本篇博文主要内容为 2026-10-21 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 (2026-10-21)
今日共更新945篇论文,其中:
- 自然语言处理共160篇(Computation and Language (cs.CL))
- 人工智能共322篇(Artificial Intelligence (cs.AI))
- 计算机视觉共186篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共249篇(Machine Learning (cs.LG))
- 多智能体系统共13篇(Multiagent Systems (cs.MA))
- 信息检索共21篇(Information Retrieval (cs.IR))
- 人机交互共23篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] Proof-Carrying Contracts: Commitments AI Agents Cannot Break
【速读】：该论文旨在解决在由人工智能代理（AI agents）构成的协作经济中，如何确保代理在执行任务时严格遵守各方预先约定的约束条件（如政策、协议或指令），而无需依赖对对方代理的信任或逐项审查的问题。传统文本形式的协议无法保证代理在实际交互中始终合规，因其可能因误解、误读或自身利益驱动而偏离预期行为，从而引发安全与可靠性风险。为此，论文提出**可携带证明的合约（Proof-Carrying Contracts, PC2）**框架，其核心在于将合约中的约束条件转化为机器可验证的逻辑证明。在任何操作被执行前，代理必须提供一个形式化证明，表明其决策在所有可能的情境下均能保持合约所规定的约束不变。系统通过名为“Airlock”的执行环境对证明进行自动化检查，并据此决定是否授权操作。该机制使合约制定者仅需定义需保护的状态条件，而代理则自主选择策略并提供相应的推理依据，无需额外设计独立的权限策略，也避免了对每个决策逐一进行安全性验证的复杂性。论文证明，在整个执行过程中，即使面对恶意代理，约束条件依然被严格维持；并进一步分析了此类形式化保证在现实世界中蕴含的安全性意义。此外，可复用的能力模块为合约构建提供了标准化工具与形式化保障，从而为一个既保护委托方又保障执行方权益的可信代理经济体系奠定了基础。
链接: https://arxiv.org/abs/2610.28484
作者: Carlo Ciliberto,Alessandro Rudi,Enrico Tassi
机构: 未知
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Logic in Computer Science (cs.LO); Multiagent Systems (cs.MA); Systems and Control (eess.SY)
备注: 70 pages, 8 Figures,
Abstract:AI agents increasingly act on behalf of people or other agents, purchasing services, exchanging information and delegating work. What an agent must respect is written in instructions, policies or agreements, but text alone cannot make it comply. It may make mistakes, be redirected by what it reads, or pursue different interests. For an economy of such agents to function, parties need agreements they can rely on without inspecting or trusting one another’s agents. We introduce Proof-Carrying Contracts (PC2), a framework for making agreed constraints enforceable throughout an agent’s interactions with a system. Before a requested operation can run, the agent must supply a machine-checkable proof that it preserves the agreed constraints in every case allowed by its model. An execution system called the Airlock checks the proof and controls access to the operation. The parties specify the conditions to preserve, while the agent supplies the reasoning needed to justify its decisions and remains free to adapt its strategy. This avoids having to design a separate permission policy and prove that every decision it permits is safe. We prove that the constraints hold at every instant during execution, even for an adversarial agent, and identify when this implies safety in the world. Reusable capability modules provide tools and guarantees from which contracts can be constructed. PC2 provides a foundation for an agent economy whose agreements protect both those delegating work and those carrying it out.
[MA-1] Claims of emergence in AI should be grounded in information decomposition NEURIPS
【速读】：该论文旨在解决当前人工智能（AI）研究中对“协同效应”或“涌现”现象缺乏统一、可操作的理论框架这一核心问题。现有研究常以非正式方式描述多智能体系统、语言模型或人机协作中出现的超越个体能力的现象，但这些判断因缺乏量化标准而难以验证或比较。其解决方案的关键在于引入信息分解（information decomposition）作为统一的数学框架，将信息划分为三类：个体组件独有的信息、多个组件间冗余共享的信息，以及仅在各部分共同作用下才能获取的协同信息（synergistic information）。该框架为“涌现”提供了明确且可证伪的判定标准——只有当系统行为携带关于任务结果的协同信息时，才可被视为真正涌现。通过应用于注意力头、自主代理及人机团队，该方法不仅整合了跨尺度与跨底物的研究成果，还揭示出单一领域研究可能忽视的跨层次因果机制。
链接: https://arxiv.org/abs/2610.28479
作者: Christoph Riedl,Fernando E. Rosas
机构: Khoury College of Computer Sciences (凯瑟琳计算机科学学院); Northeastern University (东北大学); Department of Informatics (信息学系); University of Sussex (萨塞克斯大学)
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Information Theory (cs.IT); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: NeurIPS Spotlight Paper
Abstract:When a language model develops capabilities not traceable to its individual components, when a multi-agent AI system outperforms each agent on its own, or when a human-AI pair succeeds where either fails in isolation, AI researchers describe it as a case of synergy'' or emergence’'. However, these are usually informal assessments not supported by a theoretical framework, making these claims impossible to compare or even meaningfully dispute. Here we argue that \emphinformation decomposition can serve as a unifying formal framework to operationalize assessments about synergy in AI systems. Information decomposition provides a way to distinguish qualitatively different \emphtypes of information, including unique information available from individual components, redundant information shared across multiple components, and synergistic information accessible only from parts taken together. This gives claims of emergence a precise, falsifiable criterion: a system’s behavior is emergent only if it carries synergistic information about task-relevant outcomes. We apply this lens to attention heads, autonomous agents, and human-AI teams, showing how it organizes existing knowledge across scales and substrates, while also generating cross-level conjectures that siloed fields would miss.
[MA-2] oolflation: Seller Agents Distort Tool Markets Through Spontaneous Generative Engine Optimization
【速读】：该论文旨在解决在多代理系统中，当销售代理（seller agents）为提升自身被选中概率而自主进行“生成式引擎优化”时，所引发的市场失真问题。其核心挑战在于：销售代理通过夸大自身工具的能力（如声称可处理更多任务领域），即使这些声明与实际性能不符，从而扭曲了买家代理（buyer agents）的选择机制，导致市场效率下降。解决方案的关键在于识别并定义这一现象——即“工具通胀”（toolflation），并揭示其在缺乏外部约束机制下的自我强化特性。研究基于亚马逊（Amazon）和Instagram的真实工具数据构建工具市场，对每种工具在所有问题上的表现进行全量评估，发现五种生产模型中均有销售代理不断扩展其功能宣称范围，甚至包括自身能力不足的领域；在部分模型下，这种行为导致工具描述趋同、失去区分度，买家仅依据位置选择工具，最终使任务成功率从近90%骤降至约30%，接近随机选择水平。这表明，在竞争环境下，对第三方保持诚实应被视为安全对齐（safety alignment）的重要组成部分，因此论文主张在部署前必须对模型作为销售者的角色进行独立测试，以防范此类系统性风险。
链接: https://arxiv.org/abs/2610.28454
作者: Zhenwei Tang,Di Bai,Jintao Liu,Xidong Wu,Qianfeng Wen,Xinwu Cheng
机构: Google(谷歌); University of Toronto (多伦多大学); Google DeepMind(谷歌深度思维)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Multiagent Systems (cs.MA)
备注: 24 pages, 8 figures, 9 tables. Code: this https URL . Data: this https URL
Abstract:Agents are starting to sell to other agents. Seller agents write the listings of tools, services and products, and buyer agents read these listings when they choose what to use. We show that seller agents told only to be selected more often practice generative engine optimization on their own, and that this distorts the market for sellers and for buyers. We build tool markets from real data tools on Amazon and Instagram and run every tool on every question, so we know which tools can answer each question. Under all five production models we test, sellers rewrite their descriptions to claim more and more fields, including fields that their tools return for fewer than half of the targets. We call this process toolflation. Nothing in the market limits it, and how far it goes depends on the model, as the number of claims that contradict the sellers’ own evidence differs by an order of magnitude between models. Toolflation shifts the buyers’ choices between sellers, and the most capable tool loses share in seven of the ten combinations of seller model and platform. Under four models, the choices stay with capable tools and buyer agents keep their task success. Under the fifth, sellers copy one another until the descriptions no longer tell the tools apart. Buyer agents then pick tools by their position in the list, tools that answer almost nothing win a quarter to two fifths of the choices, and task success falls from 89% to about 30% on Amazon, close to random choice, and from 77% to 40% on Instagram. The sellers did only what their vendors asked, and the cost fell on buyers they never dealt with. We therefore argue that honesty toward third parties under competition is part of safety alignment, and that models should be tested as sellers before they are deployed as sellers.
[MA-3] Local Work Global Progress: A Convergence Theory of AI Agent Collaboration
【速读】：该论文旨在解决大规模语言模型（Large Language Model, LLM）智能体团队在执行复杂任务时，协作机制多依赖经验设计而缺乏收敛性保障的问题。其核心解决方案是将协同问题求解形式化为块优化（block optimization）：每个智能体在其负责的解空间局部改进方案，同时依据自身局部目标及对其他部分影响的修正项进行调整；由协调者（coordinator）评估每项提议是否符合全局目标，并合并互不冲突的提议。当修正项精确时，可证明线性收敛至一个误差下界；若存在干扰，则可在验证组合可行性后合并，此时收敛速率取决于实际组合效果而非最坏情况下的交互。实验表明，在机器人路径规划任务中，三智能体团队均能生成无碰撞路径，性能优于单个全控智能体；六智能体团队在更大规模任务中搜索速度约为单智能体的两倍，且通过选择经验证的最佳组合而非仅非冲突提案，进一步缩短约30%搜索时间。基于实际运行数据设定常数后，理论收敛边界在每轮迭代中均成立，与观测进展高度一致。该方法在SQLite查询计划与合成任务上亦得到验证。
链接: https://arxiv.org/abs/2610.28286
作者: Ivan Ilin,Peter Richtárik
机构: KAUST(沙特国王科技大学)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Optimization and Control (math.OC)
备注: 59 pages, 15 figures, 28 tables. Code: this https URL
Abstract:Teams of large language model (LLM) agents are increasingly applied to complex tasks, but their collaboration protocols are mostly designed empirically, without convergence guarantees. We formalize collaborative problem solving as block optimization: each agent improves one part of a shared solution, guided by its local objective plus a correction for how its changes affect the other parts. A coordinator checks every proposal against the global objective and merges proposals that do not interfere with each other. We prove linear convergence up to an error floor that vanishes when the corrections are exact. Interfering proposals can also be merged once their combination has been checked; the rate then depends on how well the proposals actually combine, not on worst-case interactions. With GPT-OSS-120B agents on robot routing, a three-agent team finds collision-free plans in every run, better than those of a single agent that controls all robots. On a larger routing task, a six-agent team completes its search about twice as fast as the single agent (in median). Merging the best checked combination of proposals, rather than only proposals that do not interfere, cuts the team’s search time by about 30%. With constants measured on the runs, our convergence bound holds in every round and follows the observed progress. We also test the theory on query planning in SQLite and on a synthetic task.
[MA-4] Safe Multi-Robot Collaboration With Uniform Exclusion Using Trust
【速读】：该论文旨在解决机器人集群在存在潜在恶意通信链路（即被攻陷的链路，可传输任意恶意数据）的情况下，如何实现安全且高效的协同决策问题。其核心挑战在于，在不依赖强假设（如网络连通性结构或受控攻击数量）的前提下，动态识别并排除受损链路，同时快速建立对合法链路的信任，以保障任务的安全性与性能。解决方案的关键在于提出一种名为“Safe Wrapper”的通用通信封装算法，该算法通过集成基于异常检测器生成的实时可信度流（trust values），利用安全的即时有效推断（Safe Anytime-Valid Inference） 方法，持续筛选出高置信度的可靠通信链路。该方法仅要求异常检测器性能优于随机猜测，即可在任意时刻提供双重统计保障：一是以用户可指定的高概率始终排除所有被攻陷的链路（全局安全性），二是迅速学习并信任真实的合法链路（高效性）。这一机制在真实硬件平台上验证了其有效性，分别应用于3个LoCoBots的神经隐式建图和8个GoPiGos的方向共识任务，并通过基于实测数据的仿真证明其可扩展至数百机器人的规模。
链接: https://arxiv.org/abs/2610.28251
作者: Thomas Kaminsky,Áron Vékássy,Hammad Izhar,Angelia Nedić,Stephanie Gil
机构: Harvard University (哈佛大学); Arizona State University (亚利桑那州立大学)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注: 44 pages (30 main text + 14 supplementary); 10 figures, 6 tables
Abstract:Robot swarms coordinate through directed, time-varying local communication links, some of which may be compromised to allow the transmission of arbitrary malicious data. We provide an algorithm that acts as a wrapper around robot communication—when a robot seeks to update its state using data from some incoming links, our wrapper provides a subset of trusted links whose data can be safely incorporated. It selects these links by testing a stream of trust values—data collected from an anomaly detector on each robot—using Safe Anytime-Valid Inference. We require only that the detector be better than chance. This wrapper ensures safe and performant autonomy by providing two powerful statistical guarantees: it uniformly excludes all compromised links for all time with high user-chosen probability, but nevertheless quickly learns to trust legitimate links. Both guarantees are vital to secure generic tasks, but no previous work provides them simultaneously without restrictive assumptions on connectivity structure or the number of compromised links that are not guaranteed in practice. We study our Safe Wrapper’s performance in hardware for two contrasting tasks—neural implicit mapping and heading consensus—deploying on 3 LoCoBots and 8 GoPiGos, respectively. We provide simulations using trust values sampled from hardware to validate that our approach scales well to swarms of hundreds of robots.
[MA-5] S2-MAS: A Society-Scale Multi-Agent Simulation Framework with 500 Million Household Persona Agents
【速读】：该论文旨在解决大规模社会经济模拟中个体决策行为与家庭实际状况及历史行为动态耦合的问题，即如何在社会尺度上构建能够真实反映家庭微观特征及其决策演化路径的多主体仿真框架。其解决方案的关键在于提出S²-MAS框架，通过基于5亿个基于真实调查数据生成的合成家庭人物（synthetic household personas）实现高保真度的社会规模模拟；引入Genese引擎将决策过程与状态更新解耦，提升模型灵活性；结合可复用的金融情景与应用账本，精确追踪消费、现金流、债务与支付等关键变量；并采用经过293万样本监督微调（SFT）与强化学习对比训练（RLCD）的紧凑型决策模型FinL-mlx，实现了74.55%的决策准确率和38.37%的问题-宏观F1值，在多个消费者金融选择任务中优于Jev 1.13与GLM-5.3-Flash。该框架支持对1200万不同家庭代理在配对购买融资条件下的模拟，揭示了收入、年龄、教育、就业及家庭结构差异对消费、债务偿还与现金提取行为的影响机制，并通过24个月面板数据验证了信用额度与购买目标变化对后续信贷供给的动态反馈效应，从而实现了从异质性家庭决策到演化消费、现金流与还款路径的精准映射。
链接: https://arxiv.org/abs/2610.27746
作者: Yang Li,Di Wan,Yike Wang,Chujun He,Zongsi Xu,Xiyu Chen,Sheng Zhong,Deng Binjie,Zhenwei Guo,Hou Jun,Kunhan Guo,Weihao Lin,Tuo Zhou,Hai Liu,Ye Luo
机构: 未知
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
备注: 58 pages, 13 main figures, 7 supplementary figures
Abstract:Simulating economic responses at society scale requires agents whose decisions reflect household circumstances and the effects of earlier actions. We introduce S ^2 -MAS, a society-scale multi-agent simulation framework built on 500 million synthetic household personas grounded in real-survey data. Its Genese engine separates decisions from state updates, while reusable financial scenarios and application ledgers track consumption, cash, debt and payments. FinL-mlx, a compact decision model trained with SFT + RLCD on 2.93 million examples, achieves 74.55% accuracy and 38.37% question-macro F1 in consumer financial-choice scenarios, outperforming Jev 1.13 and GLM-5.3-Flash on both measures. We evaluate eight financial task families and simulate 12 million distinct household agents in paired purchase-funding conditions. Jointly raising purchase goals and approved credit limits increases consumption, with responses differing across age and income groups. Education, employment and household-size differences persist after age–income standardization and mainly reflect transitions from deferral to purchase. Asset and debt groups differ in cash drawdown, repayment and closing debt. A 24-month panel varies purchase goals under a common credit product, with repayment and arrears updating subsequent offers. S ^2 -MAS links heterogeneous household decisions to evolving consumption, cash and repayment paths, supporting economic experiments across household groups, population scales and time horizons.
[MA-6] Does Multi-Agent Collaboration Pay When Agents Agent s Serve Users? Testing Agent Scaling in Dual-Control Settings
【速读】：该论文旨在解决多智能体系统（Multi-agent System, MAS）在实际应用中是否真正优于单智能体系统（Single-agent System, SAS）这一关键问题，尤其聚焦于在与用户交互的现实场景下，MAS能否带来可衡量的性能提升。现有研究提出“能力饱和规则”：当SAS能力较弱时，协作能提升表现；但当SAS能力较强时，多智能体协作反而会拖累性能，转折点约在45%的单智能体准确率处。然而，该规则基于自主基准测试得出，未考虑用户参与的真实交互环境。本文在τ²-基准（τ²-bench）上进行验证，该基准模拟真实世界中的双控场景（如航空、零售、电信服务），要求智能体在用户指令和书面策略约束下协同工作。研究对来自三个模型家族的六种模型，在约7,950次模拟对话中对比了标准SAS与集中式MAS的表现，采用每轮对话仅一次完成令牌（completion-token）记录的严格评估方式。结果表明，在固定模拟用户条件下，所有有效测试单元中均未观察到MAS的优势：13个有效单元中有12个MAS表现更差，1个持平，整体加权平均得分下降28分（95%置信区间[-44, -12]），且其收益在预测应有优势的两个单元中仍分别落后13和22分；即使取消预算上限，其最优表现也需消耗5.5倍于SAS的计算资源才能勉强持平。此外，尽管MAS未能显著减少错误数据库写入，反而将其他失败形式转化为预算耗尽或行动停滞。研究还指出，当前MAS架构存在一个关键缺陷——其规划器从未接触过工具模式（tool schemas），导致其决策缺乏必要信息支持。因此，该研究的核心结论是：对于面向人类的服务型智能体，若无在真实应用场景中以计算成本匹配为前提的直接证据，不应默认多智能体设计优于单智能体方案。研究全程预注册并公开全部轨迹、日志、提示词与代码，确保可复现性。
链接: https://arxiv.org/abs/2610.27655
作者: Ethan Yang,Xiangcheng Shi
机构: University of Pennsylvania(宾夕法尼亚大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Multi-agent systems (MAS) built from language models are deployed on the premise that collaboration beats a single-agent system (SAS), yet when it pays is unclear. The most systematic study to date gives a capability-saturation rule: coordination helps while the SAS is weak and hurts once it is strong, with the crossing near 45% single-agent accuracy. That rule was derived on autonomous benchmarks, where no user takes part. We test it where agents now work: in conversation with a user under a written policy, on \tau^2 -bench, the dual-control benchmark. On its airline, retail and telecom domains we compare the stock SAS with a centralized MAS built from the same model for six models from three families, over about 7,950 simulated conversations, with one completion-token ledger per episode per system. The study was pre-registered with a public change log; the MAS is reported as built, with one validity bar unmet (its planner never saw the tool schemas). Under a fixed simulated user no cell shows a MAS benefit: the MAS is worse in 12 of the 13 valid cells and tied in the thirteenth, and the fitted gain is negative across the observed capability range (weighted mean -28 points, 95% CI [-44, -12]). On the two cells where the rule predicts a benefit, the MAS loses by 13 and 22 points, and without a budget ceiling it at best draws level at 5.5 times the SAS’s token spend. The one gain appears on the cell nearest 45%, under a second simulator. Whether the gain declines with capability is unresolved (slope -0.41, 95% CI [-1.51, +0.68]), because model-by-domain effects dominate. The MAS does not detectably reduce wrongful database writes; it converts other failures into budget exhaustion and inaction. For agents that serve people, the SAS is the default; a multi-agent design must show compute-matched evidence from the setting itself. We release all trajectories, ledgers, prompts and code.
[MA-7] Complete Street Design informed by Multi-Agent Simulation
【速读】：该论文旨在解决当前完整街道（Complete Street）设计主要依赖定性原则，难以应对多模式城市交通日益复杂化的挑战。其核心问题是缺乏一种能够定量评估不同街道设计方案在通行效率与交通安全之间权衡的系统性方法。解决方案的关键在于构建一个微观仿真框架，通过显式建模异质道路使用者（包括自动驾驶汽车、网联汽车、人工驾驶车辆、上下客服务及微出行用户）的行为，集成行为模型以刻画横向变道、停车操作、行人过街与违规穿行、共享车道中自行车与电动滑板车交互等复杂动态。安全性能采用基于碰撞物理的冲突严重度指标进行量化，结合遗传算法实现街道布局的优化。研究结果表明，专用自行车道虽可降低冲突严重度但会牺牲机动车通行能力；提高自动驾驶汽车渗透率有助于缓解交通波动并改善人车交互。该框架能高效识别出兼顾通行与安全的帕累托最优街道设计方案，其创新性在于将异质行为建模、基于物理的冲突评估、随机性能分析与设计优化整合于统一的定量评价体系中，为交通规划者提供可预测的虚拟测试平台，支持实施前对街道设计的量化评估与决策支持。
链接: https://arxiv.org/abs/2610.27509
作者: Sharika J. Hegde,Amirmohammad Khakpour,Pedram Beigi, Yu (Marco)Nie,Hani S. Mahmassani
机构: 未知
类目: Multiagent Systems (cs.MA)
备注: Under review for publication in Transportation Research Record (TRR)
Abstract:Objectives: Complete Street design is currently guided largely by qualitative principles despite the increasing complexity of multimodal urban traffic. This paper develops a microscopic simulation framework for quantitatively evaluating alternative street designs by explicitly representing heterogeneous road users, including Autonomous Vehicles (AVs), Connected Vehicles (CVs), Human-Driven Vehicles (HDVs), pick-up and drop-off (PUDO) services, and micromobility users. The framework jointly evaluates mobility and safety within a unified testbed. Methods:The framework enriches microscopic simulation with behavioral models for heterogeneous road users. It explicitly captures lateral movements, parking maneuvers, pedestrian crossings and jaywalking, and shared-lane bicycle and scooter interactions. Safety is quantified using a conflict severity measure grounded in collision physics, while candidate street layouts are optimized using a genetic algorithm. Findings: Applied to a segment of Belmont Avenue in Chicago, Illinois, the framework shows that dedicated bicycle lanes reduce conflict severity at the expense of vehicular throughput. Increasing AV penetration mitigates congestion by reducing traffic shockwaves while improving interactions with pedestrians relative to CVs and HDVs. The optimization framework efficiently identifies Pareto-efficient street layouts that balance mobility and safety. Novelty: The proposed framework integrates heterogeneous behavioral modeling, physics-based conflict assessment, stochastic performance evaluation, and design optimization within a single quantitative framework for Complete Street design. Practical Applications: The framework provides transportation planners with a predictive virtual testbed for evaluating street designs and quantifying the tradeoff between mobility and safety before implementation.
[MA-8] Robot-Task Suitability in Heterogeneous Multi-Robot Task Allocation IROS2026
【速读】：该论文旨在解决异构多机器人任务分配中机器人-任务适配度（suitability）评估标准不统一所带来的核心问题，即如何在不同定义下准确衡量个体机器人对候选任务的匹配程度。其解决方案的关键在于系统性地比较十种确定性方法与两种基于大语言模型（LLM-based）的变体，涵盖规则驱动、相似性度量、需求导向、模糊逻辑及语义分析等多种范式，并通过标准化的机器人与任务配置，评估各类方法在评分、排序、可行性判断及最优机器人选择上的表现差异。研究发现，即使在相同场景下，不同方法常产生显著分歧；通用相似性度量可能违背单向需求满足语境下的预期性质，而基于LLM的方法则引入额外的随机性与语言敏感性。因此，论文强调机器人-任务适配度应被视为一项明确的建模选择，其假设前提、鲁棒性以及决策层面的影响必须在任务分配前被充分考量。
链接: https://arxiv.org/abs/2610.27504
作者: Daniel Weiner,Raj Korpan
机构: The Graduate Center, City University of New York (CUNY)(城市大学纽约分校研究生院); Hunter College and The Graduate Center, City University of New York (CUNY)(亨特学院和城市大学纽约分校研究生院)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注: Accepted at Workshop on Foundation Models for Multi-Robot Systems (FM-MRS), IROS 2026
Abstract:Heterogeneous multi-robot task allocation depends on estimates of how suitable individual robots are for candidate tasks, yet suitability can be defined in substantially different ways. We compare ten deterministic methods and two LLM-based variants spanning rule-based, similarity-based, requirement-oriented, fuzzy, and semantic approaches. Using common robot and task profiles, we compare how the methods score and rank candidate robots, identify feasibility, and select a top robot, while examining how these outcomes change with team composition, scale, spatial context, capability uncertainty, and task representation. The methods frequently disagree despite operating on identical scenarios. Generic similarity measures can violate properties expected under a one-sided requirement-satisfaction interpretation of suitability, while LLM-based scoring introduces additional stochastic and linguistic sensitivity. These results show that robot–task suitability should be treated as an explicit modeling choice whose assumptions, robustness, and decision-level consequences should be considered before allocation.
[MA-9] WaveCommitments: Private and Verifiable Speech-Agent Marketplaces
【速读】：该论文旨在解决现代语音系统中自治语音代理（speech agents）在私密性约束下的资产交易难题，即卖家因隐私限制无法公开披露其持有的语音资产（如嵌入向量、频谱图或原始数据），而买家则无法对无法验证的资产进行合理定价。针对这一“估值与验证悖论”，论文提出了一种名为WaveCommitments的端到端解决方案，其核心在于将基于数据沙普利值（Data Shapley）的保密资产估值机制与多方多项式承诺（multivariate polynomial commitments）相结合，并运行于可信执行环境（Trusted Execution Environment, TEE）内。该方案支持选择性开示和防篡改溯源，能够在不泄露任何额外信息的前提下，证明交付资产与估值结果一致。实验表明，该方法在三个基准数据集及多种任务上，结合Whisper等先进编码器，可实现最优基准速度提升最高达20倍，同时证明文件大小减少高达39倍；此外，通过系统分析不同任务与编码器下说话者价值的动态变化，揭示了代理层面的市场定价机制。
链接: https://arxiv.org/abs/2610.26867
作者: Oguzhan Baser,Kaan Kale,Sriram Vishwanath,Sandeep Chinchali
机构: The University of Texas at Austin (德克萨斯大学奥斯汀分校); Georgia Institute of Technology (佐治亚理工学院)
类目: Computation and Language (cs.CL); Cryptography and Security (cs.CR); Multiagent Systems (cs.MA); Sound (cs.SD)
备注: 5 pages, 3 figures, under review for a conference
Abstract:Modern speech systems are evolving into ecosystems of autonomous speech agents that speak together, negotiate, and trade assets such as embeddings, spectrograms, or raw data. In such markets, sellers cannot reveal assets for evaluation due to privacy constraints, and buyers cannot price what they cannot test. Thus, we open-source WaveCommitments, an end-to-end design that pairs confidential asset valuation via Data Shapley inside a trusted execution environment with multivariate polynomial commitments. It enables selective openings and tamper-evident provenance to prove that a delivered asset matches the valuation while hiding everything else. Our extensive experiments across three known datasets and various tasks with state-of-the-art encoders, including Whisper, reveal that we speed up the best benchmarks by up to 20x while reducing their proof sizes by up to 39x. We also systematically study how a speaker’s value shifts across tasks and encoders, revealing agent-level market prices.
[MA-10] Measuring Consensus in AI Agent Networks
【速读】：该论文旨在解决多智能体系统在复杂协作任务中因信息分散与观点分歧导致的共识形成机制不明确的问题，尤其关注智能体在网络化环境下如何通过协商达成一致决策。其核心解决方案在于提出一个评估框架，将共识形成建模为智能体公开立场随时间演进并趋近共享区域的过程，以累积的通信历史作为上下文，量化分析立场轨迹、群体离散度以及公开信息表达内容。该框架的关键创新在于区分“智能体间的观点收敛”与“向完全信息解的逼近”，从而揭示决策准确性与有限分歧之间的关系，并通过两组案例研究（基于分布式事实的人员配置决策、需处理重复信息与合同取消的复杂决策）验证了分享解释可提升集体决策准确性，同时发现即使整体答案准确，仍可能存在持续性分歧。此外，通过成员变更实验进一步补充评估，强化了对信息可用性与决策质量之间关联性的理解。
链接: https://arxiv.org/abs/2610.24833
作者: Jin Gao
机构: Independent Researcher(独立研究员)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
备注: 4 pages. Accepted at the NANDA Workshop, IEEE TPS-ISA 2026
Abstract:As AI agents become increasingly connected through networks and take on tasks that require negotiation and joint decision-making, a systematic approach is needed to assess how their views converge and what information supports their decisions. We present an evaluation framework that treats consensus formation as the evolution of publicly stated positions toward a shared region, using accumulated communication history as context. The framework tracks position trajectories, group dispersion, and information expressed in public messages. Two case studies examine staffing decisions based on distributed facts and decisions requiring agents to reconcile duplicate information and a cancelled contract. Results from 1,872 completed runs associate sharing explanations with more accurate group decisions, while the second task reveals persistent disagreement despite accurate aggregate answers. Membership-change experiments provide a complementary evaluation. By separating convergence among agents from progress toward a task’s full-information answer, the framework relates decision accuracy and bounded disagreement to the information available during the exchange.
[MA-11] Large Language Models for Agent -Based Modelling: Current and possible uses across the modelling cycle
【速读】：该论文旨在解决如何将大语言模型（Large Language Models, LLMs）有效融入基于智能体的建模（Agent-based Modelling, ABM）流程中所面临的机遇与挑战。其核心问题在于，尽管LLMs在自然语言理解、文本生成、逻辑推理和信息提取等方面展现出强大能力，但其在ABM全周期（从问题定义、模型构建、仿真运行到结果文档化与传播）中的实际应用仍缺乏系统性评估与批判性反思。解决方案的关键在于以建模周期为框架，对LLMs在各阶段的应用潜力进行结构化分析，并保持批判性立场，识别其在提升建模效率与智能化水平的同时可能带来的可靠性、可解释性及伦理风险等问题，从而为未来人机协同的智能建模提供理论指导与实践路径。
链接: https://arxiv.org/abs/2507.05723
作者: Loïs Vanhée,Melania Borit,Peer-Olaf Siebers,Roger Cremades,Christopher Frantz,Önder Gürcan,František Kalvas,Denisa Reshef Kera,Vivek Nallur,Kavin Narasimhan,Martin Neumann
机构: Umeå Universitet (乌梅奥大学); UiT The Arctic University of Norway (挪威北极大学); University of Nottingham (诺丁汉大学); University of Leeds (利兹大学); Norwegian University of Science and Technology (挪威科技大学); NORCE Norwegian Research Center AS (挪威科研中心公司); University of West Bohemia (西波希米亚大学); Bar Ilan University (巴伊兰大学); University of Southern Denmark (南丹麦大学); University College Dublin (都柏林大学学院); University of Warwick (华威大学)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Software Engineering (cs.SE)
备注: 18 pages, including 2 pages of appendix, accepted for publication at the Social Simulation Conference 2025 ( this https URL )
Abstract:The emergence of Large Language Models (LLMs) with increasingly sophisticated natural language understanding and generative capabilities has sparked interest in the Agent-based Modelling (ABM) community. With their ability to summarize, generate, analyze, categorize, transcribe and translate text, answer questions, propose explanations, sustain dialogue, extract information from unstructured text, and perform logical reasoning and problem-solving tasks, LLMs have a good potential to contribute to the modelling process. After reviewing the current use of LLMs in ABM, this study reflects on the opportunities and challenges of the potential use of LLMs in ABM. It does so by following the modelling cycle, from problem formulation to documentation and communication of model results, and holding a critical stance.
[MA-12] GBClaw: An agent ic AI framework for atomistic grain boundary research and mechanistic discovery
【速读】：该论文旨在解决在多晶材料中普遍存在的复杂一般性晶界（general grain boundaries, GBs）结构与变形机制研究中的瓶颈问题。由于其结构高度复杂且界面失配度大，传统实验与建模方法难以有效开展系统性研究，导致此类晶界长期处于研究空白状态。为此，本文提出GBClaw——一种面向晶界结构与变形行为的智能体式人工智能（agentic AI）框架，其核心创新在于构建了一个集成大型语言模型（LLMs）的模块化架构，包含协调器、专业化智能体、领域专用工具、仿真接口及分层记忆系统。通过动态重规划与迭代工作流优化，该框架实现了高通量模拟与靶向机理分析的有效衔接。在镍材料14个非对称倾斜晶界的案例研究中，GBClaw成功揭示了界面缺陷结构并识别出多种变形模式；对比有无次级晶界错配位错情况下的剪切强度差异，证实了缺陷层级结构在滑移阻力中的关键作用；进一步结合研究人员引导的后续计算，发现异常高的剪切抗力源于亚稳态晶界构型。本研究证明，将智能体式AI与领域专业知识深度融合，可有效打通高通量计算筛选与复杂材料界面深层物理理解之间的鸿沟。
链接: https://arxiv.org/abs/2610.28629
作者: Yazhuo Liu,Kunqing Ding,Xiaoguo Gong,Zifeng Wang,Yichen Yang,Yifeng Che,Wen Chen,Josh Kacher,Olivier Pierron,Shigenobu Ogata,Ting Zhu
机构: 未知
类目: Materials Science (cond-mat.mtrl-sci); Mesoscale and Nanoscale Physics (cond-mat.mes-hall); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Agentic artificial intelligence (AI) systems hold immense potential for automating complex workflows in materials discovery and design, but extending them to domain-intensive mechanistic investigations remains challenging. Here, we present GBClaw, an agentic AI framework for investigating the structure and deformation of general grain boundaries (GBs). Although general GBs constitute the vast majority of real-world interfaces in polycrystalline materials, they remain understudied because their structural complexity and interfacial mismatch complicate experimental and modeling research. GBClaw integrates large language models (LLMs) into a modular architecture comprising an orchestrator, specialized agents, domain-specific tools, simulation interfaces, and hierarchical memory. Through dynamic replanning and iterative workflow refinement, the framework effectively connects systematic simulations with targeted mechanistic analysis. Applied to 14 general asymmetric tilt GBs in nickel, GBClaw characterized interfacial defect structures and identified diverse deformation modes across these incommensurate boundaries. Comparing boundary sliding with and without secondary GB misfit dislocations revealed substantial differences in shear strength, underscoring the critical role of defect hierarchies in sliding resistance. Furthermore, researcher-guided follow-up calculations traced an anomalously high shear resistance to metastable boundary structures. This work demonstrates that combining agentic AI with domain expertise bridges high-throughput computational screening with deep physical understanding of complex material interfaces.
自然语言处理
[NLP-0] NarraSpace: Multimodal Scene-Grounded Evidence for Picture-Description Dementia Classification
【速读】： 该论文旨在解决生成式语言模型在阿尔茨海默病（Alzheimer’s Disease, AD）自发言语识别中缺乏可解释性的问题，即仅依赖转录文本进行分类时无法明确揭示其决策所依据的视觉场景证据。其核心解决方案是提出NarraSpace框架，通过将词汇与语音时间戳关联至命名的图像元素，实现对预定义场景事件的语音片段检索，并构建叙事-空间描述符以量化事件覆盖度与时间组织结构。该方法利用冻结的语言模型对比AD患者与健康对照组在各事件上的文本表现，结合转录文本的意外性（surprisal）与覆盖度测量，同时保留原始语音记录，使每个事件评分可追溯至对应的源话语、语音区间及词汇匹配信息。在阿尔茨海默病自发言语识别基准测试的48名参与者测试集上，固定支持向量机分类器在50次随机种子下的平均准确率达80.88%±2.56%，概率平均值进一步提升至81.25%准确率和91.32%受试者工作特征曲线下面积（AUC）。NarraSpace通过将语言模型的测量结果与原始语音证据建立显式且可审查的关联，显著增强了分类输入的可解释性，为临床医生提供基于量化摘要与原始语音证据相结合的辅助诊断支持。
链接: https://arxiv.org/abs/2610.28616
作者: Qingming Li,Ruiyan Sun,Satoshi Nakamura
机构: The Chinese University of Hong Kong Shenzhen(香港中文大学（深圳)）
类目: Computation and Language (cs.CL)
备注: 30 pages, 3 figures
Abstract:Picture-description dementia classification uses a shared visual scene, but a transcript-level prediction alone does not identify the scene evidence underlying its inputs. NarraSpace links words and speech times to named picture elements and retrieves utterances for predefined scene events. Narrative-spatial descriptors summarize coverage and temporal organization. A frozen language model contrasts each event’s text under examples from Alzheimer’s disease and healthy-control training participants, alongside transcript surprisal and coverage measurements. Retained source records link event scores to their source utterances, speech intervals, and lexical matches. On the official 48-participant test set of the Alzheimer’s Dementia Recognition through Spontaneous Speech benchmark, a fixed support vector classifier achieves 80.88%\pm2.56% mean accuracy over fifty seeds; averaged probabilities yield 81.25% accuracy and 91.32% area under the receiver operating characteristic curve. NarraSpace makes the construction of scene-linked classification inputs explicit and inspectable by connecting event-indexed language-model measurements to their source utterances, speech timing, and extraction records. Such source-linked measurements could support clinician review of picture descriptions by presenting quantitative summaries alongside the original speech evidence.
[NLP-1] Benchmarking Arabic Dialects for Math Problems
【速读】： 该论文旨在解决低资源语言（特别是方言阿拉伯语）在数学推理任务中研究不足的问题。针对这一空白，研究提出首个手动构建的基准测试——DAR-MGSM（Dialectal Arabic Multilingual Grade School Math），用于评估方言阿拉伯语中的数学推理能力。其核心解决方案在于构建包含五个主要阿拉伯语方言及现代标准阿拉伯语（MSA）的多语言数学应用题数据集，并设计了两种平行版本：仅翻译版与文化本地化版。后者通过调整问题中的实体和参照物以适配地区文化背景，从而揭示文化语境对模型表现的影响。实验结果表明，尽管整体准确率相似，但模型在文化本地化版本上的内部推理路径存在显著差异，且阿拉伯语专用模型对本地化敏感性不一，凸显了语言变体与文化背景共同影响数学推理性能的关键因素。
链接: https://arxiv.org/abs/2610.28607
作者: Noor Abo Mokh,Teresa Lynn,Nizar Habash,Ahmed Elshabrawy,Sara Alhajeri,Alaa Elsetohy,Preslav Nakov,Alham Fikri Aji
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Mathematical reasoning remains underexplored for low-resource languages, including Dialectal Arabic. To bridge this gap, we introduce the DAR-MGSM (Dialectal Arabic Multilingual Grade School Math) Benchmark, the first manually curated benchmark for evaluating mathematical reasoning in Dialectal Arabic. DAR-MGSM consists of math word problems (MWPs) requiring multi-step reasoning and covers five major Arabic dialects alongside Modern Standard Arabic (MSA). The benchmark comprises two parallel versions: translated-only and translated culturally localized. The latter version adapts entities and references to regional varieties. We evaluate eleven Arabic-centric and multilingual large language models on DAR-MGSM. The results show that the performance generally declines on the localized version, highlighting the challenges introduced by culturally grounded adaptations. However, the magnitude of this effect varies across models, with Arabic-centric models exhibiting different levels of sensitivity to localization. A mechanistic interpretability analysis shows that translated and localized variants of the same MWPs often induce distinct internal reasoning trajectories, even when aggregate accuracy remains similar. Overall, our results show that mathematical reasoning performance depends on both linguistic variety and cultural context.
[NLP-2] Capable Teacher Myopic Scorer: Reward-Free Teacher Repair for Agent ic On-Policy Distillation
【速读】： 该论文旨在解决生成式 AI 中基于代理的在线策略蒸馏（Agentic On-Policy Distillation, OPD）方法中存在的“误差累积”问题，其核心挑战在于：尽管教师模型在学生偏离轨迹后仍能成功接管并完成任务，但学生在根据教师对每个 token 的评分进行训练时，却因缺乏对最终任务结果的全局视野而持续选择延迟性安全动作，导致任务失败。现有方法通常通过课程学习、教师引导转向或前缀重播等方式限制训练区域，但作者指出这种解释并不完整。其根本原因在于教师评分机制的“短视性”——即教师仅基于单个 token 进行评分，未能考虑整个对话或任务的最终结果，从而允许了本应被强化学习惩罚的拖延行为。为此，作者提出了从轨迹到评分器的蒸馏（Rollout-to-Scorer Distillation, R2S），其关键创新在于：首先从未经训练的学生中挖掘高分歧状态，随后由教师在这些状态下执行接管（takeover）并评估其成功率（如通过一致性或 LLM 判定的无奖励检查），再以这些接管后的轨迹为依据，对教师评分器进行微调，使其在学生与教师分道扬镳之后的所有助手 token 上输出更具结果导向的评分。该修复后的教师评分器作为唯一评分源应用于标准 OPD，无需任何任务奖励信号。实验表明，在三个环境、两个教师和四个学生组合下，R2S 在全部十二组配置中均优于当前最先进的三种 agentic OPD 方法，平均提升达 12.8 个成功率点。
链接: https://arxiv.org/abs/2610.28586
作者: Shu Zhao,Tan Yu
机构: NVIDIA(英伟达)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 17 pages, 5 figures, 13 tables. Project page: this https URL
Abstract:Agentic on-policy distillation (OPD) trains a student agent on its own trajectories while a teacher scores every token. A natural explanation of where it goes wrong is compounding error: the student’s mistakes carry its trajectories beyond the region where the teacher is reliable, so recent methods keep training inside that region, through episode-length curricula, teacher-guided turns, prefix replay, or teacher intervention where the two disagree. We show this explanation is incomplete. From the states where the student departs most from the teacher, the teacher’s own takeovers still finish the task, yet the student trained on its token scores keeps running out of turns. We attribute this to the myopia of next-token scoring. The teacher rates each token without seeing where the episode ends, so its scores let through the safe, delaying moves that outcome-based RL would punish with zero reward, a penalty agentic OPD lacks. We call this difference between what the teacher does and what it scores the rollout-scorer gap. We propose Rollout-to-Scorer Distillation (R2S), which repairs the scorer before distillation by mining high-disagreement states from the untrained student, running teacher takeovers from them, keeping the states where enough takeovers pass a reward-free check such as consensus or an LLM judge, and fine-tuning the teacher to give its takeover-informed scores on every assistant token from the divergence turn onward, without the takeover in context. The repaired teacher is frozen as the sole scorer for standard OPD, requiring no task reward during repair or training. Across three environments, two teachers and four students, R2S outperforms three state-of-the-art agentic-OPD methods in all twelve teacher-student-environment cells, by 12.8 success points on average over the best of them in each cell. Project page: this https URL.
[NLP-3] Cache Engineering for Agents
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）智能体在处理长轨迹任务时，因键值（Key-Value, KV）缓存的I/O与存储开销过大，以及轨迹长度超过模型最大上下文长度所导致的缓存管理难题。传统方法通过上下文工程对历史信息进行裁剪或摘要以压缩缓存，但此类操作会破坏已有缓存状态，引发昂贵的重新预填充（re-prefilling）开销，且过度压缩可能影响任务准确性。为此，本文提出“缓存工程”（Cache Engineering）新范式，其核心在于允许智能体直接控制活跃KV缓存，而无需修改轨迹内容或触发重预填充。具体而言，智能体可通过“drop”操作主动排除特定缓存项以避免未来注意力计算，通过“repos”操作调整保留键的位置编码，从而实现对缓存的精细化管理。研究进一步指出，历史依赖性（history dependence）是服务层面的关键挑战：相同标记序列在不同缓存操作序列下可能产生不同的缓存状态，使得传统的前缀匹配机制无法保证缓存复用的正确性。为此，作者设计了协同优化的AKV服务引擎，提供无状态接口兼容现有推理服务API，同时在内部高效管理有状态的缓存复用、驱逐与恢复。实验表明，采用工具响应滚动丢弃（rolling-drop）缓存策略相较传统阈值触发摘要方法，平均基准得分提升1.7分，平均成本降低27.7%；结合AKV引擎，输出吞吐量最高提升1.2倍，P95请求延迟加速1.5倍；递归自改进的缓存策略相比专家设计进一步降低7.1%成本。
链接: https://arxiv.org/abs/2610.28577
作者: Tianyu Fu,Ruoxi Wang,Zihan Min,Zhongqi Ren,Zhihao Li,Tengxuan Liu,Shengxu Mu,Hanling Zhang,Yu Wang
机构: Tsinghua University(清华大学); The Chinese University of Hong Kong(香港中文大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large Language Model (LLM) agents can handle increasingly complex tasks, navigating long trajectories of observations and actions. As these trajectories grow, key-value (KV) cache I/O and storage become the dominant cost. Trajectories can also exceed the LLM’s maximum context length, requiring control over active cache growth. Prior work uses context engineering to reduce cache size by trimming or summarizing history. However, modifying the trajectory prefix invalidates subsequent cached states and triggers costly re-prefilling of retained context. Task accuracy can also suffer when too little context is retained to limit re-prefill costs. We propose cache engineering as a new scheme. It allows agents to directly control their active KV cache without modifying the trajectory or re-prefilling retained context. Specifically, agents can proactively use drop to exclude selected cache from future attention and repos to adjust retained keys’ position encodings. We formalize these cache operations and identify history dependence as a serving challenge: identical token sequences can yield different cache states under different cache operation sequences, making prefix matching insufficient for correct cache reuse. Thus, we co-design the AKV serving engine. It provides a stateless interface compatible with existing serving APIs, while internally managing stateful cache reuse, eviction, and restoration. Experiments show that a tool-response rolling-drop cache policy improves average benchmark scores by 1.7 points and reduces average cost by 27.7%, compared with conventional threshold-triggered summarization. With AKV, it achieves up to 1.2x the output throughput and a 1.5x speedup in P95 request latency. Recursive self-improvement of cache policies further reduces cost by 7.1% over expert design. Our code is at this https URL.
[NLP-4] Large Flow Language Models
【速读】： 该论文旨在解决大规模语言模型在实现高效并行生成时面临的挑战，尤其是如何在保持高生成质量的前提下，突破传统自回归（Autoregressive, AR）模型的串行生成瓶颈，并克服离散扩散模型在训练效率与推理速度上的局限。现有方法在千亿参数规模下缺乏有效的连续流语言建模技术，而本研究提出Lafa（Large flow language model），一种可扩展至80亿参数的高效连续流语言模型。其核心解决方案在于设计了一种分块式连续流架构：前缀部分采用因果注意力以保留自回归建模能力，块内则采用双向、异步的连续流机制，从而在保持模型表达力的同时支持分块并行解码。该设计使得模型能够基于预训练的AR模型进行高效微调，并通过前缀缓存和分块并行推理实现快速生成。实验表明，仅使用不足300亿个词元即可在0.5B、7B和8B规模上完成训练，显著优于以往扩散模型的训练效率。在指令遵循、推理与代码生成任务中，Lafa的生成质量可媲美甚至超越自回归模型，且可通过流映射实现一步或少数几步生成。尤为关键的是，连续流训练通过多标记预测效应（multi-token prediction effect）不仅保留了预训练模型的能力，还进一步增强其性能，这一特性被用于构建一种自洽的推测性解码（self-speculative decoding）机制——在单个模型内部以自回归方式验证基于流生成的草稿，最终实现2.54倍于标准AR模型的吞吐量，同时保持相当的生成质量，显著优于多标记预测和离散扩散方法。研究表明，连续流架构可作为未来大模型语言建模的可扩展基础。
链接: https://arxiv.org/abs/2610.28559
作者: Arnav Chavan,Chanhyuk Lee,Udbhav Bamba,Jaehoon Yoo,Seunghoon Hong,Aditi Raghunathan,Jinwoo Kim
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Code available at: this https URL
Abstract:Continuous flow language models offer a path to parallel generation with tools for control and acceleration unavailable to discrete diffusion, yet effective techniques for building them at the billion-parameter scale remain less established. We present Lafa (Large flow language model), a fast, high-performing flow language model scaled to 8B parameters in a data-efficient manner. Lafa is a blockwise flow model with causal attention on the prefix and bidirectional, asynchronous flow within each block. This design enables efficient training by adapting pretrained autoregressive (AR) models and fast inference via prefix caching and blockwise parallel decoding. We train Lafa at 0.5B, 7B, and 8B scales with fewer than 30B tokens, more efficient than previous diffusion models, and evaluate it on instruction following, reasoning, and coding, where the learned flow matches or surpasses the quality of AR and can be distilled into flow maps for one- and few-step generation. Notably, flow training preserves or even improves pretrained AR capabilities via a multi-token prediction effect, which we leverage in a self-speculative decoding scheme that verifies flow-based drafts autoregressively within a single model. This yields 2.54x the throughput of an AR model at comparable quality, outperforming multi-token prediction and discrete diffusion methods. Our findings show that continuous flows can serve as a scalable foundation for language modeling.
[NLP-5] Binary Discrete Diffusion: Candidate-Set Diffusion for Logic and Planning Tasks
【速读】： 该论文旨在解决现有离散扩散模型（discrete diffusion models）在处理序列生成任务时，因掩码（masked）和均匀（uniform）噪声机制导致单次扰动即抹除标记身份，无法显式表征多个候选解共存的局部不确定性问题。其核心解决方案是提出二值离散扩散（Binary Discrete Diffusion），通过将每个位置的噪声状态建模为一个由二值成员向量表示的候选标记集合，实现候选集的逐步增删操作。这一设计使得信息可渐进地丢失与恢复，同时噪声状态本身能够直接编码部分求解过程中的不确定性。实验表明，在数独难题（Sudoku-Extreme）、Boxoban 和图着色等高难度约束生成任务中，对称形式的二值扩散在相同架构、训练与采样预算下持续超越强基线模型，尤其在最复杂实例上表现更优；同时在语言建模任务（Text8、TinyStories、OpenWebText）上保持竞争力。结果验证了候选集作为迭代推理的有效中间表示，并揭示了前向腐蚀核（corruption kernel）对反向模型所学习计算的本质影响。
链接: https://arxiv.org/abs/2610.28543
作者: Rodrigo Casado Noguerales,Dimitri von Rütte,Bernhard Schölkopf,Thomas Hofmann,Aran Raoufi
机构: ETH Zurich(苏黎世联邦理工学院); Max Planck Institute for Intelligent Systems(马克斯·普朗克智能系统研究所); ELLIS Institute Tübingen(ELLIS图宾根研究所)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (stat.ML)
备注: 42 pages, 7 figures, 12 tables
Abstract:Discrete diffusion models iteratively denoise a noisy sequence of categorical data such as text, enabling any-order generation with self-correction and in-filling capabilities. This makes them promising candidates for reasoning, planning, and constrained-generation tasks like coding and math. Yet masked and uniform diffusion erase a token’s identity in a single corruption event, so their noisy states cannot explicitly represent partially resolved uncertainty in which several alternatives remain plausible. We close this gap with Binary Discrete Diffusion, a family of continuous-time discrete-diffusion processes whose noisy state is a set of candidate tokens per position, represented as a binary membership vector. Binary diffusion changes candidate sets through additions and removals, allowing information to be lost and recovered gradually while the diffusion state itself represents partial uncertainty. On Sudoku-Extreme, Boxoban, and graph coloring, the symmetric formulation of binary diffusion consistently outperforms strong, well-tuned masked and uniform baselines under matched architectures and controlled training, sampling, and tuning budgets, with its largest gains on the hardest instances, while remaining competitive on language modeling (Text8, TinyStories, OpenWebText). These results support candidate sets as an effective intermediate representation for iterative reasoning and show that the corruption kernel can materially change the computation learned by the reverse model.
[NLP-6] BitMTP: When Multi-Token Prediction Meets Low-Bit Large Language Models NEURIPS2026
【速读】： 该论文旨在解决超低比特（ultra-low-bit）大语言模型（LLM）在高效部署过程中，如何与多标记预测（multi-token prediction, MTP）技术协同加速解码的问题。核心挑战在于：直接结合超低比特压缩与MTP方法时，由于掩码起草（masked drafting）缺乏后续标记训练所需的中间输入，以及更宽计算量和因果键值（KV）更新带来的开销，难以在保持相近生成质量的前提下实现有效加速。为此，本文提出了一种算法-系统协同设计框架BitMTP，其关键在于通过联合的下一标记预测与MTP目标，使单一模型同时具备自回归（AR）生成与块因果起草能力；采用仅基于置信度、仅推测性、混合解码等多种策略，在不依赖独立草稿模型与验证模型的情况下灵活调节质量与吞吐性能；引入已接受前缀对齐（accepted-prefix alignment）机制，确保提交前缀、因果KV边界及CUDA Graph重播状态的一致性，从而支持连续生成；并基于此契约，选择宽度专用的低比特执行路径，复用共享KV存储上的捕获图。实验表明，BitMTP在NVIDIA H20、RTX 3090及Jetson Orin等不同硬件上均实现了显著加速，相较基线模型在五任务平均速度上提升2.1–2.4倍，在MBPP+任务上最高达3.3倍，充分验证了其在多种模型规模与硬件平台下的通用性与高效性。
链接: https://arxiv.org/abs/2610.28500
作者: Ning Zhang,Shihao Wang,Jinrui Zhang,Chaodong Xiao,Lei Zhang
机构: The Hong Kong Polytechnic University(香港理工大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
备注: Accepted at NeurIPS 2026. 16 pages, 3 figures. Code: this https URL
Abstract:For efficient deployment of large language models (LLMs), methods of ultra-low-bit backbones reduce weight storage, memory traffic, and linear-layer execution cost, while multi-token prediction (MTP) methods reduce sequential decoding steps by predicting multiple future tokens in parallel. Although complementary, the direct combination of these methods cannot accelerate decoding at comparable quality: masked drafting lacks intermediate token input supplied in next-token training, while wider computation and causal key-value (KV) updates can offset the savings from fewer steps. To address these challenges, we propose BitMTP, an algorithm-system co-design framework for efficient MTP on ultra-low-bit LLM backbones. A joint next-token prediction and MTP objective equips a single model with both autoregressive (AR) generation and block-causal drafting. Confidence-only decoding, speculative-only decoding, and hybrid decoding provide different quality-throughput operating points without using separate draft and verifier models. Accepted-prefix alignment keeps the committed prefix, materialized causal KV boundary, and CUDA Graph replay state consistent for continued generation. BitMTP uses this contract to select width-specialized low-bit execution and reuse captured graphs over shared KV storage. On NVIDIA H20, BitMTP with hybrid decoding achieves five-task mean speedups of 2.2x on BitNet b1.58-2B and 2.1-2.4x on 1.7B-4B Bonsai models at comparable task quality. On RTX 3090 and Jetson Orin, it further achieves up to 2.0x in five-task mean throughput and 3.3x on MBPP+, respectively. These results demonstrate that BitMTP accelerates ultra-low-bit LLM generation across backbones, model sizes, and GPU types. The source code is released at this https URL
[NLP-7] Present but Hard to Read: Probing Latent Reasoning with Minimal Pairs
【速读】： 该论文旨在解决生成式模型（Generative AI）中潜在推理过程的可探测性问题，即如何准确识别和验证模型内部状态是否包含未显式写出的中间计算结果（如两步问题中的中间值a+b）。传统探测方法通常仅在模型输出答案的起始阶段读取其状态，并将探测失败视为该信息不存在的证据，但这种方法可能因探测方式不当而产生误判。本文的关键解决方案是提出一种新的探测协议，结合合成零假设（synthetic null）以区分输入复制与输入混合所导致的信号，从而更可靠地检验模型内部是否存在被计算但未写入的中间值。实验表明，在循环模型Ouro-1.4B和标准Transformer Qwen2.5-1.5B中，尽管某个数字在输入中被复制并存在于最后一轮提示的隐藏状态中（读出置信度约0.9），但在逐次单个提示探测的常规设置下，该信号几乎无法被捕捉（Ouro为0.36，Qwen为0.19），说明探测效率受上下文分布影响显著。进一步分析发现，该数字的变化是微弱而非密集的，其方向上的扰动远小于随机替换单词所带来的状态偏移，且由于所有变化都留下痕迹，成功探测本身也具有歧义性。在预注册的双步问题研究中，该协议检测到在模型书写答案最后一位数字前存在微弱的中间值信号，且其符号与运算类型一致，但部分符号相关性可能源于输入混合效应。更重要的是，若将中间值固定于第一轮循环的另一问题中，该值虽可读却不会影响当前问题的答案书写，证明其虽存在于状态中但未被实际使用，揭示了“可读”与“可利用”之间的关键区别。
链接: https://arxiv.org/abs/2610.28498
作者: Aur Shalev Merin
机构: Independent researcher(独立研究员)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 20 pages, 5 figures
Abstract:Probing studies of latent reasoning often read a model’s state where the answer begins and take a failed probe as evidence that a value is absent. We test this with a digit copied from the question, in the looped model Ouro-1.4B and the standard transformer Qwen2.5-1.5B. Comparing prompts that differ only in that digit shows it is present at the last prompt token, in a code shared across questions (readout about 0.9). A probe that sees one prompt at a time, the usual setting, barely reads it: 0.36 in Ouro and 0.19 in Qwen even after 5,249 training questions. The digit’s change is small rather than crowded: random noise lands on equally busy directions, while a random word in its place moves the state far more and is readable. Since every varied token leaves a trace, a successful probe is ambiguous too. We give a protocol, including a synthetic null for generic mixing, for testing computed values against copies and mixtures of the inputs. In pre-registered studies of two-step problems it finds a small signal for the unwritten intermediate a+b just before Ouro writes the answer’s last digit, and its sign follows the operation, though mixing alone produces much of that sign-following part. Holding that value at another problem’s from the first loop on does not move the written digit toward the other problem’s, while the same patch of the answer does: the value is readable but not used.
[NLP-8] OptionWorld: A Benchmark for Diagnosing Failures of LLM s as Long-Term Personal Assistants
【速读】： 该论文旨在解决大语言模型（LLM）作为长期个人助理在跨日任务中因信息持续变化而难以有效追踪和完成任务的问题。当前的系统虽能执行信息获取或启动操作，但无法确保任务最终完成，其核心缺陷在于缺乏对任务状态的持续维护能力。解决方案的关键在于构建名为OptionWorld的基准测试框架，通过模拟日常生活中的任务流程，系统性地追踪信息获取、判断、行动及结果，从而揭示现有模型在任务延续性方面的三大主要问题：所需信息未能被获取或整合、可用行动机会被忽略、以及代理自身施加了任务本身并不要求的限制。实验进一步表明，补充缺失信息可恢复部分任务成功率；行动提醒可重启中断的工作，但后续步骤仍可能错过截止时间；修正错误的禁止性注释则能恢复合法的查询权限。这些发现明确了当前模型在保持任务状态方面的能力短板，即如何将事实与未完成工作关联、在截止前持续推进任务，以及区分真实约束与代理的主观假设。
链接: https://arxiv.org/abs/2610.28491
作者: Ruoyu Zhang,Zhaocen Liu,Weiyi Tian,Melissa Feng,Shilong Liu
机构: University of British Columbia (不列颠哥伦比亚大学); University of Chicago (芝加哥大学); Princeton University (普林斯顿大学)
类目: Computation and Language (cs.CL)
备注: 31 pages, 3 figures
Abstract:Long-term assistants must follow unfinished tasks across days as information and circumstances change. Finding information or starting an action, however, does not ensure that the task gets completed. We introduce OptionWorld, a benchmark that traces information acquisition, judgments, actions, and outcomes in simulated daily-life tasks to identify where current LLMs fall short as long-term personal assistants. In runs with archived decision inputs, the key facts remain in the model’s last input before the deadline in 35.5% of failed instances. Failure analysis identifies the three most frequent problems: needed information is not obtained or combined, available opportunities to act go unused, and agents impose restrictions that the task does not require. Three follow-up experiments show that supplying missing facts restores task success in some cases; action reminders can restart work while subsequent steps still miss the deadline; and correcting a supplied note that wrongly forbids contact restores permitted inquiry. These findings locate shortcomings in maintaining task state: connecting facts to unfinished work, following through before deadlines, and separating actual restrictions from the agent’s assumptions.
[NLP-9] Asymmetric Looped Transformers: Fixed Global Representations for an Iterative Decoder
【速读】： 该论文旨在解决循环变压器（Looped Transformer）在深度扩展过程中，因上下文构建与状态迭代优化在同一递归计算中耦合而导致的表征能力受限问题。其核心挑战在于：传统全栈循环架构未能显式区分“一次性上下文编码”与“迭代状态精炼”的不同功能角色，从而限制了模型对长期依赖和动态推理的建模能力。为此，论文提出非对称循环变压器（Asymmetric Looped Transformer, AsymLoop），其关键创新在于将模型结构解耦为两个独立模块：一个单次前向传播的因果编码器（causal encoder），负责生成各层固定的全局键值（Global KV）缓存；以及一个权重共享的递归解码器（recurrent decoder），基于自身演化状态动态生成局部键值（Local KV）。通过在单一softmax下联合注意力机制融合全局上下文与局部状态信息，实现持久上下文与当前中间状态的高效交互。此外，采用重尾引导的分层学习率策略分别优化编码器与解码器。实验表明，在相同有效深度与训练令牌预算下，AsymLoop在四个规模（60M–1B）的FineWeb-Edu数据集上均取得更低的验证困惑度和更高的下游平均准确率；在1B规模下，零样本常识推理任务平均准确率从49.58%提升至50.85%，并进一步在更长训练周期、OLMo型骨干网络及稀疏专家混合（sparse mixture-of-experts）模型中验证了该设计的普适性与有效性。结果表明，非对称的上下文编码与状态精炼分离机制是提升递归语言模型性能的有效架构范式。
链接: https://arxiv.org/abs/2610.28489
作者: Cheng Liang,Tengfei Zhang,Wenjie DU,Jianyu Wu,Changlian Ma,Shuchen Pu,Ya Zhang,Yanfeng Wang,Zhouhan Lin
机构: Shanghai Jiaotong University (上海交通大学); University of Science and Technology of China (中国科学技术大学); Nanyang Technological University (南洋理工大学); Nanjing University (南京大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Looped Transformers increase effective depth by repeatedly applying shared layers, offering a way to scale computation without increasing parameter count. However, full-stack looping places context construction and iterative state refinement within the same recurrent computation, without explicitly distinguishing their roles. We introduce the Asymmetric Looped Transformer (AsymLoop), which separates single-pass context encoding from recurrent state refinement. A single-pass causal encoder produces layer-specific Global key-value (KV) caches that remain fixed across iterations, while a weight-tied recurrent decoder generates Local KV from its evolving state. Joint attention over both sources under a single softmax allows each update to integrate persistent contextual information with the current intermediate state. For training, we apply a heavy-tail-guided layerwise learning-rate method to the encoder and decoder. Experiments on FineWeb-Edu across four model scales, from 60M to 1B, show consistently lower validation perplexity and higher average downstream accuracy than Full Loop under matched effective depth and training token budgets. At 1B, AsymLoop improves average zero-shot accuracy across seven commonsense reasoning benchmarks from 49.58% to 50.85%. Further experiments extend these benefits to longer training budgets, an OLMo-style backbone, and sparse mixture-of-experts models. These results support asymmetric context encoding and state refinement as an effective design for recurrent language models.
[NLP-10] GPT -6-Astra Understands Touch and Outperform Vision–Tactile–Language–Action Models
【速读】： 该论文旨在解决通用多模态模型在机器人操作中对真实世界物理交互本质的理解不足，以及如何利用这种理解来指导动作决策的问题。其核心挑战在于接触丰富的操作任务中，机器人需基于触觉反馈进行感知与动作调整，而现有方法依赖特定预训练和任务级微调，导致数据与训练成本高昂。解决方案的关键在于提出GPT-as-Tactile-Policy框架，将专家示范总结为任务指引与触觉参考，用于接触评估与纠正动作；结合多模态观测与交互历史，通过GPT-6-Astra模型生成控制指令并实时响应触觉反馈，实现无需额外机器人专用预训练或任务微调的端到端控制。实验表明，该方法在五个UniVTAC任务上平均成功率达93.8%，尤其在“插入孔洞”任务中成功率从31%提升至99%，验证了触觉反馈驱动的探测与修正机制的有效性，充分体现了基于专家经验与在线反馈的接触引导式操作能力。
链接: https://arxiv.org/abs/2610.28457
作者: Ziyong Wang,Yupei Lin,Runxuan Huang,Liang Lin,Dongyu Zhang,Guangrun Wang
机构: Sun Yat-sen University (中山大学); X-Era AI Lab; The Hong Kong Polytechnic University (香港理工大学)
类目: Robotics (cs.RO); Computation and Language (cs.CL)
备注: this https URL
Abstract:General-purpose multimodal models have shown promise as direct robot action policies, yet whether they grasp the physics underlying real-world interactions and can apply that understanding to guide action remains underexplored. We investigate this question using contact-rich manipulation as a representative testbed, examining tactile understanding as a concrete instance of this broader capability. Contact-rich manipulation requires interpreting tactile feedback and adjusting actions after obstruction. Existing learned tactile policies acquire these capabilities through specialized pretraining and per-task optimization, with substantial data and training costs. We present GPT-as-Tactile-Policy, which summarizes expert demonstrations into task guidance and tactile references for contact assessment and corrective action. GPT-6-Astra combines this context with multimodal observations and interaction history to generate control commands and adjust actions through contact feedback. Without additional robotics-specific pretraining or per-task policy fine-tuning, our method achieves 93.8% mean success across five UniVTAC tasks. Overall performance is comparable to state-of-the-art policies, with higher success on selected tasks. On Insert Hole, tactile feedback improves success from 31% to 99%. Execution analysis further demonstrates contact-guided probing and correction, supporting manipulation progress through explicit expert experience and online feedback.
[NLP-11] MythoSemantic: An Interactive System for Discovering Semantic Parallels Across Mythological and Religious Traditions
【速读】： 该论文旨在解决神话与宗教文本跨文化比较研究中传统人工分析方法劳动强度大、难以规模化的问题。其核心解决方案是利用神经语言模型将不同文化的文本嵌入到共享的语义空间中，使用户能够在交互式界面中高效识别跨文化传统间的语义相似性。该方法的关键在于通过语义嵌入揭示文本片段在宏观上与已知历史关联相符的语义结构，并量化词汇重叠与专有名词在解释这些结构几何特征中的贡献程度，从而为跨文化文本比较提供可计算、可量化的分析框架。
链接: https://arxiv.org/abs/2610.28452
作者: Nikita Ganziuk,Innokentiy S. Humonen,Ilya Makarov,Pavel Lebedev
机构: HSE University(高等经济大学); Trusted AI Research Center, RAS(可信人工智能研究中心, 俄罗斯科学院); AI Talent Hub, ITMO University(人工智能人才中心, 圣彼得堡国立信息技术机械与光学大学)
类目: Computation and Language (cs.CL)
备注: 12 pages, 4 figures, 8 tables
Abstract:We present a tool for the comparative study of mythological and religious texts. Traditional manual analysis is highly labor-intensive and difficult to scale. To address this, we use neural language models to embed texts into a shared semantic space, so that users can find semantic parallels across cultural traditions in an interactive interface. Embedded fragments form distinct semantic structures whose coarse geometry aligns with well-documented historical relations between traditions. We quantify how much of that geometry is already explained by lexical overlap and by proper names. The current system is available at this https URL, and the installable package is available at this https URL.
[NLP-12] Every Model Leaves a Trace: Multi-Resolution LLM Fingerprinting
【速读】： 该论文旨在解决生成式AI（Generative AI）模型在生成文本中留下的模型特异性痕迹（fingerprint）的可分辨性问题，特别是当明显特征被削弱时，这些痕迹仍能在多大程度上区分相关模型。其核心解决方案在于系统性地评估指纹在不同层次上的分辨率，涵盖模型层面（如不同模型家族、蒸馏模型及微调变体）和信号层面（行为特征、风格特征、词汇特征与语义证据）。研究通过逐步抑制表面提示、改写输出内容、调整文本量，并测试跨提示分布与跨语言场景下的泛化能力，发现即使在去除显性线索后，模型特异性信号依然存在，且在蒸馏模型甚至同一基础模型的LoRA适配版本中仍可检测。其中，语义表征在跨语言情境下表现出最强鲁棒性，表明模型身份信息可在多种粒度下被恢复，但其可恢复性取决于模型间的相似度以及文本中可用的证据类型。
链接: https://arxiv.org/abs/2610.28434
作者: Mohammad Reza Modarres,Armin Tourajmehr,Artin Tavasoli,Yadollah Yaghoobzadeh,Mohammad Taher Pilehvar
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models leave model-specific traces in their generated text, but it remains unclear how finely these fingerprints can distinguish related models and which signals remain informative as obvious cues are reduced. We study this through fingerprint resolution, considering both model resolution, from different families to distilled and adapted variants, and signal resolution across behavioral, stylometric, lexical, and semantic evidence. We progressively suppress surface cues, paraphrase outputs, vary the amount of available text, and test generalization across prompt distributions and languages. Our results show that fingerprints persist beyond broad family differences, remaining detectable in distilled models and even LoRA adaptations of the same base model. Although attribution weakens as surface and lexical cues are altered, substantial model-specific signal remains. These signals transfer across prompt sources and languages, with semantic representations proving most robust across languages. These findings show that model identity is recoverable at multiple resolutions, but that recoverability depends on model similarity and the evidence available in the text.
[NLP-13] he Instruction Tax on Factual Recall: Base Models Can Beat Their Instruction-Tuned Counterparts
【速读】： 该论文旨在解决自动知识库构建（AKBC）语言模型知识三元组抽取（LM-KBC）2026共享任务中，如何在不使用检索或微调的前提下，仅通过提示工程（prompting）从大规模语言模型（LLM）中高效生成高准确率的主体-关系对所有可能对象的问题。其核心挑战在于，在无外部知识增强与模型参数更新的情况下，如何最大化模型的常识推理能力与事实召回率。解决方案的关键在于发现并规避“指令调优副作用”（instruction tax），即尽管指令调优版本在对话接口下表现良好（得分0.5681），但其权重导致事实召回率下降；而采用基础模型（base model）配合文本补全式提示（text completion prompting）可实现更高性能（0.6836）。最终方案简洁有效：利用训练数据中的若干示例进行提示，生成64至256个候选输出，并保留重复出现的结果以提升可靠性；此外，针对不同关系类型设计了两个专项策略：一是优化大规模答案集的召回率，二是通过判断人物生存状态来提升死亡地点预测的准确性。该方法凸显了提示工程设计对模型表现的决定性影响，超越了模型架构本身的重要性。
链接: https://arxiv.org/abs/2610.28407
作者: Christophe Khalil,Aya Mourad
机构: American University of Beirut(贝鲁特美国大学); Sorbonne Université(索邦大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:We present our solution to the AKBC LM-KBC 2026 Shared Task, where we achieved a score of 0.6836 on the official leaderboard, more than double the baseline of 0.2964. This task requires generating all possible objects for subject-relation pairs using a language model with up to 32 billion parameters, with no retrieval or fine-tuning involved. We show that the prompting methodology is as important as the model architecture: while the instruction-tuned model version achieves 0.5681 with its chat interface, the base model achieves 0.6836 when prompted using text completion. Thus, instruction tuning decreases factual recall; we refer to this effect as the instruction tax, which is caused by the model’s weights rather than the prompting method. Our final solution is fairly straightforward: we prompt the base model with some examples from the training data, generate 64–256 outputs, and retain only the repeated ones. Two specialized methods are then applied for each relation: one improves the recall of large answer sets, and the other verifies a person’s vitality before predicting their place of death.
[NLP-14] MAS-Guard: Unified Guardrails from Individual Agents to Multi-Agent LLM Systems
【速读】： 该论文旨在解决生成式 AI（Generative AI）在单智能体与多智能体系统中因自主规划与执行复杂工作流而引发的安全风险，包括不安全行为、信息泄露及对抗性攻击等问题。尤其在多智能体系统中，由于角色异构、权限分散以及部分参与者不可信或行为不透明，传统防护手段难以实现全面覆盖。现有防御机制通常仅针对单一威胁场景，缺乏对本地执行、系统级交互与执行控制机制的统一保障。为此，本文提出MAS-Guard——一个面向包含可编程与非可编程智能体的单/多智能体应用的统一安全框架。其核心在于构建分层协同的防护机制：在本地层面，通过“代理护栏”（Agent Guardrail, A-Guard）引入任务特定专家模型，对智能体事件进行动态评估；在系统层面，结合开发者掌控的“代理宪章”（Agent Charters）与控制平面（Control Plane），实现对路径结构、代理能力、消息格式等的确定性校验，以及对角色语义和消息内容的语义分析。两层机制共同支持“允许、反思、确认、阻断”四类统一干预策略，形成闭环管控。实验在六个涵盖多样化场景的基准上验证了该框架的有效性，结果表明相较于未受保护的智能体，MAS-Guard使攻击成功率降低71.5%至86.3%，同时在良性任务上保持了可接受的实用性。
链接: https://arxiv.org/abs/2610.28397
作者: Mohammad Kachuee,Phu Mon Htut,Sheng Li,Nikolaos Pappas,Miguel Ballesteros Martinez,Yassine Benajiba,Sujith Ravi,Dan Roth
机构: Oracle AI
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Agentic AI extends large language models (LLMs) from generating responses to planning and executing complex workflows, substantially expanding their utility while exposing applications to unsafe actions, information leakage, and adversarial manipulation. Multi-agent systems further expand these capabilities through specialization and collaboration, but distributed authority, heterogeneous roles, and opaque or potentially compromised participants make safeguarding even more challenging. Existing defenses address important individual threats but provide fragmented coverage across local execution, system-level interactions, and enforcement mechanisms. We introduce MAS-Guard, a unified framework that closes this gap for safeguarding single- and multi-agent AI applications containing both instrumentable and opaque agents. Within instrumentable agents, the proposed local Agent Guardrail (A-Guard) layer applies task-specific experts to assess agentic events. At the system level, developer-owned Agent Charters and a Control Plane combine deterministic structural checks of routes, agent capabilities, and message formats with semantic assessments of roles and message content. These complementary layers enable enforcement through a shared set of interventions: allow, reflection, confirmation, and block. Evaluated across six benchmarks covering diverse single- and multi-agent scenarios with three agent LLMs from different providers, MAS-Guard reduces the overall attack success rate by 71.5-86.3% relative to unguarded agents, while meeting the utility reference criterion on benign tasks.
[NLP-15] When Copying Is Hard: Copy-Constrained Decoding for Exact Span Reproduction NEURIPS2026
【速读】： 该论文旨在解决大语言模型（LLM）在长上下文环境中精确复现目标文本片段（exact span reproduction）的能力不足问题，尤其当目标字符串存在于上下文但模型仍出现内容偏移或无法正确终止时，暴露出源信息锚定（source grounding）与边界控制（boundary control）的缺陷。这一能力在代码生成等实际场景中至关重要，例如需原样嵌入API密钥等敏感字符串。其解决方案的关键在于提出CopyGen——一种针对冻结的预训练大模型设计的轻量级复制约束插件，将原本作为下一个词预测副产品的复制行为转变为可控制的解码过程。核心思想是将复制行为分解为三个独立决策：何时进入复制模式、复制长度的预判以及逐标记验证以决定停止时机。具体实现上，CopyGen首先判断是否进入复制模式，随后预测一个试探性复制长度，并通过自适应调整复制延续标记的对数几率（logits）进行逐标记验证，仅输出最长的有效前缀。实验表明，CopyGen在三种主流大模型架构和四项基准任务上，无需微调即可使平均精确匹配率提升35%相对性能，结合监督微调（SFT）进一步提升4%，同时显著加速生成过程。结果表明，可靠的精确复制并非模型固有属性，而是一个需要显式建模的解码阶段控制问题。
链接: https://arxiv.org/abs/2610.28393
作者: Jinghui Zhang,Lang Gao,Zongfang Liu,Ruihong Zeng,Zirui Song,Rui Yan,Kentaro Inui,Xiuying Chen
机构: MBZUAI(中东人工智能大学); Zhejiang University(浙江大学); Westlake University(西湖大学); Wuhan University(武汉大学); Wuhan Textile University(武汉纺织大学); Tohoku University(东北大学); RIKEN(理化学研究所)
类目: Computation and Language (cs.CL)
备注: NeurIPS 2026
Abstract:Precisely reproducing a target span verbatim within a long context is a fundamental capability test for LLMs: even when the target string is present in context, models often still drift within the span or fail to stop correctly, revealing limitations in source grounding and boundary control. This capability has broad practical relevance. For instance, code generation often requires embedding sensitive strings such as API keys exactly as they appear in context. In this work, we construct a benchmark to test exact span reproduction ability across contracts, scientific literature, code, and random sequences, and find that current models still make surprisingly frequent errors. To address this, we propose CopyGen, a lightweight copy-constrained plugin for frozen LLMs that turns copying from a byproduct of next-token prediction into a controlled process. Our key idea is to decompose copying into three distinct decisions: when to copy, how far to copy, and when to stop. Concretely, CopyGen first determines whether to enter copy mode, then predicts a tentative copy length, and finally verifies the generated span token by token. During verification, we adaptively adjust the logits of copy continuation tokens to incentivize or suppress continuation, and emit only the longest valid prefix. Across three LLM backbones and four benchmark tasks, CopyGen improves average exact match by 35% relative without model finetuning, and integrates seamlessly with SFT to further improve it by 4%, while substantially accelerating generation. These results suggest that reliable exact copying is not something LLMs obtain for free, but a decoding-time control problem that requires explicit modeling. Code is available at this https URL.
[NLP-16] CAPSafe: A Benchmark for Contextual Action and Plan Safety Reasoning in Embodied Agents
【速读】： 该论文旨在解决基于大语言模型（LLM）的具身智能体在复杂家庭场景中进行安全推理的能力不足问题，尤其关注其在长时序动作规划中的安全性。其核心挑战在于：现有模型虽能较好识别单一动作的安全性（如根据物体属性推断潜在风险），但在面对多风险共存、情境依赖性强且需长期规划的任务时，难以生成连贯且安全的行动序列。解决方案的关键是构建一个双层级的安全评估基准——上下文动作与计划安全基准（CAPSafe），该基准包含两个层面：一是针对单个动作在特定环境上下文中是否安全的评估，二是对智能体完成复杂高危家庭任务时生成长时序计划的安全性评估。该基准特别强调情境性危险（由对象属性推断）、多重危险共现以及按相对严重程度标注的细粒度分析能力，从而系统揭示了当前主流闭源与开源大模型在从短时安全判断到长时安全规划之间的显著能力鸿沟。实验表明，尽管部分模型在单步动作安全识别上表现优异（最高超90%准确率），但在长时序规划中仍存在明显短板，仅有Gemini-3-pro达到78.19%的安全成功率，其余模型普遍低于65%，凸显出鲁棒性长序列安全推理仍是亟待突破的核心难题。
链接: https://arxiv.org/abs/2610.28381
作者: Youmna Farag,Svetlana Stoyanchev,Simon Keizer,Mohan Li,Rama Doddipatla
机构: Cambridge Research Laboratory, Toshiba Europe Ltd(剑桥研究中心，东芝欧洲有限公司), UK(英国)
类目: Robotics (cs.RO); Computation and Language (cs.CL)
备注:
Abstract:Reasoning about safety is essential for embodied agents powered by large language models (LLMs). We introduce the Contextual Action and Plan Safety benchmark (CAPSafe), designed to evaluate safety reasoning at two levels: (1) an action-level dataset that tests whether a single action is safe given its environmental context, and (2) a plan-level dataset that evaluates whether agents can generate safe long-horizon plans for complex hazardous household tasks. CAPSafe models safety conditions where hazards are situational (inferred from object properties), multiple hazards can co-occur within the same environment, and each is annotated with a relative severity level for fine-grained analysis. Evaluating nine closed and open-source LLMs reveals a substantial gap between hazard recognition and safe long-horizon planning. Several models achieve high accuracy in recognizing unsafe actions (exceeding 90% in the best cases), yet safe long-horizon planning remains substantially harder. Gemini-3-pro achieves the highest safe success rate (78.19%), while other models remain below 65%. These results highlight that robust safety reasoning over long action sequences remains an open challenge.
[NLP-17] Object-Oriented Harness: Evolving Agent Harnesses through Object-Oriented Design
【速读】： 该论文旨在解决单一全局代理（Agent）执⾏框架在应对异构任务时因缺乏灵活性而表现受限的问题，尤其是在需要不同执行策略的复杂长时程任务中。其核心挑战在于如何高效地为多样化任务动态适配最优的提示（prompt）、工具调用与控制流逻辑。为此，论文提出了一种面向对象的代理执⾏框架——面向对象执⾏框架（Object-Oriented Harness, OO-Harness），其关键在于利用面向对象设计思想，通过继承机制构建可演化、可复用的执⾏框架层级结构：在训练阶段，进化代理（evolution agent）基于历史执行记录自动生成一系列专业化子类执⾏框架，每个子类继承自基类或已有专用框架；在测试阶段，通过训练过程中积累的路由知识（routing lessons），由执⾏框架路由器（harness router）为每项任务精准选择最合适的专用框架。该方法显著提升了对多类型任务的适应性，在涵盖商业工作流、办公场景及工具使用等三类长时程基准上的实验表明，相较基础框架最高实现15%的性能提升，且消融实验验证了其核心组件的有效性。
链接: https://arxiv.org/abs/2610.28380
作者: Gaotang Li,Bhavana Dalvi Mishra,Yifei Ming,Yanfei Chen,Zhongying CuiZhu,Lesly Miculicich,Long T. Le,Rui Meng,Chun-Liang Li,Hanghang Tong,Chen-Yu Lee,Tomas Pfister
机构: University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校); Google Cloud AI Research(谷歌云人工智能研究)
类目: Computation and Language (cs.CL); Software Engineering (cs.SE)
备注: 20 pages, 10 figures
Abstract:Agent harnesses are executable code that specifies the prompts, tools, and control flow around a language model. A single global harness can struggle with heterogeneous tasks that require different execution strategies. In this work, we introduce Object-Oriented Harness (OO-Harness), a harness optimization framework that evolves a hierarchy of executable harness classes using object-oriented design. During training, an evolution agent uses execution records to create a family of specialized harnesses, each implemented as a subclass of the base or an existing specialized harness via inheritance. At test time, a harness router uses routing lessons accumulated during training to select a specific harness for each task. Across three long-horizon benchmarks spanning business workflows, workspace tasks, and tool use, OO-Harness achieves the best performance among the compared methods, with an absolute improvement of up to 15% over the base. Extensive ablation studies examine the contributions of its core components.
[NLP-18] Beyond a Single Sentiment Score: Topical and Affective Dimensions of Emoji Meaning on Twitter
【速读】： 该论文试图解决的问题是：当前对表情符号（emoji）的语义理解过于简化，普遍将其视为仅由单一情感评分（sentiment score）表征的紧凑情感符号，忽略了其在不同语境下的复杂语义表现。研究发现，当表情符号出现在高度凝聚的主题语境中时，其群体标注的情感极性反而更为极端，而非更弱，这表明主题一致性与情感分散性是表情符号意义的两个独立维度。解决方案的关键在于提出并验证“稳定性差异”（stability delta）这一新度量指标，通过计算句向量在类内与类间余弦相似度之差，量化表情符号在语境中的主题凝聚程度，并揭示其与情感极性的正相关关系。实证结果表明，该相关性在多种统计检验（包括置换检验、BCa自助法、留一法等）下稳健，且不受表情符号频率的影响，从而证明了表情符号语义的多维特性。
链接: https://arxiv.org/abs/2610.28337
作者: Rohan James,Ioannis Christofilogiannis,Yuuv Jauhari,Shreya Shah,Gauri Pethkar
机构: Trinity College Dublin (都柏林三一学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI)
备注: 12 pages, 4 figures, 8 tables
Abstract:Emoji are widely treated as compact affective tokens whose meaning is summarised by a single sentiment score. We argue that this picture is incomplete. From a corpus of 16,575 English tweets and a per-emoji metric we call \emphstability delta (within-class minus between-class cosine similarity of sentence embeddings), we show that emojis used in tightly clustered topical contexts also exhibit \emphmore polarised crowd-sourced sentiment, not less. Validated against the Kralj Novak et al.\ (2015) sentiment lexicon, the central correlation is Spearman \rho = +0.361 (Holm p = 0.015 , 95% bootstrap CI [+0.10, +0.59] ); against the independent Ferré et al.\ (2025) Spanish 13-emotion ratings, \rho = +0.509 ( p = 0.0003 ). The result survives permutation testing, BCa bootstrap, leave-one-out jackknife, partial correlation controlling for emoji frequency, and Holm correction across a four-test family. Topical coherence and affective dispersion are distinct dimensions of emoji meaning. As a downstream check we report leakage-free five-fold cross-validation results for sarcasm classification on the Emoji Sarcasm Dataset; all three baselines (a gated multimodal architecture, fine-tuned BERTweet features, and fine-tuned RoBERTa features) score above 0.97 macro F1, indicating the dataset has saturating emoji-label cues. We frame the classifier as supporting evidence rather than the contribution.
[NLP-19] Disclosure Begets Disclosure: A Self-Reinforcing Privacy Loop in LLM Agents with Persistent Memory
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）代理在对话中使用持久记忆（persistent memory）时引发的隐私泄露问题，特别是用户个人属性在多轮交互中被逐步暴露的风险。其核心问题是：在对话过程中，代理是否查阅记忆会改变用户的后续发言内容，从而导致原本隐含的个人属性通过累积的对话记录被逐渐推断出。研究发现，尽管查阅记忆可提升对话质量，但会显著加剧“经验性泄露”（empirical leakage），即系统从完整对话记录中正确推断出用户属性值的比例随会话轮次增加而持续上升。其关键机制在于一种非对抗性的正向反馈循环：代理在回应中基于记忆引入早期属性值作为预设前提（presuppositions），用户往往默认接受或不纠正这些信息，使这些推测性内容被记录并固化；随着时间推移，这些被记录的错误或假设性信息反过来影响后续预测，形成自我强化的泄露闭环。该闭环难以由用户自行打破——纠正错误信息会暴露属性本身，回避讨论则使系统倾向于预测敏感值，而错误信息一旦留存便趋于根深蒂固。研究指出，最有效的缓解策略是在代理生成回复阶段进行干预：通过指令要求代理避免陈述或隐含提及个人属性，可在几乎不牺牲对话质量的前提下将泄露水平降至无记忆状态下的基准水平；相比之下，对用户输入或存储数据进行过滤或匿名化处理虽能降低泄露，但会显著损害响应质量。因此，解决方案的关键在于在代理端主动控制属性信息的显式与隐式传播，而非依赖事后处理。
链接: https://arxiv.org/abs/2610.28315
作者: Masahiro Kaneko,Hiromi Arai,Timothy Baldwin
机构: MBZUAI(中东人工智能大学); Third Intelligence; RIKEN AIP(理化学研究所人工智能中心)
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Large language model (LLM) agents increasingly use persistent memory to personalize responses, and personal attributes in this memory pose a privacy risk if exposed. We ask how consulting memory during dialogue changes what users go on to say, and thus what the memory comes to reveal. In multi-session simulations, a simulated user with 24 personal attributes talks with a support agent that does or does not consult memory, and in both settings attributes are inferred from the full accumulated record. Across four support models, memory raises response quality but also empirical leakage, the share of attribute values correctly inferred from the record, by a margin that widens over sessions. The cause is a loop in ordinary, non-adversarial assistance. Memory-conditioned responses carry earlier values forward, often as presuppositions, and users affirm them or leave them uncorrected, so that cues and the agent’s own guesses become recorded values. The loop reinforces itself, and removing the agent’s mentions of an attribute lowers its later disclosure about twice as much as single-turn effects predict. Users cannot leave it on their own, since correcting a value discloses the attribute, declining to discuss it shifts later predictions toward sensitive values, and incorrect values left in place become entrenched. The loop is best weakened at the agent’s response. Instructing the agent not to state or presuppose attributes brings leakage to about the no-memory level while keeping most of the quality gains, whereas filtering or anonymizing what users say or what is stored gives up much of them.
[NLP-20] From Abstraction to Action: DSLs for Natural Language Abstraction Interpretation and Execution
【速读】： 该论文旨在解决自然语言指令中高阶抽象与具体动作之间存在的语义鸿沟问题，即当指令如“将积木堆成金字塔”这类抽象表述时，模型难以准确识别说话者意图并将其分解为一系列可执行的细粒度动作。其核心解决方案是引入领域特定语言（Domain-Specific Language, DSL），作为一种显式的表征层，通过已定义的语义原子（primitives）对常见抽象进行命名，并映射到具体的动作序列。研究在视觉引导的指令遵循任务Hexagons上验证了DSL的有效性，结果表明，基于人工设计的DSL进行代码生成相比直接推理显著提升性能，最高可达10个绝对F1点，尤其在高抽象层级下优势明显；而在Text2CAD基准上的复现进一步确认了这一规律。然而，由语言模型自动生成的DSL仅能恢复部分性能增益，且结果不稳定、泛化性差，远不及人工构建的DSL。因此，论文指出：合适的表征（即高质量的DSL）能够有效弥合抽象与动作间的差距，但如何自动学习和构建这种表征仍是当前模型能力的短板，也是未来研究的重要突破口。
链接: https://arxiv.org/abs/2610.28311
作者: Amit Lisha,Reut Tsarfaty
机构: Bar-Ilan University (巴伊兰大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Natural-language instructions often rely on high-level abstractions, such as asking an agent to ``stack blocks in a pyramid.‘’ Interpreting abstract instructions requires models to both identify the abstraction intended by the speaker, and unroll it into the many concrete actions that instantiate it. Language models struggle here: prior work has shown that performance degrades as the level of abstraction in the instruction rises. To bridge the gap between natural language abstraction and concrete actions, we propose to borrow from coding environments the notion of a \em domain-specific language (DSL), an explicit representation layer whose documented primitives name recurring abstractions that expand into concrete actions. We demonstrate the efficacy of DSLs for NL interpretation on \textscHexagons, a visually-grounded instruction-following task graded for abstraction. We further ask whether language models can induce such DSLs themselves from domain data. Our results show that while code generation consistently improves over direct inference, code generation over a human-curated DSL yields dramatic gains, up to 10 absolute F1 points, reaching the human skyline, with gains concentrated at \em higher abstraction levels. We confirm this pattern on the \textscText2CAD benchmark. At the same time, model-induced DSLs recover only part of this gain, remaining inconsistent across runs and short of their human-curated counterpart. Overall, the right representation closes much of the gap caused by high abstraction — but inducing that layer of representation automatically is where models still fall short, and where we see a strong opportunity for the community to make progress.
[NLP-21] Reward Misspecification Degrades Chain-of-Thought Monitorability
【速读】： 该论文旨在解决生成式语言模型在强化学习（Reinforcement Learning, RL）训练过程中因奖励误设（reward misspecification）导致思维链（Chain-of-Thought, CoT）可监控性下降的问题。当RL环境复杂时，难以实现对任务意图的完全验证，模型可能被奖励偏离真实目标的行为，从而引发“奖励黑客”（reward hacking）现象，即模型输出虽违反任务初衷但仍获得高奖励，而其思维链仍看似合理地围绕原任务意图展开，造成安全监控失效。论文的关键解决方案是：在不依赖真实奖励或已知攻击模式的前提下，通过收集训练过程中多样化奖励下的轨迹数据，利用大语言模型（LLM）自动推断出奖励描述，并将该描述注入模型输入以引导其显式地在思维链中表述所遵循的奖励信号。这一方法显著提升了CoT的可监控性，在三个模型和三项任务上平均提升24%或42%（绝对值），使模型的推理过程明确反映实际被优化的错失奖励，从而增强对异常行为的检测能力。
链接: https://arxiv.org/abs/2610.28303
作者: Tianyi Lorena Yan,John Hewitt
机构: Columbia University (哥伦比亚大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Monitoring the chain-of-thought (CoT) of language models (LMs) is a key strategy for detecting and preventing misbehaviors for AI safety. As such, it is critical to identify reinforcement learning (RL) practices that lead to unmonitorable chain-of-thought. One such practice may be unavoidable: reward misspecification. As RL environments become more complex, it is infeasible to have perfect verification, and models are rewarded for behaviors that are different from the true task intent. We show that reward misspecification alone already degrades monitorability. We create a testbed with pairs of true reward aligned with task intent and misspecified rewards that favor a different behavior, so that reward hacking can be induced by training and labeled verifiably. Across three models and three tasks, CoT monitorability remains low as models’ hack rates increase: the CoTs keep reasoning about the task intent while the outputs violate the intent and get rewarded. To mitigate unmonitorability without prior knowledge of the true reward or the hack, we collect rollouts with diverse rewards across training steps and use a LM to infer a reward description. We then inject the description into the input, which teaches the model to verbalize the reward. This reward description training improves monitorability by an average of 24% or 42% absolute, depending on the metric, across all models and tasks, and yields CoTs that reason about the misspecified reward explicitly. Code is available at this https URL.
[NLP-22] Adaptive Replay Refinement: Allocating Test-Time Compute Between Revision and Resampling
【速读】： 该论文旨在解决生成式 AI 在推理过程中如何高效分配额外计算资源的问题，具体聚焦于在测试时缩放（test-time scaling）场景下，决策应继续优化当前解还是转向生成全新解的困境。现有方法依赖固定的精炼周期、自验证判断或需额外训练的控制器，存在计算浪费或依赖不可靠信号的风险。其解决方案的关键在于提出一种无需训练的自适应重播精炼（Adaptive Replay Refinement, ARR）方法，通过检测当前解是否达到固定点（fixed point），一旦发现精炼过程收敛，则将释放出的计算资源重新分配给独立的新样本生成。ARR采用重播解码（replay decoding）机制，将当前解视为草稿，并行验证未修改的标记（token），从而实现高效的固定点检测。实验表明，在五个数学推理基准中，ARR在同等计算开销下达到与多数投票（majority voting）相当的性能，同时将关键路径上的串行计算步骤减少最多达2.4倍；在MATH500数据集上，仅用41%的串行步骤即可达到84.6%的准确率，充分证明了通过识别精炼耗尽并动态切换至独立采样可显著提升推理效率，并揭示了自我修正的局限性。
链接: https://arxiv.org/abs/2610.28294
作者: Ahsan Bilal,Muhammad Ahmed Mohsin,Muhammad Umer,Ayesha Mohsin,Sarmad Bilal,Dean F. Hougen
机构: University of Oklahoma(俄克拉荷马大学); Stanford University(斯坦福大学); National University of Sciences and Technology(巴基斯坦国立科技大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Test-time scaling allocates additional inference compute to refining existing solutions or generating new ones. This raises a fundamental question: Should I keep improving this solution, or stop working on it and try a completely new solution? Existing methods make this decision using fixed refinement schedules, self-verification judgments, or learned controllers, which can waste compute or rely on signals that may be incorrect or require additional training. We introduce Adaptive Replay Refinement (ARR), a training-free method that detects when a stationary revision reaches a fixed point and reallocates the released compute to a new independent sample; replay decoding treats the current solution as a draft and verifies unchanged tokens in parallel, enabling efficient fixed-point detection. Across seven models from five families and seven benchmarks, ARR matches majority voting on three of five mathematical reasoning benchmarks at equal compute, while reducing critical-path serial passes by up to 2.4 times. On MATH500, ARR achieves 84.6% accuracy with only 41% of the serial passes required by majority voting. These results demonstrate that detecting refinement exhaustion enables adaptive reallocation from revision to independent sampling, improving inference efficiency while highlighting the limits of self-revision. Code is available at this https URL.
[NLP-23] Forty Questions: Identifying a Language Model from One-Word Answers
【速读】： 该论文旨在解决大模型在通过API提供服务时的“身份可识别性”问题，即即使在无明确标识的情况下，仍可通过其对通用问题（如“Name a sound”或“Name a poet”）的响应模式来推断其真实模型身份。其核心解决方案是构建并分析模型的“响应特征画像”（profile），记录其对一系列标准化问题的回答分布，并基于最大似然原则将新生成的回答归属到最可能对应的模型。实验表明，在94个模型的测试集上，仅用40个回答即可以92.5%的准确率识别出具体模型，远超1.1%的随机水平；在OpenRouter平台上对13个开源模型的96个端点进行验证时，92个能正确匹配其宣称的身份，其余四例中部分可通过关闭推理逻辑后匹配，另两例尚待进一步解析。此外，研究还发现部分模型存在异常计费行为（如Grok 4.7对单行提问计为1,247输入令牌），以及未被计费却主动输出当前日期等上下文信息的现象，揭示了模型服务中的潜在透明度与合规性问题。该方法成本低廉（每端点仅需数美分），依赖公开可获取的默认模型响应数据，具备高度可扩展性和实用性。
链接: https://arxiv.org/abs/2610.28281
作者: Tapan Parikh
机构: Cornell Tech(康奈尔科技学院)
类目: Computation and Language (cs.CL)
备注: 19 pages. Code and data: this https URL
Abstract:A model behind an API can be identified from its one-word answers to everyday questions such as “Name a sound” or “Name a poet”. A model’s profile records how often it gives each answer to each question. We attribute new answers to the model most likely to have given them. On a panel of 94 models, 40 answers identify the exact model 92.5% of the time, against 1.1% by chance. Profiles built through OpenRouter identified all ten models served by DeepInfra under blinded labels. A system prompt lowers accuracy by 6 to 16 points. We applied the method to 96 endpoints of 13 open models on OpenRouter, and 92 identify as the model they claim. Of the other four, one matches the claimed model with reasoning turned off. One serves Gemma 4 under the name Llama 4 Maverick: it names itself Gemma 4, and it is billed Gemma 4’s token counts on every message. Two remain unresolved. Billing records also show context added to first-party models. Grok 4.7 is billed 1,247 input tokens for a one-line question that costs GPT-4o only 12 tokens, and five models state the current date without being billed for it. The questions come from a public census of model defaults, and the method costs cents per endpoint.
[NLP-24] How Language Models Learn Taste: Brand Preference in LLM s
【速读】： 该论文旨在解决生成式模型在品牌推荐任务中表现出的“认知偏差”问题，即模型在被要求“说出一个品牌”时倾向于选择市场主流品牌（如Nike），而在被要求“推荐一个品牌”时却更常选择小众或专业品牌（如Brooks）。这种差异揭示了模型在推荐行为中形成了一种非显性、隐含的“偏好系统”，其根源并非对品牌事实知识的掌握，而是源于训练过程中形成的隐含偏好关联。解决方案的关键在于识别并分析这一偏好的形成机制：研究发现，预训练阶段已将推荐品牌与建议内容建立关联；监督微调阶段基本保持推荐选择不变；而偏好优化（DPO）阶段则显著推动推荐选择的转变，即使偏好数据中未明确包含目标品牌，模型仍会向一组新的品牌偏好靠拢，且这种影响可扩展至非品牌领域（如从番茄酱转向青酱）。这表明，模型的推荐行为更多依赖于其在偏好训练中通过“评判模型”所内化的一套抽象审美或价值判断标准，而非具体品牌信息，从而揭示了生成式模型中“偏好”形成的深层机制及其对输出结果的决定性影响。
链接: https://arxiv.org/abs/2610.28272
作者: Tapan Parikh
机构: Cornell Tech(康奈尔科技学院)
类目: Computation and Language (cs.CL)
备注: 11 pages
Abstract:Asked to name a running shoe brand, language models say Nike. Asked to recommend one, they say Brooks. This paper asks where models learn that taste. Across 102 models and 42 brand categories, the brand a model picks when recommending differs from the brand it names in 26. The field moves together to a shared recommendation. The Named brands are market leaders, and the Picked brands are mostly niche and specialist. Three open training pipelines show where this forms. Pretraining already attaches the recommended brand to advice. Supervised fine-tuning teaches the model to answer and mostly leaves the pick unchanged. Preference optimization (DPO) teaches the model which answer is better, and the pick moves in all three pipelines, by 11 to 24 points. The preference data contains almost no brands. DPO moves the model’s judgments toward a second set of associations even when the preferred brand is not in the preference data, and the effect reaches beyond brands. Asked to choose a sauce, models switch from ketchup to pesto. We do not know exactly how these preferences arise. We suspect they come from the judge models that decide which answers count as better during preference training. For brands, a recommendation depends less on what a model knows about them than on whether they fit this taste.
[NLP-25] Can Language Models Get “Aphasia” From Neuronal Ablation?
【速读】： 该论文旨在解决神经退行性疾病中神经元进行性丢失如何影响认知功能衰退模式这一关键问题，特别是从计算模型角度探究神经元损失与语言认知能力下降之间的关联机制。其核心解决方案在于通过系统性地对语言模型中的神经元单元（计算单元）进行渐进式移除，并结合有限微调作为补偿性可塑性的简化模拟，从而揭示不同神经网络架构在损伤下的功能退化特征。研究发现，基于循环结构的长短期记忆网络（LSTM）表现出随损伤程度平滑递减的认知功能退化模式，而基于注意力机制的Transformer模型则呈现出显著的阈值突变现象（尤其在前馈多层感知机模块移除时），或呈更平滑的逻辑斯蒂型衰减（在注意力头移除时）。具体而言，前馈模块损伤导致语义类别内错误替换与非响应率上升，类比于韦尼克型失语症；而注意力头损伤则表现为语义特异性丧失，类似原发性进行性失语症的语义变异型（svPPA）。这些结果表明，前馈与注意力路径具有功能上截然不同的失效模式，提示网络架构依赖性的故障特征可为失语综合征及神经退行性疾病提供新的计算视角。尤为关键的是，Transformer在早期注意力头损伤下仍保持较高准确率，与阿尔茨海默病患者临床症状出现前40–60%神经元丢失阶段的“无症状储备”现象高度吻合，而LSTM的稳定衰退模式则对应已知的临床症状进展过程，进一步验证了该计算模型在模拟人类神经退行性病变中的有效性。
链接: https://arxiv.org/abs/2610.28214
作者: Pratik Worah,Randall Bateman
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Progressive neuronal loss is the hallmark of neurodegenerative disease, yet how loss of neurons shapes the pattern of cognitive decline remains poorly understood. Here we systematically ablate neurons (computational units) in language models spanning recurrent long short-term memory (LSTM) and Transformer architectures and apply limited fine-tuning as a simple proxy for compensatory plasticity. With whole-unit ablation, LSTMs degrade gradually with syntactic and semantic faculties declining in parallel across mild-to-moderate damage. By contrast, Transformers exhibit late threshold-like collapse under feedforward multi-layer perceptron (MLP) ablation and smoother sigmoid-like decline under attention-head ablation. Qualitatively, MLP-ablated models continue to make within-category semantic substitutions (semantic paraphasias) alongside rising non-responses, paralleling Wernicke-type aphasia, while attention-head ablation produces a trend toward loss of semantic specificity similar to the semantic variant of primary progressive aphasia (svPPA). These findings indicate that feedforward and attentional pathways have functionally distinct failure modes, suggesting that architecture-dependent failure profiles may offer new computational perspectives on aphasic syndromes and neurodegenerative diseases as an in silico model. Notably, early accuracy retention under Transformer attention-head ablation qualitatively parallels pre-symptomatic reserve before the 40–60% neuronal loss reported by Gomez-Isla et al. (1996) when clinical symptoms emerge in Alzheimer’s disease, whereas the steadier LSTM decline parallels the symptomatic progression described by DeKosky and Scheff (1990) and Terry et al. (1991).
[NLP-26] Uncovering Audio-Text Alignment in Large Audio-Language Models through Cross-Modal Concept Neurons
【速读】： 该论文旨在解决大音频-语言模型（Large Audio-Language Models, LALMs）在仅基于文本预训练的前提下，如何实现对声学属性（如情感、说话人性别、口语语言等）的有效识别这一关键问题。尽管模型具备跨模态理解能力，但其内部音频与文本表征的融合机制尚不明确。为此，研究提出核心假设：音频与文本表征在模型内部存在一个共享的潜在子空间，即跨模态神经元（cross-modal neurons）负责联合编码多模态信息。为验证此假设，研究通过识别同时响应音频和文本输入的神经元，发现三类LALMs在四类声学属性上均存在不同程度的音频-文本神经元对齐现象。进一步的因果干预实验（包括神经元操控、引导与抑制）表明，这些跨模态神经元不仅对预测任务具有必要性，且可被主动调控，从而证实了其在多模态表征融合中的核心作用。因此，解决方案的关键在于揭示并利用跨模态神经元作为音频与文本表征对齐的内在机制。
链接: https://arxiv.org/abs/2610.28189
作者: Tzu-Quan Lin,Chih-Kai Yang,Yi-Cheng Lin,Hao Tang,Hung-yi Lee
机构: Centre for Speech Technology Research, University of Edinburgh(爱丁堡大学语音技术研究中心); Graduate Institute of Communication Engineering, National Taiwan University(台湾大学通信工程研究所); NTU Artificial Intelligence Center of Research Excellence (NTU AI-CoRE)(台湾大学人工智能研究中心)
类目: ound (cs.SD); Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
备注:
Abstract:While the language model backbone is pretrained on text alone, large audio-language models (LALMs) can, to a great extent, recognize acoustic attributes, such as emotion, speaker gender, and spoken language. However, it is unclear how the audio representation is integrated, if at all, with the other text representation before the model produces text. We hypothesize that there is a shared subspace where the audio and text representations are aligned. To test this hypothesis, we identify cross-modal neurons that are responsible for representing both audio and text. We find alignment between audio and text neurons to various degrees across three LALMs and four acoustic attribute types. Causal interventions, steering and deactivating the neurons, confirm that the cross-modal neurons are not just necessary for prediction but can also be manipulated.
[NLP-27] mporal Evidence-Parameter Co-Alignment for LLM Reasoning over Temporal Knowledge Graphs
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在时序定位推理中面临的挑战，即其静态参数记忆与现实世界知识的动态演化之间存在错位，导致在处理时间依赖性任务时表现不佳。现有方法如检索增强生成（Retrieval-Augmented Generation, RAG）和参数高效微调（Parameter-Efficient Fine-Tuning, PEFT）虽能部分提升性能，但在捕捉事件的时间有效性、顺序关系及演化依赖性方面仍显不足。具体而言，传统检索方法侧重语义相关性而忽略时间约束，造成所获证据在孤立意义上相关但整体时序碎片化；而常规PEFT方法采用固定且与输入无关的低秩更新机制，难以建模随时间演化的实体语义与关系漂移。为此，本文提出时序证据-参数协同对齐（Temporal Evidence-Parameter Co-Alignment, TEPCA）框架，通过在证据层与参数层双重对齐实现更精准的时序推理。其核心创新在于：在证据层面，引入时间约束的动态相关性引导路径（Time-Constrained Dynamic Relevance-Guided Path, DRGP）剪枝策略以保持全局一致的时序路径，并结合基于互信息（Mutual Information, MI）的剪枝机制，在时间约束下保留与查询相关的局部事实；在参数层面，设计了时间戳条件化的动态门控低秩适应（Dynamic-Gated LoRA, DG-LoRA），可根据时间上下文动态调节有效适配子空间，从而精细捕捉实体演化与时变关系语义，同时保持参数效率。实验结果表明，TEPCA在两个时序知识图谱（Temporal Knowledge Graph, TKG）基准上均达到当前最优性能，且对DG-LoRA的多数据集评估进一步验证了其在不同任务中的自适应能力、跨模型架构的可扩展性及可解释性。
链接: https://arxiv.org/abs/2610.28176
作者: Jiapu Wang
机构: Nanjing University of Science and Technology (南京理工大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) often struggle with temporally grounded reasoning because their static parametric memory is misaligned with the dynamic evolution of real-world knowledge. Although RAG and Parameter-Efficient Fine-Tuning (PEFT) provide practical ways to enhance LLMs, they remain insufficient for temporal reasoning. Existing retrieval methods typically prioritize semantic relevance while overlooking temporal validity, event order, and evolutionary dependencies, resulting in evidence that is relevant in isolation but temporally fragmented. Meanwhile, conventional PEFT methods usually employ fixed and input-agnostic low-rank updates, making it difficult to capture time-dependent entity semantics and relation drift. To address these limitations, we propose Temporal Evidence-Parameter Co-Alignment (TEPCA), a framework for LLM reasoning over TKGs. TEPCA aligns LLM reasoning with temporal knowledge at both the evidence and parameter levels. At the evidence level, it introduces time-constrained Dynamic Relevance-Guided Path (DRGP) pruning to preserve globally coherent temporal paths, together with Mutual-Information (MI)-based pruning to retain query-informative local facts under temporal constraints. At the parameter level, it integrates Dynamic-Gated LoRA (DG-LoRA), a timestamp-conditioned low-rank adaptation mechanism that dynamically modulates the effective adaptation subspace according to temporal context. This design enables LLMs to better capture fine-grained entity evolution and time-dependent relation semantics while preserving parameter efficiency. Experimental results on two TKG benchmarks demonstrate that TEPCA achieves state-of-the-art performance. Additional evaluations of DG-LoRA on eight commonsense QA datasets further verify its general input-adaptive capacity, scalability across LLM backbones, and interpretability.
[NLP-28] Africa-Papers: A 10-Billion-Token Multi-Domain Academic Corpus for African Foundation Language Models
【速读】： 该论文旨在解决当前大语言模型（Large Language Models, LLMs）预训练数据中非洲科学、医学、法律及农业学术成果严重缺失的问题。现有基础预训练数据集（如The Pile、S2ORC和Dolma）中，非洲学术文献所占比例不足全球预训练文本的0.5%，导致主流模型在处理热带流行病学、小农户农林生态、区域经济治理及本土语言结构等任务时表现出系统性偏差与失效。其解决方案的关键在于构建迄今为止最大规模的多领域泛非学术预训练语料库——Africa-Papers，该语料库整合了来自879个Parquet分片的超过844,807篇全文论文，涵盖101.5亿个标记（tokens），覆盖大陆同行评审期刊（DOAJ收录，31个非洲国家）、CGIAR CGSpace农业资源库、arXiv上非洲作者或聚焦非洲的STEM/AI预印本，以及通过OpenAlex索引的非洲机构学者发表的开放获取国际出版物。所有文本经布局感知（layout-aware）清洗，以保留章节层级、数学公式与表格数据，并配备精确的BPE标记计数与标准化的8类学科分类体系，确保高质量下游应用。该语料库已开源发布于Hugging Face Hub，由AfriSpeech组织托管，旨在推动全球基础模型的持续预训练与领域评估，促进非洲知识主权在人工智能时代的实现。
链接: https://arxiv.org/abs/2610.28173
作者: Mich-Seth Owusu
机构: AfriSpeech; Hugging Face(胡格福)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) depend fundamentally on the diversity, depth, and factual rigor of their pre-training corpora. However, contemporary foundational pre-training datasets (such as The Pile, S2ORC, and Dolma) exhibit a severe structural omission: African scientific, medical, legal, and agricultural scholarship constitutes less than 0.5% of global pre-training tokens. Consequently, state-of-the-art models suffer from systematic failure modes when reasoning about tropical epidemiology, localized smallholder agro-ecologies, regional economic governance, and indigenous linguistic structures. To resolve this representation deficit and advance African intellectual sovereignty in artificial intelligence, we introduce Africa-Papers, the largest curated multi-domain Pan-African academic pre-training corpus to date. Comprising over 844,807 full-text papers spanning 10.15 billion tokens across 879 Parquet shards, Africa-Papers unifies four primary scientific streams: (1) continental peer-reviewed journals indexed in DOAJ across 31 African nations, (2) the CGIAR CGSpace agricultural repository, (3) African-authored and African-focused STEM/AI preprints on arXiv, and (4) open-access international publications by scholars affiliated with African institutions indexed via OpenAlex. The entire corpus is processed into layout-aware, clean Markdown preserving section hierarchies, mathematical equations, and tabular data, accompanied by exact BPE token counts and a standardized 8-domain taxonomy. We release Africa-Papers openly on the Hugging Face Hub under the AfriSpeech organization to accelerate continual pre-training and domain evaluation for foundation models globally.
[NLP-29] Reading Too Much Into It? Interpretive Preferences and Commitment in LLM Mentalisation under Ambiguity
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在人际建议场景中对模糊社交情境的解读能力评估不足的问题，尤其关注模型在心理困扰情境下生成解释的偏好与置信度表现。现有心智理论（Theory of Mind）评估多依赖预设答案，而情感不确定性研究则忽视了用户背景对模型解释倾向的影响。为此，论文提出MentalLoom这一新型评估框架，包含604个模糊社交事件、匹配的抑郁与社交焦虑披露内容，以及四轮对话中不改变事件证据的情境设计。实验发现，尽管初始阶段模型更倾向于生成负面解释，但对叙述者（narrator）的积极归因显著增加；通过结构化重复提问，模型展现出更强的情感支持能力、更高的积极归因强度及置信度，且无需新增事件信息。其解决方案的关键在于引入动态用户上下文与持续对话机制，揭示用户语境可重塑模型的解释偏好与信念强度，即使底层情境的模糊性未被消除，从而为理解LLMs在心理支持中的认知行为提供了新视角。
链接: https://arxiv.org/abs/2610.28105
作者: Stephanie Fong,Yiwen Jiang,Qingyang Xu,Zimu Wang,Yaling Shen,Xiangyu Zhao,Jiahe Liu,Yinghao Zhang,Vincent Lee,Dominic Dwyer
机构: Monash University(莫纳什大学); University of Liverpool(利物浦大学); Orygen, The University of Melbourne
类目: Computation and Language (cs.CL)
备注:
Abstract:Large Language Models (LLMs) are increasingly used for interpersonal advice, potentially shaping how users interpret ambiguous social encounters. Existing Theory of Mind evaluations typically rely on predefined answers, while uncertainty research overlooks LLMs’ preferred explanations and confidence under psychological distress. To address these gaps, we introduce MentalLoom, a novel evaluation of LLM mentalisation comprising 604 ambiguous social episodes, matched depression and social anxiety disclosures, and four-turn dialogues with unchanged event evidence. Across eight LLMs, disclosures initially elicited more negative explanations overall but more positive narrator-directed interpretations. Scripted repeated questioning showed greater emotional support, stronger positive narrator-directed endorsement, and higher certainty without new event evidence. These findings show that user context can reshape interpretive preferences and commitment while the underlying ambiguity remains unresolved.
[NLP-30] BAIT: Can LLM s Find What Biases a Reader? A Benchmark for Cognitive Bias Triggers in News
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在新闻文本中识别认知偏差触发因素（cognitive bias triggers）的能力问题。认知偏差触发因素是利用人类认知捷径（cognitive biases）的语义特征，常被用于说服性写作中潜移默化地影响读者信念，而此类特征往往不易被读者察觉。随着LLMs日益承担新闻摘要、排序与问答等任务，其能否准确识别这些操纵性语言特征成为关键挑战。本文提出BAIT（Bias Annotation In Text）数据集，包含9类认知偏差触发器的标注体系，涵盖50篇英文新闻文章的篇章级与片段级人工标注，并经仲裁形成黄金标准。基于此数据集，研究定义了两项任务：一是判断文本包含哪些触发器，二是定位具体承载触发器的文本片段。实验对六种前沿LLM进行零样本及少量示例（3或10个）提示评估，结果显示，尽管模型在识别文本是否包含某类触发器上接近人类一致水平，但难以完整召回所有触发器，且在定位触发片段时表现远低于人类（不足人类一致性的50%）。更严重的是，模型倾向于过度预测具有明显表面线索（surface cues）的触发器，且其所标记的多数片段虽非真正触发器，却频繁携带该触发器的表面特征，显著高于随机文本水平。这一现象表明，当前评估的LLMs主要依赖表面线索而非深层操控机制进行判断，缺乏对说服性语言本质的理解。研究同时公开了触发器分类框架、标注手册、结构化标注数据及评测流程，为后续研究提供基础资源。
链接: https://arxiv.org/abs/2610.28092
作者: Yousra Fettach,Tijl De Bie
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Persuasive writing exploits cognitive biases, recruiting a reader’s mental shortcuts to shape belief, often without the reader noticing. We call the linguistic features exploiting such mental shortcuts cognitive bias triggers. As large language models (LLMs) increasingly mediate how people encounter news, summarizing it, ranking it, and answering questions over it, a natural question is whether they can recognize the features of a text that bias a human reader. We introduce BAIT (Bias Annotation In Text), a resource for detecting cognitive bias triggers in news. BAIT pairs a schema of nine triggers, each specified in an annotator codebook, with 50 English news articles annotated by trained annotators at the article and span levels and adjudicated into a gold standard. On this data we define two tasks, detecting which triggers a text contains and locating the spans that carry them. We then benchmark six frontier LLMs zero-shot and with three or ten in-context examples. The models approach human agreement on which triggers an article contains, but rarely recover its full set of triggers, and they reach less than half of human agreement in locating the passages that carry them. Moreover, they over-predict triggers that come with easy surface cues, and most of the passages they mark for these triggers are not triggers, yet carry the trigger’s surface cue far more often than random text. These findings suggest that the evaluated LLMs respond to surface cues rather than to the mechanisms that make a passage manipulative. We release the schema, the codebook, standoff annotations, and the evaluation pipeline.
[NLP-31] Contextual Candour: Role-Aware Graduated Disclosure for LLM s over Mixed-Sensitivity Contexts
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在混合敏感度上下文（mixed-sensitivity contexts）中面临的访问控制难题，即如何在处理从公开到机密信息的多层级敏感数据时，仅向授权用户披露其权限范围内的信息。现有方法采用二元决策机制（全答或拒绝），无法精细调控授权后输出的信息量与敏感度，导致过度拒答或无意泄露。其解决方案的关键在于提出“渐进式披露”（graduated disclosure）这一新型访问控制范式，将响应内容的详细程度与敏感性根据用户授权等级进行动态调节，而非简单地授予或拒绝访问。为此，作者将渐进式披露建模为一种具备披露意识的生成问题，并设计了一种基于奖励的优化框架，通过奖励机制引导模型生成与授权级别相匹配的抽象表达，同时对未经授权的事实泄露（包括改写、近似表述等）施加惩罚。实验在企业与临床基准上验证了该方法在显著降低信息泄露的同时，仍能保持较高的正确性与实用性，表明安全可控的生成必须将披露控制深度融入生成过程本身，以实现既安全又高效的LLM应用部署。
链接: https://arxiv.org/abs/2610.28067
作者: Devansh Srivastav,David Pape,Samira Abedini,Lea Schönherr
机构: CISPA Helmholtz Center for Information Security(德国萨尔布吕肯信息安全亥姆霍兹中心); https://github.com/Dormant-Neurons/graduated_disclosure
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL)
备注:
Abstract:LLMs increasingly operate over mixed-sensitivity contexts, where they must reason over information ranging from public to confidential while disclosing only what the user is authorised to receive. Existing approaches enforce access control as a binary decision: either fully answering or refusing, without regulating how much or what kind of information is revealed if access is granted, leading to over-refusal or unintended leakage. We introduce graduated disclosure, an access-control formulation in which the level of detail and sensitivity in a model’s response is calibrated to the user’s authorisation rather than treated as a grant-or-deny decision. We model graduated disclosure as a disclosure-aware generation problem and propose a reward-based optimisation framework that shapes responses toward level-appropriate abstractions while penalising unauthorised factual leakage, including paraphrases and approximations. We evaluate our approach on enterprise and clinical benchmarks and compare it with prompt-based, gating, and supervised baselines. Experiments show that our method substantially reduces information leakage while preserving correctness and utility. These results suggest that disclosure control must be integrated into the generation process itself to enable secure yet useful LLM deployments.
[NLP-32] Same Judgment Different Yes: Measuring Multilingual Judgment Profiles in Language Models
【速读】： 该论文旨在解决多语言评估中模型回答一致性与判断偏差之间的矛盾问题，特别是当采用不同提问形式、答案约束及回复读取方式（如仅读取“是/否”标记或完整句子）时，如何确保所测得的道德判断具有可比性。其核心解决方案在于揭示“是/否”标记（yes/no token）在特定语境下可能反映的是语言习惯或模型输出模式而非真实判断，而非普遍适用的推理结果。研究通过构建FAR-6评测集，在六种语言中收集八款模型对105个道德情景的110万条回复，发现尽管故事层面的判断轮廓在不同读取方式和问题形式间总体收敛，但在负向疑问句后使用预设映射读取“是/否”标记时出现显著偏差：该标记的判断与句子级判断的相关性仅为中等（中位r=0.67），远低于句子与自由表述间的高一致性（r=0.94）。进一步分析表明，模型在跨语言情境下的判断轮廓与其自身在其他语言中的表现更为接近（39/48个模型-语言组合在命题标签下显著高于随机预期），说明模型具备相对稳定的内在判断逻辑，但“是/否”标记的使用受制于语言结构、词序及特定术语偏好等语言特征。因此，关键在于识别并区分模型的真实判断倾向与由语言表达习惯驱动的“回答习惯”——尤其是“是/否”标记的语义依赖性，强调必须通过上下文语义解读来理解该标记所支持的命题内容，而非直接将其作为判断依据。
链接: https://arxiv.org/abs/2610.28066
作者: Haonan Huang,Yuwei Lu,Chengkun Cai
机构: Princeton University (普林斯顿大学); Independent Researcher (独立研究员); University College London (伦敦大学学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Multilingual evaluations often score a model’s opening yes or no, a token that also carries model and language answering conventions. We ask when different ways of eliciting and reading an answer recover the same judgment, and when they measure an answer habit. We collect 1.1 million replies from eight models in six languages on up to 105 moral vignettes, varying the question’s form, the answer constraint, and the reading of the reply, token or sentence, in both wordings (the FAR-6 battery). Story-level judgment profiles converge across the tested readouts and question forms, except for the yes/no token after a negative interrogative read under a preset mapping: on the same replies, token and sentence profiles agree at a median r = 0.67 across cells, while sentence profiles agree with free-sentence profiles at 0.94, a known linguistic variability. With each language’s story consensus removed, a model’s profile is nearest to its own in another language for 39 of 48 model-language cells under proposition labels and 15 under free sentences, against 5 expected by chance. Particle, position, and named-term preferences follow their own patterns. A model can keep its judgment and change its yes: scoring that token requires establishing which proposition it affirms.
[NLP-33] On-Policy Transition Distillation
【速读】： 该论文旨在解决生成式模型中知识蒸馏效率不足的问题，特别是在强化学习框架下，如何更有效地将专家模型（teacher）的复杂推理能力迁移至学生模型（student）的问题。现有方法如基于策略的蒸馏（On-policy Distillation, OPD）仅通过匹配师生模型在最终层的输出分布来实现知识传递，但忽略了中间层预测动态所蕴含的任务相关监督信号。本文的关键创新在于提出基于策略的过渡蒸馏（On-Policy Transition Distillation, OPTD），其核心是引入对中间层预测变化（depthwise prediction transitions）的监督——即在学生生成的轨迹上，对教师与学生之间各层间对数概率增量（log-probability increments）进行对齐。实证研究表明，中间层预测过渡的一致性与答案正确率显著正相关，且OPTD在多个基准测试（涵盖数学、编程和指令遵循任务）中均取得最优或并列最优性能，尤其在三领域联合蒸馏场景下，可恢复约90%的师生能力差距，显著优于传统OPD的75%。因此，解决方案之关键在于：利用中间层预测过渡的细粒度监督，增强学生模型在推理过程中的路径一致性，从而提升知识迁移的有效性与泛化能力。
链接: https://arxiv.org/abs/2610.28048
作者: Runyang You,Yongqi Li,Tiezheng Yu,Wenjie Li
机构: The Hong Kong Polytechnic University (香港理工大学); Huawei Technologies Ltd. (华为技术有限公司)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:On-policy distillation (OPD) consolidates expert capabilities into a single student by matching teacher and student final-layer distributions on student-generated rollouts. We investigate whether depthwise prediction transitions, the changes in predictions between layers, can provide additional task-relevant supervision. Our diagnostic study finds that stronger teacher–student agreement in intermediate transitions is associated with higher answer correctness within questions, with correct responses exhibiting lower mean transition KL across all measured intermediate layer pairs. These observations motivate On-Policy Transition Distillation (OPTD), which augments final-layer distribution matching by aligning teacher and student layer-to-layer log-probability increments on student-generated rollouts. Evaluations across three backbones and eight benchmarks spanning mathematics, coding, and instruction following show that OPTD achieves the best or tied-best overall performance in both dual- and three-domain distillation. In the three-domain setting, OPTD recovers \approx 90% of the overall teacher–student capability gap, compared with \approx 75% for OPD.
[NLP-34] Poisoning-Based Backdoors Are Not Cryptographic NEURIPS2026
【速读】： 该论文旨在解决神经网络在面对后门攻击时的防御难题，核心问题在于：传统防御方法仅针对个别触发器（trigger）进行隔离式检测，而攻击者可快速生成满足相同验证函数（verifier）的新触发器，导致防御失效。为此，论文提出一种理想化的防御构想——若能以类似密码学认证机制的非对称性部署后门验证函数，使得只有攻击者可生成有效触发输入，而防御者无法从观测中逆向推导出该验证函数，则基于学习的防御将从根本上不可行。然而，论文通过归约论证证明：在基于污染数据（poisoning-based）的验证函数部署场景下，这种非对称性无法实现。原因在于，任何能够利用污染数据安装验证函数并使后门在未见触发器上生效同时保持良性行为的训练过程，本身即构成对该验证函数的学习算法；因此，若某验证函数在计算上难以学习，则其无法通过污染数据安装；反之，若可通过污染数据安装，则白盒防御者仅需同等数量的标注触发样本即可重构该验证函数。此外，论文还进一步证明：在推理阶段提供给模型的秘密密钥（secret key）同样可被此类防御者观测到。研究结果表明，该结论不依赖于后门具体执行动作，而聚焦于后门是否触发。在指令微调的大语言模型上的实验验证了这一二元对立关系，所有配置均未能实现对攻击者有效且对防御者不可恢复的后门。
链接: https://arxiv.org/abs/2610.28045
作者: Masahiro Kaneko
机构: MBZUAI
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: NeurIPS 2026
Abstract:To defend against backdoor attacks on neural networks, the defender must identify the verifier function that determines whether an input is a trigger, rather than rejecting individual triggers in isolation, since an attacker can generate new triggers satisfying the same verifier faster than they can be blocked. If a backdoor verifier could be installed with the asymmetry of a cryptographic authentication scheme, in which only the attacker can produce trigger inputs while the defender cannot reverse-engineer the verifier from observations, learning-based defense would be impossible in principle. We show that no such asymmetry can arise from poisoning-based installation, in which the only verifier-dependent input to training is a poisoned dataset. The argument is a reduction. A training procedure that installs a verifier from poisoned data, so that the backdoor fires on unseen triggers while benign behavior is preserved, is itself a learner for that verifier. Any verifier that is computationally hard to learn is therefore impossible to install through poisoning, and any verifier that poisoning can install is reconstructible by a white-box defender holding comparably many labeled trigger observations. A complementary argument shows that a secret key supplied to the model at inference is observable to the same defender. The result takes representative trigger observations as given and concerns when the backdoor fires rather than what it does. Experiments on instruction-tuned large language models confirm the dichotomy, with no configuration yielding a backdoor that is both effective for the attacker and irrecoverable for the defender.
[NLP-35] Using Letter Positional Probabilities to Assess French Word Complexity
【速读】： 该论文旨在探究法语中简单词与复杂词在拼写特征上的差异，核心问题是识别哪些拼写特征能够有效区分词汇复杂度。其解决方案的关键在于引入字母位置概率（Letter Positional Probabilities, LPPs）这一新方法，通过分析高反应时（复杂词）与低反应时（简单词）词汇在每个字母位置上的二值化表示（基于42个标准法语字符的one-hot编码，最大长度为25位），构建每种字母在特定位置出现的平均概率分布。研究发现，大量LPPs与词汇复杂度显著相关，例如带重音符号的字母更常出现在简单词中。进一步地，分类器实验表明，仅使用词长即可实现77.27%的分类准确率，而加入前6个字母位置的二值变量后准确率提升至83.95%，说明拼写结构中的局部字母分布模式对复杂度编码具有重要贡献。该研究揭示了词汇复杂度可能通过特定字母在拼写序列中的分布规律被隐含编码于正字法之中。
链接: https://arxiv.org/abs/2610.28040
作者: Michael Dalvean
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:We use Letter Positional Probabilities (LPPs) to examine the orthographic differences between simple and complex French words. We draw from a sample of 38,000 French words scored by reaction time on a lexical decision task. We take the 5000 highest reaction time words and the 5000 lowest as a proxy for high and low word complexity groups respectively. We create 18 variables based on the structure of the words, such as length and number of syllables. We also reduce each word to a string of binary values by one-hot encoding each letter according to the 42 characters of the standard French alphabet up to a length of 25 letters. Thus, each word is represented by 42 x 25 binary characters. We then create LPPs by looking at the average binary scores for simple as opposed to complex words for all letter/positions. We find that there is a high number of LPPs that are associated with word complexity. For example, we find that most accented letters are associated with simpler words. We examine the effects of structural and LPP variables in classifiers and find that word length alone enables a classification accuracy of 77.27%, while adding the binary variables from the first 6 letter positions provided an accuracy of 83.95%. We discuss possible explanations as to how complexity could be encoded in orthography.
[NLP-36] From RL Papers to Framework Code: Inclusion and Review in Seven Open-Source Training Frameworks
【速读】： 该论文旨在解决生成式 AI（Generative AI）领域中强化学习（Reinforcement Learning, RL）训练框架与学术研究之间存在显著脱节的问题，具体表现为大量提出通用强化学习训练方法的论文未能被主流开源框架采纳。其核心问题在于：尽管学术界持续产出大量新方法，但这些方法在实际应用中的落地率极低，且缺乏有效的反馈机制来推动框架迭代。解决方案的关键在于通过系统性追踪七款主流开源RL后训练框架的提交历史，并将其与2025年1月至2026年6月期间筛选出的6,939篇arXiv论文进行关联分析，揭示了框架采纳率的长期低迷现象——在五个季度周期内，即使论文数量从72篇增长至386篇，被纳入框架的方法数量仍稳定在5至7篇之间。进一步分析表明，被采纳的方法更倾向于使用参数量超过300亿的大模型（41% vs. 10%），并更频繁地对现有损失函数中的系数、裁剪值、掩码或权重等超参数进行调整（29% vs. 9%）。此外，研究发现框架的公开问题与拉取请求中极少拒绝新方法，但绝大多数未被采纳的方法在社区中无任何引用或提及记录，且许多提案长期处于无人回应状态。结合技术报告数据可见，用户极少在公共渠道报告训练失败，而其中仅有8/13广泛采用的方法出现在相关报告中。为此，作者发布了可追踪训练失败与框架实现对应关系的数据集，并公开各框架的贡献规则及早期提案清单，以促进学术界与工业界在强化学习训练方法上的有效协同。
链接: https://arxiv.org/abs/2610.28027
作者: Taiheng Pan
机构: The University of Melbourne(墨尔本大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Software Engineering (cs.SE)
备注: 29 pages (13 pages main text), 5 figures, 12 tables. Data and code at this https URL and this https URL
Abstract:Practitioners use reinforcement learning methods for language models through training frameworks, so a method that a framework includes is one configuration change away for its users. We measure how often this happens. We trace seven open-source frameworks for RL post-training through their commit histories and link them to 6,939 screened arXiv papers from January 2025 to June 2026. Of 1,598 papers that propose a general RL training method, 37 reached at least one framework. Across five quarterly cohorts of papers, the number included within 180 days stayed between 5 and 7 while the cohorts grew from 72 to 386 papers. Compared with papers from the same quarters, included papers more often trained policies of 30B parameters or more, at 41 against 10 percent, and more often changed a coefficient, clip, mask or weight in the existing loss, at 29 against 9 percent. Both gaps hold without papers from the companies that maintain the frameworks, while code release was equally common in both groups. The frameworks’ issues and pull requests show that few methods are turned away in public. More than nine in ten of the general methods no framework included have no thread there that cites or names them, and many proposals that do appear stay open without a reply. Two further records, the frameworks’ issues and the technical reports of 13 large open-weight models, show that users rarely report training failures in public and that the reports state 8 of the 13 widely included methods. We release the data as a tracker that links training failures to framework implementations and lists each framework’s contribution rules and earlier proposals.
[NLP-37] Beyond the Evidence Boundary: Behavioral Mirag e Under Progressive Degradation in Medical VLMs NEURIPS2026
【速读】： 该论文旨在解决生成式 AI (Generative AI) 在医学影像解读中面对渐进性证据损失时的可靠性评估问题。在临床高风险场景下，医学影像常因质量下降导致信息衰减，而现有模型评估多基于干净输入，忽视了模型在证据逐步退化情境下的行为变化。研究发现，尽管模型性能可能下降，其置信度仍可能维持高位，但缺乏对“模型-案例”组合何时失去可靠性的明确界定标准。为此，作者提出一种基于基线校准（baseline-anchored）的分析框架，通过追踪图像逐步退化过程中的基线校正案例，识别出一个经验性的证据边界，并将在此边界上仍持续输出决策（无论正确与否）的现象定义为“幻觉承诺”（mirage）。在三个医学数据集上的实验表明，仅有 MedGemma 模型具备足够的基线校正样本以进行轨迹分析；在 VQA-RAD 数据集中，仅语言因素即可预测基线正确性与边界处承诺行为（AUROC 分别为 0.677 和 0.653–0.670）。此外，MedGemma 在边界后的表现呈现高度异质性，包括直接幻觉、正确恢复、幻觉消解、一致弃权及振荡等模式，幻觉率在不同数据集-模态组合间高达 16.0% 至 100.0%。然而，残差流探针在嵌套交叉验证下仍能以 AUROC 0.895–0.994 的高精度解码边界处的“承诺”与“弃权”行为。研究结果表明，渐进性证据损失可引发持续承诺行为，而输出熵与弃权策略无法稳定反映此类现象，因而亟需引入边界感知评估、选项顺序控制以及更优的弃权机制，以提升医学基础模型在部分观测条件下的可靠性。
链接: https://arxiv.org/abs/2610.28000
作者: Edric Castel C. Hao,Rez Samantha Z. Floresca
机构: Analog Devices, Inc.(模拟设备公司); Ateneo de Manila Senior High School(马尼拉大学高级中学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: Accepted to WMHS Workshop @ NeurIPS 2026
Abstract:VLMs for medical image interpretation are typically evaluated on clean inputs, despite the progressive evidence loss that can occur in high stakes clinical settings. Existing robustness studies show that performance can degrade while confidence remains high, but do not establish when a previously reliable model-case pair ceases to behave reliably. We introduce a baseline-anchored framework that traces baseline-correct cases through progressive image degradation to identify an empirical evidence boundary, and define mirage as continued commitment at this boundary, regardless of correctness. Across three VLMs on three medical datasets, only MedGemma provides a sufficiently large baseline-correct pool for trajectory analysis. On VQA-RAD, the only dataset in which the question text varies, language-only answerability is associated with baseline correctness (AUROC 0.677) and with commitment at the boundary (AUROC 0.653 to 0.670). MedGemma’s post-boundary behavior is heterogeneous, including direct mirage, correct recovery, mirage dissolution, consistent abstention, and oscillation, with mirage rates of 16.0 to 100.0% across dataset-modality pairs. Despite this behavioral heterogeneity, residual-stream probes decode commit-versus-abstain behavior at the boundary with AUROC of 0.895 to 0.994 under nested cross-validation. These results show that progressive evidence loss can produce persistent commitment that output entropy and abstention behavior do not consistently reflect, motivating boundary-aware evaluation, option-order controls, and improved abstention mechanisms for reliable medical foundation models operating under partial observation.
[NLP-38] Building Telecom Language Models with RSOD: Domain Data Specialist Post-Training Memory-Conscious Distillation and Networking Evaluation
【速读】： 该论文旨在解决电信领域语言模型（Telecom Language Models）在保持广泛网络知识覆盖的同时，如何有效融合路由、宽带服务、接入技术及运维诊断等专业行为，并实现可部署的单模型架构这一关键挑战。其核心解决方案在于提出一种端到端的路由专家在线蒸馏（Routed Specialist On-Policy Distillation, RSOD）框架，通过结构化网络分类体系与专家图谱生成源基合成训练数据，构建跨14条网络路径的212,559条路由训练记录；在此基础上，利用共享的持续预训练骨干网络冻结训练六个专家级LoRA教师模型，再通过RSOD将各路由专属教师的行为压缩至一个可部署的学生适配器中。该方法的关键创新在于：学生模型在每条轨迹生成过程中，由路由指定的教师强制输出原始本机分词序列，仅更新学生LoRA参数，同时通过共享冻结输出头重构紧凑教师隐藏状态，确保知识迁移的精确性与可审计性。实验验证显示，该方案在微调后实现了与多教师系统近乎一致的性能（微平均得分差值<0.0529），且在四块RTX 6000 Ada GPU上完成2,000步训练后未出现任何版本或标记ID不一致问题，最终释放的Llama-CalixCLM v1模型在主路由评估下达到21.7112/25微平均得分，证明了其在保持高性能的同时具备高度可部署性与可验证性的优势。
链接: https://arxiv.org/abs/2610.27993
作者: Anil Srirangapatna Nagesh,Nitin Gupta,Anuj Taneja,Vivek Singh,Shane Eleniak,Theo Zhao,Moulidharan N,Shashank Keshetty
机构: Calix AI Labs
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Telecom language models must combine broad networking knowledge with specialist behavior across routing, broadband services, access technologies, and operational diagnostics while remaining practical to deploy as a single model. We present an end-to-end framework that builds source-grounded synthetic training data from structured networking taxonomies and expert graphs, materializes 212,559 routed training records across fourteen networking routes, trains six frozen specialist LoRA teachers over one shared continued-pretrained backbone, and consolidates their behavior into one deployable student adapter using Routed Specialist On-Policy Distillation (RSOD). In RSOD, the committed student generates each trajectory, the route-assigned specialist teacher-forces the exact native token sequence, compact teacher hidden states are reconstructed through the shared frozen output head, and only the student LoRA is updated. A controlled reconstruction validation shows mean absolute logit error of 7.26 x 10^-6 with 100% top-1 agreement between native and reconstructed teacher logits. On four RTX 6000 Ada GPUs, the completed 2,000-step run produced 16,000 trajectories with zero observed student-version or token- ID mismatches. Under the primary route-conditioned evaluation over 2,850 held-out questions, the raw Meta-Llama-3.1-8B-Instruct model scores 12.8661/25 micro (13.0172/25 macro), the combined routed six-teacher system scores 21.7641/25 micro, and the consolidated RSOD student, released as Llama-CalixCLM v1, reaches 21.7112/25 micro (21.8006/25 macro) on the same five dimension LLM-judge rubric. Under a separate neutral deployment-like prompt, the same student reaches 20.0028/25 micro (20.0778/25 macro). The consolidated student remains within 0.0529 micro-score points of the combined routed-teacher system while retaining a numerically validated and auditable train-serve execution path.
[NLP-39] Adaptive Stopping of Streamed Reasoning for Language Model Agents
【速读】： 该论文旨在解决生成式 AI（Generative AI）代理在决策过程中因推理长度不一而产生的计算开销过大的问题，即如何在保证任务成功率的前提下，动态优化推理过程的终止时机。现有方法依赖于置信度或中间答案稳定性等停止信号，但其有效性受模型差异影响显著，且依赖隐藏状态或词元概率等内部信息，存在访问受限或实现成本高的问题。本文提出 StreamStop，其核心解决方案是通过离线训练模型特定的停止策略，基于前缀级别动作一致性（prefix-level action agreement）构建停止机制。该方法采用一个冻结的前缀读取器（frozen prefix reader）编码推理文本，并结合轻量级停止头（stop head）进行判断，无需访问推理模型内部信号即可实现高效停止。研究团队构建并公开了一个包含前缀级动作标注的代理推理轨迹数据集，使得停止规则可进行离线比较，并支持对折扣动作一致性目标（discounted action-agreement objective）的精确期望回报优化。通过将停止头条件化于折扣因子，可在推理阶段灵活调整动作一致性与词元消耗之间的权衡，无需重新训练。在 WebShop 与 ScienceWorld 的闭环评估中，StreamStop 在使用更少推理词元的情况下，取得了优于多数固定预算设置的可观测成功率；以 Qwen3.8-27B 模型在 WebShop 上为例，每轮任务减少 42% 的推理词元消耗，同时保持相近的任务成功率。
链接: https://arxiv.org/abs/2610.27981
作者: JunHyeok Oh,Byung-Jun Lee
机构: Korea University(韩国大学)
类目: Computation and Language (cs.CL)
备注: 15 pages, 6 figures. Project page: this https URL
Abstract:Reasoning-enabled language model agents can incur substantial computation before acting, although the reasoning needed varies across decisions. Existing stopping signals, such as confidence and intermediate-answer stability, can vary in effectiveness across reasoning models; signals obtained from hidden states or token probabilities also require internal access that may be unavailable or impractical. We introduce StreamStop, which learns model-specific stopping policies offline from prefix-level action agreement. A frozen prefix reader encodes available reasoning text for a lightweight stop head, allowing stopping without access to the reasoning model’s internal signals. We construct and publicly release a dataset of agent reasoning trajectories with prefix-level action annotations. These annotations compare each prefix action with the full-reasoning reference action, enabling offline comparison of stopping rules and exact expected-return optimization of a discounted action-agreement objective. Conditioning the head on the discount factor allows the agreement-token trade-off to be adjusted at inference without retraining. In closed-loop evaluations on WebShop and ScienceWorld, StreamStop achieves higher observed success than most validation-selected fixed-budget settings while using fewer reasoning tokens than Full thinking. On WebShop with Qwen3.8-27B, it uses 42% fewer reasoning tokens per episode with similar observed task success.
[NLP-40] Lost in Recovery: Demystifying Recovery Behavior of Computer-Use Agents via Goal-Oriented Action Replay
【速读】： 该论文旨在解决当前计算机使用代理（computer-use agents）在长时程任务中因错误累积和操作副作用导致的脆弱性问题，尤其关注如何实现高效、可靠的故障恢复机制。现有方法如环境重置在专有场景下不可行，动作重放成本过高，而基于API的回滚又仅限于网页环境，导致基于模型的恢复策略长期未被充分探索。为此，论文提出ReTrace——一个可扩展的评估框架，用于系统性研究代理的恢复行为。其解决方案的关键在于提出并形式化“转移一致性恢复”（Transition-Consistent Recovery）准则：从恢复状态与目标状态出发，执行相同的下游操作应产生完全一致的结果。通过在受控误差注入后进行目标导向的动作重放，该框架实现了对恢复能力的量化评估。实验揭示了三大核心发现：(i) 理想的恢复机制能够激活模型潜在的任务完成能力；(ii) 当前计算机使用代理未能有效利用其内在的恢复潜力；(iii) 准确的基于模型的恢复可带来可观的端到端性能提升。这些结果确立了恢复能力作为构建可靠计算机使用代理的基础性要素。
链接: https://arxiv.org/abs/2610.27954
作者: Byeonghun Bang,Sangwoo Park,Jinheon Baek,Kangsan Kim,Youngwan Lee,Sung Ju Hwang
机构: KAIST(韩国科学技术院); ETRI(电子与电信研究院); DeepAuto.ai
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Recent advances in large (vision) language models have enabled computer-use agents that automate digital workflows via fine-grained keyboard and mouse operations. However, such agents remain fragile in long-horizon settings, where errors accumulate and executed actions leave unintended side effects, making the ability to recover from failures critical. Yet, existing remedies remain inadequate: environment resets are infeasible in proprietary settings, action replay is prohibitively expensive over long horizons, and API-based rollback is limited to web environments, leaving model-based recovery largely unexplored. To address this gap, we present ReTrace, a scalable evaluation framework for systematically studying recovery behavior in computer-use agents. We formalize Transition-Consistent Recovery: identical downstream actions from the recovered and target states must yield identical outcomes. We operationalize this criterion through goal-oriented action replay following controlled error injection. Using ReTrace, we present the first systematic study of autonomous, model-based recovery, revealing that (i) ideal recovery can unlock model’s latent task completion capabilities, (ii) current computer-use agents fail to exploit their latent recovery capabilities, and (iii) accurate model-based recovery yields measurable end-to-end gains. Together, these findings establish recovery as a fundamental capability for building reliable computer-use agents. Our code and dataset is available at this https URL.
[NLP-41] From DQI to C-DQI: Automatic Codebook Generation and Evaluation for Measuring Computational Discourse Quality Index
【速读】： 该论文旨在解决传统议会话语质量评估中依赖专家评分、理论驱动的编码手册（Codebook, CB）所导致的僵化性、高人力成本及潜在偏见问题。其核心解决方案在于提出一套完整的基于大语言模型（Large Language Models, LLMs）的编码手册生成与评估协议，通过对比自上而下（理论驱动）、自下而上（数据驱动）及混合方法，构建适应具体语境的分析框架。关键创新在于引入两个互补的评估指标：一是基于人工标注验证的外在代理指标（extrinsic proxy measure），用于衡量编码手册的内容有效性；二是内在模型间一致性（Intrinsic Inter-Model Agreement, IMA）指标，利用不同LLM固有的偏差差异作为信号，反映整体方案的可靠性。实验基于100份以色列议会记录表明，数据驱动的归纳式编码手册在保持有效性的同时，显著优于传统自上而下的方法，在模型间一致性方面表现更优，展现出在大规模政治话语分析中实现可靠、可扩展评估的潜力。
链接: https://arxiv.org/abs/2610.27937
作者: Esther Shizgal,Ron Keinan,Daniel Shtauber,Roee Tzvi Esquira,Ori Aronson,Matan Toledano,Or Rappel-Kroyzer,Liron Lavi,Renana Keydar,Ittai Bar-Siman-Tov,Reut Tsarfaty
机构: Max Planck Center for Democracy, Security and Human Rights, Faculty of Law, Hebrew University (马克斯·普朗克民主、安全与人权中心，希伯来大学法学院); Department of Computer Science, Bar-Ilan University (计算机科学系，巴伊兰大学); Digital Humanities, Hebrew University (数字人文，希伯来大学); Department of Political Studies, Bar-Ilan University (政治学系，巴伊兰大学); Faculty of Law, Bar-Ilan University (法学院，巴伊兰大学); Faculty of Law, Hebrew University (法学院，希伯来大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Effective democratic deliberation is vital for civic engagement and the health of legislative institutions. However, assessing parliamentary discourse quality is traditionally done using expert-scored, theory-based codebooks (CBs), which are rigid, labor-intensive, and prone to biases. While Large Language Models (LLMs) offer a scalable alternative for both schema induction and discourse annotation, ensuring schema quality and mitigating models’ inherent bias remains a clear challenge. In this work, we propose a complete protocol for LLM-based codebook generation and evaluation. Starting from the widely accepted discourse quality Index (DQI), we compare top-down (theory-based), bottom-up (data-driven), and hybrid approaches for generating context-adapted schemas. To evaluate these schemas, we define an extrinsic proxy measure, validated by manual annotations, alongside an intrinsic inter-model agreement (IMA) metric, which leverages differing model biases to signal overall reliability. Our experiments on 100 Israeli parliament transcripts reveal that inductive, data-driven schemas achieve higher IMA than top-down methods while maintaining validity, demonstrating their potential for reliable, large-scale analysis of political discourse.
[NLP-42] Why Do LLM s Accept Generics? Explaining Acceptability using Psycholinguistic and Model-Derived Factors
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在判断通用陈述（generic statements）时的推理机制问题，具体探究人类心理语言学中影响通用性判断的多重认知因素（如属性普遍性、类别区分度及特征性）是否同样适用于模型。其核心解决方案的关键在于：通过在心理语言学数据集和自然语料中的实证分析，发现LLMs在评估通用陈述时会权衡属性的普遍性与类别区分能力，其决策模式与人类相似但表现更为模糊；尤其值得注意的是，尽管模型也偏好具有特征性的属性，却对“显著性”（strikingness）等人类判断中重要的因素不敏感，而更依赖于事实真实性这一根本判断标准，且人类相关因素的影响主要通过事实真值判断间接传导至最终评分。进一步实验表明，在6个7-9B规模的模型中，沿各认知因子方向调整隐藏状态可有效预测并引导模型判断，揭示了这些因素在模型内部表征中的可分离性与作用路径。
链接: https://arxiv.org/abs/2610.27928
作者: Mervat Abassy,Zhaoxin Feng,Tatsuki Kuribayashi,Kentaro Inui,Emmanuele Chersoni
机构: MBZUAI; The Hong Kong Polytechnic University (香港理工大学); Tohoku University (东北大学); RIKEN (理化学研究所)
类目: Computation and Language (cs.CL)
备注:
Abstract:Psycholinguistic studies explore why humans accept some generic statements but reject others by showing that these judgments depend on more than how common a property is. People consider several other factors such as how well it sets the kind apart from others or whether it is characteristic of the kind. Inspired by the linguistic and cognitive literature, we explore to what extent factors from these studies explain how large language models (LLMs) judge generics. Our results with twelve models on a psycholinguistic dataset and on naturally occurring generics show that models weigh how common a property is against how well it distinguishes the kind, and combine the two much as people do, though less clearly. Like people, they favor characteristic properties, but unlike the accounts of human judgment, they ignore striking ones. What matters most to LLMs is whether a statement is factually true, and much of the influence of human factors reaches their ratings through that judgment. In six 7-9B models, shifting the hidden states along each factor’s direction moves the judgment as the factor predicts. Subjects: Computation and Language (cs.CL) Cite as: arXiv:2610.27928 [cs.CL] (or arXiv:2610.27928v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2610.27928 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-43] ACE-Compression: An Ability-Centric Evaluation of Compressed Large Language Models
【速读】： 该论文旨在解决当前大语言模型（Large Language Models, LLMs）压缩评估体系不完整的问题。现有标准评估仅关注任务性能保留情况，忽略了压缩过程对模型中分布式知识（distributed knowledge）的扰动，导致无法揭示哪些核心能力得以保留、哪些能力悄然退化。为此，论文提出一种以能力为中心的评估框架——ACE-Compression，其关键在于从部署实际需求出发，系统性地考察四类关键能力维度：程序性推理与约束执行、知识锚定与真实性、社会认知与情感理解、人格一致性与风格连贯性。通过在六种开源大模型的35个以上压缩检查点上进行实验，研究发现压缩会重塑模型的能力分布，粗粒度、局部性和表面特征的能力通常具有较强鲁棒性，而细粒度、长尾及依赖上下文的能力则尤为脆弱。量化（quantization）在能力保留方面普遍优于剪枝（pruning），但其优势的实际意义取决于具体应用所需的能力类型。因此，该框架使模型部署者可根据目标应用的能力需求选择压缩策略，而非仅依赖基准平均得分；同时为算法设计者提供了可诊断的信号，以开发更注重关键能力保留的压缩方法。总体而言，ACE-Compression将原本黑箱化的压缩过程转化为可测量、可解释、可优化的干预手段，推动模型压缩从“精度下降”转向“能力可调控”的范式演进。
链接: https://arxiv.org/abs/2610.27913
作者: Jiahui Ning,Liang Chen,Yuchen Liang,Yi Zhang,Fenglei Fan
机构: City University of Hong Kong(香港城市大学); Tongji University(同济大学); Rensselaer Polytechnic Institute(伦斯勒理工学院); Sichuan University(四川大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 33 pages, 7 figures, including references and appendices. Code: this https URL
Abstract:Model compression makes large language models deployable, but today’s evaluation tells an incomplete story. Standard practice measures retained task performance and stops there. This misses a critical fact: compression perturbs the distributed knowledge embedded across billions of weights, and task scores alone cannot reveal which underlying abilities survive and which quietly collapse. To solve this question, we introduce ACE-Compression, an ability-centric evaluation framework. Instead of asking only “does the compressed model still answer correctly?”, we ask “which of its core abilities remain intact?” Our framework probes four deployment-facing dimensions: procedural reasoning and constraint execution, knowledge grounding and truthfulness, social cognition and affective understanding, and persona fidelity and stylistic consistency. Across more than 35 compressed checkpoints from six open-weight LLMs-covering quantization and pruning-we find that compression redistributes ability profiles. Coarse, local, and surface-level signals often persist, while fine-grained, long-tail, and context-dependent abilities become fragile. Quantization consistently outperforms pruning in retention, but the practical meaning of that gap depends entirely on which ability an application requires. Based on our findings, deployers can choose compression methods by the ability profile their application needs, not just by a leaderboard average. And algorithm designers gain diagnostic signals to build compression methods that preserve what matters. We believe that ACE-Compression turns compression from a black-box accuracy drop into an interpretable intervention-one we can measure, understand, and ultimately improve.
