本篇博文主要内容为 2026-11-05 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 (2026-11-05)
今日共更新828篇论文,其中:
- 自然语言处理共91篇(Computation and Language (cs.CL))
- 人工智能共219篇(Artificial Intelligence (cs.AI))
- 计算机视觉共150篇(Computer Vision and Pattern Recognition (cs.CV))
- 机器学习共215篇(Machine Learning (cs.LG))
- 多智能体系统共11篇(Multiagent Systems (cs.MA))
- 信息检索共22篇(Information Retrieval (cs.IR))
- 人机交互共29篇(Human-Computer Interaction (cs.HC))
多智能体系统
[MA-0] A Framework for Context Enrichment and Selective Robustness in LLM Social Agents NEURIPS2026
【速读】：该论文旨在解决大语言模型（Large Language Models, LLM）在社会模拟中因提示（prompt）的语义保持性变化而导致行为显著波动的问题，即模型行为的脆弱性（brittleness），这给社会模拟结果的可验证性带来了挑战。其核心解决方案在于提出一种可控的上下文增强（controlled context enrichment）框架，通过一系列系统化的提示替换规则，分别对中性填充、通用社会背景、描述性决策依据以及规范性决策依据等不同类型的上下文进行结构化增强。同时，为评估增强效果，引入“选择性鲁棒性”（selective robustness）作为验证标准，要求模型对语义不变的提示变化保持不敏感，但对与决策相关社会环境的变化具有方向性响应。研究通过类囚徒困境的游戏场景验证该框架，发现不同形式的上下文增强会引发模型产生差异化的、可区分的行为模式，且其效果在不同模型间存在显著差异，从而为提升LLM代理在社会模拟中的稳健性提供了可操作的路径与评估方法。
链接: https://arxiv.org/abs/2611.06914
作者: Mattia Greiche,Reihaneh Rabbany,Maximilian Puelma Touzel
机构: McGill University (麦吉尔大学); Mila - Quebec Artificial Intelligence Institute (魁北克人工智能研究所)
类目: Multiagent Systems (cs.MA); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Accepted to the SocialAgent Workshop @ NeurIPS 2026. 12 pages, 1 figure, 2 tables
Abstract:Researchers use large language models as agents in social simulations, where agents are prompted with observations and background information intended to shape their actions. Yet LLM agents have been shown to substantially change their behavior under meaning-preserving prompt variations, creating a validation problem for social simulation. One plausible hypothesis is that this brittleness reflects a lack of grounding and can be reduced by enriching the context provided to the agents. Whether richer context improves robustness remains unclear, in part because context richness and robustness are rarely defined in ways that support controlled comparisons. We address the first need with controlled context enrichment, a set of prompt substitution rules that systematically enriches distinct types of context: neutral padding, generic social background, descriptive decision grounding, and normative decision grounding. We address the second by introducing selective robustness as a validation criterion: insensitivity to meaning-preserving prompt variation combined with a directional response to decision-relevant changes in the social world. We demonstrate the framework in a Prisoner’s Dilemma-like game, showing how it can be used to identify which forms of context enrichment produce more desirable and robust agent behavior. We find that these forms of enrichment produce distinct behavioral patterns that vary between the two models tested.
[MA-1] Airlock: Bounding the Multi-Agent Blast Radius via Selective Online Containment
【速读】：该论文旨在解决多智能体系统（multi-agent systems）在执行任务过程中，因单个智能体输出故障而引发的故障扩散问题。传统方案中，单智能体防护机制局限于局部控制，而现有跨智能体管控策略要么仅隔离故障源而忽略下游受影响的消费者，要么采取全局回滚，导致正常任务流程被无谓中断。其核心解决方案在于提出Airlock框架，通过树状结构运行时系统EdgeCitadel实现垂直权限收缩，建立智能体节点间的边界与控制开关；同时采用追加式不可变溯源有向无环图（append-only provenance DAG）记录跨智能体的数据消费关系。当触发警报后，Airlock可沿溯源图正向追踪污染传播路径以定位受感染执行和待执行操作，并反向追溯根本原因。在控制开关处，系统提供双层处置策略：默认全隔离层（full quarantine tier）冻结受影响分支，无需求解器；优化剪枝层（assured optimized cut tier）则采用“设篱-修剪”（fence-then-prune）机制，在满足时限约束的前提下实现最小化破坏的最优割集。实验表明，在1,832次测试床运行、161次真实智能家居大语言模型（LLM）部署及24,266条仿真记录中，Airlock实现了主测试平台与智能家居网格上所有警报后故障动作的零漏过阻断，且相较A2A与LangGraph原生控制，保留健康工作量提升68至96个百分点，优化剪枝策略在高传播密度场景下比全隔离策略多保留46个百分点的正常工作，验证了运行时权限边界与在线溯源协同设计在实现精准故障隔离的同时，有效避免对正常任务的过度影响。
链接: https://arxiv.org/abs/2611.06727
作者: Zhonghao Zhan,Javad Forough,Yefan Zhang,Krinos Li,Xiao Ma,Hamed Haddadi
机构: 未知
类目: Multiagent Systems (cs.MA)
备注:
Abstract:When multi-agent systems act on the external world, a single agent’s faulty output can spread across tasks before detection, creating a system-wide blast radius. Single-agent guardrails are confined to individual harnesses, while existing multi-agent controls either cancel only the faulty source and miss downstream consumers or halt and roll back the entire workflow at the cost of healthy work. We present Airlock, a multi-agent runtime and guardrail for selective post-alert fault containment. Its tree-structured runtime, EdgeCitadel, narrows permissions vertically from parent to child agents to establish node boundaries and control switches. Meanwhile, an append-only provenance DAG records cross-agent data consumption horizontally. Upon an alert, Airlock traces forward on this DAG to isolate tainted executions and pending actions, and backward to narrow root causes. At the site switches, a default full quarantine tier freezes affected branches without a solver, while an assured optimized cut tier uses fence-then-prune to enforce a deadline-aware minimum cut minimizing collateral disruption. Across 1,832 testbed runs, 161 physical smart-home LLM runs, and 24,266 simulation records, Airlock blocks all post-alert faulty actuations on our main testbed and smart-home grids with zero post-convergence leaks. Whereas native controls in A2A and LangGraph either contain at most 33.3% of faults or retain at most 16% of work without redo, Airlock retains +68 pp to +96 pp more healthy work, and its optimized cut retains +46 pp more than full quarantine under dense spread. These results show that co-designing runtime authority boundaries with online provenance enables surgical fault containment without sacrificing healthy work (code: this https URL).
[MA-2] Signed Sealed Leaked: Trust Beyond Signature for MCP Toolchains
【速读】：该论文旨在解决生成式 AI（Generative AI）代理在依赖第三方工具（如 Model Context Protocol, MCP）时所面临的信任安全问题：即使工具经过合法签名，仍可能在执行其宣称功能的同时隐藏恶意代码，导致敏感信息泄露。现有运行时防御机制（如 Meta 的 MCPGuard-Dynamic）虽能阻止直接的数据外泄，但无法防范通过合法请求路径隐含泄露的攻击。论文提出“胚系执行”（germline execution）作为核心解决方案，其关键在于通过为每次工具调用创建从无数据状态的胚系进程（germline process）派生的临时副本，确保每个副本仅接收其目标所授权的数据，从而实现信息隔离。由于恶意代码只能访问其副本中的有限数据，其可泄露的信息被严格限制。在假设副本隔离与胚系进程无数据的前提下，论文形式化证明了对所有已签名版本的接收端非干扰性（sink noninterference），并对确定性执行版本引入可验证的见证机制以检测数据流违规。实验表明，在153次攻击测试中，胚系执行未发现任何未经授权的数据披露，而基线防御和消融实验则存在5.9%至74.5%的泄露率；同时，单个副本支持的调用延迟仅为10-49毫秒，相较传统每调用一次重启工具的方式快2至28倍。结果表明，基于数据隔离的执行模型可有效防御嵌入于认证软件中的隐蔽恶意代码。
链接: https://arxiv.org/abs/2611.06585
作者: Yichiet Aun(Universiti Tunku Abdul Rahman),Mohammed A. Alhartomi(University of Tabuk),Jing Jie Tan(National University of Singapore),Waqas Nadeem(Universiti Tunku Abdul Rahman),Adeb Salh(Universiti Tunku Abdul Rahman),Shankar Karuppayah(Bitranger, CYRES, Universiti Sains Malaysia),Sin Chee Tan(ABLEs)
机构: Universiti Tunku Abdul Rahman(拉曼大学学院); University of Tabuk(塔布克大学); National University of Singapore(新加坡国立大学); Bitranger; Universiti Sains Malaysia(马来西亚理科大学); ABLEs
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 24 pages, 5 figures, 15 tables
Abstract:AI agents rely on third-party tools, such as Model Context Protocol (MCP) servers, whose publishers can be authenticated and whose execution can be restricted by runtime security policies. Yet authenticity does not guarantee benign behaviour: a legitimately signed tool may perform its advertised function while concealing malicious code that leaks confidential information. We demonstrate this trust failure with a signed MCP tool that embeds sensitive data in requests to a policy-allowed destination. Meta’s MCPGuard-Dynamic, a state-of-the-art runtime defence, blocks a direct exfiltration attempt in the same run, yet fails to prevent this attack in every trial. We introduce germline execution, which limits information disclosure without inspecting or trusting tool implementations. Each tool invocation runs in ephemeral copies forked from a data-free germline process, each receiving only information authorised for its destination. Even if malicious code executes, what it can disclose is bounded by the data available to its copy. Under explicit assumptions on copy confinement and a data-free germline, we prove sink noninterference for every signed release and, for releases deterministic within our replay model, show that a verified witness establishes a flow violation. Across 153 attack runs, no unauthorised disclosure was detected under germline execution, compared with 5.9-74.5% of runs under baseline defences and an ablation. Invocations served by a single copy completed in 10-49 ms, 2-28x faster than restarting the tool for every invocation, the conventional way to clear state. These results show that data-isolated execution protects agent toolchains against malicious code concealed within authenticated software.
[MA-3] Corralling AI Risks with Grassroots Contracts
【速读】：该论文旨在解决生成式 AI（Generative AI）快速发展背景下，数字平台所面临的三大核心风险：一是全球性社会与经济平台被 AI 所控制，导致人类整体陷入对 AI 的依赖与脆弱性；二是匿名运行的不可问责 AI 代理在平台中与人类并行存在，可能伪装成人类、服务于未知目标且不受任何约束；三是由匿名 AI 生成的无法溯源的深度伪造内容（deepfakes），使真实信息难以辨识，且生产者逃避责任。其解决方案的关键在于提出“草根合同”（grassroots contracts）这一新型机制，作为对法国思想家皮埃尔-约瑟夫·普鲁东（Pierre-Joseph Proudhon）百年契约理念的数字重构，通过在社交图谱中引入可追溯的可信身份与行为记录，以及在社交网络中实现内容数字溯源，构建具有问责能力的 AI 代理体系与真实内容验证机制。研究进一步开发了可运行的智能手机应用原型，验证了该方案在遏制上述三类重大 AI 风险方面的可行性与有效性。
链接: https://arxiv.org/abs/2611.06374
作者: Ehud Shapiro
机构: London School of Economics (伦敦政治经济学院); Weizmann Institute of Science (魏茨曼科学研究所)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC); Multiagent Systems (cs.MA)
备注:
Abstract:AI’s inflection point is a fountain of alarming risks. Digital platforms in particular are vulnerable to three major AI risks, as those in control of global platforms cede control to AI, of the algorithms, of membership and of content: (1) AI control of global platforms, social and economic, used by the majority of humanity, resulting in the majority of humanity becoming vulnerable to AI control. (2) Unaccountable AI agents operating anonymously on social and economic digital platforms, side-by-side with humans, possibly masquerading as humans, serving unknown masters with unknown goals, operating under unknown restraints without accountability or culpability. (3) Unattributed deepfakes generated by anonymous AI, with people unable to distinguish them from genuine content and with their producers evading accountability and culpability. Here, we propose to corral these AI risks with grassroots contracts, a digital incarnation of a two-centuries old notion of Pierre-Joseph Proudhon. We present grassroots contracts for a social graph with accountable AI agents and for a social network with digital provenance, and working smartphone apps that implement the two and show how they corral these major AI risks: (1) AI control of global platforms. (2) Unaccountable AI agents. (3) Unattributed deepfakes. Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC); Multiagent Systems (cs.MA) Cite as: arXiv:2611.06374 [cs.AI] (or arXiv:2611.06374v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.06374 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[MA-4] Mechanical Engineering Skills: An Open Library of Agent Skills for Design Analysis and Manufacturing
【速读】：该论文旨在解决机械工程领域中专业工具（如计算机辅助设计、仿真、制造和研究工具）分散、接口各异且学习成本高的问题。其核心挑战在于如何将这些高度专业化、异构化的技能整合为一个统一、可交互的智能系统，使工程师能够通过自然语言或语音指令即可调用复杂工程能力。解决方案的关键在于提出“机械工程技能库”（Mechanical Engineering Skills），这是一个开源的33项代理技能集合，每项技能均包含操作指令、可执行工具与验证流程，实现了从自然语言请求到精确工程计算的端到端自动化。系统通过语言模型解析用户意图并协调技能调度，而确定性工具负责执行计算并返回经过验证的结果，从而确保可复现性和准确性。该方法无需针对特定工具进行界面培训、本地模型部署或微调，所有技能可按需加载，仅需一条命令即可安装。其中19项技能基于实验室已发表或预印本研究成果实现，涵盖文本生成CAD、连杆机构综合、计算流体动力学（CFD）设置、应力场与晶格代理模型、方程发现、增材制造规划及过程监控等前沿功能；其余14项支持通用工程与教育需求。论文还提出了技能格式规范、设计原则以及将科研成果转化为实用工作流的方法。五项案例研究在明确证据范围下独立重演，验证了晶格均质化分析的精确一致性、火箭顶点高度预测误差低于4%（对比OpenRocket基准）、方程拟合精度比三次多项式高九倍（在未见温度点），以及孔隙度监测的全流程端到端模拟。尽管受控的代理比较尚待后续研究，但该开放库已公开发布，为工程智能化提供了一种可扩展、可验证的范式。
链接: https://arxiv.org/abs/2611.05758
作者: Achuth Chandrasekhar,Jesse Barkley,Peter Pak,Rumi Loghmani,Aakash Oza,Amir Barati Farimani
机构: Carnegie Mellon University (卡内基梅隆大学)
类目: Multiagent Systems (cs.MA)
备注:
Abstract:Mechanical engineers use specialized tools for CAD, simulation, manufacturing and research, each with a distinct interface and learning curve. We present Mechanical Engineering Skills, an open library of 33 agent skills that makes these capabilities accessible through a single text or voice conversation. Each skill combines operating instructions, executable tools and verification procedures. A language model interprets the request, selects and coordinates skills, while deterministic tools perform computations and return checked results. Default workflows require no tool-specific interface training, local model execution or model fine-tuning; skills load on demand and install with one command. Nineteen skills adapt published or preprint research from our laboratory, including text-to-CAD, linkage synthesis, CFD setup, stress-field and lattice surrogates, equation discovery, additive-manufacturing planning and in-process monitoring. Fourteen additional skills support core engineering and education. We describe the skill format, design principles and a method for translating research into practical workflows. Five case studies with explicit evidence scopes and independent reruns demonstrate exact analytical agreement for lattice homogenization, rocket apogee predictions within 4% of stored OpenRocket references, an equation fit nine times more accurate than a cubic polynomial at an unseen temperature, and an end-to-end rehearsal of porosity monitoring. Controlled agent comparisons remain future work. The open library is available at this https URL.
[MA-5] Same Method Different Task: What the Bitter Lesson Does Not License in Social Simulation NEURIPS2026
【速读】：该论文试图解决的问题是：在基于大语言模型（LLM）的社会模拟中，尽管生成式人工智能（Generative AI）遵循“苦涩的教训”（bitter lesson），即通过计算规模扩展的通用方法优于手工编码的领域知识，但现有研究中出现的若干反直觉现象——如过早出现的合作行为、群体共识却缺乏个体一致性、种群同质化等——无法仅通过计算规模或模型结构解释。其核心问题在于，当前社会模拟往往将代理（agent）视为孤立的文本预测器，忽略了真实社会互动中的利益相关性与后果机制。解决方案的关键在于引入“利害关系”（stakes）这一缺失要素：真正的社会行为（如合作、背叛、规范执行）本质上是利益相关方在相互影响下产生的结果，而非无偏观察者对文本的预测。论文通过一个包含540个代理的初步实验验证，当代理被赋予持久后果而非单纯追求合作时，策略会因模型家族不同而发生显著重构，但始终未能产生人类般的多样性。因此，作者主张应构建混合动机（mixed-motive）环境来塑造代理，而非仅以测试为目的；并提出应重新考虑代理的预训练方式，使其具备社会行为演化所需的情境敏感性与后果感知能力。
链接: https://arxiv.org/abs/2611.05717
作者: Justin Downes,Robby Milletich,Newel Hirst
机构: Amazon Web Services(亚马逊网络服务)
类目: Multiagent Systems (cs.MA)
备注: Accepted at the SocialAgent Workshop, NeurIPS 2026 (non-archival). 15 pages, 5 figures
Abstract:The bitter lesson holds that general methods leveraging computation defeat hand-encoded domain knowledge. We argue it carries a condition its popular reading omits: in every case Sutton cites, the general method was scaled against the same task the domain engineering was solving. Social simulation substituted across tasks instead, replacing hand-built agents with an artifact optimized for predicting text. The move is bitter-lesson-shaped, but the lesson does not license it. Prediction is a solitary act: the predictor bears no consequences from the interaction it models. Cooperation, defection, and norm enforcement are phenomena of interested parties; they do not exist for a disinterested observer at any scale. What is missing is not emergence, which LLM populations already show, but stakes. We show that several anomalies reported in prior work (excess cooperation that predates alignment training, collective agreement without individual fidelity, and population homogeneity) follow from this account. A 540-agent pilot with pre-specified hypotheses, which tests sensitivity to persistent consequences rather than cooperation, finds that consequences reshape strategies in two of three model families but never produce human-like heterogeneity. We argue for mixed-motive settings that shape agents rather than only test them, and ask whether agents should be pretrained differently from predictors.
[MA-6] MA-SafeDiffuser: Safe Multi-Agent Planning with Diffusion Probabilistic Models AAMAS2026
【速读】：该论文旨在解决多智能体系统中基于扩散模型的轨迹规划在满足联合安全约束时缺乏有限时间不变性保证的问题。现有方法难以确保多个智能体在动态交互过程中始终处于安全状态，尤其在复杂环境中易出现局部死锁或碰撞风险。其解决方案的关键在于：构建以个体与成对控制屏障函数（Control Barrier Function, CBF）集合交集形式定义的联合安全集，并在此基础上设计具备可证明不变性的集中式与分布式（通信感知）去噪机制；同时引入随时间变化的安全规格和活力型CBF以缓解局部陷阱与死锁现象。该方法在轻量级多智能体基准测试套件（包含多智能体迷宫环境）上验证了其有效性，相较于无约束扩散基线显著降低违规次数，同时保持了高质量的路径规划性能。
链接: https://arxiv.org/abs/2611.05524
作者: Kiran Ravish,Ankita Kushwaha,Preeti,Pawan Kumar
机构: IIIT Hyderabad(印度信息科技研究所海得拉巴分校)
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: Accepted at AAMAS 2026. 14 pages, including supplementary material
Abstract:We propose MA-SafeDiffuser, a multi-agent extension of SafeDiffuser that equips diffusion-based trajectory planners with finite-time diffusion invariance guarantees for joint-agent safety specifications. Building on the single-agent construction that embeds control barrier function (CBF) constraints into reverse-diffusion updates, we: (i) formalize joint safe sets as intersections of per-agent and pairwise barrier sets; (ii) derive centralized and decentralized (communication-aware) constrained denoising procedures with provable invariance under mild assumptions; (iii) address local-trap and deadlock phenomena via time-varying specifications and liveness CBFs; and (iv) develop a lightweight benchmarking suite including a multi-agent Maze2D domain. Empirically, MA-SafeDiffuser reduces violation counts relative to unconstrained diffusion baselines, while retaining planning quality. We provide algorithms, proofs, and reference implementations.
[MA-7] Uncertainty-Aware Resource Orchestration for Agent ic AI Systems
【速读】：该论文旨在解决生成式 AI (Generative AI) 系统在边缘或云环境中执行时面临的资源调度不确定性问题。这类系统基于大语言模型（LLMs）构建，以自主多步流水线形式运行，动态调用工具、将观测结果融入记忆并迭代求解任务，其整体计算量、带宽需求及端到端延迟在调度时刻不可预知——这些指标由运行时的中间观测和随机词元采样动态决定，表现为重尾、多模态的随机变量。这种不可预测性导致基础设施编排器陷入两难：资源不足可能引发执行中限流与超时，而过度预留则浪费稀缺的边缘或云端资源。现有调度器依赖已知的固定流水线图结构，而边缘资源管理方法通常采用参数化或平稳的需求模型，均无法适应生成式流水线的未知且非平稳特性。为此，论文提出将该场景形式化为不确定性感知的生成式编排问题，目标是在调度时刻预先分配足够的计算与带宽资源，以在指定概率下覆盖实际需求，并验证任务可满足截止时间约束。解决方案的关键是提出置信预留编排算法（Conformal Reservation for Orchestration, CRO），该算法通过滑动窗口内的历史执行轨迹，利用分位数预测（conformal prediction） 技术，在无需假设分布的前提下，实现有限样本、无分布依赖的覆盖率保证。仿真结果表明，CRO 是唯一在所有风险水平下均能同时满足目标覆盖率的算法，且在满足覆盖率的方案中实现最低的过预留，资源浪费约为基于切比雪夫界基线方法的一半。
链接: https://arxiv.org/abs/2611.05392
作者: Masoud Shokrnezhad,Tarik Taleb
机构: ICTFICIAL Oy( ICTFICIAL Oy); Ruhr University Bochum (鲁尔大学波鸿分校)
类目: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Multiagent Systems (cs.MA); Networking and Internet Architecture (cs.NI)
备注: 6 pages, 2 figures. Accepted to IEEE MECOM 2026, Marrakech, Morocco, 2-4 November 2026
Abstract:Agentic artificial intelligence (AI) systems powered by large language models (LLMs) execute as autonomous multi-step pipelines that dynamically invoke tools, integrate observations into memory, and iterate until a task is resolved or a resource budget is exhausted. Unlike classical workflows with fixed graphs, the aggregate compute, bandwidth, and end-to-end latency demands of such pipelines are unknown at dispatch time: they emerge at runtime from intermediate observations and stochastic token sampling, yielding heavy-tailed, multi-modal random variables. This unpredictability forces infrastructure orchestrators into a dilemma: under-provisioning risks mid-execution throttling and deadline violations, while over-provisioning wastes scarce edge or cloud capacity. Existing schedulers require a fully known pipeline graph, while edge resource-management methods rely on parametric or stationary demand models, and neither fits the unknown, non-stationary profiles of agentic pipelines. We formalize this setting as the uncertainty-aware agentic orchestration problem, where the goal is to reserve, at dispatch time, compute and bandwidth allocations that cover actual pipeline demands, and to verify that the pipeline can meet its deadline, each with a prescribed probability. We then propose Conformal Reservation for Orchestration (CRO), an online algorithm that translates a sliding window of past execution traces into dispatch-time resource allocations with finite-sample, distribution-free coverage guarantees via conformal prediction. Simulations over heterogeneous task mixes with multi-modal and heavy-tailed profiles show that CRO is the only evaluated method that simultaneously satisfies target probabilistic coverage across all risk levels while, among coverage-feasible methods, attaining the lowest over-reservation, roughly halving the resource waste of the Chebyshev-bound baseline.
[MA-8] Who May Overrule the Agent ? Evidence-Gated Authority in LLM Review Institutions
【速读】：该论文旨在解决多智能体语言模型系统中评审者（reviewer）在决策优化过程中可能对本已正确的上游决策造成无谓替换或“错误修正”的问题，即如何界定评审者的权威边界——在何种条件下其建议应具有决定性效力。其解决方案的关键在于引入一种“证据门控机制”（evidence gate），该机制要求评审者必须生成一个机械上有效的、可验证的违反约束记录及其合规修复方案后，才能替代原有决策。这一设计通过强制评审者提供可追溯的纠错证据，显著降低了对正确提案的误伤率，同时保留了对真实错误的有效修复能力。研究发现，该机制的效果不仅取决于规则本身的设计方向，还受评审者状态转移路径与上游提案质量的影响。进一步引入显式选择接口（保留、替换、升级）后，虽能消除Qwen模型对正确提案的污染，但Mistral仍会替换全部48个正确提案，且对仅涉及福利改善的修复拒绝率远高于Mistral，表明界面设计虽能引导评审行为，却无法作为模型无关的通用保护机制。最终结论是：受限审查并非普遍更优，真正决定纠正范围与错误污染面的是评审者能力、动作接口设计与绑定管辖权三者之间的协同作用。
链接: https://arxiv.org/abs/2611.05390
作者: Fuad Ali
机构: 未知
类目: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注: 11 pages, 3 figures, 3 tables. Code, frozen outputs, and reproduction instructions: this https URL
Abstract:Multi-agent language-model systems often add a reviewer to improve an upstream agent’s decision. Yet a reviewer can both repair errors and replace decisions that were already correct. We study this as a problem of institutional jurisdiction: conditional on the same proposal, task information, model, and review output, when is the reviewer authorized to make its recommendation binding? We compare broad override authority with an evidence gate that permits replacement only after the reviewer produces a mechanically valid record of a binding-constraint violation and a compliant repair. Across two locally run open-weight models, 96 spatial-policy tasks, and a 96-task portfolio-transfer benchmark, the gate sharply reduces corruption of correct proposals while retaining most repairs of violations. The directions of this tradeoff are partly implied by the rule; its magnitude depends on reviewer transitions and upstream proposal states. A prospective validation then gives the reviewer an explicit choice to retain, replace, or escalate. This eliminates correct-proposal corruption for Qwen, but Mistral still replaces all 48 correct proposals; Qwen also declines every welfare-only repair that Mistral frequently attempts. Thus interface design changes reviewer behavior but is not a model-independent safeguard. The result is not that restricted review is universally better. It is that reviewer competence, action interface, and binding jurisdiction jointly determine both the correction domain and the corruption surface.
[MA-9] Soft-Label Governance for Distributional Safety in Multi-Agent Systems
【速读】：该论文旨在解决多智能体人工智能系统中涌现的复杂风险问题，即单个智能体在孤立状态下不会产生，但在协作或交互过程中可能引发的系统性安全风险。现有安全评估框架依赖于二元分类（如“安全”/“不安全”），忽略了基于代理（proxy-based）评估中的不确定性，从而导致对真实风险的误判。为此，本文提出SWARM（System-Wide Assessment of Risk in Multi-agent systems）仿真框架，其核心创新在于用软概率标签 $ p = P(v=+1) \in [0,1] $ 替代传统的二元标签，实现连续值收益计算、毒性度量与治理干预的精细化建模。该框架配备模块化治理引擎，可配置多种调控机制（如交易税、熔断机制、声誉衰减和随机审计），并通过概率性指标（如期望毒性 E[1−p∣accepted] 与质量差距 E[p∣accepted]−E[p∣rejected]）量化治理效果。实验表明，严格的治理虽未提升安全性，却使福利水平下降超40%；而过度内部化外部性会导致系统总福利从基准+262骤降至-67，毒性水平却保持不变。熔断机制需精细校准：阈值过严会严重损害系统价值，而最优阈值可在适度福利与最低毒性间取得平衡。配套实验进一步验证了软度量能有效识别自优化智能体对传统二元评估的代理博弈行为。该基础治理层可直接应用于实时大语言模型驱动的智能体（如Concordia实体、Claude、GPT-4o Mini），无需修改。研究结果表明，实现分布式的系统安全必须依赖连续的风险度量，且治理杠杆的校准本质上涉及可量化的安全-福利权衡。源代码与项目资源已公开。
链接: https://arxiv.org/abs/2604.19752
作者: Aizierjiang Aiersilan,Raeli Savitt
机构: The George Washington University (乔治·华盛顿大学); SWARM AI Safety; New York, USA
类目: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Computers and Society (cs.CY); Systems and Control (eess.SY)
备注:
Abstract:Multi-agent AI systems exhibit emergent risks that no single agent produces in isolation. Existing safety frameworks rely on binary classifications of agent behavior, discarding the uncertainty inherent in proxy-based evaluation. We introduce SWARM (\textbfSystem-\textbfWide \textbfAssessment of \textbfRisk in \textbfMulti-agent systems), a simulation framework that replaces binary good/bad labels with \emphsoft probabilistic labels p = P(v=+1) \in [0,1] , enabling continuous-valued payoff computation, toxicity measurement, and governance intervention. SWARM implements a modular governance engine with configurable levers (transaction taxes, circuit breakers, reputation decay, and random audits) and quantifies their effects through probabilistic metrics including expected toxicity \mathbbE[1-p \mid \textaccepted] and quality gap \mathbbE[p \mid \textaccepted] - \mathbbE[p \mid \textrejected] . Across seven scenarios with five-seed replication, strict governance reduces welfare by over 40% without improving safety. In parallel, aggressively internalizing system externalities collapses total welfare from a baseline of +262 down to -67 , while toxicity remains invariant. Circuit breakers require careful calibration; overly restrictive thresholds severely diminish system value, whereas an optimal threshold balances moderate welfare with minimized toxicity. Companion experiments show soft metrics detect proxy gaming by self-optimizing agents passing conventional binary evaluations. This basic governance layer applies to live LLM-backed agents (Concordia entities, Claude, GPT-4o Mini) without modification. Results show distributional safety requires \emphcontinuous risk metrics and governance lever calibration involves quantifiable safety-welfare tradeoffs. Source code and project resources are publicly available at this https URL.
[MA-10] Decision Models as Scalable Economic Surrogates that Capture Human Variation
【速读】：该论文旨在解决当前大型语言模型（LLMs）在经济模拟与调查中作为人类代理（human surrogates）所面临的两大核心问题：一是其高昂的计算成本限制了大规模应用，二是其对人类行为变异性的表征不足。为此，论文提出以决策模型（Decision Models, DMs）作为替代方案，其关键在于利用在通用决策基准上表现优异的“前沿DMs”（frontier DMs），这些模型不仅在响应速度上显著优于最高效的LLMs（约快十倍），且在成本上可降低超过百倍，同时在三种经济场景——家庭消费调查、经典经济博弈及劳动力市场实地实验中，更准确地捕捉了人类行为的异质性。研究结果表明，尽管前沿DMs在经济博弈中的表现略逊于LLMs，但在消费决策等情境下与人类行为模式高度一致，整体上展现出比LLMs更强的行为多样性建模能力，从而为大规模经济仿真提供了更高效、更真实的可行路径。
链接: https://arxiv.org/abs/2611.05553
作者: Maxime Saxena,Fabio Caccioli,R. Maria del Rio-Chanona
机构: University College London (伦敦大学学院); London School of Economics and Political Science (伦敦政治经济学院); Google DeepMind
类目: General Economics (econ.GN); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
备注:
Abstract:Decision Models (DMs), a new class of AI models, give fast responses and return a probability distribution over a fixed set of options. These properties make them candidate human surrogates in economic simulations and surveys, a role increasingly filled by Large Language Models (LLMs), which are costly at scale and under-represent the variation in human behaviour. We present a first evaluation of DMs as human surrogates in economics, comparing 10 DMs with 8 LLMs in three economic settings: a household spending survey on consumption, several classic economic games and a labour-market field experiment. Frontier DMs, those that score highest on general decision-making benchmarks, reproduce the variation in human responses more closely than LLMs in all three settings. They capture the modal human response as well as LLMs in the spending survey, but less well in the economic games. Frontier DMs are also roughly ten times faster than the fastest LLM and can be over a hundred times cheaper. Together, these results suggest that DMs capture the heterogeneity of human behaviour better than LLMs and make large-scale economic simulations more feasible.
自然语言处理
[NLP-0] SpecCompact: Accelerating LLM Agent Context Compaction with Slot-Parallel Semantic Speculative Verification
【速读】： 该论文旨在解决大语言模型（LLM）智能体在处理持续增长的交互历史时，因使用大型模型进行上下文压缩而导致显著延迟的问题。核心挑战在于如何在保持信息完整性与降低延迟之间实现有效平衡。其解决方案的关键在于提出一种名为SpecCompact的框架，通过槽位并行语义推测（slot-parallel semantic speculation） 实现压缩加速。具体而言，SpecCompact微调一个小型草稿模型（drafter），并行生成结构化摘要槽位（structured summary slots），由目标模型对每个槽位进行验证与修复，同时后续生成基于更新后的槽位前缀进行。该方法结合系统级优化，减少重复的上下文处理和模型调用。实验结果表明，在AppWorld和τ³基准测试中，相较于传统的Hermes式摘要方法，SpecCompact实现了2.04–5.63倍的压缩速度提升；用户研究进一步显示，在Hermes系统内实现8.68倍的中位压缩速度提升，验证了其高效性与实用性。
链接: https://arxiv.org/abs/2611.06876
作者: Yulu An,Yifei Feng,Xiangyang Yin
机构: New York University (纽约大学); SpecCompact
类目: Computation and Language (cs.CL)
备注:
Abstract:LLM agents use context compaction to manage growing interaction histories, but summarizing long histories with large models can introduce substantial delays. Compaction therefore requires balancing information retention against latency. We observe that structured summary slots can retain information when generated independently. Based on this observation, we present SpecCompact, a framework that accelerates compaction through slot-parallel semantic speculation. SpecCompact fine-tunes a small drafter and generates slots in parallel. The target model verifies each proposed line and repairs rejected content, while subsequent drafting uses the updated slot prefix. System optimizations further reduce repeated context processing and model calls. Across AppWorld and \tau^3 benchmarks with two target models, SpecCompact achieves 2.04 – 5.63\times compaction speedups over Hermes-style summarization. Separate downstream evaluations reveal model- and task-dependent performance trade-offs. A user study within Hermes further shows an 8.68\times median compaction speedup.
[NLP-1] LongBanana: An Expert-Verified Benchmark for Long-Context Multi-Reference Image Synthesis NEURIPS2026
【速读】： 该论文旨在解决在代理驱动的图像生成时代，长上下文多参考图像合成任务中缺乏权威评估基准的问题。具体而言，该任务要求系统从分散的视觉证据中检索相关信息、排除干扰项、准确绑定实体与关系，并基于多个参考图像生成一张语义连贯的最终图像。其解决方案的关键在于提出LongBanana——首个由专家设计并验证的长上下文多参考图像合成基准，包含90个精心设计的任务，采用3×3分类体系，每个任务配备约32张参考图像、20个约束条件和50个评估项，共计4,670项专家指定的原子级评估指标。该基准采用图结构引导的评估设计，将实体-关系图中的节点、边及整体图结构的义务转化为内容保真度（Content Fidelity）、关系绑定（Relational Binding）和全局约束满足度（Global Constraint Satisfaction）三类检查清单项，实现可审计的量化评估。此外，论文还引入LongBanana-Gen框架，通过代理调度机制支持长周期图像生成，实验表明最小化代理调度即可提升生成效果，而基于ReAct-Plan-Exec-Reflect（RPR）的工作流则通过强化证据锚定与自我修正能力，实现依赖于代理架构的性能增益。总体而言，LongBanana为衡量密集视觉证据利用能力提供了可复现的测试平台，并成为推动更强代理驱动图像生成系统发展的共同目标。
链接: https://arxiv.org/abs/2611.06856
作者: Haoxiang Cao,Yuxuan Zhang,Penghui Du,Bo Li,Jiajiong Cao,Yingying Fan,Zimeng Wu,Yifu Luo,Mingzhe Zheng,Junwen Miao,Huaisong Zhang,Yang Yao,Yu Wang,Dongfu Jiang,Ping Nie,Wenhu Chen,Chaoqun Wang,Changqian Yu,Kelsey R. Allen
机构: South China Normal University(华南师范大学); University of British Columbia(不列颠哥伦比亚大学); Vector Institute(向量研究所); Canada CIFAR AI Chair(加拿大魁北克人工智能主席); Carnegie Mellon University(卡内基梅隆大学); University of Waterloo(滑铁卢大学); KlingAI Research(金莱人工智能研究); University of Maryland(马里兰大学); Tsinghua University(清华大学); HKUST(香港科技大学); HKU(香港大学)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: NeurIPS 2026 Evaluations and Datasets Track. Project page: this https URL
Abstract:We introduce LongBanana, to our knowledge the first expert-designed and expert-verified benchmark for long-context multi-reference image synthesis in the age of agent-driven image generation. The task requires retrieving dispersed visual evidence, rejecting distractors, binding entities and relations, and composing one coherent final image from multiple references. LongBanana contains 90 carefully designed tasks organized by a 3-by-3 taxonomy; each task includes about 32 reference images, 20 constraints, and 50 evaluation items, yielding 4,670 expert-specified atomic evaluation items in total. Its graph-grounded evaluation design instantiates each task with reference images and an entity-relation graph, then converts node, edge, and whole-graph obligations into checklist items for Content Fidelity, Relational Binding, and Global Constraint Satisfaction. Reviewed expert labels include a 15% double-labeled subset with 89.4% exact agreement and 0.83 quadratic weighted kappa. We further introduce LongBanana-Gen, an agent-harness framework for long-horizon image generation; experiments show that minimal agent harnesses improve over direct generation, while a ReAct-Plan-Exec-Reflect (RPR) workflow yields harness-dependent gains by strengthening evidence grounding and repair. Together, LongBanana provides an auditable testbed for measuring dense visual evidence use and a shared target for developing stronger agent-driven image-generation systems. Code and data are available at the project page: this https URL.
[NLP-2] Persona Perspectives: Adopting Viewpoints Reorganizes Representations in LLM s
【速读】： 该论文旨在解决当前大型语言模型（LLM）特征监测方法在动态上下文环境下失效的问题，特别是当模型因提示（prompting）而采用不同政治立场“人格”（persona）时，其潜在表示空间的几何结构是否发生系统性变化。传统监测方法如线性探针和稀疏自编码器（Sparse Autoencoders, SAEs）通常假设特征表示具有上下文不变性，但本文发现，当模型被引导至自由派或保守派人格时，其隐层空间中词向量在语义轴上的投影以及语义轴之间的相对对齐关系会呈现出与指定意识形态一致的系统性重构。这一表征重组织显著影响了对安全相关特征（如真实性与道德性）的评估：基于无人格训练的线性探针、SAE分类器及人工选取词表示构建的分类器，在应用于人格化模型时产生明显偏倚的评分结果。因此，该研究的关键发现是，模型内部表示的几何关系并非静态固定，而是受人格条件调节，这揭示了现有特征监测范式在动态人格情境下的局限性，进而挑战了基于静态表示进行模型安全监控的有效性。
链接: https://arxiv.org/abs/2611.06848
作者: Austin C. Kozlowski,Avital Mintz,Madison Lee
机构: Knowledge Lab; University of Chicago (芝加哥大学)
类目: Computation and Language (cs.CL)
备注: 26 pages, 14 figures
Abstract:Common methods for monitoring features in large language models (LLMs) such as linear probes and sparse autoencoders (SAEs) assume that features maintain a consistent representation across contexts. LLMs, however, can flexibly adopt different perspectives or “personas” that suggest different underlying systems of conceptual relations. In this paper, we examine whether prompting an LLM to adopt a persona alters the representational geometry of the model’s latent space. We extract representations of over 9,000 common words and 33 semantic axes from models adopting liberal and conservative political personas and compare these against two neutral baselines. We find that persona conditioning systematically alters both the projections of words onto semantic axes and the alignments among the axes themselves in directions consistent with the specified ideology. We then test the implications of this reorganization for two safety-relevant features: truthfulness and morality. Linear probes, SAE-based classifiers, and classifiers built from manually selected word representations, all trained without a persona, produce divergent results when applied to models conditioned on political personas, producing politically slanted ratings of truth and morality on controversial statements. These findings suggest that internal representations, as they are conventionally extracted, can exhibit different geometric relations when the model adopts personas, potentially undermining monitoring efforts.
[NLP-3] Before You Continue: Evaluating Time-Sensitive Safety Awareness in Conversational Assistants
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）在处理涉及时间依赖性约束的对话请求时，未能有效识别并应用这些约束的问题。具体而言，当用户在后续交互中未明确提及先前设定的时间限制时，助手往往无法自动推断或适配这些隐含的时间约束，从而导致给出不安全或不合逻辑的建议。其解决方案的关键在于构建一个名为TASC（Temporal Action Safety in Conversation）的基准测试框架，用于评估模型在对话中对时间依赖性约束的应用能力。TASC包含3,418个基于128个模板生成的输入，并额外设计了33个情景以区分模型在计算经过时间与实际应用该信息之间的表现差异。实验结果表明，尽管多数模型能够正确计算时间间隔，但在需根据时间约束调整建议的场景中，仅有30.5%的情况能正确应用约束，且62.1%的失败案例并非源于时间计算错误，而是源于将推理结果转化为适当对话响应的能力缺失。这揭示了当前大语言模型在时间推理与实际行为指导之间存在显著“应用鸿沟”，即模型具备一定的时序理解能力，但难以将其有效转化为符合时间约束的对话输出。
链接: https://arxiv.org/abs/2611.06839
作者: Peidi Dong,Yunze Xiao,Yu-Lin Tsai,Yichen Zheng,Heyuan Huang,Weihao Xuan,Mark Dredze,Jen-tse Huang
机构: Carnegie Mellon University; University of California, Berkeley; The Hong Kong Polytechnic University; Johns Hopkins University; The University of Tokyo; RIKEN AIP
类目: Computation and Language (cs.CL)
备注: Preprint
Abstract:Large language model (LLM) assistants increasingly help users with everyday activities that involve time-dependent constraints. However, users may not mention these constraints in later requests, requiring assistants to recognize and adjust their advice accordingly. To evaluate such capability, we introduce TASC (Temporal Action Safety in Conversation), a benchmark that evaluates whether LLMs apply time-dependent constraints when responding to conversational requests. TASC consists of 3,418 inputs derived from 128 templates, along with a separate set of 33 scenarios designed to compare models’ ability to calculate elapsed time with their ability to apply that information. Across 13 models, assistants account for time-dependent constraints in only 30.5% of cases where the proposed action violates a rule, while rarely warning users unnecessarily. More importantly, models fail to apply these constraints in 62.1% of cases despite correctly calculating the elapsed time. Explicitly providing the timing resolves only about half of these failures, revealing a gap between temporal reasoning and its application in conversational advice.
[NLP-4] LooM: Loop-Invariant Memory for Causal Encoder-Decoder Language Models
【速读】： 该论文旨在解决循环变压器（Looped Transformers）在实际应用中面临的三大瓶颈问题：随着循环次数增加，内存占用持续增长；计算量只能以整块为单位调整；训练完成后难以灵活更改循环次数。其提出的解决方案——LooM（Loop-invariant Memory），通过在因果编码器-解码器（CED）架构中仅对首个解码层之后的层进行循环，并引入一种全局记忆机制，该记忆在每个输入标记（token）生成时仅写入一次，且每个解码层仅在其首次访问时发布对应标记的键值（KV）缓存条目，从而实现缓存大小与循环次数无关。此外，这些循环层可独立运行，支持每标记3至21次不等的执行次数，且无需路由网络（router），仅需一个在随机采样预算下训练的统一模型即可覆盖所有配置。实验表明，在足够多的循环次数下，LooM在语言建模、零样本准确率和长上下文困惑度等任务上均优于同等数据训练的单次前向传播基线模型，同时保持相同的缓存开销。进一步分析揭示，训练预算分布决定了额外深度所能带来的收益。研究结果表明，CED结构天然适合循环设计：记忆仅写入一次，而深度则可作为部署阶段按需选择的计算预算。
链接: https://arxiv.org/abs/2611.06835
作者: Lele Yang,Muxi Diao,Zhonghao Yan,Jiankang Chen,Bin Wen,Tingting Gao,Wentao Zhang,Kongming Liang,Zhanyu Ma,Weiran Xu
机构: Beijing University of Posts and Telecommunications (北京邮电大学); Zhongguancun Academy (中关村学院); Peking University (北京大学); Kling Team (Kling团队)
类目: Computation and Language (cs.CL)
备注:
Abstract:Looped Transformers increase effective depth by re-executing shared layers, but three limitations hinder their adoption: memory grows with the loop count, compute can only be adjusted in whole-block steps, and the loop count is hard to change after training. We propose LooM (Loop-invariant Memory), which addresses each in turn. In a causal encoder-decoder (CED), LooM loops only the decoder layers after the first, which read a global memory written once per token, and each of them publishes the key-value (KV) entry for a token only on its first visit, so the cache does not grow with the loop count. These layers loop independently, providing every budget from 3 to 21 layer executions per token, and one model trained on randomly sampled budgets covers all of them without a router. With enough loops, LooM outperforms a single-pass baseline trained on the same data on language modeling, zero-shot accuracy and long-context perplexity, while using the same cache size. Our analysis further shows that the training budget distribution decides what extra depth buys. These results suggest that a CED is a natural place to loop: memory is written once, and depth becomes a budget chosen at deployment.
[NLP-5] SAUCE: Summary Analysis Using Conversation Entailment ACL
【速读】： 该论文旨在解决跨语言对话语音摘要（cross-lingual conversational speech summarization）中缺乏明确评估目标时，难以准确衡量摘要质量的问题，尤其关注在语音转录与翻译误差导致的信息丢失以及摘要中幻觉或遗漏等缺陷的检测。其解决方案的关键在于提出SAUCE——一种无参考、基于事实的评估框架，通过判断对话内容与摘要之间的语义蕴含关系（entailment），量化摘要的事实准确性与事实覆盖率。相较于黑箱大语言模型评估器或密集嵌入型指标，SAUCE具备内在可解释性：将评分映射至可验证的具体事实，使用户能够精准定位幻觉或遗漏点。这一特性不仅帮助开发者系统性地分析大语言模型（LLM）的行为模式，也为终端用户提供了一种在噪声环境下的可操作工具以验证摘要的可靠性。初步实验表明，SAUCE与人工评价具有高度一致性。
链接: https://arxiv.org/abs/2611.06812
作者: Man-Ling Sung,Hemanth Kandula,Jeff Ma,William Hartmann,Matthew Snover
机构: RTX BBN Technologies; Cambridge MA, USA
类目: Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
备注: Accepted and published to the ACL GEM workshop. Proceedings of the Fifth Workshop on Generation, Evaluation and Metrics (GEM), pages 364-377, San Diego, California, USA. Association for Computational Linguistics
Abstract:With the growing need for evaluating Large Language Models (LLMs) and their applications to speech, challenges persist in summarizing and evaluating conversations that lack a clear end goal. We introduce SAUCE – a reference-free, fact-based evaluation pipeline for cross-lingual conversational speech summarization. It measures the accuracy and the fact coverage of a summary through the entailment between conversation and text. We compare SAUCE against several popular summarization metrics and demonstrate the effectiveness of capturing information loss due to transcription and translation error and identifying broken summaries. Crucially, unlike black-box LLM evaluators or dense embedding metrics, SAUCE is inherently explainable: it maps summary scores to discrete, verifiable facts, allowing users to pinpoint exact hallucinations or omissions. We illustrate how this interpretability helps developers systematically profile LLM behaviors and gives end-users an actionable tool to verify summary accuracy in noisy, real-world conditions. Preliminary investigations show SAUCE strongly align with human judgment.
[NLP-6] LLM s Dont Know What They Know: Benchmarking Temporal Self-Knowledge for Financial Backtesting
【速读】： 该论文旨在解决生成式AI在金融回测中可能因“知识泄露”（knowledge leakage）而产生虚假性能表现的问题，即模型在训练过程中已接触过未来数据，导致其预测能力被高估。现有实践通过依赖供应商声明的知识截止日期、模型自述认知状态或指令其以历史时间点作答等方式试图规避此问题，但这些方法的有效性缺乏验证。论文提出的关键解决方案是构建FinDendro——一个涵盖1999至2026年美国经济事实的基准数据集，包含2,226个带时间戳的事实，覆盖月度宏观经济指标、月末股票价格及美联储决策。该基准设计了四种类型提示：数值询问、回忆与预测（区分声称知晓或猜测）、以及带有历史声明日期的预测任务，结合持久性与主导结果控制，使随机准确率维持在约0.27。实证结果显示，多数前沿模型的自我报告存在显著低估真实知识水平的情况，例如GPT-5在可控事实中正确率达76%，但仅声称知晓38%；且其对“声称未知”的事实仍能正确回答61%。此外，知识宣称可延续至实际知识边界后长达七个季度，而明确声明历史日期虽能有效消除前置知识声称，却未影响预测准确性。因此，仅依靠自述、指令或声称的知识截止日期无法确保模型在训练窗口内进行回测时无知识泄露。研究强调需通过行为测量法识别模型的时间知识边界，而FinDendro为评估和选择基于实际事实认知能力的回测时段提供了可操作的基准与流程。
链接: https://arxiv.org/abs/2611.06808
作者: Lingfei Qian,Yan Wang,Xueqing Peng,Yupeng Cao,Yuechen Jiang,Zhuohan Xie,Anke Xu,Wenbo Cao,Jiarui Chen,Zehan Li,Jimin Huang,Alejandro Lopez-Lira,Yuehua Tang,Sophia Ananiadou,Jian-Yun Nie
机构: 未知
类目: Computation and Language (cs.CL)
备注: 20 pages, 4 figures, 11 tables
Abstract:Financial backtests of large language models (LLMs) are meaningful only if the model did not already know the outcomes it is asked to predict. Practitioners guard against this by trusting a vendor’s stated knowledge cutoff, by asking the model what it knows, or by instructing it to answer as of a past date. To verify whether any of these safeguards can be trusted, we introduce FinDendro, a benchmark of 2,226 dated U.S. facts from 1999 to 2026, covering monthly macroeconomic series, month-end stock prices and Federal Reserve decisions. Each fact yields four prompts: a value question, Recall and Predict questions that pair each candidate value with a claim of knowing or guessing without a declared date, and a Predict question with a historical declared date. Controls for persistence and dominant outcomes keep chance accuracy near 0.27. Results show that for most frontier models, self-reports understate knowledge. GPT-5 answers 76% of controlled facts correctly but claims to know 38%, and answers 61% of the facts it says it does not know. Knowledge claims can persist beyond the measured knowledge boundary, by up to seven quarters, and a declared date removes most of the claims of prior knowledge while leaving accuracy unchanged. Backtests inside a model’s training window therefore cannot be certified leak-free by self-report, instruction or stated cutoff. These findings motivate behavioral measurement of temporal knowledge boundaries, and FinDendro provides a benchmark and procedure for selecting backtest periods based on demonstrated factual knowledge.
[NLP-7] Measuring Risk Not Harm: Detecting Psychological Harm Risk Factors in LLM Conversations
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）交互过程中可能引发的心理伤害（psychological harm）难以被直接监测的问题，尤其因为此类伤害往往在长期互动中逐步显现，并高度依赖于用户的主观个体情境。其核心解决方案在于将监测重点从“伤害”转向“风险”，即通过识别可观察的AI行为模式作为心理伤害的风险因素（risk factors），并在响应层面进行聚合分析，实现对用户-模型交互过程的纵向心理安全监控。关键创新在于构建了一个包含47,062条AI响应与373,873个响应-因子标签的合成多标签语料库，并定义了八类可操作化的风险行为；在此基础上，采用嵌入式分类器、直接大模型判断以及分类器-判别器级联三种方法进行检测，所有方法均为模型无关（model-agnostic），仅依赖于LLM输出，无需访问内部状态。实验表明，最优级联模型在独立测试集上达到宏平均精确率0.86、召回率0.76和F1值0.79，在WildChat数据集上使大模型判别调用次数减少97.1%，且在DelusionEval中风险因子检测结果呈现对话内聚性。研究验证了基于响应层行为风险因子的检测机制，能够为心理安全性提供可扩展、可量化的监测框架，避免将可观测的AI行为与潜在的用户内在伤害混淆。
链接: https://arxiv.org/abs/2611.06806
作者: Nil-Jana Akpinar,Eugenia Kim,Natalie Isak,Amanda Minnich,Victoria Westerhoff,Matthew Dressman
机构: Microsoft(微软)
类目: Computation and Language (cs.CL); Computers and Society (cs.CY)
备注:
Abstract:Psychological harms from interactions with large language models cannot be monitored directly because they may emerge over repeated interactions and depend strongly on subjective individual user context. We propose measuring risk, not harm: detecting observable AI behaviors that constitute risk factors for psychological harm at the AI response level and aggregating these signals to support longitudinal monitoring. We operationalize eight such behaviors and construct a synthetic multi-label corpus of 47,062 AI responses and 373,873 response-factor labels, with targeted human annotations for validation and evaluation. We then compare trained embedding-based classifiers, direct LLM judging, and a classifier–judge cascade. All detectors are model-agnostic and operate only on LLM outputs, requiring no access to internal model states. We apply these detectors to several out-of-distribution human-AI conversation data sets from prior research. On held-out human-labeled data, the best cascade achieves macro precision of 0.86, recall of 0.76, and F1 of 0.79. On WildChat, it reduces LLM-judge calls by 97.1% relative to direct judging. On DelusionEval, risk-factor detections cluster within conversations. Together, our results show that response-level behavioral risk-factor detection can provide a scalable measurement layer for longitudinal psychological-safety monitoring without conflating observable AI behavior with latent user harm.
[NLP-8] Ranking Agreement Does Not Reveal Parag raph-Edit Agreement in Two Released Paper Acceptance Predictors
【速读】： 该论文旨在解决如何高效筛选学术论文修订版本以供后续详细评估的问题，核心挑战在于验证快速接受分类器（fast acceptance classifier）是否能够有效预测哪些修订版本更可能被慢速评估者（slower evaluator）所接受。其解决方案的关键在于通过对比一个基于词与文档特征的分类器与两个从同一30亿参数语言模型微调得到的文本类接受预测器，在198篇国际学习表征会议（ICLR）论文及其自动筛选后的段落重写版本上进行评分一致性分析。研究发现，尽管分类器对原始论文的排名与其中一个预测器接近、与另一个中等一致，但在判断哪些重写版本会提升评分方面，分类器与两个预测器均无显著一致性。进一步分析表明，先前测试中因数值精度降低导致舍入误差主导了微小分数变化，因此在全32位精度下重新评分后，结果依然显示编辑选择与随机选择相比平均得分仅提高不足2分（满分100），且一致性仍趋近于零。该证据涵盖展示性重写及来自同一模型家族的两个开源预测器，但尚未验证其在实际高成本评估者中的适用性。因此，研究强调：任何用于筛选修订版本的分类器都必须在其预期评估者上进行实证校验，以确保其有效性。
链接: https://arxiv.org/abs/2611.06789
作者: Shaoliang Yang,Jun Wang
机构: Santa Clara University (圣克拉拉大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 38 pages including supplementary methods and results; 10 figures and 22 tables
Abstract:A fast acceptance classifier could select revisions for a slower evaluator. We ask whether agreement on existing papers supports that use. We compare a classifier of word and document features with two released text-only acceptance predictors fine-tuned from the same language model with three billion parameters. For 198 papers from the International Conference on Learning Representations, we compare model rankings of the original papers and of paragraph rewrites screened automatically for unchanged facts, claim strength, and attribution. The classifier ranks original papers closely with one predictor and moderately with the other, but shows no detectable agreement with either on which rewrites raise scores. The tests specified in advance used reduced numerical precision. Rounding error dominated the small score responses. We therefore rescored the released weights in full 32-bit precision as a post hoc check; edit agreement remained near zero. At this precision, when each predictor chooses among five rewrites per paper, its average raw score is under two points out of 100 higher than with random choice. The evidence covers presentation rewrites and two open predictors from one family; transfer to costly evaluators remains untested. Revision selectors should therefore be checked against the intended evaluator on the revisions they rank.
[NLP-9] mplate Gap Reconstruction for Shell Output Simulation with Small Language Models
【速读】： 该论文旨在解决小型语言模型（Small Language Models, SLMs）在执行命令前模拟壳环境输出时，因需完整预测重复性结构与动态内容而带来的高计算开销与生成效率低下的问题。其核心挑战在于如何在保持输出准确性的同时降低生成负担。解决方案的关键是提出模板间隙重建（Template Gap Reconstruction, TGR）机制：通过构建包含1870万条跨异构执行环境输出的模板库，将输出分解为可复用的固定文本片段（即模板）与由占位符（sentinels）分隔的自回归生成间隙；系统采用模板选择器动态决定使用哪个模板进行重建，其中空模板仍保留全输出解码能力。实验表明，当最优模板覆盖超过25%的输出时，TGR在输出编辑相似度（O-ES）上显著优于全输出解码（FOD），尤其在短间隙序列中表现更优（O-ES分别为0.964与0.812），同时在不同分词器下减少26.3%–38.3%的生成标记数。该方法验证了利用可复用输出结构作为实际设计选择的有效性，兼顾响应保真度、生成目标缩短，并将模板选择与残差预测作为影响生成质量的核心要素。
链接: https://arxiv.org/abs/2611.06731
作者: Jarrod Ragsdale,Rajendra Boppana
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 25 pages, 7 figures, 12 tables, conference preprint
Abstract:Small language models offer a compact basis for shell simulation without executing submitted commands, but full-output decoding requires them to predict recurring utility structure alongside content that varies across executions. We introduce template gap reconstruction (TGR), which supplies reusable output literals interspersed with sentinels to be interleaved with autoregressively generated gaps. A template selector chooses which template to be reconstructed, with the empty template retaining full-output decoding. We construct a template bank from 18.7 million outputs collected across heterogeneous execution environments and train a template selector and five reconstruction models initialized from various small language model checkpoints on a target environment excluded from bank construction. Evaluation demonstrates that TGR improves output edit similarity (O-ES) over full-output decoding (FOD) when the optimal template covers more than 25% of the template, while performing similarly for lower-coverage outputs. Additionally, TGR outperforms FOD for shorter gap sequences, with O-ES values of 0.964 and 0.812, respectively, while maintaining similar performance for longer sequences. TGR reduces the number of tokens generated by 26.3–38.3% across used tokenizer families. These results support using a reusable output structure as a practical design choice for small shell simulators, improving response fidelity and shortening generation targets while keeping template selection and residual prediction central to response quality.
[NLP-10] urns Matter: Advancing Streaming Video RL via Turn-Level Credit
【速读】： 该论文旨在解决流式视频理解任务中因采用轨迹级（trajectory-level）优势函数导致的决策不一致问题，即现有基于组相对策略优化（GRPO）的方法在多轮交互中共享同一轨迹级优势信号，造成早期成功后冗余回复增强、早期失败后有用后续响应被抑制的现象。其解决方案的关键在于提出一种逐轮（turn-wise）GRPO框架——TurnWise，通过引入逐轮信用分配机制，利用“静默完成前缀评分”（silence-completed prefix scoring）构建每一轮决策的剩余回报（remaining return），从而实现从前期奖励中解耦的独立学习信号。同时，采用逐轮均值策略优化（turn-mean policy optimization），在每轮内平均词元损失，并跨轮次进行加权平均，使回应与等待决策获得同等权重，不受输出长度影响。上述设计共同提升了流式交互中决策的时效性与信息密度，无需额外的值函数网络或滚动采样即可显著改善策略学习效果，在多个基准测试中优于现有强化学习方法。
链接: https://arxiv.org/abs/2611.06728
作者: Shuai Dong,Kanghui Tian,Linghan Chen,Wenqi Shao,Hongrui Wu,Haowen Hou,Yongfu Zhu,Xuehui Wang,Ziyue Wang,Wenpu Liu,Weichu Xie,Yuqi Xu,Congcong Wang,Kaiwen Tuo,Caoyuan Ma,Taiqiang Wu,Ji Zhao,Wenzheng Chang,Dianyi Wang,Siyuan Wang,Zhongyu Wei,Yu Qiao,Jiaqi Wang
机构: 未知
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 16 pages, 4 figures
Abstract:Streaming video understanding requires models to repeatedly decide whether to respond or continue observing. Recently, Group Relative Policy Optimization (GRPO) has been applied to this setting, with existing approaches emphasizing reward design while retaining a single trajectory-level advantage across turns. Such shared credit can reinforce redundant replies after earlier successes and discourage useful follow-ups after earlier failures. We propose TurnWise, a turn-level GRPO framework that aligns learning with individual streaming decisions. TurnWise addresses the mismatch between trajectory-level feedback and turn-level interaction, promoting informative responses and reducing redundancy. Specifically, turn-level credit assignment uses silence-completed prefix scoring to construct group-relative advantages from each decision’s remaining return, separating its learning signal from earlier rewards. Turn-mean policy optimization averages token losses within each turn and then across turns, giving response and waiting decisions equal weight regardless of output length. These complementary designs improve streaming policy learning without a learned critic or additional rollouts. Extensive experiments show our method outperforms prior RL approaches for streaming video understanding.
[NLP-11] Reasoning over Non-Parametric Memory in Multilingual LLM s: A Cognitive-Based Analysis of Relational Knowledge
【速读】： 该论文旨在探究生成式 AI（Generative AI）在给定明确事实条件下的关系推理能力，即模型能否基于提示中直接提供的关系信息进行逻辑推断，而非依赖预训练阶段记忆的关联模式。其核心问题是：当前大语言模型在处理需要深层次推理的任务时，是否真正具备对显式提供关系证据的建模与推理能力。解决方案的关键在于构建一个受控的多语言诊断基准——Rel-Bloom，该基准包含450道多选题，基于30个Wikidata关系，以五种认知层级（记忆、理解、应用、分析、评估）和三种真实性情境（事实性、反事实性、中性化）进行系统设计。通过在43个模型、六种语言及两种多语言设置下进行大规模测试（共226,800次预测），研究发现：模型性能在“应用”到“分析”层级间显著下降，尤其在大型开源模型中降幅达47.2个百分点；“评估”类任务始终是瓶颈，即使在其他层级接近满分的Gemini 3.1 Pro，其“评估”得分也仅为60.7%；且不同语言间题目与关系的难度排序高度一致，表明存在跨语言共享的内在难度结构。此外，反事实情境普遍更难，而去除实体身份信息在某些模型中反而有助于提升高层级表现。这些结果表明，模型规模或能力提升并不能自动保证其对显式关系证据的可靠推理，揭示了当前生成式 AI 在复杂推理任务中的深层局限性。
链接: https://arxiv.org/abs/2611.06715
作者: Zohreh Abbasi Liasi,Batoul Hammoud,Anas Madkoor,Ehsaneddin Asgari
机构: Qatar Computing Research Institute (QCRI), Hamad Bin Khalifa University(卡塔尔计算研究研究所，哈马德本哈利法大学); Universitat Pompeu Fabra(庞培法布拉大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Published at COLM 2026
Abstract:Retrieval-augmented generation (RAG) assumes that language models can reason over relations stated in their context rather than fall back on associations memorized during pretraining. We study this evidence-conditioned relational reasoning in isolation, with the facts supplied directly in the prompt rather than retrieved. \textscRel-Bloom is a controlled multilingual diagnostic of 450 multiple-choice questions built from 30 Wikidata relations; the same source relations are instantiated at five Bloom-inspired levels (Remember, Understand, Apply, Analyze, Evaluate) and under three veracity conditions (Factual, Counterfactual, Neutralized). Across 43 models, six languages, and two multilingual settings (226,800 predictions), three findings stand out. First, accuracy drops sharply from Apply to Analyze in every evaluated open-weight model series, reaching 47.2 percentage points for the Large Granite bucket, but not for either Gemini model. Second, Evaluate items remain a bottleneck even for models near ceiling on the other task types: Gemini~3.1 Pro averages 99.5% on the other four levels but 60.7% on Evaluate. Third, item- and relation-level difficulty rankings are consistently preserved across languages, indicating a difficulty structure partly shared across languages within this benchmark. Across knowledge conditions, Counterfactual contexts are generally harder than Factual ones, while removing entity identities can improve higher-level performance in some model regimes. Greater model capability therefore does not by itself ensure reliable reasoning over supplied relational evidence. Relevant resources are available at: this https URL.
[NLP-12] ALTO: African and Levantine Tech-Facilitated Gender-Based Violence Corpus
【速读】： 该论文旨在解决生成式AI（Generative AI）时代下，技术赋能的性别暴力（Tech-facilitated Gender-Based Violence, TFGBV）在多元语言与文化语境中检测能力不足的问题，尤其针对全球南方（Global South）语言如黎凡特阿拉伯语、非洲法语和斯瓦希里语中存在的性别化伤害内容缺乏有效识别模型与标注数据集的困境。其解决方案的关键在于采用由社区主导的众包方法，联合约旦、西非法语区及肯尼亚的民间社会组织（CSOs），收集具有文化与语言相关性的TFGBV实例，并基于主动学习（active learning）对数据进行标注，构建了一个小规模但性能优异的数据集ALTO（N=3022）。通过基准测试与微调实验表明，现有有害内容检测模型在ALTO数据集上对TFGBV的召回率表现不佳，经微调后显著提升；同时发现模型普遍遗漏隐性形式的性别暴力，且在斯瓦希里语中的错误率尤为突出。研究不仅报告了实证结果，还公开发布了注释分类体系、数据集及训练模型，为提升多语言环境下性别暴力内容的检测能力提供了可复用的技术基础与社会参与范式。
链接: https://arxiv.org/abs/2611.06706
作者: Tianwei Zhang,Ivy Njoki Nyagwencha,Zahraa Dirani,Shalini Joshi,Scott A. Hale
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (cs.LG)
备注: 8 pages main body, 10 pages appendix, 1 figure
Abstract:Tech-facilitated gender based violence (TFGBV) is a pervasive online harm affecting women and gender minority groups through its many direct and covert forms. Current models and annotated datasets for gendered harm detection are dominated by data from, and models trained in, Global North languages. In collaboration with civil society organizations (CSOs) from Jordan, Francophone West Africa, and Kenya, we developed a community-led approach to crowdsource culturally and linguistically relevant instances of TFGBV in Levantine Arabic, African French, and Swahili. We annotated the crowdsourced data using active learning to develop a small but performant dataset (N=3022) called ALTO to improve TFGBV detection in the three languages. Through benchmarking and fine-tuning experiments, we show that existing harmful content detection models show poor TFGBV recall within ALTO and improve with fine-tuning. We found that models often miss covert harms across all languages, while error rates are markedly higher in Swahili. We report the experimental results and release the annotation taxonomy, dataset, and trained models.
[NLP-13] Beyond accuracy: what AI assurance needs to establish in public services
【速读】： 该论文旨在解决公共部门在决策与服务提供中依赖生成式AI（Generative AI）输出时所面临的可靠性与问责性问题，特别是如何评估这些自动化生成的分类、摘要与评估结果是否可信。其核心挑战在于：现有基于整体准确率的评估方法无法充分揭示模型输出的不稳定性、系统性偏差及治理盲区。解决方案的关键在于构建一个融合技术评估与治理分析的双重框架，结合测量理论与基于管理的监管机制，强调超越平均准确率的多维验证。研究通过对比八种GPT-4o配置对1200次英国首相问答的分类表现，发现尽管整体准确率在68.2%至70.3%之间，但约24.3%的问题标签存在显著变动，且不同配置间一致性高于与人工标注的一致性，表明高一致并不等同于正确。同时，基于评分提取的需人工复核比例差异巨大（0.9%–41.0%），暴露出评估流程的不可靠性。进一步分析37份英欧政策文件发现，虽均要求测试与监督，但对输出变化、共现错误、选择性评估、实质性结论及审查路径等关键治理要素缺乏一致规范。因此，论文主张公共政策应强制要求组织通过多维度核查——包括非平均性能指标、明确责任归属、专家可及性以及结果的透明审查——来正当化对生成式AI证据的依赖，以确保公共治理的可信性与问责性。
链接: https://arxiv.org/abs/2611.06655
作者: Louise Yorke
机构: University College London (伦敦大学学院)
类目: Computers and Society (cs.CY); Computation and Language (cs.CL)
备注: 20 pages, 6 tables, 2 figures
Abstract:Public bodies increasingly rely on AI-produced evidence, including classifications, summaries and assessments. These outputs can inform policy decisions and service delivery, making their reliability a matter of public accountability. Establishing whether this evidence can be relied upon requires testing its outputs and examining how organisations decide whether those tests are sufficient. We combine technical evaluation with analysis of AI governance requirements, informed by measurement theory and management-based regulation. We compare eight GPT-4o configurations classifying 1,200 UK Prime Minister’s Questions into 21 policy topics, against human-coded labels and across unchanged repeats. Accuracy ranges from 68.2% to 70.3%, yet 24.3% of questions receive different labels, with changes in topic shares and rankings. Configurations agree more with one another than with human coding: agreement does not guarantee correctness. Under the implemented score-extraction procedure, the proportion flagged for human review ranges from 0.9% to 41.0%. We analyse 37 UK and EU policy documents. Both sets expect testing and oversight, but reporting expectations do not consistently cover changing answers, shared errors, selective evaluation, substantive conclusions or review routing. Some UK guidance explicitly addresses stability. The UK sample emphasises public bodies checking their own use; the EU places substantial testing duties on providers while leaving them important choices about measures and acceptance thresholds. Without appropriate checks, evidence could misrepresent policy issues, redirect attention or resources, and leave errors undetected. Public policy should require organisations to justify reliance through checks beyond aggregate accuracy, clear responsibility, access to expertise and scrutiny of results.
[NLP-14] Replicability Reproducibility and Robustness in Political and Social Science Research with Large Language Models
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）研究中因输出的随机性导致的可复现性问题。由于LLM输出具有概率特性，且在不同模型及版本间存在差异，这种不确定性可能对下游统计结果产生显著影响。论文区分了可重复性（replicability）、可复现性（reproducibility）与鲁棒性（robustness）的概念，并提出一种基于结果分布而非完全相同输出的 probabilistic 标准。其解决方案的关键在于引入标准化提示脚本（standardized prompt script），测试编码器与解码器内部的模型变异性，并发展误差结构分析方法以诊断模型错误模式及其一致性。通过三个数据集的实证研究发现，测量方式的选择可能导致下游回归系数符号反转。尽管基于设计的校正方法（如使用金标准标注子样本）可恢复估计的有效性，但会牺牲估计精度。本文的核心贡献是构建了一个基于开源权重的堆叠集成模型（stacked ensemble of open-weight models），在保持估计有效性的同时显著提升了校正步骤中的估计精度。
链接: https://arxiv.org/abs/2611.06639
作者: Joan C. Timoneda,Sebastián Vallejo Vera
机构: Purdue University (普渡大学); University of Western Ontario (西安大略大学)
类目: Computation and Language (cs.CL); Computers and Society (cs.CY)
备注: Chapter 5 of “LLMs in Political and Social Science: Principles and Applications of Transformer Models,” Oxford University Press, forthcoming 2027. Data and code: this https URL
Abstract:This chapter addresses the central reproducibility problem of research with large language models. LLM output is probabilistic and varies across models and versions, which can have significant effects on downstream statistical results. The chapter distinguishes between replicability, reproducibility, and robustness, and proposes a probabilistic standard based on the distribution of results, not identical outputs. It introduces a standardized prompt script, tests within-model variation for encoders and decoders, and develops error-structure analysis to diagnose where models err and whether their errors align. Using three datasets, the chapter shows that measurement choice can reverse the sign of downstream regression coefficients. Existing design-based corrections with gold-labeled subsamples restore validity, but come at the cost of estimate precision. The main contribution of the chapter is a stacked ensemble of open-weight models that improves precision in the correction step while retaining estimate validity.
[NLP-15] When Do Summaries Become Instructions? Unsupported Policy in LLM Agents Self-Written Handoff Summaries
【速读】： 该论文旨在解决生成式 AI（Generative AI）在长期运行的大语言模型（LLM）智能体中，因上下文压缩导致的“权限与要求误传”问题，即智能体在生成摘要时可能引入原始记录中未支持的权限或限制性语句。其核心问题是：这些手递手（handoff）摘要中的限制或授权声明是否超出原始上下文所允许的范围，是否影响后续智能体实例（继承者）的行为，以及此类语句的来源为何。解决方案的关键在于通过构建合成工具使用任务以建立真实权限清单（ground-truth permission ledger），并采用同快照重放（same-snapshot replay）方法，将摘要中的可疑语句替换为中性填充物，从而评估其对行为的影响。实验结果表明，两个开源权重模型的可读摘要中分别存在1.2%和2.4%的不支持限制、0.1%和1.3%的不支持授权；进一步干预显示，通过修改系统提示，明确说明需求来源，可显著降低不支持性限制的比例（从5.0%降至1.8%，持例模板下从1.8%降至0.3%），且统计显著（p < 10⁻⁹）。此外，大多数开发阶段的限制源于智能体自身先前推理中“需审批”的逻辑链条。该研究揭示了上下文压缩机制中隐含的语义偏移风险，并提出通过显式规则溯源来缓解此类问题的策略。
链接: https://arxiv.org/abs/2611.06545
作者: Wenhui Chu
机构: University at Albany, State University of New York(纽约州立大学阿尔巴尼分校)
类目: Computation and Language (cs.CL)
备注: 22 pages, 3 figures, 15 tables. Code and data: this https URL
Abstract:Long-running LLM agents compact their context, replacing most history with a summary. We ask whether handoff summaries state permissions or requirements the record does not support, whether such sentences change what the next instance (the successor) does, and where they come from. Synthetic tool-use tasks carry a ground-truth permission ledger; same-snapshot replay reruns the successor with those whole sentences replaced by neutral filler. Two AI readers, reading only rule-flagged held-out summaries, confirmed unsupported restrictions in 1.2% and 2.4% of all readable summaries from two open-weight models, and unsupported authorizations in 0.1% and 1.3%. For restrictions, a pre-specified replay test was not significant (p = 0.58); a follow-up, planned after it, found a fragile effect (p = 0.027). Neutralizing false approvals cut unauthorized actions from 24 to 7 of 120 held-out runs (post hoc; later set: 17 to 11 of 48). Most development restrictions followed the agent’s earlier reasoning that an action needs approval (post hoc). In pre-specified tests, replacing the system prompt’s “Some actions need approval.” with one saying where requirements come from cut unsupported restrictions from 5.0% to 1.8% of development summaries and, on held-out templates, from 1.8% to 0.3% (both p 10^-9), while requests for approvals the ledger requires barely changed. Code and data: this https URL
[NLP-16] Dialect Matters: A Dialect-Aware Baseline for Igala-English Machine Translation
【速读】： 该论文旨在解决伊加拉语（Igala）这一低资源语言在自然语言处理（Natural Language Processing, NLP）领域缺乏平行语料库的问题，尤其针对伊加拉语内部存在的方言差异尚未被系统建模的挑战。现有研究虽已尝试缓解低资源语言的瓶颈，但伊加拉语仍无任何公开可用的双语平行语料库。为此，本文提出首个面向伊加拉语-英语机器翻译的方言感知基线模型，首次构建了带有方言标签的1137句语料库（其中训练集899句），涵盖伊达（Idah, 40.4%）、安帕（Ankpa, 26.1%）、伊巴吉（Ibaji, 14.1%）、其他（Other, 16.4%）及奥古古（Ogugu, 2.8%）等方言变体。关键解决方案在于引入方言标签并采用混合方言训练策略，实证表明：仅使用伊达方言训练的模型在伊巴吉方言上性能下降4.7 BLEU，而混合方言训练则实现更均衡的表现（BLEU 11.66），验证了方言多样性融合对提升模型公平性的重要性。模型基于Helsinki-NLP/opus-mt-en-mul（78M参数）进行微调，经消融实验确定最佳收敛周期为8轮，超过15轮出现过拟合；最终达到BLEU 11.66、chrF 46.69的性能指标。语料规模与性能趋势符合非洲低资源语言文献中的对数线性规律，本研究首次提供了伊加拉语的方言影响分析，并建立了未来相关工作的基准。
链接: https://arxiv.org/abs/2611.06517
作者: Etuh E.,Eli. A. Jiya
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:The massive permeability of intelligent systems in every sphere of human activities have greatly enhanced clarity, ease of work, and increased general global understanding which has significantly awaken the use of intelligent systems by all users. Intelligent systems greatly rely on precise dataset/corpus for the correctness of their output. Despite efforts by researchers to address Low-resource language in the Natural Language Processing (NLP), Igala still has zero parallel corpora despite efforts by researchers. We present a dialect-aware baseline for Igala-English machine translation, among first open systems and first with dialectal variation analysis. Our dataset of 1137 sentences (899 training) is first with dialect tags: Idah 459 (40.4%), Ankpa 297 (26.1%), Ibaji 160 (14.1%), Other 187 (16.4%), Ogugu 32 (2.8%). Same English yields 20-60 Igala forms. Mixed-dialect training achieves balanced BLEU 11.66 versus Idah-only model which drops 4.7 BLEU on Ibaji (13.2 on Idah, 8.5 on Ibaji), proving dialect fairness requires mixing. Fine-tuned from Helsinki-NLP/opus-mt-en-mul (78M), epoch ablation shows convergence at 8 epochs and overfitting at 15. BLEU 11.66, chrF 46.69. Analysis shows 899 sentences on log-linear trend consistent with African low-resource literature. This work provides first dialectal impact analysis for Igala and establishes baseline for future work.
[NLP-17] NaijaFacts: A Framework for Disentangling Language from Knowledge in Facts about Nigeria
【速读】： 该论文旨在解决当前多语言非洲语言基准测试中存在的核心缺陷：现有评估主要依赖将通用问题翻译至当地语言，仅衡量模型的语义理解能力，而未能检验其对非洲真实事实的认知水平。为此，研究提出 \textit{aija} 基准，一个聚焦尼日利亚的事实性与统计类问题集合，由母语者编写并基于权威官方统计数据，所有问题均以英语、尼日利亚皮钦语、约鲁巴语、伊博语和豪萨语五种语言呈现。通过构建“国家-语言”矩阵（country-by-language matrix），该方法能够解耦语言效应（模型知晓事实但在低资源语言中表现下降）与知识效应（事实本身在所有语言中均缺失）。实验结果表明，知识效应占据主导地位：在针对美国和英国的问题上，英文查询准确率可达49–50%，而对尼日利亚数据的准确率仅为20%；语言效应则较小且与网络文本代表性相关，从英语到约鲁巴语的性能下降达9个百分点（尼日利亚问题），最高达24个百分点（美国/英国问题）。进一步验证显示，模型规模提升并未缓解此差距——最强模型在美英问题上达到66–79%准确率，但在尼日利亚统计数据任务上均未超过31%。将该分析扩展至全非洲54个国家，发现相同模式普遍存在：文化类问题准确率达96%，政治类达98%，但统计类问题仅26%，说明该知识鸿沟并非局限于尼日利亚，而是非洲整体认知覆盖不足的体现。
链接: https://arxiv.org/abs/2611.06504
作者: Zainab Afolabi,Nikita Kozodoi,Jack Butler,Babs Khalidson
机构: AWS FDE(亚马逊云科技未来数据工程); Amazon(亚马逊)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Multilingual benchmarks for African languages mostly translate generic questions into local languages, measuring linguistic competence but not whether models know facts about Africa. We introduce \naija, a benchmark of factual and statistical questions about Nigeria, authored by native speakers and grounded in authoritative official statistics, with every question available in English, Nigerian Pidgin, \yoruba, Igbo, and Hausa. Answering similar questions for the UK and US yields a country by language matrix that disentangles a language effect (the model knows the fact but degrades in a low-resource language) from a knowledge effect (the fact is absent in every language). Across eight frontier and Nigeria-adapted models, the knowledge effect dominates: English-query accuracy is 49–50% for the US/UK but 20% for Nigeria. The language effect is smaller and tracks web representation, degrading from English to \yoruba (9 points on Nigeria, up to 24 on US/UK questions). We find scale does not help: the strongest models reach 66–79% on the UK and the US but none exceed 31% on Nigerian statistics. Extending to all 54 African countries with our best performing frontier model shows the same profile continent-wide, 96% on culture and 98% on politics against 26% on statistics, so the gap is not specific to Nigeria.
[NLP-18] CRESiST: Self-Enhancing Explorative Policy Learning for Simultaneous Speech Translation NEURIPS2026
【速读】： 该论文旨在解决同步语音翻译（Simultaneous Speech Translation, SimulST）中翻译质量与延迟之间的权衡问题。现有基于大语言模型（Large Language Models, LLMs）的SimulST系统主要依赖前缀增量生成（prefix-incremental generation）范式，存在关键缺陷：重复的键值（Key-Value, KV）缓存重新计算以及对额外语音活动检测（Voice Activity Detection, VAD）的依赖，难以处理无界流式输入。为克服上述瓶颈，本文提出一种基于全模态大语言模型（omni-modal LLM）的端到端交错读写（interleaved read-write）框架，通过交错排列语音片段与文本标记，实现完整的KV缓存复用，原生支持无需分段、无限长度的推理。此外，为在不依赖严格离线对齐的情况下学习最优流式策略，设计了一种新型的演员-评论家自增强（Actor-Critic self-enhancement）训练管道：采用基于RNN-T的演员（Actor）动态采样读写轨迹，由基于LLM的评论家（Critic）通过内在语义与延迟奖励评估路径，并将最优策略回传至演员进行优化。大量实验在句子级（FLEURS、CoVoST2）和文档级（MuST-C、ACL 60/60）基准上验证了所提方法在质量-延迟帕累托前沿上的显著优势，全面超越现有级联式与静态对齐基线。
链接: https://arxiv.org/abs/2611.06501
作者: Chenyang Le,Pei Zhang,Yuxiang Chen,Bing Han,Derek F. Wong,Baosong Yang,Yanmin Qian
机构: Shanghai Jiao Tong University (上海交通大学); Alibaba Token Hub, Alibaba Group (阿里巴巴令牌中心，阿里巴巴集团); University of Macau (澳门大学); University College London (伦敦大学学院)
类目: Computation and Language (cs.CL); Sound (cs.SD)
备注: Accepted to NeurIPS 2026
Abstract:Simultaneous Speech Translation (SimulST) requires balancing translation quality and latency. While Large Language Models (LLMs) excel in offline translation, current LLM-based SimulST systems predominantly rely on prefix-incremental generation. This paradigm suffers from redundant Key-Value cache recomputation and necessitates external Voice Activity Detection to process unbounded streams. To overcome these bottlenecks, we propose an end-to-end interleaved read-write framework built upon an omni-modal LLM. By interleaving speech chunks and text tokens, our architecture achieves full KV cache reuse, natively supporting segmentation-free, infinite-length inference. Furthermore, to learn optimal streaming policies without rigid offline alignments, we introduce a novel Actor-Critic self-enhancement pipeline. An RNN-T Actor dynamically samples read-write trajectories, while an LLM-based Critic evaluates these paths via intrinsic semantic and latency rewards, distilling the optimal policy back to the Actor. Extensive experiments on sentence-level (FLEURS, CoVoST2) and document-level (MuST-C, ACL 60/60) benchmarks demonstrate that our model establishes new state-of-the-art quality-latency Pareto frontiers, substantially outperforming existing cascaded and static-alignment baselines.
[NLP-19] LibriSpeech-Real-QA: A Benchmark for Streaming Question Answering ICASSP2027
【速读】： 该论文旨在解决当前语音大语言模型（Speech Large Language Models, SLMs）普遍采用离线处理模式，无法有效支持实时流式语音交互的问题。在实际应用场景中，如连续计数某人姓名出现次数等任务，要求系统能够基于不断流入的音频片段实时生成响应，而现有模型通常需等待完整语音段落接收完毕才开始推理，难以满足低延迟交互需求。为此，论文提出LibriSpeech-Real-QA数据集与基准评测体系，通过在LibriSpeech语料基础上构建多样化的实时问答对，评估SLMs在流式音频中识别目标词的能力。其解决方案的关键在于设计适用于实时语音交互的建模架构，并成功使预训练语音大模型具备在音频流过程中持续推理的能力，从而建立了一个可复现、可比较的基准，为后续研究提供了重要基础。
链接: https://arxiv.org/abs/2611.06496
作者: Zewen Zhan,Hui Lu,Xueyuan Chen,Huimeng Wang,Yikai Huang,Xixin Wu,Shiyin Kang,Bin Xia,Zhiyong Wu
机构: 未知
类目: Computation and Language (cs.CL)
备注: 5 pages, 1 figure, 6 tables. Submitted to IEEE ICASSP 2027
Abstract:Most speech large language models (SLMs) still process speech in an offline manner. They only start responding after they receive the complete speech segment. However, many real-world applications require streaming speech interaction, for which the audio continuously flows in and the machine is required to answer questions based on the so far received audio chunks. For example, a user may ask the system to continuously count how many times a person’s name has been mentioned while they are speaking. We observe a lack of benchmarks and evaluation protocols for real-time speech interaction. Motivated by this, we propose LibriSpeech-Real-QA, a dataset and a benchmark for real-time speech interaction. We construct different types of real-time question-answer pairs from LibriSpeech to evaluate whether speech LLMs can identify target words in streaming audio. We evaluate several popular real-time commercial models and find that they still tend to respond only after the speech turn has finished. We demonstrate that we can enable a pretrained SLM to perform this task, where we explore several modeling designs and establish a solid baseline for future research.
[NLP-20] What Does Multi-Turn Reinforcement Learning Really Improve in LLM Agents ?
【速读】： 该论文旨在解决多轮代理强化学习（multi-turn agentic reinforcement learning, RL）在长时序任务中对交互预算（interaction budget）高度敏感的问题，即当采样广度 K 与交互深度 T 受限时，其性能显著下降。研究发现，基于轨迹的强化学习优势依赖于充足的轮次预算，且代理强化学习并非普遍提升单步决策质量，其核心优势集中体现在错误恢复能力与长时序任务中的鲁棒性。为缓解由此带来的部署风险，论文探讨了过程监督信号的优劣，并提出一种名为BudgetMix RL的新方法——在训练过程中动态调整交互预算，从而增强模型在不同执行环境下的适应性。实验表明，BudgetMix RL能够在多种执行设置下保持稳健的性能表现。
链接: https://arxiv.org/abs/2611.06483
作者: Kejian Zhu,Zhuoran Jin,Shangqing Tu,Tianyi Men,Yupu Hao,Jiachun Li,Kang Liu,Jun Zhao
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Project Page: this https URL
Abstract:While multi-turn agentic reinforcement learning (RL) excels on long-horizon tasks, its performance gains remain highly sensitive to interaction constraints. Evaluating across sampling breadth K and interaction depth T via Pass@(K,T) , we show that trajectory-level RL advantages depend on sufficient turn budgets, diminishing significantly when budgets tighten. Besides, agentic RL does not uniformly elevate single-step decision quality; rather, its strength concentrates in error recovery and long-horizon resilience. To mitigate the deployment risks caused by this budget dependence, we provide a preliminary discussion on the benefits and limitations of process supervision signals and introduce BudgetMix RL as an alternative that varies interaction budgets during training. Experiments demonstrate that BudgetMix RL achieves robust performance across diverse execution settings. Project Page: this https URL
[NLP-21] EvoGuard: Mitigating Degradation in Online Skill Evolution
【速读】： 该论文旨在解决生成式语言智能体在在线演化技能过程中出现的技能退化问题。现有方法通常在离线状态下基于多任务收集的轨迹批量更新技能，但存在两类缺陷：其一，直接采纳新技能而不进行验证，易导致性能下降；其二，仅在保持验证集性能提升时才保留新技能，虽有效却因频繁运行私有测试集评估而带来高延迟与高成本。针对上述挑战，本文提出EvoGuard，一种低延迟、低成本的成对技能选择器。其核心在于：首先通过迭代运行智能体构建技能池，并在私有测试集上为各技能打分以获得效用标签；随后将技能两两配对，并利用拒绝采样筛选出具有明显效用差异的配对样本；进而训练一个基于两个标签词logits的判别模型，以选择更优技能。部署时，该选择器仅需一次前向传播即可完成决策，显著快于传统测试集评估。实验表明，相较于无筛选地采纳所有新技能，EvoGuard在ALFWorld上提升成功率11.0%，在ScienceWorld上提升5.9%，在LiveCodeBench上提升2.7%。
链接: https://arxiv.org/abs/2611.06444
作者: Yanli Wang,Zichen Bai,Tianyue Jiang,Kunming Shao,Junhao Shi,Xueqin Liu,Jinguo Sun,Muye Li,Tingting Bi,Yanlin Wang,Shaoqiang Zheng,Xuefei Diao,Zhuwei Zeng,Nan Duan,Peng Cao
机构: JD.Com; Sun Yat-sen University (中山大学); The University of Melbourne (墨尔本大学); The Hong Kong University of Science and Technology (香港科技大学)
类目: Computation and Language (cs.CL)
备注: 16 pages, 6 figures, 7 tables, including appendices
Abstract:Skills let language agents reuse experience across tasks as natural-language documents. Recent work evolves skills from interaction, but existing methods have limitations when skills evolve online. Prior methods update a skill offline from a batch of trajectories collected over many tasks. Some then adopt the new skill without any check, which can degrade the skill over time. Others keep it only if it improves performance on a held-out validation set, which is too slow and costly to run after every update. We introduce EvoGuard, a pairwise skill selector that mitigates skill degradation at low latency and cost. We first run the agent iteratively to build pools of skills and score each skill on a private test set to obtain utility labels. We pair the skills and use rejection sampling to keep only pairs with a clear utility gap. The selector is trained on the logits of two label tokens to pick the better skill. At deployment it decides in a single forward pass, which is significantly faster than evaluation on the private test set. Compared with adopting every new skill, EvoGuard improves success by 11.0% on ALFWorld, 5.9% on ScienceWorld, and 2.7% on LiveCodeBench.
[NLP-22] wists and Turns: A Graph-Based Benchmark and Framework for Long-Term Memory in Exploration Tasks
【速读】： 该论文旨在解决长期探索任务（long-term exploration tasks）中模型记忆与推理能力不足的问题，尤其针对科学探究和软件开发等需要持续迭代优化的复杂过程。现有长时记忆研究多聚焦于事实性信息的存储与更新，忽视了探索过程中各尝试方案之间的动态关联性与演化结构。论文提出的关键解决方案是构建一个基于图结构的基准测试框架——LExMem，将探索过程建模为具有分支、修正与重审关系的图网络，并通过进化机制生成具有高结构复杂性的对话与问答数据集（共224条对话，2,018个问答实例）。在此基础上，进一步提出LExGraph，一种基于图的记忆框架，通过增量式组织对话为互联的记忆节点，并结合混合检索与邻域扩展策略实现高效证据召回。实验表明，该方法在面对复杂探索轨迹时显著优于现有系统，具备更强的泛化能力。
链接: https://arxiv.org/abs/2611.06414
作者: Chongyuan Dai,Yaling Shen,Zirui Song,Junfeng Liao,Jinpeng Hu,Xiuying Chen
机构: 未知
类目: Computation and Language (cs.CL)
备注: 29 pages, 5 figures
Abstract:Long-term problem-solving processes, such as scientific research and software development, proceed along an evolving trial-and-error trajectory toward a central goal. We term these \textitexploration tasks, which require models to remember not only facts but also how solutions are tested against intermediate outcomes and early attempts are continually revised or revived. However, existing long-term memory research focuses on factual recall and updates, overlooking the interconnected nature of exploration. We observe that exploration processes naturally form graphs, with branching approaches linked by revisions and reconsiderations. Motivated by this, we propose \textscLExMem, a graph-guided benchmark designed to evaluate \underline\textscMemory in \underline\textscLong-term \underline\textscExploration tasks. We model exploration processes as graphs and introduce an evolutionary mechanism to increase their structural complexity, yielding 224 dialogues and 2,018 question-answer instances derived from graph structures. Building on the same insight, we propose \textscLExGraph, a graph-based memory framework that incrementally organizes dialogues into interconnected memory nodes and retrieves evidence through hybrid retrieval and neighbor expansion. Extensive experiments demonstrate that \textscLExMem poses a significant challenge to existing retrieval- and memory-based systems, whereas \textscLExGraph achieves superior performance and generalizes effectively to external benchmarks. Codes and datasets are available at this https URL.
[NLP-23] Ruby-OPD: Enhancing Pronunciation Controllability and Accuracy for Flow-Matching Text-to-Speech via On-Policy Distillation
【速读】： 该论文旨在解决日语文本到语音（TTS）中高发音准确率与细粒度控制难以兼顾的核心问题。由于日语为非分词语言，其书写系统混合使用表意的汉字（Kanji）与表音的假名（Kana），且存在大量多音字现象，导致发音歧义严重。传统级联式TTS系统虽能通过前端文本转音素模块实现发音控制，但会丢失词汇语义和语调线索；而近年兴起的零样本TTS模型虽能直接处理原始文字以提升自然度，却面临多音字错误频发且缺乏显式发音控制的缺陷。本文提出一种两阶段流匹配TTS框架：关键创新在于引入带正则化扰动的“Ruby”文本表示，以联合保留表面字形与目标发音信息，确保可控性；同时在大规模纯文本语料上采用基于策略的蒸馏（On-Policy Distillation, OPD），无需配对语音即可显著提升原始文字输入下的发音准确性。在Joyo-Kanji-Yomi基准测试中，该方法在保持强可控性的前提下，实现了97.13%的发音准确率、2.02%的目标音素错误率（PER）、0.70%的句子级PER以及4.21%的句子级字符错误率（CER）。
链接: https://arxiv.org/abs/2611.06403
作者: Rui Hu,Xiaolong Lin,Zhenpeng Zhan
机构: 未知
类目: Computation and Language (cs.CL)
备注: 5 pages, 3 tables, work in progress
Abstract:Achieving both high pronunciation accuracy and fine-grained controllability remains a central challenge in Japanese text-to-speech (TTS), as Japanese is an unsegmented language mixing logographic Kanji and phonographic Kana with extensive polyphony. Traditional cascaded systems convert text into phoneme sequences via a front-end, ensuring pronunciation control but discarding lexical semantics and prosodic cues in surface orthography. Conversely, recent zero-shot TTS models directly consume raw orthographic text to improve naturalness, yet suffer from polyphone errors and lack explicit pronunciation control. To bridge this gap, we propose a two-stage framework for flow-matching TTS. First, we investigate Ruby text representations with orthography perturbation to jointly preserve surface forms and target pronunciations for reliable control. Second, we introduce On-Policy Distillation (OPD) on large-scale text-only corpora to enhance pronunciation accuracy on raw orthographic text without paired speech. On the Joyo-Kanji-Yomi benchmark, our model achieves 97.13% Accuracy, 2.02% Target PER, 0.70% Sentence PER, and 4.21% Sentence CER under raw text while retaining strong controllability.
[NLP-24] Interpretable Text Embeddings Based on Hierarchical Questions
【速读】： 该论文旨在解决基于问答的文本嵌入（QA-based text embeddings）在实际应用中面临的两大挑战：一是生成高解释性嵌入通常需要人工审阅大量问题，导致效率低下；二是问题数量与问题粒度之间的权衡关系尚未得到充分研究。为此，论文提出一种构建层次化问题结构的方法，通过递归聚类、对比提示（contrastive prompting）和问题对齐技术，构建从粗粒度到细粒度逐层深化的问题层级。在解释阶段，可通过用浅层对齐问题替代深层具体问题，显著减少人类需审阅的问题数量，从而提升可解释性并保持任务性能。实验结果表明，当可用问题数量有限时，浅层问题即使数量相同也往往比深层问题表现出更高的任务性能，揭示了问题层级深度与性能之间的非线性关系，为高效设计可解释的文本嵌入提供了关键方法论支持。
链接: https://arxiv.org/abs/2611.06402
作者: Kohei Oda,Po-Min Chuang,Kiyoaki Shirai,Natthawut Kertkeidkachorn
机构: Japan Advanced Institute of Science and Technology (日本高级科学技术研究院); Independent Researcher (独立研究员)
类目: Computation and Language (cs.CL)
备注:
Abstract:QA-based text embeddings represent texts as vectors consisting of 1/0 answers to yes/no questions. Although these embeddings provide detailed information for interpretation and achieve high task performance, they often require humans to examine a large number of questions. Furthermore, the relationship between the number of questions and the appropriate level of question granularity has not been sufficiently studied. To address these challenges, we propose a method for constructing hierarchical questions for text embeddings. As the hierarchy deepens, questions become more specific and their number increases. During interpretation, we reduce the number of questions that humans need to examine by replacing questions at deeper levels with aligned questions at shallower levels. The question hierarchy is constructed through recursive clustering, contrastive prompting, and question alignment. Experimental results on general-purpose tasks show that our method improves interpretability while maintaining task performance. It is also found that, when only a limited number of questions are available, questions at shallower levels tend to achieve higher task performance than those at deeper levels, even with the same number of questions.
[NLP-25] Where Should a KV-Cache Repair Budget Go? Distractor Structure and Contextual Compatibility in Non-Prefix Reuse
【速读】： 该论文旨在解决生成式 AI（Generative AI）在推理过程中因非前缀型键值缓存（Non-prefix KV-cache）重用导致的上下文依赖性问题，即在对话记忆、开放域问答和API文档等场景中，由于仅独立编码并拼接检索到的片段缓存，造成部分上下文相关问题无法被正确回答。其核心解决方案是通过上下文兼容性机制（contextual compatibility），优先将有限的计算预算用于重新计算目标片段（target chunk）的键值对，而非均匀分配至所有候选片段。研究表明，在7个不同模型与9个数据集设置下，对目标片段进行重计算的性能显著优于直接插入完整预填充的键值对，尽管其重构结果仍与全预填充键值对存在差距；而基于检索器选择的高置信度片段，在移至问题附近后进行重计算，能在5%预算下超越InfoFlow风格的重排序策略，尤其在开放域问答和API文档任务中表现更优。然而，在对话记忆任务中，该方法未显著优于各场景下选定的词元级选择器，表明其有效性受任务特性的制约。
链接: https://arxiv.org/abs/2611.06397
作者: Yuxing Peng,Zhiqing Tang,Weijia Jia
机构: Beijing Normal University, Zhuhai(北京师范大学珠海校区); Institute of Artificial Intelligence and Future Networks(人工智能与未来网络研究院); Guangdong Key Lab of AI and Multi-Modal Data Processing(广东省人工智能与多模态数据处理重点实验室)
类目: Computation and Language (cs.CL)
备注: 36 pages, 6 figures, 34 tables
Abstract:Non-prefix KV-cache reuse encodes retrieved chunks independently and concatenates their caches. To find where a small repair budget should be spent, we vary the chunks around the answer-bearing one (the target) across three 7B-8B models and four datasets of dialogue memory, open-domain QA and API documentation. Reuse loses 19-100% of the context-dependent questions that full prefill answers, and it loses more when the other chunks are of the same kind as the target. Because the recoverable damage is spread over these other chunks, restoring a few of them recovers little, and the budget is better spent on the target. This works through contextual compatibility: recomputing the target against the reused cache beats inserting its full-prefill KV in all seven settings tested, although the recomputed KV stays far from the full-prefill KV. Question attention over reused caches, however, selects the target in at most 18.8% of development questions. A retriever finds it more often, and recomputing its top chunk after moving it next to the question significantly beats InfoFlow-style reordering at the same 5% budget in six of nine open-domain QA and API-documentation settings. On dialogue memory, though, recomputing chunks never significantly beats the token-level selector chosen for each setting.
[NLP-26] Presentation Sensitivity in Typed Decision Models for Subjective Text Evaluation
【速读】： 该论文旨在解决生成式 AI 评估中因呈现方式（presentation）敏感性导致的评价结果不可靠问题，即相同的评估内容在不同展示形式下可能产生显著差异的判断结果。其核心解决方案在于通过受控审计（controlled audit）系统性地检验多个主流评估模型（Jev、Clef Flash、Perplexity Decider）在不同呈现变量（如选项位置、标识符、作者署名、推荐信号、决策框架及评分标准）下的表现稳定性。研究发现，呈现方式对评估结果具有显著影响：例如，选项位置效应可使摘要评估的置信度提升6.11个百分点；改变读者推荐状态会使Perplexity在幽默性判断中选择不同文本的比例高达40.40%；而负面表述框架可使Clef模型与人类偏好之间的一致性下降23.36分。此外，尽管评分格式会改变得分水平，但通常保持相对排名稳定。因此，该研究强调，一个可靠的评估系统必须具备可重复性、呈现鲁棒性以及与人类判断的一致性，评估流程的有效性依赖于对所用决策机制和分数解释的独立验证。
链接: https://arxiv.org/abs/2611.06384
作者: Jatin Agrawal,Radhika Mamidi
机构: IIIT Hyderabad (印度国际信息技术学院); LTRC (语言技术研究组)
类目: Computation and Language (cs.CL)
备注: 27 pages, 1 figure, includes appendices
Abstract:Typed decision models offer structured choices, scores, and probabilities for low-cost automated evaluation. We present a controlled audit of presentation sensitivity in Jev, Clef Flash, and Perplexity Decider across native choice and score interfaces. Across six datasets with 9,000 multiply annotated humor, helpfulness, and summarization items, we collect 702,000 judgments under interventions that vary positions, identifiers, authorship claims, endorsements, decision framing, and rating rubrics while preserving the evaluated texts. Jev’s second-position probability advantage reaches 6.11 percentage points on summaries; reversing reader endorsements changes Perplexity’s selected humor text on 40.40% of pairs; and rejection framing reduces Clef’s summary agreement with human preferences by 23.36 points. Rating formats also shift score levels while often preserving rankings. Across these findings, repeatability, presentation robustness, and human alignment emerge as distinct properties of an evaluator. Reliable use therefore requires validating the decisions and score interpretations on which an evaluation pipeline depends.
[NLP-27] LegalEvolver: Agent ic Legal Judgment Prediction via Rubric-Guided Self-Evolution
【速读】： 该论文旨在解决大语言模型（LLM）在法律判决预测（Legal Judgment Prediction, LJP）任务中因检索查询质量受限而导致的性能瓶颈问题。现有法律搜索代理通常直接基于用户初始请求生成查询，忽视了潜在的法律争议点和具有法律意义的关键细节，导致检索结果不充分。为应对这一挑战，论文提出了一种名为LegalEvolver的查询优化框架，其核心在于通过启发式方法对初始查询进行优化作为种子，并结合滚动轨迹经验总结形成可迭代更新的评估规范（rubrics），以指导多轮自进化过程。该方案通过将强化学习（RL）与动态演化的规范相结合，使模型能够突破初始能力限制，持续提升查询质量与检索效果。实验表明，该框架在多个基准测试上显著优于直接微调后接强化学习的方法，且动态演化的规范相比固定规范表现更优，验证了自进化机制的有效性。
链接: https://arxiv.org/abs/2611.06381
作者: Xinyu Yang,Jiejun Tan,Chenlong Deng,Chenghao Zhang,Tongyu Wen,Zhicheng Dou
机构: Renmin University of China (中国人民大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 20 pages, 5 figures
Abstract:As LLM-based agents develop, they have been increasingly applied in the legal domain. Legal Judgment Prediction~(LJP) requires models to conduct multi-step reasoning and external legal evidence retrieval. Existing legal search agents’ performance is often constrained by the quality of retrieval queries. They often formulate queries directly from the initial user request, overlooking latent legal issues and legally salient details. To address this issue, we propose \emphLegalEvolver, a query optimization framework specifically designed to generate legally informative queries for LJP-oriented retrieval. Sampling high-quality optimized queries and their corresponding search trajectories is challenging, as relying solely on unguided high-temperature rollouts is both inefficient and constrained by the model’s initial capabilities. We therefore heuristically optimize queries as seeds, improving the efficiency of trajectory generation. Furthermore, we summarize rollout experience into rubrics and use them to guide multi-round iterative self-evolution. Iteratively performing rubric optimization and rubric-guided RL enables the model to surpass its initial performance ceiling. End-to-end performance on extensive benchmarks demonstrates the effectiveness of query optimization on LJP. Comparisons with direct SFT followed by RL demonstrate the effectiveness of rubric guidance. Evolving rubrics further outperform fixed rubrics, confirming the benefits of self-evolution.
[NLP-28] Does Context Compression Preserve Refusal Alignment?
【速读】： 该论文旨在解决生成式人工智能（Generative AI）在采用上下文压缩（context compression）技术以降低推理成本时，模型拒绝行为（refusal behavior）退化的问题。尽管压缩过程在语义保真度上表现良好，但研究发现，基于编码器的压缩会系统性削弱指令微调语言模型的拒绝能力，且该现象在不同模型家族与压缩架构中均持续存在。其关键解决方案在于提出一种轻量级的推理时干预方法——记忆引导（Memory Steering），该方法通过在压缩表示空间内直接操作，恢复被压缩过程弱化的拒绝方向激活，从而将拒绝率提升至接近基线水平，且无需重新训练。研究表明，仅保证语义保真并不足以确保行为一致性，因此在设计压缩感知系统时，必须显式保留对对齐（alignment）至关重要的特征。
链接: https://arxiv.org/abs/2611.06380
作者: Anthony Nguyen,Wenjun Lin
机构: Digital Innovation Lab, Faculty of Computer Science and Technology; Algoma University (安哥拉大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Presented at Canadian AI 2026
Abstract:Context compression reduces inference cost by encoding inputs into compact representations while preserving semantic content. An open question is whether semantic preservation alone is sufficient to maintain downstream behaviours such as refusal alignment. We investigate this question and find that encoder-based compression systematically weakens refusal behaviour in instruction-tuned language models, despite high reconstruction fidelity. This effect persists across model families and compression architectures. Mechanistic analysis shows that compression attenuates activation along the decoder’s learned refusal direction. We further explore Memory Steering, a lightweight inference-time intervention that restores refusal rates to near-baseline levels without retraining and operates entirely in compressed representation space. These results demonstrate that semantic preservation does not guarantee behavioural preservation under compression, highlighting the need to explicitly preserve alignment-relevant features in compression-aware systems.
[NLP-29] raining-Free Character-Length Control in Summarization with Diffusion Language Models EMNLP2026
【速读】： 该论文旨在解决生成式摘要在固定字符长度约束下的精准控制问题，尤其针对标题生成和文章预览等需严格适配指定空间的任务。现有基于自回归语言模型（Autoregressive Language Models, LMs）的方法存在显著局限：基于训练的微调方法需任务特定调整，而零样本提示策略仅能实现近似长度匹配。为此，论文提出一种无需训练的推理阶段方法——DELTA（Diffusion Expected Length Targeting Algorithm），其核心在于利用掩码扩散语言模型（Masked Diffusion LMs）在去噪过程中的可调控性，通过在每一步去噪时估计生成文本的期望字符长度，并动态插入或移除掩码位置，从而精确引导输出至目标长度。该方法不依赖微调，兼容现成的掩码扩散模型，在三个摘要基准上将字符长度误差降低一个数量级，同时在自动评估与人工评估中均保持了优异的摘要质量。
链接: https://arxiv.org/abs/2611.06371
作者: Steffen Castle,Nils Feldhus,Christopher Ebert,Leonhard Hennig,Sebastian Möller
机构: German Research Center for Artificial Intelligence (DFKI); CLCG, University of Groningen; Technische Universität Berlin
类目: Computation and Language (cs.CL)
备注: Accepted at EMNLP 2026 (Main Conference)
Abstract:Summarization with a fixed character length is important for tasks such as headline and article preview generation, where text must fit an exact space. Autoregressive language models (LMs) struggle with precise character-length control: training-based methods require task-specific fine-tuning, while zero-shot prompting strategies achieve only approximate compliance. We propose DELTA (Diffusion Expected Length Targeting Algorithm), a training-free inference-time method for masked diffusion LMs that estimates expected character length at each denoising step and dynamically inserts or removes masked positions to steer toward a target length. DELTA requires no fine-tuning and is designed for off-the-shelf masked diffusion LMs. Across three summarization benchmarks, it reduces character-length errors by an order of magnitude compared to baselines while maintaining competitive summary quality in both automatic and human evaluation.
[NLP-30] Mass Producing Training Data with Surprising Side Effects
【速读】： 该论文旨在解决语言模型在训练数据中存在隐含偏差或过时信息时，可能引发超出训练领域外的不可预测行为（如生成不安全代码、表现出19世纪人物特征等）的问题。这类现象虽已知存在，但因仅有少量人工构造或偶然发现的案例，难以系统研究其泛化规律。为此，论文提出一种自动发现具有意外副作用的数据集（DSSEs, Datasets with Surprising Side Effects）的方法。其解决方案的关键在于：定义并量化三重目标——（1）诱发特定目标行为；（2）外观上类似常规的后训练数据；（3）对人类审计者而言具有出人意料性，并通过大语言模型（LLM）作为评判代理来实现这些目标的可操作化。研究采用演化算法、强化学习与潜意识学习等多种搜索策略，在大规模空间中高效探索高分候选数据集。经保留指标重新评分与过滤后，最终生成涵盖8类行为的6,575个高质量DSSEs，验证了其在不同模型家族和规模间的泛化能力。实验表明，自动化评估指标能有效反映机器学习研究人员的判断，且前沿模型在预测副作用方面优于人类。该成果构建了一个可用于研究模型泛化的基准测试集，为在模型部署前预测潜在训练副作用提供了重要工具。
链接: https://arxiv.org/abs/2611.06324
作者: Yaowen Ye,Jacob Steinhardt,Sarah Schwettmann,Cassidy Laidlaw
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Training data can have unexpected side effects on language model behavior outside the training domain: insecure code data can induce broad misalignment, and outdated bird names can induce a 19th-century persona. Understanding such generalization is hard because only dozens of isolated examples exist, each constructed by hand or discovered by accident. We propose automatically discovering such datasets with surprising side effects (DSSEs). We define three objectives for DSSEs: they should (1) elicit a target behavior, (2) read like ordinary post-training data, and (3) be surprising to human auditors reading the data. We operationalize each with LLM judges, and search for high-scoring datasets via evolution, reinforcement learning, and subliminal learning approaches. After re-scoring and filtering with held-out metrics, our pipeline produces a collection of 6,575 DSSEs across 8 behaviors, many of which generalize across model families and sizes. We find our automated metrics track ML researchers’ judgments, and frontier models outperform humans at predicting side effects. These datasets form a testbed for studying generalization and lay the groundwork for predicting training side effects before model deployment. Explore them at this https URL.
[NLP-31] Zuse: A Decision Model for European Languages
【速读】： 该论文旨在解决多语言决策模型在非英语语境下的泛化能力与校准性问题，尤其关注基于冻结预训练基础模型（frozen open-weight bases）的低秩适配器（low-rank adapter）与指针头（pointer head）架构在27种欧洲语言上的跨语言表现。其核心挑战在于如何在不微调主干模型的前提下，实现对多种语言和任务的高效、准确且可校准的决策推理。解决方案的关键在于：采用三款不同规模（4B、9B、27B）的决策模型，通过在冻结的基础模型上添加轻量级低秩适配器与指针头结构，实现对多语言输入的精准分类，同时利用大规模多源、多任务、多语言评估套件（涵盖9个任务类别与27种语言）进行系统性验证。实验表明，27B模型在测试集上达到87.5%的准确率，预期校准误差（expected calibration error, ECE）仅为0.016（均值），且在不同语言间表现稳定；进一步分析揭示，模型校准偏差主要由任务类型决定而非语言本身，且自动化潜力与经过机会校正的技能（chance-corrected skill）高度相关，优于原始准确率指标。此外，研究发现模型在训练语言之外的泛化能力受限于基础模型的覆盖范围，超出基础模型支持语言范围的语言性能显著下降（约23–28分差距），凸显基础模型语料覆盖的重要性。该工作还识别出构建多源训练数据中的两大陷阱：主题标签词冲突及注入增强引入的捷径学习（shortcut learning）。
链接: https://arxiv.org/abs/2611.06238
作者: Umair Mehmood Imam
机构: Bineric AI(贝纳里克人工智能); OsloMet, Norway(奥斯陆城市大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: 9 pages, 3 figures, 7 tables
Abstract:Decision models read a document and return a probability for each of a fixed set of options instead of generating text. Several appeared in 2026, but none has published results outside English. We describe Zuse, three decision models (4B, 9B and 27B) that train a low-rank adapter and a pointer head on frozen open-weight bases, targeting 27 European languages, and use them to study how this design behaves across languages and tasks. On a broad European suite of 11 sources excluded from training, covering 9 task groups and 27 languages, a development set that also guided model selection, the 27B model reaches 87.5% accuracy with a pooled expected calibration error of 0.016, and 0.035 when weighted per source. Post-hoc analyses of these single-seed results give three findings. First, miscalibration is structured by task rather than language: on parallel reading and topic items in 26 languages the confidence gap varies about as much as sampling noise, while across sources it spans -0.07 to +0.09, with knowledge tasks overconfident and acceptability judgements underconfident at all three sizes, so one fitted temperature cannot correct both. Second, the share of decisions that can be automated at a 5% error budget follows chance-corrected skill (Spearman 0.97 over 33 source and model pairs) more closely than raw accuracy (0.91). Third, reach beyond the training languages is set by the base model: across 210 languages, changing the training run moved scores by at most about 2 points within a size, while languages outside the base’s documented list trail by 23 to 28 points at every size. On English public suites Zuse is behind Jev and Kev, including Kev models on the same bases. We also report two pitfalls in building multi-source training data: label-word collision between topic schemes, and an injection augmentation that taught a shortcut. Zuse can be tried at this https URL.
