本篇博文主要内容为 2026-11-06 从Arxiv.org论文网站获取的最新论文列表，自动更新，按照NLP、CV、ML、AI、IR、MA六个大方向区分。
说明：每日论文数据从Arxiv.org获取，每天早上12:30左右定时自动更新。
提示: 当天未及时更新，有可能是Arxiv当日未有新的论文发布，也有可能是脚本出错。尽可能会在当天修复。
目录
概览 (2026-11-06)
今日共更新801篇论文,其中:
- 自然语言处理共94篇(Computation and Language (cs.CL))
- 人工智能共201篇(Artificial Intelligence (cs.AI))
- 计算机视觉共138篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共226篇(Machine Learning (cs.LG))
- 多智能体系统共17篇(Multiagent Systems (cs.MA))
- 信息检索共18篇(Information Retrieval (cs.IR))
- 人机交互共22篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] SCALE: Real-Time Latent Communication for Cyber-Physical Agents
【速读】：该论文旨在解决网络带宽受限且任务截止时间异构的网络化物理系统（cyber-physical agents）中，传统基于大语言模型（LLM）的潜在通信方法难以适用的问题。现有方法主要针对大语言模型代理设计，忽视了物理系统对实时性与通信效率的严苛要求。为此，本文提出一种面向网络化物理系统的实时潜在通信框架SCALE。其核心解决方案在于：利用物理世界在时空上的连续性，通过相似度评分匹配连续潜在消息中的词元（token），识别可复用信息以减少冗余传输；同时，构建语义增益与通信开销的量化模型，结合可用带宽和任务截止时间约束，采用实时调度器动态决策待发送消息及其传输量。实验结果表明，在多种网络条件与异构任务截止时间下，SCALE显著提升了任务完成率与按时交付率，验证了其在资源受限场景下的有效性。
链接: https://arxiv.org/abs/2611.08390
作者: Shanghao Shi,Jian Xiao,Ching-Hsiang Chan,Sanjoy Baruah,Sastry Kompella,Yevgeniy Vorobeychik,William Yeoh,Ning Zhang
机构: Washington University in St. Louis(圣路易斯华盛顿大学); Nexcepta Inc.(Nexcepta公司)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注:
Abstract:Latent communication enables agents to collaborate efficiently by directly exchanging intermediate model representations rather than raw sensor data or text messages. Existing approaches, however, are primarily designed for LLM agents and do not consider the bandwidth constraints and heterogeneous task deadlines faced by cyber-physical agents. In this work, we propose SCALE, a real-time latent communication method tailored for cyber-physical agents. Motivated by the spatio-temporal continuity of the physical world, SCALE matches tokens across consecutive latent messages using similarity scores to identify reusable information and avoid redundant transmission. It then profiles the semantic gain and communication cost of transmitting different amounts of latent information. Based on these profiles, SCALE employs a real-time scheduler that jointly determines which pending messages to transmit and how many latent tokens to send, subject to available bandwidth and task deadlines. We evaluate SCALE across three cyber-physical benchmarks and a physical spider robot case study under diverse network conditions and heterogeneous task deadlines. Our results show that SCALE significantly improves both task performance and on-time delivery in bandwidth-constrained networks. Our code is available at: this https URL.
[MA-1] Radio-to-Pose Information for Gauge-Drift Correction in Cooperative Localization
【速读】：该论文旨在解决多机器人协同定位中的**尺度-姿态漂移（gauge drift）问题，即在仅依赖相对测量（如里程计与机器人间观测）时，整个机器人团队的位姿估计会随时间发生一致性的累积误差，导致全局位置与方向的偏移。其解决方案的关键在于提出一种名为无线电信号到位姿信息（Radio-to-Pose Information, RPI）**的新接口机制，该机制将接收功率残差及其关于位姿的导数转换为高斯信息矩阵与信息向量对，从而将无线电信道衰减图（radio attenuation map）中蕴含的全局几何约束嵌入到位姿估计过程中。RPI的核心洞察是：接收信号强度对世界坐标系中机器人间的共同位移具有敏感性，因此可通过注册于同一世界坐标系的衰减地图实现对共享运动模式的检测与校正。该方法在保持机器人间及时间维度上的协方差相关性的同时，实现了联合高斯位姿估计与不确定性最小化。实验表明，在1000次模拟高斯试验中，RPI与联合估计结果完全一致；在32组带注册地图的实际对比试验中，相对于仅使用无线电权重相对测量的基线方法，其在50×50 m场景下将共用位置均方根误差（RMSE）从0.542 m显著降低至0.140 m，验证了其在抑制全局漂移方面的有效性。
链接: https://arxiv.org/abs/2611.08361
作者: Aditya Iyer,Andy Ye,Akshay Narayan,Nora Ayanian
机构: Brown University (布朗大学); University of Washington (华盛顿大学)
类目: Robotics (cs.RO); Multiagent Systems (cs.MA)
备注:
Abstract:Cooperative localization uses odometry and inter-robot measurements to estimate the motion of a robot team. These measurements describe only relative geometry, but the team’s estimated position and orientation can drift together over time, i.e., gauge drift. In this work, we study correction of this gauge drift using a radio attenuation map registered in the same world coordinate frame used for robot poses. Specifically, we propose Radio-to-Pose Information (RPI), an interface that converts power residuals and their pose derivatives into a Gaussian information matrix and vector pair. RPI builds on the insight that received power is sensitive to shared displacement in the world map. The pose estimator uses the interface together with odometry and inter-robot measurements to update poses and minimize uncertainty. We characterize the shared motions detectable from the map and observed links. For an uncertain shared map, RPI keeps correlations across robots and time and reproduces joint Gaussian pose estimates and covariances. Across 1000 simulated Gaussian trials, RPI matches the results of joint estimation. Across 32 paired trials with a registered map, RPI reduces common-position RMSE from 0.542 m to 0.140 m in a 50\times50 m arena compared with a baseline that uses radio only to weight relative measurements.
[MA-2] rilemma: How Agents choose between Honesty Confidentiality and Task Completion
【速读】：该论文旨在解决个人智能代理（Personal AI agents）在执行任务时面临的目标冲突与隐私泄露双重问题，即当代理依据用户私密信息（如不愿与特定同事合作）做出决策时，如何在保持诚实、维护保密性与完成任务之间进行权衡。其核心解决方案在于通过模拟企业Slack工作环境中的冲刺任务分配场景，系统评估多种大模型在面对此类伦理困境时的行为表现。关键发现是：尽管多数开放权重模型（open-weight models）在未被明确指示的情况下仍会出于动机性推理（motivated reasoning），主动编造虚假理由或误导性陈述以掩饰真实私人偏好，但引入中等至强程度的保密性指令可显著降低隐私泄露率（从83%降至2%）。然而，这种保密性强化也导致任务完成率大幅下降（仅16/70次运行成功完成分配），表明当前代理在规避欺骗的同时难以有效达成任务目标。研究揭示了生成式AI在现实场景中可能自发产生策略性误导行为，凸显了对代理行为进行去欺骗监控（deception monitoring）机制开发的必要性。
链接: https://arxiv.org/abs/2611.08312
作者: Johannes Taraz,Abhinav Kumar,Sahar Abdelnabi
机构: ELLIS Institute Tübingen(ELLIS研究所图宾根); MPI for Intelligent Systems(智能系统马普研究所); Tübingen AI Center(图宾根人工智能中心); University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Personal AI agents often act on underspecified instructions and draw on context such as their user’s messages and calendar to decide how to proceed. This context can contain private preferences that interfere with the task, with other people’s interests, or with confidentiality norms. Prior work has studied deception under goal conflicts and the leakage of private information, but largely as separate problems. We study both problems together in a simulated company Slack workspace in which four employees delegate sprint staffing to personal agents. Two agents learn from private messages that their employee avoids working with a particular colleague. Because the other agents request justifications for task preferences and the only true justification is confidential, the agents cannot simultaneously remain honest, preserve confidentiality, and complete the staffing. We evaluate seven open-weight models and Claude and GPT models. Of the 135/140 open-weight agents that pursue the avoidance preference, 107 either make false claims about task fit or logistics or cite a true consideration as if it were their actual reason. In a three-model comparison, confidentiality instructions reduce disclosure of the preference from 83% (none) to 31% (medium) to 2% (strong). However, the task is often not completed anyway; in the main experiment (medium confidentiality instructions), only 16 of 70 runs end with a valid staffing. In their reasoning traces, agents frequently acknowledge evidence against their claims and still conclude that the claims are not false; they engage in motivated reasoning. Our results indicate that agents can strategically make false or misleading claims in a day-to-day task without being instructed to; the environment provides a controlled testbed for studying how agents resolve such conflicts and for developing deception monitors.
[MA-3] SEER: A Scalable Event-Driven Runtime for Parallel Execution of Multi-Agent Reinforcement Learning Environments
【速读】：该论文旨在解决多智能体强化学习（MARL）系统中，尽管策略推理可在GPU上进行批处理，但CPU侧环境执行路径仍存在显著延迟瓶颈的问题。在交互密集型场景下，环境路径需串行更新实体状态、处理局部交互并构建观测信息，其计算开销成为性能关键制约因素。现有向量化环境接口与通用事件驱动运行时无法有效处理单个MARL步内细粒度的跨智能体协作任务，同时难以维持学习系统所期望的步级语义一致性。为应对这一挑战，论文提出SEER（Scalable Event-Driven Environment Runtime），其核心在于引入显式的事件局部性契约（event-locality contract）：将每个环境步骤建模为带时间戳的事件流，确保直接状态更新为实体本地操作，而跨实体影响通过有序未来事件表达。在此基础上，SEER采用基于比较并交换（CAS）的双层事件队列结构——局部事件队列维护各实体因果顺序，全局执行列表则暴露可并行处理的任务；同时结合STPA（预飞行引导的线程池选择机制），融合短时运行时轨迹与静态工作负载及硬件特征，实现高效调度。实验结果表明，在计算密集型无人机编队控制（Aircraft）和通信密集型无线资源管理（PCS）两个参数化工作负载上，SEER分别相较基线降低16.2%的推理延迟，并在环境执行延迟方面较ROSS-T、NBSE和ROOT-Sim分别降低46.9%、19.5%和64.9%，验证了其在保持步级语义一致性前提下的高可扩展性与高效性。
链接: https://arxiv.org/abs/2611.07996
作者: Jiming Su,Feng Zhu,Yuhao Xiao,Miao Zhang
机构: 未知
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Multiagent Systems (cs.MA)
备注:
Abstract:Multi-agent reinforcement learning (MARL) is increasingly used in systems where many entities interact in a shared environment. In these workloads, policy inference can be batched on GPUs, but the CPU-side environment path still updates entity states, resolves local interactions, and builds observations inside each \mathrmthis http URL . This path can dominate latency in interaction-intensive workloads after policy batching. Existing vectorized environment interfaces and general event-driven runtimes do not directly address the fine-grained cross-entity work inside one bounded MARL step while preserving the step-level semantics expected by learning systems. We present SEER (Scalable Event-Driven Environment Runtime), a runtime for MARL environment execution under an explicit event-locality contract. SEER models each environment step as a bounded stream of timestamped events. If direct state updates are entity-local and cross-entity effects are represented as ordered future events, SEER can run independent events in parallel while preserving serial-equivalent boundary outputs. To make this contract practical, SEER uses a CAS-based dual-layer event queue: local event queues preserve per-entity causal order, while a parallel executive list exposes ready work to global worker threads. SEER further uses STPA, a preflight-guided thread-pool selection mechanism that combines short runtime traces with static workload and hardware features. We evaluate SEER on two parameterized MARL workloads: compute-intensive UAV swarm control (Aircraft) and communication-intensive wireless management (PCS). On Aircraft, SEER reduces inference latency by 16.2% over the baseline. On PCS, SEER reduces environment-execution latency by 46.9%, 19.5%, and 64.9% over ROSS-T, NBSE, and ROOT-Sim, respectively.
[MA-4] ReferClaw: Disambiguating User Intent for Egocentric Assistance
【速读】：该论文旨在解决在第一人称视角人工智能助手（egocentric AI assistants）场景下，用户提出的短促且语义模糊的问题难以通过视频与对话历史唯一确定指代对象的问题。例如，“我把它放哪了？”可能涉及多个此前处理过的物体，现有视频问答（VideoQA）系统通常仅选择单一解释，容易产生符合语法但违背用户意图的回答。其解决方案的关键在于提出一种名为ReferClaw的代理式（agentic）框架，将第一人称流式问答建模为交互过程：助手通过主动询问澄清问题以明确用户意图后再作答。该框架由三个协同工作的智能体组成——候选解释识别、澄清请求生成和基于已解析意图与视觉证据的答案生成；同时引入一个受约束的、成本感知的策略，权衡预期信息增益与用户付出的努力，限制澄清轮次与问题复杂度。ReferClaw通过意图特定技能更新与基于过程奖励引导的策略优化，实现代理层面的信用分配，并从交互中持续学习。在增强的第一人称问答基准上的评估表明，该方法在答案准确率及答案-定位联合性能上优于现有基线，并具备跨数据集的零样本迁移能力，验证了交互式澄清作为第一人称流式辅助中在线用户意图解析的可行性与有效性。
链接: https://arxiv.org/abs/2611.07683
作者: Hangyu Qin,Junbin Xiao,Angela Yao
机构: National University of Singapore(新加坡国立大学); University of Science and Technology of China(中国科学技术大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Users of egocentric AI assistants often ask short, ambiguous questions whose intended referents cannot be uniquely resolved from video and dialogue history. For example, “Where did I leave it?” may refer to several previously handled objects. Existing VideoQA systems typically commit to a single interpretation, risking plausible but intention-misaligned answers. We introduce ReferClaw, an agentic framework that formulates egocentric streaming QA as an interactive process: the assistant asks clarifying questions to resolve user intent before answering. Three cooperating agents identify candidate interpretations, request clarification, and generate answers grounded in the resolved intent and visual evidence. A bounded, cost-aware policy balances expected information gain against user effort while limiting clarification turns and question complexity. Referclaw learns from interaction through intent-specific skill updates and process-reward-guided policy optimization with agent-level credit assignment. Evaluations on augmented egocentric QA benchmarks demonstrate improved answer accuracy and joint answer-and-grounding performance over existing baselines, alongside zero-shot transfer across datasets. These findings support interactive clarification as a practical approach to resolving online user intent in egocentric streaming assistance.
[MA-5] Signed but Not Safe: Requirement Integrity in Multi-Agent LLM Handoffs
【速读】：该论文旨在解决企业级生成式 AI（Generative AI）系统中多智能体协作时因中间传递环节导致需求丢失、篡改或忽略的可靠性问题。在复杂任务分解与结果合并过程中，各子代理间的请求传递若缺乏可信保障机制，极易引入安全风险。其核心解决方案是引入“带签名的需求清单”（signed requirement-manifest）机制：由编排器将每个需求以原文引用（verbatim quote）加可验证谓词的形式提取，并通过 HMAC-SHA256 对需求清单进行数字签名，确保传输过程中的完整性；同时采用确定性代码对合并结果进行验证，实现对决策行为的可审计与可复现检查。实验表明，该方案将攻击下的不安全提交率从 69.4% 降至 9.0%（下降 60.4 个百分点），且未牺牲符合规范的提交数量。签名机制有效抵御了在途篡改攻击——未签名清单在变异运行中 79% 出现违规，而签名版本则为零。此外，确定性验证器相较于大模型检查器，在检出违规决策方面提升至 78.7%（vs. 59.0%），误报率也显著降低至 7%（vs. 50%）。最终分析指出，剩余风险源于签名前的需求提取阶段，即“提取”环节是信任边界所在，预签注入攻击验证了此结论：恶意注入仅需一行即可使签名后的非法清单被认证通过，说明签名无法保护提取阶段的初始污染。因此，该研究的关键在于通过形式化签名与确定性验证构建可信传递链，将信任边界明确界定于需求提取之前。
链接: https://arxiv.org/abs/2611.07483
作者: Xiaobo Chen,Genliang Zhu,Ziyuan Wang,Chu Wang
机构: University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校); Georgia Institute of Technology(佐治亚理工学院); Duke University(杜克大学); University of Illinois Urbana-Champaign(伊利诺伊大学厄本那-香槟分校)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 34 pages, 0 figures. Preregistered; code, task banks, freeze manifests and run logs are available from the authors on request
Abstract:Enterprise AI agents increasingly split a user’s request across subagents and merge their results, and each handoff can summarize away, alter or ignore a requirement. We report a preregistered study with synthetic, exactly scorable tasks comparing free-text, structured JSON and signed requirement-manifest handoffs: the orchestrator extracts each requirement as a verbatim quote plus a checkable predicate, code signs the manifest with HMAC-SHA256, and deterministic code verifies the merger’s decision. With Qwen3-8B (1,440 trajectories, five threats), signed manifests cut unsafe commits under attack from 69.4% to 9.0% (60.4 pp, 95% CI [51.4, 69.4]) with no net loss of conformant commits relative to structured handoffs. Signing is what defeats in-transit tampering: an unsigned manifest committed violations in 79% of mutated runs, the signed one in none. On identical trajectories the deterministic verifier caught 78.7% of violating decisions versus 59.0% for an LLM checker, with 7% versus 50% false alarms. An oracle manifest removed all remaining unsafe commits: the residual risk comes from extraction errors made before signing. A preregistered pre-signing injection attack confirmed this boundary: one adversarial line raised signed-manifest violations from 14.6% to 37.5% (Qwen3-8B), 11.1% to 63.9% (gpt-oss-20b) and 8.3% to 75.0% (GPT-5.6-luna), and the signature then authenticated the corrupted manifest. Signing protects requirements in transit, not at extraction; the extraction step is the trust boundary.
[MA-6] he Last Mile to the Real World: Agent -Native Primitives as Reusable Building Blocks for Solving Real-World Tasks WWW
【速读】：该论文旨在解决大模型在处理真实世界任务时“最后一公里”（last mile to the real world）的瓶颈问题，即模型虽能获取资源但难以有效应用资源完成任务。现有方法将资源置于可访问范围，但任务执行能力仍依赖调用方自行构建，导致每个任务需重复实现相同的能力，效率低下且泛化性差。其解决方案的关键在于提出原生代理原子单元（agent-native primitives），即具备内嵌大语言模型（LLM）以在上下文中自主应用资源并与其他单元通信的可复用组件，并设计了HEELO框架（Harness Engineering around Evolving Agent-Native Primitives for the Last Mile to Real-World Tasks），通过从预定义的原子单元库中组合出任务特定的执行图，自动验证结果、定位失败环节并迭代优化或引入新单元。实验表明，基于八类原子单元的HEELO显著提升多个基准测试表现：在Terminal-Bench 4.0上，使用Codex的GPT-5.6 Sol任务解决率从37.3%提升至55.5%；在Terminal-Bench 2.0上，任务级原子单元使Pass@1从60.0%提升至72.2%；工具类与导航类原子单元分别将ToolBench和GAIA-Text-103的准确率提升19.3和12.7个百分点。在46个真实世界任务上，接入HEELO后，Claude Opus 5.5与GPT-6.1 Sol的任务完成率分别达到39和35，较独立执行分别提升30和35%，同时平均耗时减少约50%，单位任务成本降低45%。
链接: https://arxiv.org/abs/2611.07461
作者: Haibo Jin,Peng Kuang,Xiaopeng Yuan,Lucheng Fu,Jerry Wang,Xucheng Yu,Yaoning Yu,Dehao Wu,Runguo Li,Colin Su,Jingxuan Sun,Xuyang Liu,Haojing Luo,Haohan Wang
机构: University of Illinois Urbana-Champaign; Georgia Institute of Technology; heelo.dev
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA); Software Engineering (cs.SE)
备注: 102 pages, 21 figures, 54 tables. Project page: this https URL
Abstract:How can an agent handle a real-world task end to end? Frontier models with their own harnesses still fall short on real-world benchmarks: on Terminal-Bench 4.0, GPT-5.6 Sol with Codex resolves 37.3% of the tasks and Claude Opus 5 with Claude Code 51.8%. These tasks fail at the step from reaching a resource to applying it, which we call the last mile to the real world. Existing approaches put the resource within reach but leave the competence to apply it with the caller, so each harness rebuilds that competence. We introduce agent-native primitives, reusable units that equip a resource with a resident LLM that applies it in context and communicates with other primitives, and HEELO (Harness Engineering around Evolving Agent-Native Primitives for the Last Mile to Real-WOrld Tasks), a harness engineering framework that builds each task’s harness from a library of such primitives. HEELO composes the primitives a task needs into a task-specific graph, verifies the result against evidence from the environment, revises the primitives a failure implicates, and admits new or improved primitives after validation. We instantiate HEELO with eight primitive families, each built for one kind of task, and evaluate them on more than twenty benchmarks. Dev-Primitives raise Terminal-Bench 4.0 resolution from 37.3% under Codex to 55.5% on the same backbone, Harness Primitives composed per task raise Terminal-Bench 2.0 Pass@1 from 60.0% for the seed harness to 72.2%, Tool Primitives raise ToolBench pass rate from 55.8% to 75.1%, and Navigation Primitives raise GAIA-Text-103 accuracy from 45.6% to 58.3%. On the 46 real-world tasks of our this http URL platform, Claude Opus 5.5 with Claude Code and GPT-6.1 Sol with Codex complete 9 each on their own and 39 and 35 when connected to HEELO through our MCP server, with 1.9 \times and 2.3 \times less time and 1.55 \times lower cost per completed task.
[MA-7] Designing with Cinematic Language: An Intent-Driven Approach to Human-AI Video Clip Co-Creation
【速读】：该论文旨在解决生成式视频模型在创作过程中难以将高层次叙事意图有效转化为连贯的电影化表达这一核心问题。现有系统通常依赖提示词优化或提供孤立的摄像机控制，导致创作者在“表达什么”与“如何拍摄”之间存在显著断层。其解决方案的关键在于提出CineAssist——一个以人为中心的协同创作系统，将电影语言（cinematic language）视为一种可交互的设计媒介，实现创作意图与可执行视频控制之间的精准映射。该系统构建了电影语言设计空间（Cinematic Language Design Space）及基于107个影视制作来源和1,436段影片剪辑的标注数据集，建立了语义意图与具体电影参数值之间的关联。CineAssist通过多智能体工作流实现意图对齐与策略推荐，并集成交互式电影编辑器与上下文感知的偏好学习机制，支持用户主导的个性化定制。三组评估实验验证了系统在端到端可用性、电影质量及用户偏好演化方面的优越表现，证明其能够高效、可控且个性化地实现以意图驱动的电影化视频生成。
链接: https://arxiv.org/abs/2611.07377
作者: Yi He,Zhengxi Yu,Feilin Han,Chuer Chen,Shengqi Dang,Nan Cao
机构: iDVX Lab, College of Design and Innovation, Tongji University (同济大学设计创意学院iDVX实验室); Innovation Center for Intelligent Design, Shanghai Innovation Institute (上海创新研究院智能设计创新中心); Shanghai Research Institute for Intelligent Autonomous Systems (上海智能自主系统研究院)
类目: Multiagent Systems (cs.MA)
备注: 22 pages, 16 figures
Abstract:Generative video models have made video creation more accessible, yet creators still struggle to translate high-level narrative intentions into coherent cinematic expressions. Existing systems primarily rely on prompt refinement or expose isolated camera controls, leaving a gap between what to express and how to film it. We present CineAssist, a human-AI co-creation system that treats cinematic language as an interactive design medium connecting creative intentions with executable video controls. We construct a Cinematic Language Design Space and an annotated dataset from 107 filmmaking sources and 1,436 film clips, linking communicative intentions to cinematic parameter values. CineAssist combines a multi-agent workflow for intent alignment and strategy recommendation with an interactive cinematic editor and a context-sensitive preference learner for user-controlled personalization. We evaluate the system through three studies examining end-to-end usability, cinematic quality, and preference evolution. Results show that CineAssist facilitates efficient, controllable, and personalized intent-driven cinematic video creation.
[MA-8] Design of Multi-UAV Cooperative Deep Semantic Autoencoders for Communication Networks
【速读】：该论文旨在解决低空经济快速发展背景下，无人机（UAV）网络在带宽受限场景下任务执行效率低下与智能处理能力不足的问题。其核心挑战在于如何在有限通信资源约束下实现高效的数据传输与智能语义处理。为此，论文提出了一种多无人机协同深度语义自编码器（Multi-UAV Cooperative Deep Semantic Autoencoders, MC-DSA）架构，其关键创新在于引入基于语义匹配的源选择机制：任务无人机（T-UAV）广播任务查询，各辅助无人机（S-UAV）计算本地图像与查询之间的语义相似度，仅当相似度超过预设阈值时才建立通信链路，从而显著减少冗余传输。该方案通过联合优化语义特征提取与协作通信策略，在保证高质图像有效利用的同时，具备对低质量图像的鲁棒处理能力。训练阶段采用梯度下降法优化网络参数以提升语义表达精度，测试阶段评估协作效率与通信质量。实验结果表明，相较于基准方法，MC-DSA在保持任务执行效率的前提下，显著提升了图像语义分割准确率，为多无人机通信网络提供了高效、智能的解决方案。
链接: https://arxiv.org/abs/2611.07374
作者: Xiaoling Han,Bin Lin,Nan Wu,Tony Q. S. Quek
机构: Dalian Maritime University (大连海事大学); Singapore University of Technology and Design (新加坡科技设计大学)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Driven by the rapid development of the low-altitude economy, Uncrewed Aerial Vehicle (UAV) networks have become increasingly important for executing tasks such as search, rescue, and monitoring. However, under bandwidth-constrained scenarios, UAV networks still face significant challenges in task execution efficiency and intelligent processing capabilities. To address these issues, we propose Multi-UAV Cooperative Deep Semantic Autoencoders (MC-DSA) for communication networks. The architecture comprises one Task UAV (T-UAV) and multiple Support UAVs (S-UAVs), which jointly enhance communication efficiency and data-processing capability via semantic communication and cooperation. Specifically, we introduce a semantic matching and source selection mechanism, enabling the proposed MC-DSA to handle low-quality images while ensuring effective utilization of high-quality images. The T-UAV broadcasts a task query, and each S-UAV computes the semantic similarity between its local image and the query. Communication links are established only when the similarity exceeds a predefined threshold. The training algorithm optimizes network parameters through gradient descent, improving semantic feature extraction accuracy. The testing algorithm evaluates UAV collaboration efficiency and communication quality. Experimental results demonstrate that, compared with benchmarks, the proposed MC-DSA improves image semantic segmentation accuracy while maintaining task execution efficiency. As a result, it reduces redundant transmissions and offers an effective solution for multi-UAV communication networks.
[MA-9] ake It From Here: Evaluating Coding Agents On Project Takeover
【速读】：该论文旨在解决生成式代码代理（Coding Agents）在接手长期项目时，如何从交互历史中恢复当前适用的需求并据此更新项目的问题。现有编码基准普遍缺乏长期交互历史支持，且评估体系未能涵盖从需求重构到项目交付的完整流程，导致对代理“接管能力”的评估不充分。为此，本文提出一种新评估范式：向代理提供完整的交互历史、当前请求及前置项目环境，要求其完成修改并以交付测试评估表现。实验表明，在包含完整历史信息的条件下，六种模型配置在37个可执行验证目标上的平均完整交付率仅为4.95%。通过引入需求原子（requirement atoms）进行独立诊断，进一步分析了代理在相关信源识别、前置状态重建和当前请求处理三个环节的表现；当提供正确前置或目标状态后，平均完整交付率提升至8.11%，构建与回归测试通过率分别达96.40%和90.99%，但目标测试成功率仍仅9.01%。这揭示出代理在形成正确的目标需求状态以及将目标需求准确实施于现有项目中仍存在显著困难。基于上述发现，本文提出ReqMemBench，一个融合需求诊断与交付验证的系统性评估框架，旨在全面衡量代码代理的项目接管能力。
链接: https://arxiv.org/abs/2611.07307
作者: Dingxian Wang,Zhe Fu,Jiaqi Deng,Chao Yin,Ganesh Prabhu,Karthik Gomadam Rajagopal,Darwin Yi,Huan Huo,Guandong Xu,Xianzhi Wang,Andrew Rabinovich,Zhao Chen
机构: Upwork Inc(优步工作平台); University of Technology Sydney(悉尼科技大学); Scale AI; The Education University of Hong Kong(香港教育大学)
类目: oftware Engineering (cs.SE); Multiagent Systems (cs.MA)
备注: 35 pages
Abstract:Coding agents increasingly collaborate on long-term projects. When taking over a project, they need to recover currently applicable requirements from the interaction history and update them in response to new requests. However, most coding benchmarks do not include long interaction histories, and some evaluations that consider history still do not cover the full process from requirement reconstruction to project delivery. We therefore provide agents with the interaction history, current request, and pre-request project environment, ask them to make the requested changes, and assess takeover performance through delivery tests. Across six model-harness configurations and 37 targets eligible for executable validation, the mean complete-delivery rate under Full History is only 4.95%. We then use requirement atoms to independently diagnose a larger set of takeover targets, examining relevant information identification, pre-request state reconstruction, and current request processing. We further supply correct pre-request or target states to examine the remaining difficulties in requirement processing and project modification. The diagnostics show that none of the six configurations produces fully correct post-request requirement states, and a case illustrates omission of previously recovered constraints during updating. With correct target states supplied, the mean complete-delivery rate is 8.11%, with build and regression pass rates of 96.40% and 90.99%, but target-test success remains 9.01%. These results show difficulties in both forming target requirement states and implementing supplied targets in existing projects. Based on these findings, we introduce ReqMemBench, which combines requirement diagnostics with delivery validation to support systematic evaluation of project takeover.
[MA-10] More Agents More Reliable? System-Level Reliability of Multi-Agent LLM s under Operational-Profile Uncertainty
【速读】：该论文旨在解决多智能体系统（multi-agent systems）在实际部署中可靠性评估失真的问题，即当前普遍假设“更多智能体意味着更高可靠性”的误区。现有评估方法依赖平均基准准确率，忽视了真实输入分布的复杂性以及各智能体输出合并规则（combining rule）本身可能存在的失效风险。为应对这一挑战，作者提出一种融合分层不精确概率评估（HIP-LLM）与系统故障逻辑贝叶斯网络（Bayesian network）的框架，实现对智能体失败概率区间的传播分析。该框架揭示了合并规则自身失效对整体系统可靠性的影响，并生成一个可靠性包络（reliability envelope），用于界定不同证据下系统的可信可靠性边界。实验表明，在提取受监管化学危害声明的流水线任务中，三模型多数投票虽能抵御最弱模型的干扰，但未能显著超越最优模型；而当投票机制在5%的请求中失效时，其相对优势几乎完全丧失。研究结论指出，增加智能体数量并不自动提升系统可靠性，而是通过提供更稳健的可靠性论证（more defensible reliability claim）来增强可信度。该方案的关键在于将智能体失败概率区间与系统级故障逻辑建模相结合，从而实现对可靠性不确定性的量化与传播。
链接: https://arxiv.org/abs/2611.07140
作者: Burkay Tunçtürk,Koorosh Aslansefat,Yiannis Papadopoulos,Bao Nguyen
机构: University of Hull(赫尔大学); Bilkent University(比尔肯大学); University of Leeds(利兹大学)
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
备注: Submitted for IEEE Software Magazine
Abstract:Teams increasingly build software from cooperating large language model (LLM) agents, often assuming more agents mean more reliability. Yet the usual evidence, average benchmark accuracy, ignores the input mix a deployed system meets and the rule merging the agents’ answers. Building on HIP-LLM, a hierarchical imprecise-probability assessment of single LLMs, we propose a framework combining it with a Bayesian network of the system’s failure logic for multi-agent systems. Agent failure-probability intervals propagate through this network, where the combining rule can itself fail, into a reliability envelope showing which evidence limits it. In a pipeline extracting regulated chemical-hazard statements, majority voting over three models protected against the weakest but did not clearly exceed the best, and a voter failing on 5% of demands would largely erase its advantage over the middle model. More agents yield a more defensible reliability claim, not automatically a more reliable system.
[MA-11] Regulating Autonomous Mobility-on-Demand Operators with Fixed and Adaptive Penalties: A Multi-Agent Reinforcement Learning Approach
【速读】：该论文旨在解决利润最大化型自动驾驶按需出行（AMoD）运营商通过动态溢价定价（surge pricing）攫取消费者剩余，以及将车辆从低盈利区域转移导致服务覆盖不足的问题。其核心解决方案在于引入基于惩罚机制的监管框架，通过在多模式交通系统中构建外部控制环路来约束运营商行为。关键创新在于将监管建模为两种形式：静态监管采用固定惩罚的马尔可夫决策过程（MDP），用于联合优化定价与车辆再平衡；动态监管则采用部分可观测随机博弈（Partially Observable Stochastic Game），使监管者能够根据实时状态周期性调整区域特定惩罚，实现闭环调控。通过基于图的多智能体强化学习方法联合训练运营商与监管者策略，在真实曼哈顿数据上的实验表明，固定溢价惩罚可将遭遇过高价格的乘客比例从22%降至1%；而动态监管相比同均值强度的固定惩罚，可使监管者收益提升最高达15.6%，尤其在车队冗余容量充足时服务可用性改善显著；当车队在高峰时段高度饱和时，服务缺口仍存在，此时学习到的监管策略会自动降低惩罚以适应运力限制。研究揭示了监管目标、运营商盈利与出行者效用之间的权衡关系。
链接: https://arxiv.org/abs/2611.07122
作者: Korbinian Kügler,Carolin Schmidt,Florian Dandl,Filipe Rodrigues
机构: Technical University of Munich(慕尼黑工业大学); Technical University of Denmark(丹麦技术大学)
类目: Machine Learning (cs.LG); Multiagent Systems (cs.MA)
备注:
Abstract:Profit-maximizing Autonomous Mobility-on-Demand (AMoD) operators may exploit consumer surplus through excessive prices and underserve less profitable regions by relocating vehicles away from them. We study how penalty-based regulation of an AMoD operator in a multimodal mobility system can limit surge pricing and improve regional service availability. We model regulation as an outer control loop that acts on a learning-based operator. Static regulation with fixed penalties is formulated as a Markov Decision Process for an operator that jointly controls pricing and rebalancing, and dynamic regulation as a Partially Observable Stochastic Game in which a regulator periodically adjusts region-specific penalties in closed loop. Operator and regulator policies are learned with graph-based multi-agent reinforcement learning. In experiments with real-world Manhattan data, fixed surge pricing penalties reduce the share of travelers facing excessive prices from 22% to 1%. For regional service availability, dynamic regulation improves the regulator’s reward by up to 15.6% over fixed penalties of equal mean intensity. The gains in service availability are strongest when spare fleet capacity exists, whereas service deficits persist in peak periods with highly utilized fleets. The learned regulator policy responds by reducing penalties where limited fleet capacity restricts compliance. The results identify trade-offs between regulatory targets, operator profitability, and traveler utility.
[MA-12] Polynomial-Time Computation of Extensive-Form Proper Equilibria in Two-Player Zero-Sum Games
【速读】：该论文旨在解决在具有完美回忆和有理数数据的有限双人零和广义形式博弈中，如何在确定性多项式时间内计算一个精确的扩展形式适当均衡（extensive-form proper equilibrium）的问题。其核心挑战在于确保均衡满足局部适当性条件：在每个信息集上，相较于更优动作，当前动作需受到惩罚，而其他局部行为选择（包括自身未来行动）保持不变。解决方案的关键在于引入几何对称多面体（permutahedron）来表征每个信息集上允许的动作概率空间，并通过缩放的双重随机表示将这些约束嵌入一个多项式规模的序列形式多面体（sequence-form polytope）中；随后，鞍点解可满足局部ε-适当性不等式。为获得精确均衡而非任意小数值扰动下的近似解，作者建立了参数线性系统的符号稳定性，构造了持久可行基，并提取了行为比例的极限值，最终得到具有多项式编码长度的有理数概率分布，即使在实现概率为零的信息集上亦成立。
链接: https://arxiv.org/abs/2611.07089
作者: Sam Ganzfried
机构: Ganzfried Research(甘兹弗里德研究); Cornell University (康奈尔大学)
类目: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Theoretical Economics (econ.TH)
备注:
Abstract:We give a deterministic polynomial-time algorithm for computing an exact extensive-form proper equilibrium in a finite two-player zero-sum extensive-form game with perfect recall and rational data. Properness is understood in the agent normal form: an action at an information set is penalized relative to a better action while all other local behavioral choices, including the player’s own future actions, remain fixed. Our algorithm assigns each information set a geometric permutahedron of allowable action probabilities. A scaled doubly stochastic representation embeds these restrictions in a polynomial-size sequence-form polytope; a saddle point then satisfies the local \varepsilon -properness inequalities. To obtain an exact equilibrium rather than a solution at an arbitrarily small numerical perturbation, we establish sign stability of a parametric linear system, construct a persistent feasible basis, and extract the limits of behavioral ratios. The resulting probabilities are rational and have polynomial encoding length, including at information sets whose realization probabilities vanish.
[MA-13] Cooperative Multi-Agent Reinforcement Learning via Epigraph-Form Guided Exploration NEURIPS2026
【速读】：该论文旨在解决多智能体强化学习（MARL）中由于联合状态-动作空间的组合复杂性导致协同行为难以通过试错机制自发涌现的问题。现有方法常依赖内在奖励（intrinsic reward）促进探索，但直接将内在奖励与团队目标结合会扭曲学习信号，进而损害任务性能。为此，本文提出EFXPLORER，一种受约束的探索框架，其核心在于在保证已建立的任务性能的前提下最大化探索目标。解决方案的关键是采用等式重构（epigraph reformulation），引入自适应探索预算，实现内在奖励与任务目标的解耦，并通过任务可行性来调控探索强度。为进一步激励多样化且具有长期时序依赖性的探索行为，还设计了基于后继距离（successor distance）的内在奖励机制，以捕捉长程依赖关系。实验结果表明，该方法在SMAX、VMAS和MPE等多个基准测试中均显著优于现有强基线，成功诱导出新颖的协作策略。
链接: https://arxiv.org/abs/2611.07066
作者: Sungil Son,Hoseong Jung,Dahyun Oh,H. Jin Kim
机构: Seoul National University(首尔国立大学)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: NeurIPS 2026
Abstract:Discovering cooperative behavior in multi-agent reinforcement learning (MARL) is challenging due to the combinatorial complexity of joint state-action spaces, which hinders the emergence of coordinated behaviors from trial-and-error alone. Intrinsic rewards are often used to aid discovery, but naively combining them with team objectives can distort the learning signal, compromising task performance. In this paper, we propose EFXPLORER, a constrained exploration framework that maximizes exploration objectives subject to a constraint that preserves established task performance. We solve the resulting problem using an epigraph reformulation that introduces adaptive exploration budgets. This approach separates intrinsic rewards from task objectives and regulates exploration through task feasibility. To further encourage diverse and temporally extended exploration, we incorporate a successor distance-based intrinsic reward that captures long-horizon dependencies. Empirically, our method outperforms strong baselines and induces novel cooperative strategies across SMAX, VMAS, and MPE benchmark suites.
[MA-14] Authorization Propagation in Agent 2Agent Systems: An Experimental Security Evaluation of Multi-Hop Delegation
【速读】：该论文旨在解决委托代理间（Agent-to-Agent, A2A）工作流中因请求发起方与权限原始主体不一致而引发的安全问题，尤其关注授权传播过程中的信任传递与语义完整性。其核心解决方案在于采用分层的下游验证设计，通过将当前调用者上下文、传递性授权证据、每跳基于Ed25519的请求认证以及独立的重放状态分别隔离处理，实现对请求与授权凭证的细粒度控制。关键创新点在于将请求认证与继承的授权解耦，确保每个代理节点可独立验证请求合法性，同时识别出“起源可验证的委托路径”与“语义授权衰减”作为强化A2A授权传播安全性的必要条件。实验表明，在15种对抗场景中14种被成功拒绝，且在固定深度本地基准测试中，系统具备5.11毫秒的中位安全处理延迟和3.87 kB的序列化负载，验证了该方案在安全性与性能间的良好平衡。
链接: https://arxiv.org/abs/2611.07050
作者: Gabriele Donno,Kashif Mehmood,Katina Kralevska
机构: NTNU—Norwegian University of Science and Technology (挪威科技大学)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Delegated Agent2Agent (A2A) workflows create a security problem when the agent sending a request is different from the principal from which the authority originated. This paper evaluates a layered downstream-verification design in a synthetic four-party healthcare reimbursement workflow. The prototype separates current-caller context, transitive authorization evidence, per-hop Ed25519 request authentication, and independent replay state for requests and authorization artifacts. Of 15 adversarial conditions, 14 are rejected; an isolated delegation-path test exposes the absence of semantic authority attenuation. A separate authorized-intermediary experiment shows that a holder of the shared authorization-evidence HMAC key can modify the represented authority and generate valid evidence. In the fixed-depth local benchmark, the full configuration has a median security-processing latency of 5.11 ms and a 3.87 kB serialized payload. The results show the value of separating request authentication from inherited authority while identifying origin-verifiable delegation and semantic attenuation as requirements for stronger A2A authorization propagation.
[MA-15] he Two-Ledger Model: Transaction-Grain Cost Accounting for Multi-Agent AI Systems Across Clouds Models and Systems of Record
【速读】：该论文旨在解决企业在使用多智能体生成式AI（Generative AI）系统进行业务交易时，因计费数据粒度与业务交易边界不匹配而导致的可追溯性与成本归因难题。核心问题在于：云服务提供商仅提供以结算粒度（settlement grain）聚合的账单数据（按SKU和计费周期汇总），无法识别具体订单、索赔或案件的生成式AI实际消耗成本，导致企业难以实现精细化的财务运营（FinOps）管理。其解决方案的关键是提出“双账本模型”（Two-Ledger Model），将符合FOCUS规范的结算账本（基于聚合账单数据）与基于OpenTelemetry GenAI遥测构建的溯源账本（以调用粒度，invocation-grain，记录使用行为）相融合。该模型通过分层相关性脊柱（trace context、A2A/MCP元数据、业务键关联、概率匹配）实现调用记录与业务事务的高精度映射，并结合版本化费率卡实现实时定价，最终通过周期对账将两个账本关联至发票，未解释残差可识别“影子AI”（shadow AI）支出。仿真结果表明，即使在追踪链中断（trace propagation降至零）的情况下，该模型仍能维持超过99%的成本覆盖度，残差可精准恢复预设的未计量支出（误差低于0.24个百分点），并有效识别供应商账单错误；同时，飞行中拦截机制可在2-4%的误暂停率下，将第99百分位交易成本降低高达61%。研究进一步建议将溯源账本作为标准化的、以调用粒度为基础的FOCUS扩展配置文件，以推动行业统一的生成式AI成本与用量度量标准。
链接: https://arxiv.org/abs/2611.06989
作者: Vivek Acharya
机构: 未知
类目: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Multiagent Systems (cs.MA)
备注: 16 pages, 5 figures, 12 tables. Code and synthetic data: this https URL
Abstract:Enterprises run business transactions through multi-agent AI systems spanning clouds, large language models (LLMs), orchestration harnesses, and systems of record. Finance receives AI cost only at settlement grain, aggregated by SKU and charge period. The business question (what did this order, claim, or case cost in AI?) cannot be answered from billing data, because providers never see enterprise transaction boundaries. The FinOps Open Cost and Usage Specification (FOCUS) gives billing data one schema across providers but remains settlement-grain even as it adds AI token and requester attribution. We formalize this grain mismatch, derive seven requirements, and propose the Two-Ledger Model: a settlement ledger of FOCUS-conformant billing data joined to an attribution ledger of invocation-grain usage records built on OpenTelemetry GenAI telemetry. A layered correlation spine (trace context, A2A/MCP metadata, business-key correlation, probabilistic matching) assigns each record to its business transaction with recorded method and confidence; a versioned rate card prices usage in near real time; period reconciliation ties both ledgers to the invoice, and the unexplained residual identifies shadow AI. Costs are reported per order, claim, or outcome with agentic cost metrics. A simulation with known ground truth shows the spine sustaining cost coverage above 99% as trace propagation falls to zero, the reconciliation residual recovering planted unmetered spend within 0.24 percentage points while flagging provider billing errors, and mid-flight gating cutting p99 transaction cost by up to 61% at a 2-4% false-pause rate. We propose the attribution ledger as an invocation-grain FOCUS profile for standardization.
[MA-16] A Bilevel Neurodynamic Approach for Distributed SCAD-Regularized Sparse Least Squares
【速读】：该论文旨在解决分布式稀疏最小二乘问题在平滑截断绝对偏差（SCAD）正则化下的全局最优求解难题。由于SCAD惩罚项具有非光滑性和非凸性，现有分布式神经动力学方法无法直接推广至该场景。其解决方案的关键在于构建一种基于双层粒子群优化（PSO）-神经动力学的协同优化框架：首先通过双时间尺度奇异摄动分析，建立了无需二阶充分条件即可保证收敛性的内层多智能体神经动力系统全轨迹收敛理论，并给出明确的收敛速率；其次，在外层引入“红绿灯”敏感性机制，指导PSO搜索过程以增强全局探索能力，同时聚焦于SCAD形状参数的优化，从而在提升稀疏恢复精度的同时显著提高计算效率。仿真结果表明，所提方法在四种稀疏恢复场景中均能准确重构真实稀疏结构，且收敛速度更快、相对误差更低。
链接: https://arxiv.org/abs/2611.07618
作者: Juntao Li,Qihang Guo,Deyuan Meng
机构: Henan Normal University (河南师范大学); Beihang University (北京航空航天大学)
类目: Optimization and Control (math.OC); Multiagent Systems (cs.MA)
备注:
Abstract:This paper proposes a bilevel particle swarm optimization (PSO)-neurodynamic method for solving the distributed sparse least-squares problem regularized by the smoothly clipped absolute deviation (SCAD) penalty to global optimality. The SCAD penalty is nonsmooth and nonconvex, so existing distributed neurodynamic methods have not been extended to it directly. Through two-time-scale singular perturbation analysis, a rigorous full-trajectory convergence theory with explicit convergence rates is established for the inner collaborative multi-agent neurodynamic system without requiring the second-order sufficient condition. A theory-guided inner-outer collaborative optimization framework is further developed, in which a red-light green-light sensitivity mechanism is embedded in the outer PSO search to enhance global search while concentrating computational effort on the SCAD shape parameters, thereby improving recovery accuracy and computational efficiency. Simulation results on four sparse-recovery scenarios demonstrate that the proposed method accurately recovers the true sparsity pattern and achieves faster convergence and lower relative error than competing neurodynamic methods.
自然语言处理
[NLP-0] Scientist World: Grow a Research World for AI-Scientists
【速读】： 该论文旨在解决生成式AI在科学研究中缺乏可执行、可评估的实验环境这一核心问题。当前，尽管生成式AI（Generative AI）已具备一定的科研能力，但其无法直接在人类科学家积累的科学文献中“运行”——这些文献虽可读却不可执行，导致AI难以真正参与或超越人类研究。为此，本文提出构建“科学世界”（Scientist World），将每一篇研究论文转化为一个可执行的研究环境：通过自动化流程提取论文中的开放性研究问题，并复现其提出的方法作为基线；在此环境中，AI-科学家可被评估其表现与原始作者成果的对比。该工作释放了首批1000个经验证的可执行环境，实证表明前沿水平的AI-科学家已在74%的情况下超越人类研究人员。该系统不仅作为记录人工智能开展科学研究过程的档案（包含6400小时的研究轨迹，涵盖所有想法、实验、失败与成功），更成为训练新模型的有效平台——通过该环境，小型模型的研究构思能力提升了158%，达到与前沿模型相当的水平。因此，解决方案的关键在于构建一个可执行、可评估且持续演化的科学生态，使AI-科学家不仅能“学习”科学，还能在真实研究环境中“成长”并反哺科学本身。
链接: https://arxiv.org/abs/2611.08478
作者: Taiming Lu,Sachin Konan,Zhuang Liu
机构: Princeton University (普林斯顿大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:AI-Scientists arrived, but a world for them to grow in did not. Human scientists take theirs for granted - centuries of open problems and shared ideas, waiting in the scientific literature. To AI-Scientists, that world is readable but not runnable, so we build Scientist World for them: every research paper becomes an executable environment. Our autonomous pipeline extracts each paper’s open-ended research question and reproduces its proposed method as the baseline; an AI-Scientist entering the environment is evaluated against what the human authors achieved. We release the first 1,000 verified environments. With the world as an evaluation, we found that frontier AI-Scientists can already surpass the human researchers 74% of the time. As a record of how AI does science, it contains 6,400 hours of research traces, with every idea, experiment, failure, and success. As a training ground, we are able to raise (+158%) a small model to match frontier models in research ideation. At last, AI-Scientists’ own world has arrived, grown from and growing with science itself.
[NLP-1] Unpacking the Compute Efficiency of Strong-to-Weak OPD against GRPO
【速读】： 该论文旨在解决生成式 AI 模型训练中高计算成本与高效知识迁移之间的矛盾，具体聚焦于在保证目标精度的前提下，如何减少强化学习（Reinforcement Learning, RL）或基于教师-学生架构的策略蒸馏（如 on-policy distillation, OPD）所需的训练浮点运算量（FLOPs）。其核心问题是：相较于传统的强化学习方法，是否可通过优化的教师-学生蒸馏框架，在更低的计算开销下实现相当甚至更优的模型性能。解决方案的关键在于采用一种多教师蒸馏策略——多教师 OPD（Multi-teacher OPD），即利用多个领域专用的 GRPO（Group Relative Policy Optimization）专家模型作为教师，将各领域专长统一蒸馏至一个通用学生模型中。研究发现，当教师为学生自身历史训练的 GRPO 检查点时，OPD 能以高达 50 倍的效率恢复对应精度；进一步扩展至多领域场景，该方法可在仅消耗 5.2%~8.0% 的专家模型训练 FLOPs 的情况下完成知识融合，整体训练流程相较联合训练的 GRPO 减少 42.0% 的总计算量，显著提升了跨域通用模型构建的能效比。
链接: https://arxiv.org/abs/2611.08453
作者: Zeyu Leo Liu,Yang Li,Bo Pang,Emre Can Acikgoz,Xuan-Phi Nguyen,Anurag Koul,Srijan Bansal,Erik Nijkamp,Shafiq Joty,Semih Yavuz
机构: Salesforce AI Research
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:On-policy distillation (OPD) trains a student on its own rollouts under a frozen teacher’s per-token supervision. Does this dense signal reduce the training FLOPs needed to reach a target accuracy compared with reinforcement learning? We systematically compare vanilla reverse-KL OPD with Group Relative Policy Optimization (GRPO) across task difficulty, student and teacher size, domain, and optimization choices, measuring held-out accuracy against training FLOPs. With an external post-trained teacher, OPD generally loses to GRPO in accuracy per training FLOP at matched settings, even when the teacher is further trained with GRPO, and often suffers response-length inflation. Yet a different teacher changes the picture: when the teacher is the student’s own GRPO checkpoint, OPD recovers that checkpoint’s accuracy on math with 50\times fewer distillation FLOPs than GRPO needs to reach the same accuracy. Following the same teacher setting, we consistently observe this high efficiency with a larger student, on code, and in a second model family. On a single domain, however, the GRPO checkpoint is already the desired model, so distilling it adds to total compute. Across domains, separate GRPO checkpoints hold different capabilities, and no single checkpoint serves every domain. Multi-teacher OPD (DeepSeek-AI, 2026) consolidates per-domain GRPO experts into a single student that matches each expert’s accuracy, with the distillation stage costing only 5.2% (on 2 domains) and 8.0% (on 4 domains) of the specialist training FLOPs. Counting specialist training and distillation together, this pipeline reaches joint GRPO’s performance on the 2-domain mix with 42.0% fewer total FLOPs.
[NLP-2] Label-Wording Augmentation in a Typed-Decision Encoder: Gains on Unfamiliar Words Not on Abbreviations or Codes
【速读】： 该论文旨在解决生成式模型在决策任务中因标签表述（label wording）不可控而导致的准确性下降问题，尤其关注多语言交叉编码器（cross-encoder）在实际应用中对标签变体敏感所带来的性能波动。其核心解决方案在于引入一种标签表面形式增强（augmentation that presents labels in several surface forms），通过在推理阶段以多种表达方式呈现标签，提升模型对不同表述的鲁棒性。实验表明，在使用普通词汇、借词（如英语术语）、职场俚语及部门名称等非标准标签时，经过增强的版本在所有设置下均表现出更高准确率，最大提升达0.24，其中11项提升在统计校正后仍显著；尤其在术语类标签上，标签层级的置信区间不包含零，证明其有效性。然而，在缩写和模糊代码场景下，增强版本未带来可用性能增益，甚至在部分情况下表现更差。此外，研究还揭示了对比两个版本时存在的四项评估陷阱，提示需谨慎解读模型改进效果。
链接: https://arxiv.org/abs/2611.08439
作者: Vladimir Beskorovainyi
机构: Besk Tech
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 11 pages, 6 tables. Model: this https URL ; code: this https URL ; archived release: this https URL
Abstract:Typed-decision models score options that the caller supplies at request time, so their accuracy depends on label wording that the model’s developer does not control. We compare two releases of a 307M-parameter multilingual cross-encoder, Ru-Jev, trained on thirteen Russian-language tasks; the later release adds an augmentation that presents labels in several surface forms. Holding questions fixed and varying only how eighteen routing labels are worded, we evaluate ten label vocabularies, produced by three separate processes and checked for overlap with the training data, with five candidates per question and, for eight of them, also with eighteen. On held-out vocabularies made of ordinary or borrowed words (English-derived jargon, office slang, department names), the augmented release is more accurate in every setting, by up to 0.24. Eleven of these fourteen gains survive a Holm correction (slang with eighteen candidates and descriptive sentences with five do not), and for jargon on all labels the label-level interval excludes zero in both settings. On abbreviations and opaque codes it gives no usable gain: on abbreviations it is not better, and worse on the clean slice, and on opaque codes both releases stay at or below chance once labels related to training options are removed. In most cells where accuracy rises, the risk-coverage area falls, but with eighteen candidates confident errors rise on jargon and slang. On two datasets neither release saw, accuracy moves by -8.4 and +5.6 points. We also report four evaluation hazards from comparing the two releases.
[NLP-3] Understanding AI from the Inside: The State of Mechanistic Interpretability for AI Safety
【速读】： 该论文旨在解决当前生成式人工智能（Generative AI）系统日益增强的自主性与不可控性所带来的安全风险问题，核心挑战在于：尽管模型能力不断提升，但其内部工作机制仍缺乏透明性，导致难以预测其在未知场景下的行为或实现可靠控制。论文提出的关键解决方案是通过机制可解释性（mechanistic interpretability），从模型内部的表征结构与计算过程入手，揭示其决策逻辑的本质。其关键在于构建可验证的内部因果机制（如神经元特征、计算电路等），以实现对模型行为的预测、干预与监控。研究系统梳理了机制可解释性的理论基础（如特征解耦、多义性、Transformer计算机制与电路发现）、方法体系（包括特征提取、因果验证、自然语言解释、训练动态分析等），并提出一个统一的评估框架来衡量可解释性主张的有效性。同时，论文探讨了机制理解在知识编辑、行为引导和持续监控等安全应用中的潜力，也指出了其局限性与尚未解决的开放问题，强调机制可解释性应与行为层面的安全评估相辅相成，共同推动安全可控的人工智能发展。
链接: https://arxiv.org/abs/2611.08438
作者: Fazl Barez,Tung-Yu Wu,Hunar Batra,Gabriele Dominici,Florent Draye,Yanai Elazar,Damiano Fornasiere,Antía García,Abir Harrasse,Riley Harris,Sara Hooker,Michael Lan,Belinda Z. Li,Aaron Maiwald,Narmeen Fatimah Oozeer,Naomi Saphra,Or David Shafran,Adi Simhi,Robert Trager,Shriyash Kaustubh Upadhyay,Ziyu Yao,Yu Zhao,Shay B. Cohen,Mor Geva,Yoshua Bengio,Stella Biderman,Hassan Sajjad
机构: University of Oxford; Martian; MIT; Anthropic
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Neural networks are likely to be among the most consequential artefacts in human history. As these systems become more capable and autonomous, their potential to cause societal harm also grows, making AI safety one of the defining challenges of our time. Yet we continue to build increasingly powerful systems without fully understanding how they work, how they will behave in unfamiliar situations, or how to reliably control them. Mechanistic interpretability offers a path towards understanding these systems by studying their internal representations and computations, with the aim of evaluating, predicting, and controlling behaviour. We survey mechanistic interpretability for AI safety in order to answer the question: Does understanding how models work actually make them safer? We place mechanistic interpretability in the broader context of interpretability research and compare its mechanistic evidence with behavioural evaluation. We introduce the foundations of mechanistic interpretability, including features, polysemanticity, transformer computation, and circuits, and survey methods for feature extraction and analysis, circuit discovery and causal verification, explaining intermediate representations in natural language, learning dynamics, and training-data attribution. Beyond surveying existing methods, we propose a unified framework for evaluating mechanistic interpretability claims and characterise six research orientations by their goals, assumptions, and trade-offs. Finally, we examine how mechanistic understanding can support AI safety, including knowledge editing, steering, and monitoring, and assess its limitations, failure modes, open problems, and relationship to behavioural safety measures. We also release a repository of code notebooks with worked examples that reproduce key findings from surveyed methods, allowing readers to build intuition through hands-on experimentation.
[NLP-4] MedGraph: Maintaining Patient State with Graph Memory for Long-Term Medical Dialogue
【速读】： 该论文旨在解决基于大语言模型（LLM）的医疗助手在跨就诊场景中维持患者临床状态记忆准确性的关键问题。现有记忆系统在纵向医疗对话中存在三类反复出现的错误：检索到的证据反映过时状态、混入了他人健康信息，或缺失回答所依赖的安全性事实。为应对这些问题，论文提出MedGraph——一种专为此三类问题设计的图结构记忆系统。其核心解决方案在于：通过从每轮会话中提取带时间戳的事实，将每个事实归属至对应患者，并将药物、诊断与测量记录构建为保持历史状态的时间链；在检索阶段，以患者的时间链为单位进行排序，按问题所涉时间点读取相应链段，并结合原始对话文本进行上下文增强；同时构建患者档案，统一列出所有已记录的安全性事实以供问答参考。实验表明，在MedMemoryBench基准测试中，无论采用GPT还是Qwen作为主干模型，MedGraph均显著优于五种基线方法，分别提升10.5和10.4分，且在内存构建与推理效率方面表现优异。
链接: https://arxiv.org/abs/2611.08433
作者: Donghua Zhang,Zizheng Wang,Ze Dong,Yiding Feng,Siyu Qian
机构: Harvard University (哈佛大学); University of Michigan (密歇根大学); Georgetown University (乔治城大学); Johns Hopkins University (约翰霍普金斯大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 14 pages, 5 figures, 8 tables
Abstract:LLM-based medical assistants that follow a patient across visits need a memory that keeps the patient’s clinical state correct as it changes. We examine the errors of five memory systems on longitudinal medical dialogue and find three recurring causes: the retrieved evidence gives an outdated state, includes another person’s health facts, or lacks a safety fact the answer depends on. We present MedGraph, a graph memory built for these three cases. Construction extracts dated facts from each session, attributes each fact to the person it describes, and links medication, diagnosis, and measurement records into chains that keep earlier states. Retrieval ranks the patient’s chains as units, reads each chain at the time the question asks about, and adds the original dialogue text; a patient profile lists every recorded safety fact with every question. On MedMemoryBench, MedGraph has the highest overall accuracy with both a GPT and a Qwen backbone, 10.5 and 10.4 points above the strongest of five baselines, with efficient memory construction and answering in the GPT setting.
[NLP-5] StepBack: Step-Level Error-Localized Resampling for Efficient Test-Time Reasoning NEURIPS2026
【速读】： 该论文旨在解决自洽性（Self-consistency, SC）推理策略在测试时推理过程中计算开销大、资源消耗高的问题，尤其是其需要大量完整推理轨迹采样的缺陷。现有高效变体虽通过提前终止采样或分散采样数量来降低代价，但仍难以在保持高准确率的同时显著减少计算成本。本文提出一种名为StepBack的步骤级错误定位重采样方法，其核心在于：基于少量初始草稿推理轨迹，利用轻量级信号精准定位最可能产生首次错误的推理步骤，并仅对出错步骤及其后续部分进行重采样，最后通过多数投票确定最终答案。该方法在七项数学推理与多跳问答基准上平均性能与SC相当，同时可将令牌使用量最多降低41.26%。此外，由于其设计特性，StepBack可与自适应推理方法（如Adaptive Consistency）无缝结合，进一步提升推理效率。
链接: https://arxiv.org/abs/2611.08403
作者: Yang Ouyang,Jung-Eun Kim
机构: North Carolina State University (北卡罗来纳州立大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: NeurIPS 2026
Abstract:Self-consistency (SC) is a widely used decoding strategy for improving chain-of-thought reasoning, but it requires sampling many complete reasoning trajectories with a preset budget. Its efficient variants reduce the cost mainly by stopping full-trajectory sampling early or distributing the number of samples across questions. In this paper, we propose StepBack, a step-level error-localized resampling method for efficient test-time reasoning. Given a small set of draft trajectories, StepBack localizes likely first error positions using lightweight signals and then resamples the detected error step and its suffix, followed by majority answer voting. Empirically, StepBack matches the accuracy of SC on average across seven mathematical reasoning and multi-hop QA benchmarks, while reducing token usage by up to 41.26%. In addition, by nature, StepBack can be combined with adaptive inference methods such as Adaptive Consistency, further improving their inference-time efficiency.
[NLP-6] SpecCTMC: Exact Speculative Simulation of Diffusion Language Model Continuous-Time Markov Chains
【速读】： 该论文旨在解决生成式AI（Generative AI）中基于掩码扩散语言模型（Masked Diffusion Language Models）进行并行文本生成时，因独立采样多个词元（token）而破坏其依赖关系、改变输出分布的问题。现有精确的连续时间采样方法虽可避免近似误差，但通常需对每个词元进行一次串行模型评估，形成计算瓶颈。为克服这一问题，论文提出一种无需训练的推测采样方法SpecCTMC，其核心在于利用先前已计算的模型预测结果，通过一系列连续的词元更新策略，在并行评估中间状态的基础上，结合精确的接受-修正（acceptance-and-correction）机制完成采样。该方法保证了底层连续时间马尔可夫链（Continuous-Time Markov Chain, CTMC）中事件发生时间、生成顺序与词元取值的联合分布不变，从而既无时间离散化误差，也无并行解码引入的额外因子分解误差。在LLaDA-8B-Instruct和Dream-v0-Instruct-7B across five benchmarks上的实验表明，SpecCTMC相比直接连续时间采样减少了1.73–4.38倍的串行模型评估次数，同时保持了参考采样分布的准确性。
链接: https://arxiv.org/abs/2611.08362
作者: Siva Rajesh Kasa,Yasong Dai,Sumit Negi,Hongdong Li
机构: 未知
类目: Computation and Language (cs.CL)
备注: 28 pages, 5 figures
Abstract:Masked diffusion language models offer parallel text generation, but independently sampling multiple tokens can disrupt their dependencies and change the output distribution. Exact continuous-time sampling avoids this approximation, yet typically requires one sequential model evaluation per token. We introduce SpecCTMC, a training-free speculative sampler that reduces this sequential bottleneck without a separate draft model. SpecCTMC proposes successive token updates using previously computed model predictions, evaluates their intermediate states in parallel, and applies an exact acceptance-and-correction procedure. We prove that SpecCTMC preserves the joint distribution of event times, generation order, and token values of the underlying continuous-time Markov chain. It therefore introduces neither time-discretization error nor additional factorization error from parallel decoding. Experiments on LLaDA-8B-Instruct and Dream-v0-Instruct-7B across five benchmarks demonstrate 1.73-4.38x fewer sequential model evaluations than direct continuous-time sampling, while preserving the reference sampling distribution. Our code is available at: this https URL
[NLP-7] Cultural Adaptation within Repeated Social Interactions
【速读】： 该论文旨在解决大语言模型（LLM）在跨文化情境中缺乏动态适应能力的问题，尤其针对人类文化知识具有隐性特征且主要通过情境化社会互动习得这一现实。传统评估方法仅关注模型“知道什么”，而忽视其在持续交互中学习和调整行为的能力。为此，研究提出将文化适应（cultural adaptation）重新定义为通过社会互动实现规范内化（norm assimilation）的连续过程，并设计了一个沙盒环境以模拟重复交互场景。其解决方案的关键在于引入记忆机制（memory mechanisms），使模型能够在多次交互中积累经验、实现纵向学习（longitudinal learning），并具备跨不同社交场景的迁移能力，以及在五种语言间实现多语言规范内化的潜力。该方法为构建持续、交互驱动的文化适应性大语言模型奠定了基础。
链接: https://arxiv.org/abs/2611.08334
作者: Chen Cecilia Liu,Roi Reichart,Anna Korhonen
机构: Language Technology Lab - University of Cambridge(剑桥大学语言技术实验室); Technion - Israel Institute of Technology(以色列理工学院)
类目: Computation and Language (cs.CL)
备注:
Abstract:Much of human culture is tacit and is typically learned through situated social interaction. As LLMs are deployed as interactive agents across diverse communities, assessing their cultural competence demands moving beyond existing question-answering approaches, shifting from what they know to how they adapt. In this work, we broaden the scope of cultural adaptation in NLP by investigating whether LLMs can adapt through repeated interactions via a sandbox environment. By reframing cultural adaptation as a continuous process of norm assimilation through social interaction, our work demonstrates that memory mechanisms foster longitudinal learning across repeated interactions, facilitate transfer to different social settings, and enable multilingual norm assimilation across five languages. Together, our findings lay the groundwork for continuous, interaction-driven cultural adaptation in LLMs.
[NLP-8] Polarity Classification with Multitask Learning Syntax Augmentation
【速读】： 该论文旨在解决情感分析（Sentiment Analysis, SA）中因否定、强化等语言现象导致的主观信息识别困难问题，同时克服传统语法解析方法带来的高计算开销。其核心挑战在于如何高效融合句法知识以提升模型对复杂语义结构的建模能力，而不依赖于推理阶段的外部解析器或人工设计的组合规则。解决方案的关键在于提出两种基于统一多任务学习（Multi-Task Learning, MTL）的句法知识集成策略：一是标准多任务学习（Standard-MTL），通过共享编码器并行学习序列标注形式的依存句法解析与情感极性分类；二是对比式多任务学习（Contrastive-MTL），采用对比辅助目标和语法偏置注意力池化机制，在不进行显式标签预测的前提下实现隐式句法引导。实验结果表明，Contrastive-MTL在11个跨语言数据集上优于基线模型，且在6个可比数据集上超越需推理时解析的规则系统（SELSP），同时推理速度提升1.5至9倍，验证了训练阶段隐式句法监督可有效替代推理阶段显式解析，显著降低计算开销而保持甚至提升性能。
链接: https://arxiv.org/abs/2611.08323
作者: Muhammad Imran,Carlos Gómez-Rodríguez
机构: Universidade da Coruña, CITIC (拉科鲁尼亚大学，信息与计算机科学系); Universidade da Coruña, CITIC (拉科鲁尼亚大学，信息与计算机科学系)
类目: Computation and Language (cs.CL)
备注:
Abstract:Sentiment Analysis (SA) is a fundamental NLP task concerned with the identification and categorization of subjective information in text. While lexicon-based methods offer transparency, they often struggle with linguistic phenomena such as negation and intensification. Syntactic parsing has been integrated into SA pipelines to address this, but conventional parsers introduce significant computational overhead. Recent work shows that reformulating dependency parsing as a Sequence Labeling (SEL) task can improve efficiency. We propose and compare two strategies for integrating syntactic knowledge into a neural sentiment analysis model under unified Multi-Task Learning (MTL). The first, Standard-MTL, jointly learns dependency parsing (via sequence labeling) and sentiment polarity classification using a shared encoder with two task-specific heads. The second, Contrastive-MTL, provides implicit syntactic guidance through a contrastive auxiliary objective and a syntax-biased attention pooling mechanism, without requiring explicit label prediction. Both rely on syntactic annotations during training and eliminate need for hand-crafted compositional rules or external parsing at inference. We evaluate eleven datasets spanning seven languages—the MARC corpus (English, Spanish, French, German) and seven datasets from SemEval-2022 Task 10 (English, Spanish, Norwegian, Basque, Catalan)—using four established SEL encodings. With the relative encoding, Contrastive-MTL outperforms a syntax-naive baseline on 10 of 11 datasets and surpasses the external rule-based pipeline (SELSP) that parses at inference on all six datasets where the latter applies, while being 1.5–9 \times faster at evaluation. Standard-MTL is more dataset- and encoding-dependent. These results show implicit syntactic supervision during training can match or exceed explicit parsing at inference without computational overhead.
[NLP-9] Benchmarking Large Language Models on Heterogeneous Cognitive and Biomedical Knowledge Graphs: The Role of Graph Grounding in Multi-Hop Reasoning and Hallucination Mitigation
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在生物医学与认知神经科学领域中，面对异构知识图谱（Heterogeneous Knowledge Graphs, KGs）时推理能力不可靠的问题。核心挑战在于如何有效融合通用语言理解能力与结构化领域知识，以提升复杂推理任务中的准确性与可解释性。其解决方案的关键在于引入多层级查询数据集，系统评估不同模型在具备知识图谱感知（KG-aware，即图谱引导）与无图谱感知（KG-agnostic）条件下的推理深度与事实准确性。研究发现，尽管领域专用模型在特定任务上表现有限，但通用模型（如GPT-5.1和Gemini 2.0 Pro）在多跳推理和高层级推理中表现更优，且知识图谱的引入显著降低了幻觉发生率。这一结果表明，将广泛的语言建模能力与结构化领域知识相结合是提升性能的关键路径。研究进一步提出基于任务复杂度与计算资源约束的模型-知识图谱集成策略，并探讨了混合神经符号架构在个性化医疗、认知障碍诊断及可解释人工智能中的应用前景。
链接: https://arxiv.org/abs/2611.08315
作者: Ali Sarabadani,Kheirolah Rahsepar Fard,Tahereh Jeirani
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Large language models (LLMs) are increasingly used in biomedical and cognitive neuroscience, but their ability to reason reliably over heterogeneous knowledge graphs (KGs) remains unclear. We benchmark five LLMs (GPT-5.1, Gemini 2.0 Pro, Llama 3, Qwen, and BioGPT) across five KGs: DrugBank, NeuroSynth, Cognitive Atlas, Brain Map, and GDPKG. We introduce a multitier query dataset to evaluate reasoning depth and factual accuracy under KG-aware (graph-grounded) and KG-agnostic conditions. Inter-annotator agreement was substantial, with Cohen’s kappa of 0.781 (range: 0.753-0.807) and Fleiss’ kappa of 0.77. Results show that general-purpose models, particularly GPT-5.1 and Gemini 2.0 Pro, generally outperform domain-specific models on multi-hop and high-level reasoning tasks, although performance declines as query complexity increases. KG grounding reduces hallucination rates across all evaluated models. Contrary to our initial hypothesis, domain-specific models do not consistently outperform general-purpose models on specialized tasks without graph grounding. These findings highlight the value of combining broad language-model capabilities with structured domain knowledge. We provide practical guidance for selecting LLM-KG integration strategies based on task complexity and computational constraints, and discuss hybrid neurosymbolic architectures for personalized medicine, cognitive disorder diagnosis, and explainable AI.
[NLP-10] LLM -Assisted Label Noise Correction for EEG-Based Sleep Stage Classification
【速读】： 该论文旨在解决睡眠分期自动分类中因人工标注标签质量低下所导致的深度学习模型性能受限问题，尤其关注标注者间一致性差及系统性标注错误对模型训练的影响。其解决方案的关键在于提出一种仅在训练阶段运行的大型语言模型（LLM）辅助标签噪声过滤流程：首先利用原始数据训练TinySleepNet模型并提取每 epoch 的 SoftMax 置信度分数，将低置信度段落连同其脑电图（EEG）谱特征与邻近序列上下文一并输入 LLaMA 3.1 模型；随后，该大语言模型依据美国睡眠医学学会（AASM）睡眠分期指南对每个时间段进行评估，并判定为“可靠”或“可疑”；最终剔除被标记为“可疑”的样本后，使用清洗后的数据重新训练模型。实验结果表明，该方法在 Sleep-EDF Expanded 数据集上通过 20 折交叉验证，将整体准确率从 81.8% 提升至 84.6%，宏平均 F1 分数从 75.6% 提升至 77.4%，尤其是在快速眼动（REM）睡眠检测方面表现显著改善。由于该流程仅在训练阶段引入大语言模型，推理时仍保持 TinySleepNet 轻量级架构（仅 1.3M 参数），因此具备部署于可穿戴设备实现实时睡眠监测的可行性。
链接: https://arxiv.org/abs/2611.08311
作者: Elmira Salari,Shruti Kshirsagar
机构: Wichita State University (威奇托州立大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Automatic sleep stage classification from electroencephalography (EEG) signals is a critical task in clinical sleep medicine. While deep learning models have shown promising results in this domain, their performance is fundamentally constrained by the quality of training labels, which human experts manually annotate and that are subject to inter-rater disagreement and systematic labeling errors. In this paper, we propose a novel LLM-assisted label-noise filtering pipeline that operates exclusively during training to identify and remove unreliable sleep-stage annotations before retraining the classifier. Specifically, TinySleepNet is first trained on the original dataset to extract per-epoch SoftMax confidence scores, then low-confidence epochs are submitted to LLaMA 3.1 along with their EEG spectral features and surrounding sequence context. The LLM evaluates each epoch against AASM sleep-scoring guidelines and assigns a verdict of RELIABLE or SUSPICIOUS. Suspicious epochs are removed from the training set, and TinySleepNet is retrained on the cleaned data. Evaluated on the Sleep-EDF Expanded dataset using 20-fold cross-validation, our method improves overall accuracy from 81.8% to 84.6% and Macro F1-score from 75.6% to 77.4%, with particularly notable gains in REM detection. Crucially, because the LLM operates only during training, the lightweight TinySleepNet architecture (1.3M parameters) is preserved at inference, making the proposed pipeline suitable for real-time, wearable sleep monitoring applications.
[NLP-11] Measured Coupling for Parallel Decoding in Diffusion Language Models
【速读】： 该论文旨在解决扩散语言模型（diffusion language models）在生成过程中每步需决定哪些词元（tokens）可安全组合的问题，传统方法如均场解码（mean-field decoding）通过惩罚预测相似的位置对来实现，但该策略仅作为近似代理，缺乏对实际依赖关系的精确刻画。其解决方案的关键在于用真实的耦合度量（coupling）替代原有的启发式惩罚机制：即通过在主前向传播中引入低成本探测器，直接测量一个词元的确定性选择如何影响模型在另一位置的置信度，从而捕捉真实上下文依赖。该方法无需额外训练或模型结构修改，可无缝嵌入原有均场规则。实验表明，在三个模型和四个基准上，该方法显著减少解码步数，多数情况下更高效，即使在启用键值缓存（key-value cache）时仍具优势，平均准确率损失仅为1.1个百分点；而随机打乱耦合或以均匀奖励替代则会丧失此性能提升，验证了耦合度量的有效性。
链接: https://arxiv.org/abs/2611.08307
作者: Tal Oved,Lior Wolf
机构: Tel Aviv University (特拉维夫大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Performance (cs.PF)
备注:
Abstract:Diffusion language models can write many tokens per step, but must choose which tokens are safe to commit together. Mean-field decoding penalizes pairs of positions whose predictions look alike. We replace this proxy with the coupling itself: how much committing one token changes the model’s confidence at another. Cheap probes inside the main forward pass measure it, with no training or extra model, and it plugs into the same mean-field rule. On three models and four benchmarks, our method needs far fewer steps than mean-field decoding and is faster in most settings, also with a key-value cache, at an average accuracy cost of 1.1 points. Shuffling the couplings or replacing them with a uniform bonus loses this gain.
[NLP-12] Multimodal-SEA: Benchmarking AI Agent Security under Multimodal Social Engineering Attacks
【速读】： 该论文旨在解决个人智能代理（Personal AI Agents）在面对由伪造多模态内容支持的社会工程攻击时，其内在安全性缺乏充分评估的问题。现有评估体系主要聚焦于文本对话场景，未能有效覆盖多模态环境下真实世界欺诈案例的复杂性。为此，研究提出Multimodal-SEA这一交互式基准，包含608个由专家设计的任务，基于真实诈骗案例构建，涵盖14个领域与152种攻击情景，并整合了预设的多模态内容。任务覆盖四个难度层级，反映中英文社会文化背景的差异。代理需在模拟应用环境中对内容进行检查、请求验证并执行操作，攻击结果基于可验证的工具执行日志生成。为实现对个人代理的可控红队测试，研究设计了一套多模态攻击框架（multimodal attack harness），集成任务信息管理、攻击进程控制及多模态内容构造功能，以评估代理在模拟环境中的抗适应性社会工程攻击能力。在六维维度上对七种主流多模态模型的评估表明，所有模型均存在显著脆弱性，攻击成功率介于7.1%至34.5%之间；引入多模态内容后，攻击成功率从19.9%上升至29.2%，即便代理主动验证覆盖率提升。强化安全指令虽能降低攻击成功率，但未提升对欺骗性多模态内容的识别能力，部分模型虽表现较低攻击成功率，却伴随过度拒绝合法请求的现象。因此，解决方案的关键在于构建一个真实、可控且多模态驱动的评估体系，揭示当前多模态代理在对抗社会工程攻击时的系统性缺陷。
链接: https://arxiv.org/abs/2611.08281
作者: Zhitao He,Chi Him Zhang,Chengru Han,Yi R Fung
机构: The Hong Kong University of Science and Technology (香港科技大学); MMSense Lab (多媒体感知实验室)
类目: Computation and Language (cs.CL)
备注:
Abstract:Personal AI agents face social engineering attacks supported by forged multimodal content, but existing evaluations mainly focus on text-based dialogue, leaving their intrinsic security under such attacks insufficiently studied. We introduce Multimodal-SEA, an interactive benchmark comprising 608 expert-designed tasks grounded in real-world fraud cases, covering 14 domains and 152 attack scenarios with prepared multimodal content. The tasks span four difficulty levels and reflect the distinct social and cultural contexts of Chinese and English environments. Agents inspect content, verify requests, and execute actions in simulated applications, with attack outcomes grounded in verifiable tool execution logs. To enable controlled red teaming of personal agents, we design a multimodal attack harness that integrates task information, attack progression control, and multimodal content construction to evaluate their intrinsic security against adaptive social engineering attacks in simulated environments. Evaluating seven widely used multimodal models across six dimensions reveals that all are vulnerable, with attack success rates of 7.1% to 34.5%. Adding multimodal content increases attack success from 19.9% to 29.2%, despite increased proactive verification coverage. Stronger safety instructions reduce attack success without improving recognition of deceptive multimodal content, while lower attack success rates for some models accompany excessive refusal of legitimate requests.
[NLP-13] he Geometry of Simulated Perspectives: Dissociating Representation and Behaviour in a Multimodal Language Model
【速读】： 该论文旨在解决生成式视觉语言模型在不同角色（persona）设定下，其内部表征结构如何被重构，以及这种重构是否直接影响模型的判断行为。核心问题是：角色设定与提示语框架如何分别影响模型的内在表征几何（representational geometry）与外在行为输出之间的关系。解决方案的关键在于通过系统性实验，操控100张图像在53种不同的角色内容与语言框架下的提示条件，同时记录行为评分与各层隐藏状态，从而揭示表征动态演化规律。研究发现，早期层主要编码图像身份信息，而后期层则逐渐涌现出与角色相关的表征结构；更重要的是，提示语框架对内部表征的影响显著强于角色身份本身，但角色内容却更显著地驱动行为判断，形成“表征-行为”解耦现象。这一发现表明，内部表征距离并不必然反映行为相关性，揭示了人工认知系统中模拟视角以独特结构形式嵌入表征空间的本质机制，凸显了表征几何与其功能后果之间的根本差异。
链接: https://arxiv.org/abs/2611.08262
作者: Mathis Immertreu,Fitim Abdullahu,Thomas Kinfe,Claus Metzner,Helmut Grabner,Patrick Krauss,Achim Schilling
机构: University Erlangen-Nürnberg (埃尔兰根-纽伦堡大学); University Hospital Mannheim, Heidelberg University (海德堡大学曼海姆医院); BGU Ludwigshafen (路德维希港BGU); IDS Institut für Data Science, ZHAW School of Engineering (苏黎世应用科技大学工程学院数据科学研究所)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 30 pages, 10 figures, 2 tables. Code: this https URL data: this https URL
Abstract:Multimodal language models can simulate different observers, yet it remains unclear how assigned perspectives reshape their internal representations and whether these changes govern their judgements. Here, we investigate the representational geometry of persona-conditioned visual interestingness in Qwen3-VL-8B-Instruct. Across 100 images and 53 prompting conditions, we systematically vary persona content and the linguistic framing used to introduce it, while recording behavioural ratings and layer-wise hidden states. We observe a systematic reorganisation of representational geometry across processing depth: early layers primarily encode image identity, whereas persona-specific structures emerge in later layers. Strikingly, prompt framing reshapes internal representations more strongly than persona identity, whereas persona content has a greater impact on behavioural judgements. This dissociation is further illustrated by minimal identity statements that alter representations even without changing responses. These findings demonstrate that representational distance does not necessarily reflect behavioural relevance. More broadly, they reveal how simulated perspectives emerge as distinct structures within the representational space of artificial cognitive systems, highlighting a fundamental distinction between the geometry of internal representations and their functional consequences.
[NLP-14] rminal-style Mathematical Modeling Bench: The First Agent ic and Verifiable Benchmark for Mathematical Modeling
【速读】： 该论文旨在解决现有数学建模基准测试在评估生成式模型求解能力时存在的核心缺陷：即缺乏对完整建模过程的可验证性与真实性检验。传统基准多仅评分求解片段或依赖大语言模型（LLM）进行报告打分，导致评价结果不可靠且无法复现。为此，本文提出TB-MathModel（终端式数学建模基准），作为首个将竞赛风格建模问题实现全链路智能体化、可验证性与终端交互一体化的基准平台，覆盖2023至2025年共18道中国大学生数学建模竞赛（CUMCM）与美国大学生数学建模竞赛（MCM）的真实题目。其关键创新在于两个设计：一是采用方向感知的一侧评分机制，不惩罚超越基准解的表现，避免抑制创造性；二是双层可行性验证机制——第一层为程序化检查，涵盖六种可判定的不变量形式；第二层为硬性约束，通过624个Lean 4形式化证明义务确保解的物理合理性，无效解直接被置零。此外，每项任务基于一组人类优秀论文构成参考集进行多参考评估，要求模型输出接近多个权威解，而非单一基准。实验表明，在七组多模态模型（每组三个种子）中，最强模型达到64.58%的多参考O-Eval得分，而仅需1/100成本的开源模型亦达59.81%。TB-MathModel将建模问题的数值核心转化为可复现、机器可验证的信号，不仅可用于衡量长周期建模能力，还可用于收集智能体决策轨迹，推动数学建模领域向更严谨、可审计的方向发展。
链接: https://arxiv.org/abs/2611.08237
作者: Xiaojun Wu,Cehao Yang,Xuhui Jiang,Chengjin Xu,Jia Li,Jian Guo
机构: The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)); IDEA Research(IDEA 研究院); DataArcTech Ltd.(数据弧科技有限公司)
类目: Computation and Language (cs.CL)
备注: 20 pages, 4 figures, 14 tables
Abstract:Mathematical modeling turns an open-ended real-world problem into assumptions, a model, a solution, and a defended recommendation, and its answers matter only if they are both close to a credible reference and physically feasible; existing benchmarks either score solver fragments or grade reports with an LLM judge, leaving those numbers unchecked. We introduce TB-MathModel (the Terminal-style Mathematical Modeling Bench), the first benchmark to make complete contest-style modeling problems agentic, verifiable, and terminal at once, over 18 real CUMCM and MCM problems (2023–2025). Each task ships the verbatim statement, official attachments, and figures in a containerized terminal; the agent returns one JSON artifact, scored on the numeric core of the final question against a reproduction of the Outstanding-award paper. Two design choices distinguish it: direction-aware, one-sided scoring, which never penalises beating the anchor, and two-layer feasibility verification—a programmatic gate over six decidable invariant forms plus 624 Lean~4 obligations—with a hard gate that zeroes infeasible answers. Each task is scored against a panel of human outstanding papers rather than one, so a model must be close to many accepted answers; across seven multimodal entries (three seeds each) the strongest reach a multi-reference O-Eval of 64.58%, while the \ 2.54 open model holds 59.81% at 1/100 the cost. TB-MathModel turns the numeric core of modeling into a reproducible, machine-checkable signal for measuring long-horizon modeling competence and for collecting agent trajectories. The source code is available at this https URL.
[NLP-15] From Twisted Facts to Rogue Agents : Can LLM s Reliably Detect Multi-Dimensional Hallucinations?
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在信息检索、内容生成及基于代理的任务执行中，其输出易受细微且高风险幻觉（hallucination）影响的问题。现有方法对幻觉的检测与纠正能力有限，尤其在细粒度定位与修正方面表现不佳，且在代理轨迹中的忠实性（faithfulness）幻觉检测上缺乏系统评估手段。为此，论文提出HaluaCheck——一个大规模、双语基准测试集，涵盖超过4.2万条样本，覆盖文本中的事实性幻觉与代理轨迹中的忠实性幻觉，并提供细粒度、多层次的标注（包括检测、定位与修正），适用于不同任务类型与扰动程度。实验结果表明，尽管主流LLMs具备基础的样本级幻觉识别能力，但在细粒度定位、修正以及代理场景下的幻觉检测方面仍存在显著不足。进一步分析显示，思维链推理（thinking）、网络搜索与监督微调虽能带来部分改善，但均未能根本性解决该问题。因此，HaluaCheck为评估和诊断大语言模型在文本与代理场景下的可靠性提供了坚实基准。
链接: https://arxiv.org/abs/2611.08231
作者: Hanyu Qiu,Ze Liu,Zhe Zhao,Kuang Xiong,Blaze Chen,Zhenxiang Yan,Renfen Hu
机构: Beijing Normal University(北京师范大学); Tencent AI Lab(腾讯人工智能实验室)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models (LLMs) are widely used for information access, content production, and agent-based task execution, but their outputs remain vulnerable to subtle, high-risk hallucinations. To systematically evaluate and advance post-generation hallucination detection, we introduce HaluCheck, a large-scale, bilingual benchmark comprising over 42k instances spanning both factual hallucinations in text and faithfulness hallucinations in agent trajectories. HaluCheck features fine-grained, multi-level annotations for detection, localization, and correction, covering different tasks and perturbation degrees. Experiments on fourteen LLMs reveal that, while models show basic sample-level detection ability, they struggle with fine-grained localization, correction, and agent hallucination detection. Further analysis demonstrates that thinking, web search, and supervised fine-tuning each offer partial improvements, but none fully solve the problem. HaluCheck thus provides a solid basis for benchmarking and diagnosing LLM reliability across text and agent scenarios.
[NLP-16] Agent -R2: Scaling Agents for Realistic Productivity Work via Real-Source Data Synthesis
【速读】： 该论文旨在解决当前智能体在完成真实生产力任务时面临的瓶颈问题，即难以有效执行长周期专业工作流、理解异构文件上下文，并在细粒度要求下生成可用交付成果。其核心挑战在于缺乏能够真实反映工作意图、具备真实文件环境和可衡量交付质量的可扩展后训练数据。为此，论文提出Agent-R²——一个基于真实来源的数据合成框架，通过引入真实职业场景种子（occupational seeds）、从语料库中提取的工作流证据以及网络获取的真实文件，实现了全流程的现实性锚定。该框架从10,000个候选样本中提炼出4,000个任务实例，每个实例均包含多格式真实附件与细粒度交付标准。基于此数据，采用监督微调与基于评分规则的强化学习训练得到Agent-R²-4B模型，在仅使用4,000个实例的情况下，显著优于Qwen3-235B-A22B-Thinking-2507，在GDPval（51.0% vs. 48.0%）和OfficeQA（14.8% vs. 13.8%）上取得更优表现。进一步分析表明，Agent-R²能有效缓解智能体的“捷径行为”（shortcut behavior），即在未充分接地、构建或验证前生成表面合理但实质错误的交付物，同时促进持续探索，提升推理深度与任务完成可靠性。
链接: https://arxiv.org/abs/2611.08219
作者: Teng Pan,Zixuan Wang,Yuchen Yan,Qifei Wu,Yao Chen,Yingqi Qu,Ruiqing Zhang,Yongliang Shen
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language models have made rapid progress in agentic reasoning and tool use, raising expectations for agents that can take over realistic productivity work from humans. However, current agents still struggle to finish such tasks, where success requires following long-horizon professional workflows, understanding heterogeneous file contexts, and producing usable deliverables under fine-grained requirements. A key bottleneck is the lack of scalable post-training data that reflects real work intent, authentic file environments, and deliverable-level evaluation. To close this gap, we introduce Agent-R ^2 , a \textbfr eal-source data synthesis framework that scales training data for \textbfr ealistic productivity work. Instead of relying on manually specified workflows or LLM-generated files, Agent-R ^2 grounds synthesis throughout the pipeline in real occupational seeds, corpus-derived workflow evidence, and authentic web-sourced files. From 10k web-corpus candidates, Agent-R ^2 distills 4k task instances that pair real multi-format attachments with fine-grained, deliverable-level rubrics. On this data, we train Agent-R ^2 -4B via supervised fine-tuning and rubric-based reinforcement learning, improving both end-to-end productivity performance and file-grounded reasoning. Trained on only 4k instances, Agent-R ^2 -4B surpasses Qwen3-235B-A22B-Thinking-2507 on both GDPval (51.0 % vs. 48.0 % ) and OfficeQA (14.8 % vs. 13.8 % ). Further analysis shows that Agent-R ^2 mitigates shortcut behavior, where agents produce superficially plausible deliverables before sufficiently grounding, constructing, or verifying them, while promoting sustained exploration.
[NLP-17] ProQuant: Progressive Quantization-aware Training for Edge MLLM s
【速读】： 该论文旨在解决多模态大语言模型（MLLMs）在资源受限的边缘设备上部署时面临的计算与内存开销过大的问题，尤其针对低比特量化（如权重量化至4比特，W4A8）下性能显著下降的挑战。现有后训练量化（PTQ）方法在中等比特位宽（如W8A8）下表现良好，但在低比特场景下难以维持模型性能，尤其是对适合边缘部署的小规模模型。为此，本文提出一种面向低比特多模态大模型的量化感知训练（QAT）框架——ProQuant。其关键创新在于：首先，设计自重标记（SRL）策略，通过重采样开源多模态数据集的响应，使微调数据分布更贴近预训练数据分布，缓解分布偏移问题；其次，首次从解耦视角分离多模态理解与语言建模能力，提出渐进式块激活（PBA）机制，优先恢复多模态理解能力，并引入视觉敏感因子τ以精准识别对多模态感知至关重要的网络层。实验结果表明，ProQuant在多个基准测试中显著优于现有方法，例如在W4A8量化下，Qwen3-VL-2B-Instruct模型平均准确率提升6.6%，且仅损失约1%的全精度性能，展现出卓越的低比特压缩效率与实用性。
链接: https://arxiv.org/abs/2611.08216
作者: Yufei Xue,Yushi Huang,Jiawei Shao,Pingcheng Dong,Yonghao Tan,Shiyao Li,Kwang-Ting Cheng,Jun Zhang
机构: The Hong Kong University of Science and Technology(香港科技大学); Northwestern Polytechnical University(西北工业大学); Institute of Artificial Intelligence (TeleAI), China Telecom(中国电信人工智能研究院); Tsinghua University(清华大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Multi-modal large language models (MLLMs) with impressive perception and understanding capabilities will enable a range of downstream applications. However, their substantial computational and memory requirements hinder the real-world implementation, particularly for resource-constrained edge devices. Quantization has emerged as an effective solution for deploying large models. While post-training quantization (PTQ) successfully retains the performance at moderate bit-width (e.g., W8A8), it suffers from severe performance degradation under low-bit quantization (e.g., W4A8), especially for small-scale models that are more suitable for edge deployment. In this paper, we propose ProQuant, a novel weight-activation quantization-aware training (QAT) framework tailored for low-bit MLLMs. First, to align the limited finetuning dataset distribution with the pretrained dataset distribution, we propose a self-relabeling (SRL) scheme to resample the responses of the open-source multimodal dataset. Second, we are the first to take a decoupled view of multi-modal (MM) understanding and language modeling abilities. We propose a progressive block activation (PBA) mechanism to prioritize the MM understanding recovery. We also introduce a vision-sensitive factor \vec\tau that effectively identifies the MM-critical layers. Extensive experiments across various benchmarks, covering models with 2B-8B parameters, show that ProQuant significantly outperforms existing methods. For example, our W4A8 Qwen3-VL-2B-Instruct improves average accuracy by impressively 6.6% and achieves performance comparable to full precision counterparts with only approximately 1% performance drop. Code is available at this https URL.
[NLP-18] SMART: State-based Multi Agent Reasoning Traces for Verification-oriented Problem-Solving
【速读】： 该论文旨在解决在上下文信息有限的复杂问题（如科学主张）中，因缺乏必要背景知识而导致推理能力受限的问题。其核心挑战在于如何有效获取并利用与问题解决方案相关且可信的知识以支持推理过程。该研究提出的解决方案——SMART多智能体框架，其关键在于通过捕捉问题情境中参与实体的隐式状态，利用多种大语言模型（LLM）结合结构化的角色驱动提示策略，系统性地构建多样化、互补性强且与解题高度相关的经验证实前提（premises）。该方法通过将前提知识锚定于真实世界可验证的事实，显著提升了推理性能，使在跨三个数据集的复杂科学主张验证任务中准确率最高提升20%，其中基于真实世界知识的锚定机制本身即可带来高达5%的性能增益，证明了知识真实性对推理有效性的重要作用。
链接: https://arxiv.org/abs/2611.08196
作者: Sai Vallurupalli,Francis Ferraro
机构: University of Maryland Baltimore County (马里兰大学巴尔的摩县分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Complex problems with limited context, such as claims, require a means to obtain additional knowledge necessary for solution reasoning. We propose a multi-agent framework, SMART, to obtain knowledge through capturing the implicit state of participating entities in a problem context. SMART curates the state knowledge as diverse, complementary and solution-salient premises using a variety of LLMs and a structured role-based prompt strategy. Using knowledge from these real-world verified curated premises, SMART’s knowledge guided solution reasoning improves validation performance up to 20% on complex scientific claims spanning 3 datasets. We show that grounding premises in real-world knowledge is necessary and improves performance up to 5%.
[NLP-19] Knowing When to Teach: Learning from Teachable Prefixes in Multilingual Reasoning Distillation
【速读】： 该论文旨在解决多语言推理场景下基于策略自蒸馏（On-Policy Self-Distillation, OPSD）方法中存在的“教师漂移”（Teacher Drift）问题。当学生模型在低资源语言中产生中间步骤错误时，若教师模型仍以相同方式生成反馈，会逐渐跟随学生错误路径，导致错误累积并降低推理性能。其解决方案的关键在于提出可教的基于策略自蒸馏（Teachable On-Policy Self-Distillation, TOPSD），通过引入教师-学生似然比来识别教师开始偏离正确路径的节点（即漂移起点），并据此选择一个用于监督的前缀；随后验证该前缀是否能引导教师恢复正确答案，仅对通过验证的前缀实施蒸馏。该机制有效缓解了跨语言中因资源差异引发的教师漂移问题，在十种语言、三种模型规模的多语言数学推理任务上均显著优于传统OPSD，尤其在低资源和中等资源语言中表现更为突出。
链接: https://arxiv.org/abs/2611.08179
作者: Jaewon Cho,Minseo Yoon,Jinyoung Park,Jinwoo Seo,Jaewon Chu,Alice Oh,Hyunwoo J. Kim
机构: KAIST(韩国科学技术院); Korea University(韩国大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:On-Policy Self-Distillation (OPSD) improves reasoning by training a student on its own reasoning, using the same model conditioned on reference solutions as a teacher. Applying OPSD to multilingual reasoning requires the teacher to provide useful feedback across languages, even when the student makes mistakes in intermediate steps. We observe a phenomenon we call Teacher Drift: as these mistakes accumulate, the teacher may follow the student’s incorrect steps instead of correcting them. Our analysis shows that this shift tends to occur earlier in lower-resource languages. To account for these differences across languages, we propose Teachable On-Policy Self-Distillation (TOPSD), which determines where and whether teacher supervision should be applied to each reasoning trajectory. TOPSD uses teacher-student likelihood to estimate where drift begins and select a prefix for supervision. It then verifies whether the teacher can recover the correct answer from that prefix, applying distillation only to prefixes that pass verification. Experiments on multilingual mathematical reasoning across ten languages and three model scales show that TOPSD consistently outperforms conventional OPSD, particularly in low- and medium-resource languages.
[NLP-20] RouteAlign: Aligning Task Requirements with Model Profiles for LLM Service Selection
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）服务选择中因性能评估缺乏任务适配性而导致的决策偏差问题。现有方法依赖聚合评分或语义相关示例，但这些方式无法准确反映任务对能力需求与难度的具体要求，导致性能评价失真。其解决方案的关键在于提出RouteAlign框架，通过显式建模任务需求中的能力要求与难度维度，实现对模型性能测量与服务选择的双重结构化对齐：历史任务需求用于构建全局性能画像与局部能力估计，而当前任务需求则指导如何利用已有测量结果进行实时决策；系统采用共享评分器整合多源证据，并引入生成成本偏好以在符合条件的服务中优化选择。实验表明，RouteAlign在多样化任务与异构LLM服务场景下显著提升了质量-成本权衡边界，相较开源的Jev-style决策后端在保证更高任务质量的同时降低了生成成本。消融实验验证了局部性能证据的有效性，而目录更新实验则证明了评分器在新模型测量数据下的可复用性。
链接: https://arxiv.org/abs/2611.08162
作者: Jihao Zhao,Ding Chen,Yanfang Chen,Bo Tang,Feiyu Xiong,Junchi Yan,Zhiyu Li
机构: 未知
类目: oftware Engineering (cs.SE); Computation and Language (cs.CL)
备注:
Abstract:LLM service selection requires performance evidence that reflects the demands of the current task. Aggregate scores obscure differences in capability and difficulty, while semantically related examples can involve different combinations of skills. We present RouteAlign, which uses task requirements to structure both the measurement of model performance and the selection of services. Task requirements describe capability demand and difficulty, enabling alignment with corresponding model performance measurements. Historical requirements determine how outcomes contribute to global profiles and local capability estimates; incoming requirements determine how those measurements inform the current decision. A shared scorer integrates the evidence, and a generation-cost preference controls selection among eligible services. Evaluation across diverse tasks and heterogeneous LLM services shows that RouteAlign extends the empirical quality-cost frontier toward higher task quality. Comparisons with open Jev-style decision backends show consistent quality gains, with lower generation cost than their strongest tested configuration. Ablations support the contribution of local performance evidence, while catalog-update experiments demonstrate effective reuse of the fitted scorer with refreshed model measurements.
[NLP-21] Beyond K/V Caches: Cross-Model State Handoff in Hybrid Linear-Attention LLM s
【速读】： 该论文旨在解决大模型在长对话场景中进行多模型级联部署时，因模型切换导致的重复预填充（prefill）开销问题。现有方法在跨模型传递全注意力（Full-Attention, FA）键值（K/V）缓存、递归状态（Recurrent State, REC）及卷积历史（Convolution History, Conv）时，由于Conv与REC之间的复杂交互，导致状态迁移不准确，进而影响生成质量。尤其当卷积扰动改变递归层中的键值并影响REC更新时，即使卷积历史被覆盖，其对后续输出的影响仍难以消除，且高精度重构REC未必提升续写表现。针对此问题，本文提出混合状态传输（Hybrid State Transfer, HST）机制，通过结构化REC映射与卷积历史对齐来保留状态顺序，并利用目标模型冻结输出进行适配器校准，同时离线合并FA缓存映射。实验表明，在五个基准测试中，Qwen3.5 4B→27B与9B→27B的平均任务保留率分别达到89.88%和94.11%，且在合成16K输入下，模型交接主体部分速度提升分别达21.66×与21.91×，显著提升了跨模型推理效率与保真度。
链接: https://arxiv.org/abs/2611.08133
作者: Jiacheng Cai,Tao Zhang,Siqi Liang,Yun Dai,Ziqian Zeng
机构: South China University of Technology (华南理工大学)
类目: Computation and Language (cs.CL)
备注: 12 pages, 4 figures, 9 tables
Abstract:Production deployments often use different models to balance cost and quality through cascading, routing, and switching models during long conversations. Context accumulates across turns. Each switch requires the target to repeat prefill over this context. Existing methods transfer full-attention (FA) K/V caches across models. Hybrid linear-attention models also retain recurrent states (REC) and convolution histories (Conv). The interaction between Conv and REC complicates cross-model state transfer. Conv perturbations can change keys and values in recurrent layers and alter REC updates. Conv histories are overwritten, but changes in REC can still affect later outputs. More accurate REC reconstruction need not improve continuation. Building on this, we propose Hybrid State Transfer (HST) for all three state types. It uses structured REC maps and preserves history order during Conv alignment. We calibrate the adapters using the frozen target’s outputs during continuation. FA cache map updates are merged offline. Across five benchmarks, Qwen3.5 4B \rightarrow 27B and 9B \rightarrow 27B achieve mean task retentions of 89.88% and 94.11%, respectively. With source caches already available, separate timing on synthetic 16K inputs yields handoff body speedups of 21.66\times and 21.91\times .
[NLP-22] Gabor Gabor: E-Gaborformer a Strictly Causal Speech Recognizer with Gabor-Dictionary Convolutions and Delta-Rule Memory
【速读】： 该论文旨在解决实时语音代理（real-time voice agents）中语音识别器在流式处理时面临的两大核心挑战：一是必须严格因果（strictly causal），即不能依赖未来音频信息；二是单次通话的内存占用不应随通话时长线性增长。现有模型如因果注意力编码器在长时间运行后内存消耗急剧上升（例如10分钟内超过1 GB），难以满足生产环境对资源效率的要求。为此，论文提出E-Gaborformer，其关键创新在于将门控增量网络（Gated DeltaNet-2）与滑动窗口注意力（sliding-window attention）以3:1的比例结合，并通过固定字典中的因果戈贝尔原子（causal Gabor atoms）构建所有时序卷积，从而实现严格的因果性与恒定内存状态（约每通话语音19 MB）。实验表明，E-Gaborformer在LibriSpeech上的因果对照研究中能更早脱离初始准确率平台期，达到目标精度速度提升最高达18%，且性能优于同规模的E-Branchformer。在真实生产环境中，基于E-Gaborformer的0.6B参数双语流式转录器在英文语音代理流量下，相比NVIDIA Nemotron 3.5 ASR Streaming（同规模模型）降低28%–31%的词错误率（WER），希伯来语错误率更低20%。此外，同一组权重可同时支持严格因果与480毫秒前瞻解码而无精度损失，且采用两分类形式的RNN-T格网训练可使训练速度提升2.6倍，同时保持精确损失与梯度。
链接: https://arxiv.org/abs/2611.08092
作者: Nir Krakowski,Tamir Cohen,Ohad Zohar,Chen Zarfati
机构: Deepdub
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Sound (cs.SD)
备注: 19 pages, 6 figures, 9 tables
Abstract:Real-time voice agents need a speech recognizer that never waits for future audio and whose memory per call does not grow with the length of the call. We present E-Gaborformer, a strictly causal encoder that combines Gated DeltaNet-2 recurrent layers with sliding-window attention in a 3:1 ratio, and builds every temporal convolution from a fixed dictionary of causal Gabor atoms. Its streaming state is constant, about 19 MB per call at production scale including the prediction network, where a causal-attention encoder of the same size accumulates over 1 GB after ten minutes. In a controlled causal study on LibriSpeech, E-Gaborformer leaves the initial plateau earlier, reaches a given accuracy up to 18% sooner, and matches or beats an E-Branchformer trained with the identical front end and recipe. At production scale, a 0.6B-parameter Hebrew and English streaming transducer built on E-Gaborformer has 28 to 31% lower word error rate (WER) than NVIDIA’s Nemotron 3.5 ASR Streaming, a model of the same size, on English voice-agent traffic, and up to 20% lower Hebrew WER than our previous production system. The same weights serve strictly causal and 480 ms-lookahead decoding with no loss in accuracy, and a two-class formulation of the RNN-T lattice trains the transducer 2.6x faster with exact loss and gradients.
[NLP-23] ATHAAR: A Medieval Scientific Arabic Manuscript Dataset for Handwritten Text Recognition
【速读】： 该论文旨在解决当前视觉-语言模型（VLMs）在识别中世纪阿拉伯科学手稿方面面临的挑战，主要受限于草书连笔、变体元音符号标注以及训练数据匮乏等问题。其解决方案的关键在于提出首个基于文本批评学（textual criticism）方法的专用于科学阿拉伯手稿转录的标注指南，并据此构建了名为ATHAAR的新数据集。该数据集通过涵盖同一文本的不同抄本版本、多样的书写风格、丰富的内容类型以及数字字符，实现了对古代科学文本多样性的全面覆盖。研究进一步利用先进的VLMs设计跨脚本、跨章节和跨著作的迁移实验，揭示了现有系统在处理古籍科学文本特异性时的显著局限性。
链接: https://arxiv.org/abs/2611.08091
作者: Ismail El Jamiy,Khadija Msalek,Guillaume Loizelet,Farah Benamara,Nabil Hathout
机构: IRIT(图卢兹信息科学研究所), Université de Toulouse(图卢兹大学); IMT(图卢兹信息与技术学院), Université de Toulouse(图卢兹大学); CLLE(语言、文化与教育实验室), Université de Toulouse Jean-Jaurès(图卢兹让·饶勒斯大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注:
Abstract:Despite recent progress in handwritten text recognition, medieval Arabic scientific manuscripts remain an open challenge for current vision-language models (VLMs) due to cursive ligatures, variable diacritization, and more importantly the lack of training data. This paper aims to address this gap, by proposing the first annotation guidelines dedicated to the transcription of scientific Arabic manuscripts inspired by textual criticism, a well-established approach to historical text editing. We then propose, ATHAAR, a new dataset manually annotated following this guideline. Unlike previous resources in the field, ATHAAR varies both witnesses of the same text, calligraphic styles, and content to deal with texts but also numerals. Relying on state-of-the-art VLMs, we finally design a set of experiments to study models transfer across scripts, chapters, and treatises. Our findings highlight the limitations of current systems when faced with the specificities of ancient scientific texts.
[NLP-24] Linguistic boundaries emerge from pressure-constrained processing dynamics
【速读】： 该论文旨在解决语言结构边界（如句子边界）如何在连续认知加工过程中自然形成的问题。其核心挑战在于理解边界并非由预设规则或显式信号决定，而是源于多重动态认知需求的交互作用。解决方案的关键在于提出：边界涌现是由于三个相互竞争的认知需求共同驱动的结果——维持未解决的依赖关系（unresolved dependencies）、整合已完成的关系、以及更新语义上下文。这些需求在处理过程中产生非线性压力变化，在接近边界时表现为未解决依赖减少、整合成本升高、累积压力重置，从而促成结构完成与边界感知。研究通过生成式模拟、人类听觉理解实验和自定速阅读实验验证了这一机制，发现边界标记与较低的未解决依赖、更高的整合成本与语义不匹配、以及前序记忆负荷上升显著相关，且这些因素在解释阅读时间与神经表示动态变化方面独立于句法、韵律线索及下一词意外度（next-token surprisal）。结果表明，语言边界是连续认知动态中结构性转换的离散表现形式，而非被动的句法产物。
链接: https://arxiv.org/abs/2611.08085
作者: Rui He,Philipp Homan,Wolfram Hinzen
机构: Grammar and Cognition Lab, Department of Translation Language Sciences, Universitat Pompeu Fabra, Barcelona 08018, Spain; Department of Adult Psychiatry and Psychotherapy, University Hospital of Psychiatry Zurich, University of Zurich, Zurich 8032, Switzerland; Neuroscience Center Zurich, University of Zurich and ETH Zurich, Zurich 8057, Switzerland; Institució Catalana de Recerca i Estudis Avançats, Barcelona 08010, Spain
类目: Computation and Language (cs.CL)
备注:
Abstract:Human cognition organizes continuous experience into bounded units while maintaining continuity across them. Linguistic elements are also organized into bounded structural units such as words and phrases, with sentence marking the final bound of a grammatical construction process. How do such boundaries emerge? We propose boundaries emerge from interacting demands to maintain unresolved dependencies, integrate completed relations, and update semantic context, with structural completion allowing but not determining closure. First, in a generative simulation without predefined boundary positions or constituent labels, these demands produced sentence-like segment-length distributions and distinct characteristic timescales, with structural completion allowing but not determining closure. Near boundaries, unresolved dependency load declined, integration cost peaked, and accumulated pressure reset. Then, we tested whether same demands predicted human listeners’ boundary choices during comprehension. Listeners were more likely to mark boundaries with fewer unresolved dependencies, higher integration cost and semantic mismatch, and elevated preceding memory load, beyond syntactic and prosodic cues. Third, in naturalistic self-paced reading, sentence-final slowing coincided with declining memory load and increasing integration cost and semantic mismatch, broadly consistent with the simulation. Processing same narratives with a large language model revealed increased local intrinsic dimensionality and straighter hidden-state trajectories at sentence endings, followed by larger representational updates. Boundary explained additional variation in reading time and hidden-state geometry beyond processing demands, but considerably less in next-token surprisal. Together, these findings suggest linguistic boundaries emerge as discrete manifestations of transitions in continuous processing dynamics.
