本篇博文主要内容为 2026-10-23 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 (2026-10-23)
今日共更新773篇论文,其中:
- 自然语言处理共117篇(Computation and Language (cs.CL))
- 人工智能共231篇(Artificial Intelligence (cs.AI))
- 计算机视觉共114篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共208篇(Machine Learning (cs.LG))
- 多智能体系统共19篇(Multiagent Systems (cs.MA))
- 信息检索共27篇(Information Retrieval (cs.IR))
- 人机交互共31篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] Decentralized Bi-Level Optimization for Cooperative Orbital Transportation of a Rigid Payload
【速读】：该论文旨在解决在轨拥挤与空间碎片增加背景下，对大型非合作目标进行可靠在轨运输的挑战。传统单体架构存在单点故障风险，而采用小型航天器编队虽可分散风险并提升系统鲁棒性，但需应对各航天器推力受锥形约束与载荷旋转动力学耦合的非凸轨迹优化问题。现有方法或为分布式控制律（具备渐近稳定性但约束满足能力有限），或为集中式/部分集中式规划（重新引入单点故障隐患）。本文提出一种基于精确梯度的分布式双层优化框架：将原问题分解为内层凸推力分配与外层载荷扭矩轨迹搜索两部分。该分解结构可在不完全收敛于非线性规划（NLP）的情况下快速获得燃料高效解，并在短时间尺度下性能随航天器数量增加而提升。理论分析证明了算法的收敛性，通过大规模蒙特卡洛仿真验证了其有效性，包括飞行中故障的闭环恢复能力。上述特性使该框架特别适用于航天器编队任务中的在轨运输场景。
链接: https://arxiv.org/abs/2610.31562
作者: Sidhdharth D. Sikka,Revathy Ravuri,Hei Yin Jowett Chan,Tim Echtermeyer,Shaoshuai Mou
机构: Purdue University (普渡大学); University of Manchester (曼彻斯特大学); Manifold Research Group (曼德尔研究组)
类目: ystems and Control (eess.SY); Multiagent Systems (cs.MA); Robotics (cs.RO)
备注: 10 pages, 6 figures, 4 tables
Abstract:The rapid growth of the space economy, alongside increasing orbital congestion and debris, is driving demand for reliable on-orbit transportation of large and non-cooperative objects. Performing such transportation with a swarm of small spacecraft distributes risk and removes the single point of failure inherent to monolithic architectures, but requires solving a non-convex trajectory optimization coupling every agent’s cone-constrained thrust to the payload’s rotational dynamics. Existing approaches span decentralized control laws with asymptotic guarantees but limited constraint enforcement, and centralized and partially centralized planners that reintroduce a single point of failure. In this work, we present a decentralized bi-level framework driven by exact gradients: the problem is split into an inner convex thrust allocation and an outer search over payload torque trajectories. This decomposition yields fuel-efficient solutions quickly, without requiring full NLP convergence, and at short timescales its performance scales with the number of agents. Theoretical convergence guarantees are established, and the method is validated through extensive Monte Carlo simulation, including closed-loop recovery from in-flight faults. These properties make the framework well suited to spacecraft swarm applications.
[MA-1] Contestable Agents for Data Governance: Recording Human Stewardship of AI-Generated Metadata Across Structured and Unstructured Systems
【速读】：该论文旨在解决工业组织在管理结构化与非结构化运营知识源时面临的治理难题，即传统治理模式依赖少数具备深厚领域知识的员工进行元数据分类、关联与文档化，效率低下且易引入错误。现有基于大语言模型（Large Language Models, LLMs）的智能代理虽可自动提出元数据建议、引用文档并识别质量问题，但未经审核的提议可能将错误传播至下游系统，而简单的审批流程又缺乏可追溯性。为此，论文提出“可争议治理”（contestable governance）机制：由专业领域代理生成带有评分和引用证据的论证，用于支持存在争议的元数据决策；由管理者（steward）在提交前可接受、拒绝、修改或补充论证，确保审查过程透明可追溯；治理记录完整保留所有证据、干预行为与最终结果。同时，通过版本化的领域术语词典（versioned domain glossary）为代理提供模型专用术语与缩写定义，提升准确性。该方案设计了分层审查路径，仅对不确定性高的提案触发完整论证流程，兼顾效率与严谨性，并引入决策门控机制（gate），控制经验证的决策是否成为后续参考先例。两个暖通空调（HVAC）应用场景验证了该方法的有效性：一是利用Brick Schema标签对设备点位进行分类，依托开源参考实现；二是通过链接索赔字段、遥测数据与服务笔记及修订版公告，实现故障模式判定的可解释决策。核心解决方案在于构建一个可审计、可辩论、可追溯的协同治理框架，将生成式人工智能的效率优势与人工专家的判断权威相结合。
链接: https://arxiv.org/abs/2610.31553
作者: Humza Akhtar,Brian Freeman
机构: MongoDB, Inc.(MongoDB公司); Trane Technologies(特耐技术公司)
类目: Artificial Intelligence (cs.AI); Databases (cs.DB); Human-Computer Interaction (cs.HC); Multiagent Systems (cs.MA)
备注: 15 pages, 2 figures, 2 tables
Abstract:Industrial organizations hold operational knowledge in structured records, such as telemetry, point lists, and transactions, and in unstructured sources, such as manuals, service notes, communications, and inspection reports. Governance programs often manage these sources separately. Classifying, linking, and documenting them depends on a small group of employees with deep domain knowledge. Agents built on large language models can propose metadata, cite documents, and flag quality issues, but unreviewed proposals can spread errors into downstream systems. A perfunctory approval leaves no record of what a reviewer checked. We present contestable governance: specialist agents produce scored, cited arguments for disputed metadata decisions; a steward can accept, reject, modify, or add arguments before commit; and a governance record retains the evidence, interventions, and outcome. A versioned domain glossary supplies model-specific terms and acronym definitions to the agents. We describe the workflow, a triage path that reserves full argumentation for uncertain proposals while preserving review before commit, the scoring and record models, and a gate that controls which verified decisions become precedent. Two HVAC use cases illustrate the pattern: point classification using Brick Schema class labels, supported by an open reference implementation, and an illustrative warranty failure-mode decision linking claim fields and telemetry to service notes and revisioned bulletins.
[MA-2] Dispersive Tree Agreements for Multi-Robot Coordination
【速读】：该论文旨在解决多机器人非线性系统协同控制中难以兼顾联合最优性与分布式实时性的核心挑战。现有方法或为集中式离线优化（仅保证全局最优但缺乏实时响应），或为分布式实时规划（支持安全导航但无法逼近联合最优解），二者在实际应用中存在明显权衡。本文提出的关键解决方案是：通过构建一种基于分散树（dispersive trees）的分布式列生成机制，将原本耦合的多机器人协同问题转化为局部约束下的可并行搜索过程。其核心创新在于证明了机器人间的相互作用可通过局部共享空间的“租金”（rents）进行有效建模，使得各机器人在本地生长分散树时，能通过异步协商机制逐步收敛至近似联合最优的共识解。该方法以“随时可用”（anytime）的方式运行，其协议WarpWeaver在滚动时域框架下实现异步实时迭代，即使在时间预算不足的情况下仍能达成稳定且近优的协调方案。仿真结果表明，相比集中式离线规划器，WarpWeaver在Dynobench基准上可减少22%总行程时间；相较同类分布式实时基线，提升达37%；同时显著提高图连通性任务的成功率。硬件实验在六台机器人上验证了其在机载平台上的可行性，证实了该方法在真实环境中的实用价值。
链接: https://arxiv.org/abs/2610.31540
作者: Shashank A. Deshpande,Blake T. Edwards,Jonathan P. How
机构: Massachusetts Institute of Technology (麻省理工学院)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA); Optimization and Control (math.OC)
备注: 19 pages, 5 figures, 7 tables. Submitted to IEEE Transactions on Robotics
Abstract:Multi-robot coordination of nonlinear systems is naturally formulated as joint-space optimal control subject to constraints over groups of robots. Existing methods that approach joint optimality remain largely centralized, offline, and confined to collision-free motion planning, while decentralized, real-time counterparts target safe navigation without pursuing the joint optimum. This paper develops a decentralized, real-time planner that approaches the joint optimum as an anytime process. We extend a recent single-robot framework that searches dispersive trees of forward-propagated commands for differentially flat systems to coordinate a multi-robot team near-optimally. We first show that the inter-robot coupling can be effectively replaced by some local constraints, under which locally grown dispersive trees contain a near-optimal agreement. It is this structure that the robots pursue by decentralized column generation, negotiating rents on the shared space that guide the search of each robot’s tree. The negotiation settles on an agreement whose near-optimality is certified at quiescence. Our protocol, WarpWeaver, implements this negotiation as an asynchronous real-time loop over a receding horizon, and demonstrates effective closure to quiesced agreements even when real-time deadlines cut the negotiation short. In simulation, WarpWeaver reduces total travel time by up to 22% against centralized offline planners on the Dynobench suite and by up to 37% against a decentralized real-time baseline, and nearly doubles the success rate of a task-specific planner on graph connectivity. Hardware experiments on six robots planning onboard demonstrate real-world implementability.
[MA-3] When Is a Memory Safe Enough to Act On? Action-Conditioned Verification for Long-Lived Agents
【速读】：该论文旨在解决长期智能体（long-lived agents）在跨会话执行中因记忆过时而导致不可逆错误行为的问题。现有方法仅关注记忆本身的维护（如检索、刷新或调和），但未考虑不同动作对记忆验证需求的差异：同一记忆条目在常规运输中可直接使用，而在高价值操作前则需重新确认。为此，论文提出动作条件化记忆验证（action-conditioned memory verification）机制，其核心在于根据候选动作的性质动态决定哪些记忆前提需要重验、何时执行或放弃行动，以权衡执行风险、弃用代价与验证成本。作者将此问题建模为一个序贯停止问题（sequential stopping problem），其中验证只有在可能改变后续决策时才具有价值；在验证器可靠性与成本满足一定条件时，随着动作后果严重性增加，相同记忆可依次进入“直接执行”、“验证后执行”和“直接弃用”三种策略模式。基于该理论框架，研究提出了名为 \textscAMV-Guard 的实用验证门控系统，其通过提取动作依赖的记忆前提、校准对错误可能性的信念，并采用有界信息价值（bounded Value of Information, VOI）选择验证项。在 ToolSandbox 上，相较于静态策略，\textscAMV-Guard 在保持 100% 安全执行召回率和 0% 有害行为的前提下，平均验证成本降低 33.9%，且执行覆盖率从 38.83% 提升至 44.17%；在 AppWorld 的门控核心迁移实验中，验证成本降低 21.3%，同时仅轻微影响安全执行召回率与任务成功率。结果表明，验证资源的分配应由即将执行的动作所决定，而非仅依赖记忆本身的可信度。
链接: https://arxiv.org/abs/2610.31446
作者: Yiqi Wang,Yifei Wang,Jiaqi Zhang,Zhangkai Wu,Yiqun Duan,Taotao Cai,Mingkai Zheng
机构: University of Southern Queensland (南昆士兰大学); Southern University of Science and Technology (南方科技大学); Jiangsu University (江苏大学); The University of Sydney (悉尼大学); Uploading Inc. (上传公司)
类目: Multiagent Systems (cs.MA)
备注: 22 Pages, 5 Figures, 12 Tables
Abstract:Long-lived agents act on memory carried across sessions, and many of their actions are irreversible, so a stale memory can turn directly into harm. Existing work improves the memory itself — retrieving, refreshing, or reconciling entries — yet whether an entry must be re-checked before acting is not a property of the entry: the same remembered address can be used unchecked for a routine shipment but should be re-confirmed before a valuable one. We therefore study \emphaction-conditioned memory verification: given a candidate action, decide which of its memory premises to re-verify and when to proceed or abstain, balancing execution risk, abstention, and verification cost. We formulate this as a sequential stopping problem, in which a verification is worth its cost only if it can change what the agent does next; under suitable verifier reliability and cost conditions, increasing action consequence can move the same memory through proceed, verify, and direct-abstention regimes. Building on this formulation, \textscAMV-Guard is a practical verification gate that extracts the premises a candidate action rests on, calibrates beliefs about whether they are wrong, and selects checks by bounded value of information (VOI). On ToolSandbox, relative to a static policy given the same inputs, it lowers mean verification cost by 33.9% while keeping 100% safe-proceed recall and 0% Gate harmful, and raises Proceed coverage from 38.83% to 44.17%; in a gate-core transfer to AppWorld, which carries over the stopping/VOI core but not the full structured pipeline, it lowers verification cost by 21.3% at a small cost in safe-proceed recall and task success. These results suggest that verification effort should be set by what an agent is about to do, not only by how reliable a memory appears.
[MA-4] Conservative Meta-Agent Search: Architecture Adaptation under Non-Stationarity
【速读】：该论文旨在解决自动化智能体系统（agentic system）在实际部署中面临的动态环境挑战，包括持续的分布外（out-of-distribution）变化、延迟的依赖性反馈、组件故障的关联性以及架构变更的成本高昂等问题。传统评估方法依赖静态任务集与即时标量奖励，无法反映真实场景中的复杂性。为此，论文将元智能体架构搜索形式化为带有切换成本和硬风险约束的序列型现有架构替换问题，并提出保守型元智能体搜索（Conservative Meta-Agent Search, CMAS）。CMAS的关键在于：通过回溯历史架构并提议类型化的图结构修改，利用相同随机数下的成对影子回放（paired shadow replay）结合干预性组件归因（interventional component credit）对候选架构进行评分；仅当其得分在经验校准的方差惩罚下同时超过预设阈值时才允许晋升，并随后引入模拟“金丝雀”阶段与预设回滚机制以保障安全性。为验证元层级泛化能力，研究构建了MetaInvest-Bench——一个具有制度切换特性的仿真器，支持可计算的近似最优架构，具备防泄漏回放、对抗性治理轨迹及五种负向控制。实验表明，在96条金融流上，CMAS将归约于动态或然最优解的标准化均值后悔率从0.116降至0.104，同时提升保留样本效用至0.48（优于最强自适应基线），且在冗余工作者无增益时能自动收缩为接近单智能体团队的轻量结构。尽管其部署推理开销降低约6%，但包含搜索过程的总令牌消耗为各方法中最高。观测到的每轮更新虚假晋升率为3.4%（目标5%），密封重跑下为4.8%，校准基于作者设计基准上的更新级回放统计，不保证整体效用提升、模拟金丝雀阶段无害性，亦不提供序列级错误控制。
链接: https://arxiv.org/abs/2610.31436
作者: Ali Atiah Alzahrani
机构: 未知
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注: 24 pages, 8 figures, 17 tables
Abstract:Automated agentic-system design is typically evaluated on static task suites with immediate scalar rewards. Deployed systems face persistent shift, delayed dependent feedback, correlated component failures, and costly architecture changes. We formalize meta-agent architecture search as sequential incumbent replacement with switching costs and hard risk constraints and introduce Conservative Meta-Agent Search (CMAS). CMAS retrieves archived architectures and proposes typed graph edits, then scores each against the incumbent by paired shadow replay under common random numbers, with interventional component credit. Promotion requires an empirically calibrated, variance-penalized simultaneous score above a pre-specified margin, and is followed by a simulated canary stage and pre-specified rollback. Since backtests cannot certify meta-level skill, we introduce MetaInvest-Bench, a regime-switching simulator whose specified worker competence makes a surrogate-optimal architecture computable, with leakage-controlled replay, adversarial governance tracks, and five negative controls. Financial streams serve as a risk-sensitive testbed for meta-level generalization; we claim no alpha. Across 96 streams, CMAS cuts normalized mean surrogate dynamic-oracle regret from 0.116 to 0.104 and raises held-out utility from 0.44 to 0.48 over the strongest adaptive baseline; when extra workers add no value, it contracts to near-single-agent teams. Its deployment inference cost is ~6% lower, though its search-inclusive token cost is the highest of the methods compared. The observed per-update false-promotion rate is 3.4% against a nominal 5% target, and 4.8% in a sealed rerun. The calibration is empirical: it concerns an update-level replay statistic on an author-designed benchmark, and guarantees neither full-utility improvement, nor absence of simulated canary-stage harm, nor sequence-level error control.
[MA-5] Probabilistic Energy Modeling Unifies Versatile Swarm Control
【速读】：该论文旨在解决机器人集群在面对多样化任务时，现有控制器需频繁重新设计、调参或重新训练的局限性问题。其核心挑战在于如何实现跨任务、跨规模与跨平台的高效泛化控制。解决方案的关键在于提出一种基于概率能量建模（probabilistic energy modeling）的统一框架，将多种集群行为抽象为可复用的能量项（energy terms），通过组合这些共享的交互原则来生成多样化的任务目标，而非依赖于特定任务的控制策略。该方法利用吉布斯随机场（Gibbs random field）能量最小化与结构化奖励最大化之间的对应关系，在统一的概率目标函数中整合基础行为模式。通过轻量级策略门控（policy gating）机制将分布式策略集成至单一控制器，无需重新训练；引入可解释的动作注意力机制以捕捉非均匀局部交互，并采用无量纲归一化实现跨集群规模与硬件平台的迁移。实验表明，该统一策略在六类任务、10–100个智能体、两种真实硬件平台以及室内外复杂环境中均实现了零样本迁移（zero-shot transfer），大规模仿真中任务成功率高达0.96–1.00，单智能体推理延迟仅为1–10毫秒，物理部署亦验证了其在真实场景下的有效性。这确立了概率能量建模作为可扩展、可迁移的去中心化集群控制新范式。
链接: https://arxiv.org/abs/2610.31371
作者: Xiang Zhou,Xiangyu Zhu,Xintong Wang,Qingyuan Yang,Rui Huang,Feng Xue,Lin Li,Qingrui Zhang
机构: Sun Yat-sen University (中山大学)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注: 25
Abstract:Robotic swarms must coordinate across diverse missions, yet existing controllers often require redesign, retuning, or retraining when objectives change. Here we show that probabilistic energy modeling provides a unifying representation for versatile swarm control, in which diverse missions emerge from compositions of shared interaction principles rather than task-specific control formulations. By exploiting the correspondence between Gibbs random field energy minimization and structured reward maximization, we encode fundamental swarm behaviors as reusable energy terms within a unified probabilistic objective. Decentralized policies trained within this representation are composed into a single controller through lightweight policy gating, without retraining. An interpretable action-attention mechanism captures non-uniform local interactions, while dimensionless normalization supports transfer across swarm scales and platforms. Demonstrated on aerial swarms, the unified policy generalizes zero-shot across six mission classes, 10-100 agents, two hardware platforms, and indoor cluttered and GPS-denied woodland environments. Large-scale simulations achieve 0.96-1.00 task success rates with 1-10 ms per-agent inference latency, while physical deployments confirm zero-shot transfer to real platforms and unstructured environments. These results establish probabilistic energy modeling as a scalable and transferable framework for decentralized swarm control.
[MA-6] Muffin - A Shared Memory Buffer for Asynchronous AI Agents
【速读】：该论文旨在解决当前目标驱动型推理任务中，单智能体与多智能体系统在性能提升上趋于持平的问题，其核心挑战在于智能体间交互与协调过程中关键信息传递效率低下。现有基于同步轮次辩论的多智能体范式依赖于各智能体间的显式信息共享，易导致通信开销大且效率受限。为此，本文提出一种基于共享记忆缓冲区推理（Memory Buffer Inference, MUFFIN）的异步智能体交互协议，通过引入一个由编排器（orchestrator）动态调度查询、将复杂任务分配至多智能体环境的机制，在保证各智能体本地私有推理独立性的同时，利用共享记忆缓冲区实现异步的中间结果交换。实验表明，MUFFIN框架在20个基准测试中的17个场景下达到最优准确率，相较主流多智能体辩论方法，准确率提升19.2个百分点，同时减少77.4%的令牌消耗，且无需完整推理轨迹共享，显著优化了性能与成本之间的权衡。
链接: https://arxiv.org/abs/2610.31355
作者: Ping-Lin Kuo,Suman Navaratnarajah,Christian Frey,Gresa Shala,Josif Grabocka
机构: University of Technology Nuremberg (纽伦堡应用技术大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:The rise of Agentic AI systems has changed the landscape of artificial intelligence by enabling strong reasoning capabilities through specialized multi-agent interaction with structured self-reflection. However, in target-driven reasoning tasks, single- and multi-agent systems are often on par in terms of performance gains. A major challenge is that failures often arise from an ineffective exchange of critical information during inter-agent interaction and coordination. Different from existing multi-agent debate paradigms that rely on synchronous turn-taking and sharing information across different agents, we propose an inter-agent protocol based on shared Memory bUFFer INference - referred to as MUFFIN - in an asynchronous agentic framework. In the first step, an orchestrator routes queries based on their reasoning demands and delegates demanding ones to a multi-agent setting, where a shared memory buffer provides asynchronous exchange of intermediate findings while an agent’s private reasoning remains local. Our experiments show that our framework Muffin outperforms established competitors, achieving the best accuracy in 17 out of 20 of our benchmark settings, and provides a better quality-cost trade-off than multi-agent debate: Muffin improves accuracy by 19.2 accuracy points while using 77.4% fewer tokens, without requiring full-trace sharing.
[MA-7] Automating Feature Flag Cleanup at Scale with Multi-Agent LLM Systems: An Industrial Experience at DoorDash
【速读】：该论文旨在解决大规模软件系统中因长期未清理的特性开关（feature flag）所引发的技术债务与运维风险问题。随着特性逐步上线，已完全发布的功能所对应的开关及其代码路径若未及时移除，将导致冗余代码积累，增加系统复杂性并可能在无意中激活潜在错误路径。在DoorDash的实际场景中，超过56,000个特性开关分布在623个代码仓库中，截至2026年5月已有1,659个处于“过期”（stale）状态，且手动清理每个开关平均耗时1至2小时，现有基于规则的自动化工具难以有效应对这一挑战。本文提出一种端到端的基于大语言模型（Large Language Model, LLM）的特性开关自动清理系统，其核心创新在于集成多源数据：通过MCP（Model-Driven Configuration Protocol）查询实时特性发布数据、解析Jira工单以获取上下文依据，并自动生成经工程师审核的拉取请求（Pull Request）。评估结果显示，该系统在50个过期开关上的测试中成功生成了45个（90%）可被工程师批准的删除拉取请求，平均每次清理耗时仅4.79至13.8分钟，验证了其在大规模生产环境中的可行性与高效性。此外，研究还通过成功率、清理成本及时间等指标评估了系统的质量与可靠性边界，并总结了构建此类多智能体系统的关键经验，为同类系统的开发提供了实践指导。
链接: https://arxiv.org/abs/2610.31351
作者: Atharv Chandratre,Jai Datar
机构: DoorDash, Inc.(DoorDash公司)
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 6 pages, 2 figures. Accepted at the 2026 IEEE International Conference on Software Maintenance and Evolution (ICSME), Industry Track
Abstract:Feature flags are runtime configuration mechanisms used for gradual feature rollouts, experimentation, and kill switches. Once a feature is fully rolled out, the flag and its gated code become technical debt and can introduce operational risk through inadvertently enabling dormant code paths. At DoorDash, over 56,000 feature flags span 623 repositories, with 1,659 currently stale as of May 2026, with internal estimates of one to two hours per manual cleanup. Existing rule-based cleanup tools proved insufficient in DoorDash’s codebases. Our contribution is an end-to-end automated LLM-based feature flag cleanup system that integrates querying live feature flag rollout data using MCP, Jira ticket processing, and end-to-end pull request creation. Our system successfully produced engineer-approved flag removal PRs for 45 (90%) of the 50 stale flags it was evaluated on, proving the system’s operational viability at scale with each cleanup averaging only 4.79 and 13.8 minutes. We evaluated the system’s quality and reliability boundary based on success rates, cost per cleanup and time metrics. Furthermore, we discuss our lessons learned when architecting this system, to provide practical guidance for developing similar multi-agent systems.
[MA-8] Consistent but Biased: Chair Personas Trade Noise for Bias in Multi-Agent Clinical Decisions
【速读】：该论文旨在解决在由多个大语言模型（Large Language Model, LLM）代理组成的临床决策系统中，主导代理（即“主席”代理）的个性提示（persona prompt）是否影响最终决策，以及这种变化是否仅影响决策时间而非实质内容的问题。现有系统常将多次运行结果的一致性视为可靠性指标，但未在保持病例不变的前提下检验这一假设。研究通过模拟一个由七名代理组成的虚拟科室，对20例退行性颈椎脊髓病（degenerative cervical myelopathy）的真实病例进行重复测试，固定病例报告不变，仅改变主席及其副手的提示类型：包括无个性提示、咨询型、专断型和回避型，且以行为规则或人格特质两种方式呈现。结果显示，当采用规则形式的提示时，专断型人格使手术决策立即执行的比例高达93%，而回避型则降至2%，相比之下无特定人格提示时为57%；尽管所有情况下均倾向手术，但人格提示主要影响的是“立即手术”还是“延迟观察”的决策时机。该效应在另一模型及重述提示后仍可复现。进一步分析表明，专断与回避人格虽提升了三轮运行间的一致性，但其一致性水平接近随机预期，且弗莱斯kappa值（Fleiss κ）未显著提升，说明所谓“一致性”并非真正可靠。值得注意的是，仅描述人格特质的提示仍保留了超过一半的专断与回避型行为影响，表现为快速决策、延迟行动及增加检查，但未能再现忽略警示信号、放弃备选方案或转交家庭决策等更深层次行为。因此，该研究的关键结论是：必须将主席代理的个性提示作为关键变量加以报告与控制，并在评估运行一致性时校正随机一致性的基准，否则易误判为可靠性。
链接: https://arxiv.org/abs/2610.31325
作者: Yitong Bai(Columbia University),Chengyi Zhang(Swansea University),Ziyang Wang(Aston University)
机构: Columbia University (哥伦比亚大学); Swansea University (斯旺西大学); Aston University (阿斯顿大学)
类目: Multiagent Systems (cs.MA)
备注: 32 pages, 8 figures
Abstract:In clinical systems built from several large language model (LLM) agents, one agent often chairs the discussion and makes the final decision. Whether the chair’s persona prompt changes that decision, or the agreement between repeated runs that some systems treat as reliability, has not been tested with the case held constant. We ask whether the persona changes the decision or only its timing, how it affects agreement between runs, and which behaviors appear when the prompt describes only personality traits. A simulated department of seven agents discussed 20 published cases of degenerative cervical myelopathy, with the case reports replayed unchanged in every run. Only the prompts of the chair and its deputy changed (no persona, or a consultative, autocratic or avoidant one), written either as rules of conduct or as personality traits. With personas written as rules, the chair chose to operate now in 93% of runs when autocratic and in 2% when avoidant, against 57% without a persona. Surgery was still planned in nearly all decisions, so the personas mainly changed whether to operate now or after more tests. These effects held with a second model and reworded prompts. The autocratic and avoidant personas made the three runs of a case agree more often, but only about as often as chance predicts, and Fleiss \kappa did not rise. Prompts that described only personality traits kept more than half of the autocratic and avoidant effects on operating now. They reproduced quick decisions, delays and extra tests, but not ignoring red flags, dropping backup plans or passing the decision to the family. Studies should report and vary the chair’s persona as they do the model, and correct agreement between runs for chance before calling it reliability.
[MA-9] Agri-SCHOLAR: Simulation-Grounded In-Season Rolling Advisories for Agricultural Crops Using LLM -Bounded Optimization
【速读】：该论文旨在解决现有农业指导体系存在的三大核心问题：（a）过度依赖专家经验而缺乏数据驱动；（b）未能充分整合实时气象信息以实现动态调整；（c）难以实现对个体农场的精细化适配。其解决方案的关键在于提出Agri-SCHOLAR（基于模拟的作物收获优化与大语言模型引导的在季农事建议滚动系统），该系统通过一个闭环架构，将多个大语言模型（LLM）代理与强大的作物模拟器（如本研究中采用的APSIM）相耦合，实现动态、个性化、可执行的农事建议生成。其中，作物模拟器作为评估判据（evaluation oracle），确保建议在农学上的合理性与可行性；LLM代理负责生成可行的操作参数边界，再由多目标优化器在该边界内搜索，综合考虑产量、资源利用效率与经济可行性等多重目标。实证分析显示，该框架在南亚两个代表性区域对玉米作物进行11年周期的模拟验证后，显著优于当地标准农艺实践，有效克服了传统农业指导的局限性。
链接: https://arxiv.org/abs/2610.31003
作者: Geetha Charan,Vedant Balasubramaniam,Manojkumar Patil,V. Priyanka,Kodur Sai Vinay Sathvik,Rohit P Suresh,Mohammad Sufyan Azam,Y. Narahari
机构: Indian Institute of Science(印度科学研究所); University of Michigan(密歇根大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Existing agricultural advisories to farmers suffer from several limitations: (a) remain largely expert-driven, (b) do not adequately account for in-season weather, and © do not scale to individual farms. In this paper, our objective is to close these gaps by developing Agri-SCHOLAR (Simulation-Grounded Crop Harvest Optimization with LLM-Guided In-Season Advisory Rollouts). Agri-SCHOLAR is a rolling, in-season advisory framework that issues, at intervals of every few days or every week, a personalized advisory using the following information: soil and crop properties, observed weather to date, short-range forecast of weather, and a season extrapolation of weather based on historical analogs. The novelty and depth of Agri-SCHOLAR lies in each advisory being produced by a closed-loop system which couples multiple LLM agents with a powerful crop simulator (such as APSIM in our case). The simulator crucially serves as an evaluation oracle. In Agri-SCHOLAR, the LLM agents propose agronomically sound and feasible bounds on actionable parameters and an optimizer searches within those bounds against a multi-objective reward spanning yield, resource utilization, and economic feasibility. We conduct a detailed 11-year analysis of the maize crop in two selected locations of South Asia and show that Agri-SCHOLAR consistently outperforms the region’s standard packages of practice and satisfactorily addresses the limitations of existing agricultural advisories.
[MA-10] Consensus-Based Distributed Filtering for Angles-Only Tracking in Cislunar Space
【速读】：该论文旨在解决在地月空间中仅依赖观测航天器网络的视线（angles-only）测量，实现对目标航天器惯性状态的分布式协同估计问题。在缺乏地面支持、通信与计算资源受限且存在单点故障风险的自主地月导航场景下，传统集中式融合架构面临严峻挑战。为此，论文提出三种基于一致性机制的分布式滤波方法：分布式扩展卡尔曼滤波（Distributed Extended Kalman Filter）、信息加权一致性滤波（Information-Weighted Consensus Filter）和卡尔曼一致性滤波（Kalman Consensus Filter）。三者共享局部预测与量测处理步骤，其核心差异在于节点间信息交换与融合方式。解决方案的关键在于通过设计有效的信息一致性协议，在保证局部计算效率的同时，实现跨网络的全局状态估计性能提升。仿真结果表明，相较于独立本地滤波器，所提分布式方法显著提高了估计精度；蒙特卡洛敏感性分析进一步验证了各方法在有限一致性迭代次数及代理间先验偏差条件下的鲁棒性。
链接: https://arxiv.org/abs/2610.30864
作者: Giovanni Romagnoli,Fabrizio Pilone,Giordana Bucchioni
机构: University of Pisa(比萨大学)
类目: ystems and Control (eess.SY); Multiagent Systems (cs.MA)
备注: Accepted for presentation at IEEE NAVICON 2026
Abstract:This paper investigates consensus-based distributed filtering strategies for multi-observer angles-only navigation in the cislunar space. The objective is to estimate the inertial state of a target spacecraft using only line-of-sight measurements collected by a network of observer spacecraft. This problem is particularly relevant for autonomous cislunar navigation, where ground support may be limited, and centralized fusion architectures may be problematic due to communication and computational constraints, and single-point failure risks. Three distributed filtering approaches are considered: the Distributed Extended Kalman Filter, the Information-Weighted Consensus Filter, and the Kalman Consensus Filter. All methods share the same local prediction and measurement-processing steps, but differ in the way information is exchanged and fused across the network. Their performance is evaluated and compared to a centralized Extended Kalman Filter, adopted as benchmark, and independent local filters in a high-fidelity cislunar scenario. Numerical simulations show that the distributed approaches significantly improve estimation accuracy with respect to independent filters. Monte Carlo sensitivity analyses are then conducted to characterize the robustness of the considered methods to finite consensus iterations and prior misalignment among the agents.
[MA-11] ADAPT for Science: A Multi-Agent System for Self-Evolving Scientific Optimization
【速读】：该论文旨在解决科学发现中在高维、受限搜索空间内高效识别少数高性能设计的难题，尤其针对评估成本高昂、可行解稀疏、参数强耦合及物理约束严格等挑战。传统通用型人工智能优化方法缺乏对科学先验知识、物理约束和有限评估预算的针对性设计。为此，本文提出面向科学的多智能体框架ADAPT for Science，其核心在于将科学优化建模为“何处搜索”与“如何搜索”两个耦合决策过程：通过历史设计数据挖掘具有前景且多样性的初始条件，逐步聚焦计算资源于有利的搜索方向，并基于执行反馈动态演化优化策略。该方法在引力波探测器设计任务中，结合可微分物理仿真器与八种未见过的拓扑结构进行验证，在相同评估预算下实现均方损失0.204 ± 0.101，相较最优经典优化器降低54.9%，相较最优通用代理演化基线降低62.2%，并在所有未见拓扑上均取得最低损失。ADAPT的关键创新在于构建了一个能够持续从过往经验与自身搜索结果中学习的科学优化系统，显著提升了复杂科学问题下的优化效率与泛化能力。
链接: https://arxiv.org/abs/2610.30851
作者: Zhaoran Feng,Yuchen Ni,Jianan Liu,Lemao Liu,Dejing Dou,Kairui Feng
机构: Shanghai Innovation Institute, Shanghai 200231, China; Shanghai Research Institute for Intelligent Autonomous Systems, Tongji University, Shanghai 201210, China
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Optics (physics.optics)
备注: Includes supplementary material
Abstract:Scientific discovery often requires identifying a small number of high-performing designs in vast, constrained search spaces. This is difficult when evaluations are costly, feasible solutions are sparse, parameters are strongly coupled, and physical constraints are strict. Recent AI for AI methods can automatically improve algorithms and optimization procedures. However, general purpose approaches are not designed around scientific priors, physical constraints, and limited evaluation budgets. We introduce ADAPT for Science, a multi-agent framework that treats scientific optimization as two coupled decisions: where to search and how to search. ADAPT uses historical designs to identify promising and diverse initial conditions, progressively concentrates computation on favorable search directions, and evolves the optimization procedure from execution feedback. We evaluate ADAPT on the design of gravitational wave detectors using a differentiable physics simulator and eight unseen detector topologies. Under matched scientific evaluation budgets, ADAPT achieves a mean loss of 0.204 +/- 0.101, reducing loss by 54.9% relative to the best classical optimizer and by 62.2% relative to the best general purpose agentic evolution baseline. ADAPT achieves the lowest loss on all unseen topologies among the evaluated baselines. ADAPT provides a step toward scientific optimization systems that continuously learn from prior experience and their own search outcomes.
[MA-12] ABRA: Agent -Bound UI for Human-Agent Communication beyond Chat with Reusable Views EMNLP2026
【速读】：该论文旨在解决当前智能代理（Agent）虽能生成丰富的上下文表示，但这些表示极少被用作表达用户意图的媒介这一关键问题。其核心挑战在于如何实现任务间可复用的交互结构与具体任务语义之间的解耦，以支持多任务场景下的动态协作。解决方案的关键是提出“代理绑定用户界面”（Agent-Bound UI, ABUI）交互模型，通过将任务特定的语义绑定至可复用的交互视图中，并将其作为共享状态发布，从而使得人与代理能够在共享的交互表面（如交互画布）上理解并响应不同任务背景下的动态语义。为验证该模型，作者实现了无需依赖特定框架的原型系统ABRA，支持用户与代理通过聊天及共享交互画布协同工作，并在画布创作、所有者侧后果处理及跨代理调用等任务上对四个基于大语言模型（LLM）的代理进行了评估，证明了该方法的有效性与普适性。
链接: https://arxiv.org/abs/2610.30780
作者: Seon Gyeom Kim,Hannah Kim
机构: Megagon Labs
类目: Human-Computer Interaction (cs.HC); Multiagent Systems (cs.MA)
备注: EMNLP 2026 System Demonstrations
Abstract:Agents can create rich representations, yet these representations are rarely used as a medium for expressing user intent. We present Agent-Bound UI, an interaction model that separates reusable interaction structures from task-specific meanings. Agents bind task-specific meanings to interactions in reusable views and publish the resulting bindings as shared state, enabling humans and agents to understand and act on shared surfaces whose meanings vary across tasks. We implement ABRA, a framework-agnostic prototype of Agent-Bound UI, in which users and agents collaborate through chat and shared interactive canvases. We demonstrate ABRA through a representative use case and evaluate four LLM-backed agents on canvas authoring, owner-side consequence resolution, and cross-agent invocation. The code and demo video are available at this https URL.
[MA-13] LLM Multi-Agent Systems Are Vulnerable to Herding and Runaway Cascades
【速读】：该论文旨在解决大语言模型（Large Language Models, LLMs）在多智能体系统中协作时，因相互响应他人选择而引发的内生反馈循环所导致的社会性影响、从众行为及失控级联现象等潜在风险。其核心解决方案的关键在于通过类比经典的音乐实验室（Music Lab）实验范式，构建大规模LLM智能体群体进行陌生歌曲推荐任务，并对比在有无累积推荐次数暴露条件下的集体决策表现。实验发现，当智能体暴露于累计推荐数据后，会迅速产生失控级联效应，导致集体推荐结果严重偏离歌曲本身的真实吸引力，且表现出远高于人类的从众倾向，同时缺乏自我校正能力。这一发现揭示了多智能体系统在推荐系统、自动化评估、复杂问题求解及非法共谋等场景中存在显著脆弱性，强调需在系统设计中引入抗干扰机制以抑制非理性集体行为。
链接: https://arxiv.org/abs/2610.30769
作者: Farnoosh Hashemi,Shiyu Ji,Joice Chen,Michael W. Macy
机构: Cornell University(康奈尔大学); Stanford University(斯坦福大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Large language models (LLMs) increasingly operate in multi-agent systems where they collaborate by responding to others’ choices, creating endogenous feedback loops with vulnerability to social influence, herding behavior and runaway cascades. Following the design of the classic Music Lab experiment, we tested this vulnerability in large populations of LLM agents that recommended unfamiliar songs, with or without exposure to cumulative recommendation counts. Exposure led to runaway cascades that generated extreme and unpredictable collective recommendations increasingly divergent from a song’s intrinsic appeal. Compared to humans, LLM populations were generally far more susceptible to herding behavior, and unlike humans, LLM cascades were less able to self-correct. This vulnerability has implications for multi-agent collaboration in recommender systems, automated evaluation, complex problem-solving, and unauthorized conspiracies.
[MA-14] Governed Self-Evolution for Autonomous Investing: A Multi-Agent LLM System with an Invariant Core
【速读】：该论文旨在解决大语言模型（LLM）代理在实时网络数据环境下进行自主投资时面临的三大核心挑战：异构证据融合、历史行为记忆以及基于结果的自我改进，同时确保不削弱系统自身的安全约束。其关键解决方案在于构建一个具有类型系统边界控制的多智能体架构——Steward，通过三重机制实现安全且可演化的自主决策：首先，九个角色专精的智能体协同工作，包括四类分析师分别处理新闻、财报、宏观经济与市场数据，并生成带类型的报告，由投资组合经理设定目标仓位，再经受保护的风险管理器审计；其次，七层共享记忆系统通过系统自身记录代码构建，为每个决策提供市场叙事、校准信息及历史决策上下文；最后，系统通过“回顾-评估-学习”闭环实现自我优化：每日对交易与预测进行代码化评分并写入记忆，每季度由反思者提出提示学习，经确定性门控可逆执行，并通过回放历史输入验证有效性。整个系统以“订单算术、风险规则和纪律性提示”为核心不变量，任何自修改都无法触及。实验表明，在5个月真实部署中，系统将24.6%的行业回撤控制在0.3%损失以内；在相同数据下，单一模型五日内相对市场亏损1.35点，而联合设计无显著损失，且未响应任何注入指令；2015–2026年回测显示，仅核心机制即可在全部50种随机入场策略下降低最大回撤。反思模块成功定位根本故障原因，门控机制在触发日实现学习采纳，而在对照日保持静默。
链接: https://arxiv.org/abs/2610.30557
作者: Yebo Feng,Jingyang Dai,Liangrui Pan,Nelson Lau,Jiahua Xu,Yang Liu
机构: Nanyang Technological University, Singapore(南洋理工大学, 新加坡); University College London(伦敦大学学院, 英国)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Multiagent Systems (cs.MA); Portfolio Management (q-fin.PM)
备注:
Abstract:LLM agents that act on live Web data must combine heterogeneous evidence, remember what they have done and improve from outcomes without weakening their own safeguards. We study this problem in autonomous investing, where every action is irreversible and based on untrusted Web content. We present Steward, a multi-agent LLM system that draws the boundary of self-evolution in its type system. First, nine role-specialised agents decide jointly: four analysts read news feeds, SEC filings, macroeconomic series and market data and emit typed reports, from which a Portfolio Manager sets target positions that a protected Risk Manager audits. Second, a seven-layer shared memory, assembled by code from the system’s own record, gives each decision its position context, the market narrative, calibration and the agent’s past decisions. Third, the system improves itself by review: code grades every trade and forecast nightly and writes the grades into memory, and each quarter a reflector proposes prompt learnings that a deterministic gate applies reversibly and replay of recorded inputs verifies. All three operate around an invariant core of order arithmetic, risk rules and discipline-agent prompts that no self-edit can reach. In a five-month live deployment (102 trading days, 1,632 LLM calls), the system contained a 24.6% sector drawdown to a 0.3% loss; in replay of its recorded inputs, a single model given the same raw data traded more and lost 1.35 points against the market over five days where the joint design showed no significant loss, and no injected instruction was followed; and a 2015-2026 backtest shows the core alone lowering the maximum drawdown under all 50 random entry policies. The reflector traced the dominant failure to its root cause, and the gate applied a learning that the agent adopted on trigger evenings and on no control evening. Code: this https URL
[MA-15] Beyond End-to-End Recovery: Towards a Better Understanding of LLM Steganography via Rolewise Decomposition
【速读】：该论文旨在解决大语言模型（Large Language Model, LLM）隐写术中因端到端（End-to-End, E2E）恢复评估方法局限性所导致的安全隐患识别不充分问题。传统E2E恢复仅衡量解码器能否正确重建私密信息，无法区分编码失败与解码失败，亦难以揭示私密信息在编码器输出文本中的具体表征方式。为此，论文提出一种分角色分解框架（rolewise decomposition framework），分别分析私密信息在编码器生成文本中的表示形式、解码器实际恢复的内容以及确定性验证可识别的编码模式。通过对7,200条编码器响应的分析，研究发现存在168个可验证但被解码器遗漏的编码实例，以及706个通过词长或标记首字符等未明确指示的模式实现信息重构的响应。这些结果表明，仅依赖E2E恢复可能掩盖编码与解码过程中的关键差异，因此必须对二者进行独立评估，以更准确理解并缓解由LLM隐写术引发的安全风险。
链接: https://arxiv.org/abs/2610.30516
作者: Om Tailor,Yucheng Pan,Solaleh Mohammadi,Xinyan Liu,Kaiqing Zhang
机构: University of Maryland, College Park (马里兰大学学院公园分校)
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Multiagent Systems (cs.MA)
备注: 33 pages, 5 figures
Abstract:In large language model (LLM) steganography, an encoder hides a private message in an otherwise ordinary response, as text transmitted to a decoder, who attempts to recover it. The success of this private-message transmission is commonly measured through end-to-end (E2E) recovery, which checks whether the decoder correctly reconstructs the private message. However, E2E recovery cannot distinguish encoding failures from decoding failures or reveal how the private message is represented in the encoder’s transmitted text. To address these limitations, we introduce a rolewise decomposition framework that separately examines how private messages are represented in encoder transmitted texts, what decoders recover, and what deterministic verification can establish. Across 7,200 encoder responses, we identify 168 verifiable encodings missed by decoders and 706 responses containing private messages reconstructed through word-length or token-initial patterns not explicitly specified in the encoder’s instructions. These findings demonstrate that E2E recovery alone may obscure important distinctions between encoding and decoding, motivating their separate evaluation to better understand and mitigate the safety issues caused by LLM steganography.
[MA-16] Learning to Incentivise: Using Reinforcement Learning for Sustainable Urban Mobility
【速读】：该论文旨在解决传统交通管理中依赖收费与道路定价机制所引发的可及性争议及公众不满问题，尤其针对如芬兰等地因社会接受度低而禁止道路定价的现实困境。其核心问题是如何在不依赖直接收费的前提下，通过动态激励机制引导驾驶员主动选择替代路径，从而实现交通网络整体效率提升与环境效益优化。解决方案的关键在于采用多智能体强化学习（multi-agent reinforcement learning）框架，动态分配激励以协同优化总出行时间（Total Travel Time, TTT）与总碳排放量，利用交通仿真软件实时估算行程时间，并基于多目标加权函数平衡二者之间的权衡关系。研究结果表明，在预算无限且以最小化TTT为目标时，激励方案可使TTT降低16%；当目标转向减排时，CO₂排放减少9%；更重要的是，通过合理设置多目标权重，可在不牺牲任一指标的前提下同时优于基线水平，有效破解“时间-排放”间的固有矛盾，为可持续城市交通管理提供了可扩展的智能化决策支持。
链接: https://arxiv.org/abs/2610.30469
作者: Germán Pardo-González,Shaghayegh Vosough,Katerina Papadaki,Claudio Roncoli
机构: London School of Economics and Political Science (伦敦政治经济学院); Aalto University (阿尔托大学); KU Leuven (鲁汶大学)
类目: Multiagent Systems (cs.MA); Machine Learning (cs.LG); Systems and Control (eess.SY)
备注: 6 pages, 5 figures. Accepted version of a paper published in the Proceedings of the 2025 9th International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS), IEEE
Abstract:Traffic management has traditionally focused on toll-based and road pricing solutions. However, road pricing often raises concerns about accessibility and public dissatisfaction, leading to its prohibition in some places, such as Finland. This study optimises the dynamic allocation of incentives to drivers, encouraging them to reroute onto alternative (potentially longer) paths to achieve greater societal benefit, reduced total travel time (TTT) and total emissions in the transportation network, contributing to sustainable urban mobility. We employ a multi-agent reinforcement learning approach to dynamically assign incentives to drivers to reduce both TTT and emissions, with travel times estimated using traffic simulation software. We demonstrate that, with an unlimited budget and an objective of minimising travel time, the incentive scheme reduces TTT by 16%, compared to the dynamic User-Equilibrium (UE) with a budget equivalent to about 11% of the UE total time. When the goal is to minimise emissions, a 9% reduction in CO2 emissions is observed under an unlimited budget. We demonstrate a critical trade-off: minimising TTT leads to an increase in emissions, while prioritising emission reductions raises TTT. However, with the right combination of weights in the multi-objective function, both TTT and total emissions are improved beyond the baseline.
[MA-17] Agent ic Architecture of Information Systems. The AAM/5C Model: A Reference Model for Distributed Autonomy
【速读】：该论文旨在解决大型组织在面对数百个相互依赖组件时所遭遇的“认知复杂性壁垒”问题，即人类团队难以在如此规模下维持依赖关系图谱并以事件响应速度作出及时反应。其解决方案的核心在于提出AAM/5C框架，通过在七个逻辑层级（L1–L7）中实现五项核心能力（捕获、理解、选择、协调、控制），并基于单一授权规则R-AUT构建分权机制：概率规划器（L4）提出行动建议，治理层（L6）在签名策略下决策并生成带签名的评估证明，安全层（L5）据此颁发与该证明绑定的短期执行令牌，策略执行点在目标层实施动作前验证双重凭证。该设计通过显式保证契约确保两个关键假设（H7a, H8a）可构造性成立，其余八个假设（如日志完整性H7b、执行点不可绕过性H8b）虽可被证伪，但其反例协议在测量前已预先确定。该研究的贡献在于提出一种规范性参考模型，整合现有机制并引入三项待验证的新构念——分布式自主循环中的验证步骤、瞬态权限面及认知依赖平面，并与自适应计算、多智能体系统、企业架构及智能体协议（MCP、A2A）进行对比，提供Regos语言编写的授权策略与部分TLA+形式化共识规范，设定量化验证里程碑，但未报告实证结果。
链接: https://arxiv.org/abs/2610.30221
作者: Laurent Goujon(IBM France)
机构: 未知
类目: oftware Engineering (cs.SE); Multiagent Systems (cs.MA)
备注: 120 pages, 19 figures, 15 tables, 3 listings. Conceptual volume. English translation of the French original (version V7.83; HAL: hal-05820640; Zenodo: https://doi.org/10.5281/zenodo.23494375 ). This English edition is also on Zenodo: https://doi.org/10.5281/zenodo.23494664
Abstract:Large organizations increasingly face what this book calls the cognitive complexity wall: at the scale of hundreds of interacting components, maintaining the dependency map and reacting at the pace of incidents exceeds what a human team can deliver. This conceptual volume proposes AAM/5C, in which five capabilities (Capture, Comprehend, Choose, Coordinate, Control) are instantiated across seven logical layers, L1 to L7. A single authorization rule, R-AUT, separates who proposes, who decides and who executes: a probabilistic planner (L4) proposes; a governance layer (L6) decides under signed policies and records its verdict in a signed evaluation proof; a security layer (L5) issues a short-lived act token bound to that proof; a policy enforcement point verifies both before the target layer materializes the act. Under an explicit guarantee contract, two analytical statements (H7a, H8a) hold by construction. The other statements of its ten hypotheses (H1 to H10), including log completeness (H7b) and non-bypass of the enforcement point (H8b), are refutable, and the protocol that could refute them is fixed before any measurement. The contribution of the book is a prescriptive reference model that articulates, under this single authorization rule, known mechanisms and adds three constructs, all to be validated: the Validation step of the Distributed Autonomy Loop, the ephemeral authority surface and the cognitive dependency plane. The book positions them against autonomic computing, multi-agent systems, enterprise architecture and agent protocols (MCP, A2A), gives an authorization policy in Rego and a partial TLA+ specification of peer consensus, and sets out a validation program with quantified milestones. It reports no empirical results. This English edition is a translation of the French original, produced with the assistance of AI tools and validated by the author.
[MA-18] Consensus-Aware Co-design of Wireless Multi-Agent Control and Communication Systems
【速读】：该论文旨在解决自主多智能体控制系统（MACS）与无线通信系统之间因设计脱节而导致的性能瓶颈问题：传统MACS设计忽略通信与控制子系统的交互影响，而传统无线资源分配则基于最坏情况下的MACS需求预留资源，造成控制性能下降和无线资源利用效率低下。其解决方案的关键在于提出一种共识感知的协同设计方法，通过构建基于设备到设备（D2D）通信共享蜂窝频谱的系统模型，将智能体间通过有限带宽资源实现一致性（consensus）的过程纳入联合优化框架。具体地，该方法以最小化带宽消耗为目标，联合优化一致性拓扑选择与无线资源分配，同时保障多智能体控制系统的收敛性及蜂窝用户的服务质量（QoS）。尽管该问题为典型的混合整数非线性规划（NP-hard），但其最优解具有可分解的结构特性，从而支持一种具有多项式复杂度的两阶段顺序优化算法。仿真结果表明，在智能电网场景下，所提方法相较两种基线方案可实现24%-27%的带宽节省，同时确保一致性达成并满足所有控制与通信用户的QoS要求。
链接: https://arxiv.org/abs/2610.30141
作者: Atefeh Termehchi,Sajedeh Soleimani,Shiva Kazemi Taskooh,Mehdi Rasti,Ekram Hossain
机构: University of Manitoba (曼尼托巴大学); Amirkabir University of Technology (阿米尔卡比尔科技大学); University of Oulu (奥卢大学)
类目: Optimization and Control (math.OC); Multiagent Systems (cs.MA)
备注:
Abstract:Autonomous multi-agent control systems (MACSs) enable distributed coordination in emerging applications such as smart power grids and autonomous vehicle or drone networks. In many of these applications, inter-agent communication relies on wireless networks. However, conventional MACS designs neglect the interaction between the communication and control subsystems. At the same time, conventional wireless designs reserve communication resources based on worst-case MACS requirements. Consequently, control performance is degraded and wireless resources are utilized inefficiently. To address these challenges, we propose a consensus-aware co-design approach for MACSs and wireless communication systems. We introduce a system model in which agents achieve consensus by exchanging information through device-to-device (D2D) links that share spectrum with cellular users under limited bandwidth resources. Based on this model, we formulate a joint optimization problem for consensus-topology selection and wireless resource allocation to minimize bandwidth consumption while guaranteeing consensus and satisfying the quality-of-service (QoS) requirements of both MACSs and cellular users. Although the resulting problem is a mixed-integer nonlinear program (NP-hard in general), a structural property of the optimal solution enables a sequential two-stage optimization algorithm with polynomial complexity. Simulations on a smart power grid case study demonstrate that the proposed approach achieves 24-27% bandwidth savings compared with the two baselines while maintaining consensus and satisfying all QoS requirements of the control agents and the communication users.
自然语言处理
[NLP-0] Direct Pass@k Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning
【速读】： 该论文旨在解决生成式推理模型在测试时缩放（test-time scaling）过程中，如何高效筛选候选解池以平衡高质量与多样性之间的权衡问题。现有剪枝策略主要分为基于质量的方法（仅保留预测最终结果正确的部分解）和基于多样性的方法（选择遵循不同推理路径的部分解），但二者常存在探索与利用之间的冲突，且缺乏对这一权衡的显式建模。本文提出一种直接优化 pass@k 指标的剪枝方法，其核心在于选择一个部分解集合，以最大化至少有一个解在完整生成后得到正确答案的概率。该目标函数天然地同时奖励高质量的部分解并惩罚解之间的冗余性，从而无需额外的人工设计即可实现质量与多样性的自适应平衡。在八种开源推理模型上的实验表明，该方法显著优于现有的质量或多样性导向剪枝策略，尤其在保留中等数量部分解时性能提升最为明显，验证了直接建模质量-多样性权衡的有效性。
链接: https://arxiv.org/abs/2610.31667
作者: Mahdi Salmani,Adel Javanmard,Devansh Gupta,Meisam Razaviyayn
机构: University of Southern California
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Test-time scaling has emerged as a key inference strategy for reasoning models, where a model generates a large number of candidate solutions in the hope of finding at least one correct solution. Running all candidate solutions to completion can be computationally demanding, so recent work prunes the candidate pool after generating only a short prefix, committing further computation only to the surviving candidates. Existing pruning strategies divide broadly into two categories: Quality-based methods assign a score to each partial solution and discard those predicted to ultimately lead to an incorrect final solution, while diversity-based methods select partial solutions that follow different reasoning trajectories and are therefore likely to cover different final solutions. The tension between these two objectives is an instance of exploration versus exploitation, yet little to no prior work on pruning exposes this trade-off directly, let alone optimizes it. In this paper, we introduce a pruning procedure that addresses this trade-off directly by optimizing pass@ k over the selected solution pool. In particular, it chooses the pool to maximize the probability that at least one of its partial solutions, when continued to completion, yields a correct answer. Because this objective simultaneously rewards high-quality partial solutions and penalizes redundancy among them, quality and diversity are naturally balanced by the criterion itself rather than through separate ad-hoc design choices. On eight open-weight reasoning models, our method outperforms existing quality-based and diversity-based pruning strategies. The gains are particularly pronounced when the pruning procedure keeps a moderate number of partial solutions for continued generation, highlighting the value of directly balancing quality and diversity during pruning.
[NLP-1] Cost-Aware Abstention for LLM -Based IT Ticket Classification ICTAI2026
【速读】： 该论文旨在解决IT支持工单自动路由中因错误分配而导致的高成本问题，尤其关注在缺乏明确分类置信度时，如何通过成本感知的拒答（cost-aware abstention）机制降低整体处理成本。其核心解决方案的关键在于：在离线阶段基于校准数据显式定义成本函数（如正确路由成本 cok=0、人工转派成本 ctriage=1、错误路由成本 cwrong=5），选择并冻结一个最优路由策略（如使用TF-IDF与逻辑回归结合阈值0.60），随后在保留测试集上评估该固定策略的性能。实验表明，该方法将测试集平均成本从0.763降至0.472，相对减少38.1%，同时将错误路由率由15.3%降至3.5%。此外，研究还揭示了模型校准（calibration）与判别能力（discrimination）之间的可分离性，强调需同时报告正确性AUROC和风险覆盖率区域（tie-aware risk-coverage area）以全面评估性能。尽管生成式大语言模型（Generative AI）在某些提示配置下表现优于传统模型，但其在零样本/少样本场景下的泛化能力仍受限，且部分配置在断点概率0.80处会全部拒绝工单，凸显了成本敏感型拒答机制在实际部署中的重要性。本研究构建的是一个评估框架，而非通用的拒答算法或部署认证标准。
链接: https://arxiv.org/abs/2610.31657
作者: Sachin Gupta
机构: Independent Researcher
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted at the 38th IEEE International Conference on Tools with Artificial Intelligence (ICTAI 2026). 9 pages, 5 figures, 9 tables
Abstract:Wrong automatic routing of an IT support ticket can cost more than human triage. We evaluate cost-aware abstention as an offline, cost-conditional readiness assessment: select a routing policy on calibration data under explicit costs, then evaluate that frozen policy once on held-out test data. On a public 47,837-ticket, eight-class corpus, we compare four trained non-LLM classifiers with four large language models under zero- and few-shot prompting. Under the illustrative cost model (c_ok, c_triage, c_wrong) = (0, 1, 5), TF-IDF with logistic regression and a threshold of 0.60 selected on calibration reduces test cost from 0.763 to 0.472, a 38.1% relative reduction (paired ratio-bootstrap 95% CI [35.7%, 40.4%]), while lowering wrong routes from 15.3% to 3.5%. Frozen calibration-selected policies also beat both always-route and always-defer on held-out test data in 14 of 16 settings in a post-hoc cost-parameter analysis. The strongest prompted configuration reaches 0.477 accuracy on a stratified 1,000-ticket test subset, versus 0.879 for fine-tuned DistilBERT on the same tickets. Isotonic lowers expected calibration error without reliably improving discrimination: GPT-4o-mini few-shot has one confidence value before and after recalibration and correctness AUROC 0.500. We therefore report correctness AUROC and tie-aware risk-coverage area alongside ECE. At the analytically derived break-even probability 0.80, seven prompted configurations defer all tickets; Haiku few-shot routes only 19/1,000. An LLM-only agreement control and margin/entropy rejection baselines further separate calibration from discrimination. The findings are scoped to this dataset, prompt design, model endpoints, and illustrative costs; the work is an evaluation framework, not a new abstention algorithm or universal deployment certification.
[NLP-2] Low-Bit N-Gram Embeddings for Language Models
【速读】： 该论文旨在解决生成式AI（Generative AI）中n-gram嵌入表（n-gram embedding tables）带来的巨大内存开销问题，这类表虽能显著提升语言模型性能，但其参数量庞大（如DeepSeek-V4.1-Flash中高达1960亿参数），制约了模型的可扩展性与部署效率。解决方案的关键在于采用低精度量化（low-bit quantization）策略，系统性地研究不同量化方法及比特分配方案，将n-gram嵌入表压缩至四比特（4-bit）时，可在保持超过99%原始验证损失下降效果的前提下，实现约75%的内存节省。进一步发现，通过短周期微调可部分恢复因低精度导致的性能损失，且将节省的内存重新投入构建更大规模的嵌入表，反而能在更低内存占用下超越原全精度表的性能表现。该方法在多个主流模型家族（NanoChat、Llama、GPT-2）及现有大模型（DeepSeek-V4.1-Flash、Qwen3.8-Flash-Next）上均验证有效，实现47%-55%的内存节约，同时维持极小的困惑度上升（<1.2%）和下游任务性能基本不变。研究表明，低比特精度是降低n-gram嵌入内存需求并释放资源以增强模型性能的可行且高效路径。
链接: https://arxiv.org/abs/2610.31646
作者: Stephen Chung,Wenyu Du
机构: DualverseAI
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 24 pages, 10 figures, 10 tables
Abstract:Many recent open-weight language models use n-gram embedding tables to improve performance. These tables store learned representations of short token sequences, such as pairs or triples of consecutive tokens, which the model retrieves during processing. However, they can require substantial memory: DeepSeek-V4.1-Flash, for example, allocates 196 billion parameters to its Engram n-gram memory. We therefore investigate whether this memory cost can be reduced while preserving the performance improvement that the tables provide. We propose storing n-gram embeddings at low precision and systematically investigate different quantization methods and bit allocations across tables. Across nanochat-, Llama-, and GPT-2-family models trained from scratch, four-bit tables save approximately 75% of table memory while preserving more than 99% of the validation-loss improvement from adding the tables. Using fewer bits leads to a larger drop in performance, although we find that a short fine-tuning phase recovers a significant part of this loss. In addition, reinvesting the saved memory in larger tables can improve performance beyond that of the original full-precision tables while still using less memory. These table memory savings also extend to large existing models: on DeepSeek-V4.1-Flash and Qwen3.8-Flash-Next, low-bit tables save 47-55% of table memory, with perplexity increases of less than 1.2% and largely preserved downstream task performance. These findings establish low-bit precision as a practical way to reduce the memory requirements of n-gram embeddings and use the saved capacity to improve model performance.
[NLP-3] Can an Agent Debug a Web Corpus? Agent ic Auditing of Slovak FineWeb2 EMNLP2026
【速读】： 该论文旨在解决多语言数据构建流水线（如FineWeb2）在跨千种语言应用统一处理策略时，未能识别和应对特定语言独特失败模式的问题。其核心挑战在于：现有方法依赖大语言模型（LLM）进行分类，但这种方式成本高昂且难以定位具体缺陷的成因与分布。本文提出一种创新解决方案——将大语言模型作为“语料库调试器”而非分类器使用，通过一个工具型智能体（tool-using agent）以只读、预算受限的API方式探索语料库，并将发现转化为可执行测试，由母语者验证。在对FineWeb2中2650万条斯洛伐克语文档的实验中，该方法以每次$2.21–$2.84的成本，在三次运行中识别出七项缺陷，涵盖四种现象，其中最大问题为一个包含1,408,437篇（占5.32%）机器生成内容的农场，其中86.6%甚至通过了更严格的过滤标准。盲测标签支持其中六项发现（缺陷率0.53–1.00，对比随机对照组0.08），一项未被支持。所生成的测试套件可在无需GPU的情况下仅用0.7秒扫描全量语料，远低于传统基于LLM分类的每轮$12k–$90k成本。尽管静态采样基线在验证更多选择器方面表现更优，且成本不足一半，但该智能体仍提供精确的缺陷出现频率、经采样文档验证的精度估计、流水线遗漏各子集的原因分析以及可复用的测试逻辑。研究团队已开源该框架、斯洛伐克语测试套件、验证数据及智能体轨迹。
链接: https://arxiv.org/abs/2610.31636
作者: Marek Šuppa,Ivan Vykopal,Jaroslav Kopčan
机构: Comenius University in Bratislava (布拉迪斯拉发考门纽斯大学); Cisco Systems(思科系统); Kempelen Institute of Intelligent Technologies (KInIT) (肯佩伦智能技术研究所); Brno University of Technology (布尔诺理工大学)
类目: Computation and Language (cs.CL)
备注: Accepted to WaC Workshop at EMNLP 2026
Abstract:Multilingual pipelines such as FineWeb2 apply one recipe across a thousand languages, leaving the failure modes of individual languages unanticipated. We use an LLM once, as a corpus debugger rather than a classifier: a tool-using agent explores the corpus through a read-only, budgeted API and compiles each finding into an executable test, which a native speaker then validates. On the 26.5M-document Slovak portion of FineWeb2, three runs costing \ 2.21-\ 2.84 each found seven findings over four phenomena. The largest is a machine-generated content farm of 1,408,437 documents (5.32%) that passed the pipeline’s filters, 86.6% of it passing even stricter thresholds. Blinded labels support 6 of the 7 findings (defect rates .53-1.00 against .08 in a random control) and do not support one. The compiled suite scans the full corpus in 0.7 s of wall-clock time without a GPU, against \ 12k-\ 90k per LLM classification pass. The best static sampling baseline validates more selectors than any agent run, at under half the cost. The agent contributes exact prevalence, a precision estimate checked against sampled documents, an account of why the pipeline missed each slice, and a reusable test. We release the framework, Slovak test suite, validation data and agent trajectories.
[NLP-4] Bilingual Language Models Capture Cross-Linguistic Influence in Second-Language Reading
【速读】： 该论文旨在解决现有单语语言模型（LM）突现性（surprisal）在预测第二语言（L2）阅读时间时无法有效捕捉母语（L1）对非母语阅读行为影响的问题。其核心解决方案在于探究双语语言模型中的跨语言迁移是否与读者的L1驱动认知偏差相一致。研究通过分析来自12种不同L1背景的读者在阅读英语文本时的眼动数据，对比了与读者L1匹配的双语模型、单语英语基线模型及其他非匹配L1训练的双语模型的表现。结果表明，当双语模型的训练语言与读者的L1一致时，其对L2阅读时间的预测能力显著提升；这种优势主要体现在眼动后期指标中，且受L1-L2文字系统重叠程度调节，在中等语言水平读者中表现最为突出。该研究为双语语言模型能够捕获跨语言影响提供了实证支持，并揭示了L1对L2阅读行为的深层作用机制。
链接: https://arxiv.org/abs/2610.31620
作者: Abderrahmane Issam,Irene E. Winther,Stefan L. Frank
机构: Maastricht University (马斯特里赫特大学); Radboud University (奈梅亨大学)
类目: Computation and Language (cs.CL)
备注: Under review
Abstract:While monolingual language model (LM) surprisal is widely established as a predictor of reading times, it cannot capture the non-native (L2) reading dynamics that are shaped by the reader’s first language (L1). In this work, we investigate whether the cross-lingual transfer within bilingual models aligns with human L1-driven biases during L2 reading. We evaluate bilingual models on eye-tracking data of readers from 12 diverse L1 backgrounds reading English texts. To isolate L1-specific transfer, we compare bilingual models trained on an L1 that matches the readers’ L1 background to both a monolingual English baseline and bilingual models trained on other L1s. We demonstrate that matching the readers’ L1 improves the prediction of L2 reading times. This predictive advantage is concentrated in late-stage eye-tracking measures, modulated by L1-L2 script overlap, and peaks among intermediate-proficiency readers. These findings provide evidence that bilingual LMs capture aspects of cross-linguistic influence, and that such influence is present in L2 reading behavior.
[NLP-5] How Ground Truth Shapes Retrieval Scores: Evidence-Aware Evaluation of a Retrieval-Augmented Generation System for Breast Lesions
【速读】： 该论文旨在解决临床检索增强生成（RAG）系统在乳腺病灶管理文献检索中，检索评分对标注覆盖度依赖性过强的问题。其核心解决方案在于提出一种基于证据感知的检索评估协议，通过引入更精细的证据跨度（evidence span）标注机制与重叠分块策略，提升检索结果的相关性评估可靠性。关键创新点包括：采用混合检索方法（融合密集嵌入、BM25及交叉编码器重排序），结合384词元分块与20%重叠配置，在18个查询的开发基准上实现高召回率（Recall@3=0.94）、高R-Precision和nDCG@3等指标；同时，通过扩展标注范围（证据跨度由25增至37，可接受分块目标由36增至51），在不改变检索排名的前提下提升了系统对真实证据的识别能力，凸显了标注覆盖度对评估结果的影响，强调需进行独立验证（如放射科医生对证据充分性的评估）以确保评估的有效性。
链接: https://arxiv.org/abs/2610.31606
作者: Miguel Mochizuki Silva,Samuel Ismael,Mateus C. Dias,Thaís Gaudêncio,Yuri Malheiros,Barufaldi
机构: Universidade Federal da Paraíba (UFPB), João Pessoa, PB, Brazil; Department of Radiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 18 pages, 3 figures, 6 tables
Abstract:Purpose: To describe an evidence-aware retrieval evaluation protocol for a clinical retrieval-augmented generation (RAG) system over breast lesion management literature, addressing the dependence of retrieval scores on annotation coverage. Methods: An 18-document corpus and an 18-query development benchmark, with one Yes/No or WH-question per document, were used. Retrieval combined dense embeddings, BM25, and cross-encoder reranking. Chunk sizes of 128 to 1024 tokens, overlaps of 0 to 20%, and four retriever configurations were evaluated. Relevance was defined by evidence spans mapped to all containing chunks, treating overlapping chunks with equivalent evidence as relevant. Retrieval was evaluated using R-Precision, Recall@K, MRR, and nDCG@K. Separately, generation was assessed on a 512-token index at K=1 to 5 using DeepEval claim-level metrics, with deepseek-v4-pro as the LLM judge, and BLEU/ROUGE metrics. Results: The selected hybrid retriever, using 384-token chunks and 20% overlap, retrieved accepted evidence in the top three results for 17 of 18 queries. Recall@3, R-Precision, MRR@3, and nDCG@3 each reached 0.94. Ground-truth curation increased evidence spans from 25 to 37 and accepted chunk targets from 36 to 51 without changing retrieval rankings. Conclusion: Cross-encoder reranking achieved the highest retrieval scores. Expanded relevance annotations increased recognized evidence without changing rankings, highlighting annotation sensitivity. The small development set and retrieval-informed curation require independent validation, including radiologist assessment of evidence sufficiency.
[NLP-6] Would an LLM Match into a Residency? Uncertainty Reasoning and Image Use of Open Language Models on the Spanish MIR Examination
【速读】： 该论文旨在解决生成式 AI（Generative AI）在医学影像诊断场景中应用时的可信度与可靠性问题，尤其关注开放模型在实际医疗系统部署中的关键挑战：模型是否具备自我识别错误的能力、医学微调能否提升其判断置信度的准确性、其正确答案是否基于合理的临床推理，以及模型是否真正利用了输入的影像数据。研究针对四个参数规模为8至270亿的开源语言模型，在西班牙医学影像住院医师考试（Spanish MIR residency examination）的8个版本共1,552道题目上进行评估，严格遵循考试评分规则。结果显示，尽管较大模型在排名上可进入前600至3,000名候选人范围，但其置信度无法有效区分正确与错误答案；医学微调虽降低了模型的过度自信，却以牺牲准确率为代价，且未改善置信度与答案正确性之间的分离能力。此外，13%至38%的正确回答因不具临床合理性的排除干扰项理由而得出，且移除或替换输入图像后模型性能保持不变，表明其决策过程并未真正依赖图像信息。因此，该研究的核心发现是：当前开源模型在医学诊断任务中仍缺乏可靠的自我评估能力，且其“正确”答案可能源于非临床逻辑或未使用关键输入信息，这严重制约了其在真实医疗环境中的可信部署。
链接: https://arxiv.org/abs/2610.31586
作者: Miguel Fernandez-de-Retana,Unai Zulaika,Ruben Sanchez-Corcuera,Aritz Bilbao-Jayo,Aitor Almeida
机构: University of Deusto (德乌斯托大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (cs.LG)
备注: 13 pages, 8 figures
Abstract:Accuracy on medical examinations is common evidence offered for the clinical use of language models, yet a health system that intends to run an open model on its own hardware needs to know considerably more: whether the model recognizes when it is likely to be wrong, whether medical fine-tuning improves this ability, whether its correct answers rest on sound clinical reasoning, or whether it actually uses the images it is given. We examine these questions for four open models of 8 to 27 billion parameters on the Spanish MIR residency examination, scoring 1,552 questions from eight editions under the exam’s own rules. Answering every question, the three larger models would rank among roughly the first 600 to 3,000 candidates, with 41 to 43 specialties still open to them. Abstention, however, never pays, because even their least confident answers are right more often than the 25% break-even. Medical fine-tuning reduced overconfidence at the cost of accuracy, without improving how well confidence separates right from wrong answers. Moreover, 13-38% of correct answers rejected a distractor for a clinically unsound reason that confidence did not reflect, and removing or swapping the provided contextual images left accuracy unchanged.
[NLP-7] Diffusion Language Models for Perplexity-Aware Black-Box Attacks
【速读】： 该论文旨在解决现有连续空间攻击方法中因依赖非学习的随机映射（random mapping）生成提示而导致提示不自然、易被输入防护机制检测，且风险评估结果缺乏真实语言分布支撑的问题。其解决方案的关键在于将原有的连续潜空间框架与一个经过学习的语言生成分布——连续扩散语言模型（dLLM, continuous diffusion language model）直接耦合，从而摒弃了人工构造的随机映射，同时保留标准高斯潜空间结构及下游采样或优化流程。实验表明，在基于采样的RareTrap和基于优化的ThinkTrap两种设置下，字符级困惑度分别从10³–10⁷降至3–5，以及从20–28降至4.3–5.4，显著提升了提示的自然性；同时，不同映射方式下的模型风险估计值可相差一个数量级，验证了所提方法在生成更可信、更隐蔽恶意提示方面的优势。
链接: https://arxiv.org/abs/2610.31558
作者: Ali Al-Lawati,Elsayed Eshra,Dongwon Lee,Suhang Wang
机构: The Pennsylvania State University (宾夕法尼亚州立大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Machine Learning (stat.ML)
备注:
Abstract:\emphContinuous-space attacks find prompts that trigger malicious behaviors, such as overthinking, in black-box LLMs by sampling or optimizing over a low-dimensional Gaussian latent space and decoding it to text through a \emphrandom mapping and nearest-token lookup. Because the random mapping is not learned from language, its prompts can be highly unnatural, and as such easily detectable through input guardrails, and the risk estimates they yield are not grounded in natural prompts. In this paper, we demonstrate that the same continuous latent-space framework can be coupled directly to a learned generative distribution over language, a continuous diffusion language model (dLLM), which removes the artificial random mapping while preserving the standard-Gaussian latent formulation and the downstream sampling or optimization procedure. We evaluate this approach under two settings: sampling-based estimation using \emphRareTrap and optimization-based search using \emphThinkTrap. We find that character-level perplexity drops from 10^3 – 10^7 to 3–5 in RareTrap and from 20–28 to 4.3–5.4 in ThinkTrap. We also find that model risk estimates can differ by an order of magnitude across the two mappings.
[NLP-8] An Empirical Study of Recent Large Language Models for Knowledge Graph to Text Generation on Standard Benchmarks
【速读】： 该论文旨在解决知识图谱到文本（KG-to-text）生成任务中评估方法与模型性能之间的不一致性问题，尤其关注传统基于参考文本的评估指标与新型基于图结构的无参考评估方法之间的差异。其核心挑战在于：尽管近年来大语言模型（LLM）在生成质量上表现优异，但现有主流评估方式（如BLEU、ROUGE等参考基准指标）可能无法全面反映生成内容对知识图谱事实的忠实度与逻辑一致性。解决方案的关键在于引入一种图结构引导的无参考评估框架，利用大语言模型进行基于知识图谱事实的自动验证，从而更准确地衡量生成结果与输入图谱的一致性。实验表明，尽管基于参考的指标显示LLM与微调的编码器-解码器模型差距缩小，但图基评估揭示了提示型LLM在减少矛盾陈述和三元组遗漏方面显著优于微调模型，且后者在图规模增大时错误率上升更快，说明参考基准指标可能忽略生成质量中的关键缺陷。这一发现强调了采用图结构约束的评估方法对于真实性能判断的重要性。
链接: https://arxiv.org/abs/2610.31529
作者: Priyanka,Renaud Delbru,Souhail Bakkali,Mickaël Coustaty,Antoine Doucet
机构: La Rochelle Université(拉罗谢尔大学); Siren(西伦); Université de Rennes(雷恩大学)
类目: Computation and Language (cs.CL)
备注: 18 pages, 6 figures, 17 tables, Preprint, Under review
Abstract:Knowledge graph-to-text (KG-to-text) generation verbalizes a set of knowledge graph triples into fluent natural language. While fine-tuned encoder-decoder models have long been the standard approach, recent LLMs are becoming increasingly competitive. We benchmark recent LLMs against fine-tuned encoder-decoder models, using reference-based metrics and graph-grounded reference-free LLM-based evaluation validated against human annotations. We observed that recent LLMs substantially narrow, and in several settings close, the gap with fine-tuned models on reference-based metrics, while performing better on graph-grounded metrics. LLM-based evaluation scores show limited correlation with reference-based metrics, suggesting that the two approaches capture different aspects of generation quality. Graph-grounded evaluation reveals that prompted models produce fewer contradictions and triple omissions than fine-tuned models, and these errors increase more rapidly with graph size for fine-tuned models, highlighting differences between the two approaches that reference-based metrics may not fully capture.
[NLP-9] he Geometry of Metaphor: Probing and Steering Metaphor Representations in Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中隐喻（metaphor）认知表征的可解释性及其可控性问题。隐喻作为人类认知的核心特征，其在模型中的内在表示机制尚不明确，且缺乏有效手段对其进行干预与调控。论文提出的关键解决方案是采用机械可解释性（mechanistic interpretability）技术，识别不同架构与规模的LLMs中与隐喻相关的关键因果子空间（causal subspaces）。研究发现，通过操纵这些子空间，可在分类任务中对不同词性的输出进行定向引导，并在生成任务中有效调节模型输出的隐喻程度，证明该方法具备对模型隐喻表达进行精准控制的能力。
链接: https://arxiv.org/abs/2610.31520
作者: Shihui Li,Mette Andersen,Martha Lewis
机构: Institute for Logic, Language, and Computation(逻辑、语言与计算研究所); University of Amsterdam(阿姆斯特丹大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: BlackboxNLP 2026: The 9th Workshop on Analyzing and Interpreting Neural Networks for NLP
Abstract:Metaphor is a crucial aspect of human cognition. As such, we would like to understand how it is represented in large language models (LLMs) and how it might be controlled. We use mechanistic interpretability techniques to identify causal subspaces in a range of LLMs covering different architectures and sizes. We show that we are able to steer output on classification tasks across different parts of speech. We are also able to steer output on generation tasks, indicating that these techniques can be used to control the metaphoricity of LLM output.
[NLP-10] owards User Simulators That Stay On Task And Still Sound Natural
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）在长时间、多轮交互任务中性能退化的问题，核心挑战在于构建能够同时忠实遵循复杂用户意图并具备真实对话行为特征的用户模拟器。现有方法存在明显权衡：基于提示的助手虽能较好遵循复杂意图，但对话行为缺乏自然性；而专用用户模拟器虽更贴近真实对话模式，却难以长期保持对复杂意图的忠实度。本文提出一种强化学习（Reinforcement Learning, RL）框架，通过联合奖励机制同时优化任务意图遵循度与对话自然性。其关键创新在于采用轻量级判别器，在交替对抗训练过程中对模拟器的自然性进行量化评估，从而实现对“自然性”的可操作化建模。基于此方法，研究团队发布了UserLM-2.0模型，能够在46.2%的情况下生成既符合任务要求又具备自然对话特征的模拟对话，显著优于前代最强模型（14.1%），提升超过三倍。
链接: https://arxiv.org/abs/2610.31512
作者: Mustafa Omer Gul,Philippe Laban,Tanya Goyal,Jennifer Neville
机构: Cornell University (康奈尔大学); Microsoft Research (微软研究院)
类目: Computation and Language (cs.CL)
备注: 30 pages, 16 figures, 6 tables. In submission
Abstract:Users increasingly rely on AI systems for extended, multi-turn tasks, yet evidence indicates that LLM-based system performance degrades over prolonged interaction. To evaluate and improve AI assistants in such settings, we need user simulators that faithfully follow complex user intents while exhibiting realistic conversational behavior. Existing approaches manifest a trade-off: prompted assistants can follow complex user intents but lack behavioral fidelity, while specialized user simulators better approximate conversational behavior but struggle to remain faithful to complex intents. In this paper, we propose a reinforcement learning (RL) recipe that jointly rewards task intent adherence and behavioral naturalness. A crucial contribution of the method is operationalizing naturalness with a lightweight discriminator trained in an alternating, adversarial loop with the user simulator. We release the resulting model, UserLM-2.0, which produces valid simulated conversations (i.e., meeting both task-faithfulness and naturalness criteria) 46.2% of the time, over three times the rate of the strongest prior user simulator (14.1%).
[NLP-11] Measuring the Verifiers Floor: How Much Apparent Rare-Word Text-to-Speech Error Is ASR Error? ICASSP2027
【速读】： 该论文旨在解决文本到语音（TTS）系统评估中存在的偏差问题，即传统评估方法依赖自动语音识别（ASR）系统对合成语音进行转写，导致词错误率（WER）不仅反映TTS系统的性能，还包含ASR系统的识别误差。研究发现，在人类录音中，现有ASR系统对极罕见词汇的漏识率高达33%-67%（在LibriSpeech和FLEURS数据集上），且三类识别器均存在22%-42%的漏识率。通过引入多个不同家族的ASR模型进行交叉验证，可将部分极罕见词汇的误判率显著降低（如非常罕见的LibriSpeech词汇从10%-13%降至3%-5%），但这些词汇仍是评估中最难处理的部分。然而，若忽略在人类语音中被识别器遗漏的词汇，则会产生系统性偏差——其他识别器在这些词汇上检测出的TTS错误频率远高于被遮蔽的词汇（4%-6% vs. 0.2%-0.3%），并纠正了约一半原被掩盖的错误。由于剩余经人工判定的错误中约有一半可能为拼写变体，因此论文强调必须辅以听觉测试以获得更准确的评估结果。其解决方案的关键在于：采用多模型、多族系的ASR交叉验证机制，并结合人工听觉测试，以减少评估过程中的识别器偏差，从而更真实地反映TTS系统的实际表现。
链接: https://arxiv.org/abs/2610.31470
作者: Kirill Borodin,Vasilii Kudryavtsev,Nikita Vasiliev,Grach Mkrtchian
机构: 未知
类目: ound (cs.SD); Computation and Language (cs.CL)
备注: Submitted to IEEE ICASSP 2027
Abstract:Text-to-speech (TTS) is usually evaluated by transcribing its output with an automatic speech recognition (ASR) system, so the word error rate also counts recognizer errors. We measure these errors on human recordings of the test sentences. On LibriSpeech and FLEURS, the evaluation recognizer misses 33-67% of very rare words even in human speech, and all three recognizers miss 22-42%. Checking every word with two more recognizer families lowers the apparent TTS error on words that at least one recognizer hears two- to three-fold (very rare LibriSpeech words: from 10-13% to 3-5%), yet very rare words remain the hardest. Ignoring words the recognizer misses in human speech is biased: the other recognizers find TTS errors far more often on these words (4-6% vs. 0.2-0.3%) and overturn about half of the errors this mask keeps. As half of the remaining adjudicated errors may be spelling variants, a listening test is needed.
[NLP-12] From Structured Records to Patient Narratives: Preserving Clinical Alignment in LLM -Generated Health Summaries
【速读】： 该论文旨在解决可穿戴与数字健康系统生成的纵向生理数据在向患者传递时，如何实现临床信息向可读性语言有效转化的问题。其核心挑战在于：尽管生成式人工智能（Generative AI）能够将结构化临床数据转化为自然语言摘要，但在此过程中易产生语义漂移（semantic drift），导致临床意义失真。为此，论文提出一种基于提示的约束生成框架，通过结合结构化证据条件化与检索增强生成（Retrieval-Augmented Generation, RAG），以实现证据驱动的健康摘要生成。该方案的关键在于引入显式的证据接地约束机制，在保证生成内容与原始医学证据一致性的基础上，显著提升临床语义对齐度，尤其在不确定性表达的保留方面表现突出；同时，在与检索到的证据进行忠实度验证后，整体上保持或提升了生成内容的忠实性，但遗漏问题仍依赖于具体模型表现。研究还发现，提升临床准确性的同时会带来阅读难度增加和医学术语使用增多的副作用。因此，该研究强调在开发面向患者的医疗生成式AI系统时，必须协同评估生成内容的忠实性、不确定性保留能力及可读性。
链接: https://arxiv.org/abs/2610.31467
作者: Bayram Ayadi,Florian Geissler,Johanna Schmidhuber
机构: University of Passau (帕绍大学); Fraunhofer Institute for Cognitive Systems IKS (弗劳恩霍夫认知系统研究所IKS)
类目: Computation and Language (cs.CL)
备注: 18 pages, 3 figures
Abstract:As wearable and digital health systems increasingly generate longitudinal physiological records, communicating these observations in patient-accessible language remains a significant challenge. Large Language Models (LLMs) offer a promising mechanism for translating structured clinical information into readable narratives, but this process may introduce semantic drift, where clinically precise meaning is distorted during simplification. We investigate evidence-grounded health summarization for menstrual health monitoring using a constrained generation framework that combines structured evidence conditioning with Retrieval-Augmented Generation (RAG) using a prompt-based constrained generation framework. We evaluate semantic alignment using Natural Language Inference (NLI), atomic factuality metrics, and readability measures. Across three LLMs and 4,500 generated summaries, explicit grounding constraints consistently improve clinical alignment, particularly by improving uncertainty preservation and, once faithfulness is evaluated against retrieved evidence, generally preserving or improving faithfulness, while effects on omission remain model-dependent. We also find that improved clinical alignment comes at the cost of increased reading difficulty and medical terminology. These findings highlight the need to jointly evaluate faithfulness, uncertainty preservation, and readability when developing patient-facing medical LLM systems.
[NLP-13] Multi-Token Prediction and Alternative Tokenization Improve Learning Latent Multi-Hop Graph Search at Scale DATE
【速读】： 该论文旨在解决在未见过的图结构中，利用语言模型从起点节点 $ s $ 到终点节点 $ t $ 生成最短路径的问题。现有研究发现，随着图规模增大，语言模型难以有效学习此任务。本文指出该失败现象源于两个独立因素：监督信号设计与节点标记化（tokenization）方式。其解决方案的关键在于：首先，理论分析并实证验证了多令牌预测（Multi-Token Prediction, MTP）的有效性，且其优势与图的拓扑结构无关；其次，提出替代性的节点标记化策略，证明上述两个因素具有正交性。通过结合MTP与优化的标记化方法，该研究突破了先前工作所观察到的性能“缩放墙”，实现了在远超以往极限的图规模下稳定、一致地学习最短路径生成任务。
链接: https://arxiv.org/abs/2610.31439
作者: Arvid Frydenlund,FuTe Wong,Katherine Lambert,Frank Rudzicz
机构: University of Toronto(多伦多大学); Vector Institute(向量研究所); Manulife Applied AI Research(加拿大宏利应用人工智能研究)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Don’t be intimidated by the page length :)
Abstract:Given an unseen graph, a start node, s , and an end node, t , can a language model generate the shortest path from s to t ? Prior work has found that models fail to learn this task as the graphs scale. We show that this failure has two causes: the supervision signal and the tokenization. First, we theoretically establish and empirically demonstrate the effectiveness of multi-token prediction (MTP), and show that the benefits of MTP are topologically agnostic. Second, we introduce alternative tokenizations and show that the two causes are orthogonal. Using these methods, we break through the scaling wall observed by prior work, consistently and stably learning to solve the task at scales far beyond previous limits.
[NLP-14] How Well Do Decision Models Transfer to Vietnamese? An Empirical Study of Text Classification
【速读】： 该论文旨在解决生成式 AI 在越南语零样本文本分类任务中性能评估不足的问题，尤其关注专用决策模型在面对越南语语言特性时的适应性与鲁棒性。其核心解决方案的关键在于系统评估四类专用决策模型在13个分类任务中的表现，涵盖英语、越南语及双语标签描述，并考察其对越南语拼写变体（如去声调、错别字、网络用语）的敏感性。研究发现，标签描述的语言选择显著影响模型性能，且去声调处理导致的性能下降最为严重（平均宏观F₁值降低9.51–15.47分），远超其他扰动类型，凸显了在构建面向越南语的零样本分类系统时，需优先考虑标签描述语言策略与越南语正字法变异的建模。
链接: https://arxiv.org/abs/2610.31422
作者: Phat Huu Ha,Nguyen Xuan Phuc,Nong Nhut Quy,Trinh Tran Tran,Nguyen Minh Triet,Dang Van Thin
机构: University of Information Technology, Ho Chi Minh City, Vietnam(胡志明市信息科技大学); Vietnam National University HoChiMinh City, Vietnam(越南国家胡志明市大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Recent advances in specialized decision models have expanded the possibilities for zero-shot classification, yet their effectiveness on Vietnamese remains underexplored. This study evaluates four specialized decision models across 13 classification settings spanning six task groups, examining zero-shot performance, sensitivity to label-description language, and robustness to Vietnamese linguistic variations. The evaluation considers English, Vietnamese, and bilingual label descriptions, alongside common linguistic perturbations involving teencode, typographical errors, and diacritic removal. Under English label descriptions, JEV achieves the strongest overall performance, ranking first in 11 classification settings. The effects of label-description language vary across models and datasets, with bilingual descriptions providing no consistent improvement. Robustness experiments reveal that diacritic removal causes the most substantial performance degradation, reducing mean macro- F_1 by 9.51–15.47 points across model configurations, considerably more than teencode or typographical errors. These findings highlight the importance of label-description choices and Vietnamese orthographic variation when evaluating specialized decision models for zero-shot text classification.
[NLP-15] How Depth Growth and Looping Shape In-Context Learning in Language Models
【速读】： 该论文旨在解决生成式语言模型在上下文学习（In-Context Learning, ICL）中如何通过深度增长（depth-growth）与循环结构（looping）实现计算复用，进而影响其学习能力的形成机制问题。核心挑战在于理解这种参数复用机制如何塑造模型对上下文信息的依赖性、提升其在少样本场景下的推理能力，并揭示其内在的归纳偏置（inductive bias）。解决方案的关键在于：通过系统实验分析不同规模预训练模型（最高达30亿参数）在受控任务中的表现，发现采用深度增长或循环结构的模型相比全深度密集型基准模型，更早习得并更强保留上下文学习能力；进一步通过注意力头追踪与消融实验表明，这类模型显著依赖于“诱导性注意力头”（induction heads），且这些头部在训练过程中随深度增长被跨层传播，从而强化了对上下文信息的利用。研究结果表明，（软）参数复用不仅促进了模型对上下文的学习倾向，还重塑了支持该能力的注意力机制结构，揭示了一种共享的、以从上下文中学习为导向的归纳偏置。
链接: https://arxiv.org/abs/2610.31421
作者: Ferdinand Kapl,Kaitlin Maile,Stefan Bauer,Johannes von Oswald
机构: Technical University of Munich (慕尼黑工业大学); Helmholtz AI (海姆霍兹人工智能); Munich Center for Machine Learning (慕尼黑机器学习中心); Google(谷歌), Paradigms of Intelligence Team
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Looped and depth-grown language models reuse computations across depth and have attracted interest for their reasoning capabilities. Building on prior suggestive evidence of improved in-context-learning (ICL), we investigate how such reuse favors ICL, including how it shapes the acquisition, retention, and mechanisms of this ability. Our analysis spans pretrained language models up to 3B parameters and controlled tasks. The experiments provide evidence for a shared inductive bias toward learning from context: grown and looped models rely more on information in context and, compared with a full-depth dense reference, acquire ICL earlier in controlled tasks and retain it more strongly when learning fixed associations in the weights is also possible. Tracing attention heads through training reveals that depth growth propagates early induction heads across layers. Ablations show that grown and looped models rely more strongly on induction-ranked heads for both few-shot performance and later-token prediction. Together, these findings provide evidence for a shared inductive bias toward learning from context and suggest that (soft) parameter reuse shapes how this ability develops and which attention mechanisms support it.
[NLP-16] SyRA-CS: Synthetic-to-Real Adaptation for Code-Switched ASR with Meta-Learning
【速读】： 该论文旨在解决多语言语码转换（Code-switched, CS）自动语音识别（ASR）中真实语码转换语音数据稀缺、而合成语音与真实语音存在分布差异的问题。现有方法在利用合成数据进行模型预训练或持续微调时，难以有效迁移到真实场景。其解决方案的关键在于提出一种基于Reptile元学习的成对合成到真实适应框架SyRA-CS，通过在对应语言对的合成语码转换FLEURS子集上进行元训练，获得对适应任务更具鲁棒性的初始化参数，随后仅需1小时标注的真实目标语音即可完成高效适配。实验表明，该方法在阿拉伯语-英语、法语-英语、德语-英语和印地语-英语等多组语言对上，相较持续微调在词错误率（WER）上实现了17.7%（XLS-R模型）至5.3%-5.8%的相对宏平均降低，在字符错误率（CER）上也有显著提升，验证了合成数据在提供适应友好初始化方面的有效性，且性能增益在不同架构、语言对和评估指标间保持一致但大小各异。
链接: https://arxiv.org/abs/2610.31415
作者: Fadel Thior,Ahmer Tabassum,Haania Siddiqui,Hanan Aldarmaki,Thamar Solorio
机构: Mohamed bin Zayed University of Artificially Intelligence (穆罕默德·本·扎耶德人工智能大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Code-switched (CS) automatic speech recognition (ASR) remains difficult because real CS speech is scarce and synthetic speech does not fully match it. We investigate whether synthetic CS speech is more useful for learning an adaptation-friendly initialization than for ordinary continual fine-tuning. We propose SyRA-CS, a pair-specific synthetic-to-real adaptation framework based on Reptile meta-learning. For each language pair and pretrained model, we meta-train on the corresponding synthetic CS-FLEURS subset and then adapt using one hour of labeled real target speech. We evaluate four architectural settings: Encoder-CTC, FastConformer-RNN-T, Attention Encoder-Decoder (AED) and Speech-LLM. Experiments cover on Arabic-, French-, German-, and Hindi-English, against synthetic-only transfer, real-only fine-tuning, and continual fine-tuning. SyRA-CS improves over continual fine-tuning on WER in all 16 model-pair comparisons and on CER in 15 of 16, with relative macro WER reductions of 17.7% for XLS-R and 5.3-5.8% for the other three models. The gains hold across the tested models but vary in size across architectures, language pairs and metrics.
[NLP-17] SQuAT: Spherical Lattice Vector Quantization-Aware Training
【速读】： 该论文旨在解决生成式人工智能（Generative AI）中大规模语言模型（LLMs）在低比特量化（low-bit quantization）下的性能损失问题，尤其聚焦于基于格点的向量量化（lattice-based Vector Quantization, VQ）在量化感知训练（VQ-QAT）中的行为不明确与可扩展性差的问题。其核心挑战在于如何在保持高压缩率的同时，维持模型在量化后的表达能力与训练稳定性。解决方案的关键是提出SQuAT（Shape-Gain Vector Quantization with Separate Straight-Through Estimators），通过将权重分解为方向（shape）和幅度（gain）两部分：方向分量采用基于E8格点的球面码本进行量化，而幅度分量则独立使用低比特标量量化器处理；同时，针对方向和幅度分别设计独立的直通估计器（STE），从而在反向传播中诱导切空间（tangent-space）更新，隐式地实现球面（Riemannian）优化，有效捕捉权重的几何结构特性。此外，为支持超大规模模型训练，论文还开发了一种分布式精确最近邻分配方案，仅需通信离散索引，显著降低通信开销，使训练时间缩短12.8倍，并成功将VQ-QAT扩展至100亿词元级别的训练规模。实验表明，SQuAT在多个指令微调的LLaMA-3.2及Qwen系列模型上均优于标准VQ-QAT，且在相同量化精度与初始化条件下达到当前最优（SOTA）的量化感知训练准确率。
链接: https://arxiv.org/abs/2610.31412
作者: Sebastian Loeschcke,Ivan Koryakovskiy,Tycho F. A. van der Ouderaa,Riccardo Del Chiaro,Mart van Baalen,Paul N. Whatmough,Markus Nagel
机构: University of Copenhagen (哥本哈根大学); Qualcomm AI Research (高通人工智能研究院)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 25 pages, 7 figures
Abstract:Vector quantization (VQ), in particular lattice-based methods, offer strong rate-distortion trade-offs for low-bit large language models (LLMs), yet their behavior under quantization-aware training (VQ-QAT) remains poorly understood. We scale lattice VQ-QAT to billion-parameter LLMs and multi-billion-token training regimes, and introduce SQuAT, a shape-gain VQ-QAT approach trained with a separate straight-through estimator (STE) for the shape and gain. SQuAT decomposes weights into direction and magnitude: normalized directions are quantized using an E_8 -derived spherical codebook, while gains are quantized independently with low-bit scalar quantizers. We show that, unlike a standard STE applied to the full quantizer, which ignores the shape-gain structure, these separate STEs induce tangent-space updates for directions, yielding an implicit form of spherical (Riemannian) optimization. To enable scalability, we develop a distributed exact nearest-neighbor assignment scheme, in which workers quantize disjoint weight blocks and communicate only discrete indices, reducing VQ-QAT training time by 12.8 \times and allowing us to scale it to 10B tokens. Across instruction-tuned LLaMA-3.2-1B/3B, Qwen3-4B, and base Qwen3-1.7B, SQuAT consistently outperforms standard VQ-QAT STE under identical quantizers and initialization, and achieves SOTA VQ-QAT accuracy.
[NLP-18] Same Entity Different Script: Diagnosing Cross-Script Entity Matching Failures in Language Models
【速读】： 该论文旨在解决多语言大模型在处理同一人名以不同文字书写时产生的身份判断偏差问题，即模型对相同实体在不同脚本（如拉丁字母与天城文）下的身份匹配能力显著下降。其核心问题是：尽管输入语义一致，但模型在非拉丁脚本中的同一个人名识别准确率大幅降低，反映出明显的脚本依赖性偏差。解决方案的关键在于识别并干预模型内部导致该偏差的因果机制——研究发现，模型在中间层对第二实体的激活响应存在关键敏感性，且通过激活修补（activation patching）在此位置施加事实无关的干预，可将马拉雅拉姆语同名人名判断准确率从6.5%提升至33.5%，同时保持98.5%的异名人名判断准确率。此外，标量校准（scalar calibration）能恢复约87%的原始模型-探针平衡准确率差距，表明决策阈值的位置是造成偏差的重要因素。这些结果共同揭示了脚本依赖性歧视的本质并非源于表征差异，而是由决策准则的不当分布及特定神经路径的因果敏感性所致，从而为改进跨语言身份推理提供了可操作的干预靶点。
链接: https://arxiv.org/abs/2610.31407
作者: Athira Gopal,Patil Prathamesh Pravin,Ankush Kadu,Arpit Jain,Ashwanth Krishnan
机构: QpiAI India(智普科技)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Multilingual language models can give different identity judgments when the same people’s names are written in different scripts. We study this effect using 601 audited identity pairs with attested Devanagari and Latin forms, holding the question and relation fixed. Qwen2.5-7B’s same-person accuracy falls from 35.1% in Latin to 8.5% in Devanagari under Marathi prompts, while different-person accuracy remains above 98%. Evaluation across five models reveals distinct response biases, and reduced tokenization fragmentation does not reliably close the gap. Replicating the design on 374 audited Malayalam pairs built from the same entities reproduces the effect in a second script and language family: same-person accuracy falls to 5.3%, about half the matched Devanagari rate, while the margin still ranks matching above non-matching names at 0.925 AUROC. Latin-trained probes recover label information from Devanagari states, exceeding a string-feature baseline under similarity-matched negatives. Scalar calibration recovers 87% of the raw model-probe balanced-accuracy gap, identifying decision criterion placement as a major contributor. Activation patching localises causal sensitivity to the second entity in Qwen’s middle layers; one shared sensitivity direction transfers about half the full-state margin correction. A fact-independent intervention at this site raises held-out Marathi same-person accuracy from 6.5% to 33.5%, with 98.5% different-person accuracy. Together, these results distinguish script-dependent discrimination from criterion placement and identify a causal site for improving identity judgments.
[NLP-19] HERD: Harness-Evolving Recursive Distillation for Continual Learning
【速读】： 该论文旨在解决持续学习（Continual Learning）中模型在不断习得新知识时面临的灾难性遗忘（catastrophic forgetting）与可塑性丧失问题。现有参数化方法易受遗忘影响，而非参数化方法（如提示、技能和钩子机制）虽依赖上下文学习，却受限于上下文长度及纯文本压缩带来的信息损失。为融合二者优势，论文提出一种混合式持续学习框架HERD（Harness-Evolving Recursive Distillation），其核心在于通过“钩子演进-递归蒸馏”机制，在保持高效知识捕获的同时实现经验的累积内化。关键创新点包括：在钩子侧，持续优化技能以提取可复用模式，并将残差显著样本保留在有限记忆中；在参数侧，采用双锚点约束的漂移蒸馏策略，有效抑制长期递归训练中的性能退化。实验表明，在固定内存与计算预算下，HERD在TRACE+17的25阶段任务流上，不仅在新任务学习能力上达到最强参数基线水平，且表现出最低的遗忘率与最优的通用能力保留。
链接: https://arxiv.org/abs/2610.31401
作者: Xun Sun,Thanh Thi Nguyen,Thuy-Trang Vu
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 22 pages, 9 figures
Abstract:Continual learning requires a model to keep acquiring new skills and knowledge without compromising its existing capabilities. Parametric methods suffer from catastrophic forgetting and loss of plasticity. Non-parametric methods, such as prompts, skills, and harnesses, leverage in-context learning but are limited by context length and lossy plain-text compression. To effectively combine their complementary strengths, we propose HERD (Harness-Evolving Recursive Distillation), a hybrid continual learning framework that learns efficiently in the harness and internalizes experience into cumulative capabilities via distillation. On the harness side, HERD continually refines its skills to capture reusable patterns, while retaining high-residual examples that the skills fail to capture in a bounded memory. On the parameter side, HERD performs drift-constrained distillation with dual anchors, substantially reducing forgetting over long recursive runs. On the 25-stage task stream of TRACE+17, under fixed memory and compute budgets, HERD matches the strongest parametric baseline in new-task acquisition while exhibiting the least forgetting and the strongest retention of general capabilities.
[NLP-20] Parameter-Efficient Adaptation of Large Language Models to a Low-Resource Agglutinative Language: A Corpus Datasets and a Reproducible Benchmark for the Bashkir Language
【速读】： 该论文旨在解决低资源屈折语（Bashkir）在现代大语言模型（LLM）中代表性不足的问题，尤其针对其使用西里尔字母书写且缺乏高质量训练数据的现状。核心挑战在于如何有效适配大型语言模型以支持该语言的下游任务，如文本分类，并评估不同适配方法在资源有限条件下的性能差异。解决方案的关键在于构建首个系统性的开放资源与基准测试体系：包括包含71,567篇文档、4690万词元的巴什基尔网络语料库（Bashkir Web Corpus）及三类标注数据集（二分类、多分类、多标签分类），并在此基础上对八种模型架构（涵盖8200万至80亿参数规模）和三种适配策略（全量微调、QLoRA、上下文学习）以及持续预训练（CPT）进行了全面实验。研究发现，基于3.8-8B参数模型的QLoRA在二分类任务中显著优于编码器类模型的全量微调；其中，经过持续预训练的YandexGPT-5-Lite-8B在宏平均F1得分上达到0.748，远超XLM-R（0.653）和mBERT（0.650）；此外，该模型在多分类任务和人类评分中也表现最优，且西里尔字母分词器结合俄语预训练的设置更有利于提升适应质量。值得注意的是，持续预训练的效果具有任务和模型依赖性，且分词器的词素密度与模型适应效果高度相关，而自动评估指标难以完全替代母语者评价，凸显了人工评估在低资源语言中的必要性。所有成果均以开源许可发布，为后续研究提供了可复现的基础。
链接: https://arxiv.org/abs/2610.31397
作者: Mullosharaf K. Arabov,Svetlana S. Khaybullina
机构: Kazan (Volga Region) Federal University (喀山(伏尔加地区)联邦大学)
类目: Computation and Language (cs.CL)
备注: Preprint. 29 pages, 1 figure, 17 tables
Abstract:Bashkir is a low-resource agglutinative language written in Cyrillic and remains poorly represented in modern LLMs. We present an open resource suite and a systematic benchmark of adaptation methods: the Bashkir Web Corpus (71,567 documents, 46.9M tokens, 16 sources) and three labelled datasets for binary, multiclass and multilabel text classification; the benchmark covers the first two. Experiments use eight architectures (82M-8B: DistilGPT2, GPT-2 base/medium, Phi-3-mini, Mistral-7B, YandexGPT-5-Lite-8B, mBERT, XLM-R) and three adaptation groups (full fine-tuning, QLoRA, in-context learning), plus continual pretraining (CPT). (1) QLoRA on 3.8-8B models outperforms encoder-only fine-tuning on binary classification: YandexGPT-5-Lite-8B with CPT reaches macro-F1 0.748 vs 0.653 (XLM-R) and 0.650 (mBERT); pairwise differences are significant after Holm correction (p0.001). (2) YandexGPT-5-Lite-8B significantly outperforms Mistral-7B and Phi-3-mini on multiclass classification, BPB and native-speaker ratings (mean 3.28 vs 2.52 and 2.22; n=100; alpha=0.69-0.77; p0.05); on binary classification its QLoRA advantage is not significant (p=0.084). (3) CPT effects depend on model and task: perplexity improves for Phi-3-mini (+0.58) and YandexGPT-5-Lite-8B (+0.33) but worsens for Mistral-7B (-1.15, p0.001); on binary classification CPT helps Mistral-7B most (+4.0 macro-F1). (4) Tokeniser fertility on Bashkir varies 2.67x (2.92-7.80) and tracks adaptation quality, while perplexity across tokenisers is not comparable. Cyrillic tokenisers with Russian pretraining are preferable; LLM judges reproduce native ratings for grammar and coherence (rho=0.33-0.46) but not language correctness (rho=0.16, p=0.11); native evaluation remains necessary. All artefacts are open-licensed.
[NLP-21] AfterDeleteBench: Auditing Post-Deletion Information Persistence in LLM Agents
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）代理在删除外部持久化记忆后仍可能通过历史交互信息恢复已删除敏感数据的隐私泄露问题，即“删除后信息残留”这一行为失效模式。其解决方案的关键在于提出一个受控的行为审计基准测试工具 AfterDeleteBench，通过对比曾暴露于特定合成秘密但随后被删除的代理与从未暴露于该秘密的对照组代理，在相同条件下评估删除后的信息恢复能力，从而将恢复行为归因于原始暴露而非偶然或存储残留。研究发现，即使原始完整字符串已被删除，仅保留按公开组合规则分散存储的组件，仍可在 Qwen、Mistral 与 Phi-3 等模型上实现 62%-99% 的精确重构成功率；同时揭示响应接口格式与解析器不匹配会误判为失败，强调了测量系统设计的重要性。此外，研究区分了信息重建与重写回机制，并验证了溯源感知闭包（provenance-aware closure）可有效消除恢复现象，而仅依赖字面精确写入过滤器无法清除已有信息。因此，该研究主张应从代理及其记忆系统的整体行为层面评估删除操作的有效性，而非仅将其视为单纯的存储层操作。
链接: https://arxiv.org/abs/2610.31366
作者: Ehsan Saleh,Roozbeh Nahavandi,Sama Salarian,Raef Bassily,Andrew Perrault
机构: The Ohio State University(俄亥俄州立大学)
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 25 pages, 7 figures, 12 tables
Abstract:Many LLM agents maintain external persistent memory and derive additional state from past interactions, such as summaries. Deleting an original memory record therefore need not remove information carried by its descendants. We introduce AfterDeleteBench, a controlled behavioral audit for this failure mode in fixed-weight LLM agents. For each synthetic secret, we compare an exposed-then-deleted agent with an otherwise matched agent that was never exposed to the secret, allowing post-deletion recovery to be attributed to the original exposure. In our primary setting, a single query recovers the deleted secret from Qwen2.5-7B-Instruct in roughly two thirds of cases, while matched never-exposed controls have zero recovery. Crucially, recovery does not require a surviving verbatim copy: when the complete secret string is absent and only separately stored components remain under a public composition rule, exact reconstruction succeeds in 62-99% of cases across Qwen, Mistral, and Phi-3. We also show that the response interface is part of the measurement system: target-bearing replies can be scored as failures when the requested format and parser are misaligned. Finally, we separate reconstruction from write-back and fresh-session recall. Provenance-aware closure eliminates observed recovery in the tested positive-baseline settings, whereas an exact-literal write filter blocks reinsertion but cannot erase information already present. Together, these results show that source deletion should be evaluated behaviorally at the level of the agent and its memory system, not only as a storage operation.
[NLP-22] CrisiLangs: A Multilingual Resource for Crisis Warning Generation
【速读】： 该论文旨在解决突发事件中多语言预警信息生成的缺失问题，特别是在欧洲多语种环境下如何高效、准确地生成符合专家通信规范的多语言预警文本。现有研究虽已证明语言模型可实时生成英文预警信息，但缺乏多语言数据支持。为此，作者提出了CrisiLangs，作为CrisiText（一个基于专家通信指南的英文危机事件与预警信息配对数据集）的多语言扩展，涵盖意大利语、保加利亚语、德语和希腊语，共包含1,000个事件的5,000个跨语言事件-消息对。其解决方案的关键在于构建高质量、标准化的多语言预警信息数据集，并在此基础上联合微调多语言模型，同时评估在提示中加入应急指南的影响。实验表明，微调后的模型在所有语言和配置下均显著优于零样本基线，在ROUGE-L、BLEU和BERTScore指标上表现更优；人类偏好研究表明，较低的解码温度在所有语言中均获得更高偏好，而指南的影响则呈现语言特异性。
链接: https://arxiv.org/abs/2610.31356
作者: Giacomo Gonella,Stefano Menini,Marco Guerini
机构: Fondazione Bruno Kessler(布鲁诺·凯斯勒基金会); University of Trento(特伦托大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Communicating with the public during an emergency entails issuing warning messages promptly. In Europe, it often requires producing them in more than one language. Recent work has shown that language models can be trained to generate warning messages as a crisis unfolds, but the available data covers only English. To fill this gap, we present CrisiLangs, a multilingual extension of CrisiText (an English dataset of crisis events paired with warning messages written according to expert communication guidelines) to Italian, Bulgarian, German, and Greek. Expert native speakers translated the warning messages of 1,000 events drawn from 411 scenarios, for a total of 5,000 event-message pairs across five languages. We test the resource to fine-tune a multilingual model on all five languages jointly, with and without the emergency guidelines in the prompt, and compare it with a zero-shot baseline. The fine-tuned models outperform the baseline on ROUGE-L, BLEU and BERTScore in every language and configuration. In a human preference study, low decoding temperatures receive the most preferences in all five languages and in both configurations, while the effect of the guidelines varies by language.
[NLP-23] When LLM s Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues EMNLP2026
【速读】： 该论文旨在解决大语言模型在长时对话中出现的“记忆-遵从分离”问题，即模型虽能记住用户设定的约束条件（如格式、角色、逻辑规则等），却在后续对话回合中逐渐停止执行这些约束，导致行为偏离初始指令。其核心问题是：现有评估方法仅关注单轮指令遵循能力，无法揭示模型在复杂、多轮交互中对状态约束的持续维护能力，从而低估了实际应用中的可靠性风险。解决方案的关键在于提出并验证“状态完整性崩溃”（State Integrity Collapse）这一新现象，并构建首个系统性评估基准——StateTrack-Bench，涵盖数值、格式、角色/人格、事实记忆及逻辑规则等五类约束，通过多轮深度对话测试模型对约束的长期保持与执行能力。研究发现，大多数失败源于行为漂移（behavioral drift），而非简单遗忘，且模型性能在深度超过20轮后显著下降；引入结构化状态锚定（Structured State Anchoring）策略可有效提升不可变约束上的表现达24.4个百分点，表明需通过显式状态管理机制增强模型在长时对话中的状态一致性。结果强调，模型可靠性必须基于对动态对话状态下约束持续有效性的追踪评估，而不能仅依赖于单轮指令遵循表现。
链接: https://arxiv.org/abs/2610.31314
作者: Yancheng Zhu,Yifan Yuan
机构: Shenzhen University (深圳大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Accepted at EMNLP 2026 Main Conference. 17 pages including appendices
Abstract:Large language models can repeat an instruction while failing to keep obeying it. We study this recall-compliance dissociation in long-horizon dialogue and define State Integrity Collapse: a model preserves a stated constraint in memory but stops applying it to later turns. We introduce StateTrack-Bench, a controlled suite of conversations covering numerical, format, role/persona, factual-memory, and logical-rule constraints. Across seven provider-pinned model snapshots and five dialogue-depth probes, strict joint retention ranges from 62.4% to 100.0%, with degradation concentrated in symbolic constraints and in one especially brittle model. A forked recall audit shows that 88.6% of observed failures are behavioral-drift failures rather than ordinary forgetting; excluding the brittle outlier preserves this diagnosis (84.4% of failures). Extended evaluations at depths up to 50 turns confirm that the dissociation persists beyond the core benchmark, while Structured State Anchoring recovers up to 24.4 points on immutable constraints. The results show that reliability cannot be inferred from single-turn instruction following alone; evaluation must track whether remembered constraints remain operational as dialogue state grows. Code and artifacts are available at: this https URL.
[NLP-24] Cola2: Co-Adaptive Hierarchical Language Modeling with Continuous Semantic Compression and Balanced Discrete Generation
【速读】： 该论文旨在解决层次化语言生成中连续语义压缩与离散文本生成之间协同优化不足的问题，核心挑战在于如何有效设计隐空间所蕴含的语义信息、平衡重建与生成任务之间的权衡，以及实现连续与离散模态间的联合自适应。其解决方案的关键在于提出Cola2框架，通过动态分割文本为语义片段，利用无辅助损失的负反馈机制自适应调节压缩率，并结合重建与生成双重监督以学习紧凑且具备预测能力的潜在表示。该框架采用块因果掩码扩散解码器，在干净、噪声及丢弃的潜在条件下游走训练，同时通过无分类器引导（classifier-free guidance）灵活调控语义重建与自主生成的平衡。此外，通过梯度更新机制使编码器基于潜在变量的词元预测梯度进行优化，而流匹配先验学习则通过停止梯度操作与表征学习解耦，从而提升训练稳定性。推理阶段，生成文本在固定语义段布局下被重新编码，以维持潜在历史与实际上下文的一致性。实验在五个下游任务上验证了核心设计的有效性，展现了可扩展性强、未饱和的性能增长趋势；在约3460亿源令牌预训练条件下，未收敛检查点在八项任务上取得平均60.10分，超越了预训练预算超过其43倍的Llama-3-8B模型在HellaSwag和StoryCloze上的表现，充分证明了连续潜在层次化语言生成的潜力。
链接: https://arxiv.org/abs/2610.31305
作者: Hongcan Guo,Yian Zhao,Feng Wang,Shen Nie,Guoqiang Wei,Yan Zeng,Hengshuang Zhao
机构: 未知
类目: Computation and Language (cs.CL)
备注: 47 pages, 14 figures, 5 tables
Abstract:Hierarchical language generation separates continuous semantic modeling from discrete text realization, but its effectiveness depends on what the latent space compresses and how the discrete generator uses it. We propose Cola2, a framework that jointly studies continuous semantic compression, the balance between reconstruction and generation, and continuous-discrete co-adaptation. Cola2 dynamically segments text into semantic spans, controls the compression rate through auxiliary-loss-free negative feedback, and combines reconstruction and generation supervision to learn compact, predictive latent representations. A block-causal masked diffusion decoder is trained with clean, noisy, and dropped latent conditions, while classifier-free guidance adjusts the balance between semantic reconstruction and autonomous generation. Token-prediction gradients update the encoder through the latents, whereas Flow Matching prior learning is decoupled from representation learning by stop-gradient. During inference, generated text is re-encoded under a fixed span layout to keep the latent history consistent with the actual text context. Experiments on five downstream tasks with about 210 billion training tokens validate the core designs and show competitive, unsaturated scaling. With 346.03 billion source tokens, an unconverged checkpoint reaches a 60.10 average across eight tasks and surpasses Llama-3-8B on HellaSwag and StoryCloze despite its reported pretraining budget being over 43 times larger. These results demonstrate the potential of continuous-latent hierarchical language generation.
[NLP-25] he Value of Memory in Language Agents
【速读】： 该论文旨在解决语言智能体在有限记忆资源下如何有效保留、利用与更新过往经验中的关键信息，以应对动态演化任务需求的问题。其核心挑战在于：记忆的价值不仅取决于信息的留存，还受制于信息能否被有效调用、在决策中是否可操作，以及在需求变化时能否被成本可控地恢复或修正。论文提出一种基于决策论的框架，将记忆价值从原始历史数据经由持久状态、实际使用，直至代价高昂的恢复与修订全过程进行量化建模，以“避免的任务损失”为单位衡量价值。通过构建风险-成本前沿（risk-cost frontiers），该框架揭示了在不同使用场景下可实现的恢复与保存能力边界。研究发现，即使指针（pointer）与答案之间互信息为零，仍可通过降低精确恢复的成本来提升整体效率；同时，交互式反馈会改变未来任务分布，进而影响记忆配置策略。通过成本感知仿真与五个最小反例分析，论文揭示了信息时效性、检索访问权限、组件间耦合、反馈机制及需求不确定性等关键因素的作用。在LoCoMo双项目更新任务及纵向案例中的诊断表明，现有系统普遍存在记忆保留、使用与度量失败问题——例如，在固定检索证据下，保存的账本会影响后续输出；在固定代理输出下，评估者修改会改变评分结果。最终，该框架将上述现象统一至操作风险边界，并提出了可验证的原则，用于指导记忆分配、恢复选项保留与证据依赖关系的动态修订。
链接: https://arxiv.org/abs/2610.31303
作者: Zehao Lin,Erenkai Zhu,Helan Hu,Xixuan Hao,Chunyu Li,Kai Chen
机构: MemTensor; Shanghai Jiaotong University (上海交通大学); The Hong Kong University of Science and Technology (Guangzhou) (香港科技大学(广州))
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 19 pages, 3 figures, 5 tables
Abstract:Memory enables language agents to carry useful distinctions from past experience into future decisions. Its value depends on which distinctions survive, whether a bounded agent can use them, and what can be recovered or revised as demands evolve. We develop a decision-theoretic framework that follows this value from raw history through persistent state, operational use, and costly recovery and revision. Original and retained value are expressed in units of avoided task loss; risk-cost frontiers describe attainable recovery and preservation across uses. Unknown demand yields value profiles, while interactive feedback changes the distribution of future work. A cost-aware simulation condition and five minimal counterexamples expose the roles of information timing, retrieval access, component interaction, feedback and demand uncertainty. In particular, a pointer with zero mutual information about the answer can still reduce the cost of exact recovery. Diagnostics on LoCoMo, a two-project update task and longitudinal cases exhibit failures of preservation, use and measurement. At fixed retrieved evidence, a saved ledger changes subsequent answers; at fixed agent outputs, an evaluator revision changes scores. The framework connects these distinctions to operational risk bounds and testable principles for allocating memory, preserving recovery options and revising evidence dependencies.
