract:We study kernel-based learning of analytic operators from finite samples, with a focus on how the regularity of the target should determine the choice of the kernel. We establish embeddings of analytic operators into vector-valued reproducing kernel Hilbert spaces and derive finite-sample, high-probability guarantees that quantify the uncertainty due to random sampling: in the noiseless setting, suitable analytic kernels yield rates exponential in a fractional power of the sample size, while Sobolev-type kernels yield polynomial rates. We apply the theory to the solution operator of a parametric elliptic PDE, where the training data are themselves finite-element approximations, and obtain bounds that separate the sampling and discretization contributions to the error and prescribe how the mesh size should be coupled to the sample size. The results are illustrated numerically.
[LG-107] Half-precision inference of AI weather prediction models: negligible loss of skill at a fraction of the cost
链接: https://arxiv.org/abs/2611.09573
作者: Stefan Siegert
类目: Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG); Performance (cs.PF)
*备注: 40 pages, 15 figures, 9 tables. Code: this https URL
Abstract:AI weather prediction models are released with 32-bit (fp32) weights and run in fp32 by default, although current GPUs compute much faster in 16-bit formats. We run five publicly available models – Aurora, Pangu-Weather, FourCastNet, SFNO and FastNet – on a single 16 GB GPU from 52 initial dates in 2024, in fp32 and with their networks in half precision (fp16), and verify the forecasts against ERA5. The physical state must stay in fp32: converting the whole model rounds it in physical units, which in fp16 overflows and in bf16 discards meaningful information. With the state, its normalisation and denormalisation in fp32, running the network in fp16 changes the global RMSE by at most 1.4%, and by about 0.5% or less for four of the five models, equivalent to less than one hour of lead time up to day 8 and at most two hours at day 10. The differences show no systematic pattern by region, scale or for tropical and extratropical storms, apart from a small tropical penalty in FourCastNet, and fp16 leaves the CRPS, spread and calibration of Aurora’s ensemble almost unchanged. No value overflowed fp16’s numeric range (+/-65,504); the largest stay 38 to 544 times below it. A 10-day forecast in fp16 is 1.3-3.5 times faster than in fp32 and uses 1.4-3.7 times less energy. Sensitivity experiments show that rules of thumb from other applications of reduced precision carry over only partly: the first and last layers dominate for some models but not for others, and no type of layer is sensitive on its own. For running AI weather models on a single GPU, fp16 with the state in fp32 is practically lossless at a fraction of the cost. bf16, often preferred for its wider range, is as fast but less accurate than fp16 for every model, and simple 8-bit quantisation costs far more skill.
[LG-108] urtLES: A Large-Scale Benchmark for Turbulent 3D Neural PDE Surrogates
链接: https://arxiv.org/abs/2611.09567
作者: Jarno Platenburg,Armand Kassaï Koupaï,Paola Cinnella,Patrick Gallinari
类目: Fluid Dynamics (physics.flu-dyn); Machine Learning (cs.LG)
*备注:
Abstract:Machine-learning surrogates are increasingly used to accelerate computational fluid dynamics, yet progress is limited by the lack of benchmarks capturing realistic, time-dependent turbulent flows. We introduce TurtLES, a 14 TB dataset of 3D turbulent cylinder wakes in the shear-layer transition regime, computed via high-fidelity implicit large-eddy simulations. The dataset contains 400 trajectories of 400 time steps each, with 3-11 million points per frame, spanning variations in geometry, Reynolds number, and angle of attack. Unlike existing datasets, TurtLES combines (i) physically realistic turbulence with an active energy cascade driven by vortex stretching, (ii) dense irregular meshes, and (iii) long-horizon temporal dynamics. We benchmark state-of-the-art neural operators on autoregressive rollouts and identify key failure modes, including temporal instability, loss of small-scale energy, and challenges in learning on large irregular domains. TurtLES provides a new benchmark for evaluating surrogate models on 3D turbulent wakes at industrially relevant mesh resolution.