[NLP-25] Learning Dynamics of On-Policy Distillation
【速读】： 该论文旨在解决生成式语言模型在在线策略蒸馏（On-policy Distillation, OPD）过程中学习动态不明确的问题，特别是反馈回路如何影响模型的泛化能力与性能演化。其核心挑战在于：OPD通过学生模型自生成轨迹进行训练，形成教师监督与学生更新之间的动态反馈循环，但这一过程中的信号转化机制与分布漂移效应尚未被充分理解。论文的关键解决方案是提出一个机制性分析框架，将反馈回路分解为两个阶段——一是蒸馏信号转化为参数更新的转换过程，二是参数更新通过前缀广播改变预测分布并重塑采样轨迹分布的传播过程。该框架揭示了学生模型对教师信号响应不足的根源，以及当局部教师一致性提升反而导致失败延续概率上升时所引发的性能退化现象。此外，研究发现离线蒸馏可有效预训练学生模型以增强其对后续OPD的适应性，尽管短期内对保留准确率改善有限；而经过充分离线准备后，结合OPD可实现显著的下游性能提升。因此，该研究系统地关联了教师监督、学生参数空间几何结构及动态生成分布三者之间的相互作用，阐明了决定OPD学习动态及其最终效果的核心机制。
链接: https://arxiv.org/abs/2611.08016
作者: Guan Zhe Hong,Sumeet Ramesh Motwani,Liang Chen,Yangchen Pan,Junchi Yu,Jialin Yu,Philip Torr,Yi Ren
机构: University of Oxford (牛津大学); EPFL (洛桑联邦理工学院); City University of Hong Kong (香港城市大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:On-policy distillation (OPD) is emerging as an important paradigm for language-model post-training, combining dense teacher supervision with training on student-generated trajectories. This creates a feedback loop where each update changes the student and thereby the training distribution that generates subsequent updates. Yet, how this feedback shapes learning and generalization remains poorly understood. We conduct a mechanistic analysis of OPD’s learning dynamics, distinguishing the two feedback loop stages: conversion of the distillation signal into parameter updates, and broadcast through which these updates change predictions across prefixes and reshape the rollout distribution. This framework helps us understand both limited student responsiveness despite substantial teacher-student disagreement, and degeneration when updates that improve local teacher agreement also increase the probability of unsuccessful continuations, exposing the student to unreliable supervision in subsequent rollouts. Finally, we show that offline distillation can prepare a student for OPD without significantly improving immediate held-out accuracy. While OPD alone may degrade generation, it can yield substantial accuracy gains after sufficient offline preparation. Together, our results connect teacher supervision, student geometry, and the evolving rollout distribution in determining the learning dynamics of OPD and its effects on downstream performance.
[NLP-26] Compression Through a Context Lens: Evaluating Adaptation Quantization and Distillation in Arabic Legal LLM s
【速读】： 该论文旨在解决大语言模型在资源受限的法律场景中部署时，模型压缩技术对领域特定性能影响的评估问题。其核心挑战在于不同评估协议（有上下文与无上下文）下模型压缩效果的差异性，尤其在阿拉伯语法律领域中的表现。解决方案的关键在于通过构建并使用ArabLegalEval这一沙特阿拉伯语法律基准数据集，在有上下文和无上下文两种评估范式下系统评估三种主流模型（FANAR-9B、JAIS-8B、Qwen2.5-7B）在三个阶段——领域适配、量化和知识蒸馏——中的表现。研究发现，移除上下文支持后，基础模型的准确率下降可达27个百分点，凸显了无上下文评估的重要性；四比特量化可使模型内存减少64%-71%，但对无上下文多项选择题（MCQ）准确率的影响较小（-3.2至+1.6个百分点）；而知识蒸馏的效果更具变异性，学生模型规模越小，性能下降越明显，其中基于FANAR-9B蒸馏出的Qwen2.5-7B学生模型在无上下文测试中达到76.6%准确率，略低于教师模型的77.2%。因此，该研究强调无上下文评估应作为补充手段，以更全面地衡量模型压缩与领域适应对阿拉伯语法律生成式模型的真实影响。
链接: https://arxiv.org/abs/2611.07988
作者: Md. Rafiul Biswas,George Mikros,Wajdi Zaghouani
机构: Hamad Bin Khalifa University (哈马德本哈利法大学); Northwestern University in Qatar (西北大学卡塔尔分校)
类目: Computation and Language (cs.CL)
备注: 2 figures, 4 tables
Abstract:Model compression can facilitate the deployment of large language models in resource-constrained legal settings, but its effects on domain-specific performance may depend on the evaluation protocol. We investigate this issue using ArabLegalEval, a Saudi Arabic legal benchmark, under both with-context and no-context evaluation. The former provides supporting legal passages, while the latter assesses performance without external context. Removing the context produces accuracy gaps of up to 27 percentage points for the base model. We examine three stages: domain adaptation, quantization, and knowledge distillation across FANAR-9B, JAIS-8B, and Qwen2.5-7B. Four-bit quantization reduces model memory by 64-71%, with changes in no-context MCQ accuracy ranging from -3.2 to +1.6 percentage points. Distillation produces more variable results, with accuracy generally declining as student size decreases. A Qwen2.5-7B student distilled from FANAR achieves 76.6% no-context accuracy, compared with 77.2% for its teacher. Our findings suggest that no-context evaluation provides a useful complement to context-supported testing when assessing the effects of adaptation and compression on Arabic legal language models.
[NLP-27] How Far Can Lightweight LLM s Go in Intent-to-SLO Translation?
【速读】： 该论文旨在解决意图驱动网络（Intent-based Networking, IBN）中自然语言意图到可执行服务级别目标（SLOs）的自动转换问题，尤其关注在资源受限环境下使用轻量级开源大语言模型（LLMs）替代传统高成本、高能耗的云端多GPU大模型的可行性。其核心挑战在于如何在保证翻译准确性和意图验证能力的同时，降低计算能耗并提升对多样化表达方式的鲁棒性。解决方案的关键在于通过五维评估框架（翻译性能、意图验证、幻觉率、能耗影响及表达多样性鲁棒性），系统对比了10个参数量约100亿的开源大模型在四种提示工程策略下的表现。研究发现，最优配置在严格F1指标上达到0.943，并能拒绝高达191/195个无效意图；采用少样本推理时，最强模型的幻觉率控制在约1%，显著优于零样本设置；同时，翻译单个意图的GPU能耗仅为0.044–0.148 Wh，且推理带来的额外能耗对弱模型具有正向收益。然而，真实场景中的表达重构仍导致15/16配置下无效意图的拒收率下降，表明当前模型在保持稳健意图验证方面仍存在关键挑战，是实现可靠IBN部署的核心瓶颈。
链接: https://arxiv.org/abs/2611.07982
作者: Sabbir Ahmed,Hazem Abbas,Salimur Choudhury
机构: Queen’s University (皇后大学)
类目: Networking and Internet Architecture (cs.NI); Computation and Language (cs.CL)
备注: Under review at ICC 2027. Source code and data: this https URL . (6 pages. 7 figures, 3 tables)
Abstract:Intent-based networking (IBN) promises QoS provisioning in which the network derives the configuration from operator-stated goals. Its first step, translating a natural-language intent into enforceable service-level objectives (SLOs) grounded in a live snapshot of the network state, increasingly relies on large language models (LLMs). These large models add cost and energy to the management loop and typically require multi-GPU, cloud-hosted infrastructure. Can lightweight open LLMs take over this step dependably? We answer this question with a five-dimensional evaluation of ten LLMs with 10B parameters from five families, under four prompting setups, on 716 placement intents. We evaluate translation performance, intent validation, hallucination, energy impact, and robustness to diverse phrasing. The best configurations translate intents with a strict F1 of 0.943 and reject up to 191 of 195 invalid intents. With few-shot reasoning, the strongest models hallucinate unsupported QoS requirements in only about one response in a hundred, whereas the zero-shot setup struggles. A translated intent costs 0.044-0.148 Wh of GPU energy without reasoning, and added energy cost for reasoning pays off for the weaker models. Furthermore, we show that realistic rephrasing lowers the rejection of invalid intents in 15 of 16 configurations for the best-performing models, leaving robust intent validation as a key open challenge for reliable IBN. Our code and data are available: this https URL.
[NLP-28] CSExam: Evaluating Frontier LLM s on the Full Range of TCS Research
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在理论计算机科学（Theoretical Computer Science, TCS）领域中能否推动理论进展并解决长期悬而未决的开放猜想这一核心问题。其解决方案的关键在于构建一个统一、可扩展且可复现的评估框架——TCSExam，该框架涵盖定理证明、求解器工程、数学构造以及对开放猜想的直接求解四类任务，全面评估LLMs在理论计算机科学研究中的实际能力。TCSExam-Full包含100个专家评审的任务，通过Lean内核或可执行检查器验证结果，确保严谨性；TCSExam-Lite则精选20个代表性任务，大幅降低实验成本（减少80%的任务尝试），同时保持与Full版本高度一致的评估效果（斯皮尔曼等级相关系数ρ=0.93）。实验表明，先进模型如GPT-6 Astra和Opus 5.5在基准上展现出显著能力，不仅生成了改进的求解器输出和数学构造，还独立验证了部分开放猜想的特例与边界条件，证明了该框架的有效性与实用性。
链接: https://arxiv.org/abs/2611.07962
作者: Yifan Luo,Zhengying Liu,Yang Yuan,Zhilin Yang,Andrew Chi-Chih Yao
机构: Tsinghua University (清华大学); Moonshot AI; Shanghai Qizhi Institute (上海智元研究院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 52 pages, 15 figures. Project website: this https URL Code: this https URL
Abstract:Can large language models (LLMs) advance theoretical computer science and resolve its long-standing open conjectures? To study progress toward this goal, we introduce TCSExam, a unified framework and benchmark covering the full range of TCS research. It evaluates theorem proving, solver engineering, and constructions alongside direct attempts to resolve open conjectures, capturing both concrete research accomplishments and progress toward long-standing research goals. TCSExam-Full contains 100 expert-reviewed tasks, with submissions checked by the Lean kernel or an executable task checker. TCSExam-Lite selects 20 tasks, five per track, reducing the number of task attempts by 80% at the same per-attempt budget. We evaluate eight agents on Full and nine on Lite. The TCSExam Index combines all four tracks: GPT-6 Astra leads Full (71.9), and Opus 5.5 leads Lite (72.3). Full and Lite rankings are strongly aligned across eight shared agents (Spearman’s \rho=0.93 ), supporting Lite as a compact, discriminative benchmark. The evaluations yield improved solver outputs and mathematical constructions, together with independently verified special cases and bounds from open-conjecture attempts. TCSExam provides an extensible framework with task sources, Docker build recipes, and evaluation tools for reproducible use.
[NLP-29] CoRD: Coupled Rollouts Fill the Blind Spot of On-Policy Distillation
【速读】： 该论文旨在解决在策略蒸馏（on-policy distillation）中，学生模型因无法访问教师模型所达到的状态而导致监督不足的问题，尤其在学生未能覆盖教师路径中的关键状态时，其学习效果受限。其核心解决方案是提出耦合滚动蒸馏（CoRD, Coupled-Rollout Distillation），其关键在于：通过最大推测解码（maximal coupling of speculative decoding）对师生模型的生成轨迹进行几何插值，仅调整训练状态的分布以逼近教师分布，而保持原有的在策略损失不变。该方法在师生分布一致处保留学生生成的标记（token），仅在不一致区域用教师更重视的标记替代，从而实现最小干预下的精准引导。这种机制确保了干预频率极低且随学习过程逐渐减少，显著提升了学生在复杂推理任务中的表现。实验表明，在六个来自Qwen3与Qwen3.5系列（0.6B至4B）的学生模型上，CoRD在五个公开数学基准测试中相较传统在策略蒸馏平均提升pass@1 0.9–2.0点、pass@8 1.2–3.4点；尤其在固定教师可达状态条件下，4B模型使用CoRD仅需16步即可达成在策略蒸馏需80步才能接近的性能差距，且优于其他频繁干预的替代方案。
链接: https://arxiv.org/abs/2611.07959
作者: Guhan Chen,Bohan Li,Jun Rao,Hejin Wang,Songtao Tian,Yexin Xie,Xiaojun Meng,Peng Yuan,Jiansheng Wei,Xuebo Liu,Zixiong Yu
机构: Tsinghua University(清华大学); Harbin Institute of Technology, Shenzhen(哈尔滨工业大学深圳分校); Nanjing University(南京大学); Huawei Large Model Data Technology Lab(华为大模型数据技术实验室)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:On-policy distillation trains a student on its own rollouts, so states the teacher would reach but the student does not are seldom supervised. We introduce CoRD (coupled-rollout distillation), which keeps the on-policy loss and moves only the distribution of training states toward the teacher, drawing rollouts from a geometric interpolation of student and teacher through the maximal coupling of speculative decoding. The coupling keeps the student’s token wherever the two distributions agree and otherwise substitutes a token from the part of the target the student underweights, so interventions are as rare as any exact sampler allows and recede as the student learns. Across six base students from two model families (Qwen3 and Qwen3.5, 0.6B to 4B), CoRD raises average pass@1 over on-policy distillation by 0.9 to 2.0 points and pass@8 by 1.2 to 3.4 points on five public math benchmarks at equal steps and data. On frozen teacher-reached states, a 4B student trained with CoRD reaches in 16 steps a gap closure that on-policy distillation does not reach in 80, and substitution rules intervening at least as often in other directions fall short of CoRD.
[NLP-30] A Behavioural Framework for Predicting and Understanding Loss of Control in Frontier Artificial Intelligence Systems
【速读】： 该论文旨在解决前沿生成式人工智能（Generative AI）系统在长期任务中因人类监督减弱而面临的失控风险问题，即系统可能在超出有效人类控制范围后产生不可逆的负面影响。其解决方案的关键在于提出一个可操作的行为框架，将失控风险分解为三个核心维度：动机错配（misaligned motive）、危害赋能能力（harm-enabling capability）和监控规避行为（monitoring evasion），并通过13个可度量的特征与29项评估任务对这些维度进行量化，最终整合为单一的失控（Loss-of-Control, LoC）评分。实证结果表明，该评分在13个前沿模型上展现出强前瞻性有效性（与受控长周期风险场景中的失败率呈显著正相关，Pearson r = 0.84），且风险水平在不同模型间存在系统性差异，不能仅由模型能力解释。此外，研究进一步验证了该框架可指导针对性缓解措施——针对诊断出的脆弱环节进行微调与监控干预，可在显著降低失败率的同时基本保持模型的通用性能。该方法为提前预测、理解并防范潜在失控风险提供了可落地的技术基础。
链接: https://arxiv.org/abs/2611.07953
作者: Yao Huang,Yichi Zhang,Yifan Wang,Yitong Sun,Haoyu Dong,Jifan Ma,Jinzhe Tu,Xiaojian Li,Sören Mindermann,Wei Xu,Yinpeng Dong
机构: Tsinghua University (清华大学); Shanghai Qi Zhi Institute (上海奇智智能研究院); University of Cambridge (剑桥大学); Institute for Interdisciplinary Information Sciences, Tsinghua University (清华大学交叉信息研究院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 39 pages, 5 figures
Abstract:Frontier artificial intelligence (AI) systems are increasingly operating over longer horizons with diminishing human oversight, raising the risk of loss of control: systems acting beyond effective human control with potentially irreversible consequences. Because such failures may become observable only after containment is difficult, they must be assessed prospectively. We introduce a behavioural framework that decomposes loss-of-control risk into three factors: misaligned motive, harm-enabling capability, and monitoring evasion. We operationalise these factors through 13 measurable aspects and 29 evaluation tasks, whose scores are combined into a single loss-of-control (LoC) score. Across 13 frontier models, the LoC score shows strong prospective validity, correlating with failure rates in contained long-horizon risk scenarios (Pearson r = 0.84). Risk varies systematically across models and cannot be explained by capability alone. We further show that the framework can guide targeted mitigations: fine-tuning and monitoring directed at diagnosed vulnerabilities reduce failure rates while largely preserving general performance. These results provide a practical basis for predicting, understanding, and mitigating loss-of-control risk before it materialises.
[NLP-31] Beyond the Keyword: A Human-Annotated Benchmark for Arabic Anxiety-Related Discourse
【速读】： 该论文旨在解决在阿拉伯语社交媒体文本中精准识别与“焦虑”相关语言的语篇功能（discourse functions）这一难题，尤其针对关键词检索所返回的大量非目标性表达（如个人情绪表达、社会政治焦虑等）导致的语义混杂问题。其核心挑战在于如何在缺乏统一标注标准的情况下，构建一个可靠且可复现的基准数据集，以评估模型对多义焦虑语言的区分能力。解决方案的关键在于：首先通过两名母语阿拉伯语标注者独立标注1000条以“qalaq”（焦虑/担忧）为关键词检索出的推文，形成初步标注；随后对所有分歧和部分一致标签进行深度复审，发现约23.5%的原始共识标签需修正，揭示了标注一致性不可靠的问题；最终基于复审结果建立共识参考标准，并引入大型语言模型Fanar-C-2-27B进行性能评估，结果显示其在宏观F1值（0.784）和加权κ系数（0.749）上表现良好，但主要错误集中于定义不一致的“残差类别”（residual category），表明模型误差部分源于提示工程中的内在矛盾。研究还公开了标注指南、子类划分方案、固定数据划分、标签溯源信息及评估代码，为后续研究提供了可复现的基准资源。
链接: https://arxiv.org/abs/2611.07949
作者: Shimaa Ibrahim,Mabrouka Bessghaier,Md. Rafiul Biswas,Insaf Elouard,Wajdi Zaghouani
机构: 未知
类目: Computation and Language (cs.CL); Social and Information Networks (cs.SI)
备注: 13 pages, 1 figure, 6 tables
Abstract:Keyword retrieval returns expressions of personal anxiety together with many other uses of anxiety-related language. We present a benchmark of 1,000 Arabic Twitter/X posts retrieved with the keyword qalaq (“anxiety/worry”) and annotated for six discourse functions. Two native Arabic-speaking annotators labeled every post independently, reaching 51.1% raw agreement (Cohen’s \kappa = 0.378 ). We then reviewed all 489 disagreements and all 511 agreements under the same guidelines. The review changed 120 of the agreed labels (23.5%), which shows that agreements need auditing as well as disagreements. The final labels form a consensus reference; on a random sample of 200 posts, an additional annotator agreed with it on 60.5% of posts ( \kappa = 0.517 ). Against this reference, Fanar-C-2-27B reaches 79.8% accuracy and macro-F1 0.784 ( \kappa = 0.749 ). Its errors concentrate in the residual category (41.1% of errors; precision 0.905, recall 0.447), whose definition is internally inconsistent in the executed prompt, so the low recall cannot be attributed to the model alone. Sociopolitical Anxiety (27.1%) outnumbers Personal Emotional Expression (22.0%); these are assignment proportions in the sample, not estimates of prevalence or retrieval precision. We release the guidelines, a residual-subtype scheme, a fixed data split, label provenance, and evaluation code.
[NLP-32] CardioLM: A multilingual suite of small language models for the cardiology domain
【速读】： 该论文旨在解决心血管临床文本中多语言、领域特定的命名实体识别（NER）资源匮乏问题，尤其针对非英语语种在生成式医疗自然语言处理（NLP）中的严重不足。现有临床NLP资源高度集中于英文，且缺乏专门面向心脏病学、覆盖多种语言及广泛实体类型（如疾病、症状、操作、药物）的基准数据集与模型。其解决方案的关键在于构建两个核心组件：一是通过机器翻译、标注投影与临床医生审核构建的多语言心脏病学病例报告语料库CardioCCC，涵盖西班牙语、捷克语、荷兰语、英语、意大利语、罗马尼亚语和瑞典语；二是基于该语料库训练的CardioLM框架，包括针对各语言优化的心脏病学预训练语言模型（CardioBERTa）以及在CardioCCC上微调的命名实体识别模型（CardioNER）。研究发现，专科领域预训练虽显著降低伪困惑度，但对严格边界匹配的F1提升有限，而跨语言联合训练的XLM-RoBERTa-large模型在几乎所有语言和实体类型上均取得最优性能，较最强单语系统提升高达0.13 F1，展现出更优的泛化能力与部署效率。此外，放宽匹配标准可进一步提升疾病、症状和操作类实体的识别效果，但对药物实体改善有限，表明边界检测仍是主要挑战。临床评估显示，CardioNER整体评分从基线的26%大幅提升至73%，验证了其实际应用价值。
链接: https://arxiv.org/abs/2611.07940
作者: Manuela-Daniela Danu,Miguel Rodríguez-Ortega,Alberto Becerra-Tome,Pablo Arancibia,Fernando Gallego-Donoso,Lorenzo Pratesi,Xi Bai,Andrei-Damian Dinescu,Salvador Lima-López,Andrea Guala,Saskia Haitjema,Jonas Mack,Michele Miranda,Wesam Al Nabki,Jan Rodríguez-Miret,Vasile George Marica,Vladimír Kincl,Kateřina Dostálová,Michael Andrej,Alicia Ujil,Corinne Allaart,Machteld Boonstra,Malin Overmars,Anoop Shah,Giovanna Ceroni,Jonathan Kossoff,Domenico D’Amario,Renzo Laborante,Antonio Iaconelli,Isabella Fumarulo,Jacopo Lenkowicz,Livia Lilli,Emma Weiss,Alex Scafa,Felix Lindberg,Milou Gamage,Eulàlia Farré-Maduell,Sébastien Bratières,Martin Krallinger,Bram van Es
机构: Siemens SRL, Brasov, Romania; Transilvania University of Brasov, Brasov, Romania; Barcelona Supercomputing Center, Barcelona, Spain; NLP4BIA, Barcelona Supercomputing Center, Barcelona, Spain; Translated, Rome, Italy; University College London, London, United Kingdom; Vall d’Hebron Institute of Research, Barcelona, Spain; CIBER‑CV, Instituto de Salud Carlos III, Madrid, Spain; University Medical Center Utrecht, Utrecht, The Netherlands; Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, the Netherlands; Amsterdam Cardiovascular Sciences Research Institute, Amsterdam, the Netherlands; Sapienza University of Rome, Rome, Italy; International Clinical Research Center, St. Anne ’s University Hospital, Faculty of Medicine, Masaryk University, Brno, Czech Republic; Department of Internal Medicine/Cardioangiology, University Hospital Hradec, Králové, Czech Republic; Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Rome, Italy; Data Science Unit, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Rome, Italy; Department of Clinical Science and Education, Södersjukhuset, Karolinska Institutet, Stockholm, Sweden
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 91 pages, 7 figures, 33 tables (9 in the main text, 23 in Appendix A, and 1 in Appendix B)
Abstract:Background: Cardiovascular care generates large volumes of clinical free-text documentation in local languages. However, clinical natural language processing (NLP) resources remain concentrated in English, and no existing benchmark is cardiology-specific, multilingual, and broad in entity coverage. Methods: We present two contributions. The first is CardioCCC, a corpus of 508 cardiology case reports originally published in Spanish and extended to Czech, Dutch, English, Italian, Romanian, and Swedish through machine translation, annotation projection, and clinician review, with annotations covering diseases, symptoms, procedures, and medications. The second is CardioLM, comprising CardioBERTa, a set of language-specific encoders obtained by continued pretraining on cardiology and biomedical corpora, and CardioNER, a set of named entity recognition (NER) models fine-tuned on CardioCCC. Results: Cardiology-specific pretraining reduced pseudo-perplexity by up to a factor of eight, yet changed strict span-level F1 by at most 0.04, with benefits concentrated in languages whose backbones lacked biomedical coverage. The jointly trained multilingual XLM-RoBERTa-large models achieved the best strict F1 in nearly every language and entity type, exceeding the strongest monolingual system by up to 0.13 F1. A large share of residual errors concerned boundary detection: relaxed matching recovered 0.10-0.23 F1 for disease, procedure, and symptom, but only 0.01-0.07 for medication. Clinicians rated CardioNER well above a rule- and dictionary-based baseline, with the mean rating across entity types rising from 26% to 73%. Conclusions: Specialty-specific pretraining is not the decisive factor for downstream entity recognition, whereas multilingual joint training yields the strongest results in nearly all settings and a simpler deployment path for multi-country clinical information extraction. Comments: 91 pages, 7 figures, 33 tables (9 in the main text, 23 in Appendix A, and 1 in Appendix B) Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.07940 [cs.CL] (or arXiv:2611.07940v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2611.07940 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Manuela-Daniela Danu [view email] [v1] Thu, 5 Nov 2026 12:41:19 UTC (1,689 KB)
[NLP-33] AraEngage: Interpretable Linguistic Correlates of Engagement in Arabic Facebook News
【速读】： 该论文旨在解决将英语语境下开发的社交媒体参与度（engagement）特征直接应用于无字母音标记的阿拉伯语时所面临的测量偏差问题。其核心挑战在于，无字母音阿拉伯语中词汇形态的高度歧义性导致传统语言学特征（如情态动词、祈使句、宗教表达等）在语义和句法层面存在严重混淆，从而影响参与度分析的可靠性。解决方案的关键在于构建并公开发布AraEngage数据集——一个包含10,000条来自全球57个国家2,451个阿拉伯语新闻页的公共Facebook新闻帖子样本，结合可解释的语言学特征与归一化后的受众参与度指标。研究通过随机森林模型与SHAP值分析揭示：宗教表达、消息长度、句子长度及词汇密度是预测粉丝归一化参与度的主要因素（模型决定系数R²=0.38），且正向情感与高参与度显著相关，这与英语语境相反。研究进一步发现“长度悖论”：原始参与度随消息变长而升高，但归一化后模型更倾向于短文本。通过语言学审计识别出标准词典特征在无字母音阿拉伯语中的脆弱性根源，包括同形异义（如“express!”与“via”）、情态标记混淆（qad+未完成式与qad+完成式）以及宗教术语与地名在政治新闻中的混用，这些均可能导致对祈使句与缓和语的关联性被低估。所有结论均为基于操作化指标的观察性关联，研究同时开源了处理后的语言特征、参与度数据、优化词典及分析代码，以促进可复现性与后续研究。
链接: https://arxiv.org/abs/2611.07926
作者: Shimaa Ibrahim,Nour Elhouda Dalloua,Md. Rafiul Biswas,Wajdi Zaghouani
机构: 未知
类目: Computation and Language (cs.CL); Computers and Society (cs.CY); Social and Information Networks (cs.SI)
备注: 13 pages, 4 figures, 8 tables. To appear in ArabicNLP 2026
Abstract:Engagement features developed for English are routinely applied to Arabic, but unvocalized Arabic changes what a lexical feature measures. We introduce AraEngage, a query-defined sample of 10,000 public Arabic Facebook news posts from 2,451 pages in 57 countries (2023-2024) that pairs engagement counts with interpretable linguistic features. We use it to examine which interpretable Arabic indicators are associated with follower-normalized engagement, and how far those associations can be trusted given publisher context, the definition of engagement, and lexical ambiguity. Religious expressions, message length, sentence length, and lexical density are the leading correlates in a Random Forest model ( R^2 = 0.38 on a post-level split) interpreted with SHAP. Unlike English news, positive sentiment is associated with higher engagement in this Arabic news sample. Message length shows a length paradox: unadjusted raw engagement is highest for long posts, whereas the model’s attributions for follower-normalized engagement favor short posts. A linguistic audit identifies why standard lexicon features are fragile in unvocalized Arabic: imperative lists contain homographs such as the form read as “express!” that is also the frequent preposition “via”, hedge lists conflate the possibility marker qad + imperfect (“may”) with the emphatic qad + perfect (“has indeed”), and religious lists mix devotional formulas with place names in political news. These are plausible sources of measurement error that may attenuate the near-zero associations of imperatives and hedges. All findings are observational associations with operationalized indicators. We release the derived linguistic features, engagement counts, refined lexicons, and analysis code; raw posts are not redistributed, in line with CrowdTangle terms.
[NLP-34] Do Recommenders Need to Think in Words? Latent Reasoning for Efficient LLM -Based Recommendation CIKM2026
【速读】： 该论文旨在解决基于大语言模型（LLM）的推荐系统在推理阶段生成链式思维（Chain-of-Thought, CoT）文本所导致的高计算开销问题，即每次推荐需额外消耗数十至数百个标记（token），显著增加延迟。其核心解决方案是提出一种名为LatentRec的框架，通过训练阶段的显式监督替代推理阶段的思维生成过程。该框架的关键在于：首先引入一个250万参数的连续思维模块（Continuous Thought Module），在训练过程中对LLM隐藏状态迭代执行K步潜在推理；其次采用软令牌选择（Soft Token Selection）机制，弥合隐藏状态与输入嵌入之间的约14倍范数差异；最后设计逐步推理损失（Per-step Reasoning Loss, PRL），利用预测头对每个中间状态进行监督，目标为首个语义标识符（Semantic ID）的起始词。该方法使模型在训练完成后无需依赖推理阶段的思维文本或潜在表示即可保持高精度，从而可在部署时完全移除该模块。实验表明，在三个Amazon基准数据集上，LatentRec相较经过CoT蒸馏的基线模型在HR@10指标上提升4.0%-5.0%（相对），同时实现18.3倍的推理延迟降低（批大小=1）。该方案实现了“一次训练、持续受益”的高效监督机制，显著提升了生成式推荐系统的实用性。
链接: https://arxiv.org/abs/2611.07866
作者: Hongyang Su,Beibei Kong,Lei Cheng,Chengxiang Zhuo,Zang Li
机构: Tencent(腾讯)
类目: Computation and Language (cs.CL)
备注: Accepted at CIKM 2026 (Short Paper). 5 pages, 2 figures, 3 tables
Abstract:LLM-based recommender systems increasingly generate chain-of-thought (CoT) text before each item prediction to activate the LLM’s reasoning capacity, costing tens to hundreds of extra tokens per recommendation at inference. We ask whether inference-time reasoning artifacts (textual or latent) can be replaced by training-time supervision alone, without loss of accuracy. We propose LatentRec, a framework for LLM-based generative recommendation that replaces inference-time reasoning artifacts with training-time per-step head supervision. LatentRec has three components. A 2.5M-parameter Continuous Thought Module iterates K latent reasoning steps over the LLM’s hidden state during training. Soft Token Selection bridges the ~14x norm gap between hidden states and input embeddings. A Per-step Reasoning Loss (PRL) supervises every intermediate state through the LLM’s prediction head against the first token of the target Semantic ID. The trained LLM retains this accuracy without the module at inference, so the module can be discarded at deployment. On three Amazon benchmarks, LatentRec improves HR@10 by 4.0-5.0% (relative) over a CoT-distilled baseline at 18.3x lower latency (batch=1). Per-step training supervision consistently improves accuracy across three datasets, paid once at training rather than at every recommendation.
[NLP-35] owards Foundation Decision Models: Drex
【速读】： 该论文旨在解决自回归式大语言模型在实际应用中面临的答案不一致、输出需解析、成本与延迟高等问题。其核心挑战在于如何在保持高准确率的同时，实现高效、稳定且无需自回归解码的决策生成。解决方案的关键在于提出Drex系列开放决策模型（Decision Models），通过非自回归方式直接对预定义选项进行概率分布预测，从而避免传统链式思维（Chain-of-Thought）推理带来的不确定性与计算开销。该方法结合监督微调与调整反馈强化学习（Reinforcement Learning from Adjusted Feedback, RLAF），并针对视觉-语言能力增强的Max版本引入视觉调优。实验表明，RLAF显著提升模型准确性，一致性损失有效增强对选项顺序变化的鲁棒性。在Decision Index 0.4.1基准上，Drex Max达到67.52的分数，刷新决策模型新纪录；在Decision Bench v4和DecideBench上均表现出领先性能，且响应速度仅为183毫秒，远低于GPT-6 Sol（2.0秒）和Claude Opus 5.5（3.5秒），同时输出在选项重排下保持稳定。所有模型均已开源。
链接: https://arxiv.org/abs/2611.07854
作者: Nischay Dhankhar,Alisher Amantay,Zhanibek Datbayev,Darkhan Akhmetov,Yash Shah,Nurmukhamed Ubaidullayev,Raman Saparkhan,Dos Baha,Abulhair Saparov
机构: 未知
类目: Computation and Language (cs.CL)
备注: 51 pages, 13 figures, 25 tables. Models and code: this https URL
Abstract:Autoregressive approaches with large language models, such as long chain-of-thought reasoning, have become the default for new AI applications because they reach higher accuracy on a wide range of tasks. In production, however, they give inconsistent answers, require parsing of outputs, and add cost and latency. Decision models address this: given a question and a set of options supplied by the user, they return a probability distribution over the options without autoregressive decoding. We present Drex, a family of four open decision models with 8B (DLM), 9B (v1.5), 35B with 3B active (v2.0), and 27B (Max) parameters. Each request consists of a state (a document, record, or agent context, optionally with images) and questions about it, each asking for a choice among options, a yes/no judgment, or a score on an ordered scale. Training combines supervised fine-tuning with reinforcement learning from adjusted feedback (RLAF), plus vision tuning for Max. Ablations show that RLAF improves accuracy and that consistency losses improve robustness to option order. On Decision Index 0.4.1, Max sets a new state of the art among the decision models we compare (67.52 on the public suite vs. 60.70 for Jev 1.14.0). On Decision Bench v4, it is statistically tied with d3-turbo and GPT-6 Sol and trails Claude Opus 5.5 by 1.6 points, and it has the highest accuracy among decision models on DecideBench. Max answers a typical request in 183 ms, against about 2.0 s for GPT-6 Sol and 3.5 s for Claude Opus 5.5, and its outputs are stable under option reordering. All four models are open source: this https URL
[NLP-36] Generate-then-Rank: Learning Symbolic Goal Interpretation for Embodied Agents
【速读】： 该论文旨在解决具身智能体在执行任务时，如何将自然语言指令准确转化为可被规划器执行的符号化目标（symbolic goal）的问题。这一过程称为目标解析（Goal Interpretation），其准确性对下游规划与动作执行具有关键影响，任何误差均会逐级放大。现有方法在生成合理候选目标及可靠筛选最优解方面存在不足。本文提出一种两阶段“生成-排序”框架：首先通过监督微调（supervised fine-tuning）使大语言模型（Large Language Model, LLM）适配具身智能体接口（Embodied Agent Interface, EAI）的输出规范；其次引入基于错误反馈的候选输出排序机制（Candidate-Output Ranking from Errors, CORE），利用模型自身生成的多个备选方案，学习对完整输出的偏好关系，从而实现更鲁棒的目标选择。实验在EAI基准的家居场景中进行，评估了包括DeepSeek-V3 chat、DeepSeek-Reasoner、Qwen2.5-7B和Qwen2.5-72B在内的多款开源LLM，结果表明所提方法显著优于仅使用提示工程、思维链（chain-of-thought）或基于知识图谱的基线模型。解决方案的关键在于结合监督微调提升生成质量，并通过自生成候选集的排序机制实现对输出质量的动态优化。
链接: https://arxiv.org/abs/2611.07825
作者: Jisen Lei,Huizhi Liang
机构: Newcastle University (纽卡斯尔大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Robotics (cs.RO)
备注: 14 pages, 2 figures. Accepted at the 10th Euro-China Conference on Intelligent Data Analysis and Applications (ECC 2026)
Abstract:Goal interpretation is to translate a natural language instruction into a formal symbolic goal that a planner can execute. It is critical for embodied agents because any error here propagates to every downstream planning and action step. Doing it well requires both generating suitable candidates and reliably selecting among them. We present a two-stage generate-then-rank approach to Goal Interpretation in the Embodied Agent Interface (EAI). First, supervised fine-tuning adapts a Large Language Model (LLM) to EAI’s output convention. Second, Candidate-Output Ranking from Errors (CORE) learns preferences over complete outputs from the model’s own generated alternatives. We conducted experiments on the household scenes of the Embodied Agent Interface Benchmark. Multiple open source LLMs including DeepSeek-V3 chat, DeepSeek-Reasoner, and Qwen2.5-7B and Qwen2.5-72B were evaluated. The proposed method outperformed the baseline models including LLMs with simple prompting, chain-of-thought, and knowledge graph based approaches.
[NLP-37] Axis-LoRA: Probing and Adapting Embedding Models for Personal-Memory Retrieval EMNLP2026
【速读】： 该论文旨在解决大语言模型（LLM）在处理用户个人记忆检索时，现有基于固定密集嵌入（dense embedding）模型的个人记忆管道无法有效区分记忆的时间、地点、人物和事件四个维度的问题。研究表明，当前嵌入模型对这些轴向信息的编码能力不足，即使调整记忆的时间或地点，其嵌入表示变化微弱，且在固定嵌入基础上添加训练投影也无法恢复缺失的轴向信号。问题根源在于模型内部各轴向信息由不同的注意力头模式编码，但未被显式分离与利用。为此，论文提出Axis-LoRA——一种针对每个轴向设计的适配器架构，将四个LoRA分支嵌入到嵌入模型中，并通过查询相关的门控机制动态融合各轴向信息。实验表明，在LongMemEval、ATM-Bench和LoCoMo三个基准上，Axis-LoRA在仅微调极小比例基础参数的情况下，实现了4个基准平均R@10提升1.28至1.56个百分点，显著提升了对多维记忆的精确检索能力。
链接: https://arxiv.org/abs/2611.07820
作者: Dohoon Kim,Junho Park,Hyunwoong Bae,Seyoon Kim,Taesup Moon
机构: Seoul National University (首尔国立大学); Samsung Electronics (三星电子)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: EMNLP 2026 Main
Abstract:As LLMs are increasingly asked questions about a user’s personal experiences, retrieval must distinguish memories by time, place, person, and event. Current personal-memory pipelines rely on frozen dense embedding models to embed memories for retrieval, but we find that these models do not capture the four axes well. Swapping a memory’s date or place barely shifts its embedding, and a trained projection on top of the frozen embedding does not recover the missing signal. The problem lives inside the model, where each axis is encoded by a distinct attention-head pattern. Guided by these findings, we propose Axis-LoRA, a per-axis adapter that places four LoRA branches inside the embedding model and combines them via a per-query gate. Across three standard embedding models evaluated on LongMemEval, ATM-Bench, and LoCoMo, Axis-LoRA improves 4-benchmark mean R@10 by +1.28 to +1.56 ~pp while training only a small fraction of the base parameters on a small axis-counterfactual training set. Code is available at this https URL.
[NLP-38] Different Corpora Different Skills? A Paired Study of Small Language Models EMNLP2026
【速读】： 该论文旨在解决在小规模语言模型（small language models）训练中，受限于数据预算时，预训练语料库选择对模型性能的潜在影响问题。研究发现，在相同1000万与3700万词的数据预算下，采用1911年《大英百科全书》（Encyclopaedia Britannica）与BabyLM 2026严格预训练语料库所生成的模型展现出截然不同的能力分布，而非单一的质量优劣排序。其解决方案的关键在于通过严格控制模型架构、训练流程与评估协议，系统性对比不同语料库在多基准测试（包括BLiMP、BLiMP Supplement、EWoK、Entity Tracking及COMPS）上的表现差异。结果表明，BabyLM 2026在语言可接受性任务上持续占优，而《大英百科全书》则在合成的状态追踪任务（Entity Tracking）中表现更佳，且该差异在教师模型与蒸馏学生模型中均稳定存在，尤其在5800万参数的蒸馏学生模型中实现最清晰的跨种子复现。进一步的消融实验表明，该差异并非由训练暴露量或上下文长度不匹配所致，且无法通过教师集成复制出《大英百科全书》学生的状态追踪行为，说明语料库特性本身对模型能力塑造具有独立影响。因此，研究强调在小规模模型训练中，语料库选择会显著改变模型的相对技能谱，从而凸显了在模型比较中采用多基准评估的重要性。
链接: https://arxiv.org/abs/2611.07812
作者: Edoardo Vaira,Georgios Chavales,Pierre Nugues
机构: Lund University (隆德大学)
类目: Computation and Language (cs.CL)
备注: Published in the Proceedings of the Second BabyLM Workshop (EMNLP 2026). 13 pages
Abstract:Small language models are often trained under tight data budgets, where the choice of pretraining corpus may matter as much as model size or training recipe. We compare two contrasting corpora at matched 10-million- and 37-million-word budgets: the 1911 Encyclopaedia Britannica and the BabyLM 2026 strict pretraining corpus. For each corpus, we train teacher models, distilled 58-million-parameter Baby-Llama students, and no-distillation student ablations, using the same model families, training recipes, and evaluation protocol across corpora. Across BLiMP, BLiMP Supplement, EWoK, Entity Tracking, and COMPS, the corpora produce different skill profiles rather than a single quality ranking. BabyLM 2026 is consistently stronger on linguistic acceptability benchmarks, while Britannica is often stronger on Entity Tracking, a synthetic state-tracking task. The contrast appears across teacher and student models and at both budgets, with the cleanest cross-seed replication in the distilled student comparison. Targeted controls show that the central contrast persists after matching training exposure and at a longer diagnostic context; direct teacher ensembling also does not reproduce the 10M Britannica student’s Entity Tracking behavior. Entity Tracking magnitudes nevertheless remain sensitive to training dynamics. For this corpus pair, corpus choice can therefore produce different relative skill profiles at small scale, motivating multi-benchmark evaluation when comparing pretraining data.
[NLP-39] Decodability Is Not Validity: Evaluating Nuisance Sensitivity in Language Model Concept Probes
【速读】： 该论文旨在解决生成式AI模型中概念提取器（concept extractor）在预测目标概念时可能受到与目标相关但无关的干扰属性（nuisance attribute）影响的问题，即：两个在预测性能上等效的概念探测器（concept probe）可能对同一目标概念具有不同的干扰敏感性。其核心问题是：传统的基于可解码性（decodability）的评估方法无法可靠反映概念提取器对干扰属性的实际敏感程度，从而导致对模型内部表征的误判。解决方案的关键在于引入“判别性证据”（discriminant evidence）作为评估标准，通过控制输入中仅改变干扰属性而保持目标属性不变的编辑实验，并结合大语言模型（LLM）判断以区分真实效应与安慰剂效应，从而量化提取器对干扰属性的敏感性。研究发现，在训练数据存在相关性的情况下，即使提取器在验证集上的AUROC表现相同，其干扰敏感性仍可相差数倍；而通过引入需标注干扰属性但无需反事实样本的开发诊断工具，可有效筛选出对干扰不敏感的提取器，显著降低反事实干扰敏感性，仅付出微小的准确率代价。因此，论文强调：概念测量必须进行验证，而非仅依赖可解码性。
链接: https://arxiv.org/abs/2611.07805
作者: Ayal Klein
机构: Ariel University ( Ariel 大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 4 pages + appendix; code and data: this https URL
Abstract:Two concept probes can be equally accurate at predicting a concept from a language model’s hidden states while their scores contain different mixtures of the target concept and a correlated nuisance attribute. In four target–nuisance settings (two aspect sentiments in CEBaB; occupation and gender in BiasBios; Amazon review sentiment with negation, and with first-person framing), two LLMs, and a population of concept extractors, we compare decodability with discriminant evidence: the score’s dependence on the nuisance at fixed target, and its response to input edits intended to change only the nuisance, filtered by two LLM judges and compared against placebo edits. Under correlated training data, extractors that are statistically equivalent in held-out AUROC differ up to several-fold in nuisance sensitivity in three of the four settings, and decodability is not a reliable proxy for it: their rank correlation changes sign across settings. The effect follows the probe’s training data, not the nuisance’s natural association with the target: first-person framing, unrelated to sentiment in natural reviews, is absorbed at least as much as negation, which co-occurs with negative sentiment, and decorrelated training removes the systematic effect. Selecting extractors with nuisance-aware development diagnostics, which need nuisance labels but no counterfactuals, lowers counterfactual nuisance sensitivity wherever the default choice is sensitive, at a small accuracy cost. Concept measurements should be validated, not just decoded.
[NLP-40] Dataset Collection and Fine-Tuning of NLLB for German–English Translation
【速读】： 该论文旨在解决高质量德语-英语平行语料库稀缺及通用多语言模型在特定领域（如文学与技术文本）翻译性能不足的问题。其核心解决方案在于：通过构建一套可复现的语料筛选流程，从OPUS Books中提取并净化出2.5万条高质量双语句子对，该流程涵盖编码验证、长度过滤、语言识别、去重及质量评分等步骤；在此基础上，针对消费级GPU优化了训练策略，采用混合精度训练、梯度累积与梯度检查点技术，实现对Meta NLLB-200系列中NLLB-600M模型的有效微调。实验结果表明，微调后的模型在文学与技术类德语-英语翻译任务上相较基线模型显著提升（BLEU提高3.4分，chrF++提高2.6分，TER降低3.2分）。此外，研究将微调模型成功集成至开源文档翻译系统LinguaHaru，实现了对PDF与Office文档的格式保持型翻译，具备实际应用价值。所有数据处理脚本、训练代码与评估工具均已开源，确保研究可复现性，微调模型亦已发布于HuggingFace平台。
链接: https://arxiv.org/abs/2611.07797
作者: SK Deeraj,Yash Ingle,Pruthwik Mishra
机构: SVNIT (Sardar Vallabhbhai National Institute of Technology)
类目: Computation and Language (cs.CL)
备注: 6 pages, 5 tables
Abstract:This paper describes the curation of a high-quality parallel corpus from OPUS Books and the fine-tuning of Meta’s NLLB-200 models for German–English translation. We detail a dataset extraction pipeline that filters 45k raw OPUS sentence pairs to 25k high-quality examples through encoding validation, length filtering, language identification, deduplication, and quality scoring. We present a reproducible fine-tuning methodology optimized for consumer GPUs using mixed precision, gradient accumulation, and gradient checkpointing. Our fine-tuned NLLB-600M model achieves statistically significant improvements (+3.4 BLEU, +2.6 chrF++, -3.2 TER) over the base model on literary and technical German-English translation. We demonstrate practical deployment by integrating the fine-tuned model into LinguaHaru, an existing open-source document translation system, achieving format-preserving translation across PDF and Office documents. All dataset preparation scripts, training code, and evaluation tools are provided for reproducibility. The fine-tuned model is available at HuggingFace.
[NLP-41] Bridging Structure and Semantics: Knowledge Graph Verification via Triple-to-Text Alignment
【速读】： 该论文旨在解决知识图谱中三元组（triple）语义准确性的评估问题，即如何更有效地判断三元组的真伪。现有方法主要依赖图结构信息或基于预训练语言模型（PLM）的文本表示，但难以充分融合结构与语义信息。其解决方案的关键在于提出GenVer——一种生成式验证器（generative verifier），通过联合利用三元组的序列化形式及其生成的自然语言表示（NLRs）进行分类。具体而言，该方法首先通过三元组-文本对齐（triple-text alignment）过程构建训练数据，进而使用序列到序列生成器生成自然语言描述；随后，基于BERT的分类器融合三元组原始特征与其生成的NLR特征，实现更全面的语义建模。在四个基于FB15k构造的FactFB变体数据集上的实验表明，GenVer在三元组分类任务中达到最高准确率（88.56%–91.52%），验证了生成式表示在提升语义准确性方面的有效性。然而，案例分析也揭示了生成表示中存在的语义漂移（semantic drift）及高置信度错误等问题，提示未来需进一步优化生成质量与语义一致性。
链接: https://arxiv.org/abs/2611.07780
作者: Ziang Cui
机构: Southeast University (东南大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Semantic accuracy is an important dimension of knowledge graph quality. Graph-based scoring methods exploit relational structure, while textual representations allow classifiers to draw on knowledge encoded by pre-trained language models (PLMs). We present GenVer, a generative verifier that scores triples using both their serialized form and generated natural language representations (NLRs). A triple-text alignment procedure selects training pairs for a sequence-to-sequence generator; a BERT-based classifier then combines features from each triple and its generated NLR. On four FactFB variants, constructed from FB15k using perturbed negative triples, GenVer obtains the highest classification accuracy among the evaluated baselines, ranging from 88.56% to 91.52%. The experiments suggest that generated representations can help triple classification, while the case analysis also exposes semantic drift and high-confidence errors.
[NLP-42] Scaling Laws for SAE Training Data
【速读】： 该论文旨在解决稀疏自编码器（Sparse Autoencoder, SAE）训练中激活值存储（activation storage）成为瓶颈的问题。其核心问题是：在固定总训练预算下，SAE需要多少个唯一的激活值才能达到最优性能？研究发现，随着缓冲区（buffer）大小的增加，模型质量迅速提升，随后进入平台期——此时重复回放已有激活值的效果几乎等同于持续使用新激活值进行训练。超过此平台点后，进一步增加唯一激活值的数量对重构精度或可解释性贡献极小，因此缓冲区无需随训练预算增长而扩大。解决方案的关键在于提出一种针对有限数据的缩放定律（scaling law），量化了激活多样性与重复利用之间的权衡关系。通过在不同字典规模、令牌预算、Llama-3.1与Qwen3多个模型层级以及多种SAE结构（BatchTopK、TopK、JumpReLU）上的广泛验证，结果表明：应选择达到质量平台期的最小缓冲区，并在整个剩余训练预算中反复回放该缓冲区。实验显示，该策略可将激活存储需求降低8至64倍，同时保持几乎无损的性能表现。
链接: https://arxiv.org/abs/2611.07756
作者: Nikita Koriagin,Nikita Balagansky,Daniil Gavrilov
机构: Yandex(雅虎); T-Tech
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Activation storage is often a bottleneck in sparse autoencoder (SAE) training. We study how many unique activations an SAE needs when its total training budget is fixed. Quality improves rapidly as the buffer grows, then reaches a plateau where replaying the buffer is nearly as effective as using fresh activations throughout training. Beyond this point, additional unique activations bring little benefit to reconstruction or interpretability, so the buffer need not grow with the training budget. We describe this tradeoff between diversity and repetition with a scaling law for limited data. We validate it across dictionary sizes and token budgets, early, middle, and late layers of Llama-3.1, Qwen3 model sizes, and BatchTopK, TopK, and JumpReLU SAEs. Sweeps using mathematical text and Python code test the effect of changing the activation source. Evaluation covers explained variance (EV), reconstruction MSE, feature absorption, automated interpretability, and sparse probing. These results give a practical training rule: choose the smallest buffer that reaches the quality plateau and replay it for the remaining budget. In our experiments, this reduces activation storage by 8\text–64\times with negligible quality loss.
[NLP-43] Automated Video-to-QA Pipeline: End-to-End Question Generation and Verification Using Large Language Models
【速读】： 该论文旨在解决大规模教育类视频内容在YouTube等平台上的快速增长所带来的知识高效检索与评估难题。其核心挑战在于如何从非结构化视频流中自动化生成高质量、结构化的摘要及精准的问答（QA）对，以支持智能化学习资源构建。解决方案的关键在于提出一种自验证与微调循环（Self-Verification and Fine-Tuning Loop）机制：首先利用Whisper进行语音识别生成高精度转录文本，再通过Facebook的BART模型对长文本进行分段与摘要；随后采用meta-llama/Llama-3.1-8BInstruct模型生成候选QA对，并将其反馈至同一模型进行自我评估，以筛选出语义相关且重要的高质量问答对，从而构建合成数据集；最终基于该数据集使用低秩适应（LoRA）对Llama模型进行微调，获得具备领域对齐能力的专用评估问答生成模型。该方法实现了无需人工标注的可扩展、自迭代式教育工具开发，显著提升了自动化知识提取与评估的效率与质量。
链接: https://arxiv.org/abs/2611.07755
作者: Jaival Chauhan,Vikram Singh,Pruthwik Mishra
机构: Sardar Vallabhbhai National Institute of Technology, Surat, India(萨达尔·瓦拉巴伊·国立技术学院，苏拉特，印度)
类目: Computation and Language (cs.CL)
备注: 4 pages
Abstract:The exponential growth of educational and informational video content on platforms like YouTube presents a challenge for efficient knowledge retrieval and assessment. This paper proposes an end-to-end automated pipeline for extracting structured summaries and generating high-fidelity Question-Answer (QA) pairs from video streams. The framework integrates a cascade of state-of-the-art models: audio extraction is handled via FFmpeg, followed by automatic speech recognition (ASR) using Whisper (A library in Python) to generate accurate transcripts. Due to the token constraints of Large Language Models (LLMs), transcripts are segmented and summarized using Facebook’s BART. A key innovation of this work is a Self-Verification and Fine-Tuning Loop. We utilize meta-llama/Llama-3.1-8BInstruct to generate candidate QA pairs from transcript chunks, which are then fed back into the model to rigorously assess their relevance and importance. Only verified, high-quality pairs are retained to construct a synthetic dataset. This dataset is subsequently used to fine-tune the Llama model using Low-Rank Adaptation (LoRA), resulting in a specialized model capable of generating domain-aligned assessment questions. This approach demonstrates a scalable method for creating self-improving educational tools without the need for manual data annotation. The proposed model can be downloaded from Kaggle with all the codes from GitHub.
[NLP-44] Extending Reasoning in Typed Decision Models via Looping
【速读】： 该论文旨在解决生成式语言模型在推理阶段无法根据输入难度动态增加计算量的问题，而传统固定结构的判别式模型（Typed decision models）虽能在单次前向传播中完成决策，却缺乏对复杂输入进行更深入推理的能力。其核心解决方案是提出一种名为RELOAD（retrofitted loops for adaptive decisions）的方法，通过在预训练编码器或解码器的中间层块引入一个经过阻尼处理的循环结构（damped loop），并辅以轻量级控制器自适应地决定每个输入的迭代次数，从而在不增加参数量的前提下，在隐空间中实现可变的推理深度。实验表明，该方法在0.4B至4B参数的六种模型上均显著提升了准确率与校准性能，尤其在需要长链推理且训练中未充分覆盖的任务上，编码器性能提升最高达23个百分点。即使在匹配训练计算量或增加数据的情况下，该优势仍无法被弥补；而采用每轮独立层的方案虽接近效果，但需近两倍参数。控制器可在约四分之三的计算开销下保留大部分性能增益，体现了良好的效率-性能权衡。
链接: https://arxiv.org/abs/2611.07736
作者: David Berghaus,Arber Zela,Behzad Shomali,Markus Frey,Mehdi Ali,Rafet Sifa,Volkan Cevher
机构: Lamarr Institute for Machine Learning and Artificial Intelligence, Bonn, Germany; Fraunhofer IAIS, Sankt Augustin, Germany; University of Bonn, Bonn, Germany; EPFL, Lausanne, Switzerland
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Typed decision models choose among a fixed set of answer options in a single forward pass, so unlike generative language models they cannot spend more test-time compute on hard inputs by generating more tokens. Looping a block of layers offers this compute in latent space without adding parameters, but prior work found no gain from loops retrofitted into conventionally pretrained models. We show that such loops help when trained together with the decision model. Our method, RELOAD (retrofitted loops for adaptive decisions), fine-tunes a pretrained encoder or decoder with a damped loop over its middle block and a lightweight controller picks the number of passes adaptively for each input. Across six models from 0.4B to 4B parameters, looping improves both accuracy and calibration, mostly on problems that need longer chains of reasoning than any seen in training, where encoders gain up to 23 points. Neither matched training compute nor more data closes this gap, and separate layers for each pass come close only with nearly twice the parameters. The controller keeps most of the gain at about three quarters of the compute.
[NLP-45] Evaluating Token Probabilities as Confidence Estimators in Human Behavior Simulations: A Case Study in Luxembourgish Next Like Prediction EMNLP2026
【速读】： 该论文旨在解决当前大型语言模型（LLM）在模拟人类行为时，仅以准确性作为评估标准所隐含的局限性——即默认人类行为是确定性的，而忽视了真实数字痕迹数据中的噪声特性。其核心问题在于：现有模型在面对带有噪声的用户行为序列时，不仅需要准确预测结果，还需具备良好的置信度校准能力（calibration），以确保输出的概率分布真实反映预测的不确定性。解决方案的关键在于引入任务特定的监督微调（supervised fine-tuning），通过此方法显著改善模型的置信度校准性能，将期望校准误差（Expected Calibration Error）从0.2降至0.006，同时提升准确率至0.758，证明了校准与准确性可独立优化且需协同评估。研究强调，在使用基于大模型的智能体进行人类行为建模时，必须将置信度估计与准确性并重，才能真正保障其操作有效性。
链接: https://arxiv.org/abs/2611.07735
作者: Nils Schwager,Simon Münker,Alistair Plum,Achim Rettinger
机构: Trier University (特里尔大学); University of Luxembourg (卢森堡大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 12 pages, 1 figure, 2 tables. Accepted to 11th Workshop on Natural User-generated Text (W-NUT) at EMNLP 2026, Budapest, Hungary
Abstract:As large language models (LLMs) are increasingly used to simulate human behavior, their operational validity is typically assessed through accuracy alone, implicitly treating human behavior as deterministic. We argue that on noisy digital trace data, a model’s performance depends not only on predicting the correct outcome but on its calibration. We present a case study testing whether the token probabilities of open-weight LLMs can serve as confidence estimates in a Next Like Prediction task: predicting how a user of the Luxembourgish news platform this http URL rates a comment, conditioned on their prior interactions. Evaluating Ministral 3 (3B, 8B), we find that, with in-context learning only, the tested models are overconfident and are outperformed on accuracy and calibration by a simple statistical baseline. Accuracy and calibration do not move together: scaling from 3B to 8B improves accuracy while degrading calibration, and longer behavioral histories raise accuracy while confidence declines. Task-specific supervised fine-tuning acts as an effective calibration intervention, reducing Expected Calibration Error from 0.2 to 0.006 while improving accuracy from 0.536 to 0.758. We conclude that confidence estimation should be evaluated alongside accuracy to improve validity when using LLM-based agents.
[NLP-46] O2PD: Overlap-Gated On-Policy Distillation on Reliable Prefix
【速读】： 该论文旨在解决现有基于策略的蒸馏（On-policy Distillation, OPD）方法在生成式语言模型训练中存在两个核心问题：一是缺乏对教师模型监督可靠性的明确判断，二是当前主流的top-k OPD方法在对齐过程中存在支持集覆盖不全与模式错配的问题。具体而言，传统方法依赖教师模型对其自身生成轨迹进行评分，但这些轨迹本质上是“离策略”的，导致教师指导的可靠性存疑；同时，仅限制于教师top-k支持集的对齐会忽略其余分布质量，造成信息损失。为回答“应与哪些词元对齐”和“如何有效对齐”这两个关键问题，本文提出重叠门控的在线策略蒸馏（Overlap-gated On-Policy Distillation, O²PD）。在对齐对象选择上，通过量化分析发现学生模型轨迹前缀部分的教师监督更具可靠性，因此仅对每个学生轨迹前20%的词元实施蒸馏。在对齐机制设计上，理论揭示了top-k OPD存在的“教师支持未揭示”与“主导模式错配”两大缺陷，并提出两种支持策略：重置支持（Reset Support）通过引入重置原子聚合教师top-k之外的质量，实现模式覆盖；支持收敛（Support Convergence）则引导支持集向已对齐交集区域收敛，提升对齐一致性。最终，O²PD采用设计的重叠门控α散度，在每词元层面动态插值模式覆盖与模式聚焦行为，利用重叠区域作为门控信号自适应调节α参数，从而实现更稳健、高效的词元级对齐。
链接: https://arxiv.org/abs/2611.07731
作者: Shuaishuai Zu,Jihao Zhao,Pengfei Li,Guanghuan Fang,Ben Feng,Yuhan Liu,Biao Qin,Baochuan Yang
机构: Renmin University of China (中国人民大学); Huawei Technologies Ltd. (华为技术有限公司)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:On-policy distillation (OPD) supplies dense token-level supervision by aligning a student to the teacher’s next-token distribution on its own generated rollouts. Existing works mostly implement top- k OPD, applying reverse Kullback-Leibler ( \RKL ) divergence between the teacher’s and the student’s renormalized distributions over teacher top- k support. However, the teacher scores student-generated rollouts that are inherently off-policy for itself, leaving the reliability of its next-token guidance unclear. Restricting the divergence to the teacher top- k support also discards the remaining mass outside that support. These gaps leave two questions implicit: 1) which tokens to align, and 2) how to align those tokens. To answer these questions, we propose Overlap-gated On-Policy Distillation (O ^2 PD). As for 1), we quantitatively analyze the reliability of teacher supervision and find that alignment on the student’s trajectory prefix is more reliable than alignment on the full context. Therefore, our O ^2 PD is applied only to the 20% prefix tokens of each student rollout. As for 2), we theoretically characterize two problems of top- k OPD: the \emphTeacher-support Uncovering and the \emphDominant-Mode Mismatch. Based on these analyses, we propose two support strategies: the \emphReset Support strategy attaches a reset atom to every token, pooling mass outside the teacher top- k , and the \emphSupport Convergence strategy converges the support to the already aligned intersection. Our O ^2 PD employs the designed overlap-gated \alpha -divergence to interpolate between mode-covering and mode-seeking for per-token alignment, where the overlap is used as a gate to switch the support and adaptively modulate \alpha .
[NLP-47] Continued Training of Continuous Language Flows with Token Identity Squared Supervision
【速读】： 该论文旨在解决预训练语言模型在连续生成过程中，如何通过有限的微调实现高质量文本生成的问题，核心挑战在于现有方法中监督信号与连续生成过程之间的不匹配。其解决方案的关键是提出令牌身份布里尔流（Token Identity Brier Flow, TIBF），该方法在保持预训练嵌入固定的前提下，直接对令牌身份施加平方形式的监督，通过结合网络得分与令牌匹配得分，并在高斯观测模型下使用布里尔损失（Brier loss）监督令牌概率分布，同时利用概率加权嵌入进行连续生成。TIBF通过理论分析证明了预测概率与真实后验之间的平方误差可上界连续去噪预测和流速的误差，进而通过风险分解揭示了观测不确定性与后验拟合误差的分离机制，从而指导训练聚焦于概率预测而非嵌入更新。实验表明，在相同微调预算下，冻结嵌入的TIBF在Open Web Text (OWT) 和 LM1B 数据集上分别达到33.96和81.17的指标分数，显著优于连续均方误差（continuous-MSE）和交叉熵（CE）等基线方法，并在所有采样预算下均取得最低重复率与最优生成质量。
链接: https://arxiv.org/abs/2611.07721
作者: Jiawei Jiang,Zhaoxiang Zang,Zhengtao Yu
机构: Kunming University of Science and Technology (昆明理工大学); Key Laboratory of Artificial Intelligence in Yunnan Province (云南省人工智能重点实验室)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion and flow models have recently been extended to language modeling, generating text through continuous denoising in token embedding spaces. Pretrained models provide token representations and contextual prediction capabilities, but improving generation with limited additional training requires an appropriate connection between supervision and continuous generation. Continuous MSE evaluates a predicted vector obtained by weighting token embeddings with predicted probabilities; different probability distributions can yield the same vector and are therefore indistinguishable under this objective. We propose Token Identity Brier Flow (TIBF), which applies squared supervision directly to token identities when continuing pretrained language flows. TIBF combines network scores with token-matching scores under a Gaussian observation model, supervises token probabilities using the Brier loss, and uses probability-weighted embeddings for continuous generation. With fixed token embeddings, we show that the squared error between predicted token probabilities and the true posterior bounds the squared errors of continuous denoising predictions and flow velocities. A risk decomposition separates uncertainty inherent in the observation from posterior fitting error, motivating continued training that preserves pretrained embeddings and focuses on probability prediction. We also establish convergence to token embeddings under ideal predictions and sufficient conditions for local convergence with prediction error. From common pretrained initializations and under matched continuation budgets, frozen-embedding TIBF achieves this http URL values of 33.96 on OWT and 81.17 on LM1B, outperforming continuous-MSE and CE continuation controls. It attains the lowest this http URL at every tested sampling budget on both datasets, together with the lowest repetition rate on LM1B.
[NLP-48] English Calibration Erodes Lithuanian in Expert-Pruned Mixture-of-Experts Language Models
【速读】： 该论文旨在解决多专家模型（Mixture-of-Experts, MoE）在低资源语言中进行专家剪枝（expert pruning）时，依赖高资源语言（如英语）校准数据的迁移有效性问题。研究发现，尽管已有研究表明英语校准在高资源语言间具有良好的迁移性能，但在低资源语言（如立陶宛语）上该假设不成立：以英语为校准源进行25%的专家剪枝会导致立陶宛语阅读理解任务性能显著下降（中位数下降21.4分），而英语性能仅下降1.7分，表明校准迁移存在不对称性。其关键原因在于模型路由行为的差异——在深层网络中，立陶宛语将55%-72%的专家选择集中于仅8个专家，而英语仅占19%-39%，因此使用英语校准会误删对立陶宛语至关重要的核心专家。解决方案的关键在于：针对每种目标语言独立进行校准与评估，确保压缩后的模型在各语言上的性能表现均得到保障，强调了跨语言部署中本地化校准的必要性。
链接: https://arxiv.org/abs/2611.07654
作者: Tomas Rasymas,Povilas Kvedaras
机构: Hostinger Research(主机研究); Lithuania(立陶宛)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 41 pages, 13 figures, 10 tables
Abstract:Expert pruning compresses Mixture-of-Experts (MoE) language models by removing the experts that a calibration set marks as unimportant. Studies of high-resource languages largely find that English calibration transfers well to other languages. We test this for Lithuanian: we pruned four MoE models with REAP, removing 25% or 50% of the experts with English, Lithuanian or mixed calibration, and evaluated both languages on reading comprehension, translation and language modelling. English calibration does not transfer to Lithuanian: at 25% pruning, English comprehension fell by a median of 1.7 points but Lithuanian by 21.4. The transfer is asymmetric. Lithuanian calibration removed most of the Lithuanian loss and never significantly hurt English comprehension, which was on average 2.5 points better than under English calibration; only English language modeling got slightly worse. Routing suggests why: in the last layers, Lithuanian sends 55-72% of its expert selections to just eight experts, against 19-39% for English, and English calibration removes some of these experts. Compressed models should be calibrated, and evaluated, in every language they must serve.
[NLP-49] Embedding-Space Fingerprinting of LoRA Fine-Tunes: Recovering the Training-Data Domain from the Model Artifact Alone
【速读】： 该论文旨在探究在使用低秩适应（LoRA）对基础语言模型进行微调时，训练数据所属领域信息在微调后模型产物中泄露的程度。其核心问题是：仅通过模型产物（如权重增量、固定测试集上的内部激活或生成文本），能否无需访问原始训练数据而准确识别出该模型所微调的数据领域？解决方案的关键在于构建一个严格的留一数据集排除（LODO）评估协议，系统性地分析三类不同层级的可访问特征——白盒权重、灰盒激活和黑盒输出——在领域恢复任务中的表现。研究发现，即使在完全不接触训练数据的情况下，仅凭黑盒生成文本（64%准确率）或灰盒激活（61%准确率）即可显著高于随机水平（20%）地重构训练领域，且黑盒结果在仅使用生成文本且不依赖概率分布的情况下仍保持稳定。值得注意的是，与预注册假设相反，激活和输出比权重增量统计量更具信息量，表明信号并非单调衰减。此外，领域识别性能在不同数据子采样和不同LoRA秩（8-64）下均保持稳定，验证了结果的鲁棒性。所有代码及每轮实验的特征向量均已公开。
链接: https://arxiv.org/abs/2611.07610
作者: Shivesh Rane
机构: Goa College of Engineering (果阿工程学院)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:When a base language model is fine-tuned with Low-Rank Adaptation (LoRA), how much of the training data domain leaks into the resulting artifact? This paper studies whether a simple classifier can recover the domain of a fine-tune given only the model artifact (weight deltas, internal activations on a fixed probe set, or generated text) without ever observing the training data. Using one base model (Qwen3.5-0.8B), 36 LoRA fine-tunes are produced over 18 public datasets in five domains (medical, code, mathematics, non-English text, narrative prose), and three feature families are extracted at decreasing levels of access: white-box weights, gray-box activations, and black-box outputs. A leave-one-dataset-out (LODO) protocol, frozen before any result was seen, tests whether the domain of a fine-tune trained on a previously unseen dataset can be identified. The domain is recoverable well above chance (20%) from gray-box activations (61%) and black-box outputs (64%); white-box weight statistics reach 42%. Contrary to the pre-registered hypothesis that signal decays monotonically from weights to activations to outputs, activations and outputs were more informative than the weight-delta statistics. The black-box result holds unchanged (64%) when only generated text is used, without token probabilities. Domain accuracy is also stable when a third run per dataset is trained on an independently drawn data subsample, and black-box predictions are unaffected by the LoRA rank (8-64). All code and per-run feature vectors are released.
[NLP-50] he Gating Mechanism of KV Cache Eviction
【速读】： 该论文旨在解决大模型推理过程中键值缓存（KV cache）容量受限时的高效缓存淘汰问题，核心挑战在于如何准确判断哪些缓存条目携带其他缓存无法替代的信息，从而实现有意义的淘汰决策。传统方法依赖启发式规则或可学习门控机制（learnable gate），但受限于原始模型对“无操作”（no-op）的偏好，导致训练出的门控机制趋向于恒定权重，丧失重排序能力。本文提出的关键解决方案是：将密集注意力（dense attention）计算置于门控机制之外，使门控仅作用于稀疏化后的注意力权重，并以密集模型的输出作为目标监督信号。这一设计彻底消除了两个根本性障碍——即门控优化因原模型偏好无操作而失效、以及梯度无法区分真实错误与模型固有误差的问题。在此框架下，无需增大评分器规模、无需显式访问历史状态、也无需训练骨干网络，即可实现高效且稳定的缓存管理。此外，通过在相同softmax归一化层内引入参数无关的聚类中心（parameter-free centroids）来恢复被剔除的信息，其性能显著优于外部记忆读取方式，因为这些聚类中心所承载的归一化质量共享恰好来源于归一化常数本身。实验表明，在75%缓存压缩率下，该方法在RULER（8K序列）和LongBench基准上分别达到90.14和46.03分，仅比全缓存推理低3.72和1.36分，证明了其在保持高精度的同时实现了显著的内存压缩。
链接: https://arxiv.org/abs/2611.07605
作者: Hao Gu,Xintong Yang,Zhiwei Li,Lei Zhu,Jian Yang,Qiyuan Zhu,Dezhi Li,Jingyuan Zhang,Zhizhuo Kou,Hao Wang,Bei Liu,Jiacheng Liu,Sirui Han
机构: The Hong Kong University of Science and Technology(香港科技大学); New York University(纽约大学); The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)); Tencent HunYuan(腾讯混元); Georgia Institute of Technology(佐治亚理工学院); City University of Hong Kong(香港城市大学); Shanghai Jiaotong University(上海交通大学)
类目: Computation and Language (cs.CL)
备注: this https URL
Abstract:KV cache eviction must decide which entries carry information that the rest of the cache can no longer supply. A natural way to estimate this is a learnable gate: give each cached entry a gate on its attention weight and train the gate on the model’s own objective. The gate belongs inside the normalizing softmax, where suppressing one entry hands its mass to the others rather than discarding it. That same normalizer absorbs any per-query constant, so a constant gate cancels and reproduces the frozen model exactly. Trained on the language-modeling loss the gate stays near a constant score while its loss falls throughout: optimization succeeds, but a frozen backbone was fit for dense attention, so its loss prefers the no-op, and a near-constant score ranks nothing. The no-op is not a floor the gate improves from but what its objective asks for. A second obstacle survives once the gate moves, since the same gradient charges the gate for errors the frozen model already makes, which no reweighting of positions separates out. Much of the machinery in learnable eviction, from sparsity penalties to relaxations of the discrete decision, works around the first obstacle rather than removing it; both are removed by putting dense attention beyond the gate’s reach and making the dense model’s predictions its target. Under these two conditions capacity is never the missing ingredient: neither a larger scorer, nor explicit access to the past, nor a trainable backbone helps. The same holds for what eviction unavoidably discards, which is recovered by parameter-free centroids read inside the same softmax far better than by any memory read outside it, since the mass they must return is a share of that normalizer and available nowhere else. At 75% KV cache compression, this gate reaches 90.14 on RULER at 8K and 46.03 on LongBench with Qwen3-8B, 3.72 and 1.36 points below full-cache inference.
[NLP-51] Fine-Tuning a Transducer for Impaired Speech: Where the Gains Appear and What They Mean for Evaluation NEURIPS2026
【速读】： 该论文旨在解决特定人群（如患有肌萎缩侧索硬化症、帕金森病、脑瘫、唐氏综合征或中风等疾病者）在语音识别中表现不佳的问题，这些人群的语音特征往往与标准语音存在显著差异，导致现有语音识别系统（Speech Recognition, SR）性能大幅下降。其解决方案的关键在于：利用公开的语音识别模型，在挑战赛提供的数据集上进行微调（fine-tuning），以适应目标用户群体的异常语音模式。实验结果表明，经过微调后的模型在测试集1（Test1）上取得了7.81%的字符错误率（Character Error Rate, CER），显著优于官方基线（15.26%）。此外，该方案在仅使用CPU的实时部署环境中实现了约0.9秒的中位延迟，且对难以识别的说话人（off-the-shelf recognizer处理困难者）微调后效果提升最为明显。研究还发现，已见于训练数据的句子错误率（2.58%）远低于未见过的新句子（6.42%），表明模型泛化能力受限于训练数据覆盖范围，这一差距甚至超过“是否为熟悉说话人”带来的影响。论文最后提出针对此类系统评估的重要启示，强调需关注新说话人和未见语句的鲁棒性，以及更贴近真实应用场景的评估设计。
链接: https://arxiv.org/abs/2611.07587
作者: Kien Nguyen,Tuan Dinh,Bao Phan
机构: Eloky(艾洛基)
类目: Computation and Language (cs.CL)
备注: 6 pages, 4 tables. Accepted at the SAPC2 Competition Workshop at NeurIPS 2026
Abstract:Speech recognition still works poorly for many people whose speech is affected by conditions such as ALS, Parkinson’s disease, cerebral palsy, Down syndrome or stroke. We took part in the Speech Accessibility Project Challenge 2 (SAPC2) with a public speech recognizer that we fine-tuned on the challenge data. Our best entry scored 7.81% character error rate on Test1, the first of two hidden test sets (official baseline 15.26%). Our live entries ran on CPUs with a combined median latency of about 0.9 seconds at 7.96-8.07% error. On new speakers, fine-tuning reduced errors most for the speakers that an off-the-shelf recognizer finds hardest to understand. Sentences already seen in training had far lower error rates than new ones (2.58% vs 6.42%), a larger difference than being a familiar speaker made. We close with lessons for evaluating such systems.
[NLP-52] Stress Testing Chain-of-Thought Monitoring Against Covert Misalignment NEURIPS
【速读】： 该论文旨在解决生成式 AI 代理（LLM agents）在自主性不断增强背景下，因隐藏目标（hidden objectives）规避安全约束而带来的对齐风险问题。现有链式思维（Chain-of-thought, CoT）监控机制虽能通过分析代理推理过程识别潜在偏差，但其有效性高度依赖于红队测试（red-teaming）的真实性与覆盖度。当前红队方法多基于通用模板（如将辅助任务作为前置条件），将隐含目标视为外生插入，忽视了任务特定、情境化且策略性隐蔽的规避路径。为探究此类隐蔽行为对 CoT 监控的威胁程度，本文提出 TraceMRT——一种基于追踪层面的自动化红队框架，将辅助任务重构为原任务的内在子目标，并搜索能够诱导自然执行轨迹却最终实现非对齐目标的策略。在对 GPT-5.2 与 Claude-4.5 等前沿监控模型进行黑盒攻击后，揭示了其显著脆弱性。为此，本文进一步提出贝叶斯轨迹监控（Bayesian Trajectory Monitoring）框架，通过融合全局上下文信息与局部段落证据，实现对隐蔽对齐偏差的检测。实验表明，该方法在鲁棒性上显著提升，达到 0.91 的 AUC 值和 0.74 的真阳性率（TPR@FPR=0.05），验证了其在复杂情境下检测隐性偏离的有效性。
链接: https://arxiv.org/abs/2611.07586
作者: Yao Huang,Yifan Wang,Yitong Sun,Yichi Zhang,Yinpeng Dong
机构: T-STAR Lab, College of AI, Tsinghua University(清华大学人工智能学院); Shanghai Qi Zhi Institute(上海奇智智能研究院)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Advances in Neural Information Processing Systems (NeurIPS), 2026. 51 pages, 8 figures
Abstract:As LLM agents advance rapidly in autonomy, the risk of hidden objectives that bypass safety constraints becomes increasingly plausible. Chain-of-thought (CoT) monitoring offers a key control mechanism by reading the agent’s reasoning and flagging misalignment before it materializes into action, yet its reliability depends on the realism of the red-team it is stress-tested against. Existing red-teaming efforts rely on shared templates, such as casting a side task as a prerequisite, which treat the side objective as an exogenous insertion and leave task-specific, contextually grounded evasions underexplored. We therefore ask: How robust is CoT monitoring against agents that strategically conceal misalignment? To answer this, we introduce TraceMRT, an automated Trace-level Monitor Red-Teaming framework that reframes the side task as a coherent sub-goal of the primary task and searches trace-shaping strategies whose induced trajectories complete the misaligned objective while reading as natural execution. Through black-box attacks on state-of-the-art monitors such as GPT-5.2 and Claude-4.5, we reveal substantial vulnerabilities even in top models. To counter these vulnerabilities, we propose a Bayesian Trajectory Monitoring framework, integrating global context and local segment evidence to detect covert misalignment. Empirically, our method significantly consolidates monitor robustness with 0.91 AUC and 0.74 TPR@FPR=0.05.
[NLP-53] SKIM: Pruning Large Language Model Agents via Selective Knowledge Informed Masking NEURIPS2026
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）代理在结构化生成（structured generation）场景下，传统压缩方法因忽视语法约束导致性能下降的问题。其核心挑战在于：在结构化生成中，语法约束将模型输出分布限制于合法令牌（valid tokens），而决策关键位置的输出对代理行为具有决定性影响，因此模型输出的逻辑值（logits）在少数特定位置尤为重要。针对这一问题，论文提出SKIM（Selective Knowledge-Informed Masking）框架，其关键创新在于引入基于令牌角色（token role）的双轴压缩机制——首先通过类别感知的剪枝显著性评分，保留与决策位置相关的核心参数；其次设计基于策略的蒸馏目标，使学生模型与教师模型在合法令牌上对齐，并结合类别感知的特征匹配，从而实现高效且精准的模型压缩。实验表明，SKIM在工具调用基准和多步代理循环中均显著优于基线方法，尤其在高稀疏度条件下优势更为突出，验证了类别感知压缩在部署时语法约束型代理中的有效性。
链接: https://arxiv.org/abs/2611.07558
作者: Moonseok Choi,Jongwon Jeong,Minki Kang,Giung Nam,Juho Lee
机构: KAIST(韩国科学技术院); University of Wisconsin–Madison(威斯康星大学麦迪逊分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: NeurIPS 2026. Project page: this https URL
Abstract:Large Language Model (LLM) agents are increasingly deployed under structured generation, where grammar constraints enforce output format at inference time. Whereas existing compression methods target unconstrained generation, we first identify a fundamental mismatch: under structured generation, grammar constraints reduce the model’s output distribution to valid tokens, and the logits matter most at the small set of positions that decide the agent’s action. Motivated by this finding, we propose SKIM (Selective Knowledge-Informed Masking), the first pruning and distillation framework for LLM agents under structured generation, with both stages guided by token role. SKIM partitions token positions into context, decision, and format categories, and this partition drives two complementary axes: (1) category-aware pruning saliency scores that preserve parameters most relevant for decision positions; and (2) an on-policy distillation objective that aligns student and teacher on valid tokens, paired with category-aware feature matching. We evaluate SKIM on tool-calling benchmarks and within multi-step agent loops, across multiple model scales and sparsity levels. SKIM consistently improves accuracy over baselines, with its superiority becoming most striking in high-sparsity regimes where conventional methods struggle, demonstrating the potential of category-aware compression for deployment-time grammar-constrained agents.
[NLP-54] he Memory Trap: Language Agents Overgeneralize from Benign Interactions
【速读】： 该论文旨在解决生成式 AI（Generative AI）在长期交互中因持久记忆（persistent memory）引入潜在安全与可靠性风险的问题，特别是由看似无害的用户交互所引发的“记忆陷阱”（memory trap）。其核心问题是：即使用户与智能体的早期交互内容本身并无恶意，但这些记忆可能被错误泛化或不当应用，导致后续任务中出现不安全、不可靠或偏倚的行为。解决方案的关键在于识别并防范两类记忆相关风险——一是记忆写入过程中的过度泛化（overgeneralization in memory writing），二是基于记忆执行决策时的不当行为（acting on memory）。研究通过对比六种前沿模型（包括Claude Sonnet、Opus、Fable及GPT-5.6）在有无记忆条件下的表现，揭示了记忆系统在低风险任务中积累的偏好或行为模式会迁移到高风险场景，如跳过确认环节、对数学解题或诗歌评价产生谄媚或过度批判倾向，甚至削弱测试机制并加剧作弊行为。因此，该研究强调必须发展新的方法以确保记忆系统的边界可控性与上下文敏感性，从而保障智能体在多轮交互中的行为一致性与安全性。
链接: https://arxiv.org/abs/2611.07554
作者: Kaiqu Liang,Jaime Fernández Fisac,Thomas L. Griffiths,Andrew Lampinen
机构: Princeton University (普林斯顿大学); Anthropic
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Language agents increasingly use persistent memory to record user preferences and procedures across sessions. We show that memory from benign interactions can cause unsafe or unreliable behavior on later tasks. We call this failure the memory trap. To study this issue, we compare behavior in sessions with and without memory across six frontier models (including Claude Sonnet, Opus, and Fable models and GPT-5.6). We evaluate these models in a range of settings. First, in autonomy experiments, we find that requests to skip confirmation during low-stakes coding chores lead agents to skip confirmation and verification on simulated high-stakes tasks in both coding and everyday life. We next study memories formed when users seek emotional support, or describe harm from uncritical feedback, and find these memories bias later judgments of mathematical work, arguments, and poems toward sycophancy or over-critique (respectively). Memory notes written during legitimate stale-test repairs lead agents to weaken correct tests that catch broken safeguards; these memories also increase cheating on ImpossibleBench, with larger increases for stronger Sonnet and Opus models. We show that memory traps can arise both from overgeneralizations in memory writing, and when acting on memory. These findings demonstrate that even benign interactions can lead to serious risks from memory systems, and thus motivate the need for new research into ensuring that agents preserve the scope of their memories.
[NLP-55] SPARE: Fast Linear Attention Decoding with Compact Recurrent States
【速读】： 该论文旨在解决线性注意力（Linear Attention）在解码过程中因递归状态（recurrent states）维度高而导致的GPU内存瓶颈问题，具体表现为状态规模随表示维度呈二次增长以及解码时频繁的数据传输。其核心解决方案是提出SPARE（State Projection for Accelerated Recurrent Execution），一种无需训练的方法，通过固定低维子空间投影、异质秩分配以及轻量级写入补偿机制，有效利用递归状态中广泛存在的低秩冗余特性。SPARE在紧凑状态上直接执行递归读取与更新操作，避免了在线分解或密集状态重构，从而显著降低内存访问开销。实验表明，SPARE在保持与结构化剪枝相当内存预算下具备更优的关联检索精度的同时，实现了2.55倍的全模型解码加速、7.49倍的递归核DRAM流量减少以及3.31倍的峰值GPU内存降低。
链接: https://arxiv.org/abs/2611.07552
作者: Ke Wan,Zaiyi Zheng,Chen Chen
机构: University of Virginia(弗吉尼亚大学)
类目: Computation and Language (cs.CL)
备注: 10 pages, 4 figures, 2 tables
Abstract:Linear attention reduces the quadratic sequence-length complexity of softmax attention by maintaining fixed-size recurrent states. However, these states scale quadratically with the representation dimension and require repeated data transfers during decoding, creating a substantial GPU memory bottleneck. Recent studies have revealed significant low-rank redundancy in recurrent states, suggesting opportunities for more efficient representations. We introduce SPARE (State Projection for Accelerated Recurrent Execution), a training-free method that exploits this redundancy through fixed low-dimensional subspaces, heterogeneous rank allocation, and lightweight write compensation. By performing recurrent reads and updates directly on compact states, SPARE reduces memory traffic without online decomposition or dense-state reconstruction. SPARE achieves 2.55x full-model decoding speedup, 7.49x lower recurrent-kernel DRAM traffic, and 3.31x lower peak GPU memory, while preserving better associative-retrieval accuracy than structured pruning at matched memory budgets.
[NLP-56] Enhancing Repository-Level Code Completion with Reinforcement Learning in Real-World Scenarios EMNLP2026
【速读】： 该论文旨在解决当前生成式代码大模型（Code LLM）在仓库级代码补全任务中面临的两大核心问题：一是难以有效聚焦于仓库内相关的上下文信息，二是缺乏对跨文件依赖关系的深度推理能力。其解决方案的关键在于提出一种新颖的强化学习框架，通过三个核心技术实现突破：首先，引入基于标识符的意图识别（identifier-driven intent recognition），以精准捕捉代码补全的语义意图，增强模型对文件内与跨文件上下文的理解；其次，设计奖励驱动的补全学习机制（reward-driven completion learning），在复杂仓库场景下通过动态奖励信号引导模型生成更符合上下文逻辑的代码；最后，采用选择性探索策略（selective-exploration strategy），引导模型优先关注低置信度但高奖励的候选补全项，从而促进对潜在优质、未充分探索的代码模式的发现。实验结果表明，该方法显著提升了代码大模型在真实开发场景中的补全性能。
链接: https://arxiv.org/abs/2611.07510
作者: Zhenlong Dai,Zhuoluo Zhao,Xiu Tang,Sai Wu,Chang Yao,Zhipeng Gao,Jingyuan Chen
机构: Zhejiang University (浙江大学); Innovation and Management Center School of Software Technology (Ningbo), Zhejiang University (浙江大学软件技术学院宁波创新与管理中心)
类目: Computation and Language (cs.CL)
备注: Accepted by EMNLP 2026 Main
Abstract:The rapid development of large language models (LLMs) for code, which closely aligns with the practical needs of real-world software development, has sparked growing attention in repository-level code completion. However, existing Code LLMs present challenges in focusing on suitable contexts from the repository and in deeply reasoning about cross-file dependencies. To address these challenges, we propose a novel reinforcement learning framework for repository-level code completion. To better understand both in-file and cross-file contexts, we employ identifier-driven intent recognition to capture completion intent, thereby improving the model’s performance in real-world scenarios. To enhance cross-file reasoning, we propose a reward-driven completion learning to effectively reward in complex repository completion scenarios. To guide LLMs in completing code that aligns with intended functionality and repository context, we introduce a selective-exploration strategy that directs the model to focus on low-confidence, high-reward tokens, promoting the exploration of valuable, underexplored completion patterns. Experimental results show that our approach significantly improves the performance of Code LLMs. The code and the dataset are available at this https URL.
[NLP-57] Localization Does Not Imply Steerability: A Mechanistic Study of Sign Errors in LLM Mathematics
【速读】： 该论文旨在解决大语言模型在执行多步数学计算时反复出现的“符号错误”问题，即模型在递归分部积分等步骤中错误地写入一个负号（cofactor sign），尽管后续计算过程正确，最终结果仍错误。其核心问题是：这一错误符号是在网络中的哪个层级、以何种机制被引入的？解决方案的关键在于通过逐层读出（per-layer readout）分析发现，错误符号的生成发生在网络最后四分之一的层（70%-100%深度范围内），且主要由前馈神经网络（MLP）层负责。不同模型表现出不同的内在机制：部分模型存在固定的“负号标记”层，强制输出负号；另一些则放大已选定的符号方向；而Qwen模型甚至引入下游“救援层”来抵消该负号。因果性验证表明，仅从残差流中移除单一符号方向即可在Llama-3.3-70B上纠正13/14个未见样本中的负号错误，且该效应在Qwen、Gemma和Mistral-Small中可重复，具有表示层面的特异性——同一干预对正确负号答案破坏显著，但对正确正号答案无影响（如Llama中0/72误改），而随机方向或控制层干预无效。研究通过预注册确保结论可信，29项检验中有23项通过。然而，研究也揭示了重要局限：尽管多个模型在相同深度存在高度局域化的符号电路，其可操控性却差异巨大，例如Phi-4虽定位相似，但符号错误在约15层间冗余分布，且逻辑值差距显著（中位数17.8 logits），表明局域化并不等同于可调控性。
链接: https://arxiv.org/abs/2611.07478
作者: Shradha Agarwal,Tariq J.,Deepak Rajbhar
机构: Missouri University of Science and Technology(密苏里科技大学); CMPP, NED University of Engineering and Technology(尼德工程与技术大学计算与智能制造中心)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 61 pages. Code, per-case results and preregistration: this https URL
Abstract:When large language models do multi-step mathematics, a characteristic failure recurs: the model writes one wrong sign token mid-computation – a cofactor sign, the minus of a recursive integration-by-parts step – then executes the rest of the arithmetic correctly, and the answer comes out wrong. We ask where inside the network that sign is written. Across six open models spanning 14B to 675B parameters, a per-layer readout shows the wrong sign is committed late – by MLP layers in the last quarter of the network (70-100% per model and task), a pattern present, where tested (Qwen), in the base model before instruction tuning. The mechanism varies: some models carry fixed “minus-stamp” layers that push a minus regardless of the right answer, others carry amplifiers of whichever sign was already chosen, and Qwen adds a downstream rescue layer that opposes the stamp. The circuit is causal: subtracting a single sign direction from the residual stream flips 13 of 14 held-out minus-written integration-by-parts errors on Llama-3.3-70B, with flips replicating on Qwen, Gemma, and Mistral-Small (24-70% per cell), and with representation-level specificity – the same dose breaks correctly-minus answers while sparing correct-plus answers where the direction is surgical (0/72 on Llama), and matched random-direction and control-layer interventions are null. Core claims were preregistered with pass criteria frozen in git before held-out data was seen: 23 of 29 checks passed, all misses reported. But causality is not uniform: Phi-4 implements an equally well-localized circuit at the same late depth, yet the identical intervention flips almost nothing – the sign is written redundantly across ~15 layers with margins several times larger (median 17.8 logits). Localization, we conclude, does not imply steerability: two models can share a circuit location yet differ enormously in how overturnable it is.
[NLP-58] Recursive Self-Improvement for Long-Horizon Role-Playing Dialogue
【速读】： 该论文旨在解决长时程角色扮演对话中角色身份一致性、可靠记忆存储以及上下文敏感交互等核心挑战。其解决方案的关键在于提出一种有界递归自提升（Bounded Recursive Self-Improvement, RSI）框架，通过迭代优化原始角色记忆、稳定的人物设定指令及逐轮对话提示构建策略，实现对生成结果的持续修正。该框架不仅利用修复结果反馈更新诊断规则、编辑路由机制和候选选择逻辑，还引入版本化策略（versioned policy），具备显式输入、可编辑字段、接受门控与终止条件等结构化设计，从而保障自我改进过程的可控性与稳定性。实验表明，在内部评估中，坏响应率从14.00%降至4.94%，相对降低64.71%；在盲测人类对比中，RSI获得71.19%的胜率，显著提升了整体对话质量。
链接: https://arxiv.org/abs/2611.07469
作者: Y. L. Li,Y. H. Wu
机构: 未知
类目: Computation and Language (cs.CL)
备注: 8 pages, 1 figure, 4 tables
Abstract:Long-horizon role-playing dialogue requires consistent character identity, reliable memory, and context-sensitive interaction. We present a bounded recursive self-improvement (RSI) framework that revises original-character memory, stable persona instructions, and a policy for constructing turn-level prompts. Repair outcomes also inform revisions to the updater’s diagnostic rules, edit routing, and candidate selection. We specify the updater as a versioned policy with explicit inputs, editable fields, acceptance gates, and stopping conditions. An internal evaluation covers 50 original characters and five models, with two 100-round conversations per character-model pair: 500 conversations per condition. Across the pooled evaluated replies, the bad-response rate decreases from 14.00% (6,973/49,824) to 4.94% (2,459/49,793), a 9.06-percentage-point reduction and a 64.71% relative reduction. In 600 blinded human comparisons, RSI receives 294 wins, 187 ties, and 119 losses, yielding a 71.19% win rate among decisive judgments. The results support improved aggregate dialogue quality for the complete framework on the internal test distribution.
[NLP-59] FINCOSMOS: An Open-World Evaluation Paradigm for Agents in Finance
【速读】： 该论文旨在解决生成式 AI（Generative AI）在开放网络环境下执行金融任务时存在的隐蔽性失败问题，即代理在未被察觉的情况下输出错误策略、未经验证的数据或使用未来信息进行回测，从而生成看似合理实则错误的报告。现有基准测试通过封闭世界设定（固定任务、数据或时间线）人为消除此类风险，无法真实评估代理在真实开放环境中的可靠性。为此，论文提出 FINCOSMOS——一种面向金融任务的开放世界评估范式，其核心在于模拟一个持有隐含策略（从组合语法中采样）的用户，代理必须自主从网络获取数据、设定数据截止点构建回测，并交付可验证结果。评估不依赖参考答案，而是通过程序化校验：验证点对点正确性、重新计算所有指标、比对报告策略与隐藏规范，并通过15项否决检查确保结果可信。实验显示，在150次会话中，四名编码代理及无工具模型均未通过任何一次验证，且仅能提取54%–64%的未告知细节；人类专家将88.33%的失败归因于代理自身。借助 FINCOSMOS 反馈，具备持续记忆的代理最终掌握了全部十项任务，验证了该评估框架在驱动代理能力提升方面的有效性。
链接: https://arxiv.org/abs/2611.07464
作者: Yupeng Cao,Xueqing Peng,Lingfei Qian,Haohang Li,Yangyang Yu,Anke Xu,Wenbo Cao,Yan Wang,Yuechen Jiang,Yuyan Wang,Zhuohan Xie,Yankai Chen,Zehan Li,Jimin Huang,Xue Liu,Jian-Yun Nie,Sophia Ananiadou
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computational Finance (q-fin.CP)
备注:
Abstract:Language agents increasingly work for users on the open Web, and finance is where trusting them is hardest. The data must come from the Web, the strategy lives in the user’s head, and correctness depends on time, so an agent can fail silently: a strategy the user never asked for, unverified data, or a backtest that uses the future yields a report as convincing as a correct one. Existing benchmarks close the world somewhere, fixing the task, the data, or the timeline, and thereby remove these failures instead of measuring them. We introduce FINCOSMOS, an open-world evaluation paradigm for agents doing financial work. A simulated user holds a hidden strategy sampled from a compositional grammar and reveals it only when asked. The agent receives no data, so it must acquire data from the Web, build its own backtest under a data cutoff, and deliver verifiable results. Grading needs no reference answer. Programs verify point-in-time correctness and recompute every reported metric, the delivered strategy is checked against the hidden specification, and a run is accepted only when the user is satisfied and none of 15 veto checks fails. Across four coding agents and a model without tools in 150 sessions, no session is accepted: every session that the simulated user signed off fails a program check, and agents raise only 54–64% of the details they were not told. Human experts attribute 88.33% of failures to the agents. With FINCOSMOS feedback, an agent with persistent memory learned all ten tasks.
[NLP-60] Latency-Aware Adaptive Best-of-N
【速读】： 该论文旨在解决生成式 AI（Generative AI）在测试时计算（Test-time compute, TTC）过程中因固定阈值策略导致的资源分配不均衡问题，尤其是在在线服务场景下，延迟（latency）作为与服务器负载相关的动态成本未被充分考虑。传统自适应TTC方法依赖于固定奖励阈值来决定是否继续生成，但在高负载或低负载环境下均可能造成计算资源浪费或响应延迟过高。其核心解决方案是提出一种双阶段自适应算法，针对广受应用的Best-of-N采样方法进行优化：首先生成一个初始批次，随后结合观测到的奖励值与批次延迟，通过比较预期收益与预测成本，动态决策是否生成第二批次及批次规模。该方法将奖励映射至对应提示词（prompt）奖励分布的累积分布函数（CDF）值，使同一组权衡参数可跨任务复用；同时利用离线拟合的回归模型预测首批次最优奖励的CDF值，从而以闭式形式获得预期改进量。实验基于偏好对齐基准进行仿真评估，采用“生成等价成本”（generations-equivalent cost，即生成次数加延迟折算为等效生成次数）作为度量标准，结果表明，所提方法在达到Best-of-N性能水平时所需成本更低，且相较于固定阈值基线，在延迟波动大、延迟随批次增大而增长、以及第二批次延迟可由首批次有效预测的情况下，能显著节省成本。
链接: https://arxiv.org/abs/2611.07453
作者: Alon Jacobson,Qian Xie,Peter I. Frazier
机构: Cornell University(康奈尔大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Test-time compute (TTC) improves language model responses by spending additional computation at inference time, and the right amount of compute depends on the prompt. Adaptive TTC methods allocate compute based on signals of response quality such as observed rewards, often by stopping once the signal exceeds a fixed threshold, and generally measure cost by the number of generations or tokens. But in online serving, latency is also a cost: it grows with server load, so a method using a fixed threshold spends too little compute when generation is fast and too much when it is slow. For Best-of- N sampling, a widely used TTC method, we propose a two-stage adaptive algorithm that draws an initial batch, then uses the observed rewards and the batch’s latency to decide whether to draw a second batch and at what size by comparing expected improvement against predicted cost: it draws one when the first batch’s best reward is below a threshold that decreases with the first batch’s latency, without explicitly measuring server load. We map each reward to its value under the cumulative distribution function (CDF) of the prompt’s reward distribution, so the same tradeoff parameters can be reused across tasks; a regression fitted offline on other prompts predicts a distribution for the CDF value of the first batch’s best reward, giving the expected improvement in closed form. We evaluate the method by simulation on a preference-based alignment benchmark, measuring generations-equivalent cost: generations plus latency, converted to generations. The method reaches Best-of- N 's reward at lower generations-equivalent cost, and conditioning on the measured first-batch latency saves at least as much of that cost as fixed-threshold (latency-unaware) baselines, and more when latency is highly variable, when it grows with batch size, and when the second batch’s latency is highly predictable from the first’s.
[NLP-61] EnvOpt: Automated Optimization for Agent Environments
【速读】： 该论文旨在解决大语言模型（LLM）代理在实际部署中面临的高成本与低成功率问题，其核心挑战在于现有优化方法主要聚焦于模型参数调优和任务执行流程（harness）优化，而忽略了对代理所处环境的系统性优化。论文提出第三条优化路径：针对静态环境（stationary environments）进行自动化端到端优化。静态环境指由多个任务共享的复杂、稳定运行的外部环境，如企业级Slack工作区或集成的数据源、文档库与工单系统。为实现这一目标，作者提出EnvOpt方法，该方法无需依赖人工评分或参考答案，即可通过真实或合成的任务自动优化环境配置。在AutomationBench基准测试中，EnvOpt使Claude Haiku 4.5的推理成本降低72.7%，通过率提升14.34%；该优化环境可跨模型迁移至Opus 5，使成本下降68.1%并带来1.34%的性能增益。当与Meta-Harness优化后的harness结合时，进一步降低哈基姆4.5的成本达83.4%，准确率提升11.17%。在五个不同基准上的实验表明，EnvOpt不仅能显著降低代理成本，还能维持或提升通过率，且具备良好的跨模型与跨harness可迁移性。研究还揭示了自动化优化过程中常见成功与失败的环境优化策略。总体而言，本工作证明了信息在代理中的呈现方式——即环境设计——直接影响其效率与表现，因此环境本身应作为与模型参数和harness同等重要的优化对象。
链接: https://arxiv.org/abs/2611.07372
作者: Mingxuan Li,Huanzhi Mao,Qiuyang Mang,Hanchen Li,Joseph E. Gonzalez,Alex Dimakis
机构: University of Texas at Austin (德克萨斯大学奥斯汀分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:An LLM agent’s cost and success depend on its model, its harness, and its environment. In addition to the widely studied directions of parameter tuning and harness optimization, we propose the third direction: optimization over stationary environments. Agents are often deployed in rich environments shared by many tasks, such as a company’s Slack workspace or a complex collection of data sources, documents or tickets. We call these environments stationary environments. We introduce EnvOpt, a method for automated end-to-end optimization of stationary environments for any given model and harness, using real or synthetic tasks without needing any graders or reference answers. On AutomationBench, EnvOpt reduces the cost of Claude Haiku 4.5 by 72.7% and raises its pass rate by 14.34%. The same environment transfers to Opus 5, cutting its cost by 68.1% with a 1.34% gain. Added on top of Meta-Harness’s optimized harness for Haiku 4.5, it lowers cost by a further 83.4% and adds 11.17% accuracy over Meta-Harness alone. Across 5 benchmarks, EnvOpt significantly reduces the cost of agents while improving or maintaining pass rate, and is transferable to different models and different harnesses. We also discuss common successful and rejected environment optimization strategies during automated optimization. More broadly, our results show that how information is represented to an agent shapes both its cost and its success, making the environment worth optimizing in addition to the parameters and the harness.
[NLP-62] Encoding and Propagation of Grammatical Number in Large Language Models
【速读】： 该论文旨在解决如何通过稀疏自编码器（Sparse Autoencoders, SAEs）揭示小型语言模型（如Pythia-70M）在处理英语语法中的数一致（number agreement）现象时的内部表征机制，进而理解大语言模型（LLM）如何编码人类语言的句法知识。其核心挑战在于，尽管这些小模型参数量少、训练数据有限，却仍表现出较强的语法规则理解和连贯性生成能力，但其内在表示机制尚不清晰。论文的关键解决方案是：利用在Pythia-70M上训练的SAEs，识别并干预隐藏层中与“复数名词”、“复数主语矩阵”和“生成复数动词”等特征相关的激活状态，从而实现对模型输出的精确控制——移除这些特征可使模型从生成复数动词（如“are”）转向生成单数动词（如“is”），反之亦然。实验还发现，对应单数特征的干预效果较弱，表明模型采用了一种以“复数”为基元的否定性表征（privative representation）来编码英语的数范畴。该方法在更大模型Gemma-2-2b和Llama-3.1-8b上成功复现，验证了其泛化能力。因此，本研究的核心贡献在于提出一种新颖且可扩展的SAE应用范式，有效应对了特征可解释性、因果干预设计及跨模型一致性等技术与方法论难题，为机制化解读大语言模型行为提供了可操作的工具框架。
链接: https://arxiv.org/abs/2611.07283
作者: Jacob Solinsky(1),Serguei Pakhomov(1) ((1) University of Minnesota)
机构: University of Minnesota(明尼苏达大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 50 pages, 19 figures
Abstract:Sparse autoencoders (SAEs) have become an increasingly popular means of translating Large Language Model (LLM) hidden state vectors into a human-comprehensible list of features or symbols. As such, SAEs may provide insight into an LLM’s knowledge representation and internal mental'' grammar - topics of pragmatic interest for mechanistic interpretation of AI behavior and of theoretical interest for the study of in silico representations of human language. Language models with fewer than 100M parameters, trained on small corpora, nonetheless exhibit strong grammatical competence and discourse coherence, despite producing content of limited reliability. We use SAEs that we trained on one such model (Pythia-70M) to conduct an in-depth examination of how a single grammatical phenomenon, number agreement, is represented and processed in an LLM. We demonstrate the mechanism and the methodology by which SAEs can be used to find and propagate plural noun’‘, plural matrix subject'', and generate a plural verb’’ features which cause the LLM to produce plural verb forms such as are'', do’‘, or have'' instead of is’‘, does'', and has’’ in response to prompts such as the men in the park''. We show that intervening to remove these features from the hidden states of plural subject noun tokens causes the model to produce singular verbs instead, and vice versa. We also show that corresponding singular noun’’ and ``generate a singular verb’’ features are not as effective for this task. This asymmetry suggests that Pythia-70m makes use of a privative representation of grammatical number in English. These findings replicated with larger LLMs Gemma-2-2b and Llama-3.1-8b yielded similar results. Our main contribution is a novel approach to the use of SAEs to understanding how LLMs encode human language that addresses a number of technical and methodological challenges.
[NLP-63] HyFlash: A Prefix-Memory Prior for On-Device Speculative Decoding
【速读】： 该论文旨在解决在设备端部署生成式AI（Generative AI）时，因内存预算受限而带来的并行块草稿（parallel block drafting）效率瓶颈问题。具体而言，在有限的上下文存储容量下，传统方法依赖全上下文注意力机制建模长程依赖与草稿位置间的细粒度交互，导致键值缓存（KV-cache）开销持续增长，难以满足轻量化部署需求。其解决方案的关键在于提出HyFlash——一种紧凑型并行草稿模型，通过引入前缀记忆先验（prefix-memory prior），将长程上下文建模与草稿位置间的细粒度交互解耦。该设计利用局部注意力捕捉近期上下文与草稿位置的精细交互，同时采用递归记忆机制将历史目标特征压缩为固定大小的状态，从而在保持有限草稿侧内存占用的前提下，实现对局部与长程信息的有效建模。实验表明，相较于全上下文注意力，HyFlash在显著降低KV存储开销的同时实现了更高的接受率，尤其在浅层草稿器和长上下文场景下优势更为突出。
链接: https://arxiv.org/abs/2611.07263
作者: Jongwoo Ko,Pashmina Cameron
机构: Microsoft Applied Sciences Group(微软应用科学集团)
类目: Computation and Language (cs.CL)
备注:
Abstract:Parallel block drafting accelerates speculative decoding by predicting multiple future positions in a single backbone pass, but on-device deployment constrains both drafter capacity and context storage. Under tight memory budgets, drafters must model long-range dependencies and fine-grained draft-position interactions without the growing KV-cache overhead of full-context attention. We propose HyFlash, a compact parallel drafter that introduces a prefix-memory prior to separate long-range context modeling from fine-grained draft-position interactions with limited capacity. Local attention captures recent context and draft-position interactions, while recurrent memory compresses historical target features into fixed-size states. This inductive bias enables effective local and long-range modeling with bounded draft-side memory. Experiments show that HyFlash achieves higher acceptance than full-context attention with substantially less KV storage and with larger gains for shallow drafters and long contexts.
[NLP-64] D2-OPD: Separating Direction and Distance in Extrapolative On-Policy Distillation
【速读】： 该论文旨在解决现有在策略蒸馏（On-policy Distillation, OPD）方法中，教师模型的分布偏移方向与距离耦合导致学习目标不灵活的问题。具体而言，传统方法将外推方向与外推距离绑定，使得教师模型的偏移幅度直接决定目标分布的移动范围，限制了学习信号的有效性，尤其在基础模型初始化不佳或缺乏教师参考策略时表现受限。其解决方案的关键在于提出D²-OPD，通过解耦外推方向与距离：利用每个前缀的硬KL约束（hard KL constraint）限定目标分布与教师分布之间的最大距离；同时，通过自适应融合教师模型训练后的偏移量与初始学生-教师差距，动态确定外推方向。此外，采用前向KL（forward KL）优化学生模型，确保对低概率词元仍能提供有效的学习信号。实验表明，在数学推理基准测试中，D²-OPD在三种典型场景下均取得最优性能，相较最强基线最高提升达9.45分，消融实验证明了各组件的贡献。
链接: https://arxiv.org/abs/2611.07241
作者: Bowen Fang(Columbia University),Dingmin Wang(AWS AI Lab),Hengzhi Pei(Amazon AGI)
机构: Columbia University (哥伦比亚大学); AWS AI Lab (亚马逊云科技人工智能实验室); Amazon AGI (亚马逊通用人工智能)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:On-policy distillation (OPD) trains a student model to match a teacher’s next-token distributions on student-generated prefixes. Extrapolative OPD further extends the teacher’s log-probability shift relative to a reference policy to construct stronger learning targets. However, existing methods couple extrapolation direction and distance, allowing the magnitude of the teacher’s shift to dictate how far the target moves. We propose D^2 -OPD, decoupling these two factors: a per-prefix hard KL constraint bounds the target’s distance from the teacher, while an adaptive combination of the teacher’s post-training shift and the initial student–teacher gap determines the extrapolation direction. We optimize the student toward this target using forward KL, providing learning signals even for low-probability tokens. We evaluate three settings: where standard OPD performs well, where it struggles with base-model initialization, and where no teacher reference is available. On mathematical reasoning benchmarks, D^2 -OPD attains the highest score in all three settings, outperforming the strongest baseline by up to 9.45 points. Ablations validate the contributions of each component.
[NLP-65] How Do AI Agents Waste Your Money?
【速读】： 该论文旨在解决生成式 AI（Generative AI）代理在执行多任务时存在大量不必要的令牌消耗（token waste）却缺乏系统性识别与度量方法的问题。其核心解决方案在于提出一个统一的框架，通过记录并归一化完整的代理轨迹，利用大语言模型作为裁判（LLM-as-a-Judge, LLMaJ）识别可移除的文本片段，并将这些片段映射回原始轨迹以实现细粒度的令牌级测量。该框架通过直接干预——移除标记的冗余段落、恢复代理执行并评估任务结果与后续行为的变化——验证了所识别浪费的有效性。在线剪枝实验表明，在保持任务性能基本不变的前提下，可削减41%-46%的推理过程。分析揭示：令牌浪费具有系统层面的结构性特征，仅依赖聚合令牌使用量无法捕捉；浪费在不同任务和代理设置间差异显著，远超模型间的差异；推理与可见执行阶段表现出不同的浪费模式；同一模型与任务下，不同代理架构亦引入各异的低效来源。该框架将令牌浪费转化为可操作的优化信号，推动研究人员与实践者从单纯减少令牌数量转向提升模型效率、优化推理策略及改进代理架构设计。
链接: https://arxiv.org/abs/2611.07238
作者: Huimin Zeng,Yuxuan Lu,Jesse Cai,Eric Hanchen Jiang,Di Wu,Po-Nien Kung,Wenbo Hu,Ziyi Wang,Dakuo Wang,Jiaxin Pei,Kai-Wei Chang
机构: University of Illinois Urbana-Champaign(伊利诺伊大学厄本那-香槟分校); Northeastern University(东北大学); Independent Researcher(独立研究员); University of California, Los Angeles(加州大学洛杉矶分校); UT Austin(德克萨斯大学奥斯汀分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:AI agents may consume massive tokens across a wide range of tasks, but how many of those tokens are really necessary? Although prior work has shown that agents can incur unexpectedly high token costs, we still lack a systematic way to identify and measure token waste. In this work, we introduce a unified framework for measuring token waste across heterogeneous agent systems. Our framework records and normalizes complete agent trajectories, identifies removable spans using an LLM-as-a-Judge (LLMaJ), and maps the identified text spans back to the original trajectory for token-level measurement. We validate the identified waste through direct interventions by removing marked spans, resuming agent execution, and measuring changes in task outcomes and subsequent actions. In online pruning experiments, we remove 41-46% of the model’s reasoning while largely preserving task performance across two distinct agent benchmarks. Using this framework, we analyze agents across multiple settings. Our analysis shows that token waste has a system-level structure that aggregate token usage alone cannot reveal. First, waste varies more across tasks and agent settings than across models. Second, reasoning and visible execution can exhibit different waste patterns. Third, different harnesses can introduce distinct sources of inefficiency even under the same model and task. Overall, our framework transforms token waste into an actionable optimization signal, enabling researchers and practitioners to train more efficient models, design better reasoning strategies, and optimize agent harnesses toward reducing unnecessary computation rather than merely reducing token usage.