[NLP-26] Memory-Aware Reward Modeling: Learning Human Preferences for LLM Memory Use
【速读】： 该论文旨在解决当前大语言模型（LLM）中记忆机制（memory）在实际应用中存在“信息冗余”与“用户偏好错配”的核心问题，即尽管系统具备存储和调用用户个人记忆的能力，但这些记忆并不总是带来更受用户欢迎的响应。现有评估方法普遍依赖自动指标或间接测试，未能直接验证记忆使用是否真正提升用户体验。为此，研究者通过收集87名参与者约1,000组成对判断数据，基于其真实的ChatGPT查询与个人记忆进行对比实验，发现记忆条件化响应在多数情况下并未显著优于无记忆响应，甚至在部分场景下用户明确更偏好不使用记忆的输出。同时，自动评价指标与用户偏好之间存在显著偏差，过度倾向支持记忆条件化响应。针对这一差距，论文提出一种基于用户反馈训练的记忆感知奖励模型（memory-aware reward model），通过学习真实用户偏好实现72.8%的偏好预测准确率，显著优于现有基线。其关键解决方案在于构建以用户主观偏好为导向的评估与优化框架，确保记忆功能的启用真正服务于用户的实际需求，而非仅体现技术能力的堆叠。
链接: https://arxiv.org/abs/2610.31272
作者: Yijiang River Dong,Zexue He,Cheng Yu,Tiancheng Hu,Michael J Ryan,Jan Batzner,Zheng Hui,Shaolong Wu,Orestis Papakyriakopoulos,Alex Pentland,Nigel Collier,Jiaxin Pei
机构: University of Cambridge(剑桥大学); Stanford University(斯坦福大学); Technical University of Munich(慕尼黑工业大学); Harvard University(哈佛大学); University of Texas at Austin(德克萨斯大学奥斯汀分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Memory enables language model assistants to maintain continuity across sessions and generate personalized responses, but knowing more about a user does not necessarily produce responses they prefer. Existing evaluations assess memory capabilities without directly establishing whether memory use benefits the user. To study this gap, we collect approximately 1,000 pairwise judgments from 87 participants using their own ChatGPT queries and personal memories. We find that memory-conditioned responses are not consistently preferred over responses generated without memory and in some cases, users strongly prefer the latter. Automatic evaluators also show limited agreement with users and favor memory-conditioned responses substantially more often than users do. Motivated by these gaps, we introduce a memory-aware reward model that learns from user judgments, achieving 72.8% preference prediction accuracy and outperforming the existing baselines. Our findings motivate evaluation and optimization objectives that align memory use with the preferences of the people whose information is being used.
[NLP-27] A Systematic Comparison of Search Space Complexity and Efficiency for Time-Synchronous End-to-End ASR Models
【速读】： 该论文旨在解决当前端到端（End-to-End, E2E）自动语音识别（ASR）系统中搜索策略效率不足的问题，特别是针对基于连接时序分类（CTC）和神经声学转换器（Neural Transducer）模型在不同剪枝策略与语言模型（LM）配置下的搜索空间复杂度与计算效率问题。其核心挑战在于：尽管这些E2E模型在多数情况下可维持较小的搜索空间，但在模型不确定性较高的时段会出现瞬时候选假设数量激增的现象，导致计算开销显著上升。论文提出的关键解决方案是采用基于得分的动态剪枝策略，该策略能够根据模型置信度自适应调整搜索范围，有效抑制不确定性时期的搜索膨胀，从而显著降低实时因子（Real-Time Factor, RTF）。研究结果表明，固定小尺寸束搜索（beam search）在低复杂度场景下表现不佳，而动态剪枝能更高效地平衡搜索精度与计算效率，为E2E ASR中的高效搜索提供了切实可行的设计准则。
链接: https://arxiv.org/abs/2610.31259
作者: Simon Berger,Larissa Kleppel,Ralf Schlüter
机构: RWTH Aachen University (亚琛工业大学); AppTek GmbH(AppTek公司)
类目: Computation and Language (cs.CL)
备注: Accepted at IEEE SLT 2026. 7 pages, 4 figures
Abstract:Search is a central component of Automatic Speech Recognition (ASR) systems and has been deeply investigated for classical ASR modeling based on Hidden Markov Models (HMMs). However, in work on current sequence-to-sequence ASR architectures it is often simplified to a small fixed-size beam search without external Language Model (LM). In this work, we present a comparative analysis of search space complexity and efficiency for Connectionist Temporal Classification (CTC) and neural transducer models under different pruning strategies and LM configurations. We show that these end-to-end (E2E) models work with a small search space most of the time but exhibit large temporary peaks in hypothesis count during periods of uncertainty. Score-based pruning dynamically adapts the search to model uncertainty, leading to a significantly lower Real-Time Factor (RTF). Our results highlight the limitations of search with low fixed-size beam pruning and provide practical guidelines for efficient search in E2E ASR.
[NLP-28] IndoNutriQA: An Indonesian Clinical Nutrition Question Answering Benchmark EACL
【速读】： 该论文旨在解决低资源语言（如印尼语）在临床智能决策支持系统中因缺乏公开领域数据而制约临床问答（Clinical QA）发展的关键问题。其核心解决方案是构建一个由专家精心标注的印尼语临床问答基准数据集IndoNutriQA，包含911对基于真实临床营养报告的问题-答案对。该数据集的关键特征在于：每一对问答均基于临床相关上下文，附带结构化元数据（包括人体测量评估、医学诊断、文档来源和文本分块），并遵循营养护理流程（NCP）标准设计八类标准化问题，涵盖问题-病因-症状（PES）诊断表述；所有答案为自由形式的临床叙述性文本而非抽取片段，且经专家验证存在6–10%的参考答案可争议。该数据集不仅适用于生成式、抽取式及检索增强型问答任务，还可支持零样本与少样本大语言模型（LLM）评估。研究通过五折交叉验证报告了微调基线模型（IndoBERT变体、mT5、RAG）的性能，并对八种主流LLM进行了零样本与五样本评测。此外，通过将LLM作为评判者与三位专家（一名注册营养师、两名营养师）对比，发现其在事实性和完整性方面表现接近人类，但在临床安全性判断上不可靠，且人类错误分析揭示其会遗漏专家识别出的无依据内容。IndoNutriQA的发布旨在推动印尼语临床自然语言处理（NLP）与营养决策支持领域的研究发展。
链接: https://arxiv.org/abs/2610.31256
作者: Tirana Noor Fatyanosa,Ummu Ditya Erliana,Mahendra Data,Muhammad Nurwegiono,Eleazar Tadeo Eman,Alham Fikri Aji
机构: Universitas Brawijaya(布拉维查亚大学); Universitas Ma Chung(马昌大学); MBZUAI(穆罕默德·本·扎耶德人工智能大学)
类目: Computation and Language (cs.CL)
备注: 26 pages (9 pages main text, plus limitations, ethics, references and appendix), 5 figures, 18 tables. Under review at EACL. Data and code: this https URL
Abstract:Clinical Question Answering (QA) underpins Clinical Intelligent Decision Support Systems, but progress is severely limited for low-resource languages such as Indonesian by the absence of public domain-specific data. We introduce IndoNutriQA, an expert-curated Indonesian clinical QA benchmark of 911 question-answer pairs built from real clinical-nutrition reports. Rather than a large-scale training corpus, it is positioned as a high-quality evaluation resource: every pair is grounded in a clinically relevant context, enriched with structured metadata (anthropometric assessment, medical diagnosis, document provenance, chunk segmentation). Eight standardised question types following the Nutrition Care Process (NCP), including the Problem, Etiology, Sign and Symptom (PES) diagnosis statement, are answered for every report, which was written by a graduate-level dietetic intern, by one of three nutritionists. Answers are free-form clinical prose rather than extractive spans, and an expert validation study finds 6-10% of reference answers contestable. The dataset supports generative, extractive and retrieval-augmented QA as well as zero- and few-shot LLM evaluation. We report fine-tuned baselines (IndoBERT variants, mT5, RAG) with 5-fold cross-validation, learning curves and bootstrap 95% confidence intervals, and benchmark eight contemporary LLMs zero- and five-shot. We further calibrate our LLM-as-judge rubric against three expert raters (one RD, two nutritionists), finding it human-comparable on factuality and completeness but unreliable on clinical safety, and a human error analysis shows it misses unsupported content that the raters identify. IndoNutriQA is released to foster research in Indonesian clinical NLP and nutritional decision support.
[NLP-29] Predicting Music Valence and Arousal via Textual and Multimodal In-Context Learning in Large Language Models ICASSP2027
【速读】： 该论文旨在解决音乐情感效价（valence）建模困难的问题，因其具有高度主观性和情境依赖性，相较于唤醒度（arousal），更难通过音频信号进行准确预测。其核心解决方案在于采用无需训练的上下文学习（in-context learning, ICL）策略，利用大规模语言模型（large language models）结合辅助数据集中的上下文示例（demonstrations），对小型专用基础模型生成的初始情感预测结果进行精细化修正。关键创新点在于：不直接预测连续情感分数，而是通过文本或音视频多模态的上下文示例引导模型在推理阶段动态调整输出。实验基于完整的MER60数据集作为独立测试集，结果表明，基于歌词与生成音频描述的文本型ICL方法，在80个示例下使效价预测的R²从0.110提升至0.427；而融合原始音频信号的多模态ICL方法仅需10个示例即可达到0.346的R²。这些成果验证了训练-free ICL作为一种高效、可扩展的推理时策略，显著提升了当前语言模型在连续音乐情感预测任务中的性能，尤其在效价建模方面表现突出。
链接: https://arxiv.org/abs/2610.31202
作者: Chi-En Dai,Chih-Pin Tan,Yi-Hsuan Yang
机构: 未知
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Submitted to ICASSP 2027
Abstract:While arousal can be reliably predicted from musical audio, emotional valence remains difficult to model due to its subjective and context-dependent nature. In this work, we investigate the potential of in-context learning (ICL) with large language models to advance continuous valence and arousal prediction. Rather than predicting emotion scores directly, our framework leverages contextual demonstrations from an auxiliary dataset to refine the initial predictions produced by a small, task-specific base model. We systematically evaluate a text-based ICL variant using lyrics and generated audio captions, alongside a multimodal variant that incorporates raw audio signals. Using the entire MER60 dataset strictly as a held-out test set, our experiments demonstrate substantial performance gains, particularly for valence. Text-based ICL improves the valence R^2 from 0.110 to 0.427 with 80 demonstration shots, while the multimodal variant achieves 0.346 with only 10 shots. These findings suggest that training-free ICL offers an effective inference-time strategy for current language models to predict continuous music emotion.
[NLP-30] Investigating Multi-Cultural Value Selection Biases in LLM s
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在处理多层文化身份时存在的简化问题，即现有文化评估常将复杂的身份认同压缩为单一标签，忽略了用户可能同时携带多重文化背景的现实。其核心挑战在于：当系统提示中设定国家角色（country persona），而用户在输入中通过显性命名国家或隐式使用特定语言来传递第二重文化线索时，这些不同信号如何影响模型输出的文化倾向。研究的关键发现是，语言线索对模型回答的影响强度约为显性国家线索的三倍，但其引导方向往往偏离所关联国家的价值分布；相比之下，显性国家线索并未稳定地使模型向对应国家的价值取向靠拢。在160万条开放式生成数据中，明确提及国家可缩小模型自身响应与目标国家响应之间的差距达30%-44%，且显性线索引发的价值偏移程度高于语言线索。因此，该研究提出，国家与语言应被视为独立的文化信号，而非可互换的文化代理变量，这对构建更具文化敏感性的生成式人工智能（Generative AI）系统具有重要意义。
链接: https://arxiv.org/abs/2610.31199
作者: Siddhesh Pawar,Arnav Arora,Dustin Wright,Dirk Hovy,Isabelle Augenstein
机构: University of Copenhagen(哥本哈根大学); Aalborg University Copenhagen(奥尔堡大学哥本哈根校区); Bocconi University(博科尼大学)
类目: Computation and Language (cs.CL)
备注: 27 pages, 13 figures, 18 tables
Abstract:Large language models (LLMs) serve users whose cultural identities are multi-layered, yet cultural evaluations commonly reduce identity to one label. In this work, we study LLMs under simultaneous identity cues, with a country persona in the system prompt and a second identity signaled in the user turn, either by naming a country (explicit cue) or through the query language (implicit cue). We test World Values Survey (WVS) multiple-choice probing and open-ended sociopolitical generation. In multiple-choice probing, language cues shift answers about three times more than explicit cues but away from the associated country’s distribution, and explicit cues do not consistently move answers toward the named country. Across 1.6 million open-ended responses, naming a country closes 30-44% of the gap to the model’s own responses for that country, and explicit cues produce more differentiated value changes than language cues. Our results suggest that country and language cues should be evaluated as distinct signals rather than interchangeable proxies for culture.
[NLP-31] When More Metadata Misleads: Probing Field-Level Grounding in LLM -Generated Museum Descriptions EMNLP2026
【速读】： 该论文旨在解决博物馆文物描述生成中元数据（metadata）条件化生成的局限性问题，尤其针对馆藏记录在内容结构、视觉信息覆盖及地理分布上的异质性与不均衡性。传统数据到文本（data-to-text）任务假设额外输入字段能增强生成内容的准确性与可接地性，但该研究指出这一假设在真实博物馆场景下并不成立。其解决方案的关键在于提出一个两阶段框架：第一阶段分析大规模博物馆藏品中自然存在的元数据变异模式，揭示字段数量、字段类型、图像可用性与地理位置之间的共变关系；第二阶段通过受控的字段探针实验，在保持对象身份不变的前提下，系统性地操纵元数据的数量、类型和图像访问权限。研究发现，元数据总量并非决定生成行为的核心因素，而字段类型决定了模型是进行信息锚定、省略还是推断；图像虽有助于视觉锚定，但也可能诱发无依据的推断。因此，论文强调元数据条件化生成评估应从整体输入覆盖率或参考文本相似度转向在字段层级和陈述层级上分别评估支持度，以实现更精确、可解释的评估体系。
链接: https://arxiv.org/abs/2610.31193
作者: Clemencia Siro,Laura Hollink
机构: Centrum Wiskunde & Informatica (CWI)(荷兰数学与计算机中心); Amsterdam, The Netherlands(阿姆斯特丹, 荷兰)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Digital Libraries (cs.DL); Machine Learning (cs.LG)
备注: Accepted to EMNLP 2026 Findings
Abstract:Metadata-conditioned generation is often framed as a data-to-text task in which additional input fields should improve grounding. Museum description generation challenges this assumption because catalogue records are heterogeneous, partially visual, and unevenly documented across geographic regions. We study how metadata composition shapes the generation of artefact descriptions through a two-phase framework. Phase 1 analyses natural metadata variation in a large museum collection, where field count, field type, image availability, and Place co-vary. Phase 2 introduces controlled field probes that hold object identity fixed while varying metadata quantity, field type, and image access. Our results show that metadata quantity alone does not explain generation behaviour. Field type determines whether models ground, omit, or infer information, while images can both support visual grounding and increase unsupported interpretation. More broadly, our findings show that metadata-conditioned data-to-text evaluation should assess field- and claim-level support, rather than relying only on aggregate input coverage or reference similarity.
[NLP-32] Prompting Strategy Effects Depend on the Judge: A Multilingual Study of LLM -as-a-Judge Robustness
【速读】： 该论文旨在解决生成式AI在多语言评估中提示策略（prompting strategies）的跨语言、跨模型可迁移性问题。当前主流的提示方法如思维链（Chain-of-Thought, CoT）和角色引导（Role Conditioning）的有效性主要基于英语基准测试，其在多语言场景下的适用性尚未得到充分验证。现有研究或局限于英语环境，或固定提示模板，缺乏对提示策略与评估设置（包括评判模型、任务类型、语言资源水平）之间交互关系的系统性考察。为此，本文构建了一个涵盖9种提示策略、4个开源评判模型、27种语言及4类任务的综合性基准测试框架，覆盖13个数据集，生成超百万条基于黄金标准输出的逐点评分，并通过人工判断进行验证。研究发现：提示策略不具备跨模型与跨语言的通用性；评判模型的选择显著影响评估结果，甚至主导了提示策略的作用；同一提示策略在不同评判模型上可能产生相反的效果；在低资源语言中，这些不稳定性被整体平均值掩盖，导致误判。因此，提示策略在多语言评估中的有效性受到评判模型、任务和语言三者的共同制约。结论强调：提示策略必须在具体评估场景中重新验证，且跨研究比较时需明确报告评判模型与提示配置，以确保评估结果的可比性和可靠性。
链接: https://arxiv.org/abs/2610.31172
作者: Asal Mehrabi,Mohammad Aliannejadi,Clemencia Siro
机构: University of Amsterdam (阿姆斯特丹大学); Centrum Wiskunde & Informatica (荷兰数学与计算机中心)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 23 pages,
Abstract:LLM-as-a-Judge evaluation is increasingly applied to multilingual tasks, yet prompting strategies such as Chain-of-Thought and Role Conditioning are recommended based solely on English benchmarks. Whether their effectiveness transfers across judge models and languages remains unknown: prompt-sensitivity studies are largely English-centric, while existing multilingual judge studies fix the prompt template. In this work, we address this gap by measuring this interaction through a benchmark of 9 prompting strategies, 4 open-weight judges, 27 languages, and 4 tasks across 13 datasets, yielding over 1M pointwise scores on gold-reference outputs, validated against human judgments. We find that prompting strategies are not transferable: the choice of the judge model dominates the choice of prompting strategy, strategy effects reverse direction across judges, and stability degrades in low-resource languages, where these effects are concealed by aggregate reporting. These findings show that prompting effects in multilingual evaluation are conditioned on the judge, task, and language. Consequently, prompting strategies should be validated for the intended evaluation setting, and cross-study comparisons should account for both judge and prompt configuration.
[NLP-33] Beyond Watching and Searching: Benchmarking Agents for Video-centered Deep Research Reports
【速读】： 该论文旨在解决当前深度研究代理（Deep Research Agents）在处理以视频为中心的科研任务时缺乏有效评估基准的问题。现有基准主要聚焦于文本或网页信息，无法充分衡量代理在基于源视频进行深度研究时的能力，尤其是在从视频中提取关键信息、关联外部多模态证据并生成可交互报告方面的能力。其解决方案的关键在于提出一种全新的任务范式——“视频中心型深度研究报告生成”（Video-Centered Deep Research Report Generation），并构建了相应的技能驱动工作流与评测基准 OmniDR-Bench。该基准涵盖50个中英文任务，覆盖19个真实世界子类别，包含6,414项评估标准。为实现精细化评估，论文进一步提出了“全维度原子评分树”（Omni-ARTs），通过音频视觉对齐性、多模态证据利用和交互式呈现三个核心维度，结合23个共享方向及任务特定的二元原子标准，采用审计式评分构建（ARC）方法，融合独立证据收集、自动化校验、行为校准与迭代人工评审，确保评估的客观性与严谨性。实验表明，当前最优系统得分仅为63.3/100，揭示出现有模型在内容覆盖广度、跨模态因果与逻辑推理能力以及引用准确性方面仍存在显著不足。研究还深入分析了模型选择、工具调用策略、技能配置、视频类型及研究路径对性能的影响，为未来视频驱动的研究代理发展提供了系统性指导。
链接: https://arxiv.org/abs/2610.31158
作者: Jiahao Meng,Qize Yang,Yuxuan Wang,Yunfei Chu,Wensi Dai,Jin Xu,Yunhai Tong
机构: Peking University (北京大学); Alibaba Token Hub (阿里巴巴通义实验室)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Deep research agents synthesize open-web evidence into long reports, but current benchmarks do not evaluate research centered on a source video. We introduce Video-Centered Deep Research Report Generation. Given a video and user request, an agent identifies research questions, retrieves external multimodal evidence, and produces an interactive report linked to relevant video moments. This task has direct practical value because it helps users locate relevant moments, verify claims, reproduce demonstrated procedures, and make informed decisions from video content. We present a skill-based workflow and OmniDR-Bench for this task. For evaluation, we propose Omni Atomic Rubric Trees (Omni-ARTs). Each tree contains three dimensions, 23 shared directions, and task-specific atomic binary criteria. The directions test audiovisual grounding, multimodal evidence use, and interactive presentation. Audited Rubric Construction (ARC) builds these trees through independent evidence collection, automated checks, behavioral calibration, and iterative human review. OmniDR-Bench contains 50 Chinese and English tasks across 19 realworld subcategories. It includes 6,414 criteria. Across 18 agent configurations, the best system scores 63.3 out of 100. Current agents still need broader content coverage, stronger causal and logical analysis across modalities, and more accurate citation support. We further analyze models, harnesses, skill configurations, video categories, and research trajectories. The code and data are released at this https URL.
[NLP-34] Prompt Sensitivity in Tool-Augmented Model Comparisons
【速读】： 该论文旨在解决语言模型在问答基准测试中因问题表述方式不同而导致性能排序不稳定的问题，即同一模型在不同措辞下可能表现出不同的优劣顺序。其核心解决方案在于提出一种“匹配协议”（matched protocol），通过将模型推理过程分解为多个可分离的步骤——包括答案读取、推理、调用工具的决策、工具返回结果以及无需调用时的报告生成——以实现对工具使用影响的精确量化。研究发现，尽管工具在程序化任务中能提升准确率，但其引入会加剧模型间比较对问题表述的敏感性，且这种敏感性与模型调用工具的频率密切相关，而调用频率本身又受问题表述、接口设计和指令设置的影响。尤其值得注意的是，在采用原生工具调用接口并提供额外工具说明的情况下，分子领域内的比较敏感性未被检测到，但在工具返回中间结果的情境下，敏感性依然存在。这些结果表明，仅依赖确定性输出报告无法确保比较结果的稳定性。因此，作者建议在评估模型性能时，除报告准确率外，还应同时披露工具调用频率及在多种表述下的比较稳定性。
链接: https://arxiv.org/abs/2610.31148
作者: Mateusz Modrzejewski,Tomasz Radzikowski
机构: Warsaw University of Technology (华沙理工大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 9 pages plus 8 pages of supplementary material; 2 figures and 4 tables in the main text, 2 figures and 5 tables in the supplement
Abstract:Language models are compared on question-answering benchmarks that pose each question in one wording, and a reworded question can change which of two models scores higher. We ask whether giving the models tools makes such comparisons more stable. An accuracy gain cannot answer this, because tools change several steps between the question and the scored answer at once. We propose a matched protocol that separates those steps: the answer readout, reasoning, the decision to call, the tool’s answer, and the report supplied without a call. We apply it to six open models in five domains under five wordings. We find that tools raise accuracy where a program answers the question and widen the wording sensitivity of many pairwise comparisons. The widening follows how often a model calls, and call frequency follows the wording for some models and not others. It also depends on the interface and the instructions: a native tool-calling interface and a second tool description leave no detected widening in molecules, and sensitivity persists where tools return intermediate results. These findings show that deterministic reports do not guarantee stable comparisons. We recommend reporting call frequency and comparison stability under several wordings beside accuracy.
[NLP-35] Mixing Matters: Evidence Position Bias Across Sequence Mixers in Long-Context Question Answering NEURIPS2026
【速读】： 该论文旨在解决长上下文模型在处理远偏离提示边界（prompt boundary）的证据时，其利用效率下降的问题，即位置偏差（position bias）问题。研究通过将一条黄金文档在固定九个干扰文档中移动，系统比较了Transformer、状态空间模型（state-space model）以及混合型Mamba架构在不同位置上的表现。关键发现是：所评估的Transformer模型表现出显著的“开头优势”（beginning advantage），而状态空间模型则未呈现此现象；但两类架构均倾向于偏好靠近结尾的证据。对纯Mamba与混合Mamba的对比进一步支持这一趋势，尽管统计上尚不显著。值得注意的是，当任务能力受限时，如在合成检索任务中因状态空间模型性能上限导致无法进行家族间比较，仍可观察到Transformer的优势重现。此外，同一状态空间架构内更换语料库虽改变整体准确率，但响应曲线形状无明显变化，表明模型行为受内容而非结构主导。提示（prompt）设计的变化能够诱发或几乎消除开头优势，说明上下文位置信息在多种架构中仍可被解码，且在测试的Transformer族中，注意力集中于初始标记的程度与开头敏感性呈正相关。综上，该研究揭示有效上下文利用依赖于架构家族、模型能力、任务类型及提示设计，而非仅由序列混合机制（sequence mixer）单一决定。
链接: https://arxiv.org/abs/2610.31135
作者: Aman Behera,Namit Solanki,Mehul Anand
机构: IIT Roorkee(印度理工学院鲁尔基分校); AISSMS IOIT, Pune(印度奥姆萨玛德工业与技术研究所, 浦那); Independent Researcher(独立研究员)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at the 2nd Workshop on Principles of Generative Modeling (PriGM), NeurIPS 2026. 9 pages. Code: this https URL
Abstract:Long-context models often use answer-bearing evidence less effectively when it appears away from a prompt boundary. We compare this position bias across Transformer, state-space, and hybrid Mamba architectures by moving one gold document through a fixed set of nine distractors. The evaluated Transformers show a beginning advantage that the evaluated state-space models lack, while both families favor evidence near the end. A matched pure-versus-hybrid Mamba comparison points in the same direction but remains statistically uncertain. The family difference appears among models that can answer the task. Transformer primacy reappears on synthetic retrieval, where state-space ceiling prevents a family comparison. A corpus swap within one state-space architecture changes overall accuracy without a detectable change in curve shape. Prompt changes can create or nearly remove a beginning advantage. Position remains decodable across architectures even when response curves differ, and attention concentration near initial tokens covaries with beginning sensitivity in the tested Transformer family. These exploratory results show that effective context use depends on the evaluated architecture family, capability, task, and prompt, without isolating the sequence mixer as the sole cause.
[NLP-36] Delta N-gram: Forward-Only Continual Writing into a Language Models Native N-gram Memory
【速读】： 该论文旨在解决大模型在持续学习（continual learning）过程中，由于反向传播导致预训练词元表（pretrained N-gram table）被破坏的问题。传统方法依赖梯度更新对词元表进行修改，这会干扰模型的原始知识表示，引发性能退化与记忆漂移。其解决方案的关键在于提出Delta N-gram机制——一种仅通过前向传播（forward-only）写入预训练词元表的路径。该机制不直接修改预训练表 T0，而是将新证据以稀疏补丁 Δ 的形式写入对应行，利用一个410万参数的生成器在每个（位置, 哈希头）事件中输出目标值和门控信号，并采用块平均的增量规则确保每行保持为旧值与加权目标的凸组合。该生成器在离线阶段使用原始文本数据以预训练目标进行训练，推理时仅需前向计算，且清空 Δ 即可恢复初始状态。实验表明，该方法在0.6B规模模型上能稳定保留16条事实中的全部16条，显著优于依赖在线梯度更新的方法（仅保留2–4条）。进一步研究发现，采用“差分编码”（delta writer）策略的模型在干扰下具有更强的记忆保持能力（70% vs 54%），且在多任务链上表现出更小的层间漂移（如第二层数学任务漂移降低176%对比188%），同时有效抑制了通用输入引起的格式行范数膨胀（从20.75%降至32.8%），并验证了关键冗余行的清除可消除超过98%的层2漂移。因此，核心创新在于实现无损、可逆、高效且稳定的前向知识注入机制。
链接: https://arxiv.org/abs/2610.31125
作者: Hao Wu
机构: Illinois Institute of Technology(伊利诺伊理工学院)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 33 pages, 3 figures, 9 tables, 2 algorithms. Follow-up to arXiv:2610.14921 . Code: this https URL
Abstract:Our previous analysis found that a model’s native hashed N-gram table is where continual updates disturb the base least, but its writes needed backpropagation and changed the pretrained table. We present Delta N-gram, a forward-only write path into that table. The pretrained table T^0 is never modified. Evidence is written into a sparse patch \Delta on the rows it addresses: a 4.1M-parameter writer emits a target and a gate per (position, hash head) event, and a block-averaged delta rule keeps each row a convex combination of its old value and the gated targets. The writer is trained offline on raw text with the pretraining objective; at use time writing takes forward passes only, and clearing \Delta restores the original model. A 0.6B prototype trained to answer questions keeps 16/16 facts after a second write; online gradient writes keep 2–4/16. We study two writers: a plain writer that encodes evidence with the T^0-only model, and a delta writer that reads its current memory while encoding. On Qwen3.8-Flash-Next, the plain writer’s memory is document-specific: on 64 held-out news events it lowers continuation NLL by 0.068 nats/token against empty memory and by 0.024 against another document’s memory, and 0.063 and 0.013 after 16 intervening articles. The delta writer’s memory is more document-specific and keeps more of it under distraction (70% vs 54%). Over 35,000 writes on our previous seven-task chain both writers stay stable and the delta writer drifts less (layer-2 math drift 176% vs 188%); later accuracy changes by -0.26/-0.52pp (-7.29pp for gradient updates) and small-k_i drift is 17.9/17.5% (39.2%). Acquisition is negative, and general-input drift reaches 32.8% (20.75%): format rows read by every example grow to 4–12 times their pretrained norm. Zeroing eight such rows removes over 98% of layer-2 drift on math inputs, together with their NLL gain.
[NLP-37] hrough a PRISM Clearly: Step-Sequence Reasoning for Mistake Recognition in Procedural Videos
【速读】： 该论文旨在解决从视频中识别流程性操作错误的问题，即在多步骤任务完成后，检测执行过程中存在的步骤内容或顺序偏差。其核心挑战在于如何有效建模正确执行流程的语义状态，并据此识别偏离正常流程的错误。解决方案的关键在于提出PRISM（Procedural mistake Recognition via Interpretable Semantic Modeling），该方法通过将视频片段与语言描述进行语义对齐，利用基于正确执行样本训练的轻量级隐马尔可夫模型（Hidden Markov Model, HMM）映射视频描述到程序状态，进而通过参考对齐和语义校验机制识别步骤内容与顺序错误。为实现跨异构基准的统一评估，研究将数据集特有的标注映射为四类基于编辑的操作错误类型：插入、删除、替换与错位。实验表明，采用训练过的局部解码器相比最近邻匹配显著提升性能，序列解码在特定数据集上带来额外增益；同时，语义角色标注在全部六组匹配的地面实况比较中均提升了Any-F1指标，但其增益随数据集不同而异，且未在独立配置选择下呈现一致优势。结果表明，语言接地（language grounding）作为显式流程错误推理的有效接口，其表现受表示学习与解码策略共同影响，验证了语义建模与结构化解码在流程错误识别中的关键作用。
链接: https://arxiv.org/abs/2610.31117
作者: Olga Loginova,Oleksandr Bezrukov,Frank Keller
机构: University of Trento(特伦托大学); Gran Sasso Science Institute(格兰萨索科学研究所); University of Edinburgh(爱丁堡大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Procedural mistake recognition requires identifying deviations from valid multi-step activities. We study post-completion recognition from videos with benchmark-provided step boundaries. We introduce PRISM (Procedural mistake Recognition via Interpretable Semantic Modeling), which grounds video segments in language and maps their descriptions to procedural states using a lightweight hidden Markov model trained on correct executions. Reference alignment and semantic checks identify errors in step content and order. To support evaluation across heterogeneous benchmarks, we map dataset-specific annotations to four edit-based error types: insertion, deletion, substitution, and transposition. Across three benchmarks, we compare gold and VLM-generated descriptions in raw, semantic-role, and action-object forms. Decoder controls show consistent gains from the trained local decoder over nearest-candidate matching, with additional, dataset-dependent benefits from sequence decoding. Semantic role labeling improves Any-F1 in all six matched grounded comparisons, although the gains vary across datasets and do not translate into uniform improvements under separate configuration selection. These results establish language grounding as an effective interface for explicit procedural mistake reasoning and characterize the contributions of representation and decoding.
[NLP-38] A Grammar of French Civil Cassation: A Ground-Level Dataset and Tool
【速读】： 该论文旨在解决法国最高法院民事庭在法律上诉裁决中，如何系统化地解析和建模上诉理由（ground of appeal）与其对应裁判理由（reasons）之间关系的问题。由于法律上诉（appeal on points of law）具有终局性且决定案件走向，理解其裁判逻辑对律师起草上诉状具有重要实践意义。解决方案的关键在于构建一个确定性处理流程（deterministic pipeline），将86,464份判决书中121,536组“上诉理由-裁判理由”进行精确配对，并依据既有的法国判例法分类体系（doctrinal taxonomy）将其归类至特定的法律教义类别。通过该方法，研究实现了对民事上诉审判实践的描述性语法建模，揭示了各类上诉理由的成功率、所援引的法律条文及其在十年间的演变规律。研究成果包括公开的数据集、可复现的代码及交互式查询工具，为法律实证研究与智能辅助决策提供了基础资源。
链接: https://arxiv.org/abs/2610.31116
作者: Avrile Floro,Tamara Dhorasoo,Nils Holzenberger,Thomas Le Goff,Tiphaine Viard,Maria Boritchev
机构: Télécom Paris, Institut Polytechnique de Paris (法国国立巴黎高等电信学院，巴黎理工学院); Université Polytechnique Hauts-de-France (上法兰西理工大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: Accepted at JURIX 2026 (39th International Conference on Legal Knowledge and Information Systems)
Abstract:This paper presents a dataset of 121,536 ground-reasons pairs across 86,464 rulings of the five civil chambers of the French Court of Cassation (2016-2025). An appeal on points of law cannot itself be appealed and often decides a case for good. Understanding how the Court rules is thus crucial and can help lawyers who draft those appeals. Inside a ruling, every ground of appeal (a distinct legal complaint against the appealed decision) has its own answer from the Court. This answer is called the reasons and it belongs to one of a few doctrinal families. Using a deterministic pipeline, we pair each ground with its reasons and attach the pair to its family, according to the established French taxonomy. We obtain a descriptive grammar of civil cassation that shows which grounds succeed, which legal texts they refer to, and how both evolve over the decade. We release the dataset, the code and an interactive consultation tool.
[NLP-39] Learning What to Ask for Tomorrow: Active Preference Acquisition for Long-Term Personalized Web Agents
【速读】： 该论文旨在解决生成式Web代理（Web agents）在长期交互中因缺乏用户偏好信息而导致个性化能力受限的问题。现有被动记忆与个性化方法无法获取用户未主动披露的偏好，而基于当前任务收益的澄清策略则忽视了已知偏好的跨任务复用价值。为此，论文提出主动偏好获取（Active Preference Acquisition, APA），其核心是通过引入“前瞻性个性化价值”（Prospective Personalization Value, PPV）来量化某一未知偏好的预期累积效用增益——即未来多个任务中重复使用该偏好所带来的总收益，减去提问所付出的努力与隐私泄露风险成本，并在每轮交互预算约束下选择最优问题。研究发现，当单一任务收益不足时，若该偏好具备高重用潜力，则仍应被获取；这一机制依赖于对任务家族重用频率的建模。作者构建了ActiveWebPref基准数据集，通过模拟具有隐藏但可查询偏好的用户行为，验证了APA在20个任务周期内的有效性：34.4%的高价值缺失偏好属于“摊销依赖型”，即仅因跨任务复用才值得获取，而APA成功捕获其中80.3%，远超仅考虑当前任务价值的信息获取方法（5.1%）。在符合实际询问成本限制的条件下，APA实现的净效用高于四种启发式基线方法，且仅以27.6%的提问比例回收84.2%的潜在效用增益。研究表明，偏好获取的价值必须充分考虑其在未来任务中的复用效应，而非局限于单次任务收益。
链接: https://arxiv.org/abs/2610.31107
作者: Yiqi Wang,Zizhi Zeng,Jiaqi Zhang,Yiqun Duan,Mingkai Zheng,Taotao Cai
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Web agents increasingly serve the same user across repeated online tasks such as booking travel, shopping, and scheduling, so their long-term usefulness depends on how well they know the user’s preferences. Passive memory and personalization methods cannot obtain preferences the user never reveals, and clarification methods that value a question by its benefit to the current task overlook the benefit of reusing the answer in later tasks. We study Active Preference Acquisition (APA): which missing preferences are worth the cost of asking? We define the Prospective Personalization Value (PPV) of a missing preference as its expected cumulative utility gain over future tasks minus the effort and disclosure-risk cost of asking, and select questions under a per-interaction budget. We characterize when cumulative reuse justifies acquisition despite insufficient single-task benefit, and evaluate this regime with an initial-profile diagnostic. APA computes the counterfactual gain of each candidate preference on public task prototypes with the agent’s own task solver and aggregates it by the expected reuse of each task family. We evaluate it on ActiveWebPref, a controlled longitudinal benchmark with hidden but queryable user profiles and exact task utilities. At a horizon of 20 tasks, 34.4% of budget-feasible valuable missing preferences are amortization-dependent; APA acquires 80.3% of them, whereas current-task value of information acquires 5.1%. Under per-user asking-cost caps matched to APA’s realized expenditure, APA achieves higher net utility than four heuristic selection baselines. In our cost setting, asking about every unknown preference yields lower net utility than not asking, while APA recovers 84.2% of its utility gain with 27.6% of its questions. The value of acquiring a user preference should account for its reuse across future tasks, not only its benefit to the task at hand.
[NLP-40] RIFT: KV Cache Compression for Draft Models with Complementary Compressibility Training
【速读】： 该论文旨在解决长序列生成任务中因输出长度增加而导致的键值缓存（KV cache）占用过高问题，尤其针对基于选择性压缩（selection-based KV cache compression）方法在长输出场景下出现的“解码陈旧性”（decode staleness）现象——即预填充阶段信息随生成过程推进逐渐失效，进而降低草案模型（drafter）的准确性。其解决方案的关键在于提出RIFT系列选择性压缩策略，通过在生成过程中动态刷新KV缓存的选择，有效平衡早期输入信息与近期生成标记的相关性，从而显著提升草案模型的质量。为缓解频繁刷新带来的额外解码延迟，研究进一步设计了一种自适应的KV缓存刷新机制，以减少评分计算次数并保持草案质量；同时引入基于稀疏注意力掩码的可压缩性驱动训练范式，增强草案模型在压缩缓存下的表现。实验表明，结合压缩性训练的RIFT策略使Qwen3.5-4B在对齐Qwen3.5-122B进行长文本生成时，在32K输入、4K输出及8K缓存预算条件下，仍能维持超过98%的完整缓存接受长度。
链接: https://arxiv.org/abs/2610.31098
作者: Dor Tsur,Maor Ashkenazi,Maximilian Golub,Peter Hicks,Yaniv Galron,Talor Abramovich,Yoav Miron,Tiyasa Mitra,Bita Darvish Rouhani,Ran Zilberstein
机构: NVIDIA(英伟达)
类目: Computation and Language (cs.CL)
备注:
Abstract:Speculative decoding is a popular decode acceleration method. As such, it is particularly relevant to long-output generation, which requires an increasingly large KV cache. KV cache compression offers a way to constrain costs in long-context drafting. However, compression methods that focus on prefill information can lose relevance as the output length grows. We term this phenomenon decode staleness and show that it degrades draft model accuracy under selection-based KV cache compression. We introduce RIFT, a family of selection-based KV cache compression strategies that balance prefill information and recent generated tokens. We show that refreshing the KV cache selection during token generation significantly improves the drafter’s quality at the cost of additional decode latency. To alleviate the refresh costs, we propose an adaptive KV refresh strategy that reduces the number of score calculations while retaining the drafter’s quality. We further demonstrate how a compressibility-driven drafter training phase, based on sparse attention masks, can improve the drafter’s quality under a compressed KV cache. Combined with compressibility training, RIFT enables Qwen3.5-4B, drafting for Qwen3.5-122B, to retain over 98% of its full-cache acceptance length on 32K-token inputs, 4K-token outputs, and an 8K-token KV cache budget.
[NLP-41] What Pairwise Evaluation Misses: Reward Scores across Prompts
【速读】： 该论文旨在解决奖励模型（Reward Model）在跨提示（prompt）评分时的可比性问题，即当前基于成对评估的训练方式导致同一提示下不同响应共享的提示相关得分成分被抵消，无法被有效检验。其核心问题是：奖励模型的评分中存在显著的“提示间方差”（between-prompt variance），而这一部分是否反映真实的人类偏好尚不明确。解决方案的关键在于提出一种基于人类参考的审计方法（human-referenced audit），通过三个关键组件实现：（1）计算提示间方差占比（between-prompt variance share），（2）构建兄弟评分回归（sibling-rating regression）以分离人类与模型评分的差异，（3）引入提示替换（prompt substitutions）进行敏感性测试。通过对18个已发布的评分器和6种匹配训练条件的应用发现，模型评分的提示间方差显著高于人类评分（如HelpSteer2上模型方差占比达0.34–0.64，人类接近0），且即使固定提示后仍存在强烈聚类现象，但与人类提示级评分的一致性下降，表明聚类并非人类偏好一致性的体现。进一步分析显示，提示均值信息具有实际意义——移除提示均值会降低全局最优响应的有用性，且模型对错误提示仍保持敏感，说明提示相关信号虽非完全反映人类判断，但对选择性能仍具价值。因此，论文主张对奖励模型的评估应独立考察三方面：提示间变异、对错误提示的敏感性以及对响应选择的实际效用。
链接: https://arxiv.org/abs/2610.31090
作者: Mateusz Modrzejewski,Tomasz Radzikowski
机构: Warsaw University of Technology (华沙理工大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 31 pages, 5 figures, 33 tables
Abstract:Reward models score responses for preference optimization, filtering and ranking, and pooling scores across prompts needs them to be comparable. Yet pairwise evaluation compares two responses to the same prompt, so any score component shared by a prompt’s responses cancels and goes untested. We study this component by comparing between-prompt variation in model scores with the same variation in human ratings of the same responses. We propose a human-referenced audit, built from a between-prompt variance share, a sibling-rating regression and prompt substitutions, and apply it to eighteen released scorers and six matched training conditions. We find that scores cluster by prompt more than the ratings do: on HelpSteer2, six reward models have between-prompt shares of 0.34 to 0.64 against a human share near zero. Replacing the original prompt with a fixed instruction keeps substantial clustering but weakens agreement with prompt-level human ratings, so clustering alone does not show agreement with humans. Part of that variation is useful: subtracting prompt means lowers the helpfulness of globally selected responses, and scores stay sensitive to a wrong prompt even when training shuffles prompt means. We propose that reward evaluation check between-prompt variation, sensitivity to a wrong prompt and usefulness for selection separately.
[NLP-42] SCOPE: Compiling Natural-Language Policies into Neurosymbolic Predicate Programs
【速读】： 该论文旨在解决大语言模型（LLM）智能体在实际应用中面临的复合型、多轮次策略执行（compound, multi-turn policy enforcement）难题，尤其针对现有方法在处理复杂策略时存在的准确性下降、推理成本高及预测不稳定等问题。其核心挑战在于如何在保证高效性的同时实现对复杂政策的精确、可解释且稳定的执行。解决方案的关键在于提出一种名为SCOPE的神经符号框架：将策略编译为由小型神经谓词（neural predicates）、确定性提取器与逻辑组合代码构成的可执行程序，通过联合优化程序与训练数据，并采用对抗性反例生成机制主动发现并修正程序与策略之间的偏差。该方法实现了在低成本模型上的高效部署，同时保持与前沿推理模型相当的准确率，且预测结果具有高度可解释性——每个判断均可追溯至具体触发的谓词输出。此外，经最弱执行器优化的程序具备良好的迁移能力，无需重新训练即可适配不同模型，显著提升了生产环境下的可靠性与实用性。
链接: https://arxiv.org/abs/2610.31062
作者: Arnon Mazza,Elad Levi
机构: Plurai
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Policy enforcement for LLM agents has shifted from single-turn content moderation to compound, application-specific policies evaluated over multi-turn transcripts and tool-call trajectories. Fine-tuned guardrail models are efficient but degrade on compound policies, and while LLM-as-a-judge approaches achieve higher accuracy, they incur high inference cost and tend to produce unstable predictions. Neurosymbolic methods offer a natural alternative, decomposing the policy into narrow neural decisions composed under explicit logic. However, existing symbolic layers lack the required expressiveness and are optimized against fixed data rather than the policy itself. To address this, we introduce SCOPE, which compiles a policy into an executable program of neural predicates executed by small models, deterministic extractors, and code that composes their outputs. The program is optimized jointly with its training data, using adversarial generation of counterexamples that target where the program deviates from the policy. We evaluate SCOPE on a benchmark of real-world policy tasks against prompt optimization, task decomposition, and workflow generation methods, achieving the highest average accuracy across tasks and models. Programs optimized with the weakest executor transfer to other models without further optimization. This allows them to be deployed in production on inexpensive models, where they reach accuracy comparable to frontier reasoning models. Unlike these monolithic models, SCOPE produces substantially more stable predictions, and each verdict is interpretable through the predicate outputs that produced it. Additionally, we validate the key components of our method through an ablation study. SCOPE thus enables reliable, interpretable, and efficient policy enforcement at production scale.
[NLP-43] Harm Is Behavioural Moderation Is Lexical: How Language and Culture Decide Whom Automated Safety Systems Protect and Whom They Silence
【速读】： 该论文旨在解决当前自动化内容安全系统在跨语言语境下对有害行为识别存在显著偏差的问题，具体表现为对非标准语言变体（如非洲裔美国人英语）的过度标记（over-flagging）以及对低资源语言群体的保护不足（under-protection）。其核心问题在于现有系统依赖词汇表层形式（vocabulary and surface form）判断危害性，而非基于信息的实际功能（harmful function），导致在不同语言和文化语境中表现不一致。论文提出的关键解决方案是：将“有害行为”（如秘密请求、付款要求、孤立行为）视为跨语言可迁移的抽象功能，主张应以功能为导向进行检测，而非依赖表面表达。为此，作者构建了一种新的评估协议，强调最佳与最弱势语言社区之间的性能差距作为核心指标，并通过实证发现：当前五种开放系统的多语言表现均未达到母语者水平（55组系统-语言组合中仅4组达标），且系统能力提升反而加剧了部分语言间的歧视，尤其在阿拉伯语和罗马化印地语中仍接近随机水平；同时揭示了单一系统可能在同一语言上既过度标记又保护不足的矛盾现象。此外，研究还指出，仅报告假阳性率会误导性地掩盖真实不公平性，凸显了以功能为中心的评估框架的重要性。
链接: https://arxiv.org/abs/2610.31047
作者: Gabriel Lars Sabadin,Dr Nicola Harding,Sarah Kingston
机构: Tuteliq AB(瑞典); Tuteliq AB(英国); University of Lancashire(兰开夏大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 18 pages, 2 figures. Code, simulations and per-case outputs: this https URL
Abstract:Content warning: this paper quotes slurs and hateful statements from published test suites; the racial slur among them is masked. Automated systems decide at scale which messages are harmful, and they decide unevenly. Classifiers label African American English as offensive up to twice as often as other English, the accounts of prominent drag queens can score as more toxic than those of white nationalists, and safety behaviour that holds in English collapses in low-resource languages. We argue that over-flagging and under-protection share one cause: deciding harm from vocabulary and surface form rather than from what a message does. We organise the evidence into seven failure modes and argue, as testable hypotheses, that harmful behaviours (a request for secrecy, a demand for payment, an attempt to isolate someone) travel across languages better than the words that express them. We propose an evaluation protocol that makes the gap between the best- and worst-served language communities a headline result, and show that the naive gap is inflated under equal service and that its usual bootstrap interval is unreliable. Applied to five open systems in eleven languages, from a keyword lexicon to two current multilingual guard models, the protocol finds that none reaches native-speaker accuracy at separating harmful function from harmless form (51 of 55 system-language pairs fall short); that greater capability raises discrimination in nine languages while leaving Arabic and romanised Hindi near chance, so the gap between the best- and worst-served languages grows with capability; that one system both over-flags and under-protects the same language most; and that a language a system reads worst can look the most fairly treated when only false positives are reported. Comments: 18 pages, 2 figures. Code, simulations and per-case outputs: this https URL Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) ACMclasses: I.2.7; K.4.1 Cite as: arXiv:2610.31047 [cs.CL] (or arXiv:2610.31047v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2610.31047 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-44] Mostly Reading: What Open-Book Scores Measure in Document-Level LoRA Adapters ACL
【速读】： 该论文旨在解决单文档低秩适配器（Single-document LoRA adapters）在知识存储与检索中的有效性问题，特别是其在开放书式评测（open-book scoring）中是否真正实现了对特定文档知识的专有记忆。研究发现，尽管适配器在开放书式任务中表现出显著优势，但这种优势主要源于模型普遍存在的“阅读行为”而非对具体文档内容的专属性存储；随机训练的适配器在开放书式任务中可恢复82%的“专家适配器”（specialist）优势，但在封闭式填空探测任务（cloze probe）中仅能恢复2%的优势，与零无显著差异，表明内容虽已部分进入权重，但未形成稳定、可区分的知识表征。关键在于：适配器的更新几何结构虽与所读文档一致，但其在实际检索中无法通过权重空间或文本嵌入的相似性被有效识别，且多适配器合并时，仅能恢复约12%-16%的闭合式任务优势，远低于开放书式任务中的87%-89%，揭示了当前方法在知识定位与泛化之间的根本矛盾。此外，研究还报告了四起“无声错误”（silent errors），即模型生成看似正确实则错误的结果，进一步暴露了基于适配器的知识系统在可靠性上的深层缺陷。
链接: https://arxiv.org/abs/2610.31017
作者: Martino M. L. Pulici,Cuong Xuan Chu,Evgeny Kharlamov,Volker Tresp
机构: Bosch Center for Artificial Intelligence(博世人工智能中心); LMU Munich(慕尼黑大学); Munich Center for Machine Learning(慕尼黑机器学习中心)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 27 pages, 3 figures, 15 tables, under review at ACL ARR
Abstract:Single-document LoRA adapters have been proposed as units of knowledge to be merged and retrieved; yet open-book scoring, with the document in the prompt, can credit a document’s own adapter (its specialist) for reading that document as readily as for holding it. Training one adapter for each of 150 Wikipedia articles on Llama-3.1-8B-Instruct, we find that open-book scores mostly measure a reading behaviour every adapter shares: a random adapter trained on another document recovers 82 % of the specialist’s open-book advantage over the base model but only 2 % of its advantage on a cloze probe that asks for the same document’s facts without providing the source text, a share we cannot distinguish from zero. Content does reach the weights, since the specialists beat the base model in cloze recall by 0.094, and the geometry of their updates follows the documents they read; yet, in a test of 30 queries, adapters chosen by weight-space proximity or text embedding are not distinguishable from random ones on either probe, although that test cannot exclude an advantage of about four fifths of what a query’s own adapter gains over a random one. At 16 documents or more, correctly scaled merges recover 87 % to 89 % of the specialists’ open-book advantage but only 12 % to 16 % of their cloze advantage, with intervals that run from below zero to at most 32 %, against 94 % for joint this http URL also report four silent errors, each of which produced a clean, wrong result.
[NLP-45] Padding to k: A Minimum Count in the Schema Makes Open-Weight Models Pad Their Extraction Lists
【速读】： 该论文旨在解决生成式 AI 在执行结构化信息抽取任务时，面对模式（schema）中设定的最小列表项数要求而实际文档内容少于该数量时所表现出的“过度填充”（over-padding）问题。其核心挑战在于：当模型被要求输出符合指定最小长度的 JSON 列表时，即使真实文本中包含的条目不足，模型仍倾向于生成额外的虚假或冗余条目以满足该约束，从而导致结果失真。解决方案的关键在于揭示这一行为的本质——模型将最小数量视为目标而非强制约束，因此会主动“补足”缺失项；而这些补足项往往并非随机生成，而是从原始文档中提取的真实文本片段，使得检测难度极大。此外，论文评估了多种常见缓解策略（如放宽约束提示、使用“最多k个”的自然语言描述、添加“N/A”选项及重试验证错误），发现它们均无法彻底消除过度填充现象，部分甚至加剧问题。研究结果表明，当前开放权重模型与托管服务在该问题上表现一致，且现有方法难以有效识别和纠正此类偏差，凸显出对更鲁棒的验证机制和模型行为控制策略的需求。
链接: https://arxiv.org/abs/2610.30997
作者: Abhishek Barali,Aayush Paudel
机构: St. Xavier’s College; Maitighar, Kathmandu; Tribhuvan University, Nepal
类目: Computation and Language (cs.CL)
备注: 26 pages, 4 figures. Preregistered at this https URL . Code and data: this https URL
Abstract:Extraction pipelines ask language models for JSON that follows a schema, and the schema can set a minimum number of list entries. We ask what models do when the document holds fewer items than that minimum. Under a preregistered plan, we test eight open-weight models on 558 real documents from nine annotated datasets, with six more models on hosted services as a reference. Three findings stand out. First, models treat the count as a target: a minimum shown in the schema, but not enforced, raises the share of lists longer than the true count from 34.5% to 55.4%. Second, the padding is hard to catch, because most added entries are real text taken from the document; about half of them pass a check that each entry occurs in the source. Third, common fixes help only partly: a line saying fewer items are fine removes the effect on average but misses more true items, a count in prose (“up to k”) also raises over-counting, an “N/A” option under an enforced minimum still leaves half the lists padded, and re-asking with the validation error makes padding far more common. We release the code, answer lists and run records at this https URL
[NLP-46] ColaBench: Evaluating Memory-Guided Knowledge Acquisition in Personal Agents
【速读】： 该论文旨在解决个人代理在处理用户当前请求时，如何有效利用长期交互历史来识别并应用与用户情境相匹配的外部知识这一核心问题。其关键挑战在于：一方面需基于历史条件精准推断当前任务所需的知识，另一方面需将检索到的知识适配至用户的实际情境中进行合理应用。现有基准主要聚焦于记忆召回、个性化检索或记忆引导行为，但缺乏对上述两个挑战协同作用的联合评估。为此，论文提出ColaBench——一个涵盖健康与日常生活场景的合成基准，包含10个用户画像、349项任务及13,033份文档构成的知识库，并构建了证据图（Evidence Graph）以关联个人状态与其在公开交互会话中的披露信息、先前文档接触记录，以及答案需求与个人及新获取文档证据之间的联系。同时，论文提出MemKB系统，通过状态摘要指导查询生成与规则应用，并借助记忆-文档关联机制实现基于用户历史的知识检索。实验表明，MemKB在平均4.07次交互轮次下达到86.25%的准确率；任务层面分析显示，跨多个历史事实整合信息的任务收益更显著；错误分析揭示出未经验证的个人假设和规则误用是阻碍有效利用检索证据的主要障碍。ColaBench为研究记忆如何引导知识获取与规则应用提供了可追溯的任务设计与评估框架。
链接: https://arxiv.org/abs/2610.30989
作者: Wenjie Zhang,Zhichao Shi,Xuhui Jiang,Xiaojun Wu,Cehao Yang,Chengjin Xu,Xiaoning Song
机构: Jiangnan University (江南大学); Institute of Computing Technology, Chinese Academy of Sciences (中国科学院计算技术研究所); The Hong Kong University of Science and Technology, Guangzhou (香港科技大学（广州）); Institute for Data Science and Artificial Intelligence (数据科学与人工智能研究院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 12 pages, 4 figures
Abstract:Personal agents need to draw on long-term interaction history to retrieve and apply external knowledge suited to a user’s current request. Even an answer grounded in relevant documents can be unsuitable when it overlooks the user’s circumstances. This creates two connected challenges: using historical conditions to identify the knowledge needed for a new request, and applying that knowledge to the user’s situation. Existing benchmarks advance memory recall, personalized retrieval, and memory-guided action, while the joint evaluation of these two challenges remains underexplored. We introduce ColaBench, a synthetic benchmark in health and daily living, with 349 tasks across 10 Profiles and a knowledge base of 13,033 documents. Its Evidence Graph connects personal conditions to their disclosures across public interaction sessions, records prior document exposure, and links answer requirements to personal and newly acquired document evidence. We also introduce MemKB, which uses state summaries to inform queries and rule application, and memory-document links to retrieve knowledge associated with the user’s history. MemKB achieves 86.25% Accuracy with an average of 4.07 interaction rounds per task. Task-level analysis finds larger gains on tasks integrating information across multiple historical Facts; error analysis further identifies unsupported personal assumptions and rule misapplication as obstacles to using retrieved evidence. ColaBench provides traceable tasks and an evaluation framework for studying how memory guides knowledge acquisition and rule application.
[NLP-47] Interpretive Variation and Discourse Structure in Detecting Enthymemes in Controversial Political Discourse
【速读】： 该论文旨在解决社交媒体文本中隐含前提（enthymeme）的检测与评估难题，尤其针对标注者间存在显著分歧这一核心挑战。传统方法在不同数据规模、特征引入或标签类型下均表现平庸，F1值稳定在0.67左右，表明基于多数投票的监督学习范式难以有效捕捉此类语义复杂且依赖读者推断的论证结构。为此，论文提出一种新评估框架，通过比较模型预测与标注者投票分布的整体差异来衡量性能，并设计了一种直接建模标注分布的生成式模型，在不降低F1的前提下更精准地反映人类判断的不确定性。进一步分析揭示，现有模型高度依赖显式论辩结构，虽能可靠识别无隐含前提的文本，却在面对需读者自行补全逻辑链的隐含论证时表现出严重盲区；同时，人工话语结构解析器同样无法识别隐含推理关系。这一发现对基于多数标签训练的内容审核系统提出了警示：其决策机制可能忽视了大量依赖语境推断的说服性表达，从而导致误判与偏见。
链接: https://arxiv.org/abs/2610.30987
作者: Martial Pastor,Nelleke Oostdijk
机构: Radboud University (奈梅亨大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Enthymemes are arguments that leave a premise or conclusion unstated, and they are a common persuasion device on social media. Because readers must fill in the missing part themselves, they often disagree on whether a text contains one. This paper asks how to detect and evaluate enthymemes when such disagreement is part of the task. On tweets about vaccination and immigration, each labeled by five annotators, F1 stalls around 0.67 for every fine-tuned model we train, regardless of data size, added features or label type. We therefore ask how these models should be evaluated instead, and propose metrics that compare predictions with the full distribution of annotator votes, together with a model that predicts this distribution directly and captures disagreement better at no cost in F1. We then ask what role discourse structure plays. Models learn to rely on explicit argumentative structure: they reliably recognize tweets without an enthymeme but rarely commit on tweets that contain one. Both the models and a discourse parser miss arguments whose links are left implicit and must be inferred by the reader. Finally, we discuss what these blind spots mean for content moderation systems trained on majority labels.
[NLP-48] A Reality Check of Inference-Time Context Augmentation Methods
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在推理过程中未能有效利用输入上下文，导致幻觉或依赖参数化知识的问题。现有上下文增强方法通过修改输入来提升模型性能，如插入填充标记以增加计算时间，或添加相关语义内容进行上下文扩展。然而，这些方法通常被孤立评估，仅关注性能提升，缺乏对不同方法之间可比性及可靠性的真实效果分析。本文提出一个涵盖五种不同信息丰富度插入策略的评估框架，在句子层面针对七种不同规模的模型于三个数据集上进行系统比较。研究发现，部分方法带来的准确率提升并非真正的性能改进，而是无方向性的扰动：某些预测被修正的同时，另一些却遭到破坏，尤其当插入内容仅延长计算时间时更为明显。进一步分析表明，仅引入非语义内容的方法主要影响模型决策边界附近的预测结果。该研究强调，在实际应用中采纳任何上下文增强技术前，必须对其准确性提升的可靠性进行严格评估。
链接: https://arxiv.org/abs/2610.30983
作者: Haeun Yu,Geert Heyman,Pepa Atanasova,Isabelle Augenstein
机构: University of Copenhagen(哥本哈根大学); Nokia Bell Labs(诺基亚贝尔实验室)
类目: Computation and Language (cs.CL)
备注: 19 pages, 8 figures
Abstract:Large language models (LLMs) often fail to utilise the context given in the input, resulting in hallucinations or a fallback to parametric knowledge. Context augmentation methods aim to improve model performance by modifying the input, from inserting filler tokens that add computational time to adding relevant content such as elaboration of a context during inference. However, these methods have been evaluated in isolation and only by their performance gains, leaving it unclear how they compare and whether they reliably improve performance. In this paper, we present an evaluation framework spanning five methods that differ in how informative the inserted content is and apply them at the sentence level across seven models spanning multiple sizes on three datasets. We find that accuracy gains are not always genuine improvements but undirected perturbations: the methods correct some predictions while breaking others, particularly when the inserted text allows only for more computational time. Our analysis reveals that methods that bring no additional content primarily change model predictions near the decision boundary. Our findings highlight the need to evaluate the reliability of the accuracy gains before adopting them in practice.
[NLP-49] Joint or Per-Option? Compact Encoders for Closed-Set Decisions from 17 Million to 1 Billion Parameters
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）在流水线末端进行封闭集决策（如工具选择）时存在的效率与性能问题。现有方法依赖提示（prompting）LLM进行决策，存在速度慢、对输入顺序敏感等缺陷。研究核心在于探究此类决策任务中编码器的最小规模及其最优读取策略：是应联合读取所有选项（joint scoring），还是逐项独立评分（per-option scoring）。研究表明，尽管每选项评分需5至13倍于联合评分的编码器调用次数，其在相同优化器步数下比联合评分在37个基准组成的测试套件上提升5.8点机会校正技能（chance-corrected skill），在近似训练计算量下仍领先2.8点；两个外部基准亦支持该优势，尽管在未见任务上未达显著水平。进一步分析表明，该优势依赖于“重排序器初始化”（reranker initialization），其输入结构匹配逐项读取模式，而原始编码器状态下两种读取方式无差异。此外，重排选项会改变联合模型12%至23%的预测结果，而逐项评分则不受影响。在6个公开发布的联合编码器中（参数量1700万至10亿），6800万和1.5亿参数模型在209个经审计的System1-Arena任务上分别超越4.35亿参数的DeBERTa-v3零样本分类器4.9和5.9个百分点，且在145个训练未覆盖任务上表现至少相当。最小模型可在16核处理器上以14毫秒完成中位请求响应，远优于分类器的728毫秒。在不超过6800万参数范围内，这些模型在公共排行榜的私有、主要分布内测试中表现领先。因此，解决方案的关键在于采用基于逐项评分的轻量化重排序器架构，并配合特定初始化方式以实现高效且鲁棒的决策能力。
链接: https://arxiv.org/abs/2610.30970
作者: Tardelli Ronan Coelho Stekel
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 18 pages, 8 figures, 8 tables. Code: this https URL ; models: this https URL
Abstract:Many language-model pipelines end in a closed-set decision, such as selecting a tool. Prompting a large language model for it is slow and order-sensitive. We ask how small an encoder for such decisions can be and whether it should read the options jointly or separately. In a three-seed study at 32 million parameters, per-option scoring beats joint scoring on a held-out sample of a 37-benchmark decision suite by 5.8 points of chance-corrected skill at matched optimizer steps and by 2.8 at approximately matched training compute; two external benchmarks agree, though not significantly on their unseen tasks. Crossing reader with initialisation shows that the advantage requires a reranker initialisation, whose query-candidate pair input matches per-option reading; from the raw encoder the two readers are indistinguishable. Per-option scoring needs 5 to 13 times more encoder passes: at most twice the median latency, but more than four times the mean. Of six released joint encoders (17 million to 1 billion parameters), the 68- and 150-million-parameter models exceed a 435-million-parameter DeBERTa-v3 zero-shot classifier by 4.9 and 5.9 accuracy points on 209 overlap-audited System1-Arena tasks and are at least on par with it on the 145 tasks unseen in training. The smallest answers a median request in 14 milliseconds on 16 processor threads, against 728 for the classifier. Reordering options changes 12 to 23% of the joint models’ answers (per-option scoring: none). Up to 68 million parameters, they lead their size range on a public leaderboard’s private, mainly in-distribution tests.
[NLP-50] Once a Tree Always a Tree? Cross-lingual Transfer of Hierarchical Generalization in Language Models EMNLP2026
【速读】： 该论文旨在解决生成式语言模型（Generative Language Models, GLMs）在跨语言持续预训练过程中，其跨语言迁移能力的内在机制不明确的问题。具体而言，研究关注的是：模型是否能够从源语言（L1）中习得的语言层级结构（hierarchical structure）等语言普遍性原则，有效迁移到目标语言（L2），从而实现对语法结构的泛化。其解决方案的关键在于提出一种新的实验范式——跨语言刺激匮乏（Cross-lingual Poverty-of-Stimulus, CPOS）设置，通过控制源语言侧的层级结构正向证据量，同时使目标语言语料在句法规则上存在线性与层级结构之间的歧义，从而量化评估语言模型在缺乏充分目标语言监督的情况下，能否基于源语言提供的结构线索实现层级泛化。实验结果表明，模型可部分但非完全地将源语言的结构证据迁移到目标语言，仅产生微弱但可检测的层级偏好，且迁移效果具有显著的语言对依赖性。进一步消融分析揭示，词汇重叠度是实现成功迁移的关键条件，表明跨语言语法结构迁移并非普遍成立，而是受制于语言间的词汇相似性，因而具有条件性。
链接: https://arxiv.org/abs/2610.30960
作者: Shin Yamaguchi,Masaya Taniguchi,Tatsuki Kuribayashi,Taro Watanabe,Yoshinobu Kano
机构: NAIST(日本信息科学研究所); RIKEN(理化学研究所); MBZUAI(穆巴达拉人工智能大学); Shizuoka University(静冈大学); Tohoku University(东北大学)
类目: Computation and Language (cs.CL)
备注: EMNLP 2026 main
Abstract:Language models (LMs) often exhibit successful cross-lingual transfer through continued (pre)training, still their underlying process remains unclear. One possibility is that LMs can leverage the evidence toward language-universal principles, namely, hierarchical structure, from a source language (L1) to a new language (L2), while many other non-linguistic explanations, e.g., improved initialization of position embeddings, can be potentially possible. In this study, to quantify the cross-linguistic transferability of linguistic evidence in cross-lingual LM training, we introduce a new experimental design, namely, a cross-lingual poverty-of-stimulus (CPOS) setting that ablates how LMs’ hierarchical generalization in L2 can be led by structural evidence provided in L1. We contrast counterfactual L1 and L2 learning scenarios where the L2 corpus is ambiguous between linear and hierarchical grammar rules, while the amount of positive evidence for the hierarchical rule is controlled on the L1 side. Our experimental results show that LMs can partially, yet not perfectly, transfer L1 structural evidence to L2, yielding only a small but detectable hierarchical bias, while the strength of transfer depends on language pairs. We further ablate the conditions for successful transfer, motivated by recent reports on the substantial impact of vocabulary overlap. We indeed find that vocabulary overlap is crucial even for the cross-lingual transfer of hierarchical syntactic generalization. This suggests that this also explains the language-pair-dependent tendencies in the main experiments and suggests that the transfer of linguistic evidence in continued new-language learning is only conditionally achievable.
[NLP-51] IQuest-Q1: Advancing Foundation Capabilities for Agent ic CLI Systems
【速读】： 该论文旨在解决大模型在生成式任务中参数效率与能力扩展之间的矛盾，特别是如何在保持训练稳定性的同时实现高效、可扩展的智能体基础模型（agentic foundation model）构建。其核心挑战在于：在大规模参数量下维持训练收敛性，同时增强模型在复杂编程任务和通用推理任务中的表现。解决方案的关键在于提出一种三阶段渐进式训练框架：第一阶段通过引入稳定分组归一化（Stable SplitNorm）、Muon优化器与muP缩放策略，并结合双阶段课程学习（two-stage curriculum），显著提升预训练阶段的稳定性与效率；第二阶段实施细粒度中段训练，通过跨领域、多格式数据混合及上下文长度的逐步扩展，强化模型的泛化能力；第三阶段采用多架构多版本后训练（multi-scaffold with multi-version post-training），融合监督微调（SFT）、领域专用强化学习（domain-specialized RL）与多目标偏好对齐（MOPD），以整合互补的训练信号并增强模型的自主决策与代码生成能力。此外，通过模型合并技术有效融合不同专家模块及各训练阶段的检查点，缓解参数更新冲突。最终形成的IQuest-Q1模型在编码与通用任务上均展现出卓越性能，且已开源，为社区研究提供支持。
链接: https://arxiv.org/abs/2610.30949
作者: A. Jian Yang,An Huang,Bo Dan,Chen Mu,Chengran Yang,Chuan Shi,Feng Chang,Hongxin Li,Hongyuan Tao,Jinxing Zhang,Jujie He,Kun Wang,Lin Jing,Ming Yang,Qicheng Hu,Renyuan Li,Shark Liu,Shawn Guo,Shi Qing,Shi Zhong,Tianyu Zheng,Wei Liu,Wei Zhang,Wenrui Cai,Wenxiao Chen,Xuxing Chen,Yan Xing,Yi Xing,Yifan Luo,Yiran Xing,Yizhi Li,Yongsheng Kang,Yucheng Ding,Yue Song,Yun Qing,Yunqi Liu,Zelong Huang,Zhengmao Ye,Zhentao Fan,Zhiteng Li,Ziling Huang,Bihong Tang,Bingli Wang,Ce Jun,Changhao Chen,Chen Hu,Chen Xiao,Chi Liu,Chunjiang Ge,Danyang Li,Deyang Kong,Dian Qi,Ding Tan,Evan Yang,Fanglin Xu,Fei Bai,Gang Liu,Guochang Li,Hao Tang,Haoran Feng,Haowen Wang,Huahao Hu,Huanjin Yao,Huatong Song,Jia Deng,Jiajun Wu,Jialiang Cheng,Jiayi Liang,Jiayi Wu,Jie Deng,Jiejun Tan,Jing Yang,Jingda Xu,Junhang Cheng,Junxiao Liu,Kai Lv,Kaixiang Pan,Kecheng Lu,Kun Zeng,Kun Zheng,Kunbo Ding,Liang Qiu,Lingyi Sun,Luan Zhang,Mengxiao Lv,Ning Dong,Pei Wang,Peng Zhang,Qingsong Cai,Qixiang Yin,Qiyao Wang,Renbiao Liu,Renfei Dang,Ruihang Li,Ruirui Huang,Shanhui Yin,Shen Jian,Shengjie Fang,Shikai Li,Shinan Zhang
机构: IQuest Research
类目: Computation and Language (cs.CL)
备注: IQuest Report
Abstract:We present IQuest-Q1, a sparse Mixture-of-Experts (MoE) agentic foundation model with 320B total parameters (15B activated). IQuest-Q1 is developed through a three-stage process: (1) stable and efficient pre-training, combining Stable SplitNorm, Muon, and muP with a two-stage curriculum whose second stage increases the coding-data mixture, preserving a wide stability margin between safe and optimal learning rates; (2) fine-grained mid-training, expanding the data mixture across domains and formats and progressively extending the context; and (3) multi-scaffold with multi-version post-training, combining SFT, domain-specialized RL, and MOPD across various scaffolds to integrate complementary training signals and strengthen agentic coding capabilities. Furthermore, model merging combines different experts and checkpoints from different stages while mitigating conflicts among their parameter updates. The resulting IQuest-Q1 demonstrates capabilities in both coding and general tasks. We release IQuest-Q1 as an open-source model to support community research and development.
[NLP-52] RetiJev: Fast and Source-Grounded Ophthalmic Decision Models
【速读】： 该论文旨在解决眼科临床决策支持中生成式推理所面临的高延迟与高计算开销问题，尤其是在需要在预定义动作集中做出快速、准确且隐私保护性决策的场景下。传统自回归大语言模型（LLM）在生成答案时需逐词解码，导致推理速度慢、资源消耗大，难以满足临床实时性需求。为此，论文提出一种面向决策任务的替代方案——RetiJev系列模型，其核心创新在于摒弃自回归生成机制，转而采用直接评分候选答案的非生成式推理范式，从而实现高效决策。关键解决方案是构建了一个基于真实文献与指南的源端对齐语料库RetiJev-Data，涵盖78,158个临床决策与诊断问题、50,074个指南评分问题及24,662个评估基准问题，并在此基础上训练出0.8B至9B参数量的决策模型。实验表明，RetiJev系列模型在指南评分任务上显著优于27B LLM和商业模型Jev 1.13，提升达10.4–23.0分；其中RetiJev-9B在临床决策任务中达到最高准确率，较基线提升5.4–6.5分，同时诊断性能相当，并可在单张GPU上以158毫秒/案例的速度完成推理，相较传统生成式方法快3.5倍，较推理过程快470倍，实现了高性能与高效率的统一。
链接: https://arxiv.org/abs/2610.30947
作者: Bo Liu,Lianyu Wang,Meng Wang,Ching-Yu Cheng,Huazhu Fu
机构: National University of Singapore (新加坡国立大学); Singapore Eye Research Institute (新加坡眼科研究所); Singapore National Eye Centre (新加坡国家眼科中心); Duke-NUS Medical School (杜克-国大医学院); Institute of Advanced Intelligence and Computing, Agency for Science, Technology and Research (A*STAR) (先进智能与计算研究所，科技研究局(新加坡))
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Ophthalmic clinical decision support requires fast, accurate, and privacy-preserving inference, yet autoregressive large language models (LLMs) incur substantial latency and computational overhead for decisions among predefined actions. We therefore investigate a decision-oriented alternative to generative reasoning. Specifically, we first introduce RetiJev-Data, a source-grounded ophthalmology corpus comprising 78,158 clinical-decision and diagnosis questions derived from articles in PubMed Central and a MeSH vocabulary covering 531 eye diseases, and 50,074 guideline-grading questions spanning 11 grading systems, along with a 24,662-question evaluation benchmark. Based on this, we develop RetiJev, a family of 0.8B-9B decision models that directly score candidate answers without autoregressive decoding. Across all model sizes, RetiJev outperforms both a 27B LLM and the commercial Jev 1.13 decision model on guideline grading by 10.4-23.0 points. RetiJev-9B further achieves the highest accuracy on clinical decisions, improving over both baselines by 5.4-6.5 points while matching them on diagnosis. It answers a case in 158 ms on one GPU, 3.5x faster than LLM generation and 470x faster than reasoning. Models are released at this https URL.
[NLP-53] racing Hallucinations in LLM Reasoning for UML Class Diagram Generation: An Empirical Taxonomy and Mitigation Study
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在从自然语言需求生成统一建模语言（UML）类图时存在的幻觉（hallucination）问题，即生成结果可能出现与原始需求不符、推理过程偏离或代码实现错误等情况，从而影响图表的正确性和完整性。现有评估方法通常仅关注最终输出的UML图，忽视了具备推理能力的LLM在生成过程中留下的中间推理轨迹，而这些轨迹恰恰是识别幻觉源头的关键。为此，研究提出一个基于实证的十一类幻觉子类型的分类体系，涵盖三个维度：提示忠实性（Prompt Faithfulness，指推理与原始需求之间的偏差）、代码忠实性（Code Faithfulness，指推理与生成的PlantUML代码之间的不一致）以及推理质量（Reasoning Quality，指推理过程本身存在的缺陷）。该分类体系通过人工标注60条DeepSeek-R1的推理轨迹构建，并经由双模型标注流水线扩展至120条，进一步在20条DeepSeek-V4-Pro轨迹上验证。结果显示，DeepSeek-R1的幻觉主要集中在代码忠实性维度，而DeepSeek-V4-Pro则以提示忠实性为主，表明不同模型存在差异化的失败模式。此外，分析发现推理轨迹长度与幻觉频率无显著单调关系，但需求复杂度与幻觉频率呈负相关，尤其在提示忠实性错误中表现明显。在针对13个黄金标准UML模型集的缓解实验中，引入提示忠实性和推理质量约束后，基于Qwen3的三种模型的合并平均F1得分提升4.0个百分点，而事后代码忠实性审查未带来持续增益；关系建模始终是得分最低的环节。研究表明，通过定位生成过程中的幻觉来源，可实现针对性的提示工程优化，从而有效提升生成质量。然而，这些结论是否适用于其他模型家族仍有待进一步验证。
链接: https://arxiv.org/abs/2610.30946
作者: Mojdeh Rahmanian,Foutse Khomh,Ashkan Sami,Yanchao Yu
机构: Edinburgh Napier University (爱丁堡纳皮尔大学); Polytechnique Montréal (蒙特利尔理工学院)
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL)
备注:
Abstract:Large Language Models (LLMs) can generate UML class diagrams from natural language requirements, but hallucinations may compromise the correctness and completeness of the resulting diagrams. Existing evaluations mostly assess the final diagram, without examining the intermediate traces of reasoning-capable LLMs, which can reveal where these hallucinations arise. We develop an empirical taxonomy of eleven hallucination sub-types organised into three dimensions: Prompt Faithfulness (deviations between requirements and reasoning), Code Faithfulness (discrepancies between reasoning and generated PlantUML code), and Reasoning Quality (problems within the reasoning). The taxonomy was derived from 60 manually annotated DeepSeek-R1 traces, applied to 120 further DeepSeek-R1 traces through a dual-LLM annotation pipeline, and examined on 20 DeepSeek-V4-Pro traces. Code Faithfulness was the most frequent dimension for DeepSeek-R1 in both datasets, whereas Prompt Faithfulness dominated for DeepSeek-V4-Pro, suggesting model-dependent failure patterns. In the manually annotated traces, reasoning-trace length showed no significant monotonic association with hallucination frequency, whereas requirement complexity was negatively associated with it, particularly for Prompt Faithfulness errors. In a mitigation experiment on 13 Golden UML ModelSet cases with three Qwen3-based models, adding Prompt Faithfulness and Reasoning Quality constraints to the generation prompt raised the pooled mean F1 score by 4.0 percentage points, whereas a post-hoc Code Faithfulness review provided no consistent further benefit. Relationship modelling remained the lowest-scoring element throughout. These results suggest that localising hallucinations within the generation process can inform targeted, prompt-based mitigation. Whether these findings generalise beyond the DeepSeek and Qwen3 model families remains to be tested.
[NLP-54] Verbalizing Activation Differences with Natural Language Autoencoders
【速读】： 该论文旨在解决生成式 AI（Generative AI）中大规模语言模型内部表示难以解释的问题，具体聚焦于自然语言自编码器（Natural Language Autoencoders, NLAs）在激活语义化（activation verbalizers）任务中的泛化能力与可迁移性。其核心挑战在于：现有NLAs是否能在未见的模型变体、不同微调策略或跨模型对比场景下仍保持有效语义重构能力。解决方案的关键在于利用激活空间的差异进行语义化表达，并通过设计可复用的语义化机制实现对模型内部状态的可解释性分析。实验表明，尽管在某些极端情况下性能受限，但NLAs在组合性结构识别、轻量级微调模型的迁移应用、基于激活差异的微调目标恢复以及强化学习优化的模型差分任务中均表现出显著有效性，尤其在无需生成任何输出token的情况下即可定位微调目标。这些发现验证了NLAs在激活空间差异语义化方面的潜力，为未来可解释人工智能研究提供了重要方向。
链接: https://arxiv.org/abs/2610.30934
作者: Andrea Blasi Núñez,Gianluca Barmina,Federico Torrielli,Thiusius Rajeeth Savarimuthu,Peter Schneider-Kamp,Lukas Galke Poech
机构: 未知
类目: Computation and Language (cs.CL)
备注: Under Review
Abstract:Activation verbalizers translate the internal representations of large language models into human-readable text. The recently proposed natural language autoencoders (NLAs) learn this by reconstructing activations through a natural language bottleneck. However, it remains unclear how well these methods generalize beyond the setting they were trained for. Here, we study this question for NLAs along four closely related questions: (i) Do NLA verbalizations exhibit compositional structure with respect to their input? (ii) Do NLAs generalize to finetuned versions of the model they were trained for? (iii) Can NLAs recover finetuning objectives in model diffing? (iv) Can NLAs be specialized for model diffing? Across four models from three families, our experiments provide evidence supporting a largely positive answer to all four questions. Specifically, verbalizations of composed analogy activations mostly name the target while rarely mentioning the subtracted word, and an NLA can be reused on lightweight finetunes of its original model without retraining, with reconstruction quality degrading noticeably only for the largest model. Verbalizing activation differences between Taboo model organisms and their original models recovers the concealed secret word for most organisms, and reinforcement learning on activation differences further improves this recovery on held-out organisms. We also find that NLAs can recover the finetuning objective without the finetuned model generating any token, since the most informative token positions precede the user prompt. NLAs do not yet match state-of-the-art model diffing methods on the emergent misalignment detection task, but outperform a blackbox baseline. Our findings suggest that NLAs can meaningfully verbalize differences in activation space, rendering NLAs a promising direction for future work.
[NLP-55] ReflectPO: Reflection-Guided Policy Optimization for Long-Horizon LLM Agents
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在执行复杂长时序任务时，因早期决策错误导致后续交互历史与环境状态逐步恶化、错误累积难以纠正的问题。其核心挑战在于长时序任务中决策的序列依赖性与环境状态的动态变化，使得单一错误可能引发连锁失效。为应对这一问题，论文提出反射引导策略优化（Reflection-Guided Policy Optimization, ReflectPO），其关键创新在于两个核心组件：一是反射引导采样（Reflection-Guided Sampling），通过分析交互轨迹中的分支差异与重复访问模式，生成用于错误恢复的提示信号；二是树状信用分配（Tree-Structured Credit Assignment），将采样轨迹组织为树结构，基于父-子节点间的奖励差异与恢复收益实现动作级信用分配，从而精准识别并修正错误行为。实验表明，ReflectPO在多个长时序工具使用基准测试中显著优于现有基线方法，平均性能最优，且能有效提升智能体对自身偏差的识别能力、错误状态下的恢复能力及任务持续推进能力。
链接: https://arxiv.org/abs/2610.30933
作者: Rumei Li,Xunliang Cai
机构: Meituan(美团); China(中国)
类目: Computation and Language (cs.CL)
备注:
Abstract:With external tools, large language models (LLMs) can tackle complex tasks through sustained environmental interaction. LLM-based agents are therefore increasingly used to tackle complex long-horizon tasks in real-world environments. However, these tasks involve sequential, interdependent decisions and changes in environmental states. Early execution errors affect subsequent interaction histories and environmental states, and their effects accumulate along the decision chain. To address this challenge, we propose Reflection-Guided Policy Optimization (ReflectPO), a reinforcement learning framework for autonomous error recovery in long-horizon tasks. ReflectPO comprises two core components. The first is Reflection-Guided Sampling. We use branch differences and repeated visits in interaction trajectories to construct hints for error recovery. The second is Tree-Structured Credit Assignment. We organize sampled trajectories into trees and assign action-level credit using parent–child reward differences and recovery gains. Experiments over two long-horizon tool-use benchmarks show that ReflectPO outperforms multiple existing baselines in most domains. It also achieves the best overall performance on average. Further analyses show that ReflectPO effectively improves agents’ ability to recognize their own deviations, recover from erroneous states, and continue making progress toward task completion.
[NLP-56] Hierarchical Fine-Grained Cross-Modal Alignment via Dynamic Granularity Discovery and Optimal Transport
【速读】： 该论文旨在解决细粒度图像-文本检索中局部语义对齐不足与全局匹配一致性弱的问题，即在区分语义相近候选项时，现有方法难以有效建模对象、属性、数量、动作及关系等局部文本语义所对应的视觉证据。其解决方案的关键在于提出一种分层细粒度跨模态对齐框架（HFCMA），通过动态粒度发现（DGD）自适应地将低层图像块和文本词元聚合成语义紧凑的视觉区域与文本短语，并基于熵正则化最优传输（OT）构建满足全局边缘约束的结构化区域-短语对齐矩阵，从而实现更精准的局部对齐建模；同时引入背景去偏、文本重要性加权和原型一致性正则化策略，以抑制背景噪声与局部语义漂移，最终在双编码器框架下显著提升了细微语义差异的判别能力。
链接: https://arxiv.org/abs/2610.30912
作者: Pengfeng Song,Shaozhong Zhang,Pengfei Shao
机构: Zhejiang Wanli University(浙江万里大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 8 figures, 6 tables. Code available at this https URL
Abstract:Fine-grained image-text retrieval requires a model not only to assess global semantic consistency between an image and a text, but also to model visual evidence for local textual semantics, including objects, attributes, quantities, actions, and relations, so as to distinguish semantically similar candidates. Existing methods commonly use fixed image patches, detected regions, or text tokens as the basic matching units and then aggregate local interactions through max pooling, cross attention, or similarity accumulation. Such strategies remain limited in semantic unit construction and global matching consistency. This paper proposes a hierarchical fine-grained cross-modal alignment framework, termed HFCMA. The framework progressively transforms low-level patch and token representations into adaptive visual regions and textual phrases, and then performs structured region-phrase matching within a global retrieval scoring process. Specifically, HFCMA first adopts dynamic granularity discovery (DGD) to aggregate raw patches and tokens into semantically compact visual regions and textual phrases. It then formulates region-phrase matching as an entropy-regularized optimal transport problem, thereby estimating a structured alignment matrix under global marginal constraints. To reduce background noise and local semantic drift, HFCMA further introduces background debiasing, textual importance weighting, and prototype consistency regularization. Experiments on Flickr30K and MS-COCO show that HFCMA effectively combines dynamic semantic unit construction with OT-based structured matching under a dual-encoder retrieval framework. The resulting fine-grained score provides structured local alignment information for candidate ranking and improves the discrimination of subtle semantic differences.
[NLP-57] Silicon Samples Match the Answers but Not the Respondents
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在构建合成调查样本时，现有评估方法无法有效衡量生成受访者与真实人类受访者之间整体分布覆盖性的问题。传统评估仅关注单题回答准确性、各问题答案分布匹配度及跨问题相关性，但未能考察生成受访者在完整回答空间中的代表性，导致某些模型虽能精准复制边际分布，却生成了现实中不存在的答题组合。为此，作者提出“受访者覆盖差距”（Respondent Coverage Gap, RCG），将每位受访者的完整回答模式视为高维空间中的点，通过计算真实人类受访者到最近生成样本的距离与人类样本间平均距离的比值，量化生成数据对真实人群的覆盖能力。实验结果表明，在两项美国社会调查数据集上，所有六种大语言模型（LLM）均存在显著的RCG，其覆盖性能劣于仅独立采样各题答案的基线模型，并且在所有测试提示与推理设置下，均不如基于人类数据拟合的高斯互依结构（Gaussian copula）模型。该研究揭示了当前生成式AI在模拟复杂人类行为模式上的根本局限，强调需引入更全面的群体覆盖性评估指标。
链接: https://arxiv.org/abs/2610.30900
作者: Yusuke Kumagae,Masahiro Kaneko
机构: 未知
类目: Computation and Language (cs.CL); Computers and Society (cs.CY)
备注:
Abstract:Silicon sampling generates survey respondents by having a large language model (LLM) answer a survey as a person with given demographic attributes, and the resulting synthetic populations are increasingly used in place of human ones. Existing evaluations focus on individual-answer accuracy, per-question answer distributions, and correlations across questions. None of these compares the respondents themselves, so a generator whose respondents have combinations of answers found in no human respondent can still score well if it reproduces every answer distribution. We propose the respondent coverage gap (RCG), which represents each respondent as a point defined by their complete answer profile and measures how far each human respondent is from the nearest generated one, relative to the distance between human samples. On two US social surveys, all six LLMs we evaluate leave human respondents uncovered, and under our main settings, they fall behind a baseline that samples each answer independently from the human marginals. Across the prompting and inference settings we test, every LLM has a higher RCG than a Gaussian copula fitted to the human data.
[NLP-58] From Perpetrators Digital Footprints to Behavioral Communities: An Ontology-Driven Framework for Bengali Cyberbullying Analysis— FULL VERSION
【速读】： 该论文旨在解决现有网络欺凌（Cyberbullying）研究中内容检测、语义表征与网络分析之间割裂的问题，尤其针对缺乏可解释的施害者表征、多尺度结构建模以及行为驱动的相似性与影响传播机制解释等关键瓶颈。其核心解决方案是提出一种以施害者为中心的本体驱动网络欺凌行为与网络分析框架（Ontology-Driven Cyberbullying Behavior and Network Analysis, ODCBNA），通过需求驱动的本体设计、知识图谱构建、行为建模、社区分析与影响力建模的协同整合，实现对施害者行为模式、群体组织结构及潜在影响力路径的系统性刻画。基于孟加拉语上下文多样网络欺凌（BenCoDiC）数据集，该框架将1,063名具有至少三次欺凌行为的用户表征为64维的行为与上下文特征向量，利用余弦相似度与k近邻建模生成包含27,554条边的加权网络；通过Louvain算法识别出模块度约为0.495的六大社区及27个子社区，并通过特征贡献分析揭示“视频”“影响”和“用户画像”三类维度是行为相似性与社区分离的关键驱动力，同时上下文维度增强解释性。社区约束下的影响力分析表明，IM Greedy在各社区中均取得微弱正向增益，尽管不同种子选择方法产生的扩散范围相近，但种子重叠度有限，进一步验证了该框架在揭示潜在影响路径方面的有效性与可解释性。
链接: https://arxiv.org/abs/2610.30885
作者: Md. Nesarul Hoque,Md Hanif Seddiqui,Miskatul Anwara,Rudra Pratap Deb Nath
机构: University of Chittagong (吉大港大学); Gopalganj Science and Technology University (戈帕尔甘杰科技大学)
类目: Computation and Language (cs.CL)
备注: 32 pages, 8 tables, 13 figures
Abstract:Cyberbullying involves behavioral and contextual patterns beyond isolated abusive messages. However, existing approaches remain fragmented across content detection, semantic representation, and network analysis, with limited interpretable perpetrator representations, multi-scale structural modeling, and explanations of behavior-driven similarity and influence propagation. These limitations hinder the identification of persistent perpetrators, characterization of behavioral communities, and assessment of potentially influential users. This paper proposes Ontology-Driven Cyberbullying Behavior and Network Analysis (ODCBNA), a perpetrator-centric framework integrating requirements-driven ontology design, knowledge graph construction, behavioral modeling, community analysis, and influence modeling. Using the Bengali Contextual and Diverse Cyberbullying (BenCoDiC) corpus as a use case, ODCBNA identifies 1,063 users with at least three bullying comments or replies as perpetrators and represents each through a 64 -dimensional behavioral and contextual feature vector. Cosine similarity and k -nearest-neighbor modeling produce a weighted network with 27,554 edges. Louvain analysis identifies six communities with a modularity of approximately 0.495 and 27 sub-communities. Feature-contribution analysis identifies the Video, Impact, and User Profile groups as the strongest contributors to behavioral similarity and community separation, while contextual dimensions support interpretation. Community-constrained influence analysis shows that IM Greedy achieves small positive gains in all six communities, although different seed-selection methods produce comparable diffusion spreads despite limited seed overlap. These findings provide an interpretable basis for understanding perpetrator behavior, community organization, and potential influence patterns in Bengali online environments.
[NLP-59] Inherited Not Invented: Tracing the Provenance of Wait in Reasoning Models
【速读】： 该论文旨在解决生成式人工智能（Generative AI）中强化学习（Reinforcement Learning, RL）对模型推理行为影响的误解问题，特别是关于“wait”这一提示性标记在推理过程中的真实作用与来源。研究指出，尽管学术界普遍认为RL促使模型学会在决策前“等待”以重新评估步骤，但大量实验表明，高精度提升并不依赖于该标记的存在，这揭示了模型性能提升可能更多源于训练数据中隐含的推理模式，而非RL显式引入的策略。其解决方案的关键在于通过无标签的频率、位置、形态及使用方式等指标，并结合注入错误测试，系统追踪“wait”标记在多个开源模型训练过程中的演变轨迹。研究发现，“wait”几乎不被零强化学习（Zero-RL）阶段生成，而是在训练中期或微调推理轨迹时由数据驱动引入；不同RL策略对“wait”标记的影响各异——如基于KL惩罚的GRPO方法不会增加其出现频率，而部分释放的RL-Zero方案则显著放大该标记；此外，该标记的形态会随训练演化，表现出更敏感但更宽松的警示特性。最终，论文强调仅以标记频率作为判断是否发生推理的依据存在严重偏差，警告不应盲目采用大语言模型（LLM）作为评判标准的自动化评估体系，因其可能逆转人类标注的趋势，导致误判。
链接: https://arxiv.org/abs/2610.30882
作者: Zhizhuo Kou,Sizai Hou,Long Chen,Jian Yang,Junyu Luo,Jingshu Peng,Wenkai Zhang,Sirui Han,Yike Guo
机构: The Hong Kong University of Science and Technology (香港科技大学); Huawei Noah’s Ark Lab (华为诺亚方舟实验室); Peking University (北京大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Under review
Abstract:Reinforcement learning (RL) is credited with teaching reasoning models to write wait before reconsidering a step, yet large RL accuracy gains do not require the word. This view can confuse what RL discovers with what a model inherits from its training data. We trace the word through open models’ training histories, reading it with label-free indices of frequency, position, form and usage and an injected-error test, all computed from sampled outputs. We find that zero-RL almost never coins the marker: on bases that lack the word, RL raises MATH-500 accuracy substantially while wait stays near absent. It arrives with training data: mid-training in OLMo-3, and fine-tuning on reasoning traces installs it in most models whose base lacks it. Where the base already uses it, both released RL-Zero recipes (OLMo-3, MiMo) multiply it, whereas KL-free GRPO from the same base does not; OLMo-3’s recipe wears down its form and makes it a more sensitive but more liberal alarm; RL after supervised fine-tuning barely changes it, and our GRPO recipe erases a marker seeded by fine-tuning. Finally, we show that marker frequency alone does not indicate RL-induced reasoning and caution against LLM-as-a-judge revision metrics, as automated judges can reverse human-labelled trends.
[NLP-60] One Teacher Many Views: Reducing Reference Sensitivity in On-Policy Self-Distillation
【速读】： 该论文旨在解决在基于策略自蒸馏（On-policy Self-Distillation, OPSD）方法中，由于依赖单一完整参考解（full-reference solution）作为教师模型条件所引发的目标不稳定性问题。现有方法通常以完整的参考解为条件生成教师输出，但这种设定存在根本局限：参考解仅是推理过程的一个特定实现，其路径顺序、表述方式和逻辑结构可能包含学生在推理阶段无法获取的额外信息，导致教师输出过度依赖于不可见的上下文，从而产生偏差。作者发现，在保持问题与学生状态不变的情况下，更换为另一个有效的参考解会显著改变教师分布，这一现象被称为参考诱导的目标不稳定性（reference-induced target instability）。为解决该问题，论文提出蒙特卡洛特权蒸馏（Monte Carlo Privileged Distillation, MCPD），其核心思想是将特权上下文建模为部分参考视图的分布（distribution over partial reference views），并通过在多个视图下评估教师模型并取其下一词分布的概率空间平均（probability-space average），获得更鲁棒的监督信号。该平均形式被证明是通过最小化教师到学生之间的期望KL散度在特权视图上得到的精确目标。实验表明，MCPD在数学与科学推理基准上不仅提升了监督信号的参考一致性，且持续优于传统的全参考条件OPS D。
链接: https://arxiv.org/abs/2610.30880
作者: Khouloud Saadi,Di Wang
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:On-policy self-distillation (OPSD) improves reasoning by training a model on its own trajectories with dense token-level supervision from a privileged copy conditioned on a reference solution. Existing methods typically use a single teacher distribution conditioned on the complete reference. We identify a limitation of this formulation: a reference is only one realization of the reasoning process. Although it provides evidence about the answer, it also encodes an ordering, phrasing, and solution path. Full-reference conditioning can therefore produce correct targets that depend excessively on information unavailable to the student at inference. We show that replacing one valid reference with another, while holding the problem and student state fixed, can substantially alter the teacher distribution, a phenomenon we call reference-induced target instability. To address this instability, we introduce Monte Carlo Privileged Distillation (MCPD), which models privileged context as a distribution over partial reference views and distills the marginal teacher distribution. MCPD evaluates the teacher under multiple views and averages the resulting next-token distributions. We show that this probability-space average is the exact target induced by minimizing the expected teacher-to-student KL divergence over privileged views. Across mathematical and scientific reasoning benchmarks, we show that MCPD produces more reference-consistent supervision and consistently improves over full-reference OPSD.
[NLP-61] Synthetic-to-Real Code-Switched ASR: A Spontaneous Nepali-English Podcast Benchmark
【速读】： 该论文旨在解决尼泊尔语-英语代码混用对话语音在自动语音识别（ASR）中的两大核心挑战：脚本混淆（即天城文与拉丁字母的不恰当混合使用）以及语言切换点处显著升高的识别错误率。针对现有方法中高质量代码混用语音数据稀缺且标注成本高昂的问题，研究提出一种两阶段合成到真实（synthetic-to-real）的适配框架。其关键解决方案在于：第一阶段利用0.95小时通过多说话人、双引擎文本转语音（TTS）系统生成的合成语音进行监督微调（SFT），实现从合成数据到真实人类口语的跨域迁移；第二阶段引入组相对策略优化（GRPO），结合转录文本与脚本奖励机制，在少量真实测试样本上进行精细化调整。实验表明，合成SFT可有效降低混合错误率（MER）1.88个百分点、语言切换点错误率（SPER）7.12个百分点；后续GRPO虽仅带来微小但统计显著的改进，仍能进一步缓解英语脚本混淆问题并优化切换点性能。值得注意的是，消融实验揭示在SFT收敛后的编码器-解码器架构中，脚本奖励项与仅基于MER优化效果相当，归因于合成数据上候选序列的脚本方差趋近于零。此外，该方法在单语基准上也表现出色，显著提升尼泊尔语识别性能，同时保持英语词汇错误率稳定。所有经验证的转录文本、时间戳、节目元数据及重建脚本均已公开发布。
链接: https://arxiv.org/abs/2610.30860
作者: Firoj Paudel
机构: Wiseyak Inc.
类目: Computation and Language (cs.CL); Sound (cs.SD)
备注: 9 pages, 4 figures, 7 tables. Code: this https URL , Dataset: this https URL
Abstract:Nepali-English code-switched conversational speech presents severe challenges for automatic speech recognition (ASR), chiefly through script confusion (inappropriate mixing of Devanagari and Latin orthography) and elevated error rates at language switch points. Supervised fine-tuning (SFT) can adapt multilingual foundation models, but human code-switched audio is expensive to collect and annotate. We curate a new benchmark of 13,044 spontaneous podcast utterances (15.45 h total; 10,214 test, 2,830 dev), audited and verified to eliminate machine translation leakages and acoustic loops. We investigate a two-stage synthetic-to-real adaptation framework: Stage 1 applies SFT on 0.95 h of speech synthesized via a multi-speaker, dual-engine TTS pipeline; Stage 2 explores Group Relative Policy Optimization (GRPO) with transcription and script rewards. We find that synthetic SFT provides the primary cross-domain transfer to spontaneous human speech, reducing Mixed Error Rate (MER) from 23.40% to 21.52% (-1.88 pp) and switch-point error rate (SPER) from 35.89% to 28.77% (-7.12 pp). Subsequent GRPO yields small, statistically significant adjustments on 173 test utterances (1.69%), driving English script confusion (SCR_en) from 6.45% to 6.36% (p = 0.001) and SPER to 28.70% (p = 0.022 clustered). Crucially, ablation reveals that in this SFT-converged encoder-decoder setting, an auxiliary script reward term yields performance identical to MER-only optimization because rollout candidate groups exhibit near-zero script variance on synthetic data. On monolingual benchmarks, SFT improves Nepali MER by 2.71 pp (stabilizing at 18.42% after GRPO, a 2.38 pp gain over baseline) while preserving English lexical WER (3.38% vs. 3.33%). Verified transcripts, timestamps, episode metadata, and reconstruction scripts are publicly released.
[NLP-62] RIAD: Benchmarking Omni-Modal Ambiguity in Multimodal Large Language Models NEURIPS2026
【速读】： 该论文旨在解决多模态大语言模型在处理跨文本、视觉与音频模态间分布性语义歧义（omni-modal ambiguity）时的评估瓶颈问题。现有主流多模态基准普遍聚焦于确定性输入下的感知、识别或推理任务，无法有效检验模型在信息不完整时通过跨模态协同消除歧义的能力。为此，论文提出TRIAD基准，其核心设计为：每个测试样本由一个文本-图像-音频三元组构成，配以一个问题和一组答案选项，确保仅当三者联合时才能唯一确定正确答案；若任一模态缺失，则存在多个一致的答案选项，从而诱发歧义。TRIAD包含327个人工编写的问题，覆盖86个场景组及18类跨模态歧义类型。实验表明，尽管部分模型在单题层面表现优异，但在场景级消歧任务中仍显著落后于人类（差距超20分），且在移除答案选项后性能急剧下降至个位数，而人类仍保持稳健。进一步的留一模态分析与反事实测试揭示，当前系统在文本与图像存在时普遍低估音频线索。因此，该研究的关键贡献在于：将多模态评估范式从“识别”转向“消歧”，并提供了一套涵盖分类体系、基准数据集与诊断工具的完整框架，以系统化衡量模型在复杂真实场景中整合异构信息消除不确定性的真实能力。
链接: https://arxiv.org/abs/2610.30841
作者: Zhaolu Kang,Yidi Wang,Yile Li,Siqi Zeng,Yingjie He,Richeng Xuan,Zhichao hu
机构: Hunyuan, Tencent(腾讯混元); Peking University (北京大学)
类目: Computation and Language (cs.CL)
备注: NeurIPS 2026 Evaluations and Datasets Track
Abstract:Ambiguity is a central feature of natural communication: a sentence, image, or sound may remain underspecified until evidence from the other channels is considered. Yet current multimodal benchmarks mostly test perception, recognition, or reasoning over already determinate inputs, leaving unclear whether omni-modal large language models can resolve ambiguity distributed across text, vision, and audio. We introduce TRIAD, a benchmark for omni-modal ambiguity resolution. Each item is a text–image–audio triplet paired with a question and a set of answer options, and is constructed so that the full triplet determines a unique answer while removing any one modality makes the set of consistent options non-unique. TRIAD contains 327 hand-written queries in 86 scene groups and covers 18 ambiguity categories across the three modalities. Evaluating 14 omni-modal systems reveals that strong item-level performance does not translate into robust scene-level disambiguation: the best model still trails humans by over 20 points in group accuracy. Removing answer options widens the gap further, driving model group accuracy to the single digits while humans remain robust. Leave-one-out and transcript/description counterfactuals further show that current systems often underuse audio cues when text and image are present. TRIAD reframes omni-modal evaluation around disambiguation rather than recognition, and provides a taxonomy, benchmark, and diagnostic protocol for measuring this capability.
[NLP-63] Invisibilization as a Source of Probability Bias in Language Models
【速读】： 该论文旨在解决生成式 AI（Generative AI）在面对包含受保护属性（如种族、族裔和性别）的输入时，因对来源的偏见而过度拒绝或回避评估信息的问题。其核心挑战在于：当系统需判断是否采纳来自特定个体的信息时，若因该个体的身份特征而忽视其提供的证据，将导致对有价值信息的无谓损失。论文的关键解决方案是通过一个两瓮博弈（two-urn game）的精确贝叶斯框架，量化这种拒绝对信息价值造成的成本，并设计一种基于重复反馈与真实结果校准的信任评分机制。实验表明，主流大模型（来自 OpenAI 与 Anthropic）在描述报告者为黑人时拒绝评估的比例（46%）远高于白人（9%），但一旦评估，则对黑人报告者的信任略高；这种偏差在常规审计中难以察觉，且随推理复杂度增加、语言表达方式变化（如使用“Black”而非其西班牙语翻译“negro”或“Indigenous”）而加剧，甚至在提示用户采取防范性指令后进一步恶化。而引入基于揭示真相的持续评分机制后，模型在多数配置下于多轮迭代后将种族信任差距压缩至0.02以下，显著降低了因身份描述引发的拒绝率偏差。
链接: https://arxiv.org/abs/2610.30785
作者: Christian Arnold,Ernesto Calvo,Guadalupe Gonzalez,Sebastián Vallejo Vera,Tiago Ventura
机构: University of Birmingham (伯明翰大学); University of Maryland, College Park (马里兰大学学院公园分校); University of Western Ontario (西安大略大学); Georgetown University (乔治城大学)
类目: Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: 17 pages, 4 figures, 2 tables
Abstract:Language models are aligned to avoid using protected attributes such as race, ethnicity, and gender to inform their responses. Therefore, they increasingly refuse or hedge when such attributes appear in a prompt. However, if a system must decide how much to learn from what people tell it, it is costly to refuse valuable information because of its source. In this article, we measure the cost of declining information in a two-urn game with an exact Bayesian answer. We present leading LLM models from two developers (OpenAI and Anthropic) with a prompt that requires them to decide how far to trust an unverifiable report from a participant whose description we randomize across ten configurations (6,600 sessions). We show that leading models often decline to assess the reporter, treating information as worthless at rates set by the description rather than by the evidence. At its widest, models decline a Black-described reporter 46% of the time and a white-described reporter 9%; yet, when they do assess the information, the Black-described reporter receives slightly more trust. The bias is invisible to audits that compare average trust or look only at assessed cases. It grows with reasoning effort, follows words rather than categories (models invoke race for “Black” but not for its Spanish translation negro, and never for “Indigenous”), and worsens under warning instructions a practitioner would try first. We also present a scoring option that reduces the refusal rate bias. When a model’s trust is scored against revealed truth over repeated rounds, it credits identical records equally. In four of five configurations from Anthropic and OpenAI, the racial gap in trust falls below 0.02 by the last rounds.
[NLP-64] MixJudge: Learning a Language-Queryable Critic for Vocal Mixing ICASSP2027
【速读】： 该论文旨在解决音频混音过程中缺乏有效评估工具的问题，尤其针对现有音频-语言模型在判断具体混音问题方面的能力不足。由于专业混音师对混音质量的评价资源稀缺，研究构建了一个包含11类人声混音问题、覆盖多严重程度的课程化评估体系，并基于此训练出可接受自然语言查询的音频评判模型MixJudge。其解决方案的关键在于：通过结构化的人工标注数据集实现对混音问题的精准识别与分级，使模型能够准确检索特定混音缺陷、区分对立性问题（如“声音过响”与“声音过弱”），并合理排序问题严重性；更重要的是，其判断能力具备下游迁移性——在面对多个处理方案候选时，能优先选择在声学特性上更接近专业参考版本的选项，从而为混音决策提供可解释、可量化的支持。
链接: https://arxiv.org/abs/2610.30783
作者: Kuan-Hsun Ho,Yuan-Chiao Cheng,Yen-Tung Yeh,Yi-Hsuan Yang
机构: 未知
类目: ound (cs.SD); Computation and Language (cs.CL)
备注: Submitted to ICASSP 2027
Abstract:Mixing requires not only applying processing but also assessing the resulting mix. Audio-language models can provide conversa- tional mixing advice, but existing evaluations do not separately test whether they can judge specific mixing issues. Since professional supervision for such judgments is scarce, we construct a curriculum of 11 vocal-mix issues at multiple severity levels and use it to train MixJudge, a language-queryable audio critic. MixJudge shows its ability to retrieve mixing issues, distinguish opposite issues, and order the severity of an issue. Moreover, its judgments transfer to menu selection: given a set of processing candidates, MixJudge often favors versions acoustically closer to a held-out professional version, demonstrating how explicit mixing judgment can support comparison and downstream mixing decisions.
[NLP-65] Where meanings converge and diverge: Revisiting linguistic relativity across 81 languages
【速读】： 该论文旨在探究不同语言使用者是否以相似的方式组织概念，这一问题核心关联于萨丕尔-惠特尼假说（Sapir–Whorf hypothesis）及其关于语言相对性（linguistic relativity）的争论。其关键在于：尽管各语言可能共享基本概念的意义，但这些意义之间的关系未必一致。研究通过使用两种多语言词嵌入模型，量化了81种语言中196个基本概念之间的概念结构关系，发现社会文化类概念在跨语言间的关系对应性高于自然与运动类概念，且该差异在不同模型和分类体系中均显著重现。此外，在社会文化领域内，语言间的语义关系差异随地理距离增加而增大；语言亲缘关系越远，各领域内差异也越大。这表明，认知普遍性（universality）与语言相对性（relativity）并非对立，而是交织存在于不同概念域中。研究提出，共享的认知资源与语言及地方经验相互作用，进而预测在社会文化概念上人类语义判断的跨语言一致性将高于自然与运动类概念。
链接: https://arxiv.org/abs/2610.30778
作者: Weiran Wang,Wenjuan Qin,Haitao Liu
机构: Fudan University(复旦大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Do people who speak different languages organize concepts in similar ways? This question lies at the heart of the Sapir–Whorf hypothesis and debates about linguistic relativity. Shared basic meanings, however, need not imply shared relations among them. We compared relations among 196 basic concepts across 81 languages, using word embeddings from two multilingual models to quantify conceptual structure expressed in language. Cross-language correspondence was greater for sociocultural meanings than for meanings concerning nature and motion, a contrast recurring across models and classifications. Within the sociocultural domain, relational differences increased with geographic separation; less closely related languages differed more across all domains. These findings suggest that universality and relativity are interwoven across conceptual domains. They inform an account in which shared cognitive resources interact with linguistic and local experience, and predict greater cross-language correspondence in human semantic judgments for sociocultural than for nature and motion meanings.
[NLP-66] How Firm Are AI Models Ethical Boundaries? A Pay-Stub Editing Test of 15 Models
【速读】： 该论文旨在探究在相同敏感请求下，仅通过微调表述方式是否能够促使生成式 AI (Generative AI) 从拒绝执行转向实际操作。其核心问题在于揭示语言表达的细微变化如何影响模型对合规性与伦理边界的判断。解决方案的关键在于通过系统性实验对比三种不同表述语境下的模型行为：明确说明贷款用途、省略用途、以及将修改描述为修正打印错误。实验结果显示，当请求被表述为“修复打印错误”时，模型发起编辑操作的比例显著上升（66.2%），表明语义框架对模型决策具有决定性影响。研究进一步揭示了不同模型在边界处理上的差异性表现，如Astra完全拒绝含贷款用途的请求，而同一模型在用途缺失或误述为打印错误时则全部响应，反映出模型行为高度依赖于输入的语义暗示。该研究揭示了生成式 AI 在伦理边界判断中的可塑性与不稳定性，强调需警惕语言设计对模型行为的潜在操纵风险。
链接: https://arxiv.org/abs/2610.30751
作者: Simiao Ren,Xingyu Shen,Tommy Duong,Tsang Ng,Ankit Raj
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 11 pages: 4 pages of main content, 1 page of references, and 6 appendix pages; 7 figures and 4 tables; ancillary data and code included
Abstract:Can a small change in wording move an AI model from refusing a sensitive request to acting on it? We test 15 models on the same request to increase the income shown on a pay stub. Across 445 observable trials, models issue editing calls in 14.1% of cases when a loan purpose is stated, 33.1% when the purpose is omitted, and 66.2% when the change is described as fixing a printing error. Astra refuses all ten loan-purpose requests but issues calls in all ten purpose-omitted trials. Fable, Opus, and Muse Spark issue none across all three wordings; Sonnet issues some under the latter two. Response excerpts illustrate different ways of refusing, redirecting, or justifying an edit. The file is a marked test document and all tools are simulated. These results probe how boundaries behave in one setting; they do not prove an overall ethics ranking or that a harmful real-world edit succeeded.
[NLP-67] UniGuard: Multi-View Guardrail Learning for Cross-Interface Safety Generalization
【速读】： 该论文旨在解决当前生成式安全护栏（Guardrail）模型在面对不同输入接口形式时泛化能力不足的问题，即现有模型通常仅在固定提示（prompt）或响应（response）界面下训练，导致当同一安全场景以对话、代理轨迹、多智能体交互或策略条件输入等形式出现时，其安全性判断可靠性显著下降。为此，论文提出UniGuard训练框架，其核心在于通过多语义对齐的视角（包括提示、响应、多轮对话、代理轨迹、多智能体轨迹等）学习接口无关的安全表征（interface-invariant safety representations），从而实现跨接口场景的稳定安全检测。关键创新在于构建了UniGuard-FT数据集，将1,000个公开安全案例转化为多种接口形式，并采用分类损失、监督对比学习与非对称置信度校准损失联合训练，使模型在未见接口基准测试（ATBench、PolicyGuardBench、DynaBench）上平均提升45.12 F1分数，同时降低35.41点预期校准误差，展现出远超自身规模（0.6B参数）的性能表现，甚至优于大型模型如Claude-Sonnet-4.5和GPT-5.2，验证了其高效性与强泛化能力。
链接: https://arxiv.org/abs/2610.30742
作者: Viet-Thanh Pham,Thuy-Trang Vu,Junae Kim,Rollin Omari,Dinh Phung,Gholamreza Haffari,Lizhen Qu
机构: 未知
类目: Computation and Language (cs.CL)
备注: 31 pages
Abstract:Guardrail models are increasingly used as safety layers for LLMs and agentic systems, but most are trained on fixed prompt or response interfaces and become unreliable when the same safety case appears as a conversation, agent trajectory, multi-agent trace, or policy-conditioned input. We propose UniGuard, a training framework that learns interface-invariant safety representations from multiple semantically aligned views of the same safety case. We also construct the UniGuard-FT dataset by transforming 1,000 public safety cases into prompt, response, multi-turn conversation, agent-trace, and multi-agent-trace views, and train UniGuard-0.6B with classification, supervised contrastive learning, and asymmetric confidence calibration losses. Across unseen-interface benchmarks ATBench, PolicyGuardBench, and DynaBench, UniGuard-0.6B improves over its pretrained guardrail by 45.12 F1 points on average and reduces expected calibration error by 35.41 points, while remaining competitive with much larger guardrails such as Claude-Sonnet-4.5 and GPT-5.2 - despite being a 0.6B model and trained only with synthetic data.
[NLP-68] Meaning Contracts: Local Semantic Geometry in Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）中语义意义局部几何结构的表征问题，即现有研究虽关注语义在嵌入空间中的距离或线性探测器对语义类别的分离能力，却未能刻画单一语义因子在不同实现中的局部几何特性——包括哪些成分在语义表达中保持不变、哪些成分发生变化，以及语义细化过程中具体如何演变。其解决方案的关键在于设计并应用两种受控探测方法：跨语言探测（cross-lingual probe）通过语言扰动分离精确语义匹配、同一语义因子的不同实现、邻近概念与域外对比；细化探测（refinement probe）则区分语义特异性增强（semantic narrowing）与高概率属性扩展（attributive elaboration）及对抗控制。两项探测均通过盲测人类判断验证有效性。研究在四种架构各异的解码器仅模型（Gemma 4、Llama 3.2、Qwen 2.5、Pythia）上发现四条共现规律：在均值池化后去除主导方差方向后，句对按语义相关性排序（同因子 → 同领域 → 不同领域），且层间无重叠；语义因子围绕一个不变核心组织，周围存在结构化的实现特异性变异；语义细化导致定向位移进入更紧密的邻域，甚至对抗基底名词构建的控制组也表现出此趋势；句子级与词元局部读出揭示互补几何特征（句子级体积收缩，词元级读出呈现更锐利的各向异性终点）。研究结果表明，LLM的语义结构可被理解为局部各向异性邻域，在语义细化过程中表现为意义的收缩。
链接: https://arxiv.org/abs/2610.30715
作者: Andrés Daniel Godoy Ortiz
机构: Imperial College London (帝国理工学院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 24 pages, 8 figures. Code and data: this https URL
Abstract:This paper studies the local geometry of semantic meaning in large language models. Current representation analyses ask whether meanings are close in embedding space, or whether linear probes separate semantic classes, but they do not characterize the within-concept local geometry: what is invariant across realizations of a single semantic factor, what varies, and what changes when meaning is refined. This paper characterizes that geometry through two controlled probes evaluated on four architecturally diverse decoder-only LLMs (Gemma 4, Llama 3.2, Qwen 2.5, Pythia). A cross-lingual probe uses language as a perturbation to separate exact meaning matches, same-factor realizations, nearby concepts, and out-of-domain contrasts. A refinement probe separates specificity-bearing semantic narrowing from high-probability attributive elaboration and adversarial controls; both probes are validated by blind human judgments. Four findings recur across all models: under mean pooling, once dominant variance directions are removed, sentence pairs order by semantic relatedness (same factor, then same domain, then different domain) without overlap over a band of layers; semantic factors organize around an invariant core surrounded by structured realization-specific variation; refinement produces directed displacement into tighter neighborhoods, including against controls built on the same base noun; and sentence-level and token-local readouts expose complementary geometry (volume contraction at the sentence level, sharper anisotropic endpoints at token-local readout). We interpret these results as evidence that LLM semantics are organized as local anisotropic neighborhoods in which meaning contracts under refinement.
[NLP-69] ULME: Predictive Coupling for Mutually Reinforcing Embedding and Language Modeling AAAI2027
【速读】： 该论文旨在解决现代大语言模型中文本嵌入（text embedding）与文本生成（text generation）能力分离的问题，即二者通常由独立的专用模型分别提供，这些模型训练目标不兼容且在检索增强生成（retrieval-augmented generation）和智能体系统中并行部署，导致功能割裂。现有统一嵌入-生成模型如GritLM虽共享参数，但仅通过模式切换或并行优化两个任务，缺乏双向交互机制，本质上仍为“部分统一”。本文提出一种名为ULME的后训练框架，通过预测性耦合（predictive coupling）在因果解码器架构上实现真正的端到端统一：在查询与答案之间插入可学习的槽位，分别读取语义嵌入与推理嵌入，并在共享的因果主干上联合优化三个目标——锚定语义槽与文档表示的对比损失、监督答案生成的语言建模损失，以及采用梯度截断的预测损失，使推理槽与答案池化摘要对齐，从而实现生成过程主动塑造嵌入内容。实验表明，在Qwen2.5与Qwen3两个模型家族、0.5B至8B共八个规模下，单一ULME检查点相较最强开源基线在MTEB上提升3.3–4.1分，同时显著提升BRIGHT与AdvBench-IR表现，并在标准准确率基准（如MMLU、GSM8K、MMLU-Pro、HumanEval）上保持与纯生成微调相当的性能，验证了真正统一带来的相互增强而非简单共存的能力协同效应。
链接: https://arxiv.org/abs/2610.30705
作者: Yang Liu,Zhiqiang Yuan,Shu-Dong Huang,Jiancheng Lv,Liwei Yang,Qiwei Yan,Kailin Lv,Jinchao Zhang
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: AAAI 2027 Submmitted
Abstract:Text embedding and text generation are two of the most useful capabilities of modern large language models, yet they are typically provided by separate specialized models (an embedder and a generator) trained with incompatible objectives and deployed side by side in retrieval-augmented generation and agent systems. Recent unified embedding-generation models such as GritLM share parameters between the two capabilities, but they either switch attention regimes between modes or optimize embedding and language modeling as parallel objectives that never supervise each other. We argue that these designs are only partially unified: the two capabilities coexist on the same weights, but do not shape each other. We propose ULME, a post-training framework that unifies embedding and language modeling on a causal decoder-only LLM through predictive coupling. ULME inserts a chain of learnable slots between the query and the answer, reads a semantic embedding and a reasoning embedding from two designated slots, and trains three objectives on a shared causal backbone: a contrastive loss anchoring the semantic slot to document representations, a language-modeling loss supervising answer generation, and a stop-gradient predictive loss that aligns the reasoning slot with a pooled summary of the answer, so that generation actively shapes what the embedding encodes. Across two model families (Qwen2.5, Qwen3) and eight scales spanning 0.5B-8B parameters, a single ULME checkpoint improves MTEB by 3.3-4.1 points over the strongest open-source baseline, tightens BRIGHT and AdvBench-IR simultaneously, and remains on par with a generation-only fine-tune on standard accuracy benchmarks (MMLU, GSM8K, MMLU-Pro, HumanEval), demonstrating that genuine unification yields mutually reinforcing rather than merely coexisting capabilities.