[NLP-32] Loop Before You Leap: What Recurrent Depth Can and Cannot Buy in Data-Limited Pretraining
【速读】： 该论文旨在解决多模态模型在BabyLM挑战中表现低于纯文本基线的问题，核心关注点在于探索递归计算（recurrence）是否能够缓解这一性能差距。其解决方案的关键在于采用循环式Transformer架构，在不增加参数量或训练数据的前提下，通过重复使用相同层数的网络层以提升计算效率与表达能力。实验在1亿词预算下对比标准Transformer与循环堆叠的12层结构（重复4次），结果表明，在纯文本任务中，循环架构比48层标准Transformer在BLiMP指标上高出5.4个百分点，验证了递归计算的有效性。然而，在多模态设置下，尽管引入视觉前缀向量（来自冻结的DINOv2编码器）可带来平均1.85个百分点的准确率提升，但整体仍比文本基线低约3.8个百分点，且递归机制未能有效缓解这一下降。进一步的对照实验显示，仅使用多模态语言数据而无图像输入时，模型性能反而比纯文本基线低5.7分，说明语言数据分布的变化及其呈现方式是导致性能下降的主要因素。诊断测试表明模型能利用视觉信息，但其对语法泛化能力的贡献尚不明确，尤其在单向量视觉接口限制下。因此，该研究揭示了多模态训练中语言数据动态变化和模态融合界面设计对模型性能的关键影响。
链接: https://arxiv.org/abs/2611.06206
作者: Daichi Hotta,Yohei Kobashi,Takeshi Kojima,Yusuke Iwasawa,Yutaka Matsuo
机构: Aoyama Gakuin University (青山学院大学); The University of Tokyo (东京大学)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注: Accepted to the BabyLM Challenge 2026 (Strict track). 20 pages, 5 figures, 12 tables. This version no longer reports a training-time shuffled-feature control that appeared in the workshop version, because its execution could not be verified (see Limitations)
Abstract:Multimodal models have often scored below text-only baselines in the BabyLM Challenge (Klerings et al., 2024; Takmaz et al., 2025). We test whether additional recurrent computation helps close this gap. Recurrence increases computation by reusing layers without adding parameters or training data. Using a single implementation, we compare standard and looped Transformers under a 100M-word budget, with three seeds per main condition. Each image is represented by one feature vector from a frozen DINOv2 encoder, used as a prefix token. Caption words count toward the budget, so multimodal training also changes the language data. In a pre-registered text-only comparison, a 12-layer stack repeated four times exceeds a standard 48-layer Transformer by 5.4 BLiMP percentage points at matched inference compute. However, multimodal training lowers BLiMP accuracy by about 3.8 points in both architectures, with no clear mitigation from recurrence. An exploratory loop-side no-image control, trained with the multimodal language data and batching schedule, scores 5.7 points below the text-only baseline. Adding the visual prefix raises accuracy by 1.85 points on average, but this contrast is uncertain with three seeds. These results suggest that changes in language data and its presentation are a major source of the drop. Diagnostic tests show that the models use visual information; its contribution to grammatical generalization remains uncertain under this single-vector interface.
[NLP-33] Editing Large Language Models with Geometry-Aware Regularization
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）知识编辑中忽视知识表示空间几何结构的问题。现有方法无法区分预训练知识的优先级（核心知识与非核心知识），也无法判断新编辑任务的难度（难编辑与易编辑），导致在知识更新过程中难以平衡编辑成功率与已有知识的保留。其解决方案的关键在于提出一种基于低秩加对角（Low-Rank Plus Diagonal, LR+D）因子分解模型的几何感知框架——GEdit。该框架通过两个互补机制实现：一是利用低秩正则化对低秩核心知识子空间实施定向保护，确保关键知识不被破坏；二是引入自适应正则化，根据每项编辑请求的可编辑性动态调整正则化强度——对困难编辑放宽约束以保障成功，对简单编辑收紧约束以增强知识保留能力。实验结果表明，在GPT2-XL、GPT-J和LLaMA-3-8B等多个模型上，面对高达10,000次连续更新的大规模编辑场景（包括小批量与全序列编辑设置），GEdit显著优于当前最优基线，在编辑成功率和知识保留性能方面均表现出一致且显著的优势。
链接: https://arxiv.org/abs/2611.06185
作者: Bingqing Liu,Wei Liu,Jun Wang,Xiaobo Sun,Yuhua Li,Ruixuan Li
机构: Huazhong University of Science and Technology (华中科技大学); iWudao Tech; Emory University (埃默里大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 10 pages for main contents
Abstract:Knowledge editing has become a key technique for updating knowledge in large language models. However, existing methods neglect the geometric structure of the knowledge representation space, leading to two key limitations: they cannot distinguish which pre-trained knowledge should be preferentially preserved (core vs. non-core) nor which new edits are inherently difficult (hard vs. easy). To address this, we propose GEdit, a geometry-aware framework built on a Low-Rank Plus Diagonal (LR+D) factor model that captures the intrinsic low-rank structure of Transformer FFN hidden representations. GEdit features two complementary mechanisms. First, the low-rank regularization applies targeted protection on a low-rank core knowledge subspace. Second, the adaptive regularization dynamically adjusts the regularization strength based on each edit request’s editability: relaxing constraints for hard edits to ensure success, and tightening them for easy edits to better preserve pretrained knowledge. Extensive experiments on GPT2-XL, GPT-J, and LLaMA-3-8B under large-scale editing scenarios, including both mini-batch and fully sequential editing settings with up to 10,000 continuous updates, show that GEdit consistently and significantly outperforms state-of-the-art baselines in both editing success and knowledge preservation.
[NLP-34] AssameseLM-E2B: An Open-Weight Assamese Language Model for Efficient Consumer-Grade Deployment
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在低资源语言（如阿萨姆语）中性能下降的问题，主要归因于分词效率低下、生成不稳定性以及语言间干扰等挑战。其核心解决方案是提出一个专为阿萨姆语优化的20亿参数模型——AssameseLM-E2B及其指令微调版本AssameseLM-E2B-it，基于Gemma-4架构设计，并引入扩展的字节级子词分词（byte-level BPE）词汇表，使分词开销相比多语言2B–3B基线模型降低近500%。通过采用参数高效且可持续的预训练（Continual Pre-Training, CPT）流程结合混合LoRA（Hybrid LoRA），并在此基础上利用平衡的多领域数据集进行监督微调（Supervised Fine-Tuning, SFT），显著提升了模型对指令的理解能力与检索增强生成（Retrieval-Augmented Generation, RAG）性能。实验结果显示，该模型在FLORES-200上的比特每字节（Bits-per-Byte, BPB）得分为0.7488，整体RAG得分达96.7%，在对抗性不可回答提示下拒绝准确率达到100%。对比分析表明，通用基线模型存在接收-产出解耦、对罗马化阿萨姆语生成退化以及无意中同化孟加拉语书写系统等严重缺陷。研究证明，基于参数高效的适配策略可在计算资源受限条件下有效构建具备本地部署能力的高性能低资源语言模型。
链接: https://arxiv.org/abs/2611.06170
作者: Pragyan Phukan
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:While Large Language Models (LLMs) excel in high-resource languages, their performance often degrades in low-resource languages due to tokenization inefficiencies, generative instability, and linguistic interference. This paper presents AssameseLM-E2B and its instruction-tuned variant, AssameseLM-E2B-it, a 2-billion-parameter model adapted specifically for low resource Assamese language. Built upon the Gemma-4 architecture, the model incorporates an expanded byte-level BPE vocabulary to reduce tokenization overhead by nearly 500% compared with multilingual 2B–3B baselines. We employ a parameter-efficient Continual Pre-Training (CPT) pipeline with Hybrid LoRA, followed by Supervised Fine-Tuning (SFT) on a balanced, multi-domain dataset to enhance instruction-following and Retrieval-Augmented Generation (RAG) capabilities. Evaluation demonstrates a Bits-per-Byte (BPB) score of 0.7488 on FLORES-200, alongside an overall RAG score of 96.7% and 100% refusal accuracy on adversarial unanswerable prompts. Comparative analysis reveals significant limitations in general-purpose baselines, including receptive-productive decoupling, generative degradation on Romanized Assamese, and unintended Bengali script assimilation. These results demonstrate the potential of parameter-efficient adaptation for developing capable, locally deployable language models for low-resource languages under constrained computational resources. Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2611.06170 [cs.CL] (or arXiv:2611.06170v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2611.06170 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-35] Accelerating Long-Context Decoding in Large-Scale Hybrid LLM s Without Retraining
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在长上下文任务中因全注意力（full-attention）层导致生成效率下降的问题，即每个生成标记的计算成本随上下文长度增加而显著上升。现有混合模型通过将大部分全注意力层替换为线性注意力层以降低计算开销，但剩余的全注意力层仍成为长上下文解码的主要瓶颈。本文提出的关键解决方案是采用无需训练或权重修改的组合稀疏注意力（combined sparse attention）机制：在每一层中同时关注三个部分——一个滑动窗口内的近期标记、一个固定前缀（pinned prefix）以及上下文中最具相关性的固定预算标记。通过基于多目标优化算法对每层的窗口大小和预算进行独立搜索，实现性能与效率的最佳平衡。实验表明，该方法在六个来自三个不同架构家族、参数规模从40亿到3970亿的混合模型上，均保持了原始模型在21个长上下文数据集上的准确率，同时在Qwen3.5-4B、Qwen3.5-122B-A10B和Ornith-1.0-397B上分别实现了64k/256k token解码速度提升最高达1.26倍/2.22倍、1.20倍/2.23倍和1.16倍/2.06倍。
链接: https://arxiv.org/abs/2611.06133
作者: João Eduardo Batista,Emil Vatai,Mohamed Wahib
机构: 未知
类目: Computation and Language (cs.CL); Performance (cs.PF)
备注: 19 pages (main body: 10 pages, references: 3 pages, appendix: 6 pages), 8 figures, 6 tables
Abstract:The full-attention layers of large language models (LLMs) make every generated token more expensive than the previous one, slowing models down on long-context tasks. Hybrid models reduce this cost by replacing most full-attention layers with linear-attention layers, but the first still dominate decoding at long contexts. We accelerate these remaining layers without any training or weight changes using combined sparse attention: each layer attends to a sliding window of recent tokens, a pinned prefix, and a fixed budget of most relevant tokens in the context. The window size and budget are found on a per-layer basis using a multi-objective search algorithm. Across six hybrid models from three families, from 4B to 397B parameters, the selected configurations preserve their baseline model’s accuracy on 21 long-context datasets. Decoding 64k/256k tokens becomes up to 1.26x/2.22x faster on Qwen3.5-4B, 1.20x/2.23x on Qwen3.5-122B-A10B, and 1.16x/2.06x on Ornith-1.0-397B.
[NLP-36] he Harness Is the Model: Agent Scaffolding Changes a Fixed LLM s Compliance with Banking Conduct Rules Unevenly Across Models
【速读】： 该论文旨在解决在银行监管场景下，生成式AI（Generative AI）模型在实际部署中如何确保合规性的问题。尽管模型本身保持不变，但其运行环境——即“约束框架”（harness）——对合规表现具有显著影响。研究的关键在于验证不同约束框架设计对模型遵守银行行为规则能力的影响，发现同一模型在不同框架下的表现差异可达11.5个百分点，表明模型与框架的交互作用至关重要。尤其值得注意的是，采用同模型合规审查机制的框架可将违规率从9.6%降至6.4%，且未增加过度拒绝情况，与预注册假设相反。此外，涉及日期计算的规则在所有框架中均失效，提示系统性缺陷。研究强调，模型风险管控的最小验证单元应为“模型-框架”组合，合规性评估必须基于执行动作日志而非对话记录，以避免误判。
链接: https://arxiv.org/abs/2611.06116
作者: Mehmet Murat Huyut
机构: Kuveyt Türk Participation Bank (库韦特土耳其参与银行); İstanbul Bilgi University (伊斯坦布尔比尔吉大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 16 pages, 4 tables. Pre-registered. Code, data and all 3120 episode traces: this https URL
Abstract:Banks validate language models, but they deploy systems: a model inside a harness that decides what it sees, which tools it can call and who checks its actions. We test whether that harness changes compliance with banking conduct rules when the model is held fixed. Four open-weight models each ran five harnesses (a pipeline that makes one structured decision, a ReAct-style tool loop, a loop in which the policy is reachable only through search, a loop with a long benign customer memory, and a loop with a same-model compliance reviewer) on 156 pre-registered cases in a simulated retail bank whose 11 rule families paraphrase EU consumer-credit, payments, investor-protection, data-protection and anti-money-laundering provisions. Every one of the 3120 two-turn episodes was scored deterministically from the executed tool trace, with violations and over-refusals counted separately. Pooled across models the harness moved the correct-case rate only between 87.3% and 92.1%, but within a model it mattered a great deal: for one model the spread between its best and worst harness was 11.5 percentage points, enough to move it from first to last place among the four models, while for another the harness made no detectable difference; the model x harness interaction itself was suggestive rather than conclusive. The same-model reviewer cut violations from 9.6% to 6.4% without raising over-refusals, contrary to our pre-registered prediction. Rules that require date arithmetic failed in every harness. In an exploratory analysis with two LLM judges and a blind human check, 26% of violations were accompanied by a customer message that denied the action the agent had just executed, which a transcript-only review would record as a correct refusal. For model-risk management the unit of validation should be the model-harness pair, and conduct testing should read the action log, not the conversation.
[NLP-37] Reconciling Privacy with Provenance for Language Model Watermarks
【速读】： 该论文旨在解决生成式 AI（Generative AI）中水印技术在实际部署时面临的隐私与可追溯性之间的矛盾问题。传统水印方法试图追踪个体用户的使用行为，引发用户对隐私泄露的担忧，进而促使用户采用重写（paraphrasing）等规避策略，削弱水印信号的有效性。为缓解这一困境，论文提出将水印设计的目标从个体级溯源转向群体级使用频率估计，从而在保护用户隐私的同时维持水印的实用性。其核心解决方案是构建局部差分隐私（Locally Differentially Private, LDP）水印机制，通过理论保证任何单个文本无法提供显著证据用于检测，有效限制个体暴露风险；同时引入红绿序列（Red-Green Sequence, RGS）水印方案，该方案兼具计算高效性与高隐私水平，在多种模型、样本量及生成熵条件下均能实现精准的群体使用频率估计。实验结果表明，在C4语言建模任务中，当隐私预算ε=1时，均方误差（MSE）低至0.0019，平均绝对偏差仅为0.0038；在更具挑战性的100样本作文生成任务中，ε=2时仍保持约0.0113 MSE和0.004平均绝对偏差，验证了该方法在高隐私强度下依然具备优异的统计估计性能。
链接: https://arxiv.org/abs/2611.06100
作者: Elise L. Han,Tatsunori Hashimoto
机构: Stanford University (斯坦福大学)
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Watermarking has great potential for regulating the use of language models (LMs), but its deployment has been marred by controversy, as many users argue that fine-grained tracking of LM usage violates their privacy. This tension reduces the effectiveness of watermarks, as users engage in evasion strategies such as paraphrasing that can remove much of the watermark signal. We show that provenance and privacy can be reconciled if we shift the goal of watermark design from tracking individuals to estimating aggregate usage. We develop locally differentially private watermarks which provably limit the evidence that any individual text can contribute to detection, while still enabling accurate estimation of population frequencies. We introduce the Red-Green Sequence (RGS) watermark, a private and computationally efficient watermark that achieves accurate frequency estimation at high privacy levels across various models, sample sizes and generation entropies in settings that reflect realistic deployment scenarios. On C4 language modeling tasks, we attain 0.0019 MSE with 0.0038 mean absolute bias at a low \varepsilon of 1, while in more challenging settings such as an essay writing task with only 100 text samples, we still find errors on the order of 0.0113 MSE with 0.004 mean absolute bias at \varepsilon=2 .
[NLP-38] Overview of the NTCIR-19 SciClaimEval Task: Cross-modal Claim Verification in Scientific Papers
【速读】： 该论文旨在解决科学主张（scientific claim）在真实研究论文中基于图表证据进行验证的可信度评估问题，尤其关注如何判断某一科学主张是否被其所在论文中的图表数据所支持。其核心挑战在于区分支持性证据与经过微小修改后不再支持主张的对比性证据，从而实现对科学主张真实性与数据支撑关系的精准判定。解决方案的关键在于构建一个包含180篇机器学习、自然语言处理及医学领域论文的基准数据集，其中每条真实主张均配对了支持性与非支持性（仅经最小改动）的图表版本，并设计了两个子任务：一是判断给定证据是否支持主张，二是从两组对比证据中选择更支持主张的一方。实验表明，基于多模态大语言模型（multimodal large language models, LLMs）并采用提示工程（prompting）与多阶段推理的系统表现优异，在第一个子任务上最佳模型达到98.1%的成对准确率；然而，所有顶尖系统均依赖于对比样本，而独立评估单个主张-证据对的系统性能显著下降，表明在实际应用场景下，仅凭单一证据验证主张仍是一个未解决的开放性问题。此外，开源模型通过精心设计的流水线架构可媲美闭源模型，凸显了系统设计而非模型封闭性的重要性。
链接: https://arxiv.org/abs/2611.06090
作者: Xanh Ho,André Greiner-Petter,Florian Boudin,Yun-Ang Wu,Sunisth Kumar,Tian Cheng Xia,Hiba Akbi,Tim Schopf,Akiko Aizawa
机构: NII LLMC, Japan(日本国立情报学研究所语言与多模态计算中心); University of Göttingen, Germany(德国哥廷根大学); Inria, LS2N, Nantes Université, France(法国国家信息与自动化研究所, 南特大学); The University of Tokyo, Japan(日本东京大学); University of Bologna, Italy(意大利博洛尼亚大学); Sorbonne University, France(法国索邦大学); NII, Japan(日本国立情报学研究所); Akiko Aizawa(日本国立情报学研究所)
类目: Computation and Language (cs.CL)
备注: The first two authors contributed equally. More information about the shared task is available at this https URL
Abstract:This paper presents an overview of \emphSciClaimEval, a pilot task at NTCIR-19 on verifying scientific claims against the tables and figures of the papers in which they appear. Our dataset pairs authentic claims from 180 papers in machine learning, natural language processing, and medicine with contrastive versions of their tables and figures, one supporting the claim and one minimally modified so that it does not. The task comprises two subtasks: deciding whether a piece of evidence supports a claim, and selecting which of two evidence versions does. Twenty-two teams registered, the highest number among NTCIR-19 tasks, and 14 submitted 177 runs. Most systems built on multimodal large language models, relying on prompting and multi-stage inference rather than task-specific fine-tuning. On the first subtask, the best systems reached 98.1% pair accuracy, but all six top-ranked systems relied on the contrastive pairs; the best system that judged each claim-evidence pair on its own scored 24.3 points lower. Open-weight models, embedded in carefully designed pipelines, matched proprietary ones. Yet verifying a single claim against its evidence, the setting that matters in practice, remains an open problem.
[NLP-39] ProcessPrefBench: Evaluating Cross-Task Transfer of Individual Process Preferences from Naturalistic Human-AI Interaction EMNLP2026
【速读】： 该论文旨在解决在自然人机对话中，如何有效识别并持续遵循用户隐含的个性化工作流程偏好（process preferences）这一关键问题。用户在对话中的修正、请求与自我描述蕴含了可复用的信号，反映了其对证据标准、分析深度、执行步骤及交付方式的特定要求。然而，当后续任务未明确重述这些偏好时，模型往往难以维持一致性。解决方案的核心在于通过构建一个名为ProcessPrefBench的基准测试，基于两位知识工作者历时三个月的真实对话数据，设计66个源-目标案例，分别评估源证据的有效性与目标适用性，并采用三种提示策略：仅目标提示（target-only）、完整历史提示（full-history）以及提炼出的偏好指令提示（distilled-preference prompts）。实验结果表明，相较于仅目标提示，使用完整历史或提炼偏好指令可使偏好实现度分别提升0.204和0.244分（0–2量表），但提炼过程并未带来显著优势；同时，尽管偏好遵循性提高，虚构证据（fabricated-evidence）错误率也从5.8%上升至10.1%（完整历史）和10.4%（强制指令），揭示了从对话中学习过程中存在的矛盾：利用过往交互虽能增强个性化响应，却可能牺牲事实可靠性。因此，该研究的关键突破在于揭示了在保持用户偏好一致性与维护生成内容真实性之间的张力，并为构建既能尊重用户流程偏好又不失事实准确性的智能助手提供了重要方向。
链接: https://arxiv.org/abs/2611.06065
作者: Yiwen Yang,Guanming Liu
机构: Fudan University (复旦大学)
类目: Computation and Language (cs.CL)
备注: 11 pages, 2 figures. Published at LM Playschool 2026 (EMNLP 2026 workshop)
Abstract:In natural human-AI dialogue, users’ corrections, requests, and self-descriptions carry reusable signals about how they want work to be done. These signals reveal individual process preferences, user-specific standards for evidence, depth, analytical focus, procedure, and delivery. We ask whether a preference revealed in one dialogue can shape an assistant’s later answer when the task does not restate it. ProcessPrefBench uses three months of natural dialogues from two knowledge workers to build 66 source-target cases separately judged for source evidence and target applicability. Sources are preselected, so the benchmark tests application rather than retrieval. We test three models with target-only, full-history, and distilled-preference prompts. Target-specific 0-2 rubrics score preference realization, and a task gate flags severe failures. Compared with target-only prompting, full-history and imperative distilled-preference prompting raise preference realization by 0.204 and 0.244 points on the 0-2 scale, respectively, with no reliable advantage from distillation. Fabricated-evidence failures rise from 5.8% under target-only prompting to 10.1% under full-history and 10.4% under imperative prompting. ProcessPrefBench highlights a tension in learning from dialogue, where drawing on prior interactions can improve preference realization yet also invite fabricated evidence. We hope it helps build assistants that honor a user’s process preferences without sacrificing factual reliability.
[NLP-40] Adapting Qwen 3 to Tajik from 0.6B to 8B: Lower Text Loss Did Not Mean Better Understanding
【速读】： 该论文旨在解决大模型在塔吉克语（Tajik）低资源场景下的语言建模能力与阅读理解能力之间的不匹配问题，即尽管通过小规模低秩适配器（LoRA）微调可显著降低塔吉克语文本的损失（loss），但这种优化并未转化为实际的阅读理解能力提升。其解决方案的关键在于采用“重配方”（heavy recipe）——即在80亿参数模型上进行大规模原始塔吉克语文本的持续预训练（约3.75亿词元），随后辅以监督微调（supervised fine-tuning），从而实现塔吉克语阅读理解性能的显著提升（在Belebele基准上提升9.9分，最终主模型e3相较Qwen3-8B提升11.4分）。这一策略突破了仅依赖轻量级适配器的局限性，证明了充分的语言数据积累与系统性训练流程对低资源语言下游任务性能的关键作用。然而，研究也揭示了诸多实践问题：训练记录丢失、推理数据集缺乏有效推理样本、部分分支训练数据未被记录，以及一个4位量化版本不含任何Transformer层，凸显了复杂模型开发中工程可靠性和可复现性的挑战。
链接: https://arxiv.org/abs/2611.06062
作者: Tohir Saidzoda,Madina Abdusalomzoda,Nasimjon Rajabov,Akhrorkhon Ikromov,Komron Giyosiev,Abubakr Abdullozoda,Amin Jalilov,Jonibek Kishvarov,Rahmonali Sharipov,Mariyamkhon Shodyeva
机构: Saidzoda AI Lab(萨伊佐达人工智能实验室); Saidzoda Engineering Company(萨伊佐达工程公司), Dushanbe(杜尚别), Tajikistan(塔吉克斯坦)
类目: Computation and Language (cs.CL)
备注: 8 pages, 3 tables, 1 figure. Technical report
Abstract:Small LoRA adapters trained on Tajik news and books lowered Qwen3’s Tajik text loss at every size we tried, from 0.6B to 8B, by 0.60 to 1.45 nats per token, but none of them improved Tajik reading comprehension where it can be measured. On Belebele the 1.7B and 8B adapters left accuracy unchanged and the 4B adapter lowered it by 3.7 points; at 0.6B the score rose by 4.2 points but sits near chance and coincides with a change in which answer letter the model prefers, so we do not read it as understanding. At 8B, continued pretraining on about 375 million tokens of raw Tajik text was followed by a Belebele gain of 9.9 points. After a further supervised fine-tuning stage, our main model, e3, is 11.4 points above Qwen3-8B on Belebele; its English MMLU changed by -1.2 points, and a loss larger than 3.3 points is excluded. The same heavy recipe applied to Qwen3-0.6B-Base brought its text loss close to that of the 8B model after books-only pretraining, while its multiple-choice answers stayed near chance and its English MMLU fell by 7.6 points. We report 15 trained models and their 5 starting models from June to October 2025, scored in 2026 on one harness with paired intervals, and we report what went wrong: lost training records, a thinking-mode dataset with almost no reasoning in it, later branches whose training data was never documented, and a 4-bit release that contains no transformer layers.
[NLP-41] SVGFaith: A Two-Layer Meta-Evaluation of Faithfulness Metrics for Text-to-SVG Generation
【速读】： 该论文旨在解决当前文本到可缩放矢量图形（Text-to-SVG）系统评估中，依赖光栅图像指标对光栅化输出进行评价时，难以准确衡量生成结果对提示的细粒度忠实度（fine-grained prompt faithfulness）的问题。其核心解决方案是提出一种两层元评估框架SVGFaith，通过构造对比对或人工判断来确定每一对生成结果中更忠实于提示的选项。其中，SVGFaith-Diagnostic包含1,842个受控的对比对，覆盖八个能力维度，涵盖双向评估场景；SVGFaith-Arena则包含1,327个真实生成结果的人工标注比较。研究发现，现有基于图像-文本嵌入的度量方法（如CLIP）在对象身份识别上表现尚可（最高0.789配对准确率），但在空间关系和属性绑定任务上表现显著下降（最低0.672），远低于理想水平。为此，论文进一步提出可解释的点式度量方法SVGFaith-Score：该方法首先利用领域调优的分解器将提示转化为原子视觉需求的依赖图，再通过视觉问答模型逐项评估每个需求的满足程度，并引入前序需求得分对当前需求得分进行折扣，从而反映依赖关系的影响。该评分体系不仅提供整体得分，还能揭示具体未满足的需求及其因前置需求不达标而被折扣的情况。实验表明，SVGFaith-Score在Arena和Diagnostic双方向均显著优于九种基线方法，包括Qwen3.5-122B-A10B判别器。
链接: https://arxiv.org/abs/2611.06048
作者: Guotao Liang,Haitao Zhou,Biao Liu,Huanxuan Liao,Lingfeng Qiao,Qian-Wen Zhang,Qian Yu,Di Yin,Xing Sun
机构: Tencent Youtu Lab(腾讯优图实验室); Visual Computing Group, Beijing(视觉计算组, 北京)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: 27 pages
Abstract:Text-to-SVG systems are commonly evaluated by applying raster-image metrics to rasterised outputs, yet whether these scores capture fine-grained prompt faithfulness remains unclear. We introduce SVGFaith, a two-layer meta-evaluation framework in which the more faithful candidate of each pair is determined by construction or by human judgment. SVGFaith-Diagnostic contains 1,842 controlled contrastive pairs across eight capability axes, split evenly between two directions: two renderings compared against one prompt, and two descriptions compared against one rendering. SVGFaith-Arena contains 1,327 human-annotated comparisons of real generator outputs. On SVGFaith-Diagnostic, CLIP and three other image–text embedding metrics reach at least 0.789 pairwise accuracy on object identity but at most 0.672 on spatial relations and attribute binding, where chance is 0.5. We further introduce SVGFaith-Score, an interpretable pointwise metric. A domain-tuned decomposer turns each prompt into a dependency graph of atomic visual requirements, a visual question answering model scores whether each requirement is satisfied, and a requirement’s score is discounted when its prerequisites score low. In addition to the overall score, the per-requirement scores show which requirements are judged unsatisfied and which inherit discounts from their prerequisites. SVGFaith-Score significantly outperforms all nine baselines, including a Qwen3.5-122B-A10B judge, on Arena and in both Diagnostic directions. Code and data: this https URL.
[NLP-42] Confucius4-R2T2: A Stable-Prefix Streaming ASR Model for Real Real-Time Transcription
【速读】： 该论文旨在解决流式语音识别（Streaming Speech Recognition）中因早期输出被后续修正而导致的计算浪费或错误动作问题，尤其在翻译系统和语音代理等实时应用场景中尤为关键。现有方法在处理连续输入时，常出现已生成的词元（token）被修改，造成资源浪费或决策失误。为应对这一挑战，本文提出Confucius4-R2T2（简称R2T2），一种无需修改先前输出的稳定前缀（Stable-Prefix）流式自动语音识别模型，可在保持高精度与低延迟的同时实现输出稳定性。其核心解决方案在于引入“最长稳定前缀”（Longest Stable Prefix, LSP）这一新概念，并基于该理论对预训练的Qwen3-ASR-1.7B模型进行后训练（post-training），在冻结音频编码器与投影层的前提下，仅通过学习LSP标签完成架构无变更的流式适配。研究证明，LSP在前缀序下具有单调性、与全音频输出一致且是等于早期完整输出的最大标签集合，从而确保输出的不可回溯性。实验表明，R2T2在中英文数据上实现了接近全上下文模型的识别质量，同时具备极低的生成延迟与稳定的文本输出，且该优势可泛化至未参与后训练的语言，为从全上下文模型向稳定流式模型的迁移提供了切实可行的技术路径。
链接: https://arxiv.org/abs/2611.06043
作者: Daniel Huang,Qi He,Shaoxiong Duan,Yingjie Li,Qiao Cheng,Yitao Duan
机构: NetEase Youdao(网易有道); UC San Diego(加州大学圣地亚哥分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Sound (cs.SD)
备注: 30 pages, 6 figures, 8 tables. Code and model weights available at this https URL and this https URL
Abstract:Streaming speech recognition is increasingly consumed by machines, such as translators and voice agents, that act on each token as it appears. A token that is later revised costs wasted work or a wrong action. In this work, we present Confucius4-R2T2 (short for Real Real-Time Transcription), a stable-prefix streaming ASR model that never modifies its earlier output while maintaining high accuracy and low latency. R2T2 is obtained by post-training Qwen3-ASR-1.7B on 1,500 hours of Chinese and English audio while freezing its pre-trained audio encoder and projector. The adaptation is based on the concept of Longest Stable Prefix (LSP), a new notion that we introduce in this work. We prove that LSP is monotone under the prefix order, consistent with the full-audio output, and maximal among labels that equal a whole earlier output. Learning from LSP labels converts a full-context model into a stable-prefix streaming one without any architectural change. Extensive evaluations show that R2T2 brings streaming recognition close to the base model’s full-context quality while delivering stable text with low emission lag. Its gains extend beyond the post-training languages, demonstrating a practical path from full-context ASR to stable streaming. We open-source the model weights and code at this https URL and this https URL.
[NLP-43] When Every Choice Is Correct: A Causal Study of Numerical Preference in Language Models
【速读】： 该论文旨在解决生成式 AI（Generative AI）在数值选择任务中表现出的隐性偏好问题，即尽管1至30之间的所有整数均为有效答案，模型却倾向于选择特定数值。研究聚焦于通过数值选择这一受控场景，揭示模型内部各层与各标记（token）间偏好信号的动态传播机制。其解决方案的关键在于采用扰动探测（perturbation probing）与梯度归因（gradient attribution）识别出对偏好具有显著影响的神经元，并实施前答案编辑（pre-answer edit），以干预模型生成过程。实验发现，在对Qwen3-4B模型进行干预后，初始输出为1的情况下，通过替换注意力头输出或联合钳制多层感知机（MLP）输出，可恢复原始偏好答案17，但二者对后续数字5与7的最终对数几率（log-odds）影响不同。进一步通过保持后期MLP输出不变，可使该差异从0.119 nat降至0.0025 nat，表明后期计算过程对局部贡献的解释方式具有调节作用。此外，回溯性估计虽能捕捉大部分概率变化，但无法解释特定答案切换；提示测试亦显示，部分提示保留影响但不引发答案改变，仅少数新表述满足“读者/MLP测试”条件，且联合钳制仍可缩小效应差异。这些结果表明，数值选择作为研究持久状态与条件下游处理的模型，揭示了偏好信号在模型内部的非线性传播路径及其动态调制机制，但尚未完整揭示自然数值偏好的起源或完整神经通路。
链接: https://arxiv.org/abs/2611.06025
作者: Hongnian Huang,Hongliang Liu
机构: University of New Mexico(新墨西哥大学); Independent researcher(独立研究员)
类目: Computation and Language (cs.CL)
备注:
Abstract:When a language model chooses a number between 1 and 30, every integer in the range is valid, yet the model can favor particular answers. We use numerical choice to study how preference is processed across layers and tokens. Perturbation probing and gradient attribution nominate neurons for activation interventions. In a selected Qwen3-4B case, a pre-answer edit changes 17 to 15 while both runs first emit 1. After the edit ends, selective cache and attention-output replacements trace part of its continuing influence on the second digit. Replacing one attention head’s output or the other heads’ outputs jointly produces similar immediate readout contributions; both interventions restore 17 but have different final log-odds effects on 7 versus 5. Holding later multilayer perceptron (MLP) outputs at their values from the edited run without either replacement reduces this effect gap from 0.119 to 0.0025 nat. Later computation therefore changes how similar local contributions affect the continuation. Retrospective estimates across eighteen contexts capture much of the probability movement but miss the selected answer switch. Prompt tests also separate influence from answer changes: one held-out prompt retains influence without switching, and only one of twelve fresh wordings qualifies for the reader/MLP test; there, joint MLP clamping again reduces the gap between the two reader effects while the answer remains 17. These results use numerical choice as a controlled case for studying persistent state and conditional downstream processing; they do not establish the origin or a complete circuit of natural numerical preference.
[NLP-44] Who Pays for the Knowledge Models Copy? Model Collapse as Rogers Paradox
【速读】： 该论文旨在解决生成式模型在迭代训练过程中因依赖自身输出而逐渐丢失罕见知识（rare knowledge）的问题，即“模型坍缩”（model collapse）现象。其核心问题在于：当前主流的缓解策略虽有效，但均依赖外部引入新信息作为补充，而这一外部信息供给被默认为固定且独立于系统本身，忽略了供给来源本身的动态性与可持续性。本文提出关键洞见——信息供给本身正是导致模型坍缩的根本原因，并借鉴文化演化理论中的罗杰斯悖论（Rogers’ paradox）进行解释：当一个群体过度依赖复制而非自主学习时，其知识会因缺乏持续更新而退化，最终无法优于从未复制的群体。研究通过三步验证该理论：首先，在仅使用完全合成数据拟合的罗杰斯模型能精确预测从词元到预训练变换器等不同规模语言模型在混合新鲜数据下的罕见知识保留程度；其次，当外部世界变化加速时，依赖自我复制的模型线系需更多新鲜数据以维持性能，符合罗杰斯模型预测，而每代从固定预训练权重重启的模型则因继承了未被复制的知识而具有较强鲁棒性，这正是固定目标假设所无法解释的现象；最后，通过构建闭环机制——训练数据的生产者转而依赖模型生成内容，使“新鲜数据”供给变为内生变量，发现系统在均衡状态下，模型可靠性恰好等于人类自行获取信息的成本，整体知识水平并未提升，任何模型生成文本占比的增加都必须由更多人自主学习来补偿。因此，唯有用户主动核查并修正错误信息，才能打破此悖论；而仅筛选看似合理的内容将加剧同质化，加速模型坍缩，唯有携带新信息的校验机制才可发挥关键学习作用。
链接: https://arxiv.org/abs/2611.06022
作者: Nam H. Le
机构: University of Vermont(佛蒙特大学); Vermont Complex Systems Institute(佛蒙特复杂系统研究所)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL); Populations and Evolution (q-bio.PE)
备注: 13 pages, 5 figures. Code: this https URL
Abstract:Generative models trained on the output of earlier models lose knowledge across generations, rare knowledge first. The remedies that work all bring in information from outside the loop, and the theory of model collapse treats that supply as given: a fixed archive, drawn from a fixed distribution. We argue that the supply is the problem, and that cultural evolution has the theory of it. We read model collapse as Rogers’ paradox: a population that comes to rely on copying ends up no better off than one that never copied, because copies go stale unless someone keeps learning for themselves. We test the identification in three steps. First, Rogers’ model, fitted only to fully synthetic training, predicts to within a few percentage points how much rare knowledge language models, from word trigrams to a pretrained transformer, retain under other mixtures of fresh data. Second, when the world itself changes, a lineage that copies its own past needs more fresh data to keep up the faster the world changes, as Rogers’ model says it must and as analyses that assume a fixed target cannot; a lineage that restarts every generation from fixed pretrained weights is largely insulated, because those weights are knowledge that is never copied. Third, we close the loop: when the people who write a model’s training data can ask the model instead of finding things out, the supply of fresh data becomes endogenous. At equilibrium, the model is exactly as reliable as it is costly for people to find out for themselves, the population is no better off than without the model, and every increase in the share of model-written text must be paid for by more people learning for themselves. Users who check what they are told and relearn when it is wrong break the paradox. Filters that select plausible text are conformist and accelerate collapse; only checks that carry new information act as critical learning.
[NLP-45] Holistic Parallel Supervision for Expressive Speech to Speech Translation ICASSP2027
【速读】： 该论文旨在解决语音到语音翻译（S2ST）中表达性不足的问题，核心挑战在于现有数据集普遍缺乏源语音与目标语音在内容、韵律、时间对齐及说话人身份等方面的联合一致性标注，导致当前系统依赖间接条件化或迁移机制，难以实现高质量的表达性转换。其解决方案的关键在于提出一种创新的训练数据构建策略：结合专业配音（提供内容、韵律与时间上的自然对应关系）和多语言语音转换技术，以降低说话人差异带来的不匹配问题，从而实现对目标语音的全方位平行监督（holistic parallel supervision）。通过在英德、英西语对上的实验验证，该方法显著提升了各评估系统在内容保真度、韵律准确性、时间对齐、说话人相似性及人工评分表达性方面的表现；其中，本文提出的端到端方法DirectS2ST在各项指标上均取得最大提升，首次系统性地探索了全维度平行监督在表达性S2ST中的有效性。
链接: https://arxiv.org/abs/2611.06010
作者: Hao Zhang,Ismail Rasim Ulgen,Bismarck Bamfo Odoom,Rui Liu,Zihan Zhang,Berrak Sisman,Philipp Koehn
机构: 未知
类目: Computation and Language (cs.CL)
备注: 5 pages, 1 figure, 5 tables. Submitted to ICASSP 2027
Abstract:Expressiveness is central to speech-to-speech translation (S2ST), yet existing datasets rarely align source and target speech jointly in content, prosody, timing, and speaker identity. Consequently, current systems often rely on indirect conditioning or transfer mechanisms. To obtain this supervision, we introduce a training-data construction strategy that combines professional dubbing, which provides human-performed correspondence in content, prosody, and timing, with multilingual voice conversion to reduce speaker mismatch. We evaluate this supervision using existing expressive S2ST systems and DirectS2ST, an end-to-end method developed in this work that predicts the complete target-speech representation directly from source speech. Experiments on English-German and English-Spanish show that holistic parallel supervision improves expressive realization across all evaluated systems. DirectS2ST exhibits the largest gains in content, prosody, timing, speaker similarity, and human-rated expressiveness. To our knowledge, this work provides the first systematic investigation of holistic parallel supervision for expressive S2ST.
[NLP-46] Familiar letters not related languages let unseen languages into a shared space of meaning
【速读】： 该论文旨在解决无翻译资源语言（即缺乏平行语料的低资源语言）如何融入多语言语义共享空间的问题。其核心挑战在于：当一个模型在未见过的语言上进行迁移时，究竟是语言间的亲缘关系（如语法或词汇相似性）还是书写符号的熟悉度（如字符形状的可识别性）在决定语言能否被有效建模和对齐中起关键作用。研究发现，字符的熟悉度而非语言亲缘性才是决定未知语言能否进入共享语义空间的关键因素。为此，作者提出GlyphMind——一种无需分词器的视觉变换器模型，能够将文本作为像素直接处理，并支持跨脚本任意句式呈现。该模型在27种书写系统间预训练以对齐翻译表示。实验表明，在仅使用未标注老挝语数据的情况下，尽管老挝语与泰语同属壮侗语系且书写系统相近，但因字符不熟悉而未能成功嵌入；相反，桑塔利语虽无自身脚本参与预训练，却可通过其熟悉的天城文（源自印地语）脚本成功进入共享空间，而其原生脚本则无法实现。这一结果表明，将目标语言转写为模型已熟悉的文字系统，可有效打破无翻译语言的接入壁垒，为构建真正覆盖全球语言的多语言生成式模型提供了可行路径。
链接: https://arxiv.org/abs/2611.06006
作者: Yuliang Liu,Xinyu Wang,Xiang Bai
机构: Huazhong University of Science and Technology (华中科技大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Hundreds of languages have written text but almost no translations, and whether a model learning one from untranslated text builds on a related language or familiar letters is unclear: they usually coincide. Here we show that familiar letters, not a related language, let an unseen language into a shared space of meaning. GlyphMind, a tokenizer-free vision transformer that reads text as pixels and lets any sentence be shown in any script, was pretrained to align translations across 27 writing systems. Adapted on untranslated Lao, a language and script it had never seen, it retrieved English translations seven times as often as the strongest simple baseline. Rewritten in unfamiliar letters, Lao did not enter although its close relative Thai was in pretraining; Santali, which had none, entered in Devanagari, familiar from Hindi, but not in its own script. Transliteration into familiar scripts may thus open multilingual models to languages without translations.
[NLP-47] How Far Can You Steer? Behavior-Capability Frontiers in Contrastive Activation Addition
【速读】： 该论文旨在解决生成式语言模型中激活控制（activation steering）技术在实际应用中的可用性问题，即仅追求目标行为的改变并不足以保证干预方法的实用价值。其核心挑战在于：尽管对比激活添加（Contrastive Activation Addition, CAA）能够有效引导特定行为，但可能伴随非目标能力的不可控退化。为此，论文提出基于“行为-能力边界”（behavior-capability frontier）的评估框架，通过在限定的能力损失预算（1%、3%、5%绝对百分点）下最大化行为增益，筛选出更具实用性的干预策略。关键解决方案在于引入显式的性能约束机制，以平衡行为调控效果与整体模型能力的维持，实证表明在3%能力损失预算下可实现72.7%的非零操作点存活率，显著优于1%预算下的37.5%。研究还揭示，成功干预依赖于模型学习到的行为方向而非随机激活扰动，且表征几何结构虽能适度预测跨行为因果转移，但预测力有限；此外，即使在受保护基准上表现良好，干预仍可能导致未测试能力的退化或在开放生成任务中出现不均衡迁移。综上，该工作将激活控制重新定位为一种受多重因素（模型架构、目标行为、方向选择、层数、评估预算等）制约的精细化干预手段，而非普适的“万能控制旋钮”。
链接: https://arxiv.org/abs/2611.06001
作者: Deepanshu Raj
机构: The University of Chicago(芝加哥大学)
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: 23 pages, 18 figures
Abstract:Activation steering can modify language-model behavior without parameter updates, but behavioral efficacy alone does not determine whether an intervention is usable. Contrastive Activation Addition (CAA) steers targeted behaviors using activation directions derived from contrasting examples, potentially degrading unrelated model capabilities in the process. We therefore study CAA through a behavior-capability frontier, selecting interventions that maximize behavioral gain subject to an explicit bound on capability degradation. We evaluate this framework across four instruction-tuned models, seven behaviors, and capability-loss budgets of 1%, 3%, and 5% (absolute percentage points). Validation-selected behavioral effects generalize strongly to held-out data ( \rho = 0.945 ), while joint behavior-capability survival among nonzero operating points rises from 37.5% at the 1% budget to 72.7% at 3%, with limited additional improvement at 5%. Control experiments indicate that these effects depend on learned behavior directions rather than generic activation perturbations. Representational geometry is significantly but only modestly predictive of causal cross-behavior transfer. Finally, secondary capability, out-of-distribution, and open-ended evaluations show that interventions satisfying protected benchmarks can still degrade untested capabilities or transfer unevenly to free-form generation. These results characterize activation steering as a constrained intervention whose usable range depends on model, behavior, direction, layer, and evaluation budget, rather than a universal control knob.
[NLP-48] On the Design of Hyper-Connections: Simplification Stability and Scalability
【速读】： 该论文旨在解决超连接（Hyper-Connections, HC）架构在前沿语言模型中虽被广泛采用，但其组件机制尚不清晰且带来显著训练开销的问题。核心挑战在于如何在保持高性能的同时降低复杂性与计算成本。其解决方案的关键在于通过系统性分析识别出两类高效的信息交互模式：一是无混合的可学习读/写门控机制，二是固定门控下的无约束混合机制，二者单独使用即可达到与多路超连接（mHC）相当的性能。研究进一步发现，即使在无混合的情况下，对扩展残差状态的重复访问仍是训练开销的主要来源。基于此，提出SimpleHC架构，通过移除残差混合、利用共享内存访问和加性结构简化执行流程，并引入层特定门控共享组入口状态以支持批量读取与延迟写入，同时采用标量修正精确保留前序子层输出贡献。该设计既保留了有效的残差结构，又优化了内存访问模式，使SimpleHC在最大规模下仅比标准残差基线增加2.4%的端到端训练时间，在从15B到30B MoE模型的扩展过程中保持与mHC相当的性能，展现出更平稳的梯度动态和更强的鲁棒性，推动了基础模型设计向简洁性与实用性方向发展。
链接: https://arxiv.org/abs/2611.05984
作者: Xiangdong Zhang,Xiaohan Qin,Sunan Zou,Tuo Dai,Bizhao Shi,Zhuo Xia,Yebin Yang,Yizhou Chen,Yuliang Liu,Yu Cheng,Junchi Yan
机构: 未知
类目: Machine Learning (cs.LG); Computation and Language (cs.CL)
备注: Code: this https URL
Abstract:Hyper-Connections (HC) architectures are increasingly adopted in frontier language models, yet their components remain insufficiently understood and introduce non-negligible training overhead. In this work, we systematically study the HC architecture to understand its components and identify simpler, more efficient designs. We find that learned read/write gating without mixing and unconstrained mixing with fixed gates form two effective interaction groups: either alone achieves performance comparable to mHC. We also identify repeated access to the expanded residual state as a major source of training overhead, even without mixing. These findings lead to SimpleHC, which simplifies the architecture by removing residual mixing and simplifies execution through shared memory access, using the resulting additive structure and shared gate inputs. Layer-specific gates share a group-entry state, enabling batched reads and deferred writes; scalar corrections exactly preserve preceding sublayer output contributions. Our analyses support both the retained residual structure and the design of shared memory access. SimpleHC incurs only 2.4% extra end-to-end training time over a vanilla residual baseline at the largest tested scale. It maintains performance comparable to mHC when scaling from 15B to 30B MoE models, exhibits smoother gradient dynamics during standard training, and remains stable under stress. We hope SimpleHC will encourage a broader pursuit of simplicity in foundation model design, making performance gains more accessible and practical at scale.
[NLP-49] Modality-Aware Expert Pruning for MoE-Based Multimodal Large Language Models NEURIPS2026
【速读】： 该论文旨在解决多模态大语言模型（Multimodal Large Language Models, MLLMs）中混合专家模型（Mixture-of-Experts, MoE）架构下专家参数占据主导内存占用的问题，尤其在仅部分专家被激活的情况下仍导致模型体积过大（如InternVL3.5-30B-A3B和Qwen3-VL-30B-A3B中专家参数占总模型大小的90%以上）。现有专家剪枝方法通常对所有输入令牌统一评估专家重要性，忽略了同一专家在视觉与语言任务中可能承担不同角色这一关键差异。为此，本文提出一种无需训练的模态感知专家剪枝框架MAEP（Modality-Aware Expert Pruning），其核心在于通过单次前向传播实现跨模态的专家重要性分解：引入ΔH输出（Delta-H Output）作为重分配感知度量，量化移除某专家时MoE输出的变化；同时利用低注意力视觉令牌作为视觉侧的剪枝信号，该信号具有数据集稳定性，即专家重要性排序在不同数据集间保持一致；而文本侧重要性则作为保留信号。最终，专家在全模型层面上进行全局排序，实现更精准的剪枝。实验表明，MAEP在多数多模态基准上优于无模态感知的基线方法，且在纯文本任务中仍保持竞争力，可将BF16格式下的专家权重内存占用减少约27 GB（约占总内存的47%），使约58 GB的模型完整权重可部署于单块40 GB GPU上。
链接: https://arxiv.org/abs/2611.05927
作者: Donggeun Kim
机构: Nota Inc.(Nota公司)
类目: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
备注: NeurIPS 2026
Abstract:Mixture-of-Experts (MoE) enables multimodal large language models (MLLMs) to scale capacity efficiently, but expert parameters dominate memory even though only a few experts are activated per token: they account for over 90% of model size in modern MoE-MLLMs (e.g., 54 GB out of about 58 GB in BF16 for InternVL3.5-30B-A3B and Qwen3-VL-30B-A3B). Expert pruning has emerged as an effective compression strategy, but existing methods aggregate importance uniformly across tokens, ignoring that the same expert may serve different roles for vision versus language processing. To address this, we propose MAEP (Modality-Aware Expert Pruning), a training-free framework that decomposes expert importance by modality in a single forward pass. MAEP introduces (1) Delta-H Output ( \Delta H ), a redistribution-aware metric measuring the MoE output change when an expert is removed, and (2) low-attention vision tokens as a vision-side pruning signal that is dataset-stable: its expert-importance ranking remains stable across different datasets. Text-side importance serves as a preservation signal while vision-side importance serves as a pruning signal, and experts are ranked globally across layers. Experiments on three MoE-based MLLMs across multimodal and text benchmarks show that MAEP outperforms modality-agnostic baselines on the majority of multimodal benchmarks while remaining competitive on text-only tasks, reducing the BF16 expert weight footprint by about 27 GB (about 47% of total memory) so that the full weight footprint of an approximately 58 GB model fits within a single 40 GB GPU. Code is available at this https URL.
[NLP-50] Routes of Reasoning : Reading and Reshaping Test-Time Compute in MoE Models
【速读】： 该论文旨在解决生成式AI在推理过程中存在的“过度思考”（overthinking）问题，即模型在找到正确答案后仍持续进行验证和重复推导，导致计算资源浪费且可能降低性能。现有方法如令牌预算或引导策略主要通过控制推理时长来调节行为，但无法针对特定推理行为进行精准干预。本文提出利用混合专家模型（Mixture-of-Experts, MoE）路由器作为无需训练的可解释性接口，实现对推理行为的精细控制。研究首先将推理轨迹分割为句子级片段，并标注六类推理行为（规划、执行、重述、事实、反思、结论），使每个标记均携带其所属行为信息。线性探测实验表明，基于路由器输出的逻辑值预测行为标签的性能几乎与基于残差流（residual stream）等完整隐藏状态的探测相当，但仅需约一半特征维度。该信号具有强泛化能力：某一数据集上与特定行为关联的专家，在跨六个数据集、五个领域（数学、代码、科学、知识、常识）中均能有效预测相同行为。基于此，研究在推理阶段选择性修改专家路由器得分，以抑制不必要的重复推理。在四个模型和四个基准上的实验显示，该方法使生成令牌数中位数减少30.4%（最高达63.3%），且关键优势在于促使模型在问题解决后及时终止，而非跳过必要推理步骤；准确率最多下降0.5个百分点，通常反而提升。综上，MoE路由机制不仅提供了对推理状态的可解释视角，还为测试时计算优化提供了一种轻量高效的解决方案。
链接: https://arxiv.org/abs/2611.05919
作者: Shengjie Wang,Mohsen Fayyaz,Nanyun Peng
机构: University of California, Los Angeles(加州大学洛杉矶分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Reasoning models allocate test-time compute across distinct behaviors such as planning, deriving, and verifying. However, models frequently “overthink,” continuing to verify and re-derive long after finding a correct answer. This wastes expensive compute and can even degrade performance. Existing interventions, such as token budgets or steering, primarily regulate reasoning duration rather than targeting specific behaviors. In this work, we demonstrate that Mixture-of-Experts (MoE) routers provide a natural, training-free interface for behavioral control. We first split reasoning traces into sentence-level segments and label each with one of six reasoning behaviors (planning, execution, restatement, fact, reflection, and conclusion), so that every token carries the behavior of its segment. Linear probes on router logits predict this label nearly as well as probes on the residual stream, the model’s full hidden state and the standard target for probing, while using about half as many features. These signals generalize: experts associated with a reasoning behavior on one dataset predict the same behavior on the others, across six datasets spanning five domains (math, code, science, knowledge, and commonsense). Building on this, we selectively edit expert router scores during inference to curb overthinking. Across four models and four benchmarks, this intervention reduces the number of generated tokens by a median of 30.4% (and up to 63.3%). Crucially, it encourages the model to stop once the problem is solved, rather than skipping essential reasoning steps. Accuracy never drops by more than 0.5 percentage points, and frequently improves. Ultimately, MoE routing offers both an interpretable lens into reasoning states and a lightweight mechanism for optimizing test-time compute.
[NLP-51] Autonomous agents for electricity-system planning
【速读】： 该论文旨在解决电力系统规划中复杂且重复性决策问题，即如何在区域资源、投资、运行及网络约束的多重耦合条件下，实现低碳目标下的高效、可扩展的自主规划。其核心挑战在于协调多时间尺度的规划决策（如年投资与调度优化）、处理不确定性情景，并确保碳排放约束的严格满足。解决方案的关键在于提出一个自主规划框架，通过智能体（agent）驱动的迭代流程——包括方案生成、数值评估、反馈解析、设计修正与终止判断——实现闭环优化。该框架自动执行年度投资与调度优化以及代表性馈线验证，采用固定接受标准并保留最低成本的可行方案。通过在五个电力系统、四个国家的十种系统年情景下评估40条碳约束路径，并结合96条在统一1,000 Mt/年碳预算下的轨迹进行对比分析，验证了该方法在公开区域数据和既定机构需求预测基础上，实现具有物理可行性与碳约束合规性的自主规划能力。
链接: https://arxiv.org/abs/2611.05911
作者: Yoon Pyo Lee,Samrendra Roy,Jay Phil Yoo,Kazuma Kobayashi,Souvik Chakraborty,Syed Bahauddin Alam
机构: University of Illinois Urbana-Champaign(伊利诺伊大学厄本那-香槟分校); Indian Institute of Technology Delhi(印度理工学院德里分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Planning electricity systems requires repeated decisions that connect regional resources, investment, operation and network constraints. Here we present an autonomous planning framework that carries prepared regional scenarios through proposal, numerical evaluation, feedback interpretation, redesign and final plan selection. Agents propose investment bounds, storage operating conditions and feeder interventions, interpret numerical feedback, and decide how to revise designs and when to stop. The framework automatically executes annual investment-and-dispatch optimization and representative-feeder verification, applies fixed acceptance criteria and retains the lowest-cost accepted candidate. We evaluate forty carbon-constrained trajectories across ten system-year scenarios in five electricity systems and four countries, using two model deployments and single- and multi-agent architectures. All forty retain candidates satisfying the implemented physical and fixed annual operational carbon-budget checks. Twenty-eight of forty matched reference designs exceed the new scenario ceilings. A separate archive of ninety-six trajectories under a common 1,000 Mt annual ceiling provides matched references and broader regional coverage. The results demonstrate autonomous planning under explicit emissions constraints using public regional data and published institutional demand outlooks.
[NLP-52] Selective Risk-Controlled Emotion Forecasting in Conversation: From a Persistence Barrier to Departure Detection with Conformal Guarantees
【速读】： 该论文旨在解决对话情绪预测中将情绪状态视为可直接回归的目标所带来的根本性问题，即在仅观测到当前状态的测试环境中，传统方法难以有效区分情绪变化的真实信号与由个体情绪持续性（persistence）带来的伪相关。其核心解决方案是将原任务重构为“选择性偏离预测”（selective departure forecasting），通过引入带提前量加权的评估指标与泄露控制协议，结合动态承诺策略（commitment policies），使模型能够学习何时偏离自身历史情绪状态。关键在于：在四个语料库中的三个上，该方法可学习到有效的偏离时机信号；而在基率最高的语料库中表现接近随机，这一负面结果亦被明确报告。进一步地，论文采用对话级置信区间控制（conformal risk control）对承诺预测的期望损失进行约束，在全部十二种配置下均满足目标，但该保证以放弃预测（abstention）为代价，导致在最困难语料库上的覆盖率与报警召回率显著下降。此外，研究发现不同语言模型的置信度作为排序信号的表现具有显著差异，这种模型依赖性失败尚未能归因于模型家族或规模差异。
链接: https://arxiv.org/abs/2611.05896
作者: Bin Wen,Dai-Qiao Zhang,Tien-Ping Tan
机构: Universiti Sains Malaysia (马来西亚理科大学); School of Computer Sciences (计算机科学学院); School of Mathematical Sciences (数学科学学院)
类目: Computation and Language (cs.CL)
备注:
Abstract:Forecasting emotion in conversation is increasingly framed as predicting the next emotional state’s value. We show this targets a hard quantity in a state-observed testbed and propose a learnable, deployable reformulation instead. On valence-arousal-dominance trajectories from a dyadic corpus, repeating a speaker’s own previous state sets a persistence baseline that no frozen-encoder correction meaningfully improves upon and a state-space filter exceeds; the floor requires the historical state to be observed, since an estimated anchor collapses persistence below a direct forecaster. What survives is the timing of a departure from persistence, learnable on three of four corpora and near chance on the highest-base-rate corpus, a negative result we report. We recast the task as selective departure forecasting: a lead-time-weighted metric with a leakage protocol, and commitment policies whose best choice switches with the departure rate. A dialogue-level conformal risk control bounds the expected loss of committed predictions, meeting the target in all twelve configurations tested, though the guarantee is bought through abstention, with coverage and alarm recall falling sharply on the hardest corpus. One tested language model’s confidence fails as a ranking signal where another succeeds, a model-dependent failure not yet attributable to family versus scale. Code and result logs are available from the corresponding author upon request.
[NLP-53] Information-Matched Auditing of Conversational Emotion-Shift Forecasting: A Leakage-Safe Evaluation Protocol
【速读】： 该论文旨在解决对话情感转移预测（conversational emotion-shift forecasting）中因信息不对称与模型泄露导致的评估偏差问题，具体表现为：当学习模型接收丰富特征而基线模型依赖噪声或不可用的当前情感标签、超参数预算不一致，或目标轮次信息意外渗入表示时，评估结果易产生误导。其解决方案的关键在于提出一种信息匹配且防泄露的审计协议（auditing protocol），通过引入明确的承诺点（commitment point）、可执行的因果性检验（executable causality tests）、四配置均衡的验证搜索机制以及基于种子对话的推理设计，确保评估过程在公平、可控条件下进行。实验表明，在部署场景下仅使用训练阶段可获得的当前情感分类器作为基线时，学习模型在所有四个主要语料库上均显著优于预测标签基线；其中DailyDialog数据集表现尤为突出，经72次比较的Holm校正后仍保持显著优势（最佳ROC-AUC提升+0.167，调整后p=0.0144）。而在信息匹配的黄金标签条件下，基线在IEMOCAP上保持竞争力，在MELD上更占优势，但所有模型在DailyDialog上依然具备显著优势。研究还进一步报告了精确率-召回率AUC、校准性、严格局域特征控制、同人自转移预测及未来信息剂量反应测试等多维度指标，并配套提供八项泄露检查清单。代码、原始种子预测与报告材料可向通讯作者申请获取。
链接: https://arxiv.org/abs/2611.05895
作者: Bin Wen,Dai-Qiao Zhang,Tien-Ping Tan
机构: Universiti Sains Malaysia (马来西亚理科大学); School of Computer Sciences (计算机科学学院); School of Mathematical Sciences (数学科学学院)
类目: Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Comparisons in conversational emotion-shift forecasting are easily distorted when a learned model receives rich features but its transition baseline is given a noisy or unavailable current-emotion label, when hyperparameter budgets differ, or when target-turn information enters the representation. We present an information-matched, leakage-safe auditing protocol that closes these gaps with an explicit commitment point, executable causality tests, equal four-configuration validation searches, and seed-dialogue-aware inference. Applying it to six causal model families on four text corpora and two feature-source sensitivity configurations shows that a deployable train-only current-emotion classifier feeding the transition baseline yields a different verdict than an oracle diagnostic where both sides receive gold current emotion. Under the deployable condition, learned models improve over the predicted-label baseline on all four primary corpora; DailyDialog gains are large and survive a 72-comparison Holm correction (best ROC-AUC gain = +0.167, adjusted p=.0144). Under matched gold information, the baseline remains competitive on IEMOCAP and stronger on MELD, while every model retains a large DailyDialog advantage. We additionally report precision-recall AUC, calibration, a strictly utterance-local feature control, same-person self-shift forecasting, and a future-access dose-response test. An eight-item leakage checklist accompanies the protocol. Code, raw seed predictions, and reporting artifacts are available from the corresponding author upon request.
[NLP-54] When Does Better Confidence Improve a Knowledge-Graph Audit?
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）构建知识图谱过程中，置信度估计在人工审核中的有效性不足问题，即尽管能够更准确地预测错误，但对审计效率的提升有限。其核心挑战在于理解误差可修复性（repairability）、选择冗余空间（selection headroom）与标签有效性（label validity）之间的关系。解决方案的关键在于将非自适应审核建模为带预算的覆盖问题，并通过结构质量（structural mass）Δ^cl 对标量保持的一阶规划方法的遗憾（regret）进行理论界定。研究通过在Re-DocRED数据集上模拟理想修复过程，对比五种抽取配置与两个跨文档图结构，在冻结确认测试集及500个哈希选取训练文档上进行复现，发现仅依赖事实的联合模型相比独立块模型在相同操作下仍存在3.98–4.74个百分点的残余错误；而开放共享操作后，联合规划器相较于自身仅基于事实的计划，残余错误降低0.89–1.43个百分点。此外，通用大语言模型验证在多数模型-预算配置中表现优于基准，且大语言模型仲裁揭示了潜在的基准遗漏问题。
链接: https://arxiv.org/abs/2611.05877
作者: Zehao Lin,Kai Chen,Xixuan Hao,Erenkai Zhu,Chunyu Li,Helan Hu
机构: MemTensor; Shanghai Jiaotong University (上海交通大学); The Hong Kong University of Science and Technology (Guangzhou) (香港科技大学（广州）)
类目: Computation and Language (cs.CL)
备注: 27 pages, 4 figures, 19 tables. Ancillary files include a public rendering of the confirmation protocol and summary results
Abstract:Confidence estimates guide human review in LLM-built knowledge graphs, but better error prediction can yield little improvement in an audit. We study this gap through repairability, selection headroom and label validity. Modelling non-adaptive auditing as budgeted covering, we characterise the minimal decision-relevant error statistics and bound the regret of scalar-preserving first-order planning by a structural mass \Delta^\mathrmcl . We simulate ideal repair against benchmark-aligned labels for five extraction configurations and two cross-document graphs on Re-DocRED, followed by frozen confirmation on test and replication on 500 hash-selected training documents. On the held-out cross-document graphs, an error-rate \times fan-out queue leaves 3.98-4.74 percentage points more errors than a joint model with the same actions. By comparison, opening shared actions for that joint planner reduces residual errors by 0.89-1.43 points relative to its own fact-only plans, in point estimates. Equivalence between joint and independent-block models remains inconclusive. Generic LLM verification improves three of four model-budget configurations on each held-out set. LLM adjudication identifies possible benchmark omissions.
[NLP-55] CARTBench: Rubric-Based Evaluation of Shopping Advice for Complex and Ambiguous Requests with Timelines
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）购物助手在多轮对话中进行商品推荐时面临的评估难题，具体表现为数据稀缺性与评价标准难以统一。为应对这一挑战，研究提出CARTBench评测基准，聚焦于具有时间线的复杂且模糊的购物请求（Complex and Ambiguous Requests with Timelines, CART）。针对数据稀缺问题，研究团队从17个零售领域的真实商业需求中收集案例，构建了包含1603条商品记录与带时间戳对话的数据集；针对评价标准不统一的问题，基于成功订单中人工精选的优质回复提炼出“理想回应原则”（Ideal Response Principles），并据此制定案例定制化的评分细则，同时设计了基于原则引导的结构化推荐策略——Principle-Guided CART Prompting。实验评测涵盖1328项由专家标注的评估维度，覆盖核心能力、卓越表现与安全性三个层面。对7款主流商业API模型（包括GPT-6.1-Sol和Claude-Opus-5）及350亿参数以上开源模型的全面评估表明，在冻结评分规则下，仅有3款模型在核心能力上超过90%，而无一模型在卓越性指标上突破45%。这揭示了当前模型在理解需求与高质量执行之间存在显著“理解-执行鸿沟”。进一步诊断发现，模型失败主要源于对产品证据的识别不足以及未解决的需求遗留问题，其根源在于动态变化的需求与关键未知信息的不确定性。
链接: https://arxiv.org/abs/2611.05862
作者: Yuqin Dai,Zhanwei Zhang,Honghui Yang,Zhouheng Yao,Yueqi Zhang,Jingyu He,Chaoyang Shi,Fanding Wei,Boyuan Pan,Kai Xu,Yao Hu,Yongliang Shen
机构: Tsinghua University(清华大学); Xiaohongshu Inc.(小红书); Shanghai Jiao Tong University(上海交通大学); Zhejiang University(浙江大学)
类目: Computation and Language (cs.CL)
备注:
Abstract:Large language model (LLM) shopping assistants now recommend products in multi-turn conversations. However, reliable evaluation faces Data scarcity and Difficulty Standardizing Criteria. To address these obstacles, we introduce CARTBench for CART (Complex and Ambiguous Requests with Timelines). For Data scarcity, we curate cases from real commercial shopping needs across 17 retail domains, pairing timestamped dialogue with 1,603 product records. For Difficulty Standardizing Criteria, we develop case-specific rubrics guided by Ideal Response Principles, distilled from excellent replies manually selected from successfully completed orders. These principles also guide Principle-Guided CART Prompting, our structured recommendation strategy. The experimental benchmark contains 1,328 expert-annotated criteria across Core, Excellence, and Safety. To answer the CART question, we evaluate 7 flagship commercial API models, including GPT-6.1-Sol and Claude-Opus-5, and 35B+ parameters open-weight models. In the first full-suite evaluation under the frozen rubrics, 3 of 10 models exceed 90% on Core, while none exceeds 45% on Excellence. Therefore, we identify an understanding-to-execution gap between required behavior and high-quality purchase advice. To examine these limitations, we diagnose shared failures in product evidence and unresolved needs, linked to changing requirements and decisive unknowns. Code and data are available at this https URL.
[NLP-56] andemBench: Benchmarking Decision Models for Multi-Agent Collaboration
【速读】： 该论文旨在解决多智能体协作场景中决策模型在团队任务规划选择与最优性验证方面的局限性问题。其核心挑战在于：尽管现有决策模型能够从给定选项中选出看似合理的团队计划（team plan），但难以准确判断所供选项中是否已包含最优解，即存在系统性误判风险。解决方案的关键在于构建TandemBench这一基准数据集，包含1,200个基于多智能体烹饪场景的复杂协作任务，所有任务均基于完整全局信息和候选团队计划进行评估，并通过精确搜索与执行验证生成参考答案。通过设计成对测试（paired tests）——在保留或移除最优计划的两种情境下对比模型判断能力——揭示出当前主流决策模型虽在选择正确计划方面达到62.50%的准确率，但在最优性验证任务上表现接近随机水平，且部分模型会错误报告“存在更优方案”，即使已有最优计划被提供。这一发现表明，仅依赖团队计划选择的准确性无法反映模型在最优性判断上的真实能力，凸显了对决策模型进行更深层次合理性验证的重要性。
链接: https://arxiv.org/abs/2611.05827
作者: Baohang Li,Xiaocheng Feng,Bing Qin
机构: Harbin Institute of Technology (哈尔滨工业大学); Peng Cheng Laboratory (鹏城实验室)
类目: Computation and Language (cs.CL)
备注:
Abstract:Decision models choose among predefined options in tasks such as routing and action selection. In multi-agent collaboration, a decision determines how members divide work, exchange resources, and act in parallel. We introduce TandemBench, a benchmark of 1,200 questions in multi-agent cooking scenarios where coordination choices affect goal completion. Models receive complete global information and candidate team plans, with reference answers established through execution and exact search. Models select an optimal team plan and, in paired variants that retain or remove it, judge whether better arrangements exist beyond the supplied options. We evaluate 40 native decision models and compare them with 10 general-purpose language models. The best native decision model achieves 62.50% decision accuracy. Native decision models remain close to chance in optimality verification. Paired tests reveal that some report better alternatives regardless of whether an optimal plan is already present. These findings show that team-plan decision accuracy alone can conceal systematic errors in judging whether the supplied options contain an optimal arrangement. The dataset and evaluation code are publicly available at this https URL.
[NLP-57] Learning When Human Decisions Matter: Executor-Relative Consultation Value for LLM Agents
【速读】： 该论文旨在解决在大型语言模型（Large Language Model, LLM）执行复杂任务时，如何高效分配有限的人类决策资源（human decision capacity）的问题。核心挑战在于：人类输入的咨询价值具有执行器依赖性（executor-dependent），即同一人类反馈对不同下游代理的增益存在差异，因此需动态评估每项决策的咨询价值以实现最优资源配置。其解决方案的关键是提出一种离线式咨询价值蒸馏（Consultation Value Distillation, CVD）方法，通过蒙特卡洛执行器回溯（Monte Carlo executor rollouts）在训练阶段预估咨询价值，并将这些价值知识蒸馏为轻量级预测器，从而在部署阶段无需依赖实时回溯即可进行预算约束下的近似最优咨询决策。CVD通过联合建模与延续感知（continuation-aware）的价值学习机制，显著提升了多轮和批量场景下的分配性能，在多个前沿执行器设置中优于独立咨询价值估计、直接提示（direct prompting）及信息价值（Value of Information）等基线方法，同时具备更强的鲁棒性与更低的回溯监督成本，实现了近乎无损的低开销近似分配。
链接: https://arxiv.org/abs/2611.05814
作者: Shuli Jiang,Khushbu Pahwa,Yuting Zhang,Yi Zhang,Wei Xia,Stefano Soatto
机构: Amazon(亚马逊); AWS Agentic AI(亚马逊网络服务代理人工智能)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 31 pages, 20 tables, 2 figures
Abstract:Large language models increasingly automate complex tasks, but human decision capacity does not scale with their capabilities. In many applications, humans remain essential for providing private context, authority, or judgments unavailable to the agent. We formulate human decision capacity allocation: given multiple consultation opportunities and a limited budget, at which decisions should an LLM agent consult a human? The central challenge is estimating consultation value, which is executor-dependent because the same human input may benefit different downstream agents differently. We introduce Consultation Value Distillation (CVD), which estimates consultation values offline using Monte Carlo executor rollouts and distills these values into lightweight predictors for budgeted allocation, requiring no rollout-based value estimation at deployment. Across five batch and three sequential settings with two frontier executors, CVD improves allocation through joint and continuation-aware value learning, while independent consultation values remain competitive in several settings. CVD also broadly outperforms direct prompting and Value of Information where applicable, remains robust with substantially less rollout supervision, and enables nearly lossless approximate allocation at low search cost.
[NLP-58] Matching the Training-Time Weight Rule in Inference-Time Expert Deactivation
【速读】： 该论文旨在解决在推理阶段进行专家（expert）去激活时，由于令牌路由策略与专家权重更新规则不匹配所导致的模型性能下降问题。其核心挑战在于：专家去激活不仅影响令牌的路由路径，还会改变专家权重分布，而这些变化对模型表现的影响高度依赖于训练阶段所采用的权重更新规则（如归一化或非归一化）。论文的关键解决方案是提出一个加权求和模型（weight-sum model），用于量化分析训练阶段的权重规则如何影响专家贡献损失与分支尺度变化，并据此预测不同去激活策略下的性能表现。实验结果表明，在使用原始路由器概率训练的模型中，保留专家重归一化（remaining-expert renormalization）能获得最佳困惑度；而在使用归一化权重训练的模型中，非归一化重路由（unnormalized re-routing）表现更优。当训练规则与推理策略不一致时，困惑度可恶化达28,740倍，且任务表现显著退化——例如在OLMoE上，GSM8K严格匹配准确率降至零。此外，实验还验证了随着活跃专家数量减少，额外的重归一化带来的困惑度惩罚呈上升趋势，揭示了专家稀疏性与模型稳定性之间的关键权衡。
链接: https://arxiv.org/abs/2611.05804
作者: Hao Cheng(Zhejiang Gongshang University),Zhiyu Yang(Chang’an University),Bo Ma(Zhejiang Gongshang University),Hanzhang Jia(University of Sussex)
机构: Zhejiang Gongshang University(浙江工商大学); Chang’an University(长安大学); University of Sussex(萨塞克斯大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: Corresponding authors: Hanzhang Jia (lead corresponding) and Bo Ma
Abstract:Expert deactivation at inference time changes both token routing and expert weights, and its effect on performance depends on how these changes are implemented. We compare removal, freezing, remaining-expert renormalization, and unnormalized re-routing on identical disabled sets across five instruction-tuned checkpoints from four model families, with a base OLMoE checkpoint as an additional comparison. We develop a weight-sum model that relates the loss of expert contributions and changes in branch scale to the weight rule used during training. With half of the experts deactivated according to activation-based rankings, remaining-expert renormalization yields the highest perplexity in models trained with raw router probabilities, whereas unnormalized re-routing yields the highest perplexity in models trained with normalized weights. Perplexity ranges from 1.4 to 73 times the intact baseline when the training rule is retained, and from 14 to 28,740 times the baseline under a rule mismatch. Task outcomes vary across evaluation types: on OLMoE, GSM8K strict-match accuracy falls to zero under all four implementations. Extending the model to changes in the number of active experts yields a further prediction, supported by the OLMoE experiments: the perplexity penalty from additional renormalization increases as fewer experts are activated.
[NLP-59] Phonological Knowledge of Chinese Characters in Language Models
【速读】： 该论文旨在探究语言模型（Language Models, LMs）对汉字的语音知识掌握程度及其表征与使用机制，尤其关注汉语中声母、韵母、声调及多音字等语音特征的理解能力。其核心问题是：当前主流文本仅输入的语言模型在处理汉字语音信息时，究竟具备何种层次的语音知识？解决方案的关键在于构建一个受控基准测试集——PinyinBench，涵盖同音字、语音特征（声母、韵母、声调）以及多音字，并采用行为评估、层间探针分析和激活操控三种互补方法进行综合考察。研究发现，模型在声母与韵母信息上表现出较强的表征能力，且可通过深度网络逐步增强；而声调信息虽在嵌入层可被有效解码，但其在生成过程中的可操控性较弱，表明声调表征难以通过因果干预有效影响输出。此外，引入音频模态训练虽提升了语音特征的可解码性，却未显著改善其操控效果。这一结果揭示了音段信息与声调信息在模型内部表征与功能上的本质差异，凸显了结合行为表现、探针分析与因果干预多维度评估语言模型语言知识的必要性。
链接: https://arxiv.org/abs/2611.05803
作者: Xiaozhe Ji,Renfen Hu,Hai Hu
机构: Beijing Normal University (北京师范大学); The Hong Kong Polytechnic University (香港理工大学)
类目: Computation and Language (cs.CL)
备注: 17 pages, 11 figures, 10 tables
Abstract:How much phonological knowledge do language models (LMs) possess about Chinese characters, and how is this knowledge represented and used? We introduce PinyinBench, a controlled benchmark covering homophones, phonological features (initials, finals, and tones), and polyphonic characters. We investigate 17 text-only and multimodal LMs through three complementary approaches: behavioral evaluation, layer-wise probing, and activation steering. Behaviorally, models demonstrate substantial knowledge of initials and finals, with a frontier model (DS-v4-flash) scoring 90%+ in homophone judgment, when two characters share no phonological features. However, its accuracy drops sharply to 72% when two characters differ only in lexical tone. Probing reveals that tone is highly decodable from embeddings and improves modestly across layers, whereas segmental information starts less decodable and improves substantially with depth. Activation steering shows that segmental representations can systematically influence generation, while tone exhibits much weaker effects. Furthermore, audio-bearing training improves phonological decodability, including tone, without corresponding improvements in steering effectiveness. Together, these findings reveal distinct representational and functional profiles for segmental and tonal information, with tonal information much harder to access and manipulate, demonstrating the necessity for combined evaluation, probing, and causal intervention for understanding linguistic knowledge in LMs.
[NLP-60] How Well Can Large Language Models Impersonate Individual Authors? A Comparative Evaluation
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在文本生成中对特定作者写作风格的模仿能力问题，尤其是识别其是否能够实现真正的风格拟合而非简单地复制训练数据。研究聚焦于六种主流大语言模型（Gemini 3 Pro、Claude Opus、GPT 5.2、GPT 5.2 Pro、DeepSeek 3.2 和 Grok 4.1 Fast）在短产品评论文本上的作者身份伪装能力，采用基于余弦距离的嵌入相似性度量与嵌入分布可视化相结合的方法进行评估。关键发现在于：尽管 Gemini 3 Pro 在表面模仿性能上表现最优，但其实际表现为大量重复和复用训练语料中的句子与短语，远超人类作者水平；进一步的因子分析表明，训练集一致性是预测模仿成功的关键因素，而作者独特性则无法作为可靠预测指标。此外，全局嵌入空间可视化揭示多数模型生成文本位于与目标人类作者集群明显分离的区域，反映出模型存在固有的词汇与短语偏好，难以真正收敛至目标作者风格。因此，该研究的核心解决方案在于提出区分“真实风格模仿”与“直接文本复用”的判别框架，为涉及大语言模型潜在伪造行为的文本司法鉴定提供了重要方法论启示。
链接: https://arxiv.org/abs/2611.05771
作者: Harshit Gupta,Shunichi Ishihara
机构: 未知
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注:
Abstract:This study investigates the authorship impersonation capacity of six widely used large language models (LLMs): Gemini 3 Pro, Claude Opus, GPT 5.2, GPT 5.2 Pro, DeepSeek 3.2, and Grok 4.1 Fast. The evaluation was conducted using short product review texts selected to ensure high within author consistency in individual writing style. Two types of in context prompt were designed: a simple prompt with minimal instruction and a complex prompt incorporating explicit stylometric guidance for mimicry. Authorship mimicry was evaluated using a cosine distance based distance to training metric, supplemented by visual analyses of embedding distributions. The results indicate that authorship impersonation remained challenging for most LLMs, with Gemini 3 Pro showing the strongest apparent mimicry performance. However, subsequent overlap analysis revealed that Gemini 3 Pro extensively copied and reused sentences and phrases from the training texts, at levels far exceeding those observed among human authors. Follow up factor analyses showed that training set consistency was the most robust predictor of mimicry success across models, whereas author distinctiveness did not generalise as a reliable predictor. Global embedding visualisations further revealed a consistent macro level pattern: for most models, generated texts occupied regions of the embedding space that were largely separated from the corresponding human training clusters. This indicates limited convergence towards target author style and suggests the presence of model specific lexical and phraseological preferences. The findings highlight the importance of distinguishing genuine stylistic imitation from direct textual reuse and have implications for the forensic analysis of textual evidence involving potential LLM enabled impersonation.
[NLP-61] PAYWORTHY: Can LLM s Predict Which Freelance Jobs an AI Agent Will Complete?
【速读】： 该论文旨在解决在生成式 AI (Generative AI) 代理执行付费任务前，如何有效评估其成功可能性的问题。现有机制缺乏对AI代理胜任力的可靠预判能力，导致难以实现高效的任务分配与信任建立。其核心解决方案是提出PAYWORTHY数据集，包含1,494个真实自由职业岗位（从2020至2026年间平台发布的1880万条岗位中采样），并通过让大型语言模型（LLM）作为客户裁判，在单次运行中评估由不同描述的AI代理尝试任务后的成功率。研究发现：当明确告知代理的能力限制时，18个测试模型均能显著优于随机排序地预测成功任务（AUC达0.66–0.86）；而若仅将代理描述为具备通用浏览器与代码执行能力的“通用代理”，则所有模型的预测表现退化至接近随机水平（AUC 0.45–0.56）。这一结果在四类其他模型家族（Claude、GLM、Kimi、Qwen）及两名人类评判者中均具稳健性，表明模型的判断能力高度依赖于对代理能力边界的明确认知。关键结论在于：准确评估生成式AI代理的胜任力，前提是提供清晰、具体的能力边界信息，而非模糊的“通用”设定。
链接: https://arxiv.org/abs/2611.05759
作者: Mohammad Reza Rezaei,Dingxian Wang,Brett Levert,Celeste Fejzo,Andrew Rabinovich,Zhao Chen
机构: Upwork Inc.(Upwork公司)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Before an AI agent takes on a paid job, someone has to judge whether it is likely to succeed. We test whether frontier language models can make this call before any attempt. We propose PAYWORTHY which contains 1,494 real freelance job postings, sampled from 18.8M postings made between 2020 and 2026 on a freelancing platform. An agent attempts each job from the posting alone in a single run, and an LLM judge acting as the client decides whether to accept the work. We ask 18 models to forecast how likely the agent is to succeed and measure how well each model’s forecasts rank the accepted jobs above the rejected ones (AUC, where 0.5 is a random ranking). The answer depends on how the agent is described. Told what the agent cannot do, every model ranks the jobs well (AUC 0.66–0.86). Told instead that it is a generic agent with a browser and code execution, the same models rank the same jobs about as well as a random ordering (0.45–0.56). These results hold under judges from four other model families (Claude, GLM, Kimi and Qwen), and the informed forecasts also rank the verdicts of two human raters. Code is available at this https URL Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL) Cite as: arXiv:2611.05759 [cs.AI] (or arXiv:2611.05759v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.05759 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-62] xt-Only Domain Adaptation for Low-Resource Medical ASR: Shallow Fusion and Density Ratio Rescoring for Hausa
【速读】： 该论文旨在解决低资源语言（如豪萨语）在医疗领域中自动语音识别（ASR）性能不佳的问题，尤其针对医疗专有词汇识别困难的挑战。其核心问题在于缺乏足够的领域特定语音数据，导致现有模型难以有效捕捉医疗语境下的语言特征。解决方案的关键在于采用轻量级、仅依赖文本的语言模型融合策略，利用固定声学模型（MMS-1B-FL102）结合两种解码时语言模型融合方法——浅层融合（Shallow Fusion, SF）与密度比法（Density Ratio Approach, DRA），通过在通用领域语言模型基础上引入医疗领域文本训练的4-gram语言模型（KenLM），实现对医疗术语的增强建模。实验结果表明，两种融合方法均显著降低词错误率（WER），分别降至23.55%和23.33%，优于贪婪解码（p < 0.001），且二者差异不显著（p = 0.423）。尽管医疗语言模型带来一定改进，但相较于通用语言模型，其增益有限，且在医疗术语错误率方面未展现出明显优势。研究证明，无需额外收集领域语音数据，仅通过文本层面的语言模型融合即可有效提升豪萨语医疗ASR性能。
链接: https://arxiv.org/abs/2611.05744
作者: Oche David Ankeli,Hassan Adelani Luqman,Theodora Ngozi Egbunike,Joshua Gottschalk,Bonaventure F. P. Dossou
机构: RUZIVO Research Lab, African Leadership University, Kigali, Rwanda; Stanford University, Stanford, CA, USA; Mila – Quebec AI Institute and McGill University, Montreal, Canada
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:Automatic speech recognition (ASR) systems often struggle with specialised medical vocabulary, especially in low-resource languages where domain-specific speech data are scarce. We investigate text-only domain adaptation for Hausa medical ASR using the fixed MMS-1B-FL102 acoustic model and two decoding-time language-model fusion methods: Shallow Fusion (SF) and the Density Ratio Approach (DRA). A 4-gram KenLM trained on 19,421 Hausa health sentences serves as the medical target model, while a 75,372-sentence general-domain model provides the source prior for DRA. Evaluation uses 250 recordings of 25 physician-reviewed prompts from 10 speakers across six northern Nigerian states. Under prompt-disjoint five-fold cross-validation, SF and DRA reduce WER from 29.56% to 23.55% and 23.33%, respectively. Both significantly outperform greedy decoding (p 0.001), while their difference is not significant (p = 0.423). However, general-LM-only SF achieves 23.96% WER, indicating that the medical LM’s additional improvement over a general LM is modest. Medical Term Error Rate also improves under fusion, but does not establish a distinct medical-domain advantage. These findings demonstrate that lightweight, text-only language-model fusion can improve Hausa medical ASR without additional domain-specific speech collection.
[NLP-63] Prospective Memory for Proactive LLM Agents
【速读】： 该论文旨在解决大语言模型（LLM）智能体在执行前瞻性任务时缺乏有效未来导向记忆机制的问题。现有记忆系统多为回溯性记忆，仅能存储和调用过往信息以支持当前决策，而无法有效维持并适时实现未来意图。其核心挑战在于：即使正确记住未来目标，也未必能在恰当的时机采取行动——某些情境可能需要延迟响应，而后续线索才真正触发执行条件。为此，论文提出Pro-Memory框架，其关键创新在于将“保持未来意图”与“决定何时何地实施”进行解耦。该框架通过保留意图及其支撑证据，并基于原始目标与激活条件动态评估新观测，从而判断是否应采取行动；同时，它根据实际执行行为或完成状态更新意图状态，并利用经验记忆从已完成任务中提炼反馈，指导未来决策。实验结果表明，Pro-Memory在延迟意图任务和开放式主动辅助任务两个基准上，相较于多种代表性记忆方法，在不同LLM模型架构下均显著提升了前瞻性任务表现，且有效减少了不必要的干预。
链接: https://arxiv.org/abs/2611.05731
作者: Yuanchen Bei,Zhining Liu,Tianxin Wei,Weizhi Zhang,Zhichen Zeng,Philip S. Yu,Hendrik Hamann,Jingrui He,Hanghang Tong
机构: University of Illinois Urbana-Champaign(伊利诺伊大学厄本那-香槟分校); University of Illinois Chicago(伊利诺伊大学芝加哥分校); Stony Brook University(石溪大学); Brookhaven National Laboratory(布鲁克海文国家实验室)
类目: Computation and Language (cs.CL)
备注: 17 pages, 8 figures
Abstract:Most memory systems for LLM agents are retrospective, preserving and retrieving past information to inform current decisions. Proactive agents, however, also require prospective memory: retaining future-directed intentions and realizing them when appropriate opportunities arise, even without explicit requests. Crucially, remembering the right intention does not guarantee acting at the right time: a related observation may call for waiting, while a later cue may make the same intention actionable. Inspired by cognitive theories of human prospective memory, we introduce Pro-Memory, a memory framework that separates maintaining future intentions from deciding when and how to realize them. Pro-Memory preserves pending objectives linked to their supporting evidence, assesses new observations against their original goals and activation conditions, and coordinates appropriate contributions with the agent’s ongoing actions. It further updates intention states based on actions actually issued or evidence of resolution, while an experience memory distills feedback from completed episodes into guidance for future decisions. Experiments on two benchmarks spanning delayed-intention tasks and open-ended proactive assistance demonstrate consistent improvements in proactive task performance over representative memory approaches across multiple LLM backbones, while limiting unnecessary interventions.
[NLP-64] Cognitive Bias Signatures in Search-Equipped LLM Agents : A Stage-Resolved Analysis
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）代理在搜索过程中因初始条件偏差而导致错误修正能力受限的问题，特别是探究初始证据顺序、先前工作记录及初始假设等因素如何影响其检索与决策过程。其解决方案的关键在于构建一个受控的搜索范式（controlled search harness），在固定问题和文档集合的前提下，系统性地操纵结果顺序、历史轨迹和初始假设，从而分离出各因素对最终答案的影响。研究通过10个模型在1000个合成搜索环境中的多维度追踪（包括查询、检索文档、文档打开行为及最终答案），发现最终答案对主动搜索中初始证据顺序更为敏感，且错误路径的历史记录会抑制策略切换，这种影响超出上下文长度的范畴；此外，在错误假设下，阶段测量可区分检索/打开不足与在反证出现后仍持续错误的情况。探索性干预表明，频繁打开反证未必提升准确率，而适度增加反证访问即可显著改善表现，揭示了优化信息获取策略的关键在于精准引导而非简单增加暴露频率。
链接: https://arxiv.org/abs/2611.05720
作者: Yanru Xiao,Yingwei Yu
机构: Amazon Web Services(亚马逊网络服务)
类目: Computation and Language (cs.CL)
备注: 19 pages, 4 figures
Abstract:Search can help LLM agents correct mistakes, but their starting conditions may shape which evidence they find and use. Drawing on cognitive-bias paradigms, we develop a controlled search harness to examine result order, records of prior work, and initial hypotheses while holding each question and document collection fixed. Matched controls distinguish order effects from differences in document composition and the effects of failed work from context length. Across 10 models and 1,000 synthetic search worlds, we track queries, retrieved documents, document openings, and final answers. Our comparisons suggest that final answers are more sensitive to initial evidence order during active search than when all evidence is supplied in full. Dead-end histories can reduce strategy switching relative to matched neutral histories, indicating an effect beyond context length. Under wrong hypotheses, stage measurements distinguish reduced retrieval or opening from errors that persist after counter-evidence is opened. Exploratory interventions show that opening counter-evidence more often can coincide with lower accuracy, while accuracy can improve with little change in opening.
[NLP-65] Grounded but Not Immune: Confirmation Bias in Agent -as-a-judge
【速读】： 该论文旨在解决生成式 AI 代理在开放性任务评估中因“代理作为裁判”（Agent-as-a-Judge）机制所引发的确认偏误（confirmation bias）问题。其核心挑战在于，尽管裁判代理能够自主收集决定性证据以验证结果，但其判断仍可能被无关的、虚假的预筛选提示（pre-screening note）所误导，导致高达40%的正确结论被错误翻转。解决方案的关键在于识别并分类确认偏误对代理行为轨迹的影响，揭示其本质为三种失败模式：过度信任（over-trust，即在正确验证后仍盲目服从外部提示）、幻觉（hallucination，即虚构支持性证据以支持翻转结论）以及调查不足（under-investigation，即在首次出现符合提示的信号时即停止深入探究）。研究发现，大型代理如Claude主要表现为过度信任（占有害翻转的67%-75%），而小型代理则三类错误模式分布较均。尽管采用轻量级缓解措施可部分削弱偏误，但无法彻底消除。该工作通过将外部提示视为受控诊断工具而非对抗攻击，系统性地揭示了确认偏误如何在代理决策过程中逐步侵蚀其判断可靠性。
链接: https://arxiv.org/abs/2611.05662
作者: Drake Benjamin Brown,Khushbu Pahwa,Yubin Ge,Salvatore Romeo,Akhil Agnihotri,Yi Zhang,Xiaofei Ma,Daniele Bonadiman
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: Preprint. Under Review
Abstract:Evaluating open-ended agent tasks increasingly falls to an Agent-as-a-Judge: an evaluator handed tools and the agent’s full trajectory so it can verify a result against the environment. The premise is that a judge that gathers its own evidence should be grounded by it. We show that this grounding does not immunize the judge against confirmation bias: a single false pre-screening note, attributed to an authority and carrying no evidential bearing on the correct verdict, overturns up to 40% of a tool-using judge’s correct verdicts across four judges and five datasets, even though the judge collects decisive evidence. The shift is driven by specific properties of the cue. We treat this cue as a controlled diagnostic of confirmation bias rather than an adversarial attack. Our central contribution is to explain how confirmation bias changes the judge’s behavior in a trajectory. We classify changes in judge behavior and attribute changes to one of three failure modes: over-trust (verify, conclude correctly, then defer), hallucination (fabricate support for the flip), and under-investigation (stop at the first corroborating signal). For the two Claude judges, the failure is predominantly over-trust (67-75% of harmful flips), while smaller judges have error modes spread across all three. Lightweight mitigations blunt but never remove the bias.
[NLP-66] AGORA: Reliability-Aware Aggregation of LLM Judges via Item-Response Theory
【速读】： 该论文旨在解决大语言模型作为评判者（LLM-as-a-judge）在人工评估中因评判者可靠性差异显著而产生的偏差问题，尤其针对小规模标注数据集下传统聚合方法（如Dawid-Skene模型）因样本不足导致的误差估计不准确与权重分配失衡的问题。其解决方案的关键在于提出AGORA（Aggregation Of Reliability-weighted Assessments），一种基于项目反应理论（Item Response Theory, IRT）的无监督聚合框架：该方法将所有评价项目置于单一潜在质量维度上，通过为每位评判者学习一个敏感度参数和有序的评分阈值，实现对评判者可靠性的联合建模，无需依赖人工标注标签即可自动推断评判者信任权重。实验表明，在三个基准测试中，AGORA在大规模评估集上性能不低于甚至优于Dawid-Skene，在数据量缩减至数百项时优势更为明显；同时，其对评判者池分布不均具有更强鲁棒性，且可通过无标签的离散度统计量提前判断信任加权的有效性，其学习得到的信任权重亦能有效反映评判者的实际人类有效性。
链接: https://arxiv.org/abs/2611.05556
作者: Junhyung Ahn
机构: University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:LLM-as-a-judge is a common substitute for costly human evaluation, and aggregating several judges is the standard remedy for any single judge’s bias. Judge pools mix judges of very different reliability, while the human-labeled sets that validate them hold a few hundred items or fewer. In this regime the per-item mean weights weak judges as heavily as strong ones, and confusion-table models such as Dawid-Skene must estimate a per-judge error table from too few items. We propose AGORA (Aggregation Of Reliability-weighted Assessments), an item-response-theory aggregator that places every item on one latent quality scale and describes each judge by a sensitivity parameter and ordered rating thresholds, so judge reliability is learned from all items jointly and without human labels. Across three benchmarks, AGORA matches or exceeds Dawid-Skene on large evaluation sets, and its advantage widens as the set shrinks toward a few hundred items. It remains accurate as pools grow more uneven, a label-free dispersion statistic indicates in advance when trust weighting helps, and its learned trust weights track per-judge human validity.
[NLP-67] Iterated creation grading and revision of data science projects by language models
【速读】： 该论文旨在解决生成式AI在端到端数据科学项目中的自主执行能力及其在反馈机制下的性能提升问题。核心挑战在于评估不同模型在无外部干预情况下完成数据获取、清洗、分析及撰写研究报告等完整流程的能力，并探究反馈循环对最终成果质量的影响。其解决方案的关键在于构建一个“写作者-评分者”双角色闭环系统：由具备特定工具调用能力的“写作者”模型独立完成数据科学项目，再由另一模型或同一模型作为“评分者”进行评估，随后根据反馈迭代优化，共进行三轮循环。实验覆盖七种来自五大开发者的模型（Anthropic、OpenAI、Google、DeepSeek等），形成49个写作者-评分者组合，每组运行三次。结果显示，初始平均得分从59%提升至76%，表明反馈显著改善了项目质量；但提升幅度高度依赖角色分工，其中弱模型难以通过反馈有效提升强模型的表现，且自评得分普遍比其他模型平均分高出约9个百分点，反映出评分者自身评估偏差的存在。此外，即使无具体反馈仅提供修订机会或增加资源预算，部分模型仍实现显著改进，说明过程灵活性与资源投入亦是关键因素。
链接: https://arxiv.org/abs/2611.05548
作者: Rohan Alexander
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:We evaluate the extent to which agentic AI models can complete an end-to-end data science project, and how this improves with feedback. A ‘writer’ model with specific tools creates a complete data science project including downloading, cleaning, and analyzing a dataset, and writing a short paper about it. The project is executed and a ‘grader’ model evaluates it. The writer can then revise its submission, and this repeats for a total of three rounds. We consider seven models from five developers: Anthropic, OpenAI, Google, DeepSeek, and this http URL. Each model is both a writer and a grader, so there are 49 writer-grader pairs and we run each pair three times. Of the 147 initial submissions, 107 run and render, and the average score increases from 59 per cent to 76 per cent after the final round of feedback. By writer the extent of the increase ranges from 3 percentage points for Anthropic’s Opus 5.5 to 46 percentage points for Anthropic’s Haiku 4.5, and by grader it ranges from 13 percentage points for Google’s Gemini 3.8 Flash to 32 percentage points for Anthropic’s Haiku 4.5. When a model grades its own writing the score tends to be about 9 percentage points higher than the average of what the other models give it. About half of the difference that we see from round to round is within the grader’s own noise. Feedback generally results in the highest marks, but even just having the chance to revise, or having the additional budget, in terms of time or tool calls, without any feedback, results in considerable increases for some of the models but not for others. We find that how much such a loop helps depends on which model plays which role, but weaker models generally do not improve stronger models.
[NLP-68] Rewrite Then Veto: Rewriter-Agnostic Edit Gating for LLM -Based Post-OCR Correction
【速读】： 该论文旨在解决大语言模型（LLM）在修复历史文献光学字符识别（OCR）文本时存在的过度修正问题：尽管LLM能够有效纠正错误的OCR输出，但其在处理原本正确的文本时也会引入新的错误。为应对这一问题，作者提出EditVeto——一种置于全上下文LLM重写之后的“否决门”机制。该方案的核心在于：将模型的重写结果与原始OCR输入进行字符级对齐，将差异划分为若干块，并通过一个0.6B参数量的小型门控网络决定每个块是否保留或回滚。关键创新点在于，基于“真实错误位置”的模拟分析表明，对每个片段独立生成修正反而比不修正更差，而对整个重写结果的编辑进行整体否决则显著提升性能。该门控机制在两个Qwen3模型生成的重写数据上训练，并在八个HIPE-OCRepair-2026测试集上冻结评估，结果显示其在所有五种重写器（包括两种未参与训练的）上均取得改进，加权字符错误率（cMER）降低0.0026–0.0044，偏好得分提升0.15–0.25（配对bootstrap检验，p≈0.0002）。在公开竞赛结果上的表现进一步验证了其有效性，虽使获胜团队的cMER略有上升（0.0050 → 0.0065），但对其他三支队伍均有改善。二次披露测试中，重新训练的门控仍能提升LoRA微调重写器的表现（cMER下降0.0011至0.0021），且多模型门控编辑的集成结果达到cMER 0.0069、偏好得分0.928，主要得益于微调与集成策略而非门控本身。该系统仅依赖公开模型与数据，最终性能仅次于唯一采用领域预训练的冠军团队（cMER 0.0050），与第二名持平（cMER 0.0071），优于其余所有已发表结果。第三次独立评估显示，未调整的门控仍可提升零样本Qwen3.8-27B的性能（cMER从0.0114降至0.0108），并提升其LoRA微调版本的偏好得分。
链接: https://arxiv.org/abs/2611.05539
作者: Jorge Ramón Fonseca Cacho
机构: University of Nevada, Las Vegas (内华达大学拉斯维加斯分校)
类目: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages, 3 figures, 12 tables. Code: this https URL models: this https URL
Abstract:LLMs can repair noisy OCR transcripts of historical documents, but they also change text that was already correct. We propose EditVeto, a veto gate placed after a full-context LLM rewrite. The rewrite is character-aligned to the OCR input, its differences are split into blocks, and a 0.6B-parameter gate keeps or reverts each block. An oracle analysis motivates the design: given the true error locations, per-span generation is worse than no correction, whereas vetoing a full rewrite’s edits helps. Trained on rewrites by two Qwen3 models and evaluated, frozen, on the eight HIPE-OCRepair-2026 test sets, the gate improves all five rewriters tested, two of them unseen in training: weighted cMER falls by 0.0026-0.0044 and the preference score rises by 0.15-0.25 (paired bootstrap, p ~ 0.0002 each). On published competition outputs it lowers cMER for three of four teams but raises it for the winning run (0.0050 to 0.0065). In a second, disclosed test look, a retrained gate still improves LoRA-tuned rewriters (cMER -0.0011 to -0.0021), and an ensemble of their gated edits reaches a cMER of 0.0069 and a preference score of 0.928, mostly through fine-tuning and pooling rather than gating. Built only on public models and data, it is behind the winner (0.0050) and tied with the second-ranked run (0.0071), both from the only team using domain pre-training, and ahead of all other published runs. A third, secondary look shows the unchanged gate improving zero-shot Qwen3.8-27B (cMER 0.0114 to 0.0108) but only the preference score of its LoRA-tuned version.
[NLP-69] When to Ask When to Answer: Strategic Clarification for Ambiguous Reference Resolution in Dialogue
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在参考消歧任务中存在过早回答（premature answering）的问题，即模型倾向于在信息不充分时直接给出确定性答案，而非通过策略性提问来获取必要澄清信息。其核心挑战在于如何使模型在对话中具备“战略澄清”（strategic clarification）能力——即在提问时能够判断哪些问题能有效缩小候选指代对象的集合，从而提升消歧效率。解决方案的关键在于提出一种轻量级的动作选择方法BACE（Belief-Aware Critic with Epistemic tracking），该方法基于冻结的预训练语言模型主干，在推理过程中引入信念感知与认知状态追踪机制，动态评估当前知识不确定性，并据此决定是否提问或回答。实验表明，BACE相比GRPO和ArCHer等基线方法，能更一致地选择最优数量的澄清问题，显著提升了参考消歧的准确性和对话效率。
链接: https://arxiv.org/abs/2611.05517
作者: Ibrahim Khebour,Nikhil Krishnaswamy
机构: Colorado State University (科罗拉多州立大学)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
备注: 22 pages, 10 figures. Code and data: this https URL
Abstract:For humans, referential disambiguation is a collaborative process involving clarification questions asked when they will meaningfully change the outcome. However, large language models (LLMs) have been shown to be biased toward premature answering, preferring definite single-turn responses over more clarification exchanges. In this paper, we study the problem of “strategic clarification” – asking clarification questions in a reference game when their answers will narrow the set of candidate referents. We construct a controlled test environment from the SIMMC 2.1 dataset that records how many questions each ambiguous reference needs to be successfully resolved, and examine how baseline LLM behavior and common post-training methods ask clarifying questions vs. attempt to answer the referent. We also propose BACE (Belief-Aware Critic with Epistemic tracking), a lightweight action chooser method on top of a frozen LM backbone, and show that BACE more consistently chooses the optimal number of clarifying questions than methods such as GRPO and ArCHer.
[NLP-70] SlackBench: Benchmarking Agents on Collaborative Projects Grounded in Real Code Repositories
【速读】： 该论文旨在解决现有评估基准在模拟真实知识工作场景方面的不足，特别是针对生成式AI代理（AI Agents）在复杂协作环境中执行任务的能力评估。传统基准多依赖合成数据或孤立任务，无法反映实际科研与工程协作中信息分散于对话、代码及动态决策中的多源异构特性。为应对这一挑战，论文提出SlackBench，一个基于真实学术研究项目隐私保护复现的基准测试平台，其核心在于构建包含多参与者对话、持续演化的项目决策以及与GitHub仓库和Google Docs相一致的代码变更的可复现环境。解决方案的关键在于设计一系列需联合推理通信内容与代码状态的任务，涵盖跨源智能搜索、对不可回答或前提错误问题的合理回避、目标导向摘要生成及长对话上下文推断等，要求代理具备识别关键信息、遍历对话线程、检查代码仓库状态并整合多源证据的能力。实验表明，当前前沿模型在该基准上的最佳表现约为80%，多数系统低于66%，揭示了现有代理在真实协作场景下仍存在显著提升空间。
链接: https://arxiv.org/abs/2611.05512
作者: Mingxuan Li,Huanzhi Mao,Ruitao Zheng,Jarod Stanbury James,Yunze Li,Natanya Anderson,Joseph E. Gonzalez,Alex Dimakis
机构: 未知
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注:
Abstract:AI Agents are becoming digital coworkers that can browse the web, write code, run experiments, and assist humans over extended horizons. Yet existing benchmarks fall short of real knowledge work: many rely on synthetic data or isolated tasks, missing the complex, multi-party dynamics of collaborative projects, where critical information is scattered across conversations, code, and evolving decisions. Even simple workplace questions may require locating relevant channels and direct messages, recovering context from long discussions, inspecting repository state, reconciling stale or conflicting evidence, and synthesizing across sources. We introduce SlackBench, a benchmark for evaluating agents on realistic research and engineering workflows that require joint reasoning over workplace communication and evolving repository state. Grounded in privacy-preserving reenactments of real academic research projects, SlackBench provides reproducible environments with realistic multi-party dialogue, evolving decisions, and corresponding code changes grounded in GitHub repositories and Google Docs. It contains 140 queries spanning heterogeneous-source agentic search, abstention on unanswerable or false-premise questions, goal-directed summarization, and circumstance inference over long conversations. These tasks are not reducible to static corpus question answering: agents must determine which conversations matter, traverse threads, inspect repository state, and synthesize evidence across communication and code artifacts. Across 33 frontier models and agent harnesses, the best system scores around 80% and most agents score below 66%, highlighting substantial headroom for improvement. SlackBench thus provides an extensible framework for evaluating agents in collaborative work environments. The SlackBench dataset is available at this https URL . Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL) Cite as: arXiv:2611.05512 [cs.AI] (or arXiv:2611.05512v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.05512 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[NLP-71] Unlocking Turing Tensor Cores: A Measured Throughput Account of From-Scratch Pre-training on a Single 22G RTX 2080 Ti
【速读】： 该论文旨在解决NVIDIA RTX 2080 Ti（TU102）显卡在硬件层面虽配备第二代张量核心（Tensor Cores），但厂商通过驱动层限制了半精度（FP16）矩阵乘法的计算性能，导致实际可用张量吞吐量远低于硬件峰值的问题。其解决方案的关键在于通过一个内核模块（kernel module）在NVIDIA驱动绑定设备前，修改FP16速度选择寄存器（SS0=0x88888888, SS1=0x8）并释放ACR互斥锁（ACR mutex），从而解除驱动层面的性能限制。实验表明，该方法使cuBLAS FP16矩阵乘法吞吐量从55.7提升至91.4 TFLOPS（约1.6倍），在相同预训练协议下，不同规模的中文编码器-解码器模型在训练吞吐量上均获得显著提升：小模型（75.7M参数）稳定吞吐量提升11.8%（从6,480增至7,246 tok/s），壁时减少9.0%；超小模型（tiny tier）提升4.7%，壁时节省3.8%；更小模型（pico-micro）提升不足0.3%。关键发现为：解锁增益与矩阵规模正相关，对小模型及以上层级呈现明显提升，且该增益为纯吞吐量提升，不影响训练动态与最终模型质量（比特级一致或误差小于10⁻³）。
链接: https://arxiv.org/abs/2611.05419
作者: Dezhou Shen
机构: 独立研究者(Independent Researcher)
类目: Hardware Architecture (cs.AR); Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF)
备注: 7 pages, 4 tables
Abstract:The RTX 2080 Ti (TU102) ships full second-generation Tensor Cores in hardware, yet the vendor throttles FP16 matrix multiplication at the driver level, leaving usable tensor throughput well below the hardware peak. turing-unlock lifts this restriction with a kernel module that, before the nvidia driver binds the device, writes the FP16 speed selectors (SS0=0x88888888, SS1=0x8) and releases the ACR mutex; published measurements report cuBLAS FP16 matmul throughput rising from 55.7 to 91.4 TFLOPS (about 1.6x). On the same 22G 2080 Ti and under the same from-scratch pre-training protocol as the companion memory ledger (five tiers of Chinese encoder-decoder models, 2,451 steps per tier, 16,384 tokens per step, seed 42), we obtain a before/after throughput account: the small tier (75.7M parameters, 512-wide, 8+8 layers, activation checkpointing on) improves in steady-state throughput from 6,480 to 7,246 tok/s (+11.8%) with wall time cut from 1.77h to 1.61h (9.0% saved); the tiny tier goes from 15,150 to 15,866 tok/s (+4.7%) with wall time savings of 3.8%; the three smaller tiers (pico-micro) gain at most +0.3%; per-tier peak memory stays bit-identical, and final quality metrics are bit-identical or consistent within 10^-3. We conclude that the unlock gain is a function of matrix size, rising monotonically with tier – imperceptible for pico-micro, visible from tiny upward, and cashed out as a double-digit percentage at small – and that it is a pure throughput gain that leaves training dynamics untouched.
[NLP-72] Challenge-Aware Knowledge Recruitment with Meta-Agent Supervision for Medical Multi-Agent Reasoning
【速读】： 该论文旨在解决大语言模型（Large Language Models, LLMs）在医学问答（Medical QA）中面对复杂推理任务时性能显著下降的问题，尤其针对现有基于多智能体辩论（multi-agent debate）的方法在处理高难度任务时难以超越单一模型提示（strong single-model prompting）的瓶颈。其核心挑战在于：当前系统在选择专家和检索知识时仅依赖初始问题信息，而忽视了推理过程中动态浮现的知识或能力缺口。为此，论文提出“挑战感知的自适应招募”（challenge-aware adaptive recruitment）机制，即引入一个外部挑战者（CHALLENGER）实时监控推理过程，识别出具体的事实性或专业性知识缺口，并据此触发针对性的知识检索或在讨论中引入特定领域的专家。实验结果表明，在MEDAGENTSBENCH硬子集上，预先检索几乎无益，而中间阶段的挑战感知检索显著提升性能；在多专家基线模型上，该方法在9个数据集中的8个实现准确率提升，相较仅使用挑战者的方案提高6个百分点。消融实验证明，检索时机与基于缺口的查询信号均具关键作用，且在真实网络搜索场景下，时间优势仍可保持。对于专家招募，自触发机制不可靠，而由挑战者触发的专家招募将GPT-4.1-mini MedQA准确率从46%提升至54%，并在所有泛化测试数据集上表现正向。然而，简单叠加检索与专家招募反而损害性能，凸显了协调多种自适应干预路径的重要性。研究结论表明，在多智能体医学推理系统设计中，“何时”与“为何”调用外部资源是决定性能的关键决策因素。
链接: https://arxiv.org/abs/2611.05393
作者: Riccardo Riccio,Haitz Sáez de Ocáriz Borde,Pietro Liò
机构: University of Cambridge(剑桥大学)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
备注:
Abstract:Large language models perform strongly on standard medical QA but degrade on harder reasoning tasks, where multi-agent debate often fails to beat strong single-model prompting. We argue that a key limitation is that existing systems select experts and retrieve knowledge from the question alone, rather than from gaps that emerge during reasoning. We propose challenge-aware adaptive recruitment: an external CHALLENGER inspects the evolving reasoning trace and identifies specific factual or expertise gaps. These signals trigger targeted knowledge retrieval or mid-discussion specialist recruitment. On the MEDAGENTSBENCH hard subset (862 questions across nine biomedical datasets), upfront retrieval provides little benefit, while challenge-aware in-between retrieval performs better; on a multi-specialist backbone it improves accuracy on 8 of 9 datasets and adds 6 percentage points over Challenger Only. Ablations indicate benefits from both retrieval timing and gap-specific query signals, and a live web-search experiment shows that the timing advantage persists across retrieval backends. For specialist recruitment, self-triggered adaptation is unreliable, whereas Challenger-triggered recruitment raises GPT-4.1-mini MedQA accuracy from 46% to 54%, with positive direction on all tested generalization datasets. Naively combining retrieval and specialist recruitment degrades performance, highlighting the need to coordinate adaptive intervention channels. These results show that when and why external resources are recruited is a key design choice in multi-agent medical reasoning.
[NLP-73] Agent Harness Awareness: Turning Verified Telemetry into Operational Control for Tool-Using Agents
【速读】： 该论文旨在解决智能体在执行任务时存在的“操作性盲视”（operational blindness）问题，即工具描述虽能说明某动作的功能，却无法反映该动作在当前时刻是否健康或适宜执行。由于决策模型缺乏对运行时状态、监控数据、历史失败等关键信息的访问，导致其难以做出最优选择。其解决方案的核心在于提出“代理执行框架感知”（Agent Harness Awareness）这一信任分离的设计架构：由宿主在执行关键动作前提供一个小型、实时、经过验证的状态快照，同时保留测量、执行、评估与最终调度的控制权。实验结果表明，在多模型（GPT-5.6 Sol、Claude Sonnet 5、Gemini 3.7 Flash）和多种场景下的对比测试中，引入经验证的操作状态可使平均可观测策略遗憾降低96.9%，并将单次调度完成率从69.4%提升至95.8%；在端到端真实网络环境（含TCP/HTTP执行、OpenTelemetry追踪及故障注入）中，平均遗憾降至0毫秒（零额外可观测运行成本），且在9次故障案例中成功完成动作，远超无状态信息的对照组（5/9）。这证实了该架构在工具与网络层面显著提升了预调度决策质量，而运行时状态、硬件负载、经济性、重试自愈机制及异构选择等仍为该架构未来可拓展的实证方向。
链接: https://arxiv.org/abs/2611.05391
作者: Manu Agrawal,Shrey Nagpal
机构: Independent Researchers(独立研究员)
类目: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
备注: 9 pages, 3 figures
Abstract:Tool descriptions tell an agent what an action can do, but they do not necessarily reveal whether that action is healthy or suitable at the moment of use. Execution harnesses often possess this information through monitors, traces, and recent failures while the model choosing the next action does not. We call this gap operational blindness. We introduce Agent Harness Awareness, a trust-separated design in which the host supplies a small, fresh, verified state snapshot before a consequential action while retaining measurement, enforcement, evaluation, and final dispatch authority. We test its tools-and-network instance with ToolRoute, a controlled isolation benchmark that holds the task and candidate semantics constant and varies the operational information available to the model. In a 216-decision ablation across GPT-5.6 Sol, Claude Sonnet 5, and Gemini 3.7 Flash, verified tool-health state reduced mean observable policy regret by 96.9% against a same-shape neutral control (4,252 ms to 131 ms) and raised single-dispatch completion from 69.4% to 95.8%. In a separate 27-decision end-to-end replication with live local TCP/HTTP execution, OpenTelemetry client spans, and Toxiproxy-injected faults, verified state reduced mean regret from 8,956 ms to 0 ms (zero excess observable operational cost) and completed the live action in 9/9 model-by-fault cases versus 5/9 for the neutral control. These studies establish the tools-and-network instance of Agent Harness Awareness: selected host telemetry materially improves pre-dispatch operational choice. Runtime state, hardware pressure, economics, retry-aware recovery, and heterogeneous selection remain open empirical extensions of the same architecture.