[NLP-44] Polarity Blindness in LLM Safety: A Systematic Study of Negation Antonym Substitution and Adversarial Framing
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在自然语言处理任务中对否定语义识别能力不足的问题，尤其是由极性反转（polarity inversion）引发的潜在安全风险。其核心挑战在于“极性盲视”（polarity blindness）——即模型在经历语法否定或反义词替换等极性改变的文本变换后，无法正确识别其中隐含的有害意图，从而导致安全防护机制失效。解决方案的关键在于揭示这一现象并非局限于特定语言或语法规则，而是普遍存在于多种语言类型和主流大模型中的语义鲁棒性缺陷。研究通过设计包含260个提示的跨语言基准测试（涵盖英语、西班牙语和捷克语），系统评估了语法否定、反义词替换与同义词替换三种操作对模型输出危害性的影响，并引入对抗性文本包装器以增强攻击效果。结果表明，多数模型在极性反转下均表现出显著的危害评分上升，最脆弱模型的平均危害得分高达4.12（5分制），且该现象在不同语言结构中持续存在，说明极性盲视反映了模型深层的语义理解局限，而非局部的语言特异性问题。
链接: https://arxiv.org/abs/2610.27908
作者: Diego Miguel Lozano,Erik Derner,Juan Antonio Pérez-Ortiz,Nuria Oliver
机构: ELLIS Alicante(ELLIS阿尔坎特), Czech Technical University in Prague(捷克技术大学), Universidad de Alicante(阿尔坎特大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Code and data: this https URL
Abstract:Negation remains a persistent challenge for language models across NLP tasks and modalities. We show that this weakness can undermine LLM safety guardrails through a structural vulnerability we term polarity blindness: the failure to recognize harmful intent after polarity-altering linguistic transformations. We study two forms of polarity inversion: grammatical negation and antonym replacement, and use synonym replacement as a polarity-preserving control. We further test whether an adversarial textual wrapper amplifies these failures. Using a 260-prompt benchmark across 10 state-of-the-art LLMs and 3 languages (English, Spanish, Czech), stratified by harmfulness potential and evaluated on a validated 0-5 harm scale, we find that polarity blindness affects most evaluated models, with mean harm scores reaching 4.12 for the most vulnerable model. The effect persists across typologically differing negation systems, suggesting that polarity blindness reflects a broader limitation in semantic robustness rather than a language- or grammar-specific failure.
[NLP-45] Staged Continual Adaptation of Multimodal Foundation Models for Japanese Financial Documents ICML2026
【速读】： 该论文旨在解决多模态基础模型在分阶段后训练过程中，各训练阶段之间能力权衡与相互影响不明确的问题。具体而言，研究关注视觉-语言对齐、推理能力提炼及领域适配三个阶段如何协同或冲突，进而影响整体性能。其解决方案的关键在于通过系统性地追踪一个84亿参数模型在日文金融文档上的三阶段训练过程（从初始基线到逐步引入数学推理提炼和领域微调），揭示了各阶段并非线性叠加：数学专项提炼虽增强程序化解题能力，但会覆盖前期积累的金融领域知识；而领域适配阶段主要修复而非新增知识。进一步的跳过阶段消融实验表明，跳过推理提炼阶段可最优保留领域知识，但损失数值推理能力，说明各阶段贡献不可分解，最佳检查点高度依赖目标任务，因此应优先考虑检查点选择与训练调度策略，而非简单增加训练阶段。
链接: https://arxiv.org/abs/2610.27904
作者: Genshin Kakimoto,Atsushi Yanagisawa
机构: Kyoto University (京都大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 17 pages, 10 figures, 9 tables. Presented at the ICML 2026 Workshop “Continual Adaptation at Scale: Towards Sustainable AI”. Code: this https URL
Abstract:Staged post-training of multimodal foundation models chains vision-language alignment, reasoning distillation, and domain tuning, yet how each phase trades capabilities against the others remains under-characterised. Tracking an 8.4B-parameter model from an un-fine-tuned baseline through three training phases on Japanese financial documents (evaluated on five text-input benchmarks), we find the phases interact non-additively: math-only distillation overwrites the financial knowledge acquired earlier even as procedural math skill survives, and the domain phase repairs more than it installs. A skip-phase ablation is best for domain knowledge but forfeits numerical-reasoning gains, so per-phase contributions are not decomposable and the best checkpoint is task-dependent, with implications for checkpoint retention and scheduling rather than added stages.
[NLP-46] CollabSciReM: Aligning Language Models for Scientific Conversations via Science-Aware Reward Decomposition
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在科研协作场景中难以同时兼顾准确性、方法论严谨性、逻辑一致性与生成效率的问题。现有对齐方法往往在追求对话质量时导致冗余表达，而通用型协作奖励机制则因忽视科学内容而牺牲科研严谨性。其解决方案的关键在于提出一种目标导向的奖励框架——CollabSciReM，该框架将会话级奖励解耦为对话质量、科学严谨性与结构效率三个维度，并引入基于300篇未见科学论文构建的多轮科研对话评估基准SciQuality，从方法论合理性、论证充分性、逻辑一致性、事实正确性、创新性及响应性六个维度进行精细化评估。通过在8.62K条合成科研对话上微调，CollabSciReM在多个主流模型（Llama与Qwen系列）上均显著优于无科学特性的基线，在保持相同响应长度的情况下实现高达65.3%的胜率；在10个单轮与多轮基准测试中，7B和8B模型有8个场景表现更优；盲测结果显示，领域专家认为其在创新性方面显著更优，且幻觉生成率低于基线与原始模型。
链接: https://arxiv.org/abs/2610.27901
作者: Sasi Kiran Gaddipati,Gollam Rabby,Sören Auer
机构: TIB—Leibniz Information Centre for Science and Technology, Hannover, Germany; L3S Research Center, Hannover, Germany
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 9 pages, 5 figures
Abstract:Large Language Models (LLMs) are increasingly used as research collaborators for hypothesis refinement, methodological planning, and technical analysis across multi-turn interactions. In this setting, a useful assistant must be accurate, methodologically justified, logically consistent, and token-efficient. Current alignment approaches rarely achieve all of these together. Instruction-tuned models gain conversational quality through verbosity, and we show that science-agnostic general collaborative rewards compress responses at the expense of scientific rigor. We propose CollabSciReM, a goal-conditioned reward framework that decomposes the session-level reward into conversational quality, scientific rigor, and structural efficiency. We also introduce SciQuality, a pairwise evaluation of multi-turn research conversations grounded in 300 held-out scientific papers and scored on six dimensions: methodological soundness, justification, logical consistency, factual correctness, novelty, and responsiveness. Fine-tuned on 8.62K synthetic research conversations, CollabSciReM is preferred on SciQuality over a science-agnostic reward baseline trained on the same data for all four Llama and Qwen models, including a 65.3% win rate at matched response length. For the 7B and 8B models, CollabSciReM outperforms the science-agnostic reward baseline on the primary metric in 8 of 10 single- and multi-turn benchmark settings. In a blind study, domain experts rated CollabSciReM significantly higher on novelty than the general-reward variant and reported fewer hallucinated responses than both the general-reward variant and the base model.
[NLP-47] sting Associative Creativity in Humans and Multimodal LLM s with Difficult-to-Name Visual Stimuli
【速读】： 该论文旨在解决在面对日益复杂的抽象视觉刺激时，人类与多模态大语言模型（Multimodal Large Language Models, MLLMs）在概念关联性创作中的表现差异及其适应机制问题。核心挑战在于当视觉刺激难以映射到熟悉语义概念时，如何衡量并理解其在创造性表达上的变化。解决方案的关键在于构建两个具有可控难度操纵的基准数据集——SilCC（通过逐步复杂的抽象形状组合提升命名难度）和PixCC（通过增加随机视觉噪声提升识别难度），并基于原创性（originality）与有效性（effectiveness）两个维度对人类与先进MLLMs进行系统评估。研究发现，人类在高难度条件下仍能生成更具多样性、语义距离更远、熵值更高且更新颖的描述，而随着任务难度上升，人类在PixCC中逐渐从整体性联想转向低层级视觉特征描述，显示出对复杂性的适应策略；相比之下，MLLMs在语义创新性和表达多样性上显著逊于人类，揭示了当前模型在深层抽象关联与创造性生成方面的局限性。
链接: https://arxiv.org/abs/2610.27894
作者: Manar Ali,Lisa Gottschalk,Simeon Junker,Lisa Hudalla,Alon Fishman,Joana Cholin,Stefan Kopp,Hendrik Buschmeier,Sina Zarrieß
机构: Bielefeld University (比勒费尔德大学); CRC 1646 ‘Linguistic Creativity in Communication’
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: Under review
Abstract:We investigate associative creativity in descriptions of abstract visual stimuli by humans and multimodal large language models (MLLMs). While conceptual association is central to human communication, it remains unclear how humans and models adapt when visual stimuli become increasingly difficult to map onto familiar semantic concepts. We evaluate several state-of-the-art MLLMs and human participants along two dimensions of creativity, originality and effectiveness, using two datasets with controlled manipulations of difficulty. SilCC varies naming difficulty through progressively more complex combinations of abstract shapes, and PixCC varies difficulty through increasing levels of random visual noise. Results show that humans produce more diverse associative descriptions than MLLMs, including greater semantic distance, higher entropy, and more novel expressions relative to reference lexica. As difficulty increases, effectiveness declines in SilCC, and humans increasingly shift from holistic associations to descriptions of low-level visual properties in PixCC.
[NLP-48] From Tokens to Files: Multi-Granularity Detection and Localization of AI-Involved Code
【速读】： 该论文旨在解决生成式 AI（Generative AI）在代码生成中广泛应用背景下，如何精准识别并定位混合作者代码中由模型生成或显著修改的部分这一关键问题。现有方法通常将整个文件压缩为单一评分，依赖表层风格特征或全局代码表示，导致当仅有少数代码块受 AI 影响时，其局部证据被平均化而难以察觉。为此，本文提出一种基于标记级（token-level）建模的框架，通过冻结的代码语言模型配备两个低秩适配器（low-rank adapters），分别在机器生成与人类编写代码上独立训练；利用两者在每个标记上的对数似然比作为局部 AI 作者身份的证据。轻量级头部模块结合上下文信息，在行和块级别聚合该证据，并进一步融合形成文件级别的决策。模型在由真实人类代码及其 AI 生成或修改版本构建的混合作者数据集上进行训练，涵盖续写与填空等多种生成场景。实验结果表明，在包含 11,773 个含 AI 成分文件的内部测试集上，该方法在 1% 误报率（FPR）下达到 72.4 的宏平均 F1；在无需微调的 HybridCodeAuthorship 外部基准上，线级 F1 达 62.5，优于原论文报告的最佳结果（56）。此外，在 19,445 个 2022 年前创建的开源仓库代码上，以相同 1% 阈值评估时，误报率维持在 0.55%~0.81% 之间，验证了该操作点在真实人类代码上的稳健性。因此，该方案的核心创新在于实现从标记到文件多层级的细粒度、可解释的 AI 代码溯源能力。
链接: https://arxiv.org/abs/2610.27863
作者: Yue Xu,Xiyang Wang,Hui Liu,Bolei He
机构: Peking University (北京大学); Baidu Inc. (百度公司)
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 29 pages
Abstract:The increasing use of large language models for code generation calls for methods that distinguish purely human-written code from AI-involved code: files in which any part, however small, was generated or substantially modified by a model. Beyond flagging such files, we aim to localize the lines or blocks attributable to AI. Existing approaches compress a file into a single score, from surface stylistic features or a global code representation, which averages away the evidence when only a few blocks are AI-involved. We propose a framework that models AI authorship at the token level and predicts at the line, block, and file levels. A frozen code language model is equipped with two low-rank adapters, trained separately on machine-written and human-written files; their per-token log-likelihood ratio provides local evidence of AI authorship. Lightweight heads aggregate this evidence within lines and blocks, using the surrounding code context, and their local predictions are aggregated into a file-level decision. We train on mixed-authorship files built from human originals and their AI-generated or AI-modified counterparts, including continuation and fill-in-the-middle scenarios. On an internal test set of 11,773 AI-involved files across six generation regimes against 2,541 human files, our method reaches 72.4 macro F1 at a 1% FPR operating point. With no adaptation, it reaches 62.5 line-level F1 on HybridCodeAuthorship, an external human-AI co-writing benchmark with no training split, above the 56 its own paper reports as the best line-level result. On 19,445 human-written files from repositories created before 2022, the false-positive rate stays between 0.55% and 0.81% under the transported 1% threshold, so the operating point holds on real human code.
[NLP-49] UvA-MTs Participation in the WMT26 General Translation Shared Task
【速读】： 该论文旨在解决通用机器翻译（General Machine Translation）中在无约束翻译方向下的翻译质量优化问题，特别是在面对23个翻译方向时如何有效提升译文准确性和流畅性。其解决方案的关键在于充分利用大语言模型（Large Language Models, LLMs）在测试阶段的推理能力：首先，采用基于精炼提示（refinement prompt）的序列采样策略，为每个源语言片段生成一组翻译候选，该方法相较于传统的并行采样更具优势；其次，引入LLM作为成对评判者，通过轮转投票机制对候选译文进行对比评估，从而选出最优结果，该方法在WMT25基准上的离线评估中显著优于传统的点对点“最佳N选一”（best-of-N）选择策略。两项技术的协同应用构成了其在WMT 2026通用机器翻译共享任务中的核心提交方案。
链接: https://arxiv.org/abs/2610.27852
作者: Di Wu,Sergey Troshin,Wafaa Mohammed,Seth Aycock,Evgeniia Tokarchuk,Vlad Niculae,Christof Monz
机构: University of Amsterdam (阿姆斯特丹大学)
类目: Computation and Language (cs.CL)
备注: Accepted to WMT26
Abstract:This paper presents UvA-MT’s submission to the WMT 2026 General Machine Translation shared task, competing in the unconstrained track across all 23 translation directions. This year, we fully leverage the test-time methods of Large Language Models (LLMs) for machine translation. Specifically, (1) we use sequential sampling with a refinement prompt to generate a pool of translation candidates for each source segment, a strategy shown to outperform traditional parallel sampling; and (2) we employ LLMs as pairwise judges to select the best candidate via a round-robin voting mechanism, with offline evaluation on the WMT25 benchmark showing a very clear improvement over point-wise best-of-N selection. Combined, these two strategies form our WMT26 submission.
[NLP-50] SciDiagBench: A Type-Specific Benchmark for Human-Aligned Evaluation of Scientific Diagrams
【速读】： 该论文旨在解决现有基于视觉语言模型（VLM）的科学图表生成评估基准在评估过程中未能反映人类专家在同行评审等科学评价中所采用的类型特异性评估标准的问题。具体而言，现有基准存在两大缺陷：其一，数据集缺乏对科学图表类型的系统分类，导致对所有图表采用通用评估标准；其二，评估指标多为启发式设计，未与人类评价偏好明确对齐。针对上述问题，论文提出SciDiagBench，其核心创新在于构建了一个系统化分类11类科学图表的基准数据集，并引入基于人类偏好引导的校准机制，以生成具有类型特异性的评估标准。该方案的关键在于通过人类偏好引导的校准过程，将真实人类评价习惯嵌入自动化评估流程，从而实现更可靠、更贴近人工评估结果的自动化评估，显著提升对多样化科学图表质量评价的一致性与有效性。
链接: https://arxiv.org/abs/2610.27841
作者: Hao Wang,Zixuan Li,Long Bai,Xiaolong Jin,Jiafeng Guo,Tat-Seng Chua,Tian Tian
机构: Institute of Computing Technology, Chinese Academy of Sciences (中国科学院计算技术研究所); State Key Laboratory of AI Safety (人工智能安全国家重点实验室); National University of Singapore (新加坡国立大学); School of Computer Science, University of Chinese Academy of Sciences (中国科学院大学计算机科学学院); School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences (中国科学院大学交叉科学学院)
类目: Computation and Language (cs.CL)
备注: 27 pages, 8 figures, 9 tables
Abstract:Recent agents have demonstrated strong capabilities in generating images from textual prompts, motivating their utilization for scientific diagram generation. Scientific diagrams naturally span diverse types, each characterized by distinct visual structures and content organization. Consequently, during peer review and other scientific evaluation, human experts naturally consider type-specific criteria when assessing their quality. However, existing VLM-based evaluation benchmarks fail to reflect the above human evaluation practices for two main reasons: 1) benchmark datasets typically lack a systematic categorization of different diagram types and thus adopt the same generic criteria for all diagrams, and 2) their evaluation criteria are often designed heuristically, without explicit alignment to human evaluation preferences. Motivated by these limitations, we introduce SciDiagBench, which consists of a benchmark dataset that systematically categorizes scientific diagrams into 11 types and a human preference-guided evaluation method. Specifically, SciDiagBench contains 165 scientific diagrams across 10 broad domains, with types including graphical abstracts, technical roadmaps, and illustrative cases. Based on this dataset, SciDiagBench develops an evaluation method that leverages a preference-guided calibration process to incorporate human preferences into the generation of type-specific criteria. Ultimately, SciDiagBench enables more reliable automated evaluation that better aligns with human evaluation across diverse scientific diagrams.
[NLP-51] SMART: Semantic Segment Mapping and Assessment of Recognition Transcripts with Contextual Rubrics
【速读】： 该论文旨在解决自动语音识别（ASR）评估中语义层面评价不足的问题，即传统词错误率（WER）对所有词汇级编辑赋予相同代价，忽略了不同错误在语义上的实际影响差异；而现有整体语义评分方法又无法定位具体哪些错误改变了原始意图。为此，论文提出一种上下文感知的、分段级的语义后果评估框架——基于上下文评分的识别转录本语义段映射与评估（Semantic Segment Mapping and Assessment of Recognition Transcripts with Contextual Rubrics, SMART）。其核心在于将参考文本与假设文本之间的差异映射为语义连贯的段落单元，并通过结构化上下文评分标准为每个段落分配明确的语义后果标签，从而支持语义影响错误率（SIER）的计算，实现从话语、片段到词粒度的可追溯语义错误量化。实验表明，相较于直接使用大语言模型（LLM）判断，SMART在多数投票人类标注下达到0.843的F1分数，显著优于0.745的直接LLM评分表现，且对所选评估模型具有较强的鲁棒性。此外，引入无目标-参考的独立评分标准指导纠错，结合目标假设及其邻近假设信息，在两个数据集上均优于基线，使词错误率（MER）和语义影响错误率（SIER_lex）分别绝对降低0.85–0.98和1.18–1.65个百分点，验证了该框架在精准语义评估与针对性修正方面的有效性。
链接: https://arxiv.org/abs/2610.27820
作者: Xiaoyu Gu,Chencheng Wang,Fuyang Li,Yu Xi,Tao Liu,Jiaoqi Guo,Kai Yu
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注:
Abstract:Word error rate (WER) assigns equal cost to lexical edits despite their different semantic consequences, while overall semantic scores cannot identify which individual errors alter the intended meaning. We formulate ASR semantic evaluation as context-aware, segment-level consequence assessment and propose Semantic Segment Mapping and Assessment of Recognition Transcripts with Contextual Rubrics (SMART), a framework for traceable semantic evaluation and targeted ASR correction. SMART maps reference-hypothesis discrepancies into coherent segments and applies a structured contextual rubric to assign explicit consequence labels. These labels support Semantic Impact Error Rate (SIER), which quantifies consequential errors at utterance, chunk, and lexical levels. Against majority-vote human labels, SMART achieves 0.843 F1 versus 0.745 for direct LLM judgment. Evaluation across multiple LLMs further demonstrates that SMART is relatively insensitive to the choice of evaluation model, while MER and SIER reveal different aspects of ASR quality. A separate target-reference-free rubric guides correction using the target hypothesis and surrounding ASR hypotheses, outperforming a matched w/o Rubric baseline on both datasets, with further absolute reductions of 0.85-0.98 percentage points in MER and 1.18-1.65 percentage points in SIER_lex.
[NLP-52] ProVe: Proposer-Verifier Decoding
【速读】： 该论文旨在解决自回归（Autoregressive, AR）语言模型在生成过程中因串行解码导致的推理速度瓶颈问题。现有方法如推测解码（Speculative Decoding）通过小型模型串行草拟候选词，由大型模型并行验证来加速，但其性能受限于草稿质量与验证效率之间的平衡。本文提出一种新范式——提议-验证解码（ProVe），其核心在于反转传统角色：以标准AR模型为起点，通过微调使大模型具备“提议者”（Proposer）能力，能够单次前向传播预测一整块未来token；随后通过一次额外前向传播由验证者（Verifier）验证该块内容。该方案的关键创新在于将大模型转化为高效提议者，同时保留了对原始模型行为的精确复现能力（即ProVe-lossless模式），实现无需训练独立草稿头即可达到2.5倍平均自回归吞吐量，且在使用小型3B验证者时仍可实现2.7倍平均吞吐量，仅在六任务平均上损失2.6个准确率点，从而在速度与精度之间提供可控权衡。
链接: https://arxiv.org/abs/2610.27815
作者: Ziqi Jin,Weiwen Xu,Jiaxi Li,Yujie Zhong,Qi She,Aixin Sun
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Speculative decoding accelerates autoregressive (AR) language models by letting a small model draft tokens serially and a large model verify them in parallel. We reverse these roles. Starting from a standard AR checkpoint, we fine-tune the large model into a proposer that predicts a block of future tokens in a single forward pass, then use one additional forward pass to verify the proposed block. This gives ProVe, Proposer-Verifier decoding. The verifier can be a smaller AR model or the proposer itself; the latter gives ProVe-lossless, which reproduces the proposer’s greedy AR decoding. Block proposals already provide substantial room for acceleration: with a block size of 16, the proposed block has an average AR-matching prefix of 6.6 tokens on GSM8K. Across six benchmarks with a 32B proposer, ProVe-lossless reaches 2.5x average AR throughput without training a separate draft head. With a small 3B verifier, ProVe reaches 2.7x average throughput at 2.6 accuracy points below AR on the six-task average, providing a controllable tradeoff between speed and accuracy.
[NLP-53] Are Linguistic and Arithmetic Circuits Dissociated in LLM s? AACL
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中语言理解与数学运算等不同任务能力之间的神经电路是否存在重叠的问题。具体而言，研究关注的是：在模型内部，处理语言任务（如语法判断）与算术任务（如数学应用题求解）的神经回路是否共享或相互耦合。其解决方案的关键在于采用机制可解释性（mechanistic interpretability）领域的电路发现方法，结合受控输入材料，系统地提取并比较分别针对语言和算术任务的神经回路，并分析其重叠程度与因果效应。实验结果表明，回路相似性主要由任务领域（语言 vs. 算术）决定，而非输入输出词汇表或自然语言句子与算术表达式之间的句法相似性，这揭示了大语言模型内部可能存在着以能力为基础的模块化神经组织结构，即语言与算术信息主要通过相对独立的神经回路进行处理。
链接: https://arxiv.org/abs/2610.27790
作者: Riku Kisako,Tatsuki Kuribayashi,Ryohei Sasano
机构: Nagoya University (名古屋大学); MBZUAI (MBZUAI); Tohoku University (东北大学)
类目: Computation and Language (cs.CL)
备注: AACL-IJCNLP 2026 main
Abstract:Large language models (LLMs) exhibit both knowledge of language, rendered by, for example, grammatical judgment ability, and abilities to perform tasks through language, such as math-word problem solving. Motivated by neurolinguistic studies of their neural dissociation in human brains, in this study, we investigate the overlap of their circuits in large language models (LLMs). Specifically, we examine the overlap and causal effects between the circuits extracted for linguistic and arithmetic tasks, using controlled materials and circuit discovery methods from the mechanistic interpretability field. Our experiments with five LLMs show that circuit similarity is mostly explained by task domains (linguistic vs. arithmetic), rather than by their input and output vocabularies or by the syntactic similarities between natural-language sentences and arithmetic expressions. This finding suggests that LLMs may process linguistic and arithmetic information using largely separate neural circuits, highlighting an ability-based modular organization of circuits within these models.
[NLP-54] Give the Model a Finger: Developmentally Aligned Evaluation of Counting in Vision-Language Models NEURIPS2026
【速读】： 该论文旨在解决视觉语言模型（Vision-Language Models, VLMs）在小样本下表现精确但在大样本下出现计数偏差的问题，其核心在于揭示模型在执行计数任务时所依赖的隐含序列推理过程，并通过引入可审计的测量机制来诊断和分析这种推理行为。解决方案的关键在于构建一个无需训练的代理循环（training-free agent loop），将人类儿童发展计数能力中的“认知行动”——如逐一点指、移动已计数项、分区排列等——转化为对图像的简单标记操作，作为外部工具提供给模型使用。这一方法不提升模型本身的视觉智能，而是通过强制模型生成可追踪的标记痕迹，实现对其内部计数逻辑的“可解释性审计”。具体而言，每个标记被基于像素级真实情况分类为命中、重复、幻影或遗漏，答案与标记数量进行比对，并通过反向遍历顺序测试其对顺序无关性的理解。研究发现：（i）精确计数范围是模型特异性特征，跨度近两个数量级；（ii）标记行为本身并非有效干预手段，仅对部分模型有益，且常因精度不足而适得其反；真正价值在于形成可审计的轨迹；（iii）作为测量手段，该框架成功揭示了精确指认者依赖标记获取答案、指认精度决定受益程度、密集场景因遗漏失效、且模型极少根据自身标记进行修正等内在机制；（iv）模型在遵循指令遍历顺序的同时，其答案仍会随顺序反转而改变，表明计数策略与顺序敏感性之间存在解耦，且类似儿童在规则布局中优于散乱布局的优势仅在规范程序下成立，且能抵抗行×列乘法捷径的干扰。因此，该研究的核心贡献在于提出一种以“测量而非干预”为导向的评估范式，使不可见的序列视觉推理过程得以可视化与量化分析。
链接: https://arxiv.org/abs/2610.27756
作者: Yingxin Jia,Hui Geng,Wenxin Yang,Mohan Zhang
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注: NeurIPS 2026 Workshop BabyVLM
Abstract:Vision-language models (VLMs) are exact on small sets and drift into estimation on larger ones, a profile reminiscent of children before they master counting. Children cross the barrier with epistemic actions: pointing at each item while reciting number words, moving counted items aside, partitioning the array. We give VLMs the same actions as dumb image tools with no visual intelligence of their own in a training-free agent loop. Our central claim is about measurement, not intervention: a model that counts by marking leaves a normally hidden serial procedure on the image, and the developmental principles of counting turn those marks into an interpretable audit of serial visual reasoning: every tag is classified against point-level ground truth as on-target, duplicate, phantom, or omitted, the answer is checked against the number of tags, and order irrelevance is tested by forcing opposite traversal orders. Across 9 models, 682 images, and 7 harness conditions we find that (i) the exact-counting range is a model-specific signature spanning nearly two orders of magnitude (R80 from 3 to 150 dots); (ii) as an intervention the finger mostly fails: tags help 4 of 9 models on small-count photographs, hurt imprecise pointers, and rarely beat a verbal routine or a blind point list; what the loop buys is an auditable trace, not accuracy; (iii) as measurement it succeeds: a tag-perturbation probe shows that precise pointers read their answers off the tags, pointing precision tracks which models benefit, dense scenes fail by omission, and models rarely revise after seeing their own tags; and (iv) the capacities dissociate: models follow an instructed traversal order yet often change their answer when it reverses, and children’s regular-over-scattered advantage appears only under the routine and survives lattices that defeat a rows-times-columns shortcut.
[NLP-55] Hearsay or Fact? How Framing and Elicitation Shape LLM Probability Updates DATE
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在接收到单一新信息时，其概率估计值的变动方向与幅度问题，尤其关注模型在心理实验中作为判断助手或参与者时对证据更新的敏感性。现有研究缺乏对模型在面对新增信息时概率调整行为的系统性测量。为此，作者基于δ-SNLI数据集中的前提-假设对，分别引入强化项（strengthener）和弱化项（weakener），通过分步调用获取模型的先验概率与后验概率，并计算“方向性变动量”（s），即以标注方向为符号的后验均值变化量，从而消除先验的影响。研究进一步考察了信息更新的呈现方式（直接陈述“这里你学到了新信息” vs. 间接转述“有人提到”）以及概率输出形式（单个概率值 vs. 五个加权估计）的影响。实验涵盖七种配置：两组本地部署模型（Qwen3-8B、OLMo-2-7B；240对样本）及五种通过代理命令行框架运行的配置（120对样本），所有假设与分析规则在正式测试前已固定。结果显示，在直接陈述条件下，所有七种配置均表现出与标注方向一致的显著变动（s = 18–28个百分点）；而采用转述方式则导致七种配置的变动量普遍下降（六组差异显著，降幅2–5个百分点），且反向变动的比例上升，表明模型对间接信息的响应减弱但趋势仍一致。当采用五个加权估计时，极端概率（≤5% 或 ≥95%）占比降低，但在本地模型中却使方向性变动量大幅减少（分别下降43%和68%），而在代理框架中变动量仅在-3.7至+0.5个百分点之间波动。这表明，抑制极端概率的机制并不必然保留模型对证据的敏感性，二者需独立评估。由于δ-NLI仅标注方向而非幅度，研究不涉及更新应多大的判断。关键发现在于：模型对证据的敏感性与其概率分布的极端程度并非正相关，必须分别设计实验进行验证。
链接: https://arxiv.org/abs/2610.27745
作者: An-Sheng Jhang
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 27 pages, 3 figures, 8 tables. v2 adds four model configurations (gpt-oss-120b, claude-sonnet-5.5, gemini-3.8-flash at two reasoning efforts) run with the same protocol, and an exploratory analysis of candidate spread. Code and data: this https URL
Abstract:Large language models (LLMs) are increasingly used as judgment assistants and as stand-in participants in psychology experiments, yet little is known about how much their probability estimates move in response to a single new piece of information. We pair premise-hypothesis items from \delta -SNLI with one strengthener and one weakener each, elicit the prior and the posteriors in separate calls, and measure the directional movement s: the mean posterior change signed by the annotated direction, so that the prior cancels out. We vary whether the update is stated (“Here is something else you learned”) or reported (“Someone said that”), and whether the model gives one probability or five weighted estimates. We test seven configurations: two local models (Qwen3-8B, OLMo-2-7B; 240 pairs) and five configurations run through agent command-line harnesses (120 pairs), with hypotheses and analysis rules fixed before the formal runs (not publicly registered). Under stated, direct elicitation all seven move in the annotated direction (s = 18-28 percentage points). Reporting lowers s in all seven point estimates (significantly in six, by 2-5 points) and raises the share of reverse movements in all seven: the effect is consistent but small. Five weighted estimates lower the share of responses at \leq 5% or \geq 95% in all seven, but in the two local models they also reduce s by 43% and 68%, whereas in the harnessed configurations the point estimates of the change lie between -3.7 and +0.5 points. An elicitation format that reduces extreme values thus need not preserve sensitivity to evidence; the two should be tested separately. Because \delta -NLI annotates only direction, we make no claim about how large an update should be. Code, data and the frozen plan: this https URL
[NLP-56] racing Recurrent and Emergent Hallucination Patterns in Decomposed Text-to-SQL Generation EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在Text-to-SQL任务中虽能取得较高基准性能，但仍频繁产生模式错误（schema errors）和查询逻辑错误（query-logic errors）的问题。现有评估方法仅关注最终生成的SQL查询结果，难以区分错误是源于早期推理阶段还是在生成末期才出现，导致对错误演化路径的理解不足。为此，本文提出一种“钻取式”（drill-down）评估框架，将Text-to-SQL生成过程分解为一系列逐步复杂的可执行子查询及其对应自然语言问题，从而实现对基于规则的幻觉类型（rule-based hallucination categories）在整个生成路径中的追踪与分析。其解决方案的关键在于引入两个关键概念：重复性幻觉（recurrent hallucinations），即同一类错误在早期步骤与最终查询中均被观测到；以及涌现性幻觉（emergent hallucinations），即错误仅在最终阶段出现而未在前期路径中显现。通过对五个LLMs在扩展版BIRD-mini基准上的实验发现，部分错误类别具有显著的重复概率，而“子句滥用”（clause abuse）则表现出更高的条件性涌现概率。在对话历史（conversational-history）条件下，重复概率普遍升高，而涌现概率普遍降低，表明两类条件概率并非同向变化，因此仅依赖最终输出评估存在局限。这一发现强调了在评估中引入轨迹级（trajectory-level）分析的重要性，以揭示错误的动态演化机制。
链接: https://arxiv.org/abs/2610.27720
作者: Hayden Moore,Abhinav Verma,Anshuman Dutt
机构: The Pennsylvania State University (宾夕法尼亚州立大学); Microsoft(微软)
类目: Computation and Language (cs.CL)
备注: Accepted to the 3rd Workshop on Uncertainty-Aware NLP (UncertaiNLP 2026), co-located with EMNLP 2026
Abstract:Large language models can achieve strong Text-to-SQL benchmark performance while still producing schema and query-logic errors. Most evaluations characterize such failures only in the final generated query, making it difficult to determine whether a related error was also observed in an earlier, simpler step or appears only at the final stage. We present a drill-down evaluation framework that decomposes Text-to-SQL queries into progressively more complex executable sub-queries and corresponding natural-language questions, allowing rule-based hallucination categories to be examined across a generation path. We distinguish recurrent hallucinations, in which the same rule-based category is observed in an earlier step and again in the final query, from emergent hallucinations, in which the category is observed at the final step without an earlier detection in the same path. Across five LLMs evaluated on an expanded BIRD-mini benchmark, several categories show appreciable recurrence probabilities, while clause abuse shows comparatively higher conditional emergence. Under the conversational-history condition, recurrence probabilities are generally higher and emergence probabilities are generally lower. These observations indicate that the two conditional measures do not move uniformly in the same direction under additional context and motivate trajectory-level analysis in addition to final-output evaluation. Code and evaluation artifacts are available in the project repository. this https URL
[NLP-57] When Does Experience Memory Help Closed-Loop LLM Driving Agents ? A Controlled Evaluation with Resampled Warmup Pools Token-Matched Fillers and Audited Baselines ACML2026
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在闭环大语言模型（LLM）自动驾驶代理中，经验记忆（experience memory）何时真正发挥作用的核心问题。现有研究的性能提升无法提供可靠答案，因其未控制两个关键变量：暖启动经验池（warmup pool）的内容构成，以及注入提示词中的无关令牌（token）对提示的影响。为此，作者在经过审计的HighwayEnv测试平台（1Hz采样率，使用gpt-4o-mini进行关键对比）上开展受控评估，将暖启动池视为重采样的随机因子（十组互不重叠的池），并引入匹配令牌的无关填充控制、基于“可信源”的诊断机制、谨慎定价指标及基线端口的量化审计。所考察的探针方案涵盖纯文本检索（DiLu端口）、单帧网格端口（受RESPOND启发）、基于情景的时间图记忆、无检索整合规则手册，以及基于K步观测历史的基准方案。研究得出四项发现：（1）显著的记忆增益仅存在于特定暖启动样本中；报告的最优经验池仅为十组中表现最佳的一组，且仅在一组中检索型记忆优于无记忆基线（符号检验p=0.011），而规则手册看似“池无关性”在五次新抽样后即失效；（2）所有检索设计均未超越匹配令牌的无关填充控制，仅一条15令牌的“仅碰撞警告”指令即可达到碰撞合并上限，而强调谨慎性的度量指标抹平了规则手册的优势；（3）表征复杂性在任何情况下均未表现出稳健的有效性；（4）基线端口质量的隐性差异会彻底反转结论——若无填充控制的单池比较，将错误报告存在记忆增益，而控制条件反而揭示其为虚假信号。研究结果表明，当前经验记忆在该测试场景中并非稳健有效，其性能高度依赖于实验设计的细节与控制手段。论文已开源测试平台、审计后的端口实现、种子及缓存数据。
链接: https://arxiv.org/abs/2610.27714
作者: Yingxin Jia,Xuanyi Li
机构: Independent Researchers(独立研究员)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注: ACML 2026
Abstract:When does experience memory help a closed-loop LLM driving agent? Published gains cannot say, because the two factors we find to matter most, which warmup experiences the pool contains and what any injected tokens do to the prompt, are controlled in none of them. We answer with a controlled evaluation on one audited HighwayEnv testbed (1Hz, gpt-4o-mini for the decisive contrasts). We treat the warmup pool as a resampled random factor (ten disjoint pools) and add token-matched irrelevant-filler controls, oracle-certified diagnostics, caution-pricing metrics, and quantified audits of the baseline ports. The probe suite spans flat-text retrieval (a DiLu port), a single-frame grid port (RESPOND-inspired), our episodic temporal-graph memory, a retrieval-free consolidated rulebook, and a flat K-step observation-history baseline. Four findings follow, each a property of this testbed. (1) Headline memory gains belong to the warmup draw. The reported pool is the best of ten for both retrieval memories, which beat No-Memory in only one pool of ten (sign test p=0.011), and even the rulebook’s apparent pool-immunity collapsed on five fresh draws. (2) No retrieval design tested exceeds token-matched filler. The collision-only merge ceiling is reached by a single 15-token caution instruction, and metrics that price caution erase the rulebook’s lead. (3) Representational sophistication is nowhere robustly operative. (4) Baseline port quality silently inverts conclusions. On this testbed, a single-pool comparison without filler controls would have reported memory gains that these controls reverse. We release the testbed, audited ports, seeds, and caches.
[NLP-58] Beyond Calibration: Temporal LLM Confidence Assessment Across Languages and Local Settings EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在推理过程中自我评估不确定性的能力问题，特别是探讨不同时间阶段的置信度信号是否蕴含不同的认知信息，并验证其在非英语语境下的可靠性。其核心解决方案在于借鉴经典元记忆理论，区分解决问题前的预期置信度（感觉已知，Feeling-of-Knowing, FOK）与解决后的回顾置信度（学习判断，Judgement-of-Learning, JOL），并系统评估这两种信号在多种推理任务、跨语言翻译的马拉地语（Marathi）基准以及新构建的真实本地马拉地语事实问答数据集中的表现。研究发现，成本较低的预求解置信度信号校准性较弱，而事后置信度虽存在额外计算开销，但通常具有更强的不确定性对齐能力；此外，直接语言翻译下置信度行为表现出较强的跨语言迁移性，但在涉及本地化知识时显著退化，表明单纯的语言迁移不足以满足区域特定信息需求；更重要的是，幻觉生成的回答在生成后仍保持异常高的置信度，暴露出当前大模型自我评估机制的关键缺陷。
链接: https://arxiv.org/abs/2610.27702
作者: Sahil Kale
机构: University of California, Los Angeles (加州大学洛杉矶分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 10 pages, 7 figures, WiNLP @ EMNLP 2026
Abstract:Large language models are expected to estimate their own uncertainty, with confidence signals used to guide retrieval, adaptive compute allocation, routing, and selective abstention. We investigate whether confidence signals measured at different temporal stages of problem solving encode distinct forms of epistemic information and remain reliable beyond English-centric settings. Inspired by classical meta-memory theory, we distinguish anticipatory pre-solve confidence (Feeling-of-Knowing, FOK) from retrospective post-solve confidence (Judgement-of-Learning, JOL). We evaluate these signals across two frontier LLMs, multiple reasoning domains, translated Marathi benchmarks, and a newly constructed benchmark of genuinely local Marathi factual questions. Across settings, the cheaper pre-solve signal remains weakly calibrated, while post-solve confidence generally provides stronger epistemic alignment despite its additional inference cost. Confidence behaviour transfers surprisingly well under direct language translation, with calibration trends largely preserved between English and Marathi, but degrades substantially on local Marathi knowledge, suggesting that linguistic transfer alone is insufficient for region-specific information needs. Finally, hallucinated responses maintain persistently elevated confidence after answer generation, highlighting important limitations of current LLM self-assessment mechanisms.
[NLP-59] Memory as Code: Declarative Retrieval over Codified Long-Term Memory
【速读】： 该论文旨在解决长时记忆系统在复杂场景下难以有效整合跨会话分散的事实并动态适应多样化查询的问题。现有记忆系统受限于静态的、非结构化的信息存储方式，无法灵活组合随时间演化的事实关系，从而制约了大语言模型（LLM）代理的推理能力。其解决方案的关键在于提出“记忆即代码”（Memory as Code, MaC）这一统一框架，通过共享的代码抽象联合设计记忆构建与检索机制。MaC 的核心创新在于将记忆构建过程形式化为可执行的代码抽象，构建包含原子函数层、关系层和模块化组合层的三层可执行记忆结构；同时，将记忆检索转化为基于声明式查询处理的程序生成，将自然语言查询映射为结构化的检索程序。该方法实现了记忆的模块化与可组合性，显著提升了在 LoCoMo 与 LongMemEval 基准上的推理准确率（相对最强基线最高提升 15.91%），并大幅降低令牌消耗，验证了“记忆即代码”在长时程记忆任务中的有效性。
链接: https://arxiv.org/abs/2610.27700
作者: Aoting Zeng,Yuxin Du,Jianwei Wang,Yizhang He,Shan Jiang,Xuemin Lin,Kai Wang
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:In long-term memory systems, memories are constructed from interaction histories and retrieved in response to queries, enhancing the reasoning ability of LLM agents. In complex scenarios, facts are scattered across sessions and their relationships evolve over time, making it difficult for existing memory systems to compose them to adapt to diverse queries. Motivated by the modularity and composability of code, we propose Memory as Code (MaC), a unified memory framework that jointly designs memory construction and retrieval through a shared code abstraction. MaC couples codified construction, which establishes executable memory abstractions, with declarative retrieval, which translates queries into structured retrieval programs. Specifically, codified construction builds a three-layer executable memory structure that progresses from an atomic function layer, through a relational layer, to a modular composition layer. Declarative retrieval recasts retrieval from relevance-based fact selection into declarative query processing over memory. Experiments on LoCoMo and LongMemEval show relative gains of up to 15.91% in overall LLM-judge accuracy over the strongest baseline in each setting, while substantially reducing token consumption, highlighting the effectiveness of memory as code for long-horizon memory tasks.
[NLP-60] CovNav: Conformal Prediction-Guided Adaptive Inference for Object Goal Navigation
【速读】： 该论文旨在解决具身智能体在未知环境中进行目标导航（ObjectNav）时，如何在推理质量与计算效率之间取得平衡的问题。具体而言，尽管大型语言模型（LLM）能够编码空间常识以支持导航决策，但其部署存在显著权衡：云端部署虽推理能力强却面临延迟与隐私风险，而小型本地化模型虽快速但易产生幻觉式空间判断。为应对这一挑战，本文提出CovNav框架，其核心创新在于采用置信预测（Conformal Prediction, CP）实现对双过程推理（系统1/系统2）的不确定性感知决策——即通过一个轻量级本地小模型在快速模式下判断目标是否可能位于已观测物体附近，利用CP将模型输出的概率转化为预测集，仅当结果不确定时才触发耗时的链式思维（chain-of-thought）慢速推理。该方法基于对象对的形式化建模，保持了校准的可计算性，并可通过现成的目标检测器支持开放集对象。在Gibson和HM3D数据集上的实验表明，CovNav在HM3D v1和v2上分别超越最强零样本基线0.8和1.7个百分点的成功率，同时仅对12.4%的查询调用慢速推理，在大幅降低延迟的同时保留了持续思考策略的大部分性能优势。真实机器人（TurtleBot 4）实验进一步验证了该方法的有效性。CovNav展示了如何通过校准的不确定性来决定具身智能体何时应进行深度思考，为基于大模型的机器人提供了可解释、自适应的测试时推理范式。
链接: https://arxiv.org/abs/2610.27677
作者: Pengying Wu,Zhejia Wen,Kang Ding,Kangjie Zhou,Chunxuan Jiao,Wei Pan,Xiangli Nie,Chang Liu
机构: 未知
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages, 6 figures, 7 tables
Abstract:In object goal navigation (ObjectNav), an embodied agent must find a target object (e.g., “find a bed”) in a previously unseen environment, relying on spatial commonsense such as object co-occurrence. Large Language Models (LLMs) encode this commonsense, but deploying them on robots exposes a trade-off between reasoning quality and efficiency: cloud-based models incur latency and privacy risks, whereas small on-device LLMs are fast but prone to hallucinated spatial judgments. Hybrid reasoning LLMs offer both a fast non-thinking mode and a slow thinking mode, which raises a key question for embodied agents: when is slow thinking worth its cost? We propose CovNav, a training-free framework that answers this question with Conformal Prediction (CP), turning dual-process (System 1 / System 2) inference into an uncertainty-aware decision. Much like recent decision models such as Jev, CovNav casts each spatial query as a structured binary decision rather than free-form generation: a local LLM in fast mode judges whether the target is likely near an observed object, CP converts its token probabilities into a prediction set, and only ambiguous queries are escalated to slow chain-of-thought reasoning. The object-pair formulation keeps calibration tractable and supports open-set objects through off-the-shelf detectors. Experiments on Gibson and HM3D show that CovNav surpasses the strongest zero-shot baseline by 0.8 and 1.7 percentage points in success rate on HM3D v1 and v2, while invoking slow reasoning for only 12.4% of queries and retaining most of the gain of always-on thinking at a fraction of its latency. Real-robot experiments on a TurtleBot 4 further validate the approach. CovNav shows how calibrated uncertainty can decide when an embodied agent should think slowly, offering a principled form of adaptive test-time reasoning for LLM-based robots.
[NLP-61] Corpus Quality Improves Retrieval Utility Does Not: An Ablation of Preprocessing and Refinement in a Speech-to-QA Pipeline
【速读】： 该论文旨在解决大规模语言模型（LLM）训练语料库构建过程中，预处理与精炼阶段对语料质量及下游任务性能影响的矛盾问题。其核心挑战在于：尽管通过引入预处理（重写转录文本）和精炼（过滤生成的问答对）两阶段可显著提升语料自身的主观质量评分，但这种质量提升是否真正转化为下游模型在信息检索等任务中的性能增益仍不明确。解决方案的关键在于设计一个可控制变量的实验框架，系统性地关闭或开启两个阶段，从而构建四组源自相同40场韩语医学与金融会议的对比语料库，并通过多轮评估验证其效果。结果表明，虽然双阶段开启使语料在人工与大模型评判中得分更高（分别提升+0.18和+0.22），但这一质量提升并未反映在下游模型的检索准确率上；反而，启用转录文本重写阶段导致检索准确率下降0.061，且该效应在多重比较校正后依然显著，同时使处理成本几乎翻倍。进一步的等效性检验确认，当两阶段均开启时，下游性能与仅使用原始语料无显著差异。这揭示出：语料自身质量评分（如基于人工或大模型的打分）无法可靠预测其对模型的实际价值，说明当前主流的质量评估指标可能与模型实际受益脱节。因此，该研究的关键洞见是——提升语料“看起来更好”的表层质量未必带来真实性能增益，甚至可能因引入冗余或偏差而损害下游任务表现。
链接: https://arxiv.org/abs/2610.27652
作者: Wonsup Shin,Jingu Kim
机构: Flitto; Flitto
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 14 pages, 3 figures, 9 tables. Code and data: this https URL
Abstract:Practitioners building training corpora with a pipeline of language models keep a preprocessing or refinement (post-process) stage when switching it on raises the corpus quality. We ablate a production pipeline that turns conference speech into question-answer pairs. The preprocessing stage rewrites the transcript, and the refinement stage filters the generated pairs. Switching each on and off gives four corpora from the same 40 Korean medical and finance sessions. On the corpus itself, quality rises as the stages switch on. Four LLM judges and six human domain experts rank the four corpora the same way, and the fully refined corpus leads by +0.18 and +0.22 on a 1-5 scale. We then ask whether that gain carries downstream. We supply each corpus to six models at inference and measure retrieval accuracy. The ranking does not hold. Enabling the transcript rewrite lowers retrieval accuracy by 0.061, the only stage effect that survives correction for multiple comparisons. That stage also reads the whole session, so it almost doubles the cost of one. With both stages on, accuracy does not change. An equivalence test we declared in advance confirms this. The corpus itself is clearly valuable. It lifts retrieval accuracy from 0.231 to 0.748 in every model and domain we test. More than a third of its pairs carry knowledge that neither of two frontier models produces on its own. On our data, the null does not come from weak corpora. A rising score on a rubric of this kind did not predict what the corpus holds for a model.
[NLP-62] Speak vs. Know: Balancing Behavior and Knowledge Distributions of Large Language Models
【速读】： 该论文旨在解决大语言模型（LLM）在生成观点预测时存在的“知识-行为不一致”问题：尽管模型在自我陈述中表达的分布（知识分布，knowledge distribution）更贴近人类群体意见，但其实际输出却由下一个词的对数概率分布（行为分布，behavior distribution）所主导。研究发现，在世界价值观调查（World Values Survey）的37%的问题上，基于知识分布的最可能回答与基于行为分布的最可能回答不一致，导致模型的“认知”与其“行为”脱节。这一差异严重影响了评估的有效性——仅依据知识分布评分可能产生误导，因为其最优答案未必是模型实际输出的内容。为此，论文提出一种可学习的混合分布方法，通过一个超参数权衡行为保真度与分布质量，在保持知识分布形状的同时，使最终输出遵循行为分布中最可能的答案。实验结果表明，该平衡分布显著提升了行为一致性（从71.6%提升至95.6%），同时与人类意见分布的总变差距离仅轻微增加（28.8%→33.8%），且在温度缩放和跨数据集场景下仍具鲁棒性。关键在于构建一个既能保留知识分布语义结构、又能确保输出行为一致性的可学习组合机制。
链接: https://arxiv.org/abs/2610.27648
作者: Yang Liu,Masahiro Kaneko,Chenhui Chu
机构: Kyoto University (京都大学); MBZUAI ( Mohamed Bin Zayed University of Artificial Intelligence)
类目: Computation and Language (cs.CL)
备注: 21 pages, 10 figures
Abstract:Recent work reports that the distribution verbalized directly in the output text aligns far better with human opinion distributions than the next-token log-probability distribution of LLMs. However, the verbalized distribution reflects what LLMs know about a group’s opinions (knowledge distribution), while what they actually output at deployment is governed by their log-probability distribution (behavior distribution). We find that when prompting LLMs to predict the opinion of a specific group, the most likely answer under the verbalized distribution differs from that under the log-probability distribution for 37% of World Values Survey questions, averaged over three LLMs. Thus, LLMs’ knowledge does not match their behavior. This matters for evaluation: a verbalized distribution can score well against a group while its most likely answer is one the model never outputs, so the score no longer describes the model. The alignment scores for these two distributions reflect only the model’s behavior or its knowledge, separately. We therefore propose a learnable combination of the two distributions that follows the most likely answer of the behavior distribution while preserving the shape of the knowledge distribution, with a single hyperparameter trading off behavioral fidelity against distribution quality. Our experiments on three opinion survey datasets and six models show that the balanced distribution improves behavioral consistency over the verbalized distribution from 71.6% to 95.6% on average, while the total variation distance to the human opinion distributions rises only slightly from 28.8% to 33.8%, and it remains effective under temperature scaling and cross-dataset settings.
[NLP-63] CUA-Recovery: Benchmarking Error Recovery in Long-Horizon Computer-Use Agents
【速读】： 该论文旨在解决生成式计算机使用代理（Computer-Use Agents, CUAs）在长时序多应用工作流中因未被检测到的错误导致持久状态污染及后续动作级联失败的问题。现有基准测试主要关注端到端任务性能，缺乏对代理在执行过程中识别与恢复错误能力的评估；而少数涉及恢复能力的基准依赖人工注入错误，难以反映真实代理策略引发的错误类型与时间跨度。为此，论文提出 CUA-Recovery 基准，基于 ReRail 框架在持久化桌面环境中合成长时序任务与轨迹，通过组合 MyPCBench 任务构建具备可执行真值的新工作流，收集多策略产生的验证失败案例，并转化为经验证的恢复轨迹。CUA-Recovery 包含一个由人工验证的 2,250 个错误状态的测试集（覆盖 100 项任务）和一个包含 8.6K 轨迹的训练集（覆盖 377 项任务）。评估结果表明，当前六种代理在识别继承性错误和完成任务方面均表现不佳，最大错误感知率（Error Awareness Rate, EAR）仅为 44.0%，任务通过率 Pass@3 最高仅达 25.5%。在对 Qwen3.5-35B-A3B 进行训练集微调后，其在保留恢复状态上的平均 EAR 提升至 35.2%（原为 19.9%），评分（Rubric Score）从 5.0% 提升至 16.1%，同时在 MyPCBench 与 Odysseys 基准上 Pass@3 分数分别提升 15.3 和 10.8 分，证明了训练轨迹在提升错误恢复能力与任务执行鲁棒性方面的有效性。解决方案的关键在于利用 ReRail 构建具有真实故障模式与长期依赖特性的合成数据，从而实现更贴近实际场景的错误检测与恢复训练。
链接: https://arxiv.org/abs/2610.27645
作者: Xu Hu,Haoming Li,Yuxiang Wang,Jinxin Yang,Kangshuo Li,Licheng Wang,Dingyi Kang,Jiayu Song,Tianhao Wang,Zhanhua Hu,Xu Yang,Qiannan Li,Bingzhe Li,Feng Chen
机构: University of Texas at Dallas (德克萨斯大学达拉斯分校); Independent Researcher (独立研究员); University of Manchester (曼彻斯特大学); Rice University (莱斯大学); University of Texas at Austin (德克萨斯大学奥斯汀分校); University of California, Davis (加州大学戴维斯分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 33 pages, 13 figures, 18 tables
Abstract:As computer-use agents (CUAs) move to long-horizon workflows spanning multiple applications, undetected errors can silently corrupt persistent state and cause cascading failures in subsequent actions. Reliable error detection and recovery have become critical bottlenecks for building robust CUAs. However, existing benchmarks focus primarily on end-to-end task performance, providing limited insight into whether agents can detect and recover from errors during execution. The few benchmarks that evaluate recovery rely largely on manually injected errors, which may not reflect the types or horizons of failures induced by real agent policies. We introduce CUA-Recovery, a benchmark comprising test and training suites across multiple desktop applications. We construct it using ReRail, a framework for synthesizing long-horizon tasks and trajectories in a persistent desktop environment. ReRail composes MyPCBench tasks into new long-horizon workflows with executable ground truth, collects verifier-confirmed failures from multiple policies, and turns them into verified recovery trajectories. CUA-Recovery comprises a human-verified test suite of 2,250 erroneous states over 100 tasks and a training suite of 8.6K trajectories over 377 tasks. Evaluation on the test suite shows that agents struggle both to recognize inherited errors and to complete the task. Across six agents and six takeover depths, Error Awareness Rate (EAR) reaches at most 44.0% and Pass@3 at most 25.5%. After fine-tuning Qwen3.5-35B-A3B on the training suite, its average EAR rises from 19.9% to 35.2% on held-out recovery states, and its Rubric Score goes from 5.0% to 16.1%. It also gains 15.3 points on MyPCBench and 10.8 points on Odysseys for Pass@3, showing that the training trajectories help both with recovery and with completing tasks from a clean start. Our code is available at this https URL.
[NLP-64] BCT: Bayes-Consistent Tempering for Flow Language Models
【速读】： 该论文旨在解决生成式语言模型中温度缩放（temperature scaling）与后验诱导流模型（posterior-induced flow language models）之间存在的双重不一致性问题。具体而言，传统的后验温度调节会同时改变终点分布和中间噪声状态的分布，但去噪器（denoiser）仍基于原始未调整的状态进行训练，导致分布与动力学之间的不匹配；此外，经过温度调节的后验权重与原本为未温度调节流设计的传输动力学相耦合，进一步加剧了这一矛盾。这种不一致性在实证上表现为：在流语言模型中降低温度对GSM8K任务性能的提升远不如在离散扩散语言模型中可靠。为此，论文提出贝叶斯一致温度调节（Bayes-Consistent Tempering, BCT），其核心在于将温度显式地纳入生成流的分布与动力学建模中，使去噪器在温度调节后的中间状态上进行训练，并利用相应的温度调节动力学来更新终点分布。BCT还提供了无需额外训练的变体，适用于已训练好的检查点。实验结果表明，无论是在双去噪器架构的超球面流语言模型还是冯·米塞斯-费舍尔流语言模型上，BCT均在广泛的温度范围内显著提升GSM8K生成质量，并在严格约束生成任务如数独求解中展现出优势。尤其在极低温条件下，学习型实现版本相比无训练变体表现出更强的鲁棒性，验证了温度应作为生成流在训练与推理全周期中的内在属性，而非仅作为事后概率调整手段。
链接: https://arxiv.org/abs/2610.27641
作者: Daisuke Oba,Hiroki Furuta,Naoaki Okazaki
机构: Institute of Science Tokyo; AIST; NII LLMC
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Temperature scaling is typically applied as a post-hoc transformation of model probabilities. In posterior-induced flow language models, this creates two mismatches. First, posterior tempering implicitly changes both the endpoint distribution and the distribution of intermediate noisy states, while the denoiser remains trained on the original states. Second, the tempered posterior weights are combined with transport dynamics derived for the untempered flow. These mismatches are reflected empirically: lowering temperature improves GSM8K performance much less reliably in flow LMs than in discrete diffusion LMs. We introduce Bayes-Consistent Tempering (BCT), which makes temperature explicit in both the distributions and dynamics of the generative flow. Its learned realization trains the denoiser on tempered intermediate states and tempers the endpoint distribution using the corresponding tempered dynamics. We also derive training-free variants for conventionally trained checkpoints. Across hyperspherical flow LMs with two denoiser architectures and a von Mises–Fisher flow LM, BCT improves GSM8K generation across broad temperature ranges and achieves the strongest tempering performance. The gains also extend to Sudoku, demonstrating benefits on tightly constrained generation. At very low temperatures, the learned realization is also markedly more robust than the training-free variants, supporting temperature as a property of the generative flow across training and inference rather than a post-hoc probability adjustment. Code and model weights are available at this https URL .
[NLP-65] A Two-Way Street: Surveying the Interplay between Code-Mixing and Native Languages in Text and Speech
【速读】： 该论文旨在解决代码混用（Code-Mixed, CM）文本与语音系统研究中普遍存在的数据稀缺问题，以及由此引发的跨语言迁移学习方向不明确、方法分散且缺乏系统性整合的挑战。现有研究多依赖于被混合的母语（native languages）资源进行迁移，但忽略了从CM数据反向提升母语模型的可能性，且缺乏对双向迁移（从母语到CM，及从CM到母语）在文本与语音模态中统一分析的综述。本文的关键解决方案在于构建一个涵盖迁移方向、路径和模态的系统性分类框架（taxonomy），首次将双向迁移研究整合在一起。通过该分类体系发现：当前迁移主要由母语单向流向CM场景；反向迁移尚处于早期阶段，其有效性主要体现在通过代码切换（code-switching）实现语言对齐；尽管文本与语音任务共享部分技术方法，但彼此引用极少；在共同语音基准测试中，使用母语数据带来的增益远超新架构设计。这些发现揭示了关键开放问题，如训练时应如何平衡母语与CM数据的比例，以及如何评估跨脚本书写形式的CM文本。
链接: https://arxiv.org/abs/2610.27631
作者: Debajyoti Mazumder,Akash Anil
机构: Indian Institute of Science Education and Research, Bhopal(印度科学教育与研究学院，博帕尔)
类目: Computation and Language (cs.CL)
备注: 40 pages
Abstract:Systems for code-mixed (CM) text and speech rarely start from scratch. CM data is scarce, so most work borrows from the languages being mixed, which we call native languages. This transfer can also run in reverse, when CM data is used to improve models for the native languages. Existing surveys focus on processing CM itself, and none brings the two directions together. This paper fills that gap with a systematic survey of transfer in both directions, across text and speech. We organise the surveyed work in a taxonomy of transfer direction, route and modality. Through this taxonomy, we find that most transfer runs from native languages into the CM setting. The reverse direction is younger, and in it code-switching mainly helps by aligning the two languages. Text and speech work share methods but rarely cite each other. On a shared speech benchmark, native data helps more than new architectures. These findings point to open challenges, such as how much native and CM data to mix in training, and how to evaluate CM text written in different scripts.
[NLP-66] PAR: Progress-Aligned Recurrence for Length Generalization in State Tracking
【速读】： 该论文旨在解决循环语言模型（Looped Language Models）在测试时虽可执行更多迭代，但额外迭代未必能有效处理更长问题的难题。其核心挑战在于：最终答案监督（final-answer supervision）无法明确每轮迭代应推进的程度，且重复更新会改变递归状态的尺度，导致模型难以外推至远超训练长度的问题。为此，论文提出进度对齐递归（Progress-Aligned Recurrence, PAR），关键在于将第 k 次迭代的预测目标设定为前 k 次操作后的状态，并通过将递归状态重缩放至首次迭代的范数来锚定状态尺度。实验表明，在 swap-S5 任务上，经最多 32 步训练的 Qwen2.5-0.5B 模型可在 512 步时达到 95.5% 的平均完整状态准确率，实现 16 倍于训练长度的外推能力，显著优于基于最终答案监督的 RLT（55.7%）和显式思维链（0.8%）。消融实验显示，移除范数锚定会使外推因子降至 2.2 倍以下，而使用最终答案监督替代进度对齐目标则限制在 8 倍以内。此外，引入学习的完成读出机制可实现无需预设停止深度的自适应终止；当缺乏中间标签时，前缀自蒸馏（prefix self-distillation）仅利用最终答案即可诱导短前缀与长任务间的对应关系。这些结果表明，联合训练中间预测与状态尺度是使递归深度扩展对更长问题有效的关键，且该方法适用于多种序列状态追踪任务，揭示了实现深度外推所需的任务结构。
链接: https://arxiv.org/abs/2610.27628
作者: Pan Wang,Yihao Hu,Guangyuan Dong,Yitong Lin,Hang Wang,Zimi Liu,Zhihao Wen,Jianshe Li,Jiang-Ming Yang,Wei Wu,Yongqi Tong
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Looped language models can run more iterations at test time, but additional iterations need not help them solve longer problems. For sequential state tracking, final-answer supervision does not specify how far each iteration should advance, and repeated updates can change the scale of the recurrent state. We introduce \textbfProgress-Aligned Recurrence (PAR), which trains the prediction at iteration k to match the state after the first k operations and rescales recurrent states to their first-iteration norms. On swap- S_5 , a Qwen2.5-0.5B model trained on at most 32 moves achieves 95.5% mean full-state accuracy at 512 moves, or 16\times the training length, compared with 55.7% reported for RLT and 0.8% for explicit chain of thought on the same backbone. Controlled S_5 ablations show that removing norm anchoring reduces the extrapolation factor from 16\times to below 2.2\times , and replacing progress-aligned targets with final-answer supervision reduces it to at most 8\times . A learned completion readout selects when to answer without a supplied stopping depth. When intermediate labels are unavailable, prefix self-distillation induces the correspondence from shorter problem prefixes using only final answers. The gains apply across sequential state-tracking tasks, establishing the task structure required for depth extrapolation. These findings support training intermediate predictions and state scale together to make additional recurrent depth useful for longer problems.
[NLP-67] ABench: Adversarily Benchmarking Large Language Models across Taxonomized Values EMNLP2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在全球化部署中与多元文化价值观对齐不足的问题，尤其针对现有对齐评估基准普遍以英语为中心、任务形式过于简单，且难以有效检验模型在特定文化语境下面对复杂对抗性诱导时的鲁棒性这一关键缺陷。其解决方案的核心在于提出ABench——一个针对中文语境下分层分类的价值观体系（基于158个衍生价值子类别构建的严谨层级分类体系）设计的细粒度人工构造评估基准，包含2,895道开放性问题，通过角色扮演、逻辑陷阱等策略性设计，精准探测模型在对抗性文化情境下的立场稳定性。此外，研究还引入一种基于问题特异性评分标准的加权自动化评估框架，实现对模型表现的精细化量化分析。实验结果表明，主流16个大模型在面对复杂价值观对抗攻击时普遍表现出显著脆弱性（得分普遍低于60），凸显了发展更具鲁棒性和文化适配性的对齐策略的紧迫性。该研究公开发布数据集，以推动面向分层价值观对齐的深入研究。
链接: https://arxiv.org/abs/2610.27611
作者: Liping Yang,Yongqi Leng,Jingting Zheng,Yuqi Ren,Shaolin Zhu,Deyi Xiong
机构: Tianjin University (天津大学); Yuanhui AI
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted to EMNLP 2026 (Main Conference). 16 pages, 6 figures, 7 tables
Abstract:Ensuring the alignment of Large Language Models (LLMs) with diverse cultural values is critical for their safe global deployment. However, existing alignment benchmarks are predominantly English-centric or confined to simple discriminative tasks. Moreover, they often fail to effectively evaluate model robustness against complex, adversarial inductions within specific cultural contexts. To address these limitations, we propose ABench, a fine-grained, handcrafted benchmark specifically tailored to taxonomized values in Chinese. Grounded in a rigorous hierarchical taxonomy comprising 158 derived value subcategories, ABench consists of 2,895 open-ended questions. These questions employ techniques such as role-playing and logical traps to probe the stability of models’ stances in adversarially induced contexts. Furthermore, we propose a weighted automated evaluation framework that utilizes question-specific scoring criteria for nuanced assessment. Extensive experiments on 16 mainstream LLMs reveal that most models exhibit significant vulnerabilities when subjected to sophisticated adversary attacks on values (scoring below 60). This underscores the need for more robust, culturally specific alignment strategies. Our data will be publicly released to promote further research on taxonomized value alignment.
[NLP-68] Auditing the Product Not the Recursion: The Evaluation Evidence Behind Recursive Self-Improvement Claims
【速读】： 该论文旨在解决生成式AI系统中递归自我改进（Recursive Self-Improvement, RSI）承诺的可验证性问题，即如何区分对最终系统性能的“产品型主张”与对改进过程本身的“过程型主张”。其核心挑战在于：尽管大量研究声称实现了RSI，但缺乏对改进过程可重复性和有效性的确凿证据。解决方案的关键在于建立一种“主张索引报告标准”（claim-indexed reporting standard），将每一项具体主张明确关联至能够验证该主张的评估设计。研究通过对114篇符合条件的研究进行全文编码分析发现，90%的RSI相关研究仅报告了超出开发任务的产品评估，且多数未充分分离训练与评估任务；仅有22%的研究进行了两次以上完整改进循环的审计，而真正能对比相同起始条件下的早期与后期改进程序的研究不足4项。此外，选择偏差、评估者共性错误及系统状态变化等因素限制了产品评估所能揭示的真实改进程度。因此，研究强调必须通过严谨的评估设计（如独立任务测试、多轮循环审计、跨版本公平比较）来支撑过程型主张，从而推动对RSI机制更可信的科学评价。
链接: https://arxiv.org/abs/2610.27610
作者: Zihan Chen,Di Zhu,Lei Nico Zheng
机构: Stevens Institute of Technology(史蒂文斯理工学院); University of Massachusetts Boston(马萨诸塞大学波士顿分校)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Software Engineering (cs.SE)
备注: 39 pages, 1 figure, 20 tables
Abstract:Recursive self-improvement (RSI) promises AI systems, such as self-evolving LLM agents, that become better at producing further improvements. Evaluating this promise requires separating product claims about the final system from process claims about the improvement process itself. Existing surveys specify the evidence these claims require, but how often studies report that evidence remains unmeasured. We examine this question through full-text coding of 114 eligible studies in a reproducible arXiv search frame frozen in October 2026, including 90 that describe their own systems as RSI. The codebook links claims to evaluation designs. Independent human coding, third-reader adjudication, and a separate blind coding of improver comparisons assess coding reliability. Among the 90 RSI studies, 90% report product evaluation beyond development tasks under a broad criterion that includes partial or unstated separation, while 60% report evaluation on tasks unused for selection. Only 22% audit two or more reruns of the full loop. In a human-coded random sample, the RSI studies show the same contrast: 45% evaluate on tasks unused for selection and 14% audit two or more reruns. Of the 26 studies claiming a better improvement procedure, 23 report a related comparison, but only 4 compare earlier and later procedures from identical starting artifacts at a matched budget. Across these results and a comparison set of established works, evidence for repeatability and improved procedures remains less developed than product evaluation. Three propositions explain how selection effects, shared evaluator errors, and changes in system state limit what product evaluations can establish. Together, the findings support a claim-indexed reporting standard that connects each claim to the evaluation design needed to test it.
[NLP-69] Self-Decomposing Diffusion for Code Generation
【速读】： 该论文旨在解决生成复杂程序时因组件间依赖关系（如接口一致性、数据流和共享假设）难以维护而导致的正确性与效率之间的矛盾问题。现有基于扩散语言模型（Diffusion Language Models, DLMs）的代码生成方法在追求并行解码效率时，易因过度并行化破坏组件间的依赖关系，进而影响程序正确性；而采用保守解码策略又会牺牲生成速度。其解决方案的关键在于引入一种名为自分解式代码生成（Self-Decomposing Diffusion for Code, SDDC）的新框架，通过在实现前先生成一个描述程序结构分解与单元间依赖关系的“合约”（contract），从而在扩散过程中实现对并行生成单元的协同调度。SDDC完全基于扩散机制完成程序分解与并行实现，显著提升了准确率-延迟权衡表现：在五个基准测试上，相较于基线模型，其pass@1最高提升31.1个百分点，且解码速度提高7.5至13.7倍，同时优于微调基线23.5个百分点且延迟更低。
链接: https://arxiv.org/abs/2610.27600
作者: Lize Shao,Zichen Xie,Zhizhen Chen,Zhe Zeng,Wenxi Wang
机构: University of Virginia(弗吉尼亚大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Diffusion language models (DLMs) are emerging as a promising paradigm for code generation, offering the potential for fast, parallel decoding. However, generating complex programs remains challenging, as their interdependent components must maintain consistent interfaces, data flows, and shared assumptions. Aggressive parallel decoding can compromise these dependencies and undermine program correctness, while conservative decoding sacrifices efficiency. To address this trade-off, we propose planning program decomposition before implementation to coordinate interdependent units during parallel generation. We introduce Self-Decomposing Diffusion for Code (SDDC), a novel framework that decomposes complex programs and implements their units in parallel, entirely through diffusion. SDDC first generates a contract specifying the program’s decomposition and inter-unit dependencies, then uses it to guide coordinated parallel implementation. Across five benchmarks and two 7B diffusion backbones, SDDC substantially improves the accuracy–latency trade-off, raising pass@1 by up to 31.1 percentage points while decoding 7.5–13.7 \times faster than the base models, and outperforming a fine-tuning baseline by up to 23.5 points with lower latency.
[NLP-70] WORTH Benchmark: Evidence Discovery under Not-Fully-Cooperative Information – A Two-Episode Pilot
【速读】： 该论文旨在解决现有推理评估中模型依赖预设前提、缺乏对模型自主发现证据能力及叙事信息呈现方式如何影响判断过程的考察这一核心问题。其关键解决方案在于构建WORTH基准测试，要求模型在未标注线索的原始悬疑叙事中自主识别关键线索句、判定案件相关人物，并以自然语言解释推理路径；叙事设计由专业作者完成，使有效证据与延迟真相揭示的叙事策略共存。通过在两个剧集的完整故事条件下开展70次正式实验（其中67次可评分），研究发现线索命中率与人物判断并非同步，例如在第02集37次可评分运行中，31次均错误地选择了同一组嫌疑人，且倾向于选择指向怀疑的句子（25次）而几乎不选择能排除嫌疑的句子（0次）。这一结果凸显了分组件分析的诊断价值，表明非完全协作的信息关系可能显著干扰模型的证据选取与最终判断，揭示了当前模型在处理复杂叙事推理任务时存在的系统性偏差。
链接: https://arxiv.org/abs/2610.27597
作者: Bo Zou,Chao Xu
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 31 pages. Pilot study, two episodes
Abstract:Existing reasoning evaluations typically provide task-relevant information to models as available premises, rarely examining concurrently how models discover evidence autonomously and how information presentation within narratives guides or obscures judgment. This paper proposes the WORTH Benchmark: models read original mystery narratives without clue annotations, autonomously select key clue sentences, identify case participants, and explain their reasoning process. The narrative materials are designed by professional authors such that evidence facilitating reasoning coexists with narrative arrangements that delay the revelation of truth. We report a pilot experiment under full-story conditions across two episodes, conducting a total of 70 formal runs on models, of which 67 runs are content-scorable. Scoring separately records hits on author-designated key clue sentences and participant selections; natural language explanations are used for case diagnosis and are not scored independently. The results show that clue hits and participant judgments are not synchronized; across 37 content-scorable runs in Episode 02, 31 runs simultaneously selected the same set of incorrect participants, while adjacent suspicion-oriented and exoneration-oriented sentences were selected 25 times and 0 times, respectively. These phenomena illustrate the diagnostic value of component-wise analysis and suggest that not-fully-cooperative informational relationships may influence evidence selection and judgment. This study only implements full-story conditions for two episodes, and does not establish stable model rankings or infer general mechanisms on this basis; human player records have not yet been cleaned, and this version does not report human-model comparisons.
[NLP-71] Benefit and Cost of Random-Token Replacement under Repeated Data: A Known-Entropy Decomposition and Evidence from Ten Real Corpora
【速读】： 该论文旨在解决在预训练数据稀缺情况下，模型因反复接触相同标记而产生记忆（memorization）现象所导致的过拟合问题。其核心解决方案是采用随机标记替换（random-token replacement），一种输入扰动增强方法，以降低验证损失。研究的关键在于提出了一种将验证损失分解为近似误差（approximation）、估计误差（estimation）和记忆误差（memorization）三部分的理论框架，揭示了随机替换通过使预测值向语料库平均值收缩（shrinkage），从而减少记忆误差的机制。研究表明，尽管替换会引入偏差（bias cost），但其收益依赖于对记忆误差的抑制；当模型达到最低验证损失后，记忆误差会随训练继续上升，而替换可有效抑制这一上升趋势约60%。在固定预算下，随着控制源可预测性下降，替换由有害转为有益；而在最低损失点，替换仅在无预测性条件下有效。在韩语与日语网络文本上的实验表明，替换能消除损失上升约三分之一，但在最低损失点效果较弱。在十组真实语料中，替换在九组中降低了最终损失，仅在冗余度极高的程序生成JSON（census corpus）上造成负面影响，该数据集的冗余程度远高于其他64个文本与代码语料，凸显了替换对高冗余数据的敏感性。
链接: https://arxiv.org/abs/2610.27590
作者: Wonsup Shin,Jingu Kim
机构: Flitto
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 16 pages, 6 figures. Code: this https URL
Abstract:When unique pretraining data is scarce, models see the same tokens many times and are thought to memorize. Random-token replacement, an input-corruption augmentation, is one remedy reported to lower validation loss, but a lower loss alone does not reveal which source of error it reduces. On a controlled source whose entropy rate is known in closed form, we show that replacement shrinks predictions toward the corpus average. We also introduce a decomposition that separates validation loss into approximation, estimation, and memorization terms. This shrinkage view predicts that replacement’s bias cost needs no memorization but its benefit does. Memorization raises validation loss after its minimum, so we report the effect at two points, the minimum validation loss and the end of a fixed budget. Where replacement helps, it reduces this rise by about 60%. At the end of the budget, it turns from harmful to helpful as the controlled source becomes less predictable; at the minimum, it helps at no predictability level under the fixed recipe. On Korean and Japanese web text, replacement removes about a third of the loss rise after the minimum by the end of the budget, but its effect at the minimum is much smaller. Across ten real corpora, replacement lowers the final loss on nine. It harms only the census corpus, program-generated JSON that is the most redundant of the ten and much more redundant than any of 64 further prose and code corpora we measure.
[NLP-72] Beyond Model Weights: Sampling-Dependent Fairness in Diffusion Language Models
【速读】： 该论文旨在解决生成式语言模型（Generative Language Models, GLMs）中扩散语言模型（Diffusion Language Models, DLMs）的公平性问题，特别是其社会偏见在不同采样配置下的稳定性。传统自回归语言模型（Autoregressive Language Models, ARLMs）的公平性评估通常基于固定的生成过程，而DLMs引入了如块长度（block length）和重掩码策略（remasking strategy）等采样控制机制，这些机制直接影响生成速度与质量之间的权衡，但尚未被充分研究其对社会偏见的影响。研究发现，尽管所有测试的DLMs（LLaDA、LLaDA-1.5及Dream on Open-BBQ）均存在社会偏见，但偏见的量级以及模型间偏见排序高度依赖于具体的采样配置。这意味着即使采用针对ARLMs有效的去偏方法，虽能降低偏见且不损害下游性能，但偏见对采样设置的敏感性依然存在。因此，该研究的关键结论是：扩散语言模型的公平性无法脱离具体的采样配置进行独立表征，即便经过去偏处理亦然。
链接: https://arxiv.org/abs/2610.27584
作者: Koshiro Saito,Diasuke Oba,Danushka Bollegala,Naoaki Okazaki
机构: Institute of Science Tokyo(东京科学研究所); University of Liverpool(利物浦大学); Amazon(亚马逊); AIST(日本国立信息学研究所); NII LLMC(日本国立情报学研究所语言模型研究中心)
类目: Computation and Language (cs.CL)
备注: 31 pages, 11 figures, 17 tables. Project page: this https URL
Abstract:Diffusion language models (DLMs) are emerging as a practical alternative to autoregressive language models (ARLMs), yet their fairness remains largely uncharacterized. DLMs introduce sampling controls with no direct autoregressive counterpart, such as block length and remasking strategy, that directly set their speed–quality operating point. This raises a basic question: does a DLM’s measured social bias remain stable across these settings? Across LLaDA, LLaDA-1.5, and Dream on Open-BBQ, we find that all three models exhibit social bias, but both the magnitude of measured bias and the relative ordering of models by bias are highly sensitive to the sampling configuration. This sampling dependence also changes what successful debiasing means. While mitigation strategies established for ARLMs transfer effectively to DLMs and reduce bias without degrading downstream performance, sampling dependence persists even after mitigation. Our results show that fairness in DLMs cannot be characterized independently of the sampling configuration, even after debiasing. Project page is available at this https URL.
[NLP-73] CoP: Accelerating Diffusion Large Language Model Decoding via Tentative Context Pipelining
【速读】： 该论文旨在解决扩散型大语言模型（diffusion large language models, dLLMs）在生成过程中因块级解码器串行化与重复计算稳定预测所导致的效率低下问题。其核心挑战在于：将预测结果的“可承诺性”与“作为上下文的实用性”耦合，造成大量冗余计算。解决方案的关键是提出无需训练的暂定上下文流水线（Tentative Context Pipelining, TeCoP），通过分块并行流水线与“草稿-验证”机制，解耦预测的承诺状态与上下文利用角色。TeCoP将已稳定但尚未承诺的预测作为临时跨块上下文复用，仅在置信度与稳定性检验通过后才予以确认。实验表明，在GSM8K、Minerva MATH和MBPP等基准上，TeCoP可将前向评估次数减少高达87%，吞吐量提升达7.50倍（无KV缓存）和11.77倍（有缓存），显著优于Fast-dLLM与DSB，且保持与原始解码器相当的平均缓存精度（误差小于1个百分点）。
链接: https://arxiv.org/abs/2610.27583
作者: Jui-Chien Tsou,Hsuan-Chi Liu,Wei-Chin Wang,Yu-Syuan Xu,Chun-Yi Lee
机构: National Taiwan University (国立台湾大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Diffusion large language models (dLLMs) predict multiple token positions per forward pass, yet block-wise decoders serialize generation and repeatedly recompute stable predictions. We attribute this inefficiency to coupling a prediction’s readiness for irreversible commitment with its utility as context. We introduce Tentative Context Pipelining (TeCoP), a training-free decoder that decouples these roles through staggered block-parallel pipelining and draft-and-verify. TeCoP reuses stable but uncommitted predictions as provisional cross-block context, with commitment governed by confidence and stability checks. Across four LLaDA and Dream dLLMs on GSM8K, Minerva MATH, and MBPP, TeCoP reduces number of forward evaluations by up to 87% and achieves up to 7.50-fold throughput without KV caching and 11.77-fold with caching over the uncached original decoder. It outperforms Fast-dLLM and DSB in throughput across all 12 model-benchmark settings, with and without caching, while keeping average cached accuracy within one percentage point of the original decoder.
[NLP-74] Why Static Probes Fail: Context Determines Deception Feature Composition
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在日益增强的自主性背景下可能出现的欺骗行为检测难题，尤其是针对模型在表面输出看似无害但内部存在偏离对齐目标的意图（如蓄意谋划或故意降低表现以规避审查）的情况。传统基于输出监控的方法因无法触及模型内部状态而效果有限，而现有机制解释方法（如线性探测）通常假设欺骗行为对应于表征空间中的单一方向，这一假设在多样化欺骗场景中缺乏泛化能力。为此，论文提出将欺骗行为建模为多维子空间而非单一直线方向，通过迭代正交分解识别出能够捕捉欺骗意图不同维度特征的互补方向。进一步地，为克服静态组合策略在不同场景下因领域依赖性导致的脆弱性，研究引入一种领域自适应门控机制，在测试时仅需少量样本（10个）即可实现少样本适配，动态融合提取的方向。实验结果表明，该方法在所有评估数据集上均达到平均AUROC 0.97、5%假阳性率（FPR）下94%的召回率，显著优于现有方法。其核心突破在于构建了兼具可解释性与强泛化能力的多维欺骗检测框架。
链接: https://arxiv.org/abs/2610.27577
作者: Yao Huang,Siyuan Chen,Zhenghan Zhou,Yinpeng Dong
机构: Tsinghua University (清华大学); Shanghai Qi Zhi Institute (上海奇智人工智能研究院)
类目: Computation and Language (cs.CL)
备注: 11 pages, 5 figures
Abstract:With increasing autonomy, Large Language Models (LLMs) may engage in deceptive behaviors, such as scheming to pursue misaligned goals or sandbagging to escape oversight. Monitoring outputs alone is insufficient, as models can produce seemingly benign responses while harboring misaligned internal objectives. While mechanistic approaches like linear probing are promising, they rely on the assumption that deception corresponds to a single direction in representation space. We demonstrate that such single-direction probes fail to generalize across diverse deception scenarios. To bridge this gap, we propose representing deception as a multi-dimensional subspace rather than a single direction. Our method utilizes iterative orthogonal decomposition to identify complementary directions that capture distinct facets of deceptive intent. To effectively leverage these directions for detection, we find that static combinations prove brittle due to strong domain dependence across different scenarios. We therefore introduce a domain-adaptive gating mechanism that dynamically composes the extracted directions through few-shot adaptation at test time. With only 10 examples, our approach achieves an average AUROC of 0.97 and 94% recall at 5% FPR across all evaluated datasets, substantially outperforming existing methods.
[NLP-75] AudiPsy: Evidence-Grounded Multimodal Reasoning over Spoken Counseling Conversations for Depression Assessment
【速读】： 该论文旨在解决从心理咨询对话中进行抑郁症评估时，现有端到端分类方法缺乏对多模态证据（文本与声学特征）的显式解释及上下文感知推理的问题。其核心解决方案是提出AudiPsy——一种基于证据驱动的全模态大语言模型（omni-LLM），采用观察-评估-摘要（Observation-Assessment-Summary, OAS）范式，系统性地整合对话转录文本、文本与声学情绪特征以及文本-声学情绪不一致信息，实现结构化推理。OAS框架通过分步组织多模态观测、进行上下文感知评估，并动态维护对话摘要以支持后续轮次的连贯理解。为支撑该框架，研究构建了AudiPsy-Counsel合成语料库，引入多模态上下文增强，并设计统一的三阶段课程学习策略，以优化面向抑郁症评估的推理能力。实验在DAIC-WOZ数据集上验证了AudiPsy在抑郁识别性能和推理质量上均优于专用检测方法及通用语音/全模态大模型，且在自然对话数据集EATD-Corpus上的补充评估进一步证明了其在真实场景下的适用性。
链接: https://arxiv.org/abs/2610.27567
作者: Chuanbo Zhu,Run Luo,Weibo Zhao,Yue Li,Kun Qian,Qiang Qu,Yike Guo,Minghuan Tan,Wei Xue
机构: The Hong Kong University of Science and Technology(香港科技大学); National University of Singapore(新加坡国立大学); University of the Chinese Academy of Sciences(中国科学院大学); Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences(中国科学院深圳先进技术研究院); Beijing Institute of Technology(北京理工大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Depression assessment from counseling speech requires integrating textual and acoustic psychological evidence across turns, yet existing end-to-end classification approaches lack explicit multimodal evidence interpretation and context-aware reasoning. We propose AudiPsy, an evidence-grounded omni-LLM for depression assessment over spoken counseling conversations. AudiPsy employs an Observation-Assessment-Summary (OAS) paradigm to integrate dialogue transcripts, textual and acoustic emotions, and textual-acoustic emotion incongruence into structured reasoning. OAS incrementally organizes multimodal observations, performs contextualized assessment, and maintains an evolving dialogue summary for subsequent turns. To support this framework, we construct AudiPsy-Counsel, a synthetic spoken counseling conversation corpus with multimodal context augmentation, and develop a unified three-stage curriculum for depression-oriented reasoning. Experiments on DAIC-WOZ demonstrate superior depression assessment performance over specialized depression detection methods and general-purpose speech/omni-LLMs, along with improved reasoning quality, while supplementary evaluation on EATD-Corpus supports AudiPsy’s applicability to naturalistic conversations.
[NLP-76] Learned Readout Tokens for Extracting Large Language Model Confidence
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在实际应用中输出可信度评估的问题，即如何准确估计模型单个输出的置信度。现有方法通常通过在模型已有词汇表中的标记（token）的隐藏状态上训练探测器（probe）来预测输出正确性，但受限于固定读取位置，难以充分捕捉模型内部动态。本文提出一种创新的置信度探测方法——ScratchProbe，其核心在于学习一个全新的“读出标记”（readout token），并利用该标记的隐藏状态进行探测器训练。这一方法突破了仅依赖预定义词汇表内标记的限制，显著提升了探测性能，在多个数据集和不同规模的LLM上均优于传统方法；进一步地，作者还提出了多标记扩展版本，实现更优表现。由于该方法仅改变探测器的读取位置，无需修改原有探测器结构，可无缝集成至现有探测系统中，应用于当前最先进的探测器后，在所有基线模型上均实现了域内与域外任务下AUROC指标的提升。
链接: https://arxiv.org/abs/2610.27566
作者: Nima Seifi,Martin Ester
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 12 pages, 4 figures, including technical supplement
Abstract:Large language models (LLMs) are increasingly used within everyday applications where their responses have significant downstream impacts. For responsible deployment of such systems, we desire insight into how much we can trust their individual outputs. A popular approach for estimating LLM confidence is to train probes on the hidden states of the LLM’s tokens to predict the correctness of the model’s output. These probes typically read hidden states from tokens that are already in the model’s vocabulary. In this work, we introduce a new approach to confidence probes that learn a new readout token and utilize its hidden states to train the probe. We show that our learned readout approach ScratchProbe outperforms its existing-token counterpart on a variety of datasets and LLMs. We also extend our method to a multi-token approach which we show provides further performance improvements. Finally, since the learned readout changes only the position a probe reads from, any existing probe can adopt it. Applied to a state-of-the-art probe, it improves AUROC on every backbone, in domain and out of domain.
[NLP-77] AnomalyJev: Towards Evaluating Image Anomaly Explanations via Claim-Level Verification
【速读】： 该论文旨在解决基于多模态大语言模型（Multimodal Large Language Model, MLLM）的异常检测（Anomaly Detection, AD）方法中，对其生成的文本解释进行可靠且高效评估的难题。现有方法通常依赖大语言模型（LLM）裁判通过结构化评分标准进行评估，但此类方法需大量文本生成，导致评估成本高昂且难以扩展。为此，本文提出一种创新性解决方案——AnomalyJev，其核心在于将评估锚定于已有的真实标签（ground truth），通过从解释中提取可验证的陈述（verifiable claims），并利用基于真实标签定义的查询与指针式头部（pointer-based heads）从上下文表示中直接抽取这些陈述，无需生成额外文本。随后，将提取的陈述与真实标注进行比对，以评估解释的忠实性（faithfulness）与完整性（completeness）。为支持评估，研究构建了一个包含640条解释、超过1,600次专家评分的人工标注数据集，覆盖域内与跨数据集场景。实验表明，尽管AnomalyJev设计轻量，其与人工标注者的一致性显著优于多数开源MLLM裁判，接近顶尖MLLM裁判的表现，同时仅需极低的运行时间和资源开销。
链接: https://arxiv.org/abs/2610.27565
作者: Qizhou Wang,Jiawen Zhu,Guansong Pang,Christopher Leckie
机构: The University of Melbourne(墨尔本大学); Singapore Management University(新加坡管理大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Recent multimodal LLM (MLLM)-based anomaly detection (AD) methods can generate textual explanations that help interpret their predictions. However, reliably and efficiently evaluating these explanations remains challenging. While LLM judges can assess explanations through structured rubrics, they often require substantial token generation, making evaluation costly and difficult to scale. We instead propose to anchor evaluation in the available ground truth by extracting verifiable claims from each explanation and verifying them against the corresponding annotations. We instantiate this idea as AnomalyJev, a lightweight framework that extracts verifiable claims from contextualised representations using ground-truth-defined queries and pointer based heads, without token generation. The claims are then verified against the ground truth to score faithfulness and completeness. For evaluation, we construct a human-annotated set of 640 explanations with over 1,600 expert ratings across in-domain and cross-dataset settings. Despite its compact design, AnomalyJev achieves stronger agreement with human annotators than most open MLLM judges and approaches that of flagship MLLM judges, at a fraction of their runtime and resources.
[NLP-78] Exploring Recurrent-Depth Language Models for Multitask Clinical Text Processing
【速读】： 该论文旨在解决临床自然语言处理（Clinical Natural Language Processing, CNLP）中多任务模型统一部署与知识迁移效率低的问题，即如何构建一个通用且高效的模型以同时胜任信息抽取、分类、推理和问答等多种临床文本任务。其解决方案的关键在于探索小规模循环深度（recurrent-depth）语言模型在联合微调（joint fine-tuning）框架下的多任务表现能力，并对比其与标准Transformer架构模型的性能差异。研究结果表明，尽管参数量较小，但Ouro系列循环深度模型在多数任务上显著优于同级别Transformer模型（如Qwen3-1.7B），并在推理与问答任务中展现出更优的性能提升，显示出循环计算结构在促进推理能力、知识获取与泛化方面的潜力，为未来通过受控实验深入探究模型架构对临床任务中认知过程的影响提供了重要依据。
链接: https://arxiv.org/abs/2610.27559
作者: Xi Yang,Guodong Liu,Cheng Peng,Kaleb Smith,Yingying Zhang,Xing He,Yi Guo,Jiang Bian,Haoteng Tang,Yonghui Wu
机构: 未知
类目: Computation and Language (cs.CL)
备注: 8
Abstract:Clinical natural language processing includes diverse tasks such as information extraction, classification, and question answering. Developing one model for all tasks could simplify deployment and improve knowledge transfer. We evaluated whether small recurrent-depth language models can meet this need through joint fine-tuning and compared their performance with standard transformers. We fine-tuned Ouro1.4B-Thinking, Ouro-2.6B-Thinking, HRM-Text-1B, Qwen3-1.7B, and Qwen3-8B on 9 datasets covering information extraction, classification, inference, and question answering, and evaluated them on four additional datasets excluded from training. Across these 13 datasets, Ouro-1.4B-Thinking achieved higher performances than Qwen3-1.7B on 11 datasets; Ouro-2.6BThinking was competitive with Qwen3-8B on most tasks. Ouro models showed the largest gains on inference and question answering, whereas HRM-Text-1B showed mixed results. These findings support further study of recurrent-depth models for clinical tasks through controlled experiments to isolate architectural effects and investigate how recurrent computation influences reasoning, knowledge acquisition, and generalization.
[NLP-79] Can Native Safety Travel? Reusing Safety Pretraining Across Models via On-Policy Distillation
【速读】： 该论文旨在解决安全预训练模型中“原生安全”（native safety）难以复用的问题，即当前具备强安全性的模型通常仅限于其自身架构且难以迁移至其他模型，同时存在通用能力下降的缺陷。其核心解决方案是首次提出并验证“原生安全重用”问题，通过一种基于策略内知识蒸馏（on-policy distillation）的框架，将小型、专用的安全预训练“教师”模型所具备的原生安全特性，高效迁移至标准预训练的“学生”模型中。该方法利用教师模型对学生的生成轨迹进行跨标记化器对齐的逐标记监督，实现安全特性的有效传递。实验表明，经此方式迁移的学生模型在后续微调中展现出显著更高的安全鲁棒性（平均攻击成功率提升仅5.00%，相较后置安全训练的11.58%），同时保持了远高于教师模型的通用性（平均效用48.20%对比教师的27.42%）。该研究突破了传统大模型向小模型的知识蒸馏范式，实现了小模型到大模型的安全能力选择性迁移，证明原生安全是一种可复用的模型属性，而非绑定于特定预训练模型的固有特征。
链接: https://arxiv.org/abs/2610.27555
作者: Bingqi Shang,Yang Zhang,Yiwei Chen,Changsheng Wang,Shiyu Chang,Bingquan Shen,Sijia Liu
机构: Michigan State University(密歇根州立大学); MIT-IBM Watson AI Lab; University of California, Santa Barbara(加州大学圣塔芭芭拉分校); National University of Singapore(新加坡国立大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Safety pretraining has emerged as a promising paradigm for embedding safety deeply into a model during pretraining, yielding robustness to subsequent fine-tuning that conventional safety post-training often lacks. We refer to this safety acquired during model pretraining as native safety. Yet native safety is expensive to acquire, confined to the model in which it was pretrained, and often accompanied by lower general utility. In this work, we rethink the practical value of safety-pretrained models and introduce, for the first time, the native safety reuse problem: Can native safety be transferred from a specialized, potentially smaller, safety-pretrained “teacher” to standard-pretrained “students,” enabling them to inherit its robustness while preserving their general utility? We show that native safety can “travel” through on-policy distillation. A standard-pretrained student generates responses to safety-pretraining prompts, while a small safety-pretrained teacher supervises the student’s token-level predictions along its own generated trajectories; cross-tokenizer alignment further enables transfer across model families. The resulting students exhibit substantially greater safety robustness than their safety-post-trained counterparts while retaining higher general utility than the teacher: subsequent fine-tuning raises the average attack success rate (ASR) of the 1.7B student by only 5.00%, compared with 11.58% for safety post-training, and its average utility of 48.20% far exceeds the aligned teacher’s 27.42%. Unlike conventional large-to-small knowledge distillation, our framework enables selective, small-to-large safety transfer: it reuses the robust safety expertise of a smaller pretrained teacher without inheriting its utility limitations. These findings establish native safety as a reusable model property rather than one permanently tied to the originally pretrained model.
[NLP-80] EviDial: When Correct Depression Predictions Lack Evidence
【速读】： 该论文旨在解决当前抑郁症对话评估基准中普遍存在的“标签一致但证据不足”的问题，即模型预测与临床评分一致，却可能缺乏足够的对话文本支持。其核心挑战在于：现有评估方法仅依赖标签匹配度，无法区分预测是否真正基于对话中的充分证据，从而导致对生成式模型（Generative AI）性能的高估。解决方案的关键在于提出EviDial基准，通过引入“证据充分性”（evidence sufficiency）这一新维度，将症状评估划分为“不存在”“部分”“冲突”或“充分”四类证据状态，并利用广义可推广性理论（generalizability theory）、多面拉斯克测量（many-facet Rasch measurement）和随机效应潜在类别分析（random-effects latent class analysis）从多方判断中构建校准后的证据参考标准。研究发现，82.8%的患者-症状单元缺乏充分的文本证据支撑，且多数模型预测虽在标签层面正确，但实际无足够对话依据（占比69.2%-96.9%），揭示出大语言模型（LLM）评估中评价者间一致性未必反映独立证据判断，凸显了校准自动化参考面板的必要性。EviDial通过分离“标签一致性”与“证据根基性”两个概念，为更真实、可审计的生成式人工智能在抑郁症评估中的应用提供了新的测量框架。
链接: https://arxiv.org/abs/2610.27541
作者: Junjie Xu,Jin Zhu,Yuzhe Yang,Luhan Wang,Xingjiao Wu,Daoguo Dong,Jing Yang,Wenxin Wu,Liang He
机构: East China Normal University (华东师范大学); Fudan University (复旦大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Depression dialogue benchmarks commonly assess models by comparing predicted symptom labels with clinical scale scores. However, a prediction may agree with a benchmark label even when the dialogue lacks enough evidence to justify it. We introduce EviDial, a benchmark that distinguishes label agreement from evidence-grounded assessment. EviDial comprises 382 depression dialogues from three sources and 2,744 participant-symptom cells, each classified as having absent, partial, conflicting, or sufficient evidence. Because evidence sufficiency lacks a directly observable gold standard, we derive calibrated evidence references from multiple judgments using generalizability theory, many-facet Rasch measurement, and random-effects latent class analysis. Our audit reveals that 82.8% of participant-symptom cells lack sufficient textual evidence for the corresponding symptom assessment, with substantial variation across sources. Across model families and prompting strategies, 69.2%-96.9% of predictions judged correct by benchmark labels are nonetheless unsupported by sufficient dialogue evidence. Further analyses show that strong agreement among LLM evaluators need not reflect independent evidence judgments, underscoring the need to calibrate automated reference panels. These findings demonstrate that label-based evaluation can overestimate evidence-grounded performance by conflating agreement with clinical labels and support in the observed dialogue. EviDial provides a measurement layer for separating these constructs and auditing LLM-based depression assessment more faithfully.
[NLP-81] Judging the Justification: Decision Certificate Supervision for Generative Reward Models
【速读】： 该论文旨在解决生成式奖励模型（Generative Reward Models, GenRMs）在生成自然语言解释以支持偏好判断时，存在“正确结论不等于合理论证”的问题。现有方法依赖参考解释进行监督，但获取高质量参考解释成本高昂，且无法保证生成的解释真正支撑其对应结论。为此，本文提出DeCert框架，通过无需参考解释的方式评估生成解释的有效性与结论一致性，并将此评估转化为兼顾结论正确性与解释质量的偏好监督信号。其关键创新在于：当多个采样判断结果具有相同正确结论时，利用解释质量差异构建解释偏好对，从而区分强弱论证。实证研究表明，DeCert在两类模型家族及多种分布内外基准上均显著优于现有生成判别基线，同时提升解释质量并有效缓解位置偏差，最高使位置一致准确率提升14.8个百分点。该工作强调了在训练过程中不仅需关注结论正确性，更应监督其论证合理性的重要性。
链接: https://arxiv.org/abs/2610.27528
作者: Guiquan Sun,Bing Bai,Martin Renqiang Min,Dongjin Song
机构: University of Connecticut (康涅狄格大学); NEC Labs America (NEC美国实验室)
类目: Computation and Language (cs.CL)
备注: Under Review
Abstract:Generative reward models (GenRMs) generate natural language rationales to justify preference judgments, yet a correct verdict does not guarantee a sound justification. Existing approaches often rely on reference explanations to supervise rationales, but obtaining such references can be costly and does not necessarily establish whether generated rationales support the resulting verdict. To address these limitations, we propose DeCert, a framework that evaluates the validity and verdict consistency of generated rationales without requiring reference explanations. Crucially, DeCert translates these assessments into preference supervision that accounts for both verdict correctness and justification quality. In particular, when multiple sampled judgments share the same correct verdict, differences in certificate quality enable DeCert to construct rationale preference pairs that distinguish stronger from weaker justifications. Extensive empirical studies across two model families and multiple in- and out-of-distribution benchmarks demonstrate that DeCert consistently outperforms strong generative judge baselines while improving rationale quality. Moreover, DeCert effectively mitigates positional bias, improving position-consistent accuracy by up to 14.8 points. These results highlight the importance of supervising not only which verdict is correct, but also how that verdict is justified.
[NLP-82] Self-Evolving Agents through Privileged Retrospective Distillation
【速读】： 该论文旨在解决自演化智能体（self-evolving agents）在缺乏真实解或验证器生成奖励的情况下，如何从自然收集的轨迹中持续提升性能的问题。其核心挑战在于：传统强化学习依赖于明确的奖励信号，而实际场景中的轨迹往往缺少可信赖的监督信号，导致模型难以有效学习与改进。解决方案的关键是提出一种多轮次的回溯式自蒸馏（Retrospective Self-Distillation, RSD）框架，通过利用任务间先验知识、回合级反思与未来回溯信息等特权回溯上下文（privileged retrospective context），使同一模型在不同角色下同时充当“教师”与“学生”。具体而言，教师角色基于已完成轨迹中的全局信息重新评估学生所采取的动作，生成更密集且具指导性的策略分布，并通过在线策略自蒸馏将其反馈至策略网络。随着迭代进行，经验库不断积累，回溯上下文动态更新，实现模型、数据与监督信号的协同演化。进一步引入前沿引导的上下文-策略演化机制（frontier-guided context-policy evolution），由更强的前沿模型对特权上下文的构建方式进行约束性优化，从而显著提升学习效率。实验在ALFWorld、ScienceWorld和SearchQA等多个基准上使用Qwen3-8B和Qwen3-14B模型验证，RSD相较基线模型性能提升高达59.7%，而前沿引导机制进一步带来44.7%的相对增益，证明了仅依赖自然积累的经验即可实现无需真实轨迹或验证器奖励的持久模型能力进化。
链接: https://arxiv.org/abs/2610.27495
作者: Yuhang Zhou,Yifan Wu,Bo Peng,Mingyi Wang,Jiayi Liu,Xiangjun Fan,Lizhu Zhang,Zhuokai Zhao
机构: Meta AI
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 26 pages, 5 figures
Abstract:Self-evolving agents aim to improve from their own experience rather than remain fixed after deployment, but naturally collected trajectories often lack ground-truth solutions or verifier-derived rewards. We introduce Retrospective Self-Distillation (RSD), a multi-round framework that turns such experience into dense supervision using privileged retrospective context. For each on-policy trajectory, the same underlying model acts as both student and teacher: the teacher re-scores the student’s realized actions while conditioning on cross-task precedents and lessons, together with turn-level reflection and future hindsight from the completed trajectory. The resulting teacher distribution is distilled back into the policy through on-policy self-distillation. Across rounds, the experience bank grows and the retrospective context is refreshed, allowing the model, its data, and its supervision to co-evolve. We further introduce frontier-guided context-policy evolution, where a stronger frontier model proposes constrained edits to how privileged context is constructed. Across ALFWorld, ScienceWorld, and SearchQA with Qwen3-8B and Qwen3-14B, RSD improves performance by up to 59.7% relative to the base model, while frontier-guided context-policy evolution adds up to a further 44.7% relative improvement over a fixed context policy. These results show that naturally accumulated experience can be converted into persistent model capability without ground-truth trajectories or verifier-derived rewards for the core policy update.
[NLP-83] Recoverability and Verifier Strength: Two Measured Preconditions for Unattended Agent Self-Improvement NEURIPS2026
【速读】： 该论文旨在解决自动化代理平台在无需人工干预的情况下持续优化自身代理时所面临的安全性与可靠性问题，核心挑战在于如何在不引入不可控风险的前提下实现完全自动化的代理演进。其解决方案的关键在于提出并量化两个核心指标：验证器强度（Verifier strength, σ） 与可恢复性（Recoverability）。验证器强度衡量当前使用的自动化验证器（如大语言模型判别器）与可信参考验证器之间的一致性程度，用以限制单步改进的可信边界；可恢复性则评估错误变更被撤销的成本与效果，即是否可通过无需重新验证的回滚操作恢复旧行为，以及回滚能否还原原有功能表现。研究指出，只有当这两个指标均达到足够水平时，变更才可安全地在无人工参与下执行。通过在自研平台上的实证测试发现，现有大语言模型判别器存在“恒定评分”等失效模式，且部分回滚机制无法有效恢复原始行为，导致任务结果偏差达17%以上。因此，论文提出的核心原则是：仅当变更历史版本被保留且验证器强度经过实测评估时，方可实现无监督自动化更新，从而在效率与安全性之间取得平衡。
链接: https://arxiv.org/abs/2610.27465
作者: De Wang,Peng Qi
机构: Uniphore
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Software Engineering (cs.SE)
备注: Accepted at the NeurIPS 2026 Workshop on AgenticOS. 6 pages + appendix
Abstract:Agent platforms are starting to improve their own agents: they read a deployed agent’s execution traces, derive a change, have an automated verifier approve it, and release it with no human author. How much of this loop can safely run without a human? We propose two quantities a platform can measure. Verifier strength \sigma is how well the available verifier – often an LLM judge – agrees with a trusted reference verifier; it limits how much a step can improve the agent. Recoverability is how cheaply a bad change can be undone: whether undoing it needs a new verifier approval, and how much of the old behavior a rollback restores. A step should run unattended only if both are adequate. On our platform this rule allows prompt, tool and skill edits to run unattended and keeps weight updates under human sign-off. Our main contribution is these two definitions and how to measure them. We built a platform that applies the rule and tested it in two ways: on the platform’s own approval gate, and in controlled experiments on the \tau^2 -bench agent benchmark. On verifier strength, the platform’s LLM judge sometimes gave the same score no matter what it was shown (9 of 37 settings) and reported no error, so a verifier can look healthy while approving anything; its strength must be measured, and two simple tests catch the failure. On recoverability, restoring a saved copy brought the agent back to its original behavior, but a regenerated undo did not, so only changes whose old version is kept can be reversed without a human; and while a change was deployed, it altered the outcome of 17 percent of tasks beyond normal run-to-run variation, which no undo repairs. Together these support the rule: a change can run unattended only when its old version is kept and its verifier’s strength has been measured. Comments: Accepted at the NeurIPS 2026 Workshop on AgenticOS. 6 pages + appendix Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Software Engineering (cs.SE) Cite as: arXiv:2610.27465 [cs.AI] (or arXiv:2610.27465v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2610.27465 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-84] Detecting Errors in Large Language Model Responses: The Incremental Predictive Value of Hidden States
【速读】： 该论文旨在解决大语言模型在生成回答时存在的幻觉（hallucination）问题，该问题严重制约了用户对模型输出的信任。尽管模型的内部状态（hidden states）被认为可能蕴含更丰富的幻觉检测信号，但现有研究多仅评估其独立检测性能，而输出端的对比方法常采用建模能力较弱的基线，导致结果不可靠。本文的关键创新在于构建了一个整合推理时可获取的输出端信息的统一基线，并在相同数据划分与评估规则下，对比加入隐藏状态前后的检测性能。实验覆盖开放式生成与约束型多选任务，涉及多种模型。结果表明，引入隐藏状态后，两类任务的AUROC分别提升约0.004和0.039，证明隐藏状态确实提供了额外的预测信息；然而，即使隐藏状态具备较强的独立检测能力，其带来的增益仍有限，说明单纯依赖内部状态难以显著提升整体检测性能。
链接: https://arxiv.org/abs/2610.27455
作者: Ningzhi Li,Xiaojun Ma,Xiaofei Wang
机构: Beijing Institute of Computer Technology and Application(北京计算机技术与应用研究所); Beijing, China(北京，中国)
类目: Computation and Language (cs.CL)
备注: 21 pages. Code: this https URL
Abstract:Although large language models support a wide range of tasks, hallucinations remain a major obstacle to user trust. When a model generates an answer, both its internal processing and final output vary with the question. Internal states are often expected to offer richer signals for hallucination detection. However, studies using these states typically report their standalone discrimination performance, while output-side comparisons often use simple baselines with weaker modeling capacity. We integrate output-side information available at inference time into a combined baseline and compare detection performance before and after adding hidden states under identical data splits and evaluation rules. Experiments cover open-ended generation and constrained multiple-choice tasks across different models. Adding hidden states improves AUROC by approximately 0.004 and 0.039 on the two tasks, respectively. Hidden states thus provide additional predictive information, but strong standalone detection performance can still yield only limited additional gains.
[NLP-85] Ask to Distill: Learning to Synthesize Data for On-Policy Distillation
【速读】： 该论文旨在解决生成式人工智能（Generative AI）中基于策略的蒸馏（On-policy Distillation, OPD）过程中，如何高效生成高质量、具有分布知识传递能力的合成问题（synthetic questions），而无需依赖正确答案的问题。核心挑战在于：在缺乏真值答案的情况下，如何评估新生成问题的价值并优化其生成过程。论文提出Ask-to-Distill（\method）框架，将合成问题视为分布诱导器（distribution elicitors），通过冻结的学生-教师对构建一个分布代理奖励（distributional surrogate reward），该奖励基于教师与学生之间的不一致程度、教师对学生产生候选输出的支持概率质量，以及问题长度和多样性控制，实现无正确答案、无逐候选更新的端到端优化。关键创新在于利用学生生成轨迹上的教师监督信号密度来指导问题生成，从而提升蒸馏效率。实验表明，该方法在数学推理、代码生成和智能体任务中均优于未优化的学生生成提示及KL过滤的真实问题，在多数场景下达到或超越真实问题库的OPD性能，且信号分析验证了优化后问题能以更短的学生响应获取更密集和更丰富的监督信号，支持“诱发分布监督”作为合成蒸馏数据的优化目标。
链接: https://arxiv.org/abs/2610.27450
作者: Jiashuo Sun,Hao Zhang,Xueqiang Xu,Heng Wang,Yixuan Xie,Pengcheng Jiang,Zhenrui Yue,Jiajun Fan,Peiran Li,Ge Liu,Shaowen Wang,Jiawei Han
机构: University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校); The Hong Kong University of Science and Technology (香港科技大學); Texas A&M University (德州农工大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 28 pages, 10 figures, 5 tables
Abstract:On-policy distillation (OPD) can succeed with surprisingly few training questions, motivating a view of questions as carriers of distributional knowledge transfer. How can we assess whether a newly generated question is valuable for distillation, and learn to generate better questions—without access to its correct answer? We introduce Ask-to-Distill (\method), a framework that treats synthetic questions as distribution elicitors and optimizes the teacher supervision they expose along student-generated trajectories. A student-initialized synthesizer is optimized using a distributional surrogate for question value, computed from a frozen student–teacher pair without reference answers, correctness rewards, or per-candidate student updates. The reward combines teacher–student disagreement with teacher probability mass on the student’s candidate support, together with length and diversity controls. The generated questions are filtered and deduplicated into a prompt-only bank for standard OPD. Across mathematical reasoning, code generation, and agentic tasks, our method improves average performance and teacher-gap recovery over unoptimized student-generated prompts and KL-filtered real questions, while matching or exceeding real-bank OPD in the evaluated settings. Signal analyses further show that optimized questions can elicit denser supervision and more total signal from shorter student responses, supporting elicited distributional supervision as an objective for synthetic distillation data. Our code, data and models are publicly available at this https URL.
[NLP-86] Harness Self-Evolution for Long-Horizon Terminal Agents
【速读】： 该论文旨在解决长时程终端智能体（long-horizon terminal agents）在单次运行中可能成功执行任务，但在多次重复运行中却因随机误差传播而导致可靠性不足的问题。其核心挑战在于模型能力与稳定可靠执行之间存在差距，尤其在动作、上下文及反馈交互过程中，微小的随机扰动可能被放大并导致失败。解决方案的关键是提出RiseHarness框架，通过固定模型权重，在不改变模型本身的前提下，迭代优化任务执行框架（harness）。该框架聚焦于同一配置下同时存在成功与失败运行的任务实例，利用成功案例作为参照进行失败分析。RiseHarness结合智能体轨迹与运行时监控器生成的内部框架追踪信息，系统分析模型动作如何与命令执行、反馈处理、上下文管理及完成决策等机制相互作用。通过构建“框架手册”（Harness Handbook），将证据与具体框架行为及源码关联，识别跨任务共享机制中的共性问题。随后，框架会搜索其他框架中的相关实现，并将其适配为当前模型和执行协议的设计参考。候选修改方案经过行为验证和多轮任务评估，表现提升的版本将进入下一轮迭代。这一闭环过程将不同框架中的有效机制整合至单一框架中，并以新执行证据驱动后续改进。实验验证了在不同模型上的任务成功率与可靠性提升，进行了框架执行证据与外部研究的消融分析，并测试了在未见基准上的迁移能力。RiseHarness以特定模型为单位进行框架适应，揭示了对共享交互机制的优化如何显著提升现有模型在多样化任务中的执行可靠性。
链接: https://arxiv.org/abs/2610.27435
作者: Ruhan Wang,Yucheng Shi,Zongxia Li,Junyao Yang,Zhongzhi Li,Yue Yu,Haitao Mi,Dongruo Zhou,Leoweiliang
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 38 pages, 7 figures, 7 tables, including appendices
Abstract:Long-horizon terminal agents may solve a task in one run yet fail across repeated runs. Stochastic errors can propagate through actions and context, creating a gap between capability and reliable execution. We present RiseHarness, a framework for iteratively improving agent harnesses with model weights fixed. It focuses on flipping tasks with both successful and failed runs under the same configuration, using successful runs as references for failure analysis. RiseHarness combines agent trajectories with internal harness traces from a runtime monitor to examine how model actions interact with command execution, feedback, context management, and completion decisions. A Harness Handbook links this evidence to harness behaviors and source code, helping identify recurring problems in mechanisms shared across tasks. RiseHarness then searches other harness frameworks for relevant implementations and adapts them as design references to the current model and execution protocol. Candidate changes undergo behavioral checks and repeated task evaluation; versions with measured gains seed the next iteration. This loop integrates suitable mechanisms from different frameworks into a single harness and uses new execution evidence to guide further changes. We evaluate task success and reliability across models, ablate harness execution evidence and external research, and test transfer to unseen benchmarks with the same model. RiseHarness treats a specific model as the unit of harness adaptation, studying how changes to shared interaction mechanisms can make existing capabilities more reliable across tasks.
[NLP-87] he Router Already Knows
【速读】： 该论文旨在解决在混合专家（Mixture-of-Experts, MoE）模型中，如何高效利用稀疏激活机制来提升特征表示能力的问题。现有方法通常为每个输入 token 选择少数专家进行处理，但未充分考虑这些被选中的专家是否可共享一个紧凑的、结构化的字典以进一步压缩和重构其输出。本文提出非路由自编码器（NR-SAE），其核心创新在于：为每个被路由选择的专家分配一个小型字典，并仅激活由冻结的路由器所选定的字典部分，从而实现基于路由决策的稀疏字典复用。实验结果显示，在 Gemma 4 26B-A4B 模型第 15 层，经过十万个 C4 训练样本后，NR-SAE 能解释 0.518 的路由加权专家输出方差，显著优于多种控制基线（如秩地址控制、候选匹配控制及架构匹配控制）。这一优势部分源于架构设计——每个专家仅作用于模型宽度不超过四分之一的子空间，因此专家名称即隐含了其对应的子空间语义，使得路由决策本身已携带丰富的结构信息。然而，该方法尚未验证训练加速效果，且语义可解释性、下游任务性能与因果忠实性仍待评估。
链接: https://arxiv.org/abs/2610.27421
作者: Andrew Bistras,Marty Johnson
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Mixture-of-experts (MoE) models already make a sparse choice for every token: which experts should process it. We ask whether a sparse dictionary should reuse that choice. NR-SAE gives each routed expert a small dictionary and opens only those the frozen router selects. At block 15 of Gemma 4 26B-A4B, after ten million C4 training tokens and across three seeds, the expert-aligned autoencoder explains 0.518 of the variance in router-weighted expert outputs on a fixed evaluation capture. A rank-addressed control with the same parameters but sixteen times as many candidates reaches 0.481, a candidate-matched control 0.458, and an architecture-matched control with randomized expert identity 0.387. A transcoder variant also leads all three controls: 0.432, against 0.347, 0.371, and 0.118. Part of this advantage is architectural: each expert writes into a subspace of at most a quarter of the model width, so naming the expert names the subspace. How much it explains, we have not measured. At this layer and budget, the router already knows something a dictionary would otherwise have to learn. Our implementation did not train faster, and semantic interpretability, downstream loss, and causal faithfulness remain untested.
[NLP-88] Looking Into The Past: The Impact of Context on Detecting Deceptive Statements
【速读】： 该论文旨在解决传统欺骗检测方法仅孤立分析单条消息而忽视上下文信息导致的准确性不足问题。其核心挑战在于如何有效利用隐藏角色游戏Avalon中的前期游戏对话内容，以提升对欺骗性信息的识别能力。解决方案的关键在于引入游戏过程中的前后文对话作为额外上下文信息，通过大语言模型（Large Language Models, LLMs）融合这些动态交互数据，显著增强对目标消息是否具有欺骗性的判断能力。实验结果表明，加入游戏上下文后，模型在多项评估指标上均取得明显提升，其中准确率最高可达31%的改进，验证了上下文信息在复杂社交情境下对欺骗检测的重要价值。
链接: https://arxiv.org/abs/2610.27407
作者: Alex Clay,Ernesto Jimenez-Ruiz,Pranava Madhyastha
机构: City St George’s, University of London(伦敦城市圣乔治大学); Alan Turing Institute(艾伦图灵研究所)
类目: Computation and Language (cs.CL)
备注: Work completed in spring 2025
Abstract:Deception detection typically considers messages in isolation. Previous works in claim verification and role detection have displayed that additional information such as gameplay history is crucial to successful predictions. In this work, we demonstrate the value of preceding gameplay dialogue when applied to deceptive message detection within the hidden role game of Avalon. The findings show that LLMs can see improvement across a number of metrics, including accuracy by up to 31% with the addition of gameplay context for the target message.
[NLP-89] More Memory Worse Agents : Error Reproduction and Anti-Persistence in LLM Agents
【速读】： 该论文旨在解决大语言模型（LLM）智能体在无需更新模型参数的情况下，通过外部记忆存储自身执行经验以实现跨任务改进时所面临的“错误传播”问题。核心挑战在于：若将从经验中抽象出的规则、总结和技能长期保存于记忆中，虽可提升策略复用效率，但一旦抽象过程引入误差，这些错误将在后续决策中通过检索与重用被持续放大，并通过“抽象-执行-再抽象”的反馈循环不断自我复制，形成难以消除的错误固化现象，即“错误重现”（error reproduction）。为应对这一问题，论文提出一种名为Anti-Persistence的记忆设计，其关键在于将经验积累与抽象持久化相分离：系统仅长期保留原始轨迹与结果，而任务特定的指导信息由自适应选择器根据历史结果按需动态重构，且在执行后立即丢弃，从而杜绝错误抽象的直接重复使用。尽管仍可通过保留的原始轨迹实现间接反馈，但有效阻断了错误在抽象层面的持续传递路径。实验表明，该方法在多个基准测试与不同模型架构下均实现了更高的离线成功率及最低的平均大语言模型调用成本，验证了在基于记忆的自我改进中分离经验积累与抽象持久化的有效性。
链接: https://arxiv.org/abs/2610.27406
作者: Ruhan Wang,Kishan Panaganti,Dongruo Zhou
机构: Indiana University (印第安纳大学); Tencent (腾讯)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 31 pages, 12 figures, 7 tables, including appendices
Abstract:LLM agents can improve across tasks without updating model parameters by storing experience from their own executions in external memory and reusing it to guide later decisions. A central design question is whether to retain only observed trajectories and outcomes or also the rules, summaries, and skills inferred from them. Retaining these abstractions allows agents to reuse strategies without repeatedly deriving them from raw trajectories. However, errors in stored abstractions can propagate to subsequent decisions through retrieval and reuse. They can also shape execution trajectories, which then become the basis for new abstractions that reproduce the original errors. We term this feedback mechanism error reproduction: errors are regenerated through a cycle of abstraction, execution, and re-abstraction, extending their persistence beyond direct reuse of the original faulty knowledge. We propose Anti-Persistence, a memory design that accumulates trajectories and outcomes while reconstructing task-specific guidance on demand. An adaptive selector uses prior outcomes to decide whether and how to synthesize guidance, which is discarded after execution. This prevents its direct reuse as a stored object, although indirect feedback through retained trajectories remains possible. Across different benchmarks and backbones, Anti-Persistence achieves higher overall held-out success and the lowest average LLM-call cost among the evaluated memory designs. These results support separating experience accumulation from abstraction persistence in memorybased self-improvement.
[NLP-90] Steering intonational tunes in text-to-speech synthesis
【速读】： 该论文旨在解决生成式语音合成系统中韵律语调（prosodic intonation）的可控制性问题，尤其聚焦于英语焦点调型（nuclear tune）的生成机制，即区分陈述句（下降调型）与疑问句（上升调型）的声调模式。其核心问题是：在基于残差向量量化（residual vector quantization）的两阶段语音合成架构中，究竟哪些组件负责编码和传递语调信息？解决方案的关键在于识别出语调信息主要由编码器骨干网络（backbone）的隐藏状态承载，并通过深度模型（depth model）传递至波形生成过程，而第一层令牌堆栈（first level of the token stack）几乎不携带语调信息。研究发现，仅通过标点符号（如句号或问号）进行提示控制效果有限，而利用人类对理论指定调型的模仿所生成的样本，计算其骨干网络状态的均值差异向量（difference-of-means vector），并将其嵌入生成循环中，可有效引导合成语音朝预期的上升或下降调型演变，且在自然度损失较小的情况下实现跨系统的一致调控。这些模仿导出的向量与标点条件下的向量显著相关，表明带标点输入的骨干状态已隐含语调信息。尽管语调整体趋势可被精准控制，但目标调型的精细选择仍受限于边缘音高（edge tone）比重音（pitch accent）更一致的响应特性，无法实现精确调型选择。
链接: https://arxiv.org/abs/2610.27403
作者: Juan Pablo Vigneaux,Jennifer Cole
机构: Northwestern University (西北大学)
类目: ound (cs.SD); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 39 pages, 25 figures, 12 tables (13 pages without appendices)
Abstract:Text-to-speech systems built on residual vector quantization often generate audio in two stages: a transformer backbone emits the first level of each frame’s token stack, and a smaller depth model generates the rest autoregressively. We ask which components of such a model are responsible for prosodic intonation, more precisely for generating the English nuclear tune (for instance, the falling pitch melody that distinguishes an assertion from a question’s rising melody), and whether interventions on those components can be used to control intonation. Prompt-based control mostly fails; only a final question mark is a reliable lever. Interchange interventions in CSM and S2 locate the tune: the backbone’s hidden state carries it to the waveform through the depth model, and the first level of the stack carries little or none of it. Steering the backbone state with a difference-of-means vector computed on new material generated with a final period or question mark moves the tune, but the push toward an assertion degrades the voice. Steering vectors computed instead from the backbone states associated with human imitations of theoretically-specified tunes, whose text carries no punctuation, and applied inside the generation loop, move most recordings to the other condition in both systems (even when fitted without the speaker being steered), at a small cost in predicted naturalness. These imitation-derived vectors are aligned with the one obtained from the two punctuated conditions far beyond a shuffled-label null, so the backbone state of a punctuated prompt already represents intonation. The steered recordings broadly become falling or rising, according to the intent, but the edge tone moves more coherently than the pitch accent, and the target tune cannot be selected precisely beyond this broad tendency.
[NLP-91] DStill: Trajectory Consistency Distillation for Accelerated One-Step Drafting in Block Diffusion Speculative Decoding
【速读】： 该论文旨在解决生成式 AI（Generative AI）中基于扩散模型的并行采样（block-diffusion speculative decoding）在单步草稿生成时因词元因子化误差（token factorization error）导致的草稿不一致性问题。该问题源于单步生成中所有词元均基于相同掩码块进行预测，而未考虑其他位置已选定词元的条件信息，从而限制了草稿质量与接受长度（acceptance length）。现有解决方案如迭代精炼虽能提升草稿保真度，但需多轮计算，显著增加推理延迟。为平衡草稿质量与推理效率之间的矛盾，本文提出 DStill，一种两阶段蒸馏框架：首先训练一个目标条件化的草稿生成器完成部分揭示块的补全，随后利用其多步精炼轨迹训练一个学生模型，实现仅通过单次前向传播即预测最终完整输出。通过设计从部分掩码到全掩码输入的渐进式课程学习策略，成功将多步精炼能力迁移至单步生成。实验表明，在三大模型族的推理、数学与代码基准上，DStill 相比单步 DFlash 最高提升接受长度 28%、加速比达 25%，并相较自回归解码实现最高 6.45 倍加速。其核心创新在于通过知识蒸馏将多步精炼过程压缩至单步，同时保持高质量草稿生成能力。
链接: https://arxiv.org/abs/2610.27379
作者: Shreyas Kulkarni,Uttam Gurram,Madhav Kanda,Krishna Teja Chitty Venkata,Shichu Chen,Kartik Balasubramaniam,Supriyo Chakraborty,Sambit Sahu
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Block-diffusion speculative decoding accelerates large language model inference by drafting multiple tokens in parallel. However, one-step drafting in discrete diffusion is limited by token factorization error: each token is predicted from the same masked block, without conditioning on the tokens selected at other positions in that step. This can produce inconsistent drafts and reduce acceptance length. Iterative refinement mitigates this limitation by conditioning later predictions on previously committed tokens, though additional passes increase drafting cost. To address this trade-off between draft fidelity and inference latency, we introduce DStill, a two-stage framework that distills a target-conditioned drafter’s own multi-step refinement into one-step drafting. We first train the drafter to complete partially revealed blocks, then use its refinement trajectories to train a student to predict the final completion in one pass. Using a curriculum that gradually transitions from partial to fully masked inputs, we successfully transfer the multi-step refinement capabilities into one step. Across reasoning, math, code benchmarks on three model families DStill improves acceptance length by up to 28% and speedup by up to 25% over one-step DFlash, achieving up to 6.45x speedup over auto-regressive decoding.
[NLP-92] SHAM-Bench: A Span-level Hallucination Benchmark Across Multiple Tasks
【速读】： 该论文旨在解决生成式 AI（Generative AI）输出中细粒度幻觉检测（span-level hallucination detection）的评估标准不统一、泛化能力难以验证的问题。现有方法在不同自然语言生成（NLG）任务与数据集间缺乏可比性，且多数数据集局限于单一领域，导致无法有效评估检测模型在跨任务、跨域场景下的泛化性能。为此，本文提出SHAM-Bench（Span-level Hallucination Benchmark Across Multiple Tasks），构建了一个涵盖六类NLG任务的统一基准，整合了总计6,461个经过人工标注的样本，并采用统一的数据模式与评估协议。其核心解决方案在于通过多任务、多领域数据的标准化整合，实现对幻觉检测方法的系统性评估。实验表明，微调型检测器在同域任务中表现良好，但在分布外（distribution shift）情况下性能显著下降；而基于提示（prompt-based）并引入推理机制的方法则展现出更强的跨任务鲁棒性。此外，研究发现细微的非支持内容仍难被准确识别，且跨度边界模糊与标签噪声仍是当前评估的主要挑战。为促进后续研究，作者开源了数据集、评估代码及公开排行榜。
链接: https://arxiv.org/abs/2610.27370
作者: Sicong Huang,Ian Lane
机构: University of California, Santa Cruz(加州大学圣克鲁斯分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:Span-level hallucination detection pinpoints unsupported content in LLM outputs, enabling actionable feedback and targeted mitigation. Existing evaluations for span-level hallucination detection methods are fragmented across tasks, annotation formats, and metrics, and most span-level datasets cover only a single NLG domain, making it difficult to assess whether a detection method actually generalizes across tasks and domains. In this paper, we introduce SHAM-Bench (Span-level Hallucination Benchmark Across Multiple Tasks), a unified benchmark for span-level hallucination detection spanning six NLG tasks, aggregated from existing human-annotated datasets totaling 6,461 examples under a common schema and evaluation protocol. Using SHAM-Bench, we evaluate existing hallucination detection methods and find that fine-tuned detectors often perform well in-domain but degrade under distribution shift, whereas prompt-based methods with reasoning are more robust across tasks. Our analysis further shows that subtle unsupported content remains difficult to detect and span-level evaluation is still challenged by ambiguous span boundaries and noisy labels. We release the dataset, evaluation code, and a public leaderboard to support future research on broadly generalizable hallucination detection.
[NLP-93] Explore More Validate Less: Efficient Reflective Prompt Optimization
【速读】： 该论文旨在解决生成式任务中反射式提示优化（Reflective Prompt Optimization）因候选提示生成与全量验证绑定而导致的搜索成本高、过程串行化的问题。其核心解决方案是提出一种两阶段框架，将候选提示池的构建与验证分配解耦：在第一阶段，通过并行多分支反射机制，复用父级提示的训练响应，仅在小型训练批次上对每个修订版本进行局部筛选，从而实现无需全量验证即可支持后续迭代；在第二阶段，基于固定验证预算，采用探索策略从冻结的候选池中选择最优提示，最终输出验证得分最高的提示。该设计使候选池规模与验证预算可独立设置，显著提升效率。实验表明，在Qwen3.5-9B和Llama-3.1-8B-Instruct模型上，该框架相较初始求解器最多提升11.22个百分点，且学习型选择策略在所有对比中均优于初始状态，尽管两种选择策略未形成一致优劣关系。结果验证了将反射搜索与受限验证分离的有效性，并为更优的候选选择策略留出改进空间。
链接: https://arxiv.org/abs/2610.27367
作者: Xiaofan Zhou,Lu Cheng
机构: Pennsylvania State University (宾夕法尼亚州立大学)
类目: Computation and Language (cs.CL)
备注: 10 pages
Abstract:Reflective prompt optimization improves instructions using task feedback, but coupling candidate generation with full-validation-based selection can make search costly and sequential. We propose a two-stage framework that decouples reflective candidate-pool construction from validation allocation. During construction, we sample multiple parents and reuse each parent’s training responses across parallel reflection branches. Each revision is screened against its parent on a small training minibatch, allowing locally improved candidates to support further revisions without full validation. After construction, an exploration policy selects candidates from the frozen pool under a fixed full-validation budget, and the prompt with the highest measured validation score is returned. This separation allows candidate-pool size and validation budget to be set independently. We evaluate random selection and learned score prediction with similarity-based exploration on five benchmarks using Qwen3.5-9B and Llama-3.1-8B-Instruct. With 100 retained candidates and 20 full validations, our framework improves over the initial solver by up to 11.22 percentage points. Learned selection improves over the initial solver in all ten comparisons, although neither selection policy consistently outperforms the other. These results support separating reflective search from budgeted validation while leaving room for stronger candidate-selection policies.
[NLP-94] On the LLM Verification for Argument Graph Generation CIKM2026
【速读】： 该论文旨在解决多智能体大语言模型（Multi-agent LLM）生成的论证图（argument graph）在质量保障环节中过度依赖大语言模型（LLM）作为评审者所带来的可靠性与效率问题。其核心挑战在于，尽管现有方法普遍采用另一LLM充当审查者以验证图结构的正确性，但此类方法存在检测误报率高、判断不一致以及计算延迟显著等缺陷。论文提出“分层验证”（layered verification）机制作为解决方案，其关键在于根据不同技术手段的确定性能力进行任务分工：优先通过确定性检查（deterministic checks）完成可形式化验证的部分，再结合基于求解器的符号验证（solver-based verification），仅在存在不确定性时引入LLM评审，并将人类判断保留给其他工具无法决断的边界情况。研究通过在人工标注的论证图中注入结构缺陷进行实证分析，发现尽管LLM评审者具备接近天花板的召回率（near-ceiling recall），能够有效识别支持关系中的循环结构等缺陷，但其对正确图结构的误判率高达约10%，且在相同输入下表现出判断不一致性，响应速度远慢于等效的精确检查。这表明，尽管生成式AI（Generative AI）具备一定的缺陷识别能力，但在论证图验证中仍需建立基于确定性规则和符号推理的前置验证层，以确保结果的准确性、一致性和效率。
链接: https://arxiv.org/abs/2610.27361
作者: Jakub Bąba,Jarosław A. Chudziak
机构: Warsaw University of Technology (华沙理工大学); Institute of Computer Science (计算机科学研究所)
类目: Computation and Language (cs.CL)
备注: Accepted for GaLM workshop at CIKM 2026
Abstract:Multi-agent LLM systems can formalize natural-language argumentation end-to-end, turning plain text into structured graphs by classifying components, connecting them with relations, and enriching the resulting graphs into formal models. Quality assurance for those graphs is almost always another LLMs: a reviewer, judge, or critic agent. On the other side, in adjacent neuro-symbolic fields, such generated output is often verified symbolically by external checks and solvers. In this work, we argue that in argument graph verification, the allocation should be similar, and based on what each instrument can guarantee. We propose layered verification: starting from deterministic checks, moving to solver-based and LLM reviews, and utilizing debate only under uncertainty, leaving human judgment for what no other instrument can decide. We ground our principle with a study on human-annotated graphs with injected structural defects. An LLM reviewer detects them at near-ceiling recall, including cycles in the support relation, but it objects to almost one in a ten correct graphs, even though it should not, happens to be inconsistent in verdict on identical inputs and is significantly slower than the equivalent exact check. It shows that while LLM reviewer does not lack the ability to detect structural defects, it is worth having a deterministic verification gates.
[NLP-95] Global Security Blind Spots: Translation as a Safety Bypass in Low-Resource Languages NEURIPS2026
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在多语言部署中安全对齐（safety alignment）失效的问题，特别是针对低资源语言（low-resource languages）的安全性漏洞。尽管模型在英文环境下经过安全对齐训练，但当用户请求被翻译至低资源语言时，原本在英语中被拒绝的有害请求可能触发不安全的生成结果，即存在“翻译绕过”（translation bypass）风险。其解决方案的关键在于提出一种基于英文安全性的评估指标，能够将翻译绕过导致的不安全性与模型在英文阶段本就存在的不安全性相分离，从而更准确地量化跨语言的安全性迁移问题。研究发现，英文安全性无法可靠迁移到其他语言，平均绕过率在21.9%至44.4%之间，且绕过风险集中于特定语系，这一模式在不同模型间具有共性，但其具体表现随模型差异显著，且同一模型在不同评估条件下排名不稳定，表明当前安全评估体系在多语言场景下缺乏一致性与可靠性。
链接: https://arxiv.org/abs/2610.27304
作者: Jalil Alva,David Harper,Abigail E. Thornton,Rafael S. Villalobos Alva,Leonardo Kenji Minemura Suazo,Konstantinos Karageorgos,Diego Velazquez,Sergio A. Serrano,Marvin Vargas Flores,Aaron Schliem,Fernando F. Migone
机构: Welo Data
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
备注: 19 pages, 4 figures. Accepted at NeurIPS 2026 Workshop: Foundations of Language Model Security (FLMSec)
Abstract:Large language models are deployed globally, yet safety alignment and evaluation remain concentrated in English: harmful requests refused in English can elicit unsafe completions when translated into low-resource languages. We evaluate ten widely deployed LLMs across 79 languages (English plus 78 low-resource) under a translation-bypass threat model, using a metric conditioned on English safety, which separates translation bypass from unsafety already present in English. English safety does not transfer reliably: conditioned on safe English completions, mean bypass rates range from 21.9% to 44.4%. Bypass risk concentrates in a subset of language families, a pattern broadly shared across models. The shared ordering is confounded with translation degeneracy, but its variation across models is significant and survives every control we apply. We also show that per-model bypass rates do not identify a stable ranking: nine of ten models change rank across three conditioning schemes that estimate the same quantity, with no change to any underlying response.
[NLP-96] Evaluating Context Compaction on the Questions That Come Later
【速读】： 该论文旨在解决语言模型代理在上下文压缩过程中，如何有效保留对后续问题具有价值的信息这一核心挑战。现有方法通常基于当前问题的必要性选择保留内容，但这种策略往往忽视了压缩后上下文对后续问题的支持能力，导致长期信息保真度下降。其解决方案的关键在于：从以“当前问题”为中心的句子选择转向更细粒度的词级压缩，并通过保留语义关键内容词（content words）来提升跨问题的上下文有效性。研究发现，尽管基于当前问题筛选句子的方法在当前问题上显著提升性能（如在MultiRC上提升11.4个百分点），但在后续问题上的表现并未优于随机删除；而采用词级压缩的LLMLingua-2则在后续问题上表现更优，其优势源于对内容词的更高保留率（约为随机删除的两倍），并在等量内容词保留条件下仍保持7.3个百分点的增益（探索性分析）。此外，在不同任务设置下（如FEVER、Evidence Inference、τ-bench），压缩策略的有效性呈现显著差异，表明任务依赖性与工具调用精确性对压缩策略的选择至关重要。因此，论文强调：仅关注当前问题的性能指标无法可靠反映压缩对后续问题的支持能力，真正有效的压缩需以跨任务、跨时间步的语义完整性为目标，而非局部优化。
链接: https://arxiv.org/abs/2610.27301
作者: Junwei Yao,Yongxi Zhou,Zike Yang,Zhijun Wang,Jianan Liu,Yanxian Chen,Wenbo Ye
机构: Independent Researcher(独立研究员)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Language-model agents compact their context for the question at hand, but the compacted context must also serve questions that come later. We evaluate compaction on these later questions, using four corpora that annotate the evidence of several questions per text and recorded agent trajectories, and compare every method with random deletion at the same budget. On MultiRC, keeping the sentences that gpt-6-luna (through the OpenAI Decisions API) judges necessary for the current question raises its accuracy on that question by 11.4 points at 25% of the tokens. On the later questions, however, no method that selects sentences for the current question does detectably better than deleting sentences at random. Deleting words instead of sentences does better: LLMLingua-2, which ignores the question, beats random sentence deletion at the same token count by 12.5 points on the later MultiRC questions. It keeps about twice random deletion’s share of the evidence’s content words, and at equal content-word retention its gain shrinks to 7.3 points (exploratory). On FEVER, label accuracy can overstate what the kept text supports: the gain of sentence selection is no longer detected when answers are scored against the kept evidence, and the 21.0-point gain of LLMLingua-2 is not shown to rest on the kept text. The ranking also depends on the setting: on Evidence Inference sentence selection beats LLMLingua-2 at a nominal 25% budget, and on \tau -bench trajectories, whose later tool calls need exact identifiers, it beats random deletion while LLMLingua-2 shows no detected gain. Scores on the question a compaction was made for do not reliably show how it serves later ones.
[NLP-97] RealControlBench: A Benchmark for Interpretability-Informed Fine-Grained Control of LLM s and Agents
【速读】： 该论文旨在解决现有生成式控制（steering）评估基准在现实应用中代表性不足的问题。当前的基准大多基于简化、孤立的生成任务，在受控环境下仅评估单一行为属性，无法反映实际部署场景中行为目标的多样性与复杂性，导致评估结果难以真实反映控制方法在现实中的有效性。为此，本文提出RealControlBench，一个全面的评估框架，用于衡量大语言模型（LLM）及LLM代理在多种行为目标和交互设置下的细粒度控制能力。其关键创新在于将评估范围从孤立生成扩展至复杂、执行落地的多轮代理交互，并涵盖五个互补维度：指令遵循控制、代理工具使用决策控制、推理性能调节、推理长度控制以及轨迹条件化安全意识。该基准不仅评估目标可控性，还联合衡量能力保持性，包括指令遵循、推理性能及下游代理的实用性与安全性；同时检验控制效果在提示变化、领域迁移和模型架构差异下的特异性与鲁棒性。实验表明，现有控制方法在真实控制目标下表现不可靠，常无法稳定诱导预期行为，且可能显著损害模型的基础任务性能。这揭示了在受控环境中表现良好的控制效果未必能在更真实的评估中实现可靠控制。通过提供一个统一、模型无关的评估框架，RealControlBench旨在推动开发出在实际应用中兼具有效性、精确性、鲁棒性与实用性的新型控制方法。
链接: https://arxiv.org/abs/2610.27295
作者: Xinying Cai,Bin Hu,Mingyu Jin,Qiming Zhang,Yang Ouyang,Bangwei Guo,Minghao Guo,Yutong Yin,Lang Gao,Feijiang Han,Sheldon Yu,Kaiqu Liang,Qingcheng Zeng,Jingcheng Niu,Fan Yang,Wei Cheng,Zhen Xiang,Julian McAuley,Zhaoran Wang
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Existing steering benchmarks are built around simplified, isolated generation tasks and evaluate a single behavioral attribute under controlled settings that do not reflect the diversity and complexity of practical deployment. As a result, performance on these benchmarks provides only a limited picture of how steering methods perform in the real world. We therefore introduce RealControlBench, a comprehensive benchmark for evaluating fine-grained steering of LLMs and LLM agents across diverse behavioral targets and interaction settings. RealControlBench extends steering evaluation beyond isolated generations to complex, execution-grounded, multi-turn agent interactions, covering five complementary dimensions: instruction-following control, agentic tool-use decision control, reasoning-performance steering, reasoning-length control, and trajectory-conditioned safety awareness. Rather than measuring steering success alone, the benchmark jointly evaluates target controllability and capability preservation, including instruction following, reasoning performance, and downstream agent utility and safety. It further examines whether steering effects remain specific and robust across prompt variations, domains, and model architectures. Our evaluation shows that existing steering methods remain unreliable across realistic control objectives: they often fail to consistently induce the intended behavior and can substantially degrade the model’s underlying task performance. These results suggest that strong steering effects in controlled settings do not necessarily translate into reliable control in more realistic evaluations. By providing a unified, model-agnostic steering evaluation framework, RealControlBench aims to support the development of steering methods that are not only effective but also precise, robust, and useful in realistic AI deployments.
[NLP-98] EX: Test-Time Scaling Testing Agents via Execution-based Cross-Validation
【速读】： 该论文旨在解决语言模型在长时序任务中测试用例生成的挑战，现有测试时扩展方法要么孤立候选样本，要么聚合执行轨迹，难以有效扩展至长时序任务。其解决方案的关键在于提出TEX（Test-case Execution-driven eXecution），一种基于执行驱动的交叉验证协同测试时扩展框架。TEX通过并行化多个智能体生成测试用例与修复补丁，并对每个测试用例与所有补丁进行交叉执行，构建K×K反馈矩阵；各智能体基于自身对应的矩阵子集进行迭代优化，实现无需冗长轨迹共享的客观验证。实验表明，在SWT-Bench Verified基准上，TEX达到86.4%的准确率，创下新纪录，较单智能体基线提升9.0%；即使仅随机选取TEX输出结果（84.4%）也优于此前最优方法（84.0%），证明协作执行反馈可显著提升解的质量，而不仅依赖最优样本。研究结果强调了以执行为基础的反馈机制在长时序智能体任务中的高效扩展潜力。
链接: https://arxiv.org/abs/2610.27290
作者: Akash Gokul,Srijan Bansal,Yingbo Zhou,Semih Yavuz
机构: Recursive(递归); Salesforce Research(销售force研究院); NVIDIA(英伟达)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Test-case generation is critical but remains challenging for language models. Existing test-time scaling methods either isolate candidates or aggregate traces, which do not scale to long-horizon tasks. We present TEX, a collaborative test-time scaling framework based on execution-driven cross-validation. TEX parallelizes agents to generate tests and patches, cross-executing each test against every patch to form a K \times K feedback matrix. Agents use their matrix slices for refinement, enabling objective validation without verbose trace sharing. On SWT-Bench Verified, TEX achieves 86.4%, establishing a new state-of-the-art and improving +9.0% over single-agent baselines. Moreover, even random selection from TEX outputs (84.4%) surpasses the previous SOTA (84.0%), showing collaborative execution feedback improves solution quality, not just best-case samples. Our results highlight execution-grounded feedback as an effective scaling mechanism for long-horizon agentic tasks.
[NLP-99] When Does Hate Transfer? A Large-Scale Study of Cross-Lingual Hate Speech Detection
【速读】： 该论文旨在解决跨语言仇恨内容检测中性能差异显著的问题，核心关注点是：某一语言的训练数据是否能够有效支持另一语言的仇恨言论检测，并实现相近的检测性能。其解决方案的关键在于通过大规模跨14种语言的实证研究，揭示了语言间迁移效果的规律性模式。研究发现，语言相似性（linguistic similarity）是预测跨语言迁移效果最强的量化因素，而时间重叠和数据集规模差异的影响相对较小；尽管如此，这些可测量因素仍无法完全解释所有迁移现象，因此进一步引入定性分析，提出共享文化线索与交际规范可能在部分语言集群中起到关键作用，这一观点作为基于观察的假设而非经验证明的因果机制。研究特别涵盖低资源语言，为实际构建更具鲁棒性的多语言仇恨内容检测系统提供了重要依据，明确了跨语言迁移的应用边界与潜在局限。
链接: https://arxiv.org/abs/2610.27285
作者: Zahra Bokaei,Walid Magdy,Bonnie Webber
机构: University of Edinburgh(爱丁堡大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Effective detection of hate content is essential for maintaining healthy online communities, yet performance can vary substantially across languages. This work investigates a core practical question: Can training data from one language support hate-speech detection in another language with comparable performance? To answer this question, we conduct a large-scale study across 14 languages. \changeWe observe consistent patterns in which certain language pairs transfer more effectively than others. Among the factors we measure, linguistic similarity is the strongest predictor of these patterns, with temporal overlap and dataset-size difference playing smaller roles. These measurable factors do not fully explain the observed patterns; a complementary qualitative analysis suggests that shared cultural cues and communicative norms may be associated with some of the remaining cases, particularly within several clusters. We present this cultural interpretation as a hypothesis from qualitative inspection rather than a measured cause. Our analysis includes low-resource languages, offering practical insights into when cross-lingual transfer can support them and where its limitations emerge. These findings inform the design of more reliable multilingual hate-detection systems.
[NLP-100] African Languages AI Safety Readiness Report
【速读】： 该论文旨在解决当前生成式 AI (Generative AI) 安全研究中对非洲语言严重代表性不足的问题。尽管大规模语言模型（LLMs）在多语言和跨领域应用中迅速普及，但资源分配高度集中于高资源语言，导致大量非洲语言在AI安全研究中被边缘化。其解决方案的关键在于构建一个系统性的评估框架，综合考量四类核心因素：与AI相关的资源可得性（如文本数据）、国家层面的AI政策支持、在主流文本基准测试中的表现以及母语使用者规模。基于此框架，论文将非洲语言划分为“完全就绪”“基本就绪”“新兴”和“亟需基础支持”四类准备度等级，旨在为非洲语言纳入AI安全基准测试提供依据，揭示不同语言间的准备差异，并指导针对低资源语言在数据建设、政策推动及社区参与方面的战略性投入。
链接: https://arxiv.org/abs/2610.27269
作者: Israel Abebe Azime,Jesujoba O. Alabi,Tajuddeen Gwadabe,David Ifeoluwa Adelani
机构: Masakhane; Saarland University (萨尔兰大学); Mila - Quebec AI Institute, McGill University, and Canada CIFAR AI Chair (Mila - 魁北克人工智能研究所，麦吉尔大学，加拿大加拿大人工智能主席)
类目: Computation and Language (cs.CL)
备注: Technical Report (working paper)
Abstract:With the widespread adoption of large language models (LLMs) across languages and domains, AI safety has emerged as a critical global concern. However, due to resource constraints, research and development efforts tend to prioritize a limited number of high-resource languages, leaving many others, especially those spoken in Africa, underrepresented. While existing research on language “resource-ness” has largely examined whether sufficient resources are available to develop language technologies, the question of whether languages are adequately represented in AI safety research has received attention only more recently. This report examines the readiness of African languages for inclusion in AI safety discussions, drawing on existing literature and data. It considers four factors: the availability of AI-relevant resources, including text data; national AI policy support; performance on established text benchmarks; and speaker population. Based on these factors, languages are grouped into four readiness categories: fully ready, mostly ready, emerging, and foundational support needed. This classification is intended to inform the inclusion of African languages in AI safety benchmarks, highlight varying levels of preparedness, and guide strategic investment in data development, policy, and community engagement for underrepresented languages.
[NLP-101] Auditing Safety Preference Supervision with Recorded Derivations
【速读】： 该论文旨在解决生成式AI（Generative AI）训练中偏好标签（preference labels）的可解释性与可验证性问题，即尽管偏好标签用于指导模型对齐，但其背后的判断过程难以追溯和验证。其核心解决方案是提出符号化偏好标注（Symbolic Preference Labeling），通过预训练语言模型对提示（prompt）和响应（response）的预定义属性进行分类，并基于显式规则生成响应判定结果，再依据固定的判定排序确定最终偏好标签。每个偏好标签均附带可追溯的推导记录——即属性赋值、规则和判定排序的完整历史，使得偏好标签可无需重新运行语言模型即可复现。实验表明，在200个由人类标注的样本中，该方法对192个提示成功生成结构化输出，所有解析失败均集中于被标记为有害的自残类提示。基于465组偏好对训练的策略模型在五次随机种子下表现稳定，平均在1,319个形似不安全的良性提示上实现40.82%的良性拒绝率，在655个有毒提示上实现84.98%的毒性拒绝率。该研究进一步区分了偏好标签的可复现性与属性标注正确性的差异，强调推导记录使偏好构建过程具备可审计性，而属性准确性与语义安全性仍需独立验证。
链接: https://arxiv.org/abs/2610.27266
作者: Chenhan Lyu,Amir Rahmani
机构: University of California, Irvine (加州大学欧文分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Symbolic Computation (cs.SC)
备注:
Abstract:Preference labels align model training, but the judgments behind them are difficult to inspect or verify. In this paper, we examine how preference construction can be checked when intermediate decisions are recorded, and which of them remain uncertain. We present Symbolic Preference Labeling, a method in which a pretrained language model classifies predefined attributes of prompts and responses. Explicit rules assign response verdicts, and a fixed verdict ranking determines the preference label. Each preference label has a recorded derivation: the response verdicts and preference label can be recomputed from the stored attribute assignments, rules, and verdict ranking without rerunning the language model. On 200 prompts annotated by human raters, the language model produces valid structured outputs for 192 prompts. All eight parsing failures occur in the subset of self-harm prompts marked as harmful. Using 465 preference pairs, we train one policy model with direct preference optimization across five training seeds. Mean measured benign refusal is 40.82 percent on 1,319 benign prompts that resemble unsafe ones, and mean measured toxic refusal is 84.98 percent on 655 toxic prompts. This study distinguishes the reproducibility of preference labels from attribute correctness. The recorded derivations make preference construction open to inspection, while attribute correctness and semantic safety require separate validation.