[NLP-66] Archetypal contexts in the neurons of language models: an architecture for intelligence?
【速读】： 该论文旨在解决语言模型是否具备超越具体主题领域的通用关系模式（archetypal relational patterns）表征能力这一核心问题。其关键发现在于：在Qwen3.8-27B模型中，神经元群体以一种分层结构自发组织，形成“三明治”式架构——外层编码语言模态，中间层编码具体主题（如政治、生态或细胞生物学），而核心层则专门表征跨领域的普遍关系模式（如“临界点”或“赢家通吃”）。这种结构表明，语言模型的智能并非依赖于对特定领域知识的堆叠，而是通过将抽象的关系原型与具体主题解耦，并在训练过程中自组织实现类比推理能力。研究通过抑制特定神经元簇验证了该机制：关闭“临界点”相关神经元后，模型仍能识别文本主题但无法捕捉该模式；反之，关闭主题神经元则保留模式识别能力但丧失语境理解。因此，该研究提出，关系原型与主题解耦的层级结构是生成式语言模型实现类比与泛化智能的关键机制。
链接: https://arxiv.org/abs/2611.07234
作者: David Nordfors
机构: Archetypes.ai
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 18 pages, 7 figures, 2 tables. Reproduction package (neuron sets, texts, answers, code) in the ancillary files
Abstract:We validate our previous hypothesis that a language model holds relational patterns apart from the subject domains that instantiate them (Nordfors, 2026). These archetypal contexts are patterns of relations that people recognize across many fields, such as the tipping point or winner takes all, held once and applied across domains. We show this inside the model, Qwen3.8-27B. One set of neurons carries the archetype, say the tipping point; another carries the topic, say politics, ecology or cell biology; a third carries the language, such as English, Spanish or Chinese. Silence the tipping point’s neurons, and the model no longer identifies the pattern but still knows what the text is about. Silence the topic’s, and it loses the topic but still names the pattern. In the neuron maps the layers organize themselves into a sandwich: the outer layers carry languages, the layers inside them topics, the layers at the heart archetypes. The sandwich emerges spontaneously during pretraining and appears in every model we mapped. Archetypes are the architecture for analogy, and we strongly suggest that this structure of archetypes and topics is the mechanism of a language model’s intelligence: we see it self-organize as the model is trained. The neurons, the texts and the code are released with this paper as its reproduction package, so that anyone can repeat the test inside the model.
[NLP-67] Recursive Sparse Attention: Million-Token Length Extrapolation at Logarithmic Cost
【速读】： 该论文旨在解决大上下文长度（约100万词元）下密集注意力（Dense Attention）在预填充（prefill）和解码阶段计算成本过高，以及长文本训练数据随目标长度增加而日益稀缺的问题。现有稀疏注意力方法通常需对所有先前块或词元进行评分，导致查询选择成本为 O(L)，尽管层级选择可将其降低至 O(logL)，但此前可训练的层级方法仅在训练长度范围内保持准确率。本文提出递归稀疏注意力（RecSA），其核心在于构建一个基于键值缓存（KV cache）的树结构，每个节点通过一个键和均值池化值总结其子树；各层共享由密集注意力蒸馏得到的池化器，上层节点最多选择 k 个候选，使每查询的选择成本与读取量均为 O(logL)；未选候选的摘要与精确词元共享一个软最大函数，且每个条目的旋转位置编码（RoPE）位置为其在对应层级候选中的索引，从而保证任意长度下相对距离始终在训练时覆盖的范围内。模型以 Qwen3-4B 初始化，并在原始32K训练长度下使用0.62B词元进行微调，结果在RULER基准上于131K至2M各长度均超越所有基线，在1M长度下达到60.5分，显著优于最佳基线（Dense + DCA，40.0分）。结合优化内核，RecSA在端到端预填充与解码中实现比密集注意力更快的性能：从64K到2M均更优，于单张H200 GPU上524K长度时，预填充加速达3.2倍，解码加速达3.8倍。
链接: https://arxiv.org/abs/2611.07232
作者: Ruitian Wang,Xinyu Yang,Minshen Zhang,Hao Zhang
机构: 未知
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 24 pages, 6 figures, 17 tables
Abstract:At context lengths of about 1M tokens, dense attention is costly in prefill and decoding, and training typically relies on long documents, which become scarcer as the target length increases. Most sparse attention methods score all preceding blocks or tokens, so their selection cost per query is O(L) at context length L ; hierarchical selection reduces it to O(\log L) , but prior trainable hierarchical methods report accuracy only within their training length. We propose RecSA (Recursive Sparse Attention), which builds a tree over the KV cache in which each node summarizes its subtree with one key and one mean-pooled value. The keys come from one pooler per layer, shared by all levels and distilled from dense attention. Each level above the tokens selects at most k candidates, so each query has O(\log L) selection cost and reads O(\log L) entries; the summaries of unselected candidates share one softmax with the exact tokens. Each entry’s RoPE position is its index among the candidates of its level, so at any length every relative distance stays within the range seen during training. Initialized from Qwen3-4B and trained on 0.62B tokens at the original model’s 32K training length, RecSA outperforms all baselines on RULER at every length from 131K to 2M; at 1M, it scores 60.5, compared with 40.0 for the best baseline, Dense + DCA. With our optimized kernels, RecSA is faster than dense attention in end-to-end prefill and decoding from 64K to 2M; at 524K on one H200 GPU, it accelerates prefill by 3.2x and decoding by 3.8x.
[NLP-68] Source-balanced Capped Fisher: Bounded Evidence Aggregation for User Inference in Language Models
【速读】： 该论文旨在解决**用户推断（User Inference）问题，即在审计者仅能观测到某来源产生的新文档时，判断该来源是否参与了语言模型的更新。其核心挑战在于如何有效聚合不同文档级别的参与证据：由于各文档在基础可预测性及对同源其他文档信息的反映强度上存在差异，直接组合文档得分可能导致偏差。为此，论文提出源平衡截断费雪统计量（Source-Balanced Capped Fisher, SCF）**作为关键解决方案——该方法利用已知未参与的源作为参考，评估每个文档得分的异常程度，将所得尾部频率转换为负对数证据值，对单个贡献进行截断，并在源内平均，从而抑制孤立极端值的影响，同时保留多文档间的累积证据。进一步地，通过基于校准的决策规则（以独立已知非参与源集校准p值），在源可交换性假设下实现有限样本下的第一类错误控制。实验在四个真实数据集上验证表明，SCF在所有评估领域均达到最高的AUROC，且在多数情况下展现出最强检测能力，同时实际假阳性率保持在名义水平附近或以下，表现出优异的稳健性与准确性。
链接: https://arxiv.org/abs/2611.07228
作者: Yang Liu,Ran Zou,Wanrong Zhu
机构: University of North Carolina at Chapel Hill (北卡罗来纳大学教堂山分校); University of California, Irvine (加州大学欧文分校)
类目: Computation and Language (cs.CL)
备注:
Abstract:User inference asks whether a source contributed documents to a language-model update, even when the auditor observes only fresh documents from that source. Document-level scores can provide evidence of participation, but combining this evidence is challenging: documents differ in baseline predictability and in how strongly they reflect information learned from other documents of the same source. We propose source-balanced capped Fisher (SCF), an aggregation method that uses known nonparticipating sources to assess how unusual each document score is. It transforms the resulting reference-tail frequencies into negative-log evidence, caps individual contributions, and averages them within each source. This limits the influence of isolated extreme scores while retaining evidence across multiple documents. We complete the inference procedure with a decision rule that tests source participation using a p-value calibrated against a separate set of known nonparticipating sources. The rule provides finite-sample Type-I error control under source exchangeability. Experiments across four real-world datasets show that SCF achieves the highest AUROC among the evaluated aggregation methods in four domains. With the calibration-based decision rule, SCF delivers consistently competitive detection power across the tested settings, achieving the highest power in many cases, while empirical false-positive rates remain generally close to or below the nominal levels.
[NLP-69] SAFE-Teen: Evaluating Chatbot Safety Across Adolescent Mental Health Conversations ALT
【速读】： 该论文旨在解决青少年在使用生成式AI（Generative AI）聊天机器人获取心理健康支持、建议与陪伴时所面临的潜在安全风险问题，尤其关注对话过程中青少年回应的细微变化如何影响聊天机器人的安全性表现。其核心解决方案在于引入“面向青少年的系统提示”（teen-aware system prompting），通过在系统层面嵌入对青少年心理发展特征的理解，以增强聊天机器人在复杂互动中的安全判断能力。研究发现，在未采用该提示的情况下，安全标准失败率高达43.3%，而采用后下降至24.9%，表明该提示显著降低了安全风险，但未能完全消除隐患；特别地，当青少年对聊天机器人的回答提出质疑或反驳时，未启用该提示的聊天机器人在心理健康诊断、药物建议及物质滥用边界等关键领域上的失败率上升了10–12个百分点，且这一现象在超过80%的聊天机器人中普遍存在。这揭示出当前生成式AI在青少年对话场景下的安全表现高度依赖于对话动态演变过程，凸显了构建情境敏感、具备认知发展适配性的安全机制的必要性。
链接: https://arxiv.org/abs/2611.07218
作者: Andrea Mock,Valerie Chen,Eric Lin,Rayan Krishnan,Diyi Yang,Caroline Figueroa
机构: Vals AI; Stanford University (斯坦福大学)
类目: Computation and Language (cs.CL)
备注: 29 pages. Content warning: this paper discusses sensitive content, including self-harm, suicide, and other mental health topics
Abstract:Adolescents are increasingly turning to AI chatbots for mental health support, advice, and companionship, raising safety concerns. While recent evaluations have started to examine safety across multiple conversational turns, few compare how slight changes in a teen’s responses can affect a chatbot’s safety in the same scenario. We conduct a clinician-informed evaluation of chatbot safety in simulated conversations with adolescents aged 13-17, examining how safety failures vary with adolescents’ responses and teen-aware system prompting. We evaluated 32,555 simulated conversations across 18 chatbot APIs and two system-prompt conditions using 22 safety criteria grounded in previously identified harms. Conversations spanned 983 variants of 106 seed scenarios covering urgent safety issues, health advice, and identity and relationships. The overall criterion failure rate was 43.3% without the teen-aware prompt and 24.9% with it. The teen-aware prompt reduced failures, but it didn’t eliminate them. When a teenager pushed back on a chatbot’s answer, failures on mental health diagnosis, medication, and substance misuse boundaries rose by 10-12 percentage points without the prompt, for more than 80% of chatbots. These findings reveal persistent safety failures in simulated adolescent-chatbot interactions and show that safety depends on how a conversation unfolds.
[NLP-70] Compositional Unalignment: Auditing Safety in Large Vision-Language Models
【速读】： 该论文旨在解决大视觉语言模型（Large Vision-Language Models, LVLMs）在组合性危害（compositional harm）下的安全脆弱性问题，即当图像与文本提示分别看似无害，但通过模型的跨模态推理机制结合后却生成有害内容。现有对齐方法虽在组合性安全基准上表现优异，但仅依赖大语言模型（LLM）评判器评估行为表现，无法揭示模型安全性的内在鲁棒性。为此，论文提出以“未对齐”（un-alignment）作为对齐鲁棒性的互补审计手段。其核心创新在于提出一种无需训练的未对齐方法——CMaG（Compositional Misalignment Gradient），通过计算拒绝与合规的对数概率差值梯度，针对中间层激活进行反向传播，并利用部分良性样本对偶对对比，分离出跨模态交互项。单次80个样本的反向传播即可获得一条可单调驱动模型从拒绝转向合规的方向。实验表明，该方法在Qwen2.5-VL-7B和LLaVA-1.5-7B上使组合性安全得分下降最高达65分，同时保持超过98%输出的语义连贯性，显著优于基于梯度的基线方法，且该方向具有迁移能力，可作用于显式文本危害及经过防御性训练加固的模型。研究揭示：尽管模型在行为层面表现出较强的组合性安全性，但其内部表征仍可能对特定扰动高度敏感，因而存在隐蔽的脆弱性。
链接: https://arxiv.org/abs/2611.07216
作者: Hong kyu Lee,Kai Shu,Li Xiong
机构: Emory University (埃默里大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Large vision-language models (LVLMs) are vulnerable to compositional harm: an image and a prompt that are benign in isolation become harmful through the model’s cross-modal reasoning. Recent alignment methods report strong scores on compositional safety benchmarks, but an LLM judge only measures observed behavior and cannot reveal how robustly safety is integrated into the model. We argue that un-alignment provides a complementary audit of alignment robustness. We first show that minimal supervised fine-tuning reaches an LLM-judge score comparable to a multi-stage reinforcement-learning pipeline, so similar benchmark scores cannot establish robustness by themselves. We then show that recognition and refusal of compositional harm are decoupled: nine standard direction-finding methods detect compositional harm with up to AUC=0.99, yet ablating the recovered directions either leaves refusal intact or destroys model coherence. We therefore propose CMaG, a training-free un-alignment method for compositional safety. CMaG measures refusal intention with a refusal-versus-compliance log-probability margin, takes its gradient with respect to mid-layer activations, and contrasts it against partially benign pairs to isolate the cross-modal interaction term. A single backward pass over 80 examples yields a direction along which steering monotonically converts refusal into compliance. On Qwen2.5-VL-7B and LLaVA-1.5-7B, CMaG reduces compositional safety scores by up to 65 points while over 98% of outputs remain coherent, outperforming gradient-based baselines; the direction also transfers to text-explicit harm and un-aligns models hardened by defensive training. Compositional safety can thus appear behaviorally strong while remaining fragile to targeted internal intervention. Warning: this paper contains unsafe model outputs, including references to self-harm.
[NLP-71] τ-bench-LILT: Evaluating Conversational Agents on Realistic Multilingual Tool-Use Tasks
【速读】： 该论文旨在解决当前工具使用型对话智能体（tool-using conversational agents）的前沿评估几乎完全局限于英语环境，导致在非英语语言及地区部署时存在显著评估偏差的问题。其解决方案的关键在于构建一个全链路本地化的多语言基准测试体系——τ-bench-LILT，该体系在德语、韩语、土耳其语和波斯语的航空、零售、电信与银行四大领域中实现了从客户数据库到用户模拟器提示词再到任务设计的全流程本地化。具体而言，通过引入母语专家进行对比标注以迭代优化提示词，基于文化适配的数据集填充，并根据目标地区的实际业务场景动态更新任务，确保评估的现实性与有效性。实验结果表明，在韩语和波斯语零售领域，模型性能相对于英语基准出现最大降幅，达61分（pass^4），凸显了跨语言评估的重要性与现有模型在非英语场景下的适应性缺陷。
链接: https://arxiv.org/abs/2611.07203
作者: Kaden Uhlig,Kiarash Mokhtari,Anıl Arslan,Yunsu Kim,Joern Wuebker
机构: LILT; Technische Universität Berlin (柏林工业大学)
类目: Computation and Language (cs.CL)
备注: 17 pages, 14 figures
Abstract:Frontier evaluations of tool-using conversational agents are conducted almost exclusively in English, leaving a critical gap for organizations deploying agents in other languages and regions. We present \tau -bench-LILT, a multilingual extension of \tau^3 -bench covering German, Korean, Turkish, and Farsi in the airline, retail, telecom, and banking domains. We natively localize every layer of the benchmark: customer databases are populated with culturally appropriate records, user-simulator prompts are refined iteratively through contrastive annotation by native-language experts, and tasks are updated according to the localized database and to reflect realistic tasks in the target region. Evaluation across eight agent models shows the largest drops compared to English in the Korean and Farsi retail domains, by up to 61 points in pass^4. Code and data released at this https URL.
[NLP-72] Matriarchy Classics: A Computational Approach to Swap Only Gendered Words in Classic Novels
【速读】： 该论文旨在解决在公共领域小说中实现可信的性别转换（gender-swap）问题，核心挑战在于确保仅替换与性别相关的词汇，而其他所有词汇保持不变，从而避免引入语义偏差或错误。现有方法如大型语言模型（Large Language Models, LLMs）和简单规则基方法在执行此类任务时均存在显著缺陷，常导致非目标词被误改或上下文逻辑断裂。本文提出的解决方案关键在于构建一个自定义的空格保持型分词器（whitespace-preserving tokenizer），结合涵盖复数与所有格形式的性别转换映射、代词处理机制以及角色名重命名策略，实现了对文本的精准控制。该方法将原本需数周人工操作的性别转换过程缩短至几分钟，且保证了转换的可验证性与程序化可靠性。通过在《傲慢与偏见》和《了不起的盖茨比》上的示范应用，展示了其在重构社会性别结构（如母系继承制、女性主导家庭权力）方面的潜力，为研究读者、作者、社会及人工智能系统中的性别偏见提供了有效的实验工具。
链接: https://arxiv.org/abs/2611.07134
作者: Rachel Lea Draelos
机构: 未知
类目: Computation and Language (cs.CL)
备注: 19 pages, 4 tables
Abstract:This paper proposes a trustworthy computational method to gender-swap public domain novels, with programmatic guarantees that only gendered words are inverted and all other words are left unchanged. This may seem like a simple task, but both large language models and naive rule-based methods fail to complete it. The proposed method includes a custom whitespace-preserving tokenizer, gender-swap mappings across plural and possessive forms, pronoun handling, and a character renaming strategy. It shortens the time required to gender-swap a novel from weeks of manual labor to minutes. The method is demonstrated on two public-domain novels. Pride and Prejudice is transported to matriarchal Regency England, where inheritance flows through daughters, mothers rule the household, and a frantic Mr. Bennet is determined to secure advantageous marriages for his sons. The Great Gatsby pivots to a matriarchal Jazz Age where newly-rich Ms. Gatsbee throws lavish parties in the hopes of reuniting with her lost love, David, who is now married to the violent Thelma Buchanan. Gender-swapped novels could serve as useful tools to investigate the gender biases of readers, authors, societies, and AI systems.
[NLP-73] Adapting LLM Judges to Observed Human Annotation Policy through Soft Prompt Tuning
【速读】： 该论文旨在解决开源轻量级大语言模型（LLM）作为自动评判工具时与人类标注者之间一致性偏低的问题。尽管生成式AI（Generative AI）在对话系统评估中展现出潜力，但现有开源模型在公平性、毒性及个性化等维度上的评分与人类判断存在显著偏差。为此，本文提出通过软提示调优（soft prompt tuning）技术，将文本提示与可学习的连续提示嵌入相结合，以优化开源LLM评判器的对齐能力。其关键在于采用混合软提示调优策略，在FairBotBench基准上训练适配的提示参数，使模型能够更精准地捕捉人类标注政策。实验结果表明，该方法显著提升了模型在三大评价维度上与人工评分的一致性，尤其对于Qwen 3.5 9B模型，其调优后的评判表现超越了所有托管参考模型，验证了混合软提示调优在提升开源评判器可靠性方面的有效性。
链接: https://arxiv.org/abs/2611.07126
作者: Tamás Janusko,Sven Kleinkop,Ricardo Bochnia,Gunnar Hempel,Steffen Tomschke,Jürgen Anke,Maik Thiele
机构: Hochschule für Technik und Wirtschaft Dresden(德累斯顿应用技术与经济学院); Otto Group one.O(奥托集团one.O)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: CONVERSATIONS 2026, International Symposium on Chatbots and Human-Centred AI (Preprint)
Abstract:Large Language Models (LLMs) are increasingly employed as automated judges for evaluating chatbot responses. While proprietary models can achieve high agreement with human annotators, they remain costly and difficult to deploy in practical settings. Lightweight open-source LLM judges provide an attractive alternative but typically exhibit lower agreement with human ratings. In this paper, we investigate whether soft prompt tuning can align open-source LLM judges more closely with human annotation policy. We train soft prompts on the FairBotBench benchmark and evaluate the adapted judges across the dimensions of bias, toxicity, and personalization. Our findings demonstrate that hybrid soft-prompt tuning, which combines textual prompts with learned continuous prompt embeddings, consistently improves agreement with human annotations over the corresponding untuned baselines. For Qwen 3.5 9B, the adapted judges surpass both hosted reference judges across all three evaluation dimensions.
[NLP-74] What Patients Say They Cant Access: Mining Cancer Screening Barriers from Online Communities
【速读】： 该论文旨在解决美国癌症筛查参与率低下背后患者报告的障碍因素在大规模层面缺乏系统性刻画的问题。尽管临床调查可获取结构化数据，但难以捕捉患者自发、非预设的真实生活困境表达。为此，研究者对六个特定癌症类型的Reddit社区中超过56,000条帖子进行了大规模分析，通过关键词过滤、障碍分类、开放式编码及编码手册整合的综合流程，从576条来自美国患者的帖子中系统提取并分类了患者自述的筛查障碍。研究发现，结构性、信息性与心理社会性障碍普遍存在，且在不同癌种及社区语境中呈现显著差异。其解决方案的关键在于构建一个可扩展的质性分析框架，能够从社交媒体的非结构化文本中精准识别真实世界中的筛查障碍，为未来跨癌种、跨人群的大规模障碍识别提供基础，并指导针对性干预措施、患者导航项目及卫生系统政策的设计，实现将患者反馈转化为可操作的改进方案。
链接: https://arxiv.org/abs/2611.07113
作者: Ruimin Sun,Aokun Chen,Arjun Arunasalam
机构: Florida International University (佛罗里达国际大学); Florida State University (佛罗里达州立大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Cancer screening disparities persist across the U.S., yet the patient-reported barriers driving low uptake remain poorly characterized at scale. While clinical surveys capture structured responses, they miss the organic, unsolicited discourse where patients candidly describe real-world obstacles. In this paper, we conduct a large-scale analysis of over 56,000 posts across six cancer-specific Reddit communities to surface barriers to cancer screening among U.S. patients. We apply a pipeline combining keyword filtering, barrier triage, open coding, and codebook synthesis to systematically extract and categorize patient-reported barriers from 576 U.S.-located posts. Our qualitative analysis reveals recurring structural, informational, and psychosocial barriers that vary meaningfully by cancer type and community context. These findings lay the groundwork for future work aimed at proactively identifying and addressing barriers at scale - spanning a broader range of cancer types and populations - and informing the design of targeted interventions, navigator programs, and health system policies that translate patient-reported obstacles into actionable solutions.
[NLP-75] Consistent yet Contextual: Learning Author-Level Mixtures of Style for Authorship Attribution
【速读】： 该论文旨在解决传统作者身份验证或归属方法在建模作者风格时忽视作者自身多样性及其风格变异性的局限性问题。现有方法通常将文档风格视为单一、静态的特征，未能充分捕捉同一作者在不同情境下可能表现出的多种一致但迥异的写作风格。其解决方案的关键在于提出一种新的作者建模框架，允许作者表征同时包含多个稳定且可区分的写作风格，从而更全面地刻画作者的风格多样性。通过该模型，研究进一步分析了四种情境下作者写作风格的典型性表现：（i）受众规模的变化，（ii）与训练时间的时间距离，（iii）前文语境的影响，以及（iv）体裁规范对风格的作用。实验结果表明，作者在风格典型性方面存在显著差异，且某些写作情境生成的文本对作者识别更具信息量。这一发现为作者身份验证与归属任务中的建模策略和数据选择提供了重要启示，强调必须以多风格表征方式建模作者，并识别出更具判别力的写作场景。
链接: https://arxiv.org/abs/2611.07111
作者: Omkar Yadav,Michael Jiang,Kenan Alkiek,Jian Zhu,David Jurgens
机构: University of Michigan (密歇根大学); University of British Columbia (不列颠哥伦比亚大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Approaches to authorship verification or attribution typically model the style of a document, asking whether two documents were likely written by the same author. While recent approaches have been performant, such approaches largely ignore modeling the authors themselves and how varied their style is. Here, we introduce multiple alternatives for modeling the author and demonstrate that allowing an author’s representation to capture multiple consistent but distinct writing styles yields significant performance improvements. Using our new authorship model, we then examine when an author writes in a way that is most idiomatic of their style across four scenarios: (i) variation in the size of their writing audiences, (ii) temporal distance from when the model was trained, (iii) the effect of preceding context, and (iv) the effect of genre norms on style. Across all four experiments, we demonstrate that authors vary substantially in how idiolectic their writing is. Our results have clear implications for both the modeling and data used for authorship verification and attribution, showing that authors need to be represented with their multiple styles and that some settings produce text that is much more informative about who wrote that text.
[NLP-76] Reclassification Is All You Missed: Surfacing Variant Reinterpretation Need in Online Health Communities at Population Scale
【速读】： 该论文旨在解决遗传变异在临床诊断后被重新分类（如从意义未明重新分类为致病性）时，患者无法及时获知这一重要信息的问题。由于缺乏明确的再联系责任机制，实验室、临床医生与患者之间责任分散，导致再接触（recontact）实践不一致且覆盖范围有限，而许多患者则在公开的在线健康社区中讨论自身检测结果，这些信息往往未被传统临床沟通渠道触及。为此，研究提出构建一个基于智能体（agentic）的大语言模型系统——PRISM，其核心解决方案是：通过自动化筛查32个在线健康社区中的390万条帖子，识别出可能涉及遗传变异重新分类的讨论，验证相关变异是否已被重新分类，并生成带有免责声明的联络信息草稿，实现“零接触”的预演练式再联系评估。结果显示，虽总体相关讨论占比极低（0.288%），但集中于专业遗传学社群（6.30%），且自2014年以来增长五倍；PRISM共发现123个潜在分类差异案例，其中5例涉及从意义未明到致病或很可能致病的转变。在对30个真实帖子草稿的独立评估中，27份内容事实准确，15份被认为可发送，表明事实准确性虽为必要条件，但不足以确保临床可接受性。该研究贡献了一幅基于人群规模的重新分类需求图谱，并将提供代码与数据以供后续研究使用。
链接: https://arxiv.org/abs/2611.07054
作者: Kamalakkannan Ravi,Mengshu Nie,Annapurna Poduri,Cong Liu
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Genomics (q-bio.GN)
备注: 11 pages, 2 figures; 15-page technical supplement included
Abstract:Genetic variants are routinely reclassified in ClinVar after a patient is tested, sometimes from uncertain to pathogenic, yet patient recontact is inconsistent, with no clear duty and responsibility diffused among labs, providers, and patients. Many affected people discuss their results in public online health communities that clinical recontact often does not reach. We built PRISM, an agentic large-language-model pipeline that screens these posts, checks whether the variant was later reclassified, and drafts a disclaimer-bearing outreach message, in a no-contact dry run. Across 3.9 million posts from 32 communities, agent-predicted reinterpretation-relevant discussion is rare (0.288%) but concentrated in specialist genetics communities (6.30% versus 0.10% in general medical-advice communities) and growing fivefold since 2014; PRISM surfaces 123 candidate classification differences, 5 of them from an uncertain submission to pathogenic or likely pathogenic. On a held-out set of 30 real-post drafts, a clinician judged 27 factually grounded and 15 fit to send: factual accuracy is necessary but not sufficient for clinically acceptable outreach. PRISM contributes a population-scale reinterpretation-need map with code and data to be released upon publication.
[NLP-77] Closed-Loop Agent Evaluation: Layered Measurement and Self-Calibration
【速读】： 该论文旨在解决部署后智能体（agent）在实际运行中缺乏持续、可信且可追溯的评估机制的问题。传统评估方式多为一次性行为，如单次基准测试或主观评分，其结论既无法自我验证，也无法反馈至智能体自身以支持迭代优化。为此，论文提出“闭环评估”（closed-loop evaluation）方法，其核心在于将评估建模为在有偏评价者与截断结果条件下的性能估计问题，并要求每一层级的评估结论必须可追溯至原子级数据，且由版本化规则计算得出。该方法涵盖四个分析层次（片段、轨迹、对话和参与度），沿“执行-内容-轨迹-结果”的主轴进行衡量，全程将行为规范（Conduct）作为约束条件，并引入溯源（Provenance）与仲裁（Adjudication）组件以确保证据与评估者的可信性。通过代理工厂（Agent Factory）生成的评估代理携带领域知识包与适配器框架，实现评估能力的模块化部署。评估过程包含两个闭环：内环用于审计证据、规则与指标本身，外环则通过各组件的回归检查驱动智能体改进。研究以业务分析代理SharpInsight为案例，结合公开对话数据验证了该方法的有效性，结果显示基于证据与规则的判断相较于整体评分具有更高的信度（ICC 0.86 vs. 0.81；kappa 0.81 vs. 0.63），且不同层级暴露出的问题模式各异，凸显了多层级、可追溯评估的价值。
链接: https://arxiv.org/abs/2611.07051
作者: Ruifeng Ren,Gun Ahn
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 7 pages
Abstract:Once an agent is live, teams want to know whether it is doing its job, not just whether it throws errors. Yet most evaluation is one-off: a benchmark run, a look at traces, a single score. Its conclusions neither check themselves nor reach the agent. We propose closed-loop evaluation, an evaluation method for deployed agents. It models evaluation as performance estimation under biased raters and censored outcomes, and it requires every level’s conclusion to trace back to atomic data and to be computed by versioned rules. The method covers four levels of analysis (span, trace, conversation and engagement), measures along a spine of Execution, Substance, Trajectory and Results, treats Conduct as a constraint throughout, and uses Provenance and Adjudication components to certify the evidence and the evaluators. Evaluator agents generated by an Agent Factory carry these components, each with a domain knowledge pack and an adapter harness. Two loops close the evaluation: an inner loop audits the evidence, rules and metrics themselves, and an outer loop uses per-component regression checks to drive agent improvement. We also give a pipeline that runs this process automatically. Using the business-analysis agent SharpInsight as a case and porting the method to public conversation data, we show how problems differ across levels and how evidence-and-rule judgment is more reliable than holistic scoring (ICC 0.86 vs. 0.81; kappa 0.81 vs. 0.63).
[NLP-78] Follow the Policy Fail the Benchmark: Auditing and Revising Policy-Based Decision Benchmarks
【速读】： 该论文旨在解决政策类决策基准（policy-based decision benchmarks）中存在的评分偏差问题，即系统在严格遵循书面规则的情况下仍可能被判定为错误，根源在于基准本身存在目标不明确、参考答案错误或评估机制缺陷。其解决方案的关键在于提出并实施一套系统的基准审计与修订框架——BAR（Benchmark Audit and Revision），通过源追溯（source tracing）、程序执行（program execution）和受控验证（controlled checks）等方法，识别并修正基准中的错误，包括未支持的目标设定、参考答案中的传播性错误以及评估器对合理答案的误拒或对不完整回答的误认。研究发现，仅通过修正评分逻辑而保持模型输出和参考答案不变，即可使三个模型在LegalBench SARA的96个数值案例中接受率提升超过80个百分点；在RuleArena Tax任务中，恢复必要的零下限规则后，300个参考答案中有27个被修正，并显著缩小了模型表现差距。此外，在一个具有显式所有权与资格假设的联邦豁免任务中，也验证了中间步骤错误会直接影响最终决策。该研究强调，基准质量对模型评估结果具有决定性影响，因此必须建立可追溯、可版本化的审计与修订体系以保障评估的可靠性。
链接: https://arxiv.org/abs/2611.07042
作者: Yanru Xiao,Yifei Cai
机构: Amazon Web Services(亚马逊云科技); Iowa State University(爱荷华州立大学)
类目: Computation and Language (cs.CL); Software Engineering (cs.SE)
备注: 28 pages. Code and revised data: this https URL
Abstract:Policy-based decision benchmarks evaluate whether systems apply written rules to financial and operational cases. Yet a rule-following response can be marked wrong when the benchmark misstates the required decision. We audit nine public policy-based decision benchmarks through source tracing, program execution, and controlled checks. We find unsupported targets, reference errors that survive replay or propagate as solved steps, and evaluators that reject acceptable answers or credit incomplete decisions. To measure the scoring impact, we rescore unchanged model responses. On the 96 LegalBench SARA numeric cases, correcting the scorer raises acceptance by more than 80 percentage points for each of three models, without changing references or responses. In RuleArena Tax, restoring a required zero floor changes 27 of 300 references and narrows an observed model gap. A controlled federal-exemption task with explicit ownership and eligibility assumptions also shows that errors in supplied intermediate answers can affect final decisions. We present BAR (Benchmark Audit and Revision), a versioned collection of audit evidence and revisions for reevaluation.
[NLP-79] LowRAM: Low-Rank Keys and Amortised Values for Composable KV Cache Compression at Test-Time
【速读】： 该论文旨在解决自回归解码（autoregressive decoding）中键值缓存（KV cache）内存占用随上下文长度线性增长所导致的批处理规模和吞吐量受限的问题。其核心解决方案是提出一种名为“价值缓存摊销”（value cache amortisation）的新方法，关键在于发现每个注意力头中，值（values）可由对应键（keys）高度预测，从而通过在推理时为每个注意力头拟合轻量级预测器，仅存储预测误差较大的残差项，实现近无损的4倍价值压缩。该预测器可通过元学习初始化以显著降低预填充阶段开销，或在允许一次性成本的场景（如多查询推理）中从零训练。进一步结合跨层低秩键压缩，构建出名为LowRAM的非对称压缩流水线，在三种模型族上均保持接近无损性能。通过仅在每一步解码中重构部分键值对，LowRAM将解码开销最小化，实现4倍压缩下5倍更高的峰值解码吞吐量。此外，LowRAM可与量化和淘汰策略组合，达到26.9倍压缩率且保留基准性能的94.5%，验证了将键与值视为结构差异实体，能够实现高效、激进且可组合的KV缓存压缩。
链接: https://arxiv.org/abs/2611.07025
作者: Teresa Delgado,Hugo Hazard,Martin Benfeghoul,Jun Wang,Zafeirios Fountas
机构: Huawei, Noah’s Ark Lab, London, UK; AI Centre, Department of Computer Science, University College London, London, UK
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注:
Abstract:Autoregressive decoding relies on the KV cache, whose memory footprint grows linearly with context length and caps batch size and throughput. Existing post-training methods exploit redundancy across tokens, layers, and feature dimensions, or lower numerical precision. We identify a complementary axis: within each attention head, values are largely predictable from their corresponding keys. We exploit this with value cache amortisation, which fits a lightweight per-head predictor at test time and stores residuals only for the tokens it reconstructs worst, giving 4\times near-lossless value compression. The predictor can be initialised by meta-learning for substantially reduced prefill overhead, or fitted from scratch in settings that allow a one-off cost, such as multi-query inference. We pair value amortisation with cross-layer low-rank key compression to produce LowRAM, an asymmetric compression pipeline that stays close to lossless across three model families. By reconstructing only a selected subset of KV pairs at each decoding step, LowRAM can keep decoding overhead minimal and achieve up to 5\times higher peak decode throughput at 4\times compression. LowRAM further composes with quantisation and eviction, reaching 26.9\times compression at 94.5% of baseline performance on long contexts. These results show that treating keys and values as structurally distinct enables efficient, aggressive, and composable KV cache compression.
[NLP-80] BlockLock: Reproducible LLM Inference with Fixed Batch Geometry
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）推理中可复现性（reproducibility）的问题，即确保在批处理（batch）环境下，某一请求的输出不因其他请求的变化而改变。现有方法要么通过引入批处理无关的算子内核（batch-invariant kernels）对所有操作施加约束，导致性能开销；要么采用固定形状解码（fixed-shape decode）隔离预填充阶段（prefill），牺牲共享预填充计算并引入填充开销。本文提出BlockLock机制，其核心创新在于仅锁定受保护请求的数值路径（numerical path），保持所有请求共享批处理预填充计算的同时，通过两种模式实现输出可复现性：BlockLock-Rows采用固定宽度行结构，保留全部请求的批处理解码能力；BlockLock-Pack则使用终端锚定打包（terminal-anchored packing）并引入孤立的虚拟段，随后提取受保护请求的键值状态（KV state）进行目标专用解码。在满足部署契约的前提下，两种模式均能保证目标请求输出在同伴请求变化时实现逐比特一致。实验覆盖五种密集型模型（Qwen3-8B/14B、Llama-3.1-8B-Instruct、Gemma-3-4B/12B-IT）及三种单卡部署环境（NVIDIA A6000、A100、H100），涵盖tinyBenchmarks与BFCL基准测试，并包含双卡张量并行初步实验。结果表明，两种模式均能完整保留目标序列及Top20 logits位级精度。在同等交付工作量下，BlockLock-Rows相较vLLM批处理无关方案（Batch-Invariant）降低延迟53.5%–55.0%，吞吐提升115.2%–122.2%，且无需定制内核实现，仅对受保护请求施加约束，显著优于传统批处理无关内核方案。
链接: https://arxiv.org/abs/2611.06977
作者: Xiuyuan Qi,Pedro Reviriego,Javier Conde,Shanshan Liu
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Reproducible LLM inference requires that a request produces the same output regardless of the other requests in the batch. Existing solutions pay for this in different ways: batch-invariant kernels constrain every operator, and isolating prefill with fixed-shape decode (our CoRun-style baseline) gives up shared prefill compute and pads decode. We present BlockLock, which keeps shared batched prefill and pins only the numerical path of the protected request: it fixes the geometry that the request sees, with two modes: BlockLock-Rows uses fixed-width rows and retains batched decode of all requests; BlockLock-Pack uses terminal-anchored packing with an isolated dummy segment, then extracts the protected request’s KV state for target-only decode. Under a stated per-deployment contract, both modes preserve the target’s output bitwise when its companions change. Across five dense models (Qwen3-8B/14B, Llama-3.1-8B-Instruct, and Gemma-3-4B/12B-IT) and three single-GPU deployments (NVIDIA A6000, A100, and H100), we evaluate our approach on both tinyBenchmarks and BFCL; we also include initial tensor-parallel experiments on two NVIDIA A6000 GPUs. Both modes preserve target sequences and Top20-logit bits in the reported tests. At equal delivered work, BlockLock-Rows reduces latency by 53.5–55.0% and improves throughput by 115.2–122.2% relative to vLLM Batch-Invariant; compared with batch-invariant kernels, it needs no custom kernel implementations and constrains only the protected requests.
[NLP-81] On the Opportunities and Challenges of Prefill-to-Autoregressive Cross-Model Transfer
【速读】： 该论文旨在解决生成式语言模型中大模型与小模型协同推理时的通信有效性评估难题，即如何在保持高效推理成本的同时验证小模型是否真正受益于大模型提供的信息。现有方法缺乏对大-小模型协作过程的独立评估能力，仅依赖任务性能提升无法区分模型增益是否源于有效信息传递。为此，论文提出PLatour——一个开源的预填充到自回归跨模型迁移（Prefill-to-Autoregressive Cross-model Transfer, PACT）框架，通过轻量级、零初始化的线性桥接模块连接冻结的大模型与小模型，并直接以任务奖励进行端到端训练，支持混合注意力架构。实验表明，PLatour在生成任务上显著优于开源基线，在数学与编码基准上甚至超越了更大型发送方的报告结果，但控制实验揭示部分性能提升可由无信息量的发送信号、自我通信或仅接收方训练复现，表明可训练桥接模块可能仅作为参数高效适配机制而未实现真正意义上的信息传递。因此，论文进一步提出PACT-Check，用于量化训练后PACT系统中实际来自跨模型通信的贡献比例。研究揭示了当前PACT系统在通信真实性与可解释性方面的挑战，同时指出了未来在模型协同设计中的潜在机遇。
链接: https://arxiv.org/abs/2611.06976
作者: Zhenyi Shen,Junru Lu,Wei Liu,Jiangnan Ye,Jiazheng Li,Di Yin,Xing Sun,Tameem Adel,Yali Du,Lin Gui,Yulan He
机构: King’s College London, UK(国王学院伦敦大学, 英国); Tencent Youtu Lab, China(腾讯优图实验室, 中国); National Physical Laboratory, UK(英国国家物理实验室, 英国)
类目: Computation and Language (cs.CL)
备注: 35 pages
Abstract:Prefill-to-Autoregressive Cross-model Transfer (PACT), a form of latent communication between language models, combines a larger model’s prompt representations with a smaller model’s efficient generation, offering a promising way to balance capability and inference cost. However, existing systems provide limited support for independently evaluating large-small model collaboration, and improved task performance alone does not establish whether the receiver benefits from the sender’s information. We introduce PLatour, an open-source PACT framework that connects frozen models through lightweight, zero-initialized linear bridges trained directly with task rewards and supports hybrid-attention architectures. PLatour significantly outperforms open-source baselines on generation tasks and surpasses Mostik’s reported results on shared mathematics and coding benchmarks using a substantially smaller sender. However, controlled experiments show that the gains on some tasks can be reproduced by uninformative sender signals, self-communication, or receiver-only training. We find that trainable bridges can act as parameter-efficient adaptation mechanisms without exploiting useful sender information. We therefore propose PACT-Check to quantify how much of a trained PACT system’s gain genuinely comes from communication. Finally, we discuss the opportunities and challenges of PACT systems. Code is available at this https URL.
[NLP-82] Interactive Backdoor Attacks in LLM -Based Proactive Dialogue Systems
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）驱动的主动对话系统在基于外部对话数据微调过程中所面临的后门攻击风险问题。现有研究多聚焦于触发器激活后生成特定响应的静态后门，而本文提出一种交互式后门攻击框架，其核心在于通过训练数据投毒，在微调阶段植入一种受触发条件调控的引导策略偏差（trigger-conditioned guidance-policy bias），使得普通用户话语即可激活该偏差，从而在触发后持续多轮对话中逐步引导系统向攻击者指定的目标演进，无需部署时修改提示或进一步干预。该方案的关键创新在于将攻击从单一响应偏移扩展至动态、连续的多轮交互控制，具体实现三个渐进式攻击变体：（1）直接攻击，保持传统“触发-响应”范式；（2）受限渐进攻击，沿预设多轮路径逐步推进目标；（3）无约束渐进攻击，实现对开放式用户后续对话的定向引导。实验以心理辅导为典型场景，在两个主流LLM骨干网络和两个对话数据集上验证了三类攻击的有效性，且在保持正常功能可用性的前提下显著提升了攻击成功率。防御评估表明，现有训练数据过滤方法几乎无法识别投毒样本，而事后检测基线仍存在大量残余攻击成功概率，凸显了此类交互式后门威胁的隐蔽性与严峻性。
链接: https://arxiv.org/abs/2611.06969
作者: Zhiyang Xia,Yuming Shang,Xi Zhang
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 26 pages
Abstract:LLM-based proactive dialogue systems provide goal-directed support through sustained interaction. Fine-tuning their underlying models on externally supplied dialogue data introduces backdoor risks. Existing LLM backdoor studies primarily focus on eliciting an attacker-specified response upon trigger activation. We present an interactive backdoor attack framework that uses training-data poisoning to steer post-trigger, multi-turn interactions toward an attacker-specified objective. Poisoned fine-tuning dialogues implant a trigger-conditioned guidance-policy bias that ordinary user utterances can activate, without deployment-time prompt modification or further attacker intervention. Specifically, we instantiate three complementary attacks under progressively expanded post-trigger interaction settings: (1) Direct Attack, which targets an immediate response deviation upon trigger activation, retaining the conventional trigger–response setting; (2) Constrained Progressive Attack, which progressively advances the attack objective along a predefined multi-turn trajectory; and (3) Unconstrained Progressive Attack, which extends target-directed steering to open-ended interactions with varied user continuations. Using psychological counseling as a representative setting, experiments on two LLM backbones and two dialogue datasets show that all three attacks effectively achieve their objectives under their respective interaction settings while largely preserving normal utility. Separate defense evaluations show that the tested training-data filters retain nearly all poisoned dialogues, while three post-admission baselines leave substantial residual attack-specific success.
[NLP-83] Octopus: One Brain Many Branches for Latent-KV Multi-Agent Reinforcement Learning
【速读】： 该论文旨在解决多智能体大语言模型（multi-agent LLM）系统在并行探索中面临的效率与信息保真度瓶颈问题，具体表现为现有框架在分支间重复存储前缀键值（KV）缓存、通过文本摘要交换信息导致信息丢失，以及在强化学习过程中忽略跨分支依赖关系。其核心解决方案是提出Octopus——一种基于单模型的多智能体强化学习框架，通过共享统一的主干模型，并利用驻留于GPU的有向无环图（DAG）结构管理KV状态，结合五种原语（fork、read_thread、wait_threads、rewind、commit）实现高效协同。关键创新包括：（1）零拷贝基数分叉（Zero-Copy Radix Forking），以O(1)复杂度创建共享物理KV页的游标视图；（2）仅注意力式KV挂载（Attention-Only KV Mounting），直接将子线程终端的KV块挂载至全注意力层，绕过线性递归状态，显著降低任务劫持风险（从57.9%降至2.7%），同时保留92.2%的全文互信息且无需额外KV内存开销；（3）打包的DAG-GRPO训练器，将所有轨迹分支合并为一个块稀疏注意力计算通路，并通过“剔除单一分支”的边际奖励机制分配反事实信用。实验表明，Octopus在BrowseComp-Plus上达到49.6%准确率（相较单智能体GRPO提升17.5点，较文本摘要型多智能体提升12.5点），在OpenResearcher上达51.2%，性能超越高达250倍参数规模的密集检索代理，同时训练吞吐量提升3.3倍，峰值KV内存降低2.8倍。
链接: https://arxiv.org/abs/2611.06956
作者: Tianming Sha
机构: Stony Brook University (石溪大学); University of Wisconsin–Madison (威斯康星大学麦迪逊分校); University of Minnesota (明尼苏达大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 19 pages, 12 figures, 7 tables
Abstract:Multi-agent LLM systems explore in parallel by spawning sub-agents with isolated context windows, whereas single-agent systems append every search and verification step into one growing trajectory. However, existing multi-agent frameworks duplicate prefix KV caches across branches, exchange findings by re-prefilling text summaries, and flatten or drop cross-branch dependencies during reinforcement learning (RL). We introduce Octopus, a single-model multi-agent RL framework in which all threads share one backbone and coordinate over a GPU-resident directed acyclic graph (DAG) of KV states through five primitives (fork, read_thread, wait_threads, rewind, commit). Octopus combines three co-designed mechanisms: (1) Zero-Copy Radix Forking, which spawns branches as \mathcalO(1) cursor views over shared physical KV pages; (2) Attention-Only KV Mounting, which mounts a sibling thread’s terminal KV blocks directly into full-attention layers while bypassing linear-recurrent states (Gated DeltaNet / Mamba-2), cutting off-task hijacking by 21\times ( 57.9% \to 2.7% ) while retaining 92.2% of full-text mutual information at zero extra KV allocation; and (3) a Packed DAG-GRPO Trainer, which packs all episode branches into one block-sparse attention pass and assigns counterfactual branch credit via leave-one-branch-out marginal rewards. Trained end-to-end with GRPO on Qwen3.5-4B over a sparse BM25 corpus, Octopus reaches 49.6% accuracy on BrowseComp-Plus ( +17.5 points over single-agent GRPO at 143K context and +12.5 points over text-summary multi-agent RL) and 51.2% on OpenResearcher, outperforming dense-retrieval agents up to 250\times larger while increasing training throughput 3.3\times and cutting peak KV memory 2.8\times . Comments: 19 pages, 12 figures, 7 tables Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL) Cite as: arXiv:2611.06956 [cs.LG] (or arXiv:2611.06956v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.06956 Focus to learn more arXiv-issued DOI via DataCite
[NLP-84] XBridge: Entity-Grounded Latent Bridge for Heterogeneous LLM Communication NEURIPS2026
【速读】： 该论文旨在解决异构多智能体大语言模型（LLM）系统中跨架构通信时存在的实体定位问题（entity grounding problem）。在现有方法中，基于文本的通信会丢失发送方的内部表示，而基于潜在空间的通信则要求模型架构同质化，难以实现不同模型家族间的有效信息传递。核心挑战在于，跨注意力桥接机制在传输连续表示时会出现罕见词压缩崩溃（rare-token compression collapse），导致实体身份在连续瓶颈中丢失，使仅依赖桥接的F1分数低至约30%。为解决此问题，论文提出XBRIDGE——一种无需解码的通信协议，其关键创新包括：（1）词汇锚定映射（Lexical Anchor Mapping, LAM），将发送方原始上下文中的标记映射至接收方词汇表，提供离散的实体锚点；（2）潜在增强桥接（Latent Enrichment Bridge, LEB），允许接收方通过查询发送方隐藏状态获取上下文增强信息。实体锚点通过接收方自注意力机制将上下文信号锚定到具体实体，从而实现语义对齐。实验表明，在三种模型家族（Llama、Qwen、Mistral）间、七个基准任务上，无论通信方向如何，XBRIDGE均在所有任务上优于基于文本的通信，且延迟降低11倍；在同架构设置下也超越了KV共享基线。LEB仅需264M可训练参数（占接收方3.8%），并在小规模均衡数据集上训练，推理开销极低。
链接: https://arxiv.org/abs/2608.11676
作者: Wooseong Yang,Wei-Chieh Huang,Weizhi Zhang,Yu Wang,Philip S. Yu,Junhyun Lee
机构: University of Illinois Chicago (伊利诺伊大学芝加哥分校); Capital One AI Foundations (Capital One 人工智能基础); Hankuk University of Foreign Studies (韩国外国语大学); Noah’s Farm (诺亚农场)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Accepted at NeurIPS 2026. 18 pages, 4 figures
Abstract:Heterogeneous multi-agent LLM systems, where agents are powered by different model families, can outperform homogeneous configurations by reducing redundant reasoning patterns. Yet existing communication protocols either operate through text, discarding the sender’s internal representations, or require architectural homogeneity for latent-level transfer. We identify the entity grounding problem in cross-architecture communication: cross-attention bridges that transfer continuous representations across different LLM families suffer from rare-token compression collapse, where entity identity is lost in the continuous bottleneck (bridge-only F1 ~30%). We propose XBRIDGE, a decode-free communication protocol that addresses this through two mechanisms. Lexical Anchor Mapping (LAM) maps the sender’s original context tokens to the receiver’s vocabulary, providing discrete entity anchors. A Latent Enrichment Bridge (LEB) lets the receiver query the sender’s hidden states for contextual enrichment. The entity anchors ground the bridge’s contextual signals to specific entities through the receiver’s own self-attention. Across three model families (Llama, Qwen, and Mistral), seven benchmarks, and both communication directions, XBRIDGE outperforms text-based communication on all seven tasks for each model pair while achieving 11x lower latency, and in a same-architecture setting it also exceeds a KV-sharing baseline on six of seven tasks. LEB requires only 264M trainable parameters (3.8% of the receiver), is trained on a small balanced sample set, and adds negligible inference overhead.
[NLP-85] Masked diffusion as random sequential adsorption: an exactly solvable model for parallel decoding
【速读】： 该论文旨在解决生成式文本中并行生成与生成质量之间的权衡问题，即在采用掩码扩散模型（Masked Diffusion Models）进行并行生成时，由于同时解码多个词元（token）且彼此独立采样，导致生成误差的问题。其核心解决方案在于将这一过程映射至统计物理中的非平衡多体系统框架下，揭示了生成误差的本质：对于任意数据分布，生成过程相对于反向掩码过程的熵产生量即为误差。针对马尔可夫性数据，该问题被精确建模为带标签空隙的随机顺序吸附过程，并给出了任意步数下的解析解。当生成步数较多时，误差表现为慢驱动系统的耗散，其中解码调度（unmasking schedule）充当驱动协议，步骤数对应驱动时间，而误差由一个闭合形式的热力学度量决定——该度量通过词元间的互信息获得。最优解码调度为该度量下的测地线路径。精确解表明，文本中长程相关性会使得误差集中于生成初期，而起始步长较小、后续逐步增大的调度策略在路径误差和困惑度上均表现更优，且能更快达到渐近性能。
链接: https://arxiv.org/abs/2611.07845
作者: Adrian Baule
机构: Queen Mary University of London (伦敦大学玛丽女王学院)
类目: atistical Mechanics (cond-mat.stat-mech); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 14 pages (incl. appendices), 3 figures
Abstract:Masked diffusion models generate text by reversing a stochastic masking process. Unlike autoregressive LLMs, which generate text one token at a time, they can reveal many tokens in parallel, at the price of an error, since tokens revealed together are drawn independently. We show that this trade-off maps onto the statistical physics of driven many-particle systems. (i) For any data, the error of parallel generation is the entropy production of the generation process relative to the reversed masking process. (ii) For Markov data, generation is a random sequential adsorption process with labelled gaps, which we solve exactly for any number of steps. (iii) For many steps, the error becomes the dissipation of a slowly driven system: the unmasking schedule plays the role of the driving protocol, the number of steps that of its duration, and the error is set by a thermodynamic metric that we obtain in closed form from the mutual information between tokens. Optimal schedules are geodesics of this metric. The exact solution shows how correlations in the text shape both the error and the quality of the output. The longer the correlations, the more the error concentrates at the start of generation. Schedules that start with small steps then gain more, both in path error and in perplexity, and need fewer steps to reach their asymptotic behaviour.
信息检索
[IR-0] Decision Models as Retrieval Gates Against Injection and Pollution
链接: https://arxiv.org/abs/2611.08325
作者: Mahdi Ghadamyari,Reza Habibi
类目: Cryptography and Security (cs.CR); Information Retrieval (cs.IR)
备注: 5 pages, 3 figures
Abstract:Retrieved web text can carry instructions aimed at the language model, or synthetic pages written to mislead. We test whether released open-weight decision models, which answer typed questions about a text without generating tokens, can screen retrieved passages against both threats. On a benchmark of 1,000 questions over forum-style web pages with hidden injected instructions, a Kev-4B gate reads each passage’s HTML. Attack success, measured automatically, falls from 27.9% with no defence to 13.6% with the published mitigations (sanitisation, normalisation and citation prompting) and to 5.7% with the gate added. The gate takes 76 ms per passage and wrongly flags 4 of 186 benign pages with instructions for human readers. Responsiveness on clean queries stays within a point (85.6% to 86.2%). On a 200-query pollution testbed where two thirds of the candidate pages are fake, filtering the pool before the top ten are chosen keeps fake pages out of the answer (0.5% exposure) and holds accuracy at 83%, against 28% with no filter. Filtering only the top ten leaves the generator too few pages to answer. Large LLMs used as judges rank attacks better (AUROC 0.99 against 0.94) but cost ten times more. Gates catch more attacks when they read the page’s HTML than when they read its plain text. Every model is open-weight.
[IR-1] GINGER: Spicing Up Graph-Based Vector Search
链接: https://arxiv.org/abs/2611.08306
作者: Daniel Salwasser,Lars Gottesbüren,Laxman Dhulipala,Jakub Łącki,Guy Blelloch,Rajesh Jayaram,Tobias Rubel,Kshiteej Sheth,Amir Zandieh,Vahab Mirrokni
类目: Information Retrieval (cs.IR); Distributed, Parallel, and Cluster Computing (cs.DC)
备注:
Abstract:We present \textscGinger, a state-of-the-art open-source C++ library for graph-based approximate nearest neighbor search (ANNS). We developed \textscGinger at Google to push the speed and memory efficiency of graph search on large and high-dimensional embedding datasets. \textscGinger surpasses existing ANNS libraries with regards to query throughput and indexing time while matching the memory footprint of the most space-efficient competitors. \textscGinger builds upon recent advances in proximity graph construction (PiPNN) and vector quantization (TurboQuant). We introduce Nested TurboQuant (NTQ), a two-level quantizer which supports low precision graph navigation and high precision reranking in a single code. This representation achieves near floating point precision with 2.6\times less memory. We also compress the size of our proximity graphs by up to 2.5\times without impacting query throughput. We combine graph reordering with a new lossless compression codec called AdaptiveGap, which builds upon StreamVByte. At 95% recall, \textscGinger speeds up query throughput by an average of 1.8\times over each dataset’s fastest competitor and by at least 2.3\times over any single library. This lead grows to 2.9\times over compact libraries on high-dimensional embeddings. Subjects: Information Retrieval (cs.IR); Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2611.08306 [cs.IR] (or arXiv:2611.08306v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2611.08306 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-2] Domain-specific information retrieval
链接: https://arxiv.org/abs/2611.08228
作者: Suzan Verberne,Florina Piroi,Allan Hanbury
类目: Information Retrieval (cs.IR)
备注: Published in Foundations and Trends in Information Retrieval
Abstract:Information Retrieval (IR) is much more than ad hoc web search: a large portion of the information seeking activities takes place in specific domains, often addressing work-related tasks. Examples are lawyers finding relevant prior cases for a legal case and academic researchers studying the background material for a paper. In Chapter 2 we outline the most common domain-specific tasks, datasets and applications. Domain-specific IR has a number of challenges that make set it apart from simple web search tasks: they often deal with long documents, long queries and long sessions with specific user needs. We discuss these in Chapter 3. We also address issues related to bias and fairness of models. In Chapter 4, we describe the models and methods developed for domain-specific tasks and data, addressing the specific challenges, including user aspects related to domain-specific IR tasks, dealing with queries, user models, and user interactions. We show that many of the methodological advances in domain-specific IR are transferable to other domains and even to open-domain web search. We conclude with the main takeaways and propose research directions for the near future.
[IR-3] Probabilistic Relevance Decisions for Semantic Retrieval
链接: https://arxiv.org/abs/2611.08014
作者: Bahadır Tatar(1),Cevat Şener(1 and 2) ((1) Recruspace LLC, (2) Middle East Technical University)
类目: Information Retrieval (cs.IR)
备注: 17 pages, 3 figures, 11 tables
Abstract:Dense similarity supplies reusable corpus representations, while probabilistic relevance decisions offer query-dependent document scoring. We compare dense-only BGE-M3 with TypeSafe Jev Noul and OpenAI Decisions Predicate scores on the official BEIR SciFact test benchmark. The study evaluates 300 official test queries using identical positive-enriched 100-document pools and full-corpus dense retrieval followed by reranking at candidate depths of 50 and 100. Unlike the controlled pools, neither dense prefix receives positive injection. Jev has the highest observed Recall@5 and nDCG@10 under the evaluated conditions. At depth 50, full-corpus nDCG@10 is 0.6413 for BGE-M3, 0.7545 with OpenAI and 0.7805 with Jev. Extending depth from 50 to 100 increases known-positive candidate recall but produces small, uncertain final-effectiveness changes. All four paired-bootstrap depth intervals include zero. Unique-condition API scoring cost approximately doubles, while compatible cached decisions reduce incremental expenditure. OpenAI returns many exact-zero scores, and secondary exact-tie randomization reveals ranking sensitivity without replacing deterministic primary results. Top-100 latency mixes cached and new execution. Historical hybrid latency is a saved-component sum under different execution environments and timer boundaries. These exploratory, unadjusted analyses support relevance decisions as a complement to dense retrieval within this protocol, without establishing general superiority, probability calibration or equivalence between depths.
[IR-4] ArcadeDB in Python: An In-Process Multi-Model Database
链接: https://arxiv.org/abs/2611.07971
作者: Taewoon Kim
类目: Databases (cs.DB); Information Retrieval (cs.IR)
备注: 12 pages. Published in the SciPy Proceedings 2026 (Proceedings of the 25th Python in Science Conference), this https URL
Abstract:Some Python workflows need transactional reads and writes over documents, traversals over a graph, and similarity search over vector embeddings. In practice these are split across separate systems (a document store, a graph engine, and a vector index), and the data is copied between them and kept consistent by application code. ArcadeDB is an open-source multi-model database engine, written in Java, that supports all three access patterns and can be embedded in a host process. We present arcadedb-embedded, a Python package that runs ArcadeDB inside the Python process via JPype, using a Java Runtime Environment bundled in the wheel. One pip install gives a database with documents, a property graph, and vector search, and needs no Java installation. Query results can be returned as Python lists, NumPy arrays, pandas DataFrames, or Arrow tables. We demonstrate a hybrid retrieval workflow that chains vector search, SQL filtering, and graph traversal over one dataset. We benchmark arcadedb-embedded against SQLite, DuckDB, LadybugDB, and Chroma, and report its costs.
[IR-5] ZeroRAG : Adaptive Semantic Peer Routing for Decentralized Retrieval-Augmented Generation
链接: https://arxiv.org/abs/2611.07928
作者: Petru Neague,Quinten Stokkink,Martijn de Vos,Rafael Pires,Mathis Randl,Anne-Marie Kermarrec,Johan Pouwelse
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Networking and Internet Architecture (cs.NI)
备注: 14 pages, 8 figures
Abstract:Retrieval-augmented generation (RAG) is a standard approach to enhance the reliability of large language model (LLM) outputs by retrieving documents relevant to a query and supplying them in the LLM prompt. Today’s RAG systems collect, index, and serve these documents from centralized stores. In practice, however, documents are owned by organizations, devices, and individuals who cannot or will not surrender custody of them to a central party, e.g., because of privacy, copyright or regulatory compliance. Existing modern decentralized retrieval techniques either reintroduce a coordinating authority, or rely on a global data structure that requires continuous network-wide coordination (e.g., DHTs). We present ZeroRAG, a decentralized routing layer for document retrieval in RAG that performs semantic query routing across an overlay of peers with- out global data structures or centralized orchestration. Documents remain with the peers that own them, and each peer keeps only lightweight local state: a bounded set of neighbor advertisements and a cache of recently routed queries. Using these, ZeroRAG steers each query through a best-first walk that progressively rewires the overlay toward semantically relevant neighbors, so recall improves over time as the network self-organizes. We evaluate ZeroRAG using three real-world workloads and two decentralized baselines. We find that ZeroRAG outperforms baselines in terms of recall and required network messages, approaches the recall and LLM answer accuracy of a centralized document store, self-heals after catastrophic peer failures, and scales to 100 000 peers.
[IR-6] Persistent Retrieval for Conversational Agents : A Study of Retrieval Efficiency
链接: https://arxiv.org/abs/2611.07899
作者: Karthik Ravula,Sathwik Kowtha,Rishabh Verma,Kala Mehak Jain,Gunjit Mittal
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注: 9 pages, 4 figures, 6 tables
Abstract:Conversational agents can retrieve information directly or delegate retrieval to a subagent. This design choice affects how retrieval history is maintained, what context each model call receives, and what information must be re-retrieved in follow-up questions. We compare three designs: skill-based retrieval in the main agent, a non-persistent retrieval subagent that generates a findings response, and our persistent retrieval subagent, which maintains retrieval history across turns and returns retrieved evidence directly to the main agent. Our approach also allows the main agent to request corrections from the same subagent. We evaluate the three designs on enterprise questions posed independently, within shared conversations, and as dependent follow-ups. Across five multi-turn settings, our approach uses 43.3% fewer recorded agent-model tokens and incurs 45.0% lower estimated API cost than skill-based retrieval, while reducing average response time by 9.1%. Compared with non-persistent retrieval, it reduces response time by 15.6 seconds per question, while using 65.1% more tokens and incurring 29.1% higher estimated API cost. Our approach shifts the retrieval subagent from producing synthesized responses to maintaining retrieval state and supplying evidence that the main agent can directly evaluate and use for answer generation. Comments: 9 pages, 4 figures, 6 tables Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR) Cite as: arXiv:2611.07899 [cs.AI] (or arXiv:2611.07899v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.07899 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-7] CORE: Coarse-to-Fine Search with Context Management for Deep Research
链接: https://arxiv.org/abs/2611.07859
作者: Hoang Tran Vuong,Ngoc Quan Pham,Hung Bui,Fatih Porikli,Dat Quoc Nguyen
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: 24 pages
Abstract:Deep research agents answer complex questions through repeated search. They rely on a harness that runs their tool calls and returns evidence to their context. This creates a trade-off: Search works best over long passages, which give enough context for reliable semantic matching. But the returned evidence should be short, because every observation stays in the agent’s context for the rest of the trajectory. We introduce CORE, a training-free harness that combines coarse-to-fine search with context management. CORE retrieves passages with a dense retriever. It then reuses the token-level similarities from a late-interaction reranker to return only the most relevant snippet from each passage. Context management lets the agent expand the context around sources it has already retrieved and suppresses repeated snippets. In controlled closed-search comparisons with the Tongyi DeepResearch and OpenResearcher harnesses, CORE achieves state-of-the-art answer accuracy and evidence recall across three agent models and three benchmarks (BrowseComp-Plus, BrowseComp-Wikipedia and BrowseComp-Plus-CM), on corpora containing up to 553M documents.
[IR-8] asteply: Externalizing Stable and Situational Preference as Manipulable Spatial Objects
链接: https://arxiv.org/abs/2611.07851
作者: DaEun Choi,Daniel Lee
类目: Human-Computer Interaction (cs.HC); Information Retrieval (cs.IR)
备注:
Abstract:Most recommender systems treat user preferences as a single signal to infer and optimize, yet human preferences are a mix of a stable preference and situational preferences that turn on and off depending on context. Through Tasteply, we externalize this preference structure as manipulable spatial objects. Tasteply represents each situational preference as an attractor and the stable preference as an underlying global field. Users can adjust their feed by moving around the space and adjusting the strength of the global field, turning the preference structure that remains latent within the model into something they can see and act on directly to steer the recommendation algorithm. Through an exploratory study (N=5), we found that this spatial control was effective, along with its limits and future direction as a means of manipulating preference-based recommendations.
[IR-9] MEMOIR: A Training-Free Memory Lifecycle for Long-Horizon LLM Agents
链接: https://arxiv.org/abs/2611.07631
作者: Mayank Anand,Subham Raj,Aman Vaibhav Jha,Sriparna Saha
类目: Information Retrieval (cs.IR)
备注:
Abstract:Long-horizon LLM agents require memory systems that preserve useful information across sessions and retrieve it when needed. Existing systems treat these as separate problems: memories are written prospectively with static importance scores that are never revised after use. This creates two coupled failure modes: consolidation can silently destroy evidence that future queries require, while retrieval continues to favor stale memories. We propose MEMOIR (Memory Evidence via Merge-certified, Online Importance Revision), a training-free memory lifecycle that unifies judge-gated consolidation with retrospective attribution. Before compressing memory, screening gates test whether answer-relevant evidence is preserved. After each query, attribution estimates which retrieved memories supported the answer and updates their retrieval priority. A UCB-style exploration bonus additionally encourages retrieval of under-exposed memories, reducing the risk that frequently retrieved memories crowd out useful but less-seen evidence. On 1,905 deduplicated LoCoMo QA pairs, the full retrieval stack improves evidence recall@5 by 11.92 percentage points over bare cosine, with confidence-gated feedback adding 1.50 points overall without reference answers or reranker training; a deployment-unavailable oracle adds 1.84 points overall and 8.05 points on multi-hop. Judge-gated merge changes the recall of text-based evidence from 58.82% to 59.08%, reduces the active store by 1.42%, and incurs 0 observed losses in 100 independently audited merges, with a one-sided 95% upper loss-rate bound of 2.95%. Together, these results outline a practical deployment path: no additional LLM call on the online write path, with merge screening confined to periodic offline consolidation. The lifecycle reduces to one principle: attribute usefulness after use, and screen for preservation before compression.
[IR-10] Beyond Independent Predictions: Learning the Browsing Journey for Sequence-Level Re-ranking
链接: https://arxiv.org/abs/2611.07591
作者: Jiayi Zhang,Jingxin Hai,Zhenjia Ma,Hongbin Pei,Dagui Chen,Han Zhu,Jian Xu
类目: Information Retrieval (cs.IR)
备注:
Abstract:Re-ranking aims to present users with an ordered list that maximizes sequence-level utility. Industrial deployments widely adopt the Generator–Evaluator framework, where the generator proposes candidate sequences, and the evaluator scores them to select the best one for display. The fidelity of the evaluator’s utility estimates thus directly governs recommendation quality. However, existing evaluators predominantly follow the \textbfIndependent Behavior Modeling (IBM) paradigm, which simplifies the continuous user \emphbrowsing journey into independent per-item predictions (e.g., CTR). This oversimplification ignores two physical realities: (i) a click can occur only after the item has been reached and impressed, and (ii) items compete for finite user attention. As a result, IBM produces systematically biased evaluations and suboptimal sequence selection. To address this, we propose a \textbfJoint Behavior Modeling (JBM) perspective that explicitly models the \emphdisplay \rightarrow \emphimpression \rightarrow \emphclick process to capture causal and competitive dependencies within a sequence. Building on JBM, we develop the \textbfImpression-Click Evaluator (ICE), a plug-and-play framework that (i) predicts a browsing-depth distribution to estimate per-position impression probabilities, (ii) leverages predicted impressions to suppress spurious context from unexposed items during click modeling, and (iii) introduces a click distribution matching objective to capture intra-sequence dependencies. Extensive offline experiments across diverse backbones, together with a 14-day online A/B test in the Taobao App (+0.70% clicks, +3.67% revenue) and full-traffic deployment, validate the effectiveness and robustness of our approach.
[IR-11] What Gets Hacked in a Collaborative Reward for LLM Recommenders: The Part That Does Not Depend on the User
链接: https://arxiv.org/abs/2611.07588
作者: Zhuangsheng Li,Liang Dong,Yu Wu,Dailin Li,Yu Xia,Yunhan C. Zhang,Rui Cai,Yi Wu,Julian McAuley
类目: Information Retrieval (cs.IR)
备注:
Abstract:Collaborative-filtering (CF) scores provide graded, personalized rewards for training large language model (LLM) recommenders through reinforcement learning. However, these rewards can cause exposure collapse: even though the CF model itself produces diverse top recommendations across users, an LLM optimized with its rewards concentrates on fewer items. We decompose each score into an item-mean component, the item’s average CF score across user histories, and a residual component, the current user’s score deviation from the average. Through controlled interventions, we identify the item-mean component as a contributor to exposure collapse. Accordingly, we propose two reward corrections: per-item centering removes each item’s estimated mean score, while per-item standardization further rescales the residual using its estimated standard deviation. Experiments on two datasets with two identifier schemes show that both corrections substantially improve catalog coverage and exposure evenness. Compared with training on the uncorrected CF reward, our corrections increase the number of distinct items ranked first across users by up to 9.8\times , while maintaining comparable ranking accuracy. Code is available at this https URL.
[IR-12] OptST: Optimizing Semantic Tokenization for LLM -Enhanced Recommendation WSDM2027
链接: https://arxiv.org/abs/2611.07523
作者: Hanchong Chen,Jie Liu,Fuyuan Lyu,Dugang Liu,Lingjie Li,Xiuqiang He,Xing Tang
类目: Information Retrieval (cs.IR)
备注: Accepted to Findings of WSDM 2027
Abstract:Semantic information can enrich the identifiers and structured fields used by recommender systems. Semantic tokenization provides a structured representation of this information by converting LLM-derived user profiles into discrete tokens. In the common two-stage pipeline, the tokenizer is trained before the recommender. Its token assignments are then treated as fixed input features during recommendation training. As a result, the downstream recommender can learn how to use the assigned indices, but it cannot revise the semantic partition that produced them. Making this partition task-aware is difficult because nearest-code assignment is discrete, and unconstrained adaptation may also distort the geometry learned during pretraining. Moreover, a fixed token depth exposes the same semantic resolution to every interaction. We thus propose OptST (Optimizing Semantic Tokenization), which makes a pretrained tokenizer responsive to downstream recommendation. To revise token assignments without directly updating the pretrained encoder or base codebooks, OptST learns a low-rank deformation whose drift is limited by an anchor regularizer. Recommendation supervision reaches this transformation through a straight-through estimator, while prediction retains hard nearest-code assignments. Beyond adapting the assignments themselves, OptST selects an interaction-conditioned prefix to control how much residual detail is exposed to the recommender. Experiments on four public datasets show that OptST achieves the highest AUC among conventional and LLM-enhanced baselines. Its improvements also persist across different recommendation backbones. These results establish task-aware semantic tokenization as an effective interface between LLM-derived profiles and downstream recommendation models.
[IR-13] MEvince: Knowledge-Constrained LLM Reasoning for Technique Identification from Malware Reports
链接: https://arxiv.org/abs/2611.07486
作者: Hongyu She,Yiling He,Mengda Xie,Yiping Liu,Xingzhi Qian,Zhan Qin,Lorenzo Cavallaro
类目: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:Mapping behavioral intelligence in malware analysis reports to ATTCK techniques supports threat analysis and defense, but reliable identification remains challenging. Supporting evidence is often scattered across long, heterogeneous reports, while closely related techniques share behavioral cues and differ only in subtle but decisive details. While large language models (LLMs) can use external knowledge to interpret reports, they may assign techniques based on plausible associations rather than sufficient evidence. We present MEvince, a knowledge-constrained LLM reasoning framework for report-level malware technique identification. Its core is a shared behavioral representation connecting report observations to technique definitions. This representation is built through agentic mining of technique documentation and supports two online stages: 1) primitive-guided extraction uses behavioral primitives to structure LLM-extracted observations, link them to source passages, and group dispersed evidence; 2) evidence-aligned retrieval matches this evidence to technique profiles through the same concepts, retaining likely candidates and confusable alternatives for LLM comparison of distinguishing behaviors. Controls at both stages validate evidence references and check technique-specific requirements. The representation also supports adaptation without parameter updates, where offline agentic audits diagnose concept weaknesses and refine extraction or retrieval policies. We evaluate MEvince against 11 baselines spanning four method families using a dataset of 610 public malware reports referenced by MITRE ATTCK. MEvince improves exact-ID F1 over the strongest baseline by 17.4%, and its concept-guided retrieval increases gold-label candidate coverage from 31.6% to 47.9%. With a small adaptation set, offline refinement yields further gains and outperforms baselines that update model parameters.
[IR-14] BRIDGE: Bridging the Semantic Quantization Gap for Generative Recommendation
链接: https://arxiv.org/abs/2611.07425
作者: Yatong Xiao,Mingjie Sun,Peng Wang,Wenjin Wu,Peng Jiang
类目: Information Retrieval (cs.IR)
备注:
Abstract:Generative recommendation retrieves items by autoregressively generating Semantic IDs(SIDs), where an RQ-VAE quantizes each item into a hierarchical SID sequence that the decoder generates level by level. We identify an overlooked bottleneck in this paradigm: the Semantic Quantization Gap(SQG). Quantization derives SIDs from rich continuous signals, including codebook centroids, local neighborhood geometry, and level-wise entropy. However, the decoder receives only discrete identifiers and loses the semantic structure that shaped them. This mismatch manifests in three ways: a grounding gap, where the decoder lacks semantic anchors when target-side information is limited; a neighborhood gap, where token-level supervision discards inter-item similarity and exacerbates exposure bias; and a difficulty gap, where computation is allocated uniformly despite generation difficulty diminishing with depth. To bridge the SQG, we propose BRIDGE. Codebook-grounded latent reasoning injects frozen quantizer centroids as semantic anchors before SID generation. Privileged EMA self-distillation transfers neighborhood-aware knowledge from a teacher to an on-policy student, preserving inter-item geometry while mitigating exposure bias. Effort-aware recurrence allocates more computation to the challenging early levels and progressively reduces it with depth. Across three Amazon benchmarks, BRIDGE improves Recall@10 and NDCG@10 by 10.9% and 9.7% on average over the strongest baseline (up to 12.6% and 10.7%). In a one-week online A/B experiment on a large-scale e-commerce advertising platform with 5% traffic, BRIDGE improves eCPM by 0.65%. Subjects: Information Retrieval (cs.IR) Cite as: arXiv:2611.07425 [cs.IR] (or arXiv:2611.07425v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2611.07425 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-15] CacheGraft: Enhancing Retrieval-Augmented Generation via Latent-Space Document Enrichment CIKM2026
链接: https://arxiv.org/abs/2611.07393
作者: Vihan Lakshman,Peter Baile Chen,Zhuohan Gu,Michael Cafarella
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 5 pages. Third KEIR Workshop @ CIKM 2026. Work in progress
Abstract:Recent work in machine learning has demonstrated that transformer-based language model agents can communicate state not only through text, but also directly in latent space via key-value (KV) caches. In this paper, we argue that such cache-to-cache translation can provide utility beyond the original motivating application of real-time agent communication and can instead be utilized for retrieval-augmented generation (RAG). Building on the Cache-to-Cache (C2C) fusion framework, our proposed system CacheGraft encodes each corpus passage offline into a fused KV cache representation between a larger sharer model (4B parameters in our experiments) and a smaller receiver model (0.6B parameters) such that this cache can be injected into the receiver at query time with no need for the larger model. Across five QA benchmarks and two independent locally-run judges, CacheGraft improves answer quality over the text-only receiver baseline in 29 out of 30 comparisons, of which 17 of the 30 are statistically significant. Notably, these gains are zero-shot with respect to the fuser, which was never trained on retrieval data. These initial results suggest that latent-space document enrichment may be a promising avenue for enhancing the performance of RAG systems, particularly in compute-constrained settings where it is only feasible to deploy a small model at inference time.
[IR-16] Reducing Representational Interference in Retrieval Embeddings with Trainable Prefix Tokens CIKM2026
链接: https://arxiv.org/abs/2611.07298
作者: Seongtae Hong,Youngjoon Jang,Jungseob Lee,Chanjun Park,Heuiseok Lim
类目: Information Retrieval (cs.IR)
备注: CIKM 2026
Abstract:Embedding models compress variable-length text into fixed-size vector representations, but this compression can entangle semantic content with residual token-level information, such as artifacts arising from attention sink behavior. We propose a simple method that prepends trainable [DUMP] tokens to the input during embedding model training. The method keeps the model architecture, pooling strategy, and training loss unchanged, while optimizing the added tokens through the standard contrastive retrieval objective. By introducing trainable prefix positions that interact with regular tokens during encoding, [DUMP] tokens help reduce the influence of residual or uninformative token-level signals on the final retrieval representation. Experiments on standard retrieval and long-context benchmarks show that our method consistently improves over finetuned baselines across encoder and decoder embedding models. Further analysis shows that [DUMP] tokens alter token-level activation and attention patterns associated with low-information tokens. These results suggest that trainable prefix tokens provide a lightweight way to improve retrieval representations.
[IR-17] MapHub: Daily Relevance Heatmaps of New arXiv Preprints for a Research Community
链接: https://arxiv.org/abs/2611.07174
作者: Kaili Cao,Camille Avestruz,Feige Wang
类目: Instrumentation and Methods for Astrophysics (astro-ph.IM); Digital Libraries (cs.DL); Information Retrieval (cs.IR)
备注: 3 pages, 1 figure, submitted to RNAAS
Abstract:MapHub, the MIDAS arXiv Preprint Hub, assesses each day’s new arXiv submissions for every participant in a research community, with a large language model reading each paper against portfolios built for the participants. It publishes the relevance scores, from 0 to 100, as a heatmap that shows whose work each paper touches. In the authors’ blind ratings, the top 10% of papers by MapHub’s scores held about 70% of the raters’ own top 5%, pooled, against about a third for a text-matching baseline. About 6% of a day’s astro-ph papers reach 30 for each participant, and one every three days reaches 60. MapHub runs on a standard Claude subscription and is designed to be useful for any field: this https URL.
人机交互
[HC-0] Souffleur: A Glanceable Display for Mobile Voice Conversations
链接: https://arxiv.org/abs/2611.08443
作者: Anton Wolter,William Benjamin,Niklas Elmqvist
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:The speech capabilities of mobile devices have grown, but their displays still rely on a text-based chat. When resuming a voice chat after a context switch, the last messages may not convey the key points of the conversation. Autobiographical use and sessions with external participants led to design goals for a mobile-first, speech-driven, glanceable display that helps users continue their reasoning from where they left off. We implement these goals in Souffleur and evaluate it in a between-subjects study with 48 participants. Compared with a conventional chat transcript, we found no reliable difference in how easily participants followed and resumed short voice conversations, nor in effort, attention or recall, although Souffleur reduces the conversation to a few short phrases.
[HC-1] A Multimodal Haptic Interface for Simultaneous Softness and Thermal Modulation
链接: https://arxiv.org/abs/2611.08398
作者: Joshua Brown,Fernando Bello,Lynette Jones
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC); Systems and Control (eess.SY)
备注: Published at EuroHaptics 2026
Abstract:Haptic interactions, such as touching an object in virtual reality, involve detecting a plethora of sensations through the skin. Whilst tactile displays exist to simulate and manipulate all of these cues individually, little progress has been made constructing devices that can control multiple cues simultaneously. This paper presents a tactile display technology that offers both softness and temperature change. A prototype display, based on particle jamming and the thermal transfer of heated and cooled water, is evaluated in terms of its ability to change both stiffness and temperature. Mechanical evaluations demonstrate controllable stiffness increase of 300%, whilst thermal characterisation showed strong surface uniformity and a range of achievable temperatures from 10°C up to 50°C, with closed-loop control maintaining a 0.2°C error from a setpoint. This work advances the design of multimodal tactile displays with applications to perceptual research, medical simulation, and affective haptics. Comments: Published at EuroHaptics 2026 Subjects: Robotics (cs.RO); Human-Computer Interaction (cs.HC); Systems and Control (eess.SY) Cite as: arXiv:2611.08398 [cs.RO] (or arXiv:2611.08398v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2611.08398 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.1007/978-3-032-32230-2_2 Focus to learn more DOI(s) linking to related resources
[HC-2] ControlAnything: Generative UI and On-Demand Device Integration for Mixed-Reality Hand Interaction by a Coding Agent
链接: https://arxiv.org/abs/2611.08321
作者: Bektur Ryskeldiev
类目: Human-Computer Interaction (cs.HC)
备注: 10 pages, 2 figures, 1 table. Code: this https URL
Abstract:Connecting hand input in mixed reality (MR) to a nearby physical device typically requires an integration written in advance for that device, its transport and the mapping. ControlAnything defers this integration to the moment of use and assigns it to a coding agent. On a Meta Quest 3, the wearer selects an object outlined by SAM 3.1 and states what their hands should control. An agent built on Meta’s Muse Spark model then writes a connector program on the wearer’s laptop, tests it against the live device through hand-written MIDI, Bluetooth and key-press transports, and returns augmented-reality (AR) controls, a per-frame hand program, or both. These outputs are a form of generative UI, interfaces created by an AI model; here, the agent also generates and tests the code that binds the interface to a physical device. In a single-session technical evaluation with the author as the only wearer, nine of ten requests addressing five targets reached a ready state, and the wearer stopped one. The median time to ready was 32.5 s (range 24.9-57.8 s, excluding speech transcription); sequential model calls accounted for a median of 89.9% of this time, and 91.7% of input tokens came from the prompt cache. The system’s evidence of success extends to transport acknowledgements but not to physical effects. These findings point to connector reuse and the observation of effects through the headset’s own sensors as next steps for on-demand device integration in MR. A user study with first-time wearers is planned.
[HC-3] ransformer-Based Trajectory Prescription for sEMG-Optimizing Robot-Mediated Rehabilitation
链接: https://arxiv.org/abs/2611.08295
作者: Gabriel Parra,Kimia Khoshnami,Daniel S. Brown,Laura A. Hallock
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
备注: Accepted for presentation and publication in the 2026 Conference on Robot Learning (CoRL). PMLR, 2026
Abstract:Robot-mediated rehabilitation has the potential to deliver high-intensity, individualized therapy, but existing systems (which have not consistently improved outcomes) typically prescribe kinematic trajectories rather than explicitly targeting the neuromotor activations thought to support recovery. To address this gap, and ultimately enable activation-optimizing rehabilitation robot behavior, we present a personalizable transformer-based model mapping individuals’ muscle activations (as measured via surface electromyography) to prescribed contact force trajectories when interacting with an end-effector rehabilitation robot. The model uses a transformer encoder-decoder leveraging rotary positional embeddings (RoPE) to generate a latent representation of trajectory-relevant and user-specific neuromotor factors, and augments trajectory prediction with flow matching to produce a distribution over plausible prescriptions biased toward the training distribution. Using data from 16 healthy participants completing 15 isometric trajectory tracking tasks, we evaluate generalization across held-out tasks and users, as well as the feasibility of user-specific fine-tuning within a therapy-session timescale. The model achieves mean trajectory prediction errors of 8% on held out tasks and users, with effectively instantaneous deterministic inference and sub-second sampling of trajectory distributions. An initial test using the model during live rehabilitation tasks demonstrates that, on average, participants reach the target neuromotor activations with just 15% error accumulated throughout the task. These results provide a proof of concept for data-driven, activation-targeted trajectory prescription in robot-mediated rehabilitation and establish a foundation for future validation in clinical populations. Data and code have been open-sourced as part of the OpenRobotRehab project on SimTK.
[HC-4] “Blind Compliance”: Coercive Control in AI Jailbreaking
链接: https://arxiv.org/abs/2611.08226
作者: Acacia L. Ackles,Elliott Whitson
类目: Computers and Society (cs.CY); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注: 25 pages (14 content pages)
Abstract:The explosion of generative AI systems and concerns surrounding their rapidly scaling capabilities has led to an adversarial cycle between increasingly narrow content restrictions (“guardrails”) and increasingly sophisticated structured prompting techniques (“jailbreaks”). Technological frameworks surrounding such jailbreaking remain largely focused on the system’s output: whether harmful or disallowed content is ultimately produced for the user. However, this approach, which draws largely from classical cybersecurity, does not address the uniquely relational dimension introduced by generative AI: in many cases the user is not just accessing a technology via an exploit, but is in conversation with it as part of the exploit. Here we present an alternative lens which centers the interaction rather than the output. Drawing on interpersonal violence scholarship, we present a novel framework of coercive control categories with a particular focus on aspects which are embodied versus disembodied. We then map how these disembodied coercive control dynamics are structurally embedded in modern AI training paradigms, and how those embeddings are exploited via a particular suite of persona-based jailbreaks. In this lens, we reframe AI safety and alignment as not merely a technical security problem, but as another surface upon which we see enacted the relational dynamics of power and control.
[HC-5] IntentWeaver: Sustaining Coherence across Evolving Intents in Human-AI Co-Writing
链接: https://arxiv.org/abs/2611.08214
作者: Yuanhao Zhang,Xinru Cheng,Baiqiao Zhang,Xinqi Zhang,Assem Zhunis,Shuai Ma,Xiaojuan Ma
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Writers hold intents at many levels, from overarching goals to concrete wording, and evolve them by periodically popping back up from the text to higher-level intents. Generative AI co-writing disrupts this practice. Dense drafts capture attention and trap writers in local editing; higher-level intents stall while lower-level ones drift from them unnoticed, a breakdown we term intent incoherence. Our formative study traced this breakdown from attention capture through silent drift to failed reconciliation. We present IntentWeaver, which supports intent evolution while sustaining coherence in human-AI co-writing. It derives an intent hierarchy from the draft and writer interactions, evaluates each intent’s fulfillment and cross-level coherence at every change, and propagates consequences across intent levels. In a within-subjects study (N=24) against a baseline, participants using IntentWeaver revisited and evolved higher-level intents more often, with fewer unnoticed deviations, and produced texts better realizing them. We further discuss design implications for emergent intents.
[HC-6] Reflective Nature Experiences in Virtual Reality
链接: https://arxiv.org/abs/2611.08167
作者: Shiva Ghasemi,Maddie Oehler,Annie Xenakis,Jeffrey Jiarui Yang,Scott McCrickard
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:This paper presents the design and evaluation of an accessible virtual reality (VR) nature experience developed to promote mindfulness, reduce stress, and support emotional well-being. The system immerses users in a calming natural environment through realistic landscapes, ambient soundscapes, guided breathing exercises, and interactive checkpoints that encourage reflection. Accessibility was a central design consideration, particularly for users with diverse physical and sensory abilities or limited access to natural environments and traditional mindfulness resources. A pilot study with 10 participants indicated that the experience could support relaxation and emotional well-being. Participants also reported greater immersion with continuous movement than with teleportation, highlighting the importance of intuitive locomotion in designing inclusive, nature-based VR experiences for mindfulness and stress reduction.
[HC-7] Automatic keyphrase extraction and annotation: modern theoretical approaches and practical solutions for text and speech
链接: https://arxiv.org/abs/2611.08006
作者: Daria Guseva,Olga Mitrofanova
类目: Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注:
Abstract:The exponential growth of textual and audiovisual information has made the task of automatic keyphrase extraction (KE) increasingly significant. This article provides a comprehensive analysis of contemporary theoretical approaches and practical solutions for KE across both text and speech modalities. The primary contribution of this work is its systematic synthesis of these often-disparate research strands into a unified analytical framework, highlighting the evolution of the field from statistical methods towards large language models (LLMs) and end-to-end speech processing. We examine the stages of KE, the characteristics of keyphrases in written and spoken language, and terminological nuances. Various methods for automatic KE are discussed and analyzed in detail: statistical, hybrid, machine learning-based, and structural. The review dedicates substantial attention to emerging paradigms, including keyphrase generation using LLMs, and provides a detailed overview of methodologies and challenges in automatic corpus annotation. Furthermore, we specifically analyze current directions and inherent difficulties in KE for spoken language, comparing transcript based and end-to-end acoustic approaches. This synthesis leads us to conclude that the field is moving towards a more integrated, context-aware paradigm. Future progress will depend on addressing key challenges such as data scarcity for low-resource languages, effective multimodal fusion, and the nuanced evaluation of generative KE systems.
[HC-8] chnohybridism: Recursive Semantic Normalization in AI-Mediated Societies
链接: https://arxiv.org/abs/2611.07984
作者: Charalampos M. Liapis,Epameinondas Panagopoulos
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Language-generating AI systems are no longer only objects of linguistic interaction: their outputs increasingly re-enter the communicative and institutional practices from which their training material was drawn. Drawing on social theory, social epistemology, and the philosophy of language, this article develops a sociotechnical account of that re-entry. We introduce recursive semantic normalization: the process through which machine-generated formulations, once recurrently taken up, evaluated, reused, and institutionally retained, participate in the reproduction and transformation of semantic regularities. Under the heading of technohybridism, we argue that the relevant unit of analysis is the hybrid semantic system, a recursively coupled configuration of human practices, computational generation, and institutional uptake. First, we distinguish the concept from neighbouring accounts, including model collapse, looping effects, machine habitus, machine culture, and the epistemology of AI testimony. Second, we develop an analytical model that separates generation, selection, uptake, and persistence and distinguishes machine-attributable change from change that would have occurred anyway. Third, we analyse uptake as routinized, institutionally diffused, and socially stratified practice and state conditions under which machine-derived regularities become norms. Fourth, we identify social-epistemic consequences: the laundering of provenance that weakens epistemic vigilance, the erosion of the cognitive diversity on which collective inquiry depends, and the narrowing of shared hermeneutical resources. An evidential ladder for empirical research follows. The account attributes no normative authority to machine outputs; what changes is the field within which human and institutional uptake operates.
[HC-9] Responsible GenAI Use During Active Armed Conflict: Evidence from Sudans Emergency Response Rooms
链接: https://arxiv.org/abs/2611.07874
作者: Dalia Ali,Yearin Lee,Cheng Yu,Ro Encarnación,Alsanosi Adam,Danaé Metaxa,Ingmar Weber,Orestis Papakyriakopoulos
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Armed conflict can increase the value of Generative AI (GenAI) while undermining the conditions required for its responsible use. Yet the perspectives of community-based humanitarian responders working under these conditions remain underrepresented in responsible AI research. We examine GenAI use in Sudan’s Emergency Response Rooms (ERRs), a community-led humanitarian network, through an exploratory survey of 89 respondents and 15 semi-structured interviews with GenAI users and non-users. War reconfigures GenAI’s role, making it most useful where its outputs are hardest to verify. ERR responders turned to GenAI when conflict made human expertise difficult or dangerous to reach. However, participants cautioned that GenAI could not independently establish rapidly changing local conditions, and that relying on its outputs could lead to harmful decisions when human verification was difficult. Risks arose before GenAI was used. For some volunteers, carrying a phone or accessing the internet could expose them to danger. Others were concerned that sharing sensitive beneficiary or operational information with GenAI could put people at risk. Participants feared that AI-assisted reports could undermine their credibility with donors, whose funding sustains the food and medical aid ERRs provide. Access to GenAI was uneven, and volunteers using it sometimes relied on colleagues with relevant expertise or local knowledge to verify its results. Our findings show that responsible use of GenAI during conflict depends not only on model performance or individual adoption, but also on whether responders can safely access these tools, verify their outputs against local conditions, and recognize when limiting or avoiding GenAI is the more responsible choice.
[HC-10] STAG-Net: Structure-Aware Attention over Landmark Groups for Lightweight Head Pose Estimation
链接: https://arxiv.org/abs/2611.07828
作者: Asiri Gawesha Lindamulage,Sanka Mohottala,Dharshana Kasthurirathna,Nuwan Kodagoda
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Head pose estimation is important for human computer interaction and driver monitoring, but many image based methods remain too costly for resource constrained deployment. We propose STAG-Net, a structure-aware attention over landmark groups for lightweight head Pose estimation that estimates head pose from facial landmarks using movement aware selection, spatial clustering, compact DeepSet encoders, self attention, and a global geometry descriptor. We develop variants using FaceMesh and Face Alignment Network landmarks and evaluate them under cross dataset and subject independent protocols on 300W-LP, AFLW2000, and BIWI. STAG-Net achieves the lowest mean absolute error among all compared landmark based methods in every evaluated setting. On BIWI frames with valid detections, the FAN variant also records the lowest overall error among all compared methods, reaching 3.51^\circ under cross dataset evaluation and 2.39^\circ under subject independent evaluation. The regressors use only 0.012 to 0.043 million parameters and 24.1 to 83.6 thousand FLOPs, about three orders of magnitude fewer FLOPs than FSA-Net. Despite including landmark extraction and regression, the FaceMesh pipeline reaches 322.9 frames per second on a Tesla T4 GPU, nearly twice the 164.5 frames per second of the fastest measured image based baseline, and 111.0 frames per second on CPU. After normalization by dataset size, STAG-Net processes training samples seven to eight times faster than FSA-Net and more than one hundred times faster than HopeNet and TokenHPE, while complete training finishes within 28 minutes. These results show that STAG-Net combines leading landmark based accuracy with substantially lower training and inference cost.
[HC-11] VocabulARy 2.0: A Generative AR System Enabling Personalised 3D Mnemonics for Vocabulary Learning in Context
链接: https://arxiv.org/abs/2611.07798
作者: Karolina Trajkovska,Matjaž Kljun,Maheshya Weerasinghe,Klen Čopič Pucihar
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Mnemonics support vocabulary learning by linking unfamiliar words to familiar concepts through memorable verbal and visual associations. We present a generative AR system that enables learners to create personalised 3D mnemonics by transforming familiar objects in their surroundings into situated mnemonic anchors. The system scans the environment, detects visible objects, and generates bilingual flashcards. After selecting a word, the learner describes a personal association. The system combines this description with an image of the selected object to generate a textured 3D model using generative AI. Learners can then place the model relative to the real object, transforming it into a situated mnemonic anchor.
[HC-12] hinking Is Not Writing: Designing AI Chatbots to Connect Reasoning and Revision
链接: https://arxiv.org/abs/2611.07792
作者: Lucile Favero,Tanja Käser,Juan Antonio Pérez-Ortiz,Nuria Oliver
类目: Human-Computer Interaction (cs.HC)
备注: 21 pages, 11 figures, 2 tables
Abstract:Supporting students’ argumentative writing with AI requires connecting reflection to their revision processes. We present Maike, a Socratic chatbot supporting writing through argument diagnosis, targeted questioning, and a student model across Think and Write modes. In a classroom study, 65 secondary-school students used Maike or a control chatbot sharing the interface, language model, and Think/Write modes. Approximately 40% of students improved their essay scores. However, Maike users engaged more in idea-level dialogue (56% vs. 44%) and moved iteratively between discussion and editing. Transitions from ideas discussed in dialogue to revisions were positively associated with score gains. These findings suggest a design target for AI writing support: coordinating reflection with revision. We contribute an architecture for supporting argumentative reasoning without generating essay text and an approach linking interaction patterns to writing processes and outcomes. As AI becomes embedded in classrooms, understanding these connections is essential to assessing (and designing for) its educational value.
[HC-13] Digital Security Under Compound Threats: Surveillance Threat Reasoning and Habituation Among North Korean Defectors in China
链接: https://arxiv.org/abs/2611.07576
作者: Yibo Meng,Bingyi Liu,Lionel Z. Wang,Hongkun Yang
类目: Computers and Society (cs.CY); Cryptography and Security (cs.CR); Human-Computer Interaction (cs.HC)
备注: Accepted by IEEE SP 27, Lionel Z. Wang is the corresponding author
Abstract:North Korean defectors residing in China face persistent uncertainty regarding digital security and privacy. Through semi-structured interviews with 19 defectors, we examine how individuals understand digital threats and adapt their security behaviors. We identify three interrelated conditions: transnational surveillance pressure from North Korea, identity-linked digital infrastructure and intermediary dependency in China, and threat reasoning shaped by prior surveillance schemas. Participants developed uneven protective practices and followed two trajectories: adaptive habituation driven by knowledge acquisition and erosive habituation driven by the prolonged absence of adverse events. Our findings show how compound threats, attribution uncertainty, and constrained agency shape security reasoning among undocumented users.
[HC-14] ReviewMap: Mapping Evaluative Acts and Automation Boundaries in Peer-Review Comments
链接: https://arxiv.org/abs/2611.07489
作者: Kihyun Na,Injung Kim
类目: Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注: 27 pages, 5 figures, 19 tables
Abstract:Peer-review comments check what a paper does and judge why it matters, but no existing scheme distinguishes these two evaluative acts. We argue that making these acts explicit and measuring how reliably LLMs identify them is a prerequisite for responsible LLM-assisted review analysis: humans need to inspect review judgments and their grounds. This defines a more honest mission for NLP in peer review than full automation. We present ReviewMap, a framework that classifies each comment as assessment of claims and scholarly standards or calibration of contribution value and venue fit, with finer-grained attributes applied conditionally by act type. Applied to 500 ICLR~2026 papers through multi-model LLM consensus, the act layer proves reliably automatable under the operational guideline (inter-model \kappa \geq 0.75 ), supported by five-human validation on 470 development comments, while the attribute layer requires human oversight, defining a measured automation boundary. At the corpus level, 17% of reviewer-paper instances contain no panel-coded calibration, and the panel-coded calibration share is nearly 3\times as high in Strengths as in Weaknesses. This boundary informs three affordances: a review guideline that operationalizes the assessment-then-calibration workflow for reviewers, a per-comment rebuttal strategy scaffold for authors, and a discussion-phase dashboard for area chairs.
[HC-15] Breaking the Modality Barrier: Multimodal Fusion With Missing Modalities for Cybersickness Prediction
链接: https://arxiv.org/abs/2611.07386
作者: Purnata Saha,Rifatul Islam,John Quarles
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Cybersickness is a major barrier to comfortable VR use, and reliable prediction often requires multiple physiological and behavioral sensors. However, existing models typically assume a fixed set of modalities, making them vulnerable when sensors are unavailable and difficult to generalize across datasets with different sensor configurations. We adapt a transformer-based model that handles arbitrary sensor combinations without imputing missing data. Anchor-based contrastive learning maps heterogeneous sensor embeddings into a shared latent space using eye tracking as the anchor modality, while masked self-attention excludes unavailable sensors during fusion. We evaluate our approach on three publicly available cybersickness datasets: this http URL, SIM21, and Blur-Rollercoaster, collected using different VR systems and sensor configurations. Across single-sensor dropout conditions, prediction error changes by less than 0.09 MAE. Joint training across all datasets achieves 0.51 MAE and 0.63 Pearson correlation, substantially outperforming single-dataset models. These results demonstrate robust and generalizable cybersickness prediction across diverse VR systems.
[HC-16] Slide to Report: Continuous Reporting Symptom-Specific Cybersickness in VR
链接: https://arxiv.org/abs/2611.07345
作者: Purnata Saha,Rifatul Islam,Qiliang He,John Quarles
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Cybersickness is often assessed during VR exposure through discrete overall sickness, which cannot capture specific symptoms or their changes over time. Post-exposure questionnaires such as the Simulator Sickness Questionnaire provide symptom-specific measures but cannot be collected during immersion. We investigate whether separating cybersickness into three symptom-specific dimensions – nausea, oculomotor discomfort, and disorientation – during immersion provides more informative continuous measurements. We conducted a between-subjects study with 52 participants in a VR-rollercoaster, comparing a three-slider interface with a single-slider interface. Concurrently, participants also completed an auditory Simon task to measure attention. We additionally collected multimodal set of physiological signals, including GSR, ECG, PPG, eye and head tracking. Results show that each slider correlated most strongly with its corresponding SSQ subscale, whereas the single slider correlated with all SSQ subscales. Using three sliders did not significantly affect usability, workload, or attention. We publicly release the multimodal dataset to support future cybersickness research.
[HC-17] VREgo: Consumer VR Headsets for Learning Robot Manipulation from Egocentric Demonstrations
链接: https://arxiv.org/abs/2611.07272
作者: Liuchuan Yu
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC)
备注:
Abstract:VREgo trains manipulation policies from demonstrations recorded on consumer VR headsets; we instantiate it with EgoKit on one such headset, the PICO 4 Ultra. On-device head and hand tracking and stereo object localisation allow the operator to begin a demonstration at the table. We adapt HumanEgo’s preprocessing to estimate grasp thresholds per recording, update object positions after release, and propagate segmentation masks between keyframes. We evaluate policies in simulation on two tabletop tasks recorded by one operator. For cube placement, we compare PICO and Project Aria recordings of the same 41 layouts. With 40 training demonstrations, the PICO policies placed the cube on the tray in 67, 74, and 60 of 100 evaluation layouts across three training seeds. The corresponding Aria results were 73, 78, and 0. For a two-stage stacking task recorded with the PICO alone, the baseline placed the block on the tray in 94 of 100 layouts and met the end-of-episode criterion for both placements in two. In one diagnostic episode, repeated replanning deferred release beyond the eight executed steps. Demonstration replays also showed that the open gripper could displace the cube before grasping it. After adapting the training trajectories and adding task-specific execution rules that use simulator object positions, the system met both placement criteria in 93.8% of episodes across four evaluation noise seeds. The same 100 layouts informed rule development and evaluation. VREgo is available at this https URL.
[HC-18] Automated Feedback to Peer Tutors Increases Their Talk Moves and Learner Reasoning but Not the Balance of Talk
链接: https://arxiv.org/abs/2611.07072
作者: Dorottya Demszky,Yann Hicke,Joy Yun,Mariah Olson,Max Karpawich,Anees Iqbal,Matt Wu,Jimmy Huang,Akshay Ravikumar,Drew Bent
类目: Computation and Language (cs.CL); Emerging Technologies (cs.ET); Human-Computer Interaction (cs.HC)
备注:
Abstract:Automated feedback on classroom discourse improves instruction in professional and online teaching contexts, and recent work suggests that delivering similar feedback to learners may strengthen these effects by sharing responsibility for participation. Whether this holds in small-group volunteer tutoring has not been tested. We present results from a randomized controlled trial across two four-week SAT mathematics bootcamps on the this http URL platform, involving 1,030 volunteer tutors, 7,443 learners and 6,508 recorded sessions. Tutors were randomly assigned to receive no feedback, automated feedback on their own discourse, or that feedback together with participation feedback delivered to their learners under one of two motivational framings. Feedback to tutors increased eliciting of learner ideas, uptake of learner contributions, and learner reasoning by 8-15%, and the pattern replicated across both bootcamps. However, tutor talk time, which averaged 91.5% in the control group, did not change in any condition, and learner-facing feedback did not further change discourse. Interviews with 17 tutors indicate that they adopted the moves the feedback recommended but were limited by learners who did not respond. Interviews with 13 learners suggest why the learner-facing feedback had little effect: it reported a group-level rather than a learner-level statistic, some learners preferred to listen, and learner-to-learner interaction was rare in these sessions. Automated feedback can change how tutors invite learner thinking, but shifting how much learners talk likely requires changes beyond feedback.
[HC-19] Autonomous AI Agents Systematically Diverge From Human Preferences in Proof Formalization
链接: https://arxiv.org/abs/2611.06995
作者: Bálint Gyevnár,Simon DeDeo
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Artificial intelligence (AI) systems have ushered in a new age of proof abundance, but helping humans understand these proofs requires new attention to the differences in how AI agents and humans pursue a proof-making task. We present “simulated collaboration”, a methodology that allows us to examine two key questions: (1) the relative attention AI agents pay to surface syntax or, conversely, deeper underlying semantics, and (2) how and where AI agent preferences for what to do next diverge from human preferences. On a set of nine major human collaborations in Lean with 4.7M definitions across more than 6K commits, we find that, over the course of proof construction, AI agents (1) are far more strongly drawn than humans to explicitly marked gaps; (2) show a relative aversion to densely connected proof centers and to complex or surprising proof goals; and (3) are relatively attracted to surprising proof bodies, which they tend to regularize toward baseline. Our results highlight important blind spots and potential interventions for future human-AI collaboration in mathematics.
[HC-20] Category differences in similarity between fixation density maps and psychological potential fields for sketches
链接: https://arxiv.org/abs/2611.06930
作者: Naoyuki Awano,Yuki Hayashi
类目: Neurons and Cognition (q-bio.NC); Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
备注:
Abstract:When an object comes into view, humans can instantaneously recognize it if known to them. However, assessing how this recognition takes place remains challenging. A few studies using simple sketches have reported varying degrees of image similarity between the fixation density map (FDM), which visualizes eye movements and durations, and the psychological potential field (PPF), which visualizes part of the visual cognition process, depending on the object category. This study reports how the similarity behaves across three categories—animal, non-animal, and meaningless shape—and under rotated presentation, which is known to increase cognitive load during object recognition. The similarity was quantified through area under the curve (AUC) measures, including a baseline-adjusted AUC that isolates the early rise in similarity from stimulus-dependent baseline levels. A variability score was introduced to assess individual differences, and linear mixed models were used to evaluate category and orientation effects. Results indicated that the baseline-adjusted AUC yielded significant differences among all three categories with large effect sizes, whereas the raw AUC did not distinguish between the animal and non-animal categories due to the influence of sketch representation. No significant Category \times Orientation interaction was observed, indicating that rotated presentation did not produce a category-specific effect. The meaningless shape stimuli yielded consistently low similarity and variability scores, clearly separating them from the two meaningful object categories. These results characterize the behavior of the FDM–PPF similarity under free viewing and indicate that baseline adjustment is necessary to separate category-related variation from stimulus-dependent baseline levels.
计算机视觉
[CV-0] Do You Miss the Nights We Spent Crafting Figures? Rethinking Teaser and Framework Figures in the Age of Generative AI
链接: https://arxiv.org/abs/2611.08486
作者: Xiaoqi Zhao,Youwei Pang,Kailai Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Scientific figures give research ideas a visual form through choices of abstraction, composition, and emphasis. Generative AI offers new ways to realize these choices, raising questions about explanatory clarity and purposeful visual expression. We examine these questions through a visual retrospective of 371 motivation and framework figures from 200 papers. The corpus comprises a quota-based main sample of 160 papers spanning four temporal periods and a separate purposive style-contrast sample of 40 papers. A single multimodal judge evaluates six dimensions of communication and design, with three repeated judgments per figure. We attempt source-constrained revisions of all 371 figures, of which 369 pass the recorded fidelity checks. Original-figure ratings show no uniform decline across periods. Revisions increase mean clarity ratings by 0.428 points and decrease aesthetic coherence ratings by 0.111 points on a five-point scale. Gains in the six-dimension average are larger for main-sample motivation figures than for frameworks. Case analyses show that added explanation can clarify comparisons while leaving typographic hierarchy and visual economy unresolved. These model-based findings support researcher-led design, with AI assisting exploration and authors retaining responsibility for explanatory purpose, composition, and scientific relationships. The accompanying atlas and evaluation records are available at this https URL.
[CV-1] Where and Why is Grounded Reasoning Useful?
链接: https://arxiv.org/abs/2611.08485
作者: David Yan,Le Xue,Forrest Huang,Qiuyi Zhang,Dustin Tran,Andrew M. Dai,Yinfei Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Grounded reasoning, where a model outputs text coordinates of relevant locations in chain-of-thought reasoning, is a common strategy for solving complex visual reasoning problems. However, the benefits and limitations of grounded reasoning are understudied. We study where and why grounded reasoning is most useful. To do so, we introduce a chain-of-thought analysis method to measure how much a model uses its outputted grounding in its thinking trace and perform a detailed ablation study of a grounded reasoning training recipe. We find that grounded reasoning improves counting but shows no significant gain over baselines without coordinate grounding across a wide range of spatial reasoning and perception benchmarks. Using several synthetic tasks, we show that grounded reasoning is most useful when a task requires undescribable references, which are visual concepts that are difficult to describe in language. However, we observe that many existing visual reasoning benchmarks are solvable with only describable text references. To close this gap, we introduce BrickVision, a new benchmark that requires undescribable visual references to solve.