[LG-109] Graph representations of dislocation networks
链接: https://arxiv.org/abs/2611.09534
作者: Gaurav Lathiya,Markus Stricker
类目: Materials Science (cond-mat.mtrl-sci); Machine Learning (cs.LG)
*备注: 28 pages, 15 figures
Abstract:The mechanical response of a crystal is governed by the evolving network of dislocations it contains, yet quantitatively linking the `state of that network’ to its yield stress remains challenging. Here we represent the dislocation networks obtained from discrete dislocation dynamic simulation as graphs, using five distinct graph representations, characterize each network through a compact set of topological properties, and predict the yield stress from these graph properties and the applied strain rate using a standard random forest model evaluated with strain rate stratified cross-validation. The model predicts the yield stress with R^2\approx0.86-0.90 across all box sizes and R^2\approx0.91-0.94 for larger box sizes, and this accuracy is consistent across all five representations, indicating that the topology-stress relationship is a robust feature of the graph properties rather than an artefact of any single graph representation. We propose graph topology as a compact and robust route for connecting the microstructural state of a dislocation network to its mechanical response in addition to established variables.
[LG-110] Geometry-Aware Score-Repellent Monte Carlo NEURIPS2026
链接: https://arxiv.org/abs/2611.08909
作者: Jie Hu,Lingyun Chen,Do Young Eun
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: To appear at NeurIPS 2026
Abstract:The recently proposed Score-Repellent Monte Carlo (SRMC) method improves MCMC sampling by tilting the target density \pi in \mathbbR^d with the factor \exp(-\alpha,\theta_n^\top s(x)) , where the score function s(x) = \nabla_x \log \pi(x) , and \theta_n is the score average from past samples. However, a single scalar repellence strength \alpha\ge0 cannot handle both the steep and the flat directions in anisotropic energy landscapes at once. We introduce \emphGeometry-Aware SRMC (GA-SRMC), which generalizes the Euclidean alignment to \theta^\top M s(x) for any symmetric positive-definite matrix M . A stochastic-approximation central limit theorem under a trace constraint on M identifies the trace-normalized inverse Fisher M_\rm opt\propto S^-1 as the unique minimizer of the worst-direction covariance bound, elevating the Fisher information S=\textCov_\pi(s,s) from a proposal-side preconditioner (e.g., FisherMALA, natural gradient) to the bound-optimal repulsion-side matrix, with a single online estimate \widehat S_n serving both roles at no additional computational cost. The full-Fisher online metric M_n converges almost surely to M_\rm opt and satisfies a CLT, providing a theoretical guarantee for the online implementation of the GA-SRMC framework. On an 8 -target simulation spanning the condition number \kappa\in[1.3,100] and dimension d\in[2,785] , GA-SRMC dominates SRMC in ESS by \boldsymbol4.3\times on anisotropic Gaussian ( \kappa=100 ) and \boldsymbol6.8\times on MNIST under the MALA baseline, by \boldsymbol16\text–32% on three well-conditioned logistic-regression posteriors, and by up to \boldsymbol+48% under FisherMALA; gains peak on the most anisotropic targets, and posterior-mean MSE is lower than with SRMC in all 16 full-Fisher implementations.
[LG-111] Reweighting for Transfer Classification: Exact Risk Conversion and Minimax Adaptation
链接: https://arxiv.org/abs/2611.08876
作者: Aoran Chen,Yang Feng
类目: atistics Theory (math.ST); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 87 pages, 5 figures. Main text (32 pages) followed by the Supplementary Material
Abstract:A source can make a training mixture easier to learn while making its accuracy less valuable for the target. We study how the training problem a source creates and the finite-sample error of learning it jointly determine target-risk guarantees under reweighting. For binary classification, for each pair of laws and each positive target weight, we derive the sharpest target excess hinge-risk bound implied by mixture excess alone, uniformly over clipped predictors. This exact conversion curve is determined by density-weighted local votes. A two-sample oracle inequality for weighted support vector machines supplies the mixture excess, and the composition becomes explicit for Gaussian kernels. On a tilted-boundary class, we establish the all-procedure minimax classification rate [n+s_+^2/(m^-1+\tau^2)]^-1 in the transferable strength s_+ and the tilt \tau , separating source sampling uncertainty ( m^-1 ) from relation uncertainty ( \tau^2 ). A target-checked Gaussian-SVM selector adapts to unknown s_+ and \tau up to a logarithmic factor for fixed dimension, and retains the target-only rate up to the same factor against arbitrary independent sources under the stated target model. On a separate smooth opposition class, the minimax risk within the family of exact Gaussian SVMs fitted to the original labels with nonnegative weights is exponentially larger than the unrestricted minimax risk whenever the target sample is at most half the source sample, despite arbitrary data-dependent weighting and tuning. The value of a source thus depends on the training problem it creates, on what accuracy there means for the target, and on whether the permitted use of the source can implement the needed correction.