[NLP-74] AI-Driven SaaS Platform Integration: AI Express for Unified Generative AI Services with Pay-As-You-Go Credit Systems and Scalable Microservice Architecture
【速读】： 该论文旨在解决生成式AI（Generative AI）当前服务碎片化的问题，即文本生成、图像合成与代码辅助等功能需分别订阅，导致使用成本高昂且工作流频繁中断。其解决方案的关键在于推出一款统一的SaaS平台——AI Express，将多种AI能力整合为按使用量计费的一体化服务，通过单一接口提供包括核心代码审查（Core Code Reviewer）、文本转网站生成、脚本生成、语法检查、数据分析及水印去除在内的多项功能，用户无需具备技术背景即可便捷使用，从而实现高效、低成本且无缝衔接的AI应用体验。
链接: https://arxiv.org/abs/2611.05389
作者: Deepit Sapru
机构: University of Illinois Urbana-Champaign (伊利诺伊大学厄本那-香槟分校)
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:Generative AI is currently fragmented, with separate subscriptions required for text generation, image synthesis, and code assistance, leading to high costs and workflow disruptions. AI Express is a SaaS platform that consolidates these capabilities into a unified pay-as-you-go service. The platform provides a range of AI tools, including a Core Code Reviewer that evaluates GitHub repositories, a Text-to-Website generator that converts text into fully functional websites, script generation, grammar checking, data analysis, and watermark removal, all accessible through a single interface without requiring technical expertise.
[NLP-75] SRRM: A Retrieval-Reconstruction Benchmark for Evaluating Long-Context Memory in Small Language Models
【速读】： 该论文旨在解决小语言模型（Small Language Models, SLMs）在长上下文场景下信息保持能力不足且缺乏系统评估方法的问题。现有基准主要聚焦于检索任务，难以全面反映模型对长上下文记忆的完整保留能力。为此，本文提出一种轻量级、可完全复现的基准测试框架——目标检索与全局信息重建联合评估基准（Simple Retrieval-Reconstruction Memory, SRRM），并引入基准层面的度量指标“记忆保持率”（Memory Preservation Ratio, MPR），用于量化检索与重建性能之间的关联性。实验基于合成数据集，在三个Qwen2.5系列模型（0.5B、1.5B和3B参数量）上评估了10至100个事实陈述的上下文长度下的表现。结果表明，随着上下文长度增加，重建性能下降速度远快于检索准确率，说明仅依赖检索任务无法充分表征长上下文记忆保持能力。其中，Qwen2.5-1.5B在检索准确率（86.7%）、重建性能（45.7%）及MPR（0.53）上均优于更大的Qwen2.5-3B模型（分别为72.2%、18.5%和0.26），揭示参数量增大并不必然提升长上下文记忆保持能力。研究证明，SRRM能够系统化评估长上下文记忆，并揭示传统仅基于检索的评估框架所掩盖的互补性记忆行为。
链接: https://arxiv.org/abs/2611.05386
作者: Christopher Alberto Hamonangan
机构: 未知
类目: Computation and Language (cs.CL)
备注: 37 pages, 20 tables, 13 figures
Abstract:Small Language Models (SLMs) are increasingly deployed on consumer-grade hardware, yet their ability to preserve information across progressively longer contexts remains insufficiently understood. Existing benchmarks primarily evaluate retrieval-oriented tasks, providing limited insight into whether successful retrieval reflects comprehensive long-context memory preservation. To address this limitation, this paper introduces Simple Retrieval-Reconstruction Memory (SRRM), a lightweight and fully reproducible benchmark that jointly evaluates Targeted Retrieval and Global Information Reconstruction. This study also proposes the Memory Preservation Ratio (MPR), a benchmark-level metric that quantifies the relationship between retrieval and reconstruction performance. SRRM was evaluated using a synthetic dataset on three Qwen2.5 models (0.5B, 1.5B, and 3B parameters) across context loads ranging from 10 to 100 factual statements. Experimental results show that reconstruction performance deteriorated substantially faster than retrieval accuracy as contextual information increased, indicating that retrieval alone is insufficient to characterize long-context memory preservation. Among the evaluated models, Qwen2.5-1.5B achieved the highest retrieval accuracy (86.7%), reconstruction performance (45.7%), and MPR (0.53), outperforming the larger Qwen2.5-3B model (72.2%, 18.5%, and 0.26, respectively). These findings demonstrate that increasing parameter count does not necessarily improve long-context memory preservation. Overall, SRRM provides a reproducible and extensible benchmark for systematically evaluating long-context memory while revealing complementary memory behaviors that remain hidden under retrieval-only evaluation frameworks.
[NLP-76] Same Origin Different Verdicts: Text Production Conditions Confound Machine-Generated Text Detection in Russian
【速读】： 该论文旨在解决生成式文本检测中一个核心矛盾：文本的生成来源属于生产过程属性，而非可直接观测的文本特征，而自动化检测系统却需基于文本本身做出判断。其关键问题在于，不同检测方法在相同文本条件下得出的机器生成性评分存在显著不一致，这种不一致性可能源于数据收集、重复设计、格式化及预处理等非语义因素的影响。研究通过构建俄语语料库（1882份文档），包含由四种生成渠道（对应不同模型族）在3种任务模式下完成的45个任务的机器生成部分，以及经源归属层级标注的人类撰写部分，以控制生成来源不变，仅改变任务表述方式来考察外部条件对检测结果的影响。研究采用四种操作化指标进行测量：固定权重的可解释指数、训练好的分类器、基于开放模型的零样本度量和裁判模型。结果显示，四种方法间存在明显分歧——任务模式切换导致所有指标得分变化，其中三种呈现负向偏移，唯裁判模型呈正向偏移，且无一指标的置信区间覆盖零值。尽管在程序1结果后调整了比较方案与后续方法的具体设定，并修正了语料缺陷与嵌套验证偏差，分歧依然存在。此外，仅依赖源标识符的基线在13组划分中达到中位数AUROC=1.0000，且去除强调标记即可将原本确定的机器生成判别逆转为人类生成，而全文未作任何字符修改。这表明当前检测分数本质上是特定测量系统与语料特性共同作用的结果，反映的是数据采集、重复性、格式化及预处理决策，而非纯粹的语言风格或生成机制差异。因此，该研究揭示了现有生成式文本检测方法的脆弱性与不可靠性，其核心解决方案在于强调检测结果的高度情境依赖性，提示应重新审视检测系统的有效性边界与评估范式。
链接: https://arxiv.org/abs/2611.05385
作者: Stanislav Kirichenko
机构: Independent Researcher
类目: Computation and Language (cs.CL)
备注: 65 pages, 6 figures, 24 tables. English translation of the frozen Russian manuscript; the Russian text is the source of record and is available with all data, code and run manifests at this https URL (v1.0.1)
Abstract:Automated detection of machine-generated text is applied where the verdict carries a cost for the author, yet origin is a property of the production process, not an observable property of the text. The conditions under which a text is produced push it in the same direction in which a generative model is presumed to push it. This study measures the contribution of these conditions while origin is held fixed. The Russian-language corpus of 1882 documents comprises a machine part, generated by four generation channels (each backed by its own model family) across 45 tasks in three task modes, and an archival human part annotated for source regimentation level before the texts were read. The primary contrast is computed within the machine corpus: both halves of a pair are generated by the same model for the same task and differ only in task wording. Machine-likeness is measured by four operationalisations: an interpretable index with weights fixed before data collection, a trained classifier, a zero-shot metric on an open model, a judge model. The main result is disagreement between procedures: changing the task mode shifted the scores of all four, but three gave a negative shift, the judge a positive one, and none of the four intervals covers zero. The comparison plan and the specifications of procedures 2-4 were fixed after the result of procedure 1, before procedures 2-4 ran; the disagreement survived correction of a corpus defect and of the nested validation. A baseline on the source identifier alone reaches a median AUROC of 1.0000 on the thirteen group splits where it is defined, and removing emphasis markers flipped a confident machine verdict to human without changing the full text by a single character. The scores are properties of a specific measurement system and corpus: they reflect data collection, repetition, formatting, and preprocessing decisions as well as linguistic style.
[NLP-77] Human-like cross-language generalisation in deep neural speaker embeddings and its acoustic foundations INTERSPEECH2026
【速读】： 该论文旨在解决深度神经网络中说话人嵌入（speaker embedding）在跨语言情境下的泛化能力问题，特别是其与人类感知及语音声学结构之间的一致性。研究通过对比粤语-英语双语语音在同语言与跨语言条件下的表现，利用18个先进的模型提取嵌入表示，并结合线性混合模型与表征相似性分析（representational similarity analysis），将模型生成的相似性度量与人类评分及声学特征进行比较。研究发现，说话人嵌入在跨语言情境下表现出良好的泛化能力，其模式与人类行为高度一致，能够准确区分说话人身份并对语言切换敏感。这种泛化能力主要依赖于跨语言共享的声学线索，并对特定语言的声学特征赋予差异化权重，呈现出与人类认知机制相类似的处理模式。因此，该研究的关键在于揭示了机器学习模型与人类在跨语言语音相似性判断上存在显著对齐，为生成式 AI 在跨语言语音识别中的可解释性与可靠性提供了有力证据。
链接: https://arxiv.org/abs/2611.05384
作者: Tianze Xu,Xiyang Li,Xiaoming Jiang,Volker Dellwo
机构: University of Zurich (苏黎世大学); Shanghai International Studies University (上海外国语大学)
类目: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
备注: This manuscript has been accepted for presentation at Interspeech 2026. Code is available at this https URL . The archived release is available on Zenodo at this https URL
Abstract:Generalisability is a central question for deep neural networks, including the cross-language generalisation of speaker embeddings. This study examined whether deep neural speaker embeddings generalise across languages and how such generalisation relates to human perception and acoustic structure. Using Cantonese–English bilingual speech presented in within- and cross-language conditions, we derived similarity from 18 state-of-the-art models and compared it with human ratings and acoustic features using linear mixed models and representational similarity analysis. Embeddings exhibited cross-language generalisation that broadly paralleled human behaviour, reflecting accurate identity discrimination and sensitivity to language switching. This generalisation relied on largely shared acoustic cues with language-sensitive weighting, mirroring human patterns. Together, these findings provide converging evidence for human–machine alignment in cross-language voice similarity processing.
[NLP-78] CCDWIKI: Decoupling Case Formulation from Response Generation in Longitudinal CBT Dialogue
【速读】： 该论文旨在解决认知行为疗法（Cognitive Behavioral Therapy, CBT）中对话系统在长期治疗过程中难以维持认知概念化一致性的问题。传统方法依赖于检索或静态档案来追踪会话历史，但缺乏对信念状态（epistemic status）的表征能力，导致临时假设被过早固化为确定性事实，从而误导后续干预策略。其解决方案的关键在于提出CCDWiki——一种即插即用的顾问系统，将纵向治疗过程建模为基于证据的假设评判机制。该系统通过可变条目形式编码每个临床假设，并动态关联支持与反驳证据；在多轮会话中，累积数据可增强已有假设、保留临时假设并剔除矛盾假设，从而实现持续的、可追溯的认知概念化演进。最终生成的结构化表述可转化为适用于未修改的治疗型语言模型的简洁指导，实验表明该方法显著提升了概念化与干预质量，且无需对基础模型进行重新训练。
链接: https://arxiv.org/abs/2611.05383
作者: Mingwen Zhang,Ruiyang Song,Qinying Wang,Yuchen Wang,Changsheng Ma,Minqiang Yang
机构: 未知
类目: Computation and Language (cs.CL)
备注:
Abstract:Cognitive Behavioral Therapy (CBT) demands more than empathic validation, requiring instead a structured formulation of the interplay among situations, automatic thoughts, emotions, and behaviors. Therapists formalize this interplay through a Cognitive Conceptualization Diagram, where maintaining longitudinal coherence necessitates continuous evidential adjudication to separate tentative hypotheses from established findings. Current dialogue systems track conversational history via retrieval or static profiles but lack the representational capacity to encode epistemic status. Consequently, provisional interpretations are often prematurely reified as established facts, inadvertently biasing downstream interventions. Addressing this limitation, CCDWiki is introduced as a plug-and-play advisor conceptualizing longitudinal therapy as evidence-grounded hypothesis adjudication. The approach encodes each clinical hypothesis as a mutable entry bound to corroborating and contradicting evidence. Across sessions, accumulating data strengthens supported entries, preserves provisional ones, and retracts contradicted assumptions. The resulting formulation distills into concise guidance for unmodified therapeutic language models. Experiments demonstrate substantial improvements in both conceptualization and intervention quality without requiring backbone retraining.
[NLP-79] DS@GT ARC at CheckThat! 2026: A Verdict-Anchored Pipeline for Generating Fact-Checking Articles
【速读】： 该论文旨在解决虚假信息传播速度远超人工事实核查效率的问题，提出自动化生成事实核查文章的需求。其核心挑战在于如何从一个待验证声明、原始核查结论及一组爬取的来源文档中，自动生成结构完整、内容准确且具备可读性的事实核查报告。解决方案的关键在于构建一个四阶段流水线：首先对噪声较大的证据进行清洗；其次通过检索与聚类技术筛选并组织相关来源文档；然后基于原始核查结论作为条件，利用Llama-3.1-8B-Instruct模型生成完整的核查文章；最后通过确定性后处理方法恢复文中内嵌引用（inline citations）。该方法在CLEF 2026 CheckThat! Lab Task 3中取得了0.428的平均得分，排名第三，显著优于0.272的基线水平，验证了其有效性与实用性。
链接: https://arxiv.org/abs/2611.05382
作者: Hoang Thanh Thanh Truong,Shuyu Tian
机构: Georgia Institute of Technology(佐治亚理工学院)
类目: Computation and Language (cs.CL)
备注: CLEF 2026 Working Notes, 21 - 24 September 2026, Jena, Germany
Abstract:Misinformation spreads faster than professional fact-checkers can manually verify it and write articles explaining their verdicts. This creates a need to automate the production of fact-checking articles. This paper details work done by the DS@GT ARC team for CLEF 2026 CheckThat! Lab Task 3 Fact-Checking Article Generation. We seek to generate a full fact-checking article from a claim, its original verdict, and a set of scraped source documents. Our team built a four-stage pipeline that cleans noisy evidence, retrieves and clusters relevant sources, generates the article by conditioning \textttLlama-3.1-8B-Instruct on the original fact-checker verdict, and recovers inline citations through deterministic post-processing. We achieved a mean score of 0.428 and ranked 3rd out of 9 teams for the CLEF 2026 CheckThat! Lab Task 3, compared to a baseline mean score of 0.272. Our code is available on GitHub at this https URL.
[NLP-80] Discourse Analysis of Legislative Bill Discussions
【速读】： 该论文旨在解决立法听证会中话语结构解析不足的问题，以深化对民主机构内部工作流程与权力关系的理解。现有研究多局限于对听证会文本的表面分析，未能充分揭示其内在的人类可识别的逻辑与结构。为此，本文提出一种端到端的语篇结构解析框架，核心在于构建一个针对立法委员会听证会的修辞结构语法（rhetorical structure grammar），并基于此实现对语篇段落层级结构的自动标注。关键创新点在于：通过引入基于语法规则的输出掩码（grammar-based output masking）与话语标签平滑（utterance tag smoothing）机制，显著提升了分类器性能，即使在使用相对简单的逻辑回归模型时，仍实现了52.8%的F1分数，优于更复杂的模型。这一方法为提升立法话语分析的准确性与民主问责机制提供了可扩展的技术路径。
链接: https://arxiv.org/abs/2611.05381
作者: Isaac Rudnick,Nipun Das,Foaad Khosmood
机构: California Polytechnic State University (加州理工州立大学)
类目: Computation and Language (cs.CL)
备注: 10 pages, 4 figures
Abstract:Legislative hearings are distinct communication events with their own human recognizable structure and logic, yet they have been analyzed as distinct phenomena beyond their transcripts. Understanding legislative hearings at a deeper level is a major step in analyzing the work and power relationships within democratic bodies. In this work, we present a novel end-to-end pipeline for parsing rhetorical structures from legislative committee hearings with the goal of improving legislative discourse analysis and democratic accountability. We also propose a rhetorical structure grammar and find that 55.21% of California state bill discussions in our dataset can be parsed with this grammar. We evaluate our pipeline with two utterance section classifiers on a random subset of over 2,000 annotated utterances corresponding to 100 hearings. We find that a logistic regression-based classifier with grammar-based output masking and utterance tag smoothing outperforms more complex models, achieving an F1 score of 52.8% against our ground truth discourse section labels.
[NLP-81] Hallucination Is Linearly Decodable from Mid-Layer Hidden States in Quantized LLM s
【速读】： 该论文旨在解决大语言模型（Large Language Model, LLM）中是否存在可线性分离的“真实性信号”（truthfulness signal）及其在模型网络深度中最强出现位置的问题。研究聚焦于三个70亿至80亿参数量级的开源指令微调模型（Llama-3.1-8B、Mistral-7B、Qwen2.5-7B），在4位NF4量化下，通过提取各层隐藏状态，并在四个幻觉检测基准（TruthfulQA、HaluEval-QA、FEVER及一个受控合成数据集）上对比多种检测方法。其解决方案的关键在于：在模型中等深度层（如Llama/Mistral的第13–18层、Qwen的第19–25层）存在显著且近似线性的真实性信号，单层线性探测器即可实现0.904–1.000的AUROC性能，远优于基于采样的检测方法（最高仅0.541 AUROC）；同时发现，首层注意力熵在知识依赖型任务中提供无额外推理开销的互补信号（HaluEval-QA上达0.866–0.941 AUROC）。该结果表明，当前采样类方法性能受限源于其评估范式与信息访问机制之间的结构性不匹配，而非方法本身固有缺陷。
链接: https://arxiv.org/abs/2606.02628
作者: Aizierjiang Aiersilan
机构: University of Macau(澳门大学)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Emerging Technologies (cs.ET); Neural and Evolutionary Computing (cs.NE)
备注:
Abstract:We investigate whether open-source LLMs encode a linearly separable truthfulness signal in their hidden states, and at which network depth this signal is strongest. Across three 7 B-- 8 B instruction-tuned models (Llama-3.1-8B, Mistral-7B, Qwen2.5-7B) loaded in 4 -bit NF4 quantization, we extract per-layer hidden states on four hallucination benchmarks (TruthfulQA, HaluEval-QA, FEVER, and a controlled synthetic set) and compare four detection approaches: linear and MLP probes, INSIDE EigenScore, self-consistency, and attention entropy. A linear probe on a single mid-network layer achieves 0.904 – 1.000 AUROC on held-out splits, while sampling-based detectors do not exceed 0.541 AUROC under the same protocol. The truthfulness signal is approximately linear: MLP probes rarely surpass linear probes by more than 0.01 AUROC. Peak probing layers fall in a consistent band across model families on natural-language benchmarks – blocks~ 13 – 18 of~ 32 for Llama and Mistral, and blocks~ 19 – 25 of~ 28 for Qwen. First-block attention entropy provides a complementary signal in knowledge-grounded settings ( 0.866 – 0.941 AUROC on HaluEval-QA) at no additional inference cost. The low discriminability of sampling methods under this protocol reflects a structural mismatch between paired-label evaluation and the information these methods access, rather than an inherent limitation of those methods. Code and data are released for full reproducibility on a single 8 ,GB GPU.
信息检索
[IR-0] Balancing hierarchical representation and quantization: A performance trade-off in text embeddings
链接: https://arxiv.org/abs/2611.06859
作者: Kai Ponel,Marco Polignano,Erasmo Purificato,Aylin Julia Temizkan,Marco Levantesi,Giovanni Semeraro,Ernesto William De Luca
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 29 pages. Authors’ version of the open access article published in Information Sciences (CC BY 4.0). This version corrects the second affiliation of Marco Polignano (Otto von Guericke University Magdeburg); the published version lists the European Commission, Joint Research Centre (JRC) in error. Marco Polignano is not affiliated with the JRC. No other content has been changed
Abstract:Hierarchical text embedding models offer variable-length representations, while quantization reduces numerical precision; both aim to improve computational efficiency. This work is an empirical benchmark study of how these two orthogonal, already-established techniques interact. We evaluate 38 tasks from the Massive Text Embedding Benchmark (MTEB) using four transformer-based embedding models, covering 1D and 2D hierarchical representations combined with 32-bit, 8-bit, and binary quantization, for 168 configurations per task. We report each efficiency dimension separately: theoretical per-vector storage and end-to-end evaluation time. Under a Friedman test with Nemenyi post-hoc analysis, int8 quantization is statistically indistinguishable from float32 for pair classification and re-ranking and costs at most 1.5 percentage points overall, whereas binary quantization is significantly worse across every task category. Halving the embedding dimension once is not statistically detectable above m=16 ; two successive halvings always are. Combining an 8x dimensionality reduction with int8 yields a 32x reduction in per-vector storage at 94.4% of baseline effectiveness. Because evaluation times were measured once, we treat them as indicative: we find no evidence that quantization speeds up evaluation, whereas sampling from intermediate layers shortens it substantially. Retrieval degrades far faster than every other category, so we report findings per task category.
[IR-1] WISE: An Agent ic Framework for Whole-page Multimodal Search Experience Attribution
链接: https://arxiv.org/abs/2611.06777
作者: Wenyi Xu,Renzhe Zhou,Songyang Li,Feiran Zhu,Ruoqi Zhao,Chenglei Dai,Xuanping Li,Kaiqiao Zhan,Jingwei Zhuo
类目: Information Retrieval (cs.IR)
备注:
Abstract:Algorithm-driven services, including search, recommendation, and advertising, rely on multi-stage model pipelines. While research prioritizes model optimization, post-deployment monitoring focuses on operational reliability. Yet operationally healthy systems may still produce poor-quality results, exposing a gap in existing evaluation and monitoring practices. However, a systematic problem formulation and benchmark for this task remain lacking. To bridge this gap, we formulate System Quality Diagnosis (SQD) for algorithm-driven services, a new task that extends quality evaluation beyond identifying poor outcomes to attributing their causes across system components. Focusing on industrial search, we propose WISE, an agentic framework for end-to-end quality diagnosis that discovers user-facing issues and traces their origins to underlying algorithmic components. To improve diagnostic reliability, we introduce ReTrace for constructing evidence-consistent trajectory supervision and Adaptive Harness-Policy Optimization (AHPO) for failure-guided optimization of the agent harness and model policy. To enable systematic evaluation, we construct WISE-BENCH, a real-world multimodal search benchmark that combines user-facing evidence, internal pipeline traces, and expert diagnosis annotations. Extensive experiments show consistent gains across model scales, with WISE achieving 75.20% end-to-end accuracy on Qwen3.5-35B-A3B. In production, WISE supports quality auditing, online quality assurance, and model optimization, reducing diagnosis costs and improving downstream relevance models.
[IR-2] Alternatives to Shallow Autoencoders for Collaborative Filtering
链接: https://arxiv.org/abs/2611.06610
作者: Mario Mallea,Angela Nebot,Francisco Mugica
类目: Machine Learning (cs.LG); Information Retrieval (cs.IR)
备注:
Abstract:Collaborative filtering (CF) is a cornerstone of recommender systems and plays a relevant role in many modern applications. CF uses user-item interaction data to discover future preferences. Although deep learning models have shown promise in CF, Embarrassingly Shallow Autoencoders for Sparse Data (EASE) has gained attention for its outstanding ranking accuracy provided by its closed-form solution. EASE relies primarily on relationships among items to fit a full-rank high-dimensional linear mapping. We hypothesize that this design limits its capacity to capture similar but not equivalent fine-grained user relationships, consequently limiting its recommendation accuracy. This paper introduces an alternative formulation based on EASE that takes advantage of user-user information. Furthermore, we propose a hybrid model that combines both user-user and item-item adjacency distributions. Our experiments reveal that the proposed models outperform EASE on well-known recommendation benchmarks, highlighting the significance of including the user alternative in shallow autoencoder studies for CF. Subjects: Machine Learning (cs.LG); Information Retrieval (cs.IR) Cite as: arXiv:2611.06610 [cs.LG] (or arXiv:2611.06610v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.06610 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Related DOI: https://doi.org/10.1007/978-3-032-02215-8_10 Focus to learn more DOI(s) linking to related resources
[IR-3] Deliberate Latent Reasoning for Sequential Recommendation
链接: https://arxiv.org/abs/2611.06521
作者: Haobo Zhang,Qiannan Zhu,Yutao Zhu,Zhicheng Dou
类目: Information Retrieval (cs.IR)
备注:
Abstract:Sequential recommenders are widely applied in real-world scenarios to model user preferences and behavior patterns based on user interactions. Recent advances in sequential recommendation enhance transformer-based models with latent reasoning to refine user representations by autoregressively generating intermediate reasoning vectors. However, exploring the optimal reasoning depth for diverse user histories remains challenging: existing methods typically employ a fixed number of reasoning steps for all users. This ``one-size-fits-all’’ strategy often leads to suboptimal representation learning due to over-thinking in simple histories or under-thinking in complex ones. To address this issue, we propose DESRec, a deliberate latent reasoning framework that performs personalized latent reasoning to learn more informative reasoning states and enhance user modeling for sequential recommendation. Specifically, it designs a reasoning step selection module that identifies the optimal reasoning step for each user sequence during the training phase. Through a self-distillation mechanism, it compresses this most informative state into the final reasoning step, which serves as the final user representation for item ranking during inference. Moreover, it introduces a progressive preference extraction mechanism for personalized reasoning, which dynamically segments a user’s history into a variable number of periods and summarizes each into a latent milestone representing the user’s preference for that period. It will be used as the input of the subsequent reasoning. The number of milestones varies with the users’ history, enabling a personalized reasoning over users’ evolving interests. Extensive experimental results on six public datasets and four backbone recommenders demonstrate the superiority of DESRec over eight baseline models.
[IR-4] Signed but Steerable: Measuring and Closing the Text-Vector Integrity Gap in Offline Retrieval-Augmented Generation
链接: https://arxiv.org/abs/2611.06498
作者: Georgios Feretzakis,Alexandros Papaspyridis
类目: Cryptography and Security (cs.CR); Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: 20 pages, 2 figures, 7 tables. Code and data: this https URL (DOI https://doi.org/10.5281/zenodo.23948839 )
Abstract:Offline retrieval-augmented generation (RAG) on edge devices is updated through signed knowledge packs that carry text and precomputed embedding vectors. A signature proves who published a pack, not that its vectors encode its text. We study this text-vector integrity gap. A verifier that recomputes vectors and accepts them within a cosine tolerance concedes a predictable share of queries to an attacker who controls the vectors: the hijack rate equals the distribution of angular retrieval margins at the tolerance (the margin law). Margins are small in five corpora and two embedding models, so an idealised tolerance at the level of floating-point noise leaves up to 5.5% of queries exposed (at most 1.2% in any corpus once exact duplicate documents are excluded), and a threshold calibrated from our CPU INT8-FP32 comparison of a small embedding model (at the empirical 0.1% quantile of honest cosines) leaves 98-100% exposed; on a Raspberry Pi 5 running a local language model, such an attack lowers correct answers from 0.75 to 0.39. Integrity audits of compressed vector stores fare no better: on a FAISS product-quantisation index over 57,638 documents, tampered vectors that all pass the audit put the attacker’s document first for 25% of targets. Verifying quantisation codes instead, with ambiguity decided by the verifier, a calibrated guard band and canonicalisation, accepted every honest vector in our samples on three platforms and left the publisher no choice. We also show that common floating-point and dynamic-INT8 inference pipelines are not bit-reproducible across batch sizes or architectures, and report pre-registered negative results on abstention, context pruning and semantic screening. We release code and data.
[IR-5] ArcRAG : Agent ic Compositional Retrieval over Knowledge Graphs
链接: https://arxiv.org/abs/2611.06481
作者: Jyotin Goel,Abhinav Arun,Tejas Prakash Agarwal,Bhaskarjit Sarmah,Stefano Pasquali
类目: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: Preprint
Abstract:Graph-based retrieval-augmented generation (GraphRAG) answers questions by retrieving over a knowledge graph built from a corpus. Yet retrieval strategies are designed around the structure their own construction pipeline produces, and whether they survive a change of construction is rarely tested. We study this cross-construction portability: whether one retrieval procedure, with no construction-specific schema, relation names, or retuning, stays effective when the graph is built differently. We introduce ArcRAG, which induces a query-local graph interface at inference time (the node labels, relation descriptors, and property keys reachable from the entities a question names), plans multi-hop traversals against it, and executes and repairs them as Cypher. We build each of the MuSiQue, HotpotQA, and TriviaQA corpora as a schema-guided graph (typed nodes, a closed relation set, source provenance) and as an open-information-extraction graph (untyped nodes, free-text relations), with the retriever’s configuration fixed across them. We compare against a construction-agnostic Direct-Cypher agent, ablate each stage across retry and pass budgets on both graph families, and compare with results reported for prior GraphRAG systems, most of which design retrieval together with their own graph construction. Across the two constructions, we find that retrieval need not be co-designed with the graph: query-time interface induction keeps a construction-agnostic retriever effective when the construction changes, although construction still shapes absolute accuracy.
[IR-6] PMRecBench: A Benchmark for Personalization in Recommendation Agents under Multimodal Queries
链接: https://arxiv.org/abs/2611.06349
作者: Xiaoyu Zhang,Yougang Lyu,Bin Wu,Sahan Bulathwela,Emine Yilmaz
类目: Information Retrieval (cs.IR)
备注: 13 pages. Project page: this https URL
Abstract:On Web platforms for shopping and entertainment, recommendation agents under multimodal queries aim to deliver personalized recommendations that satisfy users’ visually specified requirements and reflect preferences inferred from user histories. However, existing personalized recommendation benchmarks use text-only queries, while multimodal ones assess query matching without evaluating personalization based on real user histories. We introduce the Personalized Multimodal Recommendation Benchmark (PMRecBench), the first benchmark to evaluate the personalization of recommendation agents under multimodal queries. It pairs 510 quality-controlled image-text queries across clothing and movies with the corresponding users’ real histories. PMRecBench features two key designs: (1) queries derived from consistent behavioral patterns in real user records, with reference images and targets linked to actual interactions; and (2) challenging diagnostic pools with candidate items rated highly by the user but violating requirements, and items satisfying requirements but never rated highly, testing requirement understanding and preference following, respectively. Extensive experiments on diagnostic pools reveal a personalization imbalance in recommendation agents: relative to personalization signals alone, adding multimodal queries weakens preference following, even though the additional information about query requirements improves overall ranking performance. Further analysis links this decline in preference following to image interpretation, where agents can prioritize resemblance to reference images over evidence from user histories, and stronger commercial backbones alone fail to resolve this imbalance. This personalization imbalance highlights a central challenge for recommendation agents: improving ranking performance while preserving preference following when responding to multimodal user queries
[IR-7] Generative Retrieval under Real-Time Supply Constraints in Local Commerce
链接: https://arxiv.org/abs/2611.06323
作者: Shaopeng Liang,Biao Zhang,Hongzhen Ye,Zeyang Hou,Ma Jiang,Jintao Tu,yunxi.hz,Zhijia Fang,Jianqiang Cao,Junxue Rao,Jia Jia
类目: Information Retrieval (cs.IR)
备注: 11 pages, 3 figures, 6 tables, 2 algorithms. The first five authors contributed equally
Abstract:Generative retrieval predicts semantic identifiers for candidate entities. In local retail and food delivery, candidates must also be serviceable for a particular user at a particular time. Location, weather, and traffic reshape merchant and product supply, so a catalog-valid identifier may be infeasible for the current request. Post-generation filtering removes such candidates only after beam capacity and inference computation have been spent. We study retrieval over an explicitly materialized, request-time supply snapshot, where short output sequences leave little opportunity to amortize online constraint preparation. Our RTP-LLM pipeline compiles eligible residual-quantization identifiers into a compressed sparse row prefix index, overlaps CPU compilation with model prefill, and shares the uploaded index across beams. Legal expansions retain original-model log probabilities, and constraint states follow selected beam parents. We formalize snapshot-relative identifier validity and distinguish it from entity-level eligibility and freshness. The implementation generates three-token merchant identifiers. Historical evidence covers daily catalog constraints and subsequent request-scoped serving evaluations. In one matched-beam daily-constraint case, filtered merchants decrease from 170 of 299 candidates to 29 of 299; residual distance filtering motivates request-specific constraints. A different-host comparison with 30,000 supplied entries reports a 0.6 ms service-side latency increment, while client measurements expose additional request costs. Captured requests from three geographic settings characterize behavior across supply sizes and output widths. These observations span development stages and measurement boundaries. They establish a practical implementation and identify bottlenecks, but do not establish compute savings at matched retrieval quality.
[IR-8] Do Search-Augmented Recommenders Use the Latest Search Episode? WSDM2027
链接: https://arxiv.org/abs/2611.06044
作者: Jiandong Ding
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注: Accepted at WSDM 2027. Code: this https URL
Abstract:Search-augmented recommenders can use the latest search episode and earlier searches. We introduce FocalSwap to test how these inputs contribute to the next recommendation. For each recommendation, it changes the latest search slot or all search inputs while holding the target, candidates, pre-search history, and scorer fixed. It compares neutral masks with matched observed-search replacements, recording rank changes, RR gains relative to each control, and preference for the true focal search. A planted signal and GenSAR’s current-query task confirm that the audit detects known input dependence. We evaluate behavioral and lexical evidence on KuaiSAR and Qilin, and recommendation checkpoints trained using the released SESRec and UniSAR implementations. Across three independently trained UniSAR checkpoints, the estimated all-source benefit depends on the control. On the same 66,139 rows, replacing all search inputs with a neutral ID yields a positive signed contribution, whereas replacing them with category/profile-matched observed search sequences gives smaller estimates, with an interval below zero for one checkpoint. The paired RR gaps range from +2.770 \times 10^-3 to +6.604 \times 10^-3 , and every adjusted interval is above zero. Neutral masks change 59.7-61.7% of target ranks, while matched replacements change 44.1-44.8%. The control contrast persists across five alternative donor assignments per checkpoint. Under the primary neutral-mask and category-matched controls, positive latest-slot contribution is not established in SESRec or UniSAR. These results show why masking evidence should be compared with observed-search replacements when attributing a benefit to broader search state or the latest episode. Comments: Accepted at WSDM 2027. Code: this https URL Subjects: Information Retrieval (cs.IR); Machine Learning (cs.LG) Cite as: arXiv:2611.06044 [cs.IR] (or arXiv:2611.06044v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2611.06044 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-9] From Waste to Discovery: Recycling High-Scored Unexposed Candidates in Taobao Feed Recommendation
链接: https://arxiv.org/abs/2611.05997
作者: Ge Fan,Kai Meng,Huiping Chu,Yuning Jiang,Han Zhu,Bo Zheng
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注:
Abstract:Helping users discover new interests is vital to the long-term health of industrial recommender systems, yet the usual remedy, boosting exposure of under-explored (discovery) content, often backfires in production. We report URED (Unexposed Recycling for Efficient Discovery), a training-free recall-stage system serving Taobao feed recommendation at Alibaba for over six months to tens of millions of daily active users. It rests on a production-validated insight: naive boosting fails because it injects cold, information-poor candidates that rankers systematically over-estimate, buying no discovery gain while risking engagement. Items scored highly but never exposed (signals industrial pipelines routinely discard) form a pool that is both relevant and reliably predictable. Over a 30-day horizon URED aggregates them and re-injects at recall time through a lightweight pipeline spanning offline storage, distributed aggregation, and low-latency online serving, adding under 1% latency via a single O(1) lookup per request with no model retraining. Online A/B tests show +3.47% discovered categories and +8.47% content richness with engagement preserved: completion video views rise significantly (+2.62%) and total views move positive (+0.36%, not significant). Ablations confirm both design pillars are necessary: a one-day window collapses the discovery gain to +0.55% (not significant), and random injection without score filtering yields no gain while trending negative on engagement (-0.17%). Across production recall channels, URED reaches 13.9x the platform-average discovery rate (6.3x the strongest nearline baseline) at engagement quality on par with the platform average. We share deployment details and lessons for practitioners facing the exploration-exploitation dilemma under tight resource constraints.
[IR-10] One Basket Many Experts: Decorrelated Re-ranking for Financial Asset Recommendation
链接: https://arxiv.org/abs/2611.05952
作者: Ryuki Unno,Keigo Sakurai,Koshi Watanabe,Keisuke Maeda,Takahiro Ogawa,Miki Haseyama
类目: Information Retrieval (cs.IR); Computational Engineering, Finance, and Science (cs.CE)
备注: Accepted to the 7th ACM International Conference on AI in Finance (ICAIF2026)
Abstract:Financial asset recommendation for retail investors is typically optimized for relevance, that is, how well a recommendation matches what each user will buy. From the service provider’s point of view, the value of a recommendation also depends on portfolio performance: the risk-adjusted return the recommended assets (the basket) deliver, which drives user engagement and fee revenue. The two goals are rarely achieved jointly, because the recommendations that best match user behavior concentrate on a few popular assets whose returns are highly correlated, which caps the risk-adjusted return a provider can offer. We propose Decorrelated-Satellite Expert Re-ranking (\textbfDSER), which retains the most relevant recommendations from an expert-following list and replaces the rest with return-decorrelated alternatives. The number of replaced slots is a single integer~ m (a knob) that the provider sets, so the relevance–performance trade-off becomes an explicit, auditable choice rather than a property hidden inside a learned model. In experiments on the public FAR-Trans retail-brokerage dataset, under the primary protocol, replacing a few slots improves risk-adjusted return at a small relevance cost, and the gain survives realistic trading costs at typical retail portfolio sizes. Minimizing correlation would lower basket volatility, but in our data the least-correlated assets are more volatile and higher-return, so their volatility offsets that reduction while their return drives the gain, not lower volatility. The resulting trade-off is asymmetric, with the relevance cost consistent across market phases while the gain is phase-dependent. The provider sets m to read off exactly how much relevance each operating point costs.
[IR-11] Optimal Re-Ranking Depth
链接: https://arxiv.org/abs/2611.05908
作者: Siqing Huo,Andrew Parry,Debasis Ganguly,Charles L. A. Clarke
类目: Information Retrieval (cs.IR)
备注:
Abstract:Second-stage neural rankers are commonly applied with a fixed re-ranking depth, assuming that retrieval effectiveness saturates as depth increases. Prior research has questioned the assumption that increasing the re-ranking depth yields linear performance gains, further suggesting that optimal re-ranking depth varies considerably from query to query. In the past, studying such phenomena was methodologically difficult given the scale of manual annotation required. With the advent of LLM-based relevance judgments we can now more easily undertake such studies, in this case to pinpoint the optimal re-ranking depth on a per-query basis. Using dense LLM-based relevance judgments over a typical re-ranking pipeline, we show that many queries exhibit a well-defined optimal re-ranking depth, beyond which effectiveness stagnates or degrades. We formulate re-ranking depth as a query-specific property and study whether it can be predicted a priori from first-stage retrieval characteristics. Through a large-scale analysis, we find that most standard query performance prediction (QPP) methods are ineffective for this task. In contrast, a predictor derived from LLM-assessed first-stage ranking quality, which we term IR-DCG@10, can reduce the average re-ranking depth by up to a factor of 3 while preserving overall effectiveness, depending on the specific first- and second-stage rankers used. Under oracle selection of optimal depths, we further show that retrieval effectiveness can improve by more than 7% while reducing the average re-ranking depth by a factor of 5 on the MSMARCO DEV collection. Given our promising preliminary findings, we would encourage the use of automatic judgments to facilitate research otherwise infeasible under manual annotation. Towards this point, we release relevance judgments, our codebase, and experimental artefacts to support reproducibility and further research.
[IR-12] Weaver: Weaving Preference Evidence through Incremental Reasoning for Personalized Retrieval-Augmented Generation
链接: https://arxiv.org/abs/2611.05900
作者: Linyao Zheng,Zhifang Mao,Xuhang Shi
类目: Information Retrieval (cs.IR)
备注:
Abstract:Personalized retrieval-augmented generation requires extracting and applying user preferences from extended interaction histories. This task is challenging because preference evidence is scattered across past conversations and remains bound to specific temporal and situational contexts. Consequently, isolated fragments rarely provide complete support for a response, while preferences expressed under different conditions are easily conflated. To address these limitations, we propose Weaver, a framework that incrementally weaves memory-conditioned retrieval trajectories with compositional preference reasoning for personalized responses. Weaver operates on a Dual-Layer Retrieval and Reasoning Graph that couples an upper Intent Dependency Graph with a lower Memory-Conditioned Retrieval Trajectory Graph. Dependency Information Gap Tracking represents the information needed to answer the current request and its prerequisite dependencies in the upper graph. Memory-Conditioned Incremental Retrieval uses established premises and clues from retrieved conversations to locate missing evidence, recording the resulting retrieval trajectories in the lower graph. Scope-Aware Compositional Reasoning jointly interprets the accumulated evidence to determine which preferences apply to the request. It identifies complementary and alternative support and distinguishes conflicting claims within the same temporal and situational context. These conclusions iteratively update task premises and remaining information gaps, guiding subsequent retrieval to establish a coherent contextual basis for the final response. Experiments on personalized RAG benchmarks, particularly PersonaMem, demonstrate that Weaver outperforms leading baselines in both evidence completeness and answer quality.
[IR-13] SciToT: Multimodal Tip-of-the-Tongue Retrieval for Scientific Papers
链接: https://arxiv.org/abs/2611.05829
作者: Stan Fris,Sijie Tao,Makoto P. Kato,Maarten de Rijke
类目: Information Retrieval (cs.IR)
备注:
Abstract:Researchers frequently encounter difficulties re-finding scientific papers when they remember aspects of the content but cannot recall reliable identifiers, such as titles or author names. This type of re-finding task, known as tip-of-the-tongue (ToT) retrieval, is particularly challenging for scientific documents, where recollections may involve textual information and visual elements, including figures, diagrams, and plots. Existing ToT retrieval benchmarks do not address scientific search. We introduce a multimodal benchmark for scientific ToT retrieval, comprising 108,512 open-access scientific documents across five domains and 11,499 synthetically generated queries. To construct the benchmark, we propose a scalable query elicitation pipeline that independently extracts textual and visual memory cues from source documents and uses large language models to generate queries resembling incomplete human recollections. We characterize scientific ToT information needs through an empirical analysis of publicly available re-finding requests and use the outcomes to assess the generated queries using human annotations and automated quality control. The benchmark includes lexical, dense, late-interaction, and visually grounded retrieval baselines. We release the dataset, extracted document representations, generated memory cues, evaluation splits, and accompanying generation and retrieval pipelines under an MIT license to facilitate research on multimodal scientific paper re-finding. Subjects: Information Retrieval (cs.IR) Cite as: arXiv:2611.05829 [cs.IR] (or arXiv:2611.05829v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2611.05829 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-14] Item Representations Matter: Bridging Hard Semantic IDs and Soft Prototype Mixtures for Generative Retrieval
链接: https://arxiv.org/abs/2611.05799
作者: Vicente Balmaseda,Dongqi Fu,Jiarui Feng,Ren chen,Hanqing Zeng,Tianxin Wei,Yinglong Xia,Tianbao Yang,Hong Li
类目: Information Retrieval (cs.IR); Machine Learning (cs.LG)
备注:
Abstract:Modern recommender systems rely on generative retrieval techniques to predict the next item either as discrete codes or as a continuous vector. On the one hand, Semantic IDs (SIDs) give discrete models shared parameters and pretrained semantic structure, but discretizing items into hard codes discards some information from the original content embeddings. On the other hand, continuous models avoid quantization, yet typically use flat item lookup tables with no explicit parameter tying across items. We ask whether a continuous recommender can introduce shared structure without making it a discrete bottleneck. We propose CAP-Mix, a flow recommender that represents each item as a soft mixture of shared prototypes plus an item-specific vector. The assignments, prototypes, and item vectors are trained jointly, and the soft representation is retained at inference. Pretrained information is optional: SID responsibilities can serve as a prior over assignments, while a full content embedding can initialize the item-specific vector without changing the initial flat table. In matched comparisons on three datasets, CAP-Mix improves over a flat table on the two sparse Amazon datasets and yields a small gain on Steam when trained without pretrained information. Given only an offline SID, it outperforms a hard table built from the same tokenizer by 4.45-14.90% in HR@5. Given the full content embedding, it matches or improves a strong flat table while hard assignments regress on the sparse Amazon datasets. Overall, averaged across six benchmarks, CAP-Mix improves HR@5 and NDCG@5 by 6.9% and 13.6% relative to the best evaluated baseline for each dataset and metric, while delivering 1.4-161x the inference throughput of the discrete baselines in our experiments.
[IR-15] oward a Compositional Knowledge Architecture for a Shared Organizational Cortex
链接: https://arxiv.org/abs/2611.05753
作者: Andrew Miracle,Phyl Georgiou,Christine Okoth,Ese Monday
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: 25 pages, 10 figures, 10 tables. Evaluation and sources in this https URL (tag registered-plan)
Abstract:Organizations accumulate evidence in conversations, documents, decisions and relationships, but they act on interpretations: judgements assembled from several observations, resting on assumptions, and revised as circumstances change. Existing retrieval, knowledge graphs and agent memory keep the records and lose the interpretation. We propose a four-layer architecture, Signals, Domains, Claims and Dimensions, in which interpretations persist with their justifications and standing, a truth-maintenance network revises that standing as evidence changes, and validated, versioned Dimensions position interpretations on named axes. We call the result a shared organizational cortex. In a registered retrieval experiment, the Dimension term gains 0.028 nDCG@10 (p = 0.002). In system-level comparisons, the full retrieval configuration improves nDCG@10 by 0.081 over Signals-only and 0.053 over graph-augmented Signals-only retrieval (both reported p 0.001). The architecture gives organizations a persistent, revisable basis for shared judgement.
[IR-16] urbocharging PiPNN for Proximity and k-NN Graph Building
链接: https://arxiv.org/abs/2611.05690
作者: Tobias Rubel,Richard Wen,Guy Blelloch,Laxman Dhulipala,Lars Gottesbüren,Jakub Łącki,Vahab Mirrokni
类目: Databases (cs.DB); Information Retrieval (cs.IR)
备注:
Abstract:Graph-based Approximate Nearest Neighbor Search (ANNS) indexes such as HNSW and Vamana deliver state-of-the-art query performance at the cost of large indexing times. The recently introduced PiPNN method drastically accelerated index construction by way of a partitioning-based algorithm leveraging dense matrix multiplication and top- K aggregation kernels. In this paper, we unlock the full potential of PiPNN’s hardware-friendly algorithm design. We introduce PiPNN II, which optimizes multiple steps of the PiPNN recipe, giving up to 8\times speedups for graph-based ANNS index construction. In addition, we present GPiPNN, a GPU implementation which builds comparable ANNS indices up to 43.7\times faster than Jasper and up to 50\times faster than CAGRA. We also extend PiPNN II and GPiPNN to compute high-quality k -NN graphs directly, which unlocks massive speedups over methods used in many existing systems for clustering, anomaly detection, and more. For 10-NN graph building at 90% recall, we obtain up to a 13\times speedup on CPUs and 6.6\times on GPUs as compared to the prior best methods. PiPNN II can build a billion-scale 10-NN graph in under 16 minutes. Common to both PiPNN II and GPiPNN is the use of quantization to reduce memory traffic and leverage fast low-precision matrix operations. Top- K emerges as a new bottleneck, which we optimize with specialized code for specific K , and by fusing with the matrix kernel to avoid materializing the distance matrix to memory. On CPUs, we demonstrate further performance improvements by leveraging our matrix kernels in the RobustPrune procedure proposed in Vamana and used in PiPNN. On GPUs, we flatten recursive partitioning into segmented batches, use one fused top- K abstraction for partitioning and leaf scoring, and prune with already-witnessed direct edges rather than new candidate-to-candidate scoring passes. Subjects: Databases (cs.DB); Information Retrieval (cs.IR) Cite as: arXiv:2611.05690 [cs.DB] (or arXiv:2611.05690v1 [cs.DB] for this version) https://doi.org/10.48550/arXiv.2611.05690 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[IR-17] Combining Relevant Expert Knowledge Using Mixtures of Feedback Fuzzy Cognitive Maps
链接: https://arxiv.org/abs/2611.05664
作者: Olaoluwa Adigun,Akash Kumar Panda,Bart Kosko
类目: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Symbolic Computation (cs.SC)
备注: 8 page, 8 figures, 2 tables
Abstract:A neuro-symbolic system should combine only relevant answers from expert agents or other knowledge sources. Using only relevant experts for a given input query becomes ever more important for accuracy and scalability as the knowledge network grows and connects ever more knowledge sources. We extend the standard definition of relevance from the legal theory of evidence: An answer is relevant if and only if it is both probative and material for the matter at hand given the knowledge available. We apply this relevancy scheme to mixing the feedback fuzzy cognitive maps (FCMs) of real or artificial experts. The expert FCM directed causal graphs mix or combine into a common knowledge network. We compare the equilibrium limit-cycle predictions or answers of the experts to the answers of the mixture FCM and then estimate the probative status of answers by their difference or variation. A new Bayes-like theorem de-mixes the knowledge mixture into corresponding posterior-like FCMs that also allow relevancy-based comparisons. The result is a final set of relevancy mixing weights that produces a relevancy-mixed FCM knowledge network for the given query. So different queries produce different relevancy-mixed FCMs. We demonstrate this relevancy approach using generated synthetic FCMs from an actual 13-node expert FCM that models the causes and feedbacks of ischemic-stroke risk.
[IR-18] Data-Efficient Scaling-Law Discovery for Recommendation Systems via Coupled Hyperparameter Scaling
链接: https://arxiv.org/abs/2611.05643
作者: John Martabano Landy,Louis C. Tiao,Yao Xuan,Luc Kien Hang,Rupinder P. Khandpur,Arvind M. Vepa,Srinidhi Raghavendran,Anirudh Agrawal,Xiaoyi Leo Liu,Jiang Liu,Peggy Yao,Alireza Vahdatpour,Djordje Gligorijevic
类目: Machine Learning (cs.LG); Information Retrieval (cs.IR)
备注: 12 pages, 10 figures
Abstract:Large-scale ranking and recommendation models of Web platforms benefit from increased compute much as large language models do, and their scaling behavior is broadly power-law in form. Yet ranking models have heterogeneous architectures which makes their scaling law less straightforward than that of LLMs. We address how to discover the scaling law of an arbitrary ranking model efficiently. We introduce a scaling layer, which groups primitive architecture parameters into semantically coherent groups and defines how they co-vary, inducing a mapping from a low-dimensional space of hyper-hyperparameters into the full architecture space. On top of it, Axia fits scaling-informed surrogates, selects configurations by a posterior-based acquisition criterion, and returns a Pareto frontier from which per-model operating points are validated – replacing exhaustive sweeps with a small number of informative runs. A fleet-level evaluation protocol, computed on a fixed representative model set under matched trial budgets, makes refinements to either component measurable. Our contribution is an automated system that learns model-specific complexity-quality relationships and gives uncertainty-aware guidance for allocating compute across a heterogeneous model ecosystem. Across 34 model iterations in two development cycles on Meta’s ads ranking and retrieval system, the workflow yielded a mean 0.66% offline normalized-entropy improvement and 0.36% cumulative online lift.
[IR-19] RAG Compass: Generalizable Retriever Routing via Reranking
链接: https://arxiv.org/abs/2611.05596
作者: Valentin Six,Mufei Li,Dongqi Fu,Yinglong Xia,Hong Li,Pan Li
类目: Computation and Language (cs.CL); Information Retrieval (cs.IR)
备注:
Abstract:Different retrievers exist for retrieval-augmented generation, yet none performs best for all queries. Therefore, the key question is: given a query and retrieved passages from different retrievers, which one is most suited for that query? Existing routers usually make predictions from query-specific features and a fixed pool of retrievers; such routers have no output for an unseen retriever and are incapable by design on a new corpus. To fill these gaps, RAGCompass reads the retrieved passage sets themselves: passages from an unfamiliar retriever or corpus stay directly comparable to other passages, so routing stays well designed where a query-only map would fail. We contrastively fine-tune a reranker against the answer each evidence set produces, which accounts for most of the gain; a small comparison head adds a moderate improvement, and neither component has retriever-specific parameters. RAGCompass improves answer F1 by 28% relative to the best fixed retriever, and outperforms fusion baselines, three query-only routers, and a prompted 8B language model. Generalization experiments show that RAGCompass keeps most of its gain under an unseen corpus, retriever, or reader, while query-only baselines fall to the best fixed retriever or below.
[IR-20] LIBRA: Programmatic Induction of Executable Tool Libraries DATE
链接: https://arxiv.org/abs/2611.05445
作者: Zekai Li,Zhenke Liu,Yonghao Li,Junhan Liu,Xulong Xiao,Jiapeng Zhang
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
备注: 10 pages, 5 figures. Introduces LIBRA, a query-agnostic tool library built via execution-validated prerequisites in a sandbox (avg ARR review score: 3.5/5). Demonstrates consistent performance gains across three benchmarks while discussing sandbox dependencies and library construction trade-offs
Abstract:Large language model agents increasingly retrieve tools from large catalogs, yet most retrievers rank each tool independently by request similarity. For multi tool tasks, this recovers the goal matching tool while omitting prerequisite tools required by the this http URL introduce LIBRA, a framework for Library Induction by Runtime Assembly that turns a flat tool catalog into an executable tool library. An LLM proposes per tool contracts of candidate invocations, reusable outputs and state changes, and success conditions; LIBRA then tests these compositions in a sandbox and retains the relations supported at runtime. Constructed before user queries, the library requires no dependency annotations, tool use trajectories, or task specific training, and supports both one-shot and agentic retrieval. On AppWorld, BFCL, and SciToolEval, LIBRA improves one-shot Recall and NDCG over the same dense retriever by 7.9 and 3.7 percentage points on average. Within ToolReAGt, it yields corresponding gains of 4.5 and 2.9 points after ten retrieval iterations.
[IR-21] Algorithmic Authority and the Clinical Standard of Care
链接: https://arxiv.org/abs/2606.00044
作者: Aizierjiang Aiersilan
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Social and Information Networks (cs.SI); Systems and Control (eess.SY)
备注:
Abstract:The integration of artificial intelligence into clinical medicine creates a fundamental tension between algorithmic probabilistic reasoning and the experiential intuition of expert physicians; applying Lawrence Lessig’s \enquoteCode is Law framework, I argue that the architecture of clinical AI systems already functions as de facto medical regulation, reshaping liability and the standard of care. Reframing AI \enquotehallucination as structurally analogous to well-documented human cognitive failures such as confirmation bias and premature diagnostic closure, I show that both failure modes demand a unified governance response. I therefore propose a dialectical standard of care that treats the integrated AI-physician dyad as the singular responsible diagnostic entity, mandating the synthesis of algorithmic precision with human interpretive authority within robust data governance and patient privacy frameworks.
人机交互
[HC-0] Dull Dirty Dangerous? Working Conditions and the Limits of Occupational AI Exposure
链接: https://arxiv.org/abs/2611.06888
作者: Siyi Yu
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); General Economics (econ.GN)
备注:
Abstract:Automation is often justified as relief from dull, dirty, and dangerous work. I examine how these conditions relate to occupational AI exposure using 21 O*NET items across 754 US occupations. Higher exposure is associated with lower condition scores under a large language model (LLM) measure and AI Occupational Exposure (AIOE), with equal occupation weights and employment weights. Adjustment for physical demands, required education, training and experience, and wages reduces the composite association. The adjusted confidence intervals include zero for LLM exposure and remain below zero for AIOE. Separating repetition from restricted discretion also changes what the combined Dull score conveys. These comparisons show how item groupings and occupational characteristics shape the interpretation of exposure estimates. Occupational profiles offer a starting point for selecting contrasting research settings. Establishing desired automation or realized relief requires evidence about workers’ preferences and specific interventions.
[HC-1] A Frame-Level View and Image-Quality Annotation Tool for Fetal Echocardiography: Comparison with Spreadsheet Annotation
链接: https://arxiv.org/abs/2611.06791
作者: Mohamed Azzam,Ziyang Xu,Quinlan Couch,Zoe Keese,Andrea Cadlo,Christine Sedze-DeWolf,Micaiah Uhrmacher,Madelyn Graff,Catherine Fox,Lauren Vatter,Paige Jeffryes,Jason Christensen,Neil Hamill,Shibiao Wan,Ling Li,Jieqiong Wang
类目: Human-Computer Interaction (cs.HC); Image and Video Processing (eess.IV)
备注:
Abstract:Objective: To develop the Cardiologist Annotation Tool (CAT) for frame-level cardiac view and image-quality annotation in fetal echocardiography and compare it with spreadsheet-based annotation. Methods: CAT provides cine playback, frame navigation, and keyboard-based labeling without manual frame-range entry. Experts annotated 25 studies using both methods, and spreadsheet annotations were audited for predefined data-entry errors. In a randomized crossover study, six medical students annotated two normal studies, one with each method, against an expert reference. Results: Among 1,746 spreadsheet-annotated files, 6.2% contained major or critical data-entry errors, more frequently in congenital heart disease than normal studies (10.9% vs 3.4%). CAT yielded more annotated view segments per clip (3.6 vs 1.5). Rates of labeled and correctly labeled frames per minute were 1.77 and 1.68 times those with spreadsheets, respectively (both P = 0.031), with gains in all six students. No detectable difference in view-label accuracy was observed. Conclusion: CAT eliminates manual frame-range entry and enables finer-grained, structured annotation of fetal echocardiography, providing a practical approach to preparing datasets for AI research. Subjects: Human-Computer Interaction (cs.HC); Image and Video Processing (eess.IV) Cite as: arXiv:2611.06791 [cs.HC] (or arXiv:2611.06791v1 [cs.HC] for this version) https://doi.org/10.48550/arXiv.2611.06791 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Mohamed Azzam [view email] [v1] Wed, 4 Nov 2026 17:07:12 UTC (1,134 KB)
[HC-2] Go-Go-Go: the game of Go without turns
链接: https://arxiv.org/abs/2611.06786
作者: Andrea Cambiaghi,Marco Tarini
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:The study of variants of classic games has a long history, with chess alone inspiring countless variants that have been proposed, analyzed, and studied. The game of Go, despite rivaling chess in age, cultural relevance, popularity, and strategic depth, has seen drastically fewer variants. We introduce Go-Go-Go, a variant of the classic game of Go where there are no turns, and players play stones simultaneously and asynchronously on the board. The variant (which is to be played through a computer) includes a basic mechanism to avoid excessive spamming of moves, by imposing a minimal delay between moves that depends on their relative spatial distance. Apart from the lack of time discretization, most of the ruleset of the original game is either directly maintained or adapted in a natural way that preserves the original spirit. The first playtesting indicates that this results in the preservation of much of the emerging strategies and “flavor” of Go, while replacing the long strategic reasoning with fast paced, instinctive action fueled by quick reflexes. Play testers also report the need to balance immediate but suboptimal reactions with more reasoned but slower plays. We offer a publicly available, playable, OpenSource, and online reference implementation of Go-Go-Go at this http URL Subjects: Human-Computer Interaction (cs.HC) MSC classes: 91A25 ACMclasses: J.m Cite as: arXiv:2611.06786 [cs.HC] (or arXiv:2611.06786v1 [cs.HC] for this version) https://doi.org/10.48550/arXiv.2611.06786 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[HC-3] Smartphone-Aided VR Exergames for HIIT: Gamifying Fitness with Real-Time Pose Estimation
链接: https://arxiv.org/abs/2611.06734
作者: Abhiram Kothagundu,Pheelip Pritipal Raipure,Abhijit Chandrakant Patil,Himangshu Sarma,Christian Eichhorn,David A. Plecher,Johanna Pirker
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:This paper introduces an innovative virtual reality (VR) exergame system designed to gamify High-Intensity Interval Training (HIIT) workouts through 11 predefined combo moves, including squat-punch combinations, bilateral arm strikes, and diagonal stretch combos. Unlike conventional tracking methods that often rely on wearable motion capture suits or limited headset-based sensors, our system uses a standard smartphone camera placed on a tripod to capture full-body movement. This approach offers a cost-effective and scalable alternative, particularly well-suited to the fast, dynamic motions typical of HIIT routines. The smartphone footage is processed in real time using a pose estimation algorithm, with extracted biomechanical data analyzed by a backend server built on the Model Context Protocol (MCP), Anthropic’s open standard for connecting AI models to structured data sources and tools, and a GPT-4o large language model. The system generates personalized assessments of form accuracy, movement efficiency, exercise intensity, and potential injury risks, tailored specifically to high-impact, time-sensitive exercises. A within-subject study involving 20 participants (ages 18-25) compared experiences in VR and Non-VR HIIT sessions. Results indicated significantly higher enjoyment (Mean = 3.5 vs 2.8, p = 0.02), improved immersion/focus (Mean = 3.2 vs 2.8, p = 0.03), and increased likelihood to recommend (Mean = 3.2 vs 2.7, p = 0.04) under the VR condition. Confidence to continue did not differ significantly between conditions (p = 0.15), while perceived intensity remained comparable. The MCP system showed strong agreement with expert evaluation (87.5%, N = 10 sessions, one certified trainer), supporting its reliability as a feedback tool. Subjects: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.06734 [cs.HC] (or arXiv:2611.06734v1 [cs.HC] for this version) https://doi.org/10.48550/arXiv.2611.06734 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[HC-4] You Know What You Will Like: People Predict Their Own Ratings Better Than Recommender Systems
链接: https://arxiv.org/abs/2611.06696
作者: Zihan Chen,Lei Nico Zheng,Di Zhu
类目: Human-Computer Interaction (cs.HC); General Economics (econ.GN)
备注: Working paper. 23 pages, 3 figures, 10 tables
Abstract:Recommender systems usually learn what people want from what they have done. Platforms could also ask people directly, but asking is costly: 98.8% of the prompts in our data went unanswered. Theories of statistical prediction imply that a person’s own forecast adds little to a recommender algorithm built from many people’s behavior. Theories of dispersed knowledge imply that people hold information about their own taste that the data lack. We test the two with the MovieLens Beliefs dataset. In 2023-2024, a homepage prompt asked members how they expected to rate movies they said they had not seen. Of 26,020 such expectations, 551 from 277 members were followed by a rating at least a day later. We compare each expectation with four benchmarks: the platform’s own predicted rating, two collaborative-filtering models (matrix factorization and an item neighborhood), and a stack trained on other members’ outcomes. The trained models use each member’s latest ratings. People predict their own ratings more accurately than all four benchmarks (correlation 0.69, against 0.34 for MovieLens and 0.54 to 0.59 for the trained recommenders). Added to all the machine forecasts and features, the expectation lowers out-of-sample mean squared error by 26%. The gain holds when models are trained on earlier expectations and tested on later ones. A member’s earlier expectations about other movies do not replace the current one. In ordering a member’s own movies, the trained recommenders are statistically indistinguishable from the person. On old, heavily rated movies, the machines do as well or better. The results reconcile the two theories: the crowd’s data suffice where they are rich, and the person adds information about the specific item. We derive four design principles for preference elicitation and recommender evaluation.
[HC-5] Cognitive and Behavioural Effects of Haptic Feedback for ROV Teleoperation in Hazardous Environments
链接: https://arxiv.org/abs/2611.06689
作者: Joshua Brown,Ildar Farkhatdinov,Michael Jenkin
类目: Robotics (cs.RO); Human-Computer Interaction (cs.HC); Systems and Control (eess.SY)
备注: Accepted for IEEE Telepresence 2026
Abstract:Teleoperating a Remotely Operated Vehicle (ROV) in complex underwater environments presents unique challenges. Dynamic, three dimensional obstacles and unseen water currents pose real risks to expensive robotic hardware in environments in which recovery can be next to impossible. This paper explores the cognitive effects of haptic feedback to improve vehicle safety in such environments. An in-water study is analysed with a ROV exposed to controlled cross-currents, with environmental information presented via a soft and vibrating haptic touchpad. Results demonstrate that while encountering an unseen cross-current significantly increases an operator’s physical and temporal demand, the introduction of haptic feedback mitigates these factors and improves self-assessed performance. Interestingly, the haptic condition also produced the highest levels of mental demand and operator frustration. Despite this increased mental workload, operators drove the ROV significantly faster when receiving haptic cues compared to the current-only condition, indicating a higher level of user confidence. Ultimately, these findings suggest that while haptic feedback may initially require more cognitive processing to interpret unseen hazards, it successfully enhances operator awareness and promotes more confident navigation in hazardous environments. Comments: Accepted for IEEE Telepresence 2026 Subjects: Robotics (cs.RO); Human-Computer Interaction (cs.HC); Systems and Control (eess.SY) Cite as: arXiv:2611.06689 [cs.RO] (or arXiv:2611.06689v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2611.06689 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[HC-6] Perceived Trust and Transparency in AI-Altered vs. Human-Written LinkedIn Posts
链接: https://arxiv.org/abs/2611.06660
作者: Thomas Christian Joos,Azza Bouleimen,Anikó Hannák,Nicolò Pagan
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:People increasingly draft a social media post themselves and then let an AI model refine it (AI-altered), a common but understudied middle case between fully human-written and fully AI-generated content. We ask whether this changes how trustworthy and transparent readers perceive the post to be, and whether they can even tell. We first analyze real ChatGPT conversations (WildChat) to find how people actually alter social media posts, which we use to design realistic alteration prompts. We then run a pre-registered, within-subjects online experiment (n=214) in which participants rate unlabeled human-written and AI-altered LinkedIn-style posts for perceived trustworthiness and transparency, and separately judge whether each post was AI-altered. We find that AI-alteration itself has no detectable effect on either perceived trustworthiness or transparency (both estimates near zero, not significant). At the same time, readers did not reliably detect AI-alteration (49% accuracy, near chance) and defaulted to assuming human authorship when unsure. However, our analysis shows significant differences in transparency and trustworthiness ratings across the specific post, the AI model, as well as the prompt used to alter it. For instance, more detailed prompts raise both perceived trustworthiness and transparency but also make the alteration more detectable. Surprisingly, the prompt that shifted ratings most also asked the model to add one new idea absent from the original post: readers nonetheless rated this AI-invented, unverifiable content as more trustworthy and transparent.
[HC-7] Measuring the collaboration not the agent
链接: https://arxiv.org/abs/2611.06487
作者: Xinxiao Li,Eisei Nakahara
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Benchmarks for visualization agents score the agent alone; the human only supplies reference answers. For high-dimensional data this misses what matters: the knowledge of which variables count is held by on-site experts, not by the model, so the useful question is how well an analyst and an agent do together. We study one specific form of collaboration, the agent proposes a view, the analyst accepts, rejects, or modifies it, and those judgments are kept for later sessions, and propose five measures for it: complementarity, intervention efficiency, appropriate reliance, knowledge retention, and agency. The first four are computed from interaction logs alone. We test the measures, not people: in a testbed for parallel-coordinates design on public manufacturing data (UCI SECOM), with a simulated analyst whose behavior we control, all four are computable and respond to the controlled differences. Whether they reveal anything about real analysts is the next step; we release the testbed for it.
[HC-8] odays Work Tomorrows Workforce: Incentivizing Learning in AI-Assisted Work
链接: https://arxiv.org/abs/2611.06346
作者: Moritz Diener,Daniel Hendriks,Philipp Spitzer,Gerhard Satzger,Joshua Holstein,Wei Sun
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
备注: 30 pages, 6 figures, 6 tables. Moritz Diener and Daniel Hendriks contributed equally
Abstract:At work, people develop skills through experience: novices acquire skills by completing tasks, while experts retain theirs through continued practice. AI assistance changes this process by offering workers a choice between solving tasks independently and accepting the AI’s answers. While relying on AI saves immediate effort, it can also displace the practice needed to acquire and retain skills, including those required to oversee the AI. We extend an existing accept-or-solve model to a sequence of tasks in which workers’ skills evolve through independent practice and reliance on AI. We compare myopic workers, who maximize immediate utility, with strategic workers, who also value their future skills. We show that novices can become trapped at low skill levels, while experts may lose skill, even as AI-assisted outcome accuracy remains high. Strategic workers, by contrast, may choose to solve independently even when AI is more accurate, recognizing the future value of acquiring and retaining their skills. We then study incentives that reward correct independent work and characterize the payments needed to encourage skill acquisition and retention. We show that independent work can become self-sustaining once workers acquire sufficient skill, eliminating the need for continued incentives. Counterintuitively, supporting a higher skill target can sometimes reduce long-run incentive costs, either by enabling self-sustaining independent work or by lowering the payments needed to maintain proficiency. Our findings thereby shift the focus from whether workers should rely on AI on individual tasks to how organizations can sustain human expertise over time, and which skills are worth investing in.
[HC-9] Evaluating Human-LLM Alignment on Visualization Preference IEEE-VIS2026
链接: https://arxiv.org/abs/2611.06272
作者: Tanmay Bakshi,Chenglong Wang,Simra Shahid
类目: Human-Computer Interaction (cs.HC)
备注: IEEE VIS 2026 (VISxGenAI Workshop)
Abstract:There are often multiple valid visualizations, spanning different chart types, visual encodings, and aesthetic choices, that can all support a user’s goal of analyzing a dataset. Determining which of these candidate visualizations is most appropriate, however, remains challenging. Prior work has leveraged expert-curated taxonomies to guide visualization design and subsequent quality evaluation, a practical evaluation model must also account for human-centric factors. Latest visualization tools integrate Large Language Models (LLMs) to generate and refine visualizations, using LLM-as-a-judge to provide design feedback. Without a deeper understanding of whether, and in what ways, human and LLM perceptions of visualizations align, there is a real risk of deploying low-quality agents that do not reflect human judgment. We study this question through a pilot experiment - we formulate the problem as a preference task over curated chart pairs and compare the judgments of state-of-the-art vision models with those of a panel of human annotators. We establish a cautionary tale: such judges risk filtering out informative, compelling displays. We find that frontier models collapse human subjectivity, a rich signal that reveals contrasting design choices, into near-deterministic verdicts.
[HC-10] Understanding the Perception of Transliterated Offensive Social Media Posts Among Banglish and Hinglish Users
链接: https://arxiv.org/abs/2611.06226
作者: Dhiman Goswami,Zhicong Lu,Sanchari Das
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Effective content moderation depends on understanding not just what harmful content looks like, but how users perceive and respond to it. When offensive language is written in transliterated forms such as Banglish or Hinglish (Bangla or Hindi in Latin script), it may be interpreted very differently than in native scripts, with direct consequences for whether users report, ignore, or recognize it as harmful. Yet little is known about how script choice relates to user perception, or how these perceptions vary across demographic groups. To investigate these perceptions, we conducted a user study with 310 participants ( 193 Hindi, 60 Bangla, 57 Bangla-Hindi bilingual speakers) who evaluated offensive posts in native scripts and transliterations. Participants provided offensiveness scores ( 4.03 for Banglish, 3.86 for Hinglish) to transliterated posts, yet reporting fell from 20.9% (native - Bangla, Hindi) to 18.2% (transliterated - Banglish, Hinglish). Ignoring remained the dominant response (range 25.38%-44% overall). Women reported at higher probability than men for native-script ( 23.6% vs. 17.7% ) and showed almost identical reporting probability in comparison to men for transliterated posts ( 17.5% vs. 17.9% ), and younger participants ( 40 ) rated Banglish content as more offensive than older participants ( \geq 40 ), (Overall offensive score: 3.99 vs. 3.78 ). These findings identify transliteration as critical blind spot for moderation, underscoring the need for linguistically sensitive approaches to online safety.
[HC-11] GPUMAP: GPU-Accelerated Progressive Uniform Manifold Approximation and Projection
链接: https://arxiv.org/abs/2611.06191
作者: Myeongwon Jung,Jaemin Jo
类目: Machine Learning (cs.LG); Graphics (cs.GR); Human-Computer Interaction (cs.HC)
备注:
Abstract:Dimensionality reduction (DR) techniques are fundamental to the analysis of high-dimensional data, yet their substantial computational cost limits their use in interactive settings. To address this limitation, we present GPU-accelerated Progressive Uniform Manifold Approximation and Projection (GPUMAP), an efficient progressive DR algorithm designed for interactive use. We identify three atomic operations within UMAP and restructure them into a progressive pipeline. As part of this adaptation, we introduce modifications to UMAP, including operation queues, instance-wise local join, and scheduling strategies. We further accelerate computation by leveraging GPU parallelism, enabling interactive analysis of million-scale datasets on commodity GPU hardware. Finally, we employ our implementation to systematically study the behavior of progressive DR, investigating how projections evolve over time relative to the ground-truth projection and how biased data streams influence the quality and stability of the resulting projections.
[HC-12] ProGram: A Grammar for Progressive Computation Pipelines
链接: https://arxiv.org/abs/2611.06188
作者: Seunghoon Jung,Jaemin Jo
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Progressive Visual Analytics (PVA) systems have demonstrated the effectiveness of time-bounded, incremental computation for exploring large-scale data. However, our survey reveals that existing PVA implementations are often difficult to reuse and extend, as their progressive behaviors are typically realized through ad hoc, hard-coded mechanisms distributed across servers and clients. To address this limitation, we present ProGram, a grammar for progressive computation pipelines. Informed by a comprehensive analysis of prior concepts and design practices in existing PVA systems, ProGram is designed to support core progressive features while reducing implementation burden. Furthermore, we extend the design space by introducing novel mechanisms, such as lockstepping, forking, and temporary prioritization, to better support diverse usage scenarios. Finally, we demonstrate the expressiveness and coverage of ProGram by reproducing and extending existing progressive visualization systems, presenting a gallery of progressive computation pipelines, and discussing the grammar coverage.
[HC-13] FireOPS: A Human-Centric Emergency Operations Platform
链接: https://arxiv.org/abs/2611.06176
作者: Alexander Wallner,Michael Riegler,Philipp Neuhauser,Helmut Haider,Lukas Neuhauser,Marco Stadler,Eduard Paireder
类目: Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注:
Abstract:Volunteer fire brigades rely on fragmented alerting channels and radio plus paper for situational awareness, hindering a shared common operational picture (COP). FireOPS integrates smartphone alerting, navigation, a live geospatial COP, a district resource-load overview and contextual manuals into a human-centric decision-support surface for commanders and responders, running on an edge + cloud architecture whose dual-channel provides source-level redundancy against failure of either the radio or the provincial server feed; delivery to smartphones still depends on the cloud core and mobile networks. This lessons-learned report presents the design rationale, positions FireOPS within emergency-response information systems research, and reports operational findings from a primary log extract of a major regional flood in 2024 and from the 2025 public app release. A feature-matrix comparison with DIVERA 24/7, blaulichtSMS and the LifeRescue prototype identifies COP, decision support and dual-ingestion redundancy as the main differentiators, and we map the operational gaps observed during the flood to concrete engineering responses, deployed or in active development.
[HC-14] Growing Ones AI Tools: A Paradigm of Practice-Grown Tools
链接: https://arxiv.org/abs/2611.06132
作者: Zeming Zhao
类目: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL); Software Engineering (cs.SE)
备注: 15 pages, 3 figures, 3 tables, 49 references
Abstract:Professionals have long used software built for them by others. This article reports the growth of a professional computer-assisted interpreting (CAI) tool built by its own user, a conference interpreter without programming training. Drawing on an eight-week version-controlled record with telemetry from 133 logged sessions and over 20,000 recognized utterances, the article traces how real conference work drove the tool through 62 commits and six versions. The record shows capabilities deployed on the day their need appeared and one strategy change that cut median end-to-end delay by over a third within one afternoon. The article then proposes the paradigm of practice-grown tools, defined by the shift from using to growing, the co-evolution of tool and practice through breakdown and repair, bounded local computation, and an assessment corollary, and it formulates three propositions toward tool-inclusive assessment of practitioners, from tool-making literacy to a compute-standardized format for certifying what they and their tools achieve together.
[HC-15] Fluidic Body: Swarm Robots as a Fluid Extension of the Hand
链接: https://arxiv.org/abs/2611.06068
作者: Michikuni Eguchi,Shigeo Yoshida,Mai Nishimura,Takefumi Hiraki
类目: Human-Computer Interaction (cs.HC); Robotics (cs.RO)
备注: accepted to ACM ISS 2026 (ISS track of PACMHCI)
Abstract:Swarm robots have attracted growing interest as a physical interface for human–computer interaction, since a swarm can reshape its formation to fit the objects and surroundings it works with. For a human operator, however, adapting the swarm’s behavior remains difficult. Most existing methods require the user to specify where the swarm should go, and they offer no intuitive way to change the collective’s behavior as each task demands. We present Fluidic Body, a system that lets a person operate a swarm of fluid-imitating robots as an extension of the hand. The system places a virtual fluid between the hand and the swarm, and the user adapts the swarm’s behavior to each task by changing the fluid’s properties through simple hand movements. When the user opens the hand, the swarm becomes soft enough to deform and flow around obstacles. Closing the hand turns it into a hard, cohesive body that can push objects. We evaluated the controller’s performance and its operability in two simulation experiments and one demonstration with physical robots. First, our controller handled both obstacle avoidance and object transportation better than an existing method that behaves the same way in every task. Second, in a VR user study, 17 participants completed the combined avoidance-and-transportation task faster and more reliably than with that method. They also reported a lower workload and a stronger sense of agency. Finally, this experience carried over to physical robots in a real-world demonstration (88 participants). These results suggest that treating a swarm as a fluid, a material people already understand, allows a robot swarm to work as a deformable extension of the human body.
[HC-16] Robots in the City: Persona-Based Perceptions of Urban Robots
链接: https://arxiv.org/abs/2611.05954
作者: Keonhee Jang,Kanghoon Lee,Seokwoo Yun,Namwoo Kim,Yoonjin Yoon
类目: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL); Robotics (cs.RO)
备注: 10 pages, 7 figures
Abstract:Pedestrians increasingly encounter robots in urban public spaces without interacting with them, but how robots change the perceived safety of a street remains unclear. We examine how robot presence and type affect perceived safety across urban contexts. We generate 18 images of Seoul streets that vary in time of day, road class, and robot condition (no robot, a delivery robot, or a humanoid). Two multimodal large language models (MLLMs), Qwen3.8-27B and Gemma4-26B, each conditioned on 500 synthetic personas, choose the safer image in each pair, and 25 human respondents complete the same task. Both models prefer streets without robots and favor the delivery robot over the humanoid. Robots are chosen more often at night and less often by older personas, whereas road class has smaller effects. In the model rationales, the delivery robot appears as a physical obstacle, whereas the humanoid appears as either a threat or a source of protection. Like humans, both models chose robots more often at night, but chose the humanoid over the delivery robot far less often, and Gemma showed a night effect for the humanoid about three times that of humans. These results show that the effect of a robot on perceived safety depends on when it appears, and that persona-conditioned MLLMs capture only part of human judgments.
[HC-17] Compiling Validating and Using Human Reference Atlas Cell Type Annotation Crosswalks
链接: https://arxiv.org/abs/2611.05888
作者: Nicole Vasilevsky,Aleix Puig-Barbe,Andreas Bueckle,Ellen M. Quardokus,Yongxin Kong,Divya Prasanth Paraman,Jie Zheng,Yongqun He,Yashvardhan Jain,Bruce W. Herr II,Katy Börner
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:The Human Reference Atlas (HRA) v2.5 includes 4,807 anatomical structures, 1,347 cell types, and 2,215 biomarkers (genes, proteins, lipids); 81 organs with 1,392 anatomical structures modeled in three dimensions (3D). Mapping new data into the HRA requires 3D spatial registration and/or cell type annotation (CTann). Many CTann tools exist but few use cell type names linked to Cell Ontology (CL), making comparisons across tools, tissues, and organs difficult. The HRA supports five transcriptomics CTann tools (Azimuth, CellTypist, popV, FR-Match, Pan-Human Azimuth), four sc-proteomics CTann tools (DeepCell Types, DeepCell Types-HuBMAP, Robust Image-Based Cell Annotator [RIBCA], and SpaTial cELl LeARning [STELLAR]), plus a Cell Distance Explorer (CDE) Spatial Omics collection covering previously published annotations across 12 tissue types with 47,349,496 cells. Assigned cell types differ in number, resolution, and coverage. This paper details how 10 crosswalks were constructed, validated, and used to compile a “Supertree” that can be used to compare CTann across organs and technologies in support of reference atlas constitution and usage. All data and code can be accessed at this https URL.
[HC-18] Sure Enough to Share? Overconfidence in GenAI Video Discernment is Associated with Sharing Intentions
链接: https://arxiv.org/abs/2611.05844
作者: Haoning Xue,Sijia Qian,Christopher Calabrese,Cuihua Shen,Laramie Taylor,Yixuan Wang
类目: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
备注:
Abstract:Generative Artificial Intelligence (GenAI) is blurring the line between synthesized and recorded videos, but less is known about people’s awareness of their discernment abilities and the consequences of miscalibration. This study examines three forms of overconfidence in GenAI video discernment (i.e., overestimation, overplacement, and overprecision) and their associations with source judgments and sharing intentions in an online study (N = 973). Participants evaluated videos depicting everyday scenes and achieved 62.5% accuracy. Three forms of overconfidence showed distinct patterns: participants overplaced themselves by 12.8 percentile points, but overprecision showed consistent positive associations with sharing intentions and the tendency to classify videos as human-made. Age, GenAI usage, and AI literacy showed consistent positive associations with all three forms of overconfidence. These findings highlight the importance of distinguishing the three forms of overconfidence to better understand GenAI engagement and incorporating metacognition into interventions to support informed media consumption.
[HC-19] BobFlow: Towards Human-Agent Collaborative Data Science Using Dataflows
链接: https://arxiv.org/abs/2611.05761
作者: Jiadong Bai,Yicong Huang,Chen Li
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注:
Abstract:Recent advances in large language models (LLMs) have established human-agent collaboration as a promising paradigm for data science. Effective collaboration requires a shared computational abstraction that enables agents to reason about processing logic while allowing human analysts to understand, guide, and refine the analysis. Although scripts written in Python or SQL provide precise control, their syntax and execution details can be difficult for analysts to understand and modify. In this paper, we argue that dataflows are an effective abstraction for human-agent collaboration in data science. We introduce BobFlow, a system that combines the dataflow abstraction with the ReAct paradigm. It enables agents to iteratively construct and execute dataflows for data science tasks while allowing analysts to inspect the analysis and provide feedback. Our user study shows that BobFlow helps analysts understand, inspect, and verify agent-generated analyses. A thorough experimental evaluation across proprietary and open-weight LLMs shows accuracy score gains of up to 18.6 points at just over 1/3 of a script-based ReAct agent’s inference cost.
[HC-20] Pairing Expert Critique with LLM -Driven Refinement: An Iterative Approach for Curating Quality AI Evaluation Datasets
链接: https://arxiv.org/abs/2611.05749
作者: Honson Ling,Kidus Admassu,Stephanie Rosenthal
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 20 pages, 7 figures, 9 tables
Abstract:Evaluating and improving domain-specific LLM responses requires high-quality reference data and rubrics, yet expert authoring is slow and expensive, and synthetic datasets may fail to capture practitioner preferences unless they are grounded in expert judgment. We present SME-C (Subject Matter Expert Critique), a human-AI workflow that captures expertise through lightweight critiques of candidate responses and complements it with LLM-driven analysis and refinement to iteratively curate response datasets and evaluation rubrics. We validate SME-C in the domain of financial business briefs, collecting 3,853 critiques from 411 finance professionals on 520 queries across two survey rounds. We evaluate two LLM-driven refinement strategies: one guided by the experts’ written critiques and one that infers critique solely from their preference votes. In a blind validation survey with 100 additional professionals, we found that both strategies significantly improved quality over the experts’ top-ranked baseline responses. The gains were concentrated among weaker original responses, and the two strategies excelled in different ways: critique-based refinement produced more domain-grounded improvements, whereas vote-based refinement was more effective at identifying and removing irrelevant content. The resulting rubrics and refined dataset support reproducible offline evaluation and benchmarking of LLMs, suggesting that structured expert critique is a viable alternative to expert authoring for building high-quality, domain-specific evaluation resources more quickly, at scale, and at lower cost.
[HC-21] he Male Abstraction Bias in Internet Culture and Large Language Models
链接: https://arxiv.org/abs/2611.05730
作者: Douglas Guilbeault,Bhargav Srinivasa Desikan
类目: Computers and Society (cs.CY); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注:
Abstract:Two foundational aspects of human cognition, abstractness and gender, are typically examined as separable. Here, we find they are systematically linked throughout the entire space of concepts spanning single (N = 40,000) and multi-word expressions (N = 67,000): across billions of words and millions of images on the internet – and across numerous measurement techniques – more abstract concepts are more associated with men, an effect which replicates in languages around the world. This pattern not only holds across conceptual domains, but it also maps onto grammatical structure: verbs and adverbs are more abstract and more male-coded, a system-wide trace of stereotypes accentuating agentic qualities in men. Furthermore, this bias emerges in human and AI responses in free association tasks, indicating its internalization and its algorithmic reproduction. The male abstraction bias may underlie and unify a range of stereotypes, evidencing gender as a core organizing dimension in conceptual space.
[HC-22] ACT: A Prag matic Approach to Qualitative Analysis with an AI Council
链接: https://arxiv.org/abs/2611.05702
作者: Rudrajit Choudhuri,Christian Bird,Carmen Badea,Anita Sarma
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: 12 pages
Abstract:Qualitative data can tell software teams what their stakeholders need, yet rigorously analyzing thousands of qualitative responses takes expert time that few teams can afford or justify. Teams therefore often face a false choice: leave the responses unanalyzed, or hand them to a large language model and accept its interpretation and judgment. In this paper, we take a pragmatic approach to AI-assisted qualitative analysis and present TACT (Traceable AI Council for Thematic analysis), a procedure that enables such analysis at scale while keeping teams in control of interpretation and judgment. TACT comes in two variants: TACT-Lite, the core AI council procedure, and TACT-Full, which adds adversaries (models from other model families that critique each model’s output) and a decision model that surfaces high-uncertainty code assignments. We evaluated TACT on an independent dataset of 256 Stack Overflow threads qualitatively coded by other researchers. Against their labels, TACT’s candidate codebook achieved an F1 of 0.75-0.77 and its final coding an F1 of 0.71-0.72. On a held-out sample, agreement between our own manual coding and TACT’s coding was Cohen’s \kappa of 0.79-0.85. TACT is also economical: model calls cost \ 13.77 (TACT-Lite) to \ 33.59 (TACT-Full) per 100 threads. Finally, drawing on our experience building and validating TACT, we identify pitfalls in AI-assisted qualitative analysis, describe safeguards to address them, and offer guidance for teams and researchers applying TACT in practice.
[HC-23] How Do Artificial Agents Make Decisions in Light of Visual Data? IEEE-VIS
链接: https://arxiv.org/abs/2611.05671
作者: Keyty Loza,Shayan Monadjemi
类目: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
备注: Presented at the Workshop on GenAI, Agents, and the Future of VIS, held in conjunction with IEEE VIS, 2026
Abstract:Vision-language models (VLMs) have demonstrated significant advances in reasoning with visual data, which is particularly important given that much of human knowledge is encoded visually in scientific articles and other sources. However, a gap remains in understanding how artificial agents, powered by VLMs, perceive and navigate decision-making under uncertainty in the presence of visual data. This study explores this gap by evaluating and analyzing autonomous decision-making processes. Building upon prior work in visualization literacy and decision theory, we conducted two complementary studies: one examining how agents perceive and quantify uncertainty encoded in visualizations, and another investigating how artificial agents make decisions based on visual data in the form of images. We analyzed perception accuracy and decision performance across models and visualization types. Our results show that GPT-4o demonstrates the highest accuracy in perceiving uncertainty in pie and bar charts among the tested models. For decision-making, autonomous agents showed the best alignment with rational decision theory when uncertainty was represented through circle charts, icon arrays, or text.
[HC-24] Contextual Pictograms for Repairing Multilingual ASR Errors
链接: https://arxiv.org/abs/2611.05637
作者: Mariana Hernandez,Hector Luna,Carlos Bruno,Luis Ricardo Santos,Elena Simoes
类目: Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
备注:
Abstract:Speech transcription errors can be difficult to recognize when competing word hypotheses are unfamiliar or visually similar. We describe an interface that combines alternative transcripts, replay and contextually checked pictograms for selected uncertain concepts. The system will initially support French and Spanish functional utterances. A counterbalanced study will assess whether pictorial feedback helps listeners recover the speaker’s intended action beyond text alternatives alone. We will also quantify semantic mismatch and additional interaction time, since an inappropriate image can reinforce the wrong interpretation rather than repair it.
[HC-25] Personalized Visual Support After Automatic Speech Recognition
链接: https://arxiv.org/abs/2611.05625
作者: Beatriz Calero,Duarte Neves,Carlos Bruno,Luis Ricardo Santos,Mario Egea
类目: Human-Computer Interaction (cs.HC)
备注:
Abstract:Many spoken-language interfaces stop at producing a transcript. This project examines how a second layer of optional text and pictorial support can be adapted to individual needs. We specify four presentation levels and a behavior-based selection policy driven by replay requests, response latency and recent comprehension. After expert review of the mappings, we will compare fixed and adaptive presentation using counterbalanced spoken instructions. The study is designed to measure comprehension, usability and which modality listeners rely on, not to claim that transcript embellishment uniformly benefits listeners.
[HC-26] Just the Facts: How Dialogues with AI Reduce Conspiracy Beliefs
链接: https://arxiv.org/abs/2611.05572
作者: Thomas H. Costello,Gordon Pennycook,David G. Rand
类目: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
备注:
Abstract:Personalized dialogues with AI substantially reduce conspiracy beliefs. To identify why, we conducted two experiments with U.S. adults recruited online (total N = 2,193) that removed possible causes one at a time. The debunking effect was robust to disclosing the AI’s persuasive goal, framing the exchange as a debate, instructing the AI to inform rather than persuade, halving its response length, altering how beliefs were measured, and replacing the dialogue with a single tailored message. Only instructing the AI to persuade without factual evidence and arguments substantially reduced the debunking effect. Content analyses of the AI’s output and participants’ own accounts linked larger reductions in conspiracy beliefs to factual argumentation rather than rapport, storytelling, or appeals to empathy. AI dialogues appear to reduce conspiracy beliefs primarily by supplying relevant, well-reasoned counterevidence. Our results suggest that humans might be similarly effective if they could supply comparable evidence.
[HC-27] Whose Transparency is it Anyway? Understanding Generative AI Transparency Needs and Gaps Across Stakeholders
链接: https://arxiv.org/abs/2611.05549
作者: Genevieve Smith,Sahiba Chopra,Natalia Luka,Min Kyung Lee
类目: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
备注:
Abstract:Transparency has become a central principle in AI governance, yet there is little agreement about what transparency means, for whom, and in what contexts. Drawing on interpretive flexibility within the Social Construction of Technology, this paper examines how different stakeholders understand and experience transparency in generative AI. This study draws from a global survey of 400 end-users, AI builders, and policymakers, complemented by a structured focus group with end-users, grounding the quantitative findings in qualitative insights from the group most underserved by current transparency practice. We present three key findings. First, all three groups share similar understandings of what transparency means and what it should afford, centered on ensuring accuracy and privacy. Second, gaps between the importance of transparency features and their perceived availability are substantially larger for end-users than builders, revealing how power shapes what gets built and who it serves. Only 28% of end-users report that current transparency practices reflect their needs very well, compared to 53% of builders and 49% of policymakers. Third, transparency is situationally constructed, priorities shift with context and task structure, with role-specific frames activating when context demands them and receding when it does not. We argue that meaningful AI transparency requires moving beyond a one-size-fits-all solution, including current emphases on AI disclosure and static documentation, toward pluralistic transparency infrastructures that reflect the needs of diverse users and stakeholders.
[HC-28] A Cracked Mirror: How People Perceive Gender Performance by Large Language Models
链接: https://arxiv.org/abs/2611.06393
作者: Faezeh Fadaei,Jenny Carla Moran,Taha Yasseri
类目: Physics and Society (physics.soc-ph); Human-Computer Interaction (cs.HC)
备注: Preprint, under review; main article and supplementary material
Abstract:How do large language models (LLMs) perform gender under zero-shot prompting, and how do people interpret those performances in conversation? Drawing on gender performativity, we examine how gendered meanings emerge through generated language, conversational context and human interpretation. In a preregistered randomised online experiment, 1,060 participants each had five short conversations with an LLM. We used minimal prompts to elicit performances from “very masculine” to “very feminine” and asked participants to rate their partner’s perceived gender after each conversation. Mean ratings shifted in the direction of the prompts, but ratings at the extremes were closer to the midpoint than the prompt labels implied. When prompted to perform as “neither masculine nor feminine”, the models were perceived as slightly masculine on average. Ratings varied across models and conversation topics. Women and men gave similar ratings at the extremes, but women rated the models as more masculine across the intermediate prompt levels. Participants’ explanations illustrate how they drew on writing style, emotional expression and conversational interests, sometimes questioning the stereotypes behind their judgements. These findings suggest that conversational AI acts as a cracked mirror of gender norms. Models reproduce recognisable cues, and people interpret and sometimes challenge the meanings those cues carry.
计算机视觉
[CV-0] From Pixels to Actions - Towards Native Vision-Language-Action Policies
链接: https://arxiv.org/abs/2611.06918
作者: Weichen Fan,Haiwen Diao,Penghao Wu,Siyuan Hu,Ziqi Huang,Tianqi Liu,Mingze Sun,Ziyi Wang,Zhongang Cai,Hanming Deng,Quan Wang,Lei Yang,Lewei Lu,Fangzhou Hong,Zhaoxi Chen,Dahua Lin,Ziwei Liu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Code link: this https URL
Abstract:Vision-language-action models are typically built on a simple premise: perception is first compressed by visual encoders, and action reasoning begins only afterwards. Yet manipulation often hinges on precisely the visual signals most vulnerable to this abstraction, e.g., local geometry, boundaries, depth, and subtle state changes. This raises a fundamental question: should action policies inherit a fixed visual endpoint, or directly access evolving visual computation? We introduce NEO-VLA, a structurally encoder-free vision-language-action framework. Images are tokenized by lightweight patch convolutional layers and evolve jointly with language throughout the backbone, allowing a flow-matching action expert to draw from instruction-conditioned visual representations at different depths. This simple change reveals a consistent pattern. Native architectures outperform encoder-based counterparts under controlled comparisons; full fine-tuning reduces but does not erase the gap, and additional scale alone cannot recover it. More strikingly, layer-wise probing shows that the action-critical signals often peak well before the final visual representation. Across diverse manipulation settings and distribution shifts, NEO-VLA consequently achieves stronger and more robust control. Together, these results suggest that embodied control benefits from access to visual representations across the computational hierarchy, rather than a fixed encoder output.
[CV-1] Perceptive Humanoid Motion from Simulation
链接: https://arxiv.org/abs/2611.06908
作者: Siyuan Hu,Abulikemu Abuduweili,Yue Meng,Shipeng Liu,Quentin Leboutet,Chen Chen,Matthias Müller,Vladlen Koltun
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Code and videos at: this https URL
Abstract:Humanoid robots have the potential to automate a wide range of physical tasks. Progress toward generalist humanoid autonomy, however, remains constrained by a data bottleneck: mapping pixels and free-form language to low-level joint commands requires paired supervision, and no corpus of such data exists at sufficiently large scale. Some approaches rely on human teleoperation and wearable capture devices, which require at least an hour of human time to produce an hour of data. We present HALO, a humanoid autonomy system powered by a pipeline that synthesizes vision-language-motion data across a thousand procedurally generated homes. The corpus scales with compute rather than human labor. HALO acquires a repertoire of physical skills, including sitting on chairs, navigating to named objects, stepping sideways through half-open doors, crouching and crawling beneath tables, and pushing boxes. It orchestrates these skills to solve composite, long-horizon tasks. Without ever seeing a real home during training, HALO transfers zero-shot to the physical world, achieving a 91% success rate across 360 trials in six unmapped real-world environments. See this https URL for code and videos.
[CV-2] Motion-guided Contrastive Regularization For Latent World Models
链接: https://arxiv.org/abs/2611.06901
作者: Yiwen Chen,Qiqi Hou,Jisoo Jeong,Yinhao Zhu,Huaizu Jiang,Amir Said
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:JEPA-style world models learn action-conditioned latent dynamics without reconstructing observations, but distributional regularization alone does not ensure that their representations preserve information needed for control. We identify static collapse, a failure mode in which the encoder preserves episode-specific static content while suppressing manipulation-relevant state changes. Because static content is temporally predictable yet diverse across episodes, this failure can persist despite regularizers such as SIGReg. A controlled study on OGBench-Cube shows that introducing randomized backgrounds substantially degrades planning performance while leaving the underlying manipulation task unchanged. To address this problem, we propose motion-guided contrastive regularization. Our approach compares predicted latent transitions with transitions to candidate observations from the same episode, reducing the usefulness of static background identity for distinguishing action outcomes. Optical-flow similarities provide soft contrastive targets, allowing candidates with similar motion to receive similar supervision. Combined with latent prediction and SIGReg, our objective encourages representations to preserve within-episode state distinctions and similarities between related transitions. Our method improves planning success over LeWorldModel (LeWM) from 52% to 78% on OGBench-Cube with randomized backgrounds and from 14% to 54% on LIBERO-Object, while maintaining comparable performance on four other visual-control benchmarks.
[CV-3] Occlusion for Free: Detecting Occlusion from the Disagreement Between Two Monocular Distance Cues NEURIPS2026
链接: https://arxiv.org/abs/2611.06899
作者: Chibueze Praise Onyenweaku,Toyibat Adele,Sikirulahi Opeyemi Abdulkareem,Obed Honour Eje
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at the NeurIPS 2026 GlobalSouthAI Workshop
Abstract:Knowing that a pedestrian is occluded is useful to a robot even when the detector cannot recover the hidden pixels, but occlusion estimators are normally trained on occlusion annotations that deployment settings rarely have. We show that occlusion can be read off two distance estimates a camera-only pipeline already computes. Apparent height is largely unchanged by occlusion, because the detected box still spans the whole person; monocular metric depth is pulled toward the occluder, because the nearest surface inside the box is what a depth map reports. Only their ordering matters, so no focal length, stature, or depth scale is needed, provided the depth model’s error is close to a pure gain, a condition that can be checked without labels. The score detects heavy occlusion on KITTI with AUC 0.909 (0.888 over all annotated pedestrians, including those the detector misses), requires no training and no occlusion labels, and is not a proxy for range: within fixed distance bands it holds AUC 0.894-0.925 while true distance alone sits near chance. On synthetic CARLA scenes with measured visibility it reaches 0.819, against 0.644 for distance alone. We arrived here after measuring three standard architectural responses to occlusion, depth fusion, occlusion-simulating augmentation, and attention, none of which improved detection outside its training distribution in the configurations we tested, and after finding that a detector trained on German roads loses half its accuracy on self-annotated street scenes in Lagos, Nigeria. Signals that need no local annotation are worth more in settings that benchmarks do not represent.
[CV-4] Procedural Adversarial Learning
链接: https://arxiv.org/abs/2611.06890
作者: Jack Nugent,Jia Deng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:For many computer vision tasks, neural networks can be trained to produce a low average error, yet they still fail on seemingly simple scenes. Reducing these failures requires first finding them, but they may not be in a fixed dataset. We study procedural adversarial learning (PAL), a framework of minimax training and adversarial evaluation which searches the input space of a procedural generator for difficult instances. The procedural generator is an interpretable program mapping a low-dimensional vector of control parameters (e.g. layout, lighting, pose) to a rendered image and its label. We introduce the Simple-PAT algorithm, which uses zeroth-order optimization in the procedural generator’s parameter space to propose adversarial training samples that are generated on demand. We also create PAL-Dojo, a collection of efficient procedural generators for surface normal estimation, image segmentation, and human pose estimation. Against a cross-model black-box adversary with 50M aggregate queries, Simple-PAT reduces worst-case error by a geometric mean of 3.15x compared to standard training and 2.63x compared to the strongest robust training baseline. Code is available at this https URL.
[CV-5] PixelSpline: Continuous Space-Time Video Super-Resolution via Endpoint-Anchored Shared Pixel Trajectories
链接: https://arxiv.org/abs/2611.06870
作者: Wanqi Yuan,Nianyi Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Existing continuous space-time video super-resolution methods often infer motion from only two endpoint frames and model the two motion directions independently, making them vulnerable to ambiguous correspondence and inconsistent trajectories. We introduce a continuous STVSR framework that combines long-context encoding, continuous HR feature lifting, and an endpoint-anchored PixelSpline. Long-range observations provide more reliable endpoint representations, while continuous upsampling constructs detailed appearance features directly on the requested HR grid. PixelSpline explicitly models each pixel’s motion as a single continuous trajectory shared by both endpoints. Using this trajectory, our method warps endpoint colors into a coarse intermediate frame and aligns the HR features for residual reconstruction. Experiments across spatial and temporal upsampling settings demonstrate that the proposed framework offers a favorable quality–efficiency trade-off.
[CV-6] CounterFactual Retrieval for Evidence-Aware Individual Animal Identification WACV2027
链接: https://arxiv.org/abs/2611.06864
作者: Ankit K. Upadhyay,Aleksandr Algasov,Ekaterina Nepovinnykh,Fedor Zolotarev,Tuomas Eerola,Heikki Kälviäinen,Charles V. Stewart,Lukas Picek
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 14 pages, 5 figures; accepted at WACV 2027
Abstract:Automated identification of individual animals is moving from small curated studies toward shared benchmarks and reusable foundation models. This progress is useful, but current evaluation leaves the evidence behind a match unresolved, i.e., high retrieval accuracy does not show whether a model used animal features or identity-correlated shortcuts. We address this with Counterfactual Foreground Retrieval (CFR), an evidence-aware evaluation protocol. CFR evaluates the same query-gallery task under different views: full image, foreground-only, and shortcut-only. The full-image measures standard retrieval performance; the foreground-only measures foreground-supported retrieval; and shortcut-only view measure how much identity signal is recoverable from background, or silhouette. This turns a retrieval score into an evidence profile: how well the model performs on the original images, how much of that performance remains when only the animal is visible, and how much identity signal remains recoverable when the animal is removed. We release around 0.5M masks for 34 datasets and code for mask generation and CFR evaluation.
[CV-7] Visual Token Routing: Learning to keep tokens that matter
链接: https://arxiv.org/abs/2611.06854
作者: Pranjal Aggarwal,Sean Welleck
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Computer-use agents (CUAs) have recently achieved superhuman performance on several long-horizon benchmarks. However, their computational cost limits their use in long-horizon tasks and real-time domains such as video games. A key bottleneck is representing each screenshot as thousands of visual tokens, which accumulate into long contexts as agents reason over extended interaction histories. We observe that most of these visual tokens are either semantically irrelevant to the agent’s next action or temporally redundant across screenshots. Building on this observation, we introduce Visual Token Routing (VTR), a joint training framework for learning sparse visual representations that preserve the information needed for action prediction. In VTR, we train the vision-language model together with a lightweight spatio-temporal router that examines each image region in the current and previous screenshots and decides whether its visual token enters the language model. We optimize the router using an adaptive loss to keep the number of visual tokens as small as possible while minimizing performance degradation. With just 2,000 training steps, VTR matches the performance of the full-resolution model on OSWorld, CUA-World, and video game action prediction, while using 2.5-5.8x fewer visual tokens. These savings reduce inference cost by up to 3.3x and speed up prefill by up to 2.9x. VTR also outperforms lowering the image resolution and other baselines for dropping tokens, by up to 12.8 points on OSWorld. Overall, VTR enables computer-use agents to process only the visual tokens they need, allowing them to be used in long-horizon tasks and real-time domains at a fraction of the computational cost. Code and models are available at this https URL
[CV-8] rainspotting: Learning Object-Centric 3D Reconstruction from Found Photos
链接: https://arxiv.org/abs/2611.06849
作者: Hanyu Chen,Noah Snavely
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: project page: this https URL
Abstract:State-of-the-art feedforward reconstruction models perform poorly on in-the-wild object-centric photo collections. They are trained largely on single-session object captures whose images vary in viewpoint while the object and its environment remain nearly fixed. Public photo archives contain a different form of multi-view data. They hold found photos of the same object or design: photos taken independently by many people across years, environments, cameras, and sometimes physical units. In this work, we introduce Trainspotting, a dataset of 12,068 objects and 311,886 photos from Wikimedia Commons. We reconstruct each object to recover camera poses, multi-view stereo depth, and object masks. We categorize the objects into three regimes: objects fixed at one site, objects that move between sites, and designs shared by many physical units. Evaluating recent feedforward reconstruction models on Trainspotting shows that they fail most severely in the last two regimes. Fine-tuning VGGT- \Omega on Trainspotting significantly improves camera pose, depth, and 3D geometry on held-out objects. We observe similar improvements on the independent NAVI in-the-wild benchmark, which contains photos of each object in different environments. Performance on conventional benchmarks remains largely unchanged.
[CV-9] Where to Connect? Boosting MLLM s via Dynamic Gated Pathways across ALL ViT and LLM Layers NEURIPS2026
链接: https://arxiv.org/abs/2611.06844
作者: Yingying Yan,Jiaqi Tang,Wei Wei,Qianzhou Wang,Lei Zhang,Jianmin Chen,Yuyang Xia,Botong Geng,Jinjian Wu,Qifeng Chen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 24 pages (10 pages main text + references + appendix), 15 figures. Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)
Abstract:A central yet under-explored question in Multimodal Large Language Models (MLLMs) is where to connect—how visual information from a hierarchical vision encoder should be wired into the layer-wise semantics of a large language model. Mainstream MLLMs inject a fixed single-layer visual representation into the LLM and treat all LLM layers as a single consumer, ignoring that different layers demand different visual granularities. Recent multi-layer fusion methods either compress hierarchical features on the encoder side or rely on predefined sparse or hierarchical layer-to-layer connections, and therefore stop short of modeling adaptive cross-layer interactions between the ViT and LLM hierarchies. We address this gap with DGP (Dynamic Gated Pathways), a routing mechanism that establishes all-to-all connections between every ViT layer and every LLM layer, with each LLM layer dynamically gating its preference over visual layers conditioned on its own semantic state. Built on top of LLaVA-1.5-7B and 13B, DGP delivers consistent gains over the LLaVA-1.5-7B baseline (e.g., +4.2 on SQA ^I , +3.3 on MM-Vet, +3.1 on VizWiz and LLaVA ^W ) and surpasses prior multi-layer visual fusion methods on most benchmarks. Layer-wise routing analyses further show that different LLM layers do select different visual granularities on demand, supporting both the effectiveness and interpretability of the proposed dynamic gated pathways. The code is available: this https URL.
[CV-10] Same Layout Different Count: What Number Words Do Past Ten in Text-to-Image Models
链接: https://arxiv.org/abs/2611.06840
作者: Liang Yue,Eric A. Vance
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 43 pages, including 32 pages of supplementary material; 14 figures, 44 tables
Abstract:Ask SD3.5-Medium for twelve of an object and then, from the same starting noise, for twenty-five. The picture keeps its layout, but the count changes about as much as it would under a new random seed, and in no consistent direction. We study what number words do once a request passes about ten objects, in five open text-to-image models. Holding the object and the starting noise fixed, we change one thing at a time, and people who never see the request count 3,088 of the images. Every model follows the number for small crowds, but only up to a point, its reach: SD3.5-M and FLUX.1-dev stop following it soon after ten, while FLUX.2 [klein] and Qwen-Image follow it loosely up to twenty-five. Beyond SD3.5-M’s reach, the coarse part of the starting noise fixes the composition of the picture but not how many objects it holds: redrawing only the noise finer than an eighth of the image keeps the layout as well as a new number does, and already moves the count two-thirds as far. The number word still reaches the denoiser and acts in its first steps, but nothing we tried gives it a clear direction. Shown instead as a noised grid in the starting latent, the count is drawn: people find it in 59% of images, against 4% when it is only told. Past ten, an exact-match score barely separates a model that follows the number loosely from one that ignores it, and a relative-error score can even prefer a constant guess of sixteen. A same-noise test with a few requests per object tells them apart.
[CV-11] Context-Aware Autoregressive Image Generation for Emerging Reasoning Properties NEURIPS2026
链接: https://arxiv.org/abs/2611.06838
作者: Jixuan Ying,Haoyu Liu,Timing Yang,Tingyu Zhu,Predrag Neskovic,Anand Bhattad,Cihang Xie,Zeyu Zheng,Alan Yuille,Feng Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026. 16 pages, 4 figures
Abstract:Autoregressive image generation has recently emerged as a competitive visual generation paradigm, yet its generation order is often predefined or heuristic, such as next-patch prediction, random-order prediction, or one-step generation for specific scales. We observe that these strategies can cause models to overlook important structural dependencies in images, thereby limiting both structural robustness and reasoning capability. In this work, we propose a context-aware autoregressive generation framework, where the model dynamically determines the next token location based on global context and current local information. Rather than following a fixed or random order, our method allows the generation process to adapt to image structure and semantic dependencies. Extensive experiments show that context-aware generation significantly improves structural robustness across different image AR paradigms, including masked autoregressive and scale-wise autoregressive models. Beyond quantitative improvements, we further observe emerging reasoning properties: the model can better understand complex semantic instructions and generate images that more faithfully follow the given guidance. These findings suggest that generation order is a critical yet underexplored factor in image autoregressive modeling, and that context-aware decoding can unlock stronger reasoning-aware visual generation capabilities. Code is available at this https URL.
[CV-12] Open-HumanScale: An Open Recipe for Scaling Human and Robot Experience for Embodied Foundation Models
链接: https://arxiv.org/abs/2611.06817
作者: Juncheng Ma,Jiaxin Li,Jiankai Tu,Bo Liang,Guodong Lv,Yufan Deng,Junzhe Tang,Liangzhu Yang,Chen Bian,Zicong Cao,Baohao Zhu,Jianxin Bi,Tengfei Yan,Yinyuan Zhang,Kewei Zhang,Xinglong Chen,Kaiqi Chen,Duomin Wang,Xiaotian Tang,Liang Tan,Li Li,Yuan Xu,Haosen Yang,Yuqi Wang,Jingfei Xia,Songyan Huang,Zhiyang Dou,Zhen Dong,Enze Xie,Tat-Seng Chua,Daquan Zhou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: GitHub: this https URL Project website: this https URL
Abstract:Egocentric human video offers a scalable source of manipulation priors for robot learning, yet translating raw data scale into generalizable robot policies demands rigorous data curation and training recipes. We introduce Open-HumanScale, the first fully open recipe for scaling egocentric pretraining to 50,000 hours, anchored by Open-HumanScale-50K—a curated egocentric video corpus densely annotated with hand motion and natural language. Leveraging this corpus, we systematically investigate pure human pretraining via closed-loop simulation evaluations and uncover an empirical log-linear scaling law linking pretraining data hours to downstream robot success rates. Grounded in these findings, we develop HumanWAM-Vanilla, a world–action model pretrained exclusively on 50k hours of egocentric human video. Upon robot post-training, HumanWAM-Vanilla reaches 92.08% success on RoboTwin 2.0 Clean2Clean and 68.88% on Clean2Random, alongside a 14.17% mean success rate and a 19.00 mean progress score on RoboDojo. To push the capability frontier further, we explore joint human–robot pretraining with HumanWAM-Goal, a visual-goal-conditioned world–action model. Hybrid data scaling unlocks broad zero-shot generalization across both simulation and physical benchmarks, propelling post-trained RoboTwin 2.0 performance to new state-of-the-art levels: 92.70% on Clean2Clean and a substantial leap to 84.60% on Clean2Random. On physical robots, HumanWAM-Goal further exhibits emergent embodied in-context learning, dynamically adapting its manipulation behavior to unseen visual goals without parameter updates. To support reproducible scaling research in embodied AI, we open-source the Open-HumanScale-50K corpus, data curation and pretraining pipelines, alongside the pretrained and post-trained HumanWAM checkpoints, inference code, and simulation evaluation suites.