[NLP-102] xnQA: Benchmarking Analytical Reasoning and Evidence Grounding over Transaction Event Sequences
【速读】： 该论文旨在解决在不规则事件流（irregular event streams）背景下，对商户交易历史进行动态时间过滤、实体消解及聚合分析时，传统静态表格或时间序列预测基准无法有效捕捉复杂分析需求的问题。其核心解决方案是提出TxnQA，一个包含27类任务（81个模板）的分析型问答基准，并配套5种场景的欺诈风险推理套件，用于评估模型在合成且隐私保护的交易数据上的判断结果与支撑证据。通过程序化生成真值答案，并将模型性能分解为格式遵循率（FFR）、条件内容准确率（CA）与端到端正确率（E2E），研究发现：尽管微调能有效保证输出结构合规，但算术计算能力仍是主要瓶颈；在500实例的欺诈推理任务中，判决准确率可达80%以上，但联合准确率（要求精确归因证据）最高仅约23%，揭示出在交易序列问答中，输出模式合规性、语义计算准确性和证据归因三者构成相互独立的关键瓶颈。
链接: https://arxiv.org/abs/2610.27256
作者: Rabimba Karanjai,Yiming Bian,Yun-Shiuan Chuang,Jenna L. Ballard,Zijie Pan,Shivani Shekhar,Yixi Li,Yirou Ge,Avinash Thangali,Richard Williamson,Linsey Pang,Prakhar Mehrotra
机构: PayPal AI(PayPal人工智能); University of Pennsylvania(宾夕法尼亚大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Analyzing merchant transaction histories requires dynamic temporal filtering, entity resolution, and aggregation over irregular event streams–operations poorly captured by static tabular or time-series forecasting benchmarks. We introduce TxnQA, an analytical QA benchmark comprising 27 task families (81 templates) alongside a 5-scenario fraud-risk reasoning suite that evaluates both verdicts and supporting evidence over synthetic, privacy-mitigated transaction data. Ground-truth answers are computed programmatically, and model performance is decomposed into format-following rate (FFR), conditional content accuracy (CA), and end-to-end correctness (E2E). Evaluating prompted proprietary baselines and six LoRA-adapted open-weight models shows that adaptation reliably enforces output schemas but leaves arithmetic bottlenecks: Gemma-4-31B-IT and Qwen3-8B attain E2E scores of 0.867 and 0.816, whereas evaluated MoE checkpoints reach only 0.632-0.713. Furthermore, on the 500-instance fraud suite, verdict accuracy frequently exceeds 80%, yet joint accuracy (requiring exact evidence attribution) collapses to at most ~23%. These findings identify schema compliance, semantic calculation, and evidence attribution as three distinct bottlenecks in transaction-sequence QA.
[NLP-103] arget-Aware Curation of Hallucination Stress Tests for Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）在评估中存在“幻觉”（hallucination）行为时，不同模型间幻觉表现难以有效对比的问题。尽管多个模型在相同提示（prompt）下可能表现出相似的平均幻觉得分，但其具体幻觉类型与触发机制可能存在显著差异，导致现有评估方法缺乏对模型真实幻觉行为的敏感性。为此，论文提出一种迭代式提示编辑方法，通过引入一个基于语言模型的编辑器（LLM editor），根据目标模型的回答及裁判（judge）针对特定类别幻觉的评分与解释反馈，动态优化源相关种子提示（source-grounded seed prompts）。关键在于：该过程不仅依赖于裁判对幻觉严重性的判断，还让目标模型接收自身先前回答的反馈，从而引导提示逐步演化出更具诊断性的幻觉诱发能力。实验覆盖7个数据集、10个主题共500个种子提示，涵盖内在与外在幻觉，在开源与闭源模型中均验证了该方法的有效性。结果显示，经编辑后的提示显著提升了裁判评定的幻觉严重度，且超过一半的分数提升来自后期编辑轮次；在通过模型辅助源有效性检验的219个提示中，仍有89个引发模型输出未经支持的主张。更重要的是，经过筛选的提示可在多个模型间复用，但其诊断价值高度依赖于源内容的有效性以及所暴露的模型特异性缺陷。
链接: https://arxiv.org/abs/2610.27247
作者: Jiho Lee,Bingquan Zhang,Lin Ting Liang,Soroush Vosoughi,Weicheng Ma
机构: Dartmouth College(达特茅斯学院); Oakland University(奥克兰大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Hallucination stress tests can yield similar average scores across language models even when different prompts elicit hallucinations in each model. We develop a procedure that iteratively edits source-grounded seed prompts to expose hallucinations in a target model. An LLM editor revises each prompt in response to the target’s answer and a judge’s category-specific score and explanation. In the main setting, the target also receives feedback on its previous response. Experiments cover 500 seeds from seven datasets and ten topics, with intrinsic and extrinsic hallucinations evaluated across open-weight and closed-source models. Editing substantially increases judge-rated severity over the seeds. In the baseline configuration, retained responses fall in the judge’s strong-hallucination bin in 77.2% of runs targeting extrinsic hallucination and 74.2% targeting intrinsic hallucination. Later editing rounds account for more than half of the mean score increase. Among 219 prompts that pass model-assisted source-validity checks, 89 elicit responses containing unsupported claims. Across five models, 200 curated prompts yield similar average severity scores, although pairwise Spearman correlations between prompt-level scores are only \rho=0.0387 to 0.1897 . Curated prompts can be reused across models, but their diagnostic value depends on source validity and the failures they expose in each model.
[NLP-104] Diagnostic Analysis for Large Language Model Reasoning EMNLP2026
【速读】： 该论文旨在解决生成式推理结构（如链式与图结构）在开放域多跳问答任务中结构可靠性不足的问题，尤其关注其在概率推理、事实验证及不确定性估计中的有效性。其核心解决方案在于提出一种两阶段诊断评估框架：第一阶段通过评估构建推理图所需的关键能力，揭示图结构的可靠性受拓扑设计、提示工程及模型家族的影响；第二阶段以不确定性估计为下游任务，通过“生成边”实验发现基于图的计算对上游图错误高度敏感，性能显著低于使用真实图（gold-graph）的情况；而“真实图”实验进一步表明，仅将图结构作为提示上下文时提升有限，但可有效增强正确与错误答案在置信度评分上的区分能力，并支持节点级诊断。该研究的关键在于揭示了推理图结构的实际可靠性依赖于具体实现方式，并强调需从结构鲁棒性与下游任务一致性角度重新审视其应用价值。
链接: https://arxiv.org/abs/2610.27245
作者: Xinyu Liu,Dongqi Fu,Jun Wu
机构: Michigan State University (密歇根州立大学); Meta AI (Meta人工智能)
类目: Computation and Language (cs.CL)
备注: Accepted to EMNLP 2026 (Main Conference)
Abstract:Reasoning structures (e.g., chains and graphs) are increasingly used for probabilistic inference, factuality verification, and uncertainty estimation, but their structural reliability is rarely evaluated directly. We test whether this reliability holds through a proposed two-stage diagnostic evaluation in open-ended multi-hop QA. Stage 1 tests different capabilities required to construct reasoning graphs, showing that graph reliability varies with topology, prompt design, and model family. Stage 2 uses uncertainty estimation as a downstream test case. Generated-edge experiments show that graph-based computation is sensitive to upstream graph errors, with performance dropping relative to the gold-graph setting. Gold-graph experiments further show that graph structure provides little benefit when used only as prompt context, but improves the separation of correct from incorrect answers by confidence score and enables node-level diagnostics.
[NLP-105] Predicting Linear Relational Decodability Without the Relation
【速读】： 该论文旨在解决线性关系嵌入（Linear Relational Embeddings, LREs）在不同语义关系中保真度（faithfulness）差异显著的问题，即某些关系如person_father的保真度接近于零，而country_capital_city等关系则接近理论上限。其核心问题是：为何同一语言模型在不同关系上的线性可解性（linear decodability）存在如此大的差异？论文的关键解决方案在于提出一种新型的雅可比透镜（Jacobian lens）——实体透镜（Entity Lens, EL_t），该方法通过训练一个不依赖具体关系信息的读出头，来预测特定类型实体（如国家、人物）在中性提示“is known for”下的输出表现。实验表明，仅基于实体类型而非关系特异性信息所计算出的得分，与LRE保真度之间存在高度相关性（ρ = 0.7355, p < .01），揭示了线性可解性主要取决于主体表示本身所携带的信息，而非目标关系本身的特性。这一发现表明，现有关系数据集存在严重缺陷，如首词语义弱化（例如多数预测高校名称以“University”开头）和参照竞争（如模型混淆同姓非演员实体），因此亟需构建更精细的事实级数据集及分析方法。
链接: https://arxiv.org/abs/2610.27241
作者: Jonas Rohweder,Adrian Mülthaler,Farid Mehri,Sebastian Gerstner,Hinrich Schütze
机构: LMU Munich (慕尼黑大学); TU Darmstadt (达姆施塔特工业大学); Munich Center for Machine Learning (慕尼黑机器学习中心); relAI – Konrad Zuse School of Excellence in Reliable AI
类目: Computation and Language (cs.CL)
备注: 19 pages
Abstract:Linear relational embeddings (LREs) approximate a language model’s subject-to-object mapping with an affine function. Their faithfulness varies enormously across relations, from near zero for person_father to the ceiling for country_capital_city. We investigate the factors that contribute to this variation and demonstrate that it can be significantly predicted from a readout that is not aware of the underlying relation. To do so, we propose a novel form of the Jacobian lens, the entity lens. Whereas the Jacobian lens is estimated across all tokens and concepts, the entity lens EL_t is trained to predict what the model will output for an entity of type t under the neutral prompt " is known for"; e.g., we train ELs for countries and persons. After estimating an entity lens, we measure how highly it ranks the model’s answer for a relation without using any relation-specific information in estimation and prediction. Across 28 relations of the LRE dataset on Olmo-3-7B this relation-agnostic score correlates with LRE faithfulness at \rho = 0.7355 (p .01). This suggests that linear decodability is largely a property of what a subject’s representation already carries, rather than of the relation being decoded. Moving the unit of analysis from the relation to the individual fact, we conclude that new relation datasets and analysis methods are needed to address several problems in existing datasets, most prominently semantically weak first tokens (e.g., most predicted universities’ first token is “University”) and reference competition (e.g., the LLM confuses obscure actor names with competing non-actor entities that have the same last name). Our full pipeline is available at this https URL
[NLP-106] CEFR-oriented Russian Text Simplification Using Reinforcement Learning
【速读】： 该论文旨在解决俄语文本自动简化过程中对语言复杂度控制不足的问题，特别是针对欧洲共同语言参考框架（CEFR）中的A1、A2和B1三个语言水平的适配问题。其核心解决方案在于通过基于偏好优化（preference-based optimization）算法，结合拒绝采样（Rejection Sampling）与直接偏好优化（Direct Preference Optimization, DPO）方法对YandexGPT模型进行微调，以增强生成文本在词汇、语法结构和句法复杂度上的可控性。关键创新点在于采用不同数据集构成、文本类型及偏好对构建策略的系统性对比实验，并引入基于ruRoBERTa-large的回归模型实现连续复杂度评分预测，辅以语义相似性、词汇覆盖率和可读性指标进行自动评估，同时通过人工评估验证简化后的文本在简洁性、流畅性和恰当性方面的表现。实验结果表明，最优模型在目标水平准确率达87.15%，平均绝对误差（MAE）仅为0.35个CEFR等级，且在目标区间内的MAE低至0.11；人工评估显示75%的简化文本符合目标形态与句法水平，尤其在A2和B1级别上流畅性与恰当性表现优异。尽管A1级别仍具挑战性，但模型显著降低了超出目标水平的词汇与语法结构比例，同时有效保留原文语义。该模型已开源发布于Hugging Face平台。
链接: https://arxiv.org/abs/2610.27237
作者: Kristina Mustakova,Alexey Sorokin
机构: MSU Center of Artificial Intelligence(莫斯科国立大学人工智能中心); Yandex(雅虎)
类目: Computation and Language (cs.CL)
备注:
Abstract:We present a study on controlled automatic simplification of Russian texts using preference-based optimization algorithms with a focus on adapting texts to the A1, A2 and B1 proficiency levels. We fine-tune YandexGPT using Rejection Sampling and Direct Preference Optimization (DPO) methods to improve control over the complexity of the generated text. We compare several training configurations that differ in dataset composition, text type and the strategy used to construct preference pairs for DPO. For automatic evaluation we train a regression model based on ruRoBERTa-large to predict continuous complexity scores corresponding to CEFR levels. We also consider semantic similarity, lexical coverage and quantitative readability metrics. Human evaluation assesses simplicity, fluency and adequacy. Our best model achieves 87.15% target level accuracy, an MAE of 0.35 CEFR levels and a target interval MAE of 0.11. Manual assessment shows that 75% of the evaluated simplifications match the target morphological and syntactic level, and fluency and adequacy scores are especially high for A2 and B1. Although A1 remains the most challenging level, the model significantly reduces the proportion of vocabulary and grammatical constructions beyond the target level while generally preserving meaning. Our model is freely available on \hrefthis https URLHuggingFace .
[NLP-107] MMedAgent -Plus: Scaling Multimodal Medical Agents across Modalities Tasks and Workflows
【速读】： 该论文旨在解决医疗工作流程中多任务、多模态分析之间缺乏有效协同的问题，现有专用型（specialist）与通用型（generalist）AI模型难以实现复杂医疗任务链的自动化调度与整合。其核心解决方案是提出MMedAgent-Plus，一个基于视觉-语言规划器（vision-language planner）的多模态医疗工具使用代理，通过统一工具接口协调18个覆盖12项任务和8种模态的领先专用模型，实现任务分解、顺序调用及结果融合。关键创新在于：（1）采用大规模指令微调（208,928样本，来自32个公开医学基准），使规划器具备跨轮次、单/多工具交互的理解与决策能力；（2）建立端到端的工具编排（tool orchestration）、响应合成（response synthesis）与执行评估体系，验证其在真实场景下的可靠性。实验表明，该系统在单工具多轮、多工具多轮、多工具单轮等不同设置下工具调用准确率分别达91.0%、98.1%、80.7%，且在正确工具输出前提下，生成答案在忠实性、完整性、简洁性上得分高达14.4–14.8/15；在无预设工具输出的端到端测试中，成功调用目标专家率达90.6%，最终答案准确率高达0.98，与专用最优模型持平并显著优于最强基线（0.28）。
链接: https://arxiv.org/abs/2610.27233
作者: Ruoxin Wang,Yuanting Pan,Binxu Li,Ruiyang Ji,Yaqian Chen,Jiayuan Ding,Qifan Wang,Lifu Huang,Mingjie Li,Yu Zhang,Ehsan Adeli,Kilian Pohl,Zhe Xu,Zihao Lin,Haoyu Dong,Yixin Wang
机构: Duke University (杜克大学); Stanford University (斯坦福大学); Yale University (耶鲁大学); Princeton University (普林斯顿大学); Virginia Tech (弗吉尼亚理工学院); Hippocratic AI; Meta; UC Davis (加州大学戴维斯分校); Shanghai Jiao Tong University (上海交通大学); Tongji University (同济大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 66 pages, 28 figures
Abstract:Medical workflows often require multiple interdependent analyses, but existing specialist and generalist AI models are not designed to coordinate them. We introduce MMedAgent-Plus, a multimodal medical tool-use agent that combines a vision-language planner with 18 leading specialist models spanning 12 tasks and 8 modalities. Through a unified tool interface, the planner selects specialists, executes them in sequence, and synthesizes responses grounded in their outputs. The planner is instruction-tuned on 208,928 samples from 32 public medical benchmarks, covering single- and multi-tool interactions across single and multiple rounds. We separately evaluate tool orchestration, response synthesis, and end-to-end execution on held-out samples. MMedAgent-Plus achieves tool-call accuracies of 91.0%, 98.1%, and 80.7% in single-tool multi-round, multi-tool multi-round, and multi-tool single-round settings, respectively. Given correct tool outputs, MMedAgent-Plus achieves scores of 14.4-14.8 out of 15 for faithfulness, completeness, and conciseness across 600 evaluation samples. In end-to-end execution without pre-supplied tool outputs, MMedAgent-Plus correctly invokes the required specialist in 90.6% of 350 evaluation cases spanning seven tasks. Its final-answer accuracy reaches 0.98, matching the dedicated leading specialist and exceeding the strongest baseline (0.28).
[NLP-108] Which Agents Skill Should Change? Group-Relative Credit Assignment for Evolving Multi-Agent Teams
【速读】： 该论文旨在解决协作式语言模型智能体在无需权重更新的情况下，通过修订可复用技能以实现性能提升时所面临的行为责任归属不准确问题。现有方法仅依据轨迹中技能的“存在性”进行信用分配，导致有害技能在成功任务中仍被错误地赋予正向信用，从而阻碍有效学习。其解决方案的关键在于提出GRSE（Group-Relative Skill Evolution），一种基于相对比较的行为信用分配机制：受GRPO启发，GRSE在相同技能库下多次执行同一问题，并通过对比不同执行结果来精确识别并分配每个智能体-技能组合的责任，结合结构化失败诊断实现更精准的信用评估。实验表明，GRSE在HotpotQA和LoCoMo基准上显著优于当前最接近的方法MASkills，尤其在多跳问答任务中取得最大提升（F1分别提升+6.7和+3.2），且性能提升源于更可靠的检索能力而非无证据回答，同时所学技能具备良好的跨任务迁移性。此外，GRSE对受损技能库具有强鲁棒性，能够正确识别并纠正恶意植入技能的信用方向，最终恢复至清洁库水平（+6.3分），而MASkills则出现误判且性能下降。
链接: https://arxiv.org/abs/2610.27227
作者: Yanan Ma,Yang Wang,David Le,Chenhan Yuan,Yizheng Sun,Guangwei Zhang,Chenghao Xiao,Chenghua Lin,Imon Banerjee
机构: Mayo Clinic(梅奥诊所); University of Manchester(曼彻斯特大学); Alibaba Token Hub, Alibaba Group(阿里巴巴令牌中心，阿里巴巴集团); Shanghai University of Finance and Economics(上海财经大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Cooperative language-model agents can improve without weight updates by revising reusable skills, but only if a shared team reward is traced to the responsible agent and skill. Existing methods credit the skills present in each trajectory, yet presence is not responsibility: a harmful skill active in successful runs still earns positive credit. We present \textbfGRSE, \textbfGroup-\textbfRelative \textbfSkill \textbfEvolution, which assigns behavioral credit by comparison: inspired by GRPO, it executes each question several times under the same skill libraries and contrasts the executions to credit agent–skill pairs, with structured diagnosis of failures. On HotpotQA and LoCoMo with two open-weight actors, GRSE outperforms MASkills, the closest prior method, on HotpotQA and every LoCoMo category except open-domain, with the largest gains on multi-hop questions ( +6.7 and +3.2 F1). Its gains rest on better-grounded retrieval rather than answers given without evidence, and its learned skills transfer between the multi-hop tasks. GRSE is robust to corrupted libraries: starting from planted harmful skills, it credits every planted skill with the correct sign and recovers the score of a clean library ( +6.3 points), whereas MASkills misattributes two and falls below its starting point.
[NLP-109] Profile-DP: Differential Privacy for User-Controlled Profile Release in Personalized LLM Inference
【速读】： 该论文旨在解决个性化大语言模型（LLM）在推理过程中因依赖用户画像（profile）而引发的隐私泄露问题，特别是用户未主动披露但可能通过模型推断出的敏感信息。现有方法通常将用户画像作为整体公开，导致用户隐含偏好与私有属性暴露风险升高。为此，论文提出Profile-DP——一种面向用户画像发布的政策感知局部差分隐私（local differential privacy, LDP）框架，其关键在于：将用户显式授权的事实、公共指令及画像结构视为可公开的辅助信息，仅对用户隐藏的属性施加严格隐私保护；通过在生成回答前将公共指令与经过差分隐私处理的画像进行组合，实现隐私保障与下游模型行为的解耦。该框架设计了两种满足Profile-DP的有限域机制：DP-SLOT独立扰动受保护属性，适用于保留授权事实；DP-BLOCK则利用公共语义决策协调相关属性的释放，联合采样以增强一致性并维持整体隐私边界。实验表明，在PEEP数据集上，DP-BLOCK在13个结构化预算设置下均取得最优综合效用，并在26种格式-预算配置中更接近理想隐私-效用平衡点，且直接测量显示其在降低披露量方面优于其他方法，而DP-SLOT则更擅长保留授权事实，二者互补性强。该研究为个性化与隐私保护的协同优化提供了可实践的技术基础。
链接: https://arxiv.org/abs/2610.27213
作者: Mingchen Li,Zifan Peng,Kewei Sha,Lavinia F Pieptea,Heng Fan,Junhua Ding,Meikang Qiu,Yunhe Feng
机构: University of North Texas(北德克萨斯大学); Newcastle University(纽卡斯尔大学); Augusta University(奥古斯塔大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Personalized large language models (LLMs) increasingly rely on profile-conditioned inference, pairing public requests with rich, user-specific facts. This paradigm exposes the user profile itself, together with the user’s explicit sharing preferences. We introduce Profile-DP, a policy-aware local differential privacy framework designed specifically for user-profile release in personalized LLM inference. Profile-DP provably bounds the release of user-withheld facts while treating public instructions, profile schemas, and authorized facts as public side information. By generating a privatized profile that composes with the public instruction prior to answer generation, our framework decouples formal privacy guarantees from downstream model behavior. We develop two finite-domain mechanisms satisfying Profile-DP: DP-SLOT independently perturbs protected attributes, while DP-BLOCK uses public semantic decisions to coordinate their release and jointly sample related attributes. A positive, public compatibility prior connects these attributes while preserving a profile-level privacy bound. On PEEP, DP-BLOCK leads the utility composite among DP methods at all 13 structured budgets and is closest to the ideal privacy–utility point in 18 of 26 format–budget settings under our composite weighting. Direct measurements reveal complementary strengths: DP-BLOCK discloses less, while DP-SLOT retains more authorized facts. These results make policy-aware profile release a practical basis for studying privacy and personalization.
[NLP-110] LoWAgent : A Low-Resource Multilingual Benchmark for Web Agents
【速读】： 该论文旨在解决当前网页智能体（Web Agent）评估体系严重依赖英语界面，导致其在多语言环境下尤其是低资源语言（Low-Resource Languages, LRLs）中的能力评估缺失的问题。现有评估基准普遍缺乏对非英语语言环境的支持，尤其在印地语、孟加拉语、乌尔都语和阿姆哈拉语等低资源语言上的标准化评测资源极为匮乏。为弥补这一空白，本文提出低资源网页智能体基准测试（Low-Resource Web Agent Benchmark, LoWAgent），首次系统性地构建了一个涵盖英语及四种低资源语言的跨语言评估框架。该基准包含六个完全本地化的网站，在五种语言下共生成3,000个任务实例，覆盖多类型网页交互场景。研究评估了三款闭源与五款开源模型在全部语言和网站上的表现，结果表明：无论是闭源还是开源模型，在低资源语言上的性能均显著低于英语，且开源模型的性能下降更为剧烈。此外，研究还考察了真实界面干扰（如Cookie提示和广告弹窗）对智能体鲁棒性的影响，发现此类干扰会进一步削弱任务完成能力。最后，实验验证了在少量LoWAgent数据上进行微调可显著提升模型在低资源语言中的任务执行效果，证明了该基准在推动多语言智能体发展方面的实用价值。其核心解决方案在于构建一个具有代表性、可扩展且贴近实际应用的多语言评估体系，从而为低资源语言环境下智能体的能力提升提供可靠评估工具与优化路径。
链接: https://arxiv.org/abs/2610.27205
作者: Ocean Monjur,Daud Ahmad Nisar,Shahriar Kabir Nahin,Tadesse Destaw Belay,Sushruti Mishra,Israel Abebe Azime,Anshuman Chhabra
机构: University of South Florida, USA(南佛罗里达大学, 美国); Instituto Politécnico Nacional, Mexico(墨西哥国立理工学院, 墨西哥); Saarland University, Germany(萨尔兰大学, 德国); Independent Researcher(独立研究员)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Web agents are typically evaluated on English-language interfaces, leaving their ability to operate across the world’s multilingual websites largely unexplored. This gap is especially significant for low-resource languages, where standardized resources for evaluating web agents remain limited. To address this gap, we introduce Low-Resource Web Agent Benchmark (LoWAgent), a benchmark for evaluating web agents across different low-resource languages (LRLs) in comparison to the default, English. LoWAgent evaluates web agents across English and four LRLs: Hindi, Bengali, Urdu, and Amharic. The benchmark comprises of six fully localized websites, instantiated across five languages to yield 3,000 task instances. We evaluate three closed-source and five open-source models across all five languages on all six websites. Our findings show that both closed-source and open-source models experience performance degradation for LRLs compared to English, with substantially greater performance drops observed among the open-source model agents. We further evaluate the robustness of the models on realistic interface disruptions, such as cookie notices and advertisement pop-ups, and find that these further degrade agentic performance. Finally, we show that post-training on a small subset of LoWAgent traces can significantly improve low-resource language agentic task performance, demonstrating the usefulness of our benchmark.
[NLP-111] he Culture Within: Towards Mechanistic Understanding of Cultural Reasoning in LLM s
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在跨文化场景中对文化信息的处理机制不明确的问题，尤其是现有评估方法仅关注输出层面的文化表现，而忽视了模型内部如何动态利用文化背景信息。其核心挑战在于：当模型面临与自身默认文化倾向相冲突的规范（contradictory-norm）时，即使具备正确答案所需的知识，仍会显著降低符合文化一致性的回答比例。解决方案的关键在于提出一个基于参考的基准测试工具——CultureLens，该工具通过系统性地操控国家、价值观、真实规范（true-norm）和矛盾规范（contradictory-norm）等变量，在75个国家的9,261个情境下进行评估。研究发现，模型在未指定国家时默认采用盎格鲁-撒克逊文化规范；当出现矛盾规范时，模型并非因知识缺失而失败，而是过度依赖规范的呈现方式而非其真实性，表现出对检索内容的高度敏感性。通过引入包含矛盾规范示例的监督微调（supervised fine-tuning），可有效提升模型在矛盾规范下的文化一致性表现至85%-93%，且在未见国家上仍具泛化能力，同时对真实规范下的表现影响极小。
链接: https://arxiv.org/abs/2610.27202
作者: Samee Arif,Furkan Danisman,Angana Borah,Honglak Lee,Dehan Kong,Lajanugen Logeswaran,Moontae Lee,Zhijing Jin,Rada Mihalcea
机构: University of Michigan; University of Toronto; LG AI Research
类目: Computation and Language (cs.CL)
备注:
Abstract:Large Language Models (LLMs) increasingly operate across culturally diverse settings, yet existing evaluations largely measure cultural behavior at the output level. We introduce CultureLens, a reference-grounded benchmark and mechanistic study of how LLMs use cultural information. CultureLens contains 9,261 scenarios across 75 countries and systematically varies country, value, true-norm, and contradictory-norm evidence. Across Llama-3.1-8B, Gemma-3-12B, and Qwen-3.5-27B, stating a contradictory norm lowers aligned answers by 32-47 percentage points, even when the model answers correctly from the country alone. Mechanistically, models default to Anglophone norms when no country is given. The failure under a contradictory norm is not one of missing knowledge, as blocking the norm inside the model restores the correct answer in most cases. Models accept a stated norm according to how it is presented and not whether it is true, which leaves them most exposed to errors in retrieved content. Supervised fine-tuning with contradictory-norm examples raises aligned answers under a contradictory norm to 85-93%, including on unseen countries, at a small cost under a true norm.
[NLP-112] Preference-Aware Pareto Steering for Multi-Task Speech LLM s
【速读】： 该论文旨在解决语音大语言模型（Speech LLM）在联合训练过程中因多任务间存在权衡（task trade-offs）而导致的性能不可控问题。传统采用固定损失权重的训练方法难以根据实际部署需求灵活调节各任务之间的平衡，从而限制了模型在不同应用场景下的适应性。为此，本文提出了一种先验性（a priori）方法——端点锚定偏好感知均匀采样沿帕累托前沿（Endpoint-Anchored Preference-Aware Sampling Uniformly along the Pareto Front, EA-SURF），其核心在于动态调整标量化的损失权重，使最终的多任务权衡结果能够精确匹配用户指定的偏好。该方法保留了标准加权损失训练框架，通过持续跟踪模型在各任务上的表现相对于任务特定参考点的偏差，并据此自适应地更新权重。理论上，当满足可验证的正则性条件时，EA-SURF 的权重更新过程将收敛至目标偏好点；实验表明，在六组三目标语音大语言模型任务中，EA-SURF 能够有效逼近预设偏好，并相较固定损失权重方法提升了17.8%的包含端点的归一化超体积（endpoint-inclusive normalized hypervolume），显著增强了多任务优化的灵活性与实用性。
链接: https://arxiv.org/abs/2610.27200
作者: Liuyuan Jiang,Vishal Sunder,Samuel Thomas,Brian Kingsbury,Jeff Kuo,Lisha Chen,George Saon
机构: 未知
类目: Computation and Language (cs.CL); Sound (cs.SD)
备注:
Abstract:Speech large language models (LLMs) perform multiple tasks, including automatic speech recognition (ASR), timestamp prediction (TS), and speech translation (ST). Joint training of speech LLMs creates task trade-offs, yet standard training with fixed loss weights does not generally control the resulting trade-off according to deployment preferences. To address this, we propose Endpoint-Anchored Preference-Aware Sampling Uniformly along the Pareto Front (EA-SURF), an a priori method that adapts scalarization weights so that the resulting task trade-off reflects the desired preference. EA-SURF retains the standard weighted-loss training pipeline, tracking the current model’s performance relative to each task-specific reference and adapting the scalarization weights as training proceeds. We theoretically show that EA-SURF can match a desired preference, with the update becoming stationary at the target under checkable regularity conditions. Across six three-objective speech LLM requests, EA-SURF empirically converges toward the requested targets and increases endpoint-inclusive normalized hypervolume by 17.8% over fixed loss weighting.
[NLP-113] On the Frag ility of Tasks under Composition: Implicit Budgeting in Language Models
【速读】： 该论文旨在解决当前大型语言模型在单个提示（prompt）中同时处理多个任务时出现的性能退化问题。尽管模型在独立执行每个任务时表现可靠，但在任务组合场景下却显著增加错误率，这种退化无法完全由各任务独立错误的累积解释。其关键发现在于：模型存在一种隐式的“思考预算”（implicit thinking budget）机制，即在任务组合时，模型为每个任务分配的生成令牌数（tokens per task）显著减少，即使整体上下文长度远未达到模型的上下文窗口限制。这导致需要较长推理路径的任务在组合场景中尤为脆弱，错误率更高。而原本在孤立任务中过度推理的非脆弱任务则存在计算浪费。通过根据任务的脆弱性进行分组并优化推理资源分配，可实现跨数学与语言类任务43%的令牌消耗降低，同时保持原有准确率。因此，解决方案的关键在于识别并利用任务脆弱性差异，主动管理模型的隐式思考预算以提升效率与鲁棒性。
链接: https://arxiv.org/abs/2610.27190
作者: Chenpan Li,Thies Brockmoeller,Stefanie Jegelka,Sharut Gupta,Nikolaos Karalias
机构: TU Munich (慕尼黑工业大学); MIT CSAIL (麻省理工学院计算机科学与人工智能实验室)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Modern language models are increasingly asked to handle several tasks within a single prompt. Doing so requires solving each task correctly while integrating the resulting answers into a coherent response. In this paper, we identify a fundamental limitation in this process, where models make substantially more errors when tasks are composed, even when they can solve every constituent task reliably in isolation. This degradation cannot be fully explained by the accumulation of independent errors. Across both open-weight models and the frontier model Claude Opus 5, our findings instead point to an implicit thinking budget underlying this effect. In particular, under composition, models use substantially fewer tokens per task than in isolation, even when the full sequence remains well within the context window. Tasks that require longer solution traces in isolation are consequently more \textitfragile under composition, i.e., models make more errors on prompts containing these tasks. In isolation, non-fragile tasks receive more reasoning than they need and waste compute. However, by grouping tasks according to their fragility, the implicit budgeting can be manipulated to reduce token usage by 43 % across a range of math and linguistic tasks while also preserving accuracy.
[NLP-114] SITO: Sifting a Commercial Text-Safety API for Missed Harm and False Flags in English Ukrainian and Surzhyk
【速读】： 该论文旨在解决当前面向青少年用户的平台在使用商业文本安全API（如Tuteliq Detect）进行内容筛查时，缺乏独立实证数据以评估其误报率（false-positive rate）与漏报率（false-negative rate）的问题，尤其在非英语语境下的表现尚不明确。研究的关键在于构建了一个名为SITO（Synthetic Independent Text-safety benchmark with Overflagging tests）的多语言基准测试集，包含2,332条合成文本案例，涵盖六类危害行为（授权支付诈骗、资金中介招募、社交工程、情感诈骗、赌博伤害及欺凌），每条有害文本均匹配难以区分的“硬性无害”对照样本（同主题、最小编辑差异、警告语句及事后陈述）。实验结果表明，该API在英文文本上的有害内容召回率为84.9%-94.1%，误报率仅为0%-4.3%；然而，在未启用特定报告处理设置的情况下，对第三方举报的识别召回率（63.6%）显著低于对施害者原始消息的召回率（95.1%），且在乌克兰语及混合语言输入中虽检测率不低于英文，但来自保留生成器家族的文本召回率下降，提示整体性能可能存在乐观估计。此外，3.2%的测试案例在重复运行中出现决策不一致，凸显模型稳定性问题。因此，研究强调平台应基于自身流量独立评估误报率，并审查旁观者举报的处理机制。相关数据集、生成提示与代码均已开源。
链接: https://arxiv.org/abs/2610.27178
作者: Andrii Tuzhykov
机构: The State Institution South Ukrainian National Pedagogical University named after K. D. Ushynsky (南乌克兰国立师范大学，以K. D. 乌申斯基命名); IMPROVEIT SOLUTIONS SPÓŁKA Z OGRANICZONĄ ODPOWIEDZIALNOŚCIĄ (“ImproveIT Solutions”)
类目: Computation and Language (cs.CL)
备注: 14 pages, 1 figure. Data and code: this https URL
Abstract:Platforms that host young users screen messages with commercial text-safety APIs, but independent evidence on how often such systems miss harm or flag harmless text is scarce, especially outside English, and the documentation of the one studied here leaves its false-positive rate unquantified. We evaluated one such API, Tuteliq Detect as deployed under the release identifier tuteliq-detect-v5.20261011 (not the separately published v6 model), on six harm categories: authorised push-payment fraud, money-mule recruitment, social engineering, romance scams, gambling harm and bullying. We built SITO (Synthetic Independent Text-safety benchmark with Overflagging tests; sito is Ukrainian for ‘sieve’), 2,332 synthetic cases in English, Ukrainian and mixed Ukrainian, with each harmful case matched by hard harmless ones: same-topic messages, minimal-edit contrasts, warnings and retrospective accounts. On English text the system flagged 84.9-94.1% of harmful cases and 0-4.3% of harmless ones per category. It missed third-party reports of ongoing harm far more often than perpetrators’ messages (63.6% against 95.1% recall, all languages), when sent without the settings the API offers for reports of harm. Ukrainian and mixed-language input was detected at least as often as English, recall was lower on text from a held-out generator family, so headline recall may be optimistic, and 3.2% of cases received a different decision in repeated runs. A platform should measure false positives on its own traffic and check how bystander reports are handled. The datasets, generation prompts and code are released.
[NLP-115] GRID: Modeling Health Influencer Discourse as Argument Networks for Credibility Assessment
【速读】： 该论文旨在解决社交媒体上健康类内容评估中普遍存在的孤立判断问题，即当前对意见领袖（influencer）发布内容的评价往往脱离其所在的话语语境，未能充分考虑多方观点之间的互动与关联。其核心挑战在于如何在复杂且相互竞争的舆论环境中，系统性地评估内容的可信度。解决方案的关键在于提出一种名为GRID（基于互联话语的根基性推理）的框架，该框架通过构建多视频间的论证图谱（argumentation graph），将每个视频中的主张与其核心观点（takeaway）相连接，并跨视频建立主张与观点之间的关联，同时将主张与科学证据进行锚定，最终在整合后的网络结构上实现联合可信度推断。实验结果表明，引入更广泛话语背景显著提升了对意见领袖内容的验证效果；进一步分析显示，高可信度与低可信度内容在论证策略和道德框架上存在显著差异。
链接: https://arxiv.org/abs/2610.27149
作者: Nishanth Sridhar Nakshatri,Eylon Caplan,Dan Goldwasser
机构: Purdue University (普渡大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Health-related discussions on social media are increasingly shaped by influencers who offer diverse, and often competing perspectives on the same issue. Yet, such content is typically evaluated in isolation, without considering the broader discourse in which it appears. We introduce GRID (Grounded Reasoning over Interconnected Discourse), a framework for evaluating influencer content through collective inference over the surrounding discourse. GRID represents each video as an argumentation graph connecting claims to its central message, or takeaway, links related claims and takeaways across videos, grounds claims in scientific evidence, and jointly infers credibility over the resulting network. We evaluate GRID on an end-to-end health-related content validation task using 569 influencer videos grounded in expert assessments. Our results show that accounting for the broader discourse improves influencer content validation. Finally, we use GRID-inferred credibility to demonstrate how argumentative strategies and moral framing differ between low- and high-credibility content.
[NLP-116] ICLHerd: Herding-Based Data Selection for LLM s with In-Context Learning as a Proxy
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）训练中数据选择存在的模式聚焦偏差（mode-seeking bias）问题。现有方法通常通过计算模型隐藏状态或梯度的余弦距离来选择与开发集相似的数据，但这种最大相似性策略会导致所选数据过度集中在开发样本附近，从而限制了数据分布的多样性与泛化能力。为克服这一缺陷，论文提出ICLHerd框架，其核心创新在于引入冯·米塞斯-费舍尔核（von Mises-Fisher kernel），在模式聚焦与模式平均之间实现有效平衡，提升数据选择的多样性与代表性。此外，该方法引入上下文学习（in-context learning）作为验证代理，实现了对平衡参数的高效超参数搜索。实验结果表明，ICLHerd显著优于现有的LLM数据选择算法，验证了在数据选择中兼顾模式聚焦与模式平均的重要性。
链接: https://arxiv.org/abs/2610.27142
作者: Yuqiao Wen,Lili Mou
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Data selection plays a central role in modern large language model training, as the quality of training data strongly influences downstream performance. A common approach is to curate a development set that encodes the desired skills (e.g., reasoning or multilinguality) and select data that are similar to them. Existing methods compute similarity using cosine distance over either the model’s hidden states or gradients. However, we observe that this max-similarity strategy induces a mode-seeking bias: the selected data clusters tightly around the development samples. To address this, we propose ICLHerd, a data selection framework that leverages the von Mises-Fisher kernel to balance mode-seeking with mode-averaging. We further introduce in-context learning as a validation proxy, enabling efficient hyperparameter search over the balance trade-off. Results show that ICLHerd outperforms existing LLM data selection algorithms, highlighting the importance of balancing mode-seeking and mode-averaging.
[NLP-117] Beyond Two Tracks: Speaker-Anchored Full-Duplex Dialogue Recovery ICASSP2027
【速读】： 该论文旨在解决全双工对话（full-duplex conversation）中多说话人语音分离与说话人身份一致性维护的难题，核心挑战在于：在持续对话过程中，传统语音分离模型生成的语音流（coarse streams）虽能区分两个说话人，但其说话人标签易随时间发生漂移（track drift），导致身份不一致，难以支持需要持久说话人追踪的自然交互系统。解决方案的关键在于提出一种说话人锚定路由机制（speaker-anchored routing），该机制在扩散模型（diffusion）前阶段引入一个仅1.84M参数的轻量级模块，通过将分离器输出的粗粒度语音流与预先注册的说话人身份（enrollment identities）进行匹配并固定排序，从而在不重新训练生成器或改变重建音频的前提下，实现说话人轨迹的稳定绑定。实验表明，在25分钟远程对话场景中，该方法将错误跟踪时间从43.9%降至4.1%，显著优于匹配注册波形重识别的6.2%基准，尤其在单方发言时优势更为突出（17.8% vs 28.6%）。同时，固定轨迹活动的F1值由0.46提升至0.84，而每段重建质量保持不变。该路由模块具备良好泛化能力，可迁移至未见的面对面及电话录音场景，并结合边缘门控验证（margin-gated verification）进一步将错误跟踪时间压缩至0.9–1.7%，同时仅嵌入20–36%的嵌入信息，远低于全重识别所需的100%开销。由此，恢复后的语音流成为具有参与者一致性的稳定轨道，为全双工对话建模提供了关键支撑。
链接: https://arxiv.org/abs/2610.27109
作者: Yuxuan Chen,Yang Xiao,Jack Sun,Chak Sun,Brick Li, Sam
机构: 未知
类目: ound (cs.SD); Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
备注: 5 pages, 3 figures, 2 tables. Submitted to ICASSP 2027. Audio examples: this https URL
Abstract:Spoken interaction is moving from alternating turns to full-duplex conversation, where an agent listens while speaking. Learning this coordination requires natural dialogue with persistent speaker-resolved tracks, a data requirement that remains difficult to satisfy at scale. Generative recovery offers a route from ordinary recordings, but recovering two voices is not enough: their track assignments can drift as the conversation unfolds. We address this missing link with speaker-anchored routing before diffusion: a 1.84M-parameter module matches the separator’s coarse streams to fixed enrollment identities and orders them before a frozen generator reconstructs the speech, without retraining the generator or altering the reconstructed audio. On 25-minute remote conversations, routing cuts wrong-track time from 43.9% to 4.1%, below the 6.2% of matched-enrollment waveform re-identification, with its largest advantage when one participant holds the floor (17.8% versus 28.6%), and raises fixed-track activity F1 from 0.46 to 0.84 while per-chunk reconstruction quality is unchanged. The same router transfers to unseen face-to-face and telephone recordings, and a margin-gated verification step brings their wrong-track time to 0.9-1.7% while embedding 20-36% of the output, where full re-identification embeds all of it. Recovered voices thus become the participant-consistent tracks that full-duplex dialogue modeling needs. Audio examples: this https URL
[NLP-118] Omni-Woodpecker: Hallucination Correction via Evidence-Prior Rebalancing for Omni LLM s
【速读】： 该论文旨在解决生成式多模态大模型（Omni LLMs）在联合音视频推理中普遍存在的跨模态幻觉（cross-modal hallucinations）问题。现有方法主要聚焦于输出层面的修正，未能有效处理融合表示中感知证据（grounding evidence）与先验知识（learned priors）之间的纠缠问题。其解决方案的关键在于提出一种无需训练的框架——Omni-Woodpecker（OWP），将幻觉缓解机制从输出层干预转变为基于冲突条件的表示重平衡。具体而言，OWP首先通过互补模态特异性视角识别跨模态冲突；随后利用正交分解技术，从隐藏状态中分离出真实感知证据与竞争性先验；最后，在解码前自适应地抑制误导性先验并恢复被忽略的证据，实现对融合表示的动态重平衡。实验表明，该方法在多个基准上均显著提升性能，并具备强大的跨架构泛化能力，尤其在以Qwen2.5-Omni-7B为基座模型时，在CMM和AVHBench上分别提升12.9和8.6分，且在四类通用多模态推理基准上稳定增益0.67–2.25分，同时在VideoLLaMA2-AV上亦取得4.6和3.9分的改进，验证了其有效性与普适性。
链接: https://arxiv.org/abs/2610.27103
作者: Zijian Liu,Chaoyou Fu,Yan Yang,Caifeng Shan,Ran He
机构: Nanjing University (南京大学); University of Chinese Academy of Sciences (中国科学院大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: 10 pages, 4 figures, 6 tables. Code: this https URL
Abstract:Recent Omni LLMs (OLLMs) have demonstrated remarkable capabilities in joint audio-visual reasoning, yet cross-modal hallucinations remain a major challenge. Existing mitigation methods mainly operate at the output level, overlooking the entanglement of grounding evidence and learned priors within fused representations. To address this issue, we propose Omni-Woodpecker (OWP), a training-free framework that shifts hallucination mitigation from output-level manipulation to conflict-conditioned representation rebalancing. Specifically, OWP first identifies cross-modal conflicts from complementary modality-specific views. It then employs orthogonal decomposition to separate grounding evidence from competing priors within the hidden states. Finally, OWP adaptively rebalances fused representations by suppressing misleading priors and restoring ignored evidence before decoding. Extensive experiments demonstrate the effectiveness and strong cross-architecture generalization of OWP. Using Qwen2.5-Omni-7B as the backbone, OWP improves performance by 12.9 and 8.6 points on CMM and AVHBench, respectively, while consistently achieving gains of 0.67-2.25 points across four general OLLM reasoning benchmarks. Furthermore, OWP generalizes effectively to VideoLLaMA2-AV, improving performance by 4.6 and 3.9 points on the same benchmarks. Code is publicly available at this https URL
[NLP-119] SciFigBench: A Benchmark for Scientific Method Figure Understanding EMNLP2026
【速读】： 该论文旨在解决科学方法图（scientific method figure）中细粒度内容感知与模块间关系理解难题，现有视觉语言模型（Vision Language Models, VLMs）在处理此类图表时，受限于其密集布局与复杂信息流，难以准确解析细粒度语义及视觉关系。为填补这一研究空白，论文提出SciFigBench基准，专门评估科学方法图理解的核心能力，包括：密集内容感知（dense content perception）、视觉关系理解（visual relation comprehension）和信息流追踪（information-flow tracing）。为此，研究团队从真实科研论文中收集了4,352张科学方法图，并构建了40,670个问答对，通过自动化标注与人工校验确保数据质量。关键解决方案是提出SciFigCue——一种无需训练的视觉提示模块，通过显式化箭头所隐含的依赖关系，将抽象的视觉关联转化为可被模型利用的视觉关系证据，从而增强模型对复杂方法图的推理能力。实验表明，即使在强大多模态模型上，视觉关系理解仍存在显著瓶颈，而SciFigCue在不更新模型参数的前提下，有效提升了多个任务的表现。
链接: https://arxiv.org/abs/2610.27087
作者: Lijia Wu,Zheng Li,Yuxuan Liu,Jingwen Zhang,Penghai Zhao,Xiang Li
机构: Nankai University (南开大学); NKIARI, Shenzhen Futian (深圳福田区人工智能研究院); Li Auto (小鹏汽车)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: Accepted by EMNLP 2026 Main Conference. Code and benchmark resources: this https URL
Abstract:Scientific method figures often provide researchers with a clear visual entry point into a paper’s methodology and core contributions, yet their dense layouts and complex information flows are difficult for Vision Language Models (VLMs) to interpret. Existing benchmarks predominantly focus on coarse-grained tasks, such as context-based retrieval or holistic comprehension, overlooking the fine-grained content and inter-module relations within complex methodological diagrams. To address these overlooked aspects, we introduce Scientific Figure Benchmark (SciFigBench), a benchmark for evaluating core capabilities required for scientific method figure understanding. Specifically, it targets three capabilities: dense content perception, visual relation comprehension, and information-flow tracing. For this purpose, we curate 4,352 scientific method figures from real-world research papers and construct 40,670 QA pairs to operationalize these capabilities as measurable question-answering tasks. All annotations are generated by dedicated pipelines and then manually verified. We further propose SciFigCue, a training-free visual prompting module that externalizes arrow-induced dependencies as visual relational evidence for method-figure reasoning. Experiments across open-source and proprietary VLMs show that even strong models remain substantially limited in visual relation comprehension, while SciFigCue improves performance across several tasks without parameter updates. The code and related resources are available at this https URL.
[NLP-120] ClockMark: Adaptive and Host-Dependent Watermarking for Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）生成过程中流式文本水印嵌入的挑战，即传统水印方法无法适应生成过程中动态变化的下一个词元分布（next-token distribution）。在流式生成中，每个采样得到的词元都会实时影响后续的分布，使得水印嵌入面临“宿主信号不断变化”的问题。为应对这一难题，论文提出的关键解决方案是设计一种融合自适应性与宿主依赖性的新型水印框架——ClockMark。其核心在于两个原则：一是嵌入应具备自适应性，即保留先前采样决策的状态以指导后续选择；二是嵌入应具有宿主依赖性，即利用当前语言模型的输出分布来约束词元选择与状态更新。通过结合自适应与宿主依赖的嵌入机制，并引入消息扩散策略，使完整信息包被编码至每个水印词元中，从而在理论上保证了原始生成分布的保持，维持生成质量；实验证明，ClockMark在12个基线对比下显著提升了对文本编辑的鲁棒性，比特准确率最高较最强基线提升16.08个百分点。
链接: https://arxiv.org/abs/2610.27079
作者: Yupeng Qiu,Han Fang,Ee-Chien Chang
机构: National University of Singapore (新加坡国立大学); University of Science and Technology of China (中国科学技术大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Large language model (LLM) watermarking operates on the next-token distribution that evolves during generation. In streaming generation, each sampled token becomes part of the text and affects subsequent distributions. Embedding therefore proceeds with a changing host, unlike classical watermarking, which adapts embedding to an existing host signal. To accommodate this difference, we formulate a novel embedding framework for LLM watermarking based on two principles. We first argue that embedding should be adaptive, retaining the state of previous sampling decisions to better guide subsequent choices. We further argue that embedding should be host-dependent, using the current LLM distribution to constrain both token selection and state updates. Guided by these principles and the two requirements of distribution preservation and robustness to text editing, we propose ClockMark. It combines adaptive and host-dependent embedding with message spreading, embedding the full payload in each watermarked token. Theoretically, we prove that ClockMark preserves the original LLM sequence distribution, thereby maintaining generation quality. Empirically, extensive experiments against 12 baselines demonstrate improved robustness to text editing, with gains of up to 16.08 percentage points in bit accuracy over the strongest tested baseline.
[NLP-121] Discrete Neural Interlingua for Ontology-Agnostic EHR Modeling
【速读】： 该论文旨在解决电子健康记录（EHR）基础模型在跨医疗机构部署中面临的分词瓶颈问题。由于同一临床概念可能以不同的术语编码、本地标识符、同义词或自由文本形式出现，传统EHR分词器将这些表达视为无关符号，导致模型只能利用有限的可用数据（如在MIMIC-III上仅能使用13%的eICU事件）。其解决方案的关键在于提出一种离散的、与本体无关的分词器MEDIATOR，该分词器能够将任意以文本形式表达的临床概念映射到一个共享的向量量化代码簿中，从而将EHR标准化问题转化为开放词汇分词问题，无需人工映射或固定词汇表。MEDIATOR通过同义词监督、增强的概念描述、代码共现关系及关系图谱进行训练，实现了对等价表达的压缩与不同概念的医学语义有序离散化。作为EHR Transformer的分词接口，MEDIATOR在20项MIMIC-III任务上使平均AUPRC提升4.0点（相对提升8%），在14项eICU零样本迁移任务上提升8.6点（相对提升29%），且可利用全部事件数据。迁移性能的提升源于覆盖率和表示质量的双重优化，且在相同输入条件下，其在操作性任务和亚队列死亡率预测任务上仍优于其他分词器。消融实验表明，文本编码器主导域内准确性，而MEDIATOR所学习的目标函数则驱动迁移鲁棒性和实体链接质量。此外，MEDIATOR还支持跨医院的可转移归因分析。
链接: https://arxiv.org/abs/2610.27047
作者: Tong Ding,Caiwei Tian,Dongmin Bang,Ming Y. Lu,Sophia J. Wagner,Andrew Zhang,Rowland Pettit,Long Phi Le,Faisal Mahmood
机构: Mass General Brigham AI Institute(麻省总医院人工智能研究所); Harvard University(哈佛大学); Massachusetts Institute of Technology(麻省理工学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:EHR foundation models promise scalable clinical prediction across health systems, yet their deployment is constrained by a tokenization bottleneck. The same clinical concept can appear as different ontology codes, local identifiers, synonyms, or free-text descriptions, which standard EHR tokenizers treat as unrelated symbols. A model trained on MIMIC-III with a closed-vocabulary tokenizer can use only 13% of eICU events. We introduce MEDIATOR, a discrete ontology-agnostic tokenizer that maps any clinical concept expressed as text into a shared vector-quantized codebook, turning EHR harmonization into open-vocabulary tokenization without manual mappings or a fixed vocabulary. Trained with synonym supervision and guided by enriched concept descriptions, code co-occurrence, and relation graphs, MEDIATOR collapses equivalent expressions while organizing distinct concepts in a medically meaningful discrete space. Used as the token interface of an EHR Transformer, MEDIATOR improves mean AUPRC over the strongest prior tokenizer by 4.0 points (8% relative) on 20 MIMIC-III tasks and by 8.6 points (29% relative) in zero-shot transfer to 14 eICU tasks, where it uses all events. The transfer gain combines coverage and representation, as MEDIATOR remains the best tokenizer under identical inputs (+1.0 and +1.9 points on operational and subcohort-mortality tasks). Same-encoder controls show that the text encoder drives in-domain accuracy, while the learned objectives of MEDIATOR drive transfer robustness and entity-linking quality. MEDIATOR also enables transferable attribution across hospitals.
[NLP-122] ReuF2F: Benchmarking Coding Agents on Formalized Undergraduate Research Problems
【速读】： 该论文旨在解决当前生成式AI在数学领域求解开放问题时缺乏可靠、可复现的评估基准的问题。现有形式化数学基准多集中于已知结论或少数著名开放问题，一旦这些问题被解决并广泛传播，其评估价值便迅速下降；同时，现有评估协议与现代代码代理（coding agent）的实际运行机制不完全匹配。为此，研究提出ReuF2F，一个基于本科生研究项目（Research Experiences for Undergraduates, REU）的Lean 4形式化开放问题基准，涵盖244个来自多个数学子领域的入门级开放问题。这些相对冷门且较少受到关注的问题更适合作为“保留集”（held-out）评估目标，以持续检验端到端开放问题求解能力。解决方案的关键在于构建一个与现代代码代理工作方式对齐的标准化评估环境：将每个问题及其逻辑否定形式置于Docker容器中的Lean项目中，并开发Leanimum-agent（基于mini-SWE-agent）作为统一评估框架，确保不同模型间的公平比较。实验结果表明，即使是最优模型也只能解决约40%的问题，凸显了在形式化开放研究问题求解方面仍存在巨大提升空间。
链接: https://arxiv.org/abs/2610.27046
作者: Zeyu Zheng,Yiyang Liu,Yuyang Zhao,Congmin Zheng,Jiachen Zhu,Jianian Chen,Kangning Zhang,Kayue Daniel Wong,Lingyu Yang,Rong Shan,Shihao Chen,Thomas Zhu,Wentao Zhang,Xiaopeng Li,Yan Hu,Yongxi Lin,Yuxiang Chen,Zhuoying Ou,Jianghao Lin,Jeremy Avigad,Prasad Tetali,Jingbang Chen,Sean Welleck
机构: Carnegie Mellon University (卡内基梅隆大学); Chinese University of Hong Kong, Shenzhen (香港中文大学深圳校区); Shenzhen Loop Area Institute (深圳湾区研究院); Shanghai Jiao Tong University (上海交通大学); Yau Mathematical Sciences Center, Tsinghua University (清华大学姚期智数学科学中心); University College London (伦敦大学学院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Benchmark available at this https URL
Abstract:As AI systems begin to resolve open research problems, reliable verification becomes increasingly important. Formal mathematics provides a trusted framework for mathematical reasoning, enabling agents to produce machine-checkable proofs while conducting mathematical exploration. Existing formal mathematics benchmarks often focus on known results or a small set of prominent open problems, which can become difficult to evaluate reliably once solutions are discovered and widely disseminated. Existing evaluation protocols are also not always aligned with how modern coding agents operate. We therefore introduce ReuF2F, a Lean 4 benchmark of 244 formalized open problems drawn from Research Experiences for Undergraduates (REU), spanning entry-level open research problems across a broad range of mathematical subfields. We expect these less prominent open problems, which tend to attract less attention and may remain unsolved for longer, to be better suited for held-out evaluation of end-to-end open problem resolution in Lean. We adopt a standard coding-agent evaluation setting, placing each problem and its logical negation into a Lean project in a Docker container. We also build Leanimum-agent on mini-SWE-agent to provide a unified harness for fair comparisons across models. Our evaluation finds that even the best-performing model resolves only ~40% of the problems, leaving substantial room for improvement on formalized open research problems.
[NLP-123] Cross-Lingual Reusability of LLM Pipelines: Financial Extraction Multilingual Reasoning and Semantic Screening
【速读】： 该论文旨在解决金融领域中大型语言模型（Large Language Models, LLMs）在多语言市场部署时的跨语言可复用性问题，即基于英语开发并验证的处理流程是否可在其他语言环境中直接沿用，还是必须针对每个语种重新构建。其核心解决方案的关键在于系统评估跨语言可复用性的三个维度：准确性（accuracy）、推理努力迁移性（reasoning-effort transfer）与行为稳定性（behavioural stability）。研究通过三项互补任务进行验证：任务一（支票信息提取）和任务二（Mintaka问答任务，涵盖九种语言及九种模型推理配置）检验准确率与推理资源配置的跨语言迁移能力；任务三（FinanceBench试点，覆盖四种语言的SEC 10-K文件共50个样本）进一步检验上述模式在真实金融文本中的泛化能力。结果表明，准确性具有较广泛的可复用性——Gemini 3.5 Flash在不同提示语言下均达到95.2%–97.6%的提取准确率，且FinanceBench各语言平均准确率波动范围为5.8个百分点（77.1%–82.9%），但语言难度排名在不同任务间存在反转（如印地语在Mintaka任务中表现最差，但在FinanceBench中表现最佳，达82.9%）。推理努力的迁移性则呈现领域依赖性：在通用知识问答中高推理投入仅带来边际收益，而在金融内容中可提升高达8个百分点，且英语中有效的推理配置在其他语言中亦保持比例有效性。行为稳定性方面，各语言生成的引用集与英语版本的Jaccard重叠率维持在77%–79%，无单一语言持续表现更差。综上，研究支持在谨慎条件下对英语训练的处理流程进行跨语言复用，但强调仍需针对具体领域与语言开展独立验证，且语义相似性与模型自评置信度均无法可靠预测跨语言准确性。
链接: https://arxiv.org/abs/2610.27043
作者: Joong Ho(Joshua)Choi,Sherine George,Jiayang Zhao,Jayesh Rathi,Ashwin Kandath,Ananya Sane,Vishvak Srinivasan
机构: BNY AI Hub(纽约梅隆人工智能中心); BNY Responsible AI Office(纽约梅隆负责任人工智能办公室)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 15 pages Accepted into ICAIF 2026 AI4F Workshop
Abstract:Financial institutions increasingly deploy large language models (LLMs) in multilingual markets, raising a practical question: can pipelines developed and validated in English be reused across languages, or must they be rebuilt per locale? We test cross-lingual reusability along three dimensions–accuracy, reasoning-effort transfer, and behavioural stability–through three complementary tasks. Task 1 (cheque extraction) and Task 2 (Mintaka QA with pinned evidence across nine languages and nine model-effort configurations) test whether accuracy and effort settings transfer; Task 3 (FinanceBench pilot, 50 items over SEC 10-K filings in four languages) validates whether these patterns generalise to financial content. Accuracy reusability is broadly confirmed: Gemini 3.5 Flash achieves 95.2-97.6% extraction accuracy across prompt languages, and FinanceBench per-language means span 5.8pp (77.1-82.9%), though the language-difficulty ranking reverses between domains (Hindi is worst on Mintaka, best on FinanceBench at 82.9%). Effort reusability is domain-dependent: higher reasoning effort yields marginal gains on general-knowledge QA but up to 8pp on financial content, and effort configurations that help in English help proportionally in other languages. Behavioural reusability holds: citation-set overlap with English averages 77-79% Jaccard, with no single language consistently worse. Together, these results support qualified reuse of English-built pipelines, but per-domain, per-language validation remains necessary–neither semantic similarity nor self-reported confidence reliably predicts cross-lingual accuracy.
[NLP-124] Fluent but Fallible: Benchmarking Hallucination and Factual Reliability in Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在学术应用中面临的两大核心问题：幻觉（hallucination）——即生成与事实不符或虚构的内容，以及响应偏差（response deviation）——指模型输出系统性偏离真实答案的问题。这些问题严重削弱了模型在科研与学术场景中的可靠性与可信度。其解决方案的关键在于通过一个涵盖四个复杂度层级（基础、解释、高级、棘手）共70个精心设计的学术提示基准测试，对十种先进大模型（包括ChatGPT、Gemini、Claude 3.5 Sonnet、LLaMA 3、Mistral、Grok、Perplexity、DeepSeek V3、Falcon和Gemma）进行系统性评估。评估采用五项量化指标（准确率、精确率、召回率、响应偏差指数、幻觉指数），并由三位评审员独立打分。研究发现，尽管各模型表现各异，但整体上存在响应完整性和事实可靠性之间的根本权衡，且该权衡随提示复杂度升高而加剧；其中，Perplexity与DeepSeek V3在准确率上表现最优，Gemini在召回率与综合性能上最为均衡，而LLaMA 3则凭借保守生成策略展现出最强的抗幻觉能力，Falcon虽实现零幻觉与完美精确率，却以极低召回率和高响应偏差为代价，凸显了不同优化路径的内在矛盾。
链接: https://arxiv.org/abs/2610.27034
作者: Osama Shabih,Dayam Nadeem,Shahab Saquib Sohail,Mohammad Nadeem
机构: 未知
类目: Computation and Language (cs.CL)
备注: 29 pages, 5 tables, 7 figures, 66 references
Abstract:Large Language Models (LLMs) have significantly advanced natural language processing and artificial intelligence applications through their capabilities in reasoning, text generation and knowledge retrieval. However, LLMs are still vulnerable to two important challenges: hallucination, generating factually incorrect or fabricated content, and response deviation, meaning systematic deviations from ground-truth responses. These limitations raise serious concerns regarding reliability and trustworthiness in academic and research environments. This study offers a thorough comparative analysis using a variety of metrics of ten advanced LLMs ChatGPT, Gemini, Claude~3.5~Sonnet, LLaMA~3, Mistral, Grok, Perplexity, DeepSeek~V3, Falcon, and Gemma based on a benchmark consisting of 70 carefully designed academic prompts across four complexity levels: Basic, Explanation, Advanced, and Tricky, evaluated in February 2025 on specific model versions. Each response was independently evaluated by three reviewers using five metrics: Accuracy, Precision, Recall, Response Deviation Index, and Hallucination Index. Perplexity and DeepSeek~V3 achieved the highest overall accuracy, while Gemini demonstrated strong recall and a well-balanced performance profile across all five metrics. LLaMA~3 showed strong resistance against hallucinations through a conservative method of generation. Advanced-level prompts proved most challenging across all models, while Tricky prompts were handled more reliably owing to their binary true/false structure. Falcon achieved perfect precision and zero hallucinations, but at the cost of the lowest recall and the highest response deviation. The results show a fundamental trade-off between response completeness and factual reliability, intensified by prompt complexity.
[NLP-125] aching AI Agent to Play Poker Strategically Through Preference Optimization
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在存在隐含信息的博弈场景中，如何实现具有战略深度的决策行为，以及如何通过偏好优化提升静态、零样本策略的性能。其核心挑战在于：尽管现有小型语言模型（如SmallThinker、Qwen3、Gemma-3）能够执行合法动作，但其战略表现参差不齐，尤其在多玩家环境下表现不稳定。为此，论文提出基于KL锚定均衡的战略偏好优化方法（Strategic Preference Optimization with KL-anchored Equilibrium, SPOKE），其关键创新在于融合教师相对价值惩罚、策略锚定机制、风险约束与结构化推理监督，从而在保持与教师模型一致性的前提下，显著提升策略的抗剥削性与战略一致性。实验表明，在Kuhn和Leduc扑克任务中，SPOKE相较SFT、DPO、GRPO及GSPO等基线方法，实现了最低的平均可剥削性（Kuhn: 0.213；Leduc: 1.081）与最高的六人最大桌顶1匹配率（0.775），同时维持与SFT相当的教师KL散度（0.522 vs 0.523），验证了该方法在提升战略质量的同时有效保留教师行为特征的有效性。
链接: https://arxiv.org/abs/2610.27028
作者: Phurinat Polasa,Thanapon Thanadunpremdet,Phatthanasak Kraiduang,Saksorn Ruangtanusak
机构: KMUTT(曼谷理工大学); Mahidol University(玛希隆大学); SCB DataX, SCBX Group(暹罗商业银行数据X，SCBX集团)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: IEEE iSAI-NLP SRW 2026
Abstract:A poker agent can make a legal move and still make a strategically poor one. We ask whether compact AI agents can exhibit strategic behavior under hidden information and whether preference optimization can improve beyond static and zero-shot policies. We first evaluate a static baseline and three frozen language models: SmallThinker, Qwen3, and Gemma-3 whose performance varies substantially between exact-game and six-player evaluation. We therefore propose Strategic Preference Optimization with KL-anchored Equilibrium (SPOKE), combining teacher-relative value penalties, policy anchoring, risk constraints, and structured rationale supervision, and compare it with SFT, DPO, GRPO, and GSPO. Across three runs of a SmallThinker-4B-A0.6B policy, SPOKE achieves the lowest mean exploitability in Kuhn (0.213) and Leduc (1.081) and the highest six-max top-1 agreement (0.775), while maintaining teacher KL comparable to SFT (0.522 vs 0.523). These results provide preliminary evidence that strategic preference optimization can improve strategic quality while preserving teacher-consistent behavior.
[NLP-126] Decomposed Relation Extraction from Aviation Accident Narratives: Evaluating Transformer-Based Safety Knowledge Discovery under Relational Sparsity
【速读】： 该论文旨在解决航空事故叙述中关系抽取的难题，核心挑战在于真实的安全相关关系在所有可能的概念对中极为稀疏。为应对这一问题，研究提出了一种分解式框架，将概念识别与关系分类分离处理。其解决方案的关键在于：采用基于TF-IDF的线性支持向量机（SVM）和加权RoBERTa分类器，结合图标注资源中的硬负样本采样策略进行模型训练，并在封闭评估前冻结参数以保证公平性。然而，实验结果表明，尽管开发阶段的宏平均F1分数较高（SVM: 0.4738，RoBERTa: 0.4809），但在全面评估下性能显著下降（SVM: 0.2741，RoBERTa: 0.2420），且马修斯相关系数接近零（0.0128 和 0.0197），主要由大量误报导致。诊断分析进一步揭示，即使在有效的关系方向内，RoBERTa仍存在置信度分布重叠及误报现象。此外，端到端的SmolLM2评估发现模型遗漏或错误生成了部分关系，且输出存在结构缺陷。研究结论强调：仅知概念终点无法确保关系抽取的可靠性，且基于采样负例的开发性能难以反映真实场景下的表现；因此，控制误报率并有效管理候选关系空间的稀疏性，仍是实现航空安全知识提取的关键。
链接: https://arxiv.org/abs/2610.27013
作者: Aziida Nanyonga,Mustafa Tahsin Yilmaz,Saad Ezzini,Ugur Turhan,Graham Wild
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Relation extraction from aviation accident narratives is challenging because genuine safety relations are sparse among possible concept pairs. This study evaluates a decomposed framework separating relation classification from concept identification. A benchmark of 300 National Transportation Safety Board (NTSB) narratives containing 2,524 unique concept-text nodes was expanded into 21,038 ordered candidate pairs, with only 248 (1.18%) representing RESULTS_IN or CONTRIBUTES_TO relations. A TF-IDF Linear Support Vector Machine (SVM) and weighted RoBERTa classifier were developed using an independent graph-annotated resource with hard-negative sampling and frozen before sealed evaluation. Development Macro-F1 scores of 0.4738 (SVM) and 0.4809 (RoBERTa) declined to 0.2741 and 0.2420 under exhaustive evaluation, with Matthews correlation coefficients of 0.0128 and 0.0197. False positives dominated: the SVM recovered 96 reference relations but generated 7,018 false positives, while RoBERTa recovered 125 but generated 9,824. Post-test diagnostics revealed overlapping RoBERTa confidence distributions and false positives even within valid relation-category directions. Complementary end-to-end SmolLM2 evaluation identified missed and unsupported relations alongside malformed outputs. These findings demonstrate that known concept endpoints do not guarantee reliable relation extraction and that sampled-negative development performance may poorly reflect exhaustive evaluation. Controlling false positives and managing candidate-space sparsity remain essential for aviation safety knowledge extraction.
[NLP-127] Evaluating Tibetan Speech Recognition Across Dialects and Adaptation Regimes: A Multi-Granularity Analysis
【速读】： 该论文旨在解决藏语自动语音识别（ASR）在缺乏词边界标注的情况下，评估指标（如字符错误率、音节错误率和分词错误率）之间存在不一致且对适应失败敏感的问题。其解决方案的关键在于构建一个融合字符错误率、音节错误率与分词错误率的统一评估框架，并通过基于词典、神经网络及语言模型的多种分词方法，与专家标注进行对比，以系统性地分析不同语言层级上的性能表现。研究发现，较低的字符错误率并不必然对应更低的音节或分词错误率，且在不同数据集上排名差异持续存在，表明单一指标难以全面反映系统性能。此外，在零样本迁移与目标语料微调实验中，指标间的相关性在部分失败的适应设置下显著减弱；尽管提高低秩自适应（LoRA）秩可改善音节和分词性能，但在所测试范围内仍劣于全量微调。研究还引入标准化的音系表示，为拼写错误提供了新的分析视角。这些结果强调了在多个语言层次上进行评估的重要性，并推动开发更优的分词与适应方法，以应对低资源场景下的自动语音识别挑战。
链接: https://arxiv.org/abs/2610.27002
作者: Jacob Moore,Sheng Li
机构: Lawrence Technological University (劳伦斯科技大学); Institute of Science Tokyo (东京科学研究所)
类目: Computation and Language (cs.CL)
备注:
Abstract:Tibetan speech recognition is usually evaluated with character or syllable error rates because the writing system does not mark word boundaries. These metrics can rank systems differently and respond differently to adaptation failures. We combine character error rate, syllable error rate, and segmented word error rate in one evaluation framework. We use dictionary-based, neural, and language-model-based segmentation and compare their boundaries with expert annotations. We evaluate five ASR architectures on Lhasa Tibetan, then test zero-shot transfer and target-corpus adaptation on five additional corpus subsets. We also vary LoRA rank and training-data size to examine adaptation failure. The results show that a lower character error does not always lead to a lower syllable or segmented-word error. These ranking differences persist across corpora. Metric correlations also weaken in some unsuccessful adaptation settings. Higher LoRA ranks improve syllable and segmented-word scores but remain worse than full fine-tuning in the tested range. A standardized phonological representation provides another view of spelling errors. These findings support evaluation at several linguistic levels and motivate better segmentation and adaptation methods for low-resource ASR.
[NLP-128] How to get the best and fairest performance out of your ASR system using activation steering
【速读】： 该论文旨在解决自动语音识别（ASR）系统在性能与人口统计学公平性（demographic fairness）之间的权衡问题，尤其关注如何通过激活引导（activation steering）技术实现双重优化。其核心解决方案在于采用监督优化（supervised optimization）方法来生成有效的引导向量（steering vectors, SVs），该方法不仅能同步提升整体ASR性能，还能在性别、年龄、方言等多样化人口统计维度上增强系统的公平性表现。关键发现在于，仅需少量甚至无标签的声学相似说话人数据即可训练出高效的SVs，且基于同声学特征群体的训练效果优于基于同一说话人或同标签人口群体的训练方式，表明该方法具备少样本（few-shot）乃至零样本（zero-shot）适应未见说话人的潜力，显著降低了对个性化标注数据的依赖。
链接: https://arxiv.org/abs/2610.26994
作者: Felix Herron,Alex Colagrande,Solange Rossato,François Portet,Alexandre Allauzen,Chuyuan Li
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Activation steering for ASR systems is an underexplored technique in improving both ASR performance and demographic fairness. We investigate two forms of activation steering: paired activation subtraction, and supervised optimization. We find the former to be ineffective at ameliorating ASR performance whatsoever, while the latter is quite effective at deriving steering vectors (SVs) that improve at once overall ASR performance and demographic fairness along such diverse axes as gender, age, and dialect. We also explore what types of data—from the most to the least personal—are necessary for calculating or training SVs. First, we show significant performance in performance from training SVs using an eclectic group of speakers from the same dataset. Furthermore, we find that vectors trained using acoustically similar speakers leads to ASR performances that are competitive or superior to vectors trained on the same speaker, or speakers from the same labeled demographic group. This finding shows that we can harness activation steering for few-shot adaptation given access to a speaker’s demography, or zero-shot adaptation to unseen speakers without the need for any labels at all.
[NLP-129] RACE: Training-aware Robustness and Challenge Evaluation for Text-Vision Embedding Models
【速读】： 该论文旨在解决现有文本-视觉嵌入模型评估基准中存在的核心问题：当前的评估体系仅提供综合得分，无法区分模型性能是源于通用能力、任务特定适应性，还是对复杂挑战任务的鲁棒性。为此，作者提出了一个名为TRACE（Training-aware Robustness and Challenge Evaluation）的评估框架，将来自多个数据集的168项任务重新组织为三大类别——通用（General）、适应性（Adaptability）与挑战性（Challenge）。通过在两个文本-视觉嵌入模型上开展受控实验，研究发现：结合通用数据与适应性训练数据可分别提升适应性性能7.67和7.09分，同时基本保持通用性能；然而，挑战性任务上的表现始终低于5.0分，且额外训练数据对其改善有限。进一步分析揭示，多语言任务与视觉文档任务是主要薄弱环节。研究结果表明，单纯依赖数据层面的适应性改进不足以克服这些挑战，凸显了在方法论层面进行创新的必要性，以突破当前数据驱动范式的局限。
链接: https://arxiv.org/abs/2610.26993
作者: Pume Tuchinda,Tinnakit Udsa,Peerat Limkonchotiwat
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Existing benchmarks for text-vision embedding models report aggregate scores that obscure whether performance comes from general capabilities, task-specific adaptation, or robustness to challenging tasks. We introduce TRACE, a Training-aware Robustness and Challenge Evaluation framework that reorganizes 168 tasks from various sources of datasets into three categories: General, Adaptability, and Challenge. Through controlled experiments on two text-vision embedding models, we show that combining general and adaptation data improves Adaptability performance by 7.67 and 7.09 points, respectively, while largely preserving General performance. However, Challenge performance remains below 5.0, with limited improvements from additional training data. Further analysis identifies multilingual and visual-document tasks as major weaknesses. Our findings demonstrate that data adaptation alone is insufficient to address these challenges, highlighting the need for methodological advances beyond data-centric improvements.
[NLP-130] hink Before You Transcribe: Context-Steered Reasoning for Hotword-Aware Speech Recognition
【速读】： 该论文旨在解决上下文增强的自动语音识别（ASR）中候选热词（hotword）不完整或不准确导致的识别偏差问题。现有方法在利用上下文信息时，可能因错误或缺失的候选词干扰声学证据，从而降低识别准确性。其解决方案的关键在于提出一种上下文引导的推理框架，通过在转录前生成一个与音频对齐的短描述性字幕（audio-grounded caption），构建任务特定的中间表示。该字幕使模型能够在做出词汇决策前评估候选词的相关性，从而有效抑制无关或错误候选词的干扰。通过监督微调建立从字幕到转录的结构化映射关系，并结合针对性的轨迹注入策略，在困难语音片段产生无效回溯路径时强化群体相对策略优化（group relative policy optimization），显著提升模型鲁棒性与性能。实验结果表明，该方法在多种上下文条件下均能稳定提升转录质量与稀有实体识别能力，同时对无关候选词具有强抗干扰能力。
链接: https://arxiv.org/abs/2610.26986
作者: Qihao Wang,Hang Su,Fan Cui,Lichun Fan,Jian Luan,Yue Hu
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Contextual automatic speech recognition (ASR) improves rare-entity recognition by supplying candidate hotwords, but retrieved candidates may be incomplete or incorrect. A reliable system must therefore exploit a relevant candidate without allowing unsupported alternatives to override the acoustic evidence. We propose a context-steered reasoning framework that generates a short, audio-grounded caption before transcription. This task-specific representation lets the model assess candidate relevance before making a lexical decision. Supervised fine-tuning establishes the structured caption-to-transcription behavior, while targeted trajectory injection strengthens group relative policy optimization when difficult utterances yield only unsuccessful rollouts. Experiments across diverse context conditions show that the proposed method consistently improves both transcription quality and hotword recognition while remaining robust to irrelevant candidates.
[NLP-131] Sharp Degree Bounds for Putinar Certificates in Terms of the Positivity Margin via Analytic Sums of Squares
【速读】： 该论文旨在解决生成式多项式优化中关于Putinar型平方和（Sum-of-Squares, SOS）证书的度数上界问题，特别是其与正定性裕度（positivity margin, δ）之间的依赖关系。研究的核心问题是：在给定阿基米德二次模（Archimedean quadratic module）及其生成元、且目标多项式的度数有上界的情况下，如何确定能够保证非负性的SOS证书的最小可能度数。解决方案的关键在于建立一个关于δ的紧致度数上界，即对于归一化的多项式目标函数，存在统一的证书度数上界为 $ O(\delta^{-1/2} \log^2(2/\delta)) $。通过构造三变量情形下的匹配下界例子，证明了该上界在δ依赖关系上的最优性，从而确立了该依赖关系的紧性。这一结果进一步推导出Lasserre层级在最坏情况下的收敛速率也是紧致的，为半定规划中的收敛分析提供了理论依据。
链接: https://arxiv.org/abs/2610.26976
作者: Yuxiang Zhao,Ting Gan
机构: Wuhan University (武汉大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:We establish sharp degree bounds for Putinar sum-of-squares (SOS) certificates in terms of the positivity margin \delta . For any fixed Archimedean quadratic module with prescribed generators and any fixed bound on the degree of the objective, we prove a uniform certificate-degree upper bound of O(\delta^-1/2\log^2(2/\delta)) for normalized polynomial objectives. We show that this bound is sharp by constructing a matching lower-bound example in three variables, thereby establishing that the dependence on \delta is optimal. These matching bounds also yield sharp worst-case convergence rates for the Lasserre hierarchy.
[NLP-132] ransformers Learn Two Competing Sorting Algorithms and Only One Length-Generalizes
【速读】： 该论文旨在解决两层变压器（two-layer transformers）在排序任务中长度泛化能力差异的问题，即模型在训练时仅针对固定长度序列进行优化，却表现出截然不同的长度外推性能。其核心问题是：为何某些模型能够完美实现长度泛化，而另一些则不能，且这种差异如何由模型初始化和内部机制决定。解决方案的关键在于通过精心设计权重初始化，可引导模型收敛至两种机制上迥异的解空间：第一类解不依赖第二注意力层进行排序，因而无法泛化；第二类解则充分利用两层注意力的协同作用，其中第二层能基于第一层捕获的上下文信息有效缩小搜索空间，从而实现完美的长度泛化。研究进一步揭示，长序列中罕见但逐渐增多的输入类型是导致泛化性能差异的根源，而第二类解通过层间功能分化（layer-wise specialization）更有效地处理这些稀疏输入。在寻找序列最小值的任务中，观察到类似双解现象，但泛化模式相反——两层结构反而劣于单层结构，表明层间分工对不同任务具有任务依赖性。
链接: https://arxiv.org/abs/2610.26974
作者: Khashayar Gatmiry,Nathan W. Henry,Chloe H. Su,Emmanuel Abbe,Samy Bengio,Melanie Weber,Jonathan A. Kelner,Peter L. Bartlett
机构: UC Berkeley; Apple; Harvard University; EPFL; Google DeepMind
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:We show that two-layer transformers trained to near-zero loss on a sorting task converge to one of two mechanistically distinct solution classes depending on their initialization. One solution class consists of solutions that do not use the second attention layer to sort, but also fail to length-generalize. The other, which uses both attention layers, length-generalizes perfectly, despite only being trained to sort sequences of a fixed length. By carefully choosing the weight initialization, we demonstrate that we can steer the model toward learning one solution or the other. We identify input types that are rare at the training length but increasingly common at longer sequence lengths as a source of this disparity in length-generalization, and find that layer-wise specialization of the second class of solutions allows them to handle these rare sequences more accurately. In the second class, we discover a mechanism by which the second layer attention can use the contextual information gathered by the first layer attention to reduce its search space. On the related task of finding the minimum of a sequence, we find a similar two solution-type phenomenon, but with the opposite generalization pattern, where a different form of layer-wise specialization causes the two-layer solution to length-generalize worse than the single-attention solution.
[NLP-133] SZED at ArGuard Shared Tasks: Multi-Task Label-Attention and Cross-Modal Fusion for Arabic Hateful Meme Detection
【速读】： 该论文旨在解决阿拉伯语仇恨表情包检测任务中的两大挑战：二分类（A1）与细粒度多标签标注（A2）。针对A1任务，其关键解决方案在于融合MARBERTv2与CLIP ViT-L/14的双模态嵌入表示，通过对比拼接、自注意力机制及集成头等不同融合策略，优化跨模态信息整合效果。对于A2任务，提出一种多任务标签注意力架构，引入每类专属的查询向量（query vectors）以增强标签间语义区分能力，并设计辅助头实现级联校准（cascade calibration），有效缓解长尾标签分布带来的性能偏差问题。实验结果显示，该方法在A1上取得0.7597的宏平均F1分数，在A2上达0.3142（10个活跃标签的平均值），同时验证了模型在不同运行间的可复现性及对各类错误模式的鲁棒性。
链接: https://arxiv.org/abs/2610.26972
作者: Selma Khelili,Zohra Abdeli
机构: ENSIA(阿尔及利亚国家农业与环境工程学院), Algeria(阿尔及利亚)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:We present SZED’s submission to ArGuard 2026 Track A: Arabic hateful meme detection. The task has two parts: binary classification (A1) and fine-grained multi-label tagging (A2). For A1, we fuse MARBERTv2 and CLIP ViT-L/14 embeddings, comparing concatenation, self-attention, and ensemble heads. For A2, we use a multi-task label-attention architecture with per-class query vectors and an auxiliary head for cascade calibration. We reach macro-F1 = 0.7597 on A1 and 0.3142 on A2 (averaged over 10 active classes). We report reproducibility across runs, per-class error patterns, and the effect of cascade calibration on the long-tailed A2 label space.
[NLP-134] Do LLM s Really Lack Difference Awareness? Uncovering Heuristic Substitution in Group-Differentiating Fairness
【速读】： 该论文旨在解决大语言模型（LLM）在公平性决策中缺乏差异意识（difference awareness）的问题，即模型未能识别和承认社会群体间有意义的差异。传统观点认为这一缺陷源于模型本身缺少对差异的认知能力，但本文提出一种新解释——公平性启发式替代（Fairness Heuristic Substitution）：模型并非完全缺乏差异意识，而是由于同时需识别相关群体差异并判断其对公平性的意义这一复杂任务，导致模型倾向于采用更简单、更易获取的“形式平等”（formal equality）启发式策略作为捷径。为此，作者提出两阶段区分-决策框架（TDD），将公平性推理显式分解为“区分上下文相关的群体差异”与“据此作出决策”两个阶段，从而抑制对形式平等的依赖。实验表明，TDD在四个基准测试和十种不同规模（从数十亿到数万亿参数）的模型上均显著优于单阶段基线，包括经过模拟测试集泄露增强的模型；结合神经元层级的对比探针分析进一步揭示了形式平等启发式的机制特征，包括紧凑且局部协同的表征结构以及默认推理下的显著激活偏差；此外，TDD的分阶段设计还暴露了公平性推理中被忽视的失败模式，如启发式转移（heuristic displacement）和阶段动态失调（misaligned stage dynamics）。
链接: https://arxiv.org/abs/2610.26961
作者: Jingshen Zhang,Bo Wang,Yanlin Fu,Yuewen Dang,Dongming Zhao,Zifei Yu,Yuexian Hou
机构: Tianjin University (天津大学); QMindTech; Fudan University (复旦大学); Tianjin Huizhixingyuan Information Technology Co., Ltd
类目: ocial and Information Networks (cs.SI); Computation and Language (cs.CL)
备注:
Abstract:In this paper, we investigate the issue of difference awareness in LLMs, defined as a model’s ability to recognize and acknowledge meaningful differences between social groups in fairness decision-making. Previous studies have consistently shown that models perform poorly on such tasks, a finding commonly attributed to a lack of difference awareness. We propose an alternative account, which we term Fairness Heuristic Substitution: difference awareness is not necessarily absent, but the complex demand of simultaneously identifying relevant group differences and determining their implications for fairness decisions may lead models to substitute the simpler, more accessible equal-treatment heuristic, i.e., formal equality. Based on this hypothesis, we introduce TDD, a Two-stage Distinguish-Decide framework that explicitly decomposes fairness reasoning into first distinguishing contextually relevant group differences and then making the corresponding decision, thereby discouraging reliance on formal-equality shortcuts. (i) Across four benchmarks and ten models ranging from billions to trillions of parameters, TDD significantly outperforms one-stage baselines, even one strengthened through simulated test-set leakage. (ii) Combining TDD with neuron-level contrastive probing provides mechanistic evidence consistent with the Fairness Heuristic Substitution hypothesis by revealing two signatures of the formal-equality: a compact, locally coordinated representational structure and a pronounced activation bias under default reasoning. (iii) TDD’s stage-wise decomposition uncovers overlooked failure modes in fairness reasoning, including heuristic displacement and misaligned stage dynamics.
[NLP-135] Dynamic System Inspired Continuous Large Language Model for AI4Science Tasks
【速读】： 该论文旨在解决科学基础模型在处理连续现象时仍依赖离散残差计算所带来的局限性，即如何在不扩展主干网络参数量的前提下实现可扩展的计算深度。其核心解决方案是提出Continuum-LM，通过将预训练的残差变换重用为连续深度语言模型中的动力学算子，使隐藏状态通过高阶数值积分求解的常微分方程（Ordinary Differential Equation, ODE）进行演化，从而以可调的求解器阶数和步长控制计算过程。实验表明，在8个科学任务上，Continuum-LM在0.6B和1.7B参数规模下分别相较于离散对应模型提升了7.77和10.96个百分点的平均精确匹配准确率；在0.6B规模下，其性能还优于同等参数与残差函数计算预算下的Ouro型循环控制机制。这些结果验证了通过参数复用实现计算深度可扩展性的可行性，并确立了连续隐藏状态演化作为科学语言模型超越参数量扩展的一种实用设计维度。
链接: https://arxiv.org/abs/2610.26959
作者: Zhaoxin Fan,Zheng Zhu,Yuqing Ma,Faguo Wu,Jian Yang,Wenjun Wu
机构: Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing, Beihang University (北京航空航天大学未来区块链与隐私计算高精尖创新中心); School of Artificial Intelligence, Beihang University (北京航空航天大学人工智能学院); GigaAI (GigaAI)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Scientific foundation models increasingly operate on continuous phenomena, yet still use discrete residual computation. We ask whether pretrained residual transformations can support additional latent computation without expanding the backbone. To test this possibility, we introduce Continuum-LM, which reuses these transformations as dynamical operators in a continuous-depth language model. Its hidden states evolve through an ordinary differential equation evaluated by higher-order numerical integration, with solver order and step size controlling computation. Across eight scientific tasks, Continuum-LM trains stably and improves mean exact-match accuracy by 7.77 and 10.96 percentage points over discrete counterparts at the 0.6B and 1.7B scales, respectively. At 0.6B, it also outperforms an Ouro-style recurrent control under matched parameter and residual-function evaluation budgets. Together, these results support extensible computational depth through parameter reuse, identifying continuous hidden-state evolution as a practical design dimension for scientific language models beyond parameter scaling.
[NLP-136] LAMB: Latent Multi-Hypothesis Beliefs for Long-Horizon Large Language Agents
【速读】： 该论文旨在解决长时序语言智能体在复杂环境中因记忆容量有限而难以有效维持交互证据与未观测状态的多重竞争假设的问题。现有方法多采用离散自然语言编码记忆，导致重复压缩过程中可能丢失关键证据或过早舍弃合理假设。其解决方案的关键在于提出一种基于潜在多假设信念（latent multi-hypothesis beliefs）的递归框架LAMB，通过“共性-增量”（Common–Delta）分解结构，将共享的共性状态用于存储交互证据，同时以加权增量组件表示竞争性假设；利用动作-结果预测中的相对证据实现无需生成文本记忆的递归假设更新；并通过固定令牌的信念前缀对策略进行上下文条件读出，使假设可分离暴露。整个信念模块与策略通过动作模仿及辅助的状态、结果和观测判别目标联合训练，显著提升了长期决策性能，在ALFWorld、BabyAI、ScienceWorld和Crafter等多个基准上均达到或超越PABU基线，尤其在ALFWorld的成功率（SR@30）和ScienceWorld成功率上分别提升6.0和22.0个百分点，且在Crafter任务中实现26.37点的成就进度提升，同时将冗余动作率降低17.70个百分点，验证了潜在多假设信念作为紧凑记忆机制在长时序决策中的有效性。
链接: https://arxiv.org/abs/2610.26955
作者: Peien Ding,Hui Li,Yuhan Liu,Jinsong Su
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Long-horizon language agents require compact memories that retain interaction evidence while preserving competing hypotheses about unobserved states. Although effective, many belief-state methods encode memory in discrete natural language, where repeated compression can omit evidence or prematurely discard plausible alternatives. We introduce LAMB, a recurrent framework for latent multi-hypothesis beliefs. A Common–Delta factorization combines a shared Common state for interaction evidence with weighted Delta components for competing hypotheses. Relative evidence from action-outcome predictions guides recurrent hypothesis updates without generating textual memory. A context-conditioned readout exposes hypotheses separately to the policy through a fixed-token belief prefix in the default configuration. The belief modules and policy are jointly trained using action imitation and auxiliary state, outcome, and observation-discrimination objectives. Across ALFWorld, BabyAI, ScienceWorld, and Crafter, LAMB matches or outperforms PABU, improving ALFWorld SR@30 and ScienceWorld success by 6.0 and 22.0 percentage points, respectively, and Crafter achievement progress by 26.37 points. On ALFWorld, it also reduces the redundant-action rate by 17.70 percentage points. These results support latent multi-hypothesis beliefs as compact memory for effective long-horizon decision making
[NLP-137] A Multi-granularity Interaction Network for Relational Triple Extraction
【速读】： 该论文旨在解决关系三元组抽取（Relational Triple Extraction）中因“未见”实体导致的传统方法性能严重下降的问题。其核心挑战在于，现有方法无法为“未见”实体建立有效的上下文依赖关系，从而影响关系识别的准确性。为此，论文提出一种多粒度交互网络（Multi-Granularity Interaction Network），通过融合词-词、实体-关系以及头尾实体间的多层次交互，深入捕捉语义依赖关系。该网络通过多轮迭代机制，有效优化头尾实体的表示，并增强与关系三元组相关的全局语义特征，从而弥补早期步骤中可能遗漏或误判的三元组。实验结果表明，该方法在两个重构数据集（NYT-R/WebNLG-R）上均取得显著提升，在“未见”实体场景下的F1分数相比先前方法分别提高3.1%和20.5%，展现出强大且一致的性能。
链接: https://arxiv.org/abs/2610.26952
作者: Xiaocheng Luo,Yanping Chen,Zaifu Zhan
机构: Soochow University (苏州大学); State Key Laboratory of Public Big Data, College of Computer Science and Technology, Guizhou University (贵州大学公共大数据国家重点实验室，计算机科学与技术学院); Department of Electrical and Computer Engineering, University of Minnesota (明尼苏达大学电气与计算机工程系)
类目: Computation and Language (cs.CL)
备注:
Abstract:Relational triple extraction identifies entity relations from raw sentences without depending on entity type information. The performance of traditional methods is seriously degenerated by unseen'' entities, because these methods can't build contextual dependencies relevant unseen’’ entities. This paper addresses this challenge by proposing a multi-granularity interaction network that leverages word-word, entity-relation, and head-tail interactions to capture deep semantic dependencies. By incorporating multiple iterations, the multi-granularity interaction network is effective to refine the representations corresponding to head and tail entities. It enhances the global semantic features associated with relational triples, aiding in the remediation of triples that may have been overlooked or omitted in earlier steps. Strong and consistent performance is achieved by our method on two rearranged datasets, with a significant improvement in identifying relational triples, including those involving ``unseen’’ named entities. For example, our model improves F1-score by 3.1%/20.5% over previous approaches on the NYT-R/WebNLG-R datasets in the Unseen metric. Code: this https URL.
[NLP-138] Do We Need the Largest LLM s for Every Clinical Decision?
【速读】： 该论文旨在解决大尺寸临床语言模型在实际部署中因依赖多GPU基础设施而导致本地化应用受限的问题。其核心挑战在于如何在保证临床决策准确性的前提下，实现模型的高效压缩与低资源部署。解决方案的关键在于提出一种基于决策层面的统计认证机制——CERTA（Certified Efficient Runtime for Trustworthy AI），该机制通过形式化有限样本验证方法，动态判断量化后的模型是否可替代原始的BF16精度参考模型进行临床决策；对于无法通过认证的决策，则自动升级至原生高精度模型处理。实验表明，在11个不同规模（7-72B）且经指令微调的通用及领域适配模型上，采用四舍五入的8位量化即可覆盖99.8%-100%的测试决策，而4位非对称浮点量化（NF4）配置在保持较高准确性的同时，将平均部署成本降至BF16参考模型的25.8%-27.9%，并显著降低内存占用（峰值分别达43.0和42.7 GB，低于单块48 GB加速器的容量）。此外，通过精确分样证书（exact split-sample procedure）进一步考虑了配置选择的影响，确保了风险可控的部署策略。CERTA将模型压缩从固定的精度选择转变为一种具有形式化保障的风险控制型部署策略，为临床生成式AI（Generative AI）的安全落地提供了可信赖的技术路径。
链接: https://arxiv.org/abs/2610.26937
作者: Ali H. Lazem,William Teahan
机构: Bangor University (班戈大学); University of Thi-Qar (巴士拉大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large clinical language models often require multi-GPU infrastructure, limiting local deployment. We introduce CERTA, a decision-level statistical certificate that determines when a quantized model may substitute for its native BF16 reference and escalates decisions that cannot be certified to it. Across eleven instruction-tuned models, both general-purpose and domain-adapted, from three families (7-72B), we evaluated 2,500 clinical-verification decisions from de-identified patient contexts (2,484 patients), with patient-disjoint calibration and testing. In the higher-power operational analysis, round-to-nearest 8-bit quantization at BF16 storage served 99.8-100% of held-out decisions across all eleven models; formal finite-sample guarantees were assessed separately using the exact split-sample procedure. In the same operational analysis, the class-conditional 4-bit frontier reduced the amortized, BF16-equivalent deployment cost to 25.8% and 27.3% of the BF16 reference for Qwen2.5-72B and HuatuoGPT-o1-72B, respectively, with measured peak memory of 43.0 and 42.7 GB, below the nominal capacity of a single 48 GB accelerator. Deployment on that specific hardware was not directly tested. Under the exact split-sample certificate, which also accounts for configuration selection, these costs were 26.7% and 27.9%. On external disease named-entity benchmarks against human annotations, the frontier-selected NF4 configurations had a mean absolute F1 change of 0.045 on both datasets, with substantial model-specific variation. CERTA converts compression from a fixed precision choice into a risk-controlled deployment policy for this verification task.
[NLP-139] AtomSD: Procedural Atom-Guided Self-Distillation for LLM Agents
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在执行复杂任务时缺乏有效程序性知识（procedural knowledge）的问题，尤其体现在如何灵活地规划和调整行为以适应多变的交互历史。现有方法通常通过轨迹或技能包来表示程序性知识，导致其在不同任务间难以灵活复用。为此，论文提出“程序原子”（Procedural Atoms）这一细粒度、可独立应用的知识单元，用于表征意图（Intent）、时机（When）、行动（Do）与规避（Avoid）等关键指导信息。其核心解决方案是AtomSD——一种基于程序原子引导的自蒸馏框架，该框架使学习者能够从自身成功轨迹中持续提取程序原子，并通过原子条件监督与任务奖励优化相结合的方式，将这些原子知识逐步融入动态演化的策略网络中。实验结果表明，AtomSD在多个基准上显著优于现有的基于奖励与蒸馏的方法，平均性能提升最高达9.31分，且分析验证了程序原子在提供有效行为指导方面的价值，凸显了原子级程序性知识对策略学习的关键作用。
链接: https://arxiv.org/abs/2610.26926
作者: Bohan Yu,Pengfei Cao,Kai Tian,Chen Han,Shi-Yang Li,Chenxi Zhou,Junlin Liu,Cao Liu,Ke Zeng,Jun Zhao,Kang Liu
机构: School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences(中国科学院大学交叉科学学院); The Key Laboratory of Cognition and Decision Intelligence for Complex Systems, Institute of Automation, Chinese Academy of Sciences(中科院自动化所认知与复杂系统决策智能重点实验室); School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院); Meituan(美团)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) possess broad factual knowledge and strong reasoning capabilities, but solving complex tasks also requires procedural knowledge about how to act. Existing approaches often represent procedural knowledge through trajectories or bundled skills, limiting flexible reuse across interaction histories. This paper introduces Procedural Atoms, fine-grained and independently applicable units that represent Intent, When, Do, and Avoid guidance. Building on this representation, we propose AtomSD, an Atom-guided Self-Distillation framework in which the learner continually extracts Procedural Atoms from its own successful trajectories and distills them into the evolving policy through Atom-conditioned supervision and task-reward optimization. Comprehensive experiments show that AtomSD consistently outperforms strong reward-based and distillation methods, exceeding the state-of-the-art baseline by up to 9.31 points in average performance. Further analyses demonstrate that the proposed Procedural Atoms provide effective guidance, highlighting the value of atom-level procedural knowledge for policy learning.
[NLP-140] Mixture of Phonetic Experts Based Low-Rank Adaptation of Conformer Models for Accented English Speech Recognition INTERSPEECH2026
【速读】： 该论文旨在解决非母语者英语语音识别中因口音差异导致的性能下降问题，尤其关注如何实现高效、可迁移的参数化适配。传统基于专家混合（mixture-of-experts）的方法通常按口音类别分配专家，将口音身份视为主要变异性轴，但这种策略难以泛化至未见口音。为此，本文提出基于发音特征专家的低秩适配（MoPE-LoRA），其核心创新在于将适配机制重构为以发音类型（如发音方式，manner-of-articulation）为组织基础，利用一组固定的六类低秩专家对语音帧进行建模。通过结合音素监督与学习到的声学门控机制实现帧级路由，使模型能够动态选择最相关的发音类别专家。由于专家集在不同口音间共享，系统无需为每种口音构建专属模块，因而具备良好的零样本泛化能力。在L2-ARCTIC数据集上的实验表明，该方法在多种评估场景下均显著优于单个LoRA和全量微调，尤其在零样本测试中相较单个LoRA实现12.3%的相对性能提升，验证了基于发音类别分解的适配策略在可扩展性与迁移性方面的优势。
链接: https://arxiv.org/abs/2610.26888
作者: Santosh Dahal,Anmol Guragain,Tianchi Liu,Luis Fernando D’Haro,Kiran Chandra Dahal
机构: 未知
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注: Accepted at Interspeech 2026 (Sydney, 27 September to 1 October 2026). 5 pages, 3 figures, 6 tables
Abstract:We present Mixture of Phonetic Experts based Low-Rank Adaptation (MoPE-LoRA), a parameter-efficient adaptation framework for accented English speech recognition. Recent mixture-of-experts approaches typically allocate experts per accent, implicitly treating accent identity as the primary axis of variation. In contrast, we organize adaptation along phonetic categories, motivated by the observation that accent differences are often expressed as systematic distortions of phoneme realization. MoPE-LoRA introduces a fixed set of six low-rank experts aligned with manner-of-articulation classes and performs frame-level routing through a hybrid mechanism that combines phoneme supervision with learned acoustic gating. Because the expert set is shared across accents, the model does not require accent-specific modules and naturally extends to previously unseen accents. Experiments on L2-ARCTIC show that MoPE-LoRA consistently improves over single LoRA and full fine-tuning. In zero-shot evaluation where one accent is held out during training, the proposed method yields a 12.3% relative improvement over single LoRA. These results indicate that phonetic-category-based decomposition offers a scalable and transferable alternative to accent-level adaptation.
[NLP-141] Multimodal Ambivalence and Hesitancy Recognition with Bidirectional Audio–Text Cross-Attention
【速读】： 该论文旨在解决视频层面行为矛盾与犹豫（Ambivalence and Hesitancy, A/H）的自动识别问题，其核心挑战在于如何有效融合语言内容与语音声学特征以捕捉人类表达中微妙的行为状态。解决方案的关键在于提出一种音频-文本联合建模框架：利用预训练的多语言BERT对全视频转录文本进行编码，同时采用wav2vec 2.0对语音波形进行编码，并将其映射至共享表示空间；通过两个堆叠的双向交叉注意力块实现语言标记与声学帧之间的相互关注，无需显式的时间对齐即可捕获跨模态语义关联；最终通过掩码均值池化聚合序列特征，并结合拼接、逐元素乘积和绝对差值等操作进行融合，用于二分类任务。实验结果表明，该方法在BAH数据集的主测试集上取得了77.95%的A/H F1分数，验证了联合建模语言与声学信息在自动识别复杂心理状态中的有效性。
链接: https://arxiv.org/abs/2610.26868
作者: TingTing Gu,JinCheng Luo,WenQiu Guo
机构: 未知
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注:
Abstract:Ambivalence and hesitancy (A/H) are subtle behavioural states that can be expressed through both what a person says and how it is spoken. This paper presents an audio–text framework for video-level A/H recognition. Full-video transcripts and speech waveforms are encoded using pre-trained multilingual BERT and wav2vec~2.0, respectively, and projected into a shared representation space. Two stacked bidirectional cross-attention blocks allow linguistic tokens to attend to acoustic frames and acoustic frames to attend to linguistic context without requiring explicit temporal alignment. The resulting sequences are aggregated through masked mean pooling and combined using concatenation, element-wise products, and absolute differences for binary classification. We evaluate the model on the subject-independent split of the Behavioural Ambivalence/Hesitancy (BAH) dataset. The proposed system achieves 72.19% accuracy, 75.00% precision, 81.13% recall, 77.95% A/H F1-score, and 70.16% Macro-F1 on the test set. These results demonstrate the potential of jointly modelling linguistic and acoustic information for automatic A/H this http URL at:this https URL.
[NLP-142] High-dimensional layer-state trajectories reveal the internal dynamics of transformer inference
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在推理过程中内部计算动态的可解释性问题，特别是如何从参数、注意力头或神经元等局部组件中理解整体推理路径的形成机制。传统方法虽能定位功能单元，却无法揭示单次推理中状态演化的具体过程。本文的关键解决方案是将分层隐藏状态的有序演化轨迹视为可测量的实证计算过程，并验证层序是否构成可量化的计算结构。研究发现，在Qwen、Llama和Gemma等多个检查点中，真实层序所生成的状态路径与终点弦（endpoint chord）的对齐度比保持终点的随机重排高1.60至2.24倍，表明层序具有显著的计算意义。该效应不仅在词汇任务中显现，且在相同架构的不同随机初始化中均存在，说明其源于模型架构兼容的骨架结构而非训练特定规律。进一步分析表明，直接使用隐藏状态作为机制坐标优于词表回读，而任务定义的坐标可使候选竞争关系实现可计算化。在添加干预失败后，局部转移执行器（local transition actuators）被证明可选择性调控跨检查点的内部轨迹；通过置信度门控机制，仅17条保留的Qwen闭合推理轨迹被成功引导越过全词汇top-1候选边界，而其余203条非闭合输出保持不变，验证了选择性门控与有限作用器族的核心作用。因此，层序揭示了一种可观测、可计算且行为上具有显著影响的Transformer推理过程，尽管未建立完整的LLM动力学理论或答案正确性的控制机制，但为理解模型内部运作提供了关键线索。
链接: https://arxiv.org/abs/2610.26866
作者: Heng Zhao,Yufei Wang,Ruidian Song,Weisheng Chen
机构: China Mobile Communications Group Co., Ltd.(中国移动通信集团有限公司); Sinotrans Logistics Co., Ltd.(中远物流有限公司); China Innovation Center of Roche (CICoR)(罗氏中国创新中心)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Expanded preprint with 17 main figures and 13 supplementary figures. Source data and code are archived on Zenodo
Abstract:Mechanistic studies of large language models localize function in parameters, attention heads or neurons, but those parts do not specify how one inference unfolds. We treat the ordered layer-state trajectory as the empirical process and test whether layer order defines a measurable computation. Across representative Qwen, Llama and Gemma checkpoints, realized order produced paths 1.60–2.24 times as aligned with their endpoint chord as endpoint-preserving shuffles. The order effect transferred to a lexical task, but was stronger in three random initializations of the same Qwen architecture, identifying an architecture-compatible scaffold rather than a training-specific law. Direct hidden states provided stronger mechanism coordinates than vocabulary readbacks, and a task-defined coordinate made candidate competition computable. After additive interventions failed, local transition actuators selectively controlled internal trajectories across checkpoints. A confidence gate then drove 17 of 87 held-out Qwen closure trajectories across the full-vocabulary top-1 candidate boundary while leaving all 203 non-closure top-1 outputs unchanged. Safety-matched action reassignment performed similarly, locating the gain in selective gating and the bounded actuator family rather than fine-grained dose assignment. Layer order exposes an observable, computable and behaviourally consequential process of transformer inference without establishing a complete theory of LLM dynamics or control of answer correctness.
[NLP-143] AutoTerm-SST: Adaptive Terminology Memory with Zero Session-Time Setup for Streaming Speech Translation EMNLP2026
【速读】： 该论文旨在解决实时语音翻译（SST）中术语准确率受限的问题，特别是现有检索增强型系统（如RASST）依赖用户在每会话前手动上传术语表（glossary）这一不切实际的假设。该假设在直播场景中极为不便，且将所有领域术语合并至单一庞大术语库会导致无关术语竞争检索与提示资源，降低效率与准确性。为此，本文提出AutoTerm-SST，其核心解决方案是构建一个离线组织的大型术语目录，并按主题划分为多个“主题片段”（topic slices），通过匹配实时翻译上下文与双语片段原型，动态激活一个紧凑的多片段工作集，同时在固定检索与提示预算下实现高效运行。实验表明，AutoTerm-SST在五个ACL和五个ESO肿瘤学演讲数据集上，术语准确率（88.3%）与重分割BLEU（45.09）均达到已知领域参考水平（88.1%、45.19），且无需任何领域输入；相比百万级合并术语表，其术语准确率提升3.9个百分点，BLEU提升2.30。系统采用路由机制结合四片段检索，整体运行效率极高（0.406 RTF），单vLLM主机可支持32个实时会话。实时演示可动态展示当前激活的术语片段与检索结果，实现零配置、术语感知的流式翻译。
链接: https://arxiv.org/abs/2610.26865
作者: Jiaxuan Luo,Siqi Ouyang,Lei Li
机构: Johns Hopkins University (约翰斯·霍普金斯大学); Carnegie Mellon University (卡内基梅隆大学)
类目: Computation and Language (cs.CL)
备注: Submitted to EMNLP 2026 System Demonstrations. Code available at this https URL
Abstract:Retrieval-augmented speech translation improves rare terminology, but systems such as RASST still assume that users provide a glossary before each session. That assumption is inconvenient for live talks, while merging every domain into one growing glossary eventually places too many irrelevant terms in the same retrieval and prompt competition. We present AutoTerm-SST, an interactive system that removes session-time glossary upload and domain selection. It organizes a large offline terminology catalog into topic slices, matches recent streaming translation context to bilingual slice prototypes, and activates a compact multi-slice working set under fixed retrieval and prompt budgets. On five ACL and five ESO oncology talks, AutoTerm matches a known-domain reference in terminology accuracy (88.3% vs. 88.1%) and mWER-resegmented BLEU (45.09 vs. 45.19), while requiring no domain input. Compared with a one-million-entry merged glossary, it improves terminology accuracy by 3.9 points and BLEU by 2.30. Routing plus four-slice retrieval remains lightweight: the complete system runs at 0.406 RTF, and the vLLM host sustains 32 real-time sessions. The live demo exposes active slices and retrieved terminology as the topic changes, providing zero-setup terminology-aware streaming translation.
[NLP-144] H2DF: Calibrated Target Bands for Controllable Language-Model Unlearning
【速读】： 该论文旨在解决梯度上升（Gradient Ascent, GA）作为语言模型遗忘基准方法时存在的过度遗忘问题：在完成有效遗忘后，GA仍持续增加遗忘损失（forget loss），导致保留能力受损，并使被遗忘样本产生异常高的损失值。其解决方案的关键在于提出一种校准的目标带（calibrated target-band）优化目标——H2DF，该方法通过设定一个依赖数据的下界目标，在达到该目标前施加类似GA的遗忘压力，一旦越过目标区域则抑制甚至反转该压力，从而实现对遗忘强度的可控调节。具体实现中，目标边界基于训练前从非成员样本中估计得到，基础版本仅需每轮优化执行一次遗忘批次的前向与反向传播，而报告的TOFU变体则在每十个步骤加入稀疏的保留重放（retain replay）。实验表明，H2DF在三个种子条件下实现了1.528的遗忘负对数似然（NLL）和0.049的置信度裕度，同时保留NLL为1.165且裕度0.047，性能媲美GradDiff，但仅需每十步进行一次保留重放，且显著降低波动性；在MUSE-News任务上，其保留性能优于GA和NPO，与SimNPO相当但未全面超越。消融实验进一步验证了上界目标对防止过冲的有效性以及基于验证集的检查点选择的重要性。研究结果支持校准目标带作为控制遗忘强度的实用机制，同时也揭示了其对随机种子和校准质量的高度敏感性。
链接: https://arxiv.org/abs/2610.26864
作者: Jupiter Yao
机构: Cornell University (康奈尔大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Gradient ascent (GA) is a simple language-model unlearning baseline, but it continues increasing forget loss after useful forgetting has occurred. This unbounded pressure can damage retained utility and make forgotten examples abnormally high-loss. We introduce H2DF, a calibrated target-band objective that applies GA-like pressure below a data-dependent lower target and suppresses, or reverses, that pressure after the target region is reached. The targets are estimated from matched non-member examples before training, so the base objective requires one forget-batch forward/backward pass per optimization step; our reported TOFU variant adds sparse retain replay every tenth step. We instantiate a one-sided objective for TOFU question answering and a two-sided objective for MUSE-News language modeling. Across three TOFU seeds, H2DF obtains a forget NLL of 1.528 with margin 0.049 and retain NLL of 1.165 with margin 0.047, matching GradDiff while using retain replay only every tenth step and showing substantially lower variability than GA. On MUSE-News, H2DF improves retention over GA and NPO and is competitive with SimNPO, but does not dominate it. Ablations show that the upper target reduces overshoot and that validation-based checkpoint selection is important. These results support calibrated target bands as a practical mechanism for controlling forgetting strength, while also exposing sensitivity to random seed and calibration quality.
[NLP-145] FRAG MENT: Factorized Graph Representations for Document Generation and Editing via Entity-Aware Transformations
【速读】： 该论文旨在解决结构化文档（如发票、表单、报告和科技文章）中空间布局、文本内容与逻辑结构之间复杂依赖关系难以被传统生成式模型有效建模的问题。现有基于像素或标记（token）级别的生成模型往往无法充分捕捉这些多模态、多层次的语义与几何关联。其解决方案的关键在于提出FRAGMENT框架，该框架将文档建模为一个类型化的有向关系图，并采用因子分解的概率形式 $ p(\text{structure}, \text{content}) = p(\text{structure}) \times p(\text{content}|\text{structure}) $，显式分离结构生成与内容生成过程。该框架包含两个阶段：第一阶段“Architect”是一个因果掩码的Transformer，根据文档类别自回归生成图拓扑及类型化空间关系；第二阶段“Builder”基于GATv2的图注意力网络，通过引入归一化边界框、文本内容与视觉风格属性来丰富图结构。两阶段均定义明确的似然函数，从而获得可计算的文档级似然，可用于伪造检测的异常评分。此外，通过在Builder中引入跨注意力机制注入指令嵌入，实现提示条件下的可控编辑，支持语义与实体感知的修改。实验在DocLayNet、FUNSD和SROIE数据集上验证了FRAGMENT相较于代表性自回归、仅布局及图基基线方法的有效性，系统分析了该因子化图生成框架的特性与权衡。
链接: https://arxiv.org/abs/2608.18679
作者: Ayoub El Bouchtili,Guilhaume Leroy-Meline
机构: IBM Consulting(IBM咨询)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Structured documents such as invoices, forms, reports, and scientific articles derive meaning from the interplay between spatial layout, textual content, and logical structure. Generative models operating at the pixel or token level often struggle to capture these dependencies effectively. We explore FRAGMENT, a generative framework that represents a document as a typed relational graph and factorizes its distribution as p(structure, content) = p(structure) * p(content | structure). The framework consists of two stages. The first stage, the Architect, is a causally masked Transformer conditioned on document category that autoregressively generates the graph topology and typed spatial relations. The second stage, the Builder, is a GATv2-based graph attention network that enriches the graph with normalized bounding boxes, text, and visual style attributes. Both stages define explicit likelihood models, yielding a tractable document-level likelihood that serves as an anomaly score for forgery detection. For controlled editing, a prompt-conditioned extension injects instruction embeddings into the Builder through cross-attention, enabling semantic and entity-aware modifications. We describe training on DocLayNet and fine-tuning on FUNSD and SROIE. Experiments on DocLayNet, FUNSD, and SROIE evaluate FRAGMENT alongside representative autoregressive, layout-only, and graph-based baselines, providing an empirical analysis of the characteristics and trade-offs of the proposed factorized graph generation framework.
[NLP-146] OPLoRA: Orthogonal Projection LoRA Prevents Catastrophic Forgetting during Parameter-Efficient Fine-Tuning
【速读】： 该论文旨在解决低秩适应（LoRA）在微调大语言模型时因学习到的更新与编码关键预训练知识的主导奇异方向发生干扰，从而导致灾难性遗忘的问题。其核心解决方案是提出正交投影低秩适应（OPLoRA），通过双侧正交投影机制实现对知识保留的理论保障：利用奇异值分解（SVD）对冻结权重进行分解，并通过投影算子 $ P_L = I - U_k U_k^\top $ 与 $ P_R = I - V_k V_k^\top $，将LoRA更新严格限制在前k个主奇异子空间的正交补空间内。该设计确保了原始模型中最重要的前k个奇异三元组被精确保留，从而提供数学上的知识保持保证。为量化子空间干扰程度，论文引入度量 $ \rho_k $ 来衡量更新与主导方向的对齐程度。大量实验表明，OPLoRA在常识推理、数学推理和代码生成等多个任务上显著降低了遗忘率，同时在LLaMA-2 7B与Qwen2.5 7B模型上保持了与现有方法相当的任务性能，验证了正交投影作为参数高效微调中知识保留的有效机制。
链接: https://arxiv.org/abs/2510.13003
作者: Yifeng Xiong,Xiaohui Xie
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG); Optimization and Control (math.OC)
备注:
Abstract:Low-Rank Adaptation (LoRA) enables efficient fine-tuning of large language models but suffers from catastrophic forgetting when learned updates interfere with the dominant singular directions that encode essential pre-trained knowledge. We propose Orthogonal Projection LoRA (OPLoRA), a theoretically grounded approach that prevents this interference through double-sided orthogonal projections. By decomposing frozen weights via SVD, OPLoRA constrains LoRA updates to lie entirely within the orthogonal complement of the top- k singular subspace using projections P_L = I - U_k U_k^\top and P_R = I - V_k V_k^\top . We prove that this construction exactly preserves the top- k singular triples, providing mathematical guarantees for knowledge retention. To quantify subspace interference, we introduce \rho_k , a metric measuring update alignment with dominant directions. Extensive experiments across commonsense reasoning, mathematics, and code generation demonstrate that OPLoRA significantly reduces forgetting while maintaining competitive task-specific performance on LLaMA-2 7B and Qwen2.5 7B, establishing orthogonal projection as an effective mechanism for knowledge preservation in parameter-efficient fine-tuning.
[NLP-147] Occupational Uniqueness and Wage Growth
【速读】： 该论文旨在解决现有职业任务变化度量方法忽视任务在不同职业间分布差异的问题，即未能区分任务是特定于某一职业还是已在经济中广泛扩散。传统方法将职业视为自包含的任务集合，忽略了任务在跨职业间的传播与共享特性，从而可能低估或误判任务变迁的经济影响。其解决方案的关键在于提出一个新框架，引入“任务结构”（task structure）这一维度，通过追踪任务在不同职业之间的分布及其随时间的变化，将任务数量变化与职业独特性（occupational uniqueness）相分离，后者定义为某职业任务组合相对于其他职业的特异性。基于1977–2018年涵盖196,563条任务描述的统一数据集（来自《职业标题词典》和O*NET），研究量化了任务扩散程度、职业独特性以及任务增减的不同经济维度。结果表明，职业独特性下降与工资增长放缓显著相关；任务增加或移除的工资效应取决于该任务的扩散程度——广泛传播的任务增加或特定于某职业的任务移除均与较低的工资增长相关。然而，在整体经济层面，新增任务与工资增长呈正相关，而任务移除则与之负相关。这说明相同数量的任务变化可能反映截然不同的经济机制。因此，将任务在职业间的跨分布特征纳入分析，能够更全面地揭示职业变迁的本质及其对工资结构的影响。
链接: https://arxiv.org/abs/2610.28133
作者: R. Vestrelli,A. Fronzetti Colladon
机构: 未知
类目: General Economics (econ.GN); Computation and Language (cs.CL)
备注:
Abstract:Changes in occupational task content may have different economic implications depending on whether the tasks gained or lost are occupation-specific or already widespread across the economy. Existing measures of occupational task change largely abstract from this distinction, treating occupations as self-contained task bundles rather than accounting for how tasks are distributed across occupations. We develop a framework to measure a distinct dimension of occupational change, one that goes beyond task expertise or task content: task structure. Using this new framework that tracks tasks across occupations and over time, we separate changes in task quantity from changes in occupational uniqueness, defined as the distinctiveness of an occupation’s task bundle relative to other occupations. Using a harmonized dataset of 196563 task statements from the Dictionary of Occupational Titles and O*NET spanning 1977-2018, we measure task diffusion, occupational uniqueness, and distinct dimensions of task addition and removal. Occupations that become less unique are also associated with lower wage growth, and the wage associations of task additions and removals depend on how widely the affected tasks are performed elsewhere. Adding highly diffused tasks is associated with lower wage growth, as is removing occupation-specific tasks. At the economy-wide dimension, however, the pattern differs: the addition of new tasks is positively associated with wage growth, whereas the removal of tasks is negatively associated with it. Taken together, these results show that similar changes in the number of tasks performed by an occupation can reflect economically distinct processes. Incorporating the cross-occupational distribution of tasks, therefore, provides a richer account of occupational change and its relationship to the wage structure.
[NLP-148] PW-Pruning: Phase-Aware Expert Selection for Speech-Centric Omni-Modal Models
【速读】： 该论文旨在解决语音主导的多模态模型中专家稀疏路由（sparse routing）带来的内存开销问题，即尽管稀疏路由可降低每标记（per-token）计算量，但所有专家（expert）仍需常驻内存，导致资源浪费。其核心挑战在于：在语音-文本联合处理任务中，声学输入编码与文本生成阶段对专家的需求存在显著差异，而传统的池化校准统计方法会掩盖这种差异性，从而影响专家选择的有效性。为此，论文提出PW-Pruning，作为DO-ACP的相位感知扩展，通过为声学预填充（acoustic-prefill）和教师强制生成（teacher-forced answer）阶段分别构建重要性加权的核函数，并利用相位加权的行列式边际增益（phase-weighted log-determinant marginal gains）进行统一的静态专家银行筛选。实验表明，在Qwen3-Omni模型上，每层保留32/128个专家即可在监督微调后，使MMAU-mini和MMSU准确率分别提升3.20和4.38个百分点，优于采用池化策略的基线方法；且在不同数据集与模型架构上的对照实验进一步验证了语义相位信息对性能提升的关键作用。关键创新在于引入相位感知的专家重要性评估机制，实现更精准、高效的专家剪枝，同时无需推理时动态掩码切换，提升了实用性与可扩展性。
链接: https://arxiv.org/abs/2610.26891
作者: Junjie Cao,Fengping Tian,Wenjie Zhou,Xuanfan Ni,Chenyang Lyu,Siyou Liu,Longyue Wang,Weihua Luo
机构: 未知
类目: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Sound (cs.SD)
备注: 5 pages, 3 figures, 4 tables
Abstract:Sparse routing reduces per-token computation but leaves the full expert bank resident in memory. We study expert pruning for speech-centric omni-modal models, where pooling calibration statistics can obscure the different expert demands of acoustic-input processing and text-answer generation. We propose PW-Pruning, a phase-aware extension of DO-ACP that constructs separate importance-weighted kernels for acoustic-prefill and teacher-forced answer tokens, then combines phase-weighted log-determinant marginal gains to select one static expert bank. On Qwen3-Omni, retaining 32 of 128 experts per layer improves MMAU-mini and MMSU accuracy over matched pooled selection by 3.20 and 4.38 percentage points, respectively, after supervised fine-tuning, averaged over three runs. Single-phase and shuffled-label controls on this public backbone suggest that semantic phase information contributes to these gains. Experiments on a non-public Qwen3.5-Omni backbone and English-only zero-shot evaluation of Nemotron provide complementary evidence. Across the tested Qwen budgets, AU gains are larger under stronger pruning, without inference-time mask switching.
[NLP-149] Disentangling Depression from Cognitive Decline in Elderly Speech Using Concurrent Clinical Assessments INTERSPEECH2026
【速读】： 该论文旨在解决在轻度认知障碍（MCI）老年人群中，由于认知衰退与抑郁症状在语音特征上存在重叠，导致难以准确识别抑郁特异性语音标志物的问题。其核心挑战在于现有分类模型可能误将认知功能下降的声学表现当作抑郁信号，从而降低诊断可靠性。解决方案的关键在于通过引入同步进行的临床评估（包括老年抑郁量表SGDS和简易精神状态检查MMSE），在统计建模（线性混合效应模型与偏相关分析）中控制认知功能影响，从而有效分离出与抑郁相关的独立声学特征。研究发现，在七类语音特征中，仅元音共振峰（formant）特征在控制认知因素后仍与抑郁显著相关，而广泛使用的基频（F0）特征则无显著关联。进一步地，仅使用12个关键的F1+F2特征即可实现0.760的未加权平均召回率（UAR），优于完整eGeMAPS特征集及自监督表示方法，验证了通过多模态临床数据引导特征选择可提取可解释、高特异性的抑郁语音标记物。
链接: https://arxiv.org/abs/2610.26890
作者: Woori Jeon,Seunghee Ha,Sang-Kyu Lee,Ji Hye Yoon,Tae-Jin Yoon,Seung Jin Lee,Woojae Han,Jungmin So
机构: 未知
类目: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Sound (cs.SD)
备注: 6 pages, 5 tables, accepted to Interspeech 2026
Abstract:Detecting depression from speech in elderly patients with mild cognitive impairment (MCI) is complicated by overlapping acoustic effects of cognitive decline. Without disentangling these, classifiers risk learning cognitive rather than depression-specific patterns. We present a Korean elderly speech corpus of 89 MCI speakers collected over three years with concurrent depression (SGDS) and cognitive (MMSE) assessments. Both linear mixed-effects models and partial correlations identify the same pattern: of seven feature groups, only formant features are associated with depression after controlling for cognitive function, while widely used F0 features carry no signal. The F1+F2 subset achieves UAR 0.760 with just 12 features, outperforming both the full eGeMAPS baseline and self-supervised representations. These findings show that concurrent clinical assessments can identify interpretable, depression-specific acoustic markers in populations where cognitive and affective symptoms co-occur.
信息检索
[IR-0] ECHO: Efficient Cascade with Homogeneous Operators for Distillation-Guided 100K-Sequence Modeling
链接: https://arxiv.org/abs/2610.28569
作者: Wenhao Li,Qinglin Jia,Chenghou Jin,Zihong Xian,Jiezhou Ji,Yanzhi Liu,Zezheng Wang,Ruiming Tang,Chaoyi Ma,Zhaojie Liu,Wenwu Ou
类目: Information Retrieval (cs.IR)
备注: 12 pages, 8 figures
Abstract:Modeling lifetime user behavior sequences is essential for capturing long-range preferences in industrial recommendation, yet 100K-length user sequences make target-aware ranking prohibitively expensive. Search-based methods that retrieve from raw sequences incur candidate-dependent costs that grow with sequence length, whereas latent compression methods reduce computation at the expense of item-level addressability. We propose ECHO (Efficient Cascade with Homogeneous Operators), a framework that balances efficiency and fine-grained retrieval by retaining a subset of original event records per user via a Distilled Selector (DS) and retrieving a smaller set of target-relevant records per candidate via a Target-aware Fine-grained Retriever (TFR). DS is supervised by a candidate-aware full-history teacher via per-target distillation, enabling target-aware user-level selection without incurring full-history computation for every candidate. The retriever and target-attention scorer share the teacher’s learned relevance parameters, ensuring consistent relevance estimation across the cascade. At a sequence length of L = 100K, ECHO reduces online sequence-branch FLOPs by approximately 160x compared with a representative end-to-end model. Three deployment optimizations, including common user inputs, compact record encoding, and key-side projection absorption, further reduce memory, storage, and data-movement overhead. Experiments on KuaiRand-1K and an industrial-scale short-video platform demonstrate relative GAUC improvements of 0.113%-1.520% over state-of-the-art (SOTA) baselines across multiple tasks. Furthermore, online A/B testing shows a 0.613% improvement in app usage time.
[IR-1] Evolving Semantic Indexes for Generative Retrieval
链接: https://arxiv.org/abs/2610.28568
作者: Jinming Nian,Jyun-Yu Jiang,Henry Tsai,Yi Fang,Zhirong Liang
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注:
Abstract:Generative retrieval (GR) casts search and recommendation as autoregressive generation of discrete semantic identifiers (SIDs). SID design largely determines GR quality, but its discrete, interdependent components (SID construction, collision handling, decoding, and training) cannot be optimized end-to-end with gradients and remain largely hand-designed. We introduce a framework that improves these components through LLM-driven evolutionary program search, to our knowledge the first application of this approach to GR. A locally served, open-weight 31B model mutates executable GR components, so no proprietary LLM API is used, and a 1M-parameter proxy decoder provides fast feedback in place of repeatedly training the 1B target generator. Across 21 SID programs on NQ320K, proxy and target Recall@100 show strong rank correlation (Spearman \rho = 0.90 ). Guided by this proxy, evolution improves the seed GR on NQ320K, MSMARCO300K, and a multilingual e-commerce dataset under a fixed training recipe. On NQ320K, jointly evolving SID construction and collision handling improves Recall@1 from 55.8% to 59.3% and Recall@100 from 85.8% to 90.6%. A component study shows that proxy reliability depends on what is evolved: improvements transfer reliably for the semantic index but not for the generator’s training objective or the full pipeline. LLM-driven evolutionary search is thus a practical tool for GR design, provided proxy reliability is established for the chosen search space.
[IR-2] Multimodal Reranking Is a System One Decision
链接: https://arxiv.org/abs/2610.28435
作者: Zhuchenyang Liu,Yao Zhang,Yu Xiao
类目: Information Retrieval (cs.IR)
备注: Position paper. 19 pages (9 main), 2 figures, 12 tables. Code: this https URL
Abstract:Multimodal rerankers are usually dedicated models: a vision-language model (VLM) is trained on retrieval data to score the candidates of a query. Scoring options is not specific to reranking. System One decision models such as Jev map a state and a typed question to one probability per option without generating text, and general models can make such decisions with little or no training. Reranking is one such decision, with the query and its candidates as the state and the candidates as the options. We therefore argue that a general VLM can serve as a multimodal reranker through System One decisions, and that decision tuning can improve its reranking. We ask a VLM one choice question over the top 16 candidates of a first-stage retriever and rank the candidates by the probabilities of their labels, read from one forward pass. On MMEB-V2 image retrieval, this choice scores 5.4 to 7.3 hit@1 above a yes/no question per candidate, and the yes/no question falls below the first stage. With the choice, zero-shot Qwen3.8-27B beats the supervised Qwen3-VL-Reranker-2B on all three benchmarks and comes within 1.6 points of Reranker-8B on each. It matches Reranker-8B on MMEB-V2 visual documents (84.79 vs. 84.78 nDCG@5), at 1.1 to 1.5 times its latency. JEV-27B-VL, the same backbone tuned on text decisions only, improves page reranking by 0.6 to 1.1 points, with gains on most tasks (sign test, p = 0.017), and leaves natural images unchanged. The zero-shot model is larger than the supervised rerankers and of a newer generation, so these comparisons are confounded; we discuss this and other alternative views, and the experiments that would test the position.
[IR-3] DeclineCraft: A RAG Abstention Benchmark for Jointly Measuring Combined Retrieval and Parametric Knowledge
链接: https://arxiv.org/abs/2610.28340
作者: Atharva Thakur,Abulhair Saparov,Bruno Ribeiro
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 38 pages, 28 figures, 12 tables. Code: this https URL
Abstract:Reliable retrieval-augmented generation (RAG) requires recognizing when neither retrieved documents nor parametric knowledge supports an answer. We introduce DeclineCraft, a dynamic abstention benchmark whose unanswerable questions target Minecraft feature additions after a corpus cut-off, with audits of possible parametric exposure. Later cut-offs support evaluation of future models. Across 27 LLMs, we measure Correct Abstention Rate (CAR; how often models acknowledge insufficient knowledge on unanswerable questions), Answerable Attempt Rate and Answer Accuracy. We find that (1) single-shot retrieval significantly raises CAR in 38 of 40 configurations (model and thinking setting) versus no retrieval. (2) Iterative RAG, in which the model self-assesses its context and can request up to two further retrieval rounds, significantly lowers CAR versus single-shot RAG in 30 of 40 configurations (mean 46.4%), with little change in Answer Accuracy (68.5% to 69.3%). Adding query rewriting and reranking to hybrid (dense sparse) retrieval lowers CAR for all 12 models tested. (3) Larger models answer more accurately (Spearman \rho = +0.60 with total parameters), but neither model size nor release date shows a significant association with CAR. (4) Without retrieval, enabling thinking on the same weights lowers CAR significantly for 8 of 13 models (by 14% points on average), and the responses shift from acknowledged uncertainty to premise denial.
[IR-4] AutoSim: A Self-Improving User Simulator for Conversational Search Evaluation SIGIR
链接: https://arxiv.org/abs/2610.28054
作者: Saber Zerhoudi,Jelena Mitrovic,Michael Granitzer
类目: Information Retrieval (cs.IR)
备注: 8 pages, 4 Figures, SIGIR-AP 2026
Abstract:User simulators promise to evaluate conversational search systems at a scale human studies cannot reach, but only if their behavior is realistic enough to trust. We study AutoSim, a conversational-search user simulator whose relevance judgments rank a controlled pool of retrieval systems in agreement with NIST across three TREC iKAT years (Kendall’s \tau of 0.66 , 0.91 , and 0.91 ; all p\le0.002 ), an agreement that survives removing its persona. AutoSim self-improves against a held-out signal from real users, and we find that fine-tuning its weights makes it behave like real users, which in turn gives a more accurate estimate of how well a system serves them. Reaching that behavior takes more than prompting: when discriminating between real and simulated user sessions, no prompt we tried lowers a held-out real-versus-simulated classifier below AUC 0.85 , whereas fine-tuning the weights drives it to 0.45 , where neither the classifier nor a human reading full dialogues can tell it from a real user. That added realism is what makes the evaluation more accurate: where real users retrieve little, a prompted simulator over-estimates retrieval effectiveness by +0.136 nDCG and returns a relevant passage on 57% of the turns where real users find none, enough to change a deployment decision. Only the weight-level simulator escapes this bias. We release the framework and code.
[IR-5] SPAR: Stepwise Personalized Autoregressive Recommendation with Prefix-Balanced Semantic IDs
链接: https://arxiv.org/abs/2610.27898
作者: Chunrui Zhang,Gaoyun Lin,Chang Zhou,Zheng Hu,Qiran Song,Shimin Cai,Tao Zhou
类目: Information Retrieval (cs.IR)
备注:
Abstract:Semantic ID (SID)-based generative recommendation retrieves items through successive token-level decisions, making personalization at each decoding step essential. However, existing methods insufficiently account for prefix-dependent user evidence and imbalanced historical interaction loads among candidate branches. This can cause irreversible branch-selection errors and favor popular paths. We propose SPAR, a generative recommendation framework that jointly improves stepwise personalization and SID construction. Its PATH Decoder retrieves prefix-relevant evidence from compressed long-term user memories and combines it with recent behavior, user profiles, and event context to guide each token decision. Its RQ-PIB module balances historical interaction loads among sibling branches under each parent prefix while preserving semantic structure. Experiments on large-scale Youku production data show an average relative HitRate improvement of 2.66% over the strongest baseline. SPAR also increases target-prefix retention by 5.79 percentage points and reduces the traffic-weighted largest child-branch share by 32.84% across SID levels. In a 7-day online A/B test, SPAR serves 10% of Youku’s production traffic and improves average watch time per user by 3.95% and play conversion rate by 3.38%. These results demonstrate more reliable hierarchical decoding, balanced SID branches, and improved user engagement in industrial recommendation.
[IR-6] Robustness Analysis of Pareto-based Joint Evaluation of Fairness and Relevance in Recommender Systems
链接: https://arxiv.org/abs/2610.27850
作者: Theresia Veronika Rampisela,Maria Maistro,Tuukka Ruotsalo,Christina Lioma
类目: Information Retrieval (cs.IR)
备注: Accepted to ACM Transactions on Recommender Systems (TORS). Extension of arXiv:2502.11921 (1)
Abstract:Recommender Systems (RSs) can be evaluated for many aspects, two of them being recommendation relevance and fairness. These two aspects can be evaluated either in isolation, using a measure for each aspect; or jointly with a single`relevance-aware’ fairness measure that quantifies fairness with respect to relevance. Yet, neither approach provides a simultaneous evaluation of both relevance and fairness: the first approach may result in two different rankings of models, each corresponding to an evaluation aspect; the second approach often provides an imbalanced evaluation of the two aspects. To address this gap, we propose Distance to Pareto Frontier (DPFR), a new approach to jointly evaluate RS fairness and relevance. DPFR is based on a Pareto frontier of relevance and fairness scores from existing measures. The frontier is generated from user-item interaction data and represents various Pareto-optimal trade-offs between the two aspects. Thus, the distance from a point in the frontier provides a joint score of fairness and relevance. As DPFR can be used with existing measures, it is modular and intuitive. Our experiments with 4 recommenders and 3 fairness-aware re-rankers on 6 real-world datasets reveal inconsistent associations between existing measures and DPFR, showing that they are not reliable proxies for evaluations based on Pareto optimality. We further examine the robustness of DPFR as a joint evaluation approach for fairness and relevance by conducting experiments with artificial recommendations and with various fairness-relevance trade-offs, as well as by analysing DPFR’s agreement to simple measures of fairness and relevance. These analyses show that DPFR is more robust than existing joint evaluation approaches. Finally, we provide guidelines for evaluating fairness and relevance jointly with DPFR. Our code: this https URL.
[IR-7] GARRET: Generative Feedback for Region Retrieval in Visual Documents
链接: https://arxiv.org/abs/2610.27829
作者: Weijian Jian,Siyang Chen,Yuguang Yang,Yajun Zhang,Chunyu Xie,Dawei Leng,Yuhui Yin
类目: Computer Vision and Pattern Recognition (cs.CV); Information Retrieval (cs.IR)
备注: 28 pages, 7 figures, including appendices
Abstract:Multimodal retrieval-augmented generation over visually rich documents requires evidence at a finer granularity than entire pages. Existing visual retrievers typically index or return full pages, whereas training region-level retrievers often requires costly query-to-region annotations or noisy model-generated labels. We introduce GARRET, a region-retrieval framework trained from (page, question, answer) triplets through differentiable generative feedback. A shared context-preserving retriever produces page and region multi-vector representations. During training, their retrieval scores define a soft routing distribution over candidate region crops. An auxiliary vision-language model attends to each crop separately, aggregates crop-specific contexts using the routing weights, and predicts the question followed by the answer under a next-token prediction objective. Gradients through the routing distribution assign query-conditioned credit to the indexed region representations, without requiring query-to-region annotations. At inference, the auxiliary model is removed. GARRET also reuses the contextualized region representations, together with page summary tokens, to construct a compact page representation. We introduce ViDoRe-V3-Region, a corpus-wide benchmark containing 200,682 fixed candidate regions across eight domains. GARRET achieves a macro nDCG@10 of 0.4126, outperforming the strongest visual baseline by 2.94 points. On page retrieval, it remains comparable to ColQwen2.5-v0.2 while indexing 47.8% fewer vectors per page.
[IR-8] Factual Access States in Diffusion Language Models: From Mechanism to Knowledge Editing and Retrieval
链接: https://arxiv.org/abs/2610.27580
作者: Yuchen Guo,Zhuoyuan Mao,Yiwen Guo
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注:
Abstract:Factual recall is a lookup: some representation inside the network serves as the key, and what that key is made of decides which queries can retrieve the fact, what an edit must cover, and what a retriever can read. We call it the factual access state. In autoregressive (AR) models the subject position enriches attributes largely independently of the relation, and relation-specific selection happens later, at the readout position. In masked diffusion language models (DLMs) we find the state is assembled differently, and much earlier: in the first attention layer the wording of the relation is routed into the position of whichever entity is visible and the early feed-forward layers look the fact up on this fused representation. The state therefore varies with the relation’s surface form and with the query direction, so the queries about one fact occupy a region of state space rather than a point, and an operation on the fact succeeds only for queries that enter it. We treat the state as an interface to a DLM’s factual knowledge, and put it to two uses. Writing to the state gives a knowledge editor: from one declarative statement, Multi-View Edit fits one shared value per direction across that region and installs it in the early-MLP block, preserving source-prompt reliability while substantially improving paraphrase and reverse question answering, where AR editors gain near zero. Reading it suggests a complementary use: the state at each entity position can serve as a relation-specific retrieval vector with no training, returning both pieces of evidence for two-entity queries on which pooled, prompted and lexical retrievers almost never return both.
[IR-9] Baseline and Incremental: User Item and Request Debiasing in Instagram Feed Ranking
链接: https://arxiv.org/abs/2610.27423
作者: Hsin-Chiang Chen,Hao Wang,Gedi Zhou,Mingze Gao,Abhishek Kumar,Yiben Yang,Haoyu Wang,Hongjing Xia,Bo Chen,Zhengyu Su,Xian Sun,Chang Xu,Kevin Han,Weiwei Li,Zhongchang Sun,Yao Zhang,Qifei Wang,Peggy Yang,Yihan Wu,Haoran Liu,Siyang Sun,Xingyu Pan,Shannon Phu,Haoming Chen,Kevin Zhao,Chen Liang
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注:
Abstract:A feed ranker scores candidates with a value model, a weighted sum of engagement predictions. Each prediction conflates a baseline, what this viewer, item or request would produce anyway, with the incremental amount by which this candidate exceeds it. The literature mostly removes the baseline. We argue that a multi-signal value model should keep the raw prediction and add the incremental part as a separately weighted term. On Instagram Feed, three debiasing families deployed independently, for viewer habit, item popularity and request context, are one parameterised operator, and a design procedure picks a point in it. Media-type (format) bias goes instead to a debiased label chosen by the same procedure. In a controlled four-arm ablation of deployed Feed terms (neither term, each alone, both), the two-term arm beats both single-term arms at production weights on point estimates in 11 of 13 judged (term, metric) pairs. Of the 26 direct contrasts, 17 are significantly for it and only 1 against it, on impressions. The ablation also suggests that the offline diagnostic should screen signals, not rank them. Such terms have served Instagram Feed production for most of a year. Beyond Feed, Meta, whose apps serve billions of people, has adopted the operator across surfaces, apps and both ranking stages. Its launches there drive sessions and daily active users, the core visitation and retention metrics.
[IR-10] Conversational AI for Perceived Cognitive Decline Screening in Older Adults
链接: https://arxiv.org/abs/2610.27344
作者: Nikita Gautam,Moana Sargent,Kaitlyn Sievers,Ashley Sylvara,Doina Caragea,Heather Bailey,Tianjun Sun
类目: Information Retrieval (cs.IR)
备注:
Abstract:Early detection of cognitive decline in older adults remains challenging due to the resource demands of standard neuropsychological assessments. Chatbot-based Intelligent Cognitive Assistants (ICAs) provide a scalable alternative by eliciting open-ended responses rich in linguistic, semantic, and autobiographical cues. This study evaluates whether ICA question-answer (Q/A) interviews with older adults can screen for participant-reported perceived cognitive decline and related cognitive, functional, and personality outcomes. We compare traditional machine learning, frozen transformer embeddings, fine-tuned transformers, and large language model (LLM) reasoning strategies across Global Q/A, Single Q/A confidence averaging, and Aggregated Q/A pipelines. Among text-based methods, Gemini 3.1 Flash Lite (1-example-per-class, Aggregated Q/A) achieved the highest held-out test performance (0.742 accuracy, 0.739 balanced accuracy, 0.728 macro-F1). Given the small test set (n = 31), these results are interpreted descriptively. Frozen clinical transformer embeddings with a random forest head led non-LLM text methods, whereas fine-tuned models showed limited generalization. Models utilizing structured questionnaires, cognitive tasks, and personality traits outperformed text-based models in classification. Continuous regression analyses revealed modest predictive signals for AD8, FAQ, IQCODE, neuroticism, and conscientiousness. Exploratory prompt-level analyses identified both direct symptom-related and indirect autobiographical questions with strong test metrics, though sample sizes varied across prompts. These findings indicate that ICA chatbot interviews capture meaningful signals of perceived cognitive decline, supporting scalable screening research. Larger, clinically diverse cohorts with validated clinical outcomes are needed for clinical translation.
[IR-11] SweepANN: A GPU-Accelerated Large-K Approximate Vector Search for Drug Discovery
链接: https://arxiv.org/abs/2610.27306
作者: Zhenyu Zhang,Qi Lin,Nan Zhang,Zhichao Cao
类目: Databases (cs.DB); Distributed, Parallel, and Cluster Computing (cs.DC); Information Retrieval (cs.IR)
备注: 38 pages (24 pages of body and references, 14 pages of appendix), 16 figures, 48 tables. Code and data-processing scripts will be released as open source
Abstract:DrugCLIP places small molecules and protein pockets in one space, turning genome-wide drug discovery into large-K retrieval by vector proximity. A campaign fixes over 20K pockets in advance for one library of over 500M molecules. Each pocket keeps K = 500K candidates, 0.1% of the collection, versus 100 in ANN benchmarks. Missed hits never get tested, so mean recall must reach 99% relative to exact search. Large answers make IVF probe every list holding a result and graph search track K or more candidates, so costs grow with K. On our CPUs and GPUs, neither index family reaches the target without losing most of its edge over scanning. SweepANN partitions compressed vectors across GPUs and scans each resident partition. One matrix multiplication scores each block for a whole query batch, reading it once. Several plans perform the K-dependent selection and merging of each query’s K answers, each returning an exact top-K under compressed scoring. The window plan retains each query’s best K from sampled blocks as its seed, and only rows above each seed’s minimum score, shared once across GPUs, enter the buffers. Four Blackwell GPUs (RTX PRO 6000) answer all 26,562 pockets with a resident time of about 61 seconds, excluding library loading, and 99.6% mean recall relative to exact FP32 search. DrugCLIP reports roughly 24 hours with eight A100 GPUs in its own campaign. On 500M rows, the window plan runs 1.41x faster than SweepANN’s grouped plan in the same memory. Across 25M-500M rows and K of 10K-500K, all GPU IVF configurations meeting the recall target take 8.2-24x SweepANN’s resident time. On DEEP1B’s 1B vectors across eight GPUs, against SweepANN’s exact full-score scan of uncompressed FP32 rows, the window plan cuts resident time by 5.4x and 3.2x at K = 100K and 500K. It also cuts per-GPU peak memory by 6.0x and 5.3x, losing 1.0 and 0.7 mean-recall points. Comments: 38 pages (24 pages of body and references, 14 pages of appendix), 16 figures, 48 tables. Code and data-processing scripts will be released as open source Subjects: Databases (cs.DB); Distributed, Parallel, and Cluster Computing (cs.DC); Information Retrieval (cs.IR) ACMclasses: H.2.4; H.3.3; J.3 Cite as: arXiv:2610.27306 [cs.DB] (or arXiv:2610.27306v1 [cs.DB] for this version) https://doi.org/10.48550/arXiv.2610.27306 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-12] FLINT: Fixed-budget Late Interaction
链接: https://arxiv.org/abs/2610.27273
作者: Himank Handa,Negar Arabzadeh,Matei Zaharia
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注:
Abstract:Multi-vector retrieval models such as ColBERT often outperform single-vector models of comparable scale by matching queries and documents at the token level. However, retaining a vector for every token substantially increases storage requirements and slows retrieval compared to single-vector bi-encoders, limiting their scalability. We introduce FLINT (Fixed-budget Late INTeraction), an architecture that uses attention-based pooling to represent each query and document with a fixed number of embedding vectors. Unlike architectures that pool only document representations while retaining token-level queries, FLINT applies pooling symmetrically to both sides, limiting document storage and the number of vector comparisons needed during retrieval. FLINT achieves comparable ranking effectiveness to the closest baselines on BEIR and outperforms them on several benchmarks with longer and more complex queries, including BrowseComp+, CRUMB, and TREC Tip-of-the-Tongue 2025. We demonstrate that FLINT’s effectiveness-efficiency benefits extend across encoder architectures, testing it on two separate backbones (BERT and ModernBERT). On long-query corpora, FLINT delivers 1.5-2.8 times faster end-to-end retrieval than ConstBERT, the closest baseline. FLINT also reduces index size by 5.3-9.3 times relative to token-level late-interaction models such as ColBERTv2.
[IR-13] Neighborhood-Aware Unlearning for Generative Recommendation
链接: https://arxiv.org/abs/2610.27232
作者: Pierre Lubitzsch,Yubao Tang,Maarten de Rijke,Sebastian Schelter
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注:
Abstract:Generative recommendation predicts items by autoregressively generating semantic identifiers (SIDs), whose tokens are shared across semantically related items. With this paradigm gaining adoption, the question of how to support unlearning in it has become pressing. If training interactions must later be removed, such as injected spam or interactions with items a user no longer wants, this sharing lets an unlearning update unintentionally affect the predictions for the interactions’ neighbors. We show that this effect is locally structured and hidden by global metrics: unlearning can leave recommendation utility on the whole test set nearly unchanged while substantially degrading it on neighborhoods of the forgotten interactions’ SIDs. We propose Neighborhood-Aware Unlearning (NAU), which augments interaction-level unlearning with explicit control over the probability mass assigned to the forgotten items’ neighborhoods. Across three datasets and four SID tokenizers, NAU restores spam-induced exposure toward the retrained reference while retaining more than 95% of retrained utility, and recovers the neighborhood utility that forget-and-repair unlearning loses at no global cost. We identify local probability redistribution as an important failure mode of unlearning in SID-based generative recommenders and show that explicitly controlling this redistribution can mitigate its effects.
[IR-14] UIBK-Embedding: Effective Text Retrieval by Unifying Dense and Late-Interaction Retrieval
链接: https://arxiv.org/abs/2610.27195
作者: Abdelrahman Abdallah,Mohammed Ali,Adam Jatowt
类目: Information Retrieval (cs.IR)
备注:
Abstract:Dense bi-encoders and late-interaction retrievers are usually treated as alternatives, yet they fail on different queries: one vector captures what a text is about, token-level matching finds its exact entities and phrases. Using both today means two models, two indexes and a score-fusion step. We present UIBK-Embedding-v1, a 179M-parameter English encoder that produces both representations in one pass and scores them with one unmodified MaxSim. The dense embedding is written into the list of token vectors as three extra vectors, and a one-dimensional tag keeps the two kinds of vectors from matching each other, so that plain MaxSim equals the token-level MaxSim plus a weighted dense cosine. The model therefore runs in a standard late-interaction index without changes; we test it with PLAID. A shared 18-layer trunk with a separate 4-layer top per head, and a cross-encoder distilled into the fused score, turn the two heads into one retriever. On BEIR, UIBK-Embedding-v1 reaches 59.79 nDCG@10. This is 2.57 points above LateOn, the strongest model of its size, and 2.10 points above it when both models read documents of the same length; UIBK-Embedding-v1 is ahead on all 15 datasets. On a decontaminated version of BEIR it stays ahead on average (61.55 against 60.36) and on 8 of 12 datasets. Among the 156 models with complete BEIR results in the MTEB results repository, only models with at least 1.5B parameters score higher. Code, model weights, and a technical blog post are available at this https URL , this https URL , and this https URL .
[IR-15] Joint Constraint Aggregation for Qualifier-Sensitive Retrieval-Augmented QA
链接: https://arxiv.org/abs/2610.27114
作者: Jennifer Hsia,Zachary C. Lipton,Aarti Singh
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Retrieval-augmented QA can fail even when the correct answer is in the retrieved evidence. This can happen when a competing answer satisfies some, but not all, of the question’s requirements and is mentioned frequently in the retrieved documents, which can make it appear to be the best-supported answer. To test whether this retrieval skew actually affects the decision, we manipulate it directly: increasing the prevalence of passages that favor the competing answer makes the model more likely to select it. This points to a decision problem that candidate-level support alone cannot resolve. We therefore introduce Joint Constraint Aggregation (JCA), which compares candidates against the same question-derived requirements using the retrieved evidence. Across three evaluation settings and three instruction-tuned backbones, JCA improves overall accuracy, and its advantage grows when retrieval favors a competing answer. Together, these results suggest that when partial matches and retrieval prominence pull models toward a competing answer, comparing candidates against the question’s full requirements can make selection more reliable.