[LG-112] Beyond the Lower-Star Filtration: Alternative Filtration Methods for Persistent Homology
链接: https://arxiv.org/abs/2611.08833
作者: Innocent Abaa,Umar Islambekov
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Algebraic Topology (math.AT); Applications (stat.AP)
*备注: This is the published version of the article. Published in International Journal of Topology, MDPI, DOI: https://doi.org/10.3390/ijt3040025 . Link to the official version: this https URL
Abstract:In the lower-star filtration (LSF), simplices are assigned filtration values according to the maximum-vertex rule, with each simplex entering the filtration when its highest-valued vertex appears. A limitation of this construction is that some higher-dimensional simplices may enter simultaneously with their boundary faces, causing certain topological features to be filled immediately upon formation and thus not appear in the resulting persistence diagram. To address this limitation, we propose three alternative filtration methods: the (\varepsilon)-lower-star filtration (ELSF), the (L_p) filtration (LpF), and the softmax filtration (SMF). ELSF introduces a dimension-dependent delay to the maximum-vertex rule, allowing higher-dimensional simplices to enter the filtration after their boundary faces, whereas LpF and SMF use aggregation functions that depend on all vertex values rather than solely on their maximum. By modifying how filtration values are assigned, these methods allow simplices to enter the filtration more progressively, so that topological features such as loops and voids may persist over longer filtration intervals rather than being filled immediately. Consequently, the resulting persistence diagrams may include persistence features that are not observed under LSF, potentially providing useful information for distinguishing data populations. Experimental results on simulated and real-world datasets demonstrate that the proposed filtration methods are computationally efficient and can provide greater sensitivity to structural differences in the data than LSF.
[LG-113] High-Dimensional Off-Policy Evaluation without Sample Splitting using an Onsager Correction
链接: https://arxiv.org/abs/2611.08794
作者: Saksham Jain,Guillaume Braun,Masaaki Imaizumi
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注:
Abstract:Reward models used for off-policy evaluation (OPE) are commonly selected using held-out data. When sample size and context dimension grow proportionally, the difference in test risk between models fitted on the full sample and on a fixed fraction of it need not vanish. Held-out evaluation can therefore consistently estimate the test risk of the model trained on the smaller sample, but because the relative ranking of candidate reward models can change after full-data refitting, held-out selection can be suboptimal. We develop an Onsager-corrected, target-policy-weighted criterion that estimates the full-data reward model’s out-of-sample risk under the target policy using only its training observations. This enables early stopping along the full-data gradient-descent trajectory without sample splitting. For neural networks with fixed depth and hidden width under Gaussian contexts, single-index rewards, context-dependent multi-index target policies, and known context-independent logging, we prove uniform consistency of our criterion over a fixed finite set of gradient-descent iterates and a near-oracle guarantee for the selected reward model. The same correction yields a doubly robust estimator that consistently estimates the true policy value without sample splitting, even when reward-model prediction error does not vanish. Experiments show decreasing risk-estimation error as dimension grows at a fixed sample-to-dimension ratio, and improved reward-model selection relative to held-out baselines. Empirically, our correction also reduces the root mean squared error of doubly robust value estimates computed from the same data used to fit and select the reward models.
[LG-114] Blind Noise Equalization for Large Data Matrices under General Variance Profiles
链接: https://arxiv.org/abs/2611.08784
作者: Boris Landa,Francesco Strino,Fabio Parisi,Ruiqi Li,Yuval Kluger
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Signal Processing (eess.SP); Statistics Theory (math.ST); Methodology (stat.ME)
*备注:
Abstract:Low-rank approximation is a basic tool for inference from large and noisy data matrices. However, when noise variances differ across entries, a setting known as heteroskedasticity, the noise can distort the observed spectrum, producing spurious components and obscuring weak signals. To address these challenges, we develop BRINE (Blind Resolvent-based Iterative Noise Equalizer), a data-driven procedure that iteratively rescales rows and columns to equalize the average noise variance across both dimensions. Under suitable conditions, this procedure restores the Marchenko-Pastur law for the noise spectrum and can amplify signal components without prior knowledge of the signal or noise structure. A key idea underlying our approach is to equalize the diagonal entries of the data resolvent as a proxy for equalizing the unknown noise variances. At the population level, we identify a potential function that characterizes this equalization and motivates an iterative minimization procedure. We prove convergence of the iterations to the true equalizer for arbitrary positive variance profiles, with a geometric rate under mild conditions. The updates can be estimated directly from noisy data, and we establish probabilistic error bounds for the scaling factors. Simulations and experiments across diverse biological data modalities and preprocessing pipelines illustrate the effectiveness of our approach.