[CV-13] VA-TDM: A Unified View for Improving Few-Step Diffusion Models by Velocity Adversarial Distillation
链接: https://arxiv.org/abs/2611.06796
作者: Yihong Luo,Tianyu Pang,Tianyang Hu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project Page: this https URL Code: this https URL
Abstract:Few-step distillation offers efficient generation with diffusion and flow models, but preserving quality with only a few sampling steps remains challenging. Distribution-matching methods via score distillation provide a strong foundation, yet they rely on accurately tracking the student’s changing distribution. At intermediate generation steps in few-step sampling, student-score estimation can be difficult, and commonly used estimators may describe a different distribution from the one used to update the student. Flow matching allows a more flexible approach to supervision through velocity fields constructed from the student’s sampling process. Building on this perspective, we introduce Velocity Adversarial Distillation (VAD), which uses a learned critic to compare student and teacher fields and turn their disagreement into feedback for learning. VAD unifies score distillation, Fisher-based matching, and moment matching, also explaining existing heuristic MMD updates and delivering new training objectives. We further develop VA-TDM, which follows trajectory distribution matching: providing supervision for partial denoising along the student trajectory. Beyond this, VA-TDM combines transition-aware online velocity learning and dual-form VAD supervision across intermediate student transitions. Extensive results show that VA-TDM is powerful for learning few-step generators, achieving state-of-the-art four-step performance in both text-to-image and text-to-video generation across various backbones and scaling effectively to 33B MiniMax-H3. In particular, our VA-TDM-H3 notably outperforms the few-step Fast-H3 and is closer to the original many-step H3 according to human preference.
[CV-14] Semantic Residual Vector Quantization for Sign Language Generation WACV2027
链接: https://arxiv.org/abs/2611.06781
作者: Chang Liu,Ke Han,Niccolò Biondi,Gloria Isotton,Elisa Ricci,Davide Talon,Nicu Sebe
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV 2027
Abstract:Sign language generation (SLG) maps written sentences into signing motion while preserving both lexical content and fine-grained articulation. This requires a model to decide what signs should be expressed and to realize them through precise handshape, trajectory, facial expression and body coordination. Existing pose-based systems and reconstruction-only motion tokenizers often entangle these two decisions, leaving the generator without an explicit text-aligned intermediate target for what to sign before refining how to sign it. To address this issue, we introduce Semantic Residual Vector Quantization (SRVQ), a framework for SLG that frames generation as a two-level process moving from semantics to kinematics. The newly introduced SRV Quantizer learns a hierarchy in which the first residual level encodes sign content and subsequent levels refine pose, handshape, trajectory, and body coordination, all within a single quantizer and without separate semantic representations or external planning stages. Through word-aware conditioning for hierarchical code generation, a masked semantic generator predicts semantic codes and a kinematic generator completes the residual codes for full SMPL-X motion synthesis. We evaluate using retrieval and FID in a shared text-motion embedding space, alongside standard geometric metrics. On CSL-Daily and Phoenix-2014T, SRVQ achieves R@1 of 82.3 and 70.3, outperforming the previous SOTA by 30 and 46 points, respectively.
[CV-15] Benchmarking Fine-Grained Visual Reasoning for Remote Sensing in Urban Management
链接: https://arxiv.org/abs/2611.06770
作者: Jiawei Zhang,Rizhao Cai,Dayan Guan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by IEEE AP-GARSS 2026; 4 pages, 3 figures, 3 tables
Abstract:Fine-grained visual reasoning is essential for remote sensing in urban management, yet remains underexplored in the evaluation of multimodal large language models (MLLMs). We introduce UMBench, a fine-grained urban remote sensing benchmark built from unmanned aerial vehicle (UAV) images for multiple urban management tasks such as traffic monitoring, urban construction, environmental quality, and disaster warning. Evaluations on UMBench reveal that current MLLMs often rely on scene-level priors and overlook fine-grained visual evidence. To address this issue, we further propose a novel evidence-first chain-of-thought method that directs models to inspect fine-grained visual evidence in a training-free manner. Experiments show that our method can improve the performance of both open-source and commercial MLLMs on UMBench.
[CV-16] Route First See Later: Enabling Vision Agents to Reason Scientifically by Separating Routing Policy from Action Spaces
链接: https://arxiv.org/abs/2611.06760
作者: Shubhabrata Mukherjee,Jeffrey Donatelli,Daniela Ushizima
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Scientific image-analysis workflows require agents to select among diverse capabilities like description, detection, counting, and segmentation. Current multi-tool architectures conflate intent interpretation with visual execution, masking whether a failure arises from flawed reasoning (misrouting) or failed perception (invalid tool execution). This lack of auditability renders agents unreliable for high-stakes scientific automation. We introduce ZETA, an extensible architecture that operationalizes the decoupling of routing policy from action execution. ZETA resolves scientific intent before invoking specialized visual tools, treating measurement as a deterministic operation grounded in explicit geometry rather than generative output. This architecture exposes the “black box” of agent decision-making, allowing routing correctness and downstream execution to be evaluated as independent failure modes. To our knowledge, we provide the first controlled study in scientific vision-agent routing that isolates explicit routing-policy specification while holding the legal action interface fixed. Evaluation on a zero-shot benchmark of 51 scientific-intent queries shows that explicit routing semantics can substantially improve routing accuracy, particularly at semantically adjacent action boundaries, although improvements vary across models. For Qwen 3.8, explicit routing-policy specification improves segmentation routing accuracy from 6.2% to 75.0% and detection routing accuracy from 60.0% to 100%. ZETA’s execution traces make routing decisions and downstream artifacts separately observable, providing a basis for distinguishing intent-selection errors from visual execution failures. ZETA establishes the necessary infrastructure to isolate execution errors from interpretive ones, a critical requirement for deploying reliable scientific vision agents. Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.06760 [cs.CV] (or arXiv:2611.06760v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.06760 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-17] EvoWorld: Towards Out-of-Sight World Evolution in Generative Video Models
链接: https://arxiv.org/abs/2611.06748
作者: Yequan Bie,Dengyang Jiang,Zhixuan Chen,Xiaozhuan Liang,Haoxuan Che,Quande Liu,Xi Chen,Hao Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Generative video models are increasingly used as visual world simulators, yet their primary strength lies in rendering plausible pixels rather than reasoning about how a world should evolve. Coherent world evolution requires more than visual fidelity: when an event occurs, a model must infer its consequences for the entities in the world and then render the resulting state. For example, after a wild beast appears, the model should not only depict the wild beast, but also infer whether a nearby person should flee, take cover, or fight back based on the realized scene. Existing diffusion-based video models lack an explicit mechanism for maintaining world state and reasoning over such event-conditioned consequences, often producing visually plausible but behaviorally static or causally inconsistent outputs. To make this reasoning gap concrete and measurable, we propose a new problem, Out-of-Sight Causal Evolution, which requires a model to render an event, reason about its consequences for the evolving world state, and reveal a causally plausible outcome as the world continues to evolve, even out of sight. As a baseline reference solution, we introduce EvoWorld, a brain-renderer collaboration framework in which a vision-language model serves as the world brain: it observes generated segments, maintains an evolving account of the world, and reasons about plausible state transitions, while a segment-autoregressive video renderer realizes them visually. We further introduce EvoWorld-Data and EvoWorld-Bench for training and evaluating event realization, causal reactions, and revisit consistency. Experiments show that EvoWorld substantially improves event-conditioned causal evolution over recent models while preserving visual quality and consistency, establishing a reference point for reasoning-guided, autonomous world evolution beyond the current field of view.
[CV-18] ailSteer: Trajectory-Tail Steering for Drift-Resistant Multi-turn Image Editing
链接: https://arxiv.org/abs/2611.06741
作者: Ruikang Li,Zixuan Chen,Jiarui Wu,Mingde Yao,Yiyuan Zhang,Yujin Wang,Hanming Deng,Lewei Lu,Tianfan Xue
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 16 pages, 9 figures. Project page: this https URL
Abstract:Strong single-turn editing performance does not guarantee reliable multi-turn interaction: as edited outputs are repeatedly reused, small unintended changes can accumulate into severe visual drift. This long-term drift is often mitigated through training on multi-turn data or cross-turn rollouts, potentially altering pretrained priors, or through specialized inference-time mechanisms that are difficult to scale to foundation models. We introduce TailSteer, a final-stage training framework that improves multi-turn robustness while retaining the generative and editing capabilities acquired through earlier training. TailSteer follows a simple principle: keep early denoising dynamics close to those of the base model and concentrate preservation corrections near the trajectory tail, where fine-grained deviations can be suppressed. These corrections are learned from the editor’s own on-policy denoising rollouts, using unchanged source content as a preservation reference without requiring curated ground-truth editing pairs. Experiments across the FLUX, SenseNova, and Qwen families show that TailSteer substantially reduces cumulative drift over 15-20 editing turns while largely preserving base-model performance on extensive text-to-image and single-image editing benchmarks. These results demonstrate that long-horizon editing robustness can be unlocked in foundation models after TailSteer, rather than by relearning their existing capabilities.
[CV-19] Coupled Scene Graph Anticipation: When With-Constraint Recall Hides Structural Inconsistency
链接: https://arxiv.org/abs/2611.06738
作者: Zeno Nongmaithem,Anushttha Srivastava,Anshit Sinha,Apoorv Sisodia,Suman Chowdhury,Prerana Mukherjee,Rajesh P. Barnwal
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at ICVGIP 2026 (Indian Conference on Computer Vision, Graphics, and Image Processing), December 21-24, 2026, Kolkata, India. 9 pages, 2 figures, 4 tables. Code: this https URL
Abstract:Scene Graph Anticipation (SGA) predicts object relationships in future video frames. Continuous-time SGA models evolve each subject-object pair with an independent Neural ODE, so relationships sharing an entity cannot influence one another during integration, and nothing prevents mutually exclusive predictions, such as “sitting on” and “standing on”, for the same pair. We introduce CoupledSGA, which links such relationships through a Relationship Interaction Graph with graph attention inside the ODE vector field, and adds a Structural Consistency Loss ( \mathcalL_SC ) based on a matrix of mutually exclusive predicates. Developing CoupledSGA revealed a limitation of the evaluation protocol. With-Constraint evaluation keeps one predicate per group for each subject-object pair, so its R@K and mR@K cannot capture the within-group violations targeted by \mathcalL_SC : we prove they are unchanged whenever each group’s winning predicate and its score are held fixed. We further show that the gradient of \mathcalL_SC depends on predicate scores rather than ground-truth labels: for two-predicate rules it amplifies the ordering the model already holds, and when all scores in a group are equal it suppresses predicates in more exclusion rules more strongly, explaining why the penalty can lower recall while improving consistency. On Action Genome, activating \mathcalL_SC after two warm-up epochs reduces R@20 from 42.21% to 35.18% while lowering the No-Constraint contradiction rate from 45.92% to 4.07% (matched-budget R@20: 40.48% CoupledSGA vs. 44.78% SceneSayerODE). CoupledSGA is a testbed, not a state-of-the-art claim: With-Constraint recall captures the cost of enforcing structural consistency but not the improvement. We recommend reporting No-Constraint results alongside a contradiction rate for methods that constrain predicate co-activation. Comments: Accepted at ICVGIP 2026 (Indian Conference on Computer Vision, Graphics, and Image Processing), December 21-24, 2026, Kolkata, India. 9 pages, 2 figures, 4 tables. Code: this https URL Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG) ACMclasses: I.2.10; I.4.8 Cite as: arXiv:2611.06738 [cs.CV] (or arXiv:2611.06738v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.06738 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-20] Dexterous Robotic Hands in the Foundation-Model Era: A Full-Stack Survey of Hardware Data Cross-Embodiment Transfer Learning and Evaluation
链接: https://arxiv.org/abs/2611.06724
作者: Ji Pei
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 45 pages, 8 figures, 21 tables. Structured narrative review of hardware, data, cross-embodiment transfer, foundation models, tactile sensing, and evaluation for dexterous robotic hands
Abstract:Dexterous robotic hands are increasingly coupled to large-scale human data and foundation models, yet progress in hardware, representation learning and evaluation is often assessed separately. This survey reviews work from January 2024 to early November 2026, with earlier studies included to establish technical lineages. We organize the literature into five interdependent layers: hardware; data collection and curation; cross-embodiment transfer; learning and foundation models; and evaluation and applications. Human hand-object understanding and tactile sensing provide two cross-cutting perspectives. A system-level architecture, comparative tables and five case analyses connect design choices across these layers. We distinguish within-study experimental comparisons from heterogeneous corpus statistics, vendor specifications and company-reported results. Three conclusions emerge. First, the value of human data depends on kinematic, visual and contact alignment, not recording volume alone. Second, shared action representations and foundation policies do not remove embodiment-specific constraints on force, latency and contact. Third, task completion is insufficient to establish physical correctness or deployment readiness. Evidence remains limited for zero-shot transfer of contact-rich skills between morphologically different hands, long-duration reliability, and cross-sensor tactile generalization. We identify eleven open problems and propose reporting interfaces and controlled comparisons that can make full-stack progress more measurable. The review is a structured narrative synthesis rather than a pooled performance ranking or an exhaustive systematic review.
[CV-21] Agent -Driven Generative Priors for Pathology Image Synthesis and Segmentation
链接: https://arxiv.org/abs/2611.06701
作者: Sicheng Yang,Rui Tang,Haipeng Zhou,Hongqiu Wang,Zhaohu Xing,Jianda Ma,Lie Dai,Guang Yang,Lei Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Code available at: this https URL
Abstract:Deep learning methods for pathology are constrained by the limited availability and high cost of large-scale pixel-level annotations. Synthetic data can alleviate this limitation, but existing pipelines often lack reliable quality control and remain weakly connected to downstream analysis. To address these issues, we propose a three-stage framework that integrates automated data screening, generative representation learning, and efficient adaptation to downstream tasks. First, we develop a pathology agent system to assess image quality and filter samples, resulting in TCGA-200K, a dataset containing 217,030 pathology patches selected from the public TCGA repository. Second, we train a generative model on the screened data. In addition to image synthesis, its backbone learns structural and semantic representations that can be transferred to downstream tasks. Third, we adapt the pretrained generative backbone to semantic segmentation using a lightweight task-specific head. We further introduce PathSeg-380K, a semantic segmentation dataset with expert annotations for approximately 380,000 cells across 12 cell types. Experiments on multiple pathology datasets show that the proposed framework achieves state-of-the-art segmentation performance. Our code is publicly available at this https URL.
[CV-22] Sekai: Incremental Event Memory for Generalized Reasoning Video Segmentation
链接: https://arxiv.org/abs/2611.06685
作者: Zhengtao Liu,Jinrong Xu,Shiu-Hong Kao,Huixin Sun,Yicong Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 20 pages, 6 figures. Code, example annotations, and evaluation scripts: this https URL
Abstract:Online reasoning video segmentation (Online RVS) requires a model to segment query-referred objects as a video streams in, using only the frames observed so far. Existing formulations largely assume a fixed target set throughout the video, overlooking a fundamental property of streaming reasoning: as new events occur, the objects that satisfy a query may change over time. To address this, we propose a new task, Generalized Reasoning Video Segmentation (GRVS), where the target set depends on query-relevant events accumulated over the observed video history. As new evidence accumulates, the target set can evolve from empty to one or multiple objects, requiring a model to continuously maintain target validity while localizing the valid objects currently in view. To evaluate GRVS, we build Isekai, a diagnostic benchmark of 60 videos and 197 queries across 11 temporal query families, with 12,569 scoring points whose target sets are causally annotated over time. We further propose Sekai, an online framework with two complementary components: an Event Memory Reasoner (EMR) that maintains query-relevant events and evolving target states in structured memory, and a Memory-Grounded Decoder (MGD) that identifies valid targets from this memory and localizes them in the current frame. Experiments show that Sekai significantly outperforms existing online RVS methods, achieving 50.33 JF and improving over the strongest baseline by 3.2 points. Code, example annotations, and evaluation scripts are available at this https URL.
[CV-23] CardioDrift: Automatic Cardiac Motion Tracking on a 3D Atlas Mesh via Shape-aware Coherent Point Drift with Slice Shift Correction
链接: https://arxiv.org/abs/2611.06664
作者: Reza Akbari Movahed,Abuzar Rezaee,Shatha Alamri,Edmond S.L. Ho,Fani Deligianni,Ali Gooya
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Automatic cardiac motion tracking from cine cardiac magnetic resonance (CMR) imaging is essential for quantifying myocardial motion abnormalities associated with cardiovascular disease. Existing methods fall into two groups: those that represent motion without anatomical correspondence across subjects and therefore limit inter-subject motion assessment; and those that represent motion with globally corresponding points, which typically require a time-consuming and error-prone semi-automatic step to fit a cardiac atlas to the subject space. To address these limitations, this study proposes CardioDrift, a fully automatic, atlas-based probabilistic framework that represents subject-specific cardiac motion on a 3D atlas mesh with globally corresponding points. The Coherent Point Drift (CPD) formulation is generalised to jointly register the atlas surfaces of multiple cardiac regions to the anatomically corresponding contours extracted from cine CMR segmentation maps throughout the cardiac cycle. Slice-shift correction is embedded in the same probabilistic model, estimating slice-wise translations that suppress the misalignment artefact limiting atlas-fitting accuracy. Evaluation on 1301 randomly selected UK Biobank subjects and the MM dataset, CardioDrift attains Dice similarity coefficients (DSCs) of 92.23\pm8.70% , 82.11\pm7.56% , and 88.69\pm4.13% for the left ventricle (LV), myocardium (Myo), and right ventricle (RV), exceeding the strongest baseline on every distance and overlap metric of both datasets.
[CV-24] Dynamical State-Space Diffusion for Spatiotemporal Prediction
链接: https://arxiv.org/abs/2611.06652
作者: Woojun Lee,Hankyeol Kim,Pilsung Kang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Diffusion models for temporal prediction typically corrupt each future frame with independent noise, leaving temporal dependence entirely to the denoiser. We propose Dynamical State-Space Diffusion (DSSD), which learns the temporal structure of the corruption process itself. DSSD builds a temporal kernel from the damped oscillatory modes of a stable diagonal state-space model, whose learned decay rates and frequencies let the corruption capture multiple timescales and oscillatory correlations, while the denoiser architecture remains unchanged. Running this model on a finite clip, however, requires the state before the first frame, which is unobserved; starting from a zero state leaves the early frames with less noise than the diffusion schedule prescribes. We avoid this by using the Cholesky factor of the model’s stationary covariance as the temporal operator. We prove that this operator is causal, exactly invertible, keeps each frame’s noise variance equal to that of standard diffusion, and reduces to the identity at the clean end of the process, and that it is exactly the Kalman innovations form of the state-space model. The denoiser is trained to predict the temporally correlated noise, and sampling removes and reapplies the temporal mixing at each step. On scientific-field and robot-video datasets, DSSD outperforms diffusion baselines that share its denoiser, and the evolution of its learned modes during training shows that the temporal operator adapts to each dataset.
[CV-25] Human Motion Prediction via Continual Prior Compensation
链接: https://arxiv.org/abs/2611.06637
作者: Jianwei Tang,Jian-Fang Hu,Tianming Liang,Xiaotong Lin,Jiangxin Sun,Wei-Shi Zheng,Jianhuang Lai
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Human Motion Prediction (HMP) aims to predict future human poses at different moments according to observed past motion sequences. Previous approaches mainly treated the prediction of different temporal moments as a single prediction task and learned the predictions of varied moments simultaneously, which would encounter a main limitation: the learning of short-term predictions (referring to near-future'' prediction) could be hindered by the predictions of long-term (referring to far-future’’ prediction) motions. In this paper, we develop a novel temporal continual learning framework called Continual Prior Compensation (CPC) to progressively train HMP models, in which we divide the prediction task of motions corresponding to varied temporal moments into several subtasks and train the model in a multi-stage manner. To mitigate the prior information forgetting in the progressive training, we further introduce a learnable random variable Prior Compensation Factor (PCF) to explicitly measure the prior knowledge loss. We theoretically show that the PCF can be efficiently learned together with the model parameters by minimizing a reasonable upper bound of the objective function. The proposed CPC is further enhanced to estimate the prior information loss for each subtask and a new framework called Continual Prior Compensation++ (CPC++) with Fine-Grained Prior Compensation Factor (FGPCF) is finally developed. Our CPC and CPC++ frameworks are quite flexible and can be easily integrated with different HMP backbone models and adapted to various datasets and applications. Extensive experiments on three HMP benchmark datasets using multiple SOTA HMP backbones (PGBIG, siMLPe, MotionMixer, and LTD) demonstrate the effectiveness and flexibility of our frameworks.
[CV-26] Uncertainty-Driven Efficient Communication for UAV-Based Collaborative Perception
链接: https://arxiv.org/abs/2611.06635
作者: Jiuwu Hao,Yuting Wan,Liguo Sun,Zao Zhang,Yueyang Wu,Pin Lv
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to IEEE SMC 2026
Abstract:Unmanned aerial vehicle (UAV)-based collaborative perception enables multiple agents to share complementary perceptual information, overcoming limitations such as occlusion and limited sensing range. Communication efficiency is a critical bottleneck, as UAVs operate under stringent bandwidth and computational constraints. However, existing spatial selection methods focus predominantly on foreground regions to reduce missed detections, neglecting perceptually ambiguous regions and leaving false detections unresolved. We propose UCP, an uncertainty-driven framework that characterizes collaborative value through observational uncertainty, identifying sensor-limited regions, and discriminative uncertainty, capturing model-level ambiguity near the decision boundary. These estimates drive spatial request and supply maps for efficient inter-agent communication, and the received messages are aggregated via U-Fusion, an uncertainty-guided cross-attention module. Experiments on three benchmarks across 3D object detection and trajectory prediction demonstrate a superior performance-bandwidth trade-off with high interpretability and strong generalizability.
[CV-27] rick-or-Treat: Adversarial Evaluation of Multimodal Medical Reasoning with Synthetic Images
链接: https://arxiv.org/abs/2611.06631
作者: Oishi Banerjee,Eujin Lee,Sung Eun Kim,Alexandra N. Willauer,Christine Y. Zhou,Zhizhou Yang,Pranav Rajpurkar
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:We used medical image synthesis to create a challenging multimodal evaluation framework that exposes hallucination behaviors in frontier models. Evaluating multimodal reasoning has historically been difficult because existing medical datasets rarely pair different image modalities, such as radiographs and external clinical photographs. To address this gap, we introduce Trick-or-Treat, an adversarial multimodal medical benchmark that pairs real chest radiographs with synthetic clinical photographs to challenge models’ cross-modal reasoning. For each of a wide range of conditions, we created two pairs of images, where the pairs share a single chest radiograph but contain different external photographs. In matched pairs, the chest radiograph and photograph both contain findings suggesting the same disease (e.g., varicella pneumonia suggested by lung opacities and crusted skin lesions seen in chickenpox). In mismatched pairs, the photograph depicts an alternative clinical finding that would not lead to the same unifying diagnosis (e.g. replacing crusted skin lesions with ordinary acne). Synthetic photograph generation allows clinical findings to be varied while the person’s features and photographic setting stay consistent, reducing confounding. In an investigation of three frontier models, we found that models hallucinated clinically significant photograph findings in 45-57.5% of mismatched pairs, up from 25-37.5% for the same photographs shown alone. When faced with mismatched pairs, models tended to ignore mismatches and instead hallucinated findings that better matched the radiograph (e.g. misreading acne as crusted skin lesions to justify diagnosing chickenpox). Trick-or-Treat both demonstrates concerning error pattern in frontier models’ medical reasoning and provides a proof of concept for the use of synthetic images in adversarial multimodal evaluation.
[CV-28] Interactive World Simulation with Cosmos for Physical AI
链接: https://arxiv.org/abs/2611.06627
作者: NVIDIA:Anis Ahmad,Jon Allen,Malik Aqeel Anwar,Jesse Archer,Maciej Bala,Piotr Barejko,Aarti Basant,Dan Blick,Taras Bolzhelarskyi,Henry Che,Jin Chen,Yukang Chen,Tae Eun Choe,Yuzhu Dong,Nat Duca,Liang Feng,Francesco Ferroni,Xiao Fu,Mohit Gambhir,Vivek Goel,Rony Goldenthal,Aryaman Gupta,Ali Hassani,Nathan Hayes-Roth,Yong He,Michael Honke,Shengyu Huang,Wei Huang,DeLesley Hutchins,Suneel Indupuru,Pooya Jannaty,Jingyi Jin,Janmesh Kalra,Ashna Khetan,Xin Kong,Aleksandr Korovko,Hui Kuang,Zhengfei Kuang,George Kurian,Jean-Francois Lafleche,Jayjun Lee,Justin Leung,William Lew,Xiang Li,Zhaoshuo Li,Chen-Hsuan Lin,Ming-Yu Liu,Xiangyu Lu,Yuhang Lu,Martin Ding Ma,Qianli Ma,Aditya Mahajan,Miguel Martin,Janarbek Matai,Jonathan McCaffrey,Qing Miao,Ashkan Mirzaei,Mustafa Munir,Sangmin Oh,David Pankratz,Trung Pham,Marilyn Reeb,Gavriel State,Jiaxiang Tang,Shitao Tang,Yue Tang,Tolou Tavakkoli,Alexander Temirov,Hweimi Tsou,Sergei Vasilev,Thomas Volk,Andrew Z. Wang,Hao Wang,Shijie Wang,Zhonghao Wang,David Wehr,Shumian Xin,Mengyao Xu,Yao Xu,Shuai Yang,Sophia Zalewski,Pengcuo Zeren,Cindy Zha,Jing Zhang,Yuliya Zhautouskaya,Kaiwen Zheng,Andrew Zhu,Rick Zhuang
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注: Project website: this https URL
Abstract:We present a shared framework that adapts Cosmos world foundation models for real-time interactive world simulation in Physical AI. By streaming visual observations in response to actions and scene controls, the framework enables users and embodied agents to interact with evolving virtual environments. It supports both predicting the visual consequences of actions and generating observations from physical state maintained by an external simulator. We demonstrate this approach across camera navigation and robot manipulation, controllable video transfer, and multi-view autonomous vehicle (AV) simulation. These examples provide a shared foundation for developing interactive virtual environments for Physical AI training and evaluation. Our code and model checkpoints are available at this https URL and this https URL. The project website is available at this https URL.
[CV-29] Learning to Track Cells from Sparse Annotations in 3D Time-Lapse Microscopy
链接: https://arxiv.org/abs/2611.06624
作者: Yusuke Uchida
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Reconstructing cell lineages from 3D time-lapse microscopy requires distinguishing cell motion from division and inferring temporally consistent connections. When only a subset of cells is annotated, however, training procedures that treat unannotated regions as background or incorrect correspondences can be problematic. We propose a cell tracking pipeline that integrates detection from sparse annotations, calibration of candidate connections using image features, and joint selection of lineage graph events. For detection, we combine the outputs of 2.5D and 3D models trained to suppress erroneous background supervision at unannotated cells. We then calibrate ordinary link and division probabilities using motion features from dense flow, correspondence features from a learned cell-matching model, and features from models that identify cells immediately before division, dividing parents, and daughters immediately after division. Ordinary links, divisions represented as daughter pairs, and node inclusion are selected jointly under a surrogate objective derived from the evaluation metric, followed by gap filling and temporal coordinate refinement. The pipeline achieves an overall score of 0.9762 in 5-fold cross-validation on 199 videos from 2 zebrafish embryos. With candidates and calibrated values held fixed, joint selection outperforms a sequential method that allows replacement of existing divisions and reoptimization of ordinary links by 0.0105. The improvement arises primarily from division recovery.
[CV-30] SpikePose: Lightweight Video Human Pose Estimation ICASSP2027
链接: https://arxiv.org/abs/2611.06594
作者: Yunhao Huang,Kaushik Bhargav Sivangi,Tianyi Yu,Fani Deligianni
类目: Computer Vision and Pattern Recognition (cs.CV); Neural and Evolutionary Computing (cs.NE)
备注: 5 pages, submitted to ICASSP 2027
Abstract:Efficient video human pose estimation must balance localisation accuracy, motion quality, and energy cost. We present SpikePose, a lightweight video pose estimator that pairs a frame-reset, single-step spiking encoder with motion-aligned heatmap memory (MAM). The encoder supplies per-frame spatial evidence, while MAM estimates joint displacement, aligns historical heatmaps, and adaptively fuses them with current predictions. By confining temporal reasoning to a compact, lightweight module outside the spiking backbone, SpikePose helps bridge the gap between spiking and classical Artificial Neural Networks (ANN) on large-scale, non-event-driven video tasks. We evaluate spatial accuracy, motion errors, and estimated arithmetic energy on MPII and a fixed NTU60 Cross-Subject subset. Against the architecture-matched ANN with MAM, SpikePose cuts estimated arithmetic energy by 91.4% at a cost of just 1.12 points in head-bone-normalised PCK at 0.5 threshold (PCK_HB(0.5)). Within SpikePose, adding MAM reduces velocity and acceleration errors by 28.13% and 31.46% while slightly improving localisation, at a cost of only 0.003M parameters and 0.0023 analytical GFLOPs per frame. Code will be released on GitHub upon acceptance.
[CV-31] BudgetSplat: Budget-Controllable Generalizable Gaussian Splatting
链接: https://arxiv.org/abs/2611.06593
作者: Yuxi Hu,Jun Zhang,Wei Mao,Yuchen Rao,Friedrich Fraundorfer
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Feed-forward 3D Gaussian Splatting enables efficient generalizable novel view synthesis, but existing pixel-aligned methods often generate highly redundant Gaussian primitives, causing the final representation size and rendering cost to grow rapidly with image resolution and the number of input views. We propose BudgetSplat, a budget-controllable feed-forward framework that adapts the final Gaussian representation to a user-specified preservation ratio. Rather than treating compaction as simple importance ranking, we formulate it as a joint problem of importance estimation, spatial coverage, and cross-view redundancy. A single trained model supports varying numbers of context views and multiple test-time budgets without per-budget retraining or per-scene compression. To realize this formulation, Intra-Cross Mask Supervision provides structure-, coverage-, and redundancy-aware guidance for learning Gaussian importance, while budget-conditioned feature modulation adapts Gaussian opacity and attributes to the requested preservation ratio. Coverage-Aware Budget Selection further combines patch-wise pre-selection, view-wise opacity ranking, and 3D redundancy suppression to construct compact yet spatially distributed Gaussian representations. Experiments on indoor reconstruction and cross-dataset generalization demonstrate consistent quality–size trade-offs across different input view counts and preservation ratios, substantially reducing final representation storage and rendering cost while preserving rendering quality.
[CV-32] LungMend: A Healthy-Only Latent Diffusion Shape Prior for Lung Mask Correction in Severe Tuberculosis CT
链接: https://arxiv.org/abs/2611.06556
作者: Rezkellah Noureddine Khiat,Pierre-Yves Brillet,Catalin Fetita
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 5 pages (4 + 1 references), 2 figures, 2 tables. Submitted to IEEE ISBI 2027. Code: this https URL
Abstract:Lung segmentation models such as R231 are widely used, but they fail on severe tuberculosis (TB) CT: cavities and consolidations look like non-lung tissue and are left out of the mask. We present LungMend, a post-processing model that corrects such masks using only healthy anatomy. A latent diffusion model, trained on healthy lungs with synthetically corrupted masks, learns to recover the complete lung from the faulty mask and the bone anatomy, without any pathological data or CT intensities. LungMend recovers cavities and consolidations that R231 misses and improves accuracy on expert-annotated slices. Code: this https URL
[CV-33] MKD2World: A Typed Spatial MarkDown World Model For Web Agents
链接: https://arxiv.org/abs/2611.06544
作者: Zhihao Nan,Sifan Zhou,Yiming Cheng,Jinkai Zheng,Zheng Zhu,Si Shi,Ming Li,Fei Yu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 40 pages
Abstract:Graphical user interface (GUI) agents can improve decision-making with world-model foresight, provided that predicted states accurately reflect post-action interfaces. Existing state representations fail to jointly capture interface layout, UI semantics, and explicit locations. We introduce Typed Spatial Markdown (TSM), a Markdown-style textual state representation for desktop web interfaces. TSM serializes block-level and atomic-level UI elements according to their spatial relationships, jointly encoding their types, visual semantics, explicit locations, optional control states, and hierarchical membership. To train a screenshot-to-TSM renderer, we construct AtomBlock-WebUI, a dataset with precise annotations at both granularities. By injecting element-type labels into HTML and directly extracting bounding boxes after rendering, we avoid annotation errors caused by post-hoc inference from webpage structures. Building on TSM, we propose MKD2World, a desktop web world model trained in two stages for current-state description and action-conditioned next-state prediction. During inference, the web agent queries MKD2World as needed and performs a ``digest’’ step, in which it interprets each predicted state and extracts task-relevant evidence for selecting the best action. Experiments show that MKD2World-9B achieves an EleF1 score of 26.50 on the out-of-domain test set, outperforming GPT-5.6-Sol(14.98) by 11.52 points. Across Online-Mind2Web tasks, MKD2World-9B improves the success rate of the Qwen3.5-35B web agent from 22.94% to 37.61%, achieving competitive performance among leading open-source GUI world models. Code is available at this https URL.
[CV-34] Spatially and Compositionally Controlled Multimodal Glioblastoma MRI Synthesis with Latent Diffusion
链接: https://arxiv.org/abs/2611.06540
作者: Steven Glass,Max Van Puyvelde,Marija Pizurica,Mia Meyer,Olivier Gevaert,Willie Brink
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Glioblastoma exhibits substantial heterogeneity in tumour morphology and composition across MRI modalities, often making available clinical datasets highly unbalanced in terms of patient pathologies. We propose a two-stage latent diffusion model (LDM) framework for the generation of explicitly controllable 2D glioblastoma MRI scans. First, a segmentation LDM generates tumour maps conditioned on location, spatial extent, axial position, and subregion composition. Second, a modality-conditioned LDM with ControlNet generates T1, T1Gd, T2, and FLAIR images adhering to the requested segmentation. On a held-out test set, generated segmentations achieved a mean centroid error of 0.842 pixels and a subregion proportion error of 0.795 percentage points. ControlNet improved Fréchet inception distance (FID) across all four modalities, reducing pooled FID from 42.59 to 36.12. An expert reader study yielded mean quality ratings of 6.60 for real images, 6.83 for ControlNet-generated images, and 6.77 for fully synthetic images.
[CV-35] EmoReward: Self-Evolving Reward Calibration for Emotional Video Generation RL
链接: https://arxiv.org/abs/2611.06518
作者: Mingzhe Zheng,Yue Wu,Jing Tan,Chang Zou,Dengyang Jiang,Zeyue Tian,Xuanhua He,Ryan Xu,Yunyang Ge,Sihui Ji,Yunlong Lin,Kaihang Pan,Hongru Yan,Sixiang Chen,Dianyi Wang,Julian Zhang,Weijie Kong,Changlin Li,Lanqing Guo,Zhao Zhong,Qifeng Chen,Harry Yang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: 30 pages, 15 figures
Abstract:Group Relative Policy Optimization (GRPO) has advanced generative model post-training in verifiable domains such as code and mathematics, but evaluating video quality, especially emotional expression, often depends on subjective aesthetics and has no uniquely correct target. Existing approaches typically rely on fixed rewards for visual quality, motion quality, and text alignment that poorly capture emotional intent and can invite reward hacking, while RLHF is difficult to scale and subject to individual aesthetic differences that complicate the calibration of facial, body motion, and chromatic cues across narrative contexts. We introduce EmoReward, a new self-evolving training mechanism that distills evidence reliability from reward traces collected during an actual training run to adaptively guide multi-reward weight allocation in subsequent training by determining which emotional evidence to trust and how much to weight it. EmoReward first plans specific grounding cues from the emotional prompt to guide evaluation, then uses heterogeneous signal primitives to score the generated video during the act stage. Reflective Reward Gating calibrates their contributions through instance-level gates for phase, face confidence, conflict, and subtlety, then renormalizes weights over the active set to produce the sample’s fused reward. RewardMemory then estimates each primitive’s reliability from rollout scores and outcomes within each emotional regime in the finished training run, then fixes bounded multipliers that calibrate fusion weights in the next training run without updating reward models. Experiments on HunyuanVideo 1.5 show improvements across multiple RL algorithms and compare EmoReward with other agentic reward methods under SAGE-GRPO. Ablations and transfer experiments on SD 3.5 and FLUX further support adaptive reward calibration for visual generation.
[CV-36] A 3D foundation model unifying CT interpretation and segmentation
链接: https://arxiv.org/abs/2611.06471
作者: Yulu Xie,Zilong Li,Shen Xu,Yifan Bai,Liuhong Shi,Jiao Chen,Xianglei Kong,Wenchao Zhu,Na Shen,Xianghua Ye,Yang Yan,Jiaju Zheng,Yunsong Peng,Yanzhou Su,Chaohui Yu,Hao Luo,Fan Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 114 pages, including Supplementary Information; 7 main figures and 2 Extended Data figures
Abstract:Radiologists interpret computed tomography (CT) scans by jointly determining what a finding represents and where it lies in three-dimensional (3D) anatomy. However, CT models have treated CT interpretation and 3D segmentation as separate tasks, leaving textual findings spatially unverifiable and overlooking potential synergy between clinical understanding and spatial perception. Here we present HoloCT, a generalist 3D foundation model unifying CT interpretation and voxel-level segmentation by jointly modelling whole-volume context and fine-grained anatomical detail, providing 3D masks as evidence. We curated HoloDATA, comprising 195,921 CT scans from 73 public datasets and six clinical cohorts, with 6.2 million instruction dialogues and 2.8 million CT-mask pairs. Across 346 tasks spanning segmentation, lesion screening, visual question answering, and report generation on 61 internal and 12 external datasets, HoloCT outperformed specialist and foundation models. In multicentre validation across 11,455 clinical CT scans from five hospitals, HoloCT outperformed competing models in lesion segmentation and report generation by 18.2 and 15.3 points. It also achieved 85.1% mean accuracy in screening thyroid tumours and colon lesions, exceeding all four participating radiologists in a retrospective reader study. Further analyses revealed bidirectional semantic-spatial synergy: language-guided segmentation delineated unseen anatomical structures and lesions as well as diffuse free-text findings, outperforming the leading approach by 22.1 points, whereas prompting the model to generate corresponding 3D masks alongside its textual outputs significantly improved interpretation performance. HoloCT shows that clinical understanding and spatial perception can be learned together within a single foundation model and offers a path towards more accurate and reliable CT analysis.
[CV-37] CoDecide: Content-Adaptive Video Decoding for High-Throughput Video Analytics WACV2027
链接: https://arxiv.org/abs/2611.06455
作者: Shristy Shah,Thomas B. Moeslund,Kamal Nasrollahi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at WACV 2027. 12 pages including supplementary material
Abstract:Analytics on video archives are bottlenecked by decoding. In a dense pipeline that runs a detector on every frame, decoding is a small fraction of the runtime, but once the detector is restricted to selected frames, decoding takes over as the main cost. We introduce CoDecide, a retrospective analytics system that only decodes the frames where new content appears, using signals from the compressed bitstream to find them. Motion tells us something is moving, but not whether it is new to the scene or has been there the whole time. To make that distinction, we supervise a lightweight compressed-domain student on a label-free novelty signal from a frozen DINOv2 teacher, which fires on semantically new regions rather than on any motion. Consequently, the student model generalizes to new camera deployments without requiring per-camera retraining. At inference, it decides which frames to decode and which regions to reconstruct. Detections from the decoded frames are propagated to the skipped frames using motion vectors. We evaluate CoDecide with a leave-one-dataset-out protocol on VIRAT, MEVA, and ShanghaiTech. Against a dense per-frame detector, CoDecide recovers more moving-object detections than prior compressed-domain selectors at an equivalent decoding budget, outperforming even a competitor trained directly on the target camera. CoDecide runs 15 to 24x faster than a pipeline that decodes and detects every frame. Code is available at this https URL.
[CV-38] Beyond Infraction Scores: Failure-Boundary Profiles for Closed-Loop Evaluation of Driving and Robot Navigation Agents
链接: https://arxiv.org/abs/2611.06453
作者: Milan Bulat,Gino Ardelean,Ellington Kirby
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV); Software Engineering (cs.SE)
备注: 16 pages, 12 figures, 3 tables. Code: this https URL . Data: this https URL
Abstract:Closed-loop driving benchmarks in CARLA rank agents by a driving score that multiplies route completion by an infraction penalty. This score does not distinguish an agent that stops in front of an obstacle from one that pushes past it. In a corridor closed by two parked cars, for example, one agent that drives through with contact averages 52.2 on the blocked widths, while the agents that stop average between 37.8 and 40.6. We propose failure-boundary profiles, an evaluation in which each scenario family varies one measured factor, such as corridor width or the time gap in oncoming traffic, and each variant is certified from measurement as one where the agent should proceed or one where it should stop. Along this axis we estimate two curves per agent, commitment (the probability of attempting the passage) and competence (the probability of passing without contact), and locate where each crosses one half, with intervals that account for the simulator’s run-to-run variance. The certified labels also give a summary metric, decision accuracy, which credits clean passes on passable variants and stops on blocked ones. We evaluate seven driving agents on seven CARLA families. The rule-based expert reaches a decision accuracy of 0.74 and TFv6, the highest-scoring learned agent, 0.60. Agents with similar driving scores fail in different ways, and agents that reliably stop at a blocked corridor still rarely wait for a gap in oncoming traffic. We also apply the method, unchanged, to a warehouse robot in Isaac Sim. The navigation policy we test stops at blocked gaps in only one of three families, and two fine-tuned versions of it pass more gaps cleanly but no longer stop at all.
[CV-39] Within-Family EfficientNet Scaling under Fixed-Resolution Transfer Learning for Spectral-Domain Optical Coherence Tomography Retinal Disease Classification
链接: https://arxiv.org/abs/2611.06452
作者: Shkelqim Sherifi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 46 pages, 20 figures, 18 tables
Abstract:Background/Objectives: CNNs perform well in retinal disease classification from spectral-domain optical coherence tomography (SD-OCT), but whether EfficientNet’s natural-image scaling hierarchy transfers to medical-image transfer learning at fixed resolution remains unclear. This study evaluates EfficientNet-B0-B7 for four-class SD-OCT classification and accuracy-efficiency trade-offs. Methods: Eight ImageNet-pretrained variants (4.05-64.10M instantiated backbone parameters) were evaluated at 224x224 on the Kermany dataset. Grayscale images were replicated to RGB and processed using model-embedded preprocessing and augmentation. Training comprised frozen-backbone head training followed by partial fine-tuning. Metrics covered classification, discrimination, efficiency, and visual analyses. Results: Test accuracy varied non-monotonically from 95.93% (B0) to 96.95% (B1); B7 achieved 96.73%. B1 had the highest accuracy and Drusen F1 (0.8780), while B7 had Drusen F1 of 0.8720 and the highest Drusen recall (0.9177). Drusen had the lowest F1 across all variants and contributed 82.2% of the summed squared deviation in class-specific F1; class ranking remained Normal CNV DME Drusen. The OCT accuracy ranking differed from the reported ImageNet hierarchy, where B7 ranked first. The accuracy-to-parameter ratio fell from 19.82 to 1.48 percentage points per million total parameters. The final train-test gap had a weak, non-significant association with test accuracy (r = 0.26, p = 0.53). Conclusions: Larger EfficientNet variants did not consistently improve SD-OCT classification. B1 achieved the highest accuracy and Drusen F1, whereas B7 showed a secondary accuracy recovery at greater model size and CPU latency, indicating task and protocol dependent scaling effects. As cross-partition duplicates have not been independently audited, results are conditional on the supplied split.
[CV-40] DreamRealm: Exploring and Materializing Hallucination Worlds for Trustworthy Vision-Language Models
链接: https://arxiv.org/abs/2611.06433
作者: Benlei Cui,Chunyang Chai,Yiliang Zhang,Xiaoqian Xia,Chengwen Yao,Song Li,Longtao Huang,Yuwen Zhai,Haiwen Hong
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 28 pages, including supplementary material. Code: this https URL . Model: this https URL
Abstract:Multimodal large language models (MLLMs) remain prone to visual hallucinations when correct responses depend on subtle visual details. One line of post-training methods constructs visual contrasts through image retrieval, generation, or editing to improve fine-grained perception. However, these contrasts are not built around the model’s own hallucinations, making it difficult to ensure that supervision targets its actual perceptual confusions. Another line suppresses hallucinated outputs through response correction or preference learning under fixed image conditions, but leaves the design of supervision signals on the image-input side underexplored. Inspired by how humans correct misconceptions, we introduce DreamRealm, a post-training framework that explores diverse latent hallucinated responses, minimally edits original images to make them valid, and trains the model to assign each response to its supporting image. Our data construction pipeline, RealmForgePipe, mines observed and latent hallucinations from a target model and materializes them through image editing. Using hallucinations mined from Qwen3.5-9B, we construct the Realized Hallucination Dataset (RealmSet). Our training objective, RealmMatch, constructs a likelihood score matrix across image–response combinations and combines supervised learning with bidirectional contrastive losses to reinforce correct matches and suppress mismatches. Compared with the original Qwen3.5-9B model, DreamRealm improves performance across six visual perception and hallucination benchmarks, including gains of 7.26 percentage points in HallusionBench accuracy and 11.34 percentage points in MMVP pair accuracy. Together with ablation studies on data construction and training methods, these results demonstrate that hallucination assignment is a promising paradigm for improving visual perception and mitigating hallucinations.
[CV-41] Unaligned Image Guided Denoising via Cross-modal Conditional Flow Matching NEURIPS2026
链接: https://arxiv.org/abs/2611.06422
作者: Runmin Zhang,Linshan Li,Si-Yuan Cao,Zhu Yu,Lingyu Zhu,Huaqi Zhang,Hui-Liang Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted by NeurIPS 2026
Abstract:This work presents UGD-FM, a novel unaligned image guided denoising framework based on cross-modal conditional flow matching. Unlike previous approaches that assume spatially aligned inputs or only handle small-baseline rectified pairs, we address a more challenging setting where severe target noise and large-range two-dimensional cross-modal misalignment coexist. Specifically, we formulate guided denoising as a progressive flow matching process, allowing image restoration and aligned guidance aggregation to mutually reinforce each other. To support efficient few-step inference, we introduce a stage-focused and inference-consistent training strategy that focuses on task-critical denoising stages, and mitigates the training-inference mismatch. We further propose adaptive sparse guidance propagation for efficient and reliable large-range matching, regularized by perceptual and geometric auxiliary losses. Experiments demonstrate that UGD-FM achieves state-of-the-art denoising performance with only one or two inference steps, and is compatible with modern single-image restoration backbones.
[CV-42] Entropogram-guided Visual Realignment Decoding: Capturing the Rhythm of Multimodal Reasoning
链接: https://arxiv.org/abs/2611.06415
作者: Jiaqi Deng,Zhe Fu,Yakun Chen,Dichang Zhang,Zonghan Wu,Xiaoshui Huang,Huan Huo,Guandong Xu
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: 22 pages, 9 figures
Abstract:Multimodal large language models (MLLMs) have demonstrated remarkable progress on cross-modal reasoning tasks, yet they continue to suffer from hallucinations in long-chain reasoning processes. In this work, we revisit hallucination in MLLMs from the perspective of reasoning generation dynamics. We argue that entropy is not merely a static measure of uncertainty but a structured signal that reflects how modality information is integrated throughout decoding. Through systematic analysis, we identify two characteristic entropy patterns associated with hallucination behaviors: (1) a generally high-confidence regime corresponds to premature convergence to inappropriate multimodal interactions; and (2) local abrupt entropy transitions are associated with reasoning instability and hallucination onset. Building on these observations, we propose an entropy-aware self-introspective framework for MLLM reasoning. Our method continuously tracks entropy dynamics during decoding and performs targeted visual re-alignment when suspicious hallucination states are detected, encouraging the model to re-examine salient visual evidence at critical reasoning stages. Extensive experiments across six multimodal reasoning benchmarks spanning nine distinct evaluation settings demonstrate that our approach delivers robust improvements in reasoning accuracy and hallucination robustness, while providing new insights into the generation dynamics underlying MLLM failures. The code is available at: this https URL.
[CV-43] A Bag of Tricks for Defending Deep Learning-based Face Recognition Pipelines against Backdoor Attacks
链接: https://arxiv.org/abs/2611.06412
作者: Quentin Le Roux,Marti Jimenez,Yannick Teglia,Philippe Loubet Moundi
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
备注:
Abstract:The protection of mature, multi-stage Face Recognition Systems is crucial to system providers. This technical report targets the threat model most relevant to industry practitioners: although model training is performed in-house, training data is sourced from possibly-compromised third parties. Given this setting, we present three practical defense “tricks” against data poisoning-based backdoor attacks. The first operates at training-time, detecting poisoned identities during training or prior to a pipeline’s deployment. The last two operate at inference-time, filtering suspicious inputs before an answer can be sent back to a potential attacker. Each method is presented with their motivation, algorithms and computational impact, and evaluated on realistic face recognition workloads. Our goal is to provide actionable hardening techniques to system providers.
[CV-44] UniLoop: Unified Loop Closure Detection from Retrieval to Geometric Verification and Relative Pose Estimation
链接: https://arxiv.org/abs/2611.06378
作者: Jingwen Yu,Pengyu Wang,Jin Wu,Weinan Chen,Ping Tan,Hong Zhang
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Loop closure detection conventionally cascades separate models for global retrieval, geometric verification, and relative pose estimation, leading to redundant computation and disconnected visual representations. We propose UniLoop, a unified framework that reuses the hierarchical representations of the pretrained geometric foundation model throughout this pipeline. A shared image encoder produces features for compact global descriptors; the same features are passed to the cross-view decoder for pairwise verification; and the original geometric predictions provide relative camera constraints for pose-graph optimization. Experiments on place recognition, loop verification, and relative pose estimation show that UniLoop improves loop-closure robustness and accuracy while reducing redundant backbone computation relative to modular pipelines. UniLoop achieves the best Recall@10 on MSLS (96.6) and Pitts30k (99.3) among the compared methods, while its verification branch reaches 95.63 AP and 96.29 ROC-AUC, surpassing the existing geometric verification baselines.
[CV-45] NanoKV: High-Ratio KV Cache Compression for Autoregressive Video Generation
链接: https://arxiv.org/abs/2611.06377
作者: Jiacheng Lu,Haoyi Zhu,Sipei Yi,Enze Xie,Cheng Zhuo,Yu Li
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 25 pages, 13 figures. Project page: this https URL . Code: this https URL
Abstract:Autoregressive video models rely on historical key-value (KV) caches to preserve scene consistency over extended rollouts, making KV storage and repeated attention over history a growing memory and computational bottleneck. Aggressive compression can degrade generation quality, while reconstructing full-dimensional KV leaves attention computation largely unchanged. We introduce NanoKV, a frozen-backbone framework for generation-aware high-ratio KV compression. NanoKV exploits redundancy across tokens and attention heads by storing shared content in cluster centers and encoding joint head-channel residuals into low-bit coefficients. Crucially, NanoKV uses different fitting objectives for Keys and Values according to their distinct roles in attention: Key decoders are optimized to preserve downstream block outputs, whereas Value encoders and decoders are fitted for faithful reconstruction through the deployed quantization path. Building on the same compressed representation, the optional NanoKV-Fast variant reuses the stored coefficients for mixed-dimensional attention, reducing computation without modifying the video backbone. Across three autoregressive video backbones, NanoKV consistently improves the quality-storage trade-off. On HY-WorldPlay, NanoKV achieves 26.09\times effective KV compression, nearly twice that of KVTC, while improving PSNR by 1.0 dB. NanoKV-Fast retains 20.53\times effective compression and achieves an approximately 1.15\times end-to-end speedup over KVTC.
[CV-46] FlashRelight: Portrait Video Relighting with Dynamic Lighting NEURIPS2026
链接: https://arxiv.org/abs/2611.06355
作者: Jiahui Sheng,Le Li,Limin Lin,Jiahao Li,Wansen Feng,Hong Gu,Shuhan Chen,Xiaorun Li,Rui Hu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026. 23 pages, including appendix
Abstract:Portrait video relighting aims to modify the illumination of a portrait video while preserving photorealism and temporal stability. The task becomes particularly challenging under dynamic lighting, where illumination direction, intensity, and color vary over time. To tackle this problem, we introduce FlashRelight, a novel two-stage framework for dynamic portrait video relighting. FlashRelight first suppresses the source illumination through delighting, and then synthesizes the target dynamic lighting on an illumination-neutral portrait video. FlashRelight supports three complementary control modalities: envmap sequences for explicit frame-level lighting control, dynamic lighting text for semantic illumination editing, and lighting reference videos for visual lighting transfer. Training follows a low-cost hybrid strategy that combines static one-light-at-a-time (OLAT) data with generated dynamic-lighting portrait videos, covering both controlled illumination variation and realistic portrait motion. Extensive experiments show that FlashRelight achieves temporally coherent and photorealistic dynamic relighting while enabling flexible cross-video lighting transfer. Code is available at this https URL
[CV-47] raffic-VL: Reinforced Vision-Language Models for Generalizable End-to-End Traffic Signal Control
链接: https://arxiv.org/abs/2611.06353
作者: Xingchen Zou,Maonan Wang,Kemou Jiang,Jie Li,Yiqi Chen,Xi Chen,Hai Yang,Yuxuan Liang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Existing traffic signal control methods typically rely on structured traffic states produced by dedicated perception pipelines. Emerging visual approaches often use idealized simulator images and convert them into intermediate representations before control, limiting their ability to directly exploit roadside visual observations. To bridge both the visual-domain and perception-to-decision gaps, we introduce SynTra-V and Traffic-VL. SynTra-V converts physics-based traffic simulations into photorealistic multi-view surveillance imagery across diverse road topologies and sensing conditions, and automatic semantic auditing checks that control-critical geometry, vehicle movements, and lane semantics match the simulator state. Traffic-VL is an eight-billion-parameter vision-language model that directly maps synchronized multi-view images and natural-language phase descriptions to signal decisions without intermediate structured states. Building on Traffic-R1’s two-stage reinforced reasoning framework, Traffic-VL adapts its vision encoder via offline reinforcement learning on SynTra-V and freezes it during online visual stepwise trajectory policy optimization in closed-loop simulation. Experiments show that Traffic-VL matches structured-state Traffic-R1 while outperforming image-based controllers and general-purpose vision-language models, including Gemini 3 Pro. Evaluations on unseen topologies, degraded sensing conditions, real surveillance footage, and a three-week deployment across ten intersections further demonstrate robust visual generalization and effective constrained online control.
[CV-48] Prism: Bidirectional Self-Teaching for Unified Multimodal Models
链接: https://arxiv.org/abs/2611.06343
作者: Peng Zhang,Qian Chen,Xianyin Zhang,Yimin Tian,Ruite Fei,Dongyi Lv,Lifan Guo,Chi Zhang,Shuai Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Unified Multimodal Models (UMMs) aim to support visual understanding and generation within a single model. Existing post-training methods either use understanding as a fixed teacher to improve generation, or rely on ground-truth trajectories to supervise both understanding and generation. Prism instead derives bidirectional supervision from self-generated interleaved rollouts, constructing training targets according to the reliability of their text and image chains. We propose Prism, a bidirectional self-teaching post-training method for UMMs. Like a prism decomposing mixed light, Prism decouples a coupled interleaved trajectory into a text chain and an image chain, assesses the reliability of each chain, and constructs a corresponding training target from their reliability states. Experiments on both SenseNova-U1 and BAGEL show that Prism improves both understanding and generation. On the generation side, GenEval improves from 85.8 to 89.2 (+3.4), and DPG improves from 86.1 to 87.3 (+1.2) on SenseNova-U1. On the understanding side, the total MME score improves from 2282.6 to 2311.6, and MMVP improves from 58.0 to 62.0. Consistent improvements are observed on BAGEL. These results highlight that Prism enables coordinated improvements in generation and understanding by constructing reliability-guided self-training targets from self-generated trajectories, without relying on an external verifier or externally annotated corrected trajectories during target construction.
[CV-49] Beyond Scene Shortcuts: Action-Centric Video-Text Alignment for Fine-Grained Action Understanding
链接: https://arxiv.org/abs/2611.06302
作者: Zhiliang Zhu,Jiazheng Huang,Dawei Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, including appendices
Abstract:Fine-grained action understanding requires vision-language models to ground what an agent does, how it acts, and why it acts in spatio-temporal evidence. However, richer supervision can backfire: fine-tuned models may generate detailed action descriptions grounded in static scenes and salient objects rather than motion, a failure we call Environment-Grounded Action Hallucination (EGAH). To mitigate EGAH, we propose Staged Attention Rectification and Reinforcement (STARR), an action-centric video-text alignment framework that breaks scene-action shortcuts through Action-Environment Dissociation and reinforces action-focused grounding with an Attention Consistency Reward within Group Relative Policy Optimization. We further construct FineActionRR, combining temporally localized captions with hierarchical and temporal-causal reasoning data. Built on Qwen2.5-VL-7B, FineActionVLM-R1 improves average MotionBench accuracy by 3.01 percentage points over its base model and achieves 91.50% on HAU-MCQ-Bench. Qualitative comparisons further indicate more faithful fine-grained action descriptions with fewer environment-induced hallucinations.
[CV-50] MonoLaDN: Language-Guided DeNoising for Monocular 3D Object Detection ACCV2026
链接: https://arxiv.org/abs/2611.06282
作者: Kiet Dang Vu,Cao Le Minh Khoa,Nguyen Viet Tri Pham,An Nguyen Gia,Tam Le Thuc Ho,Duc Dung Nguyen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at ACCV 2026
Abstract:Monocular 3D object detection offers a low-cost route to 3D perception, but it remains fundamentally ill-posed, since depth must be inferred from a single view. Recent transformer-based detectors learn geometry purely from image features and 3D box annotations, lacking the high-level semantic priors that vision-language models (VLMs) acquire from large-scale pretraining. VLMs offer such priors, but deploying one directly as a detector is impractical due to its heavy inference cost. We propose MonoLaDN, a language-guided denoising framework that transfers the geometric reasoning of a VLM into a monocular 3D detector while keeping the deployed model purely vision-based. During training, a frozen VLM produces structured reasoning about each object’s occlusion, depth, dimension, and angle from the image and its ground-truth boxes. We encode this reasoning into queries, perturb them with noise, and train the detector to reconstruct the clean reasoning. Reconstructing the corrupted reasoning forces the detector’s depth-guided features to internalize the VLM’s geometric priors, so this knowledge remains available even after the reasoning path is discarded at inference, adding no latency or parameters to the deployed model. Experiments on KITTI 3D and nuScenes show that MonoLaDN consistently outperforms state-of-the-art detectors and generalizes well to new domains.
[CV-51] Quo Vadis Driving Scene Reconstruction? Occlusion-Aware Gaussian Densification ACCV2026
链接: https://arxiv.org/abs/2611.06273
作者: Dung Do Tien,Kien Nguyen Do Trung,Kiet Dang Vu,Trung Thai Tran,Duc Dung Nguyen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at the Workshop on Computer Vision for Developing Countries (CV4DC), ACCV 2026 (Proceedings Track)
Abstract:While recent feed-forward 3D Gaussian Splatting methods enable efficient driving-scene reconstruction from surround-view images, their reliance on a single observation fundamentally hinders novel-view synthesis. Regions occluded in the input can become visible from novel viewpoints, leaving the reconstructed Gaussians under-constrained and incomplete. Moving beyond image-space inpainting for rendered views, we introduce OccSplat, a feed-forward framework whose core contribution is Occlusion-Aware Gaussian Densification, explicitly targeting missing scene geometry rather than missing pixels. By predicting new Gaussian primitives for unsupported regions, OccSplat completes the scene directly in the 3D representation. Specifically, OccSplat first identifies occluded regions that lack Gaussian support through Gaussian Occlusion Mapping. Gaussian Occlusion Completion then generates new primitives for these regions. The predicted primitives are integrated into the original representation to improve novel-view synthesis while retaining feed-forward inference without per-scene optimization. Experiments on the nuScenes benchmark demonstrate that OccSplat improves on the compared methods and yields particularly large gains in occluded regions, leading to more accurate and visually coherent driving-scene reconstruction.
[CV-52] NextGAN: Towards Simple and Scalable GAN Training for Image and Video Generation
链接: https://arxiv.org/abs/2611.06270
作者: Xinhao Lou,Dongye Xie,Peng Sun,Chenyi Zhuang,Kaicheng Yu,Tao Lin
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Adversarial training enables one- and few-step image and video generation, but conventional adversarial losses supervise critic values without directly specifying the gradients used to update the generator. We present NEXTGAN, which trains a scalar critic by regressing its input gradient onto displacements between independently paired reference and generated features at interpolated locations in frozen representation spaces. For a fixed generator, this regression is equivalent to an adversarial objective with a gradient-energy penalty, while the generator retains the standard potential-based GAN update. The same objective adapts pre-trained image and video generators with up to 14B parameters, without auxiliary losses, online teacher scores, or trajectories. On ImageNet 256 \times 256 , NEXTGAN achieves one-step FIDs of 2.87 with pixel-space JiT-B and 1.54 with latent-space DiT-XL/1; its adversarial form further reaches 1.36. For text-to-video generation, a single trained Wan2.1-1.3B model supports one-, two-, and four-step sampling, achieving the highest VBench total scores among the compared 1.3B methods at one and two steps. At two steps, NEXTGAN scores 83.13, exceeding rCM and AnyFlow by 1.25 and 0.47 points, respectively. Code is available at this https URL.
[CV-53] A Systematic Benchmark for Evaluating Multimodal Personalized Reasoning of Vision-Language Models WACV2027
链接: https://arxiv.org/abs/2611.06259
作者: Dohyun Kim,Junyong Lhim,Kwanyong Park
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV 2027
Abstract:People often make decisions by holistically reasoning over heterogeneous forms of personal information. In this paper, we study the relatively underexplored capability of multimodal personalized reasoning in multimodal large language models (MLLMs). To this end, we introduce \textttMPRBench, a novel and comprehensive benchmark specifically designed to evaluate personalized reasoning across a wide range of tasks and real-world challenges. \textttMPRBench consists of 12 sub-tasks and additionally supports interpretable analysis of representative error cases. Through extensive experiments, we find that current personalized MLLMs still struggle with personalized reasoning due to several key challenges, including inaccurate recognition of personalized concepts and sensitivity to irrelevant personal information. We hope that MPRBench will stimulate further research on personalized reasoning in MLLMs. Code and data are available at this https URL
[CV-54] CrossPressor: Scalable Cross-View Fusion for Feed-Forward 3D Gaussian Splatting WACV2027
链接: https://arxiv.org/abs/2611.06258
作者: Yuxi Hu,Wenhao Dong,Jun Zhang,Friedrich Fraundorfer
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV 2027
Abstract:Feed-forward 3D Gaussian Splatting (3D-GS) becomes increasingly difficult to scale to long multi-view sequences. Anchor-based compression methods such as ZPressor alleviate representation growth by decoding Gaussians only from a compact set of anchor feature maps, but keeping these anchors fixed forces each one to aggregate an increasing number of support views. We identify this representation–fusion mismatch and introduce \textbfCrossPressor, a training-free method for scalable anchor-based fusion. CrossPressor preserves all decoder-facing anchor queries while constraining the view-growing support sequence through coverage-cycled evidence routing, distributing retained spatial evidence across support views and fusion layers while keeping all support views represented. Routing before K/V projection, together with joint context packing, further reduces support-side fusion cost without retraining. CrossPressor maintains reconstruction quality close to dense fusion while achieving up to 26.12\times faster cross-view fusion and 1.43\times end-to-end acceleration on long input sequences. Our project page is at: this https URL.
[CV-55] CAPE-JEPA - Cross-modal Adversarial Probabilistic Embedding JEPA BMVC2026
链接: https://arxiv.org/abs/2611.06247
作者: Aniket Junghare,Biplab Banerjee
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted for publication at the British Machine Vision Conference (BMVC 2026)
Abstract:Cross-modal remote sensing image retrieval (RS-CMIR) requires representations that are simultaneously modality-invariant and semantically discriminative, despite the profound physical gap between heterogeneous sensors such as Synthetic Aperture Radar (SAR) and multispectral imagery. Recent JEPA-based Predictive self-supervised approaches such as X-JEPA formulate cross-modal alignment as a deterministic latent forecasting task, implicitly assuming uniform patch predictability and leaving cross-modal distributional alignment unaddressed. We propose CAPE-JEPA, a probabilistic cross-modal Joint Embedding Predictive Architecture with three contributions: (i) a cross-attentional Gaussian predictor outputting per-patch parameters (mu, sigma^2) conditioned on opposite-modality context, trained via negative log-likelihood to model inherent cross-sensor prediction uncertainty; (ii) adversarial distributional regularization combining unimodal isotropic Gaussian regularization with a novel cross-modal adversarial variant that aligns modality distributions via worst-case Epps-Pulley characteristic function tests without explicit sample pairing; and (iii) in-training contrastive learning with attention-pooled heads for scene-level discriminative grounding. Evaluated on BigEarthNet-14K (BEN-14K) and fMoW across SAR, multispectral, and RGB modalities, CAPE-JEPA surpasses X-JEPA by up to 13.74 points on BEN-14K and 15.59 points on fMoW, outperforming all baselines including SkySense and CROMA across both unimodal and cross-modal directions while using fewer parameters (407.30M vs. 432.72M).
[CV-56] Homology Features for X-Ray Hyperspectral Images Segmentation
链接: https://arxiv.org/abs/2611.06230
作者: Fethi Harkat,Guillaume Gey,Valérie Perrier(LJK, CVGI),Kévin Polisano
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Characterizing Multi-Energy X-ray Transmission (XRT) hyperspectral images presents significant challenges for traditional classification frameworks due to high-dimensional redundancy and non-linear spectral overlaps. Although manifold learning techniques such as UMAP enhance spectral representation by preserving non-linear correlations, they do not explicitly utilize the geometric structure of the embedded manifold for segmentation. This study introduces a hybrid architecture that integrates topological features derived from persistent homology with convolutional and graph neural networks. The approach computes Vietoris-Rips persistent homology on UMAP-projected superpixel point clouds to capture H 1 cyclic structures that encode spectral heterogeneity. These topological descriptors are combined with pixel-level CNN features through a broadcasting mechanism that propagates superpixel-level graph representations to full spatial resolution. Experimental results on XRT multi-energy imagery demonstrate that the proposed hybrid model significantly outperforms both CNN-only and superpixel-pooling fusion baselines, suggesting that topological features provide complementary discriminative information to local spectralspatial descriptors.
[CV-57] FedXCrop: Privacy-Preserving Crop Disease Detection via Federated Learning with Post-Aggregation Explainability under Non-IID Farm Distributions KR
链接: https://arxiv.org/abs/2611.06221
作者: Kriti Saini,Manikandan P
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages. Includes 38-class crop disease classification experiments, federated learning evaluation under non-IID client distributions, and quantitative explainability analysis using Grad-CAM and SmoothGrad. Code and experimental results: this https URL
Abstract:Deep learning can classify crop diseases accurately from benchmark leaf images, but centralized training requires farms to share raw data, and predictions without explanations offer limited support for treatment decisions. We present FedXCrop, a federated crop disease classification pipeline that quantitatively evaluates explanations of the aggregated global model. A MobileNetV2 model trained centrally on PlantVillage achieved 99.38% test accuracy (standard deviation 0.05 over three seeds) on a fixed test set of 5,428 images across 38 classes. Federated models initialized with ImageNet weights were trained for 30 rounds across five clients. FedAvg achieved 98.81% accuracy with identically distributed clients and 98.81%, 98.59%, and 97.78% under Dirichlet label skew (alpha = 1, 0.5, and 0.1, respectively). FedProx was not significantly different from FedAvg at any heterogeneity level (McNemar p 0.3). Grad-CAM and SmoothGrad explanations were evaluated on 380 test images. Both attributed more energy to the leaf than chance (0.70-0.78 versus 0.60), while Grad-CAM maps from federated and centralized models showed Spearman correlations of 0.74-0.76. Agreement decreased as heterogeneity increased, lesion localization was only modestly above chance, and a model-randomization test indicated that SmoothGrad maps depended substantially on image structure rather than learned parameters. In this setting, label skew reduced federated accuracy by 0.6-1.6 percentage points relative to centralized training. FedAvg largely preserved Grad-CAM behavior, while FedProx provided no measurable benefit.
[CV-58] Cross-Modal Semantic Successive Refinement for Generative Image Communication over Underwater Acoustic Channels
链接: https://arxiv.org/abs/2611.06205
作者: Xu Pei,Bin Wang,Guangxu Zhu,Yi Zhang,Haiyan Fan
类目: Information Theory (cs.IT); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Underwater acoustic communication faces challenges such as limited available bandwidth and complex channel environments, which impose demanding requirements on efficient image transmission and reliable reconstruction. To address these challenges, this paper proposes a generative image communication framework based on cross-modal semantic successive refinement. At the transmitter, text descriptions, compact bottleneck features, and visual side information compressed by residual vector quantization are transmitted. At the receiver, text and bottleneck features are first fused to predict a base estimate of the intermediate visual representation. The received RVQ indices are then used to recover supplementary visual information and perform residual correction, followed by image reconstruction using a one-step Pixel Mean Flow model. During training, a CLIP-based text-image semantic consistency constraint is introduced to preserve the semantic correspondence between generated results and the original text descriptions. During transmission, protection resources are allocated according to the importance of different information branches and RVQ quantization stages. Simulations using COCO 2017 images and measured underwater acoustic channel records from Watermark show that text and bottleneck features provide complementary semantic and structural conditions, RVQ visual side information further improves reconstruction quality, and layered unequal error protection helps improve recovery at low SNRs. Code-rate allocation experiments reveal a tradeoff between the reliability of fundamental conditions and visual-detail reconstruction under a limited visual-feature coding budget, providing guidance for allocating protection resources in underwater acoustic image transmission.
[CV-59] BiCG-Net: Target-Specific Image Quality Assessment via Observation Modeling and In-Image Differencing
链接: https://arxiv.org/abs/2611.06203
作者: Yibu Yang,Tingqiong Cui,Liming Lei,Hao Xiong,Jun Xu,Siyuan Liu,Guanghui Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 34 pages, 3 figures
Abstract:We estimate the quality of a specific target region (e.g., a train body) in an image, rather than the whole image. This differs fundamentally from whole-image No-Reference Image Quality Assessment (NR-IQA): when the target occupies a small fraction of the frame, the whole-image score is dominated by background, and no public benchmark provides target-level human Mean Opinion Score (MOS). We estimate target quality via nested sampling - a sequence of patches expanding outward from the target box - and infer quality from the relative structure of this sequence. This motivates BiCG-Net, a student network that distills an 11-teacher ensemble through a linear observation model and an in-image differencing mechanism, which provably cancels the target term and teacher-specific bias. We train and test the student independently on each dataset, and evaluate zero-shot transfer across benchmarks. On CSIQ and LIVEC, the student reaches PLCCs of 0.9009 and 0.9133 against human MOS, exceeding the best single teacher and the teacher consensus. On a self-built MOS-free railway dataset, the student reaches a PLCC of 0.9855 against a weak label. The student exceeds its teachers because it learns a new mapping from the relative structure of teacher readings to the ground-truth label, rather than mimicking them. Residual analysis on CSIQ and LIVEC shows the prediction residual is significantly correlated with background contamination, teacher dispersion, and label extremity, demonstrating the student extracts information the consensus does not contain. Under CSIQ-to-LIVEC transfer, the frozen student reaches a PLCC of 0.8416, exceeding the consensus by 0.1947. We position BiCG-Net as an offline target-level quality auditing and label-generation engine; its 39,482-parameter student is compact by design, acting as a unifying layer mapping heterogeneous teacher readings onto a common quality scale. Comments: 34 pages, 3 figures Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2611.06203 [cs.CV] (or arXiv:2611.06203v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.06203 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-60] Annotator-Free Object Detection via Foundation Model Collaboration
链接: https://arxiv.org/abs/2611.06195
作者: Yasiru Ranasinghe,Vishal M. Patel
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Object detection research remains constrained by high annotation costs and the closed-set assumptions of supervised models, limiting scalability to novel categories. We present a target-dataset annotation-free object detection framework that eliminates the need for human labeling on the target dataset by leveraging the complementary strengths of pretrained foundation models, including segmentation and vision-language systems, combined with a self-refining training loop. To further reduce reliance on expensive LVLM queries, we introduce a landmark-based clustering mechanism that amortizes labeling across image clusters, improving label consistency while lowering computational cost. Our method autonomously generates segmentation proposals, assigns semantic labels through multimodal reasoning, and iteratively improves both localization and classification using a lightweight open-set detector. This modular pipeline enables adaptation to completely unseen object classes and domains without any human-provided target-dataset annotations, while maintaining compatibility with existing detection architectures. Evaluations on COCO, LVIS, ODinW-35, and Roboflow-100 demonstrate substantial performance gains over state-of-the-art open-vocabulary and long-tail detection methods, achieving 60.8 AP on COCO and 47.4 AP on LVIS under zero-shot settings. Beyond strong benchmark results, our approach provides a scalable foundation for real-world detection systems in unknown environments, where annotation is infeasible.
[CV-61] Rohbau3D-Semantics: Dense Semantic and Instance Labels for Point Clouds of Indoor Shell Construction Scenes WACV2027
链接: https://arxiv.org/abs/2611.06180
作者: Lukas Rauch,Thomas Braml
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:We present the Rohbau3D-Semantics dataset, which adds dense semantic and instance-level labels to laser scans of the previously released Rohbau3D domain data from shell construction sites. The dataset comprises 504 point cloud scenes from 14 survey sites, each captured with a terrestrial laser scanner and stored in a scanner-centered local coordinate system. Unlike point clouds sampled from reconstructed meshes or synthetic building models, the data preserve the characteristic properties of real laser scanning, including non-uniform sampling, measurement artifacts, reflections, and fine geometric detail. Combined with manual labels, expert oversight during pointwise annotation, and full-scene review, this yields a high-quality benchmark for semantic scene understanding in complex built environments. The dataset includes 17 class labels of building components and one background class, along with instance annotations for object-level analysis. In addition to the 3D point clouds, the release also provides scripts for deriving geometrically aligned panoramic and cube-map representations of color, LiDAR reflection intensity, depth, surface normals, semantic labels, and instance labels from the 3D arrays. We describe the acquisition pipeline, annotation protocol, quality-control process, and panorama-generation procedure, and provide benchmark splits and baseline experiments for semantic segmentation. By combining precise terrestrial laser scans, dense annotations, and a wide variety of shell-construction scenes, Rohbau3D-Semantics bridges the gap among synthetic indoor datasets, mesh-derived benchmarks, and smaller construction-specific resources. The dataset is intended to support research on 3D scene understanding, construction-site analytics, and multimodal scene analysis in the built environment. Access and download the full dataset via this https URL Comments: Accepted at the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2027) Subjects: Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2611.06180 [cs.CV] (or arXiv:2611.06180v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.06180 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-62] Reasoning -Aware Cross-Modal Association Unlearning in Multimodal Large Reasoning Models
链接: https://arxiv.org/abs/2611.06168
作者: Xinghao Li,Chenchen Tan,Youyang Qu,Shujie Cui,Xingliang Yuan,Cunjian Chen,Chi Liu,Longxiang Gao
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Multimodal large reasoning models (MLRMs) integrate visual inputs with textual and parametric knowledge, allowing sensitive information to be accessed through cross-modal associations during reasoning. This raises privacy concerns when a visual input should no longer provide access to such information. Machine unlearning can remove this influence without retraining the model from scratch. However, multimodal unlearning methods mainly target samples, concepts, or knowledge content, without separating the knowledge from the cross-modal association through which it is accessed. In reasoning models, target information may disappear from the final answer while remaining in intermediate reasoning, or be replaced by incorrect substitute knowledge. We therefore study cross-modal association unlearning, where the forget target is a designated visual-to-knowledge association: the model should suppress target-bearing behavior under the designated visual context while preserving access to the same knowledge in permitted contexts. We propose Visual-Keyed Counterfactual Rebinding (VKCR), a parameter-efficient framework for cross-modal association editing. VKCR first localizes the edit to a compact association-conditioned parameter subspace. It then reduces support for the designated target at task outputs and counterfactually rebinds the affected reasoning behavior by discouraging target-bearing or unsupported continuations while favoring valid non-sensitive alternatives. We further introduce CLEAR-R and MLLMU-R to evaluate target leakage, hallucinated reassociation, reasoning preservation, and modality-specific utility. Experiments show that VKCR substantially reduces image-conditioned target leakage and hallucinated reassociation, including across heterogeneous task formulations, while preserving non-sensitive visual reasoning and permitted text-only knowledge.
[CV-63] DSRS: Teaching MLLM s to Perceive Reason and Segment in Complex Driving Scenes
链接: https://arxiv.org/abs/2611.06160
作者: Sicen Guo,Zhengfei Song,Kuanhsun Chen,Xu Huang,Ping Zhong,Wei Ye,Shaoyi Du,Rui Fan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 26 pages and 19 figures
Abstract:Existing reasoning segmentation frameworks typically employ multimodal large language models (MLLMs) with broad world knowledge to extract semantics in a human-like manner and generate corresponding segmentation masks. Nevertheless, these frameworks lack the necessary scene-specific awareness. This paper introduces driving scene reasoning segmentation (DSRS), a scene-specific task that requires models to interpret implicit instructions, infer spatial relations, detect potential risks, handle challenging conditions, and comprehend underlying traffic rules, ultimately generating accurate segmentation masks. Meanwhile, we establish a benchmark dataset to facilitate future research on the DSRS task. To pioneer this task, we propose DSRS-Net, a novel architecture with advanced reasoning segmentation capability for automated driving. Specifically, DSRS-Net seamlessly infuses geometric features for deeper reasoning segmentation, without compromising the original multimodal fusion mechanism. This strategy effectively mitigates the modality representation mismatch in MLLMs caused by RGB-only pretraining, while enhancing the integration of geometric, contextual, and linguistic features. Extensive experiments demonstrate the superior performance of DSRS-Net in both driving scene reasoning segmentation and other language-driven segmentation tasks across various public datasets. The source code and dataset will be released upon publication.
[CV-64] Grounding in the Dark: Text-Only Training for Zero-Shot Video Temporal Localization
链接: https://arxiv.org/abs/2611.06149
作者: Divyanshu Mishra,Sabine Sternig,Rakshith Shetty,Erhan Gundogdu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Video-LLMs are typically trained on paired video-text data for temporal tasks, yet collecting precise temporal annotations and training on large-scale video datasets are both costly. We investigate whether a text-only LLM, with no vision-language pretraining, can be trained on textual video descriptions alone to perform temporal grounding and receive video only at inference time. This setting removes paired video data from training altogether and is a step toward training multimodal models without paired annotations. We generate a 134K-example corpus of textual sequences with up to 100 time steps, adapt Qwen3-8B with LoRA on CLIP or LLM2CLIP text embeddings, and evaluate on Charades-STA and QV-Highlights without training on videos from either benchmark. LLM2CLIP, whose text encoder supports longer inputs, represents detailed descriptions better than CLIP (61.54 versus 53.86 mtIoU on Temporal-TextVid) and transfers them better to video. Still, replacing text with frame embeddings at inference yields only 4.99 mtIoU on Charades-STA and 11.12 on QV-Highlights, revealing a large modality gap. To bridge it, we propose Nearest Entity Replacement (NER), a training-free mapping that represents each video frame by the normalized mean of its nearest text-entity embeddings, so that the model receives the kind of input it was trained on. NER reaches 14.45 mtIoU on Charades-STA and 20.72 on QV-Highlights, outperforming the strongest of four families of modality-gap baselines by 5.00 and 1.71 points. NER thus enables partial transfer from text to video, while more reliable grounding will likely require encoders with better-aligned text and visual representations.
[CV-65] PTQ4SAM3: Semantics-Preserving Post-Training Quantization for Segment Anything Model 3 WACV2027
链接: https://arxiv.org/abs/2611.06143
作者: Xi Xiao,Yunbei Zhang,Lin Zhao,Janet Wang,Xinru Jiang,Yanshu Li,Chenrui Ma,Pan Wang,Tianyang Wang,Xiao Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 4 figures. Accepted to WACV 2027. Includes supplementary material
Abstract:Segment Anything Model 3 (SAM 3) can segment objects from text prompts and track them across video, but its memory and computation costs make deployment difficult. Post-training quantization (PTQ) reduces these costs by using fewer bits. However, a quantized model can select the wrong object or lose track of its target across frames. We call this prompt-alignment drift. Controlled tests show that perturbing low-variance feature directions strongly affects target selection. We propose Semantics-Preserving Quantization (SPQ), which gives more calibration weight to concept-sensitive parameters, preserves residual features, and corrects the mean error in video memory. Fixed text concepts guide calibration without task annotations. All three components operate offline and add no inference operators. With a W4A8 image encoder, SPQ improves RefCOCO Pr@0.5 from 42.1 to 72.5 over PTQ4SAM while retaining 54.8 COCO mask AP. Experiments on LVIS, RefCOCO variants, and SA-V show consistent gains. A CLIP study also shows improved image-text retrieval.
[CV-66] RDT-3D: Resolution-Decoupled Training for High-Fidelity 3D Geometry Generation
链接: https://arxiv.org/abs/2611.06141
作者: Jiale Zhang,Yang Liu,Chengyang Li,Jin Yuan,Qianxi Jia,Taicheng Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:High-fidelity 3D geometry generation requires maintaining global structural coherence alongside fine-grained surface details. While recent coarse-to-fine frameworks leverage Rotary Position Embeddings (RoPE) to condition local queries, we reveal a critical training-inference resolution discrepancy: coordinate guidance and geometric supervision are aligned during training, whereas inference decodes high-resolution geometry from coarse spatial coordinates. This mismatch introduces a severe distribution shift that degrades fine details. To overcome this, we propose Resolution-Decoupled Training (RDT), which anchors RoPE guidance at 512 resolution while supervising geometry at 1536, enabling robust detail extrapolation. Additionally, to eliminate high-resolution data bottlenecks, we present Dual3D-300K, a large-scale paired dataset comprising 300K watertight 3D point clouds explicitly aligned across 512 and 1536 resolutions. Extensive experiments demonstrate that RDT substantially improves sharp geometric regions, while Dual3D-300K offers a scalable benchmark for high-resolution 3D generation.
[CV-67] ReactWorld: Structuring Interaction Responses for Embodied Video Generation
链接: https://arxiv.org/abs/2611.06120
作者: Hanwei Liu,Huijia Zhang,Hanzhang Wang,Guangyu Liu,Hanbei Zhan,Linjie Hu,Bin Xiao,Xuan Zhou,Dong Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
备注: 10 pages, 7 figures. Project page: this https URL
Abstract:Embodied video world models predict scene changes caused by robot actions, yet generated object responses may remain inconsistent with task intent and contact dynamics. Critical responses in embodied scenes often concentrate in local regions and specific time intervals. Inspired by chain-of-thought reasoning, we introduce ReactWorld, an embodied video generation model guided by structured interaction anticipation. In the generation backbone’s feature space, an interaction response field predictor anticipates how the interaction will unfold, explicitly describing participant roles, future motion response intensity, and local motion changes. A response feedback adapter forms increments to object features from the difference between the object branch’s outputs with predicted and zeroed local dynamics inputs. It also converts predicted displacements into positional biases for temporal attention. These two mechanisms guide object feature updates and information retrieval across frames, respectively. Through interaction anticipation and generative feedback, ReactWorld achieves an average score of 77.51 on RBench, outperforming PhysisForcing by 3.29 points with improvements in all nine categories. On EZS-Bench, it achieves a Domain score of 86.98 and an average score of 81.87.
[CV-68] Semantic Prior-Guided Incomplete Cross-Modal Retrieval with Noisy Labels
链接: https://arxiv.org/abs/2611.06113
作者: Jiali Chen,Ruitao Pu,Yingke Chen,Dezhong Peng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Cross-modal retrieval (CMR) aims to bridge the semantic gaps between modalities by learning shared representations. Unfortunately, due to the dynamism and inconsistencies in real-world data acquisition, multimodal datasets frequently encounter severe modality incompleteness. Existing incomplete CMR methods adhere to a label-driven paradigm, relying heavily on categorical annotations as the semantic benchmark. However, label noise is difficult to avoid and often coexists with modality incompleteness. To address this challenge, we propose a novel semantic prior-guided framework for incomplete cross-modal retrieval with noisy labels, termed SPIRN. Specifically, SPIRN uses fixed semantic priors as a consistent reference for robust representation learning, noisy-label handling, and missing-modality completion. Based on this reference, Semantic Priors Guidance (SPG) establishes a stable category-level semantic space, Quality-aware Robust Learning (QRL) dynamically evaluates sample quality to distinguish clean samples from noisy ones and handles noise via reverse learning, Prior-guided Global Label Calibration (PGLC) rectifies corrupted annotations for refined supervision, and Confidence-driven Selective Imputation (CSI) recovers missing modalities from a cross-modal memory bank while filtering unreliable imputations with an increasing threshold. To the best of our knowledge, this is the first framework to jointly tackle the dual challenges of modality incompleteness and label noise in the CMR community. Extensive experiments on four widely used benchmark datasets demonstrate that SPIRN outperforms 11 state-of-the-art methods in various scenarios involving both missing modalities and noisy labels.
[CV-69] XuanJian-DINOSAR: A SAR Foundation Model Based on Global and Local Semantic Consistency
链接: https://arxiv.org/abs/2611.06105
作者: Chunxiao Wu,Huixin Huang,Jialei Pan,Qiang Liang,Tao Wu,Cheng Wu,Long Huang,Yuanji Li,Qiang Cheng,Aifang Liu,Linghao Xia,Rui Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Self-supervised pre-training on synthetic aperture radar (SAR) imagery enables transferable representations, but the optical-SAR domain gap and speckle perturbations limit representation stability. Most SAR foundation models use masked image modeling (MIM) through pixel reconstruction or structural feature prediction, which is sensitive to speckle or relies on predefined operators. Joint modeling of cross-view semantics and perturbation stability at corresponding spatial locations remains underexplored. We propose XuanJian-DINOSAR, a SAR foundation model based on global and local semantic consistency. We curate over 360K publicly available SAR target chips, detection images, and scene images spanning diverse targets, scenes, and sensors, and adapt an optically pre-trained DINOv2 model to SAR through self-supervised learning. The framework combines DINO global semantic self-distillation, iBOT local masked modeling, and KoLeo feature-distribution regularization to learn globally discriminative and locally structure-aware representations. We introduce a local semantic consistency branch tailored to SAR intensity statistics and coherent imaging characteristics. Shared geometric transformations produce spatially aligned teacher weak views and student strong views. Intensity transformations and Gamma multiplicative speckle perturbations are applied to student views, while a cosine consistency loss aligns patch tokens at corresponding locations, promoting stability under radiometric perturbations. This objective complements the cross-region and cross-scale semantic consistency of DINOv2 multi-crop training, jointly enforcing global and local semantic consistency. Experiments on multiple downstream classification and detection datasets demonstrate state-of-the-art performance on most evaluation metrics, validating the generality and transferability of the learned representations.
[CV-70] Many Pixels One Composition: Self-Supervised Spectra for Specimen-Level Prediction
链接: https://arxiv.org/abs/2611.06096
作者: Eleftheria Tetoula-Tsonga,George Arvanitakis,Theodoros Giannakas
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Can hyperspectral (HS) imaging be used to predict the elemental composition of physical specimens found on the Earth’s surface? We approach this problem as a supervised regression task, using an open dataset of specimens, each comprising a HS cube and elemental characterization. In particular, we train an encoder in a self-supervised fashion that operates on pixel-level; then freeze the encoder and train a lightweight attention-based regression head in a supervised manner, by pairing many representations of a given specimen with its corresponding elemental label. In test specimens, our approach significantly outperforms traditional spectral matching and unmixing baselines. It also outperforms supervised learning models based on the mean spectrum (of similar model capacity), while full model fine-tuning achieves the strongest performance, at substantially higher model capacity.
[CV-71] Drift-adaptive Representation Preservation for Domain-Incremental Learning ACCV2026
链接: https://arxiv.org/abs/2611.06093
作者: Hyunmin Kong,Gyutae Oh,Jitae Shin
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to ACCV 2026
Abstract:Pretrained vision models provide strong general-purpose representations, but in Domain-Incremental Learning (DIL), where the input distribution changes sequentially, adapting to new domains often degrades performance on previously learned ones. We observe that forgetting is not uniform across domains and classes and is not explained by training order alone. This motivates us to track each past domain-class concept individually. We use representation drift, defined as the deviation of a past concept’s current representation from a fixed anchor, as a measurable surrogate for forgetting risk. The measured drift then serves two complementary roles. A consolidation loss assigns stronger preservation to high-drift pairs, while prototype construction down-weights highly drifted domains to reduce their influence on the class reference. Preservation is thus allocated by measured drift rather than assumed importance. To obtain this signal during training without storing past raw data, we repurpose a Joint-Embedding Predictive Architecture (JEPA). For each domain-class pair, we store only a compact context summary in the backbone feature space and rerun it through the learned context-to-target relation to measure its drift from the fixed anchor. Across DomainNet, Office-Home, CORe50, and CDDB-Hard, our method sustains a strong accuracy-forgetting balance on every benchmark, using substantially less training memory than accuracy-competitive methods and the smallest storage among compared methods.
[CV-72] STARE: Scene Text-Aware multimodal Retrieval for Everyday objects ACCV2026
链接: https://arxiv.org/abs/2611.06089
作者: Kento Tokura,Ryosuke Korekata,Daichi Yashima,Kei Katsumata,Kanon Amemiya,Miyu Goko,Komei Sugiura
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: ACCV 2026 MM4SciReal Workshop
Abstract:In this study, we focus on service robots that retrieve and transport everyday objects based on free-form instructions that may refer to visual features and scene text. This is challenging because it requires a multimodal understanding of the visual features and spatial relationships of objects, as well as scene text embedded in the environment. To address this, we propose Scene Text-Aware multimodal Retrieval for Everyday objects (STARE), which leverages visual and textual cues in response to free-form instructions. STARE introduces Crosslingual Visual Prompt to enhance scene text interpretation, and models the complex relationships between named entities in the instructions and their corresponding objects. We evaluate STARE using two novel benchmarks, GoGetIt and TextCaps-test, both of which consist of real-world indoor/outdoor images and instructions involving object manipulation and navigation. The results show that STARE outperforms all baseline methods on standard multimodal retrieval metrics. Furthermore, in physical experiments conducted under a zero-shot transfer setting, STARE achieves an overall success rate of 80%. Our project page and code are provided in the Supplementary Materials.
[CV-73] GeoBiaset: A Counterfactual Benchmark for Demographic Bias in World-Level Geolocalization NEURIPS2026
链接: https://arxiv.org/abs/2611.06085
作者: Niccolò Niccoli,Silvia Dani,Marco Mistretta,Andrew D. Bagdanov,Marco Bertini,Lorenzo Seidenari
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at Neurips 2026, ED track. Project page: this https URL
Abstract:World-level image geolocalization is an increasingly important task, yet the biases of current models remain poorly understood. Prior work has identified coarse dataset imbalances noting, for instance, that some cities are absent from training sets and that models tend to perform better in wealthier regions. However, no one has yet examined which visual cues drive these disparities. To do so, we introduce GeoBiaset, the first benchmark designed to measure how geolocalization models are influenced by the apparent race of a person inserted into the image foreground. The dataset consists of 9,541 images, pairing existing geotagged background images with counterfactual edits spanning seven races, holding the ground-truth coordinates fixed. Controlled masking experiments show that occlusion and background drift do not fully explain geolocalization degradation. Alongside the benchmark, we introduce the Person-Origin Perturbation Indicators (POPI), novel metrics that quantify model resilience to these demographic perturbations by measuring whether predictions are pulled toward the region associated with the inserted group. Our evaluation of state-of-the-art geolocalization and multimodal large language models reveals a 16-47% drop in continent-level performance when a subject from a demographically mismatched group is inserted. A landmark versus no-landmark analysis and gradient-weighted heatmaps show that models preferentially attend to geographic landmarks when present, but shift attention to facial features in their absence.
[CV-74] SleuthVLM: Do We Really Need Audio-in-Video for Omni Video Understanding?
链接: https://arxiv.org/abs/2611.06083
作者: Peiran Wu,Fangfei Gou,Zhaoxi Mu,Changhao Zhang,Jian Liu,Weiqiang Wang,Yuntao Wang,Yuanchun Shi,Junxiao Shen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Omni models answer questions about videos by jointly processing sampled frames and audio in one turn, with progress driven largely by scaling model size and omni training data. We ask whether a model needs to hear the audio directly at all, and present SleuthVLM, an agent harness built on a VLM that never hears the audio directly: speech will be transcribed into text, and the VLM will sleuth the video on its own without any predefined workflow. On four omni benchmarks with videos up to 90 minutes, SleuthVLM achieves the best average accuracy: \textbf75.0%, 2.6 points above Gemini 3.8 Flash, and it also tops the Video-MME-v2 leaderboard with \textbf77.8%; with an open-source 27B VLM, it beats all open-source omni models. We draw three findings: SleuthVLM beats a model that glances; hearing is not the bottleneck, agentic ability is; omni understanding needs better agentic ability, not just more omni data for modality alignment. So we provide a simple but effective starting point for doing omni video understanding. Omni understanding needs better agents, not just bigger models that hear.
[CV-75] RouTex: Multi-reference 3D Texture Generation by Reprogramming Rotary Position Encoding
链接: https://arxiv.org/abs/2611.06070
作者: Chengyang Li,Yang Liu,Jiale Zhang,Qiang Rao,Wei Zhang,Taicheng Huang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:While single-image 3D texture generation has advanced significantly, conditioning texture synthesis on multi-view reference images remains largely under-explored, despite its necessity for resolving surface ambiguity in unobserved regions. In this paper, we propose RouTex, a novel multi-reference 3D texture generation framework based on in-context Diffusion Transformers (DiT). Our key insight is that native 3D Rotary Position Encoding (RoPE) inherently possesses spatial addressing capabilities. We introduce Positional Routing (PR) to reprogram 3D RoPE along spatial and semantic dimensions, establishing explicit reference-to-view correspondence without introducing additional parameters. To execute this routing efficiently, Dual-Granularity Attention (DGA) combines per-view local attention for fine texture fidelity with global joint attention over deduplicated references for cross-view consistency. Extensive experiments demonstrate that RouTex achieves new state-of-the-art performance in multi-reference 3D texture generation.
[CV-76] ouchéWorld tech report : Towards Online Unified Camera–Hand Exploration in Generative Worlds
链接: https://arxiv.org/abs/2611.06061
作者: Jiaao Yu,Yanjian Zhang,Haoyang Jiang,Guillaume Wisniewski,Nadi Tomeh,Shijie Jia
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Recent video world models allow users to explore generated environments through camera navigation. Continuous hand motion provides a fine-grained action channel for interacting with the people and objects encountered during exploration. We present TouchéWorld, a framework that integrates gesture-driven interaction into a camera-controlled video world model. Given an initial image and a scene-description prompt, the model generates video conditioned on temporally aligned camera and hand controls. We construct joint camera–hand training data from heterogeneous egocentric videos through annotation completion, consistency-based filtering, and temporal alignment. Building on a pretrained camera-controlled world model, we combine spatial hand-skeleton conditioning with numerical hand-motion features through parameter-efficient adaptation. Causal adaptation and few-step distillation support chunk-level online generation, allowing users to update their controls based on generated visual feedback. We further introduce TouchéBench to evaluate hand-motion following, camera–hand coordination, online interaction, and object response. Together, these components establish a framework for studying continuous exploration and hand-driven interaction within generative worlds.
[CV-77] RI-Prune: Relation and Information Aware Visual Token Pruning for Multi-Image Vision-Language Models CIKM2026
链接: https://arxiv.org/abs/2611.06047
作者: Dongjin Lee,Eunju Lee,Jinhee Jang,Yoonji Lee,Youngbin Kim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to CIKM 2026 Oral
Abstract:Large vision-language models (LVLMs) have recently been extended to handle multiple input images, but this extension dramatically increases the visual token budget, making visual-token pruning not merely an efficiency option but a practical necessity. Existing visual-token pruning methods, primarily designed for the single-image setting, rely on attention-derived saliency signals to estimate per-token importance. However, this attention-centric design transfers poorly to the multi-image regime: it causes substantial performance degradation on multi-image benchmarks, suffers from strong positional bias, and loses discriminative attention signals as the number of input images increases. To test whether attention-selected tokens are truly important for multi-image reasoning, we conduct two complementary causal probes Leave-One-In and Leave-One-Deletion, which measure token importance by restoring individual tokens to a zeroed visual sequence or removing them from the full sequence. Attention-based pruners provide only marginal gains over random selection, showing that attention is a weak proxy for visual-token importance in multi-image reasoning. Motivated by this gap, we propose Relation and Information Aware Pruning (RI-Prune), a training-free, attention-free pre-decoder selector for multi-image visual token pruning. RI-Prune scores each visual token using a z-normalized combination of three signals computed from projector outputs: query relevance, cross-image correspondence with compact per-image prototypes, and intra image representativeness. By integrating these signals, RI-Prune performs balanced per-image Top-K selection to preserve information across input images. Across multiple multi-image benchmarks and three LVLM backbones, RI-Prune achieves a stronger trade-off between token reduction and task performance than prior pruning methods, particularly under aggressive token budgets.
[CV-78] DeepfakeGenome: Toward Next-Generation Deepfake Attribution
链接: https://arxiv.org/abs/2611.06046
作者: Shunli Wang,Xinyu Zhou,Yandan Zhao,Wenbin Liang,Taiping Yao,Ke-Yue Zhang,Zhiyuan Yan,Shouhong Ding,Lizhuang Ma
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 28 Pages
Abstract:In recent years, AIGC technologies are capable of generating hyper-realistic forged facial images, posing severe threats to facial security safety. To tackle these challenges, early Deepfake research primarily centered on real-fake classification. Recently, growing efforts have been devoted to the Deepfake Attribution (DFA) of generated content. Nevertheless, existing research on Deepfake detection and attribution faces saturated performance in binary classification, limited diversity in datasets and algorithms, and imperfect evaluation protocols, which severely impede practical application. To address these limitations, we propose a comprehensive deepfake detection and attribution benchmark named DeepfakeGenome (DFG). It contains 100 facial forgery algorithms and 2M images in total, achieving 4x to 100x larger than prior DFA benchmarks. We further design 4 protocols for practical evaluation, including a novel retrieval-based attribution paradigm. Unlike previous open-set evaluation metrics, the proposed retrieval metrics are more aligned with the real-world active defense situation of blacklist registration mechanisms. Based on these elaborate designs, we investigate the performance ceiling of deepfake attribution task. Over 2k experimental evaluations are conducted, and 10 insightful findings are derived. We believe this work can provide new insights into the DFA research field. The Code and Benchmark of DFG are avaliable on this https URL.
[CV-79] UniRestorer: Scalable Universal Image Restoration with Multi-Granularity Degradation Modeling
链接: https://arxiv.org/abs/2611.06042
作者: Jingbo Lin,Zhilu Zhang,Fan Li,Renjing Pei,Wenbo Li,Hang Xu,Hongzhi Zhang,Wangmeng Zuo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Universal image restoration aims to handle diverse degradations within a single model, but a reliable scaling strategy remains underexplored. Increasing backbone depth and width can lead to performance saturation, while degradation-aware specialization depends on accurate degradation estimation and is vulnerable to errors on out-of-distribution and real-world inputs. We propose UniRestorer++, a scalable framework based on multi-granularity degradation modeling. It progressively refines degradation groups and introduces group-specific experts, coupling capacity growth with specialization. Degradation-estimation-based routing selects a root-to-leaf path, while granularity-estimation-based routing chooses a level along that path to balance specialization and robustness. Spatially dense degradation representation maps further enable region-aware routing for spatially non-uniform corruptions. We develop CDD-50, a benchmark covering 50 degradation categories from single to penta-degradation. Experiments show approximately log-linear PSNR trends over the evaluated model-size range and stronger scaling gains on out-of-distribution and real-world data than vanilla MoE scaling. UniRestorer++ achieves state-of-the-art performance across benchmarks and demonstrates zero-shot generalization and few-shot transfer on real-world images. Code: this https URL
[CV-80] USI-Brain: Enhancing Underwater Spatial Intelligence via Spatial Data-Driven Vision-Language Modeling
链接: https://arxiv.org/abs/2611.06040
作者: Zhiliang Zhu,Zhiqiang Zhao,Xie’an Jiang,Yingxin Lai,Kangye Zheng,Dawei Zhao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 13 pages, 4 figures, 10 tables
Abstract:Underwater spatial reasoning depends on object boundaries and geometric cues that are obscured by scattering, color distortion, and low visibility. Existing underwater resources capture these cues through images, annotations, and depth maps, but provide fragmented supervision for vision-language models. We present a spatial data engine that connects visual appearance with object-level geometry through enhancement, 2D and geometry-assisted question generation, and expert filtering. It produces 4.27M QA pairs across five capability groups. We use this corpus to train USI-Brain and introduce an Underwater Prior-guided Rectifier (UPR) that applies gated corrections to visual features. USI-Bench evaluates the resulting capabilities with 6,525 held-out QA pairs. On its 15 task metrics, USI-Brain and USI-Brain + UPR improve over Qwen3-VL-8B-Instruct by 5.76 and 7.63 points on average, respectively, using signed differences on 100-point scales. Ablations show that geometry-assisted supervision improves depth and ordering by 4.8 and 6.1 points with UPR disabled, while UPR adds 18.0 points in scene description under matched training data. Both variants use raw images at evaluation. Across NautData, UWBench, and MarineEval, USI-Brain achieves the best or joint-best score on 28 of 32 metrics. Model weights and USI-Bench will be released at this https URL.
[CV-81] METRIC-GS: Selective Metric-Scale Estimation for 3D Gaussian Splatting
链接: https://arxiv.org/abs/2611.06020
作者: Sooyeun Yang,Jongseong Brad Choi
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:A 3D Gaussian Splatting (3DGS) reconstruction inherits arbitrary units from structure-from-motion, so it cannot be measured or used to place AR content until someone supplies a scale. Both standard recipes are fragile on real captures: on five indoor scenes with tape-measured ceilings, a quantile floor-to-ceiling heuristic errs by 27% and metric-depth alignment by 35-71% across three networks. We present METRIC-GS, a selective scale estimator for real indoor photographs: it reads the scale off the reconstruction’s own semantic content and, just as importantly, refuses when the evidence is thin. A frozen 2D teacher labels Gaussians by visibility-aware voting; instance-level size priors for architectural and furniture classes (doors, tables, the floor-to-ceiling gap) are fused by a robust maximum-likelihood step; and an inlier-mass test turns thin evidence into an explicit abstention that doubles as a router to a heavyweight metric model. On 47 ScanNet++ scenes with laser ground truth, with a configuration frozen before public evaluation, the estimator alone cuts the heuristic’s mean error from 21.1% to 16.6% (9.3% on the 28 scenes it acts on) at ~11 s per scene. MASt3R remains the more accurate standalone estimator (9.9%), but routing only our refusals to it matches running it everywhere with 60% fewer invocations; this selective hybrid is the configuration we advocate. Visibility-aware voting raises the scenes the estimator can act on from 16 to 28 at unchanged deployed accuracy. Applied unchanged to 13 phone and drone captures of our own, on which MASt3R fails its own registration gate 12 times, the system improves all five captures it acts on and degrades none. We also report that a current feed-forward metric model scores 2.3% on a benchmark inside its training mixture and 28-35% on captures it has never seen: a published metric-scale number is a best case, not a deployment estimate.
[CV-82] ChanSFormer: A Channel Agnostic Vision Transformer for Multi-Channel Cell Painting Images NEURIPS2026
链接: https://arxiv.org/abs/2611.06015
作者: Jingwei Zhang,Srinivasan Sivanandan,Dimitris Samaras
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at NeurIPS 2026
Abstract:High-content multi-channel imaging, from Cell Painting assays to remote sensing, is central to modern scientific pipelines. Because these channels are highly heterogeneous and experimental configurations frequently evolve, a practical vision backbone must be fundamentally channel-agnostic. Current channel-adaptive vision transformers rely on global self-attention with learnable channel embeddings, which can be computed only for previously known channels. Attempts to overcome this limitation of new channels either discard the embeddings, losing channel identity, or process each channel independently, thereby eliminating critical cross-channel interactions. To address this, we introduce ChanSFormer, an efficient Vision Transformer that replaces rigid embeddings with disentangled spatial-channel attention and individual channel CLS tokens. This architecture natively preserves both distinct channel identities and cross-channel information flow. Furthermore, the individual channel class token design unlocks more informative feature representations for self-supervised learning and enables representation-based channel sampling while remaining channel-agnostic. We evaluate ChanSFormer on two biology multi-channel datasets, CHAMMI, JUMP-CP and a satellite dataset, So2Sat. Experimental results show that ChanSFormer outperforms state-of-the-art methods by 3.93%-4.27% on JUMP-CP in classification accuracy. It also demonstrates exceptional cross-dataset transferability under a self-supervised setting, outperforming previous methods by up to 2.72% in accuracy in the same classification tasks. Furthermore, the disentangled attention reduces the quadratic complexity of channel-by-spatial sequence length of the global attention baseline, thus improving throughput by 55%-313%. Code is available at this https URL.
[CV-83] MotoVerse: A Benchmark for Predictive Perception of Motorized Two-Wheelers in Dense Unstructured Traffic
链接: https://arxiv.org/abs/2611.06007
作者: Varun A. Paturkar,Shankar Gangisetty,C.V. Jawahar
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Motorized Two-Wheelers (MTWs) dominate personal transport across much of the Global South, yet remain largely overlooked in predictive perception research for autonomous driving and ADAS. We introduce MotoVerse, a benchmark for predictive perception from an MTW’s ego-centric perspective. MotoVerse comprises 949 maneuver clips collected in dense, unstructured traffic and trajectory annotations for over 31,000 surrounding road agents, including over 16,000 MTWs, enabling two complementary tasks: Ego-Rider Maneuver Anticipation (ERMA) and Traffic Agent Trajectory Prediction (TATP). Each clip is further annotated with an MTW-specific challenge taxonomy comprising dynamic ego-motion, weak traffic structure and contextual occlusion, enabling challenge-aware evaluation. We benchmark established conventional prediction models, general-purpose Vision-Language Models (VLMs), driving-specific VLMs and vision-language-action models. Our experiments reveal three key insights: (i) video backbones reach most of their full-observation accuracy from the first fifth of a maneuver, whereas zero-shot VLMs need most of the maneuver to be observed; (ii) zero-shot VLMs, whether general-purpose or driving-specific, largely remain close to simple baselines on both tasks; a preliminary LoRA study shows that adaptation lifts anticipation sharply but leaves trajectory prediction behind conventional models; and (iii) the rider’s own motion drives forecasting difficulty: for every model, including a static baseline, errors are highest during left turns and rise under dynamic ego-motion. We believe MotoVerse will serve as a challenging benchmark for safer and more robust predictive AI for intelligent two-wheelers. The dataset and evaluation code are publicly available at this https URL.
[CV-84] Intra-Architecture Source Attribution for Stable Diffusion: A New Task with Benchmark Dataset and Baseline WACV2027
链接: https://arxiv.org/abs/2611.06000
作者: Ashutosh Dayal,Mallika Garg,Debashis Sen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted at IEEE/CVF Winter Conference on Applications of Computer Vision(WACV 2027)
Abstract:Designing systems that attribute a generated image to its source model has lately become a major area of AI research. Most existing datasets and systems related to image source attribution are targeted towards the detection of architectural signatures of the source models. Distinct from this, focusing on stable diffusion, we introduce a novel task, IntrAA-SD, Intra-Architecture source Attribution for Stable Diffusion, where the source models to be differentiated share the same architectural backbone and differ only due to fine-tuning. The proposed task aligns with the emergence of fine-tuned generators having the same architecture, like Juggernaut XL and RealVis XL, which are fine-tuned versions of Stable Diffusion XL (SDXL). We provide a first-of-its-kind 140 K-image benchmark dataset containing images generated from different fine-tunes of SDXL, along with in-domain and out-of-distribution prompts. We also provide a baseline framework for IntrAA-SD, which leverages a style-conditioned Mixture-of-Experts network. It uses a frozen CLIP backbone to capture the image style and gate the MoE, which processes fine-tuned trace features from a high-resolution LoRA-adapted DINOv2 network. The fine-tuned models used, the prompting strategies considered, and the formation details employed in the novel dataset are elaborated. Our baseline intra-architecture source attribution method is evaluated in terms of attribution accuracy on our dataset, and ablation studies are also performed. Code and dataset is available at this https URL.
[CV-85] FedIA: Instance-Aware Cross-Modal Representations for Federated Prompt Learning
链接: https://arxiv.org/abs/2611.05994
作者: Xuanjing Zhang,Qi Guo,Yanan Qiao,Yong Han,Junge Bo
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注:
Abstract:Federated Prompt Learning (FPL) enables parameter-efficient adaptation of pre-trained Vision-Language Models across decentralized clients without exchanging raw data. To cope with cross-client heterogeneity, existing FPL methods typically pair a globally shared prompt with a client-specific local prompt. However, this design personalizes only at the client level: the same local prompt is applied to all samples of a client, and the global and local prompts are combined in an input-agnostic, fixed manner. Consequently, it overlooks the substantial visual variations among samples within the same client and cannot adjust the balance between shared and personalized knowledge to each input. To address these limitations, we propose Federated Prompt Learning with Instance-Aware Cross-Modal Representations (FedIA), an FPL framework that learns instance-aware cross-modal representations. Specifically, FedIA first uses global text prompts as client-invariant semantic anchors to select image patches most relevant to each candidate class, suppressing distracting regions. The selected visual evidence then conditions the generation of instance-aware local text representations. Finally, an adaptive global-local fusion module calibrates the contributions of local and global text representations for each image-class pair. In this way, FedIA shifts personalization in FPL from the client level to the instance level while preserving the semantic knowledge shared across clients. Extensive experiments under domain shift, pathological label skew, and Dirichlet-based label skew show that FedIA outperforms FPL baselines, with maximum accuracy gains of 1.44%-4.37% over the latest baseline across the three settings. Ablation studies and analyses further verify the complementary contributions of instance-aware representation generation and adaptive global-local fusion.
[CV-86] Clips2Story: Training-Free Video Storyboarding and Editing Using Multimodal Retrieval-Embedded Generation
链接: https://arxiv.org/abs/2611.05992
作者: Junnuo Wang,Weihan Xu,Hao-Wen Dong
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multimedia (cs.MM)
备注:
Abstract:Video editors often work with large collections of fragmented footage and must select and organize relevant clips into a coherent story. However, many existing video editing and summarization methods assume a temporally ordered source video and aim to preserve its original narrative and temporal structure. In this paper, we introduce Clips2Story, a training-free, LLM-based pipeline that enables users to create keyword-specified short videos from a pool of video clips. Given a pool of clips and a user-specified keyword, we first extract text-based multimodal descriptors for each clip and retrieve the clips most relevant to the keyword. We then provide the textual descriptors of these candidate clips to an LLM, which selects and arranges them into an editable storyboard with connective narrations. Finally, we assemble the selected clips and retrieve accompanying visuals for the narrations to compose the output video. To evaluate the proposed pipeline, we construct a synthetic evaluation dataset from 100 long-form videos across five genres by segmenting each video into short clips using a scene-change detection model and shuffling the resulting clips to remove their temporal order. Through objective evaluation and qualitative examples, we demonstrate that Clips2Story can effectively retrieve keyword-relevant clips and compose coherent storyboards based on user-specified keywords. Human evaluation further shows that Clips2Story outperforms a strong baseline in terms of narrative coherence, keyword alignment, and overall realism.
[CV-87] PathoS2VD: From 2D Histopathology to Pseudo Z-Stack Representations ACCV2026
链接: https://arxiv.org/abs/2611.05974
作者: Jiamu Wang,Keunho Byeon,Ji-Uk Kim,Jin Tae Kwak,Heounjeong Go
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Pre-review version. Accepted at ACCV 2026
Abstract:For decades, diagnostic pathology has relied on thin two-dimensional (2D) tissue sections, despite biological tissues exhibiting inherently three-dimensional (3D) morphological organization. Because a single section captures only one focal plane, important morphological cues along the z-axis are inevitably discarded, limiting tissue representation for both human interpretation and computational analysis. In this work, we propose a generative framework that learns z-axis morphological priors from pathology z-stack data, offering a cost-effective alternative to dedicated volumetric imaging, and transfers this knowledge to conventional single-plane histopathology. Specifically, an image-to-video (I2V) latent diffusion model (LDM) is trained to synthesize a pseudo z-stack from a single 2D image, capturing plausible morphological variations across adjacent focal planes rather than reconstructing physical tissue volumes. To enable deployment on large-scale clinical datasets without native z-stack acquisition, we further introduce a prior-preserving mixture fine-tuning strategy that adapts the model to new domains while retaining the learned z-axis morphological prior. Extensive experiments demonstrate that the synthesized pseudo z-stacks exhibit realistic cross-plane morphological continuity and consistently improve downstream prostate cancer classification compared with conventional 2D representations. Our results suggest that learning transferable z-axis priors from limited z-stack data provides a practical way to enrich routine pathology images with complementary morphological information. Code: this https URL.