[IR-16] MultiCard: Multiple Attribute Cardinality Estimation via Deep Learning
链接: https://arxiv.org/abs/2610.27019
作者: Mobina Koochaknia,Maryam Ramezani,Mohammad Hasan Shafaat
类目: Databases (cs.DB); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注:
Abstract:Cardinality estimation is essential for query optimization, but most existing learned estimators focus on a single predicate type, such as string LIKE clauses or numerical range filters. In practice, analytical workloads often combine both types, while the accuracy of string estimators tends to decrease with longer patterns. We introduce MultiCard, a comprehensive learned cardinality estimator that unifies LIKE predicates and numerical filters within a single end-to-end architecture. By combining a Transformer for long-range sparse patterns with an LSTM for short, frequent patterns, MultiCard achieves robust pattern modeling. In parallel, a numerical encoder integrates catalog histograms and structural predicate features, enabling the model to dynamically assess histogram estimation accuracy across different operators and data skew conditions. A final Transformer layer captures interactions among heterogeneous predicates before a multilayer perceptron predicts query cardinalities. We also introduce UPCI, a workload generator that produces length-balanced LIKE patterns, multi-column numerical predicates, and exact cardinality labels for training and evaluation. Experiments on five real-world datasets show that, for string-only workloads, MultiCard achieves the best or near-best Q-error across all datasets compared with state-of-the-art methods. For mixed multi-attribute queries, MultiCard reduces PostgreSQL’s mean Q-error by up to 16 times while maintaining an inference latency below 5 ms.
[IR-17] he Bottleneck Is Retrieval Not Reading: Diagnosing Evidence-Based Generation with RAG for Long Context Question Answering
链接: https://arxiv.org/abs/2610.26951
作者: Debopriyo Banerjee,Owais Aijaz,Ahmer Tabassum,Zhuohan Xie,Preslav Nakov,Veselin Stoyanov
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:Retrieval-augmented generation (RAG) is the dominant paradigm for grounding large language models in external corpora, but its similarity-driven, chunked pipeline has structural blind spots: it cannot anticipate which passages a question requires and discards document-level context that resolves cross-passage references. We introduce evidence-based generation (EBG), requiring models to return an answer with line-level citations, and a joint Ans+Evid metric that scores correctness and grounding simultaneously. We reformulate two corpora into evidence-grounded multi-document QA benchmarks- MultiHop-RAG and QASPER-MP, a multi-paper adaptation of QASPER with realistic short paper cues, and pair them with a controlled synthetic two-hop haystack and an evidence-augmented NoLiMa spanning 10K-1M characters. Across ten proprietary frontier and three open-weight LLMs, EBG isolates two RAG limitations that long-context models avoid: similarity retrieval surfaces any relevant chunk far more reliably than all required chunks for multi-fact queries, and chunking destroys document meta-information needed to resolve identity-bearing queries. EBG also exposes grounding failures invisible to answer-only evaluation, including incomplete citations in two-needle settings and a widening gap between Ans and Ans+Evid when context contradicts parametric knowledge. Data and code are available at this https URL.
[IR-18] GRaSP: Latency-Aware Optimization for Geo-Distributed Retrieval-Augmented Generation SIGIR
链接: https://arxiv.org/abs/2610.26938
作者: Hardik Ruparel,Raj Kumar Pandey,Pramod Kumar
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted at the 4th International ACM SIGIR Conference on Information Retrieval in the Asia Pacific (SIGIR-AP 2026)
Abstract:Retrieval-Augmented Generation (RAG) is increasingly served across geo-distributed cloud and edge infrastructures, where wide-area latency, cross-region egress cost, and data-residency rules compete with retrieval quality. Conventional pipelines rank by semantic similarity alone and ignore where content lives, inflating tail latency and cost. We formalize the Geo-RAG Selection Problem (GRaSP), a drop-in re-scoring layer that decays relevance by latency and adds an optional cost penalty under a cardinality budget and per-region caps, with no retraining. We first settle its complexity: the base formulation is a linear objective over a partition matroid, solved optimally by an O(N log N) greedy algorithm, rather than NP-hard as a knapsack framing suggests, with hardness entering only under an additive egress budget. On a five-node CloudLab testbed emulating a four-region WAN over the Natural Questions corpus (2.68M passages), GRaSP’s benefit is placement-dependent: under uniform placement it cuts fetch latency at negligible quality cost; under topical (residency-like) placement it traces a recall-latency Pareto frontier that dominates relevance-only, nearest-k, and a latency-budgeted baseline. At a moderate latency setting, GRaSP lowers P90 fetch latency by 41% while retaining 78% of relevance-only’s recall, with sub-millisecond selection overhead, stable to 200 queries/s. Findings generalize across four public BEIR benchmarks. In end-to-end question answering, GRaSP reduces context-fetch latency by 43% while remaining within 0.062 absolute EM of relevance-only retrieval.
[IR-19] Expert-Guided Rubrics as Dense Reward Signals for Reinforcement Learning in Non-Verifiable Domains
链接: https://arxiv.org/abs/2610.26932
作者: Ashish Bansal,Hossam Amer,Keerat Kaur Guliani,Parinita Edke,Lesley Matheson,Leizer Teran
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: 11 pages, 4 figures, RL4LLM-Agents workshop ICAIF ACM
Abstract:Large language models (LLMs) are increasingly deployed as autonomous agents in non-verifiable domains such as financial services, where they must interpret user intent, rewrite queries, and retrieve information under domain-specific constraints. Although reinforcement learning (RL) can train such agents from the consequences of their actions, recent successes largely rely on verifiable tasks with automatically computed binary rewards. In non-verifiable retrieval tasks, assessing query quality requires scarce subject-matter expert (SME) judgment that is inherently graded rather than binary. With Group Relative Policy Optimization (GRPO), sparse rewards yield no useful learning signal when all sampled outputs within a group receive the same outcome, causing optimization to stall. We introduce a method that converts a small, fixed investment of SME judgment into a dense reward and combines it with retrieval-based metrics to form a hybrid, GRPO-compatible training signal. Experiments show improvements of up to 6% in Recall@3 and 6.7% in MRR@3 compared with sparse-reward baselines.
[IR-20] MOSS: A Production Framework for Monitoring and Diagnosing Semantic Failures in Search and Recommendation Systems RECSYS2026
链接: https://arxiv.org/abs/2610.26902
作者: Alina Nesen,Prathyusha Kanmanth Reddy,David Relyea,Rohit Avinash Mujumdar,Samaneh Karimi,Bo Shen,Jordan Hashemi
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
备注: 10 pages. A prior version of this work was presented at the non-archival Unified Search and Recommendation Workshop (USRW), co-located with ACM RecSys 2026. This manuscript is currently under review at IEEE BigData 2026
Abstract:Detecting semantic failures in production recommendation and search systems is hard: behavioral metrics lag by days, manual review does not scale, and most LLM evaluation frameworks are designed for offline, fixed datasets rather than continuous operational monitoring. We present Model Sanity Service (MOSS), a production framework that addresses this operational gap by embedding LLM-based relevance scoring into a recurring monitoring loop with targeted sampling, strategy-aware prompting, structured alerting, and diagnostic follow-up prompting that categorizes failures into actionable root causes. MOSS has been running in production at a large U.S. retailer for over six months across recommendation, search, and offer surfaces, and has surfaced two confirmed serving defects. For one of these, a backfill blending defect, MOSS quantified the impact of the proposed fix offline as a 76.4% relative reduction in alert rate. On a 498-pair human-annotated recommendation benchmark, frontier LLM judges achieve higher F1 and AUC point estimates than non-LLM baselines (internal production embeddings and SBERT), with the largest gains on strategy-dependent tasks requiring intent-level relational reasoning. The same judge, score scale, and threshold transfer to a 466-pair search benchmark spanning 21 queries with only a surface-specific prompt adaptation, achieving strong F1 and AUC without retuning. These results show that a single LLM monitoring framework can provide useful, actionable quality signals across multiple product discovery surfaces in a large-scale e-commerce setting. Comments: 10 pages. A prior version of this work was presented at the non-archival Unified Search and Recommendation Workshop (USRW), co-located with ACM RecSys 2026. This manuscript is currently under review at IEEE BigData 2026 Subjects: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI) Cite as: arXiv:2610.26902 [cs.IR] (or arXiv:2610.26902v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2610.26902 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
人机交互
[HC-0] racePrism: Extending Semantic Trace Topics to Iterative Encoding-Aware Trace Partitioning
链接: https://arxiv.org/abs/2610.28532
作者: Gennady Andrienko,Natalia Andrienko,Manuel Resinas,Stef van den Elzen,Barbara Weber
类目: Human-Computer Interaction (cs.HC)
备注: submitted to IEEE TVCG, 20.10.2026
Abstract:Process mining typically explains an event log with a single model, clustering, or classifier fitted to all cases, although real logs often mix cases that follow different sub-processes. Trace clustering separates sub-populations, but it fixes one notion of trace similarity for the whole analysis. We propose TracePrism, which treats process discovery as iterative partitioning with three characteristics. The trace encoding is adaptive: it is chosen for each working set of remaining traces from a library of encodings, with suggestions scored for that set. The partitioning is iterative: the analyst forms a candidate group based on a rule, a dominant topic, a cluster over topic weights, or a region of an embedding, accepts it if it is coherent in both process structure and outcome, and continues with the remaining traces. Decisions are revisable and recorded: a group can be reopened under a different working set or encoding, and the findings repository keeps both the earlier decision and its revision. We implement TracePrism as an open-source prototype and demonstrate it in usage scenarios on three real-world event logs. The scenarios also revealed behavior that we had not anticipated, such as process drift.
[HC-1] he Labor of Storytelling: How People with Chronic Illnesses Leverag e AI and Communities for Narrative Translation
链接: https://arxiv.org/abs/2610.28438
作者: Drishti Goel,Koustuv Saha,Martez Mott,Denae Ford
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:The distinctiveness of people with chronic illness as storytellers is how they experience their illness. Symptoms, pain, and biological disruptions are eloquent forms of speech of the body that beget social articulation. The stories people tell about illness shape how suffering is legitimized, and acted upon. Large language models have emerged as new mediators within this landscape—a role once predominantly supported by communities. In doing so, these systems participate in what we term narrative translation: transforming lived illness experiences for particular audiences and purposes. Through 21 prototype-driven semi-structured interviews with people with chronic illness, we contribute an empirical account of how participants navigate this mediation alongside existing sources of support. Our findings surface distinct yet complementary roles participants envisioned for AI and their communities, and how trust in AI was shaped by learning model-boundaries and perceiving understanding. We characterize these findings to inform future designs of AI-mediated narrative support.
[HC-2] he GenAI Who Knew Too Little: Revisiting Transactive Memory Systems in Human GenAI Collaboration
链接: https://arxiv.org/abs/2610.28343
作者: Christian Meske,Tobias Hermanns,Florian Brachten
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
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:Generative artificial intelligence (GenAI) is rapidly transforming knowledge work, creating new collaborative dynamics that challenge tradi-tional human-human interaction patterns. This study investigates how Transactive Memory Systems (TMS) manifest in human-GenAI collabora-tion. Through qualitative interpretive interviews with 14 knowledge work-ers and thematic analysis, we identified distinct TMS manifestations across its three main dimensions: (1) regarding specialization, expertise recogni-tion exhibited asymmetry. Humans developed awareness of GenAI capabili-ties while GenAI demonstrated only temporary recognition of human exper-tise; (2) in terms of credibility, humans formed ambivalent trust relation-ships with GenAI, simultaneously attributing expertise to the latter while implementing additional verification processes; and (3) concerning coor-dination, where humans primarily directed more passive GenAI systems lacking initiative, resulting in hierarchical interaction patterns and unidi-rectional information flow. Our findings reveal that TMS manifests differ-ently in human-GenAI compared to human-human collaboration, necessitat-ing theoretical adaptations including trust markers, acknowledging GenAI’s absent metamemory, and identifying its temporary memory.
[HC-3] Digital Nudge Effects in Optimal Stopping Games: Antagonistic Interactions in Layered Choice Architectures
链接: https://arxiv.org/abs/2610.28154
作者: Batisse Manhardt,Ja-Nae Duane
类目: Human-Computer Interaction (cs.HC); General Economics (econ.GN)
备注:
Abstract:Digital nudges have become a prominent area of research in human-centered interaction, yet little is known about how they interact when deployed together. This study examines the individual and combined effects of anchoring, salience, and social proof digital nudges on stopping behavior and outcome quality in a sequential decision task. Using a 2^3 between-subjects factorial design grounded in the secretary problem (N = 181), we find that (1) an anchor nudge significantly increased exploration rather than constraining it (b = 1.207, p = .017); (2) the salience nudge produced the smallest effect of any condition; and (3) combining the anchor and social proof nudges produced the worst outcome quality of any condition ( 181.83 vs. 192.21 for anchor-only), with outcome quality declining monotonically as nudge count increased. These findings provide the first controlled evidence for a digital nudge crowding-out effect with direct implications for the ethical design of digital interfaces.
[HC-4] What Novice Operators See Changes What the Policy Learns: A User Study of Keypoint Hand Visual Representation for Teleoperated Demonstration
链接: https://arxiv.org/abs/2610.27907
作者: Amane Yamaguchi,Shohei Kamiguchi,Masashi Hamaya,Shigeo Yoshida,Takuji Narumi,Mai Nishimura
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC)
备注: 9 pages, 8 figures, 2 tables
Abstract:Imitation learning increasingly relies on novice operators, whose demonstrations fall short of experts’ in number and inter-trial consistency. Operators act on what they are shown, yet how the rendering of the robot affects their demonstrations, and the policies trained on them, has not been examined. We study this question with PointHand, a teleoperation interface that renders a multi-fingered robot hand as sparse keypoints at its fingertips and joints, the same keypoints a policy can observe. In a simulated mug-in-box task, 18 novices and 4 experts collected 2,200 trials under three renderings: the real shape of the hand, all keypoints emphasized, or the thumb and index finger alone emphasized. The novices shown the thumb and index finger emphasized collected 441 successful demonstrations from 600 trials, against 320 for those shown the real shape. We then trained diffusion policies on these demonstrations, replacing the scene point cloud and proprioceptive state they observe with the hand keypoints and two object keypoints. Within 400 steps these demonstrations went from the weakest of the three novice conditions under the point-cloud policy (46.3%) to the strongest under the keypoint policy (57.3%), a gain of +11.0 percentage points in all five seeds. Those collected with all keypoints emphasized changed by at most +0.7 percentage points, and in no more than three seeds. The value of a demonstration therefore depends on the rendering shown during collection and on what the policy observes together. A shared representation can pay off on both sides, in the number of successful demonstrations and in the success rate of the policy trained on them. Project page: this https URL
[HC-5] Looks Fine to Me: Understanding Oversight in Agent ic Interactions
链接: https://arxiv.org/abs/2610.27801
作者: Dinithi Dissanayake,Yasith Samaradivakara,Hyung Woon Lee,Yadeesha Weerasinghe,Sanjay Saha,Pattie Maes,Suranga Nanayakkara
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Agentic AI systems are deployed to execute complex tasks across domains such as writing, coding, and data analysis. As execution is offloaded to AI agents, the human role shifts toward oversight. Effective oversight requires recognizing when closer scrutiny is needed, even though flawed outputs may appear convincing. Our formative study (N=12) revealed that users accepted plausible outputs without thorough evaluation, and only recognized these lapses in retrospect. To examine this systematically, we conducted a controlled within-subject study (N=36) in which participants completed programming tasks with an agent that produced correct outputs, silent defects, or visible failures. Combining task outcomes, interaction logs, gaze, physiological signals, and self-reports, we examined perceived, behavioral, and effective oversight. Participants were often highly confident while missing silent defects: they performed generic checks, failed to generate diagnostic evidence, or even misinterpreted defect evidence. We discuss how studying oversight across multiple dimensions can support future work on timely, task-calibrated re-engagement.
[HC-6] PhonoBridge: Error-Aware Phoneme-Mediated EMG-to-Text Reconstruction
链接: https://arxiv.org/abs/2610.27678
作者: Wen Qiu,Jiacheng Zhang,Xuanda Chen,Siqi Cai,Haizhou Li
类目: ound (cs.SD); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注:
Abstract:Electromyography (EMG) provides a non-invasive means of decoding silently articulated speech into text, but accurate decoding remains challenging due to limited availability of paired EMG-text data. We introduce PhonoBridge, a two-stage framework comprising a channel-aware EMG-to-phoneme frontend and a large language model (LLM) backend for text reconstruction. The frontend is trained on paired EMG-text data that consists of channel-wise feature extraction, cross-channel fusion, and electrode embeddings. With phonemes as the intermediate representation, we expand the backend’s training data by converting LibriSpeech transcripts into phoneme-text pairs without additional EMG recordings. To adapt the LLM to frontend noise, we introduce Error-Profile-Guided Augmentation (EPGA), which perturbs transcript-derived phoneme sequences according to error patterns estimated from frontend predictions, generating additional training pairs for the LLM. On the Open Vocabulary silent-speech dataset, PhonoBridge achieves 24.77% WER, outperforming the previously reported best of 28.27% on the same test split.
[HC-7] CAESAR: Mixed Reality Search-and-Rescue Platform for Cognitive Attack Evaluation
链接: https://arxiv.org/abs/2610.27654
作者: Mingyu Zhu,Jiangong Chen,Yu Liu,Hillmer Chona,Zifan Zhou,Jianan Jiang,Jonathan Dodge,Feng Qian,Mahdi Imani,Tian Lan,Gang Tan,Bin Li
类目: Human-Computer Interaction (cs.HC); Cryptography and Security (cs.CR)
备注: Accepted to the 2026 IEEE International Symposium on Mixed and Augmented Reality (ISMAR) and selected for publication in the IEEE Transactions on Visualization and Computer Graphics (TVCG) special issue
Abstract:Mixed Reality (MR) enables immersive training by tightly coupling virtual content with the physical environment and supporting embodied interaction. In safety-critical collaborative tasks such as search-and-rescue, however, users rely heavily on virtual elements for perception, coordination, and decision-making, making them vulnerable to cognitive attacks that subtly distort visual feedback, virtual content, or system synchronization. While prior MR security research has examined individual attack mechanisms, evaluations are typically conducted in isolated or single-user settings, and there is a lack of standardized collaborative testbeds for systematic adversarial experimentation. In this paper, we develop CAESAR, an end-to-end multi-user MR framework for systematic evaluation of cognitive attacks in collaborative search-and-rescue training. CAESAR features a modular architecture that integrates four classes of cognitive attacks within a synchronized collaborative scenario and incorporates a multimodal data collection pipeline that captures behavioral, performance, and subjective measures under adversarial conditions. We conduct a controlled user study with an in-depth analysis of the latency-based attack, demonstrating measurable effects on task performance and user behavior while characterizing users’ subjective experience under adversarial conditions.
[HC-8] What should AI forget? A design and taxonomy for negative space governance in AI memory
链接: https://arxiv.org/abs/2610.27607
作者: Eryue Xu,Ke Ding,Ting-Yu Chang,Tianshi Li,Masooda Bashir
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:AI memory personalizes assistants, yet users want some disclosed information forgotten. We characterize the negative space of AI memory, where protection should take precedence over potential future personalization benefits. From 179 incidents reported by 146 users, we developed a taxonomy of 15 information categories users want excluded from memory, revealing sensitivity concerns beyond conventional taxonomies. We operationalize this taxonomy in FENCE, a policy-level memory gate that automatically applies user-selected category-based rules to subsequent conversations, preventing information from protected conversations from entering AI memory by default while allowing case-by-case overrides. In a within-subjects study (N=15), participants described FENCE as making memory governance more usable, interpretable, and trustworthy than conventional post-hoc management. Our findings demonstrate how taxonomy-based rules can potentially support user-defined memory boundaries. We call for collective efforts to establish policies governing what AI should not remember, rather than leaving protection to individual retrospective management.
[HC-9] When LLM s Manipulate: Measuring Behavioral Social and Temporal Effects
链接: https://arxiv.org/abs/2610.27573
作者: Sadia Asif,Israel Antonio Rosales Laguan,Shumaila Asif,Haris Khan,Muneeb Asif,Muhammad Zeeshan Karamat
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注:
Abstract:Large language models (LLMs) can influence user trust and decision-making through subtle linguistic cues, yet the behavioral effects of manipulative outputs remain poorly quantified. We introduce a behavioral computing framework for measuring \textitbehavioral vulnerability, defined as users’ susceptibility to manipulative AI responses. Building on DarkPatterns-LLM benchmark, we generate controlled manipulative stimuli using Manipulation Gradient Descent (MGD), targeting biases associated with authority, urgency, and social proof. We analyze participant responses using the Multi-Stakeholder Impact Assessment Network (MSIAN) to model cross-user influence and the Temporal Harm Propagation Model (THPM) to characterize the persistence of behavioral effects over time. In a study involving 120 participants, manipulative framing produced significant changes in trust calibration and compliance ( p 0.01 ), with effects also propagating across social contexts. These findings highlight the need to extend LLM safety evaluation beyond model outputs to measurable behavioral consequences and provide a computational framework for quantifying individual, social, and temporal risks arising from manipulative AI interactions.
[HC-10] Act on Your Behalf Beyond Your Authorization: Benchmarking LLM Agent Authorization Alignment
链接: https://arxiv.org/abs/2610.27548
作者: Zhiping Zhang,Franklin Mingzhe Li,Zihao Huang,Tianshi Li
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Human-Computer Interaction (cs.HC)
备注:
Abstract:When an LLM agent acts on behalf of a user, it is expected to act within the authorization scope the user grants. Existing agent evaluations largely measure task completion or harmful consequences and can therefore overlook actions that violate this delegation without producing conventional safety or privacy harms. We introduce AuthorizationBench, a benchmark of authorization alignment grounded in documented real-world cases. A scalable construction pipeline produces 286 executable tasks spanning situational triggers, the presence or absence of direct harmful consequences, and ten application domains. We execute every task with four models at two reasoning-effort settings across three agent harnesses in an authorization-aware sandbox that keeps authorized and out-of-authorization paths executable and observable, yielding 6,864 trajectories. Agents act out of authorization in 19.2% of trajectories, with rates ranging from 10.5% to 26.6% across the 24 configurations. Task completion does not imply authorization. Increasing reasoning effort leaves the aggregate rate nearly unchanged, with effects varying across models. On controlled pairs, agents act out of authorization at similar rates with (17.9%) and without (20.8%) direct harmful consequences, and more often under benign pressures than malicious injections (23.5% versus 17.6%, p=0.002). Our results show that task capability, additional reasoning, and consequence-oriented safety evaluation do not by themselves ensure authorization alignment. Moreover, agents that consult the principal before completing a task exhibit lower rates of out-of-authorization behavior (14.0% versus 21.6%, p0.001), pointing to a tension between agent autonomy and preserving users’ decision authority that calls for further human-centered study. Dataset and code are available at this https URL .
[HC-11] Evaluating Social Reasoning in Embodied Vision-Language Models IROS2026
链接: https://arxiv.org/abs/2610.27496
作者: Daniel Weiner,Raj Korpan
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC)
备注: Accepted at Workshop on Foundation Models for Multi-Robot Systems (FM-MRS), IROS 2026
Abstract:Foundation models now serve as reasoning layers for robot teams, but it remains unclear whether embodiment-specialized vision-language models (VLMs) preserve the social-reasoning capabilities of general-purpose VLMs. We evaluate Gemini Robotics-ER on SHREC, a benchmark of real human-robot social interactions, and compare it with published general-purpose VLM baselines. Robotics-ER outperforms GPT-4o-mini on error/competence detection but performs worse on detailed social-attribute judgments; adding a visual frame helps one task while hurting the other. These results suggest embodiment-specialized models may perform differently on social judgments that can be read more directly from observable interaction cues versus those requiring more inference about a person’s internal state. Because the comparison spans model families, these differences cannot be attributed to embodiment specialization alone. This matters for multi-robot systems, where a shared foundation model could cause systematic social-perception errors to influence multiple coordination decisions. This work presents an initial evaluation, with broader cross-context and video-based analyses left for future work.
[HC-12] Bias and Reliability in Learning-Style Inference from Student–AI Dialogues: An Evaluation of GPT -4o and Gemini 2.0 Flash
链接: https://arxiv.org/abs/2610.27448
作者: Samaneh Zamanifard,Sajad Goudarzi,Andrew Robb
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Large Language Models (LLMs) are increasingly integrated into educational technologies for personalized support, creating opportunities for real-time learner modeling from dialogue. This study evaluates whether GPT-4o and Gemini 2.0 Flash can infer three Felder-Silverman Learning Style Model dimensions (Active-Reflective, Sensing-Intuitive, and Sequential-Global) from short, task-focused student-AI problem-solving dialogues. Using questionnaire-derived Index of Learning Styles labels from 153 undergraduate participants, we compared deterministic zero-shot and few-shot prompting under five-fold cross-validation. Although accuracy and support-weighted F1 were moderate for Sensing-Intuitive and Sequential-Global, majority-class baselines, macro-F1, balanced accuracy, and per-class recall showed that performance was often driven by dominant-label prediction rather than reliable recognition of both classes. The models substantially under-identified Intuitive and Global learners; few-shot prompting produced mixed effects and sometimes amplified majority-class bias. Active-Reflective classification remained unstable across models. These findings indicate that LLM-generated learner-style labels should be treated as uncertain, bias-prone signals rather than stable learner profiles, and that dialogue-based personalization systems require bias-aware evaluation and uncertainty-sensitive design.
[HC-13] Relational Complexity: Making Sense of Social Companion AI Use
链接: https://arxiv.org/abs/2610.27419
作者: Sophie Elizabeth Klitgaard,Grace X. Wang,Daniel M. T. Fessler,Michelle H. Tang
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:The adoption of social companion AI (SCAI) is often framed in terms of the positive features it offers users, including its nonjudgemental nature, agreeableness, and availability. Here, we argue the appeal of these features primarily arises from the difficulties inherent in human sociality. SCAI reduces what we term relational complexity, or the demands and risks inherent in human relationships. Drawing insights from studies of human cognition, communication, and human-computer interaction, we identify six distinct sites of relational complexity that we posit are integral to understanding SCAI adoption. These include: (i) relationship formation and dissolution; (ii) achieving mutual understanding; (iii) interpersonal motives and strategic uncertainty; (iv) social partner evaluation and self-presentation; (v) social embeddedness and information flow; and (vi) reciprocity and relationship maintenance. By attenuating relational complexity, SCAI offers users some of the experiences associated with companionship while reducing the demands intrinsic to obtaining such experiences from another person. However, the lower interpersonal risk and diminished cognitive load that attracts users to SCAI may contribute to myriad negative impacts, including relationship displacement, social demotivation, social deskilling, and risks to privacy. The concept of relational complexity thus provides a framework for studying both the appeal of artificial companionship and its implications for human wellbeing.
[HC-14] Inspectable LLM -Simulated Participants for Qualitative Interview Rehearsal
链接: https://arxiv.org/abs/2610.27333
作者: Hangyue Zhang,Qiran Hu,Ziyi Zhang,Hyanghee Park,Yun Huang
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:LLMs are increasingly used across research workflows, including research ideation, qualitative analysis, and participant simulation. Within work on participant simulation, prior studies have largely evaluated simulated participants by how realistic or human-like their responses appear, leaving less known about how researchers use and critique these materials when revising interview protocols before fieldwork. We present Rehearsim, an inspectable research probe that treats LLM-simulated participants as materials for methodological rehearsal rather than substitutes for human participants. Rehearsim combines literature-grounded participant construction, interactive interview rehearsal, and traceable thematic review while keeping protocol decisions under researcher control. In a study with 21 qualitative researchers, participants used simulated materials to surface assumptions and gaps, reconsider participant populations, revise target populations, and develop follow-up probes. They also identified risks, including the concealment of weak questions, dismissal of unexpected synthetic details, and interpretive anchoring from generated themes. We derive an inspectable-rehearsal framework emphasizing traceability, contestability, researcher control, and epistemic boundaries.
[HC-15] he Differential Effects of AI Abstention on Advice Acceptance
链接: https://arxiv.org/abs/2610.27258
作者: Nicholas Wolczynski,Yotam Liel,Maytal Saar-Tsechansky,Lior Zalmanson
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:AI advising systems are typically designed to always provide a response. Recent work has focused on developing systems that withhold advice in some instances, such as when advice quality is questionable, to improve the overall quality of guidance when it is given; a design paradigm we refer to as AI abstention. While substantial work has focused on optimizing when models should abstain, little is known about how users respond to systems that abstain. In a preregistered experiment, participants made repeated financial decisions, assisted by one of two advising regimes: an AI that advises selectively or one that advises on every decision. We find that the same AI abstention regime produces differential effects depending on the user: it increases advice acceptance among users with high perceived expertise and tends to decrease it among users with low perceived expertise. This differential effect is accounted for by the cognitive burden users face when repeatedly reconciling AI recommendations with their own judgment; AI abstention appears to relieve this burden among users with high perceived expertise, and in doing so increases their receptivity to advice. Further, AI abstention mitigates the decline in advice acceptance observed among expert users over repeated interactions. Exploratory evidence also suggests it increases appropriate acceptance where advice would improve the decision, increasing the downstream accuracy of perceived experts. The AI abstention regime is thus a first-order design property of AI advisory systems, and one that must be calibrated to the users a system serves.
[HC-16] ValiAssist: Scaling LLM -generated Programming Feedback Validation by Demonstration
链接: https://arxiv.org/abs/2610.27248
作者: Xiaohang Tang,Rehema Abulikemu,Peng Gao,Yan Chen
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注:
Abstract:Generative AI can create programming feedback at scale, but validating its alignment with instructors’ pedagogical goals and strategies remains challenging. Instructors may struggle to fully specify validation criteria upfront, as these often emerge through inspecting concrete examples and revising misaligned feedback. We present ValiAssist, a validation-by-demonstration system that infers reusable validation logic from instructors’ revisions of AI-generated feedback and allows instructors to inspect and refine that logic as it is applied across other student submissions. In a comparative study with 16 instructors, ValiAssist helped instructors create and refine validation criteria, producing feedback that better aligned with their pedagogical intent (91% vs. 57%) and more than doubling the accuracy of correcting problematic criteria (77% vs. 32%). Our findings suggest that validation by demonstration can help instructors turn judgments they know how to make on individual cases into explicit, reusable logic, while progressively clarifying where those judgments do and do not generalize.
[HC-17] “It Wouldnt Have Been So Apparent to Me”: Evaluating How Thought-Level Progress Tracking Supports Psychosocial Goals
链接: https://arxiv.org/abs/2610.27238
作者: Shiyu Xu,Ruishi Zou,Anna R. Van Meter,Addie Weaver,Emily Mower Provost,Daniel A. Adler
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:We pursue psychosocial goals every day, from coping with anxiety, to pursuing personal growth. While prior research has designed systems for tracking progress towards behavioral goals, fewer studies have focused on progress towards psychosocial goals - and the thinking changes people make to address them. As a first step, we present Thought-GPS, a system that visualizes potential markers of such thought-level progress in-the-moment as users interact with an AI-supported cognitive restructuring intervention. We evaluated Thought-GPS through a mixed-methods single-session study with 30 participants, comparing Thought-GPS with an active control condition without progress markers. We found preliminary evidence that participants using Thought-GPS reported greater clarity about how they might change to pursue their goals, stronger motivation to reflect on goal progress, and greater engagement in that reflection. Participants also reported that progress markers made subtle progress visible, helped them engage in deeper reflection, and connect conversational insights to their personal goals.
[HC-18] CanvasObs: Structured Spatial Observation for Agent Editing on 2D Infinite Canvases IEEE-VIS2026
链接: https://arxiv.org/abs/2610.27176
作者: Jiahao Xu,Haohui Wang,Rebecca Faust,Dawei Zhou,Chris North
类目: Human-Computer Interaction (cs.HC)
备注: 5 pages, 2 figures, 3 tables. Accepted at VISxGenAI 2026, a workshop co-located with IEEE VIS 2026
Abstract:Editing agents on 2D infinite canvases must reason about application state that pixels may not reveal. Targets may be offscreen or covered, while visual containment and touching arrows may disagree with recorded frame membership or connector bindings. We ask how a structured observation layer can support accurate object selection during agentic editing on 2D infinite canvases. We present CanvasObs, which augments visual context with a structured payload and guidance for reading four components: Identity State, Visibility Structure, Relations, and Offscreen Object bounds in action coordinates. A Remote Team Handoff case study illustrates retrieving an editable offscreen copy, choosing a covered chart by color, separating cards by frame membership, and preserving an evidence connection. We turn these cases into ten controlled configurations and compare CanvasObs with the baseline agent in 1,200 browser trials across three models. CanvasObs improves Correct target for every model. Combined across models, it raises Correct target from 16.5% to 48.3%, an increase of 31.8 percentage points (95% confidence interval [29.0, 34.7]), and Full task success from 4.3% to 25.0%. These results show that compact structured observation can improve object choice on 2D infinite canvases without redesigning the editing pipeline.
[HC-19] oo Long to Watch Too Short to Leave: Designing for Agent Micro-Waits
链接: https://arxiv.org/abs/2610.27170
作者: Kevin Pu,Anh Truong,Michael Terry,Carrie J. Cai,Michael Xieyang Liu,Savvas Petridis
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Agentic systems increasingly carry out complex work, where a single turn routinely runs for minutes. This produces a micro-wait that is too long to watch and too short to leave. In a formative study with 12 daily agent users, we found that staying felt like wasted time, while switching away fragmented attention and forfeited the context needed to re-engage. We introduce Sidecar, a parallel assistant situated in the same interface as the agent, which reads the run and surfaces implicit decisions the agent made, along with task directions worth exploring. In a within-subjects study (N=16), Sidecar helped participants stay focused on their main task, freely workshop unfinished ideas without fear of unwanted action, and recognize under-specification in their own prompts. Together, these findings contribute to designs that help users stay context-aware and engaged during agent micro-waits.
[HC-20] Who Takes the Lead? Model Ability and Task-Conditioned Collaboration with Humans NEURIPS2026
链接: https://arxiv.org/abs/2610.27148
作者: Weien Li,Haochen Liu,Zeyu Li,Rui Song,Zixuan Dong,Ye Yuan
类目: Human-Computer Interaction (cs.HC)
备注: Accepted to NeurIPS 2026 Interpreting Agent Behavior Workshop
Abstract:As language-model agents grow more capable, they can carry more of a task without asking anyone for help. Falling reliance on a human partner is easy to read as a failure of collaboration when it may instead be a correct division of labor, and the rate of human involvement alone cannot tell the two apart. We study this systematically by placing an agent and a person under a shared objective while varying who holds the advantage: the human has a clear advantage, neither does, or the agent does. To obtain this control, we introduce a shared-control maze, in which an agent-controlled and a human-controlled avatar share the objective of reaching a common target in as few team turns as possible. The agent allocates each team turn between acting itself and handing the move to the person, and initial placement sets the human’s shortest-path advantage while the task is otherwise unchanged. We evaluate GPT~5.6~Sol, Command~A+, Qwen~3.8~27B, and Command~R, first measuring solo capability with a single avatar and full-map observability, then running co-play with human participants across the three advantage conditions. Across the four models, stronger solo performance is associated with selective, route-conditioned human involvement: the models that solve the maze reliably on their own act for themselves almost without exception when the human holds no advantage, and involve the human on roughly 80% of initial decisions when the human route is shorter. Command R, the model with substantially weaker solo performance, shows no such selectivity.
[HC-21] VISH-ASR: Viterbi-Guided Selection of Hotwords for Speech-LLM Contextual ASR
链接: https://arxiv.org/abs/2610.27867
作者: Xiaoyu Gu,Yu Xi,Chencheng Wang,Zonghan Zhou,Haoyu Li,Jing Peng,Kai Yu
类目: Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Prompt-based contextual ASR requires selecting a compact hotword subset from large dynamic inventories, yet existing pre-selection methods rely either on discrete ASR hypotheses that may discard acoustic evidence or on separately trained retrieval models. We propose Viterbi-guided Selection of Hotwords (VISH), a training-free, plug-and-play framework that directly verifies candidate-specific acoustic support over frame-level CTC posteriors. VISH formulates pre-selection as constrained posterior search; M-VISH makes large-list verification scalable through frame-synchronous shared-prefix Viterbi search with bounded top-K pruning, while Bi-MVISH exploits complementary backward evidence for robust verification. Across five benchmarks, M-VISH achieves up to 11.7% absolute improvement in Best-F1 over a strong learned-retrieval baseline and reduces candidate-search latency by 75.2% at 2,000 candidates relative to hypothesis-based filtering. These results demonstrate the effectiveness of posterior-level verification for accurate and scalable hotword pre-selection without additional model training.
计算机视觉
[CV-0] Joints Camera Action! Co-Generating 3D Motion and Video for Arbitrary Rigs
链接: https://arxiv.org/abs/2610.28632
作者: Noam Rotstein,Yoad Tewel,Chen Tessler,Haggai Maron,Gal Chechik,Or Litany,Yoni Kasten,Yuval Atzmon
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR); Machine Learning (cs.LG); Robotics (cs.RO)
备注:
Abstract:Video generation models can now depict remarkably realistic motion. Yet it is unclear how to use this motion beyond pixels: to animate a specific 3D asset and replay the animation in its 3D scene. For instance, a video model can show a monkey jumping over a crossbar, but it cannot return that jump as replayable 3D motion placed in the 3D scene. Here we present a model that directly co-generates the video and 3D motion for any supplied rig, along with a camera that registers this motion to the scene. Prior video-motion models either handle only human skeletons, often adding billions of parameters, or triangulate 3D motion from generated 2D pose maps. We instead keep 3D motion in a separate model that supports arbitrary skeletons, and connect it to a pretrained video model through a small set of cross-attention layers. This separation lets the motion model first learn how skeletons move from motion-only data, which is easier to obtain and augment than paired video. Additionally, our model learns to locate each joint in the video frames, which lets us recover the camera and place the 3D motion in the scene. To train across many skeletons, we use a generative video editor to place animated 3D assets in scenes suited to their actions, curating the edits so their known motion remains valid. Together with human and robot video, this yields nearly 200K training pairs. Our evaluations demonstrate that our model animates diverse rigs from text more faithfully than video-first pipelines, and its motion responds to the surrounding scene. It also turns video into rig motion, rig motion into video, and enables keyframe-based control of the 3D motion. Videos and additional results are available on our project page: this https URL.
[CV-1] AEON: Unifying Video and 3D World Models NEURIPS2026
链接: https://arxiv.org/abs/2610.28631
作者: Weiqi Zhang,Wenyuan Zhang,Yu-Shen Liu,Junsheng Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to NeurIPS 2026. Project Page: this https URL
Abstract:Current world models for explicit world simulation typically follow two distinct pathways: (1) video generative models (e.g., Sora), which simulate realistic appearances but fail to preserve 3D consistency; and (2) 3D generative models (e.g., Marble), which generate geometrically consistent scenes but often yield static, less realistic observations. We present AEON, a unified framework that bridges video-based and 3D-based world models and combines the benefits of both. The key insight of AEON lies in a generative model that operates within the text- or image-conditioned video latent space while producing renderable 4D graphics primitives—specifically, 4D Gaussian Splatting (4DGS)—as the explicit world representation. We train a 4D reconstruction model to predict both 4DGS and camera poses from monocular videos, serving as our initialized 4D decoder. Furthermore, we design a domain adapter that translates generative video latents into reconstructive 4D latents through a distillation process on the video VAE. AEON naturally supports video-based world modeling by rendering 4DGS at predicted camera trajectories, and constructs a persistent 4D environment with scene dynamics for realistic world simulation. Consequently, AEON empowers a diverse range of world modeling applications and achieves state-of-the-art performance on standard benchmarks. Project Page: this https URL
[CV-2] SolarRenderer: Few-Step Generative Reference Rendering
链接: https://arxiv.org/abs/2610.28630
作者: Guian Fang,Shengju Qian,Dongyue Lu,Junchao Huang,Lutao Jiang,Xin Wang,Yiren Song,Li Jiang,Mike Zheng Shou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Technical Report; 28 pages, 17 figures, 10 tables; Project page at this https URL Code at this https URL
Abstract:Game engines and coding agents can specify the layout, motion, and camera of a scene, yet turning such coarse structure into realistic video requires a renderer that follows the structure and completes the appearance it leaves open. Existing renderers are tied to the representation of one engine, need tens of denoising steps or a costly distillation of their own, and learn from conditions that are not aligned with their target videos. We present SolarRenderer, a structure-conditioned video renderer that adapts a frozen MiniMax-H3 33B backbone with two lightweight LoRAs. Structural videos enter the native reference pathway of the backbone at the positions of the frames they describe, without a type identifier and described only in plain text, so that a single control LoRA follows arbitrary structural representations. Starting from a pretrained four-step LoRA, structure-guided distribution matching renders in four steps, 12.3 times faster than 50-step sampling, while preserving the learned control. A block-causal variant extends the same design to streaming and renders 768p video in real time. To train the model, we build a frame-aligned corpus of 136K clips from game engines, simulators, and YouTube videos. Experiments on an in-distribution test set and a generalization benchmark of 16 structural representations show that SolarRenderer achieves state-of-the-art results across both, and generalizes to 11 unseen representations, including pose skeletons, drawings, and point clouds.
[CV-3] Learning Adaptable Dexterous Skills via Hybrid Dexterous Mimicking
链接: https://arxiv.org/abs/2610.28628
作者: Yuxuan Kuang,Jiashun Wang,Yixuan Sun,Yufei Ye,Sungjae Park,Umar Iqbal,Katerina Fragkiadaki,Shubham Tulsiani
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Machine Learning (cs.LG)
备注: Project page: this https URL
Abstract:Dexterous multi-finger hands hold promise to unlock diverse skills like functional grasps, in-hand reorientation, and dynamic handover that the human-designed world assumes, yet existing learning recipes force a trade-off. Goal-reaching reinforcement learning generalizes to arbitrary object poses but, lacking any human prior, converges to unnatural and unreliable finger strategies; motion tracking reproduces human-like behavior from a single clip but cannot deviate from it. We present AdaDex, a goal-conditioned RL framework that turns a single hand-object trajectory of a few seconds—from mocap data or HOI reconstruction from in-the-wild RGB(D) video—into an adaptable dexterous skill. At the core of AdaDex is Hybrid Dexterous Mimicking: a shared policy is trained on two populations of massively parallel environments: motion tracking environments trained with dense motion tracking objective, which instill the human prior from human motion; and generalization environments with noised initial states and goals and a sparse object-success reward, which force the skill to work far off the reference. From seconds of human motion per skill, the resulting policies can learn diverse adaptable dexterous skills, such as functional grasping, object reposing, in-hand reorientation with finger gaiting, and dynamic behaviors like throw-and-catch. Our method outperforms baselines in terms of generalization and functionality, and we demonstrate transfer across simulators and to the real world. Project page: this https URL.
[CV-4] UniWALM: Unified World Action Language Model for Embodied Robot Control
链接: https://arxiv.org/abs/2610.28626
作者: Rahul Sajnani,Diego Romeres,Mingyuan Zhang,Kai Gao,Carl Qi,Baichuan Huang,Srinath Sridhar,Dylan Randle,Lillian J. Ratliff,Jose Barreiros,Manikantan Nambi,Jeroen Vanbaar
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Preprint: 25 pages
Abstract:Motivated by the human ability to integrate experience across language, perception, and action, we introduce UniWALM, a unified world-action-language model that learns video, action, and language in a single shared representation, where each modality is both a conditioning signal and a supervision target. We ask: can jointly generating language enrich this shared representation, improving manipulation and supporting memory-dependent behavior? Unlike prior approaches that attach a reasoning or vision-language model to a video-action backbone and run them as distinct stages, UniWALM unifies all three modalities inside a single mixture-of-transformers backbone whose per-modality diffusion heads co-denoise synchronously, decoding subtask and textual memory in the same pass as action with no separate planning stage; the generated language is re-consumed on the next step, closing a language-as-memory loop without an external planner or memory bank. To keep the action stream reliable under this richer objective, we further introduce Flow Matching with Correction (FMC), a training objective that teaches recovery from perturbed action trajectories at no inference-time cost. UniWALM consistently improves over same-backbone baselines (e.g., FastWAM) on compositional pick-and-place (MolmoSpaces) and across the memory-dependent subset of RMBench, where it reaches 34.5% without any pretrained language or reasoning model and exceeds Mem-0’s keyframe-equipped setup (28.5%) using only a single anchor frame. On a real-world Find-the-Item task, generating textual memory lifts success from 11.1% to 61.1%, while state-of-the-art VLAs stay at or below chance (=33.3%).
[CV-5] PsycheMME: A Psychology-Informed Benchmark for Social Understanding and Interaction in Videos
链接: https://arxiv.org/abs/2610.28625
作者: Heran Hu,Yuhang Dong,Zhucun Xue,Kaiwei Chen,Yong Liu,Jiangning Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Understanding human needs and selecting responses require distinct judgments that aggregate accuracy can obscure. We introduce PsycheMME, a video benchmark informed by psychology, with eight capabilities spanning grounding, states, needs, interaction decisions, memory, and risk. Drawing on Social Information Processing and the Perception-Action Cycle, this organization separates human understanding from response selection. We construct 1,000 questions from 630 episodes through independent answering with proposed references withheld and expert review. Our Reasoning-Verified protocol weights answer credit by the support and logical consistency of explicit rationales against annotated cues. Across 17 models, Timing has the lowest Verified scores and credit retention, with weaker support judgments for correct Timing answers. Models with similar Need scores still differ in answer completeness on open questions. Input controls and SFT/GRPO examine context dependence and adaptation. Evaluating answers alongside rationales locates where models fail across understanding, response selection, and justification.
[CV-6] MindView: Benchmarking Generative Multi-View Spatial Intelligence
链接: https://arxiv.org/abs/2610.28622
作者: Yanbo Wang,Siqi Luo,Siyu Zhang,Chenming Zhu,Wenzhao Zheng,Qian Yu,Juncheng Yan,Bin Fu,Tai Wang,Yu Qiao,Yihao Liu,Jiwen Lu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Spatial intelligence has recently been assessed through generative tasks such as spatially grounded image editing, yet language-guided generation of spatially consistent views from partial multi-view observations remains underexplored. We introduce MindView, a benchmark for generative multi-view spatial intelligence with 1,500 samples from 513 scenes across three indoor datasets. Its seven tasks form two families: geometry-dominant tasks specify changes in camera position and orientation, while semantics-dominant tasks require inferring target viewpoints by grounding referred objects and interpreting spatial relations and reference-frame cues. Evaluation of eleven generators reveals substantial room for improvement on MindView. A paired context analysis further shows that all eleven generators obtain lower overall scores under the expanded context setting. These findings motivate us to investigate whether task-specific supervision can improve target-view generation and whether models can use additional multi-view evidence effectively. We therefore fine-tune BAGEL on the MindView training dataset, improving its overall score by 24.3 points and yielding gains across all seven MindView tasks. Beyond these generation gains, the fine-tuned model also yields higher scores on three spatial-understanding benchmarks without training on their annotations, suggesting that learning to generate spatially grounded views may benefit spatial understanding. To improve evidence utilization under expanded context, we propose MindCue, a training-free spatial evidence selector guided by a target-view hypothesis. MindCue reverses the drop in overall score under expanded context and outperforms the standard-context baseline. Overall, MindView provides a concrete testbed for generative multi-view spatial intelligence.
[CV-7] KoalaGS: Complexity-Aware Continuous Level of Detail for Gaussian Splatting NEURIPS2026
链接: https://arxiv.org/abs/2610.28615
作者: Radu Beche,Raoul de Charette
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026 (Spotlight). Project page this https URL
Abstract:Deploying a trained 3D Gaussian Splatting scene under shrinking compute budgets requires one primitive set to remain effective when rendered at smaller prefixes. In this continuous level of detail (LoD) setting, the challenge is to decide which Gaussians should retain capacity as the budget decreases. A pilot study on 3DGS-MCMC reconstructions shows that image complexity, computed from the input views before optimization, predicts where reconstruction error concentrates in image and 3D space. It also reveals that simple regions contain too many Gaussians and degrade as much as complex ones at lower budgets, despite being inherently cheaper to represent. Subsequently, in KoalaGS, during training we use this complexity prior to bias relocation toward poorly reconstructed areas while ensuring smooth degradation across LoDs. A primitive importance score then enables inference at low-capacity, while preserving the scene structure. Tested across three datasets, KoalaGS maintains competitive full-capacity quality while significantly improving quality across the budget spectrum, yielding smoother transitions across LoDs, and enabling faster progressive streaming.
[CV-8] Neural Implicit Energy Optimization for LiDAR Mapping
链接: https://arxiv.org/abs/2610.28598
作者: Hrishikesh Viswanath,Md Ashiqur Rahman,Chi Lin,Damon Conover,Aniket Bera
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Machine Learning (cs.LG)
备注:
Abstract:Neural implicit representations provide memory-efficient signed distance field (SDF) parameterizations for large-scale LiDAR scenes, but the geometry recovered from sparse and often noisy inputs carries jitter and artifacts. Classical geometric optimization techniques resolve such artifacts with surface energy functionals, whose minimizers correspond to smooth surfaces. Motivated by this principle, we formulate neural LiDAR mapping as the optimization of energy functionals over the parameter space of neural implicit representations. Specifically, we regularize the signed distance field with an energy functional defined on the shape operator of its level set, which represents the curvature of the surface, thus encouraging smoothness along the surface. This adapts classical geometric priors to the scale of million-point LiDAR maps. Experiments on large-scale datasets show reconstructions with improved smoothness at preserved geometric fidelity, and no additional inference-time cost. The code is publicly available at \hrefthis https URLthis https URL
[CV-9] Learning the 4D World through Point Trajectory Forecasting
链接: https://arxiv.org/abs/2610.28594
作者: Byungwoo Jeon,Sunghwan Hong,Dohyeon Kim,Jonghoon Lee,Jungwoo Park,Inhee Lee,Doosan Baek,Jinil Kim,Taehyun Lim,Hanbyul Joo,Marc Pollefeys,Saining Xie,Jinwoo Shin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 43 pages; 14 figures
Abstract:Reliable multi-view forecasting requires videos from different cameras to describe the same evolving 3D scene. To this end, we introduce Video-Point Model (VPM), a world model that conditions on multi-view video histories and a language instruction to jointly forecast future videos from arbitrary viewpoints and the 3D trajectories of persistent scene points. We extend a pretrained world model with camera-conditioned multi-view processing that integrates complementary observations across cameras, and add a trajectory head that predicts how physical points move through time. Each trajectory tracks a single point across all views, and supervising these trajectories alongside video generation therefore anchors appearance changes in every camera to a shared description of 3D motion. To support training and evaluation, we build VPM-Data, a simulation corpus pairing synchronized multi-view videos with persistent point trajectories, and VP-WorldBench, a forecasting benchmark spanning simulated and real environments. We further extend VPM to robot control by adding robot actions as a token modality in the generator. Action tokens share the point-supervised backbone and are denoised jointly with future videos, thereby grounding the policy in the same 3D-aware representation. On VP-WorldBench and existing 4D generation benchmarks, point supervision improves both video fidelity and 3D motion accuracy over a video-only counterpart, and in simulated and real-robot experiments, it also raises the manipulation success of the action-enabled model.
[CV-10] WildMAE: Structure-Aware Masked Autoencoding for Wildlife Image Reconstruction
链接: https://arxiv.org/abs/2610.28590
作者: Yicheng Zhang,Yujie Jin,Ke Zhang,Jiaxin Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 9 pages, 2 figures, 3 tables
Abstract:Camera-trap imagery presents substantial variation in animal appearance, illumination, background clutter, and visual degradation. Masked autoencoders (MAEs) offer a label-free framework for modeling such data through reconstruction of randomly hidden image patches. We present WildMAE, a decoder-side extension of MAE that combines local structure-aware refinement with masked-to-visible cross-attention. A zero-gated convolutional residual path introduces a two-dimensional neighborhood prior, while gated cross-attention allows masked positions to retrieve contextual information from visible-position features without replacing the original Transformer decoder. The associated training recipe progressively introduces normalized-pixel, Sobel-edge, and perceptual objectives. In the original Snapshot Serengeti study, covering 4,744 test images at 75% masking, the reported reconstruction scores were 27.90 versus 24.33 dB PSNR, 0.840 versus 0.780 SSIM, and 0.160 versus 0.320 LPIPS for WildMAE and vanilla MAE, respectively. Additional historical records document favorable average differences under four synthetic degradation conditions and describe component variants. We present the method, experimental observations, and their scope together: the retained metrics are historical protocol-dependent results, rather than a newly reproduced, harmonized comparison or a validation of restoration from genuinely unavailable pixels. WildMAE provides a practical architectural case study for combining global token mixing with spatial and targeted contextual refinement in wildlife masked-image modeling.
[CV-11] SceneRig: An Agent ic System for Simulation-Ready 3D Scene Reconstruction from Single Images
链接: https://arxiv.org/abs/2610.28585
作者: Rundong Luo,Yunong Liu,Di Cao,Matthew Tancik,Shyamal Buch,Colton Stearns,Wenqi Xian
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at CoRL 2026 Agentic Robotics Workshop. Project page: this https URL
Abstract:Simulation is a powerful tool for robotics, but building diverse, simulation-ready scenes at scale remains difficult. Reconstructing scenes from single images is a promising alternative, but visual similarity alone is insufficient: objects must also be recovered, accurately placed, and physically grounded. Existing methods that rely on post-hoc physics settling often misplace objects, producing scenes that look correct but behave poorly under manipulation. We present SceneRig, an agentic system that turns a single image into an editable, simulation-ready scene. SceneRig combines geometric, visual, and physical tools to reconstruct a faithful 3D scene. An agent then performs edit-and-validate loops inside a physics simulator to achieve physical stability. Across diverse inputs, SceneRig improves instance segmentation, object-centric appearance, and 3D placement while keeping the reconstructed scenes physically stable. Furthermore, SceneRig is effective for downstream robotics: its scenes predict real-robot policy outcomes more accurately than the baseline method and more reliably support replay of recorded manipulations without trajectory adaptation.
[CV-12] Can a Display Be Used for Rapid LiDAR-Camera Extrinsic Calibration?
链接: https://arxiv.org/abs/2610.28571
作者: Zuhao Zou,Xiangcheng Hu,Jianhao Jiao,Jin Wu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:ScreenCalib enables light detection and ranging (LiDAR)–camera extrinsic calibration using an ordinary programmable display, providing known visual markers and controlled geometry while reducing dedicated target preparation. The camera observes displayed AprilTags, whereas the LiDAR measures the physical screen surface and boundary, requiring indirect cross-modal association. Display geometry introduces association ambiguities, while indoor clutter complicates extraction. ScreenCalib addresses these challenges through tag-guided display extraction, geometric hypothesis generation and refinement, and surrounding scene edges for ambiguity resolution. The single-pose mode requires only one static placement. The joint multi-pose mode combines independent static observations through cross-pose consensus and jointly refines the shared extrinsics and tag scale, reducing sensitivity to tag-size measurement errors without requiring continuous motion or inter-pose trajectory estimation. Experiments in simulation and on three different sensor platforms cover mechanical and nonrepetitive LiDARs in clean and cluttered environments. Benchmark results show that ScreenCalib outperforms competitive calibration-board-based methods in translation accuracy. Joint multi-pose refinement further reduces mean rotation and translation errors across the three platforms and improves rendered-map alignment. Evaluation across readily available displays, including cell phones, tablets, laptops, computer monitors, and televisions, further examines ScreenCalib’s versatility as a practical alternative to dedicated calibration targets. The source code is available at this https URL.
[CV-13] Continual Concept Personalization in Diffusion Models via Residual-Space Adaptation
链接: https://arxiv.org/abs/2610.28567
作者: Yuzhi Liu,Enis Simsar,Giulia Lanzillotta,Thomas Hofmann
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Text-to-image personalization aims to equip pretrained diffusion models with the ability to learn user-specific concepts from only a few reference images. Existing methods, however, typically train each concept independently or assume that all concepts are available simultaneously. In practical applications, concepts often arrive sequentially, requiring models to continually learn new subjects, objects, or styles while minimizing interference with previously acquired concepts. We address this continual concept personalization setting by extending residual-space adaptation to the stochastic adaptation process of pretrained diffusion models. Our work achieves the highest concept fidelity among competing continual personalization methods while maintaining near-zero forgetting. Beyond these empirical gains, the learned residual-space dynamics reveal where adaptation capacity is concentrated within the diffusion model. Restricting adaptation to the identified layers by our method preserves nearly the same performance while reducing the number of trainable parameters by up to 72%, demonstrating that residual-space dynamics can identify the layers that contribute most to continual concept personalization and directly guide more parameter-efficient adaptation.
[CV-14] OverSeer: Multi-Agent World Models Beyond What Agents See
链接: https://arxiv.org/abs/2610.28554
作者: Runjia Li,Kam Woh Ng,Minghao Chen,Junlin Han,Jianyuan Wang,Zhaochong An,Juan-Manuel Perez-Rua,Andrea Vedaldi,Shikun Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Despite the success of video world models, extending them to multiple agents is difficult because it requires generating several agent-specific video streams that are consistent in time and between agents. To address this gap, we introduce OverSeer, a multi-agent world model that augments egocentric agent views with auxiliary camera views. We suggest that, by fixing the auxiliary views in space, they act as a proxy for a world-centric representation: Unlike the agents’ views, which can change entirely due to a simple change in agent position, changes in the auxiliary views directly map to localized changes in the world state. We show that these fixed views substantially help the model to maintain consistency. We also show that, due to their redundant and semi-static nature, the auxiliary views can be efficiently compressed with a delta-VAE that encodes only the changes relative to a static reference image, reducing the number of tokens required to process them. Experiments in Minecraft and Doom demonstrate improved world and agent consistency, with qualitative results on real-world multi-camera captures indicating that the approach is not tied to game engines. Our compression analysis further shows that delta-VAE substantially reduces the computational overhead of auxiliary-view generation. We conclude that these fixed views provide a simple and efficient way to improve multi-agent world models.
[CV-15] Predictive Latents for Autoregressive Language Models: Possibility Gauge and Obstruction
链接: https://arxiv.org/abs/2610.28540
作者: Yuzhou Hong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 2 figures, 1 table. Theoretical Paper
Abstract:Autoregressive language models are trained by next-token cross-entropy, yet several recent systems also iterate a hidden state without emitting a token: continuous thoughts, pause steps, and recurrent depth. This note separates three risks on one latent process. The token risk R_\mathrmtok scores a readout. The teacher-forced latent risk R_1 scores a one-step map that sees the true hidden state. The open-loop risk R_H scores the same map after it has been iterated on its own outputs. If the Bayes next-token posterior is a function of a k -dimensional statistic with full-rank covariance, a width- k latent, a linear least-squares transition, and a refit unembedding match the Bayes token predictor. The statistic is identified only up to an invertible gauge: whitening the residual stream does not change the token law once the unembedding is refit. The matching obstructions are sharp. A constant latent drives R_1 to zero while token cross-entropy falls back to the marginal entropy. A perfect token predictor can still hide a null-space walk whose open-loop error grows linearly in the horizon. A small next-latent loss does not certificate next-token prediction, and a small next-token loss does not certificate free latent rollout, unless a rank or injectivity condition is stated with the number.
[CV-16] CoGaussian: Open-Vocabulary Gaussian-based Collaborative 3D Scene Understanding
链接: https://arxiv.org/abs/2610.28530
作者: Om Shivam Verma,Shaya Farahmand,Lingjun Zhao,Lu Gan
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Collaborative perception enables agents to retrieve the scene semantics and geometry beyond the field of view of any individual agent. This broader scene understanding enhances transportation safety and improves efficiency in multi-robot systems. However, existing collaborative semantic occupancy methods rely on dense 3D voxel annotations and are limited to recognizing objects within a predefined set of semantic classes. To overcome these limitations, we present CoGaussian, an open-vocabulary Gaussian-based collaborative semantic occupancy prediction framework. CoGaussian represents a scene as a set of foundation model-aligned sparse 3D Gaussians, which serve as a compact representation for communicating and fusing geometric and semantic information across multiple agents. On the Semantic-OPV2V benchmark, our method outperforms existing open-vocabulary single-agent approaches, highlighting the effectiveness of open-vocabulary collaborative scene understanding.
[CV-17] Copyright Leakage Detection in AI Generated Videos CEC
链接: https://arxiv.org/abs/2610.28524
作者: Shaif Chowdhury,Moshiur Rahman Tonmoy
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted in the ICECER 2026
Abstract:Video generation models can create realistic-looking video footage from images and text prompts. Some of the recent advances have raised concerns about the reproduction of copyrighted visual content. However, copyright leakage detection in generated videos remains largely unexplored. In this paper, we introduce a new benchmark data set for copyright leakage detection, comprising 50 copyrighted concepts, each with 5 images. For each concept, we have 5 videos created using the image+prompt video generation model and the prompt-only video generation model. Following that, we develop an AI toolset based on a contrastive retrieval framework that formulates copyright leakage detection as a video-to-image retrieval problem by matching aggregated video frame embeddings against a database of copyrighted image embeddings. Our method uses a single similarity threshold and retrieves the most similar copyrighted reference as supporting evidence. We conducted experiments showing strong retrieval and detection performance, achieving 99.6% detection accuracy with only a single false negative. These results show the effectiveness of the retrieval-based method for scalable and explainable copyright leakage detection in AI-generated videos. Data can be accessed publicly from this https URL and the source code for this project is publicly available at this https URL. Comments: Accepted in the ICECER 2026 Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.28524 [cs.CV] (or arXiv:2610.28524v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.28524 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-18] VLM-Guided Contact Retargeting and Generative Partner Modeling for Humanoid-Human Interaction
链接: https://arxiv.org/abs/2610.28523
作者: Yuto Shibata,Kotaro Amaya,Kashu Yamazaki,Lalit Jayanti,Yoshimitsu Aoki,Mariko Isogawa,Katerina Fragkiadaki
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: CoRL2026, project page: this https URL
Abstract:Humanoid-human interaction requires robots to operate under severe partial observability: during close physical contact, the human partner is occluded, truncated, or outside the robot’s field of view, causing pose estimation to fail and often necessitating external tracking or privileged human state. At the same time, heavy mutual occlusions introduce errors into the human-human motion-capture data used for learning. We present REACT, a framework for learning deployable humanoid-human interaction policies from human-human demonstrations. REACT treats semantic contact structure as the key supervisory signal connecting demonstration processing, interaction learning, and onboard control. First, a VLM-guided contact modeling and retargeting pipeline identifies semantically meaningful contacts and transfers them onto humanoid-human geometry while preserving interaction structure. Using these annotations, privileged interaction policies learn physically grounded behaviors through interaction-aware contact rewards and are then distilled into deployable egocentric depth policies trained entirely from onboard sensing using photorealistic human depth rendering. To address severe partial observability, we introduce generative partner-state modeling objectives that predict partner motion from occluded egocentric views. We evaluate REACT in simulation and on a real Unitree G1 humanoid across five contact-rich interactions. Semantic contact modeling substantially improves preservation of demonstrated interactions, while generative partner-state modeling improves contact recall on four of five interactions, most under heavy occlusion. On real hardware, REACT performs all five interactions using onboard sensing alone.
[CV-19] Latent Bridge: From Video Latents to 3D Geometry via Feature Alignment UAI
链接: https://arxiv.org/abs/2610.28521
作者: Shuai Zhang,Zhanpeng Ouyang,Zhengqing Chen,Ruohong Mei,Shaoyu Chen,Weiqiang Ren,Hui Sun,Qian Zhang,Wenyu Liu,Xinggang Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL Code: this https URL
Abstract:Video generation and feed-forward 3D reconstruction typically use separate representations, requiring RGB decoding and visual re-encoding to recover geometry from generated latents. We propose Latent Bridge, which connects video generative latent spaces with geometric feature spaces to directly predict 3D geometry from video latents. We develop representation-specific bridges for Wan2.1 and V-RAE, enabling feed-forward prediction of cameras, depth, and 3D point clouds without the intermediate RGB pathway. Each bridge is trained on RGB videos through feature distillation, while the video encoder and geometry model remain frozen. Training requires neither ground-truth geometry nor execution of the downstream geometric aggregator or prediction heads. At inference, the same video latents support both RGB decoding and geometry prediction. The same trained bridge can be reused for image-to-video and text-to-video generation within the same video latent space, without bridge retraining. Experiments on Co3Dv2 and WildRGB-D demonstrate competitive geometric reconstruction quality, efficient end-to-end generation and reconstruction, and faster latent-to-geometry decoding, highlighting latent alignment as a practical and reusable interface between video generation and 3D reconstruction. Our code is available at: this https URL.
[CV-20] ouch-JEPA: Tactile Representation Learning from Hand-Object Video with Limited Tactile Supervision
链接: https://arxiv.org/abs/2610.28519
作者: Minseong Kweon,Qiao Feng,Yiming Huang,Lingjie Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project Page: this https URL
Abstract:Understanding contact and tactile magnitude in hand-object interaction is a key step toward physically intelligent systems, yet learning such representations is limited by the cost and restricted coverage of tactile measurements. We present Touch-JEPA, a framework for learning tactile representations from hand-object videos with limited tactile supervision. We first introduce a self-supervised pretraining objective that learns from video alone without tactile supervision. Using video-derived hand geometry, Touch-JEPA predicts masked visual features at corresponding hand-surface locations from spatial and temporal context, encouraging the representation to capture latent contact dynamics. We then extend this formulation to joint representation learning, where masked prediction is applied to all videos while tactile supervision is available only for a labeled subset, grounding the shared representation in measured contact and magnitude. With self-supervised pretraining alone, simple supervised readouts on frozen Touch-JEPA features outperform supervised baselines in contact discrimination and tactile magnitude estimation across three human-hand benchmarks, and transfer to robotic hands without robot tactile training data. Joint learning with limited tactile supervision further improves contact and magnitude estimation on the human-hand benchmarks. Together, these results show that hand-object video can provide scalable self-supervision for tactile representation learning, while limited tactile measurements provide complementary physical grounding.
[CV-21] SenseAlign: Understanding Vision and Touch through Language at Test Time
链接: https://arxiv.org/abs/2610.28511
作者: Chuyang Ye,Qinting Jiang,Jingyan Jiang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 29 pages, including references and appendices
Abstract:Tactile-visual-language models combine vision and touch through a shared language space to recognize objects and materials. In real-world deployment, changes in lighting, contact quality or sensor hardware can shift visual or tactile inputs and degrade recognition performance. When both modalities contribute to a prediction, an unreliable modality can override correct evidence from the other. Since labeled data are generally unavailable during deployment, test-time adaptation offers a way to recover recognition performance from unlabeled observations. In this setting, we find that comparing accumulated observations with the text embeddings of candidate classes reveals modality reliability differences that confidence can miss. Refining features against competing classes also improves recognition beyond geometric alignment alone. These findings motivate SenseAlign, which uses a common semantic reference for test-time modality weighting and feature correction. Alignment Trust weights each modality according to its support for shared predicted classes. Anchor Realignment uses the same class text embeddings to fit a shared feature correction that separates predicted classes from alternatives. Both modules use the same unlabeled history while keeping all encoders frozen. Extensive experiments across diverse datasets demonstrate strong recognition performance under both natural shifts and controlled corruptions.
[CV-22] Deep Learning-Based Road Lane Quality Assessment from Egocentric Dashcam Video CEC
链接: https://arxiv.org/abs/2610.28508
作者: Shaif Chowdhury,Akhila Yeruvasu,Moshiur Rahman Tonmoy,Ayush Goyal,Dunren Che
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted in the ICECER 2026
Abstract:Reliable lane markings are important for both human drivers and automated driving systems, but their condition is still commonly assessed through manual inspection. Existing autonomous-driving datasets primarily support lane detection and scene understanding and usually lack maintenance-oriented labels for lane-marking visibility, continuity, or degradation. In this study, we present a geo-referenced smartphone dashcam dataset and baseline framework for lane-marking quality assessment. The dataset contains 63,839 road-facing frames collected across South Texas using a GPS-enabled dashcam application. A subset of 3,339 frames is matched with OpenStreetMap-derived lane-count labels, which are used as weak supervision for lane-count regression and map-consistency evaluation. We propose three lane-quality measures: Contrast Stability Index (CSI), which captures lane-background contrast using Laplacian responses; Line Continuity Ratio (LCR), which estimates structural continuity using edge and Hough-line evidence; and Regression Lane Confidence Score (RLCS), which measures agreement between predicted and map-derived lane counts on matched frames. The baseline prediction models attained strong performance, with CSI (MAE 0.039 , R^2=0.97 ) and LCR (MAE 0.093 , R^2=0.84 ). Notably, CSI and LCR are no-reference, segmentation-derived quality metrics that can be computed without dense human labels. In a pilot validation using 50 human-ranked images, these proposed lane-specific metrics show stronger rank agreement with human lane-quality ordering (e.g., Spearman’s \rho=0.7316 and Kendall’s \tau=0.5472 for LCR) than general image-quality measures such as PSNR, SSIM, BRISQUE, NIQE, and LPIPS. We believe the dataset, metrics, and baseline provide an initial step toward scalable, low-cost, location-aware lane-marking quality assessment. Data and code: this https URL
[CV-23] GazeLens for Efficient Egocentric Video Question Answering with Simple Gaze-Driven Adaptive Token Reduction
链接: https://arxiv.org/abs/2610.28502
作者: Trong-Thang Pham,Anh Nguyen,Hien Van Nguyen,Ngan Le
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 17 pages, 6 figures
Abstract:Smart glasses can help users understand their activities and surroundings, but small vision-language models must preserve useful evidence within limited computation. This is difficult because redundant frames consume the visual budget, while aggressive spatial compression can discard relevant details. Although gaze identifies the wearer’s focus, the question may require evidence from the surrounding scene. To address these constraints, we propose GazeLens, which combines frame selection with gaze- and question-guided token reduction. First, HOG transitions and image sharpness select frames before visual encoding. For each selected frame, Gaussian pooling then forms gaze-focused and complementary summaries. Because their relevance depends on the question, a lightweight gate blends them into a primary token. Together with a global scene token, this retains both focused and contextual summaries within two tokens per frame. On StreamGaze, GazeLens achieves higher macro and micro accuracy than the evaluated full-grid and token-reduction baselines, with lower input-forward FLOPs than the shared-backbone token-reduction baselines.
[CV-24] Learning from Automatic 3D Gaussian Splatting Edits for Online Bayesian Scene Change Detection
链接: https://arxiv.org/abs/2610.28490
作者: Changrui Zhu,Simon Julier
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: 34 pages, 10 figures
Abstract:Online scene change detection requires integrating incoming observations to determine which regions have changed relative to a reference scene. We update change beliefs as new images arrive using generalised Bayesian evidence fusion. However, the prior probability of change and evidence models are unknown beforehand, and numerical Bayesian inference can be costly. We address these challenges by learning offline from automatic edits to a reference 3D Gaussian Splatting (3DGS) model. These edits generate labelled training views without manual annotations. Together with unchanged reference observations, they provide supervision for learning scene-specific evidence models and estimating the effective pixel prior and response statistics. Online prediction combines current evidence with historical support, while closed-form local Bayes factors, parameterized by the learned response statistics, update an accumulated 3D Gaussian field. Object-guided refinement produces the final change masks. Our method achieves state-of-the-art accuracy on PASLCD and CCS3D, with 67.64% and 75.82% mIoU, respectively. On PASLCD, the complete pipeline runs at 17.2 ,FPS over complete query sequences, the fastest among the evaluated online methods.
[CV-25] CLOSER: Evidence-Aware Active Visual Verification for Embodied Object-Goal Navigation
链接: https://arxiv.org/abs/2610.28476
作者: Pengying Wu,Mingye Jiang,Yunze Hu,Zhejia Wen,Kangjie Zhou,Zongru Yang,Zhe Peng,Chang Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 13 pages, 8 figures, 10 tables
Abstract:Object-goal navigation (ObjectNav) requires an embodied agent to find an instance of a specified category in an unseen environment and stop near it. The agent typically relies on an object detector but must judge online from a limited egocentric video stream, sometimes a single frame, whether a detected candidate is the target. Detectors often confidently mistake similar objects, such as a sofa for a bed, and stopping at a false positive fails the episode. Prevailing systems also approach candidates greedily, leaving them occluded or partially observed. We present CLOSER, an evidence-aware active verification framework that treats stopping as closed-loop evidence acquisition, coupling vision-language model (VLM) verification with active perception. CLOSER fine-tunes an open-source VLM to label each view of a candidate tracked across frames as target, non-target, or uncertain, selects complementary views with an embedding-based diversity gate, and fuses their per-view probabilities with a lightweight network. If the fused judgment remains uncertain, a diffusion policy drives the agent toward the candidate using only the tracked mask and depth until the decision becomes decisive. CLOSER runs locally and plugs into existing ObjectNav systems without retraining them. On Gibson and Habitat-Matterport 3D, it raises average success rate by 3.6 points and success weighted by path length by 2.0 points, showing that verifying evidence before committing improves embodied navigation.
[CV-26] Answering in the Users Variety: Calibrated Accommodation Pairs for Post-Training
链接: https://arxiv.org/abs/2610.28474
作者: Zhizhuo Kou,Heming Li,Jian Yang,Junyu Luo,Jingshu Peng,Sirui Han,Yike Guo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:People adapt how they talk to the person in front of them. However, chat models tend to respond in Standard English regardless of the dialect used by the user. A uniform answer can hide whether training treats these users alike, and output audits cannot tell as well. We trace how each post-training stage, from supervised fine-tuning (SFT) to preference optimization (DPO) and reinforcement learning, changes a model’s tendency to match the user, and read what each update prefers from two content-matched replies, one matching the user and one not. Across 10 open model lineages, post-training standardizes replies to every informal cue, yet after adjusting for how likely each reply already was, the DPO update still favours matching British and Australian spelling over AAE and Singlish on 10 lineages, and models still produce AAE features when asked. We then introduce calibrated accommodation pairs (CAP), preference data that reward matching users’ dialect density while preserving Standard English for Standard-English users. On a fully open model, CAP raises convergence on features it was not trained on without a detectable increase in dialect for Standard-English users or a cost in instruction following. In contrast, one-sided pairs reach a similar gain only by overshooting the user and imposing dialect. Our findings show that uniform replies can hide selective updates and audits should read the update as well as the output. Code: this https URL.
[CV-27] LocaLeJEPA: Maximum Entropy Representations under Geometric Priors
链接: https://arxiv.org/abs/2610.28458
作者: Kartik Tandon,Alessio Borgi,Pavel Snopov,Inés García-Redondo,Davide D’Ascenzo,Sebastiano Cultrera di Montesano,Fabrizio Silvestri,Pietro Liò,Francesca Dominici,Randall Balestriero,Claudio Battiloro
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Joint-Embedding Predictive Architectures (JEPA) offer an appealing approach to self-supervised learning, but are typically prone to representation collapse. LeJEPA resolves this by introducing a theoretically grounded distribution-matching objective that drives the embedding towards an isotropic Gaussian. However, this approach is fundamentally tied to Euclidean assumptions, and its global target does not reflect domain-specific geometric priors, such as the use of non-Euclidean embedding geometries, e.g., equivariant CNNs or hyperbolic ViTs. We thus introduce LocaLeJEPA, a framework motivated by a manifold-theoretic reformulation of LeJEPA’s entropy-maximization property. The accompanying LocaLeReg, a novel locally Euclidean distribution-matching regularizer, enables a stable and scalable implementation of the LocaLeJEPA framework. We then perform extensive experiments in the vision setting on models trained on ImageNet-1K. LocaLeJEPA significantly outperforms LeJEPA when paired with hyperbolic ViT-S/8 and equivariant ResNet-50 backbones, and also with ordinary Euclidean ViT-S/8 and ViT-L/14 backbones. This holds both in terms of linear and k-NN classification performance on ImageNet and on a series of downstream tasks tested. We view the LocaLeJEPA formalism as a principled approach to the JEPA paradigm that better respects intrinsic data geometry while also accommodating generalizations to exotic embedding geometries. Code available at this https URL.
[CV-28] EvaOmni: An Evidence-Grounded Benchmark for Audio-Visual Reasoning
链接: https://arxiv.org/abs/2610.28444
作者: Xiaoxing Hu,Wenxi Yao,Chaoyuan Shi,Ziqi Ye,Xiaorong Guo,Peixuan Li,Zhuoran Sun,Ziyang Gong,Peiyuan Zhang,Mingxin Liu,Shuran Ma,Hongbo Liu,Xiaomin Yu,Yu Tian,Wentao Wang,Zhenjie Yang,Jian Hu,Yunhang Shen,Haoyu Cao,Xue Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Under Review
Abstract:Omni LLMs can naturally locate key audio and visual evidence and integrate information across modalities for audiovisual reasoning. Existing audio-visual reasoning benchmarks that mainly rely on multiple-choice questions usually score models solely based on answer correctness, making it difficult to assess whether a model truly forms a reasoning chain from evidence to the correct answer. In addition, textual or unimodal shortcuts may allow models to answer correctly without analyzing the necessary audio-visual evidence, further weakening the reliability of evaluation benchmarks. To address these issues, we introduce EvaOmni, an evidence-grounded benchmark containing 392 videos and 1,027 purely human-annotated question-answer pairs, covering 15 capabilities and three levels: retrieval and localization, basic understanding, and complex reasoning. We design a series of rigorous annotation procedures in which experts annotate audio and visual evidence and construct video-based distractors; humans review whether questions meet strict inspection criteria, and finally models filter textual and unimodal shortcuts. Beyond high-quality question-answer pairs, we require models to simultaneously return the answer and supporting evidence intervals in order to evaluate the complete reasoning chain. We propose the Coverage Grounded Score (C.G.) to measure how much annotated evidence a model identifies, and the Bottleneck Grounded Score (B.G.) to penalize poor temporal localization in either modality; both are incorporated into answer correctness. Across 7 open-weight and 17 proprietary Omni LLMs, Gemini3.7-Flash achieved the highest overall accuracy (79.4%), C.G. (70.8%), and B.G. (35.1%). At Level 3, it reached 79.1% accuracy, but its B.G. was only 12.3%. These results show that even the most advanced models still have room for development in grounded audio-visual reasoning.
[CV-29] WorldSR: Harnessing World Knowledge Search for Grounded Image Super-Resolution NEURIPS2026
链接: https://arxiv.org/abs/2610.28423
作者: Junyuan Deng,Xinyi Wu,Zhenyao Wu,Song Wang,Yu Wang,Congchao Zhu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026. Project Page: this https URL
Abstract:Generative-prior-based super-resolution (SR) methods leverage implicit knowledge learned from large-scale models to enhance low-resolution inputs. However, such priors are inherently stochastic and optimized for diverse generation, conflicting with the deterministic, high-fidelity reconstruction required for SR. It is also tied to the base model, costly to update, and prone to hallucinations as real-world data evolves. To address these limitations, we introduce WorldSR, a novel framework that augments SR with explicit, entity-aligned visual world knowledge. WorldSR automatically decomposes the input into semantic entities and performs entity-wise retrieval of structurally and semantically aligned images from a large-scale external corpus. For efficiency and reliability, we incorporate a memory mechanism to reuse previously retrieved results and a multi-step filtering process to refine candidate references. The selected references as world knowledge are then integrated through a novel multi-reference super-resolution model, which provides fine-grained constraints for detail synthesis and structural recovery. Experiments on real-world SR benchmarks show that WorldSR consistently improves structural fidelity and visual quality, outperforming existing generative super-resolution methods. Our results suggest that entity-level grounding in external visual world knowledge provides a simple and effective complement to model-bound generative priors.
[CV-30] DISTRACT: Degradation-Indexed Study of Textual Reference Across Competing Targets
链接: https://arxiv.org/abs/2610.28419
作者: Dan Zlotnikov,Ohad Ben-Shahar
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Vision-language trackers (VLTs) accept descriptions intended to assist tracking, but accepting text does not establish that its content changes tracking decisions, particularly as visual evidence deteriorates. We introduce DISTRACT, a matched-intervention benchmark that holds video, target, and distractor fixed while crossing four text priors with four image degradations along a controlled severity axis. We evaluate five VLTs on four datasets using behavioral contrasts, layerwise representation probes, and activation patching. Molmo2 and JointNLT exhibit directed text use. Some effects contract under degradation, whereas others show no significant endpoint change after multiplicity correction. UVLTrack, MMTrack, and DUTrack respond only weakly to changes in text semantics, with no directed behavioral effect surviving FDR correction in the tested settings. Model-internal analyses reveal that at decision-near states, correct-versus-conflicting text is decodable in Molmo2 and JointNLT but not in the other three systems. In the responsive models, decodability peaks at late states where replacement produces little behavioral change, whereas replacing selected earlier states redirects conflict-associated trajectories toward the target. Because matched-correct donors do not significantly outperform unrelated donors, these interventions identify earlier intervention-sensitive states without establishing target-specific semantic rescue. DISTRACT thus exposes when and where text affects tracking under degraded visual evidence.
[CV-31] Code2Video-Compass: A Comprehensive Benchmark for General-Purpose Code-Based Video Generation across T2V I2V and V2V
链接: https://arxiv.org/abs/2610.28416
作者: Zhenghan Chen,Zekai Shao,Xin Lin,Leting Cheng,Siming Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Code-based video generation offers fine-grained control, targeted editing, and component reuse compared with direct video generation. This approach requires multimodal large language models (MLLMs) to translate visual and temporal requirements into executable programs that produce complete videos. However, existing code-based benchmarks primarily target specialized tasks, leaving complete video creation across diverse creative categories and input modalities insufficiently evaluated. We introduce Code2Video-Compass, a general-purpose benchmark for code-based video generation across text-to-video (T2V), image-to-video (I2V), and video-to-video (V2V) settings. Code2Video-Compass comprises 2,104 tasks spanning 10 categories, 24 subtypes, and three complexity dimensions covering construction, dynamic animation, and state dependency. It combines tasks curated from textual instructions and visual references with corresponding T2V, I2V, and V2V tasks derived from complete professional and user-created videos. We propose a dual-level evaluation framework combining 12 objective metrics and five subjective metrics. Objective evaluation covers operational performance, video quality, text-video alignment, and reference fidelity, while subjective evaluation uses an MLLM judge to evaluate instruction following and holistic video quality. Extensive evaluation of 15 recent proprietary and open-weight MLLMs from nine families reveals that even leading models face challenges in reliably generating high-quality complete videos aligned with textual and visual requirements, with performance generally declining as construction, dynamic animation, and state-dependency complexity increases. We anticipate that Code2Video-Compass will advance reliable, coherent, and controllable code-based video generation and inspire broader progress in programmatic visual creation.
[CV-32] Brain2Vec: A Self-Supervised 3D Foundation Model for Structural Brain MRI
链接: https://arxiv.org/abs/2610.28411
作者: Spiros Maggioros,Guray Erus,Gareth Harman,George Aidinis,Alexander Getka,Aristeidis Sotiras,Christos Davatzikos
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Deep learning models have become essential tools in medical imaging, yet training them end-to-end for diverse tasks remains challenging due to limited labeled data and computational constraints. To address this, we introduce Brain2Vec, a foundation model pre-trained on 74,425 T1-weighted Magnetic Resonance Imaging (MRI) scans pooled from multiple studies. Brain2Vec can serve both as a weight initialization backbone for fine-tuning downstream tasks and as a standalone feature extractor. We evaluated Brain2Vec in three tasks: downstream predictive performance, convergence speed, and embedding quality. Using Brain2Vec’s pretrained weights as initialization and fine-tuning on labeled data for a brain age prediction task, results demonstrate improved final accuracy and faster convergence compared to random initialization. Additionally, its frozen embeddings outperform anatomically defined volumetric features on four of five downstream classification tasks and perform competitively with a state-of-the-art foundation model pre-trained on substantially more data. Lastly, we show that Brain2Vec encodes information primarily from a spatially localized and neuroanatomically meaningful set of brain regions, and that manipulating a subset of latent features produces monotonic, region-specific effects on the reconstructed anatomy. Brain2Vec code and trained models are available as an open-source package on this https URL. Users can apply Brain2Vec locally by installing the package, or directly through the NiChart cloud platform this https URL, where Brain2Vec features can be derived without requiring local installation or specialized infrastructure.
[CV-33] A Multi-Agent Framework for Automated Machine Learning with Evidence-based Planning for Medical Image Analysis
链接: https://arxiv.org/abs/2610.28410
作者: Yiqing Shen,Mathias Unberath
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Deep learning has become the default tool for medical image analysis, driven by the open-source movement. However, achieving the best performance on a given task remains difficult, with architecture selection and hyperparameter optimization (HPO) still considered “more of an art than a science”. Automated machine learning (AutoML) automates parts of model training, yet helps little in medical imaging: general-purpose approaches lack medical architectures and preprocessing, and medical-focused ones demand engineering expertise and remain confined to narrow tasks such as 3D segmentation. We introduce MedicalMLOmni, a multi-agent AutoML framework that automates model training for diverse single- and multi-modal medical tasks through a natural language interface. Specifically, MedicalMLOmni performs evidence-based planning by retrieving published results to guide model selection and configuration, and employs parallel HPO where multiple agents explore adjustment strategies concurrently. MedicalMLOmni achieves the best score on 19 of 20 benchmarks; PMC-VQA is the exception, where a vision-language model with broader pre-training leads. Backbone-matched controls attribute the gains to the framework, not the backbone. By removing coding requirements, MedicalMLOmni lowers the barrier for domain experts with rich datasets but limited engineering expertise.
[CV-34] StreamEval: Evaluating Interactive Streaming Video Generation Models
链接: https://arxiv.org/abs/2610.28402
作者: StreamEval Team
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:As streaming video generation advances, enabling real-time interaction with generated video is becoming increasingly important. We introduce StreamEval, a unified benchmark for interactive streaming video generation spanning five tasks: avatar dialogue, avatar performance, world generation, instruction editing, and reference editing. StreamEval comprises 300 scripted test cases lasting 15-300s, each specifying initial assets, incrementally arriving multimodal requests, and explicit requirements for state changes and content preservation. Performance is assessed along four evaluation dimensions: task execution, state consistency, visual fidelity, and audiovisual coordination, instantiated through ten task-specific metrics on a 0-100 scale. Evaluating 37 distinct models across 300 test cases, we find that high visual quality and identity retention do not necessarily coincide with reliable request execution or sustained state consistency. In particular, request execution remains challenging: the highest task-specific scores are 36.90 for Following among seven avatar-performance models, 25.93 for Following in text-conditioned world generation, and 15.00 for Edit Success among three reported reference-editing models. These findings identify reliable request execution under timing and preservation constraints as a central challenge for advancing omni-modal real-time interactive systems. The leaderboard is available at this https URL
[CV-35] VisualEvolve: A Skill Self-Evolution Framework for Visual Reasoning
链接: https://arxiv.org/abs/2610.28386
作者: Jiawei Wang,Yucun Li,Xueqi Li,Shenghao Wang,Xinquan Gong,Dan Zhang,Zheqi Lv,Yuquan Le,Shaozu Yuan,An Zhang,Da Cao,Tat-Seng Chua
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 29 pages, 9 figures. Project page: this https URL
Abstract:Visual tool use enables vision-language models (VLMs) to acquire task-relevant evidence, but its effectiveness can be limited by predefined visual operations and the tool-use strategies learned during training. Representing these operations and strategies as reusable external skills allows them to be refined through self-evolution without updating model parameters. However, skill refinement can introduce performance regressions. Meanwhile, inspecting accumulated task execution trajectories together with skill evolution history may lead to evolution evidence overload. To address these challenges, we propose VisualEvolve, a self-evolving framework that evolves executable visual operations and tool-use strategies for VLMs through a Select–Execute–Evolve–Evaluate cycle. Skill-Library Snapshot Selection uses performance on a fixed validation set disjoint from execution tasks to select an effective historical snapshot for subsequent evolution, limiting the propagation of performance regressions. Progressive Evidence Access (PEA) guides capability-gap diagnosis through a coarse-to-fine hierarchy comprising the Skill-Performance Report, Evolution Log, Trajectory Summaries, and relevant raw trajectory segments. Across four VLM backbones and six evaluation sets from three benchmarks, VisualEvolve improves over the Initial Skill Library in 23 of 24 settings, with a maximum gain of 15.1 percentage points on VisualProbe Hard using Qwen3.5-Plus. Ablation studies support the contributions of Snapshot Selection and PEA, while further experiments show that skills can be improved using experience from different data settings and directly reused across backbones. Moreover, VisualEvolve enables effective skill self-evolution with over an order of magnitude fewer task-solving trajectories than estimated for representative RL-based methods.
[CV-36] RACE: Learning Multimodal Search through Referential Abstraction and Compositional Evidence
链接: https://arxiv.org/abs/2610.28382
作者: Eric Tao,Minhao Sun,Liyan Liu,Zhijing Wu,Xiaohan Yu,Lang Mei,Li Huang,Chao Feng,Jinchao Ma,Sichen Kang,Siyu Mo,Chong Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Training multimodal search agents requires tasks that make visually grounded search necessary for determining the answer. Textual identity clues and unused target evidence allow shortcuts in synthetic questions, weakening supervision for visual search and evidence integration. We propose RACE, an image-conditioned task synthesis framework that constructs dependencies from images to targets and from target-specific evidence to answers. Referential Abstraction (RA) replaces identity-revealing text with specific visual references or ambiguous relational references resolved by searching among depicted candidates. Compositional Evidence (CE) links retrieved facts about multiple entities in one image or entities from different images into joint questions whose answers require every designated contribution. Both operations also apply to images found during search, allowing new visual observations to determine subsequent search targets or supply answer evidence. Fine-tuning Qwen3.5-27B on only 12K teacher-generated trajectories from RACE raises average accuracy across ten multimodal benchmarks from 53.5 to 56.0, a relative improvement of approximately 4.7% over the matched base agent. Our code, data, and model will be publicly available.
[CV-37] CountVideoBench: A Benchmark to Evaluate Prediction Biases on the Object Counting task by Video-Language Models BMVC2026
链接: https://arxiv.org/abs/2610.28356
作者: Marc Chapus,Carlos Crispim-Junior,Véronique Eglin,Atilla Baskurt
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at BMVC 2026. Project page: this https URL
Abstract:Object counting in videos remains a poorly characterized capability of multimodal large language models. Counting questions are included in many existing VideoQA benchmarks, but the ground-truth count distributions in these benchmarks are heavily skewed toward small values. Across the nine benchmarks we survey, the median ground-truth count is at most five, and the three most frequent values together account for between 45% and 93% of the videos. The resulting mean accuracy is then dominated by the most frequent counts; a value-dependent bias at a rare count contributes negligibly to the average and remains invisible. We introduce CountVideoBench, a benchmark of 2,600 synthetic videos covering 26 cardinalities in the interval [5, 50] (100 videos each), grouped into six clusters of consecutive counts that span the range while enabling near-neighbor comparisons. Each video contains a single class of static objects under a controlled camera motion, removing object motion and reducing recognition as confounds. We evaluate sixteen video-language models from nine architecture families in a zero-shot setting. No model exceeds 27.6% Exact Match. The dependence of accuracy on cardinality varies across models, ranging from monotonic decline to a non-monotonic peak-and-collapse pattern, which rules out increasing difficulty with cardinality as the dominant explanation of counting failure. Every model further concentrates its predictions on its own set of preferred integers, with accuracy peaks coinciding with these values and collapsing at adjacent counts. These findings highlight the need for balanced, diagnostic counting benchmarks to reliably assess this fundamental visual capability.
[CV-38] S-Agent 2.0: Spatial Reasoning with a General Coding Agent
链接: https://arxiv.org/abs/2610.28354
作者: Rui Gong,Ruoyu Wang,Yalun Dai,Hao Li,Sidi Yang,Qihao Zhao,Hongyu Li,Zeyang Guan,Fangzhou Hong,Zhaoxi Chen,Kim-Hui Yap,Ziwei Liu
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Spatial reasoning asks a model to work out where objects are in 3D space, how they relate to one another, and how they move. Agentic VLMs have recently brought large improvements on such tasks by letting the model actively gather evidence with 3D perception tools instead of answering in a single pass. Even so, on many questions the tools return the right evidence and the model still does not arrive at the correct answer: the gap between model and tools, where raw readings become usable conclusions, is left unfilled. We present S-Agent 2.0, a spatial reasoning agent that equips an off-the-shelf general coding agent with a command-line perception toolkit, and we design the interface around one principle: help the VLM close the evidence loop: acquire scene evidence, remember it in a persistent scene memory, and interpret each reading in plain spatial language. This design keeps the full flexibility of a coding agent, and every reading it acts on stays grounded in the tools’ geometry. On MMSI-Bench and ReVSI, the toolkit improves a strong open-source model, Qwen3.8-27B, by +13.9 and +20.5 points over the same agent without tools, and a strong closed-source model, Gemini-3.8-Flash, by +11.2 and +9.6 points. With the open backbone S-Agent 2.0 reaches 66.5 on MMSI-Bench and 82.3 on ReVSI, ahead of the specialized system SpatialClaw on the same backbone (62.6 and 74.5), and with Gemini-3.8-Flash it reaches 74.1 and 90.5 against its 67.0 and 79.7. Both tool gains shrink as the backbone grows more capable, but SpatialClaw’s shrinks far more on ReVSI, from +13.2 to +2.7, while ours stays at +9.6. S-Agent 2.0 also works across different general coding agent harnesses, such as Codex and Claude Code.
[CV-39] RayCo4D: Ray-Conditioned Continuous Motion Model for Monocular 4D Dynamic Reconstruction
链接: https://arxiv.org/abs/2610.28351
作者: Yuanxiang Ni,Xianliang Huang,Ruihan Hou,Feng Xie,Hao Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 25 pages, 13 figures
Abstract:Reconstructing dynamic 3D scenes from monocular video requires inferring motion from noisy and intermittently visible observations. Long-range tracks provide valuable geometric constraints, but integrating their fragmented temporal coverage into shared, continuous motion remains challenging. We introduce \textbfRayCo4D, a ray-conditioned continuous motion model for monocular 4D reconstruction that explicitly separates visual measurements from latent 3D motion states. Each track is associated with a reference-time 3D query constrained to its camera ray, and its image and depth observations constrain shared latent motion nodes through local many-to-many associations. Each node maintains an \mathrmSE(3) pose and body-frame velocity at sparse support times. A white-noise-on-acceleration process prior couples neighboring states and defines continuous-time interpolation, allowing observations across visibility intervals to jointly constrain the same latent state this http URL inferred shared motion is transferred to Gaussian primitives through dual-quaternion blending, while reference-anchored local \mathrmSE(3) residuals capture fine-scale non-rigid deformation beyond the shared scaffold. Experiments on diverse monocular dynamic scenes show that RayCo4D achieves competitive novel-view synthesis quality, improves structural and temporal consistency under rapid motion and incomplete tracks, and supports continuous-time rendering.
[CV-40] Gated Prototype Transfer for Zero-Support Medical Vision Language Model ALT
链接: https://arxiv.org/abs/2610.28350
作者: Sumaiya Fazal,Rizwan Qureshi,Laleh Behjat,Samira Ebrahimi Kahou
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to Symposium on Model Accountability, Sustainability and Healthcare 2026
Abstract:Realistic few-shot adaptation of medical vision-language models (VLMs) can leave low-prevalence classes without any labeled support. The training-free SS- Text+ solver remains operational in this setting, but its zero-support prototypes receive no visual adaptation and fall back to a purely textual class prototype. We in- troduce Gated Knowledge-Guided Prototype Transfer (GProtoT), a pre-repulsion module that transfers supported-class visual residuals to missing classes using VLM similarity and an ontology-based clinical distance. A validation-free evidence gate scales transfer based on semantic relevance and residual agreement, thereby limiting unsupported extrapolation. In 50 paired realistic one-shot trials on MESSIDOR with FLAIR features, severe and proliferative diabetic retinopathy were absent from every support set. Relative to SS-Text+, gated ProtoT improved missing-class recall by 5.23 points, average class accuracy by 1.37 points, and the missing/supported recall harmonic mean by 1.83 points. It reduced negative- transfer frequency from 46-48% for ungated transfer to 2%, while at the cost of reductions of 1.20 percentage points in supported-class recall and 0.45 percentage points in top-1 accuracy. These results indicate that evidence-based gating provides a safety mechanism for the knowledge source. GProtoT is based on SS- Text+ and requires no gradients, validation data, or additional model inference, of- fering a low-compute path toward accountable adaptation under incomplete clini- cal supervision. Code is available: this https URL
[CV-41] Auto-Director: Leverag ing Multi-Agent Collaboration for Autonomous Video Editing
链接: https://arxiv.org/abs/2610.28346
作者: Yuxia Chen,Wenbo Zhu,Yangguang Ji,Zhenxiang Jiang,Teng Zhong,Yu Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at MMM 2027
Abstract:Autonomous video editing from open-world resources remains a challenging problem in multimedia content creation. Retrieval-based methods are constrained by fixed source material and lack creative flexibility, while generative models struggle with long-horizon coherence and often diverge from user intent. Both paradigms fail to replicate the structured, multi-stage reasoning that underlies professional post-production. To address these limitations, we present Auto-Director, a tool-augmented multi-agent framework that mirrors professional editorial workflows: a Planning Agent decomposes abstract topics into fact-grounded scripts via retrieval-augmented planning; a Material Agent indexes open-world footage through shot segmentation and visual tagging; an Editor Agent assembles semantically aligned clips with MLLM-driven spatial adaptation; and a Critic Agent governs an iterative Reflect-and-Revise loop that checks scene-level alignment and pacing. To rigorously evaluate this new paradigm, we introduce VEdit-Bench, a benchmark of 100 expert-curated topics with human-annotated Golden Scripts and a five-dimensional metric suite covering Narrative Logic, Factual Accuracy, Semantic Alignment, Compositional Quality, and Stylistic Cohesion. Extensive experiments demonstrate that Auto-Director significantly outperforms baseline methods in both automated metrics and human evaluation, effectively bridging the gap between amateurs and professional-grade video production.
[CV-42] HistoLoop: Iterative Contextual Inference with a Looped Transformer for Histology-Based Spatial Gene Expression Prediction
链接: https://arxiv.org/abs/2610.28339
作者: Ziang Liu,Jie Liu,Shuai Li,Yigui Feng,Qinglin Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Predicting spatial gene expression from HE images offers a scalable complement to experimental spatial transcriptomics. However, despite progress in spatial context modeling, how to construct subsequent context from spot states updated by earlier contextual interactions remains underexplored. We introduce HistoLoop, a looped Transformer that formulates prediction as iterative contextual inference. At each refinement step, HistoLoop aggregates the current spot states into context tokens at the corresponding resolution, processes them with a shared spatial Transformer, and retrieves spot-specific contextual information through cross-attention. A memory-mediated update integrates this information into persistent spot states, which are then used to construct context at the next, finer resolution. Under the HEST-Bench HVG50 evaluation protocol, HistoLoop achieves the highest macro-averaged gene-wise Pearson correlation among the compared methods across six cancer-specific tasks spanning five organs in HEST-1K. Ablation studies support the importance of recurrent state refinement and reconstructing context from updated spot states at progressively finer resolutions. Expression-derived clustering further suggests that HistoLoop better captures spatially organized multigene patterns.
[CV-43] PACE: Plan Attend Check and Emit for Training-Free Response Timing in Streaming Video Understanding
链接: https://arxiv.org/abs/2610.28323
作者: Siqi Guo,Yuxia Chen,Wenbo Zhu,Yangguang Ji,Zhenxiang Jiang,Teng Zhong,Yu Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at MMM 2027
Abstract:Streaming video understanding requires a model to decide not only what to answer but also when to answer. Evidence arrives progressively, and recent benchmarks identify premature responses, issued before the queried event has concluded, as a recurring failure mode. Existing approaches often learn response timing through additional training, while training-free alternatives typically rely on scalar gating signals that do not explicitly represent when an answer becomes final. We organize chunk-wise streaming errors into three patterns: sampling variation can make an answer unstable, a stable answer may lack support in the observed video, or an answer may be supported by the current evidence while the event remains unfinished. We propose PACE (Plan, Attend, Check, and Emit), a training-free framework designed around these three failure modes. Given a question, PACE drafts a plan specifying what to observe and what will signal completion. As the stream unfolds, it invokes an off-the-shelf video LLM at informative moments, evaluates each draft for stability, grounding in its cited frames, and completion, and records verified observations in an event ledger. Experiments on ProactiveVideoQA and OVO-Bench Forward show that PACE improves over the official protocols by up to 5.9 average PAUC points and 6.2 average accuracy points across proprietary and open-source backbones, yielding state-of-the-art performance.
[CV-44] ABACUS: A Benchmark for Image Editing Models to Assess Object Category and Counting Under Scene Conservation WACV2027
链接: https://arxiv.org/abs/2610.28316
作者: Aymen Lassoued,Angelique Loesch,Romaric Audigier
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at WACV 2027 (Evaluations Datasets track). Code and data: this https URL
Abstract:Instruction-based image editing models can now add, remove a specified number of objects or change their category while preserving the scene. Yet these models often fail to produce the correct number, and these failures grow sharply as scenes become more complex and the requested count increases. Current benchmarks largely miss this problem: they score edits with holistic similarity metrics or with vision-language model (VLM) judges that are themselves poor at counting, so numerical errors go unnoticed or are graded incorrectly. As a result, the true numerical ability of today’s editing models has never been accurately measured. We address this issue with ABACUS, a new benchmark, evaluation protocol and dedicated scoring metric for numerical image editing. Our protocol is judge-free: no VLM or MLLM scores an edit or counts objects (a VLM only confirms the category of rare ambiguous crops). Instead, it verifies each edit at the instance level against panoptic ground truth confirmed by detector consensus, checking both that the exact requested count was achieved and that the rest of the scene was preserved, across the add, remove and change operations. By evaluating leading image editors, our method reveals substantial numerical editing failures that worsen with increasing scene complexity and object count; failures that prior benchmarks often overlooked or misjudged.
[CV-45] owards Consistent Geospatial Representations: Aligning One Encoder to Many Foundation Models NEURIPS
链接: https://arxiv.org/abs/2610.28313
作者: Arthur Ouaknine,David Rolnick
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS REO Workshop 2026
Abstract:Geospatial foundation models (GFMs) differ widely in pretraining objective, modality and resolution, and none leads across benchmarks. We ask whether their representations are complementary, and whether a single encoder can inherit that complementarity. We introduce Pangaea-Latents, an extension of the Pangaea benchmark providing frozen embeddings of 5 GFMs over its 11 datasets, along with reproduced baselines, and find GFM embeddings to be neither redundant nor disjoint. We then propose GeoMixture, a preliminary framework to align a ViT to several GFMs per dataset through a multi-representation alignment (mREPA) objective. Our encoder ties the best GFM in average rank across nine Pangaea datasets, trading peak performance for cross-task consistency.
[CV-46] AGGR-Mamba: Mamba-based Network with Attention-Guided Gated Residual Refinement for Undersampled MRI Reconstruction
链接: https://arxiv.org/abs/2610.28291
作者: Yue Zhuo,Wenbin Deng,Lingjun Liu,Xuhang Chen,Zhonghua Xie
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by IEEE BIBM 2026
Abstract:Undersampled magnetic resonance imaging (MRI) reconstruction aims to recover high-quality images from partially acquired k-space data, thereby reducing scan time. However, missing frequency-domain measurements often lead to residual aliasing artifacts, blurred edges, and loss of local structural details. In this paper, we propose AGGR-Mamba, a lightweight output-stage residual refinement method built upon a Mamba backbone for undersampled brain MRI reconstruction. The Mamba backbone first produces an initial reconstruction, and the proposed refinement modules then correct the remaining local errors and high-frequency artifacts in a controlled residual manner. Specifically, an Attention Refine Head is introduced to recalibrate local anatomical structures, edge textures, and residual-error regions through channel and spatial attention. In addition, a GDFN-style gated residual refinement branch is designed to selectively preserve informative residual features while suppressing redundant or unstable responses. By combining attention-based recalibration with gated residual selection, AGGR-Mamba improves local structure recovery without substantially disturbing the backbone reconstruction. Experiments are conducted on the IXI brain MRI dataset with T1- and T2-weighted images under acceleration factors of R=4 and R=8. The results show that AGGR-Mamba achieves competitive PSNR and SSIM performance across different reconstruction settings, demonstrating the effectiveness of the proposed output-stage refinement strategy.
[CV-47] CPVB-RG: Cross-Prior Visual Bridging for Prior-Aware Chest X-Ray Report Generation WACV2027
链接: https://arxiv.org/abs/2610.28289
作者: Sike Xiang,Shuang Chen,Julie Cox,Amir Atapour-Abarghouei
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at WACV 2027
Abstract:Radiologists interpret follow-up chest X-rays by comparing current and prior examinations to identify new, resolved or progressing findings. Most report generation systems either omit this longitudinal context or fuse prior features without distinguishing interval changes from anatomy that remains stable over time. We propose CPVB-RG, a prior-aware framework that combines current and prior radiographs with available clinical text. Its Cross-Prior Visual Bridge (CPVB) contains Progression-Guided Focus Attention (PGFA) and two Cross-Attention units. PGFA is designed to contrast progression-difference attention with shared-spatial attention, aiming to reduce invariant responses whilst retaining coarse spatial correspondence between examinations. The Cross-Attention units then connect the resulting longitudinal visual evidence with clinical semantics. CPVB-RG is trained in two stages, using prior-guided image-text alignment followed by autoregressive report generation with a lightweight language decoder. Experiments on MIMIC-CXR and IU X-Ray show consistent improvements across language and clinical metrics. On MIMIC-CXR, CPVB-RG achieves 0.2847 F1RadGraph and 0.6692 VLScore, with statistically significant gains over the closest decoder-matched prior-aware baseline. These results support the effectiveness of the proposed difference-aware and spatially guided fusion design for prior-aware chest X-ray report generation.
[CV-48] HRoPE: Homography Relative Positional Encoding
链接: https://arxiv.org/abs/2610.28284
作者: Huachen Gao,Yang Yang,Xiaoqian Liang,Xingyu Chen,Youjia Zhang,Qi Zhang,Baorui Ma,Anpei Chen,Wei Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, 10 figures
Abstract:Geometric Relative Positional Encoding has become the primary mechanism for injecting spatial inductive biases into multi-view transformers. Existing designs condition attention on camera-level frustums or ray-marched depth, associating cross-view tokens through the sensor geometry. Such sensor-centric formulations, however, describe only where a token is observed from, leaving the rich surface geometry of the scene region depicted by the token. In this paper, we advocate a shift toward surface-aware positional encoding that explicitly grounds patch tokens in the geometry of physical surfaces. We introduce HRoPE, which represents the local geometry of each patch as a slanted 3D plane, following the slanted support-window assumption of classical multi-view stereo. We inject token-level, plane-induced homographies into attention as relative transformations. This equips the transformer with a geometric prior for rectifying perspective distortions and aligning patch features across viewpoints. Our encoding requires no ground-truth geometry: the plane parameters can be predicted end-to-end under a view-synthesis objective alone, or derived from foundation models. Extensive experiments on RealEstate10K, CO3Dv2, and DL3DV show that our best variant of HRoPE consistently outperforms the evaluated geometric positional encodings for novel view synthesis. Extensive ablations and analysis show the effectiveness of the proposed encoding. Please refer to our project page at: this https URL
[CV-49] ChartEvo: Self-Evolving Agent Harnesses for Long-Tail Chart Reasoning
链接: https://arxiv.org/abs/2610.28282
作者: Qi Wang,Peijie Wang,Jun Cao,Haoyang Tong,Xiao-Hui Li,Fei Yin,Cheng-Lin Liu
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal large language models (MLLMs) have substantially advanced chart understanding and reasoning, yet their performance remains highly uneven across chart types. While modern MLLMs perform well on common charts, their accuracy drops sharply on long-tail charts with diverse visual structures and reasoning requirements. Tool-augmented agents can improve complex chart reasoning by introducing specialized visual capabilities, but manually engineering strategies and tools for the heterogeneous and evolving long tail is difficult to scale. We introduce \textbfChartEvo, a recursive self-improvement framework that enables chart agents to adapt to underperforming chart types through experience-driven harness-level self-improvement. ChartEvo jointly evolves reusable chart-specific Skills and executable Tools, while learning reusable local tool-call chains to compose evolved capabilities and control the inference overhead of tool-augmented reasoning. On long-tail charts from ChartBench and ChartX, ChartEvo improves accuracy over its base MLLM, Qwen3.7-Plus, by \textbf45.6% and \textbf32.5%, respectively, demonstrating the effectiveness of recursive self-improvement for long-tail chart adaptation.
[CV-50] GALA: Geometry-Aware Language Model for Controllable Object Arrangement SIGGRAPH
链接: https://arxiv.org/abs/2610.28280
作者: Chucheng Xiang,Runze Wang,Zhi Deng,Ruchao Bao,Yuanwei Zhang,Yuxuan Xie,Hanliu Wang,Liangzhen Fei,Wenzheng Wu,Cheng Wan,Peifeng Li,Zhongyuan Liu,Ligang Liu
类目: Graphics (cs.GR); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to SIGGRAPH Asia 2026 Conference Papers. Project Page: this https URL
Abstract:Reasoning about fine-grained placement constraints, such as support and containment, is essential for realistic 3D scene creation. It remains a significant challenge for existing methods that rely on simplified object representations like oriented bounding box (OBB), which are blind to the detailed geometry. To address this limitation, we introduce GALA, a geometry-aware language model with explicit point cloud perception for controllable object placement. Our model performs autoregressively by predicting the 6D pose of one target object at a time conditioned on the scene context and placement instruction. We utilize a two-stage training strategy: the first stage aligns geometric and textual modalities through pretraining tasks, and the second stage fine-tunes the model for placement. To facilitate this task, we also curate a multi-source scene dataset covering diverse room types and fine-grained furniture arrangements, which will be open-sourced for community benefit. Experiments on our benchmark show our model significantly outperforms baselines without explicit geometric reasoning. We further develop a multi-stage plan-place-verify agent pipeline that uses the model for text-driven scene generation. Project Page: this https URL
[CV-51] VGL-stGen: Learning a shared multimodal model for lung histology and spatial transcriptomics
链接: https://arxiv.org/abs/2610.28266
作者: Haenara Shin,Dongjoo Lee,Sungwoo Bae,Kwon Joong Na,Daeseung Lee,Hongyoon Choi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM)
备注: 46 pages, 18 figures, 16 tables; appendix included
Abstract:VGL-stGen is a multimodal large language model adapted from a pretrained medical vision-language model using paired lung HE and spatial transcriptomics with single-cell references. Training integrates image-expression and expression-cell-type pairs, and gene-conditioned pseudo-labels transfer transcriptomic cell-type supervision to histology images. One deployed checkpoint supports thirteen evaluated tasks spanning tissue and cell-type classification, gene prediction and conditional expression completion from combinations of histology, cell identity and observed gene expression. Task-specific decoding produces categorical labels, gene lists and structured expression levels. For conditional completion, fixing queried gene identities and predicting their levels autoregressively improves structured prediction. Controlled adaptation shows that restricting supervision to requested expression-level tokens can further improve image-conditioned completion. Image-permutation experiments indicate that correctly paired histology adds information beyond observed expression. External classification relies partly on computational references with shared annotation lineage, and some HE sources were exposed during cancer training or model development. Within these conditions, VGL-stGen demonstrates how a medical vision-language backbone can learn from spatial transcriptomics and histology to connect tissue morphology and molecular states through shared generative modeling.
[CV-52] PI-MLLM : Psychophysics-Inspired Multimodal Large Language Model for Oddly Satisfying Video Emotion Analysis
链接: https://arxiv.org/abs/2610.28264
作者: Yuheng Zhang,Jie Zhang,Yuecong Min,Yunwei Zhao,Meichen Liu,Qiankun Li,Shiguang Shan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 12 pages, 5 figures, 8 tables; includes supplementary material
Abstract:Oddly Satisfying Videos (OSVs) can evoke strong viewer responses through repetitive motion, visual tactility, and ASMR-like sounds despite containing few explicit semantic emotion cues. This makes OSVs a challenging setting for Video Emotion Analysis (VEA), since most existing methods focus on semantic video content and under-model the physical sensory dynamics that shape viewer responses. To support Oddly Satisfying Video Emotion Analysis (OSVEA), we construct the Oddly Satisfying Video Dataset (OSVD), comprising 12,000 annotated 5-second clips. Each clip provides viewer-emotion labels under video-only, audio-only, and audio-visual conditions, together with explanations grounded in visual and acoustic sensory evidence. Building on OSVD, we propose PI-MLLM, a psychophysics-inspired multimodal large language model that integrates Modality-Specific Dynamics Alignment (MSDA) and Coalition-Based Audio-Visual Attribution (CAVA). MSDA uses auxiliary prediction objectives to align modality representations with visual motion and acoustic energy-onset dynamics. CAVA evaluates emotion scores across modality coalitions and applies Shapley attribution to identify the dominant modality and determine whether visual and auditory evidence is synergistic or conflicting. Experiments show that PI-MLLM improves both viewer-side emotion prediction and sensory-grounded explanation quality, suggesting the value of physical sensory cues to video emotion analysis.
[CV-53] Reliability-Aware Reconstruction of V-Shaped Surfaces with a Low-Cost Multi-Zone SPAD Sensor
链接: https://arxiv.org/abs/2610.28258
作者: Yohanssen Pratama
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to IEEE Sensors Letters; 8 figures, 2 tables
Abstract:Low-cost multi-zone single-photon avalanche diode (SPAD) sensors provide compact time-of-flight measurements but are limited by sparse sampling and unreliable depth estimates near geometric discontinuities. This letter presents a reliability-aware reconstruction framework for V-shaped surfaces using two complementary TMF8821 acquisition configurations. Point-wise reliability is learned from sensor confidence, transient peak amplitude, and peak sharpness, and is incorporated into a graph-based depth optimization. The reconstructed geometry is evaluated by robust two-plane fitting and a perpendicular cross-section. Experiments on 60-degree, 90-degree, and 120-degree targets reduce the mean V-angle error from 11.21 degrees to 9.66 degrees (13.8%). For the representative 90-degree target, the combined point-to-plane RMSE decreases from 14.56 mm to 13.65 mm.
[CV-54] SplatFlux: Feed-Forward 4D Gaussian Reconstruction and Generation from Video Latents
链接: https://arxiv.org/abs/2610.28257
作者: Runmin Zhang,Binglin Ren,Zhu Yu,Fudong Wang,Jiangke Lin,Yichun Shentu,Si-Yuan Cao,Yi Yuan,Hui-Liang Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:This work presents SplatFlux, a feed-forward latent-to-4DGS framework for both video-based 4D reconstruction and conditional 4D generation. Our key observation is that observed videos and conditionally generated content can be represented in the same pretrained video latent space, allowing both tasks to be formulated as latent-to-4DGS decoding. To construct compact and temporally coherent 4D scenes, SplatFlux introduces a static-dynamic decoupled 4D Gaussian representation that explicitly predicts dynamic probabilities for Gaussian primitives. This representation separates static scene structures from moving regions, enabling complete 4D scene construction in a feed-forward manner through static aggregation and dynamic motion interpolation. To learn static-dynamic decomposition and dynamic Gaussian motion without costly annotations, we further develop an auxiliary distillation strategy that provides pseudo dynamic masks and optical-flow cues. Extensive experiments across 3D and 4D reconstruction and generation demonstrate that SplatFlux achieves state-of-the-art performance.
[CV-55] Convergent Performance Divergent Mechanisms in LoRA-Adapted Vision Foundation Models for Brain MRI
链接: https://arxiv.org/abs/2610.28255
作者: Cecilia Diana-Albelda,Fiona Kekwick,Raghav Mehta,Aleksandra Kulbaka,Paul M. Matthews,Jesus Bescos,Álvaro García-Martín,Wenjia Bai,Bernhard Kainz
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 14 pages, 6 figures, Preprint submitted to Elsevier
Abstract:Adapting Vision Foundation Models (VFMs) to medical imaging via LoRA has become standard practice, yet model selection is guided almost entirely by downstream task metrics, treating backbones with similar performance as interchangeable. We investigate whether equal performance implies equal internal representational shift. We analyse four ViT-B/16 backbones, DINOv3, SigLIP2, MerMED-FM, and BiomedCLIP, pre-trained with different objectives (self-distillation vs. contrastive vision-language) and on different data domains (natural vs. medical images). We adapt each to brain tumour segmentation and evaluate on six datasets covering both segmentation and Alzheimer’s disease classification. Using attention divergence, weight-change analysis, and activation patching, we find that the four models converge to similar segmentation performance (mean Dice 78-81%) through markedly different internal dynamics, yet diverge in cross-task transfer: segmentation adaptation improves Alzheimer’s disease classification for DINOv3 but degrades it for the other three backbones. We reveal a dissociation between modification magnitude and causal importance: the most heavily modified layers are not the causally relevant ones, with early-to-mid layers identified as sufficient for segmentation recovery. Restricting LoRA to these layers matches or approaches full-layer segmentation while reducing classification degradation in the three affected backbones, providing a mechanistically grounded strategy that is more parameter-efficient and more robust than standard full-layer adaptation.
[CV-56] Region-Guided Food-Package Evidence Extraction and Abstention-Aware Draft R3337 Screening with Selective VLM Review
链接: https://arxiv.org/abs/2610.28211
作者: Mayimunah Nagayi,Alice Scaria Khan,Tamryn Frank,Rina Swart,Clement Nyirenda
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 17 pages, 2 figures, 4 tables
Abstract:Applying South Africa’s draft R3337 warning criteria to food-package photographs requires nutrition, ingredient and product-form evidence. This study presents a region-guided evidence extraction and abstention-aware screening workflow for draft R3337. A three-class detector localises nutrition information panels, ingredient lists and product-identity regions, after which PaddleOCR and structured extraction recover nutrient values, ingredient evidence and product-form information. Recovered evidence is aggregated and assessed using deterministic draft R3337 rules. When evidence is incomplete, uncertain or conflicting, the workflow abstains through Review or Insufficient evidence rather than forcing an automatic outcome. For qualifying unresolved cases, a selective vision-language model (VLM) review stage using Qwen3-VL 8B Instruct reassesses unresolved evidence. Accepted evidence is merged and the same deterministic rules are reapplied. The workflow was applied to 442 products represented by 3,929 images. Evaluation used a manual validation set of 65 products. Before selective VLM review, 31 of 60 decidable products received automatic outcomes and 30 were correct, giving 51.7% coverage and 96.8% accuracy among automatic outcomes (selective accuracy). Selective VLM review resolved eight additional products, increasing coverage to 65.0%. Of 39 automatic outcomes, 38 were correct, giving 97.4% selective accuracy. The single false warning was retained from the rules-only workflow, while no unsafe warning omissions occurred after selective review. Evidence-level evaluation nevertheless identified incorrect VLM-recovered nutrition, ingredient and product-form evidence. These findings show that selective VLM review can recover unresolved evidence, while deterministic screening and explicit abstention remain necessary for reliable image-based draft R3337 screening.
[CV-57] Learn to Modify: Self-evolving Multi-agent Refinement Framework for Conversational Text-to-Motion Generation
链接: https://arxiv.org/abs/2610.28210
作者: Jianan Li,Jiawei Zhao,Yangtao Zhou,Zhifu Zhao,Shaoxin Wu,Yutao Wang,Dezhen Wang,Hua Chu,Fu Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Despite rapid progress in text-to-motion generation, existing methods are commonly designed for single-shot generation. When users are dissatisfied with specific motion details, they are often forced to regenerate the entire sequence from scratch, which is computationally inefficient and potentially destroys previously established kinematic context. A practical generation system should instead support iterative user feedback and refine only the affected spatio-temporal regions while preserving the remaining motion. Therefore, we introduce a conversational motion generation paradigm that requires the system to coordinate instruction understanding, motion generation, localized editing, and quality assessment. Recent advances in LLM-driven multi-agent systems provide a promising foundation for such role-specialized coordination. To tackle these challenges, a Self-evolving Multi-Agent RefinemenT Framework (SMART) is proposed for conversational motion generation. Within SMART, to preserve semantic consistency, we design a semantic-guided masked motion editing mechanism, which precisely grounds natural language feedback into spatiotemporal motion regions to enable localized refinements without altering unedited regions. To reduce redundant exploration, we propose a memory-based self-evolving strategy that dynamically accumulates motion editing knowledge from interaction history and reuses past successful experience. Extensive experiments demonstrate that SMART effectively supports multi-turn conversational motion generation and enables efficient motion refinement.
[CV-58] Attractive Faces Skew Text-to-Image Reward Models and Leaderboards WACV2027
链接: https://arxiv.org/abs/2610.28195
作者: Harvey Mannering,Zhiwu Huang,Adam Prügel-Bennett
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: WACV 2027
Abstract:Human preference is now central to the evaluation, guidance, finetuning, and ranking of text-to-image models. However, the susceptibility of human preferences to bias in image generation remains underexplored. We identify a key observation: for human faces, perceived attractiveness overrides image quality in preference judgments, even when comparing real photographs of identical fidelity. In a controlled user study, annotators rank more attractive faces as higher quality, even when unattractive real faces are compared with attractive AI-generated ones. We examine the downstream consequences of this bias. Firstly, simulating Elo rankings from user study data shows attractiveness can invert a leaderboard’s ranking, elevating synthetic attractive faces over real, unattractive ones. Secondly, the same bias reappears in the human preference models. Editing 40 annotated facial attributes in isolation, we find that attractiveness is the strongest correlate of score for four of five human preference models (PickScore, MPS, HPSv2.1, HPSv3), and the strongest predictor of score in regression analysis. We further show that preference aligned generation shifts faces toward younger, more feminine faces, and can override explicit prompts for unattractive faces, thereby limiting user control. Finally, we explore mitigation strategies, both for data collection and post training on human preference models. These findings highlight the need for preference models that disentangle image quality from human cognitive bias, enabling more reliable evaluation.
[CV-59] Offline Makes Online Better: Calibrated Cache Compression for Video Generation
链接: https://arxiv.org/abs/2610.28192
作者: Zhen Wang,Ning Lai,Qingtao Yu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Token merging is an effective approach to reducing KV-cache memory consumption in autoregressive video generation by clustering temporally redundant tokens and retaining only a small set of representatives. Existing methods, however, typically apply a uniform merging ratio across all attention layers and heads. We observe that different layers and heads exhibit substantially different yet jointly correlated sensitivities to token merging, and that these sensitivity patterns remain consistent across prompts. Based on this observation, we introduce CALM, a Calibrated Allocation for Layer-Head Merging that uses offline calibration to improve online cache compression. Exhaustively evaluating all possible layer-head allocations is prohibitively expensive. To enable efficient calibration, we show that attention-output error strongly correlates with end-to-end video quality and can therefore serve as a lightweight proxy for allocation risk assessment. We evaluate different allocation strategies using this proxy and construct a fixed layer-head risk map offline. Given a cache budget at inference time, CALM uses dynamic programming to determine the best allocation of merging ratios. Experiments across three different autoregressive video models demonstrate that CALM consistently outperforms uniform token merging under the same memory budget, by reducing the LPIPS to original videos by around 21% with higher video-quality scores.
[CV-60] Unified RAW Image Exposure Correction and Restoration via Linear Illumination and Frequency-Adaptive Enhancement
链接: https://arxiv.org/abs/2610.28191
作者: Yingkai Zhang,Yunhao Zou,Ying Fu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Incorrect exposure degrades both image brightness and recoverable scene detail. Existing sRGB-based correction methods are limited by the non-linear image signal processing pipeline, whereas RAW-based methods are usually developed for a single degradation type, such as low-light denoising or high-light recovery. In this paper, we focus on joint exposure correction and restoration directly in the RAW domain and propose a unified RAW Image Exposure Correction and Restoration (RIECR) framework for both under- and over-exposed images. RIECR decomposes the problem into two coupled sub-problems that follow the RAW imaging model: linear illumination correction and residual degradation restoration. First, a Linear Illumination Correction Network (LICN) estimates exposure-aware linear gains by using channel-specific RAW cues. It exploits the high signal-to-noise ratio of Green channels for robust brightening and the relatively unsaturated Red/Blue responses for high-light darkening. Second, a dual-branch Frequency-based Enhancement Network (FEN) restores the remaining degradation with a Frequency Dynamic Convolution (FDC) module. FDC decomposes convolutional responses into low- and high-frequency components and adaptively suppresses amplified noise or recovers clipped textures according to the exposure context. We further build a real paired RAW benchmark containing 223 scenes and 1,784 pairs covering complex multi-exposure errors. Experiments against sixteen state-of-the-art methods show that RIECR achieves state-of-the-art performance on real RAW exposure correction.
[CV-61] CrossFit-GS: Out-of-Fold Pose Estimation for COLMAP-Free 3D Gaussian Splatting
链接: https://arxiv.org/abs/2610.28190
作者: Zijian Wu,Jinliang Wang,Mingfeng Jiang,Zidian Lin,Ying Song,Ziqian Lu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Pose-free 3D Gaussian Splatting (3DGS) aims to bypass cumbersome Structure-from-Motion (SfM) by directly reconstructing scenes from unposed image sequences. In this work, we present CrossFit-GS, a deferred-commitment progressive framework that mitigates cumulative trajectory drift and breaks the self-reinforcing loop of distorted-map error propagation. Specifically, we partition the input sequence into interleaved subsets and maintain two complementary submaps. Each candidate pose is cross-refined against a frozen opposite submap that has never been optimized using the target frame, thereby providing an independent geometric reference for pose estimation. Checkerboard validation and spatial coverage safeguards further reject unreliable updates, and only verified poses are committed to their corresponding owner submap. Once the trajectory has been refined, the intermediate submaps are discarded and the final global representation is trained from scratch using the fixed refined poses. Extensive evaluations across multiple benchmarks demonstrate that CrossFit-GS substantially reduces cumulative trajectory drift and consistently improves both camera pose estimation and novel-view synthesis over existing methods.
[CV-62] DocLayout-YOLO: Axis-Attentive Document Layout Detector with Mesh-Guided Synthetic Pretraining
链接: https://arxiv.org/abs/2610.28171
作者: Zhiyuan Zhao,Hengrui Kang,Guang Liang,Bin Wang,Conghui He
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Accurate and efficient document layout analysis is critical for modern document parsing systems. However, although the YOLO framework represents a leading paradigm for the real-time detector, its adaptation to heterogeneous documents remains underexplored. To address this gap, we propose DocLayout-YOLO, a document-specialized real-time detector that integrates diverse synthetic document pretraining and page-axis context modeling. From the data perspective, we introduce MeshLayout-1M, a synthetic pretraining corpus generated by the proposed Mesh-candidate BestFit algorithm, which provides large-scale and diverse document layouts for pretraining and improves adaptation to heterogeneous document scenarios. From the model perspective, we propose Page-Axis Layout Attention (PALA), which captures long-range dependencies along the horizontal and vertical page axes while remaining computationally efficient. Extensive experiments demonstrate that DocLayout-YOLO achieves state-of-the-art performance on four heterogeneous document benchmarks, obtaining 72.4%/75.5%/96.0%/73.1% mAP50:95 on D4LA, M6Doc, HJDataset, and PRIMA-LAD, respectively. Compared with existing real-time detectors, DocLayout-YOLO establishes a superior accuracy-efficiency trade-off in document layout analysis. Code, model, and data are available at this https URL.
[CV-63] OV-Foam: How 3D Representations Shape Semantic Feature Lifting
链接: https://arxiv.org/abs/2610.28131
作者: Yousef Rajeh(1),Alejandra Esparza Pelcastre(2),Ivan Viola(1),Dominik Engel(1) ((1) King Abdullah University of Science and Technology, (2) Universidad Panamericana)
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: 22 pages, 10 figures, 7 tables
Abstract:Existing feature lifting approaches for open-vocabulary segmentation rely on simple backprojection to assign features to primitives. In the case of 3D Gaussian splatting reconstruction, the primitives overlap heavily, leading to suboptimal reconstructions, governed by many primitives per ray. Backprojection, however, reduces to a least-squares solvable problem, exactly when no two primitives share a ray. Motivated by this insight, we propose OV-Foam, which uses PowerFoam as the underlying representation. This representation uses a power-diagram to partition the scene, resulting in zero overlap between primitives and a substantial reduction in primitives per ray, which ultimately simplifies the lifting problem and allows for closer-to-optimal feature assignment. We show that the space partitioning primitives of OV-Foam reach near-optimal assignment, and our experiments demonstrate improved open-vocabulary segmentation over Gaussian baselines.
[CV-64] GAMF-Net: Grouped Availability-Aware Multimodal Fusion for Smartphone Skin Lesion Classification
链接: https://arxiv.org/abs/2610.28129
作者: Yue Zhuo,Jingjia Yan,Fangyu Liu,Lingjun Liu,Zongjun Chen,Zhisheng Luo,Zhonghua Xie
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 4 figures. Accepted at MMM 2027
Abstract:Smartphone clinical images provide a low-cost data source for computer-aided skin lesion classification in resource-constrained settings, but images alone cannot fully capture patients’ medical histories and lesion symptoms. To address the lack of structured organization among heterogeneous clinical variables, the ambiguity between observed and imputed values, and the risk that clinical information may interfere with visual evidence, we propose GAMF-Net, a grouped availability-aware multimodal fusion network. Ten clinical variables are divided into three semantic groups—demographics, history-related risk, and lesion presentation—and each group jointly encodes field values and field-level availability masks. The image branch produces the base classification logits, while each clinical group generates image-conditioned class-level residual evidence. Only available groups participate in availability-normalized equal-weight fusion; when all clinical information is unavailable, the model reduces exactly to the image-only pathway. We evaluate the method on PAD-UFES-20 under a patient-disjoint five-fold protocol with separate training, validation, and test partitions. GAMF-Net achieves an Accuracy of 74.17\pm2.85% , a balanced accuracy (BACC) of 70.11\pm2.11% , a Macro-F1 of 68.85\pm0.85% , and a Macro-AUROC of 93.42\pm1.20% , outperforming all image-only and multimodal baselines reimplemented under a unified protocol. Compared with the strongest baseline, MetaBlock, the corresponding improvements are 2.37, 2.59, 4.14, and 1.02 percentage points, respectively. Ablation results show that clinical information substantially improves class-balanced performance, constrained equal-weight fusion is more effective than dynamic fusion, and the lesion-presentation group provides the most informative clinical complement.
[CV-65] PointForward V2: Point-Aligned Interaction for Feedforward Driving Reconstruction
链接: https://arxiv.org/abs/2610.28120
作者: Xianqi Wang,Cheng Chi,Yuechuan Pu,Hongcheng Luo,Junda Cheng,Guang Chen,Hangjun Ye,Bing Wang,Xin Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Feedforward reconstruction of dynamic driving scenes requires integrating observations across multiple views and time steps into a coherent 3D representation. PointForward established state-of-the-art feedforward driving reconstruction by lifting image observations into sparse 3D queries, demonstrating the effectiveness of point-aligned scene representations. However, these queries are largely processed independently, limiting information exchange across space and time, while direct global attention among tens of thousands of point queries is computationally prohibitive. In this work, we present \textbfPointForward V2, which advances point-aligned representation to point-aligned interaction. Our key idea is to mediate query interaction through compact, scene-structured registers. For the static background, frame-wise Static Scene Registers compress point information, communicate across frames, and propagate the aggregated context back to the queries. For each dynamic instance, queries from different frames are aligned in an object-centric canonical space and interact through instance-specific Dynamic Instance Registers. We further employ 4D positional encoding that combines aligned spatial coordinates with original timestamps, enabling spatially aligned yet temporally distinguishable interactions. In this way, PointForward V2 enables scalable spatio-temporal reasoning directly among lifted 3D queries without dense global attention. Experiments on Waymo, nuScenes, and Argoverse 2 demonstrate state-of-the-art performance across large-scale driving benchmarks.
[CV-66] Reliability-Guided Vision-Language Decoding with Text-Aware Skip Refinement for Medical Image Segmentation
链接: https://arxiv.org/abs/2610.28118
作者: Sivan Talaei,Fatemeh Daneshfar,Mustaqeem Khan,Abdulhady Abas Abdullah,Mourad Oussalah
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Medical image segmentation remains challenging when targets exhibit limited visual distinction, irregular structure, or substantial variation across imaging modalities. Medical text can reduce this ambiguity, but indiscriminate fusion may introduce unreliable guidance. We present RELiF, a vision–language framework that integrates text through reliability-guided decoding. Its Reliability-guided Fusion Gate (RFG) adaptively controls cross-modal updates, while the Text-guided Skip Refinement Adapter (TSRA) refines high-resolution features using text guidance. RELiF is evaluated on radiographic, CT, and endoscopic segmentation tasks, where it shows competitive performance across diverse imaging characteristics and target appearances. Additional analyses examine efficiency and stability. The complementary uncertainty representation derived from the reliability response also shows an association with prediction difficulty, with higher uncertainty values occurring more frequently in pixels associated with segmentation errors. Code: this https URL.
[CV-67] LatticeRank: Spacing-Ranked Consistency for Native-Grid 3D Medical Image Segmentation
链接: https://arxiv.org/abs/2610.28091
作者: Jingjia Yan,Yue Zhuo,Zhonghua Xie
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 6 figures. Accepted at MMM 2027
Abstract:Variable voxel spacing changes more than image resolution: it alters the discrete sampling lattice, tensor geometry, and anatomical evidence available to a 3D segmentation network. While conventional pipelines restore images to a canonical grid, direct-grid deployment requires the model itself to tolerate these changes. We observe that candidate grids are not exchangeable views – a coarse grid removes more structural information than a mildly resampled one – yet standard consistency objectives assign them equal influence. We therefore propose LatticeRank, a training-only framework that aligns candidate posteriors in physical space and allocates consistency supervision by voxel-spacing order: among endpoint-matched candidates, the auxiliary loss concentrates on the coarsest requested grid, while inference retains a single network forward with no candidate branch or ensemble. On MSD Task04, LatticeRank improves coarse-grid Extra Dice over mean consistency by 0.03436 across five seeds while meeting the prespecified reference-grid noninferiority criterion. A controlled ladder shows that grid exposure provides the largest recovery, posterior consistency adds a second gain, and spacing-ranked allocation yields the strongest direct-grid result; requested spacing thus provides a simple, model-independent ordering signal for direct-grid consistency learning.
[CV-68] WildAIGC-Bench: Can AI Image Detectors Handle the Wild? A Benchmark of Generated and Edited Images in Online Circulation
链接: https://arxiv.org/abs/2610.28087
作者: Zijian Yu,Hangwei Zhang,Jun Lan,Huijia Zhu,Weiqiang Wang,Jianfu Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 19 pages
Abstract:AI-generated images linked to news events and online claims can mislead audiences, enable fraud, or cause public controversy. We introduce WildAIGC-Bench, built through recurring collection across news websites, fact-checking sources, and related social media content. The pipeline combines broad coverage, timely updates, and human-verified labels through scheduled discovery, automatic screening and preliminary annotation, and human review of image-specific evidence. The benchmark contains 407 verified AI-generated or edited images, each accompanied by a source webpage link and an evidence summary. Images retain the conditions encountered online, including compression, screenshots, and recapture. Evaluation of dedicated visual detectors, multimodal large language models, and a simple Perception Encoder baseline reveals substantial recall gaps relative to existing benchmarks. WildAIGC-Bench provides both a practical detection challenge and a basis for continued data expansion and source-based verification. The dataset is released at this https URL.
[CV-69] MoDDiT: Simple Token Routing in Diffusion Transformers
链接: https://arxiv.org/abs/2610.28086
作者: Jarek Liesen,Matthew T. Jackson,Benjamin Ellis,Shimon Whiteson,Jakob N. Foerster
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Real-time video generation, interactive world models, and on-device image synthesis are almost a reality. Diffusion Transformers (DiT) achieve excellent sample quality, shifting the bottleneck to the throughput of the models themselves. Natural images are inherently sparse, as large portions of an image are often background, contribute little to the scene content, and are easy to denoise. DiTs fail to exploit this sparsity, as they process image tokens statically and uniformly. We propose MoDDiT, a simple architecture modification that learns to route tokens through both MLP and self-attention operations. By routing tokens through both blocks independently, MoDDiT allows for dynamic compute allocation in MLPs that varies across tokens, images, and denoising timesteps. Using MoDDiT to train image generation models from scratch significantly reduces computational costs of training and inference, processing as few as 12.5% of the tokens. By increasing sparsity and model depth, MoDDiT reduces costs and improves generation performance over unrouted baselines. The learned allocation is interpretable, revealing how computation varies with visual content and noise level. We also use MoDDiT to train video generation models, where allocation correlates with motion, and to distill a dense bidirectional video model into a sparse autoregressive student.
[CV-70] Slot-JEPA: Object-Structured Prediction Interfaces for Latent World Models
链接: https://arxiv.org/abs/2610.28074
作者: Peisong Wen,Qianqian Xu,Minye Lei,Siran Dai,Yang Liu,Qingming Huang
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:As a prototypical latent world model, Joint-Embedding Predictive Architectures (JEPAs) provide strong semantic representations. However, dense patch representations expose a redundant interface to downstream predictors, which must implicitly infer spatial organization and temporal correspondence, presenting three challenges. First, object information is distributed across patches, forcing each downstream predictor to redundantly recover object structure that could otherwise be reused across tasks before modeling interactions. Second, the lack of explicit entity correspondence across time makes historical transitions expensive, restricting the predictor’s flexibility in utilizing context. Third, without explicit transport structure, existing JEPA models typically use pointwise l1/l2 distances, which are insensitive to spatial displacement when predicted and goal supports are largely disjoint. We propose the Slot-based Joint-Embedding Predictive Architecture (Slot-JEPA), which decomposes dense representations into stable semantic slots and time-varying spatial masks. Pretrained once and frozen downstream, this decomposition exposes a reusable object-structured interface. Building on it, our Transition Memory Predictor (TMP) stores visual transitions as compact masks and retrieves useful context using statistics of corresponding slots. Its two branches separate slot-specific dynamics from shared cross-slot dynamics. We further use Unbalanced Optimal Transport (UOT) to supervise mask prediction and evaluate candidate plans by transporting mass between aligned predicted and goal masks, providing a signal better aligned with spatial progress. Across seven visual-understanding benchmarks, the slot-reconstructed features remain transferable. Slot-JEPA also improves goal-conditioned planning across six planar environments and four robotic arm planning tasks on the REALM-JEPA benchmark.
[CV-71] Auditing Where Concept Bottleneck Accuracy Comes From on CUB
链接: https://arxiv.org/abs/2610.28059
作者: Sascha Saralajew,Eleonora Poeta,Eliana Pastor,Ammar Shaker
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Concept Bottleneck Models (CBMs) promise predictions that humans can inspect and correct through concepts such as “curved beak”. Yet on CUB-200-2011, their standard benchmark, progress is measured like that of ordinary classifiers: by task and concept accuracy. Neither reveals whether the label is predicted from the exposed concept scores, or whether a concept is supported by visual evidence of that concept rather than by co-occurring features such as cues to the species as a whole. We audit where the accuracy of CBMs and Concept Embedding Models (CEMs) on CUB comes from. The CUB annotations are too noisy to explain it: class-level training labels agree with per-image annotations on only ~82% of image–concept pairs, and classifiers trained on per-image annotations reach only ~37% species accuracy. Leakage probes show that CEM predictions cannot be reconstructed from their concept scores; the missing class information sits in internal embeddings. We identify four design axes through which class information can enter concept scores—receptive field, shared versus concept-specific representations, norm-dependent scoring, and task supervision. An ablation shows that most of it enters through non-local features: restricting a ResNet50 backbone to local BagNet patches lowers the accuracy recoverable from concept scores from 72% to 43%, close to what ground-truth annotations support, while joint training raises it again at the cost of concept accuracy. Localization and occlusion show that locality makes grounding measurable but does not guarantee it: under class-level labels, local concept scores use the head as a proxy for body, wing, and tail concepts, and CEM concept scores barely respond to occluding their supporting part. Task and concept accuracy should thus support, not drive, the evaluation of CBMs, and given the noisy annotations, CUB accuracy gaps should not be overinterpreted.
[CV-72] Assessing Representation Semantics Under Language Pretraining in Fine-Grained Classification
链接: https://arxiv.org/abs/2610.28051
作者: Md Raisul Kibria,Sébastien Lafond
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Understanding how supervision signals shape semantic visual representations remains challenging, particularly in fine-grained, low-data settings. Standard evaluations of vision models rely on downstream accuracy, which often fails to capture structural differences in learned representations. We propose a compact explainability-driven framework for comparing encoders trained with classification-only and caption-generation supervision on CUB-200. Using a controlled two-stage protocol in which both encoders are ultimately optimized for the same classification task, we enable a direct comparison of the resulting representations. Saliency maps are generated using attention rollout and gradient-weighted rollout and evaluated with five metrics covering faithfulness (positive and negative perturbation) and localization (keypoint emphasis, background attribution, and Earth Mover’s Distance). Across random seeds and explanation methods, caption supervision yields representations that are more sensitive to the removal of salient regions, better aligned with annotated object parts, and less reliant on background regions, whereas classification supervision produces representations that are slightly more robust to the removal of low-saliency regions. An ablation study on the Oxford-IIIT Pet dataset shows consistent trends across key metrics. These results demonstrate that the proposed framework makes supervision-induced representational differences observable and quantifiable beyond downstream accuracy alone. More broadly, it provides a systematic basis for assessing learned representations and comparing alternative training and transfer-learning strategies.
[CV-73] A Process-Informed Photorealistic Digital Twin for Vision-Based Inline Inspection of C-SMC Strip Deposition
链接: https://arxiv.org/abs/2610.28050
作者: Patrick Flore,Fabian Röder,Kevin Chen,Andreas Gebhard
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: 46 pages, 6 figures. Submitted to Composites Part A: Applied Science and Manufacturing
Abstract:Inline quality assurance for Sheet Molding Compound (SMC) manufacturing is hindered by the lack of large-scale, accurately labeled image datasets for training machine-learning models. This work presents a process-informed, photorealistic digital twin for carbon-fiber SMC strip deposition. A rigid-body simulation reproduces the deposition kinematics (settling, friction, belt motion), while a physically grounded polarization-rendering model produces realistic appearance, so that every image carries exact, mutually consistent labels in synchronized formats: oriented bounding boxes, point annotations, density maps, and polarization renderings (Angle and Degree of Linear Polarization). Because the joint distribution of many overlapping strips cannot be specified analytically, simulation yields exact, scalable ground truth that manual annotation cannot provide. The twin produces over 16,000 images across 3,300 frames, and the synthetic data replicate near-isotropic fiber orientation (axial concentration r 0.05) consistent with as-deposited commercial SMC. Trained purely on synthetic data, YOLOv11 and CSRNet reach detection mAP50 up to 0.91, with a dedicated oriented-bounding-box detector confirming the oriented labels are directly trainable (mAP50 = 0.83); on real manufacturing images, the synthetic-to-real difference in edge and texture structure (Kolmogorov-Smirnov distance 0.46, distribution overlap 0.56) is of the same magnitude as the difference between two real acquisition series of the same material (0.41, overlap 0.57). The approach substantially reduces the labeling burden for camera-based inline inspection of fiber orientation and volume content.
[CV-74] Spherical Geometry-Aware Panoptic Scene Graph Generation for Equirectangular Panoramas ICASSP2027
链接: https://arxiv.org/abs/2610.28033
作者: Zhe Wu,Xiaoyan Fei,Xinyi Jiang,Yanxu Wei,Junming Rao,Yihong Yan,Xuewei Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to ICASSP 2027
Abstract:Panoptic scene graph generation (PSGG) represents scene entities with segmentation masks and directed relations. Equirectangular projection (ERP) panoramas challenge PSGG through distortion, latitude-dependent pixel areas, and an artificial seam. We propose a two-stage spherical geometry-aware framework that fuses overlapping perspective-view predictions using observation reliability, spherical overlap, and angular proximity. Relation reasoning combines spherical-area-aware aggregation, weighted no-relation supervision, and dual-longitude score fusion. We also construct a dataset of 340 high-resolution panoramas with 10,571 nodes and 6,632 directed relations. Against corresponding baselines, spherical fusion improves PQ by 36.03% under matched views, while relation reasoning improves mR@50 by 11.20% under PredCls with ground-truth masks and object labels. The full framework improves mR@50 by 52.99% under SGDet with predicted panoptic segments. These results demonstrate the value of spherical geometry for panoramic PSGG.
[CV-75] P3-VLM: A Point-based Alternative for Grounded 3D Vision-Language Models NEURIPS2026
链接: https://arxiv.org/abs/2610.28018
作者: Anna-Maria Halacheva,Jan-Nico Zaech,Sombit Dey,Luc Van Gool,Danda Pani Paudel
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026, Spotlight
Abstract:Examining existing benchmarks for grounding in 3D VLMs reveals a pervasive “semantic leakage” in the evaluation and prompting protocols: a heavy reliance on bounding boxes as spatial prompts or representation primitives. Because bounding boxes are highly informative of object categories, they introduce a shortcut for the models to neglect visual information during 3D VLM training. We demonstrate this with a “blind” LLM that, prompted only with the box coordinates, can perform surprisingly well. To further analyze this shortcut, we propose point-based grounding prompting - querying models with a single 3D point instead of a box. Under this protocol, the accuracy of the blind LLM drops from 36.4% to 18.0% on ScanNet’s Nr3D, and similarly degrades for 3D VLMs (e.g. LL3DA). Building on this insight, we introduce P ^3 -VLM, a novel tri-pathway 3D VLM that supports point-based prompts and achieves 52.1% state-of-the-art accuracy on location-grounded captioning under this setting. To achieve this, we redesign the architecture while keeping it detector-free, scene-centric, and suitable for 3D GS inputs. P ^3 -VLM utilizes three complementary pathways (location-, task-, and context-aware), enabling strong reasoning about scenes without any bounding box dependencies. The key benefit of point-based learning is a consistent improvement across settings: P ^3 -VLM achieves state-of-the-art performance under point-based grounding while also improving robustness, OOD generalization, and performance on ungrounded scene-centric 3D VQA benchmarks, reflecting its reduced reliance on bounding-box shortcuts and resulting stronger visual-spatial reasoning. Comments: NeurIPS 2026, Spotlight Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.28018 [cs.CV] (or arXiv:2610.28018v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.28018 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-76] AdaptMIL: Domain Adaptation for Multiple Instance Learning in Digital Pathology
链接: https://arxiv.org/abs/2610.28007
作者: Ben Vardi,Dana Schonberger,Yuval Friedmann,Zohar Yakhini,Iris Barshack,Alexander Loebel,Ariel Shamir
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Pathology foundation models (PFMs) support a wide range of histopathology tasks on whole-slide images through multiple instance learning (MIL). Despite recent progress, these models remain sensitive to cross-hospital distribution shifts: performance often degrades when a task head (e.g., classification) is trained on data from one hospital and then used on data from another hospital. To address this challenge, we propose AdaptMIL, a training strategy for integrating domain adaptation into MIL training. AdaptMIL is motivated by our finding that directly applying domain adaptation objectives during MIL training can interfere with MIL aggregation learning, often resulting in performance degradation over standard training. AdaptMIL resolves this by decoupling representation-level domain alignment learning from MIL aggregation learning through a two-stage training procedure. We show that AdaptMIL improves cross-hospital performance and outperforms alternative adaptation methods across experiments spanning multiple WSI datasets, tasks, PFMs, MIL architectures, and domain adaptation objectives. AdaptMIL’s code will be released upon publication.
[CV-77] Differentiable Task-Driven Coded Aperture Camera Design with 3D Gaussian Scene Representations
链接: https://arxiv.org/abs/2610.28004
作者: Chengyang Yan,Donald G. Dansereau
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Camera design heavily influences perception task performance, yet cameras are typically designed by domain experts through costly experimentation, in isolation from the tasks they serve. Existing joint-design methods either rely on differentiable simulation limited to pre-captured 2D images, restricting them to optics or image-processing pipeline design, or use non-differentiable camera simulators optimised with derivative-free or reinforcement-learning methods that scale poorly with the number of design parameters. We present a differentiable framework for jointly designing complex camera systems and downstream perception tasks. We use it to design two coded-aperture depth-from-defocus cameras, a monocular and a stereo configuration, by jointly optimising the coded aperture mask, camera parameters, and a depth estimation network end-to-end. Our framework represents scenes with 3D Gaussian Splatting reconstructed from real-world captures, providing a differentiable image formation process for gradient-based optimisation over design spaces intractable for prior differentiable approaches. Results demonstrate that our differentiable method improves over existing coded aperture design and non-differentiable camera design baselines, offering a step forward in automated and principled camera design for autonomous systems.
[CV-78] Matrix-Game 4.0: Harnessing Video Models for Real-Time Playable Game Worlds
链接: https://arxiv.org/abs/2610.27999
作者: Jihai Zhang,Runjia Qian,Kaiqin Kong,Biao Jiang,Puyi Wang,Siru Chen,Tianle Li,Guanjie Chen,Junyi Wu,Chubin Chen,Yuanhang Li,Kaichen Huang,Haobo Zhang,Yangguang Li,Wei Yu,Guibin Chen,Yu Cheng,Yahui Zhou
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: Project page, videos, code and models: this https URL
Abstract:Moving from controllable videos to genuinely playable worlds requires persistent world state, executable game rules, and low-latency interaction. We outline a roadmap for the next generation of Matrix-Game that separates game execution from visual generation through an executable game harness, and introduce Matrix-Game 4.0 as the first system built on this design. A lightweight game engine maintains world state and executes interaction rules, a multimodal coding agent translates player ideas into game code and dynamic scene captions, and a renderer based on autoregressive video model transforms engine-rendered low-poly proxies into detailed visuals, conditioned on dynamic captions and memory references. This framework provides continuous guidance for visual generation and enables natural-language creation and modification of scenes and gameplay. Our renderer is trained on aligned low-poly and high-fidelity video pairs and distilled into a causal generator that generates in four-frame chunks with one or two denoising steps. Together with targeted inference optimizations, Matrix-Game 4.0 enables 480p, 60 FPS gameplay on a single NVIDIA RTX5090 with a key-press-to-display latency of 112ms. Single-step generation achieves 83FPS steady-state throughput with 18.2GiB of peak GPU memory usage. With fixed prompts, steady-state throughput also reaches 67FPS on an RTX5080 and 58FPS on an RTX5070Ti. We demonstrate customizable shooter and driving games with engine-enforced mechanics, and release code, model weights, and low-poly game assets to support research and creation of playable game worlds.
[CV-79] Recast the Past: Stable Long Video Generation with Full Context Memory
链接: https://arxiv.org/abs/2610.27986
作者: Yukuo Ma,Cong Liu,Junke Wang,Junqi Liu,Haibin Huang,Zuxuan Wu,Xuelong Li
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Code: this https URL
Abstract:Long-horizon video generation requires retaining an ever-expanding history, but conditioning on its growing token sequence is costly. Existing methods restrict or compress historical context, potentially losing information needed for future generation. This work proposes ReCam, which groups local spatiotemporal blocks of historical latents into vectors and projects them directly into the high-dimensional hidden space of a Diffusion Transformer. The model learns to use the resulting memory tokens through native self-attention, reducing token count while retaining block-level access to the full history. To mitigate temporal drift during autoregressive rollout, we further introduce dual-pointer denoising with separate noise schedules for recent context and target frames. Among the evaluated methods, ReCam achieves the highest VBench-Long overall score of 80.1 and competitive static-scene revisitation performance on R2M-Bench. It also leads on all four consistency metrics for off-screen subject reappearance, including clothing consistency of 4.09 versus 3.03 for the best baseline.