[LG-115] Compressed Per-Player Local Methods for Multiplayer Federated Learning: Exact Convergence via Increment Sharing
链接: https://arxiv.org/abs/2611.08686
作者: Bhavika Balani,TaeHo Yoon,Nicolas Loizou
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:Multiplayer Federated Learning (MpFL), introduced by Yoon et al. (2025), extends classical federated learning to settings where clients are rational players in a game with individual objectives and strategies, seeking an equilibrium. A natural baseline is Per-Player Local SGD (PEARL-SGD), where each player performs several local SGD steps before exchanging strategies through a central server. However, the original PEARL-SGD formulation does not extend cleanly to compressed communication. Directly compressing the players’ strategies can prevent convergence to the equilibrium, even with deterministic gradients. To address this issue, we reformulate PEARL-SGD so that players communicate local strategy increments rather than the strategies themselves. Without compression, the two formulations are equivalent, but with compression, increment sharing preserves exact convergence, allowing us to design the communication-efficient compressed PEARL-SGD (C-PEARL-SGD). Beyond compression, we also consider variance reduction as another axis of improvement, introducing PEARL-SVRG and its compressed variant C-PEARL-SVRG. Under the original MpFL assumptions, we establish convergence guarantees for all variants through a unified parametric inequality. Numerical experiments, including distributed spectrum sensing, demonstrate the practical benefits of our methods over vanilla PEARL-SGD.
[LG-116] Just CaloCHILL: fast hadronic shower generation in a learned latent space
链接: https://arxiv.org/abs/2611.08606
作者: Henry Day-Hall,Frank Gaede,Gregor Kasieczka,Katja Krüger,Martina Mozzanica
类目: Instrumentation and Detectors (physics.ins-det); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); High Energy Physics - Phenomenology (hep-ph); Data Analysis, Statistics and Probability (physics.data-an)
*备注: 25 pages, 19 figures, 11 tables. Code: this https URL , data: this https URL
Abstract:Calorimeter showers are the most expensive part of detector simulation, and highly granular calorimeters, in which a single hadronic shower leaves thousands of energy deposits, make them more expensive still. We present CaloCHILL (Calorimeter Hits In Learned Latents), a generative model that does not produce these deposits one by one: instead it generates a compact summary of the shower in a learned latent space and expands it to full resolution only at the end. The complete shower, including the arrival time of every hit, is built in three stages: a flow-matching model first sets the number of hits and the earliest hit time in each calorimeter layer, a mean-flow model then generates the latent summary, and the decompressor of a variational autoencoder turns it into the full point cloud, giving every hit its position, energy and time. Sparse attention along the shower axis keeps this compression affordable for showers with thousands of hits. On positive-pion showers in the calorimeter of the International Large Detector, a proposed detector for the FCC-ee, CaloCHILL reproduces the main shower observables, including their time structure. This new model is 600 times faster than its predecessor, while improving on all accuracy metrics considered.
[LG-117] Personalized health coaching via coachable health age and counterfactual wearable trajectories
链接: https://arxiv.org/abs/2611.08568
作者: Simon A. Lee,Zeinab Esmaeilpour,Menglian Zhou,Marius Guerard,Sohrab Sidhu,Alexandru Solot,Hamid Palangi,Pramod Rudrapatna,Ming-Zher Poh,Max A. Xu,Daniel McDuff,Dimitrios Antos,Yuzhe Yang,Jeffrey N. Chiang,Anthony Z. Faranesh,Lindsey Sunden,Mark Malhotra,Shwetak Patel,A. Ali Heydari,Ahmed A. Metwally
类目: Quantitative Methods (q-bio.QM); Machine Learning (cs.LG)
*备注:
Abstract:Personalized health coaching requires moving beyond retrospective wearable summaries toward quantitative, actionable behavioral prescriptions. We introduce Coachable Health Age (CHA), an interpretable health-age metric built from modifiable physiological and behavioral signals passively captured by consumer wearables. Across the WEAR-ME (N=2,780), All of Us (N=22,748), and longitudinal Fitbit (N=1,464) cohorts, CHA aligns with established cardiometabolic health constructs and improves prognostic discrimination for incident cardiovascular and metabolic disease beyond age, sex, and body mass index. To enable prospective behavioral optimization, we developed Sensor World Model (SensorWM), an action-conditioned wearable world model pretrained on over 2.2 million person-weeks of longitudinal sensor data from approximately 100,000 participants. SensorWM accurately forecasts multi-week physiological responses to candidate lifestyle changes ( R^2=0.664 ), outperforming generic time-series foundation models while preserving downstream clinical signal for diabetes, hypertension, and other health outcomes. Finally, we integrate CHA and SensorWM into an agentic counterfactual coaching system to synthesize personalized 12-week coaching plans. In blinded paired reviews, certified lifestyle coaches preferred counterfactual-grounded plans over plans grounded in wearable history alone for overall quality (71.4%, P 0.001) and anticipated adherence (71.0%, P 0.001), and board-certified clinicians judged 99.2% of counterfactual-grounded plans to meet clinical safety criteria (vs. 97.6% of comparison plans). Automated evaluations across the full cohort (N=2,780) also favored counterfactual-grounded plans. By closing the loop between longitudinal tracking, trajectory prediction, and counterfactual coaching, our framework provides a scalable path to systematically improve cardiometabolic health.