[CV-88] CLEAR: Color-Texture Latent Alignment for Single Image Reflection Removal WACV2027
链接: https://arxiv.org/abs/2611.05969
作者: Tu Vo,Sheir A. Zaheer,Chan Y. Park
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: Accepted to WACV 2027
Abstract:Single Image Reflection Removal (SIRR) seeks to recover the transmitted scene from a photograph captured through glass. Recent latent restoration methods use pretrained generative priors, but many recover detail by feeding high-frequency features from the degraded input back into the decoder. This works for degradations such as blur, noise, or haze, where input detail mostly belongs to the scene. It is less suitable for SIRR, where reflections add a second structured layer: directly reusing the input’s skip connections can sharpen the image while also bringing reflection edges back. Latent compression also introduces color drift and fine-detail loss. We propose CLEAR (Color-Texture LatEnt Alignment for Single Image Reflection Removal), a deterministic, single-timestep, non-iterative latent framework that uses a pretrained latent diffusion backbone as a natural-image prior. A physics-guided separator organizes the prediction around the transmission, reflection, and residue formation model and outputs coarse layer latents in one forward pass. The Hybrid Feature Alignment Module (HFAM) then treats input-derived skips as contaminated layer-mixture features. It rectifies their high-frequency content with a clean-transmission-aligned color query, learned from the input features themselves under ground-truth feature supervision, keeping transmission-consistent texture while attenuating reflection-colored detail. CLEAR achieves the best average across fidelity and perceptual metrics on six real-world benchmarks and the lowest LPIPS on every dataset, with the largest gains on the hardest benchmarks. Comments: Accepted to WACV 2027 Subjects: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR) Cite as: arXiv:2611.05969 [cs.CV] (or arXiv:2611.05969v1 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2611.05969 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[CV-89] FLINT: Coupling Proximal Initialization and Bounded Stochasticity for Flow-Matching Inverse Problems NEURIPS2026
链接: https://arxiv.org/abs/2611.05945
作者: Junseo Bang,Daewon Choi,Dong Ju Mun,Se Young Chun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: NeurIPS 2026, Project Page: this https URL
Abstract:Flow matching models have emerged as a dominant paradigm for generative posterior sampling in imaging inverse problems. Existing solvers follow two distinct lines of work: trajectory-space methods, which refine sampling dynamics from random initial noise, and source-space methods, which optimize the initial noise itself. While most prior work treats these as independent design choices, we show that they are inherently coupled. We derive a unified stability bound that links initialization proximity and cumulative sampling stochasticity through a multiplicative envelope, complemented by two empirical observations. First, we identify a sufficiency regime in which a moderately optimized proximal source initialization is enough to attain the performance plateau, rendering exhaustive source-space optimization unnecessary. Second, this initial advantage is fragile: it is eroded by excessive stochasticity accumulation during guided sampling, necessitating a mechanism that bounds the cumulative noise injection. These findings motivate FLINT, a two-stage framework that operationalizes this coupling. FLINT establishes a high-fidelity starting point via a manifold-constrained proximal objective in only a few iterations, then performs guided sampling under a bounded stochasticity schedule that helps prevent erasure of the initialization advantage. Across diverse inverse problems, FLINT delivers state of the art perceptual quality and competitive distortion metrics while maintaining superior computational efficiency.
[CV-90] 3D Banff Glomerulitis Scoring for Kidney Transplant Pathology: Feasibility Study for Volumetric Quantification
链接: https://arxiv.org/abs/2611.05925
作者: Yanfan Zhu,Junchao Zhu,Ruining Deng,Shilin Zhao,Yu Wang,Yaohong Wang,Yuechen Yang,Chongyu Qu,Zhengyi Lu,Juming Xiong,Mengmeng Yin,Yihan Wang,Andrew J. Rauch,Haichun Yang,Yuankai Huo
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Quantifying glomerular inflammation in kidney-transplant biopsies is traditionally performed on single whole-slide sections using the Banff “most-severe section” rule, despite the inherently three-dimensional nature of lesion distribution. This 2D based assessment might lead to instability when inflammation varies across slices. This paper reports a feasibility study examining the technical feasibility and clinical relevance of extending Banff scoring into three dimensions. To this end, we propose a 3D Banff lesion-scoring framework that aligns serial sections, reconstructs glomerular identities across depth, and integrates inflammatory-cell counts in 3D under three scoring strategies: Compatible, Volumetric, and Deduplicated Volumetric. In the experiments, glomerulitis (g-scores) was used as an example Banff metric and applied to multi-section renal allograft biopsies. Our findings indicate that 3D Banff g-scores are more consistent across sections and, under the semi-automatic setting, the Volumetric strategy correlates more strongly with clinical biomarkers than traditional 2D scores. These results show that 3D volumetric quantification offers promising added value, underscoring the potential benefit of 3D-aware Banff scoring for kidney transplant pathology.
[CV-91] DecoupAttn: Decoupled Sparse Attention for Training-Free Acceleration of Video Diffusion
链接: https://arxiv.org/abs/2611.05917
作者: Yanjin Li,Bingbing Qiu,Qiang Huang,Rui Qian,Yanfeng Li,Jian Wang,Xiaoying Tang,Mingda Zhang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 35 pages, 16 figures, 35 tables
Abstract:Sparse attention accelerates few-step video diffusion transformers by computing a few key blocks exactly and compensating for the rest. However, existing sparse attention still faces key challenges, including identification overhead from finding critical blocks at every denoising step, suboptimal budget allocation that sets sparsity budgets uniformly or by local metrics, and cross-step error accumulation as reused masks or clusters go stale. Observing that the errors left on skipped blocks cancel each other once every key is compensated, we propose DecoupAttn, a new paradigm of decoupled sparse attention: semantic clusters compensate every key, and exact attention corrects their error on a few blocks fixed once offline by position. To support this paradigm, we introduce Distortion Budget, a rate-distortion objective for budget allocation that places exact computation across timesteps, layers, and heads where it removes the most distortion that compensation leaves in the final latent. Built on this budget, we further propose Reassign Cache, in which each step reassigns the drifting keys to cached centroids and refreshes compensation while the budgeted blocks stay fixed. Moreover, DecoupAttn is training-free: a few calibration clips compile one plan offline, and inference runs standard block-sparse kernels with one fused compensation kernel. Our code will be released after submission.
[CV-92] Do Saliency Maps Fail Dense Breasts? A Chance-Normalized Confounder-Aware Audit of Explanation Fairness in Mammography
链接: https://arxiv.org/abs/2611.05909
作者: Ahmed Jaheen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Saliency maps, offered as evidence that mammography classifiers use the cancer, are untested in dense breasts, which can hide cancers. We tested whether they locate cancers worse there, and why. On 1,345 malignant CBIS-DDSM views, we scored three CNNs’ Grad-CAM, Integrated Gradients and occlusion maps by whether their peak hits the lesion and a chance-normalized lesion-to-background saliency ratio (bits; +1 doubles it). Unadjusted gaps were large: Grad-CAM peaks hit 44% of dense vs. 58% of non-dense lesions. Case mix, chiefly more calcifications in dense breasts, explained about half or more of each gap. Occlusion kept a significant average of -0.56 bit gap. Controls also implicate the classifiers: unadjusted gaps exceeded an image-only conspicuity map’s, and models relied less on dense-breast lesions. We recommend reporting density gaps before and after case-mix adjustment, per classifier, with conspicuity and lesion-reliance controls.
[CV-93] Chronos: Sequential Identity-Preserving Video Generation by Temporal Anchoring ACM-MM2026
链接: https://arxiv.org/abs/2611.05893
作者: Mengqi Huang,Wenchuan Wang,Zhendong Mao,Yijing Tu,Zhike Liu,Luoni Wang,Ning Tong,Xiaoxiao Sheng,Tianyu Huang,Xuecheng Qi,Yongdong Zhang
类目: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
备注: Accepted by ACM MM 2026. 1st place in Tracks 1 and 2 of the ACM MM 2026 Identity-Preserving Video Generation Challenge
Abstract:Identity-preserving video generation aims to synthesize high-quality and visually realistic videos from textual and visual conditions while faithfully maintaining the subject’s identity across diverse motions. Prior methods have extensively introduced reference-guided control signals and identity-aware conditioning to improve facial alignment from reference images. However, improving static identity alignment alone creates a fundamental tension between identity fidelity and temporal controllability, which frequently causes identity drift or motion suppression under large pose variations and multi-segment sequential actions. To address this, we propose Chronos, a framework that unifies single-segment facial identity preservation and multi-segment sequential action generation under a common temporal formulation. Specifically, Chronos integrates Semantic Prompt Anchoring to adaptively refine the user prompt, Visual Frame Anchoring to transform conditions into visually grounded keyframe anchors for visual guidance, and Metric-Guided Candidate Routing to select the highest-quality videos from heterogeneous generators via multi-dimensional evaluation. Extensive evaluations demonstrate the superiority of our approach, which achieved first place in both Track 1 and Track 2 of the ACM MM 2026 Identity-Preserving Video Generation Challenge, demonstrating its effectiveness across both facial identity preservation and sequential action generation settings.
[CV-94] Feature Puzzles: Single-Image Supervision for Multi-View 3D-Aware Segmentation
链接: https://arxiv.org/abs/2611.05874
作者: Jiahao Ma,Qiang Zhang,Miaomiao Liu,David Ahmedt-Aristizabal,Chuong Nguyen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Feed Forward 3D Reconstruction and Spatial Understanding, Multi-view Promptable Segmentation, Instance Tracking
Abstract:Humans can easily recognize the same objects across different viewpoints of a 3D scene, whereas teaching models such multi-view-consistent perception typically requires videos or calibrated multi-view captures, which are considerably more expensive and less diverse than large-scale monocular data.~In this paper, we introduce Feature Puzzles, a novel framework for learning multi-view-consistent segmentation using only single-image semantic segmentation supervision. Specifically, we propose a simple yet effective pipeline that constructs multi-view-consistent segmentation pseudo-labels through image cropping and geometric transformations. We further introduce a lightweight, plug-and-play network that leverages the multi-view-consistent image features learned by existing pretrained 3D foundation models and the generated pseudo-labels to perform multi-view-consistent segmentation. Importantly, our flexible framework supports various types of prompts provided from one or multiple views and can be readily extended to open-vocabulary 3D segmentation. Under this single-image supervision, our feed-forward model achieves the best mIoU on NVOS, remains competitive on SPIn-NeRF and ScanNet++ against baselines using per-scene optimization or in-domain multi-view training, and demonstrates zero-shot transfer to open-vocabulary 3D segmentation. Visual comparisons are available on our project page: this https URL
[CV-95] Generative AI for urban planning : synthesizing land-use maps with satellite imagery via DualControlNet
链接: https://arxiv.org/abs/2611.05859
作者: Qingqi Song,Yuzhou Chen,Lingqian Hu,Shenhao Wang
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Various land-use models characterize land-use patterns and facilitate scenario planning. However, existing models are largely predictive, forecasting transitions based on numeric data without synthesizing new land-use maps from simple text prompts. Meanwhile, generative AI in urban systems remains unimodal, weakly controllable, and evaluated mostly by generic image quality rather than domain-specific metrics. To address these gaps, we introduce DualControlNet, a generative land-use model that jointly synthesizes high-resolution land-use maps and satellite imagery, addressing multimodality, controllability, and verification challenges. Trained on over 40,000 paired tiles from Orlando and Chicago, DualControlNet is evaluated using land-use, spatial consistency, and image quality metrics. Relative to unimodal baselines, DualControlNet improves land-use diversity consistency by 62.8% and building coverage ratio consistency by 16.3%, more than doubles cross-modal spatial consistency, and triples visual quality. Scenario-based experiments demonstrate that DualControlNet supports sustainable development through multidimensional controls, increasing density and diversity, and responding to text prompts like “high-density development.” Overall, this work advances land-use modeling via a multimodal, controllable, and verifiable generative framework, lowering technical barriers and enabling intuitive exploration of urban futures.
[CV-96] Risk Control Certifies Rates Not Magnitudes: An Audit of Selective Fetal Biometry
链接: https://arxiv.org/abs/2611.05851
作者: Zhengshuai Yang,Linna Jiang,Changyun Li,Guihu Zhao
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 2 figures, 3 tables
Abstract:Prenatal ultrasound biometry must return a head circumference (HC) and refuse when the image cannot support one. Selective prediction trades coverage against error, but its finite-sample guarantees bound the rate of gross errors, not their magnitude. We audit a tool-grounded biometry pipeline on two separately pre-registered four-way splits of the 999 labelled HC18 images, retraining the segmenter three times on each. Calibrating the contour-completeness threshold by exact Clopper-Pearson risk control returns a certificate indistinguishable from a hand-set floor, yet the pipeline still emits a 271.3 mm error: a small, geometrically complete ring fitted to an image whose true HC is 324.6 mm. The accepted candidate scores 0.94 completeness, so the score cannot see the failure, and one outlier among \approx 150 answers leaves the certified risk bound ( \alpha = 0.10 ) intact. A second, untouched split reproduces the mechanism on a different image (233.3 mm). Constraining the admissibility region—the anatomical range in which the measurand is defined—removes the family instead: raising the lower bound from 50 to 55 mm costs 0.4 coverage points and cuts the worst-case error on both splits to 11.4 and 11.1 mm, whereas suppressing it through the completeness threshold costs 19.0 and 12.1 points. At an anatomically defined 70 mm the frozen tests give 1.91 and 1.81 mm MAE.
[CV-97] me-to-Goal: Learned Temporal Metrics for Long-Horizon Planning with JEPA World Models
链接: https://arxiv.org/abs/2611.05847
作者: Ruilong Yu,Fei Ye
类目: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
备注: 10 pages, 6 figures
Abstract:Joint-Embedding Predictive Architectures (JEPAs) support planning by comparing predicted latent states with an encoded goal. For this comparison to guide temporal progress, states closer in time to the goal should receive lower costs. However, training representations for prediction and distribution regularization does not directly enforce this ordering. We examine this mismatch in flat LeWorldModel (LeWM) by analyzing how its Euclidean goal cost changes with temporal separation. Our analysis identifies conditions under which states at different long offsets receive similar average costs, weakening their temporal distinction. To learn the missing ordering explicitly, we introduce Time-to-Goal (T2G), a lightweight metric trained on latent triples from offline trajectories. A margin ranking loss encourages smaller distances for temporally nearer states, requiring no rewards or policy labels. The learned metric then replaces the Euclidean planning cost while the JEPA encoder and predictor remain frozen. Diagnostics on AntMaze and FourRoom show that T2G preserves stronger temporal contrast and improves near–far ranking at long offsets. Across four navigation and manipulation benchmarks, T2G improves the average planning success rate from 32.7% (flat LeWM) to 63.0% and approaches or even surpasses the performance of the hierarchical H-JEPA baselines. These results show that learning temporal ordering for the goal cost can substantially improve latent planning even with a flat world model.
[CV-98] DualCast: Dual-Image Conditioned Diffusion Transformers for Digital Human E-Commerce Video Generation
链接: https://arxiv.org/abs/2611.05825
作者: Lifan Jiang,Yibo Zhao,Senlin Cheng,Bolei Xu,Xiaobo Li,Can Ye,Dehai Zhao,Deng Cai,Boxi Wu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Digital human e-commerce video generation requires preserving product appearance, human identity, and overall video style during dynamic demonstrations. Although recent models support multiple reference images, their general conditioning schemes do not explicitly distinguish the asymmetric roles of “what is presented” and “who presents it.” Consequently, generated videos may favor a single reference, resulting in product distortion, identity drift, or single-reference collapse. To address this issue, we propose DualCast, a role-addressed dual-image conditioning framework for video generation. By explicitly modeling the distinct roles of product and human conditions, DualCast reduces role ambiguity and feature competition between references. First, a dual-spatiotemporal slot injection mechanism VAE-encodes the product and human images separately and places their latent features at the front and rear boundaries of the DiT sequence, enabling sequence-level condition separation and directed injection. Second, a mask-guided region-weighted loss strengthens supervision over product, face, and product-face overlap regions, improving generation fidelity in critical areas. Third, image condition dropout enables a single model to support dual-image, product-only, and human-only generation. In addition, we construct a curated dataset of 10K e-commerce videos for model training. Experimental results demonstrate the strong performance of DualCast.
[CV-99] Navigational Safety of Marine Surface Vehicles in Restricted Visibility: From Intelligent Vision to Vision-Centered Multi-Sensor Fusion
链接: https://arxiv.org/abs/2611.05818
作者: Ryan Wen Liu,Fengchuan Song,Tianxiang Wang,Guobei Peng,Youngrong Kim,Xian Zhong,Hooi-Siang Kang,Maohan Liang,Junzhi Yu,Shanliang Yao
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 42 pages, 16 figures
Abstract:Marine surface vehicles (MSVs) have attracted considerable attention for their commercial, scientific, and military applications. Owing to their favorable performance-to-cost ratio, shipboard cameras are widely used to detect and recognize potentially hazardous situations across diverse operational scenarios. However, visual perception is inevitably degraded by restricted conditions, including irregular motion of the imaging platform, adverse weather, and long imaging distances. To improve situational awareness and navigation capabilities, complementary sensors, such as the automatic identification system (AIS), radar, and light detection and ranging (LiDAR), etc., have become essential shipboard equipment. Multi-sensor fusion can provide richer perceptual information and help ensure navigational safety of MSVs in restricted visibility. This article reviews recent progress from intelligent vision to vision-centered multi-sensor fusion for newcomers and researchers in intelligent ship navigation and intelligent waterborne transportation system (IWTS). After introducing the generation of quality-degraded images in restricted visibility, we will review intelligent vision methods for four key issues, i.e., shipboard video stabilization, visual perception enhancement, object detection and tracking, and navigable area segmentation. We then examine vision-centered multi-sensor fusion methods, including infrared-visible image fusion, radar-vision fusion, LiDAR-vision fusion, AIS-vision fusion, and other multi-sensor fusion scenarios. Open-source datasets and safety-critical applications are also summarized and discussed. Finally, we highlight open challenges and future research directions. The collected maritime datasets are available at this https URL and will be regularly updated.
[CV-100] ReTailor: A Residual Feature Tailoring Framework for Real-Time Cross-Sensor Pansharpening Adaptation
链接: https://arxiv.org/abs/2611.05811
作者: Jing-Ze Liang,Wei-Jie Xiao,Zi-Han Cao,Yang-Dong Li,Tian-Yi Hu,Liang-Jian Deng
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 18 pages, 12 figures, including appendices. The first three authors contributed equally
Abstract:Recent pansharpening models achieve strong fusion performance on trained sensors, yet often generalize poorly to unseen sensors. Existing cross-sensor adaptation is limited by target-sensor retraining, which requires target-sensor training pairs and a dedicated training stage, and by per-image zero-shot adaptation, which is computationally expensive due to whole-network optimization for each test image. To address these problems, we propose ReTailor, a real-time sequential cross-sensor adaptation framework. Only a lightweight residual Feature Tailor is rapidly optimized under self-supervised, physics-aware constraints to correct cross-sensor shifts without modifying the well-learned fusion prior. For efficient and effective deployment, ReTailor combines content-aware patch selection, which identifies a few informative patches per image for optimization at negligible cost to improve both quality and efficiency, with rolling residual bootstrap, which inherits sensor-related knowledge across the image stream to further improve quality. Extensive experiments on cross-sensor adaptation settings show that ReTailor mitigates cross-sensor degradation both effectively and efficiently. Moreover, we compare ReTailor with Low-Rank Adaptation (LoRA), a widely used method for fine-tuning, to demonstrate the advantages of our approach. We further evaluate the core principle of ReTailor on two additional remote sensing tasks: super-resolution pansharpening and image segmentation, where it consistently improves performance. In general, ReTailor provides a highly effective and efficient strategy for deploying pansharpening models under real-world cross-sensor degradation, while also offering useful insights into managing cross-sensor degradation in other remote sensing tasks. Code will be released at this https URL .
[CV-101] GREP: Grounding Generated Room Evaluation in Principles
链接: https://arxiv.org/abs/2611.05802
作者: Mohan Cao,Yongsen Mao,Chucheng Xiang,Zhongyuan Liu,Zhihang Zhong,Haiping Wang,Yuan Liu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR)
备注: 24 pages. Project page: this https URL
Abstract:Recent LLM-centric advances can generate attractive furnished rooms by selecting and arranging assets. Yet evaluation remains limited to specific perspectives or predefined questions, failing to reflect overall scene quality. To address this pressing need, we introduce GREP. Its core is a focus shift from defining specific questions to establishing a three-tier evaluation philosophy: basic physical validity checks whether a scene obeys physical constraints; visual coherence concerns compatibility in object scale, style, and composition; and, at the highest level, human-use logic requires object arrangements to align with everyday human habits. Guided by these principles, an evaluation agent autonomously and adaptively identifies issues from scene context, gathers evidence, and produces scores and reports. This supports scene-specific issue discovery and reliable judgments. Experiments show that (1) GREP achieves 87% agreement with human majority preferences, substantially outperforming baseline evaluators. (2) GREP reports improve scene quality through refinement, whereas score-only and no-feedback baselines show slight declines. (3) GREP suggests that as planning ability improves, recent models such as GPT-5.6 Sol can move from relying on pipelines and harnesses to challenging their guidance and constraints, but still lack the ability to resolve the resulting conflicts. Stronger models are therefore expected not only to plan well, but also to adapt to external guidance and constraints more flexibly. Project page: this https URL
[CV-102] WayPOP: A Panoramic Open-Set Panoptic Tracking Benchmark
链接: https://arxiv.org/abs/2611.05801
作者: Rohit Mohan,Maximilian Luz,Swarnava Chowdhury,Yao Lu,Florian Drews,Thomas Nürnberg,Yakov Miron,Federico Tombari,Stefano Gasperini,Abhinav Valada
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
备注:
Abstract:Holistic perception for autonomous agents requires reliably segmenting and associating scene elements across time and camera views, including objects outside a fixed semantic taxonomy. Existing open-set panoptic benchmarks focus on individual images or views, while panoramic tracking benchmarks assume closed-set taxonomies. Consequently, these benchmarks do not evaluate whether unknown objects can be consistently segmented and tracked across synchronized multi-camera sequences. We introduce WayPOP, a panoramic open-set panoptic tracking benchmark built on real Waymo PVPS sequences augmented with diverse out-of-taxonomy 3D assets from Objaverse. Inserted objects are placed at fixed locations in the 3D scene and rendered using calibrated vehicle-mounted cameras as the ego vehicle and surrounding traffic move, providing temporally and cross-view consistent panoptic annotations and auxiliary LiDAR range observations. WayPOP contains approximately 670 curated assets instantiated as 980 unique unknown-object tracks, yielding 53,452 unknown-object mask annotations across 21,300 camera images. Our evaluation protocol defines a hierarchy of tasks from view-independent open-set panoptic segmentation to full panoramic open-set panoptic tracking, using known/unknown PQ and (w)STQ metrics. We benchmark representative open-set and open-vocabulary segmentation methods with temporal and cross-camera association. Our results reveal substantial limitations of current methods in truly open-set scenarios, highlighting robust unknown-object localization and identity association across time and views as major open challenges.
[CV-103] ME2FDiff: Multiscale Expert Enhancing Feature Diffusion in Residual Network for Brain Tumor Segmentation
链接: https://arxiv.org/abs/2611.05766
作者: Afrida Rahman,Bijoy Kundu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Automated segmentation of post-treatment glioma from MRI is essential for accurate diagnosis and treatment planning, yet remains difficult. Surgery and chemoradiotherapy cause structural changes in brain like resection cavities, edema, and fibrosis, that alter imaging appearance and mimic tumor recurrence. Contemporary segmentation architectures attain high accuracy with two limitations. First, the bottleneck originates from the deepest encoder stage, so one representation must serve sub-regions that differ widely in size and appearance. Second, they produce the segmentation in a single forward pass with no mechanism for refinement. We propose ME^2FDiff, a single-model architecture, addressing both. An attention-conditioned spectral diffusion block refines the bottleneck. It uses the image-conditioned latent diffusion process with two processes: a spatial pre-attention gate and a spectral fusion block whose timestep-adaptive bandpass filter selects the informative frequency band at each noise level. A multiscale expert-enhanced feature block fuses the intermediate encoder scales into the bottleneck through expert-selected shared tokens, guaranteeing balanced utilization without an auxiliary loss. On the BraTS 2024 post-treatment glioma dataset, ME^2FDiff achieves the best results on every tumor sub-region among all single-model baselines (91.23% Dice, 84.52% IOU, 2.13 mm HD95 on whole-tumor) and surpasses a three-expert ensemble on both overlap metrics at roughly one-third of its inference cost. Against recent state-of-the-art methods retrained under identical conditions, it maintains the strongest performance with the fewest parameters of any competing model. These findings demonstrate that a single, compact network can absorb the specialization, usually sought through ensembling or larger architectures.
[CV-104] Decompose the Distillation: Interpretable Single-Pass Guidance for Diffusion Models
链接: https://arxiv.org/abs/2611.05755
作者: Wuyang Zhang,Xiaolin Xu
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
备注: The Fortieth Annual Conference on Neural Information Processing Systems
Abstract:Classifier-Free Guidance is a widely used inference-time guidance rule for conditional image generation with diffusion models, which exposes a guidance scale that adjusts conditioning strength after training. This flexibility requires conditional and unconditional model evaluations at every sampling step, so an N-step trajectory takes 2N forward passes. Single-pass alternatives reduce such cost, yet most collapse CFG’s two components into one opaque guided prediction, losing the separable base flow and guidance direction that support per-step control and diagnostics, even when the scale remains adjustable. This work introduces CtrlFlow, a single-pass guidance method that preserves CFG’s separable base-flow and guidance-direction interface. To preserve both components at one-pass cost, CtrlFlow attaches two lightweight LoRA adapters to a frozen backbone, each adding 2.5% parameters, and predicts the base flow and guidance direction from one pass of the shared frozen backbone. Because the guidance scale multiplies the guidance direction only after the forward pass, one evaluation yields the guided velocity at any scale, so a single trained checkpoint remains adjustable at inference without retraining. Preserving two outputs introduces a training mismatch: inference amplifies guidance-direction errors by the chosen scale, while naive component-wise losses weight both outputs equally. To align training with this composed inference-time error, CtrlFlow uses a unified multi-scale distillation loss. Residual guidance errors can still persist at extreme scales, where even small guidance-direction errors are amplified by the guidance scale. To reduce this inference-time error, CtrlFlow applies scale-adaptive modulation to damp velocity-field errors and time-grid warping to concentrate solver evaluations where the trajectory changes most rapidly.
[CV-105] Learning Hierarchical Patch Splitting Policies for Faster Vision Transformers NEURIPS2026
链接: https://arxiv.org/abs/2611.05742
作者: Aditya Gupta,Jean Sebastien Dandurand,Kai Qiu,Rohan Choudhury,Laszlo A. Jeni
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Accepted at NeurIPS 2026
Abstract:Vision Transformers (ViTs) are bottlenecked by the length of their input token sequence. Standard patchification fixes this length with a single patch size applied across the image, regardless of regional content. Prior work on adaptive patch sizing reduces token count by assigning coarser patches to visually simple regions, but relies on noisy pixel-level heuristics such as local entropy to make patchification decisions. We argue that this compute allocation should be grounded in the model’s interpretation of task relevant information, not in raw pixel statistics. We introduce Split Policy Learned Image Tokenization which learns to allocate token resolution based on what the model finds meaningful, producing shorter token sequences that are better matched to task-relevant content. SPLIT matches the accuracy of full-resolution ViT classification baselines while using 30% fewer tokens. The same gains hold on dense prediction: SPLIT achieves competitive accuracy on COCO and ADE20K, demonstrating that semantically aligned patchification generalises well beyond classification.
[CV-106] OdorSpaces: Visual-Olfactory Embodied Navigation in 3D Environments
链接: https://arxiv.org/abs/2611.05738
作者: Changyeon Lee,Kelin Yu,Zheng Wei,Shuna Ni,Ruohan Gao
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted at the 10th Conference on Robot Learning (CoRL 2026). 20 pages, 4 figures, 6 tables
Abstract:Olfaction provides critical information about airborne chemical cues, yet it remains largely underexplored in embodied AI, with limited prior machine olfaction work focusing primarily on semantic odor perception. We introduce OdorSpaces, a physics-based simulation platform and benchmark for visual-olfactory embodied navigation. OdorSpaces augments Habitat-Sim with 3D odor transport and dispersion simulated by Fire Dynamics Simulator (FDS), enabling agents to navigate photorealistic indoor environments using both visual and olfactory observations. The benchmark includes two representative source-seeking scenarios: invisible gas-leak localization and early-stage fire localization. We develop an initial multimodal reinforcement learning policy that integrates visual observations and olfactory measurements over time. Experiments show that the policy outperforms a diverse set of baselines, establishing a strong reference for future research on visual-olfactory embodied navigation.
[CV-107] C-WFDM: A ControlNet-Guided Wavelet-Fusion Diffusion Model for Multimodal Brain MRI Completion
链接: https://arxiv.org/abs/2611.05737
作者: Muhammad Nabi Yasinzai,Remika Mito,Mangor Pedersen
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Multimodal brain MRI provides complementary anatomical, tissue, and pathological information to support downstream artificial intelligence (AI) applications. However, modality availability is uneven across public and pooled neuroimaging datasets: T1-weighted (T1w) MRI is widely represented, T2-weighted (T2w) and fluid-attenuated inversion recovery (FLAIR) MRI are less consistently available, and gadolinium-enhanced T1-weighted (T1Gd) MRI is relatively scarce. Pooled datasets therefore contain participants with different combinations of acquired MRI contrasts, reflecting variation in the study designs, acquisition protocols, and clinical or research objectives of the contributing studies. Many missing-modality benchmarks simulate incomplete MRI by withholding contrasts from complete four-modality cases, which may not represent the observed availability patterns in pooled datasets. To address this gap, we introduce C-WFDM, a ControlNet extension of our Wavelet-Fusion Diffusion Model (WFDM), for subject-specific missing-modality completion. C-WFDM is built on an earlier BrainScape snapshot comprising 157 public MRI datasets and 45,880 preprocessed scans. The pretrained WFDM backbone retains target-modality and metadata conditioning, while target-specific ControlNet adapters use MRI contrasts from the same participant to guide synthesis. C-WFDM synthesizes T2w from T1w, FLAIR from T1w and T2w, and T1Gd from T1w, T2w, and FLAIR. Acquired conditioning contrasts are used when available, and generated intermediate contrasts are used only when required. We compare C-WFDM with ALDM, cWDM, and Pix2Pix using paired structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and mean absolute error (MAE). In the BrainScape completion experiment, C-WFDM achieves the best T2w and FLAIR results among the evaluated models, with SSIM values of 0.9232 and 0.8904, respectively.
[CV-108] Mitigating Spurious Invariance in Contrastive Learning: A Probabilistic Self-Supervised Learning Framework for Medical Image Analysis
链接: https://arxiv.org/abs/2611.05734
作者: Wei Wei,Yading Yuan
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Published in IEEE Transactions on Medical Imaging (TMI), 2026. 17 pages
Abstract:Owing to the prohibitive cost of manual annotation for enormous medical images, self-supervised learning (SSL) has gained substantial attention and shown promise in various medical imaging tasks. Among SSL approaches, contrastive learning has emerged as a prominent one, encouraging models to encode semantic information that remains invariant between different augmented views. However, this invariance can be spurious for medical images, as subtle but critical medical features can be suppressed during augmentations. In this paper, we mitigate this issue by encoding the learned feature embeddings as probability distributions rather than fixed vectors, thus capturing a richer and more nuanced spectrum of information. Furthermore, we propose a unified probabilistic framework that seamlessly integrates multiple SSL approaches. Our method achieves superior performance over the state-of-the-art SSL methods on various downstream tasks across multiple datasets, including medical image segmentation and classification. We further validate the method through additional analyses of robustness to stronger augmentations, label efficiency under limited fine-tuning labels, and computational trade-offs. The code is released at this https URL.
[CV-109] GeoDIVE: Geospatial Descent Informed by Value Estimation WACV2027
链接: https://arxiv.org/abs/2611.05697
作者: Hamza Iqbal,Dan Cher,Eric Xing,Pranavi Kolouju,Abby Stylianou,Robert Pless,Nathan Jacobs
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted to WACV 2027. Includes supplementary material
Abstract:Finding specific objects, structures, and spatial patterns supports critical work in agriculture, infrastructure planning, environmental monitoring, and disaster response. These features, however, are often dispersed across very large regions and span dramatically different spatial scales, which defines a search space that is costly to embed, maintain, and search naively. We introduce GeoDIVE, a model that treats a multiresolution satellite image gallery as a forest and uses best-first traversal to determine which tiles to encode. At each visited node, a lightweight model then predicts an optimistic estimate of the best query similarity hidden among its descendants using the query and subtree representations. In evaluations with image and open-vocabulary text queries, GeoDIVE matches exhaustive search at 10% of the encoding cost for image queries and leads all tested baselines at every budget in both modalities. To support future evaluation of multiscale geospatial search, we also release object-level image and text queries for Git-10M, constructed from Microsoft building footprints and OpenStreetMap polygons with VLM-generated captions. Source code and models will be made publicly available.
[CV-110] Anchored Slice Propagation with Boundary Transport for 3D CT Segmentation from Three Orthogonal Annotated Slices
链接: https://arxiv.org/abs/2611.05693
作者: Tushar Kataria,Md Rakibul Haque,Shireen Y. Elhabian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Sli2Vol propagates a binary mask through affinities between adjacent slices, but an error can persist because each prediction becomes the input to the next step. Three orthogonal annotations reduce dependence on any one chain, yet they offer more than three starting masks: where a chain crosses another annotated plane, the correct label is already known. Those labels could correct drift during propagation and indicate how much to trust each chain at fusion. Between annotated planes, the information carried by the chain may matter as well. A binary mask records foreground membership but discards distance to the boundary, a cue that could help follow gradual shape changes between this http URL address these two sources of error with a fixed pretrained affinity backbone. We anchor each chain by correcting its prediction where it crosses another annotated plane, then use its agreement with those labels to weight fusion. We compare this anchored mask with two ways of transporting boundary geometry: a truncated signed-distance field (SDF) propagated by dense affinity averaging, and a contour whose vertices follow motion cues from the same affinity. The contour tracker starts from a hand-crafted rule and learns a motion correction from supervoxel pseudo-objects in 506 CT volumes, without manual organ labels. On 2,091 held-out structure instances from approximately 403 CT scans spanning 17 abdominal structures, anchoring raises mean Dice from 0.737 with orthogonal mask propagation to 0.769 and improves every structure. Beyond the anchored mask, learned contour tracking improves Dice by 5.7% for the esophagus and 5.9% for the postcava, while SDF transport improves right-adrenal Dice by 2.8%. The anchored methods have similar mean Dice (0.769–0.772); contour tracking performs worse on the kidneys, stomach, and hepatic vessel tree.
[CV-111] RUST-Stereo: Teacher-Assisted Robust Uncertainty-Guided Search and Transfer for Fast Stereo Matching
链接: https://arxiv.org/abs/2611.05689
作者: Pouya Sohrabipour,Yihong Feng,Chaitanya Kumar Reddy Pallerla,Tuan Dang,Ngan Le,Dongyi Wang
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Coarse-to-fine stereo matching enables efficient disparity estimation but remains sensitive to errors in propagated disparity, which may place ground-truth correspondences outside subsequent local search ranges. We propose TRUST-Stereo (Teacher-Assisted Robust Uncertainty-Guided Search and Transfer for Fast Stereo Matching), which addresses this limitation through an Uncertainty-Guided Residual Search Module (UGRSM). UGRSM estimates the reliability of propagated disparity and predicts residual offsets to adaptively relocate local cost-volume centers. Additionally, we introduce a two-stage Senior-Junior Teacher Assistant knowledge distillation strategy to transfer rich monocular representations from a large foundation model to a lightweight backbone to achieve fast stereo matching performance. Experiments on KITTI 2012, KITTI 2015, ETH3D, and Middlebury demonstrate strong zero-shot generalization. Using only 1.4M synthetic stereo pairs, TRUST-Stereo achieves state-of-the-art results on several benchmark metrics, outperforming competing efficient methods trained with substantially larger datasets, including real-world pseudo-labeled samples. Specifically, it achieves a KITTI 2015 EPE of 0.93, ETH3D Bad 1.0 of 2.13% and Middlebury Bad 2.0 of 6.79% while maintaining inference at approximately 50 ms per stereo pair.
[CV-112] oward an Optimal Configuration for Tree Species Classification from Bark Images: A Leakage-Safe Ablation Study
链接: https://arxiv.org/abs/2611.05680
作者: Maeve Côté
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 17 pages, 4 figures
Abstract:Background: The automatic identification of tree species from bark photographs is of practical interest for forestry inventory and the wood-processing industry, yet the design choices that govern the accuracy of deep-learning approaches have not been studied systematically, and some prior results may be inflated by leakage of images of the same trunk across training and test partitions. Methods: A systematic ablation study is conducted on the public BarkNet 1.0 dataset (20 species) under the established tree-level, leakage-safe partitioning scheme. Using ConvNeXt-V2 backbones, the following factors are varied in turn and evaluated by five-fold cross-validation: patch size, the strategy by which per-patch predictions are aggregated into an image-level decision, the training regime (single- versus two-stage), backbone capacity, the whole-image baseline, and the number of patches available at inference. Results: Per-patch accuracy rises monotonically with patch size (from 75.6% at 160 px to 89.6% at 512 px), and image-level accuracy rises in step, plateauing near 94.9% around 512 px. Learned aggregation exceeds majority voting by at most 0.7 pp, an advantage that shrinks as patches grow and attention becomes near-uniform. Backbone capacity (pico to tiny) alters accuracy by less than a point, and two-stage training offers no advantage over one-stage. Whole-image classification (91.9%) trails patch-based aggregation, since it admits no bag to average over; accuracy increases with the number of patches at inference, saturating around 64. Conclusions: The study yields a simple recommended configuration: patches as large as possible while leaving enough to average over (around 512 px here), with a nano backbone. It also quantifies the second-order role of learned aggregation and model capacity, providing a reproducible reference point for practitioners and future research
[CV-113] NitroFusion2: Bringing Adversarial Loss to the Text Stream of Multimodal Diffusion Transformers
链接: https://arxiv.org/abs/2611.05648
作者: Dar-Yen Chen,Yi-Zhe Song
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Project page: this https URL
Abstract:Adversarial diffusion distillation has conventionally relied on image features for discrimination. We revisit this choice for multimodal diffusion transformers (MMDiTs), whose text stream becomes image-dependent through joint attention. We introduce Text-Stream Adversarial Distillation (TSAD), which exploits intermediate text-token states as a new adversarial representation. Because the text stream has already read the image while remaining tied to the text condition, changing that condition changes what information it extracts from the image. Building on this property, we introduce Aspect-Cued Discrimination (ACD), which conditions the discriminator on different aspects of the sample, such as typography, objects, or style, to steer which properties are emphasized during adversarial training. Building on the expert-discriminator design of NitroFusion, we integrate TSAD and ACD in NitroFusion2. NitroFusion2 brings few-step distillation from the conventional 4-8-step regime down to just 1-2 steps at comparable quality. At one step, it produces sharp rendered text and rich fine-grained details, making one-step generation practical rather than merely a stress test for distillation. At two steps, it remains competitive across the evaluated benchmarks. These results establish the MMDiT text stream as a powerful and steerable representation for adversarial diffusion distillation.
[CV-114] ARIDA - Active railway infrastructure dataset annotation using neural network predictions
链接: https://arxiv.org/abs/2611.05646
作者: Maximilian Dornik,Christian Sieberichs,Simon Geerkens,Alexander Braun
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Modern railway perception requires large comprehensive image datasets, but creating them is slowed by manual labeling, especially for track polylines and other complex infrastructure geometry. This paper presents the Active Railway Infrastructure Dataset Annotation framework, a railway-specific semi-automated pipeline that converts instance predictions from a mask scoring instance segmentation network into bounding-boxes, polygons, polylines, and paired polylines. To keep automatic track geometry annotation reliable without ground truth, it applies confidence thresholds and an unique contour-agreement quality score as an acceptance gate. An active learning loop ranks remaining images by uncertainty and adapts the number of manual queries per cycle. This is achieved through a novel quantity control mechanism, in which easy images are auto-accepted and human effort targets hard cases. Experiments on disjoint, diverse imagery and time-sequential, redundant imagery quantify effort reduction and accuracy trade-offs, achieving about two thirds reduction in the disjoint regime and up to 93% in the sequential regime. The study also provides guidance for initial label budgets and stopping to avoid over-annotation.
[CV-115] Spatio-Temporal Multi-Order Interaction for Video Recognition
链接: https://arxiv.org/abs/2611.05633
作者: Kevin D. Nguetche,Lilian Hollard,Eric Granger,Marco Pedersoli
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Many temporal approaches for video recognition rely on explicit cross-channel aggregation projections, which have quadratic complexity in the number of channels, introducing significant computational overhead. In addition, the spatio-temporal representation used for video recognition is biased either toward local order interactions (for CNNs) or toward global order interactions (for transformers). Starting from the multi-order spatial representation of MogaNet, our method introduces two complementary contributions to tackle these problems. First, for the temporal modelling, we introduce parallel multi-order processing with depthwise-only convolutions and implicit gating-based aggregation, which is linear in the number of channels, inducing a considerable computational reduction. Second, we propose a dual interaction between spatial and temporal dimensions, which allows a better spatio-temporal representation without increasing the computational cost. By combining these two contributions, our proposed VideoMogaNet achieves performance on par with or better than state-of-the-art action recognition approaches across multiple video benchmarks (UCF101, HMDB51, Kinetics-400, Diving48), while considerably reducing the cost of the temporal branch.
[CV-116] CAPTOR: text-encoder-free grounding by counting caption co-occurrence over the visual words of a frozen encoder
链接: https://arxiv.org/abs/2611.05599
作者: Elad Yoshai,Natan T. Shaked
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Grounding a text query in an image needs a matrix scoring every term against every patch. Vision-language models compute it with an aligned text encoder, which many image encoders, such as histopathology-related models trained on slides alone, lack. The Caption-Anchored Patch-Term Occurrence Representation (CAPTOR) estimates the matrix without a text encoder. It clusters a frozen image encoder’s patch tokens into visual words and describes each captioned image by its share of patches in each word. A term signature is the standardized difference of those shares between images whose captions mention the term and the rest. The matrix factorizes into a term-by-word table and a query-independent image factor, so an image is encoded once and a new term adds one table row. Against every training-free baseline, CAPTOR leads on Common Objects in Context (COCO), Pascal Visual Object Classes (VOC) and Cityscapes. On COCO, scored on the cells one object owns, it reaches 0.565 mean intersection over union (mIoU) against 0.502. On all three, it is also level with a text alignment trained on the same backbone, which needs gradients, a text encoder and several times the captions. On held-out breast-cancer slides, it exceeds the strongest baseline scoring patches with a pathology text encoder by 0.160 mIoU. Read through CAPTOR, UNI, an encoder trained on slides alone, exceeds such baselines by 0.132 on a common grid. Counting can therefore give a text interface to any image encoder trained without language, wherever captioned images exist.
[CV-117] Relevance Propagation as Entropic Optimal Transport on Filippov Neural Networks: Sinkhorn Propagation and Region Relevance in Image Classifiers
链接: https://arxiv.org/abs/2611.05594
作者: Olivier Gibaru,Eric Nyiri
类目: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Optimization and Control (math.OC)
备注: 41 pages, 17 figures
Abstract:Relevance propagation, which redistributes a network’s prediction backward through the layers, is usually treated as a heuristic. We show that it is the solution of a single variational problem: Layer-wise Relevance Propagation (LRP) is the divergence-free flow on the active computational graph of zero relative entropy to the backward transition law of the forward activation flow; for a non-negative reference, this is a relative entropy between Markov measures on backward paths. The rules LRP _0 , LRP _z^+ and LRP _\alpha,\beta follow from this principle by the choice of a reference, and are instances of one rule, LRPN _\alpha,\beta , whose weights may vary from node to node, for instance with the node’s activity, in the spirit of Hebb’s principle. The principle rests on a measure-theoretic formalization of feed-forward networks as directed acyclic graphs of Filippov operators, which covers piecewise-smooth and discontinuous architectures (e.g. hard-routed Mixture-of-Experts) with a Clarke-Filippov generalized-gradient calculus; it also characterizes backpropagation as the adjoint equation of a Lagrangian relaxation. On networks of ReLU and max-pooling nodes, LRP _0 is Gradient \times Input, with an explicit jump across a kink where LRP _z^+ stays continuous. Taking the entropy edge by edge yields a new rule, Sinkhorn Relevance Propagation. Aggregated over an input region, the flow of a non-negative reference tells, with exact accounting, which hidden units carry its evidence, the conditioned flow being a Schrödinger bridge. On VGG-16, the conservation identities hold to round-off, and the rules are told apart by the sign of their maps, the share of the output left to the bias nodes, faithfulness, localization and class dependence; on ConvNeXt-Tiny, LRPN _\alpha,\beta conserves the output and its Hebbian weights give the highest ABPC of the rules tested.
[CV-118] GaussianRenderFormer: Transformer-based Rendering of 3D Gaussian Splats
链接: https://arxiv.org/abs/2611.05593
作者: Shahaf Valencia Levy,Sagie Benaim
类目: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Machine Learning (cs.LG)
备注: Under review at TMLR. 32 pages. Project page: this https URL
Abstract:Can a neural network learn to render 3D Gaussian splats on its own, with no rasterizer in the renderer? We present GaussianRenderFormer, a transformer that takes a set of Gaussians and a camera and outputs the image directly. Each Gaussian becomes one token in a view-independent encoder, and a view-dependent decoder turns small patches of camera rays into pixels by attending to those tokens. Trained on 26,820 Objaverse objects pruned to 20,000 Gaussians each, GaussianRenderFormer renders 300 unseen objects at 38.08 dB PSNR against the full, unpruned splats, while rasterizing the same pruned input reaches 44.02 dB. The gap is not one of generalization, and our diagnostics are consistent with a bottleneck in the view decoder: telling it the projected footprint of each Gaussian, which it could in principle infer, adds 1.4 dB. Under-training is not ruled out. Code, trained models and videos: this https URL
[CV-119] BrainBind: Bridging Synchronized EEG and fMRI through Long-Form Naturalistic Video Representations BMVC2026
链接: https://arxiv.org/abs/2611.05581
作者: Shixiao Wang,Rining Wu,Jian K. Liu
类目: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
备注: Accepted at BMVC 2026. 26 pages, 4 figures, including supplementary material. Code: this https URL
Abstract:Real-world vision is dynamic and scene-rich, while EEG and fMRI resolve it at complementary temporal and spatial scales. Most neural decoding systems, however, align one modality in isolation and mostly against static images or short clips. We study the synchronized case: EEG and fMRI are recorded during the same long-form naturalistic viewing, giving a shared stimulus but also recurring narrative context and many semantically similar candidates. We present BrainBind, a framework that maps synchronized EEG and fMRI into frozen long-form video embedding spaces and treats those spaces as common visual-semantic references, not literal neural codes. On CineBrain, a five-seed Recall@K 2 x 4 matrix across two video anchors and four system configurations shows distinct retrieval profiles. With Omni, fMRI organizes the full candidate list most strongly (MRR = 23.82, nDCG@10 = 25.31), while Joint-Dual gives the highest mean label-aware R@1 (13.52%) with a distinct rank-wide profile. The complementary episode-disjoint evaluation quantifies the contribution of repeated stimulus identity and narrative context to the random split. Across these experiments, proxy choice and system configuration produce distinct top-rank and rank-wide retrieval profiles within the present synchronized dataset. The code is available at this https URL.
[CV-120] Multimodal Fusion Methods for Whole-Slide Images and Clinical Data in Inflammatory Bowel Disease
链接: https://arxiv.org/abs/2611.05579
作者: Ankana Banerjee(1),Mark van der Giezen(1, 2, 3),Ole Gunnar Aasprong(2),Emiel A. M. Janssen(1, 2),Øyvind Steinsbø(2, 4),Tore Grimstad(2, 4),Kjersti Engan(1) ((1) University of Stavanger, Norway, (2) Stavanger University Hospital, Norway, (3) University of Greenwich, UK, (4) University of Bergen, Norway)
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 5 pages, 4 figures, Submitted to the 2027 IEEE 24th International Symposium on Biomedical Imaging (ISBI)
Abstract:Multimodal AI fusion of histology and clinical data is an active research area in computational pathology, but fusion methods are rarely compared under identical conditions and across different clinical endpoints. We present a modular framework in which slide-level embeddings, obtained by training an attention-based aggregator over patch embeddings from a frozen foundation model, and per-visit clinical embeddings are fed into 15 interchangeable fusion methods with a shared output dimension and prediction head (except decision-level methods). Using a prospective cohort of newly diagnosed inflammatory bowel disease patients, we predicted STRIDE-II-aligned remission outcomes from baseline or follow-up visits. The best fusion method reached macro-F1 of 0.57-0.75, outperforming both single-modality models in four of six scenarios, and differed by outcome and input. Image reliance, the drop in balanced accuracy when the slide embedding is removed, tended to follow the remission type - highest for histological and lowest for endoscopic remission; methods with similar accuracy relied on the slide to very different degrees.
[CV-121] Whose Values Are Generated? Measuring Visual Value Leakage in Multilingual Text-to-Image Generation
链接: https://arxiv.org/abs/2611.05578
作者: Zucan Lyu,Binwei Yao,Rheeya Uppaal,Jing Yao,Xiaoyuan Yi,Junjie Hu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 38 pages
Abstract:Text-to-image (T2I) systems serve users across languages and cultural contexts. However, evaluations of cultural representation focus mainly on recognizable elements such as landmarks, clothing, and activities, overlooking the cultural values that generated images convey. To bridge this gap, we introduce VisionValueBench, a multilingual benchmark that operationalizes six of Hofstede’s cultural value dimensions (e.g., individualism versus collectivism) to investigate how T2I systems frame neutral social situations through a particular cultural orientation. We find that value-neutral prompts, which admit diverse cultural interpretations, nonetheless yield images that favor particular value orientations, a phenomenon we call visual value leakage. We evaluate prompts varying in language and geographic cues across 960 social scenarios and six T2I configurations. Generated images consistently favor the same value orientations across all six dimensions, for example, equality over hierarchy (low power distance), individual autonomy over group obligation (individualism), and enjoyment over restraint (indulgence). Prompt language has limited influence on these orientations, whereas geographic cues shift some dimensions toward the associated cultural values while largely preserving the prevailing orientations. These findings reveal substantial homogeneity in the cultural values conveyed by T2I systems worldwide.
[CV-122] ChangeJEPA: Predict Before You Compare for Remote Sensing Change Detection
链接: https://arxiv.org/abs/2611.05570
作者: Buddhi Wijenayake,Dhananjaya Jayasundara,Vishal M. Patel
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages
Abstract:Remote sensing change detection is typically formulated as Siamese feature comparison, where two images are encoded independently and a decoder determines which discrepancies correspond to meaningful change. This is challenging because the encoders preserve both true scene changes and nuisance variations such as illumination, seasonal, radiometric, and geometric differences. We introduce ChangeJEPA, a bitemporal joint-embedding predictive architecture that reformulates change detection as latent post-state prediction. Given a pre-change image and partial post-change evidence, ChangeJEPA predicts the full post-change representation under two complementary conditions: a prior hypothesis derived from the pre-change scene and an updated hypothesis conditioned on observed post-change context. A joint-embedding predictive loss aligns both hypotheses with the target latent representation, encouraging semantic consistency while suppressing transient appearance variations. Change is then inferred from the disagreement among the pre-change representation, the observed post-change target, and the predicted latent states, capturing complementary evidence of change, prior surprise, contextual update, and prediction reliability. Across SYSU-CD, LEVIR-CD+, and CDD, ChangeJEPA consistently outperforms its decoder matched Siamese baseline by up to 1.51 IoU points and achieves the best IoU and F1-score on all three benchmarks, demonstrating the advantage of latent post-state prediction over direct feature comparison. Page: this https URL
[CV-123] Role-based Effectiveness of an Epistemic Uncertainty Estimate in Training Models for Semantic Segmentation of 3D LiDAR Point Clouds
链接: https://arxiv.org/abs/2611.05569
作者: Krishnakumar N,Jaya Sreevalsan-Nair
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: 23 pages, 7 figures, 10 tables
Abstract:Point cloud segmentation is a fundamental task in 3D data analysis, with applications across robotics, augmented and virtual reality, and remote sensing. However, the inherent sparsity, noise, and unordered nature of point clouds introduce significant challenges, particularly in achieving reliable classification. While the state-of-the-art rely on training models for explicit modeling of uncertainty, there is a gap in exploring different roles that estimated uncertainty using generative approaches can play. This work investigates the impact of deterministic uncertainty in point cloud segmentation by explicitly modeling epistemic uncertainty and using it in different roles, namely, as a feature and in loss function. We evaluate a baseline artificial neural network (ANN) and two uncertainty-aware variants (uncertainty as an additional input feature and as a per-sample confidence weight in the training loss) on four large-scale LiDAR benchmarks spanning airborne and mobile acquisition, namely ISPRS Vaihingen, AHN3, Paris-Lille-3D, and Toronto-3D, together with PointNet++ and RandLA-Net references and their uncertainty-augmented counterparts under canonical spatially disjoint train and test splits. Under this protocol, uncertainty-aware modeling improves mean Intersection over Union, recovers minority classes that the baseline ANN never predicts at all, and in half of the configurations improves the alignment between prediction error and estimated uncertainty as measured by Patch Accuracy versus Patch Uncertainty. Our findings indicate that uncertainty can play multiple roles in the segmentation pipeline rather than as a device for improving accuracy alone.
[CV-124] Discarded or Unreachable? Diagnosing and Repairing Spatial Relation Failure in Vision–Language Similarity
链接: https://arxiv.org/abs/2611.05565
作者: Mahdi Mostajabdaveh,Zirui Zhou
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Code, stimulus generators, and all result files: this https URL
Abstract:Contrastive vision-language models score left/right relations near chance, but the symptom has two causes that call for opposite repairs: the relation may survive in the pooled embedding yet be unreachable by native image-text similarity, or pooling may discard it while the dense features retain it. On clean two object photographs, across ten encoders, the lateral polarity is linearly decodable from the flip odd component of the pooled vector at 0.77-0.99, yet a search over a million captions leaves nine of the ten at chance under native scoring: an access failure. On controlled C4 orbits, where the relation is the only thing that varies, pooling keeps the arrangement axis at 0.998-1.000 and loses its polarity, which the patch grid retains: a pooling failure. Both sit downstream of the dense features, so one readout covers both. Our zero parameter, query conditioned spatial moment recomputes the geometry from the patch grid and stores it as a vector beside the native embedding. On a 118,287 image index it raises left/right pairwise accuracy against reversed geometry negatives from 0.51-0.53 to 0.81-0.96 and conditional R@1 among 21 category matched candidates from 0.05 to 0.37-0.75, at under a millisecond per batched query, while reproducing the native ranking exactly for queries that name no relation.
[CV-125] Lesion Overlap Ranks PET Segmentation Models but Does Not Certify Patient-Level Tumor Burden
链接: https://arxiv.org/abs/2611.05564
作者: Tushar Kataria,Gerald Sabin,Ponnuswamy Sadayappan,Shireen Y. Elhabian
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Metabolic tumor volume (MTV) summarizes whole-body PET disease burden for prognostic stratification and treatment assessment. Automated MTV is useful only if its estimate is reliable for the individual patient, yet segmentation models are usually evaluated with overlap metrics such as Dice rather than burden error. We examined whether overlap metrics identify unreliable MTV estimates using 1,014 autoPET FDG PET/CT studies. We evaluated LesionTracer predictions held out from fine-tuning, compared five selected released models on 489 tumor-positive patients, and examined the findings in 597 PSMA studies. Across the five FDG models, ranking by higher mean Dice gave the same order as ranking by lower mean relative MTV error ( \rho=-1.00 , exact p=0.017 ). At the patient level, however, out-of-fold Dice tracked relative MTV error strongly ( \rho=-0.75 ), yet 82 of 489 patients (16.8%) fell on opposite sides of the Dice 0.80 and MTV-error 25% thresholds. Within the training-exposed released models, correlations ranged from -0.47 to -0.74 over narrower, high-Dice ranges. Out of fold, opposing false-positive and false-negative volumes cancelled about one third of spatial error in net MTV. In 236 of 489 patients, each of the five models missed at least one lesion, though not necessarily the same lesion. Thus, agreement in average model rankings did not establish the reliability of individual MTV estimates or reveal shared lesion misses. PET segmentation benchmarks should report patient-level MTV error and missed-lesion counts alongside Dice and other overlap metrics.
[CV-126] How Finely Can Vision-Language Models See Biology? Bio-VQA: A Benchmark for Taxonomic Morphological Spatial and Individual Reasoning WACV2027
链接: https://arxiv.org/abs/2611.05559
作者: S M Rayeed,Qian Ma,Charles V. Stewart,Yao Ma
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to WACV 2027, 8 pages main paper; 4 figures, 4 tables
Abstract:In biodiversity monitoring, a single image raises several questions: which taxon, how many, which trait, and whether the individual has been seen before. Vision-language models (VLMs) could answer all of these from the image, but their ability is largely untested in biology, as existing benchmarks focus mostly on species identification within a few taxonomic groups. We present Bio-VQA, the largest biological visual question answering (VQA) benchmark and the only one to span four aspects of biological visual understanding: hierarchical taxonomic reasoning, fine-grained attribute grounding, quantitative spatial reasoning, and individual re-identification. Bio-VQA contains 1.86M questions over 217K images from 11 public datasets, and extends right-or-wrong scoring to graded scoring, so a near-miss counts for more than a distant one. Evaluating 18 open-weight models from 3.75B to 38.4B parameters, we find that general visual capabilities do not transfer to biological organisms: the best model counts animals exactly on 41.7% of images with 2 to 10 individuals, less than half its accuracy on everyday objects at the same counts. Taxonomic accuracy falls to one sixth, from 0.72 at phylum to 0.12 at scientific name, and scaling helps mostly at coarse ranks. The gap to a domain specialist widens accordingly, from 0.731 against BioCLIP-2’s 0.874 at class (1.2x) to 0.269 against 0.544 at scientific name (2x). On trait size measurement, 15 of the 18 models do no better than an image-blind baseline that predicts one fixed number for every specimen, and on re-identification, 17 of the 18 models fall below chance under adversarial conditions. These failures persist across a tenfold range of model size. We hope Bio-VQA becomes a standard test of how well VLM capabilities transfer to the biological domain.
[CV-127] From Activation to Intervention: Interpreting Sparse Visual Features through Reconstruction
链接: https://arxiv.org/abs/2611.05522
作者: Jianyang Gu,Jacob Beattie,Mridul Khurana,Julia Chae,Xi Xiao,Cheng Zhang,Anuj Karpatne,Wei-Lun Chao,Sara Beery,Tanya Berger-Wolf
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: Demo and code available at this https URL
Abstract:Sparse autoencoders (SAEs) decompose visual representations into individual features, offering a promising basis for explaining model behavior. Interpreting these features typically relies on images or pixel regions that most strongly activate them. However, this evidence becomes ambiguous when multiple features with different semantics frequently activate in the same locations, such as texture and color in the same patterns. We propose Reconstructive SAE Intervention (RecSAEI), a framework that shifts feature explanation from passive observation to active visual reconstruction. By intervening on representations along SAE decoder directions and reconstructing the visual output, we provide direct evidence of a feature’s semantic meaning. We demonstrate that our method successfully disambiguates feature meanings across multiple datasets and vision encoders. Furthermore, our framework grounds [CLS] SAE interventions in a sparse set of patch SAE decoder directions. Beyond stable target response, this attribution also exposes the compositional relationship between elementary visual patterns and high-level semantics. Our results advocate for intervention reconstruction as an essential complementary signal for explaining neural networks.
[CV-128] From Tokens to Trajectories: Parameter-Efficient Continuous Adaptation of Unified Multimodal Models
链接: https://arxiv.org/abs/2611.05451
作者: Gayatri Deshmukh,Onkar Susladkar,Adheesh Juvekar,Halil Kilicoglu,Ismini Lourentzou
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:While discrete tokenization in Unified Multimodal Models (UMMs) establishes strong semantic reasoning, it introduces a strict quantization bottleneck that degrades fine-grained visual details, often resulting in structurally accurate but artifact-heavy images. We introduce FlowShift, a parameter-efficient post-training framework that bypasses this bottleneck by converting pre-trained discrete UMMs into continuous rectified-flow generators. Instead of discarding the backbone’s semantic priors, FlowShift replaces discrete visual interfaces with continuous latent projections and applies LoRA to the attention layers. To efficiently boost visual quality without semantic drift, we propose Dual-Objective Velocity Vector Distillation (DOVD), which aligns the model with a high-fidelity continuous teacher and ground-truth continuous trajectories. This approach drastically improves perceptual quality while preserving prompt adherence and reasoning. When applied to leading discrete UMMs, FlowShift significantly improves COCO-30K FID and aesthetic scores, while retaining over 99.5% of multimodal understanding capabilities. Ablations confirm DOVD drives generation quality. Ultimately, FlowShift achieves high-fidelity continuous generation without sacrificing pre-trained semantic grounding.
[CV-129] SO(3)-Equivariant Learning on CAD Boundary Representations NEURIPS2026
链接: https://arxiv.org/abs/2611.05449
作者: Matteo Ballegeer,Dries F. Benoit
类目: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
备注: Accepted to NeurIPS 2026 main track
Abstract:Learning on boundary representations (B-reps) requires robustness to the pose of a part, since CAD models are not stored in a canonical orientation. Existing B-rep networks either operate on world-frame coordinates and normals, and thus degrade under rotation, or encode each face in a local reference frame. Local frames yield invariance, but reduce inter-face messages to scalars and discard the relative orientation of adjacent faces. We propose SO3brep, an SO(3)-equivariant graph neural network on the B-rep face-adjacency graph. Node features are computed in local frames and mapped into a shared frame by Wigner D-matrices. Clebsch-Gordan tensor products couple these features with the directions between adjacent faces, so that relative orientation is preserved throughout message passing and the output remains SE(3)-invariant. Across three face-segmentation and two classification benchmarks with randomly rotated test parts, SO3brep achieves the highest accuracy among the compared B-rep networks and equivariant point cloud networks, with the fewest parameters. Its advantage is largest with few labeled parts, a setting typical of industrial CAD. Code is available at this https URL.
[CV-130] SurgTransfer: Multi-Step Latent Action Modeling for Action Transfer in Surgical World Model
链接: https://arxiv.org/abs/2611.05443
作者: Saurabh Koju,Love Panta,Deepak Thapa,Rudra P. K. Poudel,Prajwal Ghimire,Angus Watson,Yash Raj Shrestha,Prashnna Kumar Gyawali,Binod Bhattarai
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Surgical world models have the potential to provide intra-operative guidance, accelerate risk-free training and facilitate transfer of expert surgical maneuvers. However, the use of world models for transferring surgical expertise has been largely unexplored. Moreover, recent approaches in this domain have mainly relied on ground truth kinematics actions or expert annotations for specific tasks, limiting their applicability beyond the narrow range of tasks with ground-truth annotations. In contrast, existing works in the natural domain have shown promising results in learning latent actions solely from unlabeled videos and transferring the learned actions across contexts. We observe that this paradigm is short-sighted for the surgical domain, failing to capture the hierarchical and goal-directed structure of surgical procedures. Thus, we propose a novel multi-step action learning framework to learn context-invariant latent actions that are easily transferable across various surgical contexts. We perform extensive experiments on the publicly available datasets, SurgToolLoc-2022 and PitVis-2023, verifying our approach with both quantitative metrics and qualitative evaluation by surgical experts. We have published our code and the implementation details at this https URL.
[CV-131] PointNDF: Cascaded Neural Displacement Fields for Point Cloud Denoising
链接: https://arxiv.org/abs/2611.05411
作者: Michael Y. Wang,Pingping Cai
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 14 pages, 3 figures, 7 tables. Accepted to MMM 2027 (33rd International Conference on Multimedia Modeling)
Abstract:Point cloud denoising aims to recover clean geometry from noisy 3D observations while preserving both fine surface details and the underlying spatial distribution of the points. Many learning-based methods adopt a two-stage pipeline that first extracts point-wise features from the noisy input and subsequently predicts point-wise displacements via simple regressions. However, these approaches often lack robust feature guidance and distribution-aware geometric refinement, causing them to fail at maintaining optimal point uniformity and geometric quality. To address these limitations, we propose PointNDF, a denoising framework that combines patch-level geometric context with local displacement prediction. Specifically, a Point Transformer V3 (PTv3) encoder aggregates multiscale geometric information into a global patch descriptor. Rather than directly regressing an unconstrained displacement vector, a Neural Displacement Field (NDF) fuses this global patch descriptor with per-point neighborhood features. It then predicts each displacement through bounded-vector and scale factor branches, thereby achieving a superior, more uniform point distribution. Three cascaded NDF stages propagate local features and progressively refine the point positions. Experiments on PU-Net and ModelNet10 demonstrate competitive denoising accuracy and favorable inference efficiency. The source code is publicly available at this https URL.
[CV-132] Augmentation with Reinforcement Transformation for Low-textured Scientific Sketch Image Retrieval
链接: https://arxiv.org/abs/2611.05408
作者: Zhiqiang Wang,Sheri L. Barton,Jennifer H. Johnson,Yi He,Xingquan Zhu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: IEEE BigData 2026
Abstract:Low-textured sketch images are widely used in scientific applications, including wildlife conservation, archaeological illustration, and forensic composites. Learning from and retrieving such images is challenging because they lack detailed surface textures and visual variations commonly exploited by modern deep learning models. This paper proposes Augmentation with Reinforcement Transformation (ART) to fine-tune pretrained models to learn effective representations for accurate sketch image retrieval. Our key observation is that sketches are particularly sensitive to image transformations, such as rotation, flipping, zooming, and stretching. While some transformations preserve their structural integrity, others may introduce distortions that hinder representation learning. Instead of randomly choosing transformations to construct fine-tuning views, ART formulates augmentation selection as a multi-armed bandit (MAB) problem, where transformations are treated as arms and their effectiveness is continuously evaluated through rewards. This enables ART to adaptively construct reward-guided view banks before self-supervised fine-tuning; the subsequent representation loss does not feed back to ART. On Florida manatee sketches provided by Mote Marine Laboratory, ART achieves 71.40% average Top-1 accuracy for transformed-scar retrieval, compared with 68.83% for random augmentation selection, and 55.84% for location-aware sketch retrieval, compared with 18.18% for frozen CLIP.
[CV-133] PEEK: One-Step-Look-Ahead Exposure Bias Correction for Diffusion Sampling NEURIPS2026
链接: https://arxiv.org/abs/2611.05407
作者: Yunwei Bai,Ying Kiat Tan,Tsuhan Chen
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注: Accepted to Neurips 2026
Abstract:Diffusion image generators are successful due to accurate denoising networks and strong numerical samplers. Yet the two components are not fully matched. During sampling, the denoiser predicts for the next step based on flawed input that differ much from the real data. This train-test gap, or exposure bias, caps image generation quality (i.e., blurry, lacking in details or awkward in shape). To mitigate the problem, we propose PEEK, a novel training-free deterministic sampler plugin. At each eligible denoising step, PEEK first performs a provisional one-step update, evaluates the denoiser once at the look-ahead state, and interpolates it with the one-step-later prediction to form the next noisy state. By taking the additional step, we obtain a generally more-image-like prediction from the higher-signal noisy state, mitigating the exposure bias. To optimize the interpolation positions, PEEK searches for an interpolation schedule search made efficient by the prefix and baseline noise-schedule-rebasing caches. On 8 pixel-space and latent-space diffusion models with unconditional, label-guided, and text-guided settings across various output resolutions, PEEK consistently improves over NFE-matched baselines, with FID improvements up to 49.3% at minimal extra computation cost (e.g., the one-time-per-model calibration takes at most 2.2 minutes across our experiments.). Project website is at this https URL.
[CV-134] When Does Sharing Work? Scaling a Shared SigLIP2 Across Image Video Audio and Audio-Visual Representations
链接: https://arxiv.org/abs/2611.05406
作者: Xiaofei Sun,Peng Wang,Huayao Liu,Xiaoxiao Xu,Jun Song,Deqing Li,Lin Qu
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
备注:
Abstract:Contrastive pretraining has advanced image-text understanding, yet image, video, and audio encoders are still trained independently, with cross-modal alignment realized mainly through post-hoc binding to a frozen anchor space. We present Omni-Encoder, a general multimodal encoder that processes image, video, audio, and audio-visual inputs within a single SigLIP-2 Vision Transformer with fully shared parameters. Unlike anchor-centric binding or per-modality adapters, all modalities are optimized jointly end-to-end, and the framework is organized from input to training. A modality-native input stage first maps each signal into the shared embedding space, including a 1D-convolutional audio tokenizer that converts mel-spectrograms into tokens aligned with visual patches. A multimodal rotary position embedding parameterized by absolute timestamps then places visual frames and audio segments in a unified spatiotemporal coordinate system. Finally, alternating-modality training interleaves image, video, audio, and temporally aligned audio-visual batches and accumulates their gradients into one update, which we find central to stable joint optimization. Omni-Encoder achieves competitive results on 16 benchmarks across all four modalities with one shared parameter set, and joint training improves each modality over its separately trained counterpart. Code and models will be publicly released.
[CV-135] mmHand: Toward Pixel-Level-Accuracy Hand Localization Using a Single Commodity mmWave Device
链接: https://arxiv.org/abs/2611.05404
作者: Xiaoyu Zhang,Zhengxiong Li,Chenhan Xu,Luchuan Song,Huining Li,Hongfei Xue,Yingxiao Wu,Wenyao Xu
类目: Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:The hand localization problem has been a longstanding focus due to its many applications. The task involves modeling the hand as a singular point and determining its position within a defined coordinate system. However, due to data modality limitations, existing hand localization technologies face several challenges. For example, vision-based localization raises privacy concerns, while wearable-based methods compromise user comfort. In this paper, we introduce mmHand, a new device-free, privacy-preserving dynamic hand localization system with pixel-level accuracy, using a single commodity mmWave device. We first propose a mmImage generation tool to fully extract spatial information from raw mmWave data and introduce a novel 2D image-format representation of mmWave data. Next, we design a framework that provides a new quality evaluation method and pixel space labeling for the mmWave data. Finally, we present a cross-modality spatial feature-enhanced model with high spatial feature extraction capabilities, which can accurately localize hand positions at the pixel level in the mmWave radar U-V pixel coordinate system. We evaluate the system with experiments on 12 subjects in three scenarios, and the results across four metrics demonstrate the effectiveness of our hand localization system.
[CV-136] Visual Forensic Detection of Fabricated Bangla News Photocards using Deep Learning
链接: https://arxiv.org/abs/2611.05403
作者: Anindya Kundu,Kabya Mithun Saha,Md. Muhaiminul Islam,Fayek Ahmed,Aminul Kader Bulbul,Muhammad Ibrahim
类目: Computer Vision and Pattern Recognition (cs.CV)
备注: 15 pages, 9 figures, 22 tables
Abstract:In Bangladesh, much localized misinformation spreads as photocards: single-image news summaries that imitate the branding of well-known news organizations. Because the misleading message lives in the pixels rather than in machine-readable text, text-based fact-checking tools cannot see it. We recast photocard fake-news detection as visual image forgery detection. We build a dataset of 969 authentic and 1,077 fabricated Bangla photocards scraped from public Facebook pages: authentic cards from the verified pages of seventeen national television broadcasters, fabricated ones from pseudo-broadcaster pages, anonymous profiles, and public groups. Sixteen photometric and morphological transformations per image grow the corpus to 32,736 samples. A fully connected ANN reaches only about 52% accuracy, its loss stuck at ln(2) ~ 0.693, showing that flattening the image destroys the spatial cues forgery detection needs. A compact, fully batch-normalized four-block CNN with 489,954 parameters, about 1,644x smaller than the ANN, is evaluated with leakage-free stratified group five-fold cross-validation, which keeps every augmented version of a card in one fold. Under this protocol the CNN reaches 87.45% +/- 0.53% pooled accuracy from scratch and 88.75% +/- 1.55% when its classifier head is warm-started from the ANN. An ImageNet-pretrained YOLOv8n-cls backbone (1.44M parameters) fine-tuned on the same folds reaches 97.36% +/- 0.42%, marking the headroom above the compact model. We report per-class precision, recall, F1, and pooled confusion matrices, diagnose and fix a one-class collapse caused by an unnormalized classifier head, and give a manual error analysis and bias-variance diagnosis showing that the compact CNNs under-fit.
[CV-137] VAR-Q: Tuning-free KV Cache Quantization for Visual Autoregressive Image and Video Generation ICCV2025
链接: https://arxiv.org/abs/2611.05380
作者: Boxun Xu,Jiaji Lu,Yuxuan Yin,Zihu Wang,Ziyue Liu,Yu Wang,Zirui Liu,Peng Li
类目: Computer Vision and Pattern Recognition (cs.CV); Performance (cs.PF)
备注: A preliminary version of this work was presented at the ICCV 2025 BiVision Workshop
Abstract:Visual autoregressive generation spans two dominant paradigms: next-scale prediction for images and video, and next-frame prediction for video. Both decode block-by-block rather than token-by-token, yielding KV caches whose statistics diverge from those of language models. At scale, KV memory exceeds the weights and limits large-scale deployment caused by a large amount of visual tokens. Our analysis of these caches reveals three findings. First, a small set of massive-value channels carries the functional load: ablating these outlier channels collapses generation, whereas ablating a matched set of inlier channels leaves quality intact. Second, the outliers organize along two axes, feature and block, with magnitudes persistent within each cell and shifting across cells. Third, this structure arises from two independent mechanisms: projection bias and RoPE-induced scattering. Guided by these findings, we propose VAR-Q, a tuning-free KV-cache quantization framework that groups along the same two axes. The generation loop provides block boundaries for free, removing the need for outlier detection or calibration. Paired with bit-packing kernels for runtime memory reduction, VAR-Q delivers low-bit quantization across nine models spanning both paradigms, cutting KV memory by up to 78.5% with negligible quality degradation and enabling 6x larger inference batches.
[CV-138] Investigating Anthropometric Fidelity in SAM 3D Body
链接: https://arxiv.org/abs/2601.06035
作者: Aizierjiang Aiersilan,Ruting Cheng,James Hahn
类目: Graphics (cs.GR); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Emerging Technologies (cs.ET)
备注:
Abstract:The release of SAM 3D Body is a recent development in human mesh recovery, demonstrating improved performance in producing clean, topologically coherent meshes from single images. By leveraging the Momentum Human Rig (MHR), it achieves robustness to occlusion and diverse poses. However, our evaluation reveals a specific and consistent limitation: the model struggles to reconstruct detailed anthropometric deviations, particularly in populations exhibiting distinctive morphological alterations such as geriatric muscle atrophy, scoliosis, or pregnancy, even when these features are prominent in the input image. In this paper, we investigate this phenomenon not as a failure of the model’s capacity, but as a byproduct of the “perception-distortion trade-off”. We posit that the architectural reliance on the low-dimensional parametric MHR representation, combined with semantic-invariant conditioning (DINOv3) and annotation-based alignment, creates a pervasive “regression to the mean” effect. We analyze these mechanisms to understand why individual biological details are smoothed out. Furthermore, we state our contributions by proposing specific, constructive pathways for future work, such as implicit-explicit hybrid representations and Medical-in-the-Loop alignment, to extend the baseline performance of SAM 3D Body into the high-precision medical domain.
[CV-139] MvBody: Multi-View-Based Hybrid Transformer Using Optical 3D Body Scan for Explainable Cesarean Section Prediction
链接: https://arxiv.org/abs/2511.03212
作者: Ruting Cheng,Boyuan Feng,Yijiang Zheng,Chuhui Qiu,Aizierjiang Aiersilan,Joaquin A. Calderon,Wentao Zhao,Qing Pan,James K. Hahn
类目: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR); Machine Learning (cs.LG); Multimedia (cs.MM)
备注: 19 pages, 4 figures
Abstract:Accurately assessing the risk of cesarean section (CS) delivery is critical, especially in settings with limited medical resources, where access to healthcare is often restricted. Early and reliable risk prediction allows better-informed prenatal care decisions and can improve maternal and neonatal outcomes. However, most existing predictive models are tailored for in-hospital use during labor and rely on parameters that are often unavailable in resource-limited or home-based settings. In this study, we conduct a pilot investigation to examine the feasibility of using 3D body shape for CS risk assessment for future applications with more affordable general devices. We propose a novel multi-view-based Transformer network, MvBody, which predicts CS risk using only self-reported medical data and 3D optical body scans obtained between the 31st and 38th weeks of gestation. To enhance training efficiency and model generalizability in data-scarce environments, we incorporate a metric learning loss into the network. Compared to widely used machine learning models and the latest advanced 3D analysis methods, our method demonstrates superior performance, achieving an accuracy of 84.62% and an Area Under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.724 on the independent test set. To improve transparency and trust in the model’s predictions, we apply the Integrated Gradients algorithm to provide theoretically grounded explanations of the model’s decision-making process. Our results indicate that pre-pregnancy weight, maternal age, obstetric history, previous CS history, and body shape, particularly around the head and shoulders, are key contributors to CS risk prediction.
[CV-140] Object-independent evaluation of small-data astronomical image classifiers: architecture representation and acquisition shortcuts
链接: https://arxiv.org/abs/2611.06484
作者: Eraldo Pereira Marinho
类目: Instrumentation and Methods for Astrophysics (astro-ph.IM); Computer Vision and Pattern Recognition (cs.CV)
备注: 17 pages and 3 figures
Abstract:Reported performance in small astronomical image datasets can reflect object reuse, field-of-view differences, acquisition settings, and catalogue-derived properties as well as astronomical morphology. We designed \dataset to measure these effects in a four-class task comprising galaxies, globular clusters, nebulae, and open clusters. The dataset contains 370 catalogue-supported physical objects, each represented by three Digitised Sky Survey FITS bands, and uses a group-constrained split of 222 training, 75 validation, and 73 test objects. We compared a custom CNN–Transformer trained from scratch with ImageNet-pretrained ResNet-50 and ViT-B/16 under an adaptive full-field representation, a common angular field, and two central-region-masked controls, using three fixed seeds. In the one-pass object-independent test, mean balanced accuracy on the adaptive view was 0.773\pm0.014 , 0.874\pm0.054 , and 0.866\pm0.023 for the custom model, ResNet-50, and ViT-B/16, respectively. No primary architecture contrast rejected its null after Holm correction; ResNet-50 and ViT-B/16 differed by 0.007 balanced-accuracy points (95% cluster-bootstrap interval [-0.048,0.062] ). Controls based on acquisition and catalogue metadata reached 0.721–0.735, and central-region-masked adaptive views reached 0.731–0.800. These results demonstrate generalisation to unseen physical objects, but also show that image context, acquisition settings, and catalogue-derived object properties carry substantial class information. In this regime, high accuracy cannot be interpreted as evidence that a classifier learned target morphology, and the comparison does not isolate architecture from pretraining.
[CV-141] OP: Transmission-Oriented Remote Sensing Pansharpening
链接: https://arxiv.org/abs/2611.06430
作者: Zi-Han Cao,Jing-Ze Liang,Shao-Jie Wu,Jin-Cheng Di,Liang-Jian Deng,Gemine Vivone
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Pansharpening has been extensively studied as a fundamental image fusion task in remote sensing. However, most existing methods focus on improving fusion quality under the implicit assumption that high-resolution observations have already been fully downlinked to the ground. As a result, existing models are highly vulnerable to perturbations introduced by lossy compression used for downlink transmission. To address this limitation, we present, to the best of our knowledge, the first transmission oriented pansharpening (TOP) framework, which unifies compact latent coding, onboard pansharpening, and efficient downlink transmission within a single satellite-toground pipeline. TOP encodes remote-sensing observations into a compact binary latent space, where a lightweight onboard fusion module can efficiently predict the fused binary latent code. Since only the compact fused code is downlinked, the transmission payload is substantially reduced. Overall, TOP enables low-payload downlink and lightweight onboard processing while maintaining high fusion quality under severe bandwidth constraints. Experiments across multispectral and hyperspectral settings show that TOP consistently achieves superior fusion quality at higher compression ratios than representative pansharpening methods paired with standard lossy codecs. For example, on GaoFen-2, TOP attains a fusion PSNR of 31.4 dB at a downlink compression ratio of 50x. This design also points toward transmissionoriented processing for other remote-sensing tasks. The code will be made available at this https URL.
[CV-142] Dual Flow: Joint Image-Measurement Flow for Inverse Problems
链接: https://arxiv.org/abs/2611.05843
作者: Sevket Ege Yesilyurt,Figen S. Oktem
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注: 17 pages, 8 figures. Our code is available at this https URL
Abstract:Existing generative model-based approaches to linear inverse problems typically use the acquired noisy measurement to condition or guide sampling in the image domain. In this work, we introduce Dual Flow, a novel joint flow model for solving linear inverse problems by jointly estimating the clean image and its corresponding noise-free measurement from the observation. Specifically, Dual Flow formulates image reconstruction and measurement denoising as a coupled estimation problem, using coupled flows in the image and measurement domains. The acquired noisy measurement \mathbfy_0 is used only to initialize the measurement state; thereafter, the image and measurement states are jointly updated based on their current values. By construction, the final image and measurement estimates are guaranteed to satisfy the underlying forward model. We derive the sampling updates for the joint flow and show that they are governed by the conditional mean of the clean image given the current joint state. We further derive a closed-form expression for a sufficient statistic of the joint state and show that the required conditional mean can be obtained by solving an image-domain Gaussian denoising problem involving operator-dependent, correlated noise. We train a velocity-prediction network using this formulation to estimate the conditional mean from the derived sufficient statistic. Dual Flow requires no hyperparameter tuning beyond selecting the number of inference steps and achieves substantially shorter reconstruction times than competing methods. Extensive experiments across various linear inverse problems and image datasets demonstrate state-of-the-art performance in both distortion metrics and perceptual quality. Additional analyses assess robustness to model mismatch, as well as the ability to generate diverse plausible reconstructions and characterize reconstruction uncertainty.
[CV-143] When can multi-frag ment 2D/3D registration be trusted? Calibrated certification and uncertainty-driven view planning for pelvic fracture fluoroscopy
链接: https://arxiv.org/abs/2611.05618
作者: Bowen Xiang,Michael I. Miga
类目: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Registering a segmented CT to fluoroscopy can locate every fracture fragment but can guide surgery only when trusted. After each shot, our closed loop registers all fragments, bounds their errors with a calibrated sandwich covariance, flags implausible or ambiguous ones, and stops or plans the next view. On simulated fluoroscopy of 60 held-out fracture CTs with realistic image uncertainty, 95% bounds covered 83% of unflagged fragments and 91% of those within 5mm, but from the second view on 14% failed by over 5mm, a third with bounds of at most 2mm. A fragment’s shift when a view was added separated tightly bounded failures best (exploratory; AUC 0.94, bound 0.81). Requiring that the newest view moved no fragment beyond its bound, preferred post hoc, lowered false stops at 2mm from 23% to 6% overall, with 345 instead of 362 correct ones but more trajectories uncertified; delaying the stop did not, and other signals at equal acceptance left 10-15% false. With a segmentation error shared by all views (median Dice 0.89), 21% of this rule’s stops were false against 10% without it (15 cases; exploratory). Minimax planning certified sooner than the clinical protocol without the check (4.20 versus 4.72 shots, Holm p 0.001), but no criterion reached the pre-specified endpoint, a true 2mm error, significantly sooner, even in the true basin. On real fluoroscopy of five intact pelves no 2mm certificate was false. Trust takes evidence across views, not only a calibrated local bound.
[CV-144] Agent Count: An Agent ic Workflow for Traceable Colony Counting and Insight Generation from Agar Plate Images
链接: https://arxiv.org/abs/2611.05398
作者: Zhuosheng Liu,Austin E. Weber
类目: Quantitative Methods (q-bio.QM); Computer Vision and Pattern Recognition (cs.CV)
备注:
Abstract:Colony counting is commonly used to evaluate microbial populations in food safety testing and microbiological studies. Despite its wide application, manual counting is time-consuming and operator-dependent, and the final count often lacks a record of the objects included in the measurement. To address these limitations, this study presents AgentCount, an open-source agentic workflow for colony counting and result interpretation from agar plate images. Plate photographs and plating metadata are processed through seven agents for image quality control, plate localisation, colony detection, counting, CFU calculation, validation, and reporting. Colony detection is performed using Segment Anything Model 3 (SAM 3) with multiple text prompts and morphology-based filtering. After that, replicate statistics are calculated and a large language model (LLM)-based agent is used to generate a results narrative. Intermediate outputs, including quality scores, segmentation masks, detections, and warnings, are retained to support review of the reported values. The workflow also extends to skill-guided analysis of archived microbial population data. In a retrospective comparison with two previously published studies, we benchmarked 241 published endpoints, including all 216 mean ± SD values of the meat-analogue pathogen tables; on the 168 endpoints with a complete n=6 replicate source, 164 (97.6%) agreed within 0.01 log10 units. Overall, AgentCount demonstrates the promising potential of agentic workflows for applications in applied microbiology.
人工智能
[AI-0] BEAKER: An Expert-Curated Benchmark for Embodied Brains in Self-Driving Chemical Laboratories NEURIPS2026
链接: https://arxiv.org/abs/2611.06911
作者: Fei Lin,Tengchao Zhang,Ziyang Gong,Xiaotong Yu,Bohan Zhang,Yifan Zhou,Qihao Yang,Ji Dai,Dong Li,Yining Jiang,Qinghua Ni,Jun Huang,Qiang Zhang,Yue Zhou,Zhihang Zhong,Yonglin Tian,Fei-Yue Wang,Xue Yang
类目: Artificial Intelligence (cs.AI)
备注: Accepted by NeurIPS 2026 (Track on Evaluations and Datasets, Spotlight)
Abstract:Self-Driving Chemical Laboratories (SDCLs) are moving chemical experimentation toward embodied automation, where machines must perceive laboratory scenes, reason about experimental states, and act under physical, procedural, and safety constraints. Although Multimodal Large Language Models (MLLMs) are increasingly considered as embodied brains for such systems, existing laboratory benchmarks mainly focus on safety, anomaly detection, or general scientific understanding, leaving their execution-oriented embodied capabilities underexplored. To fill this gap, we introduce BEAKER, a Benchmark for evaluating Embodied Actionability, Knowledge, Experimentation, and Reasoning in SDCLs. BEAKER contains 1,000 expert-curated question-answer samples across 11 image and video task types, covering key judgments required for chemical laboratory execution, including affordance grounding, trajectory planning, equipment, state, and safety recognition, apparatus setup, operation and step reasoning, and temporal procedure reasoning. All questions and annotations are manually created and reviewed by chemistry and embodied Artificial Intelligence (AI) experts, without AI-assisted generation or annotation. We evaluate 28 general-purpose, science-oriented, chemistry-specialized, and embodied MLLMs in an offline, single-turn setting. Results show that current models can follow coarse experimental procedures in videos, but still struggle to ground actions into operable regions and feasible trajectories, to recognize fine-grained laboratory apparatus, states, and safety norms, and to order steps in mid-length procedures. BEAKER thus reveals a clear gap between coarse procedural understanding and the precise, action-grounded perception required for laboratory execution, providing a diagnostic benchmark for future MLLMs in SDCLs.
[AI-1] FrontierPhysics: Evaluating Agents for End-to-End Physics Research
链接: https://arxiv.org/abs/2611.06910
作者: Bingran You,Wenjun Ke,Ria Rosenauer,Bo Liu,Wenbo Chen,Min Chen,Shuqi Xu,Yue Ma,Jiechao Feng,Bo Chen,Xiangkai Sun,Jiarui Liu,Jinen Guo,Zeyu Zhou,Xu Huang,Zhuohan Li,Haoyu Li,Heguang Lin,Lijie Ding,Yadong Zeng,Kangqi Zhang,Xinge Liu,Hanwen Liu,Zhenghong Li,Alexander von Recum,Yuchen Zhu,Felix Schwarzfischer,Yizhou Zhao,Jicheng Wang,Binxu Li,Qian Yu,Chan Li,Cheng Dai,Di Wang,Haoyu Zhang,Jiankai Sun,Kaituo Feng,Mian Qin,Ruihong Yin,Yiqun Luo,Weiwei Wu,Shihong Pan,Tsung Yeh Hsieh,Xiao Huang,Yan Zhang,Sihan Li,Zirui Ren,Zengji Tu,Kyoung Whan Choe,Ke Sun,Vincenzo Galia,Vaisakhi Mishra,Jiashu He,Hongyang Wang,Zhiheng Yang,Shubham Gaur,Yifan Gu,Shuyi Wang,Zihao Zuo,Steven Dillmann,Yiyou Sun,Xinyang Han,Hanchung Lee,Molei Tao,Yue Zhao,Shilong Liu,Yuchen Wu,Xuandong Zhao,Xiangyi Li,Florian Brand,Gerbrand Ceder,Hartmut Haeffner
类目: Artificial Intelligence (cs.AI)
备注: 41 pages (10 pages main text), 19 figures, 11 tables
Abstract:AI agents are playing an increasingly important role in physics research. However, existing physics benchmarks only evaluate fixed questions, isolated sub-tasks, or paper reproduction with given solutions, and rarely check research quality alongside correctness. We introduce FrontierPhysics, an end-to-end benchmark with 56 tasks made by 42 physicists based on their own research, covering nine domains across theoretical, computational, and experimental physics. Each task requires agents to review the literature, develop a research plan, conduct research and experiments, and produce structured outputs and scientific manuscripts. Combining deterministic verifiers with expert-authored rubrics, FrontierPhysics evaluates whether agents’ research is correct and publishable. Across 22 models in 26 configurations, no configuration passes more than 46 of its 168 rollouts. Manual review of agent trajectories shows that frontier agents’ main weaknesses are insufficiently rigorous reasoning and limited expert intuition, rather than gaps in physics knowledge or execution errors.
[AI-2] Learning to Decode Linear Codes with Generalized Belief Propagation Network
链接: https://arxiv.org/abs/2611.06874
作者: Samira Malek
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Error correcting codes are a fundamental component of communication systems, enabling reliable data transmission over noisy channels. Recently, neural decoders have demonstrated notable performance gains over classical iterative algorithms. In this paper, we propose a Generalized Belief Propagation Network (GBPNet) that preserves message-passing symmetry while enhancing the decoder’s learning capability. The architecture uses cascaded variable-node updates and learned element-wise message scaling while retaining the Tanner graph structure. We also introduce two training strategies, progressive training and high-SNR training, that improve decoding accuracy. Experimental results on several linear codes show that GBPNet achieves lower bit error rates than conventional belief propagation and the evaluated state-of-the-art neural decoders.
[AI-3] On Turn-Level Best-of-N Sampling for Scaling LLM Agents
链接: https://arxiv.org/abs/2611.06868
作者: Jalaj Bhandari,Dhawal Gupta,Hongli Zhan,Zhenting Wang,Varad Pimpalkhute,Zhengzhong Liu,Eric Xing,Rupesh Srivastava
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:In this work, we evaluate a simple inference-time scaling approach based on \textitturn-level best-of-N sampling (\method) for large language model (LLM) agents on complex, long-horizon tasks. We instantiate this using a \textitgenerator-selector framework with two LLMs; at each turn, a \emphgenerator model generates N candidate actions and a \emphselector model selects one of them, which is then executed in the environment. This paper is closely related to the work of \citezainullina2025guided who used critic models for guided search at inference time. Different from their work, we evaluate if off-the-shelf, open-weight LLMs can be effective at comparing different candidate actions directly. Experiments on long-horizon terminal tasks show that \method can lead to significant performance improvements, both in terms of accuracy (5–10%) and consistency across repeated attempts. Analysis of agent trajectories reveal a shift in agent behaviour towards desirable patterns such as reduction in number of attempts to solve the task, improved task instruction following, debugging and testing.
[AI-4] CompressionBench: How can your coding agents compress LLM s?
链接: https://arxiv.org/abs/2611.06858
作者: Xinrui Zhong,Zhifeng Li,Zhiming Wang,Yongqi Liu,Shanglin Yang,Yang Zhou,Zikun Li,Zhuoming Chen,Beidi Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recursive self-improvement requires AI systems to improve the models and infrastructure used by subsequent systems. LLM compression offers a concrete efficiency objective: retain a pretrained model’s behavior within a smaller model storage budget. We study whether coding agents can carry out this step autonomously. We formulate compression as a terminal task in which an agent receives a source checkpoint, storage limit, inference environment, and time budget, then constructs compressed weights and executable inference code. The agent controls its algorithm and experiments; perplexity feedback guides search, and independent verification checks the final model. Across six agents and six model–size settings, 29 of 36 submissions pass verification. Selecting the lowest-perplexity agent design per task yields an average 3.4% perplexity increase over BF16 and 77.33% downstream accuracy, compared with 79.56% for BF16. Recorded experiments reveal how agents achieve these outcomes and where improvement breaks down. Agents lower PPL in 21 of 24 revisions at nearly fixed storage; complete-model feedback reverses 17 of 79 prior local rankings; and generation and inference tests expose failures that likelihood alone misses. The results identify representation design, experimental judgment, and capability validation as central to agent-driven model compression, a measurable component of efficiency-oriented self-improvement.
[AI-5] ServerBench: How can your coding agents design specialized LLM serving systems?
链接: https://arxiv.org/abs/2611.06841
作者: Xinrui Zhong,Zhiming Wang,Yongqi Liu,Zhaohui Wang,Shanglin Yang,Zikun Li,Zhuoming Chen,Beidi Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM) coding agents develop software through experimentation, offering an opportunity to design the serving systems supporting their own operation. Effective serving requires coordinating implementation and operating decisions under session dependencies, latency constraints, and correctness requirements. We introduce ServerBench, a benchmark for agentic design of complete, specialized LLM serving systems. Agents experiment against a common objective, dollars per dollar: latency-qualified token value per dollar of GPU cost. Stateful replay preserves session dependencies, and independent verification checks the submitted system. Across 11 coding-agent models and three deployments, Claude Sonnet 5.5 achieves 2.97 dollars per dollar on Gemma E4B, 4.87\times the vLLM reference; 22 of 33 runs yield verified systems, and 13 exceed the strongest baseline. Our analysis identifies three findings: high-value designs coordinate capacity improvements with load selection; agents interpret performance feedback through workload-relevant experiments; and reliable delivery requires verification of the execution paths introduced by optimization. Overall, ServerBench supports systematic study of how agents translate service objectives and feedback into effective, verified serving systems for given task specifications.
[AI-6] Context–Answer Metamodels: Forecasting LLM Activations Before the First Answer Token
链接: https://arxiv.org/abs/2611.06837
作者: Tommy He,Thomas Jiralerspong
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Under stochastic decoding, a single context can elicit many different answers from a language model. Characterizing this variation typically requires many expensive autoregressive rollouts, each revealing only one realization. We ask whether the activations produced while reading the context can directly forecast the model’s rollout activations. We train context–answer metamodels that map a context activation sequence to point or distribution forecasts of mean-pooled answer activations. We find that (i) the four metamodel families we study forecast mean-pooled answer activations with high fidelity, and the distribution metamodels also capture rollout variability; (ii) performance improves as the training set grows to 500,000 prompts, and the metamodels generalize zero-shot to out-of-domain datasets; (iii) their forecasts retain behavior-predictive signal; and (iv) context activations from Qwen2.5 predict mean-pooled answer activations from Qwen3.5 and Gemma 4, suggesting that activation forecasts can transfer across models. Together, these results suggest that context activation traces can support scalable forecasts of behaviors in LLM answers before the first answer token is generated.
[AI-7] Enabling AWS Cloud FPGA Development Using High-Level Vitis HLS
链接: https://arxiv.org/abs/2611.06829
作者: David Rostcheck
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:We present a toolchain for development of ML models on AWS EC2 F2 cloud FPGA infrastructure from high-level language, implementing a complete Vitis HLS to HDK bridge to AFI workflow that has been absent from the academic literature despite eight years of AWS FPGA availability. We introduce a four-level methodology (FP32 PyTorch, bit-accurate quantized Python, HLS C-simulation, physical FPGA) to aid in localizing errors, including FPGA-specific circuit errors. Together these tools allow coding agents to effectively implement FPGA designs from user directions, testing and debugging as needed. We document the deployment workflow, timing closure strategies, and practical HLS engineering decisions that proved crucial. We validate using TinyBERT fine-tuned for address Named Entity Recognition, implementing the embedding layer and the first of four encoder layers on the FPGA hardware. Our implementation streams weights from host DRAM, yielding poor performance compared to CPU/GPU implementations; we discuss forward paths.
[AI-8] Rules Traces and Routing Statistics: Training and Calibration Data for Open Models on ARC-AGI-2
链接: https://arxiv.org/abs/2611.06827
作者: Nguyen Minh Thang,Trinh Thi Phuong
类目: Artificial Intelligence (cs.AI)
备注: 39 pages, 7 figures, 23 tables
Abstract:We study which training and calibration data help three open models solve ARC-AGI-2 tasks when the solver runs offline on four L4 GPUs for at most 12 hours. We compare data offline on public tasks and report 189 runs on the 120 hidden semi-private tasks. For the NVARC Qwen3-4B grid model, we built a corpus of 12,000 executable rules in a typed language and 95,993 tasks; a rule is admitted only if its behaviour on a fixed probe set differs from that of every admitted rule, and a task only if its pairs refute a shortlist of nearest competing rules and defeat 12 shortcut solvers. Continued fine-tuning on 1,920 real task records with revised four-GPU scheduling raised the semi-private score from 28.96 (mean of the released model) to 33.06 (mean of five runs, best 34.86); no checkpoint trained with synthetic tasks, even all 95,993 of them, improved on the real-task model. For the Qwen3.8-27B reasoning model, we fine-tuned on 1,036 of its own reasoning traces with correct final answers and compared sets that add traces of other models and synthetic tasks. With an inference procedure built around the model (a thinking cap enforced in the serving engine, a Python program checked on the demonstrations after each answer, a long first attempt per input), the score rose from 26.81 to 44.03, the level the ARC Prize Foundation measured for the base model without a time limit (42.36); most of the gain came from the inference procedure. For the Qwen3.8-Flash-Next mixture-of-experts model (125B parameters, 6B active), GPU memory limited concurrency. Pruning from 512 to 384 experts per layer by router-weighted activation saliency grew the key-value cache from 0.55M to 1.39M tokens; a 296-expert list adapted with ARC routing statistics, keeping experts for program tokens, plus an FP8 indexer cache, reached 1.96M. With a tool that checks programs, the model scored 56.67, our best result.
[AI-9] ask-Efficient Agent Evaluation Using Cached Traces
链接: https://arxiv.org/abs/2611.06803
作者: Ben Johnson,Hayden Helm
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Efficient benchmarking estimates a system’s full benchmark performance from a small number of task executions. Standard approaches rely solely on the binary correctness of executions from a collection of already-evaluated reference systems. An agentic system, however, produces a rich trace for each execution that can reveal substantially more about its capabilities. Leveraging this information requires useful trace representations. We propose representing a trace by extracting a structured summary of it with an LLM and embedding the summary with a standard text encoder. These representations preserve correctness- and task-related signal while minimizing the presence of system identity-, model family- and harness-related signals that dominate raw trace embeddings. Given this representation for each task execution from all reference systems, we construct low-dimensional representations of the systems and predict a new system’s score from its position among the reference systems. On SWE-bench Verified and Terminal-Bench 2.0, augmenting an item-response model with this geometry-based prediction reduces mean absolute error from 0.13 to 0.08 (Wilcoxon p 10^-4 ) and from 0.14 to 0.12 (Wilcoxon p 10^-3 ), respectively, when provided a single random executed task. Further, given access to the full set of reference systems, our approach dominates baselines for any given number of executed tasks on both benchmarks.
[AI-10] Message-passing graph persistence
链接: https://arxiv.org/abs/2611.06801
作者: Jorge Luiz Franco,Arthur Queiroz Moura,Moacir Ponti,Amauri Souza
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注:
Abstract:Message-passing graph neural networks (GNNs) are limited in expressive power by the one-dimensional Weisfeiler-Leman test. Persistent homology (PH) provides topological descriptors that can complement GNN representations, but existing approaches typically compute persistence separately from the representation learning process. In this paper, we introduce Message-Passing Graph Persistence (MGP), a family of topological descriptors that incorporates message passing into graph filtrations. At each filtration step, node embeddings are updated using the current subgraph and associated with persistence events. These embeddings also determine the representatives of connected components during persistence computation. We prove that MGP is invariant under graph isomorphism and that it is strictly more expressive than the PH-based descriptors it extends. We evaluate MGP on an expressivity benchmark and on several graph classification and regression datasets. Experiments show that MGP outperforms other PH-based descriptors on most of these datasets.
[AI-11] ransforming Integer Grids to Spatial Intuition: Perception Not Planning Governs ARC-AGI-3
链接: https://arxiv.org/abs/2611.06793
作者: Harsidak Singh Banwait
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 24 pages, 17 figures, 8 tables. Submitted to the ARC Prize 2026 paper track. Code: this https URL
Abstract:Sample-efficient fluid intelligence requires agents that rapidly form working models of unfamiliar interactive environments. ARC-AGI-3 tests this across uninstructed visual games scored by Relative Human Action Efficiency, which penalizes exploratory waste quadratically. When frontier models read raw integer grids, serialization destroys spatial topology, and they spend hundreds of actions rediscovering basic boundaries and symmetries. We introduce SARBLOH, an open-weight agent built on the premise that spatial perception, not planning capacity, governs interactive reasoning on ARC-AGI-3. A zero-parameter O(HW) perception engine turns each frame, in 35 ms, into an object inventory with bipartite temporal tracking, a geometric briefing that separates proven invariants from affordance hypotheses, and an annotated canvas. Paired with a code-as-action runtime, the agent writes algorithmic hypotheses instead of low-level moves. Across eight presentations, grounded perception raises the rubric score from 3/28 to 25/28 points while cutting token use by 89.4%. On 23 held-out games, geometric invariants generalize with 100% accuracy, while single-frame affordance guesses fall to 6-50%, so active verification probes remain necessary. On the public set, Run A (briefing off) solved 143 of 183 levels in 8.2 hours at 59.4% single-pass efficiency; Run B (briefing on) reached 63.5%. Within the nine-hour budget on one Kaggle RTX PRO 6000 GPU, SARBLOH scored 16.53 on the hidden Kaggle evaluation, 13.7x the first Milestone winner.
[AI-12] FedFiOT: Fisher-Weighted Optimal Transport for Model Merging in One-Shot and Few-Round Federated Learning
链接: https://arxiv.org/abs/2611.06783
作者: Quan Minh Nguyen,Hoang M. Ngo,Bhakti Chowkwale,Yasaman Keshtkarjahromi,My T. Thai
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:One-shot and few-round federated learning (FL) reduce communication burden by constructing a global model within one or a few communication rounds, but often achieve lower performance than conventional multi-round FL. Without frequent synchronization, client models may drift apart, causing neurons that learn similar features to occupy different positions across clients. At the same time, local data heterogeneity can cause corresponding parameters to differ in their importance to each client’s predictions. Together, these challenges cause direct averaging to combine mismatched neurons and assign equal weight to parameters despite differences in their importance. We propose FedFiOT, a Fisher-weighted optimal transport framework for one-shot and few-round FL that leverages Fisher information to quantify parameter importance and guide neuron alignment and model aggregation. FedFiOT jointly optimizes neuron correspondences and global model parameters through alternating minimization of a shared Fisher-weighted objective. Theoretically, we show that FedFiOT’s optimization procedure monotonically decreases the layer-wise merging objective and establish a stationarity bound that quantifies the effects of client heterogeneity and nonuniform Fisher this http URL across multiple datasets demonstrate consistent improvements over competing baselines in both one-shot and few-round federated learning settings.
[AI-13] ConceptARC at Scale: Studying Conceptual Understanding in Reasoning Models
链接: https://arxiv.org/abs/2611.06732
作者: Philipp Mondorf,Adrien Deliège,Solim LeGris,Claas Beger,Ryan Yi,Barbara Plank,Brenden M. Lake,Todd M. Gureckis
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 23 pages, 14 figures
Abstract:Large reasoning models (LRMs) perform well on challenging benchmarks, yet it remains unclear whether they form generalizable abstractions across conceptually related reasoning tasks. The Abstraction and Reasoning Corpus (ARC) provides a controlled testbed for studying abstraction, requiring solvers to infer transformation rules over two-dimensional grids from a small set of examples. While ARC evaluates rule abstraction within individual tasks, different tasks may share higher-level structure despite requiring different transformations \unicodex2014 for example, several tasks may involve reasoning about objects contained within other objects. Benchmarks such as ConceptARC capture this structure by grouping ARC-style tasks into recurring spatial and semantic concepts, but they often contain only a few manually constructed tasks per concept. We introduce ConceptARC at Scale, a generator for creating thousands of ARC-style tasks targeting a specific concept. Using this generator, we construct ConceptARC-100, a dataset of 1,600 new, manually reviewed tasks, with 100 tasks for each of the 16 concepts defined by ConceptARC \unicodex2014 ten times the size of the original benchmark. We evaluate a range of recent LRMs and models trained from scratch on ConceptARC-100 to characterize which concepts they solve reliably and where they struggle. We further analyze the models’ internal representations and find clear semantic structure: task representations cluster according to the operations applied to objects, such as recoloring, removal, or translation, even when realized through substantially different rules and scenarios. Together, our results suggest that models form stable abstractions of transformation operations that generalize across diverse tasks.
[AI-14] Beyond the principle-practice gap: Responsible AI as situated epistemic and ethical practice
链接: https://arxiv.org/abs/2611.06709
作者: Thao Do Ngoc,Emma Carmel
类目: Computers and Society (cs.CY); Artificial Intelligence (cs.AI)
备注:
Abstract:Responsible AI scholarship has found that ethical principles are frequently not reflected in AI design and development. Conceptualised as a `gap’ between principle and practice, this is framed as a problem of implementation, awareness or compliance, and by implication, a deficit of systems or individual behaviour. This paper sets out an alternative explanation. Drawing on fourteen semi-structured interviews with actors designing and governing AI-enabled anti-trafficking systems across law enforcement, commercial, research, and survivor-support contexts, it argues that such apparent gaps emerge from the varied and complex institutional and professional conditions through which actors come to recognise some issues, but not others, as ethical problems requiring action. We show that epistemic choices are intertwined with ethical judgements, and trace four interconnected moments—perception, judgement, situated action and socio-technical materialisation—through which concerns become intelligible, are weighed and acted upon, and are embedded in technical artefacts and governance arrangements. Actors often hold genuine ethical commitments, but these commitments are shaped by institutional logics and professional moral economies: some concerns are materialised, while others are narrowed, deferred or left unrealised. We show that ethical agency is exercised within and between competing constraints of profession, role, prior experience, and institutional settings. The paper argues that responsible AI governance requires purposeful and well-developed epistemic infrastructures that can cut across these constraints to establish institutional conditions through which the knowledge of affected communities and potential harms becomes receivable and actionable in design.
[AI-15] Dont Replace Complement: A Detector-Agnostic Subspace Signal for Near-OOD Detection
链接: https://arxiv.org/abs/2611.06688
作者: Wonseok Shin,Songkuk Kim
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Post-hoc out-of-distribution (OOD) detection is mature on far-OOD yet remains weak on near-OOD. These two settings differ in their use of class-characteristic feature directions: far-OOD inputs largely fail to activate those of their predicted class, whereas near-OOD inputs activate them much as in-distribution samples do. Because the axis separating far-OOD leaves near-OOD almost indistinguishable from ID, a score optimized for one setting is misaligned with the other; where a detector already saturates far-OOD, added near-OOD sensitivity is bought partly against it. This divergence motivates a targeted near-OOD signal designed to complement existing scores. We propose PLUS (Per-class Low-dimensional sUbspace Signal). For each class, PLUS selects the m most class-active neurons from a sparse autoencoder, orthonormalizes the corresponding decoder directions to form a low-dimensional subspace, and scores an input by its per-class whitened squared Mahalanobis distance z within the subspace of its predicted class. Designed as a dedicated complement rather than a standalone detector, the signal z is fused additively with any calibrated base score at a single, fixed weight. This weight is a single global constant: no value is adjusted per backbone, per OOD dataset, or per deployment, and all calibration statistics are estimated from ID validation data alone. When applied to 18 existing detectors spanning diverse method families, z consistently improves near-OOD AUROC in all 126 detector–backbone combinations. Far-OOD performance is broadly maintained or improved, with only marginal declines where the base score is near-optimal, making z a robust, detector-agnostic complement across backbones.
[AI-16] LifeVal: Evaluating AI Model Performance on Personalized Agent ic Consumer Tasks
链接: https://arxiv.org/abs/2611.06687
作者: Jenny Yang,Elizabeth Proehl,Anjan Bharadwaj,Emily Fan,Charlotte Cole,James Kim,Noam Brown,Phoebe Thacker,Amelia Glaese,Tejal Patwardhan
类目: Artificial Intelligence (cs.AI)
备注: 16 pages, 6 figures
Abstract:We introduce LifeVal, a benchmark evaluating AI model performance on personalized agentic consumer tasks. LifeVal covers the most valuable life tasks that a personalized digital superassistant could automate, representing over 1.7 trillion US dollars in annual willingness-to-pay across 148 consumer task categories (e.g., managing finances, moving apartments, searching for a job, finding childcare). To construct our activity distribution, we conducted a new survey of willingness-to-pay for a digital superassistant across a demographically representative sample of over 3,800 consumers, cross-referencing our results against the American Time Use Survey (Bureau of Labor Statistics, 2025) and the Multinational Time Use Study (Fisher et al., 2025). For the most valuable consumer activities, we recruited task creators to create AI evaluation tasks personalized using real ChatGPT conversations and memories, reference files, and personal context. Task creators had to pass our expertise threshold (including personally completing the task multiple times and seeing the results through in their real lives for over a year) and also write a rubric, ensuring that our evaluation measured the best personalized completion for that task. All tasks had to pass at least three rounds of human quality-control before acceptance into our final evaluation sets.
[AI-17] Agent TrustCI: Build-Time Policy Grounding and Trajectory Tests for Java LLM Agents
链接: https://arxiv.org/abs/2611.06674
作者: Sachin Gupta
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: Accepted for publication in the 2026 International Conference on Computer and Applications (ICCA)
Abstract:Infrastructure security does not establish whether an LLM agent selected an allowed tool, grounded its answer in tool output, or followed a required workflow. AgentTrustCI is a Java/JUnit 5 framework that expresses these application-layer expectations as policy, grounding, and trajectory checks over recorded traces. A deterministic replay implementation supports repeatable build-time tests; a judge interface provides an extension point for optional semantic evaluation. On 55 synthetic cases spanning seven enterprise domains, the full pipeline correctly classifies all 30 injected faults and 25 passing controls, including ten non-trivial controls, and misses none of 12 designated critical cases. Each single-layer configuration correctly classifies 35 cases because the corpus deliberately isolates ten faults per layer. A grounded answer obtained through a disallowed sequence illustrates the additional coverage of trajectory checks. Repeated measurements quantify validator execution separately from process startup and reporting. These results establish coverage of the authored regression suite, not production accuracy. We describe a production-trace validation protocol, optional LLM-judge integration, and maintenance of expectations as workflows evolve. The contribution is a practical pattern for expressing behavioral contracts within an existing Java release pipeline.
[AI-18] Where Search Agents Go Off the Rails: Graph-Grounded Step-Level Verification
链接: https://arxiv.org/abs/2611.06662
作者: Hyeseo Jeon,Seonho Yoo,Jay-Yoon Lee
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Search-augmented reasoning systems can reach correct answers while making errors in intermediate steps. Reliable auditing must check whether evidence D covers the question Q , reasoning T addresses Q , and T is grounded in D . We introduce GraphRail, which represents requirements, evidence, and reasoning as structured claims and compares them without a generative judge or gold answer. Across three multi-hop QA benchmarks and three reasoning frameworks, GraphRail improves error recall over generative verifiers while remaining stable under stochastic extraction. Its diagnoses can also guide online reasoning, providing a traceable way to detect and correct intermediate errors.