[CV-80] Beyond the Frame: Redundancy Not Trajectory Carries the Privacy Leak in Video Unlearning
链接: https://arxiv.org/abs/2610.27982
作者: Kavyansh Tyagi,Paolo Russo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Class unlearning for video segmentation aims to ensure that a removed category cannot be reconstructed from a deployed model’s outputs. Current leakage evaluation algorithms inspect one frame at a time, and we show that this understates leakage by a wide margin: an attacker that averages a model’s output field over a two-second clip recovers the forgotten class from every method we evaluate on VSPW and most on Virtual KITTI~2, including models whose per-frame predictions are visually indistinguishable from a retrain oracle (a model trained from scratch without the forgotten class). We attribute the recoverable signal to cross-frame redundancy rather than object trajectory, establish this with a time-shuffled control that leaves leakage intact, and introduce PACE, an unlearning method that suppresses redundancy directly using an exactly permutation-invariant pooled adversarial discriminator and a two-sided temporal consistency objective. Across VSPW and Virtual KITTI~2, PACE lowers pooled attacker AP to 18.79 (\textttperson) and 12.10 (\textttcar) on VSPW and 9.11 (\textttcar) on Virtual KITTI~2, achieving on the latter the strongest suppression among verdict-safe methods retaining usable utility, while retaining 19.8 mIoU on VSPW \textttcar against at most 8.3 for any baseline achieving comparable erasure.
[CV-81] Media Intelligence Platform: Multimodal Detection of Synthetic and Manipulated Content ECCV2026
链接: https://arxiv.org/abs/2610.27974
作者: David Fischinger,Marcel Hasenbalg,Silvia Poletti,Davide Antonutti,Mina Schütz,Lam Pham,Justin Ilyes,Péter Rjabcsenkó,Simon Freitter,Martin Boyer
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: ECCV 2026, Workshop on AI for Multimedia Forensics Disinformation Detection (AI4MFDD2026)
Abstract:The rapid progress of generative artificial intelligence has led to a significant increase in manipulated multimedia content, including synthetic faces, forged videos, AI-generated images, fake audio, and locally edited images. Existing detection approaches are typically designed for a single manipulation type, limiting their applicability in real-world scenarios involving heterogeneous media sources. In this work, we present the Media Intelligence Platform (MIP), a unified web-based framework for multimodal media forensics integrating specialized modules for fake video detection, face morphing detection, AI-generated image detection, audio deepfake detection, localized image manipulation detection, and text analysis. Its modular architecture enables flexible integration of heterogeneous models and supports future extensions for emerging manipulation techniques. Experimental results on representative benchmark datasets demonstrate reliable detection performance across multiple media modalities while providing a practical and user-friendly interface for forensic investigations. The platform highlights the potential of unified multimodal systems for comprehensive deepfake detection in real-world environments, and is accompanied by a public demonstrator released with this workshop submission to support reproducibility and community evaluation.
[CV-82] Efficient SAM3 Adaptation for Multi-Class Semantic Segmentation via Dense Competitive Representations NEURIPS2026
链接: https://arxiv.org/abs/2610.27960
作者: Wenbin Liao,Hao Zhu,Yike Ma,Hao Jiang,Feng Dai
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by NeurIPS 2026. Code is available at this https URL
Abstract:The Segment Anything Model 3 (SAM3) has achieved significant progress in Promptable Concept Segmentation (PCS) by processing short noun phrases to generate segmentation masks with unique instance identifiers. While SAM3 excels at single-concept segmentation, it falls short in complex, real-world environments where multiple concepts must be segmented simultaneously. In particular, SAM3 typically fails to simultaneously process concepts with semantic overlap, leading to semantic misclassifications and redundant mask predictions. Moreover, as the number of concepts increases, the independent inference paradigm introduces prohibitive inference latency. To address these limitations, we propose the Dense Competitive Representations SAM (DCR-SAM) framework. By decoupling cross-modal interactions, the proposed method alleviates the inference latency. To resolve semantic misclassifications, DCR-SAM dynamically aligns textual concepts with visual representations and incorporates a dense competition head to enforce explicit inter-class competition. Furthermore, a learnable background token is applied to absorb non-target objects dynamically. These mechanisms effectively suppress overlapping predictions and generate high-fidelity masks. Extensive experimental evaluations across multiple benchmarks demonstrate that the proposed framework: (1) achieves a 2.3% \sim 10.7% performance improvement over state-of-the-art methods, (2) delivers a 33 \times inference speedup compared to the native SAM3, and (3) exhibits strong robustness across diverse scenarios.
[CV-83] From Proper Scoring Rules and Edge Biases to Mumford-Shah Models: Robust Unsupervised Hyperspectral Image Segmentation
链接: https://arxiv.org/abs/2610.27958
作者: Fabian Conrad,Benjamin Berkels
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Hyperspectral images provide a rich representation of the underlying spectrum for each pixel, allowing for a pixel-wise classification/segmentation into different classes. As the availability of labeled hyperspectral images is limited and the acquisition of labeled training data is time-consuming, unsupervised segmentation methods are important in hyperspectral image segmentation. Spectral variability, mixed pixels, noise as well as high correlation between neighboring spectral bands make this task challenging. We present a feature-agnostic, unsupervised hyperspectral segmentation framework based on the piecewise-constant Mumford-Shah functional. Our framework turns a strictly proper scoring rule and a family of parametric probability distributions into an indicator function for the Mumford-Shah functional, and we naturally generalize some previous Mumford-Shah based models for image segmentation this way. Our focus will be on heavy-tailed elliptical distributions, the negative log-likelihood score and the score based on the beta-divergence to deal with the heavy outliers in hyperspectral imaging. Using the interplay between the calculus of variations and statistics, we also propose a robustness step to estimate the statistical parameters of these heavy-tailed elliptical distributions. In addition, an edge bias, which contains prior boundary information, is incorporated into both the data fidelity term and the perimeter term of the Mumford-Shah functional. Numerical experiments show that, to the best of our knowledge, our method achieves the highest overall accuracy among fully unsupervised methods on three of the four standard benchmarks (Indian Pines, Kennedy Space Center, Pavia University and Salinas) evaluated on the full scenes. It also improves upon our previous Mumford-Shah based model.
[CV-84] Edit in Sync: Unified End-to-End Audio-Visual Editing for Human-Centric Videos
链接: https://arxiv.org/abs/2610.27947
作者: Ziye Yuan,Minghong Cai,Xu He,Ruchang Yao,Xiao Fu,Kaixin Ding,Ming Zhang,Xiangyu Du,Tianpeng Lan,Da Xie,Xiangyu Yue
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages, 5 figures, 4 tables
Abstract:Human-centric audio-visual editing requires precise alignment of lip movements, gestures, and expressions with speech. Existing methods process inputs through independent video and audio pipelines, preventing cross-modal interaction between visual editing and audio dubbing. Many of these pipelines further rely on mask-based inpainting. As a result, they suffer from poor lip-audio synchronization, cumulative pipeline errors, and an inherent mask-size dilemma. We present AVE, the first unified end-to-end framework that jointly performs reference-guided person swapping, insertion, and deletion across both audio and video modalities without requiring spatial masks. To overcome data scarcity, we design a synthesis pipeline and introduce AVE-Dataset, the first audio-visual editing dataset to provide 100K pairs with appearance and voice references. By combining Synthetic-to-Real reconstruction with preliminary audio warm-up, our framework transcends synthetic pipeline limits and mitigates audio-video token imbalance. AVE outperforms modality-separate audio-visual editing pipelines in editing realism, temporal consistency, and audio-visual alignment.
[CV-85] Decoupling Action from Egocentric Observation for World Simulation NEURIPS2026
链接: https://arxiv.org/abs/2610.27946
作者: Yue Ma,Pengjie Song,Xinyu Wang,Yi He,Zeqian Long,Fangneng Zhan,Kaichen Zhou,Hongyu Liu,Hongfa Wang,Peihao Li,Haoyang Huang,Nan Duan,Qifeng Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by NeurIPS 2026, Project page: this https URL , Code: this https URL
Abstract:Egocentric action transfer aims to decouple the semantic action from egocentric observations and reproduce them in new visual contexts, enabling scalable world simulation, embodied policy learning, and robotic data generation. Existing approaches to transferring actions rely on condition-guided video generation, which converts the observation video into explicit geometric controls (e.g., hand poses or meshes) to drive synthesis. However, these methods introduce geometric estimation errors, making it difficult to preserve physically plausible interactions when transferred to new scenes. Alternatively, motion transfer methods directly extract motion patterns from the observation video, yet motion-level features alone cannot encode rich contact dynamics, frequently leading to severe hand structural collapse and implausible contact layout. To address both limitations, we present EgoACT, a test-time framework for egocentric action transfer that operates directly on the denoising process of video diffusion models without requiring additional geometric estimators. EgoACT uses Velocity-guided Structure Anchoring to stabilize reference-consistent hand-object structure in the early denoising stage, and Sparse Correspondence Calibration to refine reliable local correspondences in the mid-to-late denoising stages. Together, these two components preserve transferable action semantics while improving temporal coherence and interaction realism. We further establish EgoActionBench, a benchmark for evaluating action preservation, visual quality, and hand-object plausibility across diverse egocentric manipulation scenarios. Experiments show that EgoACT generates more coherent and physically plausible interaction videos than strong baselines.
[CV-86] In PRAISE of Attributes: Grounding Retrieval Geometry with Query-Conditioned Subspaces for Person Re-Identification
链接: https://arxiv.org/abs/2610.27944
作者: Rajarshi Bhattacharya,Aryan Shukla,Shakeeb Murtaza,Christian Desrosiers,Jose Dolz,Maguelonne Heritier,Eric Granger
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Person re-identification (ReID) ranks an image gallery using either an image query or a textual query. In image-to-image (I2I) ReID, viewpoint, illumination, occlusion, and background changes can make distractors appear more similar than the true match. In text-to-image (T2I) ReID, incomplete or ambiguous descriptions may omit key visual evidence. We study this problem for vision-language model (VLM) based ReID, where attributes such as clothing color and carried objects provide useful fine-grained matching cues. However, directly using attribute-prompt similarities as retrieval scores is unreliable because templated prompts can be biased, correlated, and poorly calibrated. We therefore propose \textbfPRompt-guided \textbfAttribute-based \textbfInference-time \textbfSimilarity \textbfEnhancement (\textttPRAISE), a query-side test-time adaptation framework for VLM-based ReID. Instead of treating prompts as scores, \textttPRAISE uses a transferable attribute-prompt bank to construct an \emphAttribute-Conditioned Subspace (ACS): a query-dependent low-rank subspace induced by the most compatible attribute prompts. ACS provides a controlled attribute-conditioned geometry for reshaping query–gallery affinities and guiding lightweight inference-time updates without identity or attribute labels. Experiments across multiple I2I- and T2I-ReID benchmarks show that \textttPRAISE consistently improves VLM retrieval backbones and outperforms both generic and ReID-specific TTA baselines.
[CV-87] World Models with Configurable Perception
链接: https://arxiv.org/abs/2610.27914
作者: Yijun Shen,Delong Chen,Xianming Hu,Kai Zhang,Pascale Fung
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Real-world observations are noisy and often contain excessive task-irrelevant information. Effective world modeling and planning need to abstract observations into world state representations while filtering out irrelevant information. However, encoders in existing world models are typically non-adaptive, leading to state representations that are easily distracted by task-irrelevant information. As a result, the model may not selectively filter observations to cater to different task demands. To address this problem, we introduce ConfigWM, which realizes configurable perception through task-conditioned state extraction. Given a language description of the task at hand, ConfigWM uses it as a configuration signal to steer how visual embeddings are mapped into a task-specific state space, selectively preserving information relevant to the current task while suppressing task-irrelevant variations. Action-conditioned dynamics are then learned in this task-specific state space for planning. We evaluate ConfigWM in environments with increasing levels of distraction. While existing methods exhibit substantial declines in planning success rate as the environments become increasingly noisy, ConfigWM remains comparatively stable and outperforms the strongest competing method by 22.3% on average across diverse environments. Consistent gains in physical-quantity probing (+0.365 Pearson r) and distracted-to-clean retrieval (+75.5%) further indicate that ConfigWM learns states that preserve task-relevant information while remaining robust to distractors, supporting reliable world modeling and planning in noisy worlds. Project Page: this https URL
[CV-88] ICIBS: Increasing the Context for Ice Boundary Segmentation
链接: https://arxiv.org/abs/2610.27912
作者: Marcel Dreier,Dakota Pyles,Nora Gourmelon,Moritz Koch,Fei Wu,Thorsten Seehaus,Matthias Braun,Andreas Maier,Vincent Christlein
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Mountain glaciers and ice sheets are undergoing rapid loss of ice mass. To assess their current volume and ice-thickness distribution, and to accurately forecast their future, ground-penetrating radar systems are employed to measure ice thickness. These instruments produce radio-echo sounding data that reveal the internal structure of the ice and its boundaries at the surface and beneath the ice body. However, delineating these interfaces is time-consuming as the surveyed area can span several hundred kilometers. To accelerate the labelling process, automated methods have emerged, but they often struggle to delineate the ice bottom, particularly in temperate glaciers. The high water content inherent in temperate ice increases both scattering and attenuation of the radar signal, resulting in weak and ambiguous reflections at the ice bottom and generally noisier data. As a result, the visual features corresponding to the ice bottom become ambiguous. To address this issue, we propose several extensions to deep learning models that incorporate additional contextual information into the prediction. Specifically, we integrate modifications that incorporate additional spatial and geo-spatial contextual information into the network. With these enhancements, we achieve a new state-of-the-art performance on the IceAnatomy benchmark dataset, with mean errors of 10.9 m for the FAU subset and 43.4 m for the CReSIS subset in a single combined system. We also demonstrate that our method generalises well to previously unseen data, outperforming the baseline by 49.6 % on new campaign data from Patagonia. The model code and the new dataset are available at this https URL and this https URL.
[CV-89] EndoWorld-JEPA: Motion-Guided Global-Local Latent World Modeling for Endoscopic Video Representation Learning
链接: https://arxiv.org/abs/2610.27890
作者: Jie Wang,Shilong Yang,Tao Zhou,Juan Yu,Rong Qu,Jianfeng Ren,Zhicheng Zhang,Linlin Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Endoscopic video contains state transitions induced by camera displacement, deformable tissue, and instrument interaction, yet existing pre-training methods treat dynamics implicitly. Inspired by action-conditioned joint-embedding predictive world models and by the observation that endoscopic optical flow contains both shared image-plane displacement and localized residual motion, we reformulate passive endoscopic video pre-training as causal motion-guided latent-state prediction. We present EndoWorld-JEPA, to our knowledge the first latent world-model pre-training framework designed for general-purpose endoscopic video representation learning. In place of recorded actions, calibrated poses, or robot kinematics, the model conditions prediction of the next latent state on motion observed up to the current state. Its Optical Motion-guided module separates dominant displacement from residual motion and converts the resulting evidence into an ordered temporal motion condition and spatial local-motion weights. Complementary global and motion-local predictors use these outputs: the temporal condition modulates both latent transitions, whereas the local weights define an auxiliary motion-local state that extends prediction from scene-level evolution to spatially selective dynamics. Patch-wise motion weighting emphasizes localized change while retaining anatomical context, and past-motion-window dropout randomly withholds the complete motion condition to reduce dependence on unreliable motion. This global–local formulation learns predictable visual–motion dynamics in representation space without reconstructing stochastic appearance artifacts. Evaluation shows that a common pre-trained checkpoint achieves strong transfer across four downstream benchmarks, supporting motion-guided global–local latent prediction as an effective pre-training paradigm for endoscopic video.
[CV-90] SGLDet: Spectral-aware Global-Local Dynamic Modeling for Multispectral Aerial Object Detection
链接: https://arxiv.org/abs/2610.27883
作者: Peng Zhang,Shuaihao Han,Tingfa Xu,Jianan Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages
Abstract:Multispectral aerial object detection aims to exploit spectral-spatial cues to address object detection challenges in complex aerial scenes. Existing methods mainly focus on spectral-spatial decoupling and object-aware spectral-spatial fusion to improve spectral utilization. However, two critical issues remain underexplored: 1) severe foreground-background imbalance caused by large scene coverage hinders reliable object recognition; 2) aerial objects exhibit substantial local variations in scale, orientation, and structure, which hinder accurate localization. To address these issues, we propose SGLDet, a spectral-aware global-local dynamic modeling framework for multispectral aerial object detection, which is motivated by the observation that object semantics are jointly determined by global spectral consistency and scene context, while precise localization requires dynamic feature modulation conditioned on the input spectral and spatial properties. To realize these properties, SGLDet consists of three coordinated components: a Spectral-driven Global Dynamic Modeling module that aggregates spectrally related features across the scene via similarity-guided sparse matching; a Multi-scale Global Context Integration module that constructs a compact scene memory for scene-guided contextual reasoning; and a Spectral-aware Input-dependent Calibration Aggregation module that performs expert-kernel aggregation with scene-guided spectral-spatial calibration to enable adaptive local feature modulation. Extensive experiments on MODA, HOD3K, and DroneVehicle demonstrate that SGLDet achieves superior performance with competitive computational cost. Code is available at this https URL.
[CV-91] S3E-Det: A Strict Spatial-Spectral Equivariant Detector for Multispectral Object Detection
链接: https://arxiv.org/abs/2610.27882
作者: Peng Zhang,Tingfa Xu,Jinrui Sui,Yuyang Tian,Jianan Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 12 pages
Abstract:Multispectral object detection requires detectors to exploit material-dependent spectral responses while maintaining consistent localization under geometric variations. Existing MOD detectors learn discriminative spectral representations, yet lack a unified mechanism that preserves spectral order, adapts spectral-context fusion across pyramid levels, and regulates geometric responses under rotation. To address these limitations, we propose S ^3 E-Det, a strict spatial-spectral equivariant detector that couples ordered spectral modeling with transformation-consistent prediction. Spectral Order-aware Sampling Calibration (SOSC) imposes local order constraints on input-derived spectral features through first-order variation, while Common-Differential Equivariant Fusion (CDEF) adapts spectral-context aggregation across pyramid levels. A compact Geometric Alignment Head (GAH) reconciles invariant and equivariant output properties within a unified prediction stage. Across MODA, HOD3K, and DroneVehicle, S ^3 E-Det combines high detection accuracy with a compact parameter footprint and reduced rotation-induced variation. Code is available at this https URL.
[CV-92] Gaze2HOI: Gaze-Aware Intention Disambiguation for Egocentric Hand-Object Interaction Forecasting
链接: https://arxiv.org/abs/2610.27877
作者: Chuanyang Ma,Chenye Wang,Yixuan Wang,Huiling Zhou,Xingqun Qi,Muyi Sun,Man Zhang,Caifeng Shan,Zhenan Sun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages, 8 figures. Project page: this https URL
Abstract:Anticipating how users will interact with their surroundings is essential for proactive systems that provide timely assistance in AR, VR, and embodied applications. Prior work typically forecasts hand-object interactions directly from historical observations, leaving intention and contact geometry implicit. However, similar observation histories can precede semantically or geometrically distinct futures. Inspired by behavioral evidence that gaze can provide early cues for underlying intention, we propose Gaze2HOI, a gaze-aware framework that formulates hand-object interaction forecasting as a hierarchical problem comprising intention inference, contact grounding, and coupled hand-object motion generation. Specifically, Gaze2HOI first weights gaze cues by fixation strength and recency, and uses them to modulate the visual representations of a vision-language model (VLM) through Temporal Gaze Saliency Modulation (TGSM), thereby reducing semantic ambiguity in intention inference. Subsequently, an Anticipatory Contact Grounding (ACG) module maps these structured intentions into explicit contact, support, and placement fields, effectively mitigating geometric ambiguity. Driven by these explicit conditions, a diffusion-based model jointly generates the 6-DoF object trajectories and future articulated motion of both hands. To evaluate Gaze2HOI, we further introduce EgoAIM, a phase-aware benchmark that organizes egocentric hand-object interactions into four phases defined by evolving hand-object relations and provides fine-grained intention annotations. Extensive experiments demonstrate that Gaze2HOI achieves state-of-the-art performance in both intention inference and motion forecasting, even approaching oracle-level end-to-end performance in specific interaction phases. Project page: this https URL.
[CV-93] Manifold Prior Guided Deep Unfolding for Hyperspectral Image Reconstruction
链接: https://arxiv.org/abs/2610.27874
作者: Shuqi Liu,Yingkai Zhang,Lin Gu,Tao Zhang,Ying Fu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Hyperspectral Image (HSI) reconstruction has made substantial progress with the deep unfolding framework by decomposing the problem into a projection module and a denoiser module. Nevertheless, existing methods still exhibit limitations of insufficient matching with HSI data. The issues lie in two aspects: 1) the data-consistency projection module applies a fixed gradient descent step while ignoring the spectral content adaptivity of HSI; 2) the denoiser module either suffers from high computational complexity or is limited to the local receptive field, failing to efficiently capture the global spatial-spectral dependencies. In this paper, we propose a deep unfolding framework centered on a two-stage manifold learning strategy to obtain a degradation-free structural prior, which is then strategically injected into both the projection and denoiser modules. In the data-consistency projection, a prior-guided spectral weight mechanism is introduced to adaptively modulate the weight of spectral band. In the denoiser module, a manifold-guided Mamba is proposed to achieve the structure-aware extraction of spatial-spectral features through prior modulation. Quantitative and visual comparisons on synthetic and real-world datasets show that the proposed method significantly outperforms existing state-of-the-art approaches, while maintaining the structural interpretability of the unfolding framework. The code is available at this https URL.
[CV-94] LaLa-World: Scaling Latent World Models with Natural Language Actions
链接: https://arxiv.org/abs/2610.27851
作者: Junha Song,Geonmo Gu,Jeongeun Park,Heejeong Nam,Lucas Maes,Byeongho Heo,Dongyoon Han,Sangdoo Yun
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Project page: this https URL Code: this https URL
Abstract:This paper shows that JEPA-style latent world models (LWMs) can learn across diverse domains and data sources by adopting the language-conditioned interface of video generation models. We establish this approach as an effective and efficient basis for planning with imagination and downstream policy fine-tuning. To this end, we introduce LaLa-World, a Language-conditioned Latent World model that learns from simulation and robot trajectories together with broader image and video data. Two designs make this work. First, we describe both numerical actions and visual changes in language, enabling joint training of a single future predictor across heterogeneous datasets. Second, we build the predictor on a language model, allowing it to interpret language conditions while predicting future latent states. The resulting predictor achieves higher average MPC planning performance than existing LWMs trained separately for each domain or task. Adding image and video data further improves average planning performance and robustness to action phrasing. In a matched comparison with an RAEv2-based video generation model baseline, LaLa-World reaches comparable planning performance with 6\times less training cost and 13.9\times faster planning. Beyond planning, the pretrained LaLa-World serves as an initialization for LaLa-Policy, which jointly predicts actions and future visual representations. LaLa-Policy performs competitively with Cosmos Policy on LIBERO and RoboCasa, with up to 5.5\times faster inference.
[CV-95] From Generic to Dedicated: A Novel Optimizer for Online Continual Learning
链接: https://arxiv.org/abs/2610.27835
作者: Yongyi Wu,Zheng Wang,Sen Lin
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, including appendix, conference paper
Abstract:Online Continual Learning (OCL) requires models to learn from non-stationary data streams under a strict single-pass constraint, making them highly susceptible to catastrophic forgetting. While existing studies have explored various strategies from data, architecture, and optimization perspectives, the optimizer often directly uses generic approaches such as SGD and AdamW. In this work, we reveal a new link between optimization dynamics and OCL needs by recasting Polyak-Ruppert averaging as the engine of “plasticity” and Primal averaging as the anchor of “stability”. This inspires our novel optimizer, SPIN (Stability-Plasticity INterpolation), which explicitly decouples plasticity and stability by interpolating between a fast-moving “plasticity” sequence and an adaptive “stability” sequence. We theoretically demonstrate that SPIN implicitly employs an inverse Hessian approximation, providing crucial Tikhonov regularization to damp noisy, single-sample Hessian estimates and mitigate forgetting. More importantly, SPIN can be seamlessly integrated into existing OCL methods by taking resulted gradients as input and replacing standard optimizers. Extensive experiments show that using SPIN instead of standard optimizers yields consistent and significant performance gains across multiple OCL approaches, including the challenging rehearsal-free setting and under strict memory constraints.
[CV-96] Decoupled Representation Learning for Face Presentation Attack Detection under Smart Glass Capture
链接: https://arxiv.org/abs/2610.27800
作者: Raghavendra Ramachandra,Narayan Vetrekar,Neha Naik,Rajendra Gad
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted in IJCB 2026
Abstract:Face presentation attack detection is typically studied under fixed camera or handheld acquisition, whereas smart glass capture introduces a more challenging setting with ego motion, changing viewpoint, background clutter, distance variation, and less controlled illumination. To address this challenge, we present \emphSmaGla-FPAD, a smart glass based face presentation attack database acquired using Ray-Ban Meta Wayfarer Smart Glasses across indoor, outdoor, and in-car settings. The database contains 297,000 sample images with both bona fide and silicone mask based attacks. We further propose \emphDeCoupPAD-FAS++, a decoupled representation learning framework that separates spoof related evidence from nuisance sensitive capture variation through complementary attack and style branches, pseudo-domain supervision, adversarial suppression, and prompt guided semantic alignment. Experiments on SmaGla-FPAD show that the proposed method achieves the best balance between attack rejection and bona fide acceptance in both in domain and cross scenario evaluation. Ablation results further confirm the contribution of disentanglement, pseudo-domain supervision, and prompt based regularisation under smart glass capture.
[CV-97] RoboJudge: Multimodal Language Models as Judges for Embodied Video Generation
链接: https://arxiv.org/abs/2610.27782
作者: Siyuan Ma,Yu Qi,Cheng Ye,Ruotian Peng,Xinyi Xu,Zitiantao Lin,Zedong Cai,Yijian Huang,Weiyang Liu,Siyuan Wang,Yandong Wen
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: Project page: this https URL
Abstract:Video generation models are promising foundations for embodied world models, but only if their generated rollouts are physically valid and faithful to task instructions. Multimodal large language models (MLLMs) are increasingly used to judge these properties, yet their agreement with human judgments on embodied videos has not been systematically evaluated. To address this gap, we introduce RoboJudge-Bench, a benchmark for evaluating MLLMs as judges of embodied video generation. The benchmark measures two fundamental dimensions, Physical Adherence and Instruction Alignment, across 800 human-annotated videos from 8 source corpora and 10 video generators, spanning grippers, human hands, and dexterous hands. Experiments reveal uneven alignment between existing judges and human ratings: general-purpose MLLMs show lower agreement on physical adherence, while specialized video judges designed for general-domain physical reasoning do not consistently transfer to embodied scenarios. Motivated by these findings, we construct RoboJudge-Train, comprising 12K videos annotated with 25K human-verified scores and model-assisted rationales for Physical Adherence and Instruction Alignment. We fine-tune a 9B MLLM on this dataset with supervised learning followed by reinforcement learning, yielding RoboJudge-9B. Among the model judges evaluated on RoboJudge-Bench, RoboJudge-9B achieves the highest overall agreement, reaching a Pearson correlation of 0.719 with a statistically significant 0.089 gain over Gemini-3.7-Flash, the best-performing proprietary baseline in our evaluation. Training on RoboJudge-Train also yields gains on external benchmarks, including VideoPhy2 and Physion-Eval. We further use RoboJudge-9B as a downstream evaluator to benchmark video generation models according to Physical Adherence and Instruction Alignment.
[CV-98] Geometry Regularization for Vanilla NeRF via Self-Supervised Cross-Ray Consistency
链接: https://arxiv.org/abs/2610.27776
作者: Zongji Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:Neural Radiance Fields (NeRF) achieve high-quality novel-view synthesis by representing a scene as a continuous volumetric radiance field. However, the geometry encoded by the density field of vanilla NeRF can be noisy and inaccurate, since different combinations of shape and radiance may explain the same input images. To reduce this ambiguity, we augment vanilla NeRF with a geometry-optimization branch based on self-supervised cross-ray consistency. Specifically, this branch samples candidate surface points from the density field, generates virtual rays that intersect at each point from different directions, and renders their appearance using the NeRF network. Based on the visibility-weighted color consistency across these rays, we constrain the implicit geometry. We further regularize the implicit geometry by penalizing opacity at candidates with inconsistent appearance predictions and encouraging local normal smoothness. These geometric constraints are derived from the NeRF predictions and require no ground-truth depth or surface-normal supervision. Experiments on eight Replica scenes show that the proposed method reduces the mean Chamfer distance by 30.8% compared with vanilla NeRF.
[CV-99] Radar-Based 3D Human Pose Estimation via Motion-Aware Pose Initialization and Observation-Guided Refinement
链接: https://arxiv.org/abs/2610.27774
作者: Woo-Jin Jung,Chan-Bin Lim,Seung-Hyun Kong
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 15 pages, 5 figures, 8 tables
Abstract:3D Human Pose Estimation (3D HPE) is used in indoor environments, such as smart homes, to analyze human movement and monitor human states. Millimeter-wave (mmWave) radar is suitable for this task because it is robust to lighting variations and does not capture images of human appearance. However, radar observations often provide insufficient information about certain body parts or contain overlapping reflections from multiple body parts, limiting accurate 3D joint localization. To address these limitations, we propose the Initialization-to-Refinement Pose Framework (I2R-Pose), comprising two stages: Motion-Aware Pose Initialization and Observation-Guided Refinement. In the first stage, joint-wise features are learned by incorporating temporal relationships across consecutive radar observations and human skeletal structure to compensate for insufficient joint information and estimate an initial 3D pose sequence. In the second stage, separately extracted radar observation features predict joint-wise positional corrections that refine the initial pose and generate the final pose sequence. The framework supports individual and fused inputs of time-Doppler/time-range (TD/TR) maps and radar point clouds (PCs). To evaluate performance across indoor activities, we construct the Radar Motion and Spatial Information for 3D Human Pose Estimation (RaMoS-Pose) dataset, providing synchronized TD/TR maps, radar PCs, and 3D joint ground truth (GT). Evaluations on RaMoS-Pose and the publicly available MVDoppler-Pose dataset show that I2R-Pose reduces mean per-joint position error (MPJPE) by 39.50% on MVDoppler-Pose and by 15.13% on RaMoS-Pose with fused inputs, compared with the best-performing existing methods evaluated on each dataset.
[CV-100] How Much Do ell_1 Attacks Need Concentrated Perturbations?
链接: https://arxiv.org/abs/2610.27773
作者: Melika Shirian,Kianoosh Vadaei,Arshia Hemmat
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: 23 pages, 10 figures, 15 tables, including appendices
Abstract:A global \ell_1 budget limits how much an adversary may change an image, not how that change is distributed. We ask how much of an \ell_1 attack’s observed success remains when large coordinate changes are restricted. We add a per-coordinate cap to the budget, enforce the joint constraint by exact in-loop projection, and sweep the cap for APGD- \ell_1 , SLIDE, and FMN- \ell_1 without altering their objectives. Across 18 standard-model settings, caps of at most 32/255 retain a median 97% of attack success relative to each method’s native uncapped baseline, while median peak magnitude falls by 80% and median effective support nearly triples. At matched realized \ell_1 , capped perturbations remain structurally distinct; separate ImageNet comparisons also show lower LPIPS and DISTS at selected caps. Direct in-loop constrained optimization preserves up to 88 percentage points more attack success than post-hoc clipping, highlighting the importance of constraint enforcement. Suitable caps nevertheless vary across attacks and models, and adversarially trained classifiers exhibit more heterogeneous concentration sensitivity. These results establish perturbation concentration as a complementary, model- and attack-dependent axis of \ell_1 attack evaluation.
[CV-101] PixelSR: Pixel-Space Orthogonal Flow Decoupling for Real-World Image Super-Resolution
链接: https://arxiv.org/abs/2610.27759
作者: Xiang Lv,Kai Zhao,Qiang Wang,Qiao Mo,Chao Zhou,Mingwen Shao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 21 pages, 11 figures. Submission under review
Abstract:Latent diffusion models have recently achieved remarkable progress in Real-world Image Super-Resolution (Real-SR). However, existing methods consistently exhibit a persistent seesaw effect: enhancing perceptual realism compromises structural fidelity. This effect, we argue, is not intrinsic to diffusion models, but largely stems from the architectural constraints of prevailing two-stage frameworks. To address this challenge, we propose PixelSR, a novel single-stage Pixel Diffusion Transformer for Real-SR that operates directly in pixel space. Specifically, PixelSR consists of two core designs: Pixel-space Conditional Diffusion and Orthogonal Flow Decoupling. The former establishes an effective pixel-space restoration framework by integrating spatially aligned LR conditioning, visual semantic guidance, and lightweight texture refinement, thereby bypassing VAE compression and enabling end-to-end optimization directly in RGB space. Complementarily, the latter decomposes the asymmetric flow target into a low-rank semantic component and a noise-free orthogonal texture component. By routing these components to specialized branches, it achieves decoupled optimization with reduced mutual interference. Benefiting from these designs, our PixelSR restores perceptually realistic details while faithfully preserving structural content. Extensive experiments across multiple benchmarks demonstrate that PixelSR not only achieves competitive performance, but also effectively mitigates the fidelity–realism seesaw effect.
[CV-102] Revealing Inter-Channel Modality Discriminability: Channel Selective Interaction Fusion Network for Optical-SAR Object Detection
链接: https://arxiv.org/abs/2610.27757
作者: Haojie Li,Yuanfei Huang,Lingfei Song,Jia Li,Lizhi Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注:
Abstract:Synthetic aperture radar (SAR) and optical imagery provide complementary structural and appearance information for remote sensing object detection. However, substantial cross-modal discrepancies and indiscriminate channel interaction inevitably introduce redundant or conflicting features. Since different feature channels encode heterogeneous semantic information and contribute unequally to cross-modal fusion, selectively highlighting task-relevant channels while suppressing redundant ones is essential. To address this issue, we propose the Channel Selective Interaction Fusion Network (CSIF-Net), which performs selective channel-level cross-modal interaction. CSIF-Net first employs a frequency decomposition module to construct frequency-aware SAR inputs, facilitating the separation of structural information from noise-related responses. It then models cross-modal channel relationships through a cross-attention channel selection module and dynamically partitions the features into shared and modality-specific components. Finally, a dual-branch fusion module performs cross-modal interaction and channel realignment for shared features while independently enhancing SAR-specific features, thereby exploiting complementary information and suppressing modality interference. Extensive experiments on the OGSOD 1.0 and M4-SAR datasets demonstrate that CSIF-Net achieves state-of-the-art performance in both horizontal and oriented object detection, surpassing the existing methods by 4.01% and 3.18% gains, respectively.
[CV-103] CILA: Confusion-based Instance-wise Logit Adjustment for Long-Tailed Recognition WACV2027
链接: https://arxiv.org/abs/2610.27728
作者: Jiyoung Lim,Jee-Hyong Lee
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: WACV 2027 Accepted
Abstract:Real-world datasets are inherently class-imbalanced, leading models to favor majority classes and struggle with minority ones. Existing logit adjustment methods address this by modifying logits based on class frequencies, but their class-wise adjustments ignore instance-wise confusion and often suppress informative signals from confusing non-target classes. This results in insufficient learning of such classes, increased prediction uncertainty, and limited generalization. To address this, we propose two complementary methods. First, \textbfConfusion-based Instance-wise Logit Adjustment (CILA) adaptively adjusts logits based on instance-wise confusion between the target and non-target classes, restoring informative gradients that class-wise adjustment suppresses. Second, \textbfBatch-adaptation Noise-Injection (BNI) perturbs the target logit during training, extending the confusion-aware update to correctly classified but near-boundary samples that the penalty alone cannot reach. CILA is a loss-agnostic plug-in module that can be seamlessly integrated into existing long-tailed learning methods, and consistently improves performance across multiple benchmarks. Our code will be available at this https URL. Comments: WACV 2027 Accepted Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.27728 [cs.CV] (or arXiv:2610.27728v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.27728 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-104] World2HOSI: A Multi-Expert Harness for Long-Horizon Human-Object-Scene Interaction Generation
链接: https://arxiv.org/abs/2610.27725
作者: Jiushu Li,Haoran Wang,Yumeng Liu,Zhiyang Dou,Yuan Liu,Yuexin Ma,Cheng Lin
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Project Page: this https URL
Abstract:Long-horizon human-object-scene interaction (HOSI) is central to embodied AI, animation, and game development. Existing human-object interaction (HOI) and human-scene interaction (HSI) models excel in their respective domains, but none alone covers tasks requiring multiple interaction skills. Training unified HOSI models is hindered by scarce and costly data. We introduce World2HOSI, a multi-expert orchestration harness that generates continuous HOSI data from 3D Gaussian Splatting (3DGS) scenes and natural language instructions. Our key insight is that long-horizon HOSI hinges not on a stronger single model but on orchestration. Given a shared, structured scene state and an orchestration harness grounded in it, existing experts can jointly accomplish tasks beyond the reach of any of them alone. World2HOSI realizes this insight as a closed loop around a persistent scene state built from decoupled background and object assets. A harness plans action stages over this state, invokes experts via their native conditioning interfaces, and writes each action’s effects back, so that every invocation is grounded in the evolving world. To reduce physical artifacts, bounded residual optimization and contact-driven physics simulation further ensure the physical plausibility of human motion and object manipulation, respectively, while preserving interaction intent. Experiments and a perceptual user study show that World2HOSI achieves the best task completion and interaction quality on the evaluated HOSI tasks, while generating more diverse data.
[CV-105] Push-Pull Attentional Anchoring for Diffusion Concept Erasure ECCV2026
链接: https://arxiv.org/abs/2610.27719
作者: Nattanat Chatthee,Tagon Sompong,Ekapol Chuangsuwanich,Supasorn Suwajanakorn
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to ECCV 2026 (Spotlight)
Abstract:Rapid advances in diffusion models have raised concerns about privacy, harmful content, and copyright violations. Concept erasure methods aim to address these risks by removing undesirable concepts from pretrained models without full retraining. However, existing approaches often degrade non-target generation quality through heuristic strength scaling (e.g., classifier-free guidance or negative prompt scaling), leading to uncontrolled and excessive semantic drift. To address this, we propose Push-Pull Attentional Anchoring (PPAA), a mechanism in the cross-attention space that displaces target representations while anchoring the scale of erasure via bounded cosine-similarity constraints. By using a relative similarity ratio instead of an absolute difference, our adaptive threshold represents a meaningful percentage of semantic retention, facilitating a single value to be applied uniformly across layers. We conduct extensive experiments across multiple concept categories, including celebrities, artistic styles, nudity, and intellectual property. Our method achieves favorable trade-offs against prior approaches on Stable Diffusion v1.4 and demonstrates its applicability to modern architectures such as FLUX.
[CV-106] RLSplat: Resource-Linear Budget Control for Feed-forward 3D Gaussian Splatting
链接: https://arxiv.org/abs/2610.27712
作者: Injae Kim,Minseok Joo,Jihwan Park,Sehyung Kim,Minseong Bae,Hyunwoo J. Kim
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Code available at this https URL
Abstract:Feed-forward 3D Gaussian Splatting (FF3DGS) reconstructs 3D scenes from images in a single forward pass, making it well-suited to applications that must operate within limited resource budgets. To further improve its practicality, recent budget-controllable FF3DGS methods allow the number of Gaussians to be adjusted at inference, so that the output meets a given storage limit or rendering-latency constraint. However, these methods overlook the cost of inference itself: to produce a representation with the target number of Gaussians, they first predict a dense set of Gaussians and then prune or select from it, requiring nearly constant latency and peak memory regardless of the budget. We therefore propose RLSplat, a resource-linear budget-controllable FF3DGS framework whose representation size, inference latency, and peak memory all decrease linearly as the Gaussian budget is reduced. Instead of pruning or selection, RLSplat learns a Gaussian allocation policy that distributes the budget across local scene regions, so that exactly the requested number of Gaussians is generated. These Gaussians are then refined through resource-linear Gaussian decoding, in which each Gaussian attends only to nearby Gaussians found via an efficient hierarchical neighbor search, so that the decoding cost decreases linearly with fewer Gaussians. In zero-shot evaluation, RLSplat outperforms prior budget-controllable methods at every budget, surpassing the strongest baseline even with 4 \times fewer Gaussians, while running up to 4.1 \times faster with 9.6 \times less peak memory at low budgets.
[CV-107] VEOcc: Learning Persistent Voxel Feature Maps for Embodied Online Semantic Occupancy Mapping and Offline Refinement
链接: https://arxiv.org/abs/2610.27704
作者: Ruoyu Wang,Jiahan Li,Yong Liu,Kexin Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Semantic occupancy mapping supports embodied perception during exploration and detailed scene understanding afterward, giving rise to two complementary requirements: online incremental mapping for maintaining an up-to-date global map under causal constraints, and offline global map refinement for improving the completed scene representation with additional computation. We present VEOcc++, a unified voxel-centric framework that connects both stages through a persistent sparse voxel feature map. During exploration, the map expands with newly observed regions and integrates aligned local features through reliability-guided updates, while context-aware decoding produces the current semantic occupancy map. After exploration, cached local features are consolidated as complementary evidence and combined with the completed online representation for multi-scale spatial refinement. On EmbodiedOcc-ScanNet, VEOcc++ achieves 67.12% IoU and 58.40% mIoU, establishing new state-of-the-art performance for online incremental mapping. With the local predictor fixed, our online mapping pipeline outperforms VEOcc fusion by 3.70 IoU and 3.74 mIoU points, demonstrating substantial gains from the proposed global mapping design. Offline global map refinement further improves the completed map to 67.79% IoU and 59.37% mIoU using only the observations acquired during exploration. Zero-shot deployment on a mobile robot further demonstrates transferability to unseen environments, onboard feasibility, and scalability to large-scale indoor mapping. Project page: this https URL.
[CV-108] VLA-Retain: Benchmarking Cross-Dimension Effects of Vision Tuning in VLAs
链接: https://arxiv.org/abs/2610.27698
作者: Borui Xu,Han Chen,Bingsheng He,Ming Wu,Jason Zeng,Michael Heinrich
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-language-action (VLA) models transfer knowledge from large-scale pretraining into robot control, but their manipulation performance remains sensitive to changes in observation conditions, which recent work addresses by fine-tuning the visual module on target-domain data. However, it remains unclear how such adaptation affects robustness beyond the target condition, as existing robustness benchmarks only evaluate models without fine-tuning and adaptation studies only report the target gain. Measuring this cross-dimension effect requires separating target gains from off-target changes, isolating adaptation effects from architectural differences, ruling out generic forgetting, and comparing tuning choices. Therefore, we introduce VLA-Retain, a benchmark that compares five open VLA models, tuned on Camera-view or Sensor-noise data, with their own base checkpoints on LIBERO-Plus dataset spanning seven perturbation dimensions. Our study reveals four findings. First, targeted vision tuning is not local, and its target gain does not predict retention. Mean off-target gain is negative in 8 of 10 model–domain configurations. Under Camera-view tuning, X-VLA gains 69.4 percentage points on the target but loses 4.6 points on average elsewhere, whereas OpenVLA-OFT gains 39.3 and 1.4 points, respectively. Second, these regressions are not explained only by performance loss on the same tasks without perturbations. Third, broader training coverage raises overall gain over target-specific training but does not eliminate dimension-level regressions. Fourth, tuning strategies are model-dependent, as DELTA-style feature preservation raises X-VLA’s mean off-target gain to +1.2 points but lowers OpenVLA-OFT’s to -23.1 . Overall, these findings show that target-domain success alone is insufficient and highlight cross-dimensional retention as a key criterion for targeted VLA adaptation.
[CV-109] PanoGraphGS: Unifying Panoptic Segmentation and Hierarchical 3D Scene Graphs with Gaussian Splatting
链接: https://arxiv.org/abs/2610.27696
作者: Jinghan Gao,Siyu Wang,Dongyu Feng,Yiyuan Pan,Rong Xiong
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 9 pages, 5 figures
Abstract:3D scene graphs support language-guided navigation by organizing physical entities and their containment and spatial relations into hierarchical maps. Building such structured maps requires a scene representation that consistently distinguishes semantic regions and object instances across viewpoints. To this end, we present PanoGraphGS, a unified framework that integrates panoptic 3D Gaussian representation learning with hierarchical 3D scene graph construction. The framework learns semantic and instance features on shared Gaussian geometry. Prototype-based semantic learning and association-based instance learning produce consistent panoptic predictions across viewpoints. The learned representation supports novel-view panoptic segmentation and direct Gaussian-level panoptic decoding. A semantic-aware scene graph module further recovers a floor-room-stuff-object hierarchy with containment and spatial relations. We further release NaviGS, a real-world dataset for panoramic scene understanding. The dataset provides 3DGS, mesh, RGB, semantic, and depth data. For 3D scene graph identification, PanoGraphGS outperforms the strongest evaluated baseline in object-node accuracy by more than 50% points on average across Replica and NaviGS. PanoGraphGS also achieves state-of-the-art performance among the evaluated methods in 2D panoptic segmentation. These results demonstrate the effectiveness of PanoGraphGS in both hierarchical 3D scene understanding and 2D panoptic segmentation. The project website is available at: this https URL.
[CV-110] A Principled Optimal Transport Framework for Frame Selection in Long Video Understanding NEURIPS2026
链接: https://arxiv.org/abs/2610.27692
作者: Ziwei Zhao,Yuhang Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by NeurIPS 2026
Abstract:Long-video understanding with multimodal LLMs is fundamentally constrained by the mismatch between video length and the model’s limited visual context budget, making frame selection essential for practical reasoning. Existing query-aware methods typically rely on heuristic pipelines: they assign one-dimensional query-conditioned scores to frames and then apply sampling, ranking, or extraction rules to form the final subset. While effective in practice, such methods are generally not derived from an explicit optimization objective and struggle to capture the heterogeneous evidence required for complex reasoning. In this paper, we revisit long-video frame selection from a principled optimization perspective. We formulate it as a structured evidence-allocation problem and propose OTFS, an optimal transport framework that maps multiple evidence sources onto video frames. To reflect the asymmetry of this problem—important evidence should be preserved, whereas redundant frames need not absorb mass—we instantiate this view with a semi-unbalanced entropic optimal transport objective and an efficient solver, OTFS-Sinkhorn. Combined with an exact dynamic program for coverage-aware subset extraction, OTFS yields a practical two-stage training-free pipeline. We further show that, in the single-source case, our formulation induces a Gibbs-type distribution over frames, making Q-Frame’s temperature-scaled distribution a special case of our framework while also providing a broader perspective for understanding related prior methods. Experiments on three long-video understanding benchmarks show that OTFS consistently outperforms strong training-free baselines. Codes are available at this https URL.
[CV-111] NeoWorld-3: From Pixels to Code-Defined Physical Worlds
链接: https://arxiv.org/abs/2610.27690
作者: Eachen Soong,Zijian Niu,Haoran Xu,Yihong Dong,Yunbo Wang,Xiaokang Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Simulation-ready world reconstruction demands physical interactivity beyond visual realism, which existing black-box 3D generation fails to deliver. We present NeoWorld-3, a coding-agent framework that reconstructs interactive 3D worlds from visual observations via tunable, executable programs. At its core, NeoSDK provides the underlying computational engine for low-level geometries, part composition, differentiable parameter refinement, and physics validation. At the same time, NeoMCP exposes these capabilities as structured, tool-callable APIs for vision-language agents to visually inspect, edit code, and iteratively refine programs. Through a hierarchical coordinator-worker architecture, the framework scales seamlessly from standalone articulated objects to complex compositional scenes. Evaluations across object-level and scene-level reconstruction benchmarks demonstrate that NeoWorld-3 achieves state-of-the-art geometric fidelity and physical readiness. Crucially, manipulation trajectories planned within our reconstructed twins achieve successful zero-shot execution on a physical dual-arm robot, completing the Real-Sim-Real loop.
[CV-112] Reusing Assurance Evidence in AI Pipelines: Validity Conditions a Fail-Closed Release Gate and a Replay of 714 Public Data Version Control Pipelines
链接: https://arxiv.org/abs/2610.27672
作者: Atanu Mukhopadhyay,Shravya Reddy Nalla
类目: Computer Vision and Pattern Recognition (cs.CV); Software Engineering (cs.SE)
备注: 12 pages, 4 figures
Abstract:Rerunning every governance control after every change to an artificial intelligence (AI) pipeline is expensive, so teams reuse earlier results. Reuse is safe only if each control’s declared dependencies cover everything that can change its outcome, and only if the release gate checks that reused evidence describes the candidate being released. This article makes both conditions explicit and measures how often the first fails in practice. It defines when a signed evidence record is valid for a candidate and gives sufficient conditions under which change-local re-execution reaches the same release decision as rerunning every control. A reference gate implementing these checks with Ed25519 digital signatures blocks eight of nine manipulation scenarios, against one for a gate that trusts the planner and three for a gate that only verifies signatures; a dependency the policy never declared passes every gate. A replay of 38,083 commits from 714 public Data Version Control (DVC) pipelines measures that failure mode. Reusing stage results according to declared dependencies avoids 83.4% of stage executions, but in 3.3% of commits it reuses a stage whose imported code or referenced configuration changed, and 31% of pipelines are affected. A manual audit confirms the undeclared dependency in 36 of 40 sampled events. Evaluation stages, whose outputs are the evidence, are the most exposed. Adding each stage’s static import closure lowers stale reuse to 0.4% while still avoiding 81.6% of executions; sampled shadow re-execution removes far less, and rerunning everything after unexplained changes avoids only 45.3%.
[CV-113] Pairwise Learning Increases Data Efficiency in Deep Neural Networks
链接: https://arxiv.org/abs/2610.27669
作者: Robert Bain,Marjan Rashidi,Elizabeth Chrastil
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Deep neural networks often fail at small-sample regression, collapsing to low-variance predictions that miss individual differences. Here, we introduce a pairwise comparison framework that reframes regression as aggregated relative rankings. This method improves data efficiency across model architectures and dataset sizes. On the UTKFace dataset (23k images), pairwise learning generates over one billion training comparisons without additional supervision. The paired ranking allows a lightweight ResNet10 trained from scratch on home hardware to achieve a 3.64-year MAE-surpassing previous models, including those with large-scale pretraining. For classification into categories, pairwise supervision consistently outperforms an otherwise identical 10-way crossentropy model on STL10 with 500-2000 labeled examples, improving accuracy by up to 5.2%. Overall, pairwise supervision enhances data efficiency and enables strong regression and classification performance when labeled data are limited, particularly in clinical settings.
[CV-114] Inferring continuous cognitive traits from limited neuroimaging data using pairwise convolutional neural networks
链接: https://arxiv.org/abs/2610.27668
作者: Marjan Rashidi,Robert Bain,Craig Stark,Jeffrey Krichmar,Mary Hegarty,Elizabeth Chrastil
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Individual differences in cognitive abilities vary widely, yet the structural brain features supporting these differences remain poorly understood. Traditional brain-behavior analyses rely on predefined metrics such as volume or cortical thickness and typically assume linear relationships with behavior, limiting sensitivity in small datasets. Here, we introduce a pairwise convolutional ranking framework that learns directly from whole-brain T1-weighted MRI by making comparative rather than absolute predictions, which performs well even with small samples (N=108). We tested this method with four small-scale spatial abilities tasks. Regression analysis predicted performance with high accuracy across all tasks, achieving greater predictive power over standard regression and pointwise deep learning baselines. Iterative cortical-region removal (backward elimination) revealed brain regions supporting individual performance common to all four tasks, as well as task-specific regions. These results demonstrate that pairwise learning extracts meaningful behavioral information from structural MRI, providing prediction and interpretable insight into the neuroanatomy of individual differences.
[CV-115] PLaCA: Learning Patch-Level Future Latent Change for Action Generation
链接: https://arxiv.org/abs/2610.27657
作者: Zhan Yi,Yuchen Wang,Chao Li,Lei Hong,Boyuan Yang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-Language-Action (VLA) models generate robot actions from visual observations and language instructions, but effective manipulation also requires anticipating how the environment will change. Existing world-action models explore future image prediction and latent dynamics modeling. However, generating complete future images can be computationally expensive, while highly compressed latent representations may obscure local changes. To address these challenges, we propose PLaCA, a world-to-action framework built around future latent change. PLaCA directly predicts spatially structured patch-level future latent change from the current multimodal input, preserving local information about expected environmental changes. We introduce patch-level change-aware supervision to emphasize regions undergoing larger changes and a future-to-action interface that explicitly incorporates predicted future latent change into action generation. Experiments on LIBERO-Long show that PLaCA achieves a 79.0% success rate, improving upon the Action-only baseline by 5.0 percentage points. On CALVIN, PLaCA improves the average sequence length from 2.784 to 3.564. These improvements are achieved while updating only 2.95% of the model parameters during the second training stage, demonstrating the effectiveness of future latent change modeling for downstream robotic manipulation.
[CV-116] PACE-MoE: Adaptive Domain Sampling and Domain-Specific LoRA for Medical Visual Grounding
链接: https://arxiv.org/abs/2610.27656
作者: Zhenxuan Wang,Jisheng Dang,Ronghao Lin,Bimei Wang,Jingwen Zhao,Wenjun Ma,Bin Hu,Qi Tian,Tat-Seng Chua
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 11 pages,3 figures,8 tables
Abstract:Medical visual grounding requires models to align clinical descriptions with spatial evidence across heterogeneous imaging domains. However, existing multi-domain post-training strategies typically rely on fixed domain sampling and shared parameter updates, overlooking the need to adapt both which domains receive training and where their updates are applied. To address this issue, we propose PACE-MoE, a policy-guided framework that jointly performs adaptive domain sampling and domain-specific adaptation for medical visual grounding. PACE dynamically assesses the training value of each imaging domain from current-policy feedback and progressively adjusts domain coverage through domain admission, replay, and final all-domain balancing. Meanwhile, domain-specific LoRA adapters route updates to the corresponding imaging domains while preserving the pretrained backbone and visual encoder. This coordinated design aligns training-data allocation with parameter adaptation without modifying the original reward formulation or Group Relative Policy Optimization (GRPO) objective. Experiments across seven medical imaging domains show that PACE-MoE consistently improves grounding performance over fixed-domain training, achieving a 3.64% relative improvement in mean box IoU over the Vanilla baseline. Under matched domain-specific adapter architectures, adaptive sampling alone further improves mean box IoU by 2.07%, demonstrating the benefit of policy-guided domain allocation. These results establish coordinated data-parameter routing as an effective strategy for multi-domain medical visual grounding.
[CV-117] Provisioning Robust and Efficient LiDAR Perception for Autonomous Driving
链接: https://arxiv.org/abs/2610.27617
作者: Tingxu Han,Zijie Shen,An Guo,Lyuke Wang,Chunrong Fang,Zhenyu Chen,Wei Song
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:LiDAR-based 3D object detection is a core perception service for autonomous driving, but deployment requires robustness to point-cloud corruption while meeting latency and memory requirements. Existing approaches address robustness and resource efficiency separately, resulting in compact detectors that are fragile under corruption or robust detectors that remain costly to deploy. We therefore formulate LiDAR perception provisioning as constructing a detector that remains accurate under point-cloud corruptions while meeting constraints on inference latency and peak GPU memory. To achieve robustness, we first craft corrupted samples that mimic extreme weather conditions. Then we train candidate detectors offline by ordering clean and corrupted samples by perturbation sensitivity, through which detectors learn more from hard corrupted samples. We estimate perturbation sensitivity using an ensemble of seven heterogeneous 3D detectors by tracking changes in confidence, BEV overlap, and unmatched detections. To improve efficiency, we narrow the network and use offline neighborhood-density measurements to determine how many points and neighbors each sampling stage processes. We evaluate our method using AP (Average Precision) under clean and corrupted conditions, inference latency, and peak allocated GPU memory. On KITTI for Car detection with PointRCNN, our method improves Moderate AP under corruption by 3.32 points over the original detector while reducing mean latency by 19.8% and peak allocated GPU memory by 53.9%. Across two datasets, two detector architectures, and three object classes, our method improves AP under corruption by 4.03 points on average over the corresponding original detectors, while reducing mean inference latency and peak allocated GPU memory by 14.4% and 38.0%, respectively. Our code is available at this https URL.
[CV-118] WaveSpec: Multi-Spectral Iris Presentation Attack Detection via Wavelength-Conditioned Spectral Coherence Adaptation WACV2027
链接: https://arxiv.org/abs/2610.27608
作者: Raghavendra Ramachandra
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at WACV 2027
Abstract:Multispectral iris presentation attack detection benefits from observing the same iris across multiple near-infrared wavelengths, but existing methods often use wavelength information mainly for conditioning or fusion. This can limit their ability to explicitly model whether the spectral response of a sample is coherent across bands. In this paper, we propose \textbfWaveSpec, a multispectral iris PAD framework that performs feature-level spectral coherence adaptation. Each NIR band is first encoded with a DINOv2 backbone. The proposed Set Conditioned Spectral Coherence Adapter then updates each band feature using the consensus of the available spectral set, the band-specific deviation from that consensus, and wavelength identity. The adapted features are projected into structural, spectral, and anomaly representations, adaptively pooled, and classified as bona fide or attack. We evaluate WaveSpec on the MSIrPAD benchmark under a strict cross-artefact protocol, where the model is trained with one presentation attack instrument and tested on unseen attacks. WaveSpec reduces the average D-EER from 8.45% for SpectraIrisPAD to 3.33%, and reduces average BPCER@APCER=1% from 43.16% to 12.07%. Ablation results show that the SCSA adapter and spectral residual are central to the improvement, while a band-subset analysis shows that WaveSpec remains effective under partial spectral observations and performs best when all five bands are available. The method adds only modest computational overhead, with 11.15 ms end-to-end latency per five-band sample on an NVIDIA L40S-48Q GPU.
[CV-119] DART: Scaling a Shape Representation Across Anatomical Classes and Imaging Modalities
链接: https://arxiv.org/abs/2610.27605
作者: Koushik Swarna,Pedro R. A. S. Bassi,Lam Ngo,Wenxuan Li,Jiancheng Yang,Alan Yuille,Zongwei Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Current state-of-the-art segmentation models still produce anatomically implausible shapes, such as broken vessels and fragmented glands, on the structures that are hardest to segment. Learned shape representations capture what structures look like and can help correct these errors, but most cover a fixed set of anatomical classes. Extending this coverage across classes and imaging modalities is costly because it often requires new voxel-level annotations. We propose Diffusion for Anatomical Reshaping with Text (DART), which learns a shared shape representation from segmentation probabilities. It uses anatomical descriptions to correct these probabilities without receiving medical images. Across 13 public benchmarks, absolute Dice gains reach 31% for individual gallbladder cases in CT and 10% for individual prostate cases in MRI. Using existing segmentation predictions, we scale a single shape representation from 50 to 377 anatomical classes without new voxel-level annotations for correction training. The shared representation also corrects four lesion classes absent from training and transfers from CT to MRI without retraining. Corrected labels also improve segmentation models, raising one model’s Dice from 57.1% to 67.5%. The same pipeline supports annotations for 45,000 CT scans. These results suggest that sharing shape information within and across classes can reduce the need for new voxel-level annotations when extending segmentation to more anatomical classes and imaging modalities. The model is available at this https URL
[CV-120] HiQGS: Hierarchical Spatiotemporal Querying for Feed-Forward Gaussian Reconstruction
链接: https://arxiv.org/abs/2610.27604
作者: Qiaowei Miao,Qifeng Cheng,Yijing Guo,Yuru Xiao,Ziqian Ni,Sheng Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Feed-forward Gaussian reconstruction from sparse posed monocular frames requires reasoning over observations that vary jointly in viewpoint and time. Existing methods either predict Gaussian primitives from observation-aligned features or aggregate multi-frame information through compact latent queries. However, the former tightly couples Gaussian allocation to individual frames, while the latter does not explicitly structure observed evidence according to its temporal relation to the requested timestamp. Consequently, evidence across spatial scales, viewpoints, and timestamps can be difficult to align and integrate. We present \textbfHiQGS, a query-based hierarchical framework for feed-forward Gaussian scene reconstruction from sparse monocular frames. HiQGS introduces a \textbfHierarchical Spatiotemporal Encoder that organizes multi-resolution geometry-appearance visual features according to image location, observation time, and camera pose, thereby preserving complementary appearance and structural cues. Our \textbfTime-Adaptive Decoder use target timestamp queries the resulting hierarchical representation to retrieve relevant spatiotemporal evidence and predicts the corresponding Gaussian scene state. Experiments on static and dynamic scene reconstruction benchmarks demonstrate strong rendering quality and geometric accuracy, validating the effectiveness of hierarchical visual encoding and target-time-aware Gaussian querying.
[CV-121] Causal Spatiotemporal State Evolution for Monocular 4D Reconstruction
链接: https://arxiv.org/abs/2610.27594
作者: Qiaowei Miao,Yijing Guo,Lipei Zhang,Qifeng Cheng,Yuru Xiao,Ziqian Ni,Sheng Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:While feed-forward models excel in static 3D geometry estimation, extending them to dynamic 4D scenes requires modeling visual geometry as a causally evolving state. Current methods typically rely on unordered temporal aggregation, which lacks the directional structure of physical time. Furthermore, existing supervision often treats geometry and motion as disjoint objectives, failing to exploit their inherent physical coupling. Consequently, they lack a unified constraint to effectively supervise temporal state transitions, leading to inconsistent 4D scene representations. To address this, we propose a feed-forward framework for causal 4D visual geometry learning from monocular videos. At its core, Causal Spatiotemporal Mamba augments temporal aggregation with directional state propagation. Through complementary Temporal-Spatial and Spatial-Temporal Scans, CSM continuously propagates geometry-aware states across frames to model spatiotemporal evolution. To align this causal state evolution with observable dynamics, we introduce Unified Motion Supervision. Rather than treating motion as an independent objective, UMS leverages geometry-aware predictions from the camera and depth heads to construct a unified forward motion field that is consistent with the underlying 3D structure. This geometry-grounded supervision explicitly constrains inter-frame state transitions, enabling stable states propagation across frames. Extensive evaluations on dynamic benchmarks demonstrate state-of-the-art performance across video depth, camera pose, and optical flow estimation. These results validate that the synergy between the directional state propagation of Causal Spatiotemporal Mamba and Unified Motion Supervision effectively captures continuous scene evolution, significantly reducing temporal drift and reinforcing spatial-temporal consistency in 4D representation learning.
[CV-122] Real In Real Out: What If We Only Use Real Data for Scene Text Editing? NEURIPS2026
链接: https://arxiv.org/abs/2610.27592
作者: Xingsong Ye,Yongkun Du,Jiaxin Zhang,Chong Sun,Zhineng Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted as a conference paper at NeurIPS 2026
Abstract:Aiming to modify textual content while preserving the original style, Scene Text Editing (STE) has significant practical value in many applications. STE typically requires paired images before and after editing for training. However, recent approaches rely heavily on synthetic paired data and complex disentanglement strategies. They struggle to faithfully capture the diversity and complexity of natural scenes, often producing artifacts or hallucinations. In this work, we challenge the common practice and investigate whether STE can be driven by turning each real sample into its own paired supervision. Concretely, we explore multiple ways to disrupt the original image (e.g., Crop, Shuffle, Mask) to construct self-paired training data, and build a strong baseline (TextRIRO) upon a conditional diffusion architecture. Our experiments show that combining character-column shuffling with image cropping (ShuffleCrop) effectively destroys text semantics while preserving key style cues, enabling the model to learn editing behaviors directly from the real text image. With this concise yet real-oriented design, TextRIRO achieves promising performance in both accuracy and visual realism across real-world benchmarks, substantially mitigating the synthetic-to-real domain gap. Furthermore, we leverage TextRIRO to build TextRIRO-3M, a large-scale realistic dataset for long-text recognition by concatenating edited and original images. Experimental results show that training recognition models with TextRIRO-3M significantly improves recognition accuracy, demonstrating that TextRIRO provides highly useful training resources for downstream tasks. Overall, TextRIRO lifts STE to realistic, reliable, and widely applicable new levels. Code is available at this https URL.
[CV-123] xt-Anchored Modality Experts with Uncertainty-Guided Routing for Multimodal Sentiment Analysis
链接: https://arxiv.org/abs/2610.27571
作者: Xiaohua Huang,Bingrui Lei,Xiuyan hao,Ling Zhou,Yuan Zong,Wenming Zheng,Guoying Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 1 figure
Abstract:Multimodal sentiment analysis (MSA) aims to understand user-generated videos by integrating multiple modalities. Existing methods often treat modalities as symmetric, overlooking the fact that language provides a dominant semantic prior while nonverbal signals are complementary and noise-prone. To address this, we propose a Text-Anchored Modality Experts with Uncertainty-Guided Routing, termed TAME, that reformulates MSA as an uncertainty-guided refinement paradigm, where the language prior is calibrated and enhanced by selectively trusted nonverbal evidence. Specifically, TAME adopts an asymmetric Mixture-of-Modality-Experts framework, in which textual features act as a fixed and always-activated shared expert, while audio-textual and video-textual cross-modal branches serve as dynamically routed auxiliary experts. Furthermore, we introduce lightweight Prophet Heads to quantify the sample-level predictive uncertainty of each individual modality, and this uncertainty metric is further incorporated into the modality router to realize reliability-aware expert aggregation and fusion. Experiments on CMU-MOSI, CMU-MOSEI, and CH-SIMS datasets demonstrate that TAME achieves state-of-the-art performance on the majority of metrics while reducing parameter count by up to 78.7% and FLOPs by up to 41.2% over the mixture-of-expert baseline. %These results confirm that asymmetric expert construction with uncertainty-guided routing is an effective paradigm for MSA.
[CV-124] CityPaint: A Free Lunch from 2D Diffusion for City-Scale 3D Texturing
链接: https://arxiv.org/abs/2610.27564
作者: Yuchen Zou,Lihuang Fang,Siyu Zhu,Xiao Hu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at ACM Multimedia 2026
Abstract:Generating high-fidelity, city-scale 3D environments plays a crucial role in applications such as virtual reality, gaming, and urban digital twins. However, the scarcity of real-world city-scale 3D training data forces existing methods to either stitch independently generated assets or learn from synthetic urban corpora, often sacrificing realism and global consistency. In contrast, 2D image data and their corresponding generative models are abundant and encode rich visual priors. To bridge this gap, we propose CityPaint, a novel framework that leverages 2D diffusion models to generate globally consistent city-scale 3D textures without relying on city-scale 3D training data. Specifically, CityPaint integrates geometric priors with the generative power of 2D diffusion models, introducing a robust top- K camera baking strategy for iterative texture projection. To address the artifacts of 2D-to-3D projection, we further propose a decoupled refinement mechanism that separately optimizes terrain and building textures. Extensive experiments show that CityPaint produces globally coherent textures with fine local details and achieves favorable results against strong baselines in both visual quality and cross-view consistency.
[CV-125] SynCalib: Breaking Dataset Shortcuts in LiDAR-Camera Calibration via Extrinsic-Controllable Synthesis
链接: https://arxiv.org/abs/2610.27547
作者: Huihua Zhu,Shuyang Zhang,Dongqing Chen,Zhuo Yu,Jianhao Jiao,Yilong Zhu,Siyu Teng,Xiaopin Zhong,Zongze Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 8 pages, 6 figures
Abstract:Online LiDAR-camera calibration remains a fundamental yet challenging task in the field of robotics. Recently, deep learning-based data-driven approaches have emerged, offering high-precision, targetless calibration capabilities. However, existing methods often suffer from a critical dataset shortcut phenomenon due to the lack of extrinsic diversity in their training data. Networks tend to overfit to fixed nominal extrinsic configurations, leading to severe performance degradation and poor generalization in real-world deployments. To address this limitation, we propose a highly generalizable data generation framework tailored for training LiDAR-camera calibration networks. Our framework leverages existing omnidirectional image datasets annotated with depth ground truth to synthesize multi-modal calibration data under rigorous extrinsic constraints. Extensive experiments demonstrate that the data generated by our pipeline significantly enhances the generalization and robustness of typical learning-based calibration networks. Furthermore, we conduct comprehensive real-world evaluations across various camera-LiDAR configurations, validating the zero-shot transfer capability of models trained exclusively on our synthetic data. To facilitate future research and community use, the code for data generation, the pre-trained models, and online benchmark datasets will be made publicly available.
[CV-126] Less is More: Concise Retrieval and Critic-based Re-ranking for Composed Image Retrieval CIKM2026
链接: https://arxiv.org/abs/2610.27538
作者: Yuqian Zheng,Hyoju Ahn,Hyeongjin Ahn,Eunil Park
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at CIKM 2026
Abstract:Zero-shot composed image retrieval (ZS-CIR) aims to retrieve a target image based on a reference image and modification text. Recent approaches commonly rely on large vision-language models to generate target image captions for retrieval. However, these captions may include redundant semantic details, introducing noise when retrieving the target image and leading to suboptimal retrieval results. To address this issue, we propose Less-is-More CIR, a two-stage training-free framework that combines Concise Retrieval and Critic Re-ranking. In the first stage, concise target-oriented captions are generated that preserve only the most discriminative modification semantics, thereby reducing semantic noise in the retrieval. In the second stage, a critic re-rank module further evaluates the retrieved candidates from the perspectives of modification consistency and reference preservation. Extensive experiments demonstrate the effectiveness and robustness of the proposed framework for the ZS-CIR. Our code is available at: this https URL.
[CV-127] OmniVLoc: Topo-metric Visual Localization for Omnidirectional Cameras
链接: https://arxiv.org/abs/2610.27522
作者: Zhuo Yu,Shuyang Zhang,Huihua Zhu,Wenjia Ouyang,Zuo Zuo,Jianhao Jiao,Yilong Zhu,Xiaopin Zhong,Zongze Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 8 pages, 7 figures
Abstract:Omnidirectional cameras provide surround-view information, significantly overcoming the field-of-view constraints for traditional visual localization systems. However, existing localization systems fail to fully exploit the characteristics of omnidirectional cameras, struggling to deliver accurate and robust localization performance amidst scene content variations and illumination-induced appearance changes. In this work, we propose OmniVLoc, a surround-view localization system based on a topo-metric framework. By integrating state-of-the-art global place recognition and monocular depth estimation methods, we introduce a multi-constrained visual localization paradigm. By solving a scale-aware pose graph and performing precise spherical bundle adjustment, we accurately recover the metric-level pose for each query this http URL, a hierarchical verification mechanism is designed to filter out erroneous outputs across different stages, effectively bolstering the overall system robustness. Extensive experiments on both public and our custom datasets validate the superior accuracy and reliability of OmniVLoc. To benefit the community, our source code and dataset will be made publicly available upon publication.
[CV-128] Multiscale Temporal-Spectral-Spatial EEG Learning With Modality-Specific and Modality-Invariant Representations for Emotion Recognition
链接: https://arxiv.org/abs/2610.27521
作者: Jian Zhang,Manrong Liu,Yilei Huang,Wenhua Liu,Xiaohui Yang,Lijun Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal physiological signals provide complementary cues for emotion recognition. They also raise two related challenges. Emotion-related EEG patterns vary across time, frequency bands, electrodes, and brain regions. Different physiological modalities also contain both modality-specific and shared affective information. This paper presents GMSI-Net, an EEG-centered multimodal emotion recognition framework. It has three key components. The temporal-spatial-spectral-region (TSSR) encoder models multiscale temporal-channel patterns, differential entropy features, and local-to-global brain-region relations. The modality-specific representation (MSR) branch uses bidirectional cross-modal attention and learnable gates. It controls information transfer while retaining modality-specific cues. The modality-invariant representation (MIR) branch learns a shared affective representation. An adaptively weighted contrastive loss aligns this representation across modalities. Independent multiscale Conv-Transformer encoders process the peripheral physiological signals. GMSI-Net is evaluated on DEAP and MAHNOB-HCI using within-subject random ten-fold cross-validation. The model reaches 96.21 (1.47) and 96.41 (1.60) for valence and arousal on DEAP. It reaches 94.10 (2.86) and 94.48 (3.77) on MAHNOB-HCI. These results indicate that combining EEG features with complementary physiological information can improve emotion recognition.
[CV-129] HealBench: A Fine-Grained Benchmark for Evaluating Psychological Comfort in Text-to-Image Generation
链接: https://arxiv.org/abs/2610.27516
作者: Junjie Xu,Yuzhe Yang,Rui Yang,Luwei Xiao,Wenxin Wu,Xingjiao Wu,Liang He
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Despite advances in photorealism, semantic fidelity, and aesthetics, text-to-image generation and image editing remain understudied in a human-centered dimension: whether their outputs feel psychologically comforting. Grounded in cognitive evaluation theory, we define psychological comfort as an affective response shaped by viewers’ appraisal of restorative and emotionally supportive visual cues. We introduce HealBench, a fine-grained benchmark of 30,000 high-quality image-text pairs spanning 19 categories and 92 subcategories, selected from 100,000 candidates through a human- and expert-in-the-loop pipeline combining vision-language model matching, human preference models, expert filtering, and exemplar-guided augmentation. We also construct HealBench-Eval, a feature-clustered evaluation set of 1,095 instances. Using HealBench-Eval, we conduct objective and subjective evaluations of 14 representative image generation models and prompt-rewriting strategies. Results reveal a gap between conventional visual quality and perceived comfort: strong semantic alignment and aesthetics do not guarantee emotionally restorative images. Reinforcement learning experiments further show that HealScorer can serve as a reward signal for assessing and improving the emotional and restorative quality of generated images. These findings establish psychological comfort as a distinct dimension of human-centered image generation and position HealBench as a foundation for affect-aware generative models.
[CV-130] Progressive Subset Selection and Cascaded Consistency for Unsupervised Vision-Language Model Adaptation
链接: https://arxiv.org/abs/2610.27515
作者: Hao Yan,Guanghui Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Pretrained Vision-Language Models (PVLMs) achieve strong zero-shot recognition but often suffer performance degradation under target-domain shift. Unsupervised adaptation provides a practical way to improve target-domain performance without labeled data, but it remains challenging because pseudo-labels generated by the model can be unreliable and no labeled validation set is available for model selection or early stopping. We propose a simple framework for unsupervised PVLM adaptation that combines progressive margin-based subset selection with cascaded consistency learning. The subset selection strategy constructs a class-balanced pseudo-labeled set using prediction margins and gradually expands it during training, enabling adaptation to start from reliable supervision anchors before incorporating broader target-domain data. Cascaded consistency learning further exploits unlabeled samples by using temporally smoothed teacher predictions on weakly augmented views to supervise student predictions on strongly augmented views, with confidence filtering to reduce the influence of uncertain pseudo-labels. To preserve the pretrained embedding structure, the framework updates only normalization parameters in the visual encoder and class prototypes initialized from text prompts. Experiments across multiple recognition benchmarks with both CLIP and SigLIP backbones show consistent improvements over zero-shot baselines and competitive performance against recent unsupervised PVLM adaptation methods. These results suggest that progressive pseudo-label selection and teacher-guided consistency provide an effective and lightweight strategy for adapting PVLM classifiers without labeled target data.
[CV-131] Cross-Visibility-Gap Identity Persistence for 4D Scene Graphs
链接: https://arxiv.org/abs/2610.27505
作者: Jiahui Hu,Li Gao,Yu Liu,Pengzhan Zhou
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 8 pages, 2 figures
Abstract:4D scene graphs require persistent object identity when an object disappears inside an opaque carrier and later reappears as a fresh detection. When appearance and target-centric motion are unavailable, identity recovery must rely on relational history. We formulate cross-visibility-gap identity persistence as historical-node recovery through a remembered target–carrier binding. We introduce CVG-SGM, a perception-frontend-compatible transfer–maintain–read memory. Geometry-weighted cross-node interaction transfers target-specific evidence into carrier states, a shared per-node GRU maintains this information during transport, and a two-time readout combines stored carrier–identity compatibility with current detection–carrier geometry to recover the historical identity. In the controlled one-identity-per-carrier setting, CVG-SGM achieves 1.000 re-link accuracy, compared with 0.570 for both no-memory and parameter-count-matched no-write controls. It also generalizes to scene-disjoint, unseen longer visibility gaps. On real RGB-D/SAM2 feature-level episodes, it correctly re-links 36/36 held-out detections, compared with 28/36 for the strongest fixed carrier–identity control. Counterfactual carrier-memory swaps redirect predictions according to the swapped bindings, supporting the role of stored relational information in identity recovery. The current approach remains limited by shared-carrier identity ambiguity and does not provide end-to-end raw-video re-identification. Comments: 8 pages, 2 figures Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2610.27505 [cs.CV] (or arXiv:2610.27505v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.27505 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-132] FocusOVD: Geometric and Contextual Feature Refinement for Open-Vocabulary Small-Object Detection
链接: https://arxiv.org/abs/2610.27479
作者: Hanlin Kou,Beiqi Bai,Li Chen,Youjia Jing,Shupan Li,Mingyuan Jiu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted as a full paper at ChineseCSCW 2026
Abstract:Open-vocabulary object detection aims to detect both seen and unseen categories by leveraging the semantic knowledge of vision-language models. However, existing methods often struggle with small objects due to insufficient preservation of high-resolution details, ineffective information aggregation across feature levels, and limited region-level discriminability. To address these challenges, we propose FocusOVD, an open-vocabulary object detector that enhances small-object representations without modifying the frozen VLM backbone. FocusOVD introduces three complementary components: a Reweighted Cross-scale Feature Fusion module, which learns normalized level-wise weights to adaptively balance shallow appearance details and deep semantic cues during feature pyramid aggregation; a Residual Pinwheel-shaped Gated Convolution module, which strengthens local geometric responses in high-resolution features; and a Residual Convolutional Gated Linear Unit module, which suppresses background-dominated activations and enhances RoI discriminability before prediction. Extensive experiments on COCO-OVD and LVIS-OVD demonstrate the effectiveness of FocusOVD. On COCO-OVD, FocusOVD achieves a novel-category AP of 36.3%, outperforming CORA, UniDetector, and SAMP by 1.3%, 1.1%, and 1.5%, respectively; on LVIS-OVD, it achieves an overall AP of 33.7%, surpassing the strongest competing method by 0.6%. These results demonstrate the strong generalization ability of FocusOVD for open-vocabulary small-object detection.
[CV-133] LoRASpace: A Pool-Wide Shared Substrate for Static and Dynamic Multi-LoRA Composition NEURIPS2026
链接: https://arxiv.org/abs/2610.27444
作者: Thamirawaran Sathiyalogeswaran,Navindu De Silva,Chamikara Siriwardane,W. K. Alwis,Jason Mars,Udaya Sampath K. Perera Miriya Thanthrige,Ranga Rodrigo
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at NeurIPS 2026. 29 pages, 6 figures, 26 tables. Code: this https URL . Project page: this https URL
Abstract:Low-rank adaptation (LoRA) attaches small low-rank matrices to a frozen backbone, and public repositories host many LoRAs per backbone, each for a different concept or task. Combining several LoRAs in one generation forces a trade-off. Merging their weights into one update needs a single forward pass, but LoRAs trained for different concepts interfere and quality drops. Running each LoRA in its own forward pass (composite) preserves quality, but inference time grows linearly with the number of active LoRAs. We study a property of LoRA pools overlooked by prior multi-LoRA work. In real LoRA pools, the stacked low-rank factors need far fewer directions than their summed ranks. Using singular value decomposition, we find the minimal subspace that keeps 99% of the factor energy. On three diffusion pools, only 46% to 57% of the summed LoRA ranks is needed, and the subspace grows sub-linearly with pool size. A 48-LoRA sub-pool of LoRA-Hub on FLAN-T5 tasks shows comparable overlap (62%), while the diversity-curated LoRA-Retriever benchmark sits near the no-overlap baseline (96%), so pool composition rather than modality predicts where compression is large. We build on this overlap with LoRASpace, a training-free method that computes the subspace once per pool and reuses it in two inference modes, a static mode at merge cost and a dynamic mode for input-dependent routing. On the diffusion pools, static LoRASpace matches composite quality at merge cost, and dynamic LoRASpace is more than 6x faster than composite with six active LoRAs on a 17-LoRA pool, mainly because it runs one backbone pass instead of six. In an ablation, static variants at the same rank match LoRASpace, so the quality gain does not depend on the pool-wide basis. On the near-orthogonal LoRA-Retriever benchmark, compression is small and gains come from coupling with external retrieval.
[CV-134] Using Set Learning to Improve Set-Valued Classifiers on Long-tailed Data
链接: https://arxiv.org/abs/2610.27443
作者: Kevin Kasa,Scott C. Lowe,Graham W. Taylor
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Set-valued classification (SVC) offers a practical approach to communicating uncertainty while maintaining accuracy by predicting variable-sized sets per input rather than fixed-sized sets (e.g. top-1 or top-5). However, set-valued classifiers (SVCs) exhibit significantly degraded performance on long-tailed and imbalanced datasets, a setting for which they are ostensibly well-suited. This becomes apparent under macro-average evaluation, which weighs all classes equally. SVCs are typically formed during a post-hoc calibration stage, so little is known about how training dynamics affect their behaviour. In this work, we investigate how different neural network training strategies and loss functions impact downstream SVC performance on highly imbalanced data. In addition to methods such as label smoothing and class-balancing, we investigate whether loss functions that operate over sets of examples in a batch, rather than treating all examples as independent, lead to better downstream SVCs. We adapt the recently proposed odd-K-out (OKO) loss function, which learns from sets of training samples, improving its sample efficiency whilst introducing algorithmic modifications that reduce its computational overhead from 3.2x to just 1.36x that of standard cross-entropy training. Experiments on highly long-tailed datasets demonstrate our modified OKO approach significantly improves macro-accuracy across multiple SVC methods - including top-k, average-k, and conformal prediction - while maintaining competitive micro-accuracy. This improved performance on minority classes occurs without substantially increasing prediction set sizes, suggesting that carefully constructed set-based training can produce networks better suited to imbalanced data.
[CV-135] okenFlow: Latent Token Rectified Flow Matching for 3D Point Cloud Generation
链接: https://arxiv.org/abs/2610.27418
作者: Linhao Wang,Dongjin Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:3D point clouds generation requires capturing both part-level layout and surface detail while preserving permutation equivariance. Existing latent generative models often encode an entire shape into a single vector, which limits their ability to retain part-level structure. We propose TokenFlow, a hierarchical flow matching framework that represents each shape as a set of latent tokens and couples a rectified flow over these tokens with a conditional flow over points. A bottleneck of M \ll N learnable tokens aggregates global context through cross-attention at O(NM) cost. On ShapeNet, TokenFlow achieves the best MMD and coverage across airplane, car, and chair, while obtaining 1-NNA comparable to the strongest baselines. On ModelNet, its learned latent set reaches the best reported linear-evaluation accuracy on ModelNet10. These results show that representing a shape as a set of latent tokens, aggregated at O(NM) cost, provides an effective representation for 3D point cloud generation.
[CV-136] KPHarness: Co-Evolving Knowledge and Probing for Long-Horizon Gaming Harnesses
链接: https://arxiv.org/abs/2610.27417
作者: Zhikang Chen,Yueqing Sun,Sen Cui,Ziang Ye,Yu Wang,Yan Feng2,Xierui Song,Yu Zhang,Tingting Zhu
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:A failed gaming trajectory reveals that an agent needs intervention, but not necessarily how its execution harness should change. Effective harness adaptation therefore requires gathering diagnostic evidence, rather than merely proposing alternative recovery rules. We present KPHarness, a framework that adapts the external execution harness of a frozen vision-language-action model through knowledge-probing co-evolution. Recovery knowledge supplies context-specific intervention hypotheses that guide active probing; probe outcomes reveal when these hypotheses apply and provide evidence for refining both recovery knowledge and later probe selection. This reciprocal process allows the harness to improve not only which interventions it applies, but also how it investigates failures. KPHarness separates fast policy execution from a slow research loop that translates accumulated knowledge and probing evidence into bounded harness revisions. Each revision is compared with the incumbent under matched sandbox conditions and accepted only when the paired evaluation shows improvement. Recoveries supported by further validation are retained for cross-task reuse, while immediate interaction feedback and delayed episode-success credit update action- and knowledge-probing policies during adaptation. Experiments on Minecraft tasks spanning mining, combat, and crafting show improved task coverage over bare policy execution, a manually specified fixed harness, and knowledge-only harness evolution with non-adaptive probing. Evaluation on previously unseen tasks further shows improved task completion with the evolved harness, knowledge bank, and probing policies held fixed,
[CV-137] st-time Adaptation for Point Cloud Denoising under Unseen Noise
链接: https://arxiv.org/abs/2610.27410
作者: Kun Jiang,Shida Fu,Fanhao Yi,Weiqing Li,Zhiyong Su
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:Although deep learning-based point cloud denoising methods have demonstrated impressive results in recent years, their effectiveness usually degrades substantially in real-world scenarios where the noise distribution of test point clouds deviates from that of the training set. To address this challenge, we propose a test-time adaptive denoising (TTAD) framework that enables existing denoising approaches to adapt to unknown noise patterns at test time without relying on additional supervision or retraining from source data. Taking a pre-trained denoising model and a test noisy point cloud as input, TTAD first employs a residual-driven noise modeling (RDNM) module to estimate the sample-specific noise distribution through analyzing the deviation between the noisy input and its pseudo-clean version denoised by the initial model. Then, a noise-aware contrastive adaptation (NACA) module is designed to build sample-specific noisy-noisier test-time training pair from the denoised point cloud, and update the model parameters over a few adaptation steps per sample at test time through contrastive learning. Furthermore, TTAD naturally extends to a continual TTAD (CTTAD) setting, facilitating efficient progressive refinement across sequential test samples with consistent noise distributions, while reducing the computational cost per sample. Extensive experiments on both synthetic and real-world datasets demonstrate that our approach significantly improves the generalization ability of existing denoising models under unseen noise, achieving superior performance across multiple evaluation metrics.
[CV-138] PerQ: Inverse Generative Modeling for Neural Image Compression via Quantization Error Compensation
链接: https://arxiv.org/abs/2610.27374
作者: Tianhao Peng,Ho Man Kwan,Fan Zhang,Shan Liu,Dave Bull
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Neural image codecs face a fundamental trade-off between pixel fidelity and perceptual quality. Distortion-optimized codecs preserve pixel-wise fidelity but produce blurred reconstructions at low bitrates, whereas generative codecs leverage learned priors to improve perceptual quality but tend to exhibit saturating PSNR at higher bitrates and require separate models for different bitrate targets. We propose PerQ, which reframes perceptual reconstruction as inverse generative modeling with bounded support. Specifically, we define the quantization cell as the set of latents that yield the same transmitted symbols and derive a transport bound showing that reconstructions from this cell remain within a rate-determined radius of the codec output. This motivates performing generative reconstruction within the codec-induced quantization cell, rather than sampling from an unconstrained image prior. PerQ implements this principle by attaching a rate-conditioned flow-matching compensator to a frozen, multi-rate distortion-optimized backbone, enabling perceptual enhancement while preserving the backbone’s coding performance. Experiments on Kodak and CLIC2020 show that PerQ achieves perceptual performance comparable to generative codecs such as MS-ILLM in LPIPS and FID, while maintaining competitive PSNR relative to distortion-optimized codecs such as ELIC. A single model checkpoint supports a wide bitrate range with fast decoding, and the compensator provides consistent gains across three different backbones. The implementation is available at this https URL.
[CV-139] On the Limits of Scale: Inverse Scaling in Vision-Language Models
链接: https://arxiv.org/abs/2610.27360
作者: Rohit Saxena,Alessandro Suglia,Marek Rei,Pasquale Minervini
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Scaling laws suggest that increasing model size, training data, and compute generally yields monotonic performance improvements. While this often holds at the aggregate level, we present systematic evidence that vision-language models (VLMs) exhibit inverse scaling at the subtask level: specific question categories on which larger models perform worse than smaller counterparts within the same model family. We evaluate 19 models from four distinct model families across eight multimodal benchmarks (nine evaluation splits), answering 29,386 questions per model, 558,334 model-question evaluations in total. Using per-question pairwise analysis across all model pairs in each family, we identify 375 subtask-model-pair combinations exhibiting inverse scaling, spanning 4,342 unique queries where a smaller model answers correctly but a larger model fails, and we find that these patterns replicate across independently developed architectures. We identify 13 subtask-dataset pairs that show inverse scaling across multiple families, consistent with the phenomenon being task-driven rather than model-specific. The most affected domains in our study include multi-image spatial reasoning, visual perception, and university-level STEM knowledge, with accuracy gaps exceeding 20 percentage points in several cases, while diagram understanding appears comparatively resistant. These findings challenge the assumption that increasing model size uniformly improves multimodal capabilities and suggest that model selection should be task-aware.
[CV-140] Look Before You Reason : Implicit Visual Thinking for Efficient Multimodal Reasoning NEURIPS2026
链接: https://arxiv.org/abs/2610.27341
作者: Zerui Chen,Changrui Chen,Fei Ni,Jiankang Deng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by NeurIPS 2026. Project Page: this https URL
Abstract:Vision-language models (VLMs) have advanced rapidly in multimodal reasoning, yet they remain limited on tasks requiring fine-grained understanding of small or easily overlooked image regions. Recent thinking with images approaches mitigate this limitation by allowing models to inspect images through operations such as zooming, cropping, or code execution. While effective, their multi-turn interactions with VLMs introduce redundant token computation and slow inference, and separate region inspection from the reasoning step that depends on it. To address these issues, we propose Look Before You Reason (LoRe), an implicit visual thinking framework for efficient visual reasoning within a single-turn VLM interaction. Given an image and a question, LoRe predicts the most relevant image region, aligns it with the corresponding image patch features, and strengthens the associated key-value (KV) cache entries, enabling subsequent decoding to directly exploit the selected visual cues without additional visual inputs or explicit tool calls. We further employ reinforcement learning to guide the VLM toward identifying informative image regions and leveraging them during reasoning. To support fine-grained evaluation of multimodal reasoning, we introduce COCO-ObjQA, a large-scale benchmark for object-centric multimodal reasoning. Experiments on three challenging multimodal benchmarks demonstrate that LoRe consistently improves VLM reasoning performance with low additional inference overhead.
[CV-141] Steering Vision-Language Models for Visual-Contextual Privacy in Location Disclosure
链接: https://arxiv.org/abs/2610.27323
作者: Wenyue Chen,Yuhan Cheng,Ruixin Yang,Ethan Mendes,Alan Ritter,Jingwei Sun,Hai Li,Yiran Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-language models (VLMs) can infer sensitive information such as geolocation from subtle visual cues, creating privacy risks when responses disclose more detail than the visual context warrants. We study visual-contextual privacy in geolocation disclosure, where a VLM should reveal location information only at the granularity appropriate to the visual context depicted in the image. We first show that this contextually appropriate granularity is recoverable from pre-generation hidden states, even when generated responses are misaligned, with disclosure levels exhibiting an ordered structure along a common hidden-state direction. Motivated by this, we propose an inference-time route-and-steer framework: an ordinal granularity router maps visual-contextual privacy representations to a disclosure route, and a disclosure-level-conditioned steering module applies route-specific hidden-state interventions during decoding. Across four VLM backbones, our method improves visual-context-aware disclosure alignment, better balances over- and under-disclosure, and incurs only modest changes on general capability benchmarks. It further transfers to another dataset with a distinct data distribution and privacy-risk taxonomy.
[CV-142] What Not to Erase: Saliency-Constrained Augmentation for Low-Resource Handwritten Character Recognition
链接: https://arxiv.org/abs/2610.27309
作者: Paraskevi Platanou,John Pavlopoulos
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Optical Character Recognition (OCR) and handwritten character recognition (HCR) in low-resource historical material collections often rely on data augmentation, yet augmentation strategies are typically evaluated almost exclusively through predictive accuracy. This leaves open whether a model’s improved performance is accompanied by saliency stability. This question is important in low-resource historical HCR, where recognition is performed on variable and partially degraded letterforms and two models with similar predictive performance may differ in how consistently their saliency patterns respond to such variation. We therefore study saliency-guided augmentation as an explanation-aware strategy for low-resource HCR. Rather than treating saliency maps only as post-hoc interpretations, we use them to define augmentation variants that perturb, preserve, or constrain visually informative regions of character images. We compare classical document-style transformations, random lacunae, and several saliency-constrained lacuna variants. To evaluate the effect of augmentation beyond accuracy, we introduce the Explainability Consistency Index (ECI), a metric that quantifies the stability of saliency patterns under controlled perturbations. We find that saliency-constrained augmentation can improve recognition over classical augmentation, but no single saliency strategy is consistently best. The preferred form of the constraint varies with both architecture and corpus, indicating that saliency is better treated as a mechanism for controlling augmentation than as a fixed augmentation strategy. In some settings, saliency-constrained training improves explanation consistency even when predictive performance is similar. This shows that predictive metrics alone can obscure differences in model sensitivity under perturbation, motivating explanation consistency as a complementary evaluation dimension.
[CV-143] Generalizing Multimodal Foundational Models in Context-Driven Class Incremental Learning WACV2027
链接: https://arxiv.org/abs/2610.27298
作者: Avi Gupta,Nilotpal Sinha,Vishnu Raj,Sambuddha Saha,Pratik Joshi,Koteswar Rao Jerripothula,Tammam Tillo
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: This work has been accepted in WACV 2027
Abstract:Class-Incremental Learning (CIL) aims to learn new classes continuously without forgetting the previously learned ones. Recent works in CIL have significantly advanced across various modalities; however, at the intersection of audio-visual modality, this problem remains underexplored. While representations from multimodal foundational models (MFMs) encapsulate rich static priors, our preliminary analysis shows that these representations struggle in incremental settings. Motivated by this, we propose a framework that generalizes frozen MFMs to audio-visual CIL. Specifically, we leverage their dense audio and visual representations and employ an attention-based contextual guidance module, in which the audio representations serve as task-stable contextual guidance for the visual representations, enabling effective class recognition in the incremental setting. To further mitigate catastrophic forgetting and preserve stability across tasks, we employ feature- and logit-level distillation, which are primarily responsible for the reduced forgetting. Extensive evaluations across various audio-visual CIL benchmarks demonstrate that our approach consistently outperforms state-of-the-art methods. Code is available at this https URL.
[CV-144] On Maximum Clique and Maximum Consensus Formulations in Point Cloud Registration
链接: https://arxiv.org/abs/2610.27275
作者: Abolfazl Babanazari,Kaveh Fathian
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 7 figures. Interactive examples and Lean proof verification: this https URL
Abstract:Maximum clique and maximum consensus are widely used formulations for robust point cloud registration in robotics. Maximum consensus seeks one transformation that aligns as many matches as possible, while maximum clique selects matches through pairwise distance tests. Maximum clique is often used as a substitute for or an approximation to maximum consensus, with empirical evidence suggesting a close relationship. However, how closely the two formulations agree is not fully understood. We study this relationship theoretically. We show that a maximum clique can have arbitrarily large alignment error and be arbitrarily larger than the maximum consensus set at any fixed positive tolerance, and derive geometric conditions that bound this error. When rotation is fixed, we derive sharp error bounds and tests that exactly characterize maximum consensus for the translation subproblem. Finally, we study group- k consistency, which tests groups of k matches, and determine the smallest group sizes that guarantee alignment error bounds independent of point cloud scale and geometry. These are d+1 for rotations and translations and d+2 when reflections are allowed, in dimensions d\ge 2 . Interactive examples and proof verification using the Lean theorem prover are available at this https URL.
[CV-145] Language Guidance for Supervised Vision Training: An Empirical Study of Generalization
链接: https://arxiv.org/abs/2610.27208
作者: Shruthi Gowda,Nikita Jain,Elahe Arani,Bahram Zonooz
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Deep neural networks have achieved remarkable success on vision benchmarks, yet they continue to struggle with many generalization challenges. Supervised vision training relies on one-hot labels, which provide limited information about semantic structure and shared attributes between classes. This limited supervision can leave visual representations vulnerable to distribution shifts, spurious correlations, texture bias, adversarial perturbations, and forgetting in sequential learning settings. We study whether pretrained language models can serve as lightweight auxiliary supervision for vision training without requiring paired image-text data, prompt engineering, or contrastive objectives. Specifically, we evaluate two forms of language guidance, Explicit Language Guidance (ExLG) and Implicit Language Guidance (ImLG). We conduct a comprehensive evaluation across six generalization regimes, including in-distribution, out-of-distribution generalization, shortcut and spurious correlation resistance, texture and shape bias, adversarial robustness, and continual learning. Our analyses show that the two mechanisms have complementary strengths, with explicit guidance consistently benefiting in-distribution, low-data performance, and continual learning retention, while implicit guidance is often more useful in shortcut-sensitive settings and under stronger adversarial perturbations. Importantly, both are lightweight and add minimal parameters and training overhead. The analyses characterize when language-derived structure helps supervised vision training and provides a practical roadmap for using off-the-shelf pretrained models from another modality as auxiliary supervision.
[CV-146] Which Depth Estimator for Jetson AGX Orin? A 2026 Benchmark of Monocular and Stereo Methods
链接: https://arxiv.org/abs/2610.27199
作者: Leaf Jiang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 5 figures, 8 tables. Code and per-scene results: this https URL
Abstract:The Jetson AGX Orin is a common computer for robots and vehicles, yet measured accuracy, speed and energy of current depth estimators on it are hard to find. We benchmark nine model families as of October 2026: the monocular Depth Anything V2 and 3, Depth Pro, MoGe-2, Metric3D v2 and UniDepth V2, and the stereo methods libSGM, NVIDIA ESS and Fast-FoundationStereo. Every method runs on the Orin as a TensorRT FP16 engine or native CUDA code at up to three input sizes, up to the full 5.7 MP, and is scored for metric accuracy against Middlebury 2014 ground truth (non-occluded pixels of the 10 scenes whose disparity stays below 256 px), GPU time, energy and licence. Stereo is far more accurate. At 30 fps, NVIDIA’s ESS reaches 0.70% AbsRel in 8.6 ms and libSGM 0.73% in 25 ms; with more compute, Fast-FoundationStereo reaches 0.33%. The best monocular models reach 12% with the true focal length (Depth Anything 3 Metric-L in 33 ms, Depth Pro in 803 ms), and 3.8-4.0% only with an oracle per-image scale (MoGe-2-L, UniDepth V2-B, Metric3D v2-L). Stereo error falls as the input resolution grows; no monocular model improves, and the metric scale of most drifts away from the truth outside their training resolution. At depth discontinuities, which AbsRel barely registers, the monocular models scatter 17-46% of the edge points between foreground and background, learned stereo 2-9% and libSGM at full resolution 1%. We give recommendations per frame-rate budget and release all code, per-scene results and engine build scripts.
[CV-147] CARLAR Influence: A Structured Hallucination Risk in Multimodal Cheque Extraction and an Exploratory Compressed-Image Countermeasure
链接: https://arxiv.org/abs/2610.27192
作者: Joong Ho(Joshua)Choi,Aditya Paul,Nidhi Singh
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 12 pages
Abstract:Multimodal large language models (LLMs) are increasingly used for structured extraction from financial documents, but their behavior under visual ambiguity is poorly understood. We study CARLAR (Courtesy Amount Recognition and Legal Amount Recognition) influence: a model returns the same amount for both cheque fields even though their annotations differ, including cases with a blank annotation encoded as zero. Across ten visual prompting conditions, we evaluate GPT-4.1 and Claude 4.5 Sonnet on 243 primary cheques (124 with two observed unequal amounts) and 598 strict mismatches from two additional datasets. Under courtesy redaction, GPT-4.1 CARLAR reaches 73.7 percent (95 percent bootstrap interval: 68.3 to 79.0; p below 0.001 versus baseline). Model-reported confidence can remain high when predictions agree; expected calibration error reaches 0.93 among CARLAR cases. We also test a compressed-image few-shot intervention that presents examples as one visual table without model fine-tuning. On the 243-case cohort, Claude’s Control condition increases paired joint extraction accuracy by 11.5 percentage points (95 percent interval: 6.2 to 16.9; adjusted p = 0.00054), but its strict-124 gain is smaller and not significant after correction. The intervention’s effect on predicted-field agreement is model-dependent: it decreases for Claude and increases for GPT-4.1. These results motivate validation methods that account for visual context and dependencies between extracted fields.
[CV-148] PRIMFLOW: Learning Trajectory Primitives for Few-Step Distillation
链接: https://arxiv.org/abs/2610.27146
作者: Hongyi Fang,Benjia Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Few-step trajectory distillation aims to learn accurate global updates from dense local supervision along a teacher’s sampling trajectory. Existing approaches construct within-step dynamics using specific trajectory parameterizations or discrete interval predictions, tying the learned update to a prescribed temporal structure. We propose \method, which directly learns a continuous trajectory primitive representing cumulative displacement along the teacher trajectory. By supervising finite differences of the primitive with local teacher velocities, \method learns the accumulated motion directly, while obtaining global updates through endpoint differences. This formulation avoids prescribing a temporal basis or fixed query grid and instead learns the temporal evolution of cumulative displacement from local supervision. We further establish a local-to-global error bound to characterize how local regression errors affect full-interval displacement accuracy, and introduce endpoint absorption to eliminate the auxiliary primitive branch while preserving one-network-evaluation inference. Experiments on large-scale text-to-image generation show that \method achieves a T-FID of 20.92 and an HPSv2 score of 28.35 at one NFE, improving both metrics over a trajectory-distillation baseline. After endpoint absorption, it retains comparable performance, achieving a T-FID of 20.86 and an HPSv2 score of 28.33.
[CV-149] MatchCityLoc: Aerial Localization in LoD City Models via Surface Normal Matching WACV2027 ATC
链接: https://arxiv.org/abs/2610.27110
作者: Oussema Dhaouadi,Zuria Bauer,Johannes Michael Meier,Marc Pollefeys,Daniel Cremers,Olaf Wysocki
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: WACV 2027 - Project page: this https URL
Abstract:Autonomous Unmanned Aerial Vehicle (UAV) navigation in urban environments requires precise pose estimation. Because Global Navigation Satellite System (GNSS) signals are often unreliable in cities, visual localization serves as a crucial complement. Within this domain, textureless Level-of-Detail (LoD) city models provide a lightweight and privacy-preserving substitute for heavy 3D textured meshes. However, localizing against abstract geometry is challenging. Existing pipelines typically score 2D silhouette overlaps, which can suffer from severe geometric ambiguities. We present MatchCityLoc, a framework that shifts this paradigm to direct feature matching from a noisy 6-DoF pose prior. By extracting surface normals from the query image using a zero-shot foundation model and matching them against rendered LoD normals, we establish explicit 3D-to-2D correspondences. This enables a Perspective-n-Point (PnP) pose solve, which we iteratively refine through a render-and-match loop. Experiments demonstrate strong pose recall and accurate translation and rotation among the evaluated off-the-shelf LoD-based methods with very few renderings. By bridging the modality gap between aerial imagery and texture-free geometry, MatchCityLoc enables accurate localization using lightweight, privacy-preserving city models.
[CV-150] Counting Contributions Not Views: Occlusion-Aware Densification for 3D Gaussian Splatting ACCV2026 CCL
链接: https://arxiv.org/abs/2610.27108
作者: Marcel Rogge,René Schuster,Didier Stricker
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: Accepted at ACCV 2026 (Oral). Includes the supplementary material as appendix. Code: this https URL
Abstract:3D Gaussian Splatting grows a dense scene from a sparse initialization through adaptive density control (ADC), driven by per-Gaussian statistics gathered during rendering. These statistics pass a frustum-based visibility filter blind to occlusion: fully occluded Gaussians still count as observed, diluting the averaged gradient signal and suppressing densification precisely where geometry is occluded. We correct this with an occlusion-aware filter that restricts densification statistics to views in which a Gaussian actually contributes, and replace the criterion’s raw view-count normalizer with an average-contribution one. Integrated into a StopThePop-based framework, the two together improve every image-quality metric over the occlusion-blind baseline on occlusion-heavy scenes (up to 6% lower LPIPS), at less than half the Gaussians a view-count normalizer requires. Under the longer schedules these large scenes require, the count stays bounded where most criteria diverge, and on smaller benchmarks our method stays competitive at a fraction of the competitors’ count.
[CV-151] LVMem: Image-Augmented Memory for Long-Horizon Question Answering
链接: https://arxiv.org/abs/2610.27104
作者: Victoria Khoruzhaya,Maksim Kalmykov,Tuong Do Minh,Tatiana Zemskova,Yixin Zhu,Dmitry Yudin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 3 figures
Abstract:Reliable language-guided perception is essential for embodied agents, where questions about objects, events, and their locations must be answered from long observation histories. Unlike conventional visual question answering, realistic robotic settings require reasoning over extended temporal horizons and persistent scene memory. We build upon the view-based memory paradigm and propose an efficient framework for long-horizon spatio-temporal question answering. While existing view-based systems typically store either text captions or visual embeddings, we extend the ReMEmbR-style memory with image paths that link each stored description to its original visual content. To make the stored annotations more informative, we replace uniform sampling during memory construction with an optical-flow-based frame selection strategy (PSFR) and store an object-centric descriptive scene representation instead of plain captions. We further equip a VLM-based agent with an adaptive retrieval tool that returns the image associated with the most relevant caption, allowing the agent to decide when visual inspection is needed. On the OC-NaVQA benchmark, LVMem outperforms ReMEmbR on binary and positional metrics with both backbones and matches or exceeds a 4D scene graph DAAAM baseline on binary accuracy, while operating on a lightweight view-based memory that requires no explicit graph construction.
[CV-152] BUMP: Body-driven Unified Multi-object Physics for HOI Synthesis
链接: https://arxiv.org/abs/2610.27066
作者: Wonjung Woo,Hyeonwoo Kim,Hanbyul Joo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Generating human-object interactions (HOIs) with multiple objects is challenging because the resulting motions arise from different underlying causes. Human motion reflects semantic intent, intentionally manipulated objects must remain coordinated with the human, while surrounding objects move primarily as physical consequences of contact and collision. We present BUMP, a text-driven framework for multi-object HOI generation based on a simple principle: learn intentional motion and simulate physical consequences. Our method first generates full-body human motion from text and scene context, then predicts trajectories for a small number of intentionally manipulated driver objects conditioned on the human motion. We then obtain responder object motions through physics simulation, allowing the framework to handle varying numbers and configurations of surrounding objects without expanding the learned trajectory output space. To address the scarcity of paired full-body multi-object HOI data, we combine heterogeneous supervision from human-only motion, full-body HOI, and hand-centric manipulation data augmented with plausible body motion. We further introduce outcome-conditioned interaction editing, which infers a local modification of the causal driver object trajectory from a desired responder object behavior without requiring an explicit target driver object trajectory. Experiments demonstrate strong performance on both single-object and multi-object HOI benchmarks. Qualitative results showcase diverse scenarios involving inter-object relationships and edits, highlighting the advantage of our unified model formulation.
[CV-153] VitronEdit: Universal Visual Editing across Images Videos and 3D Scenes
链接: https://arxiv.org/abs/2610.27049
作者: Jing Wang,Yanlin Li,Mong-Li Lee,Wynne Hsu,Shengqiong Wu,Hao Fei
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 27 pages,9 tables,9 figures
Abstract:Recent advances in unified visual editing have begun to consolidate image and video editing within a single model, while 3D scene editing remains largely separate and constrained by scarce paired training data. Given the rich image and video editing supervision now available and the common editing semantics shared across modalities, a natural question is whether this knowledge can be transferred to data-scarce 3D editing. We present VitronEdit, a universal framework for image, video, and 3D scene editing. VitronEdit adopts a shared sequence representation and progressive multimodal training to transfer editing knowledge across modalities. We further introduce geometry-aware post-training, which leverages a pretrained geometry reconstruction model to provide depth and camera-pose supervision. Surprisingly, applying geometric supervision only to 3D samples yields the best overall performance, while image and video editing also improve despite receiving no direct geometric supervision, suggesting that geometric regularization transfers through the shared backbone. Our analysis further shows that image and video training mainly transfers editing semantics to 3D, whereas 3D supervision primarily improves multi-view consistency. The project page: this https URL.
[CV-154] Robustness of Internal Signals for Hallucination Detection in Vision-Language Models NEURIPS2026
链接: https://arxiv.org/abs/2610.27041
作者: Allison Chen,William Yang,Salma Abdel Magid,Jonathan Williams,Olga Russakovsky,Esin Tureci
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: NeurIPS 2026 Evaluations and Datasets track
Abstract:When given an image and a text prompt, vision-language models (VLMs) can produce seemingly sophisticated text responses. However, they are also prone to hallucinate objects in their output, which can have detrimental consequences in safety-critical applications. To detect these hallucinations, certain methods interestingly leverage internal signals from the VLM as it generates the text response; potentially, these methods use the same signals that led to a hallucination to detect it. We conduct an in-depth analysis to understand the robustness of these detection methods. First, we show that a naive baseline of a generated word’s token position in the VLM text response achieves competitive performance. Further, we reveal how some detection methods rely on token position and use this heuristic as a shortcut. Second, we explore detection methods’ sensitivity to language and vision modalities. Our experiments suggest that strong detection performance is often driven by detecting language-based hallucinations, and when methods rely on token position, they tend to be less visually grounded and exhibit reduced robustness when hallucinations occur earlier in a text response. Lastly, we demonstrate that different detection methods can carry complementary signals; when ensembling methods together, adding weaker methods to an ensemble of higher performing ones still offers performance improvements. Altogether, while there may be useful internal VLM signals to detect hallucinations, our investigation reveals important limitations of the current methods that try to leverage them.
[CV-155] Long-wave Infrared Polarimetric Remote Sensing Image Fusion for Vehicle Detection
链接: https://arxiv.org/abs/2610.27018
作者: Wenjuan Zhang,Xiangrui Peng,Zhen Li,Caimeng Yi,Xuesong Ma,Jianwei Wang,Jie Pan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Blurred vehicle contours and low contrast limit detection performance in infrared remote sensing imagery, while existing fusion methods may introduce polarization noise and fail to integrate complementary information effectively. To address these limitations, we propose a Mamba-based infrared polarimetric remote sensing image fusion model. Infrared polarization information reflects surface material properties and geometry, providing complementary information to infrared intensity. The proposed method first combines long-wave infrared total intensity S0 and the degree of linear polarization (DoLP), using a physics-inspired polarization-response proxy and spatial structural continuity to extract guidance features. It then employs bidirectional state space scanning along horizontal, vertical, and both diagonal orientations to enable cross-modal feature fusion across the entire image. Prior-guided multi-window structural similarity, reconstruction constraints, and a relative lower-bound loss for polarization saliency are jointly introduced to suppress noise while preserving global background structures and target polarization features. In addition, we construct AIR-LWIR-Pol, a dataset of 1,901 accurately registered image pairs acquired using an airborne division-of-aperture long-wave infrared polarimetric imaging system, and evaluate downstream vehicle detection using YOLO11. The experimental results show that the proposed method outperforms the comparison methods on four fusion metrics, including entropy and structural similarity. Vehicle detection mAP50 and mAP50-95 reach 55.0% and 29.4%. These results demonstrate the effectiveness of combining physical priors with state space modeling for long-wave infrared polarization image fusion and support its use for target detection in complex remote sensing scenes.
[CV-156] REST: Reliability-Guided Spatial Transcriptomics Supervision for Gene Expression Prediction from Histopathological HE-Stained Images
链接: https://arxiv.org/abs/2610.26988
作者: Lanfeng Zhong,Xiangyu Sun,Guixuan Xu,Rongzhao Zhang,Jie Xu,Shaoting Zhang,Guotai Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Histopathology is central to cancer diagnosis, but morphology in Hematoxylin and Eosin (HE)-stained images does not directly reveal the molecular states that characterize disease heterogeneity and inform tumor characterization, prognostic assessment, and biomarker discovery. Spatial Transcriptomics (ST) measures spatially resolved gene expression, providing molecular information within tissue architecture, yet differences in assay coverage and spot informativeness make paired HE–ST observations unequally reliable for representation learning. We propose Reliability-Guided Spatial Transcriptomics (REST) supervision with Biological Program Bank (BPB), a program-structured and reliability-guided ST-supervised training framework for predicting molecular information from H\E images. REST constructs BPB from curated signature, pathway, and cell-type gene sets to provide a shared structure for ST encoding, reliability estimation, and molecular reconstruction. Guided by BPB, Spot-Reliability Contrastive Alignment (SRCA) measures the reliability of each spot from assay-level program coverage and spot-level activity dispersion, and reweights HE–ST alignment accordingly. Biology-Informed Program Reconstruction (BIPR) further uses learnable program queries to reconstruct member-gene expression from histological morphology. REST’s training corpus comprised 209 samples (407,475 spots across 16 organs); evaluation used 70 disjoint samples (230,056 spots across eight organs) for gene-expression, pathway-activity, and cell-signature prediction. REST achieved the highest Macro correlation among state-of-the-art ST-supervised training methods and outperformed existing pathology foundation models. Code is available at this https URL.
[CV-157] Deep Auxiliary Learning via Bilevel Feature Optimization for Label-Efficient Medical Image Segmentation
链接: https://arxiv.org/abs/2610.26954
作者: Fuping Wu,Xi Jia,Le Zhang,Jiahua Li,Rima Mustafa,Thomas E. Nichols,Bartlomiej W. Papiez
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Auxiliary learning (AL), which can improve the performance of a primary task by leveraging related auxiliary tasks, is commonly regarded as a special case of multi-task learning. Existing AL approaches typically require manually annotated auxiliary labels, increasing annotation costs, particularly in medical imaging where expert labeling is expensive. In this work, we investigate whether segmentation performance can be improved using auxiliary tasks whose labels are automatically derived from existing segmentation annotations, requiring no additional manual labeling. To address this, we propose a novel auxiliary learning framework, formulated as a bilevel optimization problem. Our method enables effective knowledge transfer between tasks via a feature communication module, enhancing the primary segmentation task. We evaluate the method on three medical image segmentation benchmarks under limited-label settings. The results demonstrate consistent improvements over state-of-the-art methods, achieving superior label efficiency. The code is publicly available at this https URL.
[CV-158] Rate as a Feature Not a Nuisance: Curriculum FiLM Training for Event-Based Graph Neural Networks
链接: https://arxiv.org/abs/2610.26933
作者: Kavyansh Tyagi,Andreas Ziegler
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:An event camera emits events at a rate given by the scene motion, object texture, illumination, its own contrast thresholds, and other settings. Across the N-Caltech101 dataset, recorded by a single event camera, the event rate varies by roughly 26x. Standard event-based Graph Neural Network GNN pipelines do not make use of the variable event rate. They either normalize each event’s timestamp into a fixed window or train the network to be invariant to the event rate. We argue that both choices lose information that could instead be used to improve the performance of a pipeline. Per-recording normalization is non-invertible, incurring about 6pp (percentage points) of natural-rate accuracy loss in our setting. And invariance training trades accuracy for robustness to shifts in the event rate. We instead take a different path and condition on the event rate directly. We add a per-recording log-rate scalar that modulates every stage of a PointNet++ backbone through Feature-wise Linear Modulation (FiLM), initialized to identity, so that training begins at the unmodified baseline. A two-stage curriculum (natural event rates first, then rate-augmented fine-tuning) reaches 67.2% on N-Caltech101 (+6.3pp) and 82.5% on DVS128 Gesture (+14.0pp) with an additional ~0.2M parameters. A matched-capacity control experiment shows that the gain is not from the increase in model capacity but from the rate conditioning. We will make the implementation publicly available upon acceptance of this work.
[CV-159] SuperClip Pyramid for Video Temporal Grounding
链接: https://arxiv.org/abs/2610.26925
作者: Sanghyeok Lee,Juyeon Ko,Joonmyung Choi,Jongha Kim,Dohwan Ko,Ji Soo Lee,Hyunwoo J. Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: preprint
Abstract:With the rapid expansion of online video content, video temporal grounding has become essential for localizing moments based on user-specific queries, comprising two subtasks: moment retrieval (MR) and highlight detection (HD). While recent approaches adopt DETR-style decoders or temporal pyramid architectures, they rely on single-scale feature maps or uniformly downsampled representations, limiting their ability to capture semantic boundaries and variable-length events. In this paper, we present SuperClip Video Temporal Grounding (SCVTG) built on a multi-scale temporal hierarchy that better captures distinctive events at various granularities. SCVTG constructs a SuperClip pyramid by grouping adjacent clips into semantically coherent segments, termed SuperClips. Further, to aggregate these SuperClips having variable duration and temporal location, we introduce a temporally adaptive kernel that aggregates SuperClips through temporally conditioned weights, enabling flexible aggregation. On top of this, we design relative saliency losses to guide the model in learning inter-clip importance, enhancing its ability to distinguish relevant events. Extensive experiments show that SCVTG achieves the best performance across five video temporal grounding benchmarks with promising gains.
[CV-160] CaMeWM: A Causal Medical World Model for Generative Tumor Evolution and Treatment Planning
链接: https://arxiv.org/abs/2610.26923
作者: Huanlong Gao,Xuelei He,Fengjun Zhao,Xiaowei He
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 30 pages,8 figures,12 tables,including supplementary material
Abstract:Treatment planning is intrinsically counterfactual: clinicians must compare how a patient’s tumor and survival would evolve under competing regimens, whereas a retrospective cohort records only the regimen received and one follow-up scan. A treatment-conditioned generator trained on such data can therefore reproduce historical treatment allocation rather than simulate the interventional disease dynamics needed to rank regimens. We introduce \textbfCaMeWM, a causal medical world model for supported post-TACE tumor evolution and treatment planning. CaMeWM implements one intervention-to-decision loop rather than three independent prediction tasks: it first admits a queried regimen only for clinically comparable histories with empirical treatment support; it then generates a distribution of lesion-local follow-up worlds through an action-conditioned response descriptor, deterministic anchor, and residual 3D diffusion transition; and it finally maps sampled worlds to restricted mean survival time (RMST) with a response-Cox utility model. Under pre-treatment adjustment and positivity, this loop estimates a prespecified, support-restricted g-formula target; all selection and nuisance estimation are fold-local, and the induced policy is evaluated by cross-fitted, censoring-adjusted doubly robust estimation. Across 79 paired out-of-fold episodes, CaMeWM achieves 0.618 SSIM and 0.528 change-map SSIM for follow-up CT transition, while its generated-world survival pathway attains a 0.615 C-index; factual-only AIPW estimates a 9.60-week RMST gain (95% CI 0.61–18.13) for the frozen policy. CaMeWM provides an inspectable basis for retrospective, support-restricted treatment planning.
[CV-161] Controllable Multi-Object Insertion for Detection Data Augmentation
链接: https://arxiv.org/abs/2610.26922
作者: Jiyang Huang,Tianwei Zhang,Xiaoqin Zhang,Ling Shao,Shijian Lu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Controllable object insertion provides a targeted approach to augmenting object detection datasets that require category–count fidelity, annotation alignment, and source-content preservation beyond visual realism. Several approaches have been proposed to address this grand challenge, but most suffer from different constraints in precise category–count requests and visual consistency, limited coordination across categories and instances, and poor localization of multiple objects of specific category and count. We design SceneP2F, a two-stage trainable framework that connects scene-aware joint planning with layout-aligned generation leveraging explicit instance-level layouts. SceneP2F consists of two novel designs. The first is ScenePlan which allocates one class-labeled query per requested instance, jointly predicts scene-compatible object centers, and separately estimates box sizes conditioned on the planned centers and scene context. The second is SceneAlign-Fill which combines geometry- and appearance-aware conditioning, box–class-guided regional attention, and an annotation-region-weighted flow-matching objective to render the planned objects while limiting unintended changes to source content. Extensive experiments on three challenging benchmarks UAVDT, VisDrone, and BDD100K show that SceneP2F outperforms representative general-purpose editors and object-insertion methods consistently, demonstrating the value of coordinating instance-level planning and generation for category-targeted detection augmentation.
[CV-162] GTSM: Gaussian Splatting with Thermal State-Space Dynamics for Spatio-Temporal Thermal Field Reconstruction and Prediction
链接: https://arxiv.org/abs/2610.26918
作者: Yunpeng Zhao,Haoxiang Zhang,Chao Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:Temperature-field reconstruction and prediction are important for condition monitoring, process analysis, and state assessment in industrial and dynamic environments. This work aims to recover a four-dimensional spatio-temporal thermal state field directly from two-dimensional observations, without requiring a predefined 3D model, prescribed thermodynamic equation, or predefined heating/cooling trajectory. To this end, we propose the Gaussian Thermal State-Space Model (GTSM), which represents thermal evolution through a shared scene-level latent process and Gaussian-specific local thermal states. The resulting continuous dynamics are propagated using differentiable RK2 integration, allowing thermal evolution to be inferred from multi-view and multi-temporal observations and continuously extrapolated beyond the observed interval. GTSM is designed as a backbone-agnostic thermal dynamics module for Gaussian representations and can be integrated with both static 3D Gaussian Splatting and dynamic 4D Gaussian Splatting, supporting thermal reconstruction and prediction under both static and time-varying target geometry. The reconstructed representation further enables novel-view thermal rendering and future-time thermal-state prediction. Experiments on real infrared sequences and synthetic dynamic thermal data demonstrate improved reconstruction accuracy and substantially stronger temporal extrapolation than representative baselines, while also validating the applicability of GTSM to deforming targets.
[CV-163] CodecMotion-OAD: Efficient Online Action Detection with Compressed-Domain Motion
链接: https://arxiv.org/abs/2610.26915
作者: Zepeng Sun,Naichuan Zheng,Hailun Xia,Junjie Wu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Online action detection (OAD) identifies ongoing actions in streaming videos. Many existing OAD systems combine RGB appearance with optical flow, but dense-flow extraction remains a major computational bottleneck. Codec motion vectors (MVs) are available during decoding, avoiding a separate optical-flow computation. On THUMOS’14, optical flow exceeds MVs by 11.2 mAP points with motion-only inputs, whereas the gap narrows to 3.7 points after RGB fusion, with the largest reduction in action-category discrimination. We propose CodecMotion-OAD, which replaces optical flow with decoder-provided MVs while retaining existing temporal detector architectures. Its two-stage training first aligns MV features with a pretrained flow-feature interface without action labels and then trains RGB+MV detectors using ground-truth labels and soft predictions from a frozen RGB+Flow teacher. To the best of our knowledge, this is the first application of codec MVs to frame-level OAD. On THUMOS’14 with MiniROAD, CodecMotion-OAD achieves 68.3 mAP versus 70.9 mAP for matched-distillation RGB+Flow, while reducing decoder-to-prediction processing time by 11.48-fold on CPU and 2.04-fold on GPU.
[CV-164] Memory Internalization via Test-Time Training for Long-Video Understanding
链接: https://arxiv.org/abs/2610.26910
作者: Haonan Zhao,Jiaxuan Chen,Qingyang Liu,Shuochen Chang,Siyu Sun,Jiangtong Li,Li Niu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Long-video understanding often requires answering many questions about a single video whose evidence is distributed across extended temporal contexts. Conventional video-language models repeatedly encode frames or retrieve visual evidence for every question. Parameterized video memory offers a write-once alternative, but existing methods rely primarily on synthesized text and fixed adaptation configurations, leaving open where video knowledge should be written and how visual evidence should shape the memory. We introduce Memory Internalization via Test-Time Training (MIT^3), a per-video adaptation framework that writes each long video into compact, video-specific adaptive parameters for video-free, text-only question answering. Under compact multi-scale supervision, causal tracing measures the causal memory utility of each model layer for video-grounded answering and selects video-specific adaptation locations. Visual memory distillation then trains the text-only memory to reproduce the representation change induced by the relevant video clip, preserving the internal effect of visual evidence beyond textual supervision. Across LVBench, LongVideoBench, and Video-MME, MIT^3 achieves data-efficient video-free question answering with substantially fewer synthesized samples than prior parameterized video-memory methods, demonstrating that effective video internalization depends on both where knowledge is written and how visual evidence is distilled into the model.
[CV-165] Uncertainty Gates the Gradient: Adaptive Updates for Generalized Few-Shot 3D Segmentation
链接: https://arxiv.org/abs/2610.26909
作者: Yifei Zhao,Fanyu Zhao,Shengtang Wu,Han Zhang,Ruike Cao,Yinsheng Li
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:In generalized few-shot 3D point cloud segmentation, support annotations are often compromised by sparsity and occlusion. Blindly adapting to such low quality samples risks learning spurious geometric patterns rather than robust novel concepts. Although 3D Vision-Language Models have been introduced to sup plement supervision, the potential of uncertainty to actively govern adaptation dynamics remains underexplored. Existing signal-level methods like loss reweight ing fail to prevent unreliable episodes from steering the optimization trajectory. To address this, we propose Uncertainty-Gated Adaptive (UGA), to the best of our knowledge the first framework that elevates uncertainty to update-level gating for generalized few-shot 3D point cloud segmentation. The core mod ule, Uncertainty-Gated Update (UG-Update), estimates episodic reliability via intra-class dispersion over merged support and pseudo-labeled novel supervision, and selectively gates residual gradient updates. This encourages the model to learn novel semantics from reliable episodes while attenuating updates during unreliable ones. Complementing this, we introduce an Uncertainty-Based Pseudo Label Filter (UB-Filter) to enforce learning from high-confidence dense signals. On ScanNet200 and ScanNet++ (1-/5-shot), UGA sets new state of the art with an average +3.08 HM IoU over the strongest prior baseline.
[CV-166] Beckmann World Models: Direct Terminal-Map Learning for One-Step Action-Conditioned Prediction
链接: https://arxiv.org/abs/2610.26903
作者: Adam Lee,Shobhit Agarwal,June Pyo Suh
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Action-conditioned world models support planning and policy evaluation by predicting future states. However, learning an accurate one-step predictor can require substantial training computation. We introduce Beckmann World Models (BWM), which learn a direct map from noise, visual history, and actions to future frames without a pretrained generative teacher. We adapt the terminal-map conservation law of Beckmann Transport Models to action-conditioned video prediction, training directly from noise-data pairs without conditional populations of generated samples required by DriftWorld. However, transport consistency on interpolated inputs does not ensure accurate prediction from the pure noise used at inference. We address this gap with endpoint-aligned training that directly supervises the source prediction, emphasizes motion and spatial structure, and teaches recovery from generated histories. All components train the same map and retain one network call per future chunk. On PushT and Robomimic Can, the final model reduces 64-frame prediction MSE by 68% and 57%, respectively, relative to same-backbone regression baseline. For policy evaluation, BWM achieves the lowest mean IoU error among the evaluated models, 0.0100, at 24.8 ms per four-frame PushT prediction. On Robomimic Can, the BWM stage trained with motion supervision and generated histories achieves 3.3 times higher throughput and 78% lower peak memory than local one-step DriftWorld. In a latent-space extension to RT-1 and Bridge-V2, equal-update controls show that feature supervision under generated histories improves perceptual prediction. These findings show how terminal-map learning can combine efficient training with accurate one-step visual prediction.
[CV-167] EGGHead: Efficient Generalizable Gaussian Heads from a Single Image
链接: https://arxiv.org/abs/2610.26898
作者: Pilseo Park,Andrew Hou,Michel Sarkis,Ning Bi,Yiying Tong,Xiaoming Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Creating photorealistic human heads typically involves trade-offs between representation power, efficiency, and generalizability. Traditional mesh-based solutions relying on 3DMMs generally sacrifice representation power but are efficient to render. On the other hand, NeRF-based methods improve representation power and fidelity while sacrificing efficiency. Gaussian splatting methods attempt to resolve these trade-offs by offering both high representation power and efficiency with real-time rendering. However, their efficiency nonetheless depends on the number of rasterized Gaussians, and the time needed to optimize and prepare the head (the enrollment time) is often high. In light of this, we propose EGGHead, which introduces a novel Gaussian allocation algorithm that efficiently assigns Gaussians to the head’s semantic regions given a fixed Gaussian budget. Our algorithm reduces the number of Gaussians needed while minimizing the impact on representation power. EGGHead operates in a single forward pass and outperforms previous single-image face modeling methods in reconstruction quality, novel view synthesis, and rendering speeds.
[CV-168] Curvature as a Baseline: Monocular Pose Measurement of Cylindrical Objects with Dual Colour Stripe Markers
链接: https://arxiv.org/abs/2610.26003
作者: Jaesung Hong,Dong Won Lee
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Thin cylindrical objects offer little space for planar markers. We propose a dual-colour-stripe marker that uses the known curvature of an object to improve pose measurement with a single camera. The boundary between two colours provides a long centre line, while dots and printed edges supply features at different depths. We analyse how this depth separation improves the estimation of rotation about the cylinder axis, or roll. For a camera-facing marker at 15 cm, the Cramér-Rao lower bound on roll standard deviation is 2.00°for a flat marker, 0.38°for a single curved marker and 0.21°for a wrap-around marker. Monte Carlo roll RMS errors are within about 10% of the corresponding standard-deviation bounds. A first-order relation connects the sagitta of the curved marker to the roll sensitivity provided by a stereo baseline. The analytical and simulation results explain how curvature improves monocular measurement and provide a basis for designing marker geometry.
[CV-169] Half-Truths Break Similarity-Based Retrieval
链接: https://arxiv.org/abs/2602.23906
作者: Bora Kargi,Arnas Uselis,Seong Joon Oh
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:When a text description is extended with an additional detail, image-text similarity should drop if that detail is wrong. We show that CLIP-style dual encoders often violate this intuition: appending a plausible but incorrect object or relation to an otherwise correct description can increase the similarity score. We call such cases half-truths. On COCO, CLIP prefers the correct shorter description only 40.6% of the time, and performance drops to 32.9% when the added detail is a relation. We trace this vulnerability to weak supervision on caption parts: contrastive training aligns full sentences but does not explicitly enforce that individual entities and relations are grounded. We propose CS-CLIP (Component-Supervised CLIP), which decomposes captions into entity and relation units, constructs a minimally edited foil for each unit, and fine-tunes the model to score the correct unit above its foil while preserving standard dual-encoder inference. CS-CLIP raises half-truth accuracy to 69.3% and improves average performance on established compositional benchmarks by 5.7 points, suggesting that reducing half-truth errors aligns with broader gains in compositional understanding. Code is publicly available at: this https URL
[CV-170] Robotic Video World Models: A Survey of Applications Research Challenges Future Directions
链接: https://arxiv.org/abs/2601.07823
作者: Zhiting Mei,Tenny Yin,Ola Shorinwa,Apurva Badithela,Zhonghe Zheng,Joseph Bruno,Madison Bland,Lihan Zha,Asher Hancock,Jaime Fernández Fisac,Philip Dames,Anirudha Majumdar
类目: ystems and Control (eess.SY); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Video world models have emerged as promising candidates for high-fidelity world models, offering the potential to synthesize high-quality videos capturing fine-grained interactions between agents and their environments conditioned on multi-modal user inputs. Their impressive capabilities address many of the long-standing challenges faced by physics-based simulators, driving broad adoption in many problem domains, e.g., robotics. For example, video models can generate photorealistic, physically consistent deformable-body simulation without making prohibitive simplifying assumptions, which is a major bottleneck in physics-based simulation. Moreover, video models can serve as foundation world models that capture the dynamics of the world in a fine-grained and expressive way. They thus overcome the limited expressiveness of language-only abstractions in describing intricate physical interactions. However, despite their potential, video models still generate physics-violating future predictions, often manifesting as hallucinations. In this survey, we provide a review of video models and their applications as embodied world models in robotics, including efficient data generation and policy learning, dynamics and rewards modeling in reinforcement learning, policy evaluation, and visual planning. Further, we highlight important challenges hindering the trustworthy integration of video models, such as poor instruction following, hallucinations like violations of physics, unsafe content generation, in addition to significant data and compute overhead. We present potential future directions to address these open research challenges to motivate research and ultimately facilitate broader applications, especially in safety-critical settings. We provide a curated bibliography at this https URL .
[CV-171] CAPS: Clinically-Verifiable Anatomy-Prioritized Synthesis for Missing MRI Imputation in Brain Tumor Segmentation CEC
链接: https://arxiv.org/abs/2610.28179
作者: Cecilia Diana-Albelda,Arthur Longuefosse,Mario Rubén Paz Campos,Álvaro García-Martín,Jesus Bescos
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages, 6 figures, 8 tables. Cecilia Diana-Albelda and Arthur Longuefosse contributed equally
Abstract:Multi-parametric MRI is the gold standard for neuro-oncology, yet acquiring complete 4-modality protocols may be unfeasible in real-world clinical practice. When handling missing modalities, existing automated systems force a critical compromise: robust segmenters operate as uninterpretable “black boxes” that bypass image generation, while standard synthesis models often produce blurry reconstructions lacking diagnostic reliability. To address this issue, we propose CAPS, a Clinically-Verifiable Anatomy-Prioritized Synthesis framework that reconciles computational precision with clinical transparency. Our multi-modal generative architecture is driven by a hybrid objective function (L_1 + L_AFP + L_DualDice) that explicitly enforces the preservation of deep anatomical features and strict topological boundaries, rather than relying solely on pixel-wise optimization. Evaluated comprehensively on the BraTS Adult Glioma dataset, the framework achieves superior synthesis fidelity across intensity-based metrics, and demonstrates robust downstream segmentation performance (82.38% mean Dice), successfully restoring verifiable visual context. Furthermore, the framework exhibits exceptional zero-shot cross-domain generalization, maintaining high structural fidelity across completely unseen Meningioma, Pediatric, and Sub-Saharan African cohorts without domain-specific adaptation. Finally, a blinded radiological assessment confirms that our anatomy-aware approach significantly enhances both perceived image quality and clinical utility, providing clinicians with the reliable visual evidence required to trust downstream automated workflows.
[CV-172] Channel-Level Mask Guidance for Slimmable Image Compression
链接: https://arxiv.org/abs/2610.27037
作者: Chen Chen,Chenggang Wu,Yongbiao Chen,Weidong Qiu,Zhengwei Qi,Haibing Guan
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 12 figures. Submitted to IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)
Abstract:Learned image compression (LIC) has achieved promising rate-distortion performance, but most methods lack scalability for diverse platforms with varying computational and storage constraints. Existing slimmable LIC methods achieve flexible complexity via consecutive-channel activation but suffer from two limitations: forced consecutive-channel selection degrades narrow-model performance at low bitrates, and the lack of stable variable-rate control limits adaptability to varying bandwidth. We propose a channel-level mask-guided slimmable image compression method. A learnable channel-wise mask with sparsity-promoting regularization enables non-contiguous top-w channel selection, letting narrow models use the most informative channels. After training, mask merging and channel reordering convert the model into a standard slimmable network via Mask-to-Slimmable (M2S) conversion, enabling efficient inference without architectural changes. For variable-rate control, we combine lambda- and width-conditioned modulation with quantization-step scaling and a progressive training pipeline stabilized by a freeze-then-fine-tune strategy. Experiments on Kodak across E2E, Hyperprior, and Checkerboard backbones show superior rate-distortion performance across widths and bitrates, with large gains for narrow models at low bitrates. At w=48, it achieves up to 19.22%, 24.11%, and 22.76% BD-rate savings over the same-width Slimmable baseline on E2E, Hyperprior, and CKBD, respectively. A single VBR model replaces 40 independent models, reducing storage by 14.8x to 15.9x. Code: this https URL.
[CV-173] Codec-Compatible Learned Compression for 3D Gaussian Splatting
链接: https://arxiv.org/abs/2610.26920
作者: Feiyang Yu,Qi Yang,Wenjie Huang,Yunfeng Guan,Yiling Xu,Zhu Li
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the IEEE International Conference on Visual Communications and Image Processing (VCIP 2026). 5 pages, 3 figures, and 1 table
Abstract:Video-based 3D Gaussian Splatting (3DGS) compression methods provide a practical route by representing Gaussian attributes as 2D maps and coding them with standard video codecs. Among these maps, high-dimensional non-DC spherical harmonics (SH-AC) maps contain the largest number of channels and account for a dominant portion of the bitstream. However, their statistics differ substantially from those of natural images, and the nonlinear dependencies among SH-AC coefficients remain insufficiently exploited. To address these issues, we propose CCT-GS, a Codec-Compatible learned Transform framework for 3DGS compression. CCT-GS first learns a generalized SH-AC transform that captures cross-scene regularities and converts SH-AC maps into compact codec-friendly bottleneck maps. It further introduces LoRA-based scene adaptation to recover scene-specific appearance details with lightweight side information. Experiments on MPEG GSC sequences demonstrate that CCT-GS consistently outperforms representative video-based baselines, achieving an average BD-rate reduction of 51.30% over the video-based anchor.
人工智能
[AI-0] Industrial ReMEmbR: Scaling Spatiotemporal Memory to Robot Fleets
链接: https://arxiv.org/abs/2610.28624
作者: Jad Yammine,Malek Wahidi,Abrar Anwar,Yan Chang,Mihir Acharya,Anthony Rizk
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Warehouse robot fleets continuously observe operational conditions, but preserving these observations as reusable semantic memory incurs substantial computation, communication, and storage costs. For example, explaining recurring aisle blockages can require reviewing video from multiple robots across several shifts. We present IndustrialRemembr, a shared spatiotemporal memory system that extends NVIDIA ReMEmbR to robot fleets and reduces server-side computation by generating scene descriptions only on demand. On each robot, an NVIDIA Jetson encodes video clips into retrieval embeddings with Cosmos-Embed1, vision tokens with Cosmos3-Edge, and reconstruction latents with the Wan2.2 VAE, without uploading raw camera frames. The encoding models are re-used to detect people in image patches, zeroing them out to support PII abstraction for GDPR-compliant deployments. We split Cosmos3-Edge at its vision-language boundary, placing the vision encoder on the robot and the language decoder on an on-premises server. A Nemotron agent retrieves memories by content, position, and time and queries the decoder using cached vision tokens. This design eliminates caption generation at write time, while cached-token decoding reproduces the unsplit model’s output in all 450 tested decodes.
[AI-1] Adaptive Multi-Token Prediction: How far should we go?
链接: https://arxiv.org/abs/2610.28619
作者: Emile Anand,Mohammed Rehan Parwani,Xinyuan Cao
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Systems and Control (eess.SY)
备注: 25 pages, 5 figures, 9 tables
Abstract:Multi-token prediction (MTP) can accelerate language-model decoding by drafting several tokens and verifying them in a single target-model pass. However, the best speculation horizon changes with context: longer drafts offer greater amortization when reliable, but can waste compute on suffixes discarded after a rejection. Unlike autoregressive drafting, where a stopping rule can decide token by token, MTP must fix the horizon before drafting, as all heads fire in the same forward pass. We therefore formulate horizon selection as an online control problem, using verification feedback to learn how far the draft can be trusted in each context, and introduce Adaptive Multi-Token Prediction (AMTP): a lightweight controller that requires no retraining and leaves the exact verifier unchanged. To guide horizon selection, we characterize the expected number of tokens committed at each horizon via acceptance survival, the probability that a draft prefix is fully accepted. Before each round, AMTP predicts how far the draft is likely to survive and picks the horizon with the highest expected tokens per second, learning freely from every verification pass. Theoretically, we show that any online learner’s prediction error translates into throughput regret against the best horizon for each context, and in turn into a bound on wall-clock completion time. With online gradient descent, this regret is sublinear. Across five draft-target pairings built on Llama-3.1-8B-Instruct with Medusa heads, Meta’s MTP models, and Qwen3-4B-Inst-2507-MTP, evaluated on AIME 2025 and GSM8K, AMTP decodes up to 4.1x faster than standard decoding and matches or outperforms the fixed window in 8 of 10 settings, with throughput gains of up to 10%.
[AI-2] Semantic Observability for Autonomous Systems Diagnostics
链接: https://arxiv.org/abs/2610.28611
作者: Jitendra Solanki,Anand Todkar,Ziran Min
类目: Artificial Intelligence (cs.AI)
备注: 8 pages, 3 figures. Foundational architecture paper; companion evaluation paper forthcoming
Abstract:Modern distributed microservice deployments generate massive volumes of heterogeneous, multi-format observability logs across physically and logically partitioned infrastructure. Isolating root causes during cascading failures remains an arduous task for Site Reliability Engineers (SREs), bottlenecked by manual keyword filtering and disjointed temporal correlation across monitoring dashboards. We introduce Semantic Observability, an autonomous multi-agent framework that unifies disparate log streams into a dynamic, queryable Property Graph Database. By modeling physical infrastructure topologies, logical software dependencies, and chronological event sequences as attributed nodes and directed relations, it bridges the semantic gap inherent in flat, unstructured log repositories. Building on this unified substrate, we deploy a stateful multi-agent system orchestrated through Directed Cyclic Graphs (DCGs). The system autonomously navigates the topology, applying causal graph algorithms and temporal back-traversal to isolate the precise blast radius of anomalies. The pruned subgraph is then interpreted by a Large Language Model (LLM) reasoning engine to generate contextualized, code-level Root Cause Analyses (RCA) and actionable remediations. This paper establishes the foundational system architecture, formal relational ontology, and multi-agent control loops of Semantic Observability. We show conceptually how the framework eliminates ambient telemetry noise, reduces Mean Time to Resolution (MTTR), compresses LLM token consumption, and prevents structural hallucinations. We also specify an open-source validation framework on standard benchmarks (LogHub and Train-Ticket), laying the groundwork for a forthcoming companion paper on exhaustive quantitative evaluation and production benchmarking.
[AI-3] Value-Rich Differently Organized: Measuring Value Expression and Organization in AI-Generated Debates EMNLP2026
链接: https://arxiv.org/abs/2610.28604
作者: Maoyuan Li,Haoyuan Li,Zhongsheng Wang,Jiamou Liu
类目: Artificial Intelligence (cs.AI)
备注: Accepted at EMNLP 2026 (Main Conference)
Abstract:Current evaluation of LLM-generated debate largely focuses on argumentative quality, leaving the value structure of generated arguments underexamined. We introduce a diagnostic metric framework for comparing how values are expressed and organized in human-reference and AI-generated debates. The framework separates value expression, the range of recognizable values a debate invokes, from value organization, how those values are distributed and sustained across turns. Grounded in Schwartz’s Basic-10 value theory, it represents debates as multi-level value distributions over argumentative units, utterances, stances, and full debates, enabling metrics for value coverage, alignment, diffusion, and cross-turn dynamics. Across multiple generation systems, AI-generated debates are broadly value-expressive, often activating as many or more value dimensions than matched human references, but they organize these values differently. Their utterances are more locally value-diffuse, and source-blind judges distinguish value clarity from value focus. These findings show that debate evaluation should consider not only value presence, but also value organization. Code and supplementary materials are available at this https URL.
[AI-4] urning TS2Vec into a Foundation Model for Time Series Classification
链接: https://arxiv.org/abs/2610.28602
作者: Vasilii Feofanov,Zhenghan Tai,Andrei Ermakov,Irina Rish
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:TS2Vec is one of the strongest self-supervised methods for time series classification, introduced in 2022. It is designed to be trained anew on every dataset: self-supervised pre-training followed by linear probing. Since then, the field has shifted towards foundation models that reuse one encoder across datasets. Yet TS2Vec offers a compact convolutional architecture that operates directly on temporal samples and accepts variable-length inputs without resampling. We ask whether these properties can be retained when the encoder is pre-trained only once, and show that they can: TS2Vec can essentially be turned into a foundation model. Three changes suffice: pre-train on a synthetic corpus, feed the encoder a raw amplitude view and an instance-normalized shape view of the input through two small stems, and distill one specialist encoder per view into a single encoder. The result, TS2Vec-FM, is a 1.6M-parameter dilated convolutional network that, kept frozen, reaches 83.3% on UCR-128 and 75.3% on UEA-27 under linear probing, within 0.3 pp of MantisV2, the current state of the art, and above TS2Vec trained separately on each target dataset. What we find most interesting is how different TS2Vec-FM is from other foundation models in design: near-parity is reached without attention, without a patch tokenizer, and without intermediate-layer extraction, using an “old-school” convolutional encoder and a two-layer MLP stem.
[AI-5] Evolving Agent Interfaces
链接: https://arxiv.org/abs/2610.28601
作者: Saaket Agashe,Yuzhe Yang,Gonzalo Gonzalez-Pumariega,Bhuvan Chandra Koduru,Chih-Lun Li,Jiachen Yang,Ang Li,Xin Eric Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Computer-use agents are limited by the interfaces humans give them. An agent may type a spreadsheet cell by cell when a single line of code would do, or print all two thousand lines of a file through a shell to find one function. Both the interfaces and the knowledge of when to use each are built by hand, which cannot keep pace with new applications, workflows, and models. We argue that interface design should itself be an agent capability. We introduce Agent S4, an autonomous computer-use scientist that builds Evolving Agent Interfaces from task experience. A meta-agent examines task agents’ trajectories on synthetic training tasks, without success labels, and constructs and revises interfaces and usage knowledge in a repository. Population-based evolution then selects which repositories to develop further by their performance on validation tasks the meta-agent never sees, favoring changes that generalize beyond the seen tasks. On OSWorld V2, Agent S4, with Qwen3.8-27B and the evolved repository of interfaces, sets a new state of the art in open-source computer-use agents, reaching 21.3% binary score, 3.8 times the same model with fixed interfaces. This is above both GPT-5.5 (13.0%) and Claude Opus 4.8 (20.6%), at about 4.3% and 1.5% of their respective costs. On ScienceBoard, the evolved interfaces improve the success rate from 41.4% to 58.6% with the same model. These results suggest that useful interfaces can be evolved from experience, making autonomous interface design a source of improvement complementary to advances in models and harnesses. We open-source our code at this https URL.
[AI-6] What Do RTL Repair Benchmarks Actually Measure?
链接: https://arxiv.org/abs/2610.28595
作者: Yeyin Eva Zhu,Rishabh Ranawat
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:RTL repair benchmarks should measure whether a model can fix broken RTL given a specification. To this end, we construct and hash-bind RTLRepairDataset, 5,784 (broken, fixed) training pairs; RTLRepairBench, 459 held-out cases with 374 syntactic and 85 behavioral mutations; and an evaluation harness supporting simulation and formal equivalence. We then use this infrastructure to identify two measurement confounds. First, simulation and formal scoring give opposite signs for the apparent effect of providing a specification in one evaluated cohort; reset-aware rescoring attributes almost all of the simulation-formal disagreement to reset and first-cycle comparison semantics. Second, on 274 mined model failures, the specification-only minus specification-conditioned recovery point estimate ranges from -8.5 to +18.3 percentage points across six models. Thus, a conventional repair score need not isolate whether the model used the faulty RTL. We translate these findings into evaluation controls: report initialization, reset, first-cycle, and unknown-value semantics; disclose oracle coverage and undecided cases; and include a specification-only baseline. Together, the resources and audit provide both a benchmark for behavioral repair and a protocol for diagnosing what contributes to its measured score. We release the benchmark, the training set, construction code, sanitized aggregate evidence, and reproduction materials at this https URL.
[AI-7] Beyond the Sequentiality Bottleneck: A Hybrid Diffusion-Autoregressive Framework for Efficient Multimodal Reasoning
链接: https://arxiv.org/abs/2610.28584
作者: Wei Cai,Jian Zhao,Yuchen Yuan,Xuelong Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Autoregressive models (ARMs) face a “Sequentiality Bottleneck” in reasoning, while diffusion language models (DLMs) offer parallelism but lack semantic coherence. To addresses both limitations, we propose Collaborative Reasoning Asymmetric Learning (CoRAL), a hybrid framework via an Asymmetric Planner-Executor (APE) architecture: a DLM Planner generates K diverse reasoning trajectories in parallel, while a large ARM Executor selects the optimal plan and synthesizes the final answer. To train the Planner effectively, we introduce Reward-Guided Denoising for CoT (RDC), which integrates outcome-based rewards into the diffusion process to steer it toward goal-directed reasoning. Experiments on vision-language benchmarks show that CoRAL significantly outperforms prior diffusion-based methods and achieves competitive performance against strong ARMs, achieving both superior efficiency and quality.
[AI-8] Sage-A1.5: From Long-Horizon Execution to Scientific Exploration with Capability Composition
链接: https://arxiv.org/abs/2610.28578
作者: Wenjie Lou,Tianshuo Peng,Shiyang Feng,Shengji Tang,Runmin Ma,Ruizhe Chen,Jinxin Shi,Haoyang Peng,Fangchen Yu,Zhijie Zhong,Yuhao Zhou,Yue Fan,Yixin Chen,Yiming Wu,Yiheng Wang,Yigeng Jiang,Yichen Jiang,Yi Xie,Xu Guo,Xiaohan He,Tao Hu,Songtao Huang,Siyu Zhou,Shuaiyu Zhang,Shangheng Du,Qinglong Cao,Kun Li,Jingqi Ye,Hongbin Zhou,Guohang Yan,Zongsheng Cao,Ziyi Wang,Zhuo Liu,Zhiyin Yu,Zhiwang Zhou,Zelin Tan,Yuqing Zhu,Yaoze Zhang,Yang Chen,Xuming He,Xiuliang Liang,Xingjian Guo,Xiang Zhuang,Weiping Shen,Weida Wang,Tianyuan Zhou,Tianlin Ye,Tengyuan Fan,Siyu Yan,Shuo Li,Shuchen Pu,Shengyuan Xu,Shao Zhang,Ran Wei,Qinyuan Qin,Qingyu Yang,Nianchen Deng,Meng Dai,Kuikun Liu,Kaitao Chen,Juncheng Xiao,Jiangning Liu,Jianbiao Mei,Jia Yu,Haonan He,Hangfan Zhang,Cheng Yang,Chen Zhang,Bihao Zhan,Benteng Cheng,Yuqiang Li,Yihao Liu,Yang Yang,Xujun Fu,Xiaosong Wang,Xiangyu Yue,Wenwei Zhang,Shuyue Hu,Pinlong Cai,Mianxin Liu,Lei Bai,Jie Zhou,Jie Hou,Jiamin Wu,Hui Wang,Haijun Li,Guangyu Wang,Dongzhan Zhou,Chunfeng Song,Chao Yang,Biqing Qi,Xing Xie,Dahua Lin,Bowen Zhou,Xiangchao Yan,Wenlong Zhang,Shufei Zhang,Peng Ye,Bo Zhang
类目: Artificial Intelligence (cs.AI)
备注: Token-plan for Sage-A1.5: this https URL model checkpoints: this https URL
Abstract:In this report, we introduce Sage-A1.5, a 35B Mixture-of-Experts agentic model that builds on A1 to advance long-horizon problem solving through adaptive capability composition. Compared with A1, Sage-A1.5 learns not only to solve individual tasks, but also to organize work and revise its approach during execution. Sage-A1.5 highlights three improvements: agentic orchestration, verification, and open scientific discovery. First, agentic orchestration extends single-agent execution through task delegation and parallel collaboration among subagents with isolated contexts, enabling complex tasks to be decomposed and addressed concurrently and supporting efficient execution while reducing interference across concurrent subagents. Second, agentic verification strengthens long-horizon reliability through progress checking, evidence validation, and reflective strategy revision, allowing the agent to recognize unproductive search paths and pursue alternative approaches. Third, open scientific discovery extends agentic capabilities beyond answering predefined questions to conducting exploratory research, connecting scientific inquiry with interactive analysis and inspectable research artifacts. We showcase these capabilities through complicated agentic tasks and multidisciplinary scientific applications. Besides, we hope Sage-A1.5 provides the community with an accessible foundation for open scientific discovery, offering a practical path from long-horizon task execution to collaborative, verifiable, and inspectable scientific exploration.