[LG-118] pESTO: Iterative p-greedy kernel regression with risk-driven Early STOpping
链接: https://arxiv.org/abs/2611.08555
作者: L. Bruni Bruno,P. Massa,E. Perracchione,M. Trombini
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Kernel-based greedy algorithms select suitable approximation nodes by minimizing point-wise error indicators, such as the power function; in this context, they are usually referred to as p-greedy approaches. In the presence of noisy data, deterministic criteria may fail, as the discrepancy between the true solution and its estimate is influenced by both approximation error and noise amplification, and the balance between the two is unknown. This study explicitly separates these two contributions, which are governed by the power function and a Lebesgue-type term, respectively. These bounds motivate the use of the Stein’s unbiased risk estimator (SURE) as a proxy of the mean squared error, which in turn provides a principled and computable stopping criterion for the number of nodes to select in the p-greedy procedure. Numerical experiments demonstrate that this approach effectively balances deterministic accuracy and statistical noise contributions.
[LG-119] GenStrain: Multi-Conditional Generative Diffusion Model To Synthesize High-quality Motion and Strain Maps From Cardiac Magnetic Resonance Images
链接: https://arxiv.org/abs/2611.08544
作者: Rishov Paul,Sona Ghadimi,Cristiane C. Singulane,Amit R. Patel,Frederick H. Epstein,Miaomiao Zhang
类目: Image and Video Processing (eess.IV); Machine Learning (cs.LG)
*备注:
Abstract:This paper presents GenStrain, a novel method for synthesizing high-quality cardiac motion-based strain maps from routinely collected cine cardiac magnetic resonance (CMR) images. Existing approaches typically rely on image registration to estimate motion (a sequence of displacements) from cine CMR images, but cine CMR presents very sparse trackable intramyocardial features, presenting challenges to estimate intramyocardial motion. An advanced imaging technique, Displacement Encoding with Stimulated Echoes (DENSE) CMR, is considered as the gold standard for strain assessment, but it requires additional acquisitions, which are not ideal for efficient clinical workflows. To address this gap, we introduce GenStrain, a novel framework designed to learn the data distribution of highly accurate cardiac motion captured by DENSE. In particular, we develop a diffusion process in motion space, where spatiotemporal continuity is carefully considered. To fully leverage both cine CMR images and their associated motion patterns, we introduce a multi-conditional conditioning strategy that integrates anatomical features and motion representations learned by a spatiotemporal registration model to guide the generation process. Experimental results demonstrate that our proposed model, GenStrain, outperforms existing strain analysis methods based on displacement estimates from both traditional registration approaches and state-of-the-art (SOTA) deep learning models. Our model represents a promising step toward more accurate cardiac function assessment and holds great potential for clinical translation. Our code is publicly available at this https URL.
[LG-120] ELPAC: Endpoint-Anchored Latent Progression with Stage-Varying Multimodal Coordination NEURIPS2026
链接: https://arxiv.org/abs/2611.08541
作者: Shengxian Ding,Yifei Zhang,Xinyuan Tian,Yuanshuang Guo,Yize Zhao
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注: NeurIPS 2026
Abstract:Aging cohorts increasingly combine imaging, plasma, and cognitive measurements, but most subjects provide only a partially observed cross-sectional snapshot, leaving latent stage, progression subtype, and stage-varying multimodal coordination to be inferred jointly. Existing methods estimate the cohort-level aging process, model covariance under observed covariates, or fuse modalities for prediction. However, they rarely target the joint problem of endpoint-anchored staging and progression-resolved residual coordination across modalities, even though subject staging, subtype assignment, and multimodal interplay are mutually dependent. We introduce ELPAC, a Bayesian framework with two interlocking constructions. The first integrates subtype trajectories backward from clinically reliable endpoint groups along biologically structured velocity components encoding modality-specific monotone direction and optional sequential ordering priors. The second factorizes residual multimodal variation around the inferred stage axis into sparse coordination axes whose loadings and activations expose which multimodal patterns are operative in each subtype and when they emerge, with stage-resolved variation supplying the auxiliary structure for conditional identifiability. We use a scalable amortized variational inference scheme that enables joint posterior inference at cohort scale with single-pass deployment on new subjects. On synthetic data and a harmonized ADNI+NACC/SCAN Alzheimer’s cohort, ELPAC demonstrates accurate latent-structure recovery, well-calibrated prediction intervals, and biologically interpretable multimodal coordination programs that differ by subtype.