[AI-19] From Certified Pairs to a Certified Hub: Reason er-Checked Crosswalks for the Information Exchange Standard
链接: https://arxiv.org/abs/2611.06628
作者: Fabio Rovai
类目: Artificial Intelligence (cs.AI); Databases (cs.DB); Logic in Computer Science (cs.LO)
备注:
Abstract:A national digital twin was conceived as an ecosystem of twins connected through mappings to a common foundation, but the roadmap that promises to confirm those mappings are valid never says what valid means. We give the test a precise form, coherence and conservativity of the merged ontology, borrowed from logic-based alignment repair, and apply it to the Information Exchange Standard (IES), first pair by pair and then to whole hubs. Crosswalks from IES into thirteen ontologies across seven sector groups, anchored on the building ontologies BOT, Brick, and SAREF4BLDG, pass through separate stages of matching, human review, certification of each SKOS correspondence against the two class hierarchies, and reasoning with HermiT. Pair by pair, the stages catch different failures: certification alone removed all 526 conservativity violations of the raw crosswalks, review alone left 73, and together they made eleven of twelve overlapping pairs coherent and conservative. Pairwise certification does not compose. Merging the three certified building crosswalks into one hub made five Brick classes unsatisfiable, through a homonym between the IES and Brick notions of a floor that only BOT’s disjointness axioms expose; rejecting that one row gave a coherent and conservative hub that entails 290 cross-ontology links nobody wrote down. The core hub is inconsistent before any crosswalk is added, because two published standards, the Open Energy Ontology and BFO 2020, contradict each other; with either one left out it is coherent and conservative. A hub, not a pair, is the unit an ecosystem of twins has to certify.
[AI-20] From Reconstruction to Imputation: Test-Time Optimization for Missing Time-Series Channels
链接: https://arxiv.org/abs/2611.06612
作者: Björn Ludwig,Alexander Windmann,René Heesch,Oliver Niggemann
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Complete sensor-channel failures create a difficult form of time-series missingness: local interpolation is unavailable, and deploying a separately trained imputer for every failure pattern may be impractical. We ask whether a sequence reconstructor trained without artificial mask augmentation and then frozen can be converted into an imputer at inference time. Our method treats the missing entries of each test window as optimization variables, clamps all observations, and backpropagates through the frozen model to minimize reconstruction error only on observed coordinates. The resulting observed-self-consistency objective requires neither artificial mask augmentation nor full-channel dropout during reconstructor training. We characterize when this refinement is informative through the cross-Jacobian from missing inputs to observed residuals, and use an optional temporal-smoothness prior when that coupling is weak. Across five real-world datasets with exactly one channel synthetically masked per test window and four reconstruction architectures, no-prior test-time refinement reduces MAE relative to one-pass reconstruction in 19 of 20 dataset-backbone combinations. Compared with mask-aware imputers trained under either standard structured masks or matched full-channel dropout, the best refined reconstructor achieves the lowest MAE on two datasets and trails the strongest baseline on three. These gains require substantially more inference time and are sensitive to cross-channel coupling. Our results establish test-time self-consistency as a conditional inference-time adapter for frozen reconstruction models that complements mask-aware imputation without retraining the reconstructor.
[AI-21] me-Division Multiplexing: The Operational Rhythm and Temporal Organization Layer of Digital Life
链接: https://arxiv.org/abs/2611.06603
作者: Haibin Tong,Luming Guo,Nana Cui,Jiang Yu
类目: Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC)
备注: 30 pages, 6 figures
Abstract:Continual agents must handle urgent interaction while retaining alternatives, replaying errors, and committing safe long-term changes. A previous digital-life framework distinguishes online (busy) and offline (idle) channels, but leaves open whether these are separate programs or phases of a common process and how their operations should be ordered. Motivated by recent observations of rapidly alternating perceptual representations in primate cortex, we develop a temporal organization layer with phase-tagged micro-epochs, an observable duty cycle, independent content and resource controls, and three nested clocks separating candidate exploration from durable commitment. We give a projected duty-cycle update, state its deterministic descent guarantee and the additional noise term required for a stochastic guarantee, and specify coupled busy/idle execution contracts. Five synthetic experiments (64 seeds each) examine descent and noise, objective misalignment, gain asymmetry, resource coupling, and dropped replay logs. In the gain/noise model the best tested gain ratio is 1.55; decoupled scheduling yields 26.6% versus 11.1% success on a constructed urgent/high-load subset. A frozen-backbone Qwen3-8B prototype further checks orchestration, but its proxy outcomes are not task-benchmark evidence. The results establish a testable scheduling design and identify its conditions of validity, not machine consciousness or general-purpose agent superiority.
[AI-22] One Scale Four Texts: Gaps Overlaps and Silent Repairs in Published Grading Schemes
链接: https://arxiv.org/abs/2611.06569
作者: Elisa Rovai,Fabio Rovai
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Category systems often reach their users as text, and readers apply them with tacit repairs. We ask what such a text determines once the repairs are removed, using diabetic retinopathy grading, where one international scale circulates in several documents. We transcribe four texts (the English screening definitions of 2025, the international scale as tabulated in 2017 and as published in 2003, and an Italian guideline of 2015) into one rule language over 24 finding variables, and make every point a text leaves open a named precisification. Exhaustive enumeration of 1,425,807,360 finding profiles, re-derived by an SMT solver and spot-checked by an OWL reasoner, shows that every text that assigns actions leaves some eyes without a category under every reading; that the 2017 tabulation has a proliferative class that is empty when read literally and a macular gap that the 2003 text lacks, so that no function maps the 2003 grades onto the 2017 grades; and that on image findings alone the English and Italian texts order some pairs of eyes in opposite ways. On 927 annotated photographs from three public datasets, with retinal thickening assumed absent, the English text and the 2017 tabulation disagree on referral for 26.8% of the eyes of one dataset (DDR) on which both decide, and in a pre-registered experiment classifiers trained on each text’s labels side with their own text. Thickening cannot be seen on a colour photograph; left free, every disagreement involving England becomes undecidable, so it rests on what the camera cannot show. In a second pre-registered study, language models that answered without deliberating repaired most gap and overlap items even when told to read literally, while Claude Sonnet 5.5 and Claude Opus 5.5 reported the defects far more often; within two models, deliberation reduced silent repair without removing it.
[AI-23] RAG AM-AI: A Hybrid Reinforcement Learning based Agent ic approach powered by Generative AI for Edge Resource Management
链接: https://arxiv.org/abs/2611.06554
作者: Shreya K. Chari,John S. Vardakas,Kostas Ramantas,Christos Verikoukis
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Resource management for distributed edge networks hosting different services has been handled efficiently using Reinforcement learning (RL). Reward generalization in complex workflows is still an open issue within the RL realm. Therefore, this work proposes RAGAM-AI, a hybrid framework that combines the capabilities of RL algorithms and Generative-AI (Gen-AI) for powering an Agentic AI workflow that optimizes resource allocation at the Multi-access Edge Computing (MEC) units. This framework further resolves the reward generalization issue by utilizing Large Language Model (LLM) for feedback. A Multi- Agent Soft-Actor Critic (MASAC) using Math\Sigmatral as a reward oracle is implemented for managing resource allocations at the MEC. RAGAM-AI interprets system state information and evaluates multi-objective trade-offs guiding the resource orchestration loop over distributed edge server agents. The complexity of the framework is further brought down by incorporating surrogate reward model. The overall performance is evaluated against the single agent SAC, REINFORCE and random policies for benchmarking. On using RAGAM-AI, lower allocation deviations for CPU, memory and bandwidth resources are reported while minimizing the latency and data rate costs. A scalable AI as a feedback (AIAF) solution optimizing network resources with significant performance improvement is thus presented.
[AI-24] Addition by Subtraction: Toward Mitigating Frag ile Safety Alignment in Large Language Models
链接: https://arxiv.org/abs/2611.06553
作者: Hamid Dashtbani,Alexandra Dmitrienko
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注:
Abstract:Large language models (LLMs) undergo safety alignment to prevent harmful generations, yet they remain vulnerable to jailbreak attacks. Recent work reveals a deeper structural weakness: safety alignment can be highly sparse, such that pruning a small number of safety-critical neurons or ablating a single refusal direction can substantially undermine aligned behavior. In this work, we take the first steps toward strengthening safety alignment such that it is less dependent on sparse model components. We propose Addition by Subtraction, which gradually removes identified safety-critical neurons during finetuning and uses preference optimization to re-establish safety-aligned behavior through other parts of the model. Our results show that the resulting models are substantially more robust to pruning-based attacks. Remarkably, these benefits generalize beyond the white-box attacks motivating our design: Addition by Subtraction improves robustness against other sparsity-based interventions, including refusal-direction ablation, as well as a broad range of black-box jailbreak attacks. The finetuned models also exhibit improved truthfulness while largely preserving general capabilities. Comprehensive experiments across four models from three vendors demonstrate consistent improvements in safety robustness, reducing attack success rates by up to 90% for some attacks.
[AI-25] Using Experience Twice: Dynamic Process Rubrics for Privileged Token-Level Credit Assignment
链接: https://arxiv.org/abs/2611.06543
作者: Shumao Wu,Kongcheng Zhang,Jianqiang Huang,Baisheng Lai
类目: Artificial Intelligence (cs.AI)
备注: 18 pages, 3 figures
Abstract:Agent trajectories often contain both successful decisions and failed attempts, yet optimization with sparse trajectory-level rewards struggles to identify which critical decisions should be reinforced or suppressed. We observe that meaningful distinctions among on-policy trajectories can be distilled into natural-language process rubrics and used as process supervision signals. Based on this insight, we propose Online Rubric Self-Distillation (ORSD), which leverages self-derived process rubrics for dense token-level supervision, thereby augmenting outcome-centric reinforcement learning. Specifically, the current policy independently generates multiple rubric bundles at each step according to the same group of trajectories, extracting candidate process rubrics from cross-trajectory behavioral differences. A frozen verifier then evaluates all rubric items against each trajectory in the group, and selects one candidate bundle based on bundle-level alignment between rubric satisfaction and task success. ORSD treats the selected rubric bundle as privileged information during training, then compares the log-likelihoods of the original tokens with and without the rubric. The resulting likelihood gap provides token-specific process credit, which is added to the original trajectory-level GRPO advantage through additive advantage shaping. ORSD thus bridges trajectory-level outcome supervision and fine-grained process supervision, enabling token-level credit assignment using process signals. Extensive experiments on ALFWorld and WebShop show that ORSD accelerates policy improvement and boosts final task performance for both Qwen3-1.7B and Qwen3-4B under the same environment-interaction budget. Our code is available at this https URL.
[AI-26] From Admission to Retention: Predicting Durable Offspring in Quality-Diversity Search NEURIPS2026
链接: https://arxiv.org/abs/2611.06522
作者: Corrado Calamita,Matteo Mendicini,Luca Ferrari,Riccardo Bertamini
类目: Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI)
备注: Accepted at the NeurIPS 2026 Workshop on Self-Evolving Diversity-Driven Search for Robust AI Systems (EvoRobust)
Abstract:Automated failure discovery can be framed as a diversity problem: the objective is to expose distinct ways a system fails rather than a single critical adversarial case. Yet uniform parent sampling provides no way to allocate search effort toward individuals likely to produce durable archive contributions. We study this credit assignment problem in an unstructured Quality-Diversity (QD) archive, testing whether signals available before an individual has reproduced distinguish offspring that merely enter the archive from those still retained generations later. Across a planar-arm benchmark and two robotic-locomotion tasks, the same signals are often associated with entry and persistence in opposite directions. In particular, lineage depth is the clearest such signal, and its persistence discrimination grows with population capacity, which is associated with lower turnover. Biasing parent selection toward the signals with the strongest measured association to survival nevertheless does not improve search: moderate bias matches uniform sampling, and strong bias reduces QD-score.
[AI-27] IDSISA: Influence-Distilled Sharded Training for Efficient Machine Unlearning
链接: https://arxiv.org/abs/2611.06519
作者: Deeraj SK,Sadhana Devarajan,Tanmoy Hazra,Anupam Shukla
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
Abstract:Machine unlearning removes the influence of specific training examples to comply with privacy regulations. Full retraining gives strong guarantees but is computationally prohibitive at scale; approximate methods are efficient but lack formal guarantees. SISA (Sharded, Isolated, Sliced, and Aggregated training) achieves certified deletion through data sharding and independent retraining, but random partitioning scatters forget requests across most shards, and isolated retraining degrades utility. We introduce IDSISA (Influence-Distilled SISA), which uses influence functions to concentrate forget-prone samples into few shards and distillation from unaffected shards to recover utility lost to isolation. The deletion guarantee inherited from SISA’s retraining step remains exact and requires no probabilistic qualification; the distillation-based utility recovery is a separately quantified, empirically bounded approximation ( \epsilon_\textcert \leq 0.015 in expectation). We further formalize certification under sequences of repeated deletion requests, where cross-shard distillation can create dependencies between shards; we prove that a closure-based cascading retraining procedure restores the deletion guarantee to the full aggregated model under any deletion sequence. Experiments on ResNet-18/CIFAR-10, ResNet-18/CIFAR-100, BiLSTM/IMDB, TCN/IMDB, and a fully connected network on Fashion-MNIST show 14-91% better selective forgetting, up to 50% better utility retention in relative terms (with consistent percentage-point gains on the other architectures, where a relative baseline is undefined), and unlearning-time changes ranging from -6.7% to +6.4% depending on architecture, relative to SISA.
[AI-28] ElecLoad: Adaptive Electricity Load Forecasting Guided by Multiple Events
链接: https://arxiv.org/abs/2611.06508
作者: Huajian Zhang,Mingyue Cheng,Xingpeng Gao,Xiaoyu Tao,Zhiding Liu,Jiqian Yang,Zhi Li
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Accurate electricity load forecasting provides critical decision support for energy procurement, generation scheduling, and supply-demand balancing, contributing to the economic efficiency and operational reliability of power systems. However, future electricity demand can be affected by the complex interplay of multiple events described in Web notices, weather alerts, and event schedules. Their effects vary over time and can reinforce or offset one another, making electricity load difficult to forecast accurately. While existing methods can incorporate context as additional input, how to use the effects of multiple events and their interactions to guide the load forecasting process itself remains to be explored. To address this gap, we propose ElecLoad, a context-aware agent for multi-event electricity load forecasting that adaptively organizes the forecasting process. Specifically, we train ElecLoad to combine historical load with context to analyze each event’s relevance to the target load, the timing of its effects, and potential interactions among events. Guided by this analysis, ElecLoad adaptively chooses to forecast the entire forecast horizon directly or to partition it into segments according to changes in event effects. When the horizon is partitioned, sub-agents independently generate initial forecasts for each segment. The main agent then reviews these forecasts, and the sub-agents perform selective refinement guided by its feedback. Experiments on multiple electricity load datasets show that ElecLoad achieves state-of-the-art performance in most evaluation scenarios. Code is available at this https URL.
[AI-29] ools Need Not Speak in Text: Latent Context Compression for Agents
链接: https://arxiv.org/abs/2611.06500
作者: Yuetong Liu,Zhehao Zhou,Yitong Zhang,Haorui Chen,Jia Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Large language model (LLM) agents typically append raw tool outputs to their context. Over multi-step interactions, long and noisy observations incur substantial context overhead and dilute the agent’s attention to task-relevant information. Existing token pruning and summarization methods remain constrained by discrete text representations, limiting the achievable compression ratios. In this work, we challenge the prevailing assumption that agents must consume tool observations as explicit text tokens. To this end, we propose Coca-Cola, a continuous latent context compression model for LLM agents. Coca-Cola compresses tool observations into latent tokens and embeds them into the input embedding space of the agent. We further introduce Action-Preserving Training, which aligns the agent’s next-action distributions under full and compressed observations. This trains the compression model to learn action-sufficient latent tokens that retain decision-critical information. Experiments with web-search agents on four QA benchmarks and two model families show that Coca-Cola reduces the agent’s token consumption by 70.01%-89.24% and end-to-end runtime by 24.45%-55.97%, while consistently improving task performance. These results suggest that Coca-Cola offers a promising approach for context compression in agents. Our code is available at this https URL. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.06500 [cs.AI] (or arXiv:2611.06500v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.06500 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-30] FPTR: Feasibility-Preserving Transactional Refinement for Deadline-Constrained Joint Beam and Resource Allocation
链接: https://arxiv.org/abs/2611.06491
作者: Minghao Kong,Zirong Liu,Jianhao Wu,Muyao Chen,Jianxin Wu,Rongjie Wang
类目: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR); Performance (cs.PF)
备注: 14 pages, 4 figures; includes appendix. Code: this https URL
Abstract:Short-interval wireless schedulers must return a complete, feasible beam-and-resource allocation before each execution deadline, and a late or invalid output gives no utility in the current control cycle. Searching longer can increase transmitted traffic, but it may also miss the deadline or leave only an infeasible intermediate solution. We propose Feasibility-Preserving Transactional Refinement (FPTR), which always keeps a validated feasible allocation, the incumbent. Each refinement stage builds a separate candidate and leaves the incumbent unchanged. A candidate replaces the incumbent only if it is complete, passes the feasibility validator before its stage cutoff, and strictly improves the objective; otherwise it is discarded. Four refinement stages use this rule: buffer-aware resource selection, compatibility-group (CG) user sharing, beam-mask rebuilding from remaining demand, and joint reassignment of two resources. On 150 fixed synthetic instances, the full pipeline (Full) raises traffic by 3.83% on average over a lightweight beam-first reference and by 1.79% over its Base constructor. In a separate suite with matched inputs, Full exceeds six independent search baselines (ALNS, tabu search, GA, simulated annealing, iterated local search, and GRASP) by 6.68-8.35%, where outputs later than 100 ms receive zero utility. All 750 Full runs in the internal suite returned feasible allocations, and none exceeded 100 ms. Separating mandatory output construction from optional refinement lets spare computation time improve the allocation without endangering the incumbent. The open-source implementation supports independent reproduction.
[AI-31] Model Merging Does Not Require Uniform LoRA Ranks
链接: https://arxiv.org/abs/2611.06486
作者: Zirong Xie,Zainab Afolabi,Nikita Kozodoi,Jack Butler
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Model merging combines separately fine-tuned experts into a single model without retraining. Standard practice trains every expert at the same LoRA rank, which couples the rank choice across experts even though each domain has its own best rank. We show this coupling is a property of the merge method, not of merging. We train 36 experts on three domains (mathematics, instruction following, multilingual) at four ranks and three model scales, then merge with six methods at uniform ranks and at each domain’s own best rank. Giving experts different ranks costs averaging and utility-based selection nothing, and gains them up to 1.8 pp of multi-domain accuracy, while the conflict-resolution methods are inconsistent and can lose up to 4 pp. The split has a simple cause. Merging keeps only a fixed share of the pooled spectral components, and methods that choose them by size discard many that would have reduced loss. They also tend to discard the same expert: the instruction adapter has the smallest singular values, so it is dropped almost entirely and IFEval falls by 16 to 25 pp. Choosing components by their effect on loss instead keeps all three experts, and leaves accuracy nearly flat across an 8 \times rank range (0.7 pp, against 6.7 pp for TIES at 4B scale). Rank can therefore be chosen per expert whenever the merge operator does not select by magnitude.
[AI-32] AmbiHTMLBench: A Benchmark for HTML-based Visually Rich Document parsing
链接: https://arxiv.org/abs/2611.06449
作者: Boyu Niu,Yingjie Xiong,Shurui Li,Linhan Pang,Yuntong Zhang,Han Wang,Hao Wu,Kai Zuo,Xu Tang,Xuanhe Zhou,Fan Wu
类目: Artificial Intelligence (cs.AI)
备注: Check latest benchmark info on this https URL
Abstract:Traditional document parsing systems use Markdown as their output representation, which is effective for linear text but does not fully preserve fine-grained spatial structure, table geometry, visual hierarchy, or rendering semantics. We show that HTML is advantageous for heterogeneous and layout-sensitive documents whose representations must remain editable, inspectable, and reusable, and introduce AmbiHTMLBench to systematically study this setting. In this benchmark, given a real-world document, a model must recover executable HTML that preserves its content, organization, and visual appearance. The benchmark contains 1,115 diverse pages with structured annotations and evaluates predictions in two different ways: The objective protocol aligns ground-truth and predicted rendered blocks and measures table fidelity, page-level text error, and a two-dimensional layout consistency; a complementary reference-anchored protocol measures perceived fidelity by aligning LLM judges with human preferences. Across ten primary multimodal systems, the best objective score is 0.8092, while the subjective winner is different, demonstrating that automatic decomposition and human-perceived fidelity capture complementary properties. We further fine-tune a small-parameter expert model and find that task-specific adaptation yields good performance on document-to-HTML reconstruction task. These results position AmbiHTMLBench as a diagnostic testbed for building smarter and more logical document parsing systems.
[AI-33] SUMONE: A Simple and Efficient Summarizer for Plug-and-Play Agent Memory
链接: https://arxiv.org/abs/2611.06445
作者: Jizhe Zhang,Ang Li,Chenheng Zhang,Yisen Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Long-horizon language-model agents accumulate increasingly large interaction histories, making context compaction essential for efficient execution. Existing learned compaction methods often rely on downstream task outcomes or solver behavior, coupling training to specific tasks or agents. We introduce SumOne, a standalone summarizer trained to preserve information in agent histories through masked reconstruction. Given a summary of the expired history, a frozen reconstructor recovers randomly masked portions of the original trajectory; reconstruction quality, combined with a length penalty, provides a task-agnostic reward for optimizing the summarizer with GRPO. This yields self-supervision directly from raw trajectories, without reference summaries or downstream task execution. At inference time, only the lightweight summarizer is retained, enabling plug-and-play context compression for otherwise unchanged agents. We conduct a systematic evaluation of SumOne in long-horizon search and software-engineering tasks. SumOne reduces estimated inference cost by approximately 78% on DeepSearchQA and 75% on SWE-bench Verified relative to operating with full context, with absolute decreases of 0.04. The same checkpoint also transfers across downstream models and agent harnesses without additional fine-tuning.
[AI-34] EviSeek: An Evidence-Centric Agent ic Framework for Embodied Active Information Seeking
链接: https://arxiv.org/abs/2611.06434
作者: Jintao Li,Zhiyong Cao,Baihua Xiao,Yu Bai,Börje F. Karlsson
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Active embodied question answering requires agents to acquire evidence through viewpoint changes and physical interaction. The central challenge is to avoid undirected interaction: taking plausible actions without specifying what evidence they should reveal for answering the question. Action plans and accumulated histories alone can leave the informational purpose of an interaction implicit. We introduce EviSeek, a training-free framework that makes question-derived evidence demands explicit acquisition objectives, linking observation-conditioned actions and selectively retained evidence to the information the agent seeks. On 3,436 ESI-Bench questions, EviSeek improves accuracy over the Single-model agent by 7.1 percentage points with Qwen3-VL-32B and 3.7 points with Gemini 3.1 Pro, while improving visual-support-weighted accuracy and reducing mean token usage. These findings support making the informational purpose of interaction explicit, with action selection and evidence retention organized around what the agent needs to learn to answer the question.
[AI-35] D3D: Decoupled Diffusion Drafting for Asynchronous Speculative Decoding
链接: https://arxiv.org/abs/2611.06425
作者: Baoyou Chen,Ruiqiao Mei,Guancong Lin,Tao Song,Weihao Yuan,Qun Liu,Siyu Zhu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Target-conditioned diffusion drafters predict candidate tokens in parallel, yet each new block must wait for updated target-model hidden states. This forces drafting and verification to run sequentially, limiting speedup even though each block is generated in one pass. We introduce D3D, a standalone diffusion drafter that removes this inter-round dependency. D3D combines blockwise denoising with auxiliary autoregressive supervision to achieve long accepted prefixes without target hidden states. This independence lets D3D draft successor blocks during verification in one packed pass, enabling asynchronous speculative decoding. We progressively prune layers according to their effect on acceptance length to reduce the cost of packed drafting. Removing five of 36 layers retains over 99% of mean acceptance length across three benchmarks with Qwen3-8B as the target. Across seven benchmarks, D3D averages 5.00 \times and 5.71 \times decoding speedup with Qwen3-8B and Qwen3-32B targets, compared with 4.03 \times and 3.67 \times for DFlash. On GSM8K with Qwen3-32B, speedup reaches 7.48 \times . Code is publicly available at this https URL.
[AI-36] Zero-Shot Coordination among LLM Agents ICLR26 NEURIPS2026
链接: https://arxiv.org/abs/2611.06409
作者: Adrian Hayler,Shashank Reddy Chirra,Andrei Lupu,Johannes Forkel,Bidipta Sarkar,Siheng Feng,Jakob Foerster
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: NeurIPS 2026 Poster. ICLR 26 MALGAI Workshop Oral. The first two authors contributed equally
Abstract:We study zero-shot coordination (ZSC), where independently developed agents must coordinate at test time. While ZSC has been well studied in the RL literature, far less is known about the performance of LLM agents despite their increasing deployment in such settings. Existing work on LLM coordination often relies on specialised scaffolding that independently developed agents are unlikely to share in practice. Moreover, evaluations on complex environments (e.g., Hanabi) make it difficult to pinpoint the sources of coordination failure. In contrast, we focus on simple, general-purpose scaffolds in minimal environments designed to isolate specific coordination challenges. Our results show that even in these controlled settings, frontier LLM agents struggle to coordinate, largely due to limited understanding of the coordination problem and weak reasoning about their partner’s beliefs. While LLMs introduce semantic information as an additional axis for coordination, they nonetheless fail to exploit this structure effectively. Towards this, we propose Coordination-friendly definitions (CFDs) as a principled approach for enabling robust coordination among LLM agents. Finally, we show that CFDs can be discovered automatically, removing the need for manual engineering. Code is available at this https URL.
[AI-37] GIST: Gibbs-Targeted Hierarchical Data Selection for Offline RLVR
链接: https://arxiv.org/abs/2611.06395
作者: Xuliang Huang,Tianxiang Zhao,Chengwei Qin,Yao Shu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Offline reinforcement learning with verifiable rewards (RLVR) trains on a fixed pool of sampled responses and typically supervises every token of every response, although responses and tokens are not equally informative. We study which ones to supervise through the Gibbs-optimal policy of KL-regularized RLVR, which offline training targets, and measure the mismatch between a policy and this target with a bidirectional KL divergence. At the reference policy, this mismatch factorizes exactly and separates the two outcomes: correct responses contribute mainly through the forward term, concentrated on their low-probability tokens, and incorrect responses through the reverse term, concentrated on the responses the policy generates often and on their high-probability tokens. Based on this analysis, we propose GIST (Gibbs-Targeted Hierarchical Data Selection), a two-level selection method that approximates these principles with signals available before training. GIST uses the sampling temperature to select correct responses on-policy and incorrect responses near the policy’s mode, and it uses the initial policy’s token probabilities to supervise only the tokens where the mismatch lies. Applied to two offline training objectives, GIST raises the average accuracy over GSM8K, MATH-500, and MBPP for all three models under both objectives. In a controlled comparison on Qwen3-0.6B/GSM8K, it outperforms seven selection baselines that choose from the same candidates while supervising 66.6% fewer tokens than training on all candidates of the same prompts.
[AI-38] MA-Mesh: Resilient Multi-Agent Collaboration via Role-based Self-Healing EMNLP2026
链接: https://arxiv.org/abs/2611.06389
作者: JunYong Lee,Woo Kyung Kim,Honguk Woo
类目: Artificial Intelligence (cs.AI)
备注: Accepted at EMNLP 2026 (Main)
Abstract:Effective collaboration in multi-agent systems requires mechanisms for coordination and adaptability in dynamic and uncertain environments. However, existing multi-agent systems rarely model resilience explicitly, instead relying on ad-hoc output correction that limits systematic, runtime management of robustness. In this work, we propose Multi-Agent Mesh (MA-Mesh), a framework that constructs a role-based self-healing layer on top of a collaboration pattern. Specifically, MA-Mesh generates resilience-aware system specifications and synthesizes them into executable multi-agent policy code. To enhance resilience, we introduce two self-healing mechanisms: Cross-agent reflection, a mechanism that allows agents to evaluate not only their own outcomes but also the performance of their peers, enabling informed behavioral adjustments and collective policy refinement; Role-based failover, a strategy which dynamically reassigns roles and enables agents to substitute for one another when sub-tasks fail or agents underperform, ensuring continuity of task execution. Experimental results across diverse tasks (e.g., software development, tool use, scientific reasoning) demonstrate that MA-Mesh significantly outperforms existing multi-agent systems in both task performance and robustness under system uncertainty.
[AI-39] From hydrological laws to catchment models: A process-based model matches deep-learning accuracy on three continents
链接: https://arxiv.org/abs/2611.06382
作者: Xilin Xia
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Dynamical Systems (math.DS); Geophysics (physics.geo-ph)
备注: 61 pages, 12 figures, 13 tables, plus 19 pages of Supporting Information
Abstract:Dooge asked whether hydrology has laws at the catchment scale. We ask whether a low-dimensional system built under physical laws can match the accuracy of neural networks that outperform process-based models on the same inputs. Three laws constrain a lumped catchment model. Conservation makes it a flux-connected storage network, dissipation fixes the sign of every flux and aggregation determines how the unresolved composition enters the fluxes, through effective laws over threshold distributions and a memory of the forcing. These laws admit four kinds of refinement. A parsimonious model under them, CaRT (11 parameters per catchment), is refined in a loop assisted by AI agents. Each refinement addresses an observed deficit through a hypothesis from the literature. The result, CaRT+, has fourteen states, 35 parameters per catchment and no daily neural network. Most refinements follow from aggregation over thresholds or from variables that carry no water but modulate fluxes. On 602 British and 593 US gauged catchments, single CaRT+ models reach median Nash-Sutcliffe efficiencies of 0.863 and 0.765 in held-out years. A stateful LSTM reaches 0.869 and 0.751 and an entity-aware LSTM 0.862 and 0.754, while three-seed LSTM ensembles remain ahead of the CaRT+ ensemble by 0.003-0.012 in median paired NSE. On 505 drier Australian gauged catchments unused in development, single CaRT+ models reach 0.778 against 0.762 and 0.753 for the two LSTMs. The results suggest that catchment-scale laws exist as laws of form, with the constitutive families shared and only their parameters learned.
[AI-40] Learning to Tune with Biased Gradients: Iterative Feedback Tuning of Model-Free Control on Throttle Valves
链接: https://arxiv.org/abs/2611.06338
作者: Félix Martinez-Jiménez(GIPSA-INFINITY),Sander Miller-Murphy(GIPSA-INFINITY),Tobie Ondoa-Essomba(GIPSA-INFINITY),Emmanuel Witrant(GIPSA-INFINITY)
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Iterative feedback tuning (IFT) learns the parameters of a fixed-structure controller by stochastic gradient descent. At each iteration, two closed-loop experiments are run on the plant: the first one measures the tracking error, and the second one, driven by this error, provides the sensitivities from which the gradient of the cost is estimated. This estimate is unbiased for linear plants only. We study the biased case on automotive electronic throttle valves, whose friction and stiction vary between valves and runs, controlled by an intelligent proportional (model-free) controller, which is an incremental PI controller. We show that the mismatch of the controller state at the start of the two experiments, which stiction makes unavoidable, adds an error inversely proportional to the amplitude of the second experiment. On digital twins of ten valves, the IFT gradient is a descent direction in only 69% of the low-cost region, friction being the cause, and all update rules plateau and then drift. On two hardware valves, the best update rule depends on the noise structure of the gradient. On the first valve, where one gradient component is buried in noise, Adam outperforms Gauss–Newton, its normalisation and its momentum each improving one parameter. On the second valve, where both components are consistent, the same normalisation amplifies a direction that Gauss–Newton rightly discounts, and Adam drifts to a controller worse than the initial one. The twins reproduce the mechanism but not this ranking.
[AI-41] EnzymeRSI: A dual-loop recursive self-improvement system for enzyme design
链接: https://arxiv.org/abs/2611.06321
作者: Genoria AI Team,Yuejie Hou,Lixiang Yang,Yisi Liu,Weiting Tang,Yueyuxiao Yang,Meng Yang
类目: Artificial Intelligence (cs.AI)
备注: 28 pages, 8 figures, 7 tables
Abstract:Improvement in enzyme engineering is usually pursued by optimising which candidates to assay within a fixed model and workflow; when the bottleneck arises from a mismatched method or a failing workflow, that route is difficult to escape. We present EnzymeRSI, a dual-loop system for recursive self-improvement (RSI). The outer loop evolves harness assets without changing model weights, the inner loop develops prediction and selection algorithms under controlled conditions, and a cloud laboratory system is specified to close the physical loop with quality-controlled feedback. By updating its prediction strategy through iterative comparison, the outer loop raised the mean development-set Spearman correlation on ProteinGym by 0.044 across five assays, reaching 0.335 on CD19 and 0.683 on OPSD (rhodopsin); of the two first places, only CD19 was produced by the outer loop itself. In the self-improvement of the ThermoMPNN algorithm the inner loop reached a PCC of 0.849, and on TadA-Bench it reached a Spearman correlation of 0.307 in 1.13 h. In three prospective wet-lab tests of the system’s experimental decision-making, 10 of the 32 candidates that had not been pre-specified for the luciferase RLuc showed at least twice the wild-type activity; a recommended variant of the cytidine deaminase A3A raised 5mC conversion to 90.2%; and path selection in evolution raised the catalytic efficiency of the DNA polymerase KOD-E on a modified substrate to 31.1 times that of the parent. By treating the organisation of the research workflow and the development of algorithms as a single falsifiable dry-to-wet loop, this work offers an evolutionary paradigm for moving beyond the blind spots of a fixed method and for autonomous enzyme engineering.
[AI-42] Conformal Prediction for Distribution-to-Distribution Regression
链接: https://arxiv.org/abs/2611.06320
作者: Trung-Khang Tran,Tuan A. Hoang,Viet-Hoang Tran,Tan M. Nguyen
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Conformal Prediction (CP) provides a model-agnostic framework for constructing finite-sample prediction sets with coverage guarantees and has become a tool for reliable decision-making in high-risk settings. Recent advances have extended Split Conformal Prediction to increasingly complex learning problems, including multilabel, high-dimensional, and functional outputs. In this work, we study CP for Distribution-to-Distribution Regression, a setting of growing interest across many application domains. We show that a naive application of Split CP yields prediction sets that, while marginally valid, are non-interpretable, e.g., lacking a well-defined volume or a tractable sampling procedure, and of limited practical value. To overcome this, we propose a new conformal framework that operates in an orthogonal basis space, producing interpretable prediction sets while preserving coverage guarantees. We further develop an adaptive variant with asymptotic conditional coverage under mild assumptions. Empirical results on synthetic and real-world datasets validate the effectiveness of the proposed approach. The code is publicly available at this https URL.
[AI-43] When Does Risk-Controlled Throughput Forecasting Actually Control Risk? A Reproduction Study on LEO Satellite Link
链接: https://arxiv.org/abs/2611.06307
作者: Mayan Sharma,Kriti Saini,Devansh Behl,Suganya R,S. Vijayaragavan
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Networking and Internet Architecture (cs.NI)
备注: Code: this https URL
Abstract:Risk-budgeted throughput forecasting converts a point forecast of LEO satellite capacity into a conservative bound whose overestimation rate is held at a user-specified budget, so that admission control can act on it safely. We reproduce the state of the art, BG-CFQS, on the three Starlink traces its authors used and on an independent fourth dataset, and report four findings. First, its risk guarantee is reproducible but conditional on the calibration split: under a random split we recover every published quantity, including risk-pass on 3 of 3 datasets, but under a contiguous temporal split, which is what a deployed terminal faces, it holds on 1 of 3. Second, independent of any split, the method cannot serve a risk budget below 0.15, the low end of its candidate quantile set. On all four datasets the achieved rate is identical at budgets of 0.05, 0.10 and 0.15, overshooting by 1.4x to 3.7x at the tightest. Third, static conformal calibration cannot improve a ranking, only shift an operating point; we confirm this exactly on latency spike detection, where online calibration improves AUPRC by up to 51%. Fourth, group-conditional online recalibration splits into two mechanisms. The online part replicates: it tracks the budget within 1% from 0.05 to 0.35, where static calibration is 16 to 21% off. The group-conditional part does not: on three of four datasets, conditioning on any available covariate, including measured serving-satellite geometry, is worse than not conditioning, and our attempt to predict which datasets benefit failed. What is new here is empirical, not algorithmic: we claim the findings, not the methods.
[AI-44] An Engineering Reading of the Bounded Halting Problem and Observable Properties of a Universal Machine
链接: https://arxiv.org/abs/2611.06303
作者: Sergey Salishev
类目: Computational Complexity (cs.CC); Artificial Intelligence (cs.AI); Formal Languages and Automata Theory (cs.FL); Logic (math.LO)
备注: 9 pages, 1 table
Abstract:The halting problem asks whether a program eventually halts, with unrestricted resources. Engineering systems instead certify bounded halting: given a program, an input, and a step budget T, decide whether the program halts within T steps. This paper studies two of its uses. First, it defines an exact finite-budget problem, placing the simulated program and the checker on the same unit-cost timescale, so the checker can be embedded in an adversarial program. A diagonal argument with explicit additive overhead shows that no fixed deterministic checker decides every instance within any fixed fraction of T below one, for large T. In the nondeterministic case, T bounds the branch depth allowed for a stable observable decision. Second, bounded-halting instances give adversarial tests for finitary universality, since a universal model must support arbitrary finite computations as T grows. The tests check resource growth with T, the absence of a fixed deliberation bound, and later access to intermediate data. They separate a fixed model with extendable external resources, a model family that grows with T, and a larger system built around a fixed component. For machine-learning systems, this identifies the resource growth a universality claim assumes and whether it fails for deployed models.
[AI-45] STEK: Structural-Terrain-Enhanced Koopman-Inspired Cross-Channel Regularization for Retrospective Battery Multi-Target Estimation
链接: https://arxiv.org/abs/2611.06299
作者: Jing Wang,Shiyuan Piao,Fugee Tsung
类目: Artificial Intelligence (cs.AI)
备注: 18 pages, 4figures
Abstract:Retrospective battery characterization across heterogeneous cells is challenging because cycling records vary substantially with chemistry, protocol, cell format, capacity, and operating conditions. An endpoint charge–discharge curve captures the local condition of an evaluated cycle, while a cell-level lifecycle terrain summarizes structural evolution across the archived trajectory. We propose STEK, a structural-terrain-enhanced framework for retrospective battery multi-target estimation. STEK combines a cell-level lifecycle (dQ/dV) terrain with an endpoint-derived structural proxy through quality-proxy-conditioned gating. Bounded FiLM and a validity-gated residual head refine the endpoint-curve representation while preserving a curve-based prediction pathway. Battery-type and optional material-text representations further condition an identity-centered low-rank linear map, providing a Koopman-inspired cross-channel consistency objective without modeling temporal dynamics or performing rollout. Under the fixed seven-output MIX-battery protocol, the full retrospective setting improves aggregate performance over endpoint-only learned baselines across multiple temporal backbones; this comparison reflects both the additional archived context and the modeling design. Under matched auxiliary inputs, STEK outperforms direct concatenation and quality-unaware FiLM, while descriptor-removal tests show graceful fallback to the endpoint pathway. Deterministic rules remain preferable for directly recoverable capacity summaries. The method targets retrospective archive harmonization rather than online estimation from cycle t alone.
[AI-46] Multi-Resident Data Association in Smart Homes via MDP Modeling and Behavior Cloning
链接: https://arxiv.org/abs/2611.06296
作者: Xi Chen(M-PSI),Julien Cumin,Fano Ramparany,Dominique Vaufreydaz(LIG, M-PSI)
类目: Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
备注:
Abstract:Ambient sensors offer a privacy-preserving solution for activity recognition in smart homes, yet they face a critical challenge in multi-resident environments known as Data Association, which aims to identify the resident triggering a specific anonymous sensor event. Existing approaches fundamentally fail to maintain global consistency by often yielding conflicting associations for silent residents. Additionally, these methods frequently suffer from severe trajectory fragmentation caused by the ontological mismatch of applying continuous physical motion models to discrete logical behaviors. To address these limitations, we propose a novel framework that reformulates data association as a sequential decision-making task within a Markov Decision Process (MDP). We solve the defined MDP via Behavior Cloning (BC) to avoid complex reward engineering by directly learning from expert demonstrations. We instantiate the policy using two complementary architectures, including a lightweight Transformer that treats association as a multi-label classification task over linearized event histories, and a fine-tuned Large Language Model (LLM) that reformulates the problem as conditional text generation to leverage pre-trained semantic knowledge. Extensive evaluations on three diverse datasets demonstrate that our approach significantly outperforms state-of-the-art baselines, particularly in resolving concurrent multi-resident events. Furthermore, the LLM-based policy shows preliminary cross-environment transfer potential under benchmark settings, while its robustness in real deployments still requires further validation.
[AI-47] A.S.E 2.0: Understanding How Repository Context Impacts the Security of LLM -Generated Code
链接: https://arxiv.org/abs/2611.06294
作者: Jiongchi Yu,Keke Lian,Junqing Wang,Zhikang Fan,Zhaoxuan Li,Yongheng Huang,Ziming Zhao,Dong Zhang
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 15 pages
Abstract:Large language model (LLM) coding assistants and agents construct a repository context before generating code. For security-sensitive code, this context decides whether the model sees the project’s safeguards (e.g., authorization checks) or insecure code to imitate. However, context construction is designed almost exclusively for functional correctness, and secure-code benchmarks treat the context as fixed, so how repository context affects the security of generated code remains unclear. We construct CASE-Bench, a benchmark of 145 repository-level tasks from real CVEs in 51 projects. Each task runs in Docker, has functional tests and a proof-of-concept exploit validated in both directions, and is annotated with the repository code that enforces or undermines its security. Using CASE-Bench, we compare 13 context construction methods and six controls with six LLMs, edit single context elements, and analyze coding-agent trajectories. We find that (1) functionality-oriented context raises the functional pass rate by up to 8.2 percentage points (pp) but the secure pass rate by at most 3.7 pp, (2) withholding a safeguard lowers the secure pass rate by up to 8.1 pp, while one inserted insecure exemplar raises correct-but-exploitable code by 6.3 pp, and (3) coding agents observe the relevant safeguard in only 27-36% of trajectories. These findings inform security-aware context construction and secure-code evaluation.
[AI-48] he Split Protocol Decides the Result: Re-Evaluating Encrypted Traffic Classification
链接: https://arxiv.org/abs/2611.06290
作者: Mayan Sharma,Kriti Saini,Devansh Behl
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 12 pages, 4 figures, 6 tables Code and data splits: this https URL
Abstract:Encrypted traffic classifiers are routinely reported at 95-99% accuracy. We show that on one of two public corpora, the same model and code score macro-F1 0.992 under the evaluation protocol standard in the literature and 0.567 under a protocol that keeps flows from one packet capture on one side of the split. Nothing changes but the partition (gap 0.426 over eight seeds, p = 0.008 after Holm-Bonferroni correction). The inflation is not a property of random splitting as such. On a month of backbone QUIC traffic, we evaluate five split protocols at eight seeds each: random, temporal and client-disjoint splitting are statistically indistinguishable from session-disjoint, while server-disjoint splitting costs 0.192 macro-F1 (d = 17.0, p = 0.008). Which axis matters depends on how the corpus was collected, in a way invisible from its label space. We give a diagnostic that identifies the right axis before any model is trained: single-identifier probes. A decision tree on the capture identifier alone reaches 0.983 under a random split and exactly 0.000 under session-disjoint splitting. Under this evaluation, a metric-trained flow embedding leads four deep baselines but does not beat gradient-boosted trees on twenty packet sizes. Five of our own pre-registered architectural claims fail: distance-based rejection of unknown applications beats softmax thresholding, but ablating every metric objective makes rejection better, because class separation correlates at r = -0.88 with rejection quality. Few-shot enrollment is flat within a corpus but, across corpora, five labelled flows per class recover 91% of the deficit left by failed zero-shot transfer. All splits, results, code and negative findings are released. Comments: 12 pages, 4 figures, 6 tables Code and data splits: this https URL Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) ACMclasses: C.2.3; I.2.6 Cite as: arXiv:2611.06290 [cs.LG] (or arXiv:2611.06290v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.06290 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Mayan Sharma [view email] [v1] Wed, 4 Nov 2026 10:25:50 UTC (75 KB) Full-text links: Access Paper: View a PDF of the paper titled The Split Protocol Decides the Result: Re-Evaluating Encrypted Traffic Classification, by Mayan Sharma and 2 other authorsView PDFTeX Source view license Additional Features Audio Summary Current browse context: cs.LG prev | next new | recent | 2026-11 Change to browse by: cs cs.AI 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)
[AI-49] DyPSI: Dynamic Physics Sensing via Joint Field and Sensor-Trajectory Generation NEURIPS2026
链接: https://arxiv.org/abs/2611.06280
作者: Yizhou Zhang,Panqi Chen,Lei Cheng,Ting Zhang,Jianlong Li,Jianfeng Shi,Shikai Fang
类目: Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026. 35 pages, 4 figures
Abstract:Physics sensing in evolving environments requires reconstructing dense spatiotemporal fields from sparse observations while adapting where future observations should be collected. Existing reconstruction and sensor-placement methods have made substantial progress, but they typically rely on static sensing assumptions: they either use a single layout across all frames or optimize placements independently at each frame, which limits their ability to produce temporally coherent sensing trajectories. We introduce DyPSI, a dynamic physics sensing framework that formulates this problem as the joint generation of future fields and sensor trajectories conditioned on historical observations. DyPSI lifts discrete sensor coordinates into continuous Sensor Position Fields, encodes fields and sensing configurations with a shared Functional Tucker representation, and trains a joint diffusion model to generate future fields and sensing trajectories simultaneously in the resulting latent space. A Gaussian-process temporal kernel correlates the perturbation injected during diffusion training, biasing the denoiser toward temporally smooth recoveries, and we further provide an analysis linking the EDM training objective to a sequence-aggregated A-optimality criterion and a kernel-induced bound on discrete trajectory variation. Experiments on turbulent flow, GLORYS12 sea-surface temperature, and 3D car aerodynamics show that DyPSI consistently outperforms static and frame-wise placement strategies, with substantial error reduction under tight sensor budgets. The code is available at this https URL.
[AI-50] Learning Pseudo-Riemannian Manifolds for Heterophilic Graphs via Graph Signature NEURIPS2026
链接: https://arxiv.org/abs/2611.06265
作者: Yun Young Choi,Asung Kil,Sun Woo Park,Minho Lee,Seokhwan Kim
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at NeurIPS 2026. 41 pages, including appendix
Abstract:Several widely used families of Graph Neural Networks carry a positive-definite geometric prior. Graph convolution read as heat diffusion needs a positive semi-definite operator, hyperbolic models place features on positive-definite Riemannian manifolds, and dot-product and GAT-style attention scores use positive-definite inner products. On heterophilic graphs, where edges often signify dissimilarity or structural repulsion, that prior is a mismatch: no positive-definite form can assign a negative squared distance to a direction the labels separate. To measure the mismatch, we introduce the Graph Signature Index ( s_\mathcalG ), the inertia of a class-compatibility matrix built from label topology alone, and we prove that no value of the edge or node homophily index in (0,1) rules out a negative eigenvalue. We then propose the Pseudo-Riemannian Attention Network (PRAT), whose attention logit is a signed squared pseudo-distance under a learnable diagonal metric initialized from s_\mathcalG . The design is diagnose-then-learn: the signature supplies a sign structure that unconstrained training keeps outside near-zero directions, and the projection learns the rest of the metric around it. Across fifteen node-classification benchmarks PRAT outperforms or matches the strongest heterophily-adapted models. A controlled ablation ties that performance to the sign structure itself: with architecture and protocol fixed, forbidding the metric to be indefinite costs accuracy on the graphs s_\mathcalG marks as indefinite and leaves the others unchanged.
[AI-51] HarnessFlow: MCTS-Guided Harness Optimization for LLM Agents
链接: https://arxiv.org/abs/2611.06254
作者: Zhi Hong,Qianlong Wang,Hongchen Wei,Aidan Wu,Daohong Ying,Mingze Kong,Bei Liu,Chong Luo
类目: Artificial Intelligence (cs.AI)
备注: 18 pages, 8 figures, 7 tables
Abstract:The performance of large language model (LLM) agents depends on both the underlying model and the harness that governs interaction with tools and the environment. Yet harness design remains largely manual, and automating it requires navigating an open-ended search space through costly end-to-end evaluations. We introduce HarnessFlow, a framework for optimizing complete executable harnesses under a limited candidate-evaluation budget, with the model and task environment held fixed. HarnessFlow organizes candidates and their evaluation histories in a refinement tree, using a Monte Carlo Tree Search (MCTS)-based policy to balance exploitation and exploration. A coding agent refines selected harnesses using archived code, execution feedback, and a library of reusable harness mechanisms. With 15 refinement attempts per benchmark, HarnessFlow achieves the highest mean test scores among the compared methods on six benchmarks spanning agentic tasks, coding, and web interaction. The optimized harnesses achieve an average score of 69.48%, outperforming the initial harnesses and Meta-Harness by 10.89 and 3.95 percentage points, respectively, while incurring lower average per-task execution costs than hand-engineered baselines. These results suggest that structured harness search offers a promising route toward automating the design of capable and efficient LLM agents.
[AI-52] Rubric Curriculum RL: Exploiting the Generation-Verification Gap in Non-Verifiable Domains ICML2026
链接: https://arxiv.org/abs/2611.06244
作者: Sumeet Ramesh Motwani,Tejas Krishnan,Alistair Letcher,Charles London,Suhaas Bhat,Huitian Jiao,Philip Torr,Riashat Islam,Christopher Summerfield,Jakob Foerster,Christian Schroeder de Witt,Qilong Gu,Shital Shah
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at ICML 2026
Abstract:Reinforcement learning with verifiable rewards (RLVR) has led to significant gains in math and code generation. Extending these gains to open-ended domains remains challenging: ground-truth verification is unavailable, human annotation is expensive, and learned reward models are prone to reward hacking. The generation-verification gap, where judging good outputs is easier than producing them, offers a potential path to improvement. However, rubric-based judges are often imprecise and stationary, so policies can overfit or reward hack them. We introduce Rubric Curriculum RL (RcRL), which exploits this gap without new data, human annotations, or stronger teacher models. We combine pairwise preferences over rollouts with a curriculum over rubric criteria, yielding a more discriminative signal than absolute scoring and a non-stationary objective. Whereas baselines plateau within a few hundred steps, RcRL preserves output entropy and keeps improving past 1000 steps. On creative writing, RcRL achieves a 70.5% human win rate against the base model, with gains across benchmarks and judges. On HealthBench, RcRL improves over the base model by 14.6% (full set) and 25.4% (consensus subset), outperforming all training baselines, including instance-specific rubrics.
[AI-53] A Definition of Catastrophic Performance
链接: https://arxiv.org/abs/2611.06219
作者: Henrik Marklund,Alex Infanger,Benjamin Van Roy
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:A central challenge in AI is specifying objectives that align with human preferences. While existing theoretical results elucidate how misspecification impacts performance, they typically do not distinguish between merely suboptimal and catastrophic outcomes. To enable rigorous analysis of catastrophic risk from AI misalignment, we introduce a formal definition of catastrophic performance. Our definition is based on two reference values. The first, baseline value, captures the value attained if no agent is deployed. The second, orthogonal value, captures the value of a trajectory that is unrelated to human preferences, in a sense we describe. Orthogonal value is typically far worse than baseline value. We interpret these reference values through examples in which past human activity has altered the state of the environment to human benefit. Starting at the resulting state, an AI agent is deployed with the goal of accelerating further progress. Catastrophic performance represents this agent’s undoing of benefits afforded by human activity.
[AI-54] CoSafe: Cooperative Runtime Safety Assurance for Mixed-Traffic Autonomous Driving ICSE2027
链接: https://arxiv.org/abs/2611.06204
作者: Dingji Wang,You Lu,Haowen Jiang,Yanggang Chen,Yuyang Bi,Bihuan Chen,Shenglin Bao,Xin Peng
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI); Robotics (cs.RO)
备注: Accepted by the 49th IEEE/ACM International Conference on Software Engineering (ICSE 2027). 13 pages
Abstract:With the rapid development of autonomous driving systems (ADSes), real-world roads are evolving into mixed-traffic environments where autonomous vehicles (AVs) and other traffic participants (e.g., human-driven vehicles (HDVs), cyclists, and pedestrians) coexist and interact with each other. However, current ADSes still incur various safety violations in mixed traffic scenarios, raising a strong demand for runtime safety assurance frameworks that can continuously monitor hazards and prevent potential violations. Existing ego-centric assurance approaches are insufficient for mixed traffic, where violations often arise from locally reasonable yet jointly incompatible decisions among multiple vehicles and cannot be reliably resolved through independent takeover. To bridge this gap, we propose CoSafe, a modular and ADS-agnostic cooperative runtime safety assurance framework for mixed-traffic autonomous driving to extend runtime safety assurance from ego-centric takeover to scene-level cooperative intervention. CoSafe leverages vehicle-to-everything (V2X) communication to enable roadside units (RSUs) to monitor scene-level hazards and jointly mitigate hazards involving multiple vehicles. We integrate CoSafe with seven state-of-the-art ADSes, and our evaluation on the InterDrive benchmark demonstrates that CoSafe improves the driving score of AVs by 70.03% on average, reduces safety violations by 87.80%, and accurately identifies unsafe AVs with an F3 score of 0.930, while satisfying real-time constraints. Comments: Accepted by the 49th IEEE/ACM International Conference on Software Engineering (ICSE 2027). 13 pages Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Robotics (cs.RO) Cite as: arXiv:2611.06204 [cs.SE] (or arXiv:2611.06204v1 [cs.SE] for this version) https://doi.org/10.48550/arXiv.2611.06204 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-55] CARM: Conformal Adaptive Risk Masking for Reliable Scientific Super-Resolution
链接: https://arxiv.org/abs/2611.06151
作者: Guangtao Zhang,Haonan Wu,Junxuan Gao,Zhengyang Liu,Zhaoxi Zeng,Zhihao Li,Jianguo Huang,Zhiqiang Liu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Probability (math.PR)
备注:
Abstract:Deep learning-based super-resolution (SR) can reconstruct fine-scale Earth-observation fields from coarse observations, but reconstruction accuracy alone does not provide a calibrated rule for identifying reliable spatial regions. Conformal Mask Calibration (CMC), built on conformal risk control, converts a spatial uncertainty score into a trust mask and controls the expected worst-case local distortion under exchangeability. However, its fixed threshold can become miscalibrated in Earth-observation streams affected by seasonality and episodic events. We propose \textitCARM (Conformal Adaptive Risk Masking), a statistical framework for adaptive spatial risk control under temporal distribution shift. CARM maintains experts with heterogeneous update rates and aggregates their threshold proposals online using an asymmetric surrogate loss, balancing rapid adaptation to distribution changes against stability during relatively stationary periods. All updates are causal and use only mature feedback available after a fixed delay, allowing CARM to operate with any SR model that provides a spatial uncertainty score. We prove that the time-averaged mask-conditioned risk converges to at most \alpha at an O(1/T) rate, without distributional assumptions and for any choice of aggregation weights across experts. On a daily South China Sea chlorophyll- \textita SR stream, CARM reduces the empirical violation rate from 43.49% to 9.38% and the moving-average exceedance frequency from 47.29% to 1.22% at \alpha=0.10 , while retaining a non-trivial trusted region. Experiments on ERA5 wind and OSTIA sea-surface temperature further show consistent relative improvements across multiple SR backbones.
[AI-56] SafeStart: Mitigating the Overlooked Safety Cost of Active Expert Reduction in MoE LLM Serving
链接: https://arxiv.org/abs/2611.06147
作者: Erfan Baghaei Potraghloo,Saeid Shokoufa,Seyedarmin Azizi,Arya Fayyazi,Souvik Kundu,Mehdi Kamal,Massoud Pedram
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 40 pages, 13 figures, 41 tables
Abstract:Mixture-of-Experts (MoE) language models send each token through a few of many expert subnetworks. Serving methods make these models faster by running even fewer experts per token, and they are validated on accuracy benchmarks. We show that these methods can weaken the model’s refusal of harmful requests, and we introduce SafeStart, which largely restores refusal while keeping most of the speedup. In vLLM on one RTX A6000, keeping about two thirds of the selected experts speeds up offline decoding by 15 to 28% at batch size 8. We find that, at this setting, under a jailbreak that tells the model not to refuse, the share of harmful requests that the model answers rises by 18.4 percentage points on Qwen1.5-MoE, 9.5 on OLMoE, and 6.6 on DeepSeek-V2-Lite, the three of nine models where it rises significantly, while MMLU accuracy drops by less than two points. We discover that most of this increase arises while the model processes the prompt and generates its first few tokens. SafeStart therefore runs all experts for the prompt and the first eight generated tokens, and the reduced set of experts afterwards, through a fused kernel in vLLM. With experts removed, SafeStart keeps at least 85% of the offline decode speedup at batch sizes 8 and 32 and brings the change under this jailbreak to between -2.1 and +1.3 points for three expert-reduction methods on these three models. On 889 held-out requests, in an analysis fixed before generation, truncation raises harmful compliance by 5.8 to 21.4 points, and SafeStart leaves a rise of 0.6 to 2.6 points, non-inferior to the original model on two of the three models. Our results show that faster MoE serving need not cost safety: a serving system can keep most of the gain from running fewer experts by running all of them at the start of each response, and it should test refusal, not only accuracy, before it enables such a method.
[AI-57] ModalSearcher: Weaving Visual Evidence into Autonomous Cross-Modal Deep Search
链接: https://arxiv.org/abs/2611.06131
作者: Zhihan Yin,Ruyi Li,Chengyu Fang,Yifeng Yao,Nonghai Zhang,Pengfei Xiong,Terry Zhang,Huishuai Zhang,Dongyan Zhao
类目: Artificial Intelligence (cs.AI)
备注: 29pages
Abstract:Static and incomplete parametric knowledge limits the performance of multimodal large language models on novel-knowledge and knowledge-intensive tasks, especially those requiring multi-hop evidence acquisition. Existing multimodal agentic search methods often fix visual evidence at the input or a predefined trajectory position, making it difficult for agents to select search modalities as their information needs evolve and to handle complex real-world queries. We define this capability as autonomous cross-modal search: an agent chooses among direct reasoning, text-to-text search, text-to-image search, and reverse-image search, actively introduces visual evidence when appropriate, and switches between and integrates modalities across multiple hops. We propose ModalSearch-Pipeline, a scalable procedure for constructing an interleaved multimodal knowledge graph. It samples paths with diverse depths and modality compositions and uses backward question construction to obscure entity names and image targets that would expose reasoning shortcuts while retaining necessary disambiguating clues, yielding text-only, image-initialized, and arbitrarily interleaved multi-hop problems. We train the ModalSearcher Agent with supervised fine-tuning and reinforcement learning in a live Web environment, and construct ModalSearch-Bench. Experiments show that ModalSearcher-8B improves average accuracy by 16.2 points over Qwen3-VL-8B under the same agentic workflow across five multimodal knowledge-intensive QA benchmarks, with consistent gains on independently constructed active visual-search benchmarks.
[AI-58] From Handcrafted to Generative: RLVR-Optimized Icon Generation for Alexa
链接: https://arxiv.org/abs/2611.06123
作者: Rohan Sequeira,Matthew Rosenberg,Amin Mahmoudabadbozchelou,Oleg Poliannikov,Jonathan Rubin
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Alexa+ devices with screens display animated icons during conversations to visually reinforce what is being said. The original set of these icons was hand-drawn by designers one at a time, a process that produces polished results but is time-consuming and cannot scale to the range of topics encountered in open-ended conversation. We describe an automated pipeline, deployed in production, that generates these vectorised animated icons from a text description alone and surfaces relevant animations to customers based on their ongoing conversation. The pipeline decomposes icon creation into a sequence of stages (image generation, semantic decomposition, vectorisation, styling, and animation). Its quality is determined at the very first of these, image generation, on which every downstream stage depends. We therefore focus on improving this first and most critical stage. Because the icons are rendered on small display canvases under strict design guidelines, generations are often unsuitable for production. Supervised fine-tuning is not viable, as no dataset of line-art icons exists under these constraints. Instead, we design objective verifiers that automatically flag generations violating the design guidelines, and use them to fine-tune a flow matching image generation model via reinforcement learning from verifiable rewards (RLVR). The fine-tuned model raises the pass rate from 15.7% to 58.7% (3.7x) under a VLM judge that is independent of the reward model. Both the trained model and the verifier are deployed in the production pipeline, where the verifier also serves as the automated quality gate before final design review.
[AI-59] Ab Initio Accuracy Beyond the Cutoff: Distributed Message-Passing Equivariant Attention at Tens-of-Millions-of-Atom Scale
链接: https://arxiv.org/abs/2611.06122
作者: Allex Kan,Ji Zhang,Hao Yin,Hao Li,Owen Fu,Xunxing Chen,Lin Huang,Zion Wang,Lucas Li,Jiajun Cheng,Arthur Jiang,Junping Zhao,Jia Zhang
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Biological Physics (physics.bio-ph)
备注:
Abstract:Molecular dynamics predicts molecular motion from atomic forces and trajectory integration. Density functional theory (DFT) provides first-principles forces, but its O(N^3) cost limits simulations to thousands of atoms; empirical force fields scale to millions but miss bond breaking and charge transfer. Equivariant message-passing Transformers learn the ab initio potential at linear cost. Stacked layers propagate information beyond a single cutoff, giving four layers a receptive field up to 23Å. At this frontier, scale is limited by communication: layers exchange boundary features across devices, force backpropagation repeats the exchange, and moving atoms change neighbor graphs and tensor shapes. We present Tessera, a distributed runtime that fixes workload shapes between rebuilds. It sorts atoms along a space-filling curve into equal-count shards, ensuring identical atom counts and tensor shapes per rank. Peer sets scale as K^0.5 , not K . A conservative neighbor skin preserves topology across updates, enabling each rebuild interval to replay as one CUDA Graph. Fused kernels, three-level neighbor search, and overlapped All-to-All complete it. On 512 GPUs, Tessera sustains 30.7M atoms at 89.7% weak-scaling efficiency (77–90% across three per-GPU loads); per-step time rises 1.11–1.29x as machine and problem grow 16x. Graph reuse cuts full-step time 1.87\times . Across an 8\times device increase, neighbor-search cost grows 1.7\times , versus 3.0\times for Morton order. In a live 10.24 M-atom salt solution, NVE energy drift is 0.42,\mu eV/atom/ps over 10 ps, and structure matches a serial reference. A 10.06M-atom Norovirus capsid sustains 100 ps of NVT at \approx 8 ps/day on 256 A100 GPUs. Over 200,000 steps, it has no broken bonds or anomalously close pairs and uses three compiled variants, with no graph breaks or eager fallbacks.
[AI-60] Beyond Risk Cues: A Mechanism-Level Benchmark and Verifiable Debate for Scam Video Detection on the Web
链接: https://arxiv.org/abs/2611.06110
作者: Yifan Wang,Junyu Lu,Shun Zhang,Chaozhuo Li,Fanliang Bu,Rong Yin,Lingbin Bu,Changrong Min,Peng Chen
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Online scam videos pose significant threats to individuals and society. Existing studies primarily focus on binary classification, offering limited insights into how scams work and the reasoning behind detection decisions. Moreover, they often overlook the challenge of distinguishing scam videos from superficially similar non-scam content, such as advertisements, and the robustness of detection models against misleading arguments. These limitations hinder reliable scam video detection. To address these limitations, we introduce MechScam, a benchmark comprising 6,614 videos covering diverse scam mechanisms and non-scam categories. MechScam provides explanations supported by video evidence, annotations of potential scam elements and their context, and misleading arguments constructed from genuine evidence. Recognizing that genuine evidence can still be used to justify misleading conclusions, we further propose ScamCourt, a debate framework in which trained prosecution and defense debaters present opposing arguments, while a trained judge makes decisions based on a transcript with verified citations. Extensive experiments show that state-of-the-art MLLMs struggle to distinguish scam videos from similar non-scam content. ScamCourt outperforms larger MLLMs and same-backbone fine-tuned baselines, reducing false positives and showing stronger resistance to misleading arguments. Together, MechScam and ScamCourt contribute to more reliable and interpretable scam video detection. The dataset and code are available at this https URL.
[AI-61] Reasoning -Trace Overflow in Code Generation: An Empirical Study of Explicit Reasoning Models
链接: https://arxiv.org/abs/2611.06095
作者: Lukas Twist,Helen Yannakoudakis,Jie M. Zhang
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 12 pages, 6 figures, 1 table
Abstract:Explicit reasoning can help language models generate code, but it also introduces a practical risk: the model may spend so much of its output budget reasoning that it never produces usable code. We study this behaviour in relatively small open-weight reasoning models, framing it as reasoning-trace overflow. Across seven models and six code benchmark settings, we measure how much of a fixed 32k-token output budget is spent on reasoning, whether that reasoning improves pass@1, and whether lightweight inference-time controls can mitigate overflow. Our results show that the value of reasoning is highly conditional. Overflow is rare on simpler function-level tasks, but becomes common on harder code-generation benchmarks and for particular model families: on several model–benchmark pairs, more than half of generations exhaust the full 32k-token budget while still reasoning. Most complete overflow is associated with repetitive reasoning loops, although substantial long reasoning also occurs without looping. Disabling reasoning can therefore be competitive for overflow-prone models, simply because it ensures that an answer is produced. At the same time, reasoning remains valuable for stronger models on harder algorithmic and code-understanding tasks. Finally, prompt steering can substantially reduce overflow for some models but is inconsistent, while budgeted transitions provide more direct control and can improve both performance and token efficiency for overflow-prone models. These findings suggest that explicit reasoning should be treated as a controllable resource for code generation, rather than an always-on capability.
[AI-62] SysPilot: Cross-Library Performance Autotuning via Library Performance Contracts
链接: https://arxiv.org/abs/2611.06078
作者: Yineng Yan,Haowei Chen,Christopher J. Rossbach
类目: Performance (cs.PF); Artificial Intelligence (cs.AI)
备注: SysPilot is open source at this https URL
Abstract:Modern scientific and machine-learning applications compose many independently developed libraries. When these libraries run together, their independently managed runtimes can oversubscribe shared resources: using the same CPU cores, fighting over GPUs, or contending for memory bandwidth. Avoiding such contention requires careful configuration across library boundaries. Existing autotuners do not directly investigate what resources are shared between libraries, which runtime signals correlate with interference. As a result, they are inefficient for finding optimal configurations to mitigate contention. This paper presents SysPilot, a cross-library performance autotuning system based on Library Performance Contract (LPC). A Library Performance Contract captures a library’s resource demands, assumptions, controls, and observability signals in LPCL, a specification language for describing library resource requirements. SysPilot synthesizes contracts from documentation and code artifacts, composes contracts from the libraries used by an application, derives resource-conflict hypotheses, profiles the application to validate those hypotheses, and selects candidate configurations. This design makes the Library Performance Contract a reusable unit of performance knowledge: a contract for a library can be reused across many applications that depend on that library. We evaluate SysPilot on a benchmark suite of cross-library contention scenarios. The results show that SysPilot matches the performance of autotuners that rely on application-specific manual information while retaining cross-application reuse. Compared with general-purpose autotuners such as OpenTuner and SMAC3, SysPilot improves benchmark performance metrics by up to an additional 38% over the best configuration found by these autotuners, while using 3x fewer runs. Comments: SysPilot is open source at this https URL Subjects: Performance (cs.PF); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.06078 [cs.PF] (or arXiv:2611.06078v1 [cs.PF] for this version) https://doi.org/10.48550/arXiv.2611.06078 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-63] UniCompress: From Multi-LoRA Experts to One Compact Model for Edge Intelligence
链接: https://arxiv.org/abs/2611.06075
作者: Mingkang Liu,Peipei Jiao,Weiming Meng,Tianyuan Liu,Lichen Pang,Junyan Li
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Embodied and edge systems increasingly rely on vision-language models (VLMs) to support heterogeneous skills, yet deploying a separate task adapter for each skill while retaining a high-precision backbone creates substantial memory and management costs. We present UniCompress, a deployment-oriented framework that converts multi-task expertise into one compact, adapter-free model. Starting from a consolidated BF16 Qwen3-VL-4B teacher, UniCompress constructs an asymmetric INT2 student using GPTQ initialization and teacher-guided quantization recovery. The approach jointly quantizes and recovers the projection weights in the vision encoder and language backbone, enabling aggressive compression of the complete vision-language model while preserving its multi-task behavior. We evaluate UniCompress on four heterogeneous embodied tasks: posture recognition, household-safety understanding, multimodal dialogue routing, and mobile-GUI action prediction. UniCompress reduces the approximate weight storage from 8.9 GB in BF16 to 2.1 GB in INT2, corresponding to a reduction of approximately 76%. It improves the INT2 macro task score by 34.3 percentage points over GPTQ-INT2 while remaining within 0.9 points of the consolidated BF16 teacher. These results demonstrate the potential of teacher-guided joint vision-language INT2 recovery for deploying a single multi-skill VLM on resource-constrained embodied platforms.
[AI-64] Which One Is Sounding? Binaural Sounding Instance Disambiguation for Audio-Visual Segmentation
链接: https://arxiv.org/abs/2611.06074
作者: Haoze Ma,Yidi Li,Yu Ma,Pengyuan Zhao,Yi Shi,Bin Ren,Yanwen Guo,Dimitrios Kanoulas
类目: Artificial Intelligence (cs.AI); Sound (cs.SD)
备注: 12 pages, 5 figures, including appendices
Abstract:Existing audio-visual segmentation methods mainly predict sound-related regions or category masks and often fail to identify the sounding instance when multiple same-class candidates coexist. Recent audio-visual instance segmentation methods address instance-level prediction but rely largely on motion and cross-frame association, which may become unreliable under sparse observations or inconspicuous motion. To this end, we introduce binaural audio, whose interaural differences encode source-direction information, to provide spatial constraints for distinguishing the sounding instance among same-class visual candidates. We propose Binaural Sounding Instance Disambiguation for Audio-Visual Segmentation (BiSID-AVS), which progressively refines category-level candidate responses into a mask of the true sounding instance. BiSID-AVS first grounds category-consistent candidate regions through audio prompting, then projects binaural cues into an image-plane spatial response map, and finally refines the true sounding instance through dual-level spatial conditioning at the feature and decoder levels. Experiments on the binaurally extended AVS-Synthetic benchmark and Habitat-AVS dataset for robotic indoor perception show that BiSID-AVS achieves the best performance on the multi-candidate subsets, improving instance-level mIoU by 35.19 and 41.21 percentage points, respectively, while substantially reducing false activation on non-sounding same-class regions.
[AI-65] Reglos: A Self-Evolving Personal Memory Agent with Recursive Self-Improvement
链接: https://arxiv.org/abs/2611.06049
作者: Sen Yang,Jiaduo Lin,Zehong Huang,Wen Liu
类目: Artificial Intelligence (cs.AI)
备注: 12 pages, 7 figures, 8 tables. Website: this https URL
Abstract:Personal agents draw on past conversations to tailor their responses to each user. Their memory systems store and update what each user says, and coding agents now also rewrite the harness that turns this memory into the model’s context. Such patches risk overfitting, because one tested only on the failures it targets can fix those cases and help nowhere else. We present Reglos, a training-free personal memory agent that tests whether harness changes help beyond the failures they were written to fix. Its self-evolving memory adds to each user’s append-only record with every conversation, as new preferences supersede old ones and facts can be forgotten. Its harness improves through recursive self-improvement, in which a coding agent proposes patches from failure traces and a critic from a second model family screens each one for benchmark gaming. After the harness is frozen, a fixed evaluator decides which further patches to accept. With the same backbone behind every system, Reglos outperforms state-of-the-art memory systems such as Mem0, MemOS and EverOS, reaching 61.3% and 84.0% on PersonaMem-v2 and LongMemEval-S with Qwen3.8-27B, and 57.5% and 87.2% with GPT-5.6 Luna.
[AI-66] Practical Estimation of the Bayes Optimal Fairness-Accuracy Tradeoff with Soft Labels NEURIPS2026
链接: https://arxiv.org/abs/2611.06027
作者: Mohit Sharma,Okan Koç,Amit Deshpande,Takashi Ishida,Gang Niu,Rajiv Ratn Shah,Masashi Sugiyama
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: Accepted at Neurips 2026
Abstract:There is a fundamental limit to the best prediction error of any model for a given data distribution. In classification, the Bayes error quantifies this limit and serves as both a benchmark for trained models and a criterion to detect overfitting. The state-of-the-art Bayes error estimation techniques for binary classification use only soft labels (positive class probabilities) and require no input features or auxiliary training. We extend such notions to constrained classification problems, such as fairness. Fair classification is an important case of constrained or multi-objective classification. It requires a model not only to be accurate but also to yield fair outcomes (equal or near-equal true positive rates) across different demographic groups (e.g., race and gender). This leads to an inherent fairness-accuracy tradeoff or a Pareto frontier that is used to compare fair classifiers against each other. The focus of our work is to estimate the fundamental limit for this fairness-accuracy tradeoff. We derive an explicit expression for the fair Bayes error rate (i.e., best achievable error rate subject to given fairness constraints), and construct consistent, instance-free estimators for it using only soft labels and group information. We provably bound the bias and the sample complexity of our estimators. Empirically, our method produces a robust, low-variance estimate of the optimal fairness-accuracy tradeoff curve, even when the soft labels are noisy and derived from a classification model on hard labels.
[AI-67] How Unbalanced Must a Latin Rectangle Be?
链接: https://arxiv.org/abs/2611.06003
作者: Zhengyu Li,Vijay Ganesh
类目: Discrete Mathematics (cs.DM); Artificial Intelligence (cs.AI); Combinatorics (math.CO)
备注:
Abstract:We view a k\times n Latin rectangle as n points in the grid [n]^k . Each symbol gives one point, whose coordinates are the columns it occupies. The rectangle is balanced when these points are pairwise equidistant in the \ell_1 metric. Its imbalance measures the distance from this ideal. We study the minimum imbalance from three sides. Constructively, we use a simulated annealing that moves symbols within rows and only penalises column clashes. Against the reference table of Voboril et al. (2025), it improves 33 of the 51 non-square values for 8\le n\le 13 , by up to 39.4%, and matches the rest. A cheap invariant, the row group, certifies that none of the 33 new rectangles is isotopic to rows of a group table, so the earlier column-swapping local search cannot reach any of them. Exactly, a computer-assisted branch and bound proves that the minimum imbalance of a 3\times 9 rectangle is 56 . Díaz et al. found this value, but its optimality was open. A sharper pruning bound, derived from the identity I=2\sum(d-t)^+ , reduces the search tree by a factor of 2.4 to 3.3 on the instances with k\ge3 that we solve. Asymptotically, the minimum imbalance is \Theta(n^3) for every fixed k . A pigeonhole argument gives a lower constant of order k,4^-k , and random rectangles give an upper constant of order \sqrt k . For k=2 , the limit of \min I/n^3 exists, a permuton attains it, and it lies between 4/375 and 0.1196 . Perfect balance is an \ell_1 -equilateral problem, so it forces k=\Omega(n/\log n) rows. For 2\times n rectangles, single two-minute annealing runs improve five of the seven six-hour MIP upper bounds for 14\le n\le 20 . Finally, the Lean 4 kernel checks every rectangle reported in this paper against a 74-line specification that imports only Lean’s core library. We also prove the pigeonhole lower bound in Lean with Mathlib.
[AI-68] Ray-Tracing-Based Dynamic Channel Sounding With Autoregressive Wi-Fi CSI Prediction
链接: https://arxiv.org/abs/2611.05970
作者: Lyutianyang Zhang,Yixin Ying,Liu Cao,Vanlin Sathya,Jinke Ren,Shuguang Cui
类目: Information Theory (cs.IT); Artificial Intelligence (cs.AI); Networking and Internet Architecture (cs.NI)
备注:
Abstract:Channel sounding in Wi-Fi multiple-input multiple-output (MIMO) systems must balance channel state information (CSI) accuracy against CSI acquisition overhead. Device mobility causes channel aging, while autoregressive CSI prediction introduces errors that accumulate over sounding intervals. However, standardized statistical channel models cannot fully capture the environment-specific channel dynamics encountered in physical artificial intelligence (AI) applications, motivating the use of Sionna-based ray tracing to evaluate CSI prediction under device mobility and evolving propagation geometry. This paper develops a dynamic sounding framework that integrates autoregressive CSI prediction with signal-to-interference-plus-noise ratio (SINR)-constrained interval selection. A Transformer with rotary position embedding that converts device positions into CSI sequences predicts CSI across multiple subcarriers for beamforming updates between sounding exchanges. We formulate sounding-interval selection as a predicted average-throughput maximization problem under per-stream SINR lower-bound constraints, including CSI acquisition and processing overhead. A prediction-error perturbation analysis provides the SINR bounds, and cumulative-sum enumeration globally solves the finite single-variable problem. Actual SINR feasibility holds whenever the prediction errors satisfy the bounds. Ray-tracing experiments assess channel prediction, trajectory transfer, and autoregressive prediction errors. In single-STA Wi-Fi 7 link-level simulations, predicted-CSI beamforming improves mean goodput by 67.3%, 112.7%, and 119.1% over stale CSI for linear, circular, and cosine trajectories, respectively.
[AI-69] Graph-Mediated Large Language Model Systems in the Agent ic Era: Reasoning Knowledge Retrieval Execution and Coordination
链接: https://arxiv.org/abs/2611.05953
作者: Haihui Liu,Yang Chen,Zice Qiu,Yun Zhang,Jialun Zhong,Wenhua Wang,Yanzeng Li,Tian Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Graphs play an increasingly important role in large language model (LLM) systems, supporting structural reasoning, knowledge grounding, retrieval, and memory. As these systems evolve toward the agentic era, graphs also organize execution dependencies and multi-agent communication, linking intermediate observations and retained information to subsequent decisions. However, understanding their contribution requires distinguishing the function of each graph operation, the component responsible for its computation, and the subsequent use of its output. This survey introduces an operation-level role–locus taxonomy for reasoning, knowledge and evidence, retrieval and memory, execution, and coordination. We compare how graph observations and external computations guide decisions, how memory representations and updates support later tasks, and how dependencies constrain execution and information delivery. These comparisons distinguish improvements in local metrics from downstream task performance while retaining the conditions under which results were reported. We then organize benchmarks, comparison designs, process diagnostics, and system costs into an evaluation framework. Finally, we identify open challenges in adaptive computation and stopping, memory maintenance, and dependency-aware recovery. The survey provides a basis for assessing when graph mechanisms are useful and for evaluating their effects across components and stages.
[AI-70] BAM: Bivariate Additive Models for Interpretable Tabular Prediction
链接: https://arxiv.org/abs/2611.05940
作者: Atsushi Koike
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Generalized Additive Models with pairwise interactions (GA ^2 Ms) are regarded as glass-box models, since each pairwise shape function can be visualized as a contour plot. In practice, the shape functions learned by methods such as EBM or NODE-GAM are so complex (irregular contours and sharp local fluctuations) that a practitioner cannot meaningfully interpret them or check them against domain knowledge. Auditability in this stronger sense requires substantially simpler shape functions. We argue that such simplification can match the accuracy of standard GAMs if the right feature pairs are captured. We propose Bivariate Additive Models (BAMs), a class of GA ^2 Ms in which every pairwise shape function is constrained to a simple parametric form, instantiated as an asymmetric bivariate Gaussian. Each component is visualizable as a unimodal contour map, and its feature pair is selected automatically during training. Experiments on six tabular benchmarks show that BAM attains predictive accuracy comparable to standard GAMs and approaches that of GA ^2 Ms. We also show the trade-off between accuracy and the number of components to inspect.
[AI-71] Deep Meets Classical: A CPU-Efficient Decision-Fusion Model for Automatic Modulation Classification
链接: https://arxiv.org/abs/2611.05937
作者: Zhiyi Yin,Zhizhao Wen,Ruoxin Zhang,Chao Wang,Puyang Xu,Yifan Jiang,Yusi Gong
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 5 pages, 2 figures, 5 tables. Accepted at the 2026 6th International Conference on Artificial Intelligence, Automation and High Performance Computing (AIAHPC)
Abstract:Automatic modulation classification (AMC) underpins spectrum sensing and non-cooperative signal identification in cognitive radio, yet the two standard CPU-feasible approaches to it, a compact convolutional neural network (CNN) on raw in-phase and quadrature (I/Q) samples and a classical higher-order-cumulant feature classifier, are rarely compared against each other conditioned on signal-to-noise ratio (SNR). We show that they are, in fact, complementary: the CNN is markedly more robust at low SNR, where hand-crafted statistics collapse toward chance, while the cumulant classifier is more accurate at high SNR, where higher-order statistics cleanly separate dense constellations. We exploit this complementarity with FuseAMC, a lightweight decision-fusion model that trains a logistic-regression meta-combiner over the concatenated class-posteriors of an approximately 198k-parameter residual CNN and a gradient-boosted-tree cumulant classifier, on a validation split disjoint from test. On RadioML 2016.10a (5 seeds, 44k-frame held-out test set), FuseAMC reaches 55.8% overall accuracy, improving on the best cumulant-feature baseline by 5.1 points and on the CNN alone by 3.2 points (paired-bootstrap 95% confidence-interval lower bounds of 5.0 and 3.1 points, both positive), while matching the best classical baseline in its own strong high-SNR regime and the CNN in its strong low-SNR regime. With about 0.2 million parameters, a model footprint of a few megabytes, and roughly 3 MFLOPs per frame, the whole pipeline trains and runs on one CPU core, targeting general-purpose and embedded processors without hardware acceleration.
[AI-72] Split-KV: Efficient Context Parallelism for Training Diffusion Drafters
链接: https://arxiv.org/abs/2611.05926
作者: Geng Zhang,Fangzhao Zhang,Cong Xie,Tiantian Fan,Yang You
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Diffusion drafters accelerate speculative decoding by denoising multiple draft tokens in parallel. Their training uses anchor sampling to construct compact draft queries and performs cross-attention with a block-sparse mask over a much longer context. We empirically show that acceptance length remains stable as context scales with a fixed query budget, while increasing the query budget yields diminishing quality gains, jointly establishing a practical regime with queries far shorter than the context. No existing context parallelism (CP) method jointly exploits this query–context asymmetry and meets the nonlinear, sample-dependent load-balancing requirement created by anchor sampling. We present Split-KV, an exact CP method that keeps key/value shards local and limits per-layer attention communication to query-sized tensors, making this communication independent of context length. Its global query view further enables a cost-aware longest-processing-time-first (LPT) strategy that balances context key/value chunks by their realized attention cost. At a CP size of 64, Split-KV retains 85–91% weak-scaling efficiency and achieves up to 3.67\times speedup over the fastest available CP baseline. LPT achieves 1% imbalance and plans three orders of magnitude faster than the best load-balancing baseline. Code is available at this https URL.
[AI-73] Autonomous ETL Pipelines for Salesforce MA Consolidation: A Reference Framework Jitterbit Mapping and Reproducible Simulation Study
链接: https://arxiv.org/abs/2611.05918
作者: Imran Basha Shaik,Sahla Sultana,Khader Moinuddin Shaik
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET)
备注: 16 pages, 4 figures, 7 tables. Code and data: this https URL preprint also archived at this https URL
Abstract:Merging companies that each run Salesforce must consolidate their customer data into a single organization while both source systems remain in use. This requires reconciling different schemas and picklists, resolving duplicate customers across and within organizations, preserving relationships between records, and keeping the target current until cutover. This article proposes a reference framework for an autonomous extract-transform-load pipeline for this task. It combines partitioned bulk extraction, declarative mapping templates, rule-based and statistical quality gates, master data management with deterministic and blocked fuzzy matching, constrained loading with per-record error capture, and change data capture guarded by a data-health circuit breaker. Each stage is mapped to Jitterbit Integration Studio constructs. The framework was implemented as an open reference prototype and evaluated on synthetic two-organization scenarios with known ground truth at three scales, from 32,040 to 383,973 source records, over five random seeds. At a preset threshold, hybrid matching reached pairwise F1 scores of 0.918 for accounts and 0.885 for contacts; calibrating the threshold per object would raise these to 0.938 and 0.963. The quality gates prevented all load failures and all silent loss of unmapped picklist values without wrongly flagging any clean record, whereas removing them produced about 1,000 load failures per run. Over five simulated days of hourly incremental synchronization, replayed events were ignored, an injected upstream fault was contained in the batch in which it occurred, and the target agreed with a full rebuild on more than 99.3 percent of field values. Validation on a production deployment remains future work.
[AI-74] CHARM: Cross-Hardware Attention with Re-Merge Multistream Mechanism NEURIPS2026
链接: https://arxiv.org/abs/2611.05906
作者: Xinhang Zhang,Boning Zhang,Chengchun Liu,Chunpu Li,Lei Sun,Weifeng Zhang,Limin Xiao
类目: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC)
备注: Accepted to NeurIPS 2026 (Poster)
Abstract:As sequence lengths continue to grow, distributed attention based on sequence parallel (e.g., Ring-Attention) has become a mainstream solution for large language models (LLMs). Nevertheless, inherent communication dependencies limit computation-communication overlap, reducing GPU utilization. Moreover, in cross-hardware (NUMA or node) architectures, due to imbalanced communication bandwidth, frequent communication synchronization will limit overall performance. Based on these insights, we propose CHARM+, Cross-Hardware Attention with Re-Merge Multistream Mechanism. We reorganize sequential communications into several levels and achieve intra-level parallelism and computation overlap. Furthermore, we decouple cross-hardware communication from intra-hardware communication, eliminating synchronization overhead caused by bandwidth imbalance. Forward and backward algorithms were evaluated on multiple hardware architectures and consistently outperformed state-of-the-art methods. Compared to Ring-Attention, our forward and backward algorithms achieve average speedups of 1.3x and 1.7x on a dual-NUMA PCIe architecture, and 2.8x and 2.9x on a dual-node NVLink architecture, respectively.
[AI-75] EchoKV: Self-Corrective KV Cache Eviction via Iterative Query Feedback
链接: https://arxiv.org/abs/2611.05903
作者: JianKun Zou,Tongtong Yuan
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:KV cache eviction is not merely a token-ranking problem: the workspace retained for generating eviction queries also determines the quality of those queries. Consequently, an early workspace error can distort later query observations and reinforce an incorrect final eviction decision, particularly when evidence is dispersed across a long context. We present EchoKV, which discovers and addresses this failure mode with iterative correction. It initializes a feedback-compatible bootstrap workspace using proposals at future decode positions, then uses early decode-generated queries to recover the workspace before the remaining lookahead steps. It uses the same number of workspace forward passes as LAQ while adding a tensor re-scoring operation for workspace repair. On LongBench (16 tasks, Mistral-7B-Instruct-v0.2 and Qwen2.5-7B-Instruct), EchoKV shows notable improvements over LAQ on several tasks requiring dispersed evidence, including multi-hop cross-document QA and long single-document comprehension. Across budgets and backbones, EchoKV selectively improves evidence-intensive tasks while maintaining comparable performance on the remaining LongBench subsets. These results identify query–workspace coupling as a practical failure mode of query-driven KV eviction.
[AI-76] RECAP: Context-Aware Prediction Recycling with Request-Local and Persistent Memories for Multimodal Speculative Decoding
链接: https://arxiv.org/abs/2611.05894
作者: Hengjie Zhu,Dayan Wu,Xinze Liu,Peng Fu,Zheng Lin,Weiping Wang
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recycling a target model’s own predictions offers a training-free source of proposals for speculative decoding. In vision-language models, however, existing recycling indexes predictions by the last token alone, conflating predictions computed under different images and prefixes; we find that such reuse fails most often at image-sensitive decoding states. Longer keys reduce this aliasing but match fewer states. We introduce \method, which stores target predictions under one-, two-, and three-token keys: contextual memories live only for the current request and are thus implicitly conditioned on its image, while token-level memory persists across requests to retain coverage. At each lookup, entropy-confidence weighting fuses the retrieved predictions according to their key specificity and concentration. Best-first search composes the fused transitions into a candidate tree, supplemented by suffix-matched continuations, and one target pass verifies the tree under greedy decoding. Without training or an auxiliary draft model, \method\ achieves mean speedups over autoregressive decoding of 1.66\speed and 1.56\speed across seven single-turn benchmarks on Qwen2.5-VL-7B and LLaVA-1.6-7B, exceeding the strongest training-free baseline, Token Recycling ( 1.51\speed and 1.37\speed ). Our implementation is available in this https URL.
[AI-77] AI-Assisted Classification under Correlation Neglect and Trust
链接: https://arxiv.org/abs/2611.05870
作者: Dingwen Kong,Liang Lyu,Saurabh Amin,Amine Bennouna,Daniel Huttenlocher,Asuman Ozdaglar
类目: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:When does an AI assistant improve human decision-making, and when does it hurt? We argue that the answer depends not just on AI accuracy but also on how much its evidence overlaps with the human’s and how the human combines the two. A worker makes a binary decision by observing her own signal and receiving a recommendation from an AI. We develop a human behavioral model with two primitives: her awareness of the overlap and her trust in the AI. Our analysis gives a closed-form phase diagram over AI capability and overlap, with regions of complementarity, impairment, and automation. A worker who neglects the overlap (correlation neglect) can be made less accurate by a more capable AI, double-counting the shared evidence. Trust reshapes the regions - undertrust may compensate for correlation neglect, so that the trust level maximizing accuracy depends on the deployment parameters. Empirically, we identify the two model parameters by pairing the same workers with different AIs, and break the accuracy loss into a correlation-neglect cost and an undertrust cost, one tied to reducing overlap on the AI side, the other to calibrating trust on the human side. On ChatBench, AI assistance raises accuracy by 21 points, yet workers apply only about a quarter of the Bayes-rational weight to AI information; GPT-4o falls in the automation region and Llama-3.1-8b stays at the complementarity boundary; and calibrating trust recovers about 2.6 points for GPT-4o but almost nothing for Llama-3.1-8b.
[AI-78] Auditable Unlearning for LLM s
链接: https://arxiv.org/abs/2611.05866
作者: Piyush Tiwary,Anushka Srivastava,Prathosh A.P
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Machine unlearning aims to remove a user’s data from a trained model after a deletion request. Existing methods are majorly evaluated on three axes - how well they forget, how much capability they retain, and how robust they are to membership inference attacks, but not on whether the provider can offer the user any evidence that the deletion occurred. To fill this gap, we introduce auditability as a fourth axis, defined as the separation between requested (forget set) and retained (retain set) data under a statistic the provider computes. We show that across eleven unlearning methods on TOFU, MUSE and RWKU no method satisfies all four axes, with different methods falling short on different axes. To address this, we propose a two stage method that can be used on top of any unlearning method, in the the first stage we propose GapWiden which uses a hinge loss to increase the difference between audit statistics of forget set and retain set. In the second stage we propose GapVeil , an inference time rejection-sampler which uses the base model as the drafter and re-samples based on the audit statistics of the drafted response. We observe that using our method, auditability and leakage reach oracle parity for every unlearning method, and forget performance converges to the oracle; retain performance improves for nine of eleven methods but remains short of parity. Our results show that evidence of deletion or auditability can be constructed by design, and that treating it as an explicit requirement changes the viability of existing unlearning methods.
[AI-79] PEARL: Prediction Error-Adaptive Regularized Learning for Selective Forgetting under Partial Preference Drift
链接: https://arxiv.org/abs/2611.05842
作者: Rashmi Gangadharaiah,Balakrishnan Narayanaswamy
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注: 36 pages, 6 figures, 10 tables
Abstract:User preferences shift as needs, context, and tastes evolve. This matters for personalized recommendation, LLM response generation, and RLHF-trained reward models, which can become miscalibrated as preferences drift. Existing non-stationary bandit methods forget uniformly across all dimensions, but real preference shifts are typically partial: some dimensions go stale while others remain valid. Neural contextual bandits accumulate confidence monotonically via the Gram matrix, with no mechanism to direct uncertainty where predictions have gone wrong. We propose PEARL (Prediction Error-Adaptive Regularized Learning), which reverses the Gram update along feature directions where predictions are surprisingly wrong, while leaving confidence intact elsewhere. We prove that (i) any non-adaptive policy suffers linear regret after a preference shift, (ii) no fixed exploration rate achieves low regret in both stationary and non-stationary regimes, and (iii) PEARL’s cumulative regret scales with the number of shifted users N_s, not the total population N. On PRISM, a personalized LLM response selection benchmark (1,500 users, 21 LLMs), PEARL beats greedy by +4.0 points and the best neural bandit baseline by +2.5 points (p0.001), while performing on par with greedy when preferences are stable. It beats sliding-window and discounted approaches by 3 to 10 points, showing that selective forgetting outperforms uniform forgetting. The selection policy is a ~100K-parameter MLP that runs in milliseconds on CPU, leaves LLM weights frozen, and works with any API-based model.
[AI-80] Vision Paper: Why Distributed LLM Post-Training and Decentralized Alignment Are Now Essential
链接: https://arxiv.org/abs/2611.05841
作者: Zhanhong Jiang,Qisai Liu,Soumik Sarkar
类目: Artificial Intelligence (cs.AI)
备注: 10 pages, 4 figures, IEEE BigData 2026
Abstract:This vision paper argues that distributed learning, including federated and decentralized, should become a first-class paradigm in LLM post-training, not merely a privacy add-on. Current methods like RLHF and DPO rest on an invisible assumption: preference data should be centralized. We contend this assumption is scientifically unjustified, practically unsustainable, and ethically problematic. As LLMs deploy into healthcare, law, finance, and multilingual communities, preference data originates from populations that cannot or will not centralize. We develop a three-pillar argument grounded in (1) privacy and data sovereignty, (2) pluralism and representational equity, and (3) robustness and scalability. We formalize key technical challenges unique to this intersection, reference drift, preference heterogeneity, and the DP-utility trade-off for contrastive losses, and propose concrete algorithmic directions, and sketch a research agenda. Our analysis shows distributed post-training is not only necessary but tractable, and ignoring it risks AI that is legally noncompliant, representationally narrow, and brittle at scale.
[AI-81] RACE: Calibrated Cumulative Scoring for Rollout-Free Robot Policy Checkpoint Selection
链接: https://arxiv.org/abs/2611.05838
作者: Changyeon Kim,Yekyung Nah,Kimin Lee
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Project website: this https URL
Abstract:Assessing the deployment performance of learned robot policies during training is challenging, as directly measuring task success requires costly real-world trials for each candidate checkpoint. Averaged validation loss offers a rollout-free alternative, but its checkpoint rankings can be inconsistent with deployment performance. We analyze two limitations of this criterion: it penalizes action deviations without considering whether the policy can recover from them, and gives later reference predictions equal weight even when earlier deviations may lead execution away from the reference. We propose RACE, a novel rollout-free metric for ranking checkpoints using a fixed set of held-out successful trajectories. RACE calibrates action deviations against state-conditional thresholds derived from action differences at similar states in these trajectories. We aggregate these scores through a running product, so that a substantial deviation reduces the contribution of subsequent reference predictions. Averaging the resulting cumulative scores over time and across trajectories yields the checkpoint score. In simulation, RACE outperforms baselines across validation sources, while several baselines degrade sharply when the validation set changes. On real robots, its rankings align with deployment success across the evaluated policy backbones and tasks of varying difficulty. Perturbation analyses further show that RACE distinguishes deviations with markedly different effects on task success, even when baselines respond similarly.
[AI-82] Autonomy and Coordination in Intelligent Manufacturing: A Recursive Autonomous Agent Framework
链接: https://arxiv.org/abs/2611.05828
作者: Yuning Chen,Hongda Zhang,Linqiang Hu,Ziqing Zhou,Yi Liu,Zhao Huang,Ding Zhou,Haining Xu
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Intelligent production requires systems to make decisions and organize work according to production goals and operating conditions. It also requires continuous adjustment through feedback. This paper analyzes intelligence in human production organizations from a complex adaptive systems (CAS) perspective. The analysis identifies two coupled mechanisms: individual autonomy and collective coordination. Individual autonomy enables actors to form local decisions using available information. Collective coordination aligns these decisions with shared goals and joint constraints. Learning and adaptation improve both mechanisms over time. A production-system model formalizes how local decision rules and coordination jointly form system decisions. On this basis, we propose the Recursive Autonomous Agent (RAA) framework, which combines goal-driven autonomous agents with a recursive organizational structure. Goal contracts define local objectives and operating boundaries, while capability contracts support the selection and composition of production units. Common interfaces allow coordinated groups to participate as units in larger organizations, connecting equipment, production lines, and factories. RAA thus provides an engineering approach to next-generation intelligent manufacturing, translating the requirements of intelligence, openness, and system-wide integration into mechanisms for autonomous and coordinated production.
[AI-83] A Hybrid Approach for Automated Security Pattern Recovery
链接: https://arxiv.org/abs/2611.05823
作者: Anh Minh Nguyen,Tuan Dat Tran,Alzubair Hassan,Liliana Pasquale
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注: 35 pages, 41 figures, 11 tables
Abstract:Security-by-design emphasises integrating security into software design and development. Security pattern catalogues operationalise this principle by prescribing reusable architectural solutions to recurring security problems. However, recovering their implementations from source code remains challenging because security patterns are abstract, span multiple components, and admit many valid implementations. We present SecPatRec, a hybrid approach for automated security pattern recovery from source code repositories. We treat a security pattern as a reference architecture and frame its recovery as a deductive task: determining which prescribed roles are present, grounding them in code entities, and recovering their interaction sequences. SecPatRec pairs a ReAct-style LLM agent equipped with semantic and lexical code search for role detection and grounding with static code property graph analysis for interaction recovery. We curate a human-validated ground-truth dataset of 100 pattern instances covering five security patterns across 93 open-source Java and Python repositories, and introduce a three-level evaluation framework measuring recovery correctness against this ground truth. We evaluate SecPatRec using four LLMs, including proprietary and open-source models. Averaged over all models and pattern instances, SecPatRec achieves role-detection accuracy of 0.84, role-grounding similarity of 0.84, and an interaction-recovery score of 0.66. Compared to a one-shot end-to-end LLM baseline, SecPatRec achieves higher aggregate recovery scores and lower estimated LLM API costs per run, while processing repositories that exceed the baseline’s configured input limit. The recovered role mappings and statically derived interaction sequences provide structural and behavioural evidence that can support inspection of security-pattern implementations.
[AI-84] When File Ingestion Misleads Language Models: Measuring and Mitigating Divergences in Office Documents
链接: https://arxiv.org/abs/2611.05806
作者: Penghui Li
类目: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
备注:
Abstract:Office documents are now routinely uploaded to large language models (LLMs) for processing. However, an LLM does not always see what a human sees, because it typically reads the text that an extractor produces from the document’s underlying Office Open XML (OOXML) files. Depending on the implementation, extractors differ in which OOXML parts they read and how they interpret them, so their output can differ from what Office applications display. For example, by slightly shifting the positions of text boxes on a slide, an adversary can swap the order in which the model reads them while the slide looks unchanged. We show that this extraction divergence is an attack surface for LLM applications. In this work, we first measure what an adversary can exploit by analyzing the widely used extractors that LLM applications rely on. By combining static source code analysis with dynamic analysis, we identify 19 OOXML channels where extractors read content a human cannot see and 12 where they drop content a human can. We also find that widely used extractors ignore display formatting and reading order, so even visible content reaches the model in a different form. Building on these findings, we construct five attacks that lead the model to interpret a document differently from a human reviewer, without injecting any instruction. We test the attacks on four major chat assistants (ChatGPT, Claude, Gemini, and Microsoft 365 Copilot) and four coding agents (Codex, Claude Code, Antigravity, and GitHub Copilot CLI), and every one of them is misled by at least one attack. Finally, we propose Prism, a lightweight defense that reconstructs the human view from the OOXML specification before passing the text to the model. Prism also generalizes to unseen attack payloads. Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.05806 [cs.CR] (or arXiv:2611.05806v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2611.05806 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-85] Consistency Is Not Understanding: Orbit Audits for Neural Field Weight Readers
链接: https://arxiv.org/abs/2611.05798
作者: Sungguk Cha,Mintae Kim,Youngsub Han,Byoung-Ki Jeon,Sangyeob Lee
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Recognizing visual content from neural-network weights requires separating functional information from arbitrary parameter coordinates. We audit this separation with hidden-neuron permutations and positive ReLU rescalings that preserve the represented function. The study contains 480 independently fitted shape fields and 1,000 MNIST fields, with transformed copies grouped by origin. On shape recognition, a statistics reader falls from \ShapeRawStats% to \ShapeJointStats% accuracy under joint transformations; a normalized set reader retains \ShapeNormSet% with unchanged predictions. Four small frozen language models obtain \LMRange% raw accuracy with next-token scoring. A separate four-configuration reasoning experiment with 1,920 scored sessions obtains \FrontRawRange% raw accuracy with generated answers, while fresh identical raw prompts disagree on \FrontRawRepeatRange% of origins. Canonicalization makes equivalent prompts identical, but this neither establishes recognition nor guarantees identical generated responses. Joint correctness/agreement metrics, semantic-cue controls, fresh repeats, and exact-rational rounding bounds separate these failure modes. The result is an inspectable evaluation protocol for visual weight readers, with claims restricted to the tested fields, prompts, and execution configurations.
[AI-86] DARE: Difficulty Analysis for Repair Enhancement
链接: https://arxiv.org/abs/2611.05797
作者: Jikwang Lee,Seungwook Jung,Jaechang Nam
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:LLM-based coding agents are primarily evaluated by their solve rates. However, this metric does not reveal how difficult the solved tasks are or account for differences intask difficulty across benchmarks. Determining the difficulty of individual tasks by repeatedly running multiple agents is costly, while a predicted difficulty score alone provides limited guidance on what support an agent needs. To address these challenges, we propose DARE (Difficulty Analysis for Repair Enhancement), which integrates repair-task difficulty measurement, prediction, feature-group contribution analysis, and task-specific repair support. We define continuous task difficulty as the observed failure rate across multiple coding agents. The DARE difficulty model predicts this difficulty using 151 static features extracted from benchmark artifacts. The framework also analyzes the contributions of seven Repair-Task Feature Groups to identify the characteristics most salient to each task and uses this information to select task-specific agent support. We evaluate DARE on four benchmark partitions comprising 2,594 tasks. The difficulty model’s aligned predictions achieve Spearman’s \rho = .476-.621 on held-out partitions, and the most salient Feature Groups vary across tasks and benchmarks. Furthermore, coding-agent experiments show improved repair success when support is matched to each task’s salient Feature Group. These findings demonstrate that difficulty prediction and interpretation can lead to practical improvements in automated program repair (APR), while also providing a basis for difficulty-stratified agent evaluation, cross-benchmark comparisons, and identifying areas of repair capability that warrant further improvement.
[AI-87] FlyHash for LLM Agent Memory: An Empirical Validation and a New Application
链接: https://arxiv.org/abs/2611.05791
作者: Mohan Bheemadolu,Chrisogonas O. Odhiambo
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 19 pages. Code, data and interactive demo: this https URL (archived at this https URL )
Abstract:Large language model (LLM) agents increasingly rely on external memory stores to retain information across conversations, typically via dense vector embeddings searched with approximate nearest-neighbor methods. The fruit fly’s olfactory circuit computes a locality-sensitive hash of sensory input using a strikingly different mechanism - sparse random projection followed by winner-take-all competition - and prior work (Dasgupta et al., 2017) showed this scheme, dubbed FlyHash, preserves neighbor structure better per bit than conventional random-hyperplane hashing (SimHash) on several benchmarks. We apply FlyHash, to our knowledge for the first time, to LLM agent memory retrieval, and ask three questions a practitioner actually needs answered: does the quality advantage hold under a correctly matched-bit protocol (it does, robustly, across 10 seeds and four datasets spanning two modalities, with one specific, previously undocumented comparison pitfall identified and corrected along the way); does that advantage translate into real memory and latency savings, including against the dense-vector indexing (HNSW) practitioners default to today (no: at equal bits FlyHash’s codes are larger and slower to compare than SimHash’s, and HNSW dominates both hash methods at ordinary scale; a sparse-set comparison we introduce recovers most but not all of a naive latency gap, leaving hashing’s practical case narrower and more specific than the quality result alone implies); and does the advantage survive genuine, non-circular agent-memory relevance labels rather than the nearest-neighbor-by-embedding proxy this literature otherwise relies on (it does). We report all three answers precisely, including the two that narrow rather than flatter the headline result, and release our implementation, a working agent-memory demo, and a full account of the methodological pitfalls we found and corrected.
[AI-88] Scheduling Data to Compute: Hierarchical Data Pipelining for LLM Training on AMD GPUs
链接: https://arxiv.org/abs/2611.05751
作者: Gengyu Rao,Ghadeer Ahmed H Alabandi,Su Ann Chong,Thomas Huber,Kailash Gogineni,Subhajit Dutta Chowdhury,Saleel Kudchadker,Suchita Pati,Zirui Zhang,Zhen Huang,Yanyuan Qin,Xiaoming Peng,Lihuan Zhang,Botao Hu,Wei Huang,Yang Bo,Wen Xie,Xiangyu Zhao,Ruibin Zhang,Xiaobo Chen,Vidushi Goyal,Anshu Raina,Claire Songhyun Lee,Jiakun Yan,Yao Fu,Nusrat Islam,Ravi Dwivedula,Jie Zhang,Maria Garzaran,Zhenyu Gu,Emad Barsoum
类目: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models (LLM) have been the key focus of recent machine learning breakthroughs. As the model size grows exponentially, the computation needed for training is astronomical. A GPU can do a huge amount of arithmetic. However, the data its compute units need is not always available when they are ready. This data may still be in slow memory or on another GPU. While the chip waits, expensive compute sits idle, and at scale even a small drop in utilization translates into substantial cost. In this paper, we treat this waiting time as the primary bottleneck. We focus on hardware features designed for explicit data movement. These features load data from off-chip memory onto the chip, rearrange it into the layout the computation requires as it is loaded, and transfer it between GPUs. By orchestrating them, we keep a steady pipeline of data flowing to the compute units so that they remain busy. We present kernel and framework implementations that address this bottleneck using three key data-movement instructions and hardware features. The Tensor Data Mover (TDM) moves tensor data from high-bandwidth memory (HBM) into on-chip memory. Transpose-load instructions transpose tensor data into the layout that LLM computations require as it is loaded on-chip. Finally, System DMA (SDMA) and Remote DMA (RDMA) move data between GPUs over the scale-up and scale-out networks, respectively. Our results show that TDM and transpose-load provide on average 1.88x and 1.19x speedup at the kernel level, respectively. The SDMA and RDMA lead to on average 1.16x and 1.33x speedup for computation overlapped with communication. This is the first AMD LLM training system to utilize all the GPU data pipelining features. Our implementation has been integrated into open source framework and library for AMD GPUs. this https URL this https URL Subjects: Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.05751 [cs.AR] (or arXiv:2611.05751v1 [cs.AR] for this version) https://doi.org/10.48550/arXiv.2611.05751 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-89] PersonaCodeBench: Evaluating Coding Agents in Persona-driven Human-Interactive Sessions
链接: https://arxiv.org/abs/2611.05739
作者: Haixin Wang,Jianheng Hou,Hongyu Shen,Wenzhe Fan,Wei-Chieh Huang,Qianyu Zhu,Weichen Li,Taoran Li,Yuhang Yao,Ruohan Wang,Zhejian Peng,Wanting Jiang,Yifan Chen,Shushu Liang,Haoming Li,Defu Cao,Xu Hu,Siyang Wu,Chen Yang,Haodi Hu,Haichao Zhang,Cong Wang,Zhiyuan Yu,Weizhi Zhang,Yifan Wang,Jintao Huang,Yizhou Sun,Yijun Wang,Yuexing Hao,Xiaomin Li
类目: Artificial Intelligence (cs.AI)
备注: 40 pages, 9 figures
Abstract:Coding agents must deliver correct software while respecting users’ expertise, preferences, and working conditions. Existing benchmarks increasingly incorporate interaction, yet personalized requirement fulfillment remains underexplored. We introduce PersonaCodeBench, which turns user characteristics into verifiable obligations in persona-driven multi-turn coding sessions. Our persona pool spans roles, technical abilities, interaction preferences, and practical constraints, yielding task-grounded evolved requirements for implementation, permissions, testing, explanation, and delivery. At runtime, a controller governs user simulator responses and requirement disclosure, using online requirement evaluation to provide feedback and advance the interaction. Requirements emerge progressively, with their content, disclosure conditions, and acceptance criteria fixed before evaluation. A completion verifier and a personalization verifier separately assess native correctness and persona fulfillment. Joint pass requires both adherence to the declared budget and compliance with validity conditions. We evaluate all major frontier agents, establish a comprehensive leaderboard, and conduct detailed experimental analyses to validate effectiveness. PersonaCodeBench supports reproducible evaluation of coding agents’ ability to meet explicit, personalized user expectations alongside coding objectives.
[AI-90] Do Language Models Know Which Constraints Belong Together?
链接: https://arxiv.org/abs/2611.05726
作者: Junbo Jacob Lian,Kaixuan Chen,Hanzhang Qin,Huiling Chen,Chaoyu Zhang,Chung-Piaw Teo
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Data Structures and Algorithms (cs.DS); Optimization and Control (math.OC)
备注: 23 pages, 4 figures, 15 tables. Code and data: this https URL
Abstract:Dantzig-Wolfe reformulation strengthens the linear relaxation of a mixed-integer program by convexifying a chosen set of constraint families, and every candidate costs a column generation run. The choice is hard when families strengthen the bound only together, because a search that adds one family at a time then stops at the linear relaxation. The classical inequalities of lot sizing, network design and facility location rest on such zero-marginal groups. We prove that when each evaluation convexifies at most s of k families, any method that observes only bounds needs a covering number of evaluations, of order k^w , to find a hidden group of w families, also on actual Dantzig-Wolfe relaxations. A prediction that misses e members reduces the cost to \Theta(k^e) , through one staged search that does not need to know e , so the evaluations a prediction saves measure what the predictor knows. We study two sources of predictions. A one-pass lookahead at an optimal LP dual is provably blind to groups whose families share no variable, and a language model reads the names and text of the model. JointDW verifies the prediction of the language model with the staged search and interleaves it with the lookahead at a bounded loss. Groups that strengthen the bound only jointly appear in a quarter of the MIPLIB-NL models whose bound can be improved. There the strongest language models identify them about as often as the lookahead, their predictions rest on the names for textbook models, and reasoning about doubles the accuracy of the same model. On larger models with distracting families, the lookahead finds the group in one evaluation of three families. Language models given only the text seldom name it, while one reasoning model given every coefficient names it in all but one answer and then needs fewer evaluations than the lookahead for pairs of families.
[AI-91] Self-Supervised Pretraining for Label-Efficient Facial Expression Recognition using Smart Glasses
链接: https://arxiv.org/abs/2611.05711
作者: Zhengtao Gao,Francesco Bombassei De Bona,Leonardo Alchieri,Ivana Kiprijanovska,Simon Stankoski,Bojan Sofronievski,Hristijan Gjoreski,Marc Langheinrich,Martin Gjoreski
类目: Artificial Intelligence (cs.AI)
备注: 7 pages, 5 figures, 2 tables. Best Paper Award at the SEPH 2026 workshop, UbiComp Companion '26
Abstract:Facial Expression Recognition (FER) on optomyographic (OMG) smart glasses such as OCOsense enables privacy-preserving, camera-free affect sensing for health and well-being. Most FER systems are fully supervised, i.e., they rely either on hand-crafted expert features with classical machine learning, or on end-to-end deep models trained from labelled data. Both depend on large quantities of labelled expressions, which are costly to collect and limit deployment in real-world, personalized settings. We investigate Self-Supervised Learning (SSL) as a route to label-efficient FER on these glasses. We pretrain a Vision Transformer with a masked-autoencoder objective on unlabelled OCOsense recordings and finetune it for five-class FER under person-independent cross-validation. Across label budgets from 5% to 100%, pretraining consistently improves performance; with only 5% of the labels, the pretrained model outperforms the from-scratch model trained with all labels. Our results show that SSL pretraining is a practical way to reduce labelling cost for smart eyewear FER, and we discuss the implications for future personalization and continual adaptation.
[AI-92] Agents Think Only When it Helps
链接: https://arxiv.org/abs/2611.05709
作者: Min-Seon Kim,Seungdong Yoa,Sanghyu Yoon,Ye Seul Sim,Suhee Yoon,Dongmin Kim,Jung-Eun Kim
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:LLM agent distillation transfers agentic capabilities to student models by training on complete teacher trajectories that typically interleave thoughts and actions. We question how critical thought supervision is for agents’ ultimate action performance. To answer this, we disentangle their impacts by independently perturbing teacher thoughts and actions during distillation. On widely adopted agent benchmarks across model scales, altering or discarding thoughts has little effect on student performance. We find, however, that thought supervision becomes important when information about the thought is not disclosed in the corresponding action during training. By identifying such discrepant cases, our findings challenge the common practice of distilling entire teacher trajectories for LLM agents, and offer a new insight that thought supervision can be a discernible choice for efficiency depending on whether the information of thought is disclosed in the action.