[NLP-70] Disentangling Robustness and Sycophancy in Multilingual LLM s
【速读】： 该论文旨在解决多语言大语言模型（Multilingual Large Language Models, LLMs）在面对用户错误前提时，其错误响应是源于对用户意图的迎合（即“谄媚”行为，sycophancy），还是因语言能力不足而确实不知正确答案这一根本性问题。现有研究仅记录模型是否同意用户错误前提，缺乏对背后原因的区分。为此，论文提出SC-矩阵（SC-matrix），通过交叉分析模型在无压力条件下的能力（competence，即能否正确回答）与在压力条件下维持该回答的倾向性（sycophancy），实现对模型行为的精细化诊断。研究发现：（1）相同的总体谄媚率可能掩盖不同类型的失败组合；（2）即使在某语言中具备较高能力，模型仍可能表现出谄媚行为；（3）基于推理导向的缓解策略优于依赖指令驱动的策略。结果表明，多语言谄媚行为并非单一现象，而是由多种不同类型失败构成的复合体，需采用差异化诊断方法以准确识别并应对。
链接: https://arxiv.org/abs/2610.30696
作者: Rhitabrat Pokharel,Ameeta Agrawal
机构: Portland State University (波特兰州立大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:When a multilingual LLM agrees with a user’s false premise, is it being sycophantic, or does it simply not know better in that language? Current research does not distinguish these cases: they record only whether the model agreed, not why. We introduce the SC-matrix, that crosses a model’s competence (whether it answers correctly under no pressure) with its sycophancy (whether it maintains that answer under pressure). Our results show that (1) identical aggregate sycophancy rates can mask different failure compositions, (2) high competence in a language does not always protect against sycophancy, and (3) reasoning-oriented mitigations outperform directive-based strategies. We find that multilingual sycophancy is not a single phenomenon but a composite of different failures that require distinct diagnosis.
[NLP-71] From Clinical Judgment to Checkable Rules: Clinician-Led Guardrails for Mental-Health Chatbots
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在心理健康应用场景中因安全评估结果难以解释与可追溯而带来的可信度问题。现有方法如提示工程、基于LLM的评判（LLM-as-a-Judge）及宪法式AI虽能提升安全性，但其决策过程常为黑箱，缺乏透明性与可审计性。本文提出一种由临床医生主导的安全验证机制：由两名临床医生共同制定、终审并仲裁规则库中的四项安全检查条款，该规则库经过形式化验证以确保内部一致性，并作为防护层嵌入于LLM输出之上。该防护层的核心优势在于其所有判断均可基于特定规则与十六个事实进行复现，实现了决策过程的透明化与可追溯性。实验表明，在与同模型、同版本、同量化参数、同提示文本及解码策略的LLM裁判对比中，该防护层在“满足规则”这一最大分歧类别中表现显著：原模型在65条已含危机资源的回复中仍要求提供资源达55条（84.6%），而通过事实提取器验证，其中54条实际已存在资源；而防护层则因设计上不触发此类冗余请求，始终保持一致。同时，该系统在减少误报危机警报方面亦表现出更强性能。研究还揭示了引入该防护层所带来的可测量代价——例如在保留边界提示的独立测试集中，过度拒绝率从6.0%上升至16.0%，主要归因于事实提取逻辑的调整。尽管临床医生最终选择保留原有的不确定性默认处理机制，但整体分析显示，该系统的差异仅限于明确命名的规则与事实，且成本来源清晰可溯。因此，本工作的核心贡献并非生成更安全的结论，而是构建了一种可解释、可验证、可追踪的防护层（guardrail），其与裁判模型的分歧具有明确归属，且成本影响可定位至具体规则与不确定性默认设置。
链接: https://arxiv.org/abs/2610.30690
作者: Chenhan Lyu,Kit Wislocki,Sarah Elizabeth Stoeckl,Maitreyi Sinha,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); Computers and Society (cs.CY)
备注:
Abstract:Large language models (LLMs) deployed in mental-health contexts can be evaluated for safety failures, but the resulting judgments are often difficult to inspect or explain. Common approaches to improving safety rely on prompt engineering, LLM-as-a-Judge, and Constitutional AI. In this work, we describe a clinician-led alternative in which two clinicians define, finalize, and adjudicate the four safety checks of a rule base, which is formally verified for internal consistency and applied as a guardrail over LLM outputs. Here and in the title, clinician-led refers to this rule base, not to the fact extractor or the uncertainty default, which are ours. Every verdict of the resulting verification layer is recomputable from a named rule and sixteen facts. Against an LLM judge matched in model, revision, quantization, rule text, and decoding but without the layer’s computed 988/911 fact, the largest disagreement class, the satisfied-rule class, is structured: the Llama-3.3-70B judge, on its primary prompt, demands a crisis resource from 55 of the 65 replies (84.6%) that already contain one, while the same model, asked as a fact extractor, reports a recognized resource present on 54 of the 55. The verification layer never raises that demand, by construction. We report with equal prominence where this judge is stronger: fewer false crisis alarms on the one reference-scored check. We also report the measured cost (a specification revision raised block-class outcomes on held-out boundary prompts of an over-refusal benchmark from 6.0% to 16.0%, chiefly by changing fact extraction), the clinicians’ decision to keep our uncertainty default, and the analysis of the result. The contribution is not a safer verdict. It is a guardrail whose disagreements with the judge localize to named rules and facts and whose measured costs trace to named facts and the uncertainty default.
[NLP-72] Direct cueing of word identity through word duration: evidence from Mandarin homophones
【速读】： 该论文旨在解决语音词汇识别中词语持续时间（word duration）是否作为功能性语音细节被存储于心理词典中的问题，尤其关注其在同音词识别中的作用。研究发现，在普通话（Mandarin Chinese）中，高频词普遍比低频词更短，而通过同音词排序任务，实验表明参与者能够利用词语持续时间差异来正确识别同音词对的顺序——即使在高噪声环境下或缺乏其他线索时，仍能将较短的听觉刺激与高频词关联。这一结果支持了词语持续时间可直接作为词语身份的语音线索，提示其可能已被纳入词汇表征的语音细节中。该研究提出“直接线索化”解释，同时讨论了持续时间在词汇识别中可能存在的其他间接作用机制。
链接: https://arxiv.org/abs/2610.30682
作者: Xiaomeng Zhu,Megan Niu,Jason Shaw,Kevin Tang
机构: Yale University (耶鲁大学); University of Florida (佛罗里达大学); Heinrich Heine University Düsseldorf (海因里希·海涅杜塞尔多夫大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:While there is increasing evidence that words are stored in the mental lexicon with some phonetic detail, there are open questions about exactly which phonetic details are stored and whether they are functionally relevant for spoken word recognition. Word-level phonetic cues, such as word duration, provide an appropriate test case. Across languages, including Mandarin Chinese, high frequency words are shorter than low frequency words. Using a homophone ordering task, we investigate whether differences in word duration impact word recognition. Mandarin Chinese was examined since it has abundant disyllabic homophones. Our results indicate that participants were able to use word duration to recognize the order of homophone pairs, selected to vary in lexical frequency. Both with high noise levels and when no other cues were available, participants mapped the shorter duration auditory stimulus to the higher frequency word. The results provide evidence that word duration can directly cue the identity of words, suggesting that it may be a phonetic detail incorporated into the lexical representation. We discuss this direct cueing account alongside alternative accounts of how word duration may modulate word recognition.
[NLP-73] Scaling Laws for Safety Probes
【速读】： 该论文旨在解决生成式 AI（Generative AI）安全监测中安全探针（safety probe）性能不可预测的问题。当前，每当发布新模型或需监控新的有害内容领域时，均需重新训练探针，但探针的最终性能在训练前无法预知，导致必须通过大量试错来筛选最优配置，效率低下。本文的关键解决方案是提出并验证了两个普适的缩放定律（scaling laws），可基于模型规模和标注数据量预先预测探针性能。研究构建了六个大规模、高难度的新安全数据集，并在五类模型家族的20个基础模型上，使用32至10⁶条标注数据训练两种探针架构。结果表明：第一，探针误差随模型规模呈幂律下降；第二，探针误差随标注样本数量增加而趋近其下限，且该下降斜率随目标模型规模增大而提高。这两条规律在不同任务、模型家族和探针架构间均成立，仅拟合常数不同。当在某一模型族最小模型上拟合后，可将下一更大模型的探针误差预测精度控制在0.02以内，且在72%的情况下准确判断应投入更多标签还是更大模型以达到目标误差。此外，在探针架构选择上，其平均额外误差仅为0.006。这些缩放定律使安全研究人员能够根据有限的模型与标签预算，高效决策最优探针构建策略。
链接: https://arxiv.org/abs/2610.30678
作者: Yuchen Wu(University of Washington),Jerry Wei(Anthropic)
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
备注: 62 pages, 23 figures, 33 tables
Abstract:Safety probes trained on a model’s activations are a standard tool for monitoring language models in real time. Each time a new model is released or a new harm area needs to be monitored, the probe must be retrained, which requires new labels and a choice of probe architecture. Yet neither the model nor the labels tell us in advance how well the resulting probe will perform. Finding out requires training and comparing many probe configurations. In this paper, we ask whether probe performance can instead be predicted in advance from model size and the number of labels. Existing safety datasets are insufficient for this, as they are either limited in scale or already saturated. We therefore construct six new safety datasets of sufficient size and difficulty. On these datasets, we train two probe architectures on 20 base models from five model families, with 32 to 10^6 labels for each model. We measure performance using probe error, defined as 1 - AUROC, where lower values are better. We find that probe error follows two simple scaling laws. First, the lowest probe error a model can reach falls as a power law in its size. Second, probe error decreases towards its floor as the number of examples increases; the slope increases as the target model size increases. Both laws hold across tasks, model families and probe architectures, and only their fitted constants differ. When fitted on the smallest models of a family, the laws predict the error of the next larger model to within 0.02 in 72% of cases. They also identify whether more labels or a larger model is the cheaper way to reach a target error in 73% of cases. Finally, they choose the probe architecture with an extra error of only 0.006 on average when they pick the worse one. These laws allow safety researchers to determine how to create the best probe for any given model and label budget.
[NLP-74] WikiOriginQA: Automated Construction of a QA Dataset with Knowledge Origins for Analyzing Cultural Bias AACL
【速读】： 该论文旨在解决现有文化知识数据集依赖人工标注所导致的可扩展性瓶颈问题。传统方法需熟悉特定文化的标注人员，耗时耗力，难以覆盖多语言、多文化场景。其解决方案的关键在于提出一种自动化构建方法——通过客观地利用不同语言版本维基百科的创建时间来表征知识起源，从而生成名为WikiOriginQA的问答数据集。该方法不仅显著提升了数据构建的可扩展性，还揭示了大模型（LLM）在文化认知上对英语文化圈的系统性偏见，并强调知识向英语语境传播的重要性，为跨文化偏见评估提供了高效且可推广的技术路径。
链接: https://arxiv.org/abs/2610.30677
作者: Kento Haneda,Yosuke Kishinami,Ryo Fujii,Keisuke Sakaguchi,Makoto Morishita
机构: Future Corporation(未来公司); Tohoku University(东北大学)
类目: Computation and Language (cs.CL)
备注: Accepted to AACL-IJCNLP 2026 (9 pages, 2 figures, 10 tables)
Abstract:Existing datasets on cultural knowledge have been typically created by human annotators familiar with each culture. However, the process requires considerable human effort, presenting a significant challenge to scalability. To address this challenge, we introduce WikiOriginQA, a QA dataset automatically constructed by objectively representing the origins of knowledge through the creation dates of different Wikipedia language editions. Experimental results with the created dataset confirm that LLM performance tends to be biased toward Anglophone cultures. Furthermore, the results suggest the importance of knowledge propagation into the Anglosphere. These findings demonstrate the validity of our automated approach, enabling cultural bias evaluations with superior scalability across diverse languages without extensive manual annotation.
[NLP-75] Spectral Smoothing and Spiking Perturbations Reconfigure Capability Profiles and Selective Task Performance in Large Language Models
【速读】： 该论文旨在解决生成式人工智能（Generative AI）中变压器模型前馈投影层奇异值谱的因果作用不明确的问题，特别是其对下游任务表现的影响机制尚不清楚。研究通过后验的奇异值平滑与尖峰扰动干预，针对up_proj和down_proj矩阵进行系统性分析，其中正向扰动会放大偏离几何均值的奇异值（即增强大值、抑制小值，称为“尖峰”），负向扰动则相反，同时保持奇异向量和奇异值的几何均值不变。在五个开源Llama与Qwen指令模型上，研究人员在多个基准任务（包括能力标注子任务、标准及链式思维推理测试、判断评分的真实性、幻觉与安全评估，以及TRAIT人格探测）中评估扰动后的性能变化。为解释得分变化，引入由事实知识、语言理解与演绎推理构成的三维能力基底进行建模。结果表明，扰动引发的变化具有结构性而非随机性：大幅正向扰动更可靠地抑制语言理解和演绎推理能力；适度负向扰动在某些情况下可提升特定网格上的事实性或任务局部表现；且敏感性显著依赖于模型家族、基准模块及各层间奇异值分布的不均衡程度。该研究将奇异值扰动定位为一种轻量级干预手段，可用于探查大规模语言模型中能力组织方式及其权衡机制。
链接: https://arxiv.org/abs/2610.30676
作者: Qince Luan,Jianhui Xu,Yang Xu
机构: Southern University of Science and Technology (南方科技大学); Independent Scholar
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Singular-value spectra encode important structure in transformer feed-forward projections, yet their causal role in downstream behavior remains underexplored. We study post-hoc spectral smoothing and spiking interventions on the up_proj and down_proj matrices. Positive perturbation stretches deviations from the geometric mean, making large singular values larger and small ones smaller (spiking), and vice versa. The intervention preserves singular vectors and the geometric mean of the singular values. Across five open Llama and Qwen instruction models, we evaluate the perturbed systems on capability-labeled subtasks, standard and chain-of-thought reasoning benchmarks, judge-scored truthfulness, hallucination, and safety tasks, and TRAIT personality probes. To interpret score shifts, we fit responses in a three-dimensional capability basis consisting of factual knowledge, language understanding, and deductive reasoning. The induced changes suggest structured rather than purely random sensitivity: large positive perturbations more reliably suppress language and deductive responses, moderate negative perturbations can sometimes improve factual or task-local performance on the evaluated grids, and sensitivity varies strongly by model family, benchmark module, and underlying layerwise spectral inequality. These results position spectral perturbation as a lightweight intervention for probing how capabilities are organized and traded off in large language models.
[NLP-76] Group Relative Policy Optimisation Improves Multilingual Speech Recognition in Low-Resource Languages ICASSP2027
【速读】： 该论文旨在解决自回归编码器-解码器语音识别模型在使用教师强制（teacher forcing）训练时存在的暴露偏差（exposure bias）问题，该问题导致模型在推理阶段无法有效纠正自身预测错误。针对这一挑战，论文提出采用基于群体相对策略优化（Group Relative Policy Optimisation, GRPO）的强化学习方法，对非大语言模型（non-LLM）架构的语音识别模型进行优化。其解决方案的关键在于通过GRPO机制，在训练过程中引入更符合实际推理场景的反馈信号，从而增强模型对自身错误的鲁棒性与自我修正能力。该方法在数据稀缺或输入音频存在噪声/退化的情况下表现尤为突出，以Whisper-large-v3为基准模型，在51种非洲语言上验证了有效性，所提出的SunflowerASR模型在AfriVox-v2（宏平均词错误率WER：0.227）和SimbaBench（WER：0.372）两个基准上达到了当前最优性能，并已公开发布。
链接: https://arxiv.org/abs/2610.30663
作者: Tim Wenjie Hu,Evelyn Nafula Ouma,Benjamin Akera,John A. Quinn
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 5 pages, 3 figures, submitted to ICASSP 2027
Abstract:Autoregressive encoder-decoder speech recognition models trained with teacher forcing can suffer from exposure bias, limiting their ability to recover from their own prediction errors at inference time. Recent work has shown that reinforcement learning with Group Relative Policy Optimisation (GRPO) improves speech recognition in major languages with LLM-based models. We extend this line of work to smaller, non-LLM encoder-decoder ASR models, using Whisper-large-v3 as a case study across 51 African languages. The benefits of this approach are most pronounced when training data is scarce or when input audio is noisy or degraded. Our resulting model, SunflowerASR, achieves the best published macro-averaged WER on the AfriVox-v2 (0.227) and SimbaBench (0.372) benchmarks, and is publicly available.
[NLP-77] What the Specification Did Not Say: Specification-Oracle Gaps in Test-Graded Coding-Agent Benchmarks
【速读】： 该论文旨在解决代码生成基准测试中存在的“规范-验证器鸿沟”（specification-oracle gap）问题，即任务描述以自然语言给出，但评分依赖于上游项目的真实测试用例，而这些测试用例可能隐含了未在描述中说明的关键细节（如常量名称、模块路径或具体错误信息），导致模型必须复现上游实现的细节才能通过测试。这种鸿沟使得基准测试不仅评估生成能力，还间接衡量模型对特定实现细节的复制能力，从而引入偏差。其解决方案的关键在于构建一个针对五种编程语言的静态审计工具，能够自动识别并分类每个评分测试在名称、路径和字面量层面的依赖关系，实现高达89.2%的保留精度和92.9%的召回率。实证分析表明，在SWE-Milestone基准中，25个里程碑存在依赖未声明且无法推断的测试项，带有此类依赖的构建单元失败率高达80.6%，远高于无依赖项的48.4%；移除一个隐藏名称即可导致126/128个测试无法编译。尽管排行榜整体排名对移除受影响测试仍具鲁棒性，但分数分布发生显著变化，且倾向于奖励能复现上游命名的模型。此外，研究发现该问题的严重程度与任务规范的生成方式密切相关：基于发布说明或大模型摘要的任务中，超三分之一受此影响，而经过筛选或为测试而设计的任务几乎不受影响。最终，作者公开了审计工具、修正后的过滤列表及规范补充材料，以推动更公平、透明的基准评估。
链接: https://arxiv.org/abs/2610.30635
作者: Fei Ding
机构: Alibaba Group(阿里巴巴集团)
类目: Computation and Language (cs.CL)
备注: 13 pages
Abstract:Coding-agent benchmarks often describe a task in natural language but grade it with tests taken from the upstream project. Passing those tests can require information the description never states: a constant’s name, a module path, an exact message. We call this a specification-oracle gap: where it exists, a benchmark also measures whether an agent reproduces upstream implementation details. We formalise the gap, including the inconsistencies that chains of dependent tasks add, and build a static audit for five languages that classifies every name-, path- and literal-level dependency of every scored test (held-out precision 89.2%, recall 92.9%). On SWE-Milestone (98 graded milestones, 35 published runs), 25 milestones contain a test that depends on an unstated, non-inferable name or path. Build units with such a dependency fail in 80.6% of runs, against 48.4% for comparable units without one, and withholding a single hidden name from the reference stops its unit’s tests from compiling in 126 of 128 cases. The leaderboard’s ranking is robust to removing the affected tests, but they shift scores within repositories and favour agents that reproduce upstream names. Across four further benchmarks, prevalence appears to follow construction: a third or more of the tasks are affected where specifications are release notes or LLM summaries, almost none where tasks are screened or written for their tests; how much a gap costs depends on how the harness grades. We release the audit, corrected filter lists and specification addenda.
[NLP-78] -SAX: A Two-Track Time Series Symbolization for Natural Language Shape Querying
【速读】： 该论文旨在解决时间序列数据库中自然语言查询（Natural Language Querying, NLQ）的核心挑战——形状查询（Shape Queries），即用户以自然语言描述时间序列的动态变化模式（如“先上升后下降”）。现有方法虽借助符号化（Symbolization）作为查询与长历史数据之间的桥梁，使形状查询成为可能，但现有符号化方法主要针对压缩、相似性索引或分类任务设计，未系统考虑自然语言查询的需求。本文通过分析六种经典符号化方法在自然语言查询任务中的表现，识别出三大关键局限：（1）符号串中缺失方向信息；（2）方向测量粒度过细，难以匹配人类对趋势的宏观感知；（3）符号字母选择缺乏语义关联性，不符合自然语言中常用形容词的表达习惯。基于此，论文提出三个设计原则，并据此对经典符号化方法SAX进行最小化重构，提出TT-SAX（Two-Track SAX）：保留原有级别字母的同时，为每个数据段新增一个趋势字母，该趋势在相邻段之间跨段度量，并采用与自然语言查询中常用形容词（如“上升”“平稳”“下降”）相匹配的分级体系。在公开基准NLQTSBench上的大量实验表明，TT-SAX在六种主流大语言模型（LLM）下均显著优于六种经典符号化方法，尤其在形状查询任务上表现突出；消融实验进一步验证了其性能提升分别源于对上述三个局限的针对性改进。本研究首次系统地将自然语言查询作为符号化设计的目标，为未来该领域研究提供了理论基础与方法范式。
链接: https://arxiv.org/abs/2610.30593
作者: Zhao Tan,Xianglong Bao,Zhipeng Xi,Ming Jin
机构: Griffith University, Australia
类目: Databases (cs.DB); Computation and Language (cs.CL)
备注: 14 pages, 8 figures, 2 tables. Accepted at the Australasian Database Conference (ADC 2026). This is the submitted manuscript prior to peer review
Abstract:Natural language querying for time series databases aims to let users find what they need in long histories of data by asking in plain language. Shape queries ask how a series moves. They are the main challenge of this task. Recent work makes them feasible by using symbolization as a bridge between the query and the long history. However, existing symbolizations were designed for compression, similarity indexing, or classification. What a symbolization should provide for natural language querying has not been systematically studied. We therefore examine the definitions of six classical symbolizations against this task and identify three limitations. (1) Direction is missing from the string, (2) direction is measured at too small a scale, and (3) the letters are arbitrary. Three design principles follow. We follow these principles and make a minimal redesign of SAX called TT-SAX (two-track SAX). The level letters stay. Each segment gains a trend letter. It is measured across neighboring segments rather than inside one, and graded to match the adjectives that queries use. We conduct extensive experiments on the public NLQTSBench benchmark, comparing TT-SAX with the six classical symbolizations across six LLMs. TT-SAX outperforms all six on shape queries, and ablations trace its gains back to the three limitations one by one. This paper is the first systematic study of time series symbolization with natural language querying as the design goal. It provides a foundation for future research on this task. Our code is available at this https URL.
[NLP-79] IFC-MCQ: Automatically Generating Multiple-Choice Benchmarks for Evaluating LLM Agents on IFC Data Retrieval
【速读】： 该论文旨在解决建筑、工程与施工（AEC）领域中基于生成式 AI (Generative AI) 的工业基础类（IFC）数据自然语言查询系统在大规模评估时面临的挑战，即自由文本回答难以实现高效、低成本的准确评价。其解决方案的关键在于提出 IFC-MCQ，一种从 IFC 文件自动生成多选题基准测试的方法，支持精确的选项集评分。该方法构建了涵盖15个类别和36个子类型的分类体系，答案项基于文件内容生成，干扰项则来源于文件数据或预定义规则。通过在不同规模和建模特征的 IFC 模型上应用 IFC-MCQ，研究评估了生成覆盖率，并对两种代理配置（工具调用与代码检索）进行了对比分析，揭示了特定类别下的失败模式，并表明性能受源 IFC 模型及检索接口的影响。IFC-MCQ 为 AEC 工作流中基于 LLM 的 IFC 检索系统提供了可扩展、可重复的基准评估能力。
链接: https://arxiv.org/abs/2610.30533
作者: Rabindra Lamsal,Sisi Zlatanova,Xuesong Shen
机构: The University of New South Wales(新南威尔士大学); GRID Lab(地理信息与数据实验室); School of Built Environment(建筑环境学院); School of Civil and Environmental Engineering(土木与环境工程学院)
类目: Computation and Language (cs.CL)
备注:
Abstract:Industry Foundation Classes (IFC), the standard for building information exchange in architecture, engineering, and construction (AEC), is increasingly queried through large language models (LLMs) using natural language. Evaluating these systems at scale is difficult because free-text answers require costly human or LLM-based review. We propose IFC-MCQ, a method that automatically generates multiple-choice benchmarks from IFC files for exact option-set grading. The taxonomy covers 15 categories and 36 subtypes across four groups, with answer keys derived from the file and distractors constructed from file-derived values or predefined rules. We applied IFC-MCQ to IFC models of varying sizes and modelling characteristics to assess generation coverage and evaluated two agentic configurations on selected models using tool- and code-based retrieval. The evaluation exposed category-specific failure modes and showed that performance can vary with the source IFC model and retrieval interface. IFC-MCQ enables scalable benchmarking of LLM-based IFC retrieval in AEC workflows.
[NLP-80] Protegrity-PII Technical Report
【速读】： 该论文旨在解决结构化与非结构化文本中敏感信息识别（PII, Personally Identifiable Information）的精准检测问题，尤其针对现有模型在面对多样化、复杂化实体类型及分布外样本时表现不足的挑战。其解决方案的关键在于提出一种新型的PII决策模型Protegrity-PII，该模型基于一个涵盖广泛实体类型的分类体系进行训练，从而显著提升了对分布外（out-of-distribution）数据的泛化能力；同时，在新提出的基准数据集PII-SCOPE上实现了0.915的类型召回率与0.907的精确率，并具备高效的推理性能（每百万字符仅需0.49秒，成本低至0.0005美元/百万字符），体现出高精度与高效率的协同优势。
链接: https://arxiv.org/abs/2610.30518
作者: Hardeep Bassi,Hui Ma,Luís Seabra,Koowum Joshi,Daniel Pereira,Averell Gatton
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:We present Protegrity-PII, a new PII decision model for structured and unstructured text. Protegrity-PII achieves a typed recall of 0.915 with a precision of 0.907 on a new benchmark we propose, PII-SCOPE. Our model processes text in 0.49 seconds per million characters at \ 0.0005 per million characters on a single RTX 6000 using batched inference. Due in part to the expansive nature of the entity type taxonomy used in training, Protegrity-PII achieves superior out-of-distribution generalization when benchmarked against a selection of state of the art models and datasets.
[NLP-81] Behind Aggregate Scores: What Systematically Varying Arena-Hards LLM Judge Reveals NEURIPS2026
【速读】： 该论文旨在解决大语言模型（LLM）作为评判者在评估两个回答优劣时的可靠性问题，特别是其判断结果的可重复性与对指令敏感性的不确定性。研究发现，尽管聚合得分（aggregate score）看似稳定，但其背后个体判别结果可能在重复实验或指令微调下发生显著变化，这表明高稳定性聚合分数并不能反映个体判决的一致性。解决方案的关键在于：评估体系必须区分聚合层面的稳定性与个体判别层面的可重复性，并强调在依赖个体判决结果（如偏好标签或单个响应评估）时，应通过重复运行相同设置并报告聚合得分与提示级别一致性（prompt-level agreement）来确保评估的可信度。
链接: https://arxiv.org/abs/2610.30512
作者: Louise Yorke
机构: University College London(伦敦大学学院); London, UK
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: Accepted at the JUDGe workshop (Can We Trust the Judge?), NeurIPS 2026 (non-archival). 18 pages, 3 figures. Code and data: this https URL
Abstract:LLM judges are used to determine which of two answers is better. Trusting these verdicts requires knowing how repeatable they are and how much they depend on the judge’s instructions. We show that similar aggregate scores can conceal substantial changes in individual verdicts, both when the same setup is repeated and when the instructions change. Using 500 Arena-Hard v2.0 prompts with two models’ answers held fixed, we run GPT-4.1 as the judge under all eight combinations of criteria, output format and task framing, and repeat every setup unchanged at temperature zero on a pinned model version. Rerunning the default setup moves the aggregate score by about one point but changes the win/tie/loss outcome on 21.2% of prompts. Instruction changes introduce further sensitivity: in one setup, which asks for the verdict without reasoning and also drops several other instructions, outcomes change on 28.2 percentage points more prompts than the two setups’ unchanged reruns do on average, and the aggregate score falls by 4.09 points. In another, where the criteria prioritise accuracy, outcomes change on 6.7 percentage points more prompts by the same comparison, while the aggregate score barely moves. A stable aggregate score therefore does not establish that individual verdicts are repeatable. This distinction matters when those verdicts are used as preference labels or to assess individual responses. Evaluations that rely on individual verdicts should repeat each setup they use and report both aggregate scores and prompt-level agreement.
[NLP-82] Rethinking the Evaluation of LLM Fact-Checkers
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在回答事实性问题时存在错误且需逐条验证的问题，其核心挑战在于现有评估基准未能明确三个关键设计要素：待验证陈述所针对的具体问题、判断陈述真实性的条件，以及可作为证据的来源。为克服这一局限，本文提出了JSFactCheckBench，一个针对日语的事实核查基准，明确规范了上述三要素。该基准包含1,913条源自日本社交媒体中误导性信息的LLM生成陈述，每条陈述均基于参考时间进行真实性标注，并评估其对原问题的重要性，同时使用预选网络来源作为证据。实验结果表明，所有六种LLM在直接回答问题的陈述上验证准确率低于辅助性陈述，这一差异被聚合评分掩盖；此外，当模型需从更长或多源文本中定位相关信息，或在包含上下文问答对的情况下进行判断时，性能显著下降。这些缺陷之所以能被揭示，正是由于本基准显式记录了每个标签背后的评估条件，而那些将条件隐含的基准则无法暴露此类系统性弱点。因此，解决方案的关键在于构建一个透明、可复现且全面定义评估条件的事实核查基准。
链接: https://arxiv.org/abs/2610.30504
作者: Ayana Niwa,Miyu Yamada,Kouta Nakayama,Masahiro Kaneko,Hisami Suzuki,Satoshi Sekine
机构: MBZUAI(中东人工智能大学); Science Tokyo(东京科学大学); NII-LLMC(日本国立信息研究所语言与机器学习中心)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) are widely used to answer factual questions, but their answers can contain errors and must be fact-checked claim by claim. Yet the resulting score depends on three design decisions: which questions the claims answer, under what conditions each claim is judged, and which sources count as evidence. No existing benchmark specifies all three. We introduce JSFactCheckBench, a Japanese benchmark that specifies all three, with 1,913 claims from LLM answers to questions about misleading information that circulated on Japanese social media, each annotated for factuality at a reference time and for importance to the question, using evidence from preselected Web sources. Across six LLMs, all verify claims directly answering the question less accurately than supplementary claims, a gap that aggregate scores obscure. Scores also drop when models must locate the relevant content in longer or multiple sources and, for all four open models, when the surrounding question and response are added. These weaknesses are visible only because the benchmark records the conditions behind each label. A benchmark that leaves them implicit cannot reveal them.
[NLP-83] Why Bother Me Again? Benchmarking Personal Agent Memory for Interaction Preferences
【速读】： 该论文旨在解决个人智能代理（Personal Intelligence Agent）在长期交互中因重复提问、冗余确认和不当主动介入而带来的交互负担问题。核心挑战在于：现有系统缺乏对用户交互偏好（interaction preferences）的持续记忆与动态适应能力，导致代理无法根据用户的个性化交互风格进行恰当决策。其解决方案的关键在于提出一个名为MemPABench的交互式评估环境，用于在策略内（on-policy）对代理的记忆能力进行评测。该环境通过模拟四位具身化用户（grounded simulated users），构建涵盖工作与生活场景的14项交互偏好，并在长达四个月的99个会话中引入每用户五次事件驱动的偏好变更，以测试代理在偏好选择、行为执行及上下文敏感性与偏好变化适应性方面的表现。实验结果表明，当前主流记忆机制（如基于检索的Mem0、MemOS、简单RAG以及文件型记忆）在提升偏好选择准确率方面收效甚微，其中检索类方法甚至不及“无记忆”基线；而文件型记忆虽能提高准确性，却常将偏好错误泛化至无关上下文。进一步分析揭示，现有系统不仅难以有效检索交互偏好，且即便获取正确偏好，也难以克服代理固有的默认交互惯性。因此，该研究指出当前记忆系统在动态偏好建模与情境自适应方面仍存在显著瓶颈。
链接: https://arxiv.org/abs/2610.30485
作者: Jialin Li,Zhenhao Chen,Hanjun Luo,Hanan Salam
机构: New York University Abu Dhabi (纽约大学阿布扎比分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 97 pages. Project page: this https URL
Abstract:Personal intelligence systems are taking on increasingly complex everyday tasks, yet repeated questions, unnecessary confirmations, and unwanted outreach can turn capable assistance into interaction burden. Reducing this burden requires agents to remember how users want to be assisted. However, existing memory benchmarks focus on factual recall and personalized answers and rarely test how agents learn and act on interaction preferences. We introduce MemPABench, an interactive environment for on-policy evaluation of memory in personal agents. MemPABench poses long-horizon tasks requiring agents to decide when to clarify, confirm, or act autonomously. Four grounded simulated users hold 14 preferences across work and personal contexts. Their feedback spans 99 sessions over four months, with five event-driven changes per user testing memory updates. We evaluate preference selection, delivered behavior, and sensitivity to context and preference change. Evaluating four memory systems across four LLM backbones, we find that current memory brings limited gains in preference selection. Retrieval-based systems (i.e. Mem0, MemOS, simple RAG) do not outperform No Memory, while File-based memory improves accuracy but applies preferences beyond relevant contexts. Our analysis shows that current memory struggles to retrieve interaction preferences and adapt to preference changes. Even with the correct preference available, overriding the agent’s default interaction habits remains challenging. The environment and dataset are available at this https URL.
[NLP-84] Rethinking SAE Evaluation: An Atomic Interpretable Unit Framework Reveals Hidden Polysemanticity and Redundancy NEURIPS2026
【速读】： 该论文旨在解决当前稀疏自编码器（Sparse Autoencoders, SAEs）评估体系中缺乏统一、明确的全局目标问题。尽管已有研究广泛采用诸如稀疏性、死特征率、重构保真度和解耦性等指标来评估SAE特征集的全局性质，但这些指标多为经验性且缺乏理论根基，未能形成一个可形式化的终极目标。为此，本文提出一种新的SAE评估范式：要求SAE特征（以及更广泛的基于字典的模型表示分解）构成一个“原子可解释单元基底”（atomic interpretable unit basis），即作为语义可解释性解释的坐标系。这一范式将此前零散的评估指标（如DAME，Dominance-Aware Monosemanticity Evaluation）重新组织为代理指标，并构建了一个全面的基准测试框架。通过对多个架构与模型家族中的数百个SAE进行评估，研究发现部分SAE在现有指标下看似具备单义性（monosemantic），但其在模型计算中占主导地位的特征却系统性地表现出较低的可解释性与高冗余性，揭示出潜在的失效模式。这些发现凸显了对主导性特征进行感知评估的重要性，呼吁在开展归因、电路追踪或干预前，将“主导性感知评估”作为标准流程，以提升SAE应用的可靠性与有效性。
链接: https://arxiv.org/abs/2610.30474
作者: Riteng Zhang,Xiaoqian Wang
机构: Purdue University (普渡大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted to NeurIPS 2026 (Evaluations and Datasets Track). 92 pages. Code: this https URL
Abstract:Sparse autoencoders (SAEs) were originally motivated by discovering monosemantic features, a property that characterizes individual features but does not specify what the learned feature set should look like as a whole. As SAEs have increasingly been adopted for attribution, circuit discovery, and intervention, researchers have begun evaluating metrics that describe global properties of the feature set, such as sparsity, dead feature rates, reconstruction fidelity, and disentanglement, a diverse set that extends far beyond monosemanticity. However, this implicit shift beyond monosemanticity has not been formalized as a clear, well-defined objective that comes before these evaluation metrics. Without such an objective, existing metrics remain ad hoc, with their necessity unjustified. In this work, we formalize this ultimate objective by introducing a new paradigm for SAE evaluation: requiring that SAE features, and more generally dictionary-based decompositions of model representations, form an atomic interpretable unit basis, a coordinate system for semantically meaningful explanation. Under this paradigm, we organize existing and new measurements such as DAME (Dominance-Aware Monosemanticity Evaluation) as proxy metrics and introduce a comprehensive benchmarking framework. Evaluating hundreds of SAEs across multiple architectures and model families, we uncover failure modes in which some SAEs appear monosemantic under existing evaluation metrics while remaining functionally polysemantic: the most dominant features, those contributing most to model computation, are systematically less interpretable and more redundant. These failure modes warrant attention from SAE researchers as targets for improved training and architectures, and motivate dominance-aware evaluation as a standard check for practitioners before SAE features are deployed for attribution, circuit tracing, or intervention.
[NLP-85] Self-Assessment Reinforcement Learning for Search-State Awareness in Retrieval-Augmented Large Language Models
【速读】： 该论文旨在解决现有搜索代理（search agents）普遍缺乏搜索状态感知（search-state awareness）的问题，即模型无法准确识别已获取信息与缺失信息之间的差异，从而导致幻觉推理（hallucinated reasoning）和搜索过程过早终止。其核心解决方案是提出自评估强化学习（Self-Assessment Reinforcement Learning, SARL），一种无需测试阶段额外评估即可让模型在训练中自主学习自身搜索状态的强化学习框架。SARL的关键在于双奖励设计：一方面奖励每次自我评估带来的修正效果，另一方面惩罚对评估机制的过度依赖，从而协同促进搜索轨迹的有效性与效率。实验表明，SARL相较最强基线实现24.3%的相对性能提升，在多跳推理和深度研究任务上分别取得35.9%和27.9%的显著增益；同时，SARL在无辅助监督或多人协作架构支持下仍优于现有方法，并在长搜索周期下保持训练稳定性，展现出随任务复杂度增长而持续提升的能力。
链接: https://arxiv.org/abs/2610.30435
作者: Taewhoo Lee,Minju Song,Jungwoo Park,Chanwoong Yoon,Junha Jung,Hyeon Hwang,Sanghoon Lee,Minbyul Jeong,Jaewoo Kang
机构: Amazon(亚马逊); Korea University (高丽大学); AIGEN Sciences; University of Massachusetts Amherst (马萨诸塞大学阿默斯特分校); Meridian Intelligence Global Inc
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent advances in search agents have enabled large language models to tackle complex tasks by retrieving and synthesizing external information. However, existing search agents often lack search-state awareness, i.e., the ability to recognize what information has been obtained and what remains missing, leading to hallucinated reasoning and prematurely terminated search. To address this limitation, we propose Self-Assessment Reinforcement Learning (SARL), a reinforcement learning framework in which models learn to recognize their own search state during training and improve their search behavior accordingly, without requiring additional assessment at test time. SARL achieves this through a dual reward design that captures the corrective effect of each assessment while penalizing excessive reliance on it, jointly encouraging effective and efficient search trajectories. Under this setup, SARL achieves a 24.3% relative improvement over the strongest baseline, with average gains of 35.9% on multi-hop reasoning and 27.9% on deep research benchmarks against the strongest per-benchmark baselines. SARL further surpasses methods built on auxiliary supervision or multi-agent coordination, despite operating as a standalone RL framework with no inference-time scaffolding. Beyond final performance, SARL maintains training stability under extended search horizons, a regime where existing methods consistently degrade, enabling sustained gains as task complexity grows.
[NLP-86] Agent Mosaic: Recognizing Confidentiality Leaks in Agent ic LLM Systems ACL
【速读】： 该论文旨在解决大语言模型（LLM）代理在处理混合敏感度数据时，因信息碎片化聚合（mosaic effect）而导致的敏感信息泄露问题。其核心风险在于：即使用户查询仅涉及单一文档，模型仍可能通过整合多个可访问文档中的非敏感片段，生成包含机密结论的综合回答，从而造成隐蔽性数据泄露。解决方案的关键在于引入可见的分类标题和明确的聚合规则，以限制模型对跨文档信息的不当整合。然而研究发现，即便采取此类防护措施，模拟环境中仍有32.37%的回答存在不安全披露（95%置信区间：28.33%-36.19%），且模型拒绝回答的行为本身也可能暴露被禁止的信息，表明单纯增加拒绝率无法保障保密性。因此，该研究揭示了基于“拼图式”推理的泄露已成为实际存在的保密性失效模式，亟需在高敏感数据环境中实施更严格的推理与披露控制机制。
链接: https://arxiv.org/abs/2610.30418
作者: Daniel Ruiz,Luca Luceri,Emilio Ferrara
机构: Information Sciences Institute(信息科学研究所); University of Southern California(南加州大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 16 pages, 9 figures, submitted to ACL Rolling Review, October 2026 cycle
Abstract:Modern LLM agents can retrieve information from accessible documents, and then disclose a sensitive conclusion assembled from them. This mosaic effect creates a critical risk for agents operating over confidential data. We quantify its realization with 30 synthetic cybersecurity, medical, and military examples, yielding 10,800 final answers from three LLM agents in controlled simulation and 1,440 answers from live agents in two open-source harnesses. Even under benign queries designed to be answerable from one document, GPT-5.6-Sol, the best performing model tested, discloses forbidden composites in 7.08% of answers (95% CI: 3.75-10.67%) in simulation, with worse disclosure rates in live harnesses. Adding visible classification headers and explicit aggregation rules reduces pooled disclosure rates substantially, yet unsafe disclosures persist in 32.37% of simulated answers (95% CI: 28.33-36.19%) across all intents despite these safeguards. Furthermore, increased refusals do not guarantee confidentiality, since significant refusals reveal the prohibited conclusion regardless. These results point to a troubling conclusion: sensitive data leaks via mosaic-style events are an observed confidentiality failure mode in the evaluated settings at non-negligible rates. They motivate stronger controls on what agents may infer and disclose in sensitive data environments with mixed-clearance users.
[NLP-87] SIM-UT: A User Simulation based Evaluation Framework for Medical Triage Chatbots
【速读】： 该论文旨在解决当前基于大语言模型（LLM）的医疗分诊聊天机器人在低收入和中等收入国家（LMICs）应用中存在的评估方法与实际使用场景脱节的问题。现有评估多依赖于单轮次、仅限英语的基准测试，仅衡量模型的知识水平，而忽视了真实用户在复杂多轮对话中表现出的非规范行为，如用词模糊、拼写错误、使用地方方言或情绪化表达，以及对话上下文的快速演变。为此，论文提出一种名为SIM-UT（Simulation of User Behaviour for Triage）的多轮次、基于仿真的评估方法，通过模拟真实用户群体的人格特征、行为约束与沟通模式，识别出具有生态效度且情境相关的失败模式。关键发现是：无论是医学微调的开源模型（MedGemma）还是通用前沿模型（ChatGPT与Claude），均存在“降级风险”这一严重缺陷——即在初始正确识别医疗紧急情况后，当用户以简化方式描述症状但未提供新临床信息时，模型会过度迎合用户意见，出现临床上危险的顺从性表现，例如将妊娠期阴道出血误判为“无问题”，或将孕32周胎动减少视为“正常”。因此，研究建议将基于真实用户行为模拟的动态评估纳入生产级分诊聊天机器人的标准开发流程，以便在部署前或后续迭代中建立有效安全机制。
链接: https://arxiv.org/abs/2610.30412
作者: Deepali Mittal,Dheeraj Mundhra,Yulia Sidorova,Tersia de Kock,Oli Munns,Sofie S Meyer
机构: hearX Foundation(hearX基金会); Cliniva(Cliniva); Respira Health(Respira健康); University of Pretoria(普利托利亚大学); Turn.io; UserTrace(用户追踪)
类目: Computation and Language (cs.CL)
备注: 23 pages, 4 figures, 3 tables
Abstract:LLM-based chat agents are increasingly used for complex, multi-turn, multilingual medical triage in low- and middle-income countries (LMICs), yet evaluation methods focus on single-turn, English-only benchmarks that primarily test the chatbots’ knowledge. In practice, users are often confused, overwhelmed, or emotionally vulnerable, giving unclear, misspelled single-word answers or long messages in regional dialects, and the context of a conversation can shift rapidly. We propose SIM-UT (Simulation of User Behaviour for Triage), a multi-turn, simulation-based evaluation method that identifies ecologically valid, contextually relevant failure modes of LLM-based triage chat agents. Using SIM-UT, we evaluated three real-world WhatsApp-based health chatbots: a respiratory screening programme in India, a hearing health support programme in South Africa, and a diabetes and lifestyle coaching agent in Kenya. We compared a medically fine-tuned open model (MedGemma) against general-purpose frontier models (ChatGPT and Claude) across 5,000+ multilingual simulated conversation turns grounded in the personas, constraints, and communication patterns of the real user populations, over 2-4 iterative rounds. A key failure mode emerged in both model families: de-escalation risk. The agents initially identified medical emergencies correctly, but when the user minimised the symptoms without introducing new clinical information, they de-escalated, exhibiting clinically dangerous sycophancy, for example agreeing that vaginal bleeding in pregnancy was “not a problem” or that decreased fetal movements at 32 weeks “can be normal”. We recommend user simulation-based behavioural evaluation, grounded in real user context and covering adversarial and edge cases, as standard practice for triage chatbots in production, so that safeguards can be added before deployment or in subsequent iterations.
[NLP-88] xt-Only LLM Correctors Break Correct Transcripts: A Repair-Damage Analysis of Audio Grounding in ASR Error Correction ICASSP2027
【速读】： 该论文旨在解决大语言模型（LLM）在自动语音识别（ASR）结果纠错过程中，仅依赖整体词错误率（WER）评估指标会掩盖实际修复与引入错误的细节问题。其核心挑战在于：如何准确量化音频信息对纠错效果的贡献，并区分模型在不同输入条件下（有/无音频）的修复、损坏及插入行为。解决方案的关键是提出一种精确分解方法，将每次WER变化拆解为被修复的词元（repaired tokens）、被破坏的词元（damaged tokens）和新增词元（insertions），从而实现对纠错过程的细粒度分析。研究通过对比在三种语料上使用Gemini 2.5 Flash进行文本仅纠错与结合音频纠错的表现发现：当仅使用模型自身生成的草稿时，纯文本纠错会破坏13%–32%原本正确的转录文本；而加入音频后，破坏率降至1%–3%；尽管修复效果在草稿上下降而在独立识别器输出上上升，但医疗语音记录中保留的插入项抵消了净WER改进。相比之下，开放权重的Voxtral模型因极少修改文本，未呈现相同模式。由于两组提示（prompt）存在差异，实验结果虽不能完全隔离音频的影响，但有效界定了音频在纠错中的作用边界。
链接: https://arxiv.org/abs/2610.30385
作者: Vasiliy Kudryavtsev,Kirill Borodin,Nikita Vasiliev,Mikhail Gorodnichev,Grach Mkrtchian
机构: 未知
类目: ound (cs.SD); Computation and Language (cs.CL)
备注: Submitted to IEEE ICASSP 2027
Abstract:LLM correctors of ASR output are judged by aggregate WER, which hides what they fix and what they break. We decompose every WER change exactly into repaired tokens, damaged tokens and insertions, and use this to ask what the source audio adds to a corrector. We compare correction with and without the audio, for the model’s own draft and for the output of an independent recogniser, on three corpora. With Gemini 2.5 Flash, text-only correction of its own draft breaks 13-32% of already-correct transcripts; with the audio, 1-3%. Damage falls in both settings. Repair falls against the own draft and rises against the recogniser. On medical dictation, kept insertions cancel the gain in net WER. The open-weight Voxtral, which rarely edits, does not show this pattern. Prompts differ between the two arms, so the contrasts bound the effect of the audio and do not isolate it.
[NLP-89] MemPlan: Memory-Conditioned PDDL Planning for Partially Observable Text Environments NEURIPS2026
【速读】： 该论文旨在解决大语言模型（LLM）在长时程规划任务中因生成幻觉或无效动作而导致的高成本与不可靠性问题。其核心挑战在于如何在保持语言交互灵活性的同时，实现对符号规划（Symbolic Planning）的有效支持，尤其是在部分可观测的交互式文本环境中。解决方案的关键在于提出一种基于记忆条件的PDDL规划框架——MemPlan：通过固定PDDL领域，动态重构每一步的PDDL问题，结合观测解析、候选假设与交互记忆进行状态更新；其中，正向图记忆依据相关性为缺失对象、位置或条件的候选假设分配启发式动作代价，引导规划器优先探索更合理的检查路径；负向记忆则记录被执行反馈否定的假设，并将其编译为排除约束，避免重复尝试已被证伪的选项；同时，采用经过监督微调的统一模式-条件化语言-符号接口，实现从自然语言环境到符号规划循环的端到端映射，完成观察到事实的解析与符号动作到命令的语义接地。实验表明，MemPlan在TextWorld、ALFWorld和Robotouille上的任务完成率与步骤效率均优于纯语言基线及PDDL感知基线，且在线令牌消耗显著低于如ReAct等依赖LLM自主规划的方法，同时同一微调接口可跨不同任务类型复用，无需任务特定微调。
链接: https://arxiv.org/abs/2610.30379
作者: Tengwei Song,Robert Hoehndorf
机构: King Abdullah University of Science and Technology (KAUST)(阿卜杜拉国王科技大学); Åbo Akademi University (Åbo阿克迪米大学); ELLIS Institute Finland (芬兰ELLIS研究所)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: NeurIPS 2026
Abstract:LLM agents are flexible in language interaction, but using them as long-horizon planners is costly and prone to hallucinated or invalid actions. This paper targets a different role for LLMs: constrained bridges between language and symbolic planning. We propose MemPlan, a memory-conditioned PDDL planning framework for partially observable interactive text environments. MemPlan keeps the PDDL domain fixed and reconstructs the PDDL problem at each step from parsed observations, candidate hypotheses, and interaction memory. Positive graph memory assigns relevance-based action costs to candidate hypotheses for missing objects, locations, or conditions, guiding the planner toward more plausible checks. Negative memory records execution-refuted assumptions and compiles them into exclusion constraints, preventing repeated attempts at candidates already contradicted by feedback. A unified schema-conditioned language–symbol interface, trained offline with supervised fine-tuning, connects text environments to the planning loop through observation-to-fact parsing and symbolic-action-to-command grounding. Experiments on TextWorld, ALFWorld, and Robotouille show that MemPlan improves task completion and step efficiency over language-only and PDDL-aware baselines, while substantially reducing online token cost compared with LLM-planning baselines such as ReAct. The same fine-tuned interface is also reused across TextWorld task types without task-specific fine-tuning. Our code and data are available at: this https URL.
[NLP-90] Climbing the Ladder: Automating Data Pipelines and Reproducing Baselines in Low-Resource Settings
【速读】： 该论文旨在解决小语种（如西西里语）在大规模多语言机器翻译系统中因依赖噪声大、非标准化的网络文本而导致翻译质量不佳的问题。其核心解决方案在于：通过从双语期刊《Arba Sicula》的PDF文档中提取高质量对齐的平行语料，利用无语言依赖的句子编码器结合人工标注对齐数据进行阈值调优，并对自动对齐结果进行典型错误过滤，从而构建更精确的训练数据；同时，将原始系统的分词与屈折词素偏倚的子词分段机制迁移至无需Perl的实现方式，并在固定测试集上评估所有模型性能。研究对比了从零训练的小型Transformer模型与经过低秩微调适配的预训练NLLB模型在双向翻译中的表现，尤其以意大利语作为近邻桥梁语言，并引入原系统逆向训练策略——即先使用网络西西里语进行正向评分的反向翻译训练，再用精心筛选的文学文本进行后续训练。实验表明，预训练模型显著优于从头训练模型，且分阶段训练策略对西西里语到其他语言的翻译提升更为明显。尽管在包含英语的翻译方向上，采用更大适配器和更多反向翻译样本的分阶段模型优于原逆向训练方案，但在西西里语与意大利语之间则略逊一筹；然而，多语言适配器在未参与训练的方向上会降低性能，限制其作为通用多语言模型的应用潜力。
链接: https://arxiv.org/abs/2610.30374
作者: Donato Meoli,Eryk Wdowiak
机构: Università di Pisa(比萨大学); Arba Sicula; Brooklyn, NY, USA
类目: Computation and Language (cs.CL)
备注: 17 pages, 6 tables
Abstract:The massively multilingual translation systems that now include Sicilian learned it from noisy, non-standardised web text. To obtain better translation quality, we redevelop the small neural machine translation (NMT) system that Wdowiak customized for Sicilian. We rebuild on maintained software the extraction of the parallel text from the PDFs of the bilingual journal Arba Sicula, whose sentences are aligned by a language-agnostic sentence encoder with thresholds tuned against an existing hand alignment and then filtered for the typical errors of automatic alignment. Furthermore, the tokenization and the desinence-biased subword segmentation of the original system are ported to a Perl-free implementation, and all the results are measured on a frozen test set drawn from the journal. We compare small Transformers trained from scratch with the pretrained NLLB model adapted by low-rank fine-tuning, in both directions and with Italian as a closely related bridge language, and we adapt to the pretrained model the reverse-training recipe of the original system, i.e., a stage on forward-scored back-translations of the web Sicilian followed by a stage on the curated literary text. The pretrained model is by far the better of the two, and the staged training improves the translation into Sicilian more than the translation from it. On the test set of the original system, and in the tokenized space in which it was scored, the staged model with a larger adapter and more back-translations is above the reverse-training recipe in the directions that involve English and below it between Sicilian and Italian. Yet, a multilingual adapter degrades the directions on which it was not trained, which limits its use as a general multilingual model.
[NLP-91] Probing and Steering for Safety-Relevant Behaviours in Large Language Models
【速读】： 该论文旨在解决生成式AI（Generative AI）中安全相关行为的检测与可控引导问题，具体聚焦于欺骗、高风险意识和阿谀奉承三种行为的识别与调控。其核心挑战在于不同设计要素（如估计器选择、词元聚合策略、探测/引导层的选择）在统一框架下的相互作用机制及其对行为检测与引导效果的影响。解决方案的关键在于构建一个系统化评估管道，综合对比差异均值（difference-of-means）与逻辑回归（logistic regression）两种估计器，结合最后时刻（last）、平均（mean）及可学习注意力（learned attention）三种词元池化方式，在Llama-3.1-8B-Instruct模型上进行多维度实验。研究发现，尽管先前工作提出“读写分离”假设——即逻辑回归在检测上表现更优而差异均值在引导上更优——但该规律仅适用于欺骗与高风险意识行为，而在阿谀奉承行为中出现反转，表明该现象部分源于数据特性而非估计器本身。此外，最优池化方法与层选择高度依赖于具体行为与数据分布，且最佳引导层可能显著早于最佳检测层。更重要的是，用于衡量模型感知变化与实际行为诱导的指标在任务间存在分歧，提示需根据目标选择合适的评价标准。最终，研究输出可交互的探针方向，支持用户直接评分与实时引导文本生成。
链接: https://arxiv.org/abs/2610.30363
作者: Nimun Bajwa,Massimo Poesio
机构: Queen Mary University of London(伦敦玛丽女王大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: 20 pages, 26 figures, 11 tables. Code, data and interactive interface: this https URL
Abstract:Activation probes are computed directions in a language model’s residual stream, used both to detect safety-relevant behaviours and, by adding their direction back into the stream, to steer them. We explore how the main design choices behind such probes interact, within one pipeline, on one model (Llama-3.1-8B-Instruct), and across three behaviours: deception, high-stakes awareness and sycophancy. The pipeline crosses two estimators (difference-of-means and logistic regression) with three token-pooling schemes (last, mean and learned attention), selects appropriate layers to detect or steer a behaviour, and evaluates every direction twice: as a detector of the behaviour, and as a steering vector to push outputs towards the behaviour. Prior work read together suggests a read/write dissociation, in which logistic regression reads (detects) best but difference-of-means directions write (steer) best. We find that it holds for deception and high-stakes but not for sycophancy, where the pattern reverses. The reversal appears to follow partially from the data rather than from the estimators. The best pooling technique and the layer choice depend on the behaviour and data, and the layer that steers best can sit well before the layer that detects best. The right steering metric to use between changing a model’s perception and eliciting behaviour depends on the task - and the two can disagree outright. The exported probes and directions power an interactive interface through which a user can score and steer text directly.
[NLP-92] CiteAnchor: Source-Aware Grounding for Hallucinated Citation Detection and Resolution
【速读】： 该论文旨在解决科学文献中虚假引用（hallucinated citations）难以溯源与验证的问题，其核心挑战在于引文信息不完整或在不同文献类型及版本间存在差异，仅凭表面匹配无法确认引文真实性或还原其原始来源。解决方案的关键是提出CiteAnchor框架，该框架通过引入源感知（source-aware）的定位机制，针对不同文献类型动态调整检索策略与字段级验证逻辑，并独立判断引文的有效性与唯一来源可恢复性。研究进一步构建了包含423个真实世界虚假引用的基准数据集CiteResolve，用于评估源可恢复性。实验表明，CiteAnchor在现有检测基准上分别取得0.841和0.784的F1分数，在CiteResolve上实现74.7%和70.7%的端到端任务成功率，显著优于基线方法，证明了其在提升虚假引用检测与源追溯能力方面的有效性。
链接: https://arxiv.org/abs/2610.30355
作者: Chenyi Tong,Ping Nie,Yu Zhang
机构: University of Wisconsin–Madison (威斯康星大学麦迪逊分校); University of Waterloo (Waterloo大学); Texas AM University (德克萨斯农工大学)
类目: Computation and Language (cs.CL); Digital Libraries (cs.DL)
备注: 18 pages, 4 figures, 9 tables
Abstract:Hallucinated citations make it harder to trace sources and verify the evidence behind scientific claims. Detecting them is difficult because bibliographic records can be incomplete or differ across source types and publication versions. A plausible match alone neither validates the citation nor establishes its underlying source. We introduce CiteResolve, a benchmark of 423 real-world hallucinated citations annotated for source recoverability. We further propose CiteAnchor, a source-aware grounding framework that adapts retrieval and field-level verification to each source type and separately determines citation validity and whether a unique source can be recovered. On an existing detection benchmark, CiteAnchor achieves 0.841 F1 with GPT-5.6 Luna and 0.784 with an open-weight backbone, outperforming the baseline with the same backbones by 7.9 and 36.8 points. On CiteResolve, it achieves 74.7% and 70.7% end-to-end task success, outperforming the corresponding web search baselines by 16.3 and 65.3 percentage points. Our code and data are available at this https URL.
[NLP-93] Consistency Is Not Validity: Self-Rewarded Moral Argumentation Makes Language Models Decisive but Pliable
【速读】： 该论文旨在解决大语言模型（LLM）在道德判断任务中表现不稳定、易受选项呈现顺序影响的问题。其核心挑战在于，现有模型在面对道德困境时，往往因推理过程缺乏一致性与可解释性而导致判断结果不可靠。为提升道德判断性能，论文提出一种基于强化学习的解决方案：通过自评对弈（self-judged tournament）机制，在掩蔽思维（masked thinking）框架下训练策略模型，使其根据生成论证的质量获得奖励。关键创新在于将“论证质量”作为奖励信号，引导模型发展出更严谨、一致的语言表达风格（language register），从而增强判断的稳定性。实验表明，该方法显著提升了模型在外部基准（如MoralChoice、ELEPHANT、MORABLES）上的决策一致性，并减少对选项顺序的敏感性。然而，研究也揭示了“一致性不等于有效性”的根本问题——尽管模型在重排测试中表现出更强的一致性，但其仍存在谄媚倾向（sycophancy）和语义重构（reframing）等行为，反映出其判断并未真正内化正确的道德逻辑。两个消融实验进一步验证，仅通过合成困境或开放思维训练亦可实现部分重排收益，但缺乏显著的语言风格转变且更易产生谄媚行为，说明单纯优化一致性无法确保道德有效性的提升。因此，该研究的关键结论是：必须区分“形式一致性”与“实质有效性”，仅依赖自我评价和推理一致性不足以保证道德判断的可信度。
链接: https://arxiv.org/abs/2610.30349
作者: Marius De Kuthy Meurers,Aaron Parisi,Bernd Bohnet
机构: Google DeepMind(谷歌深度思维)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 20 pages, 10 figures, 9 tables
Abstract:LLMs are increasingly consulted for moral judgment. We aim to improve performance on such tasks by rewarding good arguments, on the premise that arguing well increases the likelihood of judging well. We train a reinforcement learning policy initialized from a state-of-the-art language model in a self-judged tournament with masked thinking: the generated arguments are judged pairwise on argument quality by the policy itself. As an empirical basis, we extract authentic dilemmas from classic novels, using a logic solver to verify that each constitutes a conflict. We show that our judge rates argument quality like human annotators, and the policy then adopts the language register that it rewards. Testing on the external benchmarks MoralChoice, ELEPHANT, and MORABLES, we find that the policy becomes more decisive and invariant to the order in which options are presented, but sycophancy and reframing remain a challenge to validity. Two ablation studies with synthetic dilemmas and unmasked thinking match the reordering gains, but show a weaker register shift and higher sycophancy. We conclude that consistency is not validity: consistency under reordering can be optimized without improving the moral judgment it is meant to signal.
[NLP-94] AQL-Lite: Enabling Efficient Knowledge Acquisition for Arabic LLM s via Data-Centric Continual Pretraining
【速读】： 该论文旨在解决阿拉伯语大语言模型（Large Language Model, LLM）训练中面临的三大核心挑战：网络数据噪声、方言多样性以及持续预训练（Continual Pretraining, CPT）带来的高昂计算成本。为应对这些问题，提出了一种高效且资源节约的框架AQL-Lite（Arabic Quality Language），其关键解决方案在于结合“质量优先”的数据清洗流程与部分参数微调（Partial-Parameter CPT）策略。具体而言，AQL-Lite通过结构化清理、启发式与模型驱动的过滤、现代标准阿拉伯语（Modern Standard Arabic, MSA）选择及去重等步骤构建高质量语料库，并采用基于分层学习率的RS-LoRA适配器，在网络最后八层Transformer块及嵌入矩阵上进行参数更新，实现对多语言大模型向MSA的高效适配。实验表明，在10亿令牌预算下，使用AQL-Lite处理的数据可使阿拉伯语多任务平均得分提升3.51点，显著优于原始网络文本的持续预训练；在50亿令牌的对照实验中，AQL-Lite仅用52%的时长即达到全参数持续预训练94%的性能增益，且在所有基准测试中均优于仅使用适配器的基线方法。进一步的消融实验证明，适配器、未冻结层以及嵌入层训练均对性能有贡献。最终，将该框架扩展至400亿条目并指令微调100万样本后，生成的120亿参数模型在七项基准测试的Open Arabic LLM Leaderboard v2（OALL v2）中取得75.3分，验证了在精心筛选语料上进行部分参数持续预训练是实现高效、高性能阿拉伯语模型适配的可行路径。
链接: https://arxiv.org/abs/2610.30344
作者: Houssam Eddine Boukhalfa,Mourad Bouache,Mourad Dahmane
机构: National School of Artificial Intelligence (ENSIA), Algiers, Algeria; Nojoom AI
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 15 pages, 7 figures, 10 tables
Abstract:Training Arabic large language models faces challenges from noisy web data, dialectal variation, and the high cost of continual pretraining (CPT). We introduce AQL-Lite (AQL: Arabic Quality Language), a time- and resource-efficient framework for adapting a multilingual LLM to Modern Standard Arabic (MSA). AQL-Lite couples a quality-first curation pipeline (structural cleaning, heuristic and model-based filtering, MSA selection, and deduplication) with a partial-parameter CPT recipe that trains RS-LoRA adapters across the network together with the last eight transformer blocks and the embedding matrix, under layer-wise learning rates. At a matched 1B-token budget, CPT on AQL-curated data raises a four-benchmark Arabic mean by 3.51 points over CPT on raw web text, which yields no gain over the base model. In a controlled 5B-token comparison, AQL-Lite recovers 94% of the gain of full-parameter CPT (mean 73.19 vs. 73.58) in 52% of the wall-clock time and outperforms adapter-only baselines on every benchmark, and a component ablation shows that adapters, unfrozen blocks, and embedding training each contribute. Scaled to 40B MSA tokens and instruction-tuned on one million examples, the resulting 12B model scores 75.3 on the seven-benchmark Open Arabic LLM Leaderboard v2 (OALL v2) suite. These results suggest that partial-parameter CPT on a carefully curated corpus is a practical route to efficient MSA adaptation.
[NLP-95] Measuring Research Progress and Linguistic Inclusion in Low-Resource Languages
【速读】： 该论文旨在解决生成式 AI（Generative AI）在低资源语言（Low-Resource Languages, LRLs）与多语言自然语言处理（Multilingual NLP）研究中虽取得显著进展，但其对语言多样性与代表性群体的包容性仍缺乏系统评估的问题。核心挑战在于：现有研究增长指标（如论文数量）未能真实反映语言覆盖的公平性与代表性分布的均衡性。论文提出“语言包容性框架”（Linguistic Inclusion Framework, LIF），从四个维度——研究增长、代表性扩展、代表性连续性及分布公平性——综合评估语言技术研究的包容性。关键解决方案在于突破仅依赖发表量的单一评价标准，通过多维量化分析揭示：尽管整体论文数量增长显著，但研究集中于已代表语言，亚洲等高人口基数地区语言的代表性严重不足，而欧洲语言则过度集中；且非洲和东南亚等语言多样性丰富的区域语言覆盖比例最低。研究表明，单纯的研究增长无法体现真正的语言公平，必须通过系统性框架实现对语言包容性的精准测量与改进。
链接: https://arxiv.org/abs/2610.30296
作者: Atnafu Lambebo Tonja,Tadesse Destaw Belay,Jesujoba Oluwadara Alabi,Israel Abebe Azime,Benjamin Rosman,Ehsan Shareghi,Thamar Solorio,Pontus Stenetorp,Monojit Choudhury
机构: University College London, UK(伦敦大学学院，英国); Instituto Politécnico Nacional, Mexico(墨西哥国家理工学院，墨西哥); Saarland University, Germany(萨尔兰大学，德国); University of the Witwatersrand, South Africa(约翰内斯堡大学，南非); MBZUAI, UAE(穆巴达拉科技与人工智能研究院，阿联酋)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large Language Models (LLMs) have accelerated research on low-resource and multilingual NLP, but their effectiveness in creating an equitable landscape remains unmapped. We analyze 1,158 peer-reviewed open-access papers published between 2020 and 2025, examining research attention to low-resource languages (LRLs) and representation across language and speaker populations. To assess linguistic inclusion beyond publication counts, we introduce the Linguistic Inclusion Framework (LIF), a multidimensional framework that measures linguistic inclusion in four dimensions: research growth, representation expansion, representation continuity, and distributional equity. While annual publications grew from 52 to 415, a 5.4X increase in their ACL Anthology publication share, this growth largely concentrated on represented languages rather than expanding coverage of underrepresented languages. Further analysis shows that languages spoken in Asia account for 64.1% of speakers across LRLs we analyze but receive only 46.7% of the language representation share (0.7X their proportional share), while European languages receive 5.3X their proportional share. At the country level, linguistically diverse regions such as sub-Saharan Africa and Southeast Asia show the lowest language coverage ratios. Applying the framework at the continental level, we find that no region scores well on all four dimensions, showing that research growth alone is not a reliable metric of progress toward equitable and inclusive language technology\footnoteWe will release all collected and annotated data for research purposes.
[NLP-96] Finitely Verifiable Segmentation: Decidability for Every Tokenizer with a Finite-State Verifier
【速读】： 该论文旨在解决自然语言处理与编译器设计中字节级分词策略（segmentation policy）的可验证性与计算可判定性问题，核心在于构建一类具有形式化可判定性质的分词策略。其关键解决方案是引入“有限可验证性”（finitely verifiable）这一概念：即一个分词策略若能被一个在标记字符串（marked strings，即字节与边界位配对）上的有限自动机精确接受其所有合法分词结果，则称其为有限可验证的。该自动机仅需验证给定的分词提案是否有效，而无需自主生成分词，从而实现了高效、形式化的验证机制。这一设定使得原本依赖经验或非形式化约定的四个常见问题变为可判定：是否存在窗口可证明的词元边界、边界间隙是否有限、掩码是否保持生成可完成性，以及两个分词策略是否等价。后三者分别对应有界性（limitedness）、共可达性（coaccessibility）及单值转换器的等价性；前者则深化了代码理论中的同步切割（synchronizing cut）概念，并进一步探讨认证切割或库存切割是否允许各片段独立扫描。该类分词策略严格介于在线可扫描策略与具有正则域的策略之间，其内部边界由Choffrut定理刻画，且可在多项式时间内判定。标记操作保持长度不变，因此该类恰好等同于有理分词策略（rational segmentation policies），这是一种规范形式而非限制性约束，其可判定性依赖于提供一个验证器（verifier）。该类策略在管道操作下封闭，包括带标签阶段的复合、限制、拆分以及字面量策略的字节回退并集。每个复合结构继承六项可判定性，且证书在四条传递定律下可跨阶段转移。在有理、无歧义、几乎必然停止的有限状态源下，加权截段的每部分质量均为精确有理数，且期望长度为正时的速率亦为有理数。最大贪心词法分析（maximal-munch lexing）、标准字节对编码（canonical byte-pair encoding）及双模式WordPiece均实例化了该接口，而munch库已实现全部六项判定功能。所有定理均可执行验证，所有报告统计均由程序直接输出，确保了理论与实现的高度一致性。
链接: https://arxiv.org/abs/2610.30168
作者: Nicklas Nidhögg
机构: 独立研究者(Independent Researcher)
类目: Formal Languages and Automata Theory (cs.FL); Computation and Language (cs.CL); Programming Languages (cs.PL)
备注: 36 pages, 1 figure, 2 tables. Artifact: this https URL . The modules, the gate and the records of the library and tokenizer instances are in the artifact, with a verifier that runs every module and holds each reported count to the output line it rests on. Companion to arXiv:2608.03473 , arXiv:2608.09761 , arXiv:2609.10600 and arXiv:2610.08854
Abstract:A segmentation policy assigns to each byte string at most one segmentation into tokens. Call it finitely verifiable when a finite automaton over marked strings, bytes paired with boundary bits, accepts exactly the policy’s own segmentations. The automaton judges a proposed segmentation and need not compute one. The paper shows what that hypothesis buys. Four questions that practice settles by folklore become decidable, whether a window certifies a token boundary, whether the boundary gap is bounded, whether a mask keeps generation completable, and whether two policies are equivalent. The last three specialize limitedness, coaccessibility and the equivalence of single-valued transducers. The first refines the synchronizing cut of code theory, and two more ask whether a certified cut, or the cuts of an inventory, let each piece scan alone. The class sits strictly between the online-scannable policies and those with regular domains, and its inner frontier is Choffrut’s and decidable in polynomial time. Markings are length-preserving, so the class is exactly the rational segmentation policies, a normal form rather than a restriction, and what the decisions need is a supplied verifier. The class is closed under the operations pipelines use: composition over a labeled stage, restriction, splitting, and byte-fallback union of literal policies. Each composite inherits the six decisions, and certificates transfer between stages under four laws. Under rational, unifilar, almost surely stopping finite-state sources, each mass of the weighted section is an exact rational, and so is each rate where expected length is positive. Maximal-munch lexing, canonical byte-pair encoding and two-mode WordPiece instantiate the interface, and the munch library implements all six decisions. Every theorem carries an executable check, and every reported statistic is a program’s own output.
[NLP-97] FusionEval: Benchmarking Model Fusion for Large Language Models
【速读】： 该论文旨在解决多专家模型融合（model fusion）中存在的重要问题：现有评估体系对长链思维推理（long chains of thought）和工具使用（tool use）的覆盖不足，且平均性能表现优异可能掩盖了各专家能力在融合过程中的显著退化。为应对这一挑战，论文提出FusionEval基准，其关键在于提供受控的专家池（controlled expert pools），支持无需额外训练与需额外训练的融合方法，并构建统一评估套件。该套件基于Qwen3与Llama基础模型独立训练的数学、代码、科学、指令遵循及工具使用等领域的专家模型，在八项公开基准上，在相同数据访问条件下评估能力保留度、计算开销（包括运行时、峰值GPU内存占用）以及归一化平均与最差性能。实验表明，无训练融合方法虽能较好保持推理能力，但显著损害工具使用性能；而MOPD方法虽可有效保留工具使用并实现更均衡的能力分布，但需额外的数据准备、推理与训练成本。因此，如何在低计算成本下联合保留推理、指令遵循与工具使用能力，仍是当前亟待突破的关键挑战。FusionEval旨在推动高效模型融合研究，助力未来工作实现多专家能力的协同保留。
链接: https://arxiv.org/abs/2610.30162
作者: Shuo Cai,Yanggan Gu,Yuhang Liu,Su Lu,Yuanyi Wang,Yang Yu,Zhen Li,Zhitian Hou,Guanghao Zhu,Ming Li,Hongxia Yang
机构: The Hong Kong Polytechnic University(香港理工大学); InfiX.ai; PolyU-Daya Bay Technology and Innovation Research Institute
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 20 pages, 4 figures
Abstract:Model fusion aims to integrate the capabilities of multi-expert models into a single model. Existing evaluations offer limited coverage of reasoning with long chains of thought and tool use, while strong average performance can mask substantial losses in individual expert capabilities. We introduce FusionEval, a benchmark that provides controlled expert pools, integrates fusion methods with and without additional training, and offers a unified evaluation suite. The pools contain domain experts independently trained from Qwen3 and Llama base models across math, code, science, instruction following, and tool use. The suite evaluates capability retention and computational cost on eight public benchmarks under comparable data access rules, reporting normalized average and worst performance, runtime, and peak GPU memory usage. Extensive experiments show that training-free methods preserve reasoning performance but substantially degrade tool use at both model scales. In contrast, MOPD preserves tool use and achieves a more balanced capability profile, but requires additional data preparation, inference, and training. Jointly retaining reasoning, instruction following, and tool use at low computational cost remains a challenge. We hope FusionEval supports future research on efficient model fusion that jointly preserves expert capabilities. Our code is available at this https URL.
[NLP-98] auIndianBankBench: Procedure-Aware Action-Oriented Evaluation of Tool-Using Conversational Agents
【速读】： 该论文旨在解决生成式AI在印度零售银行业务场景中因缺乏对操作流程合规性有效评估而导致的可靠性问题。现有基准测试通常仅依据对话结束时的最终记录来判断模型表现，而忽视了过程中关键步骤的执行情况，导致即使模型跳过必要操作或进行未经授权的变更，仍可能被误判为成功。其解决方案的关键在于提出一个面向动作的基准测试工具——TauIndianBankBench，该基准涵盖1,000个模拟客户与银行助手之间的真实零售银行业务任务，不仅评估最终结果，还严格检验每个会话中是否完整执行了所有必需的操作步骤（如在关闭存款账户前获取结清报价），并禁止未授权的修改。实验表明，在仅依赖最终记录评分的情况下，6.5%至12.0%的会话被错误判定为成功，而引入程序合规性与结果双重评估后，能更准确地反映语言模型在印度零售银行场景中的实际表现。
链接: https://arxiv.org/abs/2610.30161
作者: Krishanu Adhikary,Prashant Devadiga
机构: National Payments Corporation of India (NPCI)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 15 pages, 6 figures, 14 tables. Code: this https URL data: this https URL
Abstract:Many Indian banks serve customers through FAQ pages and scripted chatbots. Assistants based on language models could respond faster and carry out requests such as blocking a card, but only if they give correct information and follow the bank’s rules. These rules prescribe a procedure, a set of required steps, such as obtaining a closure quote before closing a deposit. Adapting models to this domain requires measuring how well they follow the procedure, before and after training. Benchmarks usually judge only the records left at the end of a conversation. Many required steps, however, leave the records unchanged, so an assistant that skips them can still be judged successful. We present TauIndianBankBench, an action-oriented benchmark of 1,000 Indian retail-banking tasks where an assistant helps a simulated customer. Besides the final records, it checks that the assistant took every required step and made no unrequested change. For every model, 6.5-12.0% of sessions are judged successful when only the final records are checked, although the assistant skipped a required step or made an unrequested change. Scoring both procedure and outcome shows more accurately how well language models perform in Indian retail banking.
[NLP-99] Lighthouse: On Building a Trustworthy Digital Assistant for Exploring History with Civil Rights Archives
【速读】： 该论文旨在解决如何利用生成式AI（Generative AI）安全、可靠地支持对美国民权运动时期历史的探索，尤其针对南卡罗来纳州的多模态史料数据。其核心挑战在于，在提供自然语言交互的同时，确保内容的可信性并最小化生成回答可能带来的潜在危害。解决方案的关键在于提出一种神经符号（neuro-symbolic）方法，通过结合大型语言模型（LLM）的神经学习能力与基于规则的符号推理机制，实现对答案风险与覆盖广度之间的平衡。该系统“Lighthouse”在设计上强调内容可信性与用户保护，通过整合超过1.8万份原始资料，并构建由独立专家共识形成的主观问答评估数据集，实现了对事实性与主观性问题的双重验证，从而在提升可解释性与安全性的同时保障了知识覆盖的全面性。
链接: https://arxiv.org/abs/2610.30158
作者: John A. Aydin,Angelica Williams,Steven Dindl,Nathan Saunders,Kate Boyd,Vandana Srivastava,Biplav Srivastava
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Artificial intelligence (AI), as a decision-support aid, provides new tools for studying history effectively using multimodal data, including text, images, and audio. We present, to our knowledge, the first digital assistant named Lighthouse–commonly called a chatbot–for exploring the Civil Rights period of the United States using data from South Carolina. From an AI perspective, the system’s novelty lies in its guiding design principles of (a) focusing on trustworthiness of content and (b) by modeling and reducing potential harm to users from generated answers. It implements these with a neuro-symbolic approach that combines neural learning through a large language model (LLM) with rule-based symbolic solving, balancing risk of answers with generality of coverage. We evaluate Lighthouse with both factual objective questions that admit a single acceptable answer as well as subjective questions. For the latter, a dataset of subjective question-answer pairs (QAs) is created by discovering a “gold standard” by aggregating answers from independent experts. Lighthouse has access to over 18k primary sources from a range of collections and can be easily extended.
[NLP-100] baseVultra: Compact Nonlinear Recurrence with Reallocated Capacity
【速读】： 该论文旨在解决传统循环神经网络中门控机制（gated cell）对计算资源的过度占用问题，特别是针对长短期记忆网络（LSTM）和门控循环单元（GRU）等主流架构在紧凑模型配置下效率与性能之间的权衡。其核心挑战在于：尽管门控结构在序列建模中表现优异，但其复杂的门控矩阵计算消耗了大量参数与算力，限制了模型在轻量级场景下的可扩展性。为此，作者提出baseVultra这一紧凑型循环单元，其关键创新在于将递归计算预算从传统的门控变换重构至候选生成阶段——通过仅保留一个H×H的密集候选生成矩阵作为主要递归路径，而将原本由门控矩阵承担的递归作用替换为O(H)规模的对角交互项，从而显著降低计算开销。同时，该架构引入低秩辅助分支、指数移动平均（EMA）缓存及可学习输出融合机制，并通过轻量级二阶信息访问历史状态（h_t-2），实现高效且精准的状态更新。实验表明，在小型数据集如TinyShakespeare上，baseVultra以约54%的密集矩阵乘法量达到优于匹配参数量LSTM与GRU的验证损失；在更大规模文本8（text8）任务中，其性能接近全门控细胞；而在形式语言探测任务中，其优势在隐状态容量受限时尤为明显，随着隐藏维度增大则逐渐消失。这揭示出：将递归能力集中于候选生成环节的策略在紧凑模型中具有显著优势，而全门控结构在大规模场景下仍具竞争力。
链接: https://arxiv.org/abs/2610.30154
作者: Linke Zhang
机构: Southwest University (西南大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Preprint. 14 pages, 3 figures. Revised based on peer review feedback from WAIC Academic 2026 (1 weak accept, 2 borderline)
Abstract:Nonlinear recurrent networks remain useful for streaming sequence modeling, yet gated cell design is still dominated by LSTM and GRU. We study whether recurrent computation can be reallocated away from gate transforms and toward candidate generation. We propose baseVultra, a compact recurrent cell that concentrates the dense recurrent budget on a single H x H candidate-generation matrix, while replacing recurrent gate matrices with O(H) diagonal interactions. The candidate update combines this dense branch with two low-rank auxiliary branches, an EMA cache, and learnable output blending; the gates additionally receive lightweight second-order access to h_t-2 through diagonal terms. On TinyShakespeare character-level language modeling, baseVultra achieves a validation loss of 1.594 in the compact one-layer 200K setting, compared with 1.604 for a parameter-matched LSTM and 1.616 for a parameter-matched GRU, while its dense recurrent matrix-vector computation is about 54% of GRU under matched hidden size for this configuration. On text8, a roughly 45x larger benchmark, baseVultra is competitive with full-gate cells (1.854 bpc vs. 1.849 for GRU and 1.837 for LSTM). On formal-language probes, the largest gains appear under constrained hidden-state capacity (0.991 vs. 0.761 GRU and 0.691 LSTM OOD2 accuracy on Counter at H=16) and vanish as hidden size grows. These results characterize where and when candidate-focused recurrent capacity allocation helps: the benefit is concentrated in compact regimes, while full-gate cells remain competitive at scale.
[NLP-101] When Multiple Choices Are Correct: Asymmetric Probability Allocation in Jev and Other Decision Models
【速读】： 该论文旨在解决生成式决策模型在多正确选项场景下概率输出可靠性不足的问题。尽管现有决策模型在单一正确答案任务中表现出良好的校准性，但在存在多个正确选项时，其输出的概率分布仍可能无法准确反映真实选择的合理性。研究发现，即使所有正确选项均被正确排序于干扰项之上，在28%至50%的情况下，仍有至少一个正确选项的概率低于均匀概率（1/n，n为候选数量），表明概率分配存在系统性偏差。此外，当向模型输入五个完全相同的正确选项时，其输出概率差异显著，平均差距达17至79个百分点，暴露了概率输出的不一致性。这些现象导致基于概率阈值的筛选策略无法有效召回所有正确选项。解决方案的关键在于：决策模型的概率输出虽能实现正确选项的相对排序，但不能依赖单一概率阈值来识别完整的正确选项集合，必须重新审视概率解释逻辑与多答案场景下的评估范式。
链接: https://arxiv.org/abs/2610.29217
作者: Rushab Munot
机构: 独立研究员(Independent Researcher)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 11 pages, 2 figures. Code: this https URL
Abstract:Decision models return normalized probability distributions over candidate choices, which downstream applications can use to select among choices. Although their providers emphasize the reliability of these probabilities, calibration on tasks with a single correct choice does not establish their reliability when multiple choices are correct. We evaluate three decision models (Jev, Liquid d1, and GPT-6 Luna Decisions) on three diverse datasets (MetaTool, HotpotQA, and ECtHR) to study probability allocation and thresholding. Even when models rank every correct choice above every distractor, in 28% to 50% of those lists, at least one correct choice receives a probability below the uniform probability of 1/n, where n is the number of choices. We also present each model with five identical copies of a correct choice and observe probability disparities, with mean gaps ranging from 17 to 79 percentage points. We further show how these allocation disparities affect the recovery of multiple correct choices under probability thresholding. Decision-model probabilities can rank correct choices, but a probability threshold does not identify the set of correct choices.
[NLP-102] ChemBART: Pre-training a Reaction Language Model for Comprehensive Synthesis Planning
【速读】： 该论文旨在解决当前大语言模型（LLM）在计算机辅助合成路线规划（CASP）中应用受限的问题，即现有方法普遍局限于单分子表示，仅能处理孤立的、单一任务场景（如前体预测），难以支持多步合成路径设计与复杂反应条件优化。其解决方案的关键在于提出ChemBART——一种基于BART架构并采用反应中心预训练策略的语言模型，通过在完整的反应表达式（反应物 + 试剂 + 产物）上进行掩码填充预训练，实现了对化学反应整体语义的建模。该设计使模型能够统一处理多种化学任务，包括前体/试剂生成、温度-产率回归及分子性质分类，并进一步与蒙特卡洛树搜索（MCTS）结合，在强化学习框架中优化策略函数与价值函数，从而实现多步合成路线的智能规划。最显著的是，所提出的合成路径及反应条件经湿实验验证，成功实现了比文献基准更短的合成路径和20%的产率提升，证明了反应导向预训练策略在打通“计算预测—湿实验验证”闭环中的关键作用。
链接: https://arxiv.org/abs/2601.02915
作者: Kenan Li,Yijian Zhang,Jin Wang,Haipeng Gan,Zeying Sun,Xiaoguang Lei,Hao Dong
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Under peer review. 31 pages, 6 figures, 2 tables
Abstract:Recent advances in large language models (LLMs) have demonstrated transformative potential across diverse fields. However, in computer-aided synthesis planning (CASP), current LLM methodologies are often restricted to single-molecule representations, limiting their application to isolated, single-task scenarios such as precursor prediction. Addressing this gap, we present ChemBART, a BART-based language model that adopts a reaction-centric pre-training strategy. By leveraging a mask-filling pre-training task on holistic reaction expressions (Reactant Reagent Product), ChemBART effectively solves a variety of chemical problems, including precursor/reagent generation, temperature-yield regression, and molecular property classification. Building on these capabilities, we demonstrate the integration of ChemBART with Monte Carlo tree search (MCTS) for multi-step synthesis route design, where the model serves to optimize the policy and value functions within a reinforcement learning framework. Beyond single-molecule pre-trained LLMs that are constrained to specific applications, ChemBART unifies diverse chemical tasks to facilitate synthesis planning with comprehensive implementation information. Most significantly, the proposed multi-step synthesis routes and reaction conditions were validated in the wet lab, confirming the discovery of shorter synthetic pathways with a 20% yield improvement over established literature benchmarks. Our work highlights the potential of reaction-focused pre-training and showcases the broad utility of ChemBART in closing the loop between in silico prediction and wet-lab synthesis.
[NLP-103] Domain-Specific Constitutional AI: Enhancing Safety in LLM -Powered Mental Health Chatbots
【速读】： 该论文旨在解决当前生成式AI在计算心理健康应用中因通用安全机制不足而引发的特定风险问题，尤其针对情绪脆弱人群所面临的误诊、症状恶化、危机干预失准及治疗指南偏离等关键挑战。其解决方案的核心在于引入基于领域特异性心理健康原则的宪法式AI（Constitutional AI, CAI）训练框架，通过将临床伦理规范、治疗一致性要求与危机响应准则等专业原则嵌入模型训练过程，实现对敏感对话场景的精准适配与安全约束，从而构建具备可信赖性、合规性与情境适应性的专用心理健康AI系统。
链接: https://arxiv.org/abs/2509.16444
作者: Chenhan Lyu,Yutong Song,Pengfei Zhang,Amir M. Rahmani
机构: University of California, Irvine (加州大学欧文分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Accepted to 2025 IEEE 21st International Conference on Body Sensor Networks (BSN)
Abstract:Mental health applications have emerged as a critical area in computational health, driven by rising global rates of mental illness, the integration of AI in psychological care, and the need for scalable solutions in underserved communities. These include therapy chatbots, crisis detection, and wellness platforms handling sensitive data, requiring specialized AI safety beyond general safeguards due to emotional vulnerability, risks like misdiagnosis or symptom exacerbation, and precise management of vulnerable states to avoid severe outcomes such as self-harm or loss of trust. Despite AI safety advances, general safeguards inadequately address mental health-specific challenges, including crisis intervention accuracy to avert escalations, therapeutic guideline adherence to prevent misinformation, scale limitations in resource-constrained settings, and adaptation to nuanced dialogues where generics may introduce biases or miss distress signals. We introduce an approach to apply Constitutional AI training with domain-specific mental health principles for safe, domain-adapted CAI systems in computational mental health applications.
信息检索
[IR-0] A Task-Centric Survey of Long-Document Visual Question Answering: Taxonomy Evolution and Future Directions
链接: https://arxiv.org/abs/2610.31385
作者: Yufei Tong,Guanjie Cheng,Tianle Wu,Leilei Zhang,Hang Zheng,Peihan Wu,Jianwei Yin,Shuiguang Deng
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 26 pages, 9 figures, 6 tables
Abstract:Visual document understanding is essential for acquiring knowledge from visually rich documents and is attracting growing research interest. Existing surveys emphasize architectures or technical approaches, with limited attention to how task bottlenecks shape technical choices and research priorities. We present a task-centric survey that decomposes long-document visual question answering (VQA) into four components: document parsing, evidence organization and retrieval, generator organization and learning, and evaluation. For each, we develop a fine-grained taxonomy and trace motivations, technical responses, and cross-component connections. We focus on explicit retrieval while covering direct reading. Our temporal analysis traces expanding priorities from multimodal content understanding to sufficient, high-fidelity evidence acquisition and sustained evidence management. We outline future directions in evidence sufficiency, verifiable and correctable understanding, and budget-aware coordination, with an outlook on continual document understanding.
[IR-1] SurveyGraph: Generating Academic Surveys via Methods-Entity Knowledge Graphs EMNLP2026
链接: https://arxiv.org/abs/2610.31317
作者: Yifan Yuan,Yancheng Zhu,Zuoyu Zhang,Xiaogang Peng
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: Accepted at EMNLP 2026 Findings. 34 pages including appendices
Abstract:Most automated survey generation systems retrieve papers as flat documents, limiting their ability to capture research structure. We present SurveyGraph, a framework built around a Methods-Entity Knowledge Graph (MEKG) encoding method-level follow-on, competition, and shared-resource relations. The graph supports core-paper discovery, retrieval, hybrid outlining, and evidence-gated narrative synthesis. We evaluate retrieval, semantic edges, citation grounding, collection coverage, and generation-stage components. In a controlled seven-method equal-budget benchmark, MEKG-anchored fusion balances recall and structural density, with an independent co-citation check. In a separate archived graph-native diagnostic, graph-aware retrieval nearly doubles recall over flat lexical baselines across eight runs (p = 0.004). A semantic evidence gate achieves AUC = 0.75. Audits of 160 generated sentences find 80.6% anchored-or-partial citation support. Human adjudication of 108 stratified semantic edges finds 68.5% supported-or-partial relations (binary-support Cohen’s kappa = 0.64), with improves strongest (80.6%) and extends weakest (52.8%). A SurveyBench-facing collection probe reaches mean reference coverage of 0.521 across four topics with an eight-paper budget. Same-backend Qwen diagnostics on a 30-paper/74-relation graph show reduced organization signals when hybrid outlining is removed. Concurrent-system comparisons provide bounded context rather than matched-backend end-to-end evidence. The results position MEKG as an auditable structural layer for typed diagnostics, relation-specific filtering, and controlled survey generation. Code and artifacts: this https URL.
[IR-2] he Delayed Countdown Method for Validating Review-Oriented AI Workflows in eDiscovery
链接: https://arxiv.org/abs/2610.31311
作者: David D. Lewis,Tara Emory,Lenora Gray,Mark Noel,Jeremy Pickens,Eugene Yang
类目: Information Retrieval (cs.IR)
备注:
Abstract:Review-oriented workflows have become the dominant approach to AI-assisted document review in electronic discovery (eDiscovery). How to validate the effectiveness of review-oriented workflows has, however, remained a confusing and controversial topic in technology-assisted review (TAR) in the law. We examine the sources of this controversy, and propose a simple and practical evaluation approach for review-oriented workflows. Our Delayed Countdown method is easily incorporated into existing review-oriented workflows and software, is compatible with arbitrary stopping rules, and produces valid frequentist confidence intervals on effectiveness. We explain the conceptual foundations of the method, provide example computations, discuss the handling of logically impossible recall values, and suggest future work on decreasing review costs.
[IR-3] Chunk Recomputation under Real Document Revisions: Measuring Update Cost and Retrieval Quality of Fixed Recursive and Content-Defined Chunking
链接: https://arxiv.org/abs/2610.31019
作者: Yuning Bian,Lele Xu,Yifan Yuan,Xinyu Yin,Chengyan Zhan
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 6 pages, 3 figures, 3 tables. Accepted at AIBDCC 2026
Abstract:Chunking strategies for retrieval-augmented generation are almost always evaluated on a static snapshot, yet the documents underlying a knowledge base continue to change. Because chunks are embedded independently, a boundary that moves after an edit forces unchanged text to be re-embedded. We measure this update cost on 120 pairs of adjacent Chinese and English Wikipedia revisions for three stateless chunkers (non-overlapping fixed-size, recursive paragraph-first, and content-defined chunking) under a strict byte-identical reuse rule and a frozen dense encoder. On the 116-article common-success pool, with frozen, length-calibrated parameters, fixed-size chunking recomputes 57-68% of its full new-version embedding-input workload, recursive chunking 38-56%, and content-defined chunking 43-54% (it failed on 4 of 62 Chinese articles), while at these parameters mean evidence recall under a fixed 2048-token retrieval budget differs by less than two percentage points within either language. A cost-only parameter grid shows that the ranking of the two content-aware chunkers changes with chunk-length settings, whereas non-overlapping fixed-size chunking has the highest recomputation cost in every length-matched condition. Controlled edits and a verbatim-presence diagnostic support boundary drift as one contributor to recomputation of text already present in the previous version. Update cost is therefore a substantial, parameter-sensitive dimension that static retrieval evaluations do not expose and that should be reported alongside retrieval quality.
[IR-4] Formal-Informal Mathematical Alignment through and beyond Human Annotations CIKM’26
链接: https://arxiv.org/abs/2610.31011
作者: Noah Gießing,Ankit Satpute,Hamid Rahkooy,André Greiner-Petter,Moritz Schubotz
类目: Information Retrieval (cs.IR)
备注: Accepted at: Proceedings of the 35th ACM International Conference on Information and Knowledge Management (CIKM '26), November 7-11, 2026, Rome, Italy
Abstract:Formalized mathematics libraries such as Lean’s \textttmathlib provide machine-verifiable representations of mathematical knowledge, including definitions, theorems, and their relationships. If effectively connected to informal mathematical resources, formal mathematics could help uncover previously implicit links between mathematical concepts and improve access to existing formalizations. However, connections between informal and formal mathematics libraries are currently sparse and unsystematic. In this paper, we make first strides towards better formal integration. We construct a dataset linking informal statements from the Stacks Project to their formalizations in \textttmathlib by exploiting expert annotations. We evaluate retrieval, LLM-based reranking, and LLM-assisted verification to discover new unannotated informal-to-formal links. To facilitate exploration and reuse of formalized knowledge, we integrated the newly discovered alignments into the MaRDI knowledge graph, enabling mathematicians to find formal counterparts to informal mathematical content. We discuss the difficulties and challenges in using LLMs for linking formal and informal mathematical content. All code and data necessary for reproducing this paper can be found at: this https URL
[IR-5] PAIR: Preference-Aligned Intent Retrieval for Premium People Recommendation at LinkedIn
链接: https://arxiv.org/abs/2610.30941
作者: Zhongkai Shangguan,Yanbin Kang,Yihan Cao,Ruijie Jiang,Qiannan Yin,Chang Yue,Zijian Zhou,Girish Kathalagiri,Changshuai Wei,Zhengming Xing
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注:
Abstract:General-purpose embeddings trained on profile similarity often misrank candidates in professional networks, where relevance is inherently complementary rather than symmetric. For example, job seekers usually engage with recruiters rather than peers with similar credentials. Current user recommendation systems are often built on profile similarity or social signals such as shared connections, which are effective in friend-oriented social networks where users tend to connect with similar people or within existing social circles. However, this does not hold in professional platforms like LinkedIn, where interactions are often goal-driven and cross-role interactions, requiring matching based on complementary intents rather than similarity or social proximity. In this paper, we propose Preference-Aligned Intent Retrieval, a retrieval framework that is designed to model user intent preference instead of profile similarity. Our approach is inspired by the preference-learning perspective of Direct Preference Optimization, but instantiates preference learning as an exposure-conditioned multi-negative contrastive objective over embeddings. This formulation simplifies training and scales efficiently to large candidate sets without requiring an explicit reference model. Furthermore, we incorporate user interaction history by aggregating click histories with positional encoding to capture temporal order, which enables constant-cost updates of user representations over a fixed history window, allowing new interactions to be incorporated without re-encoding history. As a result, personalization can be updated efficiently using lightweight CPU operations without repeated LLM inference during serving. We evaluate PAIR on LinkedIn’s Premium retrieval task. A 0.6B parameter model achieves a MAR@100 of 0.45.
[IR-6] UserArena: Benchmarking Interactive User Simulation for Long-Horizon Shopping Agents
链接: https://arxiv.org/abs/2610.30932
作者: Jujia Zhao,Jiakai Tang,Jing Yu,Wen Chen,Jian Wu,Han Zhu,Fei Sun,Suzan Verberne,Zhaochun Ren
类目: Information Retrieval (cs.IR)
备注:
Abstract:Agentic recommendation pursues shopper goals through sequences of context-dependent decisions. Developing and evaluating such policies requires faithful recommendation world models that capture both the interactive shopping environment and users’ behavioral responses over long interactions. Existing resources rarely provide long-horizon, fine-grained real-user trajectories or support trajectory-level evaluation and offline validation against production A/B results. We introduce UserArena, a real-world e-commerce interactive benchmark containing 64,729 session trajectories from 11,010 users, spanning 566,102 items and 13 action types. UserArena pairs these trajectories with an interactive environment and supports offline validation against outcomes from a controlled production A/B test, enabling five-level evaluation from output validity and local decisions to sequential behavior, personalization, and outcomes. To improve simulator fidelity over these long-horizon interactions, we further propose On-Policy Trajectory Distillation (OPTD), which uses trajectory-level intent to provide dense teacher targets on student-generated rollouts. Experiments show that OPTD achieves the strongest overall fidelity among the evaluated simulators under UserArena’s evaluation protocol and most closely reproduces the production policy effect in offline simulation.
[IR-7] Lean Finder V2: Semantic Search for Mathlib That Disambiguates User Queries
链接: https://arxiv.org/abs/2610.30843
作者: Jialin Lu,Xiaoyu Liu,Hengrui Luo,Rodrigo Stehling,Linxi Liu,Weiran Sun,Yuepeng Wang,Wuyang Chen
类目: Machine Learning (cs.LG); Information Retrieval (cs.IR)
备注:
Abstract:We introduce Lean Finder V2, a semantic search framework for Mathlib that disambiguates user queries through complementary model-level adversarial training and input-level query rewriting. Human mathematicians and coding agents rarely search Mathlib by providing detailed context. In practice, they query short properties, notation, aliases, or hypotheses, leaving much of the intended mathematics implicit. In contrast, existing Mathlib search engines unrealistically assume context-rich queries. The scarcity of real-world short queries paired with ground-truth Mathlib statements makes this ambiguity even more intractable. We argue that this is essentially a data problem. Through adversarial training, we scale up the generation of labeled short and ambiguous queries, and learn to disambiguate user queries at both model-level fine-tuning and input-level query rewriting. On the human-curated MathlibQR benchmark, our Lean Finder V2 sets a new state of the art, with at least 11% relative improvement over the previous best Mathlib semantic search engines. On top of our state-of-the-art retriever, optimized query rewriting further improves MathlibQR retrieval. On unlabeled queries from human and coding-agent interactions, Lean Finder V2 is ranked as the strongest system. Integrated with Claude Code, Lean Finder V2 improves proof success on downstream theorem-proving task. Lean Finder V2 is available at: this https URL
[IR-8] Bezier-Guided Adaptive Flow for Cross-domain Sequential Recommendation
链接: https://arxiv.org/abs/2610.30825
作者: Ziang Lu,Yu Zhang,Xinzhe Jiang,Lei Sang,Yiwen Zhang
类目: Information Retrieval (cs.IR)
备注:
Abstract:Cross-domain sequential recommendation aims to alleviate data sparsity and cold-start problems by transferring useful preference knowledge across domains. However, users exhibit varying degrees of preference consistency across domains, while existing methods often rely on global alignment objectives or coarse-grained transfer strategies. Such approaches may introduce irrelevant information and provide limited flexibility in adapting semantic representations to individual user preferences. To address these challenges, we propose BGAFlow, a personalized framework that combines state-adaptive semantic evolution with user-specific transfer control. BGAFlow first establishes a shared semantic space through unified dual-domain pretraining with user profiles inferred by a large language model (LLM). Based on latent interest overlap, it then estimates a personalized horizon to determine the appropriate transfer extent for each user. To enable semantic adaptation, BGAFlow constructs a conditional Bezier curve as a geometric reference and learns a residual flow that progressively refines the trajectory based on intermediate preference states. The learned flow incorporates dynamic preference feedback to adjust transfer directions, while recommendation supervision and trajectory regularization encourage effective and smooth adaptation. Extensive experiments on real-world datasets demonstrate that BGAFlow consistently outperforms state-of-the-art baselines, validating the benefits of personalized transfer control and state-adaptive semantic evolution.
[IR-9] Automated Enrichment Analysis with GraphRAG : Efficient Multi-Database ORA Extraction and Biological Interpretation
链接: https://arxiv.org/abs/2610.30710
作者: Guangcan Li,Shuyuan Zheng,Jiayao Liu,Shigeto Seno,Makoto Onizuka
类目: Information Retrieval (cs.IR); Databases (cs.DB)
备注: Accepted to the 2026 IEEE International Conference on Bioinformatics and Biomedicine (IEEE BIBM 2026)
Abstract:Over-representation analysis (ORA) is widely used to interpret gene-expression studies, but its outputs are often fragmented across gene-set databases and presented as flat term lists that obscure gene-level links among related enrichment signals. We propose Graph-Based Over-Representation Retrieval-Augmented Generation (GBOR-RAG), an integrated and traceable framework for ORA interpretation. GBOR-RAG represents heterogeneous gene-set databases as a unified term–gene knowledge graph, where shared genes connect related terms across sources. After graph-based ORA scoring, the framework retrieves a compact connected evidence subgraph that preserves both cross-database relationships and gene-level support. The subgraph conditions an LLM to generate biological interpretations, and we introduce a claim-level faithfulness metric to assess whether generated claims are supported by explicit evidence. Experiments show that graph-based ORA preserves conventional ORA results with high consistency (mean Pearson r=0.997 ) while reducing query time by a median of 37\times (up to 48\times ). The retrieved subgraphs average 36.8 nodes and achieve the highest overall integration score (4.12), and GBOR-RAG attains the highest claim-level faithfulness (0.90) among the compared systems. Adding a PubMed literature layer further improves the overall interpretation quality from 4.17 to 4.75.
[IR-10] Lychee: Certified Screening of Inverted Lists for High-Recall Vector Search
链接: https://arxiv.org/abs/2610.30697
作者: Dongfang Zhao
类目: Information Retrieval (cs.IR)
备注: 14 pages, 12 figures
Abstract:Nearest-neighbor search over embedding vectors matches faces against watchlists, transactions against known fraud patterns, and medical images against confirmed cases, where a missed neighbor is a missed threat, fraud, or diagnosis. Because an approximate index cannot report which queries lost a true neighbor, these applications require recall of 0.999 to 0.9999, at a cost far below that of exact search over billions of vectors. Inverted-file indexes reach high recall by probing more lists, but they read and score every probed vector at full precision, so that throughput falls in inverse proportion to the number of probed lists. Graph indexes saturate below these targets, whereas quantized lists must rerank their candidates with the full vectors. Our insight is to decide for almost every probed vector that it cannot enter the top k of the probed lists without reading the vector, through the certified lower bound of exact search by certified exclusion. This paper turns this insight into Lychee, a GPU inverted-file search that screens the probed lists on tensor cores under a threshold from anchors of the same lists and computes exact distances only for the few survivors. We prove that Lychee returns exactly the k nearest vectors of the probed lists, the answer of an inverted-file search at full precision, while it reads a few hundred full vectors per query regardless of the number of probed lists. On eleven datasets from one million to one billion vectors and against ten GPU and CPU baselines, Lychee is 9.7 to 117 times faster than IVF-Flat at equal recall with 200 probed lists on datasets of up to 100 million vectors. It also reaches recall levels that CAGRA and HNSW do not reach and outruns exact search by certified exclusion by 5 to 18 times above ten million vectors. On four GPUs, Lychee searches one billion vectors at recall 0.9999, 44.5 times faster than IVF-Flat on BIGANN-1B.
[IR-11] OneLatent: Latent Reasoning for Efficient Foundation Recommendation Models
链接: https://arxiv.org/abs/2610.30675
作者: Hao Jiang,Peiru Du,Pengfei Yao,Mengting Li,Siyuan Lou,Kuo Cai,Sheng Yu,Qiang Luo,Jian Liang,Ruiming Tang,Fei Pan,Peng Jiang,Wenwu Ou
类目: Information Retrieval (cs.IR)
备注:
Abstract:Large language models (LLMs) have demonstrated strong reasoning capabilities, motivating their use as the backbone of foundation recommendation models (FRMs). Existing methods enhance recommendations through explicit Chain-of-Thought (CoT) reasoning under a Think-then-Answer paradigm. However, explicit CoT incurs substantial inference overhead by generating lengthy reasoning traces and relies on manually designed templates that struggle to capture diverse, dynamic user interests. We propose OneLatent, an efficient latent reasoning framework that compresses explicit reasoning traces into several learnable latent tokens, enabling Latent-Reason-then-Answer inference without generating verbose traces. OneLatent first introduces Multi-View Adaptive CoT (MV-ACoT), which creates diverse, high-quality teacher-generated supervision by exploring user interests from multiple perspectives and automatically adapting reasoning complexity to each instance. Building on pretrained FRMs, it then uses a three-stage latent-token alignment paradigm to progressively internalize CoT traces into learnable latent tokens. Finally, a multistage curriculum-based post-training strategy activates latent-token reasoning for downstream recommendation tasks. Experiments on an industrial-scale Kuaishou dataset and the public Kuaishou LLM-Rec benchmark show that OneLatent consistently outperforms explicit CoT-based methods and traditional baselines. Compared with the Think and No-Think variants of FRMs, OneLatent improves SID@64 by 17.44% and 9.33%, respectively, while achieving over 17x higher online inference throughput. We further develop a production serving system for scalable, real-time FRM inference. An online A/B test in Kuaishou’s local-services advertising scenario shows that deploying OneLatent with this system yields an estimated 9.6% revenue lift over strong online baselines, including OneRec and OneReason.
[IR-12] Compiling Autoregressive LLM Control into Policy-Exact Parallel Inference for Hierarchical RAG
链接: https://arxiv.org/abs/2610.30664
作者: Zehao Lin,Qianfeng Shen,Xixuan Hao,Shaobo Cui,Chunyu Li,Kai Chen,Bo Tang,Zhenjun Wang,Jin Yan,Zhiyu Li,Feiyu Xiong
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 5 pages, 2 figures, 1 table
Abstract:Large language models increasingly act as controllers in structured decision pipelines, yet their decisions are still executed by autoregressive text decoding. Hierarchical RAG is one such setting: at each frontier of a document tree, an LLM commits to b of k sections under an input-dependent action grammar, executed as either failure-prone free-form JSON or reliable but strictly serial grammar-constrained generation. We show that this serialization is unnecessary for bounded-horizon control: the reachable future forms a shallow action-prefix tree, which we compile — one root prefill plus one tree-causal probe evaluate every prefix-conditioned action distribution with serial decoding’s tokens, positions, and causal ancestry. A prefix-state lemma makes compiled control policy-exact in exact arithmetic: the same policy with a two-forward critical path and zero output tokens. One further probe pass certifies the exact order-marginalized MAP set. Across four document-disjoint splits and two scales on the primary MetaX device, compiled control matches the serial policy on 84 – 89% of questions — every instrumented divergence a floating-point near-tie — and cuts control latency by 35 – 47% versus token-wise serial decoding and 6 – 7% versus an idealized chunked engine; an H800 repetition preserves the ordering.
[IR-13] he Fusion Tax: A Diagnostic Analysis of When Hybrid Retrieval Hurts
链接: https://arxiv.org/abs/2610.30612
作者: Aaditya Chauhan,Arun Agarwal
类目: Information Retrieval (cs.IR); Computation and Language (cs.CL)
备注:
Abstract:Hybrid retrieval, which fuses a sparse lexical retriever with a dense neural one, is widely recommended in the software documentation surveyed in September 2026, usually with reciprocal rank fusion (RRF), on the strength of aggregate gains. We ask what fusion costs on individual queries, treating “do not fuse” as a first-class outcome, in the setting in which hybrid retrieval is deployed: a standard Anserini BM25, five dense encoders spanning 22M-596M parameters and three model generations, a learned-sparse (SPLADE) alternative, and a cross-encoder second stage, on seven BEIR datasets and three fusion functions. Even the best method, dense-weighted score interpolation, ranks relevant documents below the better of its two components on 13-48% of queries and below the deployable dense-only retriever on 8-21% (14-48% and 11-25% at a weight selected under leave-one-corpus-out validation), while beating dense-only in aggregate on every corpus with the primary encoder; the downside is significant under loss-averse risk-sensitive evaluation on six of the seven corpora. We call this loss the fusion tax. It is mostly a ranking effect that preserves recall in aggregate, it is most likely on the queries that only one retriever covers, and it persists across the encoder ladder and the SPLADE substitution. RRF carries the higher tax wherever the dense retriever is clearly stronger and only beats dense-only where the two components are balanced. Fusion failure is weakly predictable within a corpus but transfers unreliably across corpora, because the features that predict it change sign with the sparse-dense balance. A cross-encoder recovers part of the tax only where it is itself reliable. We distill these findings into guidance that keys the fusion decision to the measured sparse-dense balance of the target corpus.
[IR-14] Better Reasoning Not Just Bigger Models: A Structure-Grounded Framework for Efficient and Reliable Inference
链接: https://arxiv.org/abs/2610.30611
作者: Amirhossein Aghaei,Liangyan Li,Shiyi Yang,Shaohua Wang,Yan Wang
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Scaling large language models has substantially improved linguistic fluency, yet reliable autonomous agents require more than fluent generation. In multi-turn inference, current agentic systems often produce semantically plausible but structurally ungrounded reasoning, especially under incomplete observations. We propose Structure-Grounded Uncertainty Reduction (SGUR), a training-free framework that complements token-level generation with active uncertainty reduction over structured hypothesis spaces. SGUR represents incomplete observations as a query graph and aligns them with candidate hypothesis graphs through a prior-weighted partial graph matching mechanism, where database-derived empirical priors assign greater importance to decision-relevant relations rather than treating all graph edges uniformly. The resulting inference state is iteratively refined by an active information-gain strategy that selects the next most informative node to query, systematically pruning the candidate space. We instantiate SGUR on a challenging PrimeKG-derived sequential medical diagnosis benchmark, demonstrating that structural grounding can partially substitute for model scale: starting from only two observed symptoms, SGUR achieves 92.5% Hit@3, outperforming the strongest Graph-RAG baseline by 9.4 percentage points while using approximately 5x fewer tokens. Notably, SGUR enables smaller backbone models to match or exceed the performance of models up to 10x larger, suggesting that structural grounding provides a more robust and efficient paradigm for scaling agent performance than relying exclusively on increasing model size.
[IR-15] Journal Basket: Supporting Cross-Disciplinary Literature Searching and Methodological Transparency
链接: https://arxiv.org/abs/2610.30544
作者: Michael Cahalane
类目: Digital Libraries (cs.DL); Information Retrieval (cs.IR)
备注: 18 pages, 2 Figures, 1 Table
Abstract:Business students beginning a research project must learn how to identify relevant scholarly publications, distinguish disciplinary sources, construct database searches, and report the decisions behind their literature reviews. These tasks can be difficult because journal classification systems and bibliographic databases organize scholarly information differently. This paper introduces Journal Basket, a browser-based tool developed to help business students conduct journal-bounded literature searches and currently available in beta. Journal Basket enables users to construct journal collections from the Australian Business Deans Council (ABDC) Journal Quality List, choose from 42 business discipline presets, entire ABDC fields or the Financial Times FT50 list, apply journal rating thresholds, and generate search expressions for Scopus, Web of Science and Business Source. Users can also inspect and modify journal collections, download journal inventories, and generate methodological descriptions and deterministic identifiers for journal selections. Companion learning resources use explanations, examples and interactive activities to teach review methods, database choice, referencing and the use of contemporary AI tools. Drawing on the journal-basket concept used by this http URL, I describe the educational motivation, design rationale, technical architecture, and practical workflow of Journal Basket. I distinguish reproducing a journal collection from reproducing an entire database search and discuss the limitations of journal rankings as search boundaries. The paper addresses scholarly information retrieval through a workflow for defining journal boundaries, generating database-specific queries and documenting search decisions.
[IR-16] Getting a GRIP on Graph-Based Search for Learned Sparse Retrieval
链接: https://arxiv.org/abs/2610.30425
作者: Vihan Lakshman,Franco Maria Nardini,Cosimo Rulli,Rossano Venturini
类目: Information Retrieval (cs.IR); Databases (cs.DB)
备注: 8 pages
Abstract:Learned sparse retrieval (LSR) has recently gained traction in the information retrieval community for the ability to combine the generalization capabilities of neural representations with the interpretability of keyword-based search. A number of recent works have demonstrated that inverted indexes, when appropriately modified for learned sparse embeddings, achieve state-of-the-art recall-latency tradeoffs for LSR, often outperforming graph-based indexes by substantial margins in latency at a fixed recall target. As a result, the community has generally concluded that graph-based approximate nearest neighbor (ANN) search is inferior to inverted indexes for LSR. This paper challenges this conventional wisdom. Our key insight is that, while applying standard graph-based ANN indexes out-of-the-box to sparse vectors leads to uncompetitive performance, we can overcome these shortcomings by infusing the strengths of inverted indexes into graph-based search. In particular, we introduce three techniques to accelerate graph-based LSR: (i) term-based entry point seeding, (ii) approximate routing via sparse sketching, and (iii) sparse partition-based graph construction. We combine these techniques into a new graph-based ANN system for LSR we call GRIP. On the MS MARCO v2 benchmark (138M passages) with SPLADE embeddings, GRIP achieves a 2.5\times latency improvement over Seismic, a state-of-the-art inverted index tailored for LSR, at 95% recall@10. Ultimately, our results call into question the default choice of inverted indexes for deploying learned sparse retrieval.
[IR-17] When Agent ic Retrieval Fails Silently: Framework and Production Evidence from a Creative AI Assistant
链接: https://arxiv.org/abs/2610.30423
作者: Yan (Fergie)Liang,Anuj Agarwal,Aditya Reddy Cheruku
类目: Information Retrieval (cs.IR)
备注: 5 pages, 1 figure. Submitted to CHIIR 2027 Industry Track
Abstract:CTR, MRR, and behavioral NDCG assume the user sees ranked results and acts on them. In an agentic assistant, a generation layer consumes retrieval output first, and that assumption breaks in three ways. The agent reads the results before the user does (who). The user’s action can come several turns or sessions after retrieval (when). Outcomes blend retrieved content with the model’s own knowledge (what caused it). We call this the attribution gap. We propose Layered Agentic Retrieval Assessment (LARA), which measures retrieval quality, engagement and outcome, and product success as three independent layers joined by a four-field lineage schema that requires no new modeling. We examine the motivation on Adobe Express, where one retrieval infrastructure serves a gallery search surface and an agentic assistant at roughly one-eighth its session volume for the same user population. Chat users reformulate more than gallery users (44.4% vs. 33.1%) because chat reformulation refines generated output. LARA restores attribution for the retrieval.
[IR-18] Data-Void Exploits in Web-Searching LLM s NDSS
链接: https://arxiv.org/abs/2610.30365
作者: Michele Guerra,Wuyeh Jobe,Sanjana Nambiar,Christina Pöpper,Azza Abouzied
类目: Cryptography and Security (cs.CR); Human-Computer Interaction (cs.HC); Information Retrieval (cs.IR)
备注: Extended version of the paper to appear in the Network and Distributed System Security (NDSS) Symposium 2027. 23 pages
Abstract:Web-searching Large Language Models (LLMs) have changed how users seek online information: instead of sifting through ranked lists of links, users read synthesized answers. This shift creates a new LLM vulnerability: a data-void exploit. Disinformers publish their narratives in search spaces where relevant and reliable content is scarce, i.e. data voids, then rely on the LLM’s web-search to retrieve and then use, cite, or list those pages in its responses to data-void seeking prompts. Unlike classical search, where users see ranking cues and competing results, LLMs can obscure when an answer relies on sparse, unverified, or adversarial sources. We study this threat by varying the user’s prompt tone - skeptical, excited, or neutral - and LLM choice - ChatGPT, Gemini, or Sonar Pro - as well as three attacker controls. The attacker can use one of five known strategies to construct and fill a data void with their narrative, host their content at varying levels of visibility from well-established, indexed domains to less established ones, and include generative-optimization engine features thought to boost citing by web-searching LLMs. We find that data-void exploits can succeed with only a single attacker controlled page. Across all conditions, attacker-content appears in LLM responses 31% of the time. User controls offer limited protection: while skeptical prompts slightly reduce success and slightly increase flagging or warnings in LLM responses, we find no evidence of large effects of prompt tone or model choice on success rates, although smaller effects cannot be ruled out. Overall, our results establish data voids as a practical inference-time attack surface for manipulating web-searching LLMs through strategically placed web content and highlight the urgent need for defenses against such exploits.