[LG-121] Online and Survey Representations of Well-Being: Coherence and Cross-Source Correspondence
链接: https://arxiv.org/abs/2611.08510
作者: Stefano M. Iacus,Giuseppe Porro,Haodong Qi
类目: Econometrics (econ.EM); Machine Learning (cs.LG); Physics and Society (physics.soc-ph); Applications (stat.AP)
*备注:
Abstract:Digital traces may complement population surveys in the measurement of well-being, but the two sources do not necessarily capture identical constructs. We compare the Human Flourishing Geographic Index (HFGI) with World Values Survey and European Social Survey measures across three domains using structural equation models. The results show strong within-source coherence and construct-specific cross-source correspondence. Cross-source convergence varies across domains; the observed pattern is consistent with stronger correspondence where the underlying indicators are more closely aligned conceptually. Near-perfect agreement would be unexpected, since surveys elicit standardized self-reports, whereas HFGI captures naturally occurring public expressions. HFGI should therefore be viewed as a distinct but empirically connected measurement system that complements survey-based evidence on well-being.
[LG-122] Drift-admissibility via Optimal Transport
链接: https://arxiv.org/abs/2611.08505
作者: Abigail Langbridge,Enrico Biffis,Jakub Mareček,Robert Shorten
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注: 10 pages; 3 figures; 2 tables
Abstract:We introduce structure-aware optimal transport diagnostics for assessing whether empirical time-series dynamics are compatible with a prescribed predictable-drift class. Our approach translates a pathwise property of stochastic processes into a finite-dimensional projection problem over empirical transition laws, encoding admissibility as row-wise moment constraints on local couplings and quantifying deviations via a Kullback–Leibler projection gap G_\lambda,k . This yields both a principled diagnostic and a scalable Bregman–Dykstra algorithm that retains the efficiency of Sinkhorn-type methods while incorporating structural constraints. Calibrated against a Brownian reference null, G_\lambda,k separates process classes that are indistinguishable by second-order statistics alone—including fractional Brownian motion and BSS processes sharing identical Hurst indices—without asymptotic resampling or continuous-time recovery. This positions the method as a general tool for structure-aware model validation of sequential data.
[LG-123] Electoral Intelligence: A Neural Zero-Inflated Negative Binomial Model for Forecasting Swedish Municipal Elections
链接: https://arxiv.org/abs/2611.08501
作者: Haodong Qi,Stefano M. Iacus,Pieter Bevelander,Derek S. Hutcheson
类目: Econometrics (econ.EM); Machine Learning (cs.LG); Applications (stat.AP)
*备注: 21 pages, 8 Figures
Abstract:Forecasting local elections is difficult because electoral competition is geographically heterogeneous, highly overdispersed, and characterized by many party-municipality combinations with little or no electoral activity. This article develops a bounded neural zero-inflated negative binomial (ZINB) model for forecasting party-level vote counts in Swedish municipal elections. The model combines a count component for expected votes, a zero-inflation component for structural electoral absence, municipality-specific dispersion, and learned embeddings capturing latent party, municipality, and party-municipality structure. Model training relies on data from the 2014, 2018, and 2022 Swedish municipal elections and performance is evaluated on a held-out set of municipalities. Testing results show strong predictive accuracy in both vote counts and vote shares. Explanatory diagnostics indicate that latent party structure, local party-municipality relationships, electoral persistence, organizational capacity, and candidate-electorate representational similarity are among the most important determinants of predictive performance. Following model evaluation and diagnostics, the selected specification is refit on all historical data and applied to generate forecasts for the 2026 municipal elections. The proposed framework provides a flexible approach for studying local electoral competition in multiparty systems characterized by heterogeneous party presence and substantial geographic variation.
附件下载