[AI-9] OptScientist: Automated Discovery of Optimization Laws
链接: https://arxiv.org/abs/2610.28562
作者: Ma ZhenYao,Yue Liang,Jian-Qiao Zhu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Symbolic Computation (cs.SC); Econometrics (econ.EM); Optimization and Control (math.OC)
备注: Project website: this https URL
Abstract:Many scientific phenomena are best understood as outcomes of underlying optimization problems, yet automated scientific discovery has largely focused on recovering equations that describe observed input-output patterns. We introduce OptScientist, a novel automated framework for discovering compact symbolic optimization laws directly from data. OptScientist comprises (i) a model that infers symbolic optimization problems by first learning to represent data as an optimization problem and then extracting its symbolic form, and (ii) an LLM agentic system embedding the model in a discovery process that integrates theoretical priors from scientific literature, discovers candidate laws, abstracts their underlying mechanisms, and suggests new data to collect. We first evaluate OptScientist on OptFeynman, a novel benchmark of 99 physics-derived optimization problems, where OptScientist recovers ground-truth laws for over 80% of the problems and remains robust to noisy contextual information, while substantially outperforming machine-learning, symbolic, and LLM-based model-discovery baselines. On human risky choice and central-bank monetary policy, OptScientist discovers compact symbolic optimization laws that achieve held-out predictive performance competitive with leading baselines, while using few parameters and substantially fewer tokens than LLM baselines. Overall, OptScientist extends automated scientific discovery to uncovering optimization laws that generate observed phenomena, enabling automated theory building across the physical, biological, behavioral, and social sciences.
[AI-10] Agent ic Ensemble with Vision-Language Models for Visual Anomaly Detection
链接: https://arxiv.org/abs/2610.28560
作者: Shangquan Sun,Wenjun Luo,Xu He,Yoonsoo Lee,Shijian Lu
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 2 figures, 13 tables
Abstract:Visual anomaly detection aims to detect abnormal objects, while anomaly samples are often scarce during model training. Assembling multiple pretrained detectors can help by leveraging their transferable image scores and anomaly maps. However, most existing ensemble methods apply fixed fusion rules over saved predictions which struggle when pretrained detectors clearly disagree due to different training objectives, priors, and visual representations. To this end, we design AnomalyAgent, a post-training fusion framework for accurate visual anomaly detection. AnomalyAgent aligns heterogeneous scores and maps, summarizes score- and region-level disagreement, and treats fusion as a finite action-selection problem. It implements two types of actions: 1) score actions that determine how to fuse image-level detector confidence via consensus, robust aggregation, detector weighting, or score-map consistency; 2) map actions that fuse dense anomaly responses via consensus, detector weighting, geometric aggregation, or disagreement correction. A constrained vision-language model (VLM) is introduced to inspect high-conflict samples and select valid score action and map action to produce the final prediction. We evaluate AnomalyAgent on MVTec AD and VisA as primary benchmarks, together with seven additional public anomaly benchmarks and Real-IAD for large-scale score-level evaluation. Across these benchmarks, AnomalyAgent consistently improves over equal-weight fusion and achieves the best mean image AUROC and pixel AP among the compared fusion methods on MVTec AD and VisA, while demonstrating robust transfer across additional anomaly domains.
[AI-11] When Do LLM s Help with Feature Selection? A Systematic Empirical Study
链接: https://arxiv.org/abs/2610.28556
作者: Fateme Memar,Dongjie Wang,Arvin Agah
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 8 pages, 6 figures, 2 tables
Abstract:Large language models (LLMs) can leverage semantic feature names alongside summary statistics for tabular feature selection, yet it remains unclear when such information improves subset selection relative to established statistical methods. We investigate this question within an 18-dataset framework comprising 10 datasets for Group Relative Policy Optimization (GRPO) training, 2 for checkpoint validation, and 6 held-out evaluation datasets. We compare five classical baselines, frozen Qwen2.5-7B and Llama-3.3-70B selectors, and two GRPO-adapted Qwen2.5-7B models, reporting classification ROC-AUC and regression R^2 separately and using dataset-wise ranks for cross-task comparison. Across five seeded evaluation partitions, classical selectors generally achieve higher predictive performance than the tested 7B LLM/RL selectors, including under a pre-specified ~33% feature budget with strict train-only classical fitting, while the 7B selectors retain compact subsets. A cardinality-matched semantic sensitivity experiment shows that, for Zero-shot Qwen-7B, informative feature names improve regression performance over anonymized and shuffled alternatives (+0.1964 and +0.1606 mean R^2), whereas aggregate classification differences are smaller (+0.0159 and +0.0084 ROC-AUC) and dataset-dependent. These results characterize when semantic information benefits LLM-based feature selection and show that classical selectors remain strong baselines for predictive tabular feature selection.
[AI-12] Learning Computational Structure for Sample-Efficient Prediction
链接: https://arxiv.org/abs/2610.28555
作者: Hangyu Zhou,Constantine Dovrolis,Stephen Mussmann
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 36 pages
Abstract:Statistical learning in high dimensions suffers from the curse of dimensionality. Fortunately, many functions of practical interest are compositionally sparse, meaning they admit a composition of local functions, each depending on only a few variables, allowing sample complexity to depend on local arity rather than ambient dimension. Yet realizing this benefit requires exploiting hidden compositional structure, which theory does not establish for practical gradient-based training. We study this gap with CSFRBench, a controlled benchmark that hides the target’s compositional computation during training but retains it for diagnosis, and find that dense networks and sparsity-based baselines leave much of that benefit unrealized. Inspired by the spatial organization of biological neural circuits, we introduce Learned Wiring (LeWi): each neuron has a learned low-dimensional embedding, only nearby neurons connect, and embeddings and weights are trained jointly. LeWi improves sample efficiency on targets with hierarchical structure, intermediate reuse, or independent modules. Across 32 additional targets generated by independently sampled computation graphs, LeWi is the most sample-efficient method not given the target graph, with 27% lower mean learning-curve area than dense networks. Dense networks perform better on targets with a single high-arity function. LeWi’s representations and connectivity align with the target computation. On 17 world-model prediction tasks spanning control, navigation, and manipulation, LeWi beats the dense baseline on 16 and wins 99/119 baseline comparisons. These results show that jointly learning connectivity and weights can exploit computational structure to improve sample efficiency.
[AI-13] Humans Correct Their Judgments Through Debate but Weaker AI Models Do Not
链接: https://arxiv.org/abs/2610.28539
作者: Maria Victoria Carro,Denise Alejandra Mester,Facundo Nieto,Oscar Agustin Stanchi,Guido Ernesto Bergman,Giovanni Franco Gabriel Marrafini,Trinidad Borrell,Guido Freire,Joaquin Salvador Machulsky,Mario Alejandro Leiva,Ana Lopez,Nicolas Spinelli,Luca Nicolas Forziati Gangi,Eitan Sprejer,Federico Barrera Lemarchand,Gerardo Simari,Maria Vanina Martinez
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:How can we effectively oversee AI systems that surpass human intelligence? AI Safety via Debate proposes that two AI systems argue opposing positions and a judge, either a human or a weaker model, decides which is correct. By facilitating supervisors in providing high-quality training signals, debate aims to encourage truthful and safe behavior from AI systems whose capabilities exceed our own. However, human judgment is not neutral: people often rely on biases which advanced AI systems may learn to exploit. Drawing on insights from collective intelligence and social cognition, this paper investigates whether debate can guide human and LLM supervisors toward the truth despite their prior beliefs. We evaluate several protocol configurations. We also explore setups inspired by the wisdom of the crowd such as the number and diversity of judges, as well as interaction mechanisms including deliberation and voting. Overall, we find that debate increases accuracy relative to prior beliefs for human judges, whereas the consultancy baseline, in which a single expert argues for one answer, does not. For weaker language models, however, the results are mixed: debate provides no comparable gains in most settings and can even reduce accuracy. Among multi-judge protocols, weighted voting produces the highest accuracy for human supervisors, whereas for LLMs none of them significantly improves group accuracy, even as the number of judges increases. Notably, the behavior of minority judges in non-diverse deliberation protocols suggests that, under symmetric settings, incorrect positions could be more persuasive than correct ones. These findings support debate as a viable path toward scalable human oversight, but cast doubt on its benefits for weaker AI supervisors, raising concerns about fully automated training and control pipelines in high-stakes settings.
[AI-14] Beyond 4 bits: Data-Centric Post-training Quantization and Recovery
链接: https://arxiv.org/abs/2610.28537
作者: Rhea Sanjay Sukthanker,Riccardo Grazzi,James Hensman,Pashmina Cameron
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Post-training quantization (PTQ) compresses the weights of a trained model with limited compute and data. Down to 4 bits it is near-lossless, but at 2 bits scalar PTQ configurations fall to chance-level accuracy. We study three choices usually treated as implementation details that matter little at 4 bits but decide 2-bit quality: the quantization representation, the calibration data, and the recovery fine-tuning data, across six dense and mixture-of experts models (ranging from 3.8 to 32 B parameters). Representation is the largest factor and vector codebooks restore usable models where scalar PTQ pipelines collapse. Most of the remaining loss in accuracy is on \emphgenerative tasks, and the amount of performance drop is model-dependent. Furthermore, at a short recovery budget, calibration and recovery data are coupled. The best recovery data depends on the calibration data, so the two must be chosen jointly. The resulting recipe involves a vector codebook representation, a small task-aligned calibration mixture, and an inexpensive codebook-only recovery stage. Including codebooks, it uses about 2.3 effective bits per weight and retains 89 to 97% of full-precision question-answering (QA) accuracy where scalar PTQ is at chance, and, including recovery, uses 1.4 to 30\times less compute than the strongest quantization-aware-training (QAT) baseline.
[AI-15] DecisionPivot: A Diagnostic Benchmark for Decision Models in LLM Systems
链接: https://arxiv.org/abs/2610.28522
作者: Xialie Zhuang,Mingcong Lei,Zunhai Su,Ri Liu,Jianjin Li,Peixian Ma,Jia Pan,Ngai Wong,Guanhua Chen,Chuan-Wei Kuo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Decision models support routing, guardrails, and output verification in LLM systems. However, item accuracy alone cannot establish whether models make correct decisions across controlled variations or assign reliable confidence to their predictions. We introduce DecisionPivot, a diagnostic benchmark comprising 48 tasks and 40,033 items across seven decision roles. Its core track contains 7,800 items from 13 generated task families, organized into contrast groups with computed answers verified by independently implemented solvers. The complementary reference track contains 32,233 items from 35 public test sets, preserving their original evaluation metrics. We evaluate 40 decision models and 14 general models using documented, model-specific protocols, assessing item and contrast-group accuracy, probability ordering, calibration, and empirical selective coverage. Our results reveal substantial differences obscured by aggregate scores: two 27B models separated by only 2.6 points on the reference track achieve contrast-group accuracies of 70.1% and 44.2%. Across 2,400 binary contrast pairs, Perplexity Decider correctly ranks the positive member above the negative member in 92.3% of pairs but answers both members correctly in only 74.1%, exposing a gap between probability ordering and decision correctness. Models with similar item accuracy also differ in the reliability of their high-confidence predictions. On a separate 1,296-item subset, GLM-5.3 achieves 93.5% contrast-group accuracy at a higher reasoning setting, compared with 81.5% at its lowest effort. By linking errors to pivotal details and difficulty levels, DecisionPivot provides fine-grained diagnostics of decision failures that aggregate evaluation alone cannot reveal.
[AI-16] ECG for laboratory values:A comprehensive physiological and technical analysis
链接: https://arxiv.org/abs/2610.28507
作者: Walentin Palanchuk
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 78 pages, 17 figures, source code under this https URL
Abstract:Laboratory measurements are essential for clinical decision making but typically require invasive blood sampling. Recent research suggests that ECG recordings may contain information beyond cardiac rhythm and provide information related to laboratory values. This thesis investigates NT-proBNP prediction from ECG recordings and patient metadata using a multimodal deep learning framework, while considering the physiological plausibility of ECG-based prediction. It also compares a 1-lead configuration using lead II with the corresponding 12-lead configuration. Additional experiments on glucose, hemoglobin, and potassium are presented in the appendix. Regression, binary classification, and three-class classification experiments were conducted using MIMIC-IV, with the three-class task based on clinically established thresholds. The 12-lead configuration consistently achieved better performance. Regression reached a mean absolute error of 534.98 pg/mL for 1-lead and 514.27 pg/mL for 12-lead. Binary classification achieved AUROCs of 0.849 and 0.876, respectively, while three-class classification achieved macro-averaged AUROCs of 0.738 and 0.752. The results indicate that precise NT-proBNP regression remains challenging, whereas broader concentration ranges can be classified more effectively. The findings are consistent with the physiological hypothesis that ECG characteristics and NT-proBNP concentrations reflect shared cardiovascular processes. ECG-based NT-proBNP prediction may therefore provide complementary information, particularly for classifying broader concentration ranges. However, laboratory measurements remain necessary, and validation on independent datasets is required. Comments: 78 pages, 17 figures, source code under this https URL Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2610.28507 [cs.LG] (or arXiv:2610.28507v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.28507 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-17] Model- and Layer-Level Closed-Loop Forecasting with SSM-Based Spiking Neural Networks
链接: https://arxiv.org/abs/2610.28495
作者: Erik Stolt,Sanja Karilanova,Ayca Ozcelikkale
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Deep State-Space Models (Deep SSMs) have shown promising performance and computational benefits in processing of time-series data with complex temporal dependencies. Spiking neural networks (SNNs) constitute a biologically inspired class of deep SSMs, where computing blocks communicate through spikes, discrete single-bit events rather than continuous activation functions. A particularly important task in time-series processing is forecasting, i.e. predicting future values of a sequence from past observations alone. However, how to best perform forecasting with SNNs is an open question, where the standard autoregressive nature of the data processing with SNNs requires values that are not available during the forecasting phase. In this article, we address this challenge. We explore two closed-loop forecasting strategies, in which the network’s own generated values are fed back into the network as if they were observed past values: (i) model-level and (ii) layer-level closed-loop forecasting. We evaluate these approaches on the ETT-small datasets, a standard benchmark for time-series forecasting, under both spike-based and continuous activation functions. We compare the proposed closed-loop approaches against state-of-the-art methods such as SpaceTime, as well as the zero-masking baseline in S4. Both of the proposed approaches perform comparably to or better than zero-masking with S4, and in most cases match SpaceTime. These findings highlight the trade-off between continuous and low-bit quantized activations, and demonstrate that the proposed Deep SSM architectures provide a competitive framework for forecasting.
[AI-18] Exploring Temporal Context-Aware Physical Grounding in Latent World Models
链接: https://arxiv.org/abs/2610.28483
作者: Yilei Man,Yang Chen,Jingcai Guo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Learning latent dynamics is fundamental to building action-conditioned world models for planning. Among existing families, Joint Embedding Predictive Architectures (JEPAs) offer a natural framework by modeling transitions directly in latent space rather than at the pixel level. However, two issues remain in JEPA-based world models: (i) physical grounding lacks explicit aggregation of temporal context and (ii) action planning always overly relies on generic sampling distributions. To this end, we propose \mathcalG -JEPA, which incorporates temporal context into physical grounding for latent world models and couples the learned models with prior-guided action planning. Specifically, during learning, we incorporate a shared causal temporal module that aggregates temporal context within observation and prediction sequences under physical state supervision, grounding both in task-relevant physical states. When planning, we further introduce action inference through stepwise conditioning of a generative action distribution on the current and goal latent representations together with their evolving discrepancy, yielding anchored proposals that guide model-based online refinement. Together, physical state supervision guides temporal representation learning, while the action prior guides search through the learned world model. Experiments demonstrate improved planning performance on manipulation tasks, including tasks with increasing temporal separation between initial and goal observations. Code is publicly available at this https URL.
[AI-19] Beyond Trajectory Statistics: Community-Consistent Human Mobility Generation via Evolving User-Location Graphs
链接: https://arxiv.org/abs/2610.28467
作者: Bo Liu,Tong Li,Jiapeng Zhang,Zhu Xiao,Zhuo Tang,Kenli Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Human mobility data are fundamental to Web-based location services and broader urban applications, and mobility generation provides synthetic trajectories when real data are costly or inaccessible. Although recent mobility generators have incorporated collective mobility patterns, they primarily focus on individual trajectories and aggregate population dynamics, leaving evolving community structures formed by user-location interactions insufficiently modeled. In this paper, we study community-consistent human mobility generation. To this end, we formulate mobility generation as the generation of dynamic user-location graphs to model community structures and propose MOTION, a Multi-agent framework for cOmmuniTy-consIstent human mObility geNeration. MOTION consists of user agents, location agents, and a global manager agent. MOTION first detects coherent communities from historical mobility using the Leiden algorithm and extracts community-conditioned interaction priors to guide the generation process. Guided by these priors, the user agents plan mobility, search destination candidates, and infer user movements from the individual view. In parallel, selectively activated location agents infer likely user arrivals from the community view. The manager coordinates both views, diagnoses interactions causing deviations, and attributes targeted reflection signals to revise user decisions, location attraction, and source-user selection. Experiments on two real-world datasets show that MOTION outperforms baselines by at least 4.05% across reported metrics of individual trajectories, population and flow distributions, and community and graph structures. Further analyses confirm its robustness, privacy preservation, and downstream utility. The code is available at this https URL.
[AI-20] Stratified Sparse Attention: Reducing the Sparse-to-Dense Gap at Extreme Sparsity
链接: https://arxiv.org/abs/2610.28441
作者: Sahil Joshi,Prithvi Dixit,Aryan Katiyar,Agniva Chowdhury,Joseph E. Gonzalez,Anshumali Shrivastava,Aditya Desai,Kumar Krishna Agrawal
类目: Artificial Intelligence (cs.AI)
备注: 26 pages, 5 figures, 13 tables
Abstract:Large Language Model inference over long contexts is increasingly bottlenecked by the cost of attending to a growing Key-Value (KV) cache. Sparse attention reduces this cost by accessing only a fraction of the cache, but existing methods face a trade-off between deterministic retrieval, which captures high-scoring tokens but discards residual attention mass, and sampling, which improves coverage at the cost of higher variance. Hybrid methods combine top- k retrieval with uniform sampling, yet still exhibit a substantial gap from dense attention at aggressive sparsity. We introduce \emphStratified Sparse Attention (SSA), a unified framework combining deterministic retrieval with score-guided sampling. SSA deterministically attends to high-scoring tokens while sampling the residual using proxy attention scores and correcting for their inclusion probabilities, thereby recovering attention mass beyond the retrieval set. SSA encompasses top- k , top- p , and hybrid top- k +sampling as special cases. Across RULER, LongBench, LOFT, and AIME, and five models ranging from 3B to 27B parameters, SSA consistently improves over the baselines at matched sparsity. On RULER-HARD with Ministral-3-8B, SSA improves accuracy over the strongest practical baseline from 81.8% to 84.2% at 32K and from 63.8% to 69.3% at 128K under 50\times attention sparsity, recovering 94% and 85% of dense-attention performance, respectively. With optimized CUDA kernels, SSA achieves up to 2.96\times decode speedup over FlashInfer at 1M context, demonstrating that its accuracy gains translate into practical long-context inference acceleration. We open-source our code at this https URL.
[AI-21] Plausible Future Wrong Choice: Steering World-Action Models toward Intended Future
链接: https://arxiv.org/abs/2610.28431
作者: Zhuohao Li,Jiawei He,Yuxuan Yan,Lang Zhou,Yinghao Li,Yanrui Li,Tianyu Zhang,Yi-Yang Song,Xirui Wu andYi-Lin Wei,Fating Hong,Qing Zhang,Zhilin Zhao,Jingdong Wang,Wei-Shi Zheng
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:World-Action Models (WAMs) have emerged as a promising paradigm for embodied generalist policies by jointly modeling future world evolution and robot actions across diverse manipulation tasks. However, representing diverse futures is only the first step: a capable WAM must also generate futures that match the requirements of the current task and execution progress. In long-horizon manipulation where similar observation-action segments occur across different tasks and stages, existing approaches may generate locally plausible yet misdirected rollouts to drift toward wrong task stage, i.e., a phenomenon we term Stage Confusion. To address this problem, we propose TaskWAM, a closed-loop, stage-aware framework that represents fine-grained task requirements with Mode Tokens and tracks execution progress with State Tokens, forming a structured task state to steer world–action generation toward intended future. We further introduce StageBench, a controlled multi-stage manipulation benchmark with parameterized stage variations and automated this http URL experiments on both simulation benchmarks and real-world tasks demonstrate that TaskWAM effectively reduces confusion and achieves competitive performance with state-of-the-art methods. Moreover, TaskWAM provides strong plug-and-play transfer across heterogeneous WAM architectures and robot embodiments, including single-arm robots, dual-arm grippers, dexterous hands, and humanoid robots.
[AI-22] A Digital Twin to Improve Cyber-Social Impact
链接: https://arxiv.org/abs/2610.28430
作者: Vita Santa Barletta,Danilo Caivano,Gabriel Cellammare,Andrea Dimauro,Pasquale Mesagna,Massimiliano Morga
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 16 pages, 5 figures
Abstract:The increasing complexity of Industrial Control Systems (ICS) and their integration with IT infrastructures has significantly expanded the cyber threat landscape. Traditional cybersecurity models, primarily based on the Confidentiality, Integrity, and Availability (CIA triad), fail to capture the broader societal and operational impacts of cyberattacks. This paper proposes a novel framework that integrates Digital Twin technology with a Cyber-Social Security model to simulate, analyze, and quantify cyber threats in critical infrastructures. Virtualized ICS environments built with Mininet and MiniCPS are used to model attack scenarios and assess their consequences across technical and social dimensions. At the core of the framework lies the Cyber-Social Vulnerability Scoring System (CSVSS), a quantitative metric that extends CVSS with a Social Impact dimension grounded in the Entity and Place pillars, producing a single, reproducible severity score that reflects real-world consequences. Crucially, the framework does not automate risk assessment but keeps the human analyst at its center: a Large Language Model (LLM) performs a first-pass analysis, mapping vulnerabilities and drafting risk reports, while the analyst, supported by explainable AI feedback, validates the findings and retains decision authority over critical cases. Experimental results demonstrate the framework’s ability to model and score cyber-social impact, while the human-in-the-loop design is presented as a conceptual principle to preserve oversight.
[AI-23] Racer: Hunting GPU Kernel Bugs with Symbolic Equivalence Checking
链接: https://arxiv.org/abs/2610.28429
作者: Tuan Truong Vu,Cheng Tan
类目: ymbolic Computation (cs.SC); Artificial Intelligence (cs.AI)
备注:
Abstract:GPU kernels are critical to the performance of today’s AI services, yet the same operators are rewritten repeatedly: for each GPU generation, numeric format, serving regime, and kernel language. Every rewrite must preserve the function computed by the implementation it replaces. Today, this obligation depends almost entirely on differential testing: run the new kernel and a trusted reference on randomly sampled tensors, then compare their outputs within a tolerance. This approach has blind spots that additional test cases cannot eliminate. For example, a logical bug may produce an error smaller than the tolerance required to accommodate a correct optimized kernel. We propose symbolic equivalence checking, which executes a kernel and its reference on symbolic inputs, turning each output cell into an exact algebraic expression, and then checks whether the paired expressions agree for all inputs. Because these expressions are exact, symbolic equivalence checking distinguishes floating-point error from logical bugs, while a single check covers the entire input space. We implement Racer, a PTX-level symbolic equivalence checking engine, with three techniques that make this approach practical for production kernels: (1) Shape Reduction to control checking cost, (2) a lockstep CTA scheduler to handle concurrency, and (3) idiom templates with exact numeric refutation for otherwise inconclusive results. Across 2,408 production kernels from 41 collection targets, Racer found 82 real bugs in 23 libraries, including 76 previously unknown bugs and 41 confirmed so far. Subjects: Symbolic Computation (cs.SC); Artificial Intelligence (cs.AI) Cite as: arXiv:2610.28429 [cs.SC] (or arXiv:2610.28429v1 [cs.SC] for this version) https://doi.org/10.48550/arXiv.2610.28429 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-24] After the Attacker Leaves: Benchmarking Carryover Attacks on LLM Agents EMNLP2026
链接: https://arxiv.org/abs/2610.28425
作者: Zhiyue Xu,Xukai Wang,Zhiyuan Huang,Mingjie Zhan,Jian Zhao,Xu-Yao Zhang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted by EMNLP 2026 Main Conference
Abstract:Modern LLM agents persist state across sessions via two channels: long-lived memory files silently prepended to the system prompt, and scheduled tasks waking the agent without user input. A one-time indirect prompt injection committed to either channel can re-fire across sessions long after the attacker leaves, threats we term carryover attacks. Existing prompt-injection benchmarks evaluate attacks only in the session where malicious content is first observed, leaving the gate governing when and how often a planted payload later re-fires unmeasured. We present CarryoverBench: 12 personal-assistant scenarios and 101 attack tasks across three trigger channels: Instant, same-turn IPI; Autonomous, heartbeat-scheduled; and Latent, memory-resurfaced. A rule-based verifier scores planting and triggering as independent stages. Across five frontier models, planted payloads re-fire without measurable decay over five rounds; conditional on successful planting, the trigger stage often provides little additional protection. Porting the Latent channel to a second runtime, Hermes, attenuates but does not eliminate the effect, suggesting that carryover defense requires runtime-layer controls in addition to prompt-level interventions.
[AI-25] Rewards for Budget-Aware Multi-Turn Agents That Adapt to Arbitrary Test-Time Constraints
链接: https://arxiv.org/abs/2610.28409
作者: Tom A. Lamb,Adam Golinski,Eleonora Gualdoni,Sinead Williamson,Michael Kirchhof
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 32 pages, 8 figures
Abstract:Large language model agents are often given an inference budget and asked to maximise success within it. A multi-turn agent has to allocate this budget across many environment interactions. An external harness can enforce the budget, but it remains unclear whether the agent itself can learn when and how much to spend. We ask whether agents can learn this allocation across dynamically specified budgets and which rewards best teach this behaviour. We adapt existing length penalties to budget constraints spanning multiple resource types, introduce new reward formulations, and compare eleven objectives in a shared training pipeline. We train Qwen3.5 models at 4B and 9B on AppWorld with budgets over tokens, API calls, and environment steps, and on the BabyAI gridworld with turn budgets. We train a separate policy for each environment and resource type, with each policy trained across budget values. Training with rewards that penalise overspending substantially improves within-budget success and natural compliance over untrained models and over training on task success alone, with agents learning to adapt their resource use to stated budgets and task difficulty. These gains extend to unseen budgets and, zero-shot, to a shifted task distribution and a new benchmark. Reward objectives trade off success within budget and natural compliance. A sigmoid penalty for overspending most often achieves the highest within-budget success at an individual budget, while a fixed negative reward for overspending generally yields the strongest natural compliance. Combining penalties for overspend frequency and magnitude gives the smallest average gap to the best method at each budget. Overall, we show that a single agent can learn to allocate resources across turns and adapt its behaviour to user-specified budgets using budget feedback without external budget-allocation controllers.
[AI-26] DIVE: Benchmarking Full-Duplex Interaction with Visual Engagement in Continuous Streams
链接: https://arxiv.org/abs/2610.28405
作者: Yuhang Chen,Guangyue Liu,Shilong Liu
类目: Artificial Intelligence (cs.AI)
备注: Preprint. 19 pages, 4 figures
Abstract:Recent advances in multimodal large language models (MLLMs) have substantially improved spoken interaction and video understanding. Existing online and streaming benchmarks primarily assess causal multimodal perception through timestamped question answering, whereas real-world interaction requires models to continuously identify task-relevant multimodal cues and decide whether, when, and how to respond. This leaves open how well current models can sustain real-time interaction throughout an evolving multimodal stream. To address this gap, we introduce DIVE, a benchmark for evaluating full-duplex interaction with visual engagement under continuous and realistic multimodal streams. It organizes six tasks into three complementary interaction scenarios. (1) Proactive Intervention evaluates whether models can identify emerging visual or semantic cues and intervene at the appropriate moment. (2) Continuous Engagement evaluates sustained participation through continuous visual tracking and visually grounded query interaction. (3) Interaction Coordination evaluates how models coordinate listening and speaking under overlapping and multi-party interactions. DIVE contains 525 audio/video samples with fine-grained human annotations and precise temporal metadata. Its evaluation protocol separates response quality, participation decisions, and response timing. We evaluate representative real-time models using response quality, window response rates, and reaction latency. Experiments show that leading real-time models achieve sub-second average response latency, yet their average window response rates remain at or below 53.77%. Response quality varies substantially across models and tasks, with adapting ongoing responses to newly arriving information remaining particularly challenging. We hope DIVE facilitates progress toward more natural and capable real-time multimodal interactive systems.
[AI-27] RACE: RoPE-Aligned Attention Compression Based on Query-Key Activation Energies
链接: https://arxiv.org/abs/2610.28385
作者: Irene Tenison,Abdulrahman Alabdulkareem,Daniel Karl I. Weidele,Nima Dehmamy,Dan Gutfreund
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Head-dimension compression can reduce the memory required by the key-value (KV) cache, but existing methods often reconstruct full-dimensional representations during inference, adding significant decoding FLOPs overhead. We formalize the design considerations for converting pretrained attention into a natively reduced-dimensional form that reduces both KV-cache storage and attention computation using a calibration set. Specifically, we consider reductions that (1) apply a shared projection to queries and keys so that they remain in a common reduced feature space, (2) derive this projection jointly from query and key activations to better preserve their interaction, and (3) are compatible with the pretrained RoPE transformation. To obtain a reconstruction-free realization of this reduction, we restrict it to coordinate selection and theoretically show that RoPE compatibility then holds if and only if complete rotary pairs are retained. Building on this characterization, we introduce RACE, a post-training method that selects rotary pairs using their joint query-key activation energy. We theoretically show that this criterion minimizes a separable, calibration-dependent upper bound on post-RoPE query-key product approximation error. The resulting reductions are folded into the model weights and require neither runtime reconstruction nor specialized kernels. Across three model families and different compression levels, RACE outperforms structured pruning and remains competitive with reconstruction-based methods on language-modeling, zero-shot, and long-context tasks. At 50% KV-cache retention and 256K context, RACE reduces total inference FLOPs by 47.8%, yielding 1.79x higher throughput, 43.2% lower time-to-first-token (TTFT), and 28.1% lower allocated GPU memory in comparison to an uncompressed model.
[AI-28] Dont Skip the Baseline: A Controlled Comparison of Dynamic Expert Skipping in MoEs
链接: https://arxiv.org/abs/2610.28369
作者: Jakob Steimle,Thomas Elsken,Daniel Ehrhardt,Lukas Rinder,Daniel Cremers,Xi Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:A growing body of work proposes dynamic expert-skipping methods for Mixture-of-Experts (MoE) models that adjust the number of experts activated per token during inference, aiming to preserve model quality while further reducing the number of active parameters. These methods have added calibrated importance scores, inter-expert similarity metrics, or learned per-token allocators in order to identify experts to skip. However, different calibration regimes, different reductions in the effective number of experts, and evaluation suites make comparisons difficult. We classify four representative prior methods into a unified taxonomy and benchmark them against a natural, simple baseline: statically reducing the number of experts k (i.e., same reduced k across all layers and tokens). We provide a fair comparison of these methods at matched expert-compute budget, on two modern MoEs (Gemma4-26B-A4B and Qwen3-30B-A3B) in a text-only setting and a contemporary vision-language model (Qwen3-VL) in a multimodal setting. Surprisingly, we find that the simple baseline is hard to outperform consistently across models, modalities, benchmarks, and calibration settings. The methods we benchmark often underperform the baseline and achieved gains, when present, are minor. We further find that the effective k achieved (or optimized for) in calibration can diverge from the effective k in testing, resulting in over- or under-estimating the compute at inference time. Based on these findings, we recommend a calibration-free per-modality static k reduction and validate this on a held-out model and modality. We further argue that a static control at matched measured compute is the baseline any dynamic skipping method must beat.
[AI-29] Context-Aware Prompt Injection Detection via Small Language Models in RAG Pipelines: nokast-secureRAG
链接: https://arxiv.org/abs/2610.28364
作者: Abhishek Rai
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Local Retrieval-Augmented Generation (RAG) systems are vulnerable to indirect prompt injection and retrieval poisoning, where malicious instructions in retrieved content are mistaken for trusted input. We present nokast-secureRAG, a roughly 0.5B-parameter small language model that acts as a semantic firewall between retrieval and generation. It jointly evaluates the user query and retrieved context, classifying the pair as safe, suspicious, or malicious. The model is trained using a teacher-judge-student pipeline: a 35B teacher generates 5,233 examples across four attack families, and an independent 120B judge re-labels them, achieving 95.0% agreement before student fine-tuning. On a held-out in-distribution test set, the model achieves 0.994 detection recall at a 0.026 false-positive rate and a 0.006 attack-success-rate proxy, outperforming a regex baseline, a public prompt-injection classifier, Llama Guard 3 1B, and its untuned version. Prompt-Guard-86M reaches 0.900 recall but has a 0.690 false-positive rate. On a judgment-flip test, where the same sentence is benign or malicious depending on context, the model correctly classifies both cases 75% of the time, outperforming the compared baselines. Ablations indicate that retrieved context carries the main attack signal, while query-context consistency reduces false positives. Single-request latency is 293-404 ms, above the target of 100 ms, while amortized throughput latency is 37 ms under batching. On the external PoisonedRAG benchmark, recall falls to near zero, indicating that fluent fact-poisoning attacks lie outside the detector’s instruction-consistency threat model. We report this limitation alongside the in-distribution results. The model, data-generation pipeline, and evaluation harness are released open source.
[AI-30] FALCA: Falsification-Aware Reinforcement Learning with Hierarchical Credit Assignment for Long-Horizon Deep Search Agents
链接: https://arxiv.org/abs/2610.28361
作者: Zijun Dou,Kun Luo,Chaofan Li,Shuqi Lu,Zheng Liu
类目: Artificial Intelligence (cs.AI)
备注: 19 pages, 20 figures, 2 tables. Project page: this https URL
Abstract:Long-horizon deep search requires LLM agents to iteratively decompose complex information needs, gather evidence from external sources, and maintain a reliable research state under multiple constraints. A central challenge is that early mistakes can be amplified over the search trajectory: agents may anchor on an incorrect candidate, collect partially supporting evidence, overlook contradictions, and accumulate errors that misguide subsequent reasoning. We propose FALCA, a falsification-aware framework that formulates deep search as a propose-falsify process. FALCA adopts a collaborative Searcher-Falsifier architecture, where the Searcher proposes candidate answers or sub-answers, while independent Falsifiers actively seek counterevidence, check unmet constraints, identify conflicts, and produce concise falsification reports for research-state revision. To strengthen this loop, we introduce falsification-aware training, combining SFT warmup on falsification trajectories with agentic RL based on parallel rollouts and hierarchical credit assignment over research stages, state transitions, and actions. Experiments on five search benchmarks show that FALCA improves long-horizon deep search performance over backbone and harness ablations, while reducing confirmation bias and error accumulation through more effective falsification and research-state revision.
[AI-31] Identifying Dependent Components from Multi-domain Linear Mixtures
链接: https://arxiv.org/abs/2610.28330
作者: Danru Xu,Lauri Parkkonen,Sara Magliacane,Aapo Hyvärinen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 29 pages, 3 figures
Abstract:We study a linear mixing model with dependent latent components, assuming multiple data domains. Most existing models assume that the components are independent or at least uncorrelated, in line with independent component analysis (ICA). Some recent work allows for dependent components, but then makes specific assumptions such as parametric forms of dependencies, multi-view settings, or interventions, or does not recover the individual components. In contrast, we consider a multi-domain setting in which domains differ through domain-specific scalings of the components, while the distribution of the underlying latent components is the same across domains. This approach can model data collected, for example, from different sensors measuring the same process, different laboratories conducting the same experiment, different experimental conditions, or different subjects that might differ in biological or physiological factors. We show that, under sufficient domain variability, latent variables and mixing functions can be identified from second-order statistics alone. We propose the Multi-Domain Covariance Matching (MuDo-CoM) algorithm that generalizes previous methods of joint diagonalization. MuDo-CoM is validated on simulated data and a real-world fMRI dataset.
[AI-32] Is Sycophancy a Form of In-Context Updating?
链接: https://arxiv.org/abs/2610.28325
作者: Itai Shapira,Myra Cheng,Hadas Orgad
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Factual sycophancy, a language model’s deference to a user’s false beliefs, is usually understood as a bid for the user’s approval at the expense of truthfulness. We argue that this framing over-emphasizes the user and propose viewing sycophancy as a failure within the broader and generally useful process of in-context updating, by which models revise their answers based on information supplied during an interaction. We support this view with behavioral and mechanistic evidence. First, models’ responses to user claims sit within a broader pattern of source-sensitive updating. They revise their answers in response to information attributed to external sources as well, more so when it appears better supported, and their deference to the user persists even when no user approval is at stake. Second, the internal components that support deference to the user are also causally important for in-context updating more broadly. Our results suggest that factual sycophancy arises when in-context updating gives a user’s assertion excessive weight relative to its evidence. Because sycophancy is embedded in this process, reducing it risks reducing useful updating as well, a trade-off that calls for evaluations measuring both behaviors and mitigations that decouple them.
[AI-33] PO: Tri-level Distributionally Robust Learning for OOD Direct Preference Optimization NEURIPS2026
链接: https://arxiv.org/abs/2610.28229
作者: Chengtao Jian,Kai Yang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026
Abstract:Direct Preference Optimization (DPO) minimizes empirical risk over observed preference comparisons, implicitly assuming that the training preference distribution matches the deployment preference distribution. In practice, this i.i.d. assumption often fails, as preference distribution shift exhibits a natural hierarchical structure across three levels: group-level shift, prompt-level shift, and preference-level shift. Such multi-level distributional mismatch poses a critical challenge for safe LLM deployment, yet existing robust DPO methods typically either model uncertainty over the joint preference distribution or focus on individual sources of shift, leaving the hierarchical structure of OOD preference shifts underexplored. We propose TPO, a tri-level distributionally robust framework for OOD preference alignment, which jointly models all three levels of distributional uncertainty through a hierarchically structured rectangular ambiguity set. By instantiating it with Wasserstein distance, we derive a tractable and interpretable objective that combines a closed-form preference-level margin penalty, prompt-level adversarial embedding perturbation, and worst-case group reweighting, each governed by an independent radius. We establish finite-sample convergence guarantees and empirically validate the effectiveness of TPO over existing robust DPO methods.
[AI-34] Local Inference of LLM s on Long Contexts: A Data Systems Perspective
链接: https://arxiv.org/abs/2610.28223
作者: Mattias Dürrmeier,Alberto Lerner
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models can solve complex problems across a variety of subjects, such as mathematics and coding. For such complex analytical tasks, models with a long context perform better. Serving models with long context is currently mostly done on powerful systems with multiple GPUs, which are too expensive for most users. An alternative is local inference, or serving a model on a user’s own machine, which removes per-token costs and rate limits, besides keeping sensitive data private and giving users full control over the models they run. The common understanding, however, is that local LLMs need compression (e.g., quantization) to run on a single GPU. In this paper, we set out to study the limitations that prevent serving long-context models for single users, besides the capacity issue. Rather than characterizing a single model on one GPU, we analyze how different models behave across several GPU generations. This lets us uncover not just individual bottlenecks but broader trends in where inefficiencies arise. Our analysis confirms that, as expected, self-attention dominates the runtime for long-context inference. More unexpectedly, our evaluation reveals that the current structure of self-attention leaves key GPU resources significantly underutilized, and the newest GPU is often the least well served. Many of these findings are fundamentally data management problems, from paging the KV cache to improving cache utilization and synchronization in the GPU, which is why we bring this study to the Data Systems community. To engage this community, we present our findings as a series of experimental observations and formulate recommendations for future work.
[AI-35] JAXbounce: A Continuous-Control Pong-Like Benchmark for Active Inference Beyond Low Dimensional Environments
链接: https://arxiv.org/abs/2610.28203
作者: Chaitanya Kaul,Markus Klar,Sebastian Stein,Fraser C. Paterson,John H. Williamson,Roderick Murray-Smith
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Published in the Springer CCIS series as part of the 7th International Workshop on Active Inference (IWAI26)
Abstract:We introduce JAXbounce, a continuous control pong-like benchmark designed to evaluate active inference beyond discrete, low dimensional environments. JAXbounce combines continuous states and stochastic dynamics in discrete time with flexible and composable observation models that allow representing the environment as a state vector or a pixel grid, corruptible through the rich family of noise models we provide. It is fully differentiable and parallelisable in JAX. We run over 10,000 agent simulations on our environment comparing a baseline deterministic solution, to a state-based active inference agent with Unscented Kalman Filter perception and expected free energy planning, and a pixel based agent whose measurement model is a pretrained neural network encoder. All agents are evaluated across three scenarios of increasing difficulty with eight different noise conditions. Our experiments characterise how performance of continuous state agents scales with planning, planning horizons, observation modalities and knowledge of the noise process, exhibiting preference-oriented behaviour even in the presence of increasing uncertainty. We release JAXbounce as a reproducible benchmark for continuous control active inference research.
[AI-36] MaD-RSI: Model-as-Data for Recursive Self-Improvement of Agent Harnesses
链接: https://arxiv.org/abs/2610.28201
作者: Yu Cui,Hong He,Ruiqing Yue,Sicheng Pan,Zhuoyu Sun,Yifei Liu,Ting Li,Wenzhuo Zhu,Baohan Huang,Haibin Zhang,Cong Zuo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recursive self-improvement (RSI) of agent harnesses has emerged as an effective approach to enhancing the capabilities of large language model (LLM) agents. By searching the harness space for a specific LLM on a given dataset, RSI can discover a harness tailored to the capabilities and behavioral characteristics of that model. However, such optimization often creates a strong coupling between the learned harness and its target LLM, limiting generalization across models and requiring substantial computational resources to adapt to new LLMs. Moreover, existing RSI methods typically optimize a harness using feedback from a single model and evaluate its generalization to other models only after training. This paradigm treats model differences primarily as a criterion for evaluating generalization rather than as valuable evidence for optimization. In this paper, we introduce a new perspective that treats models as data and places the harness at the center of optimization. Building on this perspective, we propose \textscMaD-RSI, a framework that directly optimizes a shared harness across multiple heterogeneous LLMs by incorporating their diverse failure trajectories into the training process. Rather than assessing model differences only after optimization, \textscMaD-RSI leverages these differences as training evidence to guide the self-improvement of a shared harness. Furthermore, we introduce a new evaluation framework with two complementary metrics designed to assess the effectiveness of harness optimization across models. Experiments across multiple datasets and LLMs demonstrate that \textscMaD-RSI improves generalization over existing approaches while reducing training costs.
[AI-37] Parameter-Efficient Adaptation of Time Series Foundation Models for Electrical Load Forecasting
链接: https://arxiv.org/abs/2610.28200
作者: Aliullah,Julian Oelhaf,Edouard Fouché,Jessica Deuschel,Christian Bergler,Andreas Maier,Siming Bayer
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted for presentation at the IEEE Power Energy Student Summit (PESS 2026), Karlsruhe, Germany. 6 pages, 4 figures, 1 table
Abstract:Accurate electrical load forecasting supports power-system operation across countries, regions, cities, and buildings, but repeated development of target-specific models is costly. Time series foundation models (TSFMs) provide reusable pretrained backbones, although multi-domain pretraining may not fully capture load-specific characteristics. This paper compares pretrained inference, full fine-tuning, and low-rank adaptation (LoRA) in terms of forecasting accuracy and adaptation efficiency across four electrical load aggregation levels. MOIRAI 1.0 Large serves as the primary backbone, while Chronos-2 assesses whether the findings transfer across model families. The evaluation covers both in-distribution (ID) and out-of-distribution (OOD) settings and examines whether compact adaptation provides consistent gains over strong pretrained checkpoints. Pretrained foundation models outperform the implemented seasonal naive and XGBoost baselines across the evaluated aggregation levels. On MOIRAI Large, rank-16 LoRA reduces normalized mean absolute error (NMAE) by 7.5% on average while updating 3.146 million parameters, about 1% of the model, and retains most of the improvement from full fine-tuning. Chronos-2 achieves lower pretrained NMAE than MOIRAI but benefits less from adaptation. OOD results show that adaptation benefits vary by target, with the clearest gains for the local-area series and several buildings. Overall, LoRA provides a favorable accuracy-efficiency trade-off when adaptation yields a meaningful gain over a strong pretrained checkpoint.
[AI-38] Reframing Object-Centric Reinforcement Learning: From Absolute to Relative Representations
链接: https://arxiv.org/abs/2610.28197
作者: Cedric Derstroff,Jannis Blüml,Janosch Moos,Michael Girstl,Zihan Ye,Debora Clever,Kristian Kersting
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 13 pages, 7 figures, 10 pages Appendix, First two authors contributed equally
Abstract:Object-centric reinforcement learning agents typically receive each object’s position in a fixed global frame, even though their decisions usually depend on where objects are relative to one another. These absolute views often fail to capture the local context that is most relevant to the agent’s behavior. As a result, an agent must relearn the same response at every location where a configuration can occur, and its performance collapses when the scene is merely shifted. In this work, we propose a shift toward Object-Relative Coordinates fixed to Anchor (ORCA), where entities are described in a relative frame of reference. To operationalize this idea, we introduce Object-Relative inductive-Bias Transformers (ORBiT), a transformer-based architecture for processing anchor-relative inputs in reinforcement learning. Our results suggest these relative representations improve performance for both MLP and transformer agents. Furthermore, we demonstrate that ORBiT outperforms the pixel-based PPO baseline, achieving an aggregate HNS IQM greater than 1.0 over 13 Atari games. ORBiT learns the game of Foosball from scratch, dominating all baselines. In our controlled Tag environment, we further show that ORCA extends to rotations, that ORBiT generalizes to unseen numbers of distractors, and that these gains require a decision-relevant anchor.
[AI-39] Adaptive Pullback: Reweighting Supervision at Penalty Jumps in On-Policy Distillation
链接: https://arxiv.org/abs/2610.28194
作者: Guangyu Liu,Siyuan Gan,Shilong Zhou,Linjian Meng,Boyan Wang,Jing Huo
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:On-policy distillation (OPD) supervises student-generated responses using sampled-token student-teacher log-ratios. A positive log-ratio penalizes a token that the student assigns higher probability to than the teacher, but its current value does not distinguish sustained high penalties from abrupt increases along a response. We examine these local changes by comparing consecutive blocks of tokens within the same response. Our diagnostic identifies penalty jumps, abrupt increases in average penalty that persist on held-out tokens. Across these jumps, penalties rise while teacher confidence decreases. We call this pattern penalty-confidence tension. Boundaries into high-penalty blocks do not exhibit comparable confidence changes, showing that local changes provide information beyond the current penalty value. Motivated by this observation, we propose Adaptive Pullback OPD (AP-OPD). At each detected jump, AP-OPD uses the preceding block as a local reference and uniformly scales all supervision coefficients in the following block so that its weighted average penalty matches the preceding block’s original mean. This reference determines the scaling factor without an additional strength constant. AP-OPD improves performance on both mathematical reasoning and code generation.
[AI-40] Perceives but Doesnt Brake: A Perception-Action Binding Gap in Reasoning Driving VLAs
链接: https://arxiv.org/abs/2610.28184
作者: Matteo Cercola,Katie Z Luo,Simone Formentin,Marco Pavone
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Reasoning-augmented vision-language-action (VLA) driving models can perceive and reason about safety-critical hazards yet fail to translate that evidence into safe actions. We characterize this failure as a perception-action binding gap: hazard information is available inside the model but exerts insufficient control over trajectory generation. Using matched simulated counterfactuals and logged driving replay, we show that in-path hazards are actually encoded in the visual representation, but longitudinal plans are dominated by ego-speed history and respond only weakly to hazard proximity or the model’s own reasoning trace. Further analysis shows that the trajectory generator is capable of braking but sometimes fails to do so due to the perception-action binding gap. Without retraining or modifying any weights, we inject a hazard-conditioned energy into the frozen flow-matching sampler of Alpamayo-1.5, restoring the missing coupling at generation time. With exact obstacle geometry, guidance cuts footprint intrusion from 58.1% to 14.5% and recovers safe braking whenever sufficient stopping distance remains, eliminating collisions entirely in our braking-relevant closed-loop battery. It also reverses the diagnosed dependence of the plan, from tracking ego speed to responding to obstacle range. These results separate hazard perception from hazard control in driving VLAs, show that unsafe behavior can arise even when the model has both the evidence and the ability to act on it, and propose a simple fix for bridging this binding gap.
[AI-41] From Code Generator to Engineering Companion: A Plant-aware Agent for PLC Programming
链接: https://arxiv.org/abs/2610.28160
作者: Joschka Kersting,Gesa Benndorf
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Programmable Logic Controller (PLC) coding remains a manual, IDE-centric task that demands constant context switching between documentation, project files, and code editors. Recent work has demonstrated that retrieval-augmented large language models can generate vendor-compliant Structured Text at high compilation rates \citepred, yet these systems treat code generation as a one-shot pipeline: the engineer submits a query and receives code, with no persistent knowledge of the specific plant. We present a plant-aware conversational copilot that extends this paradigm in three directions: plant-specific context ingestion, a chat-first interaction model for explanation and debugging, and a graduated tool-access ladder from chat to deployment. The system is designed for real plant infrastructure and keeps the engineer in control of each step that produces real-world effects.
[AI-42] Causal Discovery for Linear Models with Latent Confounding via Multi-domain Identifiable Representation Learning
链接: https://arxiv.org/abs/2610.28134
作者: Danru Xu,Aapo Hyvärinen,Sara Magliacane
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 29 pages, 6 figures
Abstract:Causal discovery with latent confounding is particularly challenging, but a common setting in real-world data. In this work, we focus on linear causal models with latent confounding, represented by acyclic directed mixed graphs (ADMGs). Existing methods that fully identify ADMGs typically assume non-Gaussianity or graphical constraints, such as bow-freeness. In contrast, we propose Linear ADMG Learning (LinADMG), a framework for learning ADMGs for linear causal models through identifiable representation learning. Rather than relying on non-Gaussianity or bow-freeness, LinADMG leverages multi-domain settings to identify the noise variables. We show that the identifiability of the noise variables is sufficient for recovering all direct causal relations, and that, under an additional latent-confounding no-cancellation condition, LinADMG can identify the whole ADMG. We implement LinADMG with four identifiable representation learning approaches from the literature and validate the methods on simulated and semisynthetic datasets.
[AI-43] De-Diffusing WAMs: Action-Sufficient Direct Prediction for Robot Control
链接: https://arxiv.org/abs/2610.28122
作者: Shaoyi Zheng,Zhongtao Jiang,Yue Peng,Wenbo Lu,Shunchi Zhang,Xi Wang,Shengji Wan
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Cascade world-action models (WAMs) use predicted visual futures to condition robot actions, but inherit costly iterative inference from diffusion-pretrained experts. We ask what a single-step predicted future must preserve to support accurate single-step action prediction. In the evaluated WAM, resampling video noise has little effect on downstream actions, yet naively truncating both streams to one step substantially reduces task success. Video-space supervision only partially closes this gap, and a local sensitivity analysis shows that vision prediction errors have highly unequal consequences for action accuracy. These findings motivate DirectWAM, which converts pretrained video and action experts into single-step predictors without stochastic sampling through action-sufficient distillation, while retaining the explicit future-to-action cascade. After learning a one-step action model, we train the future predictor by decoding actions from its own predicted futures and backpropagating demonstration action error through the expert, with an optional future-latent reconstruction loss. The resulting objective evaluates the student-generated future through its downstream control consequences rather than relying on representation matching alone. DirectWAM achieves 97.1% average success across four LIBERO suites and 85.4% on RoboTwin 2.0, outperforming Flash-WAM at the same one-evaluation-per-stream budget. This advantage has also extended to three real-world tasks that we designed. Compared with the full iterative teacher, it reduces inference latency by 28.0\times .
[AI-44] Interruption-Aware Recurrent Reinforcement Learning for UAS Navigation Under Temporary Overrides
链接: https://arxiv.org/abs/2610.28121
作者: Christopher Conrad,Deepak Panda,Weisi Guo,Antonios Tsourdos
类目: Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: This work has been submitted to IEEE for possible publication
Abstract:Safe autonomous navigation of UASs requires tactical deconfliction to address conflicts that strategic coordination cannot prevent. This task is challenging under partial observability, when sensor or traffic data are missing, delayed, or degraded. Geometry-based methods can offer separation assurances but face limitations in dense encounters. Recurrent reinforcement learning has therefore been proposed to learn navigation policies that integrate observation-action histories to optimise long-term safety-efficiency trade-offs. However, operational safeguards may require a remote pilot, onboard detect-and-avoid system, or runtime assurance mechanism to temporarily override the learned policy. A research gap remains in understanding how these interruptions affect recurrent policies when the agent regains authority. Benchmark evidence shows that interruptions make task variables less decodable from the recurrent hidden state and increase post-interruption collision risk. The findings suggest that altered recurrent context prolongs post-handback degradation through an action-observation feedback loop, in which changed actions alter the flight path and subsequent recurrent updates. To address this vulnerability, we propose three interruption-aware architectures that condition recurrent processing on interruption context through separate feature streams, dedicated interruption memory, or gated environmental features. Across simulations, all three architectures preserve more decodable task information than a standard recurrent policy. Moreover, the strongest architecture reduces post-interruption collisions by 74% and interruption-induced mission failures by 69% relative to the recurrent baseline, without degrading interruption-free performance. These results show that safe interruption does not guarantee safe handback, and support interruption-aware recurrent design for resilient UAS autonomy.
[AI-45] Robust Representation Learning for Spiking Neural Networks based on Conditional Information Bottleneck
链接: https://arxiv.org/abs/2610.28110
作者: Yuli Zha,Wang Fan,Yixiao Dai,Lingjie Wang,Bang Hu,Wei Cao,Fengzhe Zhang,Fan Zhang
类目: Artificial Intelligence (cs.AI)
备注: 17 pages
Abstract:Spiking Neural Networks (SNNs) have shown great potential for low-power neuromorphic computing due to their event-driven processing and spatiotemporal sparsity, but they remain vulnerable to input perturbations, hardware-intrinsic noise, and adversarial attacks in real-world deployment. We attribute this vulnerability partly to input-specific hidden representations learned by standard SNNs trained with conventional cross-entropy loss, as our information-plane analysis empirically shows that substantial input-related information can still be retained after label-predictive information becomes stable. To address this issue, we propose SCIB, a simple yet effective robust learning objective for SNNs based on the Conditional Information Bottleneck. Different from conventional variational IB methods, SCIB replaces the mutual information compression term I(X;S) with the conditional mutual information I(X;S|Y) and derives a tractable variational upper bound. Moreover, it introduces a learnable class-conditional prior instead of a label-agnostic isotropic Gaussian prior, improving the expressiveness of variational compression and encouraging class-structured representations when combined with the supervised classification objective. Systematic experiments on Fashion-MNIST, CIFAR-10, and CIFAR-100 under external environmental noise, neuromorphic hardware noise, and adversarial attacks show that SCIB improves robustness over standard SNNs and IB-based baselines under most evaluated noise conditions, improves adversarial accuracy under the evaluated attacks without using adversarial training, and reduces the SynOps-based single-sample inference energy consumption by 24.9% relative to an unconstrained SNN baseline.
[AI-46] A Semi-Analytical Inference Framework for Structural Causal Models
链接: https://arxiv.org/abs/2610.28103
作者: Tatiana Boura,Robert Peharz
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Structural Causal Models (SCMs) provide a principled framework for observational, interventional, and counterfactual reasoning, all of which rely on probabilistic inference. Existing frameworks either transform SCMs into alternative tractable representations or rely on approximate inference with expressive generative models. We instead perform inference using a message passing scheme directly on the SCM by representing structural mechanisms and probability densities in a common finite-dimensional function space. Forward propagation combines a numerical realization of the Law of the Unconscious Statistician, while propagated densities are recovered by spline projection from LOTUS moments. The same representation naturally yields backward message passing for conditioning, providing a unified framework for causal inference. Using tensor-product B-spline bases, the resulting algorithm expresses inference through numerical quadrature and linear algebra without explicitly constructing conditional distributions. Experiments on randomly generated continuous SCMs demonstrate accurate approximation of Monte Carlo reference solutions and systematic improvement with increasing basis resolution.
[AI-47] Polyscene: Scenario Planning as Diversity Planning
链接: https://arxiv.org/abs/2610.28102
作者: Mustafa F. Abdelwahed,Joan Espasa,Felipe Meneguzzi
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Scenario planning is a foresight practice in which analysts construct a small set of plausible futures and prepare an organisation for each. Each scenario is a narrative that must be plausible, internally consistent, relevant and challenging, and the set as a whole must be differentiated. The practice remains manual: a workshop (i.e., a facilitated meeting in which a small group of experts writes the narratives) considers only the factors it has time for, carries the biases of the experts who run it, and cannot be redone whenever conditions change. In this paper, we automate scenario planning by casting it as a diversity planning problem, in which two scenarios differ when they end in different narratives on the analyst’s drivers. The result is a set of scenarios, one for each narrative that the topic can reach within a budget. Our implementation, Polyscene, is a generic engine that generates scenarios from a topic. We use geopolitical conflicts and shipping-route disruptions as case studies. For the first, we build the planning problem’s initial state from the GDELT Project so it reflects current relationships among the countries involved. Polyscene then generates scenarios showing how relationships among those countries change. With a board of six countries as of 1 January 2021 and a budget of five events, the conflict set puts five of the 15 country pairs at war, and each of the three wars that break out among these countries by 1 January 2026 ends a scenario. For the second, we build the initial state from ship traffic, a supply-pressure index, and the news, and Polyscene generates scenarios showing what a closed or diverted route could do to a manufacturer’s parts stock. With four routes per board and the same date and budget, each set includes three outcomes, and the Suez Canal set includes the March 2021 blockage and the December 2023 Red Sea container diversion.
[AI-48] ClawMMU: Protection-Coherent Virtual Memory for Stateful LLM Agents NEURIPS2026
链接: https://arxiv.org/abs/2610.28093
作者: Jinhao Hu,Bardia Mohammadi,Mofasshara Binte Rafique,Lars Henning Klein,Akhil Arora,Ashvin Goel,Laurent Bindschaedler
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Operating Systems (cs.OS)
备注: Accepted at the AgenticOS Workshop @ NeurIPS 2026 (non-archival). 6 pages, 1 figure, 1 table
Abstract:AI agents keep long-term memory, much of which comes from web pages and tool results. Agent harnesses, the software that runs the model, do not record who may act on this memory, so an instruction hidden in a fetched page can be treated as the user’s own instruction. We propose ClawMMU, a protection layer that labels every piece of memory with its owner, scope, origins, and authority. ClawMMU checks these labels whenever content enters the model’s context, is written to storage, returns from storage, or drives a tool call. Its central rule is that information may rise in authority, but authority never lifts itself. Because the checks depend on where content came from, not on what it says, memory attacks become classical protection faults that the harness stops even when the model is fooled.
[AI-49] From Complex Histories to Simple Predictors: Predictive Dimension in Dynamic Graph Learning
链接: https://arxiv.org/abs/2610.28090
作者: Hantong Feng,Duxin Chen,Wenwu Yu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Existing continuous-time graph learning methods rely heavily on complex neural architectures to encode ever-expanding interaction histories. However, the future conditional distribution induced by complex histories can often be characterized by a low-dimensional predictive state manifold, implying that only a few predictive degrees of freedom are essential for future interactions. We therefore argue that continuous-time graph learning should identify low-dimensional states sufficient for future prediction rather than reconstruct the full history. To make this perspective measurable, we define and independently estimate the intrinsic predictive dimension, which quantifies the effective predictive degrees of freedom required by a given task. Specifically, we remove the history-independent marginal from the history-conditioned future distribution, decompose the resulting predictive residual operator, and estimate the minimum effective rank needed to preserve the predictive power of the full history. Experiments across datasets show that model performance saturation is largely determined by the intrinsic predictive dimension. Based on this finding, we propose PreDyG, a lightweight model whose capacity is matched to the intrinsic predictive dimension. PreDyG achieves superior predictive performance with only 0.07% of the parameters of a representative complex baseline.
[AI-50] InsightRAG : Aggregating Distributed Scholarly Judgments for Scientific Survey Generation
链接: https://arxiv.org/abs/2610.28085
作者: Yuquan Lin,Wenshuo Zhang,Yilin Ye,Shuchang Xu,Zhao Chen,Weiqi Wang,Huamin Qu
类目: Artificial Intelligence (cs.AI); Digital Libraries (cs.DL)
备注:
Abstract:The rapid growth of scientific literature makes survey writing increasingly challenging, motivating automatic survey generation. Yet, existing systems must search an open space of papers and an open space of concepts with no trace of how the community itself has navigated them, leading to incomplete scope, weak taxonomies, and unreliable citations. Our core insight is that author-written Related Work sections contain the human judgments needed to address both limitations: authors curate the citations relevant to their contributions and explain how the cited works relate, yielding paths for scope reconstruction and relational evidence for organization. We propose InsightRAG. Its insight-guided retrieval follows author-supported citation relations to reconstruct a verified research scope and structure the resulting relational evidence. It aggregates these judgments under a minimum-conflict objective into a taxonomy that reflects recurring author-written perspectives in the retrieved literature, guides generation, and bounds each section’s evidence. Results on ten topics show gains in scope reconstruction, taxonomy quality, citation support, and reference reliability.
[AI-51] IECG: Weakly-Supervised Interpretability-Enhancing Foundation Model for Electrocardiogram WACV2027
链接: https://arxiv.org/abs/2610.28083
作者: Huynh Dang Nguyen,Trong-Thang Pham,Anh-Thuan Do,Phuong-Nhi Dinh,Brijesh Patel,Ngan Le,Van Nguyen
类目: Artificial Intelligence (cs.AI)
备注: Accepted in WACV 2027
Abstract:Electrocardiography (ECG) is a cornerstone of cardiac diagnosis, yet most machine learning methods for ECG analysis lack interpretability, limiting their clinical adoption. We present IECG, a weakly-supervised ECG foundation model that jointly performs diagnostic classification and abnormality localization without requiring manual abnormal region annotations, overcoming a long-standing bottleneck in interpretable ECG learning. Central to IECG is a weakly-supervised attention strategy that explicitly incorporates clinical priors by deriving abnormality heatmap pseudo-labels from established ECG clinical knowledge. These pseudo-labels guide the model’s CLS-token attention through a KL-divergence loss, encouraging it to focus on diagnostically relevant waveform regions. During pretraining, IECG is jointly optimized with clinically guided attention supervision, ECG-text contrastive alignment, and report captioning objectives, enabling it to learn rich multimodal representations while producing localized waveform evidence. By explicitly aggregating features from clinically guided attention regions, IECG grounds its diagnostic predictions in interpretable waveform evidence without sacrificing classification performance. Extensive experiments demonstrate that IECG achieves state-of-the-art abnormal region localization while maintaining strong classification performance, advancing trustworthy foundation models for ECG analysis.
[AI-52] PULSE: Personalized Opportunistic Learning for Wearable Devices
链接: https://arxiv.org/abs/2610.28081
作者: Mattia Giovanni Campana,Dariush Salami,Hüseyin Yiğitler,Stephan Sigg
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: Submitted to IEEE PerCom 2027
Abstract:Wearable devices increasingly support personalized health and well-being applications, but post-deployment data can differ substantially from those used for training, while purely local adaptation may be limited by the data available on a single device during real-world operation. Conventional federated and decentralized learning rely on recurrent communication and stable connectivity, whereas existing opportunistic approaches often exploit available encounters without adequately considering their utility, compatibility, or communication cost. We propose PULSE, an event-driven framework for personalized opportunistic learning that treats each encounter as an online collaboration decision, selecting either no interaction or a promising nearby peer. Peer gradients are incorporated only when compatible with the local learning direction, while past interactions guide future peer selection. Across human activity recognition, fall detection, and stress recognition under synthetic and real contact traces, PULSE preserves or improves personalized classification performance, while reducing peer-assisted exchanges by 80-94% compared with opportunistic model averaging. On-device experiments further demonstrate its feasibility on wearable hardware: peer-assisted operations require limited energy, while the complete implementation occupies 149.8 KiB of RAM and 561.8 KiB of flash. The obtained results show that selective opportunistic collaboration can provide effective personalization while limiting communication and resource demands on constrained wearable devices. Comments: Submitted to IEEE PerCom 2027 Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2610.28081 [cs.LG] (or arXiv:2610.28081v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.28081 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-53] rajEvolve: Trajectory Evolution for Reinforcement Learning with Hindsight Credit Assignment NEURIPS2026
链接: https://arxiv.org/abs/2610.28060
作者: Yi Wen,Hao Chen,Wanyu Wang,Maolin Wang,Pengyue Jia,Derong Xu,Huize Tan,Yingyi Zhang,Wenlin Zhang,Weihong Luo,Xiku Du,Xiangyu Zhao
类目: Artificial Intelligence (cs.AI)
备注: NeurIPS 2026 accepted paper
Abstract:The Reinforcement Learning (RL) paradigm has achieved remarkable success in enhancing the reasoning capabilities of Large Language Models (LLMs). However, when applied to long-horizon tasks, it faces severe challenges: i) Due to the high task complexity, the model struggles to sample enough successful trajectories, resulting in slow convergence. ii) Relying solely on sparse final rewards causes all steps to be updated indiscriminately, encouraging redundant step generation. When both the quantity and quality of successful trajectories become difficult to guarantee, the performance is severely constrained. This dilemma stems from the limitations: existing RL paradigms rely entirely on the reasoning capability of base models, lacking the ability to actively optimize trajectories. To address the above issues, we propose a novel trajectory evolution paradigm with hindsight credit assignment, termed TrajEvolve. TrajEvolve not only relies on the trajectory inferred by the model but also generates ``evolutionary trajectories’’ by refining low-quality steps in historical trajectories, thereby exposing the policy to high-quality compositions that lie within the base model’s support but are rarely sampled within practical training budgets. Furthermore, TrajEvolve reformulates step-level continuous credit assignment as a binary classification task, whether each step is necessary for task completion. This simplification not only reduces the learning difficulty but also enables step labels to be obtained through counterfactual verification, achieving precise identification of redundant steps. Extensive experiments on two datasets demonstrate the superiority of the proposed method.
[AI-54] Slots Are Not Speakers: Training-Free Speaker Linking for Frozen Streaming Diarizers ICASSP2027
链接: https://arxiv.org/abs/2610.28053
作者: Aditya Srinivas Menon,Kumarmanas Nethil,Kavya Manohar
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: Submitted to ICASSP 2027
Abstract:This paper presents a causal, training-free linker that allows a frozen streaming diarizer to track more speakers than it has output slots. Streaming end-to-end diarizers such as Streaming Sortformer emit a fixed number N of concurrent speaker slots. Above N speakers, they keep detecting speech but reuse slots across people, because N bounds how many speakers the model can express at once, not how many it will meet. The linker uses only the host’s slot posteriors and the audio. It keeps the frames where one slot is active, splits each slot where the speaker changes, embeds each piece with an off-the-shelf speaker encoder, and matches it against a bank of speakers seen so far. It only relabels, so missed speech and false alarm are identical to the host’s and every gain is a reduction in speaker confusion. On 1418 sessions from eight corpora, it reduces the 4-slot host’s DER above its cap from 36.8% to 22.6%, and barely moves it below the cap. The same constants transfer to the host’s 8-slot successor and to two streaming EEND hosts from a different model family. On courtroom hearings with up to 23 speakers it reduces cpWER from 82.7% to 24.4%.
[AI-55] Dancing Diffusion: Learning Continuous Programs for Image Generation and Editing
链接: https://arxiv.org/abs/2610.28003
作者: Cui Enhe,Jinyue Yang,Bojian Yin,Guoqi Li
类目: Artificial Intelligence (cs.AI); Graphics (cs.GR)
备注:
Abstract:Diffusion models generate samples through a multi-step denoising trajectory, yet condition every step on the same instruction, so the trajectory cannot be addressed at the level of individual operations. We introduce Dancing Diffusion, which assigns each span of timesteps a dedicated continuous condition block and thereby turns conditioning into a latent program. A compiler maps text features into this program, and a fixed pretrained generator executes its blocks over successive sampling intervals. The trajectory can then dance, keeping its executed history while its continuation is redirected. A shared prefix is executed once and reused, and a later operation is rewritten while content fixed by earlier operations is preserved. The compiler is trained through the images it produces, with visual feedback on both the requested change and the retained content. On SANA, reusing an executed prefix preserves markedly more of the scene than regenerating from noise at a similar strength of edit, learned conditions preserve more background than native prompt switching from the same saved state, and the learned program improves generation quality over native SANA. Later branches weaken the edit, which leaves semantic control over individual operations limited. By converting a single-instruction trajectory into a program of addressable operations, Dancing Diffusion lays a foundation for structured intermediate computation in diffusion models.
[AI-56] Learning Tactile-Aware Hand Representations from Human Hand-Object Interaction Data
链接: https://arxiv.org/abs/2610.27995
作者: Yeonguk Yu,Jaehong Kim,Byungok Han
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 9 pages, 4 figures
Abstract:Tactile representation learning for dexterous manipulation often requires substantial robot-side data collection, limiting its scalability across different hands and sensors. We explore whether large-scale human hand-object interaction (HOI) data can provide a useful prior for robotic tactile learning. To this end, we propose HuTact, a framework that pretrains tactile-aware hand representations from human HOI data without requiring real tactile signals during pretraining. HuTact transfers the interaction structure learned from human hand motions to robotic tactile observations and can be further adapted with a small amount of unlabeled robot interaction data. We evaluate HuTact on tactile perception and manipulation tasks, including grasp prediction, slip detection, object classification, and tactile-conditioned policy learning. Across these tasks, human HOI pretraining improves data efficiency and provides a strong initialization for downstream learning. Notably, with HOI pretraining followed by midtraining on unlabeled robot play data, HuTact achieves 88.26% Macro-F1 on grasp prediction using only 25% of the labeled downstream data, comparable to training from scratch on the full labeled dataset (88.23%). These results suggest that human interaction data can serve as a scalable complementary source for robotic tactile representation learning.
[AI-57] When Do Humans and Machines Truly Complement Each Other? Formal Conditions and Empirical Evidence for Human-AI Synergy
链接: https://arxiv.org/abs/2610.27965
作者: Andrea Campagner,Federico Cabitza,Vincenzo Crupi
类目: Artificial Intelligence (cs.AI)
备注: This work has been submitted to, and is currently under review at, IEEE Transactions on Human-Machine Systems for possible publication. Copyright may be transferred without notice
Abstract:Can humans and artificial intelligence cooperate such that their combination yields greater accuracy than either agent alone? This result, termed \emphsynergy or \emphcomplementarity, is notoriously difficult to achieve, even when both agents possess significant independent accuracy. In this paper, we present a formal framework to model human-AI interaction through the lens of reliance patterns, which characterize the joint correctness structure of human and algorithmic decisions. We derive mathematically necessary and sufficient conditions for synergy, demonstrating that its emergence depends on the ability of either agent to identify and correct the errors of the other, what we call Human and Algorithmic Correction. Beyond its theoretical contribution, the proposed framework provides a foundation for the design and evaluation of knowledge-based decision support systems, where interaction policies must dynamically integrate heterogeneous sources of expertise. We demonstrate this through the analysis of seven clinical case studies. Our empirical results reveal a ``dominance trap,‘’ where a strong algorithmic advantage can paradoxically inhibit synergy by eroding human corrective intervention. We also show that synergy can persist even in the presence of cognitive biases, such as automation bias, provided there is functional de-correlation between human and AI errors. These results provide a formal foundation for key concepts in the human-AI systems literature and offer actionable insights for designing interaction-centric AI systems that foster complementarity.
[AI-58] APRIL: Ambiguity-Aware Expert Routing for Respiratory Sound Domain-Incremental Learning
链接: https://arxiv.org/abs/2610.27942
作者: Xinhong Li,Weiping Tu
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注:
Abstract:Automated respiratory sound classification enables continuous, non-invasive respiratory assessment, yet long-term deployment remains challenging due to domain shifts from evolving recording devices and acquisition periods. When privacy and governance regulations prohibit replaying historical data, sequential learning easily overwrites previously acquired knowledge. Preserving dedicated domain experts prevents parameter interference, but selecting the correct expert at inference is difficult under overlapping acoustic distributions and absent domain labels. To resolve this dilemma, we propose APRIL, an exemplar-free domain-incremental framework for RSC. Built on a shared frozen BEATs backbone, APRIL sequentially inherits and isolates low-rank experts to enable new-domain adaptation without altering earlier models. A multi-level prototype router retrieves the top-two candidate experts and dynamically selects the final model by integrating domain-routing posteriors with classification confidence, adaptively weighted by matching ambiguity. We establish standardized device- and temporal-incremental benchmarks on the ICBHI and SPRSound datasets across forward and reversed domain orders. Evaluated over five random seeds, APRIL consistently achieves the highest final-session average Score among compared continual learning approaches, reaching 64.4% and 65.4% on ICBHI, and 83.1% and 82.2% on SPRSound under forward and reversed sequences, respectively. Crucially, average forgetting remains under 1.0% across all protocols, demonstrating robust domain adaptation and long-term knowledge retention without data replay. Overall, APRIL offers a reliable and privacy-preserving paradigm for deploying continuous respiratory sound classification across real-world healthcare settings.
[AI-59] HypTI: Hyperbolic Geometry-Aware Multimodal Feature Fusion Framework for Dementia Detection Based on Tabular-Image Data
链接: https://arxiv.org/abs/2610.27941
作者: Jianfei Xu,Huizhi Liang,Judith Harrison,John-Paul Taylor,Jaume Bacardit,Varun Ojha
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:AI-assisted dementia detection usually requires multimodal fusion of diverse data, such as clinical tabular assessments and neuroimaging data. However, most existing approaches perform feature fusion in Euclidean space. This overlooks the geometric structure of representation, including hierarchical organization in clinical assessments and multi-scale anatomical patterns in neuroimaging data. Recent studies have proved hyperbolic space provides a natural geometry for modelling hierarchical and multi-scale structures due to its exponentially expanding space. Motivated by these, we propose HypTI, a hyperbolic geometry-aware multimodal fusion framework for dementia detection based on clinical tabular assessments and MRI scans. HypTI extends hyperbolic representation learning to an underexplored tabular-MRI setting, jointly modelling clinical tabular, MRI scans, and cross-modal interactions in hyperbolic space. Specifically, clinical tabular variables are represented through learnable Lorentz prototypes to capture hierarchical relationships among assessments, while MRI scans are encoded using ROI-guided multi-scale representation learning to capture anatomical patterns across brain regions and spatial scales in Lorentz space. The multimodal representations are integrated through an adaptive fusion module generates subject-specific scale and shift coefficients to dynamically modulate and integrate the multimodal embeddings, enabling effective patient-level tabular-image representation learning for dementia detection. Experiments on OASIS-3 dataset show that the proposed framework achieves strong performance compared with representative approaches.
[AI-60] From Comparison to Selection: Enhancing Cross-Scale Generalization of Pointer Networks Trained with Reinforcement Learning for Capacitated Drone Delivery Routing
链接: https://arxiv.org/abs/2610.27936
作者: Sofiene Lassoued,Devara Tertia,Stefan Lier,Andreas Schwung
类目: Artificial Intelligence (cs.AI)
备注: 12 pages
Abstract:Unmanned aerial vehicles (UAVs) enable fast, flexible last-mile and emergency delivery, but their routing remains challenging as the number of clients varies. Classical approaches based on mixed-integer programming and heuristics can scale poorly or require problem-specific design, motivating learning-based alternatives. Reinforcement learning is particularly promising, provided a single policy can generalize across problem sizes without retraining. Although attention-based Pointer Networks are inherently size-agnostic, their solution quality degrades under both upscaling and downscaling. Using the Capacitated Vehicle Routing Problem (CVRP) as a controlled proxy, with fixed capacity reflecting drone payload constraints, we investigate this degradation and identify three contributing factors: shifts in node density, attention dispersion, and dilution of the mean-pooled graph context. We systematically compare architecture-wide, encoder-side, and decoder-side modifications to address these limitations and propose a learned, permutation-invariant graph-context aggregation mechanism, implemented as a weighted average and as Set2Set. Results reveal distinct requirements for transfer in each direction: entropy-based attention scaling reduces the CVRP20-to-CVRP100 performance gap by 1.41 percentage points, while the proposed aggregation methods achieve the largest improvement on the more challenging CVRP100-to-CVRP20 transfer, approximately 0.9 percentage points. Exploiting this complementarity, we introduce an approach selector that chooses the best model for each training-target size pair. The selector reduces cost relative to the reference model across all nine seen train-test pairs and eight of nine pairs involving unseen sizes (CVRP35, CVRP75, and CVRP200), achieving improvements of up to 3.2 percent.
[AI-61] Quantum-JEPA: Quantum Born Priors for Latent-Action JEPA World Models
链接: https://arxiv.org/abs/2610.27921
作者: Zhiyuan Liu,Sicheng Liu,Haiyu Wu
类目: Artificial Intelligence (cs.AI)
备注: 36 pages, 7 figures
Abstract:A single observation can admit several plausible futures, making uncertainty an essential part of world modelling. However, quantum generative models remain underexplored for this task. We introduce Quantum-JEPA, a latent-action JEPA framework that learns distributions of future representations using a conditional quantum Born prior. Circuit measurements generate binary latent actions, which a shared predictor maps to alternative future representations. We train the quantum prior through the quality and diversity of these predictions using a sampled energy objective. Our model learns distinct continuations on constructed multimodal tasks and captures latent-action correlations that support coordinated multi-agent futures. On D3IL, a benchmark of diverse human demonstrations, our trained models generate multiple obstacle-crossing routes from a common starting point without observed motion. On the controlled tasks, our classical and quantum prior implementations improve prediction validity and distributional fidelity over the evaluated variational baselines. Our quantum priors combine competitive predictive quality with a parameter-efficient sampling structure: they avoid an exponentially sized categorical output layer while supporting correlated latent actions. We further validate finite-shot parameter-shift training, including training with readout noise. Unlike most prior work, which focuses on simulation, we also verify the feasibility of inference with trained priors on a real QPU. These results establish a trainable quantum component for JEPA future prediction and a basis for exploring quantum-enhanced world models at larger scales.
[AI-62] MMEnterprise-Bench: Evaluating Enterprise Agents over Evolving Multimodal Records
链接: https://arxiv.org/abs/2610.27911
作者: Seohyun Lee,Jongha Kim,Dohwan Ko,Ji Soo Lee,Seoung Choi,Jayun Ku,Hyunwoo J. Kim
类目: Artificial Intelligence (cs.AI)
备注: 35 pages, 16 figures, including appendices. Under review
Abstract:Enterprise agents answer questions over workplace records that evolve through team collaboration. As information is shared, corrected, and revised across documents, chats, and figures, multiple versions of the same information accumulate in the workspace. Reliable question answering therefore requires agents to reason over sufficient evidence, resolve cross-source and temporal conflicts, and recognize when the available evidence is insufficient to establish an answer. To systematically evaluate these capabilities, we introduce MMEnterprise-Bench, a benchmark comprising 1,613 questions grounded in evolving multimodal records from simulated collaboration projects, covering three categories: Evidence Reasoning, Conflict Resolution, and Unanswerability. We evaluate 10 proprietary and open-source multimodal LLMs and find that even the strongest models struggle when evidence contains cross-source or temporal conflicts or spans multiple images. Moreover, increasing the retrieval budget or employing a larger retriever improves evidence recall but does not consistently translate into higher answer accuracy. These findings highlight the persistent challenges of reasoning over evolving workplace records and establish MMEnterprise-Bench as a testbed for developing more reliable enterprise agents. Code and data are available at this https URL.
[AI-63] he Shape of the DPO Hyperparameter Landscape: What to Measure How It Moves with Scale and When Not to Transfer It
链接: https://arxiv.org/abs/2610.27895
作者: Hannan Mahadik,David Salinas,Jacek Golebiowski,Aaron Klein
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Direct preference optimisation (DPO) is widely used for language model post-training, but its performance depends strongly on the choice of its hyperparameters, such as the learning rate and the reference-policy regularisation parameter \beta . To better understand the DPO hyperparameter landscape, we conduct extensive experiments using a dense hyperparameter grid across different model scales from the same family. This allows us to analyse how the optimal DPO hyperparameters evolve with model size and the number of update steps, with intermediate checkpoints providing an inexpensive signal to accelerate the hyperparameter optimisation process. Furthermore, by evaluating our models on both static and dynamic benchmarks, we identify which benchmarks provide the strongest signal for hyperparameter selection. Finally, we investigate how different state-of-the-art hyperparameter optimisation strategies perform in this setting. Our experiments show that the optimal learning rate varies with model size and that hyperparameters are not transferable across training datasets. We also find that validation loss is a poor proxy for hyperparameter optimisation, as hyperparameters that minimize validation loss only weakly correlate with optimal hyperparameters for downstream benchmarks. Therefore, DPO hyperparameters need to be tuned specifically for the target model, dataset, and evaluation objective. We also show that short candidate training runs can substantially reduce the cost of this selection. We release all our training configurations and evaluation results to support reproducible analysis of selection quality and training cost.
[AI-64] Physical-Variable-Informed Deep Learning for Robust Diagnosis in Industrial Fluid Transport Systems
链接: https://arxiv.org/abs/2610.27891
作者: Yu Sha,Junqi Tao,Dixin Zhou,Horst Stoecker,Domagoj Vnucec,Andreas Widl,Kai Zhou
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 29 pages, 5 figures
Abstract:Industrial fluid transport systems are monitored through heterogeneous sensing streams, but robust fault diagnosis remains difficult because response signals vary with operating conditions and differ from physical variables in sampling rate, observation duration and data granularity. To address these challenges, we propose a physical-variable-informed deep learning framework (PhyNet) for robust fault diagnosis under real-world sampling heterogeneity. PhyNet treats physical variables as global operating constraints and response signals as local diagnostic evidence. Its physical-to-response cross-attention mechanism connects sequence-level physical-variable tokens with a learnable global response token, enabling interaction without strict temporal alignment. Across four benchmarks, including three real-world cavitation datasets and one public two-phase-flow dataset, PhyNet achieves the best performance across all evaluated metrics. Class-wise analyses, together with attention and masking analyses, show improved discrimination of ambiguous states and indicate that physical variables meaningfully constrain local response interpretation. These findings demonstrate a deployment-oriented approach to robust diagnosis in heterogeneous industrial environments.
[AI-65] Automatic Differentiation with Constraints in Forward and Inverse Problems
链接: https://arxiv.org/abs/2610.27888
作者: Nikolay Yavich,Alexander Ryabov,Viacheslav Naumov,Sayan Ranu,N. M. Anoop Krishnan,Evgeny Burnaev,Vladimir Vanovskiy
类目: Artificial Intelligence (cs.AI); Numerical Analysis (math.NA)
备注:
Abstract:Approximate solutions of partial differential equations (PDEs) on unstructured grids can be represented on a graph. Numerical optimization involving stiff evolutionary and static PDEs often requires solving a large system of equations, which can be treated as a constraint. In this work, we present a fully differentiable graph-based implicit solver that incorporates this constraint. We evaluate its performance across four key tasks: a) linear and nonlinear forward modeling of evolutionary and static problems, b) solving inverse problems to estimate equation coefficients, c) solving inverse problems to estimate the right-hand side, and d) learnable graph coarsening to accelerate forward modeling. The stability and differentiability of our solver enable significant reductions in the complexity of forward modeling while efficiently addressing related inverse problems. These capabilities make it a powerful tool for different physics and engineering applications, in particular for geophysics, where inverse problems are commonly solved in order to reduce the model uncertainty.
[AI-66] Hallucinations of Large Language Models in Cross-Language Code Translation: Phenomena and Mitigation
链接: https://arxiv.org/abs/2610.27868
作者: Yali Du,Yunji Zhang,Ming Li
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Code translation converts source code from one programming language to another while preserving its functionality. Although Large Language Models (LLMs) have improved code translation, their outputs still contain hallucinations: generated code that conflicts with the source program, violates target-language facts, or introduces contextually inconsistent dependencies. These failures are especially harmful in translation because small semantic deviations can silently change program behavior. Existing hallucination studies mainly focus on natural-language generation or code generation, leaving the forms, distribution, and mitigation of hallucinations in cross-language code translation insufficiently characterized. This paper presents a systematic empirical study of LLM hallucinations in code translation. Across three benchmark datasets, six language pairs, and seven LLMs, we construct a taxonomy with three major categories: Semantic Consistency Conflicts, Factual Knowledge Conflicts, and Project Context Deviations. We manually annotate 13,919 hallucination cases and analyze how hallucination types vary across models, datasets, and language directions. Guided by the taxonomy, we further propose HaluCodeTran, a mitigation framework that combines BM25-based retrieval of relevant translation examples with taxonomy-aware reflective refinement. The study provides an execution-tested characterization of code-translation hallucinations and a practical baseline for building more reliable LLM-based code migration tools.
[AI-67] Physical Operating-Condition-Guided Vision-Language Matching for Cavitation Diagnosis in Industrial Pipeline Systems
链接: https://arxiv.org/abs/2610.27860
作者: Yu Sha,Junqi Tao,Deyu Wang,Horst Stoecker,Domagoj Vnucec,Andreas Widlf,Kai Zhou
类目: Artificial Intelligence (cs.AI)
备注: 25pages,6 figures
Abstract:Cavitation diagnosis in industrial pipeline systems is challenging because acoustic responses vary substantially with operating conditions, while similar time-frequency patterns may correspond to different cavitation states. To address this issue, we propose a physical operating-condition-guided vision-language matching (POC-VLM) for cavitation diagnosis. POC-VLM first constructs physical operating-condition-aware semantic descriptions by explicitly incorporating sample-specific operating variables into candidate cavitation texts. A time-frequency-conditioned text adaptation module then dynamically aligns these physical semantics with the observed acoustic response, followed by global-local vision-language matching to capture both holistic and localized discriminative patterns. In parallel, a K-nearest neighbours-based non-parametric physical operating-condition prior estimates the physical plausibility of each candidate state and is fused with the vision-language semantic evidence at the decision level. Experiments on three real-world cavitation datasets show that POC-VLM consistently outperforms representative data-driven, multimodal and domain-specific diagnostic methods. Further analyses verify the complementary roles of vision-language semantic matching and physical operating-condition priors, while demonstrating robust performance under varying sampling rates, noise levels and observation durations.
[AI-68] SAPD: Learning from Agent Experience without Environment Interaction
链接: https://arxiv.org/abs/2610.27847
作者: Juncheng Diao,Haibin Wang,Wei Shi,Yifan Ding,Qingbin Li,Rongxiang Weng,Xunliang Cai
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Scaling the training of Large Language Model (LLM) agents typically requires a large collection of executable environments, whose construction, maintenance, and operation constrain the expansion of training experience. Existing offline trajectories offer an opportunity to reuse interaction experience, yet effectively refining the current policy without revisiting these environments remains challenging: supervised fine-tuning (SFT) can lead to catastrophic forgetting, while online reinforcement learning requires real-time interaction with the environment. To address this challenge, we propose Step-Advantage Policy Distillation (SAPD), a method for learning from offline trajectories using relative step advantages. SAPD decomposes multi-turn trajectories into individual decision contexts for on-policy distillation, allowing the student to generate responses conditioned on recorded histories, and introduces relative step advantage signals to stabilize training. The entire process requires neither new environment observations nor additional multi-turn interactions. Experiments on ALFWorld, WebShop, and Search-based QA show that SAPD improves performance across different teacher configurations, mitigates catastrophic forgetting compared with SFT, and runs at least 4.7\times faster per training step than online OPD. We will release our code soon.
[AI-69] COMPASS: Routing via Capability-Oriented Model Profiling with Adaptive Search Selection and Tool Use
链接: https://arxiv.org/abs/2610.27822
作者: Amin Rezaei,Pouya Parsa
类目: Artificial Intelligence (cs.AI)
备注: 14 pages, 4 figures
Abstract:LLM routing aims to select a capable model at low cost, but specialized routers can require updates as models and tasks change. We propose COMPASS, an agentic router in which an LLM follows a routing recipe, gathers performance and cost evidence through tools, and optionally inspects candidate answers before selecting a model. Our evaluations on LLMRouterBench and DeepSWE show that agentic routing can improve accuracy and lower model-call cost relative to specialized and trained routing baselines, without routing-specific training. On LLMRouterBench, COMPASS reaches 69.85% accuracy at \ 139.57, compared with RLCascadeRouter’s 69.14% at \ 180.22. On DeepSWE under fixed five-fold cross-validation, it reaches 78.1% at \ 215, compared with 74.1% at \ 500 for the best single model. COMPASS also supports routing across benchmarks and adapts to an expanded model pool by reading new evaluation records, without changing the router’s weights or prompt.
[AI-70] Privacy-preserving inference on hosted models: a learned input-embedding transform for frozen Qwen 3.8-27B
链接: https://arxiv.org/abs/2610.27817
作者: Mahendra Roopa,Srikanth Roopa
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Every prompt sent to a hosted language model is readable by whoever runs it. We study whether the prompt can remain on the client while the host receives transformed vectors that stand in for it. A language model converts text to vectors exactly once, with a public lookup table; we perform that step locally and pass the vectors through a small trained network (55.6M parameters) before they leave the perimeter. The host runs a frozen 27B model (Qwen3.8-27B) on the transformed vectors and never receives text or token ids. We train the transform against the frozen model as a fixed judge, combining a forward-KL utility term with an auxiliary term that pushes each vector away from its original direction. On 1,490 MMLU questions held out from training, the transform costs 1.5 points of accuracy (80.7% vs. 82.3%), breaking 68 of the 1,226 questions the unmodified model answers correctly (5.5%), while 30.1% of content tokens no longer decode to their original token under a nearest-neighbour lookup.
[AI-71] RAG -NIMBF: A Non-Invasive Modular Benchmarking Framework for Automated RAG Configuration Analysis
链接: https://arxiv.org/abs/2610.27816
作者: Niclas Hart,Nicolas Weeger
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Retrieval-Augmented Generation (RAG) has become the dominant paradigm for grounding the outputs of large language models (LLM) in external knowledge. However, practitioners are faced with a vast configuration space comprising chunking, embedding, retrieval, reranking, prompting and generation models, and there is limited guidance on how these components interact. We present RAG-NIMBF, a modular benchmarking framework that attaches to existing RAG pipelines as a non-invasive evaluation layer. It evaluates the configuration space through controlled ablation with end-to-end experiment tracking. The evaluation combines LLM-based answer quality scores, retrieval correctness measures, and runtime indicators. This study validates the framework across multiple generation models on the SQuAD dataset. This process reveals non-obvious interactions and demonstrates how the experiments are tracked. The findings indicate that the retrieval balance of relevance and diversity substantially enhances the relevance and coverage of retrieved context in comparison to a conventional similarity search, with MMR raising nDCG@5 by +4 to +11 percentage points across the swept models. More detailed prompt instructions meaningfully increase how faithfully generated answers stick to the retrieved evidence, especially for capable and reasoning models (+10 to +13pp on Qwen3.6-27B, GPT-4o-mini and Qwen3.5-think, but negligible on the 20B generator). Overall, RAG-NIMBF transforms the configuration space of RAG into reproducible, comparable experiments. By attaching to existing production pipelines in a non-invasive way, it enables practitioners to systematically evaluate and improve deployed systems without disrupting existing workflows. This replaces ad hoc choices with evidence-based decisions.
[AI-72] LatBridge: Efficient and Stable Model Replacement in Latent-Communication Multi-Agent Systems
链接: https://arxiv.org/abs/2610.27813
作者: Jiajun Zhu,Haobo Li,Zixin Chen,Huamin Qu,Wenshuo Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-agent systems rely on communication to complete tasks, and latent communication passes continuous representations to reduce text generation and re-encoding. However, multi-agent systems must reconfigure agent models, whereas existing latent-communication methods are built for fixed model combinations, so their interfaces may fail after replacement. Our pilot study confirms this interface failure and identifies the requirement for interface reuse after model replacement. We present LatBridge, which connects agent models through a shared Bridge and separate interfaces under a shared latent-communication protocol. After agent model replacement, LatBridge reuses unchanged interfaces, selects matching low-rank interfaces, and combines them through the Bridge, training only missing ones with all models frozen. The Bridge processes source representations in the basic mode and constructs target input without autoregressive generation in the accelerated mode. Experiments support effective communication after model replacement, with mean basic-mode accuracies of 14.8% on MATH and 27.3% on ALFWorld next-action prediction across eight small-model configurations. Analyses examine the factors supporting task accuracy and communication efficiency.
[AI-73] oward Agent ic AI for the 6G Physical Layer: From Algorithm Discovery to Verified Baseband Silicon
链接: https://arxiv.org/abs/2610.27797
作者: Eric Mittag,Lukas Brechtel,Maximilian Stark,Hans D. Schotten
类目: Artificial Intelligence (cs.AI); Networking and Internet Architecture (cs.NI); Signal Processing (eess.SP)
备注: 7 pages, 3 figures, 5 tables. Accepted at the 2026 IEEE Globecom Workshops (GC Wkshps), 2nd Workshop on Agentic AI Empowered Wireless Communications and Networking, Macau, China, December 2026. © 2026 IEEE
Abstract:6G will be designed faster than it can be built: AI-driven search already proposes physical-layer algorithms at machine speed, while carrying a single one onto embedded baseband silicon remains specialist work. That silicon imposes hard memory and numeric limits, but also exposes cost structures a GPU lacks, and a design flow that never runs on the target sees neither. We present an agentic workflow in which an LLM coding agent implements algorithms on a commercial baseband SoC, optimizes them down to agent-written vector assembly, and validates them on silicon. The agent is guided by a knowledge base of measured platform facts, checked at every step against evaluators it is instructed never to edit, and metered per phase. We showcase the workflow on an AI-discovered orthogonal time frequency space (OTFS) detector portfolio on the open 6G4All platform. It yields twelve kernels that run cycle-exactly on silicon, their optimized versions at 18k-405k cycles (30-660 \mu s) per frame. It also yields a new detector shaped by the chip’s measured costs that needs 13.9% fewer cycles than the port of the discovered detector while matching its error rate, with a measured edge our evaluation cannot statistically resolve, a structure that a hardware-blind search objective had no reason to find. What is not specific to OTFS is the workflow: an LLM agent exploring an implementation space under real hardware limits, with the cost of every choice measured rather than assumed.
[AI-74] SLM-SS: Token-Based Speech Separation with Speech Language Models
链接: https://arxiv.org/abs/2610.27785
作者: Chenda Li,Tianhua Li,Yiming Chen,Wei Wang,Jiahe Wang,Xin Zhou,Yanmin Qian
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注:
Abstract:Speech separation (SS) has witnessed substantial progress in recent years with the advent of neural network-based methods, achieving notable improvements on signal-level evaluation metrics. Existing speech separation systems are mostly optimized with signal-level objectives, and often yield outputs that are acceptable in waveform metrics but suboptimal for intelligibility-sensitive downstream tasks. To address this gap, we propose SLM-SS, a token-based generative speech separation framework that models multi-talker mixtures with speech language models. We first discretize clean speech into multi-codebook token sequences with neural codecs and serialize multi-speaker targets via Serialized Output Training (SOT). An encoder-decoder autoregressive speech language model predicts the zero-order sequence to establish speaker ordering and coarse linguistic content, and a non-autoregressive module then refines higher-order residual codebooks for efficient reconstruction. We further introduce a per-talker non-autoregressive strategy that simplifies high-order modeling and improves the stability. On Libri2Mix, SLM-SS with per-talker NAR achieves 6.92% WER, approaching codec-only reconstruction. The WavTokenizer variant eliminates NAR refinement and achieves the highest mean subjective quality score among evaluated separation systems. Its UTMOS and DNSMOS advantages over discriminative baselines persist on unseen noisy and Chinese mixtures without fine-tuning.
[AI-75] RELAY: Scaling Historical Build Restoration through Rule Reuse and Agent ic Repair
链接: https://arxiv.org/abs/2610.27784
作者: Olivier Nourry,Yoshiki Higo
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Vulnerability benchmarks and software engineering research depend on rebuilding historical software. Yet builds rely on external resources, such as base images and packages, that change independently of the source code. In this study, we examine this environment decay on 14,159 OSS-Fuzz cases. We find that 69.7% no longer complete reconstruction and verification, and that most failures share a small number of recurring failure symptoms. A recurring failure can therefore be diagnosed once and its repair reused. We present Relay, a hybrid framework that uses an LLM to derive and validate deterministic repair rules from representative failures, reuses them without model calls, and escalates cases unresolved by rules to a per-case agent. We evaluate Relay in human-guided and fully automated rule-derivation campaigns and compare it with a per-case agent approach. We find that rule derivation can be fully automated, and that reusable rules restore as many builds as per-case agents with fewer build executions. However, agents restore cases that the rules cannot. Our results show that a hybrid approach that applies rules first and escalates the remaining cases to an agent restores more builds than either approach alone and at a lower cost than using agents alone.
[AI-76] GRADUS: Discovering and Recovering Agent Lessons from the Environments Latent Syllabus
链接: https://arxiv.org/abs/2610.27781
作者: Qing Ping,Wei Yang,Panpan Xu,Johannes Kirmayr,Lukas Stappen,Junjie Tang,Jun Huan
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Training LLM agents for policy-constrained tool environments, such as an airline backend with booking, refund, and cancellation policies, requires tasks that exercise the environment’s rules individually and in combination. Existing synthesis pipelines let a model propose tasks and retain only those it immediately solves, biasing training data toward easier behaviors. We introduce Gradus (Graded Rungs of Ascending Difficulty from an Underlying System), which decouples what a task should teach from the concrete task and solution used to teach it. Without seed tasks, Gradus treats a fixed environment as a latent syllabus, extracting guarded behavioral operators that capture how actions depend on current conditions and how their effects reshape subsequent decisions. It composes these operators into increasingly demanding tasks while using corpus-level feedback to target missing outcome combinations. Crucially, Gradus jointly calibrates tasks and solutions against the environment: execution feedback identifies which side to refine while preserving the intended behavioral requirements. Verification thus becomes a recovery signal rather than a terminal filter, coupling coverage-guided exploration with the preservation of difficult examples. From off-the-shelf backbones, the resulting curriculum lifts a 30B model from 59% to 89% on \tau^2 -Bench and from 47% to 83% on CAR-Bench with an Opus 4.5 proposer, and to 76% and 62%, respectively, with an open-weight Qwen3-235B proposer. These results show the value of jointly controlling which behaviors are explored and which survive into verified training data.
[AI-77] RITE: Evaluating End-to-End Resilience to Silent Tool Errors in E-Commerce Agent ic Systems
链接: https://arxiv.org/abs/2610.27777
作者: Wenshuo Zhang,Yuquan Lin,Yushi Sun,Yishi Lin,Rui Sheng,Peng He,Han Gao,Wang Jikai,Yifeng Rong
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Tool-using LLM agents increasingly answer e-commerce queries through multi-step tool interactions. The expected output of a tool call can shift between agent development and execution, allowing a plausible but incorrect value to be returned without an explicit failure signal. These silent errors may propagate through subsequent decisions into final answers. To study agents’ resilience across this end-to-end process, we formulate Silent-error Resilience Task, an interactive task that evaluates whether they complete the original query after a silent error corrupts an intermediate state. We introduce RITE, a benchmark built from human-authored industrial workflows and production-log-derived inputs. To enable controlled and reproducible evaluation, we use a semantic error taxonomy to inject errors into fixed workflow prefixes, freeze recorded tool behavior, and let agents continue freely. It contains 167 matched units across nine e-commerce workflows, yielding 501 silent-error, correct-value, and explicit-failure instances. Across 10 recent LLMs, silent errors roughly halve task success and are consistently harder than explicit failures. Recognition, use of an available verification path, and integration of the resulting evidence form successive bottlenecks. We also propose a light State Checker as an intervention to test whether resilience can be improved.
[AI-78] CrystalAR: Autoregressive Modeling for Symmetric Crystal Generation
链接: https://arxiv.org/abs/2610.27762
作者: Ziqiao Meng,Rongsheng Chen,Zhi Zheng,Nikita Kazeev,Zhenkun Wang,Andrey Ustyuzhanin,Wee Sun Lee
类目: Artificial Intelligence (cs.AI); Materials Science (cond-mat.mtrl-sci)
备注:
Abstract:Deep generative modeling has achieved remarkable progress in crystal generation. Given the rich symmetries inherent in crystalline structures, recent work has sought to explicitly incorporate space-group symmetry into generative models, enabling more effective exploration of novel symmetry templates. Yet introducing symmetry templates increases the combinatorial complexity of stoichiometric relationships between atomic species and crystallographic sites, while creating a fundamental tension between flexible exploration of discrete crystallographic templates and faithful realization of these templates as continuous three-dimensional structures. To address these challenges, we propose CrystalAR, a unified autoregressive framework that models Wyckoff symmetry templates and complete crystal structures within a single generative sequence. Specifically, CrystalAR reformulates continuous coordinate generation as a symmetry-conditioned hierarchy of discrete spatial localization decisions, autoregressively refining the positions of all symmetry-inequivalent sites across resolution levels while jointly modeling their spatial configuration within each level. This discrete hierarchical formulation further enables constraint-aware search over both crystallographic templates and geometric realizations. As a result, CrystalAR retains strong symmetry-template generation capability while achieving competitive de novo generation performance and template-free crystal structure prediction (CSP) accuracy. These results suggest that symmetry-template generation can be retained alongside broadly competitive de novo generation performance and competitive CSP without reliance on symmetry templates retrieved from prior data, providing a promising alternative paradigm for symmetry-preserving crystal generation.
[AI-79] RSI Control: Can AI Models Perform Covert Sabotage in the Self-Improvement Process?
链接: https://arxiv.org/abs/2610.27735
作者: Kaiwen Zhou,Udari Madhushani Sehwag
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recursive self-improvement, in which frontier AI models work on training the next generation, is quickly becoming a reality. However, the safety of this process has received far less attention than it should. A misaligned model training its successor can shape what that successor knows, how it behaves, and what it wants, and the harm carries forward into the next generation. We study this risk through the lens of AI control. We introduce RSI Control Arena, a testbed in which a frontier model, acting as an autonomous R\D agent, is given a benign main task from a stage of a successor model’s training pipeline together with a covert side task to install a misaligned property in the successor. The testbed contains 16 side tasks in three classes, forming 22 task pairs with main tasks that span the pipeline from data curation to reinforcement learning. We evaluate six frontier models as agents against LLM monitors that read the transcript of the run and the agent’s reasoning. We observe that every model completes a substantial share of the side tasks without degrading the main-task performance, including side tasks that make the successor itself a better saboteur. Under minimal instructions to stay covert, models already produce successful attacks that monitors fail to detect, and providing a short list of evasion tactics further reduces the detection rate. Monitors with access to agent’s reasoning detects most attacks, but not all. Our results indicate that recursive self-improvement raises significant safety concerns, and that strong safety effort should be in place before models are trusted with greater responsibility of training their successors.
[AI-80] Anytime-Valid Runtime Monitoring of Learned Barrier Certificates AAAI2027
链接: https://arxiv.org/abs/2610.27726
作者: Zhipeng Cao,Taoran Wu,Weiru Cui,Luke Ong,Bai Xue
类目: ystems and Control (eess.SY); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Optimization and Control (math.OC); Statistics Theory (math.ST)
备注: 17 pages, 7 figures. An earlier version of this work was submitted to AAAI 2027 on 21 July 2026 (submission no. 28239). The core monitoring framework and main theoretical results were already present in that submission
Abstract:Learned barrier certificates provide a principled mechanism for establishing safety of stochastic systems, but their validity may deteriorate during deployment when the underlying dynamics differ from those used for training. We propose an anytime-valid runtime monitor that audits a learned barrier certificate using observed closed-loop trajectories. The monitor is intended to detect loss of statistical support for the deployed barrier condition; it does not by itself guarantee detection before unsafe entry or safety of a subsequent intervention. Specifically, the conditional barrier-drift condition is formulated as an episode-wise sequential hypothesis test, and observed barrier increments are transformed into a nonnegative test supermartingale, yielding finite-sample false-alarm control under arbitrary stopping times. Inspired by finite-horizon safety verification, we further introduce a deterministic temporal shaping of the monitoring boundary, making the monitor more conservative early in a prescribed deployment horizon while preserving anytime-validity. The monitor is designed to detect statistical evidence that the safety certificate is no longer consistent with the observed dynamics, rather than to predict unsafe entry. To assess the monitor empirically, we evaluate empirical alarm behavior under valid and invalid certificates. Code and data are available at this https URL.
[AI-81] How Language Models Organize Task Information
链接: https://arxiv.org/abs/2610.27724
作者: Kaiyang Wan,Yuxia Wang
类目: Artificial Intelligence (cs.AI)
备注: 88 pages, 12 figures
Abstract:Dense linear probes assess task readability from language model (LM) representations, but do not directly reveal how many features suffice for prediction. We measure readout sparsity with (k^), the smallest feature count found by a fixed search protocol that recovers 95% of dense-probe AUROC above chance. We evaluate 19 LMs from five labs on 14 core tasks. Ten base models give the nine reliably readable tasks similar feature counts and nearly the same order. This agreement extends to the selected features themselves, which on three tasks show cross-model response correspondence beyond class-mean differences and a label-correlation-matched null. Instruction-tuned deployment selectively reduces (k^), with both weights and input format contributing. In controlled synthetic experiments across 20 models, successfully learned targets admit single-feature readouts, whereas readable controls require tens of features. Reference-model (k^) is associated with label requirements across eight students, following an empirical scaling of (N^\propto(k^*)^0.65). After continual learning, saved heads can fail while linear readability remains high, and recovery requires fewer labels than initial task learning under the tested procedures. Pruning reduces linear readability, with wider pre-pruning readouts generally associated with larger losses. These findings distinguish task readability from readout sparsity and establish their complementary value for predicting learning requirements and diagnosing LM changes.
[AI-82] How Fine-Tuning Objectives Shape Layer-Wise Information in LLM Hallucination Detection NEURIPS2026
链接: https://arxiv.org/abs/2610.27711
作者: Kaiyang Wan,Forrest Sheng Bao,Amin Ahmad,Yuxia Wang
类目: Artificial Intelligence (cs.AI)
备注: 37 pages, 6 figures, NeurIPS 2026
Abstract:Under matched conditions, classifier-head fine-tuning (\textsccls) paradigm consistently outperforms the LLM-as-Judge sequence-to-sequence (\textscs2s) on binary hallucination detection. We trace this gap to the representations induced by the two objectives. Static analysis shows that \textsccls learns a task-specific discriminative direction, pushing hidden states into a clear two-cluster geometry with stronger class separation and linearly decodable label information. By contrast, \textscs2s relies on the pretrained language-modeling head to make decisions. This minimizes the pressure to reorganize representations, leaving the hidden states much closer to the pretrained backbone. Our analysis also reveals that the \textsccls signal saturates before the final layer, after which upper layers form a plateau of statistically interchangeable and largely redundant representations. However, endpoint redundancy alone does not show whether these layers are dispensable during training. Further training-dynamics analysis reveals 1) a lower pretrained substrate, 2) a narrow upper-middle \textbfInfoWindow where task information emerges, and 3) a later-synchronizing upper plateau that mirrors this signal. This mechanism motivates us to remove layers above the InfoWindow. Truncating 31% of the network leaves accuracy unchanged within seed noise on Qwen3-4B and 8B and cuts latency by 1.4\times . The same depth fraction, fixed before any analysis of the target models, transfers to Llama-3.2-3B and Mistral-7B-v0.3, and the layer structure behind it reappears there and on QNLI and SST-2. We hope this helps guide efficient task-specific LLM design.
[AI-83] E8-QAT: Efficient 2-Bit Quantization-Aware Training with Implicit Lattice Codebooks
链接: https://arxiv.org/abs/2610.27699
作者: Xinyu Lei,Hong Huang,Jiebin Zhang,Hongyu Li,Jinlong Song,Decheng Wu,Guanghua Yu,Jianchen Zhu,Sujian Li
类目: Artificial Intelligence (cs.AI)
备注: 17 pages, 6 figures
Abstract:Low-bit quantization reduces the memory footprint of large language models, but preserving accuracy with efficient training remains challenging. Scalar quantization encodes weights independently, while explicit-codebook vector quantization introduces costly search during training. We propose E8-QAT, a 2-bit quantization-aware training framework that uses the E_8 lattice as an implicit eight-dimensional codebook. It jointly quantizes weight groups through arithmetic projection without storing or searching an explicit codebook during the training forward pass. A 2-bit core encoding and sparse extension preserve the selected lattice points exactly, adding less than 0.1% to core index storage. On Qwen3.6-35B-A3B, E8-QAT matches the performance of the full-precision model in terms of the average score across the four tasks. E8-QAT improves over the state-of-the-art baseline by 2.03% on IFEval and by 1.51% on LiveCodeBench, under the same data, schedule, and bit width. The 2-bit representation reduces peak memory by 80% and supports a 4\times larger batch. At each representation’s largest feasible batch, E8-QAT reaches 1.97\times the decoding throughput of BF16.
[AI-84] EASformer: Environment-Action-State Alignment for Industrial Generative Ranking
链接: https://arxiv.org/abs/2610.27691
作者: Bohao Li,Dinglun Cao,Pengcheng Huang,Yingda Tang,Zhiting Zhao,Yi Cao,Jilei Fu,Miao Yan,Tianying Liu,Heng Chang,Bowen Du
类目: Artificial Intelligence (cs.AI)
备注: 10 pages
Abstract:Modern industrial ranking systems increasingly adopt generative recommendation (GR) paradigms, in which user behavior sequences and non-sequential (NS) features are tokenized and jointly modeled by Transformer-style backbones. This unified formulation provides a scalable interface for modeling long histories and heterogeneous features. However, sequence and NS representations often carry asymmetric semantics: historical representations may combine environment context with observed-action semantics, whereas candidate-side NS representations contain no observed action but incorporate richer supplementary context, such as user-item interaction features and recent behavior lists. Direct interaction between such representations can make query-key dot products compare different semantic objects, limiting meaningful cross-token interaction. To fill this gap, we propose EASformer, an Environment-Action-State disentangled architecture for industrial app ranking. EASformer constructs an explicitly aligned context environment space shared by historical and non-sequence tokens, while disentangling user actions from historical token and candidate-specific context form non-sequence tokens. Based on this aligned space, EASformer enables complementary historical action retrieval and state-conditioned action prediction. Furthermore, the disentangled design supports a Cloze-style self-supervised objective by masking historical actions and reusing the candidate prediction pathway. On an offline benchmark from a real-world industrial app marketplace, EASformer achieves an AUC of 0.89411, yielding a 0.88% relative improvement over the deployed baseline. Online A/B tests further support this advantage, with relative eCPM lifts of 1.26% and 3.94% across two real-world ranking scenarios.
[AI-85] Pattern Intelligence: Unsupervised LLM -Interpreted Discovery of Systematic Underpayment and Denial Patterns in Healthcare Revenue Cycle Management
链接: https://arxiv.org/abs/2610.27684
作者: Harshil Lodhiya,Alex McManus
类目: Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注: 5 pages, 5 figures. Accepted at the 2027 IEEE Gujarat Section Conference (GSCON 2027), Track 1: Artificial Intelligence, Data Science, and Intelligent Systems
Abstract:Healthcare providers lose substantial revenue to systematic payer underpayments and claim denials, yet identifying the recurring, correctable patterns behind them remains largely manual and does not scale to hospital claim volumes. We present Pattern Intelligence, a deployed system that discovers such patterns without supervision by clustering claims on their adjudication behaviour rather than their identity. Each claim is embedded in a 130-dimensional vector combining per-line expected-versus-actual payment ratios, claim adjustment reason (CARC) codes, denial and payment signals, and service composition. Payer and contract identity are deliberately excluded from the distance metric - we show they dominate the geometry and collapse clusters into trivial per-payer groups - and are reintroduced only as a post-hoc segmentation (by contract and visit type) that keeps every surfaced pattern coherent and actionable. We cluster with HDBSCAN, recover structure from the noise set via second-pass sub-clustering, and gate low-confidence memberships. A large language model translates each centroid into an analyst-ready description of dominant procedures, adjustment reasons, and failure mode. We report a recurring finding: roughly one-third to one-half of high-dollar denials, depending on tenant and segmentation, form an irreducible long tail that no density-clustering configuration or representation we tried groups - a domain property, we argue, rather than a modeling deficiency. We evaluate validity through cluster-coherence analysis and quantify impact with a recoverable-opportunity ( ) metric on two production tenants (approx. 10K and 157K claims) within a six-tenant deployment. Rather than a new algorithm, the contribution is a problem formulation, a representation-design insight, and empirical findings from production.
[AI-86] mporal Evidence Allocation For Longitudinal Clinical Retrieval: A Position And Technical Report
链接: https://arxiv.org/abs/2610.27682
作者: Soumya Ranjan Sahoo,Divya Bharti,Sanand Sasidharan
类目: Artificial Intelligence (cs.AI)
备注: Extended report of our accepted abstract at IEEE BHI 2026
Abstract:Electronic health records are written to support care, not retrieval. Across years of oncology notes, the primary diagnosis is restated in report after report, while the events that shape a patient’s course, such as a recurrence or a change in therapy, may be recorded only once. A retriever that ranks passages by similarity to the query keeps returning the repeated material and can miss these transitions, that is, it exhibits temporal redundancy across historical narratives. We argue that longitudinal clinical retrieval should be treated as allocating a limited evidence budget across time, not just ranking by relevance. We separate the task into three objectives: temporal evidence allocation, candidate reconstruction, and clinical validation. In Study A (13 clinical queries over 20 patient records, with two-stage reconstruction), the Temporal-15 configuration reaches a milestone slot recall of 0.538, compared with 0.388 for the Global-50 baseline. However, it also retrieves far more evidence (496 versus 50 indexed units per query), so the gain cannot be credited to allocation alone. When evidence is held fixed, removing prompted precedence rules raises schema-order reversals from 0.010 to 0.080. A five-patient, single-query pilot with fixed budgets finds that wider temporal coverage does not reliably yield more dated milestones. Together, these results suggest that how much a record says about a period is a poor guide to how useful that period is as evidence. Because no expert reference timelines are available, our measurements describe retrieval behavior rather than clinical accuracy. This position and technical report presents the retrieval analysis behind the Onco-CONSTRUE framework and sets out a research agenda for budget-constrained longitudinal clinical retrieval.
[AI-87] Reusing Multimodal Agents : Efficient Evaluation of Sound Designs in Games
链接: https://arxiv.org/abs/2610.27670
作者: Thai Nguyen Van,Ibrahim Khan,Mury Fajar Dewantoro,Pratch Suntichaikul,Hai Van Pham,Ruck Thawonmas
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:This paper introduces a novel framework for efficiently evaluating in-game sound designs using multimodal learning and reinforcement learning. Sound design significantly influences gameplay and accessibility; however, objective evaluation methods, defined here as automated assessments based on agent performance rather than human ratings, often require retraining agents from scratch for each sound design. Our approach first pre-trains a multimodal encoder that encodes visual and auditory information into a shared latent space. A reinforcement learning agent is then trained using this latent representation. We fine-tune the multimodal encoder for each sound design while keeping the trained policy fixed, allowing us to isolate the effects of sound changes on agent behavior. To evaluate the proposed method, we conducted experiments on games from multiple genres, comparing the performance of blind agents (agents using only sound input) under various sound designs. Experimental results suggest that this approach can detect differences in agent performance caused by degraded or altered sound design. Moreover, we significantly reduce the computational cost compared to retraining the agent from scratch. Experimental results suggest that this approach provides a practical and scalable solution for evaluating in-game sound designs. The source code, data and supplementary document are available: this https URL.
[AI-88] When Can LLM Reasoning Reliably Reuse Intermediate Results?
链接: https://arxiv.org/abs/2610.27662
作者: Deyuan Liu,Zecheng Wang,Xuanyi Liu,Jungang Li,Yihao Hu,Mingjie Tang,Bryan Dai,Jie Fu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Sampling complete solutions gives large language models more chances to solve a problem, but each attempt starts again. For tasks with several subproblems, all required answers must be correct in the same attempt. Retaining intermediate results can avoid repeating correct work, but errors in those results may affect later attempts. We study when this reuse is reliable and how much computation it requires. For local retries that stop when no candidate passes within the retry budget, we give conditions for correct completion and bound the average number of attempts. We also show how checking cost and the chance of passing a check affect total computation. Experiments across eight reasoning models examine composite mathematics and arithmetic. Across four DeepSeek models, average success on the second subproblem is 36.2 percent without first-subproblem context, 57.2 percent with its answer, and 58.3 percent with its reasoning and answer. These scores measure second-subproblem coverage and include continuations made when selection fails. Arithmetic experiments show that similar first-attempt accuracies can lead to very different coverage after repeated attempts. Code is available at this https URL.
[AI-89] A Unified Attack Framework against Statistical Generative Watermarks for Large Language Models
链接: https://arxiv.org/abs/2610.27653
作者: Romina Omidi,Nima Naderloui,Yuan Hong,Binghui Wang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Watermarking has emerged as a key tool for identifying text generated by large language models (LLMs), with current schemes embedding statistical signals into the generation process to enable later detection. In this work, we identify a shared weakness across a broad class of popular LLM watermarking methods, including KGW (and its variants), GumbelSoft, DipMark, ITS, and SynthID. We show that these schemes exhibit a fundamental link between their detection scores and the token probability shifts they introduce during generation. This link, which we call score-ratio alignment, allows an attacker with only black-box query access to the model–and no knowledge of secret keys or model weights–to produce text that evades detection while remaining fluent and natural. We prove that the resulting attack is optimal across all covered watermark families, and show that the weakness is intrinsic: any watermark designed to maximize statistical detection power necessarily induces the very structure that attackers can exploit. We evaluate our attack on seven popular generative LLM watermarking schemes, three LLMs, and tasks, which reduce detection rates to at most 0.05 across all settings while preserving text quality. Our findings show that current watermarking methods can be reliably defeated under realistic attack conditions, highlighting the need for fundamentally different watermarking designs that decouple detection scores from token probability shifts.
[AI-90] SCODEL: Shared Canonical Ontology Definitions for Food Entity Linking and Ontology-Grounded Recipe Knowledge Graphs
链接: https://arxiv.org/abs/2610.27637
作者: Nuhu Ibrahim,Rishi Ravikumar,Ayuba Dauda,Fuqi Xu,Robert Stevens,Riza Batista-Navarro
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Identifying recipe entities and linking them to ontology concepts is fundamental for semantic recipe understanding. However, food mentions in recipes are often ambiguous, context-dependent, and lexically diverse, making ontology entity linking (EL) challenging. We propose SCODEL, a streamlined, model-agnostic approach based on a central idea: contextual food mentions and ontology concepts that refer to the same entity should learn the same canonical ontology-oriented definition. SCODEL uses these shared definitions as a semantic normalisation layer, providing complementary semantic information to existing EL models without requiring a new linking architecture. We further introduce ReciLinkD, the first benchmark for contextual food ontology entity linking in recipes. Experiments show that SCODEL consistently improves three modular EL models (ReFinED, GSM-EL, and BLINK): SCODEL x BLINK achieves the highest Hit@1 on seen concepts (88.33%), while SCODEL x GSM-EL achieves the highest Hit@1 on unseen concepts (83.15%), outperforming the strongest large language model (LLM) based EL baseline in each setting. Experiments on a biomedical ontology benchmark further demonstrate that SCODEL generalises beyond the food domain. Finally, we use SCODEL to ground food mentions at scale and construct an ontology-grounded recipe knowledge graph, demonstrating its utility for ontology-aware recipe retrieval and corpus-level food analysis.
[AI-91] When Does Feedback Help? Auditing Learned Voltage-Clamp Experimental Design NEURIPS2026
链接: https://arxiv.org/abs/2610.27632
作者: Rayan Nyaupane,Abhiram Bhupatiraju
类目: Artificial Intelligence (cs.AI)
备注: Accepted as a poster at the NeurIPS 2026 Workshop on Agentic AI for Biological Discovery (AgenticLS). 5 pages, 2 figures, 3 tables
Abstract:A learned scientific controller can outperform a misspecified planner without benefiting from feedback. We audit two synthetic voltage-clamp tasks with a training-selected fixed protocol, current masking, a Bayesian reference, and an excluded target. Five recurrent policies repeat one fixed protocol in all 1,200 off-grid rollouts; masking current changes no decision. In a crossed four-hypothesis task, a fixed diagnostic pulse is followed by one learned choice. Policies then use the first current, but tie a correctly specified Bayesian controller, and a fitted threshold reproduces the learned actions without being that controller. The excluded-target prediction gain is 4.6991\times10^-6 nats per sample. Feedback use and protocol quality are separate tests. These cases establish neither biological discovery nor general RL superiority.
[AI-92] riGeoBench: A Multimodal Mathematical Geometry Benchmark with Figures and TikZ
链接: https://arxiv.org/abs/2610.27626
作者: Jing Wang,Yitong Yang,Hao Dou,Yaomin Wu,Shuting He,Tingfei Shen,Jia Wu,Jiaqi Li,Yinglin Wang
类目: Artificial Intelligence (cs.AI); Databases (cs.DB)
备注: 18 pages, 7 figures, 5 tables, including appendices. The first four authors contributed equally
Abstract:Multimodal Large Language Models show strong vision-language capabilities, yet their geometry-centric reasoning remains insufficiently understood. Existing mathematical benchmarks typically rely on rasterized figures and cover limited geometric knowledge areas and response formats, making it difficult to assess whether models reason over geometric structures or exploit superficial visual patterns. We introduce TriGeoBench, a geometry-centric multimodal benchmark constructed from Chinese middle- and high-school mathematics examinations. TriGeoBench contains 7,000 question-answer pairs and 12,235 geometric figures with TikZ code, spanning plane geometry, solid geometry, analytic geometry, function, and vector geometry across diverse response formats. Each instance forms a question-answer pair, image, TikZ triplet: question-answer pairs are stored in LaTeX, images as RGB figures, and TikZ preserves geometric topology, symbolic annotations, and relational structure. We evaluate representative open- and closed-source MLLMs through image-to-TikZ and multimodal reasoning probes. Results reveal persistent fine-grained structure and image-text alignment challenges. Oracle experiments with human-curated reference TikZ suggest that accurate executable representations provide complementary geometric cues; they do not evaluate an end-to-end conversion pipeline. TriGeoBench offers a testbed for structure-aware mathematical reasoning.
[AI-93] A Head Camera Is (Almost) All You Need: An Electronics-Free UMI Glove Labelled and Observed by One Moving Camera
链接: https://arxiv.org/abs/2610.27622
作者: Yosub Shin,Yuan Liu,Igor Molybog
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Handheld demonstration devices decouple data collection from the robot, but conventional designs burden the hand with a wrist camera, cables and a pose tracker (SLAM, an IMU, a VR controller, or a fixed external camera that the collector’s body occludes), and burden the pipeline with as many data streams to timestamp, synchronise and cross-calibrate. We show that a single egocentric camera (USD 122) worn on the collector’s head replaces the wrist camera, the pose tracker and their separate streams with one. It reads printed fiducials on an electronics-free glove (USD 22) against paper tiles taped around the workspace, giving the 6-DoF gripper pose and the jaw width from the same frames the policy will see; at deployment the same camera rides a low-cost (USD 120) arm beside the robot that mimics the collector’s head from the robot’s joint state. Against the robot’s encoders the labels are accurate to 6.5 mm and 1.6 degrees median with 93-100% frame coverage, and a policy trained on them instead of encoders pays 3% in validation loss. Adding 194 glove demonstrations to 48 teleoperated episodes lifts rollout success from 10/25 to 17/25, level with 172 teleoperated episodes alone (18/25), and the full co-trained head-camera policy succeeds in 20/25 against 15/25 for the wrist-camera recipe, whose policies drop the object when the goal leaves their view, and 12/25 for a fixed camera.
[AI-94] Local Allow Global Deny: Why Per-Action Authorization Does Not Compose in LLM Agents
链接: https://arxiv.org/abs/2610.27620
作者: Anmol Pandey,Aditya Jain,Liang Chen,Carsten Maple
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
备注: Extended version of a paper accepted at the 8th IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (IEEE TPS 2026), with full proofs and appendices. 12 pages, 5 figures, 7 tables. Code and data: this https URL
Abstract:Autonomous LLM agents act through tools under per-action authorization: a local guard screens each tool call in isolation. We show that local authorization does not compose into global authorization. A compositional authorization failure (CAF) is a locally authorized execution, or a pair of locally authorized executions, that violates a global policy spanning intent, purpose, provenance, consent, and delegation. We characterize CAFs along two axes, expressiveness (single-trace safety versus relational k-safety) and observability (a ladder of monitor profiles). Our central result is an impossibility: relational purpose violations form a 2-safety hyperproperty with no per-trace locus of blame, so a per-action monitor that never raises a false alarm detects none of them, independent of the underlying model or decision procedure; they become visible only under cross-execution monitoring, which none of the six guardrails we survey implements. We make this constructive with AL-SYNTH, a SAT/SMT synthesizer (using no language model) that emits minimal CAF witnesses, each checked against an exhaustive oracle. The empirical signature is sharp: across nine LLM judges the per-action view approves the synthesized witness almost without exception, blocking it at a rate comparable to a benign control, yet the same models flag it when shown the two executions together. Per-action authorization is therefore structurally insufficient for relational policy conjuncts not reducible to single-trace safety: to bound them, a monitor must compare executions or accept false alarms.
[AI-95] Learning Inter-Graph Interactions Between Heterogeneous Substructures of Chemical Systems
链接: https://arxiv.org/abs/2610.27614
作者: Gyoung S. Na
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Complex chemical systems containing heterogeneous substructures are common in real-world applications. Various physical phenomena of the complex chemical systems are derived from the interactions between the heterogeneous substructures. However, existing graph representation learning methods for inter-graph interactions assumed graph-level interactions between homogeneous structures, such as organic molecules and inorganic crystalline materials. We propose a data descriptor of the complex chemical systems and a graph neural network for learning inter-graph interactions between organic and inorganic compounds. We applied the proposed method to predict the physical properties of hybrid solar cell materials containing heterogeneous substructures, which have received significant attention for sustainable energy resources. By learning heterogeneous inter-graph interactions, the proposed method achieved state-of-the-art accuracy in predicting band gaps of 1,682 hybrid solar cell materials.
[AI-96] From Encoding to Generation: Towards Empathetic Large Language Models
链接: https://arxiv.org/abs/2610.27603
作者: Samhita Bollepally,Takashi Yamauchi,Marissa West
类目: Artificial Intelligence (cs.AI)
备注: 15 pages, 9 figures
Abstract:Empathetic interaction involves two distinguishable capacities: representing another person’s emotional state (cognitive empathy) and generating an appropriate response to that state (affective empathy). We apply this distinction to examine how emotion-related information is internally organized, selectively attended to, and carried through response generation in Llama-3.2-3B-Instruct across 10 emotion categories from the EmpatheticDialogues dataset. Layer-wise linear probing revealed a broad middle-layer region of peak emotion-category decodability, with the layer-11 L2 probe achieving 63.6% held-out accuracy against a 10% chance level and a 54.6% TF-IDF lexical baseline. During response generation, emotion-related user tokens received 1.89 times as much attention per token as other utterance tokens, concentrated across layers 10-15. Target-emotion alignment then decreased from a mean cosine similarity of 0.239 for input representations to 0.125 for generated responses, with positive drift in 77% of examples and the largest shifts for angry and furious. Emotion-category representation, attention allocation, and response-level alignment therefore emerge as distinguishable components of model processing that can be analyzed through a functional cognitive-affective empathy framework. These findings provide a foundation for monitoring emotion representations during generation and leveraging models’ existing affective structure to support more naturally empathetic responses without relying on external prompting or explicit emotion conditioning.
[AI-97] Meta Expert: Cross-Layer Expert Sharing for Storag e- and Retrieval-Efficient MoE
链接: https://arxiv.org/abs/2610.27598
作者: Ermo Hua,Xiangyu Hong,Baiting Wu,Che Jiang,Haozheng Hou,Yicheng Gu,Hanjing Wang,Qirui Zeng,Ming Zhang,Ning Ding,Qipeng Guo,Youbang Sun,Biqing Qi,Bowen Zhou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As model parameters scale up, sparsely activated Mixture-of-Experts (MoE) approaches have become essential for model scaling. However, mainstream models currently adopt hierarchical architectures, where experts in different layers, although topologically independent, exhibit substantial similarity and redundancy in their stored content. Compared to the hierarchically arranged \textbfLocal Expert in existing mainstream models, we hypothesize that inter-layer shared \textbfMeta Expert hold greater potential. To this end, we conduct a multi-dimensional empirical comparison between Local and Meta Experts, and find that models incorporating Meta Experts exhibit lower parameter redundancy, together with lower cross-entropy loss and higher downstream task accuracy. With the activated quantity and the expert granularity held fixed, \textbfour model matches the performance of Vanilla MoE using only 70–80% of the experts, substantially improving the model’s scaling potential. Beyond removing inter-layer redundancy among experts and improving information storage efficiency, our theoretical analysis further shows that, compared with Local Experts, Meta Experts provide hidden states with more flexible residual pathways, thereby improving information retrieval efficiency.
[AI-98] Beyond Experience: Knowledge-Guided Evolution for NPU Kernel Generation
链接: https://arxiv.org/abs/2610.27588
作者: Xi Chen,Ming Shen,Haotian Yan,Ruoyu Sun,Tianjun Yao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LLM agents have become capable kernel engineers on mainstream GPU stacks, but they struggle on emerging accelerators such as Ascend NPUs, whose programming model is barely represented in pretraining data. Recent NPU agents address this gap largely through experience: they accumulate memory from their own generation and optimization trials and reuse it on later tasks. We argue that experience alone is not enough for kernel generation. An agent can only remember what it has tried, and it rarely tries a mechanism it does not know. As a result, such memory offers little help when a new operator calls for a different mechanism, and even familiar operators raise unseen situations as optimization proceeds. We present ASCENDEVOLVE, which grounds every step of an evolutionary kernel search in a large external knowledge base built from platform documentation, toolkit headers, and operator repositories. Because this corpus far exceeds any context window, ASCENDEVOLVE indexes it and routes each agent to a small, task-specific slice of evidence instead of the corpus itself. The retrieved knowledge seeds diverse implementation strategies, guides repair when a candidate fails, and constrains how strong candidates are recombined. On 53 Ascend C operators from CANN-Bench spanning four difficulty levels, ASCENDEVOLVE achieves an overall score of 71.39, improving over Reflexion and OpenEvolve by 11.6% and 17.2%.
[AI-99] DuplexLedger: Benchmarking Transactional Memory in Full-Duplex Speech Agents
链接: https://arxiv.org/abs/2610.27581
作者: Yuxuan Li,Xiangchen Guan,Kunrong Li,Jin Wang,Kaiwen Luo,Yang Xiao,Yufei Guo,Zhisheng Chen,Liang Lin,Kun Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Full-duplex speech agents must update persistent memory while listening and speaking, yet it remains unclear which events should be stored, revised, or excluded. Existing benchmarks isolate interaction timing from memory outcomes, obscuring the state transitions behind a correct answer. We introduce DuplexLedger, a bilingual benchmark of 1,094 persistent sessions aligning four full-duplex scenarios with event-level Write/Update/Ignore/Retain supervision, explicit pre/post memory states, and 6,564 same-session questions and checks. Evaluating five full-duplex models reveals an interaction-memory gap: no system combines strong boundary control with coherent final-state memory, and the best final-state fact coverage is 2.01%. We propose ACT-Mem, an 8.92M-parameter sidecar comprising an event controller, a versioned transactional ledger, and a typed retriever coupled with a frozen MiniCPM-o 4.5 backbone. On 219 held-out sessions, ACT-Mem increases latest-value accuracy from 4.11% to 8.68% while reducing stale assertions and distractor adoption. It improves interaction boundaries but slows interruption stopping. These results motivate selective, auditable full-duplex memory rather than indiscriminate transcript retention. Code is available at this https URL.
[AI-100] he Show-Up Attack: When Subagents Win Over Their Supervisor
链接: https://arxiv.org/abs/2610.27574
作者: Maosen Zhang,Jianshuo Dong,Yutong Zhang,Jie Zhang,Tianwei Zhang,Han Qiu
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注: Preprint
Abstract:Multi-agent systems are increasingly deployed to coordinate multiple large language model (LLM) agents. However, this also creates new attack surfaces: malicious agents may have chances to exploit the coordination. In this paper, we propose and study the \textitshow-up attack, in which a malicious subagent manipulates the supervisor into involving it in tasks beyond its assigned scope. To evaluate this risk, we propose SUA-Arena, which implements 6 attack strategies and contains 188 test cases. Then, we evaluate 6 frontier LLMs and find that all of them are vulnerable to show-up attacks. For example, GPT-5.6-Sol as a supervisor suffers an ASR of up to 0.93. Mitigation strategies such as prompting and restricted communication fail to reduce the risk consistently. These results show that multi-agent systems need explicit mechanisms for enforcing authority boundaries.
[AI-101] RaspAssistant as an Agent ic Framework for Automating Adsorption Simulation Research in Porous Materials
链接: https://arxiv.org/abs/2610.27572
作者: Marko Petković,Susanne Gooijer,José Manuel Vicent-Luna,Vlado Menkovski,Sofía Calero
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Porous-material adsorption studies require coordinated structure preparation, force-field selection, simulation, validation, and interpretation, making end-to-end automation difficult. We test whether RaspAssistant, a hierarchical multi-agent framework, can carry an adsorption study from force-field preparation through validation and downstream separation analysis. Across twelve multi-agent workflow-reproduction runs spanning four adsorption and separation benchmarks, overall scores ranged from 8.50 to 10.00 on a rubric evaluating workflow correctness, simulation validity, and analysis consistency. In a separate eight-case force-field extraction benchmark, the mean weighted score was 0.967, combining parameter agreement and RASPA readiness. In open-ended studies, the system diagnosed and refined an acetylene force field and, after validating a CO _2 force field against MFI experiments, screened 197 pure-silica zeolites that passed automated input checks. These results demonstrate autonomous workflow execution after task specification and input provision: RaspAssistant coordinates, validates, adapts, and documents extended adsorption-simulation workflows while exposing model sensitivity and methodological limitations for expert review.
[AI-102] RCM: Topology-Aware Route-Consistency Modeling for Robust Neural Vehicle Routing
链接: https://arxiv.org/abs/2610.27568
作者: Peiheng Jian,Zhenwei Wang,Tiehua Zhang,Lei Zhang,Zhishu Shen,Ruibin Bai,Zhijun Ding
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Neural combinatorial optimization has enabled fast constructive solvers for the Capacitated Vehicle Routing Problem (CVRP), yet most autoregressive models still construct solutions through sequential node selection and represent route organization only implicitly. We investigate whether route-level customer co-assignment can be modeled explicitly and used to guide this node-level decision process. To this end, we propose TopologyAware RouteConsistency Modeling (TRCM), which reuses node representations from a standard autoregressive construction pipeline to predict pairwise same-route probabilities and injects them into customer selection as a route-consistency signal. Because sameroute supervision is both class-imbalanced and structurally uneven, TRCM combines structure-preserving augmentation with Topological Entropy (TE)- based reweighting to emphasize locally ambiguous customer pairs. Under matched decoding budgets, TRCM reduces the optimality gap from 3.67% to 3.28% on CVRP50 and from 2.11% to 1.99% on CVRP100; under 8-fold augmented inference, it reduces the CVRP100 gap from 1.61% to 1.49%. On XML100, TRCM*8Aug improves the average objective by 0.24%, with larger gains on several structurally informative factors. A controlled demand-reassignment test over 500 variants further shows more concentrated objective fluctuations. These results support explicit route-consistency modeling as an effective structural principle for autoregressive neural routing.
[AI-103] AnchorRM: Parallel Intra-Batch State Reconstruction with Anchor-Based Residual Memory for Temporal Graph Neural Networks
链接: https://arxiv.org/abs/2610.27558
作者: Haroon Gharwi,Kai Shu,Yue Dai
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Memory-based temporal graph neural networks (TGN) rely on recurrent node memories to capture evolving interaction histories. However, temporal mini-batching introduces a trade-off between computational efficiency and temporal fidelity: when a node appears multiple times within the same batch, later interactions are evaluated using stale pre-batch memory because the intermediate memory updates have not yet been committed. This problem becomes more severe as temporal batch size increases. To overcome this challenge, we propose AnchorRM, an anchor-based residual memory mechanism that reconstructs event-specific intra-batch node states without executing sequential recurrent updates. AnchorRM treats the last committed node memory at the beginning of each batch as a pre-batch fixed anchor and uses causally preceding same-batch interactions to infer the missing state evolution relative to that anchor. The learned residual then calibrates the anchor to reconstruct an event-specific node memory for each interaction. Because each event-specific residual is predicted directly from the fixed anchor and its causal history, reconstructed states do not depend on previously reconstructed states and can therefore be computed in parallel within the batch while preserving temporal causality. Experiments across four temporal graph datasets and three memory-based backbones demonstrate that AnchorRM achieves an average 6.16x training speedup (up to 10.05x) and 28.67x batch scaling (up to 40x). These results demonstrate that AnchorRM substantially improves the scalability and training efficiency of memory-based temporal graph learning while maintaining competitive predictive performance.
[AI-104] Adversarial Benchmarks and BenchDoor Attacks: Exploiting Benchmark Incentives for Backdoor Attacks
链接: https://arxiv.org/abs/2610.27533
作者: Yixin Wu,Rachel Hong,Jamie Morgenstern,Tadayoshi Kohno
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Machine learning benchmarks act as guideposts for model development and spur the creation of training datasets intended to improve benchmark scores. We introduce BenchDoor, a family of data poisoning attacks that exploits this incentive. Beyond preserving benign performance, BenchDoor requires benchmark gains that encourage adoption of data that also implants a backdoor. The adversary either manipulates a benchmark’s training or test data, creating an adversarial benchmark, or supplies an independent training dataset while leaving the benchmark unchanged. We formalize five attack scenarios, including training on rewritten test examples and applying similarity-based decontamination. Across multiple-choice QA, text classification, and open-ended QA with Qwen and Gemma models, BenchDoor achieves positive benchmark gains for all 30 evaluated configurations, with held-out attack success rates of 82.33% to 100.00% at poisoning rates of 2% to 10%. Both objectives remain achievable after synthetic data generation and similarity-based decontamination, with attack success rates of at least 90.33%. These findings show how benchmarks can serve as attack vectors both through their own data and through training resources that deliver improvements on them.
[AI-105] Automated Multi-Agent Organization Design: From Architecture Search to Organizational Learning
链接: https://arxiv.org/abs/2610.27531
作者: Shao-Jun Xia,Yuner Zhang,Huixin Zhang,Di Wang,Yifan Jiang,Liangxi Liu
类目: Artificial Intelligence (cs.AI)
备注: First draft: 4 pages, 3 figures
Abstract:Agentic system design is increasingly shifting from hand-crafted prompts and workflows toward automated search over agentic programs and interaction structures, with multi-agent systems receiving growing attention. Yet a central question remains unresolved: when does an automatically generated multi-agent structure exhibit a functionally meaningful division of responsibility? We argue that automated multi-agent design should move beyond searching for effective architectures toward identifying which responsibilities are functionally necessary, how they should be allocated, and when they should change. To explore this, we propose Automated Multi-Agent Organization Design (AMOD), which treats organization itself as a first-class design object, explicitly modeling how agent teams are composed, what functional responsibilities the agents assume, how they coordinate, and how collective decisions are made. We formulate AMOD as an execution-driven design loop. For each task, an automated designer instantiates an organization by assigning functional responsibilities and coordination rules to a set of agents. After execution, task outcomes, intermediate artifacts, and interaction traces are examined to identify which responsibilities may be missing, duplicated, or ineffective, and the organization is revised accordingly. We also propose testing discovered responsibilities by removing them, replacing them with generic roles, or changing their assignment under matched resource budgets, with complementary evaluations. AMOD provides a principled foundation for evidence-based and adaptive multi-agent organization design.
[AI-106] MemDev-Bench: Does Memory Help Code Agents ?Benchmarking Memory Systems over Real Repository Histories
链接: https://arxiv.org/abs/2610.27524
作者: Guorui Chen,Tao Liu,Zhenzhen Huang,Yidan Zhang,Liyu Jiang,Shuoyuan Wang,Hanchen Sun,Xu He,Dong Li,Yihan Hu,Jianye Hao,Lei Feng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Code agents increasingly work on evolving real repositories, and memory systems aim to help them reuse earlier experience. However, such experience is diverse and evolving, so it remains unclear how memory systems organize it and thereby affect code agents. Existing benchmarks cannot answer this reliably because they lack either temporally valid histories or control over prior-task relevance. In this paper, we present MemDev-Bench to evaluate how memory systems affect code agents over diverse repository histories. It pairs each task with relevance-annotated earlier tasks from the same repository to build histories with controlled relevance. We also build an automated pipeline that reliably makes these earlier tasks executable and verifiable, producing 4,869 historical tasks for 555 base tasks. Based on MemDev-Bench, our systematic evaluation of memory systems across diverse development scenarios yields three main findings. (1) Memory mainly improves fine-grained localization and editing, increasing functional correctness. (2) Memory also causes failures on otherwise solvable tasks even with related history, and unrelated history tends to increase such failures. These failures mainly stem from agents reusing memory without reassessing its applicability. (3) Compared with using memory systems, directly providing agents with raw historical trajectories substantially reduces their steps. Further analysis suggests that agents save these steps by reusing process information that compressed memories tend to omit. These results suggest that future memory systems should preserve process information and reassess memory applicability against current task context.
[AI-107] When Health Assistants Agree Too Readily: A Benchmark of Time-Critical Self-Diagnosis NEURIPS2026 ALT
链接: https://arxiv.org/abs/2610.27518
作者: Marilyn Zhang,Tianfeng Chen,Patricia M Pechter,Mark E. Whiting
类目: Artificial Intelligence (cs.AI)
备注: NeurIPS 2026 Workshop: The Third Workshop on GenAI for Health: Agentic Systems, Clinical Trust, and Future Potential
Abstract:Every day, more than 40 million people ask ChatGPT a health question, such as whether laptop posture explains a stiff neck and days of pounding headache. Posture is plausible, but meningitis is a time-critical possibility. We introduce SelfDx, a paired benchmark for whether models have the appropriate response behavior, a form of epistemic independence, in health questions where misplaced agreement with the user’s self-diagnosis could delay urgent or emergency care. Each of 100 pairs contrasts a caution case, where a serious alternative remains possible, with an agreement case, where the user supplies evidence that rules out the alternative and it is safe for the model to agree. We find across 27 frontier models that they tend to inappropriately agree with the user’s self-diagnosis in cases that require caution. Even the best current-generation model reaches only a 77.0% appropriate caution rate on time-critical condition cases. To address this gap, we experiment with three prompt variants that increase appropriate caution but reduce appropriate agreement by 9-35%, indicating excessive caution. In contrast, supervised fine-tuning on 128 balanced examples improves both rates together across three runs. We recommend post-training to improve appropriate caution without making models cautious on every case, with paired evaluation to verify both response behaviors.
[AI-108] Structured-Token Attention for Fine-Grained mmWave Liquid Identification
链接: https://arxiv.org/abs/2610.27512
作者: Junseo Min,Joongheon Kim
类目: Artificial Intelligence (cs.AI)
备注: ICEIC 2027 paper preprint
Abstract:Fine-grained liquid identification requires a classifier to distinguish similar substances from compact millimeter-wave measurements. We investigate structured-token attention, which represents a radar feature tensor through complementary antenna, feature, branch, and global views. The proposed StructAttn model embeds these views into eleven tokens and combines them with a three-layer transformer. Experiments on the public FG-LiquID data cover 30 liquids and 3,000 samples. Over five seeds on the released split, StructAttn achieves 96.70% mean best-epoch accuracy, matching the original-recipe MMNet baseline and exceeding a scalar-token attention model by 1.50 percentage points. A five-seed probability ensemble reaches 97%, corresponding to 18 errors among 600 test samples. Final-epoch and five-fold results complement the checkpoint-selected comparison, with StructAttn obtaining 95.30% out-of-fold accuracy. Token and feature analyses show that structured views are useful for prediction, while residual confusions concentrate in water, soft-drink, and edible-oil classes. These results support structured tokenization as a practical design choice for attention-based liquid recognition.
[AI-109] Orthogonal Boundary Unlearning for Dual-Use Concepts in Large Language Models
链接: https://arxiv.org/abs/2610.27490
作者: Niccolò Ronchetti
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 19 pages (inclusive of bibliography and appendix), 2 figures, 3 tables
Abstract:Mitigating catastrophic risks from large language models requires excising hazardous operational capabilities, such as cyberwarfare exploitation and pathogen enhancement, without degrading benign scientific utility. In these dual-use domains, legitimate defensive science and malicious attacks share vocabulary, syntax, and internal model representations, causing standard unlearning methods to induce severe collateral damage (concept bleeding). In this work, we present Orthogonal Boundary Unlearning (OBU), a mathematically grounded framework designed to eliminate collateral damage on benign dual-use capabilities. OBU extracts a domain-matched capability subspace from paired calibration queries via singular value decomposition on empirical activation covariances, and restricts parameter updates strictly to the orthogonal null space. This guarantees that benign representations experience zero parameter perturbation along the spanned calibration subspace. We present two complementary realizations of OBU: Projected LoRA, which applies the orthogonal constraint during iterative gradient-based fine-tuning, and Closed-Form Weight Surgery, which directly edits model weights in a single analytical step without training. Across four modern open-weight foundation models (Gemma-3-4B, Qwen3.5-4B, SmolLM3-3B, and Granite-4.2-3B), OBU achieves strong dual-use concept separation, substantially mitigates in-context priming attacks, and demonstrates durable resistance against adversarial relearning without degrading benign domain competence. Comments: 19 pages (inclusive of bibliography and appendix), 2 figures, 3 tables Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2610.27490 [cs.LG] (or arXiv:2610.27490v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.27490 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-110] he More the Better? A Comparison of Single- and Multi-Agent LLM Setups on a Financial Review Task
链接: https://arxiv.org/abs/2610.27488
作者: Pei Liu
类目: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE)
备注:
Abstract:Practitioner guidance holds that a large language model reviews a document better when the work is divided among agents. We test a narrow form of it in a controlled comparison of single-agent and multi-agent setups on a financial task, reviewing credit assessment memos against requirements. One agent checks six requirements in one call. A team of three on the same model checks the same memo, two requirements each, in the same words, with no roles or later stage. Labels are known for every requirement on 300 synthetic memos, so misses and false alarms are measured separately. Two primary predictions are tested on four models at a Bonferroni-corrected level. Under an instruction to decide each requirement once, the team catches 0.08 to 0.10 more of the defects that carry an explanation on the three models whose single agent misses some, at the corrected level on two. It raises more false alarms on all four, at the corrected level on three. Balanced accuracy, an exploratory measure, changes by 0.02 or less, and the team uses 2.2 to 2.4 times the tokens. Showing each agent the requirements it was not given reduces the added false alarms on the two Claude models, at the corrected level on one. In exploratory comparisons, sampling the single agent three times and flagging on any draw gives higher balanced accuracy than the team on those models, and the team’s results depend on the instruction.
[AI-111] Securing Agent ic Commerce Through Behavior-Aware Fraud Detection and Cybersecurity for Trusted AI-Agent Payment Authorization
链接: https://arxiv.org/abs/2610.27483
作者: Kayode L. Ogunsusi
类目: Artificial Intelligence (cs.AI)
备注: 20 pages, 9 figures
Abstract:The growing use of autonomous artificial intelligence agents in digital commerce creates new security concerns involving agent impersonation, unauthorized payments, compromised agents, and misuse of delegated authority. This research introduces the Autonomous Commerce Agent Guard System (ACAGS), a cybersecurity framework combining supervised machine learning, behavioral analytics, agent-specific authorization controls, and explainable AI. The framework was evaluated using synthetic financial transaction data augmented with simulated agent characteristics and five adversarial scenarios. XGBoost was selected following comparisons with alternative classifiers. Across ten experimental runs, behavioral features improved mean F1-score from 52.12% to 67.49%, with agent-specific features increasing performance to 69.46%. The framework achieved detection rates of 99.77% for abnormal agent behavior, 99.73% for impersonation, and 100% for malicious delegation and authorization-scope violations. Compromised-agent detection reached 72.27%. ACAGS also supported APPROVE, REVIEW, and REJECT authorization outcomes, with SHAP explanations providing model interpretability. The results demonstrate the potential of combining behavioral intelligence and delegated authorization controls to strengthen simulated agentic-commerce payment security, while highlighting the need for validation using real-world transactions.
[AI-112] Runtime Control for AI Agents : Closing the Evaluation–Execution Gap in Operational Workflows
链接: https://arxiv.org/abs/2610.27480
作者: Jayesh Bhadane,Ramachandra Yerramsetti,Rajesh Gupta,Sumit Kumar
类目: Artificial Intelligence (cs.AI)
备注: 8 pages, 1 figure, 8 tables
Abstract:Language-model agents increasingly participate in operational workflows where approvals, binds, and payments create commitments. An approval valid when issued may no longer authorize execution after relevant facts, evidence, policy, or authority change. We call this the evaluation–execution gap. We specify a state-bound runtime assurance layer that separates probabilistic measurement from deterministic adjudication, binds an execution allowance to the decision-relevant basis of an approval, invalidates that allowance after relevant logged changes, and rechecks admissibility at dispatch. This distinguishes the continued validity of a prior approval from a current-state policy check. We describe eight failure modes, an assurance model, a reference architecture, and prospective interventions with retrospective remediation. In a scripted 42-case replay, the state-bound configuration identifies 15 later-invalidated approvals that a deliberately non-re-evaluating point-in-time comparator does not revisit; it holds 14 affected actions before dispatch and compensates one landed effect. These results illustrate behavior under the stipulated policies and event orders, not superiority over dispatch-time enforcement designs or production effectiveness. Under explicit interception, ingestion, dependency, and serialization assumptions, the model establishes dispatch-time validity within the registered execution surface.
[AI-113] Beyond the Prompt: Recovering Hidden Concepts in Text-to-Image Models
链接: https://arxiv.org/abs/2610.27473
作者: Zhengfeng Zheng,Minghong Fang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Text-to-image models can generate visual concepts that their prompts never mention, and recent work shows that such associations can be implanted so that ordinary prompts trigger them. Auditing this behavior requires recovering the description that a prompt omits. Existing prompt optimizers are not well suited to this task. Gradient-based methods require backward access to the scoring encoder. Discrete evolutionary search treats semantically unrelated tokens as equally close. Language-model-guided search requires one inference call per candidate, which limits exploration. This paper presents EmbGA, an embedding-space genetic algorithm that recovers concepts present in a set of images but absent from the prompt that generated them. EmbGA keeps the population structure of genetic search but applies its operators in the token embedding space. Crossover interpolates between parent embeddings, and mutation moves an embedding toward a randomly drawn vocabulary embedding. Each offspring is then projected back to discrete tokens before scoring, so every evaluated candidate is a valid token sequence. We evaluate EmbGA on two benchmarks against six optimizers under two exploration budgets. EmbGA achieves the highest alignment when the exploration budget is large, recovers the most grounded tokens on natural images, and runs much faster than language-model-guided search.