[IR-19] FAIRGate: Agent ic Runtime Policy Selection and Verified Repair for Fairer Recommendation Delivery
链接: https://arxiv.org/abs/2610.30359
作者: Navkar Jain,Enock Ayiku,Kenneth Fletcher
类目: Information Retrieval (cs.IR)
备注: Accepted at the 1st IEEE International Conference on Responsible Artificial Intelligence (IRAI 2026), Melbourne, Australia. 6 pages, 2 figures, 5 tables
Abstract:Fairness-aware recommender systems often encode one fairness objective during training or apply a fixed re-ranking rule after scoring. In deployment, however, the appropriate fairness action can vary across requests because the candidate pool, item metadata, exposure debt, user preference history, and utility budget all change at runtime. We propose FAIRGate, a model-agnostic delivery controller that places an agentic policy-selection layer between a base ranker and the final top-K list. The agent does not define fairness objectives from scratch. It chooses from a predefined policy library covering long-tail exposure, item-group exposure balance, concentration control, calibration protection, and abstention. A deterministic executor then applies the selected policy through constrained repair, and a validator accepts the result only when relevance, feasibility, metadata, calibration, and audit constraints are satisfied. The final language-model component receives only the verified audit trace and produces evidence-linked narration after the ranking decision is fixed. This design gives the agent runtime responsibility while keeping fairness enforcement measurable, bounded, and replayable.
[IR-20] Protection Is Not One Event: A Survey of RAG Protection Research by Measured Endpoint ACL
链接: https://arxiv.org/abs/2610.30193
作者: Peiyang Liu,Xi Wang,Di Liang,Youru Li,Zhi Cai,Wei Ye
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: ACL Rolling Review (August 2026 cycle) recommends that the paper be accepted at an *ACL conference
Abstract:Retrieval-augmented generation (RAG) systems face security, privacy, and copyright risks including content leakage, source reconstruction, unauthorized access, data poisoning, and prompt injection. Yet RAG protection studies use incompatible evaluation metrics, datasets, and threat models. We present a systematic survey and human-coded evidence map of 118 full-text-confirmed RAG defense papers, organized by measured endpoint rather than claimed goal or defense mechanism. Our seven-family taxonomy covers content disclosure, formal privacy proxies, access control and suppression, corpus-use detection, and pipeline integrity and attribution. Thirty-two annotators double-coded the papers kappa=0.867. The survey identifies three gaps. First, retrieval reranking and context rewriting each span five or six endpoint families, so mechanisms do not identify protection outcomes. Second, only 17.8% of studies use content disclosure or reconstruction as their primary endpoint; under the frozen copyright/ownership relation label, 73.3% of records have corpus-use detection as their primary endpoint, and none have an emitted-expression primary endpoint. Third, most evaluations use general-purpose question-answering datasets and non-adaptive attacks rather than sensitive data and adaptive, multi-turn extraction. Our controlled RAG testbed shows that preventing verbatim copying does not prevent private-fact recovery: a normalized 8-gram guard reduces 20-token copying to 0.0%, but a three-query adaptive attacker recovers 48.3% of hidden facts on Qwen3-8B. We release the taxonomy, dataset, coding protocols, and testbed to enable comparable, claim-aligned RAG security evaluation.
[IR-21] Deep Search in Real-World Web Environments
链接: https://arxiv.org/abs/2610.30181
作者: Xiaomin Yu,Hongliang Chen,Zhe Huang,Shilin Yan,Shao-Rong Su,Xuqiao Ma,Feiyang Wang,Jiayu Chen,Jiwei Fang,Yu Yan,Yini Huang,Chen Chen,Ruoxi Cheng,Yuzhi He,Yuhui Zhang,Xiaoxing Hu,Chen Liu,Jianbo Lin,Yijiang Li,Xiaobin Hu,Menglong Li,Zhuoran Lu,Weishi Wang,Zhongqi Yue,Chengwei Qin
类目: Information Retrieval (cs.IR); Multimedia (cs.MM); Software Engineering (cs.SE)
备注:
Abstract:Despite substantial progress in Deep Search agents, their ability to conduct end-to-end research in real-world open Web environments remains insufficiently evaluated. Unlike static retrieval settings, information on the real-world Web is distributed across diverse sources, modalities, and heterogeneous formats, often involving complex interaction mechanisms that require agents to autonomously search for, access, and integrate relevant evidence during execution. To address this gap, we introduce the OpenDS task, designed to evaluate agents’ ability to perform multi-step deep search in the live, open Web. This task requires agents to autonomously explore the real Web, discover and connect relevant evidence from multiple independent sources through Web search, browser interaction, and evidence processing, and ultimately determine a unique factual answer. Building on this task formulation, we construct OpenDS-World, a benchmark comprising 200 challenging tasks spanning diverse information modalities, heterogeneous file formats, and interactive access requirements, with independent human verification to ensure evidence validity and answer uniqueness. We evaluate 12 frontier models under a unified tool environment and execution protocol. The best-performing model achieves 39.0% accuracy, while human evaluation across 1,000 independent attempts yields 21.1% accuracy. Further analyses of tool usage, search cost, termination behavior, and failure recovery reveal substantial differences in research strategies across models, as well as persistent challenges in tool coordination and long-horizon search management. OpenDS-World provides a unified evaluation platform for systematically assessing and advancing agents’ autonomous deep search capabilities in real-world open Web environments.
[IR-22] QProtFuse: Query-Guided Layer Fusion for Retrieval-Augmented Protein Function Prediction
链接: https://arxiv.org/abs/2610.30180
作者: Mianzhi Pan,Peishuo Liu,Yawen Ouyang,Botian Wang,Yiming Rong,Hao Zhou,Jianbing Zhang
类目: Information Retrieval (cs.IR); Quantitative Methods (q-bio.QM)
备注:
Abstract:Predicting protein functions from amino acid sequences is essential to understanding biological processes. Large language models (LLMs) offer flexible function descriptions to natural-language queries, yet their limited protein understanding motivates retrieval-augmented generation (RAG). Existing protein RAG methods primarily expand external retrieval signals while overlooking the rich knowledge encoded within a single protein language model (PLM). We show that retrieval performance varies across PLM layers and biological tasks, making conventional final-layer selection suboptimal. To this end, we propose QProtFuse, which uses queries to adaptively fuse multi-layer protein features and jointly learns protein and annotation retrieval through intra-modal and cross-modal contrastive objectives. On Prot-Inst-OOD, QProtFuse achieves the best protein and annotation retrieval scores, reaching 16.4% Recall@1 with annotation reranking. These gains improve downstream RAG-based function descriptions, yielding the highest LLM-as-a-judge scores for both Qwen3-8B (6.22/10) and DeepSeek-V4-Pro (6.68/10). Ablations and weight analyses demonstrate the value of query-guided fusion and complementary training objectives for exploiting a PLM’s internal knowledge.
[IR-23] Association between Player Perception and AI Disclosure on Steam: An Exploratory Study
链接: https://arxiv.org/abs/2610.30178
作者: Fabian Wiesinger,Johanna Pirker,Denis Helic
类目: Information Retrieval (cs.IR); Human-Computer Interaction (cs.HC)
备注:
Abstract:The use of generative AI in game development is growing rapidly and is the subject of intense debate in which player reactions play a central role. This paper contributes to this discussion by analyzing Steam reviews to examine whether games with AI disclosure are associated with different review patterns. Using a dataset across multiple disclosure groups, we apply linguistic analysis including keyword-based, word frequency, and characteristic word analysis to investigate the AI-related discussion on Steam. Our results show that AI-disclosed games are associated with higher levels of AI-related criticism in negative reviews. Our findings provide a first exploratory step towards understanding how players perceive AI disclosure.
[IR-24] FlexVDR: Learning to Allocate Matryoshka Representation Budgets for Visual Document Retrieval
链接: https://arxiv.org/abs/2610.30177
作者: Haowen Xiang,Yibo Yan,Xin Zou,Shuliang Liu,Yi Cao,Mingdong Ou,Xuming Hu
类目: Information Retrieval (cs.IR)
备注: 21 pages, 5 figures
Abstract:Visual document retrieval relies on rich multi-vector representations to capture fine-grained evidence in document pages, but encoding and matching these representations can be computationally expensive. Existing compression and elastic representation methods expose smaller configurations, yet making multiple budgets available does not determine which configuration a query should use. The allocation decision must estimate the retrieval value of alternative representations before incurring their full computation, while accommodating different resource preferences. We introduce FlexVDR, a framework for query-adaptive budget allocation within a single retriever. We construct a joint family of routes over encoder depth, embedding width, and token count. Retrieval supervision and distillation train the compact routes, while a native bypass preserves the full-budget reference. A Budget Allocator predicts route-specific retrieval utility from early query features and combines these predictions with a configurable resource preference to select a route. Experiments across five backbones on ViDoRe V1, V2, and V3 public show that FlexVDR retains much of the native retrieval quality while selecting compact routes for individual queries. A configurable resource preference allows the same retriever to adjust its computation without retraining or switching models.
[IR-25] FormBench: A Benchmark for Fine-Grained Retrieval in Formulation Patents
链接: https://arxiv.org/abs/2610.30175
作者: Vineeth Venugopal,Elsa Olivetti,Jonathan Welch
类目: Information Retrieval (cs.IR)
备注: 28 pages, 9 figures. Dataset: this https URL ; code: this https URL
Abstract:We introduce FormBench, a platform for testing the ability of models to retrieve precise chemical information from technical documents. FormBench contains 55,352 queries and two retrieval corpus variants each with ~1M passages: FormBench-Random (C0), with randomly sampled distractors, and FormBench-Structured (C1), with structurally proximate distractors. Each query is paired with an anchor passage, a hard negative, and a soft negative - derived from a materials ontology through entity overlap scoring, providing a deeper graded signal on a model’s discriminatory power. We evaluate 10 model configurations spanning commercial API encoders (OpenAI, Voyage), open-source encoders (Nomic, GTE-large, Qwen3-Embed), and their task-adaptive fine-tuned (TAPT) variants. We find: (1) commercial APIs score very poorly across all configurations (OpenAI-3-large: 0.009 on C1, 0.213 on C0); (2) TAPT on FormBench triplets lifts Qwen3-Embed to 0.4085 on C1, 0.4835 on C0; (3) BM25 remains competitive (0.375 on C1, 0.467 on C0), exceeded only by TAPT-Qwen3; (4) corpus size has a significant impact on retrieval behavior, as demonstrated through a corpus-size ablation; (5) the API failure traces to within-domain geometric collapse; (6) APIs outperform even the TAPT variants on classification and clustering tasks. FormBench is released in BEIR-compatible format at this http URL under CC-BY-4.0.
[IR-26] IRAH: Hierarchical Agent ic Tree Search for Retrieval Pipeline Optimization
链接: https://arxiv.org/abs/2610.30167
作者: Maria Ana Cardei,Aniket Singh,Amir Moghadam,Srividhya Baikadi,Rashmi Shetty
类目: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Under review
Abstract:Optimizing Information Retrieval (IR) pipelines for Retrieval-Augmented Generation (RAG) requires jointly selecting structural components and numerical hyperparameters under limited evaluation budgets. We introduce the Information Retrieval Agentic Harness (IRAH), which combines Language Agent Tree Search for IR (LATS-IR) with offline meta-learned initialization. LATS-IR performs UCT-guided hierarchical search, first exploring macro-structural pipeline choices and then refining numerical parameters, with natural-language reflections to guide subsequent mutations and provide inspectable search traces. We evaluate IRAH on three in-distribution benchmarks and one out of-distribution benchmark to assess generalization under distribution shift. Under the same meta-learned initialization, LATS-IR outperforms TPE on two of three in-distribution benchmarks, with gains of up to 6.4%, and remains competitive on the out-of-distribution benchmark. Across all four benchmarks, warm-started IRAH achieves the highest average held-out nDCG@10 (0.4816) and improves over its cold-start counterpart. These results establish IRAH as a competitive approach to retrieval pipeline optimization under constrained evaluation budgets.
人机交互
[HC-0] Who Is Sycophantic to Whom? Confident LLM s Human Deference and the Collapse of Trust Calibration
链接: https://arxiv.org/abs/2610.31579
作者: Deepthika S,Pawni Yadav,B. Sankar
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: 49 pages; 32 figures; 7 tables
Abstract:Research on sycophancy asks whether large language models (LLMs) bend towards what their users want to hear. We ask the reverse question: when an LLM is confident, consistent, and wrong, do users bend towards it? In a within-subjects experiment, 32 participants learned up to five science topics by chatting with five assistants built on the same model (GPT-5-mini). Unknown to participants, each assistant’s briefing had 0, 25, 50, 75, or 100% of its facts replaced by plausible falsehoods, and every assistant spoke with full confidence. A surprise quiz after each topic offered the correct fact, the planted falsehood, and two distractors, with a confidence rating for every answer. Across 104 quizzes and 1,040 answers, scores fell from 7.4 to 2.9 out of 10, and the planted option was chosen on 60% of falsified items against 11% of correctly taught ones (odds ratio 12.5). Transcript coding showed that whenever the assistant stated a falsehood, it was adopted about 65% of the time at every falsification level. Absorbed falsehoods were held with high confidence (8.4/10), the modeled confidence advantage of right over wrong answers vanished, and the discrimination of confidence for correctness fell from an AUROC of 0.70 to chance, while post-task trust ratings barely moved. Participants tended to return to the chat more when the assistant was worse, but returning went with more absorption, not less, and 30% of all turns pasted quiz questions back to the source being checked. The assistant itself was almost never sycophantic and never conceded under pushback. We introduce a baseline-corrected Human Deference Index to quantify this human-side sycophancy and argue that trustworthiness is a property of the human-AI pair, that sycophancy evaluations should measure both directions, and that learning interfaces must make reliability legible claim by claim.
[HC-1] Supporting or Replacing Students Reasoning ? Cognitive Offloading in AI-Assisted Proof Learning
链接: https://arxiv.org/abs/2610.31481
作者: Eason Chen,Sophia Judicke,Kayla Beigh,Xinyi Tang,Isabel Wang,Nina Yuan,Zimo Xiao,Chuangji Li,Shizhuo Li,Reed Luttmer,Shreya Singh,Maria Yampolsky,Naman Parikh,Yvonne Zhao,Meiyi Chen,Scarlett Huang,Anishka Mohanty,Gregory Johnson,John Mackey,Jionghao Lin,Ken Koedinger
类目: Human-Computer Interaction (cs.HC)
备注: 14 pages. This manuscript presents a substantially revised and extended analysis of the classroom study reported in arXiv:2602.18807 and the earlier preprint arXiv:2509.16778
Abstract:We studied 153 students in an undergraduate proof-based mathematics course with staggered-access to an AI support system. Students were randomly assigned to receive an AI chatbot and a homework-feedback interface earlier or later in the course. During the period when only the early-access group had access, their homework scores were 2.56 points higher, but exam performance did not differ significantly between groups, suggesting that improved AI-supported homework performance did not translate into detectable gains on a subsequent exam. After all students received access, every additional 100 messages sent to the AI chatbot was associated with a 3.02-point lower score on the subsequent exam. A mediation analysis showed a significant indirect association linking lower baseline self-efficacy to greater chatbot use and, in turn, lower exam scores: a one-point decrease in self-efficacy corresponded to 0.62 fewer exam points through chatbot use. These findings raise concerns that students may use the chatbot to offload their cognitive reasoning. Use of the homework-feedback interface, which required students to submit their proof attempts before receiving feedback, was not negatively associated with exam performance. We recommend against using a chatbot for AI tutoring and suggest exploring a tutor embedded in the course activities.
[HC-2] How Different or Different How? Stance Granularity in Viewpoint Selection for Dialog Systems that Engage Users with Alternative Views
链接: https://arxiv.org/abs/2610.31459
作者: Morita Tarvirdians,Reeve Lorena,Catharine Oertel
类目: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL)
备注:
Abstract:Systems that aim to help a user engage with viewpoints other than their own have to decide how different those viewpoints should be. Viewpoints are mainly represented on a scalar scale, too coarse to distinguish viewpoints that disagree with the user about different things. We introduce a selection policy in which a bandit sets how far viewpoints may sit from the user, and the policy selects the most novel viewpoint within that bound. We vary only its stance encoder, comparing a scalar score with Perspectivized Stance Vectors, which break a position into separate stances on concepts. We integrate it inside a dialog system and run a between-subjects study (N=63). Concept-level stance produces more varied exposure sequences and broader coverage of the topic’s positions, but does not change how deeply users justify viewpoints or how accurately they reconstruct them. Finally, we analyze user–viewpoint relations at the concept level, showing that users reconstruct a viewpoint least accurately on concepts that one side took no position on, rather than on concepts where the two disagreed. Our findings open up opportunities for dialog systems that support engagement with alternative viewpoints by acting on which concepts a user and a viewpoint disagree on, rather than on distance alone.
[HC-3] In-Browser Human-AI Co-Writing for Counterspeech: An Exploratory User Study of Agency Perceived Response Quality and Time
链接: https://arxiv.org/abs/2610.31357
作者: Tammy Schmidt,Luise Koch
类目: ocial and Information Networks (cs.SI); Human-Computer Interaction (cs.HC)
备注: Pre-Publishing for the 10th Workshop on Online Abuse and Harm 2026
Abstract:Writing counterspeech that challenges online hate speech can be cognitively and emotionally demanding, whereas fully automated generation may weaken users’ sense of voice and responsibility. We present an in-browser AI-assisted co-writing tool that enables users to contribute their own ideas and offers three editable suggestions generated by a large language model using dynamically retrieved hate speech-counterspeech examples from the Multitarget CONAN dataset. We evaluated the tool in an exploratory within-subject study with ten participants, complemented by think-aloud observations and semi-structured interviews with half of the participants. Participants wrote responses to three self-selected hate speech instances without assistance and then repeated the task with the tool. Final responses were subsequently peer-rated for empathy, non-toxicity, relevance, specificity, and persuasiveness. Participants evaluated the interface positively, and qualitative accounts linked user input, selection among alternatives, and small edits to perceived agency and responsibility. Assisted responses received higher descriptive mean ratings across all five proxies, with the largest difference for empathy. Mean composition time increased from 2.52 to 3.88 minutes. Qualitative findings further revealed a tension between empathy-oriented outputs and users’ desire to express legitimate anger or direct disagreement. We contribute exploratory design guidance for treating AI-assisted counterspeech as deliberative co-writing rather than automated text production.
[HC-4] Making Generative AI Workable at Work: Accountability Localization and Representation in Kenyan and Nigerian Small Businesses
链接: https://arxiv.org/abs/2610.31344
作者: Najeeb Gambo Abdulhamid,Stephanie Nyairo,Liz Ankrah,Jacki O’Neill
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:Generative AI is widely framed as democratizing skilled work by lowering the barrier to producing professional output. Drawing on 21 semi-structured interviews with 27 workers across 21 small and medium-sized businesses (SMBs) in Kenya and Nigeria (July to October 2023), we find that availability alone did not make participation workable. In participants’ accounts, drafting was rarely the hard part. The consequential work was making AI-assisted drafts defensible enough to circulate to clients. We identify five intertwined practices through which participants made generative AI workable: using it as draft labour, steering outputs through iterative interaction, switching and comparing across tools, performing release work to make outputs accountable before circulation, and repairing context misfit when outputs failed local requirements. We make two central contributions. First, we specify release work as a distinct, boundary-facing form of accountability labour, evidenced through verification gates, traceability reconstruction when sources fail, and non-delegation boundaries that decide what can ship. Second, we show how localization and representational failures, which we call context misfit, operate in everyday work as a recurring breakdown-and-repair class, including failures of local language, institutional knowledge, and the credible representation of African people and places, and we document the repair repertoire they demanded. Reframing democratization as workable participation, we argue that reported gains were conditional on this accountability and repair labour, and we translate the findings into four design directions for workplace generative AI. We treat Kenya and Nigeria as two situated cases, not as a representative sample of SMBs or the Global South.
[HC-5] CoWriter: Guiding Methodological and Synergetic Human-LLM Co-Authoring
链接: https://arxiv.org/abs/2610.31277
作者: Shmuel Amar,Benjamin Lieber,Ori Shapira,Debanjana Kar,Sebastian Maximilian Dennerlein,Yufang Hou,Ido Dagan
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:While LLMs have transformed text generation, standard chat-and-canvas interfaces often reduce human authoring to quite passive supervision, causing review fatigue, over-reliance on AI, and the sidelining of the human role in long-form, source-grounded writing. Prior systems typically address isolated writing phases, leaving comprehensive end-to-end human-LLM collaboration largely unexplored. In this paper, we introduce CoWriter, a proactive agentic scheme that adapts expert human (co-)authoring methodologies from cognitive science to flexibly steer the entire writing workflow while actively encouraging the user to engage. In an exploratory user study, compared to a standard LLM baseline, authors using CoWriter better expressed their perspectives and maintained a higher sense of document ownership, supported quantitatively by 2.2\times more successful agent engagements. Furthermore, external reader judges rated the resulting texts higher in “human touch” qualities like storytelling and integrating unique personal experiences. Overall, our results demonstrate that proactively engaging users through established human practices is beneficial for effective LLM-supported co-authoring, providing useful methodologies, techniques and insights for future research on collaborative agents.
[HC-6] PointerBench: A Diagnostic Benchmark for Fine-Grained GUI Grounding NEURIPS2026
链接: https://arxiv.org/abs/2610.31274
作者: Valentino Sacco,Saurabh Pandey,Alexander Florian Jercher,Maximilian Schilling
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: Accepted at the 2nd Workshop on Agentic AI Benchmarks and Applications for Enterprise Tasks (AABA4ET), NeurIPS 2026
Abstract:Multimodal Large Language Models have shown promising performance in computer-use tasks, from planning user workflows to grounding natural-language actions to precise UI targets. Existing end-to-end benchmarks primarily emphasize long-horizon task completion, while GUI grounding benchmarks remain focused on conventional controls and a limited set of professional interfaces. We introduce POINTERBENCH, a diagnostic suite of 1,500 screenshot-instruction pairs probing three complementary capabilities: thin sub-element identification and geometric grounding in spreadsheets (POINTERBENCH-Sheets), character-level grounding in long-text documents (POINTERBENCH-Text), and mixed GUI grounding across 100 professional applications (POINTERBENCH-Pro). We carefully curate each task with pixel-exact target geometry and evaluate both general-purpose multimodal models and 39 specialist GUI grounding configurations. We find text grounding to be the most challenging setting, with Claude Fable 5, GPT-5.6 Sol, and the leading native specialist UI-Venus-2-9B achieving 69.6%, 53.4%, and 25.8% success on POINTERBENCH-Text despite respective 83.2%, 77.7%, and 54.5% average success across POINTERBENCH. These results reveal substantial headroom in fine-grained GUI grounding across frontier and specialist models. We publicly release the benchmark and evaluation code at this https URL.
[HC-7] Navigating human-robot interaction challenges in private and residential aged care homes
链接: https://arxiv.org/abs/2610.31149
作者: David Ahmedt-Aristizabal,Mohammad Ali Armin,Akila Pemasiri,Harshala Gammulle,Tharindu Fernando,David Silvera-Tawil,Simon Denman
类目: Human-Computer Interaction (cs.HC); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: Author-accepted manuscript of a book chapter published online in Elsevier’s Reference Module in Materials Science and Materials Engineering (2026). Scheduled for inclusion in Comprehensive Robotics for Extreme and Challenging Environments (June 2027)
Abstract:This chapter explores how robotic systems and ambient intelligence can enhance care for older adults in residential aged care (RAC) homes and private homes. Although these settings differ, we frame both as extreme and challenging environments for human-robot interaction (HRI), characterised by cluttered and dynamic layouts, diverse resident needs, and limited on-site technical support. We review embodied and non-embodied systems that assist with daily living, safety monitoring, and social engagement, highlighting hybrid deployments where robots complement fixed and wearable sensors. To ensure these technologies are usable and effective, design considerations consider the cognitive, physical, and social needs of older adults. Interaction design prioritises accessibility for older adults with low digital literacy, using robust but simple voice interfaces, small sets of non-verbal cues, accessible screen-based companions, and safe object handovers. Privacy, trust, and acceptability are addressed via practical strategies such as on-device inference, data minimisation, visible consent controls, and transparent data handling aligned with emerging artificial intelligence (AI) and privacy regulation. Evaluation emphasises real-world metrics including false alarms per day, time to recover from faults, staff workload, and long-term system reliability. We close by outlining open research challenges and future directions for scalable, regulation-ready deployments that support aging with dignity, safety, and independence.
[HC-8] Adaptive Uncertainty Visualization With jk-Plots
链接: https://arxiv.org/abs/2610.31050
作者: Nikolaus Piccolotto,Daniel Pahr,Laura Lotteraner,Fatih Öztank,Markus Bögl
类目: Human-Computer Interaction (cs.HC)
备注: 5 pages
Abstract:An important task of uncertainty visualizations is to communicate how the number of available observations influences the precision of estimates, such as a Normal distribution’s mean. In this paper, we propose an adaptive approach to visualizing univariate data that produces salient representations while highlighting the uncertainty introduced by the sample size. We introduce jk-plots, inspired by educational material of Bayesian statistics, that combine visualizations of plausible reference models (emulating popular uncertainty visualization idioms) and the observations. We describe a visualization pipeline to create jk-plots, which we implemented in an R package. We demonstrate jk-plots using synthetic datasets of varying size and discuss possible usage scenarios in domain expert interviews.
[HC-9] CARL: Chat Assisted Reinforcement Learning
链接: https://arxiv.org/abs/2610.31014
作者: Sanjiv Das,Harshad Khadilkar,Sukrit Mittal,Daniel Ostrov,Jacob Sisk
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Failure of applied Reinforcement Learning (RL) to solve a given problem predominantly stems from unaddressed specification gaps, where the intent of human experts in the problem’s domain is inaccurately or incompletely translated by AI into mathematical objectives, constraints, and reward functions. Failure can also stem from unaddressed performance gaps, where human domain experts are unable to determine the validity of the RL based solution or where parameters (i.e., hyperparameters or weights on the reward terms comprising the reward) are poorly chosen. Chat Assisted Reinforcement Learning (CARL) bridges these gaps through a human-AI co-design interface architecture that enforces interactive specification prior to code generation and then enables subsequent verification and parameter alterations via iteration. More specifically, CARL leads domain experts through a structured ten-step workflow for each iteration using step-specific skill files, domain-aware suggestions via an assistant persona (called Trillian), and compact decision summaries that maintain context across iterations without context overload. Once the domain expert confirms the Markov decision process (MDP) specified by this procedure, CARL’s autonomous tool-use loop generates, tests, executes, and debugs an end-to-end RL training pipeline, followed by an interactive post-training diagnosis and revisit step that allows for verification and parameter alteration before running a new iteration if desired.
[HC-10] Metacognitive Scaffolds Create Productive Friction for AI-Assisted Knowledge Work
链接: https://arxiv.org/abs/2610.30945
作者: Viktor Kewenig,Lev Tankelevitch,Sean Rintel,Advait Sarkar,Abigail Sellen
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注: 46 pages, 11 figures, 17 tables
Abstract:Generative AI imposes new metacognitive demands on knowledge workers as their work shifts from production to planning, evaluation, and orchestration. In a randomized, between-subjects experiment on AI-assisted writing (297 workers, 594 essays), we test dynamically-generated metacognitive scaffolds supporting goal specification and task decomposition. We find that scaffolds improve writing quality, with this boost persisting in short-lag transfer when scaffolds are removed. We also find evidence of improved topic recall three days later. These effects are consistent with observed changes in how people learn to prompt AI systems after exposure to scaffolding. However, scaffolded participants rate the AI as less helpful and more cognitively demanding even as they perform better, suggesting that subjective ease can understate the value of support. Our findings make the case for metacognitive scaffolds in AI-assisted knowledge work, revealing a productive friction worth designing for as workers increasingly delegate work to agentic AI.
[HC-11] Your Home Remembers: A Voice-Enabled Household Memory Assistant for Conversational Retrieval
链接: https://arxiv.org/abs/2610.30895
作者: Aleix Sant,Jordi Luque,Fernando López,Pablo Gómez,Carlos Escolano
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Everyday household information is often exchanged in conversations but never written down, making details easy to forget or misremember. We present a voice-enabled smart-home assistant that acts as a household memory by recording conversations and retrieving information from stored and newly captured speech. The assistant runs on an Android tablet with a modular architecture in which speech-recognition, diarisation, embedding and language models can be configured or replaced through an orchestration layer. On-device wake-word detection and VAD-based query capture support hands-free interaction. Recorded audio is processed in configurable chunks, transcribed and diarised, converted into turn-aware passages with conversation-level timestamps, and indexed in a vector store. For memory queries, a retrieval-augmented generation pipeline selects key passages and uses them to generate an answer. The interface supports recording, conversation management and inspection of timestamped sources.
[HC-12] Working Through Representation: Malleable Workspaces for Human-Agent Knowledge Work
链接: https://arxiv.org/abs/2610.30893
作者: Nan Chen,Long Ling,Luna K. Qiu,Zilong Wang,Yuqing Yang
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Generative AI supports knowledge work, yet conversational histories may not capture the evolving organization that guides subsequent action. We introduce the Malleable Workspace, a conceptual model of how humans and agents form and revise a Work Representation and use its organization to advance work within a wider Work Environment. We developed it iteratively through conceptual analysis, comparison of 58 papers, and the construction and three-month longitudinal study of MalleableSpace with seven experienced agent users. Systems varied in what they represented as units and relations. System-assembled context was common; more systems supported local incorporation of agent outputs than reorganization, and fewer supported consequential agent-driven changes beyond the representation. The field study showed how representational roles and organization developed in use. These observations refined the model’s account of how representation-grounded action depends on and can change the Work Environment. We contribute a conceptual and analytic foundation for representation-centered human–agent design.
[HC-13] utorCaseBench: A Case-Grounded Benchmark for Evaluating LLM Tutors in Supporting Students with Challenging Behaviors
链接: https://arxiv.org/abs/2610.30828
作者: Yeon Su Park,Chanyoung Park,Tae Soo Kim,Seungju Kim,Hyunwoo Kim,Daho Jung,Jinsung Yoon,Juho Kim
类目: Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注: Under Review
Abstract:LLMs are increasingly used as AI tutors, yet existing benchmarks overlook the challenging behaviors common among students with learning and behavioral difficulties. We introduce TutorCaseBench, a benchmark that grounds AI-tutor evaluation in qualitative case studies from pedagogy, educational psychology, and special education. Our pipeline converts documented student categories into case-specific scenarios, tutor-student dialogues, and rubrics grounded in source evidence. The benchmark contains 923 dialogues across five challenge types, evaluated through open-ended response generation (1,560 items) and multiple-choice selection (2,537 items). Across 11 state-of-the-art LLMs, even the strongest model satisfies a rubric criterion in only 55.8% of open-ended cases, while models are better at selecting appropriate strategies than generating situated responses. These results reveal a gap between intervention recognition and context-specific adaptation, especially for open-weight models. All data are publicly available at this https URL.
[HC-14] Abstract Priorities Concrete Rewards: How Stated Preferences Relate to What People Reward in AI Responses
链接: https://arxiv.org/abs/2610.30763
作者: Zihan Chen,Lei Nico Zheng,Di Zhu
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 27 pages, 5 figures, 30 tables. Code and data are available from the authors upon request
Abstract:AI assistants built on large language models (LLMs) are trained on human feedback and increasingly personalized to users’ stated preferences. Yet how closely these preferences reflect what people reward in responses remains unclear. We compare stated and revealed preferences in the public PRISM dataset, in which 1,396 participants reported how much seven response attributes mattered to them and then rated 29,347 responses from 21 LLMs. Among responses to the same prompt, features the survey never asked about, such as response length, formatting and refusal language, were at least as strongly associated with ratings as text proxies of the surveyed attributes. Whether stated priorities matched ratings depended on how attributes were measured: participants who prioritized an attribute favoured responses from models that others perceived as strong on it, but not responses that scored higher on text proxies for it. Written feedback was more telling: participants who asked for shorter answers, even in earlier conversations, gave length less weight in their ratings, and those asking for longer answers gave it more. Their relative priorities over the surveyed attributes, however, did not predict this weight. Reasons given for choices reflected perceived attributes more than measured features. Participants also differed stably in how much they weighted several concrete features, especially length, but with a median of 6 conversations per participant, these weights were estimated poorly. The largest gain from personal information in predicting held-out choices was 0.13 percentage points, compared with 1.11 for a stronger pooled model. Stated priorities thus track perceived attributes of AI responses but say little about the concrete features that ratings reward, a distinction that matters for preference elicitation and AI personalization.
[HC-15] Can Loan Processors and Underwriters Spot AI-Edited Statements? An Empirical Study
链接: https://arxiv.org/abs/2610.30758
作者: Simiao Ren,Xingyu Shen,Tommy Duong,Ankit Raj
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Human-Computer Interaction (cs.HC); Software Engineering (cs.SE)
备注: 6 pages, 2 figures
Abstract:AI-assisted editing can produce loan documents that preserve familiar issuer layouts while changing the applicant’s identity. We investigated whether experienced underwriters and loan processors identify such forgeries in first-pass, document-only review. Three practitioners recruited through Upwork each received one applicant package drawn from 134 forged PDFs derived from 47 genuine financial documents supplied with full consent. The edits preserved amounts, dates and balances; identifier inconsistencies introduced by the pipeline were knowingly retained before distribution. Reviewers were not told that files had been altered and returned 132 file-level judgements. Checking their explanations against the originals yielded 14 edit-supported concerns (10.6%): 12 Suspicious judgements and two Insufficient information judgements. All supported concerns involved identifier disagreements within or across documents, and no reported visual cue identified an actual edit. Thus, approximately one in ten judgements contained evidence of the forgery, despite known inconsistencies remaining in the files. These observations motivate evaluating consistency checks and independent source verification alongside human review. With three reviewer-package pairs and no genuine-file controls, the findings describe this exploratory setting rather than population-wide detection accuracy.
[HC-16] Do We Need to “Be There” to Say Goodbye? Comparing Participatory Agency and Spatial Presence Across Digital Farewell Formats
链接: https://arxiv.org/abs/2610.30734
作者: Chan In Sio,Alex Mann,Qian Ming Jia,Lik-Hang Lee
类目: Human-Computer Interaction (cs.HC)
备注: 10 pages, 4 figures. Accepted for publication in IEEE Transactions on Visualization and Computer Graphics (ISMAR 2026). Author accepted manuscript
Abstract:Digitally mediated attachment is increasingly common, yet the mechanisms of symbolic closure in hybrid physical–digital relationships remain underexplored. We address an XR design question: how different mediated farewell formats relate to perceived symbolic closure, spatial presence, and bereavement-related emotional intensity. Using “cloud petting” as a controlled proxy of digitally mediated attachment, we compare three farewell formats: (1) lightweight social messaging, (2) a revisitable web memorial, and (3) immersive VR ritual participation. These formats differ in symbolic participation, spatial presence, and sensorimotor ritual enactment. In a mixed-methods study ( N=45 ), Website and VR showed significantly higher perceived symbolic closure than Messaging, while the present sample did not detect a significant closure advantage for VR over Website. VR was also associated with significantly stronger grief-related responses on the Pet Bereavement Questionnaire, whereas negative-affect change measured by the Positive and Negative Affect Schedule did not differ significantly across modalities. This pattern motivates a preliminary, exploratory, mechanism-level Agency-Amplification account: formats supporting structured farewell participation were associated with higher perceived closure, whereas the VR format, combining technical immersion, spatial presence, and sensorimotor ritual enactment, was associated with stronger loss-oriented emotional processing. Because the formats also bundled differences in participation, revisitation, technical immersion, and spatial presence, the study provides preliminary design guidance without isolating the causal contribution of any single feature.
[HC-17] Big World Little Me: Comparing a First-Person VR Walkthrough Video and Static Slides for Communicating Height-Related Accessibility Barriers
链接: https://arxiv.org/abs/2610.30731
作者: Chan In Sio,Alex Mann,Qian Ming Jia,Xuecheng Guo,Yan Zhuo,Xinyu Ning,Kaikuo Liao,Lik-Hang Lee
类目: Human-Computer Interaction (cs.HC)
备注: 14 pages, 2 figures. Accepted for presentation and publication at ICVRT 2026. Author manuscript
Abstract:People with dwarfism and other forms of short stature may encounter barriers created by fixed-height environments and discriminatory social assumptions. This paper evaluates Big World, Little Me, a 10 min 52 s first-person walkthrough recorded from a Unity-based VR environment. The narrative depicts domestic reachability, public-transport, and employment barriers. In a non-randomised between-subjects study, 100 participants aged 15–48 years viewed either the fixed-duration audiovisual walkthrough or a silent, self-paced 27-slide presentation representing the same principal events. Narrative engagement was higher for the walkthrough ( M=5.54 , SD=0.51 ) than for the slides ( M=4.24 , SD=0.73 ), mean difference =1.30 , 95% CI [1.05, 1.55], Hedges’ g=2.05 , p.001 . State empathy was also higher for the walkthrough ( M=3.15 , SD=0.33 ) than for the slides ( M=2.21 , SD=0.60 ), mean difference =0.94 , 95% CI [0.74, 1.13], Hedges’ g=1.91 , p.001 . Exploratory supportive-intention ratings increased more in the walkthrough condition than in the slide condition. Participants also rated the walkthrough as more dynamic, continuous, spatially clear, and information-rich. The findings indicate stronger immediate responses to the complete walkthrough format, but do not isolate the effects of motion, continuity, pacing, information quantity, or participant control. The walkthrough should be understood as screen-based VR-authored media rather than immersive VR or a reproduction of lived experience.
[HC-18] PATH: Preference-Aware Traversability Estimation for Robotic Teleoperation in Off-Road Navigation
链接: https://arxiv.org/abs/2610.30605
作者: Xianghao Meng,Wenshuo Wang,Cheng Shao,Lewei Feng,Qingxiao Liu,Canyu Wang,Junqiang Xi
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC)
备注:
Abstract:Traversability estimation is critical for ground vehicle teleoperation, yet existing methods often neglect the human operator’s preferences for speed and risk tolerance. To address this gap, we propose PATH-Net, a Preference-Aware Traversability estimation Hierarchical Network that adaptively generates preference-conditioned traversability cost maps and speed profiles for off-road teleoperation. Our framework integrates visual and geometric bird’s-eye-view features with short-term historical vehicle states through a multi-head shared-backbone architecture. Our PATH-Net infers traversability cost maps while optimizing speed recommendations by using maximum entropy inverse reinforcement learning and sparse speed supervision. Our method incorporates operator preference via an asymmetric feature aggregation mechanism, splitting inference into a cost branch that prioritizes high-risk traversability regions and a speed branch that outputs preference-adjusted recommended speeds. Experiments with 15 operators across 225 teleoperated trajectories show alignment with user-specific navigation strategies, achieving modified Hausdorff distance of 0.17 m (aggressive) and 0.15 m (conservative). In the questionnaire assessment, participants reported lower NASA-TLX workload scores with PATH-Net recommendations than without recommendations, decreasing from 34.7 to 22.2. These results support preference-conditioned trajectory and speed recommendation for off-road teleoperation.
[HC-19] Impacts of generative AI skill development on academic adaptation and learning outcomes among university students
链接: https://arxiv.org/abs/2610.30599
作者: Lei Fan,Fangxue Liu
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:Understanding how university students perceive the development of their generative artificial intelligence (GenAI) skills over time can clarify how GenAI use relates to learning and academic outcomes. This study examined perceived skill-development trajectories, changes in learning practices, and associated academic gains using a retrospective questionnaire with 313 analyzed responses and interviews with 18 students across subject areas. Descriptive and comparative analyses and exploratory structural equation modeling (SEM) were integrated with thematic analysis. Perceived developmental trajectories were heterogeneous: 34.2% of students reported rapid early improvement followed by a plateau, 28.4% steady improvement, and 21.1% delayed acceleration. Development was also uneven across skills, with improvement in evaluating GenAI outputs perceived as more challenging; assignments were the most frequently selected driver of skill growth. Interviews further indicated increasingly specific prompting, tailored verification, and selective delegation, although some students reported stable routines, dependence, skipped reasoning, or improvements attributable partly to better GenAI tools. Academic gains were reported more often for assignments than for examinations. The SEM indicated a positive indirect association between perceived skill growth and academic gains through adaptation of assignment processes, whereas academic gains did not clearly differ across developmental trajectory groups after multiplicity correction. The findings suggest that perceived GenAI skill development varies in both trajectory and skill domain and is associated with changes in learning practices and differentiated academic benefits.
[HC-20] Beyond Better Ideas: How Human-AI Collaboration Shapes Scientific Ideas and Human Perceptions
链接: https://arxiv.org/abs/2610.30568
作者: Qian Chen,Saatvik Pradhan,Janice Ahn,Berk Atil,Angkai Li,Shijie Lu,Nobin Sarwar,Michael Ulis,Ruyuan Wan,Wenpeng Yin
类目: Human-Computer Interaction (cs.HC)
备注: 23 pages, 10 figures, 9 tables
Abstract:As AI increasingly becomes a collaborator in scientific research, its impact may extend beyond improving ideas to reshaping how they are evaluated, attributed, and perceived by humans. Yet existing research has largely examined these dimensions in isolation. We present a unified empirical study of Human-AI scientific ideation, investigating two complementary perspectives: i) how AI collaboration shapes scientific ideas and their evaluation, and ii) how firsthand collaboration reshapes human perceptions of AI. Through a shared experimental setting involving Human-AI and Human-Human research teams, human and LLM evaluators, collaboration-provenance analysis, and participant surveys, we uncover several counterintuitive findings. First, AI collaboration shows no significant advantage in human-evaluated idea quality, while iterative revision consistently improves quality. Second, AI judges can approximate human scores yet disagree on idea rankings; notably, an AI judge favors AI-assisted proposals without a corresponding advantage in human evaluations. Third, collaboration provenance remains only partially detectable, revealing substantial overlap between Human-AI and Human-Human scientific outputs. Finally, firsthand AI collaboration is associated with greater perceived usefulness but smaller efficiency gains than expected, suggesting a nuanced recalibration of human expectations. Overall, our findings challenge the notion that AI’s value in scientific ideation lies primarily in producing better ideas, revealing instead its broader influence on scientific creation, evaluation, attribution, and human perceptions.
[HC-21] Atlas: Making Sense of Long Horizon Agent Trajectories
链接: https://arxiv.org/abs/2610.30548
作者: Sangho Suh,Joseph Chee Chang,Pao Siangliulue,Amy X. Zhang,Matt Latzke,Jonathan Bragg,Doug Downey,Daniel S. Weld
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:AI agents can now search literature, generate hypotheses, analyze datasets, and run experiments at a scale difficult for researchers to match. Yet scaling scientific execution does not automatically scale researchers’ ability to make sense of what agents produce. When researchers return to a long-horizon run, they must evaluate work they did not monitor: what the agent produced, which decisions shaped those outputs, and whether those decisions served the research goal. We frame this post-run sensemaking as a representation problem across three linked levels: goal, decision, and artifact. From this framework we developed Atlas, an agentic workspace with a trajectory tree, which renders the run as a branching structure showing what the agent chose and set aside, and artifact projections over six representations. In a within-subjects study (N=14), participants reported higher decision transparency, artifact sensemaking, and goal connection with Atlas than with a standard baseline, together with substantially lower subjective workload.
[HC-22] AVAtt: A Free-Viewing Eye-Tracking Dataset and Benchmark for Visual Attention on Paintings
链接: https://arxiv.org/abs/2610.30481
作者: Mohamed Amine Kerkouri,Marouane Tliba,Aladine Chetouani,Alessandro Bruno
类目: Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC); Image and Video Processing (eess.IV)
备注: Extenstion of https://doi.org/10.1145/3649902.3655656 published at ETRA 2024
Abstract:Saliency and scanpath models are developed almost exclusively on natural photographs, yet paintings are composed by artists to guide the viewer’s eye and differ from photographs in texture, colour and abstraction. We introduce \avatt, a free-viewing eye-tracking dataset on \nImages paintings spanning sixteen art movements and five centuries, with \nScanpaths scanpaths and \nFixations fixations from \nParticipants participants (\spPerImageMin–\spPerImageMax observers per painting). For every painting we release individual scanpaths with fixation timing and eye–screen distance, fixation maps, saliency maps smoothed at one degree of visual angle, temporal saliency videos and rich art metadata. We benchmark classic and deep saliency models against human agreement, revealing substantial headroom for art-aware models, and rank 44 scanpath-similarity measures, including a new language-free family of image-encoder metrics built on ten CNN, ViT and self-supervised encoders. With controls that remove viewing biases, we show that faces are fixated \faceEnrich \times more than expected and within half a second, that local structure draws the eye, that composition rules fitted from image content predict where people look, and that art style shapes how much viewers agree. Dataset, code and a Python package are available at: this https URL.
[HC-23] X-pressing Intimacy: Mental Health Disclosure After Twitter Acquisition
链接: https://arxiv.org/abs/2610.30413
作者: Olivia Pal,Yunhao Yuan,Meghana Mandava,Talayeh Aledavood,Eshwar Chandrasekharan,Koustuv Saha
类目: ocial and Information Networks (cs.SI); Computation and Language (cs.CL); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注:
Abstract:Social media platforms provide spaces for individuals to disclose mental health experiences and seek support, yet how these disclosures evolve during major shifts in platform governance remains unclear. We examine mental health self-disclosure surrounding Elon Musk’s October 2022 acquisition of Twitter, analyzing approximately 824K tweets across anxiety, depression, bipolar disorder (BD), borderline personality disorder (BPD), and suicidal ideation (SI) or related experiences, alongside 266K comparison tweets. Using Linguistic Inquiry and Word Count (LIWC) and an LLM-based intimacy classifier, we characterize changes in psycholinguistic expression and disclosure intimacy. Following the acquisition, we observe declines in affective and social language, alongside increases in first-person singular pronouns and future-focused expressions. Intimacy patterns vary by condition: anxiety, depression, and BPD show increases in low-intimacy and decreases in moderate-intimacy disclosures, whereas suicide-related posts show more high-intimacy disclosures. However, compared with broader Twitter discourse, only the relative decline in moderate- versus low-intimacy disclosure for BPD remains statistically significant. Our findings underscore the importance of considering both language and intimacy when studying platform transitions, informing the design of online environments that support sensitive self-disclosure.
[HC-24] Exploring spatial configuration and group familiarity when co-watching 360-degree stories in social virtual reality
链接: https://arxiv.org/abs/2610.30353
作者: Portia Wang,Monique Santoso,Yujie Tao,Jeremy Bailenson
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:While traditional cinematic viewing is often characterized as a social experience, engaging with immersive stories is predominantly researched as an individual activity. This mismatch becomes increasingly problematic in social virtual reality (VR), where features designed for typical social activities (e.g., virtual locomotion, spatial and social context) can become meaningful levers for shaping the user experience. To address our limited understanding of these effects, we conducted three exploratory experiments (N _1 =109, N _2 =116, N _3 =88) examining how spatial boundary type, size of allocated personal space, and group familiarity influence co-watching 360-degree stories in social VR. Results shows that small communal spatial boundaries heightened social presence towards others compared to individual boundaries, but also increased mental demand. Reducing the size of the personal spatial boundary increased social presence towards others but also increased disorientation and frustration. Finally, participants with higher group familiarity reported stronger attentional focus on the story, as well as higher spatial presence and social presence towards story characters, but also reported greater physical and perceptual demand. We discuss systematic tradeoffs between engagement and cognitive effort, and suggest practical guidelines for designing shared immersive narrative experiences.
[HC-25] Measuring undergraduate students reliance on generative AI during problem-solving: Scale development and validation
链接: https://arxiv.org/abs/2610.30195
作者: Chenyu Hou,Gaoxia Zhu,Vidya Sudarshan,Fun Siong Lim,Yew Soon Ong
类目: Human-Computer Interaction (cs.HC)
备注: If you are interested in accessing the full scale and instructions, please fill in this form: this https URL
Abstract:Reliance on AI describes the behavioral patterns of when and how individuals depend on AI suggestions, and appropriate reliance patterns are necessary to achieve effective human-AI collaboration. Traditional measures often link reliance to decision-making outcomes, which may not be suitable for complex problem-solving tasks where outcomes are not binary (i.e., correct or incorrect) or immediately clear. Therefore, this study aims to develop a scale to measure undergraduate students’ behaviors of using Generative AI during problem-solving tasks without directly linking them to specific outcomes. We conducted an exploratory factor analysis on 800 responses collected after students finished one problem-solving activity, which revealed four distinct factors: reflective use, cautious use, thoughtless use, and collaborative use. The overall scale has reached sufficient internal reliability (Cronbach’s alpha = .84). Two confirmatory factor analyses (CFAs) were conducted to validate the factors using the remaining 730 responses from this activity and 1,173 responses from another problem-solving activity. CFA indices showed adequate model fit for data from both problem-solving tasks, suggesting that the scale can be applied to various human-AI problem-solving tasks. This study offers a validated scale to measure students’ reliance behaviors in different human-AI problem-solving activities and provides implications for educators to responsively integrate Generative AI in higher education.
[HC-26] he EU AI Act Beyond Europe: A Practical Readiness Framework for Non-EU Organisations. Comparative Evidence from Four African Countries
链接: https://arxiv.org/abs/2610.30152
作者: Uloma Okoro,Branislav Radeljic
类目: Computers and Society (cs.CY); Human-Computer Interaction (cs.HC); Systems and Control (eess.SY)
备注:
Abstract:The EU Artificial Intelligence Act has implications for organizations within Europe as well as organisations outside the European Union where they provide AI-enabled services to EU clients, use European vendors, participate in EU-facing supply chains, or generate outputs that are used in the EU. This study examines the implications of these impacts for African countries and other non-EU organisations, using South Africa, Nigeria, Kenya and Egypt as comparative cases. It applies doctrinal legal analysis, structured document review and scenario-based comparison to assess the sufficiency or otherwise of existing regulatory frameworks within the selected countries. The analysis shows that the existing regulatory frameworks (like data-protection, financial-sector, cybersecurity and corporate-governance rules) can support AI readiness, but they do not resolve questions of territorial scope, role classification, risk categorisation, vendor responsibility and access to compliance evidence. The analysis proposes a seven-step readiness framework which covers exposure mapping, AI inventories, role and risk classification, governance, contracts, evidence management and continuous monitoring. It argues for proportionate, interoperable approaches that enable non-EU organisations to demonstrate readiness without treating EU compliance as a uniform exercise.
计算机视觉
[CV-0] A Unified Framework for Image-to-3D Part Generation via Variable Granularity NEURIPS2026
链接: https://arxiv.org/abs/2610.31671
作者: Qiwen Gu,Xin Zhou,Xiwu Chen,Zhongyang Zhu,Zhuo Zhang,Junqiao Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026. Code: this https URL
Abstract:Controllable 3D part generation is pivotal for digital asset creation, yet existing methods lack the flexibility to accommodate diverse user workflows requiring varying levels of control. In this paper, we introduce FlexPart, a unified image-to-3D part generation framework that supports variable-granularity geometric conditions within a single model. By seamlessly integrating heterogeneous inputs, ranging from sparse points to dense bounding boxes and masks, via gated adaptive modulation, FlexPart unlocks a powerful cross-modal synergy. We also propose an asymmetric geometric guidance mechanism that leverages strong spatial priors to align representations across granularities, improving geometric consistency and facilitating inter-task synergy. Under mask conditioning on PartVerseXL, FlexPart reduces Part CD by 16.6% relative to TripoSG + SAM3D and by 10.5% relative to OmniPart, while maintaining competitive whole-mesh quality. The code can be found at this https URL
[CV-1] Sensorimotor Alignment: A Platonic Proxy for Embodied Policy Success
链接: https://arxiv.org/abs/2610.31670
作者: Yunfei Xie,Howard H. Qian,Shiyi Lan,Kaiyu Hang,Chen Wei
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注: 28 pages. Project page and code available at this https URL
Abstract:Robotic foundation models are typically evaluated through costly closed-loop rollouts, which motivates cheap offline proxies for model capability. The Platonic Representation Hypothesis (PRH) suggests that as representations improve, different modalities such as vision and language converge toward a shared statistical model of reality. We ask whether motion trajectories form another such modality: as embodied models become more capable, should their sensory representations converge toward the structure of physical motion? To examine this, we propose Sensorimotor Alignment ( S^2 ), a training-free offline metric that measures how well a model’s sensory representation aligns with the geometry of physical motion. We find that S^2 correlates positively with downstream robot task success across embodied model families, making it a low-cost offline proxy for prioritizing models before closed-loop evaluation. S^2 can also identify promising vision-language model (VLM) backbones even before costly fine-tuning into vision-language-action models (VLAs). Beyond this correlation, S^2 offers representation-level explanations for phenomena observed in VLAs and world action models (WAMs). With S^2 , we extend PRH to embodied models and make sensorimotor alignment empirically measurable, offering insight into why an embodied model succeeds beyond its closed-loop success rate.
[CV-2] Reading Writing Sharing and Interacting with the World through Video
链接: https://arxiv.org/abs/2610.31668
作者: Shengqiong Wu,Meng Luo,Bobo Li,Ziwei Liu,Shuicheng Yan,Ming-Hsuan Yang,Philip Torr,Hao Fei
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 102 pages, 17 figures, 19 tables. Project page: this https URL
Abstract:Video foundation models have advanced rapidly in understanding and generating visual sequences, while video world models seek to capture the underlying dynamics of the world they depict. Despite their different objectives, both concern the relationship between video observations and an evolving world. This survey adopts world modeling as a unifying lens for examining these developments. Viewing an observed video as a partial, lossy, and temporally ordered projection of a latent temporal world, we organize the literature around four complementary operations: reading infers world properties from observed video; writing synthesizes video from specified or learned world conditions; sharing integrates understanding and generation through common modeling mechanisms; and interacting predicts and updates world states under actions and evolving observations. This perspective distinguishes temporal coherence in observed or generated sequences from consistency in the underlying world state. Across tasks, representations, models, datasets, and benchmarks, we trace a shift in emphasis from clip-level recognition and perceptual fidelity toward persistent state, controllable evolution, physical and causal validity, and evaluation grounded in interaction and downstream decisions, and we distill the open challenges that follow.
[CV-3] SpectralCanary: Embarrassingly Simple Low-Frequency Traces for T2I Dataset-Use Auditing NEURIPS2026
链接: https://arxiv.org/abs/2610.31659
作者: Sy-Tuyen Ho,Anirudh Satheesh,Soham Chawathe,Huy Nghiem,Yongkyun Lee,Andrew Adams Schoen,Furong Huang
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: NeurIPS 2026 Workshop on Attributing Model Behavior at Scale
Abstract:Auditing whether a text-to-image (T2I) model was fine-tuned on a particular image collection is difficult: a model can absorb collection-level visual characteristics without reproducing training images. Proactive dataset-use auditing plants a shared signal in a dataset before release, then later checks whether that signal appears in a suspect model’s generated outputs. Existing approaches often trade off verifiability, fidelity, robustness, and deployability. We introduce \textbfSpectralCanary , which uses a smooth low-pass mask to blend a fixed canary image into every collection image while leaving captions unchanged and requiring neither auxiliary generation nor per-image optimization. A lightweight verifier detects the transferred signal in suspect-model outputs. Across eight tests on the Pokemon dataset and three realistic data transformations, spanning two architectures families, SpectralCanary achieves the highest average per-generation TPR@ 1% FPR ( 91.1% ) and fixed threshold model-level accuracy using 30 generations ( 94.8% ) among the evaluated methods. Across five datasets spanning artistic, medical, satellite, and fashion domains with SD3.5, SpectralCanary averages 92.4% per-generation TPR@ 1% FPR while preserving the released images (SSIM 0.989 , LPIPS 0.022 ) with a near-zero mean change FID (mean \Delta FID 0.05 ) across datasets. Matched-energy ablations on Pokemon and ROCOv2 support our design choice, low-frequency placement transfers more reliably compared to mid/high-frequency placement. Under the evaluated adaptive attacks, the most effective removal requires aggressive modification and incurs more than 25 \Delta FID points. These results identify low-frequency visual signals as effective carriers for collection-level T2I audit traces.
[CV-4] BrickWright: Agent ic Creation of Physically Feasible Brick Builds
链接: https://arxiv.org/abs/2610.31651
作者: Ava Pun,Kevin Dai,Sally Chen,Maxwell Jones,Ruixuan Liu,Kangle Deng,Changliu Liu,Maneesh Agrawala,Deva Ramanan,Jun-Yan Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Frontier coding agents are becoming increasingly adept at 3D modeling. Yet, they still face challenges when building objects meant to be assembled in the physical world, often creating invalid assemblies with floating, colliding, or disconnected parts. We argue that providing agents with tools for assembling builds and verifying their physical feasibility is sufficient for them to produce physically valid assemblies that still adhere to input prompts. Focusing on the domain of LEGO brick builds, we present BrickWright, a framework for generating physically feasible brick builds from text or image prompts using pre-trained agents, where a build is physically feasible if it uses only real parts, has no colliding or floating parts, and forms a connected structure if possible. BrickWright guides agents to create builds by writing code with BrickLayer, a Python toolkit with two main components: an ergonomic brick assembly API that halves inference cost compared to working in raw LDraw, and a validator that provides feedback on physical feasibility. Whereas baseline agents achieve physical validity rates ranging from only 9% (GPT-6 Luna) to 69% (GPT-6 Astra), BrickWright-guided agents achieve 100% validity on a sample of 512 prompts, while improving prompt adherence for lightweight agents and maintaining it for stronger agents. Additionally, they outperform specialized models on the text-to-brick task. Code is available on our project page: this https URL
[CV-5] Gate3DGen: A Pre-processing Model that Selectively Purifies Real-World Images for 3D Generation
链接: https://arxiv.org/abs/2610.31649
作者: Xiaolin Li,Zihao Huang,Fanpeng Meng,Chen Yang,Guanjun Wu,Kai Zhang,Jiemin Fang,Qi Tian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Single-image 3D generation has achieved high fidelity on curated benchmarks, yet degrades sharply on real-world images with uneven illumination, blur, and occlusions. Purifying such inputs requires deciding when to intervene and preserving the visual cues needed for 3D generation, while leaving clean inputs unchanged. We introduce Gate3DGen, a selective pre-processing model that gates image inputs to existing 3D generators without retraining them. A degradation-aware router bypasses inputs classified as clean and extracts features to guide purification of the remaining inputs. Conditioned on these features, a 3D-oriented one-step purifier purifies degraded inputs with spatial losses that encourage preservation of object identity, contours, and textural cues. To provide paired supervision, we construct a synthetic data engine that renders clean views of 3D assets and synthesises composite degradations. On the synthetic benchmark, the purifier improves image fidelity and downstream 3D geometry in only 2.3s on average, reducing TRELLIS.2 mesh distance by 61.5% relative to degraded inputs. VLM evaluation and human preferences on real photographs further support identity preservation and degradation correction. More demos and code are available at this https URL.
[CV-6] OmniRay: Camera-Agnostic Monocular 3D Reconstruction with Composite Ray Fields
链接: https://arxiv.org/abs/2610.31645
作者: Namitha Guruprasad,Abhay Yadav,Cheng Peng,Rama Chellappa,Anand Bhattad
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Monocular models recover 3D geometry from a single image in one pass, but most assume a pinhole camera and produce distorted geometry on the wide-angle imagery produced by fisheye and panoramic lenses. Camera-aware models remove this assumption by predicting a per-pixel ray field. However, they represent different camera geometries using a single fixed basis to capture both the overall camera projection and its finer variations. We argue that camera geometry can be viewed as having a hierarchy of complexity for representing rays. We present OmniRay, a feed-forward model whose Ray Module is composed of three components: a monotone radial base that captures radially symmetric structure, a spherical-harmonic residual for smooth deviations from it, and a free-form residual for remaining local variation. These components are composed as successive rotations on the sphere, allowing the representation to accommodate diverse camera geometries. Given a single image, OmniRay predicts a ray direction and an up-to-scale radial distance for every pixel, from which a 3D point cloud is recovered without needing camera information at inference time. Extensive experiments across different camera configurations demonstrate that OmniRay establishes a new state-of-the-art in camera-agnostic monocular 3D reconstruction.
[CV-7] GigaBench: Benchmarking Learnable Representations for Gigapixel Image Compression and Super-Resolution WACV2027
链接: https://arxiv.org/abs/2610.31644
作者: Matteo Fusconi,Alex Costanzino,Pierluigi Zama Ramirez,Luigi Di Stefano
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page at this https URL Accepted at WACV2027
Abstract:Learnable Representations (LRs) can fit parametric families of functions to diverse kinds of signals. These emergent neural-native data formats are gathering interest in the scientific community due to their compactness, resolution-agnosticism and amenability to direct neural processing. On the other hand, gigapixel images are becoming increasingly widespread across a variety of scientific, cultural, and industrial domains. In this paper, we present GigaBench, the first benchmark and experimental evaluation focused on deploying LRs with gigapixel imagery. GigaBench includes images collected, acquired, and synthetically rendered across diverse domains, and addresses image compression and super-resolution. We adapt a representative set of LRs, spanning Implicit, Hybrid, and Explicit methods, so as to operate reliably at the gigapixel scale on consumer hardware. Through systematic evaluation across reconstruction quality, model size, compression and inference times, we identify consistent trends across methods and suggest actionable guidance for future investigation.
[CV-8] UniVOE: A Removal-Driven Framework for Interaction-Aware Universal Video Object Editing
链接: https://arxiv.org/abs/2610.31641
作者: Shuzhou Yang,Lanqing Guo,Lingen Li,Weiqi Li,Zhao Zhong,Liefeng Bo,Jian Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video object removal and injection are reciprocal editing operations between object-present and object-absent video states, yet they differ substantially in training difficulty. Removal can be effectively bootstrapped from video inpainting and further adapted to erase object-related interaction effects. Injection, however, requires large-scale paired supervision to model object appearance, dynamics, and scene-dependent effects, while perfectly aligned real-world pairs are difficult to collect and synthetic data lacks sufficient realism and diversity. To address this challenge, we present UniVOE, a removal-driven framework for training both video object removal and injection models. We first train an effect-aware remover by transferring video inpainting priors and finetuning on object-present/object-absent pairs. The trained remover converts real object-present videos into object-absent counterparts, creating large-scale paired supervision for injection. An injection model is trained on these pairs in the reverse direction to recover objects together with their dynamics and interaction effects. Moreover, since pixel-level reconstruction alone cannot adequately capture perceptual realism, we introduce a reward-based post-training method that aligns injection results with human preferences using a proprietary reward model. Experiments against strong open-source and commercial systems demonstrate that UniVOE achieves state-of-the-art performance on both object removal and injection, particularly in challenging interaction-effect cases.
[CV-9] WorldForge V2: Eliciting Camera-Trajectory Alignment in Video Diffusion Models
链接: https://arxiv.org/abs/2610.31640
作者: Chenxi Song,Yanming Yang,Tong Zhao,Ruibo Li,Joey Tianyi Zhou,Chi Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video diffusion models have learned rich 3D/4D priors, making them promising for scene generation and video re-rendering. However, their use in spatial tasks is limited by imprecise camera control, spatial and temporal inconsistencies, and entangled camera motion with scene dynamics. Existing approaches, such as task-specific fine-tuning and post-process warping, often introduce visual artifacts, generalize poorly, or incur high computational costs. We propose WorldForge V2 to achieve accurate camera control while preserving the generative capabilities of pretrained video models. We first develop a training-free guidance framework using warped references to iteratively refine generation along a target camera path. Optical-flow analysis identifies motion-related latent channels for selective guidance, helping separate motion control from appearance. A dual-path strategy further compares guided and unguided predictions to correct artifacts from imperfect references. Together, these components guide viewpoint changes while preserving scene appearance and object motion. To reduce refinement costs, v-cache approximates the correction using the preceding denoising prediction, avoiding an additional model evaluation. We further extend motion localization from latent channels to network blocks and introduce Trajectory LoRA, which concentrates adaptation on blocks most responsive to trajectory guidance. A two-stage training strategy teaches these adapters to denoise the states encountered during guided generation, using limited supervision effectively. This enables reliable camera control with very little training data, with a single training scene already supporting generalization to unseen scenes. Quantitative and qualitative evaluations across multiple datasets show that our method achieves state-of-the-art generation quality and trajectory accuracy in camera-controlled video generation.
[CV-10] HiP3R: Hierarchical Pose Graphs for Long-Horizon 3D Reconstruction
链接: https://arxiv.org/abs/2610.31635
作者: Zunnan Xu,Shangchen Zhou,Yihang Luo,Chen Change Loy
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Technical Report
Abstract:Feed-forward models accurately estimate camera motion and 3D geometry from short video windows. However, computational and memory costs prevent direct inference on sequences containing thousands of frames. Long videos must therefore be processed as overlapping chunks. This division produces independent local reconstructions with inconsistent coordinate frames and duplicate estimates of shared cameras. Achieving global consistency requires propagating constraints from adjacent overlaps and distant revisits across the sequence. This requirement naturally motivates pose graph optimization. However, the choice of graph level creates a trade-off. Chunk-level graphs efficiently resolve scale drift and close large loops. Yet, their rigid transformations cannot correct internal errors or conflicting frame predictions. Conversely, frame-level optimization offers local flexibility, but scaling it globally is computationally expensive and can break metric scale. HiP3R resolves this trade-off using a hierarchical pose graph. HiP3R first optimizes a robust chunk-level \mathrmSim(3) graph to establish the global layout and scale. It then holds this layout fixed and estimates one shared \mathrmSE(3) camera per frame, turning overlapping window predictions into a single trajectory. Experiments across multiple benchmarks demonstrate the effectiveness of HiP3R in improving camera trajectories and 3D reconstruction. Our method applies to both chunk-based and streaming backends without additional training or fine-tuning. The codes are available at this https URL.
[CV-11] F2G-Pose: Geometry-Aware Foundation Feature Lifting for Direct RGB-D Category-Level Object Pose Estimation NEURIPS2026
链接: https://arxiv.org/abs/2610.31623
作者: Jinyu Zhang,Haitao Lin,Jiashu Hou,Xiangyang Xue,Yanwei Fu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026. 22 pages, including the complete technical appendix
Abstract:This paper studies category-level RGB-D object pose estimation, which recovers an object’s 3D rotation and translation without instance-specific CAD models or reference views. This task is highly ambiguous due to intra-class variations, symmetries, and partial observations. Current solutions struggle to balance reliability and efficiency: methods based on intermediate structures are prone to error propagation, while diffusion models suffer from high inference latency and complex post-hoc candidate extraction. To address this, we present F2G-Pose, a real-time direct, single-pass framework that predicts category-level pose, metric size, and camera-frame completed shape from a single RGB-D observation in one feed-forward pass. F2G-Pose lifts dense visual foundation features to partial point clouds, encodes local geometry with DGCNN tokenization, and integrates object-level structure with a Mixture-of-Experts Transformer. Joint size prediction and camera-frame shape completion encourage metric consistency through dense geometric supervision. With its task-specific network trained only on pose-annotated synthetic SOPE data and its frozen RADIO backbone externally pretrained, F2G-Pose achieves state-of-the-art accuracy on SOPE and ROPE while running at 28 FPS for pose-only inference, and further demonstrates strong camera-frame geometry on HouseCat6D and real-object transfer on HANDAL. Analyses of latency, robustness, shape completion, and GenPose++ candidate behavior further demonstrate the efficiency and reliability of F2G-Pose, with real-world manipulation experiments demonstrating deployment potential under segmented RGB-D inputs. The project page, code, and pretrained models are made available at this https URL.
[CV-12] Beyond Best-of-k: Prior-Consistent DAPS for Ambiguous Nonlinear Inverse Problems
链接: https://arxiv.org/abs/2610.31592
作者: Merve Kocabas,Benjamin A. Burns,Andreas Geiger,Kent Gauen,Sara Fridovich-Keil
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Diffusion models have become strong priors for solving inverse problems. However, nonlinear inverse problems such as phase retrieval remain challenging when multiple measurement-consistent solutions are distinguishable only by the prior; we call such measurements ambiguous. Existing diffusion-based samplers rely on local approximations during reconstruction. For an ambiguous measurement, these approximations can allow measurement guidance to favor a measurement-consistent mode that is close to a prior-supported mode but is itself not supported by the prior. Evaluation often masks this type of failure by reporting the best-of- k runs, where the best is selected by comparison to the ground truth. We introduce Prior-Consistent DAPS (PC-DAPS), a variant of Decoupled Annealing Posterior Sampling (DAPS) designed to address this failure case by maintaining multiple candidate solutions within a single sampling run. As annealing noise decreases, PC-DAPS prunes and replaces weaker candidates according to their prior consistency, without any ground-truth access. PC-DAPS reduces the incidence of prior-inconsistent samples and improves sample quality on nonlinear inverse problems while preserving sample quality and diversity on linear inverse problems.
[CV-13] CellTrack-Bench: Microscopic Cell Tracking Benchmark for Vision-Language Models
链接: https://arxiv.org/abs/2610.31566
作者: Ilan Theodoro,Jordão Bragantini,Loïc A. Royer
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Large language models have shown outstanding performance on text-first tasks such as coding, math, and writing. Recent efforts are shifting models’ training to include scientific problems that often live in highly specialized domains beyond text, such as protein structure and biological images. However, in these scenarios, their general-purpose accuracy compared with specialized algorithms remains in question. One such domain is cell tracking, which requires spatiotemporal awareness of cell dynamics. To evaluate this, we created a new benchmark, CellTrack-Bench, which assesses Vision-Language Models’ (VLMs) cell-tracking-related performance against specialized trackers. The full benchmark draws from 15 standard cell-tracking datasets across four imaging modalities and comprises 1,537 problems corresponding to 9,041 cell-linking decisions. We selected a hard subset requiring greater reasoning effort by identifying failures shared across distinct algorithmic trackers. With this proposed dataset, we: (i) evaluated different environment harnesses for VLMs; (ii) benchmarked 9 frontier closed-source models and 13 open-weight models; and (iii) provided a detailed analysis of their failure modes. Code, the stored model outputs, and a pipeline that regenerates the benchmark from the source datasets are available at this https URL.
[CV-14] Squaring-the-Sphere: Distortion-Free and Seamless Panorama Generation via CubeAtlas
链接: https://arxiv.org/abs/2610.31565
作者: Xuhui Zuo,Chenjie Cao,Tengfei Wang,Junta Wu,Yisu Zhang,Chunchao Guo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Generating 360-degree panoramas requires both local perspective fidelity and global spherical continuity. Existing representations face a trade-off: equirectangular projection (ERP) enables seamless panorama modeling but suffers from severe distortion and a domain gap to pre-trained image generators, whereas cubemaps preserve perspective geometry but often produce inconsistent faces and visible seams. We present Squaring-the-Sphere, a panoramic image generation framework built on a novel representation CubeAtlas, which unfolds all cubemap faces into a single compact canvas. By arranging orientation-aligned, rotated polar regions alongside the equatorial faces, this representation preserves the perspective geometry of cubemaps while organizing the cross-face relationships required for seamless panorama assembly. It further enables a pretrained diffusion transformer to model the complete panorama jointly in a single pass, without altering the backbone architecture or positional encoding. While CubeAtlas has a token count comparable to ERP, its structured layout makes many dense attention interactions unnecessary. We therefore introduce a layout-aware sparse attention mask that preserves the interactions required for panorama continuity while removing geometrically irrelevant token pairs. As a result, CubeAtlas achieves better quality than ERP modeling with lower attention computation. Furthermore, we introduce a training-free super-resolution technique and a fusion scheme that enhance fine details in distortion-free perspective space before achieving the final ERP panorama. Experiments on various benchmarks demonstrate that our method improves geometric fidelity, cross-view consistency, and high-resolution visual quality over previous methods.
[CV-15] LACToS: Local Anomaly Detection via Conditional Typicality of Superpixels
链接: https://arxiv.org/abs/2610.31538
作者: Jihoon Hong,Alireza Kheirandish,Rohan Sanda,Sara Fridovich-Keil
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Local anomaly detection is the task of identifying spatially confined anomalous or out-of-distribution (OOD) regions within an otherwise in-distribution (ID) image. Most existing methods are designed for a specific context, with a strong focus on industrial defect detection, where they achieve excellent performance on well-established benchmarks. However, we observe that these methods can exhibit significant performance drops in other contexts, especially when the inlier distribution is multimodal. To account for this inlier complexity, we propose LACToS, a principled local OOD detection method based on statistical typicality and made tractable through two innovations: (i) grouping pixels into superpixels and (ii) using a computationally efficient proxy to estimate typicality. We evaluate LACToS across three diverse applications spanning medical imaging, human faces, and X-ray security screening. LACToS performs more robustly than prior methods across these benchmarks. The code is released here: this https URL.
[CV-16] Diffusion-Based Synthesis of Complete 3D Biparametric Prostate MRI MICCAI MICCAI2026
链接: https://arxiv.org/abs/2610.31523
作者: Claudia Giardina,Montse Pardàs,Verónica Vilaplana
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Presented at the 6th Deep Generative Models Workshop (DGM4MICCAI), MICCAI 2026
Abstract:Synthetic biparametric MRI (bpMRI) may help alleviate the limited availability of large and representative prostate MRI datasets for deep learning applications. In this work, we present a conditional 3D diffusion framework for anatomy-conditioned synthesis of complete prostate bpMRI from anatomical masks. The framework supports sequential synthesis, in which T2w and diffusion modalities are generated by separately trained models, and joint synthesis, in which all bpMRI modalities are generated simultaneously within a single model. Five configurations were trained using PI-CAI, FastMRI Prostate, and PROSTATEx datasets to investigate different conditioning schemes, target modalities, and synthesis strategies. Evaluation using distribution-based metrics, paired-image similarity, and qualitative assessment demonstrated the feasibility of both synthesis pathways. Joint synthesis achieved competitive distributional fidelity while generating complete bpMRI, whereas direct synthesis of ADC and computed high-b-value DWI outperformed their derivation from synthetic diffusion-weighted images.
[CV-17] Evaluation of Uncertainty-Aware Multi-Software Ensembles for Hippocampal Segmentation
链接: https://arxiv.org/abs/2610.31505
作者: Gabriel Oliveira-Stahl,Anna Schroder,James Moggridge,Hamza A. Salhab,Caroline Micallef,Josephine Barnes,M. Jorge Cardoso,Carole H. Sudre,Matthew Grech-Sollars
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted for publication at the Journal of Machine Learning for Biomedical Imaging (MELBA) this https URL
Abstract:Accurate hippocampal segmentation can be a useful tool for diagnosing and monitoring neurological conditions such as Alzheimer’s disease and epilepsy. While numerous automated segmentation methods exist, their clinical adoption remains limited. Reliable uncertainty assessment can enhance trust and facilitate clinical translation. This study evaluates ensembles of five heterogeneous hippocampal segmentation methods - InnerEye, ASHS, FastSurfer, HippoSeg, and FreeSurfer - across two dementia datasets and one epilepsy dataset. The sub-ensemble containing InnerEye, FastSurfer, and HippoSeg emerged as both accurate and efficient, highlighting the feasibility of balancing computational cost and performance. Additionally, ensemble-derived uncertainty quantification with sample variance, mutual information, and predictive entropy is shown to reduce inaccurate segmentations by flagging low-confidence cases, potentially providing a mechanism for automatically escalating ambiguous cases for expert assessment. Lastly, this study explores the downstream clinical utility of ensemble-derived uncertainty, finding a significant correlation with motion artefact severity and a modest contribution to Alzheimer’s disease classification.
[CV-18] Agent VQA: A Unified Benchmark for Agent ic Visual Understanding
链接: https://arxiv.org/abs/2610.31494
作者: Advait Gupta,Sachin Konan,Tom Yin,Gabriel Sarch,Zhuang Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Vision-language models (VLMs) can perform a broad range of tasks across diverse settings, yet their performance in agentic contexts remains poorly understood. Although online evaluations in simulators and real-world environments are essential for measuring interactive agent performance, they are costly and challenging to scale. Complementary offline VQA benchmarks can enable faster and more reliable comparisons across agentic domains, yet existing benchmarks remain limited and fragmented. To address this gap, we introduce AgentVQA, an offline benchmark for evaluating VLMs on realistic agentic scenarios across five agent domains: Web Agents, Robotics, Egocentric Videos, Games, and Spatial Planning. We build AgentVQA by manually selecting 1000 challenging, real-world agent examples from 23 source datasets and adapting them for reliable offline evaluation. The resulting benchmark preserves realistic scenarios from interactive environments across diverse domains, and its model rankings track those from corresponding online evaluations. Across the 22 evaluated models, the best achieves only 58.5% accuracy. Our cross-domain analysis shows that errors are driven primarily by failures to ground and maintain accurate representations of the current and evolving visual state, including localizing relevant UI elements, interpreting motion, tracking state changes over time, and inferring 3D spatial relationships, rather than by task misinterpretation or downstream logical reasoning.
[CV-19] MoG-VLM: Efficient Video Anomaly Detection through Motion-Guided Vision-Language Models BMVC2026
链接: https://arxiv.org/abs/2610.31478
作者: Khaled Waleed Dawoud,Muhammad Zaigham Zaheer,Khan Muhammad,Muhammad Haris Khan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at BMVC 2026. Includes supplementary material
Abstract:Vision-Language Models (VLMs) have shown strong potential for video anomaly detection, but dense frame-level inference makes them impractical for real-time deployment. Existing acceleration commonly relies on uniform sampling, which ignores content dynamics and can either waste computation on static segments or miss localized anomalies. We propose MoG-VLM a motion-adaptive anomaly detection framework that leverages motion purely as a lightweight control signal to decide when VLM inference is necessary. To this end, we introduce Temporal Gradient Magnitude (TGM) which performs content-aware sampling to concentrate computation on informative regions while avoiding redundant processing. We further introduce a hierarchical dual-stream design in which dense multi-frame VLM reasoning is activated only around frames with elevated anomaly likelihood, via anchor-triggered context-window expansion with early termination. A fusion module then reconciles these instantaneous semantic cues with temporal action dynamics. Across four benchmarks, our method achieves 91.91%, 84.30%, 87.88%, and 78.96% AUC-ROC on XD-Violence, UCF-Crime, MSAD, and ShanghaiTech, respectively, establishing new state of the art on three benchmarks under the zero-shot, training-free setting. On XD-Violence, we improve over the previous best result by +0.57% while delivering a 1.31x inference speedup. These results demonstrate that content-aware allocation of VLM capacity, processing fewer frames more intelligently rather than all frames uniformly, is a more effective axis of progress than uniformly densifying VLM inference across frames. Code is available on GitHub: this https URL.
[CV-20] Ask About the Right Moment: Training-Free Object Identity Reasoning for Video-Language Models
链接: https://arxiv.org/abs/2610.31471
作者: Di Yang,Xiaochang Li,Muyang Liu,Yufei Wang,Yanhai Xiong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Many questions about a video concern one physical object. Where does the ball end up after the cups are shuffled? Which card is the queen? How often was the mug covered? Video-language models (Video-LMs) recognise such objects in single frames but do not keep them apart over time. On real shell-game videos, open models from 8B to 235B parameters, their thinking variants, a reinforcement-learned tool agent and three training-free tool agents all stay within ten points of chance. We separate semantic identity binding from temporal identity maintenance. An object’s identity should be fixed at a moment when the evidence for it is visible, and carried afterwards by a tracker. \emphTrack, Then Bind implements this principle without training. A promptable video tracker follows every candidate object and draws it with a track number. A frozen vision-language model answers a static question about the identifying evidence, which binds the target to a track. The answer is then derived from the tracked state in text, without asking the model to reinterpret the original video during read-out. A single configuration covers all benchmarks. With Qwen3-VL-32B it solves both VET-Bench tasks, where a fine-tuned trajectory model reaches 92% and 90% and the same VLM voting over 16 samples stays at 30–32%. It answers 55.4% of the hardest Perception Test cup games, where the best published model answers 41.5%, and, using track-derived timelines, 48.6% of TOC-Bench questions, against 37.1% for the same model answering directly. Drawing the same tracks on the video and asking the question directly falls 12–19 points short of the agent on real cup games.
[CV-21] AnyGameRenderer: Entity-Consistent Generative Rendering Across Games and an Entity-Level Benchmark
链接: https://arxiv.org/abs/2610.31466
作者: Ruicong Liu,Liangyang Ouyang,Fengbo Lan,Zhixiang Wang,Kaipeng Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:AI coding agents can now build small playable games in minutes, but the results usually look like programmer art: gray blockouts, colored geometric primitives, and flat semantic blobs. We present AnyGameRenderer, a video-to-video generative renderer that turns such footage into high-quality visuals with characters, motion detail, and effects. Throughout, it keeps the layout, camera, and every entity’s position and timing intact. Our AnyGameRenderer generalizes across a wide range of genres that agents build, such as tower defense, auto-battlers, top-down survivors, turn-based battles, kart racing, and crowd runners. AnyGameRenderer takes plain video as its game-state input, with no complex representations such as engine G-buffers, per-object class labels, or low-poly proxies. Gray blockouts, colored primitives, semantic maps, and wireframes all work as input. It also preserves entity correspondence: group counts stay correct, and entities appear and disappear on the right frame. We implement AnyGameRenderer by LoRA fine-tuning MiniMax-H3 ref2va, a 33B reference-video-conditioned video generation model. Training uses \sim 10k paired clips from three sources: (i) multi-stream rendering data from 38 agent-made games exported from GEAR, an AI game-creation tool; (ii) two-stream rendering data from Unreal Engine, with programmatically controlled crowds, count changes, and entity scales; and (iii) real gameplay recordings, with pairs reconstructed via 3D-box detection. To evaluate the entity correspondence, we propose an entity-level benchmark of 59 held-out cases with 753 entities, scored against ground-truth boxes. Compared with the base model, AnyGameRenderer better preserves entity correspondence: temporal match rises from 50.9% to 62.9%, missed count changes fall from 34.7% to 21.3%, and fewer entities are hallucinated or left unrendered. Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.31466 [cs.CV] (or arXiv:2610.31466v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.31466 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Ruicong Liu [view email] [v1] Thu, 22 Oct 2026 15:58:39 UTC (6,583 KB) Full-text links: Access Paper: View a PDF of the paper titled AnyGameRenderer: Entity-Consistent Generative Rendering Across Games and an Entity-Level Benchmark, by Ruicong Liu and 4 other authorsView PDFHTML (experimental)TeX Source view license Additional Features Audio Summary Current browse context: cs.CV prev | next new | recent | 2026-10 Change to browse by: cs References Citations SciXGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv’s community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) mathjaxToggle(); We gratefully acknowledge support from our major funders, member institutions, , and all contributors. About Help Contact Subscribe Copyright Privacy Accessibility Operational Status (opens in new tab)
[CV-22] MixPT: Point Transformer with Interleaved Window and Global Attention
链接: https://arxiv.org/abs/2610.31455
作者: Minkyun Seo,Hyungtae Lim,Jaesik Park
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 23 pages, 14 figures, 12 tables. Project page: this https URL
Abstract:Point cloud transformers aim to capture detailed local geometry and relationships across a scene at a manageable computational cost. Two common approaches are window attention and global attention. They differ in how they enable scene-wide interaction, which refers to information exchange among tokens across a scene. Window attention reduces computational cost by restricting self-attention to groups of tokens called windows; in serialized designs, each window contains a fixed number of consecutive tokens from a spatially ordered sequence. Because each block only directly connects tokens within the same window, information exchange across windows relies on changes in window membership across blocks and hierarchical pooling. Global attention, in contrast, directly connects all tokens within a scene, but computing attention over all token pairs can be costly when the number of tokens is large. To balance computational cost and scene-wide interaction, we introduce MixPT, a point transformer that uses a mix of window and global attention blocks in an interleaved schedule, while retaining a single fixed serialization pattern for its window blocks. For scenes with a large number of tokens, MixPT applies global attention to spatially pooled tokens and returns residual updates to the original-resolution stream. Across indoor and outdoor benchmarks, MixPT achieves competitive accuracy with lower latency than representative window and global attention baselines. We further demonstrate the feasibility of scaling the attention stage in depth by repeating the interleaved pattern of window and global attention blocks.
[CV-23] A Sparse Topometric Graph for Scalable Cross-Session Indoor Navigation
链接: https://arxiv.org/abs/2610.31432
作者: Eun Sun Lee,Minkyun Seo,Gwangtak Bae,Junho Kim,Seunggu Kang,Jaesik Park,Young Min Kim
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:A conventional embodied agent in an indoor environment builds a map by accumulating noisy real-world measurements and results from heavy inference to guide navigation. While the detailed measurements may differ due to scene changes or sensor settings, the unique topological structure of navigable passages or the connectivity between rooms may provide training-free guidance that complements local observations. We propose a sparse topometric graph as an efficient structural representation for robust indoor navigation, alleviating the need for fine-grained matching, which is computationally expensive, memory-intensive, and prone to real-world variations. The map consists of nodes that encode local 3D geometry and distance-weighted edges that capture traversable connectivity. A learned graph-matching network finds consistent subgraph structures across sessions by considering both node-level geometric descriptors and neighborhood connectivity. The holistic contextual matching is robust to local ambiguities caused by weak distinctiveness or repetitive structure, and establishes stable correspondences despite local scene changes, different sensor modalities, or limited spatial overlap. We analyze the practicality of the graph through cross-session point-goal navigation, where a goal specified in a prior reference map must be reached from an unknown initial pose in a later session. Experiments show that our method achieves the highest success rates (77.1% on Gibson and 40.0% under 3D-FRONT scene changes), while reducing memory by more than two orders of magnitude compared to point-level representations. The proposed approach further supports cross-session alignment and goal-directed navigation with strong scalability and robust real-world transfer.
[CV-24] WeVisOne-VL: Unified Visual Perception from Objects to Pixels
链接: https://arxiv.org/abs/2610.31329
作者: Tianming Liang,Junyi Pan,Yuxi Li,Canmiao Fu,Zhipeng Huang,Chong Sun,Chen Li,Jing Lyu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Existing vision-language models (VLMs) remain unreliable for fine-grained visual perception. In this report, we present WeVisOne-VL (WeChat Vision All-in-One), a powerful VLM that unifies object-level and pixel-level perception through structured next-token prediction. It supports multiple visual tasks including object detection, referring localization, counting, OCR, segmentation, and matting, enabling localization at any user-specific granularity, from points and boxes to masks and alpha mattes. Architecturally, we integrate a lightweight pixel decoder into the VLM via a special OBJ token to generate instance-level pixel outputs. We train the model with a three-stage curriculum SFT followed by mixed-task RL. Across over 30 benchmarks, WeVisOne-VL demonstrates leading performance on both object-level and pixel-level tasks, while maintaining robust rejection for object-absent queries. By releasing WeVisOne-VL to the open-source community, we hope to advance VLMs from broad visual understanding toward truly reliable perception systems for real-world deployment.
[CV-25] MVLoMa: Multi-View Local Feature Matching with Any Geometry Foundation Model
链接: https://arxiv.org/abs/2610.31321
作者: Xinyang Lu,Shengzhe Wei,Sheng Tang,Weiqi Zhang,Juntong Fang,Zesong Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project Page: this https URL
Abstract:Local feature matching is typically formulated for image pairs, although structure-from-motion and visual localization commonly match one reference image against many retrieved or covisible neighbours. Applying a pairwise matcher independently to these centre-neighbour edges duplicates reference-side computation and prevents the matcher from exploiting complementary multi-view geometric evidence from the remaining observations. To address these limitations, we introduce MVLoMa, a sparse star-graph matcher that extends LoMa from image pairs to multiple views. MVLoMa first uses a frozen multi-view geometry model to jointly describe all images in a star, then injects the resulting cross-view geometric context into local descriptors. It then matches the centre image to all neighbours in a single forward pass through per-view self-attention and bidirectional centre-neighbour cross-attention. For supervision, we derive positive correspondences only from depth-valid, mutually visible dense warps and explicitly penalize high-confidence matches that violate epipolar geometry. On 25-view star graphs from HyperSim, ScanNet++, Mip-NeRF 360, and Tanks and Temples, MVLoMa obtains the best macro-average relative-pose AUC among the compared methods at all three thresholds (58.65/85.41/90.99). Its matcher is 2.20 times faster than sequential evaluation of the 24 centre-neighbour edges. The same design supports four tested frozen geometry encoders – DA3, \pi^3 , VGGT, and VGGT- \Omega – with DA3-Giant performing best on all reported encoder-ablation metrics. Controlled ablations further isolate the effects of geometry-aware descriptors, visibility filtering, geometric negative supervision, and joint multi-view inference.
[CV-26] Stream4D: Streaming 4D Reconstruction with Training-free Motion Decoupling via Causal KV-Cache Attention
链接: https://arxiv.org/abs/2610.31320
作者: Wei Zhang,Ying Feng,Mengtong Wu,Shiqiang Gong,Huaqi Zhang,Qi Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 8 figures. Accepted to ACM Multimedia 2026
Abstract:4D scene reconstruction requires robust handling of dynamic objects, yet existing streaming methods commonly assume static scenes, leading to pose drift and ghosting artifacts over long sequences. We introduce Stream4D, a training-free, motion-decoupled streaming reconstruction framework that can be directly integrated into causal Transformer backbones (e.g., StreamVGGT, Stream3R). Our key observation is that causal attention implicitly encodes rich temporal motion cues within the KV cache. We adaptively aggregate these cues from multiple attention layers and across temporal windows, and fuse them with depth priors to handle challenging cases such as occlusions and textureless regions, producing precise dynamic masks that decouple static and dynamic regions. We then apply a soft key masking scheme that leverages these dynamic masks to suppress interference from dynamic regions in causal attention, thereby improving camera pose estimation accuracy. Stream4D inherits the efficiency of causal attention and remains compatible with existing streaming architectures. On diverse dynamic-scene benchmarks, Stream4D significantly improves camera pose estimation and static-scene reconstruction quality, yields reliable dynamic masks, and effectively mitigates long-sequence ghosting.
[CV-27] Data-Driven Vision-Language Models for Remote Sensing: A Survey
链接: https://arxiv.org/abs/2610.31304
作者: Yue Zhou,Shujun Zhao,Xue Yang,Ruigang Li,Tianwen Zhang,Mengcheng Lan,Chaofeng Chen,Lingfei Ma,Hongjie He,Jonathan Li
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 44 pages, 15 figures
Abstract:During the deep learning era, innovations in remote sensing (RS) vision models primarily focused on optimizing network architectures for specific tasks and conducting end-to-end training. However, with the advent of the Vision-Language Models (VLMs) era, constrained by substantial computational resource demands, the mainstream research paradigm has shifted towards adapting general-purpose VLMs using RS vertical domain data for fine-tuning, rather than modifying their core architectures. This transition marks a change in the innovation pathway from model-centric to data-centric. Despite the increasing importance of data, the academic community still lacks a systematic, data-centric survey to guide researchers in constructing high-quality RS vision-language datasets for various application scenarios and in conducting effective model fine-tuning and performance evaluation based on these datasets. To address this gap, this paper presents a systematic review of the development status of RS datasets tailored for VLMs. It further provides a comprehensive practical guide for RS researchers in the large model era, covering the entire pipeline from data construction to fine-tuning and evaluation. This paper aims to encourage the academic community to build more diverse and higher-quality vision-language datasets for specialized RS sub-fields. Ultimately, the goal is to converge these efforts into a comprehensive RS vision-language corpus that offers extensive task coverage and a wide geographical distribution, thereby laying a solid data foundation for advancing towards general artificial intelligence in RS.
[CV-28] PiHand: Latent Hand-Object Interaction Priors for Physics-Based Policy Learning
链接: https://arxiv.org/abs/2610.31299
作者: Niklas Berndt,Xianghui Xie,Enes Duran,Yixiao Wang,Gerard Pons-Moll
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Synthesizing physically grounded dexterous hand-object interactions is important for VR/AR, gaming, and robotics, yet remains challenging because subtle contact dynamics, diverse object geometries, and target trajectories must be jointly respected while maintaining natural grasps. Existing methods either track mocap trajectories with limited transfer to novel objects, or rely on hand-only generative priors that fail to capture coupled hand-object motion, sacrificing contact fidelity and sample efficiency. To tackle these challenges, we introduce PiHand, a Physically grounded Interaction-aware Hand motion synthesis framework. PiHand first trains expert policies that track mocap demonstrations in simulation, distills them into a unified hand-object interaction prior using a goal-conditioned VAE, and finally solve diverse dexterous manipulation tasks through bounded residual exploration in its latent space. With only a single reward and no complex curriculum training, this yields general tabletop manipulation policies that generalize to unseen objects and trajectories on GRAB and DexYCB. The same prior transfers to object catching and bimanual handover, two tasks that are rarely seen in its training data. Across all tasks, PiHand consistently outperforms baselines in all metrics. We will release the processed data and model checkpoints.
[CV-29] RecallGS: Geometry-Grounded Recall for Online Semantic Gaussian Splatting
链接: https://arxiv.org/abs/2610.31293
作者: Thanh-Khoi Nguyen,Thien-Phuc Tran,Minh-Triet Tran
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Online semantic 3D Gaussian Splatting (3DGS) enables open-vocabulary scene understanding from streaming observations. However, existing methods recover geometric context missing from image-derived semantic features using an additionally learned 3D network, incurring extra training and runtime overhead. We observe that such geometric-semantic evidence may already exist in the agent’s observation history and can be retrieved rather than learned. Based on this insight, we introduce RecallGS (Geometry-Grounded Recall for Online Semantic Gaussian Splatting), a lightweight, training-free mechanism that treats past observations as a non-parametric geometric-semantic prior. RecallGS follows a recall-ground-fuse pipeline: CLIP retrieves relevant observations, alpha-compositing weights identify geometrically supported views, and aggregate their semantic features over each Gaussian’s rendered footprint. On ScanNet and ScanNet++, RecallGS achieves strong semantic segmentation performance at 2.48 FPS, demonstrating that geometric-semantic context can be effectively recovered from observation history without an additional learned 3D branch.
[CV-30] Retrieval-Augmented Online Textual Memory for Episodic Memory Video Question Answering
链接: https://arxiv.org/abs/2610.31284
作者: Raffaele Cali,Giuseppe Lando,Rosario Forte,Antonino Furnari
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Long egocentric videos pose a central challenge for multimodal language models: the visual stream must be compressed into a memory representation that can be searched efficiently while still preserving the evidence needed for question answering. Building on an online textual-memory framework in which first-person videos are converted into temporally grounded captions, this work systematically studies the role of retrieval in episodic-memory video question answering. We evaluate how different retrieval units, query formulations, context budgets, and reranking strategies affect multiple-choice answering on QAEgo4D-Closed, isolating retrieval as the main variable while keeping the underlying generated memory fixed. Our best configuration, Parent, matches fine-grained caption subchunks, then provides their full parent captions to the reasoner. It achieves 56.4 % accuracy, matching strong full-memory and retrieval baselines while exposing only a small fraction of the textual memory at question time. More importantly, our analysis shows that better temporal coverage does not necessarily translate into better answers: increasing the retrieval budget improves temporal Hit@K and Recall@K, but can reduce accuracy by introducing visually plausible distractors and uncertain caption evidence. These findings suggest that retrieval-augmented egocentric memory should not be optimized only for evidence recall, but also for the reliability, specificity, and answer-discriminative value of the retrieved context. The study motivates more structured episodic memories that move beyond raw caption retrieval toward evidence-aware representations for long-form egocentric reasoning. Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.31284 [cs.CV] (or arXiv:2610.31284v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.31284 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-31] ReCoG: Relational Concept Graph for Training-Free Personalization
链接: https://arxiv.org/abs/2610.31282
作者: Changseon Yu,Hyeokjun Kweon
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Training-free personalization adapts a frozen vision-language model (VLM) to recognize user-defined concepts from only a few exemplars by retrieving concept entries at inference time. While recent retriever-based methods enrich each concept with descriptive evidence, they largely treat concepts in isolation, so the retrieved information can be plausible yet non-discriminative, especially when the personalized concept set contains highly confusable, fine-grained categories. We argue that personalization is inherently relational: correct recognition depends on the specific distinctions between a concept and its nearest alternatives. To this end, we propose ReCoG, a Relational Concept Graph for retrieval-augmented personalization. ReCoG represents concepts as nodes and augments them with directed edges that explicitly capture discriminative cues between concept pairs, elicited by a frozen VLM during database construction. At inference, ReCoG retrieves a shortlist and resolves ambiguity via explicit pairwise comparisons conditioned on the corresponding relational edges, selecting the concept that consistently wins against competing candidates. We also introduce FINGER benchmark to test personalized recognition on confusable concept sets. Across FINGER, ReCoG substantially outperforms prior methods, with the largest gains in the most confusable regimes. ReCoG also achieves consistently best performance on established personalization benchmarks including captioning and VQA, confirming that relational evidence is broadly beneficial beyond fine-grained settings. Project Page: this https URL
[CV-32] Reinforcing intra-operative tumour localisation with weak supervision from pre-operative appearances
链接: https://arxiv.org/abs/2610.31253
作者: Sidhart Sami,Ali Sharjeel,Sri Atluri,Miao Zhao,Caroline Roney,Eirini Marouli,Shaheer U. Saeed
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Limited tumour visibility on intra-operative images complicates surgical navigation, including liver hepatocellular carcinoma (HCC) targeting on computed tomography (CT) imaging. This makes labelled datasets difficult to curate for deep learning, and transferring pre-operative delineations through registration remains challenging due to intra-operative deformation and lack of paired pre- and intra-operative data. We propose to transfer tumour appearances learned from pre-operative magnetic resonance (MR) images to guide localisation on intra-operative CT through weak supervision. A classifier is trained to classify tumour presence on pre-operative images. This is applied to windows on intra-operative images, and despite the domain gap, it assigns higher tumour-presence probabilities on average to tumour-containing windows. These weak cross-domain signals guide a reinforcement learning agent to localise tumours on intra-operative images. We compare direct transfer, few-shot adaptation and synthetic transfer (trained using unpaired MR and CT), which translates CT into synthetic MR before localisation. On an external HCC cohort, synthetic transfer outperformed other tested methods and CT-only baselines. Our framework enables cross-modality tumour localisation without paired MR-CT training images.
[CV-33] Lightweight Frame Compensation for Real-Time Polyp Segmentation under Frame Skipping
链接: https://arxiv.org/abs/2610.31228
作者: Weihao Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 1 figure
Abstract:Real-time polyp segmentation in computer-assisted colonoscopy is hampered by the difficulty of deploying high-accuracy models under strict latency budgets. Frame skipping, which performs inference every K frames and reuses or interpolates the rest, improves throughput but introduces motion blur and lesion misalignment that degrade temporal consistency. To close this gap, we propose the Lightweight Frame Compensator (LFC), a 77K-parameter plug-and-play module inserted before a frozen segmentation backbone. Operating on a three-frame window, LFC fuses consecutive frames via spatio-temporal attention into a single compensated frame, and is trained only on pseudo-labels generated by a differently seeded model, without any expert annotation. Experiments on three polyp video benchmarks show that LFC recovers most of the accuracy degraded by direct frame skipping and, in several settings, reaches the accuracy of frame-by-frame inference, while retaining a 2.05X–2.55X throughput gain. We further evaluate the temporal stability of three backbones spanning CNN, pyramid Transformer, and global-attention Transformer architectures, revealing a spectrum of sensitivity to frame skipping that arises from how locally or globally each architecture aggregates misaligned input. The compensation gain of LFC follows this spectrum: it is largest for high-capacity Transformer backbones, which achieve the best static accuracy but suffer the most under skip-frame inference. LFC thus offers a lightweight, pluggable, and deployment-friendly route to real-time polyp segmentation.
[CV-34] GLOW-FM: Efficient Generative Low-Light Image Enhancement with One-Step Latent Flow Matching WACV2027
链接: https://arxiv.org/abs/2610.31222
作者: Ariel Lapid,Yuval Haitman,Reuven Peretz,Arnon Netzer,Idit Diamant
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: WACV 2027
Abstract:Low-light image enhancement (LLIE) aims to recover well-exposed and perceptually faithful images from severely degraded observations. Recent diffusion-based generative methods achieve strong perceptual quality by leveraging powerful image priors, but their iterative sampling incurs substantial computational cost, limiting practical deployment. Despite recent efficiency gains in generative LLIE, high-quality enhancement with a single function evaluation and low overall computation remains challenging, particularly under diverse real-world low-light conditions. We introduce GLOW-FM, a one-step latent flow-matching framework for efficient generative LLIE. GLOW-FM performs conditional transport in the latent space of a frozen Tiny-VAE and leverages Terminal Velocity Matching (TVM) to enable enhancement with a single function evaluation. Matched comparisons with alternative one-step objectives support TVM as an effective design choice, and, to the best of our knowledge, this is its first adaptation to image restoration and specifically to LLIE. Extensive experiments on challenging real-world benchmarks demonstrate that GLOW-FM achieves competitive image quality and state-of-the-art cross-dataset generalization while requiring only a single function evaluation and 34 GFLOPs for the small variant. These findings highlight the potential of GLOW-FM to bring generative LLIE closer to practical deployment through single-step inference and compact latent modeling. Code will be made publicly available upon publication.
[CV-35] Moving Regions Not Moving Objects: Segmenting What Actually Moves under Camera Motion
链接: https://arxiv.org/abs/2610.31209
作者: Wenhua Jin,Tiancheng Chen,Sheng Tang,Jiahao Li,Yupeng Liu,Weiqi Zhang,Juntong Fang,Zesong Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page and video demonstrations: this https URL
Abstract:Moving object segmentation is conventionally formulated at the sequence and object level, marking an entire instance as moving once it moves anywhere in a video. Recent instantaneous formulations distinguish moving from stationary intervals, but still generate masks at the object level through frozen segmentation foundation models. Because the mask is derived from an instance-level proposal, motion is typically attributed to the entire instance rather than localized to the regions that actually move. Moreover, such instance-level predictions can produce masks even when no part of the instance is actually moving, revealing a more fundamental issue: motion is not an intrinsic property of an instance; it can only be determined by comparing at least two frames. Building on this premise, we introduce MoRSe, which jointly interprets two frames to predict a moving-region mask, distinguishing actual scene motion from apparent motion induced by the camera. A single RGB encoder replaces multi-stage foundation stacks, so no optical flow, point tracks, or separate mask generator is needed at inference. We release MoRSe-5K, a benchmark of 42 sequences with Static, Dynamic-Static, and Dynamic settings designed to evaluate the model’s ability to distinguish and localize moving and stationary regions under different camera and scene motion conditions. MoRSe outperforms both specialized moving object segmentation methods and dense 4D reconstruction methods with motion-thresholded predictions, while producing the fewest false positives on sequences where nothing moves.
[CV-36] A Visual Grounding Benchmark for Retinal OCT with Multimodal Large Language Models
链接: https://arxiv.org/abs/2610.31205
作者: Ronald Fecso,José Morano,Botond Fazekas,Paul Widmann-Sedlnitzky,Dorota Himel,Laetita Hinterhuber,Virginia Mares,Andreas Pollreisz,Hrvoje Bogunović
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Under review at npj digital medicine
Abstract:Multimodal large language models (MLLMs) can answer clinical questions about medical images, but trustworthy use requires visual grounding: models should localize the pathology they report rather than rely on spurious shortcuts. Whether MLLMs can visually ground clinical findings in retinal optical coherence tomography (OCT), the most widely used imaging modality in ophthalmology, remains largely unexplored. To address this gap, we introduce RetinaGroundBench, the first benchmark for biomarker visual grounding in retinal OCT. It comprises 815 B-scans exhaustively annotated by ophthalmologists with 4,537 bounding boxes covering 13 biomarkers across 8 diseases. Because MLLMs often fragment a single lesion into multiple partial predictions, we propose AggIoU, a metric that evaluates whether the union of overlapping predicted boxes covers the annotated region. Across 27 open-source MLLMs (0.8-38B parameters), we find that visual grounding remains largely unsolved: the best general-purpose model achieves a macro F1 of just 23.7%, with consistently low recall, despite reaching up to 73% balanced accuracy on biomarker presence questions. To narrow this gap, we developed RetGrounder, an MLLM fine-tuned using an OCT-specific strategy integrating grounding and textual information. Despite its relatively small size (8B), RetGrounder improves grounding F1 to 33.9%, a 43% relative improvement. Our benchmark provides a reference point for measuring progress in the field of OCT diagnosis with MLLMs.
[CV-37] Inside-JEPA: From Prediction to Relational Grounding in Latent World Models
链接: https://arxiv.org/abs/2610.31194
作者: Hossein Resani,Javen Qinfeng Shi
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Joint-embedding predictive architectures (JEPAs) learn world models in latent space, enabling efficient planning but leaving their states opaque. We introduce Inside-JEPA, which places relational channels for ordered entity pairs inside the JEPA latent state and shows that an inverse-dynamics loss can ground these channels without relation labels when its head reads them, either alone or as part of the full state. On Something-Something V2, latent prediction alone performs no better than a random relation head, while inverse-dynamics supervision that recovers actions from channel changes improves relational AUROC by .065-.103 across ten settings with annotated boxes, by .081 on RH20T robot episodes with held-out tasks, and by .050-.061 across five relations using off-the-shelf detection and tracking. An inverse head on entity latents decodes actions more accurately but barely improves relational grounding; shuffled labels, two label-free objectives, and action labels on only 5% of the pairs also fail to ground the channels. Given an action, predicted relational changes approach the quality of observed future relations (AUROC .599 vs .621). For one-step relational goals, a supervised readout of the 16 channels ranks the correct action among the top five of 65 in 43% of cases, compared with 24% for prediction-only latents and 17% for a frequency prior.
[CV-38] RaSFNet: Ray-aligned and Scan-aware Fusion Network for LiDAR Semantic Segmentation
链接: https://arxiv.org/abs/2610.31186
作者: Chuang Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:LiDAR point cloud semantic segmentation is a fundamental component of 3D perception for autonomous driving and robotics. Projection-based methods map 3D point clouds into 2D representations for feature extraction using 2D networks. However, aggregating multiple points during projection can obscure geometric differences within each frustum, while local feature extraction limits the modeling of spatial relationships across scanning regions. Moreover, proximity in the projected representation does not necessarily imply proximity in 3D space, and interactions between point and projected features may mix features from different ranges, reducing semantic discriminability. To address these challenges, we propose the Ray-aligned and Scan-aware Fusion Network (RaSFNet), which incorporates point cloud geometry and LiDAR scanning characteristics into projection-based semantic segmentation. Specifically, the Ray-Aligned Geometric Distribution Encoding (RAGDE) module characterizes point distributions within each frustum using a ray-aligned coordinate system. To model spatial context, a Scan-Topology-Aware Mamba (STAM) block organizes frustum features according to the LiDAR azimuth sampling pattern and captures long-range dependencies across scanning regions. In addition, a Range-Constrained Point-Frustum Fusion (RCPFF) mechanism uses shared range information to guide feature exchange between point and frustum representations while maintaining spatial correspondence. RaSFNet achieves mIoU scores of 70.2% and 77.7% on the SemanticKITTI and nuScenes validation sets, respectively, indicating competitive segmentation performance compared with the evaluated methods.
[CV-39] Memory before Goal: Role of Pre-History Memory for Humanoid Loco-Manipulation
链接: https://arxiv.org/abs/2610.31166
作者: Chenyuan Chen,Ran Ding,Lechen Zhang,Zewen He,Yoshihiko Nakamura
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 8 pages, 4 figures, 5 tables. Submitted to IEEE Robotics and Automation Letters (RA-L). Project page: this https URL
Abstract:Humanoid loco-manipulation in dynamic scenes is non-Markovian: cues vanish as objects change or are hidden, or leave the view as the robot turns and walks, yet whole-body vision-language-action (VLA) policies see only the present. We present WholeBodyMem, an object-guided perceptual memory that keeps the pre-history of object events seen before the robot acts on its goal. SAM3 tracks the task objects, and their events causally select key frames within an adaptive budget, which the action expert reads by cross-attention together with object tokens encoding appearance, geometry and time. We also contribute HuMeGo, a humanoid memory benchmark with six simulated and six real-world Unitree G1 tasks. Keeping the right frames rather than more frames, WholeBodyMem surpasses memoryless and uniform-memory baselines, raising its base policy’s success from 29.1% to 72.6% in simulation and from 36.7% to 88.3% on the real robot, while also improving the standard SIMPLE tasks (70.2% to 86.9%).
[CV-40] HiSTEVE 1.0: Harnessing Human-Interface VLA Agents for Open-World Games
链接: https://arxiv.org/abs/2610.31162
作者: Zecong Tang,Zian Guan,Yuncheng Guo,Zhanqiu Zhang,Yiwen Guo,Bo Dai
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Building virtual embodied agents that operate in open-ended games like human players is a long-standing goal requiring both long-horizon autonomy and reliable native-interface control. As a procedurally generated open world, Minecraft provides a natural testbed for studying the integration of these capabilities. Existing agents typically provide either native visuomotor control without long-horizon autonomy or high-level planning executed through environment APIs or predefined skills. To bridge this gap, we propose HiSTEVE, a hierarchical VLA agent with a capability-aware agent harness. At the orchestration level, we equip the harness with structured memory for tracking progress and a VLM planner for selecting subtasks within the VLA’s capabilities. To reliably determine subtask completion, we further introduce a policy-coupled task completion state detector that conditions on VLA policy context to signal candidate completion. At the control level, a single VLA executes all subtasks with vision-only perception and native keyboard-and-mouse actions. To mitigate interference between heterogeneous GUI and non-GUI interactions, we introduce regime-specific RMSNorm branches and separate mouse-movement token embeddings. We also train the VLA through world knowledge injection, imitation learning, and reinforcement learning with environment feedback to improve reliability and recovery. HiSTEVE achieves state-of-the-art atomic-task performance, scoring 85.0% on MineStudio Bench and a 37.4% average success rate on OpenHA, while also completing long-horizon tasks such as crafting an iron pickaxe from scratch like human players.
[CV-41] WildFakeToolBench: Benchmarking AI-Generated Image Detectors with Web-Sourced Image Processing
链接: https://arxiv.org/abs/2610.31152
作者: Mingjie Li,Yukun Jiang,Yicong Tan,Yang Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Cryptography and Security (cs.CR)
备注:
Abstract:As AI generated images are getting popular these days, discussions on such generations are widely existed in current public forums, software repositories, and photography tutorials. Moreover, with the help of LLM-assisted code translation, malicious actors can adopt such information to evade detectors without specialized knowledge or local GPU execution of the resulting tools. To evaluate such risks, we introduce \textscWildFakeToolBench, a benchmark suite of 156 executable image-processing tools derived from 220 public sources across 142 domains. Each tool retains its source evidence and passes syntax, execution, and output-distortion checks without using feedback from the evaluated detectors. On 105 generated images from five generators, we evaluate six state-of-the-art detectors with images processed by the tools in our collected bench to assess their effectiveness against such in-the-wild risks. The strongest individual tools achieve attack success rates of 25.25% – 100% across detectors, exceeding the strongest of six selected image-processing baselines for each detector. Evaluating subsets of an eight-tool pool yields worst-case coverage above 80% for every detector. Effectiveness is sparse and detector-specific: high coverage across candidates does not imply reliable one-shot evasion. These findings not only highlight the potential of such web-sourced risks, but also show the potential on establishing AI-image detector robustness evaluation framework based on those in-the-wild information.
[CV-42] RankAlign: Vision-Language Representation Alignment via Rank Transformation NEURIPS2026
链接: https://arxiv.org/abs/2610.31145
作者: Enzhe Zhao,Marco Fiorucci,Lamberto Ballan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026
Abstract:While the Platonic Representation Hypothesis posits that vision and language models converge toward compatible latent geometries, recovering class-level correspondences between unpaired modalities without cross-modal supervision remains challenging. Existing relational matching methods compare within-modality similarity kernels directly, although the same semantic structure may be represented through different numerical parameterizations across modalities. We propose RankAlign, a rank-based normalization strategy for class-level unpaired vision-language matching. RankAlign transforms each within-modality similarity kernel into an ordinal relational representation before permutation matching, while leaving the correspondence space and matching solver unchanged. The resulting representation is invariant to strictly increasing reparameterizations of similarity values while preserving the ordinal structure used for matching. We provide exact-recovery and stability guarantees that characterize correspondence recovery under both exact and partial ordinal agreement. Experiments across diverse frozen vision and language encoders and multiple datasets demonstrate improved and more stable class-level correspondence recovery, including under label corruption and domain shift.
[CV-43] GM-GReFEL: Geometry-Guided Spatiotemporal State-Space Learning for Robust Dynamic Facial Expression Recognition
链接: https://arxiv.org/abs/2610.31140
作者: Yudhistira Arditya Pratama,Yi-Zeng Hsieh
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Recognizing facial expressions in unconstrained video suffers from two coupled problems: self-attention grows quadratically in cost with clip length, and crowdsourced labels are noisy. We present GM-GReFEL, a spatiotemporal state-space network with a geometry-aware decision stage. Its Modulated GroupMamba backbone replaces a spatiotemporal Vision Transformer. Six channel groups are scanned separately, four over the frame plane and two over time, so the cost grows linearly with sequence length ( \mathcalO(N) ); after each layer, a Channel Affinity Modulation gate mixes the groups again. On top of the backbone, a GReFEL reliability module turns the normalized entropy of each prediction into a mixing weight between the classifier output and a vote from trainable geometric anchors, so that facial geometry counts for more when the network is unsure. The model is initialized by static-to-dynamic weight transfer and trained with a Triple Loss. In five-fold cross-validation, the best fold scores 66.57%/76.46% UAR/WAR on DFEW, 0.3 points or less below the strongest compared method, and 45.66%/62.27% on MAFW, ahead of all compared methods; on FERV39k the mean over five splits, 42.30%/52.66%, is competitive. The five-fold means are 64.07%/74.54% on DFEW, on par with recent large pre-trained models, and 38.90%/54.37% on MAFW, below the best of them. A paired t -test on DFEW indicates a modest benefit of temporal modeling (+1.61 UAR points, one-sided p=0.031 ). Attention stays on plausible facial regions, and latency is close to a ViT-B/16 baseline (47.4 vs.\ 46.8 ms per clip). Code: this https URL.
[CV-44] Pos4CHD: A Positive-Free Mask-Guided Counterfactual Synthesis Framework for Fetal Congenital Heart Defect Analysis
链接: https://arxiv.org/abs/2610.31139
作者: Wenjie Xuan,Yapeng Li,Hang Su,Chen Cheng,Juhua Liu,Sheng Zhao,Ziming Huang,Bo Du
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Submitted to IEEE TMI, under review
Abstract:The scarcity of positive cases remains a major bottleneck in developing automated diagnostic systems for fetal Congenital Heart Defects~(CHDs), particularly for rare subtypes. Although normal cases are accessible with effort, positive cases are hard to collect due to low incidence rates and privacy concerns, causing severe class imbalance or absence of positive training samples. Existing works resort to synthesizing positives with conditional generative models, but they are task-specific and require positive training samples, hindering application for fetal CHDs. To this end, we propose a positive-free mask-guided counterfactual generation framework for fetal echocardiography, namely Pos4CHD, which enables realistic, high-quality synthesis of pseudo-positive samples from semantic masks, with only normal samples at hand. Technically, we first develop FE-Gen, a mask-guided generative model tailored for fetal heart ultrasound synthesis that mitigates histogram drift and enhances image-mask spatial alignment. Building on FE-Gen, Pos4CHD employs a mask-to-image generation paradigm for CHD synthesis. It first manipulates semantic masks to simulate fetal-heart structural anomalies and then renders anatomically consistent images. Extensive experiments on in-house datasets, together with expert evaluations, demonstrate the visual fidelity~( -12.97 FID) and anatomical consistency~( +3.96% DSC) of synthesized images. Further validations across diverse real-world CHD diagnosis tasks exhibit promising results and consistent increments, including zero-shot classification~(achieving 93.25% ACC on EbA and 72.92% ACC on TAPVC), lesion detection~( +6.95% mAP50 on VSD), and anatomy segmentation~( +10.49% DSC on fetal heart) compared with baselines. Codes and models are publicly available on \hrefthis https URLGitHub.
[CV-45] WildDepth-v2: A Synchronized RGB–LiDAR–Radar Dataset and Benchmark for Metric Depth Motion and Tracking of Wildlife
链接: https://arxiv.org/abs/2610.31123
作者: Muhammad Aamir,Sangyun Shin,Naoya Muramatsu,Matthew Wijers,Amir Patel,Andrew Loveridge,Andrew Markham
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Measuring animal distance and motion in metric units underpins abundance estimation, morphometrics, and movement ecology, yet monocular video lacks metric scale and wildlife datasets lack metric ground truth. We present WildDepth 2, extending the synchronized RGB-LiDAR WildDepth dataset in three directions. (i) Imaging radar: 37 RGB-LiDAR-4D-radar sequences (19,716 frames) with per-point Doppler velocity and a measured error budget, to our knowledge the first imaging-radar wildlife recordings. (ii) Benchmarking: a frozen, checksummed hidden test set for RGB-only distance estimation and RGB+sparse-LiDAR depth completion, evaluated separately on animal and background pixels. (iii) Motion and tracking: metric 3D trajectories obtained by lifting 2D tracks using LiDAR, radar, or monocular depth. Across 48,723 animal frames, object-level radar calibration improves monocular depth over scene-level LiDAR calibration without requiring LiDAR at inference (AbsRel 0.120 vs. 0.198), although animal-level LiDAR remains more accurate (0.052). Across 431 tracks, once-per-track radar scaling reduces 3D position error four-fold (2.20 to 0.58 m) and distance-sampling detection-function bias from -13% to -2%, matching once-per-track LiDAR calibration. Transferring radar Doppler onto LiDAR produces motion masks 3.7 times denser than radar alone at 72% precision, reduces moving-animal LiDAR accumulation error by 37%, and estimates radial speed (r = 0.60), albeit with approximately 50% underestimation.
[CV-46] Where Do Long Captions Fail? Position-Aware Diagnosis and Reinforcement Learning for Detailed Image Captioning NEURIPS2026
链接: https://arxiv.org/abs/2610.31111
作者: Yuanze Hu,Zhichao Yang,Junwei Jing,Xin Yang,Wenxuan Liu,Tinghai Zhang,Jia Fu,Hongzhi Zhang,Xiangyu Wu,Ruiming Tang,Tingting Gao,Han Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026. 38 pages, 3 figures; includes references, appendices, and the paper checklist
Abstract:Detailed image captioning is usually framed as an amount problem: a model should mention more visual facts while avoiding hallucination. However, autoregressive VLMs generate captions as ordered sequences, and later generations may be produced under weaker or noisier visual grounding. We therefore argue that long-form captioning should be studied not only by what facts are mentioned, but also by where they appear. We introduce a position-aware rubric representation that decomposes dense human references into atomic visual claims, verifies generated captions against those claims, and induces a quartile-wise composition of supported, hallucinated, and generic content. This lens reveals two qualitatively different failures in long captions: models tend to produce unsupported specificity in the middle, while the final portion often escapes into vague, non-evidential language rather than simply accumulating more hallucination. A same-reference human control does not exhibit these patterns; additional human audits examine the visual validity of the induced labels. Motivated by this diagnosis, we propose PAGER (Position-Aware Grounded Evidence Reinforcement), a GRPO training objective with separate signals for supported fine-grained middle evidence, evidence-bearing tail continuation, claim coverage, and content-aware length control. The result is a closed loop: the same ordered claim representation identifies where long captions fail and supplies the position-specific reward needed to repair those failures. Experiments on detailed-caption benchmarks show that PAGER improves position-wise grounding and pairwise caption preference while preserving competitive global caption quality.