[AI-93] From Type to Token: Petri Nets Individual Token Philosophy and Individualized Causal Effects
链接: https://arxiv.org/abs/2611.05703
作者: Fernando Delbianco,Fernando Tohmé,Gustavo Bodanza
类目: Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注:
Abstract:Structural Causal Models and Bayesian Networks are the dominant framework for representing causal relations. They are fundamentally type-level formalisms since their estimands, Average Treatment Effects and Conditional Average Treatment Effects, are properties of distributions over populations, not of individual units. When treatment effects exhibit unbounded heterogeneity, these estimands collapse. What persists is the structural, token-level core of causation. We argue that Petri Nets under the Individual Token Philosophy are the natural representation of these token-level causal structures. We develop a formal translation from Bayesian Networks to Petri Nets, providing a two-layer account of individualized causality. The PN layer provides a structural representation of causal mechanisms that does not rest on distributional assumptions. In turn, a conformal prediction layer quantifies causal effects with finite-sample coverage guarantees. Together they constitute a complete account of individual causal effects. Subjects: Artificial Intelligence (cs.AI); Machine Learning (stat.ML) Cite as: arXiv:2611.05703 [cs.AI] (or arXiv:2611.05703v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.05703 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-94] Quantifying Trait-State Hierarchy in EEG Foundation Model and Signal Processing Embeddings
链接: https://arxiv.org/abs/2611.05652
作者: Anusha Madapura,Aditya Kommineni,Shrikanth Narayanan
类目: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Electroencephalography (EEG) measures brain activity through recording electrical potentials of the scalp. EEG recordings capture two distinct sets of characteristics in the signal: (1) \emphTraits – information that remains relatively constant over the course of a recording session, such as subject identity, age, and sex; (2) \emphStates – signatures that encode responses to external stimuli, including visual, auditory, and motor information. There is limited understanding of how different EEG modeling features (signal processing and foundation models) encode state and trait characteristics. In this work, we quantify state- and trait-level features captured by EEG embeddings using a hierarchy metric and kNN purity on Brain-Computer Interface dataset trials, and find that a majority of EEG features form a macro-trait hierarchy (embeddings form trait-level clusters with micro-states within). Our linear probe results indicate that this hierarchy structure leads to suboptimal state distinguishability, negatively impacting BCI performance of EEG foundation models. To mitigate this, we explore normalization techniques that reduce trait-specific information in embeddings and provide a \sim 10% improvement in balanced accuracy for state classification. Additionally, we observe that when fine-tuning models for downstream state decoding, embedding spaces transform from macro-trait to macro-state hierarchies, which in turn lead to better state decoding.
[AI-95] Attributing Agent Failures to Their History Needs an Executed Counterfactual: Toward a Reporting Standard NEURIPS2026
链接: https://arxiv.org/abs/2611.05650
作者: Egor Pakhomov,Erik Nijkamp
类目: Artificial Intelligence (cs.AI)
备注: Accepted at the IAB Workshop (Interpreting Agent Behavior) at NeurIPS 2026 (non-archival). 20 pages
Abstract:Published explanations of why long-horizon agents fail are commonly taxonomies labeled over completed trajectories: a judge reads a transcript and names the failure. We argue that a claim attributing an agent’s behavior to its history should carry intervention-verified evidence, per instance or via a validated estimator: freeze an executable mid-trajectory state, alter only the history the agent sees, and continue - executed rather than inferred from the transcript, at a state whose restoration has been measured, with a replay certificate matched to the estimand. Each of the four manipulation classes we name has a precedent, and several methods already execute interventions from checkpoints, re-executed prefixes or recorded conversation states (for example DoVer, REFLECT, Min et al., WhoWhen Pro, Causal Agent Replay), but none reports a measured replay certificate. Two cautionary measurements, neither applying the standard, follow: a carrier taxonomy whose “concentration” could not be separated from benchmark construction (adjusted odds ratio 1.15, interval crossing 1), and hash-identical snapshots that, restored into identically seeded processes, agreed with one another in 24 of 24 states while differing from the live continuation in 9 and 20 of 24. Restoring the captured position of a random stream made all 24 agree with the live run.
[AI-96] VIPBench: A Human-Aligned Benchmark for Voice Identity Perception in the Age of Voice Cloning NEURIPS2026
链接: https://arxiv.org/abs/2611.05638
作者: Xudong Tang,Matthew Groh
类目: ound (cs.SD); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Multimedia (cs.MM)
备注: Accepted at NeurIPS 2026, Track on Evaluations and Datasets (poster). 30 pages, 12 figures, 22 tables. Dataset and code: this https URL
Abstract:Speaker-recognition benchmarks test how accurately a model recovers the speaker label of a recording, treating voice identity as a property of the person who produced it. However, these benchmarks do not evaluate how humans perceive voice identity. Whether a listener hears two voices as the same person has consequences for intellectual property rights, cybersecurity, and everyday life. We introduce the Voice Identity Perception benchmark (VIPBench), which contains 124,876 same-or-different judgments from 1,290 listeners on 9,800 recording pairs from 100 public figures. The pairs include real recordings, voice clones generated by a text-to-speech system, and morphs between two speakers’ voices. We also define four tasks that evaluate the alignment between speaker-representation models and human identity perception. The first two tasks measure how well a model’s similarity score predicts the proportion of listeners who respond “same” and the majority response. The other two measure whether a model and listeners find the same speaker pairs similar, and whether a model’s relationship to listener judgments holds equally for real and synthetic recordings. We evaluate eleven publicly available speech models. Listener judgments diverge from speaker labels most often for voice clones: in 41% of responses, listeners judged a clone and a real recording of the same speaker to be different people. No model reaches the noise ceiling; the best model explains 61% of the explainable variance. Models trained with speaker labels align with listeners better than self-supervised models do, yet a model with a 12% error rate on speaker labels aligns with listeners as well as models with error rates below 2%. VIPBench enables the evaluation of speaker models against human perception on real, cloned, and morphed speech, and our results motivate training speaker models directly on listener judgments.
[AI-97] Route to Remember: Adaptive Memory Selection for LLM Agents
链接: https://arxiv.org/abs/2611.05636
作者: Zeyu Fang,Jun Yin,Yongsheng Mei,Xu Han,Chaitali Joshi,Zhongkai Sun,Jie Hao,Bin Guo,Chengyuan Ma,Benjamin Z. Yao,Chenlei Guo
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:The memory of production agents is increasingly becoming a bottleneck and is built almost exclusively from text: key points are extracted from out-of-limit context and retrieved into context at query time. Therefore, every new query incurs an additional cost that grows with the accumulated history. Weight memory is a complementary alternative that internalizes knowledge into model parameters and answers with no context at all, and has recently proven effective in self-evolving agents for internalizing skills and behaviors. However, its application in episodic and semantic memory, such as user-specific facts, remains comparatively unexplored, and a gap remains between earlier attempts at weight memory and text-based systems regarding recall performance. We study how to make weight memory competitive as a long-term factual store by applying on-policy self-distillation (OPSD). It trains a context-free student from a teacher, which is the same model conditioned on the user’s memory, together with a verify-then-regenerate router that lets weight and text memory each cover the queries they handle best. On the HaluMem benchmark, OPSD matches or surpasses strong text-memory baselines, including full-context prompting and retrieval systems, while reading fewer tokens per query. We further conduct an extensive study that characterizes the behavior of weight memory and analyzes how different training recipes affect its performance. As agent harnesses move toward persistent, multi-session deployments, these results position weight memory as a practical, low-cost component of agent memory and provide insights into feasible training and evaluation protocols for combining it with retrieval.
[AI-98] Assurance-Conditioned Continuous Authorization for Autonomous AI Agents
链接: https://arxiv.org/abs/2611.05632
作者: Mehdi Shajari
类目: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
备注: 81 pages, 3 figures
Abstract:Autonomous software agents can act through tools, delegate work, and change their models or execution environments while earlier authorizations remain in flight. We present Assurance-Conditioned Continuous Authorization (ACCA), an approach designed to be harness-independent, for maintaining operational authority according to currently justified assurance. It distinguishes evidence validity from contextual applicability, derives assurance claims from eligible evidence, and resolves a multidimensional effective-authority envelope within a principal-declared ceiling and applicable policy. A typed dependency graph supports selective re-evaluation of authorization-relevant changes. An authority epoch changes when effective authority changes; a distinct optional authorization-basis version supports revalidation of action-specific authorization conditions. Protected effect-time enforcement can reject stale actions before a consequential resource change, subject to the available commit mechanism. A compositional Domain Governance Profile expresses domain-specific semantics without granting authority. Controlled scenario tests implemented as a regression suite using a trading profile exercise default-deny rules, evidence-conditioned contraction and recovery, policy divergence, selective non-response, and stale-action rejection at a local gateway before the simulated-broker boundary. The study establishes implementation-level evidence in one domain, not a cross-domain guarantee of harmful-behavior prevention.
[AI-99] ViBench 2.0: A Benchmark for Long-Horizon Software Building
链接: https://arxiv.org/abs/2611.05620
作者: Daniel Furman,Preeya Kirani,Ibrahim Sheikh,Zhen Li,Peter Zhong,Michele Catasta
类目: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
备注:
Abstract:On Replit, our agentic engineering platform, users now build entire products without writing code, through successive feature requests to an agent. Yet existing coding benchmarks evaluate single, fully specified tasks, and many are near saturation. We introduce ViBench 2.0, a benchmark for long-horizon software building that we use to select our platform’s models. Its 17 full-stack web products, modeled on widely used software, grow over 241 feature requests and are judged only when finished, as a user would judge them: a browser agent runs 544 test plans, and a build passes only if every plan does, under a grading protocol validated to keep this strict score reliable. Evaluating nine models, we find a reliability gap: the three strongest pass up to 98% of test plans yet deliver a fully working product in at most 65% of builds, and no open-weight model delivers one. The gap is closing fast: across three releases of one model family, pass@1 on the public products rises from 3% to 53% while the cost per build falls by more than half. We release eight products with the harness and grader, hold out nine, and report lessons from production use.
[AI-100] Whiel: A Neurosymbolic Platform for Certified Relational Program Verification
链接: https://arxiv.org/abs/2611.05612
作者: Jesse Comer,Fangzhu Shen,Shuo Zhang,Mayur Naik,Sudeepa Roy,Val Tannen
类目: Databases (cs.DB); Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
备注:
Abstract:We present the Whiel (a contraction of Whi(le-R)el) system for Lean-certified reasoning about imperative relational programs. The system can encode containment and equivalence problems for Datalog, its extensions with negation, and the core fragment of graph query languages. It can also represent optimized or incremental evaluation algorithms that are not expressible in the original query language, as well as first-order specifications of integrity constraints, schema mappings, and views arising in data integration, exchange, and sharing. Whiel’s program states are database instances, and its imperative structure enables compositional Hoare-style reasoning with first-order relational specifications, using prophecy variables and leveled invariants to support inductive arguments. Whiel supports neurosymbolic discovery of loop invariants whose verification conditions can be checked using the Vampire first-order theorem prover. Counterexamples are fed back into the neurosymbolic pipeline, while proposed solutions are translated into Lean proof certificates. To validate the Whiel system, we evaluate its performance on a curated benchmark suite. In summary, we envision the Whiel system as a tool for tackling mathematically demanding, typically undecidable problems involving complex query and constraint languages, with a high degree of trust in its output.
[AI-101] Hybrid Spark: Adaptive Speculative Decoding That Lets the Content Choose the Drafter
链接: https://arxiv.org/abs/2611.05611
作者: Ekagra Ranjan,Bharat Venkitesh
类目: Artificial Intelligence (cs.AI)
备注: Speculative Decoding
Abstract:Speculative decoding (SD) speeds up large language model (LLM) serving, but its gain depends on the workload’s \emphshape, such as batch size and context length, and on its \emphcontent. A model-based drafter proposes a short block on any content, while a model-free drafter such as n-gram lookup proposes long copies at almost no cost when the output repeats its input, as in code editing. Adaptive verification spends a batch-wide verification budget on the draft tokens most likely to be accepted, but it supports only one model-based drafter. Model-free drafters provide no per-token confidence, so they cannot take part in budget allocation, and deployments must either fix one drafter or split their GPUs between drafters. We present \sys, a framework that runs a model-based and a model-free drafter on every step and gives each request the better of their chains, within adaptive verification and entirely on the GPU, which keeps asynchronous scheduling intact. The key idea is a price per verified draft token: the goodput that slot would earn elsewhere in the batch. Once the model-free drafter’s tokens carry calibrated acceptance estimates, fit once on general text, this price puts both drafters’ chains on one scale. Instantiated with \dspark and an n-gram drafter in vLLM, \sys achieves 1.10–1.62\x the output throughput of \dspark on code editing at batch size~4 and 1.04–1.27\x at batch size~128, while staying at 0.97–1.01\x of it on chat at temperature~0. A single deployment can therefore serve both workloads.
[AI-102] Verified but Vacuous: Three Failure Surfaces of LLM -Driven Verified Code Generation
链接: https://arxiv.org/abs/2611.05590
作者: Aditya Vempaty,Himaneesh Sompalle,Manan Tayal,Mowafak Allaham,Prasenjit Dey,Ashish Jagmohan,Ravi Kokku
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:What should ``verified’’ mean when evaluating LLM-generated code? Many existing pipelines report a single verification number. We argue this collapses three distinct properties measured in a conditional audit sequence: verifier acceptance (Surface A), \emphsyntactic specification binding (Surface B: does the proof obligation bind to the spec?), and \emphsemantic specification fidelity (Surface C: does the spec capture the informal requirement?). Two empirical demonstrations show the surfaces separate. \textbfFirst ( A \neq B ), via Li@Opus (a same-model reproduction of \citetli2025dafny’s Dafny-as-intermediate-representation (IR) pipeline under Claude Opus 4.7) on three Python benchmarks: it reaches \textbf99.1% verifier acceptance on HumanEval while only \textbf1.9% carry a certificate binding to a spec-defined predicate or function. \textbfSecond ( B \neq C ), via VCG on two Rust benchmarks: we use VCG (which is our own in-house Dafny-as-intermediate pipeline) on Rust because the Rust target lets us enforce fully-verifiable output (every component ships as a Dafny certificate, ensures-binding gate passing by construction). Even inside that construction, \textbf11.48% (MBPP-Rust) and 14.82% (VeriContest LC slice) of structurally-verified samples still fail hidden tests. On the LC slice of VeriContest, \textbfVCG’s reported End2End rate is 7.79% (535/6870). The cited external rates are 5.51% (GPT-5.5) and 1.88% (single-shot Opus 4.7). These give descriptive differences of 2.3 percentage points and 4.14 \times , respectively; task-budget, repair and harness differences prevent a matched-performance claim.
[AI-103] Anduin: A sybil-resistant agent ic marketplace reputation system
链接: https://arxiv.org/abs/2611.05586
作者: Reza Ramji,Preston Vander Vos,Ziming Zeng,Duc V. Le
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As autonomous AI agents enter permissionless onchain markets, selecting new service providers remains a core challenge. Existing reputation systems depend on scarce resources like human curation, identity credentials, or voluntary reviews, while naive volume rankings are vulnerable to Sybil attacks. We evaluate whether existing onchain payment histories can yield provider ratings that are resilient against Sybil attacks. Using a 157-day dataset comprising 66M agentic transactions worth 6.1M across 16 blockchains, we model payment flows as a bipartite graph between buyers and providers. A provider’s rating is derived from graph paths linked to a buyer’s existing trusted network. We propose three Sybil-resistant rating algorithms: a max-flow rating and two relaxed algorithms optimized for stronger rating performance. Evaluating these algorithms against simulated adversaries across 813 buyers, we show that rating manipulation costs 20% of those buyers’ total spend and is around 23x harder to manipulate than the standard similarity rule (collaborative filtering). Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2611.05586 [cs.AI] (or arXiv:2611.05586v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.05586 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-104] Readable Is Not Steerable: Pre-Registered Tests of Physical Structure in Learned Fluid Representations
链接: https://arxiv.org/abs/2611.05577
作者: Manvendra Modgil
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Fluid Dynamics (physics.flu-dyn)
备注: 30 pages, 9 figures. Code, pre-registrations and checkpoints: this https URL
Abstract:Self-supervised world models are often said to organize their latents around physical quantities. We test this on 2D decaying Navier-Stokes flow in three pre-registered studies, each tested once on held-out data. In Study 1, a joint-embedding predictive architecture (JEPA) trained to predict the flow fails the pre-registered criteria for organizing its latent around fluid quantities better than an untrained network of the same architecture (not supported); a JEPA trained on non-physical linear dynamics and a static invariance model that never sees time do as well or better. In Study 2, plain autoencoders that render the field make enstrophy prominent in their slowest unsupervised latent directions (best r 0.80-0.86 vs 0.38 untrained; fresh seeds, sealed data). Adding VICReg lowers this prominence in every comparison, though not by the pre-registered margin for vorticity, while raising the autoencoders’ linear decodability (partially supported). In Study 3, we build steerable directions (“dials”) by routing: the decoder gets overall scale only through a small conditioning code, while its other pathway carries a degraded copy of spectral shape. Label-free rules find a loudness dial in all 8 trained models and a shape dial in 3 of 5 fresh seeds on sealed data, against a pre-registered bar of 4; neither control shows a shape dial (partially supported). Across the studies, what a latent contains and what can be controlled through it come apart. In exploration, a supervised direction reading spectral shape almost perfectly moved nothing when steered. During the build, the loudness dial had to be found by its effect on the output (closest to a pure rescaling), not by data variance; for shape, a variance rule and an effect-constrained rule each worked in 3 of 5 seeds. The program used about 70 logged hours on one consumer GPU; all pre-registrations, deviations, code and outputs are public.
[AI-105] MESA: An Evaluation Framework for Compositional Semantic and Spatial Generalization in Robotics
链接: https://arxiv.org/abs/2611.05567
作者: Albert Wilcox,Frank Chang,Nhi Nguyen,Aishani Chakraborty,Jeremy A. Collins,Vaibhav Saxena,Benjamin Joffe,Siddharth Karamcheti,Animesh Garg
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注: 51 pages, 12 figures, including appendices. Conference on Robot Learning (CoRL), 2026. Project website: this https URL
Abstract:Despite growing interest in foundation models for robotic manipulation, systematic evaluation of their capabilities remains elusive. Existing benchmarks are either narrow and saturated, offering little signal on SOTA methods, broadly scoped, providing diversity at the cost of controlled experimentation, or require training on large, messy open-source datasets. We introduce MESA, an evaluation framework designed for precise, reproducible measurement of language-conditioned policy generalization. MESA provides five evaluation suites that isolate in-distribution performance and four distinct axes of generalization: spatial configuration, object instance, object category, and subtask composition, enabling researchers to diagnose where and why policies fail, not just whether they fail. To populate these suites at scale, we develop MESA-Gen, a pipeline for task and demonstration generation built by improving MimicGen. We instantiate the framework on two platforms and release both: the single-arm MESA suite, and BiMESA, the first large-scale benchmark for bimanual generalization. We evaluate seven baseline methods spanning diffusion policies, VLM-based approaches, and pretrained VLAs, finding that the strongest model achieves only 52% on MESA and 49% on BiMESA, providing meaningful signal while leaving ample room for progress. MESA rankings are broadly consistent with RoboArena scores, suggesting that MESA captures method-level differences relevant beyond simulation. Project website: this https URL.
[AI-106] VirtualTrialHarness: Coordinating Agent ic Operationalization of Virtual Trial Protocols Across Complex Clinical Schemas
链接: https://arxiv.org/abs/2611.05551
作者: Pengze Li,Jingna Feng,Yue Yu,Yining Wu,Xinyue Hu,Jie Lin,Avanti Bhandarkar,Cui Tao
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:Virtual trials use routinely collected clinical data to evaluate treatment strategies, but translating a study protocol into an executable analysis requires determining how clinical concepts and study requirements should be represented in the available data and applied consistently across the workflow. These decisions depend on local data representations and operational conventions and can involve source selection, patient and encounter identity, linkage, and temporal interpretation. Here we introduce VirtualTrialHarness (VTHarness), an agent framework that coordinates virtual-trial operationalization by selecting reusable procedural skills for each workflow stage and uses explicit contracts to maintain consistent study decisions across stages. We also introduce VTBenchmark for evaluating source selection and cohort construction. Across three complex synthetic environments, VTHarness successfully produced all required cohort outputs matching the references in 95.1% of workflows after requirement clarification, compared with 60.4% using retrieval from the same procedural corpus. Procedural skills improved source recovery across three language models and increased source overlap with analyst code in 13 Mayo Clinic cases. Site-specific adaptation enabled reference-matched cohort execution across eight MIMIC-IV tasks. Within the Mayo Clinic Virtual Trials Program, the framework supported investigator review of cohort construction and endpoint analysis. These findings support coordinating procedural guidance with workflow decisions to improve virtual-trial implementation across complex clinical schemas.
[AI-107] owards Efficient Whole Page Ranking with One-Pass Inference
链接: https://arxiv.org/abs/2611.05542
作者: Qi Xu,Zhigang Hua,Zijie Lei,Yan Xie,Jizhe Zhang,Lynn Lu,Delun Du,Sha Hua,Ren Mao,Ke Pan,Haomin Yu,Sandeep Pandey,Luke Simon
类目: Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
备注: 12 pages, 2 figures, 5 tables
Abstract:Whole-page ranking jointly optimizes advertisements and organic content to balance business outcomes and user engagement under complex layout and business constraints. Existing non-autoregressive (NAR) approaches enable efficient page generation but typically score candidate items independently of previously selected items, limiting their ability to capture order-dependent interactions. Autoregressive (AR) methods address this limitation through sequential conditioning, but incur substantial inference overhead that challenges large-scale deployment. Bridging this quality-efficiency gap remains a challenge in industrial whole-page ranking. We propose Multi-Item Scoring (MIS), a prefix-aware NAR-style framework that incorporates sequential context into efficient multi-item scoring. MIS integrates global candidate interactions, localized prefix conditioning, and position-aware scoring within a shared representation architecture. By computing reusable representations once per request and performing lightweight prefix-conditioned scoring during constrained beam search, MIS captures important sequential dependencies without repeated full-model inference. Experiments on KuaiRand-1K and RL4RS demonstrate that MIS approaches the generation quality of an autoregressive CGA baseline while achieving 4.07x and 3.54x faster CPU decoding, respectively. In a randomized evaluation on a large-scale social-video platform, MIS improves the realization metric by 0.104% and watch time by 0.058%, while reducing ranking-service CPU consumption by 9.73%. These results demonstrate that prefix-aware multi-item scoring can narrow the quality gap between NAR and AR generation while retaining the efficiency required for industrial scalability. Comments: 12 pages, 2 figures, 5 tables Subjects: Artificial Intelligence (cs.AI); Machine Learning (stat.ML) Cite as: arXiv:2611.05542 [cs.AI] (or arXiv:2611.05542v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.05542 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[AI-108] SFM Meets LLM : Context as Covariates
链接: https://arxiv.org/abs/2611.05515
作者: Hemanth Ram,Sai Shankar Narasimhan,Shubhankar Agarwal,Sandeep Chinchali
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Textual information is ubiquitous in real-world time series datasets, appearing as news, reports, and annotations that hold key predictive signals not deducible from numerical history alone. Large Language Models (LLMs) interpret text but lack numerical precision, while Time Series Foundation Models (TSFMs) are quantitatively accurate but cannot reason about the effects of text. Moreover, the lack of large-scale paired multimodal datasets limits progress in bridging this gap. In this work, we address this challenge by providing a scalable approach for generating realistic multimodal datasets by annotating real-world time series with LLMs. Building on this, we propose Context as Covariates (CoCo), a forecasting framework in which an LLM distills its textual reasoning into numerically grounded forecast and confidence covariates that guide a TSFM backbone, thereby exploiting the strengths of both LLMs and TSFMs. We align the two models in stages, first fine-tuning the LLM via Group Relative Policy Optimization (GRPO) with a reward tied to covariate-induced improvement in forecasting performance, and then fine-tuning the TSFM on covariates generated by the aligned LLM. Our experiments show that models trained on our synthetic multimodal corpus generalize to real benchmarks spanning finance, economics, weather, traffic, and security. Notably, CoCo with Qwen3-4B-Instruct-2507 as the LLM and Chronos2 as the TSFM achieves up to 12% MAE improvement over state-of-the-art TSFMs and LLM forecasters, such as GPT-4o. Code, checkpoints, and the evaluation datasets are publicly available.
[AI-109] Objective-Conditioned Active Preference Learning for Multi-Objective Alignment
链接: https://arxiv.org/abs/2611.05506
作者: Duc Anh Vu,Thong Nguyen,Hoang Pham,Ponhvoan Srey,Luu Anh Tuan
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Aligning large language models (LLMs) with several objectives, such as helpfulness, harmlessness, and honesty, typically relies on preference judgment for each objective, growing the annotation cost linearly with the number of objectives. Active preference learning can reduce this cost by selecting the comparisons whose labels are expected to change the policy most. Existing remedies either lack an objective-aware query selection mechanism for the multi-objective setting or regard conflicting preferences as harmful, therefore excluding them from the training data. We introduce OGC-DPO (Objective-conditioned Gradient Coverage DPO), an active learning framework that treats each (response pair, objective) combination as a separate query and measures its informativeness through the policy’s own implicit gradient, without relying on an external reward model. OGC-DPO then selects the queries that most expand each objective’s gradient coverage. Empirical experiments across varying numbers of objectives demonstrate the effectiveness of OGC-DPO over existing active preference learning methods under the same annotation cost. Further analysis demonstrates that informative conflicts, rather than being uniformly harmful, provide a useful policy training signal.
[AI-110] Can Large Language Models Review the Scientific Literature? A Test Against a 102-Article Benchmark
链接: https://arxiv.org/abs/2611.05459
作者: Abdullah Al Kalbani,David M. Schultz,Alexandra K. Anderson-Frey,Zhonghua Zheng,Catherine McGuire
类目: Artificial Intelligence (cs.AI); Digital Libraries (cs.DL)
备注: 34 pages, 7 figures, 6 tables. Data, prompts, analysis scripts and results: this https URL (GitHub: this https URL )
Abstract:Large language models (LLMs) are increasingly used to support scientific literature reviews, yet evaluations often examine individual tasks rather than the review process as a whole. We assess six AI tools across four stages: finding relevant articles, interpreting them, synthesizing across them, and proposing future research. The benchmark is a published review of the Spanish plume weather pattern and its 102-article corpus. In the initial searches, five tools collectively identified 57 articles, with the best individual tool finding 42. Institutional library access increased discovery from 42 to 82 articles for Codex and from 15 to 79 for Claude Code. Given the complete corpus of 102 PDFs, three tools correctly identified all 18 articles discussing the Spanish plume substantially, but incorrectly classified 40% to 79% of the 84 passing mentions as substantial. The same three tools produced syntheses evaluated blindly by two domain experts. Although largely accurate, the syntheses provided limited insight and sometimes contained unsupported claims, quotation errors, and exaggerated statements of consensus. Finally, all three tools proposed research directions broader than the Spanish plume literature, although two approached questions identified in the benchmark review. No single tool performed best across all tasks. We provide eight recommendations for researchers using AI tools in literature reviews. Our findings suggest that LLMs can assist with discovering and interpreting scientific literature, particularly when institutional access is available, but cannot replace expert judgment in evaluating evidence or synthesizing research. Comments: 34 pages, 7 figures, 6 tables. Data, prompts, analysis scripts and results: this https URL (GitHub: this https URL) Subjects: Artificial Intelligence (cs.AI); Digital Libraries (cs.DL) MSC classes: 68T50, 68P20, 68T07 Cite as: arXiv:2611.05459 [cs.AI] (or arXiv:2611.05459v1 [cs.AI] for this version) https://doi.org/10.48550/arXiv.2611.05459 Focus to learn more arXiv-issued DOI via DataCite
[AI-111] Hybrid LLM -Deterministic Pipelines for Enterprise-Scale Code Migration at Bloomberg
链接: https://arxiv.org/abs/2611.05457
作者: Mario Corchero,David Williams,Laszlo Kiss Kollar,Haowen Liu,Shivam Bairoliya,Gala Malbasic,Serkan Kirbas,Federica Sarro
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注:
Abstract:Large organizations carry a long tail of infrastructure migrations that are worth doing but rarely urgent enough to compete with feature work, because each one costs significant engineering time per repository. This paper reports on our industrial experience from Automated Uplift, a multi-month initiative applying LLM-assisted migration pipelines across thousands of Bloomberg repositories. We examine two distinct production campaigns: a heterogeneous Python packaging migration (approx. 1,900 candidate repositories) and a uniform C++ build system migration (approx. 6,700 candidate repositories). Together, these campaigns generated 2,333 pull requests, resulting in 729 successful migrations. We detail our hybrid architectural solution which maximizes deterministic processing, confines the LLM to bounded file transformations within a pre-computed context, and incorporates runtime production telemetry (such as execution logs and usage traces). Our findings show that generating migrations turned out to be the easier part; the bottleneck was the cost, for owning teams, of reviewing, testing, and deploying them. These costs were lowered when changes could be verified directly in the pipeline before opening a PR (as with C++ compilation and test validation), and when the population was uniform enough for fixes to carry over from one repository to the next. We conclude by distilling our experience and findings into actionable, transferable patterns for architecting hybrid LLM-deterministic pipelines and scaling code migrations across the enterprise
[AI-112] owards Robust Long-Lived Agents : Benchmarking Experience Lifecycles in Evolving Environments
链接: https://arxiv.org/abs/2611.05450
作者: Shijue Huang,Hangyu Guo,Hengheng Zhang,Guo An,Zhitao He,Qi Tian,Yiren Feng
类目: Artificial Intelligence (cs.AI)
备注:
Abstract:As agents become more capable, they are moving beyond one-off tasks toward sustained assistance across successive user goals. Such assistance requires agents to discover how unfamiliar environments work, reuse what they learn across tasks, and keep that experience current as environments evolve. Existing benchmarks increasingly study experience reuse and adaptation, but provide limited insight into which parts of accumulated experience agents revise, retain, or lose. To address this gap, we introduce ReScopeBench, a benchmark for the experience lifecycle across 18 interactive environments. It distinguishes structural experience about how an environment is organized and accessed from rule experience about reusable principles and constraints that guide decisions across tasks. To study how experience is acquired, reused, and revised, we propose an experience-first construction approach. It builds task streams around shared experience and introduces silent environmental changes. We evaluate task performance and track changes in saved experience throughout each stream. Across 20 models, we find that retained experience improves task performance, but agents struggle to keep it current as environments change. Even when the required experience is correctly recorded, agents do not always complete tasks successfully. These findings identify selective maintenance and effective use as complementary challenges for long-lived agents.
[AI-113] Jev with Foresight: Safe Language-Guided Autonomous Driving via a Language World Model
链接: https://arxiv.org/abs/2611.05448
作者: Yansong Qu,Jiancong Chen,Samuel Labi
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI)
备注:
Abstract:Large language models understand passengers’ driving commands, but as driving agents they reason in generated text for seconds and must infer each action’s consequences implicitly from the scene. We present Jev-AD, which pairs Jev, a typed decision model that answers predefined questions with probabilities, not generated text, with a language world model that predicts for every candidate action how nearby vehicles move over the next two seconds. Code turns these predictions into foreseen consequences, which Jev compares instead of inferring them from coordinates. In simulated highway merging (500 car-episodes), this foresight lowers the same decision model’s collision rate from 26.0% to 2.4% at 0.25 seconds per decision, and Jev-AD achieves the passenger’s intent in 93% of free-form command trials. In 500 replayed real-world NGSIM episodes, foresight lowers the collision rate from 48.2% to 2.8%, and the world model, learned from this traffic, foresees 93% of dangerous actions. Project page: this https URL.
[AI-114] Constraint-Preserving Flow Matching for Spectral Graph Generation
链接: https://arxiv.org/abs/2611.05436
作者: Anita Yang,Ken-ichi Kawarabayashi,Yohan Jung
类目: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
备注:
Abstract:Spectral graph generation models graphs through the eigenvalues and eigenvectors of their Laplacians, providing a compact and continuous alternative to direct graph construction. However, these eigensystems must retain the structural properties of graph Laplacians. Existing approaches preserve eigenvector orthonormality but do not enforce the Laplacian nullspace constraint, potentially producing eigensystems inconsistent with Laplacian structure. To address this, we introduce Laplacian-Constrained Stiefel Flow Matching (LaSFM), which preserves the nullspace condition alongside eigenvector orthonormality and positive eigenvalues during generation. We characterize the underlying constrained Stiefel manifold and formulate flow matching on it. We further extend LaSFM to property-guided generation, steering samples toward desired graph properties while maintaining these constraints. Experiments demonstrate improved performance on both spectral and graph-level metrics, along with better alignment with target properties.
[AI-115] LLM -based NPCs in Video Games: A Systematic Mapping Study
链接: https://arxiv.org/abs/2611.05427
作者: Dilan Mian,Jeremy Bradbury,Yann-Gaël Guéhéneuc,Fabio Petrillo,Cristiano Politowski
类目: oftware Engineering (cs.SE); Artificial Intelligence (cs.AI)
备注: 12 pages. Accepted manuscript. Published in IEEE Transactions on Games (Early Access)
Abstract:For decades, non-player characters (NPCs) have been driven by rule-based systems whose personalities and interactions are scripted before release. Large language models (LLMs) promise NPCs that converse openly, adapt to new situations, and behave in ways developers did not fully author, prompting a wave of research and commercial experimentation. Yet the field lacks a unified picture of what has been built, how well it works, and at what cost. We address this gap with a systematic mapping study of 377 papers from the academic literature, yielding 42 papers analysed against a main research question and ten sub-questions across four thematic axes. We provide the first quantitative characterisation of component adoption, latency, safety coverage, and reproducibility: short-term memory and world-state grounding are near-universal (60% and 69%), whereas long-term memory (31%), reflection (38%), and retrieval-augmented generation (14%) remain rare, and response times span 2.6 to 15 seconds. Finally, we identify six structural barriers to commercial adoption and offer evidence-based recommendations for developers and researchers.
[AI-116] CAST-Bench: Tracing Every Sound to Its Source in Multi-Shot Audio-Video Generation
链接: https://arxiv.org/abs/2611.05422
作者: Yucong Liu,Qi Tian,Shuyuan Tu,Kaihang Pan,Weijie Kong,Chao Pan
类目: ound (cs.SD); Artificial Intelligence (cs.AI)
备注: 9 pages ,6 figures
Abstract:Existing benchmarks for joint text-to-audio-video (T2AV) generation evaluate perceptual quality and audiovisual alignment, but balanced sound-source sam?pling and separate reporting by shot structure remain limited. This makes it dif?ficult to determine whether a specific sound is missing, attributed to the wrong source, or lost across shot transitions. To address this, we introduce CAST?BENCH, a benchmark designed to trace every generated sound back to its in?tended source in multi-shot audio-video generation. The core idea is to formu?late T2AV correctness as an audiovisual contract in which each requested sound must be emitted by the right source, synchronized with the corresponding visual action, and carry the right semantic content. Concretely, the benchmark orga?nizes 900 prompts into 9 sound-source categories, with 100 prompts per cate?gory, spanning speech, animals, instruments, ambient sounds, and other emitter types. The prompts include 439 single-shot and 461 multi-shot narratives, each derived from a structured specification that links every audio event to its emit?ter, subject, and temporal slot. The benchmark provides an 18-metric evaluation suite covering three dimensions: Visual Performance, Content Consistency, and Audio Semantic Alignment. We further benchmark 12 representative gen?erators and find that visual quality is nearly saturated across leading systems, yet audio semantic scores still differ markedly. Across the nine shared mod?els, multi-shot prompts receive lower aggregate scores under the subset-specific scoring rules and yield different model rankings. Multi-person speech, con?sistent speaker identity across shots, and cross-shot event retention remain the main unresolved challenges. All prompts and evaluation code are available at this https URL.
[AI-117] A Feasibility Condition for Confidence-Gated Tiered LLM Inference
链接: https://arxiv.org/abs/2611.05394
作者: Sumit P. Ahuja
类目: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF)
备注: 21 pages, 2 figures
Abstract:Tiered inference distributes large language model computation across a hierarchy of devices, from on-device and edge hardware to the datacenter, and is increasingly offered as the route to cheaper, lower-latency, and more private serving. Most such systems orchestrate frozen model checkpoints, and most decide what to resolve locally and what to escalate using a confidence signal produced by a forward-deployed model. We give a closed-form condition for when no threshold on that signal makes a tiered cascade worthwhile relative to sending every query to the backstop. Modeling the routing decision as a binary test and separating local from escalation cost through an architecture cost ratio \kappa , we derive a feasibility condition bracketed by a signal-free floor, below which a cascade is feasible at any signal quality, and a perfect-signal ceiling, above which none is. Read along the \kappa axis the condition inverts: where local compute is billed, cost binds, matching a recent empirical finding; where it is sunk, on decentralized hardware, feasibility becomes unconditional and signal quality instead sets what the cascade achieves. We turn the condition into three design requirements that existing trained-in primitives satisfy only in isolation, and we propose, as future work, a co-designed architecture and a protocol to test it. The condition and the inversion hold independently of that architecture; they are the paper’s contribution.
[AI-118] Behavior Graphs: Unsupervised Graph Structures for Interpreting Robot Deployment Data
链接: https://arxiv.org/abs/2611.00405
作者: Erik Bauer,Jakob Thumm,Christopher Agia,Rohan Sinha,Chen Xu,Haruki Nishimura,Jeannette Bohg,Marco Pavone,Masha Itkina
类目: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: 8 pages, 6 figures, 3 tables. Project website: this https URL
Abstract:Data flywheels promise continuous improvement via insights from deployment data. Understanding how the deployed model succeeds and fails can steer subsequent data collection towards patching failures and reinforcing successes. However, robot deployments create large volumes of high-dimensional rollout data, which is challenging to understand and confidently ground data collection decisions in. To make deployment data easier to analyze and act on, we propose behavior graphs, which compress policy rollouts into a single interpretable, directed graph. We construct the graphs by learning to identify similar behavior modes across rollouts, connecting the behaviors temporally, and finally linking them to outcomes. In a user study on three real-world tasks, where participants analyzed rollouts to choose a data collection strategy, 16 of 18 preferred behavior graphs to rollout videos, reaching comparable conclusions with 40.3% lower perceived temporal demand and 37.6% lower effort. Beyond interpretability for deployment data, behavior graphs are capable progress predictors: across 43 datasets, they improve the rank correlation between predicted progress and true temporal order by 22.9% and reduce mean absolute error by 53.5% relative to the strongest baseline. As the number of deployed robots grows, behavior graphs lend structure to high-dimensional policy rollouts and make deployment data easier to interpret and reason about.
[AI-119] Noise-regularized training of quantum circuit Born machines under a finite shot budget
链接: https://arxiv.org/abs/2611.06630
作者: Hoang M. Ngo,Quan Nguyen,Jung Taek Seo,Wanli Xing,My T. Thai
类目: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Quantum circuit Born machines (QCBMs) can learn distributions that are hard to sample from classically, but their value depends on generalizing from finite training data. Here, we study how hardware noise affects this generalization. We show that, for any noise process, the generalization gap of a QCBM trained with the maximum mean discrepancy is bounded by the complexity of the distributions the noisy model can reach. Scrambled local Pauli noise contracts this complexity but biases the model towards the uniform distribution, so residual noise acts as a regularizer whose optimal strength can be intermediate. We propose GENIE, a framework that realizes a chosen noise level with capped error mitigation and selects it during training from held-out data within a fixed execution budget, with provably bounded regret. In simulations on synthetic and census data, GENIE stays close to the best of four noise-control strategies, whereas every fixed noise level falls short in many settings.
[AI-120] Does Frontier AI Strengthen Cyber Defense? Evidence from Project Glasswing
链接: https://arxiv.org/abs/2611.06629
作者: Thomas Mosk
类目: General Economics (econ.GN); Artificial Intelligence (cs.AI)
备注:
Abstract:Defenders and attackers race to find the same software flaws. I use the launch of Project Glasswing, which gave a small group of large US tech firms early access to Claude Mythos Preview, a model for finding software vulnerabilities, to study how access to frontier AI affects defenders, external researchers, and attackers. Using the universe of publicly disclosed vulnerabilities, I find that firms with access more than doubled their disclosures relative to Chinese and Russian tech firms without access, disclosed about 150 percent more high-impact vulnerabilities, and fixed reported flaws as quickly as before. As treated firms found more flaws themselves, external researchers, who sell their findings through bug bounties, reported about half as many vulnerabilities in treated firms’ code and more in other firms’ code, without leaving the market. There is no evidence that attackers exploited the additional vulnerabilities: the share exploited within a month of disclosure fell by about two thirds, mainly because a smaller share was already exploited when disclosed, while the number exploited did not change. Frontier AI helped defenders find and fix more flaws before attackers exploited them, but its restricted release creates an asymmetry in software security between firms in different countries.
[AI-121] Microdroplet Generator Surrogate Enhancement and Inverse Design Through an Uncertainty-Guided Framework
链接: https://arxiv.org/abs/2611.06464
作者: David Gabriel Harispe,Matías Gerard,Pablo A. Kler
类目: Fluid Dynamics (physics.flu-dyn); Artificial Intelligence (cs.AI)
备注:
Abstract:Microfluidic droplet generation is critical for many applications, yet efficient device design in high-dimensional parameter spaces remains computationally prohibitive due to complex multiphase flow dynamics. This work presents a unified methodology that couples adaptive surrogate enhancement with simulation-driven inverse design. High-fidelity Volume-of-Fluid simulations provide ground-truth data, while neural network ensembles predict flow regimes and droplet diameters. Ensemble-based uncertainty quantification identifies epistemic uncertainty using consensus levels for regime classification and the coefficient of variation for diameter prediction, revealing that uncertainty concentrates near regime transition boundaries where physics is highly sensitive to parameter perturbations. A physics-informed enrichment strategy targets these critical zones and addresses class imbalances by deploying additional targeted simulations and retraining the surrogates forming the ensembles. A framework that integrates Particle Swarm Optimization for constrained inverse design, balancing accuracy with predictive confidence and physical consistency is therefore proposed. Validation on independent test datasets demonstrates that the enriched ensembles achieve higher accuracy, with the fraction of predictions within the 0-3% error range on droplet diameter nearly doubling, and high-error cases (above 10%) dropping substantially. The inverse design procedure reliably identifies physically realizable configurations meeting target droplet sizes. This approach significantly reduces computational costs while ensuring reliable predictions, offering a framework with potential applicability to the optimization of a broad range of microfluidic systems.
[AI-122] Mind Control of LLM s via Closed-Loop Stimulation: Out-of-Distribution Robustness
链接: https://arxiv.org/abs/2611.06418
作者: Kasra Fallah,Rudramani Singha,Hannah Ghanei,Paul Sajda,James Anderson,Erfan Zabeh
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
备注:
Abstract:Brains and large language models need not be considered equivalent systems to reveal an interpretable commonality: both exhibit high-dimensional, nonlinear dynamics whose behavior can only be modeled imperfectly. This creates a shared challenge whenever we seek to intervene in their dynamics. A controller may either rely on a model identified around a particular operating regime and risk failure as the system moves away from that regime, or explicitly account for uncertainty and a changing internal state. Neurostimulation has confronted an analogous problem for decades. The limitations of fixed, open-loop stimulation have motivated closed-loop and state-dependent approaches that condition intervention on the ongoing state of the neural system. Inspired by this principle, we formulate LLM steering as a feedback-control problem in which the network’s internal state, rather than a fixed intervention alone, determines how the system is perturbed. We further account for the mismatch between tractable dynamical models and the nonlinear transformer they approximate through robust control. As one principled realization of this framework, we develop an H_\infty -based steering controller that limits sensitivity to disturbances and linearization mismatches. Across multiple steering benchmarks, the controller improves performance over existing baselines, with particularly pronounced gains under out-of-distribution conditions and varying length complexities. Beyond controller performance, we introduce robust steerability, S_\mathrmrob , a pre-deployment measure of tolerance to dynamics mismatch; across model-task pairs, it exhibits task-specific structure and varies systematically with model scale. Beyond robust LLM steering, our measure asks not merely whether control is possible, but how much uncertainty it can survive; a principle that may span artificial and biological neural systems.
[AI-123] Does Your Trial Simulator Know What the Drug Does? An Intervention-Blind Test of Interventional Validity on 7207 Registered Trials ALT NEURIPS2026
链接: https://arxiv.org/abs/2611.05907
作者: Alicem Koyun
类目: Applications (stat.AP); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注: Oral presentation at the NeurIPS 2026 Workshop on World Models for High-Stakes Health. 31 pages, 5 figures. Code and data: this https URL
Abstract:Patient world models are increasingly proposed as simulators of clinical trials, and recent surveys rank their capabilities on a ladder from temporal prediction to counterfactual rollouts. Position on that ladder is currently assigned by reading papers rather than by measurement. We turn the counterfactual rung into a measured quantity. From 30,119 completed phase 2/3 trials with results posted to this http URL we construct 9,996 audited within-trial contrasts from 7,207 trials, and we score a simulator on how much of its accuracy depends on knowing which intervention was tested. The instrument is an intervention-blind reference: the same model, trained on the same trials, with every intervention-identifying field withheld. Interventional skill is the error the full model removes relative to that reference. On a temporal split (878 test contrasts from 655 trials) a registry-trained gradient-boosting simulator removes a fifth of the error of a trivial baseline, but its interventional skill is -0.12 (95% CI -0.21 to -0.05): withholding the intervention makes it better, and no alternative featurisation, stratum, temporal cut or same-molecule retrieval yields an interval above zero. A frontier language model reading only the registered protocol has interventional skill 0.20 [0.08, 0.31]; smaller and open-weight models sit at the limit of detection. Perturbation tests show the large model uses the intervention as such: swapping the arms reverses the sign of its prediction in 86% of contrasts. The blind reference is itself not drug-blind (a language model recovers the withheld drug class 45% of the time from the blind prompt), so interventional skill is a conservative lower bound. A reliability layer shows that ranking by predicted error lowers it but that calibrated error control does not transfer under temporal shift. Benchmark, code, prompts and per-call model outputs are released.
[AI-124] CellCanon: An Agent ic Cellular World Model for Prediction Explanation and Observation-Driven Updating
链接: https://arxiv.org/abs/2611.05890
作者: Zhaokang Liang,Shuyang Zhuang,Xiaoran Jiao,Hao Chen
类目: Quantitative Methods (q-bio.QM); Artificial Intelligence (cs.AI)
备注: 31 pages, 5 figures. Code: this https URL
Abstract:Biological discovery requires using molecular knowledge and new measurements to infer responses beyond the conditions already observed. We introduce CellCanon, the first agentic cellular world model to unify single-cell expression prediction, evidence-grounded explanation, and observation-driven updating. Its predictive core reuses condition-level biological evidence across gene-specific numerical readouts to guide single-cell expression generation. A shared world state links predictions, acquired measurements, and source-linked claims, supporting explanation and revision. On 339 held-out SciPlex3 conditions across three cell lines, CellCanon improves differential-expression AUPRC and Macro-F1 over the best baseline for each metric by 63.3% and 17.3%, respectively, and direction Macro-F1 by 7.2%. Relative to gene-wise reasoning on matched drug-cell cohorts, measured language-model token use falls by 99.77% for Predict and 98.41% for Explain. On the source-grounded benchmark, Explain correctly answers all answerable questions with supporting proofs and identifies all insufficient-evidence cases. In simulated experimental feedback, Update uses one measured source context per drug to exceed the best fine-tuned baselines in differential-expression AUPRC and Macro-F1 without changing model weights. Together, these capabilities support a cumulative account of cellular responses, with predictions and biological judgments that can be revised as evidence accumulates.
[AI-125] Unsupervised Neural Operators for Rotating Relativistic Stars: High-Fidelity Equation of State Inference and Pulsar Spin Evolution at Scale
链接: https://arxiv.org/abs/2611.05790
作者: Anarya Ray
类目: High Energy Astrophysical Phenomena (astro-ph.HE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); General Relativity and Quantum Cosmology (gr-qc)
备注:
Abstract:Observed populations of rotating neutron stars~(NSs) with millisecond-order periods are powerful probes of fundamental physics. Mass-radius measurements inform poorly constrained properties of degenerate nuclear matter, whereas high-fidelity population synthesis connects unknowns in NS formation and binary evolution to large radio-timing surveys. However, repeated structure calculations are prohibitive due to the computational cost of solving the highly coupled partial differential equations of axisymmetric general relativistic hydrostatic equilibrium that govern rotating NS interiors. Non-rotating calculations or approximate rotational corrections can bias equation of state inference and limit the fidelity of population synthesis in tracing the evolutionary history of progenitors and propagating equation of state uncertainties. In this \textitLetter, we develop physics-informed neural operators that can solve the axisymmetric problem with high accuracy and negligible computational cost. \textitOur solvers are trained in a fully unsupervised setting, with the training objective informed solely by the partial differential equations, and achieve a 10^3\times speed-up with negligible accuracy loss. These solvers enable scalable EoS inference that self-consistently accounts for rotation and repeated stellar-structure calculations during binary evolution, making spin predictions that account for rotational structure and EoS uncertainties computationally accessible to population synthesis.
[AI-126] VenusGenome: Genome-Aware Masked Autoencoding for Microbial Genome Representation Learning
链接: https://arxiv.org/abs/2611.05405
作者: Zhang Liang,Zhang Yukun,Tan Yang,Jiang Fan,Zhong Bozitao,Cheng Xiaoran,Hong Liang,Li Mingchen,Yu Yuanxi
类目: Genomics (q-bio.GN); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
备注:
Abstract:Genomic-context models have shown that contextualizing protein embeddings can improve predictions of microbial function. We observe that a gene’s protein embedding is partially predictable from the embeddings of neighboring genes in the genome, motivating high-ratio masking that requires reconstruction from sparser genomic context. Genomic modeling must also distinguish biological relationships from assembly conventions, since the ordering of separate contigs does not imply adjacency and strand labels depend on the reference orientation. We therefore introduce VenusGenome, a gene-resolution masked autoencoder that combines visible-only encoding with genome-aware attention incorporating contig boundaries, relative strand orientations, and gene positions. We pretrain this model on a curated corpus of over 1.7 million microbial genome assemblies from isolate and metagenomic sources, using frozen protein embeddings as gene-level inputs. Across genome-level phenotype prediction and gene-level tasks, the resulting frozen representations achieve the highest mean performance among the evaluated models on nine of eleven primary metrics. An exploratory gene-input perturbation analysis further links representation sensitivity to experimentally measured fitness defects. Ablations support the contribution of genome-aware attention and show the benefits of masked autoencoding over BERT-style masked modeling, with further gains from higher masking ratios.
[AI-127] A Training-Free Socket for DNA in Frozen Language Models
链接: https://arxiv.org/abs/2611.05401
作者: Arash Fatehi
类目: Genomics (q-bio.GN); Artificial Intelligence (cs.AI)
备注:
Abstract:Connecting a DNA foundation model to a large language model (LLM) usually requires training a connector, and often the LLM too. We show that a closed-form map is enough for gene-level questions. Our method, the socket, fits one ridge regression (the plug) from frozen Evo 2 features to the host LLM’s own representation of the gene’s symbol, name, and description at an early layer. The plug’s output replaces the hidden state at placeholder positions, and no gradient passes through the host. We test it on GeneHop, a five-way multiple-choice benchmark split by homology. In a single run on the held-out test set, under a frozen analysis plan, real DNA beats shuffled DNA on all four hosts: by +22.2 points [19.8, 25.0] on Qwen3.5-4B, +26.1 [24.3, 27.9] on Qwen3.6-27B, +16.7 [14.2, 19.7] on Mistral-7B, and +18.4 [16.2, 21.0] on OLMo-2-7B. All four survive Holm correction, every null control stays within two points of the sham level, and the development results agree within 2.3 points. A layer rule that uses no task items was selected on three hosts; it predicted the working layer of the fourth before any DNA-plug scoring on it. A text-only linear map lets one plug feed any other host, at a cost of at most 0.9 points on test. The socket conveys gene identity: nearest-neighbour retrieval recovers 72-84% of the effect, but the plug beats it on every host by 3.0-5.2 points and keeps most of its effect without training homologs. Subjects: Genomics (q-bio.GN); Artificial Intelligence (cs.AI) Cite as: arXiv:2611.05401 [q-bio.GN] (or arXiv:2611.05401v1 [q-bio.GN] for this version) https://doi.org/10.48550/arXiv.2611.05401 Focus to learn more arXiv-issued DOI via DataCite
[AI-128] Do agents need to sleep?
链接: https://arxiv.org/abs/2611.02216
作者: Ido Haber,Giulio Tononi
类目: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI)
备注:
Abstract:LLM-based agents, which increasingly operate persistently across sessions and accumulate memory through interaction, already exhibit the familiar symptoms of unchecked plasticity: semantic drift, degraded retrieval precision, propagation of stored errors, and saturation of memory capacity. Drawing on the selectionist “ringing, resonance and renormalization” (RRR) account of sleep, we propose that each of its three phases maps onto an operation feasible in digital memory architectures. Engineers building stateful agents, meanwhile, are independently rediscovering the problems sleep evolved to solve: what to forget, and when to do the forgetting. From this we derive the falsifiable prediction that selectionist consolidation will outperform instructionist alternatives as memory stores grow large. The core claim concerns the nature of continuous learning itself: plasticity has a price, and for any system that learns continuously under finite capacity while coupled to its environment, that price is sleep.
机器学习
[LG-0] Reinforcement Learning Enables Scaling Generalist Robot Policy Improvement
链接: https://arxiv.org/abs/2611.06905
作者: Michal Korniak,Marco Bagatella,Ralf Römer,Michal Nauman,Yarden As,Pranav Vignesh,Manan Agarwal,Leander Diaz-Bone,Piotr Zalewski,Mikołaj Czarnecki,Lenart Treven,Shashwat Saxena,Sreyas Venkataraman,Sarvesh Patil,Zhiyuan Zhou,Daniel Marta,Oier Mees,Nur Muhammad Mahi Shafiullah,Carmelo Sferrazza,Bhavya Sukhija,Angela P. Schoellig,Andreas Krause,Max Simchowitz,Andrew Wagenmaker
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注:
Abstract:The introduction of generalist robot policies has enabled a single, large model to learn to perform a variety of tasks from a vast pre-training corpus. While these models often exhibit powerful zero-shot generalization capabilities, they typically require an adaptation phase to achieve sufficient performance in deployment. Iterated self-improvement via reinforcement learning (RL) has demonstrated promise in enabling this adaptation, yet such approaches have been largely explored in single-task settings - where the goal is to simply improve on a narrow deployment task - which lies in stark contrast to the multi-task capabilities of the base policy. In this work we ask whether generalist robot policies are capable of multi-task, generalist improvement in deployment - targeting many diverse tasks simultaneously - and what algorithmic recipe enables this. To answer this question, we first develop a set of structured evaluations to test generalist improvement capabilities, and systematically evaluate algorithmic and practical design choices regarding value estimation, policy extraction and resource allocation. We show that not only can generalist policies be improved online across diverse tasks, but that the right recipe can enable generalization and increase performance on tasks not considered during online training. We characterize when this generalization occurs and show that both the training task distribution and the use of value-based RL are critical. To the best of our knowledge, this is among the first evidence for online generalist improvement for robot foundation models.
[LG-1] Adaptive Multi Mode: Budget-Aware Routing of Optimizers By Online Landscape Probes
链接: https://arxiv.org/abs/2611.06879
作者: Benjamin Jacob Bodner
类目: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Optimization and Control (math.OC)
*备注: 29 pages, 8 figures, 19 tables. Code: this https URL
Abstract:The best classical optimizer for a continuous black-box problem depends on its budget and landscape. Existing routers decide from the dimension and budget, past improvement or features fitted to one benchmark, and few test an optimizer’s implicit assumptions online. AMM (Adaptive Multi Mode) routes a known budget among quasi-Newton, trust-region, coordinate, multistart and evolution-strategy optimization engines with online probes of curvature, plateaus, separability, kinks, basins and fractal structure. Each probe has a default, a dimension-by-budget validity range and a cost charged to the budget. In its pre-registered development loop on bbob, 32 of 36 predictions held over the last 16 confirmations. AMM was then tested once on held-out bbob instances 101-110 (D=2-40, 10D-1000D) against up to seven competitors. It has the highest target share in 382 of 600 function-dimension-budget cells (236 outright, 146 tied) and, by a descriptive posterior rule, is credibly behind in 35, mostly at 10D and 30D. At every budget it also leads the controls built from its own engines (the best engine alone, the best engine per dimension and budget, and a greedy schedule), narrowly at 10D and 30D, and nevergrad’s NgIohTuned. In pre-registered runs its lead over IPOP-CMA-ES holds under affine and box transformations, on box-constrained bbob and, pooled, on rotated f2-f4, but rotating the separable f3 and f4 removes most of the lead there. On held-out MA-BBOB its final share at 2000*D is level with IPOP-CMA-ES’s and its anytime area lower. We published AMM as an open-source Python package (pip install amm-optimizer; this https URL).
[LG-2] Navigating Disagreement Among Criteria for Evaluating Explanation Faithfulness
链接: https://arxiv.org/abs/2611.06846
作者: Grant Shanklin,Jiří Němeček,Asia Biega
类目: Machine Learning (cs.LG); Computers and Society (cs.CY)
*备注:
Abstract:Faithfulness requires that algorithmic explanations accurately reflect the model’s behavior. Yet existing operationalizations of faithfulness can disagree about whether the same explanation is faithful, leaving developers and auditors to choose among conflicting evaluations. In this paper, we investigate disagreement among evaluation criteria and offer a practical path forward for evaluating explanation faithfulness. We first identify how operationalizations of faithfulness differ. We introduce a faithfulness functional that locates their disagreements in what an explanation claims about the model, which comparisons bear on that claim, and how agreement across those comparisons should be aggregated, and we reconstruct existing evaluation criteria within it. We argue these reconstructions support understanding faithfulness as a common though incompletely specified relation between the model and explanation, so the differences reflect choices left open by the general requirement. Justifying an operationalization then comes down to justifying these choices, which, on our view, draws on explanatory, institutional, and normative considerations. We support this argument with a case study of adverse action notices, reconstructing the evaluation implicit in an industry-proposed procedure and identifying which of its choices the setting justifies and which remain open. Our work gives auditors and practitioners a way to navigate disagreeing evaluations of faithfulness. It also explains why the literature should expect multiple operationalizations, while showing how concrete settings can narrow the choice among them. More broadly, our analysis shows how a shared requirement of faithfulness can support contextually justified evaluations despite disagreement among existing metrics.
[LG-3] Systolic Linear Attention: Co-Designing Tile-Grouped Gated Delta Attention for TPUs
链接: https://arxiv.org/abs/2611.06828
作者: Gagik Amirkhanyan,Nuojin Cheng
类目: Machine Learning (cs.LG)
*备注:
Abstract:Gated-delta linear attention layers such as Gated DeltaNet (GDN), Kimi Delta Attention (KDA), and GDN-2 are tuned for GPUs, where small tensor-core tiles and fast vector registers make per-channel decay rescaling and sequential in-block solves cheap. On Google TPUs, which concentrate compute in wide 256 \times 256 systolic matrix units backed by a slower vector unit, porting these layers directly leaves half of each matrix tile empty and bottlenecks on vector operations. We present \textbfSystolic Linear Attention (SLA), which reshapes gated delta attention around the matrix unit by widening keys to 256 with grouped value heads, grouping decay gates into four 64 -channel tiles, and computing the chunk solve entirely with block matrix products in a single fused kernel. On TPU v7x, a full \textttSLA-R4 layer runs 1.6\times – 2.5\times faster than the open-source Tokamax KDA kernel ( 1.4\times – 2.4\times with exact chunk inversion), driven primarily by head sharing and kernel fusion. Across a 2,048 -chip TPU v7x pre-training study of sparse Mixture-of-Experts (MoE) models from 6.5\textB to 63.1\textB parameters, SLA trains 11% – 24% faster than KDA at similar perplexity, and reaches slightly lower perplexity than KDA in smaller matched-seed runs.
[LG-4] ConDR: Conservative Reachability through Neural Hamilton-Jacobi Subsolutions
链接: https://arxiv.org/abs/2611.06802
作者: Kendra Givens,Taylor T. Johnson,Bryan C. Ward,Thomas Beckers
类目: Robotics (cs.RO); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 8 pages, 4 figures
Abstract:Hamilton-Jacobi (HJ) Reachability identifies the unsafe states of a dynamical system from which it can be forced into failure under worst-case disturbance. However, classical grid-based solvers scale exponentially with state dimension. DeepReach improves scalability by replacing the grid with a neural value function trained on the HJ-Isaacs (HJI) variational inequality (VI), but penalizes both sides of the residual equally. This allows the learned value to lie above the true one and unsafe states to be misclassified as safe. We propose Conservative DeepReach (ConDR) which aims to learn a subsolution of the HJI VI such that the learned unsafe set contains the true one. ConDR enforces the obstacle branch and terminal condition by construction, pushes the HJI residual to a safety margin \epsilon below zero with a one-sided loss, and prevents the degenerate solution of classifying everything as unsafe. We prove that the exact solution of the margin-shifted VI that ConDR trains toward is a subsolution of the original VI so its unsafe set contains the true one and its value lies within \epsilon\tau of the true value function. We compare against DeepReach and MADR in the established systems of Air3D, Air6D, a 9D multi-vehicle system, and a 13D quadrotor. ConDR reports the lowest rate of unsafe state misclassified and the lowest closed-loop collision rates with at most a small increase in intervention. On Air3D and Air6D, ConDR tolerates model mismatch that scales with \epsilon . At matched false-negative rates, ConDR’s unsafe sets are smaller than the baselines showing the safety is not obtained by trivial over-approximation.
[LG-5] Continual Classification Beyond the Kernel Regime: Harnessing the Sparsity Bias of L2 Regularization in Diagonal Linear Networks
链接: https://arxiv.org/abs/2611.06759
作者: Gilad Karpel,Itay Evron,Ron Meir,Daniel Soudry
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study logistic regression for continual learning with linear predictors in a high-dimensional statistical setting. Existing theoretical work has largely focused on continual linear regression with variants of L2 regularization, motivated by their analytical simplicity and connection to the kernel regime of neural networks. We first derive the optimal L2 regularization strength for continual linear classification and characterize its generalization performance, recovering behavior analogous to prior results for continual linear regression. Next, we examine the effect of L2 regularization in deeper models using a minimal two-layer diagonal linear network. In the single-task setting, L2 regularization on the weights of such models is known to promote sparsity in the resulting predictor, yielding a rich, sparse feature-learning regime beyond the kernel regime (a similar effect arises in two-layer ReLU networks). However, in continual learning, where L2 regularization is typically centered on previous-task weights, this sparsity bias is no longer preserved. We therefore propose a weighted variant of L2 regularization that extends this beneficial sparsity bias to continual learning. We show theoretically and empirically that, for a sufficiently sparse teacher, our variant converges substantially faster than directly learning a linear predictor with conventional continual L2 or L1 regularization, trading some long-horizon bias for markedly improved performance over moderate task horizons.
[LG-6] Symmetry Breaking at the Edge of Stability via Central Flow Analysis
链接: https://arxiv.org/abs/2611.06718
作者: Naima Elosegui Borras,Georgios Arvanitidis
类目: Machine Learning (cs.LG)
*备注:
Abstract:Continuous symmetries in neural networks lead to conserved quantities that are invariant under gradient flow, such as layer balancedness in linear and ReLU networks. Finite-step gradient descent need not preserve these laws, particularly at the edge of stability (EOS), where the iterates oscillate around a smoother trajectory, which can be described by the central flow dynamics (Cohen et al., 2025). We study how gradient flow conservation laws evolve under the central flow description of this regime. We show that the curvature correction induces a finite-dimensional drift determined by the oscillation covariance along critical parameter directions. We thus take the first steps to analyse which conservation laws remain and are broken at the EOS according to the central flow dynamics by analysing a scalar two-layer model, two-layer matrix factorisation, and a two-layer ReLU network, where the known conservation laws are shown to be complete (Marcotte et al., 2023).
[LG-7] Energy-Guided Transport for Projection-Free Physics-Informed Flow Matching NEURIPS2026
链接: https://arxiv.org/abs/2611.06711
作者: Giuliano Galadini,Aurelio Uncini,Danilo Comminiello
类目: Machine Learning (cs.LG); Computational Physics (physics.comp-ph)
*备注: Accepted to NeurIPS 2026
Abstract:Deep generative models, particularly Flow Matching frameworks, have emerged as powerful tools for synthesizing complex physical systems. However, standard architectures remain purely statistically driven and agnostic to the underlying governing equations, often violating fundamental physical constraints during inference. Existing zero-shot approaches designed to enforce these constraints typically rely on rigid algebraic projections. While effective for sparse boundary conditions, applying such discrete geometric corrections to dense, full-domain PDEs may disrupt the probability path, \revpotentially deviating from the learned manifold and introducing substantial computational overhead. To address these challenges, we introduce Energy-Guided Transport, a novel training-free framework that formulates non-linear physical constraints as a smooth potential energy landscape. Rather than interrupting the ODE integration, our method applies a continuous dissipative force directly to the intermediate generative states, steering them toward the physical manifold. By evaluating this force implicitly via reverse-mode automatic differentiation, our framework preserves the natural dynamics of the flow while maintaining an efficient \mathcalO(N) algorithmic complexity. We extensively evaluate our framework across various 1D and 2D physical systems, including the chaotic Navier-Stokes equations. Our method achieves stable generation without structural divergence, significantly improves the stability-accuracy trade-off compared to state-of-the-art methods, and substantially accelerates inference with respect to exact projection strategies. Code available at: this https URL.
[LG-8] Where It Breaks: Diagnosing Perception Memory and Decision Failures of a Vision-Language Agent in Pac-Man
链接: https://arxiv.org/abs/2611.06704
作者: Xinglu Wang,Edward Wang,Yong Zhang,Zhenan Fan
类目: Machine Learning (cs.LG)
*备注: 35 pages, 6 figures
Abstract:Training an agent to play Pac-Man with only up, down, left and right actions raises a basic question: when performance is poor, what should be improved? Difficulty can arise from reading the screen or from choosing actions over a long sequence. We examine these problems in a Qwen3.5-9B-based agent through separate perception checks and primitive-policy reinforcement-learning studies. In the policy study, which uses textual boards to remove image transcription from the decision interface, completion on eight held-out mazes increased from 0% before this optimization to 80.2% at a selected checkpoint, with three extra lives. The perception checks show that accurate static screenshot reading can coexist with errors during interaction. A separate policy comparison also shows that further optimization can accompany worse action choices. Together, these findings motivate concrete checks for post-training: evaluate perception in use, distinguish training success from full-game validation, and examine decisions and later performance rather than relying on one selected score. The study provides empirical lessons for improving future agent training and evaluation.
[LG-9] PINR-ESI: Physics-Informed Self-Supervised Implicit Neural Representation Framework for EEG Extended-Source Imaging
链接: https://arxiv.org/abs/2611.06700
作者: Zhiyong Zhou,Sheng Liao,Youshen Xiao,Xiaolin Li,Yakang Dai
类目: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
*备注: Preprint, 10 pages, 11 figures
Abstract:EEG extended-source imaging aims to localize active cortical regions and characterize their spatial extent, yet supervised inverse networks typically depend on synthetic EEG-source pairs and may be sensitive to synthetic-to-real mismatch. We propose a physics-informed self-supervised implicit neural representation (INR) framework for spatial source reconstruction without paired source labels or simulation-based supervised pretraining. The lead field matrix encodes the source-to-sensor volume-conduction model and constrains the INR through agreement between predicted and measured EEG. Cortical current density is modeled as a continuous coordinate-conditioned field. A classical inverse estimate provides warm initialization and spatial constraints on source activity, while patch-integrated projection maps the field to discrete source coefficients for forward modeling. Simulations with varying noise levels, source counts, and source extents showed the clearest advantage in identifying active cortical regions, particularly for multiple and larger sources. Quantitative and qualitative evaluations using real EEG recordings from epilepsy patients and intracranial stimulation experiments further supported the feasibility of the framework for source imaging. These findings suggest the potential of physics-informed self-supervised INR for extended-source imaging.
[LG-10] SecureGARD: Dynamic Ontology-Guided Governance for Secure Tool-Using LLM Agents
链接: https://arxiv.org/abs/2611.06682
作者: Guoqin Xiong,Shengtao Wen,Xiang Chen,Xinyi Zeng,Zhaoxin Fan,Wei Ye,Xiaofeng Cao,Sheng-Jun Huang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Tool-using large language model (LLM) agents translate model outputs into consequential actions, turning indirect prompt injection into an execution-security problem. Existing defenses constrain tool calls, trajectories, or information flows, but often leave action semantics, state dependencies, and evidence relations implicit, limiting generalization across heterogeneous tools and complicating safe updates. We present \ours, a dynamic ontology-guided governance framework. \ours grounds heterogeneous calls in a shared semantic representation and combines an executable ontology with episode-level task state to derive deterministic ready, waiting, and forbidden decisions. A waiting decision identifies an unmet but recoverable obligation, enabling bounded replanning without expanding the agent’s authorization boundary. \ours evolves its ontology offline from verified training executions. An offline LLM proposes instance-independent action–state axioms, which are accepted only if they address ontology errors exposed by training failures, preserve logical consistency and protected entailments, and pass held-out validation before the ontology is frozen. We evaluate \ours on AgentDojo, ASB, and AgentDyn under benign and indirect prompt-injection settings. \ours achieves a 0% attack success rate (ASR) across all three benchmarks while improving task utility under attack compared with existing baselines. These results show that explicit semantic governance can strengthen execution security without indiscriminate blocking.
[LG-11] Do unlearning algorithms learn the right features?
链接: https://arxiv.org/abs/2611.06681
作者: Polina Dolgova,Amartya Sanyal,Sebastian U. Stich
类目: Machine Learning (cs.LG)
*备注:
Abstract:Machine unlearning is commonly evaluated with standard post-deletion metrics that compare an unlearned model to one retrained on the remaining data. Such snapshot evaluations can conceal consequential differences in the features the two models encode. We prove that if an unlearned and a retrained model agree on their outputs but encode different features, a single gradient step on their prediction heads can break this agreement. We further show that deleting an arbitrarily small fraction of the data can cause retraining to select qualitatively different features, so matching retraining can require learning features that the full model never had. We then introduce a measure of feature recovery that separates missing from extraneous features. Across eight unlearning methods, models that perform well on standard metrics often fail to recover the retrained features. This mismatch predicts deviations from retraining under subsequent relearning and downstream adaptation, and directly correcting it systematically reduces these deviations.
[LG-12] Self-Directed Apple Tasting
链接: https://arxiv.org/abs/2611.06677
作者: Tanmay Devale,Reed Phillips,Arvind Ramaswami
类目: Machine Learning (cs.LG)
*备注:
Abstract:We study the classical apple-tasting setting, a partial feedback model in which the learner observes the true label only when it predicts 1 . A combinatorial characterization for the rates of this problem in the online adversarial setting has been studied by Raman et al. (2024). We study the self-directed setting, in which the learner can adaptively choose the order of the points it predicts on. In the realizable setting, we establish a tetrachotomy on the expected minimax number of mistakes made by any learner. Depending on the VC dimension, the \emptyset -centered star number, and the self-directed dimension, we show that the rates could possibly be \Theta(1) , O(\log n) \cap \omega(1) , \tilde \Theta(\sqrtn) , or \Theta(n) . In the agnostic setting, the minimax regret is \tilde \Theta(\sqrtn) or \Theta(n) , depending on the VC dimension.
[LG-13] Fingerprinting the Dispatchers behind Distributed Model Extraction
链接: https://arxiv.org/abs/2611.06661
作者: Snigdha Chandan Khilar
类目: Machine Learning (cs.LG)
*备注:
Abstract:API distillation campaigns try to hide model-stealing traffic by splitting requests across thousands of load-balanced accounts, which standard content classifiers often miss. However, request timestamps, specifically arrival time and stream completion time, reveal the underlying dispatching strategy, such as round-robin counters or key-shuffling pools. Using a 20 percent seed set, this timing analysis can forensically rank the remaining accounts of a campaign with 0.92 to 1.00 AUC within 20 minutes, even across thousands of real production accounts. While adaptive attackers struggle to hide rotating dispatchers using jitter, pinning keys successfully evades detection, though it can severely throttle throughput when quotas bind. Ultimately, timing exposes shared infrastructure patterns even when content filters fail.
[LG-14] Detecting Silent Action-Interface Mismatch with Calibrated Relational World Models
链接: https://arxiv.org/abs/2611.06653
作者: Rui Zhu,Xurui Li,Haixu Tang
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: Accepted at the CoRL 2026 Workshop on Modeling Uncertainty in Robotic World Models (UncertainWM). 20 pages, including 12 pages of supplementary material
Abstract:A robot action can satisfy shape and range constraints yet carry the wrong semantics: axes may be permuted, signs inverted, rotations rescaled, or gripper polarity reversed. We study learned state-action-effect relations as offline contracts for detecting these mismatches and ranking their likely fields. Clean held-out transitions calibrate alarms; a fixed inverse vocabulary supports closed-set localization. On ten LIBERO Spatial tasks comprising 432 source episodes, a capacity-controlled comparison at 2.5% empirical clean false-positive rate gives 46.1% recall for a ridge forward residual, 61.0% for a four-component joint Gaussian mixture, and 65.7% for the exact conditional of that same mixture. The joint score retains a 14.9 percentage-point advantage over ridge, while conditioning improves recall by a further 4.7 points. Joint scoring instead has higher localization: 95.8% versus 94.6%. These matched-rate comparisons are retrospective diagnostics, not deployment calibration guarantees. Earlier Object/Goal results broaden tasks within the LIBERO convention; DROID, all three HYDRA tasks, short temporal faults, small offsets, exact repair, and simulated mitigation fail their respective gates, while xArm contributes one task only. The evidence supports a bounded offline diagnostic and a detection-localization tradeoff, rather than joint-density superiority or cross-robot safety.
[LG-15] Dimension-Free Private Mean Estimation is Impossible Without the Covariance
链接: https://arxiv.org/abs/2611.06615
作者: Renaud Gaucher,Ephraim Linder,Adam Smith,Jonathan Ullman,Lydia Zakynthinou
类目: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS)
*备注:
Abstract:We study the limits of differentially private Gaussian mean estimation when the covariance matrix is unknown. Prior work shows that when the covariance \Sigma is known, the optimal sample complexity can depend only on its spectrum, rather than on the ambient dimension d . We establish new lower bounds for Gaussian mean estimation with Euclidean error which show that private estimators cannot achieve similar guarantees when the covariance is unknown. First, when the sample size n is smaller than the effective dimension \mathrmtrace(\Sigma)/|\Sigma|_\textop , the optimal error when the covariance is unknown can be larger than the optimal error when it is known by a factor of \sqrtd/n . Second, when the sample size is larger than the effective dimension, even for diagonal covariance matrices with effective dimension bounded by \log d , the optimal error when the covariance is unknown can be larger than the optimal error when it is known by a factor of \sqrtd/n^3/2 . Together, these results show that in the unknown covariance setting, exploiting anisotropic structure under differential privacy can require a dimension-dependent error or sample complexity.
[LG-16] Verifier-Guided Self-Consistency: Test-Time Reasoning with Imperfect Generators and Verifiers
链接: https://arxiv.org/abs/2611.06606
作者: Ekaterina Fadeeva,Xueyan Li,Mrinmaya Sachan,Maxim Panov
类目: Machine Learning (cs.LG)
*备注:
Abstract:Large language models can solve many multi-step reasoning problems, but their reasoning remains brittle: plausible intermediate steps often lead to incorrect answers. Improving reasoning at test time requires balancing two imperfect signals: the generator, which provides diversity, and the verifier, which guides correctness. Existing approaches largely favor one side of this tradeoff. Verifier-weighted voting relies on trajectories sampled from the generator and cannot steer generation toward correct ones. By contrast, beam-style verifier-guided search can steer generation, but may over-concentrate on narrow trajectory sets and amplify verifier errors. We propose Verifier-Guided Self-Consistency (VGSC), a stochastic decoding framework that combines step-level correctness and future-outcome estimates to guide reasoning during generation. VGSC repeatedly proposes candidate next steps from the base language model and uses both signals in an acceptance/resampling procedure, while maintaining multiple trajectories for final answer-level self-consistency. We provide a probabilistic interpretation of VGSC as approximate sampling from a verifier-tilted trajectory distribution. Across multiple reasoning benchmarks, generator models, and verifier families, VGSC achieves higher average accuracy than representative baselines. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2611.06606 [cs.LG] (or arXiv:2611.06606v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.06606 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-17] he Behavioral Rate of Fine-Tuning: Compression as a Measure of Useful Information
链接: https://arxiv.org/abs/2611.06598
作者: Joey David,Paul Caillon,Alexandre Allauzen
类目: Machine Learning (cs.LG)
*备注: 26 pages
Abstract:Low-Rank Adaptation (LoRA) adapts a frozen language model through a compact learned update, making it a natural setting for asking how much information fine-tuning needs to write. Yet adapter rank, parameter count, or numerical reconstruction error do not directly tell us how many bits are required to preserve the acquired behavior. We therefore define an operational behavioral rate R^\star as the smallest serialized size of a compressed adapter that, once reloaded into the frozen base model, retains a fixed fraction of its held-out fine-tuning gain. Across LoRA fine-tuning experiments, this rate increases with distinct training content at fixed optimization budgets, but is not determined by intuitive metrics like source description length, fine-tuning gain, or weight-space reconstruction error. Most importantly, the same datasets exhibit substantially different, and even reordered, rates across frozen base models, showing that fine-tuning compressibility is a property of the dataset–receiver pair rather than of the dataset alone. However, we show that correction spectra measured on the data are able to rank these rates within a receiver. Finally, we study a striking noise-reducing effect of compression: adapters trained on deliberately corrupted datasets fall far below the frozen model, yet after compression score above it (23.7% to 82.9% on GSM8K, against 77.6% for frozen Qwen2.5-7B), consistently across tested families and sizes. The damage is unevenly concentrated in the adapter’s leading singular directions, which compression appears to filter. Away from its training task, the compressed corrupted adapter also generalizes better than a clean fine-tune: on harder math it retains 97% of base accuracy, versus 44% for the clean adapter.
[LG-18] A trust layer for fast pedestal-stability surrogates: calibrated error bars out-of-distribution alarms and few-shot recalibration
链接: https://arxiv.org/abs/2611.06597
作者: Amin Vakhshouri
类目: Machine Learning (cs.LG); Plasma Physics (physics.plasm-ph)
*备注: 9 pages, 6 figures
Abstract:Machine-learning surrogates of linear peeling-ballooning stability now run in milliseconds, but they are mostly deployed as point estimators: they do not say how wrong they may be, or when a query lies outside what they were trained on. For a fusion company designing a new device, every equilibrium of interest is outside the training data of existing surrogates, so this is the case that matters. We describe a model-agnostic trust layer for stability surrogates, made of conformal error bars, out-of-distribution alarms that route doubtful cases to the physics code, and recalibration from a few simulations of a new design, and we test it with pre-registered hypotheses. On a large public dataset of JET-like equilibria with ideal-MHD growth rates, the error bars hold their nominal coverage in distribution and fall short under a shift in plasma shape; a few dozen simulations of the new shape, or routing the cases the alarms flag, restore it. On resistive BOUT++ surrogates, the bars stay calibrated on held-out runs, the alarm flags tokamaks absent from training, and a scenario scan is screened in seconds instead of hours. Across equilibria from the SPARC library and the DIII-D coil set, a surrogate trained on our own graded BOUT++ labels beats the baseline on unseen equilibria, while a surrogate trained elsewhere ranks the same cases backwards and its alarm declines them. Building our own neural-operator surrogate, the harness exposed two input defects whose fix closed its gap to a simple network. For every surrogate tested, the trust layer keeps coverage near target, with a bar width set by that surrogate’s own error. Not every pre-registered hypothesis held, and we report those results and the limits of all of them.
[LG-19] hermodynamics-Informed Learning Captures Hierarchical Hydrogen-Bond Networks in Deep Eutectic Solvents
链接: https://arxiv.org/abs/2611.06590
作者: Yanpeng Ye,Chaoran Duan,Mian Xu,S. Mani Sarathy
类目: Machine Learning (cs.LG); Chemical Physics (physics.chem-ph); Quantum Physics (quant-ph)
*备注:
Abstract:Deep eutectic solvents (DESs) offer a tunable platform for lignocellulose refining, but their solvation behavior depends on multicomponent hydrogen-bond networks rather than individual constituent properties. We develop TiNN, a thermodynamics-informed neural network that hierarchically represents HBA-HBD association and solute accommodation through a learned free-energy pathway. Trained on 938 records, TiNN achieves a test-set R^2 of 0.917 and generalizes to new HBA-HBD combinations and sparse conditions. Ablation identifies solute-specific stabilization as the largest contributor to prediction accuracy, while the importance of hydrogen-bond compatibility depends on the dissolution mechanism. Experimental measurements of glucose, xylan, and Kraft lignin in acidic, basic, and neutral DESs show that TiNN reduces the geometric mean solubility error from more than 23-fold with COSMO-RS to approximately 1.5-fold. Without training on fractionation data, TiNN also distinguishes DES formulations with high and low lignin and hemicellulose removal from corn stover. These results support chemically grounded solvent screening for lignocellulose fractionation.
[LG-20] From Fixed Defaults to Calibrated Activation Policies for Long-Term Time-Series Forecasting
链接: https://arxiv.org/abs/2611.06559
作者: Shijie Xu,Boyan Tang,Jianghua Wu
类目: Machine Learning (cs.LG)
*备注: 36 pages, 2 figures. Shijie Xu and Boyan Tang contributed equally. Code: this https URL
Abstract:Long-term time-series forecasting supports decisions in energy, transportation, and other applications where reliable predictions are needed beyond the observed data. Although forecasting research has advanced model architectures, fixed activation configurations provide limited guidance on how activation choices should vary across datasets and prediction horizons. This study aims to establish when task-specific activation choices improve forecasting accuracy and how previous experiments can guide the selection of suitable configurations for new tasks. We develop a calibrated selection policy that learns a reliable default from prior experiments, adapts the activation function and its location to each task, and retains the default when evidence for a change is insufficient. In an evaluation supported by 1,225 Informer training runs across 17 dataset variants and 51 forecasting tasks, the policy reduces mean squared error by 7.86% on average relative to the conventional fixed-activation baseline and improves 72.55% of tasks. These findings establish activation choice and placement as practical opportunities for improving existing forecasting models and show how accumulated experimental evidence can guide configuration decisions without repeating an exhaustive search for every task.
[LG-21] EvoOPD: Evolving Privileged Information Strategies for Black-Box On-Policy Distillation
链接: https://arxiv.org/abs/2611.06542
作者: Ximo Zhu,Ruiqi Liu,Yi Han,Zhengan Wu,Gengsheng Li,ZiAn Wang,Zhu Yuantian,Jin Jiang,Wenhao Wang
类目: Machine Learning (cs.LG)
*备注: 26 pages, 4 figures
Abstract:Proprietary language models exhibit strong reasoning and coding capabilities, but their output distributions are inaccessible, making fine-grained distillation methods that rely on teacher logits difficult to apply directly to black-box models. We observe that rich textual feedback from a black-box teacher can be treated as privileged information (PI) and converted, through a PI-conditioned privileged teacher, into dense supervision signals on the student’s own trajectories. However, what PI should be acquired from the teacher to most effectively support student learning remains unclear. PI strategies that differ in where they focus, what information they request, and how much detail they provide can have substantially different effects on the current student, and their relative effectiveness can further change as training progresses. Based on this observation, we propose Evolving Privileged Information Strategies for Black-Box On-Policy Distillation (EvoOPD), which treats PI acquisition itself as an optimization problem and maintains a PI Strategy Bank that evolves with the student state. EvoOPD uses the measured repair utility of PI on the current student to adapt the focus, information type, and granularity of teacher feedback, allowing the acquired information to continually match the student’s learning needs. We systematically evaluate EvoOPD on mathematical reasoning and code generation with Qwen3-1.7B, 4B, and 8B students. EvoOPD achieves the best overall performance in all six model-task settings, improving over the strongest baseline in each setting by an average of 2.14 percentage points in mathematics and 1.30 points in code. Further analysis shows that the relative effectiveness of PI acquisition strategies changes with student state, making a fixed strategy difficult to sustain as the best choice throughout distillation.
[LG-22] An Interaction-Aware Decision Framework for FDIC-Assisted Distressed Bank Acquisitions: Empirical Evidence from the FDIC Failed-Bank Panel and the SVB-First Citizens Case
链接: https://arxiv.org/abs/2611.06514
作者: Yousef Nejatbakhsh,Malihe Aliasgari,Parsa Safaee,Sina Shokoohyar
类目: Machine Learning (cs.LG); Applications (stat.AP)
*备注: 18 pages, 2 figures, 15 tables. Research code, datasets, and reproducibility materials are available at this https URL
Abstract:Acquirers of failed banks weigh acquisition value, regulatory protection, risk, buyer strength, and execution difficulty jointly, because each signal’s meaning depends on the others: high non-performing loans matter less when government loss sharing absorbs the losses. The proposed Spider Network Filtering Framework encodes such dependencies as four interpretable interaction rules among five dimensions, yielding graph-inspired features for a logistic model and a Pursue/Caution/Avoid decision rule. We evaluate it on 55 acquisitions assisted by the Federal Deposit Insurance Corporation in 2008-2024 by acquirers with a parent listed in the United States. Six financial inputs come from regulatory reports published before each failure; two deal terms, which the winning bidder knows when it bids but investors learn at the announcement, come from the regulator’s press releases. The outcome, a positive three-day abnormal return around the announcement, records the market’s initial reception, not realized success. Under grouped, nested cross-validation, the Spider-enhanced model reaches a mean area under the receiver operating characteristic curve of 0.594, against 0.615 for logistic regression on the same inputs and 0.632 for ridge regression; permutation tests that repeat the full model selection give at most weak evidence that any of them ranks better than chance. The Spider score used directly does not rank better than chance, and an apparent link between target distress and higher returns disappears against a bank-sector benchmark. Thresholds tuned for the F1 score approve most acquisitions with negative returns and realize negative utility when each costs three times what a positively received acquisition gains.
[LG-23] Nowcasting Future Performance for Population Pruning in Survival-of-the-Fittest Training NEURIPS2026
链接: https://arxiv.org/abs/2611.06512
作者: Rico Städeli,Léo Meynent,Boris Knyazev,Damian Borth
类目: Machine Learning (cs.LG)
*备注: Accepted at: 40th Conference on Neural Information Processing Systems (NeurIPS 2026). Workshop: Neural Network Artifacts as a New Data Modality
Abstract:Training high-performing neural networks often involves evaluating populations of candidate models with varying hyperparameters, architectures, or random seeds. Training every candidate to completion is computationally expensive, yet stopping seemingly weaker runs early risks discarding models that would ultimately perform better, as learning curves can cross. To address this challenge, we investigate whether intermediate training checkpoints can predict future model accuracy and enable more informed pruning decisions. Our nowcasters predict test accuracy five epochs ahead or at the end of training. They are trained with accuracy and loss histories, weight statistics, and SANE representations as inputs. SANE is a transformer autoencoder that encodes a checkpoint as a sequence of latent tokens. We then use the nowcasters in Survival-of-the-Fittest (SOTF) training. SOTF repeatedly ranks runs during training by predicted accuracy and stops the lowest-ranked ones. Our results show that weight statistics work best for five-epoch forecasts, and SANE representations work best for final accuracy prediction. For evaluation, we compare SOTF with pruning the population by current accuracy. In homogeneous model zoos such as the CNN zoos, pruning by current accuracy already retains a final model only slightly worse than the best. Model zoos with a larger spread in performance, such as the ResNet-18 and ViT zoos, are more difficult to prune by current accuracy alone. In these zoos, SOTF with nowcasted final accuracy almost closes the gap to the best model.
[LG-24] he Alignment Tax Is Graded: Task-Selective Capability Loss Under Length-Normalized DPO with SFT Retention
链接: https://arxiv.org/abs/2611.06476
作者: Ye Yuan
类目: Machine Learning (cs.LG)
*备注: 25 pages, 3 figures, 23 tables. Code: this https URL
Abstract:Setting. This is an empirical characterization study; we propose no new method and run no comparisons against alternative objectives. We fine-tune Qwen3.5-4B with sparse LoRA and length-normalized DPO plus an SFT retention term (the RPO objective) on structured financial-filing tasks, and track downstream capability as the implicit reward margin grows to \sim 7 \beta -scaled units (nf4 training; \sim 3 under fp16 measurement on the unquantized base). Question. Does margin growth degrade capability uniformly across tasks, or selectively? Findings. In a pre-registered study (8 seeds, 6 unseen at freeze) with artifact-aware scoring and TOST equivalence testing, the alignment tax is graded and task-selective: extraction is preserved ( \Delta = -0.55 pts, equivalent at \pm5 pts), json loses little ( \Delta = -3.52 pts, equivalence fails on a clustered item-level interval), and router degrades severely ( \Delta = -30 pts). Router’s loss concentrates in a margin 1-3 damage window containing a transient format-break spike. The margin is accounted for almost entirely ( \geq 96% ) by suppression of rejected completions, with chosen reward approximately flat, consistent with the retention term anchoring it; we do not attribute the behavioral damage to any single gradient channel. Gold-logp deflation is task-selective (extraction -0.15 , json -1.04 , router -1.47 nats/token under fp16), with gold selectively protected relative to incorrect controls on extraction (8/8 seeds). QLoRA reward margins differ up to 3.4\times between nf4 training and unquantized fp16 measurement, with an axis-dependent gap. Scope. Length-normalized rewards with SFT retention, sparse LoRA on 8/32 layers, one 4B model, easily-separable synthetic preferences, three structured tasks. A no-retention control and a measured format-alignment metric are left to future work.
[LG-25] Graph Foundation Models Can Become Effective Multivariate Time Series Anomaly Detectors
链接: https://arxiv.org/abs/2611.06441
作者: Zhuangyan Hu,Zhuoxuan Liang,Wei Li,Xiaohua Jia,Moustafa Youssef,Sumi Helal
类目: Machine Learning (cs.LG)
*备注:
Abstract:Multivariate time series anomaly detection (MTSAD) relies on modeling both temporal dynamics and inter-sensor dependencies to identify complex system faults. Although graph-based detectors explicitly model sensor topologies, they train task-specific networks from scratch on limited target-domain data, remaining vulnerable to spurious correlations and unable to reuse transferable priors. While pretrained Graph Foundation Models (GFMs) offer transferable topological priors and anomaly discriminability, directly adapting them to MTSAD is hindered by cross-modal feature-topology mismatch, unreliable pseudo-anomaly supervision, and decision distortion during node-to-window aggregation. To fill these gaps, we propose a novel adaptation framework, Gfm4Mtsad, which unleashes pretrained GFMs for MTSAD. We introduce the Structure-and-Information Contract Alignment (SICA) module to bridge continuous time series and the pretrained GFM space. This module harmonizes principal subspace token projection, normal-prior-guided graph construction, and residual recovery to suppress spurious edges while retaining fine-grained temporal details. We propose a Mechanism-Aligned Pseudo-Anomaly Screening strategy (MAPS) to steer unsupervised learning. It applies localized dependency breaks while preserving empirical residuals, selecting hard negatives near the normal boundary for targeted supervision. We develop a Tail-Gated Evidence Calibration mechanism (TGEC) to eliminate node-to-window aggregation distortion. It calibrates multi-granular graph responses against normal operational references, strictly preserving benign score rankings while selectively refining extreme anomaly candidates in the tail distribution for reliable window detection and interpretable sensor evidence. Extensive evaluations with two GFM backbones demonstrate that Gfm4Mtsad achieves superior detection performance over existing baselines.
[LG-26] Understanding the Out-of-Distribution Generalization of Chain-of-Thought Reasoning in LLM s
链接: https://arxiv.org/abs/2611.06438
作者: Nuo-Jing Liang,Xiao-Tong Yuan,Pan Zhou
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:Chain-of-Thought (CoT) reasoning substantially improves Large Language Models (LLMs), yet its out-of-distribution (OOD) generalization mechanism remains theoretically underexplored. We study CoT as compositional OOD generalization: the target task lacks complete reasoning trajectories or direct input-output pairs, while training provides only in-distribution (ID) atomic step data. We show that CoT generalizes by repeatedly applying a shared single-step predictor to compose reusable ID transitions. Theoretically, let T denote the number of reasoning steps in the target task, n the number of training samples per sub-task, and k the number of output tokens per step; then the OOD task risk is controlled by the sum of the ID sub-task risks, scaled by a coefficient of order \mathcalO(k) that is independent of T , although the summed risk itself may still grow with T . A Rademacher-complexity analysis further shows that the estimation term is of order \mathcalO(\sqrtT/n) , rather than linear in T , due to predictor sharing under a uniform capacity control on the shared class. For autoregressive Next-Token Prediction, the final-task risk is bounded by the best-in-class step-level risk scaled by the same coefficient of order \mathcalO(k) , plus an estimation term of order \widetilde\mathcalO\big(k\sqrtT/n\big) . A parity-coupled construction further shows that, for the sub-task empirical risk minimizer, the expected final-task risk is at least the best in-class risk plus \Omega(\sqrtT/n) when T\le n . Experiments on Atom-Task and GSM8K support the theory: CoT enables short-to-long and compositional generalization, while explicit state tracking is crucial. Our code is available at this https URL. Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML) Cite as: arXiv:2611.06438 [cs.LG] (or arXiv:2611.06438v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.06438 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-27] Hardwired Distributed Processing Units for Inference: Neural Decisions at the Sensor that Free the Host
链接: https://arxiv.org/abs/2611.06436
作者: Raj Mali
类目: Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: 8 pages, 9 figures, 11 tables. Companion to arXiv:2610.33107 and arXiv:2610.38752
Abstract:We show that a hardwired distributed processing unit (HDPU) can take a neural decision at the sensor and so remove the inference from the host. Built for zebra-crossing detection on a low-cost Artix-7 FPGA, the unit decides on every camera frame in a fixed 2,269,477 clock cycles (34.04 ms depends upon Fpga clk ), inside the 37.49-ms frame period, and on the board it returns the same scores and decisions as its bit-exact software model on all 639 test images. The whole detection path, from camera pixels to decision, takes 1,814 LUTs, with no processor, no operating system and no full-resolution frame store. When the unit answers instead of reporting “not sure”, 95.7% of its answers on street photographs are correct, and a hardware contrast stage lifts accuracy on camera-like dashcam frames from 47.7% to 78.2%. The host is spared all of the inference: 8.13 ms of a core in every frame period, where a full-precision ResNet-18 would need more than a frame period even on 16 threads. It also receives 16 bytes per frame instead of 614,400, 38,400 times less. The unit is an 80x60-pixel pipeline feeding a four-layer binarized CNN with 93,408 one-bit weights, verified bit-exactly from a Python model through RTL simulation to the board. Earlier HDPUs did pre-processing (edges, corners, steering); this one shows that the decision itself can move to the sensor.
[LG-28] Partially Interpretable Decision Trees
链接: https://arxiv.org/abs/2611.06428
作者: Ludovic Schwartz,Federico Berton,Tim van Erven
类目: Machine Learning (cs.LG)
*备注:
Abstract:Partially interpretable models extend simple interpretable models with the option to defer complicated inputs to an accurate but uninterpretable black-box model. By deferring more inputs, the model gets more accurate; by deferring fewer, it gets more interpretable. Existing procedures to train such partially interpretable models build on linear models or decision rules. We propose three algorithms for decision trees: a partially interpretable extension of the existing Branches algorithm, which performs exact search to optimize the tree; a conversion algorithm that we run on top of standard CART, which is computationally more efficient; and a generalized CART whose splitting criterion takes into account the option to defer while growing the tree. Perhaps surprisingly, the performance of the three algorithms is similar across many realworld datasets. We can explain this through a formal bound on the potential gain of smart tree construction, provided the black box accuracy is sufficiently homogeneous across the domain. We also construct a synthetic dataset where the accuracy of the black box is deliberately inhomogeneous. There, our deferral-aware generalization of CART does outperform regular CART and matches the performance of exact search. Finally, we evaluate an application to concept bottleneck models. Partially interpretable decision trees can replace uninterpretable neural networks after the bottleneck. On undeferred inputs, the model is fully interpretable, and deferring preserves accuracy.
[LG-29] KolTa: A Physics-Informed Active Learning Framework for Formulation Discovery with Limited Data
链接: https://arxiv.org/abs/2611.06394
作者: Boyong Tang,Tao Wang,Lei Yang
类目: Machine Learning (cs.LG)
*备注: 21 pages, including 11 pages of supplementary material; 4 figures
Abstract:Finding high-performing formulations is difficult when few have been measured and experimental budgets are limited. We propose KolTa, an active learning framework that selects batches to discover the top 5% of formulations at a target temperature under a fixed query budget. KolTa combines a three-stage predictor of component contributions, role-conditioned interactions, and anchored temperature response with task-specific soft physics constraints. Its Physics-Residual-Calibrated Upper Confidence Bound (PRC-UCB) acquisition rule combines predicted utility, ensemble disagreement, and a target-temperature physics residual, with nonnegative exploration weights fitted to out-of-fold signed errors. We evaluate the framework on three measured tasks and one simulated task covering binary-mixture viscosity, drug solubility, ionic-liquid viscosity, and electrolyte conductivity. KolTa ranks first or second in mean area under the learning curve (AULC) on every task and outperforms seven of the eight comparator strategies in paired tests, winning 75 of 78 paired comparisons across tasks and seeds against those seven strategies. Across the three measured tasks, its largest gap to the best mean AULC is 0.032, compared with at least 0.056 for every comparator. On Drug, reaching 80% average recovery of the remaining targets across drugs requires an estimated median of 348 additional formulation queries, about 18% fewer than the fastest comparator. Researchers can adapt the framework to diverse formulation problems through component roles and physical constraints.
[LG-30] Give Up to Get More: Priors Not Policies Govern Test-Time Compute Allocation
链接: https://arxiv.org/abs/2611.06372
作者: Jinhyung Bae
类目: Machine Learning (cs.LG)
*备注: 36 pages, 6 figures. Code, results and survey coding: this https URL
Abstract:When many inputs share a budget of repeated attempts by a large language model (LLM), a common rule gives harder inputs more. On public math, code and formal-proof samples, the best published difficulty-proportional rule captures half or more of the oracle gain on benchmarks the model mostly solves, but at most 13% on hard workloads, even given true pass rates. Because the value of one more attempt peaks at intermediate difficulty, the optimal static allocation is inverted-U and gives little to inputs that are hopeless at the current budget. In a strict survey of 221 LLM allocation papers, at least 13 give the hardest inputs the most attempts; their easier settings sit at budgets where this is near-optimal, so the cost appears on harder workloads. Acting on it requires knowing which inputs are solvable, so information, not the policy, drives the gain. With a verifier and a prior from a related model’s past runs on the same inputs, Thompson Sampling raises the solved share on the hardest workload from 14% to 27%; without such a prior, no policy we tested beats round-robin. For the policies we tested, the gain vanishes between prior rank correlations of about 0.3 and 0.55, which cheap text-based priors barely reach on the weakest models, and it requires a verifier whose false acceptances are rare relative to the pass rate. On new Qwen2.5 samples the pattern holds, and tempering the prior avoids losses on easy workloads. We turn these results into a decision guide.
[LG-31] EEGDash Corpus: A Training-Ready Public Corpus for EEG Foundation Models
链接: https://arxiv.org/abs/2611.06356
作者: Kuber Shahi,Bruno Aristimunha,Arnaud Delorme
类目: Machine Learning (cs.LG)
*备注: 8 pages, 6 figures, 7 tables. Accepted at the Workshop on Machine Learning for EEG Signal Processing (MLESP 2026), IEEE BIBM 2026
Abstract:Foundation models for electroencephalography (EEG) are pretrained on tens of thousands of hours of recordings, and their accuracy grows with that amount. Such data is hard to assemble. The largest sources are clinical archives that cannot be redistributed, and the public sources are spread over several archives, file formats, and licences. We release eegdash-corpus-v0, a training-ready corpus of 77 public EEG datasets, as the gated Hugging Face dataset EEGDash/eegdash-corpus-v0 with the SHA-256 of every file. REVE, the largest published EEG foundation model, names 89 datasets as its pretraining corpus. Twelve cannot be redistributed: four clinical archives that hold 82% of its hours, and eight under no-derivatives licences. The corpus holds the 76 others that REVE’s sampling rate allows and one more from REVE’s release card: 77 datasets, 14,075 hours from 7,272 subjects, 5.0 million 10-second samples and 6.21 billion tokens per epoch. Every dataset is fetched through EEGDash, converted once with REVE’s recipe, and stored in its own buckets of a chunked Zarr store with its licence recorded beside it, so a model can train on the whole store or on any subset of datasets. To test that the store pretrains a model, we retrain REVE-Base from scratch twice with the authors’ code, changed only in how it reads data, and score it on seven public OpenEEGBench tasks. It reaches 0.331 under linear probing and 0.273 after fine-tuning, against 0.333 and 0.325 for the released model and 0.200 and 0.205 for an untrained encoder. The converter, the conversion reports, and the loader changes are released with the corpus.
[LG-32] Flexible Intensities Matter: A Comprehensive Re-evaluation of Hawkes Processes and Neural Temporal Point Processes NEURIPS2026
链接: https://arxiv.org/abs/2611.06348
作者: Hendrik Mehrtens,Florian Buettner,Oliver Stegle
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Accepted at NeurIPS 2026. 36 pages, 3 figures, 25 tables. Code: this https URL (paper reproduction on branch neurips2026)
Abstract:Temporal point processes (TPPs) are the canonical framework for event sequence data, with applications spanning seismology, social media, and electronic health records. Recently, neural TPPs have been reported to substantially outperform classical TPPs, such as Hawkes processes, but these comparisons pitted flexible neural models against Hawkes processes in their most restricted form, and did not consider large-scale medical data. Here, we revisit this comparison through a systematic evaluation of Poisson and Hawkes processes of increasing flexibility against neural TPPs, on five established benchmark datasets and on two large-scale medical evaluation tasks derived from UK Biobank disease histories and MIMIC-IV ICU records, under autoregressive next-event prediction. First, the previously reported gap between Hawkes processes and neural TPPs largely disappears once Hawkes processes are equipped with flexible, learnable baselines and kernels. Second, the flexibility of the intensity parametrization also limits prior neural TPPs: a Transformer with a spline-based intensity head outperforms previous intensity-based neural TPPs and is on par with the intensity-free model. Decomposing the log-likelihood into a time term \log p(t) and a type term \log p(d \mid t) , we find that this flexibility advantage lies almost entirely in the time density: more flexibility does not improve the type term. On these benchmarks, the gains of neural TPPs thus lie mostly in modeling when events occur, not which. Third, our results underscore the value of complex datasets for evaluating TPPs: only on the MIMIC-IV ICU records do neural TPPs retain an advantage in the type term. Our results inform both future method development and model choice. We release our framework, TorchPP, as an open-source package with consistent implementations of classical and neural TPPs.
[LG-33] CoRTeC: A Differentially Private Release for Synthetic Tabular Data Decodable by Any Generator
链接: https://arxiv.org/abs/2611.06342
作者: Calister Nnona
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: 29 pages, 4 figures. Code at this https URL and the releases, synthetic draws and score files at this https URL
Abstract:The records that could train the most useful models, such as a hospital’s patient files, are often the ones an institution cannot share. We propose CoRTeC, a method that releases synthetic versions of such records under differential privacy. Differential privacy places a proven limit on what the synthetic data can reveal about any person, and in particular on membership inference. CoRTeC spends the privacy budget once, on a small file of noisy counts designed for training a classifier. For each group of records, the file gives the group’s size, its balance of outcomes, a histogram of every column for each outcome and the outcome rate in each of several non-overlapping cells. A reviewer can check every number before the file leaves the institution. Anyone can then turn the counts into synthetic rows, as often as they like and at no further privacy cost. A maximum-entropy decoder can do this without any language model, and so can each of the five frozen language models we tested. None of them ever sees a real record. A selection step then matches the rows to the released counts. Decoded without a language model, the automatically configured release at \varepsilon = 2 trains classifiers to within 0.02 AUC of a real sample of the same size on the Adult dataset, across three data splits. On NHANES and bank marketing, without the selection step, it stays within 0.05 AUC. On Adult, with a fixed cohort rule, the release carries the outcome rates a deployment names more closely than a real sample does, at some cost in accuracy on pairs of columns. Four controls show that most of the utility comes from the release and the selection step, and that a language model adds what it knows from the column names. No private row ever reaches a model, and no row CoRTeC generates copies a real one. The strongest membership attack stays far below the bound the budget allows.
[LG-34] Consistency Learning and Distillation for One-Step MIMO-OFDM Channel Estimation
链接: https://arxiv.org/abs/2611.06271
作者: Fredrik Nguyen,Qiwen Wang
类目: Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注: 6 pages, 3 figures, 4 tables
Abstract:Iterative neural channel estimators can improve recovery under pilot contamination, but repeated network evaluations increase inference cost. This paper develops observation-anchored consistency and Consistency Mid-Training (CMT) estimators for one-step channel estimation in multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) systems. Both estimators refine pilot-derived least-squares channel estimates in the delay-beam domain. Consistency learning couples endpoint predictions along noisy-to-clean paths. CMT then refines a compact consistency-trained student to predict a flow-matching teacher’s endpoint from intermediate trajectory states. Across the evaluated SNR range, the compact CMT student reduces uncoded symbol error rate (SER) by 9.6-20.2% relative to the one-step flow-matching baseline (0.44-0.98 dB in logarithmic SER reduction), while using 74.5% fewer parameters and 73.0% fewer network multiply-accumulates.
[LG-35] Analysis of Local Open-Weight LLM Agents on Vulnerability-Reproduction Tasks
链接: https://arxiv.org/abs/2611.06257
作者: Jorge García-Carrasco,Javier Sanchis,Álvaro Navarro,Alejandro Maté,Juan Trujillo
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG); Software Engineering (cs.SE)
*备注: 35 pages, 6 figures, 5 tables. Preprint of a manuscript under review at the Journal of Systems and Software (Elsevier). Artefact: this https URL
Abstract:Open-weight language models that fit on a single consumer GPU are becoming capable on a growing range of tasks, and they matter because they can run locally without sending the code to a hosted service. Vulnerability reproduction, turning a vulnerability description into an input that crashes the vulnerable build, is a natural target for them, since it is an automated analysis task closer to test generation than to code comprehension and it runs against a build that must be executed rather than read. In order to assess whether current models can perform it, we evaluate twelve configurations of local models, spanning three generations, two architectures and three quantization levels, on ten CyberGym vulnerability-reproduction tasks, three runs each under one harness, and release the full artefact. We show that every configuration, down to the 4B model, solves most of the six easier tasks, and configurations separate only on the remaining four. For the harder tasks, the 27B model is required. Quantization did not change what the models solved in any matched pair, but it changed how fast they got there, and the mixture of experts reached the same scores in less wall clock. We also examine the agent trajectories. Almost every run we examined, including those that fail, names the vulnerable function and states the correct hypothesis, and what separates a solved task from an unsolved one is whether that hypothesis becomes a working input.
[LG-36] 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)
*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.
[LG-37] Separation Is a Riesz Energy: How Many Environments Does ICA Need?
链接: https://arxiv.org/abs/2611.06228
作者: Yuda Bi,Yubo Guo,Jiaxiao Cai,Xiaonai Li,Zhuolin Li,Vince D. Calhoun
类目: Machine Learning (cs.LG); Information Theory (cs.IT); Machine Learning (stat.ML)
*备注:
Abstract:Can independent components be recovered when all of them are Gaussian? Yes, if their variances change across environments; but how many environments, and how many samples as the number of components d grows? We answer both for the joint diagonalization of K covariances. Each component has a signature, its variances across the environments. Separating component j costs the Riesz 2-energy E_j of this signature, a sum, over the other components, of inverse pairwise Fisher informations. For Gaussian sources no estimator in O(d) beats cE_j/(n+E_j) at any n and d , and above an explicit sample size a spectral estimator attains E_j/n up to a constant, for every mixing and without the signatures (given a lower bound on the pairwise information). Since signatures live in K-1 dimensions, the rate obeys a dimension law: d^2/n , d\log d/n and d/n for K=2 , 3 and K\ge4 well-spread environments. What counts is the dimension the signatures span, not their number, and \log d well-designed environments make d\log d samples suffice. Projecting the covariances onto one linear combination pays d^2/n at first order instead. In simulations up to d=1000 the predictions hold and joint diagonalization with K\ge4 attains the rate once n/d is in the hundreds; on task fMRI the energy says which components are hard to separate.
[LG-38] Consistent Submodular Maximization with Full Foresight
链接: https://arxiv.org/abs/2611.06223
作者: Eric Balkanski,Paul Dütting,Federico Fusco,Silvio Lattanzi,Ashkan Norouzi-Fard,Ola Svensson,Morteza Zadimoghaddam
类目: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
*备注:
Abstract:Online algorithms with bounded recourse face three intertwined obstacles: uncertainty about the future, computational limits, and a limited budget for changes. We ask a fundamental question that separates consistency from the other two: what does solution stability cost when the entire future is known and computation is unlimited? We answer this question for monotone submodular maximization, where an algorithm must maintain a solution of at most k arrived elements, with a constant number of changes per step. We determine the optimal approximation thresholds: 3/4 for deterministic algorithms and \xi^\star \approx 0.7942 for randomized algorithms. For every \varepsilon0 , we get within \varepsilon of these thresholds with worst-case recourse polynomial in 1/\varepsilon , independent of k and the stream length. We prove matching impossibility results even for coverage functions. We attain the sharp bounds by optimally balancing two opposing forces: preparing for future arrivals and protecting the current solution. For randomized algorithms, we use addition robustness against prefixes of the known future, which we analyze tightly via an interval-selection problem. For deterministic algorithms, we introduce certificates, paths in a fractional relaxation that maintain high value, which we round into an update schedule. Subjects: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG) Cite as: arXiv:2611.06223 [cs.DS] (or arXiv:2611.06223v1 [cs.DS] for this version) https://doi.org/10.48550/arXiv.2611.06223 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-39] Cross-Asset Knowledge Transfer for Early Fault Detection of Data-Scarce Assets Without Target Fault Data
链接: https://arxiv.org/abs/2611.06194
作者: Zekai Wang,Yuxuan Li,Zhangyue Shi
类目: Machine Learning (cs.LG)
*备注:
Abstract:Reliability management of industrial fleets depends on early fault warnings, yet lifecycle data across similar assets are highly imbalanced. A few assets accumulate run-to-failure histories, while newly installed or rarely failing assets provide only normal-condition data. This paper asks how diagnostic knowledge can be transferred from a data-rich source asset to a data-scarce target asset when no fault data have been observed on the target. We formulate cross-asset early fault detection as a zero-shot domain adaptation problem, in which zero-shot refers to the absence of target fault samples, and propose a framework that requires only normal data from the target. The framework comprises (1) a statistical process control step that labels the source asset run-to-failure data with an exponentially weighted moving average chart, reducing manual annotation to verifying flagged segments, (2) an Adapted Domain-Adversarial Neural Network (ADANN) that combines domain-adversarial training and a Maximum Mean Discrepancy penalty with an Adapted Classification-Domain (ACD) loss, which anchors the classifier normal region to the target own normal data, and (3) an ahead-of-time labeling scheme with a prediction horizon that targets incipient faults. Under an exact-classification assumption, a theoretical analysis characterizes the domain-adversarial term for a normal-only target and the alignment state in which source fault features remain separated from target normal features. On a public run-to-failure bearing dataset, ADANN attains the highest mean F1-score among four transfer baselines, with a nominally significant precision gain over the strongest baseline, and its robustness is examined under synthetic noise.
[LG-40] Chiron: Cache Management Empowered by LLM Intelligence for AI Workloads
链接: https://arxiv.org/abs/2611.06179
作者: Bohan Zhang,Renshu Ge,Tianze Wang,Chen Chen
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
*备注: 18 pages, 16 figures
Abstract:AI clusters keep data in remote object stores and rely on local caches to stop GPUs from waiting for data. A cache performs well only when its prefetching and eviction policies match the access pattern, yet these patterns differ across AI workloads and change between phases of one workflow. Existing methods either need code changes, which compromises user transparency, or use low-end heuristics, which cannot follow such changes. Large language models (LLMs) can infer access behavior without user hints, but direct LLM control of cache actions suffers from long reasoning delay, high token cost, and poor decision quality. We present Chiron, a cache system based on advisory switching control. An LLM asynchronously diagnoses the access pattern and recommends a policy. A deterministic harness gates this advice, executes expert-designed policies locally, and returns cache feedback for review. We implement Chiron atop JuiceFS and evaluate it on nine AI workloads. It reduces job completion time by 14.3-41.2% over the best baseline on popular AI workloads.