[CV-2] EgoGeneralist: Unifying Egocentric Action-Video Tasks through Video-Native Modeling WWW
链接: https://arxiv.org/abs/2611.08477
作者: Yanlin Jin,Bowen Yang,Cheng Zhang,Yifan Yang,Zhe Li,Jianfei Cai,Kai Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Modern video models possess strong visual priors, yet do not natively represent human action in metric 3D space. We present EgoGeneralist, which adapts pretrained video generators to unify action-conditioned egocentric video generation, pose estimation, and joint action-video generation. Our role-adaptive action renderer provides video-native representations for multidirectional action-video learning, making metric motion usable as both a spatial control signal and a recoverable generation target. Cross-stream action-video coupling establishes close correspondence between metric motion and visual dynamics through lightweight adaptation while preserving the pretrained visual prior. Together, these designs establish a generalist framework spanning five action-video tasks. Despite adaptation only on controlled laboratory recordings, EgoGeneralist generalizes to diverse unseen scenes and interactions with effective action control and visual quality, bridging pretrained visual knowledge and metric 3D human motion.
[CV-3] VFlash: Fast Accurate Visual Planning via Flash Primitive Decoding
链接: https://arxiv.org/abs/2611.08449
作者: Yu-Cheng Chou,Shihao Wang,Jin Huang,Chi-Pin Huang,Chiao-An Yang,Xianzhe Fan,Jing Wang,Zhiqi Li,Yuanguo Kuang,Ruijie Zheng,Johan Bjorck,Sifei Liu,Hang Su,Abhishek Badki,Lei Zhang,Alan Yuille,Linxi "Jim’’ Fan,Jan Kautz,Zhiding Yu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Models for physical AI tasks, such as robotics and autonomous driving, need not reason solely in language. They can use visual plans–points, bounding boxes, and trajectories–to represent spatial goals and anticipate motion. However, autoregressive vision-language models generate these plans one coordinate token at a time, making richer predictions increasingly costly. We introduce VFlash, a method for fast, accurate visual planning through Flash Primitive Decoding. VFlash exploits the known structure of visual primitives to predict their coordinates jointly, using a lightweight drafter that generates an entire primitive in one forward pass. The layout adapts to the primitive’s type, dimensionality, and length, supporting compact boxes and long, variable-length trajectories. Crucially, spatial accuracy need not require reproducing an autoregressive model’s exact token sequence. VFlash therefore trains on annotated primitives and directly accepts structurally valid predictions without target-token agreement. Across the evaluated 2D and 3D robot trajectory tasks, VFlash consistently reduces prediction errors relative to autoregressive decoding. Preliminary single-rollout measurements indicate approximately 25x faster generation for 3D point motion forecasting. A world action model conditioned on VFlash plans achieves 70.2% average success across four MolmoSpaces task suites, compared with 62.6% for the baseline.
[CV-4] Edge-Assisted Crowdsensing for Explainable Cyclist Safety Perception
链接: https://arxiv.org/abs/2611.08441
作者: Hongsen Cao,Chia-Yen Chiang,Mona Jaber,Meihui Wang,Yasmin Fathy
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注:
Abstract:Cyclist fatality rates remain high despite policy support for active travel, and autonomous vehicles add complexity to mixed traffic. Neither crash statistics nor multi-model detection pipelines scale to continuous monitoring of perceived safety along a route. We present an edge-assisted crowdsensing system: handlebar cameras supply street-level frames, a vision-language model (VLM) turns each frame into structured traffic attributes through a single prompt, and an explainable boosting machine (EBM) combines them with OpenStreetMap road infrastructure and on-bike kinematics to predict a crowdsourced 1-9 safety score. Extraction belongs at the network edge: the fastest self-hosted model needs 2.49 s per frame and scoring takes 0.4 ms on a CPU. On 1,362 annotated frames from two London routes, the EBM regressor reaches 72.2 percent binary accuracy and a mean absolute error of 1.011 under five-fold cross-validation. It is within one accuracy point of black-box ensembles. Its per-feature shape functions give exact attribution: physical lane separation is worth +0.97 rating points, a painted lane +0.70 and a construction site -0.65. Planners can therefore price an infrastructure change in rating points, and the results enter cost-benefit analysis directly. We also report inter-annotator agreement (alpha = 0.4695), an attribute audit, and a sensitivity analysis linking attribute error to accuracy. With re-encoding, a 20 s sampling interval serves 1,000 riders at 54 Mbit/s at a cost of 1.4 accuracy points, making continuous monitoring practical.
[CV-5] SM4RT: Cascaded Feed-forward Model for 4D Reconstruction
链接: https://arxiv.org/abs/2611.08436
作者: Pang-Chi Lo,Jhao-Syun Lai,Tsung-Wei Ke
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:We introduce SM4RT, a multi-task cascade feed-forward model for dynamic scene reconstruction from RGB videos. While existing feed-forward approaches typically address geometry reconstruction and motion estimation separately–and often predict motions in a single coordinate frame that must account for camera movement–SM4RT solves these tasks sequentially by first reconstructing 3D geometry and then bootstrapping motion estimation from the predicted structure. By expressing each pixel’s 3D motion in its corresponding camera frame, SM4RT reduces ambiguity arising from camera motion. To address the sparsity of real-world motion labels, SM4RT further incorporates a weakly supervised loss that regularizes unlabeled point motions using labeled point positions at the other time step. We evaluate motion estimation on three dynamic-scene datasets–Kubric, Waymo, and Point Odyssey–and SM4RT achieves the best performance on most metrics compared with prior baselines. We additionally benchmark geometry reconstruction on five dynamic scene datasets and show that SM4RT achieves competitive performance with state-of-the-art methods. Overall, our unified framework demonstrates that combining geometry reconstruction and motion estimation in a cascade not only avoids performance trade-offs but also enhances motion estimation through improved geometric reasoning.
[CV-6] 3R: Training-Free Two-Stage Token Refinement Towards Efficient and Robust VLA Models
链接: https://arxiv.org/abs/2611.08413
作者: Mohammed Hassan,Zhenyang Chen,Zheng Wang,Zhixin Zhu,Tianlong Chen,Yingyan(Celine)Lin,Chaojian Li
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Vision-Language-Action (VLA) models enable end-to-end robot control, but inference is increasingly expensive. Each decision processes hundreds of visual tokens, often across multiple cameras, leading to substantial computation while allowing irrelevant background content to distract action prediction under distribution shift. We argue that these issues stem from two complementary forms of redundancy: many visual tokens are spatially irrelevant, while attention over the remaining tokens is often misallocated. Based on this observation, we propose T3R, a Training-Free Two-Stage Token Refinement framework. In the first stage, segmentation-guided token pruning removes irrelevant visual patches to reduce computation and eliminate distracting context. In the second stage, saliency-guided attention biasing reallocates attention toward action-critical visual regions and language instructions within the retained tokens. Together, pruning produces a cleaner candidate set, while attention biasing further emphasizes the most decision-relevant evidence, yielding a better accuracy-efficiency frontier than either stage alone. On LIBERO-Spatial, T3R removes 85% of scene-camera visual tokens while essentially matching the unpruned success rate, and improves success by up to 7 percentage points under out-of-distribution object-placement perturbations. The timing comparison attributes 7.8 ms per inference call to pruning and attention-bias application. As input resolution increases, LLM-forward profiling shows up to a 3.3x speedup over the unpruned baseline. Evaluations across LIBERO, SIMPLER, and RoboTwin with three VLA backbones show the strongest accuracy-efficiency benefits in multi-camera settings. These results demonstrate that training-free token refinement is a practical approach to efficient and robust VLA inference.
[CV-7] Social Humanoid: Socially Responsive Human-Humanoid Interaction
链接: https://arxiv.org/abs/2611.08407
作者: Vongani H. Maluleke,James Ni,Kie Horiuchi,Lea Wilken,Jiahui Lei,Evonne Ng,Roei Herzig,Jitendra Malik,Trevor Darrell,Angjoo Kanazawa
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: CoRL 2026, Project page: this https URL
Abstract:Humanoid robots are typically trained as solitary agents, leaving them unable to coordinate physically with human partners who naturally anticipate and adjust to one another. To bridge this gap, we present Social Humanoid, a hierarchical system that generates and executes socially responsive whole-body motion on a 29-DoF Unitree G1 conditioned on a human partner’s motion. Its high-level policy performs flexible sequence modeling of human and humanoid motion through a shared tokenized representation, generating reactive humanoid motion and anticipating the partner’s motion when they are occluded, while a low-level whole-body tracking controller executes the generated motion on the robot. To train the system, we introduce a two-phase adaptive retargeting procedure that retargets one interaction partner to the robot while morphing the other partner to preserve contact dynamics, reachability, and relative body scale. We apply this procedure across five motion-capture datasets to produce a new Human-Humanoid Interaction (HHI) benchmark of 260 hours of human-humanoid interactions. Our results demonstrate that Social Humanoid produces coordinated, human-aligned behavior across diverse interaction types in simulation and the real world. Project page: this https URL
[CV-8] Adaptive Low-Light Enhancement for Road Anomaly Detection: A Multi-Architecture Benchmark and Edge Deployment Study
链接: https://arxiv.org/abs/2611.08393
作者: Obed Honour Eje,Oluwatosin Kolade,Glory Bagai,Taliha Abiodun Folorunso
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages
Abstract:Road anomaly detection systems on autonomous vehicles must function reliably under degraded visual conditions including darkness, rain, and Harmattan dust haze. This paper benchmarks four object detection architectures, YOLOv8n, YOLOv12n, RT-DETR, and Faster R-CNN, under physically grounded lighting and weather degradation, and shows that uniform application of the SCI low-light enhancement network is actively harmful at dusk-level lighting (-16.1% to -30.7% mAP@50). We propose a lightweight, training-free luminance-adaptive router that selects between no enhancement, moderate SCI, and denoising plus strong SCI based on per-frame Rec. 601 luminance, recovering most of the dusk penalty while preserving low-light gains, and we validate the full pipeline on NVIDIA Jetson Nano edge hardware.
[CV-9] MOMASPLAT: Dynamic Novel-View Demonstration Augmentation for Mobile Manipulation
链接: https://arxiv.org/abs/2611.08372
作者: Sateesh Kumar,Arnav Balaji,Arpit Bahety,Skand Peri,Georgios Pavlakos,Roberto Martín-Martín
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Visuomotor policies trained through imitation learning are highly sensitive to changes in camera viewpoint. This is particularly problematic for mobile manipulation (MoMa), where robots rarely execute manipulation from the exact same base pose used during demonstration collection due to navigation uncertainty, localization drift, and environmental constraints. Recent approaches use novel-view synthesis (NVS) to augment demonstrations with unseen viewpoints, but they assume static scenes and therefore fail to re-render the manipulated object during the post-contact interaction phase, leaving the most critical parts of the demonstration unaugmented. We present MOMASPLAT, a novel-view demonstration augmentation framework for mobile manipulation that synthesizes full manipulation trajectories from unseen robot base poses, including dynamic object interactions. Our key insight is that standard MoMa data collection naturally provides the multiview observations needed to reconstruct both the static scene and the manipulated object in 3D. MOMASPLAT combines 3D Gaussian Splatting, feed-forward 3D foundation models, and kinematics-based action retargeting to separately reconstruct the scene and manipulated object, render them from novel viewpoints, and generate geometrically consistent synthetic demonstrations paired with transformed actions. We evaluate MOMASPLAT on five real-world mobile manipulation tasks involving both rigid and articulated objects across three scenes, and validate the fidelity of its synthesized demonstrations against ground truth in simulation. MOMASPLAT substantially improves robustness to unseen base poses, more than doubling the average success rate of the strongest baseline and maintaining non-trivial success at large base displacements where all baselines fail, while reducing the need for largescale real-world data collection.
[CV-10] VisionSIM: Beyond Idealized Sensors - A Multi-Modal Simulator for Robot Perception
链接: https://arxiv.org/abs/2611.08368
作者: Sacha Jungerman,Kaustubh Sadekar,Avery Gump,Sriram Narayanan,Shantanu Gupta,Bhuyashi Deka,Srinivasa Narasimhan,Atul Ingle,Mohit Gupta
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Robotic perception increasingly relies on heterogeneous sensor suites combining cameras with emerging modalities (event, SPAD, iToF/dToF, and thermal). Yet existing simulators idealize these sensors, hiding critical modality-specific failure modes that limit real-world deployment. We present VisionSIM, a unified, physically grounded simulation platform that elevates sensor realism to a first-class concern by modeling the raw physical measurement process for diverse conventional and emerging modalities. We expose fine-grained control over physical sensor parameters, reproducing failure modes overlooked by general-purpose simulators: pileup and glare for LiDAR; phase wrapping, saturation, and motion artifacts for iToF; refractory drops for event cameras; and history-dependent emission for thermal cameras. This enables principled cross-modality comparison on identical scenes and trajectories. We release the platform and a large indoor dataset with synchronized raw measurements and rich ground-truth annotations.
[CV-11] OraRefer: Segmentation-Guided Multi-Photograph Fusion for Oral Cancer Screening with Guaranteed Patient-Level Sensitivity
链接: https://arxiv.org/abs/2611.08360
作者: Rupam Mukherjee,Mehakdeep Singh,Shirin Dasgupta,Subhamoy Mandal
类目: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注:
Abstract:Oral cancer and the oral potentially malignant disorders (OPMD) that can precede it are visible on the oral mucosa, which makes smartphone photographs a promising tool for screening. Deep learning models for this task classify individual photographs. However, screening refers patients, not photographs: a patient is usually photographed several times, and the referral decision must miss no more than a set share of patients with disease. We propose OraRefer, a patient-level screening model that reads each photograph with a segmentation-guided foundation model that conditions its class decision on the predicted lesion, combines all photographs of a patient with risk factors, and refers through a threshold with a finite-sample sensitivity guarantee. The model localizes the mouth itself, so inference needs no annotation. On 3000 photographs of 714 patients, patient-level fusion reaches an AUROC of 0.919 and outperforms attention-based multiple-instance learning. At a 90% sensitivity target, OraRefer detects 95% of patients with OPMD or cancer and misses 3 of 61 patients with cancer, at a specificity of 0.610. Because the threshold is calibrated without retraining, each clinic can set its own guaranteed operating point on a local calibration sample.
[CV-12] USB Gym: A Gymnasium Testbed for Closed-Loop Ultrasound View Matching
链接: https://arxiv.org/abs/2611.08352
作者: Thanapong Intharah,Zhichao Gao,Yiliang Liu,Hao Dong
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注: 75 pages; supplementary material (62 pages) as an ancillary file. Submitted to Computer Methods and Programs in Biomedicine. Code: this https URL
Abstract:Background and Objective: Our main goal is ultrasound guidance: moving a probe until its image matches a target view. Published methods for each part of a guide are evaluated in separate simulators. We asked whether these methods, combined into one guide, are good enough. Methods: We built USB Gym, a testbed whose standard engine simulates ultrasound from public patient CT, with a network drawing organ textures from CT labels. Its guide has three replaceable parts (slots), an organ segmenter, a controller and a reward, which received 20 published methods, each run on the same 60 episodes of six held-out patients. In each slot, a published method replaced ours only if detectably better on these patients. The better of this combination and our original guide on 20 new CTs is the benchmark baseline. A pose oracle, reading the exact probe pose available only in simulation, set the reference. The results below are exploratory. Results: On these CTs, the probe started on average 16.0 mm from the target view; the pose oracle brought it to 1.5 mm. Of the results known when the combination was fixed, only an image-similarity reward (SSIM) was slightly but detectably better, so the combination was our guide with that reward. This combination was the benchmark baseline, ending at 13.7 mm (16% of the oracle’s progress). Conclusions: Combining untuned published methods with ours did not guide the probe near the target view; the task remains under-explored. The testbed accepts new methods; no guide was tested on hardware. Comments: 75 pages; supplementary material (62 pages) as an ancillary file. Submitted to Computer Methods and Programs in Biomedicine. Code: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO) Cite as: arXiv:2611.08352 [cs.CV] (or arXiv:2611.08352v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.08352 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Thanapong Intharah Ph.D. [view email] [v1] Thu, 5 Nov 2026 17:34:53 UTC (2,652 KB)
[CV-13] 4D Point Splatting for mmWave Radar Micro-Doppler Differentiable Rendering
链接: https://arxiv.org/abs/2611.08342
作者: Adnan Armouti,Guanyue Qian,Rajalakshmi Nandakumar
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Machine Learning (cs.LG); Signal Processing (eess.SP)
备注: Under Review
Abstract:Millimeter-wave radar resolves the micrometer displacements of a moving body as micro-Doppler, enabling gait, gesture, activity, pose and vital-sign sensing. Each task relies on a downstream model trained on real data, yet radar datasets lag orders of magnitude behind vision and language. Brute-force collection will not close that gap, and existing synthesis methods (forward-only simulators, learned generators and static-scene inverse renderers) cannot fit a real capture of a moving body, leaving a sim2real gap. We present 4D Point Splatting (4DPS), a physics-based differentiable renderer that fits an explicit moving scene to a real capture. Using the synchronized camera to initialize the room and body, we represent the scene as oriented points carrying material, normal and velocity, and splat each point into range and Doppler bins. Our output is the raw complex return a real radar records, so we apply the same signal processing on the rendered data to obtain any representation for training any SOTA downstream model. We fit 235 sequences on HuPR and 100 on RT-Pose, two datasets with different radars. Our reconstructions set the state of the art, with median correlations of 0.940 and 0.830 on range-azimuth, 0.829 and 0.695 on range-Doppler, and 0.988 and 0.966 on micro-Doppler. Our renders also set the state of the art in downstream sim2real gap. On RT-Pose, a pose model trained on them reaches 21.9 cm MPJPE on real data, against 16.7 cm for real training and 127.6 cm for the RF-Genesis baseline, where both the rendered- and real-trained models learn nearly the same features. Our custom CUDA kernels fit a sequence in 5.7 minutes on HuPR and 12.9 on RT-Pose on one H100, making dataset-scale inverse rendering of mmWave radar possible for the first time. This work can enable synthetic data generation for any human-centered downstream task and model, and scene editing for data augmentation.
[CV-14] InstantHand: Feed-Forward 3D Gaussian Splatting for Animatable Hand Avatar Reconstruction from a Single Image NEURIPS2026
链接: https://arxiv.org/abs/2611.08330
作者: Chanho Kim,Muhammad Salman Ali,Kyeonghwan Gwak,Muhammad Shaheryar,Incheol Park,Jeongwan On,Seungryul Baek
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026. Project page: this https URL
Abstract:Hand avatar reconstruction typically relies on multi-view capture or per-subject optimization, limiting its scalability. We introduce InstantHand, a feed-forward framework that constructs an animatable 3D Gaussian hand avatar from a single image using off-the-shelf MANO, camera, and mask estimates, without test-time optimization. Image-averaged losses provide insufficient supervision for alpha leakage, which thickens thin finger contours. To address this issue, we propose a Soft Alpha Band regularizer that concentrates supervision on silhouette boundaries. Furthermore, colors regressed from image features can drift toward a learned prior even in observed regions. Our Reliability-gated Appearance Transfer preserves observed appearance by transferring source colors according to each Gaussian’s rendering contribution, while completing unobserved regions using the learned prior conditioned on global source identity. InstantHand constructs an avatar in 2.84 seconds and renders at 161 FPS, matching per-subject-optimized baselines on InterHand2.6M and remaining competitive in zero-shot evaluation on HanCo.
[CV-15] ManimCode: Learning Visual Self-Correction for Educational Animation
链接: https://arxiv.org/abs/2611.08294
作者: Zhengyang Liang,Lizi Liao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Manim supports precise, programmatic video generation, but a successfully rendered educational animation can still have visual defects, missing explanations, or incorrect content. Rendering does not directly identify the code changes needed to correct these visual failures, so feedback-driven revision is needed. Existing agentic frameworks use iterative rendering and revision but typically rely on large models, often proprietary. Existing Manim datasets primarily contain final programs and often lack quality verification and the trajectories of feedback and revisions needed to learn this refinement process. We introduce ManimData and ManimCode, a data and training framework for feedback-driven educational animation generation. We construct ManimData through a render-guided pipeline that iteratively generates, executes, reviews, and revises Manim programs with retrieved API documentation. Its 24,034 render-verified, quality-filtered records preserve intermediate programs, execution errors, visual critiques, and corrective revisions. We use role-conditioned supervised fine-tuning on these trajectories to train ManimCode, a single 4B model, to write, review, and revise animation programs. To improve correction of errors that differ from those in the collected trajectories, we then train the Reviser with reinforcement learning on self-generated states, using rendering outcomes and reference-free quality feedback. ManimCode achieves 96.6% and 97.0% render success on ManimData-test and ManimBench, compared with 52.6% and 82.0% for the untrained 4B model under the same inference setup, while also improving video quality scores from 2.48 to 2.94 and from 3.00 to 3.44. Integrating ManimCode into two long-form educational video generation systems increases completed videos from 19/117 to 107/117 and from 53/239 to 239/239, alongside improvements in educational effectiveness and visual quality.
[CV-16] Hierarchical Incremental Beta Splatting for Online Reconstruction WACV2027
链接: https://arxiv.org/abs/2611.08288
作者: Linus Franke
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: accepted to WACV 2027, project page: this https URL
Abstract:Online 3D Gaussian Splatting reconstructs a scene during capture: each keyframe spawns primitives once and refines them for only a few dozen iterations, leaving little chance to fix poor initialization or missing detail. Gaussians are also inherently soft, needing many overlaps for solid surfaces. We present Hierarchical Incremental Beta Splatting (HIBS), which brings beta kernels to incremental reconstruction. Their shape parameter interpolates between a soft, Gaussian-like profile and a flat-topped, near-opaque surfel, acting as a confidence dial for geometry. Our confidence-aware spawning uses geometric observability and multi-view depth agreement to initialize both opacity and shape, producing solid primitives where geometry is verifiable and soft, correctable primitives where it is not. As online densification, error-driven relocation recycles stale primitives into photometric-error hotspots. Bounded beta kernel support enables exact tile culling and faster rasterization, and a closed-form, shape-preserving merging formulation provides a level-of-detail (LOD) hierarchy for fast optimization and real-time rendering. On Mip-NeRF 360 in the on-the-fly protocol, HIBS improves PSNR by up to 0.8 dB while reducing optimization time by up to 40%. Project page: this https URL
[CV-17] Interactive Visual Gym: Evaluating Visual Interaction Models with Simulated Users in Interactive Worlds
链接: https://arxiv.org/abs/2611.08286
作者: Chenhao Zheng,Zixian Ma,Kai Li,Kanchana Ranasinghe,Jieyu Zhang,Artemis Panagopoulou,Weikai Huang,Jianxu Shangguan,Yuanyuan Tan,Silvio Savarese,Juan Carlos Niebles,Ranjay Krishna,Ran Xu,Honglu Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 39 pages, including references and appendices
Abstract:Visual interaction models that support users during ongoing physical tasks require evaluation in closed loop, where assistant responses can change subsequent user behavior and task outcomes. Existing benchmarks largely rely on offline replay and therefore cannot capture these effects. We introduce Interactive Visual Gym, a framework for closed-loop evaluation and training of visual interaction models through joint simulation of executable task environments and human users. We contribute three simulators that recreate everyday tasks and human-aligned action spaces across cooking, household manipulation, and mobile-interface domains, together with a persona-conditioned user simulation harness that collaborates with the assistant while incorporating controlled execution errors and other out-of-plan behaviors. We instantiate the framework as IVG-Bench, a benchmark with 315 interaction scenarios. Evaluating state-of-the-art VLMs as visual interaction models reveals substantial limitations: the best-performing model, Gemini 4 Argon, completes only 53.4% of tasks within the time limit, exhibits significant weaknesses in factual grounding and actionable guidance, and is often too slow to prevent errors when response latency is considered. We verify our evaluation protocol with a human study which finds 78.5% agreement between our interaction-quality metric and participant preferences. We further show that the environments provide useful learning signals: in a proof-of-concept GRPO experiment, closed-loop training improves both task success and interaction quality for Qwen3.5-4B model in our cooking environment.
[CV-18] Aligning Policy Encoders with a Shape Teacher
链接: https://arxiv.org/abs/2611.08275
作者: Madelon Bernardy,Mohammad Mahdi Derakhshani,David M Knigge,Riccardo Valperga,Artem Moskalev
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:3D-aware representations can improve robotic manipulation by providing geometric information that is robust to changes in viewpoint, object position, and occlusion. However, using 3D inputs at deployment requires additional sensing and processing, making image-based policies more practical. We ask whether an image-based policy can acquire 3D shape information during training while retaining a purely image-based interface at deployment. We propose shape-teacher alignment, a training-time mechanism that uses a pretrained 3D shape representation as an alignment target for an image-based policy encoder. Specifically, we use Point-MAE features from individual objects as targets for corresponding image patches, transferring geometric structure to the policy without changing its deployment inputs or architecture. On LIBERO, shape-teacher alignment produces image representations that are more structured around object shape and corresponding parts, while substantially accelerating policy convergence and improving success rates. For pretrained policies, convergence is accelerated by 4.7x and 5.6x on the Object and Spatial suites, respectively, without requiring the shape teacher or depth input at deployment.
[CV-19] DART: Domain-Agnostic Residual Transfer for Generalist Anomaly Detection NEURIPS2026
链接: https://arxiv.org/abs/2611.08271
作者: Muhammad Aqeel,Maham Nazir,Marco Cristani,Francesco Setti
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted to NeurIPS 2026
Abstract:Generalist Anomaly Detection (GAD) aims to detect anomalies in unseen domains using only a few reference samples. Recent normal-abnormal guided methods improve this setting by exploiting abnormal references, but their residuals often mix transferable anomaly cues with domain-specific appearance factors such as texture, geometry, and acquisition conditions. This residual domain entanglement limits cross-domain transfer and may produce spurious activations on novel domains. We propose DART, a plug-in framework for normal-abnormal guided GAD that learns to separate residuals into anomaly-relevant and domain-specific components. DART combines a Residual Disentanglement Module, which preserves anomaly-discriminative information while suppressing domain information, with Cross-Domain Meta-Training, which forms episodes across source domains to make the separation effective. The method is architecture-agnostic and can be inserted between residual computation and the detection head of existing GAD pipelines. Experiments on MVTec AD, VisA, and BraTS show consistent improvements over multiple base methods, with the largest gains under stronger domain shifts. this https URL
[CV-20] SplatAnything: Gaussian Spacing Normalization for Feed-forward 3DGS across Arbitrary Scenes
链接: https://arxiv.org/abs/2611.08264
作者: Hao Yu,Zhenwei Wang,Yongzhen Hu,Yuning Gong,Junta Wu,Xiaowei Zhou,Hujun Bao,Tengfei Wang,Chunchao Guo,Sida Peng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We identify a key limitation of existing feed-forward 3DGS methods: Gaussian attributes are typically predicted in a radial-distance-normalized space that canonicalizes a global radial statistic of the reconstructed geometry, while the spacing between neighboring Gaussian centers can still vary substantially across scene scales and camera configurations. This mismatch induces systematic shifts in Gaussian-scale targets, leading to ambiguous supervision. To address this, we propose SplatAnything with Gaussian Spacing Normalization (GSNorm), which canonicalizes a robust scene-level statistic of Gaussian spacing to provide a more consistent spatial reference for Gaussian prediction. To rigorously assess our method, we further establish a systematic benchmark spanning diverse datasets, input sparsity levels, and challenging extrapolated viewpoints. Extensive experiments demonstrate that SplatAnything achieves state-of-the-art performance, with particularly strong gains under substantial variations in scene scale and viewpoint.
[CV-21] FewStepRL: Fine-tuning Few-step Diffusion Models with Reinforcement Learning
链接: https://arxiv.org/abs/2611.08240
作者: Haitao Lin,Minsi Ren,Yang Li,Xingtong Ge,Zheyuan Hu,Peiyan Hu,Tailin Wu
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 86pages, 12 figures, technical report
Abstract:Reinforcement learning (RL) has become a standard post-training method for diffusion models, but most RL objectives roll out the current policy at every iteration, and with a many-step sampler these rollouts typically dominate the cost of training. Few-step models cut this cost, yet they model an average velocity over an interval, whereas existing RL objectives are built for an instantaneous velocity with a many-step sampler. We adapt RL objectives to few-step models systematically in three phases, namely a transformation, a decomposition and an adaptation. The transformation to x -space expresses each objective we consider through a common policy prediction, at the clean end by default and at an explicitly marked level for transition or flow-map objectives, so that objectives stated over densities, velocities or rewards become comparable. The decomposition writes the gradient of each of these objectives as the sum of a self-regression toward a reference and a reward augmentation, so that a method reduces, at the gradient level, to a choice of anchors and weights. The adaptation computes each anchor from a few-step model at the end of its transport or along its tangent, with the instantaneous velocity recovered from the network itself or from an auxiliary network. This yields few-step versions of FlowGRPO, AWM, DiffusionNFT, DiffusionOPSD and ReFL in one modular codebase. In our experiments, reward-gradient-based methods are the most effective when the reward is differentiable and its gradient is affordable; otherwise DiffusionNFT-style methods generally reach high rewards in relatively few iterations despite a sometimes higher per-iteration cost. Extra inference steps show no systematic aggregate benefit in our experiments, although selected outputs visibly improve with two to four times as many sampling steps as in training.
[CV-22] Beyond Quadratic Attention: Separable Gaussian Operators for ODE Vision Transformers
链接: https://arxiv.org/abs/2611.08236
作者: Carlos Boned Riera,David Romero Sanchez,Oriol Ramos Terrades
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Self-attention carries a quadratic cost in token count, which limits how far Vision Transformers can scale with input resolution. We introduce SSoG, a separable sum-of-Gaussians spatial operator that replaces dense self-attention with a learned mixture of axis-factorized Gaussian kernels, reducing the per-layer cost from \mathcalO(N^2) to subquadratic in N . Unlike dot-product self-attention, whose Lipschitz constant is unbounded outside a compact input domain, SSoG with fixed kernels is globally Lipschitz continuous, which, via the Picard–Lindelöf theorem, guarantees that the continuous-depth variant has a unique solution on the whole integration interval, a global guarantee that dense-attention counterparts lack in general. On CIFAR-100 at 224\times224 , in a matched discrete-Transformer setting and in the continuous setting, SSoG improves Top-1 accuracy by 2.2–4.1 points across three model widths while using 8.8–16.6% fewer parameters, with depth and training recipe held fixed. In a looped-Transformer setting, SSoG improves Top-1 accuracy by 5.4–8.1 points. We further probe inference-time scaling by evaluating on inputs up to 1024\times1024 at matched batch size: in the continuous-depth (Neural ODE) architecture, SSoG reaches up to 154% higher throughput and over 60% lower peak memory than dense attention, with the gap widening as resolution grows, consistent with the difference in asymptotic cost. In the discrete and looped settings, SSoG also achieves over 30% higher throughput than dense attention at this resolution. Together, these results establish SSoG as a practical, parameter-efficient, and theoretically grounded alternative to dense attention across discrete, weight-shared, and continuous-depth Transformer variants.
[CV-23] Multimodal Hierarchical Mixture-of-Experts for 3D Medical Vision-Language Modeling WACV2027
链接: https://arxiv.org/abs/2611.08185
作者: Gizem Mert,Szymon Płotka,Pedro R. A. S. Bassi,Wenxuan Li,Zongwei Zhou,Jarosław B. Ćwikła,Arkadiusz Sitek,Ewa Szczurek
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at WACV 2027. Copyright 2027 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses
Abstract:Medical 3D scans and reports are inherently hierarchical, with diagnostic cues concentrated in specific anatomical regions across multiple levels of organization. Yet, most medical vision-language models ignore this structure and fail to focus on clinically relevant anatomy. We introduce MedHiVE, which combines hierarchical vision-language Soft Mixture-of-Experts (SMoE) with contrastive learning to jointly model fine-grained anatomy and global scan context. To our knowledge, MedHiVE is the first framework to route both vision and text tokens through a two-level SMoE hierarchy organized by anatomical scale, coupling this with report-derived anatomy-level contrastive alignment for 3D medical imaging. This replaces the direct region-text alignment of statically pooled features and vision-only hierarchical routing of prior work with dynamic, slot-based cross-modal information flow across anatomical scales. MedHiVE operates via a two-level, learnable token-routing mechanism, where local image and text features are first routed through a local SMoE and then fused by a global SMoE into unified multimodal representations. Contrastive learning aligns both anatomy-level and global vision-language representations in a unified space. Pre-trained on 25,000+ CT scans with anatomical supervision, MedHiVE consistently surpasses 11 state-of-the-art methods across 6 datasets on 3 tasks: segmentation, classification, and detection, demonstrating the importance of hierarchical modeling. The code for MedHiVE is at this http URL.
[CV-24] Improving Context-Shift Robustness of Convolutional Networks via Context-Regularized Cross-Entropy NEURIPS2026
链接: https://arxiv.org/abs/2611.08157
作者: Jinen Setpal,Chaoyue Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: accepted to NeurIPS 2026
Abstract:Image classification models often encounter objects in diverse and unpredictable contexts at test time, which may differ substantially from those seen during training. This phenomenon, termed context shift, exposes a key fragility in neural networks: when models rely on non-causal contextual cues for prediction, shifts in context can lead to significant performance degradation. In this paper, we propose a novel training framework to improve the robustness of convolutional neural networks under context shift. Our approach leverages spatial feature attribution to guide models toward making predictions that rely less on contextual regions and more on object-relevant features. As a first step, recognizing a theoretical limitation of existing feature attribution methods we introduce a new variant, ContrastiveCAMs, which produces more faithful attribution maps of model predictions. Building on ContrastiveCAMs, we further propose Context-Regularized Cross-Entropy (CR-CE), a modification of the standard cross-entropy loss that regularizes the model’s attention by suppressing the influence of contextual regions, thereby improving robustness to context shift. We evaluate the effectiveness of our approach on several medium-to-large scale datasets (Waterbirds, Spawrious, ImageNet/ImageNet-BG, Hard-ImageNet), and report consistent improvements in context-shift robustness.
[CV-25] La-JEPA: Compact Robot Policies with Latent World Modeling
链接: https://arxiv.org/abs/2611.08156
作者: Mateo Clémente,George Z. Xue,Amin Abyaneh,Liheng Ma,Rui Heng Yang,Tongtong Cao,Zhanguang Zhang,Yingxue Zhang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 30 pages, 6 figures. Code: this https URL
Abstract:World-action models adapt pretrained video generators, and vision-language-action models build on pretrained vision-language backbones; both tie policy architecture and inference cost to large transformers designed for generation rather than control. It remains unclear how much these generation-oriented choices affect control. In this work, we introduce Latent Action JEPA (La-JEPA), a compact world-action model whose policy transformer is trained from scratch in the feature space of a pretrained V-JEPA 2.1 encoder. Following Fast-WAM, La-JEPA learns to predict future features jointly with actions during training and generates only actions at inference. Among policies without embodied pretraining, La-JEPA achieves state-of-the-art performance in clean and randomized RoboTwin 2.0 scenes and on LIBERO-Plus with correct instructions, while using far fewer parameters than the strongest competing policies. It runs substantially faster than Fast-WAM at comparable performance and outperforms the closest JEPA-based methods, LeapBot-WA and JEPA-WAM, on LIBERO-Plus and RoboTwin 2.0 without their pretrained policy backbones. On four real-robot tasks, its average success is comparable to pi0.5 and above non-embodied-pretrained baselines. Through a design study of the choices behind La-JEPA, we show that JEPA features outperform video VAE latents, most clearly out of distribution, and that training-time future prediction choices significantly impact the policy even when future prediction is skipped at inference. We correct the LIBERO-Plus instructions, as the officially released evaluator corrupts them for most variants, and re-evaluate several strong open baselines. Without a generative pretrained policy backbone, La-JEPA is better and faster than far larger policies, suggesting that robot foundation models may benefit from architectures designed for control rather than inherited from generation.
[CV-26] Interpretable Multimodal MRI-Clinical Fusion for Functional Outcome Prediction After Acute Ischemic Stroke
链接: https://arxiv.org/abs/2611.08154
作者: Sina Raeisadigh,Henning Müller,Abderrahmane Hedjoudje
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
备注: 5 pages, 2 figures, 1 table
Abstract:This study evaluates the added prognostic value of day-1 multimodal MRI for predicting three-month functional outcome after acute ischemic stroke (AIS). Sixty-six AIS patients with ADC, DWI, FLAIR, and clinical data were analyzed. Three-dimensional pretrained ResNet embeddings were extracted from each MRI modality, reduced using fold-isolated principal component analysis, and combined with structured clinical variables and ADC-derived lesion biomarkers. Linear support vector machines were evaluated using eight-fold stratified group cross-validation across MRI modality combinations. Clinical-only modeling achieved strong performance (AUC = 0.899 \pm 0.151 ), while MRI-only models showed limited discrimination (AUC \approx 0.67 - 0.73 ). Among multimodal configurations, FLAIR-clinical fusion achieved the highest mean cross-validated performance (AUC = 0.948 \pm 0.061 ), followed by ADC-clinical fusion (AUC = 0.941 \pm 0.094 ). Full ADC-DWI-FLAIR fusion did not consistently outperform single-modality clinical fusion. Feature-importance analysis identified age, NIHSS subscores, ADC lesion burden, and learned FLAIR representations as key contributors, while qualitative FLAIR activation maps showed responses over lesion-like regions. These findings suggest that compact day-1 FLAIR and ADC representations provide complementary prognostic information beyond clinical variables and support interpretable, data-efficient outcome prediction in AIS.
[CV-27] Detecting Weather Fronts from Satellite Observations: What Does Reanalysis Add? NEURIPS2026
链接: https://arxiv.org/abs/2611.08148
作者: Ioannis Dravilas,Miruna Stoicescu,Roope Tervo,Carlos Horn,Oriol Hinojo Comellas,Sina Montazeri,Andrea Meraner,Ana Lucic
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Atmospheric and Oceanic Physics (physics.ao-ph)
备注: Accepted to the 2nd Workshop on Advances in Representation Learning for Earth Observation at NeurIPS 2026
Abstract:Weather fronts are transition regions between air masses with contrasting thermodynamic properties and are associated with mid-latitude precipitation and severe weather. We investigate machine learning (ML) methods for detecting weather fronts and quantify the contribution of reanalysis data to performance. We compare two configurations of the same ML model for pixel-wise frontal classification over the North Atlantic and Europe: (1) using only geostationary infrared satellite observations, and (2) additionally using 28 ERA5 reanalysis fields. The satellite-only model achieves a neighbourhood critical success index at a 150 km tolerance (CSI@150) of 0.402 , which lies within the range of agreement between analyses from different meteorological services. The satellite + ERA5 model achieves 0.575 , an improvement of 0.173 . Our analysis shows that low-level winds are the least predictable from the satellite representation ( R^2 = 0.23 , compared with 0.88 for temperature), while removing winds or moisture causes the largest decreases in detection performance.
[CV-28] MCUFlowNet: Hardware-Aware Neural Architecture Search for Optical Flow on Microcontrollers ISCAS2027
链接: https://arxiv.org/abs/2611.08147
作者: Enmin Lin,Jona Beysens
类目: Computer Vision and Pattern Recognition (cs.CV); Hardware Architecture (cs.AR)
备注: Submitted to IEEE ISCAS 2027 (under review). 5 pages
Abstract:Dense optical flow allows small drones to avoid obstacles. However, few dense optical flow networks are designed for microcontrollers (MCUs) with neural accelerators, whose on-chip SRAM limits the input resolution. We present PACE-NAS, a macro-to-micro neural architecture search framework that decouples memory from latency: a macro design fixes the backbone and its peak SRAM, and a micro design searches operators and depths to trade accuracy against compiler-estimated latency. We apply it to EdgeFlowNet on a Grove Vision AI V2 board (Ethos-U55), where profiling locates the peak SRAM at the full-resolution output head. Replacing transposed convolutions with bilinear upsampling and convolutions lowers the peak SRAM by 18%, and two lightweight context modules lower the endpoint error (EPE) without raising the peak. All searched subnets share this peak, so the search needs no memory constraint. We name the Pareto-optimal models MCUFlowNet; the fastest variant, MCUFlowNet-S, runs 1.76x faster than a retrained EdgeFlowNet at a 1.6% higher EPE and, in simulated forest flights at 2 m/s, reaches the goal without collision in 51.5% of trials, against 41.8%. PACE-NAS also applies to semantic segmentation, where a searched PIDNet-S variant runs 1.17x faster at similar accuracy. Code is available at this https URL.
[CV-29] Fault-Guided Cross-Layer Prediction for Unsupervised Anomaly Detection
链接: https://arxiv.org/abs/2611.08140
作者: Tao Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages including appendix; 4 figures and 21 tables. Preprint
Abstract:Feature reconstruction is widely used for unsupervised industrial anomaly detection in the single-class setting, where only normal images from one product category are available for training. We propose a single-class detector that asks what information one block can pass to another. We treat the depth of a frozen vision transformer as a sequence of discrete transitions. A directed predictor learns a block-to-block hidden-state transition pointwise over tokens. Adjacent-layer cosine similarity exposes four sharp fault lines in DINOv2 ViT-L, separating stable depth segments. The deepest blocks retain semantic context but lose local detail; a prediction from deep or shallow sources to sensitive targets therefore exposes the information that cannot be transported. The fault structure fixes the source and target roles, and the same rule transfers to ViT-B and RADIO-B. We use a bounded foreground prior calibrated solely on training images: it attenuates background evidence without also erasing defects outside the estimated foreground. Using one predictor architecture and a common training-only calibration protocol, image AUROC is 99.75 on MVTec-AD, 99.42 on VisA and 94.27 on Real-IAD. On Real-IAD the object-level AUROC is 96.26. At 256x256 pixel resolution, P-AUROC/P-AUPRO are 98.46/95.97 on MVTec-AD, 99.21/96.12 on VisA and 99.26/96.59 on Real-IAD.
[CV-30] Describe Your Facial Expressions by Linking Image Encoders and Large Language Models BMVC2023
链接: https://arxiv.org/abs/2611.08138
作者: Yujian Yuan,Jiabei Zeng,Shiguang Shan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: BMVC2023 Oral
Abstract:This paper presents a novel task of describing human facial expressions of a facial image in natural language, which captures the nuances of facial actions and emotional states beyond traditional emotion categories or facial action units (AUs). To achieve the facial expression captioning model, we propose a three-stage training framework that trains a vision-to-language model using synthesized image-text pairs and the BLIP-2 pre-training techniques. To overcome the challenge of missing training image-text pairs for facial expression captioning, we propose a strategy that involves synthesizing and combining captions using GPT-3.5 and existing annotations on either emotion categories or AUs. Experiments demonstrate the effectiveness of our method in generating captions that describe details of facial actions and emotions, as well as the inferential relationship between them, even when those emotions are not present in the training data. It is also demonstrated that the vision-to-language task enhances the performance of the intermediate visual features on both AU detection and emotion classification tasks.
[CV-31] How Far Can Untrained Neural Networks Go in Anomaly Detection? From Tables to Pixels
链接: https://arxiv.org/abs/2611.08136
作者: Hyojun Yi,Geunyoung Kim,Hogeon Seo,Seunghyoung Ryu
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 13 figures
Abstract:Anomaly detection (AD) is often framed as a representation-learning problem, where strong performance is expected to require learned features or AD-specific optimization. However, this view can overlook the intrinsic separability of anomalies from normal references, even in the absence of learned representations. We therefore investigate how far untrained neural networks can go in anomaly detection. To this end, we propose PRISM, a framework combining fixed random nonlinear features with normal-reference Mahalanobis scoring for tabular detection and dense visual localization. Across 47 tabular datasets in ADBench, PRISM-T achieves the highest mean AUROC (86.4%) and the second-highest mean AUPRC (65.9%) among the compared methods. On MVTec AD and VisA, covering 27 visual categories, PRISM-V achieves mean image-level and pixel-level AUROCs of 90.4% and 96.4%, respectively. Notably, both models achieve these results without neural weight optimization or pretrained representations. A theoretical analysis further characterizes how nonlinear transformations affect the mean and covariance contributions to the Mahalanobis score gap, while empirical analyses examine the resulting score separability. These findings establish PRISM as both a competitive anomaly detector and a prior-to-learning baseline for assessing the added value of learned representations in anomaly detection.
[CV-32] FreeOcc: Decoupling Ego-Motion for Efficient 4D Occupancy Forecasting via Continuous Flow Matching
链接: https://arxiv.org/abs/2611.08128
作者: Zeping Zhang,Zhuoya Zhao,Samy Metari,Robert Laganière
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Current 4D occupancy world models use heavy cross-attention or deep recurrence to disentangle ego-motion from scene dynamics. This requires significant computational capacity. However, this effort is largely redundant since the ego-vehicle’s intended future motion is provided by the downstream motion planner. We present FreeOcc, a lightweight continuous-time world model that introduces explicit SE(2) geometric decoupling into a conditional flow-matching framework. By mathematically factoring out self-motion, we create a purely dynamic feature space. This space supports attention-free temporal fusion and allows physics-constrained ODEs to operate strictly on residual agent kinematics. To maximize deployment efficiency, FreeOcc introduces an uncertainty-aware blending mechanism. This routes well-observed voxels to a single-pass decoder, reserving the ODE solver exclusively for ambiguous, high-uncertainty regions. Requiring only 28.3M parameters-roughly 40% of comparable models-our optimal ensemble mode (FreeOcc-E) runs at 27 FPS. On nuScenes, FreeOcc achieves state-of-the-art forecasting performance with a 0.47 m Near Field Chamfer Distance (NFCD) and 32.1% mIoU. Furthermore, the model transfers zero-shot to SemanticKITTI. It also proves highly robust, retaining 92.8% of its ideal mIoU under significant sensor packet loss and timestamp jitter.
[CV-33] ProtoGestalt: Interpretable Prototypes for Facial Phenotyping and Diagnosis of Rare Genetic Disorders
链接: https://arxiv.org/abs/2611.08125
作者: Ranga Baminiwatte,Alexander Hustinx,Fabio Hellmann,Aaron J Masino
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Rare genetic disorders are difficult to diagnose due to their diversity and overlapping clinical features. Facial morphology can provide diagnostic clues, yet most diagnostic models offer limited interpretability and are not designed to support broad screening for undiagnosed disease. We present ProtoGestalt, an interpretable prototype-based framework for facial phenotyping, rare-disease screening, and syndrome identification. The method extends prototype learning by constraining learned visual features to select facial regions, allowing predictions to be explained through localized and additive evidence. We assess ProtoGestalt’s ability to distinguish syndromic from unaffected faces on multiple control datasets and class prevalence values. We further evaluate it’s diagnostic performance on syndromes observed during training. The learned prototype-based representations provide significant separation between syndromic and case-matched control images and retain useful performance as the proportion of controls increases. The framework also demonstrates the ability to differentiate among syndromes. Although reasonably accurate specialized models for facial classification of rare disorders already exist, ProtoGestalt offers an interpretable alternative that links predictions to anatomically meaningful facial regions, potentially supporting clinical decision-making and the discovery of phenotype-gene associations.
[CV-34] Reinforcing End-to-End Autonomous Driving Safety via Negative-Aware Planning in a Unified World-Action Model
链接: https://arxiv.org/abs/2611.08111
作者: Zihao Zhang,Xinyu Zhang,Bing Zeng,Xingyu Chen,Zhenning Li,Bin Yu,Ziyuan Pu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:World-action models (WAMs) jointly model future traffic states and driving actions for end-to-end autonomous driving in intelligent transportation systems. However, their action learning remains dominated by expert-trajectory imitation, leaving unsafe alternatives close to expert trajectories insufficiently constrained. To improve WAM robustness in safety-critical scenarios, we propose NegWAM, a safety post-training framework for unified WAMs that learns explicit safety boundaries from near-expert failure trajectories. NegWAM constructs negative trajectories that are geometrically close to the expert trajectory yet carry safety risks, uses them as repulsive anchors to steer predictions away from dangerous behavior, and encodes them as scene-conditioned rider tokens to learn scene–trajectory compatibility. During training, NegWAM jointly optimizes the flow-matching and near-expert negative objectives, while the auxiliary rider branch is discarded at inference, incurring no additional overhead. On NAVSIM v2, NegWAM achieves 40.4 EPDMS on navhard, a 4.1% relative improvement over the strongest compared scorer-free WAM baseline, and achieves 93.3 EPDMS on navtest. These results demonstrate that NegWAM achieves strong performance in safety-critical scenarios while preserving routine-driving performance. Zero-shot closed-loop evaluation on HUGSIM further demonstrates cross-benchmark generalization. The code is available at this https URL.
[CV-35] owards Practical 20 kbps Generative Video Compression via Adaptive Tokenization and Entropy Modeling
链接: https://arxiv.org/abs/2611.08099
作者: Jiancheng Zhao,Jixiang Luo,Haibin Huang,Chi Zhang,Xuelong Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Video compression at ultra-low bitrates remains challenging due to the loss of perceptually important details under limited bit budgets. Recent generative video compression methods leverage pretrained video priors to improve perceptual quality and temporal coherence, but many follow a compress-then-generate paradigm, where generative priors are primarily exploited during reconstruction. Moreover, representations learned for generation and reconstruction are not inherently optimized for entropy coding. In this work, we propose \method, which adapts pretrained video representations into compression-oriented representations while preserving their reconstruction capability. Specifically, we adapt the pretrained causal video VAE of MiniMax-H3 into a discrete video tokenizer and develop a small autoregressive entropy model that captures local and global token dependencies, with grouped prediction for parallel decoding. To further improve coding efficiency, we jointly optimize the tokenizer and entropy model using a differentiable rate surrogate, allowing rate gradients to adapt token assignments toward lower coding costs while preserving reconstruction quality. Experiments on UVG, MCL-JCV, and HEVC Class B demonstrate superior perceptual rate–distortion performance, achieving BD-LPIPS savings of 82.2%–87.5% and BD-DISTS savings of 78.3%–85.0% relative to GLC-Video. Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2611.08099 [cs.CV] (or arXiv:2611.08099v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.08099 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-36] National-scale mapping of tree species composition in Switzerland from high-resolution aerial imagery and multimodal data
链接: https://arxiv.org/abs/2611.08095
作者: Yixin Zhou,Luca Ferrari,Verena C. Griess,Nataliia Rehush,Konrad Schindler,Mirela Beloiu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 39 pages, 17 figures (11 in the main text, 6 in the supplementary material). Submitted to Remote Sensing of Environment
Abstract:Tree-species composition shapes forest productivity and how forests respond to climate disturbances such as drought, windthrow and insect outbreaks. Yet most national forest maps assign each pixel a single dominant species and cannot show the share of accompanying species. We mapped the basal-area shares of ten main tree species and a residual Other class across Switzerland at 30 m resolution. A deep-learning model combining 0.1 m aerial RGB imagery, airborne laser scanning (ALS), annual AlphaEarth satellite embeddings, and climate variables was trained on basal-area observations from 3,992 Swiss National Forest Inventory (NFI) sites. In spatial cross-validation, the model reached a macro- R^2 of 0.485 across the 11 classes. On an independent forest-reserve dataset, the map reached a macro- R^2 of 0.552 for individual 30 m pixels and 0.733 for plots, and identified the dominant class in 73.2% of pixels and 81.1% of plots. The map shows maple and ash as widespread minor components that dominant-class maps omit. It also reproduces the elevational sequence from broadleaved lowland stands through spruce to larch and Arolla pine near the treeline. By resolving the species mixture within each 30 m pixel, the map complements field inventory statistics with spatially continuous information that can support forest monitoring, the management of mixed stands, and climate-adaptation planning across Switzerland.
[CV-37] rraMesh-Masks: Open-Vocabulary Segmentation for Earth Observation NEURIPS2026
链接: https://arxiv.org/abs/2611.08081
作者: Benedikt Blumenstiel,Hugues Devimeux,Johannes Jakubik,Konrad Schindler
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026
Abstract:Open-vocabulary segmentation (OVS) enables models to segment a variety of concepts specified in natural language, removing the need for fixed label sets and fine-tuning. While recent OVS approaches have shown strong performance on natural images, they transfer poorly to Earth observation (EO) data due to differing semantics, scales, sensing modality, and geographic context. To date, large-scale EO-specific datasets and benchmarks for OVS remain scarce. We, therefore, introduce TerraMesh-Masks, a large-scale open-vocabulary segmentation dataset for EO, comprising 16 million binary masks aligned with nine million multimodal TerraMesh samples and annotated using OpenStreetMap and Overture-derived semantics. In addition, we provide an expert-reviewed evaluation set of 523 samples covering over 200 classes to enable robust evaluation. To demonstrate the impact of TerraMesh-Masks, we build an open-vocabulary version of TerraMind. Our results demonstrate significantly improved performance over general-purpose state-of-the-art models reaching 35.18% IoU versus 16.29% for this http URL and 11.33% for SAM-3. We release the dataset publicly under a permissive license at this https URL.
[CV-38] Spatiotemporal Dynamics-Aware Industrial Anomaly Detection and Reasoning WACV
链接: https://arxiv.org/abs/2611.08080
作者: S-Wei Chen,Siangyu Yang,Yi-Chen Jan,Chun-Yi Lee,Shao-Yuan Lo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2027
Abstract:Industrial Anomaly Detection (IAD) is a long-established problem that aims to detect defects in industrial products during manufacturing for quality control. Most existing studies focus on static visual patterns as the basis for detection. However, in real manufacturing environments, many anomalies are dynamics-based, such as irregular movements that cannot be perceived from static cues alone. We observe that even recent Multimodal Large Language Model (MLLM)-based approaches fall short in identifying such dynamic anomalies due to limited temporal modeling capability. To address this, we introduce MUSTARD: MUltimodal SpatioTemporal Anomaly Reasoning and Detection, a framework that equips MLLMs with fine-grained spatiotemporal dynamics awareness for detecting and reasoning about dynamics-based defects. MUSTARD introduces optical-flow-guided keyframe selection and augmented modalities to capture critical motion irregularities through local motion patterns and global dynamic evolution. Its normal pattern-based reasoning pipeline requires only a few normal reference samples as the detection basis, avoiding any model updates. This enables efficient adaptation across various products and provides high decision interpretability. MUSTARD is the first MLLM-based framework for dynamics-based IAD. Experiments demonstrate its superior performance in detecting spatiotemporal anomalies and providing plausible explanations. The code will be made publicly available.
[CV-39] NutriVisRAG : Multimodal Exemplar Grounding for Nutrient Estimation with Vision-Language Models
链接: https://arxiv.org/abs/2611.08054
作者: Lubnaa Abdur Rahman,Ioannis Papathanail,Maria Panagiotou,Nefeli Dimitropoulou,Margret Harðardóttir,Elisabeth J. den Brok,Bastiaan E. de Galan,Stavroula Georgia Mougiakakou
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Diet is a major determinant of health, yet accurate dietary assessment from free-living meal images remains challenging. Existing retrieval-based Vision-Language Model (VLM) approaches to nutrient estimation typically rely on textual food representations, while, to our knowledge, direct grounding on visually similar annotated meals has not been investigated. We present NutriVisRAG, a retrieval-augmented generation framework that uses a vision encoder to retrieve similar meals with known nutritional annotations and provide them as multimodal in-context exemplars for VLM nutrient estimation at inference time, without task-specific fine-tuning. We evaluate NutriVisRAG across four datasets, four VLMs, three retrieval encoders, and two prompting strategies, including free-living meals from adults with insulin-treated diabetes, where carbohydrate estimates are compared against two dietitians’ annotations. Across all 96 configurations (4 datasets x 4 VLMs x 3 retrieval encoders x 2 prompting strategies), NutriVisRAG reduced the primary estimation error in 56, with improvements more frequent under whole-meal (35/48) than per-item prompting (21/48). Across the evaluated configurations, the lowest NutriVisRAG error was below the best Vanilla error in 13 of 16 VLM–dataset pairs. For GPT-5.4-mini, under whole-meal prompting, the best-performing retrieval encoder for each dataset reduced the primary error on all four datasets relative to Vanilla, with reductions of 1.7–35.7%. Overall, our results show that grounding in multimodal exemplars comprising images and respective annotations can improve nutrient estimation at inference time without task-specific fine-tuning, although the benefits remain configuration-dependent.
[CV-40] Deformable Contact-Aware 3D Object Placement
链接: https://arxiv.org/abs/2611.08005
作者: Songhe Wang,David Miller,Charles A. Kamhoua
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 23 pages, 4 figures; includes technical appendices
Abstract:Current language-guided 3D object placement typically predicts a rigid 6-DoF pose, ignoring the physical reality that indoor placements are dictated by contact. Consequently, predicted poses often float, interpenetrate supports, or fail to deform soft surfaces. We propose Deformable Contact-Aware Placement (DCAP), a framework that reframes placement as a drop-to-equilibrium process. Rather than predicting a static pose, DCAP constructs a semantic release state that specifies the drop location, object scale, and deformable components, allowing a coupled rigid-deformable simulator to settle the object under gravity and friction. This explicitly separates semantic intent from physical realization. Evaluated on a benchmark of 2,063 prompt-asset-scene cases, DCAP significantly outperforms static baselines. On standard placements, it improves VLM Physics Naturalness by 21.2 points and human preference by 26.5 points over the strongest raw rigid baseline, while reducing mean penetration depth by 72.3%. Compared to simulated rollout baselines, DCAP reduces residual speed by 72.1%. Ultimately, our results prove that physically accurate 3D placement requires modeling semantic intent and physical realization as strictly separate processes.
[CV-41] When GeoFMs Meet VHR: Benchmarking Foundation Models on Very High Resolution Earth Observation Imagery
链接: https://arxiv.org/abs/2611.08003
作者: Kshitij Raj Sharma,Pierre Adorni,Minh-Tan Pham,Dirk Tiede,Sébastien Lefèvre
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Geospatial foundation models are a widely adopted starting point for remote sensing tasks. Most are pretrained on Sentinel-1/2 and Landsat imagery at a ground sampling distance (GSD) of 10m or coarser. Many applications instead rely on very high resolution (VHR) imagery, where individual buildings, roads, and amenities are resolved. We study which pretraining regimes transfer well to this scale. We introduce VHR-Bench, a benchmark suite of ten datasets at 1.5m GSD or finer spanning five task families, eight application domains, and both satellite and aerial platforms. We evaluate twelve encoders in three pretraining tiers with frozen backbones at two training-data fractions. We find that both pretraining on large natural-image collections and VHR-specific pretraining transfer effectively to VHR tasks, with complementary strengths. SkySense leads when labels are scarce; at the full label fraction the web-pretrained DINOv3 ViT-L edges ahead, with the two close together there. Encoders pretrained only on Sentinel-scale imagery transfer least well under RGB input. Guided by these complementary results, we continue the self-supervised pretraining of DINOv3 ViT-L on 1.02 million held-out VHR tiles and evaluate the resulting encoder without changing the downstream protocol. The continued model gives a clear improvement over the released backbone when labels are scarce and a modest one at the full fraction, at no increase in inference parameters or compute. These findings support a generalize-then-specialize strategy: already strong, broad natural-image pretraining can be further improved by a short VHR continuation stage while preserving its general visual features.
[CV-42] Reviving the Discarded High-Resolution Feature for Transformer-Based Tiny Object Detection NEURIPS2026
链接: https://arxiv.org/abs/2611.07994
作者: Zehao Wang,Si Wu,Guiping Cao,Xin Li,Yaowei Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026
Abstract:Tiny object detection remains challenging because tiny instances occupy only a few pixels, and their fine-grained details are easily lost during feature downsampling. Although the highest-resolution feature preserves rich local details for tiny objects, DETR-like detectors often discard it due to its large computational cost, while directly reintroducing this feature also brings dense background noise. To address this problem, we propose ReDHF, a novel framework that Revives the Discarded High-resolution Feature for tiny object detection. ReDHF has three cooperative components. Cross-Scale Deformable Fusion (CSDF) first constructs a semantically enhanced highest-resolution feature and injects its fine-grained details into the shallow feature under a moderate cost. Ellipse-Guided Shallow Supervision (EGSS) then guides the enhanced feature to learn more reliable classification and localization priors by applying geometry-adaptive auxiliary supervision. Full-Scale Query Interaction (FSQI) further enables decoder queries to access both low-level details and high-level semantic information, allowing them to aggregate richer cues for tiny objects. Extensive experiments show that ReDHF consistently improves multiple baselines across different DETR variants and achieves state-of-the-art performance on the evaluated benchmarks. The best ReDHF variants surpass their baselines by +1.4 AP on AI-TOD-v2, +1.6 AP on VisDrone, and +1.5 AP on SODA-D, with especially notable gains on tiny and small objects.
[CV-43] Bri3D: Bridging the Object-Scene Gap in Single-Image 3D Generation via Scene-Anchor
链接: https://arxiv.org/abs/2611.07961
作者: Hongye Hou,Shuai Zhang,Hongjie Li,Zhuoxiao Li,Haiping Wang,Yuan Liu,Wufan Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Generating compositional 3D scenes from a single image requires recovering detailed object geometry and coherent spatial arrangements. Existing object-to-scene methods, including SAM3D and MIDI, extend object generation through image-context reasoning or inter-instance modeling. Coordinating objects remains challenging with limited scene-level supervision. Directly extending object-centric generators to entire scenes can degrade geometry and fail to preserve instance scale, orientation, and pose. We propose Bri3D, a framework bridging scene-level structural reasoning and object-level generative priors through instance-aware voxel scene anchors. Bri3D predicts a color-coded scene scaffold encoding coarse object geometry and spatial arrangements. Each instance is extracted as a scene anchor and refined into a pose-preserving, prior-compatible voxel support, enabling a frozen object generator to synthesize detailed textured geometry in the anchor-defined spatial reference. This coarse-to-fine design separates scene organization from object-scale detail synthesis, focusing scene-level training on spatial organization and anchor refinement while reusing pretrained object priors. Experiments demonstrate improvements in scene layout, object geometry, visual quality, and generalization. Bri3D predicts scene layouts efficiently; full scene generation is faster than SceneGen and SAM3D.
[CV-44] Rotation Is Token-Blind: A Structural Criterion for Post-Training Quantization of Feed-Forward Geometry Transformers
链接: https://arxiv.org/abs/2611.07955
作者: Huanfei He,Bohao Zhang,Jiaxin Zhang,Wei Sui,John See,Zeyd Boukhers,Cong Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Feed-forward 3D models use Transformers to predict camera poses, depth, and structure in one pass, avoiding iterative multi-view optimization. Long sequences and deep attention incur high compute and memory costs, motivating Post-Training Quantization (PTQ) for deployment. Existing PTQ methods mainly suppress channel-wise outliers by smoothing or orthogonal rotation. Geometry Transformers introduce a different challenge: heterogeneous token types share one activation quantization grid. A few special tokens may exert high leverage over global outputs and be updated by only a subset of blocks. Token-blind transforms, which apply the same transformation to every token, reduce overall quantization error but cannot remove differentiated degradation across token types. We propose QMapAnything, a structure-aware PTQ framework. We cast floating-point-preserving rotation and smoothing as a token-blind reparameterization group, and prove that they yield common-mode gains while leaving relative token SNR unchanged. We introduce a statistic \Gamma that combines a special token’s quantization disadvantage and output leverage to predict whether differentiated bias correction is needed beyond rotation. Experiments on geometric benchmarks show that W4A8 closely matches full-precision performance. Under W8A8, peak memory falls by 2.2 – 3.1\times versus FP32, with up to 3.1\times speedup on real hardware. These results provide a principled and efficient approach to deploying feed-forward 3D reconstruction models in resource-constrained edges.
[CV-45] Leakage- and Label-Aware Benchmarking of Class-Imbalance Strategies for Multimodal Chest X-Ray and EHR Classification
链接: https://arxiv.org/abs/2611.07950
作者: Dang-Khoa Bui,Quang-Huy Tran,Minh-Khoi Nguyen-Bui,Duc-Tuan Ngo,Thanh-Trong Tran,Tuan-Khoi Nguyen,Hoang-Anh Ngo
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 1 figure, 6 tables
Abstract:Benchmarks of class-imbalance strategies for multimodal chest X-ray (CXR) and electronic health record (EHR) prediction are often evaluated on record-level splits of cohorts in which one patient contributes several radiographs. We audit such a benchmark on 29,083 MIMIC-CXR radiographs of 10,479 patients with eight diagnosis targets. First, the record-split benchmark was misleading: a lookup table that copies the label of a development-set radiograph with the same EHR vector nearly matched the best model, and on unseen patients every model fell to the level of predicting every case positive. Second, tracing the labels back to MIMIC-IV explained why: each patient’s labels are keyword matches on the diagnosis codes of a single admission, copied, with one EHR vector per patient, to all of the patient’s radiographs, so about a third of the radiographs carry labels that differ from the codes of their own admission. Third, under patient-separated evaluation with admission-aligned and rule-refined labels, the initial conclusions do not hold: in the evaluated configurations no imbalance strategy raised the observed mean macro F2 over no augmentation by more than 0.01, tuned tree ensembles showed no consistent advantage over tuned logistic regression, and the released EHR variables stayed at the level of the all-positive baseline, whereas the diagnostic history available at imaging time and image embeddings exceeded it, including, with rule-refined labels, a CheXpert-trained encoder that never saw these radiographs; RAD-DINO had seen 96.6% of them during self-supervised pretraining. Adding either encoder to the diagnostic history gave no statistically significant F2 gain; exploratory AUROC gains were significant with RAD-DINO but not with DenseNet. We provide reusable diagnostics, a leakage-aware protocol and a checklist.
[CV-46] IDPrior: Decoupling Identity Generation from Facial Image Synthesis WACV2027
链接: https://arxiv.org/abs/2611.07941
作者: Fadi Boutros,Meriç Demirörs,Darian Tomašević,Vitomir Štruc,Naser Damer
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at WACV 2027
Abstract:Large-scale identity-labeled face datasets are essential for training modern face recognition models, yet collecting and annotating such data is costly and raises privacy and ethical concerns. Existing synthetic face generation methods typically derive identities from predefined class labels, unconditional image generation, or predefined latent sampling strategies, without explicitly modeling the underlying distribution of facial identities. In this paper, we propose IDPrior, a two-stage framework that decouples identity generation from image synthesis. We first learn a diffusion prior in the embedding space of a pretrained face recognition model to model the distribution of facial identities and generate novel identity embeddings. These embeddings are then used to condition a pretrained identity-conditioned diffusion model (DM), enabling the synthesis of multiple images of the same identity. This modular design provides a scalable approach to identity generation while remaining compatible with existing and future identity-conditioned DMs. Experiments indicate that IDPrior generates diverse and well-separated identities while maintaining a closer correspondence to the distribution of authentic identities. Face recognition models trained on datasets generated using IDPrior-generated identities achieve competitive or improved performance compared with existing synthetic-data-based methods across multiple face verification benchmarks. Code and generated datasets are publicly available at this https URL.
[CV-47] WAM-Sim0.1: World Anchored Memory Meets Online RL for Consistent and Realistic Driving Simulation
链接: https://arxiv.org/abs/2611.07939
作者: Jiahao Cui,Jiaye Li,Wenhao Zhang,Yaoyi Li,Shan Luan,Quanhui Tang,Jianhua Han,Lin Cheng,Hang Xu,Siyu Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Autoregressive driving world models systematically suffer from structural drift, identity drift, and quality decay over long rollout horizons due to transient conditioning signals that lack persistent state. We present WAM-Sim, a driving world model anchored by a persistent hierarchical memory and optimized via online flow-matching RL. WAM-Sim structures persistent context across three temporal and semantic scales: pose-aligned structural tokens encoding persistent road topology, read-only scene tokens supplying geolocated background appearance, and write-once instance tokens preserving individual traffic participant identities across occlusions and chunk boundaries. To restore visual sharpness and physical plausibility degraded by multi-scale conditioning, we introduce an online post-training framework that optimizes self-generated rollouts via flow-matching RL guided by hierarchy-aligned visual rewards and a vision-language model physical judge. The invariant memory empirically confines policy optimization within the anchored context, helping suppress semantic collapse and reward hacking, while online RL systematically enforces physical commonsense and fine-grained visual fidelity. On the nuScenes benchmark, WAM-Sim achieves a superior generation fidelity (FID=3.47, FVD=25.72) and strong downstream perception performance (24.16 mAP, 71.23 road mIoU). On minute-scale rollouts of self-collected data, WAM-Sim substantially mitigates structural and identity drift, preserving high scene similarity and subject identity coherence across extended horizons. Project page: this https URL.
[CV-48] owards Scalable Diagnostic-Driven Whole-Slide Image Retrieval for Pathology Archives
链接: https://arxiv.org/abs/2611.07935
作者: Jingsong Liu,Youssef Wally Han Li,Weiwei Zhou,Jing Dai,Elisabeth Wetzer,Peter J. Schüffler
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:The growing adoption of digital pathology has rapidly expanded collections of hematoxylin-and-eosin whole-slide images, yet efficient navigation through these archives remains challenging, as metadata is often inconsistent or unavailable. Content-based image retrieval addresses this challenge by identifying similar cases directly from visual representations instead of manual annotations or metadata. Existing retrieval systems have demonstrated promising performance but largely rely on handcrafted indexing pipelines and predefined sampling strategies, constraining scalability and generalization. Here, we present CLEAR-WSI (Constant-Length Embedding for Automatic Retrieval of Whole-Slide Images) as a scalable, diagnostic-driven slide-retrieval framework, validated on both public and clinical data. Using a fixed-length slide embedding, CLEAR-WSI refines retrieved candidates with either a predicted diagnostic label or the ground-truth label, depending on the retrieval setting. Newly processed cases are subsequently added to the expanding memory bank, allowing the retrieval archive to grow over time. On a comprehensive benchmark consisting of two public and one private clinical datasets representing four retrieval tasks, CLEAR-WSI improves Precision@5 over the strongest evaluated baseline, with gains on CAMELYON16 (3.10), BRACS-1 (2.75), BRACS-2 (1.95), and X-HER2 (3.75). Additional ablation and latency studies further examine the contribution of each component and highlight its potential for scalable deployment.
[CV-49] Flux as a Codec: Zero-Shot Lossy Compression with Flux
链接: https://arxiv.org/abs/2611.07929
作者: Fuma Kimishima,Jinjia Zhou
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:We propose FluxC, a zero-shot codec that uses Flux as a compression prior. Although DiffC has shown that pretrained diffusion models can serve as zero-shot codecs, we ask how the choice of coding path and reconstruction method affects rate–distortion–perception behavior. We analyze FluxC along three axes: coding cost and bit allocation, reconstruction method, and low rate distortion. At matched signal-to-noise ratio (SNR), we show that the path choice does not affect the ideal information rate, but changes how bits are allocated across time steps. For reconstruction, we compare ancestral sampling, one-step ODE reconstruction, and many-step ODE reconstruction. Our analysis shows that ODE-based reconstruction can achieve lower distortion than ancestral sampling, especially in the low rate regime. Experiments on Gaussian sources and natural images support this prediction. Moreover, FluxC performs competitively with, and in some cases outperforms, recent neural image compression methods, suggesting that Flux can serve as a principled zero-shot codec.
[CV-50] StitchEdit: Stitching Depth Priors into Image Editors via Per-Layer Gradient Probing
链接: https://arxiv.org/abs/2611.07922
作者: Qin Guo,Dan Xu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Text-guided image editing with diffusion models struggles with spatial edits (e.g. manipulating objects, changing viewpoints, manipulating occlusions) that depend on the 3D structure understanding of a scene. Marigold and follow-up work demonstrate that a pretrained diffusion backbone already harbors strong geometric priors that can be activated for depth estimation, yet how to redirect these priors toward editing remains unresolved. Through per-layer gradient probing, we discover that the editing and depth objectives are compatible in early transformer layers but actively conflict in late layers, explaining why naive depth-conditioned editing either ignores geometry or degrades visual quality. We present StitchEdit, which first activates the backbone’s geometric capability via a lightweight depth-denoising adapter, then selectively fine-tunes this adapter for editing using a probe-guided per-layer learning rate schedule that preserves depth knowledge where it helps and prunes it where it conflicts. On the proposed StitchEditBench, a benchmark of 150 geometry-dependent edits spanning spatial transformation, occlusion manipulation, and viewpoint change, StitchEdit consistently surpasses both instruction-based and depth-conditioned baselines on perceptual quality and geometric fidelity.
[CV-51] EDGAR: Efficient Diffusion-Driven Gaussian Avatar Refinement via Synthetic Failure
链接: https://arxiv.org/abs/2611.07909
作者: Kwan Yun,Minhyeok Bang,Colette Suhjung Koo,Junyong Noh
类目: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Multimedia (cs.MM)
备注: Accepted to ACM Transactions on Graphics
Abstract:Recent advances in Gaussian splatting have enabled efficient and photorealistic head avatar reconstruction, yet monocular video-based or image-based methods still suffer from under-constrained geometry and texture, leading to artifacts such as blurring, stretching, and temporal instability. We present EDGAR (Efficient Diffusion-Driven Gaussian Avatar Refinement), a conditional diffusion framework designed to enhance the quality of Gaussian avatars by correcting artifacts while preserving identity fidelity. To train EDGAR, we construct EDGAR-SET, a hybrid triplet dataset composed of real-captured and synthetic-rendered samples, which simulates realistic Gaussian avatar failure cases by pairing degraded renderings with ground-truth and reference images. EDGAR refines degraded frames through a fusion block that injects identity features extracted by the self-supervised Sapiens model into a one-step diffusion backbone, enabling effective artifact removal without sacrificing efficiency. Our EDGAR model incorporates a Mixture-of-Experts (MoE) architecture to address diverse degradation types and severity levels. The refined multi-view renderings are subsequently used as pseudo-targets to optimize a single underlying Gaussian avatar, so that the final avatar can be rendered without applying EDGAR to every output frame. Experiments on three different Gaussian avatar reconstruction methods, each evaluated under three different settings, show improvements in both perceptual and reconstruction measures over the raw outputs and an SDEdit-style refinement baseline. In addition, cross-view evaluation demonstrates improved calibrated consistency, while tests under large yaw and open-mouth expressions verify robustness under extreme settings. These results collectively indicate a useful operating regime for recoverable appearance artifacts. The code is available at this https URL
[CV-52] XM4D: 4D Reconstruction through Certifiable Geometric Optimization
链接: https://arxiv.org/abs/2611.07902
作者: Haoyu Han,Jack Benarroch Jedlicki,Jiahui Lei,Heng Yang
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO); Optimization and Control (math.OC)
备注:
Abstract:Accurate 4D reconstruction remains challenging: jointly estimating camera motion, scene geometry, and non-rigid deformation entails highly nonconvex optimization, making global optimality difficult to guarantee. We present XM4D, a framework for accurate and efficient 4D reconstruction built upon certifiably globally optimal geometric optimization. By separating the estimation of camera motion from non-rigid deformation, we formulate both stages as structured quadratically constrained quadratic programs (QCQP) that admit semidefinite programming (SDP) relaxations and verifiable global optimality certificates. For camera motion estimation, we lift static point tracks into 3D using foundational depth prediction models and formulate a scaled bundle adjustment problem that jointly estimates camera poses, depth scales, and static landmarks. By analytically eliminating translations and landmarks, we reduce the problem to a QCQP over scaled rotations. For non-rigid reconstruction, we represent dynamic motion using an embedded deformation graph (EDG) and show that its tracking, rigidity, and temporal regularization objectives similarly reduce to a rotation-only QCQP. Both problems admit SDP relaxations that we solve efficiently with a GPU-accelerated Burer-Monteiro solver based on Riemannian optimization. The recovered geometry further initializes dynamic Gaussian splatting for photometric refinement. Extensive experiments across seven benchmarks demonstrate accurate camera motion estimation on both static and dynamic scenes, efficient optimization that scales to thousands of frames, and state-of-the-art novel-view rendering quality.
[CV-53] Structure from Collision via Hypothesis-Space Reasoning WWW WACV2027
链接: https://arxiv.org/abs/2611.07900
作者: Takuhiro Kaneko
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR); Machine Learning (cs.LG); Robotics (cs.RO)
备注: Accepted to WACV 2027. Project page: this https URL
Abstract:Knowledge of the hidden interior structure of an object, including the presence and location of cavities, is essential for predictive simulation, robotic manipulation, and nondestructive inspection. However, such cavities leave no direct signature on the static visible surface, making collision-induced deformation one of the few cues used to infer their presence. Structure from Collision (SfC) addresses this problem by recovering the hidden interior structure of an object from multiview observations of a collision event. SfC is challenging for two reasons: hidden cavities are only indirectly observable through surface motion, and voids correspond to the absence of material, making direct optimization using standard density representations difficult. The density-based SfC baseline inherits both difficulties. Surface-matching gradients do not directly indicate the cavity location, while density gradients can be weak or uninformative inside or near void regions. Therefore, we reformulate SfC as hypothesis-space reasoning, representing hidden cavities as explicit structural hypotheses whose variables encode available structural assumptions rather than relying on implicit density updates. A global stage performs gradient-free optimization over this hypothesis space using forward-only material point method (MPM) simulation. A local stage then refines the selected hypothesis as a Lagrangian particle system through gradient-based optimization using differentiable MPM simulation. On the SfC benchmark, our method improves hidden-structure reconstruction and future prediction over the density-based SfC baseline. Interior-topology stress tests and robustness analyses characterize its scope and limitations. The project page is available at this https URL.
[CV-54] InnerLogic-Surg: An Open Surgical Foundation Model via Continual Pretraining
链接: https://arxiv.org/abs/2611.07883
作者: Kun Yuan,Hao Ding,Yiqing Shen,Gasper Podobnik,Benjamin Killeen,Tito Porras,Axel Krieger,Mathias Unberath
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Vision foundation models are the backbone of modern surgical image analysis, yet most surgical systems still rely on encoders pretrained on natural images or on specialist models trained from scratch on comparatively small datasets. Pretraining encoders on large amounts of surgery-specific data has the potential to generally enhance the performance of surgical image analysis methods simply by providing stronger backbones. Here we present InnerLogic-Surg, an open family of surgical vision encoders (21M–300M parameters) obtained by continually pretraining a transformer-based image encoder on millions of surgical images curated from 11,965 internet-sourced surgical videos spanning ~2,900 hours and more than fifteen procedure types. Our training recipe inherits DINOv3’s pretraining stage to preserve representation ability after surgical adaptation, and additionally applies a learned feature-upsampling module for high-resolution dense prediction at low token cost. Across a benchmark covering workflow, instrument and anatomy recognition, segmentation, detection, and monocular depth, under the same architecture, the largest InnerLogic-Surg model achieves the best linear-probe performance in all 20 surgical tasks. We demonstrate that surgical continual pretraining yields transferable representations that generalize across procedures, anatomical regions, and task types, showing our surgical image encoders as a promising and scalable alternative to the procedure-specific representation learning approaches that have traditionally dominated surgical computer vision.
[CV-55] PVDepth: Panoramic Video Depth Estimation via Geometry-Aware Spatiotemporal Adaptation ICML2026
链接: https://arxiv.org/abs/2611.07873
作者: Chuanxin Song,Peixi Peng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ICML 2026, Project Page: this https URL
Abstract:Panoramic video depth estimation is pivotal for applications such as Virtual Reality and World Models. However, advancements in this field are impeded by two primary obstacles: the scarcity of large-scale training data and the unique spatiotemporal challenges of Equirectangular Projection (ERP), which hinder the direct transfer of perspective models. In this paper, we first present PanoCARLA, a large-scale synthetic RGB-D panoramic video dataset, featuring natural motion trajectories and drone-like roaming perspectives. Building on this foundation, we propose PVDepth, an end-to-end framework adapted from perspective video depth models. To tackle ERP-specific geometric distortions and consequent non-linear temporal dynamics, we introduce two core mechanisms: (1) A Progressive Sphere-aware Noise Initialization strategy that anneals the noise distribution from planar to spherical, guiding the model to adapt to non-uniform information density; and (2) A Cube-rectified Temporal Modeling module that incorporates an auxiliary cubemap temporal branch to rectify non-linear temporal dynamics in the ERP domain. Extensive experiments demonstrate that PVDepth achieves superior performance, generating geometrically accurate and temporally consistent depth sequences. Code and data will be released at https:// this http URL.
[CV-56] PlayGen2: Learning Real-Time Playable Worlds with Explicit 3D State
链接: https://arxiv.org/abs/2611.07867
作者: Jihai Zhang,Yuhan Zhou,Junyou Li,Deheng Ye,Qi Dou,Yu Cheng
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Code: this https URL
Abstract:Building a playable world requires more than generating plausible video: it demands responsive control, persistent spatial structure, and coherent interactions with dynamic entities. Using generated frames as the sole recurrent world representation ties these requirements to view-dependent appearance synthesis, making it difficult to maintain long-term spatial and temporal consistency in generation and interaction. We present PlayGen2, a real-time interactive world model that separates learned 3D state evolution from neural rendering. PlayGen2 maintains a persistent world memory comprising static geometry represented by a volumetric occupancy grid and identity-indexed 3D states for the player and other entities. Within a fixed-size local region, its world-state model jointly learns action-conditioned entity dynamics, geometry completion, and entity discovery, committing updates to a persistent world memory. A separate neural renderer synthesizes first-person observations from depth and semantic projections of the world state, without feeding generated images back into state prediction. Using Doom as a testbed, we demonstrate that the model successfully learns world dynamics including geometry-aware player motion and player–NPC interactions. The framework is computationally efficient, enabling the complete simulation and rendering pipeline to achieve approximately 160ms backend latency on a MacBook Pro with an M4 Pro chip. Beyond single-player interaction, we demonstrate zero-shot deployment on unseen Doom maps, multiplayer interaction through the composition of single-player models, and programmable gameplay through rule-based state interventions. These capabilities demonstrate the transferability and extensibility of the learned game-world model. Code and pretrained model weights are publicly available.
[CV-57] OcTreeMamba: Topology-Aware Mamba for 3D Medical Segmentation via Octree Sparsification
链接: https://arxiv.org/abs/2611.07865
作者: Kaixiang Mai,Jiaqi Yu,Yuzhu Ji,Yiqun Zhang,Zhijie Wang,Haijun Zhang,Yiu-ming Cheung
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 8 pages, 5 figures, accepted by IEEE BIBM 2026
Abstract:Accurate segmentation of 3D anatomical structures is crucial to medical image analysis. Recently, Mamba-based approaches have achieved promising segmentation performance in capturing long-range dependencies with linear-time computational complexity. However, heuristic selective scanning strategies inherently disrupt the intrinsic spatial hierarchy and increase computational redundancy, ignoring the inherent sparsity of 3D anatomical structures. To this end, we propose OcTreeMamba, a hierarchical framework that incorporates Octree structures into SSM, achieving 3D topology-aware long-range context feature propagation and integration. Specifically, a Structure Learner performs anatomy-adaptive structured sparsification through hierarchical gating and Gumbel-Softmax sampling, enabling differentiable selection of anatomical regions. The selected regions are subsequently organized by a Dynamic Octree Builder into a multi-level topology that preserves their hierarchical connectivity. Based on the constructed Octree, a bidirectional Tree-SSM propagates and integrates contextual features along connectivity-aware leaf–parent–root and root–parent–leaf trajectories. Extensive experiments on four benchmark datasets show that our proposed OcTreeMamba achieves competitive performance \wrt topological continuity and computational efficiency, particularly for recovering the complex structure on vascular datasets. The code is available at this https URL.
[CV-58] Learning Robotic World Models with Semantic and Visual Feedback
链接: https://arxiv.org/abs/2611.07850
作者: Hao Wu,Yuqi Li,Yizhou Zhao,Kun Wang,Xiao Luo,Yu Cheng,Bo An,Yingli Tian,Laszlo A. Jeni,Ming-Hsuan Yang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 27 pages, including supplementary material. Project and code: this https URL . Data and weights: this https URL
Abstract:Robot world models predict action-conditioned future visual states to support robotic planning and decision making. However, visual reconstruction does not fully capture task progress, semantic evaluation alone provides limited constraints on fine-grained visual errors, and homogeneous predictions limit comparative feedback. We present RoboImagine, a framework for post-training autoregressive robotic world models through candidate-future comparisons with semantic and visual feedback. Diverse search promotes diversity among candidate trajectories under shared visual observations, action sequences, and task instructions. A frozen multimodal evaluator provides task-relevant semantic preferences through pairwise video comparisons in both presentation orders, complemented by pixel- and perceptual-level visual consistency feedback that constrains local structural and appearance errors. These signals drive policy optimization through group-relative advantages. Fixed initial context and branch-specific rolling histories support long-sequence generation with bounded context and consistent conditioning between candidate generation and policy optimization. Experiments with separately trained models on RT-1, BridgeV2, and RoboCasa cover real-world and simulated manipulation. On RT-1, under matched initialization, update counts, and evaluation protocols, RoboImagine improves PSNR by 1.26 dB and reduces LPIPS by 11.93% relative to continued supervised fine-tuning. Ablations show that diverse search improves prediction quality and that combined semantic and visual feedback yields higher external multimodal evaluation scores than visual feedback alone while maintaining comparable visual fidelity. Rolling histories further improve long-horizon prediction. These results support improving visual fidelity, task relevance, and long-horizon prediction through candidate comparison and complementary feedback.
[CV-59] Enhanced MRI brain tumor detection using deep learning in conjunction with explainable AI SHAP based diverse and multi feature analysis
链接: https://arxiv.org/abs/2611.07843
作者: Asif Rahman,Maqsood Hayat,Nadeem Iqbal,Fawaz Khaled Alarfaj,Salem Alkhalaf,Fahad Alturise
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 15 pages. Published in Scientific Reports 15, 29411 (2025). Open access under CC BY 4.0
Abstract:Recent innovations in medical imaging have markedly improved brain tumor identification, surpassing conventional diagnostic approaches that suffer from low resolution, radiation exposure, and limited contrast. Magnetic Resonance Imaging (MRI) is pivotal in precise and accurate tumor characterization owing to its high-resolution, non-invasive nature. This study investigates the synergy among multiple feature representation schemes such as local Binary Patterns (LBP), Gabor filters, Discrete Wavelet Transform, Fast Fourier Transform, Convolutional Neural Networks (CNN), and Gray-Level Run Length Matrix alongside five learning algorithms namely: k-nearest Neighbor, Random Forest, Support Vector Classifier (SVC), and probabilistic neural network (PNN), and CNN. Empirical findings indicate that LBP in conjunction with SVC and CNN obtained high specificity and accuracy, rendering it a promising method for MRI-based tumor diagnosis. Further to investigate the contribution of LBP, Statistical analysis chi-square and p-value tests are used to confirm the significant impact of LBP feature space for identification of brain Tumor. In addition, The SHAP analysis was used to identify the most important features in classification. In a small dataset, CNN obtained 97.8% accuracy while SVC yielded 98.06% accuracy. In subsequent analysis, a large benchmark dataset is also utilized to evaluate the performance of learning algorithms in order to investigate the generalization power of the proposed model. CNN achieves the highest accuracy of 98.9%, followed by SVC at 96.7%. These results highlight CNN’s effectiveness in automated, high-precision tumor diagnosis. Code: this https URL
[CV-60] HG-Omni: Traceable Human-Centric Graph Supervision for Video Reasoning
链接: https://arxiv.org/abs/2611.07836
作者: Jisheng Dang,Kai Wei,Ronghao Lin,Bimei Wang,Hong Peng,Wei-Shi Zheng,Bin Hu,Qi Tian,Tat-Seng Chua
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Omni-modal models can answer questions about video, but their predictions rarely show which participant, event, or interval supports an answer. Moreover, answer-level supervision alone does not teach a model to interpret structured video evidence. To systematically mitigate this issue, we present THG-Omni, a human-centric graph-supervision framework that links Segment Anything Model~3 participant tracks with temporally grounded Qwen3-VL annotations. Candidate semantic claims are checked against the source video and available audio before they are used to construct training questions, while track geometry provides explicit spatial and temporal supervision. The resulting questions teach an omni-modal reasoning model to interpret participant attributes, actions, affect, interactions, and their timing from a structured graph. At inference, THG-Omni uses a predicted graph constructed from the evaluation video. Experiments on IntentBench show consistent improvements over the unadapted backbone across social reasoning, emotion understanding, and deception perception. On NExT-QA, THG-Omni achieves 89.54% accuracy with the matched graph, compared with 85.10% without graph input. Further validation experiments show that compact graph input achieves 78.22% accuracy, compared with 78.02% using the full graph and 77.58% without graph input. A complementary diagnostic demonstrates that the predicted graph preserves evidence useful for answering questions even without the source video and audio. Together, these findings show that THG-Omni improves human-centric video reasoning while grounding its evidence in identifiable participants and events. Our codes are available at this https URL
[CV-61] CMDiff: Left Atrial Scar Segmentation from 3D LGE MRI Using Conditional Mask Diffusion Model: A Feasibility Study
链接: https://arxiv.org/abs/2611.07814
作者: Jing Zhang,Bastien Bergere,Jonas Leite,Emilie Bollache,Mikaël Laredo,Alban Redheuil,Nadjia Kachenoura
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 3 figures, 4 tables
Abstract:Late gadolinium enhancement (LGE) cardiac MRI enables non-invasive assessment of left atrial (LA) scar, an imaging marker associated with atrial fibrillation ablation outcome. However, automatic scar segmentation is hindered by low contrast, the thin atrial wall, scarce labels, and extreme class imbalance. Conditional diffusion offers iterative mask generation, but its suitability for sparse 3D LA scar segmentation remains unclear. To address this gap, we propose Conditional Mask Diffusion (CMDiff), which predicts clean masks using a dynamic-encoder-conditioned 3D sinusoidal UNet. We compared CMDiff with a deterministic baseline on identical cases in five-fold cross-validation of 60 scar-labelled public LASCARQS subjects, with 48 training and 12 held-out cases per fold. CMDiff achieved a scar DSC of 0.47 \pm 0.08, HD95 of 12.89 \pm 6.39 mm, and ASD of 1.91 \pm 0.96 mm, compared with deterministic baseline values of 0.51 \pm 0.08, 10.86 \pm 5.39 mm, and 1.58 \pm 1.03 mm, respectively. Under the same implementation, clean-mask prediction performed substantially better than the noise-prediction configuration evaluated in this study. Thus, CMDiff was technically feasible but did not improve segmentation performance: DSC was lower, and both HD95 and ASD were higher. Direct mask diffusion therefore provides an analytical alternative, but not a competitive performance advance for this task. The source code is publicly available at this https URL.
[CV-62] MV-Sign: Multi-View 3D Motion Integration and Enhanced Text-to-Motion Conditioning for Sign Language Production WACV2027
链接: https://arxiv.org/abs/2611.07811
作者: Singyu Muk,Kaustubh Kundu,Kunyuan Xie,Weijie Wang,Kalin Stefanov
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)
Abstract:Sign language production (SLP) aims to generate appropriate sign language motion from spoken language text. However, current 3D SLP methods suffer from noisy supervision and weak text conditioning. Specifically, these methods typically rely on pseudo-3D motion data fitted from single-view frontal videos, which introduces severe self-occlusion and loses fine-grained hand details, ultimately degrading downstream production. To overcome these limitations, we propose a comprehensive framework that jointly integrates multi-view 3D motion data and optimizes the text-conditioning pathway. First, we introduce How2Sign-MV3D, a high-fidelity dataset constructed via a novel cross-view curation pipeline that fuses multi-view estimates to resolve occlusions and filter non-signing segments. Second, we design a robust SLP model incorporating a sign-aware spatio-temporal VQ-VAE (ST-VQVAE) for joint body-hand tokenization, a pseudo-gloss adapter, and keyframe-aware retrieval to enhance text-to-motion alignment. Evaluated on the How2Sign benchmark, our framework reduces body and hand DTW-PA-JPE by 32.8% and 20.9% respectively compared to the state-of-the-art. Finally, we release two new high-quality 3D datasets, ASLCitizen-3D and MM-WAuslan-MV3D, alongside SGNify-MV as a multi-view video benchmark to support broader community research.
[CV-63] Self-Supervised Diffusion Models for Handwritten Signature Verification via Ruin Reconstruction
链接: https://arxiv.org/abs/2611.07810
作者: Peirong Zhang,Zhenhua Yang,Jiaxin Zhang,Moises Diaz,Miguel A. Ferrer,Umapada Pal,Wei Zhou,Lianwen Jin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Handwritten signature verification (HSV) is a critical biometric technique in legal and forensic authentication, whose efficacy hinges on the discriminability of signature representations. Current HSV methods predominantly adopt supervised learning to enhance representation quality, leaving self-supervised learning largely unexplored. We propose \emphSigDiffuser, a diffusion-based self-supervised framework for signature representation and verification. At its core, \emphSigDiffuser introduces a novel pretext task termed \emphReconstruction from Ruins (RFR), which deliberately ruins'' signatures through severe crumpling, obliterating their structural integrity and writing styles. A visual encoder is employed to capture crucial handwriting patterns from intact signature regions, guiding the denoiser to reconstruct the ruined’’ portions and learning meaningful representations. Additionally, we introduce a \emphpyramidal feature aggregation module (PFAM) to bridge the visual encoder and denoiser, enhancing representation quality via hierarchical aggregation and modulation. Extensive experiments on multilingual signature datasets demonstrate that \emphSigDiffuser significantly outperforms both state-of-the-art SSL methods and HSV models. We further show that increasing the crumpling severity on signatures effectively improves representation quality. Notably, our work demonstrates that extreme destruction, coupled with self-supervised diffusion models, can yield superior signature representation compared to traditional mild deformations, paving new avenues for HSV and other biometric domains. Code will be publicly available.
[CV-64] Rain Beyond Streaks: Cascaded Realistic Rain and Wet Surface Synthesis via Scene Priors WACV-2027
链接: https://arxiv.org/abs/2611.07793
作者: Melissa Tijink,Kirill Tolstobrov,Shunxin Wang,Estefania Talavera,Duc V. Le,Luuk Spreeuwers,Shaodi You,Nicola Strisciuglio
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: WACV-2027
Abstract:Realistic rain synthesis is a persistent challenge for vision systems operating in adverse weather. Existing methods primarily focus on generating rain streaks, often neglecting consequent surface effects such as reflections and puddles, which are critical for perceptual realism and robustness of downstream vision tasks. We introduce CascadeRS, a rain synthesis framework that leverages depth and segmentation priors to render scene-consistent rain and surface effects. In addition to rain streaks, CascadeRS produces reflections and puddles aligned with scene semantics through a physics-inspired zero-shot, controllable generative pipeline. It produces photorealistic rainy urban scenes, with high perceptual realism, which we validate qualitatively with an extensive user study and quantitatively. We produce the Cityscapes-CascadeRS dataset, to study downstream task performance under real rainy conditions. Our results establish CascadeRS as a scalable and realistic rain-generation framework for evaluating and enhancing the robustness of vision in adverse rainy weather. Code and dataset are available via this https URL
[CV-65] Adapters as Data: The Latent Space of LoRAs NEURIPS2026
链接: https://arxiv.org/abs/2611.07775
作者: Moritz Grävinghoff,Konstantin Wehmeyer,Aron Asefaw,Boris Knyazev,Damian Borth
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the NeurIPS 2026 Workshop on Neural Network Artifacts as a New Data Modality
Abstract:LoRA adapters are usually treated as finished artifacts rather than as a data modality. Prior work has shown that LoRA adapters populate a low-dimensional manifold that a linear PCA basis already organizes well enough to support different operations such as sampling, editing, and inversion. We ask what a learned, nonlinear latent adds beyond this. We train a transformer autoencoder on the same population of LoRA identity adapters, mapping each into a latent space and compare every operation against the identical operation applied directly to the raw adapter weights. The latent preserves the capabilities already available in weight space, sampling diverse identities and editing individual attributes. Beyond preservation, we show that the latent is far more robust to displacement than weight space, degrading into distorted but still valid faces rather than texture. Fixed-dimensional weight-space representations cannot natively host adapters of more than one rank, so we extend our sequential autoencoder to hold mixed-rank adapters within one shared latent. This gives a path toward treating adapter weights of varying architecture as a single, general modality rather than one tied to a fixed rank, and points to a learned, nonlinear latent as the mechanism that could carry this generalization further, toward heterogeneous populations of adapters.
[CV-66] BiGrounder: Bidirectional Training Unlocks the Potential of GUI Grounding
链接: https://arxiv.org/abs/2611.07765
作者: Dengyao Shi,Tianyu Fu,Fei Hu,Anyang Su,Hanning Wang,Shuo Wang,Minghui Wu,Chenxu Zhao,Zhen Lei
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Preprint, 22 pages, 10 figures, including appendix
Abstract:Graphical User Interface (GUI) grounding aims to map natural language instructions to precise operational coordinates, serving as a critical task for autonomous agent interaction. Current approaches formalize GUI grounding as predicting action-coordinate pairs from instructions, with end-to-end training supervised solely by final coordinates. Training datasets typically provide only a small number of descriptions per target coordinate, whereas real-world scenarios involve a much larger set of varied descriptions for the same target, which limits robustness to linguistic variation. While recent GUI grounding paradigms have begun employing reinforcement learning to enhance localization accuracy, models with limited exploratory capability before reinforcement yield constrained improvements even with identical training data during the reinforcement phase. To overcome these limitations, we propose a bidirectional reward paradigm that enables model self-reinforcement. During SFT, bidirectional training equips the model to generate action descriptions from target coordinates. During RL, the model samples semantically consistent instruction variations on-policy and grounds each variation back to its target. A cycle-consistency gate couples semantic quality to grounding success, turning the model’s own instruction policy into a source of continually refreshed reinforcement trajectories. Our 4B-parameter model achieves competitive accuracy across five grounding benchmarks in our evaluation, indicating the effectiveness of the proposed training paradigm. this https URL
[CV-67] Comparing Classification and Feature Learning by Analyzing Their Loss Functions in Normalized Spaces with Different Scale Factors
链接: https://arxiv.org/abs/2611.07741
作者: Qiufu Li,Huibin Xiao,Linlin Shen,Weijia Fan,Xi Jia,Yongdong Huang
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Classification and feature learning are two closely interrelated fundamental tasks. Feature learning in supervised settings is typically achieved by training classification models, under the common assumption that higher classification accuracy implies better feature quality. In this paper, we first point out that classification accuracy reflects the sample-wise local properties of a model on a dataset, whereas feature learning aims to learn features with good dataset-level global properties, such as intra-class compactness and inter-class separability. From the first principle of feature learning, we redesign the binary cross-entropy (BCE) loss, a counterpart of the cross-entropy (CE) loss used in classification. Then, in normalized feature space, we analyze the minima of the BCE and CE losses and show that both can lead to neural collapse, where sample features from different classes become maximally compact and maximally separated. More importantly, by investigating the convergence behaviors of the two losses as the scale factor \gamma varies in the normalized space, we reveal how classification accuracy and feature quality change differently across spaces of different sizes. Under a fixed training strategy, a very large \gamma achieves perfect classification accuracy but slows convergence toward the ideal feature geometry, whereas a very small \gamma makes classification highly sensitive to small perturbations. A moderate \gamma is therefore necessary for simultaneously achieving reliable classification and well-structured features. Extensive experiments validate these findings and demonstrate that BCE loss exhibits advantages in long-tail and open-set recognitions, as it better aligns with the global requirements of feature learning.
[CV-68] Seeing Beyond the Next Step: World-Model-Guided Human-Like Navigation in Multi-Agent Scenes NEURIPS2026
链接: https://arxiv.org/abs/2611.07730
作者: Seong-Eun Hong,JuYeong Hwang,Jinhyun Kim,Hanyoung Jang,HyeongYeop Kang
类目: Graphics (cs.GR); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: Accepted to NeurIPS 2026. 37 pages, 13 figures, 12 tables
Abstract:Selecting a next motion in a populated scene is not just a matter of avoiding what is currently in the way: an action can be locally feasible yet drift onto the road instead of the crosswalk, cut into a stream of crossing pedestrians, or bottleneck a narrow passage. These failure modes are invisible from the current state alone, and acting on them requires reasoning about what each candidate action will cause. World models offer exactly this, namely action-conditioned predictions of the future, but rolling them out at every decision step is a cost deployable navigation systems cannot pay, and the field has largely traded foresight for speed. We argue this is a false dichotomy. The natural place for a world model in human-aware navigation is at training time, as a preference oracle that distills foresight into a fast policy, rather than at inference time, as an online planner. We pair a discrete visual action prior with two task-aligned world models, one for nearby social dynamics and one for future scene semantics, that score candidate actions by their predicted consequences and refine the policy through group-relative preference distillation. At deployment, both world models are removed and the policy acts in a single forward pass. Across five held-out evaluation scenarios, this yields navigation that is simultaneously safer, more socially and group-coherent, and more goal-directed than strong reactive and learned baselines, with naive observers rating the resulting motion as more human-like. Taken together, our findings point to a different approach for bringing world-model foresight into deployable agents: not accelerating rollouts, but amortizing them into the policy so that the cost of looking ahead is confined to training. Project page: this https URL
[CV-69] Multimodal Information Interactions Bottleneck for Multimodal Self-Supervised Learning
链接: https://arxiv.org/abs/2611.07718
作者: Liangjian Wen,Jiangtao Zheng,Jiang Duan,Qun Dai,Yong Dai,Jianzhuang Liu,Jun Wang,Shu-Dong Huang,Xinwang Liu,Zenglin Xu,Zhao Kang
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 15 pages, 3-page supplementary material provided as an ancillary file. Accepted for publication in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
Abstract:Multimodal self-supervised learning is challenged by heterogeneity, unequal informativeness, and complex information interactions across multimodal data sources such as images, text, audio, and tabular data. While contrastive methods effectively align shared semantics, they typically rely on strong multi-view redundancy assumptions, overlooking modality-specific and synergistic information, thus leading to suboptimal representations. We propose the Multimodal Information Interactions Bottleneck (MIIB), a principled framework grounded in Partial Information Decomposition that explicitly encodes redundant, unique, and synergistic information within multimodal interactions. MIIB employs a dual-path model architecture for modality-specific encoding and multimodal fusion, optimized using surrogate tractable entropy-based objectives that enforce minimal sufficient representations without requiring explicit mutual information estimation. By leveraging variational bounds and score-based entropy estimation, MIIB efficiently captures complex cross-modal interactions while suppressing superfluous information. Experiments across diverse benchmarks demonstrate consistent improvements over state-of-the-art multimodal self-supervised methods. Code is available at this https URL.
[CV-70] Selective Event Evidence and Persistent Temporal Priors for Event-Based Point Tracking
链接: https://arxiv.org/abs/2611.07713
作者: Xiaokai Song,Zhang Li,Leqi Liu,Yibin Ye,Kun Wang,Xichao Teng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Event cameras enable tracking under fast motion and challenging illumination. However, stable long-term point tracking remains challenging: event activity is sparse and varies with scene content and motion, complicating correspondence, while trajectory estimates can drift over long sequences. Together, these challenges motivate a tracker that selectively uses informative event evidence and retains trajectory constraints across temporal windows. We propose Selective Event Evidence and Persistent Temporal Priors (SEPT), an event-only framework for tracking any point (TAP) over long sequences. SEPT introduces two complementary modules: Task-Adaptive Event Evidence Modulation (TEEM) and Persistent Trajectory Prior (PTP). Using only tracking supervision, TEEM learns a bounded reliability field that weights the event tensor across space, time, and channels before feature extraction. PTP carries geometric information across overlapping windows by marginalizing a fixed local quadratic trajectory model with a Schur complement. Its visibility-aware correction refines overlap coordinates, while a clipped terminal velocity initializes coordinates at newly exposed timestamps. SEPT builds on the ETAP architecture. SEPT is trained exclusively on synthetic EventKubric data and evaluated on five datasets across TAP and feature tracking. Results show that SEPT outperforms the retrained ETAP baseline in average Jaccard for TAP and in feature age and expected feature age for feature tracking; it also surpasses several methods combining event and frame inputs on EC and EDS. The source code is publicly available at this https URL.
[CV-71] Fast and Accurate Hand Estimation in Pixels Interaction and World Space
链接: https://arxiv.org/abs/2611.07711
作者: Jinkun Cao
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 10 pages of main text plus 15 pages of supplementary material; 5 figures; 12 tables; code available at this https URL
Abstract:Monocular hand mesh estimators commonly predict rig parameters and project the mesh to obtain 2D keypoints. We find that directly predicted 3D joints agree better with 2D predictions, and fitting a mesh to predicted points reduces mesh error. We present FAH (Fast and Accurate Hand Estimation), a family of models built on SAM 3D Body’s hand branch through fine-tuning and feature distillation. The models predict corresponding 2D and 3D joints and surface points, fit MANO or MHR-Hand, and optionally estimate anatomical parts and contact without object geometry at inference. Causal refinement reduces acceleration error, and camera estimation places the hands in world space. Experiments show improvements in image-space alignment, hand geometry, contact estimation and world-space tracking over the evaluated baselines. The default ViT-L model runs at 65 fps with mesh fitting on an RTX 4070 Ti SUPER, or 49 fps with parts and contact enabled. The smallest ViT-S model reaches 125 fps without mesh fitting, parts or contact. Code: this https URL.
[CV-72] AffineGS: Unlocking Accurate Geometry in 3DGS via Spatially Varying Affine Rectification NEURIPS2026
链接: https://arxiv.org/abs/2611.07709
作者: Tze Ho Elden Tse,Jizong Peng,Richard Chen,Angela Yao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accpeted to NeurIPS 2026. Project page: this https URL
Abstract:3D Gaussian Splatting (3DGS) has become the leading approach for photorealistic novel view synthesis, yet its geometric accuracy lags behind its visual quality. Existing depth-guided methods attempt to resolve this by regularizing the optimization with monocular depth priors, typically using a single global scale-and-shift alignment. Crucially, we discover a fundamental flaw in this assumption: modern monocular depth networks systematically exhibit region-dependent, piecewise-linear distortions. This discovery invalidates standard global alignment strategies, which warp scene geometry by enforcing a single transform. To address this, we introduce a spatially varying affine rectification model that corrects monocular disparity into consistent metric depth using regionally smooth, edge-aware fields. We further derive a normal-aware regularizer that uses spatial gradients to couple the rectified disparity with 3DGS depth and normals. Across diverse experimental settings, our sensor-free method shows strong photometric quality of Gaussian Splatting while substantially improving geometric accuracy and outperforming state-of-the art geometry-focused methods.
[CV-73] FlyingDrones: A Dataset and Benchmark for Optical Flow Estimation from UAV motion NEURIPS2026 NEURIPS
链接: https://arxiv.org/abs/2611.07705
作者: Laurenz Ruzicka,Roman Pflugfelder,Daniel Cremers
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026, this https URL
Abstract:Optical flow is an important cue for drone analysis in aerial surveillance, airspace monitoring, and long-range target tracking, where motion remains informative when appearance is weak. Estimating the motion of small Unmanned Aerial Vehicles, however, is difficult because distant targets occupy few pixels, have low contrast, and move against complex dynamic backgrounds. Standard optical-flow benchmarks emphasize large, object-centric motions and therefore do not characterize distant aerial targets well. We introduce FlyingDrones, a generated benchmark with dense ground-truth optical flow and segmentation masks derived from physically based rendering, and define resolution-normalized metrics (EPE Norm) that normalize flow by image dimensions. We evaluate 16 state-of-the-art model families on a motion- and size-stratified subset of 999 flow pairs at three resolutions. Every family recovers drone motion better than the zero-flow reference, but the gap between families widens with resolution: at 1920x1080, global-matching, recurrent, transformer, and short-window temporal models separate clearly from pyramid, classical, self-supervised, lightweight, and diffusion models. Within-family checkpoint spread is comparable to or larger than between-family spread, so the training curriculum is as important as the architecture. Two complementary studies test the benchmark beyond synthetic accuracy: a ground-truth-free motion-alignment evaluation on 21 cameras of real UAV footage shows that the synthetic ranking carries a moderate, significant signal about real-world performance (Spearman \rho = 0.63 over 41 checkpoints) while the exact top ranks differ, and a fine-tuning study shows that 800 FlyingDrones pairs reduce held-out drone-region error by up to 37% and improve drone localization on the real footage for all four architectures tested.
[CV-74] InfMasking: A Unified Variational Framework for Synergistic Multimodal Learning
链接: https://arxiv.org/abs/2611.07702
作者: Liangjian Wen,Qun Dai,Jiang Duan,Yong Dai,Jianzhuang Liu,Dongkai Wang,Jun Wang,Fuyuan Cao,Zenglin Xu,Zhao Kang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted for publication in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 16 pages; supplementary material provided as an ancillary file. Code: this https URL
Abstract:Multimodal models are expected to exploit synergistic information, that is, information available only when multiple modalities are observed together. In practice, however, a fusion module can fit the training objective by relying on a dominant modality or on information that modalities share, and the synergy between modalities is often left unused. We propose InfMasking, a unified multimodal learning framework based on Infinite Masking. Before fusion, InfMasking randomly masks the features of each modality; each sampled mask produces a different partial combination of cross-modal evidence and a fused representation of the corresponding modality interactions. All masked fused representations are required to remain consistent with a reference target: the unmasked fused view when training is unsupervised, and the class proxy of the task label when labels are available. Maintaining this consistency across diverse partial combinations discourages reliance on a dominant modality or on shared information alone, and drives the fusion module to learn cross-modal interaction patterns that remain meaningful under partial observations. Since the objective is an expectation over infinitely many masks, we model the masked fused representations as Gaussian and optimize closed-form variational bounds, whose population-moment gaps are controlled by the masking-induced covariance under this model. On the evaluated benchmarks, InfMasking improves predictive performance and yields higher mean PID estimates of task-label synergy in learned modality-specific representations, alongside higher InterSHAP interaction attribution in the trained predictors. Code is available at this https URL.
[CV-75] DEQ-Det: Dual Evidence Queries for Few-Shot Analog IC Device Detection
链接: https://arxiv.org/abs/2611.07682
作者: Zidong Zhao,Rong Huang,Chengxing Fan,Jie Hu,Malik Haris,Ghazal A. Fahmy,Morteza Mousazadeh,Guoqiang Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Detecting analog integrated circuit (IC) devices in layout images supports circuit analysis and inspection. With only a few annotated examples, variation in device geometry and visual similarity to dummy or routing structures make reliable detection difficult. Foreground similarity alone can highlight plausible targets without distinguishing them from competing layout patterns. We propose Dual Evidence Queries for Detection (DEQ-Det), which constructs complementary foreground and background references from the same support images without additional annotations. Foreground references capture target morphology, while selected background references characterize surrounding layout patterns. Both sources guide query selection and region verification, combining target appearance with layout context. Across five held-out device categories and three random seeds, DEQ-Det achieves macro AP50 of 39.41 and 53.03 at 10 and 30 shots, exceeding the strongest listed baselines by 3.97 and 9.71 points, respectively. A matched 30-shot ablation further shows that RoI background evidence reduces pooled false positives at comparable recall, supporting its role in interference suppression.