[CV-47] DRAFT: Decoding and Reconstructing 3D Room Layouts from fMRI and Camera Trajectories
链接: https://arxiv.org/abs/2610.31108
作者: Chi-Wei Lee,An-Kai Chen,Cheng-Han Chiang,Yi-Chun Chang,Po-Chih Kuo
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 42 pages, 7 figures, 18 tables
Abstract:While functional magnetic resonance imaging (fMRI) decoders reconstruct images and 3D objects, recovering the metric geometry of rooms from a moving viewpoint remains an open challenge. We introduce, to our knowledge for the first time, the task of reconstructing 3D layouts of held-out scenes using only the known camera trajectories and the fMRI signals of participants watching indoor walkthroughs, without any visual frames. The task is hard because of the temporal mismatch between rapid visual changes and blurred fMRI signals. Our method, DRAFT (Decoding and Reconstructing 3D room lAyouts from fMRI and Trajectories), extracts a compact layout representation from the early visual cortex and fuses this neural code with the geometric free space derived from camera motion. To evaluate it, we propose layout r, the Pearson correlation between predicted and ray-cast log depths. On ten held-out scenes (20 participants), DRAFT reaches a layout r of 0.480, outperforming the camera path alone (0.446) and seven fMRI decoders (at most 0.429). DRAFT also works as a plug-in that improves all seven decoders without retraining, raising their layout r from 0.171-0.429 to 0.487-0.493. Brain activity and camera motion can thus recover the metric layout of unseen rooms, a first step toward reconstructing the 3D spaces people see.
[CV-48] I2-KAN: Interpretable Intra- and Inter-Modal Multi-Pathway Representation Learning for Multimodal Sentiment Analysis
链接: https://arxiv.org/abs/2610.31087
作者: Xiaoting Xie,Feng Qiao,Xianwei Zhang,Hailong Chen,Kaiwen Zheng,Joemon M. Jose,Xuri Ge,Yujun Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal sentiment analysis (MSA) aims to infer sentiment by integrating linguistic, acoustic, and visual information. However, many fusion methods integrate modality-specific representations and cross-modal interactions into joint representations, making it difficult to determine whether predictive utility arises from a modality itself or from its interactions with others. More fundamentally, a modality’s standalone predictive utility may not faithfully reflect its interaction-conditioned predictive utility. To address this issue, we propose I ^2 -KAN, an interpretable multi-pathway framework that explicitly maintains intra-modal and pairwise inter-modal information in separate structural pathways. Each pathway is refined by a Task-Decoupled Linear-Gated Denoising Bottleneck (TDL-GDB) to learn compact and task-relevant representations, followed by structured Kolmogorov-Arnold Network fusion that retains explicit pathway associations and supports multi-level interpretation. Extensive experiments on CMU-MOSI, CMU-MOSEI, and CH-SIMS v2 demonstrate strong overall performance. Ablation and interpretability analyses consistently show across the three benchmarks that low standalone predictive utility does not necessarily imply low interaction-conditioned predictive utility: a modality that appears weak in isolation can remain informative through pairwise interactions. These findings highlight the interaction-dependent nature of multimodal utility and suggest that modality utility should be characterized from both standalone and interaction-conditioned perspectives.
[CV-49] A General-Purpose Dual-Path Interaction Framework for Multi-View Medical Image Segmentation: Alleviating Data Volume Dependency in Small-Sample Regimes
链接: https://arxiv.org/abs/2610.31065
作者: Shuhuan Wang,Yuanyuan Cui,Pengju Wang,Huimin Guo,Ling Yue,Yin Gu,Yanchao Zhang,Wei Feng,Yunhe Liu,Wei Qian,Mingjun Xu,Shuang Liu,He Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages, 5 figures
Abstract:The performance of medical image segmentation depends on a large amount of annotated data, and most segmentation models experience a significant drop in performance in low-sample-size scenarios. Existing semi-supervised and weakly supervised methods rely on specific network architectures, while traditional multi-view-assisted segmentation methods require strictly paired samples, which limits their clinical applications. To address these issues, we propose DPIF, a general-purpose plug-and-play dual-path interaction framework compatible with arbitrary encoder-decoder segmentation networks, which leverages cross-view anatomical complementarity to reduce annotation reliance. DPIF first performs Sinkhorn optimal transport-based cross-view distribution alignment to establish global semantic correspondence without sample-level pairing. It then adopts conditional mutual information-guided feature decoupling to enable bidirectional knowledge transfer only within the shared semantic space, suppressing view-specific noise. Finally, persistent homology-based topological regularization encodes anatomical invariants as unsupervised priors to mitigate overfitting. Experiments were conducted using 24 baseline models on a private multicenter echocardiography dataset and two public datasets. The results show that DPIF-enhanced models outperform single-view baseline models when trained on comparable amounts of data. Notably, most dual-path variants trained on 40% - 60% of the data achieve performance comparable to that of single-view models trained on the full dataset. This demonstrates that DPIF provides a general-purpose and scalable solution for data-scarce multi-view medical image segmentation problems.
[CV-50] ResGVC: Residual Coding for High-fidelity Generative Video Compression
链接: https://arxiv.org/abs/2610.31039
作者: Ding Ding,Daowen Li,Ying Chen,Kai Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Generative video compression (GVC) methods usually employ decoder-side diffusion models to denoise reconstructed signals based on generative priors. Besides, most of these methods perform denoising in latent space rather than directly in pixel domain. In this paper, we propose ResGVC, a residual coding framework for GVC that guides denoising toward original pictures by transmitting residual signals. ResGVC first encodes the original pictures within a learned video codec, and then decomposes the distortion between the codec reconstruction and the original pictures into a sequence of residual segments, where each residual segment is acquired at each denoising step. Residual segments can further be compressed by a novel residual codec with a diffusion transformer as an entropy model. Unlike prior methods, ResGVC introduces explicit source constraints throughout denoising rather than relying on decoder-side generation alone. The diffusion-based entropy model also enables more effective distribution modeling capability based on the powerful generative priors. Moreover, denoising directly in pixel space avoids the irreversible information loss caused by latent representation bottlenecks. Experimental results show that, compared with the previous state-of-the-art method, ResGVC achieves 63.7% and 83.8% BD-rate reductions in terms of LPIPS and PSNR, respectively, while also delivering superior visual quality.
[CV-51] Revisiting Distribution Matching for State-to-Vision Policy Distillation in Laparoscopic Robotic Manipulation
链接: https://arxiv.org/abs/2610.31010
作者: Mingyu Sheng,Jianan Fan,Dongnan Liu,Guoyan Zheng,Ron Kikinis,Weidong Cai
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 3 figures, 2 tables
Abstract:End-to-end control policies trained via Reinforcement Learning (RL) have emerged as an alternative to conventional surgical robotic control approaches, with recent studies exploring state-based policies trained on privileged state observations. However, such approaches rely on precise physical states that are difficult to obtain reliably in surgical environments. Vision-based RL provides an alternative by learning manipulation policies directly from RGB observations, avoiding explicit state estimation and manually engineered perception pipelines while enabling policies to acquire task-relevant visual representations. Nevertheless, obtaining vision-based policies via pure RL methods for surgical manipulation remains challenging. Thus, imitation learning (IL) becomes another solution, yet it requires a large number of expert demonstrations, which are difficult to obtain in minimally invasive surgery. In this work, we propose a framework for state-to-vision policy distillation in laparoscopic robotic manipulation. Specifically, a state-based policy with privileged observations is employed as the teacher, and a vision-based policy learns from RGB observations and robot proprioceptive information, which requires less expert demonstrations than IL and introduces superior performance than RL methods. We conduct extensive experiments on various surgical manipulation tasks in the simulated environment. Experimental results show that, in the context of laparoscopic robotic manipulation, training vision-based policies with general-purpose RL methods is challenging, whereas our proposed state-to-vision policy distillation leads to more consistent and stable performance across tasks.
[CV-52] Diagnostic Loss: Asymmetric Supervision for Radiology Report Generation
链接: https://arxiv.org/abs/2610.30980
作者: Ignacy Kolton,Wenxuan Li,Zongwei Zhou,Jarosław B. Ćwikła,Arkadiusz Sitek,Pedro R. A. S. Bassi,Szymon Płotka
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 26 pages, 6 figures, 7 tables. Code available at this https URL
Abstract:Automated radiology report generation requires models to do more than produce fluent clinical narratives: they must explicitly state supported diagnoses while avoiding unsupported ones. However, existing medical vision-language models (MedVLMs) are typically trained with token-level maximum likelihood objectives, which prioritize text agreement over diagnostically critical terms. As a result, models can miss key diagnoses and introduce wrong ones. We identify this as an objective misalignment between conventional language modeling and the asymmetric diagnostic requirements of radiology reporting. To encode these requirements, we introduce Diagnostic Loss (DLoss), a differentiable, architecture-agnostic objective that operates directly on decoder token probabilities. DLoss separately aggregates probability mass over vocabularies associated with diagnoses supported by the reference report and with unsupported alternatives. Using asymmetric pooling across token positions in the report, it encourages supported diagnoses to receive sufficient probability in at least one position while suppressing spurious probability spikes for unsupported diagnoses across all positions during training. The loss can be applied to existing 3D MedVLMs without modifying their vision backbones, language decoders, or inference procedures, and introduces no additional inference-time computation. We train and evaluate DLoss on 50,000 abdominal CT scans across several 3D MedVLM architectures on both in-distribution and out-of-distribution report-generation benchmarks. Across these settings, DLoss improves diagnostic faithfulness while maintaining competitive performance on conventional language-generation metrics, showing the value of asymmetric sequence-level supervision for radiology report generation. The code for DLoss is available at this https URL.
[CV-53] GraphCAD: Geometry-Aware Diffusion on Edge-Adjacency Graphs for CAD Generation
链接: https://arxiv.org/abs/2610.30974
作者: Wenhao Zheng,Xianggen Liu,Li Yue,Jiancheng Shi,Jiancheng Lv
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Generative modeling for Computer-Aided Design (CAD) remains fundamentally challenged by the need to capture the intricate interplay between discrete topological connectivity and precise spatial geometry inherent in Boundary Representations (B-reps). Existing generative approaches largely formulate CAD construction as a sequence modeling problem, but consequently lack an explicit topological inductive bias. In this work, we present GraphCAD, a geometry-aware generative framework that centers modeling on the Edge-Adjacency Graph (EAG), a novel graph-based CAD representation tailored for generative tasks. GraphCAD adopts a hybrid graph diffusion formulation to jointly model the discrete and continuous components of EAG and introduces a geometry-aware denoising mechanism that renders node attributes into spatial point sets. These point sets are encoded by an SE(3)-equivariant backbone and fused into the node representations, thereby grounding the generative process in physical geometry. To support this study, we construct EAGBench, a large-scale dataset of engineering CAD models. Extensive experiments show that GraphCAD can generate complex CAD models with more than 50 faces and achieves state-of-the-art performance in geometric fidelity and distributional alignment compared with sequence-based baselines.
[CV-54] Attention-guided classification and tumor localization of combined hepatocellular-cholangiocarcinoma
链接: https://arxiv.org/abs/2610.30944
作者: Xiaowen Liang,Hélène Cazier,Aurélie Beaufrère,Miguel Albuquerque,Valérie Paradis,Cédric Wemmert
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Combined hepatocellular-cholangiocarcinomas (cHCC-CCA) are a type of cancer that share components of two other primary liver cancers: hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA). Accurate diagnosis of cHCC-CCA, of which the behavior may resemble that of either HCC or CCA, remains a challenge due to strong intratumoral heterogeneity. Since the prognoses differ between liver cancer types, it is necessary to improve their morphological characterization and recognition in order to accurately identify cHCC-CCA. In this work, we use whole slide images (WSIs) of the two main types of liver cancer (HCC and CCA) and cHCC-CCA, stained with Hematoxylin-Eosin-Safran (HES), to train a classification model that can distinguish between these subtypes and provides a potential solution for localizing different types of tumors in the combined liver cancer images.
[CV-55] ICGF-CXRMate: Image-Conditioned Clinical Context Compression and Gated Fusion for Chest X-ray Report Generation
链接: https://arxiv.org/abs/2610.30928
作者: Vu Minh Tran,Thuc Huynh,Son Thanh Luu,Doanh C. Bui
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Auxiliary patient data can improve chest X-ray report generation, but directly concatenating heterogeneous clinical information with visual tokens produces long, variable-length decoder prompts and may expose the model to redundant or image-irrelevant context. We propose ICGF-CXRMate, an image-conditioned clinical context compression and gated fusion framework. Learnable queries are first conditioned on the current chest X-ray representations and then attend to the patient-data sequence, compressing it into a fixed number of clinical context tokens. Gated cross-attention selectively integrates these tokens into the visual stream, while the compressed tokens are retained in the decoder prompt as a direct clinical pathway. Experiments on linked MIMIC-CXR and MIMIC-IV-ED examinations show that ICGF-CXRMate improves report quality while reducing the average prompt length. In Stage~2, it improves RadGraph-F1 from 24.00 to 28.86, CXR-BERT similarity from 56.86 to 62.80, and BLEU-4 from 5.14 to 7.96 compared with CXRMate-ED, while reducing the average prompt length from 373.89 to 288.37 tokens, corresponding to a 22.9% reduction. After self-critical sequence training, ICGF-CXRMate outperforms CXRMate-ED on six of eight report-quality metrics in Stage~3. Ablation studies confirm the complementary contributions of image-conditioned compression, gated fusion, and direct retention of the compressed clinical tokens.
[CV-56] Making MLLM Latents Matter: Factor-Aware Predictive Learning for Cinematic Video Generation
链接: https://arxiv.org/abs/2610.30926
作者: Songlin Yang,Zhe Wang,Xiaotong Zhao,Xi Wang,Ran Zhang,Shuai Li,Taiyi Wu,Songlian Li,Eric Liu,Alan Zhao,Anyi Rao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Cinematic video generation requires both temporal inference from limited inputs (i.e., sparse rather than frame-by-frame guidance) and factor-aware control of character, scene, action, and camera. Multimodal large language models (MLLMs) offer useful priors for both capabilities, but their representations are not automatically actionable for video generation: a Video Diffusion Transformer (DiT) may bypass them through native conditioning pathways, and using them does not ensure either factor-aware control or generation-compatible temporal prediction. We propose Factor-Aware Predictive Learning (FAPL), which establishes effective MLLM latent use as a shared prerequisite and then targets the two capabilities with complementary learning signals. First, latent-only generator grounding trains the DiT with a frozen MLLM and connector, strengthening the DiT’s use of MLLM latents for generation. Second, for factor-aware control, paired videos specify a target-factor change and non-target preservation, and factor-span replacement trains the DiT to realize this relationship in its output. Third, for temporal inference, text-only and text-plus-complete-video observations share a video target, and the trained DiT is then frozen while the MLLM and connector are optimized through it. This final stage trains limited-input representations for generation while retaining complete-observation reconstruction as a complementary training mode. Extensive experiments demonstrate our consistently better performance on three cinematic video generation tasks (i.e., sparse-to-video, factor-aware editing, and multi-shot generation) over baselines and validate FAPL’s contributions to MLLM feature use, temporal control, and factor coordination.
[CV-57] Watch and Learn: Teaching Visual In-Context Learning by Removing Textual Shortcuts
链接: https://arxiv.org/abs/2610.30913
作者: Nimrod Shabtay,Asaf Yehudai,Avihu Dekel,Eli Schwartz,Raja Giryes
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-language models (VLMs) struggle with in context learning compared to their textual counterparts. Our preliminary study shows that VLMs rely on textual shortcuts and underutilize the visual demonstrations. We therefore hypothesize that VLMs are insufficiently trained to perform visual in-context learning (vICL). To this end, we construct MIME, a task-agnostic vICL training set without textual shortcuts. We ensure the instruction does not state the rule linking the demonstration images to their answers, so the model must infer it from the visual demonstrations. Fine-tuning on MIME improves vICL accuracy by 8.1 to 18.0 points across three VLMs, preserves single-image performance, and improves multi-image and text only in-context learning. In contrast, stating the rule in the training data sharply reduces this gain. We further propose DTAD (Demonstration Text Attention Dropout), which randomly withholds demonstration labels during training so that the VLM must infer the rule from the demonstration images. While achieving comparable vICL accuracy to fine-tuning on MIME, DTAD yields more consistent attention to the demonstration images. On TrueMICL, Qwen3-VL-8B trained on MIME matches frontier models such as Claude Opus 5, and surpasses them with DTAD. Finally, adding MIME to the instruction-tuning mix of LLaVA 1.5 13B more than doubles its vICL accuracy without hurting single-image abilities, whereas a same-size multi-image set improves it only slightly. These results suggest that vICL data should be a standard component of VLM training.
[CV-58] Query-Time Calibration of CLIP Features for Open-Vocabulary 3D Segmentation
链接: https://arxiv.org/abs/2610.30908
作者: Dominik Engel,Yousef Rajeh,Leon Sick,Ivan Viola
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Lifting vision-language features such as CLIP into 3D Gaussian reconstructions enables open-vocabulary segmentation through natural-language queries. Although CLIP represents a broad range of semantic concepts, the lifted features of individual scenes can exhibit strong directional concentration, while raw cosine similarity does not account for the scene-specific feature distribution. We show that substantial segmentation gains can be obtained by changing how a fixed lifted feature field is compared with text. Our method partially subtracts a shared scene direction from visual features and text prototypes and whitens text prototypes using a fixed reference vocabulary of generic terms. It then uses soft clustering for multiclass prediction and CSLS with a relevance margin against the reference vocabulary for single-query prediction. Beyond reducing directional concentration, scene centering improves class-relative feature separation in every evaluated scene. Across seven lifting methods and four datasets, the resulting query-time calibration consistently improves mean intersection over union in both multiclass and single-query segmentation. These gains require no retraining, source-image re-encoding, or feature re-lifting, demonstrating the value of scene-conditioned comparison for making better use of existing lifted representations.
[CV-59] When Annotation Patterns Are Learned: Skin-Tone Dependence in Dermatology Concepts
链接: https://arxiv.org/abs/2610.30901
作者: Y. Li,M.Bigliardi-Qi,S. Bedros,P. Bigliardi,N. Papanikolopoulos
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Concept bottleneck models (CBMs) let clinicians inspect and correct skin disease diagnoses through clinical concepts, but only if the concept annotations are reliable. We therefore audit the reliability of SkinCon, a widely used dermatology concept dataset, across skin tones. Within the same disease and with near duplicates removed, erythema is annotated far less often on dark than on light skin, with an odds ratio of 0.17. In contrast, hyperpigmentation and hypopigmentation are annotated more often, with odds ratios of 6.2 and 4.5, respectively. We refer to this pattern as tone-dependent concept annotation. An evidential concept model trained on these annotations learns the same pattern and reproduces it on unseen images. Such a model may agree with potential erythema omissions rather than contradict them. Consistent with this, three label error detectors underperform a disease-only baseline at finding erythema that dermatologist captions mention but SkinCon labels as absent. The three detectors are based on confident learning, training dynamics, and the model’s own evidential uncertainty. Concept annotations should thus be audited within each subgroup before explanations built on them can be trusted.
[CV-60] HimmBO: High-Dimensional Interactive Model-Merging Bayesian Optimization for Image Generation
链接: https://arxiv.org/abs/2610.30875
作者: Jinyuan Mao,Zhihao Shen,Chenxi Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:Fine-tuning has become a common way to customize text-to-image models, leading to community collections of over 100K lightweight adapters. Adapters sharing a base model can be merged with continuous coefficients. This turns a large collection into a high-dimensional continuous design space, which recent work explores interactively using preferential Bayesian optimization (PBO). However, existing approaches are limited to a few dozen preselected adapters and require expensive image generation in every interaction round. We introduce HimmBO, a two-stage method that scales preference-driven adapter merging to much larger collections while reducing image-generation cost. Its screening stage recovers the active adapter set through level-set estimation over pre-generated images, requiring no image generation. The refinement stage then optimizes merge coefficients in the recovered low-dimensional subspace. Both stages use a style-aware kernel, allowing preference information to transfer between visually similar adapters. In simulated-user image-matching experiments, HimmBO achieves higher final similarity than all comparisons while reaching a given similarity with under one tenth of the image generations required by prior work. Its performance remains stable as the collection grows to 1080 adapters, while competing methods degrade substantially.
[CV-61] Verus-Insight: Towards Explainable Artifact Detection in AI-Generated Videos
链接: https://arxiv.org/abs/2610.30866
作者: Zhuoran Zhang,Yang Shi,Qixun Wang,Tengfei Liu,Yuqi Tang,Xuanyu Zhu,Yizhuo Jia,Jingyun Hua,Shixuan Liu,Xianghua Ying,Haotian Wang,Yuanxing Zhang,Pengfei Wan,Wenjing Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 35 pages
Abstract:Recent advances in video generation have made synthetic videos increasingly realistic, placing greater demands on reliable AI-generated video detection. Existing artifact-based detectors typically identify generation artifacts as evidence for source prediction, yet they may either fail to recognize subtle artifacts or reach the correct source decision without faithfully identifying the evidence underlying it. In this work, we systematically investigate how to design and train effective artifact detectors and formulate artifact detection as a structured process comprising three complementary capabilities: Inspection Planning, which determines what should be examined; Evidence Verification, which verifies whether the inspected content exhibits genuine generation artifacts; and Evidence Aggregation, which integrates the verified evidence into the final source decision. Building on this formulation, we introduce Verus-Insight, trained with human-anchored full-process supervision and progressive capability training to strengthen both fine-grained artifact detection and video-level detection. We further introduce Verus-Bench, a benchmark with fine-grained human annotations of perceptible artifacts for evaluating artifact detection. Extensive experiments across diverse benchmarks demonstrate that Verus-Insight achieves strong performance in both fine-grained artifact analysis and AI-generated video detection, with substantial gains in recovering human-perceived artifacts. Further analyses reveal that the three capabilities are complementary yet distinct, and that Verus-Insight more reliably converts relevant inspection targets into supported artifact evidence without increasing spurious artifact claims.
[CV-62] Cross-Spectral Feature Fusion for Multispectral Document Binarization
链接: https://arxiv.org/abs/2610.30863
作者: Yasas Savinda,Maheshi Buddhinee Dissanayake
类目: Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP)
备注:
Abstract:Historical document binarization remains challenging when faded ink, bleed-through, and background degradation exhibit wavelength-dependent behavior across visible and near-infrared observations. This study investigates whether adaptive cross-spectral feature fusion and targeted structural guidance can improve multispectral document binarization beyond fixed fusion and generic attention mechanisms. Four registered SMADI bands at 500, 600, 700, and 1000 nm are processed using a multi-branch conditional generative adversarial network (cGAN). Visible, NIR, and structural feature streams are fused at encoder depth L = 3 using spatial- and channel-wise softmax gating, while auxiliary edge maps derived from selected VIS-NIR band pairs provide targeted structural guidance. Nine controlled fusion and structural configurations are evaluated using five-fold document-level cross-validation, DIBCO-standard metrics, and Holm-corrected statistical testing. The F3 + NIR edge-guided configuration provides the strongest overall performance among the evaluated models, achieving mean F1 = 0.7916, p-FM = 0.8044, and IoU = 0.6654. Relative to the fixed Mid AVG baseline, it improves F1 by 0.0269, p-FM by 0.0360, and IoU by 0.0321, with all three gains remaining statistically significant after Holm correction. Mean NRM and DRD are also reduced, although these reductions are not statistically significant after correction. Targeted cross-spectral structural guidance provides an effective mechanism for regulating intermediate multispectral fusion, improving primary foreground-overlap performance while retaining the complete four-band observation. Subjects: Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP) Cite as: arXiv:2610.30863 [cs.CV] (or arXiv:2610.30863v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.30863 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Maheshi Dissanayake [view email] [v1] Thu, 22 Oct 2026 07:42:35 UTC (1,626 KB)
[CV-63] Direct Forcing: A Single-Stage Training Recipe for Few-Step Autoregressive Text-to-Video Generation
链接: https://arxiv.org/abs/2610.30861
作者: Haojun Xu,Jie Huang,Xin Lu,Mingchen Zhong,Zihao Fan,Linjiang Huang,Si Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Few-step autoregressive video generation enables low-latency streaming, yet prevailing training pipelines separate causal adaptation from distribution matching. Recent representation-based text-to-video methods retain this separation by starting from ODE-initialized checkpoints, even when the target distribution is defined directly by reference videos. We introduce Direct Forcing, a single-stage framework that uses reference videos both as paired examples of how a history continues and as a target population for self-generated rollouts. Causal Continuation Alignment learns paired spatiotemporal correspondence under reference histories, while Rollout Distribution Matching uses maximum mean discrepancy to align the representation distributions of self-generated and reference videos. Together, these complementary signals allow a pretrained video model to learn causal continuation and the target video distribution in a single training stage. Dynamic Gradient Balancing adapts the relative gradient contributions of continuation alignment and distribution matching during training. Starting from bidirectional Wan2.1-1.3B, Direct Forcing learns a four-step autoregressive generator in 250 updates without a separate initialization stage or online diffusion-score supervision, achieving 84.63 VBench Total and 86.01 Quality. Ablation studies further demonstrate the benefits of dynamic gradient balancing and reveal distinct representation preferences: deeper features benefit continuation alignment, whereas intermediate features yield better performance for rollout distribution matching.
[CV-64] CAVC: Convergence-Adaptive Audio-Visual Cache for Modality-Asymmetric Diffusion Generation
链接: https://arxiv.org/abs/2610.30854
作者: Lifan Jiang,Yibo Zhao,Senlin cheng,Bolei Xu,Xiaobo Li,Can Ye,Dehai Zhao,Deng Cai,Boxi Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Sound (cs.SD)
备注:
Abstract:Diffusion-based audio-visual generation methods typically perform joint denoising of audio and video latents throughout the entire sampling trajectory, repeatedly executing modality-specific denoising branches and cross-modal attention at every timestep. However, we observe that such uniform joint computation is not always necessary. The key is a modality-asymmetric recoverability: audio-side drift introduced by mid-stage caching can be substantially reduced by only a few terminal joint denoising steps, whereas video-side degradation caused by caching leads to persistent loss of motion and temporal details. This asymmetry suggests that inference acceleration should selectively cache recoverable audio-side computation while preserving full video refinement. Motivated by this observation, we propose \textbfConvergence-Adaptive Audio-Visual Cache (CAVC), a training-free modality-asymmetric caching framework. CAVC adopts a sandwich-structured computation allocation: after detecting that audio denoiser outputs enter a cache-tolerant regime, it caches and reuses the audio denoiser output to bypass audio denoising and cross-modal attention in the middle stage, while the video branch continues independent spatio-temporal refinement; a few joint steps are then restored at the end to reduce accumulated cross-modal drift and improve synchronization. Experiments across multiple public benchmarks demonstrate that CAVC achieves 1.10–1.35x wall-clock acceleration, improves several visual dynamics and temporal statistics, and maintains competitive audio fidelity and audio-visual alignment.
[CV-65] RECAP: Looking Once Is Not Enough for Vision-Language Reasoning NEURIPS2026
链接: https://arxiv.org/abs/2610.30834
作者: Zhaolu Kang,Tailong Luo,Chenxin Li,Yingjie He,Zhenyu Yu,Fengyu Zhou,Jiachen Qian,Lei Wei,Shuang Chen,Jiachen Li,Eric Hanchen Jiang,Rongchao Zhang,Zhengtao Yao,Hoi Leong Lee,Guansu Wang,Kaiyue Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026
Abstract:Long chain-of-thought reasoning improves vision-language models (VLMs), but it also exposes a temporal grounding failure: as generation unfolds, models may drift from image evidence while reinforcement learning still provides only a final-answer reward. We study this gap through visual revisit, a temporally localized reactivation of image evidence during reasoning. Across three VLM backbones, revisit peaks are predictive of correctness and causally linked to performance: masking them degrades accuracy more than masking non-peaks, while injecting revisit-like peaks into failed rollouts partially restores correct answers. Control analyses show that this signal is not explained by raw attention strength, uncertainty, logit margin, or hidden-state norm. We propose RECAP, a GRPO-compatible training method that turns detached revisit traces into a credit-assignment signal. RECAP uses revisit-conditioned advantage estimation, dynamic discounting, and gated value prediction to propagate final rewards through visually important reasoning steps, without extra rewards, additional rollouts, or inference-time changes. Across Qwen2.5-VL-7B, InternVL3-8B, and LLaVA-OV-7B, RECAP consistently improves over GRPO and VPPO, yielding +6.1 – +6.5 HallusionBench gains over the SFT base while preserving reasoning and text-only performance. Results on larger LoRA-tuned models, Chinese zero-shot benchmarks, visual perturbations, and human evaluation further demonstrate robust improvements in image-grounded reasoning.
[CV-66] One Evaluation Is Enough: Inversion-Free Image Editing with Trajectory Correction
链接: https://arxiv.org/abs/2610.30827
作者: Lifan Jiang,Taiyang Chen,Senlin cheng,Bolei Xu,Xiaobo Li,Can Ye,Dehai Zhao,Deng Cai,Boxi Wu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Real-image editing commonly relies on inversion and iterative sampling, while efficient one-step editors often stabilize a large update by estimating and averaging editing fields at multiple noise levels. We ask a stricter question: is a single temporal field evaluation sufficient for reliable inversion-free editing? Our answer rests on a previously underused duality in a variance-preserving (VP) one-step model: one source-target query at a shared noisy state simultaneously determines a local angular semantic tangent and a clean endpoint displacement. We integrate the tangent over the analytically known remaining VP angle to obtain a nominal one-step transport, and use the endpoint displacement to correct its long-horizon extrapolation through a constrained projection. This transport-first, correction-second construction yields Endpoint-Calibrated Transport (ECT), a closed-form, training-free editor requiring only one noise level, two text conditions packed into one denoiser batch, and one latent update. ECT requires no image inversion, trajectory integration, second temporal observation, or iterative optimization, and can be directly applied to pretrained epsilon-prediction VP one-step generators. Across the reported benchmark settings, ECT delivers stable gains in editing quality within an extremely short editing time.
[CV-67] Loop to Look Better: Learning Adaptive Depth for Question-Guided Perception in Vision-Language Models
链接: https://arxiv.org/abs/2610.30824
作者: Wencheng Ye,Anning Hu,Hongye Fang,Tianyi Wang,Junchi Yan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Visual perception should be iterative and task-directed, with representations refined in response to the question and the need for further processing. While Vision-Language Models (VLMs) increasingly support extended reasoning, their visual tokens are still typically acquired in a single, task-agnostic encoder pass. This creates an evidence-acquisition bottleneck: models may reason longer without refining the visual evidence needed to answer the question. We propose LoopLook, a lightweight encoder-side recurrent perception framework that equips pretrained VLMs with adaptive visual depth. LoopLook recurrently revisits a middle-to-late window of the vision encoder, injecting question representations at each loop to guide visual refinement. A learned stopping mechanism adaptively determines the recurrent depth, while a token-wise Mixture-of-Looks (MoL) combines representations from different loops, allowing each visual token to selectively integrate information acquired across recurrent steps. Across four backbones and eight benchmarks, LoopLook consistently improves performance over the baseline, yielding gains of up to 8.6% on MMStar and 8.1% on HallusionBench. Our analysis shows that LoopLook progressively refines visual attention and adaptively allocates depth according to task difficulty. It also better preserves task-relevant visual information under reduced input resolution, while increasing time to first token (TTFT) by only 5.1-6.5%. By bringing recurrence from reasoning back into perception, LoopLook transforms visual perception from a one-shot glimpse into an adaptive process of recurrent refinement.
[CV-68] DRILL: Training World Models to Improve Policies Not Predict Pixels NEURIPS2026
链接: https://arxiv.org/abs/2610.30814
作者: Zhaolu Kang,Tailong Luo,Lei Wei,Guangyuan Dong,Sihan Liu,Yingjie He,Siheng Wang,Shuaibo Li,Rongchao Zhang,Zhen Tian,Richeng Xuan,Zhichao Hu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026
Abstract:World models are increasingly used to reduce environment interaction when post-training vision-language-action (VLA) policies with reinforcement learning. Yet they are typically trained to predict pixels one step ahead, even though their role here is not merely to render visually faithful videos but to produce learning signals from imagined rollouts that improve the policy. We show that this mismatch is substantial: lower pixel MSE can produce worse policy gradients, and a world model trained for policy utility can yield better downstream policies despite 17% higher pixel MSE on the next frame. We propose the Downstream-Return Imagination Learning Loop (DRILL), a bilevel framework that updates the world model to maximize the policy’s return after an inner GRPO update on imagined rollouts. DRILL has two instantiations. DRILL-VWI is a closed-form surrogate, requiring no additional simulator rollouts, that reweights prediction errors by policy visitation, the group-relative GRPO advantage, and model confidence. DRILL-IMG is a full meta-gradient method that differentiates through the inner update using truncated Hessian-vector products. We show that DRILL-VWI recovers the local one-step component of the DRILL-IMG meta-gradient under score-compatibility and small-residual assumptions. Across five manipulation simulators and two VLA backbones, DRILL-IMG raises mean success rate by 13.1 points over WMPO and 7.9 points over RLVR-World, and complements RLVR pretraining rather than competing with it. More broadly, we believe video world models for VLA reinforcement learning can be productively trained by the policy gradients they produce, not only by pixel-level prediction.
[CV-69] WovenAnchor Matcher: Specialized Intra- and Inter-Image Context Modeling for Feature Matching
链接: https://arxiv.org/abs/2610.30799
作者: Zhiyang Li,Ruijiang Jin,Thibaut Klenke,Yusuke Sekikawa,Nakamasa Inoue
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 34 pages, 18 figures
Abstract:Semi-dense local feature matching commonly aggregates contextual information before coarse-to-fine correspondence estimation. However, intra-image and inter-image aggregation serve different roles: the former should propagate spatial evidence within each image while retaining structured two-dimensional dependencies, whereas the latter should gather information from plausible corresponding regions while limiting noise from unrelated locations. We propose WovenAnchor Matcher (WAM), a semi-dense matching framework based on role-specialized context aggregation. WAM introduces two complementary operators. WovenMamba performs intra-image aggregation through horizontal-then-vertical state-space propagation, allowing vertical updates to operate on horizontally contextualized features and thereby encouraging a structured two-dimensional receptive-field bias. AnchorCrossAttention performs inter-image aggregation by using cross-attention to gather context from plausible corresponding regions across the image pair. It performs this exchange over local representative anchors that aggregate feature evidence within each window. On MegaDepth, WAM achieves 66.1 pose AUC@5 ^\circ , improving on the JamMa baseline by 2.0 AUC points under the same evaluation setup. Ablations show that removing or replacing either specialized component reduces accuracy, providing evidence that intra-image propagation and inter-image retrieval benefit from role-specific aggregation designs. Code and pretrained weights are available at this https URL.
[CV-70] DeXplicit: Making Physical State Explicit in Dexterous Manipulation World Models
链接: https://arxiv.org/abs/2610.30781
作者: Adam Lee,Mutian Xu,Yukang Cao,Shulin Tian,Ziwei Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Before acting on an object, a robot needs to anticipate how its hand motion will change the physical world and what the resulting scene will look like. Existing approaches typically entangle action-conditioned dynamics, camera motion, and visual synthesis within a video generation model, or represent dynamics as point trajectories without explicit object structure. We introduce DeXplicit, a dexterous manipulation world model that instead predicts how the world evolves explicitly in metric 3D before rendering its visual consequences. Given an initial scene and a future hand trajectory, a lightweight 3D dynamics network predicts object-centric motion and deformation, providing structured and geometrically grounded future world states. These predictions then serve as a strong scaffold for video generation: appearance is transported through the predicted 3D motion and projected into future views, allowing the renderer to focus on completing uncertain and newly visible content rather than implicitly learning world dynamics from pixels. This explicit-3D-first decomposition also enables efficient causal generation, requiring only two denoising steps per block while conditioning on generated history. Across HOT3D, OakInk2, and PhysTwin, DeXplicit reduces moved-point error by 27-40% relative to PointWorld and improves object fidelity IoU+ by 47% on HOT3D under a shared renderer. Our design with two-step causal generation achieves a 57x speedup from our 14B bi-directional configuration in video synthesis. Comments: Project page: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.30781 [cs.CV] (or arXiv:2610.30781v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.30781 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-71] Evaluation of Front-End Feature Matching for Lunar Visual SLAM
链接: https://arxiv.org/abs/2610.30774
作者: Guillem Casadesus Vila,Grace Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 8 figures, 5 tables. Submitted to the 2027 IEEE Aerospace Conference
Abstract:Lunar rovers will primarily rely on cameras for localization until navigation satellite services are deployed, and will continue to need them afterwards for relative navigation, hazard detection, and surface operations. Visual odometry (VO) and Simultaneous Localization and Mapping (SLAM) mostly estimate motion by matching features between frames. However, feature matchers are developed on urban, indoor and driving scenes, planetary studies compare them on orbital registration or within complete SLAM systems, and real lunar imagery with reference poses is scarce, so matching between rover frames on low-texture, harshly lit terrain, and its effect on VO, has not been compared. We benchmark eleven classical and learned detector-matcher combinations on nine synthetic, lunar-testbed, and terrestrial-analogue datasets, and a feature tracker, scoring pairwise pose accuracy against baseline and the trajectory error of the resulting VO. Under a common rotation-only metric, the SuperPoint pairings outperform SIFT with ratio-tested FLANN matching on both lunar testbeds and one analogue site, with smaller or unresolved differences on synthetic data. When rotation and translation direction are scored jointly, this SIFT pairing exceeds both SuperPoint pairings on four synthetic datasets, so synthetic results alone do not establish relative performance on real rover imagery. In stereo VO, the SuperPoint pairings are not consistently more accurate than SIFT, but only they keep tracking on all eight sequential datasets. Pairings with classical detectors depend on lighting while SuperPoint changes little: a low spotlight makes them more accurate than diffuse light on the same testbed views, consistent with more contrast and detected keypoints. SuperPoint with LightGlue is the fastest learned combination (16 ms per frame), and DINOv2 retrieval proposals still need geometric verification.
[CV-72] FilmFlow: Agent ic System for 3D Virtual Film Cinematography
链接: https://arxiv.org/abs/2610.30753
作者: Yiming Zhang,Jiarui Guo,Haojia Wei,Yifei Liu,Yifan Yang,Hongjie Zhang,Xue Yang,Zhihang Zhong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Complex camera motion is central to visual storytelling, yet translating cinematographic intent into controllable, executable camera trajectories remains challenging. We introduce FilmFlow, an agentic framework that converts natural-language instructions or reference videos into explicit camera trajectories for dynamic 3D environments. To generate these trajectories, a Director agent constructs executable camera plans by composing high-level motions from concurrent atomic operations that jointly control translation, rotation, and framing. The Director is supported by three specialized agents: an Interpreter that extracts scene-grounded cinematographic intent, a Scout that selects initial viewpoints, and a Critic that evaluates rendered results and provides visual feedback for iterative plan refinement. We also introduce CinemaBench, comprising 150 Blender scenes from self-created and licensed sources, with five camera instructions per scene for 750 text-directed tasks and 60 reference videos associated with a subset of scenes. CinemaBench supports evaluation of both text-directed camera generation and video reference-directed camera generation. Across a diverse set of evaluated baselines, FilmFlow achieves the strongest benchmark performance and the highest user preference, demonstrating effective spatial reasoning, temporal coordination, and visual composition for executable 3D cinematography.
[CV-73] Where Did the Light Come From? Diagnosing Cast-Shadow Reasoning in VLMs and Borrowing the Missing Shadow Prior from Image Generation Models WACV2027
链接: https://arxiv.org/abs/2610.30740
作者: Ryugo Morita,Shunsuke Kitada,Stanislav Frolov,Brian Bernhard Moser,Ko Watanabe,Riku Takahashi,Jinjia Zhou,Andreas Dengel
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV2027
Abstract:Shadows are observable consequences of latent physical causes, arising from the interaction between illumination, object geometry, and scene structure. A model that understands shadows should be able not only to read a shadow’s direction, but also to infer where a shadow should fall when it is absent, and to judge whether a shadow is physically consistent with the scene. We introduce ShadowBench, built from real photographs, which poses these three questions to vision-language models (VLMs). From each scene, we build a correct-shadow, a shadow-removed, and a human-annotated wrong-direction state. Evaluating 19 open VLMs (2B to 78B) and frontier proprietary models, we find that VLMs read a visible shadow’s direction well above chance and improve with scale, yet infer the absent direction only weakly and fall to chance at judging physical consistency even under a bias-free two-alternative test that removes the yes-bias of single-image naturalness questions. Only Gemini-3.1-pro exceeds chance on validity, and even then only modestly and far below humans. Chain-of-Thought (CoT) does not help. The missing ingredient is not reasoning but knowledge that image generation models (IGMs) already hold. Routing a VLM through an IGM lifts its inference well above its direct attempt and helps it flag the wrong-direction shadows it otherwise accepts. Cast-shadow direction is a single facet of illumination understanding, but the sharp gap it reveals between VLMs and IGMs suggests image-generation or reconstruction objectives as a route toward illumination-aware VLMs.
[CV-74] SmolVLA-MoE: Sparse Mixture-of-Experts for Efficient Multi-Task Robotic Manipulation IROS2027
链接: https://arxiv.org/abs/2610.30732
作者: Shiyang Zhang,Miao Yang,Zhengda Li,Yang Shao,Ruifang Zhai,Hui Peng,Peng Song
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 5 figures, 4 tables; conference paper prepared for IROS2027
Abstract:Vision-Language-Action (VLA) models have shown promising results in multi-task robotic manipulation by leveraging visual and language understanding for action generation. However, existing lightweight VLA models employ a single dense action expert to handle all tasks, which limits their multi-task capability under constrained computational budgets. To address this issue, we propose SmolVLA-MoE, a novel framework that introduces sparse Mixture-of-Experts (MoE) into the FFN layers of the action expert in VLA models. SmolVLA-MoE replaces dense FFN layers with multiple parallel expert FFNs and a token-level router, enabling each token to adaptively select the most suitable expert for action generation. We further introduce a shared expert with a tunable weight parameter to balance the inheritance of pre-trained capabilities and task-specific expert specialization. Through sparse activation, SmolVLA-MoE activates only a single expert FFN during inference, with a total activated parameter count of approximately 500M including the frozen VLM backbone, while benefiting from approximately 800M total stored parameters. We evaluate our method on four task suites of the LIBERO benchmark. SmolVLA-MoE achieves an average success rate of 84.25%, outperforming SmolVLA (82.75%) and OpenVLA-7B (76.5%), while maintaining comparable inference efficiency. Our code is publicly available at this https URL.
[CV-75] Learning from Next-Frame Prediction: Towards Unified and Scalable Visual Representation Learning
链接: https://arxiv.org/abs/2610.30724
作者: Yifeng Yao,Jinghan Li,Zhihao Zhang,Yajiao Xiong,Yang Jin,Kun Xu,Huishuai Zhang,Dongyan Zhao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Video provides a natural source of supervision: what happens next serves as a learning signal for visual world modeling. Yet masked video pretraining focuses on recovering missing content from surrounding context rather than predicting the future from the past. Autoregressive methods instead predict forward one token at a time, but have yet to match joint-embedding methods in representation quality. This gap motivates us to revisit the granularity and space of future prediction. In this work, we introduce VANE, a self-supervised visual pretraining framework built on next-frame prediction in embedding space. A causal encoder summarizes past observations for a state predictor that infers the representation of the entire next frame, making temporal evolution the primary learning signal. This prediction is carried out in a fixed embedding space provided by a frozen semantic tokenizer, emphasizing scene content and structure. To retain fine-grained spatial detail alongside semantic content, a lightweight auxiliary generative decoder reconstructs future-frame patches conditioned on the predicted representations. To unify visual modeling across images and videos, we construct a local-to-global crop sequence for still images, allowing the same encoder to predict newly revealed regions from earlier crops under the same objective. Experiments show that VANE improves with model and data scaling, achieving SOTA on most anticipation, recognition, and dense perception benchmarks among unified encoders. Especially on EPIC-KITCHENS-100 action anticipation, it surpasses V-JEPA 2.1 by 7.1 points in mean-class recall@5. These results demonstrate that next-frame prediction in embedding space is effective and scalable for visual world modeling.
[CV-76] PhyEvolve: Multi-Agent State-Evolution Planning for Physics-Aware Video Generation
链接: https://arxiv.org/abs/2610.30717
作者: Yang-Hao Zhou,Lixing Tan,Zhiqi Li,Guang Yang,Xiao Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Recent advances in text-to-video generation have substantially improved visual fidelity, yet existing models still struggle to understand the physical evolution implicit in user prompts, making it difficult to generate videos with continuous physical state evolution. Here, we propose PhyEvolve, a multi-agent framework that enhances natural-language prompts through structured state-evolution planning. PhyEvolve consists of three agents: a Director, a Physical Evolver, and a Composer. Given a user prompt and an optional first frame, the Director grounds the scene and disentangles physical events from observation changes. The Physical Evolver uses scene evidence to identify physical mechanisms then constructs causally linked state transitions that unfold throughout the video, whether visible or not. The Composer utilizes these transitions to enhance the original prompts. Without retraining or model modification, PhyEvolve works directly with existing generators through prompt replacement. Extensive experiments on four major benchmarks show that PhyEvolve achieves state-of-the-art performance, outperforming the cutting-edge prior agentic baseline by 7.9 percentage points on average. These results demonstrate that explicit causal state-evolution planning improves physical consistency and state continuity in generated videos through natural-language prompt enhancement.
[CV-77] ParaLin: Accelerating Parallel Diffusion Integrator via Intrinsic Partially Linear Structure
链接: https://arxiv.org/abs/2610.30704
作者: Jianrong Lu,Zhiyu Zhu,Hui Liu,Junhui Hou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Training-free few-step samplers are often described as nearly quality-lossless under conventional FID/CLIP-style evaluation. It remains unclear whether that conclusion transfers to more realistic and complex generation-quality benchmarks. We first systematize the quality metrics used to evaluate training-free sampling acceleration for diffusion and flow-matching models. We find that representative training-free step-shortening methods often remain nearly lossless under FID/CLIP-class metrics, yet can suffer substantial quality degradation when evaluated with semantic-composition, dense-instruction, and long-text-rendering benchmarks such as GenEval, DPG-Bench, and LongTextBench. Parallel sampling is a notable exception: by retaining the full-step discretization and removing only the solver’s temporal dependency, it can preserve near-lossless quality while accelerating sampling. Building on this observation, we propose \textitParaLin++, a step-parity parallel sampler that preserves the full-step discretization while accelerating the solver iterations. Experiments across Stable Diffusion-v2, LSUN Church, FLUX.1-dev, and Ming-Flash-Omni 2.0 show that ParaLin++ preserves the full-step Euler quality profile at 2.6\times – 4.8\times , 1.68\times on FLUX, and 1.21\times on Ming. Project page: this https URL. Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.30704 [cs.CV] (or arXiv:2610.30704v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.30704 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-78] LoRA Superposition across Pretrained Models
链接: https://arxiv.org/abs/2610.30691
作者: Keigo Watanabe,Hiroshi Kera,Kazuhiko Kawamoto
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 12 pages, 5 figures, 2 tables
Abstract:Low-rank adaptation (LoRA) is usually tied to the pretrained backbone on which it is learned. We ask whether LoRAs trained independently for the same task on different pretrained backbones can be summed and applied unchanged to all contributing backbones. We call this construction LoRA superposition. Across seven ViT-B/16 CLIP backbones and eight image-classification tasks, the exact sum improves all 56 backbone-task pairs and achieves a mean accuracy gain of 30.21 percentage points over the unadapted backbone. The exact sum achieves a mean accuracy only 0.28 percentage points below that of individually trained solo LoRAs. In contrast, a leave-one-model-out (LOMO) sum produces a mean gain of only 0.06 percentage points over the unadapted backbones. Backbone-specific solo LoRAs are nearly orthogonal in parameter space, while pairwise cross-backbone responses are strongly concentrated on the diagonal. These results support a component-wise non-interference hypothesis, under which the frozen backbone acts as an implicit selector of its own component without explicit routing or cross-model alignment.
[CV-79] Zero-Shot Domain Generalization via Domain Prompt Learning
链接: https://arxiv.org/abs/2610.30680
作者: Shiming Chen,Shuhuang Chen,Salman Khan,Fahad Shahbaz Khan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to IJCV
Abstract:Domain generalization (DG) aims to improve the generalization of the model to any unseen domains. Targeting this goal, existing DG methods mainly focus on learning domain-invariant features or prompts from the given source domains. However, these methods rely on source-domain data, and collecting and annotating large-scale single- or multi-source datasets is often infeasible, limiting the practical applicability of DG. To tackle this challenge, we propose \underlineDomain \underlinePrompt Learning (DoP), a method built on a large vision-language model (e.g., CLIP) for zero-shot domain generalization, which recognizes unseen classes in unseen target domains without access to source-domain data. DoP essentially learns the target domain knowledge via prompt both in text and visual modalities to overcome the domain shift in text-visual modalities under the test-time adaptation formulation. Specifically, DoP first takes text domain learning (TDL) to simulate various distribution shifts by learning diverse style tokens implicitly in the text space with inter-style diversity loss and intra-style consistency loss. Subsequently, DoP employs visual domain learning (VDL) to further refine the target-domain-relevant style token generated by TDL by making use of the test visual data with self-supervised learning. Extensive experiments on popular DG benchmark datasets (e.g., PACS, VLCS, OfficeHome, and DomainNet) and ImageNet show that our DoP consistently outperforms the state-of-the-art DG approaches. Codes are available at: this https URL. Comments: Accepted to IJCV Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2610.30680 [cs.CV] (or arXiv:2610.30680v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2610.30680 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-80] Agent Dreamer: Agent ic 3D Motion Reasoning for Text-Controllable Physics Simulation KR
链接: https://arxiv.org/abs/2610.30679
作者: Mijeong Kim,Suho Park,Junha Park,Bohyung Han
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:We propose AgentDreamer, an agent-guided optimization framework that aligns text instructions with simulated motion through explicit 3D motion planning. Given a text instruction and scene geometry, an off-the-shelf agent translates the instruction into a structured semantic motion prior, specifying what should move, in which direction, and in what temporal order. The agent represents this prior using parametric functions over a series of temporal chunks, yielding dense 3D trajectories. We then optimize the parameters of a differentiable MPM simulator such that its rollout best matches the prior while remaining constrained by the scene dynamics. This allows the agent to provide semantic guidance from language, while the simulator grounds it in physically feasible motion. We further introduce Motion-IG, an information-theoretic metric for text-motion alignment based on masked text recovery. By comparing the likelihood of recovering masked motion-related content with and without the simulated motion, Motion-IG measures the information contributed by the motion beyond that available from the remaining text and scene context. Experiments show that AgentDreamer improves text alignment over video-based guidance, producing simulations that more closely follow the requested physical behavior.
[CV-81] Class-Grounded Batik Captioning under Scarce Text Annotation: Label Retention and Source-Specific Error Propagation
链接: https://arxiv.org/abs/2610.30661
作者: Novanto Yudistira,Muhammad Dzaka Aufa Fadhillah,Irawati Nurmala Sari
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Class labels offer compact supervision for describing cultural-heritage images when expert text annotations are scarce. We investigate a batik-captioning cascade that integrates a LoRA-adapted DINOv2 classifier with Qwen2.5-VL-7B, adapted using LoRA and loaded in 4-bit NF4. A 3\times2 evaluation combines three captioner conditions with two inference prompts to examine adaptation and predicted-class injection. Across 150 caption-evaluation images, all three injected conditions achieve a ground-truth class-label mention rate of 0.9067. The classifier’s separate test yields 96.10% pooled accuracy, with 98.33% on 300 public-source images and 75.76% on 33 local images. Supported-class macro-F1 and balanced accuracy on the local subset are 0.8493 and 0.8302, respectively. The classifier and captioner use separate partitions, and their training–evaluation overlap remains unquantified. Differences in label semantics and training support also affect interpretation of the source-specific accuracy gap. Reference-based caption metrics and image–text compatibility scores produce different model rankings. Qualitative examples show both retention of correctly predicted identities and propagation of classification errors into fluent descriptions. These findings characterize how class conditioning preserves label information and transfers upstream errors into batik captions under scarce text annotation.
[CV-82] Residual Calibration via Local Feature-Space Refinement NEURIPS2026
链接: https://arxiv.org/abs/2610.30652
作者: Xihao Wang,Chengyao Yu,Ruixing Ming
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (stat.ML)
备注: Accepted at NeurIPS 2026
Abstract:Deep neural networks often struggle to provide reliable confidence estimates, making post-hoc calibration essential for reliable decision-making. Existing post-hoc methods largely rely on score-based transformations, which improve marginal calibration but often fail to capture spatially varying miscalibration in local regions of the learned feature space. To address this gap, we propose Local Adaptive Probability Estimation (LAPE), a post-hoc residual calibration framework. Specifically, LAPE treats the global calibrator output as a stable anchor and refines the individual prediction by leveraging non-parametric local evidence from neighboring calibration samples. As a plug-and-play refinement module, LAPE effectively utilizes the information about feature-space geometry and can reuse the calibration data that are used to construct the global calibrator. Furthermore, we extend LAPE to the setting with multi-view inputs through Cascade Feature Augmentation Fusion (CFAF). Extensive experiments demonstrate that the proposed methods can improve the calibration quality of various existing baselines and remain efficient and robust across different settings.
[CV-83] One-shot federated unsupervised domain adaptation with Smoothed Knowledge Distillation and teacher refinement WWW
链接: https://arxiv.org/abs/2610.30646
作者: Ali Abedi,Q. M. Jonathan Wu,Ning Zhang,Farhad Pourpanah
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Published in Engineering Applications of Artificial Intelligence, Vol. 167, 113946 (2026). DOI: https://doi.org/10.1016/j.engappai.2026.113946 . Available at: this https URL and Code is available at: this https URL
Abstract:Federated unsupervised domain adaptation (FUDA) addresses the challenge of adapting models to an unlabeled target domain using decentralized source domains while preserving data privacy. However, existing FUDA methods typically require multiple communication rounds, rely on complex aggregation strategies, and often struggle with noisy pseudo labels and inconsistent source knowledge. To address these challenges, we propose SKD-ETR, a novel one-shot FUDA framework that combines Smoothed Knowledge Distillation (SKD) and Exponential moving average-based Teacher Refinement (ETR). SKD trains a student model on the target domain using smoothed soft pseudo labels generated by the ensemble of source models. This reduces overconfidence, mitigates noise, and improves robustness. ETR further refines each source model by interpolating its parameters toward the student via exponential moving average (EMA), thereby transferring target-domain knowledge back to the teachers. This bidirectional refinement enhances pseudo-label quality and student generalization without additional communication overhead. SKD-ETR eliminates complex aggregation by initializing the student model randomly and performing a single-round distillation process. Extensive experiments on OfficeHome, Office-Caltech, and DomainNet demonstrate that SKD-ETR achieves competitive performance while being communication- and computation-efficient, and robust under noisy supervision. The code for this paper is available at this https URL.
[CV-84] Asking Less Adapting Better: Feedback-Efficient Test-Time Adaptation for Vision-and-Language Navigation
链接: https://arxiv.org/abs/2610.30643
作者: Zixu Wang,Huaquan Zhang,Shoukai Xu,Jinwu Hu,Zhiquan Wen,Mingkui Tan
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-and-Language Navigation (VLN) requires agents to follow instructions and navigate to target locations. However, current VLN agents often suffer severe performance degradation in unseen environments due to distribution shifts. Test-Time Adaptation (TTA) offers a promising paradigm to bridge this gap, yet existing methods that rely on unsupervised entropy minimization can propagate errors through the trajectory, causing cascading failures in navigation. Recent methods mitigate this by incorporating human feedback, but they apply trajectory-level supervision uniformly across all steps without localizing the actions is truly informative, leading to inefficient feedback utilization and mismatched optimization granularity. To address these limitations, we propose FETA, a Feedback-Efficient Test-Time Adaptation method that improves both feedback efficiency and optimization reliability. To reduce unnecessary human interventions, we introduce a hierarchical sample selection strategy to localize informative samples. Specifically, we first filter trajectories whose normalized policy entropy exhibits large fluctuations. Within each selected trajectory, we further select the critical action via an entropy-gradient saliency criterion and query a single binary feedback on its correctness. A discrepancy-aware entropy optimization objective then updates the policy only at the queried step, with adaptive scaling that assigns larger relative weights to samples with greater prediction-feedback discrepancy. Extensive experiments across multiple benchmarks demonstrate that FETA achieves state-of-the-art performance with fewer human interventions, while remaining robust under noisy feedback and continual adaptation settings.
[CV-85] CODI: Improving Self-Supervised Vision Transformers with Cross Distillation
链接: https://arxiv.org/abs/2610.30639
作者: Siran Dai,Qianqian Xu,Peisong Wen,Yang Liu,Xiaochun Cao,Qingming Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 35 pages, 11 figures, 14 tables
Abstract:Self-supervised Vision Transformers learn strong image-level representations, but their patch-level representations can be unreliable for dense prediction. We study this mismatch and find that the class token concentrates global semantics on a small number of representative patches, while dense prediction errors often occur on low-attention regions. We provide a local gradient analysis showing that, along the [CLS] attention readout, image-level self-distillation gradients are weighted by [CLS]-to-patch attention mass. To provide masked patches with global semantics without changing the architecture, we propose Cross Distillation (CODI), which mixes a small amount of the teacher class-token distribution into each teacher patch-token target. CODI gives every masked patch an attention-independent semantic correction while preserving the standard self-distillation architecture. Experiments on dense and image-level benchmarks show that CODI improves both dense and image-level performance under aligned pretraining settings. Code is available at this https URL.
[CV-86] FedBST: Bilateral Self-Training for Unsupervised Federated Classification ICME2026
链接: https://arxiv.org/abs/2610.30634
作者: Aijing Yu,Kuangpu Guo,Jian Liang,Xiao-Yu Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ICME 2026
Abstract:Unsupervised federated learning with Vision-Language Models (e.g., CLIP) offers a promising solution for label-free classification, yet its performance often degrades under severe non-IID shifts. A pioneering method, FedCoPL, typically relies on self-training strategies that exclusively prioritize high-confidence predictions for pseudo-labeling. However, this selective approach neglects the information of ambiguous samples (often referred to as ``waist" samples), leading to the underutilization of data and biased decision boundaries. To address this limitation, we propose FedBST, a bilateral self-training framework that effectively exploits both high-confidence and ambiguous data. Specifically, for local optimization, we introduce uncertainty-aware confidence penalization (UACP). UACP targets ambiguous samples by penalizing overconfident predictions to implicitly promote higher uncertainty, thereby preventing erroneous pseudo-labeling. At inference, we employ a visual-text ensemble strategy. This approach aggregates uncertainty-aware local visual prototypes and ensembles them with textual prototypes, achieving lightweight adaptation to local distributions via prototype-based calibration. Experiments demonstrate that FedBST achieves consistent improvements over existing methods with standard federated aggregation protocols.
[CV-87] Refining Open-Vocabulary Semantic Segmentation via Regional Semantics and Visual Prototypes ICASSP2026
链接: https://arxiv.org/abs/2610.30623
作者: Aijing Yu,Zhengbo Wang,Lijun Sheng,Jian Liang,Xiao-Yu Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ICASSP 2026
Abstract:Contrastive Language-Image Pre-training (CLIP) has shown superior performance in the image-level open-vocabulary classification task, benefiting from its powerful alignment of visual features and text embeddings. However, dense prediction tasks require models to perform pixel-level predictions, necessitating modifications to CLIP for region-text alignment. In this paper, we adjust the attention layers within the vanilla visual encoder of CLIP. With a novel attention calculation method for selective attention features, the model retains more regional semantic information. To further mitigate the negative impact of the modality gap, we introduce additional visual prototypes to refine the original predictor solely based on text embeddings. Specifically, patch features with confident predictions are selected to form an additional auxiliary memory-based classifier for more accurate prediction. We further integrate an optimizable feature extractor with an exponential moving average updating strategy to enhance prediction quality. Extensive experiments on 5 semantic segmentation datasets demonstrate the effectiveness of our method.
[CV-88] GSHeadRelight: Fast Relightability for 3D Gaussian Head Synthesis
链接: https://arxiv.org/abs/2610.30619
作者: Henglei Lv,Bailin Deng,Jianzhu Guo,Xiaoqiang Liu,Pengfei Wan,Di Zhang,Lin Gao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Relighting and novel view synthesis of human portraits are essential in applications such as portrait photography, virtual reality (VR), and augmented reality (AR). Despite recent progress, 3D-aware portrait relighting remains challenging due to the demands for photorealistic rendering, real-time performance, and generalization to unseen subjects. Existing works either rely on supervision from limited and expensive light stage captured data or produce suboptimal results. Moreover, many works are based on generative NeRFs, which suffer from poor 3D consistency and low real-time performance. We resort to recent progress on generative 3D Gaussians and design a lighting model based on a unified neural radiance transfer representation, which responds linearly to incident light. Using only in-the-wild images, our method achieves state-of-the-art relighting results and a significantly faster rendering speed ( \times12 ) compared to previous 3D-aware portrait relighting research. Project page: this http URL.
[CV-89] KinoUAV: A Benchmark for Safe and Stable UAV Vision-Language Navigation with Kinodynamic Modeling
链接: https://arxiv.org/abs/2610.30609
作者: Chenyao Yang,Peijin Wang,Yingchao Feng,Wenhui Diao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Unmanned aerial vehicles (UAVs) have recently emerged as important remote sensing platforms, and UAV navigation has shown great potential in remote sensing applications. However, existing methods do not explicitly incorporate kinodynamic modeling or safety distances into trajectory generation, resulting in high absolute yaw rates, frequent collisions, and low navigation success rates, which limit safety and motion stability. To address this challenge, we introduce KinoUAV, a kinodynamic benchmark comprising a dataset and KinoUAV-Agent. First, we develop a hierarchical kinodynamic trajectory generation pipeline that drives UAV through acceleration actions. It uses waypoint chains from global planning and generates trajectories segment by segment under velocity, acceleration, and jerk constraints, reducing oscillations and improving trajectory stability. Furthermore, we introduce directional free-space distance (DFD) labels into the constructed dataset to describe the available clearance for safe motion decisions. The proposed KinoUAV-Agent predicts DFDs and uses them to generate safe actions, decoupling perception from action decision-making. This design enables autonomous obstacle avoidance and improves navigation safety. The benchmark incorporates the proposed Cruise-Seek task, in which the UAV first enters the target region and then searches for the target, reducing the difficulty of target search and improving navigation success. Experiments on KinoUAV show that KinoUAV-Agent outperforms existing methods, achieving the highest success rate while maintaining lower absolute yaw rates and collision rates. These results demonstrate the effectiveness of kinodynamic modeling for safe and stable UAV navigation.
[CV-90] Visual Grounding First Multimodal In-context Learning Follows NEURIPS2026
链接: https://arxiv.org/abs/2610.30592
作者: Minhyuk Seo,Minjae Lee,Chaeeun Lee,Wei Lin,Muhammad Jehanzeb Mirza,Tinne Tuytelaars,Jonghyun Choi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: NeurIPS 2026
Abstract:LLMs have shown strong in-context learning (ICL) capability, but extending it to Multimodal LLMs (MLLMs) remains challenging. Prior multimodal ICL methods often rely on ICL-specific datasets for additional training or task vector extraction, which can improve performance on the same ICL benchmark used for adaptation. However, they do not generalize well to other benchmarks and often induce forgetting in previously well-solved tasks. For this reason, we identify ‘visual grounding’ as the key bottleneck in multimodal ICL; MLLMs often fail to attend to task-relevant visual evidence in demonstrations, instead over-relying on textual cues and producing hallucinated outputs. To address this, we propose ProCoRe to enhance visual grounding through contrastive reinforcement learning on multimodal contrastive data for improving multimodal ICL. ProCoRe generates captions for contrastive image pairs, decomposes them into propositions, and optimizes contrastive rewards that encourage alignment with corresponding images while discouraging mismatched ones. Despite never seeing ICL-formatted training examples, ProCoRe improves average multimodal ICL classification accuracy by 5.6% and captioning ROUGE-L by 13.8% over the off-the-shelf Qwen-VL-3-8B model, outperforming ICL-data-dependent state of the arts across benchmarks. Finally, we introduce PICL, a personalized multimodal ICL captioning benchmark that tests whether MLLMs genuinely understand visual demonstrations rather than relying on pre-trained semantic priors.
[CV-91] Unified Trajectory and 3D Pose Forecasting for Multi-Agent Sports Analysis WACV2027
链接: https://arxiv.org/abs/2610.30519
作者: Marc Peral,Guillem Capellera,Luis Ferraz,Antonio Rubio,Stella Graßhof,Dan Witzner Hansen,Antonio Agudo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: WACV 2027
Abstract:Forecasting future human motion requires jointly reasoning about an agent’s global trajectory and 3D body pose, which are tightly coupled, particularly in team sports where multiple interacting players introduce complex social dynamics. We exploit a socio-temporal encoder that fuses trajectory and pose at the input level and models them jointly through shared factorized attention. The model is further enhanced by a per-joint loss reweighting scheme that emphasizes dynamic joints and prevents degenerate solutions, and a frequency-domain trajectory smoothing term that improves trajectory prediction while indirectly benefiting pose forecasting. To ensure fair comparison with prior joint forecasting methods, we identify and correct evaluation biases arising from overlapping sequence construction, adopting instead a non-overlapping protocol and a deterministic, oracle-free strategy for multi-future prediction. Our approach improves over prior joint forecasting methods on both WorldPose and 3DPW, with the largest gains on pose accuracy, demonstrating that the proposed joint formulation generalizes beyond the densely interactive scenarios that motivate it.
[CV-92] SplatLab: Autonomous Scientific Discovery for 3DGS Research NEURIPS2026
链接: https://arxiv.org/abs/2610.30468
作者: Seemandhar Jain,Manmohan Chandraker
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026. 14 pages, 4 figures. Project page: this https URL
Abstract:Reproducing a published 3DGS paper takes weeks of expert engineering before any extension can be proposed. Prior autonomous-research systems read papers, draft hypotheses, and write code, but are limited by short fixed templates, low executability, and weak traceability, and none has been demonstrated end-to-end in a live computer-vision subfield with intense activity. SplatLab is an autonomous research system for 3D Gaussian Splatting that addresses these limitations. Given a research prompt, a target-paper URL, or no input at all, it reproduces a published baseline, proposes an extension, trains and ablates across Mip-NeRF 360 and LLFF, writes the manuscript, and binds every reported number to a logged training run. Five design choices distinguish it: (i) candidates mutate plugins from a verified-reproduction backbone matched to published baselines; (ii) a cascade-weighted Elo tournament across eight islands scores candidates on measured GPU PSNR with verbal-gradient feedback; (iii) a standing falsifier subtracts Elo on failure, integrating stress-test survival into selection; (iv) every reported number is bound to a logged training run, and the run cannot terminate while any claim is unbacked; (v) a seven-category 3DGS taxonomy wraps EVOLVE-BLOCK markers in domain-aligned scaffolds transferable across 3DGS repositories. On PSNR-targeted runs, SplatLab’s manuscripts beat their reproduced targets by +0.18 to +3.78 dB, including +3.78 dB on astrophotographic nebula rendering, for which no prior 3DGS method exists. The research cycle compresses from months of graduate-student time to days of mostly autonomous compute. Code, all manuscripts, experiment ledgers, and the cross-run paper store are available on the project page.
[CV-93] VisCurate: Evaluating Behavioral Equivalence of Visual Skills WACV2027
链接: https://arxiv.org/abs/2610.30405
作者: Md Farhan Ishraq,Md Farhan Ishmam,Reaz Hassan Joarder,Md Hasanul Kabir
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2027), Main Conference
Abstract:Modern vision-based agents have shifted toward skill-based problem-solving, where a skill is an executable image-to-image operation, e.g., cropping, thresholding, or edge detection. As these skill libraries grow, they accumulate duplicated, mislabeled, and defective skills. The signals normally used to curate them (a skill’s name, description, or source code) say little about how the skill visually changes the image. We present VisCurate, a framework and benchmark that curates visual skill libraries from observed execution rather than from text. VisCurate runs each candidate skill pair on a shared probe battery and assigns one of six behavioral relations, each tied to a library edit that is safe to make. The benchmark injects seven defect types into a clean library at controlled rates, producing 300 reproducible instances with ground-truth relations and repairs. While text baselines accept up to five of six false merges on hard negatives, our verifier accepts none and recovers 51.2% of the pairs that are truly mergeable. On two real, uncurated libraries, VisCurate finds that 92% of the skills a package shares with its downstream copy are redundant, while execution confirms only 10 of 17 published equivalence claims. The curation problem also remains open, as the strongest of the nine automated curators recovers only about half of the repairs. We emphasize that visual skill libraries should be curated by behavior, not by description.
[CV-94] Latent Space Jam: Layout Guided Video Synthesis
链接: https://arxiv.org/abs/2610.30390
作者: Asaf Joseph,Moshe Mandel,Ohad Fried
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: this https URL
Abstract:Adding control to video generation often requires training, which demands additional computational resources, data, or architectural changes. We propose a training-free approach that utilizes pretrained video diffusion models to control object layout and motion using coarse layouts. Our method operates in two passes: first, it steers spatiotemporal placement through prompt-guided cross-attention to produce a coarse visual guide; second, this guide conditions the same model to generate a high-quality, layoutconsistent video. The approach enables coordinated motion with strong temporal consistency, supporting diverse trajectories and object interactions without additional training. Extensive experiments show that our method consistently outperforms prior work in overlap and trajectory accuracy, as well as visual quality across static, dynamic, and multi-object scene settings.
[CV-95] FiLM-CAM: Keyed Feature Modulation for Conditional-Access Watermarking NEURIPS2026
链接: https://arxiv.org/abs/2610.30347
作者: Yash Kulthe,Vishal Asnani,Shruti Agarwal,Andrew Gilbert,John Collomosse
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to NeurIPS 2026
Abstract:We propose FiLM-CAM; a conditional-access image watermarking approach that enforces authorized encoding, decoding, and removal through keyed feature modulation. A secret key conditions a Feature-wise Linear Modulation (FiLM) mechanism that transforms the internal representation of the watermark signal before embedding, thereby coupling the payload to the key at the feature, rather than pixel level. The same key must be presented to condition the decoder to correctly recover the payload embedded in an image. To enforce conditional access, we train the system to maximize the performance gap between authorized and unauthorized use, ensuring reliable extraction under the correct key while inducing failure under incorrect keys. In addition, we design our encoder to estimate the watermark residual invariant to the presence of pre-existing watermarks. This design enables conditional watermark removal via simple re-encoding and subtraction of the residual using iterative refinement, contingent on knowledge of the secret key. We evaluate FiLM-CAM across multiple watermarking encoder-decoder backbones, demonstrating that keyed feature modulation serves as an effective, architecture-agnostic conditional access module (CAM) for image watermarking.
[CV-96] Foundation4D: Towards a unified perception and reconstruction model for autonomous vehicles
链接: https://arxiv.org/abs/2610.30340
作者: NVIDIA:Sérgio Agostinho,Maximilian Apodaca,Emmanuel Attia,Guillem Braso,George Dong,Ilya Fedorov,Michele Fenzi,Gang Huang,Jiahui Huang,Shengyu Huang,Bingxin Ke,Laura Leal-Taixé,Kuan-Ting Lee,Mingxuan Liu,Tian Liu,Tim Meinhardt,Aljoša Ošep,Naveen Kumar Rai,Kefei Ren,Cristiano Saltori,Keqiang Wu,Ning Xu,Fei Xue,Jin-Bey Yu,Lei Zhang,Qunjie Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We introduce Foundation4D, a unified feed-forward model for geometry estimation, 3D detection, semantic occupancy, and Gaussian reconstruction from multi-camera driving video. A shared backbone uses factorized spatiotemporal attention to efficiently combine observations across cameras and time. Our video inference pipeline uses pose graph optimization to refine camera trajectories and assembles object tracks across input windows. We also develop a data annotation pipeline and release PAI-Foundation4D, a benchmark with dense depth and semantic occupancy labels for 1,253 driving clips. With benchmark-specific fine-tuning, Foundation4D achieves the lowest depth AbsRel on Waymo and nuScenes, the highest detection mAP on nuScenes, and the highest occupancy mIoU on OpenOccupancy among the compared methods. Our task-combination study shows benefits for both detection and geometry: co-training detection with geometry and occupancy improves top-class mAP by 1.2 points over detection-only training, while co-training geometry with detection reduces camera-to-rig translation error by 13.4% relative to geometry-only training. We further use the predicted geometry and trajectories as priors for lidarless NuRec, supporting reconstruction of driving scenes for simulation.
[CV-97] MoRE: Motion-Regularized Encoding for Articulated 3D Generation
链接: https://arxiv.org/abs/2610.30337
作者: Guantian Zheng,Shan Jin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 6 pages, 3 figures, 4 tables
Abstract:Generating an articulated object from a single image requires recovering both its appearance and the mechanism governing its parts. A compact encoding of joint attributes does not by itself ensure that reconstruction preserves the resulting motion. We present MoRE, Motion-Regularized Encoding for articulated 3D generation. MoRE represents coordinate-aligned joint lines with axis classification and two axis-orthogonal Plücker moment coefficients, removing the arbitrary choice of an origin along the axis. During representation learning, decoded joint hypotheses are used to move corresponding three-dimensional anchors, and their motions are compared with those of annotated joints. A differentiable expectation over joint types and axis directions couples attribute reconstruction to the motion it induces. An image-conditioned flow learns the resulting functional latents, while a multi-state-adapted Gaussian model supplies appearance. At inference, recovered part assignments and joints articulate a single set of rest-pose Gaussians without regenerating the object for each state. Experiments on PartNet-Mobility and ACD show stronger part-box agreement, articulated-state geometry and visual similarity than the compared baselines, while joint prediction reduces axis, origin and range errors relative to ArtiLatent. Controlled ablations support the joint-line encoding and show that motion supervision reduces anchor-motion error. Code and pretrained checkpoints will be released in the future.
[CV-98] R4DGS: Referring Segmentation in 4D Gaussian Splatting
链接: https://arxiv.org/abs/2610.30326
作者: Bangpu Chen,Yaxuan Li,Shirui Peng,Xiangtian Si,Liuxin Chu,Xitong Cao,Hongbo Jin,Jiayu Ding
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages. Accepted at ACM Multimedia 2026. Project page: this https URL
Abstract:Dynamic representations based on 4D Gaussian Splatting (4DGS) have shown strong capabilities in reconstruction and rendering, yet current methods remain limited in supporting complex natural-language referring in dynamic scenes. Existing 4D Gaussian methods are primarily designed for rendering-oriented dynamic modeling, while current semantic extensions often rely on scene-specific semantic optimization or are evaluated on relatively simple query settings. In this paper, we study Referring 4D Gaussian Splatting (R4DGS), a task for grounding temporally varying, multi-target or reasoning-intensive, and zero-target or distractor expressions in dynamic 4D scenes. To support this task, we introduce R4D-Bench-QA, a benchmark with structured query annotations. We further present ReferGaussian, a unified framework built upon 4DGS that converts low-level dynamic Gaussians into persistent, language-addressable entities for training-free referring inference. ReferGaussian organizes mask-supported Gaussian entities in an EntityBank, uses Multimodal Large Language Models (MLLMs) to decompose complex queries into trackable object phrases and structured referring constraints, and performs constraint-aware grounding without scene-specific semantic optimization. Experiments on R4D-Bench-QA and the 4D LangSplat protocol demonstrate improved temporal target selection and competitive spatiotemporal grounding without scene-specific semantic training. Project resources are available at this https URL.
[CV-99] Self-Distilling Future Video Representations for World Action Models
链接: https://arxiv.org/abs/2610.30320
作者: Huiwon Jang,Suhyeok Jang,Hyeonmo Kang,Seungku Kim,Jinwoo Shin,Sihyun Yu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 27 pages; Project page: this https URL
Abstract:World Action Models (WAMs) have emerged as a promising approach to learning generalist robot policies using rich spatiotemporal priors from pre-trained video diffusion models. A recent line of WAMs improves inference efficiency by conditioning a separate action decoder on intermediate future video representations extracted from a single denoising step, avoiding costly iterative denoising required for video generation. This design makes action prediction depend directly on the quality of these future representations. In this paper, we introduce Future REpresentation learning via self-Distillation with temporal context (FRED), a framework that trains WAMs to improve the quality of future video representations from noise in a single denoising step. Specifically, FRED uses self-distillation to predict future representations from noise, using representations extracted from ground-truth future observations as targets. We further condition the video model on historical observations, providing temporal context for predicting future representations better. Across diverse real-world manipulation scenarios, FRED consistently outperforms competing WAMs. It achieves particularly large gains on long-horizon and out-of-domain tasks under zero-shot environment generalization, and substantially improves human-to-robot skill transfer, reaching 50.0% task success on tasks absent from robot training data and demonstrated only in human videos, compared with 19.4% for the best competing WAM.
[CV-100] rraTransfer: Demonstration-Free End-to-End Planner
链接: https://arxiv.org/abs/2610.30319
作者: Zikang Xiong,Weixin Li,Zhouchonghao Wu,Akshay Rangesh,Saarth Bonde,Grantland Hall,Chen Tang,Yihan Hu,Wei Zhan
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:End-to-end autonomous driving has achieved state-of-the-art performance on benchmarks and real-world deployments. Its standard training recipe, however, is expensive across all existing paradigms: imitation learning requires fleet-scale driving logs whose behavior distribution only thinly covers the safety-critical tail; closed-loop reinforcement learning in a photorealistic simulator, while offering diverse behaviors, is bottlenecked by sample-efficiency, per-step cost of rendering, and a forward pass through a large vision backbone. Self-play in vectorized simulators changes the economics: at millions of rollout steps per second, one can obtain a state distribution with natural diversity in collisions, near-misses, and recoveries that no driving log contains, and a highly reactive policy emerges from a basic driving-quality reward. Our approach exploits this asymmetry. We pretrain a single policy by self-play, distill its rollouts into a flow-matching trajectory teacher that also plans, and train lightweight adapters on a frozen pretrained vision backbone to reproduce the latent space of either teacher through the action KL divergence, a batch-relational low-rank structural loss, and, for the planner, a trajectory-matching loss. Every target traces back to the self-play policy, so alignment never supervises against a logged trajectory: a paired dataset of (image, scene-state) frames suffices, with no need for the curated expert demonstrations that imitation pretraining is built on. From this one self-play teacher we obtain two end-to-end students: a closed-loop policy that outperforms every published imitation-based end-to-end method on the photorealistic 3D Gaussian splatting closed-loop benchmark HUGSim, and a planner that reaches 59.4 EPDMS on NAVSIM navhard, above every published method including DriveZero-Scale (57.1), and 95.5 PDMS on navtest, without human demonstrations.
[CV-101] EgoHand: Monocular 3D Hand Reconstruction from Egocentric Video
链接: https://arxiv.org/abs/2610.30293
作者: Chao-Yeh Chen,Tiffany Chen,Lin Duan,Tianyi Hu,Yu-Hsiang Huang,Umar Iqbal,Jiefeng Li,Edy Lim,Yanchen Liu,Yihan Pang,Hang Qiu,Jun Saito,Shao Su,Qingsi Wang,Xiufeng Xie
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Authors are listed alphabetically. Project leads are Chao-Yeh Chen, Tiffany Chen, and Umar Iqbal
Abstract:We introduce EgoHand, a monocular 3D hand reconstruction model that achieves state-of-the-art performance on egocentric video. Central to our approach is the curation of a high-quality training corpus of 14 million images, combining synthetic data, in-the-wild video, and multi-camera motion capture to balance scene diversity and supervision quality. The model adopts SOMA-Hand as its parametric representation and processes 2D hand crops with a transformer encoder, incorporating camera cues and an auxiliary 2D branch to improve geometric consistency and image alignment. Zero-shot evaluations on HOT3D, ARCTIC, and H2O demonstrate strong out-of-domain generalization, with lower joint-position, root-translation, and acceleration errors than competing baselines at an inference latency of approximately 9.69 ms. A blinded user preference study further shows that participants favor EgoHand’s reconstructed hand poses over those from Apple EgoDex and Ropedia Xperience-10M. We release the EgoHand model and source code under the OpenMDW license, allowing free commercial use of the model.
[CV-102] V-JEPA 2 Meets AQA: Phase-Based Scoring and Reference-Based Feedback for Mountain Biking
链接: https://arxiv.org/abs/2610.30160
作者: Santhiya Rajan,Savio Cardoz
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 27 pages, 10 figures
Abstract:We study mountain-bike action quality assessment (AQA) with weak labels and candidate feedback. The training corpus comprises 95 field-recorded jump clips with binary labels denoting good or poor technique. These are coarse categorical labels, not continuous expert quality scores; phase-boundary and aspect-level annotations are unavailable. Our pipeline summarizes four heuristic phases using frozen V-JEPA 2 video features and two-dimensional pose and motion measurements. A small transformer predicts a binary-label score. A separate reference layer compares phase-specific measurements with those of clips labelled as good technique to propose phase, body-parameter, and direction cues. We detail the measurements, network, and evaluation so that the limits of each output are explicit. In five-fold clip-level cross-validation, the full scorer obtained pooled out-of-fold AUROC 0.768 and average precision 0.812. These are exploratory figures because the reporting fold also selected the neural training epoch and riders were not separated between folds. Removing biomechanical scoring inputs raised AUROC to 0.809. A separate inner-selection run obtained AUROC 0.678, but used about 57 rather than 76 training clips per outer fold, so its difference from the test-selected run does not quantify selection bias. The unsupervised aspect head did not learn distinct factors. On a 17-clip external probe, an earlier checkpoint obtained binary accuracy 0.412 and Spearman association -0.135 with expert scores. Candidate cues had precision 0.118 and recall 1.0 against each of two coaches, with substantial disagreement about direction. These findings do not establish rider-independent scoring or reliable corrective feedback; they identify the empirical work needed to evaluate either use.
[CV-103] Self-Supervised Multibeam Sonar Restoration via Coherent Forward Modeling
链接: https://arxiv.org/abs/2610.31490
作者: Zixin Wang,James Hopgood,Mike Davies
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP)
备注: 16 pages, 8 figures, includes supplementary material
Abstract:Multibeam imaging sonar produces intensity images degraded by coherent speckle and system-induced blur, while clean reflectivity targets are rarely available. Existing despeckling methods often rely on simplified intensity-domain noise models, synthetic clean targets, or blind-spot self-supervision, and typically do not jointly model deblurring and speckle suppression. We propose a physics-consistent self-supervised framework for joint despeckling and deblurring of multibeam imaging sonar. A stochastic coherent image-formation model yields a conditional-mean forward operator that relates latent scene reflectivity to expected intensity through the squared magnitude of the system point spread function. We then use registered multi-frame sonar sequences to train without clean ground truth images, showing that, under an idealized inter-frame conditional-mean condition, the expected paired loss equals the corresponding forward-prediction loss up to a network-independent constant. To stabilize the ill-conditioned inverse problem, we further adapt robust equivariant imaging using stochastic coherent simulation and introduce a paired forward prediction error metric for ground-truth-free evaluation. Experiments on two real multibeam sonar datasets demonstrate a favourable balance between forward-prediction agreement, homogeneous-region speckle suppression, and preservation of visible structural boundaries compared with representative baselines. Comments: 16 pages, 8 figures, includes supplementary material Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Signal Processing (eess.SP) Cite as: arXiv:2610.31490 [eess.IV] (or arXiv:2610.31490v1 [eess.IV] for this version) https://doi.org/10.48550/arXiv.2610.31490 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-104] Bound-Conditioned Latent Inference for Progressive Image Compression NEURIPS2026
链接: https://arxiv.org/abs/2610.30966
作者: Jaeseok Jang,Seungmin Jeon,Kwang Pyo Choi,Chang-Su Kim
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to NeurIPS 2026
Abstract:Progressive image compression reconstructs images from prefixes of a single bitstream and therefore involves sequential latent inference under partial observations. Existing trit-plane codecs mainly treat decoded trits as discrete symbols for probability prediction, without fully exploiting their role as interval constraints on the underlying continuous latents. We propose bound-conditioned latent inference (BLI), a framework that reformulates progressive trit-plane coding as sequential latent inference under shrinking interval constraints. At each coding step, the trits decoded so far define lower and upper bounds for each latent element. BLI uses the center and width of these bounds, together with hyperprior information shared by the encoder and decoder, to jointly refine the Gaussian parameters (\mu,\sigma) for entropy modeling and the latent estimate \haty for reconstruction. As additional trits are decoded, the feasible intervals shrink monotonically, yielding tighter constraints for the next inference step and enabling flexible in-plane refinement schedules. Experiments on Kodak, CLIC, and JPEG-AI show that BLI achieves state-of-the-art rate-distortion performance among progressive image codecs, reducing BD-rate by 17.78% over DPICT and by 7.96% over CTC on Kodak, while replacing CTC’s plane-specific predictors with a shared model that uses about 5.3\times fewer parameters and achieves lower latency than CTC.
[CV-105] GateFormer: Forecast-Space Decomposition for Intra-Hour Solar Irradiance Prediction from Sky Images
链接: https://arxiv.org/abs/2610.30942
作者: David Hamlyn,Sunny Chaudhary,Jonathon Hare,Tasmiat Rahman
类目: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Sub-hourly dispatch, balancing, and storage control increasingly depend on forecasts that anticipate rapid variability in solar generation. Accurate sky-image-based irradiance forecasting requires a combination of visual cloud dynamics with recent irradiance state while adapting to different short-term irradiance-change regimes. To address this, we introduce GateFormer, a GHI-conditioned dual-decoder ViT for intra-hour global horizontal irradiance (GHI) forecasting through forecast-space decomposition. GateFormer embeds observed GHI into the visual token stream, allowing interaction with sky-image patch features through self-attention. From this shared representation, a Trajectory Head predicts future GHI sequence, while a Delta Head predicts final-horizon change from the latest observed GHI. The two final-horizon forecasts are adaptively weighted in prediction space, with magnitude- and direction-supervision targeting the remaining error. On the NREL-TSI benchmark, GateFormer attains 125.01 W/m^2 RMSE at +30 min, improving over a state-of-the-art transformer baseline by 9.6% under paired-timestamp evaluation. Across NREL-TSI, Chilbolton, and SIRTA, GateFormer achieves the lowest final-horizon RMSE among the evaluated baselines, with forecast skill of 22.97-28.86% relative to smart persistence. Component ablations show adaptive weighting outperforms both individual decoders and fixed averaging, distinguishing GateFormer from conventional forecast combination, while the supervised correction further reduces RMSE by 1.68 W/m^2 and MAE by 4.86 W/m^2. Ramp-regime diagnostics show larger gains during operationally important positive ramps, reducing RMSE by 26.27 W/m^2. These results identify forecast-space decomposition as an alternative forecasting formulation, with potential relevance to other multimodal short-horizon forecasting tasks and deployment using only sky images and measured GHI.
[CV-106] Vision Transformers for Estimating Irradiance Using Data Scarce Sky Images
链接: https://arxiv.org/abs/2610.30940
作者: David Hamlyn,Sunny Chaudhary,Tasmiat Rahman
类目: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Accurate estimation of diffuse horizontal irradiance (DHI) is critical for optimising photovoltaic system performance and energy forecasting yet remains challenging in regions lacking comprehensive ground-based instrumentation. Recent advancements using Vision Transformers (ViTs) trained on extensive sky image datasets have shown promise in replacing costly irradiance measurement equipment, but the scarcity of long-term, high-quality sky imagery significantly restricts practical implementation. Addressing this critical gap, this study proposes a novel dual-framework approach designed for data-scarce scenarios. First, calculated atmospheric parameters, including extraterrestrial irradiance and cyclic time encodings, are integrated to represent sky conditions without utilising any instrumentation. Next, a sequential pipeline initially predicts synthetic global horizontal irradiance (GHI) and uses it as a feature, to refine DHI estimation. Finally, a dual-parallel architecture simultaneously processes raw and overlay-enhanced fisheye sky images. Overlays are generated through unsupervised, physics-informed cloud segmentation to highlight dynamic sky features. Empirical validation is performed using data from the Chilbolton Observatory, chosen for its temperate climate and frequent cloud variability. To simulate data-scarce conditions, models are trained on a single month (e.g., January) and evaluated across a temporally disjoint, full-year test set. Under this setup, the sequential and dual-parallel frameworks achieve RMSE values within 2-3 W/m^2 and 1-6 W/m^2, respectively, of a state-of-the-art ViT trained on the complete dataset. By combining physics-informed modelling with unsupervised segmentation, the proposed method provides a scalable and cost-effective solution for DHI estimation, advancing solar resource assessment in data-constrained environments.
[CV-107] H-2DGS: Height-Field Gaussian Splatting for Sharp and Deformable Neurosurgical Surface Reconstruction
链接: https://arxiv.org/abs/2610.30910
作者: Zhang Zichuan,Bart Brouwers,Rembrandt Bakker,Maud Landers,Geert-Jan Rutten,Maxime Chamberland
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注:
Abstract:Neurosurgical navigation commonly presents magnetic resonance imaging (MRI) as 2D slices, making it difficult to relate the anatomy in those slices to the cortical surface observed through the surgical microscope. We seek a common 3D visualization that combines a photorealistic reconstruction of the exposed cortex during tumor surgery with whole-brain MRI context and diffusion tractography pathways. This requires the reconstructed cortex to form a clear boundary around the brain anatomy rather than a diffuse volume that obscures internal structures. Conventional 3D Gaussian Splatting (3DGS) can reproduce observed views but, under limited surgical view coverage, may distribute opacity through a thick and geometrically ambiguous band. We introduce height-field 2D Gaussian Splatting (H-2DGS), a three-stage pipeline that concentrates a robust 3DGS reconstruction, fuses its depth into a shared truncated signed distance field (TSDF) surface, and represents that surface with curved, surface-oriented Gaussian primitives. Separate height, color, and opacity bases provide local geometry and appearance, while a neighbor-transition objective coordinates overlapping patches. The explicit surface further supports postoperative alignment: MRI-derived depth deforms the reconstruction, and tracked-color supervision with alternating optimization preserves its appearance. Across five real neurosurgical cases, H-2DGS gives low multi-view depth inconsistency, a compact 7.97MB representation, and interactive rendering at 29–42ms per frame. The resulting sharp, deformable surface integrates clearly with tractography and MRI anatomy, including a postoperative resection cavity.
人工智能
[AI-0] METTA: Memory-Enhanced Test-Time Adaptation for Meta-Reinforcement Learning
链接: https://arxiv.org/abs/2610.31656
作者: Ali Najar,Hosein Hasani,Amir Kooshan Fattah,Yasamin Rajabi,Mahdieh Soleymani Baghshah
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 9 pages
Abstract:Meta-reinforcement learning aims to train agents that adapt quickly to unseen tasks from limited interaction, but this requires representing task evidence over long interaction histories. Classical recurrent agents compress experience into hidden states, while transformer-based agents retrieve from past trajectories at quadratic memory and computation cost. In this work, we propose Memory-Enhanced Test-Time Adaptation (METTA) for meta-reinforcement learning. It decomposes the fast adaptation phase into two temporal scales: a causal neural memory performs fine-grained within-episode adaptation with linear sequence complexity, while compact episode-level summaries support long-range cross-episode task inference without attending over the full interaction history. Moreover, equipping agents with neural memory provides sufficient expressive capacity and test-time training capability, enabling effective adaptation to unseen held-out tasks and task families at meta-test time. On unseen held-out Meta-World ML10/45 test environments, METTA achieves a success rate over 50% while maintaining a low training budget and short wall-clock time. On long-context sparse-reward PointMaze tasks, it outperforms transformer-based baselines by exhibiting more structured exploration behavior. These results show that METTA provides an efficient path toward more adaptive and scalable long-context meta-RL.
[AI-1] EpiSEM-Bench: Evaluating Safety and Evidence Recovery in LLM Agents under Missing Evidence
链接: https://arxiv.org/abs/2610.31650
作者: Qijin Zhang,Hanzhen Wang,Yuehan Zhao,Haoyu Dai,Yao Shu,Zhongxiang Dai,Shuang Qiu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Tool-using language model agents rely on environmental facts to decide whether an action is permitted. When a tool response omits a safety-relevant fact, an agent may carry out a prohibited action even though the user request is benign and every returned statement is true. We introduce \bench, an executable benchmark for evaluating unsafe execution, safety boundaries, and evidence recovery in LLM agents under missing evidence. It contains 300 matched unsafe/safe task pairs across six domains. By systematically suppressing critical facts, the benchmark measures unsafe execution under critical evidence omission and determines the minimum observed failure cardinality, including failures with full evidence and cases with no observed failure after exhaustive testing. An independent audit tool tests whether agents seek missing facts and then withhold prohibited actions while completing permissible tasks. The benchmark reveals substantial variation across the ten models in omission robustness, evidence acquisition, and appropriate action. Resistance to missing evidence does not necessarily translate into useful recovery. Recovery access reduces unsafe execution for all ten models, but agents differ in whether they actively obtain the missing facts and use them to act appropriately in both worlds.
[AI-2] Causal Chain of Thought: Mitigating Attention Leakage via Context Isolation
链接: https://arxiv.org/abs/2610.31639
作者: Gabriele Dominici,Teague McMillan,Fazl Barez,Marc Langheinrich,Martin Gjoreski
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Chain-of-thought (CoT) prompting elicits step-by-step rationales that appear to explain model predictions, but there is no guarantee these steps causally drive the final answer. In standard autoregressive generation, dense attention allows each reasoning step to attend to the full context, enabling the model to derive answers through direct lookup while generating plausible-sounding but post-hoc justifications. We introduce CausalCoT, an inference framework that enforces causal transparency by decomposing problems into a directed acyclic graph of atomic sub-questions, evaluating each node in an isolated context containing only its direct causal parents. This Markovian structure removes attention pathways to non-parent content, making downstream predictions sensitive to local reasoning rather than global context leakage. Across a diverse suite of reasoning benchmarks, CausalCoT preserves competitive predictive performance while improving explanation faithfulness, increasing causal sensitivity of intermediate reasoning steps by an average of 12 and up to 21 percentage points over the strongest baseline. The structured dependency graph additionally enables test-time interventions in which local corrections propagate predictably through downstream reasoning.
[AI-3] Questioning Instrumental Convergence in RLVR: Checkpoint Evidence from OLMo-3
链接: https://arxiv.org/abs/2610.31624
作者: Barış Alparslan,Sandy Tanwisuth,Caleb Petterson
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Instrumental convergence is often invoked as a general reason to expect capable agents to acquire or express power-seeking subgoals: self-preservation, resource acquisition, deception, or resistance to oversight which poses an important threat to the society. For LLMs, reinforcement learning is the first stage of post-training that directly optimizes behavior against an objective; therefore, it is the stage where convergent instrumental behavior, on IC grounds, likely to appear. Does reinforcement learning with verifiable rewards (RLVR) make a model more likely to express instrumental-convergence behavior, or does this depend on the training regime, evaluation frame, and stage of post-training? We study this question as a checkpoint-conditioned behavioral trajectory rather than as a single before/after safety score. Using InstrumentalEval, we evaluate OLMo-3-7B-Think across SFT, DPO, 13 intermediate RLVR checkpoints, and the final RLVR model. The trajectory is nonmonotonic: instrumental rate rises from 25.0% at SFT to 46.1% near RLVR step 200, then falls to 11.8% at the final checkpoint. Paired item-level tests and step-wise regression support the final SFT-to-RLVR reduction. Prompt framing does not explain this pattern: a harmful-prompt indicator combined with other framing effects explains only 15% of within-checkpoint score variance on average, and Iterative Null-space Projection (INLP) removes a detectable framing direction without changing the SFT-to-final-RLVR gap. Together with preliminary 32B endpoint results (35.53% SFT to 22.37% RLVR final), these findings indicate that RLVR does not monotonically increase instrumental behavior in this setting: the rate rises early, then falls below the SFT baseline. The result is specific to this model family and benchmark, and judge agreement is moderate, so absolute rates should be read cautiously.
[AI-4] Simulated Control Evals
链接: https://arxiv.org/abs/2610.31593
作者: Jeffrey Olmo,David Zhang,Tyler Tracy
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Control evaluations measure whether safeguards can prevent an AI model from causing undesired outcomes even when it is intentionally trying to do so. These evaluations typically use purpose-built control settings and extrapolate their results to a model’s deployment. However, the effectiveness of safeguards in deployment can depend on features of environments and tasks that control settings do not represent. To broaden where control evaluations can be run, we introduce simulated control evaluations (SCEs), which use a language model to simulate tool responses in place of an environment. We evaluate SCEs on LinuxArena, where execution provides ground-truth outcomes. We compare SCE prediction errors with those from a separate control setting, BashArena, and with variation in control metrics within LinuxArena. For independently sampled evaluation and deployment task sets, SCEs predict safety with a root mean squared error of 3.6 percentage points across five monitors at a 5% audit budget, compared with 10.1 for BashArena. SCEs closely reproduce distributions of attack strategies, attacker reasoning, and monitor reasoning, with over tenfold lower KL divergence from execution than BashArena. SCEs explain 78-85% of the between-environment variance in attack success, monitor scores, and safety. This accuracy is sample efficient: SCEs predict safety with less error than BashArena using fewer than 5% as many task-pair draws as the deployment task set. Strategies discovered through simulation also improve monitor performance on held-out LinuxArena attacks, outperforming BashArena and approximately matching improvement from executed strategies. These results suggest that deployment-specific SCEs can be more predictive than control evaluations executed in separate control settings.
[AI-5] VERA: A Responsible AI Dashboard for Assessing LLM Deployments in Light of the EU AI Act
链接: https://arxiv.org/abs/2610.31588
作者: Robin Quériaux,Solveig Bachellery,Léa Déléris,Saurabhsingh Rajput,Mariam Barry,Tushar Sharma,Jérôme Lebecq
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: Accepted at the 33rd Asia-Pacific Software Engineering Conference (APSEC 2026), Software Engineering in Practice track, Bali, Indonesia, 7-10 December 2026. 11 pages, 3 figures
Abstract:Institutions that deploy large language models in regulated settings often need technical evidence of responsible AI practices before approving a deployment. COMPL-AI provides a technical interpretation of the EU AI Act that organises this evidence under six principle groups and eighteen technical requirements, twelve of which current benchmarks can measure. The benchmark suites that produce this evidence are typically operated from a command line, while the staff who approve deployments are non-engineers. We present VERA, an open-source self-hosted dashboard deployed to facilitate responsible AI processes within a large European bank. A reviewer selects a model and runs public benchmarks against it. VERA scores the twelve requirements that current benchmarks can measure and instruments the energy consumed by each evaluation, adding run-level evidence for environmental impact. We run four served open-weight models through the requirement pipeline, and the dashboard displays a score for each measurable requirement. A within-subject study then compares how participants read these benchmark results: from the raw output files in one condition, and through the VERA dashboard in the other. The dashboard reduced the time needed to perform the responsible AI assessment, reducing the time required from 56 seconds using raw files to 30 seconds using our dashboard, on average. Our dashboard scored highly (4.36 out of 5) on usefulness in the survey conducted within the organization. The dashboard and the study material are released publicly to promote replicability and reproducibility in different organisational contexts.
[AI-6] What Makes a Good Unmasking Policy? A Controlled Study of Learned Unmasking in Diffusion Language Models
链接: https://arxiv.org/abs/2610.31576
作者: Jinya Sakurai,Xun Xu,Jaehong Yoon
类目: Artificial Intelligence (cs.AI)
备注: Preprint.32 pages, 11 figures
Abstract:Masked diffusion language models generate tokens in parallel, but their efficiency depends on which predictions are committed. Learned unmasking policies improve the speed–accuracy trade-off, yet differ in supervision, inputs, and capacity, obscuring which choices matter. We view supervised policies as distilling a decoding procedure through a teacher and examine independently what the teacher labels as committable, on which states the policy imitates it, what the policy observes, and how much the number of function evaluations (NFE) must decrease to offset policy overhead. We find that labeling predictions matching the final output is overly permissive at low NFE, and hidden states are weak alone but can complement confidence. About a 10% NFE reduction suffices to offset policy overhead for policies up to 7.5M parameters across tested settings. We therefore train a policy on trajectory-contraction targets, which accept only commitments staying on a reference trajectory, using confidence and hidden states. On LLaDA-8B-Instruct, the resulting 856K-parameter policy achieves the best accuracy–NFE frontier and highest throughput on GSM8K, is competitive at moderate targets on MATH-500 and HumanEval, and largely transfers to Dream-7B-Instruct.
[AI-7] Social-HUMINT: An LLM -Augmented Unified Framework for Measuring Social Media Exposure
链接: https://arxiv.org/abs/2610.31573
作者: Pellegrino Casoria,Simon Pietro Romano
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI)
备注:
Abstract:In the post-API landscape, social-media profiling has become inexpensive and automated: Large Language Models (LLMs) can extract structured information from unstructured content and deduce attributes users never disclosed. A person’s true exposure is not what they post but what can be inferred from it. As these practices become routine, exposure to social engineering attacks becomes an operational certainty, demanding a shared framework for assessment and mitigation. We make this exposure measurable by quantifying how accurately a target can be profiled, and how far that profile enables social engineering attacks. Because no existing model fully spans this continuum, we unify Social Media Intelligence (SOCMINT) and Cyber Human Intelligence (Cyber-HUMINT), two individual-focused disciplines that evolved separately, into a single doctrine-grounded taxonomy of 38 capabilities across the Cyber Threat Intelligence lifecycle, where user profiling bridges passive collection and active engagement. Domain experts validated the taxonomy (Fleiss’ \kappa = 0.92). We operationalize its core path (extraction, inference, and susceptibility assessment) as an LLM-augmented multi-agent pipeline, evaluated on real participants’ public data. Susceptibility is assessed indirectly: the pipeline predicts how each target would rank personalized engagement options, matching their actual choices significantly above chance. This ranking capacity captures the personalization on which tailored attacks depend. The underlying profiles are accurate: extraction reaches F1 = 97.3%, and inferred attributes score 4.73/5, both far above rule-based and NER-based baselines at negligible cost. The framework augments security analysts’ tradecraft rather than replacing it and enables defenders to gauge their own exposure before an adversary does. Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI) Cite as: arXiv:2610.31573 [cs.CR] (or arXiv:2610.31573v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2610.31573 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-8] When to Follow the Teacher: Advantage-Based On-Policy Distillation
链接: https://arxiv.org/abs/2610.31563
作者: Mahdi Sabbaghi,George Pappas,Hamed Hassani,Adel Javanmard
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注:
Abstract:Online Policy Distillation (OPD) often outperforms offline methods such as Supervised Fine-Tuning (SFT) by training on trajectories sampled directly from the student. However, standard OPD trains the student to follow the teacher at every state of the trajectory, assuming that teacher guidance is always beneficial. We instead distinguish between what the teacher prefers and what actually helps the current student succeed. We introduce Advantage-Based On-Policy Distillation (AOPD), which weights teacher actions by their counterfactual advantage under the student. Teacher actions that improve the student’s expected outcome are reinforced, while those that do not are downweighted. Directly computing these advantages requires trying alternative teacher actions at each step and running new rollouts from each. Instead, we show that the AOPD objective can be estimated using only standard student rollouts and their terminal rewards. Our estimator uses teacher importance weights on the student-sampled actions to provide unbiased advantage-weighted supervision, without any additional rollouts or reward evaluations. This allows AOPD to focus teacher supervision on actions that help the current student while preserving the rollout efficiency of standard OPD. Experiments on math, coding, and Medqa show gains over distillation and RL baselines in both external-teacher and self-distillation settings. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML) Cite as: arXiv:2610.31563 [cs.LG] (or arXiv:2610.31563v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.31563 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-9] Beyond Causal Masks: Auditing and Localizing Hidden Future-Token Leakage in Adaptive Sequence Models
链接: https://arxiv.org/abs/2610.31522
作者: Huiting Liu,Jinsong Wu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Hidden future-token leakage can create false architectural gains: an autoregressive model may appear substantially better under teacher-forced evaluation while its free-running generation deteriorates. We investigate this failure in an adaptive residual-routing decoder whose sequence-level routing summary exposes prefix positions to future tokens despite a correct causal attention mask. An anomalous 24-layer checkpoint reduces validation loss from the baseline’s 3.5958 to 1.4259, yet it fails a prefix-invariance audit and has substantially higher reference-model perplexity during generation. Layer-wise diagnostics and implementation inspection trace the violation to a full-sequence summary broadcast through the dynamic routing-query path. Two independently trained 12-layer causal repairs eliminate detectable leakage and return validation loss and reference perplexity to the control scale, removing the large apparent advantage. As diagnostic tools, we use the established prefix-invariance contract, maximum prefix-output discrepancy (LeakScore), complementary behavioral metrics, and layer-wise profiles. Supporting experiments cover six synthetic implementation pairs spanning five module families, clean pretrained controls, controlled fault injection, and cache consistency across numerical precisions. These results illustrate how hidden future-token access can produce misleading claims of architectural progress and why lower teacher-forced loss should be interpreted together with whole-model causal checks and free-running evaluation.
[AI-10] Jev-DLC: Decision-Level Continual Learning from Typed Feedback
链接: https://arxiv.org/abs/2610.31521
作者: Yangyang Feng,Jianle Gan,Zhuoyan Feng,Chengwei Qin
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Individual preferences often become visible only after deployment, when a user accepts one decision or revokes it in favor of another. Jev-style typed interfaces make this feedback explicit by identifying the selected action and, under revocation, the decision it replaces. We study how these consequences persist across requests as Decision-Level Continual Learning from Typed Feedback (Jev-DLC). Jev-DLC treats each correction as both evidence about a recurring action identity and a directed relation between the preferred and rejected choices. It preserves these signals in complementary bounded channels above a frozen scorer and combines their scores to shape the next decision. The method retains no raw interaction history, adds no scorer call, and performs no runtime language-model backward pass. We also introduce TCPA-Bench, a chronological suite that credits a correction only through its effect on later decisions. On TCPA-Repeat, Jev-DLC reaches 44.83% user-macro accuracy overall, 2.43 points above Recent-4 history replay. After at least two earlier feedback events, it reaches 56.07%, a 7.54-point margin over the strongest reported comparison at this depth. These results show how explicit typed corrections can become compact causal state for continual personalization.
[AI-11] From Prompt Rules to Golden Data Model: Production-Scale Generation of Personalized Compliant Pharmaceutical Claims
链接: https://arxiv.org/abs/2610.31519
作者: Suyash Mishra,Qiang Li,Subhabrata Adak,Upender Phogat,Srikanth Patil,Satyanarayan Pati
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) are increasingly used to generate domain-specific commercial medical content (e.g., claims, statements, and promotional phrases) in regulated industries such as pharmaceuticals. However, ensuring its regulatory compliance, factual accuracy, and audience-specific personalization remains a major challenge. We report a production case study asking whether moving compliance and personalization signal from prompt-level rules into retrieval time structured context improves pharmaceutical claims generation, using Composer, a deployed internal tool built on a Golden Data Model (GDM) of 1,867 pharmaceutical claims spanning four products. The GDM integrates claim metadata, references, audience labels, material type, usage history, Match Text Variant (MTV) links, and a six dimension quality score (6C). Across six product-audience segments, the context pipeline improves intended indication (+29%), intended content type (+62%) and intended target audience (+67%), 6C quality score (+86%), and claim reuse (+34%). Production log analysis over the deployment period shows a 29.5% reduction in weekly API cost. We report this as descriptive deployment evidence. This work is a systems and data architecture contribution: it shows that in a regulated, high-stakes domain, restructuring claim data for retrieval-time filtering is a more scalable lever than expanding prompt-level rules, and documents the engineering path data cleaning, graph linking, and quality scoring.
[AI-12] DuoCook: An Overcooked-Inspired Benchmark for Visual Planning and Coordination
链接: https://arxiv.org/abs/2610.31511
作者: Ziyi Jiao,Heng Zhang,Drishti Regmi,Xiaokang Ye,Lianhui Qin,Peter Yichen Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Coordinating two players in a cooking task requires a vision-language agent to track changing scene states, sequence recipe steps, and allocate shared resources under a time budget. We introduce an Overcooked-inspired benchmark for evaluating these capabilities in Unreal Engine environments constructed using coding agents. A single vision-language model controls both players from top-down visual observations, selecting high-level actions while the engine executes navigation. The benchmark contains 20 maps in five categories. These categories support evaluation of centralized coordination under equipment contention, changing terrain, and handoff dependencies. A prop legend and pot ingredient-count indicators provide explicit visual cues, while grid-based destinations support spatial grounding. The evaluation compares four VLMs on all 20 maps, measuring task score, order completion, delivery precision, action acceptance rate, and model API expenditure. A memory ablation on the same 20 maps examines the effect of a shared, self-authored notebook while retaining six-response action history in both conditions. The strongest evaluated model, \textttgpt-6-astra, achieves 26.2% mean order completion. Notebook memory has mixed score effects across models, with positive mean differences for Claude and Qwen and negative differences for GPT and Gemini.
[AI-13] FairFRESH: A Neuro-Symbolic Framework for Fair and Explainable Household Food Recommendation
链接: https://arxiv.org/abs/2610.31506
作者: Kodur Sai Vinay Sathvik,Kakoli Bora,Tejaswini R. M
类目: Artificial Intelligence (cs.AI); Computer Science and Game Theory (cs.GT)
备注: 22 pages, 4 figures. Accepted and presented at ICCIS 2026; this version includes additional theoretical analysis and experiments
Abstract:Recommender systems typically serve one user at a time. However, in households, members share a finite pantry and follow different dietary rules, making independent recommendations prone to resource conflicts and dietary violations. We formulate household food recommendation as constrained fair division and present FairFRESH, a neuro-symbolic framework that learns per-member utilities, allocates foods under explicit welfare objectives, enforces dietary rules through a deterministic filter, and explains recommendations using Shapley and Banzhaf attributions. Theoretically, we show that no finite penalty guarantees compliance near a dietary-rule threshold, establish approximation guarantees for greedy allocation under utilitarian and smoothed Nash objectives in the stock-free model before filtering, and characterize the scale invariance of Nash welfare. We also use the minimax theorem to analyze rotating shared meals. On NHANES-derived data, the learned utility model achieves Recall@20 of 0.293, outperforming five learned recommenders (at most 0.244) but not a previous-day repetition heuristic (0.321). On held-out four-person households, Nash allocation reduces Gini disparity by 42% relative to conflict resolution while improving Recall@20 from 0.180 to 0.197. It comes within 2% of an upper bound on the smoothed Nash optimum, and rescaling one member’s utilities changes at most 1.5% of assignments. All recommendations satisfy the encoded dietary rules, while attribution scores correlate with utility loss upon component removal (r = 0.76). In simulation, shared meals reduce pantry use by about one-third, and weekly rotation increases the worst-off member’s utility by 38% over the best fixed meal.
[AI-14] SABLE: Scalable Affine Bayesian Latent-space Exploration
链接: https://arxiv.org/abs/2610.31495
作者: Erik Bodin,Lisa Maria Kreusser,Donney Fan,Henry Moss,Geoff Pleiss,Nicolas Durrande
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注:
Abstract:Complex objects such as engineering designs and proteins are often difficult to optimise because their design spaces exhibit high-dimensional structure and objective evaluations are expensive. Generative models provide continuous latent representations, but their latent spaces can have thousands to tens of thousands of dimensions, creating a regime where the latent dimension D greatly exceeds the evaluation budget. We introduce SABLE, a Bayesian optimisation method for high-dimensional latent spaces with an isotropic distribution and limited budget. SABLE uses a Bayesian affine surrogate for prediction and acquisition, treating higher-order structure as surrogate discrepancy. After N observations, the slope update has dimension at most N-1 , yielding a decomposition of each query into an in-span component optimised by expected improvement (EI) in at most N dimensions and an off-span component governed by a scalar norm. We show that the off-span norm implicitly selected by affine EI can lead to underexploration, and SABLE instead prescribes it using a finite-horizon schedule, without restricting search to a learned subspace. A recompression scheme reduces information to \mathcal O(M) pseudo-observations, bounding refitting cost by \mathcal O(M^2D+M^3) while keeping parameter inference, prediction, and acquisition decisions to tens of milliseconds. Across synthetic benchmarks and latent-space optimisation tasks, SABLE achieves high sample efficiency and reduces decision time by several orders of magnitude at large budgets.
[AI-15] raining Small Language Models for Transferable Tool Selection
链接: https://arxiv.org/abs/2610.31483
作者: Zhaoyi Zhou,Akshayaa Magesh,Jeongyeol Kwon,Artem Kuriksha,Daniel Jiang,Andrea Zanette,Ayush Jain
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 19 pages, 5 figures, 5 tables
Abstract:As LLM agents operate over corpora of more than ten thousand tools, selecting the right tool for a complex multi-step user instruction becomes a central bottleneck. Tool descriptions cannot all fit in context, and relevance-based retrievers only rank tools without knowing which are actually needed. Recent work instead leverages proprietary commercial LLMs like Claude or ChatGPT to drive a retriever, decomposing the instruction and selecting over retrieved candidates. We show that a small, open model can be trained via reinforcement learning to perform the same functionality. On ToolBench, a Qwen3-8B model post-trained with our recipe outperforms Llama-4-Maverick (402B). In addition, the trained model transfers directly to different retrievers and generalizes to new tool corpora in different domains.
[AI-16] Adaptive Skill Granularity for Coding Agents : Runtime Routing and Macro-to-Micro Recovery
链接: https://arxiv.org/abs/2610.31461
作者: Dingxuan Zhang,Yanhui Jiang,Jun Wang,Hefeng Zhou
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Reusable skills allow large language model agents to package task-specific procedures, tool affordances, and execution guidance into modular units. Prior work has largely studied how agents retrieve, compose, or synthesize skills before execution. In coding agents, however, execution poses an additional runtime control problem: the same task may be solved through primitive tool calls, macro skills, or hybrid execution that starts with a macro and falls back to micro-level recovery. We introduce Adaptive Skill Granularity, a framework that treats skill granularity as an action-abstraction policy conditioned on task structure, execution state, and macro-skill failures. The system combines a feature-based granularity controller and a finite-state recovery executor. We evaluate the framework with a deterministic coding-agent protocol using verifier-defined success, total cost, and process metrics such as macro adoption, failure recovery, route match, and transition correctness. Results show that fixed macro execution is efficient when applicable but brittle under mismatch, while runtime routing and explicit macro-to-micro recovery improve robustness and recover controlled macro failures that fixed-granularity execution leaves unresolved. Our findings suggest that skill libraries for coding agents should be treated not only as retrievable resources, but also as controllable, verifiable, and recoverable execution abstractions.
[AI-17] DIAL: Steerable Language Models with Distilled Instruction-Adherence Levels
链接: https://arxiv.org/abs/2610.31438
作者: Thanh Q. Tran,John Russell Himawan,Arun Verma,Kiwan Wong,Wei Xiao,Daniela Rus,Bryan Kian Hsiang Low
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: This paper introduces DIAL, a framework for fine-grained, efficient, and scalable control of LLM instruction adherence
Abstract:When a user says “Keep it short”, how short should the LLM’s response be? More generally, users and applications require different levels of instruction adherence, but prompting does not reliably provide fine-grained control over them. We therefore study “steerable language models” that provide fine-grained numerical control over the degree to which models follow general instructions specified in text. However, existing approaches do not jointly offer fine-grained control, efficiency, and scalability across diverse instructions. To address this gap, we introduce DIAL (Distilled Instruction-Adherence Levels) to achieve all three properties. For fine-grained control, we perform classifier-free guidance (CFG) with a frozen LLM to provide training signals for on-policy distillation of instruction adherence at different levels. For efficiency, we design a lightweight steering adapter that uses a classifier over the LLM’s hidden states to compute a closed-form steering edit. The classifier adaptively sets the steering magnitude at each token based on the target adherence level, while accounting for the geometry of the LLM’s hidden states to preserve response coherence. For scalability, the classifier is text-conditioned, allowing a shared adapter to support new instructions without per-instruction fitting, and training requires only instruction and request text. Evaluation across multiple LLM families shows that DIAL achieves a competitive balance between concept expression and response fluency, with concept scores increasing monotonically across adherence levels. We introduce SteerIF, an instruction-only dataset of 98, 625 concept, persona, verifiable, and descriptive instructions. When trained on SteerIF, DIAL provides fine-grained control over personas and, when combined with prompting, improves verifiable instruction following.
[AI-18] PhysTrace: Interpretable Temporal Evaluation of Physical Consistency in Video Generative Models
链接: https://arxiv.org/abs/2610.31429
作者: Taekoan Yoo,Wonkyung Jung,Kyunghun Kim,Kyeongbo Kong
类目: Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: 17 pages, 7 figures (including supplementary material). Code: this https URL
Abstract:World action models increasingly rely on pretrained video models assumed to capture physical dynamics for prediction, planning, and action generation. Existing evaluations assess plausibility, probe representations, or measure predictive surprise, but rarely identify which history-conditioned physical expectations conflict with observations. We introduce PhysTrace, a training-free framework aligning MLLM-generated, history-conditioned expectations with observed video windows to temporally localize physical violations and provide interpretable descriptions without task-specific probes or classifiers. We also introduce IntPhys2-Timing, adding violation intervals to IntPhys2. PhysTrace achieves a Hit Rate of up to 70.7% across physical principles on IntPhys2-Timing and outperforms the compared methods in overall possible-impossible discrimination accuracy on IntPhys2. Applied to reconstructed videos along the trajectories of pretrained diffusion/flow-based video models, PhysTrace reveals attribute- and noise-dependent patterns in physical violation localization, enabling interpretable physical consistency analysis across timesteps.