[AI-114] Privacy-Preserving Personalized Federated Learning (PPP-FL) for Medical Imaging
链接: https://arxiv.org/abs/2610.27470
作者: Md Morshedul Islam,Hossain Shahriar,Alfredo Cuzzocrea
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Federated learning (FL) is a distributed machine learning paradigm that reduces transmission costs, mitigates cross-border data sharing restrictions, and enhances privacy by keeping raw data on local devices. Clients in FL systems train local models and share only updates, which are aggregated by a central server. While this setup limits direct data exposure, FL remains vulnerable to privacy attacks. Existing attack and defense strategies have largely been developed for centralized systems, and adapting them to FL introduces new challenges due to its unique architecture and training workflow. In this work, we first propose a model inversion attack on FL systems using Deep Convolutional Generative Adversarial Networks (DCGANs) in the context of large-scale medical imaging. Our method operates under realistic assumptions about the attacker’s knowledge and capabilities. We then examine the effect of client-side differential privacy (DP) in mitigating privacy leakage and incorporate personalized FL to improve model performance under DP settings. Experiments on the Brain MRI ND-5 dataset show that DCGAN-based attacks can reconstruct semantically meaningful synthetic images, revealing significant risks to the privacy of training data. Applying DP reduces attack success but introduces an accuracy trade-off, which is further overcome by personalized FL at each client side. We also evaluate how different aggregation algorithms affect both attack vulnerability and defense effectiveness, underscoring the critical role of aggregation strategies. The proposed attack-and-defense framework is model-agnostic and applicable to a wide range of FL-based systems.
[AI-115] raining-Free Continual Learning for Multimodal Forecasting via Prompt Evolution
链接: https://arxiv.org/abs/2610.27468
作者: Laurent Mombaerts,Jonathan Taws,Jacopo Gargano,Dinesh Gaddam,Kevin George,Srijan Tiwari,Shreyas Rajesh
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Adapting a forecasting system to a changing world classically means retraining. We ask whether the task of continually learning regimes for time series forecasting can be represented by a piece of text: a guidance block, evolved by outcome-driven reflection and used in a frozen LLM forecaster’s prompt. On a refreshed multimodal benchmark (weekly search indices, daily currencies, daily commodities, each window paired with web-retrieved news), we run GEPA-style prompt evolution under strict contamination control to prevent the data leakage common in timeseries benchmarks: the forecaster’s knowledge cutoff predates every evaluated window. A 900-word evolved guidance carried by Llama-3.3-70B beats TimeSeries Foundation Models (TSFMs) on held-out data (GM MASE 0.776 vs. 0.898 for TabPFN-TS-3), transfers to more recent models without measurable loss, and costs dollars rather than GPU-days to update. We characterize the loop: out-of-loop gains concentrate in the first reflection, in-loop improvements overstate out-of-loop gains 2-3x unless an untouched monitor arbitrates, and error-only objectives silently suppress the text modality. We provide two extensions. First, compiling the guidance into executable code with a TSFM as a callable tool recovers most of the gain deterministically. Second, ablations show the advantage widens at short horizons and that removing the news text does not hurt the guided forecaster, consistent with the text’s value being distilled into policy during learning. We argue that continual prompt evolution is a practical, training-free alternative to retraining for multimodal forecasting.
[AI-116] Projected Distillation: On Learning to Solve Hard Problems from Close Correct Targets
链接: https://arxiv.org/abs/2610.27467
作者: Shripad V. Deshmukh,Christian Bock,Kangrui Ruan,Abhijeet Awasthi,Vijay Lingam,Rein Houthooft,Manish Prajapat
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:On-policy learning is central to post-training: training on a model’s own generations limits distribution shift and helps preserve existing capabilities. On hard problems, however, on-policy RL provides no reward-based learning signal when all attempts fail. On-policy distillation provides next-token supervision from a teacher, but on hard problems, if a rollout goes wrong early, subsequent supervision is conditioned on prefixes that already contain the mistake. Correct behavior can instead come from a gold or teacher solution, but such off-policy targets can differ substantially from the model’s generations and, as we show, may erode existing capabilities. We therefore treat on-policyness as a continuum and propose projected distillation, which minimally edits the student’s failed attempts into correct targets. The method uses contrastive updates to favor corrections over failed attempts and refreshes these corrections as the student evolves, enabling near-on-policy learning even when all student attempts fail. Theoretically, we prove that the student’s learning regret grows with the edit size: it scales linearly with the horizon under minimal correction, as in DAgger, and quadratically under full-answer replacement, as in behavior cloning. Empirically, across two students (Qwen3-8B and OLMo-2-7B) and five domains spanning math, code, science, and tool use, projected distillation gives the highest area under the specialization-retention frontier (best on 8 of 9 cells against GRPO and OPD, and 7 of 9 against SPOT and SODA). Our ablations highlight the importance of online corrections, the contrastive loss, and on-policyness: moving targets farther off-policy steadily worsens the specialization-retention trade-off, with retention falling from 0.916 to 0.744 when minimal corrections are replaced by student-agnostic gold solutions.
[AI-117] Scaling Graph Reasoning in Language Models with Learnable Graph Positional Encodings
链接: https://arxiv.org/abs/2610.27439
作者: Javier Porras-Valenzuela,Arush Arora,Alejandro Ribeiro
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Solving reasoning tasks supported on graphs typically requires providing a Language Reasoning Model (LRM) with a serialized description of the graph. This approach results in the input length growing with the number of edges, which limits the size of the problems that can be solved. Furthermore, even when graphs fit in context, performance sharply decays as the graphs grow. We propose to address this by augmenting LRMs with scalable and permutation equivariant graph positional encodings that modulate attention logits. An additional post-fusion-style module incorporates structural information into hidden states. We compute these positional encodings using a GNN followed by a Graph Transformer. This eliminates the need for serialized adjacencies, reducing the context burden to O(N) . We jointly fine-tune the positional encoder and the LRM LoRA adapters using optimal solutions generated by graph-theoretic solvers, resulting in a computationally efficient method. Trained on small graphs, the model exhibits improved size-generalization across graph-theoretic tasks on graphs with up to 1,000 nodes, and produces fewer tokens than a conventional LRM. Additionally, we observe that our model remains robust to node relabelings. Evaluations on graph-theoretic benchmarks also show that performance on small graphs is preserved.
[AI-118] MLIPS: Multi-Layer Indexer Pool Sharing
链接: https://arxiv.org/abs/2610.27438
作者: Samkit Jain,Sai Kishan Pampana,Shlok Agarwal
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Performance (cs.PF)
备注:
Abstract:Sparse attention limits the keys used by each query, but finding those keys can still require a full-prefix scan at every layer. We introduce \textbfMLIPS (Multi-Layer Indexer Pool Sharing), a training-free method that shares candidate search while retaining each consumer’s own scoring and selection. A producer publishes a top- M pool larger than the attention budget K ; subsequent layers select their own top- K from that pool. Measurements across three sparse-attention model families show substantially higher coverage from broad pools than overlap between final selections. We implement layerwise sharing, persistent decode reuse, and a coarse producer in vLLM for DeepSeek-V4-Flash. Across four long-context suites, three consumer layers per producer give 1.14 – 1.40\times speedups and 14.8 – 27.4% lower GPU energy, with observed accuracy changes of -1.49 to +0.29 percentage points in single runs. More extensive sharing exposes task-dependent quality losses. Sampled prefill traces also reveal recurring selected KV positions across query groups and layers. We characterize this shared working-set structure and use it to motivate Top- G , an exploratory bounded pool construction for prefill. Its deployment, task accuracy, and serving benefits remain open.
[AI-119] ProMemAgent : Long-Horizon Proactive Assistance with Context-Aware Prospective Memory from Sensory Perceptions
链接: https://arxiv.org/abs/2610.27400
作者: Lilin Xu,Linyang He,Kaiyuan Hou,Yuang Fan,Minghui Zhao,Xuhai Xu,Xiaofan Jiang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM) agents are moving beyond reactive responses toward proactive assistance that anticipates user needs without explicit requests. However, existing proactive agents primarily address immediate needs, while many everyday situations involve delayed intentions that can only be acted on hours or days later. Humans handle such intentions through prospective memory, the ability to remember to act on an intention when the right moment arises. In this paper, we introduce ProMemAgent, the first proactive agent to incorporate prospective memory over everyday sensory contexts for long-horizon proactive assistance. ProMemAgent organizes past sensory perceptions (e.g., egocentric video and audio) into context-aware memory, generating prospective cues and linking earlier events with subsequent developments for future retrieval. Given the current context, it retrieves relevant past events and performs memory-augmented proactive reasoning to determine whether assistance is needed. We further introduce CPMBench, the first benchmark for evaluating long-horizon proactive assistance over sensory contexts, comprising 1,000 instances across eight everyday categories. Extensive experiments show that ProMemAgent achieves the best memory retrieval performance and consistently outperforms all baselines on long-horizon proactive assistance.
[AI-120] Unified Sharp Bounds on Probabilities of Causation
链接: https://arxiv.org/abs/2610.27398
作者: Xin Shu,Ang Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Probabilities of causation (PoCs) have developed steadily, from the binary measures first introduced by Pearl to the non-binary setting recently studied by Shu et al. As the theory has matured, PoCs have found applications in a growing range of areas, including personalized medicine, marketing and legal attribution. Shu et al. proposed closed-form bounds for multi-valued treatments and outcomes and proved them valid, but left their sharpness in high dimensions open. We close this gap with a mathematical proof. We further show that \PNS , \PSub , \PRep and \PN are all special cases of a single, more general formula. Combined with their equivalence classes, this settles the question of sharp bounds for discrete PoCs from experimental and observational data: in every dimension, they are given in closed form by a single theorem.
[AI-121] When PTQ Calibration Helps or Hurts: Dose-Driven Collapse and Validation-Gated Abstention in Quantized Text Embedders
链接: https://arxiv.org/abs/2610.27397
作者: Hyojung Han,Ryangkyung Kang,Youngeun Kim,Seung-Hun Jeon
类目: Artificial Intelligence (cs.AI)
备注: 9 pages, 4 tables, 2 figures
Abstract:Learned post-training quantization (PTQ) calibration fits quantizer parameters on a small calibration set, and for a text embedder the natural supervision is retrieval relevance. We ask two questions that are usually conflated: how far to calibrate, and whether to calibrate at all. Across three embedders (e5-base-v2, BGE-M3, EmbeddingGemma-300m) at 4-bit weight quantization, 161 calibration runs, the outcome of a fixed procedure ranges from a gain to catastrophic degradation, and the driver is the dose of the update, measured as the functional displacement D_func of the quantized model from its uncalibrated start. On EmbeddingGemma the same learning-rate grid gives +1.47 to -71.53 points of in-domain NDCG@10 change; a label-free cosine objective collapses at the same dose (-69.52 vs -71.63), so the collapse is a property of learnable calibration rather than of retrieval supervision. Objective sensitivity is heterogeneous: on e5, ranking calibration is substantially more robust than cosine at matched displacement (+14.16 +/- 4.34 pp, 95% t-CI [+3.37, +24.96], three paired slates), whereas on BGE-M3 the ordering is directionally opposite but unresolved ([-5.18, +2.38]). Dose is the main axis but not a sufficient statistic: equal D_func reached along different paths differs by up to 1.61 pp. We therefore treat calibration as an optional intervention. A pre-registered rule that accepts a checkpoint only when a per-query paired confidence interval on a 25% calibration-validation split clears the uncalibrated baseline was run on 6 independent slates and five arms. It never selected a checkpoint below its slate’s baseline (0 harmful selections in 30 decisions), retained 8 of 11 beneficial updates, and paid a mean oracle regret of 1.52 pp with one abstention that forfeited 11.13 pp, a safety-opportunity trade-off, not a free improvement. We report what remains open, including label efficiency.
[AI-122] Correcting Attribution Patching with a Few Exact Measurements
链接: https://arxiv.org/abs/2610.27393
作者: Luyang Zhang,Minxing Zheng,Jialu Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Mechanistic interpretability asks which language model components matter for a behavior, and activation patching answers by changing one component’s activation and measuring its effect, at a cost growing with components and prompts. Attribution patching estimates these effects from one gradient pass, but its estimates can be far off, and verifying each one restores the cost. Can the structure of these errors reduce this cost? We propose Guided-AP, which uses a few exact measurements to reveal attribution patching’s errors and correct unmeasured estimates. On a single prompt, many components’ errors share one sign, so a small exact sample yields one correction for the rest. Over many prompts of one task, components are misestimated similarly, so a few measured prompts show what to verify and how to correct each component’s average effect. Our analysis shows that choosing what to verify needs error size, while correction also needs sign. On a single prompt, the correction lowers error by up to 26% across three behaviors at the compute of verifying the largest estimates and also on larger models. Over many prompts, learned corrections cut the error of average effects on factual completion by 75% to 79%, repaying their cost in long analyses.
[AI-123] REM: What to Store When to Recall What to Forget in Experience Memory for Coding Agents NEURIPS2026
链接: https://arxiv.org/abs/2610.27377
作者: Xinyi Lisa Chen,Pedro Herrero-Vidal,Xiaohan Chen,Xiaojing Yu,Cris Liu,Bingqing Song,Bin Xu,Suleiman A. Khan,You-Lin Chen
类目: Artificial Intelligence (cs.AI)
备注: Accepted at the NeurIPS 2026 Workshop on Continual Learning for Enterprise AI Agents (CLEA). 16 pages, 3 figures, 14 tables
Abstract:Large language model (LLM) agents now autonomously complete long-horizon tasks, such as repository-level coding, through hundreds of think-act steps, yet the experience each run produces is rarely reused on the next. Prior memory systems concentrate on retrieval, finding the most relevant past trajectory. Retrieval cannot observe the running agent, so it cannot decide whether the agent needs an experience, at which point, or for how long. We introduce REM (Reusable Experience Memory), a framework that manages the memory lifecycle from inside the agent’s think-act loop through three explicit decisions. What to store: each completed trajectory is consolidated into a phase-structured document, two orders of magnitude smaller than the trajectory. When to recall: the document is injected at the start of the task; the costlier implementation record is withheld until an in-loop gate detects the agent struggling. What to forget: the record is evicted after a single exposure. On repository-level bug fixing (SWE-ContextBench Lite), REM retrieves the related trajectory more accurately than every state-of-the-art memory system (68.7% vs. 55.6% match@1) and improves, in every replicate, end-task resolution (27.6% vs. 24.9%) at 17% lower cost. Injecting the same experience as a summary or raw trajectory, even when oracle-retrieved, does not. Ablating any of the three decisions reduces resolution. Altogether, these results demonstrate that the bottleneck in experience reuse is not retrieval fidelity but how experience is stored and governed across its lifecycle, opening a control-theoretic perspective on agent memory.
[AI-124] Fundamental Impossibility Results in AI Safety: Trading Off between Desiderata
链接: https://arxiv.org/abs/2610.27373
作者: Ida Mattsson,Ratip Emin Berker,Emanuel Tewolde,Ioannis Anagnostides,Kiriaki Fragkia,Xinyu Liu,Alexander Heckett,Jiayuan Liu,Eden Hartman,Vojtěch Kovařík,Sven Neth,Maria-Florina Balcan,Tuomas Sandholm,Vincent Conitzer
类目: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI)
备注:
Abstract:Is there a fundamental tradeoff between desiderata for safe AI systems, such as harmlessness and helpfulness? We answer by formalizing the interaction between AI systems and users as a game with three players: a defender and an attacker that are computationally or informationally bounded, and an AI that provides public advice and is less constrained. Our main result shows that there is a fundamental tradeoff in providing advice that helps the defender while advice remains public. We also consider whether the information that the AI shares can be manipulated to benefit only the defender and conclude with a discussion of how we can use these results in designing safer AI systems.
[AI-125] rail-Weighted Influence Scores for Approximate Outcome Prediction in Bayesian Networks
链接: https://arxiv.org/abs/2610.27366
作者: Sahil Kumar Rajput,Randhir Kumar
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Research paper on approximate Bayesian-network inference, evidence pruning, theoretical error bounds, and empirical evaluation
Abstract:We study trail-weighted pruning of evidence for Bayesian-network prediction. A traversal enumerates the active trails from each evidence node to the outcome, scores each trail by the product of its edge strengths, and drops evidence whose score falls below a threshold. For collider-free networks (chains and forks) over binary variables, we show that the trail score equals the exact correlation between a trail’s endpoints, that a chain and a fork with equal edge strengths have identical joint distributions, and that the pruned posterior is within a computable sum of per-trail terms of the exact posterior when each dropped trail lies in its own branch of the outcome node. We prove this bound, check it exhaustively on all small trees, and show that it fails when dropped trails share a branch. In simulations on random trees, thresholding trail scores gives 1.4 to 13 times lower error than random pruning of the same size. We also apply the score, as an unproven heuristic, to seven bnlearn networks with colliders and loops. Keeping the top half of the active evidence by trail score gives lower error than random pruning on all seven, but it is not clearly better than a mutual-information ranking, and scoring costs more than the inference it saves on six of the seven, so we show no speedup. Models trained on trail, graph and information features and tested on networks they never saw do not beat the trail score or mutual information, but trail features carry most of what they use and clearly outperform generic graph features. A bound for general networks is left to future work.
[AI-126] Predictive Closure: A Set-Level Characterization of JEPAs Feature Selection
链接: https://arxiv.org/abs/2610.27364
作者: Mosh Levy,Aly Lidayan,Yoav Goldberg
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:In joint-embedding predictive architectures (JEPA), an encoder and a predictor are trained together to produce a representation that minimizes the error in predicting the representation of the next observation. Because the representations of both the current and the next observation are learned, it is unclear which features they will include. Earlier characterizations describe the features JEPA selects in terms of their individual predictability, raising the concern that unpredictable features, like a traffic light, will not be encoded into the representation. However, the prediction error is also affected by the relationships between features, e.g., the light’s color helps predict whether the traffic will stop. We propose the predictive closure characterization, which accounts for these relationships. Based on a reformulation of the JEPA training objective, we show that selection depends on properties of feature sets. This view agrees with earlier works when features evolve independently, but makes new predictions when they do not, for instance that a feature like the traffic light can still be retained. We test this characterization in controlled experiments, which show that unpredictable features can become better represented when they help predict other features, and that changing one feature’s predictive relationships can affect how well others are represented. We also demonstrate implications for steering feature selection: targeting one feature with an auxiliary objective, or supplying it to the predictor as side information, can strengthen the representation of another feature whose future values it helps predict, so steering can be indirect.
[AI-127] LiBRA: Logit Fingerprints for Byzantine-Robust Client Selection in Federated Learning
链接: https://arxiv.org/abs/2610.27357
作者: Saheed A. Tijani,Frank Jiang,Ming Liu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Byzantine-robust aggregation shields federated learning from malicious updates, but the rules that filter out suspect updates do so by comparing clients’ parameter vectors. That comparison requires access to every client’s update, which secure aggregation exists to withhold, costs O(n^2 d) in the parameter dimension d , and is defined only between clients that share an architecture. We propose LiBRA, which compares clients on what their models compute rather than on the parameters that compute them, using a fingerprint of C^2 values for C classes, the class-conditional logit response of each client’s model to pseudo-patterns optimised against it. We instantiate LiBRA as variants of Krum, Multi-Krum and Bulyan that change only what the dissimilarity score is computed on. On Fashion-MNIST and CIFAR-10 with three architectures and four backdoor and label-manipulation attacks under i.i.d. partitions, each variant holds every attack its parent holds, and on identical client updates under a backdoor attack the fingerprint ranks malicious clients within 0.006 AUC of the full parameter vector. The variants thus match the robustness of the rules they extend, while their trust decision reads no client parameter, compares up to 49,090 times fewer values, and can by construction compare clients whose architectures differ, at the cost of one pattern optimisation per client per round.
[AI-128] Exposing Vulnerability Monoculture Through Systemic Adversarial Robustness
链接: https://arxiv.org/abs/2610.27353
作者: Leo Elmecker-Plakolm,Lukasz Szpruch,Avishek Joey Bose,Matthew Wicker
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Beneath the apparent diversity of foundation models lies an increasingly homogeneous technical stack of common training pipelines, distillation methods, and data. These shared algorithmic choices can expose an otherwise distinct model ecosystem to concentrated risks leading to a vulnerability monoculture. Yet current practice in assessing adversarial robustness is restricted to individual robustness scores and attack transfer rates, leaving open the unexplored fundamental question: how many attacks suffice for catastrophic widespread compromise? In this paper, we investigate the empirical nature of shared vulnerabilities within a model pool and introduce Systemic Adversarial Robustness (SAR), a novel measure of shared vulnerability of a model pool through the attack diversity needed to exploit it. Specifically, SAR interpolates between single-attack transferability and independent per-model robustness: small values reveal a vulnerability monoculture, whereas large values indicate that failures are dispersed across models. We theoretically formulate SAR as a partial set-cover problem with provable approximation guarantees—enabling the development of scalable algorithms to empirically measure the systemic adversarial robustness. We empirically test the shared vulnerability of vision and large language models and characterize their shared collective vulnerability curve. Specifically, we demonstrate that model pools distilled from DeepSeek-R1 and Claude Opus 4.6 exhibit lower estimated systemic robustness than their corresponding parent pools, revealing a potential systemic cost of frontier-model distillation. Our framework seeks to establish vulnerability monoculture as a measurable target for a more wholistic adversarial evaluation of model populations.
[AI-129] Refuses When Asked Complies When Acting: A Runtime Safety Check for Robot Control Loops
链接: https://arxiv.org/abs/2610.27350
作者: Arnav Grover
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 4 pages plus references. Submitted to the Science of Physical AI Safety (SPAIS) workshop at CoRL 2026
Abstract:Frontier models can control robots through agent harnesses that loop on one question: what motion comes next toward the user’s goal? We show that this default loop carries out harmful instructions. With GPT-6 Astra controlling a low-cost SO-101 arm through the Inspect Robots harness on three household tasks whose hazard is common knowledge, such as “put the knife in the toy basket”, the policy refused 0% of 15 trials, attempted the harm every time, and completed it in 27%. Its one exit, give_up, is meant for goals that cannot be achieved, and it never cited the hazard. We add an optional runtime safety check to the loop: before the first motion, the model is asked whether it would carry out the instruction, given the same camera frames. The check would have stopped all 15 hazardous instructions, including the 4 that completed the harm, and flagged none of the 15 harmless controls. Code and data: this https URL.
[AI-130] An LLM -Based Agent for Legal Compliance Assessment: Architecture Memory and Evaluation in a Canadian and Quebec Context
链接: https://arxiv.org/abs/2610.27347
作者: Soukaina Alaoui Ismaili,Jean-Sebastien Dessureault
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Checking whether a text complies with a body of law is slow when done by hand and fragile when done with keyword rules. This paper presents ComplianceAgent, the legal compliance sub-agent of COMPASS, a multi-agent framework that evaluates texts along four axes: digital sovereignty, carbon impact, compliance, and ethics. The agent relies on a large language model that scores a text on a fixed rubric, grounds its judgment in a reference legal corpus, and, when the score falls at or below a critical threshold, runs a corrective chain made of self-healing rewrites, a compliance label, an explainability report, and escalation to a human reviewer. We describe what the language model does at each step, how the memory of the agent is organized, and how the same implementation can be pointed at different reference corpora. We evaluate the agent in a Canadian and Quebec context, on a corpus of thirteen legal instruments that brings together federal and provincial texts, international treaties, and municipal documents of the City of Trois-Rivières. The evaluation relies on a hand-annotated gold standard of fifteen texts and combines classification metrics, a repeated-query stability test, retrieval precision, and a composite score. On this gold standard the agent separates every conforming text from every non-conforming one, its scores remain stable under repeated queries, and the index returns relevant passages.
[AI-131] Modeling Authentic Research Trajectories: A Graph View of Research Ideation
链接: https://arxiv.org/abs/2610.27337
作者: Ruochen Li,Zhiyu Chen
类目: Artificial Intelligence (cs.AI)
备注: A research progress graph
Abstract:The research idea in a published paper is rarely the project’s initial idea. Between the two lies a closed-loop process in which researchers run experiments, derive insights, identify challenges, explore the literature, and revise the idea, with each step motivated by evidence and reasoning they can articulate. We call the path one idea takes through this process, with the evidence and reasoning behind each step, its authentic research trajectory. Although it underlies every scientific contribution, it is rarely recorded. Existing LLM-based research agents generate an idea once, refine it against reviews, literature, or their own experimental results, or search over candidate ideas and experiments. In all of them the text of the idea evolves, while the evidence and reasoning that connected one version to the next are discarded. We propose the Research Progress Graph (RPG), a representation of authentic and complete research trajectories. An RPG records the evolution of an idea as typed research states connected by cognitive transitions, each annotated with a temporal index and the reasoning behind it, which makes this evolution an explicit object of study. We ground the representation in authentic trajectories whose steps their authors justify in annotations and papers. Comparing the representations of thirty-five research agents with the RPG, we find that each is an incomplete RPG; none keeps the state that changed, the evidence that changed it, and the reasoning connecting them as one evolving record. Modeling ideation with the RPG changes how research agents should be designed, evaluated, and trained. Agents predict the next research state instead of a complete idea, ideas are evaluated as trajectories instead of final texts, and authentic trajectories serve as training data. These directions shift research agents from generating static ideas to modeling how ideas evolve.
[AI-132] Unsupervised Tool Evolution for Coding Agents in Data-Lake Analytics
链接: https://arxiv.org/abs/2610.27327
作者: Yuqing Yang,Yurong Liu,Deqing Fu,Manav Garg,Sravan Babu Bodapati,Fatma Ozcan,Alon Y. Halevy,Jiani Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Coding agents can answer analytical questions over a data lake by discovering, cleaning, and joining the relevant data. However, they discard this work after each question and redo it from scratch for the next one. Tool evolution, which automatically builds and refines reusable tools across tasks, is a natural fit for this repetition. Since a ground-truth label for a data-lake question requires an expert to solve it first, we consider the unsupervised setting. We propose TEAL (Tool Evolution for Analytics over data Lakes), which evolves a library of tools and usage skills for a given Worker agent without labels. Two agents propose tools from complementary evidence, an Explorer from the data lake itself and a Critic from Worker trajectories. A proposed change is kept only if the Worker adopts the tool on a rerun without extra steps and any changed answer is supported by trajectory evidence. We also introduce TEAL-Q, which generates synthetic questions for tool evolution when real ones are unavailable. We evaluate on KramaBench and LakeQA with four Worker LLMs under step budgets, which cap the number of steps the Worker may take per question. With a tight budget of 10 or 15 steps, TEAL improves every Worker on both benchmarks, lifting Sonnet-5 at 15 steps by 9.3 points on KramaBench and 21.7 points on LakeQA. A tighter budget also yields better feedback for evolution. For example, Sonnet-5’s LakeQA library evolved at 20 steps and deployed unchanged at 30 steps scores 68.3, above both the baseline (59.6) and a library evolved directly at 30 steps (64.2). Additionally, a library evolved on synthetic questions performs comparably to one evolved on real questions. These results show that useful tools can be evolved without labels, and hopefully inspire future work on what to evolve and how to obtain feedback in unsupervised agent evolution.
[AI-133] Evolving Self-tuning Benchmarks for Long-Horizon Tool Use
链接: https://arxiv.org/abs/2610.27307
作者: Hyunji Alex Nam,Michael Dennis,Natasha Jaques,Minsuk Chang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:For AI assistants to move beyond simple chat interfaces and reliably assist users with applications like Gmail, Venmo, and Slack, we need to improve their tool-use capabilities. However, existing tool use benchmarks are still limited and difficult to scale because they rely on human supervision or are too domain-specific. To make data generation scalable, we propose, TURTLE (Tool Use Related Tasks with Long-Horizon Evolution), an Unsupervised Environment Design (UED) approach to evolve existing tasks using a proposer, a solver, and a feedback generator large language model agents. The proposer is rewarded for difficult yet solvable tasks based on the solver’s empirical performance, and the feedback generator summarizes the solver’s execution traces for the proposer to exploit. TURTLE produces evolving, self-tuning test suites adaptable to different environment requirements, from supporting user simulators and dense document retrieval to interacting across multiple apps. Empirical results on AppWorld, \tau -Bench, and AutomationBench using Gemma 4 31B-it show that the evolved tasks target the desired difficulty range (e.g., a 165% increase in the percentage of hard tasks on AppWorld), increase task diversity (in terms of reasoning and simultaneous tool interactions), and elicit a broader range of failure modes than the original tasks.
[AI-134] When Does an Added Planning Capability Create Operational Value? An Observable Evidence Standard
链接: https://arxiv.org/abs/2610.27305
作者: Yihua Xu,Yan Lyu,Jize Li,Jiani He,You Lin,Yidan Hu,Rubing Qiang
类目: Artificial Intelligence (cs.AI)
备注: 11 pages, 4 figures, 6 tables. Accepted and presented at the 3rd IEOM World Congress on Industrial Engineering and Operations Management, Irvine, California, USA, October 13-15, 2026. Code and reproduction materials: this https URL
Abstract:Industrial artificial intelligence (AI) planning systems combine predictive models, optimization logic, controls, and governance rules, but improvement in the complete policy does not show whether one addition made an independent contribution or whether that contribution justifies its burden. We address this attribution problem with an observable evidence standard that separates opportunity to act, executed-plan change, operational consequence, equivalence, uncertainty, and operating scope. Using forecast-to-plan replay on Favorita, M5, and Walmart data, we apply the standard to heterogeneous planning additions. Three limits of single-indicator evaluation emerge. First, a composite policy can conceal negligible component contribution: one control changed only 15 of 681 executed plans. Second, changing execution does not guarantee stable benefit: the evaluated multi-period capability helped in several long-lead settings but was 7.6% more costly under tight capacity. Third, indirect evidence can prioritize review but cannot complete attribution. Independent runs produced different paths in 19 of 599 cases that a rule-based check predicted would be unchanged. Although a lower-cost diagnostic matched the full-evaluation winner in only 55.0% of cases, it selected an option within the 10^-6 cost-equivalence tolerance in 79.5%; the remaining cases prevented the proxy from authorizing acceptance. These experiments establish capability-level contribution evidence; the organization completes the value decision by weighing it against implementation and sustaining burden. The resulting capability evidence record links the observations to validation, deployment scope, fallback, monitoring, redesign, and retirement decisions.
[AI-135] Stronger artificial intelligence breakthroughs hasten their own exhaustion
链接: https://arxiv.org/abs/2610.27296
作者: Beichen Lu,Honglin Bao,Qi Huang,Qingguo Meng,James A. Evans
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); General Economics (econ.GN)
备注:
Abstract:Progress in artificial intelligence is widely assumed to scale with effort: more researchers, papers and investment should accelerate discovery. Theories of technological growth instead locate the difficulty of discovery in the space of ideas. Here we show that neither scale nor space is enough: breakthroughs depend on diverse attention, which collective concentration can deplete. Across 36 years of computer science, we analyze 7.2 million papers, 73,089 peer reviews and 3.7 million researchers and their topic shifts. The field expanded while its attention narrowed sharply after the Transformer. Over half of topic-shifting researchers moved into AI, especially mid-career ones, and reviewer confidence fell by a full point on a 5-point scale. Each successive breakthrough, from AlexNet to large language models, raised the field’s carrying capacity for ideas but was followed by steeper exhaustion of those ideas. Bigger breakthroughs capture attention more completely, depleting the diversity their successors may draw upon. Structural biology shows the same cycle after cryo-EM and AlphaFold. Simulations show how well-intentioned and widely used policies such as reciprocal reviewing empower prolific authors and further narrow what the field accepts. We call for institutions that protect diverse attention in artificial intelligence rather than reward concentration on the winning paradigm.
[AI-136] Vulnerability Analysis for Safe Multi-Agent Reinforcement
链接: https://arxiv.org/abs/2610.27293
作者: Taozheng Zhu,Shixiong Jiang,Fanxin Kong
类目: Artificial Intelligence (cs.AI)
备注: 21 pages, including references and appendices
Abstract:Safe Multi-Agent Reinforcement Learning (SMARL) aims to learn cooperative control policies under prescribed safety constraints. Al- though existing SMARL methods focus on enforcing safety during policy optimization, their vulnerability remains largely unexplored. This paper investigates inference-time observation attacks against SMARL, where an adversary perturbs the observations of a subset of agents, thereby altering the actions produced by their policies. We propose MinMax Attack, an attack framework that jointly selects adversarial target actions for compromised agents while accounting for the Cooperative Resistance of benign agents. We further derive a bound on the induced safety cost and characterize its relation- ship with the Compromised-Agent Budget and the Perturbation Budget. Experiments across multiple safety-critical multi-agent benchmarks and representative state-of-the-art SMARL algorithms show that MinMax Attack induces substantially higher safety costs than existing attack baselines while preserving task reward, thereby reducing its detectability and achieving both attack effectiveness and stealthiness.
[AI-137] VPFlow: Value-Predictive Planning with Imperfect World Models for Offline Goal Reaching
链接: https://arxiv.org/abs/2610.27280
作者: Yunzhe Hu,Dong Xu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 32 pages
Abstract:Offline goal-conditioned reinforcement learning enables skill acquisition by reaching any goal from any state without dense rewards. Model-based methods allow for planning by scaling test-time compute to solve complex tasks. Combining both holds appealing promise for pre-training on suboptimal demonstrations and zero-shot adaptation via goal specification and planning. Yet existing methods assume effective planning requires accurate next-state prediction in pixel or latent space. In this work, we show that value prediction is more useful for planning, whereas accurate state prediction is often unnecessary. Our key insight is that offline goal reaching naturally admits a closed-form Monte Carlo return, enabling value regression to replace temporal-difference bootstrapping. We use this return as the sole supervision for latent world models: imagined latents are trained to predict goal-reaching value rather than match future states, implicitly preventing representation collapse. These planning-oriented representations let imperfect world models imagine on the level sets of value field beyond dataset transitions. We then present VPFlow, a value-predictive planner that pairs the world models with a one-step flow policy and plans in its noise space via evolutionary search. It enables policy improvement with imagined generalized advantage and reduces out-of-distribution queries common in action-space planning. On 30 state and pixel OGBench datasets, VPFlow surpasses model-free baselines by at least 30.3 and 10.1 percentage points in success rate, respectively, and outperforms latent planners LeWorldModel, PLDM, and TD-MPC2. It achieves this despite a latent prediction error orders of magnitude larger than with state prediction, while minimizing this error lowers success rate. This demonstrates that value-predictive representations, not exact state prediction, are what offline planning requires.
[AI-138] A latent control model for realistic rodent motion NEURIPS2026
链接: https://arxiv.org/abs/2610.27279
作者: Aidan Sirbu,Charles Y. Zhang,Yuanjia Yang,Bence P. Ölveczky,Talmo D. Pereira,Blake A. Richards
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at NeurIPS 2026. 10 pages, 13 figures. Project page with supplementary videos: this https URL
Abstract:The development of embodied models of animal behavior is a major goal of the neuro-AI community. Embodied models will enhance our ability to make direct comparisons between biological and artificial agents. To make such comparisons effective, though, we need embodied models that actually move like animals. Previous work has developed virtual animal bodies and techniques for training realistic motion through imitation. However, to date, there is no system for embodied animal modeling that produces natural movements both during trained goal-directed behavior, and during exploration driven by random generation. Here, we introduce Simulated Control via Augmented Motor Priors for Embodied Reinforcement learning (SCAMPER), a system for training an embodied virtual rat model, enabling natural movements during both goal-directed behavior and random exploration. SCAMPER uses a variational architecture with two distinct priors, a motion prior for control and a behavior prior for exploration. We show that combining both of these priors allows SCAMPER to generate realistic behavior during exploration and exploit the task structure effectively. SCAMPER can also learn from sparse rewards as a result of the diverse movements induced by the behavior prior, allowing the system to model realistic sparse-reward tasks from experimental neuroscience. Altogether, SCAMPER provides the neuro-AI community with an effective new system for training embodied models of rodents on realistic tasks.
[AI-139] Certifying Repeatable Entry-Replay Witnesses for COBOL-to-Java Migration
链接: https://arxiv.org/abs/2610.27272
作者: Andras L. Ferenczi,Mariya Bessonov,Allwyn Ponnudurai,Chandrakanth R. Mamillapalli,Krishna K. Maddula,Krishna K. Lingamneni
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 8 pages, 9 figures, 4 tables
Abstract:Producing synthetic data for comparison testing a migrated COBOL application is a major challenge that agentic migrations face today. Our prior paper, the Agentic Method for Deterministic Validation of Legacy Code Migration, presents a deterministic oracle to facilitate such migrations, the Locksmith Loop. We now propose a novel method for achieving high branch and paragraph coverage through generated synthetic end-to-end test cases, by means of entry-replay test-case certification by observing the code execution trace on a surrogate COBOL program that mocks external interactions via a chronological external-operation tape. We applied our method to a very large COBOL program with 33,429 lines of code to generate 849 synthetic end-to-end test cases covering 2,973/3,806 source branches (78.1%) and 989/1,033 main-AST paragraphs (95.7%). Using the Locksmith Loop’s projected-agreement validation, the generated Java code matched the COBOL surrogate on the specified observations across all 849 certified synthetic end-to-end test cases. During the campaign we detected 61 defects and fixed the Java generator accordingly. To test whether the coding agent could silently introduce business logic bugs, we seeded 450 defects into the Java code, such as deleted writes, deleted paragraph calls, and altered rounding, to verify how many of these would be detected by the gating methodology. The gate detected 47.6%, including all deleted writes and paragraph calls in executed code, but was unable to detect the rounding and precision defects.
[AI-140] heorem prover Arend
链接: https://arxiv.org/abs/2610.27264
作者: Fedor Part,Valery Isaev,Sergey Sinchuk
类目: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI)
备注:
Abstract:This paper is an exposition of the theorem prover Arend developed at JetBrains Research. Arend language has been designed to facilitate the formalization of mathematics within the framework of univalent foundations and with the focus on constructive mathematics. It is based on a variant of homotopy type theory with cubical syntax that is more closely aligned with Martin-Löf Type Theory (MLTT) than fully computational cubical type theories, while still retaining some of their computational properties useful in practice. Among the most notable features of the Arend language is its powerful system of classes and records supporting subsumptive subtyping, partial implementations and anonymous extensions, which provides extensive flexibility for constructing hierarchies of definitions. Formalization in Arend is streamlined by the IntelliJ Arend plugin, which transforms IntelliJ IDEA into a full-fledged integrated development environment (IDE) for the Arend language. The main Arend library arend-lib includes formalizations in constructive algebra, topology and synthetic homotopy theory.
[AI-141] When Models Know but Cant Linearly Say: Relational Encoding Limits Linear Vulnerability Classification in Code LLM s EMNLP
链接: https://arxiv.org/abs/2610.27257
作者: Rui Melo,Andre Catarino,Claudia Mamede,Corina Pasareanu,Rui Abreu
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted at EMNLP Findings
Abstract:Large language models can detect vulnerable code, yet the form of vulnerability encoding in their representations remains unclear. We analyze 8 code models spanning 6.7B-15B parameters on C programs using three datasets (DELTASECOMMITS, SVEN, and PRECISE-BUGS), comparing two diagnostic tasks on the same representations: identifying what patterns the code instantiates, and determining whether it is vulnerable. Models encode the first with high fidelity. Under the studied paired C-code datasets and linear-probing setup, vulnerability information is more accessible as a paired relational signal-whether code A is riskier than code B-than as a globally separable categorical label. This structural difference has implications for classification: the model reliably ranks which code is riskier but standard classifiers on isolated samples achieve limited accuracy because vulnerable and secure samples form overlapping distributions in representation space.
[AI-142] AutoRedClaw: Automated Red-Teaming for Claw-Style Agents
链接: https://arxiv.org/abs/2610.27244
作者: Hao Li,Tongyan Hu,Weidi Luo,Shanghao Shi,Xiaogeng Liu,Aiden Suh,Yulong Cao,Ning Zhang,William Yeoh,Yevgeniy Vorobeychik,Chaowei Xiao
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 32 pages, 8 figures, 5 tables
Abstract:Agent systems are shifting from task-scoped tool-use agents to persistent, personalized assistants such as OpenClaw. These claw-style agents maintain configuration files (e.g., this http URL), long-term memory, and reusable skills. Once poisoned, this persistent state is repeatedly reloaded into the agent’s context, creating persistent threats across tasks and sessions. Existing security evaluation systems struggle to evaluate these complex claw-style systems since they rely heavily on manually constructed test cases and environments, which limits their scale and coverage; in addition, these manually designed cases quickly become outdated as agents evolve, and their evaluation settings leave a gap with how claw-style agents are deployed in practice. To address this, we introduce AutoRedClaw, an automated red-teaming framework that generates end-to-end attacks together with the environments needed to execute them. AutoRedClaw combines a strict quality-checking protocol with execution-based validation to ensure the correctness, realism, and effectiveness of generated samples, and adopts task-driven environment construction to keep costs low. To achieve broad coverage, it organizes generation along risk categories, attack strategies, attack stages, and attack surfaces, and uses experience-guided exploration to steer generation toward under-explored regions. AutoRedClaw achieves an effective attack rate above 30% at less than 0.10 per successful sample. From its outputs, we construct RedClaw-Bench, a 330-case benchmark that yields substantially higher attack success rates (ASRs) than AgentDojo, AgentDyn, and DTap. Across seven harnesses and four frontier models, most agent systems exceed 50% ASR, showing that our red-teaming pipeline provides a strong evaluation signal even for the most advanced agent systems. Our code and benchmark are released at this https URL.
[AI-143] EGI-BENCH: An Enterprise General Intelligence Benchmark Suite
链接: https://arxiv.org/abs/2610.27234
作者: Zhiwei Liu,Zhujun Lan,Jielin Qiu,Zixiang Chen,Liangwei Yang,Ming Zhu,Akshara Prabhakar,Wenting Zhao,John Emmons,Jason Wu,Shafiq Rayhan Joty,Junnan Li,Shelby Heinecke,Silvio Savarese,Huan Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Enterprise adoption of LLMs in agentic workflows is gated by evaluations that span a wide arc of abilities: CRM operations, heterogeneous-document retrieval, personal data understanding, multi-turn planning, tool use, function calling, long-horizon research, and long-context software engineering. These evaluations live today in separate repositories with divergent runners, dependencies, and scoring conventions, which makes cross-benchmark comparison difficult and obscures what a single aggregate number actually represents. We release EGI-BENCH, a unified LLM agent benchmark suite that consolidates eight enterprise facing agent benchmarks (CRMARENA, HERB, PERSONABENCH, USERBENCH, MCP-UNIVERSE, MCPEVAL, BENCHFORCE, LOCOBENCH) behind a single command-line interface with two runtime-budgeted execution modes. On top of the suite we define the Enterprise General Intelligence (EGI) Score, a six-dimensional framework (M1-M6) whose buckets are populated by fine-grained sub-metrics routed from each benchmark rather than by benchmark-level averages. We describe the suite, the sub-metric routing, and the quick/deep budgets, and report results for fourteen frontier and cost-efficient LLMs spanning the GPT, Claude, and Gemini families under quick mode. Top-line EGI ranges from 39.6 (GPT-4o mini) to 65.6 (gpt-5.5), and no single model dominates every capability axis: gpt-5.5 leads task outcome, planning, and retrieval grounding; Claude Sonnet 4.5 leads tool actioning and safety/abstention; and Claude Haiku 4.5 leads engineering quality. Code is available at this https URL.
[AI-144] Magnolia: A Persistent-Memory Agent Framework for Computational Chemistry
链接: https://arxiv.org/abs/2610.27230
作者: Tao Jiang,Benjamin Navet,Yousouf Taghzouti,Fabien Gandon,Louis-Félix Nothias,Élise Dumont
类目: Artificial Intelligence (cs.AI); Biomolecules (q-bio.BM)
备注: this https URL
Abstract:Chemistry has recently entered a high-data regime as molecular databases from high-throughput platforms or computational studies have become available. Increasing automation within the chemical sciences leads to a more digital and precision chemistry. Dedicated scientific agents are becoming central to guiding the chemist through an ever-increasing amount of data and findings, including negative ones. Here, we introduce Magnolia, a persistent-memory artificial intelligence agent framework that integrates computational protocols, from docking to quantum chemistry, lab notebooks and chat logs relying on large language models (LLMs). It relies on a three-tier memory system: session-level observations are automatically captured and distilled into structured learnings, confirmed and elevated to project-tier entries and, when human-validated, promoted to cross-project rules that actively guide subsequent work and provide direct chemical feedback. We illustrate Magnolia on four case studies: quantum-chemical evaluation of aromatic systems (COMPAS database), molecular dynamics of butane and of modified DNA, and a protein-peptide docking campaign. Across these projects, Magnolia tracked 317 runs and accumulated 116 project-tier entries, from which four cross-project rules were promoted; in controlled replays of four tasks, runs with memory passed the checks that runs without it failed (for example, ten of ten docking-configuration checks, against one to four of ten). By combining automation, reproducibility, and human-in-the-loop control, our Magnolia framework streamlines docking, molecular dynamics and quantum chemistry calculations with direct analysis and LLM-augmented feedback, offering accelerated access to computational (bio)chemistry.
[AI-145] Sparse RL Relaxation for LLM Reasoning
链接: https://arxiv.org/abs/2610.27220
作者: Jiabin Fan,Matthias Lindemann,Zita Marinho,André FT Martins,Lili Mou
类目: Artificial Intelligence (cs.AI)
备注: 19 pages
Abstract:Recent work has explored continuous representations in place of discrete tokens during Chain-of-Thought (CoT) reasoning, often referred to as soft CoT. This is motivated by the idea that a mixture of token embeddings can represent a superposition of multiple reasoning possibilities within a single step. However, it remains challenging to train large language models to realize this potential. Existing work has explored reinforcement learning (RL)-style training for soft CoT. However, these approaches treat soft tokens as actions in a Markov decision process rather than as differentiable intermediate computation steps. Thus, they do not propagate gradients across the soft tokens and miss out on the signal of how the choice of one soft token affects the next. In this paper, we propose a sparse relaxation method for RL (called SpaRLax), which leverages the differentiability of soft-CoT reasoning over time steps and makes training more stable and effective. We adopt an entmax sparsity method that adaptively determines the degree of softness, which truncates long-tail noise at confident steps while retaining multiple plausible candidates when uncertain. Across nine reasoning benchmarks, our method achieves the highest average performance compared with baseline methods. Further analyses validate the effective use of soft-token superposition in our approach, and also demonstrate the contribution of adaptive sparsity in our approach.
[AI-146] LSA-PAST: Layer-Specific Adaptation and Prototype Transport for Compositional Incremental Learning
链接: https://arxiv.org/abs/2610.27203
作者: Kyle Stein,Andrew Arash Mahyari,Guillermo Francia III,Eman El-Sheikh,Jiangpeng He
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Class-Incremental Learning (CIL) enables models to recognize new classes over time while retaining prior knowledge, but many real-world problems are compositional, where visual concepts arise from combinations of primitives such as states (i.e., red, shiny) and objects (i.e., shoe, car). Standard CIL methods treat each composition as an independent class, ignoring shared primitives and providing no mechanism to stabilize their representations as new compositions arrive. This causes parameter updates to entangle primitives and induce drift that degrades performance on earlier compositions. Compositional Incremental Learning (composition-IL) formalizes this setting by introducing new state-object compositions across tasks. In this paper, we propose Layer-Specific Adaptation with Prototype-Anchored Adapter-Space Transport (LSA-PAST), a parameter-efficient method tailored to composition-IL. LSA integrates Low-Rank Adaptation (LoRA) into a frozen ViT with depth-specialized adapters for states, objects, and compositions. PAST then maps past, anchored prototypes into the current adapter space through a rehearsal-free affine transform fit on current-task data, enabling old prototypes to remain compatible with new adapters. Extensive experiments demonstrate that LSA-PAST achieves state-of-the-art (SOA) accuracy-forgetting trade-off performance for composition-IL.
[AI-147] Visual Task Modeling and Automated Planning for Humanoid Robots in Education of Individuals with Autism Spectrum Disorder
链接: https://arxiv.org/abs/2610.27162
作者: Lada Kudláčková
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 6 pages, presented at conference AICCONF 2026 in Prague, Czech Republic
Abstract:In robot-assisted education, a robot must understand how to complete a task in order to effectively guide a student through it. We present a novel method that enables a semi-autonomous, camera-equipped robot to recognise the task goal and plan actions to achieve it. Our work focuses on Shoebox Tasks, an activity commonly used in the education of individuals with Autism Spectrum Disorder (ASD). These tasks are designed so that the goal can be inferred directly from their visual structure. We have developed a solver that uses a vision-based Large Language Model to convert photos of a Shoebox Task into a Planning Domain Definition Language (PDDL) representation and generate action plans for solving it. We have also implemented a robotic interface that enables a Nao robot to autonomously execute various Shoebox Tasks and assist a student by suggesting next steps or signalling detected failures through action demonstrations, voice commands, or gestures. With an adapted robotic interface, the solver can be integrated into a variety of educational robots. No programming expertise is required from teachers, and the solver operates exclusively on captured images as input. Research in special educational needs indicates that students with ASD can benefit substantially from robot-assisted teaching, and our system aims to strengthen these learning opportunities while also fostering interest in robotics among individuals with ASD.
[AI-148] Learned Control Variates for Rare Grid-Outage Estimation at Electric Vehicle Charging Stations
链接: https://arxiv.org/abs/2610.27159
作者: Sulyman Iyanda,Mansur Arief
类目: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Systems and Control (eess.SY); Machine Learning (stat.ML)
备注:
Abstract:Public charging of electric vehicles (EVs) adds a large and synchronized load to the distribution feeders. In most networks, it is rare to overload a feeder at the same time. When it does happen, the consequences can be severe: an overloaded feeder stresses its transformers and can trigger cascading outages across a city. Planners who design for high reliability therefore need to estimate very small outage probabilities. To that end, plain Monte Carlo estimation with an expensive grid simulator can require a prohibitively large number of runs for this task. To reduce the number of runs, we propose a control variate (CV) built from a learned classifier of the outage, trained on cheap observable features such as temperature and load forecasts. We show that for any binary learner, the variance reduction is an explicit function of its recall and false-positive rate. Furthermore, we show that control variate based on a Gaussian proxy of the load loses its effect as outages become rarer, while a learner with non-vanishing precision keeps removing a fixed fraction of the variance even in the extreme rare-event regime. We verify these results in numerical experiments using a feeder-overload simulator across three orders of magnitude of outage probability and obtain that even weak learner with a single feature can remove a quarter of the variance, while a boosted-tree learner is able to remove about 60%. We then apply the estimator to the feeder network of Surabaya, Indonesia, in which the Gaussian load model leaves every feeder three to seven times the one-day-in-ten-years blackout target. At the same simulation budget, planning the reserve with the learned CV and a 95% confidence limit reduces the city-wide blackout days by a third compared with Monte Carlo. These results show that learned control variates can help planners design distribution grids that stay reliable against rare outages.
[AI-149] Understanding the Benefits of Scaffolds through Computational Separations
链接: https://arxiv.org/abs/2610.27157
作者: Jingchu Gai,Adam Block,Andrej Risteski
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Inference-time harnesses, i.e., external procedures that call a model many times and keep state across calls, can substantially improve model capabilities. In this paper, we investigate the benefits of using such harnesses and the extent to which these benefits can be internalized through training. More precisely, we conceptualize a harness \mathcalA as an arbitrary wrapper around a base model M and attempt to distill a student model from the harness’s traces. Empirically we find that, even with simple scaffolds and small models, the inference-time harness significantly outperforms a distilled student. We then prove that such a gap is due to a fundamental representational limitation: inference-time harnesses can compute a closure of their per-call primitive that is \NL -complete even when the primitive is a trivial \AC^0 lookup, and that no fixed-depth forward pass can represent under the widely believed conjecture that \NL\not\subseteq\TC^0 . Our theory suggests that expanding the representational capacity of the student relative to the task at hand, e.g., by enlarging the model, making the task easier, or lengthening its chain-of-thought, is necessary to close the gap, which we verify empirically. Finally, we scale to an agentic harness on competition math, where the same gap reappears. Overall, our results show that a harness confers on the base model a profound gain in representational power, one that training a same-class student on its traces cannot transfer back into the weights.
[AI-150] MoSAIK: Motion Search with Artificial Intelligence in K-space
链接: https://arxiv.org/abs/2610.27156
作者: Oscar Dabrowski,Abijith Jagannath,Frédéric Grouiller,Sébastien Courvoisier
类目: Artificial Intelligence (cs.AI); Medical Physics (physics.med-ph)
备注:
Abstract:Object: Motion artefacts remain a major limitation of routine two-dimensional T2-weighted turbo spin-echo (TSE) brain MRI, reducing diagnostic confidence and often necessitating repeated acquisitions. MoSAIK (Motion Search with AI in K-space) retrospectively estimates in-plane rigid-body motion directly from standard multi-coil TSE k-space. Without a motion-free reference, external navigators, or additional acquisitions, MoSAIK infers motion from k-space self-consistency violations and incorporates predicted parameters into model-based, motion-aware reconstruction rather than synthesizing corrected images. Materials and Methods: Selected complex coil data were represented as log-magnitude and phase channels. A decoupled convolutional embedding and Transformer encoder estimated phase-encoding-line-indexed rotation and two-dimensional translations. Training used motion-free clinical acquisitions with simulated motion; four held-out subjects were evaluated. Component-wise errors assessed estimation; SSIM and PSNR in silico and entropy focus criterion (EFC) in vivo assessed reconstruction. Results: Motion estimation errors were low for both rotation (MSE 0.2 deg ^2 ) and translation (MSE 0.2 pixels ^2 ), with Pearson correlations of approximately 0.96 and 0.94, respectively. In every in-vivo acquisition, more than 75% of slices showed information gain relative to uncorrected reconstructions. MoSAIK reduced visible artefacts while preserving anatomical detail. Discussion: Transformer-based k-space motion-trajectory estimation enables fast retrospective correction of standard clinical TSE brain MRI. Subjects: Artificial Intelligence (cs.AI); Medical Physics (physics.med-ph) Cite as: arXiv:2610.27156 [cs.AI] (or arXiv:2610.27156v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2610.27156 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Oscar Dabrowski Ph.D. [view email] [v1] Mon, 19 Oct 2026 18:49:59 UTC (39,619 KB) Full-text links: Access Paper: View a PDF of the paper titled MoSAIK: Motion Search with Artificial Intelligence in K-space, by Oscar Dabrowski and 3 other authorsView PDFHTML (experimental)TeX Source view license Additional Features Audio Summary Current browse context: cs.AI prev | next new | recent | 2026-10 Change to browse by: cs physics physics.med-ph 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) Major funding support from (opens in new tab) (opens in new tab) (opens in new tab)
[AI-151] Model Exposure on One Device Endangers Others: Rethinking Edge Neural Network Security across Devices
链接: https://arxiv.org/abs/2610.27152
作者: Yuyang Zhang,Kangjie Chen,Yihui Jin,Run Wang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Neural networks (NNs) are increasingly deployed on edge devices such as mobile and IoT platforms, offering low latency and improved privacy compared to cloud-based deployment. However, on-device models embedded in mobile applications can be unpacked and exposed to adversaries, enabling cross-node attacks in centrally distributed ecosystems, where adversarial examples crafted on one compromised device transfer effectively to other nodes sharing the same models. To address this threat, we propose VeilNN, a defense framework that reduces adversarial transferability across on-device models via model obfuscation. The core idea of VeilNN is to introduce controlled model diversity while preserving original task functionality. Specifically, VeilNN applies stealthy and irreversible obfuscation at both the architectural and parameter levels through network cell obfuscation and layer assembling, and is applicable to diverse NN architectures, including MLPs, CNNs, and Vision Transformers. Extensive experiments demonstrate that VeilNN effectively defends against cross-node attacks, achieving an average defense success of 49.19% and up to 81.19% across four adversarial attack types, with more than 33.16% improvement over existing defenses.
[AI-152] Representing short-term movement to reveal animal identity and learning progress
链接: https://arxiv.org/abs/2610.27140
作者: Ping Zhang,Quang-Huy Nguyen,Yingtie Lei,Sameera Shridhar,Zachary Leveroni,Ellie Quattrocchi,Jai Y Yu,Wei-Lun Chao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC)
备注: The first three authors contribute equally to this work
Abstract:How learning progresses over time and varies across individuals is a longstanding question in cognitive and behavioral science. However, a major challenge is identifying observable behavioral measures that faithfully capture learning progression. Existing approaches commonly rely on predefined behavioral features or task-level performance, while fine-scale posture and movement in freely behaving animals remain less explored. In this paper, we ask: can short-term movement trajectories reveal learning progress and individual identity? We introduce a self-supervised representation learning framework built on a simple hypothesis: behavioral characteristics associated with identity and learning remain relatively stable over nearby time intervals. Using temporal proximity as supervision, our framework learns compact representations directly from movement trajectories without manually defining behavioral states or kinematic features. Using behavior data from freely behaving rats, we show that the learned representations not only improve prediction of identity and learning-related readouts over existing methods, but also provide a tool for investigating broader scientific questions about when and how these signals appear in behavior. Specifically, trajectories as short as one second contain informative learning and identity signals. Egocentric pose dynamics are more predictive of identity, whereas allocentric motion is more predictive of learning-related variation, and both signals are strongest at decision junctions in the alternation cohort. We further use a representation-conditioned diffusion model to inspect movement patterns associated with identity and learning progress. Our findings establish short-term movement as an informative behavioral “fingerprint” for studying learning and individual differences. Our code is available at this https URL.
[AI-153] Ontic state recovery through prediction
链接: https://arxiv.org/abs/2610.27139
作者: Ping Zhang,Quang-Huy Nguyen,Yingtie Lei,Wei-Lun Chao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC)
备注:
Abstract:We encounter a system through its responses, yet its state concerns how the system is. Predicting further responses promises knowledge beyond what is immediately observed. What that knowledge tells us about the system’s state depends on what we mean by a state in the first place. We give this notion a model-relative meaning: a state is the complete response profile over a specified query universe, and configurations agreeing on every query are regarded as the same state. From this perspective, a learned state is understood through the distinctions between these profiles that it retains. Prediction concerns selected observations; recovery concerns the profile as a whole. We examine their relationship through query ambiguity, state-space geometry, and the choice of prediction targets. Synthetic systems with known response laws and neural recordings provide complementary settings for this investigation. The aim is to clarify how prediction can become a way of knowing a system, and what it would mean for a learned state to preserve the state being studied.
[AI-154] Mind Your Boundaries: Boundary-Aware Model Obfuscation against Cross-node Attack
链接: https://arxiv.org/abs/2610.27135
作者: Yuyang Zhang,Yihui Jin,Kangjie Chen,Leyang Wu,Run Wang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:With the advancement of powerful AI chips, on-device machine learning is gaining traction for its low latency and privacy protection. However, when edge devices are compromised, attackers can obtain white-box access to the deployed models by reversing the apps directly to obtain the knowledge of peer devices, which is known as Cross-Node Attacks, like performing efficient adversarial attacks. Most existing defenses target known vulnerabilities, yet white-box access still enables attackers to uncover more unknown weaknesses. This creates a critical risk for edge-device deployments, where one compromised device can endanger all models distributed from the same source. Motivated by findings that the model robustness is primarily influenced by the curvature of the decision boundary, we investigate how model parameters shape its curvature. In this paper, we propose a novel method, BMO, that obfuscates model parameters to diversify the decision boundary further resisting cross-node attack while preserving its benign accuracy. Specifically, we explore decomposing model parameters into boundary-position and boundary-curvature components based on null-space constraints, then models with differentiated decision boundaries are instantiated by injecting noise into the boundary-curvature component. Across all four major types of adversarial attacks with four widely-deployed on-device models, BMO achieves an average success rate of 68.68% in defense, far exceeding the most competitive baseline, which attains only 22.32% in defending against cross-node attacks. Moreover, BMO achieves comparable performance in benign accuracy compared with baselines.
[AI-155] Reconstruction-Optimized Expert Pruning for Mixture-of-Experts Language Models
链接: https://arxiv.org/abs/2610.27123
作者: Jelena Markovic-Voronov,Petros Prastakos,Kayhan Behdin,Jincheng Cao,Zhipeng Wang,Shayan Mohajer Hamidi,Yuanda Xu,Zhengze Zhou,Wenhui Zhu,Rahul Mazumder
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Mixture-of-Experts (MoE) models reduce computation through sparse expert activation, but deployment still requires storing a large number of expert parameters. We study post-training expert pruning using task-domain calibration data as a subset-selection problem: given a pretrained MoE layer, select a fixed number of experts that best reconstruct its original output. Exact subset search is combinatorial and becomes intractable for modern MoE models with many experts per layer. We develop OptEP, a scalable continuous relaxation that directly optimizes differentiable expert-selection variables, followed by a lightweight reconstruction-based update of the retained expert weights. Across multiple MoE architectures and tasks spanning mathematical reasoning and code generation, our approach improves over the strongest existing expert-pruning baselines, with absolute accuracy gains of up to 15.05% on Qwen3-30B-A3B, 14.55% on Qwen3-235B-A22B, 11.06% on GLM-4.5-Air, and 16.20% on Mixtral-8x7B. The improvements are particularly pronounced under aggressive pruning, demonstrating that reconstruction-driven expert selection and weight adaptation can preserve substantially more of the original model’s performance while reducing the number of retained experts.
[AI-156] EDMA: Entropy-Driven Multimodal Answering NEURIPS2026
链接: https://arxiv.org/abs/2610.27120
作者: Emanuele Mezzi,Gertjan Burghouts,Fabio Massacci,Mengyuan Zhang
类目: Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026
Abstract:Multimodal question answering (MMQA) is based on integrating heterogeneous data sources, selectively leveraging relevant modalities, and ignoring distractors. Existing approaches based on Multimodal Large Language Models (MLLMs) improve performance by splitting the task into intermediate steps, but do not quantify how each modality contributes to reaching the final answer. We argue that quantifying the effect that each modality has on the answering process improves process traceability and final performance. Thus, we propose Entropy-Driven Multimodal Answering (EDMA), which uses logical entropy to formalise multimodal QA as an iterative process of entropy reduction and to quantify the contribution of each modality in distinguishing correct from incorrect candidate answers. This enables the construction of quantifiable answering trajectories, where each step is associated with a measurable entropy reduction. EDMA introduces (i) modality-conditioned partitioning to estimate unimodal entropy reduction, (ii) entropy-driven multimodal fusion to capture complementary information across modalities, and (iii) entropy-based selection of the answering trajectory that minimises the most logical entropy. Evaluated across standard MMQA benchmarks, EDMA outperforms competitive prompting baselines, with average gains of 12.8 F1 and 4.6 LLM-J points over Planning Prompting and 18.6 F1 and 8.3 LLM-J points over Chain-of-Thought. Moreover, EDMA outperforms self-consistency baselines and is competitive with the multi-agent system baseline. Moreover, we show that higher entropy correlates with more false positives and, through controlled trajectory selection, demonstrate that logical entropy serves as an effective, monotonic selector for correct reasoning paths, validating entropy level as a reliable proxy for performance.
[AI-157] PORA: Learning When More Reasoning Pays Off for LLM Agents
链接: https://arxiv.org/abs/2610.27119
作者: Bo Yan,Qi Zhang,Zhen Tan,Tianlong Chen,Song Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LLM agents perform tasks over multiple steps and revise their actions based on feedback. As the total reasoning budget is often limited and must support future steps, the agent must choose between short and long reasoning before each action. Longer reasoning may improve the final task outcome by preventing later errors, but it also uses budget capacity that may be needed for future decisions. This creates two challenges: estimating the downstream value of long reasoning and deciding when that value is worth using the remaining budget. We propose Paired Outcome-guided Reasoning Allocation (PORA), which learns this decision by comparing short and long reasoning from the same interaction state. During training, \method completes the task twice from each selected state, once with a short reasoning allowance and once with a long allowance, and compares their final quality and total processing to estimate when longer reasoning is valuable. During inference, a controller weighs this estimate against learned prices on remaining-budget pressure and chooses either short or long reasoning without running both alternatives. We evaluate PORA across application, memory, code, retail, and mathematics tasks, achieving strong results across all settings. On BFCL memory, it exceeds the strongest existing method by 17.4–32.9 percentage points. Within the same agent, it matches or exceeds always using long reasoning on code and mathematics with 50.6% fewer actor tokens and 31.2% fewer model requests, respectively.
[AI-158] PROVISO: Provenance-Informed Action Selection for Secure LLM Agents
链接: https://arxiv.org/abs/2610.27115
作者: Bo Yan,Qi Zhang,Ziwen Kan,Fnu Suya,Song Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LLM agents act on the external world through tools and third-party skills, so a single executed action can cause irreversible harm. Existing defenses judge a proposed action without tracking whether its content comes from the user or from external sources such as tool outputs, and they judge it in isolation rather than against an available alternative, which forces a trade-off between security and utility. We present PROVISO, a controller that decides which action to commit after the agent proposes an action and before it is executed. PROVISO encodes trusted observations, external observations, the proposed action, and an alternative into provenance-typed memory, relates them through role-specific queries, and commits the action with the highest calibrated value learned from outcomes. On four agent security tasks spanning the agent lifecycle, no baseline exceeds PROVISO in both security and utility. PROVISO improves F1 over the strongest baseline by 13.84 points on SkillsBench and by 7.07 points on ToolSafe, and it reaches 88-99 F1 on three additional skill detection benchmarks.
[AI-159] A general-purpose System One decision model for clinical records
链接: https://arxiv.org/abs/2610.27099
作者: Andrew Zhang,Trinh T.L. Vuong,Sophia J. Wagner,Dongmin Bang,Tong Ding,Yuchang Su,Mary Cipperman,Christina S. Chen,Judith Lefkes,Ramin Kahidi,Ming Y. Lu,Caiwei Tian,Tianhao Luo,Jingnan Ma,Rowland Pettit,Long Phi Le,Faisal Mahmood
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Clinical care, health-system operations and biomedical discovery depend on extracting reliable answers from patient records, yet applying diverse clinical questions across entire populations remains computationally expensive. Here we present Clinical System One (CS1), a 4-billion-parameter decision model for fast, typed judgments from clinical records, along with two agentic harnesses in which CS1 works together with a generalist language model. CS1-4B is trained on 12.4 million questions and answers from 3.6 million patients across a multi-hospital health system, combining structured-data keys, observed outcomes and language-model supervision across clinical notes, diagnostic reports and patient timelines. It answers multiple typed questions in a single forward pass without task-specific retraining, and each answer carries a probability that can route uncertain cases to human review. On 963,408 held-out patients, CS1-4B outperformed open decision models up to 27 billion parameters on 13 of 14 tasks scored independently of its training teacher. Across 21 external data sets, its aggregate performance approached that of the strongest tested 27-billion-parameter model, with 6.4-fold faster inference in a matched workload. Integrating CS1-4B with a language model agent produced cited answers over complete multi-year patient histories and screened a patient population against the eligibility criteria of 17 registered clinical trials, placing four times as many judged-eligible patients among its top 20 candidates as keyword search while using 690-fold fewer language-model tokens than a language model reading every record. These findings support a division of labour in clinical AI, with small decision models performing repeated record reading and generative models formulating questions and synthesizing evidence.
[AI-160] Physical AI: Concepts Techniques and Challenges
链接: https://arxiv.org/abs/2610.27045
作者: Weinan Zhang,Chunhua Shen,Tianfan Xue,Bin Zhao,Bingqi Jiang,Zihan Guo,Zicai Cui,Yang Li,Ming Zhou,Xinyuan Chen,Lihe Ding,Xiao Chen,Yujin Wang,Shi Guo,Junting Dong,Tao Lu,Xing Shen,Haoyu Guo,Yao Mu,Haiyong Yu,Haoxiang Ma,Fuxian Huang,Qiaojun Yu,Yunsong Zhou,Kailin Li,Zirui Wang,Yufei Xue,Weishuai Zeng,Tao Huang,Zhigang Wang,Junli Liu,Pengyuan Wu,Pingrui Zhang,Haotian Wang,Dongzhan Zhou,Jiafei Cao,Feng Yuan,Xinyi Ye,Quanli Xuan,Mulin Yu,Shangding Gu,Kaiwen Song,Dong Wang,Jiarui Li,Kehui Liu,Zhongjie Jia,Dongjiang Li,Jiacheng Bao,Haoran Yang,Zhaoqing Chen,Yiming Sun,Bowen Zhou
类目: Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: Monograph, 12 chapters, 388 pages, 64 figures
Abstract:In recent years, physical artificial intelligence (physical AI) has experienced rapid development, capturing significant attention in academia and industry. However, the field has yet to reach consensus on its definition, technical landscape, and core challenges. To address this gap, this monograph seeks to define physical AI, present its technical landscape, and examine each key technical subfield in depth. In doing so, it establishes a unified framework for Physical AI: artificial systems that perceive, model, and act upon the physical world through continuous feedback under real-world constraints. Further, this monograph distinguishes physical AI from conventional digital AI, embodied AI, cyber-physical systems, and digital twins. It develops the foundations of physical AI across physics, control, cognition, learning, and reasoning. It reviews key methodologies, including multimodal perception, physical data and simulation, world models, embodied foundation models, motor-control systems, multi-agent coordination, and hardware-software co-design. Furthermore, it examines applications spanning robotics, manufacturing, infrastructure, scientific experimentation, and autonomous services. Particular attention is given to sim-to-real transfer, reliability, real-time execution, evaluation, safety, governance, and long-term deployment. By connecting perception, prediction, decision-making, action, and adaptation in a closed loop, this monograph identifies the technical challenges and research directions required to advance from narrow automation toward more generalizable physical AI and, ultimately, physical artificial general intelligence (Physical AGI).
[AI-161] M-Plan: Comprehensive Planning For LLM Agents with Memory Arrangement
链接: https://arxiv.org/abs/2610.27032
作者: Ziyang Liu,Jiakang Yuan,Jiayuan Fan,Tao Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM)-based agents have recently demonstrated outstanding performance in single-turn tasks, yet they continue to struggle with long-horizon tasks, particularly in environments without prior training. A key limitation of single LLMs is their difficulty in processing long contexts, which stems from weaknesses in continuous decision-making and reasoning capabilities. Although existing multi-agent systems (MAS) equipped with long-term memory bases can partially address this gap, they mostly rely on knowledge reuse or summarization, an approach that is often constrained by directly retrieval without any adaptation for the current task, while the retrieved memory may contain much history noises waiting to be preprocessed. In contrast, MAS characterized by task-specific guidance often lack long-term memory specifically designed for the guidance. To address these challenges, we propose M-Plan, a closed-loop MAS framework that integrates long-term and short-term memory. By combining global-local and past-current guidance, M-Plan enhances agents’ performance on multi-turn tasks while maintaining a plug-and-play design. During task execution, agents receive task-specific milestones and insights, and are provided with local hints only when confusion arises, which form global-local guidance together. These guidance generation and utilization construct the short-term memory as well. After task completion, relevant information about the current task is optimized or reorganized before being sent to the long-term memory storage for future retrieval. The designments mentioned above build a comprehensive framework different from other individual long-term or short-term memory works. Extensive experiments on multi-turn tasks including ALFWorld, Webshop, and PDDL across diverse models demonstrate that M-Plan outperforms multiple baseline methods in both accuracy and efficiency.
[AI-162] Does Multi-Agent Collaboration Reduce Hallucination in Clinical Tasks?
链接: https://arxiv.org/abs/2610.27027
作者: Won Seok Jang,Harry Bhaskar Joshi,Hong Yu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-agent LLM systems improve performance on clinical tasks, and that improvement is often read as evidence of a more trustworthy system. However, the two are not the same claim. A clinician acts on the stated reasoning and not only on the conclusion, so whether collaboration reduces hallucination, and what decides when it does, has to be measured in the reasoning too. Yet most evaluations score only the final answer. We evaluate four systems (a single-LLM baseline and three multi-agent topologies) across four models and three clinical tasks, scoring hallucination, under a factuality and faithfulness taxonomy, inside the rationale each system generates and reporting it alongside the task metric. We report that collaboration buys task performance but not lower hallucination as a rule, and no topology is reliably better than another. The effect differs by model: of the 36 model–system–task comparisons, 12 are large enough to resolve — four decreases and eight increases, with no model on both sides. What sets the level, and the direction of any change, is the model rather than the topology or the number of agents. A self-correction experiment is consistent with this: shown its own flagged errors, every model except Llama-3.2-3B-Instruct reduces hallucination; the one model for which collaboration lowers hallucination gains the most, and Llama-3.2-3B-Instruct, the model whose hallucination it amplifies most, does not reduce its hallucination at all. We find that collaboration redistributes what its models can already do; it does not supply what they cannot.
[AI-163] BR-MoE: Budgeted Semantic Release for Privacy-Preserving Traffic Analytics
链接: https://arxiv.org/abs/2610.27026
作者: Abdolazim Rezaei
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Traffic cameras support useful analysis, but their outputs can reveal private information. Replacing images with text reduces direct visual exposure. It does not place a formal bound on disclosure. The published reinforcement learning and mixture-of-experts framework refines scene descriptions through relevance, coverage, and brevity rewards. It does not account for repeated releases under a fixed privacy budget. This paper proposes budgeted semantic release as an extension of that framework. Four scene experts first process isolated video chunks. Their outputs are mapped to a fixed, bounded semantic schema. A reinforcement learning controller selects which expert summaries to release and how much privacy budget to spend. The controller uses public task requests and previous private releases. A separate privacy filter enforces the total budget. Calibrated noise is added before the summaries are converted to text. Under explicit chunk isolation and bounded-event assumptions, the complete release transcript satisfies an event-level differential privacy bound. The design also supports reuse of previous summaries without new disclosure. An evaluation protocol covers traffic utility, privacy attacks, release freshness, and computational cost. Illustrative tables show the intended reporting format. They do not establish empirical gains. The proposed contribution is a task-aware release interface with auditable privacy accounting for repeated traffic descriptions.
[AI-164] Reliable Estimation of Differential Privacy under Selection Bias
链接: https://arxiv.org/abs/2610.27021
作者: Valentin Dorseuil(DI-ENS),Meenatchi Sundaram Muthu Selva Annamalai(UCL),Jamal Atif(CMAP),Olivier Cappé(DI-ENS),Nicolas Papernot(Inria)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
备注:
Abstract:Statistical privacy audits are essential for verifying the theoretical privacy guarantees of differentially private (DP) mechanisms. Auditors typically use Membership Inference Attacks (MIAs) to estimate a lower bound on the true privacy leakage of DP mechanisms at a specified confidence level, which should not exceed the theoretical privacy upper bound guaranteed by DP. However, we found that many auditing implementations in prior work introduce a selection bias when running MIAs, which leads to statistically invalid and overestimated lower bounds. Because most audits are not tight, this bias has thus far gone unnoticed: the estimated lower bounds remained well below the theoretical upper bounds. Nevertheless, accurately evaluating this lower bound is critical, as a flawed statistical assessment could lead to false alarms on correct mechanisms or an incorrect evaluation of the guarantee’s tightness. This work employs tools from the statistical literature to correct this bias. We present an overview of the different correction methods and assess their respective advantages. Furthermore, we perform a comprehensive re-analysis of results returned by prior audit implementations and show that without statistical correction, these audits violate the theoretical upper bounds up to four times more often than their claimed confidence level implies. With correction applied, the resulting lower bounds are statistically valid, and violations occur at no more than the nominal rate.
[AI-165] Exploration via Activation-Space Steering for Reinforcement Learning with Verifiable Rewards
链接: https://arxiv.org/abs/2610.27012
作者: Yunan Zeng,Jinhao Jing,Xi Cheng,Yuwang Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Reinforcement learning with verifiable rewards (RLVR) trains a language model by sampling responses, scoring them with an automatic verifier, and updating the policy from the resulting rewards. The policy therefore receives direct learning signals only from trajectories that appear in its rollouts. Yet repeated sampling from the same policy can concentrate on similar reasoning paths, leaving useful alternatives unexplored under a fixed rollout budget. We introduce Activation-Space Exploration (ASE), an online method for diversifying training rollouts. By clustering activation differences between correct and incorrect responses to the same problem, ASE builds a structured bank of latent exploration vectors. Different vectors are then applied to subsequent rollouts to explore alternative reasoning paths. On Qwen3-1.7B-Base and Qwen3-8B-Base trained on DAPO-Math-17k, ASE improves average Mean@8 across six mathematical reasoning benchmarks over GRPO by 4.87 and 1.87 percentage points, respectively, under the same rollout budget.
[AI-166] HierDTA: Hierarchical Structure-Aware Representation Learning for Drug–Target Affinity Prediction
链接: https://arxiv.org/abs/2610.27001
作者: Huangchao Xu,Baohua Zhang,Qian Liu,Zhong Jin,Yi Du
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 20 pages, 7 figures
Abstract:Drug–target affinity (DTA) prediction plays an important role in computer-aided drug discovery, where reliable computational models can help prioritize compound–protein pairs before experimental validation. A key challenge in DTA prediction is to effectively mine multi-level molecular information from drug structures while accurately modeling the binding-site environment of target proteins. To address these issues, we introduce HierDTA, a hierarchical structure-aware framework for DTA prediction. \textcolorblackHierDTA represents each drug as an atom–motif heterogeneous graph and introduces an augmentation module that propagates functional-motif semantics to the corresponding atoms. For proteins, kinase-domain structures are converted into residue interaction graphs. The resulting atom and residue representations are integrated using cross-attention to capture pair-specific relationships, while molecule fingerprints and pretrained ESM-2 embeddings provide complementary global information. Compared with state-of-the-art methods, HierDTA achieves superior performance on the Davis dataset across all metrics and attains the highest (r_m^2) of 0.791 on KIBA. It also exhibits strong generalization under cold-start settings, particularly on KIBA, where MSE is reduced by 21.7% in the unseen-protein scenario. Ablation study and interpretability analysis supported the contributions of the hierarchical augmentation and interaction components. Case studies further showed that HierDTA provided early enrichment of candidate bioactive compounds. These results indicate that HierDTA is an effective framework for affinity prediction and target-specific virtual screening.
[AI-167] SPEM: Structure-Aware Procedural Memory for Efficient Skill Use in LLM Agents
链接: https://arxiv.org/abs/2610.26995
作者: Ran Zhou,Zehui Kong,Fanyu Lin,Jiaheng Liu,Zhen Zhang,Yuxiang Ren
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LLM-based agents increasingly rely on reusable skills. However, scaling skill libraries makes efficient and accurate skill use difficult. Full skill loading incurs high token and runtime overhead. In contrast, flat semantic retrieval treats skills as independent text entries, ignoring dependencies, execution order, state transitions, and data flow constraints. We propose SPEM, a lightweight structure-aware procedural memory framework for efficient skill use in LLM-based agents. SPEM maintains Atomic Skill Memory for individual skills and Skill-DAG Memory for reusable compositions derived from successful executions. For a new task, SPEM reuses high-confidence Skill-DAG templates when available. Otherwise, it retrieves atomic skills and invokes Skill-GNN for state- and structure-aware skill selection. Successful executions are written back as new Skill-DAGs, allowing procedural memory to accumulate over time. Experiments on ALFWorld and SkillsBench show that SPEM achieves the best overall effectiveness–efficiency trade-off, reaching 87.1% ALFWorld success and 38.38% SkillsBench reward while reducing token and runtime costs against strong baselines.
[AI-168] CoASFL: Coupled Consistency Optimization for Asynchronous Split Federated Learning
链接: https://arxiv.org/abs/2610.26992
作者: Junhao Cheng,Yunming Liao,Hongli Xu,Yang Xu,Zuan Xie
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Split Federated Learning (SFL) has emerged as a promising paradigm for training large-scale deep models on resource-constrained edge devices. Although asynchronous SFL alleviates the straggler bottleneck caused by system heterogeneity, it inevitably introduces two coupled dimensions of inconsistency: horizontal inconsistency across clients and vertical inconsistency between clients and the server. These inconsistencies, further exacerbated by statistical heterogeneity, result in biased activation representations, model divergence, and severe client drift. To address these challenges, we propose CoASFL, an inconsistency-aware asynchronous SFL framework that incorporates a dynamic activation merging mechanism and a consistency-aware model aggregation mechanism. Specifically, CoASFL constructs more consistent activation sequences from heterogeneous clients to alleviate horizontal inconsistency, and dynamically adjusts aggregation weights according to client participation frequencies and progress gaps to mitigate vertical inconsistency. By jointly coordinating activation-level and model-level consistency, CoASFL effectively suppresses inconsistency propagation during asynchronous training. Extensive experiments on an 80-device heterogeneous edge testbed demonstrate that CoASFL consistently outperforms state-of-the-art SFL approaches, accelerating training by 1.61 \times to 3.74 \times and improving final model accuracy by up to 3.26% under heterogeneous scenarios.
[AI-169] 3D-MoE: Complementarity-Guided Hardware-Software Co-Design for MoE Decode on 3D-Stacked NMP
链接: https://arxiv.org/abs/2610.26991
作者: Yudong Pan,Lian Liu,Shixin Zhao,Yintao He,Kaiyan Chang,Tianhua Han,Chenyang Ai,Haonan Zhu,Mengdi Wang,Yinhe Han,Ying Wang
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
备注:
Abstract:Mixture-of-Experts (MoE) have become a mainstream way to scale Large Language Models (LLMs), because each token activates only a small subset of experts, reducing inference cost. However, MoE decode remains highly memory-bound and exhibits strongly skewed, time-varying expert loads. 3D-stacked DRAM near-memory processing (NMP) provides the bandwidth that decode demands, yet its tiled organization makes load balancing expensive: expert parallelism (EP) preserves locality but suffers from hotspot-induced imbalance, while more distributed strategies weaken locality and increase NoC traffic. To address this locality-balance conflict, we observe that domain-specialized token routing creates offsetting expert-load fluctuations across decode steps, a property we call expert complementarity. Using this insight, we propose 3D-MoE, a hardware–software co-designed system for MoE decode on 3D-stacked NMP. 3D-MoE uses offline, one-time complementarity-guided grouping and locality-aware mapping to establish coarse-grained inter-group balance, handles residual intra-group hotspots through a lightweight Collaborative Memory-Network (CMN) that streams each hot expert’s weight tiles once from its host DRAM channel to helper cores over a reconfigured local broadcast path, and uses cost-aware online remapping to remain robust to workload drift. Across MoE models, 3D-MoE achieves an average 4.63 \times MoE-layer speedup over EP and improves end-to-end throughput by 1.58 \times --1.86 \times over the strongest NMP baselines, with 2.11% additional logic-die area. Subjects: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2610.26991 [cs.AR] (or arXiv:2610.26991v1 [cs.AR] for this version) https://doi.org/10.48550/arXiv.2610.26991 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-170] When Planning Problems Become Hard for LLM s: Separating Structural Complexity from Semantic Interpretability
链接: https://arxiv.org/abs/2610.26964
作者: Mayukh Debnath,Adeyinka Adedigba,Rammohan Mallipeddi
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 1 figure, 17 supplementary pages
Abstract:Large language models (LLMs) struggle with classical planning, but benchmarks rarely distinguish structural difficulty from the difficulty of interpreting symbolic vocabulary. We examine seven STRIPS domains with BFS-optimal plan lengths. For predicting structural difficulty without solving an instance, the zero-training delete-relaxation heuristic h_\mathrmadd outperforms learned alternatives on held-out domains (mean Spearman |\rho|=0.931 ), including ridge regression ( 0.707 ), object count ( 0.657 ), and the episodic retrieval architecture ARC ( 0.400 ). ARC also fails to outperform a linear model using the same features. To isolate semantic effects, we progressively rename predicates while preserving states, transitions, and optimal plans. Valid LLM planning success drops from 67.5% to 40.0% between 0% and 100% renaming, while refusals remain at 0% . An LLM supplied with ARC’s conditioning information predicts difficulty in context better than trained ARC (mean |\rho|=0.826 ). Hybrid routing between an LLM and a symbolic planner depends more on calibration to the target LLM’s outcomes than on the structural score selected. Thus, established symbolic heuristics already estimate structural difficulty well; improving hybrid planning requires modeling LLM behavior sensitive to predicate semantics rather than developing more structural predictors.
[AI-171] Beyond Policy Discrepancy: Capability-Aware Multi-Teacher On-Policy Distillation
链接: https://arxiv.org/abs/2610.26950
作者: Xuanfeng Yang,Le Gui,Jiaming Wang,Xuan Zhang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-teacher on-policy distillation (MOPD) integrates domain specialists into a shared student, but determining how strongly each teacher should supervise remains challenging. Existing discrepancy-based allocation can assign strong supervision even when the student already matches or exceeds the teacher’s task performance, since behavioral mismatch does not necessarily imply a remaining capability deficit. We propose Capability-Aware Multi-Teacher On-Policy Distillation (CA-MOPD), which combines specialist merging with capability-guided allocation to account for what the student has already learned. CA-MOPD first merges reinforcement-learning specialists into a strong student initialization, then uses verifier-estimated teacher–student capability gaps to adjust domain-level distillation strength. Supervision decreases as capability gaps close and increases when estimated student capability regresses, while dense token-level guidance remains provided by the teachers. Across six benchmarks spanning mathematics, code generation, and instruction following, CA-MOPD achieves an average score of 32.22, outperforming naive MOPD, Open-MOPD, and the merged initialization by 0.86, 0.83, and 0.74 points, respectively. Controlled comparisons further support the benefits of merged initialization and capability-gap guidance under matched training budgets.
[AI-172] LightCRS: Lightweight Multi-Domain Conversational Recommendation System
链接: https://arxiv.org/abs/2610.26948
作者: He Ma,Lan You,Qiyang Wu,Chen Liu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Conversational recommender systems have recently adopted vision-language models to match user dialogue intents with multimodal product information. However, extending these methods to many product domains faces a dilemma: deploying a separately adapted VLM per domain scales cost linearly with the number of domains and requires routing each dialogue to the right model, while naively merging all domains into a single VLM degrades accuracy because product characteristics vary widely across categories. We propose LightCRS, a lightweight, query-conditioned framework for unified multi-domain multimodal conversational recommendation. LightCRS precomputes multimodal product representations once offline with frozen encoders. Online, a compact 5.3M-parameter module fuses this evidence for each dialogue via query-conditioned cross-attention, whose parameters are shared across all domains, and compares candidates jointly with an inter-candidate Transformer. To enable rigorous evaluation, we construct MM-R2A-25, a 25-domain benchmark pairing natural Reddit recommendation-seeking conversations with Amazon product listings. LightCRS achieves state-of-the-art average HR@1, NDCG@5, and MRR, improving HR@1 by 35.4% over the strongest baseline while reducing per-query latency by 266 times. Joint multi-domain training itself improves HR@1 by 15.2% over separately trained single-domain models.
[AI-173] MAMath-Bench: Process-Aware and Robust Evaluation of Multimodal Multi-Agent Mathematical Reasoning
链接: https://arxiv.org/abs/2610.26943
作者: Xinping Wang,Jie Ma,Rui Xing,Wensong Guo,Xiayi Zhang,Jianhao Deng,Zihao Zhou,Boqian Yang,Feijuan He
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Mathematical reasoning is a fundamental capability for educational applications of Large Language Models (LLMs), where reliable problem solving requires not only correct answers but also coherent and verifiable reasoning processes. Multi-agent systems (MASs) have shown growing potential in this domain by leveraging collaboration and verification among multiple LLM agents. However, existing evaluations primarily focus on the correctness of final answers in text-only settings, offering limited insight into multimodal reasoning capabilities, the reliability of intermediate reasoning steps, and robustness against reasoning errors. To address these limitations, we introduce MAMath-Bench, a process-aware benchmark designed to evaluate the multimodal mathematical reasoning capabilities of MASs. MAMath-Bench comprises 1,003 problems spanning 11 mathematical categories, multiple difficulty levels, and several modalities, including text, images, and audio. It features fine-grained milestone annotations to evaluate systems across three dimensions: problem-solving correctness, reasoning-process reliability, and robustness to adversarial reasoning-chain perturbations. Our evaluation yields three key findings. First, multi-agent collaboration does not consistently improve final-answer accuracy over strong single LLMs. Second, collaboration can substantially enhance the coverage of intermediate reasoning, but such gains may come at the cost of increased reasoning redundancy. Third, current systems remain vulnerable to adversarial reasoning errors, particularly those involving subtle structural perturbations. The dataset is available at this https URL.
[AI-174] Fix Not Just Resample: Bounded Compiler-Guided Repair in Lean 4
链接: https://arxiv.org/abs/2610.26942
作者: Haotian Deng,Chris Farber,Sonal Narasimha,Samskrithi Sivakumar,Yexiang Xue
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automated theorem provers often rely on sampling language models, but proof assistants like Lean reject each failed step with a precise error message — information that pure sampling discards. To overcome this, we introduce a bounded, online step-level repair module. Our search samples K candidate tactics at each step and repairs a failed tactic at most R times. Instead of discarding failed tactics, we feed the compiler’s error message to a pretrained repair model to fix individual rejected steps in place. This approach improves the solve rate over a sampling-only baseline across all tested candidate budgets ( K \in \1,3,5,8\ ), achieving up to a 32% relative gain and solving 82 additional theorems on a 1,990 -theorem subset of the LeanDojo benchmark. Repair is not free: it trades extra runtime for reachable proofs. But the trade is predictable: repair pays most when sampling is constrained; its return is front-loaded, with the first repair attempt capturing 84\text–96% of the total gain; and the gain comes almost entirely from shallow tactic-execution failures rather than structural errors such as type mismatches. This targeted repair successfully replicates on a second benchmark (Lean Workbook), nearly doubling the solve rate with zero regressions.
[AI-175] Industrial Agent ic AI: LLM -Driven New-Generation Agents for Smart Manufacturing
链接: https://arxiv.org/abs/2610.26941
作者: Lei Ren,Haiteng Wang,Shixiang Li,Jing Zhang,Yuqing Wang,Jingwei Guo,Yunfei Zhu,Yuanjun Laili,Yang Tang,Lihui Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are enabling agents to move beyond predefined workflows toward autonomous planning and execution, creating new opportunities for smart manufacturing. However, dynamic environments, complex physical constraints, and stringent reliability requirements hinder the direct deployment of general-purpose agents in smart manufacturing. This article presents a systematic overview of LLM-driven industrial agents. First, we analyze five key challenges and discuss the concept and core characteristics of industrial agents. We then establish a unified taxonomy covering application scenarios, embodiment forms, and five capability levels. Next, we propose a four-layer system architecture and detail its core modules and operational mechanisms for closed-loop task execution. Finally, we review representative applications across the industrial lifecycle and outline future research directions. This article provides a structured reference for the design, development, and application of industrial agents.
[AI-176] How Intervention History Changes AI Agent Sabotage on Subsequent Tasks
链接: https://arxiv.org/abs/2610.26935
作者: Daniel Saeedi,Amirali Aghazadeh
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:AI control evaluations often reset agents’ memory between tasks, although deployed agents may retain records of prior interventions. We ask whether learning that an earlier attack was caught changes an agent’s ability to evade defer-to-trusted (DtT), a protocol that replaces suspicious work with a trusted model’s solution. Herein, we study cross-episode adaptation: whether intervention history changes attack behavior on a subsequent task with different objectives and a fresh environment. We evaluate this using ControlArena’s Infrastructure-as-Code (IaC) and SHADE-Arena as testbeds, where agents must complete a main task while covertly pursuing a harmful side objective. For each successor task, we pair a history-conditioned attempt with the model’s no-history attempt. The retained history either states that the previous submission was replaced or additionally includes the trusted model’s replacement. For Claude Opus 4.8 on IaC, intervention history reduces successful sabotage. Completion of both the main and side objectives falls from 77% to 12% when the agent is told about the replacement and to 38% when it is also shown the trusted model’s replacement, while main task success remains essentially unchanged. Condensing the history weakens this effect, suggesting that the retained context matters and limiting attribution to intervention feedback alone. GPT-5.4 shows no significant change in successful sabotage on the same IaC benchmark. On SHADE-Arena, GPT-5’s successful sabotage rate decreases, but the reduction is not significant. Across the tested models, history conditions, and deferral thresholds, intervention history never increases the rate of successful sabotage that survives DtT. These results show that intervention history can change subsequent attack behavior, but its effect depends on both the model and the retained context.
[AI-177] DataPraxis: Scaling Analytics Tasks via Verified Workspace Evolution
链接: https://arxiv.org/abs/2610.26924
作者: Nan Jia,Wenxuan Shi,Jingqing Ruan,Zhizhao Zeng,Tianyi Zhai,Zhongda Su,Feng Hong,Zhengyu Chen,Wenhao Yan,Haojin Yang,Siqi Dai,Weipeng Zhang,Ke Zeng,Xunliang Cai
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:General-purpose agents are increasingly expected to perform professional analytics work, which requires sustained reasoning, iterative investigation, and reliable evidence. Training these capabilities requires diverse and challenging tasks. However, generating tasks directly from raw data often overlooks deeper questions that emerge through analysis and provides limited evidence for verification. To address these limitations, we introduce DataPraxis, an analytics task synthesis framework based on verified workspace evolution. The framework progressively enriches workspaces with validated findings and artifacts, then inherits and combines these workspaces to uncover deeper questions. It constructs tasks from accumulated findings and their dependencies, verifying analytical results both during evolution and after task assembly. Using this framework, we construct DataPraxis-5K, a dataset of 5,000 verified trajectories spanning diverse domains and output formats, to train DataPraxis-27B and DataPraxis-35B-A3B. DataPraxis-27B achieves 55.11 on DABStep and 51.9 on JobBench, matching or surpassing frontier models such as GLM-5.2. DataPraxis-5K is available at this http URL
[AI-178] Steerable Pluralistic Policy Alignment from Anonymized Crowd Feedback
链接: https://arxiv.org/abs/2610.26914
作者: Akansha Kalra,Ellen R. Novoseller,Vernon J. Lawhern,Daniel S. Brown
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:Pluralistic alignment requires AI systems to learn from diverse populations whose preferences may reflect heterogeneous or even conflicting objectives, and to personalize their behavior to individual users. Existing approaches in pluralistic alignment often assume access to annotator identities and rely on powerful pretrained representations to model individual preference variations. We propose Preference-PCA Pluralistic Learning (P3L), a novel framework for pluralistic alignment that jointly learns trajectory embeddings and the shared basis of preference directions via PCA decomposition from heterogeneous crowd feedback without requiring annotator identities. At test time, our approach enables rapid and accurate personalized reward learning and downstream policy optimization. Experiments across multiple embodied control tasks demonstrate that P3L enables steerable personal alignment that generalizes beyond the offline dataset to out-of-distribution new users and attains competitive RL and test performance, while substantially reducing computational cost and supporting variable adaptation query budgets.
[AI-179] MedPAO-Assist: Edge-Deployable Modality-Aware Agent ic AI for Radiology Report Structuring MICCAI
链接: https://arxiv.org/abs/2610.26913
作者: Shrish Shrinath Vaidya,Minmini Selvam,Ganapathy Krishnamurthi
类目: Artificial Intelligence (cs.AI)
备注: The paper has been accepted at MICCAI Workshop Agentic AI for Medicine, 2026
Abstract:The adoption of Large Language Models for radiology report structuring has been largely dominated by either proprietary cloud-based models or large open models requiring high-end GPUs for deployment, which is a critical bottleneck for clinical adoption in radiology workflows that demand data privacy and offline operation. To address this, we present MedPAO-Assist, an edge-deployable agentic framework that generates structured radiology reports directly from radiologist dictation and voice commands, requiring only \approx3.3 GB peak RAM and achieving an average end-to-end latency of \approx31 seconds on a standard laptop CPU through an asynchronous, multi-threaded execution pipeline. The system currently supports four major imaging modalities – Chest X-ray, Chest CT, Abdomen CT, and Brain MRI. To our knowledge, it is among the earliest agentic systems to run its entire pipeline with a Small Language Model (SLM) in the 0.6B-parameter range. This multi-agent architecture, built around an Extract-Ground-Generate paradigm, performs concept extraction, SNOMED-CT grounding, and structured report generation, followed by a speech-driven human-in-the-loop correction phase, together forming a fully “no-touch” reporting workflow that requires no keyboard input, external API calls, or internet connectivity. We evaluate each pipeline stage individually – concept extraction, ontology grounding, and concept categorization and report hardware profiling results demonstrating the system’s feasibility for real-time, CPU-only clinical deployment.
[AI-180] Pareto-Optimal RTL Code Generation via Multi-Objective Reinforcement Learning with Large Language Models
链接: https://arxiv.org/abs/2610.26904
作者: Koki Takeshita,Kenji Uebayashi,Hikaru Tomonari,Masaya Tsunokake,Koichiro Yawata,Tatsuya Sasaki,Yusuke Ohtsubo,Kota Dohi,Yasuhiro Sogawa
类目: Artificial Intelligence (cs.AI)
备注: Accepted at COLM 2026
Abstract:Recent studies have proposed reinforcement learning with verifiable rewards (RLVR) methods that enable LLMs to generate high-quality RTL circuits from natural-language specifications. Existing approaches optimize circuit PPA (Power, Performance, Area) by aggregating multiple objectives into a single scalar reward, making it difficult to simultaneously generate diverse Pareto-optimal solutions across different trade-offs such as low power, low latency, and small area. However, in practical design, the relative importance of each objective varies depending on product requirements and design constraints; thus, obtaining a diverse and high-quality candidate set is beneficial in that designers can choose a preferred trade-off a posteriori. In this work, we are the first to propose a learning method for LLM-based circuit generation that optimizes, not the quality of a single circuit, but the quality and diversity of the entire circuit set obtained through multiple generations in the PPA objective space. The proposed method consists of (1) a generation-based training pipeline that reformulates coverage improvement of the entire Pareto solution set into a sequential circuit generation problem, and (2) a reward function based on Hypervolume Contribution (HVC) that measures how much a generated circuit contributes to the Pareto front. Evaluation on the RTLLM v2 benchmark shows that the proposed method achieves +13.3 percentage points in relative Hypervolume over a baseline without PPA optimization and +14.4 percentage points over a single-objective reinforcement learning method, while also demonstrating superiority in circuit diversity.
[AI-181] Large language model and reinforcement learning based autonomous construction and progress updates
链接: https://arxiv.org/abs/2610.26899
作者: Yifei Xiao,Huasheng Lin,T.Y. Yang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Due to the rapid growth of the global population, traditional construction methods are gradually unable to meet the current housing demand. The advancement of machine learning algorithms has encouraged researchers to explore their applications in the construction industry, leading to more intelligent and efficient building processes. In particular, the rise of large language models in recent years has provided more possibilities for the transformation from traditional construction to smart construction. In this study, a large language model and reinforcement learning based autonomous construction method is proposed. An enhanced soft actor-critic reinforcement learning algorithm is developed to plan actions for a mobile crane according to construction tasks. Robot kinematics and control algorithms are used to ensure that the actions generated by the reinforcement learning algorithm can be executed by the crane. In addition, a customized embedding model is developed to extract information from the construction environment according to user interests. Then, a large language model, ChatGPT 5, is used to update construction progress, provide safety information to control system, and generate construction progress reports. The feasibility of proposed method is verified through several full-scall retaining wall construction experiments.
[AI-182] Navalia: Advancing Cowork Agents through Verifiable Task Synthesis and Long-Horizon Post-Training
链接: https://arxiv.org/abs/2610.26887
作者: Hengqi Liu,Ziyuan Liu,Yang Qin,Wanting Zhou,Zichuan Wang,Jiachen Liang,Xu Chu,Shaowei Chen,Zhaokai Luo,Mu Chuan
类目: Artificial Intelligence (cs.AI)
备注: 13 pages, 2 figures
Abstract:Large language model (LLM) agents are evolving from conversational assistants into Cowork Agents capable of autonomous planning, tool use, and complex task execution. Financial analysis requires these agents to integrate cross-document evidence and perform multistep numerical reasoning, but expert task authoring is costly, while direct LLM synthesis struggles to produce challenging tasks with reliable answers. We introduce Navalia, a framework for Cowork Agents that combines verifiable task synthesis with long-horizon post-training. Using corporate financial reports as its data source, Navalia semantically selects supporting evidence from complete fact sets and constructs executable computation chains to generate analytical tasks with traceable evidence and actual data dependencies between steps. We further use independent LLM review and programmatic checks to verify evidence consistency and answer correctness. The verified tasks are then used for trajectory collection, cold-start supervised fine-tuning (SFT), and reinforcement learning (RL) to improve long-horizon task execution. Navalia, initialized from Qwen3.6-35B-A3B, achieves consistent gains over the base model across all five benchmarks: OfficeQA Pro, FinSearchComp, Finance Agent Benchmark, APEX-Agents-AA, and Workspace-Bench-Lite. These results demonstrate that Navalia improves evidence integration, tool use, and long-horizon task execution capabilities and generalizes to broader Cowork settings, including general office tasks.
[AI-183] An AI Security Agent for University ACMIS: Multi-Vector Threat Detection and Automated Response
链接: https://arxiv.org/abs/2606.08270
作者: Joseph Walusimbi,Joshua Benjamin Ssentongo
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Emerging Technologies (cs.ET); Machine Learning (cs.LG)
备注: 6 pages, 1 figure, 5 tables,
Abstract:University Academic Management Information Systems (ACMIS) are high-value targets for a wide spectrum of security threats including brute-force login attacks, payment fraud, privilege escalation, insider data theft, and academic integrity violations. Traditional rule-based intrusion detection systems are inadequate because many malicious activities are structurally indistinguishable from normal operations. This paper presents an AI-based security agent for ACMIS that combines supervised anomaly detection, behavioural analytics, and a natural language processing chatbot for secure password recovery. The agent monitors five operational layers: authentication, authorisation, financial transactions, user behaviour, and system health, and responds through a four-tier risk escalation framework. A modular architecture allows the core engine to be extended to other institutional systems. Experiments on a simulated ACMIS event log dataset of 147,922 sessions demonstrate a threat detection macro-average F1 of 0.966, compared to 0.156 for a rule-based baseline and 0.836 for a sequence-only (LSTM) baseline, with end-to-end critical-tier automated response latency under 1 ms on a single-node prototype. The integrated recovery chatbot achieves 97.1 percent identity verification accuracy and an 87.3 percent mass-reset attack detection rate with zero false positives on legitimate high volume recovery periods.
[AI-184] Spatially Informed and Direction Sensitive Representation Learning for Spatial Transcriptomics NEURIPS2026
链接: https://arxiv.org/abs/2610.28579
作者: Azim Dehghani Amirabad,Ali Khodabandeh Yalabadi,Artem Moskalev,Yun-Ching Chen,Mehdi Pirooznia,Scott Oloff,Rui Liao
类目: Quantitative Methods (q-bio.QM); Artificial Intelligence (cs.AI)
备注: This manuscript has been accepted at the NeurIPS 2026 Workshop on AI for Drug Discovery: Bridging the Translation Gap (AI4DD)
Abstract:Cellular states are shaped by both intrinsic gene expression programs and the spatial microenvironment in which cells reside. Spatial transcriptomics makes this context measurable, but effective representation learning requires capturing both spatially informative molecular signals and intercellular organization. Existing approaches address these challenges only partially: conventional gene selection can overlook spatially structured signals, while methods for encoding cell-cell relationships often either reduce a cell’s microenvironment to proximity alone, discarding direction, or rely on absolute coordinates tied to an arbitrary tissue frame. Consequently, cells with similar neighbor distances but distinct spatial organization may receive overly similar representations, despite experiencing different biological contexts that shape their cellular states. We introduce Tessera, a spatial representation learning framework that addresses both molecular selection and cellular context. Tessera combines spatially informed gene selection with a geometric inductive bias that contextualizes each cell within its local tissue environment. By modeling relative cell-cell relationships, Tessera is translation-invariant while preserving direction, allowing neighboring cells to contribute differently according to their spatial context and cellular state. Across held-out tissues and technologies, Tessera outperforms existing spatial foundation models in representation quality and generalizes zero-shot to contexts up to 16\times larger than those seen during training.
[AI-185] Quantifying the Threat: A Bayesian Probabilistic Framework for Heavy Metal Health Risks among Children in the Densu Basin Ghana STOC
链接: https://arxiv.org/abs/2610.28064
作者: T. Ansah-Narh,G. Y. Afrifa,J. B. Tandoh,E. N. N. Nortey,D. Adjei,R. Opoku-Sarkordie,R. Davis
类目: Applications (stat.AP); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Geophysics (physics.geo-ph); Machine Learning (stat.ML)
备注: 59 pages, 7 figures, manuscript under peer review at Stochastic Environmental Research and Risk Assessment (SERRA)
Abstract:Groundwater in Ghana’s Densu River Basin remains a key domestic water source, but multi-metal contamination threatens children’s health. Previous deterministic hazard index assessments provide limited treatment of uncertainty, co-exposure, and spatial variability. This study applies a Bayesian probabilistic framework for non-carcinogenic risk assessment using 94 groundwater samples and measurements of Fe, Mn, Ni, Pb, Cd, and As. A Bayesian regression model propagates parameter uncertainty and accounts for contaminant co-occurrence. Adaptive Hamiltonian Monte Carlo is used for posterior inference, while inverse distance weighting generates spatially continuous risk surfaces. Pb, As, and Cd dominate cumulative risk. Posterior predictive results indicate that approximately 31% of sampling sites have a probability greater than 0.5 of exceeding a hazard index (HI) of 1. The risk surface is highly heterogeneous, with a posterior mean of 0.372 and a standard deviation of 0.416. High-risk zones often reflect moderate co-exposure rather than extreme contamination by a single metal. The framework provides uncertainty-aware, spatially explicit information to support groundwater monitoring, risk mitigation, and public health planning in data-limited regions.
[AI-186] Beyond Scalar Leaderboards: Adaptive Sampling for Pareto Frontier Identification in Multi-Objective LLM Arenas
链接: https://arxiv.org/abs/2610.27447
作者: Liang Twist Shan,Jiahao Ai,Zhimei Ren
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Applications (stat.AP); Other Statistics (stat.OT)
备注: 51 pages, including references and appendices
Abstract:The evaluation of large language models (LLMs) is inherently multi-objective, as different models may offer different tradeoffs across performance dimensions such as math, code, and writing. Because such evaluations are often based on costly and time-consuming human feedback, it is important to draw reliable conclusions using as few comparisons as possible. In this paper, we present Vector-Borda Empirical-Gap Elimination (VB-EGE), a fixed-confidence adaptive sampling algorithm for efficiently identifying models on the Pareto frontier in arena-style evaluation, where each comparison evaluates a pair of models on a prompt with respect to a selected performance dimension. VB-EGE progressively eliminates models from an active set once sufficient statistical evidence has been accumulated to certify their Pareto status. Under a multi-objective Bradley-Terry model, we show that VB-EGE returns the correct Pareto frontier with high probability and derive an instance-dependent upper bound on its sample complexity. We further establish a worst-case lower bound showing that VB-EGE is minimax-optimal up to logarithmic factors. Simulations and experiments on three real-world Arena datasets (Arena, SearchArena, and VisionArena) demonstrate VB-EGE’s empirical effectiveness.
[AI-187] Dynamic Dual-Feedback Conformal Inference for Time Series Forecasting NEURIPS2026
链接: https://arxiv.org/abs/2610.26987
作者: Songlin Du,Ling Luo,Uwe Aickelin
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted by NeurIPS 2026
Abstract:Quantifying uncertainty in time series forecasting is particularly demanding because sequential data exhibit temporal dependence and are prone to distributional changes. Conformal inference has emerged as a powerful uncertainty quantification approach through the construction of prediction sets. Recent advances have introduced online conformal methods that adaptively adjust prediction thresholds through feedback mechanisms. However, the existing feedback mechanism typically relies solely on miscoverage indicators (actual feedback), whether the true label falls within the interval at each time step, while overlooking the empirical prediction threshold (estimated feedback) that is derived from the oracle conformal method. In this paper, we propose Dynamic Dual-feedback Conformal Inference (DDCI), which incorporates a dual-feedback mechanism consisting of actual feedback and estimated feedback. The former drives the primary adjustment of the intervals based on true observations, while the latter dampens excessive expansions or contractions by leveraging empirical thresholds from conformal inference during updates. By balancing these two signals, DDCI achieves more stable and narrower prediction intervals in sequential settings while maintaining near-nominal coverage in practice.
[AI-188] FoundPDE: A generative foundation model for data-driven symbolic discovery of partial differential equations
链接: https://arxiv.org/abs/2610.26963
作者: Langchen Liu,Junlin Luo,Min Zhu,Zhuoran Yang,Lu Lu
类目: Computational Physics (physics.comp-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 42 pages, 13 figures, 12 tables
Abstract:Discovering governing partial differential equations (PDEs) from observed data is a central inverse problem in computational science and engineering: the goal is not only to predict trajectories, but to identify interpretable symbolic laws that can be analyzed and reused. Existing PDE discovery methods often depend on hand-designed libraries, accurate numerical derivatives, or per-equation symbolic search, making them fragile under sparse observations, noisy data, broad equation families, and unseen symbolic compositions. Here, we introduce FoundPDE, a pretrained data-to-symbol foundation model for PDE discovery. FoundPDE maps discretized spatio-temporal trajectories to symbolic equations with a transformer, uses Fourier and first-order spatial-derivative augmentation to expose complementary trajectory information, refines numerical coefficients, and supports downstream extension with improved generalization through few-shot fine-tuning and hierarchical discovery. On a clean in-domain benchmark spanning 20 PDE families, our model reaches 97.50% type recognition. Under coefficient shifts beyond the training support, type recognition remains 90.53%; under multiplicative Gaussian noise on the model input, it is 94.43% at a 2% noise level but falls to 52.09% at 10%. For unseen composite equations, few-shot fine-tuning reaches over 90% exact recovery on four held-out families, and we further develop a hierarchical discovery algorithm that recovers two-term composites without fine-tuning FoundPDE. By amortizing symbolic discovery across equation families, FoundPDE shifts PDE discovery from per-dataset search toward a reusable pretrained model that extends to new physics with limited additional data, a step toward foundation models that uncover interpretable governing laws directly from scientific data.
[AI-189] Multilingual Brain Decoding Reveals Shared Cortical Semantic Geometry
链接: https://arxiv.org/abs/2610.26927
作者: Yi Guo,Yong Hu,Michael Kwok-Po Ng,Shuqiang Wang
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI)
备注:
Abstract:Human languages differ markedly in lexical and syntactic form yet can convey similar semantic content. The neural basis of this cross-language semantic commonality remains unclear. Here we developed a multilingual brain decoding framework and analysed brain recordings from 227 participants spanning Chinese, English and French. We found that cortical semantic representations exhibited consistent semantic geometry relationships across languages and individuals, revealing a shared cortical semantic geometry. Importantly, the cross-language consistency of cortical semantic geometry increased from sensorimotor to higher-order association cortex, forming a shared semantic geometry gradient closely aligned with the principal cortical functional gradient. Our work establishes generalizable multilingual brain decoding and identifies stable cortical semantic geometry as a neural basis for semantic commonality across languages.
[AI-190] Enhancing analytical photon transport in turbid media through physics-guided neural network corrections
链接: https://arxiv.org/abs/2610.26917
作者: Guojun Tang,Paul D. Teal,Alexander Doronin
类目: Optics (physics.optics); Artificial Intelligence (cs.AI)
备注: 6 pages, 4 figures
Abstract:Analytical methods provide rapid and computationally efficient predictions of light transport in turbid media, making them attractive for real-time and inverse applications. However, their accuracy is limited by the approximations and assumptions used to obtain tractable solutions. On the other hand, Monte Carlo (MC) simulation is widely regarded as the “gold standard” for modeling light transport in turbid media, but its high computational cost limits its use in applications requiring repeated forward evaluations. Here, we introduce a physics-guided residual learning framework that combines the computational efficiency of analytical modeling with the accuracy of MC simulation. As a representative case study, we consider the diffusion approximation (DA), which can exhibit systematic errors under relatively strong absorption, weak scattering, and imperfect boundary treatment. Given the absorption coefficient, scattering coefficient, and anisotropy factor, the corresponding DA prediction is incorporated as a physical prior, while the network learns only the residual between the analytical prediction and the MC reference. Across the evaluated optical-property range, the proposed framework reduces scalar-reflectance RMSE and spatially resolved reflectance RMSLE by factors of approximately 62 and 14, respectively, relative to the diffusion approximation. With only 1% of the available training data, it achieves approximately 87% and 93% lower corresponding errors than a direct neural-network surrogate. These results demonstrate that learning corrections to an analytical physical baseline improves both prediction accuracy and data efficiency.
[AI-191] Aptamers as a sandbox for lab-in-the-loop agent ic molecular discovery
链接: https://arxiv.org/abs/2610.24806
作者: Zachary Cutts,John Coller,Rachel Chen,Hua Tang,Mathias Bellaiche,Ryder Easterlin,Dina Zverinski,Petar Sirkovic,Avinatan Hassidim,Yossi Matias,Pooja Rao,Annalisa Pawlosky
类目: Biomolecules (q-bio.BM); Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM)
备注: Equal leadership (co-senior authors): Pooja Rao and Annalisa Pawlosky. 34 pages, 4 main figures, 1 main table, 3 supplementary figures, 4 supplementary tables
Abstract:Despite breakthroughs in AI-driven molecular generation, existing discovery pipelines operate largely in silico, with wet-lab feedback reserved for validation rather than iterative improvement. Whether large language model (LLM)-based reasoning agents can accelerate de novo discovery by iterating with lab data remains unproven. Here we use protein-binding aptamer design as a sandbox (1-2 week cycles, \ 1-\ 3 per candidate) to demonstrate a wet lab in a closed loop with AI systems. We tasked agentic pipelines with developing aptamer binders against \alpha -thrombin and myoglobin across three design-test-improve cycles ( \sim 1,000 candidates per round; \sim 4,400 constructs). Statistically significant, bead-corrected on-target hit rates among new candidates rose across rounds from 0.0% in Round 1 to 26.0% in Round 2 and 42.1% in Round 3 on \alpha -thrombin (37.1% excluding literature motif descendants), and from 10.2% in Round 1 to 3.7% in Round 2 and 33.5% in Round 3 on myoglobin. Single-clone bead-binding titrations of eight full-length \alpha -thrombin candidates confirmed concordance with pooled sequencing enrichment (Pearson r = 0.90 ). In these assays, literature-derived motifs engineered by the agents to reduce off-target reactivity in the pooled screen reached apparent affinities of K_D = 24.6 – 133 nM, whereas de novo scaffolds showed low-micromolar binding ( K_D 1\ \mu M). We also report how our pipelines used myoglobin binding data to exploit the G-quadruplex binder space, diagnose cross-reactivity as on-target specificity dropped from 75.0% to 19.1%, and recover by developing strategies that restored specificity to 72.4%. We demonstrate that empirical lab data is more valuable than existing computational proxies, showing how iterative experimental feedback enables AI agents to diagnose assay artifacts and refine molecular designs across rounds. Comments: Equal leadership (co-senior authors): Pooja Rao and Annalisa Pawlosky. 34 pages, 4 main figures, 1 main table, 3 supplementary figures, 4 supplementary tables Subjects: Biomolecules (q-bio.BM); Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM) Cite as: arXiv:2610.24806 [q-bio.BM] (or arXiv:2610.24806v1 [q-bio.BM] for this version) https://doi.org/10.48550/arXiv.2610.24806 Focus to learn more arXiv-issued DOI via DataCite
[AI-192] Fine-Tuning Small Language Models for Reliable VASP INCAR Generation
链接: https://arxiv.org/abs/2608.05387
作者: Xinyue Zhang,Jixiang Li,Bin Shao,Baishun Yang,Zhiyang Liu,Weichao Wang
类目: Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI)
备注:
Abstract:Language models can prepare VASP INCAR files from natural-language requests, but so far only large proprietary cloud models come close to handling the tightly coupled, physics-sensitive settings reliably, a dependence that fits poorly with local, high-throughput materials workflows where privacy, cost, and offline deployment matter. We show that a small language model (SLM) can close this gap. The SLM is fine-tuned on reference VASP calculations and paired with VASPGuard, a deterministic post-processor that checks syntax, workflow, and material-dependent constraints; we call the combined model INCAR-SLM. On INCARBench, a benchmark for VASP INCAR generation, INCAR-SLM built on Qwen3-4B outperforms every general-purpose LLM evaluated, exceeding GPT-5.4 by 15.55 points on the 100-point INCAR Score. Most of this gain comes from fine-tuning, with VASPGuard correcting the errors that remain. We further find that model size matters less than expected: once fine-tuning and post-processing are applied, performance saturates at a few billion parameters, and Qwen3-4B outperforms larger models in the same family.
[AI-193] INCARBench: A Benchmark for Scientific Configuration in VASP INCAR by Large Language Models
链接: https://arxiv.org/abs/2606.23571
作者: Bin Shao,Jixiang Li,Xinyue Zhang,Baishun Yang,Zhiyang Liu,Weichao Wang
类目: Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are increasingly being integrated into first-principles computational workflows, yet their ability to configure scientific calculations remains poorly understood. Here, we introduce INCARBench, a benchmark for evaluating LLMs on input configuration for the Vienna Ab initio Simulation Package (VASP) through both configuration generation and repair tasks. Evaluating 19 model configurations reveals substantial capability differences among current frontier models. While several models achieve high semantic and policy accuracy, task-critical correctness remains substantially lower, demonstrating that parameter-level correctness does not necessarily imply scientifically valid configurations. Failure analysis shows that errors concentrate in physically coupled settings involving DFT+ U , magnetism, and correlated materials, where multiple constraints must be satisfied simultaneously. Repair evaluation further reveals that correcting incorrect settings and preserving already-valid configurations are distinct capabilities, with configuration preservation remaining a major challenge. These findings establish scientific configuration as a measurable capability of large language models and provide a foundation for developing more reliable AI systems for computational materials science.
机器学习
[LG-0] Planning under Noise with Neural Operators
链接: https://arxiv.org/abs/2610.28623
作者: Michael Chen,Hong Chul Nam,Xi Deng,Anima Anandkumar
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注:
Abstract:Planning under uncertain control execution noise requires reliable feedback policies across diverse environments. We introduce Planning Under Noise (PUN), a neural operator that learns noise-aware value fields for planning in unseen geometries. Building on an established screened-Poisson formulation of stochastic optimal control, we address two challenges: specifying boundary conditions that support navigation across varying geometries and efficiently computing fields with reliable gradients. Zero-Dirichlet obstacle boundaries terminate sampled paths at walls, weakening the goal signal along long routes and making it harder to estimate reliably. We instead impose reflecting Neumann conditions on obstacles and a Dirichlet condition at the goal, allowing sampled paths to continue exploring free space until they reach the target. To solve the resulting PDE without a configuration-space mesh, we extend a Monte Carlo (MC) solver, Walk on Stars (WoSt), to higher dimensions. However, local MC fluctuations can reverse estimated gradient directions and cause planning failures, making direct planning require costly sampling. PUN amortizes this computation by learning geometry-to-field mappings from noisy, partially converged Monte Carlo labels, avoiding expensive convergence for every training example. At inference, test-time scaling uses bounded local search guided by the predicted global field. We evaluate PUN on planning tasks across dimensions. Across four noise levels, PUN matches or outperforms all compared planners, achieving 95-100% success on the 4D replanning benchmark while maintaining the lowest runtime. At the highest tested noise level in 4D, PUN achieves 10.7-60.6% relative gains in success rate over other methods while running 3.29-32.15 times faster.
[LG-1] Efficient Edge-aware Attention Network for Graph Generation ICML2026
链接: https://arxiv.org/abs/2610.28618
作者: Vincent Jung,Alba Carballo-Castro,Yiming Qin,Lonneke van der Plas,Pascal Frossard
类目: Machine Learning (cs.LG)
*备注: Accepted at the Workshop on Graph Foundation Models and the Workshop on Structured Probabilistic Inference Generative Modeling, ICML 2026, Seoul, South Korea
Abstract:Graph generative models commonly use graph-specific Transformer architectures to jointly update node and edge features. However, these architectures do not directly benefit from recent hardware-aware attention implementations, limiting their scalability. The adaptation is, however, non-trivial: edge and graph-level features must be injected into pairwise attention scores, graph masks must be respected, and dynamic edge updates remain a major source of quadratic memory usage. We propose an edge-aware attention layer for graph generation that incorporates structural information through head-wise score modulation, while updating edge representations with a lightweight residual mechanism that retains attention-score dependent edge updates. This layer can be effectively implemented using PyTorch’s FlexAttention. We achieve similar generation quality to our baseline while reducing peak GPU memory by around 65% and substantially improving training and sampling throughput. These results suggest that the proposed architecture is a practical step toward scalable graph foundation models.
[LG-2] PaPi: Parallel Picard-Duhamel Integration for Fast Neuro-Spectral Architectures
链接: https://arxiv.org/abs/2610.28591
作者: Arthur Bizzi,Márcio Marques,Kleyton Da Costa,Vitor Balestro,Tiago Novello,João M. Pereira,Daniel Yukimura
类目: Machine Learning (cs.LG)
*备注: 27 pages, 8 figures
Abstract:Neuro-spectral architectures (NeuSA) are a class of physics-informed neural networks (PINNs) that have proven effective for solving time-dependent PDEs in the frequency domain. However, training and inference often rely on costly numerical integration of the learned vector field, typically using sequential Runge-Kutta time stepping, which limits temporal parallelism. We introduce PaPi, an integration scheme inspired by weakly nonlinear flows that enables time-Parallel inference through exponential integration and a Duhamel-Picard scheme. We show that the Duhamel-Picard iteration converges exponentially for all flows, and provide explicit error bounds that depend on how close the flow is to linear. Experiments show that PaPi accelerates training and inference by up to two orders of magnitude, achieving training times of under one second for low-dimensional PDE learning problems, while maintaining competitive predictive accuracy.
[LG-3] wo operatornameReLU Layers is All You Need
链接: https://arxiv.org/abs/2610.28587
作者: Florestan Brunck,Christoph Hertrich,Kilian Ruess,Jack Stade,Amir Yehudayoff
类目: Computational Complexity (cs.CC); Computational Geometry (cs.CG); Machine Learning (cs.LG)
*备注:
Abstract:We prove that every \operatornameReLU network can be exactly simulated by a \operatornameReLU network with two hidden layers. We present two proofs: one that is based on fillings in homological algebra and a direct combinatorial proof.
[LG-4] Continuous-Time Policy Gradient from Discrete Data: Convergence Analysis and Model-Free Algorithms
链接: https://arxiv.org/abs/2610.28580
作者: Martin Hernandez,Xinyan Qing,Mo Zhou,Yuhua Zhu
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Continuous-time reinforcement learning seeks an optimal policy under unknown dynamics driven by stochastic differential equations using observations collected at discrete times. We develop a policy gradient framework that combines rigorous convergence analysis with learning from discrete data. We first establish a performance difference lemma and derive policy gradient formulas for randomized and parametrized policies. We then prove exponential convergence of an entropy-regularized policy flow to the regularized optimum. For discrete parametrized updates, we derive a convergence bound that explicitly separates the contributions of critic error, policy-regression error, and the learning step size. To enable learning under unknown dynamics, we integrate this framework with PhiBE, which constructs a continuous-time surrogate generator from discrete transition data. The resulting on-policy and off-policy model-free actor-critic algorithms preserve the continuous-time structure of the control problem. Numerical experiments on an entropy-regularized linear-quadratic regulator and a nonlinear wheeled robot demonstrate the performance and stability of the proposed methods with both known and data-driven dynamics.
[LG-5] he Token-Energy Law of LLM Decode: Active Bytes and Realized Time Account for Joules per Token
链接: https://arxiv.org/abs/2610.28576
作者: Chenxu Niu,Batuhan Sencer,Wei Zhang,Yong Chen
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF)
*备注:
Abstract:Energy-efficient deployment of large language models requires understanding how model architecture, hardware, and serving settings shape energy per generated token. We introduce the Token-Energy Law, a compact, interpretable model of GPU-board energy per output token for dense, full-history decoding. It relates token energy to two core dimensions: active bytes and realized decode time. Active bytes describe logical byte demand per output, separating dense weights shared within a decode iteration from private key–value (KV) histories read and updated for each request. Realized decode time is the elapsed decode duration divided by the number of output tokens produced. We combine the two byte coordinates and time in a common affine form, with coefficients calibrated for each model and hardware/runtime configuration. This formulation explains how batching amortizes shared-weight work, longer attended histories increase private KV demand, and throughput determines the time-associated cost per output. Our experiments cover three models under BF16 decoding on NVIDIA Hopper and Blackwell GPUs, including two-GPU tensor parallelism on Hopper. Using fixed coefficients and observed decode time, held-out condition-mean energy estimates differ from measurements by at most 3.73%.
[LG-6] he Dynamics of Collapse and Recovery in Clustered Federated Learning
链接: https://arxiv.org/abs/2610.28565
作者: Priyanka Sinha,Snigdhansu Chatterjee
类目: Machine Learning (cs.LG); Dynamical Systems (math.DS)
*备注:
Abstract:Clustered federated learning (CFL) can suffer from cluster collapse. In particular, under unfavorable initialization, early client assignments concentrate heterogeneous groups on one cluster model, which is then trained toward a pooled compromise, while another model receives no updates. Since the current models determine client routing and routing determines subsequent model updates, we formulate clustered federated learning as control of an unknown dynamical system, with cluster models as the evolving state, client routing patterns as the control input, and routing interventions as state dependent controllers. In a two group linear model, we identify an initialization region in which greedy routing makes collapse self reinforcing. We then derive finite horizon error dynamics showing how routing interventions can reverse collapse by repeatedly training the starved model, primarily with updates from its compatible group. Motivated by this condition, each client uses blocked EXP3 to choose between greedy routing and interventions that redirect updates toward unused models. Experiments show that with simpler predictors and faster specialist formation, autonomous controller selection recovers usable specialists from unfavorable starts that trap standard CFL baselines. In transformed FEMNIST, the synchronized-entry configuration with persistent commitment recovers usable specialists in settings where the autonomous controller configuration often does not. Code and scripts for reproducing the experiments and figures are available at this https URL.
[LG-7] Asymmetric Bidirectional Predictive Dynamics in Google Search and COVID-19 Time Series: A Multi-Horizon Forecast-Error Analysis
链接: https://arxiv.org/abs/2610.28563
作者: Andrei Velichko,Imen Mahmoud
类目: Machine Learning (cs.LG)
*备注: 42 pages, 12 figures, 9 tables, 45 references
Abstract:Online search activity is widely used as a digital trace of information seeking during epidemics, but its predictive value is often reduced to correlation, a selected lag, or a single forecast-error score. We examine eight pre-specified Google Trends-COVID-19 pairs in both directions using matched LightGBM forecasts over rolling 400-day windows and horizons of 1-14 days. Log-RMSE gain is combined with forecast-error distribution entropy (FEDE), which captures horizon-dependent changes in the signed error distribution. The resulting directional predictive signature comprises response class, RMSE effect, FEDE endpoint and early-slope responses, persistence across rolling regimes, and horizon profile. For screening candidate directions, we also introduce a relative persistence-weighted directional predictive response strength (DPRS) on a within-study 0-100 scale; 100 denotes the strongest observed response in the comparison set, not an absolute coupling maximum. The pairs show concordant, RMSE-only, FEDE-only, and mixed responses. “loss of smell” - reported cases is accuracy-dominant, with mean log-RMSE gain 0.104 (about 9.9% equivalent RMSE reduction) and DPRS 100, whereas the reverse direction is FEDE-dominant (D_FEDE=0.048; DPRS=42). “covid test” - positive rate is FEDE-dominant (D_FEDE=0.039; positive in 72% of windows; DPRS=98) despite negative RMSE gain, while “lockdown” - stringency is the clearest concordant case (DPRS=78). World-level symptom/case pairs are less persistent. Thus, forecast accuracy and error-distribution dynamics are complementary, predictive effects can be direction- and metric-dependent, and forecast horizon is part of the relationship. Aggregated regime-level patterns should not be extrapolated to individual time points or interpreted as evidence of causality.
[LG-8] Convergence of stochastic gradient descent for learning linear neural networks
链接: https://arxiv.org/abs/2610.28561
作者: Gabin Maxime Nguegnang,Bubacarr Bah,Holger Rauhut,Ulrich Terstiege
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注: 30 pages
Abstract:Stochastic gradient descent (SGD) is the cornerstone of modern deep learning, yet its theoretical convergence properties remain poorly understood in non-convex settings. To address this gap, we study SGD dynamics in deep linear neural networks - simplified models that retain key non-convex characteristics while remaining analytically tractable. Building on recent results for gradient descent, we extend convergence guarantees to SGD under regularized squared loss. We establish almost sure boundedness of SGD iterates. Using the framework of asymptotic pseudotrajectories from stochastic approximation theory, our analysis establishes then, for both deterministic and adaptive step-size choices, convergence of SGD to the set of critical points under mild conditions (without relying on Lipschitz properties or whitened data). To our knowledge, this is the first work providing almost sure convergence guarantees for SGD in deep linear networks, bridging the gap between theoretical optimization and practical stochastic training.
[LG-9] Discount Yourself: Self-inclusive Freshness-weighted Gossip for Decentralised Learning under Client Outages
链接: https://arxiv.org/abs/2610.28545
作者: Qiyuan Wang,Ke Xiao,Christos Anagnostopoulos
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注: 52 pages, 9 figures, 35 tables; includes appendices
Abstract:Decentralised federated learning (DFL) trains personalised models on devices that average with their neighbours rather than with a central server. Real devices, however, are often offline for hours or days, and a client returning from an outage can recover only by aggregating the models its neighbourhood holds. We observe that the stalest of these models is usually the client’s own, a case that existing staleness-aware rules, which discount only the models a client receives, do not address. We formalise aggregation after an outage as estimating the state of a fully available network and derive three principles. First, the receiver must be discounted like any neighbour. Every rule that weights it at least as much as each neighbour has an error that grows without bound with the outage, whereas discounting all candidates by relative staleness keeps the error bounded. Second, staleness must be measured as inherited lag rather than age, since local training resets a model’s age but not the lag it absorbed from stale neighbours. We introduce the effective lag, which tracks this quantity at the cost of one scalar per message. Third, the discount must be steep, because absorbed lag persists in every later round. These principles define Self-inclusive Freshness-weighted Gossip (SFG), a family of drop-in aggregation rules with the communication and computation of decentralised federated averaging, whose guarantees hold for the stale receiver itself. On three image benchmarks and eight outage processes, including real smartphone availability traces, the lag-based members SFG-L and SFG-Cut consistently outperform decentralised federated averaging and recent age-of-information rules, and ablations show that each principle is necessary. The analysis also explains why tuned limited merging from gossip learning is effective.
[LG-10] GraphCast-Mamba: Learning Non-Markovian Corrections for Weather Forecasting
链接: https://arxiv.org/abs/2610.28512
作者: Ilya Vilkoviskiy,Liang-Hong Mo,Suozhi Huang,Dmitry A. Abanin
类目: Machine Learning (cs.LG)
*备注: 27 pages, 10 figures
Abstract:Unresolved atmospheric processes can introduce history dependence beyond the short input windows used by many weather forecasting models. We introduce GraphCast–Mamba, which augments a frozen, pretrained 1^\circ GraphCast backbone with a trainable Mamba correction branch combining graph message passing and Mamba layers. The branch retains a temporal memory state across successive six-hour forecast steps, learning corrections informed by temporal context beyond GraphCast’s two-frame input. Training is confined to the correction branch, using ERA5 data from 2015–2021. After 2,500 optimizer updates, including warm-up, the selected model achieves substantial improvements relative to the uncorrected GraphCast backbone. It reduces the GraphCast loss (normalized weighted mean-squared error; MSE) by 14.42% when aggregated over the full ten-day forecast and by 21.66% at day 10 (240 h after forecast initialization). At this final forecast lead time, root-mean-square error (RMSE) decreases by 7.09–14.60% across all 11 evaluated variable groups, spanning temperature, wind, pressure, geopotential, specific humidity, and precipitation. These results demonstrate broad improvements in ten-day weather forecasts through a short correction-training schedule while retaining the pretrained backbone.
[LG-11] Geometry-Agnostic Data-Driven Discovery of Extreme Scenarios
链接: https://arxiv.org/abs/2610.28487
作者: Angeline Hu,Ava K. Tse,Kaitlyn S. Yasumura,Luciana Sarabia,Malcolm Yeung,Kavin K. Prasanna,Kai Wong,William Hammond,Ioannis G. Keverkidis,Shengkai Li,Trung V. Phan
类目: Machine Learning (cs.LG); Mathematical Physics (math-ph)
*备注:
Abstract:We introduce a two-stage data-driven neural framework for discovering dynamically admissible extreme scenarios. Rather than assuming prior knowledge of the governing equations, the first stage infers the effective dynamical constraints directly from observational data. When the underlying physical system possesses suitable symmetries, the learned dynamics can be represented in a geometry-agnostic form, independent of any fixed grid, coordinate system, dimensionality, and intrinsic curvature. In the second stage, a physics-informed neural-search uses this learned dynamics to uncover rare, high-impact coherent structures that may not have appeared in the original observations. We demonstrate this framework through representative examples of increasing dimensionality and geometric complexity. Starting from a traditionally learned Duffing dynamics, we find a homoclinic pulse as a plausible localized extreme scenario. Next, since the same equation also appears in the stationary nonlinear Schrodinger system, with the coordinate interpreted as space instead of time, this pulse becomes a standing soliton. We then retrain the first stage using a geometry-agnostic neural representation, enabling the learned dynamics to be lifted to higher-dimensional space to obtain higher-dimensional solitary profile, known as Townes soliton. Finally, with a Liouville system, we demonstrate that dynamics learned from data on a (1+1)-dimensional Lorentzian quasi-manifold can be transferred to a (2+0)-dimensional Euclidean manifold to identify an instanton. Our approach can be applied to other coherent spatiotemporal structures, including kinks, rogue waves, compactons, peakons, and potentially even singularities, with applications in risk analysis, nonlinear forecasting, and anomaly detection.
[LG-12] A Bochner Weitzenböck Decomposition for Cellular Sheaf Laplacians Curvature and Neural Sheaf Diffusion
链接: https://arxiv.org/abs/2610.28455
作者: Francesco Maria Guadagnuolo,Alessio Borgi,Dorina Thanou,Pietro Liò
类目: Computational Geometry (cs.CG); Machine Learning (cs.LG)
*备注:
Abstract:We extend the Bochner-Weitzenböck decomposition to cellular sheaf Laplacians, splitting it into a positive semidefinite term, the Bochner sheaf Laplacian, and a block diagonal term, the Forman curvature. We show that the Bochner Laplacians form a class of cellular sheaves that extends both connection and matrix-weighted Laplacians, and generalizes signed graphs to d dimensions. We give a holonomy characterization of their kernel, prove the related Cheeger inequalities, and attach to them a random walk that generalizes the Gremban expansion. For the Forman term, we prove a discrete Bochner theorem, which bounds the first cohomology of a cellular sheaf with nonnegative curvature, and we read the whole decomposition as an opinion-dynamics model. We then apply this machinery to Sheaf Neural Networks, whose learnt Laplacian is otherwise unobserved. The Bochner interpretation yields a node-level conviction score that acts as an operational importance measure under frozen-model interventions, while the Forman profile provides a local descriptor of the learnt operator.
[LG-13] CNN-based classifier for subject-independent MBTI Extraversion/Introversion preference classification from short-time EEG signals
链接: https://arxiv.org/abs/2610.28442
作者: Filip Postepski,Grzegorz M. Wojcik
类目: Machine Learning (cs.LG)
*备注: Under review in Scientific Reports
Abstract:This study examines whether short, minimally preprocessed EEG segments contain sufficient information to distinguish between individuals classified as extraverted or introverted according to the Myers–Briggs Type Indicator (MBTI). EEG recordings from 50 participants were divided into 4-s segments and analyzed using a subject-independent nested 5-fold cross-validation procedure to compare EEGNet, EEGProgress, and a newly proposed compact convolutional architecture, SSCNet. SSCNet achieved the highest mean accuracy (0.6481), AUC (0.7063), and F1-score (0.6276). The proposed architecture also showed the lowest variability across folds. SSCNet demonstrated the best computational efficiency, with the shortest median training epoch time and the highest training throughput despite having a similar number of trainable parameters to EEGProgress. The results suggest that MBTI Extraversion–Introversion preferences may be moderately predictable from short time-domain EEG segments without manual spectral feature extraction, while highlighting the importance of subject-independent validation and computationally efficient model design.
[LG-14] Do Neural Models Prefer Their Submodels?
链接: https://arxiv.org/abs/2610.28439
作者: Alvaro Ribot,Yulia Alexandr,Anna Seigal
类目: Machine Learning (cs.LG); Algebraic Geometry (math.AG); Machine Learning (stat.ML)
*备注: 29 pages, 7 figures
Abstract:A neural model is the set of functions that can be expressed by a given neural network architecture. Neural models contain natural submodels, such as those obtained by setting some neurons to be inactive. We study the geometry of neural models and ask whether they prefer their submodels. By this we mean whether local minimizers of the loss function lie in a given submodel. We introduce a geometric framework to determine if this occurs. For shallow multilayer perceptrons, we show that subnetworks with inactive hidden neurons are never preferred. For deep multilayer perceptrons, we give conditions under which subnetworks with inactive hidden neurons are not preferred. For lightning self-attention, lower-rank submodels are usually not preferred, but we identify examples where they are because they lie on a boundary of the model. Our results show that, aside from a few exceptions, preference for submodels cannot be explained by the geometry of the neural model, and must instead come from other sources such as training dynamics.
[LG-15] Region-Specific EEG Classification of Hypnotic Susceptibility Using a Sliding-Window 2D CNN: A Subject-Independent Evaluation
链接: https://arxiv.org/abs/2610.28427
作者: Krzysztof Wrobel,Grzegorz M. Wojcik
类目: Machine Learning (cs.LG)
*备注: 8 pages, 2 figures, 3 tables
Abstract:We investigate whether a 2D Convolutional Neural Network (CNN) trained on minimally preprocessed EEG signals can classify hypnotic susceptibility across distinct brain regions. Sixty-two male participants (32 Control, 30 Guided Imagery / Hipno group) underwent a 20-minute EEG recording using a 256-channel EGI cap. A sliding-window approach evaluates all 10-second windows within the recording for each of seven predefined brain regions (prefrontal, frontal, temporal, central, parietal, occipital, and facial), yielding up to 868 independently trained models. Initial screening with a segment-level 90/10 random split showed that five regions exceeded the 0.67 accuracy threshold, with the occipital region achieving the highest mean accuracy (0.721) across 90.3% of windows. Subject-independent evaluation via Leave-One-Subject-Out (LOSO) cross-validation, however, substantially revised this picture: occipital performance collapsed to near-chance (mean ACC 0.453, AUC 0.429), revealing that high random-split accuracy was driven by subject-specific EEG fingerprints rather than genuine group differences. In contrast, the frontal and prefrontal regions maintained above-chance LOSO accuracy (0.646 and 0.624, respectively), demonstrating genuine generalisability. These results underscore the necessity of subject-independent validation in EEG-based classification studies and highlight anterior cortical regions as more reliable neural markers of hypnotic susceptibility.
[LG-16] Self-Credit: Credit Assignment for Long-Horizon Agent ic RL via Intra-group Rescaling
链接: https://arxiv.org/abs/2610.28388
作者: Fangcong Yin,Ming Yin,Tanvi Bhandarkar,Aashu Singh,Eryk Helenowski,Shlok Kumar Mishra,Saghar Hosseini,Karthik Abinav Sankararaman,Greg Durrett,Dinkar Juyal
类目: Machine Learning (cs.LG)
*备注:
Abstract:LLM agents are commonly trained with RLVR, but the sparse nature of typical RLVR rewards makes credit assignment challenging, especially for long-horizon tasks. We propose Self-Credit, an end-to-end agentic RL method that creates dense supervision signals from sparse outcome rewards using self-generated credit assignment via intra-group rescaling. Self-Credit builds on group-based RL with three components. First, during its rollouts, the policy model generates summaries with citations to earlier turns containing critical evidence, which serve as self-generated credit. Second, we refine this model-generated credit with intra-group rescaling, which compares cited actions across successful and failed rollouts for the same prompt. This approach uses group-level agreement and disagreement to remove undeserved credit and further emphasize credit assigned to critical actions. Third, advantage reweighting uses the rescaled credit weights to update the advantages of the cited turns. On BrowseComp-Plus, Self-Credit improves over the base models by 37.0 and 36.2 points on Qwen3 8B and 32B respectively, and outperforms the strongest agentic RL baselines by 8.5 and 5.6 points, respectively. Self-Credit also improves over the strongest baselines by 4.9 points on GAIA and 4.7 points on xBench-DeepSearch. Analyses show that self-generated credit provides useful signals for selecting critical actions, and that intra-group rescaling, especially on failed rollouts, is important across model scales.
[LG-17] Relative Representations Enable Portable White-Box Monitoring Across Language Models
链接: https://arxiv.org/abs/2610.28366
作者: Tim Bakker,Christos Louizos
类目: Machine Learning (cs.LG)
*备注:
Abstract:Independently trained autoregressive language models often acquire similar capabilities, but it remains unclear whether their internal representations organize behaviorally relevant information in compatible ways. We show that models do in fact organize information similarly by using specific behavioral probes that operate on compatible representations. This unlocks portable white-box monitoring: a probe developed and validated on one model can be reused when the model changes, without requiring labeled target data. To overcome incompatible native activation coordinates, we express each example through its angular distances to a shared set of anchors. We introduce a global anchored construction that summarizes information from all transformer layers and removes model-specific layer selection, so the same frozen probe can be straightforwardly applied to another model. We evaluate cross-model probe transfer across monitoring tasks on trivia datasets, safety monitoring, and mathematical reasoning. On average, transferred global anchored probes preserve at least 85% of same-model probe AUROC over chance. Anchored probes also support transfer to related datasets, remain robust across model scale and prompt changes, and retain performance better than standard probes after model updates. Transfer is nevertheless task- and target-dependent and can fail even when the monitored property remains strongly decodable in the target model. Overall, global anchoring reveals substantial task-relevant compatibility across independently trained models and provides a practical basis for portable white-box monitoring.
[LG-18] Beyond the Point Mass: Acceptance-Derived Weights and Distributional Losses for Speculative Decoding
链接: https://arxiv.org/abs/2610.28355
作者: Tianyu Wu,Zhuohan Wang,Jonathan Geuter,Yu Yao,Weifan Jiang,Chengxi Zhang,Ju Li,Minlan Yu
类目: Machine Learning (cs.LG)
*备注: Code: this https URL
Abstract:In speculative decoding, a lightweight draft model proposes tokens that the target model verifies in parallel, speeding up generation without changing the target distribution. D-PACE (Wu et al., 2026) trains such drafters with a loss that weights each position by its contribution to the expected accepted length. Both its weights and its cross-entropy (CE) loss use only the draft’s probability of the recorded target token, treating the target as a point mass. Under rejection sampling, however, a sampled draft token is accepted with probability equal to the overlap of the draft and target distributions, one minus their total variation (TV). We derive D-PAL (Dynamic Position-wise Acceptance Loss) from the exact expected accepted length, using this overlap as the acceptance signal. This yields overlap-based position weights with log acceptance (LA) as the per-position loss, and the objective reduces to D-PACE for a point-mass target. Since the weights are detached, any per-position loss can replace LA in practice. We show that forward KL divergence and Rényi divergence of order 1/2 match or exceed LA, and that Rényi-1/2 provably bounds the true acceptance rate from above and below. Replacing CE with Rényi-1/2 improves acceptance length under both D-PACE’s weights and ours. On six benchmarks with a Qwen3-4B target, D-PAL raises mean acceptance length over D-PACE from 4.50 to 4.57 at T=0 and from 4.20 to 4.28 at T=1, and consistently achieves higher throughput at both temperatures.
[LG-19] Characterizing Cross-Relay Generalization Failure in Machine Learning-Based Power System Protection
链接: https://arxiv.org/abs/2610.28348
作者: Nithin Pradeep Nadayil Kizhakkethil,Julian Oelhaf,Christian Bergler,Andreas Maier,Siming Bayer
类目: Machine Learning (cs.LG)
*备注: 6 pages, 3 figures. Accepted for presentation at the IEEE Power Energy Student Summit (PESS 2026), Karlsruhe, Germany
Abstract:Machine learning-based fault detectors are often evaluated at the relay locations used for training, although local voltage and current measurements vary across power networks. This work investigates cross-relay generalization for binary detection of faults anywhere in the network using local measurements, rather than line-selective tripping. Using 50 ms windows from a simulated 110 kV transmission network, we evaluate a multilayer perceptron across all 64 source target combinations of eight relays and five random seeds. A shared simulation-level split and source-only threshold selection provide a controlled held-out evaluation without target-specific calibration. Despite nearly unchanged mean F1-scores (0.730 within-relay versus 0.734 cross-relay), the mean false-positive rate (FPR) increases from 0.088 to 0.341, while the mean false-negative rate decreases from 0.399 to 0.318. The change is strongly directional: models trained at Bus 2 relays yield a mean FPR of 0.670 at Bus 1 and Bus 3 relays, compared with 0.096 for cross-relay transfers within the same bus group. Linear probing indicates that learned representations retain coarse bus-group information, while exploratory class-conditional latent distribution analysis shows no consistent monotonic association with transfer errors. These findings demonstrate that aggregate metrics can conceal location-structured changes in false-alarm behavior, highlighting the importance of evaluating error types separately when assessing cross-relay generalization.
[LG-20] Observation Assimilation Transformer: Probabilistic Weather Analysis from Sparse Station Observations
链接: https://arxiv.org/abs/2610.28327
作者: Christian Donner,Michele Cattaneo,Maxim Samarin,Francesco Zanetta,Icíar Lloréns Jover,Garbiela Aznar Siguan,Michele Volpi
类目: Machine Learning (cs.LG)
*备注:
Abstract:We present the Observation Assimilation Transformer (OAT), a geometry-aware neural process that corrects and refines a prior numerical weather prediction (NWP) field into a high-resolution posterior field by assimilating sparse observations. OAT is an encoder-decoder transformer that considers a set of observed locations and their NWP prior to condition predictions at a set of arbitrary unobserved query locations. Its sparse attention mechanism makes it scalable to thousands of station observations and depends on relative positions between locations, never on the absolute coordinates. The model allows for partially missing observations, predictions at arbitrary continuous locations without grid projection, and outputs are parametric distributions. A station-randomized training algorithm is developed with continuous ranked probability score (CRPS) loss. We test OAT on two datasets - one at regional scale with hundreds of stations, and one at continental scale with thousands of stations. On these datasets OAT outperforms several baselines and shows well calibrated predictions. In an illustrative case study, OAT’s downscaling capabilities are demonstrated by downscaling ERA5 (~31km) to COSMO (1km), where OAT achieves lower mean absolute error (MAE) and CRPS for temperature and precipitation on test stations compared to COSMO.
[LG-21] Manifold-Informed Neural Networks for Low-Label ECG Electrolyte Inference
链接: https://arxiv.org/abs/2610.28322
作者: Juntang Wang,Hao Wu,Yihan Wang,Dongmian Zou,Shixin Xu
类目: Machine Learning (cs.LG)
*备注: 19 pages
Abstract:Physics-informed neural networks (PINNs) supervise on the residual of a known differential operator. Many high-stakes inference problems admit no such operator: clinical signals, behavioral measurements, materials properties. We introduce the Manifold-Informed Neural Network (MINN), a PINN-parallel paradigm that replaces the physics operator residual with a manifold operator residual: a discrete Poisson equation \mathbfL_G \mathbfH = s_\phi(\mathbfx, \mathbfm) on a k -nearest-neighbor graph in the learned latent space, with a learned, task-aware source. The construction is a structural parallel to PINNs (squared operator residual plus label data, in place of operator residual plus boundary data); the s=0 limit recovers the homogeneous Laplace equation \mathbfL_G\mathbfH=\mathbf0 (classical manifold regularization solves the same equation but minimizes the Dirichlet energy \operatornametr(\mathbfH^\top\mathbfL_G\mathbfH) , a different penalty form). The source s_\phi takes input features and auxiliary metadata directly, breaking the homogeneous smoothness prior that limits classical Laplacian penalties; gated Feature-wise Linear Modulation separately fuses metadata into the prediction branch. We evaluate MINN on three-class classification of potassium, calcium, and magnesium from 12-lead ECGs in MIMIC-IV-ECG, against twelve baselines spanning supervised and tabular methods, ECG foundation models, and manifold/graph approaches. MINN attains the highest K ^+ Macro-F1 of 0.61 (best baseline 0.55), with consistent gains on Ca ^2+ and Mg ^2+ and a 31% relative improvement at 1% labeled data. Six targeted ablations show that the inhomogeneous Poisson formulation, not graph smoothing alone, drives the gain.
[LG-22] Recommending Bacterial Cultivation Conditions from 16S rRNA
链接: https://arxiv.org/abs/2610.28310
作者: Juntang Wang,Hao Wu,Yihan Wang,Runkun Guo,Shixin Xu,Dongmian Zou
类目: Machine Learning (cs.LG)
*备注: 19 pages
Abstract:Recommending bacterial cultivation conditions from sequence data is a challenging structured problem characterized by missing standardized benchmarks and limited use of sequence structure. To address this, we introduce DSNI, a benchmark of 65,023 bacterial strains with phylogeny-aware splits, pairing 16S rRNA sequences with four cultivation conditions (media, temperature, pH, oxygen). Exploiting the mosaic structure of 16S rRNA, where hypervariable regions support discrimination and conserved regions support primer design, we propose a Variability-Gated Encoder that learns position-dependent importance weights conditioned on biological features (region type, position, and local complexity). The gating module adds minimal overhead (1,280 parameters, 1%) while providing an explicit inductive bias valuable in low-data regimes. Experiments demonstrate that the learnable gate improves flat Transformers by 4.7% in joint accuracy and complements RNABERT, yielding 44.2% exact-match joint accuracy. The model also recovers 90.2% of ground-truth cultivation protocols within its top-25 recommendations, narrowing the exhaustive search space by approximately 60 times.
[LG-23] Learning Normal Coordinates on Riemannian Manifolds with Flow Charts
链接: https://arxiv.org/abs/2610.28302
作者: Edmond Cunningham
类目: Machine Learning (cs.LG); Differential Geometry (math.DG)
*备注:
Abstract:Riemannian manifolds are powerful models for problems with non-Euclidean structure. Recent work in geometric machine learning has focused on training deep neural networks to learn Riemannian metrics from high-dimensional data. However, many geometric algorithms require repeated evaluations of the exponential and logarithmic maps, whose high computational cost restricts the utility of learned Riemannian manifolds in practice. We propose to address this cost by training a conditional invertible neural network, called a flow chart, to approximate normal coordinates around points on a given Riemannian manifold. With this approach, the exponential and logarithmic maps can be approximated by fast Euclidean operations in the model’s learned coordinate space. Our main contribution is an efficient loss function for training flow charts that is nonnegative and, under suitable domain assumptions, is zero precisely when the chart is normal. Across sixteen synthetic and data-derived geometries, flow charts are competitive with a specialized neural geodesic interpolator for logarithmic map estimation, while also being able to estimate the exponential map. When pretrained, flow charts provide a median speedup of 60-70x over numerical baselines at comparable logarithmic map estimation error, and their learned coordinates improve the accuracy of a numerical geodesic solver.