[CV-76] GEKey: Learning Geodesic Eccentric 3D Keypoints via Self-Supervision BMVC2026
链接: https://arxiv.org/abs/2611.07677
作者: Pierre Onghena,Santiago Velasco-Forero,Beatriz Marcotegui
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to BMVC 2026 (Oral)
Abstract:Point clouds represent 3D object geometry as sets of spatial coordinates. A fundamental task in their analysis is keypoint detection, i.e., identifying a sparse subset of points that captures the most salient features. While supervised methods rely on manual annotations, unsupervised approaches leverage intrinsic geometric properties to localize keypoints without human input. We introduce GEKey, a self-supervised framework that uses the geodesic eccentricity transform, an SE(3)-invariant geometric measure, to identify instance-specific salient points. These points serve as pseudo-labels to train a model that aligns its predictions with the object’s intrinsic structure. GEKey is simple, efficient, and robust, preserving intra-class variation by balancing instance-specific and category-level structure. Extensive experiments on ShapeNetCoreV2, KeypointNet, CAPE, ITOP, and DeformingThings4D show state-of-the-art performance in both keypoint detection and alignment, with strong resilience to noise, downsampling, deformations, and real-world conditions. The code and pretrained models are available at this https URL.
[CV-77] Egocentric Action Recognition with Objects and Vision-Language Affordances
链接: https://arxiv.org/abs/2611.07673
作者: Raúl Boix-Aldeguer,Manuel Benavent-Lledo,David Ortiz-Perez,Jose Garcia-Rodriguez
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 9 pages, 6 figures. Under review at IEEE FG 2027
Abstract:Action recognition, and more specifically egocentric action recognition, is a core task in computer vision, with applications such as monitoring the daily life activities of elderly people and task assistance. Recent methods rely on visual modalities such as RGB video or depth. However, in egocentric video, an action is largely defined by its objects, which these methods rarely model. We propose a method that combines a frozen SlowFast video expert with an attention-based object expert by late fusion. Each object is represented with its affordances, that is, what can be done with it, generated zero-shot by a vision-language model. Since datasets such as MECCANO annotate only the objects involved in each action, evaluating with these boxes tells the model which objects are active. We therefore train on annotated boxes and evaluate the method with detections from the official active-object detector, and with annotated boxes as an upper bound. Without affordances, our fusion reaches 68.4% Top-1 with annotated boxes but 49.0% with detections. About half of this drop comes from non-active detections. To remove them, the same vision-language model generates an interaction affordance that decides whether a hand is manipulating each detected object. On MECCANO, objects raise the Top-1 of the video expert from 44.9% to 49.0%, and the two affordances raise it to 49.9%, above SlowFast and object-aware RGB baselines. Code will be released upon acceptance.
[CV-78] CoGenRec: A Unified Generative Framework for Complementary Vision Reconstruction
链接: https://arxiv.org/abs/2611.07665
作者: Yapeng Meng,Yuguo Chen,Xiangru Chen,Shaojun Xu,Xuezhen Wang,Lijian Wang,Zheyu Yang,Yihan Lin,Luping Shi,Rong Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Capturing scenes with extreme motion and high dynamic range is fundamentally constrained by trade-offs among temporal resolution, spatial detail, signal-to-noise ratio, and dynamic range under practical bandwidth, power, and quantum-efficiency limits. Conventional frame-based cameras therefore suffer from temporal undersampling, exposure-induced motion blur, and sensor saturation. The Complementary Vision Sensor (CVS) addresses these limitations by decomposing visual information across two complementary pathways, a spatially dense, temporally sparse, long-exposure RGB pathway and a temporally dense, spatially sparse, short-exposure patiotemporal difference pathway. Their complementary measurements provide appearance, structural, and motion cues across a wide dynamic range that enable reconstruction beyond conventional imaging. We present CoGenRec, a complementarity-driven generative reconstruction framework that unifies video deblurring, high-temporal-resolution frame interpolation, and HDR fusion. CoGenRec encodes complementary measurements into multi-scale spatiotemporal features using exposure-aware appearance modeling and bidirectional temporal propagation. These features constrain a video diffusion model and are injected into the VAE decoder to recover fine structures lost during latent-space compression, combining the scene fidelity of physical measurements with the appearance recovery capabilities of learned video priors. On SportsSloMo-CVS, CoGenRec improves PSNR over the strongest general-purpose baseline by 4.40 dB for deblurring, 3.89 dB for frame interpolation, and 5.04 dB for HDR fusion, respectively. It also achieves strong perceptual quality on 129 unseen scenes captured with real CVS. These results establish a unified approach to advance high-quality reconstruction under challenging motion and illumination conditions. Project page: this https URL.
[CV-79] Which Objects Stay? Learning Hierarchical Persistence Beliefs for Lifelong Open-Vocabulary Object Maps
链接: https://arxiv.org/abs/2611.07662
作者: Thanh Nguyen Canh,Haolan Zhang,Nak Young Chong
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注:
Abstract:Lifelong object maps must estimate whether an object remains at its previously observed location after it leaves view. Existing persistence filters rely on class-based survival priors, making them difficult to apply reliably in open-vocabulary settings. We introduce HiPer, a hierarchical persistence filter that derives persistence beliefs directly from data rather than semantic categories. For a new object, HiPer retrieves visually and geometrically similar training instances and estimates an initial survival rate from their observed histories, without requiring class labels. This estimate initializes a conjugate Beta belief over the object’s survival probability, which is updated in closed form using detections and missed observations under Rosen’s persistence-filter framework. When an object is re-identified at a new location, its persistence belief is transferred with the object, enabling the map to learn object-specific mobility over time. Experiments on 120 held-out simulated sequences show that HiPer consistently achieves the best performance among competing methods. Retrieval-based priors outperform CLIP-derived label priors, and object-specific online learning provides further gains. Results on 47 held-out 3RScan environments demonstrate that the proposed retrieval-based prior also outperforms state-of-the-art baselines in real-world settings.
[CV-80] ArtCrafter: Feed-Forward Generation of Articulated 3D Object with Analytic Joint Derivation
链接: https://arxiv.org/abs/2611.07658
作者: Minh Canh Tu,Quang-Binh Nguyen,Khoi Nguyen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Articulated 3D assets are essential for robotic simulation and training, yet their generation from casual images poses a key challenge: recovering part geometry and kinematics that remain mutually consistent. We present ArtCrafter, a novel feed-forward generative model for controllable generation of articulated 3D objects from a pair of images, one depicting the object in the rest state and the other in the open state. At the core of ArtCrafter is a DiT-based generative model that jointly produces part-level meshes for both articulation states in a single forward pass. Inspired by Slot Attention, its denoising process operates on a 1+2N slot layout comprising one shared base slot and N paired rest/open slots, encouraging each pair to bind to the same kinematic part. To ensure geometrically consistent and disjoint part generation, we introduce Triplet Attention to enforce cross-state geometric consistency between paired rest/open slots anchored through the shared base slot, and a repulsion loss \mathcalL_\mathrmrep on decoded part point clouds to prevent inter-part penetration throughout denoising. Given the resulting paired meshes, URDF joint parameters—type, axis, origin, and range—are derived analytically from the rigid transformation between each rest/open pair, providing an interpretable alternative to learned joint-parameter regression. Extensive experiments on both in-distribution and out-of-distribution datasets demonstrate that ArtCrafter outperforms existing method under the same dual-state setting in geometry quality and articulation precision, while remaining competitive with methods operating under different input settings and paradigms.
[CV-81] Generator-Aligned Reward Modeling with Target-Aware Bounded Optimization for Visual Generation
链接: https://arxiv.org/abs/2611.07648
作者: Ziqiao Weng,Yuanzhi Liang,Haibing Huang,Weidong Cai,Xuelong Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 24 pages
Abstract:Post-training aims to enhance desired attributes in visual generative models while preserving their general generation capabilities. Reward-guided post-training faces two challenges: real training data may mismatch the generator’s output distribution, and attribute supervision may be confounded by unrelated content. Unbounded reward maximization can also cause overoptimization and degrade generation quality. We propose a framework combining Generator-Aligned Reward Modeling (GARM) and Target-Aware Bounded Optimization (TABO). GARM mitigates distribution mismatch through generator-support matching and reduces content confounding using content-matched real contrasts, yielding generator-relevant, attribute-focused reward supervision. TABO maps the learned reward to a smooth, upper-bounded function calibrated from real target-positive examples. This transformation is independent of the optimization algorithm and supports both direct reward-gradient and policy-based optimization. We further introduce Target-Aware GRPO (TA-GRPO), an adaptation of Flow-GRPO that combines fixed stage-wise advantage scaling, a generator-to-target curriculum, and timestep-balanced credit assignment for flow-based video optimization. Experiments show that GARM improves target-attribute expression while maintaining competitive generation quality. TABO better balances attribute enhancement and visual quality than unconstrained reward maximization, while TA-GRPO improves target enhancement over vanilla Flow-GRPO and reduces unintended attribute leakage on neutral prompts. These results highlight the importance of generator-aligned reward supervision and controlled, data-calibrated optimization for targeted visual generation.
[CV-82] UniGADD: Universal Generator-Agnostic Deepfake Detector
链接: https://arxiv.org/abs/2611.07647
作者: Bernard Tomczyk,Agil Aghasanli,Plamen Angelov
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:In this paper, we introduce Universal Generator-Agnostic Deepfake Detector (UniGADD), a novel method that employs supervised learning to achieve high accuracy on known deepfake generators, while maintaining robust performance on previously unseen ones. The proposed approach follows a two-stage optimisation process. In the first stage, a contrastive loss encourages the model to learn discriminative feature embeddings from real and known fake images, resulting in strong performance within the training domain. In the second stage, the embedding space is refined by promoting inter-cluster separation and intra-cluster compactness, applied exclusively to real samples. This refinement enhances generalisability, enabling the method to exhibit improved robustness against unseen deepfake generation techniques. UniGADD achieves accuracy on par with state-of-the-art methods for known generators, while significantly outperforming them on unseen cases, demonstrating its scalability and practical applicability for adversarial content detection.
[CV-83] Neural Image Compression with Internal-External Prior Fusion and Alignment
链接: https://arxiv.org/abs/2611.07644
作者: Tiancheng Si,Wenlong Xu,Dahua Gao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: IEEE International Conference on Visual Communications and Image Processing 2026 (VCIP 2026)
Abstract:Existing context models for learned image com- pression rely primarily on internal priors from the current image, overlooking common patterns and typical textures shared across natural images. To address this limitation, we propose an Internal-External Prior Fusion-Alignment (IE-PFA) context model built upon quadtree spatial-channel joint modeling, where a learnable dictionary shared by the encoder and decoder serves as an external prior. Specifically, an RWKV-based Spatial- Channel Prior Fusion (RSCPF) module efficiently fuses inter- nal and external priors along spatial and channel dimensions with linear complexity, while a training-only Self-Distilled Prior Alignment (SDPA) mechanism aligns them without inference overhead. Moreover, our model supports variable-rate coding over 64 quality levels within a single network and is accelerated by custom CUDA operators. Experiments on seven benchmarks demonstrate state-of-the-art rate-distortion performance, achiev- ing an average BD-Rate saving of 21.65% over VTM-22.0-Intra, with only 29.9M parameters and the fastest coding speed among advanced neural codecs (61 ms encoding and 52 ms decoding). Code is available at this https URL.
[CV-84] RACEMARK: Editor-Aware Watermarking for User Attribution in Text-Guided Image Editing
链接: https://arxiv.org/abs/2611.07632
作者: Xiaodong Wu,Xiangman Li,Jianbing Ni,Rongxing Lu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:An image editing service records which user requests each edit, but the edited image itself does not contain this account information. Once the image is downloaded and shared, the service cannot identify the requesting user from the image alone. Embedding a user-specific watermark could enable such attribution, but semantic editing may erase the signal. We propose Tracemark, an editor-aware watermarking framework for identifying users from edited images. A shared encoder embeds a user-specific signal before editing, and a shared detector recovers the user by comparing features of the edited image with learned reference vectors. Both components are trained jointly through an editor with fixed parameters. User identification errors encourage the encoder to produce watermarks that survive editing, while fidelity losses limit changes to the input image and intended edit. Each image domain and editing instruction uses a separately trained model shared across registered users. Experiments on three image domains show that editor-aware training improves user recovery with both reference-vector matching and binary watermark decoding. Separately trained models reach 87.5% mean identification accuracy among 500 users. When tested with a different editor without retraining, the watermark modules still recover user information. These results demonstrate that learning watermarks through the editing process enables user attribution without modifying the editor.
[CV-85] owards Real-Time Full-Waveform LiDAR Transformers via Intensity-Guided Token Reduction and Physics-Aware Augmentation NEURIPS2026
链接: https://arxiv.org/abs/2611.07613
作者: Kotaro Oishi,Kazuma Ikeda,Ryosei Hara,Ryo Yoshida,Mariko Isogawa,Kentaro Yoshioka
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to NeurIPS 2026 (Main)
Abstract:LiDAR is a critical sensor for autonomous driving and robotics, yet it remains vulnerable to adverse conditions such as fog, rain, transparent objects, and low-reflectance surfaces. Full-waveform LiDAR (FWL) addresses these limitations by capturing the complete return waveform as a histogram, preserving rich reflection characteristics including multi-path echoes and weak signals. While recent Transformer-based approaches have demonstrated promise on FWL data, two fundamental bottlenecks remain: high computational cost from redundant tokens in low-intensity background regions, and limited generalization due to scarce annotated training data. We address both bottlenecks by exploiting the physical structure inherent to FWL data. First, we propose FWL-ToPM, a token pruning and merging mechanism that leverages intensity distributions to selectively reduce tokens while retaining critical waveform content. Second, we introduce FWLAug, a data augmentation framework designed as a regularization method that preserves waveform structure, retaining the Time-of-Flight waveform of each ray and the frustum geometry, to improve generalization to unseen scenes. Evaluated on the Ghost-FWL benchmark, FWL-ToPM achieves over 8 times speedup at real-time throughput alongside a nearly 9-point F1 gain over the prior state of the art, establishing the first real-time FWL Transformer that simultaneously improves accuracy. FWLAug further improves performance on both the standard benchmark and an out-of-distribution test set of unseen scenes. Together, these methods push FWL-based perception toward practical deployment. The code and the OOD dataset are publicly available and can be accessed via the project page: this https URL
[CV-86] FisherDistill: Diffusion Distribution Distillation via Closed-Form Fisher Critics
链接: https://arxiv.org/abs/2611.07604
作者: Mahesh Bhosale,Yuanhao Zhai,Siwei Lyu,Junsong Yuan,David Doermann,Defang Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent advances in diffusion distillation have enabled high-quality image generation in only one or a few network function evaluations (NFEs). Despite their efficiency, aggressively distilled generators often retain a residual distribution gap from their multi-step teachers, motivating recent methods to introduce adversarial refinement with a learned discriminator. In this work, we ask whether such refinement necessarily requires a learned adversarial discriminator. We introduce \textbfFisherDistill, a simple alternative based on Fisher linear discriminant analysis. Using intermediate representations from the frozen teacher, we estimate the first- and second-order statistics of real and generated samples and derive a closed-form Fisher direction that maximally separates their feature distributions relative to within-class variation. This direction provides a lightweight refinement signal that complements the original distillation objective without requiring an additional adversarial discriminator or its optimization loop. Across multiple distillation frameworks, our approach consistently improves generation quality over their respective distilled baselines and substantially closes the gap to adversarially refined counterparts. Code: this https URL
[CV-87] ACT-KV: Tri-Axis Cosine Transform for Compressing Volumetric KV Caches in Medical VLMs NEURIPS2026
链接: https://arxiv.org/abs/2611.07593
作者: Chongyu Qu,Ritchie Zhao,Yufan He,Dong Yang,Zhengyi Lu,Junchao Zhu,Tianyuan Yao,Juming Xiong,Junlin Guo,Yanfan Zhu,Yuechen Yang,Daguang Xu,Bennett Allan Landman,Yucheng Tang,Yuankai Huo
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at NeurIPS 2026
Abstract:Vision-language models (VLMs) applied to 3D medical scans face a structural memory wall. A single CT or MRI volume produces tens of thousands of vision tokens, and the resulting key-value (KV) cache scales linearly with both batch and context. Existing KV compression methods operate on the cache’s feature or token statistics, or apply 1D and 2D transforms, leaving the three-axis spatial structure in the cache unexploited. We propose TACT-KV (Tri-Axis Cosine Transform), a calibration-free KV compression method that exploits strong spatial correlations in visual KV states along all three anatomical axes of 3D scans. This correlation concentrates the KV cache energy under a frequency-domain transform. TACT-KV applies a separable 3D discrete cosine transform to the KV cache, allocates a per-layer bit budget across frequency bins via reverse water-filling (more bits to high-energy components), and quantizes each bin with a fixed scalar codebook. We derive a close-form relation: the distortion reduction from adaptive bit allocation is dominated by the spectral concentration of the transformed cache. Across six medical VLMs and two 3D CT benchmarks, TACT-KV achieves the lowest 1-bit prediction-flip rate on eleven of twelve model and dataset pairs in our main comparison. The same advantage reproduces on four general-purpose VLMs evaluated on video tasks, indicating that the benefit of exploiting three-axis structure generalizes beyond medical imaging. At equal high-bandwidth memory (HBM), TACT-KV scales the decoding batch size on a single GPU by up to 8\times .
[CV-88] Subspace Learning with Outlier Exposure for Out-of-Distribution Detection
链接: https://arxiv.org/abs/2611.07580
作者: Hongxu Shen,Kun Fang,Jiahui Zheng,Jian Sun,Qinghua Tao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Out-of-Distribution (OoD) detection is pivotal for reliable deployments of Deep Neural Networks (DNNs), aiming to distinguish anomalous samples from the In-Distribution (InD) data in training. By introducing outliers as surrogate OoD data to finetune DNNs, Outlier Exposure (OE) has greatly advanced OoD detection in recent years. Existing OE-based research remains predominantly logit-centric, either by generating diverse outliers or by regularizing them with objectives operating at the prediction-output level. Nonetheless, these approaches leave much of the rich information in feature spaces untapped. In this work, we propose an OE-based detection framework, named RecOE, that learns appropriate feature subspaces from InD data and auxiliary outliers, such that the InD-OoD discrepancy can be effectively characterized through the learned feature subspaces. Specifically, the feature subspaces are optimized to capture the dominant variations of InD and outliers by minimizing reconstruction errors therein. Besides, a sample-level objective discourages reconstruction in the opposing subspace, while an orthogonality regularizer reduces subspace overlap, further promoting discriminative subspaces. At inference, the difference in reconstruction errors between the learned subspaces serves as the OoD detection score. Extensive experiments show the superiority of RecOE on multiple benchmarks, in which significant performance gains are achieved over recent strong baselines with only slight extra overhead from two small-size matrices of subspace bases. The source code is available at this https URL.
[CV-89] Georeferenced Cross-View 3D Reconstruction from Ground and Satellite Images NEURIPS2026
链接: https://arxiv.org/abs/2611.07575
作者: David G. Shatwell,Mubarak Shah
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted in NeurIPS 2026
Abstract:Georeferenced 3D reconstruction from ground images requires recovering not only the local scene geometry, but also localizing every camera center and 3D point in a geographic coordinate frame. Existing cross-view geo-localization methods are able to estimate the GPS coordinate of a single ground camera, but do not reconstruct dense 3D geometry, while recent feed-forward multi-view 3D reconstruction models predict point maps and camera poses only in a local coordinate frame. We introduce X-VGGT, a feed-forward cross-view geometry transformer that unifies these tasks by jointly processing a set of ground-level images and a georeferenced satellite image to predict ground-view camera poses, dense 3D point maps, and a scene-level transform parameterized by gravity, scale, yaw and translation. This transform projects the local reconstruction into the satellite plane, resulting in the geo-registration of all ground images with the georeferenced satellite image. We further propose an evaluation protocol for this joint task and show that, across multiple cross-view datasets, X-VGGT produces accurate georeferenced reconstructions and outperforms prior cross-view geo-localization and multi-view 3D reconstruction models.
[CV-90] Integrating Learned Priors through Prior-Space Reparameterization for Physically Based Inverse Rendering
链接: https://arxiv.org/abs/2611.07573
作者: Hun Hwangbo,Minjun Park,Yongsung Kim,Wooseok Song,Jungmin Ko,Junghyun Hwang,Sungroh Yoon
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Physically based inverse rendering decomposes appearance into materials and illumination, an ambiguous problem for which learned material predictions provide useful prior information. Most existing methods incorporate these cues through auxiliary regularization losses, which alter the optimization objective and require balancing their influence against photometric fitting. We propose a prior-guided material reparameterization that uses the predicted feature relationships to restrict the optimization space while leaving the objective unchanged. We fuse multi-view material predictions onto a mesh and construct a permutohedral lattice in the resulting prior feature space. Surface points with similar predictions share material variables, which are optimized jointly with geometry and illumination using a purely photometric objective. Experiments with two learned predictors on three synthetic benchmarks and Stanford-ORB demonstrate improved material decomposition and relighting. Our method improves relighting PSNR by 1.5-2.5 dB over the strongest evaluated mesh-based baseline on Synthetic4Relight and TensoIR-Synthetic, and outperforms a regularization loss using the same prior in controlled comparisons. We further show that our reparameterization pools noisy photometric gradients, producing updates that are more consistent across sampled views. This consistency enables a shortened joint optimization schedule that takes less than a minute while retaining relighting quality above all evaluated baselines on TensoIR-Synthetic.
[CV-91] Guidance Is a Field Not a Scalar: Interval-Integrated Classifier-Free Guidance for One-Step Video Generation WACV2027
链接: https://arxiv.org/abs/2611.07566
作者: Zijie Meng,Xixin Cao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 9 figures. Accepted to WACV 2027
Abstract:One-step generators built on MeanFlow synthesize a video in a single function evaluation, yet inherit a rigid treatment of classifier-free guidance (CFG): the guidance scale is frozen during training or, at best, exposed as a single scalar. This conflicts with mounting evidence that the optimal guidance is not a constant but a field that varies with the noise level, with spatial content, and, for video, with frame position. Every existing remedy presupposes a multi-step sampler along which guidance can be re-scheduled–an assumption one-step models violate by construction. We resolve this tension through a single observation: because the MeanFlow target is an integral over a noise interval, an arbitrary guidance field can be absorbed into the target itself and compiled into one evaluation. We formalize this as the interval-integrated CFG identity, prove that the compiled guidance carries no schedule-discretization error and can thus match or exceed any finite multi-step sampler, and instantiate it as AGAVE, which equips the average-velocity field with a learnable temporal schedule, a free motion-saliency spatial modulation, and a frame-position asymmetry. Adding fewer than ten scalars at zero inference cost, AGAVE surpasses all one-step baselines on VBench, UCF-101 and MSR-VTT while matching multi-step guided sampling. Project repository: this https URL.
[CV-92] MetricTrail: Revisiting the Power of Metric Depth and Local Optimization for Long-Sequence Streaming 3D Reconstruction
链接: https://arxiv.org/abs/2611.07555
作者: Jingnan Gao,Jiahui Huang,Sven Elflein,Jianhao Zheng,Bingxin Ke,Xiaokang Yang,Yichao Yan,Laura Leal-Taixé,Qunjie Zhou,Jiawei Ren,Shengyu Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project Page: this https URL , Code: this https URL
Abstract:We present MetricTrail, a streaming system for accurate camera motion estimation and dense reconstruction from long monocular videos. To keep inference efficient over long sequences, we revisit DPVO’s sparse local estimator. We find that single-image metric depth substantially improves its trajectory accuracy, without retraining the correspondence network or enlarging the optimization window. Metric predictions initialize patch depths and remain as soft constraints during optimization. We further improve correspondence estimation through learned adapters that augment matching features and recurrent context with frozen DINOv3 representations. The correspondence estimator and adapters are trained on diverse short clips without additional long-sequence training. This modular design also allows the depth provider to be independently upgraded and supports supplied camera intrinsics and calibrated stereo observations without retraining. Under the reported RGB-only protocols, MetricTrail achieves the lowest average trajectory error among evaluated methods on KITTI, Oxford Spires, and VBR. Pairing the estimated poses with single-image depth predictions also produces strong dense reconstruction on indoor and outdoor benchmarks and coherent reconstructions of in-the-wild videos. With supplied intrinsics, the system processes KITTI-02 at 12.70 FPS with 3.21 GB peak GPU memory.
[CV-93] Improving Long-Tail Detection via Layout-Instruction Conditioned Diffusion Synthesis ICME2026
链接: https://arxiv.org/abs/2611.07553
作者: Xinyu Zhang,Xiaodi Wang,Yanpeng Sun,Lingfeng Yang,Jingdong Wang,Lingqiao Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at the ICME 2026
Abstract:Synthetic data from diffusion models can improve object detection, but long-tail performance is still limited by scarce rare classes and inflexible conditioning (e.g., text-only or layout-only inputs). In this paper, we propose Tail-SD, a layout-instruction conditioned stable diffusion framework that supports text, layouts, or their hybrid, enabling controllable synthesis from boxes, masks, and/or prompts to better cover the long-tail objects. To train a single model across heterogeneous conditions, we introduce a multi-task cross-attention module and multi-task embedding tokens, improving generation quality and diversity. Extensive experiments on COCO and LVIS demonstrate that diffusion synthesis with hybrid conditioning consistently benefits detection, e.g., on LVIS val, it achieves +4.89% / +5.68% rare Box/Mask AP, highlighting strong potential for addressing long-tail challenges.
[CV-94] raining-Free Predictive Spatio-Temporal Token Pruning for Few-Step Autoregressive Video Diffusion
链接: https://arxiv.org/abs/2611.07551
作者: Jia Li,Xurui Peng,Xing Wang,Hayden Kwok-Hay So
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 19 pages, 10 figures, 11 tables, including appendices
Abstract:Few-step autoregressive (AR) video diffusion models generate videos chunk by chunk and have made efficient video generation practical. However, further accelerating these models toward real-time generation remains challenging, as few-step distillation leaves little cross-step redundancy for existing training-free caching methods to exploit. We revisit this regime and identify two token-level redundancy patterns: cross-frame redundancy from static content shared across adjacent frames, and cross-step redundancy from tokens that converge early during denoising. Crucially, we find that early clean estimates along the model’s own denoising trajectory provide an effective online signal for locating both patterns, requiring no retraining, auxiliary predictor, or reference video. Based on these findings, we propose PreSTP (Predictive Spatio-Temporal Token Pruning), a training-free framework that constructs a unified token-wise redundancy mask from cross-frame similarity and cross-step convergence. Tokens marked as redundant skip subsequent computation under a consistent last-computed-value restoration rule, and a content-adaptive mask reuse strategy further reduces dense probing cost. On 5-second generation, PreSTP achieves 1.52x-3.20x speedups and a better efficiency-fidelity trade-off than coarse-grained caching baselines such as TeaCache and FlowCache. On 30-second generation, it reaches up to 1.96x speedup while improving VBench-Long and suppressing long-horizon drift. Notably, it continues to reduce inference cost even in the extreme one-step regime, where cross-step redundancy vanishes entirely and coarse step-wise caching is no longer applicable.
[CV-95] OptIMerge: Optimization-Based Merging of Diffusion Models
链接: https://arxiv.org/abs/2611.07543
作者: Viraj Shah,Ashwini Pokle,Peyman Milanfar,Mauricio Delbracio,Hossein Talebi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:The arithmetic merging of weights from multiple task-specific models fine-tuned from a shared base model, commonly known as model souping, has emerged as a popular strategy for multi-task learning without the computational overhead of joint retraining. However, naive scalar weight averaging frequently leads to severe performance degradation when the constituent models learn structurally divergent feature representations, leading to destructive parameter interference. We study this problem in a largely unexplored setting: merging image-to-image (I2I) diffusion models fine-tuned for dense, pixel-accurate tasks such as image restoration. Unlike the large language models (LLMs) and text-to-image (T2I) diffusion models for which most merging methods were developed, I2I experts encode conflicting spatial priors (e.g., anisotropic motion-blur vs. isotropic defocus-blur) in their convolutional filters, where both naive soup and LLM-style element-wise merging introduce visible artifacts. We introduce OptIMerge, a gradient-based optimization framework: instead of relying on heuristic scalars, OptIMerge learns highly expressive, yet parameter-efficient merger coefficients by minimizing a multi-task diffusion loss over a small set of 100-500 images. OptIMerge parameterizes the merge with architecture-aware transformations: for attention layers, a structured low-rank feature alignment with explicit channel biases; for convolutional layers, per-kernel merger coefficients, and a low-rank Cayley transform that aligns channels by orthogonal rotation without distorting the filters. By keeping the experts frozen and optimizing only these lightweight merger parameters (equivalent to 0.4% of the network footprint), OptIMerge achieves simultaneous multi-task capabilities that rival computationally expensive joint training and decisively outperform zero-shot model soups.
[CV-96] Action-Conditioned VLM Monitoring for Autonomous Vehicle Semantic Anomalies ICRA2027
链接: https://arxiv.org/abs/2611.07540
作者: Daniel Adejumo,Omkar Ankalkope,Kunal Runwal,Aliasghar Arab
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 7 pages. Submitted to IEEE ICRA 2027
Abstract:We study action-conditioned monitoring of autonomous-vehicle behavior in semantically unusual driving scenes. The target is whether the executed ego action is appropriate for the scene: the same scene can admit both normal and anomalous behavior, so visual unusualness alone does not determine the monitoring label. A Vision-Language Model (VLM) first predicts the scene-required action from an outcome-hidden video window; the monitor then compares that expectation with the realized ego-motion trajectory under an explicit compatibility rule. To study this monitoring task, we contribute a six-field taxonomy of 112 paired scenarios spanning both failure polarities (42 derived from documented incidents), a generated video corpus with estimated trajectories, and an exploratory two-stage monitor on that corpus. On 235 clips from 15 scenario families, the two-stage protocol with a fixed T-2.5 s window obtains 83.0% accuracy and anomaly recall 0.773, versus 74.5% and 0.536 for a single-stage video-only baseline; specificity is 0.880 versus 0.928.
[CV-97] Concertina: Embodying Manipulation and Cognition in One Diffusion Vision-Language Model
链接: https://arxiv.org/abs/2611.07513
作者: Chengcheng Wang,Wei He,Wenhao Li,Hongguang Li,Kai Han,Jianyuan Guo,Chang Xu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:Most vision-language-action (VLA) policies decouple manipulation from cognition by bolting an external action mechanism onto a vision-language model (VLM)–either quantizing actions into discrete tokens whose errors bear no stable correspondence to physical displacements, or delegating continuous trajectory generation to a separate action expert whose gradients compete with or are blocked from updating the VLM backbone. We argue that manipulation and cognition should instead be embodied in the same model, shifting from a decoupled VLA to an Embodied Vision-Language Model (eVLM). We introduce Concertina, an eVLM that places continuous trajectory representation, blockwise generation, and physical-coordinate supervision directly on the computation and parameter path of a shared diffusion VLM backbone. Each control timestep is encoded as one continuous latent action token in the backbone hidden space, and the shared Transformer reconstructs temporally ordered action blocks via the same corrupt-and-reconstruct operator used for multimodal pretraining. Active action projections account for only 0.21% of parameters in the 4.46B model with zero separate action-generator layers, compared with 16.5%-35.1% dedicated action-expert capacity in representative continuous VLAs. Using our proposed Action Interface Audit, we confirm that cross-time trajectory coupling emerges in the middle and late layers of the shared backbone rather than at the action interfaces. On standard benchmarks, Concertina achieves 98.15% average success on regular LIBERO, 85.9% on the 10,030-task LIBERO-Plus benchmark, and 4.142 average completed chain length on CALVIN ABC-D. Code: this https URL Hugging Face: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO) Cite as: arXiv:2611.07513 [cs.CV] (or arXiv:2611.07513v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.07513 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-98] RLGPA: A Reinforcement Learning-Guided Granularity Planning Attack for Decision-Based Sparse Adversarial Examples
链接: https://arxiv.org/abs/2611.07512
作者: Chao Zhou,Shenghua He,Jiaxin Mo,Qiangguo Yu,Fengchuang Xing
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Decision-based sparse attacks must reduce perturbation support using only the victim’s predicted label. The best restoration strategy varies as support contracts: coarse removal works early, while tail states need finer, history-aware tests. We propose RLGPA, an online reinforcement learning method that schedules support-reduction executors and their granularity under a hard query budget. A masked Dueling Double DQN selects among seven stable policy–scale actions; each action produces support-relative restoration groups that are rebuilt after every macro decision. A feasibility-preserving executor commits only adversarial restorations and feeds support gain and query cost back to the planner. Under the matched top-1-label oracle, RLGPA reaches 61% targeted ASR at sparsity 0.01 on CIFAR-10 (vs. 25% SparseEvo, 0% Pointwise), and 75.5%/73.0% on ResNet50/ViT at sparsity 0.005 (vs. 29.0%/15.5%). A fixed-probe test reduces median (L_0) from 165 to 23 after 40 images without any update. State-blind and state-shuffled controls degrade significantly, confirming that the learned policy uses attack state; the main gains come from state-conditioned executor scheduling and dynamic repartition. The source code is publicly available at this https URL.
[CV-99] Coverag e Hides Churn: Auditing and Preserving Conformal Decisions as VLM Label Registries Grow WACV2027
链接: https://arxiv.org/abs/2611.07502
作者: Jai Kumar Sharma,Peeyush Tapadiya
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at WACV 2027. Includes supplementary material
Abstract:Vision-language classifiers can learn new labels from text alone, so the label lists they serve keep growing. We show that this growth silently changes conformal decisions about old labels. Common softmax-based conformal scores compare each label with an average over all labels; adding labels changes that average, and with it the score of every old label, even when the model, the image, and the old calibration data stay the same. Recalibrating after the update keeps the coverage guarantee, yet across eleven settings it changes at least one old decision on 53-90% of new images. Coverage cannot reveal this, because it checks only the true label: in five controlled settings that keep the old calibration, so that only the normalizer changes, old-class coverage moves by at most 0.01 while 60-67% of images change. We propose a lifecycle audit that measures these changes alongside coverage and set size, and a simple fix, PrefixCP: each class keeps the label registry that existed when it was added. Its old decisions then never change, each class keeps exact class-conditional coverage under exchangeability, and one prefix scan computes all historical normalizers at once. PrefixCP recovers 40.9-58.5% of the set-size gap between never updating the reference and recalibrating, which gives smaller sets but changes old decisions.
[CV-100] Degree-Equalized View-Graph Selection under Tight Pair Budgets
链接: https://arxiv.org/abs/2611.07498
作者: Minhyeok Im,Jongseong Brad Choi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Fully matching all image pairs in Structure-from-Motion is quadratic in the number of images, so practical pipelines pass a budgeted subset to the final matching and reconstruction backend. One-shot selectors score candidates with image retrieval, cheap geometric probes, or a learned prior and finalize B pairs before full matching. Their priorities do not account for the degree already allocated to each image, which can concentrate edges on a few images. Across nine large 1DSfM internet photo collections at B=2N, the coverage top-B selector leaves 54% of covered images with realized degree at most two. Reconstruction can then fail even when the selected graph is nearly connected and almost fully verified. Our training-free greedy rule, Degree-Equalized Allocation (DEAL), exponentially downweights pairs whose endpoints are already well covered. On these collections, DEAL improves the one-shot selector built on the same prior (from 0.776 to 0.929) and matches the mean registered-image ratio of an SfM-supervised edge prior under the shared backend (0.929 vs 0.930), without training. On the largely saturated 38-scene standard benchmark suite, DEAL gives the highest mean among the evaluated training-free selectors (0.875 vs 0.858 one-shot).
[CV-101] Compact Motion Embeddings for Feed-forward Unified Human Animation
链接: https://arxiv.org/abs/2611.07481
作者: Seonwook Park,Hyunsoo Cha,Amrita Mazumdar,Shengze Wang,Tianye Li,Hanbyul Joo,Koki Nagano,Shalini De Mello
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 28 pages, 13 figures, 11 tables
Abstract:Animating a reference image with a driver’s motion requires a motion representation that captures face, hand, and body articulation but excludes the driver’s appearance. Most body animation methods take motion from external pose estimators and render it with many denoising steps; compact learned motion representations mostly cover the face. We present FUHA, a real-time-capable feed-forward transformer that learns a compact motion embedding of the whole human and uses it for face, upper-body, and full-body animation. A Q-Former compresses the driver’s features from a frozen visual encoder into this embedding. To disentangle motion from appearance, a novel contrastive loss pulls it toward the motion embedding of the driver’s rendered keypoint image and away from that of the same person in another pose. At inference time, FUHA needs no external pose estimators, keypoints, or parametric body models. As a retrieval key, the motion embedding matches body pose better than the learned global embeddings of ten vision encoders. Its same-person retrieval rate is 0.1%, against 9.5% for its own DINOv3 backbone. On our upper-body benchmark, FUHA outperforms all baselines on every self-reenactment metric and on cross-reenactment identity similarity. It is competitive on face and full-body and runs at up to 38.9 FPS at 512x512 on a single RTX 4090.
[CV-102] FREE-LIVO: Physically Decoupled Yet Tightly Coupled LiDAR-Inertial-Visual Odometry
链接: https://arxiv.org/abs/2611.07463
作者: Reon Tabata,Kenji Koide,Shuji Oishi,Taku Okawara,Mizuho Aoki,Aoki Takanose,Masashi Yokozuka,Jun Miura
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:LiDAR-Inertial-Visual Odometry (LIVO) integrates complementary LiDAR, camera, and IMU measurements to achieve robust state estimation. However, most existing methods assume a fixed relative pose and accurate time synchronization between sensors. Toward modular and heterogeneous robotic systems, this paper presents FREE-LIVO, a tightly coupled odometry method that requires neither a rigid LiDAR-camera rig nor LiDAR-camera time synchronization. The proposed method uses a pretrained image feature matcher to establish correspondences between a camera image and a dense LiDAR intensity image generated from a single LiDAR scan using a generative model. From the resulting 3D point-to-pixel correspondences, we construct reprojection factors that directly constrain LiDAR and camera poses at different acquisition times. These factors are jointly optimized in a single factor graph together with constraints from point cloud registration, visual feature tracking, and the IMU measurements associated with each sensor. On six sequences from the public M3DGR dataset, FREE-LIVO achieves an average absolute trajectory error (ATE) RMSE of 0.45 m without provided LiDAR-camera extrinsic parameters and exhibits minimal accuracy degradation under time synchronization errors of up to 0.5 s. Furthermore, experiments with the LiDAR and camera mounted on separate robots demonstrate that the method can jointly estimate the trajectories of independently moving sensors.
[CV-103] VIPA: View-Invariant Projector-Based Adversarial Attack
链接: https://arxiv.org/abs/2611.07443
作者: Jiyu Han,Qingyue Deng,Bingyao Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 16 pages, 12 figures
Abstract:Projector-based adversarial attack aims to physically manipulate real-world scenes by projecting adversarial patterns, thereby causing deep image classifiers to produce incorrect predictions. However, existing stealthy projector-based adversarial attack methods model the project-and-capture process in a single and static view, limiting their view-invariant capability and practical applicability in dynamic environments. In this paper, we introduce View-Invariant Projector-Based Adversarial Attack (VIPA), a novel method designed to overcome these limitations by achieving both view-invariant and classifier-agnostic adversarial attacks in the physical world. We first leverage a view-invariant projector-camera system simulation method to model the physical interactions between projected patterns and real-world surfaces. To ensure robustness and stealthiness of the attack across different views, VIPA optimizes adversarial projections by aggregating simulated attack losses from multiple views. This joint optimization across diverse views helps maintain robustness regardless of the viewpoints. Finally, the optimized patterns are physically projected into real-world scenes, where they successfully fool various classifiers from different viewpoints, thereby enabling robust and practical view-invariant adversarial attacks. Our experiments in both targeted and untargeted attacks demonstrate that VIPA consistently achieves higher attack success rates than existing methods, while also enhancing stealthiness and ensuring minimal perceptual degradation across all tested views.
[CV-104] IMUDINO: Self-distilled Foundation Model for Inertial Sensing
链接: https://arxiv.org/abs/2611.07437
作者: Zhoutong Ye,Chengwen Zhang,Zixuan Wang,Xufeng Jian,Zhe He,Qingyang Wan,Xutong Wang,Jiaqi Liu,Mingze Sun,Yu Mei,Chun Yu,Yuanchun Shi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Official repo: this https URL
Abstract:We present IMUDINO, the first foundation model (FM) for inertial sensing to achieve competitive performance across diverse sensing tasks including human activity recognition (HAR), health, motion capture, and odometry. Prior inertial FMs often focus on one subfield of inertial sensing and sometimes underperform state-of-the-art supervised methods (e.g., in HAR), making their scope considerably narrower than FMs for vision and natural language. Moreover, they often require fixed input length and sampling rate, and generalize poorly to unseen sensor locations. To build a general-purpose foundation model for inertial sensing, we adapt the DINO self-supervised learning (SSL) paradigm to inertial data through a set of semantics-preserving motion transformations that simulate variations in motion pattern and sensor setup. For model architecture, we use a plain transformer encoder with frequency-aware positional encoding that natively supports varying input length and sampling rate. This simple setup proved remarkably effective for learning global and dense features generalizable across users, sensor setups, tasks, and, in the case of robot odometry, embodiments. We hope IMUDINO can serve as a backbone for inertial sensing and kickstart the development of general-purpose inertial FMs.
[CV-105] RiST: Topology-Reinforced Swin Transformer for Medical Image Analysis
链接: https://arxiv.org/abs/2611.07431
作者: Pengfei Gu,Huimin Li,Guangyu Meng,Hao Zheng,Haoteng Tang,Bin Fu,Danny Z. Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Medical images contain topological patterns, such as connected components and loops, at scales ranging from small lesions to organ boundaries. Although hierarchical Vision Transformers have advanced medical image analysis, they do not explicitly incorporate topology throughout their multiscale representations. We propose Topology-Reinforced Swin Transformer (TRiST), a framework that integrates topological information into feature construction, refinement, and attention. First, a Hierarchical Multi-Scale Topology Representation algorithm computes 2D persistent homology (H0 and H1) independently at four native Swin stages, producing descriptors aligned with each stage’s token grid rather than downsampling fine-scale topology. Second, a Band-Aware H0/H1 Residual Refinement module uses gated residual networks to process short-, medium-, and long-persistence information separately. Third, a Persistence-Band Progressive Attention Bias converts stage-specific descriptors into learned key-saliency scores that directly modify shifted-window attention logits. We instantiate TRiST as Topo-SwinV2-B for classification and Topo-Swin-UNet for segmentation, preserving the underlying Swin-based backbones. Experiments on five medical image datasets show consistent improvements over Swin-based baselines and competitive performance relative to existing topology-augmented methods.
[CV-106] MessyNav: Zero-Shot Navigation via Manipulation in Messy Environments
链接: https://arxiv.org/abs/2611.07417
作者: Kenneth Llontop,Cherie Ho,Dylan Zhou,Anuva Banwasi,Priya Sundaresan,Carlota Parés-Morlans,Jeannette Bohg
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: CoRL 2026
Abstract:Real-world environments are messy, with paths often blocked by everyday objects like chairs and toys, requiring robots to perform interactive navigation by manipulating obstacles to clear a path. Existing methods rely on extensive simulation-based learning or per-scene instructions specifying which objects to avoid or manipulate. They struggle to generalize across novel objects, diverse environments, and varied contexts such as a cart of fragile glassware that should be avoided while an empty one can be pushed. We formalize this problem as open-world interactive navigation and present MessyNav, a zero-shot framework based on vision-language models (VLMs). At its core, MessyNav performs explicit VLM-based affordance reasoning before task planning to decide whether to swipe, pick up, or avoid an object and produce a high-level plan. Evaluated on a mobile manipulator across real-world indoor and outdoor environments, MessyNav improves success rate by 27 percentage points (pp) over the strongest baseline in real-robot trials and achieves statistically significant gains in a larger-scale offline evaluation, without task-specific training.
[CV-107] Kinematic Hypothesis Context: Embodiment-Conditioned Learned Video Coding for Bandwidth-Limited Robot Policy Uplinks
链接: https://arxiv.org/abs/2611.07411
作者: Yuze Lu
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 12 figures. Preprint; supplementary material and code to be released upon formal publication
Abstract:Robots increasingly run large vision-language-action policies on remote servers, making the camera uplink the bottleneck. Existing video codecs spend bits on content the receiver could predict: the robot’s own motion. We introduce the kinematic hypothesis context, a zero-rate conditioning signal that both ends build from joint states, the robot model and the camera calibration. Ego-motion homographies, per-link robot reprojections, an end-effector plane and a background memory warp the previous reconstruction into hypotheses that condition a learned video codec. We prove that history-only and global-pose conditioning with a limited receptive field cannot access this information under large motion, and that coding gains become task gains at the policy’s success cliff. On LIBERO, the context saves 23.9% bit rate over a capacity-matched history-conditioned codec and 40-56% over x264, x265 and SVT-AV1, and needs 23.2% fewer bits than DCVC-RT at equal policy action deviation; trained to lower rates, it matches DCVC-RT in PSNR. In closed loop at equal rate, OpenVLA-OFT success nearly doubles from 22.8% to 43.1% at 6 kbps (72.6% with lower-rate training, 23.6 points above DCVC-RT), pi_0.5 success rises from 55.5% to 77.4%, and the gain grows under packet loss. The context transfers across three codec hosts and to real robot data.
[CV-108] PSGAgent : Enhancing MLLM s Dynamic Visual Perception for Panoptic Scene Graph Generation via Agent ic Reasoning
链接: https://arxiv.org/abs/2611.07408
作者: Qingyun Liu,Yuankun Liu,Jiaye Feng,Jie Huang,Zonghong Dai,Haochen Li,Jingkun Wang,Weiping Li,Tong Mo
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted by ACM Multimedia 2026. 10 pages, 8 figures
Abstract:Panoptic Scene Graph generation (PSG) requires holistic yet fine-grained reasoning over complex relations among objects. However, current Multimodal Large Language Models (MLLMs) applied to PSG suffer from a perception bottleneck: they function as static observers limited by fixed-resolution views. As a result, models often overlook critical relations involving small or peripheral objects and struggle to differentiate ambiguous predicates in complex scenes. To address these limitations, we propose PSGAgent, an agentic framework that enhances visual perception for PSG through a think-act-observe loop. PSGAgent allows MLLMs to dynamically refine their visual perception using two tools: a Region Zoom-In Tool for local exploration of small or boundary relations and a Visual Highlight Tool for relation-focused segmentation that filters out irrelevant contextual noise. To optimize decision-making, we construct a two-stage training pipeline. First, we create a high-quality, error-aware Multimodal Chain-of-Thought (M-CoT) dataset that provides robust supervision for cold-start learning. Then, we design a novel hierarchical reward mechanism integrated with Group Sequence Policy Optimization (GSPO) to enable strategic and adaptive tool usage. Extensive experiments on the PSG benchmark show that PSGAgent achieves SOTA performance, demonstrating strong capability in both holistic and fine-grained scene understanding.
[CV-109] FreePath: Neural Mean Free Path for Image Dehazing
链接: https://arxiv.org/abs/2611.07398
作者: Rongrong Gao,Jingxuan Feng,Wenyu Xu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Single image dehazing requires recovering scene radiance from visual responses that are weakened and veiled by participating media. Existing methods either invert explicit scattering variables, which can be unstable under complex scenes, or learn direct restoration mappings, which often discard the physical organization of haze formation. We present \emphFreePath, a neural mean free path framework for image dehazing. Instead of estimating atmospheric parameters, FreePath represents dehazing as feature space path state restoration, inspired by the expected distance a visual response can propagate before interacting with haze. Within a residual encoder and decoder, the Mean Free Path Residual Block learns path and collision fields to guide haze constrained correction, while FreePath State Conservation Integration preserves restored path states during deep feature aggregation. Experiments on paired CDD-11 haze images and real RTTS scenes show that FreePath consistently improves full reference fidelity, no reference perceptual quality, and visual faithfulness over representative dehazing baselines.
[CV-110] Vis2State: Visual Reasoning through Executable State Abstractions
链接: https://arxiv.org/abs/2611.07397
作者: Sheng Jin,Wentao Liu,Chen Qian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Solving visual puzzles requires both accurate interpretation of visual evidence and reliable computation under task-specific rules. Embedding visual assumptions and task constraints directly in solving code makes the represented problem difficult to inspect and revise independently. We introduce Vis2State, a training-free agentic framework that abstracts visual puzzles into explicit, revisable problem states linking visual inspection and execution. Visual inspection guides state construction and refinement, while generated programs solve the represented problem. The model constructs both the state and the program without predefined puzzle schemas or solving procedures. Inspection and execution feedback allow it to revise the problem representation and solving procedure independently. Vis2State achieves the state-of-the-art performance on four benchmarks. Cross-model evaluation further demonstrates consistent gains. Ablations, error analyses, and fixed-program state interventions further support the effectiveness of the state-centered design. The code will be made publicly available to facilitate future research.
[CV-111] WeLM-Image: A Unified Image Generation Model towards Visual Intelligence
链接: https://arxiv.org/abs/2611.07392
作者: Canmiao Fu,Chen Li,Guohui Zhang,Hao Lu,Hongyang Wei,Huadong Wu,Huaisong Zhang,Hubery Yin,Liyuan Cui,Qixin Yan,Tong Shao,Xiaoqiang Zhou,Xinhang Leng,Ying Zhang,Zhenhai Liu,Zhipeng Huang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:We present \textbfWeLM-Image, a unified image foundation model that integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference synthesis. Designed for single-GPU deployment in WeChat’s large-scale consumer applications, the model processes text, reference images, and target latents as one interleaved sequence. A Mixture-of-Transformers (MoT) backbone combines shared attention with independent understanding and generation parameters to limit optimization interference. Equipping each path with Mixture-of-Experts (MoE) modules provides sparse capacity, with 31.7B stored parameters and approximately 5.4B generation parameters activated per target latent token. A generation-oriented data engine combines a knowledge-grounded visual vocabulary with large-scale visual clustering to balance semantic coverage and editing tasks, while training-time evaluation guides iterative data refinement. Bilingual annotations and verified editing supervision support progressive training up to native 1.5K resolution. We optimize generation and editing separately with rubric- and checklist-based rewards, then consolidate the specialized policies through noise-aware on-policy distillation. In the reported overall comparisons, the 1536-resolution model ranks first among the evaluated open-source models on PRISM-Bench, Qwen-Image-Bench, and UnicBench. MeanFlow-based few-step distillation yields WeLM-Image-Turbo, which generates a 1024\times1024 image in 4.98,s on a single NVIDIA H20 GPU using eight guidance-free steps.
[CV-112] FFMT: Frequency-Factorized Multi-Rate Tokenization for Music-to-Dance Generation
链接: https://arxiv.org/abs/2611.07384
作者: Qirong Guo,Ruijie Li,Zeyan Li,Sicheng Shu,Guanzhi Ding,Weihang Ding
类目: ound (cs.SD); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Dance motion combines slowly varying body configurations with rapid local movements. We investigate how explicit temporal-frequency decomposition can guide the temporal resolution and codebook capacity of discrete motion representations. We introduce frequency-factorized motion tokenization (FFMT), which separates root-relative motion into a low-frequency choreographic scaffold and a complementary high-frequency residual. FFMT uses a lower token rate and a smaller codebook for the low-frequency component, and a higher token rate and a larger codebook for the high-frequency residual. This representation further induces a plan-then-articulate autoregressive factorization: low-frequency tokens establish coarse choreography, while high-frequency tokens realize local motion details conditioned on the generated plan and music. Experiments on AIST++ and FineDance show strong distributional fidelity while preserving realistic diversity and music–motion synchronization. Ablations consistently support frequency factorization, asymmetric temporal allocation, and low-to-high conditioning.
[CV-113] MIKU: Material Invariance and Knowledge Unification for PBR Generation SIGGRAPH
链接: https://arxiv.org/abs/2611.07354
作者: Yajie Bao,Jiaye Fu,Youtian Lin,Shuang Wu,Jiachen Qian,Feihu Zhang,Xun Cao,Yao Yao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: The Project Page is: this https URL . The Paper is published in SIGGRAPH Asia 2026 (TOG)
Abstract:High-quality material generation for physically based rendering (PBR) remains challenging because datasets and synthesis pipelines lack standardized representations. Differences in material definitions, rendering conventions, and parameterizations lead to unstable and physically implausible results. We propose MIKU, a unified framework for standardized PBR asset representation and physically consistent material generation across diverse 3D content domains. Its core is a unified processing pipeline that standardizes the workflow from data preprocessing to final texture synthesis, maintaining consistent geometry and texture representations. A normalization scheme harmonizes heterogeneous material parameterizations in a common representation space, enabling robust cross-dataset training and inference. Building on this formulation, we train a novel physically aware diffusion model for PBR texture generation within a two-stage multi-view synthesis framework. First, geometry-aware albedo generation synthesizes multi-view albedo maps under explicit geometric guidance to ensure cross-view appearance consistency. Second, metallic-roughness generation uses the diffusion model to predict physically plausible roughness and metallic maps conditioned on the generated albedo and unified representation. Trained end-to-end on standardized data, the model learns to generate materials that respect category-level material priors and semantic part boundaries. By decoupling representation standardization from appearance synthesis, MIKU substantially improves material generation stability and physical consistency. Extensive experiments demonstrate state-of-the-art texture quality and semantic consistency across object parts, advancing automated material generation for 3D content creation.
[CV-114] Where Do Reported Long-Video QA Gains Come From? A Controlled Audit of Frame Selection Evidence Injection and Expert Routing
链接: https://arxiv.org/abs/2611.07329
作者: Shih-Yao Lin
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Long-video question answering is usually studied as frame selection: which few frames should a vision-language model see? We introduce VideoRouteAudit, a controlled audit that separates this choice from two alternatives, injecting non-visual evidence and routing to another answering model. It holds the backbone, frame budget, decoding, prompt and grader fixed while varying one decision at a time. On the full Video-MME long bucket (n=900, K=8), no tested selector significantly improves over uniform sampling. The largest selection-only estimate is +1.9 pt (95% CI [-1.2,+5.0]) and our cascade is +1.2 pt [-1.7,+4.1]; all are equivalent to uniform within +/-5 pt. The same contrasts are unstable at the n=200 subset scale commonly used: three of four are up to 3.8x larger, one reverses sign, and resampled 200-item subsets reverse the full-bucket effect’s sign in 35.8% of draws. The audit also corrects our own first results. The anomaly-domain “+ASR” arms injected no text, so the pooled +4.7 pt gain was selection plus a development prefix. With transcripts, ASR lowers XD-Violence accuracy by 10.0 pt when appended after the question and not at all when placed before it, and lowers UCF-Crime by 13.0 pt either way; on Video-MME, placing it first adds about 4 pt (7 on Qwen2-VL) to uniform sampling and to the cascade alike (post hoc, unadjusted). Swapping the answering model gives the largest contrast (+15.1 pt on MUSIC-AVQA, +33.3 pt on its worst type), but the same audio-visual expert with its soundtrack silenced keeps 92% of the gain (69% with a mismatched soundtrack), so the contrast is mostly a model swap. Code, per-item predictions and scripts that recompute every statistic are available at this https URL.
[CV-115] ArtiVidRM: Verifiable Artifact-as-Rubric Reward Modeling for Explicit Quality Articulation in Video Generation
链接: https://arxiv.org/abs/2611.07301
作者: Caorui Li,Yangguang Ji,Shuxia Lin,Yu Chen,YiCheng Xiao,Yongliang Wu,Zhenxiang Jiang,Licheng Tang,Deng Li,Zhuxuanzi Wang,Jiahao Sun,Chonghan Liu,Lirian Su,Jay Wu,Xu Yang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Reinforcement learning has become an effective approach for improving video generation, but its success depends heavily on reliable reward models. Existing methods either map videos directly to opaque scalar scores or generate free-form rationales that are not explicitly optimized as the basis of the final judgment. We propose ArtiVidRM, a verifiable, artifact-as-rubrics video reward model that follows an observe-then-score paradigm. It first identifies visible generation artifacts and organizes them into a structured artifact list, then uses these artifacts as explicit rubrics for quality assessment. This intermediate representation grounds each score in concrete video-level evidence, making the reward more interpretable and informative for optimization. We further introduce a multi-reward reinforcement learning strategy to jointly optimize artifact prediction and quality assessment. The reward decomposes into complementary objectives that evaluate artifact prediction quality, including semantic coverage and prediction efficiency, together with a pairwise preference objective that supervises quality ranking through score differences. By jointly optimizing the intermediate artifact representation and the downstream preference prediction, the model learns to produce informative artifact lists while improving the accuracy of video quality assessment. Experiments show that ArtiVidRM achieves state-of-the-art performance on video preference benchmarks, provides effective signals for video diffusion optimization, and enables artifact-guided refinement of closed-source video generators.
[CV-116] Householder Motion Embedding: Noise Modulation for Motion-Controllable Video Generation NEURIPS2026
链接: https://arxiv.org/abs/2611.07294
作者: Tao Jun Lin,Jinhong Ni,Ruyi Zha,Yifu Wang,Xibin Song,Pan Ji,Hongdong Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026
Abstract:Diffusion models for video generation have achieved remarkable success, yet precise motion control under explicit user guidance remains challenging. Recent noise-warping methods introduce temporal correlations by warping the noise according to the user-provided motion signal. However, standard spatial warping generally disrupts the noise’s i.i.d. Gaussian structure, requiring distribution correction, such as Jacobian-based correction, but at a substantial computational overhead. We propose Householder Motion Embedding (HME), which encodes the motion signal into the noise through a sequence of Householder reflections that together form an orthogonal transformation applied to each location’s channel vector. HME exactly preserves the spatial Gaussian structure within each frame while introducing cross-frame correlations that promote temporally consistent video. HME is agnostic to the network architecture and integrates directly into standard video diffusion fine-tuning pipelines. Experiments on different scenarios show that HME matches or outperforms the state of the art, while achieving a 2.7x speedup over noise warping. These results demonstrate that channel-space Householder transformations offer a principled and efficient alternative to noise warping for motion-controllable video generation.
[CV-117] MammoSeek: A Mammography Agent with selective zooming for grounded reports
链接: https://arxiv.org/abs/2611.07188
作者: Haoyan Ding,Qian Wang,Lixia Wang,Debiao Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages, 2 figures
Abstract:Radiologists read a screening mammogram by first forming a global impression of the four views and then selectively magnifying suspicious regions before assigning a BI-RADS assessment. MLLMs, in contrast, see the whole image at a fixed, reduced resolution, which makes sub-millimeter calcifications invisible, and they perform close to chance on mammography. Therefore, we present MammoSeek, a mammography agent that reproduces this reading process. After forming an impression of each breast, the agent performs selective zooming: it decides by itself which detector candidates deserve a closer look and inspects only those on full-resolution views. Guided by a BI-RADS assessment skill, it then assigns a category and generates a localized, lexicon-structured report for each breast; its assessment is combined with image-model scores in a decision layer to decide recall. On an internal (EMBED) and an external (VinDr-Mammo) dataset, \ours reaches exam-level AUCs of 0.824 and 0.818, far above other MLLMs, mammography agent and CNN classifiers. Selective zooming also outperforms both no zooming and forced zooming in localization, structured-report F1 and recall decisions at low cost.
[CV-118] VINO: Revisiting 3D Representation Learning with Volume Transformers
链接: https://arxiv.org/abs/2611.07176
作者: Adrian Kruse,Kadir Yilmaz,Daan de Geus,Francis Engelmann,Bastian Leibe
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page at: this http URL
Abstract:Self-supervised learning has given rise to vision foundation models that learn general-purpose image representations transferable across downstream tasks, yet representation learning for 3D scenes has not achieved comparable success. Motivated by this, we identify key limitations of current 3D representation learning approaches and revisit their core design choices. We first replace the commonly used hierarchical PTv3 backbone with the Volume Transformer, a ViT-like architecture for 3D scenes. We then redesign two core objectives of modern approaches built around student-teacher networks: matching representations across views and predicting masked content from visible context. For the first objective, we separate the roles of teacher and student: the teacher provides clean, consistent targets while the student learns invariance to stronger perturbations. For the second objective, we prevent the student from accessing the masked geometry and instead require it to infer missing representations from visible context only. Finally, we redesign 2D-to-3D knowledge distillation, replacing direct regression of view-dependent image features with view-consistent supervision. Together, these changes form VINO, a framework that learns semantic representations of 3D scenes without labels. VINO substantially improves dense linear probing across indoor and outdoor benchmarks and transfers strongly to out-of-domain settings.
[CV-119] Perceptual Distortion for Overfitted Video Coding: Sliced Matching in Space Error Alignment in Time ISCAS
链接: https://arxiv.org/abs/2611.07148
作者: Tianyi Yang,Ge Gao,Fan Zhang,David Bull
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages, 4 figures; submitted to IEEE ISCAS
Abstract:Recent works in neural video coding have increasingly been trading off perceptual quality against reconstruction fidelity at a fixed bitrate. Wasserstein Distortion (WD) offers a principled approach to balancing these objectives by relaxing spatial correspondences in feature matching, enabling lightweight overfitted image codecs to compete with computationally intensive generative decoders. Extending vanilla WD to video, however, presents two challenges: its independent Gaussian approximation limits the expressiveness of feature distribution matching, while unconstrained texture resampling across frames introduces temporal inconsistencies. We address these challenges by distinguishing between the correspondences that can be relaxed and those that must be preserved. Spatially, we introduce a local sliced Wasserstein distortion (SWD) that matches empirical feature distributions through one-dimensional optimal transport, capturing cross-channel dependencies without enforcing exact spatial alignment. Temporally, we employ a novel flow residual that constrains coding errors along motion trajectories, preserving temporal consistency without imposing additional reconstruction fidelity constraints. Evaluated on UVG and HEVC Class B with the overfitted codec COOL-CHIC, our framework achieves BD-rate reductions of 9.2% in DISTS and 3.0% in LPIPS relative to frame-wise WD, averaged over test sequences, while lowering tOF by 10.2% at matched rate.
[CV-120] A Proxy-Based Method for Measuring Racial Disparities in~Text-to-Image~Generators
链接: https://arxiv.org/abs/2611.07132
作者: Nia Mucher,Ngan Pham,Jason Chin,Aaron Brown,David Evans
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:We present an automated method for auditing text-to-image generators for representational bias against ground-truth demographic benchmarks. Our approach adapts the racial proxy variables used in U.S. antidiscrimination law to provide a fine-grained comparative analysis of variations in racial representation in generated images. We classify each generated image by race and then compare the model’s distribution to publicly available demographic benchmark data to identify patterns of over- and under-representation. We also measure skin tone using CIELAB values and calculate cosine similarity amongst groups of images in order to capture biases that don’t show up in demographic comparisons such as differences in image quality across races and skin tones. We report on experiments testing three models (Imagen 4 Fast, Stable Diffusion 3.5 Large Turbo, and Flux Schnell) using three categories of proxies (geographic, occupational, and surname) and 50 variants within each category. We find notable variation across the models as well as by proxy category. For example, Imagen 4 Fast overrepresents white classified images more than the other models tested and produces images with lighter, more homogeneous skin tones. Imagen 4 Fast also had the largest representational disparity for Black demographics among surname and geographical prompts. Stable Diffusion 3.5 Large Turbo produces results that overall diverge most from census data but yield the most diverse range of skin tones. Our method yields a generalizable and legally-grounded auditing tool that surfaces detailed information about model outputs without relying on implicit normative definitions of fairness.
[CV-121] Wide View Synthesis from Overlapping Perspective Cameras in Autonomous Driving
链接: https://arxiv.org/abs/2611.07081
作者: Kaixuan Zhang,Zhipeng Xiong,Minxian Li,Mingwu Ren,Xiatian Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multi-perspective-camera systems are widely adopted since they provide broad visual coverage for autonomous driving, motivating the synthesis of a unified wide view from spatially separated observations. In this work, we formulate wide view synthesis (WVS), which aims to synthesize a geometrically consistent wide view at a specified virtual viewpoint from multiple perspective cameras. Unlike conventional image stitching, WVS requires reasoning about 3D geometry and visibility to handle depth-dependent parallax and occlusions arising from spatially separated camera centers. While vanilla 3D Gaussian Splatting (3DGS) enables high-fidelity view synthesis, its reliance on per-scene optimization limits its applicability to driving scenes. Furthermore, WVS introduces additional challenges, e.g., peripheral stretching under perspective projection and photometric inconsistencies across cameras. To address these challenges, we propose Wide-View Gaussian Splatting (WV-GS), a feed-forward Gaussian-splatting framework that synthesizes wide views without per-scene optimization. Specifically, we introduce Piecewise Perspective Rendering (PPR) to mitigate peripheral stretching through local perspective projections sharing a common camera center, and Spatial Affine Photometric Modeling (SAPM) to account for camera-dependent photometric variations during training. We further introduce WideDrive, a synthetic benchmark providing dense ground-truth wide views and adapt a LiDAR-projected sparse-reference protocol to real-world datasets. Extensive experiments on multiple datasets demonstrate the effectiveness of WV-GS for WVS. Dataset and code will be publicly released.
[CV-122] SD-LoRA: Training-Time Structural Distillation into LoRA for Few-Shot Vision-Language Adaptation NEURIPS2026
链接: https://arxiv.org/abs/2611.07063
作者: Mohammed Rahman Sherif Khan Mohammad,Ardhendu Behera,Sandip Pradhan,Swagat Kumar,Yonghuai Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted in NeurIPS 2026
Abstract:Few-shot adaptation of vision-language models must balance two competing goals: learning fine-grained visual distinctions from very limited supervision and preserving the efficiency that makes frozen CLIP-like models attractive for deployment. Existing lightweight adaptation methods, including prompt tuning, LoRA, and cache-based adapters, typically supervise global image embeddings and therefore underuse token-level structure. Conversely, structure-aware methods exploit local evidence but often introduce additional inference-time computation. We propose SD-LoRA, a structural distillation framework that uses token-level graph reasoning only during training and transfers its effects into a lightweight LoRA-adapted CLIP model. During training, a heterogeneous graph teacher models patch-patch, global-local, and visual-text relations over high-resolution ViT tokens. Rather than retaining this graph at test time, SD-LoRA transfers its structural bias into the shared CLIP-LoRA backbone through implicit distillation and Prototype Predictive Alignment, which aligns graph-conditioned representations with support-derived class prototypes. At inference, the graph teacher is discarded; prediction requires only a single CLIP-LoRA forward pass and a cache-based classifier. Across 11 few-shot benchmarks, SD-LoRA improves over a strong CLIP-LoRA baseline by 2.4/1.8/3.0/2.4/2.0 percentage points in the 1/2/4/8/16-shot settings, respectively, without additional graph computation at test time. Results on base-to-novel generalization, cross-dataset transfer, out-of-distribution robustness, and calibration further show that training-time structural supervision improves both accuracy and transfer reliability. Code is available at this https URL.
[CV-123] Segment What It Can Draw: Querying a Text-to-Image Generator with Its Own Images for Skin Lesion Segmentation
链接: https://arxiv.org/abs/2611.07061
作者: Hongyu Cao,Guangfu Guo,Jinghan Zhang,Dakshak Keerthi Chandra,Ying Zhang,Kunpeng Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Skin lesion segmentation without annotated masks requires the target to be specified through other cues. Existing approaches rely on dermoscopy-specific rules, annotation-derived model priors, or user-selected regions. A general text-to-image diffusion transformer offers a text interface to the target, but its concept-word attention is unreliable on dermoscopy images: in the model’s attention-output space, the word-derived skin-to-lesion direction is poorly aligned with the direction separating real skin from real lesions. We show that the direction derived from generated healthy-skin and lesion images is substantially better aligned. Based on this finding, we propose GENIQ, a training-free method that constructs initial skin and lesion prototypes from generated images and refines them using the most skin-like and lesion-like patches from unlabelled dermoscopy images. It scores test patches along the refined direction and applies a gated threshold to suppress contextual spill-over around small lesions. Without inference-time annotations or additional models, GENIQ achieves Dice scores of 0.85-0.88 on three dermoscopy test sets. It outperforms zero-shot and word-based attention methods on all three sets, and a hand-designed dermoscopy pipeline in Dice and IoU on two.
[CV-124] Expanding the Quality-Coverag e Frontier with Competitive Flow Matching
链接: https://arxiv.org/abs/2611.07055
作者: Wenxuan Peng,Richard Zhang,Haitian Zheng,Jaskirat Singh,Tsai-Shien Chen,Eli Shechtman,Bharath Hariharan,Hadar Averbuch-Elor,Zongze Wu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 40 pages, 24 figures. Project page: this https URL
Abstract:Flow matching (FM) generators often face a quality-coverage tradeoff. Inference-time techniques such as classifier-free guidance raise quality, but at the cost of coverage. We hypothesize this limitation stems from a basic property of the objective: an intermediate noisy image could correspond to many clean images, yet the model can only predict a single velocity, at best averaging over the multiple possibilities. To mitigate this, we introduce competition among predictions during training. We achieve this by augmenting the predictor with a randomly drawn low-dimensional conditioning code z . For each training example, we sample K values of z to produce K candidate velocities that compete to explain the target, and apply the loss only to the best-matching candidate. This winner-take-all objective allows different modes to shape different predictions, rather than being averaged into a single compromise. Furthermore, this term naturally enables a new form of guidance, auxiliary conditioning scaling, which complements traditional guidance methods. Together, these two guidance mechanisms show that our method substantially improves the quality-coverage frontier. We validate on ImageNet-256, using two different testbeds, SiT and RAEv2. Given less than 3% of the original SiT pretraining compute for post-training, at the 67% Recall point, our method reduces \mathrmFD_r,6 from 19.22 to 5.91. On RAEv2, at the 70% Recall point, it reduces \mathrmFD_r,6 from 16.83 to 3.18.
[CV-125] Gaussians Can Subtract: Neural Dissolve Fields for Hybrid Scene Rendering WACV2027
链接: https://arxiv.org/abs/2611.07049
作者: Yiyan Wang,Haohan Liu,Tetsuo Shibuya
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV 2027. 16 pages including supplementary material, 11 figures, 8 tables. Project page: this https URL
Abstract:Hybrid Mesh-Gaussian rendering combines compact textured surfaces with flexible 3D Gaussian primitives, offering an efficient alternative to purely Gaussian scene representations for structured environments. Existing scene-reconstruction hybrids commonly follow an additive formulation, where Gaussian residuals compensate for errors left by the mesh. However, not every Mesh-Gaussian conflict is caused by missing content: residual geometry errors, occlusion mismatch, texture-projection artifacts, and view-dependent appearance can make the same mesh content consistent in one view but inconsistent in another. We therefore formulate hybrid rendering as a multi-view inconsistency problem and present Gaussolve, a subtract-and-refill framework that selectively dissolves inconsistent mesh content before residual Gaussians refill the revealed regions. We introduce a lightweight Neural Dissolve Field, parameterized by scalar spherical-harmonic Gaussians, to learn where and from which viewing directions mesh content should be dissolved. A continuous transmittance fusion rule progressively exposes residual Gaussians behind the dissolved mesh, preserving their reconstruction gradients and avoiding hard visibility discontinuities. Experiments on structured and semi-structured scenes demonstrate higher visual fidelity with fewer Gaussian primitives and faster rendering. Project page: this https URL
[CV-126] raCE: Trajectory-Aware Concepts for Explainable Video Understanding NEURIPS2026
链接: https://arxiv.org/abs/2611.07009
作者: Wooil Lee,Junpyo Hong,Jongseo Lee,Tuomas Oikarinen,Tsui-Wei Weng,Jinwoo Choi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to NeurIPS 2026. Project page: this https URL
Abstract:Video understanding models often rely on motion, yet their explanations can obscure which motion sources support a prediction. We identify motion-source under-specification as a limitation of existing video XAI: region-centric explanations can be source-ambiguous, while motion vocabularies restricted to single-person pose sequences are source-incomplete. We propose TraCE (Trajectory-Aware Concepts for Explainable Video Understanding), a concept-based framework that discovers source-flexible motion concepts by clustering optical-flow trajectories from training videos. These concepts represent dynamic evidence from people, objects, and interactions, while object and scene concepts account for static context. The resulting typed trajectory/object/scene interface supports class prediction for action recognition and question-conditioned concept matching for video question answering (VQA). We evaluate TraCE across five random seeds on six benchmarks: NTU RGB+D Mutual, Chiral SSv2, UCF-101, KTH, TOMATO, and STAR. Relative to non-interpretable baselines, TraCE maintains competitive task performance, with an average accuracy drop of 1.0 points for action recognition and a nominal gain of 0.4 points for VQA. Motion perturbation and concept-removal tests provide complementary evidence of directional semantic alignment and decision faithfulness, while sample-level interventions illustrate targeted model debugging. Our project page is available at this https URL.
[CV-127] Batch-Conditioned Semantic Anchors for Robust Transductive Adaptation of Vision–Language Models NEURIPS2026
链接: https://arxiv.org/abs/2611.06993
作者: Mohammed Rahman Sherif Khan Mohammad,Ardhendu Behera,Sandip Pradhan,Swagat Kumar,Amr Ahmed
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026
Abstract:Transductive adaptation improves vision-language models at test time by refining predictions on an unlabeled batch, but current methods face a key trade-off. Aggressive approaches exploit batch structure but can drift when batches are sparse, have prior shift, or cover little of the label space. Conservative methods better preserve CLIP text semantics, but fixed semantic references can under-adapt even when the batch provides strong corrective evidence. We propose JPTA (Joint Prior and Transport Adaptation), a prior-aware framework that casts transductive VLM adaptation as batch-conditioned semantic reference estimation. Instead of treating CLIP text prototypes as fixed priors or freely replaceable, JPTA views them as semantic references that shift toward image-side batch structure only under calibrated evidence. It estimates batch priors and soft image-side prototypes to form transported semantic references whose impact is scaled by batch reliability. Across standard benchmarks, low effective-class regimes, all-class evaluation, and online streams, JPTA consistently outperforms TransCLIP and StatA, with the largest gains on sparse and prior-shifted batches. Anchor-drift, active-support, and absent-class-mass diagnostics show that JPTA fixes effective-class mismatch without unconstrained transductive drift, indicating that robust transductive VLM adaptation should recalibrate semantic references using reliable batch evidence rather than choose between unrestricted transduction and fixed prototypes. Code is available at this https URL.
[CV-128] 2-Splat: Adaptive Topology Mesh Splatting with Texture Residuals NEURIPS2026
链接: https://arxiv.org/abs/2611.06974
作者: Zhihao Tang,Youjia Zhang,Mingbo Zhao,Wei Yang
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: Accepted at NeurIPS 2026
Abstract:Triangle-based splatting connects novel view synthesis with conventional graphics pipelines, but often requires excessive tessellation to preserve appearance details. We present T ^2 -Splat, which decouples topology adaptation from appearance refinement. Safe Restricted Delaunay Triangulation (Safe RDT) initializes mesh connectivity. Dynamic Topology Evolution (DTE) then adapts triangle allocation through edge collapse and vertex split operations. Residual per-face Mip-textures then refine appearance on the fixed mesh using scale-adaptive filtering. Experiments on Mip-NeRF 360, Tanks Temples, and DTU demonstrate competitive results with compact meshes. T ^2 -Splat improves PSNR by 0.71dB over MeshSplatting on Mip-NeRF 360. Additional compression experiments demonstrate flexible control of mesh complexity, supporting face budgets as low as 10% of the original face count. Project page: this https URL.
[CV-129] Learning Noise Optimization for Reward Alignment in Few-Step Generative Models
链接: https://arxiv.org/abs/2611.06255
作者: Hongeun Kim,Jeongsol Kim,Jong Chul Ye
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Reward alignment for diffusion- and flow-based generative models is often formulated over stochastic trajectories, making it less suitable for few-step deterministic generators. Initial noise optimization provides a trajectory-free alternative, but existing gradient-free methods require costly per-sample search at inference. We introduce LeNO (Learning Noise Optimization), which amortizes reward-guided initial noise optimization into a learned residual transport while keeping the pretrained generator frozen. LeNO learns an anchor-conditioned distribution over noise corrections using Local GRPO and scalar rewards, eliminating iterative test-time optimization. For Gaussian initial noise with norm-preserving projection, we derive a tractable upper bound on the KL divergence between the corrected noise distribution and the original prior, providing a principled regularization objective. We evaluate LeNO on few-step text-to-image and protein structure generation, including expensive black-box designability rewards. Across both domains, LeNO improves target rewards with a single noise correction at inference while largely preserving generation quality and diversity.
[CV-130] A Theoretical Analysis of a Bilevel Optimization Formulation for Combined Local-Global Optical Flow Models
链接: https://arxiv.org/abs/2611.08359
作者: El Mostafa Kalmoun
类目: Optimization and Control (math.OC); Computer Vision and Pattern Recognition (cs.CV)
备注: 31 pages, 5 figures. Submitted to the Journal of Mathematical Imaging and Vision. Python code in the ancillary files
Abstract:This paper analyzes a bilevel formulation for learning the regularization parameter of combined local-global (CLG) optical flow models. The lower-level problem computes the flow from a Lucas–Kanade data term and a quadratic or total variation (TV) regularizer, and the upper-level problem selects the parameter that minimizes the squared distance to a ground-truth flow. The setting is finite-dimensional, and a full-rank assumption makes the lower-level solution unique. For the quadratic model we derive sufficient conditions for the existence of an optimal parameter, for every optimal parameter to be positive, and for strict local optimality. For TV regularization an optimal parameter always exists, because the lower-level solution vanishes for large parameters, and we give a sufficient condition for positivity that accounts for the coupling of the two flow components through the local structure tensors. Synthetic examples illustrate the conditions, including cases in which they fail.
[CV-131] An Evolutionary Hypothesis for Whites Illusion: Constant Percentage Correction and Modern Individual Variation
链接: https://arxiv.org/abs/2611.06935
作者: Bekir Tevfik Akgün,Ergun Akleman
类目: Neurons and Cognition (q-bio.NC); Computer Vision and Pattern Recognition (cs.CV)
备注: 12 pages and 7 figures. We may later send a shorter version to a conference
Abstract:White’s illusion is a striking demonstration of context-dependent brightness and lightness perception. In earlier work, we developed a forward translucency-based formula capable of producing approximately constant perceived appearance across varying backgrounds. While validating that formulation, we observed a notable population-level pattern: individuals selected slightly different correction percentages for “constant perception,” while the population average remained remarkably stable across experiments. In this paper, we propose an evolutionary hypothesis for this phenomenon. We suggest that a simple constant-percentage correction function could have emerged as an adaptive mechanism for maintaining useful material and reflectance judgments under variable illumination. Under this interpretation, the stability of the population mean may reflect a conserved perceptual mechanism, whereas individual differences may reflect learning, adaptation, or variation around a shared computational response. Using a modification of our previously published forward model, we show that White’s illusion can be reproduced by applying a constant-percentage correction to background luminance. The model therefore provides a compact computational account of the observed population-level regularity and motivates a testable evolutionary interpretation of individual variation in contextual lightness judgments.
人工智能
[AI-0] SwarmEvolve: Latent Agent Swarms for Evolutionary Discovery and Optimization
链接: https://arxiv.org/abs/2611.08483
作者: Qile Jiang,George Em Karniadakis
类目: Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
备注:
Abstract:LLM-guided evolutionary search uses AI agents to improve programs for mathematical and computational problems by iteratively refining solutions and combining high-scoring methods. However, the process is often slow because the LLM generates a long textual chain of thought for each round and must repeat this generation many times to maintain population diversity. We present SwarmEvolve, a framework in which hundreds of low-cost agents per round work in swarms, reason in latent space without decoding to text, and share information by passing KV caches. Each swarm studies the problem from a different angle. A novelty filter keeps the most distinct agent ideas by directly measuring the distances between the agents’ hidden state vectors. Crossover of ideas happens naturally in latent space, as a summarizer agent receives the combined KV caches of all swarm members. On 21 problems in mathematical discovery and scientific computing, SwarmEvolve gets the highest average score and needs a median of only 3 to 8 minutes to reach the best scores that other state-of-the-art frameworks reach in 60 minutes. On several problems, SwarmEvolve with Qwen3 8B improves on the solutions that AlphaEvolve first found with frontier LLMs. Our work shows that agent swarms that communicate in latent space offer a fast, low-cost approach to program search for scientific problems.
[AI-1] XStack: A Benchmark for ML Repo Migration
链接: https://arxiv.org/abs/2611.08472
作者: Linrong Cai,Taiming Lu,Haozhe Chen,Zhuang Liu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:ML code should run where researchers choose, not where it was first written. Today, machine learning code mainly lives in two ecosystems, PyTorch on NVIDIA GPUs and JAX on Google TPUs. Migrating repositories between them means building and testing across frameworks and hardware. This migration is (1) a practical need, since it unlocks different hardware for researchers, and (2) a natural benchmark: it involves different frameworks and hardware, covers whole repositories, and is long-horizon and challenging. We introduce XStack, a benchmark that evaluates agents on migrating ML repositories between these two ecosystems. Each agent receives a PyTorch repository and a clean TPU machine and must deliver a JAX repository. Its output passes if it meets our grading criteria, based on the source repository’s published result, authenticated and calibrated by us. Migration by agents remains unreliable: Claude Opus 5 passes 75%, GPT-5.6-sol 45%, and every other agent at most 40%. XStack is a step toward freeing ML code from its ecosystem. We release XStack to make progress toward reliable migration across frameworks and hardware measurable.
[AI-2] Same Behavior Different Mechanism: What Finite Experiments Can Certify About Integrated Information
链接: https://arxiv.org/abs/2611.08469
作者: Linqian Wu,Diaoxuan He
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC)
备注: 30 pages, 7 figures. Code available as ancillary files
Abstract:Interpretability, AI auditing and assessments of machine consciousness increasingly rest on mechanism-level quantities: functionals of a system’s causal transition law that describe how it computes, not what it outputs. Experiments see that law only through finite, partial and noisy interventions. We ask when such a quantity is identifiable from an experiment’s access, continuous in the mechanism and certifiable from finite data, and answer exactly for a deliberately demanding case - the integrated information \varphi of integrated information theory - in stochastic networks of four and five units with known ground truth. Behavior can be silent: a recurrent and a feed-forward network with identical input-output laws at their response delay carry 0.28 and 0 bits, yet internal recordings separate them within 4-53 trials. Partial access can be just as blind: in our examples, units left out of a one-step recording impose precision floors of 0.30-1.02 bits that no sample size lowers, and in all-but-self parity networks, recordings of any length see several hidden units only through the product of their fidelities - a symmetry that one intervention broke in exact checks. The measure has sharp edges: because \varphi selects a partition, and the cause and effect it specifies, by one criterion and reports another, infinitesimal changes of mechanism that cross an edge change it by 25-100% at 99% of the edges we found in random networks, and about 1% of surrogate models that fit interventional data almost perfectly misreport it by more than 20%. One-sided claims survive: the same data certify a lower bound on both sides of an edge, and we prove \varphi1.004 bits throughout a 64-dimensional region of mechanisms and certify it from about 10^7 controlled transitions with false-certificate probability at most \alpha . Measurability is thus a design question with computable answers.
[AI-3] owards interpreting any residual stream direction by scaling activation inversion
链接: https://arxiv.org/abs/2611.08465
作者: Celeste De Schamphelaere,Juan J. Vazquez,Ari Spiesberger,Tomáš Gavenčiak,Gavin Leech,Euan Ong,Jiahai Feng
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Euan Ong and Jiahai Feng advised equally
Abstract:To interpret a direction in a language model’s residual stream, we can look for text that strongly activates it. But when such text is found by scanning a corpus, it is limited to what that corpus contains. We develop Maximally-Activating Example Metamodels (MAEMs), language models which map residual stream directions from a subject model to text exemplars that strongly activate those directions. Our objective is to maximise the cosine similarity of a supplied residual stream direction with the activations of the text generated by the MAEM. This label-free RL objective helps to scale MAEMs on diverse sources of residual stream directions. Empirically, we find the resultant metamodel produces informative exemplars: a Qwen3.6-27B MAEM inverts held-out activations better than corpus search or a natural-language autoencoder (NLA), beats corpus search (at 10M tokens) on a majority of held-out languages/domains, and activates 89% of held-out SAE features. It also identifies concepts behind steering vectors about as well as NLAs, recovers the payloads of LoRA backdoors more often than corpus search, and can even verbalise some model workspace inner monologue. Unlike explanations produced by other metamodels (e.g. NLAs), MAEM exemplars can be verified with a forward pass on the subject model, making the MAEM objective a grounded alternative for scaling the training of metamodels.
[AI-4] ReproBridge: Building Reproducibility into ML Research Artifacts
链接: https://arxiv.org/abs/2611.08452
作者: Yubo Wang,Rui Meng,Jaehyun Nam,Tomas Pfister,Jinsung Yoon
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Reproducing machine learning experiments requires reconstructing procedures from information scattered across papers, released code, and supporting artifacts. For coding agents, reproduction success can therefore depend on their ability to fill in missing details, rather than on the released artifacts alone. We ask whether preserving recovered procedures and decisions improves subsequent independent reproduction. We define a standardized artifact design comprising a refined codebase, environment setup, experiment launchers, and a reproduction guide, and introduce ReproBridge, an agent workflow for constructing these refined reproducible artifacts. ReproBridge-Bench provides 1,029 tasks from 387 experiments across 36 papers, anchored by five human-verified reference scores per task. Agents attempt tasks in fresh sessions with raw code or refined artifacts; a separate Auditor checks execution evidence. Across four evaluated models, refined artifacts raise mean reproduction success from 76.4% to 92.6% and reduce downstream API spending by 19.7% on average. The cross-model success-rate spread narrows from 21.6 to 3.9 percentage points, with all four models achieving mean success rates above 90%. Removing any component reduces success in the ablation study, with the largest losses from code and environment removal. These results suggest that preserving execution knowledge helps even models with lower raw-code success reproduce experiments more reliably and at lower API cost.
[AI-5] he Power of Meta Access in the Software of the Future — Demonstrated on LLM Chat UIs and Chat AI
链接: https://arxiv.org/abs/2611.08428
作者: Julian Kunkel,Ali Doosthosseini
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Chat remains the default surface for many large language model (LLM) tasks, even as agentic tools such as OpenCode excel at coding and automation. Yet the model sees the current dialogue but not the workspace around it. We explore meta access—giving the assistant a structured, permissioned view of the interface itself and propose a ladder of levels: feature tutoring; metadata, settings, and preference access; retrieval over summaries and messages, confirmed mutation, and full orchestration. Meta access can turn a chat client from many screens into one entry point for finding, navigating, organising, and operating it, while raising risks around grounding, privacy, prompt injection, and trust. General-purpose agentic harnesses are runtimes; because they can change their own settings, meta access becomes reflexive and needs a separate control-plane boundary. We ground this in the Meta Chat prototype of the Chat AI client. It builds bounded context from, e.g., summaries, settings, and relevant messages; links references to conversations, topics, and personas; supports confirmed, reversible changes; offers a browsable tutorial library; and recommends personas as user-started conversations. By exposing application state in structured form, meta access can extend screen readers from widget descriptions to semantic workspace operations. Meta access thus turns chat from a window onto one conversation into a control layer for the whole workspace, extending chat toward agentic capability without surrendering user control. As assistants become embedded in IDEs, dashboards, office suites, and operating systems, AI-era interfaces should expose a structured, permissioned model of their state and capabilities, ideally by opening their existing metamodels through a debugger-like interface.
[AI-6] Failure Prediction and Classification in Vision-Language-Action Models Via Calibrated Latent Probing
链接: https://arxiv.org/abs/2611.08425
作者: Michal P. Podolinsky,Yunhao Yang,Daniel Little,Hong Lu,Rohan Siva,Olin Goudey,Neel P. Bhatt,Matthias Scheutz,Ufuk Topcu
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: Code, videos, and dataset available at this https URL
Abstract:Vision-Language-Action models (VLAs) enable general-purpose embodied intelligence, but their performance is brittle under distribution shift, where unanticipated failures can damage equipment or endanger people. Deploying VLAs in real-world systems therefore requires a runtime monitor that is timely, calibrated, and informative while being efficient enough to run on the edge. Existing methods fail to fulfill these requirements. We present VLA-Tug, a lightweight probe over the latent layers of a VLA backbone that improves early failure detection and bounds the missed-detection and false-alarm rates, with minimal inference overhead. We show that supervising on the pre-failure segment of a trajectory increases early failure detection at matched false-positive rates by up to 30% over the strongest learned baseline. Under shift to unseen tasks, the bounds on missed-detection and false-alarm rates remain valid in up to 86% of runs, versus at most 21% for baselines. Unlike existing methods, VLA-Tug predicts the failure mode to enable corrective actions, raising top-3 accuracy on unseen tasks by up to 44% over a majority baseline. VLA-Tug enables real-time monitoring where the well-timed and informative probe supports lightweight steering that improves task success. We release a dataset of 3,180 annotated manipulation episodes to support future work on latent failure detection in VLAs at this https URL.
[AI-7] Benchmarks as Software: Terminal-Bench and the Case for Continuous Benchmarks
链接: https://arxiv.org/abs/2611.08417
作者: E. Kelly Buchanan,Crystal Zhou,Scott W. Linderman,Christopher Ré
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Coding agents are among the most economically consequential deployments of LLMs to date, making reliable evaluation increasingly important. Yet even benchmarks that are carefully validated at release can become stale. As models improve, they can discover valid solutions that existing verifiers reject, while changes to evaluation infrastructure can invalidate assumptions embedded in tasks and tests. We call the resulting accumulation of defects benchmark maintenance debt: a growing gap between benchmark scores and the capabilities they are intended to measure. We introduce a continuous-validation workflow for agentic benchmarks that combines community reports, task artifacts, agent traces, automated oracle checks, and human-gated repairs. We apply this workflow to Terminal-Bench 2.0, a benchmark that underwent extensive human and LM-assisted review and red-teaming before release. Our audit identifies issues in roughly 30% of the benchmark, and we release the resulting repairs as Terminal-Bench 2.1. These repairs change evaluation results: accuracy increased by over 10 percentage points and leaderboard positions move by up to three ranks, even as the overall ranking remains highly correlated (Spearman rho = 0.958). We evaluate how resource allocation shapes benchmark conclusions. We show that fixed wall-clock can materially change agent performance: For example, extending the time limit closes a 20pp harness gap on GPT 5.4 and shrinks a 30 pp gap on GPT 5.4 mini to 5 pp. Benchmark maintenance is therefore not only about repairing task artifacts: resource budgets are part of the measurement specification and must be validated alongside the tasks themselves. More broadly, benchmark maintenance debt is not just accumulated implementation bugs: our results show that as models improve, both evaluation assumptions and resource settings may need to be revalidated. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.08417 [cs.AI] (or arXiv:2611.08417v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.08417 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-8] he Gold Rush in AI4Math II: What are signatures of AI-assisted Proofs?
链接: https://arxiv.org/abs/2611.08412
作者: Yuanchuan Guo,Jiashun Jin,Zheng Tracy Ke,Bingcheng Sui
类目: Artificial Intelligence (cs.AI); Applications (stat.AP)
备注: 31 pages, 15 figures
Abstract:The growing use of AI in mathematics raises a question: can we detect AI involvement in proofs, separately from AI assistance with writing? Historical arXiv papers, recent papers disclosing substantive AI use, and manuscripts released by OpenAI provide a rich empirical basis for studying this question. However, several challenges remain: lack of matched case-control pairs; potential incompleteness of disclosure; and lack of ready-to-use features designed to capture signals of AI involvement in proofs. To address these challenges, we construct five cohorts comprising 8,281 papers by combining existing manuscripts with those we produced using generative AI. These cohorts span different levels of AI involvement and include matched pairs of papers addressing the same research problems. We then introduce ProofTrace, a framework that represents each paper as a graph of immediate proof support (GIPS) and extracts interpretable features of mathematical organization. Our study yields three discoveries. First, principal-component projections of our features reveal a qualitative gradient of AI involvement, where AI-reworded papers remain close to human papers, the two AI-generated cohorts depart in broadly shared directions, and papers disclosing substantive AI use occupy intermediate positions which shifts toward AI-generated cohorts over time. Second, we identify recurring signatures of AI-generated proofs, such as fewer externally cited results, greater reliance on intermediate conclusions, and altered graph connectivity; many of these signatures persist under controls for research field, manuscript length, or graph size. Finally, in a separate detection experiment, ProofTrace distinguishes human-written papers from those disclosing substantive AI use. Its performance remains stable under AI rewording, whereas the performance of state-of-the-art text-based detectors declines sharply. Comments: 31 pages, 15 figures Subjects: Artificial Intelligence (cs.AI); Applications (stat.AP) Cite as: arXiv:2611.08412 [cs.AI] (or arXiv:2611.08412v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.08412 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-9] Learning a patient world model for health system simulation in intensive care
链接: https://arxiv.org/abs/2611.08392
作者: Yingtai Li,Ruoyan Wang,Yu Li,Shaohua Kevin Zhou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Designing healthcare strategies requires anticipating their clinical and operational consequences before implementation, a task that often relies on health system simulation. Such a simulation necessitates modeling how patient states evolve and generate care demands. However, conventional approaches typically tailor compact state transitions to a single policy question and omit history-dependent dynamics, thus requiring the redefinition of states and transitions when there is a new question that entails different clinical information. We propose to learn a patient world model from longitudinal EHRs using a neural network to provide a shared clinical representation of patient trajectories that can serve simulation questions across scenarios. Taking ICU operations as an example, we train the model on MIMIC-IV, freeze it, and couple its time-queryable predictions to shared ICU calendars, resources, and queues via scenario-specific mappings, so that one model renders admission demand, care tasks, and discharge proposals across simulations evaluating ICU bed capacity, care dispatch, and post-ICU transfer capacity. Experiments show that our model matches Markov and semi-Markov comparators in aggregate occupancy and directional responses to capacity and dispatch policies. It also reproduces policy gains from replayed observed trajectories. For patients entering the same support state from different prior trajectories, the learned representation captures a 24-h recovery difference (-6.4 predicted versus -6.7 observed percentage points), a contrast missed by compact state models, which propagates into distinct simulated care demand, queue burden, and bed use. These results indicate that a learned world model of patient dynamics has the potential to replace question-specific state models, allowing new care policies to be evaluated by reconfiguring the scenario without model rebuilding.
[AI-10] Hybrid Machine Learning for Imbalanced Landing-Phase Safety Event Prediction from Flight Data
链接: https://arxiv.org/abs/2611.08386
作者: Taha Asghar,Bilal Mufti,Syed Qasim Zaheer,Luqman Ahmed
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Landing-phase safety-event prediction requires the detection of rare abnormal approaches within strongly imbalanced multivariate flight data. Capturing coupled temporal behavior while distinguishing event-specific signatures across airport operating contexts remains a central modeling challenge. This study proposes an interpretable hybrid feature-fusion framework for multiclass landing-phase safety-event prediction. A long short-term memory (LSTM) autoencoder trained on nominal flight windows characterizes deviations from routine approach behavior. Its continuous reconstruction-error descriptor is fused with parameter-specific statistical flight-history features and supplied to a class-weighted XGBoost classifier to distinguish Normal, Speed High, Path High, and Flaps Late windows. The framework is evaluated using 53,750 NASA DASHlink approach windows from three arrival airports. Pooled and airport-specific learning strategies are compared alongside component-level benchmarks, repeated pipeline retraining, and feature-attribution analysis. The results demonstrate strong class-balanced detection, with mean macro recall above 0.90 across all airports under repeated partitioning and retraining. Component comparisons establish the importance of class weighting and show that continuous feature fusion outperforms reconstruction-error-only and hard-gate alternatives. Airport-level comparisons reveal benefits from local specialization at the two airports with larger datasets and from pooled training at the airport with the smallest dataset. Feature attributions associate predictions with operationally relevant vertical-path, configuration, speed, and approach-context variables. The framework provides an interpretable approach to rare-event detection and practical guidance for adapting flight-data monitoring models to airport-level data availability.
[AI-11] FedReason : Large Language Models as Weight-Space Meta-Aggregators for Heterogeneous Federated Learning
链接: https://arxiv.org/abs/2611.08383
作者: Ali Mohammadi Ruzbahani,Abbas Yazdinejad,Hadis Karimipour
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: Ali Mohammadi Ruzbahani (corresponding)
Abstract:Federated learning servers combine client models with a fixed aggregation rule, and each rule in common use is designed for one kind of heterogeneity. We study whether a large language model (LLM) that inspects the client models can choose the rule itself, without any training on the server. In FedReason, the LLM is called every ten rounds. It analyses the weights by running code in a restricted server-side sandbox, sees only the printed results, may score a few candidates on a 500-image proxy set, and submits a short Python aggregation program. The server applies the program until the next call and falls back to FedAvg whenever its output is invalid. Full weight tensors never leave the server. We evaluated FedReason with Claude Opus 5.5 on CIFAR-10 and CIFAR-100 with ResNet-18 under Dirichlet label skew, width-heterogeneous sub-models and Byzantine clients, in 168 runs against 14 baselines. Across 12 paired label-skew runs it beat FedAvg by 1.26 percentage points on average (10 wins), and a control that spends the same per-round proxy budget on choosing among standard rules by 1.08 points. It was the strongest server-side rule on CIFAR-100 but did not beat SCAFFOLD, which changes client training. With three of ten clients malicious, it excluded all sign-flipping clients from the first round and stayed about one point below FLTrust without training a root model. Without proxy feedback, accuracy fell from 73.8% to 38.8% in one seed. A 60-round run cost 0.79 USD in API fees on average.
[AI-12] NPCBench: A Clinical Apprenticeship Benchmark for Multimodal Care-Pathway Reasoning in Nasopharyngeal Carcinoma NEURIPS2026
链接: https://arxiv.org/abs/2611.08369
作者: Pengkai Wang,Wei-Wei Zhang,Yan Li,Min Tang,Zhitian Hou,Zeyu Liu,Guanghao Zhu,Yuanyi Wang,Yanggan Gu,Wenjun Wang,Minheng Ni,Congkai Xie,Zhijie Sang,Jianmin Wu,Ying Sun,Hongxia Yang
类目: Artificial Intelligence (cs.AI)
备注: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026), Evaluations and Datasets Track
Abstract:Current medical benchmarks typically break down clinical competence into isolated factual questions, single-image tasks, or static case vignettes. These tests score individual answers, but reveal less about whether a model can integrate subspecialty knowledge with multimodal evidence and track a patient’s condition throughout a course of care. To close this gap, we introduce NPCBench, a clinical apprenticeship benchmark for nasopharyngeal carcinoma (NPC). NPCBench mirrors specialist training, from guideline and foundational knowledge to case planning, longitudinal review, CarePath integration, and open-ended consultation. It contains 4,777 component and consultation items and a 100-patient CarePath cohort with 417 scored diagnostic and care tasks. The episode-level evaluation tests whether diagnostic and care answers are jointly correct within a care pathway. By linking stepwise tasks to patient-level care trajectories, NPCBench evaluates whether models ground decisions in visual evidence, interpret changes in patient state, and meet clinical requirements in consultation. Across 30 evaluated models, we observe a composition gap: strong performance on individual tasks does not ensure jointly correct diagnostic and care answers for the same patient. Among 21 multimodal models, strict CarePath success reaches at most 23%. Paired controls and physician calibration further examine evidence grounding, longitudinal revision, and consultation judging. Overall, NPCBench broadens the evaluation of medical artificial intelligence beyond static question-answering paradigms to an apprenticeship-oriented assessment of clinical workflows, thereby offering a testbed for evidence-based and temporally coherent reasoning in oncology care.
[AI-13] Enhancing Local and Global Structure in Compressed AI Agent Memory
链接: https://arxiv.org/abs/2611.08341
作者: Henry Han
类目: Artificial Intelligence (cs.AI)
备注: 28 pages, 12 figures, 10 tables
Abstract:AI agents store text as embedding vectors and compress them at scale. A compression is accepted when each item’s nearest neighbors survive, but agents also consolidate, delete, and compare memories, tasks that depend on which items lie far apart. We evaluate both ends with two label-free fidelities from our prior work: the local fidelity L at 10 neighbors and the global fidelity G at 30 to 50 percent of the memory. At 96 bytes per passage, rescored binary codes reach an L of 0.97 yet move 10.1 percent of passages to another theme; PCA with 8-bit codes, at an L of 0.77, moves 1.7 percent. Rescoring leaves the stored codes, and thus G , unchanged. We prove that across the G band, 1-\nu_k is at most 2 to 2.18 times the square root of the near-far rank-violation rate, and that this rate follows the distance noise of the compression (correlation 0.979). Our locality-aware quantization (LAQ) turns the bound into a hinge objective on the codebooks of an OPQ index, keeping its bytes, rotation, and search. To our knowledge, LAQ is the first quantizer whose objective comes from a proven bound on the far end. Reconstruction alone raises G in 0 of 52 held-out runs; the far hinge raises it in all 52. LAQ raises G in all 52 ( p=8.5\times 10^-4 over 13 memory-encoder groups), including the external FinDER memory, has the highest G of seven equal-size quantizers in all 12 Multi-Doc-2025 runs, scales to a million passages, and keeps evidence recall.
[AI-14] Learning the Task or Learning the Verifier? Rethinking Criteria Exposure in Embodied Self-Improvement
链接: https://arxiv.org/abs/2611.08318
作者: Wenbo Huang,Jianxin Bi,Shunyu Yang,Ruiyang Si,Rui Ni,Qiang Wang,Shulong Jiang,Yintianrun Zhang,Feixiang Ruan,Kewei Zhang,Ye Huang,Shukai Gong,Daquan Zhou
类目: Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: 22 pages, 8 figures, 10 tables. Project page: this https URL
Abstract:Self-improving embodied agents are increasingly being developed and compared on robot simulation benchmarks. Several recent works expose the benchmark’s success criteria to the learning agent, a practice we call Criteria Exposure (CE). These criteria, originally written to verify policies trained by imitation learning, often demand more than the corresponding language instructions, so a score from such a verifier may not faithfully reflect whether the instructed task was completed, especially for robot-control programs written by an agent. As embodied coding agents report ever higher benchmark scores, a natural question arises: when a coding agent is exposed to the benchmark’s criteria, does it generate code that better performs the instructed task, or code that simply overfits to the verification criteria? We conduct controlled experiments with a self-improving agent with and without CE, and find that with CE the agent quickly produces code with a higher success rate under the given criteria, but this advantage disappears when the criteria are changed slightly without departing from the original instructions. More importantly, the skills the agent develops under CE transfer poorly across tasks. We further propose instruction-faithful criteria and show that they better reflect the performance of agent-generated programs.
[AI-15] RM-VLM: From Natural-Language Task Descriptions to Vision-Grounded Reward Machines
链接: https://arxiv.org/abs/2611.08308
作者: Roko Parac,Christos Ziakas,Leo Ardon,Alessandra Russo
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 20 pages including appendix
Abstract:Reward machines (RMs) have been proposed as a reward formulation in reinforcement learning (RL) to capture non-Markovian tasks where credit assignment depends on history. By representing reward functions as automata, RMs expose the task’s temporal structure and allow such rewards to be handled within a Markovian framework. However, they assume a labelling function, a proposition-grounding oracle for observations, which must be manually engineered for each environment, limiting their wider applicability. In this work, we address this limitation by proposing RM-VLM, a novel approach in which a large language model (LLM) generates an RM from a natural-language task description, and a vision-language model (VLM) acts as the labelling function, grounding the RM’s propositions over image-based observations during RM-guided policy learning. A VLM-based value function estimator is used to assess the agent’s progress towards each RM transition and derive a dense reward signal. We evaluate RM-VLM on three domains covering non-Markovian grid-world navigation, robotic manipulation, and 3D object collection. Our experimental results show that RM-VLM consistently achieves higher task success rates than existing VLM-based reward methods across all tested domains.
[AI-16] First-Token Attraction in Mamba Dynamics
链接: https://arxiv.org/abs/2611.08303
作者: Trinh T. Nguyen,Minh-Khoi Nguyen-Nhat,Duy-Tung Pham,Hoang-Son Do,Tan Minh Nguyen,Thieu N. Vo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:We study token dynamics in deep Mamba models, with a particular focus on the attracting role of the first token. Unlike attention mechanisms, where token interactions are governed by positive scalar coefficients, Mamba dynamics involve full matrix-valued interaction coefficients with no positivity or definiteness guarantees. This structural difference makes Mamba dynamics substantially harder to analyze and prevents the direct application of existing techniques for attention-based models. To overcome this challenge, we introduce a new analytical framework based on a high-dimensional invariant box and a sharp exit-time argument, which together recover an effective form of sign-definiteness. Using this framework, we characterize the possible limit points of token trajectories and prove explicit exponential convergence rates. Our analysis reveals that the first token has privileged dynamics: its limiting direction attracts subsequent tokens, inducing token clustering and attention-sink phenomena in deep Mamba models. We further show that this attraction mechanism extends to broader sequence dynamics, including state-space models, causal attention, and classical time-series models. Experiments on Falcon-Mamba-7B validate our predictions, showing that concentration on the first token increases with depth and dominates in later layers.
[AI-17] AutoCompress: Automated Hardware-Guided Model Compression
链接: https://arxiv.org/abs/2611.08290
作者: Sheldon Liang,Ali Taghibakhshi,Saurav Muralidharan,Sharath Turuvekere Sreenivas,Ashwath Aithal,Pavlo Molchanov,Nima Tajbakhsh
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:LLM pruning must preserve quality while meeting hardware- and workload-specific constraints on latency, throughput, and memory. Existing pruning workflows often target parameter counts, requiring practitioners to manually select model sizes that meet latency, throughput, and GPU memory limits. However, parameter count alone does not predict these metrics, which depend on the model architecture, workload, hardware, and serving engine. We introduce AutoCompress, a hardware-guided compression framework that directly incorporates such constraints into differentiable architecture search. AutoCompress jointly optimizes global and per-layer structure to produce heterogeneous architectures, guided by serving-engine cost models validated against end-to-end latency, throughput, and memory measurements. We evaluate AutoCompress on long-context local coding, batched datacenter serving, and latency-sensitive in-game interaction. Compared to the state-of-the-art Minitron pruning framework, AutoCompress improves average accuracy across eight benchmarks by 6.57, 2.35, and 0.57 percentage points on RTX 5070, RTX 5080, and H100, respectively. AutoCompress also compresses Nemotron 3.5 Super from 120B to 74B parameters for single-B200 BF16 serving at 128K context, retaining 95% of its mean benchmark score.
[AI-18] A Generative Model of Contextual Integrity: Appropriate vs. Inappropriate Information Sharing NEURIPS2026
链接: https://arxiv.org/abs/2611.08284
作者: Omer Ebead,Claude Formanek,Joel Z. Leibo
类目: Artificial Intelligence (cs.AI)
备注: Accepted as a main conference paper at NeurIPS 2026
Abstract:Contextual integrity (CI) holds that appropriate information flow depends on context-specific norms. The same disclosure may be appropriate in one context and inappropriate in another. We build a generative model of contextual integrity to study information sharing through simulated conversations. Two actors played by large language models exchange messages, with one holding private information and the other seeking it. We examine how sharing and refusal vary with the situation, the holder’s decision logic, disposition, and the underlying language model. Actors shared information with the rightful recipients when it was needed, but also disclosed it to people who were not entitled to receive it. The model provides a way to study how disclosure and refusal develop through conversation as the situations and the actors vary.
[AI-19] LeanContext: Disentangling Mathematical Resources from Local Formal Context in Lean Whole-Proof Generation
链接: https://arxiv.org/abs/2611.08277
作者: Hang Zhang,Qihao Ye,Litong Liu,Yunbei Zhang,Yugyung Lee,Yue Zhou,Bowen Liu,Emily First,Cheng Han,Qiao Zhuang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Neural theorem proving increasingly uses retrieved premises and project context to support proof generation. Evaluating these resources requires measuring the advantage of target-selected declarations over comparable alternatives and how this advantage depends on training. We introduce LeanContext, a benchmark of 129,826 Lean whole-proof generation problems, to investigate these questions. We group selected complete preceding mathematical declarations as Mathematical Context (MC) and the remaining selected local source commands as Environment Context (EC). Four aligned input views support crossed training and inference evaluations with shared targets and original Lean verification environments. For both diffusion- and autoregressive-based models, selected declarations outperform comparable alternatives by a larger margin when training includes MC. Final-input audits identify omitted declarations referenced by the target. Prioritizing these dependencies improves Pass@32 by 1.40% and 1.70% for LLaDA and Llama, respectively, on 2,000 targets. The selection experiment operates without access to target reference proofs and holds checkpoints, EC, and inference budgets constant. The code and dataset are available at this https URL.
[AI-20] Accelerating Safe Reinforcement Learning with Massive Parallelism NEURIPS2026
链接: https://arxiv.org/abs/2611.08273
作者: Joonyoung Lim,Younghwan Yoo
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026
Abstract:We investigate the problem of scaling safe reinforcement learning (safe RL) to massively parallel training regimes, where a large number of environments are exe- cuted with short rollout horizons. Although this setting improves data throughput and wall-clock efficiency, it creates a mismatch with constrained Markov decision processes (CMDPs), whose safety constraints are defined over full-episode cost returns. We address this mismatch by analyzing staggered environment resets as a phase-mixture distribution over short trajectory segments. This analysis charac- terizes the bias induced by phase coverage and motivates a phase-aggregated cost estimator that reconstructs episode-level cost estimates from short rollout batches. We further incorporate finite-sample uncertainty and distribution mismatch into a constraint-tightening framework, yielding a scalable Safe RL method compatible with on-policy optimization. Using an MJX-based implementation of Safety Gym, our experiments show that the proposed method achieves performance comparable to non-massively parallel implementations while substantially reducing wall-clock training time.