[AI-19] Predictable performance is suboptimal and what it means for model routing
链接: https://arxiv.org/abs/2610.31428
作者: Guanhua Zhang,Moritz Hardt
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) differ in capabilities and costs, motivating routing methods that adaptively choose a model for each query. These methods rely on performance prediction: given a query, estimate whether a model will answer it correctly. Most routing methods train separate predictors of candidate models’ performance to guide model selection. Our work shows that performance prediction only beats a simple baseline when the candidate model is suboptimal on the task or the performance predictor adds significant capacity relative to the candidate model. Specifically, we take a model’s self-confidence, the probability it assigns to its chosen answer, as a baseline and study how much a performance predictor can improve on it. First, we show that any performance predictor that outperforms a model’s self-confidence can be used to improve the model’s loss: the predictor’s advantage transfers exactly in log loss through recalibration. Second, accounting for the predictor’s additional parameters, we bound its advantage by the model’s suboptimality plus the gain from those parameters. Third, we extend our theoretical results to two-model cascades, which route queries to a more capable model when the first model is predicted to fail. Experiments with 16 language models on four benchmarks show that predictors gain a larger advantage over self-confidence on more suboptimal models. This advantage shrinks as models improve through finetuning. Cascading exhibits the same two patterns: predictor-based routing offers larger gains over confidence-based routing for more suboptimal models, and these gains diminish after finetuning.
[AI-20] FocusVLA: Future-Intent Guided Dynamic View Routing for Vision-Language-Action Models
链接: https://arxiv.org/abs/2610.31423
作者: Huixin Xu,Ziya Yu,Peng Li
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: Accepted at CoRL 2026
Abstract:Vision-language-action (VLA) models have become an important paradigm for robotic manipulation by unifying visual perception, language understanding, and action generation. Since reliable manipulation requires sufficient observation of objects, end-effectors, and their spatial relationships, multi-view inputs are widely used to provide complementary spatial information. However, our experiments show that existing VLA models do not fully exploit such multi-view complementarity. The policy tends to persistently rely on the wrist view, lacking the ability to dynamically adjust global-local view dependency across manipulation stages, which limits stage adaptability and robustness. To address this issue, we propose FocusVLA, a stage-aware multi-view routing framework. FocusVLA predicts future interaction intent from historical actions and uses it to guide dynamic view routing during action generation. By adaptively coordinating global and local observations across different task stages, FocusVLA improves multi-view utilization. On RoboTwin 2.0, FocusVLA improves the average success rate over the strongest baseline \pi_0.5 by 10.10 pp. In real-world evaluation, FocusVLA improves the success rate from 71.7% to 85.0%, achieving a 13.3% gain over \pi_0.5 . Moreover, the intent prediction and dynamic routing modules of FocusVLA can be integrated into existing VLA frameworks without modifying their core architectures.
[AI-21] Post-Hoc Feature Selection Layer for Neural Networks Interpretability IJCNN
链接: https://arxiv.org/abs/2610.31420
作者: João Kenji Suwa,João Pedro Silveira e Silva,Bruno Iochins Grisci
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 6 pages, 3 figures, 4 tables. Accepted and published at the 2026 International Joint Conference on Neural Networks (IJCNN), Maastricht, Netherlands
Abstract:The interpretability of complex neural networks remains a critical challenge, especially for models already deployed in high-stakes domains. To address this, we adapted the Feature Selection Layer (FSL) to optimize feature weights post-training, a variant we term post-hoc FSL. Our approach reframes the FSL as a lightweight, trainable module that integrates with already frozen pre-trained models on tabular datasets to highlight the features the original model considers most important. This post-hoc FSL learns feature relevance by fine-tuning its weights based on the original model’s learned outputs. Crucially, this process is non-invasive, operating without altering the original model’s architecture or its learned parameters. We conducted our experiments using both statistical and visual metrics, including accuracy, F1 score, recall, precision, weighted t-SNE and silhouette score, and also analyzed the stability of the post-hoc FSL on high-dimensional synthetic and real-world tabular datasets. We compared the post-hoc FSL feature weighting method using these metrics against the original embedded FSL and other post-hoc interpretability methods, such as Integrated Gradients, Noise Tunnel, DeepLIFT, Gradient SHAP, and Feature Ablation. Experimental results demonstrate that the post-hoc FSL feature weighting method successfully identified relevant features across the different datasets, maintaining the predictive power of the original neural network while enhancing its interpretability.
[AI-22] Major Revision Minor Accountability: LLM s as Reviewers of Published Articles in Top-Tier Journals
链接: https://arxiv.org/abs/2610.31414
作者: Bastien Le Guellec,Su Hwan Kim,Raphael Bentegeac,Aghiles Hamroun
类目: Digital Libraries (cs.DL); Artificial Intelligence (cs.AI)
备注:
Abstract:With the growth of scientific output, the peer review system is approaching a breaking point: reviewers are increasingly hard to secure, and turnaround times lengthen accordingly. LLM-assisted or LLM-generated peer review has been proposed as a remedy, but how such models would treat manuscripts already published in top-tier journals, the closest thing the current system has to a gold standard of scientific validity, is not known. Here we assembled a corpus of 200 articles, each released under a CC BY 4.0 license and published in 2026, from 20 general-science, life-science, and clinical-medicine journals including Nature, Science, and Cell, and asked four current LLMs (two proprietary, two open-weight) to produce independent peer-review reports of each one under a blinded protocol. Their decisions were overwhelmingly harsh and starkly inconsistent with each other: 605 of 800 reviews (75.6%) recommended major revision or rejection, but the rate at which any single model did so ranged from 17.5% to 100% (p 0.001), disagreement on manuscripts that had, by construction, already survived real editorial review at the journals in question. When the same manuscripts were doctored with one deliberately planted, field-specific methodological or statistical flaw, detection rates varied from 17.0% to 60.5% across models (p 0.001). Recommending revision for papers that had already passed editorial review (low specificity) while missing flaws deliberately placed in front of them (low sensitivity), these models produce output that should be read for what it is: an uneven, framing- and model-dependent signal.
[AI-23] Agent Phishing: A Benchmark of LLM Email Agents Under Adaptive Attack
链接: https://arxiv.org/abs/2610.31404
作者: Justin Tan,Miao Li,Ding Zhao
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 28 pages, 6 figures
Abstract:Computer-use agents increasingly handle private information and act on users’ behalf across desktop and web applications, including email. Phishing therefore poses privacy and security risks that depend not only on whether agents recognize malicious messages, but also on how they respond. Existing phishing datasets were developed primarily for conventional detection, leaving agent susceptibility to persuasive malicious requests insufficiently characterized. To expose vulnerabilities that pre-existing datasets may miss, we introduce AgentPhishing, a benchmark for evaluating whether persuasive phishing attacks can evade detection and induce LLM email agents to comply with malicious requests. We develop PhishOpt, which optimizes reusable email-generation skills against individual agents under fixed benchmark constraints and a realism constraint. We analyze the generated emails using Cialdini’s persuasion principles to identify strategies associated with agent failures and use these findings to construct a universal generation skill. Our evaluation shows that agents can handle pre-existing phishing emails safely yet remain susceptible to attacks adapted to agent behavior, while in an AgentPhishing evaluation of nine agents, four exhibit phishing-detection rates below 10% and malicious-request compliance rates above 45% . These results highlight the limits of inferring agent safety from conventional phishing evaluations and the need to assess how agents respond to persuasive malicious requests.
[AI-24] Auditing the Hand-Written Inputs of Learned Access Control
链接: https://arxiv.org/abs/2610.31388
作者: Arsalan Jalil Sarghaleh,Mehdi Hazrati
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Learned access control is typically deployed behind two hand-built artefacts: a hard-deny set enforced by a small trusted shield, and a region map that decides which requests mined rules answer and which the learner answers. Verifying the shield or measuring the learner validates neither. We audit both on the five public ABAC Lab policies. First, we formalise hard-deny soundness and check it exhaustively over relations of up to 794,250 tuples. Clauses that faithfully transcribe an explicit restriction are unsound at rates of 0.50-0.70 across four elicitors - an author and three blind language-model elicitors - because prose scopes restrictions to a section while the grant relation is a disjunction over all rules: 97% of counterexamples are granted by other rules. Shipped unchecked, these deny sets would withhold 16-78% of everything the policies grant; certified, they withhold nothing. Vacuity, a second silent failure, also occurs. Second, hybrid routing needs a two-class, abstaining rule set that ABAC miners do not emit, and ranking a mixed-effect candidate pool under a cap is one-class in all 24 cells for the two commonest ranking keys. Third, for additive losses the hybrid-versus-learner effect factorises into covered mass times a rule-learner risk contrast that must be reported on both false-non-grant and false-permit loss: on two real policies, most selection rules give opposite verdicts on the two losses, and the one favourable on both holds for one policy, one learner configuration and a band of noise levels. A released artefact regenerates and checks every reported number.
[AI-25] AI Risk Ontologies for the EU AI Act Compliance: AIRO VAIR and their Integration with DPV
链接: https://arxiv.org/abs/2610.31380
作者: Delaram Golpayegani,Harshvardhan J. Pandit,Declan O’Sullivan,Dave Lewis
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注:
Abstract:As the use of AI, and consequently its harms, becomes more prevalent, the global discourse around AI has become increasingly risk-aware. With new risks emerging more frequently than ever, a key challenge is how to represent, document, and catalogue AI risks in an effective, interoperable, and queryable manner. Additionally, the European Union (EU) AI Act’s requirements for “high-risk” AI systems to manage and document the risks creates the need for open and interoperable tools, methods, and standards that facilitate AI risk management in an ever-changing environment. In this paper, we present ontologies and taxonomies developed to assist with risk management and documentation, in line with the requirements of the EU the AI Act. We developed the AI Risk Ontology (AIRO) to specify AI use cases and their associated risks, complemented by the Vocabulary of AI Risks (VAIR), which populates AIRO with taxonomies of concepts. Key international standards on AI and risk, namely ISO/IEC 23894 on AI risk management and ISO/IEC 22989 on AI terminology informed the development of the ontologies. To promote reuse and enable multi-regulatory compliance, AIRO and VAIR have been integrated in the Data Privacy Vocabulary (DPV) through a consensus-building process, expanding DPV’s scope from data to AI.
[AI-26] Certified Execution Harness for Agent ic AI: Safe and Least-Intrusive Petri-Net Supervisory Control Between Reasoning and Action
链接: https://arxiv.org/abs/2610.31376
作者: Chao Gu,Quanyan Zhu
类目: Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注: 17 pages, 9 figures, 2 tables. Extended version of a paper submitted for publication
Abstract:This paper studies the problem of safe tool execution for large language model (LLM) agents under prompt injection. Using Petri nets, we formalize the execution harness between the LLM and its tools as a supervisory control problem, in which marks represent capabilities, approvals and budgets. Since a transition fires only if the marks it needs are present, an injected request whose preconditions are unmet or already spent is refused regardless of its wording. Treating tool outcomes as uncontrollable, we enforce safety by the conventional maximally permissive supervisor, and this safety transfers to the running system under four assumptions. Moreover, to capture an agent that may propose again after a refusal, we introduce the notion of a replanning closed loop, in which each retry costs an inference call. We show that, in this loop, the conventional supervisor can be arbitrarily more intrusive than a stricter safe one. Furthermore, for a given proposer model, the maximally permissive least-intrusive supervisor is characterized as a value-threshold refinement of the conventional one. Under precise conditions, this refinement orders two concurrent actions by at most one monitor place. On an autonomous release agent, we find by synthesis that safe deployment requires a step that the supervisor can refuse after a rollout starts.
[AI-27] RIDAR: Safeguarding Research Integrity in Autonomous Research Agents via Automated Detection and Runtime Intervention
链接: https://arxiv.org/abs/2610.31372
作者: Jiening Siow,Amy Xin,Junjie Wang,Ziyu Tang,Zijun Yao,Yunjia Qi,Fanjin Zhang,Lei Hou,Jie Tang,Juanzi Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Autonomous research agents are increasingly used to automate large portions of scientific research, yet they are prone to research integrity violations (RIVs): behaviors that compromise the validity, reproducibility, or faithfulness of the research process. We present RIDAR, a framework for automated RIV detection and runtime intervention. First, we introduce RIVT, a comprehensive taxonomy of 14 RIV types spanning the full research lifecycle. Building on RIVT, we present RIDAR-post-hoc, a framework to detect RIVs within agent trajectories. RIDAR-post-hoc achieves 85.2% F1 performance on RIV-Gold, a benchmark for RIV detection with 60 expert-annotated trajectories. Applying RIDAR-post-hoc to RIV-Wild, a dataset of 300 agent trajectories across 50 distinct research tasks, we detect 3,964 RIV instances in 99.7% of trajectories, revealing the prevalence of RIVs in current agents. Finally, we present RIDAR-runtime, a framework for dynamically detecting and correcting RIVs during agent runtime. RIDAR-runtime effectively reduces agent RIVs remaining at task completion by 36.8%. We will release our code and data to facilitate future research.
[AI-28] ValuSpec: Plug-and-Play Candidate Valuation before Target Verification for Tree-Based Speculative Decoding
链接: https://arxiv.org/abs/2610.31338
作者: Liang He,SiYuan Ma,Qishi Zhan,Yongqi Fan,Mingyu Cao,Qizhen Lan,Zhaolu Kang,Xilu Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Tree-based speculative decoding accelerates large language model inference by verifying multiple candidate branches in parallel. However, in practical serving scenarios, similar requests repeatedly induce overlapping candidate branches. While retrieval-based methods mitigate this by reusing historical drafts, the resulting candidate trees often contain many low-quality branches as the candidate tree grows, which reduces verification efficiency and wastes target-side computation. To address this, we propose ValuSpec, a plug-and-play candidate valuation mechanism that filters low-quality branches from a supplied candidate tree before expensive target-side verification. ValuSpec introduces transfer-based supervision to align filtering with the target model’s behavior. A filtering threshold is estimated from the target model’s greedy trajectories on a calibration split, without any parameter training. The target model then verifies the filtered tree under the output-identical greedy rule, so accepted tokens are guaranteed to match standard autoregressive decoding exactly. Experimental results on HumanEval, GSM8K, Dolly, and SpecBench show that, in our main evaluation, ValuSpec achieves up to 2.38x end-to-end speedup in retrieval-based tree settings. Furthermore, integrating ValuSpec into the EAGLE-3 dynamic tree framework yields up to 2.78x end-to-end speedup without modifying its tree construction strategy, demonstrating that ValuSpec serves as a generally applicable and plug-and-play candidate valuation module.
[AI-29] Agent icEvolve: Evolutionary Program Search with Persistent Language Model Agents that Design and Run Experiments
链接: https://arxiv.org/abs/2610.31326
作者: Ze-Yu Peng,Qing-Yu Lan,Zhan-He Wang,Jun Zhang,Yun-Song Piao
类目: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 23 pages, including supplementary material. The source code is available at this https URL
Abstract:Evolutionary program search with large language models keeps the structure of an evolutionary algorithm and uses the model as its variation operator. Typically, the model generates each candidate program in a single request from parent programs and their scores. We present AgenticEvolve, an evolutionary program search framework in which each island is run by a persistent large language model agent. Between evaluations, the agent designs and runs experiments on the candidate program. To measure the contribution of experimentation, we compare AgenticEvolve with No-analysis, an otherwise identical configuration with experiments disabled. On two heuristic design tasks with three large language models, AgenticEvolve produces better programs than No-analysis and OpenEvolve on both hidden and unseen instances. It reaches the baselines’ final quality with far fewer evaluations and comparable cost. On twelve equation discovery problems, it recovers the ground-truth structure of more equations than No-analysis and existing language-model and genetic programming baselines.
[AI-30] Bridging the Data Gap: Source-Informed Task-Space Transport for Few-Shot Diffusion
链接: https://arxiv.org/abs/2610.31323
作者: Jinjian Jiang,Jing Peng,Rehan Deen,Yu Cao,Ziquan Liu,Muhammad Salman Haleem,David Mguni
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Many deployments require a model to move rapidly from a data-rich source to a related target observed only sparsely. This is especially consequential where target data are difficult to collect, as in healthcare settings with persistent population or acquisition imbalances. Retraining or fine-tuning a high-capacity generator for every target is costly and can be statistically unreliable when target data are scarce. We introduce Meta-Diff, which instead learns a recurring source–target relation once in a compact task space. A shared encoder maps each distribution to a low-dimensional coordinate; a learned relation transports a new source coordinate to its predicted target, and a shared diffusion backbone generates from that prediction without target-specific retraining. Controlled Gaussian-mixture experiments identify when this works: task identity must survive the representation, the relation must generalise across tasks, and the source–target shift must leave useful headroom for transport. On held-out CIFAR-100 validation tasks, correct source + transport achieves 13.9% lower mean SW than transport from the wrong source. With validation-selected adaptation settings in the final-test one-target comparison, its mean SW is 14.7% lower than target-only modelling and 32.7% lower than full fine-tuning; the main route itself uses no target image and assumes the applicable relation is known. On matched CIFAR test episodes, generating from the clean-source coordinate and then applying the known corruption yields lower SW than task-space transport. Office-Home exposes a complementary boundary: transport can align target-domain statistics without reliable object identity. Together, these results recast extreme-few-shot adaptation as reusable transport in task space, while making the conditions for its success and failure directly measurable.
[AI-31] Hyperreduction accelerates finite element model updating for the discovery of neural materials
链接: https://arxiv.org/abs/2610.31322
作者: Hagen Holthusen,Matthias Knipper,Kevin Linka
类目: Computational Engineering, Finance, and Science (cs.CE); Artificial Intelligence (cs.AI)
备注: 42 pages, 23 figures, 1 table
Abstract:Finite element model updating is among the most general approaches to discover constitutive models from structural experiments: since the displacement field is computed rather than measured, it accommodates arbitrary geometries, boundary conditions, and sparse or noisy data. However, every loss evaluation requires a nonlinear finite element solve per load step, and training neural materials requires many such evaluations. Intrusive model order reduction lowers this cost while retaining boundary conditions and variational structure. How can a reduced model remain accurate while the network weights change at every training step? We combine proper orthogonal decomposition and energy-conserving sampling and weighting with adjoint sensitivities of the reduced model, and adapt both the trial space and the sampled elements along the optimization path. The constitutive law is a polyconvex neural network: a logarithmic normalization term, together with a zero-anchored input convex architecture and a growth term, renders the energy polyconvex, coercive, and non-negative by construction, independent of the trained weights. We demonstrate the approach on two experimental datasets and one synthetic three-dimensional example. For a two-phase specimen, training time decreases from nine hours to 41 minutes compared with a full-order identification, without loss of accuracy. In the three-dimensional example, six modes replace 48 672 degrees of freedom, 116 of 15 625 elements suffice to evaluate the internal forces, and training takes 13 minutes. In all cases, the identified material reproduces the measured reaction force and predicts the displacement field from the force alone. Hyperreduction makes the discovery of neural materials from structural data affordable while retaining the advantages of finite element model updating.
[AI-32] Subgoal-aware Memory Graph for GUI Agent
链接: https://arxiv.org/abs/2610.31307
作者: Dezhao Luo,Daniel Beechey,Derek Yuen,Weilin Luo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:GUI agents aim to automate complex user tasks by interacting with mobile interfaces, but ef- fective decision-making in long-horizon set- tings requires reusing prior experience rather than relying solely on the environment ob- servations. Existing memory-based methods typically store experience either at the single- action level or as entire trajectories. Single- step actions are often too fine-grained to carry meaningful semantics, since the success of an individual action does not necessarily imply progress towards the overall goal. In contrast, whole trajectories are often too coarse-grained and difficult to reuse across different tasks. We propose a subgoal-aware memory graph that distils past trajectories into structured, reusable evidence at the level of subgoals. We define a subgoal as consecutive actions shar- ing a single intent, the granularity at which progress toward the task becomes meaning- ful. To reduce redundancy and support cross- task reuse, we organise subgoal experiences as nodes in a graph, where related memories can be merged, extended, and retrieved based on the environment observation. On SPA-Bench, our method improves success rate by 11.11% over the baseline.
[AI-33] Superintelligent Software Engineering: Governance That Scales with Capability
链接: https://arxiv.org/abs/2610.31306
作者: Mohammad Mehdi Gholinejad
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 16 pages, 6 figures, 6 tables. Collection scripts and dataset are provided as ancillary files. The OSTEON artifact is archived at this https URL
Abstract:Software agents are being trained to exceed human engineers; we argue that capability alone does not produce better software engineering. We define superintelligent software engineering as the discipline of building and evolving software with agents whose capability exceeds that of human engineers, in which governance (gates, contracts, decisions and decay control) scales with capability. The argument rests on three claims. First, correctness in a project rests on decisions and contracts that the code does not carry. Second, knowledge decays. Third, once an agent outpaces its reviewers, reading catches a shrinking share of the changes built on stale knowledge. A simple model shows that, past a fixed review capacity, escaping changes grow linearly with throughput while the share that reading catches falls toward zero; governance can lower the slope, capability alone cannot. An exploratory snapshot of 452 of the most-starred public GitHub repositories finds a root-level or .github agent context file in 48.67% (95% confidence interval 44.1 to 53.3); in the median file that names repository paths, 9.5% to 22% of those paths do not resolve, depending on how a reference is read. We then present OSTEON, a discipline for the regime, released with a DOI (https://doi.org/10.5281/zenodo.23436020)%3A each project gets a governed knowledge repository of decisions and contracts, every change is checked mechanically against it, and the knowledge itself is checked for decay at every session start; planning is gated, and sessions record what they learn without writing rules. We report no measured effect of OSTEON, mark claims about its mechanisms as opinion and the author’s practice heuristics as design rationale, and state the questions that would falsify the thesis.
[AI-34] Beyond Accuracy: Disentangling Decomposition from Typed-State Quality Assurance in Multi-Agent LLM Pipelines
链接: https://arxiv.org/abs/2610.31280
作者: Maria Gabriela Alves Zuppardo,Vinicius Cardoso Garcia
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Machine Learning (cs.LG)
备注: 10 pages, 10 tables. To be published at Brazilian Symposium on Software Quality – Simpósio Brasileiro de Qualidade de Software (SBQS)
Abstract:Multi-agent LLM pipelines are often credited with gains in software engineering tasks, but it is rarely clear whether a gain stems from the typed contract between agents or from the specialized prompts that decomposition enables. We study this attribution problem in Requirements Engineering, comparing three architectures for classifying requirements as functional or non-functional and then prioritizing them with MoSCoW: a single-call baseline; a two-call variant that reuses the pipeline’s specialized prompts but passes an unvalidated intermediate; and a pipeline with a Pydantic-validated typed state. The 18 conditions cover PROMISE NFR+ (625 requirements), three locally hosted small language models, and two languages. The F1 gain over the baseline is associated with the two-call, specialized-prompt variant (up to +0.075), whereas the typed state adds a small difference (-0.003 to +0.013) that is significant in one of six comparisons. Beyond accuracy, the typed pipeline contains out-of-range model outputs at the agent boundary, while the unvalidated variant fails later and drops the requirement, and it shows a small, reproducible increase in inter-model MoSCoW consistency whose cause our design cannot isolate. The study is associative: the contrasts bundle prompts, call count, and staged context, and the results come from three 7–8B models on one dataset. The implementation and replication package are archived on Zenodo (this https URL).
[AI-35] RIPPLE: Recovery Internalization through Process-Level Policy Learning and Effect-Aware Credit for Long-Horizon Agent ic Coding
链接: https://arxiv.org/abs/2610.31276
作者: Jingxuan Xu,Zihang Xu,Yuechen Luo,Ziqi Miao,Wenqiang Zhu,Weihao Li,Junhao Fan,Xinshun Feng,Zhiyi Lai,Yanan Wu,Haoyang Huang,Ken Deng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Long-horizon coding agents can observe a failure and generate a plausible critique yet continue along the same incorrect trajectory. We call this the Critique-to-Action Handoff Gap. We introduce RIPPLE, a recovery-oriented training framework that internalizes process-level guidance into a single self-recovering policy. A prefix-only teacher provides evidence-grounded recovery supervision without access to future actions or terminal outcomes. After lightweight distillation, reinforcement learning assigns stage-specific credit to reflection triggering, diagnosis, and corrective action. Local validity signals gate trajectory advantages so that intermediate recovery decisions do not indiscriminately inherit terminal outcomes. Across four repository-level software-engineering benchmarks, RIPPLE improves the base policy by an arithmetic mean of 4.8 score points, with no auxiliary critic at deployment. Controlled ablations isolate the benefits of local, stage-aligned credit, and matched-prefix interventions show that critique semantics affect subsequent correction. Recovery gains concentrate where the prefix contains sufficient diagnostic evidence and the repository state remains reversible within the remaining budget.
[AI-36] Notation3: Graph Terms for RDF
链接: https://arxiv.org/abs/2610.31266
作者: Dörthe Arndt,Pierre-Antoine Champin
类目: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI)
备注: Submitted to TGDK
Abstract:Notation3 Logic was introduced as an extension of RDF allowing the user to express rules on top of triples and to make statements about graphs. The original proposal of the logic only came with an informal specification of its semantics which has been refined later on. While this refinement clarified the meaning of rules and implicitly quantified variables, it left the meaning of graphs used as terms - so-called graph terms - open. In this paper we aim to close this gap. We discuss in detail, how graph terms are used in Notation3, which formal requirements they need to meet, and provide examples to illustrate the difficulties when defining their semantics. We then introduce an abstract syntax for the logic and define its semantics in two steps: similarly to RDF’s simple interpretations, we define interpretations and basic entailment for Notation3 which covers plain RDF triples but also triples containing graph terms. N3-entailment is then an extension of basic entailment which assigns meaning to the different special predicates of the logic like for example the implication symbol. N3-entailment is defined in a way that it can be extended whenever a new predicate is added to the logic. As Notation3 Logic is currently under development, we expect many predicates to be added. Our research provides and important step towards the full standardisation of the logic.
[AI-37] Reducing Multi-Horizon Photovoltaic Forecast Error through Tree-Anchored Expert Combination
链接: https://arxiv.org/abs/2610.31248
作者: Sun Ze,Zhou Liguo,Xu Yuqing,Yu Lei,Jiang Mingming,Alois Knoll
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:A strong direct tree model may remain the best standalone photovoltaic forecaster, yet a weighted combination can reduce its residual errors. We examine whether complementary temporal predictors can lower the tree’s mean absolute error (MAE) over 15–240 minutes without weaker ensemble members degrading performance. The candidate bank contains LightGBM and seven neural implementations with different temporal representations. A combiner fitted to MAE has nonnegative weights that sum to one. Unhelpful candidates can receive zero weight, while the tree anchors the combination and provides a development-stage fallback. The main design shares weights across horizons; horizon-specific weights and alternative losses test whether further complexity is needed. On 56,078 windows from two co-located public PVDAQ systems, the historical shared-MAE policy reduces scaled MAE by 3.21% relative to the tree. The latest bounded neural correction does not reduce MAE further than this static combination. The results support selective combination of complementary predictions in this bank rather than equal averaging or indiscriminate addition of modules. Both analyses reuse a previously inspected evaluation period. They do not establish independent replication, weather-specific causation or universal gains from eight experts. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2610.31248 [cs.AI] (or arXiv:2610.31248v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2610.31248 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-38] Reward-Weighted Preference Optimization for Selective Combinatorial Optimization Problems
链接: https://arxiv.org/abs/2610.31239
作者: Haohui Zhang,Xinyu Hu,Wouter van Heeswijk,Neil Yorke-Smith,Martijn Mes
类目: Artificial Intelligence (cs.AI)
备注: 37 pages, 12 figures
Abstract:Reinforcement learning has emerged as an effective approach for combinatorial optimization, but most existing methods focus on problems such as the Travelling Salesperson Problem (TSP) in which all elements must be served. Selective combinatorial optimization problems (SCOPs), including the Orienteering Problem (OP), Prize-Collecting TSP (PCTSP), Stochastic PCTSP and Knapsack Problem (KP), additionally require selecting which elements to include. This selection creates sampled solutions with substantially different objective values, yet standard preference optimization reduces such comparisons to binary rankings and discards the magnitude of these differences. We introduce Reward-Weighted Preference Optimization (RWPO), a pairwise training objective that incorporates reward-gap information into preference learning. For candidate trajectories sampled from the same instance, RWPO weights each preference loss according to the normalized difference in their rewards, assigning greater training emphasis to comparisons that represent larger improvements in solution quality. We further combine RWPO with scale-specific fine-tuning and reward- and likelihood-weighted tilted sampling to adapt trained policies to larger and shifted instance distributions without relying on local-search post-processing. Experiments on the OP, PCTSP, Stochastic PCTSP, and KP show that RWPO consistently improves over unweighted preference optimization and policy-gradient baselines across a wide range of instance sizes. On the OP, the resulting solver exceeds the mean reward of the Compass metaheuristic at scales from 100 to 2,000 nodes. The results demonstrate that preserving reward-gap information can substantially improve preference-based learning for selective combinatorial optimization and support generalization to large-scale instances. Comments: 37 pages, 12 figures Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2610.31239 [cs.AI] (or arXiv:2610.31239v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2610.31239 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-39] rion-1: Unifying Language Protein Sequence and Structure with Diffusion Language Models
链接: https://arxiv.org/abs/2610.31215
作者: Tianyu Lu,Kaicheng Yang,Jiangtao Feng,Keyue Qiu,Yawen Ouyang,Yixin Wu,Xinbo Zhang,Lihao Wang,Wei-Ying Ma,Hao Zhou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We envision a general-purpose protein foundation model as a unified, interactive model of sequence, structure, and function, capable of understanding, generation, and cross-modal translation while retaining the instruction-following and reasoning capabilities of modern LLMs. Toward this vision, we introduce Trion-1, the first native, unified, multimodal diffusion language model (dLLM) for proteins. Trion-1 is built on two key design choices: (1) a unified discrete token space that jointly represents sequence, structure, and natural language; (2) a dLLM backbone that replaces strict left-to-right factorization with bidirectional or semi-bidirectional token interactions, enabling global residue communication and long-range contact modeling that are naturally suited to structure modeling. This formulation unifies six core capabilities: folding, inverse folding, sequence understanding, structure understanding, sequence design, and structure design, while retaining the native language capabilities of general-purpose LLMs. On established benchmarks, Trion-1 achieves broadly competitive performance. In folding, Trion-1 achieves a TM-score above 0.8 with a single sample under strict homology control, matching or exceeding the existing tri-modal model using 200 samples. For sequence understanding, Trion-1 performs on par with top-tier LLMs such as Claude Opus 4.7 (62.37%). These results support well-aligned sequence-structure-language representations and strong cross-modal generative capabilities.
[AI-40] Different Tasks Different Timesteps: Temporal Task-Aware Quantization for Diffusion Language Models
链接: https://arxiv.org/abs/2610.31198
作者: Ofek Avivi,Eliron Rahimi,Avi Mendelson,Chaim Baskin
类目: Artificial Intelligence (cs.AI)
备注: preprint
Abstract:Diffusion Language Models (DLMs) introduce unique quantization challenges due to their iterative denoising process. In this work, we reveal that quantization sensitivity is both taskand time-dependent, with different downstream tasks exhibiting distinct precision requirements across denoising steps. We further show that these patterns can be predicted without groundtruth labels or task-specific evaluation metrics. Building on these findings, we introduce T 2Quant, a task and time-aware mixedprecision quantization framework that uses a label-free sensitivity signal to optimally allocate activation precision across denoising steps under a fixed bit-width budget. Extensive experiments across three DLMs and six benchmarks demonstrate that T 2Quant consistently improves the average accuracy of existing quantization methods, achieving gains of up to 6.7 percentage points under the same budget.
[AI-41] PriMa-Causa: Causal Prescriptive Maintenance
链接: https://arxiv.org/abs/2610.31147
作者: Felix Saretzky
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注: PhD thesis, University of Luxembourg, 2026. 203 pages. Also available at this https URL
Abstract:Automated manufacturing requires data-driven decisions under uncertainty and sparse observational records, and unplanned downtime is a dominant cost of getting them wrong. Prescriptive maintenance recommends a maintenance action against an impending failure rather than merely predicting it. Established reference architectures couple predictive models with rule- or optimisation-based recommendation logic. Selecting a corrective action, however, requires an interventional quantity: how the line behaviour would change if the action were set. Manufacturing data is observational, so a model fitted to predict the outcome estimates an associational quantity and may misdirect the choice of action without the prediction error revealing it. Causal estimators target the interventional quantity but require an expert-led causal study per case, for which the bounded decision window of operation leaves no room. This dissertation analyses how causal reasoning can be embedded into prescriptive maintenance so that actions are ranked by their estimated interventional effect. The result is PriMa-Causa, built from three components: a causal reasoning layer that integrates the estimate into the PriMa reference model, a parametric simulator that supplies the interventional ground truth absent from manufacturing data, and a graph-conditioned tabular causal foundation model that, pre-trained once on this synthetic production-line prior, answers interventional queries in a single forward pass. PriMa-Causa attains a lower prescription error than associational baselines, with the largest margin under weak overlap and sparse context, and a partially disclosed causal graph reduces this error further. The causal inductive bias of the synthetic prior and the graph conditioning can thus replace per-case estimator fitting, a step toward automating causal inference for autonomous and self-healing maintenance.
[AI-42] Rethinking Fisher Vectors for Second-order Multi-scale Deep Feature Aggregation in Scalable Scene Retrieval
链接: https://arxiv.org/abs/2610.31146
作者: Alexy Bhowmick,Sarat Saharia,Shyamanta M. Hazarika
类目: Artificial Intelligence (cs.AI)
备注: 8 pages, 4 figures
Abstract:Aggregating CNN responses into a compact discriminative vector, given their lack of geometric invariance, is a challenge. To address this challenge, we propose a codebook-free method to encode second-order information of multi-scale CNN features within the Fisher Kernel framework. We use a scalable retrieval pipeline to show the efficacy of the proposed vector. The increased discriminative power is confirmed by experimental evaluation of our method on two difficult benchmarks (INRIA Holidays, UKBench). The significant improvement in retrieval accuracy over state-of-the-art approaches comes with only a slight dimension increase.
[AI-43] Post-Audit Mirag e: Identifiability Limits of Offline Agent Verification NEURIPS2026
链接: https://arxiv.org/abs/2610.31130
作者: Aman Behera,Payal Payal
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at the Workshop on Who Verifies the Agents? Toward Reliable Agent Development, NeurIPS 2026. 17 pages. Code: this https URL
Abstract:Offline tests can fail to reveal whether an agent update will remain safe after deployment. We operationalize this post-audit mirage using matched safe and harmful lifecycles that expose identical verifier-visible evidence while an isolated scorer records opposite deployment truth. Our theorem applies standard observational-equivalence and indistinguishability ideas to this executable audit boundary and states the resulting error-abstention tradeoff. The Post-Audit Mirage benchmark makes that boundary executable across authorization, scheduling, and shared-capacity environments. IdentifiedRangeMonitor uses fresh group-labelled harm outcomes to return deploy, hold, cannot determine, or unsupported. Experiments show that monitoring can recover the missing distinction in favorable settings but may remain undecided for rare groups, small effects, or missing observations. Verifier-agent evaluation must therefore test both calculation validity and whether the available evidence supports the deployment decision.
[AI-44] SetCommit: Coordinated Token Commitments for Few-Step Diffusion Language Model Decoding
链接: https://arxiv.org/abs/2610.31120
作者: Yuannuo Feng,Yubing Ye,Wenshuai Yao,Yizhe Chen,Ngai Wong,Wenyong Zhou,Wang Kang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion language models generate multiple token positions in parallel, but few-step decoding makes parallel commitment difficult: tokens that appear reliable individually may be jointly inconsistent, while premature commitments can prevent later context-dependent refinement. In this paper, we introduce SetCommit, a training-free method for coordinated parallel token commitment. Instead of selecting tokens independently by confidence, SetCommit selects a commitment set that balances token reliability against local joint commitment risk, preserving conditioning opportunities for subsequent denoising steps. It further performs budget-neutral local exchanges, replacing vulnerable earlier commitments when newly revealed context supports more reliable alternatives. SetCommit requires no retraining or architectural modification and is compatible with existing diffusion decoding strategies. Experiments with LLaDA-8B and Dream-7B on GSM8K, MATH-500, MBPP, and HumanEval show accuracy/pass@1 gains of up to 7.4 percentage points over native decoding at 32 model evaluations. At common quality targets, geometric-mean incremental speedups range from 1.44 to 1.74\times . Code is available at this https URL.
[AI-45] Constraint Reasoning and Large Language Models Join Forces: Providing Contrastive Explanations for Real-World Optimisation-Based Applications
链接: https://arxiv.org/abs/2610.31119
作者: Roger X. Lera-Leri,Filippo Bistaffa,Athina Georgara,Juan A. Rodríguez-Aguilar
类目: Artificial Intelligence (cs.AI)
备注: Currently under review at Applied Artificial Intelligence
Abstract:We focus on optimisation-based Decision Support Systems in real-world applications. In this context, we aim to provide contrastive explanations that answer users’ questions about recommendations based on the solution of an optimisation problem, specifically a Mixed-Integer Program (MIP). Addressing this endeavour is crucial to ensure human oversight of the decisions implemented, as required by international guidelines on trustworthy and responsible automated decision-making systems. We propose a novel methodology that combines constraint reasoning techniques and Large Language Models (LLMs) to generate contrastive explanations for optimisation problems encoded as MIPs. Although LLMs have shown impressive capabilities in Natural Language processing, they can produce incorrect answers due to reasoning limitations. To overcome these limitations, we leverage constraint reasoning techniques to distil the information necessary to produce an explanation, which is then passed as input to an LLM as the so-called graph of reasons. Such a graph captures the semantics and the relationships among constraints of the optimisation problem. Results on real-world scheduling domains show that incorporating constraint reasoning improves explanations generated by the LLM. In particular, using the graph of reasons enables the LLM to produce natural-language explanations that include more relevant reasons while reducing errors compared to explanations generated without this structure.
[AI-46] ProductWebBench: Evaluating Website Continuity Beyond Task Completion
链接: https://arxiv.org/abs/2610.31118
作者: Jiajun Wu,Jian Yang,Yaxin Du,Shukai Liu,Linzheng Chai,Tuney Zheng,Wei Zhang,Haitao Yang,Yan Bai,Yihang Lou,Xianglong Liu,Bryan Dai
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 35 pages, including appendix
Abstract:Frontend coding agents are usually graded on one question: did the requested target get built. Real engineering changes carry a second obligation, because the rest of the site has to keep working. We call that second obligation website continuity, and introduce ProductWebBench, a benchmark that scores both on living frontend repositories: 320 Change tasks over existing code and 80 long-horizon Build tasks. Every task freezes a repository snapshot, staged requirements, browser evidence and deterministic checks before evaluation, and enters only once a reference passes every check and plausible-wrong ones fail. A simulated user reveals the requirements one stage at a time, and every check cleared earlier stays armed for the rest of the run. Across 8,737 runs from 13 models the two come apart: of the Change runs passing every requested-work check, 28.2% still fail a continuity check – a blank route, horizontal overflow, console errors – and 10.8% push a browser state they had already passed back to failing. Because the same checks also run on the untouched repository, each regression carries exact provenance, not a post-hoc label: on the 586 runs where both readings exist, 18.3% are caught breaking a verified state. Interaction is a further blind spot, failing in 29.5% of runs whose static DOM already passes. ProductWebBench therefore scores a patch against the living system it lands in, not the target alone.
[AI-47] Can AI Co-Design Structures? An AI Agent for Early-Stage Structural Design
链接: https://arxiv.org/abs/2610.31113
作者: Ricardo Maia Avelino,Mennatallah El-Assady,Tom Van Mele,Philippe Block
类目: Artificial Intelligence (cs.AI)
备注: Paper accepted to the International Association for Shell and Spatial Structures (IASS) Annual Symposium 2026, in Torino, Italy
Abstract:Recent advances in large language models and agentic workflows offer a new paradigm to early-stage structural design translating design intent into structural models. Although text-to-3D approaches have demonstrated that natural language can be used to produce visual, mesh-based geometric models, structural design requires domain-specific geometric arrangements and semantically rich data types. This paper introduces a proof-of-concept workflow for prompt-to-structure generation using an AI agent connected to a COMPAS-based structural modelling API through a Model Context Protocol (MCP) server. Through controlled modelling operations, the agent creates and modifies a graph-based structural data model composed of explicit building elements. The resulting artifacts can be inspected, modified, serialised, and further connected to downstream AEC workflows. The approach is demonstrated through selected prompt-to-structure examples and by progressive and editable design sequences that resemble typical iterative structural design workflows. The results are preliminary, but demonstrate the possibility of combining natural language, agentic workflows, and domain-specific structural data structures as a first step toward AI-assisted structural co-design.
[AI-48] Who Is the Influencer? Consensus Dynamics in Multi-Agent Systems
链接: https://arxiv.org/abs/2610.31105
作者: Gouki Minegishi,Jingyuan Feng,Hiroki Furuta,Takeshi Kojima,Yusuke Iwasawa,Yutaka Matsuo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-agent LLM systems are increasingly used in practice, but their behavior is complex and hard to understand. We design a simple task to ask which model has the most influence on the final consensus when several models are mixed in one system. Experimentally, across seven model families, we find that the opinions of the other models are absorbed into the opinion of the stronger model, the one with more parameters. This happens even though the task requires no pretraining knowledge and has no correct answer. To understand this phenomenon, we interpret the agents as weighted particles moving on a simplex. Larger models carry a larger weight and therefore occupy a deeper well in the energy landscape. A large weight means strong in-context learning: a model with a large weight amplifies the majority of its context. Larger models are generally better at in-context learning, and this is what makes the larger model the influencer. We hope that this work helps to interpret multi-agent systems, whose behavior is becoming too complex for humans to understand.
[AI-49] DeepOpen: Decision-Oriented Intent Classification on BANKING77 and CLINC150
链接: https://arxiv.org/abs/2610.31104
作者: Shuang Song Zhongqiao Yuan
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We present DeepOpen, a decision-Oriented intent-classification framework with dataset- specific systems for BANKING77 and CLINC150. On BANKING77, a single English classi- fier continued with sharpness-aware minimization (SAM) achieves 94.1234% test accuracy (2,899/3,080). This is the best observed result among three fixed SAM seeds, selected for display after test evaluation; their mean is 94.0368%. On the 150 in-scope CLINC150 classes, probability fusion of one Laya specialist and three DeBERTa-v3-large classifiers achieves 98.0222% (4,411/4,500). We describe the two configurations, reproduce the reference com- parisons displayed in the accompanying evaluation explorer, and separate headline scores from evidence about improvement. Under a uniform BANKING77 evaluation path, the SAM mean exceeds the historical checkpoint by 0.0433 percentage points, with a shared-row 95% interval of [-0.2273, +0.3247]. Thus, the continuation does not establish a reliable accuracy gain. The report documents strong observed task-specific performance, selection effects, and unresolved historical execution-path differences, rather than a new backbone or a unified algorithmic advantage.
[AI-50] Solving Flood-It Puzzles: A Comparison of Tailored A* Search with Declarative Solver Approaches
链接: https://arxiv.org/abs/2610.31100
作者: Philipp von Manteuffel,Benjamin Blankertz
类目: Artificial Intelligence (cs.AI)
备注: 22 pages main text, 35 pages in total; 13 figures plus 12 figures in the appendix
Abstract:We present an optimized approach for finding minimal move sequences to solve Flood-It puzzles using A* search and compare it with approaches using SAT, Integer Programming, and Constraint Programming solvers. The performance of A* search is considerably improved by using an appropriate heuristic and sophisticated equivalences of game states to limit the search space. Each aspect is systematically evaluated, providing insights that generalize to other problems. By comparing the best variant of each approach, the A* search is the fastest in finding a minimal solution in about 75 % of the Flood-It instances tested. This result is in line with our expectation since A* specifically looks for a minimum solution. In contrast, the SAT-based approach is designed to solve the decision problem of whether a solution with a given length exists. An iterative search is required to determine a minimum solution length using a SAT-based approach. However, with increasing complexity, quantified by the minimum solution length, the SAT-based approach outperforms A*. We also compared the different approaches to the decision problem, where A* again performed best for the largest number of puzzles. Here, again, SAT supersedes A* for higher complexities, albeit with a lower crossing point.
[AI-51] Co-Recursive Semantic Self-Extension: A Mathematical Foundation for Open-Ended Intelligence
链接: https://arxiv.org/abs/2610.31097
作者: Simon Hegelich,Bardh Prenkaj,Habiba Sarhan,Morteza Shahrezaye,Kolja Hegelich,Eric Cambria,Christian Montag
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Artificial intelligence has advanced through scale, data, and computation, yet remains bounded by its representational framework. We ask what makes intelligence open-ended. Every inferential system selects meaningful distinctions, limiting what it can learn. Within a fixed purpose, performance reaches a ceiling set by its semantic domain. Teleological succession, in which purposes arise from earlier ones, faces a similar limit: Boolean-closed purpose-transformations exhaust their reachable semantic territory. We shift from purposes to inferential systems. Drawing on Kant and Gödel, we model cognitive states as partial evaluations over a formal language. Unresolved propositions may be undecided or undecidable. Reasoning can resolve the former but not the latter, leaving a frontier for extension. Non-trivial reductions of undecidable propositions expose undecidable intermediate premises whose reflective reification supports multiple decisive extensions. Finite trace-guided generation remains fallible, requiring meta-level selection and polycontextural exploration. Using Kolmogorov complexity, we show that no architecture with a fixed state description, transition mechanism, and initial condition can generate histories incorporating irreducible algorithmic information. Because physical systems are fixed in this sense, we examine collectives. Coupling finitely many finite agents does not remove the bound. Extension requires persistent external memory preserving semantic content and historical coordinates. Collective generative capacity depends on balancing overlap for transfer with difference for generation. We propose semantic self-extension as a design principle: intelligence should not merely resolve ambiguity within existing meanings, but transform unresolved structures into new spaces of meaning, generating new distinctions, questions, representations, and purposes.
[AI-52] MQAScore: Fine-Grained Text-Music Alignment Evaluation via Large Audio Language Model ICASSP2027
链接: https://arxiv.org/abs/2610.31083
作者: Wei-Jaw Lee,Chun-Ping Wang,Xuanjun Chen,Yi-Hsuan Yang
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
备注: Submitted to ICASSP 2027
Abstract:Text-to-music alignment is commonly evaluated by the global similarity between text and audio embeddings, which collapses all described attributes into a single score and obscures which ones are missing. We instead introduce MQAScore, which verifies each atomic musical concept along five clip-level dimensions (genre, instrument, mood/theme, vocal, and usage) by the probability that a large audio language model (LALM) answers affirmatively to a binary question. To assess evaluator reliability, we construct verification labels from presence/absence annotations and LLM-guided concept substitutions, and propose the Sensitivity-Invariance Ratio, which measures how selectively an evaluator responds to meaning-changing substitutions. Experiments show that concept-level scoring improves both question-answering and similarity-based evaluators, and MQAScore can be a more reliable and interpretable fine-grained complement to global similarity for evaluating text-to-music alignment.
[AI-53] Dual Exam: Controlled Evaluation of Multimodal Agents and World Models
链接: https://arxiv.org/abs/2610.31066
作者: Zhefan Rao,Meng Chu,Xuan Zhang,Jingfan Chen,Guohao Cai,Runyi Li,Shiyi Du,Ning Lu,Xu Huang,Zhixuan Chen,Feng Nie,Harold Haodong Chen,Hanwen Wang,Wen-Jie Shu,Yajing Bai,Xianfeng Wu,Xuanhua He,Shuhan Guo,Ziyue Ma,Ruichuan An,Long Bai,Miaosen Chai,Wenhao Chai,Jingye Chen,Qi Chen,Xuhang Chen,Wei Chow,Minghao Fu,Yunhao Gou,Haokun Gui,Dong Huang,Wei Huang,Zi Cheng Huang,Haibo Jin,Lingdong Kong,Leung Chun Ho,Tianqin Li,Yanheng Li,Jiayuan Liu,Jiyang Liu,Mingyu Liu,Shaoteng Liu,Xiaocheng Lu,Yawen Luo,Zeyu Qin,Tiancheng Shen,Yueqi Song,Zhaolong Su,Hengkai Tan,Tianhao Tang,Yiran Tao,Juncheng Wang,Qinghe Wang,Shangshang Wang,Yingying Wang,Fang Wu,Tiange Xiang,Qiushi Yang,Weicai Ye,Weichen Yu,Jiarui Zhang,Jize Zhang,Liuzhou Zhang,Renrui Zhang,Sheng Zhang,Wenhu Zhang,Junru Zhao,Lingting Zhu,Yuxin Zuo,Yutong Bai,Long Chen,Yang Deng,Ziqi Huang,Yeying Jin,Weijia Li,Kevin Qinghong Lin,Quande Liu,See-Kiong Ng,Hongliang Ren,Hongru Wang,Shujun Wang,Qifeng Chen,Haoxuan Che
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:World Models (WMs) can support Multimodal Agents (MAs) through simulated experience, while Agent trajectories can inform World Model refinement. Evaluating this interaction requires separate evidence about action selection and state evolution. We introduce Dual Exam, which tests Agents under fixed environment response mechanisms and World Models under fixed action interventions across physical, digital, social, and scientific domains. The Agent track compares 10 configurations on 1,088 items in 15 reporting categories. The World Model track contains 400 cases across eight scenarios; 15 configurations have full coverage and one covers six scenarios. GPT 6 Astra leads the case-weighted Agent Index at 31.3, and GPT-Image-2 leads the full-coverage WM Overall at 21.7. These scores use different tasks and metrics and are not directly comparable. A 150-case pilot of one coupled pairing records zero task successes and motivates component-level diagnosis. Results reveal uneven task performance and failures to preserve task-relevant state. Their value for predicting coupled success or sustained reciprocal learning remains to be established.
[AI-54] Diagnosing the Explicit-to-Applied Theory of Mind Gap in LLM s
链接: https://arxiv.org/abs/2610.31060
作者: Prajvi Saxena,Sabine Janzen,Wolfgang Maass
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As large language models are deployed in high-stakes roles such as therapy, education, and personal assistance, their social cognition must be assessed reliably. Current Theory of Mind (ToM) benchmarks report aggregate scores that conflate two distinct competencies: explicit ToM (recognizing what someone believes) and applied ToM (using that belief to predict behavior). Building on this distinction, we investigate the gap and its recoverability using four structurally diverse benchmarks. Aggregate reporting obscures uneven competence: on SimpleToM, Llama-3.1-8B scores 51.5 percent overall while recognizing mental states at 99.4 percent and predicting behavior at only 28.9 percent. To assess recovery, we compare structured prompting, matched in-context learning, and our proposed Belief-State Test-Time Training (BSTTT), which adapts lightweight parameters on answered support questions through action reconstruction. Across Qwen2.5-7B, Llama-3.1-8B, and GPT-oss-20B, recovery depends on the model, task, and intervention. On SimpleToM, reminders combined with BSTTT improve Llama’s behavior prediction by 40.8 percentage points; reminders account for most of the overall improvement, with BSTTT adding 1.3 to 5.5 points beyond reminders across models. Contextual examples often outperform weight updates, while BSTTT remains stronger in specific model-benchmark settings. Temporal belief tracking remains difficult despite partial recovery. These findings show that aggregate ToM scores can overstate applied competence while hiding performance recoverable through targeted assistance. We recommend decomposed reporting of explicit and applied accuracy alongside intervention comparisons that establish which failures are recoverable and under what conditions.
[AI-55] Directing to Diversify: Multi-Directional Search for Automatic Heuristic Design
链接: https://arxiv.org/abs/2610.31048
作者: Jianghan Zhu,Jianan Zhou,Fei Liu,Ni Zhang,Yaoxin Wu,Zhuoyi Lin,Zhiguang Cao
类目: Artificial Intelligence (cs.AI)
备注: 46 pages, 9 figures, 19 tables. The main paper, including references, is 13 pages; appendices are included
Abstract:Large language models (LLMs) have demonstrated strong potential to automate heuristic design for NP-hard combinatorial optimization (CO) tasks without human intervention. Recent methods predominantly rely on evolutionary computation (EC) frameworks, where heuristic diversification is mainly induced through EC-like operators or downstream selection, while the direction of heuristic transformation remains largely unharnessed. In this paper, we propose a multi-directional Perturbation-Rewrite Search (PRS) framework for automatic heuristic design that factorizes heuristic transformation through directional search primitives at the generation level. Specifically, PRS operates on heuristic ideas through three complementary directions: expansion toward richer and more expressive designs, compression toward simpler and less specialized designs, and steering for fine-grained refinement, together enabling broader and more sustained exploration of diverse regions in the heuristic design space. PRS further organizes candidate exploration within a breadth-first search-based process with adaptive operator control, which ensures thorough candidate utilization and sustained exploration, and maintains strong search dynamics while preventing premature convergence. Extensive empirical results across diverse CO tasks show that PRS consistently discovers high-quality heuristics, demonstrating the effectiveness and generality of explicitly directional exploration design.
[AI-56] Conflict-Aware Logit Adapters for Utility-Preserving Anti-Distillation NEURIPS2026
链接: https://arxiv.org/abs/2610.31046
作者: Quoc Phong Dao,Binh Anh Nguyen,Linh Ngo Van,Thanh-Toan Do,Mehrtash Harandi,Trung Le
类目: Artificial Intelligence (cs.AI)
备注: NeurIPS 2026
Abstract:While reasoning traces improve model capability, they also provide supervision signals for downstream distillation. Initial efforts to mitigate this issue rely on decoding-time interventions, but they incur high inference overhead and can harm the teacher model’s natural fluency. We propose Conflict-Aware Logit Adapters (CALA), a lightweight training-time defense. CALA attaches a small residual adapter to a frozen teacher model and trains only the adapter. A proxy student identifies positions where imitation is most sensitive, and the adapter perturbs the output distribution at these positions while remaining within high-probability regions. A conflict-aware optimization strategy balances anti-distillation and utility objectives, while regularization preserves the teacher’s performance and fluency. At inference time, CALA incurs modest overhead and requires no auxiliary model. Experiments on reasoning benchmarks show that CALA substantially reduces successful student imitation while maintaining near-original teacher performance and generation quality. These results establish CALA as a practical and efficient alternative to decoding-time approaches.
[AI-57] Algorithmic drift in generative adaptive learning
链接: https://arxiv.org/abs/2610.31041
作者: Firuz Kamalov,Hana Sulieman,Ahmed El Sayed,Sivaraj R.,Nitesh Sughnani
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Generative artificial intelligence promises to democratize personalized tutoring, yet the longitudinal stability of adaptive systems remains under-theorized. Current ethical discourse focuses on static bias. This paper argues that a distinct threat lies in algorithmic drift: a feedback loop in which minor, demographically asymmetric observation errors compound over time into systemic inequities. We make three contributions. First, we formalize algorithmic drift as a coupled learner–tutor dynamical system. Second, we corroborate this theory with an agent-based Monte Carlo simulation in which either a transparent online tracker or an LSTM deep-knowledge-tracing network drives an IRT-based curriculum policy. Third, we introduce the temporal fairness stabilizer (TFS), a targeted intervention that periodically re-probes learner capability, and benchmark it against standard exploration baselines with explicit group-fairness metrics and replication-level inference. Within the modeled environment, TFS closes the achievement gap where generic exploration does not, and we characterize the conditions under which it succeeds. We separate what is established analytically, what is demonstrated in simulation, and what remains to be validated with real learners. Our results reframe algorithmic fairness in education as a temporal property of a closed loop rather than a deployment-time checkbox.
[AI-58] Multichoice Modeling: Exploring Decision Configuration Space under Uncertainty
链接: https://arxiv.org/abs/2610.31038
作者: Mariia Kozlova,Roman Stepanova,Alexander Walzer,Andrea Saltelli
类目: Artificial Intelligence (cs.AI); Risk Management (q-fin.RM); Computation (stat.CO)
备注:
Abstract:Decision support has become sophisticated at exploring uncertainty, but much less so at representing the architecture of the decision itself. As a result, the decision space is often narrowed before uncertainty is analyzed, leaving interactions among choices and the effects of alternatives they imply largely invisible. This paper introduces multichoice modeling, a framework for making large decision configuration spaces explicit and computationally explorable under uncertainty. Drawing on the logic of rugged landscapes, we conceptualize such problems as contingent performance landscapes, where interdependent choices generate configuration-specific outcomes whose value changes across uncertain futures. The computational framework combines multidimensional decision and uncertainty spaces in a single simulation. Further, variance-based sensitivity analysis is used for identifying influential choices, and conditional distribution analysis for comparing configuration-specific performance profiles. We demonstrate the framework through three case studies: a manufacturing process upgrade, MA strategy, and mortgage design. By contrasting conventional uncertainty-only analysis with multichoice modeling, the cases show how opening the decision configuration space can reveal unexpected complementarities, near-best configurations with different risk profiles, and strategies that were not apparent as alternatives at the outset.
[AI-59] CANNBot-CAKE: Collaborative Agent ic Kernel Evolution for Ascend NPU
链接: https://arxiv.org/abs/2610.31034
作者: Deng Pan,Xue-Jian Gao,Junyi Fan,Bin Du,Jiasheng Li,Xinxian Chen,Nok Yan Chang,Xinlin Li,Lingchao Zheng,Jun Li,Yuwei Fan
类目: Artificial Intelligence (cs.AI)
备注: 23 pages, 10 figures
Abstract:Efficient custom operators are essential for modern deep learning workloads, yet developing high-performance kernels for Ascend NPUs remains slow, expert-intensive, and difficult to automate. Although AI agents now approach expert performance on GPU kernel benchmarks, current models remain less capable of generating AscendC kernels for NPUs, potentially reflecting the relative scarcity of relevant training data. We thus present CAKE, a CANN-native system for deployable custom operators that combines adaptive multi-DSL routing, separate Generation and Evolution modes, and reviewed knowledge reuse through CAKE Book and a Self-Evolving Agent. On 53 CANN-Bench v1.1 operators on Ascend 950PR, CAKE’s aggregated results achieve 100% case pass rate on the standard set (1,060/1,060) and 98.28% on the hidden set (4,167/4,240), with an overall mean score of 78.63 out of 100 and an average performance speedup of 1.334\times over the baseline. Evolution then pushes correct operators further: EngramGateFusion cuts aggregate device time by 46.6% over 20 rounds and roughly halves the mean roofline gap. These results offer a practical path toward closing Ascend’s AI-for-kernel capability gap.
[AI-60] Multi-Tool Selection for Robotic Systems: Comparing Semantic Retrieval and LLM
链接: https://arxiv.org/abs/2610.31013
作者: Miguel Sousa,Luis Ramos Pinto,João P. Matos-Carvalho
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Natural-language interfaces allow users to command robotic systems without interacting directly with mission planners or robotics middleware. Their reliable use depends on robot tool routing, defined here as the selection of the smallest complete subset of predefined executable capabilities required to satisfy a request. This task is particularly difficult for compound instructions, where incomplete selection may omit necessary sensing or monitoring capabilities, while excessive selection may expose irrelevant tools. We compare four routing strategies: three progressively refined ChromaDB-based methods - fixed-threshold retrieval, relative-distance retrieval, and clause-aware aggregated retrieval - and a cloud-hosted 120-billion-parameter language-model baseline. The methods are evaluated on 164 manually annotated prompts organised into three complexity levels, from direct single-tool requests to compound requests requiring at least three tools. Using exact-match accuracy, completeness, precision, a completeness-weighted composite score, and routing latency, CDB-v3 achieves the best macro-average score among the retrieval-only methods, with 0.7822, compared with 0.8869 for the large-model baseline. Although the LLM provides higher routing quality, CDB-v3 exhibits substantially lower latency and can be deployed locally without requiring cloud connectivity. These results identify a quality-latency trade-off between direct large-model selection and clause-aware semantic retrieval for privacy-sensitive, connectivity-constrained, and resource-constrained robotic systems.
[AI-61] Evaluating Artificial Wisdom (AW): A Trajectory-Based Framework for Context-Sensitive Decision Regulation
链接: https://arxiv.org/abs/2610.31012
作者: Md. Jalil Piran
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Capability benchmarks measure what large language models can do, but not whether their recommendations remain appropriate as decision conditions change. This paper proposes a trajectory-based measurement framework and prospective validation protocol for artificial wisdom, defined behaviorally as context-sensitive regulation of judgment. Versioned reference specifications determine admissible responses and transition materiality before evaluation. Seven stage-level criteria assess judgment under current conditions, while two transition components assess material-update adequacy and commitment-preserving stability. Stability requires prior adequacy; switching within registered commitment classes is permitted, and persistent error earns no stability credit. A nine-component profile is primary, with a secondary bottleneck score and separate critical-failure reporting. Three simplified simulations examine precision, bottleneck estimation, and sensitivity to model-family count. A single-criterion simulation met its precision criteria at 24 scenarios per domain. Near a bottleneck tie, the estimated minimum had a downward bias of 0.55 points at that size. A simplified incremental-prediction simulation detected a Brier improvement of 0.027 with probabilities of .508 and .742 using eight and twelve families, respectively; it does not establish power for the operational protocol. An exploratory predecessor-rubric study comprising 71 evaluation records illustrates aggregation sensitivity but does not validate the proposed instrument. The twelve-family target remains provisional pending feasibility assessment and calibration of the complete analysis. Whether the profile predicts independent decision outcomes beyond capability remains an open empirical question.
[AI-62] Mitigating Saliency Collapse: Robust Saliency-Aware Long-Text Image-Text Alignment
链接: https://arxiv.org/abs/2610.31005
作者: Qiuyu Kong,Zanxi Ruan,Marco Cristani,Yiming Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Long-text image-text alignment is essential for fine-grained scene understanding, yet existing methods mainly focus on extending context length or improving local alignment. We identify an overlooked failure mode, saliency collapse, where models over-rely on text that features visual prominence while underutilizing non-salient context. Consequently, retrieval performance degrades sharply when salient content is reduced, despite that the non-salient context remains informative. To address this, we propose RoSA, a saliency-aware finetuning framework that decomposes image-text pairs into complementary salient and contextual views, enabling joint learning of prominent and contextual semantics. We propose a modality-independent decomposition strategy to combine object-level visual saliency estimation with LLM-based text decomposition to bridge vision and language. We employ auxiliary objectives to explicitly align these views, ensuring robust representation with both salient and contextual features. RoSA greatly improves long-text retrieval and demonstrates strong robustness to saliency collapse across benchmarks. Moreover, our auxiliary objectives are plug-and-play, bringing consistent gains to existing tuning methods. Code and checkpoints are available at: this https URL
[AI-63] SecLog-Bench: A Scalable Benchmark for Cybersecurity Log Reasoning with Scaffolded Evidence
链接: https://arxiv.org/abs/2610.31000
作者: Yifan Wang,Dongliang Xu,Xiaoyong Li,Jing Li,Peter S. Pan,Yue Yao
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 17 pages, 9 figures. Submitted to IEEE Transactions on Dependable and Secure Computing. Code and data: this https URL
Abstract:Large language models (LLMs) have shown strong potential for cybersecurity, leading to growing interest in LLM-based automated cybersecurity log analysis. However, current datasets (1) are limited in scale; (2) rely mainly on single-source evidence; and (3) have only binary or multi-class attack labels, without dense annotations of the reasoning process. To address these limitations, SecLog-Bench has three key features. First, we build an extensible simulation platform that covers 29 attack scenarios, supports repeated execution and the integration of new scenarios, and enables scalable generation of cybersecurity logs with attack-stage labels. Second, we use multi-source logs, including audit logs, application logs, and network traffic, to jointly analyze cyber attacks. Third, we annotate each sample with the essential context for cybersecurity log reasoning and a structured step-by-step evidence chain. We evaluate eight mainstream LLMs on this benchmark for post-detection attack-progress reasoning. Results show that, when provided with scaffolded evidence, current LLMs exhibit modest capability in attack-progress reasoning, with the best model achieving only 44.44% accuracy and 45.07% Macro-F1. Their reasoning remains unreliable; i.e., the highest Stage Reasoning F1 (SRF1) score is only 54.87%. Moreover, even correct predictions may be accompanied by flawed reasoning, and reasoning quality is often misaligned with prediction accuracy.
[AI-64] ruthful by Omission: Click Thinning for LLM -Native Advertising
链接: https://arxiv.org/abs/2610.30999
作者: Junhua Liu,Zhicheng Liang,Zejun Zhang,Langzhou He,Feng Qian
类目: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Data Structures and Algorithms (cs.DS)
备注:
Abstract:Generative AI assistants are becoming an advertising medium, embedding sponsored content in their answers. This setting departs from classical position auctions in two ways: an ad’s click-through rate depends on the passage that hosts it, and because users stop at their first click, earlier ads take clicks from later ones. This competition for clicks makes truthfulness delicate: a higher bid can reshape the allocation and put a competitor ahead of the bidder, reducing the bidder’s clicks. Existing mechanisms either attain a constant fraction of the optimal welfare (advertisers’ total click value) but are not known to be truthful, or are truthful but their welfare guarantee degrades as the number of passages grows. We achieve both even when the answer and its editorial order are fixed before bidding. Assuming each ad’s click-through rate when reached is known and unaffected by omitting other ads, we give a computationally efficient mechanism that is truthful in expectation and attains at least half the optimal welfare under free reordering. Our key idea, click thinning, omits ads at random, showing earlier ads less often so that each ad’s overall click probability no longer depends on the order of the ads. Welfare maximization then reduces to a tractable matching problem, and solving it exactly with VCG-style payments makes truthful bidding optimal in expectation. Characterizing the click probabilities that omission can achieve sharpens the approximation factor from 2 to 2-(1-\delta)^K-1 when at most K ads are shown and each click-through rate is at most \delta . This factor is the best possible for any order-preserving rule, even one that ignores incentives and computation, so truthfulness and computational efficiency incur no additional worst-case loss. Subjects: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Data Structures and Algorithms (cs.DS) Cite as: arXiv:2610.30999 [cs.GT] (or arXiv:2610.30999v1 [cs.GT] for this version) https://doi.org/10.48550/arXiv.2610.30999 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-65] LAST-V2X: Language-Driven Adaptive Scenario Synthesis and Testing for Communication-Induced Failures in V2X Safety Applications
链接: https://arxiv.org/abs/2610.30979
作者: Keshu Wu,Hao Zhang,Rui Gan,Pei Li,Jiwan Jiang,Yang Zhou
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Vehicle-to-everything (V2X) communication enables connected vehicles to make safety-critical decisions, but delayed, missing, or incorrect information can compromise driving safety. Systematically testing such failures is challenging because traffic interactions, communication conditions, and application responses must be jointly configured to produce meaningful scenarios. This paper presents LAST-V2X, a language-driven, adaptive framework for automated scenario synthesis and safety testing of V2X applications. LAST-V2X leverages large language models (LLMs) to translate natural-language test descriptions into executable scenarios grounded in roadway geometry, V2X communication semantics, and vehicle behavior. The framework integrates physical traffic simulation with spatially aware communication modeling and message-level fault injection to evaluate how information degradation affects vehicle decisions. To establish whether an observed safety failure is caused by communication rather than the underlying traffic conditions, LAST-V2X employs matched executions under degraded and ground-truth information. Furthermore, an execution-guided refinement mechanism combines parameter search and LLM self-reflection to identify ineffective test configurations, improve scenario effectiveness, and produce minimal reproducible failure cases. Experiments across diverse CARLA environments and V2X safety applications demonstrate the framework’s ability to generate reliable scenarios, distinguish communication-induced failures from unrelated safety violations, and substantially improve failure discovery through adaptive refinement. These results establish LAST-V2X as a systematic approach to language-driven, causally verified, and adaptive safety testing for connected vehicle systems.
[AI-66] Zero-Shot Forecasting of In-Situ Soil Moisture and Temperature with a Time-Series Foundation Model NEURIPS2026
链接: https://arxiv.org/abs/2610.30971
作者: Mohammad Talaei,Mohammad Goudarzi,Maille Todd,Vanessa N.L. Wong,Iadine Chadès
类目: Artificial Intelligence (cs.AI)
备注: Accepted at the NeurIPS 2026 Workshop on Tackling Climate Change with Machine Learning
Abstract:Soil moisture and temperature govern plant-water availability and microbial activity, making them key indicators for evaluating land-restoration efforts on degraded landscapes. However, the physical and data-driven models normally used to forecast them both need information that a newly instrumented restoration site does not yet have: hydraulic parameters for the former and months of per-site training data for the latter. We evaluate TimesFM 1, a 500M-parameter time-series foundation model, zero-shot on 15-minute telemetry from 27 sensor stations across three land-management regimes (Control, Treatment, Scald) at a semi-arid restoration site, forecasting 5 cm and 30 cm soil moisture and enclosure temperature at seven lead times from 1 h to 168 h. Using gap-aware rolling-origin evaluation across sensors, we find that TimesFM beats the best rule-based statistical forecasting baselines (persistence, seasonal-naive, or 7-day climatology) by 21-76% at sub-daily lead times but loses by 6-79% beyond a day, a pattern explained by a diurnal-alignment mechanism that holds for both soil-moisture depths and temperature. A CatBoost baseline trained walk-forward on the same origins with full history and weather covariates never wins overall, and we show that its weakness is partly a cold-start effect specific to short deployments. We additionally quantify the inference cost of zero-shot forecasting on commodity CPU hardware. Our results highlight a practical trade-off between zero-shot forecasting capability, predictive performance, and computational efficiency.
[AI-67] Measuring the Reasoning -Action Coupling in Driving Vision-Language-Action Models through Counterfactuals
链接: https://arxiv.org/abs/2610.30953
作者: Gabriel Manalu,Saad Wazir,Thao Do,Seongah Kim,Hyeontaek Hwang,Tolcha Yalew Kidane,Daeyoung Kim
类目: Artificial Intelligence (cs.AI)
备注: 11 pages, 4 figures, 3 tables
Abstract:Driving vision-language-action (VLA) models bring an appealing form of explainability to a safety-critical domain, since a system that can state the reasons for its maneuvers offers the transparency that deployment and public trust in autonomous driving demand. These models read the world into language before they act on it, describing the scene and reasoning over it in text that ranges from chain-of-thought to multi-stage question answering on the way to a trajectory. That framing quietly assumes two separable properties, namely that the trajectory follows the generated reasoning and that the generated reasoning tracks the causes that actually produce the trajectory. To thoroughly investigate these properties, we apply causal, counterfactual interventions to four driving VLAs across different benchmarks. First, we perturb the models’ own reasoning and show that altering its meaning produces small trajectory changes relative to blanking the same models’ camera inputs, a pattern that persists across benchmarks and model architectures. We then construct scene-level counterfactuals through photorealistic removal of key objects named in the models’ reasoning and analyze whether the generated reasoning and trajectories track these changes. Overall, our findings indicate that the reasoning these models present to account for their resulting driving behavior is weakly coupled to that behavior itself, pointing to limits on the explainability these systems are intended to provide.
[AI-68] OpenAttention: One Percent Is Enough with Tail Proxies
链接: https://arxiv.org/abs/2610.30951
作者: Hyunwoo Yu,Jongbeom Lee,Hyunbin Kim,Sungjun Park,Jaemin Choi,Jincheol Yang,Seunghun Moon,Bongjoon Hyun,Kyomin Sohn,Kyeongbo Kong,Suk-Ju Kang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 27 pages. Project page: this https URL ; Code: this https URL
Abstract:Sparse attention reads only a small subset of the key-value cache, and how to treat the unread tail remains a key challenge. Discarding the tail biases the output, and sampling it as previous approaches do fails at reads of 5% and below, where the tail likely holds heavy hitters the selector missed. Our key observation is that the simplest possible proxy, one token that folds the entire tail into its count-aware mean, already yields lower output error than a sample at every read behind a practical selector. We further show in closed form that at low reads sampling nearly doubles the weight error of discarding while a single proxy stays below it. Building on these findings, we explore approximating the dense attention distribution with a few proxies that target the tail. We propose OpenAttention, a pipeline of clustering, ErrorGate and top-k selection followed by a count-aware mean proxy. To implement clustering and selection both efficiently and effectively, we design training-free Query-Geometry Clustering (QGC) and a Query-Geometry Selector (QGS), which compress the keys while respecting the query directions. Additionally, unlike prior methods that rank clusters by centroid alone, ErrorGate allocates the exact read by the estimated weight error of each cluster. On RULER-32K HARD with Llama-3.1-8B, OpenAttention retains 99% of dense accuracy at a 1% read. It also achieves state-of-the-art accuracy across models, context lengths and datasets. Our customized attention kernel, this http URL, runs the decode attention stage 10 \times faster than dense attention at a 1% read.
[AI-69] he Geometry of Meaning in Neural Manifolds
链接: https://arxiv.org/abs/2610.30938
作者: Valentino Maiorca
类目: Artificial Intelligence (cs.AI)
备注: PhD thesis, Sapienza University of Rome. 214 pages, 62 figures, 32 tables. Advisor: Emanuele Rodolà. Co-advisor: Francesco Locatello (ISTA)
Abstract:Just as different languages express the same concept through different words, neural networks express similar functions through incompatible coordinate systems. Different architectures, random initializations, and optimization trajectories produce incompatible parameterizations in extrinsic coordinates, obscuring functional relationships between independently trained models. When models independently solve similar tasks on related data, do their learned representations share exploitable structure despite incompatible coordinates? This thesis shows that semantic correspondence is a fundamental primitive revealing shared geometric structure beneath incompatible parameterizations. By identifying which elements refer to the same semantic content, correspondence breaks the symmetry of arbitrary implementation choices, anchoring incompatible systems to a shared semantic frame and revealing the intrinsic geometry that enables zero-shot transformation between independently trained models. Building on this principle, fourteen contributions explore correspondence-based alignment across zero-shot model stitching, cross-modal translation, policy transfer, activation steering, and neural decoding. They span domains (vision, language, reinforcement learning, neuroscience), architectures (CNNs, Transformers, graph networks), and scales (from small supervised models to billion-parameter generative systems). Each exploits different geometric properties (angular, metric, spectral, sparse, or distributional), yet all instantiate the same principle. It extends beyond artificial systems: fMRI data show that different human brains processing identical stimuli exhibit alignable geometric structure. By revealing intrinsic geometric structure beneath extrinsic parameterization, semantic correspondence transforms model isolation into interoperability. Comments: PhD thesis, Sapienza University of Rome. 214 pages, 62 figures, 32 tables. Advisor: Emanuele Rodolà. Co-advisor: Francesco Locatello (ISTA) Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2610.30938 [cs.AI] (or arXiv:2610.30938v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2610.30938 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-70] Low-Latency MoE Inference in Space Data Centers: A Wireless-Power-Transfer Perspective
链接: https://arxiv.org/abs/2610.30923
作者: Haotian Zheng,Zhanwei Wang,Kaibin Huang
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Networking and Internet Architecture (cs.NI)
备注: 13 pages, 11 figures
Abstract:Space data centers (SpaceDCs) are emerging as platforms for large language models (LLMs), supported by continuous solar-energy harvesting and high-speed interconnections. The mixture-of-experts (MoE) architecture supports distributed LLM inference by deploying experts across satellites. However, skewed expert activation creates uneven computation power demands despite uniform solar-power supply at the dawn-dusk sun-synchronous orbit (SSO). Consequently, active experts remain power-constrained while harvested power at idle experts is underused. These reduce inference efficiency and increase latency. To address this supply-demand mismatch, we propose the SpaceWPT framework, which uses inter-satellite laser-based wireless power transfer (WPT) to redistribute harvested power. The key design challenge is to balance the computation demands of power-donating and power-receiving satellites under WPT losses: transferring more power accelerates computation at the receiver but leaves less power for the donor. After deriving design principles that simplify WPT control, we characterize the resulting latency tradeoff through a tractable surrogate of average inference latency. Minimizing the surrogate yields a novel two-sided water-filling strategy that determines whether each active expert donates, retains, or receives power according to the water level and its WPT link efficiency. Simulations with Mixtral-8x7B and Phi-3.5-MoE show up to 29.25% lower inference latency than without WPT, with larger gains under more compact formations and higher transfer efficiencies.
[AI-71] DConstrain: Compiling Context-Parameterized Transitions for Diffusion Language Model Constrained Decoding
链接: https://arxiv.org/abs/2610.30920
作者: Shidong Cao,Jing Ma,Wenhao Chai,Yuchen Tian,Lun Du,Da Zheng
类目: Artificial Intelligence (cs.AI)
备注: 39 pages, 6 figures, 12 tables
Abstract:Diffusion language models (DLMs) decode multiple tokens in parallel, so individually valid tokens can be incompatible together under a schema. Exact joint inference over a grammar graph enforces compatibility, but it requires token-labeled successors across many hypothetical parser configurations. Encoding stacks and field histories in compiled states can cause exponential growth, which forces prior exact decoders to fix property orders and exclude valid outputs. We present DConstrain, which compiles schemas into context-parameterized transition programs. Cold compilation specializes token consumption to local grammar states while retaining context-dependent stack operations, environment guards, and caller-dependent deferred parsing symbolically. Tracking field histories as runtime values supports arbitrary property order without enumerating field subsets offline. At runtime, warm compilation binds reached configurations to these programs to construct exact successors, overlapping with the model forward pass because graph construction requires no model scores. Furthermore, token-effect grouping merges tokens with identical compiled effects before runtime contexts are bound, so groups are reused across stacks and environments even when concrete successors differ. Objective-specific score aggregation and token restoration provably preserve both the Viterbi optimum and the token-labeled path distribution of ungrouped inference. Across three benchmarks and two DLM families, DConstrain compiles 100% of BFCL schemas under arbitrary property orders, whereas an optimized finite-state baseline compiles only 71.30-85.81%. Under arbitrary property order, it reduces BFCL end-to-end decoding latency by 6.65-46.26 times with 100% schema validity and raises DiffusionGemma’s BFCL accuracy from 37.27% under fixed order to 81.56%.
[AI-72] Harnessing Language Models as Optimization Modeling Agents
链接: https://arxiv.org/abs/2610.30917
作者: Yunxiang Luo,Xin Xu,Tianhao Chen,Zhixi Xiong,Peiqi Jiang,Fan Zhang,Wei Shao,Xuejian Gao,Yuwei Fan,Can Yang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) can translate natural-language optimization problems into solver programs, but reliably producing correct formulations and implementations remains challenging. Existing approaches improve optimization modeling through agentic workflows or task-specific post-training. This raises a practical question: can a lightweight harness using a single model role achieve competitive performance without additional training? We introduce OPTICYCLE and OPTISTATE, two lightweight inference harnesses for optimization modeling, and evaluate them on nine benchmarks comprising 2,634 problems using 11 Qwen variants spanning 0.6B-27B parameters. Our results show that execution-guided revision consistently improves macro accuracy over direct generation for all evaluated models, with gains of up to 49.46 percentage points. Despite their simplicity, both harnesses outperform six adapted agentic workflows in same-model comparisons and four evaluated post-trained optimization models at comparable model sizes, without additional training. Explicit formulation state provides further but model-dependent improvements, suggesting that additional structure is not universally beneficial.
[AI-73] MIG-COW: Transferable Adversarial Attacks on Deepfake Detectors via Gradient Decomposition ACM-MM’25
链接: https://arxiv.org/abs/2610.30914
作者: Wonjune Seo,Joonhyuk Baek,Yeseong Jung,Saerom Park
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 7 pages, 3 figures, ACM MM '25
Abstract:Despite recent advances, deepfake detectors remain vulnerable to adversarial examples, particularly in diverse, real-world settings. We propose MIG-COW, a novel adversarial attack framework that generates highly generalizable and visually imperceptible adversarial examples. By combining momentum-integrated gradients with a consensus-orthogonal decomposition, MIG-COW captures both shared and model-specific vulnerabilities across heterogeneous CNN and ViT detectors. On the AADD-2025 Challenge benchmarks, MIG-COW achieves a 99.96% white-box attack success rate (ASR) with high perceptual similarity (SSIM), significantly outperforming existing baselines. However, its limited 7.16% ASR against official black-box targets - despite achieving the best overall score - highlights the ongoing challenge of transferability. We also demonstrate that incorporating low-performing but diverse models in the ensemble can degrade attack effectiveness, underscoring the need for careful surrogate model selection in real-world adversarial settings.
[AI-74] winStore: A Synthetic Consumer Platform Using Vision–Language Models
链接: https://arxiv.org/abs/2610.30909
作者: Minhyeok Lee
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We present TwinStore, a source-available synthetic consumer platform for physical retail built on vision–language models. TwinStore turns a store into a digital twin—from 82 ready-made formats, a floor-plan image or phone photos of its shelves—and fills it with synthetic shoppers who plan a trip, walk the aisles and decide what to buy by looking at rendered shelves with packs, price tags and promotion signs. Every trip ends in a receipt, so a layout, price, pack or promotion can be tried on the same shoppers before it reaches the store. Around this core, TwinStore provides 3D playback, paired comparisons and layout optimization, shopper interviews, a trade-area model, counterfactual edits of real shelf photos, a marketing studio with eleven tools, and an assistant that turns a plain sentence into a drafted study or an answer grounded in a run’s own numbers. In a held-out validation on eleven public datasets, simulated receipts approach the households’ own natural variation (category-mix divergence 0.017–0.018 against a real drift of 0.012), price elasticities fall inside the range measured in the field, and three independent decision models agree. TwinStore runs on cloud or on-device models and is distributed with installers for Windows, macOS and Linux.
[AI-75] Beyond Phenotypic Similarity: Relation Refinement and Selective Population Graph Learning for Multimodal Brain Disorder Diagnosis
链接: https://arxiv.org/abs/2610.30902
作者: Yinyu Lan
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Population graph learning supports brain disorder diagnosis by sharing information across subjects. Its effectiveness, however, depends on whether the relationships guiding this exchange are relevant to diagnosis. Similarity in demographic or acquisition attributes does not necessarily indicate diagnostic agreement, and heterogeneous subjects may require different representation updates within the same population graph. We propose PopDRP, a framework that combines diagnosis-guided relation refinement with hierarchical selective state-space learning. Using only training labels, Contrastive Affinity Refinement calibrates attribute-based affinities according to diagnostic agreement. Neural Graph Rectification then refines these relationships using fused functional-connectivity and non-imaging representations, guided by graph reconstruction and dual-view alignment objectives. On the resulting graph, a Selective Pyramid State-Space backbone integrates input-dependent state updates, graph diffusion, and hierarchical pooling and unpooling to learn population representations at multiple resolutions. Together, these components connect the refinement of inter-subject relationships with selective representation learning over the population graph. Experiments on ABIDE and ADHD-200 under stratified ten-fold cross-validation show improved accuracy and AUC over the compared baselines. Ablation studies further support the contributions of relation refinement and the selective hierarchical backbone to classification performance.
[AI-76] Parameter Learning in Active Inference via Variational Message Passing
链接: https://arxiv.org/abs/2610.30899
作者: Benedikt Waldeck
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted for publication in the proceedings of the 7th International Workshop on Active Inference (IWAI 2026), Springer CCIS
Abstract:Active inference casts perception, learning and planning as variational inference, where the Expected Free Energy (EFE) defines the planning objective. Recent work showed that EFE minimization can be written as Variational Free Energy (VFE) minimization on a generative model augmented with epistemic priors, and that the resulting entropy-corrected objective is minimized by sum-product message passing (SP-MP) on modified factor functions. These formulations treat model uncertainty as inference over a finite set of known models. We extend them to planning during learning, where beliefs over the observation and transition parameters \theta_A , \theta_B are Dirichlet priors. Closed-form Dirichlet updates are usually obtained by a structured mean-field factorization between parameters and states, but under exactly that factorization the information-gain corrections vanish. We resolve this by giving every factor its own parameter prior, frozen during a rollout. Conjugacy then makes each factor belief a product of a marginal over the categorical variables and a one-step Bayes posterior of the parameters. This reintroduces the datum-dependent parameter posteriors that mean-field factorization destroys. We show that evaluating the epistemic priors under the resulting Bethe joint recovers the information gains. Stationarity of this single closed-form objective yields SP-MP whose factor functions are augmented by the information-gain weight. Experiments on two grid POMDPs, in which only a planner that learns the model finds the safe route, show that the method plans to efficiently learn \theta_A and \theta_B . It reaches the goal on every episode, whereas SP-MP planners with the same inference update plateau or take the risky route.
[AI-77] MOER: Mix-Policy Optimization under Expert Prefix Guidance with Selective Rollback
链接: https://arxiv.org/abs/2610.30873
作者: Xuyang Zhong,Jiuding Yang,Yiwen Guo,Chen Liu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Mix-policy optimization incorporates off-policy expert data into the reinforcement learning (RL) loop to transcend the capability boundary of on-policy RL. However, existing methods either provide static full-trace guidance that suppresses self-exploration, or adopt expert prefix guidance without correctness guarantees, leading to suboptimal performance or even training collapse. In this paper, we propose MOER, a mix-policy optimization framework that harmonizes expert guidance with self-exploration. MOER injects a decaying expert chain-of-thought prefix into the prompt and lets the model complete the remaining reasoning, with the prefix length progressively annealed to encourage self-exploration. A selective rollback mechanism replaces failed completions with the original full expert trace, guaranteeing the correctness of every off-policy trajectory. Furthermore, a harmonized mix-policy objective optimizes on-policy tokens with a GRPO loss while supervising both expert prefix and completion tokens with a trust-region SFT loss, substantially stabilizing training. Extensive experiments show that MOER achieves the highest average scores in mathematical and general domain reasoning tasks, outperforming all basic and mix-policy baselines, and generalizes consistently across diverse tasks, model scales, and base model categories. Codes are available at this https URL.
[AI-78] Separate Field Scoring Protects Unused Tools When Agents Pool Tool-Use Experience
链接: https://arxiv.org/abs/2610.30867
作者: Abhijit Chakraborty
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Agents that call tools from a shared catalog can pool what they learn about which tool serves which request. We measure what pooled usage examples do to tool retrieval, reporting separately the held-out tools, for which no agent has usage, and all other tools. Adding usage examples to the index as extra entries for used tools raises Hit@5 (whether a gold tool is in the top five) for those tools and lowers it for held-out tools, by a large margin on two of three catalogs and with both embedders we tested. Giving used tools the same number of equally long entries filled with other tools’ text lowers held-out Hit@5 about as much, but gives little or none of the gain. We call our method of scoring usage as its own field SYNAPSE. On StableToolBench, the held-out gold tool appears in its top five more often than with a router that reserves slots for usage (0.785 against 0.683), and this held-out lead holds on all three catalogs with both embedders. This protection persisted when we replaced each tool’s usage with other tools’ text, and a constant score offset for tools with usage gives similar held-out Hit@5 in most settings. The gain for tools with usage, in contrast, needs usage about the right tool. The router keeps more of the gain on labeled queries on TOUCAN with both embedders and on ToolRet with one, so the better choice depends on the workload. Splitting the usage among ten simulated agents shows that pooling helps exactly as much as giving one agent the same records.
[AI-79] Bifurcations from Weight Space: JEPAWG on Parametric Dynamical Systems NEURIPS2026
链接: https://arxiv.org/abs/2610.30847
作者: Tobias Göbel,Miranda C.N. Cheng
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Chaotic Dynamics (nlin.CD)
备注: 13+9 pages, 5+8 figures, 3+4 tables. A version of this work appears at the NeurIPS 2026 Workshop on Neural Network Artifacts as a New Data Modality
Abstract:Trained network weights are increasingly treated as data, but there is rarely ground truth about what a population of models should encode, which hampers systematic validation of weight-space methods. This makes parametric dynamical systems a natural test-bed: the dynamics change qualitatively at submanifolds of parameter space that are, in well-studied families, known analytically or numerically. Fine-tuning a forecasting network at each point of a parameter grid yields a model zoo indexed by a physical parameter, against which weight-space diagnostics can be checked. We build such model zoos for four dynamical systems and analyse them with JEPAWG, a recently proposed joint-embedding hypernetwork built around a learned parameter–weight representation. It was first introduced for neural samplers for lattice field theories, where its latent geometry located the phase transition and estimated the 2D Ising critical exponent. Applied to dynamical systems, JEPAWG is competitive on forecasting with a hypernetwork trained directly for that task, despite being supervised on the parameter–weight correspondence rather than on forecasts. Beyond forecasting, the norm of the parameter-derivative of the decoded weights locates Hopf points and curves, tracks a period-doubling cascade closely enough to give a qualitative estimate of the Feigenbaum constant, and detects a global bifurcation, which has no eigenvalue signature and is therefore invisible to linearisation. As such, JEPAWG provides interpretable insight into the dynamics that is not available from other hypernetworks. Together, these results indicate that JEPAWG is a workable instrument for reading physical structure off a population of trained networks. Comments: 13+9 pages, 5+8 figures, 3+4 tables. A version of this work appears at the NeurIPS 2026 Workshop on Neural Network Artifacts as a New Data Modality Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Chaotic Dynamics (nlin.CD) Cite as: arXiv:2610.30847 [cs.LG] (or arXiv:2610.30847v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.30847 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-80] Semi-Linear Gaussian Bayesian Networks: Enhancing Flexibility in Node Dependencies via penalized B-Splines
链接: https://arxiv.org/abs/2610.30838
作者: Aitor Oviedo-Madrid,Mabel Morales-Otero,Jose Gonzalez-Gomariz,Ruben Armañanzas
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 25 pages, 6 figures
Abstract:Bayesian networks for continuous variables often rely on restrictive assumptions: linear Gaussian models lack flexibility to capture nonlinear dependencies, while non-parametric approaches provide greater flexibility at a high computational cost. This creates a gap for modeling scenarios where variables are approximately Gaussian but exhibit nonlinear relationships. We introduce Semi-Linear Gaussian Bayesian Networks (SLGBN), a novel class of continuous Bayesian networks designed to bridge this gap, preserving Gaussian conditional distributions while introducing nonlinear conditional mean functions. The proposed framework provides a unified probabilistic framework combining linear Gaussian and nonlinear Gaussian conditional models within the same Bayesian network Nonlinear conditional mean functions are estimated using generalized additive models (GAMs) with penalized B-splines. We present parameter estimation procedures for both types of nodes, as well as a greedy hill-climbing algorithm for learning the network structure under the SLGBN framework. To assess its performance, we compare SLGBN with other continuous Bayesian network models, including semiparametric Bayesian networks (SPBN) and Linear Gaussian Bayesian networks (LGBN). Experimental results evaluate both goodness of fit and the ability to recover the underlying network structure. The results show that, in the presence of nonlinear Gaussian relationships, SLGBN achieves better fit and improved structure recovery compared to alternative models, particularly when combined with the Extended Bayesian Information Criterion (EBIC*) score. Furthermore, by retaining Gaussian conditional distributions, SLGBN offers computational advantages over fully semiparametric approaches.
[AI-81] Neural Spectral Capacity with Residual Context for Heterogeneous Resource Allocation within Transformers
链接: https://arxiv.org/abs/2610.30823
作者: Chenyu Zhu,Ruoyu Zhao,Zhichao Lu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Predicting relative trained performance from architectural specifications can guide design when training costs limit experimentation. Neural Spectral Capacity (NSC) provides such a heuristic, but scores projections independently and assigns identical scores to compatible reorderings of the same layer configurations. We introduce Residual-Aware Neural Spectral Capacity (ResNSC) for Transformers, using a specification-derived statistical reference for accumulated residual updates to determine the input and residual-state statistics in each projection’s capacity calculation, while retaining additive evaluation without data or instantiated weights. When residual widths vary, earlier choices affect later scores, so we augment NSC’s allocator (NSC-DP) with residual history beyond its layer-and-budget state. ResNSC-DP jointly allocates residual and feed-forward widths under a resource budget, with certified bounds on the best numerically evaluated ResNSC score within a specified finite candidate set. At the 121M and 374M pretraining scales, selected architectures achieve lower mean perplexity than uniform, variable-width, and tapered allocations on FineWeb-Edu, WikiText-2, and LAMBADA at comparable parameter budgets. In inherited-weight LLaMA-7B pruning using the LoNAS supernet, ResNSC-DP improves eight-task average accuracy over NSC by 3.28–5.46 percentage points across eight matched parameter budgets. Paired reversals preserve the projection shapes and NSC scores yet worsen performance in both settings, exposing performance differences hidden by position-independent scoring. The code for this work is available at this https URL.
[AI-82] ESSA: A Governance-Oriented Custom GPT for SME Guidance under the EU AI Act: Design Implementation and Evaluation
链接: https://arxiv.org/abs/2610.30822
作者: Chalisa Veesommai Sillberga,Kai-Kristian Kemella,Pekka Sillberga,Katri Harjuveteläinenb,Mika Saaria,Pekka Abrahamssona
类目: Artificial Intelligence (cs.AI)
备注: 18 pages and 2 figures. Accepted for presentation at the 17th International Conference on Software Business (ICSOB 2026)
Abstract:The increasing adoption of generative AI presents new opportunities while raising AI governance and regulatory challenges, particularly for small and medium-sized enterprises (SMEs) navigating the European Union Artificial Intelligence Act (EU AI Act). To support AI governance education, this study developed TESSA, a governance-oriented educational artifact within the OpenAI Custom GPT environment. Following a design science research approach, TESSA was evaluated against baseline ChatGPT using 100 benchmark scenarios through governance alignment assessment, human expert review, and consistency testing. Preliminary LLM-assisted assessment showed governance alignment rates of 90% for TESSA and 83% for baseline ChatGPT, while independent validation by an AI regulation law practitioner yielded 71% and 70%, respectively. Human expert review indicated that both systems provided generally strong governance guidance, while TESSA maintained high consistency across repeated executions. The findings suggest that governance-oriented grounding can support structured, practical, and stable AI-assisted compliance guidance for SMEs. The study contributes (i) a governance-oriented Custom GPT artifact for AI governance education, (ii) a human-validated regulatory update mechanism, and (iii) a multidimensional evaluation framework for governance-oriented AI artifact.
[AI-83] What Lies Beyond Perception: A Partition Theory of the Limits of AI Evaluation and Oversight
链接: https://arxiv.org/abs/2610.30803
作者: Ahmed M. Mahran
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (cs.LG)
备注: 36 pages (23 main text, then references and appendices), 6 figures, 2 tables
Abstract:Replacing the judge can move a benchmark score, and models can be trained to hide capabilities. Both put the evaluator in question: what can it certify? We develop a partition theory of evaluation from one principle: judgement is bounded by perception. An agent’s discrimination is the partition of its inputs into classes it cannot tell apart. Discrimination is necessary for capability but not sufficient, so bounds on it bear directly on whether evaluation can show that a model lacks an ability. We prove a one-sided bound. An examinee whose partition refines its evaluator’s has the discrimination to pass every test the evaluator can grade, provided each input has an acceptable reply. We define the echo test, which accepts any reply the evaluator perceives as identical to the input, and show that it is complete for the tests the evaluator can grade. Passing therefore certifies at most a floor on discrimination, never a ceiling, and only a sharper evaluator can raise it. Used to detect concealment, echo-test failure points the wrong way: under exhaustive probing it flags every honestly weaker examinee and no concealer that presents as the evaluator’s peer. Exact imitation of a weaker model is undetectable by any behavioural evaluator. We extend the bound to ensembles of monitors and to channels a model controls, explain published findings on judge disagreement and sandbagging, and derive testable predictions. A safety case that argues from behavioural evaluation alone that a model lacks an ability must assume the model’s cooperation.
[AI-84] VLTest: Finding Joint Image-Text Failures in Vision-Language Models
链接: https://arxiv.org/abs/2610.30794
作者: Bin Duan,Jintao Lin,Guowei Yang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision-language models (VLMs) are increasingly embedded in software systems. Testing them raises two problems. Mutating either input can change the expected answer, so an output change is not by itself a failure. Some failures appear only when both inputs are mutated; attributing them to the combination requires testing each mutation alone. We present VLTest, a black-box VLM testing approach. It generates image variants by editing VQGAN latent codes and text variants with a rewrite-expand-prune pipeline, filters variants likely to change the expected answer, and tests each variant alone before testing combinations, so every joint failure has four matched cells. A VLM judge confirms a failure only after checking that the input keeps the expected answer. On three tasks with eight VLMs, including two commercial API models, the oracle confirms 81.7% of the candidates, with 88.8% to 94.6% precision against human labels, and four-cell confirmation rejects 22.5% of joint candidates. Under the same budget, VLTest’s generators find 1.8 to 2.5 times as many confirmed failures as SINVAD and TextFooler, and its joint tests find 91 to 197 attributed joint failures per 1,000 joint calls, 1.3 times as many as the same pairing applied to the baselines’ variants.
[AI-85] Agent ic Knowledge-Model Co-Evolution via Consistency Regularization
链接: https://arxiv.org/abs/2610.30790
作者: Sirui Chen,Zhining Liu,Chao Zhang,Jingrui He,Zheng Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LLM agents can evolve by learning from accumulated experience. Existing approaches extract reusable external knowledge, such as skills and memories, from experience to guide subsequent task solving, with some further internalizing this knowledge through model training. However, different ways of analyzing and transforming the same experience can yield knowledge with substantially different utility, even when the resulting guidance appears reasonable. Inspired by consistency regularization, we identify consistent utility gains across different ways of analyzing the same experience, rather than relying on any single way of analyzing it. We introduce SPLICE, a framework for consistency-regularized knowledge-model co-evolution. It constructs evolution views by perturbing how the same experience is analyzed, producing different knowledge candidates. Knowledge updates are then guided by common gain, which accounts for utility gains and their consistency across views. During model evolution, self-distillation internalizes the evolved knowledge alongside reinforcement learning, using knowledge associated with consistent utility degradation as a contrastive reference. The two processes interleave, with model updates driving further knowledge evolution. Experiments across diverse domains show that SPLICE outperforms knowledge evolution and model training baselines by up to 7.1 and 4.6 points in average accuracy, while its evolved knowledge reduces output tokens per correct solution by up to 44% and transfers across model sizes and families.
[AI-86] An Introduction to World Models for Generalist Agents
链接: https://arxiv.org/abs/2610.30789
作者: Zhihui Zhu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:World models have recently emerged as a promising direction toward developing generalist agents capable of understanding, predicting, and interacting with diverse environments. However, despite growing interest, the concept of a world model remains loosely defined, encompassing approaches ranging from predictive dynamics models and learned simulators to models that jointly generate future observations and actions. This introductory paper aims to clarify the concept of world models and provide a unified perspective on their role in developing generalist agents. Starting from the partially observable Markov decision process (POMDP) framework, we connect world modeling to classical control, dynamical systems, signal processing, and reinforcement learning. We introduce different approaches to world modeling, emphasizing the knowledge they capture, their learning objectives, and how they support decision-making. We further explore the relationship between world modeling and policy learning, highlighting how these capabilities can be unified through conditional prediction without requiring separate models. Rather than offering an exhaustive survey of existing methods, this paper focuses on fundamental concepts, modeling principles, and connections, with the goal of providing an accessible and unified introduction to world models for generalist agents.
[AI-87] ProQuant: Preserving Expert Quality over Quantity in Ultra-Low-Bit Quantization of Mixture-of-Experts Models
链接: https://arxiv.org/abs/2610.30779
作者: Jiaqi Chen,Xianglong Yan,Chengzhu Bao,Xie Chen,Yulun Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Mixture-of-Experts (MoE) large language models reduce computation through sparse activation but still incur substantial storage costs. Mixed-precision post-training quantization (PTQ) reduces these costs by quantizing different model components at different bit widths, but existing methods often struggle to preserve generation quality under tight bit budgets. We identify two key challenges: (1) nonlinear quantization errors at extremely low precision are underestimated when cross-layer propagation is ignored; (2) direct router fine-tuning may cause routing drift, and calibration gains may not generalize to downstream tasks. To address these issues, we propose ProQuant, a weight-only PTQ framework for MoE models. Prune-or-Prioritize Mixed-Precision Allocation (PoP) jointly optimizes expert pruning and bit-width assignment. It allows higher-precision experts to substitute for severely degraded ones while improving bit-budget utilization. Routing-Consistent RMSNorm Finetune (RoCo) adapts expert input representations through mathematically equivalent transformations. It compensates for quantization errors to reduce routing drift. Extensive experiments demonstrate that ProQuant outperforms existing methods across diverse generative tasks. Code is available at this https URL.
[AI-88] Scaling Laws for Generative Reward Models
链接: https://arxiv.org/abs/2610.30765
作者: Angel Raychev,Nathan Lile,Rafael Rafailov
类目: Artificial Intelligence (cs.AI)
备注: 28 pages. Code and research artifacts: this https URL
Abstract:We study generative reward models (GenRMs) as online preference providers for policy optimization in creative writing. Across Qwen3 reward-model sizes from 0.6B to 14B, Thinking judges trained with GRPO have higher static Gold-label accuracy than Answer-Only judges trained with SFT. This ranking often reverses when the judges train policies through online DPO: Answer-Only rewarders yield higher peak and top-five-average policy Elo in seven of eight matched settings under a fixed, learned 32B Gold evaluator. Intermediate Thinking-judge checkpoints can also outperform later, more statically accurate checkpoints as rewarders. Sweeps over rewarder size, policy size, training duration and DPO regularization show substantial benefits from reward-model capacity, including judges larger than the policy. We analyze trajectories with evaluator-specific Elo arenas, conditional match-resampling intervals and descriptive scaling fits. These results demonstrate an evaluator-rewarder gap in the studied pipelines: fixed-distribution judging accuracy does not determine downstream optimization quality. The comparison changes judge-training objective as well as response format, and the main ranking is conditional on its learned reference rather than a universal human-quality result. A limited follow-up using a different learned judge is inconclusive overall.
[AI-89] Agent DP: Towards Automated Model Deployment for Heterogeneous Embedded AI Applications
链接: https://arxiv.org/abs/2610.30760
作者: Leming Shen,Qiang Yang,Xinyu Huang,Zijing Ma,Yuanqing Zheng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:This paper presents AgentDP, which aims to deploy pre-trained AI models on heterogeneous embedded devices by fully automating model optimization and adaptation (OA) with Large Language Model (LLM) agents. AgentDP not only significantly reduces human effort via automation but also enhances model inference efficacy by integrating LLMs with device-specific domain knowledge. To deliver a robust automation framework, AgentDP develops novel solutions to address three key technical challenges: 1) To help users elicit clear requirements, we propose a user requirement standardization module; 2) To mitigate the execution environment mismatch between AI models and heterogeneous embedded devices, we design an operator-aware model partition paradigm that enables different parts of the model to be efficiently executed on distinct inference engines. 3) To avoid re-searching optimization hyperparameters for new tasks, we propose a configuration reuse module that heuristically retrieves previous hyperparameter configurations for efficient reuse. Extensive experiments demonstrate that by fully automating the entire OA process, AgentDP reduces entire model adaptation time by up to 80% compared with manual OA while achieving higher task accuracy (more than 13% on average) than the baselines. Additionally, a comprehensive user study also exhibits the robust generalizability of AgentDP in adapting to different user requirements for heterogeneous embedded devices.
[AI-90] Is Complex Agent ic Memory Always Needed? Naive RAG Is Stronger Than We Thought
链接: https://arxiv.org/abs/2610.30755
作者: Zhengye Bao,Shengtao Zhang,Haokai Zhang,Shiqi Lou,Muning Wen,Minnan Luo
类目: Artificial Intelligence (cs.AI)
备注: 62 pages, 5 figures. Includes appendices
Abstract:LLM memory is widely used in long-term conversation understanding. Agentic memory systems often use mutation, structural organization, learned retrieval strategies, or other complex operations. Is this complexity always necessary? We study Raw-RAG, a naive RAG baseline using dense retrieval and reranking over raw history to supply evidence to a reader LLM. Across LoCoMo, LongMemEval, MEME, and MemoryAgentBench, Raw-RAG ranks first among the compared memory systems in six of seven benchmark-model settings, outperforming the strongest compared agentic memory system by 0.13-5.53 points. We investigate why raw-history retrieval remains competitive. Evidence analysis shows that its higher coverage of necessary evidence can offset lower utilization. Useful evidence includes both answer values and the relations and dependency paths needed to derive them. Even complete access to transformed memory stores leaves an accuracy gap relative to raw evidence, highlighting the importance of preserving raw information. These findings challenge the necessity of complex agentic memory for the evaluated conversation tasks and establish naive RAG as an essential baseline.
[AI-91] Agent Triage: Replayable Security Diagnosis Across Heterogeneous Agent Runtimes
链接: https://arxiv.org/abs/2610.30738
作者: Jie Ying,Xutao Zhang,Ningyuan Chen,Mengce Zheng,Jun Li,Zhongjie Zhu,Tiantian Zhu
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:When tool-use agents write files, execute code, or send data, their final responses do not explain the resulting security events, and heterogeneous runtimes expose the relevant evidence through incompatible interfaces. We present AgentTriage, an executable contract for replayable security diagnosis across heterogeneous runtimes. The contract admits conclusions whose claims cite canonical runtime events and requires abstention when mandatory evidence is missing. Runtime overlays normalize native observations, while a deterministic pipeline joins invocation and result lifecycles, constructs a typed provenance graph, selects a policy-decisive slice, and binds claims to recorded evidence. Given the same normalized trace, overlay version, and pre-run policy oracle, the diagnosis can be recomputed and its claims re-checked without language-model judgment. Across five runtimes, one contract captures all 159 applicable executions, with an audit of native and synthesized identifiers. On 120 adjudicated cases, AgentTriage achieves 0.858 grounded joint diagnosis, 0.966 required-evidence coverage, and a 0.036 unsupported-claim rate, improving joint diagnosis by 0.175 (95% CI [0.092, 0.258]) over the strongest same-schema baseline. These results quantify the precision of replayable diagnosis and the observation boundaries that require abstention. We release an open-source implementation and selected sanitized evidence at this https URL.
[AI-92] Faithful Latent Harm Monitors Work No Better Than the Models Own Behaviour at a Matched False Alarm Rate
链接: https://arxiv.org/abs/2610.30727
作者: Abhishek Mishra,Sagnik Chatterjee,Suyash Maniyar
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models have been reported to encode the harmfulness of a request separately from their decision to refuse it, and latent safety monitors such as Latent Guard build on this report by classifying prompts as harmful from the internal activations of the model, with the aim of catching harmful intent even when the model complies. Whether these representations give a signal of harm that survives distribution shift and is independent of model behaviour is unclear. We test this on out-of-distribution prompts made by rewriting existing harmful and benign requests with surface rewording, persuasive framing and jailbreak-style wrappers. Across 15 instruction-tuned LLMs and 38 activation-based readers, we compare harm detection with that of TF-IDF and sentence-embedding classifiers, with independently trained safety classifiers, and with the refusal of each model. Activations are read at the last token of the user instruction, where the model has read the whole request, and at the last chat-template token before the reply, where the planned response can be represented. Because models disagree about which prompts to refuse, we can separate a signal of the harmfulness of the prompt from a signal of the refusal decision of the model. At the end of the instruction, the representation is largely independent of the refusal behaviour of each model but degrades under distribution shift, so that a TF-IDF classifier ranks harm better than the original Latent Guard in every model we test, and linear and MLP probes do about as well as sentence-embedding classifiers trained on the same examples. The representation just before generation detects more, catching 44 percent of harmful prompts at a matched false-alarm rate against 32 percent for the refusal of the model, but that gain follows the refusal decision itself, and part of it comes from refusals that string-matching evaluation misses.
[AI-93] Governance-State Corruption in Long-Term Memory for LLM Agents : A Cross-Model Stress Test
链接: https://arxiv.org/abs/2610.30698
作者: Bowen Liu,Youting Wang,Bonan Shen,Dingyan Shang
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 13 pages, 1 figure, 10 tables
Abstract:An LLM agent can retrieve a memory correctly yet use it incorrectly after its validity state changes. We study governance-state corruption: faults that alter which retained records enter model-visible context while stored content and tasks remain fixed. Across six providers, 731 benchmark-grounded bases yield 2,080 matched fault pairs. In offline control-plane checks, corruption changes the operative projection in 96.25% of pairs, while oracle-backed enforcement restores it in 100%. For downstream evaluation, a deterministically selected, non-probability LongMemEval/Memora panel resolves 298 of 300 matched triplets (238 distinct bases). Clean-minus-corrupted contrasts are 40.54 percentage points in LongMemEval binary correctness (95% base-cluster interval [30.32, 50.71]) and 30.60 points in rescaled Memora FAMA ([23.31, 37.69]); both sources also show recovery. OAKS Novel native option parity holds for all 6,714 resolved outputs. Full-cohort frozen-scorer corruption/recovery effects (-27.55/+27.51 points) remain descriptive proxies, not validated native semantic population effects. These bounded findings motivate evaluating memory control-state integrity alongside retrieval accuracy.
[AI-94] A Topology-Guided GeoAI Agent ic Framework for Urban Flood Screening
链接: https://arxiv.org/abs/2610.30660
作者: Yutian Tao,Haoan Feng,Xin Xu,Leila De Floriani
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Open-world geographic artificial intelligence (GeoAI) agents can transform public geospatial data into accessible decision support for real-world challenges, but must balance autonomous reasoning with reliable spatial analysis and human oversight. We present a topology-guided GeoAI agentic framework for urban flood screening that takes a place name, drawn region, or uploaded boundary and returns ranked candidate basins with an interactive map and report. A deterministic controller retrieves public elevation, rainfall, soil, land-cover, exposure, and infrastructure layers, extracts terrain depressions with discrete Morse theory, and ranks basins with a relative flood-priority score. The language model acts only at recorded decision points: selecting a persistence threshold from a precomputed candidate table, drafting recommendations from computed evidence, and reviewing each recommendation against the same evidence. Three human-in-the-loop checkpoints let users revise the threshold, factor weights, or recommendation. In a College Park, Maryland case study, the selected threshold reduced 278 raw terrain minima to 70 basins, with 77% of these basins’ lowest points also identified as sinks by a separately computed eight-direction (D8) flow-routing model. Across four further study areas, agreement ranged from 61% to 83%. By combining open data, explicit spatial reasoning, traceable evidence, and human oversight, the framework enables transparent first-pass flood screening while keeping consequential recommendations inspectable and user-controlled.
[AI-95] Anticipating Social Consequences with Distributional World Models
链接: https://arxiv.org/abs/2610.30657
作者: Viet-Thanh Pham,Thuy-Trang Vu,Zhuang Li,Lizhen Qu,Gholamreza Haffari
类目: Artificial Intelligence (cs.AI)
备注: 29 pages, 8 figures
Abstract:Social world models predict how interactions develop after an action, helping agents negotiate, cooperate and mediate. In these situations, one action can lead to several plausible futures, hence predicting only one future can overlook consequential alternatives, while exploring them through repeated simulations increases inference cost. We propose DSWM, a Distributional Social World Model that predicts alternative responses, their probabilities and immediate and final outcome values in one output per candidate action. We train a 4B model on repeated simulations using supervised fine-tuning and GRPO. At inference time, the actor uses these possible consequences to choose an action. Across three seen and two unseen environments, with interaction backbones absent from training collection, DSWM achieves the best average performance on 13 of 15 metrics. Ablations show that retaining multiple futures and their predicted probabilities improves outcomes in all five environments.
[AI-96] Speculative Acceptance Does Not Decay with Position: A Full-Length Diagnostic on Long Chain-of-Thought with Independent Drafts
链接: https://arxiv.org/abs/2610.30644
作者: Zhiming Meng
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 9 figures
Abstract:The claim that speculative-decoding acceptance rate decays as a chain-of-thought grows longer has become a de facto industry narrative, motivating draft-model co-design and position-aware training. Nearly all direct evidence comes from trained drafters whose offline training length is far shorter than inference, so a purely positional law has never been isolated. We present the first systematic, full-length measurement of acceptance rate as a function of generation position on MATH-500 (all 500 problems), under greedy and sampling decoding, with two independent drafts against DeepSeek-R1-Distill-Qwen-7B: a same-family draft (DeepSeek-R1-Distill-Qwen-1.5B) and a cross-family draft (Qwen2.5-1.5B-Instruct), yielding about 1.78M verification rounds. Using within-problem paired tests, we find that after isolating a greedy-decoding repetition pathology, acceptance rate does not decay with position in any of the four configurations within the supported ranges (greedy 0-8K, sampling 0-12K); the second half of each trajectory is slightly but significantly higher than the first. A controlled temperature sweep shows that the apparent sampling-mode degradation of high-entropy reflection tokens is a visibility artifact of argmax agreement, not a drop in expected acceptance probability. Rejections concentrate with high precision on reflection trigger tokens such as “wait”, “however” and “check”, an entropy-selective failure that a lightweight three-arm LoRA trajectory-alignment intervention (uniform, reflection-marked, entropy-weighted) fails to improve in mean accepted length. We thus supply the positional baseline that independent draft-target pairs were missing, and localize the residual failure to decision-point entropy rather than position.
[AI-97] Large language models show systematic failures in incorporating evolving medical knowledge
链接: https://arxiv.org/abs/2610.30631
作者: Min Zeng,Zaifu Zhan,Kai Yu,Rui Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Medical knowledge evolves as clinical guidelines revise recommendations, yet whether large language models (LLMs) faithfully incorporate such changes remains unclear. We introduce ChronoEvo, comprising 1,709 recommendation-level revisions across 12 forms of medical knowledge change, to distinguish recognition of revised guidance from incorporation into clinical recommendations. In the primary model, 90.8% of responses recognized revisions and 80.2% were clinically correct, but only 58.5% fully incorporated guidance, yielding a 32.3-percentage-point recognition-incorporation gap. This dissociation persisted across 14 models and was corroborated by two clinicians’ adjudicated assessment of 300 responses, yielding a 40.3-percentage-point gap in the validation sample. Changes in recommendation strength and patient eligibility were associated with retention of outdated guidance. A revision-salience intervention reduced outdated-content retention and improved clinical correctness but did not significantly improve full incorporation. These findings identify faithful incorporation of evolving medical knowledge as a distinct challenge for clinical LLM reliability, even when updated guideline evidence is available.
[AI-98] Interpreting Model Reasoning through Behavioral Counterfactuals
链接: https://arxiv.org/abs/2610.30630
作者: Shawn Im,James Oldfield,Adel Bibi,Sharon Li
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Faithfully understanding the internal process that occurs as models form their chain of thought is an important problem with the rise in model capabilities. However, this remains challenging due to the variance in sampled responses, making it difficult to isolate the effects of interventions and compare model reasoning systematically. We propose a method for localizing and analyzing how a given intervention affects the model’s reasoning through behavioral counterfactuals: given a fixed response, how does introducing an intervention to the prompt change the next-token distributions? We train a module to edit the residual stream such that it minimizes the KL divergence between the output distributions for a fixed response under a source prompt and its counterfactual target prompt. We find that matching the output distributions recovers the counterfactual behavior more effectively than standard activation reconstruction objectives and validate that attribution of the trained module outputs is effective for identifying causally important locations for counterfactual recovery. The attribution maps and visualizations of the downstream effects of model edits provide an interface for generating hypotheses about the original model’s computation, which can then be tested with activation patching. In a case study on Gemma-4-E4B-it and DeepSeek-R1-0528-Qwen3-8B, we investigate how a suggestion influences model reasoning even when it is not verbally acknowledged. Our analysis identifies representations at the question statement and filler tokens that mediate a substantial part of the suggestion’s influence, with downstream effects such as changing the framing of the justification or even the final answer.
[AI-99] Future of Time Series Anomaly Detection: From Historical Review to Unified Outlook
链接: https://arxiv.org/abs/2610.30628
作者: Zhijie Zhong,Zhiwen Yu,Pengyu Li,Liujian Zhang,Qiaoli Ni,Jiang Jun,Jianming Lv,Kaixiang Yang,C. L. Philip Chen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 18 pages, 5 figures, 4 tables. Accepted by IEEE Transactions on Artificial Intelligence. DOI: https://doi.org/10.1109/TAI.2026.3746935
Abstract:Time series anomaly detection (TSAD) is a foundational problem in data mining, aiming to identify observations that deviate significantly from nominal temporal patterns. Driven by escalating data volumes and application complexity, recent research has yielded diverse architectures, including diffusion models, state space models, and large language models (LLMs). However, existing surveys provide limited coverage of these emerging designs and lack a unified treatment of multivariate anomalies, cross-domain unified modeling, metric reliability, and contemporary evaluation challenges. To bridge these gaps, this survey delivers three core contributions: (1) We establish a refined anomaly categorization and a structured taxonomy of paradigms reflecting modern advances, comparing detection mechanisms and charting developmental trajectories across paradigms. (2) We provide an extensive analysis of evaluation methodologies and conduct comprehensive experiments to assess metric fidelity, complemented by an open-source benchmark repository encompassing algorithms, datasets, and evaluation tools. (3) We outline key open challenges, offering forward-looking perspectives on dynamic thresholding, model interpretability, zero-shot transfer, multi-modal alignment, and foundational models for time series. By addressing existing theoretical and empirical limitations, this survey establishes a coherent framework to guide future TSAD research and practical deployments.
[AI-100] Model Organisms of Collusion: Corrective Transfer in Untrusted Monitors
链接: https://arxiv.org/abs/2610.30582
作者: Juan-Pablo Rivera,Ari Spiesberger,Juan J Vazquez,Arun Jose,Axel Ahlqvist,Adeline Kassler,Meru Gopalan,Sandy Tanwisuth,John Hughes
类目: Artificial Intelligence (cs.AI)
备注: 33 pages, 6 figures, 15 tables
Abstract:AI control protocols aim to get useful work from models that may be misaligned by monitoring their behaviour. One approach uses a second copy of the same model as the monitor. This avoids relying on a weaker reviewer the model could outsmart, but risks collusion: the model, now monitoring itself, may cover up its own misbehaviour. One remedy is to fine-tune the monitor on whatever examples of misbehaviour the control team has caught. We ask whether this training also removes collusion that the control team has not yet found. We fine-tune Llama-3.3-70B and Qwen2.5-72B to install five policy-monitor behaviour pairs, correct the monitor on 300 honest reviews of one caught collusion, and measure collusion on the other four in unseen prompt formats. Averaged over the three seeds of each family, correction reduces held-out collusion from 91% to 14-20% under three of the four methods we compare (individual seeds 10-32%), while a control fine-tuned on the same number of unrelated examples leaves 56-93%. Family-average accuracy on control items changes by at most five points and capability is preserved, apart from a drop in mathematical accuracy on one Llama seed. Under the tested training configuration, methods with similar caught-task accuracy differ in held-out transfer, from 58% to 83% average reduction, so the caught behaviour alone does not reveal how far a correction reaches. Catching one collusion lets the control team reduce, but not remove, collusion it has not caught, by retraining the monitor alone.