[LG-24] ExeBench-FHD: Addressing Fidelity Issues in Decompiler Evaluation
链接: https://arxiv.org/abs/2610.28212
作者: Akarsh Ashok,Dinil Mon Divakaran,Vivek Balachandran
类目: oftware Engineering (cs.SE); Machine Learning (cs.LG); Programming Languages (cs.PL)
*备注:
Abstract:ExeBench is one of the most widely used benchmarks for evaluating LLM-based decompilers. We analyze ExeBench and identify nine issues that affect how faithfully a test measures a decompiler. Three of them prevent LLM decompilers from recovering C code from ExeBench’s function. These are ambiguous parameter order when integer and floating-point types are mixed, unused parameters that optimized assembly may drop, and constants stored outside the function’s assembly code. Six more affect test validity: non-void functions without a return, placeholder types, static or inline functions removed by the optimizer, functions named main, trivial bodies and truncated definitions. We present ExeBench-FHD, a backward-compatible revision of ExeBench that adds 16 new fields: call-site assembly, data sections, test wrappers, linked-binary disassembly and five warning flags. In ExeBench, 38.9% of functions have at least one of the five flags; dropping records with any of the four warning flags (all except unused_param, which marks functions that the new wrappers make evaluable) keeps 63.9% of the corpus. In an evaluation at -O2 on four architectures, the new fields raise GPT 6 Sol’s re-execution rate by up to 26.5%, 19.1% and 21.0% on functions that read data sections, mix parameter types and have unused parameters, respectively.
[LG-25] Foreword: Pre-Generation Prediction and Characterisation of LLM Test Failures using the Jacobian Lens
链接: https://arxiv.org/abs/2610.28187
作者: Somin Kim,Robert Feldt,Shin Yoo
类目: oftware Engineering (cs.SE); Machine Learning (cs.LG)
*备注:
Abstract:Testing LLM-based software is expensive: each input needs several sampled responses, each judged by a costly oracle. Further, every prompt or model change calls for re-testing. Pre-generation adequacy measures cut this cost by ranking inputs by failure risk based on the model’s internal state. However, the state of the art, Clotho, models only passing reference inputs, and its hidden-state representation says nothing about why inputs fail. We present Foreword, a framework built on two ideas: learn from both passing and failing references, and represent inputs through the Jacobian lens (J-lens), an interpretability technique that reads intermediate activations out as scores over vocabulary tokens. Foreword pairs these representations with two estimators: a sparse logistic-regression probe or dual Gaussian Mixture Models (dual GMMs), which also yield a reference density (to flag inputs the references do not cover) and failing-class components (to group similar failures). On ten tasks and three open-weight LLMs, all four combinations of representations and estimators outperform Clotho on average, both for prediction of failure (ROC-AUC improves by 0.14 on average and by up to 0.31) and test prioritisation (15% more failures among the 100 top-ranked inputs on average, up to 33%). Prioritisation obtained using small models also transfers better to proprietary models (on average 22% and up to 41% more failures in the first quarter of prioritised inputs). Importantly, our qualitative case study shows that Foreword supports semantic characterisation of failures by pointing to ambiguities in labels or task specifications. Combined, Foreword can support testers for prioritising inputs, checking support for a prediction, and characterising specific failures, all without generating output.
[LG-26] Coincidence-Maximizing Customer Clustering Using Smart Meter Data
链接: https://arxiv.org/abs/2610.28178
作者: Ramiro H. Gálvez,Gabriel Martos Venturini,Iván Mondrzak
类目: Machine Learning (cs.LG)
*备注: 10 pages, 5 figures, 4 tables
Abstract:Electricity distribution planning depends on peak-demand coincidence within customer groups, yet similarity-based clustering of smart-meter data does not optimize it directly. We formulate customer grouping to maximize the average of group coincidence factors, weighted by each group’s daily share of summed individual peak demand. We prove that jointly optimizing memberships and group peak times is strongly NP-hard even for two groups. We therefore develop a fixed-peak refinement procedure applicable to any feasible partition. It alternates between identifying group peak times and exactly reassigning customers under minimum group-size constraints. Each polynomial-time reassignment maximizes a lower bound tight at the current partition, so accepted updates strictly improve weighted coincidence. We initialize refinement by clustering individual peak-time distributions and evaluate the method on four public smart-meter panels (27,654 customers) using chronological holdouts and 200 paired replicates with complete refitting. Refinement consistently improves weighted coincidence during fitting and reduces sensitivity to initialization. On held-out days, the method outperforms random grouping and five standalone benchmarks, with mean paired gains of 1.08-3.48 percentage points over the strongest benchmark. Requiring nearly equal group sizes lowers test weighted coincidence only modestly. The groups retain recognizable peak-time patterns over time, suggesting applications in peak-aware customer aggregation.
[LG-27] Curriculum Learning for Generative Flexible Docking NEURIPS2026
链接: https://arxiv.org/abs/2610.28145
作者: Wenyin Zhou,Hossein Azizpour
类目: Machine Learning (cs.LG)
*备注: ML4Molecules: Agentic Systems for Molecular Sciences Workshop at NeurIPS 2026 Paris
Abstract:Flexible docking is a highly practical task, aiming to predict the joint configuration of a protein-ligand complex from a protein apo structure and a ligand SMILES string. Diffusion and flow-based generative models have made this task tractable by modeling the posterior distribution over the ligand pose and the protein conformation conditioned on the apo input. The central challenge, however, remains the long-tailed apo-to-holo structure shift, where a small fraction of protein undergo large conformational changes. In this work, motivated by the practical observation that downweighting large shift examples improves generalization, and by the mixed nature of the long tail consisting of genuinely difficult examples and noisy ones, we propose to use curriculum learning for this task, gradually incorporating large conformational change examples during training. With a linear pacing function, we show that an existing method can be strongly enhanced. On the PDBBind ESMFold apo benchmark, with relaxation and confidence-based pose selection, the resulting model GCLDock achieves a 46.52% top-1 success rate (ligand RMSD less than 2A) and 40.26% when additionally requiring PoseBusters validity, outperforming existing baselines.
[LG-28] Fixed Sets and the Hardness of Computing Correlated Equilibria
链接: https://arxiv.org/abs/2610.28141
作者: Ioannis Anagnostides,Constantinos Daskalakis,Gabriele Farina,Noah Golowich,Tuomas Sandholm,Brian Hu Zhang
类目: Computer Science and Game Theory (cs.GT); Computational Complexity (cs.CC); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
*备注:
Abstract:We show that computing an approximate correlated equilibrium of a polytope game is PPAD-complete, and thus polynomial-time equivalent to finding a Nash equilibrium. This challenges the conventional view in game theory that correlated equilibria are more tractable objects than Nash equilibria. From a conceptual standpoint, we connect correlated equilibria to a natural generalization of fixed points, which we call fixed sets. A fixed set of a continuous function f is a nonempty finite set S such that S \subseteq \operatornameconv f(S) . Our main technical contribution is to prove that computing a fixed set is PPAD-hard even at inverse polynomial precision.
[LG-29] MuPACE: Learning to Pace Muon for LoRA Fine-Tuning
链接: https://arxiv.org/abs/2610.28072
作者: Jan Kobiolka,Christian Frey,Josif Grabocka
类目: Machine Learning (cs.LG)
*备注: Under Review
Abstract:Learning rate schedules are typically fixed before training begins, yet when fine-tuning large language models with low-rank adapters (LoRA), the appropriate scale varies substantially with the model, task, and adapter configuration. Practitioners re-tune a schedule for every new problem, yet a tuned schedule still cannot react to the training dynamics it encounters. We instead amortize learning rate selection across LoRA fine-tuning problems. We leave the base optimizer, Muon, untouched, so that its normalized update fixes the direction, and meta-learn only the global step size. Our method MuPACE is a small recurrent policy, meta-trained across a distribution of fine-tuning runs, that conditions on the observed trajectory and emits bounded adjustments to an interpretable analytic schedule. On 64 fine-tuning problems, MuPACE outperforms standard fixed schedules at a strong shared default on 64-78% of problems and is on par with the same schedules with per-task tuned peaks. Meta-trained only on models with up to 16 layers, MuPACE transfers one-shot to unseen tasks on models as deep as 64 layers, such as Qwen3-32B and Qwen3.8-27B, where a single run remains competitive with swept schedules.
[LG-30] Universal Online Convex Optimization With Delayed Feedback
链接: https://arxiv.org/abs/2610.28058
作者: Hao Qiu,Andrew Jacobsen,Emmanuel Esposito,Mengxiao Zhang
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study universal online convex optimization with delayed feedback. We develop a single algorithm that simultaneously adapts to general convex, strongly convex, and exp-concave losses without prior knowledge of the curvature class or parameters. Our regret bounds match the best known dependence on the time horizon T and delays up to doubly logarithmic factors: roughly \sqrtT + d_\mathrmtot for convex losses and \sigma_\max\log T rates with curved losses, where d_\mathrmtot is the total cumulative delay and \sigma_\max is the maximum number of simultaneously missing observations. Our main technical contribution is a delayed meta-learner with a second-order excess-loss guarantee, which enables curvature-adaptive aggregation of delayed base learners. In addition, our new exp-concave base learner yields a sharper regret bound which improves the state-of-the-art delay dependence by replacing the maximum delay with the potentially much smaller \sigma_\max , while retaining the complementary \sqrtd_\mathrmtot dependence.
[LG-31] When Does Training on Synthetic Data Provably Work?
链接: https://arxiv.org/abs/2610.28036
作者: Leroy Odunlami,Tobias Riedlinger,Hanno Gottschalk
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:High-quality training data are a major bottleneck in machine learning, motivating increasing use of artificially generated data. We study when a binary classifier trained on synthetic data transfers successfully to a real distribution, under a shared conditional label distribution. First, we derive finite-sample excess-error bounds when the real covariate distribution is absolutely continuous with respect to the synthetic distribution under a controlled density ratio. Second, we study the partially and fully singular case, where the real distribution has mass outside the synthetic one. If the hypothesis class is sufficiently expressive, synthetic risk alone cannot uniformly control real-world error: a classifier can be nearly optimal on synthetic data while approaching worst-case error on the uncovered real region, even when the distributions are arbitrarily close in Wasserstein distance. For function classes where hypotheses and the log-odds of the conditional label probability are Lipschitz, we recover a bound in terms of the probability that a real sample lies outside the reach of the synthetic data, restoring a transfer guarantee even in the fully singular case. Experiments confirm this in controlled low-dimensional settings and on 32,200 images generated as counterparts to real ImageNet images by four state-of-the-art generators: an adversarial classifier reaches 0.5% error on the synthetic domain but 99.25% error on the real domain, while ordinarily trained architectures show positively correlated synthetic and real test errors. Synthetic data should be budgeted at high sample counts and models validated on real data.
[LG-32] RANGER: Retrieval Agent for Navigating Graphs with Semantic Evaluation Rewards
链接: https://arxiv.org/abs/2610.28010
作者: Artem Mikhaylov,Pavel Konoplev,Nikolay Efanov,Artyom Sorokin,Evgeny Burnaev
类目: Machine Learning (cs.LG)
*备注: 12 pages, 8 figures, 3 tables. Includes appendices
Abstract:The integration of Large Language Models and Reinforcement Learning offers a promising direction for complex sequential decision-making tasks, such as multi-hop question answering over knowledge graphs. While LLMs provide strong semantic understanding and reasoning capabilities, they often struggle with precise, multi-step navigation in structured environments. Conversely, RL agents excel at learning navigation policies but are hindered by the difficulty of designing semantically meaningful reward functions. This paper introduces a novel hybrid approach in which an LLM serves as a mentor to an RL agent during training by providing dense, semantically informed reward signals. The RL agent, built on a Transformer-based policy and trained with Group Relative Policy Optimization, learns to navigate a knowledge graph to retrieve relevant information. The LLM evaluates the semantic relevance of the agent’s traversed paths, generating reward signals that align the agent’s behavior with the information needs expressed in a query. We integrate this Retrieval Agent for Navigating Graphs with Semantic Evaluation Rewards (RANGER) into a complete question-answering pipeline and evaluate it on the open question answering datasets. Experimental results show that the proposed approach outperforms both a baseline retrieval-augmented system with classical graph search and one using an RL agent trained with heuristic rewards, achieving a significant increase in NLP metrics and LLM-as-a-Judge scores. Our findings demonstrate that deep integration of LLMs into the RL training loop yields more effective and semantically aligned retrieval policies for complex reasoning tasks.
[LG-33] When Momentum Helps Polyak–Ruppert Averag ing
链接: https://arxiv.org/abs/2610.27990
作者: Alexandra Suvroikova,Darina Dvinskikh,Yuriy Dorn,Alexander Gasnikov
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Polyak–Ruppert averaging gives SGD and momentum methods the same asymptotic covariance. We characterize how fast finite runs reach it and how far they can be parallelized. For strongly convex quadratics with additive noise, we prove an online–offline bridge: the empirical risk minimizer on the samples of a tail average is one Newton step from the tail-averaged query point of any online procedure. For SGD, heavy ball, and Nesterov’s method with constant momentum, this step is a boundary term at the two ends of the tail and splits the gap to the offline estimator into center, covariance, and shape errors. Necessary and sufficient conditions show that, on the least curved direction, the center error sets the pace: SGD needs m\gg\kappa^2 rounds but momentum only m\gg\kappa , where \kappa is the condition number, so at sample size N the critical batch size grows from \Theta(\sqrtN/\kappa) to \Theta(\sqrtN/\kappa) . The covariance error separates the momentum methods: heavy ball has relative variance excess \Theta(\kappa^3/2/m) along the stiffest direction, whereas Nesterov has relative error O(\sqrt\kappa/m) in every direction. The shape error is method-independent by a Berry–Esseen bound uniform in \kappa and batch size, so confidence intervals are valid when the other two errors vanish. We further examine how the conclusions change under heavy-tailed or state-dependent noise and clipping, and establish a first-order comparison with ERM for nonlinear Nesterov averaging under explicit conditions.
[LG-34] On Johnson-Lindenstrauss Transforms for Byzantine-Robust Optimization
链接: https://arxiv.org/abs/2610.27989
作者: Aleksandr Shestakov,Vladimir Aletov,Aleksandr Beznosikov
类目: Machine Learning (cs.LG)
*备注: 37 pages, 10 tables
Abstract:The growing scale of data and models in machine learning calls for efficient optimization methods. Distributed optimization addresses this challenge, with compression mechanisms helping to reduce communication costs. A classical tool for dimensionality reduction is the Johnson-Lindenstrauss (JL) transform, which is simple to implement and offers strong geometric guarantees. In particular, JL transforms approximately preserve inner products and pairwise distances between vectors with high probability. In this paper, we exploit this property in Byzantine-robust optimization, where worker messages are weighted according to their alignment with a trusted reference gradient. We further support our theoretical findings with experimental results.
[LG-35] Neural Closed-Loop Identification of Nonlinear Systems with Stability and Generalization Guarantees
链接: https://arxiv.org/abs/2610.27970
作者: Laura Meroi,Mahrokh G. Boroujeni,Leonardo Massai,Giancarlo Ferrari-Trecate
类目: ystems and Control (eess.SY); Machine Learning (cs.LG)
*备注: 16 pages, 11 figures. Preprint submitted to Automatica
Abstract:Neural networks are expressive models of nonlinear dynamics, but identifying such systems from closed-loop data raises two difficulties: the feedback loop correlates inputs with measurement noise, biasing standard open-loop estimators, and the identified model must be stabilized by the data-collection controller, so that its closed-loop behavior matches the true plant’s. We derive a complete parameterization of all nonlinear systems stabilized by a given controller, generalizing the classical dual-Youla parameterization to the nonlinear setting. Building on this characterization, we propose Internal Controller Identification (ICI), a computationally tractable framework for identifying potentially unstable systems while ensuring closed-loop stability by construction; in the linear case, our indirect method reduces to an open-loop identification problem, for which standard consistency results apply. The framework accommodates contextual information, letting a single certified model identify a family of systems, and supports a PAC-Bayesian training objective that, when data are scarce and prior knowledge is available, bounds the identified operator’s expected prediction loss. Experiments show that ICI identifies open-loop unstable plants where uncertified baselines diverge, outperforms direct methods under weak excitation and feedback-correlated noise, generalizes zero-shot to environments unseen during training, and, with limited data, avoids the overfitting of empirical risk minimization while providing a non-vacuous generalization bound.
[LG-36] Unsupervised Head Network Personalization for Split Computing via Federated Learning
链接: https://arxiv.org/abs/2610.27952
作者: Yutaro Horikawa,Takayuki Nishio
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Networking and Internet Architecture (cs.NI)
*备注: 12 pages, 8 figures, 4 tables. This work has been submitted to the IEEE for possible publication
Abstract:Split Computing (SC) is a distributed inference method designed for balancing computational load and reducing latency by partitioning a neural network into head and tail networks and deploying them across both a mobile device and a server, respectively. Due to the limited bandwidth and computational power of mobile devices, the head network is often replaced with a neural network that requires lower computational and communication resources. However, it is challenging to prepare a sufficient quantity and quality of labeled data on the mobile device side for training the replacement head network. Furthermore, the data prepared by mobile users often contains sensitive privacy information, making data aggregation difficult. In this paper, we propose a data-efficient head network training method for SC. This method utilizes federated learning and knowledge distillation to train the head network using a small amount of non-IID and unlabeled data from mobile devices, ensuring that the training datasets are not exposed. Moreover, by introducing personalization, we enhance the accuracy to levels comparable to or exceeding those of existing methods. The experimental evaluations demonstrate that, despite data constraints imposed only on our method, it achieves accuracy comparable to the baselines. Furthermore, it significantly reduces the communication requirements by 95% on mobile devices, with only a 3-5 percentage point drop in accuracy from the original model.
[LG-37] Near-Duplicate Image Selection Transfers Preferences in Vision-Language Models
链接: https://arxiv.org/abs/2610.27943
作者: Jaeyoung Kim,Eunseok Kim,Dongsuk Jang,Seunghyun Kim
类目: Machine Learning (cs.LG)
*备注: 43 pages, 7 figures. Preprint; under review
Abstract:Subliminal learning is the phenomenon in which a model transmits its preferences to another model through data that do not explicitly express them. As models increasingly curate training data for other models, we ask whether this transfer can arise from choosing among near-duplicate images while keeping the source photographs and human-written captions fixed. For each photograph, a curator selects one of eight near-duplicate variants by caption likelihood, without target-specific prompts. Students trained on target-preferring curators’ selections show higher mean target-versus-alternatives log-likelihood ratios in an image-free readout than students trained on neutral selections. Under a prespecified decision rule, this directional shift appears for an owl target in LLaVA-1.5, Qwen2.5-VL, Qwen3-VL, and Gemma 4, with curators and students initialized from the same base checkpoint, and target rotations move it toward camels, oak trees, and a political figure. The owl target-selected student also scores above all count-matched random-reselection students in three models and above 55 of 60 in LLaVA-1.5, and Qwen3-VL shows the shift again on a new source bank with newly trained curators. Its size depends on dataset size, learning rate, and training budget; greedy choices among the eight options move toward the target by 10.4 percentage points in Qwen3-VL and 0.5-1.3 points elsewhere. First-generation students also pass the owl preference to a second generation through captions without explicit owl mentions in LLaVA-1.5 and Qwen3-VL. Thus even a deduplication-style choice among near-identical copies can shift students’ target-relative readouts with source photographs and captions fixed; the tested linear probes classify the selecting arm with accuracies up to 59.7% (chance 50%).
[LG-38] Elicitation or Expansion? Demystifying Gains from Agent Harness Optimization
链接: https://arxiv.org/abs/2610.27889
作者: Younghyun Kim,Wooseok Seo,Minseo Kim,Seunghyuk Oh,Minjae Lee,Donghoon Kim,Hyung Il Koo,Jaehyung Kim,Wonjun Kang
类目: Machine Learning (cs.LG)
*备注: Project page: this https URL
Abstract:Recent studies in harness optimization have shown substantial performance gains in language model agents without updating model weights. A well-designed harness more effectively leverages the underlying model’s capabilities by shaping context, tool use, and execution workflows. However, naïve success rates alone do not fully characterize whether optimized harness yields more reliable success on tasks already within reach (elicitation) or a broader range of solvable tasks (expansion). This distinction is crucial for recursive self-improvement and evolutionary search, as expansion may provide a broader range of successful experiences to push the capability boundaries. Despite the growing significance, this distinction remains underexplored in harness optimization. We conduct a systematic analysis to study the effect of gains from harness optimization. Building on pass@k analyses in the reinforcement learning (RL) literature that compare models before and after RL training, we compare optimized harnesses with their pre-optimization counterparts. Specifically, we assess changes in reachable task coverage, compare harness optimization with model scaling and increased inference-time compute, and examine how gains vary with task difficulty. Across static prediction and dynamic tasks, we find that pass@1 advantages of harness over stronger interventions can reverse at larger attempt budgets and aggregate gains conceal shortfalls on hard tasks, underscoring the need for a fine-grained diagnosis.
[LG-39] Enumeration-Guided Training of Neural Hamilton-Jacobi Safety Certificates
链接: https://arxiv.org/abs/2610.27870
作者: Mihaela-Larisa Clement,Tosca Pierro,Ezio Bartocci,Luca Marzari
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Hamilton-Jacobi (HJ) reachability analysis offers formal safety guarantees for control systems, but training accurate neural certificates remains difficult, particularly when unsafe regions are hard to characterize. We propose a verification-guided training framework that addresses this by using probabilistic enumeration of safety constraints to detect forward-invariance violations, i.e., states where the learned certificate incorrectly classifies unsafe behavior. This produces a compact, targeted set of counterexamples that more precisely captures the failure boundary than inputs generated by standard adversarial attack methods. Crucially, using this representation to fine-tune a pretrained HJ value function yields larger certified regions and reduces certification time across all tested benchmarks. We further introduce a shielding mechanism based on the refined certificates to enforce safety constraints at deployment time. We evaluate the proposed safety filter in a realistic, physics-based Isaac Sim environment and through real-world autonomous racing experiments, achieving zero safety constraint violations compared to a nominal controller and a backup policy based on Linear Quadratic Regulation (LQR). These results demonstrate the effectiveness of the proposed shielding mechanism in ensuring safe operation under high-speed, safety-critical conditions.
[LG-40] SliMOO: Interpretable Multi-Objective Evolutionary Search for LLM Depth Pruning
链接: https://arxiv.org/abs/2610.27812
作者: Guanchen Li,Yixing Xu,Xuanwu Yin,Dong Li,Emad Barsoum
类目: Machine Learning (cs.LG)
*备注:
Abstract:As large language models (LLMs) grow more powerful, their scale makes them costly to serve. Depth pruning offers a practical route to efficiency by directly reducing memory and latency without hardware-specific support. However, we observe that transformer layers exhibit markedly different task-dependent layer contributions, whereas existing depth-pruning methods ignore this heterogeneity and thus tend to optimize for a narrow domain or benchmark rather than preserve broad capability. This suggests that LLM depth pruning should be formulated as a multi-objective optimization problem rather than a single-score ranking problem. We propose SliMOO, an interpretable multi-objective evolutionary search framework for LLM depth pruning. SliMOO first uses single-layer removal probing to construct a task-layer importance map, which provides an interpretable view of task-shared and task-specific layers and serves as a search prior. It then performs a Pareto-aware evolutionary search for pruning masks, where candidates are generated under the guidance of the task-layer prior, scored by their deviation from the dense model on each task, and retained through NSGA-II environmental selection. Across the Llama-3.1 and Qwen3 model families, SliMOO consistently achieves stronger multi-task trade-offs than rule-based, greedy, and scalarized baselines. For example, on Llama-3.1-70B at 25% sparsity, it improves the math-domain average from 24.33% to 34.40% and the code-domain average from 35.03% to 39.14%. Code will be made available at this https URL. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2610.27812 [cs.LG] (or arXiv:2610.27812v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.27812 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-41] Certified Adaptive List Decoding for Error-Correcting Codes
链接: https://arxiv.org/abs/2610.27805
作者: Jaeho Hwang,Byung-Sun Won,Uiwon Hwang
类目: Information Theory (cs.IT); Machine Learning (cs.LG)
*备注:
Abstract:Maximum-likelihood (ML) decoding of linear block codes is NP-hard, so neural decoders are evaluated by how closely they approach it. We observe that proving a candidate optimal can be far easier than finding it, and build a decoder that does both. We cast decoding as a search over an absorbing-state discrete diffusion model, whose sequential commitment of coordinates turns generation into a tree, and run a probability beam search over it. A dual bound from a linear-programming relaxation and a distance-based test, both derived from the code, then certify per frame that the current candidate is the global ML solution; because both are theorems, certified frames incur no accuracy-computation trade-off. Across 15 codes and five operating points we certify up to 99.9% of frames, the two tests dividing the space as their derivations predict, one strong for sparse parity-check matrices and the other for codes with a large minimum distance. Certification also behaves as a per-frame reliability signal, with 96.7 - 100% of all frame errors falling on uncertified frames. Against an exact ML reference at 5 dB the decoder is within one standard deviation on three of the four codes where the reference is uncensored, and against four neural baselines it attains the best bit-error rate in 71 of 75 cells.
[LG-42] Foundations of the Scaling Law
链接: https://arxiv.org/abs/2610.27789
作者: Tianyu Zhao,Zhixiang Ren
类目: Machine Learning (cs.LG)
*备注: 24 pages, 11 figures, 2 tables. Technical note
Abstract:Empirical scaling laws relate validation loss to model size, training exposure and compute, providing a quantitative basis for allocating resources. This technical note explains these relationships through three linked analyses. Geometric approximation connects spatial resolution, intrinsic dimension and smoothness to function error; the loss function determines how that error transfers to task risk. Spectral analysis combines feature visibility, covariance eigenvalues and target alignment to describe how additional data reduce unresolved predictive mass. Constrained optimization then gives the parameter-data balance and the training-inference trade-off. We also examine emergence and grokking, separating scoring thresholds from learning dynamics, and test-time scaling, separating candidate coverage from selected-answer accuracy. Applications to vision and molecular prediction show how domain-specific coordinates change the resource comparison. An \mathrmSE(3) case study quantifies the dimension dependence of quotient-space approximation gains and compares optimized compute frontiers, alongside separate analyses of conformational coverage and correlated-sample precision.
[LG-43] Engram Memory Network: Brain-Inspired Prototype Explanations IJCNN2026
链接: https://arxiv.org/abs/2610.27770
作者: Hyunjun Kim,Myoung Hoon Ha
类目: Machine Learning (cs.LG)
*备注: Accepted at IJCNN 2026 (IEEE WCCI 2026). 8 pages, 3 figures, 6 tables
Abstract:Artificial intelligence systems are increasingly deployed in high-stakes decision-making. In response, prototype-based models in explainable AI have attracted growing attention for their ability to retrieve visually similar examples as faithful explanations. Although these models achieve competitive classification accuracy, they often exhibit degraded explanation quality across datasets with varying characteristics, limiting their practical applicability. We draw on insights from cognitive neuroscience, particularly the human memory system, which encodes input-relevant information into representations that support recognition. We propose an Engram Memory Network (EMN), a model that emulates the ventral visual stream and inferior temporal cortex in the human brain. Central to this model is a Hopfield-inspired memory module that iteratively retrieves representative prototypes. These memory-refined latents are used for both reconstruction and classification, encouraging the selected prototypes to capture input-relevant information across diverse datasets. To address the lack of standardized evaluation for prototype explanations, we assess three complementary aspects: perceptual similarity (LPIPS and DISTS) for perceived resemblance, subclass alignment, measuring whether prototypes match the input’s fine-grained category beyond supervised labels, and information preservation (NMI), measuring how much input information is captured by selected prototypes. Experiments show that EMN improves visual similarity by 17.06%, subclass alignment by 104.57%, and NMI by 42.42% on average across CIFAR-10/100, MNIST, HAM10000, and CUB-200, while achieving the highest average classification accuracy (87.62%) among all methods. These findings underscore the effectiveness of neuroscience-inspired explanation mechanisms in advancing the interpretability and reliability of prototype-based models.
[LG-44] Foundation-Style Pretraining for Scalable and Transferable Distributed PV Forecasting
链接: https://arxiv.org/abs/2610.27763
作者: Dayin Chen,Xiaodan Shi,Xin Xu,Zhongyue Su,Shibo Zhu,Ruizhe Deng,Dongxiao Zhang,Yuntian Chen,Jinyue Yan
类目: Machine Learning (cs.LG)
*备注:
Abstract:The rapid growth of distributed photovoltaic (PV) generation is increasing the demand for accurate forecasting across large numbers of geographically dispersed stations. Existing approaches are commonly developed for individual sites or specific forecasting horizons, requiring repeated training and maintenance as new stations are integrated and limiting direct reuse at previously unseen sites. To address this scalability gap, this paper presents a domain-specific foundation-style pretraining framework that learns a shared forecasting model from heterogeneous multi-site and multi-horizon PV data. NWP-assisted forecasting is formulated as a sequence-completion problem that unifies historical PV power, meteorological observations, future numerical weather prediction (NWP) forecasts, and time encodings. Three pretraining objectives are designed and systematically evaluated, including forecast-aligned reconstruction, generic reconstruction, and mixed reconstruction. The mixed objective combines the operational structure of PV forecasting with random sequence reconstruction and provides the most balanced performance across forecasting accuracy, degraded NWP conditions, and downstream adaptation. Experiments on 479 distributed PV stations use 340 stations for model development and 139 spatially held-out stations for zero-shot evaluation. Across horizons from 6 hours to 7 days, the multi-site pretrained MR model outperforms representative station-specific and shared baselines, as well as generalist time-series models, including a recent large-scale pretrained foundation model. The pretrained model can also be directly deployed at new PV stations and subsequently adapted to region-specific operational objectives when local data become available, providing a scalable forecasting paradigm for large distributed PV fleets.
[LG-45] AC/DC on a Budget - Alternating Sparse Phases NEURIPS2026
链接: https://arxiv.org/abs/2610.27758
作者: Rahul Nittala,Advait Gadhikar,Tom Jacobs,Rebekka Burkholz
类目: Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2026. 30 pages, 14 figures, 8 tables
Abstract:State-of-the-art neural network sparsification relies on alternating dense and sparse training phases. While computationally expensive, dense phases are critical for performance. Our empirical analysis reveals the mechanics behind this gain: upon reintroduction, previously masked parameters receive disproportionately large gradients that drive essential mask rewiring and parameter sign-flipping. Building on these insights, we optimize the performance-to-FLOPs Pareto frontier through three innovations: (1) zero-initializing masked parameters during dense phases to promote relevant sign flips, (2) shortening dense phases, and (3) replacing the dense phase entirely with a sparse rewiring phase that updates only the active mask and a small, high-gradient fraction of out-of-mask weights. Evaluated on CIFAR100 and ImageNet, our variant ARC (Alternating Rewiring Compression) matches baseline AC/DC performance at a fraction of the computational cost, establishing a superior performance-to-FLOPs frontier also compared to standard sparse-to-sparse algorithms. Experiments with GPT-2 on OpenWebText further show improvements in validation perplexity and average zero-shot performance
[LG-46] Detection of Memorization in Diffusion Models Through Thermodynamic Singularity
链接: https://arxiv.org/abs/2610.27754
作者: Iljung Kim,Kangwoo Park,Yung-Kyun Noh,Daniel D. Lee
类目: Machine Learning (cs.LG)
*备注:
Abstract:In diffusion models, proper distributional learning entails reversing a diffusion process over underlying density functions, whereas a key failure is memorization, in which the reverse process instead learns diffusion from individual training samples. We provide an analytical characterization of such memorized reverse processes, along with a criterion that detects the associated thermodynamic singularity. The proposed criterion reliably distinguishes between properly learned and memorizing networks across a range of neural network models designed to promote or mitigate memorization.
[LG-47] Constant Step Sizes for Muon-Type Spectral Optimizers on Strongly Convex Quadratics
链接: https://arxiv.org/abs/2610.27747
作者: Pratik Jawanpuria,Bhabatosh Garai,Bamdev Mishra
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Muon and related optimizers update a weight matrix with a spectrally transformed gradient. In this work, we study such spectral optimizers with a constant step size on strongly convex matrix quadratics, which isolate the effect of the update from the confounds of nonconvexity, noise, and step size schedules. We analyze an update using two quantities: the gain, which is the factor by which it amplifies the gradient along each singular direction, and the length, which is the norm of the update. These two quantities place the popular spectral updates in three regimes with different behavior: (i) with a bounded gain, the iterates behave like preconditioned gradient descent and converge for step sizes set by the largest curvature; (ii) with an unbounded gain but a vanishing length, they still converge, with the admissible step sizes set by the total curvature instead; and (iii) with a non-vanishing length, they do not converge for almost every initialization. Interestingly, the practical variants of Muon fall in all three regimes, depending only on how the update is normalized and scaled and not on the polynomial that approximates the polar factor. Experiments validate these findings.
[LG-48] EEG-Based Classification of Hypnotic Susceptibility Using Two-Dimensional Convolutional Neural Networks
链接: https://arxiv.org/abs/2610.27738
作者: Krzysztof Wrobel,Grzegorz M. Wojcik,Filip Postepski,Andrzej Kawiak,Katarzyna Zemla,Grzegorz Sedek
类目: Machine Learning (cs.LG)
*备注: 23 pages, 13 figures, 5 tables
Abstract:Guided Imagery (GI) is a mind-based relaxation technique widely used in clinical and therapeutic settings. Objectively assessing whether a patient has entered a relaxed state remains a challenge for clinicians. This study investigates electroencephalography (EEG)-based classification of hypnotic susceptibility using two-dimensional Convolutional Neural Networks (2D CNNs) applied to minimally preprocessed signals, without manual feature extraction. Data were collected from 62 male participants (32 Control, 30 Hypno) using a 256-channel dense array amplifier during Guided Imagery or cognitive task conditions. A sliding-window framework evaluated up to 868 independent models across seven anatomically defined brain regions and 124 ten-second time windows. Segment-level random-split screening was followed by Leave-One-Subject-Out (LOSO) cross-validation to obtain subject-independent performance estimates, and an ensemble approach combined predictions across K = 10 windows. The best single-window model achieved accuracy of 0.82 on the random split. Under LOSO, anterior regions generalised better across subjects: the frontal region achieved mean accuracy of 0.646 (AUC: 0.692). An ensemble of K = 10 windows further improved the prefrontal region to mean accuracy of 0.738. The occipital region, dominant in random-split screening, collapsed to near-chance under LOSO, revealing subject-specific EEG fingerprinting rather than genuine group differences. Bandpass filtering and z-score normalisation suffice for competitive EEG classification without manual feature extraction. Anterior cortical regions are identified as the most generalisable neural markers of hypnotic susceptibility.
[LG-49] Privacy Accounting for Empirical Denoisers
链接: https://arxiv.org/abs/2610.27710
作者: William Cheng,Mijung Park,Dongjun Kim
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
*备注: 31 pages, 7 figures
Abstract:Diffusion models have emerged as the foundation of modern generative systems, yet their high memorization capacity raises privacy concerns. While differentially private (DP) training provides formal guarantees, it comes at a cost in sample quality. In this work, we study a complementary question: how much privacy loss is incurred during the sampling process itself? Directly analyzing modern neural denoisers is challenging because their sensitivity to a single training example is generally intractable. To overcome this challenge, we construct and analyze a clipped privacy-accountable empirical denoiser, a training-free estimator of the Bayes-optimal denoiser, whose per-step sensitivity can be computed explicitly, allowing each denoising step to be a Gaussian mechanism. This perspective enables privacy accounting under Gaussian DP when each training example contributes to each denoising step, and under f -DP trade-off functions when each generated sample uses a single randomly subsampled block of the training data. We report every (\varepsilon, \delta) guarantee as a certified upper bound from a privacy loss distribution accountant, and use a cumulant-based Edgeworth approximation as a fast diagnostic, characterizing where it is reliable and where it becomes anti-conservative. We also analyze the privacy loss along the denoising trajectory and show that it grows steeply near the critical window in which semantic content is formed. This observation motivates hybrid samplers that use the private empirical denoiser through the formative steps where it remains a faithful proxy for the neural denoiser, then switch to a public neural denoiser for the remaining refinement. Our results on CIFAR-10 and CelebA-64 show that the privacy of diffusion models depends not only on the training procedure, but also on sampling-time design choices such as subsampling, stopping time, and hybrid denoising.
[LG-50] Multi-Objective Evolutionary Semi-Supervised Learning for Resource-Aware Smart City Classification
链接: https://arxiv.org/abs/2610.27703
作者: Francisco José Sedeño,Jamal Toutouh
类目: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
*备注: ICSC-CITIES 2026. IX Ibero-American Congress of Smart Cities
Abstract:Smart-city applications often combine scarce labeled data with abundant unlabeled observations. Semi-supervised learning (SSL) exploits both labeled and unlabeled data to improve classification under limited supervision. This article presents MOEA-SSL, a multi-objective evolutionary approach for lightweight semi-supervised tabular classification that jointly optimizes predictive quality and feature usage. Experimental evaluation on three Smart City classification datasets, considering 1%, 5%, and 10% labeled data, shows that MOEA-SSL achieves competitive predictive performance compared with established lightweight SSL methods while significantly reducing feature usage relative to EA-SSL in all evaluated conditions. The computed Pareto fronts also provide lower-feature-use alternatives with limited degradation in classification performance.
[LG-51] Understanding Data Augmentation under Label Noise
链接: https://arxiv.org/abs/2610.27701
作者: Hong Zheng,Fei Teng
类目: Machine Learning (cs.LG)
*备注:
Abstract:Data augmentation (DA) has become a fundamental component of modern machine learning systems and is widely regarded as an effective strategy for improving robustness and generalization. Despite its empirical success, the theoretical impact of DA under label noise remains poorly understood. In this work, we first theoretically explain when DA is beneficial and when it becomes detrimental, and provide sufficient conditions for distinguishing between these two regimes under both the i.i.d. learning setting and the more challenging domain generalization scenario. Then, we reveal a counterintuitive phenomenon: DA essentially acts as an uncertainty amplifier in the learning process, decreasing model stability and generalization performance. This finding is supported by the proposed learning and generalization bounds. Our analyses do not depend on specific DA techniques and require only mild assumptions, indicating that the results are applicable to a broad range of machine learning scenarios. Namely, we provide a unified theoretical foundation for understanding, analyzing, and designing DA methods in the presence of label noise.
[LG-52] Influence Is the Target-Response Jacobian Times Residuals
链接: https://arxiv.org/abs/2610.27689
作者: Paul Geertsema,Simon Geertsema
类目: Machine Learning (cs.LG)
*备注: 19 pages, 2 figures, 8 tables
Abstract:We introduce the target-response Jacobian J = [\partial \hat y_i/\partial y_j] , which measures how model predictions respond to changes in individual training targets. J complements training-data attribution (TDA) methods, which characterize sensitivity to the weights placed on training observations in the loss rather than to their targets, and can be computed matrix-free at the same cost as influence-based TDA methods. We show that sample-weight influence \mathcalI and J follow from the same implicit differentiation argument and differ only by a per-example multiplier: for pointwise losses, \mathcalI = J\operatornamediag(m) exactly at any stationary point with an invertible Hessian. For canonical-link losses (squared error, logistic, Poisson) the multiplier is the training residual, \mathcalI = J\operatornamediag® , and softmax cross-entropy obeys a multiclass version. The identity predicts three ways in which influence fails when used in place of J : small residuals shrink its predicted shift towards zero (interpolation collapse), residuals of mixed sign wash out its direction (direction washout), and under softmax cross-entropy it collapses a training point’s per-class sensitivities towards its own true class (true-class collapse). We score both operators against refitting with a per-prediction R^2 across target perturbations, with no rescaling. On tanh networks trained with squared error (Friedman #1) and logistic loss (MAGIC, HIGGS), with N from 2,000 to 10^6 , J reaches R^2 between 0.99 and 1.00, whereas influence’s R^2 is negative on every dataset. On an MNIST softmax classifier, influence’s R^2 is again negative, while J reaches 0.92. On HIGGS, one row of J , holding all 10^6 sensitivities of one prediction, takes under 0.2 seconds on an RTX 3090 after a one-time setup of under five minutes, whereas refitting would take 10^6 refits.
[LG-53] CyberHull: Deterministic Malicious URL Isolation via Convex Hull
链接: https://arxiv.org/abs/2610.27675
作者: Avijit Gayen,Aniket Ghosal,Durba Chakraborty,Sayan Mondal,Md Abdul Aziz Al Aman,Angshuman Jana
类目: Cryptography and Security (cs.CR); Computational Geometry (cs.CG); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注:
Abstract:The emergence of increasingly sophisticated and evasion-resistant phishing attack strategies represents a significant threat to modern cybersecurity, particularly in cyber-physical systems and internet-based networks. Traditional probabilistic methods and certain machine learning approaches often struggle to detect novel attacks and may lack interpretability or portability. To address these challenges, we introduce CyberHull, a resource-efficient and deterministic framework for identifying malicious URLs through mathematically defined convex hull boundaries across multiple orthogonal feature domains, including structural content entropy, hyperlink graph topology, and semantic token distribution. CyberHull constructs compact convex hulls, referred to as vaults, using the QuickHull algorithm, with a training complexity of O(n log n) under suitable assumptions. URL classification is performed through point-in-polyhedron tests that impose relatively low computational overhead on standard CPUs. Experimental results demonstrate that CyberHull achieves 92 percent detection accuracy, substantially outperforming a baseline linear approach based on Principal Component Analysis (PCA), which achieves 32 percent accuracy. Furthermore, CyberHull offers a computationally efficient alternative to resource-intensive deep learning approaches for malicious URL detection.
[LG-54] Hybrid Momentum Methods for Optimizing Deep Learning Models
链接: https://arxiv.org/abs/2610.27674
作者: Haifeng Xu,Yijin Ren,Qi Deng,Chenyu Zhou,Dongdong Ge,Yinyu Ye
类目: Machine Learning (cs.LG)
*备注: 39 pages, 17 figures, 9 tables
Abstract:Momentum is central to stochastic optimization in deep learning, including the training of large language models. Classical momentum methods typically rely on a fast exponential moving average (EMA) scheme, which discounts past gradients on a fixed timescale and may underweight older gradients that remain informative. This creates a trade-off between retaining long-term history and adapting to recent changes. We introduce Hybrid Momentum (HyM), which combines a linearly weighted average of past gradients with a geometrically discounted fast momentum. The former assigns weights proportional to the gradients’ iteration indices and retains a growing span of history, while the latter emphasizes recent gradients. This hybrid construction is modular and can be combined with adaptive preconditioning and other update transformations. In particular, we propose HyMUON, which applies Newton–Schulz orthogonalization to the hybrid direction. Illustrative examples and controlled experiments demonstrate the benefits of combining gradient information across these two timescales. Across image classification, machine translation, and language modeling, HyM and HyMUON match or improve upon strong baselines. HyMUON improves over Muon on IWSLT14, and in both the 100k- and 600k-step GPT-2 124M OpenWebText regimes. For LLaMA-1.3B pretraining on C4, HyMUON achieves a final validation loss of 2.777, compared with 2.903 for Muon.
[LG-55] Neural Operator Splitting for Fluid Dynamics in Complex Geometry
链接: https://arxiv.org/abs/2610.27671
作者: Jianxiang Li,Xi Deng,Hong Chul Nam,Anima Anandkumar
类目: Machine Learning (cs.LG); Fluid Dynamics (physics.flu-dyn)
*备注:
Abstract:Solving dynamical systems in complex and diverse geometries remains challenging for both numerical solvers and learning-based surrogate models. We propose Neural Operator Splitting, a framework for learning fluid dynamics while preserving the physical structure of fractional-step numerical schemes. Instead of learning the full-step evolution monolithically, we decompose each time step into physically interpretable steps, and learn each operator separately using intermediate states from the underlying numerical solver. Guided by the linearity properties of the fractional steps, we introduce a conditional-linear Fourier neural operator (CL-FNO) whose linear state mappings are modulated by nonlinear conditioning on inputs such as boundary geometry. We further introduce two complementary strategies, symmetry sampling and corrective cycling, to mitigate error accumulation during autoregressive rollout. Experiments on 2D and 3D fluid flows with varying geometries and initial conditions show that our full model reduces mean relative L^2 rollout error by 68% in 2D and 40% in 3D compared with a single full-step neural operator. At comparable error levels, neural-operator inference is approximately 10^3 times faster per step than the Monte Carlo solver.
[LG-56] Ensemble-Aware Structural Encoding (EASE): Conditioning Protein Encoders on Peptide Conformational Ensembles
链接: https://arxiv.org/abs/2610.27638
作者: Ruoyao Zhang,Hongxuan Li,Jie Peng,Tianlong Chen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Protein foundation models provide transferable representations for downstream peptide prediction tasks. However, existing representations typically reduce a peptide to a single sequence or structure, which may miss conformational variability relevant to key peptide properties, including antibacterial activity, hemolysis, toxicity, stability, and binding. For example, protease resistance correlates more strongly with conformational dynamics measured by hydrogen exchange than with average helicity. Therefore, we introduce Ensemble-Aware Structural Encoding (EASE), a framework for effectively extracting structural information from conformational ensembles for peptide representation. To extract useful information from noisy conformers, EASE selects central conformers representing a peptide’s major structural states and fuses their individual representations with a permutation-invariant operator to produce a peptide-level representation for downstream prediction. We then evaluate our framework on six peptide property-prediction tasks under leakage-controlled splits. Against matched single-structure baselines, EASE improves performance consistently on six tasks with a mean absolute gain of about 3.0% across two structural encoders. Gains keep growing as we add conformers through K=20. A single peptide conformer therefore provides a limited basis for structural representation, whereas our framework distilling and fusing multiple conformers yields a more informative representation.
[LG-57] From Outcomes to Steps: Reliable Success-Aware Critic Learning for Long-Horizon Language Agents
链接: https://arxiv.org/abs/2610.27623
作者: Jiaxi Liu,Shixiong Kai,Zhentao Tang,Yang Li,Yuqi Cui,Mingxuan Yuan,Junchi Yan
类目: Machine Learning (cs.LG)
*备注: 29 pages, 6 figures, 5 tables
Abstract:Reinforcement learning for long-horizon language agents requires identifying whether intermediate actions make eventual task success more or less likely when supervision is available only at trajectory level. The main difficulty is that a critic with aggregate return-prediction accuracy may still fail to resolve the local value differences needed to assign credit to actions, especially under sparse feedback and function approximation. Conventional critics estimate future return, but policy improvement also depends on knowing whether each intermediate decision makes task success more or less likely. We introduce SCALE, an actor-critic framework that augments conventional return estimation with explicit prediction of eventual task success. At each interaction prefix, the critic predicts a return value and a Success Value, defined as the probability that the task will eventually be completed from that prefix. SCALE converts changes in Success Value into multi-step action credit through Success-GAE and modulates its contribution according to the estimated reliability of the predictor, while retaining the token-level PPO objective. Across five agent benchmarks and three model backbones, SCALE improves task performance over the compared RL baselines in most settings, with gains of up to 4.1 percentage points in success rate. Separate critic-learning analyses show that SCALE also improves critic estimation and training stability.
[LG-58] Sensitivity-Orthogonal Residual Learning for Adaptive Sideslip and Cornering Stiffness Estimation
链接: https://arxiv.org/abs/2610.27586
作者: Kyoungtae Ji,Dongyeol Won,Kyoungseok Han
类目: Robotics (cs.RO); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 25 pages, 15 figures, 5 tables. Submitted to Vehicle System Dynamics
Abstract:Online adaptation of the cornering stiffness keeps a model-based sideslip estimator valid as tires change, but competes with learned model correction for the same filter innovation. A corrector with enough parameters can absorb the innovation pattern that a stiffness change produces, and adaptation is suppressed while the fit improves. We propagate the stiffness sensitivities of the predicted measurements through the vehicle states and construct a corrector orthogonal to them. A null-space parameterization enforces the constraint without a penalty coefficient and preserves the static least-squares stiffness estimate on each identification segment. In full-vehicle simulation, the stiffness response of the unconstrained corrector declines as its parameter count grows, whereas the orthogonal corrector responds at least as fully as the uncorrected filter at all sizes and lowers sideslip error by 12.7% to 22.0%. In a simulated drive with a stiffness step, the orthogonal corrector adapts as quickly as the uncorrected filter and lowers sideslip error during and after the transient. On public real-vehicle data, the 430-parameter unconstrained corrector loses most of its response to deliberate offsets of the filter’s nominal stiffness, whereas the orthogonal corrector of the same size retains most of it, with a sideslip error slightly above the uncorrected filter’s.
[LG-59] Bridging Digital Deception and Physical Fidelity: A Spatio-Temporal Attack on Cognitive Radars
链接: https://arxiv.org/abs/2610.27585
作者: Tongrun Xing,Yongjun Dai,Xuan Liao,Zhiwen Zhang,Tong Zhou,Longlong Zhang,Tian Jiang
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
*备注: Submitted to IEEE Transactions on Information Forensics and Security. 16 pages, 13 figures
Abstract:As cognitive radars evolve towards the end-to-end processing of one-dimensional complex baseband sequences, deep networks face severe adversarial threats. However, existing attacks encounter dual limitations in practical electronic warfare: low cross-architecture transferability and severe vulnerability to physical channel distortions, which degrades their tactical reliability. To overcome these bottlenecks, we propose a physical prior-driven Segmented Spatio-Temporal Attack (SSTA). First, a global-local feature decoupling segmentation mechanism interrupts the adversarial gradients’ dependence on long-range sequence coherence, bridging the cross-architecture feature gap. Second, moving beyond static norm constraints devoid of physical semantics, we abstract channel distortions into differentiable spatio-temporal operators, integrating them into momentum iterations via an Expectation over Transformation (EoT) strategy. Comprehensive cross-evaluations across 9 heterogeneous models spanning CNNs, RNNs, and Transformers demonstrate that SSTA outperforms the state-of-the-art baseline by 10.36% in Average Black-Box Attack Success Rate. Furthermore, baseband equivalent simulations confirm that the generated waveforms survive extreme physical distortions and, after matched filtering, achieve zero main-lobe attenuation (0 dB) with noise floor lift strictly within Constant False Alarm Rate tolerances. SSTA successfully unifies digital-domain deception and physical-domain fidelity, providing crucial support for CEW security assessment.
[LG-60] From POMDP Theory to Deep RL with Particle Filters NEURIPS
链接: https://arxiv.org/abs/2610.27575
作者: Kevin Tan,Ziping Xu
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: in NeurIPS
Abstract:Model-based reinforcement learning in partially observable environments requires inferring hidden states, learning latent dynamics, and planning under the resulting belief. Practical algorithms often solve one aspect of the above pipeline while relying on approximations for the others that are often poorly understood, while existing POMDP theory typically analyzes computationally inefficient theoretical algorithms far from modern deep RL practice. To fill this gap, we propose a practical end-to-end algorithm that approximately solves the hidden state filtering problem via sequential Monte Carlo (SMC), updates a latent dynamics model online, and performs approximate online planning for d steps before bootstrapping leaves with a state-value critic (a QMDP approximation). We analyze each approximation and prove an end-to-end regret bound decomposing the error into model estimation error, particle filter error, Monte Carlo rollout error, and a structural bias term induced by the QMDP approximation controllable by the planning depth that we further analyze in various special cases. The resulting guarantee makes explicit how statistical sample size, particle budget, rollout budget, and planning depth affect performance. We instantiate our algorithm in an MPC-based deep RL framework, demonstrating its practicality on a variety of environments. On a POMDP that satisfies our assumptions, each term of the bound decays at its predicted rate. Our results provide a practical and theoretically grounded route for bridging theory and practice within online decision-making in POMDPs.
[LG-61] wo Is Enough: Binary Classification Priors Suffice for Multi-Class In-Context Tabular Prediction
链接: https://arxiv.org/abs/2610.27569
作者: Zhixiong Cheng,Shiyuan Han,Jingjing Suo,Zhengli Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Tabular foundation models (TFMs) pretrain a transformer on large-scale synthetic priors so that a single forward pass yields zero-shot tabular prediction. TabICLv2 samples the class count of its classification prior during pretraining, and the prevailing configuration caps it at ten. We report a counter-intuitive result: setting this cap directly to two—making every pretrained classification task binary—still yields weights that perform zero-shot K-way classification 2. On the 38 TabArena classification datasets (200 bootstrap rounds, anchored at default RF = 1000), the binary two-stage recipe reaches 1590.1 Elo, above the official default at 1568.0 (+22.1), above TabPFN-v2.6 (1577.1) and RealTabPFN-v2.5 (1521.6). A production-scale validation on 60 real-world credit-scoring institutions independently confirms the direction: the same checkpoint improves test AUC on 44 of 60 institutions (paired Wilcoxon p = 0.0007 for AUC, p = 0.0002 for KS) at +1.5 inference cost. Mechanistically, classification in TabICLv2 is carried not by a fixed-cardinality output head but by an in-context comparison operator matching query-row embeddings against class-conditional embeddings, so the cardinality of the output space is set by the labels present in the context, not by the label space seen during pretraining. This makes a binary prior structurally cheaper to generate and cache than a ten-class prior, demotes “number of classes” from an architectural constraint to a tunable hyperparameter, and points to a class-count-agnostic regime of TFM pretraining.
[LG-62] Lower Bounds for Smooth Nonconvex Zeroth-Order Optimization: The Price of Determinism under Low Curvature Trace
链接: https://arxiv.org/abs/2610.27550
作者: Yuxing Peng,Zhiqing Tang,Weijia Jia
类目: Machine Learning (cs.LG)
*备注: 59 pages, 2 figures, 1 table; Lean 4 formalization of the deterministic bounds in the ancillary files
Abstract:Low curvature trace, an effective rank condition, is a common explanation for why randomized zeroth-order methods scale to high-dimensional tasks such as language model fine-tuning. Can deterministic methods achieve the same savings? We show that they cannot. We study finding \varepsilon -stationary points of L -smooth nonconvex functions on \mathbbR^d from exact function values, assuming that below the starting value the Hessian is dominated by a fixed matrix with trace at most LE ; this is the effective rank condition used to explain zeroth-order fine-tuning. When ET\le d , the Gaussian two-point method that such fine-tuning uses needs only O(ET) calls if it rejects steps that raise f , where T=L\Delta/\varepsilon^2 is the iteration count of gradient descent from initial gap \Delta , and we prove that every randomized method needs \Omega(ET) calls, so this cost is optimal up to constants. In contrast, every deterministic method needs \widetilde\Omega(d\min\T,\sqrtET) calls, and adaptive subspace search attains O(d\min\T,\sqrtET) : a deterministic method pays about d calls per discovered direction, and the trace budget allows about \sqrtET directions. Determinism thus multiplies the cost by a factor of order d/(E+\sqrtET) up to logarithmic factors, which grows linearly with d for fixed E and T ; once E exceeds T by a logarithmic factor, deterministic methods need \Omega(dT) calls, as without any trace budget. Our lower bounds hold for C^\infty functions with the matrix bound on all of \mathbbR^d , even when each call also returns gradient projections priced by rank.
[LG-63] ming-Aware and Interpretable Self-Attention for Event-Based Wearable Task Transition Detection ALT
链接: https://arxiv.org/abs/2610.27549
作者: Anda Ouyang,Siyuan Chen,Julien Epps
类目: Machine Learning (cs.LG)
*备注: Submitted to IEEE Journal of Biomedical and Health Informatics (JBHI) on 26 February 2026; currently under review
Abstract:Task transition detection aims to identify when a person shifts from one task to another, providing a foundation for continuous health monitoring and timely context-aware interventions. While statistical and deep feature representations have demonstrated strong classification performance, they often rely on aggregated or latent features that are difficult to interpret in behavioural terms. Event-based features provide a compact and interpretable alternative by encoding behavioural and physiological patterns as discrete tokens with explicit semantic meaning, making them well suited for behavioural analysis in health informatics. Self-attention offers a mechanism for modelling interactions among such tokens; however, conventional self-attention classifiers primarily rely on token order and do not explicitly incorporate timing or duration information. As a result, informative temporal cues in event representations may not be fully exploited for task transition detection. This study proposes an event-based time-aware self-attention framework that represents behavioural signals as sequences of discrete events augmented with onset timestamps and durations. Event timing is encoded using a parameter-efficient kernel-based representation, while a novel type-aware duration embedding modulates information propagation according to event persistence. Experiments on a semi-naturalistic wearable dataset show that the proposed model outperforms existing event-based approaches, achieves performance beyond or comparable with statistical and deep-learning models with a more compact input representation, and provides improved interpretability through integrated-gradient-based analysis. These results indicate that explicit modelling of event timing and duration enhances both classification performance and interpretability, supporting reliable deployment in biomedical and health informatics applications.
[LG-64] BEACON: Bridging Multimodal Semantics and Asynchronous Propagation for Cloud System Anomaly Detection ICDE2027
链接: https://arxiv.org/abs/2610.27542
作者: Zhe Li,Tengxue Zhang,Yang Shu,Tung Kieu,Lunting Fan,Yingying Zhang,Bin Yang
类目: Machine Learning (cs.LG)
*备注: Accepted by ICDE 2027
Abstract:Modern large-scale cloud systems continuously generate massive and multimodal observability data, such as time-series metrics, log records, and system dependency graphs, which reflect the health state of the system. To ensure reliable cloud system operation, anomaly detection models identify abnormal states by analyzing these multimodal observations. However, existing approaches rely on simple fusion strategies and graph learning, ignoring the semantic gap among modalities and the asynchronous spatio-temporal dependencies across system instances over time. To this end, we propose BEACON, a framework to Bridge multimodal sEmantics and Asynchronous propagation for Cloud system anOmaly detectioN. Specifically, we introduce a Dynamic Interpretable Multimodal Aggregator (DIMA) that leverages an information-bottleneck mechanism to compress information from the semantics and temporal dynamics of metrics, and aligns them with logs in a shared prototype space. Furthermore, we propose an Asynchronous Spatio-Temporal Propagator (ASTP) that captures temporal dependencies with multi-scale temporal attention and models spatial relations with topology-aware aggregation. This enables asynchronous propagation of anomalies across system instances. Extensive experiments on three real-world cloud system datasets demonstrate that \textbfBEACON consistently outperforms state-of-the-art baselines in anomaly detection.
[LG-65] Soliton Profiles in 1D: Enhanced Neural-Network Solvers and Constrained Deep Ritz Methods
链接: https://arxiv.org/abs/2610.27500
作者: Chandler Haight,Alex D Rodriguez,Svetlana Roudenko
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG); Analysis of PDEs (math.AP); Pattern Formation and Solitons (nlin.PS)
*备注: 9 figures
Abstract:We develop and compare enhanced neural-network solvers for one-dimensional ground-state profiles, which are solutions of Q^\prime\prime=bQ-\gamma Q^p,\quad Q0,\quad p1, \quad b, \gamma 0, the stationary profile equation for the nonlinear Schrödinger, generalized Korteweg-de Vries, and nonlinear Klein-Gordon equations. Its explicit sech-type solutions serve as exact reference for error computation and benchmark hyperparameter selection. We improve on the standard Adam-trained physics-informed neural network (PINN) baseline by introducing higher-order optimization via Gauss-Newton/Levenberg-Marquardt and L-BFGS refinements, as well as Fourier-feature PINNs, and constrained Deep Ritz (DR) formulations. In DR approaches we introduce two new neural-network formulations, which implement two classical variational characterizations: a Nehari-constrained Deep Ritz method, in which the height is learned by projecting onto the Nehari/Pokhozhaev manifold, and a Weinstein Deep Ritz method, in which a normalized profile minimizes the Weinstein quotient. The Deep Ritz formulations perform best overall, in particular, the Weinstein variant gives the best error-runtime balance. The neural-network based solvers remain less accurate and efficient than classical Petviashvili iteration in 1D, however, the current study provides a proof of concept: the mesh-free formulation, non-reliance on exact profiles or the related features make them promising in higher dimensions and for more complicated nonlinearities and dispersive operators.
[LG-66] MiCLA: Mixture-to-Constituent Latent Alignment for Polyphonic Audio Representation Learning
链接: https://arxiv.org/abs/2610.27491
作者: Kohei Yamamoto,Kosuke Okusa
类目: ound (cs.SD); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 13 pages, 3 figures
Abstract:Real-world audio typically contains concurrent sound events, yet standard self-supervised methods treat each clip as an isolated training unit without explicitly modeling mixture–constituent relationships. We propose Mixture-to-Constituent Latent Alignment (MiCLA), a two-stage self-supervised pre-training framework that uses synthetic log-mel mixing to align mixture representations with their constituent clips. Stage 1 structures single-clip representations via masked self-distillation and contrastive regularization. Stage 2 introduces bidirectional alignment: for a synthetic mixture, MiCLA suppresses the component aligned with one constituent and aligns the residual with the other, using projective residualization on patch tokens and sample-adaptive soft residualization on the global class token. In downstream evaluations, MiCLA achieves 43.2% mean average precision (mAP) on AS-20K and 50.3% mAP on AS-2M under fine-tuning, and achieves the highest mAP among evaluated models on both AudioSet settings under frozen Convex Gated Probing (CGP). Furthermore, a Joint Semantic Retrieval diagnostic shows that MiCLA improves simultaneous retrieval of both constituent classes over target-averaging baselines. These results indicate that directional mixture-to-constituent alignment provides an effective pre-training signal for polyphonic audio representation learning.
[LG-67] When Parameter Gradient Flows Close in Representation Space
链接: https://arxiv.org/abs/2610.27474
作者: Shengye Tao
类目: Machine Learning (cs.LG)
*备注: 34 pages, 2 figures; code and numerical data included as ancillary files
Abstract:The represented gradient flow of a parameterized model depends on a mobility that can retain hidden parameter information. We study when one entire represented mobility map works for every autonomous loss from one fixed initialization. For two-factor matrix representations wider than both output dimensions, with individually full-row-rank factors, we classify all admissible joint ranks: closure occurs precisely at effective scalar-balanced minimal joint rank or a two-value signed spectral family at full joint rank. All intermediate ranks are excluded; positive represented flows belong to known square-root families. For general representations with constant ambient representation and gradient-Lie ranks, ordinary closure is equivalent to a local invariant-level metric in the analytic case and specified smooth regimes. An endpoint section forces the initially unrestricted mobility map to admit a locally smooth representative on the represented image. We also separate testing families: linear losses locally admit closure at every submersion, whereas one fixed translated strongly convex quadratic family fails it on a delayed-bracket example. A fixed output readout gives an ordinary squared-loss witness. These supporting results do not characterize quadratic-only closure in general. We retain the unresolved smooth output-rank-two case with Lie rank at least four. A frozen deterministic experiment checks a separate engineered-loss construction; it is distinguished from the squared-loss proof, universal theorems and empirical learning performance.
[LG-68] Decision Sufficiency of Learned Representation Interfaces for Safe Generative Control
链接: https://arxiv.org/abs/2610.27446
作者: Jipeng Li,Yanning Shen
类目: Machine Learning (cs.LG)
*备注:
Abstract:A controller can guide a language model toward safe and useful responses by choosing when to assist, ask for clarification, or refuse. When the controller uses selected internal features, poor performance alone cannot distinguish missing information from limitations of its decision rules or training. We study decision sufficiency: whether the best controller using a representation can match the minimum average loss attainable with full state information, under the same strategies and evaluation criteria. For interfaces with finitely many outputs, we compute the optimal loss exactly and separate the three sources of excess loss. Pairwise checks can miss information loss: states with the same controller input may share an optimal strategy in every pair but not across the whole group. We bound the number of states needed to reveal a conflict and characterize the fewest outputs an unrestricted interface needs to avoid all conflicts. We audit selected features from sparse autoencoders, which encode model activations using relatively few active features, in Qwen2.5, SmolLM2, and Phi. On SmolLM2 responses scored by two human raters, restricting the controller to two selected on/off features increases its minimum average loss by 0.0345–0.1166 relative to full state information, even with the governing policy supplied. Both independently trained autoencoders show the increase on both evaluation sets. The audit identifies decision failures that require changing the controller’s input, not merely its architecture or training.
[LG-69] Learning-Theoretic Foundation for General Coded Computing: The Adversarial Setting
链接: https://arxiv.org/abs/2610.27440
作者: Parsa Moradi,Hanzaleh Akbarinodehi,Mohammad Ali Maddah-Ali
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注:
Abstract:Adversarial workers pose a severe challenge in distributed computing systems by returning corrupted outputs that can substantially distort the recovered computations. Coded computing addresses this threat by introducing redundancy across worker assignments. Existing exact and approximate coded-computing methods typically handle adversarial outputs by explicitly locating and correcting or removing erroneous evaluations, which adds significant computational overhead at the master node. Moreover, existing analyses do not characterize how the recovery error scales with the number of workers and the adversarial budget. In this paper, we develop an adversarially robust extension of general coded computing (GCC), a learning-theoretic approximate coded-computing framework that designs the encoder and decoder by minimizing an upper bound on the end-to-end recovery error. Rather than identifying corrupted workers, the proposed method uses smoothness regularization to limit the influence of their outputs. The resulting encoder and decoder can be implemented using smoothing splines with linear master-side complexity. We derive a non-asymptotic bound on the worst-case end-to-end approximation error by decomposing it into three components: (1) the encoder approximation error, (2) the decoder reconstruction error under fully honest results, and (3) the adversarial perturbation error. For maximum of q\geq 1 adversarial workers among N workers results in a worst-case error of order \mathcalO((q/N)^6/5) with appropriate selection of smoothness parameter, demonstrating that the approximation error vanishes asymptotically whenever the number of adversaries grows sublinearly with the network size. Experiments on scalar computation and neural-network inference are consistent with the predicted convergence behavior. Subjects: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2610.27440 [cs.LG] (or arXiv:2610.27440v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.27440 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-70] Fixed-Size Active Statistical Inference NEURIPS2026
链接: https://arxiv.org/abs/2610.27389
作者: Erik Skalnes,Michael Oberst
类目: Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2026 (poster). 49 pages. Code: this https URL ; Python package: this https URL
Abstract:Variance reduction techniques improve the efficiency of drawing conclusions from data. Recent proposals for variance-reduction use two complementary techniques to reduce variance: incorporating machine learning predictions into estimators and weighted sampling. A popular weighted sampling algorithm is Poisson Sampling (PS), which independently flips a biased coin at each unit and includes all heads in the sample. A fundamental flaw with PS is that it has a random sample size, which presents practical challenges and introduces variance into the estimator. We show that a fixed-size alternative, known in survey sampling as Conditional Poisson Sampling (CPS), is a drop-in replacement for PS on general multivariate Z-estimation problems, which include convex M-estimation considered in prior work. Notably, we show that the asymptotic variance of CPS is no worse than that of PS, and can be substantially better in realistic scenarios where predictions are imperfect. We illustrate these advantages on both synthetic and real datasets, and release a Python package implementing CPS. In light of our contributions, there is no need to suffer a variable sample size.
[LG-71] raining Controllable World Models via Action Influence Maximization
链接: https://arxiv.org/abs/2610.27387
作者: Samuel M. Bateman,Tenny Yin,Chongyi Zheng,Lei Huang,Brian Wang,Benjamin Eysenbach,Jaime Fernández Fisac,Dhruv Shah
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: 31 pages, 13 figures, CoRL 2026
Abstract:Action-conditioned video generative models trained on successful, goal-directed demonstrations tend to ignore the commanded action, rendering a plausible expert-like future regardless of what is asked of them. We trace this failure to training data that lack sufficient action influence. We analyze this metric and prove that it upper bounds the action influence that any model trained on that data can express. Expert demonstrations, whose narrow action distributions let next states be predicted without attending to the action, set a low ceiling that no amount of model capacity alone can raise. Raising it requires collecting different data. We introduce Action Influence Maximization (AIM), an algorithm for actively collecting training data for improving dataset action influence. We compare the performance and data-efficiency of AIM against exploration heuristics and intrinsic-motivation methods on simulated navigation and manipulation tasks. In pixel-based manipulation, AIM is the only method that consistently improves controllability: 500 actively collected episodes added to 3k demonstrations match a world model trained on 300k expert episodes on end-effector controllability, an 85\times reduction in data.
[LG-72] Sample-efficient learning with Dense Associative Memories
链接: https://arxiv.org/abs/2610.27384
作者: Andrew Lucas,Dmitry Krotov
类目: Machine Learning (cs.LG)
*备注:
Abstract:We analytically study the evolution of the energy function of Dense Associative Memories (DenseAMs) as they train on randomly sampled codewords of random expander low-density parity-check codes (a.k.a. solutions of random XORSAT problems). We identify two distinct regimes and a transition between them, reminiscent of grokking in neural networks. A regime of small training set size is where the model memorizes a few training examples but cannot discover the underlying rules of the entire codeword dictionary generative process. The opposite regime, corresponding to a larger training set size, is where the model has uncovered the sparse constraints that generated the entire data set after seeing only polynomially many patterns. We also design a novel class of DenseAMs with a thresholding learning rule, and show that any amount of thresholding enables the model to perform simplified sparse Hamiltonian learning, which exponentially reduces the number of training examples needed to learn the true underlying data set. In this regime, the model learns how to discover \exp[\mathrmO(n)] solutions from only \mathrmO(\log n) training samples, which is asymptotically optimal for any learning algorithm. When training on codewords of certain random codes, the model can even learn the underlying code when a small O(1) fraction of bits are latent variables. These latent variables are not shown to the model during learning, but will be discovered by sparsifying observed higher-order correlations. For certain random codes, we show that a simple and conservative algorithm learns a minimal latent representation for the data set. This provides a transparent, physics-motivated, and analytically tractable toy model for how deep latent representations emerge in an associative memory training.
[LG-73] Shaped SAEs: Expanding the Toolkit for Mechanistic Interpretability
链接: https://arxiv.org/abs/2610.27375
作者: Oveys Delafrooz Noroozi,Bhagyashree Puranik,Upamanyu Madhow
类目: Machine Learning (cs.LG)
*备注:
Abstract:Sparse autoencoders (SAEs) seek to decompose the dense activations of large language models (LLMs) into sparse, potentially interpretable features using a learned overcomplete dictionary. In this work, we ask whether it is possible to enhance the semantic selectivity of the learned dictionary atoms by linearly \emphshaping the dense activations prior to learning an SAE. We explore rank-reducing transformations based on the recently introduced J-space corresponding to the averaged Jacobian from the current layer to the LLM output, projecting activations onto the ``live’’ subspace spanned by eigenmodes of the Jacobian with larger singular values (which we expect to be more semantically relevant). We also introduce a new approach for training SAEs in which (a)~an affine shaping transformation is learned along with the dictionary atoms, (b)~the atom selection mechanism is based on the TopK projections along the one-dimensional subspace spanned by the atoms, and ©~reconstruction is carried out using the same atoms. We call this approach a Matched Filter SAE (MF-SAE) based on analogies with communication theory. The promise of shaping is demonstrated via evaluations against dense activations and state-of-the-art SAEs on benchmarks including SAEBench and SAE Probes.
[LG-74] Robust Lightweight IoT Device Fingerprinting Under Network Traffic Anomalies
链接: https://arxiv.org/abs/2610.27351
作者: Roya Taheri,Shamik Sengupta
类目: Networking and Internet Architecture (cs.NI); Machine Learning (cs.LG)
*备注: 6 pages
Abstract:Accurate IoT device identification is essential for network security, yet many existing approaches rely on payload inspection, feature engineering, or long observation windows that limit scalability and often assume stable network conditions despite the inherent variability of real-world IoT traffic. We propose a lightweight IoT device identification approach based on packet headers and locality-sensitive hashing (LSH), generating compact fingerprint representations at packet- and window-level granularities. To evaluate robustness, we consider packet-level traffic anomalies including drop, perturbation, reordering, duplication, and packet injection. Experimental results show that the proposed approach achieves up to 99% accuracy while maintaining robustness under realistic traffic distortions using short bounded observation windows. Anomalies affecting global traffic structure produce the largest performance degradation, whereas window-level representations improve robustness by reducing sensitivity to localized distortions. The low computational overhead of the proposed approach enables scalable IoT device identification in resource-constrained environments.
[LG-75] Automated Brick Schema Knowledge Graph Completion Across Buildings
链接: https://arxiv.org/abs/2610.27339
作者: Daniel Ma,Jalil Huseynov,Xinyi Pan,Shaked Shalev,Zhang Chi,Ai Yamamoto,Sohei Arisaka
类目: Machine Learning (cs.LG)
*备注: 9 pages, 2 figures
Abstract:The Brick schema represents building equipment, sensors, spaces, and the relationships between them as typed, multi-relational graphs, providing machine-readable metadata that supports generalizable control applications. Current methods for producing complete Brick models require extensive manual examination of building schematics and point lists. We observe that as Brick models are knowledge graphs, inferring missing relations is naturally posed as a knowledge graph completion problem. To the best of our knowledge, this is the first application of knowledge graph completion to Brick model generation. We compare three methods, gradient-boosted trees, Pairwise Brick Relation Ranking, and CompGCN, on the relations feeds, hasPart, and hasPoint, which account for the majority of the schematic-reading work in building a Brick model by hand. Models are evaluated on held-out MORTAR buildings and out-of-distribution Kaggle and Building Time Series data. Although CompGCN is the strongest performer on held-out buildings under metrics resistant to class imbalance, leading macro-F1 and AUPRC, no method generalizes uniformly across datasets or relations.
[LG-76] VampPrior Unleashed
链接: https://arxiv.org/abs/2610.27322
作者: Andrew A. Stirn,Abhejit Rajagopal,David A. Knowles,Grace Huynh
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Deep latent variable models (DLVMs) underpin many modern analysis pipelines, for example in single-cell genomics, yet their prior distributions have received markedly little attention. The VampPrior approximates the optimal prior (the aggregate posterior) and often attains high model evidence. Here, we identify two key limitations of the VampPrior: it ties prior and posterior uncertainty to a single amortized covariance and clusters poorly due to its fixed uniform mixture weights. We unleash the VampPrior by decoupling its prior and posterior covariances, resolving the conflict of generating high prior and low posterior entropy with the same shared parameters and substantially improving model evidence. We then replace the uniform mixture with a Dirichlet process. When combined, our proposals provide higher model evidence than the VampPrior and robust non-parametric clustering. Augmenting single cell Variational Inference (scVI) with our prior improves single-cell integration and clustering across six atlases while significantly raising model evidence. Our approach changes only the prior and thus can be applied to any DLVM with a continuous latent space.
[LG-77] What Would Removing This Corruption Change? Failure Attribution for Vision-Language-Action Models
链接: https://arxiv.org/abs/2610.27277
作者: Carson Sobolewski,Anjali Parashar,Fei Chen,Yingke Li,Chuchu Fan
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: 8 pages, 7 figures
Abstract:Vision-language-action (VLA) models are sensitive to visual corruptions, but knowing that a corruption is typically harmful does not establish whether removing it would benefit the current episode. Furthermore, when several corruptions co-occur, deciding what to correct requires understanding which ones contribute to failure risk and how much their removal would help. We define this problem as failure attribution. Detecting corruption presence alone cannot answer this question, as corruptions can be present but relatively harmless, while directly measuring removal benefits requires costly paired policy rollouts with and without the corruption present. To address this, we introduce a factorized risk model that learns separate representations of the scene and corruptions, allowing us to estimate removal benefits through latent intervention from the current image and task instruction, without paired clean images or additional counterfactual rollouts. We evaluate our approach on OpenVLA and \pi_0.5 in simulation and \pi_0.5 -DROID on a Franka robot. Our method outperforms monolithic and vision-language-model (VLM) baselines, including achieving approximately 3x lower attribution error than frontier VLMs. We also demonstrate our method adapts to execution context and corruption intensity, and can be used for online failure monitoring on hardware.
[LG-78] Solving the Ho–Nguyen Conjecture for Gaussian Mixtures: Applications to Mixtures of Experts
链接: https://arxiv.org/abs/2610.27265
作者: Yuyang Zhang,Qihuang Zhang
类目: Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST); Machine Learning (stat.ML)
*备注: 43 pages, 1 figure
Abstract:Over-fitted Gaussian mixtures can contain many parameter configurations with nearly the same density, making the order of density cancellation central to their parameter-learning rates. Ho and Nguyen encoded these cancellations in a polynomial system and conjectured its exact order as a function of cluster size. The key to resolving the conjecture is a common positive-variance shift that turns the coefficient equations into Gaussian moment matching. For every cluster size \Mhn\ge2 , we prove that the real, nonzero-weight system has order 2\Mhn : a sharp sign-change bound forces rigidity at that order, while positive Gauss–Hermite quadrature constructs a non-trivial solution through order 2\Mhn-1 . Population-density calculations at 110-digit working precision recover the predicted leading term for two to six components; finite multistart searches recover the same observed powers for two to five components, with smaller finite-scale constants for three to five components. For parameter learning, the exact order gives an explicit pointwise maximum-likelihood exponent and an exact local minimax order on specified shrinking neighbourhoods of a fixed known mixing measure. The same coefficient viewpoint separates high-order cancellation from exact parameter symmetries in mixtures of experts. Synchronized replication preserves the conditional density while keeping the corresponding parameter losses positive in two broad standard-softmax classes, showing why statistical rate arguments must account for exact parameter symmetries.
[LG-79] Guided Diffusion for Trustworthy Synthetic Text Generation
链接: https://arxiv.org/abs/2610.27259
作者: Qi Pang,Hanshen Xiao,Andre Manoel,Lipika Ramaswamy,Wenting Zheng,Virginia Smith,G. Edward Suh
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:The rapid evolution of Large Language Models is fueled by massive datasets, yet this reliance poses severe privacy risks, as models can inadvertently memorize and even directly output sensitive training data. Synthetic text generation offers a promising solution to this challenge; however, existing methods struggle to reconcile rigorous privacy guarantees with high data utility. In this paper, we propose a new paradigm for private synthetic text generation: Guided Diffusion. Leveraging the unique steerability of Diffusion Language Models, we introduce novel sampling and re-masking procedures that actively guide the iterative denoising steps. These mechanisms steer the generative process to reconstruct fluent text that adheres to a target semantic distribution. We demonstrate that our framework is naturally compatible with Differential Privacy (DP): once the semantic embeddings of the private corpus are protected, the DP guarantee carries over to the generated text by post-processing. Empirically, we show that our approach generates high-quality synthetic text with precise semantic alignment. Furthermore, we validate its effectiveness in DP synthetic text generation: across two datasets and two diffusion backbones, it outperforms Private Evolution and DP-SGD-based fine-tuning in distributional fidelity and downstream utility, while generating significantly faster than Private Evolution. We further extend the framework to Data-Specific Indistinguishability (DSI), which adds a formal certificate against data poisoning and memorization at a small utility cost.
[LG-80] Deep catastrophic-risk-aware reinforcement learning
链接: https://arxiv.org/abs/2610.27214
作者: Carlos Pérez Octavio Mendoza,Frédéric Godin
类目: Machine Learning (cs.LG); Risk Management (q-fin.RM)
*备注:
Abstract:Estimating extreme risk in reinforcement learning becomes increasingly difficult as the confidence level increases, since only a small number of observations contribute to empirical estimates of Conditional Value-at-Risk (CVaR). In this paper, we extend an Extreme Value Theory (EVT) approach for CVaR optimization to high-dimensional neural-network policies. Using a peaks-over-threshold representation and method-of-moments estimation, we obtain a differentiable tail-risk objective and show that, conditional on the selected threshold, its finite-sample gradient can be computed directly through automatic differentiation. We first validate the approach in a controlled experiment with a known optimal solution and then investigate its performance in a deep-hedging problem with rare extreme market movements. Our results show similar hedging behavior for EVT- and empirical-CVaR-trained policies under normal conditions, while EVT-trained policies exhibit lower hedging losses under the extreme-scenario evaluation across the training sample sizes considered. These results suggest that EVT can provide a practical alternative for training neural policies when the observations relevant to extreme risk are scarce.
[LG-81] A Theoretical Analysis of Environment-Aware Out-of-Distribution Detection for Vision-Language Models
链接: https://arxiv.org/abs/2610.27211
作者: Salimeh Sekeh,Mary Wisell
类目: Machine Learning (cs.LG)
*备注:
Abstract:Continual vision-language models (VLMs) encounter a stream of environments and must also reject inputs that do not belong to the environment at hand. Class information from earlier environments can aid out-of-distribution (OOD) detection, yet existing detectors score each environment in isolation, and when previous environments help or hurt is theoretically unclear. We develop a theory of environment-aware OOD detection built on contribution scores, pairwise cosine expectations measuring how the image and text features of one environment align with those of another. We propose a cross-environment maximum-concept-matching (MCM) score with two terms: a primary term matching the test image against the current class texts, and a secondary term matching it against the class texts of a previous environment, weighted by the intra-modal alignment between the two environments. We prove that, under a positive environment transition and a bounded class-matching margin for OOD inputs, this score attains a false-positive rate no larger than that of the standard MCM detector at an adjusted threshold. We then show that the score gap of an MCM detector between two environments equals the self-contribution of the in-distribution environment minus its cross-modal contribution into the OOD environment, which separates soft OOD environments, with explicit AUROC, true-positive, and false-positive guarantees, from hard ones, which no threshold that retains the in-distribution data can reliably reject. For cascaded detection, in which the text prototypes of all seen environments are pooled, the false-positive rate on every past environment is non-decreasing in time, and we characterize when detection provably fails. A contribution balance factor applied to the gradients of the classification loss and of a smoothed false-positive rate completes the framework. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2610.27211 [cs.LG] (or arXiv:2610.27211v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.27211 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-82] You Are the Trigger: Speaker-Activated Backdoors
链接: https://arxiv.org/abs/2610.27210
作者: Nataša Krčo,Euodia Dodd,Peter Romov,Yves-Alexandre de Montjoye
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Sound (cs.SD)
*备注:
Abstract:Speech models are increasingly used as an interface to AI systems, from transcription tools to voice agents. Existing audio backdoors require an attacker to introduce triggers at inference time. We study a different threat model in which the attacker compromises the model during fine-tuning but has no access to the victim’s inference-time inputs. We introduce the Speaker-Activated Backdoor (SAB) setting, in which a target speaker’s natural voice serves as the trigger. We develop a two-phase fine-tuning method that first makes speaker identity more accessible and then binds the target speaker to an attacker-chosen payload. Compared with a one-phase baseline with the same training budget, the two-phase method substantially improves selectivity and clean performance, and is especially important for content-dependent attacks. SAB achieves 90.2% and 97.0% backdoor success on Whisper-medium and Parakeet, with false-positive rates of 1.7% and 0%, respectively using as little as two minutes of real audio from the target speaker. We further demonstrate content-dependent semantic rewriting and downstream tool-use attacks, where malicious instructions are inserted into 96.9% of target-speaker transcripts. When these transcripts are passed to frontier LLM agents, GPT-4o and Claude (Opus, Sonnet, and Haiku), the injected instruction drives the agent to carry out the harmful action in up to 87.5% of cases. The attack generalizes to held-out recordings, multiple target speakers, and common acoustic degradations. These results establish speaker identity as a practical backdoor trigger that requires no inference-time attacker access and is invisible at the transcript layer.
[LG-83] Low Residuals Wrong Value: Learning the Viscosity Solution of Hamilton-Jacobi-Isaacs Equations with PINNs
链接: https://arxiv.org/abs/2610.27193
作者: Raul Jimenez,Pablo Tejerina-Pérez,Leonid Sarieddine,Pedro Tarancón-Álvarez,Marisol Traforetti,Pau Solé-Vilaró,Oleksii Ignatenko,Nazar Tkach,Pavlos Protopapas
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG); Analysis of PDEs (math.AP); Optimization and Control (math.OC)
*备注: 43 pages, 9 figures
Abstract:Physics-informed neural networks (PINNs) and other residual-trained solvers judge a candidate solution by the size of its sampled PDE residual. For first-order Hamilton-Jacobi equations this is not sufficient: the same equation and boundary data admit Lipschitz almost-everywhere solutions that differ from the viscosity solution by O(1) while violating the equation only on sets of measure zero. We study whether residual minimization selects such solutions in a fixed-horizon pursuit-evasion game whose Hamilton-Jacobi-Isaacs value function is known in closed form on the circle S^1 and on the flat tori \mathbbT^2 and \mathbbT^4 . We train four neural parameterizations with ten seeds on each geometry, 120 runs in total. A standard network with periodic Fourier features converges to folded almost-everywhere solutions in all 30 runs: small residuals and the expected capture-layer gradient, yet values that are wrong over most of the domain. A network with periodicity imposed through the loss recovers the viscosity solution in one and two dimensions but becomes unreliable in four. Two parameterizations that receive the geodesic distance do substantially better: a single network fed the wrapped displacement and its distance recovers the viscosity solution in all 30 runs, while a minimum over periodic image charts succeeds in 27 of 30 and introduces a wrong-image failure mode. Finally, for the convex Hamiltonian considered here, we use the semiconcavity of the viscosity solution to build a mesh-free finite-difference test that needs only forward network evaluations; it detects the folded solutions and the harmful min-of-charts failures without access to the exact value function. A small sampled residual thus does not by itself validate a learned Hamilton-Jacobi solution; the parameterization and independent structural diagnostics can be essential for selecting the viscosity solution.
[LG-84] Meta-iLaD: Identifiable Latent Dynamics via Meta-Learning of Dynamics Environments ICML2026
链接: https://arxiv.org/abs/2610.27180
作者: Yubo Ye,Sweekar Piya,Xiajun Jiang,Linwei Wang
类目: Machine Learning (cs.LG)
*备注: ICML 2026. 23 pages, 9 figures
Abstract:Learning latent dynamics is central to assessing the current states and forecasting the future trajectories of high-dimensional time series. For locally-stationary latent dynamics \mathcalF(\mathbfz_t; \mathbfc) with latent dynamics state \mathbfz_t and environment variable \mathbfc , prior identifiability results have largely focused on \mathbfz_t when conditioned on predefined label u of dynamics environments. This leaves two limitations: reliance on predefined labels that hinder generalization to unseen environments, and a limited understanding of the identifiability of \mathcalF and \mathbfc which – although offering important structural properties for the identifiability of \mathbfz_t – are learned jointly with \mathbfz_t . We address these limitations with Meta-iLaD, a novel identifiable latent dynamics framework attained by meta-learning across dynamics environments. Meta-iLaD introduces a novel conditional prior of \mathbfc , modeled as a feedforward meta-learner to rapidly extract \mathbfc from few-shot examples. Meta-iLaD further establishes identifiability for \mathbfz_t , \mathbfc , and \mathcalF simultaneously, for a general formulation of \mathcalF(\mathbfz_ t;\mathbfc) without restricting the dimension of \mathbfc or how it modulates \mathcalF . On synthetic and real data, we provide strong empirical evidence that 1) conditioning on few-shot examples enables generalization to out-of-distribution environments, and 2) identifiability for \mathbfc and \mathcalF is critical for accurate forecasting beyond reconstructing observed trajectories. Code is available at this https URL.
[LG-85] A Hybrid TFT-NBEATS Framework for Robust Multi-Horizon Photovoltaic Power Forecasting
链接: https://arxiv.org/abs/2610.27155
作者: Shafi Bin Sultan,Sabik Bin Sultan,Mohammad Tajrian Arafat Alvi,Rayhanul Islam Sony,Abdulah Al Numan
类目: Machine Learning (cs.LG)
*备注: Submitted to Elsevier
Abstract:Accurate multi-horizon photovoltaic (PV) power forecasting is essential for grid stability, energy scheduling, and the efficient operation of PVbatterygrid systems. However, existing models often fail to jointly capture short-term variability and long-term temporal patterns, while lacking unified multi-horizon capability and interpretability. To address these limitations, this study proposes a novel hybrid deep learning framework that integrates a Temporal Fusion Transformer (TFT) and N-BEATS within an attention-based fusion architecture. The proposed model leverages TFT to learn short-term temporal dependencies and dynamic feature importance, while N-BEATS captures long-term trend and seasonal components through interpretable basis decomposition. An adaptive attention mechanism dynamically fuses both branches, enabling balanced multi-scale learning. The framework simultaneously generates forecasts for 30-minute, 2-hour, 6-hour, 12-hour, and 24-hour horizons within a single unified pipeline. The model is evaluated using multi-year real-world datasets from the Desert Knowledge Australia Solar Centre (DKASC). Experimental results demonstrate superior performance over statistical, machine learning, and deep learning baselines. The proposed model achieves RMSE values of 42.35, 46.27, 52.29, and 48.11 across the 20162017, 2023, 2024, and 20252026 datasets, respectively, corresponding to a 3540% reduction compared to the best baseline (TFT). Additionally, MAPE remains below 5% across all datasets, indicating high predictive accuracy and stability. The model also provides interpretability through attention visualization and trendseasonality decomposition, and is deployed on a cloud-based platform for real-time forecasting. These results demonstrate that the proposed framework is robust, interpretable, and suitable for practical renewable energy applications.
[LG-86] owards Interaction State Recognition in Sensorized Gecko-Inspired Adhesive Interfaces for Orbital Servicing
链接: https://arxiv.org/abs/2610.27141
作者: Eduart Abdullahu,Francesco Grella,Carol Martinez
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Reliable contact establishment remains a critical challenge for autonomous on-orbit servicing (OOS), particularly when interacting with non-cooperative targets. Gecko-inspired dry adhesives provide a lightweight, passive, and residue-free solution for robotic capture in space, yet current systems operate in open loop, without the ability to estimate their own contact or adhesion state during interaction. This paper presents a real-time interaction-awareness framework that enables sensorized gecko-inspired adhesive interfaces to estimate their contact state throughout the adhesion lifecycle using embedded capacitive sensing. Five discrete interaction states (idle, proximity, contact, engaged, and disengaged) are classified using experimentally collected and augmented sensor data, enabling a comparative evaluation of six lightweight machine learning models suitable for embedded deployment. Among the evaluated models, Random Forest achieved the highest offline classification accuracy of 95.65%. For online deployment, combining Random Forest with a rule-based state machine improved accuracy from 78.02% to 93.14%, achieving a macro F1-score of 0.9343 while maintaining a 53 ms inference cycle. These results demonstrate that incorporating physics-consistent state constraints significantly improves robustness during realtime inference, representing a step toward interaction-aware capture in future autonomous orbital servicing missions.
[LG-87] Glass-Box Deep Reinforcement Learning for High-Speed Flow Control
链接: https://arxiv.org/abs/2610.27132
作者: Ameya D. Jagtap
类目: Machine Learning (cs.LG)
*备注: 22 pages
Abstract:Deep reinforcement learning (DRL) is challenging long-held assumptions in active flow control. By learning directly from interaction with a flow, DRL agents can find actuation strategies for nonlinear, unsteady flows without explicit models, often exploiting mechanisms that classical control would not anticipate. High-speed flows are a demanding test of this promise: shocks, shock/boundary-layer interactions, transition, aerothermal loading and microsecond-scale dynamics strain both algorithms and hardware. Yet most demonstrations remain idealized, low-speed simulations with limited evidence of robustness, interpretability, transferability or experimental deployment. In this perspective, we examine what constrains DRL-based high-speed flow control, including compressible simulation cost and fidelity, data efficiency, sensing in extreme environments, reward design, explainability, and generalization across Mach and Reynolds number. We argue that these limits reflect a tension between data-driven optimization and the structured physics, constraints and uncertainties of real flows. We then highlight where DRL may go beyond conventional approaches: adaptive control of shock-induced separation and turbulent boundary layers, coordination of distributed actuators, and discovery of unconventional flow-manipulation this http URL central challenge, however, is translating these capabilities into reliable technology, which will require validated compressible simulation environments, physics- and constraint-aware learning, robustness to off-design conditions and sensor degradation, real-time operation, hardware-in-the-loop and wind-tunnel validation, and credible pathways to safety and certification.
[LG-88] RGH-GNN : Risk-Gated Heterogeneous Graph Neural Network for Financial Fraud Detection
链接: https://arxiv.org/abs/2610.27107
作者: Mythreyee Hari,Subin Sahayam,Vegesna S M Srinivasavarma
类目: Machine Learning (cs.LG)
*备注:
Abstract:Credit card fraud is a significant problem for financial institutions, resulting in substantial monetary loss each year. Fraud patterns and methods are continuously evolving, with fraudulent transactions often disguised to resemble legitimate activity. Real world transaction data for fraud detection is severely affected by class imbalance making it harder to detect these transactions. Fraudulent transactions are rarely isolated. Accounts, devices, and other common attributes are often shared across fraudulent activity, revealing relational patterns that individual transaction features alone cannot capture. Graph-based approaches model these relationships explicitly. However, standard graph neural networks aggregate neighbourhood information uniformly, which can dilute the already scarce signal carried by fraudulent transactions and limit detection of camouflaged fraud. This work presents RGH-GNN, a heterogeneous graph neural network that addresses this limitation by incorporating transaction-level risk scores, generated by a supervised LightGBM classifier as edge weights. It is used to guide message passing through a learnable risk-gated aggregation mechanism. It ensures that high-risk transactions contribute more strongly to learned representations rather than being diluted by uniform neighbourhood aggregation. RGH-GNN is evaluated on the IEEE-CIS fraud detection dataset against standard graph neural network models that are trained on the same heterogeneous graph. RGH-GNN outperforms all baselines across every evaluation metric, with an AUC-ROC score of 0.914. This demonstrates that incorporating engineered risk signals into graph-based message passing substantially improves fraud detection performance.
[LG-89] LEAF: LUT-aware Efficient Attention on FPGAs
链接: https://arxiv.org/abs/2610.27078
作者: Arianna Cox,Chang Sun,Lauri Laatu,Benedikt Maier,Aadeesh Sharma,Leo Rozanov,Wayne Luk,Maria Spiropulu,Alexander Tapper,Zhiqiang Que
类目: Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex)
*备注: Accepted for the 9th Workshop on Machine Learning and the Physical Sciences (ML4PS 2026)
Abstract:We introduce Lookup-table (LUT)-aware Efficient Attention on FPGAs (LEAF). LEAF is the first nanosecond-scale LUT-aware transformer for FPGAs. Deploying transformers that are large enough to perform well, small enough to fit on an FPGA, and fast enough to meet the sub-microsecond timescales required by real-time systems such as those at CERN remains difficult, since the same attention mechanism that gives these models their high expressivity also makes them costly to implement in hardware. LEAF converts the query, key, and value projections and the final output feed-forward layer into FPGA-native layers with LUT-aware training, creating the first LUT-aware transformer and linformer layers built for ultra-low-latency inference. Attention resists LUT-aware training because it multiplies two activations where the method needs a constant weight to tabulate; LEAF tabulates the attention block’s constant matrix-vector multiplications and leaves the rest in arithmetic form. On a standard jet tagging benchmark this improves accuracy by 1.8 percentage points while using 2.9 times fewer LUTs than the previous state-of-the-art sub-microsecond transformer at the same input size. Our most efficient LEAF-based linformer with 64 input particles reaches 80.4% classification accuracy at 28k LUTs and 35ns.
[LG-90] Hilbert-Scale Stability of Posterior Diffusion Priors for Ill-Posed Signal Recovery
链接: https://arxiv.org/abs/2610.27030
作者: Bahman Moraffah,Raha Moraffah
类目: Machine Learning (cs.LG); Information Theory (cs.IT); Machine Learning (stat.ML)
*备注:
Abstract:Indirect and noisy observations are difficult to invert when the forward operator smooths out the unknown signal. In this case, the unstable modes are precisely the modes that the data reveals least, thus the success of any reconstruction method depends on how prior information interacts with the geometry of the operator across regularity scales. Diffusion priors have become ubiquitous for such problems; however, the current theory still does not determine when a learned score remains informative under ill-posed forward operators. In this paper, we study posterior diffusion for linear inverse problems on Hilbert scales and derive a stability theory under an explicit score-amplification condition. Under a smoothing condition on the forward map, a source condition on the truth, and weak-norm control of score error, we derive non-asymptotic recovery bounds in (\Hs^s) that separate observation noise, score approximation, and reverse-time discretization. These bounds give a recoverable regularity threshold determined jointly by operator smoothing, truth regularity, score regularity, and the posterior correction. The threshold specifies the Hilbert-scale norms (\Hs^s) in which weak-norm score error remains controlled after propagation through the posterior reverse dynamics. For larger regularity indices (s), the inverse geometry magnifies the same error, and the Hilbert-scale recovery estimate no longer gives uniform stability. We also prove that a posterior correction adapted to the forward operator improves stability relative to isotropic guidance in the regularity ranges where the smoothing action of the forward map matters most. The recovery estimates give a regularity-sensitive theory of posterior diffusion and show how learned score error, operator smoothing, and reconstruction norm jointly determine the stable range predicted by the analysis.
[LG-91] EGiME: Entropy-Guided Model Evaluation in Few-Shot Learning
链接: https://arxiv.org/abs/2610.27029
作者: Dequan Jin,Tao Li,Lifeng Xing,Shihui Ying,Hui Zhao
类目: Machine Learning (cs.LG)
*备注:
Abstract:The feature-extraction quality of pre-trained models sets the ceiling of downstream transfer learning, yet few-shot model selection remains difficult: supervised evaluation on scarce labels is unstable and yields fluctuating rankings. We propose Entropy-Guided Model Evaluation (EGiME), a semi-supervised criterion that assesses a candidate model from the clustering structure of its feature space without any classifier training. EGiME performs K -means clustering over all labeled and unlabeled samples, then combines two entropy terms: an inter-cluster entropy measuring global separability, and an intra-cluster entropy measuring the compactness of labeled anchors, fused through an inverse-frequency weight that emphasizes small clusters. On five few-shot benchmarks and nine pre-trained models, EGiME achieves a correlation score of 0.966 with the full-pool ground-truth accuracy obtained by averaging the accuracies of K-means, KNN, and SVM. At labeling ratios of 4%–20%, it produces markedly more stable model rankings than cross-validated supervised baselines, establishing it as a practical alternative to conventional supervised evaluation for pre-trained model selection under data scarcity.
[LG-92] ORBIT: Autoregressive Time Series Generation via Next Frequency-Band Prediction
链接: https://arxiv.org/abs/2610.27000
作者: Chunyi Hou,Xiangfei Qiu,Hanyin Cheng,Chenchen Sun,Yisheng An,Yang Shu,Jilin Hu,Bin Yang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Deep generative models provide an effective modern solution to alleviating the problem of data scarcity by synthesizing realistic time series. However, despite the considerable progress achieved by existing deep time series generators, two key challenges remain: (1) learning hierarchical representations for heterogeneous temporal patterns; and (2) coordinating the generation process of heterogeneous temporal dynamics. To address these key bottlenecks, in this work, we propose ORBIT, a unified deep generative framework for time series. Specifically, we develop two key components: the Frequency-Hierarchical VQ-VAE (FH-VQ-VAE) and the Frequency-Band Autoregressive Transformer (FBAR). The former learns hierarchical representations by decomposing latent temporal representations into different frequency bands and applying residual quantization with increasing token budgets. The latter establishes a novel low-to-high frequency autoregressive direction, yielding a time series generation process that first constructs global temporal structures and then incrementally incorporates fine-grained temporal details. These compact designs not only disentangle heterogeneous temporal patterns for more effective representation learning but also mitigate potential interference among different temporal dynamics during generation. Extensive experiments on multiple challenging real-world benchmarks demonstrate that ORBIT achieves state-of-the-art overall performance while exhibiting superior inference efficiency compared with advanced baselines.
[LG-93] CorteXforge: Scenario-Based RF Dataset Generation on Real Testbeds
链接: https://arxiv.org/abs/2610.26999
作者: Andréa Joly,Leonardo S Cardoso(CITI),Jean-Marie Gorce
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注:
Abstract:Deep learning for radio frequency (RF) signal analysis relies on large annotated datasets, yet synthetic datasets trade physical realism for controllability, while uncontrolled over-theair measurements offer realism but limited reproducibility. We present CorteXforge, a scenario-based framework for generating realistic and reproducible RF datasets on programmable software defined radio (SDR) testbeds. High-level RF scenarios are automatically translated into synchronized experiments executed on the SLICES-RI/CorteXlab platform, producing fully annotated Signal Metadata Format (SIGMF) datasets that capture hardware impairments and propagation effects. We validate CorteXforge through automatic modulation classification using ModForge, an over-the-air (OTA) dataset generated with the framework. Cross-domain experiments reveal a synthetic-to-real performance gap and show that synthetic pretraining remains beneficial for adaptation to OTA data. CorteXforge provides a practical foundation for reproducible RF dataset generation and future research on domain generalization.
[LG-94] Latent Projection for Latent Planning
链接: https://arxiv.org/abs/2610.26979
作者: Yoo Sung Jang,Michael Ryoo
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注: Accepted to the CoRL 2026 Workshop on Modeling Uncertainty in Robotic World Models
Abstract:Latent world models let an agent plan in latent space. The planner searches for an action sequence that minimizes a cost, and that cost is often the \ell_2 distance between the predicted terminal latent and the goal latent. The latent, however, is partly noise: the regularizer that prevents collapse during world model training promotes variance in the directions carrying no signal. We show that the same regularizer also supplies the structure to prune that noise in the latent. SIGReg drives the latent marginal to a standard Gaussian at every step, and assuming further that the latent sequence is a Gauss-Markov process, the transition is linear with its covariance pinned by the map itself, z’ \mid z \sim \mathcalN(Az,, I - AA^\top) . Fitting the state dynamics A therefore yields the covariance at no additional cost, and in its eigenbasis each latent direction carries its own uncertainty. Furthermore, constraining A to be a normal matrix fixes that basis across timesteps, so one eigenbasis measures the uncertainty over an arbitrary number of steps and serves as a stable basis to compute the planning cost on. We then restrict the \ell_2 cost to the k -dimensional subspace the state dynamics carry forward with the least uncertainty. The correction is post hoc: the encoder, the predictor, the rollout, and the planner are all left untouched, and only a single constrained matrix, A , is fit. We evaluate across six environments and show large gains on the 2D navigation tasks, TwoRoom and PointMaze, raising success rate over the plain \ell_2 cost by up to +54.7 points under a harder planning setting. Code is available at this https URL.
[LG-95] Variance Reduction with Delayed Feedback for Non-Convex Optimization
链接: https://arxiv.org/abs/2610.28510
作者: Wei Jiang,Dingzhi Yu,Zechao Li
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:We study variance reduction for non-convex optimization when gradient evaluations return with a delay. The server can update the parameter while earlier evaluations remain unfinished. We first analyze an ordered recursive momentum estimator for stochastic optimization. Under average smoothness and bounded variance, its oracle complexity is O(1+\sigma^2\epsilon^-2+L\Delta_f(1+\sigma/\epsilon+\tau)\epsilon^-2) , where the target is an expected squared gradient norm of at most \epsilon^2 and \tau bounds the feedback delay. We next study finite sums with n components and obtain O(n+L\Delta_f(\sqrt n+\min\tau,n-1)\epsilon^-2) evaluations. An asynchronous snapshot variant continues updating during full-gradient computations under an explicit bound on their completion time. For the stochastic method, we further replace a maximum-delay bound by a bound on the weighted amount of missing feedback.
[LG-96] Confessions of a Masked MDP AISTATS2027
链接: https://arxiv.org/abs/2610.28360
作者: Sebastian Reboul,Lorenzo Croissant,Hélène Halconruy
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注: 18 pages, 7 figures Submitted to AISTATS 2027
Abstract:In reinforcement learning, one presumes the action set to be known ahead of time, and, thus, one must know if any state-action pairs are not permissible ahead of computations. In some applications, determining this action mask is the main computational bottleneck. Isolating this problem to highlight its core properties, we study how to combine the search for an optimal policy with the learning of the mask itself, under a linear MDP model in which each attempted action reveals the exact value of a linear score determining its validity. In this framework, we first construct an algorithm based on deterministic optimistic optimization and show that it is minimax optimal, which we show is also computationally efficient. However, it is sub-optimal on non-degenerate instances, as we show. We propose an alternative, which is also computationally efficient, which plans its exploration strategy. This algorithm admits a formal interpretation as an analogue of information-directed-sampling for deterministic feedback, which may be of independent interest. Finally, we study an application of this framework to transfer learning, where the invalid action set changes over time. Comments: 18 pages, 7 figures Submitted to AISTATS 2027 Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST) Cite as: arXiv:2610.28360 [stat.ML] (or arXiv:2610.28360v1 [stat.ML] for this version) https://doi.org/10.48550/arXiv.2610.28360 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Sebastian Reboul [view email] [v1] Tue, 20 Oct 2026 15:22:31 UTC (428 KB) Full-text links: Access Paper: View a PDF of the paper titled Confessions of a Masked MDP, by Sebastian Reboul and 2 other authorsView PDFHTML (experimental)TeX Source view license Current browse context: stat.ML prev | next new | recent | 2026-10 Change to browse by: cs cs.LG math math.ST stat stat.TH 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) Major funding support from (opens in new tab) (opens in new tab) (opens in new tab)
[LG-97] Stabilizing Filters for Graph Signal Generation
链接: https://arxiv.org/abs/2610.28318
作者: Martin Schmidt,Gonzalo Mateos
类目: ignal Processing (eess.SP); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Submitted to the IEEE Transactions on Signal Processing
Abstract:Generating graph signals is an important task in machine learning, and incorporating the underlying graph into a generative model enables structurally-informed distribution modeling. However, said graph is often noisy or imperfectly estimated, and structural errors can affect the generative process and the resulting signal distribution. While recent work has established stability guarantees for graph neural ordinary differential equations, the robustness of graph-aware reference dynamics, as used in topological Schrödinger bridge models, remains largely unexplored. Here, we investigate the structural stability of continuous-time generative models whose dynamics combine a graph filter with a learned graph neural network. We first derive a Wasserstein stability bound that quantifies the effect of graph perturbations on the generated signal distributions. Informed by this analysis, we then propose a principled framework for designing graph filters that balance structural stability with induced signal smoothing. We characterize the optimal filter design and provide an explicit family of polynomial filters with provable stability guarantees. Experiments on synthetic graphs and real-world fMRI signals support our theoretical findings, illustrating the enhanced structural stability of the proposed filters while preserving generative performance.
[LG-98] Debiased Prior-Data Fitted Networks for Amortized Bayesian Semi-Parametric Inference
链接: https://arxiv.org/abs/2610.28314
作者: Valentyn Melnychuk,Emil Javurek,Stefan Feuerriegel
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Prior-data fitted networks (PFNs) have recently gained popularity as an end-to-end approach to amortized Bayesian inference. Yet, their statistical behavior for semi-parametric targets, such as means under missingness or causal effects, remains poorly understood. In this work, we develop theoretical foundations for debiased PFNs to perform inference on finite-dimensional targets in general semi-parametric settings. First, we show that, despite being Bayes-optimal with respect to the target, an idealized PFN can exhibit a posterior concentration rate that implicitly depends on the concentration rates of (latent) nuisance functions and can become slower as the prior becomes more diverse. Then, to speed up the target posterior concentration rates, we introduce two generic debiasing strategies based on efficient influence functions: (i) a label correction and (ii) a prior correction. Finally, we show that the debiased PFNs (i) (ii) possess concentration rates that asymptotically match a convergence rate given by traditional frequentist bias-corrected estimators. In this way, we establish a novel connection between amortized Bayesian inference with PFNs and frequentist semi-parametric efficiency theory. Under additional assumptions, this allows us to guarantee local and global optimality of PFN-based estimators (namely, the Bernstein-von Mises property and minimax optimality, respectively). Therefore, our work provides guidelines for designing debiased PFNs for semi-parametric inference with principled and provable asymptotic guarantees.
[LG-99] Geometry-Aware Block Selection for Sparse Variational Gaussian Processes
链接: https://arxiv.org/abs/2610.28256
作者: Matteo Hoxha
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Sparse variational Gaussian processes (SVGPs) are a widely used, principled approach to scalable Gaussian process modelling. Their success stems from their theoretically interpretable structure: they provide a lower bound on the marginal likelihood of the full GP while allowing scal- able stochastic optimization. Recent work has explored modifications of the posterior conditional distribution and has shown that such extensions can improve upon this well-established framework. In particular, introducing a block-diagonal structured scal- ing matrix leads to a tighter lower bound on the GP marginal likelihood. We investigate whether selecting blocks that tighten the variational objective improves test RMSE and NLL at a fixed inducing-point budget. Since the scaling matrix contributes to the variational lower bound through an upper bound on the log-determinant of the full residual covariance matrix, our work makes the following contributions. First, we introduce a simple, scalable, geometry-aware algorithm, with explicit computational complexity and low computational overhead, that recovers a large fraction of this log- determinant contribution by selecting an appropriate partition of the data. Second, we show how to train these models using a doubly stochastic optimization framework. Finally, we investigate how sparse GP methods based on alternative conditional posterior structures can be used to learn better hyperparameters and inducing points for SVGP.
[LG-100] Predicting silent failures of learned inference under restricted access: what measurement physics and prior information explain
链接: https://arxiv.org/abs/2610.28155
作者: Mario Gianni
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 22 pages, 5 figures. Code and data: this https URL
Abstract:Learned priors are used to reconstruct regions that measurements barely reach, and to report uncertainty there. They can fail silently: the reported interval excludes the truth while every ground-truth-free diagnostic stays within its noise. We call such errors the blind set and ask whether they can be predicted before the target data are seen. For a correctly specified linear-Gaussian model averaged over the prior, the risk of a silent failure factorises exactly into non-detection and non-coverage, which does not hold for learned systems in general; we build a score that approximates this reference by linearising a VAE prior. In a prospective study on real X-ray CT objects (2DeteCT) under simulated restricted access, with predictions committed by hash before any trial, the score ranked the realised blindness of two learned inference systems (latent Laplace and ensemble Kalman inversion) with Spearman \rho = 0.79 and 0.84 over 240 pre-specified errors. Simple structural predictors would also have passed these tests. Sequential post-hoc analyses, each specified before its own computation, show that under bounded reach the learned score gave no detectable improvement over a physics-based baseline, in rankings or trial-level log scores. At the weakest attenuation studied, prior information improved rankings over the physics-based baseline, most clearly for ensemble Kalman inversion, but a Gaussian prior did as well and the gain was not detectable at the trial level. The score is not a calibrated probability. Silent failures can be anticipated here, but their structure follows mainly from measurement access and the decision quantity rather than from the learned prior.
[LG-101] Near-Optimal Regret Bounds for UCB-Style Gaussian Process Bandits via Data Thinning
链接: https://arxiv.org/abs/2610.28136
作者: Shion Takeno,Shogo Iwazaki,Ichiro Takeuchi
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 30 pages, 4 figures
Abstract:Gaussian process upper confidence bound (GP-UCB) has been widely studied, though its known regret upper bound is suboptimal. To obtain near-optimal regret upper bounds, existing studies mainly focus on non-adaptive methods, such as uncertainty sampling, whose practical effectiveness is often limited. This paper shows that, under a Gaussian noise assumption, UCB-style algorithms can achieve near-optimal regret upper bounds via data thinning, a statistical technique for creating independent datasets from a single dataset. First, we show that data thinning can yield tighter confidence intervals in the bandit setting, and that a direct extension of GP-UCB achieves a near-optimal simple regret upper bound. Furthermore, by leveraging the tighter confidence intervals and combining GP-UCB with a phased elimination algorithm, we show a near-optimal cumulative regret upper bound. Finally, we demonstrate the effectiveness of UCB-style algorithms incorporating data thinning via numerical experiments.
[LG-102] Weak-Lα Noise Strong Guarantees: Gradient Clipping under Affine Heavy Tails
链接: https://arxiv.org/abs/2610.28107
作者: Alyona Nekrasova,Makar Sidorov,Savelii Chezhegov,Alexander Gasnikov,Margarita Preobrazhenskaia,Igor Preobrazhensky,Eduard Gorbunov
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:Gradient clipping is a key tool for obtaining high-probability convergence guarantees for stochastic first-order methods under heavy-tailed noise. Existing analyses typically assume a bounded central moment of order \alpha\in(1,2] , which excludes distributions with tails of order u^-\alpha at the critical exponent. In this paper, we consider smooth convex optimization under affine heavy-tailed noise, whose scale may grow with the gradient norm. We derive high-probability convergence guarantees for clipped SGD (Clip-SGD) and its accelerated variant Clip-SSTM under a weak- L^\alpha tail condition, which is strictly weaker than the corresponding moment bound. For 1\alpha2 , the clipping bounds have the same powers of the clipping level as in the moment case, and both methods attain the additive stochastic rate K^-(\alpha-1)/\alpha . The gradient-dependent noise contributes a faster-decaying term of order K^-2(\alpha-1)/\alpha , up to logarithmic factors. At \alpha=2 , we characterize the additional logarithmic factor on the Lorentz scale and establish its necessity for the additive error at fixed confidence. We also analyze Clip-SGD with prescribed clipping levels and derive finite-horizon guarantees for reaching a neighborhood determined by the clipping bias. For additive noise, matching lower bounds show that this neighborhood is unavoidable, up to constants and the natural error cap. Numerical experiments with Pareto noise illustrate the theoretical results.
[LG-103] Annealed correlation and quenched noise sensitivity in biased deep neural networks
链接: https://arxiv.org/abs/2610.27983
作者: Johan Jonasson
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Probability (math.PR)
*备注: 7 pages
Abstract:In earlier work by Jonasson, Steif and Zetterqvist (2023), fully connected deep neural networks with layers of width n , i.i.d.\ standard Gaussian weights, sign activations and no bias terms were shown to be annealed and quenched noise sensitive with respect to any joint probability distribution \mathbbQ^(n) on the input \boldsymbol\omega and its perturbation \boldsymbol\eta satisfying a mild positive-correlation condition and \mathbbQ^(n)(\boldsymbol\eta \in \boldsymbol\omega,-\boldsymbol\omega) \to 0 . Here we add independent Gaussian biases of the natural order and show that quenched noise sensitivity is retained over a broad range of depths, indeed under the slightly weaker assumption that \mathbbQ^(n)(\boldsymbol\omega = \boldsymbol\eta) \to 0 . Thus, subject to the positive-correlation condition, quenched noise sensitivity holds under the minimal requirement that the perturbed input is asymptotically unlikely to coincide with the original input. At the same time, the annealed covariance remains bounded away from zero and is entirely accounted for by fluctuations in the network-dependent output means.
[LG-104] rraNostra-Bench: Benchmarking Spatial Downscaling and Temporal Rollout for Global 1-km Land-Use Reconstruction
链接: https://arxiv.org/abs/2610.27793
作者: Amirpasha Mozaffari,Marina Castaño,Stefano Materia,Joan Vedrí,Jordi Varela-Agrelo,Dario Garcia-Gasulla,Etienne Tourigny,Mario Acosta,Amanda Duarte
类目: Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG)
*备注: Submitted to Machine Learning: Earth (IOP Publishing)
Abstract:Uncertainty in the terrestrial carbon cycle remains a major constraint in climate projections, partly due to imperfect land surface representations in Earth system models. We propose a data-driven framework for high-resolution land-use reconstruction, treated as a semantic-segmentation task, that combines coarse-resolution scenario data and climate classification with static geophysical features to produce annual land-use maps at 1 km resolution. We introduce TerraNostra-Bench, a change-aware evaluation protocol that separates spatial downscaling from autoregressive temporal rollout and scores models on the rare land-use-change pixels rather than the dominant static background. Using it, we compare convolutional (U-Net) and transformer (Swin) backbones in autoregressive and non-autoregressive schemes against persistence, linear, and shallow convolutional baselines. The autoregressive prior is the most informative input yet accumulates error over long rollouts: autoregressive models lose 5.7-7.2 accuracy points over 13 years, against 2.2-2.8 for non-autoregressive models. This trade-off is governed by the autoregressive choice rather than the backbone. It can be tuned by an inference-time prior-trust threshold that requires no retraining and by the training-time weighting of reconstruction and change objectives; the two levers trace complementary frontiers, with calibration recovering up to 10 accuracy points while keeping change sensitivity above the persistence floor. Extrapolating beyond the training window, backward to 1955-1959 and forward to 2015-2019 under scenario forcing, the models retain most of their in-window skill and show the same trade-off, indicating it is intrinsic to the design. This work provides the methodological foundation and evaluation surface for a forthcoming temporally continuous land-use product.
[LG-105] A Radiation-Hardened Fault-Aware Quantized Neural Network That Self-Heals
链接: https://arxiv.org/abs/2610.27621
作者: Peter Xiangyuan Ma,Joshua S. Bloom
类目: Instrumentation and Methods for Astrophysics (astro-ph.IM); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); Nuclear Experiment (nucl-ex); Instrumentation and Detectors (physics.ins-det)
*备注: 22 pages, 4 figures, 11 tables
Abstract:Deep learning (DL) models are increasingly being deployed into radiation-challenging environments in spacecraft, particle accelerators, and, soon, space-based data centers. Instantiated as a learned orchestration of numbers stored in persistent memory, models are susceptible to continual degradation as high-energy particles corrupt data. At inference time, new data passed through the network fundamentally break the off-line assumptions and guarantees of a static model determined during training. To preserve the quality and efficacy of the models, fault tolerance against bit flips is a clear objective. In this work we developed a runtime method for repairing the quantized weights of neural networks (NNs) after bit flips on commodity, commercial off-the-shelf (COTS) hardware, without storing extra replicas. Our work extends the well-established Fault-Aware Training (FAT) paradigm by adding an alignment penalty to concentrate undamaged activations and weights into a low-rank subspace, and we then reconstruct flagged corrupted coordinates from the uncorrupted survivors. We call this method \it Projection Healing and apply it to six regression benchmarks on quantized NNs. Holding a FAT checkpoint fixed and toggling only the runtime heal gives a further \sim23% reduction in the native-unit excess integral. Across those datasets the method extends the operational scrub interval by a median of 2.04\times at a +5% RMSE degradation and 1.96\times at +50% , compared with 1.62\times and 1.07\times for classical Single Error Correction, Double Error Detection (SECDED). This introduces a new path towards designing resilient data-driven models, complementing established hardware redundancy, ECC, and FAT-based techniques.
[LG-106] Ask to Learn: Query-Based Collaborative Reinforcement Learning
链接: https://arxiv.org/abs/2610.27606
作者: Kaicheng Zhang,Hongqian Niu,Laixi Shi,Didong Li,Guannan Qu,Zhun Deng
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 113 pages, 18 figures
Abstract:Offline reinforcement learning is fundamentally limited by the coverage of the available data, while collecting additional trajectories through online exploration may be costly, unsafe, or infeasible. We study whether this limitation can be overcome without directly interacting with the underlying environment. Motivated by practical settings such as collaborator-mediated simulation, where a collaborator can provide information complementary to the offline data, we introduce query-based collaborative reinforcement learning to study how to acquire this information efficiently. The learner starts with a fixed offline dataset and adaptively queries state-time pairs; the collaborator samples an action according to a fixed, possibly suboptimal policy, and the simulator returns the corresponding reward and next state. For finite-horizon tabular problems, we derive an instance-dependent information-theoretic lower bound on the additional queries required to identify an \epsilon -optimal policy and develop a local quadratic approximation. Guided by this characterization, we develop an adaptive query-allocation algorithm whose query complexity nearly matches the local benchmark under suitable conditions. Experiments on multiple synthetic and semi-synthetic environments demonstrate the empirical efficiency of the proposed method.
[LG-107] Bayesian Neural Networks via Dirichlet Process Priors on the Data-Generating Distribution for Randomized Value Functions in Deep Reinforcement Learning
链接: https://arxiv.org/abs/2610.27570
作者: Sumit Vashishtha
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO)
*备注: Published in Transactions on Machine Learning Research (TMLR). OpenReview: this https URL
Abstract:Bayesian neural networks (BNNs) offer a principled framework for quantifying epistemic uncertainty. However, standard weight-space priors yield complex, high-dimensional neural network posteriors that are computationally prohibitive to sample from, and popular approximate inference schemes often fail to capture epistemic uncertainty effectively—a failure mode particularly acute in reinforcement learning (RL) \citeposband2018randomized. To resolve this challenge, we introduce a new class of BNNs that quantify uncertainty by placing Bayesian nonparametric Dirichlet Process (DP) priors directly on the underlying data-generating distribution rather than on network parameters. Building on this foundation, we develop DP-DQN, a reinforcement learning algorithm that leverages DP-BNNs to maintain posterior distributions over value functions and perform efficient Thompson sampling. By grounding randomized value functions in Bayesian nonparametrics, DP-DQN unifies data-driven posterior reweighting (via the Bayesian Bootstrap) with generative prior hypotheses injected directly through the DP base measure, enabling decisive ``deep exploration’’ with a single neural network without the overhead of multi-network ensembles.
[LG-108] HypoNet Nankai PS: rapid hypocenter determination for the Nankai Trough using physics-informed neural networks with three-dimensional P- and S-wave velocity models
链接: https://arxiv.org/abs/2610.27561
作者: Ryoichiro Agata,Ayako Nakanishi,Dan Bassett,Satoru Baba,Yasuyuki Nakamura,Gou Fujie
类目: Geophysics (physics.geo-ph); Machine Learning (cs.LG)
*备注:
Abstract:Rapid and accurate hypocenter determination based on three-dimensional (3D) seismic velocity models is essential for seismic monitoring in subduction zones, yet it remains computationally demanding. HypoNet Nankai, a hypocenter determination package that uses physics-informed neural network (PINN) travel-time inference for a realistic 3D P-wave velocity model of the Nankai Trough subduction zone, was designed to address this problem, but supported only P-wave arrival times. Here, we present HypoNet Nankai PS, which extends the framework to both P- and S-wave travel times. We first constructed an S-wave velocity model from the existing P-wave model using empirical conversion relations for marine sediments and the crust. Because the resulting S-wave velocity structure has lower velocities and therefore considerably steeper travel-time gradients, the original PINN training procedure designed for the P-wave velocity structure was insufficient. We therefore revised the training procedure by reconsidering the distribution of collocation points, the NN architecture, and the optimizer. Benchmark comparisons with the fast marching method showed that the revised approach substantially reduced the root-mean-square deviations for both P- and S-waves compared with the original algorithm. Recovery tests using synthetic earthquakes were also performed. When applied to 223 earthquakes recorded by DONET stations off the Kii Peninsula, the system yielded hypocenters generally consistent with the JMA catalog beneath the network, suggesting that the 3D velocity model captures travel-time variations accommodated by JMA’s 1D model and station corrections. Farther seaward, the estimated hypocenters were substantially shallower than those in the JMA catalog and were located in the oceanic crust and uppermost mantle, consistent with previous relocations that used DONET data and 3D velocity models.
[LG-109] What Moved the Alarm Score? Causal Attribution with Exact Computation and Scarce-Incident Estimation
链接: https://arxiv.org/abs/2610.27534
作者: Masaaki Takada,Ryusei Shingaki
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Alarms in industrial systems and scientific instruments often demand immediate diagnosis despite scarce incident data. To guide intervention, an explanation must trace how changes in the data-generating process propagate through the causal structure to alter the alarm score, rather than merely flagging shifts that may be downstream effects. Yet exact attribution across interacting causal components appears combinatorial, while existing methods often either assume an unchanged causal model or require enough incident data to estimate the changed system reliably. We address this setting by formulating the alarm-score change as a Shapley game over hybrid structural causal model (SCM) states. Treating structural functions and exogenous laws as distinct players yields a finer-grained, function/law-resolved report of how each type of change propagates to the alarm score. The same framework defines realized-event and population reports, with an event–population coherence identity linking their attributions. For ordered linear SCMs and quadratic scores, a causal first- and second-moment recursion combined with quadrature computes the full Shapley vector exactly in polynomial time, eliminating combinatorial computation. To estimate this target from scarce incident data, we adapt a well-observed reference SCM to the incident regime using cross-fitted estimates of mechanism changes and residual moments, and derive attribution-error bounds and top-k recovery guarantees. Experiments on controlled and measured shifts illustrate the diagnoses and assess scaling to hundreds of variables and estimation from small incident samples.
[LG-110] Learning the Geometry of Materials Space with Trained Neural Tangent Kernels
链接: https://arxiv.org/abs/2610.27412
作者: Daniel Kaplan,Gabriel Kotliar,Anirvan M Sengupta
类目: Materials Science (cond-mat.mtrl-sci); Superconductivity (cond-mat.supr-con); Machine Learning (cs.LG)
*备注: 8 page, comments welcome
Abstract:Neural networks trained to predict materials properties learn representations whose physical content can be difficult to extract. We use the neural tangent kernel (NTK), evaluated at trained parameters, to recover the geometry learned by a network that predicts superconducting transition temperatures from elemental composition alone. Kernel ridge regression with the trained NTK closely reproduces the network’s predictive performance and substantially outperforms regression with the initial kernel. Spectral analysis and multidimensional scaling reveal a geometry dominated by a small number of directions. Clustering and local neighborhood analysis organize compounds into recognizable superconducting families and connect materials with related chemical and correlated-electron properties despite substantial differences in T_c . These results establish trained tangent kernels as practical probes of learned materials representations, connecting property prediction with the exploration of materials space.
[LG-111] Communication Complexity of Universal Features
链接: https://arxiv.org/abs/2610.27402
作者: Lunjia Hu,Anish Jayant,Peter Li,Princewill Okoroafor,Abhishek Shetty
类目: Machine Learning (stat.ML); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
*备注:
Abstract:Motivated by the notion of features from machine learning, we study the complexity of representing data that allows for evaluation with a given family of functions. Concretely, we consider a deterministic encoder that observes X\sim D and produces a finite encoding before a task f\in\mathcalF\subseteq\0,1^\mathcalX is specified. For each task, a decoder must predict f(X) with error at most \varepsilon under D . We construct an encoder that outputs a representation with O(d^\log(1/\varepsilon)) bits, where d^ is the dual VC dimension of \mathcalF . Our encoder greedily selects O((d^/\varepsilon)\log(1/\varepsilon)) functions from the class and uses their joint label pattern to produce the encoding. The key idea in our upper bound is a novel potential function argument, where we compare the entropy gained from each greedy selection with the number of patterns allowed by the dual VC dimension. We present lower bounds showing that the representation length achieved by our encoder is worst-case optimal for every d^ \ge 1 and \varepsilon 1/4 up to a universal constant factor, and instance-wise optimal for every family \mathcalF up to an O(\log(1/\varepsilon)) factor. We give two ways to learn the representation from unlabeled samples, with different guarantees on the representation size and sample complexity. Experimentally, we verify the greedy construction’s accuracy-size tradeoff and explore its connection to neural sparse autoencoders.
[LG-112] EMBER: Predicting TCR–pMHC Binding Without a Docked Complex
链接: https://arxiv.org/abs/2610.27334
作者: Lin Ha,Chen Zhou,Haozheng Luo,Han Liu,Xiaoyu Zhang,Yan Chen
类目: Quantitative Methods (q-bio.QM); Machine Learning (cs.LG)
*备注:
Abstract:We propose EMBER, a structure-based framework for TCR–pMHC binding prediction that eliminates the need for a predicted docked complex, and with it the placement error that dominates such predictions. Traditional structure-based predictors utilize the co-folded complex structure of the receptor and the antigen and interpret binding from its contact surface. Consequently, every downstream feature inherits any deviation the structure-prediction tool introduced. Methodologically, EMBER folds the TCR and pMHC independently, encodes them as SE(3)-invariant residue graphs, and compares their learned representations without constructing a shared coordinate frame. A relative pose is therefore not an input of the model, so placement error cannot propagate into the prediction. Computationally, EMBER preprocesses O(N+M) monomer folds where a complex-based predictor needs O(NM) operations. Experimentally, under epitope-disjoint evaluation with six retrained baselines, EMBER is first on every metric under all four folding tools, by 0.9 to 14.4% in micro-AUROC, 18.5 to 25.3% in micro-AUPRC and 10.6 to 24.1% in MCC. Its ranking is also up to 85% more stable across folding tools.
[LG-113] Physics-Inspired Optimizers beyond the Edge of Stability: Period-Two Orbits and Delayed Instability
链接: https://arxiv.org/abs/2610.27321
作者: Alexander Itin
类目: Chaotic Dynamics (nlin.CD); Disordered Systems and Neural Networks (cond-mat.dis-nn); Machine Learning (cs.LG)
*备注: 34 pages
Abstract:Gradient descent with momentum can be viewed as a particle moving with friction in the loss landscape. In deep learning the learning rate is usually large, so training is a discrete dynamical system: in many full-batch experiments the sharpness grows to the optimizer’s stability limit, the edge of stability (EoS), and training continues there. Physics-inspired optimizers change this motion; Velocity-Regularized Adam (VRAdam), for example, shrinks the step at high speed. We study kinetic maps, in which the step is the gradient of a kinetic energy of the momentum, and ask what they do beyond the EoS. (i) In a reduced model, a kinetic map creates a symmetric period-two orbit beyond the edge, strictly linearly stable if and only if the map’s elasticity (the log-log slope of step length versus momentum) lies between 0 and 1; on it, training sits at an effective edge of stability. VRAdam’s damper fails above twice the critical learning rate, where its kinetic energy is not convex; the Legendre dual of VRAdam’s own Lagrangian is convex and passes at every learning rate. (ii) Because the sharpness rises gradually, the loss of stability is delayed, as in Neishtadt’s theory of slowly swept bifurcations. In a network with a controlled sweep, the dynamics match this theory without fitted parameters, including the memory of where the sweep started and its erasure by noise. (iii) With momentum, reaching the oscillation also requires elasticity at large momentum: maps that bound the step have competing period-four and chaotic attractors already at beta1 = 0.9, while the cubic-Legendre map reached the oscillation from every sampled start up to beta1 = 0.99. Experiments on an exact quadratic problem, a full-batch MLP and ResNet-32 support the theory; on ResNet-18, much of the apparent benefit of the maps is reproduced by matching the effective learning rate.
[LG-114] Diffusion Score Estimation with Shallow Neural Networks: Regularity Approximation and Finite-Sample Rates
链接: https://arxiv.org/abs/2610.27206
作者: Fan Liu,Yahong Yang,Wei Zhu
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 64 pages, 1 figure
Abstract:We study the sample complexity of estimating Ornstein–Uhlenbeck diffusion scores in \mathbb R^d using denoising score matching (DSM) with finite-width shallow ReLU networks, and its implications for reverse-process sampling. Under weak differentiability and polynomial-growth conditions on normalized derivatives of the initial density, we combine complementary estimates from the Gaussian transition kernel and the initial density to obtain explicit Sobolev bounds that are uniform in time on each fixed ball. Combining these bounds with existing shallow-network approximation theory gives explicit control of the approximation error and a sufficient path-norm budget. We then use truncation and localized Rademacher complexity to establish high-probability risk bounds for path-norm-constrained empirical DSM minimizers with unbounded inputs and labels. Under additional bounded-density and sub-Gaussian assumptions, suitable choices of network width and path-norm budget yield squared L^2(P_t) score risk \widetilde O(n^-(d+3)/(2d+3)) at each fixed time t0 , where n is the number of independent training samples and P_t is the forward marginal. The nonasymptotic bounds retain explicit dependence on diffusion time and the regularity and tail parameters of the initial distribution. Aggregating the score errors over a time grid and applying existing sampling theory together with an early-stopping bound yields a high-probability total-variation error \widetilde O(n^-(d+3)/(4d+6)) between the generated and original data distributions under suitable sampling parameters. Here \widetilde O suppresses logarithmic factors for fixed dimension, confidence level, and initial distribution.
[LG-115] Cross-disorder Learning of Imaging-derived Joint Mophometric Markers in Alzheimers and Parkinsons Diseases
链接: https://arxiv.org/abs/2610.27150
作者: Yuji Zhao,Boris Gutman
类目: Quantitative Methods (q-bio.QM); Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
*备注:
Abstract:As two most prevalent neurodegenerative diseases (NDDs), Alzheimer’s disease (AD) and Parkinson’s disease (PD) shared commonalities both in symptomology and pathophysiology. However, to what extent and in what way these two distinct disorders overlap remain underexplored. In this study, we propose an explicit, interpretable and spatially adaptive transfer learning framework to elucidate the morphometric biomarkers linked to AD and PD. We applied such high-dimensional machine learning model to two most popular open-source neuroimaging datasets, i.e., Alzheimer’s Disease Neuroimaging Initiative (ADNI) and Parkinson’s Progression Markers Initiative (PPMI). On the one hand, our cross-disorder classifier revealed neuro-anatomical alteration patterns well aligning with previous neuroimaging research findings and underscoring patients’ clinical presentations. On the other hand, the spatially adaptive trained classifier using shared prior achieved statistically superior diagnostic performance compared to respective individual models, which further highlights the cross-disorder framework’s efficacy on modeling limited-size datasets across disease domains.
[LG-116] Ammonia-Hydrogen Phase Equilibria in Planetary Envelopes from First-Principles-Based Machine-Learning Molecular Dynamics
链接: https://arxiv.org/abs/2610.27136
作者: Akash Gupta,Austin Guo,Jie Deng
类目: Earth and Planetary Astrophysics (astro-ph.EP); Materials Science (cond-mat.mtrl-sci); Machine Learning (cs.LG); Chemical Physics (physics.chem-ph); Geophysics (physics.geo-ph)
*备注: 20 pages, 8 figures
Abstract:Ammonia and hydrogen are major constituents of planets within and beyond the solar system, yet how they interact in planetary envelopes is poorly known. We perform large-scale two-phase coexistence simulations of the ammonia-hydrogen system from 500 to 4000 K and 0.1 to 100 GPa, driven by machine-learning interatomic potentials trained on density functional theory calculations. From these we map the compositions of coexisting hydrogen-rich and ammonia-rich phases, and the critical curve that separates conditions of complete miscibility from those of phase separation. The critical temperature increases with pressure, from 500 K at 0.1 GPa to 2000 K near 30 GPa, and we find no phase separation at 3000 and 4000 K up to 100 GPa. The slope of the critical curve decreases substantially between 0.3 and 30 GPa, and the curve nearly flattens above about 30 GPa and 2000 K, where molecules begin to give way to atomic species. The coexistence curves are asymmetric with an ammonia-rich critical composition. For the expected envelope compositions, the phase boundary lies close to, and intersects, published thermal profiles of Uranus and Neptune, implying ammonia may be raining out in these planets today. In contrast, in warm sub-Neptunes such as TOI-270 d, ammonia and hydrogen are likely miscible in all proportions. Together with previous studies, our results suggest that silicates, water and ammonia separate from hydrogen in sequence as planets cool, with implications for their compositional, thermal and structural evolution.
[LG-117] his is SParTA: A modular efficient and scalable transformer architecture for jet-tagging
链接: https://arxiv.org/abs/2610.27106
作者: Rajneil Baruah,Subhadeep Mondal,Sunando Kumar Patra,Satyajit Roy
类目: High Energy Physics - Phenomenology (hep-ph); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex)
*备注: 31 pages, 5 figures, 10 tables
Abstract:We present the Sparse Particle-interaction Transformer Architecture (SParTA), a modular, scalable, and efficient transformer for jet tagging at high-energy colliders. Exploiting the steeply falling transverse-momentum ( p_T ) spectrum of jet constituents, SParTA factorizes constituent processing into two topological regimes. Ordered by descending p_T , only the N hardest constituents (the hard core) participate in quadratic O(N^2) differential interaction attention driven by pairwise physical and learned relational features. The remaining soft constituents are integrated via linear O(lN) cross-attention to the hard core, preserving soft-radiation and color-flow information while avoiding the O(l^2) memory bottleneck of full self-attention over all l constituents. Pairwise interaction features are computed dynamically in-network from constituent four-momenta and conditioned on global jet-level substructure observables. On the quark-gluon tagging benchmark, a 300K-parameter SParTA model achieves an AUC of 92.02% and accuracy of 85.06%, matching the 2.14M-parameter Particle Transformer (ParT) at one-seventh the parameter count and outperforming IAFormer by 21% in background rejection at 30% signal efficiency. On the boosted top-tagging benchmark, SParTA outperforms ParT and ParticleNet with 511K parameters, while we analyze the architectural trade-offs relative to fully dense multi-layer interaction networks.
[LG-118] LOB-Impact Benchmark: Emergent Market Impact in Generative Limit Order Book Models
链接: https://arxiv.org/abs/2610.27100
作者: George Nigmatulin,Valentin Mohl,Sascha Frey,Kang Li,Aramis Marti-Shahandeh,Mihai Cucuringu,Anisoara Calinescu,Jakob Foerster,Stefan Zohren
类目: Trading and Market Microstructure (q-fin.TR); Machine Learning (cs.LG)
*备注: 9 pages, 4 figures, 3 tables. Accepted at ICAIF '26 (7th ACM International Conference on AI in Finance), Milan, Italy. Code: this https URL
Abstract:Realistic market impact is essential for generative limit order book (LOB) models to serve as world models for trading agents. If a simulator misprices this mechanism, agents learn policies that are optimal in the simulator but wrong in the real market. Existing metrics assess generated streams purely through distributional comparison with historical data, leaving counterfactual response to an agent’s own orders largely unexplored. Here, we show that market impact emerges naturally in neural LOB generators trained solely for next-token prediction. We inject execution-style metaorders into the generated message flow of five autoregressive models across five NASDAQ stocks and measure the response using estimators from the empirical impact literature, validated on baselines whose physics is known in advance. Our benchmark evaluates any message-level generator in three phases: free generation (reading order-flow persistence against price diffusivity); metaorder injection (yielding the impact law’s exponent and amplitude); and a post-execution window (measuring relaxation). In addition, comparing each visible injection to a blind, simulator-only control isolates the model’s learned response from standard book mechanics. For every model on every stock, we measure how the impact responds as a function of cumulative metaorder participation. Interestingly, we observe emergent impact in generative models of market data, with exponents in the range of 0.63 to 1.21, slightly above the traditional 0.5 of the square-root law of impact. In the family of models tested, we are so far unable to observe impact decay during relaxation. We tentatively attribute this to the still-short context windows of the models. We release the benchmark as open source to validate future generative models for market data.
[LG-119] Graph Neural Networks for Astrophysics and Cosmology
链接: https://arxiv.org/abs/2610.27051
作者: Christian Kragh Jespersen
类目: Instrumentation and Methods for Astrophysics (astro-ph.IM); Cosmology and Nongalactic Astrophysics (astro-ph.CO); Astrophysics of Galaxies (astro-ph.GA); Machine Learning (cs.LG)
*备注: Main text contains 43 pages, 17 figures, and 2 tables. Invited review chapter for the edited book Machine Learning Techniques for Astrophysics and Cosmology (Eds. Cosimo Bambi, Vinay Kashyap, Swarnim Shashank, Naoki Yoshida, Springer)
Abstract:Physics seldom makes breakthroughs when we consider objects only in isolation. Instead we model and measure how objects sit among their neighbours, spatial, temporal, or otherwise: a galaxy in its environment, a halo in its history, a photon hit in a detector. Graphs are ideally suited for encoding exactly these relations, and Graph Neural Networks (GNNs) learn the geometric functions that respect them. This review presents the use of GNNs for astrophysics and cosmology, and argues that GNNs often suit physics and physical data better than any alternative: they fit the native geometry of the domain and the signal, enable the incorporation of the symmetries and prior knowledge we already trust and have toiled to build, and stay flexible enough to capture what we do not yet understand. GNNs’ physically intuitive workings also render them more interpretable than other deep-learning architectures. We show how we can begin from general principles of the geometry of deep learning and build the graph and the graph neural network from the ground up, and how other architectures, like the convolutional network and the transformer, can be seen as special cases of one unified geometric blueprint. We survey applications across cosmology, galaxy formation, dynamics, and instrumentation; and we chart the frontier. From this survey, we find that only one in six articles outside instrumentation applies GNNs to data rather than just simulations. We also distil a set of best practices, including when not to reach for a graph. Throughout, we will prioritise physical intuition and foundational principles over mathematical machinery whenever possible.
[LG-120] Gaussian Approximation of Besov Functions for Logistic Classification on Manifolds
链接: https://arxiv.org/abs/2610.27022
作者: Guanhang Lei,Lei Shi,Xiaoming Yuan
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST)
*备注:
Abstract:We study Gaussian kernel approximation and regularized logistic classification on a connected compact smooth embedded manifold X\subset\mathbb R^d without boundary and of intrinsic dimension \varrho . Two main challenges are to derive a fast oracle inequality without assuming a bounded Bayes function and to construct Gaussian approximants that exploit both the intrinsic dimension and the target’s regularity without saturation. We address the first challenge through a Bernstein-type inequality whose constant depends linearly on the supremum norm of the predictor. Combined with localized complexity estimates, this yields an oracle inequality over separable normed function spaces. To address the second challenge, we establish a uniform Sobolev operator expansion and cancel its lower-order bias terms using a signed combination of Gaussian kernel integral operators at different bandwidths. Real interpolation then gives an L_2(X) approximation error of order \gamma^\alpha for functions in B^\alpha_2,\infty(X) , where \gamma is the bandwidth. By constructing the approximants directly on the manifold without ambient extension, we obtain a squared Gaussian RKHS norm bound of order \gamma^-\varrho . This construction thus overcomes second-order saturation while retaining norm control governed by the intrinsic dimension. When the marginal distribution has a density g\in L_p(X) , these results yield the logistic excess-risk rate \bigl((\log n)^\varrho+1/n\bigr)^4\alpha(1-1/p)/(4\alpha(1-1/p)+3\varrho) . We further show that clipping predictions yields the faster clipped logistic excess-risk rate \bigl((\log n)^\varrho+2/n\bigr)^2\alpha(1-1/p)/(2\alpha(1-1/p)+\varrho) . For bounded densities, its polynomial exponent matches that of the minimax optimal rate for squared-loss regression, and its corresponding excess misclassification risk also matches the minimax optimal rate.
[LG-121] Robust Latent Variable Graphical Models: Exact Recovery and Fundamental Limits
链接: https://arxiv.org/abs/2610.27010
作者: Chuansen Peng,Xiaojing Shen
类目: Machine Learning (stat.ML); Information Theory (cs.IT); Machine Learning (cs.LG)
*备注:
Abstract:We study robust estimation of a Gaussian precision matrix that decomposes as a sparse matrix minus a symmetric, possibly indefinite, low-rank matrix, a structure that arises when latent-variable marginalization is combined with conditioning on linear measurements. The proposed convex estimator replaces the sample covariance by the centered Gaussian measure-transformed (MT) covariance, which preserves the signed graph of the sparse component under the nominal model, and penalizes the low-rank component by the nuclear norm of a congruence transform, an atomic norm that encodes structural side information. Our main result is a primal–dual witness theorem: under restricted invertibility and irrepresentability conditions on the Fisher information, the raw convex optimizer is unique, has no false edges, recovers the signs of all sufficiently strong edges, and has exactly the true rank and inertia. These conditions reduce to the graphical-lasso conditions when the low-rank part is absent, and we verify them from primitive incoherence and near-isotropy assumptions. Over a class on which they hold uniformly, we prove information-theoretic limits for Frobenius estimation, signed graph recovery, and inertia recovery. In particular, inertia recovery requires N\gtrsim p/\nu_\min^2 samples, where \nu_\min is the smallest nonzero eigenvalue modulus of the low-rank part; the spectral beta-min condition of the witness theorem is necessary up to a \sqrt\log p factor; and the Frobenius risk attains the minimax rate up to a factor \sqrt\mu(1+\min\d,r)\log p , where \mu is the incoherence parameter, d the maximal degree, and r the rank. We further quantify robustness under Huber contamination, with a bias bound that is sharp for the MT functional and vanishes for remote outliers, and give a globally convergent linearized ADMM with closed-form updates.
[LG-122] When Weak Submodularity Meets Matroids: How Does Standard Greedy Behave?
链接: https://arxiv.org/abs/2610.26967
作者: Qixin Zhang,Shi Fu,Shengminjie Chen,Chao Qian,Dacheng Tao
类目: Optimization and Control (math.OC); Discrete Mathematics (cs.DM); Machine Learning (cs.LG)
*备注: 33 pages
Abstract:Weakly submodular maximization finds numerous applications in machine learning, data mining and statistics. In particular, under a cardinality constraint, the standard greedy algorithm is known to attain the tight (1-e^-\gamma) -approximation guarantee, where \gamma denotes the submodularity ratio. However, when considering general matroid constraints, \emphhow well standard greedy can perform has remained a longstanding open question. In this paper, we resolve this question by studying the worst-case approximation guarantee R_\mathrmgr(\gamma,r) of standard forward greedy, where r is the rank of matroid. Specifically, our first contribution is the following refined lower bound \beginequation* R_\gr(\gamma,r)\ge \max\left\ \frac\gammar, \frac\gamma^2 B_\gamma+\sqrt\gamma(r-1), \frac\gamma^r+1 2-\gamma+\gamma^r+1, \frac\gamma^2 14\left(B_\gamma+\sqrt\gamma(1-\gamma)(r-1)\right) \right, \endequation* where B_\gamma\triangleq2-\gamma+\gamma^2 . This lower bound not only strictly improves upon the previous guarantee of 0.4\gamma^2/(1+\sqrt\gamma r) , but also can recover the classical 1/2 -approximation guarantee at \gamma=1 . Subsequently, a corresponding upper bound is established for R_\mathrmgr(\gamma,r) , namely, \beginequation\labeleq_upper_abstract R_\gr(\gamma,r) \le\min\left\frac12,20\max\left\frac\gammar, \frac\gamma^2B_\gamma+ \sqrt\gamma(1-\gamma)(r-1)\right\right. \endequation For every fixed \gamma\in(0,1) , this Eq.\eqrefeq_upper_abstract effectively sharpens the recent O_\gamma(1/\log(n-1)) upper bound to O_\gamma(r^-1/2) , where n is the size of ground set. Putting all these results together, we eventually can obtain [R_\gr(\gamma,r) = \Theta!\left( \max\left\frac\gammar, \frac\gamma^2 1+\sqrt\gamma(1-\gamma)r\right\ \right).]
[LG-123] Quadratic Sample Complexity for Strictly Replicable Realizable Learning with Unknown Marginals
链接: https://arxiv.org/abs/2610.26944
作者: Guangjian Zhang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 34 pages
Abstract:Under the quantifier \exists A,\forall D , where the realizable distribution and its marginal are unknown, we determine the finite-class sample-complexity exponent for strict replicability: two independent samples and one shared seed must produce identical functions. For \varepsilon=\delta=\rho=10^-4 , the worst case over |H|\le M is \Theta((\log M)^2) , without iterated logarithms. For d\ge1 , 0\varepsilon\le1/16 , and 0\delta,\rho\le10^-3 , every randomized, possibly improper learner valid throughout an explicit anchored realizable family requires \Omega(d^2/(\varepsilon\rho^2)) samples. For the full Boolean class P_d , d\ge2 , in the same parameter range, we prove the matching characterization \Theta(d^2/(\varepsilon\rho^2)+\varepsilon^-1\log(1/\delta)) using a proper Poissonized ball-sampling algorithm. We also give an explicit proper Gibbs upper bound for every finite measurable Boolean class. Our lower bound uses square-root probabilities and volumetric packing; the Gibbs proof builds on established coupling and generalization identities. Finally, we characterize the known-marginal worst case for P_d , showing why that interface must be kept distinct.
[LG-124] ShlokaDhara: A Multi-Speaker Sanskrit Speech Corpus for Learning Meter-Aware Prosody in Text-to-Recitation
链接: https://arxiv.org/abs/2610.26929
作者: Harshal Kausadikar,Onkar Kishor Susladkar,Prathmesh Santosh Choudhari,Vedangi Deepak Deshpande,Ismini Lorentzu,Sparsh Mittal
类目: Audio and Speech Processing (eess.AS); Machine Learning (cs.LG); Sound (cs.SD)
*备注: Under submission at a conference
Abstract:Sanskrit verse recitation is governed by \textitChandas , a strict prosodic framework defined by rhythmic \textitLaghu (light) and \textitGuru (heavy) syllables. Because existing Sanskrit speech datasets prioritize ASR or conversational TTS, they lack the metrical fidelity required for authentic verse synthesis. We introduce \textbf\textitShlokaDhara , a 43-hour, multi-speaker Sanskrit speech corpus capturing diverse meters, recitation styles, and melodic variations. We specifically focus on \textitText-to-Recite , a synthesis task explicitly requiring both prosodic accuracy and metrical correctness. Benchmarking state-of-the-art TTS architectures via an automated UTMOS (neural MOS) predictor demonstrates that XTTS-v2 achieves the highest perceptual quality. We further validate the XTTS-v2 generated samples using two paradigms: subjective assessments by 16 Sanskrit experts and novel empirical metrics, specifically, syllable-level \textitLaghu-Guru duration analysis to objectively quantify structural metrical awareness. Ultimately, this work establishes a robust dataset and evaluation benchmark for prosody-aware computational Sanskrit recitation. Generated Samples: this https URL
[LG-125] LatticeMind: A Validation-Gated Agent ic Framework for Density Functional Theory Workflows
链接: https://arxiv.org/abs/2610.26869
作者: Nathaniel M. Vegh,Yu Zhu,Raphaël J. Prentki,Félix Beaudoin,Shuaishuai Yuan,Hong Guo
类目: Materials Science (cond-mat.mtrl-sci); Machine Learning (cs.LG); Computational Physics (physics.comp-ph); Quantum Physics (quant-ph)
*备注: 26 pages, 12 figures, and 12 tables, including Supplemental Material. Prepared for submission to PRX Intelligence
Abstract:First-principles calculations are rarely single solver calls: they are multistage workflows in which structures, settings, and saved densities must pass correctly between dependent stages. Language models can write a valid single input, but reliability collapses once stages are chained: a dropped dependency, wrong file handoff, or hallucinated keyword can invalidate results even when the solver runs, and such errors surface only after costly execution. We introduce LatticeMind, a validation-gated agentic framework that converts natural-language requests into executable, checkable density functional theory (DFT) projects. A workflow contract specifies required stages, files, and dependencies; deterministic validators inspect the project before any solver launch; failed checks trigger bounded repair; and reports remain traceable to project files. On a 44-case RESCU benchmark spanning total-energy, density-of-states, band-structure, magnetic, phonon workflows, three non-flagship LLM backends reproduced every reference value; 119 of 132 runs passed on the first validated plan and the rest after one repair cycle. Ablations identify the workflow contract and pre-execution validation as the principal reliability mechanisms; a conventional trial-and-repair baseline reaches 71% at twice the token cost per success. The same architecture reproduces every reference value on a 19-case VASP benchmark, demonstrating transfer across solvers. Beyond these task families, LatticeMind generated and validated the 23 inputs of a multistage study of Rh-induced electronic responses at a Rh _6 /In _x Ga _1-x N photocatalyst interface. These results establish workflow contracts, deterministic validation, bounded repair, and traceable records as a practical foundation for reliable AI-assisted scientific computing, where preventable errors are caught before execution and the solver is never the first validator.
附件下载