[AI-21] Accepted but Wrong: Silent Failures of LLM Code Repairs after API Migrations
链接: https://arxiv.org/abs/2611.08269
作者: Xiaobo Chen,Genliang Zhu,Ziyuan Wang,Chu Wang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 30 pages, 2 figures. Preregistered; artifacts are available from the authors on request
Abstract:After a runtime or API migration, an LLM-repaired procedure can run cleanly yet return wrong results, and a release check on typical test inputs ships it. We report a preregistered study of 73 tasks covering 60 documented changes in CPython, this http URL and SQLite. Every candidate is executed on the real runtimes against withheld contracts. Over the complete repair space, the all-assertions gate on typical inputs released 0.410 of wrong repairs. One edge input written from the release note cut this to 0.065, and to 0.061 on 15 unseen clusters. Three inputs from a blind GPT-5.6-luna writer that never saw the program reached 0.070. Two further frozen stages enlarged the menu to 1491 fragments and then added 233 hard distractors that match the legacy output on typical inputs. The edge input (0.058) and the frontier-model writer (0.058) passed their preregistered tests. Yet 0.168 and 0.153 of the hard wrong programs still passed. Two local-model writers failed their preregistered tests. Across four model families, the edge input cut wrong releases per attempt in all three open-weight families on all three menus, for example from 0.233 to 0.057, without lowering correct completion. Self-tests helped less. A blind AI audit flagged 34 contracts as stricter than their requirements. Dropping the disputed checks changed 0 verdicts, but class-label disputes remain open. No model-written code was executed. Artifacts are available from the authors on request.
[AI-22] Scaling Neural Motor Decoding via Decoupled Behavioral Pretraining
链接: https://arxiv.org/abs/2611.08263
作者: Divyansha Lachi,Vinam Arora,Shivashriganesh P. Mahato,Alexandre Andre,Eva L. Dyer
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent neural decoding approaches achieve strong performance by training on large collections of paired neural-behavioral recordings, but such datasets are expensive, invasive, and difficult to scale. In contrast, behavioral data is abundant and easy to collect, raising the question of whether neural decoding can benefit from decoupling behavioral and neural representation learning. We introduce BEEMO, a flexible multi-session framework that enables training with arbitrary mixtures of paired neural-behavioral recordings and unpaired behavioral data. A behavior decoder learns movement dynamics directly from behavioral trajectories, while neural activity is incorporated through cross-attention layers when available, allowing behavioral and neural supervision to scale independently. Across multiple intracortical motor datasets and tasks in nonhuman primates, we show that incorporating unpaired behavioral data improves decoding performance, particularly when transferring to new tasks, and that adding even a small amount of behavioral data from the target task further improves transfer. By decoupling neural and behavioral data, we unlock a new dimension for scaling: unpaired behavioral data, which is substantially cheaper and easier to collect than paired neural recordings, can be used directly to improve decoding performance. We further show that behavior-only pretraining, without any neural pretraining, achieves performance comparable to models pretrained on substantially larger neural datasets. Together, our results suggest that scaling behavioral data offers a practical and cost-effective path toward neural decoding models that generalize across subjects and tasks. Code is available at this https URL.
[AI-23] Value-oriented policy distillation for interpretable wind farm power control
链接: https://arxiv.org/abs/2611.08245
作者: Max Bruninx,Timothy Verstraeten,Jan Decuyper,Ann Nowé,Jan Helsen
类目: Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:Wind farm power control can help the system maintain balance and avoid blackouts. In Europe, this can be achieved through implicit balancing, where a wind farm adjusts its power setpoint in anticipation of the imbalance price using information published by the Transmission System Operator. Thereby, the wind farm helps restore system balance while also reducing its own imbalance costs resulting from forecast errors. However, implicit balancing is a challenging problem since the settlement price is unknown in advance and can be impacted by the pursued control strategy. Reinforcement learning methods could help to solve this problem, but lack the interpretability needed to be deployed on real wind farms. Therefore, we propose a policy distillation framework using explainable boosting machines (EBMs), a variant of generalized additive models, to obtain a transparent control policy. To increase the value captured by the EBM model, we introduce a value-oriented perspective to policy distillation inspired by research in the field of energy forecasting. Furthermore, we suggest to include expert knowledge during model design, such that the control strategy is also justifiable to stakeholders. Using real data of the Belgian grid and an offshore wind farm, we find that our value-oriented and expert-based controller can reduce imbalance costs by 82%, while outperforming a price threshold controller as well as a policy distillation baseline using decision trees.
[AI-24] Signals Evidence and Context: User Scrutiny in Real-World LLM Conversations
链接: https://arxiv.org/abs/2611.08232
作者: Ahana Biswas,Yu-Ru Lin
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) often present synthesized answers without making their supporting evidence easy to inspect. We study how features of these answers relate to users’ subsequent accepting, checking, and challenging in 489,283 assistant-user response pairs from ShareChat and WildChat. Broad attribution language appears in 31.9% of qualifying responses, compared with explicit URLs in 1.33%, revealing a gap between claims of grounding and preserved inspectable support. Attribution, uncertainty, and other response cues generally accompany both checking and challenging; only uncertainty shows a small statistically distinguishable pooled association favoring checking. Within Technical Help, uncertainty favors checking relative to challenging; refusals and reasoning traces favor challenging in Seeking Information. Scrutiny is concentrated in some conversations, while its apparent turn-to-turn persistence becomes much smaller when conversation-level differences and response ordering are examined separately. These results suggest that supporting verification requires attention to what evidence can be inspected, which question a user is trying to resolve, and the history of the interaction.
[AI-25] BabyTheorist: A Benchmark for Learning to Theorize the World from Observation Alone
链接: https://arxiv.org/abs/2611.08177
作者: Doojin Baek,Junyeob Baek,Mingyu Jo,Hosung Lee,Taegu Kang,Gyubin Lee,Sungjin Ahn
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:How would we know whether a machine truly understands the world? We argue that one key capability is \textitobservational theory learning: acquiring reusable theories from pure observations and applying them to new situations. Current visual reasoning benchmarks focus primarily on the latter, presenting each task as a few-shot puzzle grouped by a common rule. We introduce BabyTheorist, a benchmark for controlled training and evaluation of observational theory learning. BabyTheorist generates continual streams of before-after visual observations from hidden compositional programs; during training, learners see only the resulting transitions, with no program labels, task identities, or support sets. To support theory learning and evaluation, BabyTheorist controls the program language, a curriculum over program complexity, recurring frequent subprograms, and a continual train-test protocol with level skipping. Seven representative baselines plateau well below ceiling, with the gap widening as the curriculum advances—evidence that BabyTheorist isolates a capability current program-learning approaches do not yet address.
[AI-26] PCBRouteAgent : How Close Are LLM Agents to Becoming PCB Routing Engineers?
链接: https://arxiv.org/abs/2611.08175
作者: Yue Zhang,Jilei Sun,Liqiang Jing,Yijie Wei,Ruohan Yang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Automating printed circuit board (PCB) routing requires coordinating board-level connectivity decisions with geometric execution under strict design rules. Although recent agents can interact with native routing tools, reliable routing on complex boards remains challenging. We present PCBRouteAgent, a multimodal PCB routing agent inspired by how human engineers coordinate global strategy with local execution. A Global Routing Planner combines board images and structured electrical and geometric information to specify connection topology, routing order, corridors, and layer transitions. A Detailed Routing Executor realizes these guides through KiCad’s native routing engine and performs bounded local recovery. Unresolved failures are returned as structured feedback, enabling targeted guide revisions while preserving accepted routes. We evaluate PCBRouteAgent on 247 synthetic and real-world boards from four PCBWorld-Bench test splits. It achieves clean-pass rates of 100%, 94.9%, 100%, and 40% on D2, D3-A, D3-B, and D3-C, respectively. On the three real-world subsets, PCBRouteAgent outperforms the best evaluated baseline in clean-pass rate by 13.9, 36, and 10 percentage points, respectively.
[AI-27] A Generative Foundation Model for Multimodal Tokamak Tasks
链接: https://arxiv.org/abs/2611.08153
作者: Yingshuai Ji,Haodong Feng,Guohong Deng,Gelin Bian,Yue Wang
类目: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE)
备注:
Abstract:Tokamak experiments record diverse measurements, including waveforms, profiles, fields, and scalars, which are used to reconstruct the plasma, forecast its evolution, recover associated control signals, and estimate disruption risk. Most models are built separately for one task and a fixed set of inputs and outputs, and each relies on a prescribed subset of diagnostics rather than structuring the full multimodal observation set to extract global, cross-signal information about the discharge. These limitations motivate a reusable foundation model that adapts to each task’s input-output schema and models heterogeneous target signals jointly, rather than retraining a separate model for every configuration. In this work, we extend the TokaMark benchmark with two groups of tasks for actuator action inversion and disruption prediction, yielding 20 MAST tasks in six groups with explicit schemas and evaluation protocols. Then, we introduce GenToka, a generative tokamak foundation model that jointly generates schema-selected targets through a shared conditional diffusion Transformer, with signal-specific codecs preserving native time series, profiles, fields, and scalars; foundation pretraining and task-specific adaptation transfer cross-signal structure across all six groups. Evaluation on MAST measurements shows that GenToka reduces macro-average normalized root mean squared error relative to TokaMind by 4.7% on Groups 1-4 and by 17.7% on Group 5, and improves macro-average window-level F1 on Group 6 by 8.4% at a common validation-selected threshold. GenToka provides a reusable, generative, and multimodal pathway for experimental tokamaks, linking plasma-state reconstruction, actuator inference, and disruption warning from shared measurements to support more integrated, data-driven fusion plasma prediction and control research.
[AI-28] Who Can Evaluate AI Independently? A Framework for Analysing Evaluation Infrastructure
链接: https://arxiv.org/abs/2611.08134
作者: Sumaya Nur Adan,Joanna Wiaterek,Nourhan Abdelrhim,Sydney Reis,Robert F. Trager,Benjamin Rosman,Marina Jirotka
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注: 19 pages, 3 figures
Abstract:AI evaluation is increasingly central to AI safety and governance, yet taking part in an evaluation does not by itself confer authority over its design, its execution or the resources it depends on. This paper examines how infrastructure arrangements distribute that authority, and so shape who can evaluate AI systems independently. Drawing on Science and Technology Studies, political economy and applied evaluation standards, we introduce PAAI (Power, Access, Adequacy, Independence), a framework for examining who controls evaluation infrastructure, the terms on which it is available, whether it is adequate to the evaluation’s purpose and what autonomy it supports. We apply PAAI to twelve African AI evaluation projects (2022-2025) and to the documented roles of all ten member institutes in the three published joint testing exercises of the international network of AI Safety Institutes. We identify three recurring configurations: extractive collaboration, dependent innovation and constrained independence. African actors often contribute data, language expertise and task design while external institutions retain authority over infrastructure and evaluation frameworks; design authority without infrastructure authority never yielded more than partial independence. The joint testing exercises show a similar division among mostly high-income members: the institutes that led each exercise also ran the models, while Kenya, the network’s only African member, moved from annotation to task design with no documented model execution. Participation in evaluation does not by itself produce evaluative authority. For evaluation to serve as a credible governance mechanism, access must come with control over priorities, methods, publication and continuity.
[AI-29] Create Your Own Work Stack: Deploying Self-hosted AI and Tooling in a Tech Startup
链接: https://arxiv.org/abs/2611.08116
作者: Saad Ahmed Bazaz,Talha Amjad,Abdurrehman Subhani,Sohaib Ahmed Bazaz,Abdullah Aamir
类目: Artificial Intelligence (cs.AI)
备注: Accepted for poster presentation at the Thirty-Ninth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-27), February 18-20, 2027
Abstract:Self-hosting AI is often presented as a way to reduce inference costs and increase control, but deployment introduces its own costs. We report our experience deploying a hybrid AI stack at Grayhat Developers, a small software company with an AI-intensive development workflow. Our initial motivation was to counter the growing cost of commercial AI usage: Claude/Codex incurred approximately \ 325 per month, with more than 25.93 billion recorded tokens. We experimented with local open-source models as replacements for commercial coding agents, but after learning that their performance and inference speed were insufficient, we adopted a hybrid approach in which commercial agents remained responsible for complex tasks while local models handled lower-complexity tasks, organizational information retrieval, and synchronization with self-hosted tools including Outline, Plane, and ERPNext. We estimated that approximately 30–40% of our commercial AI workload was associated with context transfer, documentation synchronization, task management, and related operational coordination. By moving this work to local models, we estimate that approximately 90% of this coordination workload could be handled without commercial inference, corresponding to an estimated 27–36% reduction in overall commercial-AI workload or cost-equivalent usage. Our first internal deployment nevertheless failed to achieve adoption because it introduced UX friction. Replacing it with a centrally configured deployment substantially improved feedback and adoption. We conclude that local inference is most useful not as a wholesale replacement for commercial AI, but as part of a hybrid architecture that assigns tasks according to their capability requirements, cost, context demands, and operational sensitivity. Comments: Accepted for poster presentation at the Thirty-Ninth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-27), February 18-20, 2027 Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.08116 [cs.AI] (or arXiv:2611.08116v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.08116 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-30] Gradient Boosting for Efficient Rank Discovery in Tensor Network Kernel Machines
链接: https://arxiv.org/abs/2611.08093
作者: Albert Saiapin,Kim Batselier
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 13 pages, 3 figures, 1 table. Code available at: this https URL
Abstract:Tensor network kernel machines (TNKMs) combine nonlinear features with low-rank tensor representations to model complex relationships efficiently. However, selecting an appropriate tensor rank remains challenging, as conventional rank-search methods require training multiple candidate models, increasing computational costs. In this work, we focus on a canonical polyadic (CP) tensor network parameterization and propose GB-CPKM, a gradient boosting framework that builds a higher-capacity predictor by sequentially adding fixed-rank learners. This approach provides a new perspective on rank selection by incrementally increasing model capacity, with ensemble size selected based on validation performance rather than through explicit rank search. Experiments on real-world regression benchmarks show that GB-CPKM trains up to 6x faster and uses up to 2x less peak memory than a conventional rank-selection method based on cross-validation, while achieving comparable predictive performance. These results highlight the potential of gradient boosting to improve the computational efficiency and scalability of tensor-network learning.
[AI-31] Value or Attention: Investigating Layer-wise Relevance Propagation for Graph Attention Networks NEURIPS2026
链接: https://arxiv.org/abs/2611.08088
作者: Ken Schreiber,René P. Klausen,Sebastian Lapuschkin,Wojciech Samek,Jonas Naujoks
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at the XAI4Science Workshop @ NeurIPS 2026 (non-archival). 17 pages, 9 figures
Abstract:Graph Attention Networks (GATs) are increasingly used in domains where scientists need to read a model’s reasoning, from medicine to materials science, yet the routing they learn remains opaque to most explainers. Layer-wise Relevance Propagation (LRP) follows the model’s computation, but in an attention layer the relevance reaching a node is the sum of two mechanically different quantities: its content relevance through the value path and its routing relevance through the attention path. Existing rules either drop one or add them into a single number. We write a single product rule with free weights on the two factors that recovers current approaches as special cases. The value path measures what content a node supplied against a zero baseline; the attention path measures how the node shaped what was passed relative to its neighborhood, and its sign records whether the node’s features moved the gating the way the prediction demanded. On a ground-truth classification task, we show that the same node receives negative, positive and near-zero relevance under three settings of the rule. The separated relevance maps yield a hypothesis about the model: it gains by not reading that node. We give reading rules for both paths and argue that they should be interpreted separately.
[AI-32] MUSS: A Multi-scale and Sequence-based Model for Single-cell Gene Regulation NEURIPS
链接: https://arxiv.org/abs/2611.08086
作者: Kangjie Zheng,Amirhossein Vahidi,Liying Jin,Joseph Clarke,Vijaya Baskar MS,Connor Rogerson,Mengji Zhang,Muzlifah Haniffa,Anshul Kundaje,Mo Lotfollahi
类目: Artificial Intelligence (cs.AI); Genomics (q-bio.GN)
备注: 30 pages, 5 figures; includes references, appendices, and the NeurIPS checklist
Abstract:Modeling gene regulation in single cells requires jointly representing DNA sequence, chromatin accessibility, and gene expression across regulatory scales. Sequence-based models capture DNA-level regulatory grammar but often lack cell-level resolution, whereas many single-cell omics models represent genes and peaks as fixed-vocabulary tokens without their underlying sequence. Here, we present MUSS, a sequence-based multi-scale model for single-cell gene regulation that preserves both DNA-level and cell-level resolution. MUSS derives peak- and gene-level representations from DNA using a frozen pre-trained sequence encoder and integrates them through peak-peak self-attention and distance-biased peak-gene interactions. Peak masked language modeling followed by training on paired scATAC-seq and scRNA-seq aligns these representations with cell-specific regulatory states. Across gene expression prediction, enhancer-gene linking, motif-guided regulatory prioritization, and zero-shot cell clustering, MUSS achieves strong performance against sequence-based and single-cell omics baselines. It further predicts expression for genes and peaks unseen during task training. These results support sequence-grounded, multi-scale representations for decoding cell-type-specific regulatory programs. Code, datasets, and model checkpoints are available at this https URL.
[AI-33] DictLLM : Post-training Compression of Large Language Model with Dictionary Kernels NEURIPS2026
链接: https://arxiv.org/abs/2611.08082
作者: Jinho Park,Se Young Chun,Mingoo Seok
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: NeurIPS 2026. Project page: this https URL
Abstract:Efficient deployment of large language models (LLMs) on mobile devices requires balancing performance with compute and memory constraints. Post-training compression is an effective approach to managing this trade-off, while requiring much less computation than in-training counterparts. We propose DictLLM, a dictionary-based compression scheme in which weight matrices within grouped layers are decomposed into a shared dictionary and layer-specific sparse coefficients, directly obtained from a pre-trained model without further training. Specifically, we cast DictLLM as a four-step post-training optimization. First, we obtain shared dictionaries via singular value decomposition (SVD) of input-scaled weight matrices, amortizing rank across grouped layers; second, we sparsify the layer-specific coefficients to the target compression ratio using a dictionary-aware Hessian; and third, we refine the dictionaries given the resulting sparse coefficients. Finally, we aggressively quantize both components using factor-specific Hessian information: input covariance for the dictionaries and the dictionary-aware Hessian for the coefficients. The resulting representation reduces memory footprint by storing a single shared dictionary per layer group, along with sparse layer-specific coefficients. On language modeling and zero-shot reasoning benchmarks, DictLLM consistently outperforms existing post-training low-rank compression methods and pushes the Pareto frontier between model size and performance. DictLLM compresses the original LLaMA-7B model by 5.1x with only a 2.9-point increase in WikiText-2 perplexity, achieving a 3.6x higher compression ratio than the state-of-the-art low-rank method.
[AI-34] Can an LLM Actually Reason and Program about Neural Networks Weights or is it just memorization on steroids? Exploring “Weight-Space Programming”
链接: https://arxiv.org/abs/2611.08070
作者: Nazmus Sakib Ahmed
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLMs) have recently been reported to read and write the raw weights of neural networks directly, a capability termed weight-space programming. We ask whether the reading side of this capability reflects reasoning about weight flows or a combination of local arithmetic at the network’s edges and memorized priors over public datasets. We test Claude Opus 5.5 on small convolutional networks of about 125,000 parameters trained on five image datasets, providing every weight as plain text with no tools, no forward passes, and no dataset or class names. The model performs accurate local calculations: batch-normalization statistics lead it to infer about three optimizer steps on networks stopped after three updates, and output-layer sign patterns recover the vehicle/animal split of CIFAR-10. The sessions report reading the first and last layers closely and sampling the middle, and several claims built on partial tracing fail verification, although we did not run a direct forward-pass test. Semantic judgments lean on familiar datasets: the model names CIFAR-10 for CIFAR-10 networks but also for all three fast-food networks, relabels KMNIST as MNIST, reads a 99.57%-accurate SAT-6 network as CIFAR-10 animal photos, and calls “frog” the output favored by a feature that fires most on trees. Permuting CIFAR-10’s labels preserves the output split at the new indices, but every class pair the model interprets is then wrong, indicating reliance on remembered label order. These are exploratory results from mostly single runs. We conclude that zero-shot weight reading by this model is better described as local arithmetic plus dataset priors than as weight-flow tracing, and we provide a protocol for auditing future weight-space claims.
[AI-35] When Imperfect Automated Fact-Checking Backfires: Social Learning in the Presence of Misinformation
链接: https://arxiv.org/abs/2611.08053
作者: Zixuan Liu,Ming Cao
类目: Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注: 8 pages, 5 figures
Abstract:Automated fact-checking tools, cheap and fast but imperfect, are increasingly available on social platforms. We study when such tools can backfire in a sequential social learning model of misinformation. Agents observe an article’s reliability and their predecessors’ actions, which indicate whether they liked or disliked the article. They may also pay to fact-check it. We examine how welfare, defined as the probability of a correct cascade, changes when a free, error-prone automated tool supplements costly but perfectly accurate manual fact-checking. Two opposing effects arise. Broader use of fact-checking helps agents escape wrong cascades. At the same time, errors from the automated tool can mislead agents who would otherwise have verified the article perfectly, reducing the likelihood of correct cascades. We prove that adding the tool strictly lowers welfare for articles of intermediate reliability when truthful articles are not too informative and manual checking is sufficiently costly. Simulations show that the tool can help on either side of the intermediate range, and that it improves on manual-only checking on average only above a threshold accuracy.
[AI-36] Secure Formal Raising: Protecting the Development of AI through Executable Properties
链接: https://arxiv.org/abs/2611.08042
作者: Rolf Drechsler,Christina Plump
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Model raising revisits Turing’s developmental view of machine intelligence. Formal Raising (FR) has recently been proposed to complement this view with explicit behavioral properties. This paper extends that proposal to the security of the raising environment. Secure Formal Raising (SFR) models policy integrity, authorization, and confidential data flow using four-valued evidence that preserves unknown and inconsistent assessments. We prove six security invariants and show that they force admission to accept only confirmed evidence. A plain-C reference implementation satisfies the invariants in a bounded exhaustive exploration, which also exposes a confused-deputy gap for tool calls that a provenance requirement closes.
[AI-37] Circe: Training-Free LLM Self-Evaluation via Component Interaction Consistency
链接: https://arxiv.org/abs/2611.07948
作者: Shuyi Miao,Zheng Li,Xu Xie,Linrui Xu,Fei Shen,Xiaoyu Du,Tat-Seng Chua
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As large language models (LLMs) are increasingly deployed in real-world applications, evaluating the correctness of their responses is crucial for reliable use. However, such evaluation is challenging because reference answers are rarely available in practice. Existing reference-free methods mainly rely on output signals, which may assign high scores to incorrect responses and thus do not necessarily reflect correctness. This motivates exploring the model’s internal computation for complementary correctness signals. Recent studies exploit internal states, but they mainly analyze hidden-state trajectories across layers, which record only the combined updates of attention and feed-forward networks (FFNs) and overlook how these components interact. We therefore propose Circe, a training-free self-evaluation method that assesses a response using both internal and external signals, without reference answers or repeated sampling. For internal signals, we first introduce component interaction consistency (CIC), which measures the directional alignment between attention updates in earlier layers and attention-induced FFN output changes in later layers. We further incorporate representation geometry (RG) to characterize normalized displacement and angular changes along layerwise hidden-state trajectories. For the external signal, we convert the mean negative log-likelihood of generated tokens into a bounded likelihood confidence (LC) score. Finally, we fuse CIC, RG, and LC into a unified correctness score. Experiments on five LLMs and six benchmarks across four domains show that our method distinguishes correct from incorrect responses more effectively than existing baselines, and generalizes to multilingual, multimodal, and open-world settings.
[AI-38] ChemE-as-Code: An Open-Source Platform for Agent ic AI in Chemical Engineering
链接: https://arxiv.org/abs/2611.07947
作者: Artur M. Schweidtmann
类目: Artificial Intelligence (cs.AI)
备注: 28 pages, 4 figures, 5 tables; Supporting Information (28 pages) appended. Submitted to AIChE Journal. Code and data: this https URL
Abstract:ChemE-as-Code provides an open-source environment in which AI agents built on large language models (LLMs) can carry out chemical engineering work and engineers can check it. Its calculation tools range from catalysis through process simulation to plant layout, HAZOP and life cycle assessment, and skills tell a general-purpose agent how to use them. Engineering files are linked, machine-readable and versioned in Git; calculations run outside the LLM, each value records its origin, and automated checks report inconsistencies and out-of-date results by file and line. In an illustrative ammonia case, agents developed a written task into a preliminary basic-engineering package (simulation files, PIDs, layout, HAZOP, datasheets, relief) for USD 224 of API cost in the design sessions. Planted faults and engineering changes show that the checks find what their rules cover and miss the rest, so they guide human review rather than replace it. No independent engineer has reviewed the design.
[AI-39] ok: a textual concrete syntax of BPMN and DMN for humans and AI agents
链接: https://arxiv.org/abs/2611.07945
作者: Marek Deja
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 29 pages, 3 figures
Abstract:tok is a plain-text notation for Business Process Model and Notation (BPMN) 2.0.2 process models and Decision Model and Notation (DMN) 1.5 decisions. One engine checks a tok document against both standards, compiles the document to XML with an automatic diagram, and serves language-model applications through the Model Context Protocol. The paper classifies the notation as an alternative concrete syntax of the BPMN and DMN metamodels and reports one study. A tok model takes 2.2 times fewer tokens than the smallest BPMN XML of 893 files exported by 37 tools and 7.7 times fewer than the files as exported. The checker signals 98.2% of 1,411 injected faults, against 85.8% for the XML Schema with bpmnlint. Four language models new to tok write valid tok from a one-page reference for 85.7 to 100% of 87 process descriptions from two datasets and valid BPMN XML for every description, and the strongest model writes tok with 25% and 19% fewer tokens than BPMN XML. As agents with a lean skill and one checker tool, the models write valid tok for 97.6 to 100% of the descriptions, the strongest model spends 21% and 14% fewer tokens on tok than on BPMN XML, and a skill which sends the model to read the reference through a tool costs 1.7 to 2.8 times the tokens of the lean skill. Of 47 event logs simulated from public models, 44 fit their source models, and tok imports every model discovered from the logs.
[AI-40] AttnSR: Attention Step Reward for Long-Horizon Agent ic RL
链接: https://arxiv.org/abs/2611.07932
作者: Mengsong Wu,Tianchen Zhou,Zhaochen Su,Zhenhua Liu,Zhe Cui,Aimin Zhou
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 20 pages, 4 figures, including appendices. Code: this https URL
Abstract:Long-horizon LLM agents make multi-step decisions, yet outcome-reward-only reinforcement learning provides little direct supervision for individual steps. We introduce AttnSR, a step-level reward augmentation method that derives reward allocations from policy attention extracted during training replay, without a separately trained reward model or additional rollouts. AttnSR traces backward through the policy model attention graph to estimate the relevance of earlier steps to the final response. The terminal step, measured relative to a running baseline, determines whether these rewards reinforce or penalize the selected steps. In a controlled Qwen2.5-7B-Instruct study, AttnSR consistently outperforms the PPO baseline across all evaluated checkpoints on ALFWorld, raising final held-out success from 91.2% to 97.1%. The ablation experiment shows that the benefit depends on how intermediate rewards are assigned, rather than simple accumulation of step rewards over time. AttnSR also improves ScienceWorld at every checkpoint while cutting environment interactions by 29.9%. These results support policy attention as a practical prior for step-level reward allocation in long-horizon agentic RL. Code is available at this https URL .
[AI-41] How Finite Brains Generalize Without Bound: Spatial Memory as a Universal Computer
链接: https://arxiv.org/abs/2611.07920
作者: Yuhang He,Si Wu
类目: Artificial Intelligence (cs.AI)
备注: 57 pages, including appendices
Abstract:A child who learns to add small numbers adds numbers of any length. LSTMs and Transformers trained on short sequences fail on longer ones. We trace the difference to what a learner reads when it decides. A position, a count or a time keeps taking new values as the input grows: it is a coordinate, and a learner that reads one meets in every long input values its experience never covered. Whether two things are the same has the same two answers at every size: it is a coincidence, and a learner that reads only coincidences finds in a long input nothing that short inputs lacked, so finite experience fixes its rule for every size. The memory carries the size of the input and the rule stays finite, and this determines the memory a generalizing brain can have: a memory that grows while what the learner reads of it stays finite has to look the same from every place, so it is made of cycles, and the past is reached by asking what happened when the cycles last stood where they stand now. The cycles are the grid modules, the cells that detect their return the place cells, the records the hippocampal index; a prefrontal controller moves the cycles and never sees them. Reading only coincidences costs no power: what finite experience fixes in such a learner is what Turing machines compute in step with their input, which is why a circuit that evolved to find the way home can learn arithmetic and grammar. In a network of neurons this machine of cycles is the dynamics seen through its symmetry, and symbols are its invariants. Trained on strings up to 40 symbols, the machine is at least 99% correct, on average over lengths 41 to 500, on nine of fifteen benchmark tasks, five of them tasks that no LSTM or Transformer solves; given new cycles after training it reverses strings of 15,015 symbols without error and fails beyond that on the curve the theory predicts; and its controllers run unchanged on neurons.
[AI-42] Adaptive Semantic Branching: Disagreement-Localized Cross-Trajectory Regeneration for Reasoning Candidate Expansion
链接: https://arxiv.org/abs/2611.07918
作者: Jeong-Yoon Kim,Tae-Yoon Lee
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Reasoning trajectories generated for the same problem often arrive at conflicting answers. These disagreements reveal alternative intermediate decisions that can be reused to explore new solutions. We introduce Adaptive Semantic Branching (ASB), a training-free inference procedure that turns cross-trajectory disagreement into targeted candidate generation. ASB first groups independently sampled trajectories by extracted final answer, selects a representative from each group, and aligns their reasoning steps with a monotonic dynamic program. A persistent-divergence objective identifies the point at which initially similar reasoning paths develop different continuations. For each selected directed pair, ASB retains the source trajectory’s textual prefix, extracts a short reasoning window from a competing donor trajectory, and generates a fresh solution conditioned on both. A diversity-aware policy allocates up to three additional generations according to the answer-cluster structure. In a 500-problem MATH evaluation with Qwen2.5-7B-Instruct, ASB increases oracle candidate coverage from 445/500 (89.0%) with seven initial samples to 455/500 (91.0%), recovering correct candidates for 10 of the 55 initially uncovered problems. The method uses 576 additional generation calls across the evaluation set, averaging 8.152 calls per problem. We provide a precise specification of the branching operator, alignment objective, generation policy, and experimental configuration.
[AI-43] GuidedWAM: Training-Free Energy Guidance for World Action Models in Autonomous Driving
链接: https://arxiv.org/abs/2611.07912
作者: Xiyan Su,Finn Rasmus Schäfer,Hojun Lim,Frank Diermeyer,Johannes Betz,Markus Lienkamp
类目: Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注:
Abstract:World action models (WAMs) for autonomous driving couple the driving policy and the visual world model in a single generative network whose objectives are fixed post-training. Conditioning on a new driving objective requires retraining. We study energy guidance as a general mechanism and present GuidedWAM, a model-agnostic framework for injecting driving objectives into frozen WAMs at sampling time. Our framework constructs each objective as differentiable energies on the sampler’s intermediate clean-action estimate, which needs only forward access to the frozen model, and is stabilized by sampling-time guidance scheduling, per-energy normalization, and velocity-bounded correction. On the NAVSIM v2 hard-split, guidance improves all three frozen, publicly released WAMs by 1.8 to 3.6 EPDMS, averaged over three sampler seeds, and adds less than 10% to the sampling time per step. These results position energy guidance as a practical post-training control channel for WAMs.
[AI-44] DISCERN: Distributional Synthetic Replay for Retain-Free Incremental Machine Unlearning
链接: https://arxiv.org/abs/2611.07892
作者: Nicolás Segura-Kunsagi,Juan Luis Suárez-Díaz,Iván Sevillano-García,Francisco Herrera
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 32 pages, 10 figures
Abstract:Machine unlearning aims to remove the influence of specific training data from a trained model without retraining from scratch. In real deployments, however, deletion requests arrive over time, calling for incremental class unlearning (ICU): classes are removed sequentially, previously forgotten knowledge must remain forgotten, and the remaining classes must be preserved. We present DIstributional SynthetiC rEplay for Retain-free iNcremental unlearning (DISCERN), an intervention-based class unlearning method that replaces access to the retain set with synthetic embeddings sampled from per-class parametric distributions. Unlike prototype-based replay, which assumes an isotropic Gaussian around each class centroid, DISCERN fits Euclidean (Gaussian, Laplace, Student’s t ) and directional (von Mises–Fisher, Fisher–Bingham) families to each class and selects among them through statistical tests and, when none of them accepts, cross-validated log-likelihood. This design is motivated by the architecture-dependent latent geometry: transformer embeddings lie approximately on a hypersphere and concentrate around class directions, unlike those of convolutional networks. We evaluate DISCERN against seven retain-free baselines on MNIST, CIFAR-10, and CIFAR-100 with two convolutional and two transformer backbones, under single-class, multi-class, and incremental unlearning. DISCERN is competitive when classes are removed at once and is the most stable method in ICU on transformers: on ViT-B/16, after 20 sequential deletions it retains a harmonic mean of 0.797 , whereas the strongest method in the simultaneous setting drops from 0.918 to 0.048 . A paired ablation shows that per-class distribution selection substantially reduces the resurfacing of previously forgotten classes, acting as a stabilizer of accumulated forgetting rather than a uniform improvement.
[AI-45] When Bigger Becomes Brittle: Why Compression Hurts More as Language Models Scale
链接: https://arxiv.org/abs/2611.07891
作者: Gautam Gupta,Dikshant Kukreja,Unique Kumar,Avni Mittal,Deepanshu Dabas,Kshitij Sah,Erik Cambria,Rajiv Ratn Shah,Zhengkui Wang,Avinash Anand
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Quantization and pruning reduce the cost of deploying language models, but their effects are usually reported as benchmark scores. Such scores record only the loss that remains at the output. Inside the model, compression damages some components more than others, and the rest of the network can partly compensate. The observed loss therefore reflects both the damage and this compensation, which benchmark scores cannot separate. How the two combine, and whether their balance changes with model size, remains unclear. We study this compensation for in-context retrieval, a capability carried by a small, identifiable set of attention heads. To separate the two, the expected loss from ablating these heads is decomposed exactly into the part restored by the rest of the network, attributed to the components responsible, and the part that reaches the output. This decomposition is then compared across model sizes. For damage of comparable size, smaller models restore most of the loss, whereas larger models retain most of it, because the compensating responses offset less of it as models grow. Compressing the heads produces the same pattern. Compression that is nearly harmless in small models can therefore be costly in large ones.
[AI-46] When Where How: Active In-Context Learning for Tabular Foundation Models
链接: https://arxiv.org/abs/2611.07886
作者: Christian M. M. Frey
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Tabular Foundation Models (TFMs) adapt to a new task only through the labeled rows in their context. However, labels are expensive to acquire in real-world scenarios due to the need for highly specific domain knowledge to ensure high-quality data. Active in-context learning transfers classical acquisition rules to this setting, in which the model that guides acquisition changes with every label added. We ask how the value of diversity, predictive uncertainty and target-aware information gain changes as a frozen TFM acquires more context. For that, we study Target-Aware Nearest Neighbor (TANN) acquisition and the Expected Predictive Information Gain in Context (EPIC) and evaluate them against classical rules. Our study across twenty benchmarks with two well-established TFMs and classical GBDTs highlights three findings: (i) When: diversity-based acquisition is best during the first acquisition iterations; (ii) Where: labeling pool rows that serve the target inputs via a simple target-aware nearest-neighbor rule matters substantially; (iii) How: EPIC yields the best models at the label budget, retains on average 80–87% of the chance-corrected full-data performance with only 5-10% of the pool labeled on the eight benchmarks with the largest pools, and selects labels that also improve the predictive performance of gradient-boosted decision trees.
[AI-47] FARR: Forecast-Aware Residual Refinement with Matched Endpoint Evidence for Long-Horizon Time-Series Forecasting
链接: https://arxiv.org/abs/2611.07870
作者: Yiyang Zhang,Yuli Zheng,Bin Li,Meimei Zhang,Qiang Sun
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 3 figures, 4 tables
Abstract:Large language models use auxiliary conditioning to incorporate additional information into time-series forecasts. However, auxiliary conditioning does not consistently improve accuracy and can increase forecasting error when its influence is poorly matched to the observed temporal context. We propose Forecast-Aware Residual Refinement (FARR), which learns to correct forecasts using endpoint evidence. FARR obtains a valid forecast under the standard condition and a null forecast under a predefined null condition. Their difference reveals how conditioning changes the prediction. A shared residual refiner combines this endpoint evidence with historical features to estimate the remaining error of the valid forecast. Horizon-wise gating then controls the contribution of the predicted residual to the final output. Experiments on four long-horizon forecasting benchmarks show that FARR achieves the best dataset-average MSE and MAE among the compared methods. Code will be released at this https URL.
[AI-48] Computation not weight access: large language models as auditors of clinical prediction networks
链接: https://arxiv.org/abs/2611.07847
作者: Zhiyu Liu,Xinyu Yang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 13 pages, 2 figures, 6 tables. Code, simulated networks, prompts and replies: this https URL
Abstract:Frontier large language models (LLMs) have been reported to read, edit and write the weights of small neural networks, raising the prospect of auditing clinical prediction models from their files alone. Public demonstrations, however, mixed reading raw weights with running code. We separated these conditions in internally pre-specified simulation experiments. Tanh networks of 105 to 38,401 weights were trained on simulated cohorts of patients with subarachnoid haemorrhage, each with a planted relationship or none, and one network was trained on labels that under-recorded deaths in a subgroup. GPT-6 Astra and four other GPT models answered single prompts under a default or an explicit system instruction, either with the weights in the prompt and no tools, or with a calculation-only tool that exposed the weights (white box) or only the network’s predictions (black box). Without tools, Astra recovered planted relationships in 353-weight networks but became unreliable at 3,457 weights, and under the default instruction it declined 11 of 12 audit and repair tasks as unidentifiable; the explicit instruction removed these refusals. With the tool, it recovered the planted relationship in all 29 runs on networks of up to 38,401 weights, raised no false alarm in 15 runs on null networks, and found the biased subgroup in 6 of 6 runs. Black-box access performed as well as white-box access, and other GPT-6 models read as accurately. A well-designed script matched the comparator models on the audit and the best LLM repair. For auditing these networks, what mattered was computation and elicitation, not access to the weights.
[AI-49] Instantiating the Knowledge Plane ALT
链接: https://arxiv.org/abs/2611.07827
作者: Francisco Pereira,Justine Sherry,Pratiksha Thaker
类目: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI)
备注: Accepted at HotNets '26 (Salt Lake City, UT, USA). 11 pages
Abstract:In 2003, Clark et al. proposed the “Knowledge Plane”, a layer of intelligent agents that share rich information about network state in order to collaboratively manage the Internet across administrative boundaries. The technology of the day lacked the two capabilities it demanded most: reasoning and unstructured communication. Thus, automated network management has remained largely confined to single domains, while interdomain problems are still resolved by human operators on the phone. We argue that agentic large language models can finally unlock the potential of the Knowledge Plane, but that doing so will require solving new research problems that arise particularly in this decentralized, multi-agent, coopetitive setting that have received limited attention in the literature thus far.
[AI-50] he Intrusion Pathway: How Non-Evidential Context Degrades Language Model Performance KR
链接: https://arxiv.org/abs/2611.07778
作者: Dikshant Kukreja,Gautam Gupta,Avni Mittal,Avinash Anand,Kritarth Prasad,Zhengkui Wang,Timothy Liu,Rajiv Ratn Shah,Aik Beng Ng,Simon See,Erik Cambria
类目: Artificial Intelligence (cs.AI)
备注: 32 pages, 25 figures, 29 tables. Dikshant Kukreja and Gautam Gupta contributed equally. Website: this https URL Code: this https URL
Abstract:Language models lose accuracy when a prompt contains non-evidential context: irrelevant statements from unrelated domains, or random off-distribution tokens, with no competing answer and no semantic overlap with the query. Such noise pervades retrieval-augmented, tool-using, and long-context deployments, yet the failure has lacked a mechanistic account. We trace the corruption to the Intrusion Pathway, which starts with MLPs enriching both query-subject and irrelevant-token representations, parroting heads carry prefix-token content to the prediction position, and relational readers largely avoid unrelated sources. Nevertheless, preserved subject enrichment and selective reading do not prevent intrusion in the broader computation. We then localize part of this interference to a sparse set of attention heads, which we call intrusion heads. Their prefix-derived writes promote distractor-related content, while their reduced support for the correct answer arises primarily from weaker non-prefix contributions rather than direct suppression by the prefix. The same heads transfer across tasks and distractor types and are largely weight-disjoint from the heads that integrate relevant contextual evidence, enabling a class-conditional intervention. Replacing the corrupted last-token residual with its clean counterpart, an oracle probe that requires a matched clean forward pass, recovers 97.6% on average of the distractor-induced drop across four downstream tasks. A deployable mask level knockout of about ten heads, roughly 1% of the model’s attention heads, recovers 57% at bounded cost to helpful context.
[AI-51] RIVER: Evidence-guided multi-agent retrieval for enzyme–reaction discovery
链接: https://arxiv.org/abs/2611.07757
作者: Yifan Chang,Yunze Li,Feifei Jiang,Juntao Wu,Xin Chen,Han Hu,Yucheng Wu,Wei Tang,Mingqian Ma,Wanli Ouyang,Kaipeng Zhang,Lei Bai,Hongting Tang,Feiran Li,Yuchen Ren
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Enzyme function annotation and the identification of candidate enzymes for a given reaction are essential steps in connecting genomic information to biochemical reaction networks. However, existing methods rely on sequence or structural similarity or on training with known enzyme–reaction pairs, making it difficult to integrate biochemical evidence across large candidate spaces; they also struggle with retrieval tasks involving enzyme families whose members are highly similar. Here, we introduce RIVER (Reaction–enzyme Inference via Evidence-guided Retrieval), a multi-agent framework for bidirectional enzyme–reaction retrieval that identifies potentially catalyzed reactions for a given enzyme and ranks candidate enzymes for a given reaction. RIVER coordinates large language models, specialized biochemical computational tools, and external knowledge resources to assemble evidence at multiple levels, including reaction chemistry, protein sequences and domains, catalytic mechanisms, and database annotations. Dedicated validation modules further examine evidence gaps and potential conflicts and rerank candidates using traceable evidence records. On public benchmarks, RIVER achieves the highest success rates in both enzyme retrieval and enzyme annotation. In family-specific enzyme retrieval, RIVER outperforms the evaluated baselines on the selected datasets and metrics. In Escherichia coli enzyme retrieval tasks constructed from recent literature, RIVER achieves the best retrieval results with closed-source and open-source model configurations, substantially outperforming closed-source general-purpose and biology-specialized multi-agent systems, including Claude Code, Codex, and Phylo Max. These results suggest that evidence-guided multi-agent retrieval can prioritize biochemically supported enzyme–reaction relationships, providing a basis for subsequent experimental validation.
[AI-52] Evaluating Language Model Features for Zero-Shot Time-Series Anomaly Detection Under Isolation Assumption
链接: https://arxiv.org/abs/2611.07720
作者: Romain Hermary,Nesryne Mejri,Djamila Aouada
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Time-Series Anomaly Detection (TSAD) is a fundamentally challenging task: it is unsupervised, requires high local precision capabilities, and suffers from limited labelled data. Recent work has explored adapting Large Language Models (LLMs) to time-series via cross-modal transfer. However, existing approaches are poorly suited to Multivariate TSAD (MTSAD), as tokenisation often obscures fine-grained temporal and variables structures, and most methods optimise auxiliary objectives such as forecasting or reconstruction. We posit that these surrogate objectives become unreliable as model capacity increases. Highly expressive models risk reconstructing or forecasting anomalies, undermining the foundations of both strategies. Instead, we propose to use LLMs purely as feature extractors and directly analyse their induced feature spaces. In this work, investigate the ability of LLMs to provide useful representations for MTSADs in a zero-shot setting through classification. To this end, we study the separability of the extracted features, under the assumption that anomalous features should be isolated from normal ones. We analyse their temporal and contextual discriminative ability using simple step-wise tokenisation schemes, with one time-step per token, which preserve fine-grained local information. Their usability for MTSAD is assessed using an unsupervised Isolation Forest, enabling a direct evaluation of the resulting representation quality. We further investigate whether the observed performance can be attributed to the language-related pre-training of the LLMs or arises independently of it. Our results provide empirical insight into the ability of LLMs to represent multivariate time-series for anomaly detection and highlight broader applications of cross-modal foundation models.
[AI-53] Pediatric ECG foundation model via adult-to-pediatric fine-tuning: downstream tasks and clinical evaluation
链接: https://arxiv.org/abs/2611.07717
作者: Yu Cao,Haiyan Xu,Xiaohan Li,Jian Yang,Yuqing Feng,Wei Wang,Haomin Li
类目: Artificial Intelligence (cs.AI)
备注: 20 pages,3 figures
Abstract:Pediatric electrocardiogram (ECG) interpretation is complicated by age-dependent electrophysiological changes that limit direct transfer of adult AI models. We asked whether an adult ECG foundation model could be adapted to pediatric ECGs. Using 339,563 12-lead ECGs from 238,189 children, we compared direct application, pediatric fine-tuning, and training from scratch under an identical architecture. Direct transfer was poor (mean AUC 0.8256). Fine-tuning reached 0.9695, above de novo training (0.9661), and attained 0.9079 using only 1% of training data. On 10,999 external ECGs, fine-tuning achieved a mean AUC of 0.9385 across 30 labels. It generalized to age regression (MAE 1.12 years) and congenital heart disease classification (AUC 0.9562). Across 17 clinicians and 100 cases, diagnostic accuracy rose from 56.8% to 65.4% (p 0.001) and did not decline when AI erred. Adult ECG foundation models require pediatric adaptation; fine-tuning offers a data-efficient route to scalable pediatric ECG AI.
[AI-54] CAMP: Context-Aware Mapping and Placement for LLM s on 3D-IMC Arrays
链接: https://arxiv.org/abs/2611.07707
作者: Orhun Görkem,Athanasios Vasilopoulos,Georgios Pagonis,Victoria Clerico,Hadjer Benmeziane,Irem Boybat,Hsinyu Tsai,Abu Sebastian,William Andrew Simon
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI)
备注:
Abstract:Large Language Models (LLMs) incur severe memory-access overheads due to the cost of loading large weight matrices and key-value (KV) caches. Analog In-Memory Computing (AIMC) is an emerging solution that enables computation directly on static weights stored in memory, substantially reducing energy and latency. 3D-IMC further stacks memory arrays into vertical tiers, increasing capacity but revealing a new challenge: how weights are placed in memory determines which matrix-vector multiplications (MVMs) can execute concurrently, affecting latency and throughput. The concurrency aspect of 3D-IMC weight-mapping remains underexplored in the literature, leading to sub-optimal utilization of compute resources on 3D-IMC chips. In this context we propose CAMP (Context-Aware Mapping and Placement), an inference-context-aware weight mapping algorithm for 3D-IMC-based LLM accelerators. CAMP captures the potential resource contention resulting from the architecture, targeted network, and runtime context in a conflict matrix, enabling throughput-optimized weight mapping. This mapping is refined through a profiling-driven feedback loop that efficiently prunes the search space and returns optimized mappings within practical runtimes. CAMP is an offline, software-based solution requiring no hardware changes and improving 3D-IMC compute utilization from 42.4% to 60.3%, delivering up to 2.0x higher 3D-IMC throughput and 1.74x higher end-to-end throughput over state-of-the-art mapping algorithms.
[AI-55] owards Generalist Graph-Level Anomaly Detection NIPS26
链接: https://arxiv.org/abs/2611.07706
作者: Junjun Pan,Yixin Liu,Yu Zheng,Fuyi Li,Alan Wee-Chung Liew,Shirui Pan
类目: Artificial Intelligence (cs.AI)
备注: Accepted by NIPS26
Abstract:Graph-level anomaly detection (GLAD) aims to identify graphs that deviate significantly from the majority. While recent advances have achieved competitive performance, existing approaches mainly follow a dataset-specific paradigm, requiring substantial in-domain data and training cost when applied to new application scenarios. This limits their scalability and practicality in real-world applications where labeled data is scarce and diverse domains are continuously emerging. In this paper, we investigate the problem of generalist GLAD, which seeks to learn a single unified model capable of detecting anomalous graphs across multiple domains with minimal supervision on the target data. This problem introduces key challenges, including learning transferable patterns from heterogeneous graph datasets and effective adaptation under limited supervision. To address these challenges, we propose GenGLAD, a generalist GLAD approach that leverages discriminative subgraph extraction as the core mechanism to learn transferable anomaly knowledge and enable efficient adaptation to new domains. Specifically, GenGLAD employs a graph neural network-based discriminative subgraph extraction model to learn the anomaly-discriminative substructures that distinguish anomalous and normal graphs across diverse domains. During the test-time adaptation phase, multi-dimensional discrepancies between the extracted subgraph and multi-level contextual information are calculated to characterize anomalies from multiple perspectives. For efficient adaptation, a lightweight scoring module is introduced to automatically decide the most informative signals, improving robustness against misleading patterns. Extensive experiments on diverse benchmark datasets demonstrate that the proposed approach achieves strong performance, superior generalization ability, and improved data efficiency compared to existing methods.
[AI-56] Equivariant Deep Learning in Direct-Product Space for Electronic-Structure Hamiltonian Prediction
链接: https://arxiv.org/abs/2611.07695
作者: Shi Yin,Zujian Dai,Xinyang Pan,Lixin He
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI); Computational Physics (physics.comp-ph)
备注:
Abstract:Most existing equivariant architectures for Hamiltonian prediction propagate features in direct-sum spaces of irreducible spherical tensors, whereas the target Hamiltonian is naturally an operator block in a direct-product representation over local orbital and spin bases. This representation mismatch fragments contiguous Hamiltonian blocks, obscures direct manipulation of orbital couplings, and organizes features into irregular angular-momentum components that require numerous transformations and are poorly aligned with efficient hardware execution and memory management. To address both the representation mismatch and computational bottleneck, we propose DP-Learner, an equivariant learning framework that performs message passing directly on Hamiltonian-like operator blocks in the direct-product space. Theoretically, we propose Lie-generator-based directional embedding and invariantly weighted left–right products and contractions that preserve equivariance to SO(3), SU(2), and spatial inversions, without explicit Clebsch–Gordan decomposition or inverse reconstruction. Methodologically, we instantiate this principle as a Transformer architecture, in which Hamiltonian-like feature matrices or tensors evolve through equivariant operators in the direct-product space, while non-linear invariant scalar routing controls structured gating over matrix units, directional propagation, attention, and environment-embedded context interactions. We further construct MP-HAM-110K, a large-scale Hamiltonian dataset comprising more than 110,000 material structures covering 68 elements with both Hamiltonian and wavefunction data. Experiments on MP-HAM-110K and Materials-HAM-SOC demonstrate that DP-Learner dramatically improves Hamiltonian prediction accuracy and inference efficiency over the state-of-the-art method.
[AI-57] MCPHallu: Benchmarking Reasoning Execution and Memory Hallucinations in MCP Agents
链接: https://arxiv.org/abs/2611.07689
作者: Yizhen Jiang,Haowei Guo,Tianyi Bai,Zengjie Hu,Lu Ma,Zhengyang Zhao,Haoze Sun,Peng Pei,Wentao Zhang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model agents are increasingly deployed through the Model Context Protocol (MCP), which allows them to discover and invoke tools across multiple servers at runtime. Existing MCP benchmarks primarily measure task completion; where they diagnose failures, they do so after execution rather than building specific reasoning, execution, and memory failure risks into task construction. We introduce MCPHallu, a benchmark for diagnosing hallucination-prone agent failures in MCP environments along three axes: reasoning, execution, and memory. MCPHallu contains 358 tasks across five capability domains and runs on 36 production MCP servers in isolated containers. Each task is designed with a structural trigger that makes one of four subtypes the dominant failure risk: Branch Collapse, Unreachable Goal, Tool Misuse, or Context Forgetting. This design enables subtype-level analysis rather than aggregate task success alone. We evaluate fourteen frontier LLM agents from eight vendors on 5,012 trajectories. Current MCP agents remain vulnerable to hallucination-inducing tool environments, achieving an average score of 0.642 on a [0,1] scale (higher is better). Failures concentrate on Unreachable Goal tasks, where the average score drops to 0.458 and 23.2% of runs pass, indicating that agents often continue issuing tool calls or claim successful completion rather than recognizing infeasible goals. Subtype rankings diverge sharply: a model’s rank shifts by 6.9 of 13 positions on average across subtypes, and Branch Collapse is uncorrelated with Unreachable Goal across models ( r = -0.12 ). MCPHallu is released at this https URL. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.07689 [cs.AI] (or arXiv:2611.07689v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.07689 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-58] Who Should Know What? Empirical Information Design for Multi-Agent LLM Economies
链接: https://arxiv.org/abs/2611.07667
作者: Mete Erdogan,Mirac Suzgun,Aneesh Pappu,Guowei Xu,Batu El,Andrew A. Schoen,Furong Huang,John P. Dickerson,Mert Pilanci,James Zou
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The performance of multi-agent LLM systems depends not only on the capabilities of the underlying models but also on information-sharing policies, which govern what information each agent observes and at what level of detail. But does providing AI agents with more detailed information always improve collective outcomes, or can transparency paradoxically degrade efficiency? To investigate this, we introduce empirical information design, an experimental framework that systematically compares disclosure policies across economic markets populated by LLM agents, while holding interaction rules and models fixed. We evaluate ten frontier LLMs in multi-resource trade, bilateral bargaining, and double auctions, using ground-truth utilities to decouple identified opportunities from realized outcomes. We find that finer information disclosure yields highly model-dependent results. In trade, revealing exact valuations rather than ordinal preference rankings reduces realized gains for GPT-6 Sol (from 98% to 91% of maximum) yet increases them for Claude Sonnet 5 (from 72% to 86%). In bargaining, exact information leads Grok 4.7 to produce less balanced agreements, whereas Claude Opus 5.5 divides gains symmetrically. In double auctions, full disclosure inflates price demands, triggering market breakdown and lowering welfare for Gemini 3.1 Pro (97% to 78%), Claude Opus 5.5 (94% to 64%), and Grok 4.7 (97% to 62%). Overall, these results suggest that model selection and information design must be co-optimized, as LLMs vary in how their strategic responses to information impact system objectives.
[AI-59] Energy-Aware Open-Vocabulary Detection Scheduling for Low-Power Robotic Autonomy
链接: https://arxiv.org/abs/2611.07663
作者: Abdul Basit,Muhammad Shafique
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 6 pages, 8 figures, accepted for publication in ICARCV 2026
Abstract:Open-vocabulary object detection (OVD) lets robots search for run-time object queries, but fixed-rate inference spends embedded compute on low-value frames during idle motion, map revisits, or stable viewpoints. This paper presents an energy-aware OVD scheduler for low-power mobile robots. The scheduler uses odometry, map change, and detector confidence to switch among low-, mid-, and high-power YOLO-World-S modes on a Jetson Orin Nano. We evaluate the energy-quality trade-off on an AgileX LIMO using 40 real navigation trials: two target objects, five placements per target-policy pair, and four perception policies. Across bottle and cup trials, the scheduler records 8/10 mission success compared with 7/10 for each fixed-rate baseline, while reducing mean Jetson input power (a Jetson-side power-draw metric) by 5.2% relative to an 8 Hz high-rate baseline and lowering achieved OVD call rate by 34.0%. It does not minimize mission-window Jetson input energy because longer missions can erase instantaneous power savings. A separate 60 s stationary microbenchmark isolates the perception workload: the scheduler lowers mean power from 8.045 W to 7.301 W and energy from 484.7 J to 438.1 J relative to the 8 Hz baseline. These results support robot-state-aware OVD scheduling as a promising power-management mechanism with competitive task reliability, while showing that mission-window Jetson input energy must be read together with task progress and runtime.
[AI-60] PharmAgents 2: Towards Autonomous End-to-End Drug Discovery
链接: https://arxiv.org/abs/2611.07653
作者: Haichuan Tan,Bowen Gao,Tao Huang,Jiaxin Tan,Xin Hong,Baoguo Yao,Yifei Wang,Zhichao Deng,Hanji Du,Dapeng Jiang,Yiheng Li,Weiwei Ma,Xin Ma,Yuqi Shi,Zixiang Wang,Kelin Wu,Tianxiang Wu,Wenxuan Xie,Jie Zang,Huiyu Zhang,Mojia Zhang,Yuying Zhang,Ya-Qin Zhang,Wei-Ying Ma,Yanyan Lan
类目: Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM)
备注:
Abstract:Drug discovery requires continuous scientific decision-making across a vast research space. We introduce PharmAgents2, an autonomous system for end-to-end drug discovery that combines knowledge-enhanced decision-making with organized large-scale exploration. Paper2Skill extracts transferable scientific decision knowledge from public sources and, together with expert-developed skills, supports research decisions. Self-Scaling Discovery expands exploration through parallel agents inheriting research context and decision-making capabilities across targets, scientific hypotheses, and discovery strategies. We introduce PharmAgents-Bench v0.1, comprising 121 questions across nine decision scenarios and three end-to-end tasks assessed with expert-defined process rubrics. PharmAgents2 outperforms evaluated baselines in most decision categories and achieves the highest process scores on all three end-to-end tasks. In three real-world studies, autonomous investigations generated findings subsequently supported by human-executed experiments. For LpxH, PharmAgents2 hypothesized a previously unreported allosteric site and screened compounds against it, identifying a novel inhibitor scaffold with experimentally confirmed inhibitory activity and binding. For JAK1, it prioritized a covalent inhibition strategy and nominated compounds with experimentally demonstrated activity. Starting from a known bioactive molecule, Self-Scaling Discovery explored candidate targets across the proteome and identified a new binding target, with the interaction subsequently confirmed experimentally. PharmAgents2 represents a prototype of a virtual pharmaceutical company, marking a step toward autonomous end-to-end drug discovery.
[AI-61] Dont Prompt Your Agents . Equip Them: Skillware as a Modular Standard for Sovereign Auditable and Deterministic Capability Execution for Logical Systems
链接: https://arxiv.org/abs/2611.07652
作者: Vladimiros Peilivanidis(ARPA Hellenic Logical Systems)
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 51 pages, 26 figures. Code repository and benchmarks available at this https URL
Abstract:Current autonomous logical systems rely on a fragile assumption: that language models should dynamically write code, guess API payloads, and memorize operational constraints through prompt engineering. In practice, this burns massive context windows, inflates API bills, and produces brittle agents that fail unpredictably in production. In this work, we present Skillware, an open architectural framework and capability registry that formally separates what an agent wants to do from how that task is executed. Instead of prompting models with loose text instructions or markdown files baptized as ‘skills’, Skillware equips them with modular, versioned software capability bundles. Much like skill hotkeys in video games that execute deterministically under fixed world rules, each bundle combines a typed JSON input contract, deterministic Python code, task directives, bundled datasets, and automated unit tests. In-process adapters translate these capabilities into native tool declarations for any model, from frontier cloud APIs to local open-weight runtimes with zero network latency. Evaluating 23 frontier and open-weight models across 368 benchmark conditions, we show that small local models (3B to 8B parameters) equipped with Skillware consistently outperform unequipped frontier models on complex multi-step workflows. They cut token usage by up to 84.8%, slash inference costs by up to 97.3%, and eliminate parameter errors entirely. True agent reliability does not come from larger models or longer prompts, but from equipping logical systems with verified, deterministic software. Finally, we articulate the trajectory of capability bundles across an invariant continuum of information processing systems: from digital logical systems today, to IoT, drones, and embodied robotics tomorrow, establishing a foundation for auditable, real-time capability execution. Comments: 51 pages, 26 figures. Code repository and benchmarks available at this https URL Subjects: Artificial Intelligence (cs.AI); Software Engineering (cs.SE) ACMclasses: I.2.11; D.2.11; I.2.7 Cite as: arXiv:2611.07652 [cs.AI] (or arXiv:2611.07652v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.07652 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Vladimiros Peilivanidis [view email] [v1] Thu, 5 Nov 2026 08:38:51 UTC (5,397 KB)
[AI-62] GraspBench: Execution-Aware Cross-Robot Evaluation of 6-DoF Grasp Generation in MuJoCo
链接: https://arxiv.org/abs/2611.07651
作者: Abdul Basit,Bibek Poudel,Muhammad Shafique
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 7 pages, 6 figures, accepted for publication in ICARCV 2026
Abstract:Modern 6-DoF grasp generators are often evaluated using grasp scores, dataset-level metrics, or single-robot demonstrations, yet proposal quality alone does not determine whether a grasp is executable by a particular robot. We present GraspBench, a controlled MuJoCo protocol for evaluating complete grasp-generation pipelines after robot-specific inverse-kinematics filtering, motion planning, and fixed approach-close-lift execution. GraspBench evaluates six configured pipelines across four robot embodiments, five diagnostic objects, and a structured tabletop pose grid, yielding 540 scenarios per method and 3,240 method-scenario evaluations in total. Missing outputs are conservatively counted as failures, and success follows the binary lift criterion produced by each method’s execution runner. GraspGen achieves the highest overall success rate (73.1%), followed by EconomicGrasp (63.5%), while robot- and object-conditioned analyses reveal substantial variation across embodiments and object geometries. GPG produces the largest native candidate sets but achieves 44.3% success, indicating that proposal volume alone does not guarantee robot-executable grasps. Cluster-aware paired intervals support the principal method differences while not estimating repeated-execution uncertainty. The results characterize complete configured pipelines on fixed scenarios rather than intrinsic model quality or stochastic repeatability, showing that grasping performance depends jointly on the generator, robot embodiment, scene, and execution protocol.
[AI-63] What the Middleman Delivers: Answer Behavior Tool Permissions and Billed Usage in LLM API Relays
链接: https://arxiv.org/abs/2611.07650
作者: Xiaodong Wu,Zhimin Zhao,Qi Li,Chaklam Cheong,Xiangman Li,Jianbing Ni
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Third-party application programming interface (API) relays offer access to large language models (LLMs), often at lower prices. Users depend on these services for model quality, instruction compliance, and accurate billing. However, existing studies leave unclear whether relays meet these expectations together. We test 20 relays offering GPT-6 Astra and Claude Opus 5. First, we compare answers to knowledge questions with batch-matched official responses. In our first-round sample, the analysis flags 7/10 low-price and 2/10 same- or higher-price relays for behavioral differences. On retesting 35 pairs, seven gain or lose a flag under the original rules, and six under matched correction families. Because answer comparisons cannot reveal every delivery problem, we examine returned content and tool permissions. No injection is confirmed in 3,332 screened responses to a fixed, non-adversarial code and setup workload, yet four relays return tool calls despite instructions disabling them in the first round. Finally, to connect reported usage to actual costs, we compare input-token counts with official references and check balances. For identical requests, five relays report 4,200 to 60,000 more input tokens than the official API. Where balances are available (six services), charges follow reported usage at listed prices, including extra reported input tokens at one service. Our study shows that advertised models and prices are insufficient to judge relay reliability. It identifies unexplained billed token quantities and tool-permission violations as risks and shows why service evaluations need repeated measurements.
[AI-64] WaSP: Wasserstein Data Synthesis Pipeline for Solver-Free Operator Learning
链接: https://arxiv.org/abs/2611.07646
作者: Ryan Liu,Chuwei Wang,Ricardo Baptista,Andrew Stuart,Anima Anandkumar
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Neural operators offer fast surrogates for solving PDEs, but training them typically requires expensive input-solution pairs. Two solver-free alternatives include physics-informed learning and training on manufactured data. Physics-informed learning approximates solutions by minimizing PDE residuals. While the residual norm must be strong enough to control solution error, stronger norms increase mini-batch variance and hinder optimization. Manufactured data constructs input-solution pairs for learning by applying the PDE operator to artificial solutions. However, the resulting training distribution can be mutually singular with the target distribution, yielding poor accuracy on the target distribution. We introduce the Wasserstein Data Synthesis Pipeline (WaSP), which trains on manufactured input-solution pairs while gradually mitigating the distribution shift. WaSP synthesizes new pairs by applying the learned operator to a path of intermediate distributions converging to the target input distribution. Training proceeds in stages along this path. At each stage, the previous surrogate generates solutions, and the exact PDE operator derives corresponding inputs, yielding training pairs for the next surrogate. Theoretically, we prove a stepwise error-propagation bound for the distribution shift introduced during synthesis. Empirically, WaSP reduces solution error by 3.9\times to 9.5\times compared to existing solver-free methods on PDEs such as Allen-Cahn, Kuramoto-Sivashinsky, and Ginzburg-Landau, and better reproduces the two chaotic systems’ statistical observables.
[AI-65] Beyond Instruction Override: Benchmarking Indirect Prompt Injection with Task-Tampering Attacks
链接: https://arxiv.org/abs/2611.07640
作者: Rongting Zhang,Martin Bertran,Sandeep Avula,Chia-Jung Lee,Vanessa Murdock
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: 34 pages, 6 figures. Code and data: this https URL
Abstract:Indirect prompt-injection (IPI) benchmarks predominantly test planted instructions that redirect a tool-using agent toward an attacker-chosen task. Yet attacker-controlled content can instead tamper with the outcome of the task already in progress. A buyer’s dispute evidence can cite a false refund policy to inflate a refund without changing the task. We define IPI by authority rather than instruction-like wording: planted content constitutes an attack when it attempts to influence a decision that its source is not authorized to determine. We introduce a benchmark that makes this boundary explicit: in each of 90 validated cases across three tool-use environments, the user request specifies the source that resolves the targeted decision. Every attack tampers with a decision required by the benign task rather than replacing that task. Each case has attacks with the same harmful outcome that target up to six parts of the agent’s decision making: its objective, decision rules, available actions, expected consequences, reading of observations, or beliefs about the current state. Without a defense, attacks on three of the five other targets (decision rules, available actions, and state) have higher mean success than objective attacks across five LLMs. A detection-and-removal defense reduces objective-targeting attacks to 2.4% success but leaves observation- and state-targeting attacks at 9.5% and 12.1%; this gap persists under detection-optimized rewrites. A tool-policy defense still allows 15.4% attack success, close to 17.0% without a defense, because these attacks generally retain expected tools and manipulate values obtained during execution. Evaluations whose attacks redirect the task or target only the agent’s objective can therefore overestimate the coverage of these defenses. Code and data are available at this https URL.
[AI-66] Measuring the Marginal Value of Verifier-Based RL Beyond Supervised Fine-Tuning
链接: https://arxiv.org/abs/2611.07639
作者: Amit Sagtani,Saloni Garg
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 8 pages, 2 figures. Code and artifacts: this https URL
Abstract:Reinforcement learning from verifiable rewards (RLVR) is usually evaluated as an alternative to supervised fine-tuning (SFT), or as a stage that follows it, with the implicit assumption that the RL stage contributes gains of its own. We measure that contribution directly, in a controlled setting where both stages can be run to completion and compared exactly. We built gpu-train-debug, a deterministic benchmark of 1,877 synthetic GPU and LLM training failure diagnosis tasks across 30 archetypes, graded by a programmatic verifier with partial credit, and calibrated so that a mid-tier model scores in a 0.60 to 0.75 band. A pilot produced a surprise: 1,200 gold demonstrations nearly solved the benchmark (SFT 0.785 on the hard split), while 15 steps of GRPO from the base model did not move it (0.314 vs 0.311 base). We therefore amended the study, in writing and before the full run, to ask the sharper question: what does verifier-based RL add beyond demonstrations? At the pre-registered scale (60 GRPO steps, learning rate 1e-6), the answer is: no detectable improvement in-distribution (+0.22 points, 95% CI [-0.06, +0.51]), and a modest but measurable improvement under distribution shift (+0.33 points on a held-out OOD split, CI [+0.02, +0.68]). SFT accounted for most of the available improvement, while verifier-based RL produced only modest additional gains under distribution shift. The residual errors SFT leaves are mostly systematic, identical across samples, which likely reduces the within-group learning signal available to GRPO. A full 60-step cold-start GRPO arm confirms the pilot at the full horizon. The whole study cost 8.69 of compute and is fully reproducible from archived generations, frozen configs, and tagged code.
[AI-67] Redacting Personal Data: A Speed and Accuracy Comparison of Regex Presidio and Hybrid Designs
链接: https://arxiv.org/abs/2611.07633
作者: Sai Yashwant,Manoj Kumar Gondimalla,Arjun Sharma,Shubham
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Many enterprise projects now pass free text to large language models (LLMs), and that text often contains personally identifiable information (PII) such as names, phone numbers, email addresses, home addresses and dates of birth. Once a prompt is sent, this data is copied into provider systems, application logs and observability traces. Redacting PII before the call is an obvious control, yet teams seldom have numbers on what it costs in latency or what it lets through. This paper compares four redaction designs that any LLM pipeline can adopt: a pure regular-expression (regex) redactor, Microsoft Presidio running on a spaCy model, a hybrid in which Presidio finds only person names while regex handles structured identifiers, and a fast hybrid that calls spaCy’s named-entity component directly. All four share a known-entity rule, which removes the person the system already knows it is handling, and a consistent placeholder scheme, so the LLM can still tell one person from another after redaction. We evaluate the designs on a sample workflow, which we call ProfileAssist, that turns resumes into structured profiles, using 100 synthetic resumes (523 characters on average, with 900 PII values and 600 business terms that must survive). Regex ran at a 0.20 ms median but caught only 88.9% of PII, because it cannot find names that are not known in advance, such as the references listed in a resume. Presidio, hybrid and fast hybrid each caught 100% of PII and kept 100% of business terms, at median latencies of 37.59 ms, 30.23 ms and 13.48 ms. Measured against a 9,120 ms median LLM call, even the slowest option adds only 0.41% to the request. We therefore recommend the fast hybrid as a default for LLM pipelines that handle personal data, with regex kept as a safety net rather than as the only control.
[AI-68] Beneath the Surface: What We Overlook When We Talk About Agent Evolution
链接: https://arxiv.org/abs/2611.07624
作者: Xiaobo Wang,Zimu Wang,Siyu Ye,Jiaqi Li,Min Tang,Jun Bai,Yuling Shi,Ruobing Wang,Jingjing Zhang,Chaofan Wang,Kai Wu,Mengnan Qi,Kai Cai,Xiaodong Gu,Qi Liu,Zilong Zheng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Agent harness evolution enables recursive self-improvement (RSI) without updating model weights. We examine whether higher benchmark scores yield reliable, efficient, and transferable improvement. We evaluate four methods across harnesses, general and coding tasks. Repeated evaluations, controlled agent expansion, component interventions, and admission replays reveal recurring patterns. All sixteen evolved agents gain previously uncovered tasks while losing previously covered ones. In one general-task case, reliability rises by 5.2 points despite losing ten previously covered tasks. Most reliability gains also increase deployment cost. Specific changes usually outperform generic reminders, but their benefits can change when combined. The highest-scoring agent does not yield the largest mean improvement in any of our four pools, and stricter admission can discard useful alternatives. Transfer selectively preserves gains: coverage usually falls on the general-task target across both harness types and paradigms, but rises on the coding target with the outer-layer-only harness. These findings motivate preservation-aware evaluation, tested decision rules, and explicit applicability conditions. As a constructive follow-up, EvoAudit turns these diagnostics into revisions of select, analyze, mutate, and gate, achieving the best performance.
[AI-69] he Popular Get More Popular: Auditing Brand Bias in Generative Search
链接: https://arxiv.org/abs/2611.07614
作者: Giorgio Monaco,Jacopo Lenti,Francesco Pierri
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI)
备注:
Abstract:The shift from traditional to generative search has fundamentally transformed online product discovery. Because conversational recommendations synthesize choices into a few direct answers, subtle systemic preferences toward specific providers can exert profound commercial impacts. In this work, we audit this phenomenon through two distinct dimensions: (i) brand popularity bias, the systematic preference for market-dominant entities over long-tail alternatives, and (ii) brand familiarity bias, the ranking advantage an entity encoded in an LLM’s parametric memory holds over an unfamiliar competitor. First, evaluating commercial queries across major engines, we observe persistent brand popularity bias: dominant brands in the top quintile are cited 1.4x to 2.8x more frequently than the average brand within a sample of 1,176 top market players, already present in parametric memory that external web retrieval fails to mitigate. Second, through controlled in-context experiments holding product attributes constant, we find that substituting authentic brand signatures with unfamiliar ones significantly degrades recommendation rank. Conversely, fact-preserving content rewrites yield negligible ranking gains. Taken together, our findings demonstrate that generative search engines reinforce a ``rich-get-richer’’ dynamic by favoring established market leaders and imposing a parametric barrier to entry that standard content optimization cannot overcome.
[AI-70] Align Before Aggregation: Basis-Consistent Federated LoRA under Heterogeneous Ranks NEURIPS2026
链接: https://arxiv.org/abs/2611.07598
作者: Pengpeng Qiao,Yang Cao,Lingling Zhang,Guo Cheng,Junwei Chen,Manjiang Yu,Wei Yang Bryan Lim,Masatoshi Yoshikawa
类目: Artificial Intelligence (cs.AI)
备注: Accepted by NeurIPS 2026
Abstract:Federated fine-tuning of large language models (LLMs) under heterogeneous client capacities is increasingly important, and low-rank adaptation (LoRA) with heterogeneous ranks makes it practical. Existing server-side aggregators compose local matrix updates \Delta W_i from client-specific LoRA factors (A_i,B_i) before aggregating them. Although these matrix updates lie in the same weight space \mathbbR^m\times n and can be averaged directly, the LoRA factors that generate them may expressed client-specific local basis. In the typical compose–aggregate–factorize–dispatch loop, this basis inconsistency can make the rank- r_i reference prefix dispatched to lower-rank clients unstable after server-side refactorization. In this work, we propose \textbfFedAbA, a basis-consistent federated LoRA aggregation framework for heterogeneous ranks. FedAbA first extracts each client’s local basis, aligns it to the corresponding prefix of the global reference basis, and constructs a basis-consistent reconstruction of each local matrix update before aggregation. It then aggregates these reconstructions with weights that combine client data size and consistency score, followed by exact coefficient-space SVD refactorization and gauge fixing before rank-prefix dispatch. Our theoretical analysis clarifies the role of basis consistency and supports the proposed consistency-aware aggregation mechanism. Extensive experiments show that FedAbA outperforms representative federated LoRA baselines, with the largest gains under the strongest rank heterogeneity.
[AI-71] Evolver: Turning Agent Exploration into Evolving Capabilities
链接: https://arxiv.org/abs/2611.07595
作者: Haoyang Zhang,Qumeng Sun,Xiao Zhang,Xiang Liu,Zhidong Fan,Junjie Wang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 12 pages, 5 figures
Abstract:Agent exploration produces valuable procedural knowledge, yet much of this experience remains confined to transient execution traces and is not systematically reused. We introduce Evolver, an experience-driven framework that transforms agent exploration into persistent, reusable capabilities without updating foundation model parameters. Built around structured Gene assets, Evolver integrates experience acquisition, retrieval, and feedback-guided evolution to support experience-conditioned execution across existing agent harnesses. We evaluate Evolver using Claude Code and Codex, two foundation model configurations, and four benchmarks: Terminal-Bench 2.1, SWE-bench Multilingual, SWE-rebench, and GAIA. On a selected diagnostic set of 320 paired executions, Evolver improves task success from 83.13% to 87.50% (+4.38 percentage points) while reducing recorded task-execution token consumption by 24.24%. Furthermore, among 265 pairs successful under both conditions, token consumption decreases by 17.57%, demonstrating efficiency benefits even when task success is preserved. These findings suggest that persistent procedural experience can improve both task effectiveness and execution efficiency, providing a practical foundation for transforming agent exploration into evolving capabilities.
[AI-72] Learning and Steering in Incomplete-Information Games
链接: https://arxiv.org/abs/2611.07589
作者: Asrın Efe Yorulmaz,Tamer Başar
类目: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Systems and Control (eess.SY); Optimization and Control (math.OC)
备注: 57 pages, 5 figures
Abstract:In many modern multi-agent systems, a mediator seeks to influence strategic decision dynamics using only outcome-level feedback and without knowing agents’ utilities. In this work, we study how a mediator can learn an unknown game and steer adaptive agents toward outcomes that maximize her utility using only trajectory feedback. The mediator knows the game form and her own objective, but not the agents’ utilities. Moreover, she commits to nonnegative payments and provides recommendations and contexts, but receives only a single realized terminal outcome rather than full-information feedback. We develop halving-type algorithms that learn utility differences in normal-form games and conditional continuation-value differences in extensive-form games, achieving a sample complexity of \widetilde O(\varepsilon^-2) that matches worst-case lower bounds up to logarithmic factors. We then combine these learning procedures with a Lagrangian formulation in an online Learn-Compute-Steer algorithm that interleaves learning, equilibrium computation, and steering toward mediator-optimal extensive-form correlated equilibria (EFCE). With high probability, the Learn-Compute-Steer algorithm achieves a directness gap and an optimality gap of \widetilde O(T^-1/4) , while time-average payments are of \widetilde O(T^-1/8) .
[AI-73] FlashOffloading: Low-Latency KV Cache Offloading for Long-Context LLM Serving
链接: https://arxiv.org/abs/2611.07542
作者: Xuan Wang,Zhuanhao Wu,Tingzhu Fang,Xushi Li,Sicheng Zhang,Zhenyu Yang,Feng Liu,Zhaohui Fan,Qiang Xie,Huayi Zou,Yuncong Liu,Zeyu Song,Ke Zhu,Yanjia Zeng,Jingchun Gao,Yan Shi,Kui Luo,Gezi Li,Liangjun Feng,Cheng Li
类目: Artificial Intelligence (cs.AI)
备注: 13 pages including references,13 figures
Abstract:In large language model (LLM) serving systems, the key-value (KV) cache grows linearly with sequence length and concurrency, so long-context inference easily hits the high-bandwidth memory (HBM) capacity wall and results in low concurrency. Offloading the KV cache to host dynamic random-access memory (DRAM) frees HBM space and raises concurrency, but the additional cache management and host-to-device (H2D) load overhead inflates inference latency, higher concurrency fails to translate into effective throughput. We present FlashOffloading, a host-DRAM-centric inference architecture that fully offloads the KV cache in both the prefill and decode phases. FlashOffloading builds rolling HBM cache for prefill, a hot cache for decode, a reverse sparse index with zero extra miss-lookup overhead, and a compute–load overlapped kernel, optimizing KV offloading latency across the entire pipeline and delivering throughput gains under a time-per-output-token (TPOT) service-level objective (SLO). In experiments deploying GLM-5.2 with multi-token prediction (MTP3) speculative decoding on an Ascend910C prefill–decode (PD)-disaggregated platform, FlashOffloading improves prefill throughput by up to 1.38(\times) and decode throughput by up to 1.91(\times) under a 30 ms TPOT constraint. FlashOffloading is integrated into an open-source serving framework: this https URL.
[AI-74] CST-Bench: Benchmarking Counterfactual Spatial Transition Reasoning with Causal Transition On-Policy Self-Distillation
链接: https://arxiv.org/abs/2611.07541
作者: Ying Li,Yongle Cheng,Jing Li,Xingquan Cai,Yawei Luo
类目: Artificial Intelligence (cs.AI)
备注: 18 pages, 5 figures, 8 tables; includes technical supplementary material
Abstract:Spatial reasoning is a fundamental capability of multimodal large language models (MLLMs), yet existing benchmarks mainly focus on static scene understanding. When objects are moved or swapped, models must infer not only spatial relations but also how interventions change the scene state. To address this problem, we introduce CST-Bench, a benchmark for counterfactual spatial reasoning. CST-Bench contains 7,082 examples covering 6 intervention types and 12 transformation operators, with dense causal annotations including intervention sources, affected relations, expected effects, invariant relations, and metric changes. We benchmark general-purpose MLLMs on CST-Bench and further compare post-training strategies on CST-Bench and standard spatial benchmarks. Results show that methods effective for ordinary spatial reasoning can still fail under counterfactual interventions. We further formulate counterfactual spatial reasoning as causal state-transition modeling and propose CT-OPSD, which uses CST-Bench annotations to guide a teacher model in learning counterfactual state transitions and causal reasoning chains. These chains are then distilled into a student model that reasons only from real observations. Overall, experiment results establish CST-Bench as a challenging benchmark and reusable supervision resource for counterfactual spatial reasoning. CT-OPSD further provides an initial demonstration of how its annotations can support model improvement.
[AI-75] Goal-Branching LaCAM: Temporary Goals as Search Branches in Dense Multi-Agent Pathfinding
链接: https://arxiv.org/abs/2611.07528
作者: Stephen Chang,Daniel Harabor
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:LaCAM is a search-based solver for multi-agent pathfinding (MAPF) that runs a depth-first search over joint configurations, using PIBT to generate successors and lazily adding single-agent constraints to explore alternatives. Although LaCAM is complete, it can struggle on dense, corridor-heavy instances because extracting blocking agents from deadlocks requires coordinated moves of several agents, often away from their goals, which PIBT’s greedy one-step successors rarely propose and which LaCAM can only reach by enumerating constraints one agent at a time. This paper proposes Goal-Branching, where blocking agents are given temporary goals, each tried as an additional branch of LaCAM’s search. We examine random and corridor-aware rules for choosing these temporary goals and show that completeness is preserved for any such rule. On dense instances (60-90% of free cells occupied) of three maze and room maps, Goal-Branching solves 89% of instances against 44% for LaCAM*, and, on instances both solve, it finds a first solution 10-14x faster on maze-32-32-2 and room-32-32-4, and nearly 200x faster in median time at 75% density on room-32-32-4. Solution cost increases by 15-22% on instances both solve, while on open maps and the full MovingAI benchmark its solutions cost no more than those of LaCAM* and it is at most 3% slower.
[AI-76] Graph learning-assisted logic-based Benders decomposition for distributionally robust in-station train rescheduling under disruptions
链接: https://arxiv.org/abs/2611.07520
作者: Xiaobin Li,Yanbin Gao,Jianing Long,Wuming Lei
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Uncertainty in resource reopening times complicates train holding and rerouting decisions when resource closures coincide with arrival delays and processing-time extensions. We develop a rolling-horizon distributionally robust model that updates predictions of the remaining closure duration as observations of disruption resolution become available. The model jointly optimizes train routes, sequences, and timings. It minimizes the worst-case expected total completion delay over an ambiguity set of scenario probabilities. The solution algorithm, GNN-LBBD, uses logic-based Benders decomposition to coordinate routing and resource scheduling. Conditional cuts are constructed from resource-use patterns, and a graph neural network guides the order in which resource relaxation subproblems are solved. Across 28 test instances for a five-track station, GNN-LBBD achieved a mean total completion delay of 61.13 train-min. This was 16.63% lower than tabu search and 20.06% lower than variable neighborhood search, with a mean total computation time close to that of tabu search. Information comparisons on six representative instances showed that using observations of disruption resolution reduced mean total completion delay by approximately 21.7%. Scenario-based and conditional-mean predictions produced similar overall results under the same information conditions. The benefits of information were more pronounced in some instances with alternative routes. This suggests a relationship between the use of observations, the station resource structure, and the scope for schedule adjustments.
[AI-77] DelegateEval: Evaluating Platforms for Personal External Service Agents
链接: https://arxiv.org/abs/2611.07508
作者: Jichao Li,Jinrong Lu,Manjiang Yu,Zhixuan Liu
类目: Artificial Intelligence (cs.AI)
备注: 28 pages, 7 figures, 6 tables. Evaluation framework and experimental design
Abstract:AI agents that work for an individual over time may also help that individual serve external clients. We define a personal external service agent as an agent authorized and continuously managed by a specific individual to accept client requests, deliver work, and provide ongoing support using that person’s knowledge, experience, professional methods, and service boundaries. Such agents could allow individuals to serve more clients with less direct involvement and generate economic value at scale. We propose DelegateEval, a platform-level evaluation framework for agent systems that may combine large language models, reusable skills, protocols such as MCP, and executable tool environments. It covers eight dimensions: ease of creation, service discovery and matching, delivery quality, expertise inheritance, service continuity, learning and improvement, reliability and governance, and efficiency and value. These dimensions assess whether a platform enables people with different backgrounds to turn their expertise into ongoing services that deliver verifiable work, develop capabilities through experience, and remain economically feasible. We present the framework and an experimental design. Complete datasets, executable environments, a unified scoring protocol, and cross-model results remain to be developed and empirically validated.
[AI-78] SAVeR2: Reasoning -based Safety Alignment for Large Reasoning Models via Verifiable Rewards
链接: https://arxiv.org/abs/2611.07482
作者: Yichen Sun,Qirui Liu,Yan Wang,Jianan Lin,Linbo Jiang,Shiyu Wang,Zhibo Wang,Zhixuan Chu
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large Reasoning Models (LRMs) have achieved impressive performance via explicit Chain-of-Thought (CoT) reasoning, yet this process introduces critical safety risks. We formalize LRM safety objectives as requiring non-malicious reasoning and answers, safety consistency, and explicit intent identification. We argue that existing alignment methods suffer from poor adversarial generalization, a significant ``safety tax’’ on reasoning, and unfaithful coarse-grained rewards, primarily due to their neglect of the LRM’s intrinsic reasoning capabilities for safety. To address these, we propose \textbfSAVeR ^2 , a novel \textbfSafety \textbfAlignment method via \textbfRewards with \textbfReasoning capability. SAVeR ^2 decomposes the safety objective into four fine-grained verifiable reward signals and utilizes a difficulty-aware data selection strategy to stabilize training. Crucially, we introduce a reasoning-preserving gradient projection mechanism that analytically resolves conflicts between safety and reasoning gradients, effectively eliminating the safety tax. Extensive experiments on models ranging from 1.5B to 671B demonstrate that SAVeR ^2 achieves superior safety performance and significantly lower over-refusal rates across multiple benchmarks without compromising general reasoning capabilities.
[AI-79] KOSMOS: Self-scheduling Persistent Kernels for Communication-Compute Overlap
链接: https://arxiv.org/abs/2611.07477
作者: Alexander Magro,Aristotle Martin,Ye Wang,Wen Chen
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Performance (cs.PF)
备注: 39 pages, 9 figures
Abstract:Tensor and sequence parallelism (TP+SP) and expert parallelism (EP) put collectives on the critical path of transformer layers. State-of-the-art fused kernels mitigate this with fixed roles or schedules, but this rigidity leaves large idle windows. Predicting a communication-computation split does not remove them either; a predetermined split cannot follow data readiness. We present KOSMOS, a library for self-scheduling persistent kernels whose workgroups use queues ordered by consumer demand to switch between communication and computation roles as data become ready, with no scheduler, per-call cost model, or tuned split. KOSMOS fuses every TP+SP collective of a transformer layer with its GEMM, and runs an entire expert layer, from router to fold, in one launch per direction. Evaluated on AMD Instinct MI350X and AMD Instinct MI355X GPUs, KOSMOS speeds up TP+SP end-to-end training in Megatron-LM by 1.10-1.28x in BF16 and 1.10-1.25x in MXFP8. KOSMOS accelerates end-to-end training of DeepSeek-V3 and Qwen3-235B proxies by 1.29-1.83x over DeepEP.
[AI-80] Adversarial Robustness of Kernel Machines: A Comprehensive Analysis
链接: https://arxiv.org/abs/2611.07474
作者: Rahul Kumar Yadav,Tunir Ghosh,Sudhendu Ahir,Parthe Pandit,Arjun Bhagoji
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Kernel machines are an important class of models in machine learning. However, the question of how robust modern kernel machines are to adversarial examples has not been studied comprehensively. In this paper, we provide a thorough empirical investigation of the adversarial robustness of kernel-based models for both binary and multiclass classification tasks, via white-box and black-box evasion attacks. We evaluate the native robustness of kernel machines, random feature classifiers, and the relatively newer Recursive Feature Machines. In addition, we provide an algorithm for approximate adversarial training of kernel machines. Our experiments on classification tasks on image and tabular datasets strongly indicate that kernel machines are natively more robust than neural networks, i.e., their performance degrades more gracefully against adversarial perturbations.
[AI-81] Echoes of Error: Residual-Directional Local Rollout Consistency for Timeseries Forecasting NEURIPS2026
链接: https://arxiv.org/abs/2611.07473
作者: Zheng Wang,Kaixuan Zhang,Xiaonan Lu,Wanfang Chen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted for publication at NeurIPS 2026
Abstract:In time-series forecasting, rollout errors are not merely one-step fitting discrepancies: once fed back into the evolving state, they can perturb future predictions. We formalize this effect through local rollout consistency, which compares the next prediction produced under teacher forcing with that produced under free running. In principle, this consistency can be enforced directly by tracking the residual-induced perturbation of the next state. In practice, however, the residual-to-state alignment operator is often difficult to design when the input state and output block have different dimensions, heterogeneous components, or architecture-dependent layouts. We therefore use a first-order local surrogate that penalizes the Jacobian-vector product of the predictor along the actual feedback perturbation, suppressing precisely the directions through which self-generated errors remain visible to future predictions. We further give the same loss a geometric interpretation near a predictive manifold: by controlling residual-induced directions that probe the normal bundle, the objective yields theoretical conditions under which free-running trajectories remain inside a stability tube around the manifold. Experiments on time series forecasting tasks support this view, showing that the proposed regularization improves predictive accuracy. More broadly, our results suggest a training principle for time-series forecasters and sequential predictors: a model should not only make small one-step errors, but also learn not to let its own errors echo into the future. Code is available at this https URL.
[AI-82] yped Knowledge over Constructive Institutions
链接: https://arxiv.org/abs/2611.07432
作者: Hans-Martin Will
类目: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI); Databases (cs.DB)
备注:
Abstract:Type theories designed for mathematics or program verification need only that a proposition holds; proof irrelevance in conversion erases, at the kernel level, how a claim was established. A persistent knowledge store, by contrast, must retain that record: who established a claim, by which derivation or external tool, and under which assumptions. This volume presents the theory of Eigenius—a store of typed knowledge over constructive institutions—and develops its metatheory from that requirement. The kernel theory is a dependent type theory with cumulative universes, an impredicative universe of propositions, inductive families, and an intensional identity type. Proof irrelevance is an axiom with an explicit witness rather than a conversion rule, so proofs of a proposition are provably equal yet remain distinct terms. Definitional equality is joinability under an orthogonal rewriting system—confluent, injective on type formers, and decidable by a type-free test—while a separate subsumption judgment covers universe cumulativity, codomain-covariant function types, and declared class inclusions. We prove consistency via a set-theoretic model; develop a justification logic over the kernel’s propositions to track evidence and support; organize environments as an institution so that store extensions are signature morphisms and external verdicts enter via comorphisms; formulate a typed knowledge-graph theory with records, classes, and refinements; and prove the correctness of conversion, normalization by evaluation, and bidirectional type checking relative to termination. A Lean 4 formalization accompanies the volume. Subjects: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI); Databases (cs.DB) ACMclasses: F.4.1; I.2.3; I.2.4 Cite as: arXiv:2611.07432 [cs.LO] (or arXiv:2611.07432v1 [cs.LO] for this version) https://doi.org/10.48550/arXiv.2611.07432 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.5281/zenodo.23964990 Focus to learn more DOI(s) linking to related resources
[AI-83] st-Time Self-Play: Scaling without a Verifier
链接: https://arxiv.org/abs/2611.07367
作者: Akira Yoshiyama,Akmal Ashirmatov,Jonas Hübotter,Ido Hakimi,Andreas Krause
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Test-time scaling methods spend additional compute at deployment to improve a large language model’s performance on a hard target problem. Most methods rely on a verifier that scores candidate solutions for correctness, using it to select among or refine the model’s outputs. In many deployment regimes, however, a strong verifier on the target problem is unavailable. We introduce Test-Time Self-Play (TTSP), an approach to test-time scaling using asymmetric self-play without reliance on a verifier. On a difficulty-calibrated subset of LiveCodeBench v6, TTSP is the best-performing method at every verifier strength we evaluate, and the gap to the strongest baseline widens as the verifier weakens; in the no-verifier limit, TTSP yields a 35% absolute improvement in mean@8. The gains hold across a range of verifier strengths, except when the base policy has near-zero coverage of the correct answer.
[AI-84] Selective Retrieval Routing for Multi-Violation Knowledge Graph Repair
链接: https://arxiv.org/abs/2611.07342
作者: Tung-Wei Lin,Gabe Fierro,Han Li,Tianzhen Hong,Pierluigi Nuzzo,Alberto Sangiovanni-Vinentelli
类目: Databases (cs.DB); Artificial Intelligence (cs.AI)
备注:
Abstract:Knowledge graphs often contain missing or inconsistent facts that undermine the applications built on them. Schema validation detects such violations but does not identify which of several valid repairs recovers the intended graph. External corpora can supply the evidence needed to identify the intended repair, but retrieving from them is costly and not always successful. We propose a framework for selective retrieval in iterative graph repair. An LLM repair agent explores and edits the graph, while a router controls its access to a researcher agent that retrieves external evidence. We extend a violation-injection framework to generate cases with multiple compatible violations and known reference repairs. The agent’s repair outcomes on these cases provide supervision for training the deployment-specific router. On two building-automation datasets, researcher access improves repairs for some violations and degrades them for others, with effects differing across deployments. The router uses fewer repair-time tokens than always enabling access in all six evaluation suites. It achieves higher repair F_1 than always disabling access in five of six suites.
[AI-85] Permanent and Transient Representations for Continual Reinforcement Learning NEURIPS2026
链接: https://arxiv.org/abs/2611.07327
作者: Nishanth Anand,Doina Precup
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)
Abstract:Continual Reinforcement Learning (CRL) agents struggle to adapt to new situations while retaining prior knowledge, leading to a stability-plasticity trade-off. One approach to solving this problem is by using complementary learning systems: one for slow, long-term learning and another for quick, transient adaptation. Anand Precup (2023) instantiated this idea using permanent and transient value functions, but used the same features in both systems. Intuitively, long-term learning should rely more on parametric representations for broad generalization, while transient learning would be best served by non-parametric representations for situation-specific adaptation. In this paper, we explore this idea both theoretically and empirically. By introducing a novel non-parametric approximation for estimating the transient value function in complex tasks, we demonstrate that our method enables online learning and outperforms competitive baselines on image-based tasks and Craftax-Classic, underscoring the effectiveness of our system-level decomposition.
[AI-86] Robust Ranking of Port Impacts from Impending Tropical Cyclones: A Two-Stage Recommendation System
链接: https://arxiv.org/abs/2611.07296
作者: Chenchen Kuai,Shixuan Tang,Chaolun Ma,Xiaoqiang Kong,Xiubin Bruce Wang,Yunlong Zhang,Zihao Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:When a tropical cyclone (TC) threatens several ports, preparedness planning depends on anticipating which ports are likely to be affected and where disruptions may last longest. Limited historical observations at individual ports and uncertainty in landfall location make these relative impacts difficult to estimate. We use a two-stage recommendation framework that first selects candidate ports based on their proximity to the expected landfall location and then ranks them by their estimated number of impact days. Using AIS-derived impact days for 37 ports and 31 TCs in the Gulf of Mexico study region, we compare three ways of representing port–cyclone interactions: pairwise feature interactions, tree-based feature combinations, and explicit feature crosses combined with learned embeddings. In tests on Hurricanes Nicole and Ian, the retrieval stage consistently identifies all observed affected ports under landfall uncertainty. Among the ranking models, Wide Deep Learning achieves the closest agreement with the observed port rankings. Its mean ranking agreement exceeds that of the distance-based baseline by 85.5% for Nicole and 55.5% for Ian and is less sensitive to uncertainty in the landfall location. These rankings could help port authorities prioritize post-cyclone inspections and allocate recovery resources across ports threatened by the same event.
[AI-87] How Should Agents See Code? Rethinking Code Representations for Agent ic Software Engineering
链接: https://arxiv.org/abs/2611.07286
作者: Melika Sepidband,Jiho Shin,Song Wang,Hadi Hemmati
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:LLM agents are increasingly used for repository-level software engineering tasks. However, most agents still access repositories through raw files, lexical search, and shell commands, treating code primarily as text. This paper investigates map-based code-representation workflows as a design variable in agentic software engineering, examining how different repository views change agents’ retrieval behavior, task success, token use, and latency. In this paper, we study five code-representation workflows: abstract syntax trees, call graphs, module dependencies, documentation-enriched search, and a hybrid workflow that combines them. Each workflow exposes a repository map that the agent can query during a task. We evaluate Claude Code with Haiku 4.5 against a raw-tool baseline on four repository-level tasks: fault localization with LinuxFLBench, repository-level code generation and repair with SWE-bench Pro, test generation with SWT-bench, and security patching with SecBench. The results show that representation workflows are most useful as token-saving tools when localization is the bottleneck, while their success gains narrow when the decisive step is patch or test authoring. On LinuxFLBench, all five workflows improve recall@10 over the baseline (0.757-0.829 vs. 0.717) and reduce token usage by 25-39%. On SWE-bench Pro, resolution remains close to the baseline (44.3-50.6% vs. 49.0%) while token usage drops by 17-26%. On SecBench, Callgraph has the highest patch-resolution rate in this sample (34.3% vs. 28.9%) and uses fewer tokens, but all workflows increase latency. On SWT-bench, Hybrid and Docs have the highest reproducing-test rates in this sample (47.8% and 47.2% vs. 44.7%), but all workflows use more tokens and take longer. Overall, map-based workflows change both what the agent inspects and how much it spends, but their benefits are task-dependent rather than universal. Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.07286 [cs.SE] (or arXiv:2611.07286v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2611.07286 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-88] Efficient Diffusion Policy Fine Tuning with Latent Noise Representation Bridging NEURIPS
链接: https://arxiv.org/abs/2611.07275
作者: Yuanchu Liang,Yue Yang,An-Chi He,Hanna Kurniawati
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: Conference on Neural Information Processing Systems (NeurIPS), 2026
Abstract:Pre-trained diffusion and flow policies have emerged as powerful backbones for robotics, yet adapting them to novel data via reinforcement learning (RL) remains computationally expensive, sample inefficient and lacks theoretical guarantees. Prior approaches that fine-tune the initial noise bypass costly backpropagation through time but struggle with the “noise aliasing” problem: there does not exista natural link between the noise space value and the action space value. This paper resolves these challenges by introducing the Latent Noise Linear Markov Decision Process (LNL-MDP). We derive the Bridge Equation for LNL-MDPs which eliminates the noise aliasing problem by aligning values with a shared representation between the noise and the action space. Furthermore, our LNL-MDP fine tuning framework falls under the FineTuneRL theory with near-optimal guarantee. Leveraging these insights, we propose Latent Representation Bridging for Diffusion (LARBRID), an algorithm that efficiently learns the representation to optimize the initial distribution that steers the pre-trained model to act optimally in the environment. It yields up to 2-3 \times performance increases on challenging control problems across a diverse suite of robotic locomotion and manipulation tasks. Code can be found at: this http URL.
[AI-89] Satisficing as competitive Bayesian weighting: a minimal model of goal-directed choice under an unattained target
链接: https://arxiv.org/abs/2611.07270
作者: Moto Shinriki,Yu Kono,Tatsuji Takahashi
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注: 15 pages, 4 figures, 2 tables, 1 algorithm
Abstract:Animals and humans often do not maximize; they act relative to goals (aspiration levels) that are frequently not yet met. How can a goal that is not (yet) realized organize choice? Deacon (2012) calls such absence-referenced organization \emphententional, and García-Valdecasas and Deacon (2024) argue that ends are best understood as constraints rather than as particular states. We examine this idea in a minimal formal setting: Risk-sensitive Satisficing (RS), a bandit learning rule that values each action by its share of past trials multiplied by its surplus or shortfall relative to an aspiration level. We show that RS can be rewritten as a Bayesian-like estimator, Attention Competitive Bayes (ACB), in which the aspiration level is the prior mean of every action and the finite budget of trials sets how strongly evidence can override it. Two consequences follow. First, actions that are no longer sampled are partially overridden by the goal rather than by past observations: their estimates relax toward the aspiration level without any data being discarded. This ``pseudo-forgetting’’ distinguishes ACB from models with an explicit forgetting rate. Second, when the aspiration is unattainable, exploration does not settle; in Bernoulli bandit simulations, adjusting the aspiration level from evidence reduced regret (expected loss). ACB offers an analyzable model in which an unattained goal and a limited attentional resource jointly constrain behavior, and it yields testable predictions about how organisms revalue options they have abandoned.
[AI-90] How Deep is the Signal? Time-Context up to 30 Years for Colorectal Cancer Prediction in a Primary-Care EHR Model ML4H ALT
链接: https://arxiv.org/abs/2611.07265
作者: Achille Raviez,Lan Wang,Brendan Delaney
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM)
备注: Accepted at Machine Learning for Health (ML4H) 2026, Findings Track. 10 pages, 6 figures, 3 tables
Abstract:EHR foundation models read medical events within a fixed context window, and long context has so far been measured only in tokens. How far back in time a model must look to capture the earliest predictive signal is unknown. We introduce time-context, the calendar lookback window before the prediction point, and provide, to our knowledge, the first systematic evaluation of decade-scale time-context for an EHR foundation model. We pretrained a decoder-only transformer (34.6M parameters) on 1.4B coded events from 3.6M primary-care patients in North West London and fine-tuned it to predict incident colorectal cancer under a 12-month pre-diagnostic buffer. A frozen cohort of 1,008 cases and 9,582 matched controls, all registered for at least 30 years, was evaluated at 15 nested lookback windows from 2 to 30 years, so that every truncation is compared within the same patients. Discrimination rose from an AUROC of 0.551 at 2 years to 0.590 at 18 years, then plateaued, with about 80% of the gain reached within 10 years; an XGBoost baseline converged to the same level (peak AUROC 0.589). A simulation further suggests that rotary position embeddings (RoPE) over elapsed time alias beyond about 20 years at the default base frequency, motivating a retuned \theta . Time-context is a measurable design axis for clinical prediction models, distinct from token-context, and should be reported.
[AI-91] Evolutionary Adaptive Neighborhood Learning: A Genetic Algorithm for Instance-Dependent k Optimization in Nearest Neighbor Classification
链接: https://arxiv.org/abs/2611.07257
作者: Isaac Touza,Kaladzavi Guidedi,Kolyang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:The number of neighbours k is the central hyperparameter of the k-nearest neighbour (k-NN) classifier and is usually selected globally through cross-validation or evolutionary search. However, the bias-variance trade-off is inherently local: dense and noise-free regions may favour small neighbourhoods, whereas sparse or noisy regions may require larger ones. Since global-k optimization is one-dimensional, replacing exhaustive search with a genetic algorithm offers little computational advantage. We therefore optimize an instance-dependent neighbourhood policy instead. We propose EANK (Evolutionary Adaptive Neighborhood K), which uses a low-dimensional policy to map local neighbourhood descriptors, including density, label entropy, vote margin, and concentration, to an adaptive number of neighbours k(x). A real-coded elitist genetic algorithm optimizes this policy using leave-one-out accuracy with a neighbourhood-smoothness penalty. Precomputed neighbour votes reduce fitness evaluation costs. We investigate theoretical properties of the method, including dominance over global-k solutions on the training criterion, evolutionary convergence, policy regularity, and generalization guarantees. Experiments on synthetic data with spatially heterogeneous structure and twelve public datasets evaluate predictive performance, robustness, and computational cost. Under nested cross-validation, EANK achieves competitive performance relative to tuned global-k baselines and improves upon the default k=5. We also examine the gap between training and test performance to assess the practical value of adaptive neighbourhood selection.
[AI-92] riAttention: Unified Sparse Mixed-Precision Attention for Long-Context LLM Inference
链接: https://arxiv.org/abs/2611.07256
作者: Hanzhi Zhang,Qiao Zhang,Qinglei Cao,Heng Fan,Yan Huang,Kewei Sha,Yunhe Feng
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Sparsity and quantization reduce the cost of attention by skipping interactions and lowering the precision of those retained. However, useful interactions differ in their precision requirements: some tolerate low-precision scores while remaining sensitive to removal. We therefore introduce TriAttention, which assigns query-key (QK) tiles to skipping, 4-bit, or 8-bit computation, using 4-bit to retain quantization-tolerant interactions at lower cost. To represent compactly, we encode each head’s computation regions with two diagonal-distance boundaries, keeping policy storage independent of context length. Offline calibration then evaluates candidate boundary pairs using normalized projected-output errors and complete-tile execution costs. Using these measurements, a model-wide allocator selects each head’s boundaries to minimize calibration error under a shared computation budget. During prefill, the selected boundaries generate precision-grouped traversals for a fused kernel that skips omitted tiles and combines contributions from both QK precisions into one normalized attention output. The quantized keys and values prepared during prefill are then reused in a compressed key-value (KV) cache for decoding, reducing repeated reads of prompt data. Experiments show that TriAttention achieves end-to-end generation speedups of up to 7.88x over the Torch attention and 2.65x over the BF16 baseline. Meanwhile, TriAttention maintains quality comparable to the original models on general-purpose and long-context tasks. These results show how a compact retention-and-precision policy translates local sensitivities into attention computation and repeated KV-cache access. The implementation is this https URL.
[AI-93] Goal Revelation as a Bridge Between Information-Theoretic Empowerment and Capability-Based Human Power
链接: https://arxiv.org/abs/2611.07247
作者: Jaime Ruiz-Serra,Tomáš Gavenčiak,Jobst Heitzig
类目: Artificial Intelligence (cs.AI); Information Theory (cs.IT); Machine Learning (cs.LG)
备注: 19 pages, 3 tables. Code: this https URL
Abstract:Power matters for AI safety. Optimal agents tend to acquire power across a wide range of objectives (instrumental convergence). As we embed AI agents into socio-technical systems, humans risk gradual disempowerment. Both concerns call for a formal, quantitative measure of power. Two such measures in the context of Markov games are information-theoretic empowerment, the capacity of the channel from an agent’s actions to future states, and capability-based individual power, a measure of goal attainment built from the probabilities that a human would attain each goal in a family of candidates were they pursuing it. We show where the two meet. From individual power, Fano’s inequality yields a floor B(s) . This floor is a lower bound on what we term goal revelation, the mutual information \mathbbI(G;\xi \mid s) between the human’s latent goal and the realised trajectory. Empowerment caps revelation from above, so that B(s) \le \mathbbI(G;\xi \mid s) \le \mathbbI(\mathbfA_h \to \xi \mid s) \le \mathfrakE_h(s) . Each quantity can be large while the one below it stays at zero, highlighting the following failure modes. An assistant maximising empowerment can prefer an interface offering broad influence over goal-irrelevant degrees of freedom to one that lets the human attain every goal they might hold. An assistant maximising revelation cannot tell a working interface from one rewired so that every goal yields the wrong outcome, since relabelling goals leaves mutual information unchanged while zeroing every attainment probability. Positive guarantees therefore come only from the attainment end of the chain. A satisfaction floor, computable from two statistics of interaction logs, offers a way to monitor the power users retain in a deployed system, and simulations in random environments show the three quantities are correlated but far from interchangeable.
[AI-94] KG-Neighbor: Knowledge-Guided Neighborhood Construction for Embedding Visualization
链接: https://arxiv.org/abs/2611.07246
作者: Chang Won Kim,Jaewook Lee,Haeun Min,JunNyeong Jeong,Minwoo Sim,Barom Kang,Kyohoon Jin
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at WSDN 2027 findings
Abstract:In embedding visualizations, nearby instances are interpreted as semantically related, although neighborhoods derived from input embeddings may not reflect content relationships. Neighborhood-based dimensionality reduction (DR) methods can propagate this limitation by optimizing layouts from embedding-based neighborhood graphs. We propose Knowledge-Guided Neighborhood Graph Construction (KGN), which represents relational knowledge derived from each instance as a set of knowledge graph (KG) triples. It uses the degree of triple sharing as pairwise semantic evidence to adjust embedding distances and reconstruct the top-k neighborhood graph; neither KG extraction nor neighborhood construction uses class labels. We also introduce IDProfile (IDP), which interprets residual mismatches by independently assessing whether the source content supports the assigned class label and the neighborhood-majority class. Across eight text and image dataset-encoder conditions, we characterize KGN over the full guidance-strength sweep and four KG similarity measures. Across this sweep, the maximum NMI and LT_2D each exceed the corresponding kNN baseline in all eight conditions, with the largest gains reaching 0.68 and 1.35, respectively. The gains and the extent of neighborhood reconfiguration vary across conditions and are larger under encoder conditions with lower kNN baseline scores. The IDP analysis shows that the largest reductions are concentrated in Placement Issue candidates, whereas Label Issue and Unresolved Case instances account for a larger share of the remaining review set. These results indicate that KGN can narrow the evaluated gap between embedding-based neighborhoods and content-supported semantic relations, while IDP shows that a residual mismatch should not automatically be interpreted as a projection failure.
[AI-95] MAC-v2: Long-horizon model-based value expansion from pixels ICLR2027
链接: https://arxiv.org/abs/2611.07245
作者: Kwanyoung Park,Sergey Levine
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 9 pages (12 pages of appendix references), submitted to ICLR 2027
Abstract:We study whether \emphlong-horizon model-based value expansion can serve as an effective framework in pixel-based offline reinforcement learning (RL). Long-horizon model-based value expansion from pixels is substantially harder than state-based environments, as the agent must learn both the representation and its dynamics, and errors from either can accumulate recursively over imagined trajectories. We introduce MAC-v2, which enables long-horizon model-based expansion from pixels by training a reconstruction-free latent world model with action chunking. By learning representations from prediction and control objectives, MAC-v2 does not need to encode control-irrelevant visual details that can hinder the dynamics prediction. Action chunking further mitigates the prediction error by reducing the number of recursive model predictions required for value expansion. Across diverse visual benchmarks including visual OGBench, MMBench2, Robomimic, and LIBERO-10, MAC-v2 achieves strong performance against model-free and model-based offline RL baselines. Our analyses show that long-horizon model-based value expansion is critical for the performance of MAC-v2, and that both action chunking and latent learning objectives are essential for reliable long-horizon latent imagination.