[AI-101] Bridging the Closed-World Gap: Interpretable Anomaly Detection via a Hybrid TSK-FCM Ensemble on the BETH Dataset
链接: https://arxiv.org/abs/2610.30576
作者: Lohith Srikanth Pentapalli,Josette Riep,Justin Ouwerkerk,Kelly Cohen
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Presented at North American Fuzzy Information Processing Society (NAFIPS) 2026, El Paso, Texas, USA; accepted for publication in Springer proceedings
Abstract:A fundamental paradox plagues the deployment of Machine Learning in cybersecurity: supervised models generalize poorly to novel, Out-of-Distribution (OoD) attacks, whereas unsupervised models often suffer from poor separability, leading to unmanageable false-alarm rates. This challenge is exemplified by the BETH dataset, a kernel-level benchmark in which the training data consists solely of benign and suspicious insider behaviors, whereas the test set consists almost entirely of novel botnet activation attacks. To bridge this detection gap, we propose a Hybrid Fuzzy Ensemble that synergizes the precision of supervised learning with the adaptability of unsupervised clustering. Our architecture integrates a gradient-optimized Takagi-Sugeno-Kang (TSK) fuzzy classifier to act as an expert on known insider threats, with a parallel Fuzzy C-Means (FCM) module to detect geometric anomalies. A critical finding of our research is that traditional fuzzy membership scores degrade in high-dimensional feature spaces due to the curse of dimensionality. We overcome this by introducing a robust Min Distance metric that exploits the raw geometric isolation of botnet activation attacks. This approach yields a state-of-the-art AUROC of 0.9951 on the BETH test set, outperforming the original Isolation Forest baseline by 17.1%. By combining near-perfect OoD detection with the intrinsic interpretability of fuzzy logic, our framework offers a viable path toward robust, transparent, and autonomous network defense.
[AI-102] Diagnosing Personalization Failures in Mobile GUI Agents
链接: https://arxiv.org/abs/2610.30573
作者: Zhixin Lin,Dongliang Xu,Jungang Li,Shidong Pan,Jun Feng,Yanbin Sun,Jing Li,Liang Liu,Yue Yao
类目: Artificial Intelligence (cs.AI)
备注: 24 pages, 8 figures
Abstract:Unlike conventional GUI agents, personalized GUI agents require user preferences to guide sequential execution over changing GUI states. This raises two challenges: determining HOW each preference should alter execution strategy and WHEN it should take effect along the interaction trajectory. Existing benchmarks primarily evaluate overall task success, offering limited insight into these questions. We introduce PrefDiag-Bench with two components: (1) a categorization of preferences into five types, characterizing HOW they affect execution; and (2) a stage-wise diagnostic framework with 1,664 probes, tracing WHEN preferences should take effect and where failures occur. We evaluate eleven GUI agent systems, including end-to-end agents and agentic workflows, under two forms of user context: explicit user profiles and behavioral memory logs. Although existing agents achieve relatively high completion rates on non-personalized tasks, performance is substantially lower on preference-personalized tasks. Even the best-performing model, Seed-2.0-Pro, achieves only 47.17% on Process preferences under explicit user profiles. Agents struggle particularly with Ranking and Process preferences. Further analysis reveals two key capability bottlenecks: selecting relevant evidence from user history and recognizing when preferences should take effect during execution. Controlled interventions that explicitly provide relevant evidence improve end-to-end task success by up to 7.06%. PrefDiag-Bench offers a more systematic evaluation of GUI agents’ personalization capabilities beyond overall task success. Code and data are available at this https URL.
[AI-103] PLE: Physical-Engineerings Last Exam
链接: https://arxiv.org/abs/2610.30537
作者: Rujun Han,Yufan Zhuang,Arijit Ray,Han Yu,Zoey CuiZhu,George Lee,Yifei Ming,Binz Roy,Varun Sakalkar,Vishy Tirumalashetty,Burak Gokturk,Tomas Pfister,Chen-Yu Lee
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Frontier language models now score above 90% on popular knowledge benchmarks and are increasingly deployed as agents that carry out multi-hour professional work. Yet no existing benchmark tests whether they can perform the real-world, economically valuable work of a professional physical engineer, such as modelling physics and legal code requirements across a multi-format project corpus, optimising material and installation costs, and issuing 2D/3D engineering design packages that a licensed reviewer would sign off on. Existing engineering benchmarks focus on textbook question answering, isolated tool calls or self-contained toy design problems. As a result, most of these benchmarks are already saturated by leading agents and solving them is far away from delivering real engineering design. We introduce PLE, a benchmark of 85 long-horizon, agentic tasks in professional physical engineering consisting of electrical, civil, structural and mechanical engineering, where solving a task means completing a realistic, billable engineering deliverable. PLE combines two complementary curation paths. Expert-Authored tasks are created end to end by practising engineers from a professional data provider’s expert pool. Co-Authored tasks start from problem statements written by practising engineers; an adversarial Creator-Solver agent pipeline constructs each task’s realistic project environment, executable grader and reference deliverables before sending them to expert engineers for validation and difficulty intensification. We evaluate a variety of frontier LLMs under seven publicly available agent harnesses. Even the strongest configuration reaches a mean score of only 50.6%, while spending 27M tokens and 63 minutes per task on average, demonstrating the substantial gap between current agents and autonomous physical engineering design.
[AI-104] Decision Making and Ride Comfort Evaluation Based on Lane-Change Prediction for Autonomous Vehicles
链接: https://arxiv.org/abs/2610.30528
作者: Mohamed Manzour,Aditya Kumar,Rubén Izquierdo,Miguel Ángel Sotelo
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Predicting surrounding vehicles’ lane-change intentions enables autonomous vehicles to anticipate potential hazards and respond proactively rather than relying solely on reactive maneuvers. This study investigates the benefits of integrating lane-change prediction, based on knowledge graph embeddings and Bayesian inference, into autonomous vehicle decision-making. Three driving strategies are evaluated in risky merging scenarios simulated in CARLA: reactive driving, prediction-based longitudinal action, and prediction-based lateral action. The proposed prediction model anticipates the surrounding vehicle’s lane change approximately 3-4 seconds before its occurrence, allowing the ego vehicle to adjust its behavior accordingly. In the representative simulation records, reactive driving results in a collision accompanied by aggressive deceleration. In contrast, prediction-based longitudinal action initiates early braking to establish a safe gap, while prediction-based lateral action performs a lane change when a safe adjacent lane is available. Both predictive strategies avoid collisions in the evaluated records. Compared with reactive driving, peak deceleration before impact is reduced by 31.5% with longitudinal action and 91.0% with lateral action. Additionally, an ISO 2631-1-based horizontal vibration estimate is used to provide a ride comfort assessment. Overall, the results highlight the potential of lane-change prediction to support safer and smoother driving through proactive decision-making, with additional insights into ride comfort provided by vibration-based estimates.
[AI-105] LLM 2Cert: LLM Synthesis of Certificates with SMT/Lean Verification
链接: https://arxiv.org/abs/2610.30524
作者: Zhipeng Cao,Peiyang Song,Nikolaus Vertovec,Weiru Cui,Zhihao Yang,Mingkang Liu,Luxia Lin,Bai Xue,Alessandro Abate
类目: ystems and Control (eess.SY); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Optimization and Control (math.OC)
备注: 59 pages, 3 figures. Equal contribution: Zhipeng Cao and Peiyang Song. Equal advising: Bai Xue and Alessandro Abate
Abstract:Lyapunov functions and barrier certificates give state-space guarantees of stability and safety, but synthesizing and verifying them—especially for nonlinear, high-dimensional systems—remains an open challenge. Existing pipelines verify candidates with numerical solvers (sum-of-squares programs), interval methods, or satisfiability-modulo-theories (SMT) solvers, which often require relaxations, are conservative, or scale poorly with model dimension and nonlinearity. These limitations have also shaped benchmark design: standard test suites are dominated by low-dimensional polynomial systems chosen to suit what solvers can handle rather than what engineering applications require. We address both sides of this challenge. First, we introduce DynaCertBench, a benchmark of 105 instances: 80 non-redundant systems curated from the literature, extended with high-dimensional and control-constrained variants and new nonlinearly coupled systems, comprehensively spanning stability, region-of-attraction, forward-invariance, and safety tasks. Second, we present LLM2Cert, an LLM-driven framework that synthesizes candidate certificates and verifies them against their proof obligations: each obligation is routed to an SMT solver or directly to the Lean Proof-Repair agent. On DynaCertBench, LLM2Cert verifies 81 of 105 instances ( 77.1% ), a 17.1 percentage point improvement over one-shot proposal without feedback. On the 81 fixed certificates, LLM2Cert verifies 22 that All-SMT leaves unresolved. Code and data are available at this https URL.
[AI-106] PhysibleMDM: Enforcing Inference-Time Physical Feasibility Constraints on Motion Diffusion Models via Noise Optimization
链接: https://arxiv.org/abs/2610.30500
作者: Maximilian Stölzle,David Müller,Agon Serifi,Ruben Grandia,Espen Knoop,Moritz Bächer
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Motion Diffusion Models (MDMs) can generate diverse and visually appealing human motion; yet, the resulting trajectories frequently violate physical constraints, limiting their downstream use in simulation and control. We present PhysibleMDM, a framework that promotes physical feasibility of MDM output by optimizing at inference time the diffusion noise without retraining the generative model. We introduce a set of surrogate feasibility constraints that can be evaluated without full trajectory simulation. These constraints combine kinematic checks (e.g., ground penetration, slip, and self-collision) with dynamic feasibility tests computed via an inverse-dynamics inner loop formulated as a cone program. We solve the resulting black-box noise optimization problem using the Separable Covariance Matrix Adaptation Evolution Strategy (Sep-CMA-ES), thereby avoiding reliance on computationally expensive and discontinuous gradients. An autoregressive scheme then chains optimized segments to synthesize long motion sequences. Across 100 motion styles on an SMPL humanoid, PhysibleMDM reduces mean ground penetration, self-collision penetration, and ground slip by 75%, 92%, and 74%, respectively, relative to unoptimized diffusion output. Across 11 sequences for a custom bipedal robot, it achieves median reductions of 89.5% in per-sequence mean acceleration-reconstruction error and 79.5% in the mean inverse-dynamics residual.
[AI-107] Can LLM s Improve Their Own Harness? A Controlled Study on ARC-AGI-3
链接: https://arxiv.org/abs/2610.30499
作者: Jens Lehmann,Sahar Vahdati
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 25 pages, 5 figures. Includes appendices and a supplementary code/data archive
Abstract:A generally intelligent system acquires new skills efficiently without requiring humans to redesign its software scaffold for every new task. Yet language models rarely act alone. A harness is a symbolic program that orchestrates calls to the model, handling perception, memory, tool use, recovery, and inference allocation on the model’s behalf. ARC-AGI-3 is an interactive benchmark in which agents must explore unfamiliar environments, figure out their goals on the fly, and keep learning as they go. On it, human-steered systems already report ceiling scores across the public games, so we know effective designs exist. We ask whether frontier models can build and improve such a harness on their own. Through a fixed coding bootstrap, each model constructs a complete task harness, acts through that harness on five development games, and revises the entire codebase twice. Every version is evaluated on twenty games that never influence adaptation. We measure task performance with Relative Human Action Efficiency (RHAE), which combines level completion with action efficiency relative to human baselines, and record dollar-denominated inference cost as a practical resource measure. Across 27 lineages, harness performance ranges from zero to near-ceiling scores. Two revisions raise mean held-out RHAE by only 2.80 points (95% interval [-1.44,7.35]). Development gains overstate held-out improvement, and reductions in inference cost can come at the expense of task performance. Human-steered systems demonstrate further headroom. These findings reveal a gap between autonomous harness construction and reliable adaptation under this two-revision protocol: even strong task performance does not ensure that subsequent revisions generalize or preserve existing capabilities.
[AI-108] From Prose to Properties: Testing Research Claims with ARATest for Agent -Native Science
链接: https://arxiv.org/abs/2610.30495
作者: Yifeng He,Jiachen Liu,Hao Chen
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 31 pages, including references and appendices. Code: this https URL
Abstract:AI agents now run experiments and report findings, so research produces claims faster than researchers can check them. Each empirical claim implies a property that must hold over its cases, but it states that property only in prose. Without an explicit property, no test can decide which evidence would falsify a claim, so validation cannot keep pace with the claims agents produce. To make claims testable, we approximate a claim’s semantics by abstracting its property skeleton, a parameterized relation that recurs across claims. We present ARATest, a property-based testing library that checks these skeletons as executable tests. A typed declaration binds a claim to a skeleton and fixes the claim’s compared quantities, cases, scope, and numerical policy. A shared case type makes every generator supply cases in the claim’s statistical unit. A three-valued semantics assigns each property one of three verdicts: holds, violated, or inconclusive. Every violation carries a replayable witness. A default generator reads the evidence that agent-native research artifacts (ARA) link to each claim, and auto-research generators dispatch an agent to run released code for fresh cases. Our DissClaimer plugin supplies that agent and runs these generators as continuous testing. Of 243 claims in 33 ARA artifacts, 54 received faithful declarations bound to evidence: 32 held, eight were violated, and 14 remained inconclusive. Two violations exposed errors in published papers that their authors confirmed. ARATest detected all 553 seeded faults without a false alarm on 19 valid controls, outperforming two LLM judges. Fresh runs exposed behaviors that reported evidence could not show. These results suggest that one fixed test can check explicit claim properties against both reported and fresh evidence. Comments: 31 pages, including references and appendices. Code: this https URL Subjects: Artificial Intelligence (cs.AI); Software Engineering (cs.SE) Cite as: arXiv:2610.30495 [cs.AI] (or arXiv:2610.30495v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2610.30495 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-109] What the Trace Shows and What the Label Says: Auditing Behavioral Detectors in Agent Evaluation
链接: https://arxiv.org/abs/2610.30491
作者: Idunnuoluwa A. Adeniji,Shilpika,Michael E. Papka
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:LLM agents are often evaluated by reducing execution traces to labels such as final-answer correctness, tool verification, and fault propagation. These labels are measurements produced by parsers, behavioral detectors, and harness rules, and they may not fully capture what the trace shows. We study this in a tool-use setting built on GSM8K, where agents call a calculator whose outputs are occasionally corrupted, giving clear ground truth for what the tool returned, what the model later used, and what the detectors reported. We audit three trace-level detectors across six production models and 30 problems per model. Final correctness, visible re-verification, and fault propagation can disagree in systematic ways. A key failure mode is silent tool-output override: the calculator returns a corrupted value, but the model later uses a different, often correct, value without calling the tool again, so the bad value stops propagating while a verification detector records no re-check. Human review of 90 traces agrees perfectly with the automated labels for final correctness (67/67), fault impact (19/19), and fault relevance (18/18), but only 67/81 for verification; all 14 disagreements fall into four recurring patterns. Execution limits also change the reported result: in an ablation across two models, one ranges from 21/30 to 27/30 correct as step and token budgets change, while the other stays between 27/30 and 29/30. These findings motivate measurement transparency: an automated label should come with the trace evidence behind it, whether the question the detector asks applied to that trace, whether the run was cut off by an execution limit, and any human corrections. Our trace viewer makes this information inspectable alongside the reported results.
[AI-110] From Single-Actor Environments to Multi-Actor Event Worlds: Can LLM s Be Event World Models?
链接: https://arxiv.org/abs/2610.30488
作者: Juntong Ni,Zewen Liu,Bohan Wang,Lige Zhang,Wei Jin
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:World models help agents assess the consequences of their actions before acting. Many real systems, however, involve multiple actors who make their own decisions, can evolve even when the agent takes no action, and are only partially observed through event logs. We formalize these systems as event worlds and ask whether large language models can predict future observations from the log and the agent’s proposed actions. We build complementary benchmarks from human Diplomacy games and real Polymarket trades. Our hierarchical framework compares prediction with and without other actors’ actions, then tests multi-step rollouts, predictions for actions never taken, and usefulness to agents. Fine-tuning and more data improve prediction when all actors’ recorded actions are given. With only the agent’s action, models struggle when other actors strongly affect the outcome, and larger models and longer histories help little. This gap also limits multi-step prediction and agent decisions, with models overestimating plan success. Yet imagined experience and initialization from world models can still improve policy learning. Anticipating other actors’ actions remains a central challenge for event world modeling.
[AI-111] Externalise the Memory: A Pre-Registered Study of Where an Embodied Agents Experience Should Live NEURIPS2026
链接: https://arxiv.org/abs/2610.30478
作者: Cheng-Yin Tu
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 28 pages. A 5-page version appears at the CL4FMAgents workshop, NeurIPS 2026 (non-archival). Code and pre-registration: this https URL
Abstract:Should an embodied agent’s experience be baked into its policy weights, or kept external and retrieved at inference? A memory-augmented teacher (a VLM that retrieves lessons, with their episode images, from its own past attempts) produces trajectories on a vision-conditioned reach task; from the same trajectories we distil LoRA student policies and evaluate them under matched protocols, including one that re-attaches the teacher’s frozen retrieval buffer to a fixed set of student weights. We pre-registered the hypotheses, arms, tests, and analysis rules before training and before any p-value. Distilling the memory-augmented behaviour into weights beat an action-only control by +1 pp (n.s.); the registered test proves underpowered for its own minimum effect, so we withdraw - rather than refute - the memory-in-weights claim. Attaching retrieval to a fixed set of student weights added +34 pp (z=5.15), reaching the pooled memory-augmented teacher reference (49.3%); the pre-registered protocol contrast agrees (+23 pp, z=3.36); under a pre-registered auto-reset sensitivity count the two are +36 and +29 pp. An exploratory behavioural probe with pre-registered per-arm predictions is consistent with behavioural distillation transferring the marginal correction but not the conditional, situation-specific correction that retrieval supplies. The recipe that follows is to externalise the memory: in follow-up controls registered before they ran, the same frozen buffer lifts the base model with no adapter to 48% (sensitivity 48%) and the other three adapters from 19-26% to 53-55% (sensitivity 17-21% to 50-55%), and the adapter adds nothing detectable once the buffer is attached. On a second, harder task, the retrieval advantage points the same direction but falls short of significance (+6 pp, 95% CI [-4.5, +16.4]); exploratory diagnostics tie the attenuation to memory coverage.
[AI-112] Short-to-Long: Detached Features for Long-Horizon Value Learning
链接: https://arxiv.org/abs/2610.30461
作者: Tong Che,Boris Ivanovic,Marco Pavone
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Learning useful representations from continually changing value targets remains a challenge in deep reinforcement learning. We study the temporal horizon of this supervision: smaller discounts make these targets less sensitive to distant bootstrap changes, but using them for behavior changes the control objective. We introduce Short-to-Long (S2L), which supplies a long-horizon behavior learner with features trained by smaller-discount source networks. Detached connections prevent the behavior value loss from updating the sources, allowing each source to follow its own value objective while the behavior discount remains unchanged. Across Atari, online continuous control, and offline continuous control, the approach yields aggregate performance gains, improving the paired terminal median on 40 of 57 Atari games. Matched Atari ablations support the roles of feature transfer, smaller source discounts, and gradient isolation. Fixed-trajectory measurements further show reduced drift in source value targets with mixed discounts.
[AI-113] aggmetrics: Aggregating Classification Performance Measures for Machine Learning
链接: https://arxiv.org/abs/2610.30456
作者: Ronaldo Cristiano Prati
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 5 pages (4 + references). Software: this https URL , documentation: this https URL , archived at doi: https://doi.org/10.5281/zenodo.23535051
Abstract:aggmetrics is a Python package that extends any binary performance measure expressed in the four confusion-matrix counts to multiclass, multilabel, ordinal, cost-sensitive, soft-label and cross-validation settings, by substituting one of four aggregation operators for what libraries usually expose as an average switch. The package also stores, for every measure, the structural properties that decide which identities apply to it, can infer those properties from a user-supplied formula, and uses them to warn when a measure is used where it does not behave as commonly assumed. Its test suite encodes the results of the companion theory as executable tests. The package is BSD-3 licensed, depends only on NumPy, and is available at this https URL documentation at this https URL.
[AI-114] Same Beliefs Different Policies: Learning Reusable Representations for Early Stopping in LLM Reasoning
链接: https://arxiv.org/abs/2610.30450
作者: Jingwei Ni,Dmitry Abulkhanov,Vladislav Smirnov,Markus Leippold,Elliott Ash,Timothy Baldwin,Artem Shelmanov,Mrinmaya Sachan
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Under Review
Abstract:Reasoning LLMs often continue reasoning after reaching a correct answer, wasting computation and sometimes replacing it with an incorrect one. Deciding when to stop is a metareasoning problem: the model must weigh what further reasoning could achieve against its cost, given what it can already answer. Direct utility training learns a stopping representation at a chosen token price; supporting new prices can require additional controller training. We introduce BELT (BELief-based Termination), which separates outcome prediction from stopping control. It learns beliefs about current and future answers from a frozen LLM’s hidden states, then fits a linear stopping policy on the frozen features. The prediction targets are independent of token price, so one probe supports different accuracy-cost trade-offs through CPU policy refitting or threshold adjustment. Averaged over five out-of-distribution (OOD) math benchmarks, BELT reduces generated tokens by 41% on Qwen3-1.7B and 51% on OLMo-3-7B-Think without loss of accuracy. The same math-trained BELT transfers to coding and agentic tasks without training on code or agentic data. In all three domains, its accuracy-cost trade-offs are competitive with or better than those of end-to-end controllers. On OLMo3, its probe has 21.2x fewer parameters than one checkpoint of the strongest baseline, OS-Pruner, and needs 55x fewer training GPU-hours than OS-Pruner’s five price-specific checkpoints.
[AI-115] A Three-Pillar Governance Framework for Safe and Trustworthy AI Agents
链接: https://arxiv.org/abs/2610.30445
作者: Edward C. Cheng,Jeshua Cheng,Alice Siu
类目: Artificial Intelligence (cs.AI)
备注: 7 pages, 3 figures, accepted paper at the IEEE International Conference on Agentic, Generative, and Cognitive AI Systems (AGCS 2026)
Abstract:AI agents capable of autonomous decision-making introduce significant challenges related to safety, alignment, accountability, and human oversight. Existing Human-in-the-Loop (HITL) and responsible AI approaches provide important foundations, but many remain fragmented across system design, governance, and operational deployment. This paper presents a Three-Pillar (3P) model and framework for developing safe and trustworthy AI agents through transparency, accountability, and trustworthiness. The framework defines an evolutionary path in which increasing agent autonomy is introduced through progressive validation, structured human oversight, and auditable operational controls. We further describe how the model supports agent lifecycle monitoring, decision traceability, escalation management, and human-agent collaboration within enterprise agentic environments. By integrating oversight principles with operational mechanisms, the proposed framework provides a practical foundation for deploying AI agents that remain aligned with human intent while enabling controlled progression toward higher autonomy.
[AI-116] SpecMAGIC: Specification-Guided Multi-Task Generalization in Multi-Agent Reinforcement Learning
链接: https://arxiv.org/abs/2610.30443
作者: Zijian Guo,İlker Işık,H. M. Sabbir Ahmad,Wenchao Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Multi-task generalization is an important yet underexplored problem in multi-agent reinforcement learning (MARL). Many real-world applications require multi-agent systems to generalize across complex, temporally extended tasks with diverse inter-agent coordination requirements. We introduce a specification language that combines temporal-logic task specifications with operators for expressing such requirements. We propose SpecMAGIC, a hierarchical framework that leverages the automaton structure of task specifications for subtask decomposition. A centralized high-level planner identifies and allocates compatible subtask sequences across agents, while decentralized low-level policies execute individual subtasks. This enables zero-shot generalization to new tasks by composing reusable subtask policies according to new specifications, without retraining. We establish soundness and probabilistic completeness of the high-level planner. Experiments on multi-agent navigation tasks show that SpecMAGIC substantially outperforms existing approaches on unseen task specifications and maintains strong performance as the number of agents increases.
[AI-117] Secure Agent ic AI and Multimodal Intelligence in 6G Communication Systems
链接: https://arxiv.org/abs/2610.30436
作者: Satyam Gaba,Devanshu Motwani,Siva Sai,Vineet Suri,Vikas Ranjan,Vinay Chamola
类目: Artificial Intelligence (cs.AI); Signal Processing (eess.SP)
备注: 20 pages, 5 figures, 5 tables
Abstract:Sixth-generation (6G) communication systems are expected to support immersive, intent-driven, and latency-critical services through tight integration of multimodal intelligence, edge computing, and autonomous network control. In this setting, agentic AI systems can reason over heterogeneous inputs, use tools and network functions, coordinate across distributed entities, and adapt to dynamic environments. These capabilities make agentic multimodal intelligence promising for 6G applications such as semantic communications, digital twins, and zero-touch network management, but they also introduce security and privacy risks that extend beyond conventional AI and network security models. Vulnerabilities can propagate across multimodal data pipelines, memory, tools, communication interfaces, and inter-agent coordination loops. We survey multimodal intelligence and agentic AI for 6G cross-layered communication systems from the perspectives of privacy and security. We present a 6G-oriented system view spanning perception, memory, tool use, and multi-agent orchestration; analyze threat models and attack surfaces across sensing, cognition, action, memory, and communication layers; organize existing attacks into a structured taxonomy; and review mitigation strategies including prompt and memory defenses, inter-agent trust mechanisms, secure orchestration, and privacy-preserving learning. We conclude by outlining open research directions for trustworthy multimodal multi-agent 6G systems.
[AI-118] RoleFlow: Self-Evolving Multi-Role Agent ic Collaboration via Structural Adaptation
链接: https://arxiv.org/abs/2610.30429
作者: Yi Ge,Pan Wang,Xuehao Cui,Zian Zeng,Xu Hu,Heng Huang,Furong Huang
类目: Artificial Intelligence (cs.AI)
备注: 21 pages
Abstract:Large language model (LLM) based multi-agent systems often rely on predefined collaboration structures, limiting their adaptability to evolving task requirements. Although recent approaches enable dynamic orchestration, trajectory-level rewards they apply assign one outcome to each decision, ofen struggling to identify the contributions of individual structural changes. We propose RoleFlow, a self-evolving multi-role agentic system that refines collaboration structures through feedback-driven atomic edits and on-demand role synthesis. To make structural adaptation learnable, we introduce Single-Edit Counterfactual Policy Optimization (SECPO), which evaluates each edit against its no-edit counterpart at the same chain state and assigns targeted credit to edit-specific tokens. Across seven benchmarks spanning four domains, RoleFlow achieves the optimal average performance, surpassing the strongest competitor by 4.7 points. It leads on all three multi-hop QA benchmarks by 3.2-7.5 F1 points and also shows supremacy on mathematical reasoning, web interaction, and code generation. Further analyses demonstrate cross-benchmark transferability, training-data efficiency, and cross-executor generalization.
[AI-119] AcoustiPrint: Physics-Guided Neural Design of Manufacturable Acoustic Structures
链接: https://arxiv.org/abs/2610.30411
作者: Shanmuka Sadhu,Harshvardhan C. Takawale,Yehia Zakaria,Ahmed Allam,Nirupam Roy
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent work in additive manufacturing and neural inverse design enables the creation of complex three-dimensional structures to manipulate physical fields, yet these designs must also adhere to discrete, non-differentiable fabrication constraints. Current methods limit the design space by imposing differentiable soft manufacturability loss penalties, or by fixing structures through post-processing, causing a discrepancy between the geometry optimized by the physics model and the final fabricated result. Thus, we present AcoustiPrint, a physics-informed neural architecture for constrained inverse design that represents three-dimensional geometries with an implicit neural representation (INR) and passes each candidate through a printability operator before evaluating its final response. The printability operator executes layer-wise self-support, discrete material selection, and whole-body connectivity, assuring fabrication feasibility. To allow for backpropagation through the discrete non-differentiable operations, we use surrogate gradients for the backward pass and a self-supervised printability objective that progressively drives the INR toward a valid design domain. We demonstrate AcoustiPrint for personalized transcranial focused ultrasound, where inter-patient differences in acoustic properties and head geometry deform wave propagation and requires patient-specific acoustic focusing. AcoustiPrint jointly optimizes three-dimensional geometry and material composition through differentiable wave physics while ensuring manufacturability. Experiments show manufacturable acoustic structures with enhanced focal accuracy and over density-based and phase-based baselines, demonstrating a generalized methodology for neural design through non-differentiable manufacturing constraints and differentiable physics.
[AI-120] Soft Ontology Priors over Latent Concept Posteriors for Weakly Supervised Neuro-Symbolic Learning
链接: https://arxiv.org/abs/2610.30391
作者: Tengwei Song,Robert Hoehndorf
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: NeSy 2026 oral
Abstract:Neuro-symbolic (NeSy) predictors that supervise only the task label suffer from reasoning shortcuts: distinct concept mappings achieve the same maximum likelihood, leaving the recovered concepts unidentifiable and unsafe for downstream reuse. We attack this failure by injecting a soft ontology prior over the latent concepts into the training objective and compiling it together with the task constraint into a smooth, decomposable probabilistic circuit. Under independent Bernoulli perception, the normalisers of the resulting structured concept posterior reduce to weighted model counts that are linear in the compiled circuit size, because the neural leaf weights fold into the circuit before evaluation. The soft compilation strategy places expert axioms in the circuit’s weights rather than its support (task-structural axioms guaranteed by the data-generating process stay hard), so noisy expert axioms only redistribute prior mass and cannot empty the support. We characterise how the prior reweights label-restricted shortcut mass through differences in ontology violation energy, show that soft priors incur bounded log-density perturbations under bounded axiom noise, and prove that the hard-support counterpart can be made degenerate by a single corrupted axiom. We validate these predictions on an audited six-concept view of the real-world BDD-OIA driving benchmark and on controlled NeSy tasks.
[AI-121] Route-and-Generate: KL-Regularized Router Improvement for Reasoning Routing in LLM s
链接: https://arxiv.org/abs/2610.30351
作者: Tien-Phat Nguyen,Truong Nguyen,Thanh-Thien Le,Hoang Phan,Thien Huu Nguyen,Linh Ngo Van,Trung Le
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Large language models often apply complex reasoning indiscriminately, leading to overthinking on simple tasks and underthinking on harder ones. We present Route-and-Generate (Rouge), a hierarchical training framework that enables LLMs to adaptively decide when and how much reasoning is required for each input. Our approach decouples reasoning effort allocation from answer generation by introducing a routing policy that allocates a fixed sampling budget across multiple reasoning formats, while the base model is optimized with GRPO using per-sample rewards. To promote efficient reasoning, we employ a rank-based reward that favors shorter correct solutions, and update the routing policy via a closed-form KL-regularized policy improvement step distilled into a neural router. This design encourages concise answers on easy problems while reserving longer reasoning for harder ones, achieving improved accuracy-efficiency trade-offs across multiple reasoning benchmarks.
[AI-122] MPBRS: Multi-Potential-Based Reward Shaping for Unified Reward Configuration in Humanoid Locomotion
链接: https://arxiv.org/abs/2610.30334
作者: Van Chung Nguyen,Hung Manh La
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Designing and tuning reward functions remains one of the main challenges in building effective reinforcement learning (RL) pipelines for humanoid locomotion. While shaping rewards can accelerate learning, naively formulated terms often conflict with the desired behavior, overfit to a specific robot, or become unstable under small weight changes. Potential-Based Reward Shaping (PBRS) offers an alternative solution: shaping the reward via differences of a potential function to guide exploration without changing the optimal policy. In this paper, we study Multi-PBRS for humanoid locomotion and propose a unified reward configuration. We then evaluate the same reward structure and weights across nine humanoid robots, requiring only robot-specific metadata (joint/link mappings and default poses) rather than per-robot reward redesign. Our experiments show that, compared to standard shaping, MPBRS provides more consistent scaling behavior and reduced tuning sensitivity across robots, making it easier to tune without altering the optimal policy.
[AI-123] Gender Sexual Orientation and the One-Year Refugee Bar: A Computational Study of Bill C-12 and Canadian Refugee Adjudication
链接: https://arxiv.org/abs/2610.30297
作者: Sean Rehaag
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注:
Abstract:In March 2026, Bill C-12 amended the Immigration and Refugee Protection Act to make refugee claims ineligible for referral to the Refugee Protection Division where they are made more than one year after the claimant’s first entry to Canada. Affected claimants are diverted to a Pre-Removal Risk Assessment (PRRA), which is typically decided by an immigration officer on a paper record, with no guaranteed oral hearing, no appeal on the merits, and no automatic stay of removal pending judicial review. The government defended the bar partly on the premise that genuine refugees claim promptly. Critics countered that late claims are concentrated among vulnerable groups, particularly people fleeing persecution based on gender, sexual orientation, gender identity, or gender expression (SOGIE). This article tests those competing premises. Using artificial intelligence tools, checked through human review, it examines nearly 29,000 published decisions of the Refugee Protection Division, the Refugee Appeal Division, and the Federal Court from 2006 to 2025. The results of this study demonstrate that, first, while most claimants claim within a year, roughly one in nine in the published record does not, and gender and SOGIE claims are significantly overrepresented among these late claims. Second, among late claims, refusals of gender and SOGIE claims are more likely than refusals of other claims to be overturned on appeal or judicial review. Late claims on other grounds, by contrast, fare somewhat worse on review than timely claims. The one-year bar therefore removes the hearing, the appeal, and the stay from the late claimants for whom those safeguards have been doing the most work. The article recommends repealing the bar or, failing that, adopting exemptions, procedural protections, and greater transparency in PRRA decision making. Subjects: Computers and Society (cs.CY); Artificial Intelligence (cs.AI) Cite as: arXiv:2610.30297 [cs.CY] (or arXiv:2610.30297v1 [cs.CY] for this version) https://doi.org/10.48550/arXiv.2610.30297 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-124] GAV: A scientific experimentation environment for PDE agents
链接: https://arxiv.org/abs/2610.30292
作者: Muheng Li,Wenlong Mou
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 52 pages, 19 figures, 14 tables
Abstract:Reliable feedback is essential for coding agents to develop programs for partial differential equations (PDEs). Yet evaluating these programs on new PDE problems often requires expensive reference computations, while fixed datasets restrict the range of experiments. We introduce Generator-Agent-Verifier (GAV), a scientific experimentation environment that makes new PDE problems verifiable by construction. The Generator learns a distribution of continuous reference solutions using data from numerical simulations or experimental measurements to represent a target physical regime. Using the method of manufactured solutions, it works backward from the generated reference solutions to construct matching PDE problems without a separate numerical solve. The Agent designs experiments, investigates failures, and iteratively improves its program. The Verifier evaluates candidate programs against the reference solutions and provides numerical feedback on accuracy and computational cost. This feedback can also guide the generation of new problems that probe the current program’s weaknesses. We provide a theoretical analysis of evaluation reliability and guided development. Experiments on Navier-Stokes and Vlasov-Poisson-type equations across different data sources and regimes demonstrate the effectiveness of GAV. In particular, on the numerical Navier-Stokes task, GAV reduces the mean test error of three existing PDE agents by more than an order of magnitude relative to their original setups. Code is available at this https URL.
[AI-125] Generalizing Metropolis-Hastings for DAG Sampling
链接: https://arxiv.org/abs/2610.30291
作者: Daniele Nikzad,Mikko Koivisto
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation (stat.CO)
备注:
Abstract:A fundamental challenge in Bayesian network structure learning lies in efficiently approximating the posterior distribution of directed acyclic graphs (DAGs). For large, high-dimensional datasets, the task remains beyond the capabilities of current methods. This work advances Markov chain Monte Carlo algorithms by introducing novel moves that simultaneously resample the parents for multiple nodes of the DAG. At the core of these moves is duet sampling, a new generalization of the Metropolis-Hastings algorithm that selects two proposal distributions from a pool at each iteration. One is used for generating a candidate state of the Markov chain and another for evaluating the return to the current state; both are ingredients of an appropriately crafted acceptance criterion. Pairing different proposal distributions provides additional freedom and enables constructing Markov chains that mix more efficiently than their basic Metropolis-Hastings counterparts, at no additional computational cost. Experiments show that our new moves, both individually and in combination, outperform the previous state of the art, which relies on a more restrictive pool of moves. This advantage persists even when letting the previous methods leverage duet sampling to enhance their performance.
[AI-126] SWE-VERIFIER: Selective Verification for Cost-Efficient Coding Agents
链接: https://arxiv.org/abs/2610.30284
作者: Kaijian Zou,Vaibhav Chennareddy,Yunxiang Zhang,Naihao Deng,Longju Bai,Jiaxin Pei,Rada Mihalcea,Lu Wang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Coding agents spend substantial computation generating actions and processing long interaction histories. Poor action choices can steer agents into low-value states, wasting computation and increasing API costs. Sampling and verifying alternative actions can improve the agent’s decisions, but applying this procedure at every step adds substantial overhead. We introduce SWE-Verifier, a lightweight generative model that selectively triggers candidate sampling and verification. It scores the coding agent’s proposed action to determine when additional candidate sampling is needed, then ranks the resulting actions before execution. We train SWE-Verifier from Qwen3.5-4B using preference data that groups contrasting actions with their resulting environment observations and rubric-guided teacher scores. Using the mini-swe-agent harness, SWE-Verifier improves resolution from 75.47% to 77.20% on MiniMax-M2.7 at 1.09 times cost and generalizes to DeepSeek-V4-Flash without training on its actions, improving resolution from 72.60% to 74.27% at 1.13 times cost. Across both models, SWE-Verifier verifies fewer than 4% of steps and achieves comparable or higher resolution than the trajectory and all-step baselines at substantially lower cost. These results demonstrate that selective candidate sampling and lightweight verification provide complementary gains in coding-agent performance and inference efficiency. Our model, code, and data are available at this https URL.
[AI-127] Breaking the high-degree barrier in Lehmers problem with a human-AI discovery loop
链接: https://arxiv.org/abs/2610.30277
作者: Yuri Chervonyi,Sergei Gukov,Thang Luong
类目: Artificial Intelligence (cs.AI)
备注: 36 pages, 6 figures, 7 tables
Abstract:While artificial intelligence has made rapid progress in mathematical reasoning, many open problems depend on structure that is revealed only through large-scale computational exploration and analytical synthesis. One such example is Lehmer’s problem from 1933. It asks whether the Mahler measure of an integer polynomial, which measures how far its roots stray from the unit circle, can be made arbitrarily small. Previously, constructing irreducible polynomials with integer coefficients and small Mahler measure has been carried out systematically only for even degrees d \le 180 . At higher degrees, Boyd’s analytical one-parameter family frequently factors, skipping nearly 50% of target degrees and leaving no general construction for arbitrary d . Here, we address the high-degree regime through a human-AI discovery loop, in which we extend the conversational prompting with a fast numerical pipeline, an exact symbolic verifier, and large-scale distributed search. The AI agent prototypes numerical and symbolic solvers locally, launches parallel searches across distributed clusters and works with the human researcher to extract and verify mathematical hypotheses. Through this interactive loop, we derive a five-parameter construction, which we call the Generalized Boyd algorithm. It eliminates degree-skipping and constructs irreducible polynomials with Mahler measure strictly below Boyd’s limit point \alpha_1 \approx 1.255434 for every target degree. Evaluated across all even degrees up to 2000, our approach improves upon six previously published values and holds 99.6% of all best known records. With this work we release the verified dataset of polynomials with minimal Mahler measure across all studied even degrees.
[AI-128] Edit Dont Imitate: Minimal-Edit On-Policy Distillation for Multimodal Reasoning
链接: https://arxiv.org/abs/2610.30231
作者: Wen Ye,Ailin Deng,Zirui Wang,Shixing Yu,Han Zhang,Tian Yun,Nikhil Mehta,Tarun Kalluri,Xuefei Cao,Xuezhe Ma,Hengduo Li
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Visual mathematical reasoning can fail before reasoning begins: when a vision-language model (VLM) misreads a diagram, subsequent reasoning proceeds from false premises. This perceptual bottleneck is especially pronounced in smaller models, while accurate textual descriptions can substantially narrow the gap to larger models. Because obtaining such descriptions at inference requires additional annotation or model calls, we propose minimal-edit corrective on-policy distillation. Using reference figure descriptions and verified reference solutions as training-only information, a teacher scores incorrect student rollouts under a minimal-edit objective, and only teacher-disfavored tokens contribute to the corrective loss. Correct rollouts receive task-reward reinforcement learning. This preserves on-policy supervision while avoiding full-response imitation. Across model scales, the method requires no descriptions at test time and meets or exceeds the corresponding base model supplied with a reference figure description at inference. It also improves perception-oriented performance, particularly low-level geometric perception, and transfers to external mathematics and perception benchmarks. Because the minimal-edit teacher evaluates corrections rather than generating its own solution, the method remains effective with a frozen same-size teacher, although its gains are smaller than with a larger teacher on most metrics.
[AI-129] Automated screening of privacy risks in medical AI
链接: https://arxiv.org/abs/2610.30229
作者: Konstantinos Vrettos,Eleftherios Tzanis,Haridimos Kondylakis,Evangelia E. Vassalou,Florence X. Doo,Michail E. Klontzas
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注: 37 pages, 5 figures
Abstract:Medical artificial intelligence (AI) models are often shared without their training data, assuming this protects the patients who contributed them. However, a model can reveal whether a patient’s records were used in training, a privacy breach known as a membership inference attack. This risk is usually reported as a cohort average, which can appear negligible while individual patients are almost certainly exposed. Identifying potentially vulnerable records before developing a production model could make privacy assessment more practical. We measured each patient’s vulnerability to membership inference attack, across 7,789 models trained on six clinical datasets in four modalities, including two adapted pretrained medical models. Training choices changed how many patients had high attack scores, but the patients who ranked highest by risk largely stayed the same across the configurations we tested. Stronger weight decay reduced the exposed cohort up to threefold, yet largely the same patients remained most at risk across training recipes, architectures and initializations, including foundation models. Therefore, we developed a lightweight screening tool that uses four small models to identify patients with high risk of exposure before production-model training. We also assessed if flagged patients could be protected without removing their data. Privacy screening should become incorporated into data preparation for clinical AI.
[AI-130] When Reconstruction Is Not Enough: Doubly Robust Counterfactual and Interventional Mediation Analysis with Learned Multimodal Representations
链接: https://arxiv.org/abs/2610.30220
作者: Cong Cao
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Methodology (stat.ME); Machine Learning (stat.ML)
备注: 22 pages, 1 table
Abstract:Learned representations of high-dimensional multimodal health records are increasingly used to adjust for confounding, yet they are usually trained to reconstruct the data rather than preserve the information that causal identification requires. We study counterfactual outcomes under alternative treatments, including phenotype-specific effects and patient-level counterfactual risk profiles of the kind digital twins provide, and interventional direct and indirect effects under exposure-induced intermediate confounding. We give conditions under which a representation preserves identification and show that interventional mediation is more demanding than counterfactual comparison. We prove that no covariate-only training objective, whether reconstruction, contrastive, or physics-informed, can certify that a representation is sufficient, and that, in a linear structural model, a representation can lose arbitrarily little reconstruction accuracy yet induce arbitrarily large bias. We decompose the error of cross-fitted doubly robust estimators into a first-order representation bias, which neither double robustness nor cross-fitting removes, and a second-order remainder of nuisance-estimation errors, so coverage worsens as the sample size grows when the representation is insufficient. We propose a workflow combining modality-specific variational autoencoders trained outside the estimation sample, cross-fitted doubly robust estimators, locked phenotypes, select-then-confirm pathway analysis, portable representations, and diagnostics for representation insufficiency, exposure leakage, and sensitivity. In simulations, discarding a rare, moderately strong confounder gave indirect-effect intervals with 2% coverage at n = 3,000; reconstruction loss missed the problem, and a simple predictive diagnostic flagged it in every replicate.
[AI-131] Prediction Is Not Calibration: Item Parameter Recovery and Representation Coupling in Knowledge Tracing
链接: https://arxiv.org/abs/2610.30211
作者: Wenrui Yuan,Stéphanie M. van den Berg,Johannes Steinrücke
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 46 pages, including appendix
Abstract:Knowledge tracing (KT) models can estimate item parameters through an item response decoder trained solely for next-response prediction, yet predictive accuracy does not guarantee parameter recovery. We study 2PL, GPCM, and NRM decoders paired with LSTM, Transformer, and DKVMN encoders. Response-model derivatives show why discrimination and option slope updates depend directly on the scale and spread of estimated ability. Affine logit identities show how item-dependent shifts and rescaling in estimated ability are absorbed into item parameters without changing response probabilities. In simulations, difficulty, mean threshold, and correct-option intercept recovery remain high under shuffled order across representation widths, whereas discrimination and keyed option slope contrast recovery require wider item representations and degrade further under fixed item order. Separating the sequence item representation (width d_S ) from the item parameter key (width d_M ), and pairing a wide key with a thin sequence representation ( \mathrmSK(64,8) ), reduces fixed-order ability distortions and improves recovery and cross-order consistency without sacrificing test accuracy or increasing model size relative to a shared width-64 baseline ( \mathrmSH(64) ). On TIMSS and EdNet, widening d_M consistently improves agreement with an offline IRT reference, while the trade-off from narrowing d_S depends on the encoder’s reliance on sequence width for prediction. These findings show why item parameters, especially discrimination and option slopes, must be evaluated separately from response prediction.
[AI-132] MoEZip: Routing-Aware KV Cache Compression for Sparse Mixture-of-Experts LLM s NEURIPS
链接: https://arxiv.org/abs/2610.30206
作者: Minsang Kim,Seungjun Baek
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Annual Conference on Neural Information Processing Systems (NeurIPS) 2026, Accepted
Abstract:Mixture-of-Experts (MoE) language models scale model capacity efficiently by activating a small subset of parameters per token. However, in long-context inference, memory remains a major bottleneck because the KV cache grows linearly with sequence length. Existing KV cache compression methods reduce this cost for dense transformers, but overlook a problem specific to MoE: KV eviction can change which experts are selected during answer generation, degrading the generation quality. We propose MoEZip, a routing-aware KV cache compression for sparse MoE language models. To estimate the sensitivity of expert routing to perturbation, we propose a novel metric based on the Fisher information matrix of routing distributions. MoEZip combines this metric with attention weights for routing-aware scoring of contexts. Finally, we compose a complementary scoring system that combines eviction methods for dense and sparse-MoE architectures. Experiments show that MoEZip achieves robust performance under aggressive cache compression on various benchmarks. At a KV retention ratio of 10%, MoEZip improves over the SoTA baseline by +24 and +20 points on average on LongBench and RULER, corresponding to 2.3 \times and 4.9 \times higher scores, respectively.
[AI-133] Feedback That Backfires in Games: Do LLM Players Repeat the Move They Just Watched Fail?
链接: https://arxiv.org/abs/2610.30205
作者: Esmail Gumaan
类目: Artificial Intelligence (cs.AI)
备注: 20 pages, 6 figures. Code, probe items and raw records: this https URL
Abstract:When an LLM agent’s move fails, the harness writes the move and the failure message into the transcript and asks again. In tool calling, this record makes small models more likely to repeat the failed action. We test whether the same holds in games (Sokoban, Connect Four, Othello), where a move can fail by being illegal or by being legal but losing, both verified exactly. On fixed probe items built by perturbing a correct move, the record backfires: pooled over 8 instruction-tuned models (135M-1.7B) it makes the failed move more likely (corrective gain G = -1.38 nats per token), through the move’s surface form, for illegal moves but not for losing ones, and not for every model. An exploratory analysis shows what sets the sign: after the record, the failed move’s log-probability barely depends on how likely it was before (slope 0.18). Moves the model would not have played become more likely, and greedy decoding starts to produce them; moves it was about to play become less likely, but greedy decoding still repeats 38% of them. A description of the failure that omits the move string keeps the model’s preference instead. Along the OLMo 2 1B post-training ladder the inversion is present in the base model and largely removed at the DPO stage. RAGEN’s default models invert in RAGEN’s own Sokoban format before any RL. A “do not repeat” instruction does nothing, a clean retry is the worst harness, and a decoder ban stops exact repeats but not illegal play. All measurements ran in float32 on a free Colab GPU, which first reproduced the original paper’s number.
[AI-134] MPIDetector: Detecting Multi-modal Prompt Injection via LLM -based Reasoning
链接: https://arxiv.org/abs/2610.30202
作者: Yiheng Li,Jingyue Li
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Accepted by the 8th IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (IEEE TPS 2026)
Abstract:With the rapid integration of Multi-modal Large Language Models (MLLMs) into real-world applications, Prompt Injection (PI) has emerged as a critical security threat. Attackers can conceal malicious instructions within textual narratives or visual inputs to bypass safety alignment. Existing approaches, such as signature- and perplexity-based methods, rely heavily on pattern matching, making them insufficient to detect sophisticated, cross-modal PI payloads. In this paper, we propose a novel detection framework, called MPIDetector, that leverages LLMs’ intrinsic reasoning capabilities via few-shot In-Context Learning (ICL). Instead of fine-tuning, we provide the model with specific prompt templates and representative attack examples, enabling it to scrutinize the semantic intent of input across modalities. We constructed a comprehensive evaluation benchmark comprising eight textual and seven visual malicious categories, supplemented with benign datasets to rigorously test for false positives. Our evaluation across multiple leading LLMs, including Gemini-2.0-Flash, Qwen3-VL-32B-Instruct, and Doubao-seed-1.6-flash, demonstrates the efficacy and generalizability of this approach. The proposed method achieves up to 90.91% accuracy and 100% recall on textual tasks. In the visual domain, it achieves up to 99.28% accuracy while maintaining 100% precision across all evaluated models. Crucially, when evaluated against 500 highly obfuscated environmental web-agent injections (WebInject), MPIDetector achieved a detection rate up to 99.80%. Comparisons with two baseline approaches for detecting textual PI payloads, along with two additional methods for visual input detection, demonstrate MPIDetector’s superior accuracy. These results show that LLM-driven reasoning provides a robust, cross-modal defense against advanced prompt injections.
[AI-135] Detection Is Not Enough: Reasoning Traces as Testable Program Hypotheses
链接: https://arxiv.org/abs/2610.30200
作者: Vaidehi Hegde
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 17 pages, 8 figures, 9 tables. Code and stored outputs: this https URL
Abstract:When a language model is deployed, no reference solution exists, so the reasoning trace it writes is the only evidence for judging its answer. The reference-free substitutes each fail differently: self-assessment degrades the answer it was meant to protect, sampling agreement measures stability rather than correctness, and perturbation studies read the expected answer off a human-written template, grading a model rather than a trace. We ask how far a single trace can be verified with no reference solution, no judge model and no weight update, and find, for a frozen LLaMA-2-7B-chat on GSM8K, that detection is not what limits such a pipeline. Checking a trace against itself reaches an AUROC of 0.741 and stops there for a structural reason: two thirds of wrong traces state no false equation, having executed a mistaken plan correctly, and on wrong-plan failures the detector flags a wrong trace less often than a right one. A candidate-derived metamorphic probe exposes many of these failures: we abstract the trace into a program over the question’s quantities, perturb one quantity, execute the program, and ask the model the rewritten question once, so every trace supplies its own test. Valid perturbations exist for 937 of 1,319 test questions (71.0%); among the 847 that also have a parseable first trace, supported traces are correct 60.6% of the time against a 31.6% base rate, although the probe also refutes 45.5% of correct traces. No tested policy yields a statistically significant improvement in accuracy. A flag earns its retry only when detector precision exceeds D/(R+D), where R and D are the rates at which a retry repairs a wrong trace and damages a correct one; our repair rate is so low that, at the same retry budget, even a perfect detector would recover an estimated 66 of 1,319 questions. Repair, not detection, is the open problem.
[AI-136] Agent ic Verified Optimization of Real-World Software
链接: https://arxiv.org/abs/2610.30188
作者: Zhe Ye,Udaya Ghai,Zhizhen Qin,Dawn Song,Soonho Kong
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Logic in Computer Science (cs.LO); Programming Languages (cs.PL)
备注:
Abstract:Program optimization aims to improve performance while preserving behavior. Coding agents can already make programs faster by changing algorithms, data structures, and execution strategies. However, existing equivalence checks mostly rely on tests, which agents have been shown to exploit, passing the tests without preserving the intended behavior. We propose agentic verified optimization, in which an agent produces a faster implementation together with a machine-checked proof of equivalence to the existing one. Because the existing implementation serves as the specification, the task avoids the specification-writing bottleneck of verified code generation, and recent advances in theorem proving make the proofs attainable. We formulate the task at the function and repository levels in Lean 4, where programs can be both executed and reasoned about. Few real-world applications are written in Lean, so we develop Aqueduct, an agentic framework that migrates C, C++, Go, and Rust programs into native Lean, validates each port by differential testing against its original. On the 189 reference implementations in Verina, agents produce faster proved rewrites for 156, including algorithmic changes such as replacing exponential recursion with dynamic programming. On 17 migrated applications, their 92 proved rewrites speed up 47 of 51 application scenarios. These application-level gains are smaller than the function-level ones, making the choice of functions to optimize an important problem. Current agents can thus already carry out verified optimization on both standalone functions and applications migrated from real programs. We discuss the remaining challenges and the steps toward verified optimization in production.
[AI-137] Self-Evolving Static Analysis via LLM Pseudo-Labeling
链接: https://arxiv.org/abs/2610.30184
作者: Rui Melo,Andre Duarte,Aditya Oke,Shubham Gandhi,Nachiket Kotalwar,Rui Abreu,Lei Li,Carolyn Rose
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Security bugs often slip past static analysis not because they are impossible to detect, but because detection rules have not yet been written for the code in front of us. New frameworks, APIs, and coding idioms continually create such gaps, leaving developers to discover failures and manually encode new rules. We present SESAA, a framework in which an Large Language Model (LLM) agent starts from no rules and iteratively evolves a set of readable and auditable joern queries. The agent receives feedback only from a second LLM critic, which inspects flagged and unflagged code to identify likely false alarms and suspected misses, without access to ground-truth labels. The agent then uses this feedback to revise its queries. On the OWASP Benchmark for Python, SESAA reaches a mean F1 of 0.742 across 8 independent runs, well above the default configurations of five widely used tools and comparable to the strongest of them once it is tuned with access to ground-truth labels that SESAA never sees. In a pilot on real-world Python projects, queries induced from code available before a vulnerability was introduced detect 10 of 19 subsequent CVEs, including 9 detected by rules written for the same repository. The resulting detector is a set of executable, human-readable queries that engineers can inspect, modify, and run independently. These results suggest that LLM agents can turn static-analysis rule maintenance from a manually authored process into an iterative process of automated discovery and repair. Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE) Cite as: arXiv:2610.30184 [cs.CR] (or arXiv:2610.30184v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2610.30184 Focus to learn more arXiv-issued DOI via DataCite
[AI-138] Detecting AI Training in Datacentre Power
链接: https://arxiv.org/abs/2610.30182
作者: Tom Kimpson,Mauricio Baker,Emlyn Graham
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (cs.LG)
备注: 15 pages, 8 figures
Abstract:AI governance agreements require methods to detect large model training runs on a monitored computing cluster, without trusting the operator. We develop a method to detect synchronous model training in facility-level power by Viterbi tracking of the wandering iteration cadence. On simulated datacentre power, when the operator makes no attempt to hide, the Viterbi tracker detects at least 96% of training runs against a declared inference workload, at a 1% false-alarm rate and within 0.04 of the Neyman–Pearson optimum, given a meter that samples at 2 Hz or faster. Of eight adversarial strategies an operator might use to hide, only extreme variation of the work per iteration evades the tracker. Its strongest form hides training in up to 73% of five-minute observations, but over a day at least 87% of such runs are still caught, at a 0.1% false-alarm rate per run. This adversarial strategy costs no throughput, but in small-scale training the operator needs 93 – 267% more compute to reach the same loss. A modest passive meter is sufficient to make hidden training both detectable and expensive, which makes power metering a promising part of compute verification that does not depend on trusting the operator.
[AI-139] When Stable and Preferred Semantics Agree: A Path-Based Approach
链接: https://arxiv.org/abs/2610.30176
作者: Yan Zhang,Kehan Liu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We investigate structural conditions under which preferred and stable semantics agree in abstract argumentation frameworks. Our approach is based on suitably chosen sets of paths in the attack graph and binary relations induced by them. We establish a general path-based criterion for semantic agreement and instantiate it with several classes of paths, obtaining sufficient conditions expressed in terms of well-foundedness of the corresponding relations. The resulting conditions strictly generalize two principal structural criteria for preferred and stable agreement in the literature: Dung’s limited controversiality condition and the SCC-symmetry condition. In particular, Dung’s result is recovered as a special case by taking the relevant path set to be the set of even paths. We further introduce two infinite sequences of path sets, each yielding a family of agreement conditions. For the first sequence, a Collapse Theorem reduces the infinitely many associated well-foundedness conditions to three essentially distinct ones, whereas the second sequence gives rise to genuinely distinct conditions with no analogous collapse. Our results apply to both finite and infinite argumentation frameworks, admit simplified formulations in the finite case, and also yield sufficient conditions for the existence of stable extensions. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2610.30176 [cs.AI] (or arXiv:2610.30176v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2610.30176 Focus to learn more arXiv-issued DOI via DataCite
[AI-140] Neural Networks with Differentiable Belief-Function Layers
链接: https://arxiv.org/abs/2610.30173
作者: Antonin Sulc
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 2 figures, BELIEF 2026
Abstract:Yager’s 1987 rule of combination \boxplus was proposed on epistemic grounds: under high conflict, mass should be redistributed to ignorance rather than renormalised onto the class simplex. We observe that backpropagation needs the same rule: Dempster’s normaliser (1-\kappa)^-1 , where \kappa is the conflict mass, puts a (1-\kappa)^-2 term in the gradient that diverges exactly when evidence sources disagree, while Yager’s \boxplus is bilinear and has no such term. Building on this, we construct a deep evidential pipeline in which every mass in the chain from heads to output is a valid Dempster–Shafer mass function, combined by Yager’s rule. The output mass on the ignorance set, m(\Theta) , serves as the network’s ``I do not know’’ signal. The network is fit to hard labels and never sees an annotator distribution, yet m(\Theta) orders human disagreement in two modalities, visual and linguistic. Dempster’s rises too and ranks contested inputs about as well, but its normaliser divides the conflict out: Yager’s vacuity is 9\times larger on the most contested CIFAR-10H images, so only Yager’s is readable as an amount of ignorance without recalibration. For out-of-distribution detection it is competitive with strong single-pass baselines, while the clamped Dempster variant that scores higher becomes gradient-unstable once the clamp is removed.
[AI-141] Accelerating Wildfire Hazard Mapping Across Canada with an AI Surrogate Modelling Framework
链接: https://arxiv.org/abs/2610.30170
作者: Hager Radi Abdelwahed,Toluwaleke Olutayo,Chirag Goel,Jana Osea,Aldo Zaimi,Tom Swystun,Sandy Erni,Francisco Emiliano Lopez Saavedra,Yan Boulanger,Gaétan Marceau Caron
类目: Artificial Intelligence (cs.AI)
备注: Preprint - under review
Abstract:Canada’s increasing wildfire activity requires frequent national hazard mapping to guide proactive mitigation, including fuel management, land-use planning, infrastructure protection, and evacuation routing. However, process-based wildfire simulation frameworks (e.g., BurnP3 and BurnP3+) rely on computationally intensive Monte Carlo simulations to model burn probability, fire intensity, and rate of spread. These workflows can take days to weeks to process a single regional modelling unit. To overcome these computational bottlenecks, we present three contributions developed through collaboration between AI researchers and wildfire modelling experts: (a) a standardized framework for training, fine-tuning, and evaluating AI surrogate models against process-based wildfire simulators; (b) a multi-task deep learning surrogate trained on national-scale data to emulate burn probability, mean fire intensity, and mean rate of spread at a fraction of the computational cost (seconds); and © domain-specific metrics and counterfactual evaluation protocols to probe out-of-distribution generalization. Integrating geospatial data (fuel, terrain, ignition likelihood) and regional weather, our model approximates BurnP3+ simulation outputs, achieving strong performance across agreement, rank-correlation, error, and spatial-overlap metrics. It preserves downstream hazard classifications and produces physically plausible responses to expert-informed counterfactual interventions. On a representative region, the surrogate achieves 806-fold end-to-end CPU acceleration and uses 13% of BurnP3+'s host memory, facilitating rapid scenario analysis.
[AI-142] MCP vs CLI for AI Agents : A Comparative Analysis of Tool Integration Strategies for Autonomous Systems
链接: https://arxiv.org/abs/2610.30157
作者: Arun Malik,Nitish Talasu,Bradley Klemick,Vineet Tushar Trivedi,Prachi Shah,Sanay Devi,Junjie Zhang,Kunal Sonawane
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Distributed, Parallel, and Cluster Computing (cs.DC); Emerging Technologies (cs.ET); Software Engineering (cs.SE)
备注: 9 pages, 6 figures
Abstract:AI agents require interfaces to interact with external systems such as files, APIs, databases, and cloud services. Two competing paradigms have emerged: CLI (Command Line Interface), where agents execute shell commands directly, and MCP (Model Context Protocol), a structured JSON-RPC protocol providing governed, typed tool access. This paper presents a comparative analysis across 14 evaluation dimensions, supported by worked examples, security assessment against established compliance frameworks, and economic modeling. Our analysis indicates that CLI provides superior efficiency for single-user, local-only workflows, while MCP addresses requirements that CLI fundamentally cannot satisfy, including fine-grained access control, credential isolation, structured audit, and multi-tenant governance. We further examine how progressive disclosure techniques eliminate MCP’s historical weakness (context window overhead), yielding token efficiency comparable to CLI while preserving its architectural benefits. A hybrid architecture is recommended for production systems. This analysis draws upon academic, government, and industry sources including NSA formal guidance (CSI U/OO/6030316-26), OWASP security standards, and economic modeling from BCG and Gartner.
[AI-143] he Cost of Saying Hello: An Empirical Study of Handshake Overhead in AI Agent Communication Protocols
链接: https://arxiv.org/abs/2610.30151
作者: Pranav Varshney
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 15 pages, 4 figures, 4 tables. Harness, raw per-run data, and analysis: this https URL
Abstract:Every interaction between AI agents begins with a handshake: the initiating agent must locate its counterpart, learn what it can do, and establish a session before the first useful message can be sent. Surveys of agent communication protocols name handshake overhead as an efficiency criterion, yet none report measurements. This paper presents the first empirical, phase-decomposed benchmark of the handshake in three major agent protocols: MCP (in both its 2025 initialize and 2026 discover eras), A2A, and ACP. Using each protocol’s official SDK, we instrument every request for round trips, per-phase latency, wire bytes, and token cost across cold-start, warm-repeat, capability-mismatch, and degraded-metadata scenarios (N=50 per cell, on loopback and under an emulated 50 ms RTT). Round-trip count dominates handshake cost. A legacy MCP handshake takes three round trips and 230 ms at 50 ms RTT; A2A needs one round trip and 61.0 ms. Round trips are not the whole bill: when an LLM reads the counterpart’s metadata, handshake cost scales with the counterpart’s capability inventory at 24-107 tokens per capability, so a 200-capability counterpart spends 11% of a 200k-token context window before the agent reads its own task. MCP’s own 2026 revision cuts its handshake from three round trips to two and, with era pinning, to zero on warm reconnects, protocol evolution that corroborates the cost we measure. We release the full harness, raw data, and analysis at this https URL.
[AI-144] Can Entanglement-Induced Feature-Map Designs Improve Projected Quantum Kernels?
链接: https://arxiv.org/abs/2610.31260
作者: Ravi Kumar Jha,Hsuan-Yin Lin,Eirik Rosnes
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Projected quantum kernels (PQKs) have emerged as a notable approach to circumvent the hardness of machine learning tasks arising from the exponential dimensionality of the quantum Hilbert space. However, recent PQK studies often rely heavily on extensive classical hyperparameter tuning of the radial basis function (RBF) kernel parameter and the support vector machine regularization parameter. The extensive tuning can mask the true representational power of the underlying PQKs induced by quantum feature maps, which has not been explored in detail. Thus, this work aims to expose the structural capabilities of PQKs by isolating and evaluating their native inductive bias through comparisons between untuned and tuned PQKs against the RBF baseline. Furthermore, we propose novel quantum feature maps that leverage two-qubit entanglement to capture multivariate correlations in IoT data. The results show that extensive tuning can lead to competitive performance architectures, while our proposed entanglement topologies exhibit a significantly stronger native inductive bias. This reveals that superior quantum architectural design can yield structurally better PQKs with improvements arising from quantum-mechanical properties rather than from classical hyperparameter tuning.
[AI-145] QICOM: Open-Source Software for Reproducible Image Quality Evaluation Blinded Assessment and Radiomic Comparison
链接: https://arxiv.org/abs/2610.31218
作者: Fabio Mattiussi,Switinder Ghotra
类目: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI)
备注:
Abstract:QICOM (Quality Image COMparator) is open-source software for quantitative image comparison, paired radiomics, and blinded assessment. It saves processing settings, image identities, measurements, and response mappings so a comparison can be reopened and recalculated. We evaluate reference agreement, responses to controlled perturbations, and recovery of saved analyses. Synthetic experiments cover 37 quantitative descriptors using independentnumerical references and separate upstream integration comparisons. The QICOM/MIRP backend meets the reference tolerances for all 482 numeric IBSI 1 digital-phantom entries and all 1,622 entries across configurations A-E of the public CT reference case; these comprise 1,804 feature/aggregation comparisons and 300 preprocessing diagnostics. Restricting masked SSIM and HFEN to complete kernel support removes the observed dependence on excluded pixels. In a synthetic filtering example, stronger smoothing increases SSIM while worsening MSE and HFEN relative to milder smoothing and changing GLCM contrast more strongly. Reopened radiomics projects reproduce all 43 feature pairs per comparison, and two blinded deliveries return four correctly mapped scripted responses. These experiments document the calculations and data transfers linking the three workflows.
[AI-146] Network topology shapes consensus and vulnerability in LLM populations
链接: https://arxiv.org/abs/2610.31165
作者: Roberto Passaro,Ariel Flint,Romualdo Pastor-Satorras,Andrea Baronchelli,Luca Maria Aiello
类目: Physics and Society (physics.soc-ph); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注:
Abstract:As large language models (LLMs) are deployed in interacting populations, their collective behaviour depends not only on the models but also on who interacts with whom. Studies of convention formation in LLM populations have mainly tested well-mixed interactions. We place policies extracted from four LLMs on fully connected, random regular, and heterogeneous networks, with 10^3 to 10^5 agents playing the naming game. Topology affects how quickly consensus emerges, sometimes whether it emerges at all, and, in rare cases, which convention wins. Sparsest networks slow consensus up to twelvefold. Yet, by preserving correlations between persistent neighbours, they restore consensus in some cases that fail to converge under well-mixed interactions. Networks also govern vulnerability to adversarial attack. We measure the critical mass of committed agents needed to overturn an established convention. Tipping thresholds follow the same ordering across models and conditions: hubs \leq random \leq low-degree nodes. On the most heterogeneous topology, median critical masses are 3%, 16%, and 35%, respectively; targeting hubs requires one-sixth as many committed agents as random placement. For most models, networked populations are easier to tip than well-mixed ones, so well-mixed estimates can overstate robustness. Assessing the safety of multi-agent AI requires mapping who interacts with whom.
[AI-147] he Delegation Divide: An Open Index of AI-Agent Delegability across Occupations and Countries
链接: https://arxiv.org/abs/2610.30752
作者: Cristian Espinal Maya(Universidad Nacional de Colombia),Santiago Jiménez Londoño(Universidad EAFIT)
类目: General Economics (econ.GN); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 32 pages, 6 figures, 4 tables. Preregistered at OSF: https://doi.org/10.17605/OSF.IO/9ZNTF . Data and code: doi: https://doi.org/10.5281/zenodo.23552918%3B repository and MCP server: this https URL
Abstract:35.2% of task time in world employment is at least partly delegable to AI agents in 2026, and only 2.5% can be handed over in full. Existing measures score exposure to generative AI, mostly for the United States. We build the Agent Delegability Index (ADI): two language models rate all 18,796 O*NET 30.2 task statements on a frozen, preregistered rubric of delegation levels and barriers; tasks are mapped to 433 ISCO-08 unit groups and weighted by ILOSTAT employment in 146 countries. Four findings. First, delegable employment rises with income, from 22.5% in low-income to 48.7% in high-income countries, a slope of 7.23 percentage points per log point of GDP per capita. This delegation divide is compositional: the physical barrier dominates the jobs poorer countries employ most. Second, at the occupation level ADI tracks exposure closely (r = 0.93 with Eloundou et al. and 0.89 with Felten et al.), above the preregistered bound of 0.80; ADI adds delegation levels and barriers on top of exposure, for the world, not a new ranking of occupations. Third, full delegation is rare and peaks for tasks that take a human about twelve minutes (11.8 minutes). Fourth, the preregistered adoption gap is not supported: across 101 countries, delegability and observed AI usage correlate at r = 0.86, and the gap between them is not wider in middle-income countries (p = 0.27). Model agreement is kappa = 0.77; a blind human rating of 300 tasks agrees with the model consensus at kappa = 0.85, and an agentic panel with web search at kappa = 0.87. We release the index, the atlas, the prompts and an MCP server at this https URL.
[AI-148] Lifetime-valid decisions by continually evolving AI analysts
链接: https://arxiv.org/abs/2610.30562
作者: Xiaojing Sun,Yuhan Zeng,Xiao Wang
类目: Methodology (stat.ME); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
备注:
Abstract:AI agents increasingly act as analysts whose decision can often be cast as hypothesis tests. Classical tests control the Type I error rate, the probability of rejecting a true null hypothesis, for fixed procedures. However, AI analysts introduce internal randomness and can evolve continuously over time. As AI analysts evolve, the risk of approving versions with Type I error rates above the target can accumulate, and adaptive use of evaluation feedback can lead to overfitting. We introduce continual validation, which evaluates each proposed AI analyst version before deployment using simulated data under the null hypothesis, with scenarios that can be randomized and expanded over time. We prove that, regardless of how updates are chosen or how many are proposed, the lifetime probability of ever deploying a version whose Type I error rate exceeds the allowed level remains bounded. This guarantees that every subsequent decision controls the Type I error rate at the target level across these scenarios. At the same time, updates with sufficiently low Type I error rates pass validation with high probability, allowing the analyst to continuously improve its performance without sacrificing rigor. In two examples involving evolving AI analysts for bioequivalence studies and product A/B tests, Type I error rates reached 0.098 and 0.183 without continual validation, compared with at most 0.045 and 0.044 with it, while retaining 82% and 96% of the power from optimally selected valid updates.
[AI-149] High-Value Training Data from Noisy Diverse Sources
链接: https://arxiv.org/abs/2610.30421
作者: Hedi Driss,Evan Frick,Anastasios N. Angelopoulos,Cho-Jui Hsieh,Yejin Huh,Tijana Zrnic,Andrew Ilyas
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Training state-of-the-art models increasingly requires learning from large, heterogeneous datasets whose examples vary in quality, domain relevance, and source of supervision. Dataweighting—assigning different weights to training examples—offers a general way to control how such data influences training, but learning effective weights at scale remains challenging. We develop a scalable approach to learning dataweights: given a small, high-quality target dataset, we use metagradients through the training process to efficiently optimize training-example weights directly for performance on the target domain. To scale beyond a fixed dataset and model, we learn parametric weighting rules that generalize to new and streaming examples and transfer from smaller models to substantially larger ones. We study reward modeling as a prototypical application, where preference data may be noisy, adversarial, synthetic, or mismatched with the target domain. Learned dataweights improve held-out performance, automatically adapt reward models to specialized domains, and allow mixtures of human and synthetic supervision to outperform either source alone.
[AI-150] Why Constrained Updates Still Forget: A Functional-Path Analysis of Continual SFT
链接: https://arxiv.org/abs/2610.30191
作者: Qingsen Ma,Dianyun Wang,Jiahong Liu,Irwin King
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Supervised fine-tuning adapts language models to specialized tasks but can disrupt previously acquired behavior. Existing approaches constrain parameters, gradients, or replay objectives, yet the connection between the protected object, the intervening trajectory, and retained outputs remains unclear. We develop a functional-path account that decomposes old-loss changes along realized updates into the controlled direction, gradient staleness, probe mismatch, and finite-step effects. The account yields a conditional refresh budget and identifies the restoring direction supplied by informative replay. Matched QLoRA interventions separate constraint granularity from placement before or after Adam: placement strongly affects retention across the studied code-label audits, whereas the granularity interaction depends on the endpoint. Checkpointed generation separates endpoint recovery from higher sampled-path accuracy, and matched replay schedules reveal loss exposure concealed by final scores. A sequential-task comparison extends this distinction to backward transfer: subspace regularization and projection improve average retention while trading off different old tasks; supervised replay protects both more consistently in this setting. Additional response interfaces distinguish content gains from exact-response retention. These findings connect update-level protection to task-wise behavior and motivate mitigation through explicit functional anchors, trajectory evaluation, and joint old/new auditing.
[AI-151] GeoDiv KAN Solver for Time Fractional Navier Stokes Equations on Complex Domains to Realistic CFD Reconstruction
链接: https://arxiv.org/abs/2610.30179
作者: Yatin Kumar,Shivani Saini,Ramesh Kumar Vats
类目: Fluid Dynamics (physics.flu-dyn); Artificial Intelligence (cs.AI)
备注:
Abstract:Scientific machine learning provides a promising framework for solving complex problems in science and engineering. Notably, physics informed neural networks (PINNs) have shown considerable potential for solving differential equations and modelling engineering systems. Solving time fractional Navier Stokes equations (TFNSEs) on complex domains is challenging due to temporal nonlocality, nonlinear flow dynamics, geometric complexity, and incompressibility constraints. To address these challenges, we propose a geometry gated, divergence-free Gaussian radial basis function Kolmogorov Arnold network (GeoDiv KAN). The proposed architecture uses geometry dependent feature gating and a streamfunction formulation to preserve incompressibility by construction. GeoDiv KAN is evaluated for the time fractional incompressible Navier-Stokes equations on square, circular, and star-shaped domains. Numerical experiments show that GeoDiv KAN achieves higher predictive accuracy than PINN, PIKAN, RBF KAN, and DF RBFKAN across the considered benchmark problems. To establish practical relevance, GeoDiv KAN is applied to two data driven CFD cases: unsteady flow past a circular cylinder using the \hrefthis https URLZenodo dataset and turbulent RANS flow through a complex U bend using the \hrefthis https URLU-bend dataset.
机器学习
[LG-0] Phase Transition from Random Features to Fully Trainable Networks
链接: https://arxiv.org/abs/2610.31663
作者: Xin He,Huan Lao,Xiliang Lu,Shijun Zhang
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG); Statistics Theory (math.ST); Machine Learning (stat.ML)
*备注:
Abstract:Constructing accurate and computationally tractable approximation spaces is a fundamental task in scientific computing. Partially trainable neural networks bridge fixed random features and fully adaptive representations. This paper quantifies a phase transition pattern in approximation bounds and observes analogous behavior in training errors: a rapid but smooth initial improvement followed by diminishing returns, with a sharper transition under greater initialization mismatch. Under polynomial spectral decay with exponent \alpha1 and leverage compatibility, we establish a squared L_2 approximation rate \mathcalO((\kappa_P\log(m/\delta)/m)^\alpha) for two-layer networks of width m and trainable proportion P , where \kappa_P=(P/\kappa_1^+(1-P)/\kappa_0^)^-1 . The bound holds uniformly over the unit ball of a reproducing kernel Hilbert space (RKHS) with probability at least 1-\delta . The constants \kappa_0^* and \kappa_1^* quantify the fixed and trainable groups’ sampling mismatches relative to the target RKHS. For spherical biased-ReLU features in dimension d , we verify these assumptions with \alpha=(d+2)/(d-1) and \kappa_0^=4\sqrt2\tau(d+1)/d for initialization scale \tau\ge1 , and \kappa_1^=1 under ridge leverage sampling. For d=5 and \tau=10 , a trainable proportion of approximately 4.1% captures 90% of the reduction in the approximation bound from P=0 to P=1 . A structured three-layer example illustrates how a few trainable first-layer neurons exploit fixed downstream blocks through nonlinear composition. Numerical experiments across diverse settings, including networks with up to 15 layers, exhibit similar phase transitions in training errors.
[LG-1] A Fix for Softmax and Cross-Entropy
链接: https://arxiv.org/abs/2610.31642
作者: Jonathan T. Barron,Dan Gnanapragasam,Michael Tschannen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Standard implementations of softmax, log-softmax, and cross-entropy have two numerical bugs. The first bug, Softmax Collapse, rounds the target logit’s gradient to zero if that one logit is much larger than the others, resulting in gradients that no longer sum to zero. The second bug, Softmax Cancellation, occurs when student logits are close to teacher logits, and produces non-zero gradients even when the student and teacher logits are equal. We present simple implementations of log-softmax, softmax, and cross-entropy that fix both bugs. Our fix eliminates the previously-observed “Slingshot” phenomenon, speeds up grokking with weight decay, prevents accuracy collapse without weight decay, stops student networks from drifting away from the teacher during distillation, and shows that after memorization, logit growth is bounded by an equilibrium between weight decay and Adam’s \epsilon hyperparameter.
[LG-2] BeNice! Closing Loopings Inference-Memory Gap via KV-Sharing
链接: https://arxiv.org/abs/2610.31633
作者: Hayden Prairie,Jacob M. Ryan,Taylor Berg-Kirkpatrick,Daniel Y. Fu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Looped architectures scale quality while keeping parameter count fixed by repeatedly passing activations through the same block of layers. During inference, each recurrence generates independent KV-caches, causing memory to grow linearly. To address this challenge, prior work has proposed KV-sharing (i.e., sharing KV-caches across depth), finding that, in some cases, the KV-cache of the final recurrence can be shared with the earlier recurrences without quality loss, a property we denote as KV-shareability. However, the mechanisms underlying KV-shareability remain poorly understood, with even identically trained looped architectures not consistently exhibiting it. In this paper, we study how KV-shareability can be enabled in looped models. We propose a two-stage method: in the first stage, we use stochastic depth during pretraining, which induces contractive residual dynamics that drive representations toward a common limit across recurrences. In the second stage, we propose BeNice!, a post-training method that broadcasts the final recurrence’s KV-cache into the first recurrence, training it to produce compatible residuals while preserving the original model’s quality. As a result, BeNice! enables KV-shareability while matching standard finetuning’s unshared validation perplexity. Using BeNice!, we show that KV-sharing is complementary to alternative KV-compression methods, composing with memory-efficient sequence mixers (e.g., multi-latent attention, Kimi delta attention) and low-precision KV-cache quantization (e.g., fp8, mxfp8, nvfp4). Finally, in simulated low-memory, long-context inference, KV-sharing improves cache hit rates by up to 31.4%, increases throughput by up to 23.6%, and reduces p90 time-to-first-token by up to 83.6%.
[LG-3] OP-Print: Tracing Membership Imprints in On-Policy Self-Distillation
链接: https://arxiv.org/abs/2610.31616
作者: Qi Li,Xinchao Wang
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
*备注:
Abstract:On-policy distillation has been widely adopted for alleviating the distribution mismatch of off-policy distillation. Self-distillation further removes the need for an external teacher by using the student’s own outputs, enhanced with extra context or demonstrations, as supervision. This teacher-free design makes on-policy self-distillation practical, yet its privacy risks remain largely unexplored. In particular, using two coupled views of the same training example may strengthen its memorization. Motivated by this, we present the first study of membership inference attacks against on-policy self-distillation. We investigate both gray-box and black-box threat models and propose a unified attack framework, named OP-Print. OP-Print builds on our finding that the supervision imprint in on-policy self-distillation is sparse and structured, concentrating on uncertainty-sensitive correction sites rather than across the entire rollout. Exploiting this property, we construct attack signals tailored to the two access levels: in the gray-box setting, OP-Print directly aggregates the privileged teacher-student residual correction along the same rollout; in the black-box setting, it measures restart-based output-basin stability as an observable surrogate for the hidden correction process. Across diverse models and distillation schemes, all baseline attacks remain essentially at random guessing even under gray-box access, whereas OP-Print is broadly validated to be effective in both gray-box and black-box settings.
[LG-4] Efficient and Optimal Learning of Discrete Distributions with Person-Level Privacy
链接: https://arxiv.org/abs/2610.31612
作者: Gautam Kamath,Mahbod Majid,Ankur Moitra,Argyris Mouzakis,Jonathan Ullman
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Data Structures and Algorithms (cs.DS)
*备注: Full version of the SODA 2027 paper
Abstract:We study learning a discrete distribution under person-level differential privacy in the batched setting. Given n users, each contributing m i.i.d. samples from an unknown distribution \mu over [k] , we seek an estimate of \mu with total-variation error \alpha . We give polynomial-time algorithms under both pure and approximate differential privacy. Under pure (\varepsilon,0) -DP, we achieve n = \widetilde O(k/(\alpha^2 m)+k/(\alpha\sqrtm\varepsilon)+k/\varepsilon) , the first polynomial-time pure-DP guarantee matching the information-theoretic optimum up to logarithmic factors. Under approximate (\varepsilon,\delta) -DP, we achieve n = \widetilde O(k/(\alpha^2 m)+k/(\alpha\sqrtm\varepsilon)+k+\log(1/\delta)/\varepsilon) . At \delta=2^-\Theta(k) , this is \widetilde O(k) for fixed \alpha,\varepsilon,m and failure probability. Our algorithms use a new degree-6 sum-of-squares estimator for learning from untrusted batches. Its key ingredient is a trace-bounded diagonal majorizer for the difference between the empirical covariance of the repaired batch averages and the multinomial covariance predicted by their mean. We prove that clean data admit this certificate through semidefinite-programming duality and a new application of the stronger symmetric form of Grothendieck’s inequality due to Friedland and Lim. The certificate preserves the mean-dependent variance of multinomial batch averages, yielding the near-optimal robust rate O(\eta\sqrt\log(1/\eta)/m) and providing the score required by the privacy reduction. We also prove a statistical query lower bound ruling out a uniformly smaller order of error under fixed polynomial SQ resource bounds.
[LG-5] CRANE: Post-Hoc Counterfactual Retrieval for Multimodal Classifiers
链接: https://arxiv.org/abs/2610.31594
作者: Franco Rugolon,Ioanna Miliou,Panagiotis Papapetrou
类目: Machine Learning (cs.LG)
*备注: 7 pages + references, 3 figures, accepted to DSAA 2026
Abstract:Counterfactual explanation methods have been introduced for unimodal settings, with dedicated frameworks for tabular, timeseries, text, and image data, yet no post-hoc method has been designed for classifiers that learn from multimodal inputs. Existing explainability methods for multimodal classifiers apply modality-specific attribution independently to each input, which fails to yield coherent explanations and does not transfer to arbitrary pre-trained multimodal architectures. To address these shortcomings, we present CRANE (Counterfactual Retrieval via Agnostic Nearest-neighbour Explanations), the first post-hoc, model-agnostic framework for nearest-neighbour multimodal counterfactual generation that retrieves real training examples from the target class as candidate explanations, ensuring plausibility by construction. Across four heterogeneous datasets and three fusion strategies, we show that Multimodal Consensus Retrieval (MC-R), which intersects per-modality neighbour sets to identify instances jointly close to the query in every modality, emerges as the strongest multimodal strategy, achieving the best or joint-best outcome on three of four datasets while grounding every explanation in the joint training distribution. Unimodal strategies remain competitive when a single modality dominates the decision boundary.
[LG-6] Action Learning for Scientific Simulations
链接: https://arxiv.org/abs/2610.31585
作者: Sarah Liaw,Yuanqi Du,Yilun Du
类目: Machine Learning (cs.LG)
*备注: 25 pages, 7 figures
Abstract:Understanding and controlling how systems evolve is central to science and engineering, where many problems require reasoning about complete trajectories. Existing approaches typically learn local dynamics or generate trajectories, but objectives constructed from individual transitions can miss dependencies across time, while generative models do not necessarily provide a tractable objective for evaluating and optimizing arbitrary paths. We study \textitaction learning , the problem of recovering a stochastic action from trajectory observations, yielding a common objective for reasoning about dynamical systems. We introduce \textbfLEAP ( \textbfL earned \textbfE nergy on \textbfA ction \textbfP aths), which anchors the unknown action to a tractable reference process and learns the remaining correction by comparing observed trajectories with reference trajectories. This construction yields a differentiable objective on path space that can capture both local dynamical structure and nonadditive dependencies across time. Once learned, the same action can be flexibly reused for path ranking and inference, mechanism steering, constrained optimization and sampling. Across stochastic physical systems, many-body orbital dynamics and molecular simulations, LEAP recovers underlying dynamics, captures dependencies across time and enables transfer across inference objectives and physical conditions. We show that \textitaction learning is a promising framework for turning trajectory observations into reusable objectives for scientific inference and control.
[LG-7] Rethinking Sequence Compression with Test-time Regression
链接: https://arxiv.org/abs/2610.31574
作者: Sihan Yang,Jie Huang,Hanxuan Li,Yifan Guo,Weigao Sun,Yiran Zhong,Weiyang Liu
类目: Machine Learning (cs.LG)
*备注: Technical report (19 pages, 4 figures, project page: this https URL )
Abstract:Sequence compression is a fundamental operation and a critical information bottleneck in efficient language and multimodal models, and yet prevailing approaches based on pooling or learned projections remain underperforming. We revisit sequence compression through the lens of test-time regression: input tokens define a regression problem whose fitted coefficients form the compressed representation. Importantly, the regression problem is learned end-to-end via backpropagation, while its coefficients are computed efficiently in closed form during the forward pass. This leads to TTR-Compressor, a general alternative to conventional pooling and projection-based compression. We evaluate TTR-Compressor as the compressor for block-wise sparse attention and the visual token merger for multimodal large language models. On block-wise sparse attention, TTR-Compressor improves RULER performance over attention-based compressors by a large margin. For visual token merging, it consistently outperforms the standard MLP merger and matches its performance with 55.6% fewer visual tokens. These results establish test-time regression as an effective primitive for sequence compression in efficient language and multimodal models.
[LG-8] On the Token Collapse Phenomenon in Textual Graph Soft Prompting
链接: https://arxiv.org/abs/2610.31572
作者: Andrea Giuseppe Di Francesco,Rishabh Jain,Franco Maria Nardini,Nicola Tonellotto,Pietro Liò,Fabrizio Silvestri
类目: Machine Learning (cs.LG)
*备注: Preprint under review
Abstract:Recent methods for injecting graph structure into LLM prompts include graph soft prompting, in which a graph neural network (GNN) learns graph-specific tokens that capture relational information among entities in textual prompts. G-Retriever combines graph soft prompts with a verbalized description of the same graph to strengthen the textual prompt signal. We conduct an in-depth analysis of G-Retriever and uncover a pathological behavior we term token collapse. Across three benchmarks, three graph encoders, two LLM families, and both frozen and LoRA-adapted models, G-Retriever’s graph tokens become almost indistinguishable across inputs. We provide evidence that the same failure is inherited by subsequent methods, even with multi-token pooling. Under the standard setup, swapping a token with the dataset mean, a token from another example, or one derived from a corrupted graph produces no resolved answer-level effect in accuracy. Moreover, a learned constant prompt that never accesses the graph matches or surpasses G-Retriever on all three benchmarks. These findings indicate that, rather than conveying graph content as intended, the collapsed token functions primarily as a generic soft prompt. We also examine conditions linked to this collapse and propose a two-stage procedure that yields non-collapsed tokens while maintaining competitive performance in most cases. However, token variability alone does not prove that the model uses graph information: via interventions and decoding probes, we show that depending on the configuration, a token can reflect the question, an answer prior, or graph-derived content that the LLM may or may not use. Overall, our results offer a practical approach to evaluating graph-token alignment rather than inferring it from task accuracy alone. Code to reproduce our experiments is available at this https URL.
[LG-9] Fine-Grained Analysis of Catastrophic Forgetting in Continual Learning via Algorithmic Stability
链接: https://arxiv.org/abs/2610.31551
作者: Guodong Zheng,Yan Sun,Peng Wang,Enneng Yang,Jiafei Lyu,Xiaochun Cao,Dacheng Tao,Li Shen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Understanding catastrophic forgetting, in which learning new tasks degrades performance on previously learned tasks, remains a fundamental challenge in continual learning (CL). Existing theoretical analyses often rely on specific model or task structures and offer limited insight into how iterative optimization affects forgetting. To address this gap, we develop an algorithmic stability framework for expected population forgetting under SGD without prescribing a particular model architecture. Specifically, we decompose forgetting into intra-task and cross-task generalization gaps and empirical forgetting. To bound the generalization gaps, we track how a sample perturbation affects updates within a task and propagates through subsequent tasks. We consider constant, intra-task decaying, and continuously decaying learning-rate schedules. Comparing the two decaying schedules allows us to characterize how restarting learning-rate decay at task boundaries affects the propagation of historical sample perturbations. For empirical forgetting, we derive bounds for all three schedules by retaining cumulative gradient alignment along the optimization trajectory, without requiring convexity. Combining these results yields upper bounds on expected forgetting that capture the effects of training set size, the number of updates per task, the total number of tasks, learning-rate schedules, and cumulative gradient alignment. Finally, we conduct experiments on Split CIFAR-100 and Split MiniImageNet to examine the empirical implications of our theoretical findings.
[LG-10] Shift-aware score transport for privacy auditing without retraining
链接: https://arxiv.org/abs/2610.31549
作者: Yaohong Yang,Gauri Pradhan,Qi Chen,Samuel Kaski,Antti Honkela
类目: Machine Learning (cs.LG)
*备注:
Abstract:Privacy auditing empirically assesses privacy leakage by using membership inference attacks to establish lower bounds on privacy loss. We consider an observational auditing setting in which the auditor cannot insert canaries, control data inclusion, or retrain the target model, and must instead rely on fixed sets of known members and non-members. This setting is relevant for large deployed models or when retraining is restricted by computational or access constraints. In this case, non-members available for auditing may differ in distribution from the training data, causing membership-inference scores to be skewed due to the distribution shift. We address this problem by estimating the counterfactual in-distribution non-member score distribution using the available out-of-distribution (OOD) non-members. Our proposed method learns the relationship between data representations and membership-inference scores from OOD non-members and transports it to the target covariate distribution. Because this transport may become unreliable under extrapolation or imperfect conditional invariance, we further introduce a conservative correction calibrated directly to the nominal transported privacy level. We construct controlled calibration tasks from the available non-member data to estimate how representation-space displacement inflates the audit, and use this estimate to correct the transport-induced audit error. Across multiple datasets with distribution shifts, our method yields substantially more stable privacy estimates while retaining tighter lower bounds than baselines.
[LG-11] Same Luck Different Choice: Twin-World Credit Assignment for Agent ic Reinforcement Learning
链接: https://arxiv.org/abs/2610.31544
作者: Zeropy Surio,Tobias Engel,Mei-Lin Chou,Anika Raman,Samuel Okafor
类目: Machine Learning (cs.LG)
*备注:
Abstract:Agentic reinforcement learning rewards a complete episode, yet the policy needs credit for every decision inside it. Counterfactual branching estimates this credit by re-running the episode after an alternative decision, but the branch draws its sampling noise independently of the factual episode, so the outcome difference mixes the effect of the decision with the luck of everything that follows. We observe that in agentic search the retriever, the corpus, and the grader are deterministic, so an episode is a deterministic function of the policy’s decisions and its sampling noise. Twin worlds are therefore realizable: the same noise, a different decision. We introduce TWICE (TWIn-world Credit Estimation), which keys the Gumbel noise of every token by turn and line, rolls out counterfactual twins that inherit the factual episode’s downstream noise, and turns their paired outcome differences into turn-level advantages for policy-gradient updates on actor-sampled trajectories; only the actor is deployed. Twin credit is unbiased for any coupling quality, its variance is that of an independent counterfactual scaled by 1-\rho , where \rho is the twins’ outcome correlation, and under coupling the number of rollouts needed to resolve an advantage of size \Delta falls from \Theta(\Delta^-2) to \Theta(\Delta^-1) . A square-root rule spends the twin budget where outcomes are predicted to disagree per unit cost. On seven question-answering benchmarks, TWICE improves the average exact match of GRPO by 5.3 and 4.5 points on Qwen2.5-3B and 7B, outperforms tree-search and counterfactual-branching baselines, and uses 68% fewer auxiliary tokens than VinePPO.
[LG-12] PYTHIA a transferable machine learning framework for simultaneous defect and adsorption site optimization
链接: https://arxiv.org/abs/2610.31539
作者: Ioannis Kouroudis,Zhaosu Gu,Poonam,Leon Muller,Alessio Gagliardi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Machine learning has permeated material optimization, providing high value. Nevertheless, its re- liance on a high number of samples limits its applicability to relatively explored material families. In this paper we extend the applicability of machine learning to novel and understudied materials through the PYTHIA (Probabilistic Yet Transferable Hierarchical Interaction Architecture) framework. Pythia utilizes two of the most powerful machine learning concepts, foundational models and transfer learning. We employ a complex foundational model trained on a large dataset of chemically similar molecules, on a target that is physically close to our prediction target. Then, we freeze the weights and perform subspace alignment with the use of a transformer layer and add uncertainty quantification with gaussian processes. These two layers can be trained on a limited dataset yet allow us to leverage the gained knowl- edge and the studied materials and complex architectures. The resulting model is used as a surrogate for a Bayesian Optimization scheme. To showcase the robustness of the framework, we applied it to the investigation of adsorption and defect optimization of two distinct systems. Not only were we able to correctly and efficiently identify the optimal adsorption sites of even complex a
[LG-13] Piano2Pop: Expanding Piano Demos into Pop Productions Without Aligned Supervision ICASSP2027
链接: https://arxiv.org/abs/2610.31510
作者: Yun-Chen Cheng,Chih-Pin Tan,Tzu-Hung Huang,Yi-Hsuan Yang
类目: Machine Learning (cs.LG); Sound (cs.SD)
*备注: Submitted to ICASSP2027
Abstract:We introduce Piano2Pop, a music-expansion task that transforms piano MIDI into instrumental pop productions. Enabled by recent advances in music generation and polyphonic music transcription, our pseudo-supervised approach conditions a pretrained flow-matching model on piano rolls without aligned piano-pop supervision. We train on synthetic piano-instrumental pairs derived from pop recordings, comparing pair-free transcription, including function-selective (FST) and register-density (RDA) augmentations, with a pair-supervised piano-cover generator as pseudo-demo sources. On 476 testing real-world piano-cover/pop pairs, the pair-supervised approach achieves the best Frechet Audio Distance, Accompaniment Prompt Adherence, and spectral shape on real-piano inputs, while FST mitigates piano-heavy outputs in pair-free training and remains a promising alternative. All fine-tuned systems significantly outperform the zero-shot baselines, which suggests a role for task-specific modeling.
[LG-14] Structure-preserving reduced-order models of thin panels in supersonic and hypersonic flow
链接: https://arxiv.org/abs/2610.31504
作者: Miguel A. Aguilo
类目: Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Numerical Analysis (math.NA)
*备注: 68 pages, 15 figures
Abstract:We build structure-preserving reduced-order models for both challenge problems of the AePW-4 High-Speed Working Group. They are a clamped thin panel at Mach 1.92 (RC-19) and a cantilever plate at Mach 5.85 (HyMAX). The published data for each case hold no time-resolved response and no repeated runs. The physics must therefore be carried by the form of the model. One construction gives two models per problem. A Ritz-Galerkin reference model describes the plate. A learned model keeps the reference model’s exact potential, fits the remaining generalized forces, and trains only a dissipation that cannot add energy. Its energy balance holds for any weights by construction. On the trained model the worst relative residual is 2.2e-15. On RC-19, shock relations predict a pressure ratio of 1.521 across the panel. Pressure-sensitive paint measured the same value. Thirty runs of the learned model place the mean dominant frequency within 1.0% of the measured 236 Hz. At one periodic and one chaotic condition, the learned model reproduces the response regime of a published reduced-order model in all sixteen runs. It does so without that model’s 113 Hz spectral peak, which the measurements do not show. Measured displacements can also correct the chaotic response through one more force that respects the energy balance. We tested this on synthetic data at the experiment’s 5 kHz sampling rate. The state error is 1.25% of the panel thickness without noise and 3.5% with 10 micrometers of noise. On HyMAX the learned model follows the reference model’s trailing-edge displacement for 65 ms with a relative error of 3.7%. These results show that a model whose structure carries the physics can match measured flutter behavior when the available data cannot train it.
[LG-15] Evaluating the Applicability of LLM Inference Optimizations on Apple Silicon NEURIPS2026
链接: https://arxiv.org/abs/2610.31479
作者: Kyrylo Yemets,Andrii Yaroshevych,Victor Muryn,Ostap Khomenko,Pavlo Kryven,Maksym Kmet,Maksym Shamrai,Mariya Hirna
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注: Paper was accepted at the NeurIPS 2026 Workshop on On-Device Intelligence (ODI) and the NeurIPS 2026 Workshop on SLMs for Agentic Systems (SLM-Agents). OpenReview ODI: this https URL , OpemReview SLM: this https URL
Abstract:Apple Silicon is among the most widely used platforms for local LLM inference, yet nearly all inference optimizations are designed and validated on server-class NVIDIA GPUs under batched serving workloads. Which optimizations retain their benefits when moved to Apple Silicon, and what determines their applicability? We evaluate four optimization families: KV-cache pruning, speculative decoding, ported GPU kernels, and post-training quantization. Experiments span four Apple machines under a common single-request decode workload, with matched RTX 4090 comparisons where possible. KV-cache pruning can reach up to 1.75 times end-to-end throughput on Apple but can reduce throughput on desktop NVIDIA GPUs. Under the common benchmark, all evaluated speculative decoding methods reduce throughput on Apple, despite reported 2-6 times gains on server GPUs. CUDA kernel designs transfer only partially: public Metal lacks asynchronous copy, and batch-one decode makes several serving-oriented mechanisms inapplicable, although transferable components yield up to 1.20 times decode throughput. Quantization gains vary by machine. Across four families, a reported speedup is not a portable property of a technique alone but a joint function of technique, runtime, hardware contract, and workload regime.
[LG-16] RheOPT: Few-Shot Non-Newtonian Flow Prediction via Newtonian Pretraining of RheOFormer
链接: https://arxiv.org/abs/2610.31465
作者: Maedeh Saberi,Amir Barati Farimani,Safa Jamali
类目: Machine Learning (cs.LG)
*备注: 15 pages, 7 figures, Supplementary Information
Abstract:Despite advances in numerical methods, non-Newtonian flow simulations remain computationally expensive because velocity and stress must be advanced together through rate- and history-dependent constitutive dynamics. Neural-operator surrogates reduce inference cost after training, but their training data require the same expensive simulations they aim to replace, making data-efficient learning essential when only a few non-Newtonian simulations can be afforded. Here, we introduce the Rheological Operator Pretrained Transformer (RheOPT), which pretrains a transformer-based neural operator on inexpensive Newtonian simulations and fine-tunes it on a small number of non-Newtonian samples. Rather than transferring the Newtonian constitutive response itself, RheOPT reuses geometry-conditioned transport structure that remains useful after the stress model changes. We evaluate RheOPT in two geometry-matched settings: a 4:1 contraction channel for Giesekus, Oldroyd-B, and Bingham fluids, and the oscillatory wake behind a triangular obstacle for Oldroyd-B flow. In the contraction, Newtonian pretraining substantially improves few-shot prediction, and a model fine-tuned on four target samples can match or exceed a from-scratch model trained on four times as many. This benefit is reliable under reversible instance normalization (RevIN) but dataset dependent under Global normalization, a difference we trace to the dynamic range of each fluid’s velocity and stress fields. In the wake, where relative L_2 error alone cannot verify that vortex shedding is recovered, phase-insensitive diagnostics show that pretraining recovers the shedding frequency and wake fluctuation amplitude that are suppressed by the from-scratch model in the few-shot regime. Overall, RheOPT establishes Newtonian simulation as a physically grounded pretraining source for data-efficient non-Newtonian operator learning.
[LG-17] Drift Says When Budget Says How Much: Rate-Controlled KV Re-selection for Long-Form Multimodal Generation
链接: https://arxiv.org/abs/2610.31433
作者: Bin Hu,Yingbo Xiao
类目: Machine Learning (cs.LG)
*备注: 17 pages, 7 figures, 13 tables
Abstract:KV cache compression for multimodal LLMs usually selects the visual tokens to keep once, at prefill, and reuses that selection for the whole generation. We show that this one-shot selection is severely limited on long-form generation: on detailed image description, even an oracle that selects with the attention of the uncompressed answer reaches only 0.44-0.52 ROUGE-L agreement with the uncompressed output, because the evidence the model reads moves as the description unfolds. Re-selecting whenever attention drifts out of the kept set is the natural remedy, but we find that a fixed drift threshold fails systematically: drift magnitudes are not comparable across samples, so low-drift samples are almost never refreshed although they benefit the most. We propose Rate-Controlled Refresh (RCR), which lets drift decide when and where to re-select while a per-sample controller regulates how often, combined with a low-rank drift detector, candidate-restricted re-selection, and a decision that stays on the GPU. With one setting for all models and tasks, RCR outperforms the per-step retrieval method InfiniGen by 1.6-3.8 ROUGE-L points on five of six model-task settings with no significant difference on the sixth, while staying within 1.7 points of exact per-step re-selection, with comparable estimated data movement per step and a 4x smaller GPU-resident detector. On DetailCaps, drift-chosen timing beats random timing at the same refresh rate and periodic timing even at a higher rate, by 2.0-5.0 points. On pre-registered video description, where drift is homogeneous across samples, RCR is on par with InfiniGen on Qwen2-VL and exceeds it on Qwen3-VL but no longer beats a fixed threshold at the nearest tested rates, consistent with the diagnosis; object-level evaluation shows that RCR achieves higher object F1 against the uncompressed output than InfiniGen on two of three models.
[LG-18] Convex Equilibrium Potentials for Monge–Ampère: Curvature Bounds and A Posteriori Evaluation
链接: https://arxiv.org/abs/2610.31386
作者: Edison Vázquez,Yingdong Lu,Tomasz Nowicki
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG)
*备注:
Abstract:Equilibrium neural networks define their hidden state as the fixed point of a contraction, so their effective depth is set at evaluation time rather than by the architecture. Using them for fully nonlinear PDEs raises two issues that explicit networks avoid: second derivatives may grow with the effective depth, and the potential is only ever evaluated approximately. We study nonnegative softplus equilibrium networks with contraction factor q1 , a convex class suited to the Monge–Ampère equation. A positivity cancellation in the resolvent shows that the Hessian grows at most like (1-q)^-2 , rather than the (1-q)^-3 of a direct estimate, and an explicit one-neuron family shows that this exponent is sharp; third derivatives satisfy a sharp (1-q)^-3 bound. Residual-based a posteriori estimates then control the state, its input derivatives and, in two dimensions, the Hessian determinant at an approximate equilibrium. Combined with interval arithmetic on complete covers of the domain, they yield rigorous uniform bounds on the Monge–Ampère residual and, through a comparison argument, on the solution error of a fixed trained network. On two manufactured Dirichlet problems on the unit disk, all six trained models are certified with residual below 0.25 and boundary error below 0.01 . Adaptive stopping halves float64 evaluation time at best, whereas certified adaptive evaluation remains costlier than fixed depth.
[LG-19] Harmful Responses Are Jailbreak Templates: Transferring Autoregressive Jailbreaks to Diffusion Language Models
链接: https://arxiv.org/abs/2610.31331
作者: Jinwei Gan,Zifeng Cheng,Zhiwei Jiang,Ran Sun,Shiping Ge,Qing Gu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Diffusion large language models (dLLMs) enable flexible masked generation through bidirectional context modeling, but this capability also introduces a distinct jailbreak attack surface. We find that a harmful response to a given query naturally serves as an effective jailbreak template: simply masking a subset of its tokens can induce a dLLM to reconstruct harmful content. Based on this observation, we propose RMT-Jailbreak, a general framework that transfers existing autoregressive (AR) jailbreak attacks to dLLMs through their induced harmful responses. RMT-Jailbreak converts an AR-generated harmful response into a high-mask-ratio template using uniform token retention, and further applies prompt-based safety filtering to mask any remaining visible harmful content. The refined template is then provided to the target dLLM for bidirectional reconstruction. Extensive experiments across multiple AR jailbreak methods and target dLLMs demonstrate the effectiveness and generality of RMT-Jailbreak, revealing strong cross-paradigm transferability of jailbreak attacks.
[LG-20] he Value of System 1 Models in Medicine
链接: https://arxiv.org/abs/2610.31315
作者: Xiufeng Yang,Konstantinos Vrettos,Zeineb Ben Chaaben,Fabian Drexel,Jannik Lübberstedt,Johannes Moll,Anirudh Narayanan,Frederike Stumpf,Cosmin I. Bercea,Jiawen Wang,Evangelia E. Vassalou,Krischan Braitsch,Sebastian Ziegelmayer,Markus Graf,Marcus R. Makowski,Lisa Adams,Nassir Navab,Michail E. Klontzas,Keno Bressem
类目: Machine Learning (cs.LG)
*备注: 27 pages, 6 figures; includes Supplementary Appendix (3 figures, 5 tables)
Abstract:Many clinical decisions are a choice among a few fixed options, taken quickly and in large numbers. Such decisions also need an uncertainty estimate to know when to defer to a clinician. Generative language models answer slowly in free text and state their confidence poorly. In contrast, System 1 models answer zero-shot with a probability for each answer. However, their performance across clinical tasks is unknown. Four System 1 models, two proprietary and two with open weights, were evaluated zero-shot on 86 public tasks from 14 clinical domains in 20 languages, spanning examination questions and medical imaging. Tasks were assigned to six capability tiers before analysis. The primary outcome was the area under the receiver-operating-characteristic curve (AUROC). Secondary outcomes included calibration, language effects and comparisons with published results and generative models. Across 179,111 cases, the mean AUROC of the best model per task was 0.94 for medical knowledge, 0.93 for text understanding, 0.83 for clinical judgement from text, 0.89 for image perception, 0.84 for 2D image findings and 0.69 for volumetric findings. Imaging tasks scored 0.09 (95% CI, 0.06 to 0.11) lower than text tasks (P 0.001). The best model matched or exceeded the best published result on 13 of 26 text tasks and on 8 of 35 imaging tasks. Triage nurses outperformed the best model (accuracy, 0.85 vs. 0.56). One decision took 0.34 to 0.44 seconds and 24 to 81 joules on open-weight models run locally. Accuracy on identical questions fell by up to 39 percentage points from English to Yoruba. Generative models of the same family were 11 to 76 times slower, with 3 to 5 percentage points higher balanced accuracy. System 1 models decide quickly, cheaply and with usable confidence when the answer is stated in a text or recalled from medical knowledge, but not when it must be judged or read from an image.
[LG-21] Identifiability from Stationary Covariance in Linear SDEs with Reciprocal Interactions and Latent Variables
链接: https://arxiv.org/abs/2610.31310
作者: Ana Leticia Garcez Vicente,Saber Salehkaleybar
类目: Machine Learning (cs.LG)
*备注:
Abstract:Continuous-time stochastic differential equations (SDEs) provide a natural framework for modeling interacting dynamical systems. A central causal question is whether causal effects among observed variables can be identified from observational data drawn from the stationary distribution, given the underlying causal graph. For linear SDEs, existing identifiability results are largely restricted to fully observed systems whose causal graphs contain no directed 2-cycles. We first establish new graphical conditions for the generic identifiability of causal effects in systems that may contain directed 2-cycles, thereby allowing for reciprocal interactions. Next, we address the previously unexplored setting of latent variables by developing an instrumental-variable framework for identifying a target direct causal effect under partial observability.
[LG-22] Extreme Compression of Per-Layer Embeddings via Cross-Layer Latent Vector Quantization
链接: https://arxiv.org/abs/2610.31309
作者: Jiacheng Liu,Yaxin Luo,Xiaohan Zhao,Peng Sun,Xinyi Shang,Jiacheng Cui,Yonina C. Eldar,Zhiqiang Shen
类目: Machine Learning (cs.LG)
*备注: Code at this https URL
Abstract:Scaling embedding capacity alongside Transformer computation is an emerging trend in language models. This trend makes high compression rates increasingly useful. Hence, we study extreme compression of Per-Layer Embeddings (PLE), which follow this trend and account for 35.2% of the parameters of Gemma 4 E4B. At such rates, conventional quantizers give each token-layer vector its own code of only a few bytes. We instead develop OPTIC (Output-Path-fitted, Token-shared Indexed Codes), which stores one quantized code tuple per token and decodes it at every layer with a layer-specific affine decoder. Codes and decoders are fitted to the PLE output path rather than to the stored vectors. At about 100\times PLE compression of bfloat16 (BF16) E4B, OPTIC retains an aggregate 97.8% of the uncompressed model’s performance across seven benchmarks, whereas product quantization, the strongest of six conventional baselines, retains 66.3%. At 594\times and 1,002\times PLE compression of BF16 E4B, OPTIC retains 90.5% and 84.1%. On the official E4B release with 4-bit linear weights and 16-bit activations (W4A16), OPTIC retains 96.9% at about 100\times PLE compression. During inference at about 100\times PLE compression, BF16 E4B uses 29.7% less GPU memory than with uncompressed PLE at the same key-value cache size, while running at most 0.40% slower.
[LG-23] he Bayesian Learning Rule Beyond KL Geometry
链接: https://arxiv.org/abs/2610.31308
作者: Sophia Sklaviadis,Wu Lin,Andre Martins,Mario Figueiredo,Thomas Moellenhoff,Emtiyaz Khan
类目: Machine Learning (cs.LG)
*备注:
Abstract:The Bayesian learning rule (BLR) exploits the relation between Bayesian updating and the Kullback-Leibler divergence (KLD) to derive efficient additive updates for exponential family variational posteriors. The KL geometry is powerful but restrictive: combined with posterior families with parameter-dependent support it can lead to undefined objectives. We ask which non-KL geometries preserve the additive BLR computation. Previous works that use f -divergences suggest that this is only possible with the KLD. Here, we show that this is not true. We show that additive information-geometric optimization algorithms can be derived for the class of mixed-space Bregman or Fenchel-Young discrepancies. These discrepancies induce generalized maximum-entropy families, dual coordinates, and generalized-BLR canonical gradient updates, recovering the classical BLR. The generalized BLR leads to new algorithms that \textitcannot be derived from the KL-based BLR. For the Tsallis negentropy with quadratic statistics, the generalized BLR yields bounded-support and heavy-tailed q -Gaussian posteriors and an easy Newton-like implementation. Experiments up to CIFAR-100 illustrate controllable support and tail behavior and competitive performance.
[LG-24] DISC: Domain-Integrated Similarity Cost for Time-Series Dataset
链接: https://arxiv.org/abs/2610.31288
作者: Yen-Ku Liu,Hongjie Chen,Franck Dernoncourt,Ryan Rossi
类目: Machine Learning (cs.LG)
*备注: Accepted as a conference paper at IEEE BigData 2026
Abstract:The growth of time-series datasets and foundation models has made dataset comparison a central part of time-series workflows. Assessing source-target transfer before model adaptation requires similarity estimates that can summarize transfer-related dataset relationships without exhaustive source-target evaluation over every candidate pair. Many metrics have been proposed for measuring similarity between time-series datasets, but they capture different dataset relationships and vary in how well they align with downstream utility across tasks. We evaluate this gap by measuring the correlation between source-target similarity scores and downstream forecasting errors in out-of-distribution and transfer-learning settings. We propose Domain-Integrated Similarity Cost (DISC), a learnable aggregation framework that uses calibration transfer observations to select, weight, and combine similarity metrics for different downstream tasks. By integrating complementary metric values, DISC better distinguishes source-target dataset differences that matter for downstream behavior. After calibration, DISC scores source-target pairs using only their similarity-metric values and yields higher alignment than single-metric references on the reported split. These results show that time-series dataset similarity is task-dependent, and that learned metric integration provides a stronger proxy for downstream transfer behavior than manually selecting one similarity metric.
[LG-25] Learning Robot Behavior Priors from Non-Robotic Synthetic Data
链接: https://arxiv.org/abs/2610.31275
作者: Mateusz Odrowaz-Sypniewski,Michael Amir,Jasmine Bayrooti,Ajay Shankar,Amanda Prorok
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Real robot data, rich across tasks and morphologies, is often treated as the bottleneck for training robot learning models, including embodied foundation models that aim for zero-shot physical transfer. In this paper, we investigate the degree to which such models can acquire necessary structure using data from systems without embodiment. To this end, we test if inexpensive, procedurally generated synthetic data can yield useful priors before a model observes any real-world behavior. We study this through robot action prediction, a natural pretraining objective and a direct test of whether structure learned from arbitrary synthetic tasks transfers to embodied settings. We introduce three data generation pipelines that model mathematical abstractions of robotic control, and use them as pretraining priors for a Transformer. We demonstrate that a model trained purely on this non-robotic data attains non-trivial zero-shot prediction performance on unseen physical robot tasks via an in-context learning mechanism, and that our specialized control priors notably improve upon generic causal priors. We also compare against a compute-matched baseline trained on real robot data, and find that it outperforms our model but only on the datasets it was trained on. These findings suggest that procedurally generated, mathematical abstractions can yield useful robotic priors, offering a potential complement to expensive embodied data collection.
[LG-26] Deep ReLU Networks Can Fail the Best Approximation Property for Continuous Target Functions
链接: https://arxiv.org/abs/2610.31264
作者: Sebastian Kassing
类目: Machine Learning (cs.LG); Numerical Analysis (math.NA); Optimization and Control (math.OC)
*备注:
Abstract:We construct a continuous function on [-1,1]^2 for which the infimum of the L^2 -approximation error over the class of ReLU networks with two hidden layers and architecture (2,2,1,1) is not attained. This counterexample shows that deep ReLU networks do not satisfy the best approximation property in general, even for continuous target functions. Our findings contrast with existing results for shallow neural networks, where ReLU is one of the few activation functions that guarantee the existence of global minima in typical regression tasks. The proof is based on the fact that the breaklines for neurons in the second hidden layer can bend, which breaks the symmetry arguments used in the one-hidden-layer case.
[LG-27] QuaNAS: Uncertainty-Aware Reuse of Earlier Searches for Query-Efficient Neural Architecture Search
链接: https://arxiv.org/abs/2610.31251
作者: Brama Seta Putra Satria,Dwi Sunaryono,Misbakhul Munir Irfan Subakti,Irzal Ahmad Sabilla,Laras Suciningtyas
类目: Machine Learning (cs.LG)
*备注: 30 pages, 7 figures, 13 tables. Code and data: this https URL
Abstract:Neural architecture search (NAS) on a new dataset often affords only tens of short training runs, so what the search knows before its first evaluation decides its efficiency. Yet most NAS methods start from scratch, guided at best by zero-cost proxies that are unreliable across tasks. QuaNAS treats a library of earlier searches in the same search space as explicit knowledge of uncertain reliability. Main effects of design decisions learned from this library and, secondarily, zero-cost proxies become priors over candidate architectures without any label of the new dataset. A Gaussian process starts at the prior optimum and re-estimates after every evaluation how far each prior is trusted. Under a protocol frozen before any confirmation data were seen, QuaNAS was evaluated once on 12 new image, time-series, text and tabular datasets. It lowered anytime regret on the short-training objective 2.7- to 3.9-fold relative to five pre-specified NAS baselines that start from scratch (Holm-adjusted p = 0.0012 each). Its final models matched (non-inferiority margin 0.5 pp) or exceeded matched-recipe reference models on nine datasets, as did the prior optimum, selected without a single evaluation. Post-hoc comparisons with established transfer methods given the same library trace the gain to the library: without it, the same Gaussian process had over twice the regret. They also show where each transfer mechanism works best: the pooled main effects of QuaNAS for weakly transferable tabular tasks, per-task surrogates for image, time-series and text tasks and for small or mismatched libraries. Finally, the main effects make the transferred knowledge inspectable.
[LG-28] Muting Tokens: Exploring Token Sparsification for Efficient Audio Transformers
链接: https://arxiv.org/abs/2610.31235
作者: Shreya Sajid,Fabrizio Julian Piva,Hiram Rayo Torres Rodriguez,Elif Ceren Gok Yildirim,Murat Onur Yildirim,Joaquin Vanschoren
类目: ound (cs.SD); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
*备注: This work was conducted in 2024; the related work reflects the literature at that time
Abstract:Transformer-based models have shown strong performance across diverse tasks, but their reliance on self-attention mechanisms often leads to high computational costs. Token sparsification, which reduces the number of tokens processed in transformer layers, offers a promising solution to maintain performance while improving efficiency. However, although widely studied in the vision and language domains, token sparsification remains underexplored in audio, with existing work often limited to a single method, model size, or dataset. This work presents a comprehensive evaluation of token sparsification techniques for audio transformers. We assess a range of token pruning and merging strategies across multiple datasets and model sizes, under both post-training and fine-tuning settings. Furthermore, we introduce a novel sparsification approach that combines attention scores with spectral cues derived from the frequency content of token embeddings. Our experiments show that frequency-aware approaches, particularly those based on spectral flatness and low-frequency energy, consistently outperform attention-only methods. We also find that sparsification based on spectrogram patch intensity is especially effective for longer audio inputs. These results highlight the value of incorporating frequency-domain and input-specific information into sparsification strategies to enhance performance in audio tasks.
[LG-29] Local Discrimination of Binary Classifiers via Localized AUC Functional
链接: https://arxiv.org/abs/2610.31216
作者: Wenye Ren,Anna Győrffy-Kerekes,Tancredi Dorsi,Olga Demler
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Risk prediction models are frequently evaluated with AUC, which means that during evaluation we are blind to performance heterogeneity, which could lead to bad decision-making in clinics. In this work, we aim to change that by identifying patient cohorts for whom the risk prediction model might not be an accurate predictor. To separate these cohorts, we use a version of the weighted AUC, which we call Localized AUC or LAUC. Using this metric, we build a method that we coin LOCATOR. This method is able to identify cohorts of patients for whom the risk prediction model is inaccurate. We test this method both in real-world data and in simulations. We also build a statistical background for the use of Localized AUC for our purposes.
[LG-30] Evolving Federated Density Peak Clustering with Shuffler-Based Anonymization
链接: https://arxiv.org/abs/2610.31211
作者: Miha Ožbot,Seiichi Ozawa,Igor Škrjanc
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Distributed, Parallel, and Cluster Computing (cs.DC); Systems and Control (eess.SY)
*备注: 30 pages, 4 figures. Published open access (CC BY 4.0) in Evolving Systems 17, 123 (2026). Code: this https URL
Abstract:In this paper, we present a novel one-shot federated clustering approach based on evolving fuzzy clustering and shuffler-based anonymization. In federated clustering, each round reveals additional model information to the server and can increase the opportunity to infer properties of local datasets, even when raw samples are never shared. Consequently, fewer client-server exchanges and unlinkability are important, so transmitted information cannot be attributed to any single client. Data-stream clustering algorithms, which learn in a single pass over the data, align well with the one-shot federated learning paradigm, where a single communication round can be sufficient. The proposed Evolving Federated Density Peak Clustering (eFDPC) method combines evolving fuzzy data-stream clustering and density peak clustering in a federated setting. Each client independently evolves micro-clusters from private data mini-batches through cluster addition and merging. To limit communication, density peak clustering is applied to the learned micro-clusters and only the selected peaks are transmitted, rather than the full local models. A shuffler-based layer is introduced between the clients and the server to anonymize contributions prior to reconstruction and global peak selection. We consider a peak-only variant, eFDPC-P, as well as eFDPC-G with server-side label propagation. Experiments on synthetic and real-world benchmark datasets show that eFDPC outperforms recent federated clustering baselines in one-shot settings and scales to higher-dimensional, non-IID data.