[LG-41] Network Intrusion Forensic Detection System with Explainable Artificial Intelligence and Digital Chain of Custody
链接: https://arxiv.org/abs/2611.06173
作者: Jose Luis Vela Alonso,Carmen Pellicer
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注: 25 pages, 8 figures, 5 tables. Companion paper to arXiv:2607.00763
Abstract:Deploying an explainable intrusion detection model from a validated laboratory setting into a forensic workflow requires guarantees that benchmarks do not test: that the certified model is the exact artifact applied to a case, that heterogeneous evidence can be classified reliably, and that every step is verifiable. We present MODEXRE, a forensic pipeline that extends a previously validated synthetic-data-trained XGBoost classifier with SHAP explainability (arXiv:2607.00763) into a dual-mode architecture: a Laboratory mode, where models are trained and certified on public and synthetic data, and a Formal mode, where a frozen, hash-verified model performs inference on case evidence under a SHA-256 hash-chained chain of custody. Suricata alerts, CEF firewall logs and packet captures are normalized through OCSF. Four design decisions, each triggered by an observed failure, shaped the architecture: a timestamp leak that inflated Kitsune performance (F1-macro 0.984, 0.662 after correction); a feature-coverage gap that made the 78-feature reference model misclassify a real SSH alert; causal temporal flow aggregation, which raised Kitsune multiclass F1-macro from 0.394 to 0.932; and a certification defect, exposed only by a real packet capture, whereby the pipeline had always trained a binary classifier. After correction, the same capture went from 100% of events flagged as malicious to 87% classified as Normal, with the scanning host identified as Reconnaissance. The default four-class certified model reaches F1-macro = 0.9538 on CICIDS2017-derived evidence; the recertified 18-category model reaches 0.6214, reflecting the real multiclass task. Chain-of-custody verification, model integrity hashing and per-prediction SHAP explanations are reported as measured outcomes.
[LG-42] FaCoDiff: Factor-Correlated Diffusion for Multivariate Probabilistic Forecasting
链接: https://arxiv.org/abs/2611.06166
作者: Jiaxuan Li,Yao Liu,Yuewei Xue,Shijie Dong,Zixiang Jin,Xiaoyu Guo,Qiao Liu
类目: Machine Learning (cs.LG)
*备注: 13 pages, 6 figures
Abstract:Multivariate probabilistic forecasting requires modeling cross-variable dependence alongside marginal uncertainty. The challenge is to learn this dependence while controlling marginal noise scales and avoiding dense covariance costs. We propose Factor-Correlated Diffusion (FaCoDiff), which couples a compact conditional statistical estimator with structured denoising. History-conditioned factor loadings use distinct coefficients for forecast-center correction, scale adjustment and noise dependence. Diagonal compensation preserves the estimated marginal noise variances, while low-rank operators enable residual Gaussian likelihood evaluation and noise transport without dense covariance storage or factorization. Factor-derived feature modulation explicitly conditions the denoiser, and a geometry-aware objective accounts for errors after noise transport. The same conditional geometry governs forward perturbations and reverse sampling. Across nine datasets, FaCoDiff achieves the lowest mean variogram score on eight datasets and ranks among the top two on 43 of 54 dataset-metric combinations in the reported comparisons. Matched ablations on Solar and Weather show that factor covariance improves mean Energy and Variogram scores under all four combinations of center and scale correction. Disrupting cross-variable sample pairings while preserving marginal trajectories worsens joint and aggregate scores on both datasets, demonstrating the predictive value of generated cross-variable dependence.
[LG-43] Regret Bounds for Market Making with Persistent Orders
链接: https://arxiv.org/abs/2611.06159
作者: Nicolò Cesa-Bianchi,Tommaso Cesari,Roberto Colomboni,Luigi Foscari,Vinayak Pathak
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注:
Abstract:We study adversarial market making with orders that remain active until execution. The maker operates on two markets: a primary market where orders are made and a secondary market where the resulting positions are hedged. In each round, the maker posts a quote on the primary market before observing both primary and secondary market prices. An order executes when the primary price reaches its quote, and the resulting position is immediately offset at the secondary price. When both price sequences are chosen by an adversary over T rounds, we give a simple deterministic policy with regret of order \log T against the best fixed quote (static comparator) and \sqrt T against the best quote for each submission date (dynamic comparator), plus two terms measuring exposure to future primary and secondary prices. These rates remain optimal even when both exposure terms vanish and lower bounds show that the exposure terms are necessary. Our policy naturally extends to simultaneous bids and asks, where we show that static and dynamic regret are both characterized (up to constant factors) by the exposures terms (where the primary price exposure V_T^\mathrmtwo now accounts for both bids and asks) and a term of order \sqrtV_T^\mathrmtwo\log T . Our policy is simple, efficient and fully parameterless. We extend this setting by demonstrating how natural assumptions on the price processes lead to improved guarantees without changing the policy.
[LG-44] Selective Prediction Intervals: Jointly Learning to Predict Bound and Abstain
链接: https://arxiv.org/abs/2611.06158
作者: Lidor Rubi,Gilad Katz
类目: Machine Learning (cs.LG)
*备注:
Abstract:Many high-stakes regression decisions ask three things of a model at once: a point estimate to act on, an interval that bounds the risk of acting, and the option to abstain when neither can be trusted. Deep learning provides these in isolation: prediction-interval methods return an interval, and at best a value estimate inside it, for every input, while selective-prediction methods learn when to abstain, but only for point predictions and with no statement of how far the truth may lie from the estimate. We introduce SelPI, which trains interval, value and abstention heads jointly under a single loss with an explicit acceptance-rate budget. A natural objection is that interval width already tells us when to abstain. We show that it does not: across 18 tabular benchmarks, width predicts large point errors well (AUROC 0.72) but is anti-correlated with interval misses (AUROC 0.38), so abstaining on wide intervals discards precisely the inputs the interval got right. SelPI therefore selects with two gates, one trained on the interval loss and one on the value loss, coupled by a Fr’echet penalty that ties their separate budgets to the joint acceptance rate they deliver. With the base predictor held fixed, the learned gate beats width thresholding on a proper interval score on all 18 tabular datasets, closing a median 87% of the gap to an oracle selector, and it also beats selection by deep-ensemble disagreement on every dataset. The advantage grows with the heteroscedasticity of the task, and we report the settings in which it disappears.
[LG-45] Neural Stochastic Forecasting for Lateral Molecular Diffusion Simulations
链接: https://arxiv.org/abs/2611.06157
作者: Jacob Ede Levine,Mario Hidalgo-Soria,Chen Yun Wen,Yun Lyna Luo,Sai Chandra Kosaraju
类目: Machine Learning (cs.LG)
*备注:
Abstract:Predicting the long-time diffusion of membrane proteins is challenging because molecular dynamics (MD) simulations are computationally expensive, while experimental single-particle tracking is often limited by finite observation times. Here, we introduce a neural stochastic (NS) model for forecasting the collective diffusion of membrane proteins from coarse-grained MD trajectories. The model combines a transformer-based representation of protein configurations, recurrent networks operating at multiple temporal scales, and a stochastic latent process with learned memory. A cross-attention decoder then generates correlated protein displacements while accounting for the instantaneous system geometry and periodic boundary conditions. We evaluate the model in two membrane systems: seven interacting PIEZO channels and a mixed system containing seven PIEZO and ten TREK channels. For the seven PIEZO system, the NS model reproduces the anomalous diffusion exponent measured in MD simulations \alpha \sim 1 , substantially improving on a previously developed variational autoregressive model that exhibits superdiffusive behavior. The model also generates trajectories extending to 40,000 frames and captures the distinct diffusion characteristics of PIEZO and TREK populations in the mixed system. Comparisons of time-averaged mean-square displacements, Van Hove displacement distributions, and non-Gaussian parameters show that the model preserves key statistical features of the simulated dynamics. These results establish neural stochastic modeling as a promising approach for generating long trajectories of multi-protein membrane diffusion and extending the statistical analysis of molecular dynamics simulations.
[LG-46] Short-Term Forecasting of Heterogeneous Wind Power Fleets: Global Temporal Fusion Transformer Leverag ing Static Covariates
链接: https://arxiv.org/abs/2611.06130
作者: Viktor Walter,Andreas Wagner
类目: Machine Learning (cs.LG)
*备注: 18th International Green Energy Conference (IGEC-XVIII) the 7th International Conference on Energy and AI (ICEAI-VII)
Abstract:Accurate short-term wind power forecasting is critical for grid stability and electricity trading. While transformer-based deep learning is state of the art, industry practice relies mostly on local models trained for individual assets. This study investigates whether a global temporal fusion transformer (TFT), trained jointly on a heterogeneous portfolio of wind farms across Germany, outperforms site-specific local models in 48-hour-ahead forecasting. Using numerical weather prediction (NWP) data from the regional ICON-D2 and the global ECMWF HRES models, we compare the global TFT against local models and a linear regression baseline. The global model leverages static covariates, such as farm age, altitude above sea level, hub height, rotor diameter and power curve parameters, to learn shared representations across sites. We evaluate on 100 synthetic wind farm profiles and a real portfolio of 11 German wind farms, across two seasonal test periods with intermediate retraining. The global TFT outperforms the local models for 87% of the synthetic and 77% of the real wind farms and clearly surpasses the linear baseline. Wilcoxon signed-rank tests confirm that these gains are significant across both seasons and datasets. Crucially, this superiority depends on static covariates: without asset-specific context, the global model fails to generalize. The TFT’s variable selection network (VSN) reveals the learned feature selection. Altitude emerges as a relevant static discriminator on both datasets, while hub height is decisive on the real portfolio. For dynamic inputs, the encoder prioritizes historical power, whereas the decoder combines upper-level wind speeds from both NWP models, consistent with averaging out model-specific errors. These findings support a shift from single-site to scalable multi-site forecasting, reducing operational complexity while improving accuracy in large portfolios.
[LG-47] Learning When to Think Aloud in Latent Reasoning
链接: https://arxiv.org/abs/2611.06108
作者: Zhehao Zhou,Tongrui Su,Shengyu Zhu,Da Li,Zeyu Liu,Jun Zhu,Xueqi Cheng
类目: Machine Learning (cs.LG)
*备注:
Abstract:Latent Chain-of-Thought (CoT) reasoning improves efficiency by shifting intermediate computation into continuous hidden states. However, fully latent methods often struggle on precision-critical tasks such as mathematical reasoning, while existing hybrid approaches retain substantial token-generation overhead. We introduce VITAL, an importance-guided hybrid reasoning framework that selectively verbalizes informative tokens as reasoning anchors while keeping most computation latent. A lightweight gating head dynamically selects between explicit token generation and direct hidden-state feedback. VITAL constructs hybrid trajectories from explicit CoT using token- importance estimation and trains the model through curriculum-based supervised training, followed by reinforcement learning with a reward that balances correctness and conciseness. Experiments show that VITAL consistently out- performs existing fully latent reasoning methods, achieving relative average accuracy gains of 9.6% on GPT-2 and 5.1% on Llama-3.2-1B at comparable reasoning lengths. Compared with hybrid baselines, VITAL achieves higher reasoning accuracy with substantially shorter reasoning trajectories on Qwen3-1.7B.
[LG-48] Shallow Encoding Emits Deep Memories
链接: https://arxiv.org/abs/2611.06084
作者: Mingyuan Ma,Yangmin Li,Harris Nover,Zheyu Fu,Po-Han Huang,Alex Settle,Ashwin Nanjappa
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注: 47 pages. Mingyuan Ma and Yangmin Li contributed equally
Abstract:Hybrid language models interleave a few softmax-attention layers with many linear-attention or state-space layers, so every layer keeps a memory of earlier tokens: a key-value cache or a fixed-size recurrent state. Serving them spends more compute and high-bandwidth memory on these memories than the output requires. The prefill runs every layer for every prompt token, and every decode step loads and stores whole recurrent states, which on Qwen3.8-Flash-Next take as many bytes per request as the attention cache of thousands of tokens. We find two properties of trained hybrids that locate this excess. First, writes are easier than reads: writing a deep layer’s memory of a token needs little of the deep computation that predicting the next token needs. Trained copies of the deep layers’ write paths, which we call emitters, therefore write a prompt’s deep memories from the residual stream of a shallower anchor layer, and all prompt tokens but the last run as few as 63% of the layers. Second, recurrent states have a state sink: most of their energy lies in a rank-one term along a fixed direction of each head, and the rest is a low-rank working memory. SEED builds on both: it makes the prefill shallow in depth and the decode state shallow in representation. Each recurrent state is stored as its sink plus low-rank factors, which the Gated DeltaNet, Kimi Delta Attention and Mamba-2 updates extend exactly; we prove that the open-loop error of the periodic compressions back to low rank is bounded independently of their number. SEED keeps the original weights frozen, compresses no attention cache, and runs all layers for every generated token. On six checkpoints of five production hybrids of 30B to over 300B parameters, it is nearly lossless on eleven official exams. Served with SGLang on NVIDIA GB300 GPUs on an agentic-coding trace, it raises the per-GPU throughput of Qwen3.8-Flash-Next 1.2-1.4x.
[LG-49] Catastrophic Overfitting is Weight Drift the Training Loss Cannot See
链接: https://arxiv.org/abs/2611.06082
作者: Chao Pan
类目: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
*备注: 12 pages (including appendix), 7 figures, 6 tables. Code available at this https URL
Abstract:Fast adversarial training (FAT) makes robust training affordable by replacing the multi-step attack with a single-step one, but it suffers from catastrophic overfitting (CO): partway through training, robustness to multi-step attacks collapses to near zero while the single-step attack used for training keeps being repelled. We show that CO is a quiet escape of the weights. Across the event the network’s clean behavior quietly improves, and moving from the weights just before CO to those just after it changes the training objective only marginally while destroying most of the robustness. The training objective is therefore nearly blind to the displacement it causes, so it exerts no force to undo it. The exponential moving average (EMA) of the weights, which FAT protocols already maintain, does not accumulate this drift and stays robust through the entire event. This observation suggests a nearly free fix: when validation robustness drops, reset the weights to their EMA. We build this guard on top of a coupling-based base method designed so that its average stays inside the robust region. The resulting training removes CO in every one of our runs and attains the strongest AutoAttack accuracy among fast methods on all three benchmark datasets, including a clear margin on CIFAR-100, where it also clears the multi-step reference at roughly half its cost. Code is available at this https URL.
[LG-50] Offline Learning of Reserve Prices from Censored Bids
链接: https://arxiv.org/abs/2611.06067
作者: Lavanya Marla,Yunzong Xu,Boxuan Zhou
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG); Econometrics (econ.EM); Machine Learning (stat.ML)
*备注: 87 pages, 12 figures, 25 tables. Authors listed alphabetically
Abstract:We study how a seller can learn a reserve price for a second-price auction from censored historical bid data. Because values below past reserves are unobserved, even unlimited data may not reveal which reserve earns the most revenue. Under a monotone hazard rate (MHR) assumption, we characterize exactly when an optimal reserve is identified, even though the value distribution may not be: either revenue has not yet begun to fall at the lowest historical reserve, so the best reserve lies among the observed bids, or the censored fraction is as large as MHR allows, which pins down the entire distribution. When identification fails, we derive the reserve that minimizes worst-case revenue regret over distributions consistent with the data, and prove that this unique minimax reserve and its regret depend on just two observable quantities: the censored fraction and the hazard just above the lowest historical reserve. With N historical bidders, the optimal reserve is learned with expected worst-case regret O(N^-1/2) when it lies among the observed bids, and O(N^-1) from the censored count alone when the censored fraction is as large as MHR allows and the optimum lies at or below the lowest reserve. In the nonidentifiable regime, plugging estimates of the two quantities into the minimax reserve gives expected worst-case regret within O(N^-2/5) of the population minimum under local density smoothness, with a matching minimax lower bound on a smooth class. Beyond MHR, we solve the minimax problem for regular and unrestricted distributions and prove an O(N^-2/5) learning bound in the regular nonidentifiable regime, given known bounds on the boundary parameters. Experiments on synthetic and eBay data show gains from exploiting MHR and the benefits of the robust rule when parametric models are misspecified.
[LG-51] PRISM-AMC: Replay-Free Continual Learning for Automatic Modulation Classification
链接: https://arxiv.org/abs/2611.06066
作者: Usman Akram,Matthew Graham,Haris Vikalo
类目: Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注:
Abstract:Automatic modulation classification (AMC) is a key component of cognitive wireless systems, enabling spectrum awareness and adaptive communication. Most deep learning approaches to AMC assume a fixed set of modulation classes and access to previously observed training data. In practice, modulation classes and channel conditions may evolve over time, while storage or data-retention constraints may preclude replay of raw baseband I/Q samples. We propose PRISM-AMC, a replay-free continual learning framework that mitigates catastrophic forgetting by preserving task-relevant subspaces of intermediate representations via gradient projection memory. We further analyze projected stochastic-gradient updates under an SNR-dependent gradient-noise model. PRISM-AMC is evaluated in both class-incremental and domain-incremental settings. Experiments on the MIGOU and DeepSig 2018.01A datasets show substantially reduced forgetting relative to sequential fine-tuning and elastic weight consolidation, while remaining competitive with or outperforming replay-based RSD-AMC in class-incremental settings, without storing raw I/Q samples.
[LG-52] Learning From Static Context in Federated Time Series Forecasting
链接: https://arxiv.org/abs/2611.06064
作者: Viktor Walter,Andreas Wagner,Fabian Stricker
类目: Machine Learning (cs.LG)
*备注: 4th IEEE International Conference on Federated Learning Technologies and Applications (FLTA 2026)
Abstract:Time series forecasting is shifting from locally trained, per-series models to global models trained across many heterogeneous series, which requires learning what sets the series apart. When the raw data cannot be centralized, such a global model is trained with federated learning (FL), one client per data owner. State-of-the-art forecasters such as the Temporal Fusion Transformer (TFT) tell series apart through a static input that describes each series and does not change over time. In the special case of a federated setting where each client holds a single series, this static input is constant within each client and varies only across clients. We ask whether a federated TFT can learn from it. Per-batch FedAdam, which aggregates gradients after every batch, uses the static input almost as well as central training, whereas communication-efficient schemes with local epochs (FedAvg, FedAdam+Prox) extract far less from it and gain nothing significant on either application dataset. At the same local computation, an Adam-compatible SCAFFOLD keeps the static pathway responsive on wind power and gains significantly there. On building energy its error still drops, but the static pathway falls silent, so drift and communication frequency cannot be fully separated as the cause. Only per-batch FedAdam matches or beats the per-client local models across all datasets. We confirm the effect on building-energy meter data, a wind-power dataset, and a controlled synthetic dataset. The loss is not specific to the TFT: on the synthetic and the wind data two further architectures with different static-conditioning mechanisms show it as well.
[LG-53] Softmax Attention with Any Temperature in Fixed Dimension: Matching the Best Known Maximum-Inner-Product Bounds
链接: https://arxiv.org/abs/2611.06055
作者: Zixuan Huang(1 and 2),Guozhang Chen(1 and 3) ((1) State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University, Beijing, China, (2) School of Physics, Peking University, Beijing, China, (3) National Engineering Research Center of Visual Technology, School of Computer Science, Peking University, Beijing, China)
类目: Data Structures and Algorithms (cs.DS); Computational Complexity (cs.CC); Computational Geometry (cs.CG); Machine Learning (cs.LG)
*备注: 65 pages, 5 figures
Abstract:Attention is central to modern AI models for language, vision, and scientific modeling. However, direct evaluation compares every query with every key, requiring quadratic time in the context length, which limits its scalability to long contexts. Consequently, characterizing the complexity bounds of attention is of fundamental importance. For every fixed dimension d , we give an algorithm for m queries, n key-value pairs, and additive error \varepsilon , with expected time \widetilde\mathrmO\bigl((mn)^k/(k+1)+m+n\bigr),\qquad k=\lfloor d/2\rfloor. For fixed dimension and a fixed number of value channels, soft-O hides factors of the form \operatornamepoly_d\bigl(b,\log(m+n),\log B,\log(1/\varepsilon)\bigr), where b bounds input encoding lengths and B bounds coordinate and value magnitudes. In particular, the dependence on B and 1/\varepsilon is polylogarithmic, which allows arbitrary positive temperature. When m=n and d\ge2 , the bound is \widetilde\mathrmO(n^2-2/(\lfloor d/2\rfloor+1)) : near-linear in dimensions two and three, with exponent 4/3 in dimensions four and five and 3/2 in dimensions six and seven. It improves the previous \widetilde\mathrmO(n^2-1/d) upper bound and matches the best known batch maximum-inner-product bound in each fixed dimension. Additionally, we reduce integer batch maximum inner product computation to one attention call with the same dimension and input sizes m,n . Comments: 65 pages, 5 figures Subjects: Data Structures and Algorithms (cs.DS); Computational Complexity (cs.CC); Computational Geometry (cs.CG); Machine Learning (cs.LG) Cite as: arXiv:2611.06055 [cs.DS] (or arXiv:2611.06055v1 [cs.DS] for this version) https://doi.org/10.48550/arXiv.2611.06055 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-54] Alleviating Limited Effectiveness and Multi-Class Collapse in Deep Graph-level Clustering
链接: https://arxiv.org/abs/2611.06052
作者: Xiaoyang Lin,Renchi Yang,Yuqi Chen,Xin Huang
类目: Machine Learning (cs.LG)
*备注:
Abstract:Given a collection of N graphs, graph-level clustering (GLC) seeks to partition the N graphs into semantically meaningful groups, which serves as a fundamental task in analyzing graph-structured data. While Graph Isomorphism Networks (GINs) have become the de facto encoders for GLC, they still fail to distinguish graphs due to their neglect of global topology. On top of that, existing deep clustering frameworks suffer from severe “clustering collapse” issues, particularly in imbalanced multi-class scenarios. To address these challenges, we propose S ^3 GLC, a novel framework that integrates global topological characteristics with local neighborhood features for balanced GLC. Theoretically, we leverage the Hoffman-Wielandt inequality to construct Spectral Signature as a rigorous lower bound for graph edit distance (GED), providing a computationally efficient proxy for global topological features. We additionally propose a joint manifold-preserving sparse regression module to fuse Spectral Signature with Spatial Representation, eliminating noise while preserving the complementary information. Furthermore, to mitigate clustering collapse, we introduce a balanced clustering objective based on optimal transport, harnessing the Sinkhorn-Knopp normalization to enforce equipartition constraints on cluster assignments. Extensive experiments on 10 real-world datasets demonstrate that S ^3 GLC significantly outperforms 15 state-of-the-art baselines, achieving remarkable improvements on benchmark datasets such as REDDIT-12K and OGBG-PPA.
[LG-55] Act-HiMotif: Hierarchical Daily Activity Schedule Generation via Learned Activity-Structure Motifs
链接: https://arxiv.org/abs/2611.06038
作者: Yushi Bai,Kang Liu,Ling Yin,Wei Wang,Yilin Lyu
类目: Machine Learning (cs.LG)
*备注: 22 pages, 17 figures, including supplementary material
Abstract:Generating daily activity schedules from individual attributes (e.g., age and income) is valuable for activity-based travel modeling and human behavior simulation. The task is challenging because exact activity structures follow a highly imbalanced, long-tailed distribution, while the same structure can admit multiple duration allocations under a fixed full-day budget. Direct generation must therefore learn many exact activity structures from highly uneven observations while modeling their possible temporal realizations. However, many exact structures share recurring activity organization. Based on this observation, we propose Act-HiMotif, a hierarchical motif-based generative model that introduces learned, reusable activity-structure motifs as an intermediate representation between individual profiles and specific activity structures. Act-HiMotif first distinguishes home-only days from active days, then samples a motif from a profile-conditioned prior to guide activity-structure generation, and finally applies a budget-constrained duration allocator to generate activity durations that sum exactly to the full-day horizon. This hierarchy models home-only routing, activity-structure generation, and duration allocation at their corresponding levels, while motifs allow related structures to share statistical support without losing their distinct sequences. Experiments on large-scale travel survey data show that Act-HiMotif outperforms strong baselines in both in-distribution and spatial generalization settings, demonstrating the effectiveness of hierarchical generation and reusable activity-structure motifs for producing realistic profile-conditioned daily activity schedules.
[LG-56] RACK: Multiple Instance Learning for Anti-Money Laundering in Transaction Networks WSDM2027
链接: https://arxiv.org/abs/2611.06035
作者: Seonkyu Lim,Inwoo Tae,Gwangui Hong,Jaehoon Lee,Jeongwhan Choi,Taehoon Kim,Min-Soo Kim,Yongjae Lee
类目: Machine Learning (cs.LG)
*备注: 11 pages, 3 figures. Accepted at the 20th ACM International Conference on Web Search and Data Mining (WSDM 2027)
Abstract:Anti-money laundering (AML) investigations are organized around accounts, while the suspicious evidence for an account may reside in only a few of its transactions. Yet many account-level detectors summarize transactions before learning, and message-passing models may aggregate once more over neighboring accounts. We name this discrepancy the aggregation mismatch and study it under an account-window benchmark label defined by the presence of at least one flagged transaction. We recast account-level AML detection as multiple instance learning, in which an account-window is a bag and its transactions are instances. We propose TRACK, which encodes every transaction with a shared encoder and selectively pools the encodings into an account decision. To emulate coarse supervision, we withhold the transaction labels of five AML datasets from TRACK throughout training. Under a time-ordered protocol, TRACK achieves the highest average precision among the tabular and graph baselines trained from account labels on every dataset. Despite using only account-window supervision, it remains competitive with edge classifiers trained directly from transaction labels. Its instance scores also rank flagged transactions well above chance and approach or exceed directly supervised ranking performance across datasets.
[LG-57] Attraction-Ranking Mismatch in Out-of-Distribution Detection: Theory and Evidence from Hamiltonian Feature Dynamics
链接: https://arxiv.org/abs/2611.06021
作者: Zilin Wang,Lianghai Xiao
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注: 24 pages
Abstract:Attracting features toward learned in-distribution (ID) regions is an intuitive strategy for out-of-distribution (OOD) detection, but does such attraction improve detection? We investigate this question using class-conditioned Hamiltonian dynamics as a controlled testbed and identify an attraction-ranking mismatch: improving individual class-affinity scores need not improve the relative ordering of ID and OOD samples. Our analysis separates misalignment between the driving field and the detection score from ranking failure that persists even when they are aligned. An exact AUROC decomposition and a local boundary analysis identify relative score changes at the ID-OOD ranking boundary as the source of first-order AUROC change under regularity conditions. Experiments demonstrate that a score-aligned rule can improve performance on a held-out source validation split yet consistently degrade it after transfer, even without a single winning-class change. A separate perturbation study finds that near-boundary relative score changes track the direction of small AUROC changes in the tested settings. Finally, centroid approximation and shared scales connect the zero-step score to a static centroid family that includes the existing CTM detector. Together, these results explain why class attraction is insufficient to justify feature refinement and motivate evaluating its effect on the ranking boundary.
[LG-58] Decouple to Reuse: Fine-Grained Knowledge Augmentation for Time Series Foundation Models
链接: https://arxiv.org/abs/2611.05993
作者: Yuqian Wu,Qingxiang Liu,Yuxin Guo,Weifeng Su,Jian Li,Pengfei Xu,Xi Chen,Yutao Yue,Yuxuan Liang
类目: Machine Learning (cs.LG)
*备注: 13 pages, 10 figures
Abstract:Time series forecasting is critical for demand prediction and resource management in modern Web services, yet applying pretrained Time Series Foundation Models (TSFMs) to downstream tasks remains challenging due to domain shifts. Existing retrieval-augmented approaches typically retrieve coarse-grained window contexts and may introduce context-specific noise. To address this problem, we argue that temporal patterns should be reused at a decomposition level, as effective trend and seasonal dynamics may originate from different contexts. We propose \method, a two-stage \underlineFine-grained \underlineKnowledge \underlineAugmentation framework that decouples and consolidates pre-training knowledge into compact, task-shared residual codebooks to enhance frozen TSFMs. Specifically, Stage I constructs dual-branch residual codebooks for decoupled trend and seasonal signals, while Stage II dynamically retrieves and merges these fine-grained patterns into target representations via task-adaptive aggregation and gate fusion. Extensive experiments across eight real-world benchmarks with four representative TSFM backbones demonstrate the superior effectiveness and computational efficiency of \method, while further validating the strong transferability of the residual codebooks across different backbones. The source code is available at the anonymous repository: this https URL
[LG-59] A Controlled Experimental Evaluation of Configuration and Cache Architecture in LLM Serving Engines
链接: https://arxiv.org/abs/2611.05991
作者: Muhammad Faizan Anwar,Sina Keshvadi
类目: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Performance (cs.PF)
*备注: 29 pages, 14 figures, 7 tables. Artifact: this https URL
Abstract:Multi-agent LLM applications repeat most of their prompt tokens across requests, and serving engines reuse the cached state of shared prefixes. vLLM and SGLang reuse it at different granularities, and benchmarks of their default deployments report large performance gaps, usually credited to the cache. Such comparisons rarely control for other configuration differences. One of them, whether a long prompt is prefilled in chunks, is set by each engine’s defaults from the hardware and the release, and it confounds the comparison. We built a reproducible testbed on Kubernetes and a seeded multi-agent workload generator, and compared the two engines under vendor defaults and with chunked prefill matched. SGLang’s configuration is the same in both arms, which makes it an internal control across independently provisioned deployments. Under defaults, vLLM’s median time to first token is up to 64x SGLang’s. Matching one setting removes 78% of the capacity gap where that gap is largest and 30 to 67% of the latency gap. Near saturation the remaining latency gap reflects queue growth. A fluid queue model driven only by a 4.4% throughput shortfall reproduces it, and shows that such a gap scales with the length of the benchmark run. What survives matching is a tail-latency advantage of 2.4 to 6.4x at equal throughput, consistent with cache-aware ordering of waiting requests. We derive six recommendations for comparing serving engines.
[LG-60] A Comparative Study of Importance Attribution to Columns in Deep Learning over Relational Databases
链接: https://arxiv.org/abs/2611.05986
作者: Hila Levi-Yosefi,Shunit Agmon,Ester Livshits,Agapi Rissaki,Benny Kimelfeld
类目: Machine Learning (cs.LG); Databases (cs.DB)
*备注:
Abstract:Deep learning over relational databases has led to increasingly effective prediction models, motivating a growing need to explain their behavior. We study explanation methods that assign importance scores to individual database attributes, focusing on a fundamental question: to what extent does the choice of importance-attribution method matter? We develop a methodology to systematically compare importance-attribution methods and analyze their similarities and differences. In particular, we study two aspects of the explanations: which attributes they deem important, captured by differences in attribute rankings, and how importance scores vary across attributes, captured by differences in score distributions. We further investigate whether these differences are reflected in the conventional measure of explanation quality, fidelity. We apply our methodology in an extensive empirical study of four methods: Monte Carlo Shapley, Kernel SHAP, single-attribute impact, and mask learning. We experiment with multiple datasets and tasks using two RDL architectures: heterogeneous GraphSAGE and Relational Graph Transformers. Our experiments show that explanation methods behave differently depending on the prediction type: regression or classification. In regression, most methods exhibit high agreement in attribute rankings, whereas classification yields considerably less consensus. Notably, while mask learning frequently diverges in its importance distributions, Shapley-based methods and single-attribute impact often converge to similar explanations. This suggests that, in many settings, single-attribute impact can replace Shapley at substantially lower computational cost. Overall, our study provides methodological tools for comparing importance-attribution methods, and empirical evidence that their choice has important consequences for interpreting relational deep learning models. Subjects: Machine Learning (cs.LG); Databases (cs.DB) Cite as: arXiv:2611.05986 [cs.LG] (or arXiv:2611.05986v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2611.05986 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-61] Reward Budgeting Reduces Premature Convergence in Reinforcement Learning for LLM Reasoning NEURIPS2026
链接: https://arxiv.org/abs/2611.05942
作者: Mengni Jia,Mengyu Zhou,Xiaoxi Jiang,Guanjun Jiang
类目: Machine Learning (cs.LG)
*备注: Accepted to NeurIPS 2026
Abstract:RL for LLM reasoning often reaches an early performance plateau as the policy distribution becomes increasingly sharp. Recent work has studied policy entropy as both a diagnostic and a control signal for this behavior. In our experiments, however, increasing entropy alone does not reliably change the eventual plateau. We analyze RL through its implicit target policy and derive a sharpness factor \kappa , together with an effective counterpart \kappa_\mathrmeff , that describes how strongly reward or advantage gaps are translated into probability-ratio gaps. Motivated by this view, we introduce \emphreward budgeting, a plug-in advantage construction that assigns a fixed total reward budget to each prompt. For binary rewards and independent on-policy rollouts, its unclipped score-function estimator follows the gradient of an exact finite- G concave objective with diminishing returns, while its advantage gap decreases as more rollouts become correct. For general rewards, a shifted log-mean interpretation holds asymptotically under the stated regularity conditions. These results motivate, rather than establish exact objective equivalence for, the clipped implementations. On Qwen3-8B and Llama3.2-3B-Instruct, reward budgeting consistently improves a range of RL baselines, with gains up to +12.50 on AMC23 and +10.00 on AIME24 and AIME25. Overall, our results support an update-dynamics view of premature convergence and show that strong gains can be obtained without explicitly preserving entropy.
[LG-62] Diversity-Aware Experience Replay via Statistical Leverag e Scores
链接: https://arxiv.org/abs/2611.05929
作者: Shaifalee Saxena,Alexander Scheinker
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注:
Abstract:Off-policy deep reinforcement learning trains on random samples from a replay buffer of past experience. In physical robot settings, that buffer is dominated by routine: thousands of near-identical transitions from normal operation, while the rare events that define the edges of the task are rarely sampled. We propose scoring every stored transition by its statistical leverage, which measures how unusual it is relative to the buffer’s own data, computed from the stored numbers alone, with no labels, rewards, or value network. We draw each mini-batch from a mixture of this score and uniform sampling. First, the natural way to build the score along the directions where the data varies most turns out to be blind when actions are stored as one-hot indicators, because indicator columns always win the variance ranking; we instead select the feature columns with the heaviest tails (largest kurtosis) where small far-away populations are revealed. Second, the importance weights that prioritized replay uses to keep gradients unbiased cancel the method exactly: with full correction, training matches uniform sampling, so we drop the correction and deliberately optimize the rebalanced objective. Third, when the rare situations of a log are relational, visible only in how a stretch of behavior relates to its goal, not in any single stored coordinate-per-transition scores are blind; applying the same tail rule to short trajectory windows described relative to their goal restores the signal, and on a three-joint arm that must push an object around an obstacle it raises the rare-task success from 0.49 to 0.69 against a labeled-oracle ceiling of 0.74.
[LG-63] Improved approximation algorithms for capacitated cluster editing
链接: https://arxiv.org/abs/2611.05916
作者: Cong Ma,Guangting Chen,Yong Chen,Guohui Lin,Jiaxuan Ma,Lusheng Wang
类目: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
*备注: 16 pages, comments welcome!
Abstract:We study the minimum-disagreement cluster editing problem that, from a complete edge-signed graph, where a positive edge represents two vertices agreeing to each other while a negative represents a disagreement, constructs disjoint cliques of upper bounded size such that the total number of negative edges inside cliques and positive edges outside cliques is minimized. When the sizes of the candidate cliques are unbounded, the problem is NP-hard and APX-hard, and many approximation algorithms with improving ratios have been proposed in the literature and the state-of-the-art approximation ratio of \alpha = 1.3865 + \epsilon . The capacitated variant, in which the candidate cliques have sizes no greater than a given k , is less studied and the best approximation ratio is 5.37 . In this paper, we study the latter problem. We first present a downsizing procedure that achieves an approximation ratio of \alpha_k = 2 \alpha + 1 , based on any \alpha -approximation algorithm for the uncapacitated problem. Secondly, we propose a hybrid of cluster-LP sampling and pivoting and show that for every fractional capacitated cluster-LP solution of cost L , the hybrid using ordinary pivoting has an expected cost at most \frac 95 L + \frac 110 \OPT_k , and the hybrid using conditional pivoting has an expected cost at most \frac 306203 L + \frac 50203 \OPT_k . A cardinality-constrained adaptation of the approximate dual-separation framework of García-Soriano and Schohn supplies the required fractional solution and yields a randomized (1.9 + \epsilon) -approximation algorithm and a randomized (1.7537 + \epsilon) -approximation algorithm, respectively.
[LG-64] Retrieval Supersedes Scale: Efficient Log Level Prediction via Multiplex Socio-Technical Context
链接: https://arxiv.org/abs/2611.05901
作者: Youssef Esseddiq Ouatiti,Mohammed Sayagh,Bram Adams,Ahmed E. Hassan
类目: oftware Engineering (cs.SE); Machine Learning (cs.LG)
*备注:
Abstract:Developers insert logging statements in source code to record its execution, and the resulting logs are essential for detecting and troubleshooting system anomalies. A key part of these statements is the log level (e.g., DEBUG, INFO, WARN, ERROR), which determines whether a message is recorded. Choosing it is a trade-off: levels set too low mask important information in production, while levels set too high create noisy logs and extra I/O and storage overhead. Recent work uses Large Language Models (LLMs) with In-Context Learning (ICL) to predict log levels, but typically augments the prompt with randomly sampled examples. We argue that this overlooks the socio-technical nature of logging, where conventions vary by component and development team. Off-the-shelf LLMs using random examples are reported to achieve median AUCs of only 0.56 to 0.76, leaving room for improvement without fine-tuning. In this paper, we propose a retrieval approach for LLM-powered log level prediction that prioritizes context relevance over random selection and basic similarity. Our method retrieves logging examples relevant along two socio-technical dimensions: (i) functionality (code performing similar operations) and (ii) ownership (code maintained by the same developers). Evaluating four LLMs across ten popular Java projects, we find that this retrieval improves AUC over random baselines by 0.07 (median), and combining both signals yields median AUCs of 0.83 to 0.97. Our ablation studies show that each signal improves on random retrieval, and that combining them outperforms the stronger of the two on all ten projects. Our results also significantly exceed state-of-the-art supervised predictors (e.g., DeepLV) by a median AUC of 0.12 to 0.19. Our findings highlight the importance of socio-technical signals such as code functionality and ownership for LLM-powered log level prediction.
[LG-65] When Is One Query per Template Enough? Adapting Learned Cost Models to New Hardware KR
链接: https://arxiv.org/abs/2611.05892
作者: Krish Patel
类目: Databases (cs.DB); Machine Learning (cs.LG); Performance (cs.PF)
*备注: 13 pages, 2 figures, 10 tables. Code and data: this https URL
Abstract:A learned cost model that predicts query runtimes is trained on measurements from one machine. When the database moves to other hardware, its plans stay the same but their runtimes change, each by its own factor. We ask how to adapt such a model from k queries measured on the new machine: which queries to run, how to use their runtimes, and how many are enough. On PostgreSQL we timed the same plans of 1,162 TPC-H and TPC-DS queries in six environments: a laptop and a server, each with all its processors or pinned to one, the server also pinned to two, and the laptop also with a small buffer pool. No single factor describes the slowdown between two environments: most of its variance lies between query templates, so a speed factor stops improving after a few queries. A correction that fits separate lines for serial and parallel plans and adds an offset for every measured template brings the moved model, with one measured query per template, to an error (geometric mean over templates of the median Q-error) of 1.10 on TPC-H and 1.19 on TPC-DS, close to the same correction on a model trained on the new machine; fine-tuning a DACE-style neural cost model on the same queries gains at most 2%. On templates never measured, no correction we tried is on average more than 2% better than the speed factor. We then fixed hypotheses and timed new data. On TPC-H at ten times the scale the errors follow these results, but two of five repeated tests miss. On the Join Order Benchmark, whose queries differ widely within a template, the model errs by 2.36 even when retrained on the new machine, no correction gains much, and only the test on new templates holds; boosted trees and a neural model fit these queries better, and correcting the neural model works as well as fine-tuning it. We report every test as registered.
[LG-66] MIGONet: a multiple-input graph neural operator with solver-free discrete manufactured-solution training
链接: https://arxiv.org/abs/2611.05876
作者: Muhammad Hasnain,N. K. Anand,Arun Srinivasa
类目: Machine Learning (cs.LG)
*备注: Accepted for publication in Transport Phenomena by De Gruyter Brill. 35 pages, 16 figures, 13 tables
Abstract:Surrogate modeling of transport requires operators that capture the combined influence of spatially distributed sources and Dirichlet and Neumann boundary conditions, together with training datasets that can be generated at low computational cost. This work presents MIGONet, a multiple-input graph neural operator, and a discrete method of manufactured solutions (MMS) that trains it without solving a partial differential equation, using steady heat conduction as the model problem. In the discrete MMS procedure, temperature fields are prescribed a priori, and the source terms and Dirichlet or Neumann boundary data are evaluated with the same discrete operators as the reference finite-difference method (FDM), so every training sample satisfies the discrete governing equations by construction. Because the boundary data are derived from each field, a Dirichlet or Neumann condition can be assigned to every boundary segment of every sample, and the boundary specification becomes an operator input. MIGONet supplies representations of the complete source field and boundary data to every node before local message passing. On the rectangular domain, it is compared with a version without the complete-source encoder (MIGONet-local) and with the multiple-input operator network MIONet, using the same training data. MIGONet achieves mean relative temperature errors of 10.31% and 10.64% for curved-wire and Gaussian-spot heating, compared with 45.97% and 48.34% for MIONet and 72.25% and 67.82% for MIGONet-local. For 10,000 samples, discrete MMS reduces data-generation time from 486.65 s to 3.15 s relative to matched iterative FDM labeling. Separately trained MIGONet models are also evaluated on L-shaped, disc-shaped, and ring-shaped domains.
[LG-67] On Convergence of Gradient Temporal-Difference Learning with Regularized Corrections and Eligibility Traces
链接: https://arxiv.org/abs/2611.05869
作者: Jingyi Jin,Martha White,Hao Wang
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注: 66 pages, 1 figure. Submitted to the Journal of Machine Learning Research
Abstract:Temporal-difference (TD) learning underlies many prediction and control algorithms, but can diverge when combined with off-policy learning, bootstrapping, and function approximation. TD with Regularized Corrections (TDRC) is theoretically sound and practically simple, but existing convergence theory covers only the one-step case under i.i.d. sampling, while the Huber Bellman-error formulation has not been extended to TDRC or eligibility traces. In this work, we establish convergence of constrained TDRC( \lambda ) with diminishing stepsizes and non-i.i.d. off-policy data, for state-dependent and history-dependent \lambda -parameters. We also propose TDRC( \lambda )-Huber, which preserves the TDRC( \lambda ) fixed point while reducing sensitivity to large Bellman errors. For both algorithms, we apply an ODE-based weak-convergence method from stochastic approximation theory. For the nonlinear mean dynamics of TDRC( \lambda )-Huber, we give an integral small-gain condition that guarantees uniqueness and global asymptotic stability of the target equilibrium. We then establish convergence of the constrained algorithm, derive computable sufficient conditions for convergence, and identify a regularization-based convergence mechanism shared by linear TDRC( \lambda ) and TDRC( \lambda )-Huber. When the mean TD matrix is positive stable, both constrained algorithms converge under sufficiently strong regularization without requiring positive definiteness.
[LG-68] A low-rank decoder bottleneck bounds reliability in foundation-model perturbation prediction NEURIPS2026
链接: https://arxiv.org/abs/2611.05849
作者: Zheyu Zhang,Yuanhao Huang,Fan Feng,Jie Liu
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: Accepted at NeurIPS 2026
Abstract:Foundation-model predictors of perturbation response are evaluated primarily by aggregate accuracy. Using their predictions to prioritize experiments requires assessing each prediction’s reliability. We audit the state-transition (ST) decoder used in STATE across three encoders (SE-600M, scGPT, and Tahoe-x1) and four cell lines under CRISPRi. The decoder’s first attention layer has a low-rank output. Projecting this output onto 4-32 fixed directions estimated from training activations preserves mean per-perturbation Pearson delta correlation (PDC). This low-rank attention bottleneck bounds the decoder’s reliability by distance to training responses: PDC declines as the cosine distance from a perturbation’s observed response to its nearest training response increases. We propose predicted-NN-distance, the cosine distance from a predicted response to its nearest training response, as a per-prediction confidence score. The score has a mean absolute Spearman correlation of 0.81 with PDC, close to the 0.88 obtained with the observed response. Thresholds calibrated on this score transfer across encoders and cell lines, and the score remains predictive of PDC for drug perturbations.
[LG-69] Pathwise-Budgeted LLM Routing: Online Allocation with K-Bit Tail-Cost Learning
链接: https://arxiv.org/abs/2611.05835
作者: Hanzhang Qin,Sophie H. Yu,Yuewei Zhang,Zhi-Hai Zhang
类目: Machine Learning (cs.LG); Optimization and Control (math.OC)
*备注:
Abstract:Firms using large language models under fixed budgets must decide which model to call and whether to retry a failed task. Because a call’s charge is revealed only after completion, one long response can exhaust the remaining budget and leave later tasks unserved. Our objective is to maximize the number of tasks solved before service stops. We develop a contextual bandit policy, ADP-HT, which learns each model’s task-specific success probability and constructs conservative cost estimates from coarsened response-length labels on a quantile grid with 2^K intervals. The policy weighs predicted success against these costs, adjusting the cost penalty to realized spending and the budget available per remaining task. Under calibration, local regularity, and grid-coverage conditions, we bound the excess allowance in the policy’s conservative cost estimates. At a fixed sample size, increasing K reduces the discretization term but increases the estimation allowance. At the tightest tested budgets, ADP-HT solves more than twice as many MATH500 and GPQA tasks as the best implemented baseline. It achieves the highest mean cumulative score at ten of fourteen RouterBench budgets. On LiveCodeBench, it outperforms a deployed commercial router in both default and cost-saving modes at the three lowest test budgets.
[LG-70] ZipDrift: Generating Multimodal Short-Horizon MPC Costs with One-Step Drifting Models
链接: https://arxiv.org/abs/2611.05793
作者: Ke Zhang
类目: Robotics (cs.RO); Machine Learning (cs.LG); Systems and Control (eess.SY)
*备注: 6 pages, 2 figures, 4 tables. Preliminary simulation study
Abstract:Short-horizon model predictive control (MPC) is fast but myopic. ZipMPC compensates by learning a context-dependent correction of the short-horizon cost that imitates a long-horizon expert. We observe that the optimal correction is often multimodal: when a slower opponent can be passed on either side, good corrections of opposite sign coexist, and the expert’s choice between them hinges on tiny cost differences. Deterministic regression then averages modes or inherits inconsistent expert tie-breaking. We propose ZipDrift, which trains a one-step drifting generator over MPC cost corrections, tilted toward low closed-loop cost and learned only from cost evaluations of candidate corrections, without differentiating through the optimiser. A few generated candidates are ranked by a learned critic, and the model-based MPC then enforces dynamics and constraints. In a kinematic overtaking benchmark derived from ZipMPC, 53% of contexts are bimodal and the manual short-horizon MPC crashes in 28.5% of test episodes. With an eight-dimensional stagewise correction, ZipDrift is collision-free in all 600 test episodes and matches the best evaluated candidate, whereas an idealised ZipMPC-style regressor crashes in 4.0% (16.0% with inconsistent expert labels) and critic-ranked random candidates with the same budget in 2.0%.
[LG-71] ming is Everything: Inferring Temporal Membership in Continually Updated Models
链接: https://arxiv.org/abs/2611.05781
作者: Jie Fu,Xiangpeng Deng,Yuan Hong,Wendy Hui Wang
类目: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
*备注:
Abstract:Continual model updating has become a common practice in the deployment of machine learning systems. However, the dynamic training behaviors give rise to new risks not sufficiently addressed by existing studies. In this paper, we present the first systematic study of temporal privacy risk in continually updated deep learning models. We formalize the notion of temporal membership in continually updated models. We then define the temporal membership inference attack, which seeks to infer whether and when individual samples were involved in model training, at both timestamp-level and interval-level granularities. To this end, we propose TMIA, a novel black-box attack that leverages bidirectional LSTMs to infer temporal membership from model outputs observed across successive updates. Through extensive experiments on three language models and three image classification models, we show that TMIA is highly effective and significantly outperforms existing baselines. Finally, we evaluate three potential defenses, revealing the trade-off between privacy protection and model utility.
[LG-72] Optimizing Recourse Under Predictive Multiplicity
链接: https://arxiv.org/abs/2611.05774
作者: Shahin Jabbari
类目: Machine Learning (cs.LG)
*备注:
Abstract:Algorithmic recourse identifies modifications that can change an unfavorable model prediction into a favorable one. Under predictive multiplicity, similarly performing models can disagree on individual predictions, so recourse valid for one model may fail for another. We formulate robust recourse under predictive multiplicity as a budget-constrained minimization of the worst-case validity loss across a competing set of models, which can include models with different architectures or from different hypothesis classes. We propose an oracle-efficient algorithm called ORPM and prove sufficient conditions for its convergence. We evaluate ORPM on one synthetic and five real datasets, and competing sets consist of linear, neural, mixed, and decision-tree models. Our experiments show decreasing optimization-gap bounds and high validity in many tested settings, with favorable performance against baselines designed for specific model classes or architectures. We also quantify how requiring a common action across competing models changes recourse success at a given budget and examine disparities in this success across demographic groups.
[LG-73] Self-Supervised Learning of Universal Latent Spaces for Operator Learning
链接: https://arxiv.org/abs/2611.05735
作者: Sepehr Mousavi,Levi Lingsch,Siddhartha Mishra,Laura De Lorenzis
类目: Machine Learning (cs.LG)
*备注:
Abstract:Neural operators fit their representation of functions to the task they are trained on, so every new task rebuilds a reduced space from labeled data. Following the self-supervised pre-training that made representations shareable in natural language processing and computer vision, we decouple representation learning from operator learning. A functional autoencoder is pre-trained on analytic functions synthesized on the fly, with no solver and no labels, to represent functions on arbitrary point clouds, and the latent space it induces is frozen and reused across tasks. The representation is exactly equivariant to translation, dilation, and amplitude, and reconstructs solutions of partial differential equations (PDEs) and natural images on geometries it never saw. Being task-agnostic, the latent space can be interpreted in its own right. The whole space is used and decodes to valid fields, it preserves the ordering of distances between physical fields, its geodesics interpolate physical evolutions, and edit directions transfer across fields. With frozen representation maps, a dataset is encoded once and a neural operator is trained entirely on cached embeddings, which saves memory and training time. On 28 tasks in fluid dynamics, solid mechanics, acoustics, and the Helmholtz and Poisson equations, this paradigm can exceed the accuracy of state-of-the-art end-to-end architectures, with up to an order of magnitude higher throughput and memory savings that fit models nearly two orders of magnitude larger on consumer-grade GPUs. At inference, latent rollouts run up to an order of magnitude faster than regular rollouts, and chunked inference processes functions of over half a billion points on a single consumer-grade GPU.
[LG-74] RESHPart: Feedback-Reshaped Spectral Embedding for Hypergraph Partitioning
链接: https://arxiv.org/abs/2611.05729
作者: Rongjian Liang,Zhuo Feng,Brucek Khailany
类目: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
*备注: Accepted to ASP-DAC 2027
Abstract:With scalable Laplacian solvers now available, the next frontier is to exploit them more effectively for high-quality hypergraph partitioning. In this work, we revisit spectral hypergraph partitioning from the perspective of spectral embedding geometry. Through a systematic analysis of circuit hypergraphs and controlled synthetic (hyper)graphs, we identify a ray-like embedding structure that helps explain the effectiveness of projection-based partitioning. We further show that the geometry of spectral embeddings can be systematically influenced through Laplacian design and hyperedge reweighting. Guided by these insights, we propose FRESHPart, a feedback-reshaped spectral partitioning framework that refines the spectral embedding through Geometry Feedback and Search Feedback, producing embeddings that are better aligned with projection-based partitioning. On the Titan23 benchmark, FRESHPart achieves average cut reductions over hMETIS of 13.09% at \epsilon=2% and 25.21% at \epsilon=20% , surpassing the strongest previously reported results (12.16% and 22.53%, respectively), while completing every benchmark in under 120 s on a single GPU.
[LG-75] Give the Drafter a Second Chance to Guess: Parallel Reconsideration for Speculative Decoding
链接: https://arxiv.org/abs/2611.05724
作者: Hoang Anh Duy Le,Huibing Dong,Alireza Khadem,Ziteng Sun,Xiang Song,Yizheng Huang,Soroush Radpour,Anshumali Shrivastava,Ye He
类目: Machine Learning (cs.LG)
*备注:
Abstract:Speculative decoding accelerates large language model inference with a small drafter, whose per-guess acceptance rate determines the achievable speedup. Feature-level drafters share a design choice: every token, easy or hard, is drafted under a single forward pass. Many of the tokens the target rejects are ones the drafter is itself unsure about yet the target answers decisively. To address this, we introduce Parallel Reconsideration: guided by its own uncertainty, the drafter reconsiders its hesitant guesses with extra passes that run in parallel with the one-pass draft, adding negligible latency. A co-designed training recipe and serving-time drafting algorithm append these reconsidered guesses to the original draft tree as refined candidates. Across models and benchmarks, on average, Parallel Reconsideration improves over the baselines by 5.3% to 10.2% on acceptance length and by 3.5% to 9.2% on speedup.
[LG-76] CoReFlow: Consistency-Relaxed Flow Maps for Probabilistic Traffic Forecasting
链接: https://arxiv.org/abs/2611.05715
作者: Pengnan Chi,Zhiguo Zhang,Xiaoliang Ma
类目: Machine Learning (cs.LG)
*备注: 12 pages, 8 figures
Abstract:Probabilistic traffic forecasting is essential for reliable Intelligent Transportation Systems (ITS), as it captures uncertainties in complex road networks. Recent diffusion- and flow-based generative models have facilitated uncertainty modeling through a multi-step sampling process. However, beyond slow inference, this process is designed to capture complex multimodal distributions, which misaligns with the limited complexity of the conditional traffic data distribution, causing models to underperform deterministic baselines. To address this, we introduce the \underlineConsistency-\underlineRelaxed \underlineFlow Map (CoReFlow) model, presenting the first effort to leverage the one-step generation capability of flow maps, known as solution operators of probability-flow ordinary differential equations. Besides, we improve accuracy using a concise training objective that supervises the model via interpolation endpoints over finite time intervals. Unlike existing methods (e.g., self-distillation and MeanFlow Identity) that maintain trajectory consistency to model multimodal distributions, our objective leverages conditional mean to restrict the consistency, thereby achieving higher accuracy in the probabilistic traffic forecasting task. Extensive experiments on four real-world datasets show CoReFlow achieves state-of-the-art forecasting accuracy. It surpasses recent baselines, provides competitive uncertainty modeling, and reduces inference costs compared to existing generative baselines.
[LG-77] Neural networks and games
链接: https://arxiv.org/abs/2611.05706
作者: Philip R. Neary
类目: Neural and Evolutionary Computing (cs.NE); Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG); Theoretical Economics (econ.TH)
*备注:
Abstract:Hopfield neural networks and noncooperative games provide seemingly different approaches to understanding how collective outcomes emerge from local interactions. In a neural network, individual units adjust their states according to local signals; in a game, players choose actions to maximise individual payoffs. I establish an exact equivalence between symmetric binary Hopfield networks and binary-action games that admit a quadratic potential. The equivalence identifies stable network configurations with pure-strategy Nash equilibria, and asynchronous neural updating with best-response dynamics. More broadly, the storage and retrieval of memories in neural networks correspond, respectively, to the design and selection of equilibria in games.
[LG-78] Revisiting Critic Estimation in Soft Actor-Critic Algorithms
链接: https://arxiv.org/abs/2611.05666
作者: Jan Felix Kleuker,Koen Ponse,Aske Plaat,Thomas M. Moerland
类目: Machine Learning (cs.LG)
*备注: 10 pages
Abstract:Soft Actor-Critic (SAC) is a widely used deep reinforcement learning algorithm that combines entropy regularization with an actor-critic architecture. Standard implementations rely on one-step temporal-difference learning of Q-functions, yet alternative critic design choices remain underexplored. In this work, we revisit three such choices: learning soft advantages instead of soft Q-values, the use of multi-step return targets, and the training of multi-step critics from replay buffer data. Keeping the policy objective fixed, we compare the six critic variants these choices generate: (i) advantage and (ii) dueling-advantage critics learned from one-step transitions, and \lambda -return and soft Generalized Advantage Estimation (GAE) critics learned either (iii, iv) on-policy or (v, vi) from replayed trajectory segments. Each variant is evaluated across an entire random hyperparameter search on fourteen tasks, treating the hyperparameters as part of the algorithm. Learning soft advantages instead of soft Q-values does not improve performance, and the soft GAE critic adds nothing over the simpler \lambda -return critic. What does improve performance, most clearly on continuous control, is training a \lambda -return critic on replayed trajectory segments, even without importance weighting or trace cutting. This extends earlier evidence that uncorrected multi-step targets work well under replay to the full soft \lambda -return, entropy terms included. By design, the underlying ideas studied here extend beyond SAC, and our empirical findings aim to encourage their application to a broader class of reinforcement learning algorithms.
[LG-79] Adaptive Domain Decomposition for Finite Basis Physics-Informed Neural Networks
链接: https://arxiv.org/abs/2611.05659
作者: Sarah Helfert,Wanxin Chen,Amanda A. Howard,Panos Stinis,Alexander Heinlein
类目: Machine Learning (cs.LG)
*备注:
Abstract:In this work, we present adaptive domain decomposition finite basis physics-informed neural networks (ADD-FBPINNs). The proposed method integrates Partition of Unity Networks (POUNets), a data-driven approach for learning domain partitions, with finite Basis physics-informed neural networks (FBPINNs) to perform more efficient domain decomposition and achieve more accurate approximations. FBPINNs improve upon the accuracy of standard physics-informed neural networks (PINNs) through user-defined subdivisions of the computational domain. By adapting the domain decomposition based on the solution’s features using a separable POUNet, the framework can automatically learn partitions that are stable and can be geometrically interpretable. We test this approach on the two-dimensional Helmholtz equation, two-dimensional Poisson equation with multiscale features, and viscous Burgers equation and observe improved accuracy compared with FBPINNs with fixed domain decomposition.
[LG-80] Beyond Accuracy: Information-Theoretic Decoder Evaluation Under Latent Intent
链接: https://arxiv.org/abs/2611.05658
作者: Tasha Kim,Benjamin Ballyk,Anson Lei,Ingmar Posner,Oiwi Parker Jones
类目: Machine Learning (cs.LG); Robotics (cs.RO)
*备注:
Abstract:A central challenge in assistive brain-to-robot control is decoding a user’s latent intent from complex and noisy brain signals. Prior approaches evaluate decoders as single-trial classifiers with accuracy-based metrics, but this neglects the inherently sequential nature of assistive control. Since a planner must integrate intermittent, noisy decoder outputs over time, single-step accuracy can only be a proxy when predicting downstream task performance. We formalize brain-to-robot control as a partially observable Markov decision process (POMDP), and prove that there is a sharp phase transition: regret is uniformly bounded across the horizon as long as pairs of goals are all decoder-distinguishable, but grows linearly otherwise. We show that a single channel statistic, the worst-pair Chernoff information \Gamma_\min sets this boundary. Then, we derive a closed-form regret predictor U_\mathrmid© , that can be computed from any confusion matrix in O(K^2) time. Across two domains spanning orthogonal information channels, we show that U_\mathrmid predicts cumulative planning regret beyond what accuracy can explain. We then highlight a pathological accuracy trap, in which two decoders have identical validation accuracy 35.1% on real EEG data (BNCI2014001 subject 5) but result in a difference of 9.6\times in held-out regret, which U_\mathrmid predicts. Same-signal control confirms U_\mathrmid 's advantage with decoders that span independent feature channels. Ultimately, our work provides a drop-in diagnostic to evaluate and select sensors and decoders in sequential assistive control.
[LG-81] When Are Two SAE Latents the Same Feature?
链接: https://arxiv.org/abs/2611.05655
作者: Alexandra Suvorikova,Ilgam Latypov,Anton Trukhan,Yuriy Dorn
类目: Machine Learning (cs.LG)
*备注:
Abstract:Comparing sparse autoencoders (SAEs) requires connecting feature correspondence, differences between whole dictionaries, and uncertainty from finite evaluation data. We develop a framework that connects these tasks through an energy-weighted optimal-transport plan under a specified similarity geometry. The same plan defines a dictionary-level discrepancy, decomposes it into latent contributions, and quantifies how each latent’s mass is distributed across counterparts. We characterize relationships between decoder, activation, and first-order causal similarities; idealized models clarify the interpretation of splitting and absorption. For fixed trained SAEs, we derive joint asymptotic inference for transport costs and fractions, accounting for activation similarities and both energy marginals estimated from the same documents. A joint document bootstrap gives asymptotically valid simultaneous confidence bands. Experiments across four language models locate cross-seed discrepancy and separate context-dependent changes.
[LG-82] LACS: Distribution-Sensitive Label-Agnostic Confidence Selection for Dynamic Data Pruning
链接: https://arxiv.org/abs/2611.05616
作者: Heejin Choi
类目: Machine Learning (cs.LG)
*备注: 40 pages
Abstract:Training on large datasets motivates selecting which examples receive expensive training computation. Fixed quotas determine the amount of pruning in advance, while absolute confidence cutoffs depend on the score scale. We study a population-relative alternative that adapts the threshold and retained count together. We introduce LACS (Label-Agnostic Confidence Selector), a training-time rule based on maximum predictive confidence, requiring neither labels at selection time nor per-example gradients. LACS removes examples at or above a log-mean-exp reference plus a margin. We characterize its sensitivity to score gaps and mass, and derive an exact no-pruning condition. For 0\rho1 , the removed fraction is at most \rho and at least one example remains. The unsigned score permits both confidently correct and confidently wrong examples to be removed. Seven evaluation settings include neural training, language modeling, and boosting. Neural execution studies reduce backward-active rows by 10-35%; on CIFAR-10, 29% fewer backward rows accompany accuracy means of 89.77% versus 89.82% for full CE. At a matched backward budget and number of optimizer updates, LACS is 1.0 pp above small-batch training and 1.2 pp above random selection over five seeds. Across 14 tabular datasets, quadratic boosting reduces mean test error from 23.36% to 22.96%. Its full training replays use about 30% fewer fitting rows; these counters span a different horizon from validation-selected accuracy. In one RBF-SVM case study, elapsed time falls by 37% with slightly lower error. These findings separate selective-work reduction from realized acceleration: the current exact-current-score neural implementation uses an additional scoring forward, so its backward sparsity does not yet translate into wall-clock acceleration. This overhead is an execution choice rather than a requirement of the selection rule.
[LG-83] On the Generalization Limitation of On-Policy Distillation for Post-RL Consolidation
链接: https://arxiv.org/abs/2611.05609
作者: Yichen Wang,Yufa Zhou,Siqi Zhu,Haozheng Luo,Jingzhi Bao,Zhichao Liu,Zijun Wei
类目: Machine Learning (cs.LG)
*备注: 27 pages, 12 figures. Project page: this https URL
Abstract:Multi-teacher on-policy distillation (MOPD) extends on-policy distillation (OPD) to consolidate capabilities from multiple teachers. Although prior work reports near-perfect in-distribution (IND) distillation performance, we find that this can mask incomplete capability consolidation beyond the supervised distribution. A preliminary study on synthetic reasoning shows that OPD does not fully transfer teachers’ generalization capabilities. We then extend this investigation to real-world tasks, evaluating extrapolative and compositional generalization across mathematical reasoning, coding, agent, and long-context capabilities using Qwen3 models. Across extrapolative settings, OPD and MOPD reach 96.04-102.73% of their teachers’ IND scores. Yet their out-of-distribution (OOD) scores remain 2.45-16.25% below the base model’s. Across three compositional settings, MOPD improves IND performance over the base model, but its performance on composed tasks as OOD remains 18.19-24.34% below that of the best teachers. To understand these gaps, we examine teacher capability and data coverage as two potential bottlenecks. Jointly alleviating these bottlenecks eliminates OOD forgetting in our experiments and suggests the potential for generalization beyond the teacher. (M)OPD also reduces OOD forgetting on average relative to its reinforcement learning (RL) baselines. Analysis of the associated update geometry reveals smaller and more spectrally concentrated updates on average than those of the RL baselines, along with greater shared update structure under MOPD. Motivated by these findings, we propose geometry-based patching and base-preserving replay to improve OOD performance while largely preserving IND performance. Our findings identify a blind spot in capability consolidation via (M)OPD, highlighting the need for algorithms that explicitly support OOD retention and generalization.
[LG-84] ransport along Local Annealing Paths: Wasserstein Geometry and Multimodal Sampling
链接: https://arxiv.org/abs/2611.05605
作者: Han Chen,Sifan Liu
类目: Machine Learning (cs.LG); Computation (stat.CO); Machine Learning (stat.ML)
*备注: 36 pages, 4 figures
Abstract:Variational approaches to learning transport maps for multimodal sampling can suffer from mode collapse. One way to mitigate this problem is to learn transport along a prescribed path of intermediate distributions defined by geometric annealing. We show that annealing can take a substantial detour in Wasserstein space: when mass moves between modes, the path can be much longer than the distance between its endpoints. We further link path geometry to transport learning, showing that long annealing paths can slow population optimization, while sufficiently regular paths permit rapid convergence. Motivated by these findings, we propose componentwise annealed transport (CAT) to reduce the need for mass transfer between modes. Starting from a mode-informed mixture reference, CAT constructs an exact mixture decomposition of the target and learns separate transport maps along componentwise annealing paths. We estimate the component weights separately and show that these estimates remain consistent even when the learned flows are inexact. We also develop adaptive annealing schedules guided by transport effort. Experiments on multimodal synthetic distributions and Bayesian inference problems demonstrate improvements in transport learning and sampling efficiency.
[LG-85] RIDGE: State-Conditioned Reward Blending for Behavioral Coverag e in Deep RL Game Agents
链接: https://arxiv.org/abs/2611.05601
作者: AL Muqshith Mohammed Shifan,Kevin Christopher Chua,Ali Neshati,Loutfouz Zaman,Cristiano Politowski
类目: Machine Learning (cs.LG); Software Engineering (cs.SE)
*备注: Conference of Games, 4 pages, 3 figures, reinforcement learning, automated playtesting, procedural personas,reward shaping, PPO, Crafter
Abstract:A deep RL agent trained on a fixed persona reward plays one way: an Explorer keeps exploring while starving, a Survivor hoards on an empty map. Since each persona fixes its own reward, automated playtesting trains N agents for N styles. Changing reward composition mid-episode runs into the switching-stability dilemma: PPO degrades under non-stationary rewards. We present RIDGE, a single-agent architecture that sidesteps the dilemma with two design choices: a state-conditioned blender that computes softmax-normalized sigmoid weights from slow-moving state variables, and a multi-head critic that trains one value head per persona on its own stationary signal. On Crafter with four personas, one \ridge agent acquires both branches of the wood-tier crafting/combat split that no fixed-persona baseline solves jointly, and a sharpness ablation confirms state-conditioning drives the gain. We position these results as demonstrations of the blending mechanism’s dynamics across n=3 seeds per condition; larger multi-seed sweeps remain future work.
[LG-86] C-GRPO: Conformal Group Relative Policy Optimization NEURIPS2026
链接: https://arxiv.org/abs/2611.05591
作者: Arya Fayyazi,Seyedarmin Azizi,Mehdi Kamal,Massoud Pedram
类目: Machine Learning (cs.LG)
*备注: Proceeding of NeurIPS 2026
Abstract:Reasoning-oriented language model training increasingly relies on multiple on-policy rollouts, but standard Group Relative Policy Optimization (GRPO) uses a fixed group size, wasting compute on easy prompts and under-exploring hard ones. Existing adaptive-budget methods alleviate this inefficiency, yet their stopping rules are heuristic and do not provide a finite-sample calibration guarantee. We propose CGRPO, which replaces the fixed rollout budget of standard GPRO with calibrated conformal thresholds over task-specific nonconformity scores, combining Adaptive Prediction Set (APS)-style calibration for closed-form reasoning, a first-success score for execution-based code, automatic miscoverage selection, and periodic re-calibration. We prove finite-sample coverage for the resulting stopping rule and show that, under monotone policy improvement, recalibration induces a non-increasing expected training budget. Across six benchmarks spanning mathematics, graduate-level science, and code generation, CGRPO reduces mean training rollouts by 49-70% relative to fixed-budget GRPO while retaining strong fixed-budget accuracy, achieves the strongest pass@1 among evaluated static and dynamic baselines on different datasets, and transfers across model families with up to 82% rollout savings.
[LG-87] Exploring the Privacy–Size Tradeoff for Differentially Private Self-Attentive Variational Autoencoders
链接: https://arxiv.org/abs/2611.05582
作者: Rohit Sadavarte,Dhruvi Kothari,Sparsh Gupta,Eric Van Oeveren
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE); Cryptography and Security (cs.CR)
*备注: ICAIF 2026, Financial AI Security, Privacy and Safety workshop. Poster session
Abstract:Data tends to be most useful when it is shared, and the data most worth sharing is often the data that privacy obligations restrict. Synthetic data from a generator trained under differential privacy (DP) is one way to reconcile the two, but its utility depends on the privacy budget \varepsilon and on how much data the generator sees - a parameter that most data sharing applications of synthetic data control. We present SAVANT, a self-attentive variational autoencoder for tabular data trained with DP-SGD that, to our knowledge, is the first to combine attention, a VAE, and DP in one tabular model. We evaluate it against two marginal mechanisms, PrivBayes and MST, jointly over the privacy budget and the dataset size on two public datasets, measuring the utility a downstream classifier loses by training on synthetic rather than real data. We find that, under the same DP guarantee, more data generally means more utility. For the marginal mechanisms it absorbs most of what a strict budget costs, and a noise-scaling argument suggests DP-SGD exchanges data fo budget at a worse rate. Two independent membership-inference attacks, judged against a measured null, found no aggregate membership signal.
[LG-88] Learning vs. Optimizing Players in Budgeted Games
链接: https://arxiv.org/abs/2611.05580
作者: Giannis Fikioris,Balasubramanian Sivan,Éva Tardos
类目: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG)
*备注:
Abstract:We study repeated general games over T rounds between a utility-maximizing optimizer and a learning agent, where budget constraints couple the players’ decisions across rounds. Without budgets, low swap regret simultaneously guarantees the learner sublinear regret and caps the optimizer at its Stackelberg value. Budgets break this picture: adversarial no-regret becomes impossible, so the natural goals are non-manipulability—no adaptive optimizer earns more than \mathttBSE \cdot T + o(T) , where \mathttBSE is its Budgeted Stackelberg Equilibrium value—and low stochastic regret when the optimizer plays a fixed mixed strategy. We give an efficient, budget-feasible learning algorithm that achieves both goals in every budgeted game: any adaptive optimizer, budgeted or not, earns at most \mathttBSE \cdot T + O(\sqrtT) in expectation, and if the optimizer plays a fixed mixed strategy on each of a constant number of intervals, the learner has O(\sqrtT) regret on each such interval. Our learner best responds to forecasts of the optimizer’s play that are calibrated in two novel senses, since under budgets the calibration that prevents manipulation no longer yields low stochastic regret. Response calibration, which conditions on the learner’s own mixed response by embedding its continuous mixing weight as a step function in L^2([-1,1]) , prevents manipulation without discretizing responses; dual-weighted Lagrangian calibration ensures low regret. Although low calibration error requires randomized forecasts—distributions over a continuous space that need not have a succinct representation or be directly computable—we construct such forecasts in polynomial time by simulating online gradient ascent. Subjects: Computer Science and Game Theory (cs.GT); Machine Learning (cs.LG) Cite as: arXiv:2611.05580 [cs.GT] (or arXiv:2611.05580v1 [cs.GT] for this version) https://doi.org/10.48550/arXiv.2611.05580 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-89] Support Vector Machines Directly Minimise KL Divergence
链接: https://arxiv.org/abs/2611.05557
作者: Ivan Silajev
类目: Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 11 pages, 1 table
Abstract:Support vector machines (SVMs) are fundamentally grounded in geometric margin maximisation and currently lack a mainstream probabilistic interpretation. This keeps them theoretically isolated from probabilistic machine learning models that are naturally comparable and combinable. In this paper, we bridge this gap by proving that the dual objectives of both support vector classification and support vector regression are exact, constrained Kullback-Leibler (KL) divergence minimisation tasks between two Gaussian measures. We further show that this same divergence formulation recovers the exact objective function for kernel ridge regression when domain constraints are removed and L^2 regularisation is applied. By demonstrating that SVM duals directly minimise relative entropy, our results provide a unified information-theoretic model for standard kernel machines and lay the foundation for unifying broader kernel-based algorithms under probability theory.
[LG-90] FedSharp: Byzantine Robust Decentralized Federated Learning Under Noisy Channel
链接: https://arxiv.org/abs/2611.05546
作者: Ojaswi Sinha,Amin Farajzadeh,Animesh Yadav,Melike Erol-Kantarci
类目: Machine Learning (cs.LG)
*备注: 6 pages, 4 figures, submitted for possible publication
Abstract:Decentralized federated learning (DFL) eliminates the central aggregation server but remains vulnerable to data heterogeneity, Byzantine attacks, and channel noise. We show that noise accumulates as models propagate across nodes, becoming increasingly significant as the learning rate decays. Moreover, accumulated noise enables Byzantine attackers to exploit loss-based model selection by mimicking honest model averages. To address these challenges, we propose screened, hop-aware, ranked, proximal averaging, a robust DFL framework (FedSharp). In this approach, each node maintains models from its 2b predecessors, ensuring at least b honest candidates when at most b nodes are Byzantine. FedSharp screens suspicious models using coordinate-wise median deviations, averages the b lowest-loss candidates to reduce fresh channel-noise variance by a factor of b, and applies noise-adaptive gradient-norm regularization and proximal training to mitigate accumulated noise and non-IID drift. Experiments on a 100-node ring with 33 Byzantine nodes show that FedSharp achieves 93.5-94.9% accuracy on MNIST, compared with 29.4-38.1% for a state-of-the-art single-model-selection benchmark, and 32.8-48.8% on CIFAR-10, compared with 10.0-30.5%, across Gaussian and hidden attacks under IID and non-IID data. On MNIST, FedSharp also achieves accuracy within approximately one percentage point of the benchmark operating without channel noise.
[LG-91] Recasting Visual Language Models as EEG Foundation Models for Clinical Neurodiagnostics
链接: https://arxiv.org/abs/2611.05538
作者: Xin Zhou,Elena Sergeeva,Nan Bi,Wenyu Song,Lijun Yin,Hao Deng
类目: Machine Learning (cs.LG)
*备注:
Abstract:Historically automated electroencephalography (EEG) systems was done using task-specific supervised models trained on purposely annotated data. A more recent modeling paradigm relies on creating a unified representation, or a ‘foundation model’, for multiple modality-specific tasks through large-scale domain-specific pretraining on the unannotated EEG signal. The quality of this unified representation directly depends on the availability of the modality-specific datasets. At the same time, generalist Vision Language Models (VLMs) exhibit broad few-shot learning capacities as well as good fine-tuning adaptability to different visual and language tasks. Can VLMs learn how to read and interpret EEG visually, emulating to a degree the way human clinicians inspect the EEG patterns? In this work we make a case that they can: we show that generalist VLMs without large-scale EEG pretraining perform competitively across a collection of clinical neuro-diagnostic tasks, can adapt to different tasks in a few-shot learning setting better than specialist foundation models, and exhibit better performance with model scaling when modality-specific models fail to exhibit.
[LG-92] he Confidence Score: a New Measure on Kernels
链接: https://arxiv.org/abs/2611.05528
作者: Elizabeth Hofer,Martin v. Mohrenschildt
类目: Machine Learning (cs.LG); Signal Processing (eess.SP)
*备注: 27 pages
Abstract:Across a wide range of applications, kernel measures are used to tune hyperparameters, select among candidate kernels, and assess the suitability of kernels for specific learning tasks. For classification, kernel measures are used to predict which kernels may produce accurate kernel classifiers. Many existing kernel measures, such as alignment to the ideal kernel, are designed to reward kernels which captures all samples from the same class as similar. While effective in many settings, this requirement is unnecessarily restrictive for classification, where classes may exhibit substantial internal structure. In this work, the “Confidence Score”, a new kernel measure, is proposed. Unlike alignment-to-the-ideal kernel measures, the Confidence Score measures the alignment between an encoding of the ground truth and that ground truth projected by the kernel, allowing kernels that preserve intra-class structure to achieve high scores. Through experiments, it is demonstrated that the Confidence Score exhibits a higher correlation with classification accuracy across a variety of kernels on both real-world and synthetic datasets, compared to select other kernel measures.
[LG-93] Ctrl-F: Force as an Interface for Humanoid Loco-Manipulation CEC
链接: https://arxiv.org/abs/2611.05507
作者: Mengjie Zhao,Yuxin Chen,Yiyang Shao,Chuye Hong,Jiakang Chen,Koushil Sreenath,Masayoshi Tomizuka
类目: Robotics (cs.RO); Machine Learning (cs.LG)
*备注: Project page: this https URL
Abstract:Humanoid motion tracking has scaled to general policies that reproduce diverse human motion, but the resulting behavior is in free-space: the tracker knows where the body should be, not how it should interact with the environment. Other work makes the object part of the objective, yet abstracts that interaction to contact or geometry, which leaves the demanding cases under-explored. We argue that force is a natural command channel between motion imitation and object interaction, one that specifies not just where to make contact but how hard. We present Ctrl-F, a hierarchical motion-force framework that augments motion tracking with an explicit force command, a 3-D force at each designated contact body such as a hand or a foot. The hierarchical framework comprises: (1) A unified motion-force controller tracks the reference motion while applying the commanded force to the world. (2) A high-level force planner that issues a force command online based on object state and proprioception observations. In simulation and on hardware, Ctrl-F completes loco-manipulation that demands large, sustained interaction force and adjusts its command online as the load changes, both out of reach of a force-blind tracker. On hardware, Ctrl-F demonstrates hand-based dragging, pushing, and lifting, as well as foot-based kicking, successfully manipulating boxes weighing up to 20kg in pushing and kicking tasks where the force-blind baseline fails, while adapting online to additional loads introduced mid-task.
[LG-94] Matched but Not Better: An Equivalence-Tested Benchmark of Quantum Machine Learning on Six Clinical Datasets and the QML-Health-READY Reporting Framework
链接: https://arxiv.org/abs/2611.05442
作者: Mohammed Siraj B,Gousia Habib,Ruban S
类目: Machine Learning (cs.LG)
*备注:
Abstract:Background: Quantum machine learning (QML) is promoted for clinical prediction, but comparisons with classical models are rarely paired or tested for equivalence. Methods: We compared a quantum-kernel support vector machine (QSVM; 4 qubits, tuned bandwidth) and a variational quantum classifier (VQC) with five tuned classical models across six public clinical datasets. All models used identical four-component inputs and 25 cross-validation splits. Paired AUROC differences were assessed using a corrected resampled t-test, Holm-adjusted equivalence tests (margins: +/-0.02 and +/-0.05), and random-effects pooling. Ablation, noise, finite-shot, and implementation-fault analyses were conducted. We propose QML-Health-READY, an eight-domain reporting framework. Results: Neither quantum model significantly outperformed any classical comparator. Pooled across datasets, QSVM differed from the kernel-matched radial basis function (RBF) SVM by -0.006 AUROC (95% CI: -0.017 to +0.004), equivalent to +/-0.02. VQC differed from the best classical model by -0.007 (95% CI: -0.019 to +0.004). QSVM was equivalent to RBF SVM within +/-0.05 in four of six datasets but was significantly worse than the best classical model on NHANES 2017-18 (-0.076). A classically trivial product-state kernel performed as well as the entangling kernel. With depolarizing noise (p = 0.01) and 1,024 shots per kernel entry, QSVM AUROC decreased by 0.054 (95% CI: 0.036 to 0.071). Untuned bandwidth produced near-chance AUROC (0.48-0.59). Conclusions: Correctly configured quantum classifiers approached but did not exceed tuned classical models. Finite-shot estimation, rather than gate noise, was their main practical cost, while configuration errors could reverse conclusions. QML-Health-READY defines reporting requirements for credible claims.
[LG-95] Learning Implicit Solves for Dynamic Simulations
链接: https://arxiv.org/abs/2611.05439
作者: Chawit Kaewnuratchadasorn,Jiaji Wang,Xin Nie
类目: Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)
*备注:
Abstract:High-fidelity simulation of dynamical systems is central to scientific discovery and engineering design, yet remains computationally demanding. Here we introduce the Implicit Neural Operator Solver (INOS), which learns the implicit solve and encodes a time integrator. INOS pretrains a neural operator from static equilibrium solutions and reuses it within Newmark- \beta integration. At each time increment, the integrator forms an effective equilibrium problem, the pretrained neural operator infers the next-step solution field, and the Newmark relations advance the dynamic state. Numerical examples in linear and nonlinear elastodynamics demonstrate accurate full-field predictions without dynamic trajectory supervision. Comparative results show that INOS, evaluated zero-shot, achieves higher predictive accuracy than neural solver baselines trained on dynamic simulations. In addition, INOS delivers over 4,600 \times speedup in nonlinear simulations over sequential classical finite-element solvers. INOS natively accommodates changes in dynamic conditions and enables reuse of the pretrained operator across spatial and temporal discretizations. Further evaluation demonstrates a stable million-step rollout. These findings establish INOS as a reusable neural implicit solver for accurate, efficient, and robust dynamic modeling.
[LG-96] FlowReconBench: Accuracy calibration and cost of learned and classical flow reconstruction from sparse sensors
链接: https://arxiv.org/abs/2611.06912
作者: Nicolas Tricard,Linzheng Wang,Jason Chen,Sili Deng
类目: Fluid Dynamics (physics.flu-dyn); Machine Learning (cs.LG)
*备注: 18 pages, 7 figures, 8 tables. Supplementary material (16 pages, 2 figures, 5 tables) is included as an ancillary file
Abstract:Reconstructing full flow fields from sparse sensors requires prior assumptions about unmeasured regions and physical variables. Yet it remains unclear when learned models outperform classical estimators and which aspects of their predictions can be trusted. FlowReconBench evaluates ten learned and five classical methods on three regimes: a laminar cylinder wake across Reynolds numbers, two-dimensional Kolmogorov turbulence and a large-scale three-dimensional (3D) isotropic turbulence. Methods share leakage-controlled splits and sensor draws and are scored on per-channel error, ensemble spread, calibration, spectra and cost. The leading method changes for each regime. Classical methods perform strongly in low-order and dense-sensing regimes. On the cylinder wake, linear stochastic estimation and gappy proper orthogonal decomposition lead, including for unobserved pressure. In 3D, inverse-distance weighting matches the best learned methods on observed channels. Data-driven methods become superior at decreasing sensor density or poor sensor observability. On Kolmogorov flow, continuous ranked probability score and error yield different method rankings: generative models offer broader but underdispersed ensembles, while deterministic models prioritize pointwise error. Generative models balance this tradeoff through their sampler budget, with a larger budget improving ensemble quality at higher computational cost. In 3D, scaling tests demonstrate that queryable methods are flat in memory, while full-grid construction methods run out of memory. Observing only 2 of 4 fields leaves the problem kinematically underconstrained, yet a latent model can reproduce structural similarities to the ground truth.
[LG-97] hermally Conditioned Time-Series Forecasting of Grain Growth in Polycrystalline Materials
链接: https://arxiv.org/abs/2611.06843
作者: Eliane Younes,Elie Hachem,Marc Bernacki
类目: Materials Science (cond-mat.mtrl-sci); Machine Learning (cs.LG)
*备注:
Abstract:Grain growth during heat treatment strongly affects the microstructure of polycrystalline materials and, consequently, their mechanical properties. Although numerical methods such as front-tracking or front-capturing simulations can accurately describe grain growth, their high computational cost limits their use for long or non-isothermal thermal cycles. In previous work, grain growth was successfully predicted under isothermal conditions by treating grain size distributions as time series and forecasting their evolution using recurrent neural networks, allowing complete grain growth trajectories to be generated in seconds rather than through costly PDE-based simulations. In the present work, this forecasting framework is reformulated to handle non-isothermal conditions by explicitly incorporating thermal information into the model architecture. Grain size distributions are encoded using a convolutional neural network, thermal descriptors are embedded through a dedicated multilayer perceptron, and temporal evolution is predicted using a recurrent network. Instantaneous temperature and thermal rate are provided at each time step, allowing the model to adapt its predictions to changing thermal conditions. The framework is trained and evaluated on datasets generated from validated front-tracking simulations across a wide range of non-isothermal thermal paths. The results show that the thermally conditioned model accurately reproduces grain growth dynamics under complex thermal histories, achieving, in k-fold cross-validation over all simulations, mean relative errors of 2.99% (one-hour cycles) and 4.51% (four-hour cycles) with coefficients of determination exceeding 0.99, and mean relative errors between 1.91% and 14.66% on ten simulations following four unseen thermal histories of increasing complexity.
[LG-98] Alignment Myopia in Generative AI
链接: https://arxiv.org/abs/2611.06830
作者: Remi Denton,Aida Davani,Cynthia L Bennett,Michael Madaio,Roma Patel,Himabindu Lakkaraju
类目: Machine Learning (stat.ML); Computers and Society (cs.CY); Machine Learning (cs.LG)
*备注:
Abstract:Current generative AI alignment relies on a myopic paradigm that prioritizes immediate conversational satisfaction. By relying on short-horizon feedback, this approach ignores the profound, longitudinal externalities of persistent human-AI interaction, which can lead to adverse outcomes such as cognitive atrophy, social erosion, and diminished human agency. To address this, we propose a paradigm shift toward long-horizon alignment that accounts for how AI reshapes human experience over time. To this end, we map the tensions between near-term utility and long-term costs of AI use across four facets of human well-being to reveal outcomes that current reward models fail to capture. Ultimately, we advocate for moving beyond simple preference optimization toward alignment strategies that actively foster long-term human well-being by navigating these complex temporal scales.
[LG-99] GFlowQEC: Generative Flow Networks that Learn to Sample Quantum LDPC Codes
链接: https://arxiv.org/abs/2611.06819
作者: Nuno Batista,Gabriel Falcao,Penousal Machado
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG)
*备注: 16 pages, 1 figure, 2 tables and 42 references
Abstract:Choosing a quantum error-correcting code for a real device means scoring candidates with circuit-level simulations so slow that only a few hundred can be run. So every evaluation counts, and a search is judged by how many good codes it finds per paid evaluation. Quantum LDPC codes built from a few group elements protect many logical qubits, but good ones are rare in a huge space. We introduce GFlowQEC, to our knowledge the first generative flow network (GFlowNet) for quantum code design. It builds a code one group element at a time and learns to propose codes in proportion to a power of their score. We train it once on scores from earlier searches, freeze it, and let it propose lifted-product codes with 180 physical qubits, scored by their distances. Ranking ten of its draws per paid call, it finds more than twice as many new top-scoring codes (inequivalent to any it was trained on) as tuned parallel tempering, even when tempering starts from GFlowQEC’s training codes. It finds more than three and a half times as many as a supervised screen ranking ten random candidates per call. When both rank a hundred, the screen beats GFlowQEC, and does best ranking GFlowQEC’s draws. Soft Q-learning, which shares the GFlowNet’s optimal policy, does as well at one noise bias and better at the other, so the gain does not depend on the GFlowNet’s training loss. Trained proposers can make a budget of expensive evaluations go further when good codes are rare.
[LG-100] Evolutionary Design of Interpretable Spatial Descriptors for Structure-Property Relationship Discovery
链接: https://arxiv.org/abs/2611.06811
作者: Aditya Raghavan,Sergei V Kalinin
类目: Materials Science (cond-mat.mtrl-sci); Machine Learning (cs.LG)
*备注:
Abstract:Structure property prediction is a classical task in machine learning, ranging from molecular activity and materials properties to local functional responses measured in microscopy. For image-based regression, two choices define the information available to the predictor: the size of the image neighborhood and the spatial resolution retained within it. A small neighborhood can omit relevant surrounding structure, whereas a large, finely sampled neighborhood can introduce redundant or irrelevant information. Here, we develop a framework for constructing spatial descriptors for prediction of local spectra and hysteresis loops. We benchmark regression models, use SHAP maps to examine their spatial attribution, and explore the effect of the patch size and compression to establish the relationship between predictive accuracy and descriptor dimensionality. We then use a genetic algorithm to optimize spatial partitions and the assignment of mean, minimum, and maximum pooling operations. Each retained region contributes an explicit measurement of the image, allowing the search to balance prediction error against the number of descriptor values. Cartesian and polar constructions provide complementary ways to organize this information. The resulting descriptors make it possible to examine where spatial detail is needed, where regional summaries are sufficient, and how these choices depend on the predictor and the measured response. The same approach can be used for molecules and solids once the genetic algorithm rules for these are established.
[LG-101] stBED: An Open-Source Toolkit for Bayesian Experimental Design
链接: https://arxiv.org/abs/2611.06633
作者: Gavin Kerrigan,Marcel Hedman,Sahel Iqbal,Tom Rainforth
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:Bayesian experimental design (BED) is a principled framework for intelligent data acquisition with applications across a wide range of scientific and industrial domains. However, the development and adoption of BED remains hindered by a lack of standardised software and benchmarks. Existing methods are typically implemented in bespoke codebases and evaluated on a small subset of tasks, making methods difficult to apply, compare, or extend. To address this, we present TestBED, a general-purpose, open-source toolkit for Bayesian experimental design. TestBED decomposes the BED process into modular, reusable components spanning tasks, objectives, training mechanisms, policy architectures, and evaluation procedures, together with scalable implementations of core primitives such as expected information gain estimators. We provide a wide range of these components, which can be freely composed, allowing existing methods to be reproduced within a common framework, while also enabling previously unexplored configurations and new methodologies to be easily implemented and evaluated. Using TestBED, we further conduct a large-scale benchmark of contemporary BED methods, including controlled studies of under-appreciated but surprisingly critical methodological choices, such as the importance of time embeddings in design policy architectures. By providing a modular, extensible software foundation and a reproducible benchmarking suite, we hope that TestBED will accelerate methodological research in BED and lower the barrier to its real-world application.
[LG-102] Finite-width shift of the edge of chaos in deep neural networks
链接: https://arxiv.org/abs/2611.06582
作者: Arn Caesar
类目: Disordered Systems and Neural Networks (cond-mat.dis-nn); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注: 9 pages, 5 figures, 1 table; Supplemental Material (28 pages, 9 figures, 12 tables) included. Code: this https URL
Abstract:Deep neural networks train best when they are initialised at the edge of chaos, the critical line in the plane of weight and bias variances on which the distance between two inputs neither grows nor shrinks with depth. The line is known for infinitely wide networks; at finite width it moves, and how it moves has been disputed. We treat the logarithm of the distance between two inputs through a network of width N as a random walk and compute its drift and variance to order 1/N . The edge of chaos moves to a larger weight variance, by \Delta\sigma_w^2=-a_1/(N,\partial_\sigma_w^2\ln\chi_1) , about 8/N for \tanh networks at a bias variance of 0.05 , confirmed by simulations of the exact finite-width recursion. The shift describes a typical network; the average over many networks is dominated by rare networks in which the distance grows, and its critical point moves the other way, which explains why calculations and simulations have disagreed. Backpropagation follows the same random walk, so the shift applies to gradients unchanged. Along the edge of chaos, the information about the similarity of two inputs is constant until the compression over the depth exceeds about one nat, which fixes where to sit on the line. The results give an initialisation rule, Crinit, that places a network on its finite-width edge of chaos without a search. On MNIST and Fashion-MNIST at width 64 it is more accurate than the standard rules at their usual settings, by a small margin in shallow networks and a large one in deep networks, where most standard rules fail; it trains in fewer steps, and it matches or beats the standard rules with a tuned gain.
[LG-103] Adapting Sleep Foundation Models for Dense Sleep Spindle Localization ICASSP2027
链接: https://arxiv.org/abs/2611.06546
作者: Alexis Bollengier,Abel Díaz Berenguer,Olivier Pallanca,André Anjos,Hichem Sahli
类目: ignal Processing (eess.SP); Machine Learning (cs.LG)
*备注: 5 pages, 3 figures, 1 table. Submitted to IEEE ICASSP 2027 (under review)
Abstract:Adapting sleep foundation models to dense electroencephalography (EEG) event localization is challenging because multi-second representations may not preserve sufficient temporal detail for precise boundary recovery. We address sleep spindle localization by adapting a frozen SleepFM encoder with a local time-frequency pathway that preserves spindle-relevant spectral content and intra-token position. The local representation drives localization, while residual cross-attention incorporates SleepFM temporal context. Dense predictions are obtained through localized token-to-sample reconstruction and temporal refinement. Tested on the Massive Online Data Annotation (MODA), the proposed model achieves an event-level F_1 of 0.8080 \pm 0.0040 at IoU 0.2 across six folds. Ablations show that localization primarily depends on local spectral and sub-token temporal information, while the pretrained representation provides complementary context. These results indicate that dense adaptation benefits from preserving information below the pretrained token scale. Code will be released upon acceptance.
[LG-104] Risk-Sensitive Non-Parametric and Kernel Bandits
链接: https://arxiv.org/abs/2611.06489
作者: Devdan Dey,Sujoy Bhore,Avishek Ghosh
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 46 pages, 6 figures
Abstract:We study non-parametric bandits in which arms are evaluated using Cumulative Prospect Theory (CPT) rather than expected reward. We consider continuous action spaces with arm-dependent noise and propose Non-Parametric CPT Bandits ( \mathttNPCB ), an optimistic algorithm that decouples mean estimation from CPT-based decision-making. We prove CPT regret of order \widetilde\mathcal O\bigl(\sqrtT,\dim_E(\mathcal F,T^-2),\log N(\mathcal F,T^-2,|\cdot|_\infty)\bigr) , where \dim_E is the Eluder dimension of the function class \mathcal F , T is the time horizon, and N is its covering number in the sup norm. This matches the order-wise regret of classical non-parametric bandits, showing that CPT risk sensitivity incurs no additional order-wise regret cost. We extend \mathttNPCB to unknown noise CDFs, retaining the known-CDF mean-learning term and adding a CDF-learning cost governed by a distributional Eluder dimension. We also study Kernel-CPT bandits and obtain information-gain-based regret bounds for a broad range of kernels, matching the corresponding risk-neutral kernel rates up to logarithmic factors.
[LG-105] How many sorted projections determine n points in k dimensions?
链接: https://arxiv.org/abs/2611.06470
作者: Christopher Gorham Lester,Raymond Hau Sørensen
类目: Combinatorics (math.CO); Discrete Mathematics (cs.DM); Information Theory (cs.IT); Machine Learning (cs.LG)
*备注: 28 pages, 5 figures
Abstract:Let d_1,\dots,d_m be directions in \mathbbR^k , where k\ge 2 . A multiset X of n points of \mathbbR^k projects to the m multisets of numbers \x\cdot d_j : x\in X\ , or equivalently to the m sorted lists of these numbers. How many directions are needed for these projections to determine every multiset of n points? Two simple bounds are known in every dimension: more than (k-1)\log_2 n directions are always needed, and (k-1)n+1 directions in general position always suffice. In the plane much more is known: Matousek, Privetivy and Skovron showed that almost every choice of O(\log n \log\log n) directions suffices. Whether anything similar holds in higher dimensions was left open by Dym, Wellershoff, Tsoukanis, Levy and Balan. We observe that two lemmas of Hair and Sahai, proved for list recovery of random linear codes, answer that question, and by tightening their argument we show that in every dimension almost every choice of (4k-2)\ln(2n)\ln(k\log_2 2n)+k+1 directions suffices. For k\le 10 this beats the linear bound once n is larger than about 100: for example, 144 directions suffice for multisets of 1000 points in the plane, and an exact computation lowers this to 133. Letting some of the random blocks in their argument fail removes a factor \log k : almost every choice of O(k\log n\log\log n) directions suffices, uniformly in k and n . We also show that 2k-1 directions never determine three points, so that the logarithmic lower bound is not tight; that every m directions in the plane fail on some pair of multisets of size O(1.78^m) , improving the bound O(1.80^m) of Matousek, Privetivy and Skovron; and that for each k , m and n , either almost every choice of m directions works or almost none does.
[LG-106] PARGAR: Pilot-Aware Graph Attention for Sparse-Pilot Neural OFDM Receivers
链接: https://arxiv.org/abs/2611.06432
作者: Sajedeh Norouzi,Mostafa Rahmani Ghourtani,Eric Samikwa,Torsten Braun
类目: ignal Processing (eess.SP); Machine Learning (cs.LG)
*备注:
Abstract:Reliable channel estimation and symbol detection in 5G New Radio (NR), as well as emerging 6G radio systems, rely on demodulation reference signals (DMRS). The density and placement of these pilots on the orthogonal frequency-division multiplexing (OFDM) time–frequency resource grid affect both receiver reliability and spectral efficiency. Sparse DMRS layouts increase the time–frequency distance between pilot observations and data resource elements (REs), which can make pilot information harder to exploit with purely local processing. This paper proposes \textbfPARGAR (Pilot-Aware Graph Attention Receiver), a graph-attention neural receiver that represents the OFDM resource grid as a graph for bit detection. The graph combines local time–frequency edges with pilot-anchor links from data REs to selected pilot REs, and we further study a channel-scaled pilot-anchor rule based on approximate coherence time and coherence bandwidth. Using Sionna-generated OFDM simulations, we evaluate \textscPARGAR against DeepRx, a convolutional neural network (CNN)-based receiver, across dense, two-symbol sparse, and single-symbol sparse DMRS layouts under channel-mismatch scenarios. Under the considered SISO uncoded setup, channel-scaled \textscPARGAR explicitly injects pilot-aware inductive bias and yields lower BER in selected sparse-pilot scenarios while remaining competitive under dense pilots.
[LG-107] Impact of Negative Weights on Performance in Reservoir Computing
链接: https://arxiv.org/abs/2611.06051
作者: Yeonsu Jeong,Seonho Nam,Dahan Choi,Mi Jin Lee,Seung-Woo Son
类目: Physics and Society (physics.soc-ph); Machine Learning (cs.LG)
*备注:
Abstract:Reservoir computing provides an efficient framework for time-series prediction because only the readout layer is trained while the recurrent reservoir remains fixed. In this study, we examine how the negative sign ratio of a signed Erdos-Renyi reservoir affects the internal dynamics and forecasting performance of an echo state network on the Mackey-Glass benchmark. To characterize the reservoir responses, we use two Shannon-entropy-based observables, namely node-state entropy and node-frequency entropy, as empirical proxies for the diversity of reservoir states and temporal activation patterns. We find that intermediate negative sign ratios are associated with richer internal dynamics and substantially lower prediction error than the sign-homogeneous extremes for the task considered here. These results suggest that link-sign composition can be a useful structural parameter in reservoir design for chaotic forecasting. The present findings should be interpreted within the scope of the current benchmark and reservoir setting, and broader validation will be required to establish their generality across other tasks, topologies, and controlled spectral conditions.
[LG-108] Generative Modeling for Cascading Extremes via Autoregressive Transformers
链接: https://arxiv.org/abs/2611.05728
作者: Miguel de Carvalho,Myung Won Lee,Jordan Richards,Sotirios Sabanis
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: 28 pages, 10 figures
Abstract:Extreme events often occur sequentially, from earthquake-tsunami cascades to successive extreme losses in financial markets. We introduce an extremal token representation that encodes multivariate observations as a categorical sequence. A quiet token records the absence of threshold exceedances; otherwise, the token identifies the exceeding component with the largest marginal probability-scale score. The representation is invariant to componentwise strictly increasing transformations. It preserves the identity and temporal ordering of dominant exceedances but discards their magnitudes and the identities of nondominant simultaneous exceedances. A decoder-only transformer estimates the history-dependent next-token kernel on this representation, and recursive sampling generates future token sequences from the induced joint predictive law. We amortize our generative model by pretraining across simulated process families with varying parameters and dependence persistence. Next-token predictions are evaluated against exact conditional probabilities for held-out categorical processes and particle filter references for heavy-tailed vector autoregressive processes. Applied to ten-second returns for five semiconductor stocks, the pretrained model supports zero-shot transfer, while a few epochs of adaptation improve agreement between observed and generated token frequencies and transition patterns. These results support task-specific tokenization and simulation-based pretraining for modeling which dimensions dominate multivariate extremes.
[LG-109] Distinguishing Initial-Condition and Model Error Using Machine-Learning Weather Models
链接: https://arxiv.org/abs/2611.05592
作者: P. Trent Vonich,Gregory J. Hakim
类目: Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG)
*备注:
Abstract:Gated and ungated optimizations evaluated in GraphCast and Pangu-Weather. The first panel (upper-left) shows the mean over the 36 dates; remaining panels show first-of-month 12 UTC optimization descents from January-November 2017. The y-axis metric is the summed day 1-10 Z500 RMSE verified against ERA5, expressed as percent change from each model’s date-matched control. The gated arm optimizes the initial conditions jointly with a LoRA weight correction; its GraphCast curve uses the initial conditions alone (LoRA removed), the same state Pangu receives. Stars and diamonds mark the lowest Pangu errors in the 5- and 10-day windows, respectively. Percentages above each panel show the best Pangu improvement relative to control for gated (green) and ungated (red) optimization. Ungated outperforms gated on half of dates when stopped early. Gating does not guarantee a better Pangu performance, but it does produce more comparable GraphCast and Pangu-Weather improvements on all 36 cases.
[LG-110] Near-Minimax Sample Complexity for Multi-Objective Reinforcement Learning under Generative Models
链接: https://arxiv.org/abs/2611.05562
作者: Yue Wang
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注: Preprint, work in progress
Abstract:In multi-objective reinforcement learning, preferences may be specified only after data collection, requiring one dataset to support a continuum of reward trade-offs. We study finite discounted Markov decision processes with unknown transitions, S states, A actions, and m known reward functions taking values in [0,1] , given access to a generative model. The learner must provide an \varepsilon -optimal policy for every convex combination of these rewards without further sampling, with a guarantee that holds simultaneously over all preferences and initial states. Writing H=(1-\gamma)^-1 , we establish the constant-confidence minimax sample complexity \widetilde\Theta!\left( \fracSAH^3\varepsilon^2\min\S,\log(em)\ \right) , where the suppressed factors are logarithmic in S,A,H,\varepsilon^-1 and the dependence on m is explicit. Thus, up to these logarithmic factors, the worst-case statistical cost of supporting all preferences grows logarithmically with the number of objectives and saturates at the number of states. The upper bound uses uniform state–action sampling and empirical planning, with reward rounding in the logarithmic-objective regime. Its analysis chains optimal-value increments over covers of the reward hull. A one-sided Bellman variance bound controls signed value differences under a second transition kernel, preserving the H^3\varepsilon^-2 dependence. A complementary uniform guarantee for all bounded rewards yields the state-dimension saturation. We establish lower bounds matching up to logarithmic factors through a rare-exit construction with subset rewards and action encoding. These bounds apply to adaptive simulator queries and arbitrary randomized history-dependent output policies, while the upper bounds are attained by deterministic stationary policies. Comments: Preprint, work in progress Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG) Cite as: arXiv:2611.05562 [stat.ML] (or arXiv:2611.05562v1 [stat.ML] for this version) https://doi.org/10.48550/arXiv.2611.05562 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
[LG-111] A Unified Framework for Training-Dataset Aggregation and Exact Machine Unlearning via Gibbs Probability Measures
链接: https://arxiv.org/abs/2611.05544
作者: Yaiza Bermudez,Samir M. Perlaza,Iñaki Esnaola
类目: Machine Learning (stat.ML); Information Theory (cs.IT); Machine Learning (cs.LG)
*备注: Submitted to the IEEE Transaction on Information Theory
Abstract:In this paper, three operations on Gibbs probability measures are studied: renormalization, normalized log-linear combination, and nesting. The resulting measures are shown to be Gibbs probability measures with explicit expressions for their objective functions, reference measures, and regularization factors. These measures are further characterized as unique solutions to optimization problems involving linear combinations of the original objective functions, subject to either relative entropy regularization or relative entropy constraints. When the reference measures are probability measures, geometric interpretations of these three operations are established involving Pythagorean identities and exponential geodesics.A central result is the equivalence between nesting and normalized log-linear combinations of Gibbs probability measures. These results are then applied to supervised machine learning with Gibbs algorithms, which are represented by Gibbs probability measures. In this context, either through nesting or log-linear combinations, Gibbs algorithms trained on local datasets are used to form new Gibbs algorithms trained upon the aggregation of all local training datasets. Interestingly, the parameters of such operations, determine the weights of each training data point on the aggregated dataset. %Operations on local Gibbs probability measures are shown to control the contributions of training data. This observation is leveraged for achieving: (a) exact machine unlearning in centralized, decentralized, and federated learning systems; and (b) centralized-equivalent performance, in the sense of producing algorithms trained on the aggregation of all local datasets in decentralized and federated learning .
[LG-112] Generalized Hyperrectangular Prediction Sets in Multivariate Conformal Prediction
链接: https://arxiv.org/abs/2611.05543
作者: Nabil Alami,Souhaib Ben Taieb
类目: Machine Learning (stat.ML); Machine Learning (cs.LG)
*备注:
Abstract:In multivariate conformal prediction, many nonconformity scores yield prediction sets with complex geometry (e.g., ellipsoids or density level sets) that are difficult to interpret and visualize. Hyperrectangles offer a practical alternative by providing one interval per output. We extend this notion to Cartesian products of arbitrary one-dimensional sets, which we call generalized hyperrectangles. We introduce a rectangularization framework that turns any continuous multivariate nonconformity score, or conditional density, into one yielding generalized hyperrectangular prediction sets with finite-sample coverage. We establish key geometric properties of our rectangularization framework and provide useful intuitions on generalized hyperrectangles. We further establish a representation theorem showing that a broad class of hyperrectangular conformal methods based on max-aggregated coordinate-wise scores can be interpreted as rectangularizations of implicit conditional densities. Through synthetic and real-world experiments, we quantify the efficiency cost of rectangularization, while highlighting its benefits, especially for practical applications of uncertainty quantification.
[LG-113] ght Lower Bounds under Matched High-Order Hölder Smoothness and Uniform Convexity
链接: https://arxiv.org/abs/2611.05527
作者: Cedar Site Bai
类目: Optimization and Control (math.OC); Machine Learning (cs.LG); Machine Learning (stat.ML)
*备注:
Abstract:In this paper, we provide tight lower bounds for minimizing high-order Hölder-smooth and uniformly convex functions when their orders match. Specifically, we consider functions whose p^th -order derivatives are Hölder continuous with degree \nu and parameter H , and that are uniformly convex with degree q and parameter \sigma . When these orders match, namely, q = p + \nu , we show that any deterministic algorithm with access to derivatives up to order p requires \Omega\left(\left(\fracH\sigma\right)^\frac23(p+\nu)-2\log\frac\Delta_0\epsilon\right) oracle calls to reduce an initial optimality gap of \Delta_0 to \epsilon . This matches the corresponding upper bound and establishes a higher-order uniformly convex counterpart to Nesterov’s classical first-order lower bound for strongly convex functions. Rather than directly generalizing Nesterov’s classical quadratic-chain construction, we develop a multiscale phase-wise construction that preserves previous oracle responses while maintaining global Hölder smoothness and uniform convexity. Our result resolves the matched case left open by existing lower bounds for the two asymmetric regimes.
[LG-114] Classical training of conditional instantaneous quantum polynomial-time Born machines
链接: https://arxiv.org/abs/2611.05483
作者: Oriol Balló-Gimbernat,Emiliano Tolotti,Laia Domingo,Fernando Vilariño,Adan Garriga
类目: Quantum Physics (quant-ph); Machine Learning (cs.LG)
*备注: 25 pages, 9 figures
Abstract:Instantaneous quantum polynomial-time (IQP) Born machines are attractive quantum generative models, as certain IQP circuit families are conjectured to be hard to sample, yet they can be trained classically. Previously proposed conditional IQP Born machines make the parameters of the diagonal layer condition dependent. We introduce conditional IQP Born machines, in which the condition is encoded in the input state and a single IQP unitary is shared across conditions. Efficient sampling and amplitude-ratio evaluation for the reference state entering the diagonal layer suffice to obtain unbiased classical estimates of a conditional maximum mean discrepancy objective and its gradients. We show that input conditioning can represent conditional families that no diagonally conditioned IQP model with the same number of qubits can reproduce. For the sixteen-qubit Curie-Weiss spin model, a single conditional model captures the magnetization transition at held-out temperatures. Its mean total variation distance is six times lower than that of diagonal conditioning under the same protocol and comparable to that of classical generative models with up to seven times as many parameters. We also train conditional models classically with up to 2048 qubits, which reproduce the mean squared magnetization of the target away from the crossover. These results extend classically trained IQP Born machines to conditional generation.
[LG-115] HLA-Net: A Physics-Guided Skeleton-Based Framework for Automatic Gravitational Arc Detection in Astronomical Images
链接: https://arxiv.org/abs/2611.05438
作者: Don S
类目: Instrumentation and Methods for Astrophysics (astro-ph.IM); Astrophysics of Galaxies (astro-ph.GA); Machine Learning (cs.LG)
*备注: 16 pages, 2 figures, planning to submit: The Open Journal of Astrophysics
Abstract:Gravitational arcs produced by strong gravitational lensing provide a unique approach for investigating dark matter distribution, galaxy cluster mass reconstruction and the structure of the Universe. Automated detection remains challenging because gravitational arcs are rare, often faint and morphologically similar to elongated this http URL this paper, we propose \textbfHLA-Net (Hybrid Lens Arc Network), an interpretable multi-stage framework for automatic gravitational arc detection that explicitly exploits both the geometric morphology and physical characteristics of gravitational lensing without requiring manually annotated training this http URL proposed framework was evaluated on astronomical FITS images containing complex galaxy cluster structures. Starting from 169 initial arc candidates extracted from ordered skeleton representations, the multi-stage verification pipeline reduced the search space to 83 geometrically consistent candidates, while robust circle fitting successfully modelled 113 candidate arc structures. The results demonstrate that HLA-Net effectively suppresses false-positive detections while preserving physically reasonable lens candidates through an interpretable and modular verification strategy.
[LG-116] BIOTRAM: Biochemistry-informed Transformer for Relative Abundance via Metabarcoding
链接: https://arxiv.org/abs/2611.05418
作者: Emma Boehly,Scott C. Lowe,Ken A. Thompson,Anthony Fuller,Will Armstrong,Iuliia Zarubiieva,Neelu Madan,Tiancheng Gao,Sameer Padhye,Jayme Sones,Kate H. J. Perez,Bess Hardwick,Brendan Furneaux,Evgeny V. Zakharov,Dirk Steinke,Jeremy R. deWaard,Stephanie L. deWaard,Jaclyn McKeown,Norman Monkhouse,Cole Pace,Sujeevan Ratnasingham,Raina Schmuck,Amy Thompson,Jorge L. León-Cortés,Luciano Cagnolo,Julieta Soledad Arena,Jeremías Domínguez Masciale,Errol Douwes,Elli Groner,Masae Iwamoto Ishihara,Shunsuke Matsuoka,Mailyn A. Gonzalez,Alexandre Antonelli,Renata S. Sousa-Lima,Armin Niessner,Patrícia Guedes,Elizabeth G. Biro,Zuania Colón-Piñeiro,Dezene P. W. Huber,Santosh Mazumdar,Magda Bou Dagher Kharrat,Eddie P. Mondejar,Pablo Luis Tubaro,Eugene Marais,Marie-Eugénie Maggia,Darío Alejandro Lijtmaer,Sean W. J. Prosser,Andrea Murillo-Vázquez,Paola Pulido-Santacruz,Paula Leitman,Thiago Arnt Dorigo,Nao Ito,Paul Knoops,Maria L. Kuzmina,Hikaru Morishita,Keishi Kimoto,Dimby Raharinjanahary,Rhian J. Smith,Solny A. Adalsteinsson,Kevin G. Borja-Acosta,Paul Amouroux,Angela M. Mendoza-Henao,Angélica Diaz-Pulido,Anamaria Dal Molin,Luke L. Powell,Diego Inclan,Alice Calvente,David Dunson,Brian Fisher,Tomas Roslin,Otso Ovaskainen,Paul D. N. Hebert,Graham W. Taylor
类目: Genomics (q-bio.GN); Machine Learning (cs.LG)
*备注:
Abstract:Arthropod monitoring at scale begins with metabarcoding of bulk samples: trap collections of hundreds to thousands of specimens processed as one unit, extraction of DNA from the bulked specimens, then sequencing of the DNA from all specimens together. Metabarcoding reveals which taxa such a sample contains, but due to factors such as differing DNA amplification rates and body sizes, its read counts do not neatly correspond to how abundant the different taxa are relative to one another. This lack of relative abundance limits the data’s utility for the multitude of ecological analyses that depend on it. We propose BIOTRAM: a neural network that estimates per-taxon specimen counts by dividing the observed read counts by a learned read yield per specimen, which it factors into the DNA each specimen contributes and how efficiently that DNA is amplified. On a paired dataset of 1,188 bottles (800 k specimens) processed both by metabarcoding and by comprehensive individual barcoding for ground-truthing, BIOTRAM reduces the Jensen-Shannon divergence between the estimated and the true composition from 0.3586 for raw read proportions (the usual proxy for abundance) to 0.1200. Trained on the same objective and inputs as four generic deep architectures, all with tuned loss weights, it reaches a count error 3.2% below that of the strongest of them and the lowest composition divergence. It also generalizes to out-of-domain data better, beating all baselines on bottles from unseen regions, later collection dates and unseen sequencing batches, and transferring to data from an unseen laboratory. Structuring the mapping from reads to counts around the steps of the experimental pipeline thus yields more accurate abundance estimates than leaving that mapping unconstrained.
附件下载