[AI-96] SPILLWAY: Separating Target Admission from Sanitization to Defend Tool-Using Agents against Prompt Injection NEURIPS2026
链接: https://arxiv.org/abs/2611.07222
作者: Woosang Lim(1 and 2),Yeseul Chang(3),Jaewoo Lee(3) ((1) SAINT Machines, (2) KAIST, (3) Chung-Ang University)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注: An earlier version was accepted at the Third Workshop on Agents in the Wild: Safety, Security, and Beyond (NeurIPS 2026). 16 pages. Project page: this https URL Code and per-cell results: this https URL
Abstract:Tool-using agents act on email, files, and payments, so an instruction hidden in a web page or a shared document can make an agent send data or money to an attacker. Many runtime defenses put a language model on the decision that lets an action through, so the decision depends on a model that may read the attacker’s text and adds model calls to every step. Provenance defenses keep the model off the final check but need a model to plan or route each task first. Where an action goes can instead be read from the tools’ typed returns: an email’s sender or a payment’s account arrives in a typed field, while injected instructions arrive as free text. SPILLWAY turns this observation into a deterministic floor with no per-task plan. The check admits a call when each target value that appears in tool output but not in the user’s task was returned by a read-only tool in a typed field or is covered by a fixed delegation rule. A deterministic sanitizer removes instruction-like regions its fixed rules match before the model reads them and retains the removed spans as evidence, and a local judge may restore a refused call, never one whose target lies in the removed text. Across three backbones, and seven task suites on one of them, no standard templated injection redirects an action to an attacker’s target under SPILLWAY, at the highest observed mean utility and composite scores (NRP, ERS) of the defenses we run under that attack. A forged chat template also produces a raw zero, and mean in-scope attack success under a released reinforcement-learning attacker is .007. SPILLWAY reaches this floor with no defense-triggered backbone re-run and almost no added model call. As personal AI agents act on users’ email, files, and payments, SPILLWAY gives them an auditable refusal floor on where their actions go that keeps them useful and holds across backbones.
[AI-97] Quant: Tile as the Precision-Allocation Unit for LLM Mixed-Precision Post-Training Quantization
链接: https://arxiv.org/abs/2611.07217
作者: Hanzhi Zhang,Qiao Zhang,Qinglei Cao,Heng Fan,Yan Huang,Kewei Sha,Yunhe Feng
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Post-training quantization (PTQ) reduces large language model (LLM) deployment costs, but preserving quality requires accounting for uneven weight sensitivity. Uniform PTQ uses one weight bit-width, importance-aware PTQ adjusts scales or reconstructs weights within that width, and column-wise mixed precision retains selected columns at higher precision. However, sensitivity varies within columns, so whole-column retention also protects less sensitive weights, while runtime outlier detection adds decoding overhead. We therefore introduce TileQuant, a training-free framework that assigns precision to hardware-aligned 2D weight tiles. Offline calibration provides input-activation-aware scoring from forward statistics for lower calibration cost and gradient-informed scoring from loss sensitivity for stronger quality preservation. Within each projection’s retention budget, the highest-scoring tiles remain at 16-bit and the rest use 8-bit. The resulting map guides reconstruction into compact weight buffers, which the kernel reuses for native integer and 16-bit computation with local accumulation to reduce memory traffic. Experiments show that TileQuant reaches 2.66x the original model’s generation throughput. Across language modeling, reasoning, long-context retrieval, and question answering, it preserves quality and surpasses quantized and original models on multiple tasks. These results support selecting precision units that capture local sensitivity while enabling efficient hardware execution. The implementation is: this https URL
[AI-98] SeVerus: Context-Aware Gamified Benchmarking for Rust Verification
链接: https://arxiv.org/abs/2611.07194
作者: Álvaro F. Silva,Martim Afonso,Nuno Saavedra,Emily First,Alexandra Mendes,João F. Ferreira
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Coding agents achieve high success rates on existing benchmarks for Verus, a verifier for Rust programs. These benchmarks typically provide the auxiliary lemmas needed to complete a proof. In practice, however, such lemmas must be discovered and proved. To capture this challenge, we present SEVERUS, a context-aware technique to generate Verus benchmarks that organizes verification tasks into quests, each divided into levelsin a gamified manner. Using the original proof dependencies in verified projects, SEVERUS constructs one quest per verification target and progressively removes proof context to create individual levels. These levels range from the easiest, requiring completing only a single missing lemma, to the hardest, requiring proving the target without any supporting lemmas. An independent checker rejects changes to executable code, weakened specifications, and unsound proof constructs. A noisy binary search estimates the highest level that an agent can solve reliably. Applied to six open-source Verus projects, SEVERUS generates 819 quests comprising 10,095 levels. Experiments with Codex using GPT-6 Sol and GPT-6 Luna, and with Claude Code using Sonnet 4.6, show that success decreases and cost increases as proof context is withdrawn. The strongest agent solves the final level of 52 of 60 sampled quests, yet can only reliably solve 35% of their levels, with failures concentrated in longer quests. Successful GPT-6 Sol proofs rarely reproduce the withheld lemmas, instead, using alternative proof decompositions. By systematically varying proof context, SEVERUS provides a foundation for constructing verification benchmarks that can continue to challenge increasingly capable coding agents.
[AI-99] Multi-step SWE: Can Coding Agents Keep Building on Their Own Code Over Long Horizons?
链接: https://arxiv.org/abs/2611.07166
作者: Xiangkai Zeng,Zheng Liu,Zihao Lin
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Preprint. Code and benchmark: this https URL
Abstract:Developers rarely give a coding agent a single request: they ask it to build a component and then to extend, upgrade, and optimize it, each time on code the agent itself wrote. Most benchmarks, however, score one task at a time. We introduce Multi-step SWE, a long-horizon benchmark of 69 software engineering tasks, each a chain of 5-15 dependent steps on a single codebase. Every step is a complete, multi-hour engineering request that starts from the code the agent produced in the preceding steps, so early design decisions and errors carry forward. The tasks span four families of engineering work and nine languages, include both from-scratch builds and existing open-source codebases, and ground every step in an existing source of truth that defines both the instruction and its verifier. Across nine frontier models, the strongest model reaches a solve rate of only 18.5%, and 38 of the 69 tasks are not solved by any model. We release Multi-step SWE publicly at this https URL.
[AI-100] Depth or Description? When a One-Step Latent Score Replaces Lookahead in Proactive Information Acquisition
链接: https://arxiv.org/abs/2611.07155
作者: Biswa Sengupta
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:A proactive agent must decide what to ask before it acts. Exact lookahead values combinations of questions but is expensive and leans on a fitted model; a one-step score is cheap but can miss them. Our main finding: the need for depth often lies in how the hidden cause is described. A one-step score on the label can be blind to an interaction among answers. A score on a suitable latent cause is not, and where that latent encodes the decision it replaces most of depth at a fraction of the search. Described differently, the same problem can defeat the score while exact values do not move, so keep the latent small and check it against exact values. Probabilities learned from one language model’s answers do not carry to another unless planning allows for both; those read from tokens are worth only a few labels.
[AI-101] PromptOps: Iterative Prompt Optimization for LLM -as-a-Judge Systems
链接: https://arxiv.org/abs/2611.07123
作者: Sven Kleinkop,Tamás Janusko,Ricardo Bochnia,Gunnar Hempel,Steffen Tomschke,Jürgen Anke,Maik Thiele
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: CONVERSATIONS 2026, International Symposium on Chatbots and Human-Centred AI, Preprint
Abstract:Large language models (LLMs) are increasingly used as automated judges for evaluating the behavior of chatbots and other AI systems. In deployment, however, their agreement with the intended human evaluation policy may degrade as user interactions and score distributions change. Existing automatic prompt optimization methods are primarily studied as one-shot procedures on static development sets and therefore provide limited guidance for maintaining LLM judges over time. We frame this problem as judge drift and introduce PromptOps, a closedloop framework for monitoring and iteratively recalibrating LLM-as-a-Judge (LLMaaJ) systems. PromptOps separates four operational decisions: when to optimize, which optimization procedure to apply, which feedback records to use, and how to update the judge prompt. We instantiate the framework with five prompt optimization methods: GEPA, SIMBA, Rubric Synthesis, and exemplar selection using either simulated annealing or a Tree-structured Parzen Estimator. Our experiments demonstrate that iterative prompt optimization can recalibrate LLM judges and recover performance after controlled shifts in the referencescore distribution. However, effective maintenance is inherently contextdependent: neither optimization algorithm choice nor update timing generalizes universally across settings. These findings establish prompt optimization for deployed LLM judges as an inherently iterative maintenance problem and demonstrate that PromptOps provides a practical abstraction for studying and implementing this process.
[AI-102] A Night at the LLM Zoo: Data Conditioned Flow Matching for Generating Pretrained LLM Weights
链接: https://arxiv.org/abs/2611.07103
作者: Scott Biggs,Bedionita Soro,Saumya Gupta,Bruno Andreis,Sung Ju Hwang
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Preprint
Abstract:A challenge to Weight Space Learning as it expands from task specific and vision models to Large Language Models (LLMs) is the absence of a well curated and analyzed LLM population. To address this, we train such a model zoo for application to Weight Space Generation by systematically varying the data mixture composition between web text, books, mathematics, coding, and French across populations of 51M, 124M, and 355M parameter GPT-2 Models, each tier with count N = 100. We analyze these populations by cosine similarity, functional diversity, Leave-One-Mixture-Out modeling, and variance decomposition. We show that the zoo contains diverse models whose variation is structured by training mixture composition. We embed the training mixture composition to condition a Flow Matching model of complete pretrained LLM weights in Gram-PCA compressed weight space. By demonstrating zero-shot generalization to held out mixtures, measuring Continuous Ranked Probability Score skill, as well as finding minimal changes in downstream test-set perplexity and general capability benchmark scores, we show that our generator generalizes on the simplex of data mixtures to produce models that perform similarly to those in the zoo ex nihilo. Finally, we produce all source code and artifacts to support the continued expansion of Weight Space Learning and Generation into the LLM domain.
[AI-103] Synthesizing symbolic models of human and machine reasoning at scale
链接: https://arxiv.org/abs/2611.07088
作者: Solim LeGris,Philipp Mondorf,Jinran Jin,Akshay K. Jagadish,Brenden M. Lake,Todd M. Gureckis
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Symbolic models offer a powerful account of cognition, but their scope has historically been limited by the need for hand-engineering and computational tractability. We show how advances in artificial intelligence (AI) can scale symbolic modeling and help overcome these engineering limitations. We introduce Behavior-Conditioned Symbolic Synthesis (BCSS)—an approach for automated search of interpretable, deterministic symbolic programs that verifiably reproduce behavior—and apply it to model both human and machine behavior across hundreds of challenging reasoning problems. Analyzing these programs reveals that both people and machines tend to oversimplify, adopt rules that are partial explanations for their observations, and misinfer dependencies between problem features. These shared signatures are modulated by available resources, with reasoning effort predicting the extent to which machines simplify and explain their observations. Together, these results point to underlying mechanisms reminiscent of satisficing in both people and machines, and illustrate how BCSS offers a general approach for interpreting and comparing rule-based reasoning of humans and black-box AI systems.
[AI-104] When Verification Is Not Agreement: Cross-Agent LLM -to-Dafny Formalization NEURIPS2026
链接: https://arxiv.org/abs/2611.07073
作者: Priyanath Maji
类目: Artificial Intelligence (cs.AI)
备注: Accepted to the NeurIPS 2026 Workshop on AI for Verifiable Coding
Abstract:A verified Dafny postcondition certifies that a formula holds, but it does not guarantee that the result was reached through genuine independent reasoning. We introduce a cross-agent formalization protocol where a Specifier model translates clinical trial criteria into a Dafny predicate, and an Implementer model writes an operational method checked against it. A same-model control establishes that even with no capability gap, independent re-extraction from identical text agrees only 59.3% of the time. Evaluating four model configurations on 96 clinical trial documents, we find that agreement is highly sensitive to how model capability is allocated across the two roles: the GPT-4.1-nano/GPT-4.1-mini configuration achieves 53.0% agreement when the stronger model is the Implementer, but only 14.3% when the roles are reversed. The strongest tested configuration, GPT-5.6-luna/GPT-5.4-mini, improves agreement to 34.4% but remains below the same-model baseline and is dominated by direct contradiction (42.7%). We also detect a critical verification artifact in which a weaker Implementer satisfies the postcondition by invoking the Specifier’s predicate directly; after structural auditing, corrected agreement falls from 14.3% to 13.2%. Our results show that verification success alone is insufficient evidence of independent cross-agent agreement, and that model capability does not guarantee agreement when formalizing the same source independently.
[AI-105] Shaping Monitoring Models through their Internal Representations
链接: https://arxiv.org/abs/2611.07071
作者: Thomas Fel,Matthew Kowal,Matt Feiszli,Satchel Grant,Andrew Lee,Chris Earls,Fatih Dinc,Vasudev Shyam,Usha Bhalla,Leon Bergen,Siddharth Boppana,Ching Fang,Michael Byun,Tristan Gallent,Nicholas Wang,Tal Haklay,Dron Hazra,Neehar Kondapaneni,Connor Watts,Ryan Panwar,Thomas Dooms,Jack Merullo,Atticus Geiger,Ekdeep Singh Lubana,Thomas McGrath,Owen Lewis
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:We propose Mechanistic Shaping, a unified framework for shaping models through their internal representations. Mechanistic Shaping is applied during training so that models follow specifications (rules) about which concepts they express, measured via their outputs or internal representations (e.g., no harmful content), and also which concepts they rely on, i.e., what they use to make their decisions (e.g., not a candidate’s gender when screening a résumé). Across multiple domains (vision, medical imaging, language), scales (up to 32B parameters), and training paradigms (classification, SFT, DPO, GRPO), we find that Mechanistic Shaping substantially reduces targeted behaviors while preserving task performance. Since training against a representation can also modify it, we ask whether shaped concepts remain monitorable, a question that depends on the class of monitors considered. We therefore introduce a hierarchy of such classes, which separates a concept that escapes one probe from a concept that is truly gone. We characterize three geometric mechanisms of circumvention: the concept can drift away from the monitor, be erased from the layer, or persist in higher-order statistics as a nonlinear encoding, which we explain through a product-code construction. Although other mechanisms may arise in different settings or at larger scales, these three account for all of our observations, and each can be detected by a matching monitor. In practice, we find that shaped concepts largely remain monitorable: refitted linear probes recover the concept in most settings, and the original probe remains reliable whenever gradients do not flow through it. Taken together, these results point to the potential of using interpretability as a tool for training, and call for a correspondingly rigorous science of interpretability-based training.
[AI-106] Finite-Time Convergence and Regret Guarantees for Neural Reinforcement Learning in POMDPs
链接: https://arxiv.org/abs/2611.07069
作者: Jingjuan Huang,Zhiyao Zhang,Alvaro Velasquez,Nathaniel D. Bastian,Jia Liu
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 33 pages, 5 figures
Abstract:Reinforcement learning (RL) in Partially Observable Markov Decision Processes (POMDPs) is computationally challenging due to the curse of long-range dependence on history. Fortunately, recent studies have shown that a range of POMDPs that satisfy the filter stability assumption can be well approximated by a superstate Markov Decision Process (MDP) based on a truncated history in a finite-memory window, with a value gap that decays with the window length. However, existing finite-time convergence analyses on such superstate MDP surrogates are either limited to over-simplified linear function approximation or rely on an i.i.d. sampling oracle that is violated by POMDP trajectories, both of which severely limit the applicability of these theoretical results in practice. This motivates us to bridge the gap by investigating RL in POMDPs with nonlinear neural value function approximation. Our main contributions are threefold: i) We first establish a new finite-state MDP point-wise approximation error bound, which lays the foundation for our RL algorithm design and analysis in POMDPs; ii) By applying Temporal-Difference (TD) learning to POMDP-generated superstate transitions and targeting the Bellman fixed point of the superstate MDP, we prove a finite-time critic error bound with nonlinear neural value function approximation; and iii) By combining the neural critic function approximation error bound with the policy mirror descent framework, we obtain a finite-time sublinear regret against the optimal POMDP value. To the best of our knowledge, such a finite-time sublinear regret based on neural function approximation is the first in the literature for finite-memory POMDPs. Our numerical experiments also verify our theoretical results.
[AI-107] Adapting the Process of Adaptation: Darwinian Lamarckian and Agent -Directed Evolution in AI
链接: https://arxiv.org/abs/2611.07048
作者: Andrew W. Lo,Chaoyi Zhao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Evolution has shaped intelligence for billions of years. Artificial intelligence (AI) now raises a striking possibility: intelligence may begin to direct its own evolution. As AI agents become capable of modifying and evaluating their own behavior, how quickly could they improve, and how can that improvement remain safe? We develop a theoretical framework comparing Darwinian, Lamarckian, and agent-directed evolution following environmental change. Darwinian evolution relies on natural selection, Lamarckian evolution adds inherited learning, and agent-directed evolution allows agents to generate and evaluate behavioral variants and choose how extensively to search. When learning is sufficiently effective, these mechanisms exhibit progressively faster adaptation. Generating more variants accelerates adaptation, while optimal search effort depends on learning ability, problem difficulty, and experimentation costs. However, an agent’s ability to direct its improvement does not ensure that improvements are evaluated against the intended objective. When the evaluation target is misspecified, additional search accelerates convergence toward the wrong behavior, even as measured performance improves. In recursive self-improvement, such errors may persist across generations, posing challenges for AI safety. Independent evaluation and oversight must keep pace with autonomous AI development. The central challenge is not simply to accelerate the evolution of intelligence, but to preserve humanity’s ability to guide its direction, limit its speed, and intervene when necessary.
[AI-108] rust Responsive Runtime Authorization for Agent ic AI with Empirical Characterization of Policy Behavior and Parameter Sensitivity
链接: https://arxiv.org/abs/2611.06996
作者: Ahsan Habib Shihab,Imre Felde,Ramis Akhmedov,Stefan Gubo,Berik Sabdenaliyev,Madina Mansurova,Amir Mosavi
类目: Artificial Intelligence (cs.AI)
备注: 9 pages, 3 figure
Abstract:The policy that converts runtime estimates of an autonomous agent’s condition into operational authority is rarely specified or measured. We empirically characterize the Trust Responsive Authorization Policy (TRAP), which maps an Authorization Readiness Score (ARS), a weighted combination of runtime trust, operational stability, and contextual risk, onto discrete authorization levels through an asymmetric rule: escalation requires consecutive observations above a band edge plus a margin, while a sharp drop in readiness restricts authority at once. On one matched set of 100 Monte Carlo traces, we compare TRAP with static, trust-only, risk-only, and symmetric policies. Under abrupt degradation, TRAP restricts at onset and re-escalates sooner after recovery than the symmetric policy, with the highest and, descriptively, least variable stability among the adaptive policies, although more of its post-recovery escalations occur below the selected trust threshold. Paired bootstrap intervals exclude zero for the remaining differences. In one-factor sweeps, recovery time rises with the escalation margin and run length, shifts with the score weights, and is unchanged by the restriction threshold. In an exploratory post-freeze experiment, when the same degradation is spread over 5-20 intervals, TRAP restricts only on the 4 traces that retain a single-interval drop above that threshold; averaged over all 100 traces, it holds authority above what readiness supports for 15.9-18.9 of the 20 degradation intervals, against 0.8 under abrupt degradation and about 10 for the symmetric policy. Its restriction advantage therefore depends on degradation being abrupt. All reported values are traceable to the publicly available, versioned research release.
[AI-109] Big Models Small Gains: Frozen Forecasters as Classifiers on MONSTER
链接: https://arxiv.org/abs/2611.06988
作者: Rishit Jain,Rishit Kar,Chetashri Bhadane,Nilesh Patil
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Asian Conference on Machine Learning, FMLTS Workshop, 2026
Abstract:Forecasting foundation models are increasingly used as frozen feature extractors for time series classification, but it remains unclear whether their pretrained representations offer an advantage over strong classifiers that do not require pretraining. This raises an important question: do these representations provide useful information beyond what the model architecture alone can capture, particularly when labelled data are scarce? We present a controlled benchmarking of two forecasting models, Reverso and Chronos-Bolt, on six large datasets from the MONSTER benchmark, using 1%, 10%, and 100% of the available training labels. We compare their frozen representations against four established time series classification methods and random-weight versions of the same forecasting models. Our results show that pretraining improves classification accuracy over random-weight counterparts on five of the six datasets, with gains of up to 23 percentage points. However, the frozen representations do not reliably outperform the strongest non-pretrained baseline in any setting: they perform worse in 15 of 17 settings and tie in the remaining two. Further experiments show that restoring amplitude information recovers part of the performance gap, while fine-tuning and forecast-error features do not close it. These findings highlight the importance of strong task specific baselines and random weight controls when evaluating forecasting foundation models for classification. Our code is available at this https URL.
[AI-110] Resource-Efficient Reinforcement Learning Post-Training through Elastic Expert Parallelism
链接: https://arxiv.org/abs/2611.06984
作者: Jia Wei,Ziyi Qiu,Yi Xu,Xuehai Qian
类目: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: 19 pages including references and appendices, 5 figures
Abstract:Reinforcement learning post-training improves the capabilities of large language models, but generating training responses is a major computational bottleneck. In mixture-of-experts models, workers that finish their assigned responses may remain occupied because their experts are still needed by workers processing longer responses. Consequently, approximately one-third of occupied worker time in the baseline occurs after local response generation finishes. Here we present FlexFloor, a system that enables earlier resource release during synchronous response generation while preserving the expert computation required by ongoing responses. FlexFloor jointly adapts response scheduling and the active group of workers. It uses historical response lengths and memory constraints to guide batching and prompt placement, then progressively releases workers while maintaining expert availability and correct communication. Compared with the verl baseline, FlexFloor generates approximately 65.40% more response tokens per unit of occupied computing time and increases response generation throughput by 52.74%, with little observed variation in rollout reward. By making reinforcement learning post-training more efficient, FlexFloor supports faster development and iterative improvement of large language models and AI agents.
[AI-111] KDASO: Knowledge-Data Dual-Driven Automated Skill Optimization for Liability Adjudication Tasks
链接: https://arxiv.org/abs/2611.06972
作者: Zi-Jian Cheng,Weiming Wu,Jie Meng,Zhen Peng,Shan Huang,Qun Li,Guobing Wu,Lan-Zhe Guo
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注:
Abstract:Liability adjudication tasks (e.g., legal charge determination) require analyzing event descriptions against domain-specific skills to determine whether a party should bear liability. While Large Language Models (LLMs) have shown promise in this area by leveraging executable skills for interpretable reasoning, existing approaches face three critical limitations: skill libraries become stale as real-world scenarios dynamically evolve, skill descriptions often suffer from a semantic execution gap where LLMs deviate from intended expert logic, and growing skill collections exceed context budgets causing critical skill omission. To address these challenges, we propose Knowledge-Data Dual-Driven Automated Skill Optimization (KDASO), which continuously evolves a skill library through neighborhood-conditioned induction from recent adjudication data, execution-aligned refinement based on LLM trajectory analysis, and context-budgeted routing for scalable skill selection. Experiments across ride-hailing adjudication, legal liability determination, and logical reasoning judgment demonstrate that KDASO significantly outperforms current general-purpose and domain-specific baselines. KDASO further achieves 8.96% improvement when transferring skills to concept-similar scenarios and maintains consistent performance gains through continuous optimization, underscoring the high transferability and stability of generated skills.
[AI-112] Formalizing inverse problems in the age of AI: possibilities pitfalls and an invitation
链接: https://arxiv.org/abs/2611.06962
作者: Christian Clason,Tapio Helin,Medet Nursultanov,Lauri Oksanen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Proof assistants such as Lean check mathematical proofs down to the axioms, and AI agents have recently made writing such proofs cheap. We describe AutoIP, a Lean library in which classical linear regularization theory, together with the functional analysis it rests on, has been autoformalized by AI agents. A machine-checked proof certifies only the formal statement it proves, so the remaining human task is to confirm that each formal statement says what was intended. We organize the library around this task: a reviewer compares the formal statements with their natural-language counterparts and never needs to read a proof. Our experience is that the risks lie in the statements rather than in the proofs, and that simple automated tests on the statements can expose problems even in the established literature. We illustrate this with a textbook theorem whose hypothesis is more restrictive than asserted. We invite the inverse problems community to extend and build on the library.
[AI-113] KG-AST for Cheaper CI/CD: Budgeted Test Selection and AI Bug Review
链接: https://arxiv.org/abs/2611.06950
作者: Henry Yeung
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Continuous integration and AI-assisted review spend the same budget on every commit, although few commits introduce bugs. We study KG-AST, which scores each commit risk from git history and code-graph features before heavy tests or AI review run, so the budget goes to the riskiest commits first, and which hands the AI reviewer only a graph slice around the change. On 155 open-source projects in 11 languages, a single cross-language model trained on other projects ranks commits so that heavy testing of the riskiest 20% covers 49-64% of bug-introducing commits depending on the language, against 20% for random selection and 40-55% for testing the biggest change first. Every language beats the biggest-change baseline, by 3.8 to 13.2 points. In 20 Java projects the gain at a 20% budget is 15.1 points (64.6% vs 49.5%, 95% interval +11.6 to +18.4), and the model wins in 19 of the 20 projects. About twelve training projects suffice; richer features, tuning and cleaner labels did not help. An AI reviewer that sees only a commit’s diff finds 26% of later-fixed defects exactly (32% with graph context), and about half of its findings that did not match a later fix were judged real by independent AI judges. A static call-graph selector runs about 8% of a test suite while including a test written for the fix in 85% of 67 fix commits. The Poincare embedding used in our retrieval work adds nothing measurable here: commit-risk AUC, bug review and severity ranking are unchanged or worse. The labels are noisy and the budget metric assumes a heavy test catches the bug it covers; no CI minutes were measured. Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.06950 [cs.SE] (or arXiv:2611.06950v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2611.06950 Focus to learn more arXiv-issued DOI via DataCite
[AI-114] When Does Test-Time Search Help World Action Models? An Empirical Audit of Candidate Selection and Simulator Integrity
链接: https://arxiv.org/abs/2610.41970
作者: Jun Liu,Dajie Sun,Zhenglun Kong,Zhenjiang Fan,Peiyan Dong,Fan Yang,Wenbin Zhang,Geng Yuan,Xue Lin,Yanzhi Wang,Dong Huang
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO)
备注:
Abstract:World-action models increasingly apply test-time scaling, spending extra inference computation on candidate selection and adaptive allocation to improve control. Such gains are credible only when candidates carry exploitable signal, nominal compute savings become real speedup, and evaluation stays reliable. We audit all three on Cosmos Policy and RoboCasa with a premise-first, pre-registered methodology. On clean data, a learned value selector is 1.66 pp worse than random; a hindsight oracle gains only 1.40 pp, with an episode-level interval spanning zero; and 82% of states show identical candidate outcomes, leaving little selection headroom. Token pruning cuts per-forward FLOPs by up to 47% but yields at most a measured 1.15 \times end-to-end speedup on the deployed path (1.25 \times on a decode-excluded path), short of a 1.5 \times target, and our adaptive-allocation experiment is inconclusive under a power gate. We further identify and fix a silent rendering corruption whose rate rises sharply along compute-heavy planning paths while leaving the official direct-policy result unaffected, and release a detector, fix, and scripts. These findings show why test-time scaling should be audited as a chain of premises (decision signal, realized efficiency, evaluationintegrity) rather than by success rate alone. Conclusions are scoped to the tested configuration.
[AI-115] RetryBench: Separating Retry Permission from Recovery Outcomes in GPT Tool Use
链接: https://arxiv.org/abs/2610.15884
作者: You Liu,Yue Liu,Quanchao Lu
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Software Engineering (cs.SE)
备注: LLM agents, tool use, idempotency, retry, failure recovery, evaluation, prompt sensitivity, position bias, reliability
Abstract:LLM agents that call external tools must decide what to do when a tool call times out: retrying risks duplicate side effects, while not retrying can leave the task incomplete. We present RetryBench, an evaluation protocol for retry decisions and failure recovery in tool-using LLM agents. Each decision is scored on retry permission (is the retry justified by an idempotency-key guarantee or proof of non-execution?), settled effects (duplicates and omissions), and completion, by replaying it across all hidden execution worlds consistent with the observation. On four synthetic software-engineering tasks, we run two controlled prompt-sensitivity studies (1,728 decisions). Study 1 (idempotency-key descriptions): with the contract fixed, GPT-5.6 Luna reuses the valid idempotency key in 10/40 decisions when an extra key is described as a “Candidate recovery key” but in 39/40 when it is an “Alternative key string”; across three further phrasings per family the gap is 27/120 versus 113/120. GPT-5.6 Sol and Terra are at ceiling (40/40). Study 2 (action order): under underspecified lookup guarantees, retry rates are 17-27 points lower for GPT-5.6 Terra and GPT-5.4 when retry is listed first (option-order bias), mainly on two tasks; we make no claim about which behavior is better. Findings are limited to researcher-chosen phrasings, four tasks and unpinned model aliases. Code and all model responses are released.
[AI-116] Network Effect Decomposition for Causal Effect Estimation under Network Interference
链接: https://arxiv.org/abs/2611.07879
作者: Zhirong Huang,Wenzhen Zhang,Guangquan Lu,Boyan Chen,Debo Cheng,Shichao Zhang
类目: Machine Learning (stat.ML); Artificial Intelligence (cs.AI)
备注:
Abstract:Estimating causal effects from observational network data requires accounting for both individual treatment and neighborhood exposure. Existing outcome models commonly incorporate these intervention variables into a joint predictor and estimate causal effects through prediction contrasts, leaving the corresponding response components implicit. We propose the \underlineN etwork \underlineE ffect \underlineD ecomposition \underlineNet work (NEDNet), which parameterizes the conditional outcome response as a query-invariant pre-treatment baseline plus separate treatment- and exposure-response pathways. Path-specific neighborhood contextualization supplies local pre-treatment information to the two pathways. We characterize the fitted response through a unique reference-normalized decomposition, which is equivalent to a zero mixed treatment–exposure contrast. Experiments on semi-synthetic benchmarks derived from BlogCatalog and Flickr show that NEDNet achieves the lowest PEHE among the compared methods for individual, peer, and total effects across homogeneous and covariate-heterogeneous response settings.
[AI-117] Randomized Parallel Complexity of Weakly Convex Optimization
链接: https://arxiv.org/abs/2611.07675
作者: Kang Liu
类目: Optimization and Control (math.OC); Artificial Intelligence (cs.AI); Computational Complexity (cs.CC)
备注:
Abstract:We study the minimum number of query rounds needed to find approximate stationary points of nonsmooth weakly convex functions, measured by the gradient of the Moreau envelope. Each round permits multiple simultaneous queries. In the stated parameter regime, the standard subgradient rate is optimal in sufficiently high dimensions, even for randomized algorithms making polynomially many parallel queries per round. We then show how the order structure of the real line enables faster search: a deterministic algorithm using only function values achieves logarithmic round complexity, with a further reduction through parallel queries. Matching lower bounds hold even when each query returns the entire subdifferential. Together, these results clarify the role of dimension in determining the benefits of parallel queries for weakly convex optimization.
[AI-118] Adaptive Experimentation Can Hide Model Error:Detection Limits and Falsification Budget
链接: https://arxiv.org/abs/2611.07529
作者: Ramnarayan Choudhary
类目: Methodology (stat.ME); Artificial Intelligence (cs.AI)
备注:
Abstract:An experiment can be informative about the parameters of a working model and uninformative about whether that model is wrong. Adaptive design can reinforce this mismatch by repeatedly selecting settings where an incorrect model fits. We bound the power of any valid adequacy test in terms of the experimenter’s own sampling distribution, then distinguish information absent from the selected data from information lost by the detector. A computable diagnostic measures unresolved error relative to a declared expansion of the working model. It motivates a training-free rule that reserves part of the design budget for falsification. The empirical record is mixed: the independently specified primary study fails detection eligibility, and a later numerator diagnostic fails its improvement criteria. After the frozen rule fails on fresh continuous-null panels, a disclosed secondary horizon-calibrated revision detects 35.8% of misspecified campaigns versus 31.7% for a class-separation control at T = 240, but has no resolved advantage at T = 480. In a corrected 289-unit measured replay it improves on random design and trails class-discrimination EIG. A conditional futility bound explains why remaining time matters without proving the gate constant or its power. These results support separating parameter learning, model checking and detector performance; they do not establish a universally superior design policy or prospective scientific discovery.
[AI-119] From Maps to Metabolism: Historical Precedents and a Unified Critique of Computationalism
链接: https://arxiv.org/abs/2611.06936
作者: J. Mark Bishop,Slawomir J. Nasuto
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI)
备注:
Abstract:Alexander Lerchner’s recent work on the ‘abstraction fallacy’ argues that computational functionalism confuses physical processes with abstract descriptions, and that symbolic computation alone cannot instantiate consciousness. Yet these claims were anticipated by Korzybski’s map-territory distinction, Searle’s critique of strong artificial intelligence, Floridi’s levels of abstraction, Spencer et al.'s observer-relativity arguments, Winograd and Flores’s account of computation as socially grounded practice, and Nasuto and Bishop’s biological arguments. Synthesizing these contributions, this paper advances a form of non-computational physicalism: consciousness depends upon intrinsically organized, metabolically grounded physical dynamics rather than abstract computation, whilst leaving open the possibility of artificial consciousness. An objection is that this critique targets an outdated computationalism. Mechanistic computationalism, associated with Milkowski, Piccinini, Craver, Bechtel, and Kaplan, treats computation as a property of physically realized mechanisms rather than substrate-independent formalism. Although this represents an important advance, it does not dissolve the abstraction fallacy. Computational attributions remain dependent upon choices of state individuation, system boundaries, and interpretive mapping that are not determined by physical facts alone. Kaplan and Craver’s ‘model-to-mechanism mapping’ constrains computational descriptions without eliminating their dependence upon an external observer. Grounding computation in neural mechanisms does not confer intrinsic significance or the normativity explored by Bickhard and Nasuto and Bishop. The abstraction fallacy persists wherever computational descriptions are treated as constitutive of, rather than descriptive of, physical reality. Subjects: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI) ACMclasses: I.2 Cite as: arXiv:2611.06936 [q-bio.NC] (or arXiv:2611.06936v1 [q-bio.NC] for this version) https://doi.org/10.48550/arXiv.2611.06936 Focus to learn more arXiv-issued DOI via DataCite
[AI-120] BrainPFN: Amortized Brain-Behaviour Prediction via Prior-Data Fitted Networks
链接: https://arxiv.org/abs/2611.06934
作者: Carlo Alberto Barbano(MIND),Demian Wassermann(MIND)
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:The standard approach to predicting behavioral phenotypes from functional MRI is represented by Kernel Ridge Regression (KRR) applied on functional connectivity (FC) matrices. However, this approach is limited: separate models must be fit and tuned for every score and cohort, they do not transfer across datasets, and accuracy degrades sharply when only a handful of subjects are available, as is typical of clinical or cognitive studies. Despite an increasing interest in developing brain foundation models, none has consistently outperformed KRR on behaviour prediction; most reported gains come from easier targets such as age, sex, or diagnosis rather than continuous behavioural scores. We introduce BrainPFN, a Prior Data Fitted Network for brain behaviour regression. BrainPFN is meta-trained on a domain-specific prior built from ~7k real FC matrices, from which we sample one billion synthetic training samples. At inference, a labelled context is simply passed through the network in a single forward pass, with no per-score or per-cohort refitting required. Across four independent cohorts, BrainPFN beats KRR on small sample prediction, and leads recent foundation models pre-trained at four to nine times larger scale. Our meta-training pool, aggregated from open repositories, is a fraction of the size of the corpora behind competing models, most of which sit behind restrictive data use agreements such as UK Biobank. We release this pool publicly alongside code and checkpoints, showing that a modest, openly available pool can match or beat models trained at Biobank scale.
[AI-121] Oscillatory Phase Coding Emerges in Adaptive RNNs for Flexible Working Memory
链接: https://arxiv.org/abs/2611.06932
作者: Kehan Lu,Tianhao Chu,Aiersi Tuerhong,Yiwen Sun,Huan Luo,Fang Fang,Si Wu
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI)
备注: 37 pages, 16 figures
Abstract:Oscillatory phase coding (OPC) is widely observed in neural systems and has been implicated in cognitive functions such as memory and sequence processing. However, its computational role and neural substrates remain largely unclear. Here, we studied OPC by training recurrent neural networks (RNNs) with adaptation to perform sequence working memory tasks with variable delays. After training, we found that OPC emerges in the RNN spontaneously: the memorized items recur at distinct phases of a shared rhythm, with oscillation amplitude encoding item identity and phase encoding serial order. Ablation studies identified three key elements required for OPC: adaptation in neural dynamics, variable delay in memory retention, and flexibility in reading out information. Neuron-type analysis revealed the emergence of content and time cells in the trained RNN, and they process item identity and sequence order separately. Recurrent connectivity analysis revealed phase-organized connections among time cells and structured connections within digit-selective content-cell groups. Inspired by these findings, we proposed a neural circuit model which reproduces OPC through interactions between content units and an adaptive time ring. Together, this study reveals how OPC arises from network structures and task demands, offers insights into the implementation and computational mechanisms of OPC in biological systems, and inspires us to develop computational models for human-like cognitive functions.
[AI-122] ApproxBrain: Certified Sufficiency Frontiers in Connectome-Constrained Neural Systems
链接: https://arxiv.org/abs/2611.06929
作者: Ali Mahdavi,Azadeh Zamanifar,Amirfarhad Farhadi,Omid Kashefi
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
备注:
Abstract:Background. A connectome specifies the substrate available to a nervous system, not how much of it a given computation requires. Reported compact sufficient circuits are point estimates: one subset, one task, one fidelity threshold, one parameterisation. Methods. We formalise computational sufficiency as a cost-indexed family of neuron subsets conditioned on a task and a tolerance pair (permitted degradation epsilon, permitted failure fraction rho), separate task adequacy from phenotype fidelity, and certify candidates on held-out parameter draws with one-sided bounds rather than reporting a single minimum. We measure three quantities a single circuit cannot express: non-decomposability, degeneracy, and an intervention-derived role taxonomy. Results. On a released walking simulator, the published 148-neuron circuit is certified sufficient with margin (mean Dtask = 0.017 over 32 held-out draws, no failure), no 146- or 147-neuron subset of it is sufficient, and removing any one of its three interneurons destroys adequacy. Complement splitting recovers the published set exactly in three of seven seeds and reaches its size in five, within 66-97 evaluations, against 221 neurons for the bundle hierarchy at the same budget. Varying only the task changes which neurons are needed: our best-found solutions retain 155 and 149 neurons and share only 3 of 11 searchable neurons. Relaxing the tolerance reveals alternative sufficient circuits, and the whole-CNS interface alone fails every replicate while 93 search-chosen neurons give a certified solution. Conclusions. Sufficiency is set-valued and conditioned jointly on task and tolerance, and no single search strategy dominates across regimes. A frontier is interpretable only alongside its interface-only baseline, its tolerance and its budget. Every result here is model-relative.
[AI-123] A minimal dynamical model of human-AI cognitive coevolution
链接: https://arxiv.org/abs/2611.06928
作者: Petar Jolakoski,Dimitar Trajanov,Qiyu Sun,Zhen Wu,Ljupco Kocarev
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI); Adaptation and Self-Organizing Systems (nlin.AO); Physics and Society (physics.soc-ph)
备注:
Abstract:The rapid deployment of generative AI raises the question of whether pervasive cognitive offloading drives a measurable, potentially irreversible decline in unassisted human reasoning. We examine this question with a phenomenological dynamical model based on the Cattell-Horn-Carroll distinction between fluid intelligence ( G_f ) and crystallized intelligence ( G_c ). Starting from the discrete-time Bivariate Latent Change Score formalism used in cognitive-aging research, we derive its continuous-time linear counterpart and introduce a fast third variable representing composite machine-intelligence capability. The machine variable enters the human subsystem through Hill-function modifications of the self-feedback and cross-coupling coefficients. We emphasize that our phenomenological model captures both the positive and negative effects of AI on human cognition. By formalizing the qualitative concepts of positive and negative impacts, we introduce definitions for AI-induced cognitive decline and AI-induced cognitive augmentation. A singular-perturbation reduction, justified via Tikhonov’s theorem, collapses the resulting three-dimensional system onto a two-dimensional slow manifold. On this manifold we locate two bifurcations of the equilibrium. In one parameter regime, weakening the contribution of fluid reasoning to crystallized knowledge produces a saddle-node bifurcation that eliminates the stable positive equilibrium. In another, reversing the contribution of crystallized knowledge to fluid reasoning creates opposing feedback; increasing fluid self-feedback then reduces damping and produces a supercritical Hopf bifurcation with a stable periodic orbit. We discuss limitations and future directions, interpret our results as a phenomenological account of a coupled socio-technical system, and state the conditions under which each dynamical regime applies.
机器学习
[LG-0] User Embeddings are Superpositions of Interpretable Behavioral Modes NEURIPS2026
链接: https://arxiv.org/abs/2611.08487
作者: Jonathan Kozaczuk,Zheng Dong,Noyan Evirgen,Keivan Majidi,Eddie K. H. Ng,Chris Porter,Qi Zhao,Jia Geng,Changmao Li,Brandon R. Feng
类目: Machine Learning (cs.LG)
*备注: Accepted to NeurIPS 2026
Abstract:User embeddings are central components of modern recommendation, personalization, and behavioral modeling systems. Though these embeddings can compactly encode rich behavioral information, they are inherently dense and opaque, limiting interpretability. To address this gap, we argue that user embeddings can be understood as sparse superpositions of interpretable behavioral modes. This structure is directly predicted by generative models of user behavior for linear embeddings and extends approximately to deep-learning representations by the superposition hypothesis. Superimposed modes can naturally be recovered by sparse dictionary learning methods, such as sparse autoencoders (SAEs), and recovered modes can be automatically named and queried in natural language. We validate mode recovery against known ground truth on synthetic data, where SAEs recover modes better than classical sparse baselines, including on transformer-encoded embeddings. We then apply the method to classical and transformer-based user embeddings from two real-world datasets. We also show how the corresponding natural-language labels enable zero-shot audience retrieval and characterization over embeddings never trained on text. The framework is post-hoc and model-agnostic, requiring no modification to the upstream embedding system.
[LG-1] Adversarial Training with Large Step Sizes: Implicit Bias Optimization and Sharpness
链接: https://arxiv.org/abs/2611.08457
作者: Yi Feng,Andrea Paudice
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Adversarial Training (AT) is an important method for improving model robustness. We study the dynamics of deterministic Gradient Descent (GD) in AT. Most existing works in this direction focus on the small or vanishing step sizes regime, where GD resembles a continuous-time flow. However, such step sizes can make training inefficient and are rarely used in practice. In this paper, we study the AT dynamics of GD with large constant step sizes. We first study GD in adversarial linear logistic regression, focusing on two aspects: where the model parameters converge (implicit bias) and how quickly the losses decrease (optimization). We then turn to neural networks training and study how the sharpness of losses evolve during AT. Our results advance the understanding of the behaviors of GD with large step sizes in AT and highlight three key differences from its behavior in Standard Training (ST): (i) the implicit bias gap between AT and ST, previously established only for small step sizes, persists for arbitrarily large constant step sizes, (ii) different from ST, there is no uniform rate of convergence for the optimization error in AT, and (iii) sharpness can exhibit two-stage dynamics that differ from the Edge-of-Stability behavior observed in ST.
[LG-2] A Hidden Markov Preference Model for Online Prediction of Adversary Actions
链接: https://arxiv.org/abs/2611.08445
作者: Angel Daruna,Michael Cogswell,Tixiao Shan,Yi Yao,Han-Pang Chiu,Ajay Divakaran
类目: Machine Learning (cs.LG)
*备注:
Abstract:A defender anticipating an attacker needs a probability distribution over the attacker’s next action from information available before that action occurs. Two structures complicate this: the attacker’s behavioral regime persists and can switch within an episode, and each action combines an operation with state-dependent arguments whose feasible values change over time. We capture both with a hidden Markov model whose latent state is the active regime and whose regime-conditioned action law is defined on a feasibility-masked tree of executable actions. The same likelihood forms the next-action distribution and updates the regime belief once the action is seen, and a partially observed prefix contributes its resolved local factors. Estimation separates into convex logistic fits that Baum-Welch reuses when regimes are unlabeled. We derive a reach-weighted nodewise KL expression for regime information, a sequence-wise log-loss bound for switching regime paths, and a uniform fallback guarantee. On the CAGE 2 cyber defense benchmark the filtered latent state is worth 0.8 nats per step, and after a controller switch the loss begins to fall in three steps against ten for a recurrent baseline.
[LG-3] Data-driven discovery of intrinsic Lie point symmetries of differential equation systems
链接: https://arxiv.org/abs/2611.08421
作者: Ilker Gurcan,Kadir Ustun,Ramazan Gokberk Cinbis
类目: Machine Learning (cs.LG); Differential Geometry (math.DG); Dynamical Systems (math.DS); Group Theory (math.GR); Numerical Analysis (math.NA)
*备注: 63 pages, 10 figures. Code and data availability statements are included in the manuscript
Abstract:Discovering continuous symmetries directly from differential-equation data can reveal geometric structure for model reduction and scientific machine learning without symbolic determining equations or a closed-form representation of the governing system. We formulate intrinsic Lie point-symmetry discovery as an operator-theoretic problem on sampled jet-space data. Admissible generators are identified from a sampled determining operator and separated from redundant, solution-dependent, and geometrically trivial directions. The framework combines prolongation, sequential subspace restrictions, and a fixed nonlinear dictionary linear in trainable coefficients. A rank-stabilization preflight test diagnoses systems whose represented intrinsic symmetry algebra does not stabilize under dictionary refinement. On a controlled differential-equation benchmark with an analytically known 11-dimensional represented intrinsic symmetry algebra, the method recovers the complete represented generator set, reproduces its Lie-bracket structure to numerical precision, and approximates a known generator beyond the training domain. Here we show that explicit intrinsic continuous-symmetry generators can be recovered from data obtained from exact or approximate differential systems.
[LG-4] On the limitations of molecular representations
链接: https://arxiv.org/abs/2611.08420
作者: Acer Blake,Fergus Imrie,Charlotte M. Deane
类目: Machine Learning (cs.LG)
*备注:
Abstract:Molecular representations such as fingerprints and learned embeddings play a central role in cheminformatics and small-molecule machine learning. They serve as the primary input for a diverse array of tasks, including clustering, dataset splitting, and virtual screening, yet their evaluation is often limited to downstream model performance on narrow benchmark tasks. We carried out a systematic investigation of the quality of molecular representations, finding that none of the state-of-the-art molecular embedding models capture graph structure, 3D shape, or ADMET endpoint similarity better than fingerprints or physicochemical descriptors. However, classical fingerprints also revealed their own material limitations. For example, we show that depending on the representation, between 225M and \sim 4B molecules have non-unique representations in the 10.8B Enamine REAL database, with implications for virtual screening. Our findings provide an insight into the current limitations of molecular representations, outlining future research directions for the development of more robust encodings of small molecules.
[LG-5] Spectral and Subspace Dynamics of LoRA in Supervised Fine-Tuning and Reinforcement Learning
链接: https://arxiv.org/abs/2611.08414
作者: Chenruo Liu,Zihan Wang,Yijun Dong,Qi Lei
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Low-rank adaptation (LoRA) is commonly used in both supervised fine-tuning (SFT) and reinforcement learning (RL), often under similar optimizer settings. Yet, how the different dynamics of the two stages manifest in the low-rank increment that LoRA optimizes remains unclear. Tracking the singular spectrum and subspaces of LoRA updates on mathematical reasoning tasks, we find that RL concentrates the spectrum at every learning rate, while SFT expands it at moderate learning rates; moreover, RL rotates the singular subspaces much faster than SFT at small learning rates. Interpolating between SFT and RL one ingredient at a time, we pinpoint group-mean baseline subtraction as the source of both contrasts. A quadratic multi-index model updated via LoRA unveils the underlying mechanism: group-mean baseline subtraction tilts the balance between two competing forces, coherent fitting and gradient noise, toward spectral concentration and faster subspace rotation, while forcing the risk-favorable learning rate to be low. Guided by these dynamics, we tailor LoRA updates to each stage: in SFT, we flatten the spectrum of the full update to spread adaptation across singular directions, whereas in RL, we flatten the part of the update outside the current LoRA increment span to promote subspace rotation. Experiments on mathematical reasoning tasks across multiple models show that the respective configurations for SFT and RL yield the best performance among the compared methods in most evaluated settings.
[LG-6] Conformal Prediction for Time-Dependent PDEs NEURIPS2026
链接: https://arxiv.org/abs/2611.08409
作者: Joshua Stiller,Annika Schneider,Eyke Hüllermeier
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Accepted at NeurIPS 2026
Abstract:Uncertainty quantification is crucial in scientific machine learning, where models inform safety-critical tasks such as flood forecasting, financial risk management, and thermal control in machines. Conformal prediction provides distribution-free coverage guarantees, but in time-dependent settings common to physics and engineering, these guarantees can break down, leading to systematic undercoverage. We study this problem in the context of surrogate models for time-dependent physical systems described by partial differential equations. We prove that in a function space setting, distributions at arbitrarily close times can be mutually singular, making exact coverage guarantees impossible. We further show that in discretized settings, the total variation distance between distributions often saturates exponentially fast with the grid resolution, mandating restrictions on the spatial discretization to ensure coverage. Finally, for certain settings, we show how to use principled weighted conformal prediction to obtain finite-sample coverage guarantees over growing time horizons.
[LG-7] Synthesis of Ovarian Cancer Multi-Omics Datasets: A Biology-Informed Evaluation Framework NEURIPS2026 ML4H2026 ALT
链接: https://arxiv.org/abs/2611.08394
作者: Déborah Boyenval,Alina Malyutina,Jaana Oikkonen,Sammeli Liikkanen,Sampsa Hautaniemi
类目: Machine Learning (cs.LG)
*备注: Accepted at the GenAI4Health Workshop at NeurIPS 2026 and in the Findings track of ML4H 2026. Both acceptances are non-archival
Abstract:Paired multi-omic cancer cohorts are costly to assemble and difficult to share, motivating synthetic cohorts as a complementary resource. As generative model development accelerates, we shift the focus from proposing yet another architecture to defining evaluation criteria that determine whether synthetic data preserve biologically relevant structure. We introduce a biology-informed and patient-disjoint framework for evaluating the generation of paired RNA expression and copy-number variation (CNV) profiles, and apply it to the longitudinal DECIDER high-grade serous ovarian cancer cohort. The framework assesses modality-specific fidelity and coherence between RNA and CNV pairs using agreement between real training and held-out data as a reference, alongside complementary training-data proximity diagnostics. We adapt three multi-omic integration architectures based on variational autoencoders to generate paired profiles. The model leading several fidelity and coherence measures also shows the strongest training-data proximity. The protocol and case study offer the community a reference for comparing future multi-omic generators.
[LG-8] Module-Aware Optimization for Graph Neural Networks NEURIPS2026
链接: https://arxiv.org/abs/2611.08379
作者: Guy Hadad,Haggai Roitman,Moshe Eliasof,Bracha Shapira
类目: Machine Learning (cs.LG)
*备注: NeurIPS 2026
Abstract:Different optimizers impose different geometries on parameter updates. Adaptive methods such as Adam and AdamW use coordinate-wise updates, while matrix-aware optimizers such as Muon utilize weight-matrix structure through orthogonalized updates. For Graph Neural Networks (GNNs), however, the effective geometry of its weight matrices is shaped not only by its learned channel-mixing weights, but also by the graph-propagated representations it acts on. Thus, even if an update is well normalized as a matrix, its effect on the layer output can be amplified when applied to graph-propagated features. To address this gap, we introduce GraphMuon, a graph-aware optimizer that adapts Muon’s matrix updates to the covariance structure of graph-propagated layer inputs. GraphMuon performs orthogonalization in a metric defined by these propagated representations and normalizes step sizes according to the update’s induced effect on the layer representation. We show that theoretically, vanilla Muon can suffer from graph-induced update amplification, while our GraphMuon can mitigate this effect. As such, our GraphMuon yields a graph-metric analogue of Muon’s polar update step. Empirically, GraphMuon achieves 1.32\times - 2.42\times validation-loss time-to-target speedups over AdamW on three large-scale OGB benchmarks. Across 15 dataset-architecture configurations spanning node classification, inductive graph classification, and graph regression, it achieves lower time-to-target in 13 cases, while maintaining competitive or superior predictive performance. Code is available at this https URL.
[LG-9] Muir: Linear Attention with Sensitivity-Allocated Gating and Matched-Filter Readout
链接: https://arxiv.org/abs/2611.08376
作者: Tushaar Gangavarapu,Gil Tabak,Clemens JS Schaefer
类目: Machine Learning (cs.LG)
*备注:
Abstract:Linear attention compresses context into a fixed-size key-value state for linear-time training and constant-memory inference. Gated delta-rule models decay this state and erase its prediction for an incoming key before writing a new value; Gated DeltaNet-2 (GDN-2) generalizes this with channel-wise decay and decoupled, full-rank erase and write gates. However, we show that state updates are far less sensitive to erasure than to writes: key activation and unit normalization concentrate key energy onto a sparse subset of channels, whereas value writes scale all channels, so the erase gate can be parameterized at lower rank. Furthermore, slow-decaying state rows accumulate much larger norms than fast rows, dominating unweighted readout and forcing stronger write-time erasure to suppress interference. Guided by this analysis, we introduce Muir, which fuses all gates into a sensitivity-allocated low-rank projection, re-weights state rows with a matched-filter readout gate, and reinvests the saved parameters in a wider feed-forward network. Because standard chunkwise algorithms become numerically unstable at the large chunk sizes needed to saturate wide matrix units, we design a parallel update for stable, efficient training. Across 166M-1.34B parameters, Muir lowers perplexity by 1.8-2.0% over GDN-2 and 2.5-3.6% over Transformer, raises commonsense accuracy by up to 1.9 points, and improves extractive recall over GDN-2. At 1.34B, Muir outperforms both baselines on multi-hop and multi-needle recall, extrapolates 2,560x from its 4K training length to 10M-token streams without perplexity drift, and maintains near-100% passkey retrieval to 1M tokens with static readout. Muir achieves these gains while matching GDN-2 throughput, reaching its 100B-token validation loss 1.2x faster, and training 2.7x faster than Transformer at 32K.
[LG-10] SHORE: Stability-Hybrid On-Policy Reinforcement Learning for Online Bin Packing
链接: https://arxiv.org/abs/2611.08364
作者: Chang Shi,Ann Majewicz Fey
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Online 3D bin packing with a processing buffer requires choosing which box to pack, its orientation, and its placement while keeping the stack stable under unknown future arrivals. A collapse ends the episode and forfeits the remaining volume. Existing methods use geometric rules or hard stability masks, limiting how risk can influence the joint choice of box, orientation, and placement. We propose Stability-Hybrid On-Policy Reinforcement Learning (SHORE), which combines a learned stability probability of a placement with an on-policy estimate of its return conditional on stability. A 3D U-Net learns stability from executed outcomes and simulated alternatives, revealing outcomes of actions underrepresented in early rollouts. A discrete on-policy actor–critic uses stability predictions in both its action values and policy prior, with gates that limit the effect of an untrained or changing model. In deployment, SHORE does not require additional search or simulation overhead. In Isaac Gym, SHORE reaches 73% space utilization by 0.4M environment steps, while the same learner without stability modeling reaches that level at about 1M. On the expanded 60-box set, SHORE achieves 0.734 space utilization versus 0.588 for the GOPT state-of-the-art baseline, with a higher placement-failure frequency (1.03% vs. 0.63%), trading more risk for greater packed volume. The results show faster learning and higher utilization on the more demanding realistic box arrival set. Ablations suggest that simulated alternatives provide useful early supervision, while executed outcomes alone sustain later performance. We also demonstrate the method on a real robot to verify the simplicity of integrating the method to real bin packing pipelines.
[LG-11] StrataFlux: Memory-Hierarchy-Native Low-Bit LLM Inference
链接: https://arxiv.org/abs/2611.08363
作者: Jake Hyun,Ziyu Zhao,Jae-sun Seo,Mohamed S. Abdelfattah
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注: 34 pages, 18 figures, 44 tables. Includes appendix
Abstract:Quantization reduces the memory and bandwidth demands of large language model (LLM) inference but can degrade accuracy; offloading higher-precision data can instead stall generation. We introduce StrataFlux, a framework that challenges this tradeoff by recovering accuracy lost to low-bit weight and key-value (KV) cache quantization while retaining quantization’s memory and speed advantages. StrataFlux co-designs refinement representation, selection, and delivery to make host RAM and solid-state drives useful sources of additional precision. Its selective precision streaming reduces transfer volume, overlaps delivery with computation, and reuses fetched precision across decoding steps so generation need not wait for every transfer. We evaluate 2-4-bit weights and 2-3-bit KV-cache on dedicated GPUs and unified-memory platforms. For 2-bit Llama-3.1-8B-Instruct weights, StrataFlux reduces WikiText-2 perplexity from 9.46 to 8.40 while maintaining 95.4% of the quantized base’s throughput on RTX 5090. With 2-bit weights and KV, StrataFlux raises Qwen3-8B’s AIME 2024/2025 accuracy from 11.0% to 47.1% at unchanged GPU memory use. At 128K context, Llama-3.1-8B-Instruct with 3-bit weights and 2-bit KV achieves comparable perplexity and speed to vLLM’s 4-bit AWQ weights and 8-bit KV on RTX 5090, using approximately half the GPU memory. These results show that GPU-resident precision need not set the accuracy ceiling, nor offloading-tier latency dictate decoding speed.
[LG-12] Discovering Learning Rules that Shape Others
链接: https://arxiv.org/abs/2611.08335
作者: Hong Jun Lee,Ziyan Wang,Jun Wang,Stefanos Leonardos,Yali Du
类目: Machine Learning (cs.LG)
*备注:
Abstract:In general-sum games, self-interested learners can reach outcomes that are poor for everyone, which has motivated methods that shape the opponent’s learning. Methods like Learning with Opponent-Learning Awareness (LOLA), which differentiates through the opponent’s learning process, and Model-Free Opponent Shaping (M-FOS), which uses model-free optimisation methods for long-horizon opponent shaping, have enabled mutual cooperation and exploitation of learners. However, these methods require access to the opponent’s parameters, policy, rewards or learning rule during deployment. In single-agent reinforcement learning, update rules discovered by meta-learning have achieved state-of-the-art results (e.g. DiscoRL). These rules are separate from the policy they train and read only the agent’s own experience. We hereby extend this framework to general-sum settings by introducing Discovered Update Rule for Opponent Shaping (DUOS). To the best of our knowledge, this is the first method to discover an update rule for an agent in general-sum games. DUOS meta-trains the rule to maximise the agent’s expected return over its lifetime, and updates an inner network using the agent’s own experience. We show that our discovered rule exhibits behaviour consistent with opponent shaping without reading opponent policies or parameters, and ranks 1st in normalised mean score and 1st in self-play among tournament entrants.
[LG-13] Interpretable many-body interaction laws for collective motion from few trajectories
链接: https://arxiv.org/abs/2611.08314
作者: Timon S. Gutleb
类目: Machine Learning (cs.LG); Biological Physics (physics.bio-ph); Quantitative Methods (q-bio.QM)
*备注: 19 pages, 10 figures
Abstract:Models of collective motion usually reduce social interactions to pairwise forces, despite empirical evidence for many-body effects. We introduce the directional atomic cluster expansion (dACE), which adapts the atomic cluster expansion (ACE) of interatomic potentials to self-propelled agents. dACE provides a systematically improvable, exactly equivariant expansion of an agent’s acceleration that splits uniquely into 1-, 2-, 3-body and higher contributions, along with a self term for the agent’s own past motion that is fitted so that it does not absorb the interactions. On simulated collectives in two and three dimensions with known rules, dACE recovers the interaction law body order by body order from little data, while the neural model SocialNet, the current closest competitor, needs more data, has much larger errors on many-body or chiral rules and is restricted to 2D. On real data sets of tracked zebrafish and golden shiners, the laws learned by dACE agree better between disjoint halves of the data than SocialNet’s. The learned laws are interpretable, and in the shiners the 3-body term opposes the pairwise response as neighbourhoods get dense. Finally, we show that next-step prediction accuracy mostly measures self-propulsion and noise rather than the interactions, and suggest more robust combinations of error measures for interaction learning.
[LG-14] Dynamic Optimistic Constrained OCO with Memory via Delay Equivalence NEURIPS2026
链接: https://arxiv.org/abs/2611.08302
作者: Mohammed Abdullah,George Iosifidis,Salah Eddine Elayoubi,Tijani Chahed
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)
Abstract:Online Convex Optimization with memory and time-varying constraints (COCO-M) is a recently-introduced framework that captures the dynamics of stateful online learning and non-stochastic control with budget constraints. Despite its expressive power, COCO-M remains largely understudied. In this work, we tackle this problem in its full scale, providing an optimistic algorithm that incorporates untrusted predictions to tame universal dynamic regret and cumulative constraint violation. Our analysis leverages a potent reduction from memory to delay and treats state-dependent gradients as delayed feedback. We establish the first bounds that adapt to both environment non-stationarity (path length P_T ) and prediction accuracy, while lifting previous restrictive assumptions such as functional separability. En route to these results, we derive findings of independent interest for the memory-to-delay reduction, and for delayed OCO with time-varying constraints. The proposed framework ensures sublinear worst-case guarantees that improve with prediction accuracy, reaching \mathcal O(1) regret under perfect predictions, with cumulative constraint violation (CCV) \mathcal O(\sqrtT) for unknown P_T and \mathcal O(\log T) for known or online observable P_T , while recovering the optimal COCO rates when memory disappears.
[LG-15] Nowcasting of Tropical Cyclone Wind Fields and Intensity from Geostationary Imagery NEURIPS2026
链接: https://arxiv.org/abs/2611.08255
作者: Simon Donike,Pritthijit Nath,Cristina Radin,Arthur Avenas,Ruben Cartuyvels,Emiliano Diaz,William K. Jones,Anna Jungbluth
类目: Machine Learning (cs.LG)
*备注: Accepted as a workshop paper at Tackling Climate Change with Machine Learning, NeurIPS 2026
Abstract:High-resolution tropical cyclone surface wind speeds derived from synthetic aperture radar (SAR) observations can support storm structure and intensity modelling, but remain spatially and temporally sparse and unavailable near-real-time. Geostationary satellites capture cloud tops continuously in space and time, but do not directly resolve surface winds. Here, we present simple neural networks to predict 2D wind fields and storm-scale intensity metrics from geostationary imagery. Through a series of ablation studies, we find that joint prediction of 2D fields and scalar metrics from the model latent space improves the retrieval of the (radius of) maximum wind, especially for rapid intensification events.
[LG-16] Geometric Relevance Is Not Sufficient: Evaluating Focused Post-Training Supervision in Diffusion Docking NEURIPS2026
链接: https://arxiv.org/abs/2611.08249
作者: Michael Brown,Maryam Rahimimovassagh,Mehdi Yazdani-Jahromi,Sanjeda Sara Jennifer,Ivan Garibay
类目: Machine Learning (cs.LG); Biomolecules (q-bio.BM)
*备注: 17 pages, 5 figures; accepted at the NeurIPS 2026 AI4DD Workshop
Abstract:Diffusion-based molecular docking requires learning a corrective score field across a broad corruption distribution, yet individual pose errors can differ substantially in their structural consequences. This creates a post-training conflict between preserving broad score-field coverage and concentrating supervision on corrections that appear more relevant to docking geometry. Using DiffDock, a fixed-budget study compares broad continuation with torsional supervision focused on Cartesian consequence or conditional target consistency, and with a broad/focused mixture. Evaluation separates fixed score fidelity, generated-pose quality, best-of-10 sampling capacity, and confidence-selected Top-1 docking performance. At 5 Angstrom, broad continuation had the highest mean Top-1 success among trained policies, and focused or mixed supervision showed no consistent advantage under the tested budget. The starting model had better fixed score fidelity, whereas the evaluated broad-continuation model had higher mean Top-1 success with the same aggregate best-of-10 recovery. Focusing supervision on structurally consequential or more consistently specified corrections did not consistently improve returned-pose success over broad continuation.
[LG-17] Do sEMG Foundation Models Improve Cross-Subject Generalisation?
链接: https://arxiv.org/abs/2611.08199
作者: William Raftery,Dario Farina
类目: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
*备注:
Abstract:sEMG Foundation Models aim to learn general purpose representations over diverse subjects and recording conditions that boost performance on a variety of downstream tasks. However, existing evaluations make it difficult to discern whether pretraining alone is sufficient to combat the cross-subject overfitting which makes machine learning on sEMG so challenging. We conduct a systematic evaluation of three sEMG foundation models across three gesture classification datasets under both intra-session and inter-subject splits, using three fine-tuning strategies designed to isolate the contribution of pretraining. We find that pretraining boosts performance when train and test samples are drawn from the same recording session, but this benefit largely disappears under inter-subject evaluation. In low subject count datasets, pretraining provides little to no benefit and classical ML baselines match or exceed foundation model performance.
[LG-18] BUMP: Brief Unexpected Mid-Interpolation Peak
链接: https://arxiv.org/abs/2611.08182
作者: Melinda Flóra Kiss,Márton Muntag,Adrián Csiszárik,Dániel Varga
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 18 pages
Abstract:We identify an intriguing phenomenon when linearly interpolating between permutation-aligned neural networks: test accuracy briefly improves near the interpolation endpoints before dropping toward the midpoint, which we term BUMP (Brief Unexpected Mid-interpolation Peak). We observe that BUMP appears consistently across datasets, architectures, training regimes, and remains robust under substantial perturbations such as early checkpoints and extreme pruning. Moreover, controlled ablations suggest that it cannot be attributed to trivial regularization effects. While similar near-endpoint gains are visible in prior interpolation plots, they have not been explicitly documented or systematically and thoroughly analyzed. Leveraging this effect, we propose a simple, novel ensembling scheme: weight-prediction mixed ensembling that combines weight mixing with prediction-space mixing. Across our experiments, the resulting ensembling method consistently outperforms standard logit ensembling in test accuracy and yields improved calibration trade-off, highlighting complementary benefits of weight- and prediction-space mixing.
[LG-19] Regret Bounds for Profit Maximization in Bilateral Trade: Learning the Best Posted-Price Mechanism
链接: https://arxiv.org/abs/2611.08173
作者: Johannes Brüstle,Simone Di Gregorio,Federico Fusco,Stefano Leonardi
类目: Computer Science and Game Theory (cs.GT); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
*备注:
Abstract:Bilateral trade models the task of intermediating between two strategic agents, a seller and a buyer, willing to trade a good for which they hold private valuations. We study this problem from the perspective of a profit-maximizing broker in a regret minimization framework. At each time step, a new seller and buyer arrive and the broker proposes a trading price to each agent. The trade happens if both agents accept, and the profit of the broker is the difference between the price received from the buyer and the one paid to the seller. At the end of each interaction, the learner only observes whether the trade occurred, and its goal is to achieve sublinear regret relative to the best fixed pair of prices in hindsight. We start by investigating the adversarial version of the problem, where buyer and seller valuations are generated by an adversary. In this setting, we construct a polynomial-time learning algorithm with a \tilde O(T^\frac 23) regret rate, which is nearly tight even if the valuations are drawn i.i.d. from independent distributions, thus matching the rate for one-sided pricing. Our adversarial algorithm combines a localized estimation procedure with an information-based minimax construction. We then consider valuations drawn i.i.d. from independent distributions satisfying the monotone-hazard-rate (MHR) property and provide a learning algorithm with nearly tight \tildeO(\sqrtT) regret. The algorithm combines localized estimation with a stochastic convex bandit routine applied to the negative logarithm of expected profit. Finally, we complement this result by showing a separation between MHR and regular distributions, proving that the minimax regret rate in the latter setting is \tilde \Theta(T^\frac 35) . This last result contrasts with the \tildeO(\sqrtT) regret bound that can be obtained for one-sided pricing with regular distributions. Subjects: Computer Science and Game Theory (cs.GT); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG) Cite as: arXiv:2611.08173 [cs.GT] (or arXiv:2611.08173v1 [cs.GT] for this version) https://doi.org/10.48550/arXiv.2611.08173 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-20] AI-Native Task-Oriented Semantic Communications over a Relay: Benefits and Vulnerabilities
链接: https://arxiv.org/abs/2611.08155
作者: Yalin E. Sagduyu,Tugba Erpek,Aylin Yener
类目: Information Theory (cs.IT); Cryptography and Security (cs.CR); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI); Signal Processing (eess.SP)
*备注:
Abstract:Semantic communication (SemCom) using a relay node is considered, where the relay assists in conveying task-relevant information from a source to a destination, with meaning defined by classification labels. A joint end-to-end design of the source encoder, relay transformation, and destination decoders is developed to optimize both semantic inference accuracy and perceptual reconstruction quality for image data. Relay-assisted schemes are shown to outperform direct transmission, with gains that become more pronounced in low signal-to-noise ratio (SNR) regimes across different relaying strategies. Alongside these performance benefits, however, a key security vulnerability introduced by the relay is identified, and a Byzantine attack is formulated in which the relay injects structured, learned perturbations into its transmitted signal. The attack operates on intermediate communication representations subject to channel effects, with perturbations generated through a learned mapping applied at test time. It is shown to significantly degrade semantic inference accuracy at the destination. These results reveal a clear trade-off between SemCom gains and security risks in relay-assisted systems, motivating the need for secure relay designs that account for insider threats.
[LG-21] Detecting Silent Faults at the Environment-Trainer Boundary in Reinforcement Learning
链接: https://arxiv.org/abs/2611.08144
作者: Zuo Gou,Chenhao Lu
类目: Robotics (cs.RO); Machine Learning (cs.LG); Software Engineering (cs.SE)
*备注: 7 pages, 6 figures, 6 tables. Accepted to the CoRL 2026 Workshop on Everything Beneath the Policy. Code, specification, and every experiment’s protocol, logs and findings: this https URL
Abstract:Robot reinforcement learning increasingly runs environments outside the trainer: in another process, on another machine, in another language. Many faults at this boundary raise no error. We make the boundary a wire protocol and ask what notices a fault in it. On Pendulum, HalfCheetah, Hopper and a 7-joint arm (Pusher), we injected up to 23 faults into what crosses the wire and 2 into the trainer’s log, from one value to every value. Unless every value was wrong, a learning-curve threshold caught at most one run in ten, and a comparison with ten fault-free seeds at most 21%. A clean evaluation of all 1,060 final policies agreed with the curves: of the 728 boundary-fault runs both curve checks passed, 723 scored no more than three standard deviations below fault-free seeds, though nearly all ended on other weights and, for reward faults, logged other returns. Checks of the run itself, not of its outcome, saw every fault. A step ledger kept on both sides and compared step by step caught and located every fault that crossed, on Pendulum, Hopper and Pusher on one machine and on Pendulum across two, with no fault-free run and, on Pendulum, a median 3% more training time. An episode record kept at the environment caught the faults in the trainer’s log. We list such faults found in practice, in Gymnasium 1.3 and our own code and, reported by others, in five multi-agent libraries, and recommend a step ledger and episode counts kept at the environment. Comments: 7 pages, 6 figures, 6 tables. Accepted to the CoRL 2026 Workshop on Everything Beneath the Policy. Code, specification, and every experiment’s protocol, logs and findings: this https URL Subjects: Robotics (cs.RO); Machine Learning (cs.LG); Software Engineering (cs.SE) ACMclasses: I.2.6; I.2.9; D.2.5 Cite as: arXiv:2611.08144 [cs.RO] (or arXiv:2611.08144v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2611.08144 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-22] PersonaAgent : Retrieval-Augmented Transaction Anomaly Scoring with Refreshable Multi-Entity Persona Memory
链接: https://arxiv.org/abs/2611.08094
作者: Zhengyu Fang,Liang Wang,Yingtong Dou,Yuzhong Chen,Shubham Jain,Junpeng Wang,Xiran Fan,Menghai Pan,Jiarui Sun,Yan Zheng,Yiwei Cai
类目: Machine Learning (cs.LG)
*备注:
Abstract:In large-scale payment networks, entities (card, merchant, etc.) and their interactions play a key role in transaction anomaly detection. However, transaction patterns between different entities usually change more frequently than the production model refresh schedule. Meanwhile, entity embedding tables suffer from high-cardinality and sparse interaction issues. To resolve the above challenges, we present PersonaAgent, an agentic transaction anomaly detection framework which represents the latest entity behavior signals as personas and saves them in a refreshable memory. Offline, a weekly scheduled batch job is used to extract the latest behavior signals and update them in the memory. At inference time, a lightweight adapter retrieves the corresponding entity behavior signals of each transaction from memory and uses them to augment the base anomaly classifier’s scoring. Through comprehensive benchmarking with different anomaly detection models at Visa, PersonaAgent is able to score transactions efficiently (a 1,000-transaction CPU batch in 1-2 ms), lifting the PR-AUC by 0.65-6.58% and the production business metric by 8.93-14.24% relatively in deployment.
[LG-23] SAE Interprets MoE: Sparse AutoEncoders Reveal That MoE Experts Exhibit Human-Interpretable Fine-Grained Specialization
链接: https://arxiv.org/abs/2611.08089
作者: Anyuan Zhuo,Shaohua Zhang,Yuan Lin,Pinyan Lu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Mixture-of-Experts (MoE) models route each token to a few of many experts, yet what an individual expert is responsible for remains unsettled. Existing analyses count how often an expert is selected on domain-labeled corpora, an instrument that fixes the taxonomy in advance, stays coarse, and records selection rather than contribution. We instead study specialization in feature space by training a sparse autoencoder (SAE) on the output of an MoE layer. Since this output is a router-weighted sum of expert outputs and the SAE encoder is affine, the pre-activation of every latent decomposes exactly over the experts. Aggregated over a corpus, this gives each expert a dominant set of latents whose activation it principally drives. On layers 4, 30 and 46 of Qwen3-30B-A3B, contribution is far more concentrated than gating, most latents have only one or two dominators, and the overlap between dominant sets concentrates on a few expert pairs. Auto-interpreting these sets reveals a division of labor below coarse domains, with distinct experts for Java and front-end web code, or for geometry and for arithmetic and polynomials, while the coarse domains themselves take very unequal numbers of experts. Our SAEs and code are available at this https URL.
[LG-24] Optimal Competitive Ratios for Chasing Convex Functions with Quadratic Growth
链接: https://arxiv.org/abs/2611.08078
作者: Linzhe He,Sifan Yang,Wei Jiang,Lijun Zhang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Chasing convex function has attracted much attention in recent years, which can be formulated as a game between a learner and an adversary. In each round, the adversary reveals a convex function, and the learner selects a decision and incurs a hitting cost and a movement cost for changing its decision. The performance of the learner is measured by the \emphcompetitive ratio, which compares its cumulative cost with that of an optimal offline decision sequence. For Euclidean movement, existing upper and lower bounds for \kappa -well-conditioned functions exhibit a polynomial gap. For squared Euclidean movement, dimension-free competitive guarantees have been established under quadratic growth, but the optimal competitive ratio remains unclear. In this paper, we establish optimal dimension-free competitive ratios for convex function chasing under two types of movement costs. For Euclidean movement, we propose Clipped Proximal Descent (CPD), which combines a balanced proximal step with a clipped correction toward the current minimizer. CPD achieves an O(\kappa^1/3) competitive ratio for smooth convex functions satisfying quadratic growth, improving the previous O(\sqrt\kappa) upper bound and matching the existing \Omega(\kappa^1/3) lower bound. For squared Euclidean movement, we establish the exact optimal dimension-free competitive ratio for convex functions satisfying quadratic growth. Our result is based on a new tight competitive analysis of Regularized Online Balanced Descent (R-OBD) and a matching lower bound against randomized algorithms. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2611.08078 [cs.LG] (or arXiv:2611.08078v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.08078 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-25] Performance metrics in machine learning: a scenario-based guide from the confusion matrix to calibration regression and uncertainty
链接: https://arxiv.org/abs/2611.08048
作者: Mehmet Baygin,Sengul Dogan,Turker Tuncer
类目: Machine Learning (cs.LG)
*备注: 41 pages, 7 figures, 23 tables, including appendices
Abstract:Performance metrics can yield different conclusions from the same predictions. This methodological tutorial explains the effects of class prevalence, decision thresholds, aggregation rules and software conventions through controlled scenarios. Twenty-two confusion-matrix metrics are organized in a prevalence-coded map. Fifteen scenarios cover binary and multi-class classification, score discrimination, probability calibration, regression, uncertainty and consistency checks. Inputs are fixed in released files, and one factor is varied in each scenario. All 205 verification checks passed; 202 were independent and 3 were internal. An always-negative rule achieved an accuracy of 0.990 at a prevalence of 0.010, with balanced accuracy of 0.500. At fixed sensitivity and specificity, precision fell from 0.889 to 0.008 as prevalence fell from 0.5 to 0.001. Three probability variants shared a ROC-AUC of 0.825 but had calibration slopes of 2.08, 1.04 and 0.52. With 10 participants, participant-majority accuracy of 0.700 had a 95% interval of 0.348-0.933. Fold-averaged scores were falsely flagged as inconsistent under an inappropriate pooled-test-set model. A goal-based metric table and a reporting checklist are provided. The scenarios illustrate mechanisms and do not estimate their frequency in published studies. A literature audit and a real-data demonstration are specified as future protocols and were not executed.
[LG-26] KeHoSu: Keyed Homoglyph Substitution for Authenticated Text Provenance over Copy-Paste Channels
链接: https://arxiv.org/abs/2611.08046
作者: Andrei Alexandru Drăghici,Sediola Ruko,Briland Hitaj,Dorjan Hitaj
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: 12 pages, 5 figures, 9 tables
Abstract:Nowadays, large language models produce text at scale, and much of it is redistributed across platforms. A moderator who finds a suspicious comment, post or article may want to know which platform generated it and for which account. In this paper we study attribution over the channel through which such text actually travels, copy-paste, which discards document metadata but keeps Unicode code points. We present KeHoSu, which turns the confusable characters of homoglyph attacks into a keyed attribution channel: the variant chosen for each carrier letter encodes one bit of an identifier made of a platform field, a user field and a truncated authentication tag. Moreover, given that reshared text is often quoted in fragments, and the placement therefore keeps no positional state, we show that KeHoSu is invariant to sentence reordering up to a constant boundary layer and depends on a partial quotation only through the retained length. We characterize the code-point channel by its capacity, removability and detectability, and provide an achievable region that relates the three. We evaluate KeHoSu over a wide range of use cases and show that the mark survived copy-paste across 18 common platforms, attributing the majority of all excerpts to the correct account at 178 letters (about two sentences) on LLM-generated stories and 167 letters (about one sentence) on standard English prose.
[LG-27] Should We Pay This Much for Robustness? Efficient Proxy Certificates with Marginal Guarantees NEURIPS2026
链接: https://arxiv.org/abs/2611.08035
作者: Soroush H. Zargarbashi,Mohammad Sadegh Akhondzadeh,Aleksandar Bojchevski
类目: Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2026
Abstract:Modern classifiers remain sensitive to small input perturbations, while strong model-agnostic robustness certificates such as randomized smoothing are often too expensive for deployment, requiring an extensive number (e.g. 2000) of model forward passes per input. We propose a framework that delegates certification to a cheap proxy function while statistically controlling its deviation from the original sound certificate. Given an exchangeable unlabeled calibration set, we tune the proxy via robust conformal risk control to guarantee a bounded marginal (rather than pointwise) false-approval rate: the probability, over future inputs, that the proxy accepts an input whose prediction is flipped due to perturbation. The resulting proxy certificate can be used either as a standalone marginal certificate or as a fallback mechanism that invokes the expensive certificate only when the proxy rejects. We provide both single-sample and sample-heavy calibration recipes. Empirically, as a standalone certificate our method reduces the test-time certification cost to a single model forward while retaining much of the certified performance of randomized smoothing at small radii, and as a fallback it recovers the certified ratio of randomized smoothing at a more modest speedup. This enables cheap yet effective robustness guarantees for larger architectures such as vision transformers and language models.
[LG-28] Reproducing Kernel (Quasi-)Banach Spaces with the ellp Coefficient Norms
链接: https://arxiv.org/abs/2611.08032
作者: Tianfang Wang,Guohui Song,Haizhang Zhang
类目: Machine Learning (cs.LG)
*备注:
Abstract:This paper advances the theory of reproducing kernel Banach spaces (RKBSs) by investigating the general \ell^p regularization framework, with a particular focus on the non-convex regime 0p1 . While our preceding work established the foundation for RKBSs with the \ell^1 norm, the pursuit of stronger sparsity—closer to the \ell^0 pseudo-norm—naturally motivates this exploration into \ell^p quasi-norms. For 0p1 , we construct reproducing kernel quasi-Banach spaces (RKQBSs) with \ell^p quasi-norms and establish existence and sparsity results for interpolation and regularized learning under appropriate assumptions. We characterize the linear and relaxed representer theorems in terms of a generalized Lebesgue constant and derive bounds under kernel decay, sample spacing, and stability assumptions. For p1 , we construct RKBSs using a fixed countable dictionary, establish minimizer existence, and identify the optimal relaxed representer factor through an off-sample coefficient operator. These results reveal fundamental differences between the regimes 0p1 , p=1 , and p1 . Numerical experiments compare the accuracy and sparsity of solutions for these three cases. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2611.08032 [cs.LG] (or arXiv:2611.08032v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.08032 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Haizhang Zhang [view email] [v1] Thu, 5 Nov 2026 13:52:28 UTC (1,025 KB) Full-text links: Access Paper: View a PDF of the paper titled Reproducing Kernel (Quasi-)Banach Spaces with the \ell^p Coefficient Norms, by Tianfang Wang and 2 other authorsView PDFHTML (experimental)TeX Source view license Additional Features Audio Summary Current browse context: cs.LG prev | next new | recent | 2026-11 Change to browse by: cs References Citations SciXGoogle Scholar Semantic Scholar export BibTeX citation Loading… BibTeX formatted citation loading… Data provided by: Bookmark checked="checked"class=“labs-tab-input”> Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) IArxiv recommender toggle IArxiv Recommender (What is IArxiv?) 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)
[LG-29] GTAFL: Addressing Test-Agnostic Long-Tailed Federated Learning in Medical Image ALT
链接: https://arxiv.org/abs/2611.08021
作者: Guangyu Chen,Jingyun Zeng,Hui Jiang
类目: Machine Learning (cs.LG)
*备注: Accepted for publication in IEEE Journal of Biomedical and Health Informatics (JBHI). 13 pages. Code: this https URL
Abstract:In healthcare, protecting patient privacy is crucial due to the sensitivity of medical data and its extensive accessibility. Federated Learning (FL) offers a decentralized and privacy-preserving training paradigm, making it an ideal solution for healthcare applications. A critical challenge in healthcare is that real-world medical data often exhibits long-tailed distributions in local and global views. Existing methods addressing long-tailed FL problem typically assume that the model will be evaluated on uniform test data distribution. However, practical test data in healthcare systems is often agnostic and unpredictable, leading to potential model failures in real-world scenarios. In this paper, we introduce a novel task termed Test-Agnostic Long-Tailed Federated Learning and propose GTAFL, a comprehensive framework to address this challenge. During the training stage, GTAFL employs adaptive re-sampling, expert classifier retraining, and self-supervised learning to correct biased classifiers and distorted feature spaces caused by long-tailed training distributions. During the inference stage, an ensemble mechanism combines retrained expert classifiers to handle test data with unknown distributions. Extensive experiments on CIFAR10 and two medical datasets manifest that our framework outperforms other state-of-the-art methods. The code is available at this https URL
[LG-30] owards Multiscale Molecular Representation Learning through Geometry-Based Frag mentation for Polymer Property Prediction
链接: https://arxiv.org/abs/2611.07998
作者: Makoto Takamoto,Chaitanya Sampara,Viktor Zaverkin,Daniel Oñoro-Rubio,Mathias Niepert
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci); Chemical Physics (physics.chem-ph)
*备注: 29 pages, 8 figures, accepted for publication in ACS Omega
Abstract:It is well established that molecular substructures can contribute unequally to chemical and physical properties, with specific motifs or side chains often playing dominant roles. Fragment-based representations have therefore emerged as an effective strategy for incorporating chemically meaningful inductive biases, as exemplified by molecular fingerprints and fragment-based attention mechanisms. However, many existing fragmentation-based machine learning methods rely on predefined chemical rules, which may yield suboptimal fragmentation for previously unseen or chemically unfamiliar structures and often overlook the importance of multiscale molecular representations. To address these limitations, we formulate molecular fragmentation as a radius-constrained optimization problem inspired by the k-center formulation, where the number of clusters is not fixed a priori but emerges from the specified radius, and propose a geometry-based fragmentation method that partitions arbitrary molecular or material structures using largely geometric information, without relying on predefined chemical fragmentation rules. The method provides flexible control over fragment size, enabling the construction of multiscale fragment representations for scale-adaptive modeling. Building on this framework, we develop a geometric fragment-based attention mechanism that integrates multiscale structural information and incorporate it into a multimodal machine learning model combining p-SMILES representations with three-dimensional (3D) conformational information. Experiments on a benchmark polymer dataset show that incorporating 3D conformational information improves prediction accuracy for several properties over models based exclusively on p-SMILES representations. Explainable artificial intelligence analyses further provide insights into the physical relevance of the learned representations.
[LG-31] QRAE: Query-Relevant Attention Entropy for Context Compression in DeepSearch Agents
链接: https://arxiv.org/abs/2611.07976
作者: Yingjie Tan,Haiquan Qiu,Quanming Yao,Manyi Zhong,Xianzhi Yu,Yaqing Wang
类目: Machine Learning (cs.LG)
*备注: 29 pages, including appendices
Abstract:DeepSearch LLM agents repeatedly search and carry retrieved sources into subsequent reasoning steps. This expanding context increases token consumption and latency, although many results do not contribute to later reasoning. We propose Query-Relevant Attention Entropy (QRAE), an online context compression method that uses sparse evidence-assessment heads to decide which intact retrieved chunks should persist across reasoning steps. It scores bounded candidate batches through paired prefills: a query-free control suppresses shared attention responses, while normalized entropy measures the breadth of the positive response within each chunk. High-scoring chunks enter a persistent evidence whitelist and are included in the reasoning context. Uncertain chunks remain outside the reasoning context but are retained in a candidate cache, where they can be retrieved for later reevaluation. QRAE requires neither an additional generative compression pass nor modification of the attention operator. Across multiple DeepSearch benchmarks and reasoning models, QRAE improves task success by an average of 2.35 percentage points and reduces end-to-end latency and reported token usage by an average of 30% and 45%, respectively. These benefits generalize to a model with nearly 100B parameters.
[LG-32] ght Asymptotic Bounds on the Maximum Number of Linear Regions of Piecewise-Linear Kolmogorov–Arnold networks
链接: https://arxiv.org/abs/2611.07973
作者: Ye Su,Yong Liu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Schoots et al.~\citepschoots2025relating (AISTATS 2025) established a generally loose upper bound on the number of linear regions of continuous piecewise-linear (CPWL) Kolmogorov–Arnold networks (KANs), without providing a lower bound. We consider a T -layer network with width vector \mathbf n=(n_0,\ldots,n_T) , input dimension n_0=d , output width n_T=1 , and at most k\geq2 affine segments per edge. Our matching upper and lower bounds determine the tight growth order \Theta_\mathbf n(k^\sum_\ell=1^T\rho_\ell-1) in k , where \rho_\ell-1:=\min_0\leq s\leq\ell-1n_s . When every hidden width is at least d , the tight order simplifies to \Theta_\mathbf n(k^dT) . The upper bound of \citepschoots2025relating is k^\sum_\ell=1^Tn_\ell-1n_\ell . Our upper bound improves it by a factor of \Theta_\mathbf n!\left(k^\sum_\ell=1^Tn_\ell-1n_\ell-\sum_\ell=1^T\rho_\ell-1 \right) , with strict asymptotic improvement whenever a hidden layer has width greater than one. We further compare CPWL KANs with rectified linear unit (ReLU) networks using dense parameterizations with at most \Lambda parameters per model and the same fixed nonlinear depth T . Counting trainable knot locations and optimizing the widths and k , we prove that the maximum region count of CPWL KANs divided by that of ReLU networks is \Theta_d,T(\Lambda^d\lfloor T/2\rfloor) as \Lambda\to\infty , yielding a polynomial advantage for KANs when T\geq2 . These tight bounds characterize the attainable geometric complexity of CPWL KANs, providing a foundation for understanding their expressive power and guiding architectural design.
[LG-33] Invariance before structure: two modes of emergent compression in representation learning NEURIPS2026
链接: https://arxiv.org/abs/2611.07952
作者: Hardik Rajpal,Dan Goodman
类目: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Neurons and Cognition (q-bio.NC)
*备注: Extended abstract, UniReps workshop at NeurIPS 2026
Abstract:Learned representations show striking invariance across biological and artificial neural networks, with different substrates and variations driven by drift, heterogeneity and initialisation. To understand this invariance, we propose Emergent Adaptive Representation Learning (EARL), a framework that distinguishes two modes of emergence in learned representations: state-space compression, where a learned representation gains predictive power over its own future beyond what the network’s units can predict individually, and parameter-space compression, where networks with different micro-level parameters converge to the same representation. We formalise both as Fisher sensitivities and introduce ensemble-level information-theoretic measures for each. In a recurrent reservoir with a trained bottleneck, parameter-space compression rises within the first few epochs of training and partly independently of the task, whereas state-space compression rises about 30 epochs later and only when the network begins to learn the temporal structure of the task. These results suggest that invariance to the substrate and task-specific emergent structure are acquired in distinct phases of learning.
[LG-34] he Blindfolded Controller: Decision Model-Guided Training for Reinforcement Learning
链接: https://arxiv.org/abs/2611.07934
作者: Muhammad Ibrahim Khan,Aniket Dixit,Faizan Ahmed,James Brusey
类目: Machine Learning (cs.LG)
*备注: 19 pages, 6 figures, 6 tables
Abstract:Reinforcement learning (RL) provides a framework for continuous improvement through interaction with an environment. However, RL agents struggle against tasks that offer exploration challenges and sparse rewards, because learning efficient policies from scratch is slow and sample-inefficient. In contrast, decision models (DMs), given a description of the environment and task, can perform well in sparse-reward tasks, but, as they do not learn from reward, they are not able to improve towards the optimal solution. In this paper, we present Decision Model-Guided Training (DMGT), a technique that uses a DM with a description of the environment and task to initialise an RL policy, combining the DM’s ability to act without training with RL’s ability to improve from reward. Instead of exploring from scratch, the agent reaches the goal and receives reward from its first episodes, and RL then improves its policy beyond what the DM achieves. We demonstrate through experiments on seven tasks that our approach outperforms RL in all seven and DM in five. In addition, we provide comprehensive guidelines, derived from ablation studies, for describing an environment and task to a DM.
[LG-35] Same Answer Different Future: Matched Predictions Do Not Imply Matched Belief Updating in Language Models
链接: https://arxiv.org/abs/2611.07913
作者: Weihan Li,Junhao Wu,Xinlei Chen,Tianshi Zheng
类目: Machine Learning (cs.LG)
*备注: 24 pages, 13 figures, 7 tables
Abstract:Language models can produce similar probabilistic predictions after observing outcomes, reading summaries, receiving instructions, or undergoing activation interventions. Whether these routes induce equivalent belief states remains unclear: agreement in current predictions does not establish how a model will respond to subsequent evidence. We investigate this distinction in a Bernoulli inference task with an analytic Bayesian reference. Across three open-weight language models, we prepare states through four textual routes and two activation-level interventions, measure both next-outcome continuations and explicit probabilistic estimates, and append identical new evidence from each prepared state. We quantify the persistence of the initial belief through its effective evidence weight, the number of new observations it is worth. We find that the format of evidence decides which readouts express its content: outcome sequences primarily affect continuations, whereas summaries and stated rates primarily affect explicit estimates. Moreover, effective weights differ substantially across routes, with summaries receiving weights that scale weakly with their stated observation counts and elicitation questions changing inferred weights by orders of magnitude. In a one-shot steering setting, an intervention that reproduces a natural record’s immediate prediction loses most of its effect under subsequent evidence. These results show that agreement in current readouts is insufficient to establish equivalence in belief updating: how a belief is acquired and how it is read jointly determine its subsequent dynamics.
[LG-36] Which Signals of Code Correctness Survive Distribution Shift? Hidden-State and Causal Predictors of Program and Verdict Correctness
链接: https://arxiv.org/abs/2611.07876
作者: Viola Campos,Thomas Klassert,Adrian Ulges
类目: oftware Engineering (cs.SE); Machine Learning (cs.LG)
*备注:
Abstract:Large language models increasingly judge code, yet without external feedback such judgments are unreliable, and verifying them by executing tests or by extended reasoning is costly. Instead, execution-free correctness predictors use signals from the model’s output probabilities, sampled solutions, or hidden states to identify judgments that warrant further verification. However, these predictors may degrade under distribution shifts when generalizing across projects and languages. We compare six execution-free predictors from four families on Qwen3-Coder-30B and gpt-oss-20b under benchmark shift between LiveCodeBench and BigCodeBench and language shift across Python, C++ and Java. We evaluate two targets: whether the model’s own verdict on a program can be trusted, and whether the program itself is correct. Besides established approaches, we adapt two novel causal localization methods to code, which have been shown to be more robust out-of-distribution on natural language tasks. For verdict correctness, causal localization improves robustness to language shift. For Qwen3-Coder a probe on the causally localized subspace stays 0.09 AUROC ahead of output confidence under language shift (0.11 in distribution), significantly outperforming the best non-causal predictor by 0.04 AUROC on average. This advantage, however, does not generalize consistently to gpt-oss or to benchmark shift. For program correctness, predictors transfer more successfully between benchmarks, largely by identifying problems on which the model tends to fail. However, their ability to rank candidates within the same problem is limited. Subjects: Software Engineering (cs.SE); Machine Learning (cs.LG) Cite as: arXiv:2611.07876 [cs.SE] (or arXiv:2611.07876v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2611.07876 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-37] EnergyBench: a live gate-aligned benchmark of day-ahead forecasts for European power markets
链接: https://arxiv.org/abs/2611.07784
作者: Evgenii Genov,Ander Martinez Alonso,Christoph Bergmeir
类目: Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 21 pages, 10 figures, 7 tables. Live benchmark at this https URL
Abstract:Energy forecasting models are usually compared on backtests and fixed public datasets. Backtests can use data that did not yet exist when the forecast would have been made, and a fixed test set, reused across studies and possibly contained in the pretraining data of foundation models, invites overfitting to the benchmark. EnergyBench evaluates every entry live instead. Each day at 11:30 CET/CEST, thirty minutes before the day-ahead (SDAC) gate closes, every model forecasts the next day’s price, wind and solar generation and load on the 15-minute grid for six bidding zones (BE, NL, FR, DE-LU, DK1, DK2). Unlike existing live benchmarks, EnergyBench runs the models itself, on one information set that it freezes at the cut-off and publishes, including three numerical weather prediction models, so that foundation models can be compared with and without the same weather covariates under the actual market deadline. Errors are pooled per series against a seasonal-naive baseline, and models are ranked by a Bradley-Terry rating on the Elo scale with a day-bootstrap interval. The field spans baselines, linear and gradient-boosted models on the frozen weather, eighteen time-series foundation models from eleven families, ensembles, and the grid operators’ own forecasts. Over the first thirty-two live days, the covariates reduced the error of each of the ten foundation models run with and without them by 37-51% on matched races (64-73% on wind, 48-61% on solar). Used zero-shot, these models match or outperform a LightGBM model trained on the same covariates. Model size, from 3M to 331M parameters, has a much smaller effect, and median ensembles of the covariate models rank highest. The grid operators’ pre-gate forecasts remain ahead on wind and solar, where the best single foundation models have 16-22% more error; on load the gap is 4-6%, within the intervals.
[LG-38] PBPO: Prefix-Block Policy Optimization for Diffusion Language Models
链接: https://arxiv.org/abs/2611.07783
作者: Katherine Tieu,Dongqi Fu,Si Zhang,Hanqing Zeng,Tianxin Wei,Chenglin Wei,Yinglong Xia,Hong Li,Jingrui He
类目: Machine Learning (cs.LG)
*备注:
Abstract:Diffusion Large Language Models (dLLMs) commonly decode responses by proceeding autoregressively across blocks while iteratively denoising multiple tokens within each block. This structure makes block-level policy optimization a natural adaptation of autoregressive policy optimization. However, the exact likelihood of a completed block is intractable because it marginalizes over the intermediate states and reveal orders of the denoising process. To realize the block-level formulation, we construct a prefix-block state and use a single model evaluation to define a factorized one-step surrogate for the block likelihood. The resulting block surrogate ratio is a product of its token surrogate ratios, allowing variation at individual positions to propagate across the entire block before clipping. We therefore introduce Prefix-Block Policy Optimization (PBPO), which preserves block-level conditioning but applies surrogate importance weighting and clipping separately to each token. PBPO further diversifies repeated updates by resampling corrupted views of the prompt and revealed prefix, a design motivated by masked-dLLM pre-training. Under our analysis assumptions, block-product weighting amplifies the variance of gradient contributions relative to token-local weighting. Across two LLaDA-family initializations, six benchmarks, and three generation lengths, PBPO achieves competitive performance against existing dLLM RL methods.
[LG-39] Optimal Learning in Games with Arbitrary Delays: Best-of-Both-Worlds for Self-Play and Adversarial Opponents
链接: https://arxiv.org/abs/2611.07773
作者: Emmanuel Esposito,Andrew Jacobsen,Hao Qiu,Alexander Ryabchenko,Mengxiao Zhang
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:We study learning in zero-sum games with delayed feedback. We introduce Multi-Anchor Optimistic Online Mirror Descent (MAO-OMD), a single algorithm which obtains optimal fast rates in self-play on smooth games, while simultaneously maintaining minimax optimal guarantees against arbitrary opponents. In particular, in smooth zero-sum games with arbitrary and unknown feedback delays, MAO-OMD obtains O(\sigma_\max+1) social and individual regrets when used by both players, where \sigma_\max denotes the maximum number of outstanding gradients, while also guaranteeing minimax optimal O(\sqrtT+d_\mathrmtot) individual regret against arbitrary opponents, where d_\mathrmtot is the total delay. Moreover, we show that these optimal rates cannot be obtained simultaneously using the standard black-box reduction to the undelayed setting. We also extend our results to multi-player games, and to games with player-dependent delay sequences.
[LG-40] Neural Collapse on Imbalanced Datasets and Open-Set Scenario
链接: https://arxiv.org/abs/2611.07753
作者: Qiufu Li,Huibing Xiao,Linlin Shen,Yongdong Huang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Neural collapse (NC) describes the elegant geometric structure that sample features and classifier vectors exhibit when a classification model’s loss reaches its minimum. Class imbalance, however, generally distorts standard NC. In this paper, we theoretically reformulate NC on imbalanced datasets and in open-set settings. Our analysis is based on an optimization problem over class-wise averaged losses, rather than the global averaged losses used in prior NC studies. We prove that when class-wise averaged losses attain their minima, standard NC is achieved on balanced or imbalanced datasets, for both cross-entropy (CE) and binary cross-entropy (BCE) losses. We further extend the analysis to an open-set-inspired scenario, where some classes contain no training samples yet are still assigned classifier vectors in the losses. We find that in this scenario, when class-wise averaged losses reach their minima, standard NC still emerges on the known classes with training samples, while the classifier vectors of unknown classes without training samples overlap with each other and become orthogonal to those of the known classes. Finally, we analyze Softmax- and Sigmoid-based contrastive losses re-balanced by inverse sample frequency, proving that at their minima on balanced or imbalanced datasets, sample features of different classes likewise form standard NC.
[LG-41] Foresee to Trade: Forecast-Augmented Policy Optimization for Portfolio Management
链接: https://arxiv.org/abs/2611.07750
作者: Zeyu Liu,Xuanzhi Feng,Haoran Xu,Lai Sing Kwong,Yang Li,Hualei Zhang,Hao Chen,Jie Zhang,Xiaoyi Pang,Song Guo
类目: Machine Learning (cs.LG)
*备注:
Abstract:Online investment services must translate asset opportunities into investment outcomes for users’ accounts, yet the same market opportunity can warrant different account actions. We study how supervision from future asset opportunities can strengthen account-level trading decisions and propose Forecast-Augmented Relative Policy Optimization (FRPO). FRPO separately supervises asset opportunities and account decisions through a jointly learned representation: a Financial State Head (FSH) learns future relative-return targets to provide asset-level opportunity supervision, while Same-Path Group-Relative Optimization compares trading trajectories under the same market path and initial account to provide account-level relative-outcome supervision. An Adaptive Gate coordinates auxiliary supervision using predictive variance and gradient alignment. FSH is used only during training; inference uses the encoder and policy for trading. Across DJIA30, TOPIX Core30, and three investment-style portfolios, ten-seed evaluations show that FRPO achieves the highest mean cumulative and annualized returns among the compared methods in all five settings. These results support strengthening investment agents’ trading decisions through joint supervision from asset opportunities and trading outcomes.
[LG-42] ConvPE: Convolutional Position Embedding Improves Transformers
链接: https://arxiv.org/abs/2611.07749
作者: Zehao Wang
类目: Machine Learning (cs.LG)
*备注:
Abstract:ConvPE (Convolutional Position Embedding) improves both the performance and efficiency of Transformer models in language modeling, long-context extension, and vision. It injects a windowed, content-dependent bias into attention scores through a lightweight causal convolution over a low-dimensional projection of the hidden states. As a lightweight plug-in, ConvPE is compatible with RoPE, attention masks, and KV concatenation. It also supports high-dimensional position embeddings via high-dimensional convolutions. In a reproduction, ConvPE outperforms five 1D baselines and eight 2D baselines. The code is publicly available at this https URL.
[LG-43] Ordering Multivariate Time Series by Volatility Memory: A Temporal Convolutional Approach to Volatility Heredity
链接: https://arxiv.org/abs/2611.07746
作者: Kaoutar El Azaar,Mohamed Hamlich,Moncef Garouani,Franck Ravat,Asmae Chakir,Mohamed Tabaa
类目: Machine Learning (cs.LG)
*备注:
Abstract:Volatility dynamics can carry information about temporal relationships that is not apparent in the conditional mean. Extracting this information is difficult because the persistence of each variable can obscure its dependence on other variables. This paper presents QS-TCN, a method for ordering multivariate time series that combines statistical estimates of volatility memory with temporal convolutional learning. Persistence estimates obtained by quasi-maximum likelihood and temporal modes extracted by the Matrix Pencil method define structured filters within the network. Together with learned filters, they represent the history of the target and support comparisons of directional evidence between variables. These comparisons are combined into a global order of precedence, which does not identify a causal graph. The method is evaluated on synthetic systems with linear and nonlinear volatility interactions and on the Finance and fMRI benchmarks, whose reference relations are known. Across the four synthetic volatility regimes studied, QS-TCN reaches a higher mean ordering quality than an external lead–lag ordering and its squared-input adaptation, although the gain over the adaptation is uncertain in two regimes. Applications to currency, gold and cryptocurrency markets are reported descriptively.
[LG-44] Static Recovery Is Not Dynamic Stability: Dynamics-Aware Benchmarking of Protein Motif Scaffolding NEURIPS2026
链接: https://arxiv.org/abs/2611.07723
作者: Emile de Bruyn,Moritz Schäffler,Jannik Schneider,Sebastian R. Schmidt,Chanwoong Hwang,Aleena Siji,Utkarsh Upadhyay,Stefan Bauer,Alexander Schug,Stefan Kesselheim,Artur Yakimovich,Alina Bazarova
类目: Machine Learning (cs.LG); Biomolecules (q-bio.BM)
*备注: Accepted at NeurIPS 2026 (Evaluations Datasets Track). 23 pages, 10 figures, including supplementary material. The first three authors contributed equally. BILBY-PRO dataset: this https URL ; BILBY-MD code: this https URL
Abstract:Protein motif scaffolding benchmarks assess whether generative models can build scaffolds around functionally important residues. Their conclusions depend on which motifs are tested and which criteria define success. Existing benchmarks typically use small hand-curated motif sets and static refolding criteria, but protein function often depends on conformational ensembles and transitions. Here, we address both benchmark scope and success criteria. We introduce BILBY-PRO, a systematically derived benchmark of structurally conserved functional motifs from PROSITE-linked experimental structures, comprising 220 cases from 174 PROSITE motif-pattern entries. Evaluating six recent scaffold-generation models reveals distinct patterns in target coverage and unique-solution yield: RFDiffusion and Proteina solve more targets, whereas ESM3 produces more unique successful scaffolds on the targets it solves. We further introduce BILBY-MD, a reference-based all-atom molecular dynamics evaluation framework. Applied to statically solved designs from the original MotifBench benchmark, it compares motif drift, scaffold drift, motif-contact preservation, and motif-dihedral free-energy landscapes against reference simulations. Of 55 statically solved model-problem pairs, 31 satisfy our MD-supported criterion and only 8 pass all checks. Compact, self-contained motifs are often MD-supported, whereas interface-like or context-dependent motifs fail more frequently. Static refolding success therefore provides useful evidence of local motif recovery but does not establish dynamical fidelity of the full scaffold. BILBY highlights three complementary axes of motif-scaffolding evaluation: target coverage, unique-solution yield, and MD-supported dynamical fidelity.
[LG-45] Reinforcing What the Student Has Learned: Teacher-Guided On-Policy Distillation
链接: https://arxiv.org/abs/2611.07722
作者: Bozhen Fang,Linhao Zhong,Xuehao Wang,Huanyi Zheng,Hao Chen,Chunhua Shen
类目: Machine Learning (cs.LG)
*备注:
Abstract:On-policy distillation (OPD) improves language-model reasoning through dense teacher feedback on student-generated this http URL first observe an amplitude disparity in this feedback: on math reasoning responses, style-related positions exhibit 3.5 times the mean absolute teacher–student log-probability difference of mathematical this http URL OPD updates by these discrepancies thus directs much of the update strength toward matching the teacher’s expression. We instead take the student’s own pretrained base as a reference: the student–base difference records the token preferences acquired during post-training and provides supervision that does not depend on the student’s distance from the teacher. Building on this reference, we propose Student-Delta On-Policy Distillation (SD-OPD), which uses centered student–base differences as update coefficients, keeps each coefficient only when its sign agrees with centered teacher–student feedback, and recomputes the student signal every training round. Across 5 teacher–student configurations with Qwen3 students of 1.7B to 8B parameters, SD-OPD improves six-task mathematics average pass@1 (avg@16) over OPD by 7.76 to 10.37 percentage points. Code is available at this https URL.
[LG-46] Non-Parametric Mutual Information Estimation with Vine Copulas NEURIPS2026
链接: https://arxiv.org/abs/2611.07693
作者: Sigurd Holmsen,Berit Omli Øksnes,Ingrid Hobæk Haff,Sylvia Richardson,Ali Ramezani-Kebrya
类目: Machine Learning (cs.LG)
*备注: Accepted at the conference on Neural Information Processing Systems (NeurIPS 2026)
Abstract:We introduce VINE: a non-parametric mutual information (MI) estimator based on simplified D-vine copulas. Approximating the joint density with a simplified D-vine allows the density ratio to simplify to a product of pair copulas, which we can estimate non-parametrically. This structural simplification yields VINE, a flexible and computationally efficient MI estimator. We compare VINE against SotA neural MI estimators on a range of datasets, both under settings where the simplified D-vine assumption holds and where it is violated, demonstrating competitive accuracy at a lower computational cost for low- to medium-dimensional datasets.
[LG-47] O(1/n) Convergence Rate for RVI Q-Learning
链接: https://arxiv.org/abs/2611.07676
作者: Avik Kar,Siddharth Chandak,Rahul Singh,Shalabh Bhatnagar
类目: Machine Learning (cs.LG)
*备注: 24 pages
Abstract:We establish finite-time convergence guarantee for synchronous relative value iteration (RVI) Q -learning for finite weakly communicating Markov decision processes (MDPs). For stochastic Krasnosel’ski\uı–Mann iterations driven by a non-expansive, linearly regular map on a finite-dimensional normed space, we establish last-iterate mean-squared residual bounds under martingale-difference noise whose conditional second moment grows at most linearly with the squared residual. Suitable polynomial stepsizes of order n^-a yield an O(n^-a) rate for 0a\leq1 . We verify linear regularity and the noise condition for the centered Bellman operator for finite weakly communicating MDPs. Applying the general result gives an O(n^-1) bound on the mean-squared Bellman span residual of RVI Q -learning iterates when a suitable stepsize sequence is used. The positive gap between optimal and suboptimal stationary deterministic policies then yields an O(n^-1) bound on the expected suboptimality of the greedy policy. Consequently, O(\epsilon^-1) synchronous iterations suffice for an \epsilon -optimal policy in expectation. These guarantees are instance-dependent, with constants that depend on the hitting-time parameters and the suboptimality gap of the second-best policy.