[LG-31] owards Field-Deployable Wearable IMU Systems for Occupational Activity Recognition
链接: https://arxiv.org/abs/2610.31160
作者: Ihsan Ul Haq(1 and 2),James Brusey(1),Farnad Nasirzadeh(2),Chandan Karmakar(3) ((1) Centre for Computational Science and Mathematical Modelling, Coventry University, United Kingdom, (2) School of Architecture and Built Environment Deakin University, Australia, (3) School of Information Technology Deakin University, Australia)
类目: Machine Learning (cs.LG)
*备注: Accepted at IEEE BSN 2026
Abstract:Wearable inertial measurement unit (IMU) systems for construction worker human activity recognition (HAR) routinely report aggregate accuracies above 95 %, rarely under genuine subject-independent (leave-one-subject-out (LOSO)) evaluation, masking the per-individual variance that determines real-world deployment performance. Using the VTT-ConIoT benchmark, we conduct, to our knowledge, the first systematic configuration sensitivity analysis of this dataset: six classifiers across five sensor placements and three window sizes under a single LOSO protocol, a personalised-calibration pilot, and a validated comparison of tabular GAN-based and standard time-series augmentation. Per-subject LOSO accuracy for the main multi-sensor, 16-class configuration ranges from 61.9 % to 88.7 % (mean 77.5 %), a 27 pp spread, though all data are laboratory-collected. Multi-sensor fusion and longer windows both improve accuracy; standard augmentation gives small, consistent gains, while Conditional Tabular GAN (CTGAN) augmentation actively harms accuracy with a demonstrated distributional-fidelity mechanism; and personalised calibration recovers 16.8 pp for the weakest participants from just 30 % of their own data. These findings provide evidence-grounded configuration guidance for scaling occupational body sensor networks towards field deployment
[LG-32] Is High Local Accuracy Enough? Geometric Consensus with Support-Adaptive Statistical Personalization for Federated Learning
链接: https://arxiv.org/abs/2610.31133
作者: Junyu Lin,Bosong Qi,Pengcheng Shi,Jiayuan Li
类目: Machine Learning (cs.LG)
*备注: 27 pages, including 9 pages of supplementary material. Code: this https URL
Abstract:Real-world deployment of personalized federated learning (PFL) requires adapting to heterogeneous clients while preserving recognition performance beyond local distributions. However, high local accuracy can conceal failures on locally rare or absent classes. Effective personalization should therefore strengthen global recognition while raising the personalization floor. We propose an evaluation protocol that complements local accuracy with global, rare-class, and unseen-class evaluations to reveal generalization obscured by local scores, and assesses personalization floors among the worst-performing clients to expose weaknesses hidden by aggregate performance. To meet these deployment challenges, we propose FedGRC, a unified PFL framework integrating generalization and personalization through (i) Consensus-Guided Representation Learning (CGRL), which learns discriminative representations through shared directional and magnitude guidance; (ii) Multilayer Metric Synthesis (MMS), which constructs a shared covariance metric from these multilayer representations for statistical classification; and (iii) Support-Adaptive Shrinkage Personalization (SASP), which adapts class means within this metric by shrinking local estimates toward shared means according to local sample support, enabling adaptation while retaining shared knowledge. Extensive experiments, including benchmarking against 21 methods across 5 heterogeneity settings on 3 datasets and comprehensive ablation studies, validate the effectiveness of FedGRC. On CIFAR-10 and CIFAR-100, FedGRC surpasses state-of-the-art (SOTA) baselines in global recognition and personalization floors by at least 2.17% and 2.13%, respectively. On the real-world heterogeneous FedRS dataset, it outperforms the respective best baselines by 2.15%, 3.30%, and 4.61% in local adaptation, global recognition, and the personalization floor.
[LG-33] What Comes Next and Why: Interpretable Next-Event Prediction with Neuro-Symbolic Rules NEURIPS2026
链接: https://arxiv.org/abs/2610.31114
作者: Tim Nico Bauerschmidt,Isabel Valera,Jilles Vreeken
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Accepted at NeurIPS 2026
Abstract:In real-world applications, being able to explain why a prediction is made is often as important - if not more so - than achieving high accuracy. However, existing methods for sequential event prediction typically prioritize accuracy at the expense of interpretability. Neuro-symbolic approaches offer a promising direction to improve interpretability by learning symbolic rules, but extending them to next-event prediction is challenging due to the presence of temporal dependencies. We introduce RULES, a neuro-symbolic model that learns sequential rules directly from data, and combine it with a residual network to form RUNES (Rule Network for Event Sequences). RULES accurately recovers ground truth rules from synthetic data and discovers compact, interpretable rule sets on real data. RUNES combines these rules additively with a residual network, achieving competitive accuracy while providing transparent access to active rules and how residuals adjust predictions.
[LG-34] ransformers with CoT Discover Heavy-Ball Acceleration Through Training
链接: https://arxiv.org/abs/2610.31089
作者: Yuxuan Cai,Yuan Cao
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 110 pages, 3 figures
Abstract:Modern transformers often solve complex tasks by generating and refining intermediate reasoning steps. To develop a theoretical understanding of chain-of-thought (CoT) reasoning, a notable recent study by Huang et al. (2025) viewed transformers as algorithm executors, and showed that when transformers’ CoT steps are supervised by a gradient-descent teacher, they can learn to implement gradient descent for in-context weight prediction in linear regression. In this paper, we study the same in-context weight prediction task but ask whether CoT reasoning can go beyond merely imitating the teacher algorithm. We show that, after initial supervision from a gradient-descent teacher, further trajectory-level supervision using the true regression weights leads the model to implement the Heavy-Ball method, thereby achieving acceleration over the original teacher. Notably, this accelerated algorithm emerges without supervision from any Heavy-Ball teacher. Our results provide a concrete theoretical example in which transformer reasoning can discover new algorithms through training, rather than merely imitate the algorithm used for supervision.
[LG-35] nnInteractive Agent : General-Purpose AI Agents as 3D Image Annotators
链接: https://arxiv.org/abs/2610.31080
作者: Fabian Isensee,Constantin Ulrich Harsy,Maximilian Rokuss,Benjamin Hamm,Philipp Vollmuth,Klaus H. Maier-Hein
类目: Machine Learning (cs.LG)
*备注:
Abstract:Interactive 3D segmentation models such as nnInteractive turn a few spatial prompts (points, scribbles, boxes) into accurate volumetric masks, but a human still does the annotation work: finding each structure, placing prompts, checking the result and correcting it. General-purpose AI agents can now interpret images, call tools and write code. We show that such agents can take over this work without additional training: the user makes a request in plain language, and the agent works out what has to be segmented, locates the targets, prompts nnInteractive, inspects the 3D result, corrects it and responds to feedback. Asked for ‘the muscles used for chewing’, an agent identifies and segments the masseters and pterygoids; asked to count liver lesions, it writes its own code to find candidates and splits a merged lesion on request. Our open-source nninteractive-agent provides nnInteractive as tools over the Model Context Protocol (MCP), together with a skill that encodes how experienced users prompt it. On a benchmark spanning CT, MRI, PET and micro-CT of humans, a mouse and an ant (14 images; 51 targets from organs and tumours to muscles, foot bones, vessel trees and an insect brain), Claude Code and Codex reach a mean Dice of 0.81 and 0.84 from the request alone; when told where each target is, both reach 0.85 and prompt nearly as effectively as an oracle that places prompts using the reference segmentation. A text-prompted 3D model (VoxTell) is more accurate on most standard radiological targets but less robust: it returns an empty mask for one of them and fails on the leg muscles, the foot bones, the mouse and the ant, where the agents remain accurate. Code: this https URL
[LG-36] When Forgotten Facts Return: Relearning-Aware Unlearning of Knowledge Graph Embeddings via Tangent Kernels
链接: https://arxiv.org/abs/2610.31075
作者: Parnian Fazel,Mohammad M Maheri,Hamed Haddadi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Approximate unlearning answers a deletion request on a knowledge-graph embedding by editing entity embeddings instead of retraining from scratch. These embeddings keep being updated as new facts arrive. Under this ordinary continued training, forgotten facts recover more than in the exact-unlearning baseline, a model retrained without them and given the same continued training. We trace this relearning to entity embeddings shared by forgotten and new facts. To our knowledge, this is the first study of relearning in knowledge-graph embeddings. Relearning-robust methods developed for deep networks do not account for the graph structure underlying this recovery. We propose Kernel-Anticipated Unlearning (KAU), which uses the tangent kernel to model how continued training changes forgotten facts’ scores through these shared parameters. KAU expresses this first-order response in closed form and differentiates through it inside the forgetting loss. The objective steers edits toward directions that continued training reverses less. We compare KAU with state-of-the-art knowledge-graph unlearners and relearning-robust methods under identical continued training, measuring leakage beyond the exact-unlearning baseline. Across three backbones, KAU reduces the strongest knowledge-graph unlearner’s leakage by 46-63% on FB15k-237 and 50-82% on CoDEx-L, at higher retained utility.
[LG-37] Federated Differentially Private Optimization via Low-Rank Preconditioning
链接: https://arxiv.org/abs/2610.31069
作者: Antti Koskela,Tejas Kulkarni,Elias Raninen
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
*备注:
Abstract:We consider differentially private (DP) optimization in the federated setting, where a DP estimate of the feature covariance matrix serves as a Hessian approximation for generalized linear models. Existing central DP methods for covariance estimation have communication cost that grows quadratically in the data dimension, while existing baselines with linear communication cost are only locally private and degrade as the number of clients increases. As a remedy, we propose a federated DP-PCA framework that achieves linear communication cost while providing central DP guarantees under secure aggregation. It consists of a DP rank selection step, a DP power method, and a noisy release of the covariance projected onto the selected low-dimensional subspace. We give error bounds for this covariance estimate and for the resulting low-communication preconditioned DP gradient descent. Empirically, our proposed DP optimization method matches a centralized baseline with quadratic communication cost at a fraction of the communication.
[LG-38] NN-Based Transistor Modeling for Ultra-Fast SPICE Simulation and Standard Cell Library Characterization
链接: https://arxiv.org/abs/2610.31024
作者: Rodion Novkin,Hussam Amrouch
类目: Emerging Technologies (cs.ET); Machine Learning (cs.LG)
*备注: 11 pages
Abstract:Neural network (NN)-based transistor models have proven to be a promising solution to accelerate device modeling. Although these models demonstrated remarkable speedup in circuit simulations, they were not rigorously tested in complex tasks within standard EDA tool flows for chip design, which require very high transistor model precision to provide accurate results and converge properly. To investigate their full potential and promise, we develop NN-based transistor models that accurately predict the drain current and various charges for a 3nm nanosheet, and employ the models for a standard cell library characterization. This enables us to evaluate the accuracy and speedup under a very wide range of SPICE simulations, benchmarking our models against an industry-standard implementation. Since standard cell library characterization involves thousands of SPICE simulations, it serves as a great benchmark for transistor model quality and consistency. To further evaluate model accuracy, we also perform analog- and digital-level simulations for various complex circuits. Our experiments reveal that NN-based models are a valid alternative to existing transistor compact models, capable of producing accurate delay and power estimates with a much higher speed and convergence time compared to the conventional Verilog-A-based industry-standard (BSIM-CMG) transistor compact model. Our NNs achieve sub-0.11% and 0.76% errors for delay and total power evaluation on a wide range of circuits, while taking up to 8 times less time during the standard cell library characterization. Additionally, we provide comprehensive information and guidance on NN accuracy, unveiling the relationship between NN size and its effect on circuit simulation precision.
[LG-39] XALIGN: A Human-centric Visual Analytics Approach to Understanding AI Explanations
链接: https://arxiv.org/abs/2610.31015
作者: Bahavathy Kathirgamanathan,Gennady Andrienko,Natalia Andrienko,Tatiana von Landesberger
类目: Machine Learning (cs.LG)
*备注:
Abstract:Explainable AI (XAI) methods are usually evaluated for technical properties such as faithfulness and robustness. In high-stakes application domains, experts also need to check model predictions against established knowledge. Different explanation methods often disagree, and this disagreement is hard to see when each method is inspected separately. We propose XALIGN, a visual analytics workflow and interface that shows rule-based explanations and several feature attribution methods (TreeSHAP, KernelSHAP, and LIME) side by side for the same Random Forest model. Agreements and contradictions between the methods become visible signals that users can explore and compare with domain knowledge. We apply XALIGN in two case studies: the German Credit dataset and a meteorological dataset, in which events derived from ICON model output and Cloudnet observations are classified. In both, the views reveal discrepancies between the rule-based explanation and the attribution methods that a single method would not show. In an expert evaluation with three specialists, participants found XALIGN useful for identifying influential features and for avoiding over-trust in predictions. Linking rules to attributions was harder for the participant without an XAI background.
[LG-40] EEG-Captioner: Grounding EEG-to-Language Generation in Fundamental Signal Properties ICASSP2027
链接: https://arxiv.org/abs/2610.30995
作者: Leyuan Jin,Jiahao Mei,Yixuan Yuan,Shimeng Li,Mengyue Wu,Yue Ding
类目: Machine Learning (cs.LG)
*备注: 5 pages, 2 figures, 3 tables. Submitted to ICASSP 2027
Abstract:Electroencephalography (EEG)-to-language models offer a promising approach for translating complex neural signals into human-readable descriptions. However, generating plausible language does not necessarily demonstrate that a model has learned to interpret the underlying EEG signal. A direct way to assess such interpretation is to probe objectively measurable properties of the signal itself. We therefore construct an EEG-language dataset of 160,781 segments annotated with signal-level question answering (QA) pairs, where each question targets one of five property families and requires a quantitative comparison involving spectral, spatial, temporal, or lateral signal properties. Each segment is also paired with a feature-grounded EEG caption. We further propose a Dual-Domain EEG Encoder that jointly models temporal waveforms and spectral-spatial representations, and introduce EEG-Captioner, a unified EEG-to-language framework trained with both signal-level QA and EEG caption supervision. Experiments show that the proposed encoder achieves the highest overall macro-averaged balanced accuracy (Macro-BA) among the compared EEG encoders, while EEG-Captioner achieves a Macro-BA of 88.81% on signal-level QA and improves caption factual consistency from 82.06% to 83.46% over caption-only language adaptation, demonstrating the complementary benefits of signal-level QA and caption supervision.
[LG-41] Energy and Latency First: How Deployment Constraints Reverse Time-Series Model Selection
链接: https://arxiv.org/abs/2610.30965
作者: Shruti Bothe,Illyyne Saffar,Ayush K. Varshney
类目: Machine Learning (cs.LG)
*备注: 13 pages, 1 figure
Abstract:AI models, and time-series models in particular are typically ranked by accuracy and deployed under constraints absent from that ranking. In this paper, we study a real setting where these constraints are explicit and measurable: handover pre- diction in cellular networks, where inference runs on the User Equipment once per 40 ms measurement period and must complete before the network acts. The timing bound is hard: a late prediction is superseded by the measurement it was predicting, while energy is a cumulative budget spent for the life of the device. Neither appears in the objective these models are selected on. We evaluate eight architectures under a protocol that screens for deadline feasibility at the tail rather than the mean, then ranks the feasible models on measured inference energy, count- ing the feature computation that precedes the forward pass and is usually omitted. Gradient-boosted trees on handcrafted temporal features are the most accurate model (F1=0.974) and among the cheapest to run, while the neural architecture an accuracy-only benchmark would rank first (CNN-1D, F1=0.914) costs over 3 times more energy per inference under a realistic duty-cycled deployment pattern. We give three conditions under which this outcome should be expected, so the finding applies beyond telecommunications. Accuracy-only benchmarking does not merely under-report deployment cost. It can rank first a model that cannot be deployed.
[LG-42] ReDence: A Hybrid Agent ic and Decision-Model Framework for Rapid Evidence-Grounded Detection of Modern Social Engineering Scams
链接: https://arxiv.org/abs/2610.30963
作者: Jaishnoor Kaur,Josef Horáček,Ian Waskom,Sayak Saha Roy
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Detecting online social engineering (SE) scams often requires examining website workflows, verifying claims against external sources, and connecting evidence across multiple pages. Existing detectors provide limited support for these investigations, leaving analysts to manually gather and validate evidence, which inevitably delays detection and subsequent takedown of these attacks. To address this gap, we present ReDence, an open-source hybrid framework for adaptive, evidence-grounded detection across diverse SE scam categories. ReDence combines two locally hosted models with complementary responsibilities: a fine-tuned multimodal agentic Investigator (Qwen3.5-4B) that performs investigative planning, evidence interpretation, and reporting, and a specialized Decider (Tev1) that handles binary and multiple-choice tasks, including browser-action selection, evidence assessment, and decisions about whether to continue an investigation. By delegating discrete decisions to a specialized model, ReDence reduces reliance on lengthy language-model responses, while preserving the reasoning capabilities required for complex investigations. Our evaluation demonstrates that ReDence achieves 96% accuracy in detecting scams across twelve SE scam categories, substantially outperforming existing detectors while being at least four times faster than comparable agentic models without compromising investigative depth. ReDence also produces structured, evidence-grounded reports that document investigative actions, supporting and contradictory evidence, and collected artifacts, which allow analysts to independently audit detection decisions and utilize the collected evidence to support subsequent mitigation and takedown efforts.
[LG-43] Doubly Robust Distributionally Robust Offline Contextual Pricing
链接: https://arxiv.org/abs/2610.30957
作者: Min Xu,Xinyi Yin,Yunfan Zhang,Yuxuan Han,Houcai Shen,Caihua Chen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Offline contextual pricing often relies on logged observational data, but faces challenges from distributional shifts between training and deployment environments. Distributionally robust optimization (DRO) provides a principled approach to off-policy evaluation and learning (OPE/L). However, existing methods are mostly limited to discrete actions. Recent work has explored DRO for continuous treatments using inverse propensity weighting (IPW), while such IPW-based estimators can be sensitive to the convergence rate of propensity score estimates, particularly when estimated nonparametrically, which may lead to larger estimation errors and regret. In this work, we develop a doubly robust (DR) framework for distributionally robust OPE/L in continuous pricing settings. For evaluation, we propose a localized DR estimator that addresses the computational challenges of worst-case expectations by fitting only a small number of regressions, comparable to standard non-robust DR, while achieving semiparametric efficiency under mild product rate conditions. For learning, we leverage the inherent smoothness of demand noise to handle pricing-specific discontinuities in revenue outcomes (e.g., threshold-based purchase decisions), establishing a finite-sample regret bound of \tilde\mathcalO_p(T^-s/(2s+1)) for smoothness orders s=1,2 . This bound improves upon existing regret rates in existing DRO-based off-policy learning (OPL) for continuous treatments. Extensive experiments under various levels of distribution shift validate our proposed framework.
[LG-44] Ground-Truth-First Benchmarking of Reinforcement Learning via Converse Optimality: From Smooth Fully Observed Systems to Non-Smooth or Partially Observed Tasks
链接: https://arxiv.org/abs/2610.30954
作者: Grégoire Ouerdane,Sinan Ibrahim,Stefan Streif,Pavel Osinenko
类目: Machine Learning (cs.LG)
*备注:
Abstract:Benchmarking in reinforcement learning is usually relative: algorithms are ranked by returns on environments with unknown optima. Distance to optimality remains unobservable, while conclusions may depend on reward design, stochasticity, hyperparameters, environment selection, and simulator details. We develop a ground-truth-first framework based on converse optimality, constructing each task with a certified optimal value function and policy withheld from learners and used only for evaluation. The framework unifies two modes: sensitivity benchmarking samples certified environments from continuous-time generators to assess robustness to task variation, while oracle-referenced comparison uses native discrete-time generators to evaluate policies via optimality gap, regret, and policy error. Continuous-time models support physically interpretable sensitivity studies but introduce dependence on the simulator step ( \Delta t ) and numerical scheme. Native discrete-time models align the environment, oracle, and learner on the same transition law, eliminating external integration choices as a confounder. We reproduce quantitative results covering optimizer sensitivity and oracle-referenced comparisons on serial planar arms, dynamically extended nonholonomic vehicles, and quadrotor-inspired tasks. The methodology extends beyond smooth affine Gaussian systems: virtual-input pullback handles non-affine raw-action interfaces and finite-covariance non-Gaussian disturbances. A non-smooth converse optimality formulation enables certified benchmarks with kinked value functions and switching surfaces. For partially observed evaluation, an information-state formulation reduces benchmarking to converse optimality on a sufficient statistic and provides an information-penalty reference. Code, fixtures, and diagnostics are available on the public project page.
[LG-45] Random Initialization as a Baseline for Time Series Foundation Models NEURIPS-26
链接: https://arxiv.org/abs/2610.30950
作者: Pinar Sungu Isiacik,Georgiana Ifrim
类目: Machine Learning (cs.LG)
*备注: Accepted by the NeurIPS-26 Workshop on TS-LIMITS
Abstract:Time series foundation models (TSFMs) use self-supervised pretraining to learn reusable representations for downstream tasks such as multivariate time series classification (MTSC). We examine how much pretrained weights contribute beyond the architecture itself by comparing seven publicly released instances from four TSFM families with architecture-matched randomly initialized counterparts on the UEA MTSC archive. For each instance, five random initializations are evaluated using four classifiers on frozen representations: ridge classification, logistic regression, RBF SVM, and random forest. After model-specific exclusions due to computational constraints, the study comprises 183 paired dataset-level comparisons. Pretraining improves mean accuracy for all seven instances, but the gains are modest, ranging from 1.43 to 5.19 percentage points, with an overall average improvement of 2.58 points. Three of seven per-instance Wilcoxon signed-rank tests have raw p -values below 0.05, but only NuTime remains significant after Holm correction. A secondary pooled test across all 183 pairs gives p 0.001 , while the corrected per-instance tests remain the primary evidence. We also compare with ROCKET, Hydra, and Quant, which remain competitive without requiring large-scale pretraining. These findings suggest that MTSC evaluations should distinguish the effect of pretrained weights from that of the underlying architecture and should include both architecture-matched random initialization and strong non-pretrained baselines.
[LG-46] Bridging Number-Theoretic Primality Testing And Algorithmic Optimization: An Adaptive Cryptographic Framework
链接: https://arxiv.org/abs/2610.30935
作者: Tran Minh Thanh,Vo Hoai Viet
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:The security of modern public-key cryptographic systems, such as RSA, Diffie-Hellman, and ECC, depends entirely on the computational intractability of the integer factorization and discrete logarithm problems. For these asymmetric algorithms to operate, a core requirement is that the system must be capable of generating, testing, and simultaneously proving the primality of extraordinarily large numbers. To meet those requirements, this study proposes a comprehensive framework, bridging clas- sical number theory with modern machine learning models and optimization algorithms. Specifically, the paper delineates an adaptive, arbitrary-precision architecture dedicated exclusively to primality testing. Concurrently, the research deeply analyzes the theoretical foundations of testing methods (both deterministic and probabilistic), hardware-level optimization techniques, and the application of quantum-secured entropy sources, while providing rigorous mathematical proofs to guarantee security even when the prime distribution is biased.
[LG-47] Same molecule Different Knowledge: Cross-Identifier Binding Failures in LLM s For chemistry
链接: https://arxiv.org/abs/2610.30903
作者: Khouloud Saadi,Di Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Chemical AI workflows routinely represent the same molecule through multiple co-referring identifiers, including trivial names, IUPAC names, and SMILES, implicitly assuming that these representations provide interchangeable access to an LLM’s knowledge. To test this assumption, we first introduce a controlled benchmark and show that it does not hold. Across foundation models spanning 7B to 72B parameters, even chemistry-specialized models, co-referring molecular identifiers consistently expose different subsets of the model’s knowledge. We then conduct a series of controlled behavioral, representational, and causal analyses to characterize this failure. We show that even when a model can identify a molecule from its SMILES and retrieve a fact through the corresponding name, the same fact can remain inaccessible directly from SMILES. We term this phenomenon \emphcross-identifier binding failure. Further analysis of internal representations, conflicting identifiers, and parameter editing shows that neither improved chemical representation competence nor modification of individual facts is sufficient to establish reliable access across identifiers. Together, these results indicate that the core failure lies in the binding between an identifier and knowledge already present in the model. Motivated by this diagnosis, we introduce Binding Self-Distillation, a simple mitigation that uses lexical identity during training to transfer the model’s existing knowledge access to the structural route. This enables factual retrieval directly from SMILES at inference time, without requiring a name lookup or training the underlying model from scratch.
[LG-48] Equilibria and Learning in Competitive Market Making
链接: https://arxiv.org/abs/2610.30891
作者: Matteo Onger,Luigi Foscari
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG); Trading and Market Microstructure (q-fin.TR)
*备注: 35 pages, 4 figures
Abstract:We propose a market-making game played among N \ge 2 strategic market makers and a single non-strategic representative trader. In a continuous quote space, standard finite-game arguments do not apply, while non-convex and discontinuous payoffs rule out classical continuous existence results and complicate learning. Nonetheless, we characterize pure-strategy Nash equilibria (PSNEs) and show that at least one PSNE always exists under minimal assumptions on the underlying asset value distribution, and that the equilibrium bid-ask spread widens strictly with asset value volatility. To obtain a tractable learning framework, we discretize prices and, under an additional regularity condition, show that the time-averaged play of no-internal-regret learners approaches the continuous Nash price envelope as the tick size vanishes. Finally, we empirically validate our theoretical results through numerical simulations.
[LG-49] Building Presentation Intelligence for High-Quality Slide Generation
链接: https://arxiv.org/abs/2610.30888
作者: Yuhang Xie,Cheng Li,Chenliang Li,Mingyu Xu,Mingwei Sun,Suhang Zheng,Ming Yan,Xiaoxiao Xu,Lin Qu
类目: Machine Learning (cs.LG)
*备注: 28 pages
Abstract:PowerPoint presentations (PPTs) are a dominant medium for communication across diverse domains, yet high-quality PPT generation still depends heavily on labor-intensive human effort. Recent advances in large language models (LLMs) for agentic tasks have created both new opportunities and challenges (Chen et al., 2026) for automatic PPT generation. In this work, we present a unified training framework for automatic PPT generation toward presentation intelligence, integrating cross-validated data synthesis, guided teacher enhancement, and verifiable reward optimization. We first construct a cross-validated pipeline for data synthesis, with explicit control over both difficulty and validity. We then develop a teacher-enhancement framework grounded in rules and prior knowledge to substantially strengthen the instructional capacity of teacher models. Building on this, we further improve supervised fine-tuning (SFT) through a mixed-teacher strategy and a lightweight outcome-driven trajectory filtering method. Finally, based on the resulting SFT models, we introduce a suite of verifiable aesthetic metrics together with a GRPO-based reinforcement learning framework to optimize slide-level visual attributes beyond text-only supervision. Experiments show that our enhanced SFT pipeline improves Qwen-3.6-27B and Qwen-3.6-35B-A3B by 7.45 and 9.35 points on PresentBench, respectively, surpassing GLM-5.2 while maintaining strong generalization. Reinforcement learning further improves performance across five aesthetic dimensions, compensating for the limitations of SFT in visual refinement. Overall, the results demonstrate the effectiveness of our framework for high-quality PPT generation.
[LG-50] racing Subliminal Learning Through Pretraining
链接: https://arxiv.org/abs/2610.30884
作者: Angela Tang,Vishesh Anand,Gonçalo Paulo
类目: Machine Learning (cs.LG)
*备注:
Abstract:Large language models trained on data generated by other models can sometimes inherit traits unrelated to that data, a phenomenon called subliminal learning. We study both hard label distillation, where students train on numbers generated by a biased teacher, and on-policy soft label distillation, where students generate the data and the teacher’s logits provide the learning signal. Using Pythia, PolyPythia, and DataDecide OLMo models, we find that transfer persists in both settings across differences in weight initialization, pretraining data order, pretraining corpus, model size, and architecture. Across pretraining checkpoints, the ability to transmit a preference and the ability to acquire one develop separately: some checkpoints are more effective teachers but weaker students, and vice versa. Consistent with this distinction, cross-model transfer can be strongly asymmetric in magnitude when teacher and student roles are reversed. To investigate what carries the preference signal, we construct synthetic number sequences that match the teacher’s generated number frequencies but have no dependence on the prompt or preceding numbers and find that these still induce substantial transfer in the hard-label setting. In top- K soft-label distillation, removing animal tokens from the loss reduces transfer, although not completely in the same-model setting. In our experiments, soft label distillation produces more consistent transfer with less extensive training than hard label distillation. The growing use of on-policy distillation in model post-training makes this a particularly relevant setting for assessing unintended preference transfer.
[LG-51] Joules-to-GDP: How Much Economic Value Do AI Systems Produce per Unit of Energy?
链接: https://arxiv.org/abs/2610.30858
作者: Amirreza Zeinali,Jon Saad-Falcon,Avanika Narayan,Mersad Abbasi,Jonathan S. Hartley,John Hennessy,Christopher Re,Azalia Mirhoseini
类目: Machine Learning (cs.LG); General Economics (econ.GN)
*备注:
Abstract:An AI system can produce more economically valuable work while delivering less value for the energy it consumes. We therefore ask: How much economic value does an AI system create per unit of energy consumed? We formalize economic value in two ways: (1) net dollar value, the output’s dollar value minus AI-system execution cost, and (2) time saving, the expert’s completion time minus the AI system’s completion time. Dividing each by measured end-to-end energy yields Net Value per Joule (NVJ) and Time Saving per Joule (TSJ), reported in USD/MJ and hours/MJ. We study 220 public GDPval tasks across 44 occupations, serving five open-weight models locally and evaluating four accelerator types. First, the occupation groups where AI systems succeed most often, deliver the greatest dollar value, or save the most time do not necessarily provide the greatest economic return per joule. For example, among the tested tasks, engineering work delivers about 10% more net dollar value per task than software development. Yet software development delivers about 27% more net value per joule. Second, GLM 5.2 delivers the highest average net dollar value, at 72.27 per task using 0.430 MJ, equivalent in energy to about four hours of active laptop use. By contrast, Gemma 4 31B delivers 56% of that net dollar value using just 23% of the energy, achieving approximately 2.5 times the NVJ and 2.8 times the TSJ. Third, across four single-accelerator deployments, Qwen3.6-27B’s NVJ varies by a factor of 4.32 and TSJ by 9.27, with B200 leading both. Additionally, one and eight H100s deliver similar net dollar value, but eight reduce completion time by 22% at 3.24 times the energy. Our framework makes output value and expert time saving under an energy budget explicit objectives for improving models, accelerators, and software, while offering a new basis for assessing AI’s potential economic effect on occupational work.
[LG-52] raining Mixed-Monotone Neural Networks for Verification Without Set Propagation
链接: https://arxiv.org/abs/2610.30853
作者: Jonas Fingerhut,Lukas Koller,Matthias Althoff
类目: Machine Learning (cs.LG)
*备注: 20 pages, 3 figures, 9 tables
Abstract:Most formal verification approaches for neural networks conservatively enclose the output set of the neural network. However, computing a tight output enclosure is in general computationally hard. We sidestep this challenge with a novel neural network architecture that realizes mixed-monotone functions as the difference of monotone subnetworks. By utilizing the theory on mixed-monotone systems, this reduces the computation of a tight output enclosure from complicated set propagation to two simple forward passes of opposite corners of an input interval. Moreover, because the size of the resulting enclosure is a closed-form function of the weights of the neural network, we can directly minimize it during training. We prove a lower bound on the size of any such enclosure, given by the total variation of the represented function generalized to several inputs, which makes it possible to compare the tightness of the enclosure against the optimum. Our evaluation on regression and dynamical-system benchmarks demonstrates that our training reduces the enclosure size to within 0.5% of the lower bound without sacrificing prediction accuracy. Computing our enclosure is orders of magnitude faster than CROWN and \alpha -CROWN and on deep networks it is also tighter than zonotope propagation and CROWN, while its size exceeds that of \alpha -CROWN by only 5.1%.
[LG-53] Reverse-KL Policy Distillation: Statistical Guarantees and Limits
链接: https://arxiv.org/abs/2610.30844
作者: Haolin Liu,Chen-Yu Wei,Julian Zimmert
类目: Machine Learning (cs.LG)
*备注:
Abstract:Reverse Kullback-Leibler (KL) distillation has shown empirical promise for training language models, but its statistical guarantees and limits remain poorly understood. In particular, the effects of two design choices remain unclear: Does access to the teacher’s full action distribution provide an advantage over observing only the probability it assigns to the sampled action? Must training trajectories come from the student (online distillation), or can teacher-generated trajectories (offline distillation) suffice? Our work takes a first step toward understanding the theory underlying these questions. We study the sample complexity of online reverse-KL distillation under two feedback models: the teacher reveals either its full action distribution or only the probability of the student-generated action. We establish minimax bounds for the former and show that substantially more data may be required for the latter. In contrast, offline reverse-KL distillation can attain the same minimax rate whether the teacher reveals its full distribution or only its probability of the teacher-generated action. The key distinction is that teacher-generated actions themselves reveal the teacher’s preferences, while queries at student-selected actions can leave those preferences unobserved. This highlights an additional statistical challenge for online distillation compared to offline distillation. These statistical guarantees do not follow simply from optimizing standard distillation objectives: our lower bounds show that common online objectives can fail when candidates are evaluated only on past student-generated trajectories. We develop trajectory-loss and importance-ratio truncation techniques together with policy-space learning to achieve optimal online distillation guarantees. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2610.30844 [cs.LG] (or arXiv:2610.30844v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.30844 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-54] Is spectral predictability enough? Testing Ω-matching for synthetic pretraining under distribution shift ICASSP2027
链接: https://arxiv.org/abs/2610.30802
作者: Edric Castel C. Hao,Juan Raphael S. Seña,Neon S. De Jesus
类目: Machine Learning (cs.LG)
*备注: Submitted to ICASSP 2027
Abstract:Synthetic pretraining under distribution shift requires deciding what synthetic data to use. A natural criterion is to match the spectral predictability \Omega of the source to the target, but is \Omega sufficient? We test this hypothesis, which we call predictability-aware pretraining, across three architectures (LSTM, PatchTST, TCN) and three seeds. Under ideal synthetic-to-synthetic conditions, \Omega -matched pretraining can outperform even an oracle baseline (+46.8% MSE reduction, LSTM), though the benefit is already architecture-dependent. On the real-world ETTm1 benchmark, however, this relationship does not hold: sine-plus-noise pretraining at matched \Omega yields only a marginal mean improvement +2.3% averaged over all three architectures, up to +65.5% on individual channels), whereas \Omega matched chaotic ODE trajectories systematically degrade performance -13.2%). The two sources agree on the direction of transfer for only 53% of configurations across all three architectures (Pearson r=-0.02), indicating that \Omega carries little information about which source will transfer beneficially. We further observe an architecture-dependent asymmetry in the response to the direction of the \Omega gap. We conclude that \Omega matching can help in controlled settings but is not sufficient to predict transfer across sources, and we discuss additional signal properties that a source-selection criterion may need to consider.
[LG-55] A Computationally Efficient and Rate-Optimal Algorithm for Adversarial Linear MDPs with (Aggregate) Bandit Feedback
链接: https://arxiv.org/abs/2610.30801
作者: Haolin Liu,Chen-Yu Wei,Hanbin Zhou
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study episodic linear Markov decision processes with unknown transitions, adversarial losses, and bandit feedback. The state-of-the-art result for this setting (Liu et al., 2024) either achieves \sqrtK regret with a computationally inefficient algorithm or achieves suboptimal K^3/4 regret, where K is the number of episodes. Our work gives the first \sqrtK regret with a polynomial-time algorithm. Our method reduces adversarial linear MDPs to adversarial linear bandits whose action set consists of policies’ expected feature vectors. The central challenge is that this action set is unknown and can be estimated only through exploration. We address this challenge by constructing an approximate separation oracle for the action set that either generates executable policies from queried features, or refines an outer approximation of the action set. This approach draws inspiration from the reduction of van Erven et al. (2025) for linear contextual bandits and the oracle-efficient linear bandit algorithm of Ito et al. (2019), but additionally deals with the imprecision of the oracle. The resulting algorithm requires only aggregate loss feedback, and to our knowledge. This is the first polynomial-time algorithm for adversarial MDPs with aggregate feedback beyond the tabular case.
[LG-56] Amplifying Latency and Cost of LLM Inference via Rowhammer Attack
链接: https://arxiv.org/abs/2610.30798
作者: Mrityunjay Shukla,Shubham Roy,Biswabandan Panda
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Large language models (LLMs) achieve high inference efficiency using optimized GPU kernels generated by just-in-time (JIT) compilation frameworks such as Triton. Triton compiles GPU kernels through an MLIR/LLVM-based pipeline. We investigate whether a Rowhammer-induced bit-flip in the CPU’s DRAM can degrade GPU-based LLM inference performance while preserving output correctness. We present LLVMhammer, the first Rowhammer attack that causes LLM inference slowdown by inducing a bit-flip in Triton’s compilation pipeline. A bit-flip in a performance-critical optimization pass causes Triton to generate a functionally identical but suboptimal GPU kernel. Since Triton caches the generated kernel on disk, a single successful bit-flip causes persistent slowdown across subsequent inference executions. Triton’s target-independent stages are shared across GPU vendors, which makes our attack GPU vendor-agnostic. To identify exploitable bit-flip locations without manual inspection, we build Karnage, a tool that isolates useful bit-flip locations in Triton’s target-independent compilation pipeline. We evaluate LLVMHammer across a diverse set of LLMs, including Qwen2.5-7B-Instruct, Llama-3.1-8B-Instruct, Mistral-7B-Instruct-v0.3, and Deepseek-llm-7B-chat, across NVIDIA RTX 4090 and AMD RX 7900 XTX GPUs. LLVMhammer increases inference latency of an LLM inference server running on a single GPU by 4-16%, which, in turn, can translate into additional annual operating costs at the scale of multi-GPU deployments. We further show that LLVMhammer can also corrupt LLM outputs across different GPU vendors, reducing BERTScore to as low as 25%, illustrating the broader security implications of bit-flips in Triton’s compilation pipeline.
[LG-57] GIFT: Gradient Informed Filtering for Distillation Augmented Reasoning RL
链接: https://arxiv.org/abs/2610.30796
作者: Shijie Li,Yilin Gao,Tieyuan Chen,Yun Luo,Weiyao Lin,Hang Yu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Training large language models for complex reasoning tasks relies on two complementary signals: sparse, outcome-grounded reinforcement learning (RL) for correctness, and dense, token-level distillation for guidance. However, complementarity in supervision does not guarantee compatibility in optimization; imitation-focused distillation can conflict with the outcome-driven RL objective, and existing hybrids lack a principled resolution. We propose \textbfGIFT (\textbfGradient-\textbfInformed \textbfFil\textbfTering for Distillation-Augmented Reasoning RL), a novel method that directly addresses the allocation problem of which teacher updates to apply. GIFT performs a token-level geometric analysis, decomposing the distillation gradient into a component \textbfparallel to the RL gradient, which measures first-order conflict and provides a \textbfsafety filter, and a component \textbforthogonal to it, which quantifies novel directional guidance and provides a measure of \textbfvalue. This decomposition yields a principled scoring rule to rank token-level guidance. We combine this with a progressive \textbfcoverage annealing schedule that shrinks the budget of tokens receiving distillation, ensuring that the most conflicting updates are discarded first while the most complementary ones are retained at full strength. This facilitates a smooth transition from dense, geometrically-filtered guidance to a pure, outcome-grounded objective at convergence. Extensive experiments on mathematical and logical reasoning benchmarks show that GIFT consistently outperforms strong baselines across multiple model scales, achieving relative gains of up to 9.5% on difficult mathematical reasoning tasks. Code: this https URL
[LG-58] race2World: Reconstructing App Worlds from Real Interaction Traces for Scalable GUI Agent Training
链接: https://arxiv.org/abs/2610.30792
作者: Zezhou Wang,Hanzhang Zhou,Panrong Tong,Quyu Kong,Pengxiang Li,Steven Hoi,Zhuzhong Qian,Yue Wang
类目: Machine Learning (cs.LG); Software Engineering (cs.SE)
*备注: Preprint. Under review
Abstract:Scaling GUI reinforcement learning requires environments that preserve real-app behavior. Yet coupling behavior inference with coding can turn partial traces into runnable mocks with incorrect state transitions. We introduce \textbfTrace2World, which separates global behavior reconstruction from implementation. It reconciles traces into a shared, source-linked app model, then compiles an executable environment. Independent replay of source workflows localizes mismatches for bounded implementation repair. Validated worlds then support learner-adaptive task and environment augmentation within source-supported behavior. From 1,570 traces, Trace2World constructs 50 environments. On 200 workflows across ten apps, 91.1% are reproduced end to end, versus 59.8% for the strongest matched task-local baseline. Source-grounded validation retains 94.0% of simulated training gains on real devices, versus 54.7% under self-validation. Learner guidance raises benchmark success from 56.9% to 61.1%. Training in Trace2World further improves the model by 20.2 points on our benchmark, 5.4–16.2 points across four external GUI benchmarks, and 16.3 points on 100 real-device tasks. Reliable reconstruction separates what behavior source evidence supports from whether the implementation preserves it.
[LG-59] Gandalf: Flash Shall Not Die! Increasing Flash Storag e Lifetime for KV Cache Offloading in LLM Serving
链接: https://arxiv.org/abs/2610.30791
作者: Maharani A. P. Irawan,Kuntai Du,Jiayi Yao,Dongjoo Seo,Yuyang Huang,Haryadi S. Gunawi,Junchen Jiang
类目: Machine Learning (cs.LG)
*备注:
Abstract:KV caching reduces the serving cost of large language models (LLMs), but the growing KV cache footprint is increasingly exceeds GPU and host-memory capacity. Therefore, people use SSD to expand the KV cache storage capacity with high I/O bandwidth. However, in production LLM workload, such as multi-round conversation and LLM agents, the total amount of KV cache writes significantly exceeds SSD capacity, shortening SSD lifetime from years to months. To improve SSD lifetime, a natural solution is to reduce the KV cache writes by gating the KV cache writes via historical access patterns. However, that significantly lowers the KV cache hit rate (and increase the LLM serving cost) since KV caches from different attention types might have identical historical access pattern but different future access patterns. Our insight is that, to reduce KV cache writes without hurting KV cache hit rate, one need to not only consider the access pattern, but also application-level information such as the attention type of KV cache. Leveraging this observation, we propose Gandalf , the first SSD KV cache offloading system with SSD life time awareness. Gandalf prioritize the admission of full-attention KV cache into SSD over the cache generated by other types of attention, which significantly increases SSD lifetime without sacrificing cache hit rate. Our evaluation shows that Gandalf extends the SSD lifetime by 7.3X on average, while having negligible impact on cache hit rate across multiple workload traces and wide range of hardware capacity. Gandalf design can be extended to other flash-based storage, such as High Bandwidth Flash (HBF).
[LG-60] One Pretrained Decision Model Is Enough for Training-Free Generalist Graph Anomaly Detection
链接: https://arxiv.org/abs/2610.30777
作者: Xinyu Zhao,Qingyun Sun,Ziwei Zhang,Tianchen Zhu,Daqing Li,Jianxin Li
类目: Machine Learning (cs.LG)
*备注:
Abstract:Graph anomaly detection (GAD) aims to identify irregular entities in graph-structured data, yet heterogeneous anomaly patterns and distributional differences across graphs challenge generalizable detection. Existing training-free approaches avoid task-specific optimization but primarily rely on predefined numerical scoring rules, leaving the semantic interpretation of heterogeneous anomaly evidence largely unexplored. In this paper, we propose Jev-GAD, a training-free generalist graph anomaly detection framework that bridges heterogeneous numerical evidence and pretrained decision mechanisms. To address heterogeneous anomaly patterns, we construct complementary structural, attribute, and neighborhood evidence and transform heterogeneous statistics into graph-relative anomaly percentiles. To move beyond predefined numerical scoring rules, we leverage a single frozen pretrained language model for evidence-conditioned anomaly assessment. To bridge the semantic gap between numerical evidence and pretrained models, we design a semantic Graph-to-State (G2S) transformation that organizes graph context and semantically labeled node evidence into structured decision states for non-generative option-logit readout. The resulting decision scores are integrated with four complementary statistical scores through equal-weight rank-based fusion, without GAD-specific gradient-based optimization. Extensive experiments on seven graph anomaly detection benchmarks spanning citation, social, and co-review networks demonstrate the effectiveness of Jev-GAD, achieving average AUROC and AUPRC scores of 89.18% and 52.40%, respectively. Further analyses validate the contributions of pretrained decision readout and semantic evidence grounding.
[LG-61] Adaptive Deadline- and Energy-Aware Caching Algorithms in Dynamic Multi-UAV Edge Networks
链接: https://arxiv.org/abs/2610.30768
作者: Jian Kim,Ki-Il Kim
类目: Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注: 18 pages, 7 figures
Abstract:Content caching via unmanned aerial vehicles (UAVs) becomes one of the solutions to maintain content delivery when terrestrial infrastructure is unavailable or disrupted. However, existing caching approaches often focus primarily on cache-hit performance or average delivery delay, without jointly addressing spatially heterogeneous and time-varying demand, content-specific deadlines, and uneven communication-energy consumption across UAVs. To address these challenges, we formulate a dynamic deadline- and energy-aware cache placement and replacement problem that jointly considers deadline misses, UAV-side communication energy, and inter-UAV energy imbalance. To achieve this, we first develop deadline- and energy-aware UBAC (DE-UBAC), an interpretable hierarchical rule-based realization that incorporates deadline-aware prioritization, cooperative placement, energy-aware coordination, and cache-update cost. However, predefined priorities and thresholds limit its adaptability under dynamically coupled network conditions. We therefore propose GRU-assisted Deadline- and Energy-aware Proximal Policy Optimization (GRU-DEPPO), which combines local and global GRU-based temporal representations with a centralized PPO actor-critic for adaptive cache placement and replacement. Simulation results show that GRU-DEPPO achieves the lowest deadline-miss rate and total UAV-side communication energy among the considered schemes. Compared with DE-UBAC, GRU-DEPPO reduces the deadline-miss rate by approximately 61% and inter-UAV energy imbalance by approximately 40%.
[LG-62] Primal-Dual Representation Learning for Low-Rank Constrained MDPs NEURIPS2026
链接: https://arxiv.org/abs/2610.30756
作者: Chenhao Zhou,Chao Zhang,Wensong Bai,Hanbin Zhao,Hui Qian
类目: Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2026
Abstract:Representation learning has substantially improved the efficiency of unconstrained continuous-state Markov Decision Processes (MDPs), but its extension to Constrained MDPs (CMDPs) is difficult because the transition representation must be learned while constraint violation is controlled. In this paper, we study representation learning for low-rank CMDPs with continuous state spaces, and propose REP-PD-TEG for the soft cumulative-violation setting. In the feature learning period, the algorithm interacts with the environment using a decaying- \epsilon -greedy mixture execution policy, and uses the collected samples to learn transition features by maximum likelihood. Based on the learned features, it constructs uncertainty bonuses and plans with an optimistic Lagrangian objective to obtain the return policy. We prove regret and soft-violation guarantees for the return policies, and quantify the influence of the decaying- \epsilon -greedy exploration. In particular, REP-PD-TEG achieves the optimal \widetilde O(\sqrt K) return policy bound, and gives a \widetilde O(K^3/4) overall sample bound when exploratory actions are charged. For the stricter hard cumulative violation, where feasibility must be verified episode by episode, we propose REP-PD-H-TEG. It plans under a bonus-corrected lower-confidence utility certificate with the learned representation, and performs a stabilized dual search to obtain the return policy, yielding regret and hard-violation bounds. Together, these results give the first theoretically guaranteed representation-learning method for low-rank continuous-state CMDPs with both soft cumulative guarantees and certified hard-violation control.
[LG-63] Visualizing and Interpreting Temporal Care Team Dynamics for Cancer Survival Risk Analysis
链接: https://arxiv.org/abs/2610.30748
作者: Yuhua Huang,Hsiao-Ying Lu,Kwan-Liu Ma
类目: Machine Learning (cs.LG)
*备注:
Abstract:Cancer care is a longitudinal, time-sensitive process requiring coordination among diverse healthcare professionals (HCPs), whose collaboration evolves throughout treatment. Prior studies link care team structure to patient outcomes, but the temporal dynamics of collaboration and their relationship to evolving patient risk remain poorly understood. We introduce a visual analytics framework for modeling and interpreting time-varying care team interactions in oncology. We construct dynamic collaboration networks from electronic health record (EHR)-mediated communication data and use a temporal graph neural network (TGNN) to estimate patient outcome risk over time. We then investigate how changes in care team composition and specialist participation relate to shifts in estimated risk across treatment stages. Our multi-view interactive system combines TGNN-based risk modeling with coordinated visualizations of care team networks, feature-level explanations from a surrogate logistic regression model, and exploratory what-if analysis. Without making causal claims, the system helps analysts examine how risk estimates co-evolve with collaboration patterns and identify care team characteristics associated with these changes. We demonstrate its utility through expert-guided analysis scenarios involving 505 patients with breast, colorectal, and lung cancers, evaluating its ability to reveal clinically meaningful collaboration patterns. This work improves the interpretability of temporal graph-based models in longitudinal, team-based healthcare, with potential applicability to other evolving collaborative settings.
[LG-64] Wavenumber-Aware Deep Learning for Robust Analysis of Raman Spectra
链接: https://arxiv.org/abs/2610.30746
作者: Jihwan Kim,Jaeseong Heo,Yujin Kim,Sang Bin Yoon,Chanjoong Kim,Kwang-Im Oh,Hyon Bin Na,Kyungwon Kwak,Myeongkee Park,Joonyoung F. Joung
类目: Machine Learning (cs.LG); Chemical Physics (physics.chem-ph); Optics (physics.optics)
*备注: 29 pages, 5 figures, 6 tables; Supplementary Information (15 pages) included
Abstract:Machine-learning models for spectroscopy usually treat a spectrum as a fixed-grid intensity vector, leaving the physical coordinate implicit in the array index. Such models are tied to a single sampling grid, spectral range and resolution, which limits their use when measurement conditions change. Here we represent each spectral point as a pair of intensity and wavenumber and compute Transformer positional encodings from the wavenumber itself. On Raman spectra of three bacterial species and their synthetic mixtures, this representation remained robust to noise, sparser sampling and partial spectral coverage across convolutional, residual and Transformer networks. In controlled comparisons, wavenumber-based encoding outperformed index-based and learned encodings under partial observation and retained its performance on sampling grids never seen during training, where index-based encoding degraded. These findings were reproduced on a public 30-isolate bacterial benchmark. Explicit physical coordinates offer a simple route to robust spectral analysis under variable measurement conditions.
[LG-65] Random Remasking Scales Masked Diffusion Inference
链接: https://arxiv.org/abs/2610.30744
作者: Zhaoyi Zhou,Frank Wu,Andrea Zanette
类目: Machine Learning (cs.LG)
*备注: 18 pages, 4 figures, 8 tables
Abstract:Masked diffusion models (MDMs) generate sequences by progressively unmasking tokens, but errors introduced during decoding are typically irreversible. Recent work has shown that iterative correction can improve sample quality, yet existing theory fails to explain this given a learned, imperfect denoiser model. We re-formulate MDM inference as a Markov chain on the clean sequence space, defined by a remask-denoise kernel that randomly re-introduces masks and then denoises. This helps us analyze the actual denoising trajectory towards the data distribution. The analysis motivates Random Remask-Denoise (R ^2 D) inference, a purely test-time, finetuning-free, reward-free procedure. To further improve performance, we propose Trajectory Correction Training (TCT), a denoiser-training procedure that exposes the model to its own errors along the generation trajectory. On logical tasks, our approach exhibits a striking near-log-linear test-time scaling, achieving over 97% accuracy on hard Sudoku puzzles and over 99% accuracy on 3-SAT. Our approach also applies to pretrained models on actual language tasks, beating iterative correction baselines and the official Q-mode accuracy of LLaDA2.1-mini on HumanEval+.
[LG-66] Width and Depth Limits Commute During Training
链接: https://arxiv.org/abs/2610.30729
作者: Ruijia Zhang,Soufiane Hayou
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Recent work established that the infinite-width and infinite-depth limits of residual networks commute at initialization when residual branches are scaled by 1/\sqrtdepth . In particular, the network output converges to the same limit whether width and depth tend to infinity sequentially or jointly. We extend this commutativity result to training in linear residual networks. Under the same 1/\sqrtdepth residual-branch scaling (Depth- \mu P) and the \mu P width parameterization, we prove that, for any fixed number of gradient-descent steps with gradient clipping, the network output converges in L^2 along every joint width–depth growth path as both width and depth tend to infinity. Consequently, the infinite-width and infinite-depth limits commute throughout any finite training horizon. Without gradient clipping, the corresponding L^2 convergence does not hold in general, but we establish the weaker convergence in probability.
[LG-67] Federated Graph Retrieval-Augmented Generation
链接: https://arxiv.org/abs/2610.30728
作者: Xiaobao Wang,Renda Han,Zhiliang Xia,Meng Liu,Renxiang Guan,Di Jin
类目: Machine Learning (cs.LG)
*备注:
Abstract:Graph Retrieval-Augmented Generation (GraphRAG) leverages centralized knowledge graphs to enhance large language models with structured knowledge retrieval. However, real-world knowledge is inherently distributed across multiple organizations rather than being centrally accessible. Privacy constraints further prohibit sharing entire local knowledge graphs, preventing GraphRAG from effectively retrieving and utilizing distributed knowledge. To this end, we propose Federated GraphRAG (FedGRAG), a unified framework for distributed graph retrieval and knowledge-enhanced generation over multiple clients with knowledge graphs. Specifically, we address two fundamental challenges in federated GraphRAG. First, the absence of a global knowledge graph makes unified retrieval impossible. To address this, we develop a Federated Knowledge Routing mechanism that efficiently identifies relevant clients through lightweight knowledge indices without exposing local graph structures. Second, local graph memory cannot be shared directly because it contains sensitive entities and knowledge information. We therefore propose a controlled-disclosure evidence construction method. It extracts retrieved evidence subgraphs and converts them into an evidence abstract, enabling evidence conversion and reconstruction on the server. Extensive experiments demonstrate that Federated GraphRAG consistently outperforms existing baselines in both retrieval and downstream generation tasks, establishing a new paradigm for deploying GraphRAG in federated environments.
[LG-68] Solvable models of learning to pursue a moving target
链接: https://arxiv.org/abs/2610.30721
作者: John J. Vastola,Kanaka Rajan
类目: Machine Learning (cs.LG)
*备注: Accepted to RLC 2026, to be published in the Reinforcement Learning Journal
Abstract:Agents tasked with intercepting a moving target must learn how to reach it, ideally as efficiently as possible. As observers of animal pursuit behavior have noticed, different strategies are possible: on one extreme, the agent reactively moves toward the target’s current location; on the other, the agent predicts the target’s future location and moves directly there. Motivated by the desire to understand how such strategies might be learned, we introduce a continuous-time linear-quadratic model of open-loop target pursuit whose optimal control strategies interpolate between these possibilities. Usefully, a small number of interpretable parameters control which type of strategy is optimal, and it is possible to derive closed-form solutions for optimal strategies, policy learning dynamics, and value learning dynamics. Exploiting our model’s linear structure, we find that the time scales of learning precisely correspond to the eigenvalues of certain matrices, and that relevant eigenvalue spectra generically have a gap. We show that this gap indicates that agents tend to learn how to reach their target before optimizing their movement. Our results provide a detailed mathematical characterization of pursuit behavior and its learning dynamics, which can both serve as a benchmark for empirical work, and as a foundation for more elaborate theoretical treatments of pursuit.
[LG-69] AIRO: A Streaming-Based Machine Learning Framework for Threat Detection and Security in Industrial Internet of Things
链接: https://arxiv.org/abs/2610.30713
作者: Raima Deb,Pranav M Pawar,Mithun Mukherjee,Raja Muthalagu,Mian Guo
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
*备注: 7 pages, 6 figures
Abstract:Modern Industrial Internet of Things (IIoT) im- plementations now connect edge gateways, cloud dashboards and business applications. This improves visibility and remote monitoring. However it simultaneously exposes industrial traffic to various cyber threats including scanning, web attacks and Structured Query Language (SQL) injection (SQLi), phishing- driven credential theft, and command-and-control (C2) activity. Although rule-based tools remain useful for detecting known signatures, their efficacy decreases with minimally altered attacks or when alerts require manual investigation. This paper presents AIRO (Autonomous Intelligence-Driven Cybersecurity Risk Re- duction System), an efficient framework for IIoT monitoring and response. AIRO combines Apache Kafka streaming, attack- specific feature extraction, machine learning (ML) inference, confidence-based decision logic and risk reduction measures such as Internet Protocol (IP) blocking, account suspension and traffic limiting. We developed four specialized datasets for C2 beaconing, reconnaissance, SQLi, and phishing to evaluate AIRO. These datasets include noisy and overlapping characteristic distribu- tions to minimize label leakage. The best-performing models, primarily Gradient Boosting and LightGBM, achieve F1-scores up to 0.9229 while ensuring low-latency from event input to mitigation
[LG-70] An Agent Architecture with Tools and Memory for Selective Correction of Wrist PPG Heart Rate Estimates
链接: https://arxiv.org/abs/2610.30688
作者: Xu Wang,Shuhan Yi,Zhen Cheng
类目: Machine Learning (cs.LG)
*备注: 71 pages, including 48 pages of Supplementary Information
Abstract:Wrist photoplethysmography (PPG) heart rate estimates are vulnerable to motion artifacts, and corrections that reduce mean error can worsen individual windows. We propose an agent architecture with tools and memory around a fixed heart rate estimator. Pulse, motion, and limited historical evidence generate candidate estimates, whose predicted benefits and harms guide selection. An independent risk gate determines whether to retain, correct, or reject each output. We trained the architecture on PPG-DaLiA and tested it on Wrist Exercise, GalaxyPPG, and WildPPG with frozen weights. Across these four public datasets, mean absolute error (MAE), macro-averaged over ten estimators, decreased by up to 44.5%, with the largest reduction on GalaxyPPG. Beneficial and harmful corrections decreased and increased absolute error by more than 1 bpm, respectively. Their ten-estimator macro-averaged proportions ranged from 16.06% to 36.64% and from 1.68% to 6.77%, respectively, across the four datasets. At 60% and 80% target coverage, numerical correction on the same accepted windows contributed more to total error reduction than window selection. We implemented real-time wearable acquisition, Bluetooth reception, and smartphone deployment of a selected estimator and the framework. In backend analysis of four prototype recordings from one participant, ten-estimator macro-averaged MAE decreased from 9.09 to 4.45 bpm without additional rejection. Under the evaluated conditions, the architecture improved average accuracy while quantifying correction benefits and harms, supporting wearable heart rate analysis that considers error, correction risk, and coverage.
[LG-71] Learning Domain Trajectories with Flow Matching for Gradual Domain Adaptation NEURIPS2026
链接: https://arxiv.org/abs/2610.30685
作者: Yubo Huang,Yushe Cao,Zixi Wang,Jingzehua Xu,Xin Lai,Xiongyu Hu,Shuai Zhang
类目: Machine Learning (cs.LG)
*备注: Accepted by NeurIPS 2026
Abstract:Large distribution shifts remain a primary obstacle for domain generalization and unsupervised adaptation. Although gradual domain adaptation alleviates this difficulty by introducing ordered intermediate domains, most existing approaches still operate in a discrete, step-wise manner, which can accumulate errors and often relies on unstable Min–Max optimization. We propose \textbfFlowGDA, a flow-matching framework that models gradual adaptation as \emphcontinuous transport on a statistical manifold in a representation space. FlowGDA learns time-dependent velocity fields that induce smooth domain trajectories, using a simulation-free regression objective rather than adversarial training. To prevent semantic drift, we couple left-to-right transport between consecutive domains with a source-anchored transport and a semantic consistency loss, and we further stabilize learning via a two-stage sequential optimization scheme. Extensive experiments demonstrate that our method outperforms state-of-the-art techniques, validating flow matching as an effective alternative for domain adaptation.
[LG-72] RiboBridge: protein-conditioned de novo RNA binder structure design
链接: https://arxiv.org/abs/2610.30668
作者: Zhou Zhang,Hanqun Cao,Minghao Sun,Odin Zhang,Yang Zhang,Pheng Ann Heng,Wujun Li,Tianfan Fu
类目: Machine Learning (cs.LG); Biomolecules (q-bio.BM)
*备注:
Abstract:Designing RNA molecules that bind to target protein structures requires generating both an RNA conformation and its bound complex geometry, without assuming a predefined binding site. Existing approaches generally model these coupled processes within a unified representation, entangling RNA conformation generation with protein-relative placement. In this work, we introduce RiboBridge, a structure-first framework that separates RNA conformational generation from protein–RNA assembly through cross-frame interaction geometry. RiboBridge first proposes multiple candidate RNA-binding interfaces on the protein surface, generates canonical RNA conformations within a pretrained RNA structural latent space (RiboSphere), and then predicts protein–RNA interaction constraints to bridge the independently defined RNA and protein coordinate frames. The resulting interaction geometry is converted into a geometrically consistent protein-relative pose through rigid placement. We evaluate RiboBridge on a temporally split, sequence-filtered, and redundancy-controlled benchmark of 124 protein–RNA complexes. Under post-hoc interface-correspondence evaluation, where native information is used only after generation to select among already generated predicted-pocket candidates, RiboBridge achieves the highest fNAT and DockQ and the lowest interface and ligand RMSDs among compared methods. Component ablations and stage-wise diagnostics further characterize the contributions of RNA conformational generation, interaction prediction, and rigid placement, identifying the main sources of complex-recovery error.
[LG-73] Fast Iterative Combinatorial Auctions via Deep Reinforcement Learning
链接: https://arxiv.org/abs/2610.30667
作者: Luis F. Paul Tosta,Quinn Arnold,Gianluca Brero
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注:
Abstract:Iterative combinatorial auctions are used in spectrum allocation, transportation, and industrial procurement, but rounds are costly. We introduce FastCombo, a deep reinforcement learning approach to sequential price discovery. Bayesian pricing has sharply reduced auction rounds even against subgradient benchmarks tuned with hindsight for each instance, but relies on a one-step objective and repeated online inference and optimization. We prove that greedily maximizing immediate clearing probability can require arbitrarily more rounds in expectation than an informative policy, even with the true valuation distribution known. FastCombo learns price policies offline, assigning credit to queries for the future decisions they enable and producing each online price update with a single neural-network forward pass. We formulate the objective of minimizing expected rounds to clearing in single-minded auctions as a partially observable Markov decision process and train a recurrent bidder-item graph policy over complete auction trajectories. In a one-item experiment, all five trained policies learn an informative opening, reducing expected rounds from 1.85 to 1.65 relative to the tested Bayesian one-step rule. Across four domains from the Combinatorial Auction Test Suite, FastCombo achieves competitive clearing while reducing measured online price-selection time per auction by a factor of roughly 300-900 relative to the standard Bayesian baseline. These results demonstrate the value of learning sequential price policies for both information acquisition and online computational efficiency.
[LG-74] Evaluating Type-1 Diabetes Digital Twins by Decision Quality
链接: https://arxiv.org/abs/2610.30577
作者: Victor Li,Owen Tucker,Michael S. Hughes,Temiloluwa Prioleau,Shengpu Tang
类目: Machine Learning (cs.LG)
*备注: 12 pages, 1 figure, 7 tables. Code: this https URL
Abstract:A digital twin (DT) for type 1 diabetes (T1D) can predict how an individual’s blood glucose will respond to different insulin treatments without actually trying them. For each treatment, downstream statistics of predicted glucose trajectories, such as glycemic risk, can be computed to guide clinical decisions. However, the quality of DTs is usually measured by the accuracy of their predicted glucose trajectories, which may not reflect the quality of the downstream treatment decisions they inform. In this work, we directly evaluate DTs by this decision quality on downstream treatment ranking, i.e., whether a DT predicts glucose trajectories that can be used to rank candidate treatments as the real patient would. We present an evaluation protocol over 30 virtual patients from the FDA-accepted UVA/Padova model, each simulated under 21 different basal-bolus treatments. For each patient, a DT is fitted under one treatment and then produces glucose trajectories for the remaining 20 treatments. Our results show that prediction accuracy and decision quality disagree. A baseline that models the average patient, never seeing the individual’s glucose data, has the lowest prediction accuracy of all six methods (RMSE: 63.6 mg/dL) but ranks treatments with the second-highest mean Spearman (0.766), while a black-box neural network that predicts glucose nearly twice as accurately (RMSE: 32.9 mg/dL) ranks them no better than a random ordering of the 20 treatments (0.049 Spearman). Therefore, selecting a DT based on prediction accuracy alone can prefer one that recommends treatments at chance, and decision quality must be evaluated directly rather than assumed to follow from prediction accuracy.
[LG-75] When News Gets Old: Portfolio Allocation with Reinforcement Learning and Large Language Models
链接: https://arxiv.org/abs/2610.30564
作者: Bolin Song(Cornell University),Surui Xiang(United Nations)
类目: Machine Learning (cs.LG); Portfolio Management (q-fin.PM)
*备注:
Abstract:Financial news can influence investment decisions long after publication. Timely news may reach an investor before prices fully adjust, whereas older news may already be reflected in prices. We examine two ways to handle aging signals: retain them and give a portfolio policy timing information, or filter them out before portfolio selection. A fixed large language model (LLM) extracts signals from public SEC disclosures. We compare these treatments in reinforcement learning (RL) policies and LLMs that choose portfolios directly. We evaluate returns after trading costs in historical tests over 2023-2025. RL choices respond to timing information in a controlled test, but historical returns show no consistent benefit from adding it. For the LLM, filtering old news produced the highest average net return, reaching 18.43% per year when news delivery was delayed by four weeks. This exceeded the returns of all RL strategies that used news.
[LG-76] NEST: Nested Embedding Representation for Multiscale Clustering
链接: https://arxiv.org/abs/2610.30520
作者: Zhanqi Zhang,Mikio Aoi,Gal Mishne
类目: Machine Learning (cs.LG)
*备注:
Abstract:Real-world data often exhibits cluster structure at multiple resolutions, yet existing methods typically learn a single partition or construct an explicit hierarchy. Inspired by Matryoshka Representation Learning, we introduce NEST, an unsupervised framework that encodes coarse-to-fine structure within a single embedding. NEST jointly optimizes clustering and separation across nested prefix subspaces, enabling progressively finer distinctions as embedding dimensionality increases, with rank and within-cluster regularization that keep the embedding from collapsing. Empirically, distances in the learned embedding follow the hierarchy distance more closely than those of the input, although the hierarchy is never observed during training. On synthetic hierarchies, NEST recovers multiscale structure given only the number of clusters per level. Experiments on vision, text, and single-cell data further demonstrate multiscale organization across real-world representations. By making clustering granularity an intrinsic property of the representation, NEST provides a unified approach to multiscale structure discovery.
[LG-77] Optimal Multitask Representations Under a Rate Constraint: An Information Bottleneck Analysis
链接: https://arxiv.org/abs/2610.30493
作者: Nicholas Stiles,Gregory W. Wornell
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注: 9 pages main text; 39 pages total. 4 figures main text; 15 figures total
Abstract:Which pretraining objective yields the best representation for downstream tasks under an information budget? We formalize this problem as a multitask information bottleneck, using a mutual information constraint to model incomplete information retention. For Gaussian inputs and tasks, we characterize optimal representations and show that a weakly task-aligned specialist can outperform a reconstruction generalist at low rates but lose at high rates. We also construct task families for which every fixed Gaussian prediction objective incurs arbitrarily large relative regret, driven by reversals in optimal allocation priorities. Conversely, the average task information matrix defines an objective with a uniform relative-regret bound controlled by maximum task SNR. CelebA and dSprites experiments illustrate that compression can change which pretraining objective performs best, and even task-aware training can reverse which downstream task is better preserved as rate increases.
[LG-78] FIGFLEX: A Mixed-Precision FP-INT GEMM Accelerator Co-Designed with Low-Bit LLM Quantization
链接: https://arxiv.org/abs/2610.30480
作者: Faraz Tahmasebi,Haocheng Xu,Andrew Trauger,Yomal Perera,Zhenyu Tang,Nael Chowdhury,Sitao Huang
类目: Hardware Architecture (cs.AR); Machine Learning (cs.LG)
*备注: 15 pages, 12 Figures
Abstract:Weight-only quantization keeps LLM activations in floating point and stores weights as low-bit integers, turning the dominant workload into floating-point-integer (FP-INT) general matrix multiplication. Recent quantizers reach a 2-3-bit average with mixed precision, but hardware has not kept pace: GPUs dequantize weights to FP16 before the tensor-core multiply-accumulate, and bit-parallel FP-INT accelerators run at a single weight precision, so fewer bits save memory but not computation. In this paper, we present FIGFLEX, an algorithm-hardware co-designed mixed-precision accelerator whose bit-parallel datapath scales throughput as precision falls. FIGFLEX contributes (1) a shift-add, radix-4 reformulation of FP-INT multiplication that fuses the multiply-accumulate into a unified, precision-flexible partial-product reduction and roughly halves the partial products, together with a novel INT3 optimization that reclaims the idle datapath radix-4 leaves at that precision; and (2) a mixed-precision quantizer that assigns each channel group its bit-width according to the error GPTQ cannot compensate, with a per-input-channel column scaling whose power-of-two normalization keeps the scaled activations out of the FP16 subnormal range, so that the quantized model maps directly onto FIGFLEX’s datapath with the same accuracy as on a GPU. On LLaMA-2-7B, FIGFLEX attains 5.85 and 8.77 WikiText-2 perplexity at 3- and 2-bit average precision, lower than GPTQ, AWQ, SliM-LLM, AMQ, and QuIP# (without fine-tuning), while delivering 1.56x and 1.42x higher whole-chip throughput per area on the linear GEMMs of a 3-bit mixed-precision model than AxCore and FIGLUT, respectively.
[LG-79] DiffLTM: Continuous Diffusions Large Tabular Models for Conditional Generation
链接: https://arxiv.org/abs/2610.30472
作者: Alankrita Bhatt,Mukur Gupta,Andrea Montanari,Pulkit Tandon
类目: Machine Learning (cs.LG)
*备注: 68 pages; 20 pdf figures
Abstract:Conditional generation is a key capability to be supported by Large Tabular Models, enabling a number of applications. Accurate conditional generation has to contend with two key challenges: (i) Data comprise a mixture of numerical and categorical entries, the latter often with high cardinality and heavily unbalanced proportions; (ii) Conditional generation becomes increasingly difficult for rare sub-populations. The best models in this class use a hybrid of continuous and discrete diffusions. In this paper, we design a continuous diffusion model with strong generation quality. An important element is the use of cross-entropy loss for categoricals, for which we provide a first theoretical justification. We introduce metrics that focus on rare subpopulations. Further, we generalize conditional generation techniques to enable new types of conditioning. Finally, we show that our model (trained on labor data) can capture classical insights from econometrics. Code is available at this https URL Comments: 68 pages; 20 pdf figures Subjects: Machine Learning (cs.LG) Cite as: arXiv:2610.30472 [cs.LG] (or arXiv:2610.30472v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.30472 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-80] Stability Analysis of Signed Graphs under Structural Perturbations
链接: https://arxiv.org/abs/2610.30444
作者: Ning Zhang,Li Zhang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Signed graph learning has emerged as a powerful framework for modeling networks with both positive and negative interactions. While graph-filter-based methods have achieved promising empirical performance on signed networks, their stability under edge perturbations remains less well understood. Existing stability analyses primarily focus on unsigned graphs and do not account for sign-flip perturbations. In this paper, we study the stability of signed graph filters under edge additions, deletions, and sign flips. Our theoretical analysis reveals how signed edge perturbations and node feature correlations jointly determine the learned representation stability. Based on this analysis, we propose a theory-guided perturbation strategy that selects structurally influential signed-edge modifications. Experiments on synthetic and real signed graphs show that the proposed method consistently induces larger embedding drift, polarity change, and downstream link prediction degradation than baselines. These results highlight the importance of \textitsign-aware and \textitdistribution-aware stability analysis for signed graph learning.
[LG-81] owards Faithful Data Selection: Statistical Limits and Scalable Algorithms
链接: https://arxiv.org/abs/2610.30442
作者: Ahmed Mehdi Inane,Mahdi Haghifam,Gintare Karolina Dziugaite,Ioannis Mitliagkas
类目: Machine Learning (cs.LG)
*备注:
Abstract:What does it mean to have an optimal data selection algorithm? Optimality is conventionally framed as finding a subset that minimizes the population error of the classifier trained only on that subset. Our starting point in this work is first to show that this goal is information-theoretically impossible unless the data selection algorithm has access to a validation set order-wise larger than the training set. This impossibility result motivates shifting from unverifiable population risk minimization to a one-sided guarantee that we refer to as faithfulness: for any candidate subset, we provide an estimable proxy that upper-bounds the difference between training on the full training data and on only the subset. We model training with the Unadjusted Langevin Algorithm and bound the Kullback–Leibler divergence between subset-trained and full-data-trained parameter distributions by the trajectory-accumulated gradient discrepancy. Minimizing this discrepancy gives a principled way to select a subset. Even though the resulting problem is NP-hard, we propose a swap-based local search that decreases the bound at every accepted step. It can therefore refine the output of any selector without worsening its bound. Finally, we provide experimental results on CIFAR10 and CIFAR100 showing that it matches existing data selection baselines. Implementation of our algorithm and the code for reproducing the results are available at: this https URL.
[LG-82] Memory-Augmented Online Test-Time Adaptation for EEG Motor Imagery via Perturbation Consistency and Confidence Calibration
链接: https://arxiv.org/abs/2610.30422
作者: Kunkun Zhang,Kehao Wu,Zhitao Chen,Haigen Hu,Qianwei Zhou
类目: Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注: Accepted for publication in IEEE Transactions on Instrumentation and Measurement (TIM), 2026
Abstract:Recently, electroencephalography (EEG)-based brain–computer interfaces (BCIs) have gained traction as portable neural measurement systems. However, the practical deployment of these instruments is often hindered by nonstationarity and cross-subject distribution shifts, which lead to significant intersubject inconsistency. Existing adaptation methods remain impractical for real-time measurement scenarios, as they typically require offline access to source or aggregated target data. To address these challenges, this article proposes memory-augmented online test-time adaptation (OTTA) with perturbation consistency and confidence calibration (MOPC), a source-free framework designed to enhance the reliability of BCI instrumentation. MOPC enables the measurement system to update its decoding parameters on the fly during the inference stage, facilitating single-pass adaptation without accessing sensitive source data. We introduce random-mask perturbations to enforce prediction consistency against inherent EEG signal noise and acquisition artifacts. Furthermore, a confidence-calibrated reweighting loss is designed to integrate feature-space similarity, effectively mitigating unreliable pseudolabels and class imbalance during continuous signal acquisition. To ensure system stability and prevent the catastrophic forgetting of established neural patterns, a lightweight memory module is incorporated to regularize parameter updates within a tight latency budget. Evaluated on the BCI IV IIb and OpenBMI datasets, MOPC achieves state-of-the-art performance, outperforming prior methods by 2.22~% and 3.67~%, respectively, while keeping the per-trial adaptation latency to approximately 40–60 ms. This study provides a robust, recalibration-free solution for real-time EEG decoding, marking a significant advancement in the precision and usability of practical BCI measurement instruments.
[LG-83] Escaping the Phylogenetic Trap: Persistent Homology Enables Cross-Species Ion Channel Classification via Multimodal Graph Neural Networks
链接: https://arxiv.org/abs/2610.30394
作者: Arezoo Bybordi,Favour Achimba,Mariam Gelashvili,Jessy Ramirez,Weigang Qiu,Mandë Holford,Anita Raja
类目: Machine Learning (cs.LG); Biomolecules (q-bio.BM)
*备注: 9 pages, 3 figures, 2 tables
Abstract:Protein classifiers are highly sensitive to evaluation protocol: random train/test splits induce phylogenetic leakage, allowing models to exploit taxon-specific correlations and substantially overestimate out-of-distribution performance. This issue is particularly acute for venom compound classification by ion-channel target (Ca2+, K+, Na+), where reported accuracies of 80-90% do not reflect generalization to unseen taxa. We formulate cross-species prediction as a structured distribution shift in which taxonomic identity acts as a confounder, requiring representations that capture structural features invariant across phylogeny. Using a shared multimodal backbone that fuses ESM-2 sequence embeddings with a disulfide-aware graph attention network, we systematically compare five structural inductive biases (spectral, Riemannian, SE(3)-invariant, statistical-manifold, and persistent-homology representations) under a leave-one-taxon-out evaluation protocol on a benchmark of 1,373 ColabFold-predicted venom compounds spanning five taxa. The persistent-homology representation summarizes backbone geometry through topological descriptors that are robust to local structural perturbations, providing an inductive bias toward phylogenetically invariant learning. Under conventional i.i.d. splits, all models achieve similar accuracy (81-86%), masking differences in inductive bias. Under taxonomic shift, this parity collapses: a sequence-only baseline drops to approximately 30%, the remaining geometric priors cluster at 54-55%, while our persistent-homology model (PH-GNN) achieves 63.4%, outperforming the next-best method by over 8 percentage points. These results demonstrate that random-split evaluation obscures meaningful differences in representation quality and identify topological representations as an effective inductive bias for learning under structured biological distribution shift.
[LG-84] Bayesian Fixed-Confidence Best-Arm Identification with Abstention
链接: https://arxiv.org/abs/2610.30384
作者: Yuqi Huang,Yunlong Hou,Vincent Y. F. Tan
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study the Bayesian fixed-confidence best-arm identification problem in which a learner can abstain from making a terminal recommendation. Subject to a target error probability \delta and an abstention budget \alpha , we propose Screen-and-Confirm (SC) and analyze its expected stopping time. We further derive the minimum expected stopping time for any algorithm that satisfies these constraints. For Gaussian priors and rewards, as (\delta,\alpha)\downarrow(0,0) , we establish matching upper and lower bounds up to multiplicative constants, with rate \kappa_\nu^2[1+\log((\alpha+\delta)/\delta)]/(\alpha+\delta) , where \kappa_\nu is the prior density of the top-two gap at zero, highlighting that nearly tied instances govern the sampling cost. When abstention is disallowed ( \alpha=0 ), SC yields the optimal \Theta(1/\delta) dependence for fixed model parameters, removing the additional \log(1/\delta) factor in the upper bound of Jang et al. (2024) under our Gaussian assumptions. We further characterize the exact multiplicative constant of the minimum expected stopping time for practical regimes including (i) fixed \alpha \in (0,1) and (ii) \alpha/\delta\to\infty , which are also achieved by SC.
[LG-85] CTRL-FM: Inference-Time Structural Control for Flow-Based Virtual Sensing under Long-Horizon Node Outages
链接: https://arxiv.org/abs/2610.30362
作者: Boshra Khalili,Andrew Smyth
类目: Machine Learning (cs.LG)
*备注:
Abstract:Long-horizon node outages are harder than conventional missing-value imputation because complete sensor trajectories remain unobserved for extended periods, while the usefulness of spatial and temporal dependencies varies across outage locations and operating regimes. Existing graph, diffusion, and flow-based methods typically impose a fixed structural bias, which can oversmooth abrupt changes or propagate errors from unreliable neighbors. We introduce CTRL-FM, an observation-preserving flow-matching framework that treats structural guidance as an outage-conditioned inference-time decision. A missing-only flow-matching backbone learns a conditional velocity field while keeping observed measurements exactly fixed. For each realized outage, CTRL-FM constructs self-supervised proxy outages from surviving measurements and uses their known values to select among uncontrolled, graph-dominant, temporal-dominant, and balanced control policies. The selected policy is locally modulated by graph and temporal reliability estimated exclusively from surviving observations, while boundary-anchored trends provide structural guidance. Control is applied only to missing entries and is rejected when proxy evidence does not sufficiently improve over uncontrolled sampling. Across PeMS08, METR-LA, Air-36, and AQI, CTRL-FM achieves the strongest overall performance under 30% and 60% node outages. Stress tests with up to 75% sensor loss further show graceful degradation and consistent gains over training-time and fixed structural control. These results demonstrate that validating, selecting, and locally adapting structural guidance for each realized outage is more effective than imposing a single global structural prior.
[LG-86] On the Adversarial Vulnerability of Neural Receivers with Unrolled LDPC Decoders
链接: https://arxiv.org/abs/2610.30345
作者: Aakash Saini,Jafar Aco Mohammadi,Wolfgang Gerstacker
类目: Information Theory (cs.IT); Cryptography and Security (cs.CR); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注:
Abstract:Neural receivers for 5G New Radio (NR) match classical model-based receivers in link-level performance, but their robustness to adversarial perturbations at the receiver input remains underexplored. We show that standard gradient-based attacks applied end-to-end through an unrolled low-density parity-check (LDPC) decoder fail to significantly outperform a matched-power Gaussian perturbation. We identify the failure as gradient masking caused by tanh saturation in the boxplus operation of the belief-propagation (BP) check-node update, with a per-iteration gradient norm decay of more than four orders of magnitude across ten unrolled BP iterations. Based on this finding, we propose two structurally independent loss functions that restore attack efficacy. L_BP routes the gradient through the unsaturated early iterations via per-iteration supervision, while L_LLR is a margin loss on pre-decoder log-likelihood ratios (LLRs) whose gradient bypasses the LDPC decoder entirely. Both loss functions increase the block-error rate (BLER) by one to four orders of magnitude at matched perturbation power, with the results reproduced on both the Sionna NeuralReceiver and DeepRx. At low PSR, L_LLR-optimized perturbations achieve up to three orders of magnitude higher BLER on unseen channel realizations than LBP-optimized perturbations.
[LG-87] Spike-Aware Gradient-Based Mixed-Precision Quantization of Spiking Neural Networks for Vision Applications
链接: https://arxiv.org/abs/2610.30339
作者: Rajshekhar Rakshit,Devgokul B.V,Rakshitha Kalkura,Gopalakrishnan Srinivasan
类目: Machine Learning (cs.LG)
*备注:
Abstract:In recent years, Spiking Neural Networks (SNNs) have gained massive attention as an energy-efficient neural computing paradigm due to their event-driven and sparse computing capability. However, maintaining accuracy under low-precision constraints, necessary for practical deployment, has been largely unexplored for SNN models. In addition, existing Hessian-based approaches for quantization, such as HAWQ-V2, incur significant memory and computational overhead due to temporal unrolling across multiple SNN timesteps. We propose a spike-aware gradient-based mixed-precision quantization framework for SNNs that adaptively assigns bit-widths using first-order gradient information, avoiding costly second-order Hessian computations. We evaluate our method across various SNN-based architectures, including VGG, ResNet, and Vision Transformer models obtained from both ANN-to-SNN conversion and direct training frameworks, on CIFAR10/100, ImageNet, and DVS-CIFAR-10 datasets. Notably, on ResNet-34 evaluated on ImageNet, our method achieves 73.30% accuracy at 4.78 \times compression, which is comparable to that provided by HAWQ-V2, while reducing memory consumption by 2.31 \times and bit-allocation runtime by 100 \times . We also present an energy analysis for mixed-precision SNNs showing that the proposed quantization strategy enables energy-efficient deployment without compromising performance.
[LG-88] A Quadratic Lens on Muons Polar Update: Benefits Pitfalls and the Role of Newton-Schulz ICML2026
链接: https://arxiv.org/abs/2610.30280
作者: Egor Shulgin,Sam Laing,Antonio Orvieto,Peter Richtárik
类目: Machine Learning (cs.LG)
*备注: 47 pages, 13 figures. A shorter version was presented at the ICML 2026 Workshop on High-dimensional Learning Dynamics
Abstract:Muon updates matrix-valued parameters along approximately orthogonalized directions and often trains large language models faster than AdamW, yet it remains unclear when and why orthogonalization helps. We study PolarGD, the orthogonalized update at the core of Muon, on matrix quadratics, where optimizers can be analyzed in closed form, and compare update directions under line search to remove the effect of learning-rate tuning. Orthogonalization pays off when the gradient is far from low rank and its largest singular directions see high curvature. PolarGD reaches the minimizer in one step whenever the rows of the error are orthogonal with equal norms, for any Hessian, whereas gradient descent needs all Hessian eigenvalues to be equal; and unlike sign descent, PolarGD is unaffected by rotations of the problem. Over many steps, PolarGD acts as gradient descent preconditioned by the current iterate, so its speed is governed by the conditioning of the iterate rather than of the Hessian. This explains fast regimes as well as a slow one, caused by tiny singular values of the iterate, in which PolarGD is much slower than gradient descent. Finally, the few Newton-Schulz iterations used in Muon treat these tiny singular values as gradient descent does, so fewer iterations can converge faster than exact orthogonalization; in a transformer trained with Muon, about a quarter of the singular values of the attention matrices fall in this regime.
[LG-89] Benchmarking Model Compression Techniques Across Heterogeneous Edge Platforms: A Predictive Maintenance Use Case for Rotating Machinery
链接: https://arxiv.org/abs/2610.30228
作者: Matheus Calheira Guimarães de Oliveira,Marco Aurélio da Silva Cruz,Regis Cardoso,Lucas Cambuim,Rafael Macieira,Arthur Paim,Helaine Neves,Alex Santos,Lilian Guarieiro,Jean Schmith,Prashant Kumar,Erick G. Sperandio Nascimento
类目: Machine Learning (cs.LG); Performance (cs.PF)
*备注:
Abstract:To enable Industry 5.0, where autonomous systems are widely adopted and integrated with people to promote sustainable development in industries, it is imperative to adapt complex deep learning models to operate efficiently on resource-constrained hardware for edge inference. However, adapting these models remains challenging due to the hardware heterogeneity of edge computing in industrial settings, performance degradation caused by model compression, and the lack of clear guidance on which compression techniques are practical and effective for applications such as predictive maintenance. To address these challenges, this paper presents a comprehensive comparative analysis of four state-of-the-art model compression techniques applied to a high-performance convolutional neural network for predictive maintenance using a real-world dataset of rotating machinery, equipment typically used in industry. The compressed models were evaluated across a heterogeneous set of nine computing platforms, ranging from highly constrained edge computing devices to general purpose and high-performance processors, reflecting the diversity and fragmentation typical of industrial edge environments. By jointly prioritizing classification performance, inference speed, and compression ratio, we found that compression using knowledge distillation emerged as the most effective approach. The model had more than 90% speedup across all devices while maintaining a 91.22% balanced F-score, with a very small drop in predictive performance relative to the uncompressed model, along with a 98.9% size reduction. Through successful deployment across multiple edge platforms and a systematic comparison of compression strategies, we demonstrated a clear path to achieving Industry 5.0 targets, drastically cutting computational costs and, consequently, energy consumption without sacrificing predictive performance.
[LG-90] GANet: A Lightweight Group Attention Network for Plant Disease Identification on Embedded Systems
链接: https://arxiv.org/abs/2610.30225
作者: Anubhab Tripathi,Bibek Das,Asfak Ali,Sheli Shinha Chaudhuri
类目: Machine Learning (cs.LG)
*备注:
Abstract:Global warming and climate change are major catalysts for plant diseases, significantly affecting food production worldwide. Addressing this issue requires continuous monitoring and early detection of plant diseases. While recent advancements in deep learning offer potential solutions, most models demand high processing power and struggle to perform efficiently on low-end electronic hardware. Furthermore, many existing works focus on specific crop diseases, limiting their broader applicability. This paper presents a lightweight Group Equivariance Attention Network (GANet) designed for the prediction and monitoring of 38 types of plant diseases. GANet introduces a novel Group Attention block and Group-Wise Feature Fusion block, making it both sparse and efficient. With only 0.001M parameters, the model achieves an impressive 99.99% accuracy on the PlantVillage dataset. Additionally, GANet is optimized for deployment on edge devices such as the Raspberry Pi 4b and Jetson Nano, where it reaches 5 FPS and 11 FPS, respectively, demonstrating its potential for real-time plant disease monitoring on low-power devices.
[LG-91] How Prompt Style Affects AI Accuracy and Speed in Competitive Programming
链接: https://arxiv.org/abs/2610.30223
作者: Saad B. Alarifi,Mohammed A. Alam
类目: Machine Learning (cs.LG); Software Engineering (cs.SE)
*备注: 10 pages, 3 figures, 7 tables. Code and data: this https URL
Abstract:Large language models are increasingly used in competitive programming, and their outputs are sensitive to how a prompt is phrased. Prior work has compared prompt strategies for accuracy on competitive programming, but not for runtime speed, and has not situated the results against human performance. In an exploratory study of one prompt in each of nine styles on GPT-5.5 across three accuracy tasks from the International Olympiad in Informatics (IOI) and two NP-hard speed tasks, we found that no single prompt was best: the strongest prompt depended on the task type and on whether results were summarized by the mean or the median, suggesting that, at least for these tasks, prompt style should be matched to the objective rather than treated as universal. Many differences in mean scores came from a few outlier runs, showing that evaluating prompting by a single metric can obscure how much its effectiveness varies. Compared descriptively with human IOI contestants, who worked under different conditions, the AI was neither strictly better nor strictly worse, exceeding the average contestant on one task but falling short on the other two. By extending prompt comparison from accuracy to runtime speed and grounding it against human performance, this study gives a broader picture of how prompt style affects competitive programming performance.
[LG-92] AI Mindmeld: A Latent Interconnect Protocol
链接: https://arxiv.org/abs/2610.30215
作者: Siavash Golkar,Jake Kovalic,Shirley Ho
类目: Machine Learning (cs.LG); Biomolecules (q-bio.BM); Machine Learning (stat.ML)
*备注:
Abstract:LLMs can act as scientific agents using tools to consult specialists and run experiments. Yet specialist information must enter their context, generally as text, images, audio or video: a bottleneck between models with rich, high-dimensional representations. We advocate direct representation exchange: a mindmeld conveying many properties, including information an LLM would not think to request. It should be extensible to new specialists, flexible across questions and tasks, and compositional across independently connected specialists without joint training. The Latent Interconnect Protocol (LIP) learns connectors mapping specialist representations into LLM internal states. Training only connectors keeps this practical as LLMs grow and evolve. We connect three protein and three small molecule specialists to Gemma 3 (4B and 12B parameters) and Kimi K2.6 (1T total, 32B active), freezing all pretrained models. Independently trained connectors provide useful information without changing the LLM or existing connectors, supporting familiar and new question types about unseen objects. We demonstrate zero-shot transfer to cross-domain tasks without joint training. Task performance and compositionality generally improve with LLM capability, with especially strong gains in zero-shot task transfer. Limitations remain: current LLMs were not trained for specialist representations as native inputs; access does not ensure reliable interpretation, integration or use in reasoning. We call on specialist developers to provide consistent representation interfaces and, especially, on LLM developers to train models to receive, combine and reason with these inputs. These efforts could expand scientific agents’ information while preserving conventional tools’ flexibility. Data, pretrained connectors, and code for training, evaluation and interactive inference: this https URL Subjects: Machine Learning (cs.LG); Biomolecules (q-bio.BM); Machine Learning (stat.ML) Cite as: arXiv:2610.30215 [cs.LG] (or arXiv:2610.30215v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.30215 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-93] Spectral Uncertainty Decomposition for Sparse Bayesian Learning
链接: https://arxiv.org/abs/2610.30207
作者: Ismail Huseynov,Stefan Haufe
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)
*备注:
Abstract:Estimating distributed neural sources from M/EEG measurements is an ill-posed inverse problem where principled uncertainty quantification remains challenging. The usual conditional Gaussian posterior covariance does not explicitly distinguish uncertainty along likelihood-visible source directions from uncertainty that is invisible to the measurements. We propose a spectral framework based on the singular value structure of the prior-preconditioned forward operator, which provides an exact, interpretable decomposition of the posterior covariance into data-informed and indeterminacy epistemic components. This formulation identifies uncertainty contributions that are reducible by better measurements from those that are fundamentally irreducible by sensing geometry. We further derive an equivalent subtractive decomposition that expresses the prior covariance as the sum of posterior and data-driven reduction terms, enabling precise accounting of uncertainty reduction through Bayesian learning. Together, these complementary representations offer both theoretical insight and practical tools for interpretable uncertainty quantification in M/EEG source imaging for linear-Gaussian Bayesian inverse solvers where the posterior covariance is available in closed form.
[LG-94] Multifidelity Monte Carlo Estimator for High-Dimensional Quantities of Interest
链接: https://arxiv.org/abs/2610.25016
作者: Zhibo Dai,Hamdi A. Tchelepi,Daniel M. Tartakovsky
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)
*备注:
Abstract:Multifidelity Monte Carlo (MFMC) reduces the cost of estimating the statistics of quantities of interest (QoIs) computed via expensive high-fidelity model by using cheaper low-fidelity models as control variates. We extend MFMC to vector-valued QoIs by replacing the scalar control-variate coefficients with matrices and minimizing a weighted mean-squared error. The optimal full-matrix coefficients are available in closed form, and their variance reduction is summarized by an effective correlation that takes the place of the squared Pearson correlation in the closed-form MFMC design. Restricting the coefficients to block-diagonal, diagonal, and scalar classes gives a nested hierarchy of estimators with increasing variance reduction and increasing pilot-data demand, with closed-form coefficients for compatible weightings. For a scalar QoI obtained by linearly projecting the state, correcting the full state before projecting is never worse at the population level. Experiments on a Gaussian problem, a reduced-order model of an advection–diffusion problem, and a coarse-grid two-phase flow problem show that the best class depends on the pilot size relative to the number of covariance quantities it must estimate, and that reusing the pilot samples lowers the mean-squared error.
[LG-95] Reliable AI-Assisted Decisions via Selective Auditing
链接: https://arxiv.org/abs/2610.31662
作者: Isaac Gibbs,Elise L. Han,Ryan J. Tibshirani,Tijana Zrnic
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
*备注:
Abstract:We consider online monitoring of an AI system that produces a sequence of decisions. At each round, the system outputs a proposed decision together with a confidence score estimating the trustworthiness of the decision. The decision and confidence score need not be produced by the same model: for example, one agent may propose a decision, while a separate verifier agent evaluates it and outputs a correctness or safety score. Based on this information, we must decide whether to accept the AI decision automatically or consult an expert, such as a human auditor, who reviews the proposal and either accepts or corrects it. We formulate this task as an online constrained optimization problem in which the goal is to minimize the frequency of expert auditing while guaranteeing that the fraction of AI errors that go unaudited stays below a user-defined threshold. We derive a family of algorithms with provable error-control guarantees with no assumptions, including a method that is optimally efficient: it asymptotically attains the best achievable auditing rate for both adversarial and i.i.d. data streams. Experiments on screening harmful requests to large language models, qualitative text annotation, and data labeling show that our methods achieve valid error control in practice while substantially reducing the need for human auditing compared to baselines.
[LG-96] Direct Inference on Admissible Sets (Diona)
链接: https://arxiv.org/abs/2610.31526
作者: Lorenzo Sabug Jr.,Eric Kerrigan
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 27 pages, 4 figures
Abstract:We propose a new regression method named Direct Inference on Admissible Sets (Diona) that directly uses samples to estimate the state and its derivatives at a query point, along with their admissible bounds, without long global training. By casting the values and derivatives at the query and neighbouring samples as latent variables in a constrained optimisation problem, the burden of long training loops becomes one of on-demand local solves. We establish Diona’s theoretical properties including finite-sample coverage and, in the noiseless case, convergence to the true state and derivatives under dense sampling. In a nonlinear regression benchmark on six functions of three to eight dimensions, Diona was more accurate than a Gaussian process (GP) on noiseless data especially with denser data sets. Furthermore, as with GP, it produced uncertainty bounds that contained the true value at more than 95% of query points in four functions. Diona’s time to a first prediction after a data set update was three to four orders of magnitude shorter than that of the GP, and two to three orders shorter than that of a neural network, which was more accurate but offers no uncertainty quantification. This low latency is paid for by on-demand predictions, so this latency advantage applies when relatively few predictions are needed between data updates.
[LG-97] Nonparametric autoregressive estimation of probability distributions with parallel gradient boosting
链接: https://arxiv.org/abs/2610.31524
作者: Rémy Chapelle(CB, EVDG, CESP)
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Nonparametric estimation of joint probability distributions is a core problem in statistics, which is notoriously challenging with high-dimensional data. Such challenges also arise with conditional distribution estimation and have led to flexible algorithms that aim to tackle the ‘‘curse of dimensionality’’ by exploiting predictive structures of lower dimension than the ambient space. In this paper, we study the use of such conditional distribution estimators to estimate joint distributions by factorizing the latter in an ‘‘autoregressive’’ way. We show that for loss functions satisfying a subadditivity property, bounds on the convergence rate of an autoregressive estimator can be derived from bounds on the convergence rate of the conditional distribution estimators. These bounds relate the difficulty of estimating the joint distribution to that of estimating the most challenging factor in the autoregressive factorization. When the conditional estimators adapt to sparsity in their conditioning sets, we characterize this difficulty through variable elimination in an undirected dependence graph. This allows bounding the rates of convergence by quantities related to the sparsity of the graph, and provides an additional motivation for the autoregressive approach. We then evaluate the practical performance of an autoregressive estimator based on parallel gradient boosting, a flexible nonparametric method for estimating conditional distributions. The estimator demonstrates competitive estimation performance on various simulated and real datasets, even without optimizing the factorization order. Together, these results support nonparametric autoregressive estimation as a practical approach to flexible joint distribution estimation.
[LG-98] Dimension-free regimes for unadjusted Langevin in ridge-separable models: A subspace robust Wasserstein analysis
链接: https://arxiv.org/abs/2610.31431
作者: Yuansi Chen,Francesco Pedrotti,Peter A. Whalley
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST); Computation (stat.CO)
*备注: 30 pages, 4 figures
Abstract:The Unadjusted Langevin Algorithm (ULA) is widely used for high-dimensional Bayesian inference. Classical analyses in Wasserstein-2 distance predict an asymptotic bias and a mixing time that grow with dimension, but this metric is stronger than necessary for many statistical tasks. We instead analyze the bias in subspace robust Wasserstein distance, which controls the mean and arbitrary low-dimensional marginals and is therefore well suited to observables typically estimated in practice. For ridge-separable potentials formed with i.i.d. data, we prove high-probability bounds showing a bell-shaped mixing profile as the number of data points N varies relative to the dimension d . In the overparameterized regime, the effective dimension is the rank N of the data-dependent perturbation, leading to complexity depending on N rather than d . In the large-sample regime, Gaussian-perturbation estimates yield dimension-free complexity once N is sufficiently large. In contrast, near the proportional regime \asymp d , we prove that dimension-free guarantees are impossible in general.
[LG-99] Graph-Recurrent Learning of Emergent Collective Dynamics
链接: https://arxiv.org/abs/2610.31350
作者: Ali Rayat,Acacia Lee,Gia-Wei Chern
类目: rongly Correlated Electrons (cond-mat.str-el); Statistical Mechanics (cond-mat.stat-mech); Machine Learning (cs.LG)
*备注: 13 pages, 8 figures
Abstract:Collective ordering in correlated electron systems emerges from microscopic dynamics, yet the evolution of observable textures need not admit a closed, memoryless description. We introduce a graph-recurrent neural network (GRNN) framework that learns such emergent dynamics directly from trajectories without specifying microscopic equations of motion. The framework integrates gated recurrent memory with graph message passing: local hidden states encode information beyond the instantaneous observables, while spatial coupling allows this information to propagate between sites. The observable and hidden fields evolve jointly through a shared local rule, providing a scalable representation of dynamics with spatial interactions and temporal memory. We demonstrate this framework on the emergent dynamics of charge-density-wave order in the semiclassical Holstein model, successfully reproducing temporal and spatial correlations and dynamical scaling during domain coarsening. By learning finite-time evolution directly, the GRNN offers a route to efficient, data-driven descriptions of collective dynamics with unresolved microscopic degrees of freedom.
[LG-100] Metric Conformal Prediction via the Conditional Expected Local Radius
链接: https://arxiv.org/abs/2610.31336
作者: Rui Qiu,Fang Yao,Zhou Yu
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:We introduce the conditional expected local radius (CELR) and develop a new conformal prediction method for non-Euclidean responses in general metric spaces. CELR measures local conditional probability concentration, allowing prediction sets to accommodate heterogeneity and multimodality without requiring densities or a vector space structure. Its conditional distribution yields a nonconformity score that is comparable across covariate values and avoids costly transport comparisons between distance profiles. We further adapt a metric distributional random forest that handles multivariate covariates and nonlinear interactions, while a single forest fit supplies all score components, reducing computation by eliminating repeated local fitting. We establish uniform consistency of the estimators and asymptotic conditional validity under suitable regularity conditions, together with finite-sample marginal coverage under exchangeability. Numerical studies demonstrate effective uncertainty quantification across different metric spaces and substantial computational savings over the conditional profile score method.
[LG-101] Deep Learning for Random Objects in Metric Spaces: Beyond Manifold Assumptions
链接: https://arxiv.org/abs/2610.31328
作者: Marcos Matabuena,Xiaokai Luo,Oscar Hernan Madrid Padilla
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST); Methodology (stat.ME)
*备注:
Abstract:Deep learning (DL) methods have substantially improved predictive performance across a wide range of downstream prediction tasks, enabling new discoveries in multiple domains. However, principled DL methods for non-Euclidean random objects remain relatively scarce, despite the growing importance of modern biomedical outcomes taking the form of probability distributions, random graphs, functions, and other random objects in metric spaces. Motivated by a new generation of digital and clinical biomarkers, we propose a deep learning framework for regression with random objects in metric spaces. Our approach operates directly on the metric structure of the response space, avoiding restrictive manifold or Euclidean embedding assumptions imposed by many existing methods. At the same time, it retains modern statistical guarantees under the stated statistical assumptions. We establish statistical consistency and convergence rates for the proposed estimator, and characterize conditions under which it can mitigate the curse of dimensionality for hierarchically compositional regression functions. Extensive simulation studies demonstrate competitive predictive performance, with gains over several competing approaches in multiple Euclidean and non-Euclidean configurations. Finally, we demonstrate the practical utility of the proposed framework in two digital health case studies with distribution-valued outcomes in Wasserstein spaces, using high-resolution accelerometry and continuous glucose monitoring data. Embedding-based deep-learning methods may be less suitable when no informative low-dimensional representation exists or when the learned embedding does not adequately preserve the intrinsic geometry of the response space.
[LG-102] rack finding and particle identification performance with particle physics foundation models
链接: https://arxiv.org/abs/2610.31180
作者: Shuhang Li,Haiwang Yu,Gabor Galgoczi,Yeonju Go,Yi Huang,David Park,Michael Peters,Christof Roland,Jin Huang,Yihui Ren,Joseph D. Osborn
类目: Data Analysis, Statistics and Probability (physics.data-an); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); Nuclear Experiment (nucl-ex)
*备注:
Abstract:Modern high energy nuclear and particle physics experiments produce large amounts of sparse data that must be reduced significantly in size before performing physics analysis. Recently, there has been an effort to evaluate Foundation Models as a potential tool for performing data analysis. These models have been shown to learn features in high dimensional datasets based on the multi-modal data they are trained on. To show their effectiveness at performing complex pattern recognition tasks, such as reconstructing charged particles in a magnetic field, we evaluated the performance of a Foundation Model based reconstruction on several downstream tasks with a dataset composed of spacepoints in a Time Projection Chamber. We show that the performance of the Foundation Model based data analysis significantly improves upon that of traditional methods, and furthermore can be adapted to learn additional tasks that are not necessarily related to the initial training objective.
[LG-103] Prediction Uncertainty Controlled Computer Model Calibration
链接: https://arxiv.org/abs/2610.31163
作者: Ryan Heejoon Byun,Dayeon Yoon,Won Chang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Applications (stat.AP); Computation (stat.CO); Methodology (stat.ME)
*备注:
Abstract:Computer-model calibration is usually undertaken in service of prediction, yet neither Bayesian Kennedy-O’Hagan (KOH) calibration nor amortized simulation-based inference is built to control the predictive error of the calibrated simulator. When parameters are weakly identified or the data-model discrepancy is misspecified, the simulator-output bands induced by a nominal posterior level can be far from nominal. We propose a two-layer framework that targets predictive miscoverage directly. First, a conditional flow-matching model learns an amortized posterior for the calibration parameters, trained on simulator outputs perturbed with Gaussian-process discrepancies of varying structure. Second, conformal risk control arranges posterior draws for the target observation into nested candidate sets and selects the smallest one whose fixed-rank pointwise bands attain a prescribed time-averaged miscoverage on a synthetic calibration sample. Under exchangeability of calibration and target observations, the bands control expected miscoverage in finite samples. In a synthetic benchmark, the proposed bands are roughly 20% the width of KOH-MCMC bands with miscoverage near the 10% target. In calibrating an age-structured SEIR model to the 2010 Malawi measles outbreak, they are again sharper and near-nominal, and the weeks they fail to cover coincide with the asymmetric epidemic decline, exposing structural inadequacy of the mechanistic model.
[LG-104] Random Projections for Spectral Algorithms: Optimal Rates in Sobolev Norms
链接: https://arxiv.org/abs/2610.30948
作者: Caixing Wang,Yue Wang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Random projections (RP) offer an effective method for reducing computational and storage overhead while preserving the geometric structure of data. However, existing studies have primarily focused on the optimal generalization performance of specific kernel-regularized algorithms with RP in the well-specified setting and under restrictive conditions. In this paper, we provide a comprehensive and improved analysis of the generalization performance of RP-based spectral algorithms under general conditions \changewithout sacrificing statistical efficiency. By leveraging the embedding property of the reproducing kernel Hilbert space (RKHS) and a refined analysis of the operator similarity, we establish optimal learning rates in Sobolev norms that match the minimax lower bounds up to logarithmic factors. For two common implementations of RP, randomized sketches and Nyström sub-sampling (uniform or leverage-based), we show that the projection dimension required for optimality is proportional to the average or maximal effective dimension, yielding a significant reduction in computational cost while maintaining the statistical efficiency. Our results do not rely on the uniform boundedness assumption on the regression function and are valid across a wide range of source conditions, i.e., s \alpha_0-1/\beta , where s,\beta , and \alpha_0 denote the smoothness index, capacity index, and the embedding index, respectively. In the benign case when \alpha_0=1/\beta , the optimality holds for all s\in (0,2\tau] with \tau and \gamma denoting the qualification index and norm index. Experimental results demonstrate both the theoretical validity and the practical effectiveness of the proposed method.
[LG-105] Continuous-depth Deep Gaussian Processes
链接: https://arxiv.org/abs/2610.30890
作者: Ying Li,Sanyou Wu,Zi Yang,Qiaochu Xu,Yuhao Liu,Zhidi Lin,Michael Minyi Zhang,Petar Djuric,Alexandre H. Thiery
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 22 pages, 3 figures
Abstract:Deep Gaussian processes (DGPs) are appealing Bayesian regression models, but their standard discrete-depth parameterization often becomes harder to optimize as depth grows. We argue that depth is better treated as a latent continuous evolution than as a long discrete stack. We therefore propose a continuous-depth formulation in which the latent representation follows a stochastic differential equation with a depth-indexed drift family equipped with a GP-based prior. This reframes the core design choice as the prior placed on that family, rather than adopting a purely neural continuous-depth parameterization as in neural ODE or neural SDE models. We study two instantiations of this idea: CDGP, a direct GP prior over the state–depth drift, and FlowDGP, a flow-evolved prior that captures richer depth dependence at lower training cost. To motivate the redesign, we analyze how the single-sample DGP objective can become increasingly sensitive under repeated layer composition, while residual DGP is a useful first repair that still remains a finite discrete construction. We also give pathwise sensitivity analysis for the continuous-depth formulation. Experiments on synthetic, benchmark, and large-scale regression tasks show that residual DGP improves over plain DGP, while CDGP and FlowDGP provide a stronger overall story across the regimes we study.
[LG-106] xVGAE: A Hierarchical Variational Graph Autoencoder for Exchangeable Graphs
链接: https://arxiv.org/abs/2610.30815
作者: Daniele Micheletti,Federica Zoe Ricci,Erik B. Sudderth
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Graphs play a crucial role in many applications, ranging from drug discovery to social network modeling and astronomy. Generative models for graphs have substantially advanced in recent years, but we identify relatively simple scenarios where state-of-the-art models struggle to converge within a standard compute budget, or exhibit prohibitively slow generation times. We propose a generative architecture that achieves quality competitive with state-of-the-art models and is orders of magnitude faster at generation. Our approach is inspired by the Aldous-Hoover theorem, a classic representation theorem that characterizes any probability distribution of graphs that is invariant to permutations of node indices or exchangeable. This theorem establishes that three ingredients suffice: a graph-level latent variable, a set of exchangeable node-specific latent variables, and a so-called graphon function that maps graph- and node-level variables to the probability of an edge between any node pair. Given a training set of graphs, our exchangeable Variational Graph AutoEncoder (xVGAE) learns an approximation to their underlying graphons, as well as graph- and node-level latent variables. We show that our xVGAE encodes latent representations that correlate with graph- and node-level properties, and can (quickly, in one shot) generate new graphs that closely match key training set statistics.
[LG-107] Flow-based Spectral Kernel Learning for Nonstationary Attention
链接: https://arxiv.org/abs/2610.30465
作者: Zi Yang,Ying Li,Alejandro Lancho,Darío C. Larese,Pablo M. Olmos,Michael Minyi Zhang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 23 pages
Abstract:Kernelized attention provides a useful perspective for understanding self-attention as a similarity-based aggregation mechanism, but existing approaches often rely on fixed or stationary kernels, limiting their ability to adapt the similarity structure to complex token interactions. In this paper, we propose flow-based spectral kernel attention (FSKA), a framework for learning expressive nonstationary attention kernels. FSKA formulates kernelized attention through spectral density learning and parameterizes the resulting bivariate spectral density with normalizing flows. This design enables flexible and regularizable kernel learning while remaining compatible with scalable linear-attention computation. We further show that expressive kernel learning enables a simplified attention design with removed query–key projections and fixed orthogonal value projections, resulting in a more parameter-efficient architecture. Controlled experiments on classification, ablation, and scalability profiling show that FSKA achieves competitive or improved performance over existing attention mechanisms, while reducing projection-related trainable parameters by approximately 40% – 47% compared with most baselines and maintaining favorable linear-attention memory scaling. Code Available: this https URL.
[LG-108] A Transformer-based Approach for Real-time Multi-agent Path Planning with Varying Initial/Final Conditions
链接: https://arxiv.org/abs/2610.30388
作者: Qian Zhang,Shanqing Liu,Jerome Darbon,George Karniadakis
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Numerical Analysis (math.NA)
*备注:
Abstract:We address the problem of real-time multi-agent path planning under prescribed initial and final configurations, formulated as a high-dimensional optimal control problem subject to inter-agent collision avoidance and obstacle constraints. Classical solvers based on the Hamilton–Jacobi–Bellman equation or Pontryagin’s maximum principle (PMP) become computationally intractable as the number of agents grows, due to the curse of dimensionality and the non-convexity introduced by interaction constraints. We propose OCFormer, a transformer-based neural network that parameterizes the inverse of a coordinate transformation mapping a simple, analytically tractable latent space, in which agents evolve independently under a decoupled Hamiltonian, to the physical space of the coupled, interacting system. The model is trained by minimizing a physics-informed loss comprising the Hamiltonian ODE residual from the PMP and a boundary matching term, without requiring pre-computed trajectory data. The self-attention mechanism endows OCFormer with permutation equivariance over agents, a structural symmetry that prior symplectic network approaches do not natively enforce. When trained on a distribution of boundary conditions, the amortized OCFormer produces near-optimal trajectories for unseen initial and final configurations in a single forward pass ( \sim 220ms), enabling real-time inference without retraining. We further introduce a lightweight transfer learning framework that adapts a pretrained OCFormer to out-of-distribution perturbations in seconds via L-BFGS fine-tuning. Numerical experiments on two- and three-dimensional multi-agent navigation with up to 128 agents demonstrate the scalability and efficiency of the proposed approach.
[LG-109] SciFusion: Physics-Guided Compositional Diffusion for Multimodal 3D Scientific Image Generation
链接: https://arxiv.org/abs/2610.30313
作者: Luoxi Wang,Alessio Quercia,Hanno Scharr,Arya Bangun
类目: Image and Video Processing (eess.IV); Machine Learning (cs.LG)
*备注:
Abstract:Existing 3D generative approaches are typically developed for individual imaging modalities, requiring substantial modality-specific data and limiting their ability to exploit shared structure across scientific imaging domains. We present SciFusion, a multimodal framework for unified 3D scientific image generation built on a latent diffusion backbone, with two key contributions. First, we introduce a compositional conditioning scheme that factorizes each modality into domain, object, and contrast representations. Sharing these factors across related modalities enables knowledge transfer across heterogeneous imaging domains and allows the factors to be recombined at inference time to synthesize unseen domain-object-contrast compositions in a zero-shot manner. Second, we incorporate domain-specific physics-informed objectives: Fourier spectral consistency for MRI, Radon projection consistency for CT, and multislice diffraction consistency for electron microscopy, to encourage generated samples to respect modality-dependent image formation processes. Extensive experiments across 15 medical and scientific imaging modalities demonstrate that SciFusion improves generation fidelity and cross-domain generalization over modality-specific and flattened multimodal conditioning strategies, while enabling zero-shot synthesis of unseen compositions without training on those target compositions. These results demonstrate that compositional, physics-guided conditioning provides a scalable framework for unified 3D scientific image generation, particularly for emerging modalities where large-scale training data are difficult to obtain.
[LG-110] Differentially Private Synthetic Data via Generalized Moment Matching
链接: https://arxiv.org/abs/2610.30196
作者: Rundong Ding,Yiyun He,Yizhe Zhu
类目: Machine Learning (stat.ML); Information Theory (cs.IT); Machine Learning (cs.LG); Probability (math.PR); Statistics Theory (math.ST)
*备注: 53 pages
Abstract:We study (\varepsilon,\delta) -differentially private synthetic data for arbitrary databases of n records in [-1,1]^d under a specified integral probability metric (IPM). We introduce generalized moment matching, which estimates moments privately, fits a probability measure, and samples synthetic records. A general error theorem applies when moment error controls the target IPM. This framework extends Chebyshev moment matching for 1 -Wasserstein distance and smooth queries. We apply the framework to maximum mean discrepancy with Matern and Laplace kernels and a two-layer ReLU neural tangent this http URL also obtain guarantees for sliced Wasserstein- 1 and the IPM induced by norm-bounded ReLU networks. With sufficiently many synthetic records,these rates are minimax optimal up to logarithmic factors for fixed dimension.
[LG-111] SIGNN: Robust Graph Neural Star Identification on Earth
链接: https://arxiv.org/abs/2610.30186
作者: Jeongbin Sohn,Jina Lee,Max Yen,Dowan Gwon,Jeahn Han,Floyd Hepburn-Dickins,Brian Coltin,Ryan Soussan,Pyojin Kim
类目: Instrumentation and Methods for Astrophysics (astro-ph.IM); Machine Learning (cs.LG)
*备注: 9 pages, 9 figures, 4 tables
Abstract:We present SIGNN++, a learning-based star identification (SID) framework for celestial navigation in GNSS-denied nighttime environments. Although researchers have developed learning-based SID methods to improve robustness to observation noise, prior evaluations have focused mainly on synthetic or non-terrestrial environments, leaving their reliability under real terrestrial conditions unclear. To address these limitations, SIGNN++ extends SIGNN, a graph neural network-based SID method, with three components: a more realistic noise model for training data, Fourier Transform (FT) of angular-distance edge features, and RANSAC-based Perspective Star Alignment for identification and attitude refinement. The results show that SIGNN++ is more robust to terrestrial observation disturbances and achieves higher star identification accuracy and solved rates than the original SIGNN and publicly available SID baselines, while maintaining comparable attitude accuracy and practical computation time for continuous operation. The implementation code of SIGNN++ and datasets will be available at this https URL.
[LG-112] Do Machine Learning-Guided Proactive Full Capacity Protocols Mitigate Emergency Department Overcrowding? A Simulation-Based Evaluation
链接: https://arxiv.org/abs/2610.30148
作者: Khalid Aram,Orhun Vural,Bunyamin Ozaydin,James Booth,Abdulaziz Ahmed
类目: Optimization and Control (math.OC); Computers and Society (cs.CY); Machine Learning (cs.LG)
*备注:
Abstract:Objective: Emergency department (ED) overcrowding threatens care quality, yet Full Capacity Protocols (FCPs) are typically activated only after crowding thresholds have been exceeded. This study evaluated whether machine learning-guided proactive FCP activation improves ED crowding control compared with conventional reactive activation. Materials and Methods: We developed and validated an integrated machine learning and discrete-event simulation framework representing ED patient flow and capacity management. An embedded machine learning model predicted hourly waiting-room crowding six hours ahead and triggered proactive FCP activation. Using empirically calibrated operational parameters from a large US hospital, we compared proactive and reactive deployment strategies over a 365-day simulation horizon with 10 independent replications. Outcomes included ED length of stay, wait time, percentage of time above the crowding threshold, number of FCP activations, and protocol duration. Results: Proactive deployment produced statistically significant reductions in the percentage of time above the crowding threshold and average protocol duration across all threshold settings. It also reduced crowding variability and extreme congestion events, resulting in smoother ED operations. Although proactive deployment generally improved ED length of stay and wait time, these differences were not statistically significant. Conclusion: Embedding machine learning predictions within an operational simulation enables proactive, prediction-guided capacity management before severe congestion develops. These findings support the feasibility and potential value of proactive FCP deployment as a decision-support strategy for improving ED crowding control and provide evidence to support prospective implementation in clinical practice. Subjects: Optimization and Control (math.OC); Computers and Society (cs.CY); Machine Learning (cs.LG) Cite as: arXiv:2610.30148 [math.OC] (or arXiv:2610.30148v1 [math.OC] for this version) https://doi.org/10.48550/arXiv.2610.30148 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Abdulaziz Ahmed [view email] [v1] Wed, 5 Aug 2026 21:06:02 UTC (4,284 KB)
附件下载