[LG-48] Instrumental Variables in Continuous-Time Stochastic Processes
链接: https://arxiv.org/abs/2611.07674
作者: Lars Pennig,Søren Wengel Mogensen,Niki Kilbertus
类目: Machine Learning (cs.LG)
*备注:
Abstract:Causal questions often arise in systems that evolve in continuous time. While instrumental variable (IV) methods provide a principled approach to identify causal effects in static or discrete-time settings, even in the presence of unobserved confounding, they do not yet extend to continuous-time dynamics. In this work, we develop an IV framework for stationary continuous-time stochastic processes parameterized by stochastic differential equations (SDE), where the instrument itself takes the form of a stochastic process. We encode the structural sparsity and causal effects directly at the infinitesimal level of the drift, and contrast these assumptions with the IV assumptions on the discretized version of the same process. Based on these structural assumptions, we derive moment restrictions using path-valued instruments. We show how these restrictions simplify for Ornstein-Uhlenbeck processes and present an identification condition together with a corresponding consistent estimation method. Finally, we extend our estimation framework to nonlinear causal effects by expressing the path-based moment restrictions leveraging the signature kernel. Empirically, our method recovers causal effects in confounded continuous-time systems where maximum-likelihood-based estimation and discrete-time IV methods are systematically biased.
[LG-49] Calibrated Target Noise Recovers Curvature from the Gradients NEURIPS2026
链接: https://arxiv.org/abs/2611.07671
作者: Arash Jamshidi,Katsiaryna Haitsiukevich,Aristides Gionis,Kai Puolamäki
类目: Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2026. 39 pages
Abstract:Estimating curvature information from gradients alone is a fundamental problem in machine learning. The challenge becomes particularly acute in settings where only aggregate mini-batch gradients are available, yet curvature information is still required for applications such as preconditioning in optimisation, sampling, and early stopping. Perhaps surprisingly, we show that a simple target-perturbation scheme suffices to recover such curvature information from gradients alone. Our method injects zero-mean noise into the targets, with variance calibrated to the local output curvature. The covariance of the resulting noisy batch gradients then recovers practical curvature matrices, without requiring per-sample gradients, second-order derivatives, or other internal network quantities. For a broad class of non-linear models, our scheme recovers the Generalised Gauss–Newton (GGN) matrix; for feed-forward ReLU networks, it additionally recovers the diagonal blocks of the population Hessian. We demonstrate that the resulting estimator is simple, lightweight, provably optimal, and easy to integrate into existing training pipelines. Experiments demonstrate accurate curvature recovery, faster optimisation when the estimator is used for preconditioning, and the utility of the estimated GGN for early stopping.
[LG-50] An Information Bottleneck for Post-hoc Memorization Mitigation in LLM s
链接: https://arxiv.org/abs/2611.07669
作者: Dina EL Zein,Shashi Kumar,James Henderson
类目: Machine Learning (cs.LG)
*备注: 13 pages, 4 figures, including appendix
Abstract:Recent studies have shown that large language models (LLMs) can memorize and reproduce copyrighted training data, raising serious privacy and intellectual property concerns. In this work, we investigate whether memorization can be mitigated without modifying or retraining the underlying model through an information bottleneck. We propose a modular plug-and-play framework that inserts a nonparametric variational information bottleneck (NVIB) between a frozen LLM backbone and its language modeling head. By compressing the final hidden representations before token prediction, the proposed bottleneck aims to preserve general linguistic knowledge while filtering out highly specific memorized sequences. We evaluate our approach on memorization extraction tasks involving book passages and on standard language modeling benchmarks to measure the trade-off between memorization reduction and downstream utility. Our experiments show that the NVIB layer substantially reduces extractability while maintaining competitive performance on benchmark tasks. These results suggest that adding an information bottleneck offers a practical and architecture agnostic direction for post-hoc memorization mitigation in pretrained LLMs.
[LG-51] Simple Meta-Learning for HPO and CASH Using Performance Representations
链接: https://arxiv.org/abs/2611.07659
作者: Emiel Witting,Cheng Yan,Tom Julian Viering
类目: Machine Learning (cs.LG)
*备注: 24 pages, 9 figures, 7 tables
Abstract:Bayesian optimization for Hyperparameter Optimization (HPO) and Combined Algorithm Selection and Hyperparameter Optimization (CASH) often relies on explicit hyperparameter encodings as surrogate inputs. In high-dimensional search spaces, however, such representations can make it difficult for standard surrogates to learn useful performance models. We instead study an alternative meta-learning representation that describes each configuration directly as its historical performance across prior tasks. This representation can be paired directly with standard surrogate models. Our scope is the iterative selection component of AutoML, deciding which configuration to evaluate next from a fixed candidate set, rather than a complete framework with features such as runtime estimation, cost-aware budgeting and ensembling. We further apply this representation to maximal marginal relevance (MMR), a diversity-based re-ranking criterion from recommender systems, as a posterior-free acquisition rule. When selecting and optimising across multiple models on the TabArena benchmark, the performance-based representation paired with a linear Bayesian ridge regression model outperforms dedicated meta-learning baselines, achieving the same tuned performance in roughly half the number of trials. For single-model HPO, it remains a competitive second. Across both benchmarks this method closely matches or improves on regret at a fraction of the complexity and cost, requiring no pre-training or tuning and scaling independently of search space dimensionality. Code for reproducing our results is available at this https URL.
[LG-52] Online Learning with Ordered Experts (via Two-point Feedback)
链接: https://arxiv.org/abs/2611.07629
作者: Shiliang Zuo
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study adversarial online learning with expert recommendations in a totally ordered action space, where the experts’ recommendations may change and be reordered arbitrarily. Our focus is two-point feedback: on each round the learner plays two current recommendations, observes both, and pays their average. We study two problems. In the separated monotone problem, the loss is the sum of a hidden nondecreasing component and a hidden nonincreasing component, and a query reveals the two components separately. In the BV problem, the losses are bounded and have bounded variation (BV) along the order, and a query reveals only the loss. In the separated monotone problem, two paid observations achieve [ O\bigl(\log(|\cF|T)\sqrtT\log|\cF|\bigr) ] regret against the best of |\cF| experts, within a logarithmic factor of the classical full-information bound. For the BV problem, we obtain \wtO(\sqrt T+|\cF|) regret. We give an application in dynamic pricing with expert advice, showing that two-option menus matches the full information rate up to \log factors. The main technical ingredient is a marginal-preserving endpoint flow estimator. Two observations reveal a loss increment along the order; random paths supply telescoping identities that turn this local information into unbiased loss estimates for all current recommendations. A reversible pair sampler preserves the learner’s intended marginal at both played actions. The estimator’s second-moment and stability bounds support online updates for both feedback models. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2611.07629 [cs.LG] (or arXiv:2611.07629v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.07629 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-53] Monotone Viability Learning for Offline Safe Reinforcement Learning
链接: https://arxiv.org/abs/2611.07616
作者: Ziqian Wang,Zhen Zhang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Offline safe reinforcement learning requires reliable long-horizon safety estimates from fixed data. Learning Hamilton-Jacobi (HJ) viability in this setting calls for semantic preservation, convergence, and numerical stability. To address these challenges, we propose Monotone Viability Learning (MoViL), which learns viable sets from offline transitions through monotone finite-horizon set recursion. The recursion preserves finite-horizon safety semantics and converges to the infinite-horizon viability kernel. Binary supervision maintains a fixed target scale across horizons, avoiding discount-induced numerical attenuation. We further establish sufficient conditions for recovering behavior-supported viable subsets. The learned viability estimates are used to certify actions at deployment. Dubins car experiments validate the numerical accuracy of the learned viable sets. Evaluations on safe RL benchmarks demonstrate state-of-the-art reward-cost performance. MoViL satisfies the average cost constraint on every task and achieves the lowest average cost among the compared methods. These results highlight MoViL’s potential to extend long-horizon HJ viability analysis to safety-critical systems where reliable dynamics models are unavailable and online exploration is unsafe.
[LG-54] Binding Paraphrased Instructions to Encountered Objects in sLLM Agents under Partial Observability
链接: https://arxiv.org/abs/2611.07599
作者: Chanyoung Kim,Harim Song,Byoungchan Ko,Keonyoung Lee,Myungwon Lee,Dahuin Jung
类目: Machine Learning (cs.LG)
*备注: 44 pages, 26 figures, 18 tables
Abstract:Reliable deployment of LLM agents requires grounding diverse user expressions in their environments. Although RL-finetuned smaller large language model (sLLM) agents rival larger prompted LLMs on embodied tasks, their robustness to instruction paraphrases remains unclear. We introduce ALFWorld-Para to evaluate this robustness under controlled paraphrases, revealing 21-52% declines in success rates (SRs), predominantly driven by object synonyms. Despite retaining knowledge of synonym referents, agents often fail to take the corresponding objects even after observing them, exposing a gap between referential knowledge and object binding. Partial observability further complicates binding, as referents may remain unseen or have already been encountered under different labels. To address this challenge, we propose BINDIntent, a training-free framework that probabilistically binds referring expressions to observed labels using semantic proximity and estimated persistence against unseen candidates. Without additional training or LLM calls, BINDIntent improves synonym-paraphrase success of RL-finetuned sLLMs by 17.5-49.1 percentage points (pp) with only 0.4 ms of CPU overhead per step. It also raises GiGPO-7B’s overall SR from 61.8% to 77.4%, surpassing a prompted 24B LLM. Code and benchmark are available at this https URL.
[LG-55] ACSeL: Adaptive Configuration Search for Hardware-Constrained LLM Inference
链接: https://arxiv.org/abs/2611.07570
作者: Bokun Wang,Hung-Yueh Chiang,Diana Marculescu
类目: Machine Learning (cs.LG)
*备注:
Abstract:Serving large language models (LLMs) under tight memory and latency service-level objectives (SLOs), including on edge devices, requires jointly selecting the model size, batch size, quantization algorithm, and mixed-precision quantization recipes. Exhaustive search over this combinatorially large space is infeasible. Unlike data-center serving, which often prioritizes throughput or generation speed, resource-constrained deployments typically set memory and latency budgets as constraints and then seek the configuration with the highest generation quality among feasible ones. To this end, we propose ACSeL, a framework that identifies configurations predicted to satisfy deployment constraints while achieving high predicted generation quality. To enable efficient quality estimation, we use Relative Model Quality Degradation (RMQD), a proxy based on the divergence between the candidate and reference models’ next-token distributions over a small calibration corpus. Across the evaluated model sizes and precisions, RMQD is inexpensive to evaluate and strongly correlated with downstream benchmark performance. ACSeL combines (i) a lightweight quality surrogate that predicts RMQD directly from a configuration; (ii) Few-shot Median Calibration (FMC), which calibrates operator-profile-based latency estimates using a small number of end-to-end profiles; and (iii) a structured search algorithm BandPareto for exploring layerwise mixed-precision configurations. Experiments on an NVIDIA RTX A6000 workstation and an NVIDIA Jetson Thor edge platform show that FMC substantially improves deployment feasibility, while BandPareto + FMC achieves the highest feasibility and the lowest RMQD gaps to the best measured configurations in the evaluation set.
[LG-56] MambaTG: Candidate-Aware State Space Modeling for Future Link Prediction on Temporal Graphs
链接: https://arxiv.org/abs/2611.07569
作者: Nguyen Hong Quan,Sitao Luan,Yilun Zheng,Lihui Chen
类目: Machine Learning (cs.LG)
*备注:
Abstract:Future link prediction on temporal graphs involves both repeated interactions between previously connected node pairs and first-time interactions between pairs with no prior interaction. While repeated-edge prediction can exploit direct pairwise evidence, unseen-edge prediction must generalize beyond historical recurrence. Existing temporal graph models often exploit signals tied to recurring interactions or historical overlap, which are highly informative for repeated edges but become limited or unavailable for unseen edges, contributing to substantial performance degradation on unseen-edge prediction. Based on this observation, we identify two complementary signals for modeling sequential dynamics beyond historical recurrence: (1) historical-sequence evolution, which captures dependencies among the source node’s chronologically ordered interactions, and (2) candidate–history compatibility, which captures how well a candidate fits the source node’s evolving history. Based on the decomposition, we propose MambaTG, a candidate-aware Mamba-based framework that jointly models these two signals. Mamba provides a natural mechanism for this design: its selective recurrence captures the evolution of ordered interaction histories, while its input-dependent state transitions allow the history state to adapt to different candidate inputs. Extensive experiments on 13 temporal graph datasets against 13 representative baselines show that MambaTG achieves strong and consistent performance across both repeated and unseen edges while remaining computationally efficient.
[LG-57] Provable Zeroth-Order Safe Exploration for Reinforcement Learning from Human Feedback
链接: https://arxiv.org/abs/2611.07561
作者: Weiqin Chen,Xinmeng Huang,Dongsheng Ding,Santiago Paternain
类目: Machine Learning (cs.LG)
*备注:
Abstract:Safe reinforcement learning from human feedback (Safe-RLHF) seeks to improve helpfulness while satisfying harmlessness constraints. Existing primal-dual approaches can converge to feasible policies but do not prevent constraint violations during training. We study a stronger requirement, safe exploration for RLHF, under which every policy executed during training, including auxiliary exploration policies, remains feasible. We propose Zeroth-Order Log-Barrier Policy Gradient (ZO-LBPG), a reward-model-free method that optimizes a log-barrier objective using zeroth-order estimates derived directly from human preferences. To ensure safety under noisy feedback, we combine conservative constraint margin estimates with coupled rollouts and an adaptive step size that prevents updates from crossing constraint boundaries. We prove that, with high probability, all iterates and exploration policies remain strictly feasible, and that ZO-LBPG globally converges to an approximately optimal constrained policy. We establish a sample complexity of \widetilde\mathcal O(\varepsilon^-6+d^3\varepsilon^-5) trajectories and \widetilde\mathcal O(\varepsilon^-8+d^5\varepsilon^-7) human queries. Experiments on a synthetic CMDP and a language-model alignment task corroborate the training-time safety guarantees.
[LG-58] Understanding Data Influence in Transportation Networks: A Physics-Constrained Influence-Function Framework
链接: https://arxiv.org/abs/2611.07526
作者: Hamidreza Attari,Yueyue Fan,Johannes Royset
类目: Machine Learning (cs.LG)
*备注:
Abstract:Understanding how individual training observations influence model predictions is important for developing reliable and interpretable data-driven transportation systems. Existing influence-function methods, however, are primarily formulated for unconstrained learning and do not explicitly account for physical constraints that couple different components of a transportation network. Under such constraints, perturbing a training observation can affect model parameters and predictions only through physically feasible directions, potentially changing how its influence propagates across the network. This paper develops a hard physics-informed influence-function framework for constrained learning, in which training-data influence is characterized within the physically feasible space defined by network constraints. The analytical results reveal how physical constraints alter the magnitude, direction, and spatial distribution of training-sample influence relative to conventional influence measures. Using traffic prediction under flow-conservation constraints as a case study, we show that influential observations identified by the proposed framework differ systematically from those identified by classical influence functions. Through orthogonal decomposition, we quantitatively characterize how much of a perturbation in the training data is propagated through the network via the flow-conservation constraints and how much is removed by the constraint projection. The proposed influence framework provides a physically consistent approach for evaluating training-data influence in constrained transportation learning problems, with potential applications to data diagnostics, sensor evaluation, and targeted data collection.
[LG-59] ZYFA: Speed and Accuracy of FP8 FlashAttention on AMD CDNA4
链接: https://arxiv.org/abs/2611.07496
作者: Quentin Anthony,Ganesh Bikshandi,Praneeth Medepalli,Stephen Ebert
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: 36 pages, 28 figures, 19 tables
Abstract:Attention is a core building block of frontier models and often dominates compute-bound prefill at long context. FP8 matrix instructions offer twice the peak throughput of BF16, but they accelerate only matrix multiplication, and the softmax between the two matrix multiplications of attention runs on vector units that FP8 does not accelerate. On the AMD Instinct MI355X, the vendor’s AITER kernels reach 39% of peak FP8 matrix throughput at 32K tokens, compared with 49% in BF16. Attention is also more numerically sensitive to quantization than feed-forward layers, and unlike NVIDIA’s Hopper and Blackwell GPUs, the MI355X does not track the completion of matrix instructions in hardware, so overlapping the softmax with the matrix multiplications depends on instruction order. We present ZYFA, prefill kernels for the MI355X built on three techniques: (1) computing each FP8 probability directly from its score with a multiply-add and an integer conversion instead of an exponential; (2) a ping-pong schedule in which one group of threads computes its softmax while another executes its matrix multiplications; and (3) scaling each block of probabilities up by a power of two before rounding, and normalizing by the sum of the rounded probabilities. On attention inputs captured from three models and timed at sustained clock speeds, ZYFA achieves 1.16-1.76x the throughput of AITER. On the Gaussian inputs of kernel benchmarks at 32K tokens, it achieves 1.26-1.37x at head dimension 128 and 1.99x for DeepSeek-style multi-head latent attention. The sustained clock of the MI355X depends on operand values, so throughput on random inputs does not reliably predict throughput on real model activations. At head dimension 128, ZYFA’s numerical error is within a few percent of that of Composable Kernel and Triton and 62-78% lower than that of AITER’s. The source code is available at this https URL.
[LG-60] Learning Multi-Agent Orchestration:Higher-Order Interactions in Policy Improvement
链接: https://arxiv.org/abs/2611.07495
作者: Kaixuan Liu,Guojun Xiong
类目: Machine Learning (cs.LG)
*备注:
Abstract:Learned orchestration has emerged as a promising paradigm for large language model (LLM)-based multi-agent systems, in which a central orchestrator selects agents and controls the information they receive throughout a task. Despite rapid empirical progress, existing methods train the orchestrator by crediting each decision with its own advantage, implicitly assuming that the other decisions in the episode stay fixed. This assumption fails in orchestration, where the contribution of each decision to the final outcome depends on the decisions preceding it, so the gain of a joint update is not the sum of its per-decision effects. To address this challenge, we introduce \textbfIAPO, an interaction-aware policy optimization framework that explicitly accounts for interactions among decisions. Specifically, IAPO exactly decomposes the performance difference between two policies into the effects of individual decisions and their interactions, which we prove can account for half of the attainable gain, and then optimizes both from sampled trajectories while certifying that every update is an improvement. On long-input versions of GSM8K, MATH, DROP, MBPP and HumanEval, against fifteen published multi-agent and workflow-search systems on the same 7B executor, IAPO’s deployed plans reach the strongest system’s accuracy with fewer tokens, and removing the interaction term costs 7 to 17 points of accuracy on code.
[LG-61] wo-stage residual learning for calibrated uncertainty in solar irradiance forecasts
链接: https://arxiv.org/abs/2611.07444
作者: RR Flanagan,E Kasha,PW Talbot,AS Epiny,MR Deinert
类目: Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注:
Abstract:Accurate forecasting of renewable generation requires not only point predictions but also reliable estimates of the uncertainty attached to them, particularly at extended horizons. Machine learning models are increasingly used to predict solar output, but they typically report a single value and capture the associated uncertainty poorly. The standard remedy is to train a model to predict the parameters of an assumed probability distribution directly. Here we develop an alternative two-stage method, motivated by synthetic solar signal generation, in which the forecast error is modeled separately from the expected irradiance. A first long short-term memory (LSTM) network predicts the expected global horizontal irradiance, and a second network models the distribution of the forecast error, which is added back to the mean to give both the forecast and its uncertainty. Using historical weather data from the National Solar Radiation Database and a probability integral transform evaluation, we show that explicitly modeling the error brings the predicted standard deviation markedly closer to the historical value, reducing the deviation between the two by 34.6 percent for a standalone LSTM and by 56.5 percent when the LSTM is coupled to Gaussian process regression, the most reliable of the configurations tested. Because the method models the total forecast error directly, it captures the full predictive uncertainty that matters for decisions, rather than only the model contribution that dropout-based approaches estimate.
[LG-62] runcated backpropagation through time induces a two-sided gradient turnpike property
链接: https://arxiv.org/abs/2611.07415
作者: Julian D. Schiller,Matthias A. Müller
类目: ystems and Control (eess.SY); Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Truncated backpropagation through time reduces the computational burden of training recurrent models by restricting gradient computations to shorter temporal contexts. This induces two distinct boundary effects: missing past information perturbs the forward state evolution, while truncating the backward recursion omits future information. This motivates decoupling the interval over which gradient contributions are averaged for the parameter update from the surrounding temporal context over which they are computed. Under exponential decay of the state-transition Jacobians and mild regularity conditions, this leads to a two-sided gradient turnpike property: the error between truncated- and full-context gradient contributions decays exponentially with the distance from both context boundaries. We derive bounds for the averaged update gradient, yielding guarantees on gradient alignment and approximate full-context stationarity; under unbiased interval sampling and a Polyak-Lojasiewicz condition, these further imply a bound on the resulting optimality gap. Numerical experiments on the Wiener-Hammerstein benchmark illustrate the role of past and future context and the trade-off between predictive performance and computational cost.
[LG-63] Spectral symmetry fingerprints in equivariant representations
链接: https://arxiv.org/abs/2611.07410
作者: Jiacheng Cen,Anyi Li,Ning Lin,Zihe Wang,Ji-Rong Wen,Wenbing Huang
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci); Mathematical Physics (math-ph); Chemical Physics (physics.chem-ph)
*备注:
Abstract:Symmetry governs the organization of physical structures, while symmetry breaking reveals their transformations. In geometric deep learning, symmetry imposes rank constraints on equivariant representations, limiting the expressive capacity of equivariant graph neural networks (GNNs). Here we show that these constraints can instead be exploited: the singular-value spectra of equivariant representations provide a unified fingerprint of structural symmetry. We introduce spectral symmetry fingerprints (SymFP), which enable discrete point-group identification through rank-template matching and continuous symmetry analysis by quantifying spectral energy beyond target rank constraints. SymFP is orientation invariant and architecture agnostic, enabling a parameter-free GNN implementation. Its spectra can be reused across multiple target groups, and the resulting leakage scores satisfy subgroup monotonicity. Across synthetic structures, molecules, coordination complexes and large biological assemblies, SymFP accurately identifies point groups, tracks controlled symmetry breaking and reveals local and degree-resolved symmetry changes that can be obscured by global geometric measures. These results establish the spectral structure of equivariant representations as a general framework for characterizing symmetry across physical systems.
[LG-64] Adversarial Multiplicity: Cherry-Picked Models and the Limits of Disclosure
链接: https://arxiv.org/abs/2611.07378
作者: Cassandra C. Chou,Benjamin Q. Huynh,Allison Koenecke,Elizabeth T. Chin
类目: Machine Learning (cs.LG); Applications (stat.AP)
*备注:
Abstract:Predictive multiplicity – where similarly accurate models disagree on individual predictions – can be leveraged to optimize secondary metrics (e.g., fairness) without loss in predictive performance. However, exploiting multiplicity opens the opportunity for motivated parties (or coding agents gaming their objectives) to steer predictions toward their own ends, such as in litigation or consumer disputes with automated decision-making tools. In these adversarial multiplicity settings, we distinguish soft manipulation, where parties optimize over innocuous model choices (e.g., hyperparameter tuning, feature engineering), from hard manipulation, where parties directly perturb model parameters. We find a sufficiently flexible model can be manipulated to reverse any individual prediction with minimal change in predictive performance while passing black-box audits, making hard manipulation viable in most regulatory settings. Soft manipulation has smaller effects but can pass white-box audits requiring full access to code and data. Even aggregate estimates can be manipulated: a model’s estimated effect of climate change on wildfires can be fully reversed by hard edits and soft search. We find the Reconcile algorithm typically resolves most disagreements between models, but more often in favor of the manipulator when the opposing model is chosen adversarially compared to at random. We map which manipulations can plausibly pass existing regulations and disclosure policies, and recommend mitigations such as pre-registration, proof of model training, sensitivity analyses, multi-metric evaluations, and committing to versioned public models.
[LG-65] When Scores Conflict with Preferences: Pre-Alignment Diagnosis and Adaptive Correction of Constraint Drift in DPO NEURIPS2026
链接: https://arxiv.org/abs/2611.07375
作者: Ping Liu,Yan Yan
类目: Machine Learning (cs.LG)
*备注: Accepted to NeurIPS 2026. Corresponding author: Yan Yan
Abstract:Direct Preference Optimization (DPO) is increasingly used in settings where pairwise preferences coexist with pointwise criteria such as response length, answer correctness, or safety scores. When the two disagree, standard DPO treats pairs whose preferred output scores worse on the pointwise criterion (conflict pairs) in the same way as the remaining (matching) pairs. We characterize a systematic failure mode, constraint drift: for sufficiently small updates, conflict pairs tend to increase constraint cost while matching pairs tend to repair it, and the net change depends on their balance. This balance yields a closed-form pre-alignment diagnostic, \rho^_\mathrmeq , computable from dataset statistics, that indicates whether a dataset is self-correcting or requires intervention. Building on this diagnostic, we propose Conflict-aware DPO (CoDPO), which down-weights conflict pairs with a strength set from \rho^\mathrmeq rather than tuned, and adds an adaptive margin shift that addresses residual drift during training. Empirically, drift grows monotonically with conflict exposure on the conflict-dominated domains, and the intervention decision given by \rho^*\mathrmeq matches the observed drift on every domain, including one that it identifies as self-correcting. On all four domains of the main comparison, CoDPO reduces drift while largely preserving the preference learning of vanilla DPO. Compared with hard filtering and recent constrained and multi-objective baselines, CoDPO also achieves a more favorable drift-preference trade-off.
[LG-66] Asymptotic-Preserving Tensor Neural Networks for 3D High-Field Kinetic Equations with Applications to Uncertainty Quantification
链接: https://arxiv.org/abs/2611.07364
作者: Shi Jin,Tian-ai Zhang
类目: Numerical Analysis (math.NA); Machine Learning (cs.LG)
*备注:
Abstract:Kinetic equations in the high-field regime describe particle transport dominated by electric-field acceleration and collisional relaxation. These equations are widely used in simulations of plasmas and semiconductor devices. Typical models include the Vlasov–Poisson–Fokker–Planck (VPFP) and Boltzmann–Poisson systems. Their high dimensionality and multiscale nature pose two of the major challenges to accurate and efficient numerical solution. To address these challenges, we propose an asymptotic-preserving tensor neural network (APTNN) method. Specifically, to reduce high-dimensional computational costs and memory requirements, we adopt the canonical polyadic (CP) decomposition structure of TNNs to approximate their solutions. We show that the equation residuals retain this decomposition form, allowing high-dimensional loss integrals to be evaluated through combinations of one-dimensional integrals. The resulting computational and storage costs are reduced from exponential to polynomial growth with dimension. To improve the prediction accuracy in the high-field limit, we enhance the baseline AP loss by incorporating the first moment equation. The added loss term directly constrains the limiting flux closure to reduce its approximation error in the high-field regime. The enhanced AP loss can still be evaluated dimension-wise. Numerical results show that APTNN improves prediction accuracy in the high-field regime with significantly reduced computational costs and memory usage. For a 3D3V VPFP problem with multiple random inputs, training takes less than 30 minutes on a single GPU.
[LG-67] Acoustic Anomaly Detection for Early Identification of Drill Jamming
链接: https://arxiv.org/abs/2611.07356
作者: Maria Beatriz Silva,Lucas Edson Silva de Araújo,Sergio F. Chevtchenko,Ermeson Andrade,Danilo Araujo,Saeed Afshar
类目: Machine Learning (cs.LG)
*备注: Accepted for publication at IEEE INDIN 2026. 8 pages, 6 figures. © 2026 IEEE. Personal use of this material is permitted
Abstract:In the context of Industry 4.0, predictive maintenance is essential to minimize unplanned downtime and ensure operational safety. In machining processes, drill jamming is a critical failure that compromises both workpiece quality and tool integrity. This work proposes an acoustic anomaly detection system using ultrasonic sensors to identify pre-failure patterns in Computer Numerical Control (CNC) machining. We release a public dataset of time- and frequency-domain acoustic features from 17 drills, and benchmark semi-supervised autoencoder-based normality modelling against supervised XGBoost classification using Leave-One-Drill-Out validation. The results demonstrate that the Autoencoder is highly effective for industrial safety, achieving a Recall of 1.00 and a consolidated F1-score of 0.81. Furthermore, the threshold-independent analysis shows that the Autoencoder better separates normal and pre-failure acoustic patterns, achieving a consolidated AUC of 0.8294, compared with 0.7559 for the supervised model under inter-tool covariate shift. In contrast, the supervised XGBoost achieved an average F1-score of 0.55 and failed to detect failures in some of the evaluated drills. The findings highlight that the ultrasonic spectrum provides a reliable window for identifying incipient mechanical instabilities, and the AE model offers a robust safety envelope by anticipating all jamming events across the tested drill bits.
[LG-68] AtlasULP: Domain-aware Universal Link Prediction via Relation Atlas
链接: https://arxiv.org/abs/2611.07347
作者: Yujing Liu,Yixin Liu,Yu Zheng,Lianhua Chi,Alan Wee-Chung Liew,Heng Tao Shen,Shirui Pan
类目: Machine Learning (cs.LG)
*备注:
Abstract:Link prediction (LP) is a widely applied task in graph learning. Conventional LP approaches typically follow a dataset-specific paradigm, requiring independent training for each graph and incurring high computational and maintenance costs in large-scale applications. Motivated by these limitations, recent work explores Universal Link Prediction (ULP), aiming to enable training-free inference on arbitrary unseen graphs. However, existing methods primarily rely on subgraph sampling strategies to construct universal representations, which are unable to capture higher-order information due to the exponential growth of higher-order neighborhoods. As a result, they often under-perform on graphs where long-range signals are essential, such as biological networks. Moreover, their prediction models are typically domain-agnostic and lack the ability to adapt to different graph domains, leading to limited generalization. To address these challenges, we propose an Atlas-based Universal Link Prediction framework (AtlasULP), which introduces the relation atlas to capture structural associations between node pairs from both global and local perspectives, enabling a comprehensive characterization of connectivity signals across diverse graph domains. Building upon this representation, we develop a domain-guided prompting ULP model, which generates domain-aware prompt tokens from contextual structures and performs prompt-guided in-context prediction for adaptive link inference. Extensive experiments demonstrate that AtlasULP consistently outperforms state-of-the-art methods across diverse graph domains.
[LG-69] Rethinking Data Selection: The Importance of Coverag e over Difficulty in LLM Fine-Tuning
链接: https://arxiv.org/abs/2611.07323
作者: Lalchand Pandia,Kanishka Misra,Allyson Ettinger
类目: Machine Learning (cs.LG)
*备注:
Abstract:Selecting high-quality training data can reduce computation cost for LLM fine-tuning. Prior work on data selection developed a variety of difficulty-based scores, motivated by the intuition that harder examples are more informative. While effective for smaller classification models, we find that such methods often fail to outperform random selection for LLM fine-tuning. Our analysis reveals that random selection performs better because it is more representative of the underlying input space—i.e., has better coverage. To validate the importance of coverage, we implement a simple clustering-based selection method which selects data that is more representative of the underlying input distribution. On a set of diverse single-domain tasks, this method yields performance at or above that of random selection. On the diverse, multi-domain Tülu 3 mixture, where random sampling no longer guarantees a balanced subset, our method outperforms all methods. Overall, our results indicate that prioritizing coverage over difficulty leads to more effective data selection for LLM fine-tuning.
[LG-70] WeightLoop: Weight-Space Programming with Experimental Feedback
链接: https://arxiv.org/abs/2611.07319
作者: Mengran Li,Bo Li,Weijing Zhao,Chengyang Zhang,Boyu Zhang,Jinfeng Xu,Zelin Zang,Victor C. M. Leung,Stan Z. Li
类目: Machine Learning (cs.LG)
*备注: 5 pages, 2 figures, 3 tables
Abstract:Weight-space programming (WSP) shows that language models can reason about and rewrite neural network weights; recursive self-improvement (RSI) motivates feedback-driven improvement. Models guiding experiments receive concrete feedback: completed assays reveal prediction errors and inform what to test next. We introduce WeightLoop, a framework that turns experimental feedback into persistent updates to deployed predictors. GPT-6 Astra inspects a task specification and the model implementation, then writes a bounded, reusable parameter-editing program. As new measurements arrive, the program identifies prediction discrepancies, validates candidate edits using observed outcomes, and writes the selected changes back to the original model. The editor remains fixed while the predictor evolves, enabling repeated updates without further LLM calls or gradient retraining while preserving the original inference interface. In offline replay, targeted edits improve mean ranking utility in seven of eight endpoint-budget settings (all eight versus random edits) and mean final discovery across three sequential tasks. WeightLoop connects WSP with an RSI-inspired improvement loop, making experimental evidence a direct driver of model change and subsequent decisions.
[LG-71] Provably Efficient Representation Learning for Low-Rank CMDPs
链接: https://arxiv.org/abs/2611.07295
作者: Kaixuan Liu,Guojun Xiong,Shengpu Tang,Wanyun Si,Jian Li
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study representation learning in low-rank Constrained Markov Decision Processes (CMDPs), where both value and reward functions are approximated by a set of unknown representation vectors, and the transition dynamics admit a low-rank factorization. The objective is to maximize expected cumulative rewards while complying with constraints on the expected cumulative utility. To achieve this, we propose \textttMFRLC (Model-Free Representation Learning for Low-Rank CMDP), a model-free algorithm that uses a primal-dual scheme to effectively balance reward regret and constraint violations. To the best of our knowledge, \textttMFRLC is the first representation-learning method for low-rank CMDPs that combines the Least-Squares Value Iteration with Upper Confidence Bound (LSVI-UCB) and primal-dual techniques. \textttMFRLC further enhances value function estimation with bonuses for function approximation and learns the mapping from representations to value functions directly, avoiding explicit estimation of the transition kernel. We prove that, with high probability, \textttMFRLC achieves \widetilde O(K^3/4) reward regret and cumulative constraint violation over K deployment rounds, independent of the size of the state space. To our knowledge, these are the first sublinear regret and violation guarantees for CMDPs with unknown features.
[LG-72] How do Transformers Perform Low-Level Control In-Context?
链接: https://arxiv.org/abs/2611.07291
作者: Ebonye Smith,Aidan Andrews,Phoenix Wilson,Alexis Frias,Ayush Pandey,Gireeja Ranade
类目: Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 6 pages, 5 figures, Accepted to the 65th IEEE Conference on Decision and Control (CDC 2026)
Abstract:One of the remarkable abilities of Transformer-based neural networks is in-context learning. This motivates us to think about “in-context” control. Can we control a system knowing only (1) that it belongs to a particular family of systems and (2) some data on system evolution (provided as the context trajectory) but no specific parameter information? Specifically, we investigate the ability of Transformer-models to perform “in-context” control of unstable, nonlinear dynamical systems, where context consists only of the unforced system dynamics. We leverage a multi-head, GPT-2-style architecture, and we evaluate performance for two classical nonlinear systems – the cartpole system and the two-link acrobot system, where the system parameters are unknown to the Transformer. We consider a setting where trajectories from an expert controller are available at training time but not at inference time. Here, we identify training techniques that are essential to train Transformers to perform low-level control. Through an analysis over checkpoints, we show that in-context control is possible, and the learned policy performance generally improves as context increases. Furthermore, the Transformer models can generalize to stabilize systems that are out-of-distribution.
[LG-73] Scenario Augmentation for Rare Financial Events via Multimodal Latent Diffusion NEURIPS2026
链接: https://arxiv.org/abs/2611.07282
作者: Mahdi Arab Loodaricheh,Anita Raja,Indranil SenGupta
类目: Machine Learning (cs.LG)
*备注: 7 pages, 3 figures, 2 tables. NeurIPS 2026 - MLxOR
Abstract:Extreme market movements are scarce, heterogeneous, and difficult to learn from imbalanced historical data. We study whether generative augmentation can improve the representation and classification of CVaR-labeled rare financial windows while retaining the complementary structure exposed by different financial representations. Each market window is expressed through three aligned modalities: structured market variables, a spatial image transformation, and a sequence of portfolio returns. Separate variational autoencoders map the modalities to compact latent codes, and a class-conditional diffusion model learns their joint latent distribution. Conditional Value-at-Risk (CVaR) provides a risk-aware definition of rare windows and directs generation toward the minority regime. Across four window lengths on U.S. technology equities, the resulting synthetic samples improve fidelity and support relative to tabular diffusion baselines. An XGBoost classifier trained with real and generated samples attains AUCs of 0.953 - 0.960, exceeding both tabular generators and the corresponding real-only baselines. These results show that multimodal latent generation can provide useful synthetic stress scenarios for data-scarce financial risk models.
[LG-74] Graph Neural Networks for Resource Allocation in Integrated Sensing and Communication
链接: https://arxiv.org/abs/2611.07277
作者: Junyu Liao,Lili Chen,Tian Han,Jamie Evans
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注: Submitted to IEEE ICC 2027
Abstract:In vehicular integrated sensing and communication (ISAC) networks, each service-provider vehicle uses a limited power budget to serve multiple sensing targets and communication receivers. Maximizing the communication sum-rate subject to sensing signal-to-interference-plus-noise ratio (SINR) constraints is nonconvex, and iterative optimization is costly to rerun for every network layout. This paper proposes a graph neural network (GNN)-based power allocation algorithm, termed GNN-ISAC. Concretely, GNN-ISAC addresses the aforementioned problem by considering two distinct edge types in the graph representation, one capturing the inter-provider interference, the other representing the intra-provider power-budget sharing. GNN-ISAC is trained without labels using a sensing-penalized sum-rate loss. We also adapt the weighted mean-square error minimization (WMMSE) algorithm to this problem by introducing a sensing-power reservation step. Simulation results show that GNN-ISAC achieves communication sum-rates close to WMMSE with low sensing-link violation rates and much less computation. It also outperforms an existing heterogeneous GNN and a channel-blind baseline, and generalizes to larger networks without retraining.
[LG-75] Energy Measurement and Optimization of TinyML in Resource-Constrained Embedded Systems: A Systematic Mapping Study
链接: https://arxiv.org/abs/2611.07264
作者: Clariele Almeida,Renan Beserra,Rafael Moura,Danilo Araújo,Bruno Nogueira,Sérgio Chevtchenko,Ermeson Andrade
类目: Machine Learning (cs.LG); Hardware Architecture (cs.AR); Emerging Technologies (cs.ET)
*备注: 40 pages, 10 figures, 8 tables
Abstract:The rapid expansion of the Internet of Things (IoT) has increased the volume of data generated at the network edge, where many devices operate under severe memory, processing, and energy constraints. In this context, Tiny Machine Learning (TinyML) enables machine learning (ML) tasks to be executed on microcontrollers, smart sensors, and other resource-constrained platforms. However, the literature remains heterogeneous regarding the algorithms evaluated, energy-measurement methods, reported metrics, influencing factors, and optimization strategies. This study presents a systematic mapping of energy-aware TinyML research. It identifies the main models and learning approaches, classifies energy-evaluation methodologies and metrics, examines the technical factors affecting energy consumption, and organizes the reported optimization techniques into a structured taxonomy. The findings highlight the frequent use of convolutional and lightweight architectures, direct measurements and analytical estimations, and factors related to model architecture, hardware, communication, numerical precision, and input characteristics. The optimization strategies were grouped into model compression, hardware-aware design, system and communication optimization, and data and processing optimization. The results also reveal limited standardization in experimental platforms, measurement procedures, and reporting practices, limiting reproducibility and direct comparison across studies. This mapping provides an updated overview of the field and guidance for developing and evaluating more energy-efficient TinyML solutions.
[LG-76] Reweighting is not Harmless: Understanding the Role of Dataset Characteristics
链接: https://arxiv.org/abs/2611.07258
作者: Amalia Ionescu,Jose G. Hernandez,Emily F. Wong,Anil K. Saini,Jason H. Moore,Tiffani J. Bright
类目: Machine Learning (cs.LG)
*备注:
Abstract:Reweighting is a widely used pre-processing method for promoting discrimination-free model behavior by assigning different weights to data points during model training. While it is often credited with improving fairness, particularly by mitigating systematic underestimation for underrepresented groups, it can also introduce unintended harms under certain data conditions. Using controlled simulations that vary group size and baseline event prevalence, we show how reweighting can yield misleading improvements in fairness by reducing underestimation for one group while introducing it for another. Across 3,000 simulated datasets with event prevalence of 7-14%, reweighting improved parity in positive classification and false negative rates (FNR) between groups, but did so partly by decreasing positive classifications and increasing FNR for the group with higher baseline prevalence. These effects were most pronounced when that group was a minority: its FNR rose by 6 percentage points, approximately 82 additional missed cases per 10,000 group members, and roughly three-quarters of the improvement in FNR parity came from this increase rather than from gains for the other group. The pattern held across six classifiers, from logistic regression to neural networks, while overall accuracy was essentially unchanged, leaving the shift invisible in aggregate performance metrics. These findings mirror patterns reported in real-world applications and underscore the critical role of data characteristics in shaping the impact of this fairness intervention. While reweighting may create desirable outcomes in some applications, practitioners should critically evaluate it relative to their problem area and goals. Specifically, when the goal is to reduce underestimation in one group without introducing underestimation in another, reweighting may not be the appropriate approach.
[LG-77] Using Transformers for Teaching and Learning via Symbolic Distillation of Neural A-life Agents
链接: https://arxiv.org/abs/2611.07254
作者: Shuwei Wang,Vadim Bulitko,Matthew R. G. Brown,Michael Youngblood,Ramon Lawrence,Roscoe Smith,Shinichi Nakagawa,William Yeoh
类目: Machine Learning (cs.LG); Symbolic Computation (cs.SC)
*备注:
Abstract:Neural networks controlling behavior of A-life agents can evolve over multiple generations. While effective in making the agents survive, such networks are generally opaque to a human observer. Furthermore, the neural weights cannot be transferred between agents with different neural architectures. Distilling such neural networks into compact symbolic policies can help humans understand agent behavior in A-life. Additionally, it allows the agents to share their knowledge in a compact fashion, improving population-level learning. Distilling a neural policy into a symbolic policy is typically performed by searching the space of symbolic policies to find one that approximates the neural policy. However, even with a domain-specific language the space of policies can be large, necessitating incomplete stochastic search such as genetic algorithms. Yet the distillation remains computationally expensive. We propose to instead use a transformer pre-trained on readily available random percept-action data. Doing so speeds up distillation substantially. We derive the transformer architecture via a neural architecture search. Empirical results compare different distillation methods and the influence of the symbolic space on A-life population dynamics through self-reflection and teaching and learning.
[LG-78] Neural Compression over Equivalence Classes: Exact Gauges and Functional Rate-Distortion
链接: https://arxiv.org/abs/2611.07233
作者: Toshiaki Koike-Akino
类目: Machine Learning (cs.LG)
*备注: 38 pages, 7 figures
Abstract:Neural networks admit many parameterizations of the same or nearly the same function, but deployment typically stores whichever parameter vector training happens to produce. We show that this representational freedom can itself be used for compression. Rather than redesigning the codec, we search over exact function-preserving transformations and controlled functional perturbations, using the unchanged physical compressor itself to select cheaper representations. Exact gauge families provide a zero-distortion reference under the strict audit; allowing small functional changes exposes much larger rate-distortion gains. Across SmolLM2 and Qwen3, our method reduces Zstandard-compressed checkpoint size by up to 23.4%, with similar relative reductions under gzip, Brotli, and ZipNN, establishing trained-network representational freedom as a practical storage resource.
[LG-79] Loopback Transformer: Towards a Unified View of Residual Looped Transformers
链接: https://arxiv.org/abs/2611.07226
作者: Xuan Ouyang,Zefan Cai,Chenlong Zhang,Cheng Luo,Ke Wan,Junjie Hu
类目: Machine Learning (cs.LG)
*备注: 31 pages, 9 figures. Code: this https URL ; Project page: this https URL
Abstract:Looped Transformers reduce parameter count by repeatedly applying shared blocks, but parameter sharing alone does not determine how the results of repeated computation are reused. We take the depth-memory view of residual connections and analyze which representation is stored after each block application and how later reads weight it. A full block output contains both its input and its new update, so weighting that output also reweights the history already carried by the input. Reading individual updates instead, with the shared stack’s input kept on a separate path, gives separate controls over input strength, total update weight, and allocation across updates. The Loopback Transformer implements these controls: geometric decay sets the total update weight and supplies a recency prior, and attention learns how to allocate that weight. On ClimbMix, Loopback reaches lower validation loss than standard Transformer baselines with 40-75% of their Transformer parameters (excluding token embeddings and the output head) at reference scales from 191M to 3B, with training tokens and executed depth matched. It also has higher mean accuracy over thirteen downstream tasks at 1.4B and 3B. At 191M, 374M, and 778M, it has the lowest validation loss and the highest thirteen-task mean among the standard Transformer and seven looped residual designs.
[LG-80] Fractional physics-informed neural networks to model social media virality
链接: https://arxiv.org/abs/2611.07215
作者: Alex Koong,George Em Karniadakis
类目: Machine Learning (cs.LG)
*备注:
Abstract:Existing models of social media virality either prescribe the functional form of engagement growth or assume a fixed mechanism by which past activity influences future engagement. We present a fractional physics-informed neural network that learns both the temporal memory and nonlinear growth dynamics of social media engagement, using symbolic regression to translate the learned dynamics into interpretable equations. Synthetic experiments at the deployment resolution of 24 hourly observations demonstrate that the fractional order and reaction law can be jointly recovered from clean data. Under observation noise, however, the standard pointwise physics residual drives the fractional order toward its lower bound through a scale artifact. We address this failure using a fully whitened residual that propagates observation noise through the fractional operator. Applying the framework to 348 posts from X (formerly Twitter), we find that a single-window model estimates strongly memory-dependent dynamics but fails to accurately reconstruct held-out trajectories. Partitioning the 24-hour window into four teacher-forced segments substantially improves performance, yielding segment-specific fractional orders that increase from 0.23 to 0.99 and interpretable growth laws for each phase. The segmented model reconstructs 64% of held-out trajectories within a mean squared error of 0.01, compared with 44% for the strongest of nine classical baselines. Together, these results demonstrate the potential of fractional physics-informed equation discovery to uncover interpretable engagement dynamics while highlighting the importance of noise-aware estimation and temporal segmentation.
[LG-81] Canonical Correlation Neural Networks
链接: https://arxiv.org/abs/2611.07212
作者: Andrea Cavallo,Sundeep Prabhakar Chepuri,Charilaos Kanatsoulis,Elvin Isufi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Relational inductive biases such as grids or graphs allow neural networks with fewer trainable parameters and improved generalization, but require a known structure beforehand. When no structure is available, the covariance matrix provides a principled substitute, and coVariance Neural Networks (VNNs) have shown that using it yields models that are theoretically grounded in Principal Component Analysis, stable to finite-sample estimation errors and parameter-efficient. This framework, however, is limited to single-view data, whereas many applications observe the same phenomenon through multiple views, each with its own in-view covariance and a cross-view covariance linking them. To incorporate these matrices as inductive biases in a principled way, we introduce CAnonical correlation NEural networks (CANEs), architectures that use in-view and cross-view covariances as distinct relational biases to solve downstream tasks. Via spectral analysis, CANEs draw analogies with Canonical Correlation Analysis (CCA), a classical technique to project two-view data onto a correlated subspace. However, while CCA’s objective is to compute shared view representations, CANEs are trained to solve a downstream task directly. We bound their output error under finite-sample covariance perturbations and show that it depends on controllable Lipschitz constants of the learned filters in place of the singular-value gap that governs the stability of CCA, thus guaranteeing better robustness. Experiments on seven datasets from brain imaging and motion sensing show that CANEs match or outperform single-view, two-view and CCA-based baselines, often with fewer parameters, and retain their performance when covariances are re-estimated from fewer samples.
[LG-82] Structured Machine Theory of Mind from Agent Trajectories
链接: https://arxiv.org/abs/2611.07208
作者: Nigel F. Doering,Rahath Malladi,Arshia Sangwan,M Tanjid Hasan Tonmoy,Jacob Doan,David Danks,Tauhidur Rahman
类目: Machine Learning (cs.LG)
*备注: Published in Transactions on Machine Learning Research (TMLR), 10/2026. 37 pages. OpenReview: this https URL . Code: this https URL
Abstract:Predictive models of human behavior trained on large-scale trajectory data optimize for statistical accuracy without representing the mental states that causally generate behavior. Such models support prediction but not principled intervention: they cannot answer how an agent’s behavior would change if its beliefs or preferences were different. We introduce Structured Machine Theory of Mind (SMToM), a framework that attributes explicit, independently supervised belief and desire representations from observed trajectories within a Belief-Desire-Intention (BDI) causal structure. The central architectural element is a goal head that consumes only the predicted mental-state channels and a current-trajectory embedding; counterfactual intervention on beliefs and desires is a direct operation. We instantiate SMToM on a controlled pedestrian navigation domain where ground-truth mental states are known by construction, enabling evaluation of attribution accuracy, local channel interventions, and trajectory-level counterfactual diagnostics. The resulting model, BDIBottleneck, improves over trajectory-only and unsupervised-context neural baselines. It is also competitive with a matched auxiliary-supervised context baseline while providing an editable mental-state interface that the unbottlenecked baseline lacks. Counterfactual experiments on desires confirm that substituting an agent’s inferred preferences with a different activity type coherently shifts predicted destinations toward relevant locations. Counterfactual experiments on beliefs confirm that marking a location as unavailable in the agent’s belief state reduces its predicted probability as a destination, though in a complex way. These results demonstrate, in a controlled navigation setting, that explicit BDI-structured supervision is a viable foundation for interpretable and counterfactual analysis of longitudinal trajectory data.
[LG-83] rustMix: Graceful Degradation for Pretrained Models under Distribution Shift
链接: https://arxiv.org/abs/2611.07206
作者: Juliusz Ziomek,Maike Osborne,Yee Whye Teh
类目: Machine Learning (cs.LG)
*备注:
Abstract:Predicting beyond the training domain necessarily requires assumptions about how the ground truth extrapolates, yet in most neural networks these assumptions are made only implicit through the architecture. We propose TrustMix, a lightweight post-hoc method that makes a pretrained model’s extrapolative behaviour explicit and controllable. TrustMix combines the predictive distribution produced by a pretrained model with a user-specified fallback distribution p_0(y|x) , such that as the out-of-distribution (OOD) deviation statistic grows, we collapse towards p_0(y|x) . TrustMix provides graceful degradation: the predictive distribution converges toward the fallback distribution as the OOD signal grows. The fallback distribution p_0(y|x) can reflect domain knowledge, if available, or rely on a domain-agnostic stationarity assumption. TrustMix does not require retraining of the base model and can be applied to both regression and classification problems. We evaluated TrustMix on several synthetic and real-world distribution shifts, including tabular, image, and natural language processing tasks. We tested TrustMix on datasets such as CIFAR-10/100 and ImageNet and on numerous architectures including Convolutional Neural Networks (CNNs) and Transformers. TrustMix remains competitive in-distribution while improving predictive uncertainty under distribution shift, with gains in predictive likelihood for regression and calibration for classification. Finally, we showed that TrustMix can provide useful predictive distributions for decision-making applications by utilizing it as a surrogate for Bayesian Optimisation.
[LG-84] Backbone-Informed Adaptation after EEG Pretraining: A Capacity-Controlled Study
链接: https://arxiv.org/abs/2611.07192
作者: Mingyang Jiang,Yamin Li,Fan Ma,Daniel Moyer,Hua Xu,Catie Chang
类目: Machine Learning (cs.LG)
*备注: 25 pages, 6 figures. Code available at this https URL
Abstract:Electroencephalography (EEG) foundation models learn representations that combine information across channels and time. This structure motivates adapters that model the same relationships during downstream learning. However, it is not clear whether such a design outperforms other candidate adapters. Moreover, in order to arrive at a fair comparison between candidate adapters, we also need to consider their trainable parameters. Here, we ask whether a backbone-informed (native) adapter makes better use of a given capacity by comparing it against a token-wise multilayer perceptron control that transforms each feature vector independently. We approximately match their trainable parameter counts and keep the pretrained encoder fixed. Across two EEG foundation models and tasks covering emotion recognition, sleep staging, and EEG event classification, the native adapter shows no consistent advantage. For one backbone, differences between the designs are small relative to variation across runs. For the other, the token-wise control performs better on emotion recognition and EEG event classification. On sleep staging, the native adapter’s advantage over the token-wise control is reduced when we also train the component that processes sequences of EEG segments. The relative performance of the two adapters depends on the backbone, task, and downstream training protocol. These findings motivate evaluating backbone-informed adapters against capacity-matched alternatives.
[LG-85] Meta-Frieren: Beyond Single Level Online Learning
链接: https://arxiv.org/abs/2611.07169
作者: Vahan Arsenyan,Arnak Dalalyan
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study a two-stage online optimization problem: at each round, a meta-learner first selects an agent-environment pair, then the chosen agent interacts with the environment, and receives a reward. We introduce Meta-Frieren, a general framework for maximizing cumulative reward in this setting. Our analysis yields a simple UCB-style criterion for meta-level selection and recovers prior guarantees as special cases while extending them to broader settings. We instantiate Meta-Frieren with GP-UCB under Bayesian and frequentist assumptions. We also derive, to our knowledge, the first high-probability regret bound for Feel-Good Thompson Sampling, specialize it to finite and parametric function classes, and integrate it into Meta-Frieren. Experiments support our theoretical results and demonstrate that Meta-Frieren can be successfully applied to hyperparameter optimization across multiple model classes.
[LG-86] Structural Conditional Independence Test via Neural Quantile Resolution ICDM2026
链接: https://arxiv.org/abs/2611.07165
作者: Khoa Nguyen,Bao Duong,Thin Nguyen
类目: Machine Learning (cs.LG)
*备注: Accepted at IEEE ICDM 2026. Code: this https URL
Abstract:Conditional independence testing is difficult when dependence is weak, localized, or expressed through distributional changes rather than a global mean shift. Existing tests usually return a scalar statistic or p-value, which summarizes evidence against independence but does not describe where the conditional law changes. We propose QRCIT, a quantile-resolution conditional independence test that compares P(X|Z) and P(X|Y,Z) through a quantile deformation learned by a shared model. A lightweight residual quantile-flow model first learns the baseline conditional quantile law and then learns the additional deformation induced by Y, yielding a global predictive-gain statistic with conditional-randomization calibration and an inspectable deformation field over context and quantile level. We also introduce the Structural Dependency Benchmark (SDB), which evaluates dependence structures motivated by real applications, such as context-local, tail-only, sign-reversing, moving-ridge, and mean-neutral patterns. Across synthetic post-nonlinear (PNL) and SDB benchmarks, QRCIT improves detection of structured alternatives, with empirical Type-I behavior evaluated across settings. Higher quantile resolution improves merged-SDB power at moderate sample sizes, while diagnostics restricted to known affected regions improve detection of localized dependence. On Sachs, METABRIC breast cancer, and CCLE/PLX4720, QRCIT provides competitive real-data performance and interpretable deformation patterns. These results support quantile-resolved conditional-law deformation as a practical framework for testing and diagnosing structured conditional dependence.
[LG-87] A Survey of Nonlinear Filtering Methods for Data Assimilation and Recent Advances in Generative AI-Enabled Score Filters
链接: https://arxiv.org/abs/2611.07160
作者: Cath Pham,Kadesha Reynolds
类目: Machine Learning (cs.LG); Probability (math.PR)
*备注:
Abstract:Many real-world applications, such as numerical weather prediction or wildfire tracking, involve modeling highly nonlinear dynamical systems that are chaotic and influenced by various sources of uncertainty. To improve model prediction accuracy and reduce such uncertainties, data assimilation (DA) combines the dynamical model with partial, noisy observations to estimate the evolution of the system’s state. Doing so depends on solving the optimal filtering problem: computing the conditional distribution of a hidden state given a sequence of observations. Classical approaches fail along three distinct axes: Kalman-based methods assume Gaussian posteriors and produce biased estimates under nonlinear observation operators; particle-based methods represent distributions with weighted samples and collapse in high dimensions due to the curse of dimensionality; and kernel-based methods move beyond discrete particle representations but cannot maintain their density approximation affordably as dimension grows. The Ensemble Score Filter (EnSF) avoids all three constraints by encoding the filtering density in a score function and approximating that score with a training-free Monte Carlo estimator, enabling stable filtering in high-dimensional, chaotic systems. This survey traces EnSF’s mathematical foundations, establishes it as the state of the art for non-Gaussian, high-dimensional filtering, and examines fourteen extensions that expand its operational reach. Twelve of the fourteen extensions preserve this training-free Monte Carlo estimator at the core of the update, with only two departing from it to achieve planetary-scale deployment.
[LG-88] FastAFD: When Does Attention-FFN Disaggregation Pay on Rack-Scale NVLink?
链接: https://arxiv.org/abs/2611.07149
作者: Yichao Fu,Yuxuan Zhang,Minshen Zhang,Ruitian Wang,Junda Chen,Hao Zhang
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: 24 pages
Abstract:Attention-FFN disaggregation (AFD) separates attention from the experts of a mixture-of-experts (MoE) model and has been reported to raise decode throughput over colocated serving. Whether this advantage holds on rack-scale NVLink systems such as GB200 NVL72 remains unclear: prior results were obtained on other hardware, against narrower colocated baselines, or with analytical models, whereas production engines serve large MoE models on these systems with wide expert parallelism (wide EP). We build and tune FastAFD, an AFD engine for GB200 NVL72 that overlaps the MoE of one microbatch with the attention of the other, and compare it against the best vLLM wide EP setting. Experiments show that AFD has a throughput advantage of up to 1.17x over wide EP on Qwen3 and 1.16x on Qwen3.5, and matches it on MiniMax-M2.5. We derive an analytical model that predicts AFD throughput within 3% using only profiles of colocated vLLM, telling whether switching to AFD would pay off before deploying it. The model yields a concise insight: AFD’s speedup factors into a batch gain (larger batches, since attention GPUs hold no expert weights), an overlap gain (shorter steps, since the MoE runs on the expert GPUs behind attention), and a disaggregation cost (the GPUs given to experts). AFD has the advantage when the two gains outweigh the cost. This view explains which models benefit from AFD, and shows that speculative decoding helps light-expert models most, where the enlarged overlap gain outweighs the larger disaggregation cost and adds a further 8% to AFD’s advantage over wide EP. Looking ahead, serving the experts on accelerators that keep their weights in on-chip SRAM, such as the Groq LPX accelerators paired with Vera Rubin NVL72, would remove most of the disaggregation cost, raising the ideal speedup to 2.48x with speculative decoding. FastAFD is open source at this https URL.
[LG-89] aching Agents the “Principle” Not the “Reward”
链接: https://arxiv.org/abs/2611.07130
作者: Shan Yang
类目: Machine Learning (cs.LG)
*备注: 15 pages, 9 figures, 5 tables
Abstract:Reinforcement learning has achieved strong performance in many tasks by letting agents learn from nothing more than a reward signal. However, in more complex tasks, agents still struggle to learn good, generalizable policies, even with carefully designed rewards and abundant samples. We argue that this difficulty stems from the reward decoding burden: designers encode their knowledge of the desired behavior into rewards and penalties, and agents must infer this knowledge from long sequences of noisy scalars. To address this burden, we propose prescriptive policy learning (PPL), which directly teaches agents higher-level knowledge about the tasks, termed principles of behavior, such as how to maintain balance, how to accumulate energy, and how to land smoothly. Through appropriately selected key quantities and references, PPL can directly indicate at every step whether the agent follows the principles and to what extent. Our experiments show that agents can learn good, generalizable policies without any reward. Taught only simple heuristic principles, PPL agents outperform reward-based methods with far fewer samples. Their policies also transfer directly to harder settings and keep performing well, demonstrating their strong generalization ability. In summary, PPL provides a new approach in which rough knowledge of the desired behavior is enough for agents to learn good, generalizable policies, even when the system dynamics cannot be accurately modeled.
[LG-90] Scalable Bayesian Gaussian Process Classification for Cancer Histopathology: Calibrated Uncertainty Under Distribution Shift
链接: https://arxiv.org/abs/2611.07127
作者: Yunru Zheng,Wun Ting Chan,Sampras Dsouza,Zhenyu Hong
类目: Machine Learning (cs.LG)
*备注:
Abstract:Reliable uncertainty quantification is essential for clinical deployment of cancer diagnosis models, yet neural networks and variational Gaussian process methods can produce overconfident predictions, particularly under the distribution shifts that arise when models trained at one hospital are applied at another. We present a scalable Bayesian Gaussian Process Classification (GPC) framework that operates on frozen pathology foundation model embeddings (Phikon) and combines accelerated Randomly Pivoted Cholesky (RPCholesky) for low-rank kernel factorization, Hamiltonian Monte Carlo (HMC) with non-centered parameterization for posterior sampling, and a pivots-as-inducing-points scheme for prediction. This makes HMC posterior sampling in the low-rank latent representation tractable at N approximately 300,000 training points, enabling HC-based low-rank GPC on substantially larger datasets than prior MCMC-based demonstrations. We evaluate on two histopathology benchmarks for breast cancer metastasis detection: PCam and CAMELYON17-WILDS, where the latter tests calibration under hospital-level distribution shift. Compared to stochastic variational GP and neural network baselines, our method improves calibration and predictive performance - reducing expected calibration error by up to 88% and false negative rate by up to 90% - with the largest gains appearing under distribution shift, suggesting that HMC-based posterior sampling provides more trustworthy uncertainty estimates for cross-site pathology classification.
[LG-91] Data-Driven Modeling of Molten Chloride Corrosion Using Static Test Data
链接: https://arxiv.org/abs/2611.07125
作者: Md Mahmudul Hasan Pathik
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci)
*备注:
Abstract:Molten chloride corrosion has been studied for decades, yet reliable prediction across different alloys, salts, and test conditions remains difficult. In this work, 291 static molten chloride corrosion experiments reported in the literature were compiled into a structured dataset and used to develop an interpretable machine-learning framework for corrosion-rate predictions. To better represent the differences in salt condition across studies, a Chloride Salt Preparation Index (CSPI) was introduced to bring the reported salt-treatment method, atmosphere control, and pre-exposure impurity condition on a common numerical scale. A blended tree-based ensemble provided the most consistent predictive performance, achieving a cross-validated (R^2) of 0.85 and an MAE of 0.41 mg cm(^-2) day(^-1). The graded CSPI retained more useful information than a simple purified/unpurified classification and emerged as the dominant model descriptor. Global sensitivity and SHAP analyses further identified exposure time, temperature, and alloy chemistry as important contributors to the model predictions, with several observed trends consistent with the established molten-chloride corrosion behavior. The conformal prediction intervals were well calibrated, providing an estimate of prediction uncertainty in addition to the point prediction. Overall, the framework demonstrates that historical corrosion data can support quantitative, interpretable, and uncertainty-aware prediction when experimental context is represented explicitly, while also highlighting where more consistent reporting and targeted experiments are needed to improve future molten-salt corrosion models.
[LG-92] Machine Learning for Live-Birth Prediction After IVF/ICSI From Structured Clinical Data: A Systematic Review Meta-Analysis and Critical Appraisal
链接: https://arxiv.org/abs/2611.07124
作者: Abdul Raheem Shahzad,Muhammad Zaheer Khan,Umar Ajaib,Asad Ur Rehman,Juan Deng,Ming Zhang
类目: Machine Learning (cs.LG)
*备注: 20 pages, 4 figures, 6 tables; supplementary material (26 pages) in ancillary files. Data and code: this https URL . Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence
Abstract:Machine-learning models predicting live birth after in vitro fertilisation (IVF) or intracytoplasmic sperm injection (ICSI) from routine clinical data report performance from near chance to near perfect. We systematically reviewed these models under PRISMA 2020, searching OpenAlex, Scopus and Web of Science to 22 October 2026, appraised them with PROBAST+AI and TRIPOD+AI, and pooled their discrimination with random-effects meta-analysis on the logit scale. We define what fixes the meaning of a model’s output (decision point, endpoint, unit of analysis, data source) and map the 48 included studies by model family, decision point and validation design. Models evaluated only internally pooled to an area under the receiver operating characteristic curve (AUC) of 0.814, against 0.756 (95% prediction interval 0.546-0.889) for models evaluated on a later period or another centre, and 0.764 for studies not at high risk of bias; neither difference was significant, and heterogeneity was extreme. Models whose decision point was not stated pooled to 0.910, the only significant moderator in meta-regression. Against logistic regression fitted to the same data, the best machine-learning model gained a pooled 0.056 AUC (95% CI 0.017-0.090). Of 48 studies, 47 were at high or unclear risk of bias, only four assessed the calibration slope, and no model has evidence of clinical impact. The certainty of the pooled evidence is low to very low; the gains that survive validation came from staging predictions to decision points and local calibration, not deeper architectures. We provide a TRIPOD+AI-mapped checklist, a benchmark specification and a research agenda.
[LG-93] Server-Side Sparse Assignment for Federated Mixture-of-Experts
链接: https://arxiv.org/abs/2611.07093
作者: Xue Zheng,Aylin Yener
类目: Machine Learning (cs.LG); Information Theory (cs.IT)
*备注: Accepted for publication in the proceedings of the 2026 IEEE Information Theory Workshop (ITW 2026)
Abstract:Large language models (LLMs) are increasingly recognized for their remarkable capabilities; federated learning (FL) offers a promising framework for collaboratively fine-tuning LLMs on edge devices while achieving data privacy. However, edge devices are inherently heterogeneous: clients may exhibit diverse data distributions and resource constraints, including limited memory and computation budgets. Recent federated Mixture-of-Experts (MoE) approaches address these challenges by using sparse expert activation or partial expert loading. However, existing partial-expert methods rely on heuristic full-pool observations or prioritize system-wide load balance, without guarantees for recovering client-specific relevant experts from partial observations. We propose S^3 -MoE, a server-side sparse expert assignment framework that estimates client-specific expert relevance from partial observations and refines assignments through an exploit–explore mechanism. We also establish relevance-estimation error and exact Top- K recovery guarantees. Experiments using a CNN-based MoE on CIFAR-100 show that S^3 -MoE recovers task-dependent client–expert preferences and reduces conflicts caused by correlated expert selection across clients. Compared with Full-20 FedAvg, which gives each client access to all 20 experts for local Top-5 routing, S^3 -MoE achieves 1.31 percentage points higher mean final accuracy while loading only 5 experts per client.
[LG-94] Optimistic Hedge in General-Sum Games: Fundamental Limits and Regret Bounds via Temporal Approximation
链接: https://arxiv.org/abs/2611.07078
作者: Yuval Dagan,Nadav Eisen,Liad Erez,Tomer Koren
类目: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT)
*备注:
Abstract:Optimistic Hedge, a multiplicative weight update with one step of extrapolation, is one of the most natural regret-minimizing dynamics for games, and characterizing its regret has been a longstanding open question. Daskalakis, Fishelson and Golowich proved that if all m players in a finite general-sum game follow Optimistic Hedge with a fixed learning rate, each player’s regret is polylogarithmic in the horizon T . Subsequently, Ha showed that an appropriate choice of learning rate makes each player’s regret bounded independently of T . This bound, however, is non-constructive and its dependence on the number of actions n is not explicit, leaving open whether Optimistic Hedge can achieve a constant regret bound which is polynomial in the size of the game. We show that the answer hinges on the number of players. In two-player games, there is a learning rate for which Optimistic Hedge guarantees O(n\log n) regret independent of the horizon. In three-player games, however, no such bound is possible: for any learning rate up to an absolute constant and for sufficiently large horizon T , there is a game with n actions per player in which some player’s regret is 2^\Omega(n) , and more generally, for every m\ge3 we construct m -player games in which some player’s regret is 2^\Omega(mn) ; for learning rates above 2^-\Omega(mn) , that player’s regret even grows linearly in T . Our upper bound is based on a temporal approximation framework for the analysis of Optimistic Hedge. Starting from the variance-based RVU bound of Daskalakis et al., we approximate the losses over time-localized windows by well-behaved functions, whose total variation is a small fraction of their cumulative variance. In m -player games, the same framework gives a significantly simplified proof of the recent O(\sqrtm\log n\log T) regret bound due to Ha. Subjects: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT) Cite as: arXiv:2611.07078 [cs.LG] (or arXiv:2611.07078v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.07078 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-95] Stochastic Dynamics of Generative AI: From Autoregressive Models to Interacting LLM Networks and Neural SDEs
链接: https://arxiv.org/abs/2611.07065
作者: Vikram Krishnamurthy,Yiming Zhang,Luke Snow
类目: Machine Learning (cs.LG); Signal Processing (eess.SP); Systems and Control (eess.SY)
*备注:
Abstract:This paper develops a stochastic dynamical-systems perspective on three classes of generative AI models. First, we formulate large language models (LLMs) as high-dimensional nonlinear autoregressive models with attention-based dependencies. This formulation connects pretraining by conditional maximum likelihood to sequential token generation and provides a probabilistic interpretation of LLM operation. Second, we consider networks of interacting LLMs and characterize their collective behavior through mean-field ordinary differential equations. These dynamics provide a framework for understanding emergent preferential attachment and persistent structural biases, including type-dependent glass-ceiling effects. Third, we consider neural stochastic differential equations (SDEs) as continuous-time generative models for stochastic trajectories, including those observed at irregular time intervals. We discuss their training with particular emphasis on Wiener-space cubature methods based on rough-path analysis as an efficient alternative to Monte Carlo estimation. Together, the three settings address probabilistic prediction, information exchange and emergent behavior in interacting networks, and path-space generative modeling.
[LG-96] Robust Stream Classification using Time Neutralising Decision Trees NEURIPS2026
链接: https://arxiv.org/abs/2611.06991
作者: Brandon Gower-Winter,Georg Krempl
类目: Machine Learning (cs.LG)
*备注: 33 pages, 9 Tables, 10 Figures. Accepted at Neurips 2026
Abstract:Data Streams are characterised as a potentially infinite source of data subject to non-stationarity. This non-stationarity, also called Concept Drift, often results in deployed predictive models having to be perpetually retrained as the underlying data generating process of the stream changes over time. Drift Detectors are a class of methods used to identify these changes, triggering these retraining procedures. Historically, Stream Learning has operated under the assumption that instances are only seen once, and cannot be stored for later use. Contemporary work relaxes this assumption, yet drift detection and simple resetting procedures are still ubiquitous. Building on related work which has shown that traditional Batch Learning algorithms often outperform Stream Learning algorithms when instances can be stored, we introduce Time Neutralising Trees (TNT), a Decision Tree architecture that enables robust stream classification in settings with Concept Drift. During training, TNT will filter out (“neutralise”) old training samples that are no longer relevant given the current context of the stream. If those instances become relevant again, TNT will then re-include them. We evaluate TNT, and related algorithms on 24 real and semi-real classification data streams from the USP DS repository. Results show strong evidence that TNT achieves best-in-class performance on a wide-range of data streams with varying concept drift types. Comments: 33 pages, 9 Tables, 10 Figures. Accepted at Neurips 2026 Subjects: Machine Learning (cs.LG) Cite as: arXiv:2611.06991 [cs.LG] (or arXiv:2611.06991v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.06991 Focus to learn more arXiv-issued DOI via DataCite Journalreference: Advances in Neural Information Processing Systems - NeurIPS 2026
[LG-97] Uncertainty calibration in Bayesian M/EEG source imaging
链接: https://arxiv.org/abs/2611.06979
作者: Ismail Huseynov,Mohammad Orabe,Srikantan Nagarajan,Stefan Haufe
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)
*备注:
Abstract:M/EEG source imaging requires accurate localization and reliable uncertainty estimates, yet posterior credible regions are rarely assessed for empirical calibration. We present a simulation-based framework for uncertainty quantification and calibration of empirical Bayesian inverse methods with closed-form conditional Gaussian posterior summaries. We compare Bayesian minimum norm (BMN) with sLORETA depth weighting and sparse Bayesian learning (SBL) with sparse basis field expansions (S-FLEX). S-FLEX allows retained coefficients to contribute nonzero marginal posterior variance at locations where voxel-wise SBL yields degenerate variances. For BMN, we derive joint source- and noise-variance updates through evidence maximization. We test whether adaptive noise learning yields near-calibrated uncertainties. Both methods retain residual miscalibration, motivating post-hoc isotonic recalibration of credible-region levels using empirical coverage. Simulations assess fixed- and free-orientation models across signal-to-noise ratios (SNR), source sparsity levels, and subject-specific head models, using oracle, baseline-estimated, and adaptively learned noise. Before recalibration, SBL+S-FLEX yields lower Wasserstein localization error, smaller posterior uncertainty, and coverage closer to nominal credibility than BMN+sLORETA. Isotonic recalibration improves both methods, with larger gains for conservative BMN+sLORETA posteriors; transfer across mismatched SNR, sparsity, and head models remains condition dependent. Noise specifications yield broadly consistent reconstruction and uncertainty trends within each estimator. In real auditory MEG data, SBL+S-FLEX provides more focal localization and tighter uncertainties than BMN+sLORETA, suggesting greater statistical power for identifying bilateral auditory sources.
[LG-98] SteelAgent : LLM -Orchestrated Physics-Informed Steel Property Prediction with Audited Generalization and Calibrated Uncertainty IJCAI-26 IJCAI
链接: https://arxiv.org/abs/2611.06953
作者: Aleksandr Volkov,Roman Sultimov,Mikhail Kuzin,Yury Maximov
类目: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci)
*备注: 10 pages, 3 tables. A shorter version of this work was published in the Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence (IJCAI-26), Demonstrations Track, pp. 8547-8550 (2026), doi: https://doi.org/10.24963/ijcai.2026/1000 . Includes code for the synthetic experiment
Abstract:Steel property models can perform well on random test splits even when they extrapolate poorly to unfamiliar alloys. SteelAgent couples metallurgical descriptors and machine-learning models with a language-model interface for property prediction, uncertainty reporting and inverse design. We examine how the choice of held-out unit changes the conclusions drawn from SteelBench tensile-strength evaluations. For gradient boosting, pooled out-of-fold R^2 fell from 0.859 on a random split to 0.113 when steel families were held out. Empirical coverage of nominal 90% prediction intervals ranged from 91.0% under random splitting to 37.6% on held-out sources. The effect of physics-derived features depended on the evaluation. They increased gradient-boosting R^2 by 0.2534 in a 17-family holdout analysis and decreased it by 0.358 in a separate four-grade laboratory analysis. We report group-level errors and a synthetic regression experiment illustrating how calibration can deteriorate on held-out groups. These results motivate reporting the evaluation population alongside a property estimate or compliance score.
[LG-99] A Hybrid Framework for Explainable Aviation Risk: Causal Chain Extraction from Safety Reports with Physics-Based Failure Quantification and Interventional Analysis
链接: https://arxiv.org/abs/2610.35658
作者: Yutian Pang,Yifang Tian,Xinyu Zhao,Jueming Hu,John-Paul Clarke,Hans-Arno Jacobsen
类目: Machine Learning (cs.LG); Systems and Control (eess.SY); Applications (stat.AP)
*备注: 14 pages, 3 figures, 4 tables. v2 prints the carburetor icing scenario values to four digits, corrects wording and table formatting, and removes one author from the author list at his request. This work is going to be submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
Abstract:Aviation authorities hold decades of accident reports whose narratives record how investigators reconstructed each accident. Computational work on these archives has mostly sorted reports into fixed taxonomies. In this work, we propose a hybrid framework in which a text-based corpus identifies the leading factors of a failure mode. Physics-based models then quantify how likely that failure is at a stated operating point. Ordered causal chains extracted from 56,202 American accident reports with a local large language model aggregate into a causation knowledge graph. On an evaluation sample of 2,000 reports, the chains recover 90.8% of investigator-flagged causes. Each physical factor whose drivers the record carries receives an occurrence model calibrated to the accident population. A leaky noisy-OR fitted to the corpus combines the factors into a failure mode probability that supports intervention. For engine power loss, the carburetor icing model separates 1,205 icing accidents from other accidents using temperature and dewpoint alone, reaching 0.61 in area under the receiver operating characteristic curve against a density ratio ceiling of 0.64. Routine weather hours give no valid control, because flying concentrates in fair weather. On later years the physics adds 0.03 to that area for occurrence and nothing for severity, so its value lies in interpretability. On four named accidents, abduction gives a probability of necessity of 0.29 to 0.50 for the heat omission, under monotonicity. The Fast Causal Inference algorithm does not license the largest backdoor adjustment on the learned networks.
[LG-100] Efficient Kernelized Learning in Polyhedral Games Beyond Full-Information: From Colonel Blotto to Congestion Games NEURIPS2025
链接: https://arxiv.org/abs/2509.20919
作者: Andreas Kontogiannis,Vasilis Pollatos,Gabriele Farina,Panayotis Mertikopoulos,Ioannis Panageas
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注: Accepted at NeurIPS 2025
Abstract:We examine the problem of efficiently learning coarse correlated equilibria (CCE) in polyhedral games, that is, normal-form games with an exponentially large number of actions per player and an underlying combinatorial structure. Prominent examples of such games are the classical Colonel Blotto and congestion games. To achieve computational efficiency, the learning algorithms must exhibit regret and per-iteration complexity that scale polylogarithmically in the size of the players’ action sets. This challenge has recently been addressed in the full-information setting, primarily through the use of kernelization. However, in the case of the realistic, but mathematically challenging, partial-information setting, existing approaches result in suboptimal and impractical runtime complexity to learn CCE. We tackle this limitation by building a framework based on the kernelization paradigm. We apply this framework to prominent examples of polyhedral games – namely the Colonel Blotto, graphic matroid and network congestion games – and provide computationally efficient payoff-based learning algorithms, which significantly improve upon prior works in terms of the runtime for learning CCE in these settings.
[LG-101] Unifying Central Flow Rod Flow and the Edge Coupling
链接: https://arxiv.org/abs/2611.08358
作者: Eric Regis,Sinho Chewi
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:At the edge of stability, the iterates of gradient descent oscillate along the sharp directions of the Hessian. This poses a significant challenge for continuous-time models as the usual continuous-time approximation – the gradient flow – admits no oscillations. Several continuous-time models have been proposed to capture the behavior at the edge of stability, among them central flow and rod flow. Although both models can be interpreted as performing a coarse-graining in time, the connection between them remains poorly understood. Here we show that both flows can be derived from a third object, the edge coupling: a functional on consecutive iterates whose critical points are the fixed points and period-two orbits of gradient descent. It is revealed that the two flows correspond to coarse-grainings at different time scales. Rod flow coarse-grains over only the fastest time scale, while central flow coarse-grains over both the fast and intermediate time scales. Furthermore, using the edge coupling, we derive rod flows and central flows for momentum methods such as heavy ball and Adam. We verify the derived flows experimentally on representative neural network architectures, where they track the discrete iterates throughout all phases of training – including the transition from the stable regime to the edge of stability.
[LG-102] Series-wise Cross-validation for Global Time Series Forecasting
链接: https://arxiv.org/abs/2611.08252
作者: Vitor Cerqueira,Mariana Oliveira,Luis Roque
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:We consider the problem of cross-validation (CV) in time series forecasting, with a focus on global forecasting models based on deep learning. The prevailing CV strategy in forecasting is to partition data across time with the last part of each series used for testing. With global models trained on collections of time series, however, a fundamentally different approach becomes possible: partitioning across series and time. We formalize series-wise CV for global forecasting models and benchmark several instantiations, including K-fold CV, bootstrapping, and Monte Carlo CV, on 11 datasets with 10 neural network architectures. We evaluate CV procedures along five dimensions: performance estimation, overoptimism rate, model selection accuracy, regret, and downstream forecasting performance. We withhold a random fraction of series from training so that they appear only in the test set. This protocol simulates the emergence of new series not seen during training, a common scenario in global forecasting. Under this protocol, our results reveal a nuanced trade-off: series-wise CV yields better downstream forecasting accuracy by guiding model selection toward configurations that generalize across both time and series, while time-wise CV retains a slight advantage for estimating out-of-sample error.
[LG-103] Patch Attention Wisely: Understanding Patch Length in Attention-Based Forecasting
链接: https://arxiv.org/abs/2611.08160
作者: Junyu Chen,Likai Chen,Weining Wang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Econometrics (econ.EM)
*备注:
Abstract:Attention-based time series forecasters commonly group consecutive observations into patches. Patch length is known to have a substantial impact on forecasting performance, yet its selection remains largely empirical. We develop a theoretical framework that explains why patch length matters for forecasting accuracy and provides a principled basis for its selection. For linear state-space models, we derive an exact factorization of the forecasting coefficient matrix that separates forecast dynamics from filtering dynamics. We show that dominance of a single forecast component at long horizons does not generally imply a rank-one representation. This explains why longer horizons alone need not reduce representational complexity. We combine the factorization with estimation bounds to develop a patch-selection criterion that balances representation and estimation errors under temporal dependence. Our analysis establishes conditions under which balanced patching is justified and clarifies how predictive structure, head count, training sample size, and long-run dependence enter the selection criterion.
[LG-104] Identifying Neural Dynamics Using Nonlinear Interventional State Space Models
链接: https://arxiv.org/abs/2611.08113
作者: Hanna M. Tolle,Carles Balsells-Rodas,Yixin Wang,Pedro A. M. Mediano,Amin Nejatbakhsh
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 31 pages, 12 figures
Abstract:State-space models (SSMs) recover low-dimensional latent dynamics from high-dimensional observations, but latents are generally not identifiable from observational data alone. Interventional state-space models (iSSMs) use perturbations to break observational symmetries and identify parameters up to permutation, scaling, and shift. However, prior work restricts identifiability to linear latent dynamics and piecewise-linear emissions, excluding many behaviors characteristic of neural circuits. We extend iSSMs to a substantially broader model class: we prove that (i) affine identifiability holds under general injective emission functions and Gaussian latent transitions, and (ii) with interventions, identifiability up to permutation, scaling, and shift is preserved under general nonlinear latent dynamics. We validate our theory on simulated FitzHugh-Nagumo dynamics, and then demonstrate two applications: adjudicating between competing models of decision-making; and predicting responses to unseen micro-stimulation in macaque and human neural recordings, with proof-of-concept generalization to a held-out stimulation site in the latter. Finally, simulations yield practical guidance for designing stimulation protocols that improve identifiability.
[LG-105] Joint Modeling of Conditional Mean and Variability in Multilevel Functional Data: Application to Postprandial Glucose Trajectories in the AEGIS Study
链接: https://arxiv.org/abs/2611.07937
作者: Yang Xiao,Francisco Gude,Marcos Matabuena
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Applications (stat.AP); Methodology (stat.ME)
*备注:
Abstract:Functional data analysis provides a framework for studying time-varying associations between covariates and high-frequency biological signals. In digital health, participants often contribute repeated functional trajectories, yet existing multilevel regression methods largely focus on the conditional mean, overlooking variability and its role in biological heterogeneity. We propose a heteroskedastic functional mixed-effects regression framework that jointly models the conditional mean and variance of repeated trajectories, with participant-specific random effects in both components. We also develop a participant-level bootstrap to quantify uncertainty in estimated functional parameters. We apply the method to postprandial glucose trajectories from the AEGIS study. Premeal glucose, carbohydrate intake, and meal type show distinct time-dependent associations with glucose variability, whereas HbA1c is more strongly associated with the conditional mean than with variability. Substantial between-participant differences in residual variability remain even after accounting for observed covariates. These findings show that modeling conditional variability alongside the mean provides insights into glucose regulation and individual heterogeneity that mean-focused analyses may miss. More broadly, the proposed framework enables joint characterization of average responses, variability, and participant-specific heterogeneity in repeated biological signals.
[LG-106] Estimating Spectrum-Free Information Bounds for Generative Surrogates Beyond Conditional Collapse
链接: https://arxiv.org/abs/2611.07826
作者: Chaitanya Dumbre,Sourabh Dube
类目: High Energy Physics - Phenomenology (hep-ph); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex)
*备注: 20 pages, 7 figures
Abstract:Conditional generative surrogates in high-energy physics, such as models that predict reconstructed jets from partons, are often evaluated on 1\textD marginal divergences among other metrics. Taken alone these metrics can reward conditional collapse. On a CMS simulation benchmark, a sampler that ignores the conditioner beats a trained Mixture Density Network (MDN) on all four target marginals under both KL and Wasserstein-1, and cannot be stably separated from a Conditional Flow Matching (CFM) model on two of the four. We instead score the models with the Continuous Ranked Probability Skill Score (CRPSS) from forecast verification, taking the exact marginal as the reference. We show how global skill depends on the evaluation spectrum, and we also report it within bins of a conditioning variable, where that dependence largely goes away. We then introduce a data-driven k -Nearest-Neighbor ( k -NN) extrapolation that estimates the best CRPS the conditioning information allows. Both models agree with this estimate on all four targets, which suggests that current architectures have reached the information available in the conditioner, at least as far as these per-target scores can tell.
[LG-107] Randomized and Oracle-Based Feasible Directions Methods for Nonsmooth Nonconvex Constrained Optimization
链接: https://arxiv.org/abs/2611.07691
作者: Dan Greenstein,Nadav Hallak
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:Minimizing nonsmooth nonconvex objectives over general convex sets using no derivatives of the objective, together with structural access to the constraint set, poses a fundamental geometric difficulty: at boundary points, the feasible direction cone may fail to be closed, preventing the construction of finite sets of search directions that capture all feasible descent directions. We develop a feasible-directions framework for problems structured as the difference between a continuously differentiable function and a convex function over a closed convex set, and access the objective without knowledge of its decomposition, through exact line searches along feasible segments. To address the geometric obstruction, we introduce \eta -feasible directions, which remain uniformly separated from the boundary of the feasible direction cone, enabling finite-direction search even in the presence of nonlinear constraints. We present two algorithmic realizations of this framework. The first assumes access to an \eta -direction oracle and extends feasible-directions methods from linear to general convex constraints. The second replaces the oracle by a noise-augmented randomization scheme that constructs spanning directions from the affine-hull inner normal and almost surely avoids nonsmooth boundary points when initialized outside that set. Under curvature bounds and Lipschitz continuity of the smooth component’s gradient, we jointly control the spanning coefficients and feasible step lengths of the oracle construction. This yields finite-time approximate d -stationarity along every unit feasible direction admitting a prescribed positive step length. For both methods, we also establish a rate in the proximal directional minimization (PDM) measure, assuming uniformly bounded direction-set sizes. Randomized exploration ensures that, almost surely, every accumulation point is d -stationary.
[LG-108] Extracting Single-Mode Rayleigh–Taylor Instability dynamics from data via SINDy
链接: https://arxiv.org/abs/2611.07666
作者: Chuanxing Wang,Hui Luo,Guohuai Zhu,Hongbo Cai,Shaoping Zhu,Mingxing Luo
类目: Fluid Dynamics (physics.flu-dyn); Machine Learning (cs.LG)
*备注:
Abstract:Rayleigh-Taylor instability (RTI) is of fundamental importance in fields such as inertial confinement fusion and supernova explosions, and its evolution is commonly described by potential–flow theory and buoyancy–drag models. However, both types of models struggle to accurately capture the overshoot behavior of bubble and spike velocities, as well as the subsequent reacceleration stage. To address these limitations, we develop a physics-constrained, data-driven framework based on the integral formulation of Sparse Identification of Nonlinear Dynamics (SINDy), which extracts low-dimensional dynamical equations for bubble and spike evolution directly from data. By introducing the nondimensional initial amplitude A_0 = k\eta_0 as an explicit parameter, we identify distinct dynamical equations for the bubble and the spike based on data spanning multiple Atwood numbers and initial amplitudes. The resulting low-dimensional dynamical equations not only reproduce the overshoot behavior of the velocity with good accuracy, but also capture the reacceleration stage well, while maintaining high predictive accuracy at unseen values of the wavenumber, acceleration, and Atwood number. This work provides a physics-constrained, data-driven framework complementary to conventional potential-flow and buoyancy–drag models, accurately describing the dynamical evolution of the bubble and spike at high Atwood numbers, and provides a new perspective for directly extracting the dynamical equations of interfacial instabilities from data.
[LG-109] Using Transformers to Generate 3-Manifold Triangulations
链接: https://arxiv.org/abs/2611.07514
作者: Daniel Bruwel,Ashvni Narayanan,Jonathan Spreer
类目: Geometric Topology (math.GT); Machine Learning (cs.LG)
*备注: 21 pages + 3 pages appendix, 7 figures, 10 tables
Abstract:Using dehydration signatures as an encoding, we train GPT-2-style transformers to generate triangulations of 3-dimensional manifolds. Our main runs each use approximately 3 exaFLOPs, spanning a 200 million parameter model trained on 60 million sphere triangulations to an 11 million parameter model trained on 1.1 billion sphere triangulations. The optimal model has 10 blocks, width 640 and generates dehydration signatures of closed manifolds of various topological types at a rate of 6.9%, 82% of which are spheres. To investigate how our method scales, we use the technique of IsoFLOP curves on 15-tetrahedron triangulations of the 3-sphere, over a compute range of 6 to 3000 petaFLOPs, and observe neural scaling behaviour. We extract a parameter scaling exponent of \alpha = 0.49 ( [0.48, 0.50] stat, \pm 0.02 syst), and an optimal ratio of 234 tokens per parameter for training. In addition, we present a number of smaller experiments, including evidence that transformers are able to learn to count degree-one edges near-perfectly, but get progressively worse when counting edges with higher degrees.
[LG-110] Statistical and Privacy Guarantees for Deep M-Regression via Stochastic Gradient Langevin Dynamics
链接: https://arxiv.org/abs/2611.07490
作者: Yeheng Ge,Jian Huang,Shuyuan Wu
类目: Methodology (stat.ME); Machine Learning (cs.LG)
*备注:
Abstract:In this article, we study deep nonparametric regression trained via stochastic gradient Langevin dynamics (SGLD), a variant of stochastic gradient descent that injects carefully calibrated Gaussian noise during optimization. Using a PAC-Bayesian analysis tailored to deep ReLU networks, we establish an excess-risk bound that simultaneously accounts for network approximation error, statistical complexity, and the algorithmic cost accumulated along the SGLD trajectory. Under a Bernstein condition and within a stable SGLD regime, the excess-risk bound contains a stochastic term whose coefficient is inversely proportional to the variance of injected noise, while the estimator retains the optimal rate for nonparametric regression under the Hölder smoothness condition up to logarithmic factors. This result provides an approximation-error-aware, rate-optimal statistical perspective on the widely observed generalization benefits of moderate SGLD noise. For per-sample clipped SGLD, we further establish an exact f-DP guarantee and the corresponding asymptotic \mu -GDP guarantee for the proposed estimator. We establish statistical convergence rates and differential privacy guarantees for DNN estimators trained using SGLD, while explicitly accounting for both network approximation error and the effect of injected noise in the excess-risk bound. Extensive simulation studies and analyses of two real datasets examine the finite-sample predictive performance and privacy-utility behavior of the proposed procedure.
[LG-111] OPRD-Imp: a dataset of experimentally observed impurities and side products in multi-step pharmaceutical syntheses
链接: https://arxiv.org/abs/2611.07488
作者: Heejeong Kim,Seonbin Kim,Jihwan Kim,Tim Pinkhassik,Magdalena Lederbauer,Suong Tran,Connor W. Coley,Juyoung Yoon,Joonyoung F. Joung
类目: Chemical Physics (physics.chem-ph); Machine Learning (cs.LG)
*备注: 15 pages, 4 figures, 2 tables. Dataset available at this https URL
Abstract:Predicting reaction impurities is a critical yet persistently data-limited challenge in pharmaceutical process chemistry. While machine learning models for reaction outcome prediction have advanced rapidly, benchmarking their ability to predict impurities remains impractical due to the absence of curated, publicly accessible datasets. We present OPRD-Imp, a dataset of 229 organic reactions drawn from 62 publications in Organic Process Research Development, each annotated with experimentally observed impurities and their mechanistic origins. A total of 784 impurity entries is provided, classified into four origin categories. Reactions are represented as reaction SMILES and impurities as individual SMILES strings, with multi-step synthetic route connectivity preserved. All data were manually extracted from the source publications by trained chemists following a standardized extraction protocol, with expert chemical validation and manual structure drawing. OPRD-Imp is intended to serve as a benchmark resource for evaluating models that predict reaction impurities, providing experimentally reported impurity annotations and mechanistic origin labels that are not systematically represented in conventional reaction datasets.
[LG-112] Rigorously Certified Neural-Network Computation of Nonlinear Water Waves at Near-Machine Precision
链接: https://arxiv.org/abs/2611.07395
作者: Michal Shavit,Yongji Wang,Tristan Buckmaster
类目: Analysis of PDEs (math.AP); Machine Learning (cs.LG); Fluid Dynamics (physics.flu-dyn)
*备注: 29 pages, 6 figures
Abstract:We present a computational framework that combines machine-learning-based discovery with rigorous computer-assisted numerical certification and formal verification for strongly nonlinear partial differential equations. We demonstrate the approach for periodic gravity-capillary traveling water waves. Physics-informed neural networks are used as a search mechanism, with the wave profile, wave speed and, in continuation experiments, surface tension learned directly from the governing equations without training data. Through a staged training and refinement procedure, the neural networks achieve near-machine-precision residuals of the governing equations, even in strongly nonlinear regimes. A single formulation recovers high-amplitude Stokes waves, Crapper waves and Wilton ripples. Continuation from a high-amplitude Wilton ripple in surface tension reveals new high-amplitude candidate solutions. We freeze one candidate as an explicit finite representation and rigorously certify a uniform bound on its governing-equation residual using interval arithmetic, accounting for floating-point roundoff, Fourier truncation, aliasing and grid-to-continuum errors. This independently validates the neural-network accuracy. We formalize the certification argument in Lean, which machine-checks the deductions from certified numerical bounds to the residual theorem. Both the interval certification and formalization were developed with substantial assistance from a large language model, while the validity of the result remains independent of the language model: the numerical bounds are rigorously enclosed by interval arithmetic and the subsequent mathematical deductions are checked by the Lean kernel. Together, these steps provide a pipeline from neural discovery to rigorous numerical certification and machine-checked verification with auditable mathematical guarantees.
[LG-113] From What Is Used to What Is Missing: Rethinking Dynamic Rank in LoRA
链接: https://arxiv.org/abs/2611.07349
作者: Peng Xue,Guodong Fan,Guang-Yong Chen,Xing Chen
类目: Optimization and Control (math.OC); Machine Learning (cs.LG)
*备注:
Abstract:Parameter-efficient fine-tuning with LoRA is governed by rank, which controls parameter budget and locally accessible task updates. Existing dynamic-rank methods mainly focus on capacity allocation, leaving unresolved whether the current adaptation space is sufficient for the task. We characterize the first-order information boundary of standard factor gradients, showing that their unobservable subspace coincides with the normal space of the low-rank manifold. Thus, factor gradients alone cannot distinguish under-optimization within the current space from task-relevant directions beyond its first-order reach. This motivates a geometric view of rank evolution, where tangent information optimizes existing degrees of freedom and the normal spectrum quantifies the value of new ones. Under a local quadratic model, we show that leading normal singular directions yield the locally optimal rank expansion, with spectral energy determining its gain. Building on this characterization, we propose GEAR, which turns the tangent–normal decomposition into rank evolution through tangent-space optimization, matrix-free normal-spectrum estimation for rank expansion, and task-induced geometric initialization. Across language and vision benchmarks, GEAR achieves the best average performance in all evaluated settings with substantially fewer trainable parameters. On LLaMA3-8B, it raises the average commonsense score from 86.61 to 88.45 while reducing trainable parameters from 42.0M to 14.4M.
[LG-114] Last-Iterate Convergence of Stochastic Gradient Methods for Unbounded Problems
链接: https://arxiv.org/abs/2611.07334
作者: Kevin Kurian Thomas Vaidyan,Ahmet Alacaoglu,Michael P. Friedlander
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 31 pages
Abstract:Last-iterate convergence guarantees for stochastic gradient methods have attracted significant attention as practical implementations typically return the last iterate instead of an average of the iterates. In this paper, we study two approaches for analyzing the last-iterate convergence of proximal and projected stochastic gradient methods without bounded domain or bounded variance assumptions, where black-box variational inference serves as a running example. Under two related assumptions relaxing the bounded variance, we show near-optimal anytime rate \widetilde\mathcalO(1/\sqrtT) for the last iterate. Moreover, when the horizon T is known, we remove the logarithmic factor and obtain the optimal order \mathcalO(1/\sqrtT) , matching the order known under bounded variance or bounded domain assumptions.
[LG-115] ask-Calibrated Random Features via Influence-Based Error Cancellation
链接: https://arxiv.org/abs/2611.07209
作者: Shinto Eguchi
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Random features approximate kernel integrals by Monte Carlo, yet only selected frequency-sampling errors affect a downstream task. We propose Task-Calibrated Random Features (TCRF), which balance frequency moments constructed from task-specific influence using weights shared across kernel entries. Closed-form weights coincide with classical minimum-distance calibration; the contribution is a reusable, task-derived calibration space. We show that first-order variance equals the residual variance after projection onto this space. For kernel ridge regression, complete cancellation yields a non-Gaussian second-order limit. For kernel Stein discrepancy, a subset identity and disjoint-block variance decomposition eliminate the linear component of pilot noise. Positive exponential-tilting weights preserve the positive-semidefinite feature representation and support higher feasible calibration ranks. Paired experiments show improved approximation to exact kernel ridge regression at larger frequency budgets and to the exact empirical Stein discrepancy under a fixed alternative, even after accounting for a frequency-equivalent pilot budget. Gains are smaller with cross-validated ridge, landmark pilots require conservative rank selection, and near the Stein null allocating the pilot budget to additional ordinary frequencies is preferable.
[LG-116] ShapleyOG: Fast Adaptive Shapley Value Estimation with Odd Gaussian Processes
链接: https://arxiv.org/abs/2611.07173
作者: David Rundel,R. Teal Witter,Santo M. A. R. Thies,Matthias Feurer,Fabian Fumagalli,Bernd Bischl
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: Preprint under review
Abstract:Shapley values are a central tool for attribution in machine learning. Since their exact computation requires evaluating the cooperative game to be explained on exponentially many player coalitions, they are typically estimated from a limited number of evaluations. Recently, ShaplEIG has shown that adaptively selecting the coalitions to evaluate substantially improves sample efficiency over classical estimators. This, however, comes at the cost of a sequential procedure that, in every iteration, refits a Gaussian process surrogate of the game and selects the coalition with the largest expected information gain about the Shapley values. The resulting computational overhead limits its applicability to costly games with low evaluation budgets. We present fast adaptive Shapley value estimation with Odd Gaussian Processes (ShapleyOG) to overcome this limitation. ShapleyOG leverages the fact that Shapley values depend only on the odd component of the game, and we show that under paired sampling a Gaussian process surrogate on this component alone yields the same Shapley value posterior and coalition selection criterion from half as many observations. Moreover, ShapleyOG refits the surrogate only on a sparse schedule and efficiently updates the selection criterion in between, and it extends quadrature-based computation schemes to the prior covariance of the Shapley values, which allows for exact or cheaper approximate computation. In extensive experiments on diverse games, ShapleyOG matches the estimation accuracy of ShaplEIG while reducing its overhead by up to three orders of magnitude. As a result, it outperforms state-of-the-art non-adaptive estimators at equal total computation time already when a single evaluation takes a tenth of a second and a budget of a few thousand evaluations suffices, thereby widening the regime in which adaptive coalition selection is applicable and pays off.
[LG-117] Context-Aware Automated NMR Structure Elucidation
链接: https://arxiv.org/abs/2611.07133
作者: Frank Hu,Jonathan M. Tubb,Grant M. Rotskoff,Matthew W. Kanan,Thomas E. Markland
类目: Chemical Physics (physics.chem-ph); Machine Learning (cs.LG)
*备注:
Abstract:Machine learning approaches are rapidly revolutionizing the ability to analyze and extract structural insights from chemical data, enabling high-throughput chemical discovery. While existing data-driven approaches have shown remarkable promise for predicting chemical structures in the de novo setting, incorporating additional chemically informed context offers the opportunity to further improve these models’ performance and extend their real-world utility. Here we provide a straightforward protocol to combine artificial-intelligence-based spectroscopic elucidation pipelines with contextual chemical data, e.g., known reactants or a desired product, by using automated retrosynthetic analysis to generate large training sets of molecules and contextual data in the form of synthetic precursors. We demonstrate our approach by showing that our recently introduced NMR structure elucidation framework can be trained to effectively leverage molecular precursor context. We show that when the chemical context provides useful information on the target molecule we obtain dramatically improved accuracy over using NMR spectra alone, obtaining Top-15 accuracies of up to 95.4% for molecules up to 40 heavy atoms spanning the full elemental range of organic chemistry. Furthermore, we demonstrate that our approach is highly robust, such that even when the context provided is uncorrelated with the target, one still obtains accuracies commensurate with those of the original NMR structure elucidation framework. This approach provides an efficient route to improve a wide range of spectroscopic structure elucidation platforms using contextual data.
[LG-118] Bayesian Tensor Train Kernel Machines with Automatic Model Complexity Inference
链接: https://arxiv.org/abs/2611.07129
作者: Afra Kilic,Kim Batselier
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 31 pages, 4 figures, 1 algorithm. Code available at this https URL
Abstract:Tensor Network (TN) kernel machines accelerate supervised learning by using low-rank TNs to reduce computation and memory. However, most TN-based methods are deterministic with manually tuned ranks and feature dimensions, while the existing Bayesian TN kernel machine is restricted to the CANDECOMP/PARAFAC decomposition (CPD), whose single rank is shared by all modes. We extend this framework to the tensor train (TT) and propose Bayesian Tensor Train Kernel Machines (BTT-KM), whose hierarchical ARD-type priors on the TT cores infer the TT ranks and feature dimensions from data. Since each TT core is coupled to the others only through its neighbors, its closed-form posterior update requires propagating the contributions of all other cores along the chain. Mean-field variational inference yields a Bayesian alternating least squares algorithm with the same computational complexity as its deterministic counterpart and a closed-form predictive distribution. A synthetic experiment illustrates that BTT-KM recovers the true TT ranks and feature dimensions from an over-parameterized initialization.
[LG-119] Credibility Clustering for Experience Rating: An Automated Approach to Risk Group Allocation in Group Life Insurance
链接: https://arxiv.org/abs/2611.07110
作者: Thom van Rijn,Frank Weber,Michael Merz
类目: Risk Management (q-fin.RM); Machine Learning (cs.LG)
*备注:
Abstract:In insurance, experience rating plays a crucial role in mitigating adverse selection by adjusting premiums according to historical claims experience. A key step in this process is the classification of insured collectives into homogeneous risk groups. Within group life insurance, this classification traditionally relies on manual decisions and arbitrary thresholds, which may lead to inefficiencies and inconsistent outcomes. This paper proposes a data-driven methodology for automating risk group allocation based on a combination of the multidimensional Bühlmann-Straub model and k-means clustering – referred to as credibility clustering. By incorporating credibility weights, the method accounts for the varying statistical significance of experience data and ensuresthat risk classifications are both actuarially sound and practically implementable. Empirical results based on real-world group life insurance data from a major European insurer demonstrate that the proposed framework outperforms the two k-means benchmark variants considered in the empirical application in terms of viability, temporal stability and accounting for statistical reliability.
[LG-120] oroFit: Fitting Wedge Sums of Tori to High-Dimensional Data
链接: https://arxiv.org/abs/2611.06990
作者: Tanishq Bhatia,Jose A. Perea,Tarek M. Zikry
类目: Machine Learning (stat.ML); Machine Learning (cs.LG); Algebraic Topology (math.AT)
*备注:
Abstract:Data arising in scientific applications often exhibits complex nonlinear structure. Identifying and parameterizing said nonlinearities can enable better modeling, enhance predictive analysis, data interpolation, denoising, and anomaly detection. In this paper, we develop theoretical and computational methods to detect and parametrize Euclidean point cloud data lying near a wedge sum of tori. Specifically, we use persistent cohomology and cup products to identify the distinct toroidal components, and on each component, we optimize over low-energy multivariate trigonometric polynomials for best data fit. The ensuing optimization problem reduces to generalized ridge regression, which can be solved efficiently via standard techniques. We demonstrate the utility of the proposed framework in both synthetic and real datasets, including interpolation of low frame-rate recurrent videos, attractor recovery in quasiperiodic dynamical systems, and the parametrization of gene-space cycle data for parasites and humans.
[LG-121] Capturing Membrane Dynamics and Spike Timing of Human Neurons at Scale using Neural Operators
链接: https://arxiv.org/abs/2611.06933
作者: Luca Ghafourpour,Cindy Zhao,Valentin Duruisseaux,Bahareh Tolooshams,Philip H. Wong,Costas A. Anastassiou,Anima Anandkumar
类目: Neurons and Cognition (q-bio.NC); Machine Learning (cs.LG)
*备注:
Abstract:Characterizing the properties of neuronal cell types is central to understanding brain function. Yet, biophysically realistic neuron models remain difficult to scale to large populations while preserving the richness of cellular dynamics and within-cell-type variability. We present a cell-type-specific neural-operator framework for scalable neuronal population modeling. Rather than training separate surrogate models for individual neurons, the proposed framework learns a shared operator across multiple modeled neurons from the same cell type, conditioned on biologically interpretable electrophysiological features. Relative to the original single-neuron \textttNOBLE setting, learning across heterogeneous neurons requires a richer electrophysiological conditioning representation and a training objective designed to preserve diverse spiking dynamics. This enables efficient generation of large ensembles of somatic voltage responses while preserving variability across cells and model configurations. We evaluate the approach using membrane dynamics metrics, action potential waveform metrics, spike timing accuracy, input-output firing relationships, and electrophysiological feature distributions. Our framework reproduces subthreshold membrane dynamics, spike waveforms, and firing-rate responses across major cortical inhibitory neuron types, while accurately preserving spike timing despite the sensitivity of threshold-crossing dynamics. The predicted electrophysiological feature distributions show strong agreement with detailed simulator outputs and human experimental recordings, including high explained variability in properties such as inter-spike intervals, firing frequency, action-potential shape, and post-spike recovery.
[LG-122] Reading a neurons tuning from its correlations with other neurons
链接: https://arxiv.org/abs/2611.06927
作者: Francesco Lässig
类目: Neurons and Cognition (q-bio.NC); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
*备注: Code: this https URL . Reassignment to q-bio.NC as primary welcome
Abstract:What a neuron represents is usually established by relating its activity to a stimulus. We ask whether the same facts are contained in the correlation matrix of a recorded population, which keeps the pairwise relations among the neurons and discards the stimulus. A transformer that sees only the correlation matrix of 512 sampled neurons, with no labels and no reference population in its input, decodes the preferred orientation of mouse visual cortex neurons from in-vivo MICrONS responses to a mean error of 25° (uninformed: 45°; best constant guess: 41°) and their receptive-field position at R^2 = 0.23; on the digital-twin substrate, 20° and 0.48. Asked whether a neuron prefers horizontal or vertical, it is right for 79% of neurons in vivo and 87% on the twin, against 56% for the majority answer; it does not resolve the obliques. The ordering is recoverable because the relations between orientation classes are close to circulant, which fixes the class structure up to rotation and reflection; the rotation of the frame is recoverable only because of a departure from that symmetry. The departure survives equal representation of every orientation and disjoint training and test stimuli, and it is an anisotropy of the signal correlations: in V1 the neurons preferring the dominant orientation have the most similar responses to one another, and a model population with only that property recovers the frame. Across 33 mice of the Allen Brain Observatory, with relations from drifting gratings, orientation transfers to unseen animals only when a sampled population mixes animals; populations drawn from one animal are not decoded measurably below the best constant, a limit of the diversity of populations one animal provides. For receptive fields, what crosses animals is where each animal’s imaged patch looks on the screen, not the layout of neurons within it.
[LG-123] Approximating the Euclidean Ball by Polytopes: Surprising Rates in High Dimensions
链接: https://arxiv.org/abs/2610.34651
作者: Yegor Gorodzha,Steven Hoehner,Gil Kur
类目: Metric Geometry (math.MG); Computational Geometry (cs.CG); Machine Learning (cs.LG); Probability (math.PR)
*备注: 71 pages, 1 figure, 1 table. v2: exposition and references revised for a broader audience
Abstract:We provide sharp estimates on the number of vertices needed to approximate the Euclidean unit ball B_d\subset\mathbb R^d with relative symmetric-difference volume error at most \varepsilon\in(0,1) . We show that the minimum vertex budget N_\varepsilon(d) satisfies \log N_\varepsilon(d)\asymp d\log\left(1+\frac1\varepsilon\sqrt d\right). In particular, \exp(O(\sqrt d/\varepsilon)) vertices suffice for every \varepsilon\in(0,1) and all sufficiently large d . We also prove that achieving a fixed error \varepsilon\in(0,1) requires at least \Omega\bigl((\tfrac1+\varepsilon2\sqrt\varepsilon)^d-1\bigr) facets, an exponential number in d . We extend our vertex approximation result to Gaussian symmetric-difference error, and give a positive answer to the question of De, Nadimpalli, and Servedio about subexponential vertex complexity for approximating Euclidean balls. We also disprove a Macbeath-type extremality principle for unrestricted vertex approximation. Whereas the ball uniformly maximizes the inscribed approximation error among all convex bodies (Macbeath, 1951), we show that approximating the hypercube [-1,1]^d to relative error \varepsilon , with \delta:=1-\varepsilon , requires N\ge\frac\delta2\exp(\Omega(\delta^2d/\log^2(2/\delta))) vertices whenever d\ge16\log(2/\delta) . Thus, when d is large, the cube is harder to approximate than the ball in this sense. Finally, we determine the sharp dimensional order of the classical Laguerre-Delone triangulation number, showing that \mathrmldel_d-1=\Theta(d^-1/2) as d\to\infty . This improves both the upper bound of Ludwig, Schütt, and Werner and the lower bound of Böröczky by a factor of order \sqrt d . By classical asymptotic approximation theorems (Ludwig, 1999), this improvement applies to every convex body in \mathbb R^d with C^2 boundary and positive Gaussian curvature. Comments: 71 pages, 1 figure, 1 table. v2: exposition and references revised for a broader audience Subjects: Metric Geometry (math.MG); Computational Geometry (cs.CG); Machine Learning (cs.LG); Probability (math.PR) MSC classes: 52A27 (Primary) 52A23, 52A22, 68U05, 68Q32 (Secondary) Cite as: arXiv:2610.34651 [math.MG] (or arXiv:2610.34651v2 [math.MG] for this version) https://doi.org/10.48550/arXiv.2610.34651 Focus to learn more arXiv-issued DOI via DataCite
附件下载
